424 files changed, 83624 insertions, 20271 deletions

diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index b16c50ee769c..a7c54c843291 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -139,6 +139,7 @@ config DRM_I915
 	select BACKLIGHT_CLASS_DEVICE if ACPI
 	select VIDEO_OUTPUT_CONTROL if ACPI
 	select INPUT if ACPI
+	select THERMAL if ACPI
 	select ACPI_VIDEO if ACPI
 	select ACPI_BUTTON if ACPI
 	help
@@ -213,6 +214,8 @@ source "drivers/gpu/drm/mgag200/Kconfig"
 
 source "drivers/gpu/drm/cirrus/Kconfig"
 
+source "drivers/gpu/drm/rcar-du/Kconfig"
+
 source "drivers/gpu/drm/shmobile/Kconfig"
 
 source "drivers/gpu/drm/omapdrm/Kconfig"
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 1c9f24396002..801bcafa3028 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -12,7 +12,8 @@ drm-y       :=	drm_auth.o drm_buffer.o drm_bufs.o drm_cache.o \
 		drm_platform.o drm_sysfs.o drm_hashtab.o drm_mm.o \
 		drm_crtc.o drm_modes.o drm_edid.o \
 		drm_info.o drm_debugfs.o drm_encoder_slave.o \
-		drm_trace_points.o drm_global.o drm_prime.o
+		drm_trace_points.o drm_global.o drm_prime.o \
+		drm_rect.o
 
 drm-$(CONFIG_COMPAT) += drm_ioc32.o
 drm-$(CONFIG_DRM_GEM_CMA_HELPER) += drm_gem_cma_helper.o
@@ -48,6 +49,7 @@ obj-$(CONFIG_DRM_EXYNOS) +=exynos/
 obj-$(CONFIG_DRM_GMA500) += gma500/
 obj-$(CONFIG_DRM_UDL) += udl/
 obj-$(CONFIG_DRM_AST) += ast/
+obj-$(CONFIG_DRM_RCAR_DU) += rcar-du/
 obj-$(CONFIG_DRM_SHMOBILE) +=shmobile/
 obj-$(CONFIG_DRM_OMAP)	+= omapdrm/
 obj-$(CONFIG_DRM_TILCDC)	+= tilcdc/
diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h
index 02e52d543e4b..622d4ae7eb9e 100644
--- a/drivers/gpu/drm/ast/ast_drv.h
+++ b/drivers/gpu/drm/ast/ast_drv.h
@@ -348,8 +348,24 @@ int ast_gem_create(struct drm_device *dev,
 int ast_bo_pin(struct ast_bo *bo, u32 pl_flag, u64 *gpu_addr);
 int ast_bo_unpin(struct ast_bo *bo);
 
-int ast_bo_reserve(struct ast_bo *bo, bool no_wait);
-void ast_bo_unreserve(struct ast_bo *bo);
+static inline int ast_bo_reserve(struct ast_bo *bo, bool no_wait)
+{
+	int ret;
+
+	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
+	if (ret) {
+		if (ret != -ERESTARTSYS && ret != -EBUSY)
+			DRM_ERROR("reserve failed %p\n", bo);
+		return ret;
+	}
+	return 0;
+}
+
+static inline void ast_bo_unreserve(struct ast_bo *bo)
+{
+	ttm_bo_unreserve(&bo->bo);
+}
+
 void ast_ttm_placement(struct ast_bo *bo, int domain);
 int ast_bo_push_sysram(struct ast_bo *bo);
 int ast_mmap(struct file *filp, struct vm_area_struct *vma);
diff --git a/drivers/gpu/drm/ast/ast_fb.c b/drivers/gpu/drm/ast/ast_fb.c
index fbc0823cfa18..7b33e14e44aa 100644
--- a/drivers/gpu/drm/ast/ast_fb.c
+++ b/drivers/gpu/drm/ast/ast_fb.c
@@ -51,7 +51,7 @@ static void ast_dirty_update(struct ast_fbdev *afbdev,
 	struct ast_bo *bo;
 	int src_offset, dst_offset;
 	int bpp = (afbdev->afb.base.bits_per_pixel + 7)/8;
-	int ret;
+	int ret = -EBUSY;
 	bool unmap = false;
 	bool store_for_later = false;
 	int x2, y2;
@@ -65,7 +65,8 @@ static void ast_dirty_update(struct ast_fbdev *afbdev,
 	 * then the BO is being moved and we should
 	 * store up the damage until later.
 	 */
-	ret = ast_bo_reserve(bo, true);
+	if (!in_interrupt())
+		ret = ast_bo_reserve(bo, true);
 	if (ret) {
 		if (ret != -EBUSY)
 			return;
diff --git a/drivers/gpu/drm/ast/ast_ttm.c b/drivers/gpu/drm/ast/ast_ttm.c
index 09da3393c527..98d670825a1a 100644
--- a/drivers/gpu/drm/ast/ast_ttm.c
+++ b/drivers/gpu/drm/ast/ast_ttm.c
@@ -271,26 +271,19 @@ int ast_mm_init(struct ast_private *ast)
 		return ret;
 	}
 
-	ast->fb_mtrr = drm_mtrr_add(pci_resource_start(dev->pdev, 0),
-				    pci_resource_len(dev->pdev, 0),
-				    DRM_MTRR_WC);
+	ast->fb_mtrr = arch_phys_wc_add(pci_resource_start(dev->pdev, 0),
+					pci_resource_len(dev->pdev, 0));
 
 	return 0;
 }
 
 void ast_mm_fini(struct ast_private *ast)
 {
-	struct drm_device *dev = ast->dev;
 	ttm_bo_device_release(&ast->ttm.bdev);
 
 	ast_ttm_global_release(ast);
 
-	if (ast->fb_mtrr >= 0) {
-		drm_mtrr_del(ast->fb_mtrr,
-			     pci_resource_start(dev->pdev, 0),
-			     pci_resource_len(dev->pdev, 0), DRM_MTRR_WC);
-		ast->fb_mtrr = -1;
-	}
+	arch_phys_wc_del(ast->fb_mtrr);
 }
 
 void ast_ttm_placement(struct ast_bo *bo, int domain)
@@ -310,24 +303,6 @@ void ast_ttm_placement(struct ast_bo *bo, int domain)
 	bo->placement.num_busy_placement = c;
 }
 
-int ast_bo_reserve(struct ast_bo *bo, bool no_wait)
-{
-	int ret;
-
-	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
-	if (ret) {
-		if (ret != -ERESTARTSYS && ret != -EBUSY)
-			DRM_ERROR("reserve failed %p\n", bo);
-		return ret;
-	}
-	return 0;
-}
-
-void ast_bo_unreserve(struct ast_bo *bo)
-{
-	ttm_bo_unreserve(&bo->bo);
-}
-
 int ast_bo_create(struct drm_device *dev, int size, int align,
 		  uint32_t flags, struct ast_bo **pastbo)
 {
diff --git a/drivers/gpu/drm/cirrus/cirrus_drv.h b/drivers/gpu/drm/cirrus/cirrus_drv.h
index 7ca059596887..bae55609e6c3 100644
--- a/drivers/gpu/drm/cirrus/cirrus_drv.h
+++ b/drivers/gpu/drm/cirrus/cirrus_drv.h
@@ -240,8 +240,25 @@ void cirrus_ttm_placement(struct cirrus_bo *bo, int domain);
 int cirrus_bo_create(struct drm_device *dev, int size, int align,
 		     uint32_t flags, struct cirrus_bo **pcirrusbo);
 int cirrus_mmap(struct file *filp, struct vm_area_struct *vma);
-int cirrus_bo_reserve(struct cirrus_bo *bo, bool no_wait);
-void cirrus_bo_unreserve(struct cirrus_bo *bo);
+
+static inline int cirrus_bo_reserve(struct cirrus_bo *bo, bool no_wait)
+{
+	int ret;
+
+	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
+	if (ret) {
+		if (ret != -ERESTARTSYS && ret != -EBUSY)
+			DRM_ERROR("reserve failed %p\n", bo);
+		return ret;
+	}
+	return 0;
+}
+
+static inline void cirrus_bo_unreserve(struct cirrus_bo *bo)
+{
+	ttm_bo_unreserve(&bo->bo);
+}
+
 int cirrus_bo_push_sysram(struct cirrus_bo *bo);
 int cirrus_bo_pin(struct cirrus_bo *bo, u32 pl_flag, u64 *gpu_addr);
 #endif				/* __CIRRUS_DRV_H__ */
diff --git a/drivers/gpu/drm/cirrus/cirrus_fbdev.c b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
index 3541b567bbd8..b27e95666fab 100644
--- a/drivers/gpu/drm/cirrus/cirrus_fbdev.c
+++ b/drivers/gpu/drm/cirrus/cirrus_fbdev.c
@@ -25,7 +25,7 @@ static void cirrus_dirty_update(struct cirrus_fbdev *afbdev,
 	struct cirrus_bo *bo;
 	int src_offset, dst_offset;
 	int bpp = (afbdev->gfb.base.bits_per_pixel + 7)/8;
-	int ret;
+	int ret = -EBUSY;
 	bool unmap = false;
 	bool store_for_later = false;
 	int x2, y2;
@@ -39,7 +39,8 @@ static void cirrus_dirty_update(struct cirrus_fbdev *afbdev,
 	 * then the BO is being moved and we should
 	 * store up the damage until later.
 	 */
-	ret = cirrus_bo_reserve(bo, true);
+	if (!in_interrupt())
+		ret = cirrus_bo_reserve(bo, true);
 	if (ret) {
 		if (ret != -EBUSY)
 			return;
diff --git a/drivers/gpu/drm/cirrus/cirrus_ttm.c b/drivers/gpu/drm/cirrus/cirrus_ttm.c
index 2ed8cfc740c9..0047012045c2 100644
--- a/drivers/gpu/drm/cirrus/cirrus_ttm.c
+++ b/drivers/gpu/drm/cirrus/cirrus_ttm.c
@@ -271,9 +271,8 @@ int cirrus_mm_init(struct cirrus_device *cirrus)
 		return ret;
 	}
 
-	cirrus->fb_mtrr = drm_mtrr_add(pci_resource_start(dev->pdev, 0),
-				    pci_resource_len(dev->pdev, 0),
-				    DRM_MTRR_WC);
+	cirrus->fb_mtrr = arch_phys_wc_add(pci_resource_start(dev->pdev, 0),
+					   pci_resource_len(dev->pdev, 0));
 
 	cirrus->mm_inited = true;
 	return 0;
@@ -281,8 +280,6 @@ int cirrus_mm_init(struct cirrus_device *cirrus)
 
 void cirrus_mm_fini(struct cirrus_device *cirrus)
 {
-	struct drm_device *dev = cirrus->dev;
-
 	if (!cirrus->mm_inited)
 		return;
 
@@ -290,12 +287,8 @@ void cirrus_mm_fini(struct cirrus_device *cirrus)
 
 	cirrus_ttm_global_release(cirrus);
 
-	if (cirrus->fb_mtrr >= 0) {
-		drm_mtrr_del(cirrus->fb_mtrr,
-			     pci_resource_start(dev->pdev, 0),
-			     pci_resource_len(dev->pdev, 0), DRM_MTRR_WC);
-		cirrus->fb_mtrr = -1;
-	}
+	arch_phys_wc_del(cirrus->fb_mtrr);
+	cirrus->fb_mtrr = 0;
 }
 
 void cirrus_ttm_placement(struct cirrus_bo *bo, int domain)
@@ -315,24 +308,6 @@ void cirrus_ttm_placement(struct cirrus_bo *bo, int domain)
 	bo->placement.num_busy_placement = c;
 }
 
-int cirrus_bo_reserve(struct cirrus_bo *bo, bool no_wait)
-{
-	int ret;
-
-	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
-	if (ret) {
-		if (ret != -ERESTARTSYS && ret != -EBUSY)
-			DRM_ERROR("reserve failed %p\n", bo);
-		return ret;
-	}
-	return 0;
-}
-
-void cirrus_bo_unreserve(struct cirrus_bo *bo)
-{
-	ttm_bo_unreserve(&bo->bo);
-}
-
 int cirrus_bo_create(struct drm_device *dev, int size, int align,
 		  uint32_t flags, struct cirrus_bo **pcirrusbo)
 {
diff --git a/drivers/gpu/drm/drm_bufs.c b/drivers/gpu/drm/drm_bufs.c
index 0128147265f3..5a4dbb410b71 100644
--- a/drivers/gpu/drm/drm_bufs.c
+++ b/drivers/gpu/drm/drm_bufs.c
@@ -210,12 +210,16 @@ static int drm_addmap_core(struct drm_device * dev, resource_size_t offset,
 		if (drm_core_has_MTRR(dev)) {
 			if (map->type == _DRM_FRAME_BUFFER ||
 			    (map->flags & _DRM_WRITE_COMBINING)) {
-				map->mtrr = mtrr_add(map->offset, map->size,
-						     MTRR_TYPE_WRCOMB, 1);
+				map->mtrr =
+					arch_phys_wc_add(map->offset, map->size);
 			}
 		}
 		if (map->type == _DRM_REGISTERS) {
-			map->handle = ioremap(map->offset, map->size);
+			if (map->flags & _DRM_WRITE_COMBINING)
+				map->handle = ioremap_wc(map->offset,
+							 map->size);
+			else
+				map->handle = ioremap(map->offset, map->size);
 			if (!map->handle) {
 				kfree(map);
 				return -ENOMEM;
@@ -410,6 +414,15 @@ int drm_addmap_ioctl(struct drm_device *dev, void *data,
 
 	/* avoid a warning on 64-bit, this casting isn't very nice, but the API is set so too late */
 	map->handle = (void *)(unsigned long)maplist->user_token;
+
+	/*
+	 * It appears that there are no users of this value whatsoever --
+	 * drmAddMap just discards it.  Let's not encourage its use.
+	 * (Keeping drm_addmap_core's returned mtrr value would be wrong --
+	 *  it's not a real mtrr index anymore.)
+	 */
+	map->mtrr = -1;
+
 	return 0;
 }
 
@@ -451,11 +464,8 @@ int drm_rmmap_locked(struct drm_device *dev, struct drm_local_map *map)
 		iounmap(map->handle);
 		/* FALLTHROUGH */
 	case _DRM_FRAME_BUFFER:
-		if (drm_core_has_MTRR(dev) && map->mtrr >= 0) {
-			int retcode;
-			retcode = mtrr_del(map->mtrr, map->offset, map->size);
-			DRM_DEBUG("mtrr_del=%d\n", retcode);
-		}
+		if (drm_core_has_MTRR(dev))
+			arch_phys_wc_del(map->mtrr);
 		break;
 	case _DRM_SHM:
 		vfree(map->handle);
diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index e7e92429d10f..fc83bb9eb514 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -29,6 +29,7 @@
  *      Dave Airlie <airlied@linux.ie>
  *      Jesse Barnes <jesse.barnes@intel.com>
  */
+#include <linux/ctype.h>
 #include <linux/list.h>
 #include <linux/slab.h>
 #include <linux/export.h>
@@ -91,7 +92,7 @@ EXPORT_SYMBOL(drm_warn_on_modeset_not_all_locked);
 
 /* Avoid boilerplate.  I'm tired of typing. */
 #define DRM_ENUM_NAME_FN(fnname, list)				\
-	char *fnname(int val)					\
+	const char *fnname(int val)				\
 	{							\
 		int i;						\
 		for (i = 0; i < ARRAY_SIZE(list); i++) {	\
@@ -104,7 +105,7 @@ EXPORT_SYMBOL(drm_warn_on_modeset_not_all_locked);
 /*
  * Global properties
  */
-static struct drm_prop_enum_list drm_dpms_enum_list[] =
+static const struct drm_prop_enum_list drm_dpms_enum_list[] =
 {	{ DRM_MODE_DPMS_ON, "On" },
 	{ DRM_MODE_DPMS_STANDBY, "Standby" },
 	{ DRM_MODE_DPMS_SUSPEND, "Suspend" },
@@ -116,7 +117,7 @@ DRM_ENUM_NAME_FN(drm_get_dpms_name, drm_dpms_enum_list)
 /*
  * Optional properties
  */
-static struct drm_prop_enum_list drm_scaling_mode_enum_list[] =
+static const struct drm_prop_enum_list drm_scaling_mode_enum_list[] =
 {
 	{ DRM_MODE_SCALE_NONE, "None" },
 	{ DRM_MODE_SCALE_FULLSCREEN, "Full" },
@@ -124,7 +125,7 @@ static struct drm_prop_enum_list drm_scaling_mode_enum_list[] =
 	{ DRM_MODE_SCALE_ASPECT, "Full aspect" },
 };
 
-static struct drm_prop_enum_list drm_dithering_mode_enum_list[] =
+static const struct drm_prop_enum_list drm_dithering_mode_enum_list[] =
 {
 	{ DRM_MODE_DITHERING_OFF, "Off" },
 	{ DRM_MODE_DITHERING_ON, "On" },
@@ -134,7 +135,7 @@ static struct drm_prop_enum_list drm_dithering_mode_enum_list[] =
 /*
  * Non-global properties, but "required" for certain connectors.
  */
-static struct drm_prop_enum_list drm_dvi_i_select_enum_list[] =
+static const struct drm_prop_enum_list drm_dvi_i_select_enum_list[] =
 {
 	{ DRM_MODE_SUBCONNECTOR_Automatic, "Automatic" }, /* DVI-I and TV-out */
 	{ DRM_MODE_SUBCONNECTOR_DVID,      "DVI-D"     }, /* DVI-I  */
@@ -143,7 +144,7 @@ static struct drm_prop_enum_list drm_dvi_i_select_enum_list[] =
 
 DRM_ENUM_NAME_FN(drm_get_dvi_i_select_name, drm_dvi_i_select_enum_list)
 
-static struct drm_prop_enum_list drm_dvi_i_subconnector_enum_list[] =
+static const struct drm_prop_enum_list drm_dvi_i_subconnector_enum_list[] =
 {
 	{ DRM_MODE_SUBCONNECTOR_Unknown,   "Unknown"   }, /* DVI-I and TV-out */
 	{ DRM_MODE_SUBCONNECTOR_DVID,      "DVI-D"     }, /* DVI-I  */
@@ -153,7 +154,7 @@ static struct drm_prop_enum_list drm_dvi_i_subconnector_enum_list[] =
 DRM_ENUM_NAME_FN(drm_get_dvi_i_subconnector_name,
 		 drm_dvi_i_subconnector_enum_list)
 
-static struct drm_prop_enum_list drm_tv_select_enum_list[] =
+static const struct drm_prop_enum_list drm_tv_select_enum_list[] =
 {
 	{ DRM_MODE_SUBCONNECTOR_Automatic, "Automatic" }, /* DVI-I and TV-out */
 	{ DRM_MODE_SUBCONNECTOR_Composite, "Composite" }, /* TV-out */
@@ -164,7 +165,7 @@ static struct drm_prop_enum_list drm_tv_select_enum_list[] =
 
 DRM_ENUM_NAME_FN(drm_get_tv_select_name, drm_tv_select_enum_list)
 
-static struct drm_prop_enum_list drm_tv_subconnector_enum_list[] =
+static const struct drm_prop_enum_list drm_tv_subconnector_enum_list[] =
 {
 	{ DRM_MODE_SUBCONNECTOR_Unknown,   "Unknown"   }, /* DVI-I and TV-out */
 	{ DRM_MODE_SUBCONNECTOR_Composite, "Composite" }, /* TV-out */
@@ -176,7 +177,7 @@ static struct drm_prop_enum_list drm_tv_subconnector_enum_list[] =
 DRM_ENUM_NAME_FN(drm_get_tv_subconnector_name,
 		 drm_tv_subconnector_enum_list)
 
-static struct drm_prop_enum_list drm_dirty_info_enum_list[] = {
+static const struct drm_prop_enum_list drm_dirty_info_enum_list[] = {
 	{ DRM_MODE_DIRTY_OFF,      "Off"      },
 	{ DRM_MODE_DIRTY_ON,       "On"       },
 	{ DRM_MODE_DIRTY_ANNOTATE, "Annotate" },
@@ -184,7 +185,7 @@ static struct drm_prop_enum_list drm_dirty_info_enum_list[] = {
 
 struct drm_conn_prop_enum_list {
 	int type;
-	char *name;
+	const char *name;
 	int count;
 };
 
@@ -210,7 +211,7 @@ static struct drm_conn_prop_enum_list drm_connector_enum_list[] =
 	{ DRM_MODE_CONNECTOR_VIRTUAL, "Virtual", 0},
 };
 
-static struct drm_prop_enum_list drm_encoder_enum_list[] =
+static const struct drm_prop_enum_list drm_encoder_enum_list[] =
 {	{ DRM_MODE_ENCODER_NONE, "None" },
 	{ DRM_MODE_ENCODER_DAC, "DAC" },
 	{ DRM_MODE_ENCODER_TMDS, "TMDS" },
@@ -219,7 +220,7 @@ static struct drm_prop_enum_list drm_encoder_enum_list[] =
 	{ DRM_MODE_ENCODER_VIRTUAL, "Virtual" },
 };
 
-char *drm_get_encoder_name(struct drm_encoder *encoder)
+const char *drm_get_encoder_name(const struct drm_encoder *encoder)
 {
 	static char buf[32];
 
@@ -230,7 +231,7 @@ char *drm_get_encoder_name(struct drm_encoder *encoder)
 }
 EXPORT_SYMBOL(drm_get_encoder_name);
 
-char *drm_get_connector_name(struct drm_connector *connector)
+const char *drm_get_connector_name(const struct drm_connector *connector)
 {
 	static char buf[32];
 
@@ -241,7 +242,7 @@ char *drm_get_connector_name(struct drm_connector *connector)
 }
 EXPORT_SYMBOL(drm_get_connector_name);
 
-char *drm_get_connector_status_name(enum drm_connector_status status)
+const char *drm_get_connector_status_name(enum drm_connector_status status)
 {
 	if (status == connector_status_connected)
 		return "connected";
@@ -252,6 +253,28 @@ char *drm_get_connector_status_name(enum drm_connector_status status)
 }
 EXPORT_SYMBOL(drm_get_connector_status_name);
 
+static char printable_char(int c)
+{
+	return isascii(c) && isprint(c) ? c : '?';
+}
+
+const char *drm_get_format_name(uint32_t format)
+{
+	static char buf[32];
+
+	snprintf(buf, sizeof(buf),
+		 "%c%c%c%c %s-endian (0x%08x)",
+		 printable_char(format & 0xff),
+		 printable_char((format >> 8) & 0xff),
+		 printable_char((format >> 16) & 0xff),
+		 printable_char((format >> 24) & 0x7f),
+		 format & DRM_FORMAT_BIG_ENDIAN ? "big" : "little",
+		 format);
+
+	return buf;
+}
+EXPORT_SYMBOL(drm_get_format_name);
+
 /**
  * drm_mode_object_get - allocate a new modeset identifier
  * @dev: DRM device
@@ -569,16 +592,8 @@ void drm_framebuffer_remove(struct drm_framebuffer *fb)
 		}
 
 		list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
-			if (plane->fb == fb) {
-				/* should turn off the crtc */
-				ret = plane->funcs->disable_plane(plane);
-				if (ret)
-					DRM_ERROR("failed to disable plane with busy fb\n");
-				/* disconnect the plane from the fb and crtc: */
-				__drm_framebuffer_unreference(plane->fb);
-				plane->fb = NULL;
-				plane->crtc = NULL;
-			}
+			if (plane->fb == fb)
+				drm_plane_force_disable(plane);
 		}
 		drm_modeset_unlock_all(dev);
 	}
@@ -593,7 +608,7 @@ EXPORT_SYMBOL(drm_framebuffer_remove);
  * @crtc: CRTC object to init
  * @funcs: callbacks for the new CRTC
  *
- * Inits a new object created as base part of an driver crtc object.
+ * Inits a new object created as base part of a driver crtc object.
  *
  * RETURNS:
  * Zero on success, error code on failure.
@@ -628,11 +643,12 @@ int drm_crtc_init(struct drm_device *dev, struct drm_crtc *crtc,
 EXPORT_SYMBOL(drm_crtc_init);
 
 /**
- * drm_crtc_cleanup - Cleans up the core crtc usage.
+ * drm_crtc_cleanup - Clean up the core crtc usage
  * @crtc: CRTC to cleanup
  *
- * Cleanup @crtc. Removes from drm modesetting space
- * does NOT free object, caller does that.
+ * This function cleans up @crtc and removes it from the DRM mode setting
+ * core. Note that the function does *not* free the crtc structure itself,
+ * this is the responsibility of the caller.
  */
 void drm_crtc_cleanup(struct drm_crtc *crtc)
 {
@@ -657,7 +673,7 @@ EXPORT_SYMBOL(drm_crtc_cleanup);
 void drm_mode_probed_add(struct drm_connector *connector,
 			 struct drm_display_mode *mode)
 {
-	list_add(&mode->head, &connector->probed_modes);
+	list_add_tail(&mode->head, &connector->probed_modes);
 }
 EXPORT_SYMBOL(drm_mode_probed_add);
 
@@ -803,6 +819,21 @@ void drm_encoder_cleanup(struct drm_encoder *encoder)
 }
 EXPORT_SYMBOL(drm_encoder_cleanup);
 
+/**
+ * drm_plane_init - Initialise a new plane object
+ * @dev: DRM device
+ * @plane: plane object to init
+ * @possible_crtcs: bitmask of possible CRTCs
+ * @funcs: callbacks for the new plane
+ * @formats: array of supported formats (%DRM_FORMAT_*)
+ * @format_count: number of elements in @formats
+ * @priv: plane is private (hidden from userspace)?
+ *
+ * Inits a new object created as base part of a driver plane object.
+ *
+ * RETURNS:
+ * Zero on success, error code on failure.
+ */
 int drm_plane_init(struct drm_device *dev, struct drm_plane *plane,
 		   unsigned long possible_crtcs,
 		   const struct drm_plane_funcs *funcs,
@@ -851,6 +882,14 @@ int drm_plane_init(struct drm_device *dev, struct drm_plane *plane,
 }
 EXPORT_SYMBOL(drm_plane_init);
 
+/**
+ * drm_plane_cleanup - Clean up the core plane usage
+ * @plane: plane to cleanup
+ *
+ * This function cleans up @plane and removes it from the DRM mode setting
+ * core. Note that the function does *not* free the plane structure itself,
+ * this is the responsibility of the caller.
+ */
 void drm_plane_cleanup(struct drm_plane *plane)
 {
 	struct drm_device *dev = plane->dev;
@@ -868,6 +907,32 @@ void drm_plane_cleanup(struct drm_plane *plane)
 EXPORT_SYMBOL(drm_plane_cleanup);
 
 /**
+ * drm_plane_force_disable - Forcibly disable a plane
+ * @plane: plane to disable
+ *
+ * Forces the plane to be disabled.
+ *
+ * Used when the plane's current framebuffer is destroyed,
+ * and when restoring fbdev mode.
+ */
+void drm_plane_force_disable(struct drm_plane *plane)
+{
+	int ret;
+
+	if (!plane->fb)
+		return;
+
+	ret = plane->funcs->disable_plane(plane);
+	if (ret)
+		DRM_ERROR("failed to disable plane with busy fb\n");
+	/* disconnect the plane from the fb and crtc: */
+	__drm_framebuffer_unreference(plane->fb);
+	plane->fb = NULL;
+	plane->crtc = NULL;
+}
+EXPORT_SYMBOL(drm_plane_force_disable);
+
+/**
  * drm_mode_create - create a new display mode
  * @dev: DRM device
  *
@@ -1740,7 +1805,7 @@ int drm_mode_getplane(struct drm_device *dev, void *data,
 
 	plane_resp->plane_id = plane->base.id;
 	plane_resp->possible_crtcs = plane->possible_crtcs;
-	plane_resp->gamma_size = plane->gamma_size;
+	plane_resp->gamma_size = 0;
 
 	/*
 	 * This ioctl is called twice, once to determine how much space is
@@ -1834,7 +1899,8 @@ int drm_mode_setplane(struct drm_device *dev, void *data,
 		if (fb->pixel_format == plane->format_types[i])
 			break;
 	if (i == plane->format_count) {
-		DRM_DEBUG_KMS("Invalid pixel format 0x%08x\n", fb->pixel_format);
+		DRM_DEBUG_KMS("Invalid pixel format %s\n",
+			      drm_get_format_name(fb->pixel_format));
 		ret = -EINVAL;
 		goto out;
 	}
@@ -1906,18 +1972,31 @@ out:
 int drm_mode_set_config_internal(struct drm_mode_set *set)
 {
 	struct drm_crtc *crtc = set->crtc;
-	struct drm_framebuffer *fb, *old_fb;
+	struct drm_framebuffer *fb;
+	struct drm_crtc *tmp;
 	int ret;
 
-	old_fb = crtc->fb;
+	/*
+	 * NOTE: ->set_config can also disable other crtcs (if we steal all
+	 * connectors from it), hence we need to refcount the fbs across all
+	 * crtcs. Atomic modeset will have saner semantics ...
+	 */
+	list_for_each_entry(tmp, &crtc->dev->mode_config.crtc_list, head)
+		tmp->old_fb = tmp->fb;
+
 	fb = set->fb;
 
 	ret = crtc->funcs->set_config(set);
 	if (ret == 0) {
-		if (old_fb)
-			drm_framebuffer_unreference(old_fb);
-		if (fb)
-			drm_framebuffer_reference(fb);
+		/* crtc->fb must be updated by ->set_config, enforces this. */
+		WARN_ON(fb != crtc->fb);
+	}
+
+	list_for_each_entry(tmp, &crtc->dev->mode_config.crtc_list, head) {
+		if (tmp->fb)
+			drm_framebuffer_reference(tmp->fb);
+		if (tmp->old_fb)
+			drm_framebuffer_unreference(tmp->old_fb);
 	}
 
 	return ret;
@@ -2099,10 +2178,10 @@ out:
 	return ret;
 }
 
-int drm_mode_cursor_ioctl(struct drm_device *dev,
-			void *data, struct drm_file *file_priv)
+static int drm_mode_cursor_common(struct drm_device *dev,
+				  struct drm_mode_cursor2 *req,
+				  struct drm_file *file_priv)
 {
-	struct drm_mode_cursor *req = data;
 	struct drm_mode_object *obj;
 	struct drm_crtc *crtc;
 	int ret = 0;
@@ -2122,13 +2201,17 @@ int drm_mode_cursor_ioctl(struct drm_device *dev,
 
 	mutex_lock(&crtc->mutex);
 	if (req->flags & DRM_MODE_CURSOR_BO) {
-		if (!crtc->funcs->cursor_set) {
+		if (!crtc->funcs->cursor_set && !crtc->funcs->cursor_set2) {
 			ret = -ENXIO;
 			goto out;
 		}
 		/* Turns off the cursor if handle is 0 */
-		ret = crtc->funcs->cursor_set(crtc, file_priv, req->handle,
-					      req->width, req->height);
+		if (crtc->funcs->cursor_set2)
+			ret = crtc->funcs->cursor_set2(crtc, file_priv, req->handle,
+						      req->width, req->height, req->hot_x, req->hot_y);
+		else
+			ret = crtc->funcs->cursor_set(crtc, file_priv, req->handle,
+						      req->width, req->height);
 	}
 
 	if (req->flags & DRM_MODE_CURSOR_MOVE) {
@@ -2143,6 +2226,25 @@ out:
 	mutex_unlock(&crtc->mutex);
 
 	return ret;
+
+}
+int drm_mode_cursor_ioctl(struct drm_device *dev,
+			void *data, struct drm_file *file_priv)
+{
+	struct drm_mode_cursor *req = data;
+	struct drm_mode_cursor2 new_req;
+
+	memcpy(&new_req, req, sizeof(struct drm_mode_cursor));
+	new_req.hot_x = new_req.hot_y = 0;
+
+	return drm_mode_cursor_common(dev, &new_req, file_priv);
+}
+
+int drm_mode_cursor2_ioctl(struct drm_device *dev,
+			   void *data, struct drm_file *file_priv)
+{
+	struct drm_mode_cursor2 *req = data;
+	return drm_mode_cursor_common(dev, req, file_priv);
 }
 
 /* Original addfb only supported RGB formats, so figure out which one */
@@ -2312,7 +2414,8 @@ static int framebuffer_check(const struct drm_mode_fb_cmd2 *r)
 
 	ret = format_check(r);
 	if (ret) {
-		DRM_DEBUG_KMS("bad framebuffer format 0x%08x\n", r->pixel_format);
+		DRM_DEBUG_KMS("bad framebuffer format %s\n",
+			      drm_get_format_name(r->pixel_format));
 		return ret;
 	}
 
diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c
index ed1334e27c33..6a647493ca7f 100644
--- a/drivers/gpu/drm/drm_crtc_helper.c
+++ b/drivers/gpu/drm/drm_crtc_helper.c
@@ -189,13 +189,14 @@ prune:
 	if (list_empty(&connector->modes))
 		return 0;
 
+	list_for_each_entry(mode, &connector->modes, head)
+		mode->vrefresh = drm_mode_vrefresh(mode);
+
 	drm_mode_sort(&connector->modes);
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s] probed modes :\n", connector->base.id,
 			drm_get_connector_name(connector));
 	list_for_each_entry(mode, &connector->modes, head) {
-		mode->vrefresh = drm_mode_vrefresh(mode);
-
 		drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V);
 		drm_mode_debug_printmodeline(mode);
 	}
@@ -564,14 +565,13 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
 
 	DRM_DEBUG_KMS("\n");
 
-	if (!set)
-		return -EINVAL;
+	BUG_ON(!set);
+	BUG_ON(!set->crtc);
+	BUG_ON(!set->crtc->helper_private);
 
-	if (!set->crtc)
-		return -EINVAL;
-
-	if (!set->crtc->helper_private)
-		return -EINVAL;
+	/* Enforce sane interface api - has been abused by the fb helper. */
+	BUG_ON(!set->mode && set->fb);
+	BUG_ON(set->fb && set->num_connectors == 0);
 
 	crtc_funcs = set->crtc->helper_private;
 
@@ -645,11 +645,6 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
 			mode_changed = true;
 		} else if (set->fb == NULL) {
 			mode_changed = true;
-		} else if (set->fb->depth != set->crtc->fb->depth) {
-			mode_changed = true;
-		} else if (set->fb->bits_per_pixel !=
-			   set->crtc->fb->bits_per_pixel) {
-			mode_changed = true;
 		} else if (set->fb->pixel_format !=
 			   set->crtc->fb->pixel_format) {
 			mode_changed = true;
@@ -682,6 +677,11 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set)
 					/* don't break so fail path works correct */
 					fail = 1;
 				break;
+
+				if (connector->dpms != DRM_MODE_DPMS_ON) {
+					DRM_DEBUG_KMS("connector dpms not on, full mode switch\n");
+					mode_changed = true;
+				}
 			}
 		}
 
diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
index 9cc247f55502..99fcd7c32ea2 100644
--- a/drivers/gpu/drm/drm_drv.c
+++ b/drivers/gpu/drm/drm_drv.c
@@ -166,6 +166,7 @@ static const struct drm_ioctl_desc drm_ioctls[] = {
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_DESTROY_DUMB, drm_mode_destroy_dumb_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_GETPROPERTIES, drm_mode_obj_get_properties_ioctl, DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF(DRM_IOCTL_MODE_OBJ_SETPROPERTY, drm_mode_obj_set_property_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
+	DRM_IOCTL_DEF(DRM_IOCTL_MODE_CURSOR2, drm_mode_cursor2_ioctl, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 };
 
 #define DRM_CORE_IOCTL_COUNT	ARRAY_SIZE( drm_ioctls )
diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c
index 9e62bbedb5ad..95d6f4b6967c 100644
--- a/drivers/gpu/drm/drm_edid.c
+++ b/drivers/gpu/drm/drm_edid.c
@@ -968,6 +968,9 @@ bool drm_edid_block_valid(u8 *raw_edid, int block, bool print_bad_edid)
 	u8 csum = 0;
 	struct edid *edid = (struct edid *)raw_edid;
 
+	if (WARN_ON(!raw_edid))
+		return false;
+
 	if (edid_fixup > 8 || edid_fixup < 0)
 		edid_fixup = 6;
 
@@ -1010,15 +1013,15 @@ bool drm_edid_block_valid(u8 *raw_edid, int block, bool print_bad_edid)
 		break;
 	}
 
-	return 1;
+	return true;
 
 bad:
-	if (raw_edid && print_bad_edid) {
+	if (print_bad_edid) {
 		printk(KERN_ERR "Raw EDID:\n");
 		print_hex_dump(KERN_ERR, " \t", DUMP_PREFIX_NONE, 16, 1,
 			       raw_edid, EDID_LENGTH, false);
 	}
-	return 0;
+	return false;
 }
 EXPORT_SYMBOL(drm_edid_block_valid);
 
@@ -1706,11 +1709,11 @@ static struct drm_display_mode *drm_mode_detailed(struct drm_device *dev,
 		return NULL;
 
 	if (pt->misc & DRM_EDID_PT_STEREO) {
-		printk(KERN_WARNING "stereo mode not supported\n");
+		DRM_DEBUG_KMS("stereo mode not supported\n");
 		return NULL;
 	}
 	if (!(pt->misc & DRM_EDID_PT_SEPARATE_SYNC)) {
-		printk(KERN_WARNING "composite sync not supported\n");
+		DRM_DEBUG_KMS("composite sync not supported\n");
 	}
 
 	/* it is incorrect if hsync/vsync width is zero */
@@ -2321,6 +2324,31 @@ u8 *drm_find_cea_extension(struct edid *edid)
 }
 EXPORT_SYMBOL(drm_find_cea_extension);
 
+/*
+ * Calculate the alternate clock for the CEA mode
+ * (60Hz vs. 59.94Hz etc.)
+ */
+static unsigned int
+cea_mode_alternate_clock(const struct drm_display_mode *cea_mode)
+{
+	unsigned int clock = cea_mode->clock;
+
+	if (cea_mode->vrefresh % 6 != 0)
+		return clock;
+
+	/*
+	 * edid_cea_modes contains the 59.94Hz
+	 * variant for 240 and 480 line modes,
+	 * and the 60Hz variant otherwise.
+	 */
+	if (cea_mode->vdisplay == 240 || cea_mode->vdisplay == 480)
+		clock = clock * 1001 / 1000;
+	else
+		clock = DIV_ROUND_UP(clock * 1000, 1001);
+
+	return clock;
+}
+
 /**
  * drm_match_cea_mode - look for a CEA mode matching given mode
  * @to_match: display mode
@@ -2339,21 +2367,9 @@ u8 drm_match_cea_mode(const struct drm_display_mode *to_match)
 		const struct drm_display_mode *cea_mode = &edid_cea_modes[mode];
 		unsigned int clock1, clock2;
 
-		clock1 = clock2 = cea_mode->clock;
-
 		/* Check both 60Hz and 59.94Hz */
-		if (cea_mode->vrefresh % 6 == 0) {
-			/*
-			 * edid_cea_modes contains the 59.94Hz
-			 * variant for 240 and 480 line modes,
-			 * and the 60Hz variant otherwise.
-			 */
-			if (cea_mode->vdisplay == 240 ||
-			    cea_mode->vdisplay == 480)
-				clock1 = clock1 * 1001 / 1000;
-			else
-				clock2 = DIV_ROUND_UP(clock2 * 1000, 1001);
-		}
+		clock1 = cea_mode->clock;
+		clock2 = cea_mode_alternate_clock(cea_mode);
 
 		if ((KHZ2PICOS(to_match->clock) == KHZ2PICOS(clock1) ||
 		     KHZ2PICOS(to_match->clock) == KHZ2PICOS(clock2)) &&
@@ -2364,6 +2380,66 @@ u8 drm_match_cea_mode(const struct drm_display_mode *to_match)
 }
 EXPORT_SYMBOL(drm_match_cea_mode);
 
+static int
+add_alternate_cea_modes(struct drm_connector *connector, struct edid *edid)
+{
+	struct drm_device *dev = connector->dev;
+	struct drm_display_mode *mode, *tmp;
+	LIST_HEAD(list);
+	int modes = 0;
+
+	/* Don't add CEA modes if the CEA extension block is missing */
+	if (!drm_find_cea_extension(edid))
+		return 0;
+
+	/*
+	 * Go through all probed modes and create a new mode
+	 * with the alternate clock for certain CEA modes.
+	 */
+	list_for_each_entry(mode, &connector->probed_modes, head) {
+		const struct drm_display_mode *cea_mode;
+		struct drm_display_mode *newmode;
+		u8 cea_mode_idx = drm_match_cea_mode(mode) - 1;
+		unsigned int clock1, clock2;
+
+		if (cea_mode_idx >= ARRAY_SIZE(edid_cea_modes))
+			continue;
+
+		cea_mode = &edid_cea_modes[cea_mode_idx];
+
+		clock1 = cea_mode->clock;
+		clock2 = cea_mode_alternate_clock(cea_mode);
+
+		if (clock1 == clock2)
+			continue;
+
+		if (mode->clock != clock1 && mode->clock != clock2)
+			continue;
+
+		newmode = drm_mode_duplicate(dev, cea_mode);
+		if (!newmode)
+			continue;
+
+		/*
+		 * The current mode could be either variant. Make
+		 * sure to pick the "other" clock for the new mode.
+		 */
+		if (mode->clock != clock1)
+			newmode->clock = clock1;
+		else
+			newmode->clock = clock2;
+
+		list_add_tail(&newmode->head, &list);
+	}
+
+	list_for_each_entry_safe(mode, tmp, &list, head) {
+		list_del(&mode->head);
+		drm_mode_probed_add(connector, mode);
+		modes++;
+	}
+
+	return modes;
+}
 
 static int
 do_cea_modes (struct drm_connector *connector, u8 *db, u8 len)
@@ -2946,6 +3022,7 @@ int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid)
 	if (edid->features & DRM_EDID_FEATURE_DEFAULT_GTF)
 		num_modes += add_inferred_modes(connector, edid);
 	num_modes += add_cea_modes(connector, edid);
+	num_modes += add_alternate_cea_modes(connector, edid);
 
 	if (quirks & (EDID_QUIRK_PREFER_LARGE_60 | EDID_QUIRK_PREFER_LARGE_75))
 		edid_fixup_preferred(connector, quirks);
diff --git a/drivers/gpu/drm/drm_edid_load.c b/drivers/gpu/drm/drm_edid_load.c
index fa445dd4dc00..271b42bbfb72 100644
--- a/drivers/gpu/drm/drm_edid_load.c
+++ b/drivers/gpu/drm/drm_edid_load.c
@@ -133,8 +133,8 @@ static u8 generic_edid[GENERIC_EDIDS][128] = {
 	},
 };
 
-static u8 *edid_load(struct drm_connector *connector, char *name,
-			char *connector_name)
+static u8 *edid_load(struct drm_connector *connector, const char *name,
+			const char *connector_name)
 {
 	const struct firmware *fw;
 	struct platform_device *pdev;
@@ -186,12 +186,11 @@ static u8 *edid_load(struct drm_connector *connector, char *name,
 		goto relfw_out;
 	}
 
-	edid = kmalloc(fwsize, GFP_KERNEL);
+	edid = kmemdup(fwdata, fwsize, GFP_KERNEL);
 	if (edid == NULL) {
 		err = -ENOMEM;
 		goto relfw_out;
 	}
-	memcpy(edid, fwdata, fwsize);
 
 	if (!drm_edid_block_valid(edid, 0, print_bad_edid)) {
 		connector->bad_edid_counter++;
@@ -243,7 +242,7 @@ out:
 
 int drm_load_edid_firmware(struct drm_connector *connector)
 {
-	char *connector_name = drm_get_connector_name(connector);
+	const char *connector_name = drm_get_connector_name(connector);
 	char *edidname = edid_firmware, *last, *colon;
 	int ret;
 	struct edid *edid;
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index b78cbe74dadf..3d13ca6e257f 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -168,6 +168,9 @@ static void drm_fb_helper_save_lut_atomic(struct drm_crtc *crtc, struct drm_fb_h
 	uint16_t *r_base, *g_base, *b_base;
 	int i;
 
+	if (helper->funcs->gamma_get == NULL)
+		return;
+
 	r_base = crtc->gamma_store;
 	g_base = r_base + crtc->gamma_size;
 	b_base = g_base + crtc->gamma_size;
@@ -284,13 +287,27 @@ EXPORT_SYMBOL(drm_fb_helper_debug_leave);
  */
 bool drm_fb_helper_restore_fbdev_mode(struct drm_fb_helper *fb_helper)
 {
+	struct drm_device *dev = fb_helper->dev;
+	struct drm_plane *plane;
 	bool error = false;
-	int i, ret;
+	int i;
+
+	drm_warn_on_modeset_not_all_locked(dev);
 
-	drm_warn_on_modeset_not_all_locked(fb_helper->dev);
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head)
+		drm_plane_force_disable(plane);
 
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		struct drm_mode_set *mode_set = &fb_helper->crtc_info[i].mode_set;
+		struct drm_crtc *crtc = mode_set->crtc;
+		int ret;
+
+		if (crtc->funcs->cursor_set) {
+			ret = crtc->funcs->cursor_set(crtc, NULL, 0, 0, 0);
+			if (ret)
+				error = true;
+		}
+
 		ret = drm_mode_set_config_internal(mode_set);
 		if (ret)
 			error = true;
@@ -583,6 +600,14 @@ static int setcolreg(struct drm_crtc *crtc, u16 red, u16 green,
 		return 0;
 	}
 
+	/*
+	 * The driver really shouldn't advertise pseudo/directcolor
+	 * visuals if it can't deal with the palette.
+	 */
+	if (WARN_ON(!fb_helper->funcs->gamma_set ||
+		    !fb_helper->funcs->gamma_get))
+		return -EINVAL;
+
 	pindex = regno;
 
 	if (fb->bits_per_pixel == 16) {
@@ -626,12 +651,19 @@ static int setcolreg(struct drm_crtc *crtc, u16 red, u16 green,
 int drm_fb_helper_setcmap(struct fb_cmap *cmap, struct fb_info *info)
 {
 	struct drm_fb_helper *fb_helper = info->par;
+	struct drm_device *dev = fb_helper->dev;
 	struct drm_crtc_helper_funcs *crtc_funcs;
 	u16 *red, *green, *blue, *transp;
 	struct drm_crtc *crtc;
 	int i, j, rc = 0;
 	int start;
 
+	drm_modeset_lock_all(dev);
+	if (!drm_fb_helper_is_bound(fb_helper)) {
+		drm_modeset_unlock_all(dev);
+		return -EBUSY;
+	}
+
 	for (i = 0; i < fb_helper->crtc_count; i++) {
 		crtc = fb_helper->crtc_info[i].mode_set.crtc;
 		crtc_funcs = crtc->helper_private;
@@ -654,10 +686,13 @@ int drm_fb_helper_setcmap(struct fb_cmap *cmap, struct fb_info *info)
 
 			rc = setcolreg(crtc, hred, hgreen, hblue, start++, info);
 			if (rc)
-				return rc;
+				goto out;
 		}
-		crtc_funcs->load_lut(crtc);
+		if (crtc_funcs->load_lut)
+			crtc_funcs->load_lut(crtc);
 	}
+ out:
+	drm_modeset_unlock_all(dev);
 	return rc;
 }
 EXPORT_SYMBOL(drm_fb_helper_setcmap);
diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c
index 429e07d0b0f1..3a24385e0368 100644
--- a/drivers/gpu/drm/drm_fops.c
+++ b/drivers/gpu/drm/drm_fops.c
@@ -271,6 +271,11 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 	priv->uid = current_euid();
 	priv->pid = get_pid(task_pid(current));
 	priv->minor = idr_find(&drm_minors_idr, minor_id);
+	if (!priv->minor) {
+		ret = -ENODEV;
+		goto out_put_pid;
+	}
+
 	priv->ioctl_count = 0;
 	/* for compatibility root is always authenticated */
 	priv->authenticated = capable(CAP_SYS_ADMIN);
@@ -292,7 +297,7 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 	if (dev->driver->open) {
 		ret = dev->driver->open(dev, priv);
 		if (ret < 0)
-			goto out_free;
+			goto out_prime_destroy;
 	}
 
 
@@ -304,7 +309,7 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 		if (!priv->minor->master) {
 			mutex_unlock(&dev->struct_mutex);
 			ret = -ENOMEM;
-			goto out_free;
+			goto out_close;
 		}
 
 		priv->is_master = 1;
@@ -322,7 +327,7 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 				drm_master_put(&priv->minor->master);
 				drm_master_put(&priv->master);
 				mutex_unlock(&dev->struct_mutex);
-				goto out_free;
+				goto out_close;
 			}
 		}
 		mutex_lock(&dev->struct_mutex);
@@ -333,7 +338,7 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 				drm_master_put(&priv->minor->master);
 				drm_master_put(&priv->master);
 				mutex_unlock(&dev->struct_mutex);
-				goto out_free;
+				goto out_close;
 			}
 		}
 		mutex_unlock(&dev->struct_mutex);
@@ -367,7 +372,17 @@ static int drm_open_helper(struct inode *inode, struct file *filp,
 #endif
 
 	return 0;
-      out_free:
+
+out_close:
+	if (dev->driver->postclose)
+		dev->driver->postclose(dev, priv);
+out_prime_destroy:
+	if (drm_core_check_feature(dev, DRIVER_PRIME))
+		drm_prime_destroy_file_private(&priv->prime);
+	if (dev->driver->driver_features & DRIVER_GEM)
+		drm_gem_release(dev, priv);
+out_put_pid:
+	put_pid(priv->pid);
 	kfree(priv);
 	filp->private_data = NULL;
 	return ret;
diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
index cf919e36e8ae..603f256152ef 100644
--- a/drivers/gpu/drm/drm_gem.c
+++ b/drivers/gpu/drm/drm_gem.c
@@ -108,12 +108,8 @@ drm_gem_init(struct drm_device *dev)
 		return -ENOMEM;
 	}
 
-	if (drm_mm_init(&mm->offset_manager, DRM_FILE_PAGE_OFFSET_START,
-			DRM_FILE_PAGE_OFFSET_SIZE)) {
-		drm_ht_remove(&mm->offset_hash);
-		kfree(mm);
-		return -ENOMEM;
-	}
+	drm_mm_init(&mm->offset_manager, DRM_FILE_PAGE_OFFSET_START,
+		    DRM_FILE_PAGE_OFFSET_SIZE);
 
 	return 0;
 }
@@ -453,25 +449,21 @@ drm_gem_flink_ioctl(struct drm_device *dev, void *data,
 	spin_lock(&dev->object_name_lock);
 	if (!obj->name) {
 		ret = idr_alloc(&dev->object_name_idr, obj, 1, 0, GFP_NOWAIT);
-		obj->name = ret;
-		args->name = (uint64_t) obj->name;
-		spin_unlock(&dev->object_name_lock);
-		idr_preload_end();
-
 		if (ret < 0)
 			goto err;
-		ret = 0;
+
+		obj->name = ret;
 
 		/* Allocate a reference for the name table.  */
 		drm_gem_object_reference(obj);
-	} else {
-		args->name = (uint64_t) obj->name;
-		spin_unlock(&dev->object_name_lock);
-		idr_preload_end();
-		ret = 0;
 	}
 
+	args->name = (uint64_t) obj->name;
+	ret = 0;
+
 err:
+	spin_unlock(&dev->object_name_lock);
+	idr_preload_end();
 	drm_gem_object_unreference_unlocked(obj);
 	return ret;
 }
@@ -644,6 +636,59 @@ void drm_gem_vm_close(struct vm_area_struct *vma)
 }
 EXPORT_SYMBOL(drm_gem_vm_close);
 
+/**
+ * drm_gem_mmap_obj - memory map a GEM object
+ * @obj: the GEM object to map
+ * @obj_size: the object size to be mapped, in bytes
+ * @vma: VMA for the area to be mapped
+ *
+ * Set up the VMA to prepare mapping of the GEM object using the gem_vm_ops
+ * provided by the driver. Depending on their requirements, drivers can either
+ * provide a fault handler in their gem_vm_ops (in which case any accesses to
+ * the object will be trapped, to perform migration, GTT binding, surface
+ * register allocation, or performance monitoring), or mmap the buffer memory
+ * synchronously after calling drm_gem_mmap_obj.
+ *
+ * This function is mainly intended to implement the DMABUF mmap operation, when
+ * the GEM object is not looked up based on its fake offset. To implement the
+ * DRM mmap operation, drivers should use the drm_gem_mmap() function.
+ *
+ * NOTE: This function has to be protected with dev->struct_mutex
+ *
+ * Return 0 or success or -EINVAL if the object size is smaller than the VMA
+ * size, or if no gem_vm_ops are provided.
+ */
+int drm_gem_mmap_obj(struct drm_gem_object *obj, unsigned long obj_size,
+		     struct vm_area_struct *vma)
+{
+	struct drm_device *dev = obj->dev;
+
+	lockdep_assert_held(&dev->struct_mutex);
+
+	/* Check for valid size. */
+	if (obj_size < vma->vm_end - vma->vm_start)
+		return -EINVAL;
+
+	if (!dev->driver->gem_vm_ops)
+		return -EINVAL;
+
+	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP;
+	vma->vm_ops = dev->driver->gem_vm_ops;
+	vma->vm_private_data = obj;
+	vma->vm_page_prot =  pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
+
+	/* Take a ref for this mapping of the object, so that the fault
+	 * handler can dereference the mmap offset's pointer to the object.
+	 * This reference is cleaned up by the corresponding vm_close
+	 * (which should happen whether the vma was created by this call, or
+	 * by a vm_open due to mremap or partial unmap or whatever).
+	 */
+	drm_gem_object_reference(obj);
+
+	drm_vm_open_locked(dev, vma);
+	return 0;
+}
+EXPORT_SYMBOL(drm_gem_mmap_obj);
 
 /**
  * drm_gem_mmap - memory map routine for GEM objects
@@ -653,11 +698,9 @@ EXPORT_SYMBOL(drm_gem_vm_close);
  * If a driver supports GEM object mapping, mmap calls on the DRM file
  * descriptor will end up here.
  *
- * If we find the object based on the offset passed in (vma->vm_pgoff will
+ * Look up the GEM object based on the offset passed in (vma->vm_pgoff will
  * contain the fake offset we created when the GTT map ioctl was called on
- * the object), we set up the driver fault handler so that any accesses
- * to the object can be trapped, to perform migration, GTT binding, surface
- * register allocation, or performance monitoring.
+ * the object) and map it with a call to drm_gem_mmap_obj().
  */
 int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 {
@@ -665,7 +708,6 @@ int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct drm_device *dev = priv->minor->dev;
 	struct drm_gem_mm *mm = dev->mm_private;
 	struct drm_local_map *map = NULL;
-	struct drm_gem_object *obj;
 	struct drm_hash_item *hash;
 	int ret = 0;
 
@@ -686,32 +728,7 @@ int drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 		goto out_unlock;
 	}
 
-	/* Check for valid size. */
-	if (map->size < vma->vm_end - vma->vm_start) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	obj = map->handle;
-	if (!obj->dev->driver->gem_vm_ops) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP;
-	vma->vm_ops = obj->dev->driver->gem_vm_ops;
-	vma->vm_private_data = map->handle;
-	vma->vm_page_prot =  pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
-
-	/* Take a ref for this mapping of the object, so that the fault
-	 * handler can dereference the mmap offset's pointer to the object.
-	 * This reference is cleaned up by the corresponding vm_close
-	 * (which should happen whether the vma was created by this call, or
-	 * by a vm_open due to mremap or partial unmap or whatever).
-	 */
-	drm_gem_object_reference(obj);
-
-	drm_vm_open_locked(dev, vma);
+	ret = drm_gem_mmap_obj(map->handle, map->size, vma);
 
 out_unlock:
 	mutex_unlock(&dev->struct_mutex);
diff --git a/drivers/gpu/drm/drm_gem_cma_helper.c b/drivers/gpu/drm/drm_gem_cma_helper.c
index 0a7e011509bd..ece72a8ac245 100644
--- a/drivers/gpu/drm/drm_gem_cma_helper.c
+++ b/drivers/gpu/drm/drm_gem_cma_helper.c
@@ -21,6 +21,7 @@
 #include <linux/slab.h>
 #include <linux/mutex.h>
 #include <linux/export.h>
+#include <linux/dma-buf.h>
 #include <linux/dma-mapping.h>
 
 #include <drm/drmP.h>
@@ -32,11 +33,44 @@ static unsigned int get_gem_mmap_offset(struct drm_gem_object *obj)
 	return (unsigned int)obj->map_list.hash.key << PAGE_SHIFT;
 }
 
-static void drm_gem_cma_buf_destroy(struct drm_device *drm,
-		struct drm_gem_cma_object *cma_obj)
+/*
+ * __drm_gem_cma_create - Create a GEM CMA object without allocating memory
+ * @drm: The drm device
+ * @size: The GEM object size
+ *
+ * This function creates and initializes a GEM CMA object of the given size, but
+ * doesn't allocate any memory to back the object.
+ *
+ * Return a struct drm_gem_cma_object* on success or ERR_PTR values on failure.
+ */
+static struct drm_gem_cma_object *
+__drm_gem_cma_create(struct drm_device *drm, unsigned int size)
 {
-	dma_free_writecombine(drm->dev, cma_obj->base.size, cma_obj->vaddr,
-			cma_obj->paddr);
+	struct drm_gem_cma_object *cma_obj;
+	struct drm_gem_object *gem_obj;
+	int ret;
+
+	cma_obj = kzalloc(sizeof(*cma_obj), GFP_KERNEL);
+	if (!cma_obj)
+		return ERR_PTR(-ENOMEM);
+
+	gem_obj = &cma_obj->base;
+
+	ret = drm_gem_object_init(drm, gem_obj, size);
+	if (ret)
+		goto error;
+
+	ret = drm_gem_create_mmap_offset(gem_obj);
+	if (ret) {
+		drm_gem_object_release(gem_obj);
+		goto error;
+	}
+
+	return cma_obj;
+
+error:
+	kfree(cma_obj);
+	return ERR_PTR(ret);
 }
 
 /*
@@ -49,44 +83,42 @@ struct drm_gem_cma_object *drm_gem_cma_create(struct drm_device *drm,
 		unsigned int size)
 {
 	struct drm_gem_cma_object *cma_obj;
-	struct drm_gem_object *gem_obj;
+	struct sg_table *sgt = NULL;
 	int ret;
 
 	size = round_up(size, PAGE_SIZE);
 
-	cma_obj = kzalloc(sizeof(*cma_obj), GFP_KERNEL);
-	if (!cma_obj)
-		return ERR_PTR(-ENOMEM);
+	cma_obj = __drm_gem_cma_create(drm, size);
+	if (IS_ERR(cma_obj))
+		return cma_obj;
 
 	cma_obj->vaddr = dma_alloc_writecombine(drm->dev, size,
 			&cma_obj->paddr, GFP_KERNEL | __GFP_NOWARN);
 	if (!cma_obj->vaddr) {
-		dev_err(drm->dev, "failed to allocate buffer with size %d\n", size);
+		dev_err(drm->dev, "failed to allocate buffer with size %d\n",
+			size);
 		ret = -ENOMEM;
-		goto err_dma_alloc;
+		goto error;
 	}
 
-	gem_obj = &cma_obj->base;
+	sgt = kzalloc(sizeof(*cma_obj->sgt), GFP_KERNEL);
+	if (sgt == NULL) {
+		ret = -ENOMEM;
+		goto error;
+	}
 
-	ret = drm_gem_object_init(drm, gem_obj, size);
-	if (ret)
-		goto err_obj_init;
+	ret = dma_get_sgtable(drm->dev, sgt, cma_obj->vaddr,
+			      cma_obj->paddr, size);
+	if (ret < 0)
+		goto error;
 
-	ret = drm_gem_create_mmap_offset(gem_obj);
-	if (ret)
-		goto err_create_mmap_offset;
+	cma_obj->sgt = sgt;
 
 	return cma_obj;
 
-err_create_mmap_offset:
-	drm_gem_object_release(gem_obj);
-
-err_obj_init:
-	drm_gem_cma_buf_destroy(drm, cma_obj);
-
-err_dma_alloc:
-	kfree(cma_obj);
-
+error:
+	kfree(sgt);
+	drm_gem_cma_free_object(&cma_obj->base);
 	return ERR_PTR(ret);
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_create);
@@ -143,11 +175,20 @@ void drm_gem_cma_free_object(struct drm_gem_object *gem_obj)
 	if (gem_obj->map_list.map)
 		drm_gem_free_mmap_offset(gem_obj);
 
-	drm_gem_object_release(gem_obj);
-
 	cma_obj = to_drm_gem_cma_obj(gem_obj);
 
-	drm_gem_cma_buf_destroy(gem_obj->dev, cma_obj);
+	if (cma_obj->vaddr) {
+		dma_free_writecombine(gem_obj->dev->dev, cma_obj->base.size,
+				      cma_obj->vaddr, cma_obj->paddr);
+		if (cma_obj->sgt) {
+			sg_free_table(cma_obj->sgt);
+			kfree(cma_obj->sgt);
+		}
+	} else if (gem_obj->import_attach) {
+		drm_prime_gem_destroy(gem_obj, cma_obj->sgt);
+	}
+
+	drm_gem_object_release(gem_obj);
 
 	kfree(cma_obj);
 }
@@ -174,10 +215,7 @@ int drm_gem_cma_dumb_create(struct drm_file *file_priv,
 
 	cma_obj = drm_gem_cma_create_with_handle(file_priv, dev,
 			args->size, &args->handle);
-	if (IS_ERR(cma_obj))
-		return PTR_ERR(cma_obj);
-
-	return 0;
+	return PTR_RET(cma_obj);
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_dumb_create);
 
@@ -215,13 +253,26 @@ const struct vm_operations_struct drm_gem_cma_vm_ops = {
 };
 EXPORT_SYMBOL_GPL(drm_gem_cma_vm_ops);
 
+static int drm_gem_cma_mmap_obj(struct drm_gem_cma_object *cma_obj,
+				struct vm_area_struct *vma)
+{
+	int ret;
+
+	ret = remap_pfn_range(vma, vma->vm_start, cma_obj->paddr >> PAGE_SHIFT,
+			vma->vm_end - vma->vm_start, vma->vm_page_prot);
+	if (ret)
+		drm_gem_vm_close(vma);
+
+	return ret;
+}
+
 /*
  * drm_gem_cma_mmap - (struct file_operation)->mmap callback function
  */
 int drm_gem_cma_mmap(struct file *filp, struct vm_area_struct *vma)
 {
-	struct drm_gem_object *gem_obj;
 	struct drm_gem_cma_object *cma_obj;
+	struct drm_gem_object *gem_obj;
 	int ret;
 
 	ret = drm_gem_mmap(filp, vma);
@@ -231,12 +282,7 @@ int drm_gem_cma_mmap(struct file *filp, struct vm_area_struct *vma)
 	gem_obj = vma->vm_private_data;
 	cma_obj = to_drm_gem_cma_obj(gem_obj);
 
-	ret = remap_pfn_range(vma, vma->vm_start, cma_obj->paddr >> PAGE_SHIFT,
-			vma->vm_end - vma->vm_start, vma->vm_page_prot);
-	if (ret)
-		drm_gem_vm_close(vma);
-
-	return ret;
+	return drm_gem_cma_mmap_obj(cma_obj, vma);
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_mmap);
 
@@ -270,3 +316,82 @@ void drm_gem_cma_describe(struct drm_gem_cma_object *cma_obj, struct seq_file *m
 }
 EXPORT_SYMBOL_GPL(drm_gem_cma_describe);
 #endif
+
+/* low-level interface prime helpers */
+struct sg_table *drm_gem_cma_prime_get_sg_table(struct drm_gem_object *obj)
+{
+	struct drm_gem_cma_object *cma_obj = to_drm_gem_cma_obj(obj);
+	struct sg_table *sgt;
+	int ret;
+
+	sgt = kzalloc(sizeof(*sgt), GFP_KERNEL);
+	if (!sgt)
+		return NULL;
+
+	ret = dma_get_sgtable(obj->dev->dev, sgt, cma_obj->vaddr,
+			      cma_obj->paddr, obj->size);
+	if (ret < 0)
+		goto out;
+
+	return sgt;
+
+out:
+	kfree(sgt);
+	return NULL;
+}
+EXPORT_SYMBOL_GPL(drm_gem_cma_prime_get_sg_table);
+
+struct drm_gem_object *
+drm_gem_cma_prime_import_sg_table(struct drm_device *dev, size_t size,
+				  struct sg_table *sgt)
+{
+	struct drm_gem_cma_object *cma_obj;
+
+	if (sgt->nents != 1)
+		return ERR_PTR(-EINVAL);
+
+	/* Create a CMA GEM buffer. */
+	cma_obj = __drm_gem_cma_create(dev, size);
+	if (IS_ERR(cma_obj))
+		return ERR_PTR(PTR_ERR(cma_obj));
+
+	cma_obj->paddr = sg_dma_address(sgt->sgl);
+	cma_obj->sgt = sgt;
+
+	DRM_DEBUG_PRIME("dma_addr = 0x%x, size = %zu\n", cma_obj->paddr, size);
+
+	return &cma_obj->base;
+}
+EXPORT_SYMBOL_GPL(drm_gem_cma_prime_import_sg_table);
+
+int drm_gem_cma_prime_mmap(struct drm_gem_object *obj,
+			   struct vm_area_struct *vma)
+{
+	struct drm_gem_cma_object *cma_obj;
+	struct drm_device *dev = obj->dev;
+	int ret;
+
+	mutex_lock(&dev->struct_mutex);
+	ret = drm_gem_mmap_obj(obj, obj->size, vma);
+	mutex_unlock(&dev->struct_mutex);
+	if (ret < 0)
+		return ret;
+
+	cma_obj = to_drm_gem_cma_obj(obj);
+	return drm_gem_cma_mmap_obj(cma_obj, vma);
+}
+EXPORT_SYMBOL_GPL(drm_gem_cma_prime_mmap);
+
+void *drm_gem_cma_prime_vmap(struct drm_gem_object *obj)
+{
+	struct drm_gem_cma_object *cma_obj = to_drm_gem_cma_obj(obj);
+
+	return cma_obj->vaddr;
+}
+EXPORT_SYMBOL_GPL(drm_gem_cma_prime_vmap);
+
+void drm_gem_cma_prime_vunmap(struct drm_gem_object *obj, void *vaddr)
+{
+	/* Nothing to do */
+}
+EXPORT_SYMBOL_GPL(drm_gem_cma_prime_vunmap);
diff --git a/drivers/gpu/drm/drm_ioctl.c b/drivers/gpu/drm/drm_ioctl.c
index e77bd8b57df2..ffd7a7ba70d4 100644
--- a/drivers/gpu/drm/drm_ioctl.c
+++ b/drivers/gpu/drm/drm_ioctl.c
@@ -38,6 +38,9 @@
 
 #include <linux/pci.h>
 #include <linux/export.h>
+#ifdef CONFIG_X86
+#include <asm/mtrr.h>
+#endif
 
 /**
  * Get the bus id.
@@ -181,7 +184,17 @@ int drm_getmap(struct drm_device *dev, void *data,
 	map->type = r_list->map->type;
 	map->flags = r_list->map->flags;
 	map->handle = (void *)(unsigned long) r_list->user_token;
-	map->mtrr = r_list->map->mtrr;
+
+#ifdef CONFIG_X86
+	/*
+	 * There appears to be exactly one user of the mtrr index: dritest.
+	 * It's easy enough to keep it working on non-PAT systems.
+	 */
+	map->mtrr = phys_wc_to_mtrr_index(r_list->map->mtrr);
+#else
+	map->mtrr = -1;
+#endif
+
 	mutex_unlock(&dev->struct_mutex);
 
 	return 0;
diff --git a/drivers/gpu/drm/drm_mm.c b/drivers/gpu/drm/drm_mm.c
index 07cf99cc8862..543b9b3171d3 100644
--- a/drivers/gpu/drm/drm_mm.c
+++ b/drivers/gpu/drm/drm_mm.c
@@ -669,7 +669,7 @@ int drm_mm_clean(struct drm_mm * mm)
 }
 EXPORT_SYMBOL(drm_mm_clean);
 
-int drm_mm_init(struct drm_mm * mm, unsigned long start, unsigned long size)
+void drm_mm_init(struct drm_mm * mm, unsigned long start, unsigned long size)
 {
 	INIT_LIST_HEAD(&mm->hole_stack);
 	INIT_LIST_HEAD(&mm->unused_nodes);
@@ -690,8 +690,6 @@ int drm_mm_init(struct drm_mm * mm, unsigned long start, unsigned long size)
 	list_add_tail(&mm->head_node.hole_stack, &mm->hole_stack);
 
 	mm->color_adjust = NULL;
-
-	return 0;
 }
 EXPORT_SYMBOL(drm_mm_init);
 
@@ -699,8 +697,8 @@ void drm_mm_takedown(struct drm_mm * mm)
 {
 	struct drm_mm_node *entry, *next;
 
-	if (!list_empty(&mm->head_node.node_list)) {
-		DRM_ERROR("Memory manager not clean. Delaying takedown\n");
+	if (WARN(!list_empty(&mm->head_node.node_list),
+		 "Memory manager not clean. Delaying takedown\n")) {
 		return;
 	}
 
@@ -716,36 +714,37 @@ void drm_mm_takedown(struct drm_mm * mm)
 }
 EXPORT_SYMBOL(drm_mm_takedown);
 
-void drm_mm_debug_table(struct drm_mm *mm, const char *prefix)
+static unsigned long drm_mm_debug_hole(struct drm_mm_node *entry,
+				       const char *prefix)
 {
-	struct drm_mm_node *entry;
-	unsigned long total_used = 0, total_free = 0, total = 0;
 	unsigned long hole_start, hole_end, hole_size;
 
-	hole_start = drm_mm_hole_node_start(&mm->head_node);
-	hole_end = drm_mm_hole_node_end(&mm->head_node);
-	hole_size = hole_end - hole_start;
-	if (hole_size)
+	if (entry->hole_follows) {
+		hole_start = drm_mm_hole_node_start(entry);
+		hole_end = drm_mm_hole_node_end(entry);
+		hole_size = hole_end - hole_start;
 		printk(KERN_DEBUG "%s 0x%08lx-0x%08lx: %8lu: free\n",
 			prefix, hole_start, hole_end,
 			hole_size);
-	total_free += hole_size;
+		return hole_size;
+	}
+
+	return 0;
+}
+
+void drm_mm_debug_table(struct drm_mm *mm, const char *prefix)
+{
+	struct drm_mm_node *entry;
+	unsigned long total_used = 0, total_free = 0, total = 0;
+
+	total_free += drm_mm_debug_hole(&mm->head_node, prefix);
 
 	drm_mm_for_each_node(entry, mm) {
 		printk(KERN_DEBUG "%s 0x%08lx-0x%08lx: %8lu: used\n",
 			prefix, entry->start, entry->start + entry->size,
 			entry->size);
 		total_used += entry->size;
-
-		if (entry->hole_follows) {
-			hole_start = drm_mm_hole_node_start(entry);
-			hole_end = drm_mm_hole_node_end(entry);
-			hole_size = hole_end - hole_start;
-			printk(KERN_DEBUG "%s 0x%08lx-0x%08lx: %8lu: free\n",
-				prefix, hole_start, hole_end,
-				hole_size);
-			total_free += hole_size;
-		}
+		total_free += drm_mm_debug_hole(entry, prefix);
 	}
 	total = total_free + total_used;
 
diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c
index a371ff865a88..a6729bfe6860 100644
--- a/drivers/gpu/drm/drm_modes.c
+++ b/drivers/gpu/drm/drm_modes.c
@@ -535,6 +535,8 @@ int drm_display_mode_from_videomode(const struct videomode *vm,
 		dmode->flags |= DRM_MODE_FLAG_INTERLACE;
 	if (vm->flags & DISPLAY_FLAGS_DOUBLESCAN)
 		dmode->flags |= DRM_MODE_FLAG_DBLSCAN;
+	if (vm->flags & DISPLAY_FLAGS_DOUBLECLK)
+		dmode->flags |= DRM_MODE_FLAG_DBLCLK;
 	drm_mode_set_name(dmode);
 
 	return 0;
@@ -787,16 +789,17 @@ EXPORT_SYMBOL(drm_mode_set_crtcinfo);
  * LOCKING:
  * None.
  *
- * Copy an existing mode into another mode, preserving the object id
- * of the destination mode.
+ * Copy an existing mode into another mode, preserving the object id and
+ * list head of the destination mode.
  */
 void drm_mode_copy(struct drm_display_mode *dst, const struct drm_display_mode *src)
 {
 	int id = dst->base.id;
+	struct list_head head = dst->head;
 
 	*dst = *src;
 	dst->base.id = id;
-	INIT_LIST_HEAD(&dst->head);
+	dst->head = head;
 }
 EXPORT_SYMBOL(drm_mode_copy);
 
@@ -1017,6 +1020,11 @@ static int drm_mode_compare(void *priv, struct list_head *lh_a, struct list_head
 	diff = b->hdisplay * b->vdisplay - a->hdisplay * a->vdisplay;
 	if (diff)
 		return diff;
+
+	diff = b->vrefresh - a->vrefresh;
+	if (diff)
+		return diff;
+
 	diff = b->clock - a->clock;
 	return diff;
 }
diff --git a/drivers/gpu/drm/drm_pci.c b/drivers/gpu/drm/drm_pci.c
index 14194b6ef644..80c0b2b29801 100644
--- a/drivers/gpu/drm/drm_pci.c
+++ b/drivers/gpu/drm/drm_pci.c
@@ -278,10 +278,10 @@ static int drm_pci_agp_init(struct drm_device *dev)
 		}
 		if (drm_core_has_MTRR(dev)) {
 			if (dev->agp)
-				dev->agp->agp_mtrr =
-					mtrr_add(dev->agp->agp_info.aper_base,
-						 dev->agp->agp_info.aper_size *
-						 1024 * 1024, MTRR_TYPE_WRCOMB, 1);
+				dev->agp->agp_mtrr = arch_phys_wc_add(
+					dev->agp->agp_info.aper_base,
+					dev->agp->agp_info.aper_size *
+					1024 * 1024);
 		}
 	}
 	return 0;
diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c
index 5b7b9110254b..85e450e3241c 100644
--- a/drivers/gpu/drm/drm_prime.c
+++ b/drivers/gpu/drm/drm_prime.c
@@ -62,20 +62,125 @@ struct drm_prime_member {
 	struct dma_buf *dma_buf;
 	uint32_t handle;
 };
-static int drm_prime_add_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t handle);
+
+struct drm_prime_attachment {
+	struct sg_table *sgt;
+	enum dma_data_direction dir;
+};
+
+static int drm_prime_add_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t handle)
+{
+	struct drm_prime_member *member;
+
+	member = kmalloc(sizeof(*member), GFP_KERNEL);
+	if (!member)
+		return -ENOMEM;
+
+	get_dma_buf(dma_buf);
+	member->dma_buf = dma_buf;
+	member->handle = handle;
+	list_add(&member->entry, &prime_fpriv->head);
+	return 0;
+}
+
+static int drm_gem_map_attach(struct dma_buf *dma_buf,
+			      struct device *target_dev,
+			      struct dma_buf_attachment *attach)
+{
+	struct drm_prime_attachment *prime_attach;
+	struct drm_gem_object *obj = dma_buf->priv;
+	struct drm_device *dev = obj->dev;
+
+	prime_attach = kzalloc(sizeof(*prime_attach), GFP_KERNEL);
+	if (!prime_attach)
+		return -ENOMEM;
+
+	prime_attach->dir = DMA_NONE;
+	attach->priv = prime_attach;
+
+	if (!dev->driver->gem_prime_pin)
+		return 0;
+
+	return dev->driver->gem_prime_pin(obj);
+}
+
+static void drm_gem_map_detach(struct dma_buf *dma_buf,
+			       struct dma_buf_attachment *attach)
+{
+	struct drm_prime_attachment *prime_attach = attach->priv;
+	struct drm_gem_object *obj = dma_buf->priv;
+	struct drm_device *dev = obj->dev;
+	struct sg_table *sgt;
+
+	if (dev->driver->gem_prime_unpin)
+		dev->driver->gem_prime_unpin(obj);
+
+	if (!prime_attach)
+		return;
+
+	sgt = prime_attach->sgt;
+	if (sgt) {
+		if (prime_attach->dir != DMA_NONE)
+			dma_unmap_sg(attach->dev, sgt->sgl, sgt->nents,
+					prime_attach->dir);
+		sg_free_table(sgt);
+	}
+
+	kfree(sgt);
+	kfree(prime_attach);
+	attach->priv = NULL;
+}
+
+static void drm_prime_remove_buf_handle_locked(
+		struct drm_prime_file_private *prime_fpriv,
+		struct dma_buf *dma_buf)
+{
+	struct drm_prime_member *member, *safe;
+
+	list_for_each_entry_safe(member, safe, &prime_fpriv->head, entry) {
+		if (member->dma_buf == dma_buf) {
+			dma_buf_put(dma_buf);
+			list_del(&member->entry);
+			kfree(member);
+		}
+	}
+}
 
 static struct sg_table *drm_gem_map_dma_buf(struct dma_buf_attachment *attach,
 		enum dma_data_direction dir)
 {
+	struct drm_prime_attachment *prime_attach = attach->priv;
 	struct drm_gem_object *obj = attach->dmabuf->priv;
 	struct sg_table *sgt;
 
+	if (WARN_ON(dir == DMA_NONE || !prime_attach))
+		return ERR_PTR(-EINVAL);
+
+	/* return the cached mapping when possible */
+	if (prime_attach->dir == dir)
+		return prime_attach->sgt;
+
+	/*
+	 * two mappings with different directions for the same attachment are
+	 * not allowed
+	 */
+	if (WARN_ON(prime_attach->dir != DMA_NONE))
+		return ERR_PTR(-EBUSY);
+
 	mutex_lock(&obj->dev->struct_mutex);
 
 	sgt = obj->dev->driver->gem_prime_get_sg_table(obj);
 
-	if (!IS_ERR_OR_NULL(sgt))
-		dma_map_sg(attach->dev, sgt->sgl, sgt->nents, dir);
+	if (!IS_ERR(sgt)) {
+		if (!dma_map_sg(attach->dev, sgt->sgl, sgt->nents, dir)) {
+			sg_free_table(sgt);
+			kfree(sgt);
+			sgt = ERR_PTR(-ENOMEM);
+		} else {
+			prime_attach->sgt = sgt;
+			prime_attach->dir = dir;
+		}
+	}
 
 	mutex_unlock(&obj->dev->struct_mutex);
 	return sgt;
@@ -84,9 +189,7 @@ static struct sg_table *drm_gem_map_dma_buf(struct dma_buf_attachment *attach,
 static void drm_gem_unmap_dma_buf(struct dma_buf_attachment *attach,
 		struct sg_table *sgt, enum dma_data_direction dir)
 {
-	dma_unmap_sg(attach->dev, sgt->sgl, sgt->nents, dir);
-	sg_free_table(sgt);
-	kfree(sgt);
+	/* nothing to be done here */
 }
 
 static void drm_gem_dmabuf_release(struct dma_buf *dma_buf)
@@ -142,10 +245,18 @@ static void drm_gem_dmabuf_kunmap(struct dma_buf *dma_buf,
 static int drm_gem_dmabuf_mmap(struct dma_buf *dma_buf,
 		struct vm_area_struct *vma)
 {
-	return -EINVAL;
+	struct drm_gem_object *obj = dma_buf->priv;
+	struct drm_device *dev = obj->dev;
+
+	if (!dev->driver->gem_prime_mmap)
+		return -ENOSYS;
+
+	return dev->driver->gem_prime_mmap(obj, vma);
 }
 
 static const struct dma_buf_ops drm_gem_prime_dmabuf_ops =  {
+	.attach = drm_gem_map_attach,
+	.detach = drm_gem_map_detach,
 	.map_dma_buf = drm_gem_map_dma_buf,
 	.unmap_dma_buf = drm_gem_unmap_dma_buf,
 	.release = drm_gem_dmabuf_release,
@@ -185,11 +296,6 @@ static const struct dma_buf_ops drm_gem_prime_dmabuf_ops =  {
 struct dma_buf *drm_gem_prime_export(struct drm_device *dev,
 				     struct drm_gem_object *obj, int flags)
 {
-	if (dev->driver->gem_prime_pin) {
-		int ret = dev->driver->gem_prime_pin(obj);
-		if (ret)
-			return ERR_PTR(ret);
-	}
 	return dma_buf_export(obj, &drm_gem_prime_dmabuf_ops, obj->size, flags);
 }
 EXPORT_SYMBOL(drm_gem_prime_export);
@@ -235,15 +341,34 @@ int drm_gem_prime_handle_to_fd(struct drm_device *dev,
 	ret = drm_prime_add_buf_handle(&file_priv->prime,
 				       obj->export_dma_buf, handle);
 	if (ret)
-		goto out;
+		goto fail_put_dmabuf;
+
+	ret = dma_buf_fd(buf, flags);
+	if (ret < 0)
+		goto fail_rm_handle;
 
-	*prime_fd = dma_buf_fd(buf, flags);
+	*prime_fd = ret;
 	mutex_unlock(&file_priv->prime.lock);
 	return 0;
 
 out_have_obj:
 	get_dma_buf(dmabuf);
-	*prime_fd = dma_buf_fd(dmabuf, flags);
+	ret = dma_buf_fd(dmabuf, flags);
+	if (ret < 0) {
+		dma_buf_put(dmabuf);
+	} else {
+		*prime_fd = ret;
+		ret = 0;
+	}
+
+	goto out;
+
+fail_rm_handle:
+	drm_prime_remove_buf_handle_locked(&file_priv->prime, buf);
+fail_put_dmabuf:
+	/* clear NOT to be checked when releasing dma_buf */
+	obj->export_dma_buf = NULL;
+	dma_buf_put(buf);
 out:
 	drm_gem_object_unreference_unlocked(obj);
 	mutex_unlock(&file_priv->prime.lock);
@@ -276,7 +401,7 @@ struct drm_gem_object *drm_gem_prime_import(struct drm_device *dev,
 
 	attach = dma_buf_attach(dma_buf, dev->dev);
 	if (IS_ERR(attach))
-		return ERR_PTR(PTR_ERR(attach));
+		return ERR_CAST(attach);
 
 	get_dma_buf(dma_buf);
 
@@ -412,8 +537,10 @@ struct sg_table *drm_prime_pages_to_sg(struct page **pages, int nr_pages)
 	int ret;
 
 	sg = kmalloc(sizeof(struct sg_table), GFP_KERNEL);
-	if (!sg)
+	if (!sg) {
+		ret = -ENOMEM;
 		goto out;
+	}
 
 	ret = sg_alloc_table_from_pages(sg, pages, nr_pages, 0,
 				nr_pages << PAGE_SHIFT, GFP_KERNEL);
@@ -423,7 +550,7 @@ struct sg_table *drm_prime_pages_to_sg(struct page **pages, int nr_pages)
 	return sg;
 out:
 	kfree(sg);
-	return NULL;
+	return ERR_PTR(ret);
 }
 EXPORT_SYMBOL(drm_prime_pages_to_sg);
 
@@ -492,21 +619,6 @@ void drm_prime_destroy_file_private(struct drm_prime_file_private *prime_fpriv)
 }
 EXPORT_SYMBOL(drm_prime_destroy_file_private);
 
-static int drm_prime_add_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t handle)
-{
-	struct drm_prime_member *member;
-
-	member = kmalloc(sizeof(*member), GFP_KERNEL);
-	if (!member)
-		return -ENOMEM;
-
-	get_dma_buf(dma_buf);
-	member->dma_buf = dma_buf;
-	member->handle = handle;
-	list_add(&member->entry, &prime_fpriv->head);
-	return 0;
-}
-
 int drm_prime_lookup_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf, uint32_t *handle)
 {
 	struct drm_prime_member *member;
@@ -523,16 +635,8 @@ EXPORT_SYMBOL(drm_prime_lookup_buf_handle);
 
 void drm_prime_remove_buf_handle(struct drm_prime_file_private *prime_fpriv, struct dma_buf *dma_buf)
 {
-	struct drm_prime_member *member, *safe;
-
 	mutex_lock(&prime_fpriv->lock);
-	list_for_each_entry_safe(member, safe, &prime_fpriv->head, entry) {
-		if (member->dma_buf == dma_buf) {
-			dma_buf_put(dma_buf);
-			list_del(&member->entry);
-			kfree(member);
-		}
-	}
+	drm_prime_remove_buf_handle_locked(prime_fpriv, dma_buf);
 	mutex_unlock(&prime_fpriv->lock);
 }
 EXPORT_SYMBOL(drm_prime_remove_buf_handle);
diff --git a/drivers/gpu/drm/drm_rect.c b/drivers/gpu/drm/drm_rect.c
new file mode 100644
index 000000000000..7047ca025787
--- /dev/null
+++ b/drivers/gpu/drm/drm_rect.c
@@ -0,0 +1,295 @@
+/*
+ * Copyright (C) 2011-2013 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/errno.h>
+#include <linux/export.h>
+#include <linux/kernel.h>
+#include <drm/drmP.h>
+#include <drm/drm_rect.h>
+
+/**
+ * drm_rect_intersect - intersect two rectangles
+ * @r1: first rectangle
+ * @r2: second rectangle
+ *
+ * Calculate the intersection of rectangles @r1 and @r2.
+ * @r1 will be overwritten with the intersection.
+ *
+ * RETURNS:
+ * %true if rectangle @r1 is still visible after the operation,
+ * %false otherwise.
+ */
+bool drm_rect_intersect(struct drm_rect *r1, const struct drm_rect *r2)
+{
+	r1->x1 = max(r1->x1, r2->x1);
+	r1->y1 = max(r1->y1, r2->y1);
+	r1->x2 = min(r1->x2, r2->x2);
+	r1->y2 = min(r1->y2, r2->y2);
+
+	return drm_rect_visible(r1);
+}
+EXPORT_SYMBOL(drm_rect_intersect);
+
+/**
+ * drm_rect_clip_scaled - perform a scaled clip operation
+ * @src: source window rectangle
+ * @dst: destination window rectangle
+ * @clip: clip rectangle
+ * @hscale: horizontal scaling factor
+ * @vscale: vertical scaling factor
+ *
+ * Clip rectangle @dst by rectangle @clip. Clip rectangle @src by the
+ * same amounts multiplied by @hscale and @vscale.
+ *
+ * RETURNS:
+ * %true if rectangle @dst is still visible after being clipped,
+ * %false otherwise
+ */
+bool drm_rect_clip_scaled(struct drm_rect *src, struct drm_rect *dst,
+			  const struct drm_rect *clip,
+			  int hscale, int vscale)
+{
+	int diff;
+
+	diff = clip->x1 - dst->x1;
+	if (diff > 0) {
+		int64_t tmp = src->x1 + (int64_t) diff * hscale;
+		src->x1 = clamp_t(int64_t, tmp, INT_MIN, INT_MAX);
+	}
+	diff = clip->y1 - dst->y1;
+	if (diff > 0) {
+		int64_t tmp = src->y1 + (int64_t) diff * vscale;
+		src->y1 = clamp_t(int64_t, tmp, INT_MIN, INT_MAX);
+	}
+	diff = dst->x2 - clip->x2;
+	if (diff > 0) {
+		int64_t tmp = src->x2 - (int64_t) diff * hscale;
+		src->x2 = clamp_t(int64_t, tmp, INT_MIN, INT_MAX);
+	}
+	diff = dst->y2 - clip->y2;
+	if (diff > 0) {
+		int64_t tmp = src->y2 - (int64_t) diff * vscale;
+		src->y2 = clamp_t(int64_t, tmp, INT_MIN, INT_MAX);
+	}
+
+	return drm_rect_intersect(dst, clip);
+}
+EXPORT_SYMBOL(drm_rect_clip_scaled);
+
+static int drm_calc_scale(int src, int dst)
+{
+	int scale = 0;
+
+	if (src < 0 || dst < 0)
+		return -EINVAL;
+
+	if (dst == 0)
+		return 0;
+
+	scale = src / dst;
+
+	return scale;
+}
+
+/**
+ * drm_rect_calc_hscale - calculate the horizontal scaling factor
+ * @src: source window rectangle
+ * @dst: destination window rectangle
+ * @min_hscale: minimum allowed horizontal scaling factor
+ * @max_hscale: maximum allowed horizontal scaling factor
+ *
+ * Calculate the horizontal scaling factor as
+ * (@src width) / (@dst width).
+ *
+ * RETURNS:
+ * The horizontal scaling factor, or errno of out of limits.
+ */
+int drm_rect_calc_hscale(const struct drm_rect *src,
+			 const struct drm_rect *dst,
+			 int min_hscale, int max_hscale)
+{
+	int src_w = drm_rect_width(src);
+	int dst_w = drm_rect_width(dst);
+	int hscale = drm_calc_scale(src_w, dst_w);
+
+	if (hscale < 0 || dst_w == 0)
+		return hscale;
+
+	if (hscale < min_hscale || hscale > max_hscale)
+		return -ERANGE;
+
+	return hscale;
+}
+EXPORT_SYMBOL(drm_rect_calc_hscale);
+
+/**
+ * drm_rect_calc_vscale - calculate the vertical scaling factor
+ * @src: source window rectangle
+ * @dst: destination window rectangle
+ * @min_vscale: minimum allowed vertical scaling factor
+ * @max_vscale: maximum allowed vertical scaling factor
+ *
+ * Calculate the vertical scaling factor as
+ * (@src height) / (@dst height).
+ *
+ * RETURNS:
+ * The vertical scaling factor, or errno of out of limits.
+ */
+int drm_rect_calc_vscale(const struct drm_rect *src,
+			 const struct drm_rect *dst,
+			 int min_vscale, int max_vscale)
+{
+	int src_h = drm_rect_height(src);
+	int dst_h = drm_rect_height(dst);
+	int vscale = drm_calc_scale(src_h, dst_h);
+
+	if (vscale < 0 || dst_h == 0)
+		return vscale;
+
+	if (vscale < min_vscale || vscale > max_vscale)
+		return -ERANGE;
+
+	return vscale;
+}
+EXPORT_SYMBOL(drm_rect_calc_vscale);
+
+/**
+ * drm_calc_hscale_relaxed - calculate the horizontal scaling factor
+ * @src: source window rectangle
+ * @dst: destination window rectangle
+ * @min_hscale: minimum allowed horizontal scaling factor
+ * @max_hscale: maximum allowed horizontal scaling factor
+ *
+ * Calculate the horizontal scaling factor as
+ * (@src width) / (@dst width).
+ *
+ * If the calculated scaling factor is below @min_vscale,
+ * decrease the height of rectangle @dst to compensate.
+ *
+ * If the calculated scaling factor is above @max_vscale,
+ * decrease the height of rectangle @src to compensate.
+ *
+ * RETURNS:
+ * The horizontal scaling factor.
+ */
+int drm_rect_calc_hscale_relaxed(struct drm_rect *src,
+				 struct drm_rect *dst,
+				 int min_hscale, int max_hscale)
+{
+	int src_w = drm_rect_width(src);
+	int dst_w = drm_rect_width(dst);
+	int hscale = drm_calc_scale(src_w, dst_w);
+
+	if (hscale < 0 || dst_w == 0)
+		return hscale;
+
+	if (hscale < min_hscale) {
+		int max_dst_w = src_w / min_hscale;
+
+		drm_rect_adjust_size(dst, max_dst_w - dst_w, 0);
+
+		return min_hscale;
+	}
+
+	if (hscale > max_hscale) {
+		int max_src_w = dst_w * max_hscale;
+
+		drm_rect_adjust_size(src, max_src_w - src_w, 0);
+
+		return max_hscale;
+	}
+
+	return hscale;
+}
+EXPORT_SYMBOL(drm_rect_calc_hscale_relaxed);
+
+/**
+ * drm_rect_calc_vscale_relaxed - calculate the vertical scaling factor
+ * @src: source window rectangle
+ * @dst: destination window rectangle
+ * @min_vscale: minimum allowed vertical scaling factor
+ * @max_vscale: maximum allowed vertical scaling factor
+ *
+ * Calculate the vertical scaling factor as
+ * (@src height) / (@dst height).
+ *
+ * If the calculated scaling factor is below @min_vscale,
+ * decrease the height of rectangle @dst to compensate.
+ *
+ * If the calculated scaling factor is above @max_vscale,
+ * decrease the height of rectangle @src to compensate.
+ *
+ * RETURNS:
+ * The vertical scaling factor.
+ */
+int drm_rect_calc_vscale_relaxed(struct drm_rect *src,
+				 struct drm_rect *dst,
+				 int min_vscale, int max_vscale)
+{
+	int src_h = drm_rect_height(src);
+	int dst_h = drm_rect_height(dst);
+	int vscale = drm_calc_scale(src_h, dst_h);
+
+	if (vscale < 0 || dst_h == 0)
+		return vscale;
+
+	if (vscale < min_vscale) {
+		int max_dst_h = src_h / min_vscale;
+
+		drm_rect_adjust_size(dst, 0, max_dst_h - dst_h);
+
+		return min_vscale;
+	}
+
+	if (vscale > max_vscale) {
+		int max_src_h = dst_h * max_vscale;
+
+		drm_rect_adjust_size(src, 0, max_src_h - src_h);
+
+		return max_vscale;
+	}
+
+	return vscale;
+}
+EXPORT_SYMBOL(drm_rect_calc_vscale_relaxed);
+
+/**
+ * drm_rect_debug_print - print the rectangle information
+ * @r: rectangle to print
+ * @fixed_point: rectangle is in 16.16 fixed point format
+ */
+void drm_rect_debug_print(const struct drm_rect *r, bool fixed_point)
+{
+	int w = drm_rect_width(r);
+	int h = drm_rect_height(r);
+
+	if (fixed_point)
+		DRM_DEBUG_KMS("%d.%06ux%d.%06u%+d.%06u%+d.%06u\n",
+			      w >> 16, ((w & 0xffff) * 15625) >> 10,
+			      h >> 16, ((h & 0xffff) * 15625) >> 10,
+			      r->x1 >> 16, ((r->x1 & 0xffff) * 15625) >> 10,
+			      r->y1 >> 16, ((r->y1 & 0xffff) * 15625) >> 10);
+	else
+		DRM_DEBUG_KMS("%dx%d%+d%+d\n", w, h, r->x1, r->y1);
+}
+EXPORT_SYMBOL(drm_rect_debug_print);
diff --git a/drivers/gpu/drm/drm_stub.c b/drivers/gpu/drm/drm_stub.c
index 16f3ec579b3b..327ca19cda85 100644
--- a/drivers/gpu/drm/drm_stub.c
+++ b/drivers/gpu/drm/drm_stub.c
@@ -203,7 +203,7 @@ EXPORT_SYMBOL(drm_master_put);
 int drm_setmaster_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *file_priv)
 {
-	int ret;
+	int ret = 0;
 
 	if (file_priv->is_master)
 		return 0;
@@ -229,7 +229,7 @@ int drm_setmaster_ioctl(struct drm_device *dev, void *data,
 	}
 	mutex_unlock(&dev->struct_mutex);
 
-	return 0;
+	return ret;
 }
 
 int drm_dropmaster_ioctl(struct drm_device *dev, void *data,
@@ -451,14 +451,8 @@ void drm_put_dev(struct drm_device *dev)
 
 	drm_lastclose(dev);
 
-	if (drm_core_has_MTRR(dev) && drm_core_has_AGP(dev) &&
-	    dev->agp && dev->agp->agp_mtrr >= 0) {
-		int retval;
-		retval = mtrr_del(dev->agp->agp_mtrr,
-				  dev->agp->agp_info.aper_base,
-				  dev->agp->agp_info.aper_size * 1024 * 1024);
-		DRM_DEBUG("mtrr_del=%d\n", retval);
-	}
+	if (drm_core_has_MTRR(dev) && drm_core_has_AGP(dev) && dev->agp)
+		arch_phys_wc_del(dev->agp->agp_mtrr);
 
 	if (dev->driver->unload)
 		dev->driver->unload(dev);
diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c
index 02296653a058..2290b3b73832 100644
--- a/drivers/gpu/drm/drm_sysfs.c
+++ b/drivers/gpu/drm/drm_sysfs.c
@@ -30,14 +30,14 @@ static struct device_type drm_sysfs_device_minor = {
 };
 
 /**
- * drm_class_suspend - DRM class suspend hook
+ * __drm_class_suspend - internal DRM class suspend routine
  * @dev: Linux device to suspend
  * @state: power state to enter
  *
  * Just figures out what the actual struct drm_device associated with
  * @dev is and calls its suspend hook, if present.
  */
-static int drm_class_suspend(struct device *dev, pm_message_t state)
+static int __drm_class_suspend(struct device *dev, pm_message_t state)
 {
 	if (dev->type == &drm_sysfs_device_minor) {
 		struct drm_minor *drm_minor = to_drm_minor(dev);
@@ -52,6 +52,26 @@ static int drm_class_suspend(struct device *dev, pm_message_t state)
 }
 
 /**
+ * drm_class_suspend - internal DRM class suspend hook. Simply calls
+ * __drm_class_suspend() with the correct pm state.
+ * @dev: Linux device to suspend
+ */
+static int drm_class_suspend(struct device *dev)
+{
+	return __drm_class_suspend(dev, PMSG_SUSPEND);
+}
+
+/**
+ * drm_class_freeze - internal DRM class freeze hook. Simply calls
+ * __drm_class_suspend() with the correct pm state.
+ * @dev: Linux device to freeze
+ */
+static int drm_class_freeze(struct device *dev)
+{
+	return __drm_class_suspend(dev, PMSG_FREEZE);
+}
+
+/**
  * drm_class_resume - DRM class resume hook
  * @dev: Linux device to resume
  *
@@ -72,6 +92,12 @@ static int drm_class_resume(struct device *dev)
 	return 0;
 }
 
+static const struct dev_pm_ops drm_class_dev_pm_ops = {
+	.suspend	= drm_class_suspend,
+	.resume		= drm_class_resume,
+	.freeze		= drm_class_freeze,
+};
+
 static char *drm_devnode(struct device *dev, umode_t *mode)
 {
 	return kasprintf(GFP_KERNEL, "dri/%s", dev_name(dev));
@@ -106,8 +132,7 @@ struct class *drm_sysfs_create(struct module *owner, char *name)
 		goto err_out;
 	}
 
-	class->suspend = drm_class_suspend;
-	class->resume = drm_class_resume;
+	class->pm = &drm_class_dev_pm_ops;
 
 	err = class_create_file(class, &class_attr_version.attr);
 	if (err)
diff --git a/drivers/gpu/drm/drm_trace.h b/drivers/gpu/drm/drm_trace.h
index 03ea964aa604..27cc95f36381 100644
--- a/drivers/gpu/drm/drm_trace.h
+++ b/drivers/gpu/drm/drm_trace.h
@@ -21,7 +21,7 @@ TRACE_EVENT(drm_vblank_event,
 		    __entry->crtc = crtc;
 		    __entry->seq = seq;
 		    ),
-	    TP_printk("crtc=%d, seq=%d", __entry->crtc, __entry->seq)
+	    TP_printk("crtc=%d, seq=%u", __entry->crtc, __entry->seq)
 );
 
 TRACE_EVENT(drm_vblank_event_queued,
@@ -37,7 +37,7 @@ TRACE_EVENT(drm_vblank_event_queued,
 		    __entry->crtc = crtc;
 		    __entry->seq = seq;
 		    ),
-	    TP_printk("pid=%d, crtc=%d, seq=%d", __entry->pid, __entry->crtc, \
+	    TP_printk("pid=%d, crtc=%d, seq=%u", __entry->pid, __entry->crtc, \
 		      __entry->seq)
 );
 
@@ -54,7 +54,7 @@ TRACE_EVENT(drm_vblank_event_delivered,
 		    __entry->crtc = crtc;
 		    __entry->seq = seq;
 		    ),
-	    TP_printk("pid=%d, crtc=%d, seq=%d", __entry->pid, __entry->crtc, \
+	    TP_printk("pid=%d, crtc=%d, seq=%u", __entry->pid, __entry->crtc, \
 		      __entry->seq)
 );
 
diff --git a/drivers/gpu/drm/drm_vm.c b/drivers/gpu/drm/drm_vm.c
index 67969e25d60f..feb20035b2c4 100644
--- a/drivers/gpu/drm/drm_vm.c
+++ b/drivers/gpu/drm/drm_vm.c
@@ -43,18 +43,19 @@
 static void drm_vm_open(struct vm_area_struct *vma);
 static void drm_vm_close(struct vm_area_struct *vma);
 
-static pgprot_t drm_io_prot(uint32_t map_type, struct vm_area_struct *vma)
+static pgprot_t drm_io_prot(struct drm_local_map *map,
+			    struct vm_area_struct *vma)
 {
 	pgprot_t tmp = vm_get_page_prot(vma->vm_flags);
 
 #if defined(__i386__) || defined(__x86_64__)
-	if (boot_cpu_data.x86 > 3 && map_type != _DRM_AGP) {
-		pgprot_val(tmp) |= _PAGE_PCD;
-		pgprot_val(tmp) &= ~_PAGE_PWT;
-	}
+	if (map->type == _DRM_REGISTERS && !(map->flags & _DRM_WRITE_COMBINING))
+		tmp = pgprot_noncached(tmp);
+	else
+		tmp = pgprot_writecombine(tmp);
 #elif defined(__powerpc__)
 	pgprot_val(tmp) |= _PAGE_NO_CACHE;
-	if (map_type == _DRM_REGISTERS)
+	if (map->type == _DRM_REGISTERS)
 		pgprot_val(tmp) |= _PAGE_GUARDED;
 #elif defined(__ia64__)
 	if (efi_range_is_wc(vma->vm_start, vma->vm_end -
@@ -250,13 +251,8 @@ static void drm_vm_shm_close(struct vm_area_struct *vma)
 			switch (map->type) {
 			case _DRM_REGISTERS:
 			case _DRM_FRAME_BUFFER:
-				if (drm_core_has_MTRR(dev) && map->mtrr >= 0) {
-					int retcode;
-					retcode = mtrr_del(map->mtrr,
-							   map->offset,
-							   map->size);
-					DRM_DEBUG("mtrr_del = %d\n", retcode);
-				}
+				if (drm_core_has_MTRR(dev))
+					arch_phys_wc_del(map->mtrr);
 				iounmap(map->handle);
 				break;
 			case _DRM_SHM:
@@ -617,7 +613,7 @@ int drm_mmap_locked(struct file *filp, struct vm_area_struct *vma)
 	case _DRM_FRAME_BUFFER:
 	case _DRM_REGISTERS:
 		offset = drm_core_get_reg_ofs(dev);
-		vma->vm_page_prot = drm_io_prot(map->type, vma);
+		vma->vm_page_prot = drm_io_prot(map, vma);
 		if (io_remap_pfn_range(vma, vma->vm_start,
 				       (map->offset + offset) >> PAGE_SHIFT,
 				       vma->vm_end - vma->vm_start,
diff --git a/drivers/gpu/drm/exynos/exynos_ddc.c b/drivers/gpu/drm/exynos/exynos_ddc.c
index 4e9b5ba8edff..95c75edef01a 100644
--- a/drivers/gpu/drm/exynos/exynos_ddc.c
+++ b/drivers/gpu/drm/exynos/exynos_ddc.c
@@ -53,6 +53,8 @@ static struct of_device_id hdmiddc_match_types[] = {
 	{
 		.compatible = "samsung,exynos5-hdmiddc",
 	}, {
+		.compatible = "samsung,exynos4210-hdmiddc",
+	}, {
 		/* end node */
 	}
 };
diff --git a/drivers/gpu/drm/exynos/exynos_drm_buf.c b/drivers/gpu/drm/exynos/exynos_drm_buf.c
index 57affae9568b..b8ac06d92fbf 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_buf.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_buf.c
@@ -24,8 +24,6 @@ static int lowlevel_buffer_allocate(struct drm_device *dev,
 	enum dma_attr attr;
 	unsigned int nr_pages;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (buf->dma_addr) {
 		DRM_DEBUG_KMS("already allocated.\n");
 		return 0;
@@ -59,8 +57,7 @@ static int lowlevel_buffer_allocate(struct drm_device *dev,
 		dma_addr_t start_addr;
 		unsigned int i = 0;
 
-		buf->pages = kzalloc(sizeof(struct page) * nr_pages,
-					GFP_KERNEL);
+		buf->pages = drm_calloc_large(nr_pages, sizeof(struct page *));
 		if (!buf->pages) {
 			DRM_ERROR("failed to allocate pages.\n");
 			return -ENOMEM;
@@ -71,8 +68,8 @@ static int lowlevel_buffer_allocate(struct drm_device *dev,
 					&buf->dma_attrs);
 		if (!buf->kvaddr) {
 			DRM_ERROR("failed to allocate buffer.\n");
-			kfree(buf->pages);
-			return -ENOMEM;
+			ret = -ENOMEM;
+			goto err_free;
 		}
 
 		start_addr = buf->dma_addr;
@@ -109,9 +106,9 @@ err_free_attrs:
 	dma_free_attrs(dev->dev, buf->size, buf->pages,
 			(dma_addr_t)buf->dma_addr, &buf->dma_attrs);
 	buf->dma_addr = (dma_addr_t)NULL;
-
+err_free:
 	if (!is_drm_iommu_supported(dev))
-		kfree(buf->pages);
+		drm_free_large(buf->pages);
 
 	return ret;
 }
@@ -119,8 +116,6 @@ err_free_attrs:
 static void lowlevel_buffer_deallocate(struct drm_device *dev,
 		unsigned int flags, struct exynos_drm_gem_buf *buf)
 {
-	DRM_DEBUG_KMS("%s.\n", __FILE__);
-
 	if (!buf->dma_addr) {
 		DRM_DEBUG_KMS("dma_addr is invalid.\n");
 		return;
@@ -138,7 +133,7 @@ static void lowlevel_buffer_deallocate(struct drm_device *dev,
 	if (!is_drm_iommu_supported(dev)) {
 		dma_free_attrs(dev->dev, buf->size, buf->kvaddr,
 				(dma_addr_t)buf->dma_addr, &buf->dma_attrs);
-		kfree(buf->pages);
+		drm_free_large(buf->pages);
 	} else
 		dma_free_attrs(dev->dev, buf->size, buf->pages,
 				(dma_addr_t)buf->dma_addr, &buf->dma_attrs);
@@ -151,7 +146,6 @@ struct exynos_drm_gem_buf *exynos_drm_init_buf(struct drm_device *dev,
 {
 	struct exynos_drm_gem_buf *buffer;
 
-	DRM_DEBUG_KMS("%s.\n", __FILE__);
 	DRM_DEBUG_KMS("desired size = 0x%x\n", size);
 
 	buffer = kzalloc(sizeof(*buffer), GFP_KERNEL);
@@ -167,8 +161,6 @@ struct exynos_drm_gem_buf *exynos_drm_init_buf(struct drm_device *dev,
 void exynos_drm_fini_buf(struct drm_device *dev,
 				struct exynos_drm_gem_buf *buffer)
 {
-	DRM_DEBUG_KMS("%s.\n", __FILE__);
-
 	if (!buffer) {
 		DRM_DEBUG_KMS("buffer is null.\n");
 		return;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_connector.c b/drivers/gpu/drm/exynos/exynos_drm_connector.c
index 8bcc13ac9f73..02a8bc5226ca 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_connector.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_connector.c
@@ -34,7 +34,6 @@ convert_to_display_mode(struct drm_display_mode *mode,
 			struct exynos_drm_panel_info *panel)
 {
 	struct fb_videomode *timing = &panel->timing;
-	DRM_DEBUG_KMS("%s\n", __FILE__);
 
 	mode->clock = timing->pixclock / 1000;
 	mode->vrefresh = timing->refresh;
@@ -58,37 +57,6 @@ convert_to_display_mode(struct drm_display_mode *mode,
 		mode->flags |= DRM_MODE_FLAG_DBLSCAN;
 }
 
-/* convert drm_display_mode to exynos_video_timings */
-static inline void
-convert_to_video_timing(struct fb_videomode *timing,
-			struct drm_display_mode *mode)
-{
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
-	memset(timing, 0, sizeof(*timing));
-
-	timing->pixclock = mode->clock * 1000;
-	timing->refresh = drm_mode_vrefresh(mode);
-
-	timing->xres = mode->hdisplay;
-	timing->right_margin = mode->hsync_start - mode->hdisplay;
-	timing->hsync_len = mode->hsync_end - mode->hsync_start;
-	timing->left_margin = mode->htotal - mode->hsync_end;
-
-	timing->yres = mode->vdisplay;
-	timing->lower_margin = mode->vsync_start - mode->vdisplay;
-	timing->vsync_len = mode->vsync_end - mode->vsync_start;
-	timing->upper_margin = mode->vtotal - mode->vsync_end;
-
-	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-		timing->vmode = FB_VMODE_INTERLACED;
-	else
-		timing->vmode = FB_VMODE_NONINTERLACED;
-
-	if (mode->flags & DRM_MODE_FLAG_DBLSCAN)
-		timing->vmode |= FB_VMODE_DOUBLE;
-}
-
 static int exynos_drm_connector_get_modes(struct drm_connector *connector)
 {
 	struct exynos_drm_connector *exynos_connector =
@@ -99,8 +67,6 @@ static int exynos_drm_connector_get_modes(struct drm_connector *connector)
 	unsigned int count = 0;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!display_ops) {
 		DRM_DEBUG_KMS("display_ops is null.\n");
 		return 0;
@@ -168,15 +134,12 @@ static int exynos_drm_connector_mode_valid(struct drm_connector *connector,
 					to_exynos_connector(connector);
 	struct exynos_drm_manager *manager = exynos_connector->manager;
 	struct exynos_drm_display_ops *display_ops = manager->display_ops;
-	struct fb_videomode timing;
 	int ret = MODE_BAD;
 
 	DRM_DEBUG_KMS("%s\n", __FILE__);
 
-	convert_to_video_timing(&timing, mode);
-
-	if (display_ops && display_ops->check_timing)
-		if (!display_ops->check_timing(manager->dev, (void *)&timing))
+	if (display_ops && display_ops->check_mode)
+		if (!display_ops->check_mode(manager->dev, mode))
 			ret = MODE_OK;
 
 	return ret;
@@ -190,8 +153,6 @@ struct drm_encoder *exynos_drm_best_encoder(struct drm_connector *connector)
 	struct drm_mode_object *obj;
 	struct drm_encoder *encoder;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	obj = drm_mode_object_find(dev, exynos_connector->encoder_id,
 				   DRM_MODE_OBJECT_ENCODER);
 	if (!obj) {
@@ -234,8 +195,6 @@ void exynos_drm_display_power(struct drm_connector *connector, int mode)
 static void exynos_drm_connector_dpms(struct drm_connector *connector,
 					int mode)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * in case that drm_crtc_helper_set_mode() is called,
 	 * encoder/crtc->funcs->dpms() will be just returned
@@ -282,8 +241,6 @@ exynos_drm_connector_detect(struct drm_connector *connector, bool force)
 					manager->display_ops;
 	enum drm_connector_status status = connector_status_disconnected;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (display_ops && display_ops->is_connected) {
 		if (display_ops->is_connected(manager->dev))
 			status = connector_status_connected;
@@ -299,8 +256,6 @@ static void exynos_drm_connector_destroy(struct drm_connector *connector)
 	struct exynos_drm_connector *exynos_connector =
 		to_exynos_connector(connector);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	drm_sysfs_connector_remove(connector);
 	drm_connector_cleanup(connector);
 	kfree(exynos_connector);
@@ -322,8 +277,6 @@ struct drm_connector *exynos_drm_connector_create(struct drm_device *dev,
 	int type;
 	int err;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_connector = kzalloc(sizeof(*exynos_connector), GFP_KERNEL);
 	if (!exynos_connector) {
 		DRM_ERROR("failed to allocate connector\n");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_core.c b/drivers/gpu/drm/exynos/exynos_drm_core.c
index 4667c9f67acd..1bef6dc77478 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_core.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_core.c
@@ -27,8 +27,6 @@ static int exynos_drm_create_enc_conn(struct drm_device *dev,
 	struct drm_connector *connector;
 	int ret;
 
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	subdrv->manager->dev = subdrv->dev;
 
 	/* create and initialize a encoder for this sub driver. */
@@ -102,8 +100,6 @@ static int exynos_drm_subdrv_probe(struct drm_device *dev,
 static void exynos_drm_subdrv_remove(struct drm_device *dev,
 				      struct exynos_drm_subdrv *subdrv)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	if (subdrv->remove)
 		subdrv->remove(dev, subdrv->dev);
 }
@@ -114,8 +110,6 @@ int exynos_drm_device_register(struct drm_device *dev)
 	unsigned int fine_cnt = 0;
 	int err;
 
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	if (!dev)
 		return -EINVAL;
 
@@ -158,8 +152,6 @@ int exynos_drm_device_unregister(struct drm_device *dev)
 {
 	struct exynos_drm_subdrv *subdrv;
 
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	if (!dev) {
 		WARN(1, "Unexpected drm device unregister!\n");
 		return -EINVAL;
@@ -176,8 +168,6 @@ EXPORT_SYMBOL_GPL(exynos_drm_device_unregister);
 
 int exynos_drm_subdrv_register(struct exynos_drm_subdrv *subdrv)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	if (!subdrv)
 		return -EINVAL;
 
@@ -189,8 +179,6 @@ EXPORT_SYMBOL_GPL(exynos_drm_subdrv_register);
 
 int exynos_drm_subdrv_unregister(struct exynos_drm_subdrv *subdrv)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	if (!subdrv)
 		return -EINVAL;
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_crtc.c b/drivers/gpu/drm/exynos/exynos_drm_crtc.c
index c200e4d71e3d..9a35d171a6d3 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_crtc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_crtc.c
@@ -76,8 +76,6 @@ static void exynos_drm_crtc_dpms(struct drm_crtc *crtc, int mode)
 
 static void exynos_drm_crtc_prepare(struct drm_crtc *crtc)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* drm framework doesn't check NULL. */
 }
 
@@ -85,8 +83,6 @@ static void exynos_drm_crtc_commit(struct drm_crtc *crtc)
 {
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_drm_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
 	exynos_plane_commit(exynos_crtc->plane);
 	exynos_plane_dpms(exynos_crtc->plane, DRM_MODE_DPMS_ON);
@@ -97,8 +93,6 @@ exynos_drm_crtc_mode_fixup(struct drm_crtc *crtc,
 			    const struct drm_display_mode *mode,
 			    struct drm_display_mode *adjusted_mode)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* drm framework doesn't check NULL */
 	return true;
 }
@@ -115,8 +109,6 @@ exynos_drm_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	int pipe = exynos_crtc->pipe;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * copy the mode data adjusted by mode_fixup() into crtc->mode
 	 * so that hardware can be seet to proper mode.
@@ -139,7 +131,7 @@ exynos_drm_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	return 0;
 }
 
-static int exynos_drm_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
+static int exynos_drm_crtc_mode_set_commit(struct drm_crtc *crtc, int x, int y,
 					  struct drm_framebuffer *old_fb)
 {
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
@@ -148,8 +140,6 @@ static int exynos_drm_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 	unsigned int crtc_h;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* when framebuffer changing is requested, crtc's dpms should be on */
 	if (exynos_crtc->dpms > DRM_MODE_DPMS_ON) {
 		DRM_ERROR("failed framebuffer changing request.\n");
@@ -169,18 +159,16 @@ static int exynos_drm_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 	return 0;
 }
 
-static void exynos_drm_crtc_load_lut(struct drm_crtc *crtc)
+static int exynos_drm_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
+					  struct drm_framebuffer *old_fb)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-	/* drm framework doesn't check NULL */
+	return exynos_drm_crtc_mode_set_commit(crtc, x, y, old_fb);
 }
 
 static void exynos_drm_crtc_disable(struct drm_crtc *crtc)
 {
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_plane_dpms(exynos_crtc->plane, DRM_MODE_DPMS_OFF);
 	exynos_drm_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
 }
@@ -192,7 +180,6 @@ static struct drm_crtc_helper_funcs exynos_crtc_helper_funcs = {
 	.mode_fixup	= exynos_drm_crtc_mode_fixup,
 	.mode_set	= exynos_drm_crtc_mode_set,
 	.mode_set_base	= exynos_drm_crtc_mode_set_base,
-	.load_lut	= exynos_drm_crtc_load_lut,
 	.disable	= exynos_drm_crtc_disable,
 };
 
@@ -206,8 +193,6 @@ static int exynos_drm_crtc_page_flip(struct drm_crtc *crtc,
 	struct drm_framebuffer *old_fb = crtc->fb;
 	int ret = -EINVAL;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* when the page flip is requested, crtc's dpms should be on */
 	if (exynos_crtc->dpms > DRM_MODE_DPMS_ON) {
 		DRM_ERROR("failed page flip request.\n");
@@ -237,7 +222,7 @@ static int exynos_drm_crtc_page_flip(struct drm_crtc *crtc,
 		spin_unlock_irq(&dev->event_lock);
 
 		crtc->fb = fb;
-		ret = exynos_drm_crtc_mode_set_base(crtc, crtc->x, crtc->y,
+		ret = exynos_drm_crtc_mode_set_commit(crtc, crtc->x, crtc->y,
 						    NULL);
 		if (ret) {
 			crtc->fb = old_fb;
@@ -260,8 +245,6 @@ static void exynos_drm_crtc_destroy(struct drm_crtc *crtc)
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
 	struct exynos_drm_private *private = crtc->dev->dev_private;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	private->crtc[exynos_crtc->pipe] = NULL;
 
 	drm_crtc_cleanup(crtc);
@@ -276,8 +259,6 @@ static int exynos_drm_crtc_set_property(struct drm_crtc *crtc,
 	struct exynos_drm_private *dev_priv = dev->dev_private;
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(crtc);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (property == dev_priv->crtc_mode_property) {
 		enum exynos_crtc_mode mode = val;
 
@@ -322,8 +303,6 @@ static void exynos_drm_crtc_attach_mode_property(struct drm_crtc *crtc)
 	struct exynos_drm_private *dev_priv = dev->dev_private;
 	struct drm_property *prop;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	prop = dev_priv->crtc_mode_property;
 	if (!prop) {
 		prop = drm_property_create_enum(dev, 0, "mode", mode_names,
@@ -343,8 +322,6 @@ int exynos_drm_crtc_create(struct drm_device *dev, unsigned int nr)
 	struct exynos_drm_private *private = dev->dev_private;
 	struct drm_crtc *crtc;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_crtc = kzalloc(sizeof(*exynos_crtc), GFP_KERNEL);
 	if (!exynos_crtc) {
 		DRM_ERROR("failed to allocate exynos crtc\n");
@@ -379,8 +356,6 @@ int exynos_drm_crtc_enable_vblank(struct drm_device *dev, int crtc)
 	struct exynos_drm_crtc *exynos_crtc =
 		to_exynos_crtc(private->crtc[crtc]);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (exynos_crtc->dpms != DRM_MODE_DPMS_ON)
 		return -EPERM;
 
@@ -396,8 +371,6 @@ void exynos_drm_crtc_disable_vblank(struct drm_device *dev, int crtc)
 	struct exynos_drm_crtc *exynos_crtc =
 		to_exynos_crtc(private->crtc[crtc]);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (exynos_crtc->dpms != DRM_MODE_DPMS_ON)
 		return;
 
@@ -413,8 +386,6 @@ void exynos_drm_crtc_finish_pageflip(struct drm_device *dev, int crtc)
 	struct exynos_drm_crtc *exynos_crtc = to_exynos_crtc(drm_crtc);
 	unsigned long flags;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	spin_lock_irqsave(&dev->event_lock, flags);
 
 	list_for_each_entry_safe(e, t, &dev_priv->pageflip_event_list,
diff --git a/drivers/gpu/drm/exynos/exynos_drm_dmabuf.c b/drivers/gpu/drm/exynos/exynos_drm_dmabuf.c
index ff7f2a886a34..a0f997e0cbdf 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_dmabuf.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_dmabuf.c
@@ -71,8 +71,6 @@ static struct sg_table *
 	unsigned int i;
 	int nents, ret;
 
-	DRM_DEBUG_PRIME("%s\n", __FILE__);
-
 	/* just return current sgt if already requested. */
 	if (exynos_attach->dir == dir && exynos_attach->is_mapped)
 		return &exynos_attach->sgt;
@@ -133,8 +131,6 @@ static void exynos_dmabuf_release(struct dma_buf *dmabuf)
 {
 	struct exynos_drm_gem_obj *exynos_gem_obj = dmabuf->priv;
 
-	DRM_DEBUG_PRIME("%s\n", __FILE__);
-
 	/*
 	 * exynos_dmabuf_release() call means that file object's
 	 * f_count is 0 and it calls drm_gem_object_handle_unreference()
@@ -219,8 +215,6 @@ struct drm_gem_object *exynos_dmabuf_prime_import(struct drm_device *drm_dev,
 	struct exynos_drm_gem_buf *buffer;
 	int ret;
 
-	DRM_DEBUG_PRIME("%s\n", __FILE__);
-
 	/* is this one of own objects? */
 	if (dma_buf->ops == &exynos_dmabuf_ops) {
 		struct drm_gem_object *obj;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c
index ba6d995e4375..ca2729a85129 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c
@@ -46,8 +46,6 @@ static int exynos_drm_load(struct drm_device *dev, unsigned long flags)
 	int ret;
 	int nr;
 
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	private = kzalloc(sizeof(struct exynos_drm_private), GFP_KERNEL);
 	if (!private) {
 		DRM_ERROR("failed to allocate private\n");
@@ -140,8 +138,6 @@ err_crtc:
 
 static int exynos_drm_unload(struct drm_device *dev)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	exynos_drm_fbdev_fini(dev);
 	exynos_drm_device_unregister(dev);
 	drm_vblank_cleanup(dev);
@@ -159,8 +155,7 @@ static int exynos_drm_unload(struct drm_device *dev)
 static int exynos_drm_open(struct drm_device *dev, struct drm_file *file)
 {
 	struct drm_exynos_file_private *file_priv;
-
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
+	int ret;
 
 	file_priv = kzalloc(sizeof(*file_priv), GFP_KERNEL);
 	if (!file_priv)
@@ -168,7 +163,13 @@ static int exynos_drm_open(struct drm_device *dev, struct drm_file *file)
 
 	file->driver_priv = file_priv;
 
-	return exynos_drm_subdrv_open(dev, file);
+	ret = exynos_drm_subdrv_open(dev, file);
+	if (ret) {
+		kfree(file_priv);
+		file->driver_priv = NULL;
+	}
+
+	return ret;
 }
 
 static void exynos_drm_preclose(struct drm_device *dev,
@@ -178,8 +179,6 @@ static void exynos_drm_preclose(struct drm_device *dev,
 	struct drm_pending_vblank_event *e, *t;
 	unsigned long flags;
 
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	/* release events of current file */
 	spin_lock_irqsave(&dev->event_lock, flags);
 	list_for_each_entry_safe(e, t, &private->pageflip_event_list,
@@ -196,8 +195,6 @@ static void exynos_drm_preclose(struct drm_device *dev,
 
 static void exynos_drm_postclose(struct drm_device *dev, struct drm_file *file)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	if (!file->driver_priv)
 		return;
 
@@ -207,8 +204,6 @@ static void exynos_drm_postclose(struct drm_device *dev, struct drm_file *file)
 
 static void exynos_drm_lastclose(struct drm_device *dev)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	exynos_drm_fbdev_restore_mode(dev);
 }
 
@@ -292,8 +287,6 @@ static struct drm_driver exynos_drm_driver = {
 
 static int exynos_drm_platform_probe(struct platform_device *pdev)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	pdev->dev.coherent_dma_mask = DMA_BIT_MASK(32);
 	exynos_drm_driver.num_ioctls = DRM_ARRAY_SIZE(exynos_ioctls);
 
@@ -302,8 +295,6 @@ static int exynos_drm_platform_probe(struct platform_device *pdev)
 
 static int exynos_drm_platform_remove(struct platform_device *pdev)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	drm_platform_exit(&exynos_drm_driver, pdev);
 
 	return 0;
@@ -322,8 +313,6 @@ static int __init exynos_drm_init(void)
 {
 	int ret;
 
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 #ifdef CONFIG_DRM_EXYNOS_FIMD
 	ret = platform_driver_register(&fimd_driver);
 	if (ret < 0)
@@ -455,8 +444,6 @@ out_fimd:
 
 static void __exit exynos_drm_exit(void)
 {
-	DRM_DEBUG_DRIVER("%s\n", __FILE__);
-
 	platform_device_unregister(exynos_drm_pdev);
 
 	platform_driver_unregister(&exynos_drm_platform_driver);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h
index 680a7c1b9dea..eaa19668bf00 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h
@@ -142,7 +142,7 @@ struct exynos_drm_overlay {
  * @is_connected: check for that display is connected or not.
  * @get_edid: get edid modes from display driver.
  * @get_panel: get panel object from display driver.
- * @check_timing: check if timing is valid or not.
+ * @check_mode: check if mode is valid or not.
  * @power_on: display device on or off.
  */
 struct exynos_drm_display_ops {
@@ -151,7 +151,7 @@ struct exynos_drm_display_ops {
 	struct edid *(*get_edid)(struct device *dev,
 			struct drm_connector *connector);
 	void *(*get_panel)(struct device *dev);
-	int (*check_timing)(struct device *dev, void *timing);
+	int (*check_mode)(struct device *dev, struct drm_display_mode *mode);
 	int (*power_on)(struct device *dev, int mode);
 };
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_encoder.c b/drivers/gpu/drm/exynos/exynos_drm_encoder.c
index c63721f64aec..a99a033793bc 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_encoder.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_encoder.c
@@ -61,7 +61,7 @@ static void exynos_drm_encoder_dpms(struct drm_encoder *encoder, int mode)
 	struct exynos_drm_manager_ops *manager_ops = manager->ops;
 	struct exynos_drm_encoder *exynos_encoder = to_exynos_encoder(encoder);
 
-	DRM_DEBUG_KMS("%s, encoder dpms: %d\n", __FILE__, mode);
+	DRM_DEBUG_KMS("encoder dpms: %d\n", mode);
 
 	if (exynos_encoder->dpms == mode) {
 		DRM_DEBUG_KMS("desired dpms mode is same as previous one.\n");
@@ -104,8 +104,6 @@ exynos_drm_encoder_mode_fixup(struct drm_encoder *encoder,
 	struct exynos_drm_manager *manager = exynos_drm_get_manager(encoder);
 	struct exynos_drm_manager_ops *manager_ops = manager->ops;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		if (connector->encoder == encoder)
 			if (manager_ops && manager_ops->mode_fixup)
@@ -155,8 +153,6 @@ static void exynos_drm_encoder_mode_set(struct drm_encoder *encoder,
 	struct exynos_drm_manager *manager;
 	struct exynos_drm_manager_ops *manager_ops;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		if (connector->encoder == encoder) {
 			struct exynos_drm_encoder *exynos_encoder;
@@ -189,8 +185,6 @@ static void exynos_drm_encoder_mode_set(struct drm_encoder *encoder,
 
 static void exynos_drm_encoder_prepare(struct drm_encoder *encoder)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* drm framework doesn't check NULL. */
 }
 
@@ -200,8 +194,6 @@ static void exynos_drm_encoder_commit(struct drm_encoder *encoder)
 	struct exynos_drm_manager *manager = exynos_encoder->manager;
 	struct exynos_drm_manager_ops *manager_ops = manager->ops;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (manager_ops && manager_ops->commit)
 		manager_ops->commit(manager->dev);
 
@@ -274,8 +266,6 @@ static void exynos_drm_encoder_destroy(struct drm_encoder *encoder)
 	struct exynos_drm_encoder *exynos_encoder =
 		to_exynos_encoder(encoder);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_encoder->manager->pipe = -1;
 
 	drm_encoder_cleanup(encoder);
@@ -315,8 +305,6 @@ void exynos_drm_encoder_setup(struct drm_device *dev)
 {
 	struct drm_encoder *encoder;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head)
 		encoder->possible_clones = exynos_drm_encoder_clones(encoder);
 }
@@ -329,8 +317,6 @@ exynos_drm_encoder_create(struct drm_device *dev,
 	struct drm_encoder *encoder;
 	struct exynos_drm_encoder *exynos_encoder;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!manager || !possible_crtcs)
 		return NULL;
 
@@ -427,8 +413,6 @@ void exynos_drm_encoder_crtc_dpms(struct drm_encoder *encoder, void *data)
 	struct exynos_drm_manager_ops *manager_ops = manager->ops;
 	int mode = *(int *)data;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (manager_ops && manager_ops->dpms)
 		manager_ops->dpms(manager->dev, mode);
 
@@ -449,8 +433,6 @@ void exynos_drm_encoder_crtc_pipe(struct drm_encoder *encoder, void *data)
 		to_exynos_encoder(encoder)->manager;
 	int pipe = *(int *)data;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * when crtc is detached from encoder, this pipe is used
 	 * to select manager operation
@@ -465,8 +447,6 @@ void exynos_drm_encoder_plane_mode_set(struct drm_encoder *encoder, void *data)
 	struct exynos_drm_overlay_ops *overlay_ops = manager->overlay_ops;
 	struct exynos_drm_overlay *overlay = data;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (overlay_ops && overlay_ops->mode_set)
 		overlay_ops->mode_set(manager->dev, overlay);
 }
@@ -478,8 +458,6 @@ void exynos_drm_encoder_plane_commit(struct drm_encoder *encoder, void *data)
 	struct exynos_drm_overlay_ops *overlay_ops = manager->overlay_ops;
 	int zpos = DEFAULT_ZPOS;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (data)
 		zpos = *(int *)data;
 
@@ -494,8 +472,6 @@ void exynos_drm_encoder_plane_enable(struct drm_encoder *encoder, void *data)
 	struct exynos_drm_overlay_ops *overlay_ops = manager->overlay_ops;
 	int zpos = DEFAULT_ZPOS;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (data)
 		zpos = *(int *)data;
 
@@ -510,8 +486,6 @@ void exynos_drm_encoder_plane_disable(struct drm_encoder *encoder, void *data)
 	struct exynos_drm_overlay_ops *overlay_ops = manager->overlay_ops;
 	int zpos = DEFAULT_ZPOS;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (data)
 		zpos = *(int *)data;
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fb.c b/drivers/gpu/drm/exynos/exynos_drm_fb.c
index 0e04f4ea441f..c2d149f0408a 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fb.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fb.c
@@ -70,8 +70,6 @@ static void exynos_drm_fb_destroy(struct drm_framebuffer *fb)
 	struct exynos_drm_fb *exynos_fb = to_exynos_fb(fb);
 	unsigned int i;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* make sure that overlay data are updated before relesing fb. */
 	exynos_drm_encoder_complete_scanout(fb);
 
@@ -97,8 +95,6 @@ static int exynos_drm_fb_create_handle(struct drm_framebuffer *fb,
 {
 	struct exynos_drm_fb *exynos_fb = to_exynos_fb(fb);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* This fb should have only one gem object. */
 	if (WARN_ON(exynos_fb->buf_cnt != 1))
 		return -EINVAL;
@@ -112,8 +108,6 @@ static int exynos_drm_fb_dirty(struct drm_framebuffer *fb,
 				unsigned color, struct drm_clip_rect *clips,
 				unsigned num_clips)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO */
 
 	return 0;
@@ -225,8 +219,6 @@ exynos_user_fb_create(struct drm_device *dev, struct drm_file *file_priv,
 	struct exynos_drm_fb *exynos_fb;
 	int i, ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_fb = kzalloc(sizeof(*exynos_fb), GFP_KERNEL);
 	if (!exynos_fb) {
 		DRM_ERROR("failed to allocate exynos drm framebuffer\n");
@@ -293,8 +285,6 @@ struct exynos_drm_gem_buf *exynos_drm_fb_buffer(struct drm_framebuffer *fb,
 	struct exynos_drm_fb *exynos_fb = to_exynos_fb(fb);
 	struct exynos_drm_gem_buf *buffer;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (index >= MAX_FB_BUFFER)
 		return NULL;
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
index 8f007aaeffc3..8e60bd61137f 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
@@ -43,8 +43,6 @@ static int exynos_drm_fb_mmap(struct fb_info *info,
 	unsigned long vm_size;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	vma->vm_flags |= VM_IO | VM_DONTEXPAND | VM_DONTDUMP;
 
 	vm_size = vma->vm_end - vma->vm_start;
@@ -84,8 +82,6 @@ static int exynos_drm_fbdev_update(struct drm_fb_helper *helper,
 	unsigned int size = fb->width * fb->height * (fb->bits_per_pixel >> 3);
 	unsigned long offset;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	drm_fb_helper_fill_fix(fbi, fb->pitches[0], fb->depth);
 	drm_fb_helper_fill_var(fbi, helper, fb->width, fb->height);
 
@@ -148,8 +144,6 @@ static int exynos_drm_fbdev_create(struct drm_fb_helper *helper,
 	unsigned long size;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	DRM_DEBUG_KMS("surface width(%d), height(%d) and bpp(%d\n",
 			sizes->surface_width, sizes->surface_height,
 			sizes->surface_bpp);
@@ -238,8 +232,6 @@ int exynos_drm_fbdev_init(struct drm_device *dev)
 	unsigned int num_crtc;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!dev->mode_config.num_crtc || !dev->mode_config.num_connector)
 		return 0;
 
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimc.c b/drivers/gpu/drm/exynos/exynos_drm_fimc.c
index 4a1616a18ab7..61b094f689a7 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimc.c
@@ -175,8 +175,6 @@ static void fimc_sw_reset(struct fimc_context *ctx)
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* stop dma operation */
 	cfg = fimc_read(EXYNOS_CISTATUS);
 	if (EXYNOS_CISTATUS_GET_ENVID_STATUS(cfg)) {
@@ -210,8 +208,6 @@ static void fimc_sw_reset(struct fimc_context *ctx)
 
 static int fimc_set_camblk_fimd0_wb(struct fimc_context *ctx)
 {
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	return regmap_update_bits(ctx->sysreg, SYSREG_CAMERA_BLK,
 				  SYSREG_FIMD0WB_DEST_MASK,
 				  ctx->id << SYSREG_FIMD0WB_DEST_SHIFT);
@@ -221,7 +217,7 @@ static void fimc_set_type_ctrl(struct fimc_context *ctx, enum fimc_wb wb)
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:wb[%d]\n", __func__, wb);
+	DRM_DEBUG_KMS("wb[%d]\n", wb);
 
 	cfg = fimc_read(EXYNOS_CIGCTRL);
 	cfg &= ~(EXYNOS_CIGCTRL_TESTPATTERN_MASK |
@@ -257,10 +253,10 @@ static void fimc_set_polarity(struct fimc_context *ctx,
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:inv_pclk[%d]inv_vsync[%d]\n",
-		__func__, pol->inv_pclk, pol->inv_vsync);
-	DRM_DEBUG_KMS("%s:inv_href[%d]inv_hsync[%d]\n",
-		__func__, pol->inv_href, pol->inv_hsync);
+	DRM_DEBUG_KMS("inv_pclk[%d]inv_vsync[%d]\n",
+		pol->inv_pclk, pol->inv_vsync);
+	DRM_DEBUG_KMS("inv_href[%d]inv_hsync[%d]\n",
+		pol->inv_href, pol->inv_hsync);
 
 	cfg = fimc_read(EXYNOS_CIGCTRL);
 	cfg &= ~(EXYNOS_CIGCTRL_INVPOLPCLK | EXYNOS_CIGCTRL_INVPOLVSYNC |
@@ -282,7 +278,7 @@ static void fimc_handle_jpeg(struct fimc_context *ctx, bool enable)
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:enable[%d]\n", __func__, enable);
+	DRM_DEBUG_KMS("enable[%d]\n", enable);
 
 	cfg = fimc_read(EXYNOS_CIGCTRL);
 	if (enable)
@@ -298,7 +294,7 @@ static void fimc_handle_irq(struct fimc_context *ctx, bool enable,
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:enable[%d]overflow[%d]level[%d]\n", __func__,
+	DRM_DEBUG_KMS("enable[%d]overflow[%d]level[%d]\n",
 			enable, overflow, level);
 
 	cfg = fimc_read(EXYNOS_CIGCTRL);
@@ -319,8 +315,6 @@ static void fimc_clear_irq(struct fimc_context *ctx)
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	cfg = fimc_read(EXYNOS_CIGCTRL);
 	cfg |= EXYNOS_CIGCTRL_IRQ_CLR;
 	fimc_write(cfg, EXYNOS_CIGCTRL);
@@ -335,7 +329,7 @@ static bool fimc_check_ovf(struct fimc_context *ctx)
 	flag = EXYNOS_CISTATUS_OVFIY | EXYNOS_CISTATUS_OVFICB |
 		EXYNOS_CISTATUS_OVFICR;
 
-	DRM_DEBUG_KMS("%s:flag[0x%x]\n", __func__, flag);
+	DRM_DEBUG_KMS("flag[0x%x]\n", flag);
 
 	if (status & flag) {
 		cfg = fimc_read(EXYNOS_CIWDOFST);
@@ -364,7 +358,7 @@ static bool fimc_check_frame_end(struct fimc_context *ctx)
 
 	cfg = fimc_read(EXYNOS_CISTATUS);
 
-	DRM_DEBUG_KMS("%s:cfg[0x%x]\n", __func__, cfg);
+	DRM_DEBUG_KMS("cfg[0x%x]\n", cfg);
 
 	if (!(cfg & EXYNOS_CISTATUS_FRAMEEND))
 		return false;
@@ -380,15 +374,13 @@ static int fimc_get_buf_id(struct fimc_context *ctx)
 	u32 cfg;
 	int frame_cnt, buf_id;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	cfg = fimc_read(EXYNOS_CISTATUS2);
 	frame_cnt = EXYNOS_CISTATUS2_GET_FRAMECOUNT_BEFORE(cfg);
 
 	if (frame_cnt == 0)
 		frame_cnt = EXYNOS_CISTATUS2_GET_FRAMECOUNT_PRESENT(cfg);
 
-	DRM_DEBUG_KMS("%s:present[%d]before[%d]\n", __func__,
+	DRM_DEBUG_KMS("present[%d]before[%d]\n",
 		EXYNOS_CISTATUS2_GET_FRAMECOUNT_PRESENT(cfg),
 		EXYNOS_CISTATUS2_GET_FRAMECOUNT_BEFORE(cfg));
 
@@ -398,7 +390,7 @@ static int fimc_get_buf_id(struct fimc_context *ctx)
 	}
 
 	buf_id = frame_cnt - 1;
-	DRM_DEBUG_KMS("%s:buf_id[%d]\n", __func__, buf_id);
+	DRM_DEBUG_KMS("buf_id[%d]\n", buf_id);
 
 	return buf_id;
 }
@@ -407,7 +399,7 @@ static void fimc_handle_lastend(struct fimc_context *ctx, bool enable)
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:enable[%d]\n", __func__, enable);
+	DRM_DEBUG_KMS("enable[%d]\n", enable);
 
 	cfg = fimc_read(EXYNOS_CIOCTRL);
 	if (enable)
@@ -424,7 +416,7 @@ static int fimc_src_set_fmt_order(struct fimc_context *ctx, u32 fmt)
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:fmt[0x%x]\n", __func__, fmt);
+	DRM_DEBUG_KMS("fmt[0x%x]\n", fmt);
 
 	/* RGB */
 	cfg = fimc_read(EXYNOS_CISCCTRL);
@@ -497,7 +489,7 @@ static int fimc_src_set_fmt(struct device *dev, u32 fmt)
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:fmt[0x%x]\n", __func__, fmt);
+	DRM_DEBUG_KMS("fmt[0x%x]\n", fmt);
 
 	cfg = fimc_read(EXYNOS_MSCTRL);
 	cfg &= ~EXYNOS_MSCTRL_INFORMAT_RGB;
@@ -557,8 +549,7 @@ static int fimc_src_set_transf(struct device *dev,
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg1, cfg2;
 
-	DRM_DEBUG_KMS("%s:degree[%d]flip[0x%x]\n", __func__,
-		degree, flip);
+	DRM_DEBUG_KMS("degree[%d]flip[0x%x]\n", degree, flip);
 
 	cfg1 = fimc_read(EXYNOS_MSCTRL);
 	cfg1 &= ~(EXYNOS_MSCTRL_FLIP_X_MIRROR |
@@ -621,10 +612,9 @@ static int fimc_set_window(struct fimc_context *ctx,
 	v1 = pos->y;
 	v2 = sz->vsize - pos->h - pos->y;
 
-	DRM_DEBUG_KMS("%s:x[%d]y[%d]w[%d]h[%d]hsize[%d]vsize[%d]\n",
-	__func__, pos->x, pos->y, pos->w, pos->h, sz->hsize, sz->vsize);
-	DRM_DEBUG_KMS("%s:h1[%d]h2[%d]v1[%d]v2[%d]\n", __func__,
-		h1, h2, v1, v2);
+	DRM_DEBUG_KMS("x[%d]y[%d]w[%d]h[%d]hsize[%d]vsize[%d]\n",
+		pos->x, pos->y, pos->w, pos->h, sz->hsize, sz->vsize);
+	DRM_DEBUG_KMS("h1[%d]h2[%d]v1[%d]v2[%d]\n", h1, h2, v1, v2);
 
 	/*
 	 * set window offset 1, 2 size
@@ -653,8 +643,8 @@ static int fimc_src_set_size(struct device *dev, int swap,
 	struct drm_exynos_sz img_sz = *sz;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:swap[%d]hsize[%d]vsize[%d]\n",
-		__func__, swap, sz->hsize, sz->vsize);
+	DRM_DEBUG_KMS("swap[%d]hsize[%d]vsize[%d]\n",
+		swap, sz->hsize, sz->vsize);
 
 	/* original size */
 	cfg = (EXYNOS_ORGISIZE_HORIZONTAL(img_sz.hsize) |
@@ -662,8 +652,7 @@ static int fimc_src_set_size(struct device *dev, int swap,
 
 	fimc_write(cfg, EXYNOS_ORGISIZE);
 
-	DRM_DEBUG_KMS("%s:x[%d]y[%d]w[%d]h[%d]\n", __func__,
-		pos->x, pos->y, pos->w, pos->h);
+	DRM_DEBUG_KMS("x[%d]y[%d]w[%d]h[%d]\n", pos->x, pos->y, pos->w, pos->h);
 
 	if (swap) {
 		img_pos.w = pos->h;
@@ -720,7 +709,7 @@ static int fimc_src_set_addr(struct device *dev,
 
 	property = &c_node->property;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]buf_id[%d]buf_type[%d]\n", __func__,
+	DRM_DEBUG_KMS("prop_id[%d]buf_id[%d]buf_type[%d]\n",
 		property->prop_id, buf_id, buf_type);
 
 	if (buf_id > FIMC_MAX_SRC) {
@@ -772,7 +761,7 @@ static int fimc_dst_set_fmt_order(struct fimc_context *ctx, u32 fmt)
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:fmt[0x%x]\n", __func__, fmt);
+	DRM_DEBUG_KMS("fmt[0x%x]\n", fmt);
 
 	/* RGB */
 	cfg = fimc_read(EXYNOS_CISCCTRL);
@@ -851,7 +840,7 @@ static int fimc_dst_set_fmt(struct device *dev, u32 fmt)
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:fmt[0x%x]\n", __func__, fmt);
+	DRM_DEBUG_KMS("fmt[0x%x]\n", fmt);
 
 	cfg = fimc_read(EXYNOS_CIEXTEN);
 
@@ -919,8 +908,7 @@ static int fimc_dst_set_transf(struct device *dev,
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:degree[%d]flip[0x%x]\n", __func__,
-		degree, flip);
+	DRM_DEBUG_KMS("degree[%d]flip[0x%x]\n", degree, flip);
 
 	cfg = fimc_read(EXYNOS_CITRGFMT);
 	cfg &= ~EXYNOS_CITRGFMT_FLIP_MASK;
@@ -970,7 +958,7 @@ static int fimc_dst_set_transf(struct device *dev,
 
 static int fimc_get_ratio_shift(u32 src, u32 dst, u32 *ratio, u32 *shift)
 {
-	DRM_DEBUG_KMS("%s:src[%d]dst[%d]\n", __func__, src, dst);
+	DRM_DEBUG_KMS("src[%d]dst[%d]\n", src, dst);
 
 	if (src >= dst * 64) {
 		DRM_ERROR("failed to make ratio and shift.\n");
@@ -1039,20 +1027,20 @@ static int fimc_set_prescaler(struct fimc_context *ctx, struct fimc_scaler *sc,
 
 	pre_dst_width = src_w / pre_hratio;
 	pre_dst_height = src_h / pre_vratio;
-	DRM_DEBUG_KMS("%s:pre_dst_width[%d]pre_dst_height[%d]\n", __func__,
+	DRM_DEBUG_KMS("pre_dst_width[%d]pre_dst_height[%d]\n",
 		pre_dst_width, pre_dst_height);
-	DRM_DEBUG_KMS("%s:pre_hratio[%d]hfactor[%d]pre_vratio[%d]vfactor[%d]\n",
-		__func__, pre_hratio, hfactor, pre_vratio, vfactor);
+	DRM_DEBUG_KMS("pre_hratio[%d]hfactor[%d]pre_vratio[%d]vfactor[%d]\n",
+		pre_hratio, hfactor, pre_vratio, vfactor);
 
 	sc->hratio = (src_w << 14) / (dst_w << hfactor);
 	sc->vratio = (src_h << 14) / (dst_h << vfactor);
 	sc->up_h = (dst_w >= src_w) ? true : false;
 	sc->up_v = (dst_h >= src_h) ? true : false;
-	DRM_DEBUG_KMS("%s:hratio[%d]vratio[%d]up_h[%d]up_v[%d]\n",
-	__func__, sc->hratio, sc->vratio, sc->up_h, sc->up_v);
+	DRM_DEBUG_KMS("hratio[%d]vratio[%d]up_h[%d]up_v[%d]\n",
+		sc->hratio, sc->vratio, sc->up_h, sc->up_v);
 
 	shfactor = FIMC_SHFACTOR - (hfactor + vfactor);
-	DRM_DEBUG_KMS("%s:shfactor[%d]\n", __func__, shfactor);
+	DRM_DEBUG_KMS("shfactor[%d]\n", shfactor);
 
 	cfg = (EXYNOS_CISCPRERATIO_SHFACTOR(shfactor) |
 		EXYNOS_CISCPRERATIO_PREHORRATIO(pre_hratio) |
@@ -1070,10 +1058,10 @@ static void fimc_set_scaler(struct fimc_context *ctx, struct fimc_scaler *sc)
 {
 	u32 cfg, cfg_ext;
 
-	DRM_DEBUG_KMS("%s:range[%d]bypass[%d]up_h[%d]up_v[%d]\n",
-		__func__, sc->range, sc->bypass, sc->up_h, sc->up_v);
-	DRM_DEBUG_KMS("%s:hratio[%d]vratio[%d]\n",
-		__func__, sc->hratio, sc->vratio);
+	DRM_DEBUG_KMS("range[%d]bypass[%d]up_h[%d]up_v[%d]\n",
+		sc->range, sc->bypass, sc->up_h, sc->up_v);
+	DRM_DEBUG_KMS("hratio[%d]vratio[%d]\n",
+		sc->hratio, sc->vratio);
 
 	cfg = fimc_read(EXYNOS_CISCCTRL);
 	cfg &= ~(EXYNOS_CISCCTRL_SCALERBYPASS |
@@ -1113,8 +1101,8 @@ static int fimc_dst_set_size(struct device *dev, int swap,
 	struct drm_exynos_sz img_sz = *sz;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:swap[%d]hsize[%d]vsize[%d]\n",
-		__func__, swap, sz->hsize, sz->vsize);
+	DRM_DEBUG_KMS("swap[%d]hsize[%d]vsize[%d]\n",
+		swap, sz->hsize, sz->vsize);
 
 	/* original size */
 	cfg = (EXYNOS_ORGOSIZE_HORIZONTAL(img_sz.hsize) |
@@ -1122,8 +1110,7 @@ static int fimc_dst_set_size(struct device *dev, int swap,
 
 	fimc_write(cfg, EXYNOS_ORGOSIZE);
 
-	DRM_DEBUG_KMS("%s:x[%d]y[%d]w[%d]h[%d]\n",
-		__func__, pos->x, pos->y, pos->w, pos->h);
+	DRM_DEBUG_KMS("x[%d]y[%d]w[%d]h[%d]\n", pos->x, pos->y, pos->w, pos->h);
 
 	/* CSC ITU */
 	cfg = fimc_read(EXYNOS_CIGCTRL);
@@ -1180,7 +1167,7 @@ static int fimc_dst_get_buf_seq(struct fimc_context *ctx)
 		if (cfg & (mask << i))
 			buf_num++;
 
-	DRM_DEBUG_KMS("%s:buf_num[%d]\n", __func__, buf_num);
+	DRM_DEBUG_KMS("buf_num[%d]\n", buf_num);
 
 	return buf_num;
 }
@@ -1194,8 +1181,7 @@ static int fimc_dst_set_buf_seq(struct fimc_context *ctx, u32 buf_id,
 	u32 mask = 0x00000001 << buf_id;
 	int ret = 0;
 
-	DRM_DEBUG_KMS("%s:buf_id[%d]buf_type[%d]\n", __func__,
-		buf_id, buf_type);
+	DRM_DEBUG_KMS("buf_id[%d]buf_type[%d]\n", buf_id, buf_type);
 
 	mutex_lock(&ctx->lock);
 
@@ -1252,7 +1238,7 @@ static int fimc_dst_set_addr(struct device *dev,
 
 	property = &c_node->property;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]buf_id[%d]buf_type[%d]\n", __func__,
+	DRM_DEBUG_KMS("prop_id[%d]buf_id[%d]buf_type[%d]\n",
 		property->prop_id, buf_id, buf_type);
 
 	if (buf_id > FIMC_MAX_DST) {
@@ -1302,7 +1288,7 @@ static struct exynos_drm_ipp_ops fimc_dst_ops = {
 
 static int fimc_clk_ctrl(struct fimc_context *ctx, bool enable)
 {
-	DRM_DEBUG_KMS("%s:enable[%d]\n", __func__, enable);
+	DRM_DEBUG_KMS("enable[%d]\n", enable);
 
 	if (enable) {
 		clk_prepare_enable(ctx->clocks[FIMC_CLK_GATE]);
@@ -1326,7 +1312,7 @@ static irqreturn_t fimc_irq_handler(int irq, void *dev_id)
 		c_node->event_work;
 	int buf_id;
 
-	DRM_DEBUG_KMS("%s:fimc id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("fimc id[%d]\n", ctx->id);
 
 	fimc_clear_irq(ctx);
 	if (fimc_check_ovf(ctx))
@@ -1339,7 +1325,7 @@ static irqreturn_t fimc_irq_handler(int irq, void *dev_id)
 	if (buf_id < 0)
 		return IRQ_HANDLED;
 
-	DRM_DEBUG_KMS("%s:buf_id[%d]\n", __func__, buf_id);
+	DRM_DEBUG_KMS("buf_id[%d]\n", buf_id);
 
 	if (fimc_dst_set_buf_seq(ctx, buf_id, IPP_BUF_DEQUEUE) < 0) {
 		DRM_ERROR("failed to dequeue.\n");
@@ -1357,8 +1343,6 @@ static int fimc_init_prop_list(struct exynos_drm_ippdrv *ippdrv)
 {
 	struct drm_exynos_ipp_prop_list *prop_list;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	prop_list = devm_kzalloc(ippdrv->dev, sizeof(*prop_list), GFP_KERNEL);
 	if (!prop_list) {
 		DRM_ERROR("failed to alloc property list.\n");
@@ -1402,7 +1386,7 @@ static inline bool fimc_check_drm_flip(enum drm_exynos_flip flip)
 	case EXYNOS_DRM_FLIP_BOTH:
 		return true;
 	default:
-		DRM_DEBUG_KMS("%s:invalid flip\n", __func__);
+		DRM_DEBUG_KMS("invalid flip\n");
 		return false;
 	}
 }
@@ -1419,8 +1403,6 @@ static int fimc_ippdrv_check_property(struct device *dev,
 	bool swap;
 	int i;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	for_each_ipp_ops(i) {
 		if ((i == EXYNOS_DRM_OPS_SRC) &&
 			(property->cmd == IPP_CMD_WB))
@@ -1526,8 +1508,6 @@ static void fimc_clear_addr(struct fimc_context *ctx)
 {
 	int i;
 
-	DRM_DEBUG_KMS("%s:\n", __func__);
-
 	for (i = 0; i < FIMC_MAX_SRC; i++) {
 		fimc_write(0, EXYNOS_CIIYSA(i));
 		fimc_write(0, EXYNOS_CIICBSA(i));
@@ -1545,8 +1525,6 @@ static int fimc_ippdrv_reset(struct device *dev)
 {
 	struct fimc_context *ctx = get_fimc_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* reset h/w block */
 	fimc_sw_reset(ctx);
 
@@ -1570,7 +1548,7 @@ static int fimc_ippdrv_start(struct device *dev, enum drm_exynos_ipp_cmd cmd)
 	int ret, i;
 	u32 cfg0, cfg1;
 
-	DRM_DEBUG_KMS("%s:cmd[%d]\n", __func__, cmd);
+	DRM_DEBUG_KMS("cmd[%d]\n", cmd);
 
 	if (!c_node) {
 		DRM_ERROR("failed to get c_node.\n");
@@ -1679,7 +1657,7 @@ static void fimc_ippdrv_stop(struct device *dev, enum drm_exynos_ipp_cmd cmd)
 	struct drm_exynos_ipp_set_wb set_wb = {0, 0};
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:cmd[%d]\n", __func__, cmd);
+	DRM_DEBUG_KMS("cmd[%d]\n", cmd);
 
 	switch (cmd) {
 	case IPP_CMD_M2M:
@@ -1869,8 +1847,7 @@ static int fimc_probe(struct platform_device *pdev)
 		goto err_put_clk;
 	}
 
-	DRM_DEBUG_KMS("%s:id[%d]ippdrv[0x%x]\n", __func__, ctx->id,
-		(int)ippdrv);
+	DRM_DEBUG_KMS("id[%d]ippdrv[0x%x]\n", ctx->id, (int)ippdrv);
 
 	mutex_init(&ctx->lock);
 	platform_set_drvdata(pdev, ctx);
@@ -1917,7 +1894,7 @@ static int fimc_suspend(struct device *dev)
 {
 	struct fimc_context *ctx = get_fimc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	if (pm_runtime_suspended(dev))
 		return 0;
@@ -1929,7 +1906,7 @@ static int fimc_resume(struct device *dev)
 {
 	struct fimc_context *ctx = get_fimc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	if (!pm_runtime_suspended(dev))
 		return fimc_clk_ctrl(ctx, true);
@@ -1943,7 +1920,7 @@ static int fimc_runtime_suspend(struct device *dev)
 {
 	struct fimc_context *ctx = get_fimc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	return  fimc_clk_ctrl(ctx, false);
 }
@@ -1952,7 +1929,7 @@ static int fimc_runtime_resume(struct device *dev)
 {
 	struct fimc_context *ctx = get_fimc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	return  fimc_clk_ctrl(ctx, true);
 }
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimd.c b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
index 97c61dbffd82..3e106beca5b6 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimd.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
@@ -63,14 +63,24 @@
 
 struct fimd_driver_data {
 	unsigned int timing_base;
+
+	unsigned int has_shadowcon:1;
+	unsigned int has_clksel:1;
+};
+
+static struct fimd_driver_data s3c64xx_fimd_driver_data = {
+	.timing_base = 0x0,
+	.has_clksel = 1,
 };
 
 static struct fimd_driver_data exynos4_fimd_driver_data = {
 	.timing_base = 0x0,
+	.has_shadowcon = 1,
 };
 
 static struct fimd_driver_data exynos5_fimd_driver_data = {
 	.timing_base = 0x20000,
+	.has_shadowcon = 1,
 };
 
 struct fimd_win_data {
@@ -107,10 +117,13 @@ struct fimd_context {
 	atomic_t			wait_vsync_event;
 
 	struct exynos_drm_panel_info *panel;
+	struct fimd_driver_data *driver_data;
 };
 
 #ifdef CONFIG_OF
 static const struct of_device_id fimd_driver_dt_match[] = {
+	{ .compatible = "samsung,s3c6400-fimd",
+	  .data = &s3c64xx_fimd_driver_data },
 	{ .compatible = "samsung,exynos4210-fimd",
 	  .data = &exynos4_fimd_driver_data },
 	{ .compatible = "samsung,exynos5250-fimd",
@@ -137,8 +150,6 @@ static inline struct fimd_driver_data *drm_fimd_get_driver_data(
 
 static bool fimd_display_is_connected(struct device *dev)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO. */
 
 	return true;
@@ -148,15 +159,11 @@ static void *fimd_get_panel(struct device *dev)
 {
 	struct fimd_context *ctx = get_fimd_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	return ctx->panel;
 }
 
-static int fimd_check_timing(struct device *dev, void *timing)
+static int fimd_check_mode(struct device *dev, struct drm_display_mode *mode)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO. */
 
 	return 0;
@@ -164,8 +171,6 @@ static int fimd_check_timing(struct device *dev, void *timing)
 
 static int fimd_display_power_on(struct device *dev, int mode)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO */
 
 	return 0;
@@ -175,7 +180,7 @@ static struct exynos_drm_display_ops fimd_display_ops = {
 	.type = EXYNOS_DISPLAY_TYPE_LCD,
 	.is_connected = fimd_display_is_connected,
 	.get_panel = fimd_get_panel,
-	.check_timing = fimd_check_timing,
+	.check_mode = fimd_check_mode,
 	.power_on = fimd_display_power_on,
 };
 
@@ -183,7 +188,7 @@ static void fimd_dpms(struct device *subdrv_dev, int mode)
 {
 	struct fimd_context *ctx = get_fimd_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s, %d\n", __FILE__, mode);
+	DRM_DEBUG_KMS("%d\n", mode);
 
 	mutex_lock(&ctx->lock);
 
@@ -221,8 +226,6 @@ static void fimd_apply(struct device *subdrv_dev)
 	struct fimd_win_data *win_data;
 	int i;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	for (i = 0; i < WINDOWS_NR; i++) {
 		win_data = &ctx->win_data[i];
 		if (win_data->enabled && (ovl_ops && ovl_ops->commit))
@@ -239,15 +242,12 @@ static void fimd_commit(struct device *dev)
 	struct exynos_drm_panel_info *panel = ctx->panel;
 	struct fb_videomode *timing = &panel->timing;
 	struct fimd_driver_data *driver_data;
-	struct platform_device *pdev = to_platform_device(dev);
 	u32 val;
 
-	driver_data = drm_fimd_get_driver_data(pdev);
+	driver_data = ctx->driver_data;
 	if (ctx->suspended)
 		return;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* setup polarity values from machine code. */
 	writel(ctx->vidcon1, ctx->regs + driver_data->timing_base + VIDCON1);
 
@@ -274,6 +274,11 @@ static void fimd_commit(struct device *dev)
 	val = ctx->vidcon0;
 	val &= ~(VIDCON0_CLKVAL_F_MASK | VIDCON0_CLKDIR);
 
+	if (ctx->driver_data->has_clksel) {
+		val &= ~VIDCON0_CLKSEL_MASK;
+		val |= VIDCON0_CLKSEL_LCD;
+	}
+
 	if (ctx->clkdiv > 1)
 		val |= VIDCON0_CLKVAL_F(ctx->clkdiv - 1) | VIDCON0_CLKDIR;
 	else
@@ -292,8 +297,6 @@ static int fimd_enable_vblank(struct device *dev)
 	struct fimd_context *ctx = get_fimd_context(dev);
 	u32 val;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return -EPERM;
 
@@ -319,8 +322,6 @@ static void fimd_disable_vblank(struct device *dev)
 	struct fimd_context *ctx = get_fimd_context(dev);
 	u32 val;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return;
 
@@ -370,8 +371,6 @@ static void fimd_win_mode_set(struct device *dev,
 	int win;
 	unsigned long offset;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!overlay) {
 		dev_err(dev, "overlay is NULL\n");
 		return;
@@ -381,7 +380,7 @@ static void fimd_win_mode_set(struct device *dev,
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
 
-	if (win < 0 || win > WINDOWS_NR)
+	if (win < 0 || win >= WINDOWS_NR)
 		return;
 
 	offset = overlay->fb_x * (overlay->bpp >> 3);
@@ -418,8 +417,6 @@ static void fimd_win_set_pixfmt(struct device *dev, unsigned int win)
 	struct fimd_win_data *win_data = &ctx->win_data[win];
 	unsigned long val;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	val = WINCONx_ENWIN;
 
 	switch (win_data->bpp) {
@@ -478,8 +475,6 @@ static void fimd_win_set_colkey(struct device *dev, unsigned int win)
 	struct fimd_context *ctx = get_fimd_context(dev);
 	unsigned int keycon0 = 0, keycon1 = 0;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	keycon0 = ~(WxKEYCON0_KEYBL_EN | WxKEYCON0_KEYEN_F |
 			WxKEYCON0_DIRCON) | WxKEYCON0_COMPKEY(0);
 
@@ -489,6 +484,33 @@ static void fimd_win_set_colkey(struct device *dev, unsigned int win)
 	writel(keycon1, ctx->regs + WKEYCON1_BASE(win));
 }
 
+/**
+ * shadow_protect_win() - disable updating values from shadow registers at vsync
+ *
+ * @win: window to protect registers for
+ * @protect: 1 to protect (disable updates)
+ */
+static void fimd_shadow_protect_win(struct fimd_context *ctx,
+							int win, bool protect)
+{
+	u32 reg, bits, val;
+
+	if (ctx->driver_data->has_shadowcon) {
+		reg = SHADOWCON;
+		bits = SHADOWCON_WINx_PROTECT(win);
+	} else {
+		reg = PRTCON;
+		bits = PRTCON_PROTECT;
+	}
+
+	val = readl(ctx->regs + reg);
+	if (protect)
+		val |= bits;
+	else
+		val &= ~bits;
+	writel(val, ctx->regs + reg);
+}
+
 static void fimd_win_commit(struct device *dev, int zpos)
 {
 	struct fimd_context *ctx = get_fimd_context(dev);
@@ -498,21 +520,19 @@ static void fimd_win_commit(struct device *dev, int zpos)
 	unsigned int last_x;
 	unsigned int last_y;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return;
 
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
 
-	if (win < 0 || win > WINDOWS_NR)
+	if (win < 0 || win >= WINDOWS_NR)
 		return;
 
 	win_data = &ctx->win_data[win];
 
 	/*
-	 * SHADOWCON register is used for enabling timing.
+	 * SHADOWCON/PRTCON register is used for enabling timing.
 	 *
 	 * for example, once only width value of a register is set,
 	 * if the dma is started then fimd hardware could malfunction so
@@ -522,9 +542,7 @@ static void fimd_win_commit(struct device *dev, int zpos)
 	 */
 
 	/* protect windows */
-	val = readl(ctx->regs + SHADOWCON);
-	val |= SHADOWCON_WINx_PROTECT(win);
-	writel(val, ctx->regs + SHADOWCON);
+	fimd_shadow_protect_win(ctx, win, true);
 
 	/* buffer start address */
 	val = (unsigned long)win_data->dma_addr;
@@ -602,10 +620,13 @@ static void fimd_win_commit(struct device *dev, int zpos)
 	writel(val, ctx->regs + WINCON(win));
 
 	/* Enable DMA channel and unprotect windows */
-	val = readl(ctx->regs + SHADOWCON);
-	val |= SHADOWCON_CHx_ENABLE(win);
-	val &= ~SHADOWCON_WINx_PROTECT(win);
-	writel(val, ctx->regs + SHADOWCON);
+	fimd_shadow_protect_win(ctx, win, false);
+
+	if (ctx->driver_data->has_shadowcon) {
+		val = readl(ctx->regs + SHADOWCON);
+		val |= SHADOWCON_CHx_ENABLE(win);
+		writel(val, ctx->regs + SHADOWCON);
+	}
 
 	win_data->enabled = true;
 }
@@ -617,12 +638,10 @@ static void fimd_win_disable(struct device *dev, int zpos)
 	int win = zpos;
 	u32 val;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
 
-	if (win < 0 || win > WINDOWS_NR)
+	if (win < 0 || win >= WINDOWS_NR)
 		return;
 
 	win_data = &ctx->win_data[win];
@@ -634,9 +653,7 @@ static void fimd_win_disable(struct device *dev, int zpos)
 	}
 
 	/* protect windows */
-	val = readl(ctx->regs + SHADOWCON);
-	val |= SHADOWCON_WINx_PROTECT(win);
-	writel(val, ctx->regs + SHADOWCON);
+	fimd_shadow_protect_win(ctx, win, true);
 
 	/* wincon */
 	val = readl(ctx->regs + WINCON(win));
@@ -644,10 +661,13 @@ static void fimd_win_disable(struct device *dev, int zpos)
 	writel(val, ctx->regs + WINCON(win));
 
 	/* unprotect windows */
-	val = readl(ctx->regs + SHADOWCON);
-	val &= ~SHADOWCON_CHx_ENABLE(win);
-	val &= ~SHADOWCON_WINx_PROTECT(win);
-	writel(val, ctx->regs + SHADOWCON);
+	if (ctx->driver_data->has_shadowcon) {
+		val = readl(ctx->regs + SHADOWCON);
+		val &= ~SHADOWCON_CHx_ENABLE(win);
+		writel(val, ctx->regs + SHADOWCON);
+	}
+
+	fimd_shadow_protect_win(ctx, win, false);
 
 	win_data->enabled = false;
 }
@@ -697,8 +717,6 @@ out:
 
 static int fimd_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * enable drm irq mode.
 	 * - with irq_enabled = 1, we can use the vblank feature.
@@ -725,8 +743,6 @@ static int fimd_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 
 static void fimd_subdrv_remove(struct drm_device *drm_dev, struct device *dev)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* detach this sub driver from iommu mapping if supported. */
 	if (is_drm_iommu_supported(drm_dev))
 		drm_iommu_detach_device(drm_dev, dev);
@@ -741,8 +757,6 @@ static int fimd_calc_clkdiv(struct fimd_context *ctx,
 	u32 best_framerate = 0;
 	u32 framerate;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	retrace = timing->left_margin + timing->hsync_len +
 				timing->right_margin + timing->xres;
 	retrace *= timing->upper_margin + timing->vsync_len +
@@ -777,10 +791,6 @@ static int fimd_calc_clkdiv(struct fimd_context *ctx,
 
 static void fimd_clear_win(struct fimd_context *ctx, int win)
 {
-	u32 val;
-
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	writel(0, ctx->regs + WINCON(win));
 	writel(0, ctx->regs + VIDOSD_A(win));
 	writel(0, ctx->regs + VIDOSD_B(win));
@@ -789,15 +799,11 @@ static void fimd_clear_win(struct fimd_context *ctx, int win)
 	if (win == 1 || win == 2)
 		writel(0, ctx->regs + VIDOSD_D(win));
 
-	val = readl(ctx->regs + SHADOWCON);
-	val &= ~SHADOWCON_WINx_PROTECT(win);
-	writel(val, ctx->regs + SHADOWCON);
+	fimd_shadow_protect_win(ctx, win, false);
 }
 
 static int fimd_clock(struct fimd_context *ctx, bool enable)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (enable) {
 		int ret;
 
@@ -883,8 +889,6 @@ static int fimd_probe(struct platform_device *pdev)
 	int win;
 	int ret = -EINVAL;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (dev->of_node) {
 		pdata = devm_kzalloc(dev, sizeof(*pdata), GFP_KERNEL);
 		if (!pdata) {
@@ -949,6 +953,7 @@ static int fimd_probe(struct platform_device *pdev)
 		return ret;
 	}
 
+	ctx->driver_data = drm_fimd_get_driver_data(pdev);
 	ctx->vidcon0 = pdata->vidcon0;
 	ctx->vidcon1 = pdata->vidcon1;
 	ctx->default_win = pdata->default_win;
@@ -989,8 +994,6 @@ static int fimd_remove(struct platform_device *pdev)
 	struct device *dev = &pdev->dev;
 	struct fimd_context *ctx = platform_get_drvdata(pdev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_drm_subdrv_unregister(&ctx->subdrv);
 
 	if (ctx->suspended)
@@ -1055,8 +1058,6 @@ static int fimd_runtime_suspend(struct device *dev)
 {
 	struct fimd_context *ctx = get_fimd_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	return fimd_activate(ctx, false);
 }
 
@@ -1064,14 +1065,15 @@ static int fimd_runtime_resume(struct device *dev)
 {
 	struct fimd_context *ctx = get_fimd_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	return fimd_activate(ctx, true);
 }
 #endif
 
 static struct platform_device_id fimd_driver_ids[] = {
 	{
+		.name		= "s3c64xx-fb",
+		.driver_data	= (unsigned long)&s3c64xx_fimd_driver_data,
+	}, {
 		.name		= "exynos4-fb",
 		.driver_data	= (unsigned long)&exynos4_fimd_driver_data,
 	}, {
diff --git a/drivers/gpu/drm/exynos/exynos_drm_g2d.c b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
index af75434ee4d7..42a5a5466075 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_g2d.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
@@ -388,12 +388,9 @@ out:
 
 	sg_free_table(g2d_userptr->sgt);
 	kfree(g2d_userptr->sgt);
-	g2d_userptr->sgt = NULL;
 
-	kfree(g2d_userptr->pages);
-	g2d_userptr->pages = NULL;
+	drm_free_large(g2d_userptr->pages);
 	kfree(g2d_userptr);
-	g2d_userptr = NULL;
 }
 
 static dma_addr_t *g2d_userptr_get_dma_addr(struct drm_device *drm_dev,
@@ -463,11 +460,11 @@ static dma_addr_t *g2d_userptr_get_dma_addr(struct drm_device *drm_dev,
 	npages = (end - start) >> PAGE_SHIFT;
 	g2d_userptr->npages = npages;
 
-	pages = kzalloc(npages * sizeof(struct page *), GFP_KERNEL);
+	pages = drm_calloc_large(npages, sizeof(struct page *));
 	if (!pages) {
 		DRM_ERROR("failed to allocate pages.\n");
-		kfree(g2d_userptr);
-		return ERR_PTR(-ENOMEM);
+		ret = -ENOMEM;
+		goto err_free;
 	}
 
 	vma = find_vma(current->mm, userptr);
@@ -543,7 +540,6 @@ err_sg_free_table:
 
 err_free_sgt:
 	kfree(sgt);
-	sgt = NULL;
 
 err_free_userptr:
 	exynos_gem_put_pages_to_userptr(g2d_userptr->pages,
@@ -554,10 +550,10 @@ err_put_vma:
 	exynos_gem_put_vma(g2d_userptr->vma);
 
 err_free_pages:
-	kfree(pages);
+	drm_free_large(pages);
+
+err_free:
 	kfree(g2d_userptr);
-	pages = NULL;
-	g2d_userptr = NULL;
 
 	return ERR_PTR(ret);
 }
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.c b/drivers/gpu/drm/exynos/exynos_drm_gem.c
index cf4543ffa079..24c22a8c3364 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gem.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gem.c
@@ -132,8 +132,6 @@ void exynos_drm_gem_destroy(struct exynos_drm_gem_obj *exynos_gem_obj)
 	struct drm_gem_object *obj;
 	struct exynos_drm_gem_buf *buf;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	obj = &exynos_gem_obj->base;
 	buf = exynos_gem_obj->buffer;
 
@@ -227,7 +225,6 @@ struct exynos_drm_gem_obj *exynos_drm_gem_create(struct drm_device *dev,
 	}
 
 	size = roundup_gem_size(size, flags);
-	DRM_DEBUG_KMS("%s\n", __FILE__);
 
 	ret = check_gem_flags(flags);
 	if (ret)
@@ -249,13 +246,14 @@ struct exynos_drm_gem_obj *exynos_drm_gem_create(struct drm_device *dev,
 	exynos_gem_obj->flags = flags;
 
 	ret = exynos_drm_alloc_buf(dev, buf, flags);
-	if (ret < 0) {
-		drm_gem_object_release(&exynos_gem_obj->base);
-		goto err_fini_buf;
-	}
+	if (ret < 0)
+		goto err_gem_fini;
 
 	return exynos_gem_obj;
 
+err_gem_fini:
+	drm_gem_object_release(&exynos_gem_obj->base);
+	kfree(exynos_gem_obj);
 err_fini_buf:
 	exynos_drm_fini_buf(dev, buf);
 	return ERR_PTR(ret);
@@ -268,8 +266,6 @@ int exynos_drm_gem_create_ioctl(struct drm_device *dev, void *data,
 	struct exynos_drm_gem_obj *exynos_gem_obj;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_gem_obj = exynos_drm_gem_create(dev, args->flags, args->size);
 	if (IS_ERR(exynos_gem_obj))
 		return PTR_ERR(exynos_gem_obj);
@@ -331,8 +327,6 @@ int exynos_drm_gem_map_offset_ioctl(struct drm_device *dev, void *data,
 {
 	struct drm_exynos_gem_map_off *args = data;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	DRM_DEBUG_KMS("handle = 0x%x, offset = 0x%lx\n",
 			args->handle, (unsigned long)args->offset);
 
@@ -371,8 +365,6 @@ static int exynos_drm_gem_mmap_buffer(struct file *filp,
 	unsigned long vm_size;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	vma->vm_flags |= VM_IO | VM_DONTEXPAND | VM_DONTDUMP;
 	vma->vm_private_data = obj;
 	vma->vm_ops = drm_dev->driver->gem_vm_ops;
@@ -429,9 +421,7 @@ int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data,
 {
 	struct drm_exynos_gem_mmap *args = data;
 	struct drm_gem_object *obj;
-	unsigned int addr;
-
-	DRM_DEBUG_KMS("%s\n", __FILE__);
+	unsigned long addr;
 
 	if (!(dev->driver->driver_features & DRIVER_GEM)) {
 		DRM_ERROR("does not support GEM.\n");
@@ -473,14 +463,14 @@ int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data,
 
 	drm_gem_object_unreference(obj);
 
-	if (IS_ERR((void *)addr)) {
+	if (IS_ERR_VALUE(addr)) {
 		/* check filp->f_op, filp->private_data are restored */
 		if (file_priv->filp->f_op == &exynos_drm_gem_fops) {
 			file_priv->filp->f_op = fops_get(dev->driver->fops);
 			file_priv->filp->private_data = file_priv;
 		}
 		mutex_unlock(&dev->struct_mutex);
-		return PTR_ERR((void *)addr);
+		return (int)addr;
 	}
 
 	mutex_unlock(&dev->struct_mutex);
@@ -643,8 +633,6 @@ void exynos_gem_unmap_sgt_from_dma(struct drm_device *drm_dev,
 
 int exynos_drm_gem_init_object(struct drm_gem_object *obj)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	return 0;
 }
 
@@ -653,8 +641,6 @@ void exynos_drm_gem_free_object(struct drm_gem_object *obj)
 	struct exynos_drm_gem_obj *exynos_gem_obj;
 	struct exynos_drm_gem_buf *buf;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_gem_obj = to_exynos_gem_obj(obj);
 	buf = exynos_gem_obj->buffer;
 
@@ -671,8 +657,6 @@ int exynos_drm_gem_dumb_create(struct drm_file *file_priv,
 	struct exynos_drm_gem_obj *exynos_gem_obj;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * alocate memory to be used for framebuffer.
 	 * - this callback would be called by user application
@@ -704,8 +688,6 @@ int exynos_drm_gem_dumb_map_offset(struct drm_file *file_priv,
 	struct drm_gem_object *obj;
 	int ret = 0;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	mutex_lock(&dev->struct_mutex);
 
 	/*
@@ -743,8 +725,6 @@ int exynos_drm_gem_dumb_destroy(struct drm_file *file_priv,
 {
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * obj->refcount and obj->handle_count are decreased and
 	 * if both them are 0 then exynos_drm_gem_free_object()
@@ -788,8 +768,6 @@ int exynos_drm_gem_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct drm_gem_object *obj;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* set vm_area_struct. */
 	ret = drm_gem_mmap(filp, vma);
 	if (ret < 0) {
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gsc.c b/drivers/gpu/drm/exynos/exynos_drm_gsc.c
index 762f40d548b7..472e3b25e7f2 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gsc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gsc.c
@@ -400,8 +400,6 @@ static int gsc_sw_reset(struct gsc_context *ctx)
 	u32 cfg;
 	int count = GSC_RESET_TIMEOUT;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* s/w reset */
 	cfg = (GSC_SW_RESET_SRESET);
 	gsc_write(cfg, GSC_SW_RESET);
@@ -441,8 +439,6 @@ static void gsc_set_gscblk_fimd_wb(struct gsc_context *ctx, bool enable)
 {
 	u32 gscblk_cfg;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	gscblk_cfg = readl(SYSREG_GSCBLK_CFG1);
 
 	if (enable)
@@ -460,7 +456,7 @@ static void gsc_handle_irq(struct gsc_context *ctx, bool enable,
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:enable[%d]overflow[%d]level[%d]\n", __func__,
+	DRM_DEBUG_KMS("enable[%d]overflow[%d]level[%d]\n",
 			enable, overflow, done);
 
 	cfg = gsc_read(GSC_IRQ);
@@ -491,7 +487,7 @@ static int gsc_src_set_fmt(struct device *dev, u32 fmt)
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:fmt[0x%x]\n", __func__, fmt);
+	DRM_DEBUG_KMS("fmt[0x%x]\n", fmt);
 
 	cfg = gsc_read(GSC_IN_CON);
 	cfg &= ~(GSC_IN_RGB_TYPE_MASK | GSC_IN_YUV422_1P_ORDER_MASK |
@@ -567,8 +563,7 @@ static int gsc_src_set_transf(struct device *dev,
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:degree[%d]flip[0x%x]\n", __func__,
-		degree, flip);
+	DRM_DEBUG_KMS("degree[%d]flip[0x%x]\n", degree, flip);
 
 	cfg = gsc_read(GSC_IN_CON);
 	cfg &= ~GSC_IN_ROT_MASK;
@@ -616,8 +611,8 @@ static int gsc_src_set_size(struct device *dev, int swap,
 	struct gsc_scaler *sc = &ctx->sc;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:swap[%d]x[%d]y[%d]w[%d]h[%d]\n",
-		__func__, swap, pos->x, pos->y, pos->w, pos->h);
+	DRM_DEBUG_KMS("swap[%d]x[%d]y[%d]w[%d]h[%d]\n",
+		swap, pos->x, pos->y, pos->w, pos->h);
 
 	if (swap) {
 		img_pos.w = pos->h;
@@ -634,8 +629,7 @@ static int gsc_src_set_size(struct device *dev, int swap,
 		GSC_CROPPED_HEIGHT(img_pos.h));
 	gsc_write(cfg, GSC_CROPPED_SIZE);
 
-	DRM_DEBUG_KMS("%s:hsize[%d]vsize[%d]\n",
-		__func__, sz->hsize, sz->vsize);
+	DRM_DEBUG_KMS("hsize[%d]vsize[%d]\n", sz->hsize, sz->vsize);
 
 	/* original size */
 	cfg = gsc_read(GSC_SRCIMG_SIZE);
@@ -650,8 +644,7 @@ static int gsc_src_set_size(struct device *dev, int swap,
 	cfg = gsc_read(GSC_IN_CON);
 	cfg &= ~GSC_IN_RGB_TYPE_MASK;
 
-	DRM_DEBUG_KMS("%s:width[%d]range[%d]\n",
-		__func__, pos->w, sc->range);
+	DRM_DEBUG_KMS("width[%d]range[%d]\n", pos->w, sc->range);
 
 	if (pos->w >= GSC_WIDTH_ITU_709)
 		if (sc->range)
@@ -677,8 +670,7 @@ static int gsc_src_set_buf_seq(struct gsc_context *ctx, u32 buf_id,
 	u32 cfg;
 	u32 mask = 0x00000001 << buf_id;
 
-	DRM_DEBUG_KMS("%s:buf_id[%d]buf_type[%d]\n", __func__,
-		buf_id, buf_type);
+	DRM_DEBUG_KMS("buf_id[%d]buf_type[%d]\n", buf_id, buf_type);
 
 	/* mask register set */
 	cfg = gsc_read(GSC_IN_BASE_ADDR_Y_MASK);
@@ -721,7 +713,7 @@ static int gsc_src_set_addr(struct device *dev,
 
 	property = &c_node->property;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]buf_id[%d]buf_type[%d]\n", __func__,
+	DRM_DEBUG_KMS("prop_id[%d]buf_id[%d]buf_type[%d]\n",
 		property->prop_id, buf_id, buf_type);
 
 	if (buf_id > GSC_MAX_SRC) {
@@ -765,7 +757,7 @@ static int gsc_dst_set_fmt(struct device *dev, u32 fmt)
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:fmt[0x%x]\n", __func__, fmt);
+	DRM_DEBUG_KMS("fmt[0x%x]\n", fmt);
 
 	cfg = gsc_read(GSC_OUT_CON);
 	cfg &= ~(GSC_OUT_RGB_TYPE_MASK | GSC_OUT_YUV422_1P_ORDER_MASK |
@@ -838,8 +830,7 @@ static int gsc_dst_set_transf(struct device *dev,
 	struct exynos_drm_ippdrv *ippdrv = &ctx->ippdrv;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:degree[%d]flip[0x%x]\n", __func__,
-		degree, flip);
+	DRM_DEBUG_KMS("degree[%d]flip[0x%x]\n", degree, flip);
 
 	cfg = gsc_read(GSC_IN_CON);
 	cfg &= ~GSC_IN_ROT_MASK;
@@ -881,7 +872,7 @@ static int gsc_dst_set_transf(struct device *dev,
 
 static int gsc_get_ratio_shift(u32 src, u32 dst, u32 *ratio)
 {
-	DRM_DEBUG_KMS("%s:src[%d]dst[%d]\n", __func__, src, dst);
+	DRM_DEBUG_KMS("src[%d]dst[%d]\n", src, dst);
 
 	if (src >= dst * 8) {
 		DRM_ERROR("failed to make ratio and shift.\n");
@@ -944,20 +935,19 @@ static int gsc_set_prescaler(struct gsc_context *ctx, struct gsc_scaler *sc,
 		return ret;
 	}
 
-	DRM_DEBUG_KMS("%s:pre_hratio[%d]pre_vratio[%d]\n",
-		__func__, sc->pre_hratio, sc->pre_vratio);
+	DRM_DEBUG_KMS("pre_hratio[%d]pre_vratio[%d]\n",
+		sc->pre_hratio, sc->pre_vratio);
 
 	sc->main_hratio = (src_w << 16) / dst_w;
 	sc->main_vratio = (src_h << 16) / dst_h;
 
-	DRM_DEBUG_KMS("%s:main_hratio[%ld]main_vratio[%ld]\n",
-		__func__, sc->main_hratio, sc->main_vratio);
+	DRM_DEBUG_KMS("main_hratio[%ld]main_vratio[%ld]\n",
+		sc->main_hratio, sc->main_vratio);
 
 	gsc_get_prescaler_shfactor(sc->pre_hratio, sc->pre_vratio,
 		&sc->pre_shfactor);
 
-	DRM_DEBUG_KMS("%s:pre_shfactor[%d]\n", __func__,
-		sc->pre_shfactor);
+	DRM_DEBUG_KMS("pre_shfactor[%d]\n", sc->pre_shfactor);
 
 	cfg = (GSC_PRESC_SHFACTOR(sc->pre_shfactor) |
 		GSC_PRESC_H_RATIO(sc->pre_hratio) |
@@ -1023,8 +1013,8 @@ static void gsc_set_scaler(struct gsc_context *ctx, struct gsc_scaler *sc)
 {
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:main_hratio[%ld]main_vratio[%ld]\n",
-		__func__, sc->main_hratio, sc->main_vratio);
+	DRM_DEBUG_KMS("main_hratio[%ld]main_vratio[%ld]\n",
+		sc->main_hratio, sc->main_vratio);
 
 	gsc_set_h_coef(ctx, sc->main_hratio);
 	cfg = GSC_MAIN_H_RATIO_VALUE(sc->main_hratio);
@@ -1043,8 +1033,8 @@ static int gsc_dst_set_size(struct device *dev, int swap,
 	struct gsc_scaler *sc = &ctx->sc;
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:swap[%d]x[%d]y[%d]w[%d]h[%d]\n",
-		__func__, swap, pos->x, pos->y, pos->w, pos->h);
+	DRM_DEBUG_KMS("swap[%d]x[%d]y[%d]w[%d]h[%d]\n",
+		swap, pos->x, pos->y, pos->w, pos->h);
 
 	if (swap) {
 		img_pos.w = pos->h;
@@ -1060,8 +1050,7 @@ static int gsc_dst_set_size(struct device *dev, int swap,
 	cfg = (GSC_SCALED_WIDTH(img_pos.w) | GSC_SCALED_HEIGHT(img_pos.h));
 	gsc_write(cfg, GSC_SCALED_SIZE);
 
-	DRM_DEBUG_KMS("%s:hsize[%d]vsize[%d]\n",
-		__func__, sz->hsize, sz->vsize);
+	DRM_DEBUG_KMS("hsize[%d]vsize[%d]\n", sz->hsize, sz->vsize);
 
 	/* original size */
 	cfg = gsc_read(GSC_DSTIMG_SIZE);
@@ -1074,8 +1063,7 @@ static int gsc_dst_set_size(struct device *dev, int swap,
 	cfg = gsc_read(GSC_OUT_CON);
 	cfg &= ~GSC_OUT_RGB_TYPE_MASK;
 
-	DRM_DEBUG_KMS("%s:width[%d]range[%d]\n",
-		__func__, pos->w, sc->range);
+	DRM_DEBUG_KMS("width[%d]range[%d]\n", pos->w, sc->range);
 
 	if (pos->w >= GSC_WIDTH_ITU_709)
 		if (sc->range)
@@ -1104,7 +1092,7 @@ static int gsc_dst_get_buf_seq(struct gsc_context *ctx)
 		if (cfg & (mask << i))
 			buf_num--;
 
-	DRM_DEBUG_KMS("%s:buf_num[%d]\n", __func__, buf_num);
+	DRM_DEBUG_KMS("buf_num[%d]\n", buf_num);
 
 	return buf_num;
 }
@@ -1118,8 +1106,7 @@ static int gsc_dst_set_buf_seq(struct gsc_context *ctx, u32 buf_id,
 	u32 mask = 0x00000001 << buf_id;
 	int ret = 0;
 
-	DRM_DEBUG_KMS("%s:buf_id[%d]buf_type[%d]\n", __func__,
-		buf_id, buf_type);
+	DRM_DEBUG_KMS("buf_id[%d]buf_type[%d]\n", buf_id, buf_type);
 
 	mutex_lock(&ctx->lock);
 
@@ -1177,7 +1164,7 @@ static int gsc_dst_set_addr(struct device *dev,
 
 	property = &c_node->property;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]buf_id[%d]buf_type[%d]\n", __func__,
+	DRM_DEBUG_KMS("prop_id[%d]buf_id[%d]buf_type[%d]\n",
 		property->prop_id, buf_id, buf_type);
 
 	if (buf_id > GSC_MAX_DST) {
@@ -1217,7 +1204,7 @@ static struct exynos_drm_ipp_ops gsc_dst_ops = {
 
 static int gsc_clk_ctrl(struct gsc_context *ctx, bool enable)
 {
-	DRM_DEBUG_KMS("%s:enable[%d]\n", __func__, enable);
+	DRM_DEBUG_KMS("enable[%d]\n", enable);
 
 	if (enable) {
 		clk_enable(ctx->gsc_clk);
@@ -1236,7 +1223,7 @@ static int gsc_get_src_buf_index(struct gsc_context *ctx)
 	u32 buf_id = GSC_MAX_SRC;
 	int ret;
 
-	DRM_DEBUG_KMS("%s:gsc id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("gsc id[%d]\n", ctx->id);
 
 	cfg = gsc_read(GSC_IN_BASE_ADDR_Y_MASK);
 	curr_index = GSC_IN_CURR_GET_INDEX(cfg);
@@ -1259,7 +1246,7 @@ static int gsc_get_src_buf_index(struct gsc_context *ctx)
 		return ret;
 	}
 
-	DRM_DEBUG_KMS("%s:cfg[0x%x]curr_index[%d]buf_id[%d]\n", __func__, cfg,
+	DRM_DEBUG_KMS("cfg[0x%x]curr_index[%d]buf_id[%d]\n", cfg,
 		curr_index, buf_id);
 
 	return buf_id;
@@ -1271,7 +1258,7 @@ static int gsc_get_dst_buf_index(struct gsc_context *ctx)
 	u32 buf_id = GSC_MAX_DST;
 	int ret;
 
-	DRM_DEBUG_KMS("%s:gsc id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("gsc id[%d]\n", ctx->id);
 
 	cfg = gsc_read(GSC_OUT_BASE_ADDR_Y_MASK);
 	curr_index = GSC_OUT_CURR_GET_INDEX(cfg);
@@ -1294,7 +1281,7 @@ static int gsc_get_dst_buf_index(struct gsc_context *ctx)
 		return ret;
 	}
 
-	DRM_DEBUG_KMS("%s:cfg[0x%x]curr_index[%d]buf_id[%d]\n", __func__, cfg,
+	DRM_DEBUG_KMS("cfg[0x%x]curr_index[%d]buf_id[%d]\n", cfg,
 		curr_index, buf_id);
 
 	return buf_id;
@@ -1310,7 +1297,7 @@ static irqreturn_t gsc_irq_handler(int irq, void *dev_id)
 	u32 status;
 	int buf_id[EXYNOS_DRM_OPS_MAX];
 
-	DRM_DEBUG_KMS("%s:gsc id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("gsc id[%d]\n", ctx->id);
 
 	status = gsc_read(GSC_IRQ);
 	if (status & GSC_IRQ_STATUS_OR_IRQ) {
@@ -1331,7 +1318,7 @@ static irqreturn_t gsc_irq_handler(int irq, void *dev_id)
 		if (buf_id[EXYNOS_DRM_OPS_DST] < 0)
 			return IRQ_HANDLED;
 
-		DRM_DEBUG_KMS("%s:buf_id_src[%d]buf_id_dst[%d]\n", __func__,
+		DRM_DEBUG_KMS("buf_id_src[%d]buf_id_dst[%d]\n",
 			buf_id[EXYNOS_DRM_OPS_SRC], buf_id[EXYNOS_DRM_OPS_DST]);
 
 		event_work->ippdrv = ippdrv;
@@ -1350,8 +1337,6 @@ static int gsc_init_prop_list(struct exynos_drm_ippdrv *ippdrv)
 {
 	struct drm_exynos_ipp_prop_list *prop_list;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	prop_list = devm_kzalloc(ippdrv->dev, sizeof(*prop_list), GFP_KERNEL);
 	if (!prop_list) {
 		DRM_ERROR("failed to alloc property list.\n");
@@ -1394,7 +1379,7 @@ static inline bool gsc_check_drm_flip(enum drm_exynos_flip flip)
 	case EXYNOS_DRM_FLIP_BOTH:
 		return true;
 	default:
-		DRM_DEBUG_KMS("%s:invalid flip\n", __func__);
+		DRM_DEBUG_KMS("invalid flip\n");
 		return false;
 	}
 }
@@ -1411,8 +1396,6 @@ static int gsc_ippdrv_check_property(struct device *dev,
 	bool swap;
 	int i;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	for_each_ipp_ops(i) {
 		if ((i == EXYNOS_DRM_OPS_SRC) &&
 			(property->cmd == IPP_CMD_WB))
@@ -1521,8 +1504,6 @@ static int gsc_ippdrv_reset(struct device *dev)
 	struct gsc_scaler *sc = &ctx->sc;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* reset h/w block */
 	ret = gsc_sw_reset(ctx);
 	if (ret < 0) {
@@ -1549,7 +1530,7 @@ static int gsc_ippdrv_start(struct device *dev, enum drm_exynos_ipp_cmd cmd)
 	u32 cfg;
 	int ret, i;
 
-	DRM_DEBUG_KMS("%s:cmd[%d]\n", __func__, cmd);
+	DRM_DEBUG_KMS("cmd[%d]\n", cmd);
 
 	if (!c_node) {
 		DRM_ERROR("failed to get c_node.\n");
@@ -1643,7 +1624,7 @@ static void gsc_ippdrv_stop(struct device *dev, enum drm_exynos_ipp_cmd cmd)
 	struct drm_exynos_ipp_set_wb set_wb = {0, 0};
 	u32 cfg;
 
-	DRM_DEBUG_KMS("%s:cmd[%d]\n", __func__, cmd);
+	DRM_DEBUG_KMS("cmd[%d]\n", cmd);
 
 	switch (cmd) {
 	case IPP_CMD_M2M:
@@ -1728,8 +1709,7 @@ static int gsc_probe(struct platform_device *pdev)
 		return ret;
 	}
 
-	DRM_DEBUG_KMS("%s:id[%d]ippdrv[0x%x]\n", __func__, ctx->id,
-		(int)ippdrv);
+	DRM_DEBUG_KMS("id[%d]ippdrv[0x%x]\n", ctx->id, (int)ippdrv);
 
 	mutex_init(&ctx->lock);
 	platform_set_drvdata(pdev, ctx);
@@ -1772,7 +1752,7 @@ static int gsc_suspend(struct device *dev)
 {
 	struct gsc_context *ctx = get_gsc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	if (pm_runtime_suspended(dev))
 		return 0;
@@ -1784,7 +1764,7 @@ static int gsc_resume(struct device *dev)
 {
 	struct gsc_context *ctx = get_gsc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	if (!pm_runtime_suspended(dev))
 		return gsc_clk_ctrl(ctx, true);
@@ -1798,7 +1778,7 @@ static int gsc_runtime_suspend(struct device *dev)
 {
 	struct gsc_context *ctx = get_gsc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	return  gsc_clk_ctrl(ctx, false);
 }
@@ -1807,7 +1787,7 @@ static int gsc_runtime_resume(struct device *dev)
 {
 	struct gsc_context *ctx = get_gsc_context(dev);
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __FILE__, ctx->id);
+	DRM_DEBUG_KMS("id[%d]\n", ctx->id);
 
 	return  gsc_clk_ctrl(ctx, true);
 }
diff --git a/drivers/gpu/drm/exynos/exynos_drm_hdmi.c b/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
index 437fb947e46d..aaa550d622f0 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_hdmi.c
@@ -88,16 +88,12 @@ void exynos_mixer_drv_attach(struct exynos_drm_hdmi_context *ctx)
 
 void exynos_hdmi_ops_register(struct exynos_hdmi_ops *ops)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ops)
 		hdmi_ops = ops;
 }
 
 void exynos_mixer_ops_register(struct exynos_mixer_ops *ops)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ops)
 		mixer_ops = ops;
 }
@@ -106,8 +102,6 @@ static bool drm_hdmi_is_connected(struct device *dev)
 {
 	struct drm_hdmi_context *ctx = to_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (hdmi_ops && hdmi_ops->is_connected)
 		return hdmi_ops->is_connected(ctx->hdmi_ctx->ctx);
 
@@ -119,34 +113,31 @@ static struct edid *drm_hdmi_get_edid(struct device *dev,
 {
 	struct drm_hdmi_context *ctx = to_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (hdmi_ops && hdmi_ops->get_edid)
 		return hdmi_ops->get_edid(ctx->hdmi_ctx->ctx, connector);
 
 	return NULL;
 }
 
-static int drm_hdmi_check_timing(struct device *dev, void *timing)
+static int drm_hdmi_check_mode(struct device *dev,
+		struct drm_display_mode *mode)
 {
 	struct drm_hdmi_context *ctx = to_context(dev);
 	int ret = 0;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	* Both, mixer and hdmi should be able to handle the requested mode.
 	* If any of the two fails, return mode as BAD.
 	*/
 
-	if (mixer_ops && mixer_ops->check_timing)
-		ret = mixer_ops->check_timing(ctx->mixer_ctx->ctx, timing);
+	if (mixer_ops && mixer_ops->check_mode)
+		ret = mixer_ops->check_mode(ctx->mixer_ctx->ctx, mode);
 
 	if (ret)
 		return ret;
 
-	if (hdmi_ops && hdmi_ops->check_timing)
-		return hdmi_ops->check_timing(ctx->hdmi_ctx->ctx, timing);
+	if (hdmi_ops && hdmi_ops->check_mode)
+		return hdmi_ops->check_mode(ctx->hdmi_ctx->ctx, mode);
 
 	return 0;
 }
@@ -155,8 +146,6 @@ static int drm_hdmi_power_on(struct device *dev, int mode)
 {
 	struct drm_hdmi_context *ctx = to_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (hdmi_ops && hdmi_ops->power_on)
 		return hdmi_ops->power_on(ctx->hdmi_ctx->ctx, mode);
 
@@ -167,7 +156,7 @@ static struct exynos_drm_display_ops drm_hdmi_display_ops = {
 	.type = EXYNOS_DISPLAY_TYPE_HDMI,
 	.is_connected = drm_hdmi_is_connected,
 	.get_edid = drm_hdmi_get_edid,
-	.check_timing = drm_hdmi_check_timing,
+	.check_mode = drm_hdmi_check_mode,
 	.power_on = drm_hdmi_power_on,
 };
 
@@ -177,8 +166,6 @@ static int drm_hdmi_enable_vblank(struct device *subdrv_dev)
 	struct exynos_drm_subdrv *subdrv = &ctx->subdrv;
 	struct exynos_drm_manager *manager = subdrv->manager;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (mixer_ops && mixer_ops->enable_vblank)
 		return mixer_ops->enable_vblank(ctx->mixer_ctx->ctx,
 						manager->pipe);
@@ -190,8 +177,6 @@ static void drm_hdmi_disable_vblank(struct device *subdrv_dev)
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (mixer_ops && mixer_ops->disable_vblank)
 		return mixer_ops->disable_vblank(ctx->mixer_ctx->ctx);
 }
@@ -200,8 +185,6 @@ static void drm_hdmi_wait_for_vblank(struct device *subdrv_dev)
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (mixer_ops && mixer_ops->wait_for_vblank)
 		mixer_ops->wait_for_vblank(ctx->mixer_ctx->ctx);
 }
@@ -214,11 +197,9 @@ static void drm_hdmi_mode_fixup(struct device *subdrv_dev,
 	struct drm_display_mode *m;
 	int mode_ok;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	drm_mode_set_crtcinfo(adjusted_mode, 0);
 
-	mode_ok = drm_hdmi_check_timing(subdrv_dev, adjusted_mode);
+	mode_ok = drm_hdmi_check_mode(subdrv_dev, adjusted_mode);
 
 	/* just return if user desired mode exists. */
 	if (mode_ok == 0)
@@ -229,7 +210,7 @@ static void drm_hdmi_mode_fixup(struct device *subdrv_dev,
 	 * to adjusted_mode.
 	 */
 	list_for_each_entry(m, &connector->modes, head) {
-		mode_ok = drm_hdmi_check_timing(subdrv_dev, m);
+		mode_ok = drm_hdmi_check_mode(subdrv_dev, m);
 
 		if (mode_ok == 0) {
 			struct drm_mode_object base;
@@ -256,8 +237,6 @@ static void drm_hdmi_mode_set(struct device *subdrv_dev, void *mode)
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (hdmi_ops && hdmi_ops->mode_set)
 		hdmi_ops->mode_set(ctx->hdmi_ctx->ctx, mode);
 }
@@ -267,8 +246,6 @@ static void drm_hdmi_get_max_resol(struct device *subdrv_dev,
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (hdmi_ops && hdmi_ops->get_max_resol)
 		hdmi_ops->get_max_resol(ctx->hdmi_ctx->ctx, width, height);
 }
@@ -277,8 +254,6 @@ static void drm_hdmi_commit(struct device *subdrv_dev)
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (hdmi_ops && hdmi_ops->commit)
 		hdmi_ops->commit(ctx->hdmi_ctx->ctx);
 }
@@ -287,8 +262,6 @@ static void drm_hdmi_dpms(struct device *subdrv_dev, int mode)
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (mixer_ops && mixer_ops->dpms)
 		mixer_ops->dpms(ctx->mixer_ctx->ctx, mode);
 
@@ -301,8 +274,6 @@ static void drm_hdmi_apply(struct device *subdrv_dev)
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 	int i;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	for (i = 0; i < MIXER_WIN_NR; i++) {
 		if (!ctx->enabled[i])
 			continue;
@@ -331,8 +302,6 @@ static void drm_mixer_mode_set(struct device *subdrv_dev,
 {
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (mixer_ops && mixer_ops->win_mode_set)
 		mixer_ops->win_mode_set(ctx->mixer_ctx->ctx, overlay);
 }
@@ -342,9 +311,7 @@ static void drm_mixer_commit(struct device *subdrv_dev, int zpos)
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 	int win = (zpos == DEFAULT_ZPOS) ? MIXER_DEFAULT_WIN : zpos;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
-	if (win < 0 || win > MIXER_WIN_NR) {
+	if (win < 0 || win >= MIXER_WIN_NR) {
 		DRM_ERROR("mixer window[%d] is wrong\n", win);
 		return;
 	}
@@ -360,9 +327,7 @@ static void drm_mixer_disable(struct device *subdrv_dev, int zpos)
 	struct drm_hdmi_context *ctx = to_context(subdrv_dev);
 	int win = (zpos == DEFAULT_ZPOS) ? MIXER_DEFAULT_WIN : zpos;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
-	if (win < 0 || win > MIXER_WIN_NR) {
+	if (win < 0 || win >= MIXER_WIN_NR) {
 		DRM_ERROR("mixer window[%d] is wrong\n", win);
 		return;
 	}
@@ -392,8 +357,6 @@ static int hdmi_subdrv_probe(struct drm_device *drm_dev,
 	struct exynos_drm_subdrv *subdrv = to_subdrv(dev);
 	struct drm_hdmi_context *ctx;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!hdmi_ctx) {
 		DRM_ERROR("hdmi context not initialized.\n");
 		return -EFAULT;
@@ -440,8 +403,6 @@ static int exynos_drm_hdmi_probe(struct platform_device *pdev)
 	struct exynos_drm_subdrv *subdrv;
 	struct drm_hdmi_context *ctx;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL);
 	if (!ctx) {
 		DRM_LOG_KMS("failed to alloc common hdmi context.\n");
@@ -466,8 +427,6 @@ static int exynos_drm_hdmi_remove(struct platform_device *pdev)
 {
 	struct drm_hdmi_context *ctx = platform_get_drvdata(pdev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_drm_subdrv_unregister(&ctx->subdrv);
 
 	return 0;
diff --git a/drivers/gpu/drm/exynos/exynos_drm_hdmi.h b/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
index 6b709440df4c..724cab181976 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_hdmi.h
@@ -32,11 +32,11 @@ struct exynos_hdmi_ops {
 	bool (*is_connected)(void *ctx);
 	struct edid *(*get_edid)(void *ctx,
 			struct drm_connector *connector);
-	int (*check_timing)(void *ctx, struct fb_videomode *timing);
+	int (*check_mode)(void *ctx, struct drm_display_mode *mode);
 	int (*power_on)(void *ctx, int mode);
 
 	/* manager */
-	void (*mode_set)(void *ctx, void *mode);
+	void (*mode_set)(void *ctx, struct drm_display_mode *mode);
 	void (*get_max_resol)(void *ctx, unsigned int *width,
 				unsigned int *height);
 	void (*commit)(void *ctx);
@@ -57,7 +57,7 @@ struct exynos_mixer_ops {
 	void (*win_disable)(void *ctx, int zpos);
 
 	/* display */
-	int (*check_timing)(void *ctx, struct fb_videomode *timing);
+	int (*check_mode)(void *ctx, struct drm_display_mode *mode);
 };
 
 void exynos_hdmi_drv_attach(struct exynos_drm_hdmi_context *ctx);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_ipp.c b/drivers/gpu/drm/exynos/exynos_drm_ipp.c
index be1e88463466..b1ef8e7ff9c9 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_ipp.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_ipp.c
@@ -131,8 +131,6 @@ void exynos_platform_device_ipp_unregister(void)
 
 int exynos_drm_ippdrv_register(struct exynos_drm_ippdrv *ippdrv)
 {
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!ippdrv)
 		return -EINVAL;
 
@@ -145,8 +143,6 @@ int exynos_drm_ippdrv_register(struct exynos_drm_ippdrv *ippdrv)
 
 int exynos_drm_ippdrv_unregister(struct exynos_drm_ippdrv *ippdrv)
 {
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!ippdrv)
 		return -EINVAL;
 
@@ -162,8 +158,6 @@ static int ipp_create_id(struct idr *id_idr, struct mutex *lock, void *obj,
 {
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* do the allocation under our mutexlock */
 	mutex_lock(lock);
 	ret = idr_alloc(id_idr, obj, 1, 0, GFP_KERNEL);
@@ -179,7 +173,7 @@ static void *ipp_find_obj(struct idr *id_idr, struct mutex *lock, u32 id)
 {
 	void *obj;
 
-	DRM_DEBUG_KMS("%s:id[%d]\n", __func__, id);
+	DRM_DEBUG_KMS("id[%d]\n", id);
 
 	mutex_lock(lock);
 
@@ -216,7 +210,7 @@ static struct exynos_drm_ippdrv *ipp_find_driver(struct ipp_context *ctx,
 	struct exynos_drm_ippdrv *ippdrv;
 	u32 ipp_id = property->ipp_id;
 
-	DRM_DEBUG_KMS("%s:ipp_id[%d]\n", __func__, ipp_id);
+	DRM_DEBUG_KMS("ipp_id[%d]\n", ipp_id);
 
 	if (ipp_id) {
 		/* find ipp driver using idr */
@@ -257,14 +251,13 @@ static struct exynos_drm_ippdrv *ipp_find_driver(struct ipp_context *ctx,
 		 */
 		list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
 			if (ipp_check_dedicated(ippdrv, property->cmd)) {
-				DRM_DEBUG_KMS("%s:used device.\n", __func__);
+				DRM_DEBUG_KMS("used device.\n");
 				continue;
 			}
 
 			if (ippdrv->check_property &&
 			    ippdrv->check_property(ippdrv->dev, property)) {
-				DRM_DEBUG_KMS("%s:not support property.\n",
-					__func__);
+				DRM_DEBUG_KMS("not support property.\n");
 				continue;
 			}
 
@@ -283,10 +276,10 @@ static struct exynos_drm_ippdrv *ipp_find_drv_by_handle(u32 prop_id)
 	struct drm_exynos_ipp_cmd_node *c_node;
 	int count = 0;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]\n", __func__, prop_id);
+	DRM_DEBUG_KMS("prop_id[%d]\n", prop_id);
 
 	if (list_empty(&exynos_drm_ippdrv_list)) {
-		DRM_DEBUG_KMS("%s:ippdrv_list is empty.\n", __func__);
+		DRM_DEBUG_KMS("ippdrv_list is empty.\n");
 		return ERR_PTR(-ENODEV);
 	}
 
@@ -296,8 +289,7 @@ static struct exynos_drm_ippdrv *ipp_find_drv_by_handle(u32 prop_id)
 	 * e.g PAUSE state, queue buf, command contro.
 	 */
 	list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
-		DRM_DEBUG_KMS("%s:count[%d]ippdrv[0x%x]\n", __func__,
-			count++, (int)ippdrv);
+		DRM_DEBUG_KMS("count[%d]ippdrv[0x%x]\n", count++, (int)ippdrv);
 
 		if (!list_empty(&ippdrv->cmd_list)) {
 			list_for_each_entry(c_node, &ippdrv->cmd_list, list)
@@ -320,8 +312,6 @@ int exynos_drm_ipp_get_property(struct drm_device *drm_dev, void *data,
 	struct exynos_drm_ippdrv *ippdrv;
 	int count = 0;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!ctx) {
 		DRM_ERROR("invalid context.\n");
 		return -EINVAL;
@@ -332,7 +322,7 @@ int exynos_drm_ipp_get_property(struct drm_device *drm_dev, void *data,
 		return -EINVAL;
 	}
 
-	DRM_DEBUG_KMS("%s:ipp_id[%d]\n", __func__, prop_list->ipp_id);
+	DRM_DEBUG_KMS("ipp_id[%d]\n", prop_list->ipp_id);
 
 	if (!prop_list->ipp_id) {
 		list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list)
@@ -371,11 +361,11 @@ static void ipp_print_property(struct drm_exynos_ipp_property *property,
 	struct drm_exynos_pos *pos = &config->pos;
 	struct drm_exynos_sz *sz = &config->sz;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]ops[%s]fmt[0x%x]\n",
-		__func__, property->prop_id, idx ? "dst" : "src", config->fmt);
+	DRM_DEBUG_KMS("prop_id[%d]ops[%s]fmt[0x%x]\n",
+		property->prop_id, idx ? "dst" : "src", config->fmt);
 
-	DRM_DEBUG_KMS("%s:pos[%d %d %d %d]sz[%d %d]f[%d]r[%d]\n",
-		__func__, pos->x, pos->y, pos->w, pos->h,
+	DRM_DEBUG_KMS("pos[%d %d %d %d]sz[%d %d]f[%d]r[%d]\n",
+		pos->x, pos->y, pos->w, pos->h,
 		sz->hsize, sz->vsize, config->flip, config->degree);
 }
 
@@ -385,7 +375,7 @@ static int ipp_find_and_set_property(struct drm_exynos_ipp_property *property)
 	struct drm_exynos_ipp_cmd_node *c_node;
 	u32 prop_id = property->prop_id;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]\n", __func__, prop_id);
+	DRM_DEBUG_KMS("prop_id[%d]\n", prop_id);
 
 	ippdrv = ipp_find_drv_by_handle(prop_id);
 	if (IS_ERR(ippdrv)) {
@@ -401,8 +391,8 @@ static int ipp_find_and_set_property(struct drm_exynos_ipp_property *property)
 	list_for_each_entry(c_node, &ippdrv->cmd_list, list) {
 		if ((c_node->property.prop_id == prop_id) &&
 		    (c_node->state == IPP_STATE_STOP)) {
-			DRM_DEBUG_KMS("%s:found cmd[%d]ippdrv[0x%x]\n",
-				__func__, property->cmd, (int)ippdrv);
+			DRM_DEBUG_KMS("found cmd[%d]ippdrv[0x%x]\n",
+				property->cmd, (int)ippdrv);
 
 			c_node->property = *property;
 			return 0;
@@ -418,8 +408,6 @@ static struct drm_exynos_ipp_cmd_work *ipp_create_cmd_work(void)
 {
 	struct drm_exynos_ipp_cmd_work *cmd_work;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	cmd_work = kzalloc(sizeof(*cmd_work), GFP_KERNEL);
 	if (!cmd_work) {
 		DRM_ERROR("failed to alloc cmd_work.\n");
@@ -435,8 +423,6 @@ static struct drm_exynos_ipp_event_work *ipp_create_event_work(void)
 {
 	struct drm_exynos_ipp_event_work *event_work;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	event_work = kzalloc(sizeof(*event_work), GFP_KERNEL);
 	if (!event_work) {
 		DRM_ERROR("failed to alloc event_work.\n");
@@ -460,8 +446,6 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 	struct drm_exynos_ipp_cmd_node *c_node;
 	int ret, i;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!ctx) {
 		DRM_ERROR("invalid context.\n");
 		return -EINVAL;
@@ -486,7 +470,7 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 	 * instead of allocation.
 	 */
 	if (property->prop_id) {
-		DRM_DEBUG_KMS("%s:prop_id[%d]\n", __func__, property->prop_id);
+		DRM_DEBUG_KMS("prop_id[%d]\n", property->prop_id);
 		return ipp_find_and_set_property(property);
 	}
 
@@ -512,8 +496,8 @@ int exynos_drm_ipp_set_property(struct drm_device *drm_dev, void *data,
 		goto err_clear;
 	}
 
-	DRM_DEBUG_KMS("%s:created prop_id[%d]cmd[%d]ippdrv[0x%x]\n",
-		__func__, property->prop_id, property->cmd, (int)ippdrv);
+	DRM_DEBUG_KMS("created prop_id[%d]cmd[%d]ippdrv[0x%x]\n",
+		property->prop_id, property->cmd, (int)ippdrv);
 
 	/* stored property information and ippdrv in private data */
 	c_node->priv = priv;
@@ -569,8 +553,6 @@ err_clear:
 
 static void ipp_clean_cmd_node(struct drm_exynos_ipp_cmd_node *c_node)
 {
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* delete list */
 	list_del(&c_node->list);
 
@@ -593,8 +575,6 @@ static int ipp_check_mem_list(struct drm_exynos_ipp_cmd_node *c_node)
 	struct list_head *head;
 	int ret, i, count[EXYNOS_DRM_OPS_MAX] = { 0, };
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	mutex_lock(&c_node->mem_lock);
 
 	for_each_ipp_ops(i) {
@@ -602,20 +582,19 @@ static int ipp_check_mem_list(struct drm_exynos_ipp_cmd_node *c_node)
 		head = &c_node->mem_list[i];
 
 		if (list_empty(head)) {
-			DRM_DEBUG_KMS("%s:%s memory empty.\n", __func__,
-				i ? "dst" : "src");
+			DRM_DEBUG_KMS("%s memory empty.\n", i ? "dst" : "src");
 			continue;
 		}
 
 		/* find memory node entry */
 		list_for_each_entry(m_node, head, list) {
-			DRM_DEBUG_KMS("%s:%s,count[%d]m_node[0x%x]\n", __func__,
+			DRM_DEBUG_KMS("%s,count[%d]m_node[0x%x]\n",
 				i ? "dst" : "src", count[i], (int)m_node);
 			count[i]++;
 		}
 	}
 
-	DRM_DEBUG_KMS("%s:min[%d]max[%d]\n", __func__,
+	DRM_DEBUG_KMS("min[%d]max[%d]\n",
 		min(count[EXYNOS_DRM_OPS_SRC], count[EXYNOS_DRM_OPS_DST]),
 		max(count[EXYNOS_DRM_OPS_SRC], count[EXYNOS_DRM_OPS_DST]));
 
@@ -644,15 +623,14 @@ static struct drm_exynos_ipp_mem_node
 	struct list_head *head;
 	int count = 0;
 
-	DRM_DEBUG_KMS("%s:buf_id[%d]\n", __func__, qbuf->buf_id);
+	DRM_DEBUG_KMS("buf_id[%d]\n", qbuf->buf_id);
 
 	/* source/destination memory list */
 	head = &c_node->mem_list[qbuf->ops_id];
 
 	/* find memory node from memory list */
 	list_for_each_entry(m_node, head, list) {
-		DRM_DEBUG_KMS("%s:count[%d]m_node[0x%x]\n",
-			__func__, count++, (int)m_node);
+		DRM_DEBUG_KMS("count[%d]m_node[0x%x]\n", count++, (int)m_node);
 
 		/* compare buffer id */
 		if (m_node->buf_id == qbuf->buf_id)
@@ -669,7 +647,7 @@ static int ipp_set_mem_node(struct exynos_drm_ippdrv *ippdrv,
 	struct exynos_drm_ipp_ops *ops = NULL;
 	int ret = 0;
 
-	DRM_DEBUG_KMS("%s:node[0x%x]\n", __func__, (int)m_node);
+	DRM_DEBUG_KMS("node[0x%x]\n", (int)m_node);
 
 	if (!m_node) {
 		DRM_ERROR("invalid queue node.\n");
@@ -678,7 +656,7 @@ static int ipp_set_mem_node(struct exynos_drm_ippdrv *ippdrv,
 
 	mutex_lock(&c_node->mem_lock);
 
-	DRM_DEBUG_KMS("%s:ops_id[%d]\n", __func__, m_node->ops_id);
+	DRM_DEBUG_KMS("ops_id[%d]\n", m_node->ops_id);
 
 	/* get operations callback */
 	ops = ippdrv->ops[m_node->ops_id];
@@ -714,8 +692,6 @@ static struct drm_exynos_ipp_mem_node
 	void *addr;
 	int i;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	mutex_lock(&c_node->mem_lock);
 
 	m_node = kzalloc(sizeof(*m_node), GFP_KERNEL);
@@ -732,14 +708,11 @@ static struct drm_exynos_ipp_mem_node
 	m_node->prop_id = qbuf->prop_id;
 	m_node->buf_id = qbuf->buf_id;
 
-	DRM_DEBUG_KMS("%s:m_node[0x%x]ops_id[%d]\n", __func__,
-		(int)m_node, qbuf->ops_id);
-	DRM_DEBUG_KMS("%s:prop_id[%d]buf_id[%d]\n", __func__,
-		qbuf->prop_id, m_node->buf_id);
+	DRM_DEBUG_KMS("m_node[0x%x]ops_id[%d]\n", (int)m_node, qbuf->ops_id);
+	DRM_DEBUG_KMS("prop_id[%d]buf_id[%d]\n", qbuf->prop_id, m_node->buf_id);
 
 	for_each_ipp_planar(i) {
-		DRM_DEBUG_KMS("%s:i[%d]handle[0x%x]\n", __func__,
-			i, qbuf->handle[i]);
+		DRM_DEBUG_KMS("i[%d]handle[0x%x]\n", i, qbuf->handle[i]);
 
 		/* get dma address by handle */
 		if (qbuf->handle[i]) {
@@ -752,9 +725,8 @@ static struct drm_exynos_ipp_mem_node
 
 			buf_info.handles[i] = qbuf->handle[i];
 			buf_info.base[i] = *(dma_addr_t *) addr;
-			DRM_DEBUG_KMS("%s:i[%d]base[0x%x]hd[0x%x]\n",
-				__func__, i, buf_info.base[i],
-				(int)buf_info.handles[i]);
+			DRM_DEBUG_KMS("i[%d]base[0x%x]hd[0x%x]\n",
+				i, buf_info.base[i], (int)buf_info.handles[i]);
 		}
 	}
 
@@ -778,7 +750,7 @@ static int ipp_put_mem_node(struct drm_device *drm_dev,
 {
 	int i;
 
-	DRM_DEBUG_KMS("%s:node[0x%x]\n", __func__, (int)m_node);
+	DRM_DEBUG_KMS("node[0x%x]\n", (int)m_node);
 
 	if (!m_node) {
 		DRM_ERROR("invalid dequeue node.\n");
@@ -792,7 +764,7 @@ static int ipp_put_mem_node(struct drm_device *drm_dev,
 
 	mutex_lock(&c_node->mem_lock);
 
-	DRM_DEBUG_KMS("%s:ops_id[%d]\n", __func__, m_node->ops_id);
+	DRM_DEBUG_KMS("ops_id[%d]\n", m_node->ops_id);
 
 	/* put gem buffer */
 	for_each_ipp_planar(i) {
@@ -824,8 +796,7 @@ static int ipp_get_event(struct drm_device *drm_dev,
 	struct drm_exynos_ipp_send_event *e;
 	unsigned long flags;
 
-	DRM_DEBUG_KMS("%s:ops_id[%d]buf_id[%d]\n", __func__,
-		qbuf->ops_id, qbuf->buf_id);
+	DRM_DEBUG_KMS("ops_id[%d]buf_id[%d]\n", qbuf->ops_id, qbuf->buf_id);
 
 	e = kzalloc(sizeof(*e), GFP_KERNEL);
 
@@ -857,16 +828,13 @@ static void ipp_put_event(struct drm_exynos_ipp_cmd_node *c_node,
 	struct drm_exynos_ipp_send_event *e, *te;
 	int count = 0;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (list_empty(&c_node->event_list)) {
-		DRM_DEBUG_KMS("%s:event_list is empty.\n", __func__);
+		DRM_DEBUG_KMS("event_list is empty.\n");
 		return;
 	}
 
 	list_for_each_entry_safe(e, te, &c_node->event_list, base.link) {
-		DRM_DEBUG_KMS("%s:count[%d]e[0x%x]\n",
-			__func__, count++, (int)e);
+		DRM_DEBUG_KMS("count[%d]e[0x%x]\n", count++, (int)e);
 
 		/*
 		 * quf == NULL condition means all event deletion.
@@ -912,8 +880,6 @@ static int ipp_queue_buf_with_run(struct device *dev,
 	struct exynos_drm_ipp_ops *ops;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	ippdrv = ipp_find_drv_by_handle(qbuf->prop_id);
 	if (IS_ERR(ippdrv)) {
 		DRM_ERROR("failed to get ipp driver.\n");
@@ -929,12 +895,12 @@ static int ipp_queue_buf_with_run(struct device *dev,
 	property = &c_node->property;
 
 	if (c_node->state != IPP_STATE_START) {
-		DRM_DEBUG_KMS("%s:bypass for invalid state.\n" , __func__);
+		DRM_DEBUG_KMS("bypass for invalid state.\n");
 		return 0;
 	}
 
 	if (!ipp_check_mem_list(c_node)) {
-		DRM_DEBUG_KMS("%s:empty memory.\n", __func__);
+		DRM_DEBUG_KMS("empty memory.\n");
 		return 0;
 	}
 
@@ -964,8 +930,6 @@ static void ipp_clean_queue_buf(struct drm_device *drm_dev,
 {
 	struct drm_exynos_ipp_mem_node *m_node, *tm_node;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!list_empty(&c_node->mem_list[qbuf->ops_id])) {
 		/* delete list */
 		list_for_each_entry_safe(m_node, tm_node,
@@ -989,8 +953,6 @@ int exynos_drm_ipp_queue_buf(struct drm_device *drm_dev, void *data,
 	struct drm_exynos_ipp_mem_node *m_node;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!qbuf) {
 		DRM_ERROR("invalid buf parameter.\n");
 		return -EINVAL;
@@ -1001,8 +963,8 @@ int exynos_drm_ipp_queue_buf(struct drm_device *drm_dev, void *data,
 		return -EINVAL;
 	}
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]ops_id[%s]buf_id[%d]buf_type[%d]\n",
-		__func__, qbuf->prop_id, qbuf->ops_id ? "dst" : "src",
+	DRM_DEBUG_KMS("prop_id[%d]ops_id[%s]buf_id[%d]buf_type[%d]\n",
+		qbuf->prop_id, qbuf->ops_id ? "dst" : "src",
 		qbuf->buf_id, qbuf->buf_type);
 
 	/* find command node */
@@ -1075,8 +1037,6 @@ err_clean_node:
 static bool exynos_drm_ipp_check_valid(struct device *dev,
 		enum drm_exynos_ipp_ctrl ctrl, enum drm_exynos_ipp_state state)
 {
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (ctrl != IPP_CTRL_PLAY) {
 		if (pm_runtime_suspended(dev)) {
 			DRM_ERROR("pm:runtime_suspended.\n");
@@ -1104,7 +1064,6 @@ static bool exynos_drm_ipp_check_valid(struct device *dev,
 	default:
 		DRM_ERROR("invalid state.\n");
 		goto err_status;
-		break;
 	}
 
 	return true;
@@ -1126,8 +1085,6 @@ int exynos_drm_ipp_cmd_ctrl(struct drm_device *drm_dev, void *data,
 	struct drm_exynos_ipp_cmd_work *cmd_work;
 	struct drm_exynos_ipp_cmd_node *c_node;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!ctx) {
 		DRM_ERROR("invalid context.\n");
 		return -EINVAL;
@@ -1138,7 +1095,7 @@ int exynos_drm_ipp_cmd_ctrl(struct drm_device *drm_dev, void *data,
 		return -EINVAL;
 	}
 
-	DRM_DEBUG_KMS("%s:ctrl[%d]prop_id[%d]\n", __func__,
+	DRM_DEBUG_KMS("ctrl[%d]prop_id[%d]\n",
 		cmd_ctrl->ctrl, cmd_ctrl->prop_id);
 
 	ippdrv = ipp_find_drv_by_handle(cmd_ctrl->prop_id);
@@ -1213,7 +1170,7 @@ int exynos_drm_ipp_cmd_ctrl(struct drm_device *drm_dev, void *data,
 		return -EINVAL;
 	}
 
-	DRM_DEBUG_KMS("%s:done ctrl[%d]prop_id[%d]\n", __func__,
+	DRM_DEBUG_KMS("done ctrl[%d]prop_id[%d]\n",
 		cmd_ctrl->ctrl, cmd_ctrl->prop_id);
 
 	return 0;
@@ -1249,7 +1206,7 @@ static int ipp_set_property(struct exynos_drm_ippdrv *ippdrv,
 		return -EINVAL;
 	}
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]\n", __func__, property->prop_id);
+	DRM_DEBUG_KMS("prop_id[%d]\n", property->prop_id);
 
 	/* reset h/w block */
 	if (ippdrv->reset &&
@@ -1310,13 +1267,13 @@ static int ipp_start_property(struct exynos_drm_ippdrv *ippdrv,
 	struct list_head *head;
 	int ret, i;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]\n", __func__, property->prop_id);
+	DRM_DEBUG_KMS("prop_id[%d]\n", property->prop_id);
 
 	/* store command info in ippdrv */
 	ippdrv->c_node = c_node;
 
 	if (!ipp_check_mem_list(c_node)) {
-		DRM_DEBUG_KMS("%s:empty memory.\n", __func__);
+		DRM_DEBUG_KMS("empty memory.\n");
 		return -ENOMEM;
 	}
 
@@ -1343,8 +1300,7 @@ static int ipp_start_property(struct exynos_drm_ippdrv *ippdrv,
 				return ret;
 			}
 
-			DRM_DEBUG_KMS("%s:m_node[0x%x]\n",
-				__func__, (int)m_node);
+			DRM_DEBUG_KMS("m_node[0x%x]\n", (int)m_node);
 
 			ret = ipp_set_mem_node(ippdrv, c_node, m_node);
 			if (ret) {
@@ -1382,7 +1338,7 @@ static int ipp_start_property(struct exynos_drm_ippdrv *ippdrv,
 		return -EINVAL;
 	}
 
-	DRM_DEBUG_KMS("%s:cmd[%d]\n", __func__, property->cmd);
+	DRM_DEBUG_KMS("cmd[%d]\n", property->cmd);
 
 	/* start operations */
 	if (ippdrv->start) {
@@ -1405,7 +1361,7 @@ static int ipp_stop_property(struct drm_device *drm_dev,
 	struct list_head *head;
 	int ret = 0, i;
 
-	DRM_DEBUG_KMS("%s:prop_id[%d]\n", __func__, property->prop_id);
+	DRM_DEBUG_KMS("prop_id[%d]\n", property->prop_id);
 
 	/* put event */
 	ipp_put_event(c_node, NULL);
@@ -1418,8 +1374,7 @@ static int ipp_stop_property(struct drm_device *drm_dev,
 			head = &c_node->mem_list[i];
 
 			if (list_empty(head)) {
-				DRM_DEBUG_KMS("%s:mem_list is empty.\n",
-					__func__);
+				DRM_DEBUG_KMS("mem_list is empty.\n");
 				break;
 			}
 
@@ -1439,7 +1394,7 @@ static int ipp_stop_property(struct drm_device *drm_dev,
 		head = &c_node->mem_list[EXYNOS_DRM_OPS_DST];
 
 		if (list_empty(head)) {
-			DRM_DEBUG_KMS("%s:mem_list is empty.\n", __func__);
+			DRM_DEBUG_KMS("mem_list is empty.\n");
 			break;
 		}
 
@@ -1456,7 +1411,7 @@ static int ipp_stop_property(struct drm_device *drm_dev,
 		head = &c_node->mem_list[EXYNOS_DRM_OPS_SRC];
 
 		if (list_empty(head)) {
-			DRM_DEBUG_KMS("%s:mem_list is empty.\n", __func__);
+			DRM_DEBUG_KMS("mem_list is empty.\n");
 			break;
 		}
 
@@ -1491,8 +1446,6 @@ void ipp_sched_cmd(struct work_struct *work)
 	struct drm_exynos_ipp_property *property;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	ippdrv = cmd_work->ippdrv;
 	if (!ippdrv) {
 		DRM_ERROR("invalid ippdrv list.\n");
@@ -1550,7 +1503,7 @@ void ipp_sched_cmd(struct work_struct *work)
 		break;
 	}
 
-	DRM_DEBUG_KMS("%s:ctrl[%d] done.\n", __func__, cmd_work->ctrl);
+	DRM_DEBUG_KMS("ctrl[%d] done.\n", cmd_work->ctrl);
 
 err_unlock:
 	mutex_unlock(&c_node->cmd_lock);
@@ -1571,8 +1524,7 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv,
 	int ret, i;
 
 	for_each_ipp_ops(i)
-		DRM_DEBUG_KMS("%s:%s buf_id[%d]\n", __func__,
-			i ? "dst" : "src", buf_id[i]);
+		DRM_DEBUG_KMS("%s buf_id[%d]\n", i ? "dst" : "src", buf_id[i]);
 
 	if (!drm_dev) {
 		DRM_ERROR("failed to get drm_dev.\n");
@@ -1585,12 +1537,12 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv,
 	}
 
 	if (list_empty(&c_node->event_list)) {
-		DRM_DEBUG_KMS("%s:event list is empty.\n", __func__);
+		DRM_DEBUG_KMS("event list is empty.\n");
 		return 0;
 	}
 
 	if (!ipp_check_mem_list(c_node)) {
-		DRM_DEBUG_KMS("%s:empty memory.\n", __func__);
+		DRM_DEBUG_KMS("empty memory.\n");
 		return 0;
 	}
 
@@ -1609,7 +1561,7 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv,
 			}
 
 			tbuf_id[i] = m_node->buf_id;
-			DRM_DEBUG_KMS("%s:%s buf_id[%d]\n", __func__,
+			DRM_DEBUG_KMS("%s buf_id[%d]\n",
 				i ? "dst" : "src", tbuf_id[i]);
 
 			ret = ipp_put_mem_node(drm_dev, c_node, m_node);
@@ -1677,8 +1629,7 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv,
 	}
 
 	do_gettimeofday(&now);
-	DRM_DEBUG_KMS("%s:tv_sec[%ld]tv_usec[%ld]\n"
-		, __func__, now.tv_sec, now.tv_usec);
+	DRM_DEBUG_KMS("tv_sec[%ld]tv_usec[%ld]\n", now.tv_sec, now.tv_usec);
 	e->event.tv_sec = now.tv_sec;
 	e->event.tv_usec = now.tv_usec;
 	e->event.prop_id = property->prop_id;
@@ -1692,7 +1643,7 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv,
 	wake_up_interruptible(&e->base.file_priv->event_wait);
 	spin_unlock_irqrestore(&drm_dev->event_lock, flags);
 
-	DRM_DEBUG_KMS("%s:done cmd[%d]prop_id[%d]buf_id[%d]\n", __func__,
+	DRM_DEBUG_KMS("done cmd[%d]prop_id[%d]buf_id[%d]\n",
 		property->cmd, property->prop_id, tbuf_id[EXYNOS_DRM_OPS_DST]);
 
 	return 0;
@@ -1711,8 +1662,7 @@ void ipp_sched_event(struct work_struct *work)
 		return;
 	}
 
-	DRM_DEBUG_KMS("%s:buf_id[%d]\n", __func__,
-		event_work->buf_id[EXYNOS_DRM_OPS_DST]);
+	DRM_DEBUG_KMS("buf_id[%d]\n", event_work->buf_id[EXYNOS_DRM_OPS_DST]);
 
 	ippdrv = event_work->ippdrv;
 	if (!ippdrv) {
@@ -1733,8 +1683,8 @@ void ipp_sched_event(struct work_struct *work)
 	 * or going out operations.
 	 */
 	if (c_node->state != IPP_STATE_START) {
-		DRM_DEBUG_KMS("%s:bypass state[%d]prop_id[%d]\n",
-			__func__, c_node->state, c_node->property.prop_id);
+		DRM_DEBUG_KMS("bypass state[%d]prop_id[%d]\n",
+			c_node->state, c_node->property.prop_id);
 		goto err_completion;
 	}
 
@@ -1759,8 +1709,6 @@ static int ipp_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 	struct exynos_drm_ippdrv *ippdrv;
 	int ret, count = 0;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* get ipp driver entry */
 	list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
 		ippdrv->drm_dev = drm_dev;
@@ -1772,7 +1720,7 @@ static int ipp_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 			goto err_idr;
 		}
 
-		DRM_DEBUG_KMS("%s:count[%d]ippdrv[0x%x]ipp_id[%d]\n", __func__,
+		DRM_DEBUG_KMS("count[%d]ippdrv[0x%x]ipp_id[%d]\n",
 			count++, (int)ippdrv, ippdrv->ipp_id);
 
 		if (ippdrv->ipp_id == 0) {
@@ -1816,8 +1764,6 @@ static void ipp_subdrv_remove(struct drm_device *drm_dev, struct device *dev)
 {
 	struct exynos_drm_ippdrv *ippdrv;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* get ipp driver entry */
 	list_for_each_entry(ippdrv, &exynos_drm_ippdrv_list, drv_list) {
 		if (is_drm_iommu_supported(drm_dev))
@@ -1834,8 +1780,6 @@ static int ipp_subdrv_open(struct drm_device *drm_dev, struct device *dev,
 	struct drm_exynos_file_private *file_priv = file->driver_priv;
 	struct exynos_drm_ipp_private *priv;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
 	if (!priv) {
 		DRM_ERROR("failed to allocate priv.\n");
@@ -1846,7 +1790,7 @@ static int ipp_subdrv_open(struct drm_device *drm_dev, struct device *dev,
 
 	INIT_LIST_HEAD(&priv->event_list);
 
-	DRM_DEBUG_KMS("%s:done priv[0x%x]\n", __func__, (int)priv);
+	DRM_DEBUG_KMS("done priv[0x%x]\n", (int)priv);
 
 	return 0;
 }
@@ -1860,10 +1804,10 @@ static void ipp_subdrv_close(struct drm_device *drm_dev, struct device *dev,
 	struct drm_exynos_ipp_cmd_node *c_node, *tc_node;
 	int count = 0;
 
-	DRM_DEBUG_KMS("%s:for priv[0x%x]\n", __func__, (int)priv);
+	DRM_DEBUG_KMS("for priv[0x%x]\n", (int)priv);
 
 	if (list_empty(&exynos_drm_ippdrv_list)) {
-		DRM_DEBUG_KMS("%s:ippdrv_list is empty.\n", __func__);
+		DRM_DEBUG_KMS("ippdrv_list is empty.\n");
 		goto err_clear;
 	}
 
@@ -1873,8 +1817,8 @@ static void ipp_subdrv_close(struct drm_device *drm_dev, struct device *dev,
 
 		list_for_each_entry_safe(c_node, tc_node,
 			&ippdrv->cmd_list, list) {
-			DRM_DEBUG_KMS("%s:count[%d]ippdrv[0x%x]\n",
-				__func__, count++, (int)ippdrv);
+			DRM_DEBUG_KMS("count[%d]ippdrv[0x%x]\n",
+				count++, (int)ippdrv);
 
 			if (c_node->priv == priv) {
 				/*
@@ -1913,8 +1857,6 @@ static int ipp_probe(struct platform_device *pdev)
 	if (!ctx)
 		return -ENOMEM;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	mutex_init(&ctx->ipp_lock);
 	mutex_init(&ctx->prop_lock);
 
@@ -1978,8 +1920,6 @@ static int ipp_remove(struct platform_device *pdev)
 {
 	struct ipp_context *ctx = platform_get_drvdata(pdev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	/* unregister sub driver */
 	exynos_drm_subdrv_unregister(&ctx->subdrv);
 
@@ -1999,7 +1939,7 @@ static int ipp_remove(struct platform_device *pdev)
 
 static int ipp_power_ctrl(struct ipp_context *ctx, bool enable)
 {
-	DRM_DEBUG_KMS("%s:enable[%d]\n", __func__, enable);
+	DRM_DEBUG_KMS("enable[%d]\n", enable);
 
 	return 0;
 }
@@ -2009,8 +1949,6 @@ static int ipp_suspend(struct device *dev)
 {
 	struct ipp_context *ctx = get_ipp_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (pm_runtime_suspended(dev))
 		return 0;
 
@@ -2021,8 +1959,6 @@ static int ipp_resume(struct device *dev)
 {
 	struct ipp_context *ctx = get_ipp_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!pm_runtime_suspended(dev))
 		return ipp_power_ctrl(ctx, true);
 
@@ -2035,8 +1971,6 @@ static int ipp_runtime_suspend(struct device *dev)
 {
 	struct ipp_context *ctx = get_ipp_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	return ipp_power_ctrl(ctx, false);
 }
 
@@ -2044,8 +1978,6 @@ static int ipp_runtime_resume(struct device *dev)
 {
 	struct ipp_context *ctx = get_ipp_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	return ipp_power_ctrl(ctx, true);
 }
 #endif
diff --git a/drivers/gpu/drm/exynos/exynos_drm_plane.c b/drivers/gpu/drm/exynos/exynos_drm_plane.c
index 83efc662d65a..6ee55e68e0a2 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_plane.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_plane.c
@@ -81,8 +81,6 @@ int exynos_plane_mode_set(struct drm_plane *plane, struct drm_crtc *crtc,
 	int nr;
 	int i;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	nr = exynos_drm_fb_get_buf_cnt(fb);
 	for (i = 0; i < nr; i++) {
 		struct exynos_drm_gem_buf *buffer = exynos_drm_fb_buffer(fb, i);
@@ -159,8 +157,6 @@ void exynos_plane_dpms(struct drm_plane *plane, int mode)
 	struct exynos_plane *exynos_plane = to_exynos_plane(plane);
 	struct exynos_drm_overlay *overlay = &exynos_plane->overlay;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (mode == DRM_MODE_DPMS_ON) {
 		if (exynos_plane->enabled)
 			return;
@@ -189,8 +185,6 @@ exynos_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 {
 	int ret;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	ret = exynos_plane_mode_set(plane, crtc, fb, crtc_x, crtc_y,
 			crtc_w, crtc_h, src_x >> 16, src_y >> 16,
 			src_w >> 16, src_h >> 16);
@@ -207,8 +201,6 @@ exynos_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 
 static int exynos_disable_plane(struct drm_plane *plane)
 {
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	exynos_plane_dpms(plane, DRM_MODE_DPMS_OFF);
 
 	return 0;
@@ -218,8 +210,6 @@ static void exynos_plane_destroy(struct drm_plane *plane)
 {
 	struct exynos_plane *exynos_plane = to_exynos_plane(plane);
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	exynos_disable_plane(plane);
 	drm_plane_cleanup(plane);
 	kfree(exynos_plane);
@@ -233,8 +223,6 @@ static int exynos_plane_set_property(struct drm_plane *plane,
 	struct exynos_plane *exynos_plane = to_exynos_plane(plane);
 	struct exynos_drm_private *dev_priv = dev->dev_private;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (property == dev_priv->plane_zpos_property) {
 		exynos_plane->overlay.zpos = val;
 		return 0;
@@ -256,8 +244,6 @@ static void exynos_plane_attach_zpos_property(struct drm_plane *plane)
 	struct exynos_drm_private *dev_priv = dev->dev_private;
 	struct drm_property *prop;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	prop = dev_priv->plane_zpos_property;
 	if (!prop) {
 		prop = drm_property_create_range(dev, 0, "zpos", 0,
@@ -277,8 +263,6 @@ struct drm_plane *exynos_plane_init(struct drm_device *dev,
 	struct exynos_plane *exynos_plane;
 	int err;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	exynos_plane = kzalloc(sizeof(struct exynos_plane), GFP_KERNEL);
 	if (!exynos_plane) {
 		DRM_ERROR("failed to allocate plane\n");
diff --git a/drivers/gpu/drm/exynos/exynos_drm_rotator.c b/drivers/gpu/drm/exynos/exynos_drm_rotator.c
index 9b6c70964d71..427640aa5148 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_rotator.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_rotator.c
@@ -244,7 +244,7 @@ static int rotator_src_set_size(struct device *dev, int swap,
 	/* Get format */
 	fmt = rotator_reg_get_fmt(rot);
 	if (!rotator_check_reg_fmt(fmt)) {
-		DRM_ERROR("%s:invalid format.\n", __func__);
+		DRM_ERROR("invalid format.\n");
 		return -EINVAL;
 	}
 
@@ -287,7 +287,7 @@ static int rotator_src_set_addr(struct device *dev,
 		/* Get format */
 		fmt = rotator_reg_get_fmt(rot);
 		if (!rotator_check_reg_fmt(fmt)) {
-			DRM_ERROR("%s:invalid format.\n", __func__);
+			DRM_ERROR("invalid format.\n");
 			return -EINVAL;
 		}
 
@@ -381,7 +381,7 @@ static int rotator_dst_set_size(struct device *dev, int swap,
 	/* Get format */
 	fmt = rotator_reg_get_fmt(rot);
 	if (!rotator_check_reg_fmt(fmt)) {
-		DRM_ERROR("%s:invalid format.\n", __func__);
+		DRM_ERROR("invalid format.\n");
 		return -EINVAL;
 	}
 
@@ -422,7 +422,7 @@ static int rotator_dst_set_addr(struct device *dev,
 		/* Get format */
 		fmt = rotator_reg_get_fmt(rot);
 		if (!rotator_check_reg_fmt(fmt)) {
-			DRM_ERROR("%s:invalid format.\n", __func__);
+			DRM_ERROR("invalid format.\n");
 			return -EINVAL;
 		}
 
@@ -471,8 +471,6 @@ static int rotator_init_prop_list(struct exynos_drm_ippdrv *ippdrv)
 {
 	struct drm_exynos_ipp_prop_list *prop_list;
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	prop_list = devm_kzalloc(ippdrv->dev, sizeof(*prop_list), GFP_KERNEL);
 	if (!prop_list) {
 		DRM_ERROR("failed to alloc property list.\n");
@@ -502,7 +500,7 @@ static inline bool rotator_check_drm_fmt(u32 fmt)
 	case DRM_FORMAT_NV12:
 		return true;
 	default:
-		DRM_DEBUG_KMS("%s:not support format\n", __func__);
+		DRM_DEBUG_KMS("not support format\n");
 		return false;
 	}
 }
@@ -516,7 +514,7 @@ static inline bool rotator_check_drm_flip(enum drm_exynos_flip flip)
 	case EXYNOS_DRM_FLIP_BOTH:
 		return true;
 	default:
-		DRM_DEBUG_KMS("%s:invalid flip\n", __func__);
+		DRM_DEBUG_KMS("invalid flip\n");
 		return false;
 	}
 }
@@ -536,19 +534,18 @@ static int rotator_ippdrv_check_property(struct device *dev,
 
 	/* Check format configuration */
 	if (src_config->fmt != dst_config->fmt) {
-		DRM_DEBUG_KMS("%s:not support csc feature\n", __func__);
+		DRM_DEBUG_KMS("not support csc feature\n");
 		return -EINVAL;
 	}
 
 	if (!rotator_check_drm_fmt(dst_config->fmt)) {
-		DRM_DEBUG_KMS("%s:invalid format\n", __func__);
+		DRM_DEBUG_KMS("invalid format\n");
 		return -EINVAL;
 	}
 
 	/* Check transform configuration */
 	if (src_config->degree != EXYNOS_DRM_DEGREE_0) {
-		DRM_DEBUG_KMS("%s:not support source-side rotation\n",
-			__func__);
+		DRM_DEBUG_KMS("not support source-side rotation\n");
 		return -EINVAL;
 	}
 
@@ -561,51 +558,47 @@ static int rotator_ippdrv_check_property(struct device *dev,
 		/* No problem */
 		break;
 	default:
-		DRM_DEBUG_KMS("%s:invalid degree\n", __func__);
+		DRM_DEBUG_KMS("invalid degree\n");
 		return -EINVAL;
 	}
 
 	if (src_config->flip != EXYNOS_DRM_FLIP_NONE) {
-		DRM_DEBUG_KMS("%s:not support source-side flip\n", __func__);
+		DRM_DEBUG_KMS("not support source-side flip\n");
 		return -EINVAL;
 	}
 
 	if (!rotator_check_drm_flip(dst_config->flip)) {
-		DRM_DEBUG_KMS("%s:invalid flip\n", __func__);
+		DRM_DEBUG_KMS("invalid flip\n");
 		return -EINVAL;
 	}
 
 	/* Check size configuration */
 	if ((src_pos->x + src_pos->w > src_sz->hsize) ||
 		(src_pos->y + src_pos->h > src_sz->vsize)) {
-		DRM_DEBUG_KMS("%s:out of source buffer bound\n", __func__);
+		DRM_DEBUG_KMS("out of source buffer bound\n");
 		return -EINVAL;
 	}
 
 	if (swap) {
 		if ((dst_pos->x + dst_pos->h > dst_sz->vsize) ||
 			(dst_pos->y + dst_pos->w > dst_sz->hsize)) {
-			DRM_DEBUG_KMS("%s:out of destination buffer bound\n",
-				__func__);
+			DRM_DEBUG_KMS("out of destination buffer bound\n");
 			return -EINVAL;
 		}
 
 		if ((src_pos->w != dst_pos->h) || (src_pos->h != dst_pos->w)) {
-			DRM_DEBUG_KMS("%s:not support scale feature\n",
-				__func__);
+			DRM_DEBUG_KMS("not support scale feature\n");
 			return -EINVAL;
 		}
 	} else {
 		if ((dst_pos->x + dst_pos->w > dst_sz->hsize) ||
 			(dst_pos->y + dst_pos->h > dst_sz->vsize)) {
-			DRM_DEBUG_KMS("%s:out of destination buffer bound\n",
-				__func__);
+			DRM_DEBUG_KMS("out of destination buffer bound\n");
 			return -EINVAL;
 		}
 
 		if ((src_pos->w != dst_pos->w) || (src_pos->h != dst_pos->h)) {
-			DRM_DEBUG_KMS("%s:not support scale feature\n",
-				__func__);
+			DRM_DEBUG_KMS("not support scale feature\n");
 			return -EINVAL;
 		}
 	}
@@ -693,7 +686,7 @@ static int rotator_probe(struct platform_device *pdev)
 		goto err_ippdrv_register;
 	}
 
-	DRM_DEBUG_KMS("%s:ippdrv[0x%x]\n", __func__, (int)ippdrv);
+	DRM_DEBUG_KMS("ippdrv[0x%x]\n", (int)ippdrv);
 
 	platform_set_drvdata(pdev, rot);
 
@@ -752,8 +745,6 @@ static struct platform_device_id rotator_driver_ids[] = {
 
 static int rotator_clk_crtl(struct rot_context *rot, bool enable)
 {
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (enable) {
 		clk_enable(rot->clock);
 		rot->suspended = false;
@@ -771,8 +762,6 @@ static int rotator_suspend(struct device *dev)
 {
 	struct rot_context *rot = dev_get_drvdata(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (pm_runtime_suspended(dev))
 		return 0;
 
@@ -783,8 +772,6 @@ static int rotator_resume(struct device *dev)
 {
 	struct rot_context *rot = dev_get_drvdata(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	if (!pm_runtime_suspended(dev))
 		return rotator_clk_crtl(rot, true);
 
@@ -797,8 +784,6 @@ static int rotator_runtime_suspend(struct device *dev)
 {
 	struct rot_context *rot = dev_get_drvdata(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	return  rotator_clk_crtl(rot, false);
 }
 
@@ -806,8 +791,6 @@ static int rotator_runtime_resume(struct device *dev)
 {
 	struct rot_context *rot = dev_get_drvdata(dev);
 
-	DRM_DEBUG_KMS("%s\n", __func__);
-
 	return  rotator_clk_crtl(rot, true);
 }
 #endif
diff --git a/drivers/gpu/drm/exynos/exynos_drm_vidi.c b/drivers/gpu/drm/exynos/exynos_drm_vidi.c
index 24376c194a5e..41cc74d83e4e 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_vidi.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_vidi.c
@@ -89,8 +89,6 @@ static bool vidi_display_is_connected(struct device *dev)
 {
 	struct vidi_context *ctx = get_vidi_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * connection request would come from user side
 	 * to do hotplug through specific ioctl.
@@ -105,8 +103,6 @@ static struct edid *vidi_get_edid(struct device *dev,
 	struct edid *edid;
 	int edid_len;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * the edid data comes from user side and it would be set
 	 * to ctx->raw_edid through specific ioctl.
@@ -128,17 +124,13 @@ static struct edid *vidi_get_edid(struct device *dev,
 
 static void *vidi_get_panel(struct device *dev)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO. */
 
 	return NULL;
 }
 
-static int vidi_check_timing(struct device *dev, void *timing)
+static int vidi_check_mode(struct device *dev, struct drm_display_mode *mode)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO. */
 
 	return 0;
@@ -146,8 +138,6 @@ static int vidi_check_timing(struct device *dev, void *timing)
 
 static int vidi_display_power_on(struct device *dev, int mode)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO */
 
 	return 0;
@@ -158,7 +148,7 @@ static struct exynos_drm_display_ops vidi_display_ops = {
 	.is_connected = vidi_display_is_connected,
 	.get_edid = vidi_get_edid,
 	.get_panel = vidi_get_panel,
-	.check_timing = vidi_check_timing,
+	.check_mode = vidi_check_mode,
 	.power_on = vidi_display_power_on,
 };
 
@@ -166,7 +156,7 @@ static void vidi_dpms(struct device *subdrv_dev, int mode)
 {
 	struct vidi_context *ctx = get_vidi_context(subdrv_dev);
 
-	DRM_DEBUG_KMS("%s, %d\n", __FILE__, mode);
+	DRM_DEBUG_KMS("%d\n", mode);
 
 	mutex_lock(&ctx->lock);
 
@@ -196,8 +186,6 @@ static void vidi_apply(struct device *subdrv_dev)
 	struct vidi_win_data *win_data;
 	int i;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	for (i = 0; i < WINDOWS_NR; i++) {
 		win_data = &ctx->win_data[i];
 		if (win_data->enabled && (ovl_ops && ovl_ops->commit))
@@ -212,8 +200,6 @@ static void vidi_commit(struct device *dev)
 {
 	struct vidi_context *ctx = get_vidi_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return;
 }
@@ -222,8 +208,6 @@ static int vidi_enable_vblank(struct device *dev)
 {
 	struct vidi_context *ctx = get_vidi_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return -EPERM;
 
@@ -246,8 +230,6 @@ static void vidi_disable_vblank(struct device *dev)
 {
 	struct vidi_context *ctx = get_vidi_context(dev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return;
 
@@ -271,8 +253,6 @@ static void vidi_win_mode_set(struct device *dev,
 	int win;
 	unsigned long offset;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!overlay) {
 		dev_err(dev, "overlay is NULL\n");
 		return;
@@ -282,7 +262,7 @@ static void vidi_win_mode_set(struct device *dev,
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
 
-	if (win < 0 || win > WINDOWS_NR)
+	if (win < 0 || win >= WINDOWS_NR)
 		return;
 
 	offset = overlay->fb_x * (overlay->bpp >> 3);
@@ -324,15 +304,13 @@ static void vidi_win_commit(struct device *dev, int zpos)
 	struct vidi_win_data *win_data;
 	int win = zpos;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (ctx->suspended)
 		return;
 
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
 
-	if (win < 0 || win > WINDOWS_NR)
+	if (win < 0 || win >= WINDOWS_NR)
 		return;
 
 	win_data = &ctx->win_data[win];
@@ -351,12 +329,10 @@ static void vidi_win_disable(struct device *dev, int zpos)
 	struct vidi_win_data *win_data;
 	int win = zpos;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (win == DEFAULT_ZPOS)
 		win = ctx->default_win;
 
-	if (win < 0 || win > WINDOWS_NR)
+	if (win < 0 || win >= WINDOWS_NR)
 		return;
 
 	win_data = &ctx->win_data[win];
@@ -407,8 +383,6 @@ static void vidi_fake_vblank_handler(struct work_struct *work)
 
 static int vidi_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/*
 	 * enable drm irq mode.
 	 * - with irq_enabled = 1, we can use the vblank feature.
@@ -431,8 +405,6 @@ static int vidi_subdrv_probe(struct drm_device *drm_dev, struct device *dev)
 
 static void vidi_subdrv_remove(struct drm_device *drm_dev, struct device *dev)
 {
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	/* TODO. */
 }
 
@@ -441,11 +413,6 @@ static int vidi_power_on(struct vidi_context *ctx, bool enable)
 	struct exynos_drm_subdrv *subdrv = &ctx->subdrv;
 	struct device *dev = subdrv->dev;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
-	if (enable != false && enable != true)
-		return -EINVAL;
-
 	if (enable) {
 		ctx->suspended = false;
 
@@ -483,8 +450,6 @@ static int vidi_store_connection(struct device *dev,
 	struct vidi_context *ctx = get_vidi_context(dev);
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	ret = kstrtoint(buf, 0, &ctx->connected);
 	if (ret)
 		return ret;
@@ -522,8 +487,6 @@ int vidi_connection_ioctl(struct drm_device *drm_dev, void *data,
 	struct drm_exynos_vidi_connection *vidi = data;
 	int edid_len;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	if (!vidi) {
 		DRM_DEBUG_KMS("user data for vidi is null.\n");
 		return -EINVAL;
@@ -592,8 +555,6 @@ static int vidi_probe(struct platform_device *pdev)
 	struct exynos_drm_subdrv *subdrv;
 	int ret;
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	ctx = devm_kzalloc(dev, sizeof(*ctx), GFP_KERNEL);
 	if (!ctx)
 		return -ENOMEM;
@@ -625,8 +586,6 @@ static int vidi_remove(struct platform_device *pdev)
 {
 	struct vidi_context *ctx = platform_get_drvdata(pdev);
 
-	DRM_DEBUG_KMS("%s\n", __FILE__);
-
 	exynos_drm_subdrv_unregister(&ctx->subdrv);
 
 	if (ctx->raw_edid != (struct edid *)fake_edid_info) {
diff --git a/drivers/gpu/drm/exynos/exynos_hdmi.c b/drivers/gpu/drm/exynos/exynos_hdmi.c
index fd1426dca882..62ef5971ac3c 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmi.c
+++ b/drivers/gpu/drm/exynos/exynos_hdmi.c
@@ -83,6 +83,7 @@ struct hdmi_resources {
 	struct clk			*sclk_pixel;
 	struct clk			*sclk_hdmiphy;
 	struct clk			*hdmiphy;
+	struct clk			*mout_hdmi;
 	struct regulator_bulk_data	*regul_bulk;
 	int				regul_count;
 };
@@ -689,8 +690,6 @@ static void hdmi_reg_infoframe(struct hdmi_context *hdata,
 	u32 mod;
 	u32 vic;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mod = hdmi_reg_read(hdata, HDMI_MODE_SEL);
 	if (hdata->dvi_mode) {
 		hdmi_reg_writeb(hdata, HDMI_VSI_CON,
@@ -755,8 +754,6 @@ static struct edid *hdmi_get_edid(void *ctx, struct drm_connector *connector)
 	struct edid *raw_edid;
 	struct hdmi_context *hdata = ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (!hdata->ddc_port)
 		return ERR_PTR(-ENODEV);
 
@@ -777,8 +774,6 @@ static int hdmi_find_phy_conf(struct hdmi_context *hdata, u32 pixel_clock)
 	const struct hdmiphy_config *confs;
 	int count, i;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (hdata->type == HDMI_TYPE13) {
 		confs = hdmiphy_v13_configs;
 		count = ARRAY_SIZE(hdmiphy_v13_configs);
@@ -796,18 +791,17 @@ static int hdmi_find_phy_conf(struct hdmi_context *hdata, u32 pixel_clock)
 	return -EINVAL;
 }
 
-static int hdmi_check_timing(void *ctx, struct fb_videomode *timing)
+static int hdmi_check_mode(void *ctx, struct drm_display_mode *mode)
 {
 	struct hdmi_context *hdata = ctx;
 	int ret;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
+	DRM_DEBUG_KMS("xres=%d, yres=%d, refresh=%d, intl=%d clock=%d\n",
+		mode->hdisplay, mode->vdisplay, mode->vrefresh,
+		(mode->flags & DRM_MODE_FLAG_INTERLACE) ? true :
+		false, mode->clock * 1000);
 
-	DRM_DEBUG_KMS("[%d]x[%d] [%d]Hz [%x]\n", timing->xres,
-			timing->yres, timing->refresh,
-			timing->vmode);
-
-	ret = hdmi_find_phy_conf(hdata, timing->pixclock);
+	ret = hdmi_find_phy_conf(hdata, mode->clock * 1000);
 	if (ret < 0)
 		return ret;
 	return 0;
@@ -1042,7 +1036,7 @@ static void hdmi_conf_init(struct hdmi_context *hdata)
 	}
 }
 
-static void hdmi_v13_timing_apply(struct hdmi_context *hdata)
+static void hdmi_v13_mode_apply(struct hdmi_context *hdata)
 {
 	const struct hdmi_tg_regs *tg = &hdata->mode_conf.conf.v13_conf.tg;
 	const struct hdmi_v13_core_regs *core =
@@ -1118,9 +1112,9 @@ static void hdmi_v13_timing_apply(struct hdmi_context *hdata)
 		hdmi_regs_dump(hdata, "timing apply");
 	}
 
-	clk_disable(hdata->res.sclk_hdmi);
-	clk_set_parent(hdata->res.sclk_hdmi, hdata->res.sclk_hdmiphy);
-	clk_enable(hdata->res.sclk_hdmi);
+	clk_disable_unprepare(hdata->res.sclk_hdmi);
+	clk_set_parent(hdata->res.mout_hdmi, hdata->res.sclk_hdmiphy);
+	clk_prepare_enable(hdata->res.sclk_hdmi);
 
 	/* enable HDMI and timing generator */
 	hdmi_reg_writemask(hdata, HDMI_CON_0, ~0, HDMI_EN);
@@ -1131,7 +1125,7 @@ static void hdmi_v13_timing_apply(struct hdmi_context *hdata)
 		hdmi_reg_writemask(hdata, HDMI_TG_CMD, ~0, HDMI_TG_EN);
 }
 
-static void hdmi_v14_timing_apply(struct hdmi_context *hdata)
+static void hdmi_v14_mode_apply(struct hdmi_context *hdata)
 {
 	const struct hdmi_tg_regs *tg = &hdata->mode_conf.conf.v14_conf.tg;
 	const struct hdmi_v14_core_regs *core =
@@ -1285,9 +1279,9 @@ static void hdmi_v14_timing_apply(struct hdmi_context *hdata)
 		hdmi_regs_dump(hdata, "timing apply");
 	}
 
-	clk_disable(hdata->res.sclk_hdmi);
-	clk_set_parent(hdata->res.sclk_hdmi, hdata->res.sclk_hdmiphy);
-	clk_enable(hdata->res.sclk_hdmi);
+	clk_disable_unprepare(hdata->res.sclk_hdmi);
+	clk_set_parent(hdata->res.mout_hdmi, hdata->res.sclk_hdmiphy);
+	clk_prepare_enable(hdata->res.sclk_hdmi);
 
 	/* enable HDMI and timing generator */
 	hdmi_reg_writemask(hdata, HDMI_CON_0, ~0, HDMI_EN);
@@ -1298,12 +1292,12 @@ static void hdmi_v14_timing_apply(struct hdmi_context *hdata)
 		hdmi_reg_writemask(hdata, HDMI_TG_CMD, ~0, HDMI_TG_EN);
 }
 
-static void hdmi_timing_apply(struct hdmi_context *hdata)
+static void hdmi_mode_apply(struct hdmi_context *hdata)
 {
 	if (hdata->type == HDMI_TYPE13)
-		hdmi_v13_timing_apply(hdata);
+		hdmi_v13_mode_apply(hdata);
 	else
-		hdmi_v14_timing_apply(hdata);
+		hdmi_v14_mode_apply(hdata);
 }
 
 static void hdmiphy_conf_reset(struct hdmi_context *hdata)
@@ -1311,9 +1305,9 @@ static void hdmiphy_conf_reset(struct hdmi_context *hdata)
 	u8 buffer[2];
 	u32 reg;
 
-	clk_disable(hdata->res.sclk_hdmi);
-	clk_set_parent(hdata->res.sclk_hdmi, hdata->res.sclk_pixel);
-	clk_enable(hdata->res.sclk_hdmi);
+	clk_disable_unprepare(hdata->res.sclk_hdmi);
+	clk_set_parent(hdata->res.mout_hdmi, hdata->res.sclk_pixel);
+	clk_prepare_enable(hdata->res.sclk_hdmi);
 
 	/* operation mode */
 	buffer[0] = 0x1f;
@@ -1336,8 +1330,6 @@ static void hdmiphy_conf_reset(struct hdmi_context *hdata)
 
 static void hdmiphy_poweron(struct hdmi_context *hdata)
 {
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (hdata->type == HDMI_TYPE14)
 		hdmi_reg_writemask(hdata, HDMI_PHY_CON_0, 0,
 			HDMI_PHY_POWER_OFF_EN);
@@ -1345,8 +1337,6 @@ static void hdmiphy_poweron(struct hdmi_context *hdata)
 
 static void hdmiphy_poweroff(struct hdmi_context *hdata)
 {
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (hdata->type == HDMI_TYPE14)
 		hdmi_reg_writemask(hdata, HDMI_PHY_CON_0, ~0,
 			HDMI_PHY_POWER_OFF_EN);
@@ -1410,8 +1400,6 @@ static void hdmiphy_conf_apply(struct hdmi_context *hdata)
 
 static void hdmi_conf_apply(struct hdmi_context *hdata)
 {
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	hdmiphy_conf_reset(hdata);
 	hdmiphy_conf_apply(hdata);
 
@@ -1423,7 +1411,7 @@ static void hdmi_conf_apply(struct hdmi_context *hdata)
 	hdmi_audio_init(hdata);
 
 	/* setting core registers */
-	hdmi_timing_apply(hdata);
+	hdmi_mode_apply(hdata);
 	hdmi_audio_control(hdata, true);
 
 	hdmi_regs_dump(hdata, "start");
@@ -1569,8 +1557,7 @@ static void hdmi_v14_mode_set(struct hdmi_context *hdata,
 			(m->vsync_start - m->vdisplay) / 2);
 		hdmi_set_reg(core->v2_blank, 2, m->vtotal / 2);
 		hdmi_set_reg(core->v1_blank, 2, (m->vtotal - m->vdisplay) / 2);
-		hdmi_set_reg(core->v_blank_f0, 2, (m->vtotal +
-			((m->vsync_end - m->vsync_start) * 4) + 5) / 2);
+		hdmi_set_reg(core->v_blank_f0, 2, m->vtotal - m->vdisplay / 2);
 		hdmi_set_reg(core->v_blank_f1, 2, m->vtotal);
 		hdmi_set_reg(core->v_sync_line_aft_2, 2, (m->vtotal / 2) + 7);
 		hdmi_set_reg(core->v_sync_line_aft_1, 2, (m->vtotal / 2) + 2);
@@ -1580,7 +1567,10 @@ static void hdmi_v14_mode_set(struct hdmi_context *hdata,
 			(m->htotal / 2) + (m->hsync_start - m->hdisplay));
 		hdmi_set_reg(tg->vact_st, 2, (m->vtotal - m->vdisplay) / 2);
 		hdmi_set_reg(tg->vact_sz, 2, m->vdisplay / 2);
-		hdmi_set_reg(tg->vact_st2, 2, 0x249);/* Reset value + 1*/
+		hdmi_set_reg(tg->vact_st2, 2, m->vtotal - m->vdisplay / 2);
+		hdmi_set_reg(tg->vsync2, 2, (m->vtotal / 2) + 1);
+		hdmi_set_reg(tg->vsync_bot_hdmi, 2, (m->vtotal / 2) + 1);
+		hdmi_set_reg(tg->field_bot_hdmi, 2, (m->vtotal / 2) + 1);
 		hdmi_set_reg(tg->vact_st3, 2, 0x0);
 		hdmi_set_reg(tg->vact_st4, 2, 0x0);
 	} else {
@@ -1602,6 +1592,9 @@ static void hdmi_v14_mode_set(struct hdmi_context *hdata,
 		hdmi_set_reg(tg->vact_st2, 2, 0x248); /* Reset value */
 		hdmi_set_reg(tg->vact_st3, 2, 0x47b); /* Reset value */
 		hdmi_set_reg(tg->vact_st4, 2, 0x6ae); /* Reset value */
+		hdmi_set_reg(tg->vsync2, 2, 0x233); /* Reset value */
+		hdmi_set_reg(tg->vsync_bot_hdmi, 2, 0x233); /* Reset value */
+		hdmi_set_reg(tg->field_bot_hdmi, 2, 0x233); /* Reset value */
 	}
 
 	/* Following values & calculations are same irrespective of mode type */
@@ -1633,22 +1626,19 @@ static void hdmi_v14_mode_set(struct hdmi_context *hdata,
 	hdmi_set_reg(tg->hact_sz, 2, m->hdisplay);
 	hdmi_set_reg(tg->v_fsz, 2, m->vtotal);
 	hdmi_set_reg(tg->vsync, 2, 0x1);
-	hdmi_set_reg(tg->vsync2, 2, 0x233); /* Reset value */
 	hdmi_set_reg(tg->field_chg, 2, 0x233); /* Reset value */
 	hdmi_set_reg(tg->vsync_top_hdmi, 2, 0x1); /* Reset value */
-	hdmi_set_reg(tg->vsync_bot_hdmi, 2, 0x233); /* Reset value */
 	hdmi_set_reg(tg->field_top_hdmi, 2, 0x1); /* Reset value */
-	hdmi_set_reg(tg->field_bot_hdmi, 2, 0x233); /* Reset value */
 	hdmi_set_reg(tg->tg_3d, 1, 0x0);
 }
 
-static void hdmi_mode_set(void *ctx, void *mode)
+static void hdmi_mode_set(void *ctx, struct drm_display_mode *mode)
 {
 	struct hdmi_context *hdata = ctx;
 	struct drm_display_mode *m = mode;
 
-	DRM_DEBUG_KMS("[%s]: xres=%d, yres=%d, refresh=%d, intl=%s\n",
-		__func__, m->hdisplay, m->vdisplay,
+	DRM_DEBUG_KMS("xres=%d, yres=%d, refresh=%d, intl=%s\n",
+		m->hdisplay, m->vdisplay,
 		m->vrefresh, (m->flags & DRM_MODE_FLAG_INTERLACE) ?
 		"INTERLACED" : "PROGERESSIVE");
 
@@ -1661,8 +1651,6 @@ static void hdmi_mode_set(void *ctx, void *mode)
 static void hdmi_get_max_resol(void *ctx, unsigned int *width,
 					unsigned int *height)
 {
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	*width = MAX_WIDTH;
 	*height = MAX_HEIGHT;
 }
@@ -1671,8 +1659,6 @@ static void hdmi_commit(void *ctx)
 {
 	struct hdmi_context *hdata = ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mutex_lock(&hdata->hdmi_mutex);
 	if (!hdata->powered) {
 		mutex_unlock(&hdata->hdmi_mutex);
@@ -1687,8 +1673,6 @@ static void hdmi_poweron(struct hdmi_context *hdata)
 {
 	struct hdmi_resources *res = &hdata->res;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mutex_lock(&hdata->hdmi_mutex);
 	if (hdata->powered) {
 		mutex_unlock(&hdata->hdmi_mutex);
@@ -1699,10 +1683,12 @@ static void hdmi_poweron(struct hdmi_context *hdata)
 
 	mutex_unlock(&hdata->hdmi_mutex);
 
-	regulator_bulk_enable(res->regul_count, res->regul_bulk);
-	clk_enable(res->hdmiphy);
-	clk_enable(res->hdmi);
-	clk_enable(res->sclk_hdmi);
+	if (regulator_bulk_enable(res->regul_count, res->regul_bulk))
+		DRM_DEBUG_KMS("failed to enable regulator bulk\n");
+
+	clk_prepare_enable(res->hdmiphy);
+	clk_prepare_enable(res->hdmi);
+	clk_prepare_enable(res->sclk_hdmi);
 
 	hdmiphy_poweron(hdata);
 }
@@ -1711,8 +1697,6 @@ static void hdmi_poweroff(struct hdmi_context *hdata)
 {
 	struct hdmi_resources *res = &hdata->res;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mutex_lock(&hdata->hdmi_mutex);
 	if (!hdata->powered)
 		goto out;
@@ -1725,9 +1709,9 @@ static void hdmi_poweroff(struct hdmi_context *hdata)
 	hdmiphy_conf_reset(hdata);
 	hdmiphy_poweroff(hdata);
 
-	clk_disable(res->sclk_hdmi);
-	clk_disable(res->hdmi);
-	clk_disable(res->hdmiphy);
+	clk_disable_unprepare(res->sclk_hdmi);
+	clk_disable_unprepare(res->hdmi);
+	clk_disable_unprepare(res->hdmiphy);
 	regulator_bulk_disable(res->regul_count, res->regul_bulk);
 
 	mutex_lock(&hdata->hdmi_mutex);
@@ -1742,7 +1726,7 @@ static void hdmi_dpms(void *ctx, int mode)
 {
 	struct hdmi_context *hdata = ctx;
 
-	DRM_DEBUG_KMS("[%d] %s mode %d\n", __LINE__, __func__, mode);
+	DRM_DEBUG_KMS("mode %d\n", mode);
 
 	switch (mode) {
 	case DRM_MODE_DPMS_ON:
@@ -1765,7 +1749,7 @@ static struct exynos_hdmi_ops hdmi_ops = {
 	/* display */
 	.is_connected	= hdmi_is_connected,
 	.get_edid	= hdmi_get_edid,
-	.check_timing	= hdmi_check_timing,
+	.check_mode	= hdmi_check_mode,
 
 	/* manager */
 	.mode_set	= hdmi_mode_set,
@@ -1831,8 +1815,13 @@ static int hdmi_resources_init(struct hdmi_context *hdata)
 		DRM_ERROR("failed to get clock 'hdmiphy'\n");
 		goto fail;
 	}
+	res->mout_hdmi = devm_clk_get(dev, "mout_hdmi");
+	if (IS_ERR(res->mout_hdmi)) {
+		DRM_ERROR("failed to get clock 'mout_hdmi'\n");
+		goto fail;
+	}
 
-	clk_set_parent(res->sclk_hdmi, res->sclk_pixel);
+	clk_set_parent(res->mout_hdmi, res->sclk_pixel);
 
 	res->regul_bulk = devm_kzalloc(dev, ARRAY_SIZE(supply) *
 		sizeof(res->regul_bulk[0]), GFP_KERNEL);
@@ -1877,7 +1866,6 @@ static struct s5p_hdmi_platform_data *drm_hdmi_dt_parse_pdata
 {
 	struct device_node *np = dev->of_node;
 	struct s5p_hdmi_platform_data *pd;
-	enum of_gpio_flags flags;
 	u32 value;
 
 	pd = devm_kzalloc(dev, sizeof(*pd), GFP_KERNEL);
@@ -1891,7 +1879,7 @@ static struct s5p_hdmi_platform_data *drm_hdmi_dt_parse_pdata
 		goto err_data;
 	}
 
-	pd->hpd_gpio = of_get_named_gpio_flags(np, "hpd-gpio", 0, &flags);
+	pd->hpd_gpio = of_get_named_gpio(np, "hpd-gpio", 0);
 
 	return pd;
 
@@ -1930,6 +1918,9 @@ static struct of_device_id hdmi_match_types[] = {
 		.compatible = "samsung,exynos5-hdmi",
 		.data	= (void	*)HDMI_TYPE14,
 	}, {
+		.compatible = "samsung,exynos4212-hdmi",
+		.data	= (void	*)HDMI_TYPE14,
+	}, {
 		/* end node */
 	}
 };
@@ -1944,8 +1935,6 @@ static int hdmi_probe(struct platform_device *pdev)
 	struct resource *res;
 	int ret;
 
-	DRM_DEBUG_KMS("[%d]\n", __LINE__);
-
 	if (dev->of_node) {
 		pdata = drm_hdmi_dt_parse_pdata(dev);
 		if (IS_ERR(pdata)) {
@@ -2071,8 +2060,6 @@ static int hdmi_remove(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	pm_runtime_disable(dev);
 
 	/* hdmiphy i2c driver */
@@ -2089,8 +2076,6 @@ static int hdmi_suspend(struct device *dev)
 	struct exynos_drm_hdmi_context *ctx = get_hdmi_context(dev);
 	struct hdmi_context *hdata = ctx->ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	disable_irq(hdata->irq);
 
 	hdata->hpd = false;
@@ -2098,7 +2083,7 @@ static int hdmi_suspend(struct device *dev)
 		drm_helper_hpd_irq_event(ctx->drm_dev);
 
 	if (pm_runtime_suspended(dev)) {
-		DRM_DEBUG_KMS("%s : Already suspended\n", __func__);
+		DRM_DEBUG_KMS("Already suspended\n");
 		return 0;
 	}
 
@@ -2112,14 +2097,12 @@ static int hdmi_resume(struct device *dev)
 	struct exynos_drm_hdmi_context *ctx = get_hdmi_context(dev);
 	struct hdmi_context *hdata = ctx->ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	hdata->hpd = gpio_get_value(hdata->hpd_gpio);
 
 	enable_irq(hdata->irq);
 
 	if (!pm_runtime_suspended(dev)) {
-		DRM_DEBUG_KMS("%s : Already resumed\n", __func__);
+		DRM_DEBUG_KMS("Already resumed\n");
 		return 0;
 	}
 
@@ -2134,7 +2117,6 @@ static int hdmi_runtime_suspend(struct device *dev)
 {
 	struct exynos_drm_hdmi_context *ctx = get_hdmi_context(dev);
 	struct hdmi_context *hdata = ctx->ctx;
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
 	hdmi_poweroff(hdata);
 
@@ -2145,7 +2127,6 @@ static int hdmi_runtime_resume(struct device *dev)
 {
 	struct exynos_drm_hdmi_context *ctx = get_hdmi_context(dev);
 	struct hdmi_context *hdata = ctx->ctx;
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
 
 	hdmi_poweron(hdata);
 
diff --git a/drivers/gpu/drm/exynos/exynos_hdmiphy.c b/drivers/gpu/drm/exynos/exynos_hdmiphy.c
index ea49d132ecf6..ef04255076c7 100644
--- a/drivers/gpu/drm/exynos/exynos_hdmiphy.c
+++ b/drivers/gpu/drm/exynos/exynos_hdmiphy.c
@@ -51,6 +51,10 @@ static struct of_device_id hdmiphy_match_types[] = {
 	{
 		.compatible = "samsung,exynos5-hdmiphy",
 	}, {
+		.compatible = "samsung,exynos4210-hdmiphy",
+	}, {
+		.compatible = "samsung,exynos4212-hdmiphy",
+	}, {
 		/* end node */
 	}
 };
diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c
index 7c197d3820c5..42ffb71c63bc 100644
--- a/drivers/gpu/drm/exynos/exynos_mixer.c
+++ b/drivers/gpu/drm/exynos/exynos_mixer.c
@@ -78,6 +78,7 @@ struct mixer_resources {
 enum mixer_version_id {
 	MXR_VER_0_0_0_16,
 	MXR_VER_16_0_33_0,
+	MXR_VER_128_0_0_184,
 };
 
 struct mixer_context {
@@ -283,17 +284,19 @@ static void mixer_cfg_scan(struct mixer_context *ctx, unsigned int height)
 	val = (ctx->interlace ? MXR_CFG_SCAN_INTERLACE :
 				MXR_CFG_SCAN_PROGRASSIVE);
 
-	/* choosing between porper HD and SD mode */
-	if (height <= 480)
-		val |= MXR_CFG_SCAN_NTSC | MXR_CFG_SCAN_SD;
-	else if (height <= 576)
-		val |= MXR_CFG_SCAN_PAL | MXR_CFG_SCAN_SD;
-	else if (height <= 720)
-		val |= MXR_CFG_SCAN_HD_720 | MXR_CFG_SCAN_HD;
-	else if (height <= 1080)
-		val |= MXR_CFG_SCAN_HD_1080 | MXR_CFG_SCAN_HD;
-	else
-		val |= MXR_CFG_SCAN_HD_720 | MXR_CFG_SCAN_HD;
+	if (ctx->mxr_ver != MXR_VER_128_0_0_184) {
+		/* choosing between proper HD and SD mode */
+		if (height <= 480)
+			val |= MXR_CFG_SCAN_NTSC | MXR_CFG_SCAN_SD;
+		else if (height <= 576)
+			val |= MXR_CFG_SCAN_PAL | MXR_CFG_SCAN_SD;
+		else if (height <= 720)
+			val |= MXR_CFG_SCAN_HD_720 | MXR_CFG_SCAN_HD;
+		else if (height <= 1080)
+			val |= MXR_CFG_SCAN_HD_1080 | MXR_CFG_SCAN_HD;
+		else
+			val |= MXR_CFG_SCAN_HD_720 | MXR_CFG_SCAN_HD;
+	}
 
 	mixer_reg_writemask(res, MXR_CFG, val, MXR_CFG_SCAN_MASK);
 }
@@ -376,7 +379,7 @@ static void vp_video_buffer(struct mixer_context *ctx, int win)
 	unsigned long flags;
 	struct hdmi_win_data *win_data;
 	unsigned int x_ratio, y_ratio;
-	unsigned int buf_num;
+	unsigned int buf_num = 1;
 	dma_addr_t luma_addr[2], chroma_addr[2];
 	bool tiled_mode = false;
 	bool crcb_mode = false;
@@ -557,6 +560,14 @@ static void mixer_graph_buffer(struct mixer_context *ctx, int win)
 	/* setup geometry */
 	mixer_reg_write(res, MXR_GRAPHIC_SPAN(win), win_data->fb_width);
 
+	/* setup display size */
+	if (ctx->mxr_ver == MXR_VER_128_0_0_184 &&
+		win == MIXER_DEFAULT_WIN) {
+		val  = MXR_MXR_RES_HEIGHT(win_data->fb_height);
+		val |= MXR_MXR_RES_WIDTH(win_data->fb_width);
+		mixer_reg_write(res, MXR_RESOLUTION, val);
+	}
+
 	val  = MXR_GRP_WH_WIDTH(win_data->crtc_width);
 	val |= MXR_GRP_WH_HEIGHT(win_data->crtc_height);
 	val |= MXR_GRP_WH_H_SCALE(x_ratio);
@@ -581,7 +592,8 @@ static void mixer_graph_buffer(struct mixer_context *ctx, int win)
 	mixer_cfg_layer(ctx, win, true);
 
 	/* layer update mandatory for mixer 16.0.33.0 */
-	if (ctx->mxr_ver == MXR_VER_16_0_33_0)
+	if (ctx->mxr_ver == MXR_VER_16_0_33_0 ||
+		ctx->mxr_ver == MXR_VER_128_0_0_184)
 		mixer_layer_update(ctx);
 
 	mixer_run(ctx);
@@ -696,8 +708,6 @@ static int mixer_enable_vblank(void *ctx, int pipe)
 	struct mixer_context *mixer_ctx = ctx;
 	struct mixer_resources *res = &mixer_ctx->mixer_res;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mixer_ctx->pipe = pipe;
 
 	/* enable vsync interrupt */
@@ -712,8 +722,6 @@ static void mixer_disable_vblank(void *ctx)
 	struct mixer_context *mixer_ctx = ctx;
 	struct mixer_resources *res = &mixer_ctx->mixer_res;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	/* disable vsync interrupt */
 	mixer_reg_writemask(res, MXR_INT_EN, 0, MXR_INT_EN_VSYNC);
 }
@@ -725,8 +733,6 @@ static void mixer_win_mode_set(void *ctx,
 	struct hdmi_win_data *win_data;
 	int win;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (!overlay) {
 		DRM_ERROR("overlay is NULL\n");
 		return;
@@ -742,7 +748,7 @@ static void mixer_win_mode_set(void *ctx,
 	if (win == DEFAULT_ZPOS)
 		win = MIXER_DEFAULT_WIN;
 
-	if (win < 0 || win > MIXER_WIN_NR) {
+	if (win < 0 || win >= MIXER_WIN_NR) {
 		DRM_ERROR("mixer window[%d] is wrong\n", win);
 		return;
 	}
@@ -776,7 +782,7 @@ static void mixer_win_commit(void *ctx, int win)
 {
 	struct mixer_context *mixer_ctx = ctx;
 
-	DRM_DEBUG_KMS("[%d] %s, win: %d\n", __LINE__, __func__, win);
+	DRM_DEBUG_KMS("win: %d\n", win);
 
 	mutex_lock(&mixer_ctx->mixer_mutex);
 	if (!mixer_ctx->powered) {
@@ -799,7 +805,7 @@ static void mixer_win_disable(void *ctx, int win)
 	struct mixer_resources *res = &mixer_ctx->mixer_res;
 	unsigned long flags;
 
-	DRM_DEBUG_KMS("[%d] %s, win: %d\n", __LINE__, __func__, win);
+	DRM_DEBUG_KMS("win: %d\n", win);
 
 	mutex_lock(&mixer_ctx->mixer_mutex);
 	if (!mixer_ctx->powered) {
@@ -820,17 +826,21 @@ static void mixer_win_disable(void *ctx, int win)
 	mixer_ctx->win_data[win].enabled = false;
 }
 
-static int mixer_check_timing(void *ctx, struct fb_videomode *timing)
+static int mixer_check_mode(void *ctx, struct drm_display_mode *mode)
 {
+	struct mixer_context *mixer_ctx = ctx;
 	u32 w, h;
 
-	w = timing->xres;
-	h = timing->yres;
+	w = mode->hdisplay;
+	h = mode->vdisplay;
 
-	DRM_DEBUG_KMS("%s : xres=%d, yres=%d, refresh=%d, intl=%d\n",
-		__func__, timing->xres, timing->yres,
-		timing->refresh, (timing->vmode &
-		FB_VMODE_INTERLACED) ? true : false);
+	DRM_DEBUG_KMS("xres=%d, yres=%d, refresh=%d, intl=%d\n",
+		mode->hdisplay, mode->vdisplay, mode->vrefresh,
+		(mode->flags & DRM_MODE_FLAG_INTERLACE) ? 1 : 0);
+
+	if (mixer_ctx->mxr_ver == MXR_VER_0_0_0_16 ||
+		mixer_ctx->mxr_ver == MXR_VER_128_0_0_184)
+		return 0;
 
 	if ((w >= 464 && w <= 720 && h >= 261 && h <= 576) ||
 		(w >= 1024 && w <= 1280 && h >= 576 && h <= 720) ||
@@ -891,8 +901,6 @@ static void mixer_poweron(struct mixer_context *ctx)
 {
 	struct mixer_resources *res = &ctx->mixer_res;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mutex_lock(&ctx->mixer_mutex);
 	if (ctx->powered) {
 		mutex_unlock(&ctx->mixer_mutex);
@@ -901,10 +909,10 @@ static void mixer_poweron(struct mixer_context *ctx)
 	ctx->powered = true;
 	mutex_unlock(&ctx->mixer_mutex);
 
-	clk_enable(res->mixer);
+	clk_prepare_enable(res->mixer);
 	if (ctx->vp_enabled) {
-		clk_enable(res->vp);
-		clk_enable(res->sclk_mixer);
+		clk_prepare_enable(res->vp);
+		clk_prepare_enable(res->sclk_mixer);
 	}
 
 	mixer_reg_write(res, MXR_INT_EN, ctx->int_en);
@@ -917,8 +925,6 @@ static void mixer_poweroff(struct mixer_context *ctx)
 {
 	struct mixer_resources *res = &ctx->mixer_res;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mutex_lock(&ctx->mixer_mutex);
 	if (!ctx->powered)
 		goto out;
@@ -928,10 +934,10 @@ static void mixer_poweroff(struct mixer_context *ctx)
 
 	ctx->int_en = mixer_reg_read(res, MXR_INT_EN);
 
-	clk_disable(res->mixer);
+	clk_disable_unprepare(res->mixer);
 	if (ctx->vp_enabled) {
-		clk_disable(res->vp);
-		clk_disable(res->sclk_mixer);
+		clk_disable_unprepare(res->vp);
+		clk_disable_unprepare(res->sclk_mixer);
 	}
 
 	mutex_lock(&ctx->mixer_mutex);
@@ -945,8 +951,6 @@ static void mixer_dpms(void *ctx, int mode)
 {
 	struct mixer_context *mixer_ctx = ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	switch (mode) {
 	case DRM_MODE_DPMS_ON:
 		if (pm_runtime_suspended(mixer_ctx->dev))
@@ -978,7 +982,7 @@ static struct exynos_mixer_ops mixer_ops = {
 	.win_disable		= mixer_win_disable,
 
 	/* display */
-	.check_timing		= mixer_check_timing,
+	.check_mode		= mixer_check_mode,
 };
 
 static irqreturn_t mixer_irq_handler(int irq, void *arg)
@@ -1128,12 +1132,17 @@ static int vp_resources_init(struct exynos_drm_hdmi_context *ctx,
 	return 0;
 }
 
-static struct mixer_drv_data exynos5_mxr_drv_data = {
+static struct mixer_drv_data exynos5420_mxr_drv_data = {
+	.version = MXR_VER_128_0_0_184,
+	.is_vp_enabled = 0,
+};
+
+static struct mixer_drv_data exynos5250_mxr_drv_data = {
 	.version = MXR_VER_16_0_33_0,
 	.is_vp_enabled = 0,
 };
 
-static struct mixer_drv_data exynos4_mxr_drv_data = {
+static struct mixer_drv_data exynos4210_mxr_drv_data = {
 	.version = MXR_VER_0_0_0_16,
 	.is_vp_enabled = 1,
 };
@@ -1141,10 +1150,10 @@ static struct mixer_drv_data exynos4_mxr_drv_data = {
 static struct platform_device_id mixer_driver_types[] = {
 	{
 		.name		= "s5p-mixer",
-		.driver_data	= (unsigned long)&exynos4_mxr_drv_data,
+		.driver_data	= (unsigned long)&exynos4210_mxr_drv_data,
 	}, {
 		.name		= "exynos5-mixer",
-		.driver_data	= (unsigned long)&exynos5_mxr_drv_data,
+		.driver_data	= (unsigned long)&exynos5250_mxr_drv_data,
 	}, {
 		/* end node */
 	}
@@ -1153,7 +1162,13 @@ static struct platform_device_id mixer_driver_types[] = {
 static struct of_device_id mixer_match_types[] = {
 	{
 		.compatible = "samsung,exynos5-mixer",
-		.data	= &exynos5_mxr_drv_data,
+		.data	= &exynos5250_mxr_drv_data,
+	}, {
+		.compatible = "samsung,exynos5250-mixer",
+		.data	= &exynos5250_mxr_drv_data,
+	}, {
+		.compatible = "samsung,exynos5420-mixer",
+		.data	= &exynos5420_mxr_drv_data,
 	}, {
 		/* end node */
 	}
@@ -1186,8 +1201,7 @@ static int mixer_probe(struct platform_device *pdev)
 
 	if (dev->of_node) {
 		const struct of_device_id *match;
-		match = of_match_node(of_match_ptr(mixer_match_types),
-							  dev->of_node);
+		match = of_match_node(mixer_match_types, dev->of_node);
 		drv = (struct mixer_drv_data *)match->data;
 	} else {
 		drv = (struct mixer_drv_data *)
@@ -1251,10 +1265,8 @@ static int mixer_suspend(struct device *dev)
 	struct exynos_drm_hdmi_context *drm_hdmi_ctx = get_mixer_context(dev);
 	struct mixer_context *ctx = drm_hdmi_ctx->ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (pm_runtime_suspended(dev)) {
-		DRM_DEBUG_KMS("%s : Already suspended\n", __func__);
+		DRM_DEBUG_KMS("Already suspended\n");
 		return 0;
 	}
 
@@ -1268,10 +1280,8 @@ static int mixer_resume(struct device *dev)
 	struct exynos_drm_hdmi_context *drm_hdmi_ctx = get_mixer_context(dev);
 	struct mixer_context *ctx = drm_hdmi_ctx->ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	if (!pm_runtime_suspended(dev)) {
-		DRM_DEBUG_KMS("%s : Already resumed\n", __func__);
+		DRM_DEBUG_KMS("Already resumed\n");
 		return 0;
 	}
 
@@ -1287,8 +1297,6 @@ static int mixer_runtime_suspend(struct device *dev)
 	struct exynos_drm_hdmi_context *drm_hdmi_ctx = get_mixer_context(dev);
 	struct mixer_context *ctx = drm_hdmi_ctx->ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mixer_poweroff(ctx);
 
 	return 0;
@@ -1299,8 +1307,6 @@ static int mixer_runtime_resume(struct device *dev)
 	struct exynos_drm_hdmi_context *drm_hdmi_ctx = get_mixer_context(dev);
 	struct mixer_context *ctx = drm_hdmi_ctx->ctx;
 
-	DRM_DEBUG_KMS("[%d] %s\n", __LINE__, __func__);
-
 	mixer_poweron(ctx);
 
 	return 0;
diff --git a/drivers/gpu/drm/exynos/regs-mixer.h b/drivers/gpu/drm/exynos/regs-mixer.h
index 5d8dbc0301e6..4537026bc385 100644
--- a/drivers/gpu/drm/exynos/regs-mixer.h
+++ b/drivers/gpu/drm/exynos/regs-mixer.h
@@ -44,6 +44,9 @@
 #define MXR_CM_COEFF_Y			0x0080
 #define MXR_CM_COEFF_CB			0x0084
 #define MXR_CM_COEFF_CR			0x0088
+#define MXR_MO				0x0304
+#define MXR_RESOLUTION			0x0310
+
 #define MXR_GRAPHIC0_BASE_S		0x2024
 #define MXR_GRAPHIC1_BASE_S		0x2044
 
@@ -119,6 +122,10 @@
 #define MXR_GRP_WH_WIDTH(x)		MXR_MASK_VAL(x, 26, 16)
 #define MXR_GRP_WH_HEIGHT(x)		MXR_MASK_VAL(x, 10, 0)
 
+/* bits for MXR_RESOLUTION */
+#define MXR_MXR_RES_HEIGHT(x)		MXR_MASK_VAL(x, 26, 16)
+#define MXR_MXR_RES_WIDTH(x)		MXR_MASK_VAL(x, 10, 0)
+
 /* bits for MXR_GRAPHICn_SXY */
 #define MXR_GRP_SXY_SX(x)		MXR_MASK_VAL(x, 26, 16)
 #define MXR_GRP_SXY_SY(x)		MXR_MASK_VAL(x, 10, 0)
diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index 91f3ac6cef35..40034ecefd3b 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -36,6 +36,7 @@ i915-y := i915_drv.o i915_dma.o i915_irq.o \
 	  intel_overlay.o \
 	  intel_sprite.o \
 	  intel_opregion.o \
+	  intel_sideband.o \
 	  dvo_ch7xxx.o \
 	  dvo_ch7017.o \
 	  dvo_ivch.o \
diff --git a/drivers/gpu/drm/i915/dvo_ch7xxx.c b/drivers/gpu/drm/i915/dvo_ch7xxx.c
index 3edd981e0770..757e0fa11043 100644
--- a/drivers/gpu/drm/i915/dvo_ch7xxx.c
+++ b/drivers/gpu/drm/i915/dvo_ch7xxx.c
@@ -32,12 +32,14 @@ SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #define CH7xxx_REG_DID		0x4b
 
 #define CH7011_VID		0x83 /* 7010 as well */
+#define CH7010B_VID		0x05
 #define CH7009A_VID		0x84
 #define CH7009B_VID		0x85
 #define CH7301_VID		0x95
 
 #define CH7xxx_VID		0x84
 #define CH7xxx_DID		0x17
+#define CH7010_DID		0x16
 
 #define CH7xxx_NUM_REGS		0x4c
 
@@ -87,11 +89,20 @@ static struct ch7xxx_id_struct {
 	char *name;
 } ch7xxx_ids[] = {
 	{ CH7011_VID, "CH7011" },
+	{ CH7010B_VID, "CH7010B" },
 	{ CH7009A_VID, "CH7009A" },
 	{ CH7009B_VID, "CH7009B" },
 	{ CH7301_VID, "CH7301" },
 };
 
+static struct ch7xxx_did_struct {
+	uint8_t did;
+	char *name;
+} ch7xxx_dids[] = {
+	{ CH7xxx_DID, "CH7XXX" },
+	{ CH7010_DID, "CH7010B" },
+};
+
 struct ch7xxx_priv {
 	bool quiet;
 };
@@ -108,6 +119,18 @@ static char *ch7xxx_get_id(uint8_t vid)
 	return NULL;
 }
 
+static char *ch7xxx_get_did(uint8_t did)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(ch7xxx_dids); i++) {
+		if (ch7xxx_dids[i].did == did)
+			return ch7xxx_dids[i].name;
+	}
+
+	return NULL;
+}
+
 /** Reads an 8 bit register */
 static bool ch7xxx_readb(struct intel_dvo_device *dvo, int addr, uint8_t *ch)
 {
@@ -179,7 +202,7 @@ static bool ch7xxx_init(struct intel_dvo_device *dvo,
 	/* this will detect the CH7xxx chip on the specified i2c bus */
 	struct ch7xxx_priv *ch7xxx;
 	uint8_t vendor, device;
-	char *name;
+	char *name, *devid;
 
 	ch7xxx = kzalloc(sizeof(struct ch7xxx_priv), GFP_KERNEL);
 	if (ch7xxx == NULL)
@@ -204,7 +227,8 @@ static bool ch7xxx_init(struct intel_dvo_device *dvo,
 	if (!ch7xxx_readb(dvo, CH7xxx_REG_DID, &device))
 		goto out;
 
-	if (device != CH7xxx_DID) {
+	devid = ch7xxx_get_did(device);
+	if (!devid) {
 		DRM_DEBUG_KMS("ch7xxx not detected; got 0x%02x from %s "
 				"slave %d.\n",
 			  vendor, adapter->name, dvo->slave_addr);
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index e913d325d5b8..47d6c748057e 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -61,11 +61,11 @@ static int i915_capabilities(struct seq_file *m, void *data)
 
 	seq_printf(m, "gen: %d\n", info->gen);
 	seq_printf(m, "pch: %d\n", INTEL_PCH_TYPE(dev));
-#define DEV_INFO_FLAG(x) seq_printf(m, #x ": %s\n", yesno(info->x))
-#define DEV_INFO_SEP ;
-	DEV_INFO_FLAGS;
-#undef DEV_INFO_FLAG
-#undef DEV_INFO_SEP
+#define PRINT_FLAG(x)  seq_printf(m, #x ": %s\n", yesno(info->x))
+#define SEP_SEMICOLON ;
+	DEV_INFO_FOR_EACH_FLAG(PRINT_FLAG, SEP_SEMICOLON);
+#undef PRINT_FLAG
+#undef SEP_SEMICOLON
 
 	return 0;
 }
@@ -196,6 +196,32 @@ static int i915_gem_object_list_info(struct seq_file *m, void *data)
 	} \
 } while (0)
 
+struct file_stats {
+	int count;
+	size_t total, active, inactive, unbound;
+};
+
+static int per_file_stats(int id, void *ptr, void *data)
+{
+	struct drm_i915_gem_object *obj = ptr;
+	struct file_stats *stats = data;
+
+	stats->count++;
+	stats->total += obj->base.size;
+
+	if (obj->gtt_space) {
+		if (!list_empty(&obj->ring_list))
+			stats->active += obj->base.size;
+		else
+			stats->inactive += obj->base.size;
+	} else {
+		if (!list_empty(&obj->global_list))
+			stats->unbound += obj->base.size;
+	}
+
+	return 0;
+}
+
 static int i915_gem_object_info(struct seq_file *m, void* data)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
@@ -204,6 +230,7 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 	u32 count, mappable_count, purgeable_count;
 	size_t size, mappable_size, purgeable_size;
 	struct drm_i915_gem_object *obj;
+	struct drm_file *file;
 	int ret;
 
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
@@ -215,7 +242,7 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 		   dev_priv->mm.object_memory);
 
 	size = count = mappable_size = mappable_count = 0;
-	count_objects(&dev_priv->mm.bound_list, gtt_list);
+	count_objects(&dev_priv->mm.bound_list, global_list);
 	seq_printf(m, "%u [%u] objects, %zu [%zu] bytes in gtt\n",
 		   count, mappable_count, size, mappable_size);
 
@@ -230,7 +257,7 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 		   count, mappable_count, size, mappable_size);
 
 	size = count = purgeable_size = purgeable_count = 0;
-	list_for_each_entry(obj, &dev_priv->mm.unbound_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.unbound_list, global_list) {
 		size += obj->base.size, ++count;
 		if (obj->madv == I915_MADV_DONTNEED)
 			purgeable_size += obj->base.size, ++purgeable_count;
@@ -238,7 +265,7 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 	seq_printf(m, "%u unbound objects, %zu bytes\n", count, size);
 
 	size = count = mappable_size = mappable_count = 0;
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
 		if (obj->fault_mappable) {
 			size += obj->gtt_space->size;
 			++count;
@@ -263,6 +290,21 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 		   dev_priv->gtt.total,
 		   dev_priv->gtt.mappable_end - dev_priv->gtt.start);
 
+	seq_printf(m, "\n");
+	list_for_each_entry_reverse(file, &dev->filelist, lhead) {
+		struct file_stats stats;
+
+		memset(&stats, 0, sizeof(stats));
+		idr_for_each(&file->object_idr, per_file_stats, &stats);
+		seq_printf(m, "%s: %u objects, %zu bytes (%zu active, %zu inactive, %zu unbound)\n",
+			   get_pid_task(file->pid, PIDTYPE_PID)->comm,
+			   stats.count,
+			   stats.total,
+			   stats.active,
+			   stats.inactive,
+			   stats.unbound);
+	}
+
 	mutex_unlock(&dev->struct_mutex);
 
 	return 0;
@@ -283,7 +325,7 @@ static int i915_gem_gtt_info(struct seq_file *m, void* data)
 		return ret;
 
 	total_obj_size = total_gtt_size = count = 0;
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
 		if (list == PINNED_LIST && obj->pin_count == 0)
 			continue;
 
@@ -570,6 +612,7 @@ static const char *ring_str(int ring)
 	case RCS: return "render";
 	case VCS: return "bsd";
 	case BCS: return "blt";
+	case VECS: return "vebox";
 	default: return "";
 	}
 }
@@ -604,73 +647,187 @@ static const char *purgeable_flag(int purgeable)
 	return purgeable ? " purgeable" : "";
 }
 
-static void print_error_buffers(struct seq_file *m,
+static bool __i915_error_ok(struct drm_i915_error_state_buf *e)
+{
+
+	if (!e->err && WARN(e->bytes > (e->size - 1), "overflow")) {
+		e->err = -ENOSPC;
+		return false;
+	}
+
+	if (e->bytes == e->size - 1 || e->err)
+		return false;
+
+	return true;
+}
+
+static bool __i915_error_seek(struct drm_i915_error_state_buf *e,
+			      unsigned len)
+{
+	if (e->pos + len <= e->start) {
+		e->pos += len;
+		return false;
+	}
+
+	/* First vsnprintf needs to fit in its entirety for memmove */
+	if (len >= e->size) {
+		e->err = -EIO;
+		return false;
+	}
+
+	return true;
+}
+
+static void __i915_error_advance(struct drm_i915_error_state_buf *e,
+				 unsigned len)
+{
+	/* If this is first printf in this window, adjust it so that
+	 * start position matches start of the buffer
+	 */
+
+	if (e->pos < e->start) {
+		const size_t off = e->start - e->pos;
+
+		/* Should not happen but be paranoid */
+		if (off > len || e->bytes) {
+			e->err = -EIO;
+			return;
+		}
+
+		memmove(e->buf, e->buf + off, len - off);
+		e->bytes = len - off;
+		e->pos = e->start;
+		return;
+	}
+
+	e->bytes += len;
+	e->pos += len;
+}
+
+static void i915_error_vprintf(struct drm_i915_error_state_buf *e,
+			       const char *f, va_list args)
+{
+	unsigned len;
+
+	if (!__i915_error_ok(e))
+		return;
+
+	/* Seek the first printf which is hits start position */
+	if (e->pos < e->start) {
+		len = vsnprintf(NULL, 0, f, args);
+		if (!__i915_error_seek(e, len))
+			return;
+	}
+
+	len = vsnprintf(e->buf + e->bytes, e->size - e->bytes, f, args);
+	if (len >= e->size - e->bytes)
+		len = e->size - e->bytes - 1;
+
+	__i915_error_advance(e, len);
+}
+
+static void i915_error_puts(struct drm_i915_error_state_buf *e,
+			    const char *str)
+{
+	unsigned len;
+
+	if (!__i915_error_ok(e))
+		return;
+
+	len = strlen(str);
+
+	/* Seek the first printf which is hits start position */
+	if (e->pos < e->start) {
+		if (!__i915_error_seek(e, len))
+			return;
+	}
+
+	if (len >= e->size - e->bytes)
+		len = e->size - e->bytes - 1;
+	memcpy(e->buf + e->bytes, str, len);
+
+	__i915_error_advance(e, len);
+}
+
+void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...)
+{
+	va_list args;
+
+	va_start(args, f);
+	i915_error_vprintf(e, f, args);
+	va_end(args);
+}
+
+#define err_printf(e, ...) i915_error_printf(e, __VA_ARGS__)
+#define err_puts(e, s) i915_error_puts(e, s)
+
+static void print_error_buffers(struct drm_i915_error_state_buf *m,
 				const char *name,
 				struct drm_i915_error_buffer *err,
 				int count)
 {
-	seq_printf(m, "%s [%d]:\n", name, count);
+	err_printf(m, "%s [%d]:\n", name, count);
 
 	while (count--) {
-		seq_printf(m, "  %08x %8u %02x %02x %x %x%s%s%s%s%s%s%s",
+		err_printf(m, "  %08x %8u %02x %02x %x %x",
 			   err->gtt_offset,
 			   err->size,
 			   err->read_domains,
 			   err->write_domain,
-			   err->rseqno, err->wseqno,
-			   pin_flag(err->pinned),
-			   tiling_flag(err->tiling),
-			   dirty_flag(err->dirty),
-			   purgeable_flag(err->purgeable),
-			   err->ring != -1 ? " " : "",
-			   ring_str(err->ring),
-			   cache_level_str(err->cache_level));
+			   err->rseqno, err->wseqno);
+		err_puts(m, pin_flag(err->pinned));
+		err_puts(m, tiling_flag(err->tiling));
+		err_puts(m, dirty_flag(err->dirty));
+		err_puts(m, purgeable_flag(err->purgeable));
+		err_puts(m, err->ring != -1 ? " " : "");
+		err_puts(m, ring_str(err->ring));
+		err_puts(m, cache_level_str(err->cache_level));
 
 		if (err->name)
-			seq_printf(m, " (name: %d)", err->name);
+			err_printf(m, " (name: %d)", err->name);
 		if (err->fence_reg != I915_FENCE_REG_NONE)
-			seq_printf(m, " (fence: %d)", err->fence_reg);
+			err_printf(m, " (fence: %d)", err->fence_reg);
 
-		seq_printf(m, "\n");
+		err_puts(m, "\n");
 		err++;
 	}
 }
 
-static void i915_ring_error_state(struct seq_file *m,
+static void i915_ring_error_state(struct drm_i915_error_state_buf *m,
 				  struct drm_device *dev,
 				  struct drm_i915_error_state *error,
 				  unsigned ring)
 {
 	BUG_ON(ring >= I915_NUM_RINGS); /* shut up confused gcc */
-	seq_printf(m, "%s command stream:\n", ring_str(ring));
-	seq_printf(m, "  HEAD: 0x%08x\n", error->head[ring]);
-	seq_printf(m, "  TAIL: 0x%08x\n", error->tail[ring]);
-	seq_printf(m, "  CTL: 0x%08x\n", error->ctl[ring]);
-	seq_printf(m, "  ACTHD: 0x%08x\n", error->acthd[ring]);
-	seq_printf(m, "  IPEIR: 0x%08x\n", error->ipeir[ring]);
-	seq_printf(m, "  IPEHR: 0x%08x\n", error->ipehr[ring]);
-	seq_printf(m, "  INSTDONE: 0x%08x\n", error->instdone[ring]);
+	err_printf(m, "%s command stream:\n", ring_str(ring));
+	err_printf(m, "  HEAD: 0x%08x\n", error->head[ring]);
+	err_printf(m, "  TAIL: 0x%08x\n", error->tail[ring]);
+	err_printf(m, "  CTL: 0x%08x\n", error->ctl[ring]);
+	err_printf(m, "  ACTHD: 0x%08x\n", error->acthd[ring]);
+	err_printf(m, "  IPEIR: 0x%08x\n", error->ipeir[ring]);
+	err_printf(m, "  IPEHR: 0x%08x\n", error->ipehr[ring]);
+	err_printf(m, "  INSTDONE: 0x%08x\n", error->instdone[ring]);
 	if (ring == RCS && INTEL_INFO(dev)->gen >= 4)
-		seq_printf(m, "  BBADDR: 0x%08llx\n", error->bbaddr);
+		err_printf(m, "  BBADDR: 0x%08llx\n", error->bbaddr);
 
 	if (INTEL_INFO(dev)->gen >= 4)
-		seq_printf(m, "  INSTPS: 0x%08x\n", error->instps[ring]);
-	seq_printf(m, "  INSTPM: 0x%08x\n", error->instpm[ring]);
-	seq_printf(m, "  FADDR: 0x%08x\n", error->faddr[ring]);
+		err_printf(m, "  INSTPS: 0x%08x\n", error->instps[ring]);
+	err_printf(m, "  INSTPM: 0x%08x\n", error->instpm[ring]);
+	err_printf(m, "  FADDR: 0x%08x\n", error->faddr[ring]);
 	if (INTEL_INFO(dev)->gen >= 6) {
-		seq_printf(m, "  RC PSMI: 0x%08x\n", error->rc_psmi[ring]);
-		seq_printf(m, "  FAULT_REG: 0x%08x\n", error->fault_reg[ring]);
-		seq_printf(m, "  SYNC_0: 0x%08x [last synced 0x%08x]\n",
+		err_printf(m, "  RC PSMI: 0x%08x\n", error->rc_psmi[ring]);
+		err_printf(m, "  FAULT_REG: 0x%08x\n", error->fault_reg[ring]);
+		err_printf(m, "  SYNC_0: 0x%08x [last synced 0x%08x]\n",
 			   error->semaphore_mboxes[ring][0],
 			   error->semaphore_seqno[ring][0]);
-		seq_printf(m, "  SYNC_1: 0x%08x [last synced 0x%08x]\n",
+		err_printf(m, "  SYNC_1: 0x%08x [last synced 0x%08x]\n",
 			   error->semaphore_mboxes[ring][1],
 			   error->semaphore_seqno[ring][1]);
 	}
-	seq_printf(m, "  seqno: 0x%08x\n", error->seqno[ring]);
-	seq_printf(m, "  waiting: %s\n", yesno(error->waiting[ring]));
-	seq_printf(m, "  ring->head: 0x%08x\n", error->cpu_ring_head[ring]);
-	seq_printf(m, "  ring->tail: 0x%08x\n", error->cpu_ring_tail[ring]);
+	err_printf(m, "  seqno: 0x%08x\n", error->seqno[ring]);
+	err_printf(m, "  waiting: %s\n", yesno(error->waiting[ring]));
+	err_printf(m, "  ring->head: 0x%08x\n", error->cpu_ring_head[ring]);
+	err_printf(m, "  ring->tail: 0x%08x\n", error->cpu_ring_tail[ring]);
 }
 
 struct i915_error_state_file_priv {
@@ -678,9 +835,11 @@ struct i915_error_state_file_priv {
 	struct drm_i915_error_state *error;
 };
 
-static int i915_error_state(struct seq_file *m, void *unused)
+
+static int i915_error_state(struct i915_error_state_file_priv *error_priv,
+			    struct drm_i915_error_state_buf *m)
+
 {
-	struct i915_error_state_file_priv *error_priv = m->private;
 	struct drm_device *dev = error_priv->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct drm_i915_error_state *error = error_priv->error;
@@ -688,34 +847,35 @@ static int i915_error_state(struct seq_file *m, void *unused)
 	int i, j, page, offset, elt;
 
 	if (!error) {
-		seq_printf(m, "no error state collected\n");
+		err_printf(m, "no error state collected\n");
 		return 0;
 	}
 
-	seq_printf(m, "Time: %ld s %ld us\n", error->time.tv_sec,
+	err_printf(m, "Time: %ld s %ld us\n", error->time.tv_sec,
 		   error->time.tv_usec);
-	seq_printf(m, "Kernel: " UTS_RELEASE "\n");
-	seq_printf(m, "PCI ID: 0x%04x\n", dev->pci_device);
-	seq_printf(m, "EIR: 0x%08x\n", error->eir);
-	seq_printf(m, "IER: 0x%08x\n", error->ier);
-	seq_printf(m, "PGTBL_ER: 0x%08x\n", error->pgtbl_er);
-	seq_printf(m, "FORCEWAKE: 0x%08x\n", error->forcewake);
-	seq_printf(m, "DERRMR: 0x%08x\n", error->derrmr);
-	seq_printf(m, "CCID: 0x%08x\n", error->ccid);
+	err_printf(m, "Kernel: " UTS_RELEASE "\n");
+	err_printf(m, "PCI ID: 0x%04x\n", dev->pci_device);
+	err_printf(m, "EIR: 0x%08x\n", error->eir);
+	err_printf(m, "IER: 0x%08x\n", error->ier);
+	err_printf(m, "PGTBL_ER: 0x%08x\n", error->pgtbl_er);
+	err_printf(m, "FORCEWAKE: 0x%08x\n", error->forcewake);
+	err_printf(m, "DERRMR: 0x%08x\n", error->derrmr);
+	err_printf(m, "CCID: 0x%08x\n", error->ccid);
 
 	for (i = 0; i < dev_priv->num_fence_regs; i++)
-		seq_printf(m, "  fence[%d] = %08llx\n", i, error->fence[i]);
+		err_printf(m, "  fence[%d] = %08llx\n", i, error->fence[i]);
 
 	for (i = 0; i < ARRAY_SIZE(error->extra_instdone); i++)
-		seq_printf(m, "  INSTDONE_%d: 0x%08x\n", i, error->extra_instdone[i]);
+		err_printf(m, "  INSTDONE_%d: 0x%08x\n", i,
+			   error->extra_instdone[i]);
 
 	if (INTEL_INFO(dev)->gen >= 6) {
-		seq_printf(m, "ERROR: 0x%08x\n", error->error);
-		seq_printf(m, "DONE_REG: 0x%08x\n", error->done_reg);
+		err_printf(m, "ERROR: 0x%08x\n", error->error);
+		err_printf(m, "DONE_REG: 0x%08x\n", error->done_reg);
 	}
 
 	if (INTEL_INFO(dev)->gen == 7)
-		seq_printf(m, "ERR_INT: 0x%08x\n", error->err_int);
+		err_printf(m, "ERR_INT: 0x%08x\n", error->err_int);
 
 	for_each_ring(ring, dev_priv, i)
 		i915_ring_error_state(m, dev, error, i);
@@ -734,24 +894,25 @@ static int i915_error_state(struct seq_file *m, void *unused)
 		struct drm_i915_error_object *obj;
 
 		if ((obj = error->ring[i].batchbuffer)) {
-			seq_printf(m, "%s --- gtt_offset = 0x%08x\n",
+			err_printf(m, "%s --- gtt_offset = 0x%08x\n",
 				   dev_priv->ring[i].name,
 				   obj->gtt_offset);
 			offset = 0;
 			for (page = 0; page < obj->page_count; page++) {
 				for (elt = 0; elt < PAGE_SIZE/4; elt++) {
-					seq_printf(m, "%08x :  %08x\n", offset, obj->pages[page][elt]);
+					err_printf(m, "%08x :  %08x\n", offset,
+						   obj->pages[page][elt]);
 					offset += 4;
 				}
 			}
 		}
 
 		if (error->ring[i].num_requests) {
-			seq_printf(m, "%s --- %d requests\n",
+			err_printf(m, "%s --- %d requests\n",
 				   dev_priv->ring[i].name,
 				   error->ring[i].num_requests);
 			for (j = 0; j < error->ring[i].num_requests; j++) {
-				seq_printf(m, "  seqno 0x%08x, emitted %ld, tail 0x%08x\n",
+				err_printf(m, "  seqno 0x%08x, emitted %ld, tail 0x%08x\n",
 					   error->ring[i].requests[j].seqno,
 					   error->ring[i].requests[j].jiffies,
 					   error->ring[i].requests[j].tail);
@@ -759,13 +920,13 @@ static int i915_error_state(struct seq_file *m, void *unused)
 		}
 
 		if ((obj = error->ring[i].ringbuffer)) {
-			seq_printf(m, "%s --- ringbuffer = 0x%08x\n",
+			err_printf(m, "%s --- ringbuffer = 0x%08x\n",
 				   dev_priv->ring[i].name,
 				   obj->gtt_offset);
 			offset = 0;
 			for (page = 0; page < obj->page_count; page++) {
 				for (elt = 0; elt < PAGE_SIZE/4; elt++) {
-					seq_printf(m, "%08x :  %08x\n",
+					err_printf(m, "%08x :  %08x\n",
 						   offset,
 						   obj->pages[page][elt]);
 					offset += 4;
@@ -775,12 +936,12 @@ static int i915_error_state(struct seq_file *m, void *unused)
 
 		obj = error->ring[i].ctx;
 		if (obj) {
-			seq_printf(m, "%s --- HW Context = 0x%08x\n",
+			err_printf(m, "%s --- HW Context = 0x%08x\n",
 				   dev_priv->ring[i].name,
 				   obj->gtt_offset);
 			offset = 0;
 			for (elt = 0; elt < PAGE_SIZE/16; elt += 4) {
-				seq_printf(m, "[%04x] %08x %08x %08x %08x\n",
+				err_printf(m, "[%04x] %08x %08x %08x %08x\n",
 					   offset,
 					   obj->pages[0][elt],
 					   obj->pages[0][elt+1],
@@ -806,8 +967,7 @@ i915_error_state_write(struct file *filp,
 		       size_t cnt,
 		       loff_t *ppos)
 {
-	struct seq_file *m = filp->private_data;
-	struct i915_error_state_file_priv *error_priv = m->private;
+	struct i915_error_state_file_priv *error_priv = filp->private_data;
 	struct drm_device *dev = error_priv->dev;
 	int ret;
 
@@ -842,25 +1002,81 @@ static int i915_error_state_open(struct inode *inode, struct file *file)
 		kref_get(&error_priv->error->ref);
 	spin_unlock_irqrestore(&dev_priv->gpu_error.lock, flags);
 
-	return single_open(file, i915_error_state, error_priv);
+	file->private_data = error_priv;
+
+	return 0;
 }
 
 static int i915_error_state_release(struct inode *inode, struct file *file)
 {
-	struct seq_file *m = file->private_data;
-	struct i915_error_state_file_priv *error_priv = m->private;
+	struct i915_error_state_file_priv *error_priv = file->private_data;
 
 	if (error_priv->error)
 		kref_put(&error_priv->error->ref, i915_error_state_free);
 	kfree(error_priv);
 
-	return single_release(inode, file);
+	return 0;
+}
+
+static ssize_t i915_error_state_read(struct file *file, char __user *userbuf,
+				     size_t count, loff_t *pos)
+{
+	struct i915_error_state_file_priv *error_priv = file->private_data;
+	struct drm_i915_error_state_buf error_str;
+	loff_t tmp_pos = 0;
+	ssize_t ret_count = 0;
+	int ret = 0;
+
+	memset(&error_str, 0, sizeof(error_str));
+
+	/* We need to have enough room to store any i915_error_state printf
+	 * so that we can move it to start position.
+	 */
+	error_str.size = count + 1 > PAGE_SIZE ? count + 1 : PAGE_SIZE;
+	error_str.buf = kmalloc(error_str.size,
+				GFP_TEMPORARY | __GFP_NORETRY | __GFP_NOWARN);
+
+	if (error_str.buf == NULL) {
+		error_str.size = PAGE_SIZE;
+		error_str.buf = kmalloc(error_str.size, GFP_TEMPORARY);
+	}
+
+	if (error_str.buf == NULL) {
+		error_str.size = 128;
+		error_str.buf = kmalloc(error_str.size, GFP_TEMPORARY);
+	}
+
+	if (error_str.buf == NULL)
+		return -ENOMEM;
+
+	error_str.start = *pos;
+
+	ret = i915_error_state(error_priv, &error_str);
+	if (ret)
+		goto out;
+
+	if (error_str.bytes == 0 && error_str.err) {
+		ret = error_str.err;
+		goto out;
+	}
+
+	ret_count = simple_read_from_buffer(userbuf, count, &tmp_pos,
+					    error_str.buf,
+					    error_str.bytes);
+
+	if (ret_count < 0)
+		ret = ret_count;
+	else
+		*pos = error_str.start + ret_count;
+out:
+	kfree(error_str.buf);
+	return ret ?: ret_count;
 }
 
 static const struct file_operations i915_error_state_fops = {
 	.owner = THIS_MODULE,
 	.open = i915_error_state_open,
-	.read = seq_read,
+	.read = i915_error_state_read,
 	.write = i915_error_state_write,
 	.llseek = default_llseek,
 	.release = i915_error_state_release,
@@ -941,7 +1157,7 @@ static int i915_cur_delayinfo(struct seq_file *m, void *unused)
 			   MEMSTAT_VID_SHIFT);
 		seq_printf(m, "Current P-state: %d\n",
 			   (rgvstat & MEMSTAT_PSTATE_MASK) >> MEMSTAT_PSTATE_SHIFT);
-	} else if (IS_GEN6(dev) || IS_GEN7(dev)) {
+	} else if ((IS_GEN6(dev) || IS_GEN7(dev)) && !IS_VALLEYVIEW(dev)) {
 		u32 gt_perf_status = I915_READ(GEN6_GT_PERF_STATUS);
 		u32 rp_state_limits = I915_READ(GEN6_RP_STATE_LIMITS);
 		u32 rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
@@ -1009,6 +1225,26 @@ static int i915_cur_delayinfo(struct seq_file *m, void *unused)
 
 		seq_printf(m, "Max overclocked frequency: %dMHz\n",
 			   dev_priv->rps.hw_max * GT_FREQUENCY_MULTIPLIER);
+	} else if (IS_VALLEYVIEW(dev)) {
+		u32 freq_sts, val;
+
+		mutex_lock(&dev_priv->rps.hw_lock);
+		freq_sts = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
+		seq_printf(m, "PUNIT_REG_GPU_FREQ_STS: 0x%08x\n", freq_sts);
+		seq_printf(m, "DDR freq: %d MHz\n", dev_priv->mem_freq);
+
+		val = vlv_punit_read(dev_priv, PUNIT_FUSE_BUS1);
+		seq_printf(m, "max GPU freq: %d MHz\n",
+			   vlv_gpu_freq(dev_priv->mem_freq, val));
+
+		val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_LFM);
+		seq_printf(m, "min GPU freq: %d MHz\n",
+			   vlv_gpu_freq(dev_priv->mem_freq, val));
+
+		seq_printf(m, "current GPU freq: %d MHz\n",
+			   vlv_gpu_freq(dev_priv->mem_freq,
+					(freq_sts >> 8) & 0xff));
+		mutex_unlock(&dev_priv->rps.hw_lock);
 	} else {
 		seq_printf(m, "no P-state info available\n");
 	}
@@ -1290,6 +1526,25 @@ static int i915_fbc_status(struct seq_file *m, void *unused)
 	return 0;
 }
 
+static int i915_ips_status(struct seq_file *m, void *unused)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (!HAS_IPS(dev)) {
+		seq_puts(m, "not supported\n");
+		return 0;
+	}
+
+	if (I915_READ(IPS_CTL) & IPS_ENABLE)
+		seq_puts(m, "enabled\n");
+	else
+		seq_puts(m, "disabled\n");
+
+	return 0;
+}
+
 static int i915_sr_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
@@ -1642,27 +1897,27 @@ static int i915_dpio_info(struct seq_file *m, void *data)
 	seq_printf(m, "DPIO_CTL: 0x%08x\n", I915_READ(DPIO_CTL));
 
 	seq_printf(m, "DPIO_DIV_A: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_DIV_A));
+		   vlv_dpio_read(dev_priv, _DPIO_DIV_A));
 	seq_printf(m, "DPIO_DIV_B: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_DIV_B));
+		   vlv_dpio_read(dev_priv, _DPIO_DIV_B));
 
 	seq_printf(m, "DPIO_REFSFR_A: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_REFSFR_A));
+		   vlv_dpio_read(dev_priv, _DPIO_REFSFR_A));
 	seq_printf(m, "DPIO_REFSFR_B: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_REFSFR_B));
+		   vlv_dpio_read(dev_priv, _DPIO_REFSFR_B));
 
 	seq_printf(m, "DPIO_CORE_CLK_A: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_CORE_CLK_A));
+		   vlv_dpio_read(dev_priv, _DPIO_CORE_CLK_A));
 	seq_printf(m, "DPIO_CORE_CLK_B: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_CORE_CLK_B));
+		   vlv_dpio_read(dev_priv, _DPIO_CORE_CLK_B));
 
-	seq_printf(m, "DPIO_LFP_COEFF_A: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_LFP_COEFF_A));
-	seq_printf(m, "DPIO_LFP_COEFF_B: 0x%08x\n",
-		   intel_dpio_read(dev_priv, _DPIO_LFP_COEFF_B));
+	seq_printf(m, "DPIO_LPF_COEFF_A: 0x%08x\n",
+		   vlv_dpio_read(dev_priv, _DPIO_LPF_COEFF_A));
+	seq_printf(m, "DPIO_LPF_COEFF_B: 0x%08x\n",
+		   vlv_dpio_read(dev_priv, _DPIO_LPF_COEFF_B));
 
 	seq_printf(m, "DPIO_FASTCLK_DISABLE: 0x%08x\n",
-		   intel_dpio_read(dev_priv, DPIO_FASTCLK_DISABLE));
+		   vlv_dpio_read(dev_priv, DPIO_FASTCLK_DISABLE));
 
 	mutex_unlock(&dev_priv->dpio_lock);
 
@@ -1780,7 +2035,8 @@ i915_drop_caches_set(void *data, u64 val)
 	}
 
 	if (val & DROP_UNBOUND) {
-		list_for_each_entry_safe(obj, next, &dev_priv->mm.unbound_list, gtt_list)
+		list_for_each_entry_safe(obj, next, &dev_priv->mm.unbound_list,
+					 global_list)
 			if (obj->pages_pin_count == 0) {
 				ret = i915_gem_object_put_pages(obj);
 				if (ret)
@@ -1812,7 +2068,11 @@ i915_max_freq_get(void *data, u64 *val)
 	if (ret)
 		return ret;
 
-	*val = dev_priv->rps.max_delay * GT_FREQUENCY_MULTIPLIER;
+	if (IS_VALLEYVIEW(dev))
+		*val = vlv_gpu_freq(dev_priv->mem_freq,
+				    dev_priv->rps.max_delay);
+	else
+		*val = dev_priv->rps.max_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return 0;
@@ -1837,9 +2097,16 @@ i915_max_freq_set(void *data, u64 val)
 	/*
 	 * Turbo will still be enabled, but won't go above the set value.
 	 */
-	do_div(val, GT_FREQUENCY_MULTIPLIER);
-	dev_priv->rps.max_delay = val;
-	gen6_set_rps(dev, val);
+	if (IS_VALLEYVIEW(dev)) {
+		val = vlv_freq_opcode(dev_priv->mem_freq, val);
+		dev_priv->rps.max_delay = val;
+		gen6_set_rps(dev, val);
+	} else {
+		do_div(val, GT_FREQUENCY_MULTIPLIER);
+		dev_priv->rps.max_delay = val;
+		gen6_set_rps(dev, val);
+	}
+
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return 0;
@@ -1863,7 +2130,11 @@ i915_min_freq_get(void *data, u64 *val)
 	if (ret)
 		return ret;
 
-	*val = dev_priv->rps.min_delay * GT_FREQUENCY_MULTIPLIER;
+	if (IS_VALLEYVIEW(dev))
+		*val = vlv_gpu_freq(dev_priv->mem_freq,
+				    dev_priv->rps.min_delay);
+	else
+		*val = dev_priv->rps.min_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return 0;
@@ -1888,9 +2159,15 @@ i915_min_freq_set(void *data, u64 val)
 	/*
 	 * Turbo will still be enabled, but won't go below the set value.
 	 */
-	do_div(val, GT_FREQUENCY_MULTIPLIER);
-	dev_priv->rps.min_delay = val;
-	gen6_set_rps(dev, val);
+	if (IS_VALLEYVIEW(dev)) {
+		val = vlv_freq_opcode(dev_priv->mem_freq, val);
+		dev_priv->rps.min_delay = val;
+		valleyview_set_rps(dev, val);
+	} else {
+		do_div(val, GT_FREQUENCY_MULTIPLIER);
+		dev_priv->rps.min_delay = val;
+		gen6_set_rps(dev, val);
+	}
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return 0;
@@ -2057,6 +2334,7 @@ static struct drm_info_list i915_debugfs_list[] = {
 	{"i915_gem_hws", i915_hws_info, 0, (void *)RCS},
 	{"i915_gem_hws_blt", i915_hws_info, 0, (void *)BCS},
 	{"i915_gem_hws_bsd", i915_hws_info, 0, (void *)VCS},
+	{"i915_gem_hws_vebox", i915_hws_info, 0, (void *)VECS},
 	{"i915_rstdby_delays", i915_rstdby_delays, 0},
 	{"i915_cur_delayinfo", i915_cur_delayinfo, 0},
 	{"i915_delayfreq_table", i915_delayfreq_table, 0},
@@ -2066,6 +2344,7 @@ static struct drm_info_list i915_debugfs_list[] = {
 	{"i915_ring_freq_table", i915_ring_freq_table, 0},
 	{"i915_gfxec", i915_gfxec, 0},
 	{"i915_fbc_status", i915_fbc_status, 0},
+	{"i915_ips_status", i915_ips_status, 0},
 	{"i915_sr_status", i915_sr_status, 0},
 	{"i915_opregion", i915_opregion, 0},
 	{"i915_gem_framebuffer", i915_gem_framebuffer_info, 0},
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 3b315ba85a3e..67ec54f67afe 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -42,7 +42,6 @@
 #include <linux/vga_switcheroo.h>
 #include <linux/slab.h>
 #include <acpi/video.h>
-#include <asm/pat.h>
 
 #define LP_RING(d) (&((struct drm_i915_private *)(d))->ring[RCS])
 
@@ -956,6 +955,9 @@ static int i915_getparam(struct drm_device *dev, void *data,
 	case I915_PARAM_HAS_BLT:
 		value = intel_ring_initialized(&dev_priv->ring[BCS]);
 		break;
+	case I915_PARAM_HAS_VEBOX:
+		value = intel_ring_initialized(&dev_priv->ring[VECS]);
+		break;
 	case I915_PARAM_HAS_RELAXED_FENCING:
 		value = 1;
 		break;
@@ -999,8 +1001,7 @@ static int i915_getparam(struct drm_device *dev, void *data,
 		value = 1;
 		break;
 	default:
-		DRM_DEBUG_DRIVER("Unknown parameter %d\n",
-				 param->param);
+		DRM_DEBUG("Unknown parameter %d\n", param->param);
 		return -EINVAL;
 	}
 
@@ -1359,8 +1360,10 @@ static int i915_load_modeset_init(struct drm_device *dev)
 cleanup_gem:
 	mutex_lock(&dev->struct_mutex);
 	i915_gem_cleanup_ringbuffer(dev);
+	i915_gem_context_fini(dev);
 	mutex_unlock(&dev->struct_mutex);
 	i915_gem_cleanup_aliasing_ppgtt(dev);
+	drm_mm_takedown(&dev_priv->mm.gtt_space);
 cleanup_irq:
 	drm_irq_uninstall(dev);
 cleanup_gem_stolen:
@@ -1397,29 +1400,6 @@ void i915_master_destroy(struct drm_device *dev, struct drm_master *master)
 	master->driver_priv = NULL;
 }
 
-static void
-i915_mtrr_setup(struct drm_i915_private *dev_priv, unsigned long base,
-		unsigned long size)
-{
-	dev_priv->mm.gtt_mtrr = -1;
-
-#if defined(CONFIG_X86_PAT)
-	if (cpu_has_pat)
-		return;
-#endif
-
-	/* Set up a WC MTRR for non-PAT systems.  This is more common than
-	 * one would think, because the kernel disables PAT on first
-	 * generation Core chips because WC PAT gets overridden by a UC
-	 * MTRR if present.  Even if a UC MTRR isn't present.
-	 */
-	dev_priv->mm.gtt_mtrr = mtrr_add(base, size, MTRR_TYPE_WRCOMB, 1);
-	if (dev_priv->mm.gtt_mtrr < 0) {
-		DRM_INFO("MTRR allocation failed.  Graphics "
-			 "performance may suffer.\n");
-	}
-}
-
 static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
 {
 	struct apertures_struct *ap;
@@ -1431,7 +1411,7 @@ static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
 		return;
 
 	ap->ranges[0].base = dev_priv->gtt.mappable_base;
-	ap->ranges[0].size = dev_priv->gtt.mappable_end - dev_priv->gtt.start;
+	ap->ranges[0].size = dev_priv->gtt.mappable_end;
 
 	primary =
 		pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW;
@@ -1445,15 +1425,19 @@ static void i915_dump_device_info(struct drm_i915_private *dev_priv)
 {
 	const struct intel_device_info *info = dev_priv->info;
 
-#define DEV_INFO_FLAG(name) info->name ? #name "," : ""
-#define DEV_INFO_SEP ,
+#define PRINT_S(name) "%s"
+#define SEP_EMPTY
+#define PRINT_FLAG(name) info->name ? #name "," : ""
+#define SEP_COMMA ,
 	DRM_DEBUG_DRIVER("i915 device info: gen=%i, pciid=0x%04x flags="
-			 "%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s",
+			 DEV_INFO_FOR_EACH_FLAG(PRINT_S, SEP_EMPTY),
 			 info->gen,
 			 dev_priv->dev->pdev->device,
-			 DEV_INFO_FLAGS);
-#undef DEV_INFO_FLAG
-#undef DEV_INFO_SEP
+			 DEV_INFO_FOR_EACH_FLAG(PRINT_FLAG, SEP_COMMA));
+#undef PRINT_S
+#undef SEP_EMPTY
+#undef PRINT_FLAG
+#undef SEP_COMMA
 }
 
 /**
@@ -1468,7 +1452,7 @@ static void intel_early_sanitize_regs(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (IS_HASWELL(dev))
+	if (HAS_FPGA_DBG_UNCLAIMED(dev))
 		I915_WRITE_NOTRACE(FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
 }
 
@@ -1511,6 +1495,14 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	dev_priv->dev = dev;
 	dev_priv->info = info;
 
+	spin_lock_init(&dev_priv->irq_lock);
+	spin_lock_init(&dev_priv->gpu_error.lock);
+	spin_lock_init(&dev_priv->rps.lock);
+	spin_lock_init(&dev_priv->backlight.lock);
+	mutex_init(&dev_priv->dpio_lock);
+	mutex_init(&dev_priv->rps.hw_lock);
+	mutex_init(&dev_priv->modeset_restore_lock);
+
 	i915_dump_device_info(dev_priv);
 
 	if (i915_get_bridge_dev(dev)) {
@@ -1574,8 +1566,8 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 		goto out_rmmap;
 	}
 
-	i915_mtrr_setup(dev_priv, dev_priv->gtt.mappable_base,
-			aperture_size);
+	dev_priv->mm.gtt_mtrr = arch_phys_wc_add(dev_priv->gtt.mappable_base,
+						 aperture_size);
 
 	/* The i915 workqueue is primarily used for batched retirement of
 	 * requests (and thus managing bo) once the task has been completed
@@ -1601,6 +1593,7 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	intel_detect_pch(dev);
 
 	intel_irq_init(dev);
+	intel_gt_sanitize(dev);
 	intel_gt_init(dev);
 
 	/* Try to make sure MCHBAR is enabled before poking at it */
@@ -1626,14 +1619,6 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	if (!IS_I945G(dev) && !IS_I945GM(dev))
 		pci_enable_msi(dev->pdev);
 
-	spin_lock_init(&dev_priv->irq_lock);
-	spin_lock_init(&dev_priv->gpu_error.lock);
-	spin_lock_init(&dev_priv->rps.lock);
-	mutex_init(&dev_priv->dpio_lock);
-
-	mutex_init(&dev_priv->rps.hw_lock);
-	mutex_init(&dev_priv->modeset_restore_lock);
-
 	dev_priv->num_plane = 1;
 	if (IS_VALLEYVIEW(dev))
 		dev_priv->num_plane = 2;
@@ -1647,6 +1632,9 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	/* Start out suspended */
 	dev_priv->mm.suspended = 1;
 
+	if (HAS_POWER_WELL(dev))
+		i915_init_power_well(dev);
+
 	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
 		ret = i915_load_modeset_init(dev);
 		if (ret < 0) {
@@ -1660,7 +1648,7 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	if (INTEL_INFO(dev)->num_pipes) {
 		/* Must be done after probing outputs */
 		intel_opregion_init(dev);
-		acpi_video_register();
+		acpi_video_register_with_quirks();
 	}
 
 	if (IS_GEN5(dev))
@@ -1679,12 +1667,7 @@ out_gem_unload:
 	intel_teardown_mchbar(dev);
 	destroy_workqueue(dev_priv->wq);
 out_mtrrfree:
-	if (dev_priv->mm.gtt_mtrr >= 0) {
-		mtrr_del(dev_priv->mm.gtt_mtrr,
-			 dev_priv->gtt.mappable_base,
-			 aperture_size);
-		dev_priv->mm.gtt_mtrr = -1;
-	}
+	arch_phys_wc_del(dev_priv->mm.gtt_mtrr);
 	io_mapping_free(dev_priv->gtt.mappable);
 	dev_priv->gtt.gtt_remove(dev);
 out_rmmap:
@@ -1703,6 +1686,9 @@ int i915_driver_unload(struct drm_device *dev)
 
 	intel_gpu_ips_teardown();
 
+	if (HAS_POWER_WELL(dev))
+		i915_remove_power_well(dev);
+
 	i915_teardown_sysfs(dev);
 
 	if (dev_priv->mm.inactive_shrinker.shrink)
@@ -1719,12 +1705,7 @@ int i915_driver_unload(struct drm_device *dev)
 	cancel_delayed_work_sync(&dev_priv->mm.retire_work);
 
 	io_mapping_free(dev_priv->gtt.mappable);
-	if (dev_priv->mm.gtt_mtrr >= 0) {
-		mtrr_del(dev_priv->mm.gtt_mtrr,
-			 dev_priv->gtt.mappable_base,
-			 dev_priv->gtt.mappable_end);
-		dev_priv->mm.gtt_mtrr = -1;
-	}
+	arch_phys_wc_del(dev_priv->mm.gtt_mtrr);
 
 	acpi_video_unregister();
 
@@ -1737,10 +1718,10 @@ int i915_driver_unload(struct drm_device *dev)
 		 * free the memory space allocated for the child device
 		 * config parsed from VBT
 		 */
-		if (dev_priv->child_dev && dev_priv->child_dev_num) {
-			kfree(dev_priv->child_dev);
-			dev_priv->child_dev = NULL;
-			dev_priv->child_dev_num = 0;
+		if (dev_priv->vbt.child_dev && dev_priv->vbt.child_dev_num) {
+			kfree(dev_priv->vbt.child_dev);
+			dev_priv->vbt.child_dev = NULL;
+			dev_priv->vbt.child_dev_num = 0;
 		}
 
 		vga_switcheroo_unregister_client(dev->pdev);
@@ -1773,6 +1754,7 @@ int i915_driver_unload(struct drm_device *dev)
 			i915_free_hws(dev);
 	}
 
+	drm_mm_takedown(&dev_priv->mm.gtt_space);
 	if (dev_priv->regs != NULL)
 		pci_iounmap(dev->pdev, dev_priv->regs);
 
@@ -1782,6 +1764,8 @@ int i915_driver_unload(struct drm_device *dev)
 	destroy_workqueue(dev_priv->wq);
 	pm_qos_remove_request(&dev_priv->pm_qos);
 
+	dev_priv->gtt.gtt_remove(dev);
+
 	if (dev_priv->slab)
 		kmem_cache_destroy(dev_priv->slab);
 
@@ -1796,7 +1780,7 @@ int i915_driver_open(struct drm_device *dev, struct drm_file *file)
 	struct drm_i915_file_private *file_priv;
 
 	DRM_DEBUG_DRIVER("\n");
-	file_priv = kmalloc(sizeof(*file_priv), GFP_KERNEL);
+	file_priv = kzalloc(sizeof(*file_priv), GFP_KERNEL);
 	if (!file_priv)
 		return -ENOMEM;
 
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index a2e4953b8e8d..45b3c030f483 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -123,10 +123,14 @@ module_param_named(preliminary_hw_support, i915_preliminary_hw_support, int, 060
 MODULE_PARM_DESC(preliminary_hw_support,
 		"Enable preliminary hardware support. (default: false)");
 
-int i915_disable_power_well __read_mostly = 0;
+int i915_disable_power_well __read_mostly = 1;
 module_param_named(disable_power_well, i915_disable_power_well, int, 0600);
 MODULE_PARM_DESC(disable_power_well,
-		 "Disable the power well when possible (default: false)");
+		 "Disable the power well when possible (default: true)");
+
+int i915_enable_ips __read_mostly = 1;
+module_param_named(enable_ips, i915_enable_ips, int, 0600);
+MODULE_PARM_DESC(enable_ips, "Enable IPS (default: true)");
 
 static struct drm_driver driver;
 extern int intel_agp_enabled;
@@ -280,6 +284,7 @@ static const struct intel_device_info intel_ivybridge_m_info = {
 	GEN7_FEATURES,
 	.is_ivybridge = 1,
 	.is_mobile = 1,
+	.has_fbc = 1,
 };
 
 static const struct intel_device_info intel_ivybridge_q_info = {
@@ -308,12 +313,19 @@ static const struct intel_device_info intel_valleyview_d_info = {
 static const struct intel_device_info intel_haswell_d_info = {
 	GEN7_FEATURES,
 	.is_haswell = 1,
+	.has_ddi = 1,
+	.has_fpga_dbg = 1,
+	.has_vebox_ring = 1,
 };
 
 static const struct intel_device_info intel_haswell_m_info = {
 	GEN7_FEATURES,
 	.is_haswell = 1,
 	.is_mobile = 1,
+	.has_ddi = 1,
+	.has_fpga_dbg = 1,
+	.has_fbc = 1,
+	.has_vebox_ring = 1,
 };
 
 static const struct pci_device_id pciidlist[] = {		/* aka */
@@ -445,7 +457,6 @@ void intel_detect_pch(struct drm_device *dev)
 	 */
 	if (INTEL_INFO(dev)->num_pipes == 0) {
 		dev_priv->pch_type = PCH_NOP;
-		dev_priv->num_pch_pll = 0;
 		return;
 	}
 
@@ -454,9 +465,15 @@ void intel_detect_pch(struct drm_device *dev)
 	 * make graphics device passthrough work easy for VMM, that only
 	 * need to expose ISA bridge to let driver know the real hardware
 	 * underneath. This is a requirement from virtualization team.
+	 *
+	 * In some virtualized environments (e.g. XEN), there is irrelevant
+	 * ISA bridge in the system. To work reliably, we should scan trhough
+	 * all the ISA bridge devices and check for the first match, instead
+	 * of only checking the first one.
 	 */
 	pch = pci_get_class(PCI_CLASS_BRIDGE_ISA << 8, NULL);
-	if (pch) {
+	while (pch) {
+		struct pci_dev *curr = pch;
 		if (pch->vendor == PCI_VENDOR_ID_INTEL) {
 			unsigned short id;
 			id = pch->device & INTEL_PCH_DEVICE_ID_MASK;
@@ -464,37 +481,39 @@ void intel_detect_pch(struct drm_device *dev)
 
 			if (id == INTEL_PCH_IBX_DEVICE_ID_TYPE) {
 				dev_priv->pch_type = PCH_IBX;
-				dev_priv->num_pch_pll = 2;
 				DRM_DEBUG_KMS("Found Ibex Peak PCH\n");
 				WARN_ON(!IS_GEN5(dev));
 			} else if (id == INTEL_PCH_CPT_DEVICE_ID_TYPE) {
 				dev_priv->pch_type = PCH_CPT;
-				dev_priv->num_pch_pll = 2;
 				DRM_DEBUG_KMS("Found CougarPoint PCH\n");
 				WARN_ON(!(IS_GEN6(dev) || IS_IVYBRIDGE(dev)));
 			} else if (id == INTEL_PCH_PPT_DEVICE_ID_TYPE) {
 				/* PantherPoint is CPT compatible */
 				dev_priv->pch_type = PCH_CPT;
-				dev_priv->num_pch_pll = 2;
 				DRM_DEBUG_KMS("Found PatherPoint PCH\n");
 				WARN_ON(!(IS_GEN6(dev) || IS_IVYBRIDGE(dev)));
 			} else if (id == INTEL_PCH_LPT_DEVICE_ID_TYPE) {
 				dev_priv->pch_type = PCH_LPT;
-				dev_priv->num_pch_pll = 0;
 				DRM_DEBUG_KMS("Found LynxPoint PCH\n");
 				WARN_ON(!IS_HASWELL(dev));
 				WARN_ON(IS_ULT(dev));
 			} else if (id == INTEL_PCH_LPT_LP_DEVICE_ID_TYPE) {
 				dev_priv->pch_type = PCH_LPT;
-				dev_priv->num_pch_pll = 0;
 				DRM_DEBUG_KMS("Found LynxPoint LP PCH\n");
 				WARN_ON(!IS_HASWELL(dev));
 				WARN_ON(!IS_ULT(dev));
+			} else {
+				goto check_next;
 			}
-			BUG_ON(dev_priv->num_pch_pll > I915_NUM_PLLS);
+			pci_dev_put(pch);
+			break;
 		}
-		pci_dev_put(pch);
+check_next:
+		pch = pci_get_class(PCI_CLASS_BRIDGE_ISA << 8, curr);
+		pci_dev_put(curr);
 	}
+	if (!pch)
+		DRM_DEBUG_KMS("No PCH found?\n");
 }
 
 bool i915_semaphore_is_enabled(struct drm_device *dev)
@@ -549,6 +568,8 @@ static int i915_drm_freeze(struct drm_device *dev)
 		 */
 		list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
 			dev_priv->display.crtc_disable(crtc);
+
+		intel_modeset_suspend_hw(dev);
 	}
 
 	i915_save_state(dev);
@@ -556,7 +577,7 @@ static int i915_drm_freeze(struct drm_device *dev)
 	intel_opregion_fini(dev);
 
 	console_lock();
-	intel_fbdev_set_suspend(dev, 1);
+	intel_fbdev_set_suspend(dev, FBINFO_STATE_SUSPENDED);
 	console_unlock();
 
 	return 0;
@@ -600,7 +621,7 @@ void intel_console_resume(struct work_struct *work)
 	struct drm_device *dev = dev_priv->dev;
 
 	console_lock();
-	intel_fbdev_set_suspend(dev, 0);
+	intel_fbdev_set_suspend(dev, FBINFO_STATE_RUNNING);
 	console_unlock();
 }
 
@@ -669,7 +690,7 @@ static int __i915_drm_thaw(struct drm_device *dev)
 	 * path of resume if possible.
 	 */
 	if (console_trylock()) {
-		intel_fbdev_set_suspend(dev, 0);
+		intel_fbdev_set_suspend(dev, FBINFO_STATE_RUNNING);
 		console_unlock();
 	} else {
 		schedule_work(&dev_priv->console_resume_work);
@@ -685,7 +706,7 @@ static int i915_drm_thaw(struct drm_device *dev)
 {
 	int error = 0;
 
-	intel_gt_reset(dev);
+	intel_gt_sanitize(dev);
 
 	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
 		mutex_lock(&dev->struct_mutex);
@@ -711,7 +732,7 @@ int i915_resume(struct drm_device *dev)
 
 	pci_set_master(dev->pdev);
 
-	intel_gt_reset(dev);
+	intel_gt_sanitize(dev);
 
 	/*
 	 * Platforms with opregion should have sane BIOS, older ones (gen3 and
@@ -855,37 +876,14 @@ static int gen6_do_reset(struct drm_device *dev)
 
 int intel_gpu_reset(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret = -ENODEV;
-
 	switch (INTEL_INFO(dev)->gen) {
 	case 7:
-	case 6:
-		ret = gen6_do_reset(dev);
-		break;
-	case 5:
-		ret = ironlake_do_reset(dev);
-		break;
-	case 4:
-		ret = i965_do_reset(dev);
-		break;
-	case 2:
-		ret = i8xx_do_reset(dev);
-		break;
-	}
-
-	/* Also reset the gpu hangman. */
-	if (dev_priv->gpu_error.stop_rings) {
-		DRM_INFO("Simulated gpu hang, resetting stop_rings\n");
-		dev_priv->gpu_error.stop_rings = 0;
-		if (ret == -ENODEV) {
-			DRM_ERROR("Reset not implemented, but ignoring "
-				  "error for simulated gpu hangs\n");
-			ret = 0;
-		}
+	case 6: return gen6_do_reset(dev);
+	case 5: return ironlake_do_reset(dev);
+	case 4: return i965_do_reset(dev);
+	case 2: return i8xx_do_reset(dev);
+	default: return -ENODEV;
 	}
-
-	return ret;
 }
 
 /**
@@ -906,6 +904,7 @@ int intel_gpu_reset(struct drm_device *dev)
 int i915_reset(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
+	bool simulated;
 	int ret;
 
 	if (!i915_try_reset)
@@ -915,13 +914,26 @@ int i915_reset(struct drm_device *dev)
 
 	i915_gem_reset(dev);
 
-	ret = -ENODEV;
-	if (get_seconds() - dev_priv->gpu_error.last_reset < 5)
+	simulated = dev_priv->gpu_error.stop_rings != 0;
+
+	if (!simulated && get_seconds() - dev_priv->gpu_error.last_reset < 5) {
 		DRM_ERROR("GPU hanging too fast, declaring wedged!\n");
-	else
+		ret = -ENODEV;
+	} else {
 		ret = intel_gpu_reset(dev);
 
-	dev_priv->gpu_error.last_reset = get_seconds();
+		/* Also reset the gpu hangman. */
+		if (simulated) {
+			DRM_INFO("Simulated gpu hang, resetting stop_rings\n");
+			dev_priv->gpu_error.stop_rings = 0;
+			if (ret == -ENODEV) {
+				DRM_ERROR("Reset not implemented, but ignoring "
+					  "error for simulated gpu hangs\n");
+				ret = 0;
+			}
+		} else
+			dev_priv->gpu_error.last_reset = get_seconds();
+	}
 	if (ret) {
 		DRM_ERROR("Failed to reset chip.\n");
 		mutex_unlock(&dev->struct_mutex);
@@ -984,12 +996,6 @@ static int i915_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	struct intel_device_info *intel_info =
 		(struct intel_device_info *) ent->driver_data;
 
-	if (intel_info->is_valleyview)
-		if(!i915_preliminary_hw_support) {
-			DRM_ERROR("Preliminary hardware support disabled\n");
-			return -ENODEV;
-		}
-
 	/* Only bind to function 0 of the device. Early generations
 	 * used function 1 as a placeholder for multi-head. This causes
 	 * us confusion instead, especially on the systems where both
@@ -1218,16 +1224,16 @@ MODULE_LICENSE("GPL and additional rights");
 static void
 ilk_dummy_write(struct drm_i915_private *dev_priv)
 {
-	/* WaIssueDummyWriteToWakeupFromRC6: Issue a dummy write to wake up the
-	 * chip from rc6 before touching it for real. MI_MODE is masked, hence
-	 * harmless to write 0 into. */
+	/* WaIssueDummyWriteToWakeupFromRC6:ilk Issue a dummy write to wake up
+	 * the chip from rc6 before touching it for real. MI_MODE is masked,
+	 * hence harmless to write 0 into. */
 	I915_WRITE_NOTRACE(MI_MODE, 0);
 }
 
 static void
 hsw_unclaimed_reg_clear(struct drm_i915_private *dev_priv, u32 reg)
 {
-	if (IS_HASWELL(dev_priv->dev) &&
+	if (HAS_FPGA_DBG_UNCLAIMED(dev_priv->dev) &&
 	    (I915_READ_NOTRACE(FPGA_DBG) & FPGA_DBG_RM_NOCLAIM)) {
 		DRM_ERROR("Unknown unclaimed register before writing to %x\n",
 			  reg);
@@ -1238,7 +1244,7 @@ hsw_unclaimed_reg_clear(struct drm_i915_private *dev_priv, u32 reg)
 static void
 hsw_unclaimed_reg_check(struct drm_i915_private *dev_priv, u32 reg)
 {
-	if (IS_HASWELL(dev_priv->dev) &&
+	if (HAS_FPGA_DBG_UNCLAIMED(dev_priv->dev) &&
 	    (I915_READ_NOTRACE(FPGA_DBG) & FPGA_DBG_RM_NOCLAIM)) {
 		DRM_ERROR("Unclaimed write to %x\n", reg);
 		I915_WRITE_NOTRACE(FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
@@ -1247,21 +1253,21 @@ hsw_unclaimed_reg_check(struct drm_i915_private *dev_priv, u32 reg)
 
 #define __i915_read(x, y) \
 u##x i915_read##x(struct drm_i915_private *dev_priv, u32 reg) { \
+	unsigned long irqflags; \
 	u##x val = 0; \
+	spin_lock_irqsave(&dev_priv->gt_lock, irqflags); \
 	if (IS_GEN5(dev_priv->dev)) \
 		ilk_dummy_write(dev_priv); \
 	if (NEEDS_FORCE_WAKE((dev_priv), (reg))) { \
-		unsigned long irqflags; \
-		spin_lock_irqsave(&dev_priv->gt_lock, irqflags); \
 		if (dev_priv->forcewake_count == 0) \
 			dev_priv->gt.force_wake_get(dev_priv); \
 		val = read##y(dev_priv->regs + reg); \
 		if (dev_priv->forcewake_count == 0) \
 			dev_priv->gt.force_wake_put(dev_priv); \
-		spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags); \
 	} else { \
 		val = read##y(dev_priv->regs + reg); \
 	} \
+	spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags); \
 	trace_i915_reg_rw(false, reg, val, sizeof(val)); \
 	return val; \
 }
@@ -1274,8 +1280,10 @@ __i915_read(64, q)
 
 #define __i915_write(x, y) \
 void i915_write##x(struct drm_i915_private *dev_priv, u32 reg, u##x val) { \
+	unsigned long irqflags; \
 	u32 __fifo_ret = 0; \
 	trace_i915_reg_rw(true, reg, val, sizeof(val)); \
+	spin_lock_irqsave(&dev_priv->gt_lock, irqflags); \
 	if (NEEDS_FORCE_WAKE((dev_priv), (reg))) { \
 		__fifo_ret = __gen6_gt_wait_for_fifo(dev_priv); \
 	} \
@@ -1287,6 +1295,7 @@ void i915_write##x(struct drm_i915_private *dev_priv, u32 reg, u##x val) { \
 		gen6_gt_check_fifodbg(dev_priv); \
 	} \
 	hsw_unclaimed_reg_check(dev_priv, reg); \
+	spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags); \
 }
 __i915_write(8, b)
 __i915_write(16, w)
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 9669a0b8b440..d2ee3343c943 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -76,6 +76,8 @@ enum plane {
 };
 #define plane_name(p) ((p) + 'A')
 
+#define sprite_name(p, s) ((p) * dev_priv->num_plane + (s) + 'A')
+
 enum port {
 	PORT_A = 0,
 	PORT_B,
@@ -86,6 +88,24 @@ enum port {
 };
 #define port_name(p) ((p) + 'A')
 
+enum intel_display_power_domain {
+	POWER_DOMAIN_PIPE_A,
+	POWER_DOMAIN_PIPE_B,
+	POWER_DOMAIN_PIPE_C,
+	POWER_DOMAIN_PIPE_A_PANEL_FITTER,
+	POWER_DOMAIN_PIPE_B_PANEL_FITTER,
+	POWER_DOMAIN_PIPE_C_PANEL_FITTER,
+	POWER_DOMAIN_TRANSCODER_A,
+	POWER_DOMAIN_TRANSCODER_B,
+	POWER_DOMAIN_TRANSCODER_C,
+	POWER_DOMAIN_TRANSCODER_EDP = POWER_DOMAIN_TRANSCODER_A + 0xF,
+};
+
+#define POWER_DOMAIN_PIPE(pipe) ((pipe) + POWER_DOMAIN_PIPE_A)
+#define POWER_DOMAIN_PIPE_PANEL_FITTER(pipe) \
+		((pipe) + POWER_DOMAIN_PIPE_A_PANEL_FITTER)
+#define POWER_DOMAIN_TRANSCODER(tran) ((tran) + POWER_DOMAIN_TRANSCODER_A)
+
 enum hpd_pin {
 	HPD_NONE = 0,
 	HPD_PORT_A = HPD_NONE, /* PORT_A is internal */
@@ -112,15 +132,38 @@ enum hpd_pin {
 	list_for_each_entry((intel_encoder), &(dev)->mode_config.encoder_list, base.head) \
 		if ((intel_encoder)->base.crtc == (__crtc))
 
-struct intel_pch_pll {
+struct drm_i915_private;
+
+enum intel_dpll_id {
+	DPLL_ID_PRIVATE = -1, /* non-shared dpll in use */
+	/* real shared dpll ids must be >= 0 */
+	DPLL_ID_PCH_PLL_A,
+	DPLL_ID_PCH_PLL_B,
+};
+#define I915_NUM_PLLS 2
+
+struct intel_dpll_hw_state {
+	uint32_t dpll;
+	uint32_t fp0;
+	uint32_t fp1;
+};
+
+struct intel_shared_dpll {
 	int refcount; /* count of number of CRTCs sharing this PLL */
 	int active; /* count of number of active CRTCs (i.e. DPMS on) */
 	bool on; /* is the PLL actually active? Disabled during modeset */
-	int pll_reg;
-	int fp0_reg;
-	int fp1_reg;
+	const char *name;
+	/* should match the index in the dev_priv->shared_dplls array */
+	enum intel_dpll_id id;
+	struct intel_dpll_hw_state hw_state;
+	void (*enable)(struct drm_i915_private *dev_priv,
+		       struct intel_shared_dpll *pll);
+	void (*disable)(struct drm_i915_private *dev_priv,
+			struct intel_shared_dpll *pll);
+	bool (*get_hw_state)(struct drm_i915_private *dev_priv,
+			     struct intel_shared_dpll *pll,
+			     struct intel_dpll_hw_state *hw_state);
 };
-#define I915_NUM_PLLS 2
 
 /* Used by dp and fdi links */
 struct intel_link_m_n {
@@ -175,7 +218,6 @@ struct opregion_header;
 struct opregion_acpi;
 struct opregion_swsci;
 struct opregion_asle;
-struct drm_i915_private;
 
 struct intel_opregion {
 	struct opregion_header __iomem *header;
@@ -286,6 +328,8 @@ struct drm_i915_error_state {
 
 struct intel_crtc_config;
 struct intel_crtc;
+struct intel_limit;
+struct dpll;
 
 struct drm_i915_display_funcs {
 	bool (*fbc_enabled)(struct drm_device *dev);
@@ -293,11 +337,28 @@ struct drm_i915_display_funcs {
 	void (*disable_fbc)(struct drm_device *dev);
 	int (*get_display_clock_speed)(struct drm_device *dev);
 	int (*get_fifo_size)(struct drm_device *dev, int plane);
+	/**
+	 * find_dpll() - Find the best values for the PLL
+	 * @limit: limits for the PLL
+	 * @crtc: current CRTC
+	 * @target: target frequency in kHz
+	 * @refclk: reference clock frequency in kHz
+	 * @match_clock: if provided, @best_clock P divider must
+	 *               match the P divider from @match_clock
+	 *               used for LVDS downclocking
+	 * @best_clock: best PLL values found
+	 *
+	 * Returns true on success, false on failure.
+	 */
+	bool (*find_dpll)(const struct intel_limit *limit,
+			  struct drm_crtc *crtc,
+			  int target, int refclk,
+			  struct dpll *match_clock,
+			  struct dpll *best_clock);
 	void (*update_wm)(struct drm_device *dev);
 	void (*update_sprite_wm)(struct drm_device *dev, int pipe,
-				 uint32_t sprite_width, int pixel_size);
-	void (*update_linetime_wm)(struct drm_device *dev, int pipe,
-				 struct drm_display_mode *mode);
+				 uint32_t sprite_width, int pixel_size,
+				 bool enable);
 	void (*modeset_global_resources)(struct drm_device *dev);
 	/* Returns the active state of the crtc, and if the crtc is active,
 	 * fills out the pipe-config with the hw state. */
@@ -331,68 +392,56 @@ struct drm_i915_gt_funcs {
 	void (*force_wake_put)(struct drm_i915_private *dev_priv);
 };
 
-#define DEV_INFO_FLAGS \
-	DEV_INFO_FLAG(is_mobile) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_i85x) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_i915g) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_i945gm) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_g33) DEV_INFO_SEP \
-	DEV_INFO_FLAG(need_gfx_hws) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_g4x) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_pineview) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_broadwater) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_crestline) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_ivybridge) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_valleyview) DEV_INFO_SEP \
-	DEV_INFO_FLAG(is_haswell) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_force_wake) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_fbc) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_pipe_cxsr) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_hotplug) DEV_INFO_SEP \
-	DEV_INFO_FLAG(cursor_needs_physical) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_overlay) DEV_INFO_SEP \
-	DEV_INFO_FLAG(overlay_needs_physical) DEV_INFO_SEP \
-	DEV_INFO_FLAG(supports_tv) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_bsd_ring) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_blt_ring) DEV_INFO_SEP \
-	DEV_INFO_FLAG(has_llc)
+#define DEV_INFO_FOR_EACH_FLAG(func, sep) \
+	func(is_mobile) sep \
+	func(is_i85x) sep \
+	func(is_i915g) sep \
+	func(is_i945gm) sep \
+	func(is_g33) sep \
+	func(need_gfx_hws) sep \
+	func(is_g4x) sep \
+	func(is_pineview) sep \
+	func(is_broadwater) sep \
+	func(is_crestline) sep \
+	func(is_ivybridge) sep \
+	func(is_valleyview) sep \
+	func(is_haswell) sep \
+	func(has_force_wake) sep \
+	func(has_fbc) sep \
+	func(has_pipe_cxsr) sep \
+	func(has_hotplug) sep \
+	func(cursor_needs_physical) sep \
+	func(has_overlay) sep \
+	func(overlay_needs_physical) sep \
+	func(supports_tv) sep \
+	func(has_bsd_ring) sep \
+	func(has_blt_ring) sep \
+	func(has_vebox_ring) sep \
+	func(has_llc) sep \
+	func(has_ddi) sep \
+	func(has_fpga_dbg)
+
+#define DEFINE_FLAG(name) u8 name:1
+#define SEP_SEMICOLON ;
 
 struct intel_device_info {
 	u32 display_mmio_offset;
 	u8 num_pipes:3;
 	u8 gen;
-	u8 is_mobile:1;
-	u8 is_i85x:1;
-	u8 is_i915g:1;
-	u8 is_i945gm:1;
-	u8 is_g33:1;
-	u8 need_gfx_hws:1;
-	u8 is_g4x:1;
-	u8 is_pineview:1;
-	u8 is_broadwater:1;
-	u8 is_crestline:1;
-	u8 is_ivybridge:1;
-	u8 is_valleyview:1;
-	u8 has_force_wake:1;
-	u8 is_haswell:1;
-	u8 has_fbc:1;
-	u8 has_pipe_cxsr:1;
-	u8 has_hotplug:1;
-	u8 cursor_needs_physical:1;
-	u8 has_overlay:1;
-	u8 overlay_needs_physical:1;
-	u8 supports_tv:1;
-	u8 has_bsd_ring:1;
-	u8 has_blt_ring:1;
-	u8 has_llc:1;
+	DEV_INFO_FOR_EACH_FLAG(DEFINE_FLAG, SEP_SEMICOLON);
 };
 
+#undef DEFINE_FLAG
+#undef SEP_SEMICOLON
+
 enum i915_cache_level {
 	I915_CACHE_NONE = 0,
 	I915_CACHE_LLC,
 	I915_CACHE_LLC_MLC, /* gen6+, in docs at least! */
 };
 
+typedef uint32_t gen6_gtt_pte_t;
+
 /* The Graphics Translation Table is the way in which GEN hardware translates a
  * Graphics Virtual Address into a Physical Address. In addition to the normal
  * collateral associated with any va->pa translations GEN hardware also has a
@@ -428,6 +477,9 @@ struct i915_gtt {
 				   struct sg_table *st,
 				   unsigned int pg_start,
 				   enum i915_cache_level cache_level);
+	gen6_gtt_pte_t (*pte_encode)(struct drm_device *dev,
+				     dma_addr_t addr,
+				     enum i915_cache_level level);
 };
 #define gtt_total_entries(gtt) ((gtt).total >> PAGE_SHIFT)
 
@@ -449,19 +501,31 @@ struct i915_hw_ppgtt {
 			       struct sg_table *st,
 			       unsigned int pg_start,
 			       enum i915_cache_level cache_level);
+	gen6_gtt_pte_t (*pte_encode)(struct drm_device *dev,
+				     dma_addr_t addr,
+				     enum i915_cache_level level);
 	int (*enable)(struct drm_device *dev);
 	void (*cleanup)(struct i915_hw_ppgtt *ppgtt);
 };
 
+struct i915_ctx_hang_stats {
+	/* This context had batch pending when hang was declared */
+	unsigned batch_pending;
+
+	/* This context had batch active when hang was declared */
+	unsigned batch_active;
+};
 
 /* This must match up with the value previously used for execbuf2.rsvd1. */
 #define DEFAULT_CONTEXT_ID 0
 struct i915_hw_context {
+	struct kref ref;
 	int id;
 	bool is_initialized;
 	struct drm_i915_file_private *file_priv;
 	struct intel_ring_buffer *ring;
 	struct drm_i915_gem_object *obj;
+	struct i915_ctx_hang_stats hang_stats;
 };
 
 enum no_fbc_reason {
@@ -491,6 +555,7 @@ enum intel_sbi_destination {
 #define QUIRK_PIPEA_FORCE (1<<0)
 #define QUIRK_LVDS_SSC_DISABLE (1<<1)
 #define QUIRK_INVERT_BRIGHTNESS (1<<2)
+#define QUIRK_NO_PCH_PWM_ENABLE (1<<3)
 
 struct intel_fbdev;
 struct intel_fbc_work;
@@ -658,6 +723,7 @@ struct i915_suspend_saved_registers {
 
 struct intel_gen6_power_mgmt {
 	struct work_struct work;
+	struct delayed_work vlv_work;
 	u32 pm_iir;
 	/* lock - irqsave spinlock that protectects the work_struct and
 	 * pm_iir. */
@@ -668,6 +734,7 @@ struct intel_gen6_power_mgmt {
 	u8 cur_delay;
 	u8 min_delay;
 	u8 max_delay;
+	u8 rpe_delay;
 	u8 hw_max;
 
 	struct delayed_work delayed_resume_work;
@@ -704,6 +771,15 @@ struct intel_ilk_power_mgmt {
 	struct drm_i915_gem_object *renderctx;
 };
 
+/* Power well structure for haswell */
+struct i915_power_well {
+	struct drm_device *device;
+	spinlock_t lock;
+	/* power well enable/disable usage count */
+	int count;
+	int i915_request;
+};
+
 struct i915_dri1_state {
 	unsigned allow_batchbuffer : 1;
 	u32 __iomem *gfx_hws_cpu_addr;
@@ -812,14 +888,20 @@ struct i915_gem_mm {
 	u32 object_count;
 };
 
+struct drm_i915_error_state_buf {
+	unsigned bytes;
+	unsigned size;
+	int err;
+	u8 *buf;
+	loff_t start;
+	loff_t pos;
+};
+
 struct i915_gpu_error {
 	/* For hangcheck timer */
 #define DRM_I915_HANGCHECK_PERIOD 1500 /* in ms */
 #define DRM_I915_HANGCHECK_JIFFIES msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD)
 	struct timer_list hangcheck_timer;
-	int hangcheck_count;
-	uint32_t last_acthd[I915_NUM_RINGS];
-	uint32_t prev_instdone[I915_NUM_INSTDONE_REG];
 
 	/* For reset and error_state handling. */
 	spinlock_t lock;
@@ -875,6 +957,37 @@ enum modeset_restore {
 	MODESET_SUSPENDED,
 };
 
+struct intel_vbt_data {
+	struct drm_display_mode *lfp_lvds_vbt_mode; /* if any */
+	struct drm_display_mode *sdvo_lvds_vbt_mode; /* if any */
+
+	/* Feature bits */
+	unsigned int int_tv_support:1;
+	unsigned int lvds_dither:1;
+	unsigned int lvds_vbt:1;
+	unsigned int int_crt_support:1;
+	unsigned int lvds_use_ssc:1;
+	unsigned int display_clock_mode:1;
+	unsigned int fdi_rx_polarity_inverted:1;
+	int lvds_ssc_freq;
+	unsigned int bios_lvds_val; /* initial [PCH_]LVDS reg val in VBIOS */
+
+	/* eDP */
+	int edp_rate;
+	int edp_lanes;
+	int edp_preemphasis;
+	int edp_vswing;
+	bool edp_initialized;
+	bool edp_support;
+	int edp_bpp;
+	struct edp_power_seq edp_pps;
+
+	int crt_ddc_pin;
+
+	int child_dev_num;
+	struct child_device_config *child_dev;
+};
+
 typedef struct drm_i915_private {
 	struct drm_device *dev;
 	struct kmem_cache *slab;
@@ -941,9 +1054,9 @@ typedef struct drm_i915_private {
 			HPD_MARK_DISABLED = 2
 		} hpd_mark;
 	} hpd_stats[HPD_NUM_PINS];
+	u32 hpd_event_bits;
 	struct timer_list hotplug_reenable_timer;
 
-	int num_pch_pll;
 	int num_plane;
 
 	unsigned long cfb_size;
@@ -953,6 +1066,7 @@ typedef struct drm_i915_private {
 	struct intel_fbc_work *fbc_work;
 
 	struct intel_opregion opregion;
+	struct intel_vbt_data vbt;
 
 	/* overlay */
 	struct intel_overlay *overlay;
@@ -962,37 +1076,15 @@ typedef struct drm_i915_private {
 	struct {
 		int level;
 		bool enabled;
+		spinlock_t lock; /* bl registers and the above bl fields */
 		struct backlight_device *device;
 	} backlight;
 
 	/* LVDS info */
 	struct drm_display_mode *lfp_lvds_vbt_mode; /* if any */
 	struct drm_display_mode *sdvo_lvds_vbt_mode; /* if any */
-
-	/* Feature bits from the VBIOS */
-	unsigned int int_tv_support:1;
-	unsigned int lvds_dither:1;
-	unsigned int lvds_vbt:1;
-	unsigned int int_crt_support:1;
-	unsigned int lvds_use_ssc:1;
-	unsigned int display_clock_mode:1;
-	unsigned int fdi_rx_polarity_inverted:1;
-	int lvds_ssc_freq;
-	unsigned int bios_lvds_val; /* initial [PCH_]LVDS reg val in VBIOS */
-	struct {
-		int rate;
-		int lanes;
-		int preemphasis;
-		int vswing;
-
-		bool initialized;
-		bool support;
-		int bpp;
-		struct edp_power_seq pps;
-	} edp;
 	bool no_aux_handshake;
 
-	int crt_ddc_pin;
 	struct drm_i915_fence_reg fence_regs[I915_MAX_NUM_FENCES]; /* assume 965 */
 	int fence_reg_start; /* 4 if userland hasn't ioctl'd us yet */
 	int num_fence_regs; /* 8 on pre-965, 16 otherwise */
@@ -1020,16 +1112,13 @@ typedef struct drm_i915_private {
 	/* Kernel Modesetting */
 
 	struct sdvo_device_mapping sdvo_mappings[2];
-	/* indicate whether the LVDS_BORDER should be enabled or not */
-	unsigned int lvds_border_bits;
-	/* Panel fitter placement and size for Ironlake+ */
-	u32 pch_pf_pos, pch_pf_size;
 
 	struct drm_crtc *plane_to_crtc_mapping[3];
 	struct drm_crtc *pipe_to_crtc_mapping[3];
 	wait_queue_head_t pending_flip_queue;
 
-	struct intel_pch_pll pch_plls[I915_NUM_PLLS];
+	int num_shared_dpll;
+	struct intel_shared_dpll shared_dplls[I915_NUM_PLLS];
 	struct intel_ddi_plls ddi_plls;
 
 	/* Reclocking support */
@@ -1038,8 +1127,6 @@ typedef struct drm_i915_private {
 	/* indicates the reduced downclock for LVDS*/
 	int lvds_downclock;
 	u16 orig_clock;
-	int child_dev_num;
-	struct child_device_config *child_dev;
 
 	bool mchbar_need_disable;
 
@@ -1052,6 +1139,9 @@ typedef struct drm_i915_private {
 	 * mchdev_lock in intel_pm.c */
 	struct intel_ilk_power_mgmt ips;
 
+	/* Haswell power well */
+	struct i915_power_well power_well;
+
 	enum no_fbc_reason no_fbc_reason;
 
 	struct drm_mm_node *compressed_fb;
@@ -1059,6 +1149,8 @@ typedef struct drm_i915_private {
 
 	struct i915_gpu_error gpu_error;
 
+	struct drm_i915_gem_object *vlv_pctx;
+
 	/* list of fbdev register on this device */
 	struct intel_fbdev *fbdev;
 
@@ -1124,7 +1216,7 @@ struct drm_i915_gem_object {
 	struct drm_mm_node *gtt_space;
 	/** Stolen memory for this object, instead of being backed by shmem. */
 	struct drm_mm_node *stolen;
-	struct list_head gtt_list;
+	struct list_head global_list;
 
 	/** This object's place on the active/inactive lists */
 	struct list_head ring_list;
@@ -1271,9 +1363,18 @@ struct drm_i915_gem_request {
 	/** GEM sequence number associated with this request. */
 	uint32_t seqno;
 
-	/** Postion in the ringbuffer of the end of the request */
+	/** Position in the ringbuffer of the start of the request */
+	u32 head;
+
+	/** Position in the ringbuffer of the end of the request */
 	u32 tail;
 
+	/** Context related to this request */
+	struct i915_hw_context *ctx;
+
+	/** Batch buffer related to this request if any */
+	struct drm_i915_gem_object *batch_obj;
+
 	/** Time at which this request was emitted, in jiffies. */
 	unsigned long emitted_jiffies;
 
@@ -1291,6 +1392,8 @@ struct drm_i915_file_private {
 		struct list_head request_list;
 	} mm;
 	struct idr context_idr;
+
+	struct i915_ctx_hang_stats hang_stats;
 };
 
 #define INTEL_INFO(dev)	(((struct drm_i915_private *) (dev)->dev_private)->info)
@@ -1341,6 +1444,7 @@ struct drm_i915_file_private {
 
 #define HAS_BSD(dev)            (INTEL_INFO(dev)->has_bsd_ring)
 #define HAS_BLT(dev)            (INTEL_INFO(dev)->has_blt_ring)
+#define HAS_VEBOX(dev)          (INTEL_INFO(dev)->has_vebox_ring)
 #define HAS_LLC(dev)            (INTEL_INFO(dev)->has_llc)
 #define I915_NEED_GFX_HWS(dev)	(INTEL_INFO(dev)->need_gfx_hws)
 
@@ -1371,10 +1475,13 @@ struct drm_i915_file_private {
 #define HAS_PIPE_CXSR(dev) (INTEL_INFO(dev)->has_pipe_cxsr)
 #define I915_HAS_FBC(dev) (INTEL_INFO(dev)->has_fbc)
 
+#define HAS_IPS(dev)		(IS_ULT(dev))
+
 #define HAS_PIPE_CONTROL(dev) (INTEL_INFO(dev)->gen >= 5)
 
-#define HAS_DDI(dev)		(IS_HASWELL(dev))
+#define HAS_DDI(dev)		(INTEL_INFO(dev)->has_ddi)
 #define HAS_POWER_WELL(dev)	(IS_HASWELL(dev))
+#define HAS_FPGA_DBG_UNCLAIMED(dev)	(INTEL_INFO(dev)->has_fpga_dbg)
 
 #define INTEL_PCH_DEVICE_ID_MASK		0xff00
 #define INTEL_PCH_IBX_DEVICE_ID_TYPE		0x3b00
@@ -1435,6 +1542,7 @@ extern bool i915_enable_hangcheck __read_mostly;
 extern int i915_enable_ppgtt __read_mostly;
 extern unsigned int i915_preliminary_hw_support __read_mostly;
 extern int i915_disable_power_well __read_mostly;
+extern int i915_enable_ips __read_mostly;
 
 extern int i915_suspend(struct drm_device *dev, pm_message_t state);
 extern int i915_resume(struct drm_device *dev);
@@ -1476,7 +1584,7 @@ void i915_handle_error(struct drm_device *dev, bool wedged);
 extern void intel_irq_init(struct drm_device *dev);
 extern void intel_hpd_init(struct drm_device *dev);
 extern void intel_gt_init(struct drm_device *dev);
-extern void intel_gt_reset(struct drm_device *dev);
+extern void intel_gt_sanitize(struct drm_device *dev);
 
 void i915_error_state_free(struct kref *error_ref);
 
@@ -1486,8 +1594,6 @@ i915_enable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask);
 void
 i915_disable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask);
 
-void intel_enable_asle(struct drm_device *dev);
-
 #ifdef CONFIG_DEBUG_FS
 extern void i915_destroy_error_state(struct drm_device *dev);
 #else
@@ -1626,6 +1732,7 @@ i915_gem_object_unpin_fence(struct drm_i915_gem_object *obj)
 {
 	if (obj->fence_reg != I915_FENCE_REG_NONE) {
 		struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+		WARN_ON(dev_priv->fence_regs[obj->fence_reg].pin_count <= 0);
 		dev_priv->fence_regs[obj->fence_reg].pin_count--;
 	}
 }
@@ -1658,9 +1765,12 @@ void i915_gem_init_swizzling(struct drm_device *dev);
 void i915_gem_cleanup_ringbuffer(struct drm_device *dev);
 int __must_check i915_gpu_idle(struct drm_device *dev);
 int __must_check i915_gem_idle(struct drm_device *dev);
-int i915_add_request(struct intel_ring_buffer *ring,
-		     struct drm_file *file,
-		     u32 *seqno);
+int __i915_add_request(struct intel_ring_buffer *ring,
+		       struct drm_file *file,
+		       struct drm_i915_gem_object *batch_obj,
+		       u32 *seqno);
+#define i915_add_request(ring, seqno) \
+	__i915_add_request(ring, NULL, NULL, seqno)
 int __must_check i915_wait_seqno(struct intel_ring_buffer *ring,
 				 uint32_t seqno);
 int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf);
@@ -1705,6 +1815,21 @@ void i915_gem_context_fini(struct drm_device *dev);
 void i915_gem_context_close(struct drm_device *dev, struct drm_file *file);
 int i915_switch_context(struct intel_ring_buffer *ring,
 			struct drm_file *file, int to_id);
+void i915_gem_context_free(struct kref *ctx_ref);
+static inline void i915_gem_context_reference(struct i915_hw_context *ctx)
+{
+	kref_get(&ctx->ref);
+}
+
+static inline void i915_gem_context_unreference(struct i915_hw_context *ctx)
+{
+	kref_put(&ctx->ref, i915_gem_context_free);
+}
+
+struct i915_ctx_hang_stats * __must_check
+i915_gem_context_get_hang_stats(struct intel_ring_buffer *ring,
+				struct drm_file *file,
+				u32 id);
 int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
 				  struct drm_file *file);
 int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
@@ -1786,6 +1911,8 @@ void i915_gem_dump_object(struct drm_i915_gem_object *obj, int len,
 /* i915_debugfs.c */
 int i915_debugfs_init(struct drm_minor *minor);
 void i915_debugfs_cleanup(struct drm_minor *minor);
+__printf(2, 3)
+void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...);
 
 /* i915_suspend.c */
 extern int i915_save_state(struct drm_device *dev);
@@ -1802,7 +1929,7 @@ void i915_teardown_sysfs(struct drm_device *dev_priv);
 /* intel_i2c.c */
 extern int intel_setup_gmbus(struct drm_device *dev);
 extern void intel_teardown_gmbus(struct drm_device *dev);
-extern inline bool intel_gmbus_is_port_valid(unsigned port)
+static inline bool intel_gmbus_is_port_valid(unsigned port)
 {
 	return (port >= GMBUS_PORT_SSC && port <= GMBUS_PORT_DPD);
 }
@@ -1811,7 +1938,7 @@ extern struct i2c_adapter *intel_gmbus_get_adapter(
 		struct drm_i915_private *dev_priv, unsigned port);
 extern void intel_gmbus_set_speed(struct i2c_adapter *adapter, int speed);
 extern void intel_gmbus_force_bit(struct i2c_adapter *adapter, bool force_bit);
-extern inline bool intel_gmbus_is_forced_bit(struct i2c_adapter *adapter)
+static inline bool intel_gmbus_is_forced_bit(struct i2c_adapter *adapter)
 {
 	return container_of(adapter, struct intel_gmbus, adapter)->force_bit;
 }
@@ -1823,14 +1950,10 @@ extern int intel_opregion_setup(struct drm_device *dev);
 extern void intel_opregion_init(struct drm_device *dev);
 extern void intel_opregion_fini(struct drm_device *dev);
 extern void intel_opregion_asle_intr(struct drm_device *dev);
-extern void intel_opregion_gse_intr(struct drm_device *dev);
-extern void intel_opregion_enable_asle(struct drm_device *dev);
 #else
 static inline void intel_opregion_init(struct drm_device *dev) { return; }
 static inline void intel_opregion_fini(struct drm_device *dev) { return; }
 static inline void intel_opregion_asle_intr(struct drm_device *dev) { return; }
-static inline void intel_opregion_gse_intr(struct drm_device *dev) { return; }
-static inline void intel_opregion_enable_asle(struct drm_device *dev) { return; }
 #endif
 
 /* intel_acpi.c */
@@ -1844,6 +1967,7 @@ static inline void intel_unregister_dsm_handler(void) { return; }
 
 /* modesetting */
 extern void intel_modeset_init_hw(struct drm_device *dev);
+extern void intel_modeset_suspend_hw(struct drm_device *dev);
 extern void intel_modeset_init(struct drm_device *dev);
 extern void intel_modeset_gem_init(struct drm_device *dev);
 extern void intel_modeset_cleanup(struct drm_device *dev);
@@ -1856,6 +1980,9 @@ extern void intel_disable_fbc(struct drm_device *dev);
 extern bool ironlake_set_drps(struct drm_device *dev, u8 val);
 extern void intel_init_pch_refclk(struct drm_device *dev);
 extern void gen6_set_rps(struct drm_device *dev, u8 val);
+extern void valleyview_set_rps(struct drm_device *dev, u8 val);
+extern int valleyview_rps_max_freq(struct drm_i915_private *dev_priv);
+extern int valleyview_rps_min_freq(struct drm_i915_private *dev_priv);
 extern void intel_detect_pch(struct drm_device *dev);
 extern int intel_trans_dp_port_sel(struct drm_crtc *crtc);
 extern int intel_enable_rc6(const struct drm_device *dev);
@@ -1867,10 +1994,11 @@ int i915_reg_read_ioctl(struct drm_device *dev, void *data,
 /* overlay */
 #ifdef CONFIG_DEBUG_FS
 extern struct intel_overlay_error_state *intel_overlay_capture_error_state(struct drm_device *dev);
-extern void intel_overlay_print_error_state(struct seq_file *m, struct intel_overlay_error_state *error);
+extern void intel_overlay_print_error_state(struct drm_i915_error_state_buf *e,
+					    struct intel_overlay_error_state *error);
 
 extern struct intel_display_error_state *intel_display_capture_error_state(struct drm_device *dev);
-extern void intel_display_print_error_state(struct seq_file *m,
+extern void intel_display_print_error_state(struct drm_i915_error_state_buf *e,
 					    struct drm_device *dev,
 					    struct intel_display_error_state *error);
 #endif
@@ -1885,8 +2013,20 @@ int __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv);
 
 int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u8 mbox, u32 *val);
 int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u8 mbox, u32 val);
-int valleyview_punit_read(struct drm_i915_private *dev_priv, u8 addr, u32 *val);
-int valleyview_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val);
+
+/* intel_sideband.c */
+u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr);
+void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val);
+u32 vlv_nc_read(struct drm_i915_private *dev_priv, u8 addr);
+u32 vlv_dpio_read(struct drm_i915_private *dev_priv, int reg);
+void vlv_dpio_write(struct drm_i915_private *dev_priv, int reg, u32 val);
+u32 intel_sbi_read(struct drm_i915_private *dev_priv, u16 reg,
+		   enum intel_sbi_destination destination);
+void intel_sbi_write(struct drm_i915_private *dev_priv, u16 reg, u32 value,
+		     enum intel_sbi_destination destination);
+
+int vlv_gpu_freq(int ddr_freq, int val);
+int vlv_freq_opcode(int ddr_freq, int val);
 
 #define __i915_read(x, y) \
 	u##x i915_read##x(struct drm_i915_private *dev_priv, u32 reg);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 9e35dafc5807..d9e2208cfe98 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -176,7 +176,7 @@ i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
 
 	pinned = 0;
 	mutex_lock(&dev->struct_mutex);
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list)
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list)
 		if (obj->pin_count)
 			pinned += obj->gtt_space->size;
 	mutex_unlock(&dev->struct_mutex);
@@ -956,7 +956,7 @@ i915_gem_check_olr(struct intel_ring_buffer *ring, u32 seqno)
 
 	ret = 0;
 	if (seqno == ring->outstanding_lazy_request)
-		ret = i915_add_request(ring, NULL, NULL);
+		ret = i915_add_request(ring, NULL);
 
 	return ret;
 }
@@ -1087,6 +1087,25 @@ i915_wait_seqno(struct intel_ring_buffer *ring, uint32_t seqno)
 			    interruptible, NULL);
 }
 
+static int
+i915_gem_object_wait_rendering__tail(struct drm_i915_gem_object *obj,
+				     struct intel_ring_buffer *ring)
+{
+	i915_gem_retire_requests_ring(ring);
+
+	/* Manually manage the write flush as we may have not yet
+	 * retired the buffer.
+	 *
+	 * Note that the last_write_seqno is always the earlier of
+	 * the two (read/write) seqno, so if we haved successfully waited,
+	 * we know we have passed the last write.
+	 */
+	obj->last_write_seqno = 0;
+	obj->base.write_domain &= ~I915_GEM_GPU_DOMAINS;
+
+	return 0;
+}
+
 /**
  * Ensures that all rendering to the object has completed and the object is
  * safe to unbind from the GTT or access from the CPU.
@@ -1107,18 +1126,7 @@ i915_gem_object_wait_rendering(struct drm_i915_gem_object *obj,
 	if (ret)
 		return ret;
 
-	i915_gem_retire_requests_ring(ring);
-
-	/* Manually manage the write flush as we may have not yet
-	 * retired the buffer.
-	 */
-	if (obj->last_write_seqno &&
-	    i915_seqno_passed(seqno, obj->last_write_seqno)) {
-		obj->last_write_seqno = 0;
-		obj->base.write_domain &= ~I915_GEM_GPU_DOMAINS;
-	}
-
-	return 0;
+	return i915_gem_object_wait_rendering__tail(obj, ring);
 }
 
 /* A nonblocking variant of the above wait. This is a highly dangerous routine
@@ -1154,19 +1162,10 @@ i915_gem_object_wait_rendering__nonblocking(struct drm_i915_gem_object *obj,
 	mutex_unlock(&dev->struct_mutex);
 	ret = __wait_seqno(ring, seqno, reset_counter, true, NULL);
 	mutex_lock(&dev->struct_mutex);
+	if (ret)
+		return ret;
 
-	i915_gem_retire_requests_ring(ring);
-
-	/* Manually manage the write flush as we may have not yet
-	 * retired the buffer.
-	 */
-	if (obj->last_write_seqno &&
-	    i915_seqno_passed(seqno, obj->last_write_seqno)) {
-		obj->last_write_seqno = 0;
-		obj->base.write_domain &= ~I915_GEM_GPU_DOMAINS;
-	}
-
-	return ret;
+	return i915_gem_object_wait_rendering__tail(obj, ring);
 }
 
 /**
@@ -1676,7 +1675,7 @@ i915_gem_object_put_pages(struct drm_i915_gem_object *obj)
 	/* ->put_pages might need to allocate memory for the bit17 swizzle
 	 * array, hence protect them from being reaped by removing them from gtt
 	 * lists early. */
-	list_del(&obj->gtt_list);
+	list_del(&obj->global_list);
 
 	ops->put_pages(obj);
 	obj->pages = NULL;
@@ -1696,7 +1695,7 @@ __i915_gem_shrink(struct drm_i915_private *dev_priv, long target,
 
 	list_for_each_entry_safe(obj, next,
 				 &dev_priv->mm.unbound_list,
-				 gtt_list) {
+				 global_list) {
 		if ((i915_gem_object_is_purgeable(obj) || !purgeable_only) &&
 		    i915_gem_object_put_pages(obj) == 0) {
 			count += obj->base.size >> PAGE_SHIFT;
@@ -1733,7 +1732,8 @@ i915_gem_shrink_all(struct drm_i915_private *dev_priv)
 
 	i915_gem_evict_everything(dev_priv->dev);
 
-	list_for_each_entry_safe(obj, next, &dev_priv->mm.unbound_list, gtt_list)
+	list_for_each_entry_safe(obj, next, &dev_priv->mm.unbound_list,
+				 global_list)
 		i915_gem_object_put_pages(obj);
 }
 
@@ -1867,7 +1867,7 @@ i915_gem_object_get_pages(struct drm_i915_gem_object *obj)
 	if (ret)
 		return ret;
 
-	list_add_tail(&obj->gtt_list, &dev_priv->mm.unbound_list);
+	list_add_tail(&obj->global_list, &dev_priv->mm.unbound_list);
 	return 0;
 }
 
@@ -1880,6 +1880,10 @@ i915_gem_object_move_to_active(struct drm_i915_gem_object *obj,
 	u32 seqno = intel_ring_get_seqno(ring);
 
 	BUG_ON(ring == NULL);
+	if (obj->ring != ring && obj->last_write_seqno) {
+		/* Keep the seqno relative to the current ring */
+		obj->last_write_seqno = seqno;
+	}
 	obj->ring = ring;
 
 	/* Add a reference if we're newly entering the active list. */
@@ -2005,17 +2009,18 @@ i915_gem_get_seqno(struct drm_device *dev, u32 *seqno)
 	return 0;
 }
 
-int
-i915_add_request(struct intel_ring_buffer *ring,
-		 struct drm_file *file,
-		 u32 *out_seqno)
+int __i915_add_request(struct intel_ring_buffer *ring,
+		       struct drm_file *file,
+		       struct drm_i915_gem_object *obj,
+		       u32 *out_seqno)
 {
 	drm_i915_private_t *dev_priv = ring->dev->dev_private;
 	struct drm_i915_gem_request *request;
-	u32 request_ring_position;
+	u32 request_ring_position, request_start;
 	int was_empty;
 	int ret;
 
+	request_start = intel_ring_get_tail(ring);
 	/*
 	 * Emit any outstanding flushes - execbuf can fail to emit the flush
 	 * after having emitted the batchbuffer command. Hence we need to fix
@@ -2047,7 +2052,21 @@ i915_add_request(struct intel_ring_buffer *ring,
 
 	request->seqno = intel_ring_get_seqno(ring);
 	request->ring = ring;
+	request->head = request_start;
 	request->tail = request_ring_position;
+	request->ctx = ring->last_context;
+	request->batch_obj = obj;
+
+	/* Whilst this request exists, batch_obj will be on the
+	 * active_list, and so will hold the active reference. Only when this
+	 * request is retired will the the batch_obj be moved onto the
+	 * inactive_list and lose its active reference. Hence we do not need
+	 * to explicitly hold another reference here.
+	 */
+
+	if (request->ctx)
+		i915_gem_context_reference(request->ctx);
+
 	request->emitted_jiffies = jiffies;
 	was_empty = list_empty(&ring->request_list);
 	list_add_tail(&request->list, &ring->request_list);
@@ -2100,9 +2119,114 @@ i915_gem_request_remove_from_client(struct drm_i915_gem_request *request)
 	spin_unlock(&file_priv->mm.lock);
 }
 
+static bool i915_head_inside_object(u32 acthd, struct drm_i915_gem_object *obj)
+{
+	if (acthd >= obj->gtt_offset &&
+	    acthd < obj->gtt_offset + obj->base.size)
+		return true;
+
+	return false;
+}
+
+static bool i915_head_inside_request(const u32 acthd_unmasked,
+				     const u32 request_start,
+				     const u32 request_end)
+{
+	const u32 acthd = acthd_unmasked & HEAD_ADDR;
+
+	if (request_start < request_end) {
+		if (acthd >= request_start && acthd < request_end)
+			return true;
+	} else if (request_start > request_end) {
+		if (acthd >= request_start || acthd < request_end)
+			return true;
+	}
+
+	return false;
+}
+
+static bool i915_request_guilty(struct drm_i915_gem_request *request,
+				const u32 acthd, bool *inside)
+{
+	/* There is a possibility that unmasked head address
+	 * pointing inside the ring, matches the batch_obj address range.
+	 * However this is extremely unlikely.
+	 */
+
+	if (request->batch_obj) {
+		if (i915_head_inside_object(acthd, request->batch_obj)) {
+			*inside = true;
+			return true;
+		}
+	}
+
+	if (i915_head_inside_request(acthd, request->head, request->tail)) {
+		*inside = false;
+		return true;
+	}
+
+	return false;
+}
+
+static void i915_set_reset_status(struct intel_ring_buffer *ring,
+				  struct drm_i915_gem_request *request,
+				  u32 acthd)
+{
+	struct i915_ctx_hang_stats *hs = NULL;
+	bool inside, guilty;
+
+	/* Innocent until proven guilty */
+	guilty = false;
+
+	if (ring->hangcheck.action != wait &&
+	    i915_request_guilty(request, acthd, &inside)) {
+		DRM_ERROR("%s hung %s bo (0x%x ctx %d) at 0x%x\n",
+			  ring->name,
+			  inside ? "inside" : "flushing",
+			  request->batch_obj ?
+			  request->batch_obj->gtt_offset : 0,
+			  request->ctx ? request->ctx->id : 0,
+			  acthd);
+
+		guilty = true;
+	}
+
+	/* If contexts are disabled or this is the default context, use
+	 * file_priv->reset_state
+	 */
+	if (request->ctx && request->ctx->id != DEFAULT_CONTEXT_ID)
+		hs = &request->ctx->hang_stats;
+	else if (request->file_priv)
+		hs = &request->file_priv->hang_stats;
+
+	if (hs) {
+		if (guilty)
+			hs->batch_active++;
+		else
+			hs->batch_pending++;
+	}
+}
+
+static void i915_gem_free_request(struct drm_i915_gem_request *request)
+{
+	list_del(&request->list);
+	i915_gem_request_remove_from_client(request);
+
+	if (request->ctx)
+		i915_gem_context_unreference(request->ctx);
+
+	kfree(request);
+}
+
 static void i915_gem_reset_ring_lists(struct drm_i915_private *dev_priv,
 				      struct intel_ring_buffer *ring)
 {
+	u32 completed_seqno;
+	u32 acthd;
+
+	acthd = intel_ring_get_active_head(ring);
+	completed_seqno = ring->get_seqno(ring, false);
+
 	while (!list_empty(&ring->request_list)) {
 		struct drm_i915_gem_request *request;
 
@@ -2110,9 +2234,10 @@ static void i915_gem_reset_ring_lists(struct drm_i915_private *dev_priv,
 					   struct drm_i915_gem_request,
 					   list);
 
-		list_del(&request->list);
-		i915_gem_request_remove_from_client(request);
-		kfree(request);
+		if (request->seqno > completed_seqno)
+			i915_set_reset_status(ring, request, acthd);
+
+		i915_gem_free_request(request);
 	}
 
 	while (!list_empty(&ring->active_list)) {
@@ -2133,7 +2258,17 @@ void i915_gem_restore_fences(struct drm_device *dev)
 
 	for (i = 0; i < dev_priv->num_fence_regs; i++) {
 		struct drm_i915_fence_reg *reg = &dev_priv->fence_regs[i];
-		i915_gem_write_fence(dev, i, reg->obj);
+
+		/*
+		 * Commit delayed tiling changes if we have an object still
+		 * attached to the fence, otherwise just clear the fence.
+		 */
+		if (reg->obj) {
+			i915_gem_object_update_fence(reg->obj, reg,
+						     reg->obj->tiling_mode);
+		} else {
+			i915_gem_write_fence(dev, i, NULL);
+		}
 	}
 }
 
@@ -2193,9 +2328,7 @@ i915_gem_retire_requests_ring(struct intel_ring_buffer *ring)
 		 */
 		ring->last_retired_head = request->tail;
 
-		list_del(&request->list);
-		i915_gem_request_remove_from_client(request);
-		kfree(request);
+		i915_gem_free_request(request);
 	}
 
 	/* Move any buffers on the active list that are no longer referenced
@@ -2262,7 +2395,7 @@ i915_gem_retire_work_handler(struct work_struct *work)
 	idle = true;
 	for_each_ring(ring, dev_priv, i) {
 		if (ring->gpu_caches_dirty)
-			i915_add_request(ring, NULL, NULL);
+			i915_add_request(ring, NULL);
 
 		idle &= list_empty(&ring->request_list);
 	}
@@ -2494,9 +2627,10 @@ i915_gem_object_unbind(struct drm_i915_gem_object *obj)
 		obj->has_aliasing_ppgtt_mapping = 0;
 	}
 	i915_gem_gtt_finish_object(obj);
+	i915_gem_object_unpin_pages(obj);
 
 	list_del(&obj->mm_list);
-	list_move_tail(&obj->gtt_list, &dev_priv->mm.unbound_list);
+	list_move_tail(&obj->global_list, &dev_priv->mm.unbound_list);
 	/* Avoid an unnecessary call to unbind on rebind. */
 	obj->map_and_fenceable = true;
 
@@ -2533,7 +2667,6 @@ static void i965_write_fence_reg(struct drm_device *dev, int reg,
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	int fence_reg;
 	int fence_pitch_shift;
-	uint64_t val;
 
 	if (INTEL_INFO(dev)->gen >= 6) {
 		fence_reg = FENCE_REG_SANDYBRIDGE_0;
@@ -2543,8 +2676,23 @@ static void i965_write_fence_reg(struct drm_device *dev, int reg,
 		fence_pitch_shift = I965_FENCE_PITCH_SHIFT;
 	}
 
+	fence_reg += reg * 8;
+
+	/* To w/a incoherency with non-atomic 64-bit register updates,
+	 * we split the 64-bit update into two 32-bit writes. In order
+	 * for a partial fence not to be evaluated between writes, we
+	 * precede the update with write to turn off the fence register,
+	 * and only enable the fence as the last step.
+	 *
+	 * For extra levels of paranoia, we make sure each step lands
+	 * before applying the next step.
+	 */
+	I915_WRITE(fence_reg, 0);
+	POSTING_READ(fence_reg);
+
 	if (obj) {
 		u32 size = obj->gtt_space->size;
+		uint64_t val;
 
 		val = (uint64_t)((obj->gtt_offset + size - 4096) &
 				 0xfffff000) << 32;
@@ -2553,12 +2701,16 @@ static void i965_write_fence_reg(struct drm_device *dev, int reg,
 		if (obj->tiling_mode == I915_TILING_Y)
 			val |= 1 << I965_FENCE_TILING_Y_SHIFT;
 		val |= I965_FENCE_REG_VALID;
-	} else
-		val = 0;
 
-	fence_reg += reg * 8;
-	I915_WRITE64(fence_reg, val);
-	POSTING_READ(fence_reg);
+		I915_WRITE(fence_reg + 4, val >> 32);
+		POSTING_READ(fence_reg + 4);
+
+		I915_WRITE(fence_reg + 0, val);
+		POSTING_READ(fence_reg);
+	} else {
+		I915_WRITE(fence_reg + 4, 0);
+		POSTING_READ(fence_reg + 4);
+	}
 }
 
 static void i915_write_fence_reg(struct drm_device *dev, int reg,
@@ -2653,6 +2805,10 @@ static void i915_gem_write_fence(struct drm_device *dev, int reg,
 	if (i915_gem_object_needs_mb(dev_priv->fence_regs[reg].obj))
 		mb();
 
+	WARN(obj && (!obj->stride || !obj->tiling_mode),
+	     "bogus fence setup with stride: 0x%x, tiling mode: %i\n",
+	     obj->stride, obj->tiling_mode);
+
 	switch (INTEL_INFO(dev)->gen) {
 	case 7:
 	case 6:
@@ -2676,35 +2832,17 @@ static inline int fence_number(struct drm_i915_private *dev_priv,
 	return fence - dev_priv->fence_regs;
 }
 
-static void i915_gem_write_fence__ipi(void *data)
-{
-	wbinvd();
-}
-
 static void i915_gem_object_update_fence(struct drm_i915_gem_object *obj,
 					 struct drm_i915_fence_reg *fence,
 					 bool enable)
 {
-	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int fence_reg = fence_number(dev_priv, fence);
-
-	/* In order to fully serialize access to the fenced region and
-	 * the update to the fence register we need to take extreme
-	 * measures on SNB+. In theory, the write to the fence register
-	 * flushes all memory transactions before, and coupled with the
-	 * mb() placed around the register write we serialise all memory
-	 * operations with respect to the changes in the tiler. Yet, on
-	 * SNB+ we need to take a step further and emit an explicit wbinvd()
-	 * on each processor in order to manually flush all memory
-	 * transactions before updating the fence register.
-	 */
-	if (HAS_LLC(obj->base.dev))
-		on_each_cpu(i915_gem_write_fence__ipi, NULL, 1);
-	i915_gem_write_fence(dev, fence_reg, enable ? obj : NULL);
+	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
+	int reg = fence_number(dev_priv, fence);
+
+	i915_gem_write_fence(obj->base.dev, reg, enable ? obj : NULL);
 
 	if (enable) {
-		obj->fence_reg = fence_reg;
+		obj->fence_reg = reg;
 		fence->obj = obj;
 		list_move_tail(&fence->lru_list, &dev_priv->mm.fence_list);
 	} else {
@@ -2712,6 +2850,7 @@ static void i915_gem_object_update_fence(struct drm_i915_gem_object *obj,
 		fence->obj = NULL;
 		list_del_init(&fence->lru_list);
 	}
+	obj->fence_dirty = false;
 }
 
 static int
@@ -2841,7 +2980,6 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj)
 		return 0;
 
 	i915_gem_object_update_fence(obj, reg, enable);
-	obj->fence_dirty = false;
 
 	return 0;
 }
@@ -2883,7 +3021,7 @@ static void i915_gem_verify_gtt(struct drm_device *dev)
 	struct drm_i915_gem_object *obj;
 	int err = 0;
 
-	list_for_each_entry(obj, &dev_priv->mm.gtt_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.gtt_list, global_list) {
 		if (obj->gtt_space == NULL) {
 			printk(KERN_ERR "object found on GTT list with no space reserved\n");
 			err++;
@@ -2930,6 +3068,8 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 	struct drm_mm_node *node;
 	u32 size, fence_size, fence_alignment, unfenced_alignment;
 	bool mappable, fenceable;
+	size_t gtt_max = map_and_fenceable ?
+		dev_priv->gtt.mappable_end : dev_priv->gtt.total;
 	int ret;
 
 	fence_size = i915_gem_get_gtt_size(dev,
@@ -2956,9 +3096,11 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 	/* If the object is bigger than the entire aperture, reject it early
 	 * before evicting everything in a vain attempt to find space.
 	 */
-	if (obj->base.size >
-	    (map_and_fenceable ? dev_priv->gtt.mappable_end : dev_priv->gtt.total)) {
-		DRM_ERROR("Attempting to bind an object larger than the aperture\n");
+	if (obj->base.size > gtt_max) {
+		DRM_ERROR("Attempting to bind an object larger than the aperture: object=%zd > %s aperture=%zu\n",
+			  obj->base.size,
+			  map_and_fenceable ? "mappable" : "total",
+			  gtt_max);
 		return -E2BIG;
 	}
 
@@ -2974,14 +3116,10 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 		return -ENOMEM;
 	}
 
- search_free:
-	if (map_and_fenceable)
-		ret = drm_mm_insert_node_in_range_generic(&dev_priv->mm.gtt_space, node,
-							  size, alignment, obj->cache_level,
-							  0, dev_priv->gtt.mappable_end);
-	else
-		ret = drm_mm_insert_node_generic(&dev_priv->mm.gtt_space, node,
-						 size, alignment, obj->cache_level);
+search_free:
+	ret = drm_mm_insert_node_in_range_generic(&dev_priv->mm.gtt_space, node,
+						  size, alignment,
+						  obj->cache_level, 0, gtt_max);
 	if (ret) {
 		ret = i915_gem_evict_something(dev, size, alignment,
 					       obj->cache_level,
@@ -3007,7 +3145,7 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 		return ret;
 	}
 
-	list_move_tail(&obj->gtt_list, &dev_priv->mm.bound_list);
+	list_move_tail(&obj->global_list, &dev_priv->mm.bound_list);
 	list_add_tail(&obj->mm_list, &dev_priv->mm.inactive_list);
 
 	obj->gtt_space = node;
@@ -3022,7 +3160,6 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 
 	obj->map_and_fenceable = mappable && fenceable;
 
-	i915_gem_object_unpin_pages(obj);
 	trace_i915_gem_object_bind(obj, map_and_fenceable);
 	i915_gem_verify_gtt(dev);
 	return 0;
@@ -3722,7 +3859,7 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj,
 			  const struct drm_i915_gem_object_ops *ops)
 {
 	INIT_LIST_HEAD(&obj->mm_list);
-	INIT_LIST_HEAD(&obj->gtt_list);
+	INIT_LIST_HEAD(&obj->global_list);
 	INIT_LIST_HEAD(&obj->ring_list);
 	INIT_LIST_HEAD(&obj->exec_list);
 
@@ -3822,7 +3959,13 @@ void i915_gem_free_object(struct drm_gem_object *gem_obj)
 		dev_priv->mm.interruptible = was_interruptible;
 	}
 
-	obj->pages_pin_count = 0;
+	/* Stolen objects don't hold a ref, but do hold pin count. Fix that up
+	 * before progressing. */
+	if (obj->stolen)
+		i915_gem_object_unpin_pages(obj);
+
+	if (WARN_ON(obj->pages_pin_count))
+		obj->pages_pin_count = 0;
 	i915_gem_object_put_pages(obj);
 	i915_gem_object_free_mmap_offset(obj);
 	i915_gem_object_release_stolen(obj);
@@ -3973,12 +4116,21 @@ static int i915_gem_init_rings(struct drm_device *dev)
 			goto cleanup_bsd_ring;
 	}
 
+	if (HAS_VEBOX(dev)) {
+		ret = intel_init_vebox_ring_buffer(dev);
+		if (ret)
+			goto cleanup_blt_ring;
+	}
+
+
 	ret = i915_gem_set_seqno(dev, ((u32)~0 - 0x1000));
 	if (ret)
-		goto cleanup_blt_ring;
+		goto cleanup_vebox_ring;
 
 	return 0;
 
+cleanup_vebox_ring:
+	intel_cleanup_ring_buffer(&dev_priv->ring[VECS]);
 cleanup_blt_ring:
 	intel_cleanup_ring_buffer(&dev_priv->ring[BCS]);
 cleanup_bsd_ring:
@@ -4453,10 +4605,10 @@ i915_gem_inactive_shrink(struct shrinker *shrinker, struct shrink_control *sc)
 	}
 
 	cnt = 0;
-	list_for_each_entry(obj, &dev_priv->mm.unbound_list, gtt_list)
+	list_for_each_entry(obj, &dev_priv->mm.unbound_list, global_list)
 		if (obj->pages_pin_count == 0)
 			cnt += obj->base.size >> PAGE_SHIFT;
-	list_for_each_entry(obj, &dev_priv->mm.inactive_list, gtt_list)
+	list_for_each_entry(obj, &dev_priv->mm.inactive_list, mm_list)
 		if (obj->pin_count == 0 && obj->pages_pin_count == 0)
 			cnt += obj->base.size >> PAGE_SHIFT;
 
diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index a1e8ecb6adf6..51b7a2171cae 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -113,7 +113,7 @@ static int get_context_size(struct drm_device *dev)
 	case 7:
 		reg = I915_READ(GEN7_CXT_SIZE);
 		if (IS_HASWELL(dev))
-			ret = HSW_CXT_TOTAL_SIZE(reg) * 64;
+			ret = HSW_CXT_TOTAL_SIZE;
 		else
 			ret = GEN7_CXT_TOTAL_SIZE(reg) * 64;
 		break;
@@ -124,10 +124,10 @@ static int get_context_size(struct drm_device *dev)
 	return ret;
 }
 
-static void do_destroy(struct i915_hw_context *ctx)
+void i915_gem_context_free(struct kref *ctx_ref)
 {
-	if (ctx->file_priv)
-		idr_remove(&ctx->file_priv->context_idr, ctx->id);
+	struct i915_hw_context *ctx = container_of(ctx_ref,
+						   typeof(*ctx), ref);
 
 	drm_gem_object_unreference(&ctx->obj->base);
 	kfree(ctx);
@@ -145,6 +145,7 @@ create_hw_context(struct drm_device *dev,
 	if (ctx == NULL)
 		return ERR_PTR(-ENOMEM);
 
+	kref_init(&ctx->ref);
 	ctx->obj = i915_gem_alloc_object(dev, dev_priv->hw_context_size);
 	if (ctx->obj == NULL) {
 		kfree(ctx);
@@ -155,7 +156,8 @@ create_hw_context(struct drm_device *dev,
 	if (INTEL_INFO(dev)->gen >= 7) {
 		ret = i915_gem_object_set_cache_level(ctx->obj,
 						      I915_CACHE_LLC_MLC);
-		if (ret)
+		/* Failure shouldn't ever happen this early */
+		if (WARN_ON(ret))
 			goto err_out;
 	}
 
@@ -169,18 +171,18 @@ create_hw_context(struct drm_device *dev,
 	if (file_priv == NULL)
 		return ctx;
 
-	ctx->file_priv = file_priv;
-
 	ret = idr_alloc(&file_priv->context_idr, ctx, DEFAULT_CONTEXT_ID + 1, 0,
 			GFP_KERNEL);
 	if (ret < 0)
 		goto err_out;
+
+	ctx->file_priv = file_priv;
 	ctx->id = ret;
 
 	return ctx;
 
 err_out:
-	do_destroy(ctx);
+	i915_gem_context_unreference(ctx);
 	return ERR_PTR(ret);
 }
 
@@ -213,12 +215,16 @@ static int create_default_context(struct drm_i915_private *dev_priv)
 	 */
 	dev_priv->ring[RCS].default_context = ctx;
 	ret = i915_gem_object_pin(ctx->obj, CONTEXT_ALIGN, false, false);
-	if (ret)
+	if (ret) {
+		DRM_DEBUG_DRIVER("Couldn't pin %d\n", ret);
 		goto err_destroy;
+	}
 
 	ret = do_switch(ctx);
-	if (ret)
+	if (ret) {
+		DRM_DEBUG_DRIVER("Switch failed %d\n", ret);
 		goto err_unpin;
+	}
 
 	DRM_DEBUG_DRIVER("Default HW context loaded\n");
 	return 0;
@@ -226,7 +232,7 @@ static int create_default_context(struct drm_i915_private *dev_priv)
 err_unpin:
 	i915_gem_object_unpin(ctx->obj);
 err_destroy:
-	do_destroy(ctx);
+	i915_gem_context_unreference(ctx);
 	return ret;
 }
 
@@ -236,6 +242,7 @@ void i915_gem_context_init(struct drm_device *dev)
 
 	if (!HAS_HW_CONTEXTS(dev)) {
 		dev_priv->hw_contexts_disabled = true;
+		DRM_DEBUG_DRIVER("Disabling HW Contexts; old hardware\n");
 		return;
 	}
 
@@ -248,11 +255,13 @@ void i915_gem_context_init(struct drm_device *dev)
 
 	if (dev_priv->hw_context_size > (1<<20)) {
 		dev_priv->hw_contexts_disabled = true;
+		DRM_DEBUG_DRIVER("Disabling HW Contexts; invalid size\n");
 		return;
 	}
 
 	if (create_default_context(dev_priv)) {
 		dev_priv->hw_contexts_disabled = true;
+		DRM_DEBUG_DRIVER("Disabling HW Contexts; create failed\n");
 		return;
 	}
 
@@ -262,6 +271,7 @@ void i915_gem_context_init(struct drm_device *dev)
 void i915_gem_context_fini(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_hw_context *dctx = dev_priv->ring[RCS].default_context;
 
 	if (dev_priv->hw_contexts_disabled)
 		return;
@@ -271,9 +281,16 @@ void i915_gem_context_fini(struct drm_device *dev)
 	 * other code, leading to spurious errors. */
 	intel_gpu_reset(dev);
 
-	i915_gem_object_unpin(dev_priv->ring[RCS].default_context->obj);
+	i915_gem_object_unpin(dctx->obj);
 
-	do_destroy(dev_priv->ring[RCS].default_context);
+	/* When default context is created and switched to, base object refcount
+	 * will be 2 (+1 from object creation and +1 from do_switch()).
+	 * i915_gem_context_fini() will be called after gpu_idle() has switched
+	 * to default context. So we need to unreference the base object once
+	 * to offset the do_switch part, so that i915_gem_context_unreference()
+	 * can then free the base object correctly. */
+	drm_gem_object_unreference(&dctx->obj->base);
+	i915_gem_context_unreference(dctx);
 }
 
 static int context_idr_cleanup(int id, void *p, void *data)
@@ -282,11 +299,38 @@ static int context_idr_cleanup(int id, void *p, void *data)
 
 	BUG_ON(id == DEFAULT_CONTEXT_ID);
 
-	do_destroy(ctx);
-
+	i915_gem_context_unreference(ctx);
 	return 0;
 }
 
+struct i915_ctx_hang_stats *
+i915_gem_context_get_hang_stats(struct intel_ring_buffer *ring,
+				struct drm_file *file,
+				u32 id)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	struct drm_i915_file_private *file_priv = file->driver_priv;
+	struct i915_hw_context *to;
+
+	if (dev_priv->hw_contexts_disabled)
+		return ERR_PTR(-ENOENT);
+
+	if (ring->id != RCS)
+		return ERR_PTR(-EINVAL);
+
+	if (file == NULL)
+		return ERR_PTR(-EINVAL);
+
+	if (id == DEFAULT_CONTEXT_ID)
+		return &file_priv->hang_stats;
+
+	to = i915_gem_context_get(file->driver_priv, id);
+	if (to == NULL)
+		return ERR_PTR(-ENOENT);
+
+	return &to->hang_stats;
+}
+
 void i915_gem_context_close(struct drm_device *dev, struct drm_file *file)
 {
 	struct drm_i915_file_private *file_priv = file->driver_priv;
@@ -325,6 +369,7 @@ mi_set_context(struct intel_ring_buffer *ring,
 	if (ret)
 		return ret;
 
+	/* WaProgramMiArbOnOffAroundMiSetContext:ivb,vlv,hsw */
 	if (IS_GEN7(ring->dev))
 		intel_ring_emit(ring, MI_ARB_ON_OFF | MI_ARB_DISABLE);
 	else
@@ -353,13 +398,13 @@ mi_set_context(struct intel_ring_buffer *ring,
 static int do_switch(struct i915_hw_context *to)
 {
 	struct intel_ring_buffer *ring = to->ring;
-	struct drm_i915_gem_object *from_obj = ring->last_context_obj;
+	struct i915_hw_context *from = ring->last_context;
 	u32 hw_flags = 0;
 	int ret;
 
-	BUG_ON(from_obj != NULL && from_obj->pin_count == 0);
+	BUG_ON(from != NULL && from->obj != NULL && from->obj->pin_count == 0);
 
-	if (from_obj == to->obj)
+	if (from == to)
 		return 0;
 
 	ret = i915_gem_object_pin(to->obj, CONTEXT_ALIGN, false, false);
@@ -382,7 +427,7 @@ static int do_switch(struct i915_hw_context *to)
 
 	if (!to->is_initialized || is_default_context(to))
 		hw_flags |= MI_RESTORE_INHIBIT;
-	else if (WARN_ON_ONCE(from_obj == to->obj)) /* not yet expected */
+	else if (WARN_ON_ONCE(from == to)) /* not yet expected */
 		hw_flags |= MI_FORCE_RESTORE;
 
 	ret = mi_set_context(ring, to, hw_flags);
@@ -397,9 +442,9 @@ static int do_switch(struct i915_hw_context *to)
 	 * is a bit suboptimal because the retiring can occur simply after the
 	 * MI_SET_CONTEXT instead of when the next seqno has completed.
 	 */
-	if (from_obj != NULL) {
-		from_obj->base.read_domains = I915_GEM_DOMAIN_INSTRUCTION;
-		i915_gem_object_move_to_active(from_obj, ring);
+	if (from != NULL) {
+		from->obj->base.read_domains = I915_GEM_DOMAIN_INSTRUCTION;
+		i915_gem_object_move_to_active(from->obj, ring);
 		/* As long as MI_SET_CONTEXT is serializing, ie. it flushes the
 		 * whole damn pipeline, we don't need to explicitly mark the
 		 * object dirty. The only exception is that the context must be
@@ -407,15 +452,26 @@ static int do_switch(struct i915_hw_context *to)
 		 * able to defer doing this until we know the object would be
 		 * swapped, but there is no way to do that yet.
 		 */
-		from_obj->dirty = 1;
-		BUG_ON(from_obj->ring != ring);
-		i915_gem_object_unpin(from_obj);
+		from->obj->dirty = 1;
+		BUG_ON(from->obj->ring != ring);
+
+		ret = i915_add_request(ring, NULL);
+		if (ret) {
+			/* Too late, we've already scheduled a context switch.
+			 * Try to undo the change so that the hw state is
+			 * consistent with out tracking. In case of emergency,
+			 * scream.
+			 */
+			WARN_ON(mi_set_context(ring, from, MI_RESTORE_INHIBIT));
+			return ret;
+		}
 
-		drm_gem_object_unreference(&from_obj->base);
+		i915_gem_object_unpin(from->obj);
+		i915_gem_context_unreference(from);
 	}
 
-	drm_gem_object_reference(&to->obj->base);
-	ring->last_context_obj = to->obj;
+	i915_gem_context_reference(to);
+	ring->last_context = to;
 	to->is_initialized = true;
 
 	return 0;
@@ -444,6 +500,8 @@ int i915_switch_context(struct intel_ring_buffer *ring,
 	if (dev_priv->hw_contexts_disabled)
 		return 0;
 
+	WARN_ON(!mutex_is_locked(&dev_priv->dev->struct_mutex));
+
 	if (ring != &dev_priv->ring[RCS])
 		return 0;
 
@@ -512,8 +570,8 @@ int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
 		return -ENOENT;
 	}
 
-	do_destroy(ctx);
-
+	idr_remove(&ctx->file_priv->context_idr, ctx->id);
+	i915_gem_context_unreference(ctx);
 	mutex_unlock(&dev->struct_mutex);
 
 	DRM_DEBUG_DRIVER("HW context %d destroyed\n", args->ctx_id);
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 117ce3813681..87a3227e5179 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -786,7 +786,7 @@ i915_gem_execbuffer_move_to_active(struct list_head *objects,
 			obj->dirty = 1;
 			obj->last_write_seqno = intel_ring_get_seqno(ring);
 			if (obj->pin_count) /* check for potential scanout */
-				intel_mark_fb_busy(obj);
+				intel_mark_fb_busy(obj, ring);
 		}
 
 		trace_i915_gem_object_change_domain(obj, old_read, old_write);
@@ -796,13 +796,14 @@ i915_gem_execbuffer_move_to_active(struct list_head *objects,
 static void
 i915_gem_execbuffer_retire_commands(struct drm_device *dev,
 				    struct drm_file *file,
-				    struct intel_ring_buffer *ring)
+				    struct intel_ring_buffer *ring,
+				    struct drm_i915_gem_object *obj)
 {
 	/* Unconditionally force add_request to emit a full flush. */
 	ring->gpu_caches_dirty = true;
 
 	/* Add a breadcrumb for the completion of the batch buffer */
-	(void)i915_add_request(ring, file, NULL);
+	(void)__i915_add_request(ring, file, obj, NULL);
 }
 
 static int
@@ -885,6 +886,15 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 			return -EPERM;
 		}
 		break;
+	case I915_EXEC_VEBOX:
+		ring = &dev_priv->ring[VECS];
+		if (ctx_id != 0) {
+			DRM_DEBUG("Ring %s doesn't support contexts\n",
+				  ring->name);
+			return -EPERM;
+		}
+		break;
+
 	default:
 		DRM_DEBUG("execbuf with unknown ring: %d\n",
 			  (int)(args->flags & I915_EXEC_RING_MASK));
@@ -1074,7 +1084,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	trace_i915_gem_ring_dispatch(ring, intel_ring_get_seqno(ring), flags);
 
 	i915_gem_execbuffer_move_to_active(&eb->objects, ring);
-	i915_gem_execbuffer_retire_commands(dev, file, ring);
+	i915_gem_execbuffer_retire_commands(dev, file, ring, batch_obj);
 
 err:
 	eb_destroy(eb);
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index bdb0d7717bc7..5101ab6869b4 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -28,8 +28,6 @@
 #include "i915_trace.h"
 #include "intel_drv.h"
 
-typedef uint32_t gen6_gtt_pte_t;
-
 /* PPGTT stuff */
 #define GEN6_GTT_ADDR_ENCODE(addr)	((addr) | (((addr) >> 28) & 0xff0))
 
@@ -44,29 +42,22 @@ typedef uint32_t gen6_gtt_pte_t;
 #define GEN6_PTE_CACHE_LLC_MLC		(3 << 1)
 #define GEN6_PTE_ADDR_ENCODE(addr)	GEN6_GTT_ADDR_ENCODE(addr)
 
-static inline gen6_gtt_pte_t gen6_pte_encode(struct drm_device *dev,
-					     dma_addr_t addr,
-					     enum i915_cache_level level)
+static gen6_gtt_pte_t gen6_pte_encode(struct drm_device *dev,
+				      dma_addr_t addr,
+				      enum i915_cache_level level)
 {
 	gen6_gtt_pte_t pte = GEN6_PTE_VALID;
 	pte |= GEN6_PTE_ADDR_ENCODE(addr);
 
 	switch (level) {
 	case I915_CACHE_LLC_MLC:
-		/* Haswell doesn't set L3 this way */
-		if (IS_HASWELL(dev))
-			pte |= GEN6_PTE_CACHE_LLC;
-		else
-			pte |= GEN6_PTE_CACHE_LLC_MLC;
+		pte |= GEN6_PTE_CACHE_LLC_MLC;
 		break;
 	case I915_CACHE_LLC:
 		pte |= GEN6_PTE_CACHE_LLC;
 		break;
 	case I915_CACHE_NONE:
-		if (IS_HASWELL(dev))
-			pte |= HSW_PTE_UNCACHED;
-		else
-			pte |= GEN6_PTE_UNCACHED;
+		pte |= GEN6_PTE_UNCACHED;
 		break;
 	default:
 		BUG();
@@ -75,16 +66,48 @@ static inline gen6_gtt_pte_t gen6_pte_encode(struct drm_device *dev,
 	return pte;
 }
 
-static int gen6_ppgtt_enable(struct drm_device *dev)
+#define BYT_PTE_WRITEABLE		(1 << 1)
+#define BYT_PTE_SNOOPED_BY_CPU_CACHES	(1 << 2)
+
+static gen6_gtt_pte_t byt_pte_encode(struct drm_device *dev,
+				     dma_addr_t addr,
+				     enum i915_cache_level level)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	uint32_t pd_offset;
-	struct intel_ring_buffer *ring;
-	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
+	gen6_gtt_pte_t pte = GEN6_PTE_VALID;
+	pte |= GEN6_PTE_ADDR_ENCODE(addr);
+
+	/* Mark the page as writeable.  Other platforms don't have a
+	 * setting for read-only/writable, so this matches that behavior.
+	 */
+	pte |= BYT_PTE_WRITEABLE;
+
+	if (level != I915_CACHE_NONE)
+		pte |= BYT_PTE_SNOOPED_BY_CPU_CACHES;
+
+	return pte;
+}
+
+static gen6_gtt_pte_t hsw_pte_encode(struct drm_device *dev,
+				     dma_addr_t addr,
+				     enum i915_cache_level level)
+{
+	gen6_gtt_pte_t pte = GEN6_PTE_VALID;
+	pte |= GEN6_PTE_ADDR_ENCODE(addr);
+
+	if (level != I915_CACHE_NONE)
+		pte |= GEN6_PTE_CACHE_LLC;
+
+	return pte;
+}
+
+static void gen6_write_pdes(struct i915_hw_ppgtt *ppgtt)
+{
+	struct drm_i915_private *dev_priv = ppgtt->dev->dev_private;
 	gen6_gtt_pte_t __iomem *pd_addr;
 	uint32_t pd_entry;
 	int i;
 
+	WARN_ON(ppgtt->pd_offset & 0x3f);
 	pd_addr = (gen6_gtt_pte_t __iomem*)dev_priv->gtt.gsm +
 		ppgtt->pd_offset / sizeof(gen6_gtt_pte_t);
 	for (i = 0; i < ppgtt->num_pd_entries; i++) {
@@ -97,6 +120,19 @@ static int gen6_ppgtt_enable(struct drm_device *dev)
 		writel(pd_entry, pd_addr + i);
 	}
 	readl(pd_addr);
+}
+
+static int gen6_ppgtt_enable(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	uint32_t pd_offset;
+	struct intel_ring_buffer *ring;
+	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
+	int i;
+
+	BUG_ON(ppgtt->pd_offset & 0x3f);
+
+	gen6_write_pdes(ppgtt);
 
 	pd_offset = ppgtt->pd_offset;
 	pd_offset /= 64; /* in cachelines, */
@@ -154,9 +190,9 @@ static void gen6_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt,
 	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
 	unsigned last_pte, i;
 
-	scratch_pte = gen6_pte_encode(ppgtt->dev,
-				      ppgtt->scratch_page_dma_addr,
-				      I915_CACHE_LLC);
+	scratch_pte = ppgtt->pte_encode(ppgtt->dev,
+					ppgtt->scratch_page_dma_addr,
+					I915_CACHE_LLC);
 
 	while (num_entries) {
 		last_pte = first_pte + num_entries;
@@ -191,8 +227,8 @@ static void gen6_ppgtt_insert_entries(struct i915_hw_ppgtt *ppgtt,
 		dma_addr_t page_addr;
 
 		page_addr = sg_page_iter_dma_address(&sg_iter);
-		pt_vaddr[act_pte] = gen6_pte_encode(ppgtt->dev, page_addr,
-						    cache_level);
+		pt_vaddr[act_pte] = ppgtt->pte_encode(ppgtt->dev, page_addr,
+						      cache_level);
 		if (++act_pte == I915_PPGTT_PT_ENTRIES) {
 			kunmap_atomic(pt_vaddr);
 			act_pt++;
@@ -233,8 +269,15 @@ static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 	/* ppgtt PDEs reside in the global gtt pagetable, which has 512*1024
 	 * entries. For aliasing ppgtt support we just steal them at the end for
 	 * now. */
-       first_pd_entry_in_global_pt = gtt_total_entries(dev_priv->gtt);
+	first_pd_entry_in_global_pt = gtt_total_entries(dev_priv->gtt);
 
+	if (IS_HASWELL(dev)) {
+		ppgtt->pte_encode = hsw_pte_encode;
+	} else if (IS_VALLEYVIEW(dev)) {
+		ppgtt->pte_encode = byt_pte_encode;
+	} else {
+		ppgtt->pte_encode = gen6_pte_encode;
+	}
 	ppgtt->num_pd_entries = I915_PPGTT_PD_ENTRIES;
 	ppgtt->enable = gen6_ppgtt_enable;
 	ppgtt->clear_range = gen6_ppgtt_clear_range;
@@ -396,7 +439,7 @@ void i915_gem_restore_gtt_mappings(struct drm_device *dev)
 	dev_priv->gtt.gtt_clear_range(dev, dev_priv->gtt.start / PAGE_SIZE,
 				      dev_priv->gtt.total / PAGE_SIZE);
 
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
 		i915_gem_clflush_object(obj);
 		i915_gem_gtt_bind_object(obj, obj->cache_level);
 	}
@@ -437,7 +480,8 @@ static void gen6_ggtt_insert_entries(struct drm_device *dev,
 
 	for_each_sg_page(st->sgl, &sg_iter, st->nents, 0) {
 		addr = sg_page_iter_dma_address(&sg_iter);
-		iowrite32(gen6_pte_encode(dev, addr, level), &gtt_entries[i]);
+		iowrite32(dev_priv->gtt.pte_encode(dev, addr, level),
+			  &gtt_entries[i]);
 		i++;
 	}
 
@@ -449,7 +493,7 @@ static void gen6_ggtt_insert_entries(struct drm_device *dev,
 	 */
 	if (i != 0)
 		WARN_ON(readl(&gtt_entries[i-1])
-			!= gen6_pte_encode(dev, addr, level));
+			!= dev_priv->gtt.pte_encode(dev, addr, level));
 
 	/* This next bit makes the above posting read even more important. We
 	 * want to flush the TLBs only after we're certain all the PTE updates
@@ -474,8 +518,9 @@ static void gen6_ggtt_clear_range(struct drm_device *dev,
 		 first_entry, num_entries, max_entries))
 		num_entries = max_entries;
 
-	scratch_pte = gen6_pte_encode(dev, dev_priv->gtt.scratch_page_dma,
-				      I915_CACHE_LLC);
+	scratch_pte = dev_priv->gtt.pte_encode(dev,
+					       dev_priv->gtt.scratch_page_dma,
+					       I915_CACHE_LLC);
 	for (i = 0; i < num_entries; i++)
 		iowrite32(scratch_pte, &gtt_base[i]);
 	readl(gtt_base);
@@ -586,7 +631,7 @@ void i915_gem_setup_global_gtt(struct drm_device *dev,
 		dev_priv->mm.gtt_space.color_adjust = i915_gtt_color_adjust;
 
 	/* Mark any preallocated objects as occupied */
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
 		DRM_DEBUG_KMS("reserving preallocated space: %x + %zx\n",
 			      obj->gtt_offset, obj->base.size);
 
@@ -809,6 +854,13 @@ int i915_gem_gtt_init(struct drm_device *dev)
 	} else {
 		dev_priv->gtt.gtt_probe = gen6_gmch_probe;
 		dev_priv->gtt.gtt_remove = gen6_gmch_remove;
+		if (IS_HASWELL(dev)) {
+			dev_priv->gtt.pte_encode = hsw_pte_encode;
+		} else if (IS_VALLEYVIEW(dev)) {
+			dev_priv->gtt.pte_encode = byt_pte_encode;
+		} else {
+			dev_priv->gtt.pte_encode = gen6_pte_encode;
+		}
 	}
 
 	ret = dev_priv->gtt.gtt_probe(dev, &dev_priv->gtt.total,
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index 130d1db27e28..982d4732cecf 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -62,7 +62,10 @@ static unsigned long i915_stolen_to_physical(struct drm_device *dev)
 	 * its value of TOLUD.
 	 */
 	base = 0;
-	if (INTEL_INFO(dev)->gen >= 6) {
+	if (IS_VALLEYVIEW(dev)) {
+		pci_read_config_dword(dev->pdev, 0x5c, &base);
+		base &= ~((1<<20) - 1);
+	} else if (INTEL_INFO(dev)->gen >= 6) {
 		/* Read Base Data of Stolen Memory Register (BDSM) directly.
 		 * Note that there is also a MCHBAR miror at 0x1080c0 or
 		 * we could use device 2:0x5c instead.
@@ -136,6 +139,7 @@ static int i915_setup_compression(struct drm_device *dev, int size)
 err_fb:
 	drm_mm_put_block(compressed_fb);
 err:
+	pr_info_once("drm: not enough stolen space for compressed buffer (need %d more bytes), disabling. Hint: you may be able to increase stolen memory size in the BIOS to avoid this.\n", size);
 	return -ENOSPC;
 }
 
@@ -143,7 +147,7 @@ int i915_gem_stolen_setup_compression(struct drm_device *dev, int size)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (dev_priv->mm.stolen_base == 0)
+	if (!drm_mm_initialized(&dev_priv->mm.stolen))
 		return -ENODEV;
 
 	if (size < dev_priv->cfb_size)
@@ -175,6 +179,9 @@ void i915_gem_cleanup_stolen(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	if (!drm_mm_initialized(&dev_priv->mm.stolen))
+		return;
+
 	i915_gem_stolen_cleanup_compression(dev);
 	drm_mm_takedown(&dev_priv->mm.stolen);
 }
@@ -182,6 +189,7 @@ void i915_gem_cleanup_stolen(struct drm_device *dev)
 int i915_gem_init_stolen(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	int bios_reserved = 0;
 
 	dev_priv->mm.stolen_base = i915_stolen_to_physical(dev);
 	if (dev_priv->mm.stolen_base == 0)
@@ -190,8 +198,12 @@ int i915_gem_init_stolen(struct drm_device *dev)
 	DRM_DEBUG_KMS("found %zd bytes of stolen memory at %08lx\n",
 		      dev_priv->gtt.stolen_size, dev_priv->mm.stolen_base);
 
+	if (IS_VALLEYVIEW(dev))
+		bios_reserved = 1024*1024; /* top 1M on VLV/BYT */
+
 	/* Basic memrange allocator for stolen space */
-	drm_mm_init(&dev_priv->mm.stolen, 0, dev_priv->gtt.stolen_size);
+	drm_mm_init(&dev_priv->mm.stolen, 0, dev_priv->gtt.stolen_size -
+		    bios_reserved);
 
 	return 0;
 }
@@ -270,7 +282,7 @@ _i915_gem_object_create_stolen(struct drm_device *dev,
 		goto cleanup;
 
 	obj->has_dma_mapping = true;
-	obj->pages_pin_count = 1;
+	i915_gem_object_pin_pages(obj);
 	obj->stolen = stolen;
 
 	obj->base.write_domain = I915_GEM_DOMAIN_GTT;
@@ -291,7 +303,7 @@ i915_gem_object_create_stolen(struct drm_device *dev, u32 size)
 	struct drm_i915_gem_object *obj;
 	struct drm_mm_node *stolen;
 
-	if (dev_priv->mm.stolen_base == 0)
+	if (!drm_mm_initialized(&dev_priv->mm.stolen))
 		return NULL;
 
 	DRM_DEBUG_KMS("creating stolen object: size=%x\n", size);
@@ -322,7 +334,7 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
 	struct drm_i915_gem_object *obj;
 	struct drm_mm_node *stolen;
 
-	if (dev_priv->mm.stolen_base == 0)
+	if (!drm_mm_initialized(&dev_priv->mm.stolen))
 		return NULL;
 
 	DRM_DEBUG_KMS("creating preallocated stolen object: stolen_offset=%x, gtt_offset=%x, size=%x\n",
@@ -330,7 +342,6 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
 
 	/* KISS and expect everything to be page-aligned */
 	BUG_ON(stolen_offset & 4095);
-	BUG_ON(gtt_offset & 4095);
 	BUG_ON(size & 4095);
 
 	if (WARN_ON(size == 0))
@@ -351,6 +362,10 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
 		return NULL;
 	}
 
+	/* Some objects just need physical mem from stolen space */
+	if (gtt_offset == -1)
+		return obj;
+
 	/* To simplify the initialisation sequence between KMS and GTT,
 	 * we allow construction of the stolen object prior to
 	 * setting up the GTT space. The actual reservation will occur
@@ -371,7 +386,7 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
 	obj->gtt_offset = gtt_offset;
 	obj->has_global_gtt_mapping = 1;
 
-	list_add_tail(&obj->gtt_list, &dev_priv->mm.bound_list);
+	list_add_tail(&obj->global_list, &dev_priv->mm.bound_list);
 	list_add_tail(&obj->mm_list, &dev_priv->mm.inactive_list);
 
 	return obj;
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 0aa2ef0d2ae0..3d92a7cef154 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -70,15 +70,6 @@ static const u32 hpd_status_gen4[] = {
 	[HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS
 };
 
-static const u32 hpd_status_i965[] = {
-	 [HPD_CRT] = CRT_HOTPLUG_INT_STATUS,
-	 [HPD_SDVO_B] = SDVOB_HOTPLUG_INT_STATUS_I965,
-	 [HPD_SDVO_C] = SDVOC_HOTPLUG_INT_STATUS_I965,
-	 [HPD_PORT_B] = PORTB_HOTPLUG_INT_STATUS,
-	 [HPD_PORT_C] = PORTC_HOTPLUG_INT_STATUS,
-	 [HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS
-};
-
 static const u32 hpd_status_i915[] = { /* i915 and valleyview are the same */
 	[HPD_CRT] = CRT_HOTPLUG_INT_STATUS,
 	[HPD_SDVO_B] = SDVOB_HOTPLUG_INT_STATUS_I915,
@@ -88,13 +79,12 @@ static const u32 hpd_status_i915[] = { /* i915 and valleyview are the same */
 	[HPD_PORT_D] = PORTD_HOTPLUG_INT_STATUS
 };
 
-static void ibx_hpd_irq_setup(struct drm_device *dev);
-static void i915_hpd_irq_setup(struct drm_device *dev);
-
 /* For display hotplug interrupt */
 static void
 ironlake_enable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
 {
+	assert_spin_locked(&dev_priv->irq_lock);
+
 	if ((dev_priv->irq_mask & mask) != 0) {
 		dev_priv->irq_mask &= ~mask;
 		I915_WRITE(DEIMR, dev_priv->irq_mask);
@@ -105,6 +95,8 @@ ironlake_enable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
 static void
 ironlake_disable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
 {
+	assert_spin_locked(&dev_priv->irq_lock);
+
 	if ((dev_priv->irq_mask & mask) != mask) {
 		dev_priv->irq_mask |= mask;
 		I915_WRITE(DEIMR, dev_priv->irq_mask);
@@ -112,6 +104,215 @@ ironlake_disable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
 	}
 }
 
+static bool ivb_can_enable_err_int(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc;
+	enum pipe pipe;
+
+	assert_spin_locked(&dev_priv->irq_lock);
+
+	for_each_pipe(pipe) {
+		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
+
+		if (crtc->cpu_fifo_underrun_disabled)
+			return false;
+	}
+
+	return true;
+}
+
+static bool cpt_can_enable_serr_int(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe pipe;
+	struct intel_crtc *crtc;
+
+	for_each_pipe(pipe) {
+		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
+
+		if (crtc->pch_fifo_underrun_disabled)
+			return false;
+	}
+
+	return true;
+}
+
+static void ironlake_set_fifo_underrun_reporting(struct drm_device *dev,
+						 enum pipe pipe, bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t bit = (pipe == PIPE_A) ? DE_PIPEA_FIFO_UNDERRUN :
+					  DE_PIPEB_FIFO_UNDERRUN;
+
+	if (enable)
+		ironlake_enable_display_irq(dev_priv, bit);
+	else
+		ironlake_disable_display_irq(dev_priv, bit);
+}
+
+static void ivybridge_set_fifo_underrun_reporting(struct drm_device *dev,
+						  bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (enable) {
+		if (!ivb_can_enable_err_int(dev))
+			return;
+
+		I915_WRITE(GEN7_ERR_INT, ERR_INT_FIFO_UNDERRUN_A |
+					 ERR_INT_FIFO_UNDERRUN_B |
+					 ERR_INT_FIFO_UNDERRUN_C);
+
+		ironlake_enable_display_irq(dev_priv, DE_ERR_INT_IVB);
+	} else {
+		ironlake_disable_display_irq(dev_priv, DE_ERR_INT_IVB);
+	}
+}
+
+static void ibx_set_fifo_underrun_reporting(struct intel_crtc *crtc,
+					    bool enable)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t bit = (crtc->pipe == PIPE_A) ? SDE_TRANSA_FIFO_UNDER :
+						SDE_TRANSB_FIFO_UNDER;
+
+	if (enable)
+		I915_WRITE(SDEIMR, I915_READ(SDEIMR) & ~bit);
+	else
+		I915_WRITE(SDEIMR, I915_READ(SDEIMR) | bit);
+
+	POSTING_READ(SDEIMR);
+}
+
+static void cpt_set_fifo_underrun_reporting(struct drm_device *dev,
+					    enum transcoder pch_transcoder,
+					    bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (enable) {
+		if (!cpt_can_enable_serr_int(dev))
+			return;
+
+		I915_WRITE(SERR_INT, SERR_INT_TRANS_A_FIFO_UNDERRUN |
+				     SERR_INT_TRANS_B_FIFO_UNDERRUN |
+				     SERR_INT_TRANS_C_FIFO_UNDERRUN);
+
+		I915_WRITE(SDEIMR, I915_READ(SDEIMR) & ~SDE_ERROR_CPT);
+	} else {
+		I915_WRITE(SDEIMR, I915_READ(SDEIMR) | SDE_ERROR_CPT);
+	}
+
+	POSTING_READ(SDEIMR);
+}
+
+/**
+ * intel_set_cpu_fifo_underrun_reporting - enable/disable FIFO underrun messages
+ * @dev: drm device
+ * @pipe: pipe
+ * @enable: true if we want to report FIFO underrun errors, false otherwise
+ *
+ * This function makes us disable or enable CPU fifo underruns for a specific
+ * pipe. Notice that on some Gens (e.g. IVB, HSW), disabling FIFO underrun
+ * reporting for one pipe may also disable all the other CPU error interruts for
+ * the other pipes, due to the fact that there's just one interrupt mask/enable
+ * bit for all the pipes.
+ *
+ * Returns the previous state of underrun reporting.
+ */
+bool intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
+					   enum pipe pipe, bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	unsigned long flags;
+	bool ret;
+
+	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+
+	ret = !intel_crtc->cpu_fifo_underrun_disabled;
+
+	if (enable == ret)
+		goto done;
+
+	intel_crtc->cpu_fifo_underrun_disabled = !enable;
+
+	if (IS_GEN5(dev) || IS_GEN6(dev))
+		ironlake_set_fifo_underrun_reporting(dev, pipe, enable);
+	else if (IS_GEN7(dev))
+		ivybridge_set_fifo_underrun_reporting(dev, enable);
+
+done:
+	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	return ret;
+}
+
+/**
+ * intel_set_pch_fifo_underrun_reporting - enable/disable FIFO underrun messages
+ * @dev: drm device
+ * @pch_transcoder: the PCH transcoder (same as pipe on IVB and older)
+ * @enable: true if we want to report FIFO underrun errors, false otherwise
+ *
+ * This function makes us disable or enable PCH fifo underruns for a specific
+ * PCH transcoder. Notice that on some PCHs (e.g. CPT/PPT), disabling FIFO
+ * underrun reporting for one transcoder may also disable all the other PCH
+ * error interruts for the other transcoders, due to the fact that there's just
+ * one interrupt mask/enable bit for all the transcoders.
+ *
+ * Returns the previous state of underrun reporting.
+ */
+bool intel_set_pch_fifo_underrun_reporting(struct drm_device *dev,
+					   enum transcoder pch_transcoder,
+					   bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe p;
+	struct drm_crtc *crtc;
+	struct intel_crtc *intel_crtc;
+	unsigned long flags;
+	bool ret;
+
+	if (HAS_PCH_LPT(dev)) {
+		crtc = NULL;
+		for_each_pipe(p) {
+			struct drm_crtc *c = dev_priv->pipe_to_crtc_mapping[p];
+			if (intel_pipe_has_type(c, INTEL_OUTPUT_ANALOG)) {
+				crtc = c;
+				break;
+			}
+		}
+		if (!crtc) {
+			DRM_ERROR("PCH FIFO underrun, but no CRTC using the PCH found\n");
+			return false;
+		}
+	} else {
+		crtc = dev_priv->pipe_to_crtc_mapping[pch_transcoder];
+	}
+	intel_crtc = to_intel_crtc(crtc);
+
+	spin_lock_irqsave(&dev_priv->irq_lock, flags);
+
+	ret = !intel_crtc->pch_fifo_underrun_disabled;
+
+	if (enable == ret)
+		goto done;
+
+	intel_crtc->pch_fifo_underrun_disabled = !enable;
+
+	if (HAS_PCH_IBX(dev))
+		ibx_set_fifo_underrun_reporting(intel_crtc, enable);
+	else
+		cpt_set_fifo_underrun_reporting(dev, pch_transcoder, enable);
+
+done:
+	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
+	return ret;
+}
+
+
 void
 i915_enable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask)
 {
@@ -142,28 +343,21 @@ i915_disable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask)
 }
 
 /**
- * intel_enable_asle - enable ASLE interrupt for OpRegion
+ * i915_enable_asle_pipestat - enable ASLE pipestat for OpRegion
  */
-void intel_enable_asle(struct drm_device *dev)
+static void i915_enable_asle_pipestat(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	unsigned long irqflags;
 
-	/* FIXME: opregion/asle for VLV */
-	if (IS_VALLEYVIEW(dev))
+	if (!dev_priv->opregion.asle || !IS_MOBILE(dev))
 		return;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
 
-	if (HAS_PCH_SPLIT(dev))
-		ironlake_enable_display_irq(dev_priv, DE_GSE);
-	else {
-		i915_enable_pipestat(dev_priv, 1,
-				     PIPE_LEGACY_BLC_EVENT_ENABLE);
-		if (INTEL_INFO(dev)->gen >= 4)
-			i915_enable_pipestat(dev_priv, 0,
-					     PIPE_LEGACY_BLC_EVENT_ENABLE);
-	}
+	i915_enable_pipestat(dev_priv, 1, PIPE_LEGACY_BLC_EVENT_ENABLE);
+	if (INTEL_INFO(dev)->gen >= 4)
+		i915_enable_pipestat(dev_priv, 0, PIPE_LEGACY_BLC_EVENT_ENABLE);
 
 	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 }
@@ -181,10 +375,16 @@ static int
 i915_pipe_enabled(struct drm_device *dev, int pipe)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
-	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
-								      pipe);
 
-	return I915_READ(PIPECONF(cpu_transcoder)) & PIPECONF_ENABLE;
+	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
+		/* Locking is horribly broken here, but whatever. */
+		struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+		return intel_crtc->active;
+	} else {
+		return I915_READ(PIPECONF(pipe)) & PIPECONF_ENABLE;
+	}
 }
 
 /* Called from drm generic code, passed a 'crtc', which
@@ -334,6 +534,21 @@ static int i915_get_vblank_timestamp(struct drm_device *dev, int pipe,
 						     crtc);
 }
 
+static int intel_hpd_irq_event(struct drm_device *dev, struct drm_connector *connector)
+{
+	enum drm_connector_status old_status;
+
+	WARN_ON(!mutex_is_locked(&dev->mode_config.mutex));
+	old_status = connector->status;
+
+	connector->status = connector->funcs->detect(connector, false);
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s] status updated from %d to %d\n",
+		      connector->base.id,
+		      drm_get_connector_name(connector),
+		      old_status, connector->status);
+	return (old_status != connector->status);
+}
+
 /*
  * Handle hotplug events outside the interrupt handler proper.
  */
@@ -350,6 +565,8 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	struct drm_connector *connector;
 	unsigned long irqflags;
 	bool hpd_disabled = false;
+	bool changed = false;
+	u32 hpd_event_bits;
 
 	/* HPD irq before everything is fully set up. */
 	if (!dev_priv->enable_hotplug_processing)
@@ -359,6 +576,9 @@ static void i915_hotplug_work_func(struct work_struct *work)
 	DRM_DEBUG_KMS("running encoder hotplug functions\n");
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+
+	hpd_event_bits = dev_priv->hpd_event_bits;
+	dev_priv->hpd_event_bits = 0;
 	list_for_each_entry(connector, &mode_config->connector_list, head) {
 		intel_connector = to_intel_connector(connector);
 		intel_encoder = intel_connector->encoder;
@@ -373,6 +593,10 @@ static void i915_hotplug_work_func(struct work_struct *work)
 				| DRM_CONNECTOR_POLL_DISCONNECT;
 			hpd_disabled = true;
 		}
+		if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
+			DRM_DEBUG_KMS("Connector %s (pin %i) received hotplug event.\n",
+				      drm_get_connector_name(connector), intel_encoder->hpd_pin);
+		}
 	}
 	 /* if there were no outputs to poll, poll was disabled,
 	  * therefore make sure it's enabled when disabling HPD on
@@ -385,14 +609,20 @@ static void i915_hotplug_work_func(struct work_struct *work)
 
 	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 
-	list_for_each_entry(intel_encoder, &mode_config->encoder_list, base.head)
-		if (intel_encoder->hot_plug)
-			intel_encoder->hot_plug(intel_encoder);
-
+	list_for_each_entry(connector, &mode_config->connector_list, head) {
+		intel_connector = to_intel_connector(connector);
+		intel_encoder = intel_connector->encoder;
+		if (hpd_event_bits & (1 << intel_encoder->hpd_pin)) {
+			if (intel_encoder->hot_plug)
+				intel_encoder->hot_plug(intel_encoder);
+			if (intel_hpd_irq_event(dev, connector))
+				changed = true;
+		}
+	}
 	mutex_unlock(&mode_config->mutex);
 
-	/* Just fire off a uevent and let userspace tell us what to do */
-	drm_helper_hpd_irq_event(dev);
+	if (changed)
+		drm_kms_helper_hotplug_event(dev);
 }
 
 static void ironlake_handle_rps_change(struct drm_device *dev)
@@ -447,7 +677,6 @@ static void notify_ring(struct drm_device *dev,
 
 	wake_up_all(&ring->irq_queue);
 	if (i915_enable_hangcheck) {
-		dev_priv->gpu_error.hangcheck_count = 0;
 		mod_timer(&dev_priv->gpu_error.hangcheck_timer,
 			  round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES));
 	}
@@ -464,25 +693,48 @@ static void gen6_pm_rps_work(struct work_struct *work)
 	pm_iir = dev_priv->rps.pm_iir;
 	dev_priv->rps.pm_iir = 0;
 	pm_imr = I915_READ(GEN6_PMIMR);
-	I915_WRITE(GEN6_PMIMR, 0);
+	/* Make sure not to corrupt PMIMR state used by ringbuffer code */
+	I915_WRITE(GEN6_PMIMR, pm_imr & ~GEN6_PM_RPS_EVENTS);
 	spin_unlock_irq(&dev_priv->rps.lock);
 
-	if ((pm_iir & GEN6_PM_DEFERRED_EVENTS) == 0)
+	if ((pm_iir & GEN6_PM_RPS_EVENTS) == 0)
 		return;
 
 	mutex_lock(&dev_priv->rps.hw_lock);
 
-	if (pm_iir & GEN6_PM_RP_UP_THRESHOLD)
+	if (pm_iir & GEN6_PM_RP_UP_THRESHOLD) {
 		new_delay = dev_priv->rps.cur_delay + 1;
-	else
+
+		/*
+		 * For better performance, jump directly
+		 * to RPe if we're below it.
+		 */
+		if (IS_VALLEYVIEW(dev_priv->dev) &&
+		    dev_priv->rps.cur_delay < dev_priv->rps.rpe_delay)
+			new_delay = dev_priv->rps.rpe_delay;
+	} else
 		new_delay = dev_priv->rps.cur_delay - 1;
 
 	/* sysfs frequency interfaces may have snuck in while servicing the
 	 * interrupt
 	 */
-	if (!(new_delay > dev_priv->rps.max_delay ||
-	      new_delay < dev_priv->rps.min_delay)) {
-		gen6_set_rps(dev_priv->dev, new_delay);
+	if (new_delay >= dev_priv->rps.min_delay &&
+	    new_delay <= dev_priv->rps.max_delay) {
+		if (IS_VALLEYVIEW(dev_priv->dev))
+			valleyview_set_rps(dev_priv->dev, new_delay);
+		else
+			gen6_set_rps(dev_priv->dev, new_delay);
+	}
+
+	if (IS_VALLEYVIEW(dev_priv->dev)) {
+		/*
+		 * On VLV, when we enter RC6 we may not be at the minimum
+		 * voltage level, so arm a timer to check.  It should only
+		 * fire when there's activity or once after we've entered
+		 * RC6, and then won't be re-armed until the next RPS interrupt.
+		 */
+		mod_delayed_work(dev_priv->wq, &dev_priv->rps.vlv_work,
+				 msecs_to_jiffies(100));
 	}
 
 	mutex_unlock(&dev_priv->rps.hw_lock);
@@ -529,7 +781,7 @@ static void ivybridge_parity_work(struct work_struct *work)
 	I915_WRITE(GEN7_MISCCPCTL, misccpctl);
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	dev_priv->gt_irq_mask &= ~GT_GEN7_L3_PARITY_ERROR_INTERRUPT;
+	dev_priv->gt_irq_mask &= ~GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
 	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 
@@ -561,7 +813,7 @@ static void ivybridge_handle_parity_error(struct drm_device *dev)
 		return;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	dev_priv->gt_irq_mask |= GT_GEN7_L3_PARITY_ERROR_INTERRUPT;
+	dev_priv->gt_irq_mask |= GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
 	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 	spin_unlock_irqrestore(&dev_priv->irq_lock, flags);
 
@@ -573,25 +825,26 @@ static void snb_gt_irq_handler(struct drm_device *dev,
 			       u32 gt_iir)
 {
 
-	if (gt_iir & (GEN6_RENDER_USER_INTERRUPT |
-		      GEN6_RENDER_PIPE_CONTROL_NOTIFY_INTERRUPT))
+	if (gt_iir &
+	    (GT_RENDER_USER_INTERRUPT | GT_RENDER_PIPECTL_NOTIFY_INTERRUPT))
 		notify_ring(dev, &dev_priv->ring[RCS]);
-	if (gt_iir & GEN6_BSD_USER_INTERRUPT)
+	if (gt_iir & GT_BSD_USER_INTERRUPT)
 		notify_ring(dev, &dev_priv->ring[VCS]);
-	if (gt_iir & GEN6_BLITTER_USER_INTERRUPT)
+	if (gt_iir & GT_BLT_USER_INTERRUPT)
 		notify_ring(dev, &dev_priv->ring[BCS]);
 
-	if (gt_iir & (GT_GEN6_BLT_CS_ERROR_INTERRUPT |
-		      GT_GEN6_BSD_CS_ERROR_INTERRUPT |
-		      GT_RENDER_CS_ERROR_INTERRUPT)) {
+	if (gt_iir & (GT_BLT_CS_ERROR_INTERRUPT |
+		      GT_BSD_CS_ERROR_INTERRUPT |
+		      GT_RENDER_CS_MASTER_ERROR_INTERRUPT)) {
 		DRM_ERROR("GT error interrupt 0x%08x\n", gt_iir);
 		i915_handle_error(dev, false);
 	}
 
-	if (gt_iir & GT_GEN7_L3_PARITY_ERROR_INTERRUPT)
+	if (gt_iir & GT_RENDER_L3_PARITY_ERROR_INTERRUPT)
 		ivybridge_handle_parity_error(dev);
 }
 
+/* Legacy way of handling PM interrupts */
 static void gen6_queue_rps_work(struct drm_i915_private *dev_priv,
 				u32 pm_iir)
 {
@@ -619,23 +872,25 @@ static void gen6_queue_rps_work(struct drm_i915_private *dev_priv,
 #define HPD_STORM_DETECT_PERIOD 1000
 #define HPD_STORM_THRESHOLD 5
 
-static inline bool hotplug_irq_storm_detect(struct drm_device *dev,
-					    u32 hotplug_trigger,
-					    const u32 *hpd)
+static inline void intel_hpd_irq_handler(struct drm_device *dev,
+					 u32 hotplug_trigger,
+					 const u32 *hpd)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	unsigned long irqflags;
 	int i;
-	bool ret = false;
+	bool storm_detected = false;
 
-	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	if (!hotplug_trigger)
+		return;
 
+	spin_lock(&dev_priv->irq_lock);
 	for (i = 1; i < HPD_NUM_PINS; i++) {
 
 		if (!(hpd[i] & hotplug_trigger) ||
 		    dev_priv->hpd_stats[i].hpd_mark != HPD_ENABLED)
 			continue;
 
+		dev_priv->hpd_event_bits |= (1 << i);
 		if (!time_in_range(jiffies, dev_priv->hpd_stats[i].hpd_last_jiffies,
 				   dev_priv->hpd_stats[i].hpd_last_jiffies
 				   + msecs_to_jiffies(HPD_STORM_DETECT_PERIOD))) {
@@ -643,16 +898,20 @@ static inline bool hotplug_irq_storm_detect(struct drm_device *dev,
 			dev_priv->hpd_stats[i].hpd_cnt = 0;
 		} else if (dev_priv->hpd_stats[i].hpd_cnt > HPD_STORM_THRESHOLD) {
 			dev_priv->hpd_stats[i].hpd_mark = HPD_MARK_DISABLED;
+			dev_priv->hpd_event_bits &= ~(1 << i);
 			DRM_DEBUG_KMS("HPD interrupt storm detected on PIN %d\n", i);
-			ret = true;
+			storm_detected = true;
 		} else {
 			dev_priv->hpd_stats[i].hpd_cnt++;
 		}
 	}
 
-	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+	if (storm_detected)
+		dev_priv->display.hpd_irq_setup(dev);
+	spin_unlock(&dev_priv->irq_lock);
 
-	return ret;
+	queue_work(dev_priv->wq,
+		   &dev_priv->hotplug_work);
 }
 
 static void gmbus_irq_handler(struct drm_device *dev)
@@ -669,6 +928,38 @@ static void dp_aux_irq_handler(struct drm_device *dev)
 	wake_up_all(&dev_priv->gmbus_wait_queue);
 }
 
+/* Unlike gen6_queue_rps_work() from which this function is originally derived,
+ * we must be able to deal with other PM interrupts. This is complicated because
+ * of the way in which we use the masks to defer the RPS work (which for
+ * posterity is necessary because of forcewake).
+ */
+static void hsw_pm_irq_handler(struct drm_i915_private *dev_priv,
+			       u32 pm_iir)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_priv->rps.lock, flags);
+	dev_priv->rps.pm_iir |= pm_iir & GEN6_PM_RPS_EVENTS;
+	if (dev_priv->rps.pm_iir) {
+		I915_WRITE(GEN6_PMIMR, dev_priv->rps.pm_iir);
+		/* never want to mask useful interrupts. (also posting read) */
+		WARN_ON(I915_READ_NOTRACE(GEN6_PMIMR) & ~GEN6_PM_RPS_EVENTS);
+		/* TODO: if queue_work is slow, move it out of the spinlock */
+		queue_work(dev_priv->wq, &dev_priv->rps.work);
+	}
+	spin_unlock_irqrestore(&dev_priv->rps.lock, flags);
+
+	if (pm_iir & ~GEN6_PM_RPS_EVENTS) {
+		if (pm_iir & PM_VEBOX_USER_INTERRUPT)
+			notify_ring(dev_priv->dev, &dev_priv->ring[VECS]);
+
+		if (pm_iir & PM_VEBOX_CS_ERROR_INTERRUPT) {
+			DRM_ERROR("VEBOX CS error interrupt 0x%08x\n", pm_iir);
+			i915_handle_error(dev_priv->dev, false);
+		}
+	}
+}
+
 static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = (struct drm_device *) arg;
@@ -727,12 +1018,9 @@ static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 
 			DRM_DEBUG_DRIVER("hotplug event received, stat 0x%08x\n",
 					 hotplug_status);
-			if (hotplug_trigger) {
-				if (hotplug_irq_storm_detect(dev, hotplug_trigger, hpd_status_i915))
-					i915_hpd_irq_setup(dev);
-				queue_work(dev_priv->wq,
-					   &dev_priv->hotplug_work);
-			}
+
+			intel_hpd_irq_handler(dev, hotplug_trigger, hpd_status_i915);
+
 			I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
 			I915_READ(PORT_HOTPLUG_STAT);
 		}
@@ -740,7 +1028,7 @@ static irqreturn_t valleyview_irq_handler(int irq, void *arg)
 		if (pipe_stats[0] & PIPE_GMBUS_INTERRUPT_STATUS)
 			gmbus_irq_handler(dev);
 
-		if (pm_iir & GEN6_PM_DEFERRED_EVENTS)
+		if (pm_iir & GEN6_PM_RPS_EVENTS)
 			gen6_queue_rps_work(dev_priv, pm_iir);
 
 		I915_WRITE(GTIIR, gt_iir);
@@ -758,15 +1046,14 @@ static void ibx_irq_handler(struct drm_device *dev, u32 pch_iir)
 	int pipe;
 	u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK;
 
-	if (hotplug_trigger) {
-		if (hotplug_irq_storm_detect(dev, hotplug_trigger, hpd_ibx))
-			ibx_hpd_irq_setup(dev);
-		queue_work(dev_priv->wq, &dev_priv->hotplug_work);
-	}
-	if (pch_iir & SDE_AUDIO_POWER_MASK)
+	intel_hpd_irq_handler(dev, hotplug_trigger, hpd_ibx);
+
+	if (pch_iir & SDE_AUDIO_POWER_MASK) {
+		int port = ffs((pch_iir & SDE_AUDIO_POWER_MASK) >>
+			       SDE_AUDIO_POWER_SHIFT);
 		DRM_DEBUG_DRIVER("PCH audio power change on port %d\n",
-				 (pch_iir & SDE_AUDIO_POWER_MASK) >>
-				 SDE_AUDIO_POWER_SHIFT);
+				 port_name(port));
+	}
 
 	if (pch_iir & SDE_AUX_MASK)
 		dp_aux_irq_handler(dev);
@@ -795,10 +1082,64 @@ static void ibx_irq_handler(struct drm_device *dev, u32 pch_iir)
 	if (pch_iir & (SDE_TRANSB_CRC_ERR | SDE_TRANSA_CRC_ERR))
 		DRM_DEBUG_DRIVER("PCH transcoder CRC error interrupt\n");
 
-	if (pch_iir & SDE_TRANSB_FIFO_UNDER)
-		DRM_DEBUG_DRIVER("PCH transcoder B underrun interrupt\n");
 	if (pch_iir & SDE_TRANSA_FIFO_UNDER)
-		DRM_DEBUG_DRIVER("PCH transcoder A underrun interrupt\n");
+		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A,
+							  false))
+			DRM_DEBUG_DRIVER("PCH transcoder A FIFO underrun\n");
+
+	if (pch_iir & SDE_TRANSB_FIFO_UNDER)
+		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_B,
+							  false))
+			DRM_DEBUG_DRIVER("PCH transcoder B FIFO underrun\n");
+}
+
+static void ivb_err_int_handler(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 err_int = I915_READ(GEN7_ERR_INT);
+
+	if (err_int & ERR_INT_POISON)
+		DRM_ERROR("Poison interrupt\n");
+
+	if (err_int & ERR_INT_FIFO_UNDERRUN_A)
+		if (intel_set_cpu_fifo_underrun_reporting(dev, PIPE_A, false))
+			DRM_DEBUG_DRIVER("Pipe A FIFO underrun\n");
+
+	if (err_int & ERR_INT_FIFO_UNDERRUN_B)
+		if (intel_set_cpu_fifo_underrun_reporting(dev, PIPE_B, false))
+			DRM_DEBUG_DRIVER("Pipe B FIFO underrun\n");
+
+	if (err_int & ERR_INT_FIFO_UNDERRUN_C)
+		if (intel_set_cpu_fifo_underrun_reporting(dev, PIPE_C, false))
+			DRM_DEBUG_DRIVER("Pipe C FIFO underrun\n");
+
+	I915_WRITE(GEN7_ERR_INT, err_int);
+}
+
+static void cpt_serr_int_handler(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 serr_int = I915_READ(SERR_INT);
+
+	if (serr_int & SERR_INT_POISON)
+		DRM_ERROR("PCH poison interrupt\n");
+
+	if (serr_int & SERR_INT_TRANS_A_FIFO_UNDERRUN)
+		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A,
+							  false))
+			DRM_DEBUG_DRIVER("PCH transcoder A FIFO underrun\n");
+
+	if (serr_int & SERR_INT_TRANS_B_FIFO_UNDERRUN)
+		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_B,
+							  false))
+			DRM_DEBUG_DRIVER("PCH transcoder B FIFO underrun\n");
+
+	if (serr_int & SERR_INT_TRANS_C_FIFO_UNDERRUN)
+		if (intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_C,
+							  false))
+			DRM_DEBUG_DRIVER("PCH transcoder C FIFO underrun\n");
+
+	I915_WRITE(SERR_INT, serr_int);
 }
 
 static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
@@ -807,15 +1148,14 @@ static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
 	int pipe;
 	u32 hotplug_trigger = pch_iir & SDE_HOTPLUG_MASK_CPT;
 
-	if (hotplug_trigger) {
-		if (hotplug_irq_storm_detect(dev, hotplug_trigger, hpd_cpt))
-			ibx_hpd_irq_setup(dev);
-		queue_work(dev_priv->wq, &dev_priv->hotplug_work);
+	intel_hpd_irq_handler(dev, hotplug_trigger, hpd_cpt);
+
+	if (pch_iir & SDE_AUDIO_POWER_MASK_CPT) {
+		int port = ffs((pch_iir & SDE_AUDIO_POWER_MASK_CPT) >>
+			       SDE_AUDIO_POWER_SHIFT_CPT);
+		DRM_DEBUG_DRIVER("PCH audio power change on port %c\n",
+				 port_name(port));
 	}
-	if (pch_iir & SDE_AUDIO_POWER_MASK_CPT)
-		DRM_DEBUG_DRIVER("PCH audio power change on port %d\n",
-				 (pch_iir & SDE_AUDIO_POWER_MASK_CPT) >>
-				 SDE_AUDIO_POWER_SHIFT_CPT);
 
 	if (pch_iir & SDE_AUX_MASK_CPT)
 		dp_aux_irq_handler(dev);
@@ -834,6 +1174,9 @@ static void cpt_irq_handler(struct drm_device *dev, u32 pch_iir)
 			DRM_DEBUG_DRIVER("  pipe %c FDI IIR: 0x%08x\n",
 					 pipe_name(pipe),
 					 I915_READ(FDI_RX_IIR(pipe)));
+
+	if (pch_iir & SDE_ERROR_CPT)
+		cpt_serr_int_handler(dev);
 }
 
 static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
@@ -846,6 +1189,14 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 
 	atomic_inc(&dev_priv->irq_received);
 
+	/* We get interrupts on unclaimed registers, so check for this before we
+	 * do any I915_{READ,WRITE}. */
+	if (IS_HASWELL(dev) &&
+	    (I915_READ_NOTRACE(FPGA_DBG) & FPGA_DBG_RM_NOCLAIM)) {
+		DRM_ERROR("Unclaimed register before interrupt\n");
+		I915_WRITE_NOTRACE(FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
+	}
+
 	/* disable master interrupt before clearing iir  */
 	de_ier = I915_READ(DEIER);
 	I915_WRITE(DEIER, de_ier & ~DE_MASTER_IRQ_CONTROL);
@@ -861,6 +1212,15 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 		POSTING_READ(SDEIER);
 	}
 
+	/* On Haswell, also mask ERR_INT because we don't want to risk
+	 * generating "unclaimed register" interrupts from inside the interrupt
+	 * handler. */
+	if (IS_HASWELL(dev)) {
+		spin_lock(&dev_priv->irq_lock);
+		ironlake_disable_display_irq(dev_priv, DE_ERR_INT_IVB);
+		spin_unlock(&dev_priv->irq_lock);
+	}
+
 	gt_iir = I915_READ(GTIIR);
 	if (gt_iir) {
 		snb_gt_irq_handler(dev, dev_priv, gt_iir);
@@ -870,11 +1230,14 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 
 	de_iir = I915_READ(DEIIR);
 	if (de_iir) {
+		if (de_iir & DE_ERR_INT_IVB)
+			ivb_err_int_handler(dev);
+
 		if (de_iir & DE_AUX_CHANNEL_A_IVB)
 			dp_aux_irq_handler(dev);
 
 		if (de_iir & DE_GSE_IVB)
-			intel_opregion_gse_intr(dev);
+			intel_opregion_asle_intr(dev);
 
 		for (i = 0; i < 3; i++) {
 			if (de_iir & (DE_PIPEA_VBLANK_IVB << (5 * i)))
@@ -901,12 +1264,21 @@ static irqreturn_t ivybridge_irq_handler(int irq, void *arg)
 
 	pm_iir = I915_READ(GEN6_PMIIR);
 	if (pm_iir) {
-		if (pm_iir & GEN6_PM_DEFERRED_EVENTS)
+		if (IS_HASWELL(dev))
+			hsw_pm_irq_handler(dev_priv, pm_iir);
+		else if (pm_iir & GEN6_PM_RPS_EVENTS)
 			gen6_queue_rps_work(dev_priv, pm_iir);
 		I915_WRITE(GEN6_PMIIR, pm_iir);
 		ret = IRQ_HANDLED;
 	}
 
+	if (IS_HASWELL(dev)) {
+		spin_lock(&dev_priv->irq_lock);
+		if (ivb_can_enable_err_int(dev))
+			ironlake_enable_display_irq(dev_priv, DE_ERR_INT_IVB);
+		spin_unlock(&dev_priv->irq_lock);
+	}
+
 	I915_WRITE(DEIER, de_ier);
 	POSTING_READ(DEIER);
 	if (!HAS_PCH_NOP(dev)) {
@@ -921,9 +1293,10 @@ static void ilk_gt_irq_handler(struct drm_device *dev,
 			       struct drm_i915_private *dev_priv,
 			       u32 gt_iir)
 {
-	if (gt_iir & (GT_USER_INTERRUPT | GT_PIPE_NOTIFY))
+	if (gt_iir &
+	    (GT_RENDER_USER_INTERRUPT | GT_RENDER_PIPECTL_NOTIFY_INTERRUPT))
 		notify_ring(dev, &dev_priv->ring[RCS]);
-	if (gt_iir & GT_BSD_USER_INTERRUPT)
+	if (gt_iir & ILK_BSD_USER_INTERRUPT)
 		notify_ring(dev, &dev_priv->ring[VCS]);
 }
 
@@ -968,7 +1341,7 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 		dp_aux_irq_handler(dev);
 
 	if (de_iir & DE_GSE)
-		intel_opregion_gse_intr(dev);
+		intel_opregion_asle_intr(dev);
 
 	if (de_iir & DE_PIPEA_VBLANK)
 		drm_handle_vblank(dev, 0);
@@ -976,6 +1349,17 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 	if (de_iir & DE_PIPEB_VBLANK)
 		drm_handle_vblank(dev, 1);
 
+	if (de_iir & DE_POISON)
+		DRM_ERROR("Poison interrupt\n");
+
+	if (de_iir & DE_PIPEA_FIFO_UNDERRUN)
+		if (intel_set_cpu_fifo_underrun_reporting(dev, PIPE_A, false))
+			DRM_DEBUG_DRIVER("Pipe A FIFO underrun\n");
+
+	if (de_iir & DE_PIPEB_FIFO_UNDERRUN)
+		if (intel_set_cpu_fifo_underrun_reporting(dev, PIPE_B, false))
+			DRM_DEBUG_DRIVER("Pipe B FIFO underrun\n");
+
 	if (de_iir & DE_PLANEA_FLIP_DONE) {
 		intel_prepare_page_flip(dev, 0);
 		intel_finish_page_flip_plane(dev, 0);
@@ -1002,7 +1386,7 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg)
 	if (IS_GEN5(dev) &&  de_iir & DE_PCU_EVENT)
 		ironlake_handle_rps_change(dev);
 
-	if (IS_GEN6(dev) && pm_iir & GEN6_PM_DEFERRED_EVENTS)
+	if (IS_GEN6(dev) && pm_iir & GEN6_PM_RPS_EVENTS)
 		gen6_queue_rps_work(dev_priv, pm_iir);
 
 	I915_WRITE(GTIIR, gt_iir);
@@ -1222,11 +1606,13 @@ i915_error_state_free(struct kref *error_ref)
 	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
 		i915_error_object_free(error->ring[i].batchbuffer);
 		i915_error_object_free(error->ring[i].ringbuffer);
+		i915_error_object_free(error->ring[i].ctx);
 		kfree(error->ring[i].requests);
 	}
 
 	kfree(error->active_bo);
 	kfree(error->overlay);
+	kfree(error->display);
 	kfree(error);
 }
 static void capture_bo(struct drm_i915_error_buffer *err,
@@ -1273,7 +1659,7 @@ static u32 capture_pinned_bo(struct drm_i915_error_buffer *err,
 	struct drm_i915_gem_object *obj;
 	int i = 0;
 
-	list_for_each_entry(obj, head, gtt_list) {
+	list_for_each_entry(obj, head, global_list) {
 		if (obj->pin_count == 0)
 			continue;
 
@@ -1415,7 +1801,7 @@ static void i915_gem_record_active_context(struct intel_ring_buffer *ring,
 	if (ring->id != RCS || !error->ccid)
 		return;
 
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list) {
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) {
 		if ((error->ccid & PAGE_MASK) == obj->gtt_offset) {
 			ering->ctx = i915_error_object_create_sized(dev_priv,
 								    obj, 1);
@@ -1552,7 +1938,7 @@ static void i915_capture_error_state(struct drm_device *dev)
 	list_for_each_entry(obj, &dev_priv->mm.active_list, mm_list)
 		i++;
 	error->active_bo_count = i;
-	list_for_each_entry(obj, &dev_priv->mm.bound_list, gtt_list)
+	list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list)
 		if (obj->pin_count)
 			i++;
 	error->pinned_bo_count = i - error->active_bo_count;
@@ -1932,38 +2318,28 @@ ring_last_seqno(struct intel_ring_buffer *ring)
 			  struct drm_i915_gem_request, list)->seqno;
 }
 
-static bool i915_hangcheck_ring_idle(struct intel_ring_buffer *ring, bool *err)
+static bool
+ring_idle(struct intel_ring_buffer *ring, u32 seqno)
 {
-	if (list_empty(&ring->request_list) ||
-	    i915_seqno_passed(ring->get_seqno(ring, false),
-			      ring_last_seqno(ring))) {
-		/* Issue a wake-up to catch stuck h/w. */
-		if (waitqueue_active(&ring->irq_queue)) {
-			DRM_ERROR("Hangcheck timer elapsed... %s idle\n",
-				  ring->name);
-			wake_up_all(&ring->irq_queue);
-			*err = true;
-		}
-		return true;
-	}
-	return false;
+	return (list_empty(&ring->request_list) ||
+		i915_seqno_passed(seqno, ring_last_seqno(ring)));
 }
 
-static bool semaphore_passed(struct intel_ring_buffer *ring)
+static struct intel_ring_buffer *
+semaphore_waits_for(struct intel_ring_buffer *ring, u32 *seqno)
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
-	u32 acthd = intel_ring_get_active_head(ring) & HEAD_ADDR;
-	struct intel_ring_buffer *signaller;
-	u32 cmd, ipehr, acthd_min;
+	u32 cmd, ipehr, acthd, acthd_min;
 
 	ipehr = I915_READ(RING_IPEHR(ring->mmio_base));
 	if ((ipehr & ~(0x3 << 16)) !=
 	    (MI_SEMAPHORE_MBOX | MI_SEMAPHORE_COMPARE | MI_SEMAPHORE_REGISTER))
-		return false;
+		return NULL;
 
 	/* ACTHD is likely pointing to the dword after the actual command,
 	 * so scan backwards until we find the MBOX.
 	 */
+	acthd = intel_ring_get_active_head(ring) & HEAD_ADDR;
 	acthd_min = max((int)acthd - 3 * 4, 0);
 	do {
 		cmd = ioread32(ring->virtual_start + acthd);
@@ -1972,124 +2348,216 @@ static bool semaphore_passed(struct intel_ring_buffer *ring)
 
 		acthd -= 4;
 		if (acthd < acthd_min)
-			return false;
+			return NULL;
 	} while (1);
 
-	signaller = &dev_priv->ring[(ring->id + (((ipehr >> 17) & 1) + 1)) % 3];
-	return i915_seqno_passed(signaller->get_seqno(signaller, false),
-				 ioread32(ring->virtual_start+acthd+4)+1);
+	*seqno = ioread32(ring->virtual_start+acthd+4)+1;
+	return &dev_priv->ring[(ring->id + (((ipehr >> 17) & 1) + 1)) % 3];
 }
 
-static bool kick_ring(struct intel_ring_buffer *ring)
+static int semaphore_passed(struct intel_ring_buffer *ring)
 {
-	struct drm_device *dev = ring->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 tmp = I915_READ_CTL(ring);
-	if (tmp & RING_WAIT) {
-		DRM_ERROR("Kicking stuck wait on %s\n",
-			  ring->name);
-		I915_WRITE_CTL(ring, tmp);
-		return true;
-	}
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	struct intel_ring_buffer *signaller;
+	u32 seqno, ctl;
 
-	if (INTEL_INFO(dev)->gen >= 6 &&
-	    tmp & RING_WAIT_SEMAPHORE &&
-	    semaphore_passed(ring)) {
-		DRM_ERROR("Kicking stuck semaphore on %s\n",
-			  ring->name);
-		I915_WRITE_CTL(ring, tmp);
-		return true;
-	}
-	return false;
+	ring->hangcheck.deadlock = true;
+
+	signaller = semaphore_waits_for(ring, &seqno);
+	if (signaller == NULL || signaller->hangcheck.deadlock)
+		return -1;
+
+	/* cursory check for an unkickable deadlock */
+	ctl = I915_READ_CTL(signaller);
+	if (ctl & RING_WAIT_SEMAPHORE && semaphore_passed(signaller) < 0)
+		return -1;
+
+	return i915_seqno_passed(signaller->get_seqno(signaller, false), seqno);
 }
 
-static bool i915_hangcheck_hung(struct drm_device *dev)
+static void semaphore_clear_deadlocks(struct drm_i915_private *dev_priv)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
-
-	if (dev_priv->gpu_error.hangcheck_count++ > 1) {
-		bool hung = true;
+	struct intel_ring_buffer *ring;
+	int i;
 
-		DRM_ERROR("Hangcheck timer elapsed... GPU hung\n");
-		i915_handle_error(dev, true);
+	for_each_ring(ring, dev_priv, i)
+		ring->hangcheck.deadlock = false;
+}
 
-		if (!IS_GEN2(dev)) {
-			struct intel_ring_buffer *ring;
-			int i;
+static enum intel_ring_hangcheck_action
+ring_stuck(struct intel_ring_buffer *ring, u32 acthd)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 tmp;
 
-			/* Is the chip hanging on a WAIT_FOR_EVENT?
-			 * If so we can simply poke the RB_WAIT bit
-			 * and break the hang. This should work on
-			 * all but the second generation chipsets.
-			 */
-			for_each_ring(ring, dev_priv, i)
-				hung &= !kick_ring(ring);
-		}
+	if (ring->hangcheck.acthd != acthd)
+		return active;
 
+	if (IS_GEN2(dev))
 		return hung;
+
+	/* Is the chip hanging on a WAIT_FOR_EVENT?
+	 * If so we can simply poke the RB_WAIT bit
+	 * and break the hang. This should work on
+	 * all but the second generation chipsets.
+	 */
+	tmp = I915_READ_CTL(ring);
+	if (tmp & RING_WAIT) {
+		DRM_ERROR("Kicking stuck wait on %s\n",
+			  ring->name);
+		I915_WRITE_CTL(ring, tmp);
+		return kick;
 	}
 
-	return false;
+	if (INTEL_INFO(dev)->gen >= 6 && tmp & RING_WAIT_SEMAPHORE) {
+		switch (semaphore_passed(ring)) {
+		default:
+			return hung;
+		case 1:
+			DRM_ERROR("Kicking stuck semaphore on %s\n",
+				  ring->name);
+			I915_WRITE_CTL(ring, tmp);
+			return kick;
+		case 0:
+			return wait;
+		}
+	}
+
+	return hung;
 }
 
 /**
  * This is called when the chip hasn't reported back with completed
- * batchbuffers in a long time. The first time this is called we simply record
- * ACTHD. If ACTHD hasn't changed by the time the hangcheck timer elapses
- * again, we assume the chip is wedged and try to fix it.
+ * batchbuffers in a long time. We keep track per ring seqno progress and
+ * if there are no progress, hangcheck score for that ring is increased.
+ * Further, acthd is inspected to see if the ring is stuck. On stuck case
+ * we kick the ring. If we see no progress on three subsequent calls
+ * we assume chip is wedged and try to fix it by resetting the chip.
  */
 void i915_hangcheck_elapsed(unsigned long data)
 {
 	struct drm_device *dev = (struct drm_device *)data;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	uint32_t acthd[I915_NUM_RINGS], instdone[I915_NUM_INSTDONE_REG];
 	struct intel_ring_buffer *ring;
-	bool err = false, idle;
 	int i;
+	int busy_count = 0, rings_hung = 0;
+	bool stuck[I915_NUM_RINGS] = { 0 };
+#define BUSY 1
+#define KICK 5
+#define HUNG 20
+#define FIRE 30
 
 	if (!i915_enable_hangcheck)
 		return;
 
-	memset(acthd, 0, sizeof(acthd));
-	idle = true;
 	for_each_ring(ring, dev_priv, i) {
-	    idle &= i915_hangcheck_ring_idle(ring, &err);
-	    acthd[i] = intel_ring_get_active_head(ring);
-	}
+		u32 seqno, acthd;
+		bool busy = true;
+
+		semaphore_clear_deadlocks(dev_priv);
+
+		seqno = ring->get_seqno(ring, false);
+		acthd = intel_ring_get_active_head(ring);
+
+		if (ring->hangcheck.seqno == seqno) {
+			if (ring_idle(ring, seqno)) {
+				if (waitqueue_active(&ring->irq_queue)) {
+					/* Issue a wake-up to catch stuck h/w. */
+					DRM_ERROR("Hangcheck timer elapsed... %s idle\n",
+						  ring->name);
+					wake_up_all(&ring->irq_queue);
+					ring->hangcheck.score += HUNG;
+				} else
+					busy = false;
+			} else {
+				int score;
+
+				/* We always increment the hangcheck score
+				 * if the ring is busy and still processing
+				 * the same request, so that no single request
+				 * can run indefinitely (such as a chain of
+				 * batches). The only time we do not increment
+				 * the hangcheck score on this ring, if this
+				 * ring is in a legitimate wait for another
+				 * ring. In that case the waiting ring is a
+				 * victim and we want to be sure we catch the
+				 * right culprit. Then every time we do kick
+				 * the ring, add a small increment to the
+				 * score so that we can catch a batch that is
+				 * being repeatedly kicked and so responsible
+				 * for stalling the machine.
+				 */
+				ring->hangcheck.action = ring_stuck(ring,
+								    acthd);
+
+				switch (ring->hangcheck.action) {
+				case wait:
+					score = 0;
+					break;
+				case active:
+					score = BUSY;
+					break;
+				case kick:
+					score = KICK;
+					break;
+				case hung:
+					score = HUNG;
+					stuck[i] = true;
+					break;
+				}
+				ring->hangcheck.score += score;
+			}
+		} else {
+			/* Gradually reduce the count so that we catch DoS
+			 * attempts across multiple batches.
+			 */
+			if (ring->hangcheck.score > 0)
+				ring->hangcheck.score--;
+		}
 
-	/* If all work is done then ACTHD clearly hasn't advanced. */
-	if (idle) {
-		if (err) {
-			if (i915_hangcheck_hung(dev))
-				return;
+		ring->hangcheck.seqno = seqno;
+		ring->hangcheck.acthd = acthd;
+		busy_count += busy;
+	}
 
-			goto repeat;
+	for_each_ring(ring, dev_priv, i) {
+		if (ring->hangcheck.score > FIRE) {
+			DRM_ERROR("%s on %s\n",
+				  stuck[i] ? "stuck" : "no progress",
+				  ring->name);
+			rings_hung++;
 		}
-
-		dev_priv->gpu_error.hangcheck_count = 0;
-		return;
 	}
 
-	i915_get_extra_instdone(dev, instdone);
-	if (memcmp(dev_priv->gpu_error.last_acthd, acthd,
-		   sizeof(acthd)) == 0 &&
-	    memcmp(dev_priv->gpu_error.prev_instdone, instdone,
-		   sizeof(instdone)) == 0) {
-		if (i915_hangcheck_hung(dev))
-			return;
-	} else {
-		dev_priv->gpu_error.hangcheck_count = 0;
+	if (rings_hung)
+		return i915_handle_error(dev, true);
 
-		memcpy(dev_priv->gpu_error.last_acthd, acthd,
-		       sizeof(acthd));
-		memcpy(dev_priv->gpu_error.prev_instdone, instdone,
-		       sizeof(instdone));
-	}
+	if (busy_count)
+		/* Reset timer case chip hangs without another request
+		 * being added */
+		mod_timer(&dev_priv->gpu_error.hangcheck_timer,
+			  round_jiffies_up(jiffies +
+					   DRM_I915_HANGCHECK_JIFFIES));
+}
+
+static void ibx_irq_preinstall(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (HAS_PCH_NOP(dev))
+		return;
 
-repeat:
-	/* Reset timer case chip hangs without another request being added */
-	mod_timer(&dev_priv->gpu_error.hangcheck_timer,
-		  round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES));
+	/* south display irq */
+	I915_WRITE(SDEIMR, 0xffffffff);
+	/*
+	 * SDEIER is also touched by the interrupt handler to work around missed
+	 * PCH interrupts. Hence we can't update it after the interrupt handler
+	 * is enabled - instead we unconditionally enable all PCH interrupt
+	 * sources here, but then only unmask them as needed with SDEIMR.
+	 */
+	I915_WRITE(SDEIER, 0xffffffff);
+	POSTING_READ(SDEIER);
 }
 
 /* drm_dma.h hooks
@@ -2113,19 +2581,34 @@ static void ironlake_irq_preinstall(struct drm_device *dev)
 	I915_WRITE(GTIER, 0x0);
 	POSTING_READ(GTIER);
 
-	if (HAS_PCH_NOP(dev))
-		return;
+	ibx_irq_preinstall(dev);
+}
 
-	/* south display irq */
-	I915_WRITE(SDEIMR, 0xffffffff);
-	/*
-	 * SDEIER is also touched by the interrupt handler to work around missed
-	 * PCH interrupts. Hence we can't update it after the interrupt handler
-	 * is enabled - instead we unconditionally enable all PCH interrupt
-	 * sources here, but then only unmask them as needed with SDEIMR.
-	 */
-	I915_WRITE(SDEIER, 0xffffffff);
-	POSTING_READ(SDEIER);
+static void ivybridge_irq_preinstall(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
+
+	atomic_set(&dev_priv->irq_received, 0);
+
+	I915_WRITE(HWSTAM, 0xeffe);
+
+	/* XXX hotplug from PCH */
+
+	I915_WRITE(DEIMR, 0xffffffff);
+	I915_WRITE(DEIER, 0x0);
+	POSTING_READ(DEIER);
+
+	/* and GT */
+	I915_WRITE(GTIMR, 0xffffffff);
+	I915_WRITE(GTIER, 0x0);
+	POSTING_READ(GTIER);
+
+	/* Power management */
+	I915_WRITE(GEN6_PMIMR, 0xffffffff);
+	I915_WRITE(GEN6_PMIER, 0x0);
+	POSTING_READ(GEN6_PMIER);
+
+	ibx_irq_preinstall(dev);
 }
 
 static void valleyview_irq_preinstall(struct drm_device *dev)
@@ -2201,33 +2684,41 @@ static void ibx_irq_postinstall(struct drm_device *dev)
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 	u32 mask;
 
-	if (HAS_PCH_IBX(dev))
-		mask = SDE_GMBUS | SDE_AUX_MASK;
-	else
-		mask = SDE_GMBUS_CPT | SDE_AUX_MASK_CPT;
-
 	if (HAS_PCH_NOP(dev))
 		return;
 
+	if (HAS_PCH_IBX(dev)) {
+		mask = SDE_GMBUS | SDE_AUX_MASK | SDE_TRANSB_FIFO_UNDER |
+		       SDE_TRANSA_FIFO_UNDER | SDE_POISON;
+	} else {
+		mask = SDE_GMBUS_CPT | SDE_AUX_MASK_CPT | SDE_ERROR_CPT;
+
+		I915_WRITE(SERR_INT, I915_READ(SERR_INT));
+	}
+
 	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
 	I915_WRITE(SDEIMR, ~mask);
 }
 
 static int ironlake_irq_postinstall(struct drm_device *dev)
 {
+	unsigned long irqflags;
+
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
 	/* enable kind of interrupts always enabled */
 	u32 display_mask = DE_MASTER_IRQ_CONTROL | DE_GSE | DE_PCH_EVENT |
 			   DE_PLANEA_FLIP_DONE | DE_PLANEB_FLIP_DONE |
-			   DE_AUX_CHANNEL_A;
-	u32 render_irqs;
+			   DE_AUX_CHANNEL_A | DE_PIPEB_FIFO_UNDERRUN |
+			   DE_PIPEA_FIFO_UNDERRUN | DE_POISON;
+	u32 gt_irqs;
 
 	dev_priv->irq_mask = ~display_mask;
 
 	/* should always can generate irq */
 	I915_WRITE(DEIIR, I915_READ(DEIIR));
 	I915_WRITE(DEIMR, dev_priv->irq_mask);
-	I915_WRITE(DEIER, display_mask | DE_PIPEA_VBLANK | DE_PIPEB_VBLANK);
+	I915_WRITE(DEIER, display_mask |
+			  DE_PIPEA_VBLANK | DE_PIPEB_VBLANK | DE_PCU_EVENT);
 	POSTING_READ(DEIER);
 
 	dev_priv->gt_irq_mask = ~0;
@@ -2235,26 +2726,28 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(GTIIR, I915_READ(GTIIR));
 	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 
+	gt_irqs = GT_RENDER_USER_INTERRUPT;
+
 	if (IS_GEN6(dev))
-		render_irqs =
-			GT_USER_INTERRUPT |
-			GEN6_BSD_USER_INTERRUPT |
-			GEN6_BLITTER_USER_INTERRUPT;
+		gt_irqs |= GT_BLT_USER_INTERRUPT | GT_BSD_USER_INTERRUPT;
 	else
-		render_irqs =
-			GT_USER_INTERRUPT |
-			GT_PIPE_NOTIFY |
-			GT_BSD_USER_INTERRUPT;
-	I915_WRITE(GTIER, render_irqs);
+		gt_irqs |= GT_RENDER_PIPECTL_NOTIFY_INTERRUPT |
+			   ILK_BSD_USER_INTERRUPT;
+
+	I915_WRITE(GTIER, gt_irqs);
 	POSTING_READ(GTIER);
 
 	ibx_irq_postinstall(dev);
 
 	if (IS_IRONLAKE_M(dev)) {
-		/* Clear & enable PCU event interrupts */
-		I915_WRITE(DEIIR, DE_PCU_EVENT);
-		I915_WRITE(DEIER, I915_READ(DEIER) | DE_PCU_EVENT);
+		/* Enable PCU event interrupts
+		 *
+		 * spinlocking not required here for correctness since interrupt
+		 * setup is guaranteed to run in single-threaded context. But we
+		 * need it to make the assert_spin_locked happy. */
+		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
 		ironlake_enable_display_irq(dev_priv, DE_PCU_EVENT);
+		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 	}
 
 	return 0;
@@ -2269,12 +2762,15 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 		DE_PLANEC_FLIP_DONE_IVB |
 		DE_PLANEB_FLIP_DONE_IVB |
 		DE_PLANEA_FLIP_DONE_IVB |
-		DE_AUX_CHANNEL_A_IVB;
-	u32 render_irqs;
+		DE_AUX_CHANNEL_A_IVB |
+		DE_ERR_INT_IVB;
+	u32 pm_irqs = GEN6_PM_RPS_EVENTS;
+	u32 gt_irqs;
 
 	dev_priv->irq_mask = ~display_mask;
 
 	/* should always can generate irq */
+	I915_WRITE(GEN7_ERR_INT, I915_READ(GEN7_ERR_INT));
 	I915_WRITE(DEIIR, I915_READ(DEIIR));
 	I915_WRITE(DEIMR, dev_priv->irq_mask);
 	I915_WRITE(DEIER,
@@ -2284,16 +2780,32 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 		   DE_PIPEA_VBLANK_IVB);
 	POSTING_READ(DEIER);
 
-	dev_priv->gt_irq_mask = ~GT_GEN7_L3_PARITY_ERROR_INTERRUPT;
+	dev_priv->gt_irq_mask = ~GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
 
 	I915_WRITE(GTIIR, I915_READ(GTIIR));
 	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 
-	render_irqs = GT_USER_INTERRUPT | GEN6_BSD_USER_INTERRUPT |
-		GEN6_BLITTER_USER_INTERRUPT | GT_GEN7_L3_PARITY_ERROR_INTERRUPT;
-	I915_WRITE(GTIER, render_irqs);
+	gt_irqs = GT_RENDER_USER_INTERRUPT | GT_BSD_USER_INTERRUPT |
+		  GT_BLT_USER_INTERRUPT | GT_RENDER_L3_PARITY_ERROR_INTERRUPT;
+	I915_WRITE(GTIER, gt_irqs);
 	POSTING_READ(GTIER);
 
+	I915_WRITE(GEN6_PMIIR, I915_READ(GEN6_PMIIR));
+	if (HAS_VEBOX(dev))
+		pm_irqs |= PM_VEBOX_USER_INTERRUPT |
+			PM_VEBOX_CS_ERROR_INTERRUPT;
+
+	/* Our enable/disable rps functions may touch these registers so
+	 * make sure to set a known state for only the non-RPS bits.
+	 * The RMW is extra paranoia since this should be called after being set
+	 * to a known state in preinstall.
+	 * */
+	I915_WRITE(GEN6_PMIMR,
+		   (I915_READ(GEN6_PMIMR) | ~GEN6_PM_RPS_EVENTS) & ~pm_irqs);
+	I915_WRITE(GEN6_PMIER,
+		   (I915_READ(GEN6_PMIER) & GEN6_PM_RPS_EVENTS) | pm_irqs);
+	POSTING_READ(GEN6_PMIER);
+
 	ibx_irq_postinstall(dev);
 
 	return 0;
@@ -2302,10 +2814,9 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 static int valleyview_irq_postinstall(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
+	u32 gt_irqs;
 	u32 enable_mask;
 	u32 pipestat_enable = PLANE_FLIP_DONE_INT_EN_VLV;
-	u32 render_irqs;
-	u16 msid;
 
 	enable_mask = I915_DISPLAY_PORT_INTERRUPT;
 	enable_mask |= I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
@@ -2321,13 +2832,6 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 		I915_DISPLAY_PIPE_A_VBLANK_INTERRUPT |
 		I915_DISPLAY_PIPE_B_VBLANK_INTERRUPT;
 
-	/* Hack for broken MSIs on VLV */
-	pci_write_config_dword(dev_priv->dev->pdev, 0x94, 0xfee00000);
-	pci_read_config_word(dev->pdev, 0x98, &msid);
-	msid &= 0xff; /* mask out delivery bits */
-	msid |= (1<<14);
-	pci_write_config_word(dev_priv->dev->pdev, 0x98, msid);
-
 	I915_WRITE(PORT_HOTPLUG_EN, 0);
 	POSTING_READ(PORT_HOTPLUG_EN);
 
@@ -2348,9 +2852,9 @@ static int valleyview_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(GTIIR, I915_READ(GTIIR));
 	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 
-	render_irqs = GT_USER_INTERRUPT | GEN6_BSD_USER_INTERRUPT |
-		GEN6_BLITTER_USER_INTERRUPT;
-	I915_WRITE(GTIER, render_irqs);
+	gt_irqs = GT_RENDER_USER_INTERRUPT | GT_BSD_USER_INTERRUPT |
+		GT_BLT_USER_INTERRUPT;
+	I915_WRITE(GTIER, gt_irqs);
 	POSTING_READ(GTIER);
 
 	/* ack & enable invalid PTE error interrupts */
@@ -2402,6 +2906,8 @@ static void ironlake_irq_uninstall(struct drm_device *dev)
 	I915_WRITE(DEIMR, 0xffffffff);
 	I915_WRITE(DEIER, 0x0);
 	I915_WRITE(DEIIR, I915_READ(DEIIR));
+	if (IS_GEN7(dev))
+		I915_WRITE(GEN7_ERR_INT, I915_READ(GEN7_ERR_INT));
 
 	I915_WRITE(GTIMR, 0xffffffff);
 	I915_WRITE(GTIER, 0x0);
@@ -2413,6 +2919,8 @@ static void ironlake_irq_uninstall(struct drm_device *dev)
 	I915_WRITE(SDEIMR, 0xffffffff);
 	I915_WRITE(SDEIER, 0x0);
 	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
+	if (HAS_PCH_CPT(dev) || HAS_PCH_LPT(dev))
+		I915_WRITE(SERR_INT, I915_READ(SERR_INT));
 }
 
 static void i8xx_irq_preinstall(struct drm_device * dev)
@@ -2626,7 +3134,7 @@ static int i915_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(IER, enable_mask);
 	POSTING_READ(IER);
 
-	intel_opregion_enable_asle(dev);
+	i915_enable_asle_pipestat(dev);
 
 	return 0;
 }
@@ -2715,12 +3223,9 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 
 			DRM_DEBUG_DRIVER("hotplug event received, stat 0x%08x\n",
 				  hotplug_status);
-			if (hotplug_trigger) {
-				if (hotplug_irq_storm_detect(dev, hotplug_trigger, hpd_status_i915))
-					i915_hpd_irq_setup(dev);
-				queue_work(dev_priv->wq,
-					   &dev_priv->hotplug_work);
-			}
+
+			intel_hpd_irq_handler(dev, hotplug_trigger, hpd_status_i915);
+
 			I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
 			POSTING_READ(PORT_HOTPLUG_STAT);
 		}
@@ -2860,7 +3365,7 @@ static int i965_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(PORT_HOTPLUG_EN, 0);
 	POSTING_READ(PORT_HOTPLUG_EN);
 
-	intel_opregion_enable_asle(dev);
+	i915_enable_asle_pipestat(dev);
 
 	return 0;
 }
@@ -2872,6 +3377,8 @@ static void i915_hpd_irq_setup(struct drm_device *dev)
 	struct intel_encoder *intel_encoder;
 	u32 hotplug_en;
 
+	assert_spin_locked(&dev_priv->irq_lock);
+
 	if (I915_HAS_HOTPLUG(dev)) {
 		hotplug_en = I915_READ(PORT_HOTPLUG_EN);
 		hotplug_en &= ~HOTPLUG_INT_EN_MASK;
@@ -2952,17 +3459,14 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 			u32 hotplug_status = I915_READ(PORT_HOTPLUG_STAT);
 			u32 hotplug_trigger = hotplug_status & (IS_G4X(dev) ?
 								  HOTPLUG_INT_STATUS_G4X :
-								  HOTPLUG_INT_STATUS_I965);
+								  HOTPLUG_INT_STATUS_I915);
 
 			DRM_DEBUG_DRIVER("hotplug event received, stat 0x%08x\n",
 				  hotplug_status);
-			if (hotplug_trigger) {
-				if (hotplug_irq_storm_detect(dev, hotplug_trigger,
-							    IS_G4X(dev) ? hpd_status_gen4 : hpd_status_i965))
-					i915_hpd_irq_setup(dev);
-				queue_work(dev_priv->wq,
-					   &dev_priv->hotplug_work);
-			}
+
+			intel_hpd_irq_handler(dev, hotplug_trigger,
+					      IS_G4X(dev) ? hpd_status_gen4 : hpd_status_i915);
+
 			I915_WRITE(PORT_HOTPLUG_STAT, hotplug_status);
 			I915_READ(PORT_HOTPLUG_STAT);
 		}
@@ -3113,9 +3617,9 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->disable_vblank = valleyview_disable_vblank;
 		dev_priv->display.hpd_irq_setup = i915_hpd_irq_setup;
 	} else if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev)) {
-		/* Share pre & uninstall handlers with ILK/SNB */
+		/* Share uninstall handlers with ILK/SNB */
 		dev->driver->irq_handler = ivybridge_irq_handler;
-		dev->driver->irq_preinstall = ironlake_irq_preinstall;
+		dev->driver->irq_preinstall = ivybridge_irq_preinstall;
 		dev->driver->irq_postinstall = ivybridge_irq_postinstall;
 		dev->driver->irq_uninstall = ironlake_irq_uninstall;
 		dev->driver->enable_vblank = ivybridge_enable_vblank;
@@ -3158,6 +3662,7 @@ void intel_hpd_init(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_mode_config *mode_config = &dev->mode_config;
 	struct drm_connector *connector;
+	unsigned long irqflags;
 	int i;
 
 	for (i = 1; i < HPD_NUM_PINS; i++) {
@@ -3170,6 +3675,11 @@ void intel_hpd_init(struct drm_device *dev)
 		if (!connector->polled && I915_HAS_HOTPLUG(dev) && intel_connector->encoder->hpd_pin > HPD_NONE)
 			connector->polled = DRM_CONNECTOR_POLL_HPD;
 	}
+
+	/* Interrupt setup is already guaranteed to be single-threaded, this is
+	 * just to make the assert_spin_locked checks happy. */
+	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
 	if (dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev);
+	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 }
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 2d6b62e42daf..f2326fc60ac9 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -147,15 +147,9 @@
 #define   VGA_MSR_MEM_EN (1<<1)
 #define   VGA_MSR_CGA_MODE (1<<0)
 
-/*
- * SR01 is the only VGA register touched on non-UMS setups.
- * VLV doesn't do UMS, so the sequencer index/data registers
- * are the only VGA registers which need to include
- * display_mmio_offset.
- */
-#define VGA_SR_INDEX (dev_priv->info->display_mmio_offset + 0x3c4)
+#define VGA_SR_INDEX 0x3c4
 #define SR01			1
-#define VGA_SR_DATA (dev_priv->info->display_mmio_offset + 0x3c5)
+#define VGA_SR_DATA 0x3c5
 
 #define VGA_AR_INDEX 0x3c0
 #define   VGA_AR_VID_EN (1<<5)
@@ -265,13 +259,19 @@
 #define  MI_SEMAPHORE_UPDATE	    (1<<21)
 #define  MI_SEMAPHORE_COMPARE	    (1<<20)
 #define  MI_SEMAPHORE_REGISTER	    (1<<18)
-#define  MI_SEMAPHORE_SYNC_RV	    (2<<16)
-#define  MI_SEMAPHORE_SYNC_RB	    (0<<16)
-#define  MI_SEMAPHORE_SYNC_VR	    (0<<16)
-#define  MI_SEMAPHORE_SYNC_VB	    (2<<16)
-#define  MI_SEMAPHORE_SYNC_BR	    (2<<16)
-#define  MI_SEMAPHORE_SYNC_BV	    (0<<16)
-#define  MI_SEMAPHORE_SYNC_INVALID  (1<<0)
+#define  MI_SEMAPHORE_SYNC_VR	    (0<<16) /* RCS  wait for VCS  (RVSYNC) */
+#define  MI_SEMAPHORE_SYNC_VER	    (1<<16) /* RCS  wait for VECS (RVESYNC) */
+#define  MI_SEMAPHORE_SYNC_BR	    (2<<16) /* RCS  wait for BCS  (RBSYNC) */
+#define  MI_SEMAPHORE_SYNC_BV	    (0<<16) /* VCS  wait for BCS  (VBSYNC) */
+#define  MI_SEMAPHORE_SYNC_VEV	    (1<<16) /* VCS  wait for VECS (VVESYNC) */
+#define  MI_SEMAPHORE_SYNC_RV	    (2<<16) /* VCS  wait for RCS  (VRSYNC) */
+#define  MI_SEMAPHORE_SYNC_RB	    (0<<16) /* BCS  wait for RCS  (BRSYNC) */
+#define  MI_SEMAPHORE_SYNC_VEB	    (1<<16) /* BCS  wait for VECS (BVESYNC) */
+#define  MI_SEMAPHORE_SYNC_VB	    (2<<16) /* BCS  wait for VCS  (BVSYNC) */
+#define  MI_SEMAPHORE_SYNC_BVE	    (0<<16) /* VECS wait for BCS  (VEBSYNC) */
+#define  MI_SEMAPHORE_SYNC_VVE	    (1<<16) /* VECS wait for VCS  (VEVSYNC) */
+#define  MI_SEMAPHORE_SYNC_RVE	    (2<<16) /* VECS wait for RCS  (VERSYNC) */
+#define  MI_SEMAPHORE_SYNC_INVALID  (3<<16)
 /*
  * 3D instructions used by the kernel
  */
@@ -342,33 +342,74 @@
 #define  DEBUG_RESET_DISPLAY		(1<<9)
 
 /*
- * DPIO - a special bus for various display related registers to hide behind:
- *  0x800c: m1, m2, n, p1, p2, k dividers
- *  0x8014: REF and SFR select
- *  0x8014: N divider, VCO select
- *  0x801c/3c: core clock bits
- *  0x8048/68: low pass filter coefficients
- *  0x8100: fast clock controls
+ * IOSF sideband
+ */
+#define VLV_IOSF_DOORBELL_REQ			(VLV_DISPLAY_BASE + 0x2100)
+#define   IOSF_DEVFN_SHIFT			24
+#define   IOSF_OPCODE_SHIFT			16
+#define   IOSF_PORT_SHIFT			8
+#define   IOSF_BYTE_ENABLES_SHIFT		4
+#define   IOSF_BAR_SHIFT			1
+#define   IOSF_SB_BUSY				(1<<0)
+#define   IOSF_PORT_PUNIT			0x4
+#define   IOSF_PORT_NC				0x11
+#define   IOSF_PORT_DPIO			0x12
+#define VLV_IOSF_DATA				(VLV_DISPLAY_BASE + 0x2104)
+#define VLV_IOSF_ADDR				(VLV_DISPLAY_BASE + 0x2108)
+
+#define PUNIT_OPCODE_REG_READ			6
+#define PUNIT_OPCODE_REG_WRITE			7
+
+#define PUNIT_REG_GPU_LFM			0xd3
+#define PUNIT_REG_GPU_FREQ_REQ			0xd4
+#define PUNIT_REG_GPU_FREQ_STS			0xd8
+#define PUNIT_REG_MEDIA_TURBO_FREQ_REQ		0xdc
+
+#define PUNIT_FUSE_BUS2				0xf6 /* bits 47:40 */
+#define PUNIT_FUSE_BUS1				0xf5 /* bits 55:48 */
+
+#define IOSF_NC_FB_GFX_FREQ_FUSE		0x1c
+#define   FB_GFX_MAX_FREQ_FUSE_SHIFT		3
+#define   FB_GFX_MAX_FREQ_FUSE_MASK		0x000007f8
+#define   FB_GFX_FGUARANTEED_FREQ_FUSE_SHIFT	11
+#define   FB_GFX_FGUARANTEED_FREQ_FUSE_MASK	0x0007f800
+#define IOSF_NC_FB_GFX_FMAX_FUSE_HI		0x34
+#define   FB_FMAX_VMIN_FREQ_HI_MASK		0x00000007
+#define IOSF_NC_FB_GFX_FMAX_FUSE_LO		0x30
+#define   FB_FMAX_VMIN_FREQ_LO_SHIFT		27
+#define   FB_FMAX_VMIN_FREQ_LO_MASK		0xf8000000
+
+/*
+ * DPIO - a special bus for various display related registers to hide behind
  *
  * DPIO is VLV only.
+ *
+ * Note: digital port B is DDI0, digital pot C is DDI1
  */
-#define DPIO_PKT			(VLV_DISPLAY_BASE + 0x2100)
-#define  DPIO_RID			(0<<24)
-#define  DPIO_OP_WRITE			(1<<16)
-#define  DPIO_OP_READ			(0<<16)
-#define  DPIO_PORTID			(0x12<<8)
-#define  DPIO_BYTE			(0xf<<4)
-#define  DPIO_BUSY			(1<<0) /* status only */
-#define DPIO_DATA			(VLV_DISPLAY_BASE + 0x2104)
-#define DPIO_REG			(VLV_DISPLAY_BASE + 0x2108)
+#define DPIO_DEVFN			0
+#define DPIO_OPCODE_REG_WRITE		1
+#define DPIO_OPCODE_REG_READ		0
+
 #define DPIO_CTL			(VLV_DISPLAY_BASE + 0x2110)
 #define  DPIO_MODSEL1			(1<<3) /* if ref clk b == 27 */
 #define  DPIO_MODSEL0			(1<<2) /* if ref clk a == 27 */
 #define  DPIO_SFR_BYPASS		(1<<1)
 #define  DPIO_RESET			(1<<0)
 
+#define _DPIO_TX3_SWING_CTL4_A		0x690
+#define _DPIO_TX3_SWING_CTL4_B		0x2a90
+#define DPIO_TX3_SWING_CTL4(pipe) _PIPE(pipe, _DPIO_TX_SWING_CTL4_A, \
+					_DPIO_TX3_SWING_CTL4_B)
+
+/*
+ * Per pipe/PLL DPIO regs
+ */
 #define _DPIO_DIV_A			0x800c
 #define   DPIO_POST_DIV_SHIFT		(28) /* 3 bits */
+#define   DPIO_POST_DIV_DAC		0
+#define   DPIO_POST_DIV_HDMIDP		1 /* DAC 225-400M rate */
+#define   DPIO_POST_DIV_LVDS1		2
+#define   DPIO_POST_DIV_LVDS2		3
 #define   DPIO_K_SHIFT			(24) /* 4 bits */
 #define   DPIO_P1_SHIFT			(21) /* 3 bits */
 #define   DPIO_P2_SHIFT			(16) /* 5 bits */
@@ -394,14 +435,111 @@
 #define _DPIO_CORE_CLK_B		0x803c
 #define DPIO_CORE_CLK(pipe) _PIPE(pipe, _DPIO_CORE_CLK_A, _DPIO_CORE_CLK_B)
 
-#define _DPIO_LFP_COEFF_A		0x8048
-#define _DPIO_LFP_COEFF_B		0x8068
-#define DPIO_LFP_COEFF(pipe) _PIPE(pipe, _DPIO_LFP_COEFF_A, _DPIO_LFP_COEFF_B)
+#define _DPIO_IREF_CTL_A		0x8040
+#define _DPIO_IREF_CTL_B		0x8060
+#define DPIO_IREF_CTL(pipe) _PIPE(pipe, _DPIO_IREF_CTL_A, _DPIO_IREF_CTL_B)
+
+#define DPIO_IREF_BCAST			0xc044
+#define _DPIO_IREF_A			0x8044
+#define _DPIO_IREF_B			0x8064
+#define DPIO_IREF(pipe) _PIPE(pipe, _DPIO_IREF_A, _DPIO_IREF_B)
+
+#define _DPIO_PLL_CML_A			0x804c
+#define _DPIO_PLL_CML_B			0x806c
+#define DPIO_PLL_CML(pipe) _PIPE(pipe, _DPIO_PLL_CML_A, _DPIO_PLL_CML_B)
+
+#define _DPIO_LPF_COEFF_A		0x8048
+#define _DPIO_LPF_COEFF_B		0x8068
+#define DPIO_LPF_COEFF(pipe) _PIPE(pipe, _DPIO_LPF_COEFF_A, _DPIO_LPF_COEFF_B)
+
+#define DPIO_CALIBRATION		0x80ac
 
 #define DPIO_FASTCLK_DISABLE		0x8100
 
-#define DPIO_DATA_CHANNEL1		0x8220
-#define DPIO_DATA_CHANNEL2		0x8420
+/*
+ * Per DDI channel DPIO regs
+ */
+
+#define _DPIO_PCS_TX_0			0x8200
+#define _DPIO_PCS_TX_1			0x8400
+#define   DPIO_PCS_TX_LANE2_RESET	(1<<16)
+#define   DPIO_PCS_TX_LANE1_RESET	(1<<7)
+#define DPIO_PCS_TX(port) _PORT(port, _DPIO_PCS_TX_0, _DPIO_PCS_TX_1)
+
+#define _DPIO_PCS_CLK_0			0x8204
+#define _DPIO_PCS_CLK_1			0x8404
+#define   DPIO_PCS_CLK_CRI_RXEB_EIOS_EN	(1<<22)
+#define   DPIO_PCS_CLK_CRI_RXDIGFILTSG_EN (1<<21)
+#define   DPIO_PCS_CLK_DATAWIDTH_SHIFT	(6)
+#define   DPIO_PCS_CLK_SOFT_RESET	(1<<5)
+#define DPIO_PCS_CLK(port) _PORT(port, _DPIO_PCS_CLK_0, _DPIO_PCS_CLK_1)
+
+#define _DPIO_PCS_CTL_OVR1_A		0x8224
+#define _DPIO_PCS_CTL_OVR1_B		0x8424
+#define DPIO_PCS_CTL_OVER1(port) _PORT(port, _DPIO_PCS_CTL_OVR1_A, \
+				       _DPIO_PCS_CTL_OVR1_B)
+
+#define _DPIO_PCS_STAGGER0_A		0x822c
+#define _DPIO_PCS_STAGGER0_B		0x842c
+#define DPIO_PCS_STAGGER0(port) _PORT(port, _DPIO_PCS_STAGGER0_A, \
+				      _DPIO_PCS_STAGGER0_B)
+
+#define _DPIO_PCS_STAGGER1_A		0x8230
+#define _DPIO_PCS_STAGGER1_B		0x8430
+#define DPIO_PCS_STAGGER1(port) _PORT(port, _DPIO_PCS_STAGGER1_A, \
+				      _DPIO_PCS_STAGGER1_B)
+
+#define _DPIO_PCS_CLOCKBUF0_A		0x8238
+#define _DPIO_PCS_CLOCKBUF0_B		0x8438
+#define DPIO_PCS_CLOCKBUF0(port) _PORT(port, _DPIO_PCS_CLOCKBUF0_A, \
+				       _DPIO_PCS_CLOCKBUF0_B)
+
+#define _DPIO_PCS_CLOCKBUF8_A		0x825c
+#define _DPIO_PCS_CLOCKBUF8_B		0x845c
+#define DPIO_PCS_CLOCKBUF8(port) _PORT(port, _DPIO_PCS_CLOCKBUF8_A, \
+				       _DPIO_PCS_CLOCKBUF8_B)
+
+#define _DPIO_TX_SWING_CTL2_A		0x8288
+#define _DPIO_TX_SWING_CTL2_B		0x8488
+#define DPIO_TX_SWING_CTL2(port) _PORT(port, _DPIO_TX_SWING_CTL2_A, \
+				       _DPIO_TX_SWING_CTL2_B)
+
+#define _DPIO_TX_SWING_CTL3_A		0x828c
+#define _DPIO_TX_SWING_CTL3_B		0x848c
+#define DPIO_TX_SWING_CTL3(port) _PORT(port, _DPIO_TX_SWING_CTL3_A, \
+				       _DPIO_TX_SWING_CTL3_B)
+
+#define _DPIO_TX_SWING_CTL4_A		0x8290
+#define _DPIO_TX_SWING_CTL4_B		0x8490
+#define DPIO_TX_SWING_CTL4(port) _PORT(port, _DPIO_TX_SWING_CTL4_A, \
+				       _DPIO_TX_SWING_CTL4_B)
+
+#define _DPIO_TX_OCALINIT_0		0x8294
+#define _DPIO_TX_OCALINIT_1		0x8494
+#define   DPIO_TX_OCALINIT_EN		(1<<31)
+#define DPIO_TX_OCALINIT(port) _PORT(port, _DPIO_TX_OCALINIT_0, \
+				     _DPIO_TX_OCALINIT_1)
+
+#define _DPIO_TX_CTL_0			0x82ac
+#define _DPIO_TX_CTL_1			0x84ac
+#define DPIO_TX_CTL(port) _PORT(port, _DPIO_TX_CTL_0, _DPIO_TX_CTL_1)
+
+#define _DPIO_TX_LANE_0			0x82b8
+#define _DPIO_TX_LANE_1			0x84b8
+#define DPIO_TX_LANE(port) _PORT(port, _DPIO_TX_LANE_0, _DPIO_TX_LANE_1)
+
+#define _DPIO_DATA_CHANNEL1		0x8220
+#define _DPIO_DATA_CHANNEL2		0x8420
+#define DPIO_DATA_CHANNEL(port) _PORT(port, _DPIO_DATA_CHANNEL1, _DPIO_DATA_CHANNEL2)
+
+#define _DPIO_PORT0_PCS0		0x0220
+#define _DPIO_PORT0_PCS1		0x0420
+#define _DPIO_PORT1_PCS2		0x2620
+#define _DPIO_PORT1_PCS3		0x2820
+#define DPIO_DATA_LANE_A(port) _PORT(port, _DPIO_PORT0_PCS0, _DPIO_PORT1_PCS2)
+#define DPIO_DATA_LANE_B(port) _PORT(port, _DPIO_PORT0_PCS1, _DPIO_PORT1_PCS3)
+#define DPIO_DATA_CHANNEL1              0x8220
+#define DPIO_DATA_CHANNEL2              0x8420
 
 /*
  * Fence registers
@@ -443,6 +581,7 @@
 #define RENDER_RING_BASE	0x02000
 #define BSD_RING_BASE		0x04000
 #define GEN6_BSD_RING_BASE	0x12000
+#define VEBOX_RING_BASE		0x1a000
 #define BLT_RING_BASE		0x22000
 #define RING_TAIL(base)		((base)+0x30)
 #define RING_HEAD(base)		((base)+0x34)
@@ -450,12 +589,20 @@
 #define RING_CTL(base)		((base)+0x3c)
 #define RING_SYNC_0(base)	((base)+0x40)
 #define RING_SYNC_1(base)	((base)+0x44)
-#define GEN6_RVSYNC (RING_SYNC_0(RENDER_RING_BASE))
-#define GEN6_RBSYNC (RING_SYNC_1(RENDER_RING_BASE))
-#define GEN6_VRSYNC (RING_SYNC_1(GEN6_BSD_RING_BASE))
-#define GEN6_VBSYNC (RING_SYNC_0(GEN6_BSD_RING_BASE))
-#define GEN6_BRSYNC (RING_SYNC_0(BLT_RING_BASE))
-#define GEN6_BVSYNC (RING_SYNC_1(BLT_RING_BASE))
+#define RING_SYNC_2(base)	((base)+0x48)
+#define GEN6_RVSYNC	(RING_SYNC_0(RENDER_RING_BASE))
+#define GEN6_RBSYNC	(RING_SYNC_1(RENDER_RING_BASE))
+#define GEN6_RVESYNC	(RING_SYNC_2(RENDER_RING_BASE))
+#define GEN6_VBSYNC	(RING_SYNC_0(GEN6_BSD_RING_BASE))
+#define GEN6_VRSYNC	(RING_SYNC_1(GEN6_BSD_RING_BASE))
+#define GEN6_VVESYNC	(RING_SYNC_2(GEN6_BSD_RING_BASE))
+#define GEN6_BRSYNC	(RING_SYNC_0(BLT_RING_BASE))
+#define GEN6_BVSYNC	(RING_SYNC_1(BLT_RING_BASE))
+#define GEN6_BVESYNC	(RING_SYNC_2(BLT_RING_BASE))
+#define GEN6_VEBSYNC	(RING_SYNC_0(VEBOX_RING_BASE))
+#define GEN6_VERSYNC	(RING_SYNC_1(VEBOX_RING_BASE))
+#define GEN6_VEVSYNC	(RING_SYNC_2(VEBOX_RING_BASE))
+#define GEN6_NOSYNC 0
 #define RING_MAX_IDLE(base)	((base)+0x54)
 #define RING_HWS_PGA(base)	((base)+0x80)
 #define RING_HWS_PGA_GEN6(base)	((base)+0x2080)
@@ -467,6 +614,7 @@
 #define DONE_REG		0x40b0
 #define BSD_HWS_PGA_GEN7	(0x04180)
 #define BLT_HWS_PGA_GEN7	(0x04280)
+#define VEBOX_HWS_PGA_GEN7	(0x04380)
 #define RING_ACTHD(base)	((base)+0x74)
 #define RING_NOPID(base)	((base)+0x94)
 #define RING_IMR(base)		((base)+0xa8)
@@ -527,7 +675,11 @@
 
 #define ERROR_GEN6	0x040a0
 #define GEN7_ERR_INT	0x44040
-#define   ERR_INT_MMIO_UNCLAIMED (1<<13)
+#define   ERR_INT_POISON		(1<<31)
+#define   ERR_INT_MMIO_UNCLAIMED	(1<<13)
+#define   ERR_INT_FIFO_UNDERRUN_C	(1<<6)
+#define   ERR_INT_FIFO_UNDERRUN_B	(1<<3)
+#define   ERR_INT_FIFO_UNDERRUN_A	(1<<0)
 
 #define FPGA_DBG		0x42300
 #define   FPGA_DBG_RM_NOCLAIM	(1<<31)
@@ -583,24 +735,7 @@
 #define VLV_IIR		(VLV_DISPLAY_BASE + 0x20a4)
 #define VLV_IMR		(VLV_DISPLAY_BASE + 0x20a8)
 #define VLV_ISR		(VLV_DISPLAY_BASE + 0x20ac)
-#define   I915_PIPE_CONTROL_NOTIFY_INTERRUPT		(1<<18)
-#define   I915_DISPLAY_PORT_INTERRUPT			(1<<17)
-#define   I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT	(1<<15)
-#define   I915_GMCH_THERMAL_SENSOR_EVENT_INTERRUPT	(1<<14) /* p-state */
-#define   I915_HWB_OOM_INTERRUPT			(1<<13)
-#define   I915_SYNC_STATUS_INTERRUPT			(1<<12)
-#define   I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT	(1<<11)
-#define   I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT	(1<<10)
-#define   I915_OVERLAY_PLANE_FLIP_PENDING_INTERRUPT	(1<<9)
-#define   I915_DISPLAY_PLANE_C_FLIP_PENDING_INTERRUPT	(1<<8)
-#define   I915_DISPLAY_PIPE_A_VBLANK_INTERRUPT		(1<<7)
-#define   I915_DISPLAY_PIPE_A_EVENT_INTERRUPT		(1<<6)
-#define   I915_DISPLAY_PIPE_B_VBLANK_INTERRUPT		(1<<5)
-#define   I915_DISPLAY_PIPE_B_EVENT_INTERRUPT		(1<<4)
-#define   I915_DEBUG_INTERRUPT				(1<<2)
-#define   I915_USER_INTERRUPT				(1<<1)
-#define   I915_ASLE_INTERRUPT				(1<<0)
-#define   I915_BSD_USER_INTERRUPT                      (1<<25)
+#define VLV_PCBR	(VLV_DISPLAY_BASE + 0x2120)
 #define   DISPLAY_PLANE_FLIP_PENDING(plane) (1<<(11-(plane))) /* A and B only */
 #define EIR		0x020b0
 #define EMR		0x020b4
@@ -712,28 +847,6 @@
 #define CACHE_MODE_1		0x7004 /* IVB+ */
 #define   PIXEL_SUBSPAN_COLLECT_OPT_DISABLE (1<<6)
 
-/* GEN6 interrupt control
- * Note that the per-ring interrupt bits do alias with the global interrupt bits
- * in GTIMR. */
-#define GEN6_RENDER_HWSTAM	0x2098
-#define GEN6_RENDER_IMR		0x20a8
-#define   GEN6_RENDER_CONTEXT_SWITCH_INTERRUPT		(1 << 8)
-#define   GEN6_RENDER_PPGTT_PAGE_FAULT			(1 << 7)
-#define   GEN6_RENDER_TIMEOUT_COUNTER_EXPIRED		(1 << 6)
-#define   GEN6_RENDER_L3_PARITY_ERROR			(1 << 5)
-#define   GEN6_RENDER_PIPE_CONTROL_NOTIFY_INTERRUPT	(1 << 4)
-#define   GEN6_RENDER_COMMAND_PARSER_MASTER_ERROR	(1 << 3)
-#define   GEN6_RENDER_SYNC_STATUS			(1 << 2)
-#define   GEN6_RENDER_DEBUG_INTERRUPT			(1 << 1)
-#define   GEN6_RENDER_USER_INTERRUPT			(1 << 0)
-
-#define GEN6_BLITTER_HWSTAM	0x22098
-#define GEN6_BLITTER_IMR	0x220a8
-#define   GEN6_BLITTER_MI_FLUSH_DW_NOTIFY_INTERRUPT	(1 << 26)
-#define   GEN6_BLITTER_COMMAND_PARSER_MASTER_ERROR	(1 << 25)
-#define   GEN6_BLITTER_SYNC_STATUS			(1 << 24)
-#define   GEN6_BLITTER_USER_INTERRUPT			(1 << 22)
-
 #define GEN6_BLITTER_ECOSKPD	0x221d0
 #define   GEN6_BLITTER_LOCK_SHIFT			16
 #define   GEN6_BLITTER_FBC_NOTIFY			(1<<3)
@@ -744,9 +857,52 @@
 #define   GEN6_BSD_SLEEP_INDICATOR	(1 << 3)
 #define   GEN6_BSD_GO_INDICATOR		(1 << 4)
 
-#define GEN6_BSD_HWSTAM			0x12098
-#define GEN6_BSD_IMR			0x120a8
-#define   GEN6_BSD_USER_INTERRUPT	(1 << 12)
+/* On modern GEN architectures interrupt control consists of two sets
+ * of registers. The first set pertains to the ring generating the
+ * interrupt. The second control is for the functional block generating the
+ * interrupt. These are PM, GT, DE, etc.
+ *
+ * Luckily *knocks on wood* all the ring interrupt bits match up with the
+ * GT interrupt bits, so we don't need to duplicate the defines.
+ *
+ * These defines should cover us well from SNB->HSW with minor exceptions
+ * it can also work on ILK.
+ */
+#define GT_BLT_FLUSHDW_NOTIFY_INTERRUPT		(1 << 26)
+#define GT_BLT_CS_ERROR_INTERRUPT		(1 << 25)
+#define GT_BLT_USER_INTERRUPT			(1 << 22)
+#define GT_BSD_CS_ERROR_INTERRUPT		(1 << 15)
+#define GT_BSD_USER_INTERRUPT			(1 << 12)
+#define GT_RENDER_L3_PARITY_ERROR_INTERRUPT	(1 <<  5) /* !snb */
+#define GT_RENDER_PIPECTL_NOTIFY_INTERRUPT	(1 <<  4)
+#define GT_RENDER_CS_MASTER_ERROR_INTERRUPT	(1 <<  3)
+#define GT_RENDER_SYNC_STATUS_INTERRUPT		(1 <<  2)
+#define GT_RENDER_DEBUG_INTERRUPT		(1 <<  1)
+#define GT_RENDER_USER_INTERRUPT		(1 <<  0)
+
+#define PM_VEBOX_CS_ERROR_INTERRUPT		(1 << 12) /* hsw+ */
+#define PM_VEBOX_USER_INTERRUPT			(1 << 10) /* hsw+ */
+
+/* These are all the "old" interrupts */
+#define ILK_BSD_USER_INTERRUPT				(1<<5)
+#define I915_PIPE_CONTROL_NOTIFY_INTERRUPT		(1<<18)
+#define I915_DISPLAY_PORT_INTERRUPT			(1<<17)
+#define I915_RENDER_COMMAND_PARSER_ERROR_INTERRUPT	(1<<15)
+#define I915_GMCH_THERMAL_SENSOR_EVENT_INTERRUPT	(1<<14) /* p-state */
+#define I915_HWB_OOM_INTERRUPT				(1<<13)
+#define I915_SYNC_STATUS_INTERRUPT			(1<<12)
+#define I915_DISPLAY_PLANE_A_FLIP_PENDING_INTERRUPT	(1<<11)
+#define I915_DISPLAY_PLANE_B_FLIP_PENDING_INTERRUPT	(1<<10)
+#define I915_OVERLAY_PLANE_FLIP_PENDING_INTERRUPT	(1<<9)
+#define I915_DISPLAY_PLANE_C_FLIP_PENDING_INTERRUPT	(1<<8)
+#define I915_DISPLAY_PIPE_A_VBLANK_INTERRUPT		(1<<7)
+#define I915_DISPLAY_PIPE_A_EVENT_INTERRUPT		(1<<6)
+#define I915_DISPLAY_PIPE_B_VBLANK_INTERRUPT		(1<<5)
+#define I915_DISPLAY_PIPE_B_EVENT_INTERRUPT		(1<<4)
+#define I915_DEBUG_INTERRUPT				(1<<2)
+#define I915_USER_INTERRUPT				(1<<1)
+#define I915_ASLE_INTERRUPT				(1<<0)
+#define I915_BSD_USER_INTERRUPT				(1 << 25)
 
 #define GEN6_BSD_RNCID			0x12198
 
@@ -807,7 +963,9 @@
 #define   DPFC_CTL_EN		(1<<31)
 #define   DPFC_CTL_PLANEA	(0<<30)
 #define   DPFC_CTL_PLANEB	(1<<30)
+#define   IVB_DPFC_CTL_PLANE_SHIFT	(29)
 #define   DPFC_CTL_FENCE_EN	(1<<29)
+#define   IVB_DPFC_CTL_FENCE_EN	(1<<28)
 #define   DPFC_CTL_PERSISTENT_MODE	(1<<25)
 #define   DPFC_SR_EN		(1<<10)
 #define   DPFC_CTL_LIMIT_1X	(0<<6)
@@ -840,6 +998,7 @@
 #define ILK_DPFC_CHICKEN	0x43224
 #define ILK_FBC_RT_BASE		0x2128
 #define   ILK_FBC_RT_VALID	(1<<0)
+#define   SNB_FBC_FRONT_BUFFER	(1<<1)
 
 #define ILK_DISPLAY_CHICKEN1	0x42000
 #define   ILK_FBCQ_DIS		(1<<22)
@@ -855,6 +1014,25 @@
 #define   SNB_CPU_FENCE_ENABLE	(1<<29)
 #define DPFC_CPU_FENCE_OFFSET	0x100104
 
+/* Framebuffer compression for Ivybridge */
+#define IVB_FBC_RT_BASE			0x7020
+
+#define IPS_CTL		0x43408
+#define   IPS_ENABLE	(1 << 31)
+
+#define MSG_FBC_REND_STATE	0x50380
+#define   FBC_REND_NUKE		(1<<2)
+#define   FBC_REND_CACHE_CLEAN	(1<<1)
+
+#define _HSW_PIPE_SLICE_CHICKEN_1_A	0x420B0
+#define _HSW_PIPE_SLICE_CHICKEN_1_B	0x420B4
+#define   HSW_BYPASS_FBC_QUEUE		(1<<22)
+#define HSW_PIPE_SLICE_CHICKEN_1(pipe) _PIPE(pipe, + \
+					     _HSW_PIPE_SLICE_CHICKEN_1_A, + \
+					     _HSW_PIPE_SLICE_CHICKEN_1_B)
+
+#define HSW_CLKGATE_DISABLE_PART_1	0x46500
+#define   HSW_DPFC_GATING_DISABLE	(1<<23)
 
 /*
  * GPIO regs
@@ -963,7 +1141,10 @@
 #define   DPLL_FPA01_P1_POST_DIV_MASK	0x00ff0000 /* i915 */
 #define   DPLL_FPA01_P1_POST_DIV_MASK_PINEVIEW	0x00ff8000 /* Pineview */
 #define   DPLL_LOCK_VLV			(1<<15)
+#define   DPLL_INTEGRATED_CRI_CLK_VLV	(1<<14)
 #define   DPLL_INTEGRATED_CLOCK_VLV	(1<<13)
+#define   DPLL_PORTC_READY_MASK		(0xf << 4)
+#define   DPLL_PORTB_READY_MASK		(0xf)
 
 #define   DPLL_FPA01_P1_POST_DIV_MASK_I830	0x001f0000
 /*
@@ -1073,7 +1254,7 @@
 #define  DSTATE_PLL_D3_OFF			(1<<3)
 #define  DSTATE_GFX_CLOCK_GATING		(1<<1)
 #define  DSTATE_DOT_CLOCK_GATING		(1<<0)
-#define DSPCLK_GATE_D		0x6200
+#define DSPCLK_GATE_D	(dev_priv->info->display_mmio_offset + 0x6200)
 # define DPUNIT_B_CLOCK_GATE_DISABLE		(1 << 30) /* 965 */
 # define VSUNIT_CLOCK_GATE_DISABLE		(1 << 29) /* 965 */
 # define VRHUNIT_CLOCK_GATE_DISABLE		(1 << 28) /* 965 */
@@ -1186,6 +1367,8 @@
 #define FW_BLC_SELF_VLV		(VLV_DISPLAY_BASE + 0x6500)
 #define  FW_CSPWRDWNEN		(1<<15)
 
+#define MI_ARB_VLV		(VLV_DISPLAY_BASE + 0x6504)
+
 /*
  * Palette regs
  */
@@ -1535,14 +1718,13 @@
 					 GEN7_CXT_EXTENDED_SIZE(ctx_reg) + \
 					 GEN7_CXT_GT1_SIZE(ctx_reg) + \
 					 GEN7_CXT_VFSTATE_SIZE(ctx_reg))
-#define HSW_CXT_POWER_SIZE(ctx_reg)	((ctx_reg >> 26) & 0x3f)
-#define HSW_CXT_RING_SIZE(ctx_reg)	((ctx_reg >> 23) & 0x7)
-#define HSW_CXT_RENDER_SIZE(ctx_reg)	((ctx_reg >> 15) & 0xff)
-#define HSW_CXT_TOTAL_SIZE(ctx_reg)	(HSW_CXT_POWER_SIZE(ctx_reg) + \
-					 HSW_CXT_RING_SIZE(ctx_reg) + \
-					 HSW_CXT_RENDER_SIZE(ctx_reg) + \
-					 GEN7_CXT_VFSTATE_SIZE(ctx_reg))
-
+/* Haswell does have the CXT_SIZE register however it does not appear to be
+ * valid. Now, docs explain in dwords what is in the context object. The full
+ * size is 70720 bytes, however, the power context and execlist context will
+ * never be saved (power context is stored elsewhere, and execlists don't work
+ * on HSW) - so the final size is 66944 bytes, which rounds to 17 pages.
+ */
+#define HSW_CXT_TOTAL_SIZE		(17 * PAGE_SIZE)
 
 /*
  * Overlay regs
@@ -1691,6 +1873,12 @@
 /* SDVO is different across gen3/4 */
 #define   SDVOC_HOTPLUG_INT_STATUS_G4X		(1 << 3)
 #define   SDVOB_HOTPLUG_INT_STATUS_G4X		(1 << 2)
+/*
+ * Bspec seems to be seriously misleaded about the SDVO hpd bits on i965g/gm,
+ * since reality corrobates that they're the same as on gen3. But keep these
+ * bits here (and the comment!) to help any other lost wanderers back onto the
+ * right tracks.
+ */
 #define   SDVOC_HOTPLUG_INT_STATUS_I965		(3 << 4)
 #define   SDVOB_HOTPLUG_INT_STATUS_I965		(3 << 2)
 #define   SDVOC_HOTPLUG_INT_STATUS_I915		(1 << 7)
@@ -1702,13 +1890,6 @@
 						 PORTC_HOTPLUG_INT_STATUS | \
 						 PORTD_HOTPLUG_INT_STATUS)
 
-#define HOTPLUG_INT_STATUS_I965			(CRT_HOTPLUG_INT_STATUS | \
-						 SDVOB_HOTPLUG_INT_STATUS_I965 | \
-						 SDVOC_HOTPLUG_INT_STATUS_I965 | \
-						 PORTB_HOTPLUG_INT_STATUS | \
-						 PORTC_HOTPLUG_INT_STATUS | \
-						 PORTD_HOTPLUG_INT_STATUS)
-
 #define HOTPLUG_INT_STATUS_I915			(CRT_HOTPLUG_INT_STATUS | \
 						 SDVOB_HOTPLUG_INT_STATUS_I915 | \
 						 SDVOC_HOTPLUG_INT_STATUS_I915 | \
@@ -1967,6 +2148,10 @@
 #define   BLM_PIPE_A			(0 << 29)
 #define   BLM_PIPE_B			(1 << 29)
 #define   BLM_PIPE_C			(2 << 29) /* ivb + */
+#define   BLM_TRANSCODER_A		BLM_PIPE_A /* hsw */
+#define   BLM_TRANSCODER_B		BLM_PIPE_B
+#define   BLM_TRANSCODER_C		BLM_PIPE_C
+#define   BLM_TRANSCODER_EDP		(3 << 29)
 #define   BLM_PIPE(pipe)		((pipe) << 29)
 #define   BLM_POLARITY_I965		(1 << 28) /* gen4 only */
 #define   BLM_PHASE_IN_INTERUPT_STATUS	(1 << 26)
@@ -2540,9 +2725,7 @@
 #define   DP_PRE_EMPHASIS_SHIFT		22
 
 /* How many wires to use. I guess 3 was too hard */
-#define   DP_PORT_WIDTH_1		(0 << 19)
-#define   DP_PORT_WIDTH_2		(1 << 19)
-#define   DP_PORT_WIDTH_4		(3 << 19)
+#define   DP_PORT_WIDTH(width)		(((width) - 1) << 19)
 #define   DP_PORT_WIDTH_MASK		(7 << 19)
 
 /* Mystic DPCD version 1.1 special mode */
@@ -2646,18 +2829,20 @@
  * which is after the LUTs, so we want the bytes for our color format.
  * For our current usage, this is always 3, one byte for R, G and B.
  */
-#define _PIPEA_GMCH_DATA_M			0x70050
-#define _PIPEB_GMCH_DATA_M			0x71050
+#define _PIPEA_DATA_M_G4X	0x70050
+#define _PIPEB_DATA_M_G4X	0x71050
 
 /* Transfer unit size for display port - 1, default is 0x3f (for TU size 64) */
 #define  TU_SIZE(x)             (((x)-1) << 25) /* default size 64 */
+#define  TU_SIZE_SHIFT		25
 #define  TU_SIZE_MASK           (0x3f << 25)
 
 #define  DATA_LINK_M_N_MASK	(0xffffff)
 #define  DATA_LINK_N_MAX	(0x800000)
 
-#define _PIPEA_GMCH_DATA_N			0x70054
-#define _PIPEB_GMCH_DATA_N			0x71054
+#define _PIPEA_DATA_N_G4X	0x70054
+#define _PIPEB_DATA_N_G4X	0x71054
+#define   PIPE_GMCH_DATA_N_MASK			(0xffffff)
 
 /*
  * Computing Link M and N values for the Display Port link
@@ -2670,16 +2855,18 @@
  * Attributes and VB-ID.
  */
 
-#define _PIPEA_DP_LINK_M				0x70060
-#define _PIPEB_DP_LINK_M				0x71060
+#define _PIPEA_LINK_M_G4X	0x70060
+#define _PIPEB_LINK_M_G4X	0x71060
+#define   PIPEA_DP_LINK_M_MASK			(0xffffff)
 
-#define _PIPEA_DP_LINK_N				0x70064
-#define _PIPEB_DP_LINK_N				0x71064
+#define _PIPEA_LINK_N_G4X	0x70064
+#define _PIPEB_LINK_N_G4X	0x71064
+#define   PIPEA_DP_LINK_N_MASK			(0xffffff)
 
-#define PIPE_GMCH_DATA_M(pipe) _PIPE(pipe, _PIPEA_GMCH_DATA_M, _PIPEB_GMCH_DATA_M)
-#define PIPE_GMCH_DATA_N(pipe) _PIPE(pipe, _PIPEA_GMCH_DATA_N, _PIPEB_GMCH_DATA_N)
-#define PIPE_DP_LINK_M(pipe) _PIPE(pipe, _PIPEA_DP_LINK_M, _PIPEB_DP_LINK_M)
-#define PIPE_DP_LINK_N(pipe) _PIPE(pipe, _PIPEA_DP_LINK_N, _PIPEB_DP_LINK_N)
+#define PIPE_DATA_M_G4X(pipe) _PIPE(pipe, _PIPEA_DATA_M_G4X, _PIPEB_DATA_M_G4X)
+#define PIPE_DATA_N_G4X(pipe) _PIPE(pipe, _PIPEA_DATA_N_G4X, _PIPEB_DATA_N_G4X)
+#define PIPE_LINK_M_G4X(pipe) _PIPE(pipe, _PIPEA_LINK_M_G4X, _PIPEB_LINK_M_G4X)
+#define PIPE_LINK_N_G4X(pipe) _PIPE(pipe, _PIPEA_LINK_N_G4X, _PIPEB_LINK_N_G4X)
 
 /* Display & cursor control */
 
@@ -2715,6 +2902,7 @@
 #define   PIPECONF_INTERLACED_ILK		(3 << 21)
 #define   PIPECONF_INTERLACED_DBL_ILK		(4 << 21) /* ilk/snb only */
 #define   PIPECONF_PFIT_PF_INTERLACED_DBL_ILK	(5 << 21) /* ilk/snb only */
+#define   PIPECONF_INTERLACE_MODE_MASK		(7 << 21)
 #define   PIPECONF_CXSR_DOWNCLOCK	(1<<16)
 #define   PIPECONF_COLOR_RANGE_SELECT	(1 << 13)
 #define   PIPECONF_BPC_MASK	(0x7 << 5)
@@ -2915,6 +3103,10 @@
 #define WM3S_LP_IVB		0x45128
 #define  WM1S_LP_EN		(1<<31)
 
+#define HSW_WM_LP_VAL(lat, fbc, pri, cur) \
+	(WM3_LP_EN | ((lat) << WM1_LP_LATENCY_SHIFT) | \
+	 ((fbc) << WM1_LP_FBC_SHIFT) | ((pri) << WM1_LP_SR_SHIFT) | (cur))
+
 /* Memory latency timer register */
 #define MLTR_ILK		0x11222
 #define  MLTR_WM1_SHIFT		0
@@ -3294,7 +3486,7 @@
 #define SPRGAMC(pipe) _PIPE(pipe, _SPRA_GAMC, _SPRB_GAMC)
 #define SPRSURFLIVE(pipe) _PIPE(pipe, _SPRA_SURFLIVE, _SPRB_SURFLIVE)
 
-#define _SPACNTR		0x72180
+#define _SPACNTR		(VLV_DISPLAY_BASE + 0x72180)
 #define   SP_ENABLE			(1<<31)
 #define   SP_GEAMMA_ENABLE		(1<<30)
 #define   SP_PIXFORMAT_MASK		(0xf<<26)
@@ -3313,30 +3505,30 @@
 #define   SP_YUV_ORDER_YVYU		(2<<16)
 #define   SP_YUV_ORDER_VYUY		(3<<16)
 #define   SP_TILED			(1<<10)
-#define _SPALINOFF		0x72184
-#define _SPASTRIDE		0x72188
-#define _SPAPOS			0x7218c
-#define _SPASIZE		0x72190
-#define _SPAKEYMINVAL		0x72194
-#define _SPAKEYMSK		0x72198
-#define _SPASURF		0x7219c
-#define _SPAKEYMAXVAL		0x721a0
-#define _SPATILEOFF		0x721a4
-#define _SPACONSTALPHA		0x721a8
-#define _SPAGAMC		0x721f4
-
-#define _SPBCNTR		0x72280
-#define _SPBLINOFF		0x72284
-#define _SPBSTRIDE		0x72288
-#define _SPBPOS			0x7228c
-#define _SPBSIZE		0x72290
-#define _SPBKEYMINVAL		0x72294
-#define _SPBKEYMSK		0x72298
-#define _SPBSURF		0x7229c
-#define _SPBKEYMAXVAL		0x722a0
-#define _SPBTILEOFF		0x722a4
-#define _SPBCONSTALPHA		0x722a8
-#define _SPBGAMC		0x722f4
+#define _SPALINOFF		(VLV_DISPLAY_BASE + 0x72184)
+#define _SPASTRIDE		(VLV_DISPLAY_BASE + 0x72188)
+#define _SPAPOS			(VLV_DISPLAY_BASE + 0x7218c)
+#define _SPASIZE		(VLV_DISPLAY_BASE + 0x72190)
+#define _SPAKEYMINVAL		(VLV_DISPLAY_BASE + 0x72194)
+#define _SPAKEYMSK		(VLV_DISPLAY_BASE + 0x72198)
+#define _SPASURF		(VLV_DISPLAY_BASE + 0x7219c)
+#define _SPAKEYMAXVAL		(VLV_DISPLAY_BASE + 0x721a0)
+#define _SPATILEOFF		(VLV_DISPLAY_BASE + 0x721a4)
+#define _SPACONSTALPHA		(VLV_DISPLAY_BASE + 0x721a8)
+#define _SPAGAMC		(VLV_DISPLAY_BASE + 0x721f4)
+
+#define _SPBCNTR		(VLV_DISPLAY_BASE + 0x72280)
+#define _SPBLINOFF		(VLV_DISPLAY_BASE + 0x72284)
+#define _SPBSTRIDE		(VLV_DISPLAY_BASE + 0x72288)
+#define _SPBPOS			(VLV_DISPLAY_BASE + 0x7228c)
+#define _SPBSIZE		(VLV_DISPLAY_BASE + 0x72290)
+#define _SPBKEYMINVAL		(VLV_DISPLAY_BASE + 0x72294)
+#define _SPBKEYMSK		(VLV_DISPLAY_BASE + 0x72298)
+#define _SPBSURF		(VLV_DISPLAY_BASE + 0x7229c)
+#define _SPBKEYMAXVAL		(VLV_DISPLAY_BASE + 0x722a0)
+#define _SPBTILEOFF		(VLV_DISPLAY_BASE + 0x722a4)
+#define _SPBCONSTALPHA		(VLV_DISPLAY_BASE + 0x722a8)
+#define _SPBGAMC		(VLV_DISPLAY_BASE + 0x722f4)
 
 #define SPCNTR(pipe, plane) _PIPE(pipe * 2 + plane, _SPACNTR, _SPBCNTR)
 #define SPLINOFF(pipe, plane) _PIPE(pipe * 2 + plane, _SPALINOFF, _SPBLINOFF)
@@ -3474,6 +3666,15 @@
 #define _LGC_PALETTE_B           0x4a800
 #define LGC_PALETTE(pipe) _PIPE(pipe, _LGC_PALETTE_A, _LGC_PALETTE_B)
 
+#define _GAMMA_MODE_A		0x4a480
+#define _GAMMA_MODE_B		0x4ac80
+#define GAMMA_MODE(pipe) _PIPE(pipe, _GAMMA_MODE_A, _GAMMA_MODE_B)
+#define GAMMA_MODE_MODE_MASK	(3 << 0)
+#define GAMMA_MODE_MODE_8BIT	(0 << 0)
+#define GAMMA_MODE_MODE_10BIT	(1 << 0)
+#define GAMMA_MODE_MODE_12BIT	(2 << 0)
+#define GAMMA_MODE_MODE_SPLIT	(3 << 0)
+
 /* interrupts */
 #define DE_MASTER_IRQ_CONTROL   (1 << 31)
 #define DE_SPRITEB_FLIP_DONE    (1 << 29)
@@ -3502,7 +3703,7 @@
 #define DE_PIPEA_FIFO_UNDERRUN  (1 << 0)
 
 /* More Ivybridge lolz */
-#define DE_ERR_DEBUG_IVB		(1<<30)
+#define DE_ERR_INT_IVB			(1<<30)
 #define DE_GSE_IVB			(1<<29)
 #define DE_PCH_EVENT_IVB		(1<<28)
 #define DE_DP_A_HOTPLUG_IVB		(1<<27)
@@ -3525,21 +3726,6 @@
 #define DEIIR   0x44008
 #define DEIER   0x4400c
 
-/* GT interrupt.
- * Note that for gen6+ the ring-specific interrupt bits do alias with the
- * corresponding bits in the per-ring interrupt control registers. */
-#define GT_GEN6_BLT_FLUSHDW_NOTIFY_INTERRUPT	(1 << 26)
-#define GT_GEN6_BLT_CS_ERROR_INTERRUPT		(1 << 25)
-#define GT_GEN6_BLT_USER_INTERRUPT		(1 << 22)
-#define GT_GEN6_BSD_CS_ERROR_INTERRUPT		(1 << 15)
-#define GT_GEN6_BSD_USER_INTERRUPT		(1 << 12)
-#define GT_BSD_USER_INTERRUPT			(1 << 5) /* ilk only */
-#define GT_GEN7_L3_PARITY_ERROR_INTERRUPT	(1 << 5)
-#define GT_PIPE_NOTIFY				(1 << 4)
-#define GT_RENDER_CS_ERROR_INTERRUPT		(1 << 3)
-#define GT_SYNC_STATUS				(1 << 2)
-#define GT_USER_INTERRUPT			(1 << 0)
-
 #define GTISR   0x44010
 #define GTIMR   0x44014
 #define GTIIR   0x44018
@@ -3569,6 +3755,9 @@
 # define CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE	(1 << 5)
 # define CHICKEN3_DGMG_DONE_FIX_DISABLE		(1 << 2)
 
+#define CHICKEN_PAR1_1		0x42080
+#define  FORCE_ARB_IDLE_PLANES	(1 << 14)
+
 #define DISP_ARB_CTL	0x45000
 #define  DISP_TILE_SURFACE_SWIZZLING	(1<<13)
 #define  DISP_FBC_WM_DIS		(1<<15)
@@ -3661,6 +3850,7 @@
 				 SDE_PORTC_HOTPLUG_CPT |	\
 				 SDE_PORTB_HOTPLUG_CPT)
 #define SDE_GMBUS_CPT		(1 << 17)
+#define SDE_ERROR_CPT		(1 << 16)
 #define SDE_AUDIO_CP_REQ_C_CPT	(1 << 10)
 #define SDE_AUDIO_CP_CHG_C_CPT	(1 << 9)
 #define SDE_FDI_RXC_CPT		(1 << 8)
@@ -3685,6 +3875,12 @@
 #define SDEIIR  0xc4008
 #define SDEIER  0xc400c
 
+#define SERR_INT			0xc4040
+#define  SERR_INT_POISON		(1<<31)
+#define  SERR_INT_TRANS_C_FIFO_UNDERRUN	(1<<6)
+#define  SERR_INT_TRANS_B_FIFO_UNDERRUN	(1<<3)
+#define  SERR_INT_TRANS_A_FIFO_UNDERRUN	(1<<0)
+
 /* digital port hotplug */
 #define PCH_PORT_HOTPLUG        0xc4030		/* SHOTPLUG_CTL */
 #define PORTD_HOTPLUG_ENABLE            (1 << 20)
@@ -3734,15 +3930,15 @@
 
 #define _PCH_DPLL_A              0xc6014
 #define _PCH_DPLL_B              0xc6018
-#define _PCH_DPLL(pll) (pll == 0 ? _PCH_DPLL_A : _PCH_DPLL_B)
+#define PCH_DPLL(pll) (pll == 0 ? _PCH_DPLL_A : _PCH_DPLL_B)
 
 #define _PCH_FPA0                0xc6040
 #define  FP_CB_TUNE		(0x3<<22)
 #define _PCH_FPA1                0xc6044
 #define _PCH_FPB0                0xc6048
 #define _PCH_FPB1                0xc604c
-#define _PCH_FP0(pll) (pll == 0 ? _PCH_FPA0 : _PCH_FPB0)
-#define _PCH_FP1(pll) (pll == 0 ? _PCH_FPA1 : _PCH_FPB1)
+#define PCH_FP0(pll) (pll == 0 ? _PCH_FPA0 : _PCH_FPB0)
+#define PCH_FP1(pll) (pll == 0 ? _PCH_FPA1 : _PCH_FPB1)
 
 #define PCH_DPLL_TEST           0xc606c
 
@@ -3782,46 +3978,40 @@
 #define PCH_SSC4_AUX_PARMS      0xc6214
 
 #define PCH_DPLL_SEL		0xc7000
-#define  TRANSA_DPLL_ENABLE	(1<<3)
-#define	 TRANSA_DPLLB_SEL	(1<<0)
-#define	 TRANSA_DPLLA_SEL	0
-#define  TRANSB_DPLL_ENABLE	(1<<7)
-#define	 TRANSB_DPLLB_SEL	(1<<4)
-#define	 TRANSB_DPLLA_SEL	(0)
-#define  TRANSC_DPLL_ENABLE	(1<<11)
-#define	 TRANSC_DPLLB_SEL	(1<<8)
-#define	 TRANSC_DPLLA_SEL	(0)
+#define	 TRANS_DPLLB_SEL(pipe)		(1 << (pipe * 4))
+#define	 TRANS_DPLLA_SEL(pipe)		0
+#define  TRANS_DPLL_ENABLE(pipe)	(1 << (pipe * 4 + 3))
 
 /* transcoder */
 
-#define _TRANS_HTOTAL_A          0xe0000
-#define  TRANS_HTOTAL_SHIFT     16
-#define  TRANS_HACTIVE_SHIFT    0
-#define _TRANS_HBLANK_A          0xe0004
-#define  TRANS_HBLANK_END_SHIFT 16
-#define  TRANS_HBLANK_START_SHIFT 0
-#define _TRANS_HSYNC_A           0xe0008
-#define  TRANS_HSYNC_END_SHIFT  16
-#define  TRANS_HSYNC_START_SHIFT 0
-#define _TRANS_VTOTAL_A          0xe000c
-#define  TRANS_VTOTAL_SHIFT     16
-#define  TRANS_VACTIVE_SHIFT    0
-#define _TRANS_VBLANK_A          0xe0010
-#define  TRANS_VBLANK_END_SHIFT 16
-#define  TRANS_VBLANK_START_SHIFT 0
-#define _TRANS_VSYNC_A           0xe0014
-#define  TRANS_VSYNC_END_SHIFT  16
-#define  TRANS_VSYNC_START_SHIFT 0
-#define _TRANS_VSYNCSHIFT_A	0xe0028
-
-#define _TRANSA_DATA_M1          0xe0030
-#define _TRANSA_DATA_N1          0xe0034
-#define _TRANSA_DATA_M2          0xe0038
-#define _TRANSA_DATA_N2          0xe003c
-#define _TRANSA_DP_LINK_M1       0xe0040
-#define _TRANSA_DP_LINK_N1       0xe0044
-#define _TRANSA_DP_LINK_M2       0xe0048
-#define _TRANSA_DP_LINK_N2       0xe004c
+#define _PCH_TRANS_HTOTAL_A		0xe0000
+#define  TRANS_HTOTAL_SHIFT		16
+#define  TRANS_HACTIVE_SHIFT		0
+#define _PCH_TRANS_HBLANK_A		0xe0004
+#define  TRANS_HBLANK_END_SHIFT		16
+#define  TRANS_HBLANK_START_SHIFT	0
+#define _PCH_TRANS_HSYNC_A		0xe0008
+#define  TRANS_HSYNC_END_SHIFT		16
+#define  TRANS_HSYNC_START_SHIFT	0
+#define _PCH_TRANS_VTOTAL_A		0xe000c
+#define  TRANS_VTOTAL_SHIFT		16
+#define  TRANS_VACTIVE_SHIFT		0
+#define _PCH_TRANS_VBLANK_A		0xe0010
+#define  TRANS_VBLANK_END_SHIFT		16
+#define  TRANS_VBLANK_START_SHIFT	0
+#define _PCH_TRANS_VSYNC_A		0xe0014
+#define  TRANS_VSYNC_END_SHIFT	 	16
+#define  TRANS_VSYNC_START_SHIFT	0
+#define _PCH_TRANS_VSYNCSHIFT_A		0xe0028
+
+#define _PCH_TRANSA_DATA_M1	0xe0030
+#define _PCH_TRANSA_DATA_N1	0xe0034
+#define _PCH_TRANSA_DATA_M2	0xe0038
+#define _PCH_TRANSA_DATA_N2	0xe003c
+#define _PCH_TRANSA_LINK_M1	0xe0040
+#define _PCH_TRANSA_LINK_N1	0xe0044
+#define _PCH_TRANSA_LINK_M2	0xe0048
+#define _PCH_TRANSA_LINK_N2	0xe004c
 
 /* Per-transcoder DIP controls */
 
@@ -3890,44 +4080,45 @@
 #define HSW_TVIDEO_DIP_VSC_DATA(trans) \
 	 _TRANSCODER(trans, HSW_VIDEO_DIP_VSC_DATA_A, HSW_VIDEO_DIP_VSC_DATA_B)
 
-#define _TRANS_HTOTAL_B          0xe1000
-#define _TRANS_HBLANK_B          0xe1004
-#define _TRANS_HSYNC_B           0xe1008
-#define _TRANS_VTOTAL_B          0xe100c
-#define _TRANS_VBLANK_B          0xe1010
-#define _TRANS_VSYNC_B           0xe1014
-#define _TRANS_VSYNCSHIFT_B	 0xe1028
-
-#define TRANS_HTOTAL(pipe) _PIPE(pipe, _TRANS_HTOTAL_A, _TRANS_HTOTAL_B)
-#define TRANS_HBLANK(pipe) _PIPE(pipe, _TRANS_HBLANK_A, _TRANS_HBLANK_B)
-#define TRANS_HSYNC(pipe) _PIPE(pipe, _TRANS_HSYNC_A, _TRANS_HSYNC_B)
-#define TRANS_VTOTAL(pipe) _PIPE(pipe, _TRANS_VTOTAL_A, _TRANS_VTOTAL_B)
-#define TRANS_VBLANK(pipe) _PIPE(pipe, _TRANS_VBLANK_A, _TRANS_VBLANK_B)
-#define TRANS_VSYNC(pipe) _PIPE(pipe, _TRANS_VSYNC_A, _TRANS_VSYNC_B)
-#define TRANS_VSYNCSHIFT(pipe) _PIPE(pipe, _TRANS_VSYNCSHIFT_A, \
-				     _TRANS_VSYNCSHIFT_B)
-
-#define _TRANSB_DATA_M1          0xe1030
-#define _TRANSB_DATA_N1          0xe1034
-#define _TRANSB_DATA_M2          0xe1038
-#define _TRANSB_DATA_N2          0xe103c
-#define _TRANSB_DP_LINK_M1       0xe1040
-#define _TRANSB_DP_LINK_N1       0xe1044
-#define _TRANSB_DP_LINK_M2       0xe1048
-#define _TRANSB_DP_LINK_N2       0xe104c
-
-#define TRANSDATA_M1(pipe) _PIPE(pipe, _TRANSA_DATA_M1, _TRANSB_DATA_M1)
-#define TRANSDATA_N1(pipe) _PIPE(pipe, _TRANSA_DATA_N1, _TRANSB_DATA_N1)
-#define TRANSDATA_M2(pipe) _PIPE(pipe, _TRANSA_DATA_M2, _TRANSB_DATA_M2)
-#define TRANSDATA_N2(pipe) _PIPE(pipe, _TRANSA_DATA_N2, _TRANSB_DATA_N2)
-#define TRANSDPLINK_M1(pipe) _PIPE(pipe, _TRANSA_DP_LINK_M1, _TRANSB_DP_LINK_M1)
-#define TRANSDPLINK_N1(pipe) _PIPE(pipe, _TRANSA_DP_LINK_N1, _TRANSB_DP_LINK_N1)
-#define TRANSDPLINK_M2(pipe) _PIPE(pipe, _TRANSA_DP_LINK_M2, _TRANSB_DP_LINK_M2)
-#define TRANSDPLINK_N2(pipe) _PIPE(pipe, _TRANSA_DP_LINK_N2, _TRANSB_DP_LINK_N2)
-
-#define _TRANSACONF              0xf0008
-#define _TRANSBCONF              0xf1008
-#define TRANSCONF(plane) _PIPE(plane, _TRANSACONF, _TRANSBCONF)
+#define _PCH_TRANS_HTOTAL_B          0xe1000
+#define _PCH_TRANS_HBLANK_B          0xe1004
+#define _PCH_TRANS_HSYNC_B           0xe1008
+#define _PCH_TRANS_VTOTAL_B          0xe100c
+#define _PCH_TRANS_VBLANK_B          0xe1010
+#define _PCH_TRANS_VSYNC_B           0xe1014
+#define _PCH_TRANS_VSYNCSHIFT_B	 0xe1028
+
+#define PCH_TRANS_HTOTAL(pipe) _PIPE(pipe, _PCH_TRANS_HTOTAL_A, _PCH_TRANS_HTOTAL_B)
+#define PCH_TRANS_HBLANK(pipe) _PIPE(pipe, _PCH_TRANS_HBLANK_A, _PCH_TRANS_HBLANK_B)
+#define PCH_TRANS_HSYNC(pipe) _PIPE(pipe, _PCH_TRANS_HSYNC_A, _PCH_TRANS_HSYNC_B)
+#define PCH_TRANS_VTOTAL(pipe) _PIPE(pipe, _PCH_TRANS_VTOTAL_A, _PCH_TRANS_VTOTAL_B)
+#define PCH_TRANS_VBLANK(pipe) _PIPE(pipe, _PCH_TRANS_VBLANK_A, _PCH_TRANS_VBLANK_B)
+#define PCH_TRANS_VSYNC(pipe) _PIPE(pipe, _PCH_TRANS_VSYNC_A, _PCH_TRANS_VSYNC_B)
+#define PCH_TRANS_VSYNCSHIFT(pipe) _PIPE(pipe, _PCH_TRANS_VSYNCSHIFT_A, \
+					 _PCH_TRANS_VSYNCSHIFT_B)
+
+#define _PCH_TRANSB_DATA_M1	0xe1030
+#define _PCH_TRANSB_DATA_N1	0xe1034
+#define _PCH_TRANSB_DATA_M2	0xe1038
+#define _PCH_TRANSB_DATA_N2	0xe103c
+#define _PCH_TRANSB_LINK_M1	0xe1040
+#define _PCH_TRANSB_LINK_N1	0xe1044
+#define _PCH_TRANSB_LINK_M2	0xe1048
+#define _PCH_TRANSB_LINK_N2	0xe104c
+
+#define PCH_TRANS_DATA_M1(pipe) _PIPE(pipe, _PCH_TRANSA_DATA_M1, _PCH_TRANSB_DATA_M1)
+#define PCH_TRANS_DATA_N1(pipe) _PIPE(pipe, _PCH_TRANSA_DATA_N1, _PCH_TRANSB_DATA_N1)
+#define PCH_TRANS_DATA_M2(pipe) _PIPE(pipe, _PCH_TRANSA_DATA_M2, _PCH_TRANSB_DATA_M2)
+#define PCH_TRANS_DATA_N2(pipe) _PIPE(pipe, _PCH_TRANSA_DATA_N2, _PCH_TRANSB_DATA_N2)
+#define PCH_TRANS_LINK_M1(pipe) _PIPE(pipe, _PCH_TRANSA_LINK_M1, _PCH_TRANSB_LINK_M1)
+#define PCH_TRANS_LINK_N1(pipe) _PIPE(pipe, _PCH_TRANSA_LINK_N1, _PCH_TRANSB_LINK_N1)
+#define PCH_TRANS_LINK_M2(pipe) _PIPE(pipe, _PCH_TRANSA_LINK_M2, _PCH_TRANSB_LINK_M2)
+#define PCH_TRANS_LINK_N2(pipe) _PIPE(pipe, _PCH_TRANSA_LINK_N2, _PCH_TRANSB_LINK_N2)
+
+#define _PCH_TRANSACONF              0xf0008
+#define _PCH_TRANSBCONF              0xf1008
+#define PCH_TRANSCONF(pipe) _PIPE(pipe, _PCH_TRANSACONF, _PCH_TRANSBCONF)
+#define LPT_TRANSCONF		_PCH_TRANSACONF /* lpt has only one transcoder */
 #define  TRANS_DISABLE          (0<<31)
 #define  TRANS_ENABLE           (1<<31)
 #define  TRANS_STATE_MASK       (1<<30)
@@ -4011,10 +4202,9 @@
 #define  FDI_LINK_TRAIN_600MV_3_5DB_SNB_B	(0x39<<22)
 #define  FDI_LINK_TRAIN_800MV_0DB_SNB_B		(0x38<<22)
 #define  FDI_LINK_TRAIN_VOL_EMP_MASK		(0x3f<<22)
-#define  FDI_DP_PORT_WIDTH_X1           (0<<19)
-#define  FDI_DP_PORT_WIDTH_X2           (1<<19)
-#define  FDI_DP_PORT_WIDTH_X3           (2<<19)
-#define  FDI_DP_PORT_WIDTH_X4           (3<<19)
+#define  FDI_DP_PORT_WIDTH_SHIFT		19
+#define  FDI_DP_PORT_WIDTH_MASK			(7 << FDI_DP_PORT_WIDTH_SHIFT)
+#define  FDI_DP_PORT_WIDTH(width)           (((width) - 1) << FDI_DP_PORT_WIDTH_SHIFT)
 #define  FDI_TX_ENHANCE_FRAME_ENABLE    (1<<18)
 /* Ironlake: hardwired to 1 */
 #define  FDI_TX_PLL_ENABLE              (1<<14)
@@ -4039,7 +4229,6 @@
 /* train, dp width same as FDI_TX */
 #define  FDI_FS_ERRC_ENABLE		(1<<27)
 #define  FDI_FE_ERRC_ENABLE		(1<<26)
-#define  FDI_DP_PORT_WIDTH_X8           (7<<19)
 #define  FDI_RX_POLARITY_REVERSED_LPT	(1<<16)
 #define  FDI_8BPC                       (0<<16)
 #define  FDI_10BPC                      (1<<16)
@@ -4061,9 +4250,6 @@
 #define  FDI_LINK_TRAIN_PATTERN_IDLE_CPT	(2<<8)
 #define  FDI_LINK_TRAIN_NORMAL_CPT		(3<<8)
 #define  FDI_LINK_TRAIN_PATTERN_MASK_CPT	(3<<8)
-/* LPT */
-#define  FDI_PORT_WIDTH_2X_LPT			(1<<19)
-#define  FDI_PORT_WIDTH_1X_LPT			(0<<19)
 
 #define _FDI_RXA_MISC			0xf0010
 #define _FDI_RXB_MISC			0xf1010
@@ -4309,6 +4495,7 @@
 #define   GEN6_RC_CTL_RC6_ENABLE		(1<<18)
 #define   GEN6_RC_CTL_RC1e_ENABLE		(1<<20)
 #define   GEN6_RC_CTL_RC7_ENABLE		(1<<22)
+#define   GEN7_RC_CTL_TO_MODE			(1<<28)
 #define   GEN6_RC_CTL_EI_MODE(x)		((x)<<27)
 #define   GEN6_RC_CTL_HW_ENABLE			(1<<31)
 #define GEN6_RP_DOWN_TIMEOUT			0xA010
@@ -4370,7 +4557,7 @@
 #define  GEN6_PM_RP_DOWN_THRESHOLD		(1<<4)
 #define  GEN6_PM_RP_UP_EI_EXPIRED		(1<<2)
 #define  GEN6_PM_RP_DOWN_EI_EXPIRED		(1<<1)
-#define  GEN6_PM_DEFERRED_EVENTS		(GEN6_PM_RP_UP_THRESHOLD | \
+#define  GEN6_PM_RPS_EVENTS			(GEN6_PM_RP_UP_THRESHOLD | \
 						 GEN6_PM_RP_DOWN_THRESHOLD | \
 						 GEN6_PM_RP_DOWN_TIMEOUT)
 
@@ -4392,20 +4579,6 @@
 #define   GEN6_PCODE_FREQ_IA_RATIO_SHIFT	8
 #define   GEN6_PCODE_FREQ_RING_RATIO_SHIFT	16
 
-#define VLV_IOSF_DOORBELL_REQ			0x182100
-#define   IOSF_DEVFN_SHIFT			24
-#define   IOSF_OPCODE_SHIFT			16
-#define   IOSF_PORT_SHIFT			8
-#define   IOSF_BYTE_ENABLES_SHIFT		4
-#define   IOSF_BAR_SHIFT			1
-#define   IOSF_SB_BUSY				(1<<0)
-#define   IOSF_PORT_PUNIT			0x4
-#define VLV_IOSF_DATA				0x182104
-#define VLV_IOSF_ADDR				0x182108
-
-#define PUNIT_OPCODE_REG_READ			6
-#define PUNIT_OPCODE_REG_WRITE			7
-
 #define GEN6_GT_CORE_STATUS		0x138060
 #define   GEN6_CORE_CPD_STATE_MASK	(7<<4)
 #define   GEN6_RCn_MASK			7
@@ -4602,9 +4775,6 @@
 #define  TRANS_DDI_EDP_INPUT_B_ONOFF	(5<<12)
 #define  TRANS_DDI_EDP_INPUT_C_ONOFF	(6<<12)
 #define  TRANS_DDI_BFI_ENABLE		(1<<4)
-#define  TRANS_DDI_PORT_WIDTH_X1	(0<<1)
-#define  TRANS_DDI_PORT_WIDTH_X2	(1<<1)
-#define  TRANS_DDI_PORT_WIDTH_X4	(3<<1)
 
 /* DisplayPort Transport Control */
 #define DP_TP_CTL_A			0x64040
@@ -4648,9 +4818,7 @@
 #define  DDI_BUF_PORT_REVERSAL			(1<<16)
 #define  DDI_BUF_IS_IDLE			(1<<7)
 #define  DDI_A_4_LANES				(1<<4)
-#define  DDI_PORT_WIDTH_X1			(0<<1)
-#define  DDI_PORT_WIDTH_X2			(1<<1)
-#define  DDI_PORT_WIDTH_X4			(3<<1)
+#define  DDI_PORT_WIDTH(width)			(((width) - 1) << 1)
 #define  DDI_INIT_DISPLAY_DETECTED		(1<<0)
 
 /* DDI Buffer Translations */
@@ -4774,6 +4942,9 @@
 #define  SFUSE_STRAP_DDIC_DETECTED	(1<<1)
 #define  SFUSE_STRAP_DDID_DETECTED	(1<<0)
 
+#define WM_MISC				0x45260
+#define  WM_MISC_DATA_PARTITION_5_6	(1 << 0)
+
 #define WM_DBG				0x45280
 #define  WM_DBG_DISALLOW_MULTIPLE_LP	(1<<0)
 #define  WM_DBG_DISALLOW_MAXFIFO	(1<<1)
@@ -4787,6 +4958,9 @@
 #define _PIPE_A_CSC_COEFF_RV_GV	0x49020
 #define _PIPE_A_CSC_COEFF_BV	0x49024
 #define _PIPE_A_CSC_MODE	0x49028
+#define   CSC_BLACK_SCREEN_OFFSET	(1 << 2)
+#define   CSC_POSITION_BEFORE_GAMMA	(1 << 1)
+#define   CSC_MODE_YUV_TO_RGB		(1 << 0)
 #define _PIPE_A_CSC_PREOFF_HI	0x49030
 #define _PIPE_A_CSC_PREOFF_ME	0x49034
 #define _PIPE_A_CSC_PREOFF_LO	0x49038
@@ -4808,10 +4982,6 @@
 #define _PIPE_B_CSC_POSTOFF_ME	0x49144
 #define _PIPE_B_CSC_POSTOFF_LO	0x49148
 
-#define CSC_BLACK_SCREEN_OFFSET (1 << 2)
-#define CSC_POSITION_BEFORE_GAMMA (1 << 1)
-#define CSC_MODE_YUV_TO_RGB (1 << 0)
-
 #define PIPE_CSC_COEFF_RY_GY(pipe) _PIPE(pipe, _PIPE_A_CSC_COEFF_RY_GY, _PIPE_B_CSC_COEFF_RY_GY)
 #define PIPE_CSC_COEFF_BY(pipe) _PIPE(pipe, _PIPE_A_CSC_COEFF_BY, _PIPE_B_CSC_COEFF_BY)
 #define PIPE_CSC_COEFF_RU_GU(pipe) _PIPE(pipe, _PIPE_A_CSC_COEFF_RU_GU, _PIPE_B_CSC_COEFF_RU_GU)
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 369b3d8776ab..70db618989c4 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -192,6 +192,7 @@ static void i915_restore_vga(struct drm_device *dev)
 static void i915_save_display(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long flags;
 
 	/* Display arbitration control */
 	if (INTEL_INFO(dev)->gen <= 4)
@@ -202,6 +203,8 @@ static void i915_save_display(struct drm_device *dev)
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		i915_save_display_reg(dev);
 
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
+
 	/* LVDS state */
 	if (HAS_PCH_SPLIT(dev)) {
 		dev_priv->regfile.savePP_CONTROL = I915_READ(PCH_PP_CONTROL);
@@ -222,6 +225,8 @@ static void i915_save_display(struct drm_device *dev)
 			dev_priv->regfile.saveLVDS = I915_READ(LVDS);
 	}
 
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
+
 	if (!IS_I830(dev) && !IS_845G(dev) && !HAS_PCH_SPLIT(dev))
 		dev_priv->regfile.savePFIT_CONTROL = I915_READ(PFIT_CONTROL);
 
@@ -257,6 +262,7 @@ static void i915_restore_display(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 mask = 0xffffffff;
+	unsigned long flags;
 
 	/* Display arbitration */
 	if (INTEL_INFO(dev)->gen <= 4)
@@ -265,6 +271,8 @@ static void i915_restore_display(struct drm_device *dev)
 	if (!drm_core_check_feature(dev, DRIVER_MODESET))
 		i915_restore_display_reg(dev);
 
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
+
 	/* LVDS state */
 	if (INTEL_INFO(dev)->gen >= 4 && !HAS_PCH_SPLIT(dev))
 		I915_WRITE(BLC_PWM_CTL2, dev_priv->regfile.saveBLC_PWM_CTL2);
@@ -304,6 +312,8 @@ static void i915_restore_display(struct drm_device *dev)
 		I915_WRITE(PP_CONTROL, dev_priv->regfile.savePP_CONTROL);
 	}
 
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
+
 	/* only restore FBC info on the platform that supports FBC*/
 	intel_disable_fbc(dev);
 	if (I915_HAS_FBC(dev)) {
diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c
index d5e1890678f9..6875b5654c63 100644
--- a/drivers/gpu/drm/i915/i915_sysfs.c
+++ b/drivers/gpu/drm/i915/i915_sysfs.c
@@ -212,7 +212,13 @@ static ssize_t gt_cur_freq_mhz_show(struct device *kdev,
 	int ret;
 
 	mutex_lock(&dev_priv->rps.hw_lock);
-	ret = dev_priv->rps.cur_delay * GT_FREQUENCY_MULTIPLIER;
+	if (IS_VALLEYVIEW(dev_priv->dev)) {
+		u32 freq;
+		freq = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
+		ret = vlv_gpu_freq(dev_priv->mem_freq, (freq >> 8) & 0xff);
+	} else {
+		ret = dev_priv->rps.cur_delay * GT_FREQUENCY_MULTIPLIER;
+	}
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n", ret);
@@ -226,7 +232,10 @@ static ssize_t gt_max_freq_mhz_show(struct device *kdev, struct device_attribute
 	int ret;
 
 	mutex_lock(&dev_priv->rps.hw_lock);
-	ret = dev_priv->rps.max_delay * GT_FREQUENCY_MULTIPLIER;
+	if (IS_VALLEYVIEW(dev_priv->dev))
+		ret = vlv_gpu_freq(dev_priv->mem_freq, dev_priv->rps.max_delay);
+	else
+		ret = dev_priv->rps.max_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n", ret);
@@ -246,16 +255,25 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev,
 	if (ret)
 		return ret;
 
-	val /= GT_FREQUENCY_MULTIPLIER;
-
 	mutex_lock(&dev_priv->rps.hw_lock);
 
-	rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
-	hw_max = dev_priv->rps.hw_max;
-	non_oc_max = (rp_state_cap & 0xff);
-	hw_min = ((rp_state_cap & 0xff0000) >> 16);
+	if (IS_VALLEYVIEW(dev_priv->dev)) {
+		val = vlv_freq_opcode(dev_priv->mem_freq, val);
+
+		hw_max = valleyview_rps_max_freq(dev_priv);
+		hw_min = valleyview_rps_min_freq(dev_priv);
+		non_oc_max = hw_max;
+	} else {
+		val /= GT_FREQUENCY_MULTIPLIER;
+
+		rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
+		hw_max = dev_priv->rps.hw_max;
+		non_oc_max = (rp_state_cap & 0xff);
+		hw_min = ((rp_state_cap & 0xff0000) >> 16);
+	}
 
-	if (val < hw_min || val > hw_max || val < dev_priv->rps.min_delay) {
+	if (val < hw_min || val > hw_max ||
+	    val < dev_priv->rps.min_delay) {
 		mutex_unlock(&dev_priv->rps.hw_lock);
 		return -EINVAL;
 	}
@@ -264,8 +282,12 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev,
 		DRM_DEBUG("User requested overclocking to %d\n",
 			  val * GT_FREQUENCY_MULTIPLIER);
 
-	if (dev_priv->rps.cur_delay > val)
-		gen6_set_rps(dev_priv->dev, val);
+	if (dev_priv->rps.cur_delay > val) {
+		if (IS_VALLEYVIEW(dev_priv->dev))
+			valleyview_set_rps(dev_priv->dev, val);
+		else
+			gen6_set_rps(dev_priv->dev, val);
+	}
 
 	dev_priv->rps.max_delay = val;
 
@@ -282,7 +304,10 @@ static ssize_t gt_min_freq_mhz_show(struct device *kdev, struct device_attribute
 	int ret;
 
 	mutex_lock(&dev_priv->rps.hw_lock);
-	ret = dev_priv->rps.min_delay * GT_FREQUENCY_MULTIPLIER;
+	if (IS_VALLEYVIEW(dev_priv->dev))
+		ret = vlv_gpu_freq(dev_priv->mem_freq, dev_priv->rps.min_delay);
+	else
+		ret = dev_priv->rps.min_delay * GT_FREQUENCY_MULTIPLIER;
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	return snprintf(buf, PAGE_SIZE, "%d\n", ret);
@@ -302,21 +327,32 @@ static ssize_t gt_min_freq_mhz_store(struct device *kdev,
 	if (ret)
 		return ret;
 
-	val /= GT_FREQUENCY_MULTIPLIER;
-
 	mutex_lock(&dev_priv->rps.hw_lock);
 
-	rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
-	hw_max = dev_priv->rps.hw_max;
-	hw_min = ((rp_state_cap & 0xff0000) >> 16);
+	if (IS_VALLEYVIEW(dev)) {
+		val = vlv_freq_opcode(dev_priv->mem_freq, val);
+
+		hw_max = valleyview_rps_max_freq(dev_priv);
+		hw_min = valleyview_rps_min_freq(dev_priv);
+	} else {
+		val /= GT_FREQUENCY_MULTIPLIER;
+
+		rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
+		hw_max = dev_priv->rps.hw_max;
+		hw_min = ((rp_state_cap & 0xff0000) >> 16);
+	}
 
 	if (val < hw_min || val > hw_max || val > dev_priv->rps.max_delay) {
 		mutex_unlock(&dev_priv->rps.hw_lock);
 		return -EINVAL;
 	}
 
-	if (dev_priv->rps.cur_delay < val)
-		gen6_set_rps(dev_priv->dev, val);
+	if (dev_priv->rps.cur_delay < val) {
+		if (IS_VALLEYVIEW(dev))
+			valleyview_set_rps(dev, val);
+		else
+			gen6_set_rps(dev_priv->dev, val);
+	}
 
 	dev_priv->rps.min_delay = val;
 
diff --git a/drivers/gpu/drm/i915/i915_ums.c b/drivers/gpu/drm/i915/i915_ums.c
index 985a09716237..967da4772c44 100644
--- a/drivers/gpu/drm/i915/i915_ums.c
+++ b/drivers/gpu/drm/i915/i915_ums.c
@@ -41,7 +41,7 @@ static bool i915_pipe_enabled(struct drm_device *dev, enum pipe pipe)
 		return false;
 
 	if (HAS_PCH_SPLIT(dev))
-		dpll_reg = _PCH_DPLL(pipe);
+		dpll_reg = PCH_DPLL(pipe);
 	else
 		dpll_reg = (pipe == PIPE_A) ? _DPLL_A : _DPLL_B;
 
@@ -148,13 +148,13 @@ void i915_save_display_reg(struct drm_device *dev)
 		dev_priv->regfile.savePFA_WIN_SZ = I915_READ(_PFA_WIN_SZ);
 		dev_priv->regfile.savePFA_WIN_POS = I915_READ(_PFA_WIN_POS);
 
-		dev_priv->regfile.saveTRANSACONF = I915_READ(_TRANSACONF);
-		dev_priv->regfile.saveTRANS_HTOTAL_A = I915_READ(_TRANS_HTOTAL_A);
-		dev_priv->regfile.saveTRANS_HBLANK_A = I915_READ(_TRANS_HBLANK_A);
-		dev_priv->regfile.saveTRANS_HSYNC_A = I915_READ(_TRANS_HSYNC_A);
-		dev_priv->regfile.saveTRANS_VTOTAL_A = I915_READ(_TRANS_VTOTAL_A);
-		dev_priv->regfile.saveTRANS_VBLANK_A = I915_READ(_TRANS_VBLANK_A);
-		dev_priv->regfile.saveTRANS_VSYNC_A = I915_READ(_TRANS_VSYNC_A);
+		dev_priv->regfile.saveTRANSACONF = I915_READ(_PCH_TRANSACONF);
+		dev_priv->regfile.saveTRANS_HTOTAL_A = I915_READ(_PCH_TRANS_HTOTAL_A);
+		dev_priv->regfile.saveTRANS_HBLANK_A = I915_READ(_PCH_TRANS_HBLANK_A);
+		dev_priv->regfile.saveTRANS_HSYNC_A = I915_READ(_PCH_TRANS_HSYNC_A);
+		dev_priv->regfile.saveTRANS_VTOTAL_A = I915_READ(_PCH_TRANS_VTOTAL_A);
+		dev_priv->regfile.saveTRANS_VBLANK_A = I915_READ(_PCH_TRANS_VBLANK_A);
+		dev_priv->regfile.saveTRANS_VSYNC_A = I915_READ(_PCH_TRANS_VSYNC_A);
 	}
 
 	dev_priv->regfile.saveDSPACNTR = I915_READ(_DSPACNTR);
@@ -205,13 +205,13 @@ void i915_save_display_reg(struct drm_device *dev)
 		dev_priv->regfile.savePFB_WIN_SZ = I915_READ(_PFB_WIN_SZ);
 		dev_priv->regfile.savePFB_WIN_POS = I915_READ(_PFB_WIN_POS);
 
-		dev_priv->regfile.saveTRANSBCONF = I915_READ(_TRANSBCONF);
-		dev_priv->regfile.saveTRANS_HTOTAL_B = I915_READ(_TRANS_HTOTAL_B);
-		dev_priv->regfile.saveTRANS_HBLANK_B = I915_READ(_TRANS_HBLANK_B);
-		dev_priv->regfile.saveTRANS_HSYNC_B = I915_READ(_TRANS_HSYNC_B);
-		dev_priv->regfile.saveTRANS_VTOTAL_B = I915_READ(_TRANS_VTOTAL_B);
-		dev_priv->regfile.saveTRANS_VBLANK_B = I915_READ(_TRANS_VBLANK_B);
-		dev_priv->regfile.saveTRANS_VSYNC_B = I915_READ(_TRANS_VSYNC_B);
+		dev_priv->regfile.saveTRANSBCONF = I915_READ(_PCH_TRANSBCONF);
+		dev_priv->regfile.saveTRANS_HTOTAL_B = I915_READ(_PCH_TRANS_HTOTAL_B);
+		dev_priv->regfile.saveTRANS_HBLANK_B = I915_READ(_PCH_TRANS_HBLANK_B);
+		dev_priv->regfile.saveTRANS_HSYNC_B = I915_READ(_PCH_TRANS_HSYNC_B);
+		dev_priv->regfile.saveTRANS_VTOTAL_B = I915_READ(_PCH_TRANS_VTOTAL_B);
+		dev_priv->regfile.saveTRANS_VBLANK_B = I915_READ(_PCH_TRANS_VBLANK_B);
+		dev_priv->regfile.saveTRANS_VSYNC_B = I915_READ(_PCH_TRANS_VSYNC_B);
 	}
 
 	dev_priv->regfile.saveDSPBCNTR = I915_READ(_DSPBCNTR);
@@ -259,14 +259,14 @@ void i915_save_display_reg(struct drm_device *dev)
 		dev_priv->regfile.saveDP_B = I915_READ(DP_B);
 		dev_priv->regfile.saveDP_C = I915_READ(DP_C);
 		dev_priv->regfile.saveDP_D = I915_READ(DP_D);
-		dev_priv->regfile.savePIPEA_GMCH_DATA_M = I915_READ(_PIPEA_GMCH_DATA_M);
-		dev_priv->regfile.savePIPEB_GMCH_DATA_M = I915_READ(_PIPEB_GMCH_DATA_M);
-		dev_priv->regfile.savePIPEA_GMCH_DATA_N = I915_READ(_PIPEA_GMCH_DATA_N);
-		dev_priv->regfile.savePIPEB_GMCH_DATA_N = I915_READ(_PIPEB_GMCH_DATA_N);
-		dev_priv->regfile.savePIPEA_DP_LINK_M = I915_READ(_PIPEA_DP_LINK_M);
-		dev_priv->regfile.savePIPEB_DP_LINK_M = I915_READ(_PIPEB_DP_LINK_M);
-		dev_priv->regfile.savePIPEA_DP_LINK_N = I915_READ(_PIPEA_DP_LINK_N);
-		dev_priv->regfile.savePIPEB_DP_LINK_N = I915_READ(_PIPEB_DP_LINK_N);
+		dev_priv->regfile.savePIPEA_GMCH_DATA_M = I915_READ(_PIPEA_DATA_M_G4X);
+		dev_priv->regfile.savePIPEB_GMCH_DATA_M = I915_READ(_PIPEB_DATA_M_G4X);
+		dev_priv->regfile.savePIPEA_GMCH_DATA_N = I915_READ(_PIPEA_DATA_N_G4X);
+		dev_priv->regfile.savePIPEB_GMCH_DATA_N = I915_READ(_PIPEB_DATA_N_G4X);
+		dev_priv->regfile.savePIPEA_DP_LINK_M = I915_READ(_PIPEA_LINK_M_G4X);
+		dev_priv->regfile.savePIPEB_DP_LINK_M = I915_READ(_PIPEB_LINK_M_G4X);
+		dev_priv->regfile.savePIPEA_DP_LINK_N = I915_READ(_PIPEA_LINK_N_G4X);
+		dev_priv->regfile.savePIPEB_DP_LINK_N = I915_READ(_PIPEB_LINK_N_G4X);
 	}
 	/* FIXME: regfile.save TV & SDVO state */
 
@@ -282,14 +282,14 @@ void i915_restore_display_reg(struct drm_device *dev)
 
 	/* Display port ratios (must be done before clock is set) */
 	if (SUPPORTS_INTEGRATED_DP(dev)) {
-		I915_WRITE(_PIPEA_GMCH_DATA_M, dev_priv->regfile.savePIPEA_GMCH_DATA_M);
-		I915_WRITE(_PIPEB_GMCH_DATA_M, dev_priv->regfile.savePIPEB_GMCH_DATA_M);
-		I915_WRITE(_PIPEA_GMCH_DATA_N, dev_priv->regfile.savePIPEA_GMCH_DATA_N);
-		I915_WRITE(_PIPEB_GMCH_DATA_N, dev_priv->regfile.savePIPEB_GMCH_DATA_N);
-		I915_WRITE(_PIPEA_DP_LINK_M, dev_priv->regfile.savePIPEA_DP_LINK_M);
-		I915_WRITE(_PIPEB_DP_LINK_M, dev_priv->regfile.savePIPEB_DP_LINK_M);
-		I915_WRITE(_PIPEA_DP_LINK_N, dev_priv->regfile.savePIPEA_DP_LINK_N);
-		I915_WRITE(_PIPEB_DP_LINK_N, dev_priv->regfile.savePIPEB_DP_LINK_N);
+		I915_WRITE(_PIPEA_DATA_M_G4X, dev_priv->regfile.savePIPEA_GMCH_DATA_M);
+		I915_WRITE(_PIPEB_DATA_M_G4X, dev_priv->regfile.savePIPEB_GMCH_DATA_M);
+		I915_WRITE(_PIPEA_DATA_N_G4X, dev_priv->regfile.savePIPEA_GMCH_DATA_N);
+		I915_WRITE(_PIPEB_DATA_N_G4X, dev_priv->regfile.savePIPEB_GMCH_DATA_N);
+		I915_WRITE(_PIPEA_LINK_M_G4X, dev_priv->regfile.savePIPEA_DP_LINK_M);
+		I915_WRITE(_PIPEB_LINK_M_G4X, dev_priv->regfile.savePIPEB_DP_LINK_M);
+		I915_WRITE(_PIPEA_LINK_N_G4X, dev_priv->regfile.savePIPEA_DP_LINK_N);
+		I915_WRITE(_PIPEB_LINK_N_G4X, dev_priv->regfile.savePIPEB_DP_LINK_N);
 	}
 
 	/* Fences */
@@ -379,13 +379,13 @@ void i915_restore_display_reg(struct drm_device *dev)
 		I915_WRITE(_PFA_WIN_SZ, dev_priv->regfile.savePFA_WIN_SZ);
 		I915_WRITE(_PFA_WIN_POS, dev_priv->regfile.savePFA_WIN_POS);
 
-		I915_WRITE(_TRANSACONF, dev_priv->regfile.saveTRANSACONF);
-		I915_WRITE(_TRANS_HTOTAL_A, dev_priv->regfile.saveTRANS_HTOTAL_A);
-		I915_WRITE(_TRANS_HBLANK_A, dev_priv->regfile.saveTRANS_HBLANK_A);
-		I915_WRITE(_TRANS_HSYNC_A, dev_priv->regfile.saveTRANS_HSYNC_A);
-		I915_WRITE(_TRANS_VTOTAL_A, dev_priv->regfile.saveTRANS_VTOTAL_A);
-		I915_WRITE(_TRANS_VBLANK_A, dev_priv->regfile.saveTRANS_VBLANK_A);
-		I915_WRITE(_TRANS_VSYNC_A, dev_priv->regfile.saveTRANS_VSYNC_A);
+		I915_WRITE(_PCH_TRANSACONF, dev_priv->regfile.saveTRANSACONF);
+		I915_WRITE(_PCH_TRANS_HTOTAL_A, dev_priv->regfile.saveTRANS_HTOTAL_A);
+		I915_WRITE(_PCH_TRANS_HBLANK_A, dev_priv->regfile.saveTRANS_HBLANK_A);
+		I915_WRITE(_PCH_TRANS_HSYNC_A, dev_priv->regfile.saveTRANS_HSYNC_A);
+		I915_WRITE(_PCH_TRANS_VTOTAL_A, dev_priv->regfile.saveTRANS_VTOTAL_A);
+		I915_WRITE(_PCH_TRANS_VBLANK_A, dev_priv->regfile.saveTRANS_VBLANK_A);
+		I915_WRITE(_PCH_TRANS_VSYNC_A, dev_priv->regfile.saveTRANS_VSYNC_A);
 	}
 
 	/* Restore plane info */
@@ -448,13 +448,13 @@ void i915_restore_display_reg(struct drm_device *dev)
 		I915_WRITE(_PFB_WIN_SZ, dev_priv->regfile.savePFB_WIN_SZ);
 		I915_WRITE(_PFB_WIN_POS, dev_priv->regfile.savePFB_WIN_POS);
 
-		I915_WRITE(_TRANSBCONF, dev_priv->regfile.saveTRANSBCONF);
-		I915_WRITE(_TRANS_HTOTAL_B, dev_priv->regfile.saveTRANS_HTOTAL_B);
-		I915_WRITE(_TRANS_HBLANK_B, dev_priv->regfile.saveTRANS_HBLANK_B);
-		I915_WRITE(_TRANS_HSYNC_B, dev_priv->regfile.saveTRANS_HSYNC_B);
-		I915_WRITE(_TRANS_VTOTAL_B, dev_priv->regfile.saveTRANS_VTOTAL_B);
-		I915_WRITE(_TRANS_VBLANK_B, dev_priv->regfile.saveTRANS_VBLANK_B);
-		I915_WRITE(_TRANS_VSYNC_B, dev_priv->regfile.saveTRANS_VSYNC_B);
+		I915_WRITE(_PCH_TRANSBCONF, dev_priv->regfile.saveTRANSBCONF);
+		I915_WRITE(_PCH_TRANS_HTOTAL_B, dev_priv->regfile.saveTRANS_HTOTAL_B);
+		I915_WRITE(_PCH_TRANS_HBLANK_B, dev_priv->regfile.saveTRANS_HBLANK_B);
+		I915_WRITE(_PCH_TRANS_HSYNC_B, dev_priv->regfile.saveTRANS_HSYNC_B);
+		I915_WRITE(_PCH_TRANS_VTOTAL_B, dev_priv->regfile.saveTRANS_VTOTAL_B);
+		I915_WRITE(_PCH_TRANS_VBLANK_B, dev_priv->regfile.saveTRANS_VBLANK_B);
+		I915_WRITE(_PCH_TRANS_VSYNC_B, dev_priv->regfile.saveTRANS_VSYNC_B);
 	}
 
 	/* Restore plane info */
diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c
index 95070b2124c6..53f2bed8bc5f 100644
--- a/drivers/gpu/drm/i915/intel_bios.c
+++ b/drivers/gpu/drm/i915/intel_bios.c
@@ -212,7 +212,7 @@ parse_lfp_panel_data(struct drm_i915_private *dev_priv,
 	if (!lvds_options)
 		return;
 
-	dev_priv->lvds_dither = lvds_options->pixel_dither;
+	dev_priv->vbt.lvds_dither = lvds_options->pixel_dither;
 	if (lvds_options->panel_type == 0xff)
 		return;
 
@@ -226,7 +226,7 @@ parse_lfp_panel_data(struct drm_i915_private *dev_priv,
 	if (!lvds_lfp_data_ptrs)
 		return;
 
-	dev_priv->lvds_vbt = 1;
+	dev_priv->vbt.lvds_vbt = 1;
 
 	panel_dvo_timing = get_lvds_dvo_timing(lvds_lfp_data,
 					       lvds_lfp_data_ptrs,
@@ -238,7 +238,7 @@ parse_lfp_panel_data(struct drm_i915_private *dev_priv,
 
 	fill_detail_timing_data(panel_fixed_mode, panel_dvo_timing);
 
-	dev_priv->lfp_lvds_vbt_mode = panel_fixed_mode;
+	dev_priv->vbt.lfp_lvds_vbt_mode = panel_fixed_mode;
 
 	DRM_DEBUG_KMS("Found panel mode in BIOS VBT tables:\n");
 	drm_mode_debug_printmodeline(panel_fixed_mode);
@@ -274,9 +274,9 @@ parse_lfp_panel_data(struct drm_i915_private *dev_priv,
 		/* check the resolution, just to be sure */
 		if (fp_timing->x_res == panel_fixed_mode->hdisplay &&
 		    fp_timing->y_res == panel_fixed_mode->vdisplay) {
-			dev_priv->bios_lvds_val = fp_timing->lvds_reg_val;
+			dev_priv->vbt.bios_lvds_val = fp_timing->lvds_reg_val;
 			DRM_DEBUG_KMS("VBT initial LVDS value %x\n",
-				      dev_priv->bios_lvds_val);
+				      dev_priv->vbt.bios_lvds_val);
 		}
 	}
 }
@@ -316,7 +316,7 @@ parse_sdvo_panel_data(struct drm_i915_private *dev_priv,
 
 	fill_detail_timing_data(panel_fixed_mode, dvo_timing + index);
 
-	dev_priv->sdvo_lvds_vbt_mode = panel_fixed_mode;
+	dev_priv->vbt.sdvo_lvds_vbt_mode = panel_fixed_mode;
 
 	DRM_DEBUG_KMS("Found SDVO panel mode in BIOS VBT tables:\n");
 	drm_mode_debug_printmodeline(panel_fixed_mode);
@@ -345,20 +345,20 @@ parse_general_features(struct drm_i915_private *dev_priv,
 
 	general = find_section(bdb, BDB_GENERAL_FEATURES);
 	if (general) {
-		dev_priv->int_tv_support = general->int_tv_support;
-		dev_priv->int_crt_support = general->int_crt_support;
-		dev_priv->lvds_use_ssc = general->enable_ssc;
-		dev_priv->lvds_ssc_freq =
+		dev_priv->vbt.int_tv_support = general->int_tv_support;
+		dev_priv->vbt.int_crt_support = general->int_crt_support;
+		dev_priv->vbt.lvds_use_ssc = general->enable_ssc;
+		dev_priv->vbt.lvds_ssc_freq =
 			intel_bios_ssc_frequency(dev, general->ssc_freq);
-		dev_priv->display_clock_mode = general->display_clock_mode;
-		dev_priv->fdi_rx_polarity_inverted = general->fdi_rx_polarity_inverted;
+		dev_priv->vbt.display_clock_mode = general->display_clock_mode;
+		dev_priv->vbt.fdi_rx_polarity_inverted = general->fdi_rx_polarity_inverted;
 		DRM_DEBUG_KMS("BDB_GENERAL_FEATURES int_tv_support %d int_crt_support %d lvds_use_ssc %d lvds_ssc_freq %d display_clock_mode %d fdi_rx_polarity_inverted %d\n",
-			      dev_priv->int_tv_support,
-			      dev_priv->int_crt_support,
-			      dev_priv->lvds_use_ssc,
-			      dev_priv->lvds_ssc_freq,
-			      dev_priv->display_clock_mode,
-			      dev_priv->fdi_rx_polarity_inverted);
+			      dev_priv->vbt.int_tv_support,
+			      dev_priv->vbt.int_crt_support,
+			      dev_priv->vbt.lvds_use_ssc,
+			      dev_priv->vbt.lvds_ssc_freq,
+			      dev_priv->vbt.display_clock_mode,
+			      dev_priv->vbt.fdi_rx_polarity_inverted);
 	}
 }
 
@@ -375,7 +375,7 @@ parse_general_definitions(struct drm_i915_private *dev_priv,
 			int bus_pin = general->crt_ddc_gmbus_pin;
 			DRM_DEBUG_KMS("crt_ddc_bus_pin: %d\n", bus_pin);
 			if (intel_gmbus_is_port_valid(bus_pin))
-				dev_priv->crt_ddc_pin = bus_pin;
+				dev_priv->vbt.crt_ddc_pin = bus_pin;
 		} else {
 			DRM_DEBUG_KMS("BDB_GD too small (%d). Invalid.\n",
 				      block_size);
@@ -486,7 +486,7 @@ parse_driver_features(struct drm_i915_private *dev_priv,
 
 	if (SUPPORTS_EDP(dev) &&
 	    driver->lvds_config == BDB_DRIVER_FEATURE_EDP)
-		dev_priv->edp.support = 1;
+		dev_priv->vbt.edp_support = 1;
 
 	if (driver->dual_frequency)
 		dev_priv->render_reclock_avail = true;
@@ -501,20 +501,20 @@ parse_edp(struct drm_i915_private *dev_priv, struct bdb_header *bdb)
 
 	edp = find_section(bdb, BDB_EDP);
 	if (!edp) {
-		if (SUPPORTS_EDP(dev_priv->dev) && dev_priv->edp.support)
+		if (SUPPORTS_EDP(dev_priv->dev) && dev_priv->vbt.edp_support)
 			DRM_DEBUG_KMS("No eDP BDB found but eDP panel supported.\n");
 		return;
 	}
 
 	switch ((edp->color_depth >> (panel_type * 2)) & 3) {
 	case EDP_18BPP:
-		dev_priv->edp.bpp = 18;
+		dev_priv->vbt.edp_bpp = 18;
 		break;
 	case EDP_24BPP:
-		dev_priv->edp.bpp = 24;
+		dev_priv->vbt.edp_bpp = 24;
 		break;
 	case EDP_30BPP:
-		dev_priv->edp.bpp = 30;
+		dev_priv->vbt.edp_bpp = 30;
 		break;
 	}
 
@@ -522,48 +522,48 @@ parse_edp(struct drm_i915_private *dev_priv, struct bdb_header *bdb)
 	edp_pps = &edp->power_seqs[panel_type];
 	edp_link_params = &edp->link_params[panel_type];
 
-	dev_priv->edp.pps = *edp_pps;
+	dev_priv->vbt.edp_pps = *edp_pps;
 
-	dev_priv->edp.rate = edp_link_params->rate ? DP_LINK_BW_2_7 :
+	dev_priv->vbt.edp_rate = edp_link_params->rate ? DP_LINK_BW_2_7 :
 		DP_LINK_BW_1_62;
 	switch (edp_link_params->lanes) {
 	case 0:
-		dev_priv->edp.lanes = 1;
+		dev_priv->vbt.edp_lanes = 1;
 		break;
 	case 1:
-		dev_priv->edp.lanes = 2;
+		dev_priv->vbt.edp_lanes = 2;
 		break;
 	case 3:
 	default:
-		dev_priv->edp.lanes = 4;
+		dev_priv->vbt.edp_lanes = 4;
 		break;
 	}
 	switch (edp_link_params->preemphasis) {
 	case 0:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_0;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_0;
 		break;
 	case 1:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_3_5;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_3_5;
 		break;
 	case 2:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_6;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_6;
 		break;
 	case 3:
-		dev_priv->edp.preemphasis = DP_TRAIN_PRE_EMPHASIS_9_5;
+		dev_priv->vbt.edp_preemphasis = DP_TRAIN_PRE_EMPHASIS_9_5;
 		break;
 	}
 	switch (edp_link_params->vswing) {
 	case 0:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_400;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_400;
 		break;
 	case 1:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_600;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_600;
 		break;
 	case 2:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_800;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_800;
 		break;
 	case 3:
-		dev_priv->edp.vswing = DP_TRAIN_VOLTAGE_SWING_1200;
+		dev_priv->vbt.edp_vswing = DP_TRAIN_VOLTAGE_SWING_1200;
 		break;
 	}
 }
@@ -611,13 +611,13 @@ parse_device_mapping(struct drm_i915_private *dev_priv,
 		DRM_DEBUG_KMS("no child dev is parsed from VBT\n");
 		return;
 	}
-	dev_priv->child_dev = kcalloc(count, sizeof(*p_child), GFP_KERNEL);
-	if (!dev_priv->child_dev) {
+	dev_priv->vbt.child_dev = kcalloc(count, sizeof(*p_child), GFP_KERNEL);
+	if (!dev_priv->vbt.child_dev) {
 		DRM_DEBUG_KMS("No memory space for child device\n");
 		return;
 	}
 
-	dev_priv->child_dev_num = count;
+	dev_priv->vbt.child_dev_num = count;
 	count = 0;
 	for (i = 0; i < child_device_num; i++) {
 		p_child = &(p_defs->devices[i]);
@@ -625,7 +625,7 @@ parse_device_mapping(struct drm_i915_private *dev_priv,
 			/* skip the device block if device type is invalid */
 			continue;
 		}
-		child_dev_ptr = dev_priv->child_dev + count;
+		child_dev_ptr = dev_priv->vbt.child_dev + count;
 		count++;
 		memcpy((void *)child_dev_ptr, (void *)p_child,
 					sizeof(*p_child));
@@ -638,23 +638,23 @@ init_vbt_defaults(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = dev_priv->dev;
 
-	dev_priv->crt_ddc_pin = GMBUS_PORT_VGADDC;
+	dev_priv->vbt.crt_ddc_pin = GMBUS_PORT_VGADDC;
 
 	/* LFP panel data */
-	dev_priv->lvds_dither = 1;
-	dev_priv->lvds_vbt = 0;
+	dev_priv->vbt.lvds_dither = 1;
+	dev_priv->vbt.lvds_vbt = 0;
 
 	/* SDVO panel data */
-	dev_priv->sdvo_lvds_vbt_mode = NULL;
+	dev_priv->vbt.sdvo_lvds_vbt_mode = NULL;
 
 	/* general features */
-	dev_priv->int_tv_support = 1;
-	dev_priv->int_crt_support = 1;
+	dev_priv->vbt.int_tv_support = 1;
+	dev_priv->vbt.int_crt_support = 1;
 
 	/* Default to using SSC */
-	dev_priv->lvds_use_ssc = 1;
-	dev_priv->lvds_ssc_freq = intel_bios_ssc_frequency(dev, 1);
-	DRM_DEBUG_KMS("Set default to SSC at %dMHz\n", dev_priv->lvds_ssc_freq);
+	dev_priv->vbt.lvds_use_ssc = 1;
+	dev_priv->vbt.lvds_ssc_freq = intel_bios_ssc_frequency(dev, 1);
+	DRM_DEBUG_KMS("Set default to SSC at %dMHz\n", dev_priv->vbt.lvds_ssc_freq);
 }
 
 static int __init intel_no_opregion_vbt_callback(const struct dmi_system_id *id)
diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index 58b4a53715cd..3acec8c48166 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -84,6 +84,28 @@ static bool intel_crt_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+static void intel_crt_get_config(struct intel_encoder *encoder,
+				 struct intel_crtc_config *pipe_config)
+{
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct intel_crt *crt = intel_encoder_to_crt(encoder);
+	u32 tmp, flags = 0;
+
+	tmp = I915_READ(crt->adpa_reg);
+
+	if (tmp & ADPA_HSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+
+	if (tmp & ADPA_VSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	pipe_config->adjusted_mode.flags |= flags;
+}
+
 /* Note: The caller is required to filter out dpms modes not supported by the
  * platform. */
 static void intel_crt_set_dpms(struct intel_encoder *encoder, int mode)
@@ -127,7 +149,7 @@ static void intel_enable_crt(struct intel_encoder *encoder)
 	intel_crt_set_dpms(encoder, crt->connector->base.dpms);
 }
 
-
+/* Special dpms function to support cloning between dvo/sdvo/crt. */
 static void intel_crt_dpms(struct drm_connector *connector, int mode)
 {
 	struct drm_device *dev = connector->dev;
@@ -158,6 +180,8 @@ static void intel_crt_dpms(struct drm_connector *connector, int mode)
 	else
 		encoder->connectors_active = true;
 
+	/* We call connector dpms manually below in case pipe dpms doesn't
+	 * change due to cloning. */
 	if (mode < old_dpms) {
 		/* From off to on, enable the pipe first. */
 		intel_crtc_update_dpms(crtc);
@@ -207,6 +231,10 @@ static bool intel_crt_compute_config(struct intel_encoder *encoder,
 	if (HAS_PCH_SPLIT(dev))
 		pipe_config->has_pch_encoder = true;
 
+	/* LPT FDI RX only supports 8bpc. */
+	if (HAS_PCH_LPT(dev))
+		pipe_config->pipe_bpp = 24;
+
 	return true;
 }
 
@@ -431,7 +459,7 @@ static bool intel_crt_detect_ddc(struct drm_connector *connector)
 
 	BUG_ON(crt->base.type != INTEL_OUTPUT_ANALOG);
 
-	i2c = intel_gmbus_get_adapter(dev_priv, dev_priv->crt_ddc_pin);
+	i2c = intel_gmbus_get_adapter(dev_priv, dev_priv->vbt.crt_ddc_pin);
 	edid = intel_crt_get_edid(connector, i2c);
 
 	if (edid) {
@@ -637,7 +665,7 @@ static int intel_crt_get_modes(struct drm_connector *connector)
 	int ret;
 	struct i2c_adapter *i2c;
 
-	i2c = intel_gmbus_get_adapter(dev_priv, dev_priv->crt_ddc_pin);
+	i2c = intel_gmbus_get_adapter(dev_priv, dev_priv->vbt.crt_ddc_pin);
 	ret = intel_crt_ddc_get_modes(connector, i2c);
 	if (ret || !IS_G4X(dev))
 		return ret;
@@ -774,6 +802,7 @@ void intel_crt_init(struct drm_device *dev)
 	crt->base.compute_config = intel_crt_compute_config;
 	crt->base.disable = intel_disable_crt;
 	crt->base.enable = intel_enable_crt;
+	crt->base.get_config = intel_crt_get_config;
 	if (I915_HAS_HOTPLUG(dev))
 		crt->base.hpd_pin = HPD_CRT;
 	if (HAS_DDI(dev))
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index fb961bb81903..b042ee5c4070 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -174,6 +174,8 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 	 * mode set "sequence for CRT port" document:
 	 * - TP1 to TP2 time with the default value
 	 * - FDI delay to 90h
+	 *
+	 * WaFDIAutoLinkSetTimingOverrride:hsw
 	 */
 	I915_WRITE(_FDI_RXA_MISC, FDI_RX_PWRDN_LANE1_VAL(2) |
 				  FDI_RX_PWRDN_LANE0_VAL(2) |
@@ -181,7 +183,8 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 
 	/* Enable the PCH Receiver FDI PLL */
 	rx_ctl_val = dev_priv->fdi_rx_config | FDI_RX_ENHANCE_FRAME_ENABLE |
-		     FDI_RX_PLL_ENABLE | ((intel_crtc->fdi_lanes - 1) << 19);
+		     FDI_RX_PLL_ENABLE |
+		     FDI_DP_PORT_WIDTH(intel_crtc->config.fdi_lanes);
 	I915_WRITE(_FDI_RXA_CTL, rx_ctl_val);
 	POSTING_READ(_FDI_RXA_CTL);
 	udelay(220);
@@ -209,7 +212,7 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 		 * port reversal bit */
 		I915_WRITE(DDI_BUF_CTL(PORT_E),
 			   DDI_BUF_CTL_ENABLE |
-			   ((intel_crtc->fdi_lanes - 1) << 1) |
+			   ((intel_crtc->config.fdi_lanes - 1) << 1) |
 			   hsw_ddi_buf_ctl_values[i / 2]);
 		POSTING_READ(DDI_BUF_CTL(PORT_E));
 
@@ -278,392 +281,6 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 	DRM_ERROR("FDI link training failed!\n");
 }
 
-/* WRPLL clock dividers */
-struct wrpll_tmds_clock {
-	u32 clock;
-	u16 p;		/* Post divider */
-	u16 n2;		/* Feedback divider */
-	u16 r2;		/* Reference divider */
-};
-
-/* Table of matching values for WRPLL clocks programming for each frequency.
- * The code assumes this table is sorted. */
-static const struct wrpll_tmds_clock wrpll_tmds_clock_table[] = {
-	{19750,	38,	25,	18},
-	{20000,	48,	32,	18},
-	{21000,	36,	21,	15},
-	{21912,	42,	29,	17},
-	{22000,	36,	22,	15},
-	{23000,	36,	23,	15},
-	{23500,	40,	40,	23},
-	{23750,	26,	16,	14},
-	{24000,	36,	24,	15},
-	{25000,	36,	25,	15},
-	{25175,	26,	40,	33},
-	{25200,	30,	21,	15},
-	{26000,	36,	26,	15},
-	{27000,	30,	21,	14},
-	{27027,	18,	100,	111},
-	{27500,	30,	29,	19},
-	{28000,	34,	30,	17},
-	{28320,	26,	30,	22},
-	{28322,	32,	42,	25},
-	{28750,	24,	23,	18},
-	{29000,	30,	29,	18},
-	{29750,	32,	30,	17},
-	{30000,	30,	25,	15},
-	{30750,	30,	41,	24},
-	{31000,	30,	31,	18},
-	{31500,	30,	28,	16},
-	{32000,	30,	32,	18},
-	{32500,	28,	32,	19},
-	{33000,	24,	22,	15},
-	{34000,	28,	30,	17},
-	{35000,	26,	32,	19},
-	{35500,	24,	30,	19},
-	{36000,	26,	26,	15},
-	{36750,	26,	46,	26},
-	{37000,	24,	23,	14},
-	{37762,	22,	40,	26},
-	{37800,	20,	21,	15},
-	{38000,	24,	27,	16},
-	{38250,	24,	34,	20},
-	{39000,	24,	26,	15},
-	{40000,	24,	32,	18},
-	{40500,	20,	21,	14},
-	{40541,	22,	147,	89},
-	{40750,	18,	19,	14},
-	{41000,	16,	17,	14},
-	{41500,	22,	44,	26},
-	{41540,	22,	44,	26},
-	{42000,	18,	21,	15},
-	{42500,	22,	45,	26},
-	{43000,	20,	43,	27},
-	{43163,	20,	24,	15},
-	{44000,	18,	22,	15},
-	{44900,	20,	108,	65},
-	{45000,	20,	25,	15},
-	{45250,	20,	52,	31},
-	{46000,	18,	23,	15},
-	{46750,	20,	45,	26},
-	{47000,	20,	40,	23},
-	{48000,	18,	24,	15},
-	{49000,	18,	49,	30},
-	{49500,	16,	22,	15},
-	{50000,	18,	25,	15},
-	{50500,	18,	32,	19},
-	{51000,	18,	34,	20},
-	{52000,	18,	26,	15},
-	{52406,	14,	34,	25},
-	{53000,	16,	22,	14},
-	{54000,	16,	24,	15},
-	{54054,	16,	173,	108},
-	{54500,	14,	24,	17},
-	{55000,	12,	22,	18},
-	{56000,	14,	45,	31},
-	{56250,	16,	25,	15},
-	{56750,	14,	25,	17},
-	{57000,	16,	27,	16},
-	{58000,	16,	43,	25},
-	{58250,	16,	38,	22},
-	{58750,	16,	40,	23},
-	{59000,	14,	26,	17},
-	{59341,	14,	40,	26},
-	{59400,	16,	44,	25},
-	{60000,	16,	32,	18},
-	{60500,	12,	39,	29},
-	{61000,	14,	49,	31},
-	{62000,	14,	37,	23},
-	{62250,	14,	42,	26},
-	{63000,	12,	21,	15},
-	{63500,	14,	28,	17},
-	{64000,	12,	27,	19},
-	{65000,	14,	32,	19},
-	{65250,	12,	29,	20},
-	{65500,	12,	32,	22},
-	{66000,	12,	22,	15},
-	{66667,	14,	38,	22},
-	{66750,	10,	21,	17},
-	{67000,	14,	33,	19},
-	{67750,	14,	58,	33},
-	{68000,	14,	30,	17},
-	{68179,	14,	46,	26},
-	{68250,	14,	46,	26},
-	{69000,	12,	23,	15},
-	{70000,	12,	28,	18},
-	{71000,	12,	30,	19},
-	{72000,	12,	24,	15},
-	{73000,	10,	23,	17},
-	{74000,	12,	23,	14},
-	{74176,	8,	100,	91},
-	{74250,	10,	22,	16},
-	{74481,	12,	43,	26},
-	{74500,	10,	29,	21},
-	{75000,	12,	25,	15},
-	{75250,	10,	39,	28},
-	{76000,	12,	27,	16},
-	{77000,	12,	53,	31},
-	{78000,	12,	26,	15},
-	{78750,	12,	28,	16},
-	{79000,	10,	38,	26},
-	{79500,	10,	28,	19},
-	{80000,	12,	32,	18},
-	{81000,	10,	21,	14},
-	{81081,	6,	100,	111},
-	{81624,	8,	29,	24},
-	{82000,	8,	17,	14},
-	{83000,	10,	40,	26},
-	{83950,	10,	28,	18},
-	{84000,	10,	28,	18},
-	{84750,	6,	16,	17},
-	{85000,	6,	17,	18},
-	{85250,	10,	30,	19},
-	{85750,	10,	27,	17},
-	{86000,	10,	43,	27},
-	{87000,	10,	29,	18},
-	{88000,	10,	44,	27},
-	{88500,	10,	41,	25},
-	{89000,	10,	28,	17},
-	{89012,	6,	90,	91},
-	{89100,	10,	33,	20},
-	{90000,	10,	25,	15},
-	{91000,	10,	32,	19},
-	{92000,	10,	46,	27},
-	{93000,	10,	31,	18},
-	{94000,	10,	40,	23},
-	{94500,	10,	28,	16},
-	{95000,	10,	44,	25},
-	{95654,	10,	39,	22},
-	{95750,	10,	39,	22},
-	{96000,	10,	32,	18},
-	{97000,	8,	23,	16},
-	{97750,	8,	42,	29},
-	{98000,	8,	45,	31},
-	{99000,	8,	22,	15},
-	{99750,	8,	34,	23},
-	{100000,	6,	20,	18},
-	{100500,	6,	19,	17},
-	{101000,	6,	37,	33},
-	{101250,	8,	21,	14},
-	{102000,	6,	17,	15},
-	{102250,	6,	25,	22},
-	{103000,	8,	29,	19},
-	{104000,	8,	37,	24},
-	{105000,	8,	28,	18},
-	{106000,	8,	22,	14},
-	{107000,	8,	46,	29},
-	{107214,	8,	27,	17},
-	{108000,	8,	24,	15},
-	{108108,	8,	173,	108},
-	{109000,	6,	23,	19},
-	{110000,	6,	22,	18},
-	{110013,	6,	22,	18},
-	{110250,	8,	49,	30},
-	{110500,	8,	36,	22},
-	{111000,	8,	23,	14},
-	{111264,	8,	150,	91},
-	{111375,	8,	33,	20},
-	{112000,	8,	63,	38},
-	{112500,	8,	25,	15},
-	{113100,	8,	57,	34},
-	{113309,	8,	42,	25},
-	{114000,	8,	27,	16},
-	{115000,	6,	23,	18},
-	{116000,	8,	43,	25},
-	{117000,	8,	26,	15},
-	{117500,	8,	40,	23},
-	{118000,	6,	38,	29},
-	{119000,	8,	30,	17},
-	{119500,	8,	46,	26},
-	{119651,	8,	39,	22},
-	{120000,	8,	32,	18},
-	{121000,	6,	39,	29},
-	{121250,	6,	31,	23},
-	{121750,	6,	23,	17},
-	{122000,	6,	42,	31},
-	{122614,	6,	30,	22},
-	{123000,	6,	41,	30},
-	{123379,	6,	37,	27},
-	{124000,	6,	51,	37},
-	{125000,	6,	25,	18},
-	{125250,	4,	13,	14},
-	{125750,	4,	27,	29},
-	{126000,	6,	21,	15},
-	{127000,	6,	24,	17},
-	{127250,	6,	41,	29},
-	{128000,	6,	27,	19},
-	{129000,	6,	43,	30},
-	{129859,	4,	25,	26},
-	{130000,	6,	26,	18},
-	{130250,	6,	42,	29},
-	{131000,	6,	32,	22},
-	{131500,	6,	38,	26},
-	{131850,	6,	41,	28},
-	{132000,	6,	22,	15},
-	{132750,	6,	28,	19},
-	{133000,	6,	34,	23},
-	{133330,	6,	37,	25},
-	{134000,	6,	61,	41},
-	{135000,	6,	21,	14},
-	{135250,	6,	167,	111},
-	{136000,	6,	62,	41},
-	{137000,	6,	35,	23},
-	{138000,	6,	23,	15},
-	{138500,	6,	40,	26},
-	{138750,	6,	37,	24},
-	{139000,	6,	34,	22},
-	{139050,	6,	34,	22},
-	{139054,	6,	34,	22},
-	{140000,	6,	28,	18},
-	{141000,	6,	36,	23},
-	{141500,	6,	22,	14},
-	{142000,	6,	30,	19},
-	{143000,	6,	27,	17},
-	{143472,	4,	17,	16},
-	{144000,	6,	24,	15},
-	{145000,	6,	29,	18},
-	{146000,	6,	47,	29},
-	{146250,	6,	26,	16},
-	{147000,	6,	49,	30},
-	{147891,	6,	23,	14},
-	{148000,	6,	23,	14},
-	{148250,	6,	28,	17},
-	{148352,	4,	100,	91},
-	{148500,	6,	33,	20},
-	{149000,	6,	48,	29},
-	{150000,	6,	25,	15},
-	{151000,	4,	19,	17},
-	{152000,	6,	27,	16},
-	{152280,	6,	44,	26},
-	{153000,	6,	34,	20},
-	{154000,	6,	53,	31},
-	{155000,	6,	31,	18},
-	{155250,	6,	50,	29},
-	{155750,	6,	45,	26},
-	{156000,	6,	26,	15},
-	{157000,	6,	61,	35},
-	{157500,	6,	28,	16},
-	{158000,	6,	65,	37},
-	{158250,	6,	44,	25},
-	{159000,	6,	53,	30},
-	{159500,	6,	39,	22},
-	{160000,	6,	32,	18},
-	{161000,	4,	31,	26},
-	{162000,	4,	18,	15},
-	{162162,	4,	131,	109},
-	{162500,	4,	53,	44},
-	{163000,	4,	29,	24},
-	{164000,	4,	17,	14},
-	{165000,	4,	22,	18},
-	{166000,	4,	32,	26},
-	{167000,	4,	26,	21},
-	{168000,	4,	46,	37},
-	{169000,	4,	104,	83},
-	{169128,	4,	64,	51},
-	{169500,	4,	39,	31},
-	{170000,	4,	34,	27},
-	{171000,	4,	19,	15},
-	{172000,	4,	51,	40},
-	{172750,	4,	32,	25},
-	{172800,	4,	32,	25},
-	{173000,	4,	41,	32},
-	{174000,	4,	49,	38},
-	{174787,	4,	22,	17},
-	{175000,	4,	35,	27},
-	{176000,	4,	30,	23},
-	{177000,	4,	38,	29},
-	{178000,	4,	29,	22},
-	{178500,	4,	37,	28},
-	{179000,	4,	53,	40},
-	{179500,	4,	73,	55},
-	{180000,	4,	20,	15},
-	{181000,	4,	55,	41},
-	{182000,	4,	31,	23},
-	{183000,	4,	42,	31},
-	{184000,	4,	30,	22},
-	{184750,	4,	26,	19},
-	{185000,	4,	37,	27},
-	{186000,	4,	51,	37},
-	{187000,	4,	36,	26},
-	{188000,	4,	32,	23},
-	{189000,	4,	21,	15},
-	{190000,	4,	38,	27},
-	{190960,	4,	41,	29},
-	{191000,	4,	41,	29},
-	{192000,	4,	27,	19},
-	{192250,	4,	37,	26},
-	{193000,	4,	20,	14},
-	{193250,	4,	53,	37},
-	{194000,	4,	23,	16},
-	{194208,	4,	23,	16},
-	{195000,	4,	26,	18},
-	{196000,	4,	45,	31},
-	{197000,	4,	35,	24},
-	{197750,	4,	41,	28},
-	{198000,	4,	22,	15},
-	{198500,	4,	25,	17},
-	{199000,	4,	28,	19},
-	{200000,	4,	37,	25},
-	{201000,	4,	61,	41},
-	{202000,	4,	112,	75},
-	{202500,	4,	21,	14},
-	{203000,	4,	146,	97},
-	{204000,	4,	62,	41},
-	{204750,	4,	44,	29},
-	{205000,	4,	38,	25},
-	{206000,	4,	29,	19},
-	{207000,	4,	23,	15},
-	{207500,	4,	40,	26},
-	{208000,	4,	37,	24},
-	{208900,	4,	48,	31},
-	{209000,	4,	48,	31},
-	{209250,	4,	31,	20},
-	{210000,	4,	28,	18},
-	{211000,	4,	25,	16},
-	{212000,	4,	22,	14},
-	{213000,	4,	30,	19},
-	{213750,	4,	38,	24},
-	{214000,	4,	46,	29},
-	{214750,	4,	35,	22},
-	{215000,	4,	43,	27},
-	{216000,	4,	24,	15},
-	{217000,	4,	37,	23},
-	{218000,	4,	42,	26},
-	{218250,	4,	42,	26},
-	{218750,	4,	34,	21},
-	{219000,	4,	47,	29},
-	{220000,	4,	44,	27},
-	{220640,	4,	49,	30},
-	{220750,	4,	36,	22},
-	{221000,	4,	36,	22},
-	{222000,	4,	23,	14},
-	{222525,	4,	28,	17},
-	{222750,	4,	33,	20},
-	{227000,	4,	37,	22},
-	{230250,	4,	29,	17},
-	{233500,	4,	38,	22},
-	{235000,	4,	40,	23},
-	{238000,	4,	30,	17},
-	{241500,	2,	17,	19},
-	{245250,	2,	20,	22},
-	{247750,	2,	22,	24},
-	{253250,	2,	15,	16},
-	{256250,	2,	18,	19},
-	{262500,	2,	31,	32},
-	{267250,	2,	66,	67},
-	{268500,	2,	94,	95},
-	{270000,	2,	14,	14},
-	{272500,	2,	77,	76},
-	{273750,	2,	57,	56},
-	{280750,	2,	24,	23},
-	{281250,	2,	23,	22},
-	{286000,	2,	17,	16},
-	{291750,	2,	26,	24},
-	{296703,	2,	56,	51},
-	{297000,	2,	22,	20},
-	{298000,	2,	21,	19},
-};
-
 static void intel_ddi_mode_set(struct drm_encoder *encoder,
 			       struct drm_display_mode *mode,
 			       struct drm_display_mode *adjusted_mode)
@@ -675,7 +292,7 @@ static void intel_ddi_mode_set(struct drm_encoder *encoder,
 	int pipe = intel_crtc->pipe;
 	int type = intel_encoder->type;
 
-	DRM_DEBUG_KMS("Preparing DDI mode for Haswell on port %c, pipe %c\n",
+	DRM_DEBUG_KMS("Preparing DDI mode on port %c, pipe %c\n",
 		      port_name(port), pipe_name(pipe));
 
 	intel_crtc->eld_vld = false;
@@ -684,24 +301,9 @@ static void intel_ddi_mode_set(struct drm_encoder *encoder,
 		struct intel_digital_port *intel_dig_port =
 			enc_to_dig_port(encoder);
 
-		intel_dp->DP = intel_dig_port->port_reversal |
+		intel_dp->DP = intel_dig_port->saved_port_bits |
 			       DDI_BUF_CTL_ENABLE | DDI_BUF_EMP_400MV_0DB_HSW;
-		switch (intel_dp->lane_count) {
-		case 1:
-			intel_dp->DP |= DDI_PORT_WIDTH_X1;
-			break;
-		case 2:
-			intel_dp->DP |= DDI_PORT_WIDTH_X2;
-			break;
-		case 4:
-			intel_dp->DP |= DDI_PORT_WIDTH_X4;
-			break;
-		default:
-			intel_dp->DP |= DDI_PORT_WIDTH_X4;
-			WARN(1, "Unexpected DP lane count %d\n",
-			     intel_dp->lane_count);
-			break;
-		}
+		intel_dp->DP |= DDI_PORT_WIDTH(intel_dp->lane_count);
 
 		if (intel_dp->has_audio) {
 			DRM_DEBUG_DRIVER("DP audio on pipe %c on DDI\n",
@@ -748,8 +350,8 @@ intel_ddi_get_crtc_encoder(struct drm_crtc *crtc)
 	}
 
 	if (num_encoders != 1)
-		WARN(1, "%d encoders on crtc for pipe %d\n", num_encoders,
-		     intel_crtc->pipe);
+		WARN(1, "%d encoders on crtc for pipe %c\n", num_encoders,
+		     pipe_name(intel_crtc->pipe));
 
 	BUG_ON(ret == NULL);
 	return ret;
@@ -802,30 +404,227 @@ void intel_ddi_put_crtc_pll(struct drm_crtc *crtc)
 	intel_crtc->ddi_pll_sel = PORT_CLK_SEL_NONE;
 }
 
-static void intel_ddi_calculate_wrpll(int clock, int *p, int *n2, int *r2)
+#define LC_FREQ 2700
+#define LC_FREQ_2K (LC_FREQ * 2000)
+
+#define P_MIN 2
+#define P_MAX 64
+#define P_INC 2
+
+/* Constraints for PLL good behavior */
+#define REF_MIN 48
+#define REF_MAX 400
+#define VCO_MIN 2400
+#define VCO_MAX 4800
+
+#define ABS_DIFF(a, b) ((a > b) ? (a - b) : (b - a))
+
+struct wrpll_rnp {
+	unsigned p, n2, r2;
+};
+
+static unsigned wrpll_get_budget_for_freq(int clock)
+{
+	unsigned budget;
+
+	switch (clock) {
+	case 25175000:
+	case 25200000:
+	case 27000000:
+	case 27027000:
+	case 37762500:
+	case 37800000:
+	case 40500000:
+	case 40541000:
+	case 54000000:
+	case 54054000:
+	case 59341000:
+	case 59400000:
+	case 72000000:
+	case 74176000:
+	case 74250000:
+	case 81000000:
+	case 81081000:
+	case 89012000:
+	case 89100000:
+	case 108000000:
+	case 108108000:
+	case 111264000:
+	case 111375000:
+	case 148352000:
+	case 148500000:
+	case 162000000:
+	case 162162000:
+	case 222525000:
+	case 222750000:
+	case 296703000:
+	case 297000000:
+		budget = 0;
+		break;
+	case 233500000:
+	case 245250000:
+	case 247750000:
+	case 253250000:
+	case 298000000:
+		budget = 1500;
+		break;
+	case 169128000:
+	case 169500000:
+	case 179500000:
+	case 202000000:
+		budget = 2000;
+		break;
+	case 256250000:
+	case 262500000:
+	case 270000000:
+	case 272500000:
+	case 273750000:
+	case 280750000:
+	case 281250000:
+	case 286000000:
+	case 291750000:
+		budget = 4000;
+		break;
+	case 267250000:
+	case 268500000:
+		budget = 5000;
+		break;
+	default:
+		budget = 1000;
+		break;
+	}
+
+	return budget;
+}
+
+static void wrpll_update_rnp(uint64_t freq2k, unsigned budget,
+			     unsigned r2, unsigned n2, unsigned p,
+			     struct wrpll_rnp *best)
 {
-	u32 i;
+	uint64_t a, b, c, d, diff, diff_best;
 
-	for (i = 0; i < ARRAY_SIZE(wrpll_tmds_clock_table); i++)
-		if (clock <= wrpll_tmds_clock_table[i].clock)
-			break;
+	/* No best (r,n,p) yet */
+	if (best->p == 0) {
+		best->p = p;
+		best->n2 = n2;
+		best->r2 = r2;
+		return;
+	}
+
+	/*
+	 * Output clock is (LC_FREQ_2K / 2000) * N / (P * R), which compares to
+	 * freq2k.
+	 *
+	 * delta = 1e6 *
+	 *	   abs(freq2k - (LC_FREQ_2K * n2/(p * r2))) /
+	 *	   freq2k;
+	 *
+	 * and we would like delta <= budget.
+	 *
+	 * If the discrepancy is above the PPM-based budget, always prefer to
+	 * improve upon the previous solution.  However, if you're within the
+	 * budget, try to maximize Ref * VCO, that is N / (P * R^2).
+	 */
+	a = freq2k * budget * p * r2;
+	b = freq2k * budget * best->p * best->r2;
+	diff = ABS_DIFF((freq2k * p * r2), (LC_FREQ_2K * n2));
+	diff_best = ABS_DIFF((freq2k * best->p * best->r2),
+			     (LC_FREQ_2K * best->n2));
+	c = 1000000 * diff;
+	d = 1000000 * diff_best;
+
+	if (a < c && b < d) {
+		/* If both are above the budget, pick the closer */
+		if (best->p * best->r2 * diff < p * r2 * diff_best) {
+			best->p = p;
+			best->n2 = n2;
+			best->r2 = r2;
+		}
+	} else if (a >= c && b < d) {
+		/* If A is below the threshold but B is above it?  Update. */
+		best->p = p;
+		best->n2 = n2;
+		best->r2 = r2;
+	} else if (a >= c && b >= d) {
+		/* Both are below the limit, so pick the higher n2/(r2*r2) */
+		if (n2 * best->r2 * best->r2 > best->n2 * r2 * r2) {
+			best->p = p;
+			best->n2 = n2;
+			best->r2 = r2;
+		}
+	}
+	/* Otherwise a < c && b >= d, do nothing */
+}
+
+static void
+intel_ddi_calculate_wrpll(int clock /* in Hz */,
+			  unsigned *r2_out, unsigned *n2_out, unsigned *p_out)
+{
+	uint64_t freq2k;
+	unsigned p, n2, r2;
+	struct wrpll_rnp best = { 0, 0, 0 };
+	unsigned budget;
 
-	if (i == ARRAY_SIZE(wrpll_tmds_clock_table))
-		i--;
+	freq2k = clock / 100;
 
-	*p = wrpll_tmds_clock_table[i].p;
-	*n2 = wrpll_tmds_clock_table[i].n2;
-	*r2 = wrpll_tmds_clock_table[i].r2;
+	budget = wrpll_get_budget_for_freq(clock);
 
-	if (wrpll_tmds_clock_table[i].clock != clock)
-		DRM_INFO("WRPLL: using settings for %dKHz on %dKHz mode\n",
-			 wrpll_tmds_clock_table[i].clock, clock);
+	/* Special case handling for 540 pixel clock: bypass WR PLL entirely
+	 * and directly pass the LC PLL to it. */
+	if (freq2k == 5400000) {
+		*n2_out = 2;
+		*p_out = 1;
+		*r2_out = 2;
+		return;
+	}
+
+	/*
+	 * Ref = LC_FREQ / R, where Ref is the actual reference input seen by
+	 * the WR PLL.
+	 *
+	 * We want R so that REF_MIN <= Ref <= REF_MAX.
+	 * Injecting R2 = 2 * R gives:
+	 *   REF_MAX * r2 > LC_FREQ * 2 and
+	 *   REF_MIN * r2 < LC_FREQ * 2
+	 *
+	 * Which means the desired boundaries for r2 are:
+	 *  LC_FREQ * 2 / REF_MAX < r2 < LC_FREQ * 2 / REF_MIN
+	 *
+	 */
+	for (r2 = LC_FREQ * 2 / REF_MAX + 1;
+	     r2 <= LC_FREQ * 2 / REF_MIN;
+	     r2++) {
+
+		/*
+		 * VCO = N * Ref, that is: VCO = N * LC_FREQ / R
+		 *
+		 * Once again we want VCO_MIN <= VCO <= VCO_MAX.
+		 * Injecting R2 = 2 * R and N2 = 2 * N, we get:
+		 *   VCO_MAX * r2 > n2 * LC_FREQ and
+		 *   VCO_MIN * r2 < n2 * LC_FREQ)
+		 *
+		 * Which means the desired boundaries for n2 are:
+		 * VCO_MIN * r2 / LC_FREQ < n2 < VCO_MAX * r2 / LC_FREQ
+		 */
+		for (n2 = VCO_MIN * r2 / LC_FREQ + 1;
+		     n2 <= VCO_MAX * r2 / LC_FREQ;
+		     n2++) {
 
-	DRM_DEBUG_KMS("WRPLL: %dKHz refresh rate with p=%d, n2=%d r2=%d\n",
-		      clock, *p, *n2, *r2);
+			for (p = P_MIN; p <= P_MAX; p += P_INC)
+				wrpll_update_rnp(freq2k, budget,
+						 r2, n2, p, &best);
+		}
+	}
+
+	*n2_out = best.n2;
+	*p_out = best.p;
+	*r2_out = best.r2;
+
+	DRM_DEBUG_KMS("WRPLL: %dHz refresh rate with p=%d, n2=%d r2=%d\n",
+		      clock, *p_out, *n2_out, *r2_out);
 }
 
-bool intel_ddi_pll_mode_set(struct drm_crtc *crtc, int clock)
+bool intel_ddi_pll_mode_set(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
@@ -835,6 +634,7 @@ bool intel_ddi_pll_mode_set(struct drm_crtc *crtc, int clock)
 	int type = intel_encoder->type;
 	enum pipe pipe = intel_crtc->pipe;
 	uint32_t reg, val;
+	int clock = intel_crtc->config.port_clock;
 
 	/* TODO: reuse PLLs when possible (compare values) */
 
@@ -863,7 +663,7 @@ bool intel_ddi_pll_mode_set(struct drm_crtc *crtc, int clock)
 		return true;
 
 	} else if (type == INTEL_OUTPUT_HDMI) {
-		int p, n2, r2;
+		unsigned p, n2, r2;
 
 		if (plls->wrpll1_refcount == 0) {
 			DRM_DEBUG_KMS("Using WRPLL 1 on pipe %c\n",
@@ -885,7 +685,7 @@ bool intel_ddi_pll_mode_set(struct drm_crtc *crtc, int clock)
 		WARN(I915_READ(reg) & WRPLL_PLL_ENABLE,
 		     "WRPLL already enabled\n");
 
-		intel_ddi_calculate_wrpll(clock, &p, &n2, &r2);
+		intel_ddi_calculate_wrpll(clock * 1000, &r2, &n2, &p);
 
 		val = WRPLL_PLL_ENABLE | WRPLL_PLL_SELECT_LCPLL_2700 |
 		      WRPLL_DIVIDER_REFERENCE(r2) | WRPLL_DIVIDER_FEEDBACK(n2) |
@@ -995,7 +795,7 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 			/* Can only use the always-on power well for eDP when
 			 * not using the panel fitter, and when not using motion
 			  * blur mitigation (which we don't support). */
-			if (dev_priv->pch_pf_size)
+			if (intel_crtc->config.pch_pfit.size)
 				temp |= TRANS_DDI_EDP_INPUT_A_ONOFF;
 			else
 				temp |= TRANS_DDI_EDP_INPUT_A_ON;
@@ -1022,7 +822,7 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 
 	} else if (type == INTEL_OUTPUT_ANALOG) {
 		temp |= TRANS_DDI_MODE_SELECT_FDI;
-		temp |= (intel_crtc->fdi_lanes - 1) << 1;
+		temp |= (intel_crtc->config.fdi_lanes - 1) << 1;
 
 	} else if (type == INTEL_OUTPUT_DISPLAYPORT ||
 		   type == INTEL_OUTPUT_EDP) {
@@ -1030,25 +830,10 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 
 		temp |= TRANS_DDI_MODE_SELECT_DP_SST;
 
-		switch (intel_dp->lane_count) {
-		case 1:
-			temp |= TRANS_DDI_PORT_WIDTH_X1;
-			break;
-		case 2:
-			temp |= TRANS_DDI_PORT_WIDTH_X2;
-			break;
-		case 4:
-			temp |= TRANS_DDI_PORT_WIDTH_X4;
-			break;
-		default:
-			temp |= TRANS_DDI_PORT_WIDTH_X4;
-			WARN(1, "Unsupported lane count %d\n",
-			     intel_dp->lane_count);
-		}
-
+		temp |= DDI_PORT_WIDTH(intel_dp->lane_count);
 	} else {
-		WARN(1, "Invalid encoder type %d for pipe %d\n",
-		     intel_encoder->type, pipe);
+		WARN(1, "Invalid encoder type %d for pipe %c\n",
+		     intel_encoder->type, pipe_name(pipe));
 	}
 
 	I915_WRITE(TRANS_DDI_FUNC_CTL(cpu_transcoder), temp);
@@ -1148,7 +933,7 @@ bool intel_ddi_get_hw_state(struct intel_encoder *encoder,
 		}
 	}
 
-	DRM_DEBUG_KMS("No pipe for ddi port %i found\n", port);
+	DRM_DEBUG_KMS("No pipe for ddi port %c found\n", port_name(port));
 
 	return false;
 }
@@ -1324,7 +1109,8 @@ static void intel_enable_ddi(struct intel_encoder *intel_encoder)
 		 * enabling the port.
 		 */
 		I915_WRITE(DDI_BUF_CTL(port),
-			   intel_dig_port->port_reversal | DDI_BUF_CTL_ENABLE);
+			   intel_dig_port->saved_port_bits |
+			   DDI_BUF_CTL_ENABLE);
 	} else if (type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 
@@ -1334,7 +1120,7 @@ static void intel_enable_ddi(struct intel_encoder *intel_encoder)
 		ironlake_edp_backlight_on(intel_dp);
 	}
 
-	if (intel_crtc->eld_vld) {
+	if (intel_crtc->eld_vld && type != INTEL_OUTPUT_EDP) {
 		tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
 		tmp |= ((AUDIO_OUTPUT_ENABLE_A | AUDIO_ELD_VALID_A) << (pipe * 4));
 		I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
@@ -1352,9 +1138,12 @@ static void intel_disable_ddi(struct intel_encoder *intel_encoder)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t tmp;
 
-	tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
-	tmp &= ~((AUDIO_OUTPUT_ENABLE_A | AUDIO_ELD_VALID_A) << (pipe * 4));
-	I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
+	if (intel_crtc->eld_vld && type != INTEL_OUTPUT_EDP) {
+		tmp = I915_READ(HSW_AUD_PIN_ELD_CP_VLD);
+		tmp &= ~((AUDIO_OUTPUT_ENABLE_A | AUDIO_ELD_VALID_A) <<
+			 (pipe * 4));
+		I915_WRITE(HSW_AUD_PIN_ELD_CP_VLD, tmp);
+	}
 
 	if (type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
@@ -1366,14 +1155,14 @@ static void intel_disable_ddi(struct intel_encoder *intel_encoder)
 int intel_ddi_get_cdclk_freq(struct drm_i915_private *dev_priv)
 {
 	if (I915_READ(HSW_FUSE_STRAP) & HSW_CDCLK_LIMIT)
-		return 450;
+		return 450000;
 	else if ((I915_READ(LCPLL_CTL) & LCPLL_CLK_FREQ_MASK) ==
 		 LCPLL_CLK_FREQ_450)
-		return 450;
+		return 450000;
 	else if (IS_ULT(dev_priv->dev))
-		return 338;
+		return 337500;
 	else
-		return 540;
+		return 540000;
 }
 
 void intel_ddi_pll_init(struct drm_device *dev)
@@ -1386,7 +1175,7 @@ void intel_ddi_pll_init(struct drm_device *dev)
 	 * Don't even try to turn it on.
 	 */
 
-	DRM_DEBUG_KMS("CDCLK running at %dMHz\n",
+	DRM_DEBUG_KMS("CDCLK running at %dKHz\n",
 		      intel_ddi_get_cdclk_freq(dev_priv));
 
 	if (val & LCPLL_CD_SOURCE_FCLK)
@@ -1472,6 +1261,27 @@ static void intel_ddi_hot_plug(struct intel_encoder *intel_encoder)
 		intel_dp_check_link_status(intel_dp);
 }
 
+static void intel_ddi_get_config(struct intel_encoder *encoder,
+				 struct intel_crtc_config *pipe_config)
+{
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
+	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
+	u32 temp, flags = 0;
+
+	temp = I915_READ(TRANS_DDI_FUNC_CTL(cpu_transcoder));
+	if (temp & TRANS_DDI_PHSYNC)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+	if (temp & TRANS_DDI_PVSYNC)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	pipe_config->adjusted_mode.flags |= flags;
+}
+
 static void intel_ddi_destroy(struct drm_encoder *encoder)
 {
 	/* HDMI has nothing special to destroy, so we can go with this. */
@@ -1482,9 +1292,13 @@ static bool intel_ddi_compute_config(struct intel_encoder *encoder,
 				     struct intel_crtc_config *pipe_config)
 {
 	int type = encoder->type;
+	int port = intel_ddi_get_encoder_port(encoder);
 
 	WARN(type == INTEL_OUTPUT_UNKNOWN, "compute_config() on unknown output!\n");
 
+	if (port == PORT_A)
+		pipe_config->cpu_transcoder = TRANSCODER_EDP;
+
 	if (type == INTEL_OUTPUT_HDMI)
 		return intel_hdmi_compute_config(encoder, pipe_config);
 	else
@@ -1518,16 +1332,6 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
 		return;
 	}
 
-	if (port != PORT_A) {
-		hdmi_connector = kzalloc(sizeof(struct intel_connector),
-					 GFP_KERNEL);
-		if (!hdmi_connector) {
-			kfree(dp_connector);
-			kfree(intel_dig_port);
-			return;
-		}
-	}
-
 	intel_encoder = &intel_dig_port->base;
 	encoder = &intel_encoder->base;
 
@@ -1541,12 +1345,12 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
 	intel_encoder->disable = intel_disable_ddi;
 	intel_encoder->post_disable = intel_ddi_post_disable;
 	intel_encoder->get_hw_state = intel_ddi_get_hw_state;
+	intel_encoder->get_config = intel_ddi_get_config;
 
 	intel_dig_port->port = port;
-	intel_dig_port->port_reversal = I915_READ(DDI_BUF_CTL(port)) &
-					DDI_BUF_PORT_REVERSAL;
-	if (hdmi_connector)
-		intel_dig_port->hdmi.hdmi_reg = DDI_BUF_CTL(port);
+	intel_dig_port->saved_port_bits = I915_READ(DDI_BUF_CTL(port)) &
+					  (DDI_BUF_PORT_REVERSAL |
+					   DDI_A_4_LANES);
 	intel_dig_port->dp.output_reg = DDI_BUF_CTL(port);
 
 	intel_encoder->type = INTEL_OUTPUT_UNKNOWN;
@@ -1554,7 +1358,21 @@ void intel_ddi_init(struct drm_device *dev, enum port port)
 	intel_encoder->cloneable = false;
 	intel_encoder->hot_plug = intel_ddi_hot_plug;
 
-	if (hdmi_connector)
+	if (!intel_dp_init_connector(intel_dig_port, dp_connector)) {
+		drm_encoder_cleanup(encoder);
+		kfree(intel_dig_port);
+		kfree(dp_connector);
+		return;
+	}
+
+	if (intel_encoder->type != INTEL_OUTPUT_EDP) {
+		hdmi_connector = kzalloc(sizeof(struct intel_connector),
+					 GFP_KERNEL);
+		if (!hdmi_connector) {
+			return;
+		}
+
+		intel_dig_port->hdmi.hdmi_reg = DDI_BUF_CTL(port);
 		intel_hdmi_init_connector(intel_dig_port, hdmi_connector);
-	intel_dp_init_connector(intel_dig_port, dp_connector);
+	}
 }
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 56746dcac40f..5fb305840db8 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -46,18 +46,6 @@ static void intel_increase_pllclock(struct drm_crtc *crtc);
 static void intel_crtc_update_cursor(struct drm_crtc *crtc, bool on);
 
 typedef struct {
-	/* given values */
-	int n;
-	int m1, m2;
-	int p1, p2;
-	/* derived values */
-	int	dot;
-	int	vco;
-	int	m;
-	int	p;
-} intel_clock_t;
-
-typedef struct {
 	int	min, max;
 } intel_range_t;
 
@@ -71,24 +59,6 @@ typedef struct intel_limit intel_limit_t;
 struct intel_limit {
 	intel_range_t   dot, vco, n, m, m1, m2, p, p1;
 	intel_p2_t	    p2;
-	/**
-	 * find_pll() - Find the best values for the PLL
-	 * @limit: limits for the PLL
-	 * @crtc: current CRTC
-	 * @target: target frequency in kHz
-	 * @refclk: reference clock frequency in kHz
-	 * @match_clock: if provided, @best_clock P divider must
-	 *               match the P divider from @match_clock
-	 *               used for LVDS downclocking
-	 * @best_clock: best PLL values found
-	 *
-	 * Returns true on success, false on failure.
-	 */
-	bool (*find_pll)(const intel_limit_t *limit,
-			 struct drm_crtc *crtc,
-			 int target, int refclk,
-			 intel_clock_t *match_clock,
-			 intel_clock_t *best_clock);
 };
 
 /* FDI */
@@ -104,29 +74,6 @@ intel_pch_rawclk(struct drm_device *dev)
 	return I915_READ(PCH_RAWCLK_FREQ) & RAWCLK_FREQ_MASK;
 }
 
-static bool
-intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-		    int target, int refclk, intel_clock_t *match_clock,
-		    intel_clock_t *best_clock);
-static bool
-intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-			int target, int refclk, intel_clock_t *match_clock,
-			intel_clock_t *best_clock);
-
-static bool
-intel_find_pll_g4x_dp(const intel_limit_t *, struct drm_crtc *crtc,
-		      int target, int refclk, intel_clock_t *match_clock,
-		      intel_clock_t *best_clock);
-static bool
-intel_find_pll_ironlake_dp(const intel_limit_t *, struct drm_crtc *crtc,
-			   int target, int refclk, intel_clock_t *match_clock,
-			   intel_clock_t *best_clock);
-
-static bool
-intel_vlv_find_best_pll(const intel_limit_t *limit, struct drm_crtc *crtc,
-			int target, int refclk, intel_clock_t *match_clock,
-			intel_clock_t *best_clock);
-
 static inline u32 /* units of 100MHz */
 intel_fdi_link_freq(struct drm_device *dev)
 {
@@ -148,7 +95,6 @@ static const intel_limit_t intel_limits_i8xx_dvo = {
 	.p1 = { .min = 2, .max = 33 },
 	.p2 = { .dot_limit = 165000,
 		.p2_slow = 4, .p2_fast = 2 },
-	.find_pll = intel_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_i8xx_lvds = {
@@ -162,7 +108,6 @@ static const intel_limit_t intel_limits_i8xx_lvds = {
 	.p1 = { .min = 1, .max = 6 },
 	.p2 = { .dot_limit = 165000,
 		.p2_slow = 14, .p2_fast = 7 },
-	.find_pll = intel_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_i9xx_sdvo = {
@@ -176,7 +121,6 @@ static const intel_limit_t intel_limits_i9xx_sdvo = {
 	.p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 200000,
 		.p2_slow = 10, .p2_fast = 5 },
-	.find_pll = intel_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_i9xx_lvds = {
@@ -190,7 +134,6 @@ static const intel_limit_t intel_limits_i9xx_lvds = {
 	.p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 112000,
 		.p2_slow = 14, .p2_fast = 7 },
-	.find_pll = intel_find_best_PLL,
 };
 
 
@@ -207,7 +150,6 @@ static const intel_limit_t intel_limits_g4x_sdvo = {
 		.p2_slow = 10,
 		.p2_fast = 10
 	},
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_g4x_hdmi = {
@@ -221,7 +163,6 @@ static const intel_limit_t intel_limits_g4x_hdmi = {
 	.p1 = { .min = 1, .max = 8},
 	.p2 = { .dot_limit = 165000,
 		.p2_slow = 10, .p2_fast = 5 },
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_g4x_single_channel_lvds = {
@@ -236,7 +177,6 @@ static const intel_limit_t intel_limits_g4x_single_channel_lvds = {
 	.p2 = { .dot_limit = 0,
 		.p2_slow = 14, .p2_fast = 14
 	},
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_g4x_dual_channel_lvds = {
@@ -251,21 +191,6 @@ static const intel_limit_t intel_limits_g4x_dual_channel_lvds = {
 	.p2 = { .dot_limit = 0,
 		.p2_slow = 7, .p2_fast = 7
 	},
-	.find_pll = intel_g4x_find_best_PLL,
-};
-
-static const intel_limit_t intel_limits_g4x_display_port = {
-	.dot = { .min = 161670, .max = 227000 },
-	.vco = { .min = 1750000, .max = 3500000},
-	.n = { .min = 1, .max = 2 },
-	.m = { .min = 97, .max = 108 },
-	.m1 = { .min = 0x10, .max = 0x12 },
-	.m2 = { .min = 0x05, .max = 0x06 },
-	.p = { .min = 10, .max = 20 },
-	.p1 = { .min = 1, .max = 2},
-	.p2 = { .dot_limit = 0,
-		.p2_slow = 10, .p2_fast = 10 },
-	.find_pll = intel_find_pll_g4x_dp,
 };
 
 static const intel_limit_t intel_limits_pineview_sdvo = {
@@ -281,7 +206,6 @@ static const intel_limit_t intel_limits_pineview_sdvo = {
 	.p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 200000,
 		.p2_slow = 10, .p2_fast = 5 },
-	.find_pll = intel_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_pineview_lvds = {
@@ -295,7 +219,6 @@ static const intel_limit_t intel_limits_pineview_lvds = {
 	.p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 112000,
 		.p2_slow = 14, .p2_fast = 14 },
-	.find_pll = intel_find_best_PLL,
 };
 
 /* Ironlake / Sandybridge
@@ -314,7 +237,6 @@ static const intel_limit_t intel_limits_ironlake_dac = {
 	.p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 225000,
 		.p2_slow = 10, .p2_fast = 5 },
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_ironlake_single_lvds = {
@@ -328,7 +250,6 @@ static const intel_limit_t intel_limits_ironlake_single_lvds = {
 	.p1 = { .min = 2, .max = 8 },
 	.p2 = { .dot_limit = 225000,
 		.p2_slow = 14, .p2_fast = 14 },
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_ironlake_dual_lvds = {
@@ -342,7 +263,6 @@ static const intel_limit_t intel_limits_ironlake_dual_lvds = {
 	.p1 = { .min = 2, .max = 8 },
 	.p2 = { .dot_limit = 225000,
 		.p2_slow = 7, .p2_fast = 7 },
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 /* LVDS 100mhz refclk limits. */
@@ -357,7 +277,6 @@ static const intel_limit_t intel_limits_ironlake_single_lvds_100m = {
 	.p1 = { .min = 2, .max = 8 },
 	.p2 = { .dot_limit = 225000,
 		.p2_slow = 14, .p2_fast = 14 },
-	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_ironlake_dual_lvds_100m = {
@@ -371,21 +290,6 @@ static const intel_limit_t intel_limits_ironlake_dual_lvds_100m = {
 	.p1 = { .min = 2, .max = 6 },
 	.p2 = { .dot_limit = 225000,
 		.p2_slow = 7, .p2_fast = 7 },
-	.find_pll = intel_g4x_find_best_PLL,
-};
-
-static const intel_limit_t intel_limits_ironlake_display_port = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 1760000, .max = 3510000},
-	.n = { .min = 1, .max = 2 },
-	.m = { .min = 81, .max = 90 },
-	.m1 = { .min = 12, .max = 22 },
-	.m2 = { .min = 5, .max = 9 },
-	.p = { .min = 10, .max = 20 },
-	.p1 = { .min = 1, .max = 2},
-	.p2 = { .dot_limit = 0,
-		.p2_slow = 10, .p2_fast = 10 },
-	.find_pll = intel_find_pll_ironlake_dp,
 };
 
 static const intel_limit_t intel_limits_vlv_dac = {
@@ -396,15 +300,14 @@ static const intel_limit_t intel_limits_vlv_dac = {
 	.m1 = { .min = 2, .max = 3 },
 	.m2 = { .min = 11, .max = 156 },
 	.p = { .min = 10, .max = 30 },
-	.p1 = { .min = 2, .max = 3 },
+	.p1 = { .min = 1, .max = 3 },
 	.p2 = { .dot_limit = 270000,
 		.p2_slow = 2, .p2_fast = 20 },
-	.find_pll = intel_vlv_find_best_pll,
 };
 
 static const intel_limit_t intel_limits_vlv_hdmi = {
-	.dot = { .min = 20000, .max = 165000 },
-	.vco = { .min = 4000000, .max = 5994000},
+	.dot = { .min = 25000, .max = 270000 },
+	.vco = { .min = 4000000, .max = 6000000 },
 	.n = { .min = 1, .max = 7 },
 	.m = { .min = 60, .max = 300 }, /* guess */
 	.m1 = { .min = 2, .max = 3 },
@@ -413,7 +316,6 @@ static const intel_limit_t intel_limits_vlv_hdmi = {
 	.p1 = { .min = 2, .max = 3 },
 	.p2 = { .dot_limit = 270000,
 		.p2_slow = 2, .p2_fast = 20 },
-	.find_pll = intel_vlv_find_best_pll,
 };
 
 static const intel_limit_t intel_limits_vlv_dp = {
@@ -424,61 +326,11 @@ static const intel_limit_t intel_limits_vlv_dp = {
 	.m1 = { .min = 2, .max = 3 },
 	.m2 = { .min = 11, .max = 156 },
 	.p = { .min = 10, .max = 30 },
-	.p1 = { .min = 2, .max = 3 },
+	.p1 = { .min = 1, .max = 3 },
 	.p2 = { .dot_limit = 270000,
 		.p2_slow = 2, .p2_fast = 20 },
-	.find_pll = intel_vlv_find_best_pll,
 };
 
-u32 intel_dpio_read(struct drm_i915_private *dev_priv, int reg)
-{
-	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
-
-	if (wait_for_atomic_us((I915_READ(DPIO_PKT) & DPIO_BUSY) == 0, 100)) {
-		DRM_ERROR("DPIO idle wait timed out\n");
-		return 0;
-	}
-
-	I915_WRITE(DPIO_REG, reg);
-	I915_WRITE(DPIO_PKT, DPIO_RID | DPIO_OP_READ | DPIO_PORTID |
-		   DPIO_BYTE);
-	if (wait_for_atomic_us((I915_READ(DPIO_PKT) & DPIO_BUSY) == 0, 100)) {
-		DRM_ERROR("DPIO read wait timed out\n");
-		return 0;
-	}
-
-	return I915_READ(DPIO_DATA);
-}
-
-static void intel_dpio_write(struct drm_i915_private *dev_priv, int reg,
-			     u32 val)
-{
-	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
-
-	if (wait_for_atomic_us((I915_READ(DPIO_PKT) & DPIO_BUSY) == 0, 100)) {
-		DRM_ERROR("DPIO idle wait timed out\n");
-		return;
-	}
-
-	I915_WRITE(DPIO_DATA, val);
-	I915_WRITE(DPIO_REG, reg);
-	I915_WRITE(DPIO_PKT, DPIO_RID | DPIO_OP_WRITE | DPIO_PORTID |
-		   DPIO_BYTE);
-	if (wait_for_atomic_us((I915_READ(DPIO_PKT) & DPIO_BUSY) == 0, 100))
-		DRM_ERROR("DPIO write wait timed out\n");
-}
-
-static void vlv_init_dpio(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	/* Reset the DPIO config */
-	I915_WRITE(DPIO_CTL, 0);
-	POSTING_READ(DPIO_CTL);
-	I915_WRITE(DPIO_CTL, 1);
-	POSTING_READ(DPIO_CTL);
-}
-
 static const intel_limit_t *intel_ironlake_limit(struct drm_crtc *crtc,
 						int refclk)
 {
@@ -497,10 +349,7 @@ static const intel_limit_t *intel_ironlake_limit(struct drm_crtc *crtc,
 			else
 				limit = &intel_limits_ironlake_single_lvds;
 		}
-	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT) ||
-		   intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))
-		limit = &intel_limits_ironlake_display_port;
-	else
+	} else
 		limit = &intel_limits_ironlake_dac;
 
 	return limit;
@@ -521,8 +370,6 @@ static const intel_limit_t *intel_g4x_limit(struct drm_crtc *crtc)
 		limit = &intel_limits_g4x_hdmi;
 	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_SDVO)) {
 		limit = &intel_limits_g4x_sdvo;
-	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)) {
-		limit = &intel_limits_g4x_display_port;
 	} else /* The option is for other outputs */
 		limit = &intel_limits_i9xx_sdvo;
 
@@ -573,13 +420,14 @@ static void pineview_clock(int refclk, intel_clock_t *clock)
 	clock->dot = clock->vco / clock->p;
 }
 
-static void intel_clock(struct drm_device *dev, int refclk, intel_clock_t *clock)
+static uint32_t i9xx_dpll_compute_m(struct dpll *dpll)
 {
-	if (IS_PINEVIEW(dev)) {
-		pineview_clock(refclk, clock);
-		return;
-	}
-	clock->m = 5 * (clock->m1 + 2) + (clock->m2 + 2);
+	return 5 * (dpll->m1 + 2) + (dpll->m2 + 2);
+}
+
+static void i9xx_clock(int refclk, intel_clock_t *clock)
+{
+	clock->m = i9xx_dpll_compute_m(clock);
 	clock->p = clock->p1 * clock->p2;
 	clock->vco = refclk * clock->m / (clock->n + 2);
 	clock->dot = clock->vco / clock->p;
@@ -636,10 +484,9 @@ static bool intel_PLL_is_valid(struct drm_device *dev,
 }
 
 static bool
-intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
+i9xx_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
 		    int target, int refclk, intel_clock_t *match_clock,
 		    intel_clock_t *best_clock)
-
 {
 	struct drm_device *dev = crtc->dev;
 	intel_clock_t clock;
@@ -668,8 +515,7 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 	     clock.m1++) {
 		for (clock.m2 = limit->m2.min;
 		     clock.m2 <= limit->m2.max; clock.m2++) {
-			/* m1 is always 0 in Pineview */
-			if (clock.m2 >= clock.m1 && !IS_PINEVIEW(dev))
+			if (clock.m2 >= clock.m1)
 				break;
 			for (clock.n = limit->n.min;
 			     clock.n <= limit->n.max; clock.n++) {
@@ -677,7 +523,66 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 					clock.p1 <= limit->p1.max; clock.p1++) {
 					int this_err;
 
-					intel_clock(dev, refclk, &clock);
+					i9xx_clock(refclk, &clock);
+					if (!intel_PLL_is_valid(dev, limit,
+								&clock))
+						continue;
+					if (match_clock &&
+					    clock.p != match_clock->p)
+						continue;
+
+					this_err = abs(clock.dot - target);
+					if (this_err < err) {
+						*best_clock = clock;
+						err = this_err;
+					}
+				}
+			}
+		}
+	}
+
+	return (err != target);
+}
+
+static bool
+pnv_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
+		   int target, int refclk, intel_clock_t *match_clock,
+		   intel_clock_t *best_clock)
+{
+	struct drm_device *dev = crtc->dev;
+	intel_clock_t clock;
+	int err = target;
+
+	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
+		/*
+		 * For LVDS just rely on its current settings for dual-channel.
+		 * We haven't figured out how to reliably set up different
+		 * single/dual channel state, if we even can.
+		 */
+		if (intel_is_dual_link_lvds(dev))
+			clock.p2 = limit->p2.p2_fast;
+		else
+			clock.p2 = limit->p2.p2_slow;
+	} else {
+		if (target < limit->p2.dot_limit)
+			clock.p2 = limit->p2.p2_slow;
+		else
+			clock.p2 = limit->p2.p2_fast;
+	}
+
+	memset(best_clock, 0, sizeof(*best_clock));
+
+	for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max;
+	     clock.m1++) {
+		for (clock.m2 = limit->m2.min;
+		     clock.m2 <= limit->m2.max; clock.m2++) {
+			for (clock.n = limit->n.min;
+			     clock.n <= limit->n.max; clock.n++) {
+				for (clock.p1 = limit->p1.min;
+					clock.p1 <= limit->p1.max; clock.p1++) {
+					int this_err;
+
+					pineview_clock(refclk, &clock);
 					if (!intel_PLL_is_valid(dev, limit,
 								&clock))
 						continue;
@@ -699,9 +604,9 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 }
 
 static bool
-intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-			int target, int refclk, intel_clock_t *match_clock,
-			intel_clock_t *best_clock)
+g4x_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
+		   int target, int refclk, intel_clock_t *match_clock,
+		   intel_clock_t *best_clock)
 {
 	struct drm_device *dev = crtc->dev;
 	intel_clock_t clock;
@@ -712,12 +617,6 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 	found = false;
 
 	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
-		int lvds_reg;
-
-		if (HAS_PCH_SPLIT(dev))
-			lvds_reg = PCH_LVDS;
-		else
-			lvds_reg = LVDS;
 		if (intel_is_dual_link_lvds(dev))
 			clock.p2 = limit->p2.p2_fast;
 		else
@@ -742,13 +641,10 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 				     clock.p1 >= limit->p1.min; clock.p1--) {
 					int this_err;
 
-					intel_clock(dev, refclk, &clock);
+					i9xx_clock(refclk, &clock);
 					if (!intel_PLL_is_valid(dev, limit,
 								&clock))
 						continue;
-					if (match_clock &&
-					    clock.p != match_clock->p)
-						continue;
 
 					this_err = abs(clock.dot - target);
 					if (this_err < err_most) {
@@ -765,62 +661,9 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 }
 
 static bool
-intel_find_pll_ironlake_dp(const intel_limit_t *limit, struct drm_crtc *crtc,
-			   int target, int refclk, intel_clock_t *match_clock,
-			   intel_clock_t *best_clock)
-{
-	struct drm_device *dev = crtc->dev;
-	intel_clock_t clock;
-
-	if (target < 200000) {
-		clock.n = 1;
-		clock.p1 = 2;
-		clock.p2 = 10;
-		clock.m1 = 12;
-		clock.m2 = 9;
-	} else {
-		clock.n = 2;
-		clock.p1 = 1;
-		clock.p2 = 10;
-		clock.m1 = 14;
-		clock.m2 = 8;
-	}
-	intel_clock(dev, refclk, &clock);
-	memcpy(best_clock, &clock, sizeof(intel_clock_t));
-	return true;
-}
-
-/* DisplayPort has only two frequencies, 162MHz and 270MHz */
-static bool
-intel_find_pll_g4x_dp(const intel_limit_t *limit, struct drm_crtc *crtc,
-		      int target, int refclk, intel_clock_t *match_clock,
-		      intel_clock_t *best_clock)
-{
-	intel_clock_t clock;
-	if (target < 200000) {
-		clock.p1 = 2;
-		clock.p2 = 10;
-		clock.n = 2;
-		clock.m1 = 23;
-		clock.m2 = 8;
-	} else {
-		clock.p1 = 1;
-		clock.p2 = 10;
-		clock.n = 1;
-		clock.m1 = 14;
-		clock.m2 = 2;
-	}
-	clock.m = 5 * (clock.m1 + 2) + (clock.m2 + 2);
-	clock.p = (clock.p1 * clock.p2);
-	clock.dot = 96000 * clock.m / (clock.n + 2) / clock.p;
-	clock.vco = 0;
-	memcpy(best_clock, &clock, sizeof(intel_clock_t));
-	return true;
-}
-static bool
-intel_vlv_find_best_pll(const intel_limit_t *limit, struct drm_crtc *crtc,
-			int target, int refclk, intel_clock_t *match_clock,
-			intel_clock_t *best_clock)
+vlv_find_best_dpll(const intel_limit_t *limit, struct drm_crtc *crtc,
+		   int target, int refclk, intel_clock_t *match_clock,
+		   intel_clock_t *best_clock)
 {
 	u32 p1, p2, m1, m2, vco, bestn, bestm1, bestm2, bestp1, bestp2;
 	u32 m, n, fastclk;
@@ -1066,14 +909,24 @@ static void assert_pll(struct drm_i915_private *dev_priv,
 #define assert_pll_enabled(d, p) assert_pll(d, p, true)
 #define assert_pll_disabled(d, p) assert_pll(d, p, false)
 
+static struct intel_shared_dpll *
+intel_crtc_to_shared_dpll(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+
+	if (crtc->config.shared_dpll < 0)
+		return NULL;
+
+	return &dev_priv->shared_dplls[crtc->config.shared_dpll];
+}
+
 /* For ILK+ */
-static void assert_pch_pll(struct drm_i915_private *dev_priv,
-			   struct intel_pch_pll *pll,
-			   struct intel_crtc *crtc,
-			   bool state)
+static void assert_shared_dpll(struct drm_i915_private *dev_priv,
+			       struct intel_shared_dpll *pll,
+			       bool state)
 {
-	u32 val;
 	bool cur_state;
+	struct intel_dpll_hw_state hw_state;
 
 	if (HAS_PCH_LPT(dev_priv->dev)) {
 		DRM_DEBUG_DRIVER("LPT detected: skipping PCH PLL test\n");
@@ -1081,36 +934,16 @@ static void assert_pch_pll(struct drm_i915_private *dev_priv,
 	}
 
 	if (WARN (!pll,
-		  "asserting PCH PLL %s with no PLL\n", state_string(state)))
+		  "asserting DPLL %s with no DPLL\n", state_string(state)))
 		return;
 
-	val = I915_READ(pll->pll_reg);
-	cur_state = !!(val & DPLL_VCO_ENABLE);
+	cur_state = pll->get_hw_state(dev_priv, pll, &hw_state);
 	WARN(cur_state != state,
-	     "PCH PLL state for reg %x assertion failure (expected %s, current %s), val=%08x\n",
-	     pll->pll_reg, state_string(state), state_string(cur_state), val);
-
-	/* Make sure the selected PLL is correctly attached to the transcoder */
-	if (crtc && HAS_PCH_CPT(dev_priv->dev)) {
-		u32 pch_dpll;
-
-		pch_dpll = I915_READ(PCH_DPLL_SEL);
-		cur_state = pll->pll_reg == _PCH_DPLL_B;
-		if (!WARN(((pch_dpll >> (4 * crtc->pipe)) & 1) != cur_state,
-			  "PLL[%d] not attached to this transcoder %d: %08x\n",
-			  cur_state, crtc->pipe, pch_dpll)) {
-			cur_state = !!(val >> (4*crtc->pipe + 3));
-			WARN(cur_state != state,
-			     "PLL[%d] not %s on this transcoder %d: %08x\n",
-			     pll->pll_reg == _PCH_DPLL_B,
-			     state_string(state),
-			     crtc->pipe,
-			     val);
-		}
-	}
+	     "%s assertion failure (expected %s, current %s)\n",
+	     pll->name, state_string(state), state_string(cur_state));
 }
-#define assert_pch_pll_enabled(d, p, c) assert_pch_pll(d, p, c, true)
-#define assert_pch_pll_disabled(d, p, c) assert_pch_pll(d, p, c, false)
+#define assert_shared_dpll_enabled(d, p) assert_shared_dpll(d, p, true)
+#define assert_shared_dpll_disabled(d, p) assert_shared_dpll(d, p, false)
 
 static void assert_fdi_tx(struct drm_i915_private *dev_priv,
 			  enum pipe pipe, bool state)
@@ -1227,8 +1060,8 @@ void assert_pipe(struct drm_i915_private *dev_priv,
 	if (pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE)
 		state = true;
 
-	if (!intel_using_power_well(dev_priv->dev) &&
-	    cpu_transcoder != TRANSCODER_EDP) {
+	if (!intel_display_power_enabled(dev_priv->dev,
+				POWER_DOMAIN_TRANSCODER(cpu_transcoder))) {
 		cur_state = false;
 	} else {
 		reg = PIPECONF(cpu_transcoder);
@@ -1262,12 +1095,13 @@ static void assert_plane(struct drm_i915_private *dev_priv,
 static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 				   enum pipe pipe)
 {
+	struct drm_device *dev = dev_priv->dev;
 	int reg, i;
 	u32 val;
 	int cur_pipe;
 
-	/* Planes are fixed to pipes on ILK+ */
-	if (HAS_PCH_SPLIT(dev_priv->dev) || IS_VALLEYVIEW(dev_priv->dev)) {
+	/* Primary planes are fixed to pipes on gen4+ */
+	if (INTEL_INFO(dev)->gen >= 4) {
 		reg = DSPCNTR(pipe);
 		val = I915_READ(reg);
 		WARN((val & DISPLAY_PLANE_ENABLE),
@@ -1277,7 +1111,7 @@ static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 	}
 
 	/* Need to check both planes against the pipe */
-	for (i = 0; i < 2; i++) {
+	for (i = 0; i < INTEL_INFO(dev)->num_pipes; i++) {
 		reg = DSPCNTR(i);
 		val = I915_READ(reg);
 		cur_pipe = (val & DISPPLANE_SEL_PIPE_MASK) >>
@@ -1291,19 +1125,30 @@ static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 static void assert_sprites_disabled(struct drm_i915_private *dev_priv,
 				    enum pipe pipe)
 {
+	struct drm_device *dev = dev_priv->dev;
 	int reg, i;
 	u32 val;
 
-	if (!IS_VALLEYVIEW(dev_priv->dev))
-		return;
-
-	/* Need to check both planes against the pipe */
-	for (i = 0; i < dev_priv->num_plane; i++) {
-		reg = SPCNTR(pipe, i);
+	if (IS_VALLEYVIEW(dev)) {
+		for (i = 0; i < dev_priv->num_plane; i++) {
+			reg = SPCNTR(pipe, i);
+			val = I915_READ(reg);
+			WARN((val & SP_ENABLE),
+			     "sprite %c assertion failure, should be off on pipe %c but is still active\n",
+			     sprite_name(pipe, i), pipe_name(pipe));
+		}
+	} else if (INTEL_INFO(dev)->gen >= 7) {
+		reg = SPRCTL(pipe);
+		val = I915_READ(reg);
+		WARN((val & SPRITE_ENABLE),
+		     "sprite %c assertion failure, should be off on pipe %c but is still active\n",
+		     plane_name(pipe), pipe_name(pipe));
+	} else if (INTEL_INFO(dev)->gen >= 5) {
+		reg = DVSCNTR(pipe);
 		val = I915_READ(reg);
-		WARN((val & SP_ENABLE),
-		     "sprite %d assertion failure, should be off on pipe %c but is still active\n",
-		     pipe * 2 + i, pipe_name(pipe));
+		WARN((val & DVS_ENABLE),
+		     "sprite %c assertion failure, should be off on pipe %c but is still active\n",
+		     plane_name(pipe), pipe_name(pipe));
 	}
 }
 
@@ -1323,14 +1168,14 @@ static void assert_pch_refclk_enabled(struct drm_i915_private *dev_priv)
 	WARN(!enabled, "PCH refclk assertion failure, should be active but is disabled\n");
 }
 
-static void assert_transcoder_disabled(struct drm_i915_private *dev_priv,
-				       enum pipe pipe)
+static void assert_pch_transcoder_disabled(struct drm_i915_private *dev_priv,
+					   enum pipe pipe)
 {
 	int reg;
 	u32 val;
 	bool enabled;
 
-	reg = TRANSCONF(pipe);
+	reg = PCH_TRANSCONF(pipe);
 	val = I915_READ(reg);
 	enabled = !!(val & TRANS_ENABLE);
 	WARN(enabled,
@@ -1474,6 +1319,8 @@ static void intel_enable_pll(struct drm_i915_private *dev_priv, enum pipe pipe)
 	int reg;
 	u32 val;
 
+	assert_pipe_disabled(dev_priv, pipe);
+
 	/* No really, not for ILK+ */
 	BUG_ON(!IS_VALLEYVIEW(dev_priv->dev) && dev_priv->info->gen >= 5);
 
@@ -1525,156 +1372,86 @@ static void intel_disable_pll(struct drm_i915_private *dev_priv, enum pipe pipe)
 	POSTING_READ(reg);
 }
 
-/* SBI access */
-static void
-intel_sbi_write(struct drm_i915_private *dev_priv, u16 reg, u32 value,
-		enum intel_sbi_destination destination)
-{
-	u32 tmp;
-
-	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
-
-	if (wait_for((I915_READ(SBI_CTL_STAT) & SBI_BUSY) == 0,
-				100)) {
-		DRM_ERROR("timeout waiting for SBI to become ready\n");
-		return;
-	}
-
-	I915_WRITE(SBI_ADDR, (reg << 16));
-	I915_WRITE(SBI_DATA, value);
-
-	if (destination == SBI_ICLK)
-		tmp = SBI_CTL_DEST_ICLK | SBI_CTL_OP_CRWR;
-	else
-		tmp = SBI_CTL_DEST_MPHY | SBI_CTL_OP_IOWR;
-	I915_WRITE(SBI_CTL_STAT, SBI_BUSY | tmp);
-
-	if (wait_for((I915_READ(SBI_CTL_STAT) & (SBI_BUSY | SBI_RESPONSE_FAIL)) == 0,
-				100)) {
-		DRM_ERROR("timeout waiting for SBI to complete write transaction\n");
-		return;
-	}
-}
-
-static u32
-intel_sbi_read(struct drm_i915_private *dev_priv, u16 reg,
-	       enum intel_sbi_destination destination)
+void vlv_wait_port_ready(struct drm_i915_private *dev_priv, int port)
 {
-	u32 value = 0;
-	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
+	u32 port_mask;
 
-	if (wait_for((I915_READ(SBI_CTL_STAT) & SBI_BUSY) == 0,
-				100)) {
-		DRM_ERROR("timeout waiting for SBI to become ready\n");
-		return 0;
-	}
-
-	I915_WRITE(SBI_ADDR, (reg << 16));
-
-	if (destination == SBI_ICLK)
-		value = SBI_CTL_DEST_ICLK | SBI_CTL_OP_CRRD;
+	if (!port)
+		port_mask = DPLL_PORTB_READY_MASK;
 	else
-		value = SBI_CTL_DEST_MPHY | SBI_CTL_OP_IORD;
-	I915_WRITE(SBI_CTL_STAT, value | SBI_BUSY);
-
-	if (wait_for((I915_READ(SBI_CTL_STAT) & (SBI_BUSY | SBI_RESPONSE_FAIL)) == 0,
-				100)) {
-		DRM_ERROR("timeout waiting for SBI to complete read transaction\n");
-		return 0;
-	}
+		port_mask = DPLL_PORTC_READY_MASK;
 
-	return I915_READ(SBI_DATA);
+	if (wait_for((I915_READ(DPLL(0)) & port_mask) == 0, 1000))
+		WARN(1, "timed out waiting for port %c ready: 0x%08x\n",
+		     'B' + port, I915_READ(DPLL(0)));
 }
 
 /**
- * ironlake_enable_pch_pll - enable PCH PLL
+ * ironlake_enable_shared_dpll - enable PCH PLL
  * @dev_priv: i915 private structure
  * @pipe: pipe PLL to enable
  *
  * The PCH PLL needs to be enabled before the PCH transcoder, since it
  * drives the transcoder clock.
  */
-static void ironlake_enable_pch_pll(struct intel_crtc *intel_crtc)
+static void ironlake_enable_shared_dpll(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = intel_crtc->base.dev->dev_private;
-	struct intel_pch_pll *pll;
-	int reg;
-	u32 val;
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct intel_shared_dpll *pll = intel_crtc_to_shared_dpll(crtc);
 
 	/* PCH PLLs only available on ILK, SNB and IVB */
 	BUG_ON(dev_priv->info->gen < 5);
-	pll = intel_crtc->pch_pll;
-	if (pll == NULL)
+	if (WARN_ON(pll == NULL))
 		return;
 
 	if (WARN_ON(pll->refcount == 0))
 		return;
 
-	DRM_DEBUG_KMS("enable PCH PLL %x (active %d, on? %d)for crtc %d\n",
-		      pll->pll_reg, pll->active, pll->on,
-		      intel_crtc->base.base.id);
+	DRM_DEBUG_KMS("enable %s (active %d, on? %d)for crtc %d\n",
+		      pll->name, pll->active, pll->on,
+		      crtc->base.base.id);
 
-	/* PCH refclock must be enabled first */
-	assert_pch_refclk_enabled(dev_priv);
-
-	if (pll->active++ && pll->on) {
-		assert_pch_pll_enabled(dev_priv, pll, NULL);
+	if (pll->active++) {
+		WARN_ON(!pll->on);
+		assert_shared_dpll_enabled(dev_priv, pll);
 		return;
 	}
+	WARN_ON(pll->on);
 
-	DRM_DEBUG_KMS("enabling PCH PLL %x\n", pll->pll_reg);
-
-	reg = pll->pll_reg;
-	val = I915_READ(reg);
-	val |= DPLL_VCO_ENABLE;
-	I915_WRITE(reg, val);
-	POSTING_READ(reg);
-	udelay(200);
-
+	DRM_DEBUG_KMS("enabling %s\n", pll->name);
+	pll->enable(dev_priv, pll);
 	pll->on = true;
 }
 
-static void intel_disable_pch_pll(struct intel_crtc *intel_crtc)
+static void intel_disable_shared_dpll(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = intel_crtc->base.dev->dev_private;
-	struct intel_pch_pll *pll = intel_crtc->pch_pll;
-	int reg;
-	u32 val;
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct intel_shared_dpll *pll = intel_crtc_to_shared_dpll(crtc);
 
 	/* PCH only available on ILK+ */
 	BUG_ON(dev_priv->info->gen < 5);
-	if (pll == NULL)
+	if (WARN_ON(pll == NULL))
 	       return;
 
 	if (WARN_ON(pll->refcount == 0))
 		return;
 
-	DRM_DEBUG_KMS("disable PCH PLL %x (active %d, on? %d) for crtc %d\n",
-		      pll->pll_reg, pll->active, pll->on,
-		      intel_crtc->base.base.id);
+	DRM_DEBUG_KMS("disable %s (active %d, on? %d) for crtc %d\n",
+		      pll->name, pll->active, pll->on,
+		      crtc->base.base.id);
 
 	if (WARN_ON(pll->active == 0)) {
-		assert_pch_pll_disabled(dev_priv, pll, NULL);
+		assert_shared_dpll_disabled(dev_priv, pll);
 		return;
 	}
 
-	if (--pll->active) {
-		assert_pch_pll_enabled(dev_priv, pll, NULL);
+	assert_shared_dpll_enabled(dev_priv, pll);
+	WARN_ON(!pll->on);
+	if (--pll->active)
 		return;
-	}
-
-	DRM_DEBUG_KMS("disabling PCH PLL %x\n", pll->pll_reg);
-
-	/* Make sure transcoder isn't still depending on us */
-	assert_transcoder_disabled(dev_priv, intel_crtc->pipe);
-
-	reg = pll->pll_reg;
-	val = I915_READ(reg);
-	val &= ~DPLL_VCO_ENABLE;
-	I915_WRITE(reg, val);
-	POSTING_READ(reg);
-	udelay(200);
 
+	DRM_DEBUG_KMS("disabling %s\n", pll->name);
+	pll->disable(dev_priv, pll);
 	pll->on = false;
 }
 
@@ -1683,15 +1460,15 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 {
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	uint32_t reg, val, pipeconf_val;
 
 	/* PCH only available on ILK+ */
 	BUG_ON(dev_priv->info->gen < 5);
 
 	/* Make sure PCH DPLL is enabled */
-	assert_pch_pll_enabled(dev_priv,
-			       to_intel_crtc(crtc)->pch_pll,
-			       to_intel_crtc(crtc));
+	assert_shared_dpll_enabled(dev_priv,
+				   intel_crtc_to_shared_dpll(intel_crtc));
 
 	/* FDI must be feeding us bits for PCH ports */
 	assert_fdi_tx_enabled(dev_priv, pipe);
@@ -1706,7 +1483,7 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 		I915_WRITE(reg, val);
 	}
 
-	reg = TRANSCONF(pipe);
+	reg = PCH_TRANSCONF(pipe);
 	val = I915_READ(reg);
 	pipeconf_val = I915_READ(PIPECONF(pipe));
 
@@ -1731,7 +1508,7 @@ static void ironlake_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 
 	I915_WRITE(reg, val | TRANS_ENABLE);
 	if (wait_for(I915_READ(reg) & TRANS_STATE_ENABLE, 100))
-		DRM_ERROR("failed to enable transcoder %d\n", pipe);
+		DRM_ERROR("failed to enable transcoder %c\n", pipe_name(pipe));
 }
 
 static void lpt_enable_pch_transcoder(struct drm_i915_private *dev_priv,
@@ -1760,8 +1537,8 @@ static void lpt_enable_pch_transcoder(struct drm_i915_private *dev_priv,
 	else
 		val |= TRANS_PROGRESSIVE;
 
-	I915_WRITE(TRANSCONF(TRANSCODER_A), val);
-	if (wait_for(I915_READ(_TRANSACONF) & TRANS_STATE_ENABLE, 100))
+	I915_WRITE(LPT_TRANSCONF, val);
+	if (wait_for(I915_READ(LPT_TRANSCONF) & TRANS_STATE_ENABLE, 100))
 		DRM_ERROR("Failed to enable PCH transcoder\n");
 }
 
@@ -1778,13 +1555,13 @@ static void ironlake_disable_pch_transcoder(struct drm_i915_private *dev_priv,
 	/* Ports must be off as well */
 	assert_pch_ports_disabled(dev_priv, pipe);
 
-	reg = TRANSCONF(pipe);
+	reg = PCH_TRANSCONF(pipe);
 	val = I915_READ(reg);
 	val &= ~TRANS_ENABLE;
 	I915_WRITE(reg, val);
 	/* wait for PCH transcoder off, transcoder state */
 	if (wait_for((I915_READ(reg) & TRANS_STATE_ENABLE) == 0, 50))
-		DRM_ERROR("failed to disable transcoder %d\n", pipe);
+		DRM_ERROR("failed to disable transcoder %c\n", pipe_name(pipe));
 
 	if (!HAS_PCH_IBX(dev)) {
 		/* Workaround: Clear the timing override chicken bit again. */
@@ -1799,11 +1576,11 @@ static void lpt_disable_pch_transcoder(struct drm_i915_private *dev_priv)
 {
 	u32 val;
 
-	val = I915_READ(_TRANSACONF);
+	val = I915_READ(LPT_TRANSCONF);
 	val &= ~TRANS_ENABLE;
-	I915_WRITE(_TRANSACONF, val);
+	I915_WRITE(LPT_TRANSCONF, val);
 	/* wait for PCH transcoder off, transcoder state */
-	if (wait_for((I915_READ(_TRANSACONF) & TRANS_STATE_ENABLE) == 0, 50))
+	if (wait_for((I915_READ(LPT_TRANSCONF) & TRANS_STATE_ENABLE) == 0, 50))
 		DRM_ERROR("Failed to disable PCH transcoder\n");
 
 	/* Workaround: clear timing override bit. */
@@ -1835,6 +1612,9 @@ static void intel_enable_pipe(struct drm_i915_private *dev_priv, enum pipe pipe,
 	int reg;
 	u32 val;
 
+	assert_planes_disabled(dev_priv, pipe);
+	assert_sprites_disabled(dev_priv, pipe);
+
 	if (HAS_PCH_LPT(dev_priv->dev))
 		pch_transcoder = TRANSCODER_A;
 	else
@@ -2096,7 +1876,7 @@ static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	case 1:
 		break;
 	default:
-		DRM_ERROR("Can't update plane %d in SAREA\n", plane);
+		DRM_ERROR("Can't update plane %c in SAREA\n", plane_name(plane));
 		return -EINVAL;
 	}
 
@@ -2145,6 +1925,9 @@ static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			dspcntr &= ~DISPPLANE_TILED;
 	}
 
+	if (IS_G4X(dev))
+		dspcntr |= DISPPLANE_TRICKLE_FEED_DISABLE;
+
 	I915_WRITE(reg, dspcntr);
 
 	linear_offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8);
@@ -2193,7 +1976,7 @@ static int ironlake_update_plane(struct drm_crtc *crtc,
 	case 2:
 		break;
 	default:
-		DRM_ERROR("Can't update plane %d in SAREA\n", plane);
+		DRM_ERROR("Can't update plane %c in SAREA\n", plane_name(plane));
 		return -EINVAL;
 	}
 
@@ -2384,9 +2167,9 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 	}
 
 	if (intel_crtc->plane > INTEL_INFO(dev)->num_pipes) {
-		DRM_ERROR("no plane for crtc: plane %d, num_pipes %d\n",
-				intel_crtc->plane,
-				INTEL_INFO(dev)->num_pipes);
+		DRM_ERROR("no plane for crtc: plane %c, num_pipes %d\n",
+			  plane_name(intel_crtc->plane),
+			  INTEL_INFO(dev)->num_pipes);
 		return -EINVAL;
 	}
 
@@ -2414,7 +2197,8 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 	crtc->y = y;
 
 	if (old_fb) {
-		intel_wait_for_vblank(dev, intel_crtc->pipe);
+		if (intel_crtc->active && old_fb != fb)
+			intel_wait_for_vblank(dev, intel_crtc->pipe);
 		intel_unpin_fb_obj(to_intel_framebuffer(old_fb)->obj);
 	}
 
@@ -2467,6 +2251,11 @@ static void intel_fdi_normal_train(struct drm_crtc *crtc)
 			   FDI_FE_ERRC_ENABLE);
 }
 
+static bool pipe_has_enabled_pch(struct intel_crtc *intel_crtc)
+{
+	return intel_crtc->base.enabled && intel_crtc->config.has_pch_encoder;
+}
+
 static void ivb_modeset_global_resources(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -2476,10 +2265,13 @@ static void ivb_modeset_global_resources(struct drm_device *dev)
 		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[PIPE_C]);
 	uint32_t temp;
 
-	/* When everything is off disable fdi C so that we could enable fdi B
-	 * with all lanes. XXX: This misses the case where a pipe is not using
-	 * any pch resources and so doesn't need any fdi lanes. */
-	if (!pipe_B_crtc->base.enabled && !pipe_C_crtc->base.enabled) {
+	/*
+	 * When everything is off disable fdi C so that we could enable fdi B
+	 * with all lanes. Note that we don't care about enabled pipes without
+	 * an enabled pch encoder.
+	 */
+	if (!pipe_has_enabled_pch(pipe_B_crtc) &&
+	    !pipe_has_enabled_pch(pipe_C_crtc)) {
 		WARN_ON(I915_READ(FDI_RX_CTL(PIPE_B)) & FDI_RX_ENABLE);
 		WARN_ON(I915_READ(FDI_RX_CTL(PIPE_C)) & FDI_RX_ENABLE);
 
@@ -2517,8 +2309,8 @@ static void ironlake_fdi_link_train(struct drm_crtc *crtc)
 	/* enable CPU FDI TX and PCH FDI RX */
 	reg = FDI_TX_CTL(pipe);
 	temp = I915_READ(reg);
-	temp &= ~(7 << 19);
-	temp |= (intel_crtc->fdi_lanes - 1) << 19;
+	temp &= ~FDI_DP_PORT_WIDTH_MASK;
+	temp |= FDI_DP_PORT_WIDTH(intel_crtc->config.fdi_lanes);
 	temp &= ~FDI_LINK_TRAIN_NONE;
 	temp |= FDI_LINK_TRAIN_PATTERN_1;
 	I915_WRITE(reg, temp | FDI_TX_ENABLE);
@@ -2615,8 +2407,8 @@ static void gen6_fdi_link_train(struct drm_crtc *crtc)
 	/* enable CPU FDI TX and PCH FDI RX */
 	reg = FDI_TX_CTL(pipe);
 	temp = I915_READ(reg);
-	temp &= ~(7 << 19);
-	temp |= (intel_crtc->fdi_lanes - 1) << 19;
+	temp &= ~FDI_DP_PORT_WIDTH_MASK;
+	temp |= FDI_DP_PORT_WIDTH(intel_crtc->config.fdi_lanes);
 	temp &= ~FDI_LINK_TRAIN_NONE;
 	temp |= FDI_LINK_TRAIN_PATTERN_1;
 	temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
@@ -2750,8 +2542,8 @@ static void ivb_manual_fdi_link_train(struct drm_crtc *crtc)
 	/* enable CPU FDI TX and PCH FDI RX */
 	reg = FDI_TX_CTL(pipe);
 	temp = I915_READ(reg);
-	temp &= ~(7 << 19);
-	temp |= (intel_crtc->fdi_lanes - 1) << 19;
+	temp &= ~FDI_DP_PORT_WIDTH_MASK;
+	temp |= FDI_DP_PORT_WIDTH(intel_crtc->config.fdi_lanes);
 	temp &= ~(FDI_LINK_TRAIN_AUTO | FDI_LINK_TRAIN_NONE_IVB);
 	temp |= FDI_LINK_TRAIN_PATTERN_1_IVB;
 	temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
@@ -2852,8 +2644,8 @@ static void ironlake_fdi_pll_enable(struct intel_crtc *intel_crtc)
 	/* enable PCH FDI RX PLL, wait warmup plus DMI latency */
 	reg = FDI_RX_CTL(pipe);
 	temp = I915_READ(reg);
-	temp &= ~((0x7 << 19) | (0x7 << 16));
-	temp |= (intel_crtc->fdi_lanes - 1) << 19;
+	temp &= ~(FDI_DP_PORT_WIDTH_MASK | (0x7 << 16));
+	temp |= FDI_DP_PORT_WIDTH(intel_crtc->config.fdi_lanes);
 	temp |= (I915_READ(PIPECONF(pipe)) & PIPECONF_BPC_MASK) << 11;
 	I915_WRITE(reg, temp | FDI_RX_PLL_ENABLE);
 
@@ -3085,6 +2877,30 @@ static void lpt_program_iclkip(struct drm_crtc *crtc)
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
+static void ironlake_pch_transcoder_set_timings(struct intel_crtc *crtc,
+						enum pipe pch_transcoder)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum transcoder cpu_transcoder = crtc->config.cpu_transcoder;
+
+	I915_WRITE(PCH_TRANS_HTOTAL(pch_transcoder),
+		   I915_READ(HTOTAL(cpu_transcoder)));
+	I915_WRITE(PCH_TRANS_HBLANK(pch_transcoder),
+		   I915_READ(HBLANK(cpu_transcoder)));
+	I915_WRITE(PCH_TRANS_HSYNC(pch_transcoder),
+		   I915_READ(HSYNC(cpu_transcoder)));
+
+	I915_WRITE(PCH_TRANS_VTOTAL(pch_transcoder),
+		   I915_READ(VTOTAL(cpu_transcoder)));
+	I915_WRITE(PCH_TRANS_VBLANK(pch_transcoder),
+		   I915_READ(VBLANK(cpu_transcoder)));
+	I915_WRITE(PCH_TRANS_VSYNC(pch_transcoder),
+		   I915_READ(VSYNC(cpu_transcoder)));
+	I915_WRITE(PCH_TRANS_VSYNCSHIFT(pch_transcoder),
+		   I915_READ(VSYNCSHIFT(cpu_transcoder)));
+}
+
 /*
  * Enable PCH resources required for PCH ports:
  *   - PCH PLLs
@@ -3101,7 +2917,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	int pipe = intel_crtc->pipe;
 	u32 reg, temp;
 
-	assert_transcoder_disabled(dev_priv, pipe);
+	assert_pch_transcoder_disabled(dev_priv, pipe);
 
 	/* Write the TU size bits before fdi link training, so that error
 	 * detection works. */
@@ -3115,31 +2931,18 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	 * transcoder, and we actually should do this to not upset any PCH
 	 * transcoder that already use the clock when we share it.
 	 *
-	 * Note that enable_pch_pll tries to do the right thing, but get_pch_pll
-	 * unconditionally resets the pll - we need that to have the right LVDS
-	 * enable sequence. */
-	ironlake_enable_pch_pll(intel_crtc);
+	 * Note that enable_shared_dpll tries to do the right thing, but
+	 * get_shared_dpll unconditionally resets the pll - we need that to have
+	 * the right LVDS enable sequence. */
+	ironlake_enable_shared_dpll(intel_crtc);
 
 	if (HAS_PCH_CPT(dev)) {
 		u32 sel;
 
 		temp = I915_READ(PCH_DPLL_SEL);
-		switch (pipe) {
-		default:
-		case 0:
-			temp |= TRANSA_DPLL_ENABLE;
-			sel = TRANSA_DPLLB_SEL;
-			break;
-		case 1:
-			temp |= TRANSB_DPLL_ENABLE;
-			sel = TRANSB_DPLLB_SEL;
-			break;
-		case 2:
-			temp |= TRANSC_DPLL_ENABLE;
-			sel = TRANSC_DPLLB_SEL;
-			break;
-		}
-		if (intel_crtc->pch_pll->pll_reg == _PCH_DPLL_B)
+		temp |= TRANS_DPLL_ENABLE(pipe);
+		sel = TRANS_DPLLB_SEL(pipe);
+		if (intel_crtc->config.shared_dpll == DPLL_ID_PCH_PLL_B)
 			temp |= sel;
 		else
 			temp &= ~sel;
@@ -3148,14 +2951,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 
 	/* set transcoder timing, panel must allow it */
 	assert_panel_unlocked(dev_priv, pipe);
-	I915_WRITE(TRANS_HTOTAL(pipe), I915_READ(HTOTAL(pipe)));
-	I915_WRITE(TRANS_HBLANK(pipe), I915_READ(HBLANK(pipe)));
-	I915_WRITE(TRANS_HSYNC(pipe),  I915_READ(HSYNC(pipe)));
-
-	I915_WRITE(TRANS_VTOTAL(pipe), I915_READ(VTOTAL(pipe)));
-	I915_WRITE(TRANS_VBLANK(pipe), I915_READ(VBLANK(pipe)));
-	I915_WRITE(TRANS_VSYNC(pipe),  I915_READ(VSYNC(pipe)));
-	I915_WRITE(TRANS_VSYNCSHIFT(pipe),  I915_READ(VSYNCSHIFT(pipe)));
+	ironlake_pch_transcoder_set_timings(intel_crtc, pipe);
 
 	intel_fdi_normal_train(crtc);
 
@@ -3205,86 +3001,82 @@ static void lpt_pch_enable(struct drm_crtc *crtc)
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 
-	assert_transcoder_disabled(dev_priv, TRANSCODER_A);
+	assert_pch_transcoder_disabled(dev_priv, TRANSCODER_A);
 
 	lpt_program_iclkip(crtc);
 
 	/* Set transcoder timing. */
-	I915_WRITE(_TRANS_HTOTAL_A, I915_READ(HTOTAL(cpu_transcoder)));
-	I915_WRITE(_TRANS_HBLANK_A, I915_READ(HBLANK(cpu_transcoder)));
-	I915_WRITE(_TRANS_HSYNC_A,  I915_READ(HSYNC(cpu_transcoder)));
-
-	I915_WRITE(_TRANS_VTOTAL_A, I915_READ(VTOTAL(cpu_transcoder)));
-	I915_WRITE(_TRANS_VBLANK_A, I915_READ(VBLANK(cpu_transcoder)));
-	I915_WRITE(_TRANS_VSYNC_A,  I915_READ(VSYNC(cpu_transcoder)));
-	I915_WRITE(_TRANS_VSYNCSHIFT_A, I915_READ(VSYNCSHIFT(cpu_transcoder)));
+	ironlake_pch_transcoder_set_timings(intel_crtc, PIPE_A);
 
 	lpt_enable_pch_transcoder(dev_priv, cpu_transcoder);
 }
 
-static void intel_put_pch_pll(struct intel_crtc *intel_crtc)
+static void intel_put_shared_dpll(struct intel_crtc *crtc)
 {
-	struct intel_pch_pll *pll = intel_crtc->pch_pll;
+	struct intel_shared_dpll *pll = intel_crtc_to_shared_dpll(crtc);
 
 	if (pll == NULL)
 		return;
 
 	if (pll->refcount == 0) {
-		WARN(1, "bad PCH PLL refcount\n");
+		WARN(1, "bad %s refcount\n", pll->name);
 		return;
 	}
 
-	--pll->refcount;
-	intel_crtc->pch_pll = NULL;
+	if (--pll->refcount == 0) {
+		WARN_ON(pll->on);
+		WARN_ON(pll->active);
+	}
+
+	crtc->config.shared_dpll = DPLL_ID_PRIVATE;
 }
 
-static struct intel_pch_pll *intel_get_pch_pll(struct intel_crtc *intel_crtc, u32 dpll, u32 fp)
+static struct intel_shared_dpll *intel_get_shared_dpll(struct intel_crtc *crtc, u32 dpll, u32 fp)
 {
-	struct drm_i915_private *dev_priv = intel_crtc->base.dev->dev_private;
-	struct intel_pch_pll *pll;
-	int i;
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct intel_shared_dpll *pll = intel_crtc_to_shared_dpll(crtc);
+	enum intel_dpll_id i;
 
-	pll = intel_crtc->pch_pll;
 	if (pll) {
-		DRM_DEBUG_KMS("CRTC:%d reusing existing PCH PLL %x\n",
-			      intel_crtc->base.base.id, pll->pll_reg);
-		goto prepare;
+		DRM_DEBUG_KMS("CRTC:%d dropping existing %s\n",
+			      crtc->base.base.id, pll->name);
+		intel_put_shared_dpll(crtc);
 	}
 
 	if (HAS_PCH_IBX(dev_priv->dev)) {
 		/* Ironlake PCH has a fixed PLL->PCH pipe mapping. */
-		i = intel_crtc->pipe;
-		pll = &dev_priv->pch_plls[i];
+		i = crtc->pipe;
+		pll = &dev_priv->shared_dplls[i];
 
-		DRM_DEBUG_KMS("CRTC:%d using pre-allocated PCH PLL %x\n",
-			      intel_crtc->base.base.id, pll->pll_reg);
+		DRM_DEBUG_KMS("CRTC:%d using pre-allocated %s\n",
+			      crtc->base.base.id, pll->name);
 
 		goto found;
 	}
 
-	for (i = 0; i < dev_priv->num_pch_pll; i++) {
-		pll = &dev_priv->pch_plls[i];
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		pll = &dev_priv->shared_dplls[i];
 
 		/* Only want to check enabled timings first */
 		if (pll->refcount == 0)
 			continue;
 
-		if (dpll == (I915_READ(pll->pll_reg) & 0x7fffffff) &&
-		    fp == I915_READ(pll->fp0_reg)) {
-			DRM_DEBUG_KMS("CRTC:%d sharing existing PCH PLL %x (refcount %d, ative %d)\n",
-				      intel_crtc->base.base.id,
-				      pll->pll_reg, pll->refcount, pll->active);
+		if (dpll == (I915_READ(PCH_DPLL(pll->id)) & 0x7fffffff) &&
+		    fp == I915_READ(PCH_FP0(pll->id))) {
+			DRM_DEBUG_KMS("CRTC:%d sharing existing %s (refcount %d, ative %d)\n",
+				      crtc->base.base.id,
+				      pll->name, pll->refcount, pll->active);
 
 			goto found;
 		}
 	}
 
 	/* Ok no matching timings, maybe there's a free one? */
-	for (i = 0; i < dev_priv->num_pch_pll; i++) {
-		pll = &dev_priv->pch_plls[i];
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		pll = &dev_priv->shared_dplls[i];
 		if (pll->refcount == 0) {
-			DRM_DEBUG_KMS("CRTC:%d allocated PCH PLL %x\n",
-				      intel_crtc->base.base.id, pll->pll_reg);
+			DRM_DEBUG_KMS("CRTC:%d allocated %s\n",
+				      crtc->base.base.id, pll->name);
 			goto found;
 		}
 	}
@@ -3292,24 +3084,32 @@ static struct intel_pch_pll *intel_get_pch_pll(struct intel_crtc *intel_crtc, u3
 	return NULL;
 
 found:
-	intel_crtc->pch_pll = pll;
-	pll->refcount++;
-	DRM_DEBUG_DRIVER("using pll %d for pipe %d\n", i, intel_crtc->pipe);
-prepare: /* separate function? */
-	DRM_DEBUG_DRIVER("switching PLL %x off\n", pll->pll_reg);
+	crtc->config.shared_dpll = i;
+	DRM_DEBUG_DRIVER("using %s for pipe %c\n", pll->name,
+			 pipe_name(crtc->pipe));
 
-	/* Wait for the clocks to stabilize before rewriting the regs */
-	I915_WRITE(pll->pll_reg, dpll & ~DPLL_VCO_ENABLE);
-	POSTING_READ(pll->pll_reg);
-	udelay(150);
+	if (pll->active == 0) {
+		memcpy(&pll->hw_state, &crtc->config.dpll_hw_state,
+		       sizeof(pll->hw_state));
+
+		DRM_DEBUG_DRIVER("setting up %s\n", pll->name);
+		WARN_ON(pll->on);
+		assert_shared_dpll_disabled(dev_priv, pll);
+
+		/* Wait for the clocks to stabilize before rewriting the regs */
+		I915_WRITE(PCH_DPLL(pll->id), dpll & ~DPLL_VCO_ENABLE);
+		POSTING_READ(PCH_DPLL(pll->id));
+		udelay(150);
+
+		I915_WRITE(PCH_FP0(pll->id), fp);
+		I915_WRITE(PCH_DPLL(pll->id), dpll & ~DPLL_VCO_ENABLE);
+	}
+	pll->refcount++;
 
-	I915_WRITE(pll->fp0_reg, fp);
-	I915_WRITE(pll->pll_reg, dpll & ~DPLL_VCO_ENABLE);
-	pll->on = false;
 	return pll;
 }
 
-void intel_cpt_verify_modeset(struct drm_device *dev, int pipe)
+static void cpt_verify_modeset(struct drm_device *dev, int pipe)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int dslreg = PIPEDSL(pipe);
@@ -3319,10 +3119,53 @@ void intel_cpt_verify_modeset(struct drm_device *dev, int pipe)
 	udelay(500);
 	if (wait_for(I915_READ(dslreg) != temp, 5)) {
 		if (wait_for(I915_READ(dslreg) != temp, 5))
-			DRM_ERROR("mode set failed: pipe %d stuck\n", pipe);
+			DRM_ERROR("mode set failed: pipe %c stuck\n", pipe_name(pipe));
+	}
+}
+
+static void ironlake_pfit_enable(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = crtc->pipe;
+
+	if (crtc->config.pch_pfit.size) {
+		/* Force use of hard-coded filter coefficients
+		 * as some pre-programmed values are broken,
+		 * e.g. x201.
+		 */
+		if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev))
+			I915_WRITE(PF_CTL(pipe), PF_ENABLE | PF_FILTER_MED_3x3 |
+						 PF_PIPE_SEL_IVB(pipe));
+		else
+			I915_WRITE(PF_CTL(pipe), PF_ENABLE | PF_FILTER_MED_3x3);
+		I915_WRITE(PF_WIN_POS(pipe), crtc->config.pch_pfit.pos);
+		I915_WRITE(PF_WIN_SZ(pipe), crtc->config.pch_pfit.size);
 	}
 }
 
+static void intel_enable_planes(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+	struct intel_plane *intel_plane;
+
+	list_for_each_entry(intel_plane, &dev->mode_config.plane_list, base.head)
+		if (intel_plane->pipe == pipe)
+			intel_plane_restore(&intel_plane->base);
+}
+
+static void intel_disable_planes(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+	struct intel_plane *intel_plane;
+
+	list_for_each_entry(intel_plane, &dev->mode_config.plane_list, base.head)
+		if (intel_plane->pipe == pipe)
+			intel_plane_disable(&intel_plane->base);
+}
+
 static void ironlake_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -3339,6 +3182,10 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 		return;
 
 	intel_crtc->active = true;
+
+	intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
+	intel_set_pch_fifo_underrun_reporting(dev, pipe, true);
+
 	intel_update_watermarks(dev);
 
 	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS)) {
@@ -3362,22 +3209,7 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 		if (encoder->pre_enable)
 			encoder->pre_enable(encoder);
 
-	/* Enable panel fitting for LVDS */
-	if (dev_priv->pch_pf_size &&
-	    (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) ||
-	     intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))) {
-		/* Force use of hard-coded filter coefficients
-		 * as some pre-programmed values are broken,
-		 * e.g. x201.
-		 */
-		if (IS_IVYBRIDGE(dev))
-			I915_WRITE(PF_CTL(pipe), PF_ENABLE | PF_FILTER_MED_3x3 |
-						 PF_PIPE_SEL_IVB(pipe));
-		else
-			I915_WRITE(PF_CTL(pipe), PF_ENABLE | PF_FILTER_MED_3x3);
-		I915_WRITE(PF_WIN_POS(pipe), dev_priv->pch_pf_pos);
-		I915_WRITE(PF_WIN_SZ(pipe), dev_priv->pch_pf_size);
-	}
+	ironlake_pfit_enable(intel_crtc);
 
 	/*
 	 * On ILK+ LUT must be loaded before the pipe is running but with
@@ -3388,6 +3220,8 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	intel_enable_pipe(dev_priv, pipe,
 			  intel_crtc->config.has_pch_encoder);
 	intel_enable_plane(dev_priv, plane, pipe);
+	intel_enable_planes(crtc);
+	intel_crtc_update_cursor(crtc, true);
 
 	if (intel_crtc->config.has_pch_encoder)
 		ironlake_pch_enable(crtc);
@@ -3396,13 +3230,11 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	intel_update_fbc(dev);
 	mutex_unlock(&dev->struct_mutex);
 
-	intel_crtc_update_cursor(crtc, true);
-
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		encoder->enable(encoder);
 
 	if (HAS_PCH_CPT(dev))
-		intel_cpt_verify_modeset(dev, intel_crtc->pipe);
+		cpt_verify_modeset(dev, intel_crtc->pipe);
 
 	/*
 	 * There seems to be a race in PCH platform hw (at least on some
@@ -3415,6 +3247,42 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	intel_wait_for_vblank(dev, intel_crtc->pipe);
 }
 
+/* IPS only exists on ULT machines and is tied to pipe A. */
+static bool hsw_crtc_supports_ips(struct intel_crtc *crtc)
+{
+	return HAS_IPS(crtc->base.dev) && crtc->pipe == PIPE_A;
+}
+
+static void hsw_enable_ips(struct intel_crtc *crtc)
+{
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+
+	if (!crtc->config.ips_enabled)
+		return;
+
+	/* We can only enable IPS after we enable a plane and wait for a vblank.
+	 * We guarantee that the plane is enabled by calling intel_enable_ips
+	 * only after intel_enable_plane. And intel_enable_plane already waits
+	 * for a vblank, so all we need to do here is to enable the IPS bit. */
+	assert_plane_enabled(dev_priv, crtc->plane);
+	I915_WRITE(IPS_CTL, IPS_ENABLE);
+}
+
+static void hsw_disable_ips(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (!crtc->config.ips_enabled)
+		return;
+
+	assert_plane_enabled(dev_priv, crtc->plane);
+	I915_WRITE(IPS_CTL, 0);
+
+	/* We need to wait for a vblank before we can disable the plane. */
+	intel_wait_for_vblank(dev, crtc->pipe);
+}
+
 static void haswell_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -3430,6 +3298,11 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 		return;
 
 	intel_crtc->active = true;
+
+	intel_set_cpu_fifo_underrun_reporting(dev, pipe, true);
+	if (intel_crtc->config.has_pch_encoder)
+		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, true);
+
 	intel_update_watermarks(dev);
 
 	if (intel_crtc->config.has_pch_encoder)
@@ -3441,18 +3314,7 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 
 	intel_ddi_enable_pipe_clock(intel_crtc);
 
-	/* Enable panel fitting for eDP */
-	if (dev_priv->pch_pf_size &&
-	    intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP)) {
-		/* Force use of hard-coded filter coefficients
-		 * as some pre-programmed values are broken,
-		 * e.g. x201.
-		 */
-		I915_WRITE(PF_CTL(pipe), PF_ENABLE | PF_FILTER_MED_3x3 |
-					 PF_PIPE_SEL_IVB(pipe));
-		I915_WRITE(PF_WIN_POS(pipe), dev_priv->pch_pf_pos);
-		I915_WRITE(PF_WIN_SZ(pipe), dev_priv->pch_pf_size);
-	}
+	ironlake_pfit_enable(intel_crtc);
 
 	/*
 	 * On ILK+ LUT must be loaded before the pipe is running but with
@@ -3466,6 +3328,10 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	intel_enable_pipe(dev_priv, pipe,
 			  intel_crtc->config.has_pch_encoder);
 	intel_enable_plane(dev_priv, plane, pipe);
+	intel_enable_planes(crtc);
+	intel_crtc_update_cursor(crtc, true);
+
+	hsw_enable_ips(intel_crtc);
 
 	if (intel_crtc->config.has_pch_encoder)
 		lpt_pch_enable(crtc);
@@ -3474,8 +3340,6 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	intel_update_fbc(dev);
 	mutex_unlock(&dev->struct_mutex);
 
-	intel_crtc_update_cursor(crtc, true);
-
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		encoder->enable(encoder);
 
@@ -3490,6 +3354,21 @@ static void haswell_crtc_enable(struct drm_crtc *crtc)
 	intel_wait_for_vblank(dev, intel_crtc->pipe);
 }
 
+static void ironlake_pfit_disable(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = crtc->pipe;
+
+	/* To avoid upsetting the power well on haswell only disable the pfit if
+	 * it's in use. The hw state code will make sure we get this right. */
+	if (crtc->config.pch_pfit.size) {
+		I915_WRITE(PF_CTL(pipe), 0);
+		I915_WRITE(PF_WIN_POS(pipe), 0);
+		I915_WRITE(PF_WIN_SZ(pipe), 0);
+	}
+}
+
 static void ironlake_crtc_disable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -3509,58 +3388,51 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
 
 	intel_crtc_wait_for_pending_flips(crtc);
 	drm_vblank_off(dev, pipe);
-	intel_crtc_update_cursor(crtc, false);
-
-	intel_disable_plane(dev_priv, plane, pipe);
 
 	if (dev_priv->cfb_plane == plane)
 		intel_disable_fbc(dev);
 
+	intel_crtc_update_cursor(crtc, false);
+	intel_disable_planes(crtc);
+	intel_disable_plane(dev_priv, plane, pipe);
+
+	if (intel_crtc->config.has_pch_encoder)
+		intel_set_pch_fifo_underrun_reporting(dev, pipe, false);
+
 	intel_disable_pipe(dev_priv, pipe);
 
-	/* Disable PF */
-	I915_WRITE(PF_CTL(pipe), 0);
-	I915_WRITE(PF_WIN_SZ(pipe), 0);
+	ironlake_pfit_disable(intel_crtc);
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		if (encoder->post_disable)
 			encoder->post_disable(encoder);
 
-	ironlake_fdi_disable(crtc);
-
-	ironlake_disable_pch_transcoder(dev_priv, pipe);
+	if (intel_crtc->config.has_pch_encoder) {
+		ironlake_fdi_disable(crtc);
 
-	if (HAS_PCH_CPT(dev)) {
-		/* disable TRANS_DP_CTL */
-		reg = TRANS_DP_CTL(pipe);
-		temp = I915_READ(reg);
-		temp &= ~(TRANS_DP_OUTPUT_ENABLE | TRANS_DP_PORT_SEL_MASK);
-		temp |= TRANS_DP_PORT_SEL_NONE;
-		I915_WRITE(reg, temp);
+		ironlake_disable_pch_transcoder(dev_priv, pipe);
+		intel_set_pch_fifo_underrun_reporting(dev, pipe, true);
 
-		/* disable DPLL_SEL */
-		temp = I915_READ(PCH_DPLL_SEL);
-		switch (pipe) {
-		case 0:
-			temp &= ~(TRANSA_DPLL_ENABLE | TRANSA_DPLLB_SEL);
-			break;
-		case 1:
-			temp &= ~(TRANSB_DPLL_ENABLE | TRANSB_DPLLB_SEL);
-			break;
-		case 2:
-			/* C shares PLL A or B */
-			temp &= ~(TRANSC_DPLL_ENABLE | TRANSC_DPLLB_SEL);
-			break;
-		default:
-			BUG(); /* wtf */
+		if (HAS_PCH_CPT(dev)) {
+			/* disable TRANS_DP_CTL */
+			reg = TRANS_DP_CTL(pipe);
+			temp = I915_READ(reg);
+			temp &= ~(TRANS_DP_OUTPUT_ENABLE |
+				  TRANS_DP_PORT_SEL_MASK);
+			temp |= TRANS_DP_PORT_SEL_NONE;
+			I915_WRITE(reg, temp);
+
+			/* disable DPLL_SEL */
+			temp = I915_READ(PCH_DPLL_SEL);
+			temp &= ~(TRANS_DPLL_ENABLE(pipe) | TRANS_DPLLB_SEL(pipe));
+			I915_WRITE(PCH_DPLL_SEL, temp);
 		}
-		I915_WRITE(PCH_DPLL_SEL, temp);
-	}
 
-	/* disable PCH DPLL */
-	intel_disable_pch_pll(intel_crtc);
+		/* disable PCH DPLL */
+		intel_disable_shared_dpll(intel_crtc);
 
-	ironlake_fdi_pll_disable(intel_crtc);
+		ironlake_fdi_pll_disable(intel_crtc);
+	}
 
 	intel_crtc->active = false;
 	intel_update_watermarks(dev);
@@ -3588,24 +3460,24 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 
 	intel_crtc_wait_for_pending_flips(crtc);
 	drm_vblank_off(dev, pipe);
-	intel_crtc_update_cursor(crtc, false);
-
-	intel_disable_plane(dev_priv, plane, pipe);
 
+	/* FBC must be disabled before disabling the plane on HSW. */
 	if (dev_priv->cfb_plane == plane)
 		intel_disable_fbc(dev);
 
+	hsw_disable_ips(intel_crtc);
+
+	intel_crtc_update_cursor(crtc, false);
+	intel_disable_planes(crtc);
+	intel_disable_plane(dev_priv, plane, pipe);
+
+	if (intel_crtc->config.has_pch_encoder)
+		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, false);
 	intel_disable_pipe(dev_priv, pipe);
 
 	intel_ddi_disable_transcoder_func(dev_priv, cpu_transcoder);
 
-	/* XXX: Once we have proper panel fitter state tracking implemented with
-	 * hardware state read/check support we should switch to only disable
-	 * the panel fitter when we know it's used. */
-	if (intel_using_power_well(dev)) {
-		I915_WRITE(PF_CTL(pipe), 0);
-		I915_WRITE(PF_WIN_SZ(pipe), 0);
-	}
+	ironlake_pfit_disable(intel_crtc);
 
 	intel_ddi_disable_pipe_clock(intel_crtc);
 
@@ -3615,6 +3487,7 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 
 	if (intel_crtc->config.has_pch_encoder) {
 		lpt_disable_pch_transcoder(dev_priv);
+		intel_set_pch_fifo_underrun_reporting(dev, TRANSCODER_A, true);
 		intel_ddi_fdi_disable(crtc);
 	}
 
@@ -3629,17 +3502,11 @@ static void haswell_crtc_disable(struct drm_crtc *crtc)
 static void ironlake_crtc_off(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	intel_put_pch_pll(intel_crtc);
+	intel_put_shared_dpll(intel_crtc);
 }
 
 static void haswell_crtc_off(struct drm_crtc *crtc)
 {
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-
-	/* Stop saying we're using TRANSCODER_EDP because some other CRTC might
-	 * start using it. */
-	intel_crtc->config.cpu_transcoder = (enum transcoder) intel_crtc->pipe;
-
 	intel_ddi_put_crtc_pll(crtc);
 }
 
@@ -3685,6 +3552,77 @@ g4x_fixup_plane(struct drm_i915_private *dev_priv, enum pipe pipe)
 	}
 }
 
+static void i9xx_pfit_enable(struct intel_crtc *crtc)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc_config *pipe_config = &crtc->config;
+
+	if (!crtc->config.gmch_pfit.control)
+		return;
+
+	/*
+	 * The panel fitter should only be adjusted whilst the pipe is disabled,
+	 * according to register description and PRM.
+	 */
+	WARN_ON(I915_READ(PFIT_CONTROL) & PFIT_ENABLE);
+	assert_pipe_disabled(dev_priv, crtc->pipe);
+
+	I915_WRITE(PFIT_PGM_RATIOS, pipe_config->gmch_pfit.pgm_ratios);
+	I915_WRITE(PFIT_CONTROL, pipe_config->gmch_pfit.control);
+
+	/* Border color in case we don't scale up to the full screen. Black by
+	 * default, change to something else for debugging. */
+	I915_WRITE(BCLRPAT(crtc->pipe), 0);
+}
+
+static void valleyview_crtc_enable(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_encoder *encoder;
+	int pipe = intel_crtc->pipe;
+	int plane = intel_crtc->plane;
+
+	WARN_ON(!crtc->enabled);
+
+	if (intel_crtc->active)
+		return;
+
+	intel_crtc->active = true;
+	intel_update_watermarks(dev);
+
+	mutex_lock(&dev_priv->dpio_lock);
+
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		if (encoder->pre_pll_enable)
+			encoder->pre_pll_enable(encoder);
+
+	intel_enable_pll(dev_priv, pipe);
+
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		if (encoder->pre_enable)
+			encoder->pre_enable(encoder);
+
+	/* VLV wants encoder enabling _before_ the pipe is up. */
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		encoder->enable(encoder);
+
+	i9xx_pfit_enable(intel_crtc);
+
+	intel_crtc_load_lut(crtc);
+
+	intel_enable_pipe(dev_priv, pipe, false);
+	intel_enable_plane(dev_priv, plane, pipe);
+	intel_enable_planes(crtc);
+	intel_crtc_update_cursor(crtc, true);
+
+	intel_update_fbc(dev);
+
+	mutex_unlock(&dev_priv->dpio_lock);
+}
+
 static void i9xx_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -3708,17 +3646,22 @@ static void i9xx_crtc_enable(struct drm_crtc *crtc)
 		if (encoder->pre_enable)
 			encoder->pre_enable(encoder);
 
+	i9xx_pfit_enable(intel_crtc);
+
+	intel_crtc_load_lut(crtc);
+
 	intel_enable_pipe(dev_priv, pipe, false);
 	intel_enable_plane(dev_priv, plane, pipe);
+	intel_enable_planes(crtc);
+	/* The fixup needs to happen before cursor is enabled */
 	if (IS_G4X(dev))
 		g4x_fixup_plane(dev_priv, pipe);
-
-	intel_crtc_load_lut(crtc);
-	intel_update_fbc(dev);
+	intel_crtc_update_cursor(crtc, true);
 
 	/* Give the overlay scaler a chance to enable if it's on this pipe */
 	intel_crtc_dpms_overlay(intel_crtc, true);
-	intel_crtc_update_cursor(crtc, true);
+
+	intel_update_fbc(dev);
 
 	for_each_encoder_on_crtc(dev, crtc, encoder)
 		encoder->enable(encoder);
@@ -3728,20 +3671,15 @@ static void i9xx_pfit_disable(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	enum pipe pipe;
-	uint32_t pctl = I915_READ(PFIT_CONTROL);
 
-	assert_pipe_disabled(dev_priv, crtc->pipe);
+	if (!crtc->config.gmch_pfit.control)
+		return;
 
-	if (INTEL_INFO(dev)->gen >= 4)
-		pipe = (pctl & PFIT_PIPE_MASK) >> PFIT_PIPE_SHIFT;
-	else
-		pipe = PIPE_B;
+	assert_pipe_disabled(dev_priv, crtc->pipe);
 
-	if (pipe == crtc->pipe) {
-		DRM_DEBUG_DRIVER("disabling pfit, current: 0x%08x\n", pctl);
-		I915_WRITE(PFIT_CONTROL, 0);
-	}
+	DRM_DEBUG_DRIVER("disabling pfit, current: 0x%08x\n",
+			 I915_READ(PFIT_CONTROL));
+	I915_WRITE(PFIT_CONTROL, 0);
 }
 
 static void i9xx_crtc_disable(struct drm_crtc *crtc)
@@ -3762,17 +3700,23 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 	/* Give the overlay scaler a chance to disable if it's on this pipe */
 	intel_crtc_wait_for_pending_flips(crtc);
 	drm_vblank_off(dev, pipe);
-	intel_crtc_dpms_overlay(intel_crtc, false);
-	intel_crtc_update_cursor(crtc, false);
 
 	if (dev_priv->cfb_plane == plane)
 		intel_disable_fbc(dev);
 
+	intel_crtc_dpms_overlay(intel_crtc, false);
+	intel_crtc_update_cursor(crtc, false);
+	intel_disable_planes(crtc);
 	intel_disable_plane(dev_priv, plane, pipe);
+
 	intel_disable_pipe(dev_priv, pipe);
 
 	i9xx_pfit_disable(intel_crtc);
 
+	for_each_encoder_on_crtc(dev, crtc, encoder)
+		if (encoder->post_disable)
+			encoder->post_disable(encoder);
+
 	intel_disable_pll(dev_priv, pipe);
 
 	intel_crtc->active = false;
@@ -3845,8 +3789,8 @@ static void intel_crtc_disable(struct drm_crtc *crtc)
 	/* crtc should still be enabled when we disable it. */
 	WARN_ON(!crtc->enabled);
 
-	intel_crtc->eld_vld = false;
 	dev_priv->display.crtc_disable(crtc);
+	intel_crtc->eld_vld = false;
 	intel_crtc_update_sarea(crtc, false);
 	dev_priv->display.off(crtc);
 
@@ -3977,17 +3921,131 @@ bool intel_connector_get_hw_state(struct intel_connector *connector)
 	return encoder->get_hw_state(encoder, &pipe);
 }
 
-static bool intel_crtc_compute_config(struct drm_crtc *crtc,
-				      struct intel_crtc_config *pipe_config)
+static bool ironlake_check_fdi_lanes(struct drm_device *dev, enum pipe pipe,
+				     struct intel_crtc_config *pipe_config)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *pipe_B_crtc =
+		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[PIPE_B]);
+
+	DRM_DEBUG_KMS("checking fdi config on pipe %c, lanes %i\n",
+		      pipe_name(pipe), pipe_config->fdi_lanes);
+	if (pipe_config->fdi_lanes > 4) {
+		DRM_DEBUG_KMS("invalid fdi lane config on pipe %c: %i lanes\n",
+			      pipe_name(pipe), pipe_config->fdi_lanes);
+		return false;
+	}
+
+	if (IS_HASWELL(dev)) {
+		if (pipe_config->fdi_lanes > 2) {
+			DRM_DEBUG_KMS("only 2 lanes on haswell, required: %i lanes\n",
+				      pipe_config->fdi_lanes);
+			return false;
+		} else {
+			return true;
+		}
+	}
+
+	if (INTEL_INFO(dev)->num_pipes == 2)
+		return true;
+
+	/* Ivybridge 3 pipe is really complicated */
+	switch (pipe) {
+	case PIPE_A:
+		return true;
+	case PIPE_B:
+		if (dev_priv->pipe_to_crtc_mapping[PIPE_C]->enabled &&
+		    pipe_config->fdi_lanes > 2) {
+			DRM_DEBUG_KMS("invalid shared fdi lane config on pipe %c: %i lanes\n",
+				      pipe_name(pipe), pipe_config->fdi_lanes);
+			return false;
+		}
+		return true;
+	case PIPE_C:
+		if (!pipe_has_enabled_pch(pipe_B_crtc) ||
+		    pipe_B_crtc->config.fdi_lanes <= 2) {
+			if (pipe_config->fdi_lanes > 2) {
+				DRM_DEBUG_KMS("invalid shared fdi lane config on pipe %c: %i lanes\n",
+					      pipe_name(pipe), pipe_config->fdi_lanes);
+				return false;
+			}
+		} else {
+			DRM_DEBUG_KMS("fdi link B uses too many lanes to enable link C\n");
+			return false;
+		}
+		return true;
+	default:
+		BUG();
+	}
+}
+
+#define RETRY 1
+static int ironlake_fdi_compute_config(struct intel_crtc *intel_crtc,
+				       struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = intel_crtc->base.dev;
+	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
+	int lane, link_bw, fdi_dotclock;
+	bool setup_ok, needs_recompute = false;
+
+retry:
+	/* FDI is a binary signal running at ~2.7GHz, encoding
+	 * each output octet as 10 bits. The actual frequency
+	 * is stored as a divider into a 100MHz clock, and the
+	 * mode pixel clock is stored in units of 1KHz.
+	 * Hence the bw of each lane in terms of the mode signal
+	 * is:
+	 */
+	link_bw = intel_fdi_link_freq(dev) * MHz(100)/KHz(1)/10;
+
+	fdi_dotclock = adjusted_mode->clock;
+	fdi_dotclock /= pipe_config->pixel_multiplier;
+
+	lane = ironlake_get_lanes_required(fdi_dotclock, link_bw,
+					   pipe_config->pipe_bpp);
+
+	pipe_config->fdi_lanes = lane;
+
+	intel_link_compute_m_n(pipe_config->pipe_bpp, lane, fdi_dotclock,
+			       link_bw, &pipe_config->fdi_m_n);
+
+	setup_ok = ironlake_check_fdi_lanes(intel_crtc->base.dev,
+					    intel_crtc->pipe, pipe_config);
+	if (!setup_ok && pipe_config->pipe_bpp > 6*3) {
+		pipe_config->pipe_bpp -= 2*3;
+		DRM_DEBUG_KMS("fdi link bw constraint, reducing pipe bpp to %i\n",
+			      pipe_config->pipe_bpp);
+		needs_recompute = true;
+		pipe_config->bw_constrained = true;
+
+		goto retry;
+	}
+
+	if (needs_recompute)
+		return RETRY;
+
+	return setup_ok ? 0 : -EINVAL;
+}
+
+static void hsw_compute_ips_config(struct intel_crtc *crtc,
+				   struct intel_crtc_config *pipe_config)
+{
+	pipe_config->ips_enabled = i915_enable_ips &&
+				   hsw_crtc_supports_ips(crtc) &&
+				   pipe_config->pipe_bpp == 24;
+}
+
+static int intel_crtc_compute_config(struct intel_crtc *crtc,
+				     struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
 
 	if (HAS_PCH_SPLIT(dev)) {
 		/* FDI link clock is fixed at 2.7G */
 		if (pipe_config->requested_mode.clock * 3
 		    > IRONLAKE_FDI_FREQ * 4)
-			return false;
+			return -EINVAL;
 	}
 
 	/* All interlaced capable intel hw wants timings in frames. Note though
@@ -3996,12 +4054,12 @@ static bool intel_crtc_compute_config(struct drm_crtc *crtc,
 	if (!pipe_config->timings_set)
 		drm_mode_set_crtcinfo(adjusted_mode, 0);
 
-	/* WaPruneModeWithIncorrectHsyncOffset: Cantiga+ cannot handle modes
-	 * with a hsync front porch of 0.
+	/* Cantiga+ cannot handle modes with a hsync front porch of 0.
+	 * WaPruneModeWithIncorrectHsyncOffset:ctg,elk,ilk,snb,ivb,vlv,hsw.
 	 */
 	if ((INTEL_INFO(dev)->gen > 4 || IS_G4X(dev)) &&
 		adjusted_mode->hsync_start == adjusted_mode->hdisplay)
-		return false;
+		return -EINVAL;
 
 	if ((IS_G4X(dev) || IS_VALLEYVIEW(dev)) && pipe_config->pipe_bpp > 10*3) {
 		pipe_config->pipe_bpp = 10*3; /* 12bpc is gen5+ */
@@ -4011,7 +4069,18 @@ static bool intel_crtc_compute_config(struct drm_crtc *crtc,
 		pipe_config->pipe_bpp = 8*3;
 	}
 
-	return true;
+	if (HAS_IPS(dev))
+		hsw_compute_ips_config(crtc, pipe_config);
+
+	/* XXX: PCH clock sharing is done in ->mode_set, so make sure the old
+	 * clock survives for now. */
+	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
+		pipe_config->shared_dpll = crtc->config.shared_dpll;
+
+	if (pipe_config->has_pch_encoder)
+		return ironlake_fdi_compute_config(crtc, pipe_config);
+
+	return 0;
 }
 
 static int valleyview_get_display_clock_speed(struct drm_device *dev)
@@ -4120,7 +4189,7 @@ static inline bool intel_panel_use_ssc(struct drm_i915_private *dev_priv)
 {
 	if (i915_panel_use_ssc >= 0)
 		return i915_panel_use_ssc != 0;
-	return dev_priv->lvds_use_ssc
+	return dev_priv->vbt.lvds_use_ssc
 		&& !(dev_priv->quirks & QUIRK_LVDS_SSC_DISABLE);
 }
 
@@ -4156,7 +4225,7 @@ static int i9xx_get_refclk(struct drm_crtc *crtc, int num_connectors)
 		refclk = vlv_get_refclk(crtc);
 	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) &&
 	    intel_panel_use_ssc(dev_priv) && num_connectors < 2) {
-		refclk = dev_priv->lvds_ssc_freq * 1000;
+		refclk = dev_priv->vbt.lvds_ssc_freq * 1000;
 		DRM_DEBUG_KMS("using SSC reference clock of %d MHz\n",
 			      refclk / 1000);
 	} else if (!IS_GEN2(dev)) {
@@ -4168,28 +4237,14 @@ static int i9xx_get_refclk(struct drm_crtc *crtc, int num_connectors)
 	return refclk;
 }
 
-static void i9xx_adjust_sdvo_tv_clock(struct intel_crtc *crtc)
+static uint32_t pnv_dpll_compute_fp(struct dpll *dpll)
 {
-	unsigned dotclock = crtc->config.adjusted_mode.clock;
-	struct dpll *clock = &crtc->config.dpll;
-
-	/* SDVO TV has fixed PLL values depend on its clock range,
-	   this mirrors vbios setting. */
-	if (dotclock >= 100000 && dotclock < 140500) {
-		clock->p1 = 2;
-		clock->p2 = 10;
-		clock->n = 3;
-		clock->m1 = 16;
-		clock->m2 = 8;
-	} else if (dotclock >= 140500 && dotclock <= 200000) {
-		clock->p1 = 1;
-		clock->p2 = 10;
-		clock->n = 6;
-		clock->m1 = 12;
-		clock->m2 = 8;
-	}
+	return (1 << dpll->n) << 16 | dpll->m2;
+}
 
-	crtc->config.clock_set = true;
+static uint32_t i9xx_dpll_compute_fp(struct dpll *dpll)
+{
+	return dpll->n << 16 | dpll->m1 << 8 | dpll->m2;
 }
 
 static void i9xx_update_pll_dividers(struct intel_crtc *crtc,
@@ -4199,18 +4254,15 @@ static void i9xx_update_pll_dividers(struct intel_crtc *crtc,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int pipe = crtc->pipe;
 	u32 fp, fp2 = 0;
-	struct dpll *clock = &crtc->config.dpll;
 
 	if (IS_PINEVIEW(dev)) {
-		fp = (1 << clock->n) << 16 | clock->m1 << 8 | clock->m2;
+		fp = pnv_dpll_compute_fp(&crtc->config.dpll);
 		if (reduced_clock)
-			fp2 = (1 << reduced_clock->n) << 16 |
-				reduced_clock->m1 << 8 | reduced_clock->m2;
+			fp2 = pnv_dpll_compute_fp(reduced_clock);
 	} else {
-		fp = clock->n << 16 | clock->m1 << 8 | clock->m2;
+		fp = i9xx_dpll_compute_fp(&crtc->config.dpll);
 		if (reduced_clock)
-			fp2 = reduced_clock->n << 16 | reduced_clock->m1 << 8 |
-				reduced_clock->m2;
+			fp2 = i9xx_dpll_compute_fp(reduced_clock);
 	}
 
 	I915_WRITE(FP0(pipe), fp);
@@ -4225,6 +4277,68 @@ static void i9xx_update_pll_dividers(struct intel_crtc *crtc,
 	}
 }
 
+static void vlv_pllb_recal_opamp(struct drm_i915_private *dev_priv)
+{
+	u32 reg_val;
+
+	/*
+	 * PLLB opamp always calibrates to max value of 0x3f, force enable it
+	 * and set it to a reasonable value instead.
+	 */
+	reg_val = vlv_dpio_read(dev_priv, DPIO_IREF(1));
+	reg_val &= 0xffffff00;
+	reg_val |= 0x00000030;
+	vlv_dpio_write(dev_priv, DPIO_IREF(1), reg_val);
+
+	reg_val = vlv_dpio_read(dev_priv, DPIO_CALIBRATION);
+	reg_val &= 0x8cffffff;
+	reg_val = 0x8c000000;
+	vlv_dpio_write(dev_priv, DPIO_CALIBRATION, reg_val);
+
+	reg_val = vlv_dpio_read(dev_priv, DPIO_IREF(1));
+	reg_val &= 0xffffff00;
+	vlv_dpio_write(dev_priv, DPIO_IREF(1), reg_val);
+
+	reg_val = vlv_dpio_read(dev_priv, DPIO_CALIBRATION);
+	reg_val &= 0x00ffffff;
+	reg_val |= 0xb0000000;
+	vlv_dpio_write(dev_priv, DPIO_CALIBRATION, reg_val);
+}
+
+static void intel_pch_transcoder_set_m_n(struct intel_crtc *crtc,
+					 struct intel_link_m_n *m_n)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = crtc->pipe;
+
+	I915_WRITE(PCH_TRANS_DATA_M1(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
+	I915_WRITE(PCH_TRANS_DATA_N1(pipe), m_n->gmch_n);
+	I915_WRITE(PCH_TRANS_LINK_M1(pipe), m_n->link_m);
+	I915_WRITE(PCH_TRANS_LINK_N1(pipe), m_n->link_n);
+}
+
+static void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
+					 struct intel_link_m_n *m_n)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe = crtc->pipe;
+	enum transcoder transcoder = crtc->config.cpu_transcoder;
+
+	if (INTEL_INFO(dev)->gen >= 5) {
+		I915_WRITE(PIPE_DATA_M1(transcoder), TU_SIZE(m_n->tu) | m_n->gmch_m);
+		I915_WRITE(PIPE_DATA_N1(transcoder), m_n->gmch_n);
+		I915_WRITE(PIPE_LINK_M1(transcoder), m_n->link_m);
+		I915_WRITE(PIPE_LINK_N1(transcoder), m_n->link_n);
+	} else {
+		I915_WRITE(PIPE_DATA_M_G4X(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
+		I915_WRITE(PIPE_DATA_N_G4X(pipe), m_n->gmch_n);
+		I915_WRITE(PIPE_LINK_M_G4X(pipe), m_n->link_m);
+		I915_WRITE(PIPE_LINK_N_G4X(pipe), m_n->link_n);
+	}
+}
+
 static void intel_dp_set_m_n(struct intel_crtc *crtc)
 {
 	if (crtc->config.has_pch_encoder)
@@ -4237,24 +4351,16 @@ static void vlv_update_pll(struct intel_crtc *crtc)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_encoder *encoder;
 	int pipe = crtc->pipe;
-	u32 dpll, mdiv, pdiv;
+	u32 dpll, mdiv;
 	u32 bestn, bestm1, bestm2, bestp1, bestp2;
-	bool is_sdvo;
-	u32 temp;
+	bool is_hdmi;
+	u32 coreclk, reg_val, dpll_md;
 
 	mutex_lock(&dev_priv->dpio_lock);
 
-	is_sdvo = intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_SDVO) ||
-		intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI);
-
-	dpll = DPLL_VGA_MODE_DIS;
-	dpll |= DPLL_EXT_BUFFER_ENABLE_VLV;
-	dpll |= DPLL_REFA_CLK_ENABLE_VLV;
-	dpll |= DPLL_INTEGRATED_CLOCK_VLV;
-
-	I915_WRITE(DPLL(pipe), dpll);
-	POSTING_READ(DPLL(pipe));
+	is_hdmi = intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI);
 
 	bestn = crtc->config.dpll.n;
 	bestm1 = crtc->config.dpll.m1;
@@ -4262,72 +4368,104 @@ static void vlv_update_pll(struct intel_crtc *crtc)
 	bestp1 = crtc->config.dpll.p1;
 	bestp2 = crtc->config.dpll.p2;
 
-	/*
-	 * In Valleyview PLL and program lane counter registers are exposed
-	 * through DPIO interface
-	 */
+	/* See eDP HDMI DPIO driver vbios notes doc */
+
+	/* PLL B needs special handling */
+	if (pipe)
+		vlv_pllb_recal_opamp(dev_priv);
+
+	/* Set up Tx target for periodic Rcomp update */
+	vlv_dpio_write(dev_priv, DPIO_IREF_BCAST, 0x0100000f);
+
+	/* Disable target IRef on PLL */
+	reg_val = vlv_dpio_read(dev_priv, DPIO_IREF_CTL(pipe));
+	reg_val &= 0x00ffffff;
+	vlv_dpio_write(dev_priv, DPIO_IREF_CTL(pipe), reg_val);
+
+	/* Disable fast lock */
+	vlv_dpio_write(dev_priv, DPIO_FASTCLK_DISABLE, 0x610);
+
+	/* Set idtafcrecal before PLL is enabled */
 	mdiv = ((bestm1 << DPIO_M1DIV_SHIFT) | (bestm2 & DPIO_M2DIV_MASK));
 	mdiv |= ((bestp1 << DPIO_P1_SHIFT) | (bestp2 << DPIO_P2_SHIFT));
 	mdiv |= ((bestn << DPIO_N_SHIFT));
-	mdiv |= (1 << DPIO_POST_DIV_SHIFT);
 	mdiv |= (1 << DPIO_K_SHIFT);
+
+	/*
+	 * Post divider depends on pixel clock rate, DAC vs digital (and LVDS,
+	 * but we don't support that).
+	 * Note: don't use the DAC post divider as it seems unstable.
+	 */
+	mdiv |= (DPIO_POST_DIV_HDMIDP << DPIO_POST_DIV_SHIFT);
+	vlv_dpio_write(dev_priv, DPIO_DIV(pipe), mdiv);
+
 	mdiv |= DPIO_ENABLE_CALIBRATION;
-	intel_dpio_write(dev_priv, DPIO_DIV(pipe), mdiv);
+	vlv_dpio_write(dev_priv, DPIO_DIV(pipe), mdiv);
+
+	/* Set HBR and RBR LPF coefficients */
+	if (crtc->config.port_clock == 162000 ||
+	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_ANALOG) ||
+	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI))
+		vlv_dpio_write(dev_priv, DPIO_LPF_COEFF(pipe),
+				 0x005f0021);
+	else
+		vlv_dpio_write(dev_priv, DPIO_LPF_COEFF(pipe),
+				 0x00d0000f);
+
+	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_EDP) ||
+	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT)) {
+		/* Use SSC source */
+		if (!pipe)
+			vlv_dpio_write(dev_priv, DPIO_REFSFR(pipe),
+					 0x0df40000);
+		else
+			vlv_dpio_write(dev_priv, DPIO_REFSFR(pipe),
+					 0x0df70000);
+	} else { /* HDMI or VGA */
+		/* Use bend source */
+		if (!pipe)
+			vlv_dpio_write(dev_priv, DPIO_REFSFR(pipe),
+					 0x0df70000);
+		else
+			vlv_dpio_write(dev_priv, DPIO_REFSFR(pipe),
+					 0x0df40000);
+	}
 
-	intel_dpio_write(dev_priv, DPIO_CORE_CLK(pipe), 0x01000000);
+	coreclk = vlv_dpio_read(dev_priv, DPIO_CORE_CLK(pipe));
+	coreclk = (coreclk & 0x0000ff00) | 0x01c00000;
+	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT) ||
+	    intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_EDP))
+		coreclk |= 0x01000000;
+	vlv_dpio_write(dev_priv, DPIO_CORE_CLK(pipe), coreclk);
 
-	pdiv = (1 << DPIO_REFSEL_OVERRIDE) | (5 << DPIO_PLL_MODESEL_SHIFT) |
-		(3 << DPIO_BIAS_CURRENT_CTL_SHIFT) | (1<<20) |
-		(7 << DPIO_PLL_REFCLK_SEL_SHIFT) | (8 << DPIO_DRIVER_CTL_SHIFT) |
-		(5 << DPIO_CLK_BIAS_CTL_SHIFT);
-	intel_dpio_write(dev_priv, DPIO_REFSFR(pipe), pdiv);
+	vlv_dpio_write(dev_priv, DPIO_PLL_CML(pipe), 0x87871000);
 
-	intel_dpio_write(dev_priv, DPIO_LFP_COEFF(pipe), 0x005f003b);
+	for_each_encoder_on_crtc(dev, &crtc->base, encoder)
+		if (encoder->pre_pll_enable)
+			encoder->pre_pll_enable(encoder);
+
+	/* Enable DPIO clock input */
+	dpll = DPLL_EXT_BUFFER_ENABLE_VLV | DPLL_REFA_CLK_ENABLE_VLV |
+		DPLL_VGA_MODE_DIS | DPLL_INTEGRATED_CLOCK_VLV;
+	if (pipe)
+		dpll |= DPLL_INTEGRATED_CRI_CLK_VLV;
 
 	dpll |= DPLL_VCO_ENABLE;
 	I915_WRITE(DPLL(pipe), dpll);
 	POSTING_READ(DPLL(pipe));
+	udelay(150);
+
 	if (wait_for(((I915_READ(DPLL(pipe)) & DPLL_LOCK_VLV) == DPLL_LOCK_VLV), 1))
 		DRM_ERROR("DPLL %d failed to lock\n", pipe);
 
-	intel_dpio_write(dev_priv, DPIO_FASTCLK_DISABLE, 0x620);
+	dpll_md = (crtc->config.pixel_multiplier - 1)
+		<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
+	I915_WRITE(DPLL_MD(pipe), dpll_md);
+	POSTING_READ(DPLL_MD(pipe));
 
 	if (crtc->config.has_dp_encoder)
 		intel_dp_set_m_n(crtc);
 
-	I915_WRITE(DPLL(pipe), dpll);
-
-	/* Wait for the clocks to stabilize. */
-	POSTING_READ(DPLL(pipe));
-	udelay(150);
-
-	temp = 0;
-	if (is_sdvo) {
-		temp = 0;
-		if (crtc->config.pixel_multiplier > 1) {
-			temp = (crtc->config.pixel_multiplier - 1)
-				<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
-		}
-	}
-	I915_WRITE(DPLL_MD(pipe), temp);
-	POSTING_READ(DPLL_MD(pipe));
-
-	/* Now program lane control registers */
-	if(intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT)
-	   || intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_HDMI)) {
-		temp = 0x1000C4;
-		if(pipe == 1)
-			temp |= (1 << 21);
-		intel_dpio_write(dev_priv, DPIO_DATA_CHANNEL1, temp);
-	}
-
-	if(intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_EDP)) {
-		temp = 0x1000C4;
-		if(pipe == 1)
-			temp |= (1 << 21);
-		intel_dpio_write(dev_priv, DPIO_DATA_CHANNEL2, temp);
-	}
-
 	mutex_unlock(&dev_priv->dpio_lock);
 }
 
@@ -4355,14 +4493,14 @@ static void i9xx_update_pll(struct intel_crtc *crtc,
 	else
 		dpll |= DPLLB_MODE_DAC_SERIAL;
 
-	if (is_sdvo) {
-		if ((crtc->config.pixel_multiplier > 1) &&
-		    (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev))) {
-			dpll |= (crtc->config.pixel_multiplier - 1)
-				<< SDVO_MULTIPLIER_SHIFT_HIRES;
-		}
-		dpll |= DPLL_DVO_HIGH_SPEED;
+	if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev)) {
+		dpll |= (crtc->config.pixel_multiplier - 1)
+			<< SDVO_MULTIPLIER_SHIFT_HIRES;
 	}
+
+	if (is_sdvo)
+		dpll |= DPLL_DVO_HIGH_SPEED;
+
 	if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_DISPLAYPORT))
 		dpll |= DPLL_DVO_HIGH_SPEED;
 
@@ -4391,12 +4529,8 @@ static void i9xx_update_pll(struct intel_crtc *crtc,
 	if (INTEL_INFO(dev)->gen >= 4)
 		dpll |= (6 << PLL_LOAD_PULSE_PHASE_SHIFT);
 
-	if (is_sdvo && intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_TVOUT))
+	if (crtc->config.sdvo_tv_clock)
 		dpll |= PLL_REF_INPUT_TVCLKINBC;
-	else if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_TVOUT))
-		/* XXX: just matching BIOS for now */
-		/*	dpll |= PLL_REF_INPUT_TVCLKINBC; */
-		dpll |= 3;
 	else if (intel_pipe_has_type(&crtc->base, INTEL_OUTPUT_LVDS) &&
 		 intel_panel_use_ssc(dev_priv) && num_connectors < 2)
 		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
@@ -4422,15 +4556,9 @@ static void i9xx_update_pll(struct intel_crtc *crtc,
 	udelay(150);
 
 	if (INTEL_INFO(dev)->gen >= 4) {
-		u32 temp = 0;
-		if (is_sdvo) {
-			temp = 0;
-			if (crtc->config.pixel_multiplier > 1) {
-				temp = (crtc->config.pixel_multiplier - 1)
-					<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
-			}
-		}
-		I915_WRITE(DPLL_MD(pipe), temp);
+		u32 dpll_md = (crtc->config.pixel_multiplier - 1)
+			<< DPLL_MD_UDI_MULTIPLIER_SHIFT;
+		I915_WRITE(DPLL_MD(pipe), dpll_md);
 	} else {
 		/* The pixel multiplier can only be updated once the
 		 * DPLL is enabled and the clocks are stable.
@@ -4442,7 +4570,6 @@ static void i9xx_update_pll(struct intel_crtc *crtc,
 }
 
 static void i8xx_update_pll(struct intel_crtc *crtc,
-			    struct drm_display_mode *adjusted_mode,
 			    intel_clock_t *reduced_clock,
 			    int num_connectors)
 {
@@ -4497,20 +4624,26 @@ static void i8xx_update_pll(struct intel_crtc *crtc,
 	I915_WRITE(DPLL(pipe), dpll);
 }
 
-static void intel_set_pipe_timings(struct intel_crtc *intel_crtc,
-				   struct drm_display_mode *mode,
-				   struct drm_display_mode *adjusted_mode)
+static void intel_set_pipe_timings(struct intel_crtc *intel_crtc)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum pipe pipe = intel_crtc->pipe;
 	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
-	uint32_t vsyncshift;
+	struct drm_display_mode *adjusted_mode =
+		&intel_crtc->config.adjusted_mode;
+	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
+	uint32_t vsyncshift, crtc_vtotal, crtc_vblank_end;
+
+	/* We need to be careful not to changed the adjusted mode, for otherwise
+	 * the hw state checker will get angry at the mismatch. */
+	crtc_vtotal = adjusted_mode->crtc_vtotal;
+	crtc_vblank_end = adjusted_mode->crtc_vblank_end;
 
 	if (!IS_GEN2(dev) && adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
 		/* the chip adds 2 halflines automatically */
-		adjusted_mode->crtc_vtotal -= 1;
-		adjusted_mode->crtc_vblank_end -= 1;
+		crtc_vtotal -= 1;
+		crtc_vblank_end -= 1;
 		vsyncshift = adjusted_mode->crtc_hsync_start
 			     - adjusted_mode->crtc_htotal / 2;
 	} else {
@@ -4532,10 +4665,10 @@ static void intel_set_pipe_timings(struct intel_crtc *intel_crtc,
 
 	I915_WRITE(VTOTAL(cpu_transcoder),
 		   (adjusted_mode->crtc_vdisplay - 1) |
-		   ((adjusted_mode->crtc_vtotal - 1) << 16));
+		   ((crtc_vtotal - 1) << 16));
 	I915_WRITE(VBLANK(cpu_transcoder),
 		   (adjusted_mode->crtc_vblank_start - 1) |
-		   ((adjusted_mode->crtc_vblank_end - 1) << 16));
+		   ((crtc_vblank_end - 1) << 16));
 	I915_WRITE(VSYNC(cpu_transcoder),
 		   (adjusted_mode->crtc_vsync_start - 1) |
 		   ((adjusted_mode->crtc_vsync_end - 1) << 16));
@@ -4555,13 +4688,52 @@ static void intel_set_pipe_timings(struct intel_crtc *intel_crtc,
 		   ((mode->hdisplay - 1) << 16) | (mode->vdisplay - 1));
 }
 
+static void intel_get_pipe_timings(struct intel_crtc *crtc,
+				   struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum transcoder cpu_transcoder = pipe_config->cpu_transcoder;
+	uint32_t tmp;
+
+	tmp = I915_READ(HTOTAL(cpu_transcoder));
+	pipe_config->adjusted_mode.crtc_hdisplay = (tmp & 0xffff) + 1;
+	pipe_config->adjusted_mode.crtc_htotal = ((tmp >> 16) & 0xffff) + 1;
+	tmp = I915_READ(HBLANK(cpu_transcoder));
+	pipe_config->adjusted_mode.crtc_hblank_start = (tmp & 0xffff) + 1;
+	pipe_config->adjusted_mode.crtc_hblank_end = ((tmp >> 16) & 0xffff) + 1;
+	tmp = I915_READ(HSYNC(cpu_transcoder));
+	pipe_config->adjusted_mode.crtc_hsync_start = (tmp & 0xffff) + 1;
+	pipe_config->adjusted_mode.crtc_hsync_end = ((tmp >> 16) & 0xffff) + 1;
+
+	tmp = I915_READ(VTOTAL(cpu_transcoder));
+	pipe_config->adjusted_mode.crtc_vdisplay = (tmp & 0xffff) + 1;
+	pipe_config->adjusted_mode.crtc_vtotal = ((tmp >> 16) & 0xffff) + 1;
+	tmp = I915_READ(VBLANK(cpu_transcoder));
+	pipe_config->adjusted_mode.crtc_vblank_start = (tmp & 0xffff) + 1;
+	pipe_config->adjusted_mode.crtc_vblank_end = ((tmp >> 16) & 0xffff) + 1;
+	tmp = I915_READ(VSYNC(cpu_transcoder));
+	pipe_config->adjusted_mode.crtc_vsync_start = (tmp & 0xffff) + 1;
+	pipe_config->adjusted_mode.crtc_vsync_end = ((tmp >> 16) & 0xffff) + 1;
+
+	if (I915_READ(PIPECONF(cpu_transcoder)) & PIPECONF_INTERLACE_MASK) {
+		pipe_config->adjusted_mode.flags |= DRM_MODE_FLAG_INTERLACE;
+		pipe_config->adjusted_mode.crtc_vtotal += 1;
+		pipe_config->adjusted_mode.crtc_vblank_end += 1;
+	}
+
+	tmp = I915_READ(PIPESRC(crtc->pipe));
+	pipe_config->requested_mode.vdisplay = (tmp & 0xffff) + 1;
+	pipe_config->requested_mode.hdisplay = ((tmp >> 16) & 0xffff) + 1;
+}
+
 static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t pipeconf;
 
-	pipeconf = I915_READ(PIPECONF(intel_crtc->pipe));
+	pipeconf = 0;
 
 	if (intel_crtc->pipe == 0 && INTEL_INFO(dev)->gen < 4) {
 		/* Enable pixel doubling when the dot clock is > 90% of the (display)
@@ -4573,26 +4745,28 @@ static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
 		if (intel_crtc->config.requested_mode.clock >
 		    dev_priv->display.get_display_clock_speed(dev) * 9 / 10)
 			pipeconf |= PIPECONF_DOUBLE_WIDE;
-		else
-			pipeconf &= ~PIPECONF_DOUBLE_WIDE;
 	}
 
-	/* default to 8bpc */
-	pipeconf &= ~(PIPECONF_BPC_MASK | PIPECONF_DITHER_EN);
-	if (intel_crtc->config.has_dp_encoder) {
-		if (intel_crtc->config.dither) {
-			pipeconf |= PIPECONF_6BPC |
-				    PIPECONF_DITHER_EN |
+	/* only g4x and later have fancy bpc/dither controls */
+	if (IS_G4X(dev) || IS_VALLEYVIEW(dev)) {
+		/* Bspec claims that we can't use dithering for 30bpp pipes. */
+		if (intel_crtc->config.dither && intel_crtc->config.pipe_bpp != 30)
+			pipeconf |= PIPECONF_DITHER_EN |
 				    PIPECONF_DITHER_TYPE_SP;
-		}
-	}
 
-	if (IS_VALLEYVIEW(dev) && intel_pipe_has_type(&intel_crtc->base,
-						      INTEL_OUTPUT_EDP)) {
-		if (intel_crtc->config.dither) {
-			pipeconf |= PIPECONF_6BPC |
-					PIPECONF_ENABLE |
-					I965_PIPECONF_ACTIVE;
+		switch (intel_crtc->config.pipe_bpp) {
+		case 18:
+			pipeconf |= PIPECONF_6BPC;
+			break;
+		case 24:
+			pipeconf |= PIPECONF_8BPC;
+			break;
+		case 30:
+			pipeconf |= PIPECONF_10BPC;
+			break;
+		default:
+			/* Case prevented by intel_choose_pipe_bpp_dither. */
+			BUG();
 		}
 	}
 
@@ -4602,23 +4776,17 @@ static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc)
 			pipeconf |= PIPECONF_CXSR_DOWNCLOCK;
 		} else {
 			DRM_DEBUG_KMS("disabling CxSR downclocking\n");
-			pipeconf &= ~PIPECONF_CXSR_DOWNCLOCK;
 		}
 	}
 
-	pipeconf &= ~PIPECONF_INTERLACE_MASK;
 	if (!IS_GEN2(dev) &&
 	    intel_crtc->config.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
 		pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION;
 	else
 		pipeconf |= PIPECONF_PROGRESSIVE;
 
-	if (IS_VALLEYVIEW(dev)) {
-		if (intel_crtc->config.limited_color_range)
-			pipeconf |= PIPECONF_COLOR_RANGE_SELECT;
-		else
-			pipeconf &= ~PIPECONF_COLOR_RANGE_SELECT;
-	}
+	if (IS_VALLEYVIEW(dev) && intel_crtc->config.limited_color_range)
+		pipeconf |= PIPECONF_COLOR_RANGE_SELECT;
 
 	I915_WRITE(PIPECONF(intel_crtc->pipe), pipeconf);
 	POSTING_READ(PIPECONF(intel_crtc->pipe));
@@ -4631,16 +4799,14 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_display_mode *adjusted_mode =
-		&intel_crtc->config.adjusted_mode;
 	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
 	int refclk, num_connectors = 0;
 	intel_clock_t clock, reduced_clock;
 	u32 dspcntr;
-	bool ok, has_reduced_clock = false, is_sdvo = false;
-	bool is_lvds = false, is_tv = false;
+	bool ok, has_reduced_clock = false;
+	bool is_lvds = false;
 	struct intel_encoder *encoder;
 	const intel_limit_t *limit;
 	int ret;
@@ -4650,15 +4816,6 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		case INTEL_OUTPUT_LVDS:
 			is_lvds = true;
 			break;
-		case INTEL_OUTPUT_SDVO:
-		case INTEL_OUTPUT_HDMI:
-			is_sdvo = true;
-			if (encoder->needs_tv_clock)
-				is_tv = true;
-			break;
-		case INTEL_OUTPUT_TVOUT:
-			is_tv = true;
-			break;
 		}
 
 		num_connectors++;
@@ -4672,9 +4829,10 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	 * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
 	 */
 	limit = intel_limit(crtc, refclk);
-	ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, NULL,
-			     &clock);
-	if (!ok) {
+	ok = dev_priv->display.find_dpll(limit, crtc,
+					 intel_crtc->config.port_clock,
+					 refclk, NULL, &clock);
+	if (!ok && !intel_crtc->config.clock_set) {
 		DRM_ERROR("Couldn't find PLL settings for mode!\n");
 		return -EINVAL;
 	}
@@ -4689,10 +4847,10 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		 * by using the FP0/FP1. In such case we will disable the LVDS
 		 * downclock feature.
 		*/
-		has_reduced_clock = limit->find_pll(limit, crtc,
+		has_reduced_clock =
+			dev_priv->display.find_dpll(limit, crtc,
 						    dev_priv->lvds_downclock,
-						    refclk,
-						    &clock,
+						    refclk, &clock,
 						    &reduced_clock);
 	}
 	/* Compat-code for transition, will disappear. */
@@ -4704,11 +4862,8 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		intel_crtc->config.dpll.p2 = clock.p2;
 	}
 
-	if (is_sdvo && is_tv)
-		i9xx_adjust_sdvo_tv_clock(intel_crtc);
-
 	if (IS_GEN2(dev))
-		i8xx_update_pll(intel_crtc, adjusted_mode,
+		i8xx_update_pll(intel_crtc,
 				has_reduced_clock ? &reduced_clock : NULL,
 				num_connectors);
 	else if (IS_VALLEYVIEW(dev))
@@ -4716,7 +4871,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	else
 		i9xx_update_pll(intel_crtc,
 				has_reduced_clock ? &reduced_clock : NULL,
-				num_connectors);
+                                num_connectors);
 
 	/* Set up the display plane register */
 	dspcntr = DISPPLANE_GAMMA_ENABLE;
@@ -4728,10 +4883,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 			dspcntr |= DISPPLANE_SEL_PIPE_B;
 	}
 
-	DRM_DEBUG_KMS("Mode for pipe %c:\n", pipe == 0 ? 'A' : 'B');
-	drm_mode_debug_printmodeline(mode);
-
-	intel_set_pipe_timings(intel_crtc, mode, adjusted_mode);
+	intel_set_pipe_timings(intel_crtc);
 
 	/* pipesrc and dspsize control the size that is scaled from,
 	 * which should always be the user's requested size.
@@ -4743,10 +4895,6 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 
 	i9xx_set_pipeconf(intel_crtc);
 
-	intel_enable_pipe(dev_priv, pipe, false);
-
-	intel_wait_for_vblank(dev, pipe);
-
 	I915_WRITE(DSPCNTR(plane), dspcntr);
 	POSTING_READ(DSPCNTR(plane));
 
@@ -4757,6 +4905,33 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	return ret;
 }
 
+static void i9xx_get_pfit_config(struct intel_crtc *crtc,
+				 struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t tmp;
+
+	tmp = I915_READ(PFIT_CONTROL);
+	if (!(tmp & PFIT_ENABLE))
+		return;
+
+	/* Check whether the pfit is attached to our pipe. */
+	if (INTEL_INFO(dev)->gen < 4) {
+		if (crtc->pipe != PIPE_B)
+			return;
+	} else {
+		if ((tmp & PFIT_PIPE_MASK) != (crtc->pipe << PFIT_PIPE_SHIFT))
+			return;
+	}
+
+	pipe_config->gmch_pfit.control = tmp;
+	pipe_config->gmch_pfit.pgm_ratios = I915_READ(PFIT_PGM_RATIOS);
+	if (INTEL_INFO(dev)->gen < 5)
+		pipe_config->gmch_pfit.lvds_border_bits =
+			I915_READ(LVDS) & LVDS_BORDER_ENABLE;
+}
+
 static bool i9xx_get_pipe_config(struct intel_crtc *crtc,
 				 struct intel_crtc_config *pipe_config)
 {
@@ -4764,10 +4939,34 @@ static bool i9xx_get_pipe_config(struct intel_crtc *crtc,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t tmp;
 
+	pipe_config->cpu_transcoder = crtc->pipe;
+	pipe_config->shared_dpll = DPLL_ID_PRIVATE;
+
 	tmp = I915_READ(PIPECONF(crtc->pipe));
 	if (!(tmp & PIPECONF_ENABLE))
 		return false;
 
+	intel_get_pipe_timings(crtc, pipe_config);
+
+	i9xx_get_pfit_config(crtc, pipe_config);
+
+	if (INTEL_INFO(dev)->gen >= 4) {
+		tmp = I915_READ(DPLL_MD(crtc->pipe));
+		pipe_config->pixel_multiplier =
+			((tmp & DPLL_MD_UDI_MULTIPLIER_MASK)
+			 >> DPLL_MD_UDI_MULTIPLIER_SHIFT) + 1;
+	} else if (IS_I945G(dev) || IS_I945GM(dev) || IS_G33(dev)) {
+		tmp = I915_READ(DPLL(crtc->pipe));
+		pipe_config->pixel_multiplier =
+			((tmp & SDVO_MULTIPLIER_MASK)
+			 >> SDVO_MULTIPLIER_SHIFT_HIRES) + 1;
+	} else {
+		/* Note that on i915G/GM the pixel multiplier is in the sdvo
+		 * port and will be fixed up in the encoder->get_config
+		 * function. */
+		pipe_config->pixel_multiplier = 1;
+	}
+
 	return true;
 }
 
@@ -4779,7 +4978,6 @@ static void ironlake_init_pch_refclk(struct drm_device *dev)
 	u32 val, final;
 	bool has_lvds = false;
 	bool has_cpu_edp = false;
-	bool has_pch_edp = false;
 	bool has_panel = false;
 	bool has_ck505 = false;
 	bool can_ssc = false;
@@ -4794,25 +4992,22 @@ static void ironlake_init_pch_refclk(struct drm_device *dev)
 			break;
 		case INTEL_OUTPUT_EDP:
 			has_panel = true;
-			if (intel_encoder_is_pch_edp(&encoder->base))
-				has_pch_edp = true;
-			else
+			if (enc_to_dig_port(&encoder->base)->port == PORT_A)
 				has_cpu_edp = true;
 			break;
 		}
 	}
 
 	if (HAS_PCH_IBX(dev)) {
-		has_ck505 = dev_priv->display_clock_mode;
+		has_ck505 = dev_priv->vbt.display_clock_mode;
 		can_ssc = has_ck505;
 	} else {
 		has_ck505 = false;
 		can_ssc = true;
 	}
 
-	DRM_DEBUG_KMS("has_panel %d has_lvds %d has_pch_edp %d has_cpu_edp %d has_ck505 %d\n",
-		      has_panel, has_lvds, has_pch_edp, has_cpu_edp,
-		      has_ck505);
+	DRM_DEBUG_KMS("has_panel %d has_lvds %d has_ck505 %d\n",
+		      has_panel, has_lvds, has_ck505);
 
 	/* Ironlake: try to setup display ref clock before DPLL
 	 * enabling. This is only under driver's control after
@@ -5102,7 +5297,6 @@ static int ironlake_get_refclk(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_encoder *encoder;
-	struct intel_encoder *edp_encoder = NULL;
 	int num_connectors = 0;
 	bool is_lvds = false;
 
@@ -5111,34 +5305,28 @@ static int ironlake_get_refclk(struct drm_crtc *crtc)
 		case INTEL_OUTPUT_LVDS:
 			is_lvds = true;
 			break;
-		case INTEL_OUTPUT_EDP:
-			edp_encoder = encoder;
-			break;
 		}
 		num_connectors++;
 	}
 
 	if (is_lvds && intel_panel_use_ssc(dev_priv) && num_connectors < 2) {
 		DRM_DEBUG_KMS("using SSC reference clock of %d MHz\n",
-			      dev_priv->lvds_ssc_freq);
-		return dev_priv->lvds_ssc_freq * 1000;
+			      dev_priv->vbt.lvds_ssc_freq);
+		return dev_priv->vbt.lvds_ssc_freq * 1000;
 	}
 
 	return 120000;
 }
 
-static void ironlake_set_pipeconf(struct drm_crtc *crtc,
-				  struct drm_display_mode *adjusted_mode,
-				  bool dither)
+static void ironlake_set_pipeconf(struct drm_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
 	uint32_t val;
 
-	val = I915_READ(PIPECONF(pipe));
+	val = 0;
 
-	val &= ~PIPECONF_BPC_MASK;
 	switch (intel_crtc->config.pipe_bpp) {
 	case 18:
 		val |= PIPECONF_6BPC;
@@ -5157,20 +5345,16 @@ static void ironlake_set_pipeconf(struct drm_crtc *crtc,
 		BUG();
 	}
 
-	val &= ~(PIPECONF_DITHER_EN | PIPECONF_DITHER_TYPE_MASK);
-	if (dither)
+	if (intel_crtc->config.dither)
 		val |= (PIPECONF_DITHER_EN | PIPECONF_DITHER_TYPE_SP);
 
-	val &= ~PIPECONF_INTERLACE_MASK;
-	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE)
+	if (intel_crtc->config.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
 		val |= PIPECONF_INTERLACED_ILK;
 	else
 		val |= PIPECONF_PROGRESSIVE;
 
 	if (intel_crtc->config.limited_color_range)
 		val |= PIPECONF_COLOR_RANGE_SELECT;
-	else
-		val &= ~PIPECONF_COLOR_RANGE_SELECT;
 
 	I915_WRITE(PIPECONF(pipe), val);
 	POSTING_READ(PIPECONF(pipe));
@@ -5240,33 +5424,31 @@ static void intel_set_pipe_csc(struct drm_crtc *crtc)
 	}
 }
 
-static void haswell_set_pipeconf(struct drm_crtc *crtc,
-				 struct drm_display_mode *adjusted_mode,
-				 bool dither)
+static void haswell_set_pipeconf(struct drm_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum transcoder cpu_transcoder = intel_crtc->config.cpu_transcoder;
 	uint32_t val;
 
-	val = I915_READ(PIPECONF(cpu_transcoder));
+	val = 0;
 
-	val &= ~(PIPECONF_DITHER_EN | PIPECONF_DITHER_TYPE_MASK);
-	if (dither)
+	if (intel_crtc->config.dither)
 		val |= (PIPECONF_DITHER_EN | PIPECONF_DITHER_TYPE_SP);
 
-	val &= ~PIPECONF_INTERLACE_MASK_HSW;
-	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE)
+	if (intel_crtc->config.adjusted_mode.flags & DRM_MODE_FLAG_INTERLACE)
 		val |= PIPECONF_INTERLACED_ILK;
 	else
 		val |= PIPECONF_PROGRESSIVE;
 
 	I915_WRITE(PIPECONF(cpu_transcoder), val);
 	POSTING_READ(PIPECONF(cpu_transcoder));
+
+	I915_WRITE(GAMMA_MODE(intel_crtc->pipe), GAMMA_MODE_MODE_8BIT);
+	POSTING_READ(GAMMA_MODE(intel_crtc->pipe));
 }
 
 static bool ironlake_compute_clocks(struct drm_crtc *crtc,
-				    struct drm_display_mode *adjusted_mode,
 				    intel_clock_t *clock,
 				    bool *has_reduced_clock,
 				    intel_clock_t *reduced_clock)
@@ -5276,22 +5458,13 @@ static bool ironlake_compute_clocks(struct drm_crtc *crtc,
 	struct intel_encoder *intel_encoder;
 	int refclk;
 	const intel_limit_t *limit;
-	bool ret, is_sdvo = false, is_tv = false, is_lvds = false;
+	bool ret, is_lvds = false;
 
 	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
 		switch (intel_encoder->type) {
 		case INTEL_OUTPUT_LVDS:
 			is_lvds = true;
 			break;
-		case INTEL_OUTPUT_SDVO:
-		case INTEL_OUTPUT_HDMI:
-			is_sdvo = true;
-			if (intel_encoder->needs_tv_clock)
-				is_tv = true;
-			break;
-		case INTEL_OUTPUT_TVOUT:
-			is_tv = true;
-			break;
 		}
 	}
 
@@ -5303,8 +5476,9 @@ static bool ironlake_compute_clocks(struct drm_crtc *crtc,
 	 * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
 	 */
 	limit = intel_limit(crtc, refclk);
-	ret = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, NULL,
-			      clock);
+	ret = dev_priv->display.find_dpll(limit, crtc,
+					  to_intel_crtc(crtc)->config.port_clock,
+					  refclk, NULL, clock);
 	if (!ret)
 		return false;
 
@@ -5315,16 +5489,13 @@ static bool ironlake_compute_clocks(struct drm_crtc *crtc,
 		 * by using the FP0/FP1. In such case we will disable the LVDS
 		 * downclock feature.
 		*/
-		*has_reduced_clock = limit->find_pll(limit, crtc,
-						     dev_priv->lvds_downclock,
-						     refclk,
-						     clock,
-						     reduced_clock);
+		*has_reduced_clock =
+			dev_priv->display.find_dpll(limit, crtc,
+						    dev_priv->lvds_downclock,
+						    refclk, clock,
+						    reduced_clock);
 	}
 
-	if (is_sdvo && is_tv)
-		i9xx_adjust_sdvo_tv_clock(to_intel_crtc(crtc));
-
 	return true;
 }
 
@@ -5346,65 +5517,25 @@ static void cpt_enable_fdi_bc_bifurcation(struct drm_device *dev)
 	POSTING_READ(SOUTH_CHICKEN1);
 }
 
-static bool ironlake_check_fdi_lanes(struct intel_crtc *intel_crtc)
+static void ivybridge_update_fdi_bc_bifurcation(struct intel_crtc *intel_crtc)
 {
 	struct drm_device *dev = intel_crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *pipe_B_crtc =
-		to_intel_crtc(dev_priv->pipe_to_crtc_mapping[PIPE_B]);
-
-	DRM_DEBUG_KMS("checking fdi config on pipe %i, lanes %i\n",
-		      intel_crtc->pipe, intel_crtc->fdi_lanes);
-	if (intel_crtc->fdi_lanes > 4) {
-		DRM_DEBUG_KMS("invalid fdi lane config on pipe %i: %i lanes\n",
-			      intel_crtc->pipe, intel_crtc->fdi_lanes);
-		/* Clamp lanes to avoid programming the hw with bogus values. */
-		intel_crtc->fdi_lanes = 4;
-
-		return false;
-	}
-
-	if (INTEL_INFO(dev)->num_pipes == 2)
-		return true;
 
 	switch (intel_crtc->pipe) {
 	case PIPE_A:
-		return true;
+		break;
 	case PIPE_B:
-		if (dev_priv->pipe_to_crtc_mapping[PIPE_C]->enabled &&
-		    intel_crtc->fdi_lanes > 2) {
-			DRM_DEBUG_KMS("invalid shared fdi lane config on pipe %i: %i lanes\n",
-				      intel_crtc->pipe, intel_crtc->fdi_lanes);
-			/* Clamp lanes to avoid programming the hw with bogus values. */
-			intel_crtc->fdi_lanes = 2;
-
-			return false;
-		}
-
-		if (intel_crtc->fdi_lanes > 2)
+		if (intel_crtc->config.fdi_lanes > 2)
 			WARN_ON(I915_READ(SOUTH_CHICKEN1) & FDI_BC_BIFURCATION_SELECT);
 		else
 			cpt_enable_fdi_bc_bifurcation(dev);
 
-		return true;
+		break;
 	case PIPE_C:
-		if (!pipe_B_crtc->base.enabled || pipe_B_crtc->fdi_lanes <= 2) {
-			if (intel_crtc->fdi_lanes > 2) {
-				DRM_DEBUG_KMS("invalid shared fdi lane config on pipe %i: %i lanes\n",
-					      intel_crtc->pipe, intel_crtc->fdi_lanes);
-				/* Clamp lanes to avoid programming the hw with bogus values. */
-				intel_crtc->fdi_lanes = 2;
-
-				return false;
-			}
-		} else {
-			DRM_DEBUG_KMS("fdi link B uses too many lanes to enable link C\n");
-			return false;
-		}
-
 		cpt_enable_fdi_bc_bifurcation(dev);
 
-		return true;
+		break;
 	default:
 		BUG();
 	}
@@ -5421,78 +5552,13 @@ int ironlake_get_lanes_required(int target_clock, int link_bw, int bpp)
 	return bps / (link_bw * 8) + 1;
 }
 
-void intel_pch_transcoder_set_m_n(struct intel_crtc *crtc,
-				  struct intel_link_m_n *m_n)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe = crtc->pipe;
-
-	I915_WRITE(TRANSDATA_M1(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
-	I915_WRITE(TRANSDATA_N1(pipe), m_n->gmch_n);
-	I915_WRITE(TRANSDPLINK_M1(pipe), m_n->link_m);
-	I915_WRITE(TRANSDPLINK_N1(pipe), m_n->link_n);
-}
-
-void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
-				  struct intel_link_m_n *m_n)
-{
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe = crtc->pipe;
-	enum transcoder transcoder = crtc->config.cpu_transcoder;
-
-	if (INTEL_INFO(dev)->gen >= 5) {
-		I915_WRITE(PIPE_DATA_M1(transcoder), TU_SIZE(m_n->tu) | m_n->gmch_m);
-		I915_WRITE(PIPE_DATA_N1(transcoder), m_n->gmch_n);
-		I915_WRITE(PIPE_LINK_M1(transcoder), m_n->link_m);
-		I915_WRITE(PIPE_LINK_N1(transcoder), m_n->link_n);
-	} else {
-		I915_WRITE(PIPE_GMCH_DATA_M(pipe), TU_SIZE(m_n->tu) | m_n->gmch_m);
-		I915_WRITE(PIPE_GMCH_DATA_N(pipe), m_n->gmch_n);
-		I915_WRITE(PIPE_DP_LINK_M(pipe), m_n->link_m);
-		I915_WRITE(PIPE_DP_LINK_N(pipe), m_n->link_n);
-	}
-}
-
-static void ironlake_fdi_set_m_n(struct drm_crtc *crtc)
+static bool ironlake_needs_fb_cb_tune(struct dpll *dpll, int factor)
 {
-	struct drm_device *dev = crtc->dev;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_display_mode *adjusted_mode =
-		&intel_crtc->config.adjusted_mode;
-	struct intel_link_m_n m_n = {0};
-	int target_clock, lane, link_bw;
-
-	/* FDI is a binary signal running at ~2.7GHz, encoding
-	 * each output octet as 10 bits. The actual frequency
-	 * is stored as a divider into a 100MHz clock, and the
-	 * mode pixel clock is stored in units of 1KHz.
-	 * Hence the bw of each lane in terms of the mode signal
-	 * is:
-	 */
-	link_bw = intel_fdi_link_freq(dev) * MHz(100)/KHz(1)/10;
-
-	if (intel_crtc->config.pixel_target_clock)
-		target_clock = intel_crtc->config.pixel_target_clock;
-	else
-		target_clock = adjusted_mode->clock;
-
-	lane = ironlake_get_lanes_required(target_clock, link_bw,
-					   intel_crtc->config.pipe_bpp);
-
-	intel_crtc->fdi_lanes = lane;
-
-	if (intel_crtc->config.pixel_multiplier > 1)
-		link_bw *= intel_crtc->config.pixel_multiplier;
-	intel_link_compute_m_n(intel_crtc->config.pipe_bpp, lane, target_clock,
-			       link_bw, &m_n);
-
-	intel_cpu_transcoder_set_m_n(intel_crtc, &m_n);
+	return i9xx_dpll_compute_m(dpll) < factor * dpll->n;
 }
 
 static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
-				      intel_clock_t *clock, u32 *fp,
+				      u32 *fp,
 				      intel_clock_t *reduced_clock, u32 *fp2)
 {
 	struct drm_crtc *crtc = &intel_crtc->base;
@@ -5501,7 +5567,7 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 	struct intel_encoder *intel_encoder;
 	uint32_t dpll;
 	int factor, num_connectors = 0;
-	bool is_lvds = false, is_sdvo = false, is_tv = false;
+	bool is_lvds = false, is_sdvo = false;
 
 	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
 		switch (intel_encoder->type) {
@@ -5511,11 +5577,6 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 		case INTEL_OUTPUT_SDVO:
 		case INTEL_OUTPUT_HDMI:
 			is_sdvo = true;
-			if (intel_encoder->needs_tv_clock)
-				is_tv = true;
-			break;
-		case INTEL_OUTPUT_TVOUT:
-			is_tv = true;
 			break;
 		}
 
@@ -5526,13 +5587,13 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 	factor = 21;
 	if (is_lvds) {
 		if ((intel_panel_use_ssc(dev_priv) &&
-		     dev_priv->lvds_ssc_freq == 100) ||
+		     dev_priv->vbt.lvds_ssc_freq == 100) ||
 		    (HAS_PCH_IBX(dev) && intel_is_dual_link_lvds(dev)))
 			factor = 25;
-	} else if (is_sdvo && is_tv)
+	} else if (intel_crtc->config.sdvo_tv_clock)
 		factor = 20;
 
-	if (clock->m < factor * clock->n)
+	if (ironlake_needs_fb_cb_tune(&intel_crtc->config.dpll, factor))
 		*fp |= FP_CB_TUNE;
 
 	if (fp2 && (reduced_clock->m < factor * reduced_clock->n))
@@ -5544,23 +5605,21 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 		dpll |= DPLLB_MODE_LVDS;
 	else
 		dpll |= DPLLB_MODE_DAC_SERIAL;
-	if (is_sdvo) {
-		if (intel_crtc->config.pixel_multiplier > 1) {
-			dpll |= (intel_crtc->config.pixel_multiplier - 1)
-				<< PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT;
-		}
+
+	dpll |= (intel_crtc->config.pixel_multiplier - 1)
+		<< PLL_REF_SDVO_HDMI_MULTIPLIER_SHIFT;
+
+	if (is_sdvo)
 		dpll |= DPLL_DVO_HIGH_SPEED;
-	}
-	if (intel_crtc->config.has_dp_encoder &&
-	    intel_crtc->config.has_pch_encoder)
+	if (intel_crtc->config.has_dp_encoder)
 		dpll |= DPLL_DVO_HIGH_SPEED;
 
 	/* compute bitmask from p1 value */
-	dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
+	dpll |= (1 << (intel_crtc->config.dpll.p1 - 1)) << DPLL_FPA01_P1_POST_DIV_SHIFT;
 	/* also FPA1 */
-	dpll |= (1 << (clock->p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT;
+	dpll |= (1 << (intel_crtc->config.dpll.p1 - 1)) << DPLL_FPA1_P1_POST_DIV_SHIFT;
 
-	switch (clock->p2) {
+	switch (intel_crtc->config.dpll.p2) {
 	case 5:
 		dpll |= DPLL_DAC_SERIAL_P2_CLOCK_DIV_5;
 		break;
@@ -5575,18 +5634,12 @@ static uint32_t ironlake_compute_dpll(struct intel_crtc *intel_crtc,
 		break;
 	}
 
-	if (is_sdvo && is_tv)
-		dpll |= PLL_REF_INPUT_TVCLKINBC;
-	else if (is_tv)
-		/* XXX: just matching BIOS for now */
-		/*	dpll |= PLL_REF_INPUT_TVCLKINBC; */
-		dpll |= 3;
-	else if (is_lvds && intel_panel_use_ssc(dev_priv) && num_connectors < 2)
+	if (is_lvds && intel_panel_use_ssc(dev_priv) && num_connectors < 2)
 		dpll |= PLLB_REF_INPUT_SPREADSPECTRUMIN;
 	else
 		dpll |= PLL_REF_INPUT_DREFCLK;
 
-	return dpll;
+	return dpll | DPLL_VCO_ENABLE;
 }
 
 static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
@@ -5596,19 +5649,16 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_display_mode *adjusted_mode =
-		&intel_crtc->config.adjusted_mode;
-	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
 	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
 	int num_connectors = 0;
 	intel_clock_t clock, reduced_clock;
-	u32 dpll, fp = 0, fp2 = 0;
+	u32 dpll = 0, fp = 0, fp2 = 0;
 	bool ok, has_reduced_clock = false;
 	bool is_lvds = false;
 	struct intel_encoder *encoder;
+	struct intel_shared_dpll *pll;
 	int ret;
-	bool dither, fdi_config_ok;
 
 	for_each_encoder_on_crtc(dev, crtc, encoder) {
 		switch (encoder->type) {
@@ -5623,11 +5673,9 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	WARN(!(HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)),
 	     "Unexpected PCH type %d\n", INTEL_PCH_TYPE(dev));
 
-	intel_crtc->config.cpu_transcoder = pipe;
-
-	ok = ironlake_compute_clocks(crtc, adjusted_mode, &clock,
+	ok = ironlake_compute_clocks(crtc, &clock,
 				     &has_reduced_clock, &reduced_clock);
-	if (!ok) {
+	if (!ok && !intel_crtc->config.clock_set) {
 		DRM_ERROR("Couldn't find PLL settings for mode!\n");
 		return -EINVAL;
 	}
@@ -5643,34 +5691,31 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	/* Ensure that the cursor is valid for the new mode before changing... */
 	intel_crtc_update_cursor(crtc, true);
 
-	/* determine panel color depth */
-	dither = intel_crtc->config.dither;
-	if (is_lvds && dev_priv->lvds_dither)
-		dither = true;
-
-	fp = clock.n << 16 | clock.m1 << 8 | clock.m2;
-	if (has_reduced_clock)
-		fp2 = reduced_clock.n << 16 | reduced_clock.m1 << 8 |
-			reduced_clock.m2;
-
-	dpll = ironlake_compute_dpll(intel_crtc, &clock, &fp, &reduced_clock,
-				     has_reduced_clock ? &fp2 : NULL);
-
-	DRM_DEBUG_KMS("Mode for pipe %d:\n", pipe);
-	drm_mode_debug_printmodeline(mode);
-
 	/* CPU eDP is the only output that doesn't need a PCH PLL of its own. */
 	if (intel_crtc->config.has_pch_encoder) {
-		struct intel_pch_pll *pll;
+		fp = i9xx_dpll_compute_fp(&intel_crtc->config.dpll);
+		if (has_reduced_clock)
+			fp2 = i9xx_dpll_compute_fp(&reduced_clock);
+
+		dpll = ironlake_compute_dpll(intel_crtc,
+					     &fp, &reduced_clock,
+					     has_reduced_clock ? &fp2 : NULL);
+
+		intel_crtc->config.dpll_hw_state.dpll = dpll;
+		intel_crtc->config.dpll_hw_state.fp0 = fp;
+		if (has_reduced_clock)
+			intel_crtc->config.dpll_hw_state.fp1 = fp2;
+		else
+			intel_crtc->config.dpll_hw_state.fp1 = fp;
 
-		pll = intel_get_pch_pll(intel_crtc, dpll, fp);
+		pll = intel_get_shared_dpll(intel_crtc, dpll, fp);
 		if (pll == NULL) {
-			DRM_DEBUG_DRIVER("failed to find PLL for pipe %d\n",
-					 pipe);
+			DRM_DEBUG_DRIVER("failed to find PLL for pipe %c\n",
+					 pipe_name(pipe));
 			return -EINVAL;
 		}
 	} else
-		intel_put_pch_pll(intel_crtc);
+		intel_put_shared_dpll(intel_crtc);
 
 	if (intel_crtc->config.has_dp_encoder)
 		intel_dp_set_m_n(intel_crtc);
@@ -5679,11 +5724,18 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 		if (encoder->pre_pll_enable)
 			encoder->pre_pll_enable(encoder);
 
-	if (intel_crtc->pch_pll) {
-		I915_WRITE(intel_crtc->pch_pll->pll_reg, dpll);
+	if (is_lvds && has_reduced_clock && i915_powersave)
+		intel_crtc->lowfreq_avail = true;
+	else
+		intel_crtc->lowfreq_avail = false;
+
+	if (intel_crtc->config.has_pch_encoder) {
+		pll = intel_crtc_to_shared_dpll(intel_crtc);
+
+		I915_WRITE(PCH_DPLL(pll->id), dpll);
 
 		/* Wait for the clocks to stabilize. */
-		POSTING_READ(intel_crtc->pch_pll->pll_reg);
+		POSTING_READ(PCH_DPLL(pll->id));
 		udelay(150);
 
 		/* The pixel multiplier can only be updated once the
@@ -5691,32 +5743,25 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 		 *
 		 * So write it again.
 		 */
-		I915_WRITE(intel_crtc->pch_pll->pll_reg, dpll);
-	}
+		I915_WRITE(PCH_DPLL(pll->id), dpll);
 
-	intel_crtc->lowfreq_avail = false;
-	if (intel_crtc->pch_pll) {
-		if (is_lvds && has_reduced_clock && i915_powersave) {
-			I915_WRITE(intel_crtc->pch_pll->fp1_reg, fp2);
-			intel_crtc->lowfreq_avail = true;
-		} else {
-			I915_WRITE(intel_crtc->pch_pll->fp1_reg, fp);
-		}
+		if (has_reduced_clock)
+			I915_WRITE(PCH_FP1(pll->id), fp2);
+		else
+			I915_WRITE(PCH_FP1(pll->id), fp);
 	}
 
-	intel_set_pipe_timings(intel_crtc, mode, adjusted_mode);
+	intel_set_pipe_timings(intel_crtc);
 
-	/* Note, this also computes intel_crtc->fdi_lanes which is used below in
-	 * ironlake_check_fdi_lanes. */
-	intel_crtc->fdi_lanes = 0;
-	if (intel_crtc->config.has_pch_encoder)
-		ironlake_fdi_set_m_n(crtc);
+	if (intel_crtc->config.has_pch_encoder) {
+		intel_cpu_transcoder_set_m_n(intel_crtc,
+					     &intel_crtc->config.fdi_m_n);
+	}
 
-	fdi_config_ok = ironlake_check_fdi_lanes(intel_crtc);
+	if (IS_IVYBRIDGE(dev))
+		ivybridge_update_fdi_bc_bifurcation(intel_crtc);
 
-	ironlake_set_pipeconf(crtc, adjusted_mode, dither);
-
-	intel_wait_for_vblank(dev, pipe);
+	ironlake_set_pipeconf(crtc);
 
 	/* Set up the display plane register */
 	I915_WRITE(DSPCNTR(plane), DISPPLANE_GAMMA_ENABLE);
@@ -5726,9 +5771,46 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 
 	intel_update_watermarks(dev);
 
-	intel_update_linetime_watermarks(dev, pipe, adjusted_mode);
+	return ret;
+}
+
+static void ironlake_get_fdi_m_n_config(struct intel_crtc *crtc,
+					struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum transcoder transcoder = pipe_config->cpu_transcoder;
+
+	pipe_config->fdi_m_n.link_m = I915_READ(PIPE_LINK_M1(transcoder));
+	pipe_config->fdi_m_n.link_n = I915_READ(PIPE_LINK_N1(transcoder));
+	pipe_config->fdi_m_n.gmch_m = I915_READ(PIPE_DATA_M1(transcoder))
+					& ~TU_SIZE_MASK;
+	pipe_config->fdi_m_n.gmch_n = I915_READ(PIPE_DATA_N1(transcoder));
+	pipe_config->fdi_m_n.tu = ((I915_READ(PIPE_DATA_M1(transcoder))
+				   & TU_SIZE_MASK) >> TU_SIZE_SHIFT) + 1;
+}
+
+static void ironlake_get_pfit_config(struct intel_crtc *crtc,
+				     struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	uint32_t tmp;
+
+	tmp = I915_READ(PF_CTL(crtc->pipe));
 
-	return fdi_config_ok ? ret : -EINVAL;
+	if (tmp & PF_ENABLE) {
+		pipe_config->pch_pfit.pos = I915_READ(PF_WIN_POS(crtc->pipe));
+		pipe_config->pch_pfit.size = I915_READ(PF_WIN_SZ(crtc->pipe));
+
+		/* We currently do not free assignements of panel fitters on
+		 * ivb/hsw (since we don't use the higher upscaling modes which
+		 * differentiates them) so just WARN about this case for now. */
+		if (IS_GEN7(dev)) {
+			WARN_ON((tmp & PF_PIPE_SEL_MASK_IVB) !=
+				PF_PIPE_SEL_IVB(crtc->pipe));
+		}
+	}
 }
 
 static bool ironlake_get_pipe_config(struct intel_crtc *crtc,
@@ -5738,42 +5820,67 @@ static bool ironlake_get_pipe_config(struct intel_crtc *crtc,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	uint32_t tmp;
 
+	pipe_config->cpu_transcoder = crtc->pipe;
+	pipe_config->shared_dpll = DPLL_ID_PRIVATE;
+
 	tmp = I915_READ(PIPECONF(crtc->pipe));
 	if (!(tmp & PIPECONF_ENABLE))
 		return false;
 
-	if (I915_READ(TRANSCONF(crtc->pipe)) & TRANS_ENABLE)
+	if (I915_READ(PCH_TRANSCONF(crtc->pipe)) & TRANS_ENABLE) {
+		struct intel_shared_dpll *pll;
+
 		pipe_config->has_pch_encoder = true;
 
+		tmp = I915_READ(FDI_RX_CTL(crtc->pipe));
+		pipe_config->fdi_lanes = ((FDI_DP_PORT_WIDTH_MASK & tmp) >>
+					  FDI_DP_PORT_WIDTH_SHIFT) + 1;
+
+		ironlake_get_fdi_m_n_config(crtc, pipe_config);
+
+		/* XXX: Can't properly read out the pch dpll pixel multiplier
+		 * since we don't have state tracking for pch clocks yet. */
+		pipe_config->pixel_multiplier = 1;
+
+		if (HAS_PCH_IBX(dev_priv->dev)) {
+			pipe_config->shared_dpll = crtc->pipe;
+		} else {
+			tmp = I915_READ(PCH_DPLL_SEL);
+			if (tmp & TRANS_DPLLB_SEL(crtc->pipe))
+				pipe_config->shared_dpll = DPLL_ID_PCH_PLL_B;
+			else
+				pipe_config->shared_dpll = DPLL_ID_PCH_PLL_A;
+		}
+
+		pll = &dev_priv->shared_dplls[pipe_config->shared_dpll];
+
+		WARN_ON(!pll->get_hw_state(dev_priv, pll,
+					   &pipe_config->dpll_hw_state));
+	} else {
+		pipe_config->pixel_multiplier = 1;
+	}
+
+	intel_get_pipe_timings(crtc, pipe_config);
+
+	ironlake_get_pfit_config(crtc, pipe_config);
+
 	return true;
 }
 
 static void haswell_modeset_global_resources(struct drm_device *dev)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	bool enable = false;
 	struct intel_crtc *crtc;
-	struct intel_encoder *encoder;
 
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, base.head) {
-		if (crtc->pipe != PIPE_A && crtc->base.enabled)
-			enable = true;
-		/* XXX: Should check for edp transcoder here, but thanks to init
-		 * sequence that's not yet available. Just in case desktop eDP
-		 * on PORT D is possible on haswell, too. */
-	}
+		if (!crtc->base.enabled)
+			continue;
 
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
-			    base.head) {
-		if (encoder->type != INTEL_OUTPUT_EDP &&
-		    encoder->connectors_active)
+		if (crtc->pipe != PIPE_A || crtc->config.pch_pfit.size ||
+		    crtc->config.cpu_transcoder != TRANSCODER_EDP)
 			enable = true;
 	}
 
-	/* Even the eDP panel fitter is outside the always-on well. */
-	if (dev_priv->pch_pf_size)
-		enable = true;
-
 	intel_set_power_well(dev, enable);
 }
 
@@ -5784,68 +5891,28 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_display_mode *adjusted_mode =
-		&intel_crtc->config.adjusted_mode;
-	struct drm_display_mode *mode = &intel_crtc->config.requested_mode;
-	int pipe = intel_crtc->pipe;
 	int plane = intel_crtc->plane;
-	int num_connectors = 0;
-	bool is_cpu_edp = false;
-	struct intel_encoder *encoder;
 	int ret;
-	bool dither;
-
-	for_each_encoder_on_crtc(dev, crtc, encoder) {
-		switch (encoder->type) {
-		case INTEL_OUTPUT_EDP:
-			if (!intel_encoder_is_pch_edp(&encoder->base))
-				is_cpu_edp = true;
-			break;
-		}
-
-		num_connectors++;
-	}
-
-	if (is_cpu_edp)
-		intel_crtc->config.cpu_transcoder = TRANSCODER_EDP;
-	else
-		intel_crtc->config.cpu_transcoder = pipe;
-
-	/* We are not sure yet this won't happen. */
-	WARN(!HAS_PCH_LPT(dev), "Unexpected PCH type %d\n",
-	     INTEL_PCH_TYPE(dev));
-
-	WARN(num_connectors != 1, "%d connectors attached to pipe %c\n",
-	     num_connectors, pipe_name(pipe));
-
-	WARN_ON(I915_READ(PIPECONF(intel_crtc->config.cpu_transcoder)) &
-		(PIPECONF_ENABLE | I965_PIPECONF_ACTIVE));
-
-	WARN_ON(I915_READ(DSPCNTR(plane)) & DISPLAY_PLANE_ENABLE);
 
-	if (!intel_ddi_pll_mode_set(crtc, adjusted_mode->clock))
+	if (!intel_ddi_pll_mode_set(crtc))
 		return -EINVAL;
 
 	/* Ensure that the cursor is valid for the new mode before changing... */
 	intel_crtc_update_cursor(crtc, true);
 
-	/* determine panel color depth */
-	dither = intel_crtc->config.dither;
-
-	DRM_DEBUG_KMS("Mode for pipe %d:\n", pipe);
-	drm_mode_debug_printmodeline(mode);
-
 	if (intel_crtc->config.has_dp_encoder)
 		intel_dp_set_m_n(intel_crtc);
 
 	intel_crtc->lowfreq_avail = false;
 
-	intel_set_pipe_timings(intel_crtc, mode, adjusted_mode);
+	intel_set_pipe_timings(intel_crtc);
 
-	if (intel_crtc->config.has_pch_encoder)
-		ironlake_fdi_set_m_n(crtc);
+	if (intel_crtc->config.has_pch_encoder) {
+		intel_cpu_transcoder_set_m_n(intel_crtc,
+					     &intel_crtc->config.fdi_m_n);
+	}
 
-	haswell_set_pipeconf(crtc, adjusted_mode, dither);
+	haswell_set_pipeconf(crtc);
 
 	intel_set_pipe_csc(crtc);
 
@@ -5857,8 +5924,6 @@ static int haswell_crtc_mode_set(struct drm_crtc *crtc,
 
 	intel_update_watermarks(dev);
 
-	intel_update_linetime_watermarks(dev, pipe, adjusted_mode);
-
 	return ret;
 }
 
@@ -5867,22 +5932,69 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc,
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum intel_display_power_domain pfit_domain;
 	uint32_t tmp;
 
-	tmp = I915_READ(PIPECONF(crtc->config.cpu_transcoder));
+	pipe_config->cpu_transcoder = crtc->pipe;
+	pipe_config->shared_dpll = DPLL_ID_PRIVATE;
+
+	tmp = I915_READ(TRANS_DDI_FUNC_CTL(TRANSCODER_EDP));
+	if (tmp & TRANS_DDI_FUNC_ENABLE) {
+		enum pipe trans_edp_pipe;
+		switch (tmp & TRANS_DDI_EDP_INPUT_MASK) {
+		default:
+			WARN(1, "unknown pipe linked to edp transcoder\n");
+		case TRANS_DDI_EDP_INPUT_A_ONOFF:
+		case TRANS_DDI_EDP_INPUT_A_ON:
+			trans_edp_pipe = PIPE_A;
+			break;
+		case TRANS_DDI_EDP_INPUT_B_ONOFF:
+			trans_edp_pipe = PIPE_B;
+			break;
+		case TRANS_DDI_EDP_INPUT_C_ONOFF:
+			trans_edp_pipe = PIPE_C;
+			break;
+		}
+
+		if (trans_edp_pipe == crtc->pipe)
+			pipe_config->cpu_transcoder = TRANSCODER_EDP;
+	}
+
+	if (!intel_display_power_enabled(dev,
+			POWER_DOMAIN_TRANSCODER(pipe_config->cpu_transcoder)))
+		return false;
+
+	tmp = I915_READ(PIPECONF(pipe_config->cpu_transcoder));
 	if (!(tmp & PIPECONF_ENABLE))
 		return false;
 
 	/*
-	 * aswell has only FDI/PCH transcoder A. It is which is connected to
+	 * Haswell has only FDI/PCH transcoder A. It is which is connected to
 	 * DDI E. So just check whether this pipe is wired to DDI E and whether
 	 * the PCH transcoder is on.
 	 */
-	tmp = I915_READ(TRANS_DDI_FUNC_CTL(crtc->pipe));
+	tmp = I915_READ(TRANS_DDI_FUNC_CTL(pipe_config->cpu_transcoder));
 	if ((tmp & TRANS_DDI_PORT_MASK) == TRANS_DDI_SELECT_PORT(PORT_E) &&
-	    I915_READ(TRANSCONF(PIPE_A)) & TRANS_ENABLE)
+	    I915_READ(LPT_TRANSCONF) & TRANS_ENABLE) {
 		pipe_config->has_pch_encoder = true;
 
+		tmp = I915_READ(FDI_RX_CTL(PIPE_A));
+		pipe_config->fdi_lanes = ((FDI_DP_PORT_WIDTH_MASK & tmp) >>
+					  FDI_DP_PORT_WIDTH_SHIFT) + 1;
+
+		ironlake_get_fdi_m_n_config(crtc, pipe_config);
+	}
+
+	intel_get_pipe_timings(crtc, pipe_config);
+
+	pfit_domain = POWER_DOMAIN_PIPE_PANEL_FITTER(crtc->pipe);
+	if (intel_display_power_enabled(dev, pfit_domain))
+		ironlake_get_pfit_config(crtc, pipe_config);
+
+	pipe_config->ips_enabled = hsw_crtc_supports_ips(crtc) &&
+				   (I915_READ(IPS_CTL) & IPS_ENABLE);
+
+	pipe_config->pixel_multiplier = 1;
 
 	return true;
 }
@@ -6120,7 +6232,7 @@ static void ironlake_write_eld(struct drm_connector *connector,
 		eldv |= IBX_ELD_VALIDB << 4;
 		eldv |= IBX_ELD_VALIDB << 8;
 	} else {
-		DRM_DEBUG_DRIVER("ELD on port %c\n", 'A' + i);
+		DRM_DEBUG_DRIVER("ELD on port %c\n", port_name(i));
 		eldv = IBX_ELD_VALIDB << ((i - 1) * 4);
 	}
 
@@ -6188,16 +6300,31 @@ void intel_crtc_load_lut(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int palreg = PALETTE(intel_crtc->pipe);
+	enum pipe pipe = intel_crtc->pipe;
+	int palreg = PALETTE(pipe);
 	int i;
+	bool reenable_ips = false;
 
 	/* The clocks have to be on to load the palette. */
 	if (!crtc->enabled || !intel_crtc->active)
 		return;
 
+	if (!HAS_PCH_SPLIT(dev_priv->dev))
+		assert_pll_enabled(dev_priv, pipe);
+
 	/* use legacy palette for Ironlake */
 	if (HAS_PCH_SPLIT(dev))
-		palreg = LGC_PALETTE(intel_crtc->pipe);
+		palreg = LGC_PALETTE(pipe);
+
+	/* Workaround : Do not read or write the pipe palette/gamma data while
+	 * GAMMA_MODE is configured for split gamma and IPS_CTL has IPS enabled.
+	 */
+	if (intel_crtc->config.ips_enabled &&
+	    ((I915_READ(GAMMA_MODE(pipe)) & GAMMA_MODE_MODE_MASK) ==
+	     GAMMA_MODE_MODE_SPLIT)) {
+		hsw_disable_ips(intel_crtc);
+		reenable_ips = true;
+	}
 
 	for (i = 0; i < 256; i++) {
 		I915_WRITE(palreg + 4 * i,
@@ -6205,6 +6332,9 @@ void intel_crtc_load_lut(struct drm_crtc *crtc)
 			   (intel_crtc->lut_g[i] << 8) |
 			   intel_crtc->lut_b[i]);
 	}
+
+	if (reenable_ips)
+		hsw_enable_ips(intel_crtc);
 }
 
 static void i845_update_cursor(struct drm_crtc *crtc, u32 base)
@@ -6451,7 +6581,7 @@ static int intel_crtc_cursor_set(struct drm_crtc *crtc,
 	intel_crtc->cursor_width = width;
 	intel_crtc->cursor_height = height;
 
-	intel_crtc_update_cursor(crtc, true);
+	intel_crtc_update_cursor(crtc, intel_crtc->cursor_bo != NULL);
 
 	return 0;
 fail_unpin:
@@ -6470,7 +6600,7 @@ static int intel_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 	intel_crtc->cursor_x = x;
 	intel_crtc->cursor_y = y;
 
-	intel_crtc_update_cursor(crtc, true);
+	intel_crtc_update_cursor(crtc, intel_crtc->cursor_bo != NULL);
 
 	return 0;
 }
@@ -6791,8 +6921,10 @@ static int intel_crtc_clock_get(struct drm_device *dev, struct drm_crtc *crtc)
 			return 0;
 		}
 
-		/* XXX: Handle the 100Mhz refclk */
-		intel_clock(dev, 96000, &clock);
+		if (IS_PINEVIEW(dev))
+			pineview_clock(96000, &clock);
+		else
+			i9xx_clock(96000, &clock);
 	} else {
 		bool is_lvds = (pipe == 1) && (I915_READ(LVDS) & LVDS_PORT_EN);
 
@@ -6804,9 +6936,9 @@ static int intel_crtc_clock_get(struct drm_device *dev, struct drm_crtc *crtc)
 			if ((dpll & PLL_REF_INPUT_MASK) ==
 			    PLLB_REF_INPUT_SPREADSPECTRUMIN) {
 				/* XXX: might not be 66MHz */
-				intel_clock(dev, 66000, &clock);
+				i9xx_clock(66000, &clock);
 			} else
-				intel_clock(dev, 48000, &clock);
+				i9xx_clock(48000, &clock);
 		} else {
 			if (dpll & PLL_P1_DIVIDE_BY_TWO)
 				clock.p1 = 2;
@@ -6819,7 +6951,7 @@ static int intel_crtc_clock_get(struct drm_device *dev, struct drm_crtc *crtc)
 			else
 				clock.p2 = 2;
 
-			intel_clock(dev, 48000, &clock);
+			i9xx_clock(48000, &clock);
 		}
 	}
 
@@ -6950,7 +7082,8 @@ void intel_mark_idle(struct drm_device *dev)
 	}
 }
 
-void intel_mark_fb_busy(struct drm_i915_gem_object *obj)
+void intel_mark_fb_busy(struct drm_i915_gem_object *obj,
+			struct intel_ring_buffer *ring)
 {
 	struct drm_device *dev = obj->base.dev;
 	struct drm_crtc *crtc;
@@ -6962,8 +7095,12 @@ void intel_mark_fb_busy(struct drm_i915_gem_object *obj)
 		if (!crtc->fb)
 			continue;
 
-		if (to_intel_framebuffer(crtc->fb)->obj == obj)
-			intel_increase_pllclock(crtc);
+		if (to_intel_framebuffer(crtc->fb)->obj != obj)
+			continue;
+
+		intel_increase_pllclock(crtc);
+		if (ring && intel_fbc_enabled(dev))
+			ring->fbc_dirty = true;
 	}
 }
 
@@ -6984,6 +7121,8 @@ static void intel_crtc_destroy(struct drm_crtc *crtc)
 		kfree(work);
 	}
 
+	intel_crtc_cursor_set(crtc, NULL, 0, 0, 0);
+
 	drm_crtc_cleanup(crtc);
 
 	kfree(intel_crtc);
@@ -7411,7 +7550,7 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 		goto cleanup_pending;
 
 	intel_disable_fbc(dev);
-	intel_mark_fb_busy(obj);
+	intel_mark_fb_busy(obj, NULL);
 	mutex_unlock(&dev->struct_mutex);
 
 	trace_i915_flip_request(intel_crtc->plane, obj);
@@ -7442,28 +7581,6 @@ static struct drm_crtc_helper_funcs intel_helper_funcs = {
 	.load_lut = intel_crtc_load_lut,
 };
 
-bool intel_encoder_check_is_cloned(struct intel_encoder *encoder)
-{
-	struct intel_encoder *other_encoder;
-	struct drm_crtc *crtc = &encoder->new_crtc->base;
-
-	if (WARN_ON(!crtc))
-		return false;
-
-	list_for_each_entry(other_encoder,
-			    &crtc->dev->mode_config.encoder_list,
-			    base.head) {
-
-		if (&other_encoder->new_crtc->base != crtc ||
-		    encoder == other_encoder)
-			continue;
-		else
-			return true;
-	}
-
-	return false;
-}
-
 static bool intel_encoder_crtc_ok(struct drm_encoder *encoder,
 				  struct drm_crtc *crtc)
 {
@@ -7531,13 +7648,39 @@ static void intel_modeset_commit_output_state(struct drm_device *dev)
 	}
 }
 
+static void
+connected_sink_compute_bpp(struct intel_connector * connector,
+			   struct intel_crtc_config *pipe_config)
+{
+	int bpp = pipe_config->pipe_bpp;
+
+	DRM_DEBUG_KMS("[CONNECTOR:%d:%s] checking for sink bpp constrains\n",
+		connector->base.base.id,
+		drm_get_connector_name(&connector->base));
+
+	/* Don't use an invalid EDID bpc value */
+	if (connector->base.display_info.bpc &&
+	    connector->base.display_info.bpc * 3 < bpp) {
+		DRM_DEBUG_KMS("clamping display bpp (was %d) to EDID reported max of %d\n",
+			      bpp, connector->base.display_info.bpc*3);
+		pipe_config->pipe_bpp = connector->base.display_info.bpc*3;
+	}
+
+	/* Clamp bpp to 8 on screens without EDID 1.4 */
+	if (connector->base.display_info.bpc == 0 && bpp > 24) {
+		DRM_DEBUG_KMS("clamping display bpp (was %d) to default limit of 24\n",
+			      bpp);
+		pipe_config->pipe_bpp = 24;
+	}
+}
+
 static int
-pipe_config_set_bpp(struct drm_crtc *crtc,
-		    struct drm_framebuffer *fb,
-		    struct intel_crtc_config *pipe_config)
+compute_baseline_pipe_bpp(struct intel_crtc *crtc,
+			  struct drm_framebuffer *fb,
+			  struct intel_crtc_config *pipe_config)
 {
-	struct drm_device *dev = crtc->dev;
-	struct drm_connector *connector;
+	struct drm_device *dev = crtc->base.dev;
+	struct intel_connector *connector;
 	int bpp;
 
 	switch (fb->pixel_format) {
@@ -7580,22 +7723,66 @@ pipe_config_set_bpp(struct drm_crtc *crtc,
 
 	/* Clamp display bpp to EDID value */
 	list_for_each_entry(connector, &dev->mode_config.connector_list,
-			    head) {
-		if (connector->encoder && connector->encoder->crtc != crtc)
+			    base.head) {
+		if (!connector->new_encoder ||
+		    connector->new_encoder->new_crtc != crtc)
 			continue;
 
-		/* Don't use an invalid EDID bpc value */
-		if (connector->display_info.bpc &&
-		    connector->display_info.bpc * 3 < bpp) {
-			DRM_DEBUG_KMS("clamping display bpp (was %d) to EDID reported max of %d\n",
-				      bpp, connector->display_info.bpc*3);
-			pipe_config->pipe_bpp = connector->display_info.bpc*3;
-		}
+		connected_sink_compute_bpp(connector, pipe_config);
 	}
 
 	return bpp;
 }
 
+static void intel_dump_pipe_config(struct intel_crtc *crtc,
+				   struct intel_crtc_config *pipe_config,
+				   const char *context)
+{
+	DRM_DEBUG_KMS("[CRTC:%d]%s config for pipe %c\n", crtc->base.base.id,
+		      context, pipe_name(crtc->pipe));
+
+	DRM_DEBUG_KMS("cpu_transcoder: %c\n", transcoder_name(pipe_config->cpu_transcoder));
+	DRM_DEBUG_KMS("pipe bpp: %i, dithering: %i\n",
+		      pipe_config->pipe_bpp, pipe_config->dither);
+	DRM_DEBUG_KMS("fdi/pch: %i, lanes: %i, gmch_m: %u, gmch_n: %u, link_m: %u, link_n: %u, tu: %u\n",
+		      pipe_config->has_pch_encoder,
+		      pipe_config->fdi_lanes,
+		      pipe_config->fdi_m_n.gmch_m, pipe_config->fdi_m_n.gmch_n,
+		      pipe_config->fdi_m_n.link_m, pipe_config->fdi_m_n.link_n,
+		      pipe_config->fdi_m_n.tu);
+	DRM_DEBUG_KMS("requested mode:\n");
+	drm_mode_debug_printmodeline(&pipe_config->requested_mode);
+	DRM_DEBUG_KMS("adjusted mode:\n");
+	drm_mode_debug_printmodeline(&pipe_config->adjusted_mode);
+	DRM_DEBUG_KMS("gmch pfit: control: 0x%08x, ratios: 0x%08x, lvds border: 0x%08x\n",
+		      pipe_config->gmch_pfit.control,
+		      pipe_config->gmch_pfit.pgm_ratios,
+		      pipe_config->gmch_pfit.lvds_border_bits);
+	DRM_DEBUG_KMS("pch pfit: pos: 0x%08x, size: 0x%08x\n",
+		      pipe_config->pch_pfit.pos,
+		      pipe_config->pch_pfit.size);
+	DRM_DEBUG_KMS("ips: %i\n", pipe_config->ips_enabled);
+}
+
+static bool check_encoder_cloning(struct drm_crtc *crtc)
+{
+	int num_encoders = 0;
+	bool uncloneable_encoders = false;
+	struct intel_encoder *encoder;
+
+	list_for_each_entry(encoder, &crtc->dev->mode_config.encoder_list,
+			    base.head) {
+		if (&encoder->new_crtc->base != crtc)
+			continue;
+
+		num_encoders++;
+		if (!encoder->cloneable)
+			uncloneable_encoders = true;
+	}
+
+	return !(num_encoders > 1 && uncloneable_encoders);
+}
+
 static struct intel_crtc_config *
 intel_modeset_pipe_config(struct drm_crtc *crtc,
 			  struct drm_framebuffer *fb,
@@ -7605,7 +7792,13 @@ intel_modeset_pipe_config(struct drm_crtc *crtc,
 	struct drm_encoder_helper_funcs *encoder_funcs;
 	struct intel_encoder *encoder;
 	struct intel_crtc_config *pipe_config;
-	int plane_bpp;
+	int plane_bpp, ret = -EINVAL;
+	bool retry = true;
+
+	if (!check_encoder_cloning(crtc)) {
+		DRM_DEBUG_KMS("rejecting invalid cloning configuration\n");
+		return ERR_PTR(-EINVAL);
+	}
 
 	pipe_config = kzalloc(sizeof(*pipe_config), GFP_KERNEL);
 	if (!pipe_config)
@@ -7613,11 +7806,23 @@ intel_modeset_pipe_config(struct drm_crtc *crtc,
 
 	drm_mode_copy(&pipe_config->adjusted_mode, mode);
 	drm_mode_copy(&pipe_config->requested_mode, mode);
-
-	plane_bpp = pipe_config_set_bpp(crtc, fb, pipe_config);
+	pipe_config->cpu_transcoder = to_intel_crtc(crtc)->pipe;
+	pipe_config->shared_dpll = DPLL_ID_PRIVATE;
+
+	/* Compute a starting value for pipe_config->pipe_bpp taking the source
+	 * plane pixel format and any sink constraints into account. Returns the
+	 * source plane bpp so that dithering can be selected on mismatches
+	 * after encoders and crtc also have had their say. */
+	plane_bpp = compute_baseline_pipe_bpp(to_intel_crtc(crtc),
+					      fb, pipe_config);
 	if (plane_bpp < 0)
 		goto fail;
 
+encoder_retry:
+	/* Ensure the port clock defaults are reset when retrying. */
+	pipe_config->port_clock = 0;
+	pipe_config->pixel_multiplier = 1;
+
 	/* Pass our mode to the connectors and the CRTC to give them a chance to
 	 * adjust it according to limitations or connector properties, and also
 	 * a chance to reject the mode entirely.
@@ -7646,11 +7851,27 @@ intel_modeset_pipe_config(struct drm_crtc *crtc,
 		}
 	}
 
-	if (!(intel_crtc_compute_config(crtc, pipe_config))) {
+	/* Set default port clock if not overwritten by the encoder. Needs to be
+	 * done afterwards in case the encoder adjusts the mode. */
+	if (!pipe_config->port_clock)
+		pipe_config->port_clock = pipe_config->adjusted_mode.clock;
+
+	ret = intel_crtc_compute_config(to_intel_crtc(crtc), pipe_config);
+	if (ret < 0) {
 		DRM_DEBUG_KMS("CRTC fixup failed\n");
 		goto fail;
 	}
-	DRM_DEBUG_KMS("[CRTC:%d]\n", crtc->base.id);
+
+	if (ret == RETRY) {
+		if (WARN(!retry, "loop in pipe configuration computation\n")) {
+			ret = -EINVAL;
+			goto fail;
+		}
+
+		DRM_DEBUG_KMS("CRTC bw constrained, retrying\n");
+		retry = false;
+		goto encoder_retry;
+	}
 
 	pipe_config->dither = pipe_config->pipe_bpp != plane_bpp;
 	DRM_DEBUG_KMS("plane bpp: %i, pipe bpp: %i, dithering: %i\n",
@@ -7659,7 +7880,7 @@ intel_modeset_pipe_config(struct drm_crtc *crtc,
 	return pipe_config;
 fail:
 	kfree(pipe_config);
-	return ERR_PTR(-EINVAL);
+	return ERR_PTR(ret);
 }
 
 /* Computes which crtcs are affected and sets the relevant bits in the mask. For
@@ -7755,6 +7976,9 @@ intel_modeset_affected_pipes(struct drm_crtc *crtc, unsigned *modeset_pipes,
 	 */
 	*modeset_pipes &= 1 << intel_crtc->pipe;
 	*prepare_pipes &= 1 << intel_crtc->pipe;
+
+	DRM_DEBUG_KMS("set mode pipe masks: modeset: %x, prepare: %x, disable: %x\n",
+		      *modeset_pipes, *prepare_pipes, *disable_pipes);
 }
 
 static bool intel_crtc_in_use(struct drm_crtc *crtc)
@@ -7821,31 +8045,114 @@ intel_modeset_update_state(struct drm_device *dev, unsigned prepare_pipes)
 	list_for_each_entry((intel_crtc), \
 			    &(dev)->mode_config.crtc_list, \
 			    base.head) \
-		if (mask & (1 <<(intel_crtc)->pipe)) \
+		if (mask & (1 <<(intel_crtc)->pipe))
 
 static bool
-intel_pipe_config_compare(struct intel_crtc_config *current_config,
+intel_pipe_config_compare(struct drm_device *dev,
+			  struct intel_crtc_config *current_config,
 			  struct intel_crtc_config *pipe_config)
 {
-	if (current_config->has_pch_encoder != pipe_config->has_pch_encoder) {
-		DRM_ERROR("mismatch in has_pch_encoder "
-			  "(expected %i, found %i)\n",
-			  current_config->has_pch_encoder,
-			  pipe_config->has_pch_encoder);
-		return false;
-	}
+#define PIPE_CONF_CHECK_X(name)	\
+	if (current_config->name != pipe_config->name) { \
+		DRM_ERROR("mismatch in " #name " " \
+			  "(expected 0x%08x, found 0x%08x)\n", \
+			  current_config->name, \
+			  pipe_config->name); \
+		return false; \
+	}
+
+#define PIPE_CONF_CHECK_I(name)	\
+	if (current_config->name != pipe_config->name) { \
+		DRM_ERROR("mismatch in " #name " " \
+			  "(expected %i, found %i)\n", \
+			  current_config->name, \
+			  pipe_config->name); \
+		return false; \
+	}
+
+#define PIPE_CONF_CHECK_FLAGS(name, mask)	\
+	if ((current_config->name ^ pipe_config->name) & (mask)) { \
+		DRM_ERROR("mismatch in " #name " " \
+			  "(expected %i, found %i)\n", \
+			  current_config->name & (mask), \
+			  pipe_config->name & (mask)); \
+		return false; \
+	}
+
+#define PIPE_CONF_QUIRK(quirk)	\
+	((current_config->quirks | pipe_config->quirks) & (quirk))
+
+	PIPE_CONF_CHECK_I(cpu_transcoder);
+
+	PIPE_CONF_CHECK_I(has_pch_encoder);
+	PIPE_CONF_CHECK_I(fdi_lanes);
+	PIPE_CONF_CHECK_I(fdi_m_n.gmch_m);
+	PIPE_CONF_CHECK_I(fdi_m_n.gmch_n);
+	PIPE_CONF_CHECK_I(fdi_m_n.link_m);
+	PIPE_CONF_CHECK_I(fdi_m_n.link_n);
+	PIPE_CONF_CHECK_I(fdi_m_n.tu);
+
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_hdisplay);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_htotal);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_hblank_start);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_hblank_end);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_hsync_start);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_hsync_end);
+
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_vdisplay);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_vtotal);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_vblank_start);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_vblank_end);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_vsync_start);
+	PIPE_CONF_CHECK_I(adjusted_mode.crtc_vsync_end);
+
+	if (!HAS_PCH_SPLIT(dev))
+		PIPE_CONF_CHECK_I(pixel_multiplier);
+
+	PIPE_CONF_CHECK_FLAGS(adjusted_mode.flags,
+			      DRM_MODE_FLAG_INTERLACE);
+
+	if (!PIPE_CONF_QUIRK(PIPE_CONFIG_QUIRK_MODE_SYNC_FLAGS)) {
+		PIPE_CONF_CHECK_FLAGS(adjusted_mode.flags,
+				      DRM_MODE_FLAG_PHSYNC);
+		PIPE_CONF_CHECK_FLAGS(adjusted_mode.flags,
+				      DRM_MODE_FLAG_NHSYNC);
+		PIPE_CONF_CHECK_FLAGS(adjusted_mode.flags,
+				      DRM_MODE_FLAG_PVSYNC);
+		PIPE_CONF_CHECK_FLAGS(adjusted_mode.flags,
+				      DRM_MODE_FLAG_NVSYNC);
+	}
+
+	PIPE_CONF_CHECK_I(requested_mode.hdisplay);
+	PIPE_CONF_CHECK_I(requested_mode.vdisplay);
+
+	PIPE_CONF_CHECK_I(gmch_pfit.control);
+	/* pfit ratios are autocomputed by the hw on gen4+ */
+	if (INTEL_INFO(dev)->gen < 4)
+		PIPE_CONF_CHECK_I(gmch_pfit.pgm_ratios);
+	PIPE_CONF_CHECK_I(gmch_pfit.lvds_border_bits);
+	PIPE_CONF_CHECK_I(pch_pfit.pos);
+	PIPE_CONF_CHECK_I(pch_pfit.size);
+
+	PIPE_CONF_CHECK_I(ips_enabled);
+
+	PIPE_CONF_CHECK_I(shared_dpll);
+	PIPE_CONF_CHECK_X(dpll_hw_state.dpll);
+	PIPE_CONF_CHECK_X(dpll_hw_state.fp0);
+	PIPE_CONF_CHECK_X(dpll_hw_state.fp1);
+
+#undef PIPE_CONF_CHECK_X
+#undef PIPE_CONF_CHECK_I
+#undef PIPE_CONF_CHECK_FLAGS
+#undef PIPE_CONF_QUIRK
 
 	return true;
 }
 
-void
-intel_modeset_check_state(struct drm_device *dev)
+static void
+check_connector_state(struct drm_device *dev)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	struct intel_crtc *crtc;
-	struct intel_encoder *encoder;
 	struct intel_connector *connector;
-	struct intel_crtc_config pipe_config;
 
 	list_for_each_entry(connector, &dev->mode_config.connector_list,
 			    base.head) {
@@ -7856,6 +8163,13 @@ intel_modeset_check_state(struct drm_device *dev)
 		WARN(&connector->new_encoder->base != connector->base.encoder,
 		     "connector's staged encoder doesn't match current encoder\n");
 	}
+}
+
+static void
+check_encoder_state(struct drm_device *dev)
+{
+	struct intel_encoder *encoder;
+	struct intel_connector *connector;
 
 	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
 			    base.head) {
@@ -7907,12 +8221,23 @@ intel_modeset_check_state(struct drm_device *dev)
 		     tracked_pipe, pipe);
 
 	}
+}
+
+static void
+check_crtc_state(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc;
+	struct intel_encoder *encoder;
+	struct intel_crtc_config pipe_config;
 
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list,
 			    base.head) {
 		bool enabled = false;
 		bool active = false;
 
+		memset(&pipe_config, 0, sizeof(pipe_config));
+
 		DRM_DEBUG_KMS("[CRTC:%d]\n",
 			      crtc->base.base.id);
 
@@ -7927,6 +8252,7 @@ intel_modeset_check_state(struct drm_device *dev)
 			if (encoder->connectors_active)
 				active = true;
 		}
+
 		WARN(active != crtc->active,
 		     "crtc's computed active state doesn't match tracked active state "
 		     "(expected %i, found %i)\n", active, crtc->active);
@@ -7934,7 +8260,6 @@ intel_modeset_check_state(struct drm_device *dev)
 		     "crtc's computed enabled state doesn't match tracked enabled state "
 		     "(expected %i, found %i)\n", enabled, crtc->base.enabled);
 
-		memset(&pipe_config, 0, sizeof(pipe_config));
 		active = dev_priv->display.get_pipe_config(crtc,
 							   &pipe_config);
 
@@ -7942,16 +8267,88 @@ intel_modeset_check_state(struct drm_device *dev)
 		if (crtc->pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE)
 			active = crtc->active;
 
+		list_for_each_entry(encoder, &dev->mode_config.encoder_list,
+				    base.head) {
+			if (encoder->base.crtc != &crtc->base)
+				continue;
+			if (encoder->get_config)
+				encoder->get_config(encoder, &pipe_config);
+		}
+
 		WARN(crtc->active != active,
 		     "crtc active state doesn't match with hw state "
 		     "(expected %i, found %i)\n", crtc->active, active);
 
-		WARN(active &&
-		     !intel_pipe_config_compare(&crtc->config, &pipe_config),
-		     "pipe state doesn't match!\n");
+		if (active &&
+		    !intel_pipe_config_compare(dev, &crtc->config, &pipe_config)) {
+			WARN(1, "pipe state doesn't match!\n");
+			intel_dump_pipe_config(crtc, &pipe_config,
+					       "[hw state]");
+			intel_dump_pipe_config(crtc, &crtc->config,
+					       "[sw state]");
+		}
+	}
+}
+
+static void
+check_shared_dpll_state(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct intel_crtc *crtc;
+	struct intel_dpll_hw_state dpll_hw_state;
+	int i;
+
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i];
+		int enabled_crtcs = 0, active_crtcs = 0;
+		bool active;
+
+		memset(&dpll_hw_state, 0, sizeof(dpll_hw_state));
+
+		DRM_DEBUG_KMS("%s\n", pll->name);
+
+		active = pll->get_hw_state(dev_priv, pll, &dpll_hw_state);
+
+		WARN(pll->active > pll->refcount,
+		     "more active pll users than references: %i vs %i\n",
+		     pll->active, pll->refcount);
+		WARN(pll->active && !pll->on,
+		     "pll in active use but not on in sw tracking\n");
+		WARN(pll->on && !pll->active,
+		     "pll in on but not on in use in sw tracking\n");
+		WARN(pll->on != active,
+		     "pll on state mismatch (expected %i, found %i)\n",
+		     pll->on, active);
+
+		list_for_each_entry(crtc, &dev->mode_config.crtc_list,
+				    base.head) {
+			if (crtc->base.enabled && intel_crtc_to_shared_dpll(crtc) == pll)
+				enabled_crtcs++;
+			if (crtc->active && intel_crtc_to_shared_dpll(crtc) == pll)
+				active_crtcs++;
+		}
+		WARN(pll->active != active_crtcs,
+		     "pll active crtcs mismatch (expected %i, found %i)\n",
+		     pll->active, active_crtcs);
+		WARN(pll->refcount != enabled_crtcs,
+		     "pll enabled crtcs mismatch (expected %i, found %i)\n",
+		     pll->refcount, enabled_crtcs);
+
+		WARN(pll->on && memcmp(&pll->hw_state, &dpll_hw_state,
+				       sizeof(dpll_hw_state)),
+		     "pll hw state mismatch\n");
 	}
 }
 
+void
+intel_modeset_check_state(struct drm_device *dev)
+{
+	check_connector_state(dev);
+	check_encoder_state(dev);
+	check_crtc_state(dev);
+	check_shared_dpll_state(dev);
+}
+
 static int __intel_set_mode(struct drm_crtc *crtc,
 			    struct drm_display_mode *mode,
 			    int x, int y, struct drm_framebuffer *fb)
@@ -7988,11 +8385,10 @@ static int __intel_set_mode(struct drm_crtc *crtc,
 
 			goto out;
 		}
+		intel_dump_pipe_config(to_intel_crtc(crtc), pipe_config,
+				       "[modeset]");
 	}
 
-	DRM_DEBUG_KMS("set mode pipe masks: modeset: %x, prepare: %x, disable: %x\n",
-		      modeset_pipes, prepare_pipes, disable_pipes);
-
 	for_each_intel_crtc_masked(dev, disable_pipes, intel_crtc)
 		intel_crtc_disable(&intel_crtc->base);
 
@@ -8005,12 +8401,10 @@ static int __intel_set_mode(struct drm_crtc *crtc,
 	 * to set it here already despite that we pass it down the callchain.
 	 */
 	if (modeset_pipes) {
-		enum transcoder tmp = to_intel_crtc(crtc)->config.cpu_transcoder;
 		crtc->mode = *mode;
 		/* mode_set/enable/disable functions rely on a correct pipe
 		 * config. */
 		to_intel_crtc(crtc)->config = *pipe_config;
-		to_intel_crtc(crtc)->config.cpu_transcoder = tmp;
 	}
 
 	/* Only after disabling all output pipelines that will be changed can we
@@ -8146,15 +8540,20 @@ static void intel_set_config_restore_state(struct drm_device *dev,
 }
 
 static bool
-is_crtc_connector_off(struct drm_crtc *crtc, struct drm_connector *connectors,
-		      int num_connectors)
+is_crtc_connector_off(struct drm_mode_set *set)
 {
 	int i;
 
-	for (i = 0; i < num_connectors; i++)
-		if (connectors[i].encoder &&
-		    connectors[i].encoder->crtc == crtc &&
-		    connectors[i].dpms != DRM_MODE_DPMS_ON)
+	if (set->num_connectors == 0)
+		return false;
+
+	if (WARN_ON(set->connectors == NULL))
+		return false;
+
+	for (i = 0; i < set->num_connectors; i++)
+		if (set->connectors[i]->encoder &&
+		    set->connectors[i]->encoder->crtc == set->crtc &&
+		    set->connectors[i]->dpms != DRM_MODE_DPMS_ON)
 			return true;
 
 	return false;
@@ -8167,10 +8566,8 @@ intel_set_config_compute_mode_changes(struct drm_mode_set *set,
 
 	/* We should be able to check here if the fb has the same properties
 	 * and then just flip_or_move it */
-	if (set->connectors != NULL &&
-	    is_crtc_connector_off(set->crtc, *set->connectors,
-				  set->num_connectors)) {
-			config->mode_changed = true;
+	if (is_crtc_connector_off(set)) {
+		config->mode_changed = true;
 	} else if (set->crtc->fb != set->fb) {
 		/* If we have no fb then treat it as a full mode set */
 		if (set->crtc->fb == NULL) {
@@ -8349,12 +8746,6 @@ static int intel_crtc_set_config(struct drm_mode_set *set)
 		goto fail;
 
 	if (config->mode_changed) {
-		if (set->mode) {
-			DRM_DEBUG_KMS("attempting to set mode from"
-					" userspace\n");
-			drm_mode_debug_printmodeline(set->mode);
-		}
-
 		ret = intel_set_mode(set->crtc, set->mode,
 				     set->x, set->y, set->fb);
 	} else if (config->fb_changed) {
@@ -8365,8 +8756,8 @@ static int intel_crtc_set_config(struct drm_mode_set *set)
 	}
 
 	if (ret) {
-		DRM_ERROR("failed to set mode on [CRTC:%d], err = %d\n",
-			  set->crtc->base.id, ret);
+		DRM_DEBUG_KMS("failed to set mode on [CRTC:%d], err = %d\n",
+			      set->crtc->base.id, ret);
 fail:
 		intel_set_config_restore_state(dev, config);
 
@@ -8397,23 +8788,93 @@ static void intel_cpu_pll_init(struct drm_device *dev)
 		intel_ddi_pll_init(dev);
 }
 
-static void intel_pch_pll_init(struct drm_device *dev)
+static bool ibx_pch_dpll_get_hw_state(struct drm_i915_private *dev_priv,
+				      struct intel_shared_dpll *pll,
+				      struct intel_dpll_hw_state *hw_state)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	int i;
+	uint32_t val;
 
-	if (dev_priv->num_pch_pll == 0) {
-		DRM_DEBUG_KMS("No PCH PLLs on this hardware, skipping initialisation\n");
-		return;
+	val = I915_READ(PCH_DPLL(pll->id));
+	hw_state->dpll = val;
+	hw_state->fp0 = I915_READ(PCH_FP0(pll->id));
+	hw_state->fp1 = I915_READ(PCH_FP1(pll->id));
+
+	return val & DPLL_VCO_ENABLE;
+}
+
+static void ibx_pch_dpll_enable(struct drm_i915_private *dev_priv,
+				struct intel_shared_dpll *pll)
+{
+	uint32_t reg, val;
+
+	/* PCH refclock must be enabled first */
+	assert_pch_refclk_enabled(dev_priv);
+
+	reg = PCH_DPLL(pll->id);
+	val = I915_READ(reg);
+	val |= DPLL_VCO_ENABLE;
+	I915_WRITE(reg, val);
+	POSTING_READ(reg);
+	udelay(200);
+}
+
+static void ibx_pch_dpll_disable(struct drm_i915_private *dev_priv,
+				 struct intel_shared_dpll *pll)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct intel_crtc *crtc;
+	uint32_t reg, val;
+
+	/* Make sure no transcoder isn't still depending on us. */
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, base.head) {
+		if (intel_crtc_to_shared_dpll(crtc) == pll)
+			assert_pch_transcoder_disabled(dev_priv, crtc->pipe);
 	}
 
-	for (i = 0; i < dev_priv->num_pch_pll; i++) {
-		dev_priv->pch_plls[i].pll_reg = _PCH_DPLL(i);
-		dev_priv->pch_plls[i].fp0_reg = _PCH_FP0(i);
-		dev_priv->pch_plls[i].fp1_reg = _PCH_FP1(i);
+	reg = PCH_DPLL(pll->id);
+	val = I915_READ(reg);
+	val &= ~DPLL_VCO_ENABLE;
+	I915_WRITE(reg, val);
+	POSTING_READ(reg);
+	udelay(200);
+}
+
+static char *ibx_pch_dpll_names[] = {
+	"PCH DPLL A",
+	"PCH DPLL B",
+};
+
+static void ibx_pch_dpll_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int i;
+
+	dev_priv->num_shared_dpll = 2;
+
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		dev_priv->shared_dplls[i].id = i;
+		dev_priv->shared_dplls[i].name = ibx_pch_dpll_names[i];
+		dev_priv->shared_dplls[i].enable = ibx_pch_dpll_enable;
+		dev_priv->shared_dplls[i].disable = ibx_pch_dpll_disable;
+		dev_priv->shared_dplls[i].get_hw_state =
+			ibx_pch_dpll_get_hw_state;
 	}
 }
 
+static void intel_shared_dpll_init(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev))
+		ibx_pch_dpll_init(dev);
+	else
+		dev_priv->num_shared_dpll = 0;
+
+	BUG_ON(dev_priv->num_shared_dpll > I915_NUM_PLLS);
+	DRM_DEBUG_KMS("%i shared PLLs initialized\n",
+		      dev_priv->num_shared_dpll);
+}
+
 static void intel_crtc_init(struct drm_device *dev, int pipe)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
@@ -8436,7 +8897,6 @@ static void intel_crtc_init(struct drm_device *dev, int pipe)
 	/* Swap pipes & planes for FBC on pre-965 */
 	intel_crtc->pipe = pipe;
 	intel_crtc->plane = pipe;
-	intel_crtc->config.cpu_transcoder = pipe;
 	if (IS_MOBILE(dev) && IS_GEN3(dev)) {
 		DRM_DEBUG_KMS("swapping pipes & planes for FBC\n");
 		intel_crtc->plane = !pipe;
@@ -8519,13 +8979,8 @@ static void intel_setup_outputs(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_encoder *encoder;
 	bool dpd_is_edp = false;
-	bool has_lvds;
 
-	has_lvds = intel_lvds_init(dev);
-	if (!has_lvds && !HAS_PCH_SPLIT(dev)) {
-		/* disable the panel fitter on everything but LVDS */
-		I915_WRITE(PFIT_CONTROL, 0);
-	}
+	intel_lvds_init(dev);
 
 	if (!IS_ULT(dev))
 		intel_crt_init(dev);
@@ -8598,10 +9053,8 @@ static void intel_setup_outputs(struct drm_device *dev)
 				intel_hdmi_init(dev, GEN4_HDMIB, PORT_B);
 			}
 
-			if (!found && SUPPORTS_INTEGRATED_DP(dev)) {
-				DRM_DEBUG_KMS("probing DP_B\n");
+			if (!found && SUPPORTS_INTEGRATED_DP(dev))
 				intel_dp_init(dev, DP_B, PORT_B);
-			}
 		}
 
 		/* Before G4X SDVOC doesn't have its own detect register */
@@ -8617,17 +9070,13 @@ static void intel_setup_outputs(struct drm_device *dev)
 				DRM_DEBUG_KMS("probing HDMI on SDVOC\n");
 				intel_hdmi_init(dev, GEN4_HDMIC, PORT_C);
 			}
-			if (SUPPORTS_INTEGRATED_DP(dev)) {
-				DRM_DEBUG_KMS("probing DP_C\n");
+			if (SUPPORTS_INTEGRATED_DP(dev))
 				intel_dp_init(dev, DP_C, PORT_C);
-			}
 		}
 
 		if (SUPPORTS_INTEGRATED_DP(dev) &&
-		    (I915_READ(DP_D) & DP_DETECTED)) {
-			DRM_DEBUG_KMS("probing DP_D\n");
+		    (I915_READ(DP_D) & DP_DETECTED))
 			intel_dp_init(dev, DP_D, PORT_D);
-		}
 	} else if (IS_GEN2(dev))
 		intel_dvo_init(dev);
 
@@ -8675,6 +9124,7 @@ int intel_framebuffer_init(struct drm_device *dev,
 			   struct drm_mode_fb_cmd2 *mode_cmd,
 			   struct drm_i915_gem_object *obj)
 {
+	int pitch_limit;
 	int ret;
 
 	if (obj->tiling_mode == I915_TILING_Y) {
@@ -8688,10 +9138,26 @@ int intel_framebuffer_init(struct drm_device *dev,
 		return -EINVAL;
 	}
 
-	/* FIXME <= Gen4 stride limits are bit unclear */
-	if (mode_cmd->pitches[0] > 32768) {
-		DRM_DEBUG("pitch (%d) must be at less than 32768\n",
-			  mode_cmd->pitches[0]);
+	if (INTEL_INFO(dev)->gen >= 5 && !IS_VALLEYVIEW(dev)) {
+		pitch_limit = 32*1024;
+	} else if (INTEL_INFO(dev)->gen >= 4) {
+		if (obj->tiling_mode)
+			pitch_limit = 16*1024;
+		else
+			pitch_limit = 32*1024;
+	} else if (INTEL_INFO(dev)->gen >= 3) {
+		if (obj->tiling_mode)
+			pitch_limit = 8*1024;
+		else
+			pitch_limit = 16*1024;
+	} else
+		/* XXX DSPC is limited to 4k tiled */
+		pitch_limit = 8*1024;
+
+	if (mode_cmd->pitches[0] > pitch_limit) {
+		DRM_DEBUG("%s pitch (%d) must be at less than %d\n",
+			  obj->tiling_mode ? "tiled" : "linear",
+			  mode_cmd->pitches[0], pitch_limit);
 		return -EINVAL;
 	}
 
@@ -8712,7 +9178,8 @@ int intel_framebuffer_init(struct drm_device *dev,
 	case DRM_FORMAT_XRGB1555:
 	case DRM_FORMAT_ARGB1555:
 		if (INTEL_INFO(dev)->gen > 3) {
-			DRM_DEBUG("invalid format: 0x%08x\n", mode_cmd->pixel_format);
+			DRM_DEBUG("unsupported pixel format: %s\n",
+				  drm_get_format_name(mode_cmd->pixel_format));
 			return -EINVAL;
 		}
 		break;
@@ -8723,7 +9190,8 @@ int intel_framebuffer_init(struct drm_device *dev,
 	case DRM_FORMAT_XBGR2101010:
 	case DRM_FORMAT_ABGR2101010:
 		if (INTEL_INFO(dev)->gen < 4) {
-			DRM_DEBUG("invalid format: 0x%08x\n", mode_cmd->pixel_format);
+			DRM_DEBUG("unsupported pixel format: %s\n",
+				  drm_get_format_name(mode_cmd->pixel_format));
 			return -EINVAL;
 		}
 		break;
@@ -8732,12 +9200,14 @@ int intel_framebuffer_init(struct drm_device *dev,
 	case DRM_FORMAT_YVYU:
 	case DRM_FORMAT_VYUY:
 		if (INTEL_INFO(dev)->gen < 5) {
-			DRM_DEBUG("invalid format: 0x%08x\n", mode_cmd->pixel_format);
+			DRM_DEBUG("unsupported pixel format: %s\n",
+				  drm_get_format_name(mode_cmd->pixel_format));
 			return -EINVAL;
 		}
 		break;
 	default:
-		DRM_DEBUG("unsupported pixel format 0x%08x\n", mode_cmd->pixel_format);
+		DRM_DEBUG("unsupported pixel format: %s\n",
+			  drm_get_format_name(mode_cmd->pixel_format));
 		return -EINVAL;
 	}
 
@@ -8782,6 +9252,15 @@ static void intel_init_display(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	if (HAS_PCH_SPLIT(dev) || IS_G4X(dev))
+		dev_priv->display.find_dpll = g4x_find_best_dpll;
+	else if (IS_VALLEYVIEW(dev))
+		dev_priv->display.find_dpll = vlv_find_best_dpll;
+	else if (IS_PINEVIEW(dev))
+		dev_priv->display.find_dpll = pnv_find_best_dpll;
+	else
+		dev_priv->display.find_dpll = i9xx_find_best_dpll;
+
 	if (HAS_DDI(dev)) {
 		dev_priv->display.get_pipe_config = haswell_get_pipe_config;
 		dev_priv->display.crtc_mode_set = haswell_crtc_mode_set;
@@ -8796,6 +9275,13 @@ static void intel_init_display(struct drm_device *dev)
 		dev_priv->display.crtc_disable = ironlake_crtc_disable;
 		dev_priv->display.off = ironlake_crtc_off;
 		dev_priv->display.update_plane = ironlake_update_plane;
+	} else if (IS_VALLEYVIEW(dev)) {
+		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
+		dev_priv->display.crtc_mode_set = i9xx_crtc_mode_set;
+		dev_priv->display.crtc_enable = valleyview_crtc_enable;
+		dev_priv->display.crtc_disable = i9xx_crtc_disable;
+		dev_priv->display.off = i9xx_crtc_off;
+		dev_priv->display.update_plane = i9xx_update_plane;
 	} else {
 		dev_priv->display.get_pipe_config = i9xx_get_pipe_config;
 		dev_priv->display.crtc_mode_set = i9xx_crtc_mode_set;
@@ -8914,6 +9400,17 @@ static void quirk_invert_brightness(struct drm_device *dev)
 	DRM_INFO("applying inverted panel brightness quirk\n");
 }
 
+/*
+ * Some machines (Dell XPS13) suffer broken backlight controls if
+ * BLM_PCH_PWM_ENABLE is set.
+ */
+static void quirk_no_pcm_pwm_enable(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	dev_priv->quirks |= QUIRK_NO_PCH_PWM_ENABLE;
+	DRM_INFO("applying no-PCH_PWM_ENABLE quirk\n");
+}
+
 struct intel_quirk {
 	int device;
 	int subsystem_vendor;
@@ -8983,6 +9480,11 @@ static struct intel_quirk intel_quirks[] = {
 
 	/* Acer Aspire 4736Z */
 	{ 0x2a42, 0x1025, 0x0260, quirk_invert_brightness },
+
+	/* Dell XPS13 HD Sandy Bridge */
+	{ 0x0116, 0x1028, 0x052e, quirk_no_pcm_pwm_enable },
+	/* Dell XPS13 HD and XPS13 FHD Ivy Bridge */
+	{ 0x0166, 0x1028, 0x058b, quirk_no_pcm_pwm_enable },
 };
 
 static void intel_init_quirks(struct drm_device *dev)
@@ -9037,6 +9539,11 @@ void intel_modeset_init_hw(struct drm_device *dev)
 	mutex_unlock(&dev->struct_mutex);
 }
 
+void intel_modeset_suspend_hw(struct drm_device *dev)
+{
+	intel_suspend_hw(dev);
+}
+
 void intel_modeset_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -9082,13 +9589,13 @@ void intel_modeset_init(struct drm_device *dev)
 		for (j = 0; j < dev_priv->num_plane; j++) {
 			ret = intel_plane_init(dev, i, j);
 			if (ret)
-				DRM_DEBUG_KMS("pipe %d plane %d init failed: %d\n",
-					      i, j, ret);
+				DRM_DEBUG_KMS("pipe %c sprite %c init failed: %d\n",
+					      pipe_name(i), sprite_name(i, j), ret);
 		}
 	}
 
 	intel_cpu_pll_init(dev);
-	intel_pch_pll_init(dev);
+	intel_shared_dpll_init(dev);
 
 	/* Just disable it once at startup */
 	i915_disable_vga(dev);
@@ -9289,57 +9796,18 @@ void i915_redisable_vga(struct drm_device *dev)
 	}
 }
 
-/* Scan out the current hw modeset state, sanitizes it and maps it into the drm
- * and i915 state tracking structures. */
-void intel_modeset_setup_hw_state(struct drm_device *dev,
-				  bool force_restore)
+static void intel_modeset_readout_hw_state(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	enum pipe pipe;
-	u32 tmp;
-	struct drm_plane *plane;
 	struct intel_crtc *crtc;
 	struct intel_encoder *encoder;
 	struct intel_connector *connector;
+	int i;
 
-	if (HAS_DDI(dev)) {
-		tmp = I915_READ(TRANS_DDI_FUNC_CTL(TRANSCODER_EDP));
-
-		if (tmp & TRANS_DDI_FUNC_ENABLE) {
-			switch (tmp & TRANS_DDI_EDP_INPUT_MASK) {
-			case TRANS_DDI_EDP_INPUT_A_ON:
-			case TRANS_DDI_EDP_INPUT_A_ONOFF:
-				pipe = PIPE_A;
-				break;
-			case TRANS_DDI_EDP_INPUT_B_ONOFF:
-				pipe = PIPE_B;
-				break;
-			case TRANS_DDI_EDP_INPUT_C_ONOFF:
-				pipe = PIPE_C;
-				break;
-			default:
-				/* A bogus value has been programmed, disable
-				 * the transcoder */
-				WARN(1, "Bogus eDP source %08x\n", tmp);
-				intel_ddi_disable_transcoder_func(dev_priv,
-						TRANSCODER_EDP);
-				goto setup_pipes;
-			}
-
-			crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
-			crtc->config.cpu_transcoder = TRANSCODER_EDP;
-
-			DRM_DEBUG_KMS("Pipe %c using transcoder EDP\n",
-				      pipe_name(pipe));
-		}
-	}
-
-setup_pipes:
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list,
 			    base.head) {
-		enum transcoder tmp = crtc->config.cpu_transcoder;
 		memset(&crtc->config, 0, sizeof(crtc->config));
-		crtc->config.cpu_transcoder = tmp;
 
 		crtc->active = dev_priv->display.get_pipe_config(crtc,
 								 &crtc->config);
@@ -9351,16 +9819,35 @@ setup_pipes:
 			      crtc->active ? "enabled" : "disabled");
 	}
 
+	/* FIXME: Smash this into the new shared dpll infrastructure. */
 	if (HAS_DDI(dev))
 		intel_ddi_setup_hw_pll_state(dev);
 
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i];
+
+		pll->on = pll->get_hw_state(dev_priv, pll, &pll->hw_state);
+		pll->active = 0;
+		list_for_each_entry(crtc, &dev->mode_config.crtc_list,
+				    base.head) {
+			if (crtc->active && intel_crtc_to_shared_dpll(crtc) == pll)
+				pll->active++;
+		}
+		pll->refcount = pll->active;
+
+		DRM_DEBUG_KMS("%s hw state readout: refcount %i, on %i\n",
+			      pll->name, pll->refcount, pll->on);
+	}
+
 	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
 			    base.head) {
 		pipe = 0;
 
 		if (encoder->get_hw_state(encoder, &pipe)) {
-			encoder->base.crtc =
-				dev_priv->pipe_to_crtc_mapping[pipe];
+			crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
+			encoder->base.crtc = &crtc->base;
+			if (encoder->get_config)
+				encoder->get_config(encoder, &crtc->config);
 		} else {
 			encoder->base.crtc = NULL;
 		}
@@ -9388,6 +9875,21 @@ setup_pipes:
 			      drm_get_connector_name(&connector->base),
 			      connector->base.encoder ? "enabled" : "disabled");
 	}
+}
+
+/* Scan out the current hw modeset state, sanitizes it and maps it into the drm
+ * and i915 state tracking structures. */
+void intel_modeset_setup_hw_state(struct drm_device *dev,
+				  bool force_restore)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum pipe pipe;
+	struct drm_plane *plane;
+	struct intel_crtc *crtc;
+	struct intel_encoder *encoder;
+	int i;
+
+	intel_modeset_readout_hw_state(dev);
 
 	/* HW state is read out, now we need to sanitize this mess. */
 	list_for_each_entry(encoder, &dev->mode_config.encoder_list,
@@ -9398,6 +9900,19 @@ setup_pipes:
 	for_each_pipe(pipe) {
 		crtc = to_intel_crtc(dev_priv->pipe_to_crtc_mapping[pipe]);
 		intel_sanitize_crtc(crtc);
+		intel_dump_pipe_config(crtc, &crtc->config, "[setup_hw_state]");
+	}
+
+	for (i = 0; i < dev_priv->num_shared_dpll; i++) {
+		struct intel_shared_dpll *pll = &dev_priv->shared_dplls[i];
+
+		if (!pll->on || pll->active)
+			continue;
+
+		DRM_DEBUG_KMS("%s enabled but not in use, disabling\n", pll->name);
+
+		pll->disable(dev_priv, pll);
+		pll->on = false;
 	}
 
 	if (force_restore) {
@@ -9440,12 +9955,23 @@ void intel_modeset_cleanup(struct drm_device *dev)
 	struct drm_crtc *crtc;
 	struct intel_crtc *intel_crtc;
 
+	/*
+	 * Interrupts and polling as the first thing to avoid creating havoc.
+	 * Too much stuff here (turning of rps, connectors, ...) would
+	 * experience fancy races otherwise.
+	 */
+	drm_irq_uninstall(dev);
+	cancel_work_sync(&dev_priv->hotplug_work);
+	/*
+	 * Due to the hpd irq storm handling the hotplug work can re-arm the
+	 * poll handlers. Hence disable polling after hpd handling is shut down.
+	 */
 	drm_kms_helper_poll_fini(dev);
+
 	mutex_lock(&dev->struct_mutex);
 
 	intel_unregister_dsm_handler();
 
-
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 		/* Skip inactive CRTCs */
 		if (!crtc->fb)
@@ -9461,17 +9987,8 @@ void intel_modeset_cleanup(struct drm_device *dev)
 
 	ironlake_teardown_rc6(dev);
 
-	if (IS_VALLEYVIEW(dev))
-		vlv_init_dpio(dev);
-
 	mutex_unlock(&dev->struct_mutex);
 
-	/* Disable the irq before mode object teardown, for the irq might
-	 * enqueue unpin/hotplug work. */
-	drm_irq_uninstall(dev);
-	cancel_work_sync(&dev_priv->hotplug_work);
-	cancel_work_sync(&dev_priv->rps.work);
-
 	/* flush any delayed tasks or pending work */
 	flush_scheduled_work();
 
@@ -9520,6 +10037,9 @@ int intel_modeset_vga_set_state(struct drm_device *dev, bool state)
 #include <linux/seq_file.h>
 
 struct intel_display_error_state {
+
+	u32 power_well_driver;
+
 	struct intel_cursor_error_state {
 		u32 control;
 		u32 position;
@@ -9528,6 +10048,7 @@ struct intel_display_error_state {
 	} cursor[I915_MAX_PIPES];
 
 	struct intel_pipe_error_state {
+		enum transcoder cpu_transcoder;
 		u32 conf;
 		u32 source;
 
@@ -9562,8 +10083,12 @@ intel_display_capture_error_state(struct drm_device *dev)
 	if (error == NULL)
 		return NULL;
 
+	if (HAS_POWER_WELL(dev))
+		error->power_well_driver = I915_READ(HSW_PWR_WELL_DRIVER);
+
 	for_each_pipe(i) {
 		cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv, i);
+		error->pipe[i].cpu_transcoder = cpu_transcoder;
 
 		if (INTEL_INFO(dev)->gen <= 6 || IS_VALLEYVIEW(dev)) {
 			error->cursor[i].control = I915_READ(CURCNTR(i));
@@ -9598,46 +10123,60 @@ intel_display_capture_error_state(struct drm_device *dev)
 		error->pipe[i].vsync = I915_READ(VSYNC(cpu_transcoder));
 	}
 
+	/* In the code above we read the registers without checking if the power
+	 * well was on, so here we have to clear the FPGA_DBG_RM_NOCLAIM bit to
+	 * prevent the next I915_WRITE from detecting it and printing an error
+	 * message. */
+	if (HAS_POWER_WELL(dev))
+		I915_WRITE_NOTRACE(FPGA_DBG, FPGA_DBG_RM_NOCLAIM);
+
 	return error;
 }
 
+#define err_printf(e, ...) i915_error_printf(e, __VA_ARGS__)
+
 void
-intel_display_print_error_state(struct seq_file *m,
+intel_display_print_error_state(struct drm_i915_error_state_buf *m,
 				struct drm_device *dev,
 				struct intel_display_error_state *error)
 {
 	int i;
 
-	seq_printf(m, "Num Pipes: %d\n", INTEL_INFO(dev)->num_pipes);
+	err_printf(m, "Num Pipes: %d\n", INTEL_INFO(dev)->num_pipes);
+	if (HAS_POWER_WELL(dev))
+		err_printf(m, "PWR_WELL_CTL2: %08x\n",
+			   error->power_well_driver);
 	for_each_pipe(i) {
-		seq_printf(m, "Pipe [%d]:\n", i);
-		seq_printf(m, "  CONF: %08x\n", error->pipe[i].conf);
-		seq_printf(m, "  SRC: %08x\n", error->pipe[i].source);
-		seq_printf(m, "  HTOTAL: %08x\n", error->pipe[i].htotal);
-		seq_printf(m, "  HBLANK: %08x\n", error->pipe[i].hblank);
-		seq_printf(m, "  HSYNC: %08x\n", error->pipe[i].hsync);
-		seq_printf(m, "  VTOTAL: %08x\n", error->pipe[i].vtotal);
-		seq_printf(m, "  VBLANK: %08x\n", error->pipe[i].vblank);
-		seq_printf(m, "  VSYNC: %08x\n", error->pipe[i].vsync);
-
-		seq_printf(m, "Plane [%d]:\n", i);
-		seq_printf(m, "  CNTR: %08x\n", error->plane[i].control);
-		seq_printf(m, "  STRIDE: %08x\n", error->plane[i].stride);
+		err_printf(m, "Pipe [%d]:\n", i);
+		err_printf(m, "  CPU transcoder: %c\n",
+			   transcoder_name(error->pipe[i].cpu_transcoder));
+		err_printf(m, "  CONF: %08x\n", error->pipe[i].conf);
+		err_printf(m, "  SRC: %08x\n", error->pipe[i].source);
+		err_printf(m, "  HTOTAL: %08x\n", error->pipe[i].htotal);
+		err_printf(m, "  HBLANK: %08x\n", error->pipe[i].hblank);
+		err_printf(m, "  HSYNC: %08x\n", error->pipe[i].hsync);
+		err_printf(m, "  VTOTAL: %08x\n", error->pipe[i].vtotal);
+		err_printf(m, "  VBLANK: %08x\n", error->pipe[i].vblank);
+		err_printf(m, "  VSYNC: %08x\n", error->pipe[i].vsync);
+
+		err_printf(m, "Plane [%d]:\n", i);
+		err_printf(m, "  CNTR: %08x\n", error->plane[i].control);
+		err_printf(m, "  STRIDE: %08x\n", error->plane[i].stride);
 		if (INTEL_INFO(dev)->gen <= 3) {
-			seq_printf(m, "  SIZE: %08x\n", error->plane[i].size);
-			seq_printf(m, "  POS: %08x\n", error->plane[i].pos);
+			err_printf(m, "  SIZE: %08x\n", error->plane[i].size);
+			err_printf(m, "  POS: %08x\n", error->plane[i].pos);
 		}
 		if (INTEL_INFO(dev)->gen <= 7 && !IS_HASWELL(dev))
-			seq_printf(m, "  ADDR: %08x\n", error->plane[i].addr);
+			err_printf(m, "  ADDR: %08x\n", error->plane[i].addr);
 		if (INTEL_INFO(dev)->gen >= 4) {
-			seq_printf(m, "  SURF: %08x\n", error->plane[i].surface);
-			seq_printf(m, "  TILEOFF: %08x\n", error->plane[i].tile_offset);
+			err_printf(m, "  SURF: %08x\n", error->plane[i].surface);
+			err_printf(m, "  TILEOFF: %08x\n", error->plane[i].tile_offset);
 		}
 
-		seq_printf(m, "Cursor [%d]:\n", i);
-		seq_printf(m, "  CNTR: %08x\n", error->cursor[i].control);
-		seq_printf(m, "  POS: %08x\n", error->cursor[i].position);
-		seq_printf(m, "  BASE: %08x\n", error->cursor[i].base);
+		err_printf(m, "Cursor [%d]:\n", i);
+		err_printf(m, "  CNTR: %08x\n", error->cursor[i].control);
+		err_printf(m, "  POS: %08x\n", error->cursor[i].position);
+		err_printf(m, "  BASE: %08x\n", error->cursor[i].base);
 	}
 }
 #endif
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 70789b1b5642..26e162bb3a51 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -52,30 +52,6 @@ static bool is_edp(struct intel_dp *intel_dp)
 	return intel_dig_port->base.type == INTEL_OUTPUT_EDP;
 }
 
-/**
- * is_pch_edp - is the port on the PCH and attached to an eDP panel?
- * @intel_dp: DP struct
- *
- * Returns true if the given DP struct corresponds to a PCH DP port attached
- * to an eDP panel, false otherwise.  Helpful for determining whether we
- * may need FDI resources for a given DP output or not.
- */
-static bool is_pch_edp(struct intel_dp *intel_dp)
-{
-	return intel_dp->is_pch_edp;
-}
-
-/**
- * is_cpu_edp - is the port on the CPU and attached to an eDP panel?
- * @intel_dp: DP struct
- *
- * Returns true if the given DP struct corresponds to a CPU eDP port.
- */
-static bool is_cpu_edp(struct intel_dp *intel_dp)
-{
-	return is_edp(intel_dp) && !is_pch_edp(intel_dp);
-}
-
 static struct drm_device *intel_dp_to_dev(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
@@ -88,25 +64,6 @@ static struct intel_dp *intel_attached_dp(struct drm_connector *connector)
 	return enc_to_intel_dp(&intel_attached_encoder(connector)->base);
 }
 
-/**
- * intel_encoder_is_pch_edp - is the given encoder a PCH attached eDP?
- * @encoder: DRM encoder
- *
- * Return true if @encoder corresponds to a PCH attached eDP panel.  Needed
- * by intel_display.c.
- */
-bool intel_encoder_is_pch_edp(struct drm_encoder *encoder)
-{
-	struct intel_dp *intel_dp;
-
-	if (!encoder)
-		return false;
-
-	intel_dp = enc_to_intel_dp(encoder);
-
-	return is_pch_edp(intel_dp);
-}
-
 static void intel_dp_link_down(struct intel_dp *intel_dp);
 
 static int
@@ -118,7 +75,12 @@ intel_dp_max_link_bw(struct intel_dp *intel_dp)
 	case DP_LINK_BW_1_62:
 	case DP_LINK_BW_2_7:
 		break;
+	case DP_LINK_BW_5_4: /* 1.2 capable displays may advertise higher bw */
+		max_link_bw = DP_LINK_BW_2_7;
+		break;
 	default:
+		WARN(1, "invalid max DP link bw val %x, using 1.62Gbps\n",
+		     max_link_bw);
 		max_link_bw = DP_LINK_BW_1_62;
 		break;
 	}
@@ -344,11 +306,12 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 	 * Note that PCH attached eDP panels should use a 125MHz input
 	 * clock divider.
 	 */
-	if (is_cpu_edp(intel_dp)) {
+	if (IS_VALLEYVIEW(dev)) {
+		aux_clock_divider = 100;
+	} else if (intel_dig_port->port == PORT_A) {
 		if (HAS_DDI(dev))
-			aux_clock_divider = intel_ddi_get_cdclk_freq(dev_priv) >> 1;
-		else if (IS_VALLEYVIEW(dev))
-			aux_clock_divider = 100;
+			aux_clock_divider = DIV_ROUND_CLOSEST(
+				intel_ddi_get_cdclk_freq(dev_priv), 2000);
 		else if (IS_GEN6(dev) || IS_GEN7(dev))
 			aux_clock_divider = 200; /* SNB & IVB eDP input clock at 400Mhz */
 		else
@@ -660,6 +623,49 @@ intel_dp_i2c_init(struct intel_dp *intel_dp,
 	return ret;
 }
 
+static void
+intel_dp_set_clock(struct intel_encoder *encoder,
+		   struct intel_crtc_config *pipe_config, int link_bw)
+{
+	struct drm_device *dev = encoder->base.dev;
+
+	if (IS_G4X(dev)) {
+		if (link_bw == DP_LINK_BW_1_62) {
+			pipe_config->dpll.p1 = 2;
+			pipe_config->dpll.p2 = 10;
+			pipe_config->dpll.n = 2;
+			pipe_config->dpll.m1 = 23;
+			pipe_config->dpll.m2 = 8;
+		} else {
+			pipe_config->dpll.p1 = 1;
+			pipe_config->dpll.p2 = 10;
+			pipe_config->dpll.n = 1;
+			pipe_config->dpll.m1 = 14;
+			pipe_config->dpll.m2 = 2;
+		}
+		pipe_config->clock_set = true;
+	} else if (IS_HASWELL(dev)) {
+		/* Haswell has special-purpose DP DDI clocks. */
+	} else if (HAS_PCH_SPLIT(dev)) {
+		if (link_bw == DP_LINK_BW_1_62) {
+			pipe_config->dpll.n = 1;
+			pipe_config->dpll.p1 = 2;
+			pipe_config->dpll.p2 = 10;
+			pipe_config->dpll.m1 = 12;
+			pipe_config->dpll.m2 = 9;
+		} else {
+			pipe_config->dpll.n = 2;
+			pipe_config->dpll.p1 = 1;
+			pipe_config->dpll.p2 = 10;
+			pipe_config->dpll.m1 = 14;
+			pipe_config->dpll.m2 = 8;
+		}
+		pipe_config->clock_set = true;
+	} else if (IS_VALLEYVIEW(dev)) {
+		/* FIXME: Need to figure out optimized DP clocks for vlv. */
+	}
+}
+
 bool
 intel_dp_compute_config(struct intel_encoder *encoder,
 			struct intel_crtc_config *pipe_config)
@@ -667,17 +673,18 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
-	struct drm_display_mode *mode = &pipe_config->requested_mode;
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	enum port port = dp_to_dig_port(intel_dp)->port;
+	struct intel_crtc *intel_crtc = encoder->new_crtc;
 	struct intel_connector *intel_connector = intel_dp->attached_connector;
 	int lane_count, clock;
 	int max_lane_count = drm_dp_max_lane_count(intel_dp->dpcd);
 	int max_clock = intel_dp_max_link_bw(intel_dp) == DP_LINK_BW_2_7 ? 1 : 0;
 	int bpp, mode_rate;
 	static int bws[2] = { DP_LINK_BW_1_62, DP_LINK_BW_2_7 };
-	int target_clock, link_avail, link_clock;
+	int link_avail, link_clock;
 
-	if (HAS_PCH_SPLIT(dev) && !HAS_DDI(dev) && !is_cpu_edp(intel_dp))
+	if (HAS_PCH_SPLIT(dev) && !HAS_DDI(dev) && port != PORT_A)
 		pipe_config->has_pch_encoder = true;
 
 	pipe_config->has_dp_encoder = true;
@@ -685,12 +692,13 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 	if (is_edp(intel_dp) && intel_connector->panel.fixed_mode) {
 		intel_fixed_panel_mode(intel_connector->panel.fixed_mode,
 				       adjusted_mode);
-		intel_pch_panel_fitting(dev,
-					intel_connector->panel.fitting_mode,
-					mode, adjusted_mode);
+		if (!HAS_PCH_SPLIT(dev))
+			intel_gmch_panel_fitting(intel_crtc, pipe_config,
+						 intel_connector->panel.fitting_mode);
+		else
+			intel_pch_panel_fitting(intel_crtc, pipe_config,
+						intel_connector->panel.fitting_mode);
 	}
-	/* We need to take the panel's fixed mode into account. */
-	target_clock = adjusted_mode->clock;
 
 	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK)
 		return false;
@@ -701,12 +709,12 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 
 	/* Walk through all bpp values. Luckily they're all nicely spaced with 2
 	 * bpc in between. */
-	bpp = min_t(int, 8*3, pipe_config->pipe_bpp);
-	if (is_edp(intel_dp) && dev_priv->edp.bpp)
-		bpp = min_t(int, bpp, dev_priv->edp.bpp);
+	bpp = pipe_config->pipe_bpp;
+	if (is_edp(intel_dp) && dev_priv->vbt.edp_bpp)
+		bpp = min_t(int, bpp, dev_priv->vbt.edp_bpp);
 
 	for (; bpp >= 6*3; bpp -= 2*3) {
-		mode_rate = intel_dp_link_required(target_clock, bpp);
+		mode_rate = intel_dp_link_required(adjusted_mode->clock, bpp);
 
 		for (clock = 0; clock <= max_clock; clock++) {
 			for (lane_count = 1; lane_count <= max_lane_count; lane_count <<= 1) {
@@ -741,20 +749,21 @@ found:
 
 	intel_dp->link_bw = bws[clock];
 	intel_dp->lane_count = lane_count;
-	adjusted_mode->clock = drm_dp_bw_code_to_link_rate(intel_dp->link_bw);
 	pipe_config->pipe_bpp = bpp;
-	pipe_config->pixel_target_clock = target_clock;
+	pipe_config->port_clock = drm_dp_bw_code_to_link_rate(intel_dp->link_bw);
 
 	DRM_DEBUG_KMS("DP link bw %02x lane count %d clock %d bpp %d\n",
 		      intel_dp->link_bw, intel_dp->lane_count,
-		      adjusted_mode->clock, bpp);
+		      pipe_config->port_clock, bpp);
 	DRM_DEBUG_KMS("DP link bw required %i available %i\n",
 		      mode_rate, link_avail);
 
 	intel_link_compute_m_n(bpp, lane_count,
-			       target_clock, adjusted_mode->clock,
+			       adjusted_mode->clock, pipe_config->port_clock,
 			       &pipe_config->dp_m_n);
 
+	intel_dp_set_clock(encoder, pipe_config, intel_dp->link_bw);
+
 	return true;
 }
 
@@ -773,24 +782,28 @@ void intel_dp_init_link_config(struct intel_dp *intel_dp)
 	}
 }
 
-static void ironlake_set_pll_edp(struct drm_crtc *crtc, int clock)
+static void ironlake_set_pll_cpu_edp(struct intel_dp *intel_dp)
 {
-	struct drm_device *dev = crtc->dev;
+	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
+	struct intel_crtc *crtc = to_intel_crtc(dig_port->base.base.crtc);
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 dpa_ctl;
 
-	DRM_DEBUG_KMS("eDP PLL enable for clock %d\n", clock);
+	DRM_DEBUG_KMS("eDP PLL enable for clock %d\n", crtc->config.port_clock);
 	dpa_ctl = I915_READ(DP_A);
 	dpa_ctl &= ~DP_PLL_FREQ_MASK;
 
-	if (clock < 200000) {
+	if (crtc->config.port_clock == 162000) {
 		/* For a long time we've carried around a ILK-DevA w/a for the
 		 * 160MHz clock. If we're really unlucky, it's still required.
 		 */
 		DRM_DEBUG_KMS("160MHz cpu eDP clock, might need ilk devA w/a\n");
 		dpa_ctl |= DP_PLL_FREQ_160MHZ;
+		intel_dp->DP |= DP_PLL_FREQ_160MHZ;
 	} else {
 		dpa_ctl |= DP_PLL_FREQ_270MHZ;
+		intel_dp->DP |= DP_PLL_FREQ_270MHZ;
 	}
 
 	I915_WRITE(DP_A, dpa_ctl);
@@ -806,8 +819,8 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 	struct drm_device *dev = encoder->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
-	struct drm_crtc *crtc = encoder->crtc;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	enum port port = dp_to_dig_port(intel_dp)->port;
+	struct intel_crtc *crtc = to_intel_crtc(encoder->crtc);
 
 	/*
 	 * There are four kinds of DP registers:
@@ -833,21 +846,11 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 
 	/* Handle DP bits in common between all three register formats */
 	intel_dp->DP |= DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0;
+	intel_dp->DP |= DP_PORT_WIDTH(intel_dp->lane_count);
 
-	switch (intel_dp->lane_count) {
-	case 1:
-		intel_dp->DP |= DP_PORT_WIDTH_1;
-		break;
-	case 2:
-		intel_dp->DP |= DP_PORT_WIDTH_2;
-		break;
-	case 4:
-		intel_dp->DP |= DP_PORT_WIDTH_4;
-		break;
-	}
 	if (intel_dp->has_audio) {
 		DRM_DEBUG_DRIVER("Enabling DP audio on pipe %c\n",
-				 pipe_name(intel_crtc->pipe));
+				 pipe_name(crtc->pipe));
 		intel_dp->DP |= DP_AUDIO_OUTPUT_ENABLE;
 		intel_write_eld(encoder, adjusted_mode);
 	}
@@ -856,7 +859,7 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 
 	/* Split out the IBX/CPU vs CPT settings */
 
-	if (is_cpu_edp(intel_dp) && IS_GEN7(dev) && !IS_VALLEYVIEW(dev)) {
+	if (port == PORT_A && IS_GEN7(dev) && !IS_VALLEYVIEW(dev)) {
 		if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
 			intel_dp->DP |= DP_SYNC_HS_HIGH;
 		if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
@@ -866,14 +869,8 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		if (intel_dp->link_configuration[1] & DP_LANE_COUNT_ENHANCED_FRAME_EN)
 			intel_dp->DP |= DP_ENHANCED_FRAMING;
 
-		intel_dp->DP |= intel_crtc->pipe << 29;
-
-		/* don't miss out required setting for eDP */
-		if (adjusted_mode->clock < 200000)
-			intel_dp->DP |= DP_PLL_FREQ_160MHZ;
-		else
-			intel_dp->DP |= DP_PLL_FREQ_270MHZ;
-	} else if (!HAS_PCH_CPT(dev) || is_cpu_edp(intel_dp)) {
+		intel_dp->DP |= crtc->pipe << 29;
+	} else if (!HAS_PCH_CPT(dev) || port == PORT_A) {
 		if (!HAS_PCH_SPLIT(dev) && !IS_VALLEYVIEW(dev))
 			intel_dp->DP |= intel_dp->color_range;
 
@@ -886,22 +883,14 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		if (intel_dp->link_configuration[1] & DP_LANE_COUNT_ENHANCED_FRAME_EN)
 			intel_dp->DP |= DP_ENHANCED_FRAMING;
 
-		if (intel_crtc->pipe == 1)
+		if (crtc->pipe == 1)
 			intel_dp->DP |= DP_PIPEB_SELECT;
-
-		if (is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev)) {
-			/* don't miss out required setting for eDP */
-			if (adjusted_mode->clock < 200000)
-				intel_dp->DP |= DP_PLL_FREQ_160MHZ;
-			else
-				intel_dp->DP |= DP_PLL_FREQ_270MHZ;
-		}
 	} else {
 		intel_dp->DP |= DP_LINK_TRAIN_OFF_CPT;
 	}
 
-	if (is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev))
-		ironlake_set_pll_edp(crtc, adjusted_mode->clock);
+	if (port == PORT_A && !IS_VALLEYVIEW(dev))
+		ironlake_set_pll_cpu_edp(intel_dp);
 }
 
 #define IDLE_ON_MASK		(PP_ON | 0 	  | PP_SEQUENCE_MASK | 0                     | PP_SEQUENCE_STATE_MASK)
@@ -1290,6 +1279,7 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 				  enum pipe *pipe)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	enum port port = dp_to_dig_port(intel_dp)->port;
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 tmp = I915_READ(intel_dp->output_reg);
@@ -1297,9 +1287,9 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 	if (!(tmp & DP_PORT_EN))
 		return false;
 
-	if (is_cpu_edp(intel_dp) && IS_GEN7(dev) && !IS_VALLEYVIEW(dev)) {
+	if (port == PORT_A && IS_GEN7(dev) && !IS_VALLEYVIEW(dev)) {
 		*pipe = PORT_TO_PIPE_CPT(tmp);
-	} else if (!HAS_PCH_CPT(dev) || is_cpu_edp(intel_dp)) {
+	} else if (!HAS_PCH_CPT(dev) || port == PORT_A) {
 		*pipe = PORT_TO_PIPE(tmp);
 	} else {
 		u32 trans_sel;
@@ -1335,9 +1325,48 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+static void intel_dp_get_config(struct intel_encoder *encoder,
+				struct intel_crtc_config *pipe_config)
+{
+	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	u32 tmp, flags = 0;
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum port port = dp_to_dig_port(intel_dp)->port;
+	struct intel_crtc *crtc = to_intel_crtc(encoder->base.crtc);
+
+	if ((port == PORT_A) || !HAS_PCH_CPT(dev)) {
+		tmp = I915_READ(intel_dp->output_reg);
+		if (tmp & DP_SYNC_HS_HIGH)
+			flags |= DRM_MODE_FLAG_PHSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NHSYNC;
+
+		if (tmp & DP_SYNC_VS_HIGH)
+			flags |= DRM_MODE_FLAG_PVSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NVSYNC;
+	} else {
+		tmp = I915_READ(TRANS_DP_CTL(crtc->pipe));
+		if (tmp & TRANS_DP_HSYNC_ACTIVE_HIGH)
+			flags |= DRM_MODE_FLAG_PHSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NHSYNC;
+
+		if (tmp & TRANS_DP_VSYNC_ACTIVE_HIGH)
+			flags |= DRM_MODE_FLAG_PVSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NVSYNC;
+	}
+
+	pipe_config->adjusted_mode.flags |= flags;
+}
+
 static void intel_disable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	enum port port = dp_to_dig_port(intel_dp)->port;
+	struct drm_device *dev = encoder->base.dev;
 
 	/* Make sure the panel is off before trying to change the mode. But also
 	 * ensure that we have vdd while we switch off the panel. */
@@ -1347,16 +1376,17 @@ static void intel_disable_dp(struct intel_encoder *encoder)
 	ironlake_edp_panel_off(intel_dp);
 
 	/* cpu edp my only be disable _after_ the cpu pipe/plane is disabled. */
-	if (!is_cpu_edp(intel_dp))
+	if (!(port == PORT_A || IS_VALLEYVIEW(dev)))
 		intel_dp_link_down(intel_dp);
 }
 
 static void intel_post_disable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	enum port port = dp_to_dig_port(intel_dp)->port;
 	struct drm_device *dev = encoder->base.dev;
 
-	if (is_cpu_edp(intel_dp)) {
+	if (port == PORT_A || IS_VALLEYVIEW(dev)) {
 		intel_dp_link_down(intel_dp);
 		if (!IS_VALLEYVIEW(dev))
 			ironlake_edp_pll_off(intel_dp);
@@ -1381,15 +1411,73 @@ static void intel_enable_dp(struct intel_encoder *encoder)
 	intel_dp_complete_link_train(intel_dp);
 	intel_dp_stop_link_train(intel_dp);
 	ironlake_edp_backlight_on(intel_dp);
+
+	if (IS_VALLEYVIEW(dev)) {
+		struct intel_digital_port *dport =
+			enc_to_dig_port(&encoder->base);
+		int channel = vlv_dport_to_channel(dport);
+
+		vlv_wait_port_ready(dev_priv, channel);
+	}
 }
 
 static void intel_pre_enable_dp(struct intel_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
+	struct intel_digital_port *dport = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev))
+	if (dport->port == PORT_A && !IS_VALLEYVIEW(dev))
 		ironlake_edp_pll_on(intel_dp);
+
+	if (IS_VALLEYVIEW(dev)) {
+		struct intel_crtc *intel_crtc =
+			to_intel_crtc(encoder->base.crtc);
+		int port = vlv_dport_to_channel(dport);
+		int pipe = intel_crtc->pipe;
+		u32 val;
+
+		val = vlv_dpio_read(dev_priv, DPIO_DATA_LANE_A(port));
+		val = 0;
+		if (pipe)
+			val |= (1<<21);
+		else
+			val &= ~(1<<21);
+		val |= 0x001000c4;
+		vlv_dpio_write(dev_priv, DPIO_DATA_CHANNEL(port), val);
+
+		vlv_dpio_write(dev_priv, DPIO_PCS_CLOCKBUF0(port),
+				 0x00760018);
+		vlv_dpio_write(dev_priv, DPIO_PCS_CLOCKBUF8(port),
+				 0x00400888);
+	}
+}
+
+static void intel_dp_pre_pll_enable(struct intel_encoder *encoder)
+{
+	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int port = vlv_dport_to_channel(dport);
+
+	if (!IS_VALLEYVIEW(dev))
+		return;
+
+	/* Program Tx lane resets to default */
+	vlv_dpio_write(dev_priv, DPIO_PCS_TX(port),
+			 DPIO_PCS_TX_LANE2_RESET |
+			 DPIO_PCS_TX_LANE1_RESET);
+	vlv_dpio_write(dev_priv, DPIO_PCS_CLK(port),
+			 DPIO_PCS_CLK_CRI_RXEB_EIOS_EN |
+			 DPIO_PCS_CLK_CRI_RXDIGFILTSG_EN |
+			 (1<<DPIO_PCS_CLK_DATAWIDTH_SHIFT) |
+				 DPIO_PCS_CLK_SOFT_RESET);
+
+	/* Fix up inter-pair skew failure */
+	vlv_dpio_write(dev_priv, DPIO_PCS_STAGGER1(port), 0x00750f00);
+	vlv_dpio_write(dev_priv, DPIO_TX_CTL(port), 0x00001500);
+	vlv_dpio_write(dev_priv, DPIO_TX_LANE(port), 0x40400000);
 }
 
 /*
@@ -1451,10 +1539,13 @@ static uint8_t
 intel_dp_voltage_max(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	enum port port = dp_to_dig_port(intel_dp)->port;
 
-	if (IS_GEN7(dev) && is_cpu_edp(intel_dp))
+	if (IS_VALLEYVIEW(dev))
+		return DP_TRAIN_VOLTAGE_SWING_1200;
+	else if (IS_GEN7(dev) && port == PORT_A)
 		return DP_TRAIN_VOLTAGE_SWING_800;
-	else if (HAS_PCH_CPT(dev) && !is_cpu_edp(intel_dp))
+	else if (HAS_PCH_CPT(dev) && port != PORT_A)
 		return DP_TRAIN_VOLTAGE_SWING_1200;
 	else
 		return DP_TRAIN_VOLTAGE_SWING_800;
@@ -1464,6 +1555,7 @@ static uint8_t
 intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	enum port port = dp_to_dig_port(intel_dp)->port;
 
 	if (HAS_DDI(dev)) {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
@@ -1477,7 +1569,19 @@ intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing)
 		default:
 			return DP_TRAIN_PRE_EMPHASIS_0;
 		}
-	} else if (IS_GEN7(dev) && is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev)) {
+	} else if (IS_VALLEYVIEW(dev)) {
+		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_400:
+			return DP_TRAIN_PRE_EMPHASIS_9_5;
+		case DP_TRAIN_VOLTAGE_SWING_600:
+			return DP_TRAIN_PRE_EMPHASIS_6;
+		case DP_TRAIN_VOLTAGE_SWING_800:
+			return DP_TRAIN_PRE_EMPHASIS_3_5;
+		case DP_TRAIN_VOLTAGE_SWING_1200:
+		default:
+			return DP_TRAIN_PRE_EMPHASIS_0;
+		}
+	} else if (IS_GEN7(dev) && port == PORT_A) {
 		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
 		case DP_TRAIN_VOLTAGE_SWING_400:
 			return DP_TRAIN_PRE_EMPHASIS_6;
@@ -1502,6 +1606,101 @@ intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing)
 	}
 }
 
+static uint32_t intel_vlv_signal_levels(struct intel_dp *intel_dp)
+{
+	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_digital_port *dport = dp_to_dig_port(intel_dp);
+	unsigned long demph_reg_value, preemph_reg_value,
+		uniqtranscale_reg_value;
+	uint8_t train_set = intel_dp->train_set[0];
+	int port = vlv_dport_to_channel(dport);
+
+	switch (train_set & DP_TRAIN_PRE_EMPHASIS_MASK) {
+	case DP_TRAIN_PRE_EMPHASIS_0:
+		preemph_reg_value = 0x0004000;
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_400:
+			demph_reg_value = 0x2B405555;
+			uniqtranscale_reg_value = 0x552AB83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_600:
+			demph_reg_value = 0x2B404040;
+			uniqtranscale_reg_value = 0x5548B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_800:
+			demph_reg_value = 0x2B245555;
+			uniqtranscale_reg_value = 0x5560B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_1200:
+			demph_reg_value = 0x2B405555;
+			uniqtranscale_reg_value = 0x5598DA3A;
+			break;
+		default:
+			return 0;
+		}
+		break;
+	case DP_TRAIN_PRE_EMPHASIS_3_5:
+		preemph_reg_value = 0x0002000;
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_400:
+			demph_reg_value = 0x2B404040;
+			uniqtranscale_reg_value = 0x5552B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_600:
+			demph_reg_value = 0x2B404848;
+			uniqtranscale_reg_value = 0x5580B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_800:
+			demph_reg_value = 0x2B404040;
+			uniqtranscale_reg_value = 0x55ADDA3A;
+			break;
+		default:
+			return 0;
+		}
+		break;
+	case DP_TRAIN_PRE_EMPHASIS_6:
+		preemph_reg_value = 0x0000000;
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_400:
+			demph_reg_value = 0x2B305555;
+			uniqtranscale_reg_value = 0x5570B83A;
+			break;
+		case DP_TRAIN_VOLTAGE_SWING_600:
+			demph_reg_value = 0x2B2B4040;
+			uniqtranscale_reg_value = 0x55ADDA3A;
+			break;
+		default:
+			return 0;
+		}
+		break;
+	case DP_TRAIN_PRE_EMPHASIS_9_5:
+		preemph_reg_value = 0x0006000;
+		switch (train_set & DP_TRAIN_VOLTAGE_SWING_MASK) {
+		case DP_TRAIN_VOLTAGE_SWING_400:
+			demph_reg_value = 0x1B405555;
+			uniqtranscale_reg_value = 0x55ADDA3A;
+			break;
+		default:
+			return 0;
+		}
+		break;
+	default:
+		return 0;
+	}
+
+	vlv_dpio_write(dev_priv, DPIO_TX_OCALINIT(port), 0x00000000);
+	vlv_dpio_write(dev_priv, DPIO_TX_SWING_CTL4(port), demph_reg_value);
+	vlv_dpio_write(dev_priv, DPIO_TX_SWING_CTL2(port),
+			 uniqtranscale_reg_value);
+	vlv_dpio_write(dev_priv, DPIO_TX_SWING_CTL3(port), 0x0C782040);
+	vlv_dpio_write(dev_priv, DPIO_PCS_STAGGER0(port), 0x00030000);
+	vlv_dpio_write(dev_priv, DPIO_PCS_CTL_OVER1(port), preemph_reg_value);
+	vlv_dpio_write(dev_priv, DPIO_TX_OCALINIT(port), 0x80000000);
+
+	return 0;
+}
+
 static void
 intel_get_adjust_train(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_STATUS_SIZE])
 {
@@ -1669,6 +1868,7 @@ static void
 intel_dp_set_signal_levels(struct intel_dp *intel_dp, uint32_t *DP)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	enum port port = intel_dig_port->port;
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	uint32_t signal_levels, mask;
 	uint8_t train_set = intel_dp->train_set[0];
@@ -1676,10 +1876,13 @@ intel_dp_set_signal_levels(struct intel_dp *intel_dp, uint32_t *DP)
 	if (HAS_DDI(dev)) {
 		signal_levels = intel_hsw_signal_levels(train_set);
 		mask = DDI_BUF_EMP_MASK;
-	} else if (IS_GEN7(dev) && is_cpu_edp(intel_dp) && !IS_VALLEYVIEW(dev)) {
+	} else if (IS_VALLEYVIEW(dev)) {
+		signal_levels = intel_vlv_signal_levels(intel_dp);
+		mask = 0;
+	} else if (IS_GEN7(dev) && port == PORT_A) {
 		signal_levels = intel_gen7_edp_signal_levels(train_set);
 		mask = EDP_LINK_TRAIN_VOL_EMP_MASK_IVB;
-	} else if (IS_GEN6(dev) && is_cpu_edp(intel_dp)) {
+	} else if (IS_GEN6(dev) && port == PORT_A) {
 		signal_levels = intel_gen6_edp_signal_levels(train_set);
 		mask = EDP_LINK_TRAIN_VOL_EMP_MASK_SNB;
 	} else {
@@ -1729,8 +1932,7 @@ intel_dp_set_link_train(struct intel_dp *intel_dp,
 		}
 		I915_WRITE(DP_TP_CTL(port), temp);
 
-	} else if (HAS_PCH_CPT(dev) &&
-		   (IS_GEN7(dev) || !is_cpu_edp(intel_dp))) {
+	} else if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || port != PORT_A)) {
 		dp_reg_value &= ~DP_LINK_TRAIN_MASK_CPT;
 
 		switch (dp_train_pat & DP_TRAINING_PATTERN_MASK) {
@@ -1981,6 +2183,7 @@ static void
 intel_dp_link_down(struct intel_dp *intel_dp)
 {
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	enum port port = intel_dig_port->port;
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc =
@@ -2010,7 +2213,7 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 
 	DRM_DEBUG_KMS("\n");
 
-	if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || !is_cpu_edp(intel_dp))) {
+	if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || port != PORT_A)) {
 		DP &= ~DP_LINK_TRAIN_MASK_CPT;
 		I915_WRITE(intel_dp->output_reg, DP | DP_LINK_TRAIN_PAT_IDLE_CPT);
 	} else {
@@ -2301,11 +2504,10 @@ intel_dp_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
 			return NULL;
 
 		size = (intel_connector->edid->extensions + 1) * EDID_LENGTH;
-		edid = kmalloc(size, GFP_KERNEL);
+		edid = kmemdup(intel_connector->edid, size, GFP_KERNEL);
 		if (!edid)
 			return NULL;
 
-		memcpy(edid, intel_connector->edid, size);
 		return edid;
 	}
 
@@ -2499,15 +2701,16 @@ done:
 }
 
 static void
-intel_dp_destroy(struct drm_connector *connector)
+intel_dp_connector_destroy(struct drm_connector *connector)
 {
-	struct intel_dp *intel_dp = intel_attached_dp(connector);
 	struct intel_connector *intel_connector = to_intel_connector(connector);
 
 	if (!IS_ERR_OR_NULL(intel_connector->edid))
 		kfree(intel_connector->edid);
 
-	if (is_edp(intel_dp))
+	/* Can't call is_edp() since the encoder may have been destroyed
+	 * already. */
+	if (connector->connector_type == DRM_MODE_CONNECTOR_eDP)
 		intel_panel_fini(&intel_connector->panel);
 
 	drm_sysfs_connector_remove(connector);
@@ -2541,7 +2744,7 @@ static const struct drm_connector_funcs intel_dp_connector_funcs = {
 	.detect = intel_dp_detect,
 	.fill_modes = drm_helper_probe_single_connector_modes,
 	.set_property = intel_dp_set_property,
-	.destroy = intel_dp_destroy,
+	.destroy = intel_dp_connector_destroy,
 };
 
 static const struct drm_connector_helper_funcs intel_dp_connector_helper_funcs = {
@@ -2588,11 +2791,11 @@ bool intel_dpd_is_edp(struct drm_device *dev)
 	struct child_device_config *p_child;
 	int i;
 
-	if (!dev_priv->child_dev_num)
+	if (!dev_priv->vbt.child_dev_num)
 		return false;
 
-	for (i = 0; i < dev_priv->child_dev_num; i++) {
-		p_child = dev_priv->child_dev + i;
+	for (i = 0; i < dev_priv->vbt.child_dev_num; i++) {
+		p_child = dev_priv->vbt.child_dev + i;
 
 		if (p_child->dvo_port == PORT_IDPD &&
 		    p_child->device_type == DEVICE_TYPE_eDP)
@@ -2670,7 +2873,7 @@ intel_dp_init_panel_power_sequencer(struct drm_device *dev,
 	DRM_DEBUG_KMS("cur t1_t3 %d t8 %d t9 %d t10 %d t11_t12 %d\n",
 		      cur.t1_t3, cur.t8, cur.t9, cur.t10, cur.t11_t12);
 
-	vbt = dev_priv->edp.pps;
+	vbt = dev_priv->vbt.edp_pps;
 
 	/* Upper limits from eDP 1.3 spec. Note that we use the clunky units of
 	 * our hw here, which are all in 100usec. */
@@ -2738,9 +2941,6 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 		pp_div_reg = PIPEA_PP_DIVISOR;
 	}
 
-	if (IS_VALLEYVIEW(dev))
-		port_sel = I915_READ(pp_on_reg) & 0xc0000000;
-
 	/* And finally store the new values in the power sequencer. */
 	pp_on = (seq->t1_t3 << PANEL_POWER_UP_DELAY_SHIFT) |
 		(seq->t8 << PANEL_LIGHT_ON_DELAY_SHIFT);
@@ -2754,8 +2954,10 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 
 	/* Haswell doesn't have any port selection bits for the panel
 	 * power sequencer any more. */
-	if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) {
-		if (is_cpu_edp(intel_dp))
+	if (IS_VALLEYVIEW(dev)) {
+		port_sel = I915_READ(pp_on_reg) & 0xc0000000;
+	} else if (HAS_PCH_IBX(dev) || HAS_PCH_CPT(dev)) {
+		if (dp_to_dig_port(intel_dp)->port == PORT_A)
 			port_sel = PANEL_POWER_PORT_DP_A;
 		else
 			port_sel = PANEL_POWER_PORT_DP_D;
@@ -2773,7 +2975,85 @@ intel_dp_init_panel_power_sequencer_registers(struct drm_device *dev,
 		      I915_READ(pp_div_reg));
 }
 
-void
+static bool intel_edp_init_connector(struct intel_dp *intel_dp,
+				     struct intel_connector *intel_connector)
+{
+	struct drm_connector *connector = &intel_connector->base;
+	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+	struct drm_device *dev = intel_dig_port->base.base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_display_mode *fixed_mode = NULL;
+	struct edp_power_seq power_seq = { 0 };
+	bool has_dpcd;
+	struct drm_display_mode *scan;
+	struct edid *edid;
+
+	if (!is_edp(intel_dp))
+		return true;
+
+	intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
+
+	/* Cache DPCD and EDID for edp. */
+	ironlake_edp_panel_vdd_on(intel_dp);
+	has_dpcd = intel_dp_get_dpcd(intel_dp);
+	ironlake_edp_panel_vdd_off(intel_dp, false);
+
+	if (has_dpcd) {
+		if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11)
+			dev_priv->no_aux_handshake =
+				intel_dp->dpcd[DP_MAX_DOWNSPREAD] &
+				DP_NO_AUX_HANDSHAKE_LINK_TRAINING;
+	} else {
+		/* if this fails, presume the device is a ghost */
+		DRM_INFO("failed to retrieve link info, disabling eDP\n");
+		return false;
+	}
+
+	/* We now know it's not a ghost, init power sequence regs. */
+	intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
+						      &power_seq);
+
+	ironlake_edp_panel_vdd_on(intel_dp);
+	edid = drm_get_edid(connector, &intel_dp->adapter);
+	if (edid) {
+		if (drm_add_edid_modes(connector, edid)) {
+			drm_mode_connector_update_edid_property(connector,
+								edid);
+			drm_edid_to_eld(connector, edid);
+		} else {
+			kfree(edid);
+			edid = ERR_PTR(-EINVAL);
+		}
+	} else {
+		edid = ERR_PTR(-ENOENT);
+	}
+	intel_connector->edid = edid;
+
+	/* prefer fixed mode from EDID if available */
+	list_for_each_entry(scan, &connector->probed_modes, head) {
+		if ((scan->type & DRM_MODE_TYPE_PREFERRED)) {
+			fixed_mode = drm_mode_duplicate(dev, scan);
+			break;
+		}
+	}
+
+	/* fallback to VBT if available for eDP */
+	if (!fixed_mode && dev_priv->vbt.lfp_lvds_vbt_mode) {
+		fixed_mode = drm_mode_duplicate(dev,
+					dev_priv->vbt.lfp_lvds_vbt_mode);
+		if (fixed_mode)
+			fixed_mode->type |= DRM_MODE_TYPE_PREFERRED;
+	}
+
+	ironlake_edp_panel_vdd_off(intel_dp, false);
+
+	intel_panel_init(&intel_connector->panel, fixed_mode);
+	intel_panel_setup_backlight(connector);
+
+	return true;
+}
+
+bool
 intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 			struct intel_connector *intel_connector)
 {
@@ -2782,38 +3062,47 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_device *dev = intel_encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_display_mode *fixed_mode = NULL;
-	struct edp_power_seq power_seq = { 0 };
 	enum port port = intel_dig_port->port;
 	const char *name = NULL;
-	int type;
+	int type, error;
 
 	/* Preserve the current hw state. */
 	intel_dp->DP = I915_READ(intel_dp->output_reg);
 	intel_dp->attached_connector = intel_connector;
 
-	if (HAS_PCH_SPLIT(dev) && port == PORT_D)
-		if (intel_dpd_is_edp(dev))
-			intel_dp->is_pch_edp = true;
-
+	type = DRM_MODE_CONNECTOR_DisplayPort;
 	/*
 	 * FIXME : We need to initialize built-in panels before external panels.
 	 * For X0, DP_C is fixed as eDP. Revisit this as part of VLV eDP cleanup
 	 */
-	if (IS_VALLEYVIEW(dev) && port == PORT_C) {
-		type = DRM_MODE_CONNECTOR_eDP;
-		intel_encoder->type = INTEL_OUTPUT_EDP;
-	} else if (port == PORT_A || is_pch_edp(intel_dp)) {
+	switch (port) {
+	case PORT_A:
 		type = DRM_MODE_CONNECTOR_eDP;
-		intel_encoder->type = INTEL_OUTPUT_EDP;
-	} else {
-		/* The intel_encoder->type value may be INTEL_OUTPUT_UNKNOWN for
-		 * DDI or INTEL_OUTPUT_DISPLAYPORT for the older gens, so don't
-		 * rewrite it.
-		 */
-		type = DRM_MODE_CONNECTOR_DisplayPort;
+		break;
+	case PORT_C:
+		if (IS_VALLEYVIEW(dev))
+			type = DRM_MODE_CONNECTOR_eDP;
+		break;
+	case PORT_D:
+		if (HAS_PCH_SPLIT(dev) && intel_dpd_is_edp(dev))
+			type = DRM_MODE_CONNECTOR_eDP;
+		break;
+	default:	/* silence GCC warning */
+		break;
 	}
 
+	/*
+	 * For eDP we always set the encoder type to INTEL_OUTPUT_EDP, but
+	 * for DP the encoder type can be set by the caller to
+	 * INTEL_OUTPUT_UNKNOWN for DDI, so don't rewrite it.
+	 */
+	if (type == DRM_MODE_CONNECTOR_eDP)
+		intel_encoder->type = INTEL_OUTPUT_EDP;
+
+	DRM_DEBUG_KMS("Adding %s connector on port %c\n",
+			type == DRM_MODE_CONNECTOR_eDP ? "eDP" : "DP",
+			port_name(port));
+
 	drm_connector_init(dev, connector, &intel_dp_connector_funcs, type);
 	drm_connector_helper_add(connector, &intel_dp_connector_helper_funcs);
 
@@ -2873,74 +3162,21 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 		BUG();
 	}
 
-	if (is_edp(intel_dp))
-		intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
-
-	intel_dp_i2c_init(intel_dp, intel_connector, name);
-
-	/* Cache DPCD and EDID for edp. */
-	if (is_edp(intel_dp)) {
-		bool ret;
-		struct drm_display_mode *scan;
-		struct edid *edid;
-
-		ironlake_edp_panel_vdd_on(intel_dp);
-		ret = intel_dp_get_dpcd(intel_dp);
-		ironlake_edp_panel_vdd_off(intel_dp, false);
+	error = intel_dp_i2c_init(intel_dp, intel_connector, name);
+	WARN(error, "intel_dp_i2c_init failed with error %d for port %c\n",
+	     error, port_name(port));
 
-		if (ret) {
-			if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11)
-				dev_priv->no_aux_handshake =
-					intel_dp->dpcd[DP_MAX_DOWNSPREAD] &
-					DP_NO_AUX_HANDSHAKE_LINK_TRAINING;
-		} else {
-			/* if this fails, presume the device is a ghost */
-			DRM_INFO("failed to retrieve link info, disabling eDP\n");
-			intel_dp_encoder_destroy(&intel_encoder->base);
-			intel_dp_destroy(connector);
-			return;
-		}
-
-		/* We now know it's not a ghost, init power sequence regs. */
-		intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
-							      &power_seq);
-
-		ironlake_edp_panel_vdd_on(intel_dp);
-		edid = drm_get_edid(connector, &intel_dp->adapter);
-		if (edid) {
-			if (drm_add_edid_modes(connector, edid)) {
-				drm_mode_connector_update_edid_property(connector, edid);
-				drm_edid_to_eld(connector, edid);
-			} else {
-				kfree(edid);
-				edid = ERR_PTR(-EINVAL);
-			}
-		} else {
-			edid = ERR_PTR(-ENOENT);
+	if (!intel_edp_init_connector(intel_dp, intel_connector)) {
+		i2c_del_adapter(&intel_dp->adapter);
+		if (is_edp(intel_dp)) {
+			cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
+			mutex_lock(&dev->mode_config.mutex);
+			ironlake_panel_vdd_off_sync(intel_dp);
+			mutex_unlock(&dev->mode_config.mutex);
 		}
-		intel_connector->edid = edid;
-
-		/* prefer fixed mode from EDID if available */
-		list_for_each_entry(scan, &connector->probed_modes, head) {
-			if ((scan->type & DRM_MODE_TYPE_PREFERRED)) {
-				fixed_mode = drm_mode_duplicate(dev, scan);
-				break;
-			}
-		}
-
-		/* fallback to VBT if available for eDP */
-		if (!fixed_mode && dev_priv->lfp_lvds_vbt_mode) {
-			fixed_mode = drm_mode_duplicate(dev, dev_priv->lfp_lvds_vbt_mode);
-			if (fixed_mode)
-				fixed_mode->type |= DRM_MODE_TYPE_PREFERRED;
-		}
-
-		ironlake_edp_panel_vdd_off(intel_dp, false);
-	}
-
-	if (is_edp(intel_dp)) {
-		intel_panel_init(&intel_connector->panel, fixed_mode);
-		intel_panel_setup_backlight(connector);
+		drm_sysfs_connector_remove(connector);
+		drm_connector_cleanup(connector);
+		return false;
 	}
 
 	intel_dp_add_properties(intel_dp, connector);
@@ -2953,6 +3189,8 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 		u32 temp = I915_READ(PEG_BAND_GAP_DATA);
 		I915_WRITE(PEG_BAND_GAP_DATA, (temp & ~0xf) | 0xd);
 	}
+
+	return true;
 }
 
 void
@@ -2986,6 +3224,9 @@ intel_dp_init(struct drm_device *dev, int output_reg, enum port port)
 	intel_encoder->disable = intel_disable_dp;
 	intel_encoder->post_disable = intel_post_disable_dp;
 	intel_encoder->get_hw_state = intel_dp_get_hw_state;
+	intel_encoder->get_config = intel_dp_get_config;
+	if (IS_VALLEYVIEW(dev))
+		intel_encoder->pre_pll_enable = intel_dp_pre_pll_enable;
 
 	intel_dig_port->port = port;
 	intel_dig_port->dp.output_reg = output_reg;
@@ -2995,5 +3236,9 @@ intel_dp_init(struct drm_device *dev, int output_reg, enum port port)
 	intel_encoder->cloneable = false;
 	intel_encoder->hot_plug = intel_dp_hot_plug;
 
-	intel_dp_init_connector(intel_dig_port, intel_connector);
+	if (!intel_dp_init_connector(intel_dig_port, intel_connector)) {
+		drm_encoder_cleanup(encoder);
+		kfree(intel_dig_port);
+		kfree(intel_connector);
+	}
 }
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 624a9e6b8d71..b7d6e09456ce 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -120,7 +120,6 @@ struct intel_encoder {
 	struct intel_crtc *new_crtc;
 
 	int type;
-	bool needs_tv_clock;
 	/*
 	 * Intel hw has only one MUX where encoders could be clone, hence a
 	 * simple flag is enough to compute the possible_clones mask.
@@ -140,6 +139,12 @@ struct intel_encoder {
 	 * the encoder is active. If the encoder is enabled it also set the pipe
 	 * it is connected to in the pipe parameter. */
 	bool (*get_hw_state)(struct intel_encoder *, enum pipe *pipe);
+	/* Reconstructs the equivalent mode flags for the current hardware
+	 * state. This must be called _after_ display->get_pipe_config has
+	 * pre-filled the pipe config. Note that intel_encoder->base.crtc must
+	 * be set correctly before calling this function. */
+	void (*get_config)(struct intel_encoder *,
+			   struct intel_crtc_config *pipe_config);
 	int crtc_mask;
 	enum hpd_pin hpd_pin;
 };
@@ -177,7 +182,30 @@ struct intel_connector {
 	u8 polled;
 };
 
+typedef struct dpll {
+	/* given values */
+	int n;
+	int m1, m2;
+	int p1, p2;
+	/* derived values */
+	int	dot;
+	int	vco;
+	int	m;
+	int	p;
+} intel_clock_t;
+
 struct intel_crtc_config {
+	/**
+	 * quirks - bitfield with hw state readout quirks
+	 *
+	 * For various reasons the hw state readout code might not be able to
+	 * completely faithfully read out the current state. These cases are
+	 * tracked with quirk flags so that fastboot and state checker can act
+	 * accordingly.
+	 */
+#define PIPE_CONFIG_QUIRK_MODE_SYNC_FLAGS (1<<0) /* unreliable sync mode.flags */
+	unsigned long quirks;
+
 	struct drm_display_mode requested_mode;
 	struct drm_display_mode adjusted_mode;
 	/* This flag must be set by the encoder's compute_config callback if it
@@ -201,29 +229,67 @@ struct intel_crtc_config {
 	/* DP has a bunch of special case unfortunately, so mark the pipe
 	 * accordingly. */
 	bool has_dp_encoder;
+
+	/*
+	 * Enable dithering, used when the selected pipe bpp doesn't match the
+	 * plane bpp.
+	 */
 	bool dither;
 
 	/* Controls for the clock computation, to override various stages. */
 	bool clock_set;
 
+	/* SDVO TV has a bunch of special case. To make multifunction encoders
+	 * work correctly, we need to track this at runtime.*/
+	bool sdvo_tv_clock;
+
+	/*
+	 * crtc bandwidth limit, don't increase pipe bpp or clock if not really
+	 * required. This is set in the 2nd loop of calling encoder's
+	 * ->compute_config if the first pick doesn't work out.
+	 */
+	bool bw_constrained;
+
 	/* Settings for the intel dpll used on pretty much everything but
 	 * haswell. */
-	struct dpll {
-		unsigned n;
-		unsigned m1, m2;
-		unsigned p1, p2;
-	} dpll;
+	struct dpll dpll;
+
+	/* Selected dpll when shared or DPLL_ID_PRIVATE. */
+	enum intel_dpll_id shared_dpll;
+
+	/* Actual register state of the dpll, for shared dpll cross-checking. */
+	struct intel_dpll_hw_state dpll_hw_state;
 
 	int pipe_bpp;
 	struct intel_link_m_n dp_m_n;
-	/**
-	 * This is currently used by DP and HDMI encoders since those can have a
-	 * target pixel clock != the port link clock (which is currently stored
-	 * in adjusted_mode->clock).
+
+	/*
+	 * Frequence the dpll for the port should run at. Differs from the
+	 * adjusted dotclock e.g. for DP or 12bpc hdmi mode.
 	 */
-	int pixel_target_clock;
+	int port_clock;
+
 	/* Used by SDVO (and if we ever fix it, HDMI). */
 	unsigned pixel_multiplier;
+
+	/* Panel fitter controls for gen2-gen4 + VLV */
+	struct {
+		u32 control;
+		u32 pgm_ratios;
+		u32 lvds_border_bits;
+	} gmch_pfit;
+
+	/* Panel fitter placement and size for Ironlake+ */
+	struct {
+		u32 pos;
+		u32 size;
+	} pch_pfit;
+
+	/* FDI configuration, only valid if has_pch_encoder is set. */
+	int fdi_lanes;
+	struct intel_link_m_n fdi_m_n;
+
+	bool ips_enabled;
 };
 
 struct intel_crtc {
@@ -242,7 +308,6 @@ struct intel_crtc {
 	bool lowfreq_avail;
 	struct intel_overlay *overlay;
 	struct intel_unpin_work *unpin_work;
-	int fdi_lanes;
 
 	atomic_t unpin_work_count;
 
@@ -259,12 +324,14 @@ struct intel_crtc {
 
 	struct intel_crtc_config config;
 
-	/* We can share PLLs across outputs if the timings match */
-	struct intel_pch_pll *pch_pll;
 	uint32_t ddi_pll_sel;
 
 	/* reset counter value when the last flip was submitted */
 	unsigned int reset_counter;
+
+	/* Access to these should be protected by dev_priv->irq_lock. */
+	bool cpu_fifo_underrun_disabled;
+	bool pch_fifo_underrun_disabled;
 };
 
 struct intel_plane {
@@ -279,6 +346,18 @@ struct intel_plane {
 	unsigned int crtc_w, crtc_h;
 	uint32_t src_x, src_y;
 	uint32_t src_w, src_h;
+
+	/* Since we need to change the watermarks before/after
+	 * enabling/disabling the planes, we need to store the parameters here
+	 * as the other pieces of the struct may not reflect the values we want
+	 * for the watermark calculations. Currently only Haswell uses this.
+	 */
+	struct {
+		bool enable;
+		uint8_t bytes_per_pixel;
+		uint32_t horiz_pixels;
+	} wm;
+
 	void (*update_plane)(struct drm_plane *plane,
 			     struct drm_framebuffer *fb,
 			     struct drm_i915_gem_object *obj,
@@ -411,7 +490,6 @@ struct intel_dp {
 	uint8_t downstream_ports[DP_MAX_DOWNSTREAM_PORTS];
 	struct i2c_adapter adapter;
 	struct i2c_algo_dp_aux_data algo;
-	bool is_pch_edp;
 	uint8_t train_set[4];
 	int panel_power_up_delay;
 	int panel_power_down_delay;
@@ -426,11 +504,24 @@ struct intel_dp {
 struct intel_digital_port {
 	struct intel_encoder base;
 	enum port port;
-	u32 port_reversal;
+	u32 saved_port_bits;
 	struct intel_dp dp;
 	struct intel_hdmi hdmi;
 };
 
+static inline int
+vlv_dport_to_channel(struct intel_digital_port *dport)
+{
+	switch (dport->port) {
+	case PORT_B:
+		return 0;
+	case PORT_C:
+		return 1;
+	default:
+		BUG();
+	}
+}
+
 static inline struct drm_crtc *
 intel_get_crtc_for_pipe(struct drm_device *dev, int pipe)
 {
@@ -474,6 +565,7 @@ int intel_ddc_get_modes(struct drm_connector *c, struct i2c_adapter *adapter);
 extern void intel_attach_force_audio_property(struct drm_connector *connector);
 extern void intel_attach_broadcast_rgb_property(struct drm_connector *connector);
 
+extern bool intel_pipe_has_type(struct drm_crtc *crtc, int type);
 extern void intel_crt_init(struct drm_device *dev);
 extern void intel_hdmi_init(struct drm_device *dev,
 			    int hdmi_reg, enum port port);
@@ -488,13 +580,14 @@ extern bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg,
 extern void intel_dvo_init(struct drm_device *dev);
 extern void intel_tv_init(struct drm_device *dev);
 extern void intel_mark_busy(struct drm_device *dev);
-extern void intel_mark_fb_busy(struct drm_i915_gem_object *obj);
+extern void intel_mark_fb_busy(struct drm_i915_gem_object *obj,
+			       struct intel_ring_buffer *ring);
 extern void intel_mark_idle(struct drm_device *dev);
-extern bool intel_lvds_init(struct drm_device *dev);
+extern void intel_lvds_init(struct drm_device *dev);
 extern bool intel_is_dual_link_lvds(struct drm_device *dev);
 extern void intel_dp_init(struct drm_device *dev, int output_reg,
 			  enum port port);
-extern void intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
+extern bool intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 				    struct intel_connector *intel_connector);
 extern void intel_dp_init_link_config(struct intel_dp *intel_dp);
 extern void intel_dp_start_link_train(struct intel_dp *intel_dp);
@@ -512,7 +605,6 @@ extern void ironlake_edp_panel_on(struct intel_dp *intel_dp);
 extern void ironlake_edp_panel_off(struct intel_dp *intel_dp);
 extern void ironlake_edp_panel_vdd_on(struct intel_dp *intel_dp);
 extern void ironlake_edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync);
-extern bool intel_encoder_is_pch_edp(struct drm_encoder *encoder);
 extern int intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane);
 extern void intel_flush_display_plane(struct drm_i915_private *dev_priv,
 				      enum plane plane);
@@ -524,12 +616,14 @@ extern void intel_panel_fini(struct intel_panel *panel);
 
 extern void intel_fixed_panel_mode(struct drm_display_mode *fixed_mode,
 				   struct drm_display_mode *adjusted_mode);
-extern void intel_pch_panel_fitting(struct drm_device *dev,
-				    int fitting_mode,
-				    const struct drm_display_mode *mode,
-				    struct drm_display_mode *adjusted_mode);
-extern u32 intel_panel_get_max_backlight(struct drm_device *dev);
-extern void intel_panel_set_backlight(struct drm_device *dev, u32 level);
+extern void intel_pch_panel_fitting(struct intel_crtc *crtc,
+				    struct intel_crtc_config *pipe_config,
+				    int fitting_mode);
+extern void intel_gmch_panel_fitting(struct intel_crtc *crtc,
+				     struct intel_crtc_config *pipe_config,
+				     int fitting_mode);
+extern void intel_panel_set_backlight(struct drm_device *dev,
+				      u32 level, u32 max);
 extern int intel_panel_setup_backlight(struct drm_connector *connector);
 extern void intel_panel_enable_backlight(struct drm_device *dev,
 					 enum pipe pipe);
@@ -553,11 +647,11 @@ extern void intel_crtc_load_lut(struct drm_crtc *crtc);
 extern void intel_crtc_update_dpms(struct drm_crtc *crtc);
 extern void intel_encoder_destroy(struct drm_encoder *encoder);
 extern void intel_encoder_dpms(struct intel_encoder *encoder, int mode);
-extern bool intel_encoder_check_is_cloned(struct intel_encoder *encoder);
 extern void intel_connector_dpms(struct drm_connector *, int mode);
 extern bool intel_connector_get_hw_state(struct intel_connector *connector);
 extern void intel_modeset_check_state(struct drm_device *dev);
 extern void intel_plane_restore(struct drm_plane *plane);
+extern void intel_plane_disable(struct drm_plane *plane);
 
 
 static inline struct intel_encoder *intel_attached_encoder(struct drm_connector *connector)
@@ -565,19 +659,17 @@ static inline struct intel_encoder *intel_attached_encoder(struct drm_connector
 	return to_intel_connector(connector)->encoder;
 }
 
-static inline struct intel_dp *enc_to_intel_dp(struct drm_encoder *encoder)
-{
-	struct intel_digital_port *intel_dig_port =
-		container_of(encoder, struct intel_digital_port, base.base);
-	return &intel_dig_port->dp;
-}
-
 static inline struct intel_digital_port *
 enc_to_dig_port(struct drm_encoder *encoder)
 {
 	return container_of(encoder, struct intel_digital_port, base.base);
 }
 
+static inline struct intel_dp *enc_to_intel_dp(struct drm_encoder *encoder)
+{
+	return &enc_to_dig_port(encoder)->dp;
+}
+
 static inline struct intel_digital_port *
 dp_to_dig_port(struct intel_dp *intel_dp)
 {
@@ -607,6 +699,7 @@ intel_pipe_to_cpu_transcoder(struct drm_i915_private *dev_priv,
 extern void intel_wait_for_vblank(struct drm_device *dev, int pipe);
 extern void intel_wait_for_pipe_off(struct drm_device *dev, int pipe);
 extern int ironlake_get_lanes_required(int target_clock, int link_bw, int bpp);
+extern void vlv_wait_port_ready(struct drm_i915_private *dev_priv, int port);
 
 struct intel_load_detect_pipe {
 	struct drm_framebuffer *release_fb;
@@ -660,13 +753,9 @@ extern void assert_pipe(struct drm_i915_private *dev_priv, enum pipe pipe,
 #define assert_pipe_disabled(d, p) assert_pipe(d, p, false)
 
 extern void intel_init_clock_gating(struct drm_device *dev);
+extern void intel_suspend_hw(struct drm_device *dev);
 extern void intel_write_eld(struct drm_encoder *encoder,
 			    struct drm_display_mode *mode);
-extern void intel_cpt_verify_modeset(struct drm_device *dev, int pipe);
-extern void intel_cpu_transcoder_set_m_n(struct intel_crtc *crtc,
-					 struct intel_link_m_n *m_n);
-extern void intel_pch_transcoder_set_m_n(struct intel_crtc *crtc,
-					 struct intel_link_m_n *m_n);
 extern void intel_prepare_ddi(struct drm_device *dev);
 extern void hsw_fdi_link_train(struct drm_crtc *crtc);
 extern void intel_ddi_init(struct drm_device *dev, enum port port);
@@ -675,9 +764,7 @@ extern void intel_ddi_init(struct drm_device *dev, enum port port);
 extern void intel_update_watermarks(struct drm_device *dev);
 extern void intel_update_sprite_watermarks(struct drm_device *dev, int pipe,
 					   uint32_t sprite_width,
-					   int pixel_size);
-extern void intel_update_linetime_watermarks(struct drm_device *dev, int pipe,
-			 struct drm_display_mode *mode);
+					   int pixel_size, bool enable);
 
 extern unsigned long intel_gen4_compute_page_offset(int *x, int *y,
 						    unsigned int tiling_mode,
@@ -689,8 +776,6 @@ extern int intel_sprite_set_colorkey(struct drm_device *dev, void *data,
 extern int intel_sprite_get_colorkey(struct drm_device *dev, void *data,
 				     struct drm_file *file_priv);
 
-extern u32 intel_dpio_read(struct drm_i915_private *dev_priv, int reg);
-
 /* Power-related functions, located in intel_pm.c */
 extern void intel_init_pm(struct drm_device *dev);
 /* FBC */
@@ -701,7 +786,12 @@ extern void intel_update_fbc(struct drm_device *dev);
 extern void intel_gpu_ips_init(struct drm_i915_private *dev_priv);
 extern void intel_gpu_ips_teardown(void);
 
-extern bool intel_using_power_well(struct drm_device *dev);
+/* Power well */
+extern int i915_init_power_well(struct drm_device *dev);
+extern void i915_remove_power_well(struct drm_device *dev);
+
+extern bool intel_display_power_enabled(struct drm_device *dev,
+					enum intel_display_power_domain domain);
 extern void intel_init_power_well(struct drm_device *dev);
 extern void intel_set_power_well(struct drm_device *dev, bool enable);
 extern void intel_enable_gt_powersave(struct drm_device *dev);
@@ -719,7 +809,7 @@ extern void intel_ddi_disable_transcoder_func(struct drm_i915_private *dev_priv,
 extern void intel_ddi_enable_pipe_clock(struct intel_crtc *intel_crtc);
 extern void intel_ddi_disable_pipe_clock(struct intel_crtc *intel_crtc);
 extern void intel_ddi_setup_hw_pll_state(struct drm_device *dev);
-extern bool intel_ddi_pll_mode_set(struct drm_crtc *crtc, int clock);
+extern bool intel_ddi_pll_mode_set(struct drm_crtc *crtc);
 extern void intel_ddi_put_crtc_pll(struct drm_crtc *crtc);
 extern void intel_ddi_set_pipe_settings(struct drm_crtc *crtc);
 extern void intel_ddi_prepare_link_retrain(struct drm_encoder *encoder);
@@ -728,5 +818,11 @@ intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector);
 extern void intel_ddi_fdi_disable(struct drm_crtc *crtc);
 
 extern void intel_display_handle_reset(struct drm_device *dev);
+extern bool intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
+						  enum pipe pipe,
+						  bool enable);
+extern bool intel_set_pch_fifo_underrun_reporting(struct drm_device *dev,
+						 enum transcoder pch_transcoder,
+						 bool enable);
 
 #endif /* __INTEL_DRV_H__ */
diff --git a/drivers/gpu/drm/i915/intel_dvo.c b/drivers/gpu/drm/i915/intel_dvo.c
index cc70b16d5d42..eb2020eb2b7e 100644
--- a/drivers/gpu/drm/i915/intel_dvo.c
+++ b/drivers/gpu/drm/i915/intel_dvo.c
@@ -54,6 +54,13 @@ static const struct intel_dvo_device intel_dvo_devices[] = {
 		.dev_ops = &ch7xxx_ops,
 	},
 	{
+		.type = INTEL_DVO_CHIP_TMDS,
+		.name = "ch7xxx",
+		.dvo_reg = DVOC,
+		.slave_addr = 0x75, /* For some ch7010 */
+		.dev_ops = &ch7xxx_ops,
+	},
+	{
 		.type = INTEL_DVO_CHIP_LVDS,
 		.name = "ivch",
 		.dvo_reg = DVOA,
@@ -129,6 +136,26 @@ static bool intel_dvo_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+static void intel_dvo_get_config(struct intel_encoder *encoder,
+				 struct intel_crtc_config *pipe_config)
+{
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	struct intel_dvo *intel_dvo = enc_to_intel_dvo(&encoder->base);
+	u32 tmp, flags = 0;
+
+	tmp = I915_READ(intel_dvo->dev.dvo_reg);
+	if (tmp & DVO_HSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+	if (tmp & DVO_VSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	pipe_config->adjusted_mode.flags |= flags;
+}
+
 static void intel_disable_dvo(struct intel_encoder *encoder)
 {
 	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
@@ -153,6 +180,7 @@ static void intel_enable_dvo(struct intel_encoder *encoder)
 	intel_dvo->dev.dev_ops->dpms(&intel_dvo->dev, true);
 }
 
+/* Special dpms function to support cloning between dvo/sdvo/crt. */
 static void intel_dvo_dpms(struct drm_connector *connector, int mode)
 {
 	struct intel_dvo *intel_dvo = intel_attached_dvo(connector);
@@ -174,6 +202,8 @@ static void intel_dvo_dpms(struct drm_connector *connector, int mode)
 		return;
 	}
 
+	/* We call connector dpms manually below in case pipe dpms doesn't
+	 * change due to cloning. */
 	if (mode == DRM_MODE_DPMS_ON) {
 		intel_dvo->base.connectors_active = true;
 
@@ -440,6 +470,7 @@ void intel_dvo_init(struct drm_device *dev)
 	intel_encoder->disable = intel_disable_dvo;
 	intel_encoder->enable = intel_enable_dvo;
 	intel_encoder->get_hw_state = intel_dvo_get_hw_state;
+	intel_encoder->get_config = intel_dvo_get_config;
 	intel_connector->get_hw_state = intel_dvo_connector_get_hw_state;
 
 	/* Now, try to find a controller */
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index 6b7c3ca2c035..dff669e2387f 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -60,8 +60,9 @@ static struct fb_ops intelfb_ops = {
 static int intelfb_create(struct drm_fb_helper *helper,
 			  struct drm_fb_helper_surface_size *sizes)
 {
-	struct intel_fbdev *ifbdev = (struct intel_fbdev *)helper;
-	struct drm_device *dev = ifbdev->helper.dev;
+	struct intel_fbdev *ifbdev =
+		container_of(helper, struct intel_fbdev, helper);
+	struct drm_device *dev = helper->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct fb_info *info;
 	struct drm_framebuffer *fb;
@@ -108,7 +109,7 @@ static int intelfb_create(struct drm_fb_helper *helper,
 		goto out_unpin;
 	}
 
-	info->par = ifbdev;
+	info->par = helper;
 
 	ret = intel_framebuffer_init(dev, &ifbdev->ifb, &mode_cmd, obj);
 	if (ret)
@@ -217,7 +218,7 @@ static void intel_fbdev_destroy(struct drm_device *dev,
 int intel_fbdev_init(struct drm_device *dev)
 {
 	struct intel_fbdev *ifbdev;
-	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
 	ifbdev = kzalloc(sizeof(struct intel_fbdev), GFP_KERNEL);
@@ -242,7 +243,7 @@ int intel_fbdev_init(struct drm_device *dev)
 
 void intel_fbdev_initial_config(struct drm_device *dev)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	/* Due to peculiar init order wrt to hpd handling this is separate. */
 	drm_fb_helper_initial_config(&dev_priv->fbdev->helper, 32);
@@ -250,7 +251,7 @@ void intel_fbdev_initial_config(struct drm_device *dev)
 
 void intel_fbdev_fini(struct drm_device *dev)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	if (!dev_priv->fbdev)
 		return;
 
@@ -261,7 +262,7 @@ void intel_fbdev_fini(struct drm_device *dev)
 
 void intel_fbdev_set_suspend(struct drm_device *dev, int state)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_fbdev *ifbdev = dev_priv->fbdev;
 	struct fb_info *info;
 
@@ -274,7 +275,7 @@ void intel_fbdev_set_suspend(struct drm_device *dev, int state)
 	 * been restored from swap. If the object is stolen however, it will be
 	 * full of whatever garbage was left in there.
 	 */
-	if (!state && ifbdev->ifb.obj->stolen)
+	if (state == FBINFO_STATE_RUNNING && ifbdev->ifb.obj->stolen)
 		memset_io(info->screen_base, 0, info->screen_size);
 
 	fb_set_suspend(info, state);
@@ -284,16 +285,14 @@ MODULE_LICENSE("GPL and additional rights");
 
 void intel_fb_output_poll_changed(struct drm_device *dev)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	drm_fb_helper_hotplug_event(&dev_priv->fbdev->helper);
 }
 
 void intel_fb_restore_mode(struct drm_device *dev)
 {
 	int ret;
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	struct drm_mode_config *config = &dev->mode_config;
-	struct drm_plane *plane;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	if (INTEL_INFO(dev)->num_pipes == 0)
 		return;
@@ -304,10 +303,5 @@ void intel_fb_restore_mode(struct drm_device *dev)
 	if (ret)
 		DRM_DEBUG("failed to restore crtc mode\n");
 
-	/* Be sure to shut off any planes that may be active */
-	list_for_each_entry(plane, &config->plane_list, head)
-		if (plane->enabled)
-			plane->funcs->disable_plane(plane);
-
 	drm_modeset_unlock_all(dev);
 }
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index a9057930f2b2..98df2a0c85bd 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -602,7 +602,7 @@ static void intel_hdmi_mode_set(struct drm_encoder *encoder,
 	u32 hdmi_val;
 
 	hdmi_val = SDVO_ENCODING_HDMI;
-	if (!HAS_PCH_SPLIT(dev) && !IS_VALLEYVIEW(dev))
+	if (!HAS_PCH_SPLIT(dev))
 		hdmi_val |= intel_hdmi->color_range;
 	if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
 		hdmi_val |= SDVO_VSYNC_ACTIVE_HIGH;
@@ -658,6 +658,28 @@ static bool intel_hdmi_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+static void intel_hdmi_get_config(struct intel_encoder *encoder,
+				  struct intel_crtc_config *pipe_config)
+{
+	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	u32 tmp, flags = 0;
+
+	tmp = I915_READ(intel_hdmi->hdmi_reg);
+
+	if (tmp & SDVO_HSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NHSYNC;
+
+	if (tmp & SDVO_VSYNC_ACTIVE_HIGH)
+		flags |= DRM_MODE_FLAG_PVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_NVSYNC;
+
+	pipe_config->adjusted_mode.flags |= flags;
+}
+
 static void intel_enable_hdmi(struct intel_encoder *encoder)
 {
 	struct drm_device *dev = encoder->base.dev;
@@ -697,6 +719,14 @@ static void intel_enable_hdmi(struct intel_encoder *encoder)
 		I915_WRITE(intel_hdmi->hdmi_reg, temp);
 		POSTING_READ(intel_hdmi->hdmi_reg);
 	}
+
+	if (IS_VALLEYVIEW(dev)) {
+		struct intel_digital_port *dport =
+			enc_to_dig_port(&encoder->base);
+		int channel = vlv_dport_to_channel(dport);
+
+		vlv_wait_port_ready(dev_priv, channel);
+	}
 }
 
 static void intel_disable_hdmi(struct intel_encoder *encoder)
@@ -775,6 +805,8 @@ bool intel_hdmi_compute_config(struct intel_encoder *encoder,
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
+	int clock_12bpc = pipe_config->requested_mode.clock * 3 / 2;
+	int desired_bpp;
 
 	if (intel_hdmi->color_range_auto) {
 		/* See CEA-861-E - 5.1 Default Encoding Parameters */
@@ -794,14 +826,29 @@ bool intel_hdmi_compute_config(struct intel_encoder *encoder,
 	/*
 	 * HDMI is either 12 or 8, so if the display lets 10bpc sneak
 	 * through, clamp it down. Note that g4x/vlv don't support 12bpc hdmi
-	 * outputs.
+	 * outputs. We also need to check that the higher clock still fits
+	 * within limits.
 	 */
-	if (pipe_config->pipe_bpp > 8*3 && HAS_PCH_SPLIT(dev)) {
-		DRM_DEBUG_KMS("forcing bpc to 12 for HDMI\n");
-		pipe_config->pipe_bpp = 12*3;
+	if (pipe_config->pipe_bpp > 8*3 && clock_12bpc <= 225000
+	    && HAS_PCH_SPLIT(dev)) {
+		DRM_DEBUG_KMS("picking bpc to 12 for HDMI output\n");
+		desired_bpp = 12*3;
+
+		/* Need to adjust the port link by 1.5x for 12bpc. */
+		pipe_config->port_clock = clock_12bpc;
 	} else {
-		DRM_DEBUG_KMS("forcing bpc to 8 for HDMI\n");
-		pipe_config->pipe_bpp = 8*3;
+		DRM_DEBUG_KMS("picking bpc to 8 for HDMI output\n");
+		desired_bpp = 8*3;
+	}
+
+	if (!pipe_config->bw_constrained) {
+		DRM_DEBUG_KMS("forcing pipe bpc to %i for HDMI\n", desired_bpp);
+		pipe_config->pipe_bpp = desired_bpp;
+	}
+
+	if (adjusted_mode->clock > 225000) {
+		DRM_DEBUG_KMS("too high HDMI clock, rejecting mode\n");
+		return false;
 	}
 
 	return true;
@@ -955,6 +1002,97 @@ done:
 	return 0;
 }
 
+static void intel_hdmi_pre_enable(struct intel_encoder *encoder)
+{
+	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc =
+		to_intel_crtc(encoder->base.crtc);
+	int port = vlv_dport_to_channel(dport);
+	int pipe = intel_crtc->pipe;
+	u32 val;
+
+	if (!IS_VALLEYVIEW(dev))
+		return;
+
+	/* Enable clock channels for this port */
+	val = vlv_dpio_read(dev_priv, DPIO_DATA_LANE_A(port));
+	val = 0;
+	if (pipe)
+		val |= (1<<21);
+	else
+		val &= ~(1<<21);
+	val |= 0x001000c4;
+	vlv_dpio_write(dev_priv, DPIO_DATA_CHANNEL(port), val);
+
+	/* HDMI 1.0V-2dB */
+	vlv_dpio_write(dev_priv, DPIO_TX_OCALINIT(port), 0);
+	vlv_dpio_write(dev_priv, DPIO_TX_SWING_CTL4(port),
+			 0x2b245f5f);
+	vlv_dpio_write(dev_priv, DPIO_TX_SWING_CTL2(port),
+			 0x5578b83a);
+	vlv_dpio_write(dev_priv, DPIO_TX_SWING_CTL3(port),
+			 0x0c782040);
+	vlv_dpio_write(dev_priv, DPIO_TX3_SWING_CTL4(port),
+			 0x2b247878);
+	vlv_dpio_write(dev_priv, DPIO_PCS_STAGGER0(port), 0x00030000);
+	vlv_dpio_write(dev_priv, DPIO_PCS_CTL_OVER1(port),
+			 0x00002000);
+	vlv_dpio_write(dev_priv, DPIO_TX_OCALINIT(port),
+			 DPIO_TX_OCALINIT_EN);
+
+	/* Program lane clock */
+	vlv_dpio_write(dev_priv, DPIO_PCS_CLOCKBUF0(port),
+			 0x00760018);
+	vlv_dpio_write(dev_priv, DPIO_PCS_CLOCKBUF8(port),
+			 0x00400888);
+}
+
+static void intel_hdmi_pre_pll_enable(struct intel_encoder *encoder)
+{
+	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int port = vlv_dport_to_channel(dport);
+
+	if (!IS_VALLEYVIEW(dev))
+		return;
+
+	/* Program Tx lane resets to default */
+	vlv_dpio_write(dev_priv, DPIO_PCS_TX(port),
+			 DPIO_PCS_TX_LANE2_RESET |
+			 DPIO_PCS_TX_LANE1_RESET);
+	vlv_dpio_write(dev_priv, DPIO_PCS_CLK(port),
+			 DPIO_PCS_CLK_CRI_RXEB_EIOS_EN |
+			 DPIO_PCS_CLK_CRI_RXDIGFILTSG_EN |
+			 (1<<DPIO_PCS_CLK_DATAWIDTH_SHIFT) |
+			 DPIO_PCS_CLK_SOFT_RESET);
+
+	/* Fix up inter-pair skew failure */
+	vlv_dpio_write(dev_priv, DPIO_PCS_STAGGER1(port), 0x00750f00);
+	vlv_dpio_write(dev_priv, DPIO_TX_CTL(port), 0x00001500);
+	vlv_dpio_write(dev_priv, DPIO_TX_LANE(port), 0x40400000);
+
+	vlv_dpio_write(dev_priv, DPIO_PCS_CTL_OVER1(port),
+			 0x00002000);
+	vlv_dpio_write(dev_priv, DPIO_TX_OCALINIT(port),
+			 DPIO_TX_OCALINIT_EN);
+}
+
+static void intel_hdmi_post_disable(struct intel_encoder *encoder)
+{
+	struct intel_digital_port *dport = enc_to_dig_port(&encoder->base);
+	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
+	int port = vlv_dport_to_channel(dport);
+
+	/* Reset lanes to avoid HDMI flicker (VLV w/a) */
+	mutex_lock(&dev_priv->dpio_lock);
+	vlv_dpio_write(dev_priv, DPIO_PCS_TX(port), 0x00000000);
+	vlv_dpio_write(dev_priv, DPIO_PCS_CLK(port), 0x00e00060);
+	mutex_unlock(&dev_priv->dpio_lock);
+}
+
 static void intel_hdmi_destroy(struct drm_connector *connector)
 {
 	drm_sysfs_connector_remove(connector);
@@ -1094,6 +1232,12 @@ void intel_hdmi_init(struct drm_device *dev, int hdmi_reg, enum port port)
 	intel_encoder->enable = intel_enable_hdmi;
 	intel_encoder->disable = intel_disable_hdmi;
 	intel_encoder->get_hw_state = intel_hdmi_get_hw_state;
+	intel_encoder->get_config = intel_hdmi_get_config;
+	if (IS_VALLEYVIEW(dev)) {
+		intel_encoder->pre_enable = intel_hdmi_pre_enable;
+		intel_encoder->pre_pll_enable = intel_hdmi_pre_pll_enable;
+		intel_encoder->post_disable = intel_hdmi_post_disable;
+	}
 
 	intel_encoder->type = INTEL_OUTPUT_HDMI;
 	intel_encoder->crtc_mask = (1 << 0) | (1 << 1) | (1 << 2);
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 817f936e2666..61348eae2f04 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -49,8 +49,6 @@ struct intel_lvds_connector {
 struct intel_lvds_encoder {
 	struct intel_encoder base;
 
-	u32 pfit_control;
-	u32 pfit_pgm_ratios;
 	bool is_dual_link;
 	u32 reg;
 
@@ -88,6 +86,38 @@ static bool intel_lvds_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+static void intel_lvds_get_config(struct intel_encoder *encoder,
+				  struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 lvds_reg, tmp, flags = 0;
+
+	if (HAS_PCH_SPLIT(dev))
+		lvds_reg = PCH_LVDS;
+	else
+		lvds_reg = LVDS;
+
+	tmp = I915_READ(lvds_reg);
+	if (tmp & LVDS_HSYNC_POLARITY)
+		flags |= DRM_MODE_FLAG_NHSYNC;
+	else
+		flags |= DRM_MODE_FLAG_PHSYNC;
+	if (tmp & LVDS_VSYNC_POLARITY)
+		flags |= DRM_MODE_FLAG_NVSYNC;
+	else
+		flags |= DRM_MODE_FLAG_PVSYNC;
+
+	pipe_config->adjusted_mode.flags |= flags;
+
+	/* gen2/3 store dither state in pfit control, needs to match */
+	if (INTEL_INFO(dev)->gen < 4) {
+		tmp = I915_READ(PFIT_CONTROL);
+
+		pipe_config->gmch_pfit.control |= tmp & PANEL_8TO6_DITHER_ENABLE;
+	}
+}
+
 /* The LVDS pin pair needs to be on before the DPLLs are enabled.
  * This is an exception to the general rule that mode_set doesn't turn
  * things on.
@@ -118,7 +148,8 @@ static void intel_pre_pll_enable_lvds(struct intel_encoder *encoder)
 	}
 
 	/* set the corresponsding LVDS_BORDER bit */
-	temp |= dev_priv->lvds_border_bits;
+	temp &= ~LVDS_BORDER_ENABLE;
+	temp |= intel_crtc->config.gmch_pfit.lvds_border_bits;
 	/* Set the B0-B3 data pairs corresponding to whether we're going to
 	 * set the DPLLs for dual-channel mode or not.
 	 */
@@ -136,7 +167,10 @@ static void intel_pre_pll_enable_lvds(struct intel_encoder *encoder)
 	 * special lvds dither control bit on pch-split platforms, dithering is
 	 * only controlled through the PIPECONF reg. */
 	if (INTEL_INFO(dev)->gen == 4) {
-		if (dev_priv->lvds_dither)
+		/* Bspec wording suggests that LVDS port dithering only exists
+		 * for 18bpp panels. */
+		if (intel_crtc->config.dither &&
+		    intel_crtc->config.pipe_bpp == 18)
 			temp |= LVDS_ENABLE_DITHER;
 		else
 			temp &= ~LVDS_ENABLE_DITHER;
@@ -150,29 +184,6 @@ static void intel_pre_pll_enable_lvds(struct intel_encoder *encoder)
 	I915_WRITE(lvds_encoder->reg, temp);
 }
 
-static void intel_pre_enable_lvds(struct intel_encoder *encoder)
-{
-	struct drm_device *dev = encoder->base.dev;
-	struct intel_lvds_encoder *enc = to_lvds_encoder(&encoder->base);
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (HAS_PCH_SPLIT(dev) || !enc->pfit_control)
-		return;
-
-	/*
-	 * Enable automatic panel scaling so that non-native modes
-	 * fill the screen.  The panel fitter should only be
-	 * adjusted whilst the pipe is disabled, according to
-	 * register description and PRM.
-	 */
-	DRM_DEBUG_KMS("applying panel-fitter: %x, %x\n",
-		      enc->pfit_control,
-		      enc->pfit_pgm_ratios);
-
-	I915_WRITE(PFIT_PGM_RATIOS, enc->pfit_pgm_ratios);
-	I915_WRITE(PFIT_CONTROL, enc->pfit_control);
-}
-
 /**
  * Sets the power state for the panel.
  */
@@ -241,62 +252,6 @@ static int intel_lvds_mode_valid(struct drm_connector *connector,
 	return MODE_OK;
 }
 
-static void
-centre_horizontally(struct drm_display_mode *mode,
-		    int width)
-{
-	u32 border, sync_pos, blank_width, sync_width;
-
-	/* keep the hsync and hblank widths constant */
-	sync_width = mode->crtc_hsync_end - mode->crtc_hsync_start;
-	blank_width = mode->crtc_hblank_end - mode->crtc_hblank_start;
-	sync_pos = (blank_width - sync_width + 1) / 2;
-
-	border = (mode->hdisplay - width + 1) / 2;
-	border += border & 1; /* make the border even */
-
-	mode->crtc_hdisplay = width;
-	mode->crtc_hblank_start = width + border;
-	mode->crtc_hblank_end = mode->crtc_hblank_start + blank_width;
-
-	mode->crtc_hsync_start = mode->crtc_hblank_start + sync_pos;
-	mode->crtc_hsync_end = mode->crtc_hsync_start + sync_width;
-}
-
-static void
-centre_vertically(struct drm_display_mode *mode,
-		  int height)
-{
-	u32 border, sync_pos, blank_width, sync_width;
-
-	/* keep the vsync and vblank widths constant */
-	sync_width = mode->crtc_vsync_end - mode->crtc_vsync_start;
-	blank_width = mode->crtc_vblank_end - mode->crtc_vblank_start;
-	sync_pos = (blank_width - sync_width + 1) / 2;
-
-	border = (mode->vdisplay - height + 1) / 2;
-
-	mode->crtc_vdisplay = height;
-	mode->crtc_vblank_start = height + border;
-	mode->crtc_vblank_end = mode->crtc_vblank_start + blank_width;
-
-	mode->crtc_vsync_start = mode->crtc_vblank_start + sync_pos;
-	mode->crtc_vsync_end = mode->crtc_vsync_start + sync_width;
-}
-
-static inline u32 panel_fitter_scaling(u32 source, u32 target)
-{
-	/*
-	 * Floating point operation is not supported. So the FACTOR
-	 * is defined, which can avoid the floating point computation
-	 * when calculating the panel ratio.
-	 */
-#define ACCURACY 12
-#define FACTOR (1 << ACCURACY)
-	u32 ratio = source * FACTOR / target;
-	return (FACTOR * ratio + FACTOR/2) / FACTOR;
-}
-
 static bool intel_lvds_compute_config(struct intel_encoder *intel_encoder,
 				      struct intel_crtc_config *pipe_config)
 {
@@ -307,11 +262,8 @@ static bool intel_lvds_compute_config(struct intel_encoder *intel_encoder,
 	struct intel_connector *intel_connector =
 		&lvds_encoder->attached_connector->base;
 	struct drm_display_mode *adjusted_mode = &pipe_config->adjusted_mode;
-	struct drm_display_mode *mode = &pipe_config->requested_mode;
 	struct intel_crtc *intel_crtc = lvds_encoder->base.new_crtc;
-	u32 pfit_control = 0, pfit_pgm_ratios = 0, border = 0;
 	unsigned int lvds_bpp;
-	int pipe;
 
 	/* Should never happen!! */
 	if (INTEL_INFO(dev)->gen < 4 && intel_crtc->pipe == 0) {
@@ -319,20 +271,18 @@ static bool intel_lvds_compute_config(struct intel_encoder *intel_encoder,
 		return false;
 	}
 
-	if (intel_encoder_check_is_cloned(&lvds_encoder->base))
-		return false;
-
 	if ((I915_READ(lvds_encoder->reg) & LVDS_A3_POWER_MASK) ==
 	    LVDS_A3_POWER_UP)
 		lvds_bpp = 8*3;
 	else
 		lvds_bpp = 6*3;
 
-	if (lvds_bpp != pipe_config->pipe_bpp) {
+	if (lvds_bpp != pipe_config->pipe_bpp && !pipe_config->bw_constrained) {
 		DRM_DEBUG_KMS("forcing display bpp (was %d) to LVDS (%d)\n",
 			      pipe_config->pipe_bpp, lvds_bpp);
 		pipe_config->pipe_bpp = lvds_bpp;
 	}
+
 	/*
 	 * We have timings from the BIOS for the panel, put them in
 	 * to the adjusted mode.  The CRTC will be set up for this mode,
@@ -345,138 +295,13 @@ static bool intel_lvds_compute_config(struct intel_encoder *intel_encoder,
 	if (HAS_PCH_SPLIT(dev)) {
 		pipe_config->has_pch_encoder = true;
 
-		intel_pch_panel_fitting(dev,
-					intel_connector->panel.fitting_mode,
-					mode, adjusted_mode);
-		return true;
-	}
-
-	/* Native modes don't need fitting */
-	if (adjusted_mode->hdisplay == mode->hdisplay &&
-	    adjusted_mode->vdisplay == mode->vdisplay)
-		goto out;
-
-	/* 965+ wants fuzzy fitting */
-	if (INTEL_INFO(dev)->gen >= 4)
-		pfit_control |= ((intel_crtc->pipe << PFIT_PIPE_SHIFT) |
-				 PFIT_FILTER_FUZZY);
-
-	/*
-	 * Enable automatic panel scaling for non-native modes so that they fill
-	 * the screen.  Should be enabled before the pipe is enabled, according
-	 * to register description and PRM.
-	 * Change the value here to see the borders for debugging
-	 */
-	for_each_pipe(pipe)
-		I915_WRITE(BCLRPAT(pipe), 0);
-
-	drm_mode_set_crtcinfo(adjusted_mode, 0);
-	pipe_config->timings_set = true;
-
-	switch (intel_connector->panel.fitting_mode) {
-	case DRM_MODE_SCALE_CENTER:
-		/*
-		 * For centered modes, we have to calculate border widths &
-		 * heights and modify the values programmed into the CRTC.
-		 */
-		centre_horizontally(adjusted_mode, mode->hdisplay);
-		centre_vertically(adjusted_mode, mode->vdisplay);
-		border = LVDS_BORDER_ENABLE;
-		break;
-
-	case DRM_MODE_SCALE_ASPECT:
-		/* Scale but preserve the aspect ratio */
-		if (INTEL_INFO(dev)->gen >= 4) {
-			u32 scaled_width = adjusted_mode->hdisplay * mode->vdisplay;
-			u32 scaled_height = mode->hdisplay * adjusted_mode->vdisplay;
-
-			/* 965+ is easy, it does everything in hw */
-			if (scaled_width > scaled_height)
-				pfit_control |= PFIT_ENABLE | PFIT_SCALING_PILLAR;
-			else if (scaled_width < scaled_height)
-				pfit_control |= PFIT_ENABLE | PFIT_SCALING_LETTER;
-			else if (adjusted_mode->hdisplay != mode->hdisplay)
-				pfit_control |= PFIT_ENABLE | PFIT_SCALING_AUTO;
-		} else {
-			u32 scaled_width = adjusted_mode->hdisplay * mode->vdisplay;
-			u32 scaled_height = mode->hdisplay * adjusted_mode->vdisplay;
-			/*
-			 * For earlier chips we have to calculate the scaling
-			 * ratio by hand and program it into the
-			 * PFIT_PGM_RATIO register
-			 */
-			if (scaled_width > scaled_height) { /* pillar */
-				centre_horizontally(adjusted_mode, scaled_height / mode->vdisplay);
-
-				border = LVDS_BORDER_ENABLE;
-				if (mode->vdisplay != adjusted_mode->vdisplay) {
-					u32 bits = panel_fitter_scaling(mode->vdisplay, adjusted_mode->vdisplay);
-					pfit_pgm_ratios |= (bits << PFIT_HORIZ_SCALE_SHIFT |
-							    bits << PFIT_VERT_SCALE_SHIFT);
-					pfit_control |= (PFIT_ENABLE |
-							 VERT_INTERP_BILINEAR |
-							 HORIZ_INTERP_BILINEAR);
-				}
-			} else if (scaled_width < scaled_height) { /* letter */
-				centre_vertically(adjusted_mode, scaled_width / mode->hdisplay);
-
-				border = LVDS_BORDER_ENABLE;
-				if (mode->hdisplay != adjusted_mode->hdisplay) {
-					u32 bits = panel_fitter_scaling(mode->hdisplay, adjusted_mode->hdisplay);
-					pfit_pgm_ratios |= (bits << PFIT_HORIZ_SCALE_SHIFT |
-							    bits << PFIT_VERT_SCALE_SHIFT);
-					pfit_control |= (PFIT_ENABLE |
-							 VERT_INTERP_BILINEAR |
-							 HORIZ_INTERP_BILINEAR);
-				}
-			} else
-				/* Aspects match, Let hw scale both directions */
-				pfit_control |= (PFIT_ENABLE |
-						 VERT_AUTO_SCALE | HORIZ_AUTO_SCALE |
-						 VERT_INTERP_BILINEAR |
-						 HORIZ_INTERP_BILINEAR);
-		}
-		break;
-
-	case DRM_MODE_SCALE_FULLSCREEN:
-		/*
-		 * Full scaling, even if it changes the aspect ratio.
-		 * Fortunately this is all done for us in hw.
-		 */
-		if (mode->vdisplay != adjusted_mode->vdisplay ||
-		    mode->hdisplay != adjusted_mode->hdisplay) {
-			pfit_control |= PFIT_ENABLE;
-			if (INTEL_INFO(dev)->gen >= 4)
-				pfit_control |= PFIT_SCALING_AUTO;
-			else
-				pfit_control |= (VERT_AUTO_SCALE |
-						 VERT_INTERP_BILINEAR |
-						 HORIZ_AUTO_SCALE |
-						 HORIZ_INTERP_BILINEAR);
-		}
-		break;
-
-	default:
-		break;
-	}
-
-out:
-	/* If not enabling scaling, be consistent and always use 0. */
-	if ((pfit_control & PFIT_ENABLE) == 0) {
-		pfit_control = 0;
-		pfit_pgm_ratios = 0;
-	}
-
-	/* Make sure pre-965 set dither correctly */
-	if (INTEL_INFO(dev)->gen < 4 && dev_priv->lvds_dither)
-		pfit_control |= PANEL_8TO6_DITHER_ENABLE;
+		intel_pch_panel_fitting(intel_crtc, pipe_config,
+					intel_connector->panel.fitting_mode);
+	} else {
+		intel_gmch_panel_fitting(intel_crtc, pipe_config,
+					 intel_connector->panel.fitting_mode);
 
-	if (pfit_control != lvds_encoder->pfit_control ||
-	    pfit_pgm_ratios != lvds_encoder->pfit_pgm_ratios) {
-		lvds_encoder->pfit_control = pfit_control;
-		lvds_encoder->pfit_pgm_ratios = pfit_pgm_ratios;
 	}
-	dev_priv->lvds_border_bits = border;
 
 	/*
 	 * XXX: It would be nice to support lower refresh rates on the
@@ -953,11 +778,11 @@ static bool lvds_is_present_in_vbt(struct drm_device *dev,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int i;
 
-	if (!dev_priv->child_dev_num)
+	if (!dev_priv->vbt.child_dev_num)
 		return true;
 
-	for (i = 0; i < dev_priv->child_dev_num; i++) {
-		struct child_device_config *child = dev_priv->child_dev + i;
+	for (i = 0; i < dev_priv->vbt.child_dev_num; i++) {
+		struct child_device_config *child = dev_priv->vbt.child_dev + i;
 
 		/* If the device type is not LFP, continue.
 		 * We have to check both the new identifiers as well as the
@@ -1045,7 +870,7 @@ static bool compute_is_dual_link_lvds(struct intel_lvds_encoder *lvds_encoder)
 	 */
 	val = I915_READ(lvds_encoder->reg);
 	if (!(val & ~(LVDS_PIPE_MASK | LVDS_DETECTED)))
-		val = dev_priv->bios_lvds_val;
+		val = dev_priv->vbt.bios_lvds_val;
 
 	return (val & LVDS_CLKB_POWER_MASK) == LVDS_CLKB_POWER_UP;
 }
@@ -1072,7 +897,7 @@ static bool intel_lvds_supported(struct drm_device *dev)
  * Create the connector, register the LVDS DDC bus, and try to figure out what
  * modes we can display on the LVDS panel (if present).
  */
-bool intel_lvds_init(struct drm_device *dev)
+void intel_lvds_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_lvds_encoder *lvds_encoder;
@@ -1090,43 +915,39 @@ bool intel_lvds_init(struct drm_device *dev)
 	u8 pin;
 
 	if (!intel_lvds_supported(dev))
-		return false;
+		return;
 
 	/* Skip init on machines we know falsely report LVDS */
 	if (dmi_check_system(intel_no_lvds))
-		return false;
+		return;
 
 	pin = GMBUS_PORT_PANEL;
 	if (!lvds_is_present_in_vbt(dev, &pin)) {
 		DRM_DEBUG_KMS("LVDS is not present in VBT\n");
-		return false;
+		return;
 	}
 
 	if (HAS_PCH_SPLIT(dev)) {
 		if ((I915_READ(PCH_LVDS) & LVDS_DETECTED) == 0)
-			return false;
-		if (dev_priv->edp.support) {
+			return;
+		if (dev_priv->vbt.edp_support) {
 			DRM_DEBUG_KMS("disable LVDS for eDP support\n");
-			return false;
+			return;
 		}
 	}
 
 	lvds_encoder = kzalloc(sizeof(struct intel_lvds_encoder), GFP_KERNEL);
 	if (!lvds_encoder)
-		return false;
+		return;
 
 	lvds_connector = kzalloc(sizeof(struct intel_lvds_connector), GFP_KERNEL);
 	if (!lvds_connector) {
 		kfree(lvds_encoder);
-		return false;
+		return;
 	}
 
 	lvds_encoder->attached_connector = lvds_connector;
 
-	if (!HAS_PCH_SPLIT(dev)) {
-		lvds_encoder->pfit_control = I915_READ(PFIT_CONTROL);
-	}
-
 	intel_encoder = &lvds_encoder->base;
 	encoder = &intel_encoder->base;
 	intel_connector = &lvds_connector->base;
@@ -1138,11 +959,11 @@ bool intel_lvds_init(struct drm_device *dev)
 			 DRM_MODE_ENCODER_LVDS);
 
 	intel_encoder->enable = intel_enable_lvds;
-	intel_encoder->pre_enable = intel_pre_enable_lvds;
 	intel_encoder->pre_pll_enable = intel_pre_pll_enable_lvds;
 	intel_encoder->compute_config = intel_lvds_compute_config;
 	intel_encoder->disable = intel_disable_lvds;
 	intel_encoder->get_hw_state = intel_lvds_get_hw_state;
+	intel_encoder->get_config = intel_lvds_get_config;
 	intel_connector->get_hw_state = intel_connector_get_hw_state;
 
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
@@ -1228,11 +1049,11 @@ bool intel_lvds_init(struct drm_device *dev)
 	}
 
 	/* Failed to get EDID, what about VBT? */
-	if (dev_priv->lfp_lvds_vbt_mode) {
+	if (dev_priv->vbt.lfp_lvds_vbt_mode) {
 		DRM_DEBUG_KMS("using mode from VBT: ");
-		drm_mode_debug_printmodeline(dev_priv->lfp_lvds_vbt_mode);
+		drm_mode_debug_printmodeline(dev_priv->vbt.lfp_lvds_vbt_mode);
 
-		fixed_mode = drm_mode_duplicate(dev, dev_priv->lfp_lvds_vbt_mode);
+		fixed_mode = drm_mode_duplicate(dev, dev_priv->vbt.lfp_lvds_vbt_mode);
 		if (fixed_mode) {
 			fixed_mode->type |= DRM_MODE_TYPE_PREFERRED;
 			goto out;
@@ -1293,7 +1114,7 @@ out:
 	intel_panel_init(&intel_connector->panel, fixed_mode);
 	intel_panel_setup_backlight(connector);
 
-	return true;
+	return;
 
 failed:
 	DRM_DEBUG_KMS("No LVDS modes found, disabling.\n");
@@ -1303,5 +1124,5 @@ failed:
 		drm_mode_destroy(dev, fixed_mode);
 	kfree(lvds_encoder);
 	kfree(lvds_connector);
-	return false;
+	return;
 }
diff --git a/drivers/gpu/drm/i915/intel_opregion.c b/drivers/gpu/drm/i915/intel_opregion.c
index a8117e614009..cfb8fb68f09c 100644
--- a/drivers/gpu/drm/i915/intel_opregion.c
+++ b/drivers/gpu/drm/i915/intel_opregion.c
@@ -110,6 +110,10 @@ struct opregion_asle {
 	u8 rsvd[102];
 } __attribute__((packed));
 
+/* Driver readiness indicator */
+#define ASLE_ARDY_READY		(1 << 0)
+#define ASLE_ARDY_NOT_READY	(0 << 0)
+
 /* ASLE irq request bits */
 #define ASLE_SET_ALS_ILLUM     (1 << 0)
 #define ASLE_SET_BACKLIGHT     (1 << 1)
@@ -123,6 +127,12 @@ struct opregion_asle {
 #define ASLE_PFIT_FAILED	(1<<14)
 #define ASLE_PWM_FREQ_FAILED	(1<<16)
 
+/* Technology enabled indicator */
+#define ASLE_TCHE_ALS_EN	(1 << 0)
+#define ASLE_TCHE_BLC_EN	(1 << 1)
+#define ASLE_TCHE_PFIT_EN	(1 << 2)
+#define ASLE_TCHE_PFMB_EN	(1 << 3)
+
 /* ASLE backlight brightness to set */
 #define ASLE_BCLP_VALID                (1<<31)
 #define ASLE_BCLP_MSK          (~(1<<31))
@@ -152,7 +162,6 @@ static u32 asle_set_backlight(struct drm_device *dev, u32 bclp)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct opregion_asle __iomem *asle = dev_priv->opregion.asle;
-	u32 max;
 
 	DRM_DEBUG_DRIVER("bclp = 0x%08x\n", bclp);
 
@@ -163,8 +172,7 @@ static u32 asle_set_backlight(struct drm_device *dev, u32 bclp)
 	if (bclp > 255)
 		return ASLE_BACKLIGHT_FAILED;
 
-	max = intel_panel_get_max_backlight(dev);
-	intel_panel_set_backlight(dev, bclp * max / 255);
+	intel_panel_set_backlight(dev, bclp, 255);
 	iowrite32((bclp*0x64)/0xff | ASLE_CBLV_VALID, &asle->cblv);
 
 	return 0;
@@ -174,29 +182,22 @@ static u32 asle_set_als_illum(struct drm_device *dev, u32 alsi)
 {
 	/* alsi is the current ALS reading in lux. 0 indicates below sensor
 	   range, 0xffff indicates above sensor range. 1-0xfffe are valid */
-	return 0;
+	DRM_DEBUG_DRIVER("Illum is not supported\n");
+	return ASLE_ALS_ILLUM_FAILED;
 }
 
 static u32 asle_set_pwm_freq(struct drm_device *dev, u32 pfmb)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	if (pfmb & ASLE_PFMB_PWM_VALID) {
-		u32 blc_pwm_ctl = I915_READ(BLC_PWM_CTL);
-		u32 pwm = pfmb & ASLE_PFMB_PWM_MASK;
-		blc_pwm_ctl &= BACKLIGHT_DUTY_CYCLE_MASK;
-		pwm = pwm >> 9;
-		/* FIXME - what do we do with the PWM? */
-	}
-	return 0;
+	DRM_DEBUG_DRIVER("PWM freq is not supported\n");
+	return ASLE_PWM_FREQ_FAILED;
 }
 
 static u32 asle_set_pfit(struct drm_device *dev, u32 pfit)
 {
 	/* Panel fitting is currently controlled by the X code, so this is a
 	   noop until modesetting support works fully */
-	if (!(pfit & ASLE_PFIT_VALID))
-		return ASLE_PFIT_FAILED;
-	return 0;
+	DRM_DEBUG_DRIVER("Pfit is not supported\n");
+	return ASLE_PFIT_FAILED;
 }
 
 void intel_opregion_asle_intr(struct drm_device *dev)
@@ -231,64 +232,6 @@ void intel_opregion_asle_intr(struct drm_device *dev)
 	iowrite32(asle_stat, &asle->aslc);
 }
 
-void intel_opregion_gse_intr(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct opregion_asle __iomem *asle = dev_priv->opregion.asle;
-	u32 asle_stat = 0;
-	u32 asle_req;
-
-	if (!asle)
-		return;
-
-	asle_req = ioread32(&asle->aslc) & ASLE_REQ_MSK;
-
-	if (!asle_req) {
-		DRM_DEBUG_DRIVER("non asle set request??\n");
-		return;
-	}
-
-	if (asle_req & ASLE_SET_ALS_ILLUM) {
-		DRM_DEBUG_DRIVER("Illum is not supported\n");
-		asle_stat |= ASLE_ALS_ILLUM_FAILED;
-	}
-
-	if (asle_req & ASLE_SET_BACKLIGHT)
-		asle_stat |= asle_set_backlight(dev, ioread32(&asle->bclp));
-
-	if (asle_req & ASLE_SET_PFIT) {
-		DRM_DEBUG_DRIVER("Pfit is not supported\n");
-		asle_stat |= ASLE_PFIT_FAILED;
-	}
-
-	if (asle_req & ASLE_SET_PWM_FREQ) {
-		DRM_DEBUG_DRIVER("PWM freq is not supported\n");
-		asle_stat |= ASLE_PWM_FREQ_FAILED;
-	}
-
-	iowrite32(asle_stat, &asle->aslc);
-}
-#define ASLE_ALS_EN    (1<<0)
-#define ASLE_BLC_EN    (1<<1)
-#define ASLE_PFIT_EN   (1<<2)
-#define ASLE_PFMB_EN   (1<<3)
-
-void intel_opregion_enable_asle(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct opregion_asle __iomem *asle = dev_priv->opregion.asle;
-
-	if (asle) {
-		if (IS_MOBILE(dev))
-			intel_enable_asle(dev);
-
-		iowrite32(ASLE_ALS_EN | ASLE_BLC_EN | ASLE_PFIT_EN |
-			  ASLE_PFMB_EN,
-			  &asle->tche);
-		iowrite32(1, &asle->ardy);
-	}
-}
-
 #define ACPI_EV_DISPLAY_SWITCH (1<<0)
 #define ACPI_EV_LID            (1<<1)
 #define ACPI_EV_DOCK           (1<<2)
@@ -368,8 +311,8 @@ static void intel_didl_outputs(struct drm_device *dev)
 
 	list_for_each_entry(acpi_cdev, &acpi_video_bus->children, node) {
 		if (i >= 8) {
-			dev_printk(KERN_ERR, &dev->pdev->dev,
-				    "More than 8 outputs detected\n");
+			dev_dbg(&dev->pdev->dev,
+				"More than 8 outputs detected via ACPI\n");
 			return;
 		}
 		status =
@@ -395,8 +338,8 @@ blind_set:
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		int output_type = ACPI_OTHER_OUTPUT;
 		if (i >= 8) {
-			dev_printk(KERN_ERR, &dev->pdev->dev,
-				    "More than 8 outputs detected\n");
+			dev_dbg(&dev->pdev->dev,
+				"More than 8 outputs in connector list\n");
 			return;
 		}
 		switch (connector->connector_type) {
@@ -472,8 +415,10 @@ void intel_opregion_init(struct drm_device *dev)
 		register_acpi_notifier(&intel_opregion_notifier);
 	}
 
-	if (opregion->asle)
-		intel_opregion_enable_asle(dev);
+	if (opregion->asle) {
+		iowrite32(ASLE_TCHE_BLC_EN, &opregion->asle->tche);
+		iowrite32(ASLE_ARDY_READY, &opregion->asle->ardy);
+	}
 }
 
 void intel_opregion_fini(struct drm_device *dev)
@@ -484,6 +429,9 @@ void intel_opregion_fini(struct drm_device *dev)
 	if (!opregion->header)
 		return;
 
+	if (opregion->asle)
+		iowrite32(ASLE_ARDY_NOT_READY, &opregion->asle->ardy);
+
 	if (opregion->acpi) {
 		iowrite32(0, &opregion->acpi->drdy);
 
@@ -546,6 +494,8 @@ int intel_opregion_setup(struct drm_device *dev)
 	if (mboxes & MBOX_ASLE) {
 		DRM_DEBUG_DRIVER("ASLE supported\n");
 		opregion->asle = base + OPREGION_ASLE_OFFSET;
+
+		iowrite32(ASLE_ARDY_NOT_READY, &opregion->asle->ardy);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index 67a2501d519d..a3698812e9c7 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -217,7 +217,7 @@ static int intel_overlay_do_wait_request(struct intel_overlay *overlay,
 	int ret;
 
 	BUG_ON(overlay->last_flip_req);
-	ret = i915_add_request(ring, NULL, &overlay->last_flip_req);
+	ret = i915_add_request(ring, &overlay->last_flip_req);
 	if (ret)
 		return ret;
 
@@ -286,7 +286,7 @@ static int intel_overlay_continue(struct intel_overlay *overlay,
 	intel_ring_emit(ring, flip_addr);
 	intel_ring_advance(ring);
 
-	return i915_add_request(ring, NULL, &overlay->last_flip_req);
+	return i915_add_request(ring, &overlay->last_flip_req);
 }
 
 static void intel_overlay_release_old_vid_tail(struct intel_overlay *overlay)
@@ -1485,14 +1485,15 @@ err:
 }
 
 void
-intel_overlay_print_error_state(struct seq_file *m, struct intel_overlay_error_state *error)
+intel_overlay_print_error_state(struct drm_i915_error_state_buf *m,
+				struct intel_overlay_error_state *error)
 {
-	seq_printf(m, "Overlay, status: 0x%08x, interrupt: 0x%08x\n",
-		   error->dovsta, error->isr);
-	seq_printf(m, "  Register file at 0x%08lx:\n",
-		   error->base);
+	i915_error_printf(m, "Overlay, status: 0x%08x, interrupt: 0x%08x\n",
+			  error->dovsta, error->isr);
+	i915_error_printf(m, "  Register file at 0x%08lx:\n",
+			  error->base);
 
-#define P(x) seq_printf(m, "    " #x ":	0x%08x\n", error->regs.x)
+#define P(x) i915_error_printf(m, "    " #x ":	0x%08x\n", error->regs.x)
 	P(OBUF_0Y);
 	P(OBUF_1Y);
 	P(OBUF_0U);
diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index eb5e6e95f3c7..67e2c1f1c9a8 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -54,14 +54,16 @@ intel_fixed_panel_mode(struct drm_display_mode *fixed_mode,
 
 /* adjusted_mode has been preset to be the panel's fixed mode */
 void
-intel_pch_panel_fitting(struct drm_device *dev,
-			int fitting_mode,
-			const struct drm_display_mode *mode,
-			struct drm_display_mode *adjusted_mode)
+intel_pch_panel_fitting(struct intel_crtc *intel_crtc,
+			struct intel_crtc_config *pipe_config,
+			int fitting_mode)
 {
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_display_mode *mode, *adjusted_mode;
 	int x, y, width, height;
 
+	mode = &pipe_config->requested_mode;
+	adjusted_mode = &pipe_config->adjusted_mode;
+
 	x = y = width = height = 0;
 
 	/* Native modes don't need fitting */
@@ -104,17 +106,212 @@ intel_pch_panel_fitting(struct drm_device *dev,
 		}
 		break;
 
-	default:
 	case DRM_MODE_SCALE_FULLSCREEN:
 		x = y = 0;
 		width = adjusted_mode->hdisplay;
 		height = adjusted_mode->vdisplay;
 		break;
+
+	default:
+		WARN(1, "bad panel fit mode: %d\n", fitting_mode);
+		return;
 	}
 
 done:
-	dev_priv->pch_pf_pos = (x << 16) | y;
-	dev_priv->pch_pf_size = (width << 16) | height;
+	pipe_config->pch_pfit.pos = (x << 16) | y;
+	pipe_config->pch_pfit.size = (width << 16) | height;
+}
+
+static void
+centre_horizontally(struct drm_display_mode *mode,
+		    int width)
+{
+	u32 border, sync_pos, blank_width, sync_width;
+
+	/* keep the hsync and hblank widths constant */
+	sync_width = mode->crtc_hsync_end - mode->crtc_hsync_start;
+	blank_width = mode->crtc_hblank_end - mode->crtc_hblank_start;
+	sync_pos = (blank_width - sync_width + 1) / 2;
+
+	border = (mode->hdisplay - width + 1) / 2;
+	border += border & 1; /* make the border even */
+
+	mode->crtc_hdisplay = width;
+	mode->crtc_hblank_start = width + border;
+	mode->crtc_hblank_end = mode->crtc_hblank_start + blank_width;
+
+	mode->crtc_hsync_start = mode->crtc_hblank_start + sync_pos;
+	mode->crtc_hsync_end = mode->crtc_hsync_start + sync_width;
+}
+
+static void
+centre_vertically(struct drm_display_mode *mode,
+		  int height)
+{
+	u32 border, sync_pos, blank_width, sync_width;
+
+	/* keep the vsync and vblank widths constant */
+	sync_width = mode->crtc_vsync_end - mode->crtc_vsync_start;
+	blank_width = mode->crtc_vblank_end - mode->crtc_vblank_start;
+	sync_pos = (blank_width - sync_width + 1) / 2;
+
+	border = (mode->vdisplay - height + 1) / 2;
+
+	mode->crtc_vdisplay = height;
+	mode->crtc_vblank_start = height + border;
+	mode->crtc_vblank_end = mode->crtc_vblank_start + blank_width;
+
+	mode->crtc_vsync_start = mode->crtc_vblank_start + sync_pos;
+	mode->crtc_vsync_end = mode->crtc_vsync_start + sync_width;
+}
+
+static inline u32 panel_fitter_scaling(u32 source, u32 target)
+{
+	/*
+	 * Floating point operation is not supported. So the FACTOR
+	 * is defined, which can avoid the floating point computation
+	 * when calculating the panel ratio.
+	 */
+#define ACCURACY 12
+#define FACTOR (1 << ACCURACY)
+	u32 ratio = source * FACTOR / target;
+	return (FACTOR * ratio + FACTOR/2) / FACTOR;
+}
+
+void intel_gmch_panel_fitting(struct intel_crtc *intel_crtc,
+			      struct intel_crtc_config *pipe_config,
+			      int fitting_mode)
+{
+	struct drm_device *dev = intel_crtc->base.dev;
+	u32 pfit_control = 0, pfit_pgm_ratios = 0, border = 0;
+	struct drm_display_mode *mode, *adjusted_mode;
+
+	mode = &pipe_config->requested_mode;
+	adjusted_mode = &pipe_config->adjusted_mode;
+
+	/* Native modes don't need fitting */
+	if (adjusted_mode->hdisplay == mode->hdisplay &&
+	    adjusted_mode->vdisplay == mode->vdisplay)
+		goto out;
+
+	drm_mode_set_crtcinfo(adjusted_mode, 0);
+	pipe_config->timings_set = true;
+
+	switch (fitting_mode) {
+	case DRM_MODE_SCALE_CENTER:
+		/*
+		 * For centered modes, we have to calculate border widths &
+		 * heights and modify the values programmed into the CRTC.
+		 */
+		centre_horizontally(adjusted_mode, mode->hdisplay);
+		centre_vertically(adjusted_mode, mode->vdisplay);
+		border = LVDS_BORDER_ENABLE;
+		break;
+	case DRM_MODE_SCALE_ASPECT:
+		/* Scale but preserve the aspect ratio */
+		if (INTEL_INFO(dev)->gen >= 4) {
+			u32 scaled_width = adjusted_mode->hdisplay *
+				mode->vdisplay;
+			u32 scaled_height = mode->hdisplay *
+				adjusted_mode->vdisplay;
+
+			/* 965+ is easy, it does everything in hw */
+			if (scaled_width > scaled_height)
+				pfit_control |= PFIT_ENABLE |
+					PFIT_SCALING_PILLAR;
+			else if (scaled_width < scaled_height)
+				pfit_control |= PFIT_ENABLE |
+					PFIT_SCALING_LETTER;
+			else if (adjusted_mode->hdisplay != mode->hdisplay)
+				pfit_control |= PFIT_ENABLE | PFIT_SCALING_AUTO;
+		} else {
+			u32 scaled_width = adjusted_mode->hdisplay *
+				mode->vdisplay;
+			u32 scaled_height = mode->hdisplay *
+				adjusted_mode->vdisplay;
+			/*
+			 * For earlier chips we have to calculate the scaling
+			 * ratio by hand and program it into the
+			 * PFIT_PGM_RATIO register
+			 */
+			if (scaled_width > scaled_height) { /* pillar */
+				centre_horizontally(adjusted_mode,
+						    scaled_height /
+						    mode->vdisplay);
+
+				border = LVDS_BORDER_ENABLE;
+				if (mode->vdisplay != adjusted_mode->vdisplay) {
+					u32 bits = panel_fitter_scaling(mode->vdisplay, adjusted_mode->vdisplay);
+					pfit_pgm_ratios |= (bits << PFIT_HORIZ_SCALE_SHIFT |
+							    bits << PFIT_VERT_SCALE_SHIFT);
+					pfit_control |= (PFIT_ENABLE |
+							 VERT_INTERP_BILINEAR |
+							 HORIZ_INTERP_BILINEAR);
+				}
+			} else if (scaled_width < scaled_height) { /* letter */
+				centre_vertically(adjusted_mode,
+						  scaled_width /
+						  mode->hdisplay);
+
+				border = LVDS_BORDER_ENABLE;
+				if (mode->hdisplay != adjusted_mode->hdisplay) {
+					u32 bits = panel_fitter_scaling(mode->hdisplay, adjusted_mode->hdisplay);
+					pfit_pgm_ratios |= (bits << PFIT_HORIZ_SCALE_SHIFT |
+							    bits << PFIT_VERT_SCALE_SHIFT);
+					pfit_control |= (PFIT_ENABLE |
+							 VERT_INTERP_BILINEAR |
+							 HORIZ_INTERP_BILINEAR);
+				}
+			} else {
+				/* Aspects match, Let hw scale both directions */
+				pfit_control |= (PFIT_ENABLE |
+						 VERT_AUTO_SCALE | HORIZ_AUTO_SCALE |
+						 VERT_INTERP_BILINEAR |
+						 HORIZ_INTERP_BILINEAR);
+			}
+		}
+		break;
+	case DRM_MODE_SCALE_FULLSCREEN:
+		/*
+		 * Full scaling, even if it changes the aspect ratio.
+		 * Fortunately this is all done for us in hw.
+		 */
+		if (mode->vdisplay != adjusted_mode->vdisplay ||
+		    mode->hdisplay != adjusted_mode->hdisplay) {
+			pfit_control |= PFIT_ENABLE;
+			if (INTEL_INFO(dev)->gen >= 4)
+				pfit_control |= PFIT_SCALING_AUTO;
+			else
+				pfit_control |= (VERT_AUTO_SCALE |
+						 VERT_INTERP_BILINEAR |
+						 HORIZ_AUTO_SCALE |
+						 HORIZ_INTERP_BILINEAR);
+		}
+		break;
+	default:
+		WARN(1, "bad panel fit mode: %d\n", fitting_mode);
+		return;
+	}
+
+	/* 965+ wants fuzzy fitting */
+	/* FIXME: handle multiple panels by failing gracefully */
+	if (INTEL_INFO(dev)->gen >= 4)
+		pfit_control |= ((intel_crtc->pipe << PFIT_PIPE_SHIFT) |
+				 PFIT_FILTER_FUZZY);
+
+out:
+	if ((pfit_control & PFIT_ENABLE) == 0) {
+		pfit_control = 0;
+		pfit_pgm_ratios = 0;
+	}
+
+	/* Make sure pre-965 set dither correctly for 18bpp panels. */
+	if (INTEL_INFO(dev)->gen < 4 && pipe_config->pipe_bpp == 18)
+		pfit_control |= PANEL_8TO6_DITHER_ENABLE;
+
+	pipe_config->gmch_pfit.control = pfit_control;
+	pipe_config->gmch_pfit.pgm_ratios = pfit_pgm_ratios;
+	pipe_config->gmch_pfit.lvds_border_bits = border;
 }
 
 static int is_backlight_combination_mode(struct drm_device *dev)
@@ -130,11 +327,16 @@ static int is_backlight_combination_mode(struct drm_device *dev)
 	return 0;
 }
 
+/* XXX: query mode clock or hardware clock and program max PWM appropriately
+ * when it's 0.
+ */
 static u32 i915_read_blc_pwm_ctl(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 val;
 
+	WARN_ON_SMP(!spin_is_locked(&dev_priv->backlight.lock));
+
 	/* Restore the CTL value if it lost, e.g. GPU reset */
 
 	if (HAS_PCH_SPLIT(dev_priv->dev)) {
@@ -164,7 +366,7 @@ static u32 i915_read_blc_pwm_ctl(struct drm_device *dev)
 	return val;
 }
 
-static u32 _intel_panel_get_max_backlight(struct drm_device *dev)
+static u32 intel_panel_get_max_backlight(struct drm_device *dev)
 {
 	u32 max;
 
@@ -182,23 +384,8 @@ static u32 _intel_panel_get_max_backlight(struct drm_device *dev)
 			max *= 0xff;
 	}
 
-	return max;
-}
-
-u32 intel_panel_get_max_backlight(struct drm_device *dev)
-{
-	u32 max;
-
-	max = _intel_panel_get_max_backlight(dev);
-	if (max == 0) {
-		/* XXX add code here to query mode clock or hardware clock
-		 * and program max PWM appropriately.
-		 */
-		pr_warn_once("fixme: max PWM is zero\n");
-		return 1;
-	}
-
 	DRM_DEBUG_DRIVER("max backlight PWM = %d\n", max);
+
 	return max;
 }
 
@@ -217,8 +404,11 @@ static u32 intel_panel_compute_brightness(struct drm_device *dev, u32 val)
 		return val;
 
 	if (i915_panel_invert_brightness > 0 ||
-	    dev_priv->quirks & QUIRK_INVERT_BRIGHTNESS)
-		return intel_panel_get_max_backlight(dev) - val;
+	    dev_priv->quirks & QUIRK_INVERT_BRIGHTNESS) {
+		u32 max = intel_panel_get_max_backlight(dev);
+		if (max)
+			return max - val;
+	}
 
 	return val;
 }
@@ -227,6 +417,9 @@ static u32 intel_panel_get_backlight(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 val;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
 
 	if (HAS_PCH_SPLIT(dev)) {
 		val = I915_READ(BLC_PWM_CPU_CTL) & BACKLIGHT_DUTY_CYCLE_MASK;
@@ -244,6 +437,9 @@ static u32 intel_panel_get_backlight(struct drm_device *dev)
 	}
 
 	val = intel_panel_compute_brightness(dev, val);
+
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
+
 	DRM_DEBUG_DRIVER("get backlight PWM = %d\n", val);
 	return val;
 }
@@ -270,6 +466,10 @@ static void intel_panel_actually_set_backlight(struct drm_device *dev, u32 level
 		u32 max = intel_panel_get_max_backlight(dev);
 		u8 lbpc;
 
+		/* we're screwed, but keep behaviour backwards compatible */
+		if (!max)
+			max = 1;
+
 		lbpc = level * 0xfe / max + 1;
 		level /= lbpc;
 		pci_write_config_byte(dev->pdev, PCI_LBPC, lbpc);
@@ -282,9 +482,23 @@ static void intel_panel_actually_set_backlight(struct drm_device *dev, u32 level
 	I915_WRITE(BLC_PWM_CTL, tmp | level);
 }
 
-void intel_panel_set_backlight(struct drm_device *dev, u32 level)
+/* set backlight brightness to level in range [0..max] */
+void intel_panel_set_backlight(struct drm_device *dev, u32 level, u32 max)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 freq;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
+
+	freq = intel_panel_get_max_backlight(dev);
+	if (!freq) {
+		/* we are screwed, bail out */
+		goto out;
+	}
+
+	/* scale to hardware */
+	level = level * freq / max;
 
 	dev_priv->backlight.level = level;
 	if (dev_priv->backlight.device)
@@ -292,11 +506,16 @@ void intel_panel_set_backlight(struct drm_device *dev, u32 level)
 
 	if (dev_priv->backlight.enabled)
 		intel_panel_actually_set_backlight(dev, level);
+out:
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
 }
 
 void intel_panel_disable_backlight(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
 
 	dev_priv->backlight.enabled = false;
 	intel_panel_actually_set_backlight(dev, 0);
@@ -314,12 +533,19 @@ void intel_panel_disable_backlight(struct drm_device *dev)
 			I915_WRITE(BLC_PWM_PCH_CTL1, tmp);
 		}
 	}
+
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
 }
 
 void intel_panel_enable_backlight(struct drm_device *dev,
 				  enum pipe pipe)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum transcoder cpu_transcoder =
+		intel_pipe_to_cpu_transcoder(dev_priv, pipe);
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
 
 	if (dev_priv->backlight.level == 0) {
 		dev_priv->backlight.level = intel_panel_get_max_backlight(dev);
@@ -347,14 +573,18 @@ void intel_panel_enable_backlight(struct drm_device *dev,
 		else
 			tmp &= ~BLM_PIPE_SELECT;
 
-		tmp |= BLM_PIPE(pipe);
+		if (cpu_transcoder == TRANSCODER_EDP)
+			tmp |= BLM_TRANSCODER_EDP;
+		else
+			tmp |= BLM_PIPE(cpu_transcoder);
 		tmp &= ~BLM_PWM_ENABLE;
 
 		I915_WRITE(reg, tmp);
 		POSTING_READ(reg);
 		I915_WRITE(reg, tmp | BLM_PWM_ENABLE);
 
-		if (HAS_PCH_SPLIT(dev)) {
+		if (HAS_PCH_SPLIT(dev) &&
+		    !(dev_priv->quirks & QUIRK_NO_PCH_PWM_ENABLE)) {
 			tmp = I915_READ(BLC_PWM_PCH_CTL1);
 			tmp |= BLM_PCH_PWM_ENABLE;
 			tmp &= ~BLM_PCH_OVERRIDE_ENABLE;
@@ -369,6 +599,8 @@ set_level:
 	 */
 	dev_priv->backlight.enabled = true;
 	intel_panel_actually_set_backlight(dev, dev_priv->backlight.level);
+
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
 }
 
 static void intel_panel_init_backlight(struct drm_device *dev)
@@ -405,7 +637,8 @@ intel_panel_detect(struct drm_device *dev)
 static int intel_panel_update_status(struct backlight_device *bd)
 {
 	struct drm_device *dev = bl_get_data(bd);
-	intel_panel_set_backlight(dev, bd->props.brightness);
+	intel_panel_set_backlight(dev, bd->props.brightness,
+				  bd->props.max_brightness);
 	return 0;
 }
 
@@ -425,6 +658,7 @@ int intel_panel_setup_backlight(struct drm_connector *connector)
 	struct drm_device *dev = connector->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct backlight_properties props;
+	unsigned long flags;
 
 	intel_panel_init_backlight(dev);
 
@@ -434,7 +668,11 @@ int intel_panel_setup_backlight(struct drm_connector *connector)
 	memset(&props, 0, sizeof(props));
 	props.type = BACKLIGHT_RAW;
 	props.brightness = dev_priv->backlight.level;
-	props.max_brightness = _intel_panel_get_max_backlight(dev);
+
+	spin_lock_irqsave(&dev_priv->backlight.lock, flags);
+	props.max_brightness = intel_panel_get_max_backlight(dev);
+	spin_unlock_irqrestore(&dev_priv->backlight.lock, flags);
+
 	if (props.max_brightness == 0) {
 		DRM_DEBUG_DRIVER("Failed to get maximum backlight value\n");
 		return -ENODEV;
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index aa01128ff192..6a347f54d39f 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -113,8 +113,8 @@ static void i8xx_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 	fbc_ctl |= obj->fence_reg;
 	I915_WRITE(FBC_CONTROL, fbc_ctl);
 
-	DRM_DEBUG_KMS("enabled FBC, pitch %d, yoff %d, plane %d, ",
-		      cfb_pitch, crtc->y, intel_crtc->plane);
+	DRM_DEBUG_KMS("enabled FBC, pitch %d, yoff %d, plane %c, ",
+		      cfb_pitch, crtc->y, plane_name(intel_crtc->plane));
 }
 
 static bool i8xx_fbc_enabled(struct drm_device *dev)
@@ -148,7 +148,7 @@ static void g4x_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 	/* enable it... */
 	I915_WRITE(DPFC_CONTROL, I915_READ(DPFC_CONTROL) | DPFC_CTL_EN);
 
-	DRM_DEBUG_KMS("enabled fbc on plane %d\n", intel_crtc->plane);
+	DRM_DEBUG_KMS("enabled fbc on plane %c\n", plane_name(intel_crtc->plane));
 }
 
 static void g4x_disable_fbc(struct drm_device *dev)
@@ -228,7 +228,7 @@ static void ironlake_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 		sandybridge_blit_fbc_update(dev);
 	}
 
-	DRM_DEBUG_KMS("enabled fbc on plane %d\n", intel_crtc->plane);
+	DRM_DEBUG_KMS("enabled fbc on plane %c\n", plane_name(intel_crtc->plane));
 }
 
 static void ironlake_disable_fbc(struct drm_device *dev)
@@ -242,6 +242,18 @@ static void ironlake_disable_fbc(struct drm_device *dev)
 		dpfc_ctl &= ~DPFC_CTL_EN;
 		I915_WRITE(ILK_DPFC_CONTROL, dpfc_ctl);
 
+		if (IS_IVYBRIDGE(dev))
+			/* WaFbcDisableDpfcClockGating:ivb */
+			I915_WRITE(ILK_DSPCLK_GATE_D,
+				   I915_READ(ILK_DSPCLK_GATE_D) &
+				   ~ILK_DPFCUNIT_CLOCK_GATE_DISABLE);
+
+		if (IS_HASWELL(dev))
+			/* WaFbcDisableDpfcClockGating:hsw */
+			I915_WRITE(HSW_CLKGATE_DISABLE_PART_1,
+				   I915_READ(HSW_CLKGATE_DISABLE_PART_1) &
+				   ~HSW_DPFC_GATING_DISABLE);
+
 		DRM_DEBUG_KMS("disabled FBC\n");
 	}
 }
@@ -253,6 +265,47 @@ static bool ironlake_fbc_enabled(struct drm_device *dev)
 	return I915_READ(ILK_DPFC_CONTROL) & DPFC_CTL_EN;
 }
 
+static void gen7_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_framebuffer *fb = crtc->fb;
+	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
+	struct drm_i915_gem_object *obj = intel_fb->obj;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+	I915_WRITE(IVB_FBC_RT_BASE, obj->gtt_offset);
+
+	I915_WRITE(ILK_DPFC_CONTROL, DPFC_CTL_EN | DPFC_CTL_LIMIT_1X |
+		   IVB_DPFC_CTL_FENCE_EN |
+		   intel_crtc->plane << IVB_DPFC_CTL_PLANE_SHIFT);
+
+	if (IS_IVYBRIDGE(dev)) {
+		/* WaFbcAsynchFlipDisableFbcQueue:ivb */
+		I915_WRITE(ILK_DISPLAY_CHICKEN1, ILK_FBCQ_DIS);
+		/* WaFbcDisableDpfcClockGating:ivb */
+		I915_WRITE(ILK_DSPCLK_GATE_D,
+			   I915_READ(ILK_DSPCLK_GATE_D) |
+			   ILK_DPFCUNIT_CLOCK_GATE_DISABLE);
+	} else {
+		/* WaFbcAsynchFlipDisableFbcQueue:hsw */
+		I915_WRITE(HSW_PIPE_SLICE_CHICKEN_1(intel_crtc->pipe),
+			   HSW_BYPASS_FBC_QUEUE);
+		/* WaFbcDisableDpfcClockGating:hsw */
+		I915_WRITE(HSW_CLKGATE_DISABLE_PART_1,
+			   I915_READ(HSW_CLKGATE_DISABLE_PART_1) |
+			   HSW_DPFC_GATING_DISABLE);
+	}
+
+	I915_WRITE(SNB_DPFC_CTL_SA,
+		   SNB_CPU_FENCE_ENABLE | obj->fence_reg);
+	I915_WRITE(DPFC_CPU_FENCE_OFFSET, crtc->y);
+
+	sandybridge_blit_fbc_update(dev);
+
+	DRM_DEBUG_KMS("enabled fbc on plane %d\n", intel_crtc->plane);
+}
+
 bool intel_fbc_enabled(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -378,7 +431,7 @@ void intel_disable_fbc(struct drm_device *dev)
  *   - no pixel mulitply/line duplication
  *   - no alpha buffer discard
  *   - no dual wide
- *   - framebuffer <= 2048 in width, 1536 in height
+ *   - framebuffer <= max_hdisplay in width, max_vdisplay in height
  *
  * We can't assume that any compression will take place (worst case),
  * so the compressed buffer has to be the same size as the uncompressed
@@ -396,6 +449,7 @@ void intel_update_fbc(struct drm_device *dev)
 	struct intel_framebuffer *intel_fb;
 	struct drm_i915_gem_object *obj;
 	int enable_fbc;
+	unsigned int max_hdisplay, max_vdisplay;
 
 	if (!i915_powersave)
 		return;
@@ -439,7 +493,7 @@ void intel_update_fbc(struct drm_device *dev)
 	if (enable_fbc < 0) {
 		DRM_DEBUG_KMS("fbc set to per-chip default\n");
 		enable_fbc = 1;
-		if (INTEL_INFO(dev)->gen <= 6)
+		if (INTEL_INFO(dev)->gen <= 7 && !IS_HASWELL(dev))
 			enable_fbc = 0;
 	}
 	if (!enable_fbc) {
@@ -454,13 +508,22 @@ void intel_update_fbc(struct drm_device *dev)
 		dev_priv->no_fbc_reason = FBC_UNSUPPORTED_MODE;
 		goto out_disable;
 	}
-	if ((crtc->mode.hdisplay > 2048) ||
-	    (crtc->mode.vdisplay > 1536)) {
+
+	if (IS_G4X(dev) || INTEL_INFO(dev)->gen >= 5) {
+		max_hdisplay = 4096;
+		max_vdisplay = 2048;
+	} else {
+		max_hdisplay = 2048;
+		max_vdisplay = 1536;
+	}
+	if ((crtc->mode.hdisplay > max_hdisplay) ||
+	    (crtc->mode.vdisplay > max_vdisplay)) {
 		DRM_DEBUG_KMS("mode too large for compression, disabling\n");
 		dev_priv->no_fbc_reason = FBC_MODE_TOO_LARGE;
 		goto out_disable;
 	}
-	if ((IS_I915GM(dev) || IS_I945GM(dev)) && intel_crtc->plane != 0) {
+	if ((IS_I915GM(dev) || IS_I945GM(dev) || IS_HASWELL(dev)) &&
+	    intel_crtc->plane != 0) {
 		DRM_DEBUG_KMS("plane not 0, disabling compression\n");
 		dev_priv->no_fbc_reason = FBC_BAD_PLANE;
 		goto out_disable;
@@ -481,8 +544,6 @@ void intel_update_fbc(struct drm_device *dev)
 		goto out_disable;
 
 	if (i915_gem_stolen_setup_compression(dev, intel_fb->obj->base.size)) {
-		DRM_INFO("not enough stolen space for compressed buffer (need %zd bytes), disabling\n", intel_fb->obj->base.size);
-		DRM_INFO("hint: you may be able to increase stolen memory size in the BIOS to avoid this\n");
 		DRM_DEBUG_KMS("framebuffer too large, disabling compression\n");
 		dev_priv->no_fbc_reason = FBC_STOLEN_TOO_SMALL;
 		goto out_disable;
@@ -1633,6 +1694,10 @@ static bool ironlake_check_srwm(struct drm_device *dev, int level,
 		I915_WRITE(DISP_ARB_CTL,
 			   I915_READ(DISP_ARB_CTL) | DISP_FBC_WM_DIS);
 		return false;
+	} else if (INTEL_INFO(dev)->gen >= 6) {
+		/* enable FBC WM (except on ILK, where it must remain off) */
+		I915_WRITE(DISP_ARB_CTL,
+			   I915_READ(DISP_ARB_CTL) & ~DISP_FBC_WM_DIS);
 	}
 
 	if (display_wm > display->max_wm) {
@@ -2016,31 +2081,558 @@ static void ivybridge_update_wm(struct drm_device *dev)
 		   cursor_wm);
 }
 
-static void
-haswell_update_linetime_wm(struct drm_device *dev, int pipe,
-				 struct drm_display_mode *mode)
+static uint32_t hsw_wm_get_pixel_rate(struct drm_device *dev,
+				      struct drm_crtc *crtc)
+{
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	uint32_t pixel_rate, pfit_size;
+
+	pixel_rate = intel_crtc->config.adjusted_mode.clock;
+
+	/* We only use IF-ID interlacing. If we ever use PF-ID we'll need to
+	 * adjust the pixel_rate here. */
+
+	pfit_size = intel_crtc->config.pch_pfit.size;
+	if (pfit_size) {
+		uint64_t pipe_w, pipe_h, pfit_w, pfit_h;
+
+		pipe_w = intel_crtc->config.requested_mode.hdisplay;
+		pipe_h = intel_crtc->config.requested_mode.vdisplay;
+		pfit_w = (pfit_size >> 16) & 0xFFFF;
+		pfit_h = pfit_size & 0xFFFF;
+		if (pipe_w < pfit_w)
+			pipe_w = pfit_w;
+		if (pipe_h < pfit_h)
+			pipe_h = pfit_h;
+
+		pixel_rate = div_u64((uint64_t) pixel_rate * pipe_w * pipe_h,
+				     pfit_w * pfit_h);
+	}
+
+	return pixel_rate;
+}
+
+static uint32_t hsw_wm_method1(uint32_t pixel_rate, uint8_t bytes_per_pixel,
+			       uint32_t latency)
+{
+	uint64_t ret;
+
+	ret = (uint64_t) pixel_rate * bytes_per_pixel * latency;
+	ret = DIV_ROUND_UP_ULL(ret, 64 * 10000) + 2;
+
+	return ret;
+}
+
+static uint32_t hsw_wm_method2(uint32_t pixel_rate, uint32_t pipe_htotal,
+			       uint32_t horiz_pixels, uint8_t bytes_per_pixel,
+			       uint32_t latency)
+{
+	uint32_t ret;
+
+	ret = (latency * pixel_rate) / (pipe_htotal * 10000);
+	ret = (ret + 1) * horiz_pixels * bytes_per_pixel;
+	ret = DIV_ROUND_UP(ret, 64) + 2;
+	return ret;
+}
+
+static uint32_t hsw_wm_fbc(uint32_t pri_val, uint32_t horiz_pixels,
+			   uint8_t bytes_per_pixel)
+{
+	return DIV_ROUND_UP(pri_val * 64, horiz_pixels * bytes_per_pixel) + 2;
+}
+
+struct hsw_pipe_wm_parameters {
+	bool active;
+	bool sprite_enabled;
+	uint8_t pri_bytes_per_pixel;
+	uint8_t spr_bytes_per_pixel;
+	uint8_t cur_bytes_per_pixel;
+	uint32_t pri_horiz_pixels;
+	uint32_t spr_horiz_pixels;
+	uint32_t cur_horiz_pixels;
+	uint32_t pipe_htotal;
+	uint32_t pixel_rate;
+};
+
+struct hsw_wm_maximums {
+	uint16_t pri;
+	uint16_t spr;
+	uint16_t cur;
+	uint16_t fbc;
+};
+
+struct hsw_lp_wm_result {
+	bool enable;
+	bool fbc_enable;
+	uint32_t pri_val;
+	uint32_t spr_val;
+	uint32_t cur_val;
+	uint32_t fbc_val;
+};
+
+struct hsw_wm_values {
+	uint32_t wm_pipe[3];
+	uint32_t wm_lp[3];
+	uint32_t wm_lp_spr[3];
+	uint32_t wm_linetime[3];
+	bool enable_fbc_wm;
+};
+
+enum hsw_data_buf_partitioning {
+	HSW_DATA_BUF_PART_1_2,
+	HSW_DATA_BUF_PART_5_6,
+};
+
+/* For both WM_PIPE and WM_LP. */
+static uint32_t hsw_compute_pri_wm(struct hsw_pipe_wm_parameters *params,
+				   uint32_t mem_value,
+				   bool is_lp)
+{
+	uint32_t method1, method2;
+
+	/* TODO: for now, assume the primary plane is always enabled. */
+	if (!params->active)
+		return 0;
+
+	method1 = hsw_wm_method1(params->pixel_rate,
+				 params->pri_bytes_per_pixel,
+				 mem_value);
+
+	if (!is_lp)
+		return method1;
+
+	method2 = hsw_wm_method2(params->pixel_rate,
+				 params->pipe_htotal,
+				 params->pri_horiz_pixels,
+				 params->pri_bytes_per_pixel,
+				 mem_value);
+
+	return min(method1, method2);
+}
+
+/* For both WM_PIPE and WM_LP. */
+static uint32_t hsw_compute_spr_wm(struct hsw_pipe_wm_parameters *params,
+				   uint32_t mem_value)
+{
+	uint32_t method1, method2;
+
+	if (!params->active || !params->sprite_enabled)
+		return 0;
+
+	method1 = hsw_wm_method1(params->pixel_rate,
+				 params->spr_bytes_per_pixel,
+				 mem_value);
+	method2 = hsw_wm_method2(params->pixel_rate,
+				 params->pipe_htotal,
+				 params->spr_horiz_pixels,
+				 params->spr_bytes_per_pixel,
+				 mem_value);
+	return min(method1, method2);
+}
+
+/* For both WM_PIPE and WM_LP. */
+static uint32_t hsw_compute_cur_wm(struct hsw_pipe_wm_parameters *params,
+				   uint32_t mem_value)
+{
+	if (!params->active)
+		return 0;
+
+	return hsw_wm_method2(params->pixel_rate,
+			      params->pipe_htotal,
+			      params->cur_horiz_pixels,
+			      params->cur_bytes_per_pixel,
+			      mem_value);
+}
+
+/* Only for WM_LP. */
+static uint32_t hsw_compute_fbc_wm(struct hsw_pipe_wm_parameters *params,
+				   uint32_t pri_val,
+				   uint32_t mem_value)
+{
+	if (!params->active)
+		return 0;
+
+	return hsw_wm_fbc(pri_val,
+			  params->pri_horiz_pixels,
+			  params->pri_bytes_per_pixel);
+}
+
+static bool hsw_compute_lp_wm(uint32_t mem_value, struct hsw_wm_maximums *max,
+			      struct hsw_pipe_wm_parameters *params,
+			      struct hsw_lp_wm_result *result)
+{
+	enum pipe pipe;
+	uint32_t pri_val[3], spr_val[3], cur_val[3], fbc_val[3];
+
+	for (pipe = PIPE_A; pipe <= PIPE_C; pipe++) {
+		struct hsw_pipe_wm_parameters *p = &params[pipe];
+
+		pri_val[pipe] = hsw_compute_pri_wm(p, mem_value, true);
+		spr_val[pipe] = hsw_compute_spr_wm(p, mem_value);
+		cur_val[pipe] = hsw_compute_cur_wm(p, mem_value);
+		fbc_val[pipe] = hsw_compute_fbc_wm(p, pri_val[pipe], mem_value);
+	}
+
+	result->pri_val = max3(pri_val[0], pri_val[1], pri_val[2]);
+	result->spr_val = max3(spr_val[0], spr_val[1], spr_val[2]);
+	result->cur_val = max3(cur_val[0], cur_val[1], cur_val[2]);
+	result->fbc_val = max3(fbc_val[0], fbc_val[1], fbc_val[2]);
+
+	if (result->fbc_val > max->fbc) {
+		result->fbc_enable = false;
+		result->fbc_val = 0;
+	} else {
+		result->fbc_enable = true;
+	}
+
+	result->enable = result->pri_val <= max->pri &&
+			 result->spr_val <= max->spr &&
+			 result->cur_val <= max->cur;
+	return result->enable;
+}
+
+static uint32_t hsw_compute_wm_pipe(struct drm_i915_private *dev_priv,
+				    uint32_t mem_value, enum pipe pipe,
+				    struct hsw_pipe_wm_parameters *params)
+{
+	uint32_t pri_val, cur_val, spr_val;
+
+	pri_val = hsw_compute_pri_wm(params, mem_value, false);
+	spr_val = hsw_compute_spr_wm(params, mem_value);
+	cur_val = hsw_compute_cur_wm(params, mem_value);
+
+	WARN(pri_val > 127,
+	     "Primary WM error, mode not supported for pipe %c\n",
+	     pipe_name(pipe));
+	WARN(spr_val > 127,
+	     "Sprite WM error, mode not supported for pipe %c\n",
+	     pipe_name(pipe));
+	WARN(cur_val > 63,
+	     "Cursor WM error, mode not supported for pipe %c\n",
+	     pipe_name(pipe));
+
+	return (pri_val << WM0_PIPE_PLANE_SHIFT) |
+	       (spr_val << WM0_PIPE_SPRITE_SHIFT) |
+	       cur_val;
+}
+
+static uint32_t
+hsw_compute_linetime_wm(struct drm_device *dev, struct drm_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 temp;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct drm_display_mode *mode = &intel_crtc->config.adjusted_mode;
+	u32 linetime, ips_linetime;
 
-	temp = I915_READ(PIPE_WM_LINETIME(pipe));
-	temp &= ~PIPE_WM_LINETIME_MASK;
+	if (!intel_crtc_active(crtc))
+		return 0;
 
 	/* The WM are computed with base on how long it takes to fill a single
 	 * row at the given clock rate, multiplied by 8.
 	 * */
-	temp |= PIPE_WM_LINETIME_TIME(
-		((mode->crtc_hdisplay * 1000) / mode->clock) * 8);
+	linetime = DIV_ROUND_CLOSEST(mode->htotal * 1000 * 8, mode->clock);
+	ips_linetime = DIV_ROUND_CLOSEST(mode->htotal * 1000 * 8,
+					 intel_ddi_get_cdclk_freq(dev_priv));
 
-	/* IPS watermarks are only used by pipe A, and are ignored by
-	 * pipes B and C.  They are calculated similarly to the common
-	 * linetime values, except that we are using CD clock frequency
-	 * in MHz instead of pixel rate for the division.
-	 *
-	 * This is a placeholder for the IPS watermark calculation code.
-	 */
+	return PIPE_WM_LINETIME_IPS_LINETIME(ips_linetime) |
+	       PIPE_WM_LINETIME_TIME(linetime);
+}
+
+static void hsw_compute_wm_parameters(struct drm_device *dev,
+				      struct hsw_pipe_wm_parameters *params,
+				      uint32_t *wm,
+				      struct hsw_wm_maximums *lp_max_1_2,
+				      struct hsw_wm_maximums *lp_max_5_6)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc;
+	struct drm_plane *plane;
+	uint64_t sskpd = I915_READ64(MCH_SSKPD);
+	enum pipe pipe;
+	int pipes_active = 0, sprites_enabled = 0;
 
-	I915_WRITE(PIPE_WM_LINETIME(pipe), temp);
+	if ((sskpd >> 56) & 0xFF)
+		wm[0] = (sskpd >> 56) & 0xFF;
+	else
+		wm[0] = sskpd & 0xF;
+	wm[1] = ((sskpd >> 4) & 0xFF) * 5;
+	wm[2] = ((sskpd >> 12) & 0xFF) * 5;
+	wm[3] = ((sskpd >> 20) & 0x1FF) * 5;
+	wm[4] = ((sskpd >> 32) & 0x1FF) * 5;
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+		struct hsw_pipe_wm_parameters *p;
+
+		pipe = intel_crtc->pipe;
+		p = &params[pipe];
+
+		p->active = intel_crtc_active(crtc);
+		if (!p->active)
+			continue;
+
+		pipes_active++;
+
+		p->pipe_htotal = intel_crtc->config.adjusted_mode.htotal;
+		p->pixel_rate = hsw_wm_get_pixel_rate(dev, crtc);
+		p->pri_bytes_per_pixel = crtc->fb->bits_per_pixel / 8;
+		p->cur_bytes_per_pixel = 4;
+		p->pri_horiz_pixels =
+			intel_crtc->config.requested_mode.hdisplay;
+		p->cur_horiz_pixels = 64;
+	}
+
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
+		struct intel_plane *intel_plane = to_intel_plane(plane);
+		struct hsw_pipe_wm_parameters *p;
+
+		pipe = intel_plane->pipe;
+		p = &params[pipe];
+
+		p->sprite_enabled = intel_plane->wm.enable;
+		p->spr_bytes_per_pixel = intel_plane->wm.bytes_per_pixel;
+		p->spr_horiz_pixels = intel_plane->wm.horiz_pixels;
+
+		if (p->sprite_enabled)
+			sprites_enabled++;
+	}
+
+	if (pipes_active > 1) {
+		lp_max_1_2->pri = lp_max_5_6->pri = sprites_enabled ? 128 : 256;
+		lp_max_1_2->spr = lp_max_5_6->spr = 128;
+		lp_max_1_2->cur = lp_max_5_6->cur = 64;
+	} else {
+		lp_max_1_2->pri = sprites_enabled ? 384 : 768;
+		lp_max_5_6->pri = sprites_enabled ? 128 : 768;
+		lp_max_1_2->spr = 384;
+		lp_max_5_6->spr = 640;
+		lp_max_1_2->cur = lp_max_5_6->cur = 255;
+	}
+	lp_max_1_2->fbc = lp_max_5_6->fbc = 15;
+}
+
+static void hsw_compute_wm_results(struct drm_device *dev,
+				   struct hsw_pipe_wm_parameters *params,
+				   uint32_t *wm,
+				   struct hsw_wm_maximums *lp_maximums,
+				   struct hsw_wm_values *results)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_crtc *crtc;
+	struct hsw_lp_wm_result lp_results[4] = {};
+	enum pipe pipe;
+	int level, max_level, wm_lp;
+
+	for (level = 1; level <= 4; level++)
+		if (!hsw_compute_lp_wm(wm[level], lp_maximums, params,
+				       &lp_results[level - 1]))
+			break;
+	max_level = level - 1;
+
+	/* The spec says it is preferred to disable FBC WMs instead of disabling
+	 * a WM level. */
+	results->enable_fbc_wm = true;
+	for (level = 1; level <= max_level; level++) {
+		if (!lp_results[level - 1].fbc_enable) {
+			results->enable_fbc_wm = false;
+			break;
+		}
+	}
+
+	memset(results, 0, sizeof(*results));
+	for (wm_lp = 1; wm_lp <= 3; wm_lp++) {
+		const struct hsw_lp_wm_result *r;
+
+		level = (max_level == 4 && wm_lp > 1) ? wm_lp + 1 : wm_lp;
+		if (level > max_level)
+			break;
+
+		r = &lp_results[level - 1];
+		results->wm_lp[wm_lp - 1] = HSW_WM_LP_VAL(level * 2,
+							  r->fbc_val,
+							  r->pri_val,
+							  r->cur_val);
+		results->wm_lp_spr[wm_lp - 1] = r->spr_val;
+	}
+
+	for_each_pipe(pipe)
+		results->wm_pipe[pipe] = hsw_compute_wm_pipe(dev_priv, wm[0],
+							     pipe,
+							     &params[pipe]);
+
+	for_each_pipe(pipe) {
+		crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+		results->wm_linetime[pipe] = hsw_compute_linetime_wm(dev, crtc);
+	}
+}
+
+/* Find the result with the highest level enabled. Check for enable_fbc_wm in
+ * case both are at the same level. Prefer r1 in case they're the same. */
+struct hsw_wm_values *hsw_find_best_result(struct hsw_wm_values *r1,
+					   struct hsw_wm_values *r2)
+{
+	int i, val_r1 = 0, val_r2 = 0;
+
+	for (i = 0; i < 3; i++) {
+		if (r1->wm_lp[i] & WM3_LP_EN)
+			val_r1 = r1->wm_lp[i] & WM1_LP_LATENCY_MASK;
+		if (r2->wm_lp[i] & WM3_LP_EN)
+			val_r2 = r2->wm_lp[i] & WM1_LP_LATENCY_MASK;
+	}
+
+	if (val_r1 == val_r2) {
+		if (r2->enable_fbc_wm && !r1->enable_fbc_wm)
+			return r2;
+		else
+			return r1;
+	} else if (val_r1 > val_r2) {
+		return r1;
+	} else {
+		return r2;
+	}
+}
+
+/*
+ * The spec says we shouldn't write when we don't need, because every write
+ * causes WMs to be re-evaluated, expending some power.
+ */
+static void hsw_write_wm_values(struct drm_i915_private *dev_priv,
+				struct hsw_wm_values *results,
+				enum hsw_data_buf_partitioning partitioning)
+{
+	struct hsw_wm_values previous;
+	uint32_t val;
+	enum hsw_data_buf_partitioning prev_partitioning;
+	bool prev_enable_fbc_wm;
+
+	previous.wm_pipe[0] = I915_READ(WM0_PIPEA_ILK);
+	previous.wm_pipe[1] = I915_READ(WM0_PIPEB_ILK);
+	previous.wm_pipe[2] = I915_READ(WM0_PIPEC_IVB);
+	previous.wm_lp[0] = I915_READ(WM1_LP_ILK);
+	previous.wm_lp[1] = I915_READ(WM2_LP_ILK);
+	previous.wm_lp[2] = I915_READ(WM3_LP_ILK);
+	previous.wm_lp_spr[0] = I915_READ(WM1S_LP_ILK);
+	previous.wm_lp_spr[1] = I915_READ(WM2S_LP_IVB);
+	previous.wm_lp_spr[2] = I915_READ(WM3S_LP_IVB);
+	previous.wm_linetime[0] = I915_READ(PIPE_WM_LINETIME(PIPE_A));
+	previous.wm_linetime[1] = I915_READ(PIPE_WM_LINETIME(PIPE_B));
+	previous.wm_linetime[2] = I915_READ(PIPE_WM_LINETIME(PIPE_C));
+
+	prev_partitioning = (I915_READ(WM_MISC) & WM_MISC_DATA_PARTITION_5_6) ?
+			    HSW_DATA_BUF_PART_5_6 : HSW_DATA_BUF_PART_1_2;
+
+	prev_enable_fbc_wm = !(I915_READ(DISP_ARB_CTL) & DISP_FBC_WM_DIS);
+
+	if (memcmp(results->wm_pipe, previous.wm_pipe,
+		   sizeof(results->wm_pipe)) == 0 &&
+	    memcmp(results->wm_lp, previous.wm_lp,
+		   sizeof(results->wm_lp)) == 0 &&
+	    memcmp(results->wm_lp_spr, previous.wm_lp_spr,
+		   sizeof(results->wm_lp_spr)) == 0 &&
+	    memcmp(results->wm_linetime, previous.wm_linetime,
+		   sizeof(results->wm_linetime)) == 0 &&
+	    partitioning == prev_partitioning &&
+	    results->enable_fbc_wm == prev_enable_fbc_wm)
+		return;
+
+	if (previous.wm_lp[2] != 0)
+		I915_WRITE(WM3_LP_ILK, 0);
+	if (previous.wm_lp[1] != 0)
+		I915_WRITE(WM2_LP_ILK, 0);
+	if (previous.wm_lp[0] != 0)
+		I915_WRITE(WM1_LP_ILK, 0);
+
+	if (previous.wm_pipe[0] != results->wm_pipe[0])
+		I915_WRITE(WM0_PIPEA_ILK, results->wm_pipe[0]);
+	if (previous.wm_pipe[1] != results->wm_pipe[1])
+		I915_WRITE(WM0_PIPEB_ILK, results->wm_pipe[1]);
+	if (previous.wm_pipe[2] != results->wm_pipe[2])
+		I915_WRITE(WM0_PIPEC_IVB, results->wm_pipe[2]);
+
+	if (previous.wm_linetime[0] != results->wm_linetime[0])
+		I915_WRITE(PIPE_WM_LINETIME(PIPE_A), results->wm_linetime[0]);
+	if (previous.wm_linetime[1] != results->wm_linetime[1])
+		I915_WRITE(PIPE_WM_LINETIME(PIPE_B), results->wm_linetime[1]);
+	if (previous.wm_linetime[2] != results->wm_linetime[2])
+		I915_WRITE(PIPE_WM_LINETIME(PIPE_C), results->wm_linetime[2]);
+
+	if (prev_partitioning != partitioning) {
+		val = I915_READ(WM_MISC);
+		if (partitioning == HSW_DATA_BUF_PART_1_2)
+			val &= ~WM_MISC_DATA_PARTITION_5_6;
+		else
+			val |= WM_MISC_DATA_PARTITION_5_6;
+		I915_WRITE(WM_MISC, val);
+	}
+
+	if (prev_enable_fbc_wm != results->enable_fbc_wm) {
+		val = I915_READ(DISP_ARB_CTL);
+		if (results->enable_fbc_wm)
+			val &= ~DISP_FBC_WM_DIS;
+		else
+			val |= DISP_FBC_WM_DIS;
+		I915_WRITE(DISP_ARB_CTL, val);
+	}
+
+	if (previous.wm_lp_spr[0] != results->wm_lp_spr[0])
+		I915_WRITE(WM1S_LP_ILK, results->wm_lp_spr[0]);
+	if (previous.wm_lp_spr[1] != results->wm_lp_spr[1])
+		I915_WRITE(WM2S_LP_IVB, results->wm_lp_spr[1]);
+	if (previous.wm_lp_spr[2] != results->wm_lp_spr[2])
+		I915_WRITE(WM3S_LP_IVB, results->wm_lp_spr[2]);
+
+	if (results->wm_lp[0] != 0)
+		I915_WRITE(WM1_LP_ILK, results->wm_lp[0]);
+	if (results->wm_lp[1] != 0)
+		I915_WRITE(WM2_LP_ILK, results->wm_lp[1]);
+	if (results->wm_lp[2] != 0)
+		I915_WRITE(WM3_LP_ILK, results->wm_lp[2]);
+}
+
+static void haswell_update_wm(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct hsw_wm_maximums lp_max_1_2, lp_max_5_6;
+	struct hsw_pipe_wm_parameters params[3];
+	struct hsw_wm_values results_1_2, results_5_6, *best_results;
+	uint32_t wm[5];
+	enum hsw_data_buf_partitioning partitioning;
+
+	hsw_compute_wm_parameters(dev, params, wm, &lp_max_1_2, &lp_max_5_6);
+
+	hsw_compute_wm_results(dev, params, wm, &lp_max_1_2, &results_1_2);
+	if (lp_max_1_2.pri != lp_max_5_6.pri) {
+		hsw_compute_wm_results(dev, params, wm, &lp_max_5_6,
+				       &results_5_6);
+		best_results = hsw_find_best_result(&results_1_2, &results_5_6);
+	} else {
+		best_results = &results_1_2;
+	}
+
+	partitioning = (best_results == &results_1_2) ?
+		       HSW_DATA_BUF_PART_1_2 : HSW_DATA_BUF_PART_5_6;
+
+	hsw_write_wm_values(dev_priv, best_results, partitioning);
+}
+
+static void haswell_update_sprite_wm(struct drm_device *dev, int pipe,
+				     uint32_t sprite_width, int pixel_size,
+				     bool enable)
+{
+	struct drm_plane *plane;
+
+	list_for_each_entry(plane, &dev->mode_config.plane_list, head) {
+		struct intel_plane *intel_plane = to_intel_plane(plane);
+
+		if (intel_plane->pipe == pipe) {
+			intel_plane->wm.enable = enable;
+			intel_plane->wm.horiz_pixels = sprite_width + 1;
+			intel_plane->wm.bytes_per_pixel = pixel_size;
+			break;
+		}
+	}
+
+	haswell_update_wm(dev);
 }
 
 static bool
@@ -2120,7 +2712,8 @@ sandybridge_compute_sprite_srwm(struct drm_device *dev, int plane,
 }
 
 static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
-					 uint32_t sprite_width, int pixel_size)
+					 uint32_t sprite_width, int pixel_size,
+					 bool enable)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int latency = SNB_READ_WM0_LATENCY() * 100;	/* In unit 0.1us */
@@ -2128,6 +2721,9 @@ static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
 	int sprite_wm, reg;
 	int ret;
 
+	if (!enable)
+		return;
+
 	switch (pipe) {
 	case 0:
 		reg = WM0_PIPEA_ILK;
@@ -2146,15 +2742,15 @@ static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
 					    &sandybridge_display_wm_info,
 					    latency, &sprite_wm);
 	if (!ret) {
-		DRM_DEBUG_KMS("failed to compute sprite wm for pipe %d\n",
-			      pipe);
+		DRM_DEBUG_KMS("failed to compute sprite wm for pipe %c\n",
+			      pipe_name(pipe));
 		return;
 	}
 
 	val = I915_READ(reg);
 	val &= ~WM0_PIPE_SPRITE_MASK;
 	I915_WRITE(reg, val | (sprite_wm << WM0_PIPE_SPRITE_SHIFT));
-	DRM_DEBUG_KMS("sprite watermarks For pipe %d - %d\n", pipe, sprite_wm);
+	DRM_DEBUG_KMS("sprite watermarks For pipe %c - %d\n", pipe_name(pipe), sprite_wm);
 
 
 	ret = sandybridge_compute_sprite_srwm(dev, pipe, sprite_width,
@@ -2163,8 +2759,8 @@ static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
 					      SNB_READ_WM1_LATENCY() * 500,
 					      &sprite_wm);
 	if (!ret) {
-		DRM_DEBUG_KMS("failed to compute sprite lp1 wm on pipe %d\n",
-			      pipe);
+		DRM_DEBUG_KMS("failed to compute sprite lp1 wm on pipe %c\n",
+			      pipe_name(pipe));
 		return;
 	}
 	I915_WRITE(WM1S_LP_ILK, sprite_wm);
@@ -2179,8 +2775,8 @@ static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
 					      SNB_READ_WM2_LATENCY() * 500,
 					      &sprite_wm);
 	if (!ret) {
-		DRM_DEBUG_KMS("failed to compute sprite lp2 wm on pipe %d\n",
-			      pipe);
+		DRM_DEBUG_KMS("failed to compute sprite lp2 wm on pipe %c\n",
+			      pipe_name(pipe));
 		return;
 	}
 	I915_WRITE(WM2S_LP_IVB, sprite_wm);
@@ -2191,8 +2787,8 @@ static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
 					      SNB_READ_WM3_LATENCY() * 500,
 					      &sprite_wm);
 	if (!ret) {
-		DRM_DEBUG_KMS("failed to compute sprite lp3 wm on pipe %d\n",
-			      pipe);
+		DRM_DEBUG_KMS("failed to compute sprite lp3 wm on pipe %c\n",
+			      pipe_name(pipe));
 		return;
 	}
 	I915_WRITE(WM3S_LP_IVB, sprite_wm);
@@ -2238,23 +2834,15 @@ void intel_update_watermarks(struct drm_device *dev)
 		dev_priv->display.update_wm(dev);
 }
 
-void intel_update_linetime_watermarks(struct drm_device *dev,
-		int pipe, struct drm_display_mode *mode)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (dev_priv->display.update_linetime_wm)
-		dev_priv->display.update_linetime_wm(dev, pipe, mode);
-}
-
 void intel_update_sprite_watermarks(struct drm_device *dev, int pipe,
-				    uint32_t sprite_width, int pixel_size)
+				    uint32_t sprite_width, int pixel_size,
+				    bool enable)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
 	if (dev_priv->display.update_sprite_wm)
 		dev_priv->display.update_sprite_wm(dev, pipe, sprite_width,
-						   pixel_size);
+						   pixel_size, enable);
 }
 
 static struct drm_i915_gem_object *
@@ -2481,6 +3069,67 @@ void gen6_set_rps(struct drm_device *dev, u8 val)
 	trace_intel_gpu_freq_change(val * 50);
 }
 
+/*
+ * Wait until the previous freq change has completed,
+ * or the timeout elapsed, and then update our notion
+ * of the current GPU frequency.
+ */
+static void vlv_update_rps_cur_delay(struct drm_i915_private *dev_priv)
+{
+	unsigned long timeout = jiffies + msecs_to_jiffies(10);
+	u32 pval;
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+	do {
+		pval = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
+		if (time_after(jiffies, timeout)) {
+			DRM_DEBUG_DRIVER("timed out waiting for Punit\n");
+			break;
+		}
+		udelay(10);
+	} while (pval & 1);
+
+	pval >>= 8;
+
+	if (pval != dev_priv->rps.cur_delay)
+		DRM_DEBUG_DRIVER("Punit overrode GPU freq: %d MHz (%u) requested, but got %d Mhz (%u)\n",
+				 vlv_gpu_freq(dev_priv->mem_freq, dev_priv->rps.cur_delay),
+				 dev_priv->rps.cur_delay,
+				 vlv_gpu_freq(dev_priv->mem_freq, pval), pval);
+
+	dev_priv->rps.cur_delay = pval;
+}
+
+void valleyview_set_rps(struct drm_device *dev, u8 val)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	gen6_rps_limits(dev_priv, &val);
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+	WARN_ON(val > dev_priv->rps.max_delay);
+	WARN_ON(val < dev_priv->rps.min_delay);
+
+	vlv_update_rps_cur_delay(dev_priv);
+
+	DRM_DEBUG_DRIVER("GPU freq request from %d MHz (%u) to %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv->mem_freq,
+				      dev_priv->rps.cur_delay),
+			 dev_priv->rps.cur_delay,
+			 vlv_gpu_freq(dev_priv->mem_freq, val), val);
+
+	if (val == dev_priv->rps.cur_delay)
+		return;
+
+	vlv_punit_write(dev_priv, PUNIT_REG_GPU_FREQ_REQ, val);
+
+	dev_priv->rps.cur_delay = val;
+
+	trace_intel_gpu_freq_change(vlv_gpu_freq(dev_priv->mem_freq, val));
+}
+
+
 static void gen6_disable_rps(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -2488,6 +3137,25 @@ static void gen6_disable_rps(struct drm_device *dev)
 	I915_WRITE(GEN6_RC_CONTROL, 0);
 	I915_WRITE(GEN6_RPNSWREQ, 1 << 31);
 	I915_WRITE(GEN6_PMINTRMSK, 0xffffffff);
+	I915_WRITE(GEN6_PMIER, I915_READ(GEN6_PMIER) & ~GEN6_PM_RPS_EVENTS);
+	/* Complete PM interrupt masking here doesn't race with the rps work
+	 * item again unmasking PM interrupts because that is using a different
+	 * register (PMIMR) to mask PM interrupts. The only risk is in leaving
+	 * stale bits in PMIIR and PMIMR which gen6_enable_rps will clean up. */
+
+	spin_lock_irq(&dev_priv->rps.lock);
+	dev_priv->rps.pm_iir = 0;
+	spin_unlock_irq(&dev_priv->rps.lock);
+
+	I915_WRITE(GEN6_PMIIR, GEN6_PM_RPS_EVENTS);
+}
+
+static void valleyview_disable_rps(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	I915_WRITE(GEN6_RC_CONTROL, 0);
+	I915_WRITE(GEN6_PMINTRMSK, 0xffffffff);
 	I915_WRITE(GEN6_PMIER, 0);
 	/* Complete PM interrupt masking here doesn't race with the rps work
 	 * item again unmasking PM interrupts because that is using a different
@@ -2499,6 +3167,11 @@ static void gen6_disable_rps(struct drm_device *dev)
 	spin_unlock_irq(&dev_priv->rps.lock);
 
 	I915_WRITE(GEN6_PMIIR, I915_READ(GEN6_PMIIR));
+
+	if (dev_priv->vlv_pctx) {
+		drm_gem_object_unreference(&dev_priv->vlv_pctx->base);
+		dev_priv->vlv_pctx = NULL;
+	}
 }
 
 int intel_enable_rc6(const struct drm_device *dev)
@@ -2655,12 +3328,15 @@ static void gen6_enable_rps(struct drm_device *dev)
 	gen6_set_rps(dev_priv->dev, (gt_perf_status & 0xff00) >> 8);
 
 	/* requires MSI enabled */
-	I915_WRITE(GEN6_PMIER, GEN6_PM_DEFERRED_EVENTS);
+	I915_WRITE(GEN6_PMIER, I915_READ(GEN6_PMIER) | GEN6_PM_RPS_EVENTS);
 	spin_lock_irq(&dev_priv->rps.lock);
-	WARN_ON(dev_priv->rps.pm_iir != 0);
-	I915_WRITE(GEN6_PMIMR, 0);
+	/* FIXME: Our interrupt enabling sequence is bonghits.
+	 * dev_priv->rps.pm_iir really should be 0 here. */
+	dev_priv->rps.pm_iir = 0;
+	I915_WRITE(GEN6_PMIMR, I915_READ(GEN6_PMIMR) & ~GEN6_PM_RPS_EVENTS);
+	I915_WRITE(GEN6_PMIIR, GEN6_PM_RPS_EVENTS);
 	spin_unlock_irq(&dev_priv->rps.lock);
-	/* enable all PM interrupts */
+	/* unmask all PM interrupts */
 	I915_WRITE(GEN6_PMINTRMSK, 0);
 
 	rc6vids = 0;
@@ -2742,6 +3418,207 @@ static void gen6_update_ring_freq(struct drm_device *dev)
 	}
 }
 
+int valleyview_rps_max_freq(struct drm_i915_private *dev_priv)
+{
+	u32 val, rp0;
+
+	val = vlv_nc_read(dev_priv, IOSF_NC_FB_GFX_FREQ_FUSE);
+
+	rp0 = (val & FB_GFX_MAX_FREQ_FUSE_MASK) >> FB_GFX_MAX_FREQ_FUSE_SHIFT;
+	/* Clamp to max */
+	rp0 = min_t(u32, rp0, 0xea);
+
+	return rp0;
+}
+
+static int valleyview_rps_rpe_freq(struct drm_i915_private *dev_priv)
+{
+	u32 val, rpe;
+
+	val = vlv_nc_read(dev_priv, IOSF_NC_FB_GFX_FMAX_FUSE_LO);
+	rpe = (val & FB_FMAX_VMIN_FREQ_LO_MASK) >> FB_FMAX_VMIN_FREQ_LO_SHIFT;
+	val = vlv_nc_read(dev_priv, IOSF_NC_FB_GFX_FMAX_FUSE_HI);
+	rpe |= (val & FB_FMAX_VMIN_FREQ_HI_MASK) << 5;
+
+	return rpe;
+}
+
+int valleyview_rps_min_freq(struct drm_i915_private *dev_priv)
+{
+	return vlv_punit_read(dev_priv, PUNIT_REG_GPU_LFM) & 0xff;
+}
+
+static void vlv_rps_timer_work(struct work_struct *work)
+{
+	drm_i915_private_t *dev_priv = container_of(work, drm_i915_private_t,
+						    rps.vlv_work.work);
+
+	/*
+	 * Timer fired, we must be idle.  Drop to min voltage state.
+	 * Note: we use RPe here since it should match the
+	 * Vmin we were shooting for.  That should give us better
+	 * perf when we come back out of RC6 than if we used the
+	 * min freq available.
+	 */
+	mutex_lock(&dev_priv->rps.hw_lock);
+	if (dev_priv->rps.cur_delay > dev_priv->rps.rpe_delay)
+		valleyview_set_rps(dev_priv->dev, dev_priv->rps.rpe_delay);
+	mutex_unlock(&dev_priv->rps.hw_lock);
+}
+
+static void valleyview_setup_pctx(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_gem_object *pctx;
+	unsigned long pctx_paddr;
+	u32 pcbr;
+	int pctx_size = 24*1024;
+
+	pcbr = I915_READ(VLV_PCBR);
+	if (pcbr) {
+		/* BIOS set it up already, grab the pre-alloc'd space */
+		int pcbr_offset;
+
+		pcbr_offset = (pcbr & (~4095)) - dev_priv->mm.stolen_base;
+		pctx = i915_gem_object_create_stolen_for_preallocated(dev_priv->dev,
+								      pcbr_offset,
+								      -1,
+								      pctx_size);
+		goto out;
+	}
+
+	/*
+	 * From the Gunit register HAS:
+	 * The Gfx driver is expected to program this register and ensure
+	 * proper allocation within Gfx stolen memory.  For example, this
+	 * register should be programmed such than the PCBR range does not
+	 * overlap with other ranges, such as the frame buffer, protected
+	 * memory, or any other relevant ranges.
+	 */
+	pctx = i915_gem_object_create_stolen(dev, pctx_size);
+	if (!pctx) {
+		DRM_DEBUG("not enough stolen space for PCTX, disabling\n");
+		return;
+	}
+
+	pctx_paddr = dev_priv->mm.stolen_base + pctx->stolen->start;
+	I915_WRITE(VLV_PCBR, pctx_paddr);
+
+out:
+	dev_priv->vlv_pctx = pctx;
+}
+
+static void valleyview_enable_rps(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_ring_buffer *ring;
+	u32 gtfifodbg, val;
+	int i;
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+	if ((gtfifodbg = I915_READ(GTFIFODBG))) {
+		DRM_ERROR("GT fifo had a previous error %x\n", gtfifodbg);
+		I915_WRITE(GTFIFODBG, gtfifodbg);
+	}
+
+	valleyview_setup_pctx(dev);
+
+	gen6_gt_force_wake_get(dev_priv);
+
+	I915_WRITE(GEN6_RP_UP_THRESHOLD, 59400);
+	I915_WRITE(GEN6_RP_DOWN_THRESHOLD, 245000);
+	I915_WRITE(GEN6_RP_UP_EI, 66000);
+	I915_WRITE(GEN6_RP_DOWN_EI, 350000);
+
+	I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10);
+
+	I915_WRITE(GEN6_RP_CONTROL,
+		   GEN6_RP_MEDIA_TURBO |
+		   GEN6_RP_MEDIA_HW_NORMAL_MODE |
+		   GEN6_RP_MEDIA_IS_GFX |
+		   GEN6_RP_ENABLE |
+		   GEN6_RP_UP_BUSY_AVG |
+		   GEN6_RP_DOWN_IDLE_CONT);
+
+	I915_WRITE(GEN6_RC6_WAKE_RATE_LIMIT, 0x00280000);
+	I915_WRITE(GEN6_RC_EVALUATION_INTERVAL, 125000);
+	I915_WRITE(GEN6_RC_IDLE_HYSTERSIS, 25);
+
+	for_each_ring(ring, dev_priv, i)
+		I915_WRITE(RING_MAX_IDLE(ring->mmio_base), 10);
+
+	I915_WRITE(GEN6_RC6_THRESHOLD, 0xc350);
+
+	/* allows RC6 residency counter to work */
+	I915_WRITE(0x138104, _MASKED_BIT_ENABLE(0x3));
+	I915_WRITE(GEN6_RC_CONTROL,
+		   GEN7_RC_CTL_TO_MODE);
+
+	val = vlv_punit_read(dev_priv, PUNIT_REG_GPU_FREQ_STS);
+	switch ((val >> 6) & 3) {
+	case 0:
+	case 1:
+		dev_priv->mem_freq = 800;
+		break;
+	case 2:
+		dev_priv->mem_freq = 1066;
+		break;
+	case 3:
+		dev_priv->mem_freq = 1333;
+		break;
+	}
+	DRM_DEBUG_DRIVER("DDR speed: %d MHz", dev_priv->mem_freq);
+
+	DRM_DEBUG_DRIVER("GPLL enabled? %s\n", val & 0x10 ? "yes" : "no");
+	DRM_DEBUG_DRIVER("GPU status: 0x%08x\n", val);
+
+	dev_priv->rps.cur_delay = (val >> 8) & 0xff;
+	DRM_DEBUG_DRIVER("current GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv->mem_freq,
+				      dev_priv->rps.cur_delay),
+			 dev_priv->rps.cur_delay);
+
+	dev_priv->rps.max_delay = valleyview_rps_max_freq(dev_priv);
+	dev_priv->rps.hw_max = dev_priv->rps.max_delay;
+	DRM_DEBUG_DRIVER("max GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv->mem_freq,
+				      dev_priv->rps.max_delay),
+			 dev_priv->rps.max_delay);
+
+	dev_priv->rps.rpe_delay = valleyview_rps_rpe_freq(dev_priv);
+	DRM_DEBUG_DRIVER("RPe GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv->mem_freq,
+				      dev_priv->rps.rpe_delay),
+			 dev_priv->rps.rpe_delay);
+
+	dev_priv->rps.min_delay = valleyview_rps_min_freq(dev_priv);
+	DRM_DEBUG_DRIVER("min GPU freq: %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv->mem_freq,
+				      dev_priv->rps.min_delay),
+			 dev_priv->rps.min_delay);
+
+	DRM_DEBUG_DRIVER("setting GPU freq to %d MHz (%u)\n",
+			 vlv_gpu_freq(dev_priv->mem_freq,
+				      dev_priv->rps.rpe_delay),
+			 dev_priv->rps.rpe_delay);
+
+	INIT_DELAYED_WORK(&dev_priv->rps.vlv_work, vlv_rps_timer_work);
+
+	valleyview_set_rps(dev_priv->dev, dev_priv->rps.rpe_delay);
+
+	/* requires MSI enabled */
+	I915_WRITE(GEN6_PMIER, GEN6_PM_RPS_EVENTS);
+	spin_lock_irq(&dev_priv->rps.lock);
+	WARN_ON(dev_priv->rps.pm_iir != 0);
+	I915_WRITE(GEN6_PMIMR, 0);
+	spin_unlock_irq(&dev_priv->rps.lock);
+	/* enable all PM interrupts */
+	I915_WRITE(GEN6_PMINTRMSK, 0);
+
+	gen6_gt_force_wake_put(dev_priv);
+}
+
 void ironlake_teardown_rc6(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -3465,13 +4342,22 @@ void intel_disable_gt_powersave(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	/* Interrupts should be disabled already to avoid re-arming. */
+	WARN_ON(dev->irq_enabled);
+
 	if (IS_IRONLAKE_M(dev)) {
 		ironlake_disable_drps(dev);
 		ironlake_disable_rc6(dev);
-	} else if (INTEL_INFO(dev)->gen >= 6 && !IS_VALLEYVIEW(dev)) {
+	} else if (INTEL_INFO(dev)->gen >= 6) {
 		cancel_delayed_work_sync(&dev_priv->rps.delayed_resume_work);
+		cancel_work_sync(&dev_priv->rps.work);
+		if (IS_VALLEYVIEW(dev))
+			cancel_delayed_work_sync(&dev_priv->rps.vlv_work);
 		mutex_lock(&dev_priv->rps.hw_lock);
-		gen6_disable_rps(dev);
+		if (IS_VALLEYVIEW(dev))
+			valleyview_disable_rps(dev);
+		else
+			gen6_disable_rps(dev);
 		mutex_unlock(&dev_priv->rps.hw_lock);
 	}
 }
@@ -3484,8 +4370,13 @@ static void intel_gen6_powersave_work(struct work_struct *work)
 	struct drm_device *dev = dev_priv->dev;
 
 	mutex_lock(&dev_priv->rps.hw_lock);
-	gen6_enable_rps(dev);
-	gen6_update_ring_freq(dev);
+
+	if (IS_VALLEYVIEW(dev)) {
+		valleyview_enable_rps(dev);
+	} else {
+		gen6_enable_rps(dev);
+		gen6_update_ring_freq(dev);
+	}
 	mutex_unlock(&dev_priv->rps.hw_lock);
 }
 
@@ -3497,7 +4388,7 @@ void intel_enable_gt_powersave(struct drm_device *dev)
 		ironlake_enable_drps(dev);
 		ironlake_enable_rc6(dev);
 		intel_init_emon(dev);
-	} else if ((IS_GEN6(dev) || IS_GEN7(dev)) && !IS_VALLEYVIEW(dev)) {
+	} else if (IS_GEN6(dev) || IS_GEN7(dev)) {
 		/*
 		 * PCU communication is slow and this doesn't need to be
 		 * done at any specific time, so do this out of our fast path
@@ -3520,6 +4411,19 @@ static void ibx_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(SOUTH_DSPCLK_GATE_D, PCH_DPLSUNIT_CLOCK_GATE_DISABLE);
 }
 
+static void g4x_disable_trickle_feed(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int pipe;
+
+	for_each_pipe(pipe) {
+		I915_WRITE(DSPCNTR(pipe),
+			   I915_READ(DSPCNTR(pipe)) |
+			   DISPPLANE_TRICKLE_FEED_DISABLE);
+		intel_flush_display_plane(dev_priv, pipe);
+	}
+}
+
 static void ironlake_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -3579,10 +4483,12 @@ static void ironlake_init_clock_gating(struct drm_device *dev)
 		   _3D_CHICKEN2_WM_READ_PIPELINED << 16 |
 		   _3D_CHICKEN2_WM_READ_PIPELINED);
 
-	/* WaDisableRenderCachePipelinedFlush */
+	/* WaDisableRenderCachePipelinedFlush:ilk */
 	I915_WRITE(CACHE_MODE_0,
 		   _MASKED_BIT_ENABLE(CM0_PIPELINED_RENDER_FLUSH_DISABLE));
 
+	g4x_disable_trickle_feed(dev);
+
 	ibx_init_clock_gating(dev);
 }
 
@@ -3607,7 +4513,7 @@ static void cpt_init_clock_gating(struct drm_device *dev)
 		val = I915_READ(TRANS_CHICKEN2(pipe));
 		val |= TRANS_CHICKEN2_TIMING_OVERRIDE;
 		val &= ~TRANS_CHICKEN2_FDI_POLARITY_REVERSED;
-		if (dev_priv->fdi_rx_polarity_inverted)
+		if (dev_priv->vbt.fdi_rx_polarity_inverted)
 			val |= TRANS_CHICKEN2_FDI_POLARITY_REVERSED;
 		val &= ~TRANS_CHICKEN2_FRAME_START_DELAY_MASK;
 		val &= ~TRANS_CHICKEN2_DISABLE_DEEP_COLOR_COUNTER;
@@ -3637,7 +4543,6 @@ static void gen6_check_mch_setup(struct drm_device *dev)
 static void gen6_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe;
 	uint32_t dspclk_gate = ILK_VRHUNIT_CLOCK_GATE_DISABLE;
 
 	I915_WRITE(ILK_DSPCLK_GATE_D, dspclk_gate);
@@ -3646,11 +4551,11 @@ static void gen6_init_clock_gating(struct drm_device *dev)
 		   I915_READ(ILK_DISPLAY_CHICKEN2) |
 		   ILK_ELPIN_409_SELECT);
 
-	/* WaDisableHiZPlanesWhenMSAAEnabled */
+	/* WaDisableHiZPlanesWhenMSAAEnabled:snb */
 	I915_WRITE(_3D_CHICKEN,
 		   _MASKED_BIT_ENABLE(_3D_CHICKEN_HIZ_PLANE_DISABLE_MSAA_4X_SNB));
 
-	/* WaSetupGtModeTdRowDispatch */
+	/* WaSetupGtModeTdRowDispatch:snb */
 	if (IS_SNB_GT1(dev))
 		I915_WRITE(GEN6_GT_MODE,
 			   _MASKED_BIT_ENABLE(GEN6_TD_FOUR_ROW_DISPATCH_DISABLE));
@@ -3677,8 +4582,8 @@ static void gen6_init_clock_gating(struct drm_device *dev)
 	 * According to the spec, bit 11 (RCCUNIT) must also be set,
 	 * but we didn't debug actual testcases to find it out.
 	 *
-	 * Also apply WaDisableVDSUnitClockGating and
-	 * WaDisableRCPBUnitClockGating.
+	 * Also apply WaDisableVDSUnitClockGating:snb and
+	 * WaDisableRCPBUnitClockGating:snb.
 	 */
 	I915_WRITE(GEN6_UCGCTL2,
 		   GEN7_VDSUNIT_CLOCK_GATE_DISABLE |
@@ -3709,16 +4614,11 @@ static void gen6_init_clock_gating(struct drm_device *dev)
 		   ILK_DPARBUNIT_CLOCK_GATE_ENABLE  |
 		   ILK_DPFDUNIT_CLOCK_GATE_ENABLE);
 
-	/* WaMbcDriverBootEnable */
+	/* WaMbcDriverBootEnable:snb */
 	I915_WRITE(GEN6_MBCTL, I915_READ(GEN6_MBCTL) |
 		   GEN6_MBCTL_ENABLE_BOOT_FETCH);
 
-	for_each_pipe(pipe) {
-		I915_WRITE(DSPCNTR(pipe),
-			   I915_READ(DSPCNTR(pipe)) |
-			   DISPPLANE_TRICKLE_FEED_DISABLE);
-		intel_flush_display_plane(dev_priv, pipe);
-	}
+	g4x_disable_trickle_feed(dev);
 
 	/* The default value should be 0x200 according to docs, but the two
 	 * platforms I checked have a 0 for this. (Maybe BIOS overrides?) */
@@ -3739,7 +4639,6 @@ static void gen7_setup_fixed_func_scheduler(struct drm_i915_private *dev_priv)
 	reg |= GEN7_FF_VS_SCHED_HW;
 	reg |= GEN7_FF_DS_SCHED_HW;
 
-	/* WaVSRefCountFullforceMissDisable */
 	if (IS_HASWELL(dev_priv->dev))
 		reg &= ~GEN7_FF_VS_REF_CNT_FFME;
 
@@ -3758,65 +4657,72 @@ static void lpt_init_clock_gating(struct drm_device *dev)
 		I915_WRITE(SOUTH_DSPCLK_GATE_D,
 			   I915_READ(SOUTH_DSPCLK_GATE_D) |
 			   PCH_LP_PARTITION_LEVEL_DISABLE);
+
+	/* WADPOClockGatingDisable:hsw */
+	I915_WRITE(_TRANSA_CHICKEN1,
+		   I915_READ(_TRANSA_CHICKEN1) |
+		   TRANS_CHICKEN1_DP0UNIT_GC_DISABLE);
+}
+
+static void lpt_suspend_hw(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->pch_id == INTEL_PCH_LPT_LP_DEVICE_ID_TYPE) {
+		uint32_t val = I915_READ(SOUTH_DSPCLK_GATE_D);
+
+		val &= ~PCH_LP_PARTITION_LEVEL_DISABLE;
+		I915_WRITE(SOUTH_DSPCLK_GATE_D, val);
+	}
 }
 
 static void haswell_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe;
 
 	I915_WRITE(WM3_LP_ILK, 0);
 	I915_WRITE(WM2_LP_ILK, 0);
 	I915_WRITE(WM1_LP_ILK, 0);
 
 	/* According to the spec, bit 13 (RCZUNIT) must be set on IVB.
-	 * This implements the WaDisableRCZUnitClockGating workaround.
+	 * This implements the WaDisableRCZUnitClockGating:hsw workaround.
 	 */
 	I915_WRITE(GEN6_UCGCTL2, GEN6_RCZUNIT_CLOCK_GATE_DISABLE);
 
-	/* Apply the WaDisableRHWOOptimizationForRenderHang workaround. */
+	/* Apply the WaDisableRHWOOptimizationForRenderHang:hsw workaround. */
 	I915_WRITE(GEN7_COMMON_SLICE_CHICKEN1,
 		   GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC);
 
-	/* WaApplyL3ControlAndL3ChickenMode requires those two on Ivy Bridge */
+	/* WaApplyL3ControlAndL3ChickenMode:hsw */
 	I915_WRITE(GEN7_L3CNTLREG1,
 			GEN7_WA_FOR_GEN7_L3_CONTROL);
 	I915_WRITE(GEN7_L3_CHICKEN_MODE_REGISTER,
 			GEN7_WA_L3_CHICKEN_MODE);
 
-	/* This is required by WaCatErrorRejectionIssue */
+	/* This is required by WaCatErrorRejectionIssue:hsw */
 	I915_WRITE(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
 			I915_READ(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
 			GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
 
-	for_each_pipe(pipe) {
-		I915_WRITE(DSPCNTR(pipe),
-			   I915_READ(DSPCNTR(pipe)) |
-			   DISPPLANE_TRICKLE_FEED_DISABLE);
-		intel_flush_display_plane(dev_priv, pipe);
-	}
+	g4x_disable_trickle_feed(dev);
 
+	/* WaVSRefCountFullforceMissDisable:hsw */
 	gen7_setup_fixed_func_scheduler(dev_priv);
 
-	/* WaDisable4x2SubspanOptimization */
+	/* WaDisable4x2SubspanOptimization:hsw */
 	I915_WRITE(CACHE_MODE_1,
 		   _MASKED_BIT_ENABLE(PIXEL_SUBSPAN_COLLECT_OPT_DISABLE));
 
-	/* WaMbcDriverBootEnable */
+	/* WaMbcDriverBootEnable:hsw */
 	I915_WRITE(GEN6_MBCTL, I915_READ(GEN6_MBCTL) |
 		   GEN6_MBCTL_ENABLE_BOOT_FETCH);
 
-	/* WaSwitchSolVfFArbitrationPriority */
+	/* WaSwitchSolVfFArbitrationPriority:hsw */
 	I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) | HSW_ECOCHK_ARB_PRIO_SOL);
 
-	/* XXX: This is a workaround for early silicon revisions and should be
-	 * removed later.
-	 */
-	I915_WRITE(WM_DBG,
-			I915_READ(WM_DBG) |
-			WM_DBG_DISALLOW_MULTIPLE_LP |
-			WM_DBG_DISALLOW_SPRITE |
-			WM_DBG_DISALLOW_MAXFIFO);
+	/* WaRsPkgCStateDisplayPMReq:hsw */
+	I915_WRITE(CHICKEN_PAR1_1,
+		   I915_READ(CHICKEN_PAR1_1) | FORCE_ARB_IDLE_PLANES);
 
 	lpt_init_clock_gating(dev);
 }
@@ -3824,7 +4730,6 @@ static void haswell_init_clock_gating(struct drm_device *dev)
 static void ivybridge_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe;
 	uint32_t snpcr;
 
 	I915_WRITE(WM3_LP_ILK, 0);
@@ -3833,16 +4738,16 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 
 	I915_WRITE(ILK_DSPCLK_GATE_D, ILK_VRHUNIT_CLOCK_GATE_DISABLE);
 
-	/* WaDisableEarlyCull */
+	/* WaDisableEarlyCull:ivb */
 	I915_WRITE(_3D_CHICKEN3,
 		   _MASKED_BIT_ENABLE(_3D_CHICKEN_SF_DISABLE_OBJEND_CULL));
 
-	/* WaDisableBackToBackFlipFix */
+	/* WaDisableBackToBackFlipFix:ivb */
 	I915_WRITE(IVB_CHICKEN3,
 		   CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE |
 		   CHICKEN3_DGMG_DONE_FIX_DISABLE);
 
-	/* WaDisablePSDDualDispatchEnable */
+	/* WaDisablePSDDualDispatchEnable:ivb */
 	if (IS_IVB_GT1(dev))
 		I915_WRITE(GEN7_HALF_SLICE_CHICKEN1,
 			   _MASKED_BIT_ENABLE(GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE));
@@ -3850,11 +4755,11 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 		I915_WRITE(GEN7_HALF_SLICE_CHICKEN1_GT2,
 			   _MASKED_BIT_ENABLE(GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE));
 
-	/* Apply the WaDisableRHWOOptimizationForRenderHang workaround. */
+	/* Apply the WaDisableRHWOOptimizationForRenderHang:ivb workaround. */
 	I915_WRITE(GEN7_COMMON_SLICE_CHICKEN1,
 		   GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC);
 
-	/* WaApplyL3ControlAndL3ChickenMode requires those two on Ivy Bridge */
+	/* WaApplyL3ControlAndL3ChickenMode:ivb */
 	I915_WRITE(GEN7_L3CNTLREG1,
 			GEN7_WA_FOR_GEN7_L3_CONTROL);
 	I915_WRITE(GEN7_L3_CHICKEN_MODE_REGISTER,
@@ -3867,7 +4772,7 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 			   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
 
 
-	/* WaForceL3Serialization */
+	/* WaForceL3Serialization:ivb */
 	I915_WRITE(GEN7_L3SQCREG4, I915_READ(GEN7_L3SQCREG4) &
 		   ~L3SQ_URB_READ_CAM_MATCH_DISABLE);
 
@@ -3882,31 +4787,27 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 	 * but we didn't debug actual testcases to find it out.
 	 *
 	 * According to the spec, bit 13 (RCZUNIT) must be set on IVB.
-	 * This implements the WaDisableRCZUnitClockGating workaround.
+	 * This implements the WaDisableRCZUnitClockGating:ivb workaround.
 	 */
 	I915_WRITE(GEN6_UCGCTL2,
 		   GEN6_RCZUNIT_CLOCK_GATE_DISABLE |
 		   GEN6_RCCUNIT_CLOCK_GATE_DISABLE);
 
-	/* This is required by WaCatErrorRejectionIssue */
+	/* This is required by WaCatErrorRejectionIssue:ivb */
 	I915_WRITE(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
 			I915_READ(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
 			GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
 
-	for_each_pipe(pipe) {
-		I915_WRITE(DSPCNTR(pipe),
-			   I915_READ(DSPCNTR(pipe)) |
-			   DISPPLANE_TRICKLE_FEED_DISABLE);
-		intel_flush_display_plane(dev_priv, pipe);
-	}
+	g4x_disable_trickle_feed(dev);
 
-	/* WaMbcDriverBootEnable */
+	/* WaMbcDriverBootEnable:ivb */
 	I915_WRITE(GEN6_MBCTL, I915_READ(GEN6_MBCTL) |
 		   GEN6_MBCTL_ENABLE_BOOT_FETCH);
 
+	/* WaVSRefCountFullforceMissDisable:ivb */
 	gen7_setup_fixed_func_scheduler(dev_priv);
 
-	/* WaDisable4x2SubspanOptimization */
+	/* WaDisable4x2SubspanOptimization:ivb */
 	I915_WRITE(CACHE_MODE_1,
 		   _MASKED_BIT_ENABLE(PIXEL_SUBSPAN_COLLECT_OPT_DISABLE));
 
@@ -3924,54 +4825,45 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 static void valleyview_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int pipe;
-
-	I915_WRITE(WM3_LP_ILK, 0);
-	I915_WRITE(WM2_LP_ILK, 0);
-	I915_WRITE(WM1_LP_ILK, 0);
 
-	I915_WRITE(ILK_DSPCLK_GATE_D, ILK_VRHUNIT_CLOCK_GATE_DISABLE);
+	I915_WRITE(DSPCLK_GATE_D, VRHUNIT_CLOCK_GATE_DISABLE);
 
-	/* WaDisableEarlyCull */
+	/* WaDisableEarlyCull:vlv */
 	I915_WRITE(_3D_CHICKEN3,
 		   _MASKED_BIT_ENABLE(_3D_CHICKEN_SF_DISABLE_OBJEND_CULL));
 
-	/* WaDisableBackToBackFlipFix */
+	/* WaDisableBackToBackFlipFix:vlv */
 	I915_WRITE(IVB_CHICKEN3,
 		   CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE |
 		   CHICKEN3_DGMG_DONE_FIX_DISABLE);
 
-	/* WaDisablePSDDualDispatchEnable */
+	/* WaDisablePSDDualDispatchEnable:vlv */
 	I915_WRITE(GEN7_HALF_SLICE_CHICKEN1,
 		   _MASKED_BIT_ENABLE(GEN7_MAX_PS_THREAD_DEP |
 				      GEN7_PSD_SINGLE_PORT_DISPATCH_ENABLE));
 
-	/* Apply the WaDisableRHWOOptimizationForRenderHang workaround. */
+	/* Apply the WaDisableRHWOOptimizationForRenderHang:vlv workaround. */
 	I915_WRITE(GEN7_COMMON_SLICE_CHICKEN1,
 		   GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC);
 
-	/* WaApplyL3ControlAndL3ChickenMode requires those two on Ivy Bridge */
+	/* WaApplyL3ControlAndL3ChickenMode:vlv */
 	I915_WRITE(GEN7_L3CNTLREG1, I915_READ(GEN7_L3CNTLREG1) | GEN7_L3AGDIS);
 	I915_WRITE(GEN7_L3_CHICKEN_MODE_REGISTER, GEN7_WA_L3_CHICKEN_MODE);
 
-	/* WaForceL3Serialization */
+	/* WaForceL3Serialization:vlv */
 	I915_WRITE(GEN7_L3SQCREG4, I915_READ(GEN7_L3SQCREG4) &
 		   ~L3SQ_URB_READ_CAM_MATCH_DISABLE);
 
-	/* WaDisableDopClockGating */
+	/* WaDisableDopClockGating:vlv */
 	I915_WRITE(GEN7_ROW_CHICKEN2,
 		   _MASKED_BIT_ENABLE(DOP_CLOCK_GATING_DISABLE));
 
-	/* WaForceL3Serialization */
-	I915_WRITE(GEN7_L3SQCREG4, I915_READ(GEN7_L3SQCREG4) &
-		   ~L3SQ_URB_READ_CAM_MATCH_DISABLE);
-
-	/* This is required by WaCatErrorRejectionIssue */
+	/* This is required by WaCatErrorRejectionIssue:vlv */
 	I915_WRITE(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG,
 		   I915_READ(GEN7_SQ_CHICKEN_MBCUNIT_CONFIG) |
 		   GEN7_SQ_CHICKEN_MBCUNIT_SQINTMOB);
 
-	/* WaMbcDriverBootEnable */
+	/* WaMbcDriverBootEnable:vlv */
 	I915_WRITE(GEN6_MBCTL, I915_READ(GEN6_MBCTL) |
 		   GEN6_MBCTL_ENABLE_BOOT_FETCH);
 
@@ -3987,10 +4879,10 @@ static void valleyview_init_clock_gating(struct drm_device *dev)
 	 * but we didn't debug actual testcases to find it out.
 	 *
 	 * According to the spec, bit 13 (RCZUNIT) must be set on IVB.
-	 * This implements the WaDisableRCZUnitClockGating workaround.
+	 * This implements the WaDisableRCZUnitClockGating:vlv workaround.
 	 *
-	 * Also apply WaDisableVDSUnitClockGating and
-	 * WaDisableRCPBUnitClockGating.
+	 * Also apply WaDisableVDSUnitClockGating:vlv and
+	 * WaDisableRCPBUnitClockGating:vlv.
 	 */
 	I915_WRITE(GEN6_UCGCTL2,
 		   GEN7_VDSUNIT_CLOCK_GATE_DISABLE |
@@ -4001,18 +4893,13 @@ static void valleyview_init_clock_gating(struct drm_device *dev)
 
 	I915_WRITE(GEN7_UCGCTL4, GEN7_L3BANK2X_CLOCK_GATE_DISABLE);
 
-	for_each_pipe(pipe) {
-		I915_WRITE(DSPCNTR(pipe),
-			   I915_READ(DSPCNTR(pipe)) |
-			   DISPPLANE_TRICKLE_FEED_DISABLE);
-		intel_flush_display_plane(dev_priv, pipe);
-	}
+	I915_WRITE(MI_ARB_VLV, MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE);
 
 	I915_WRITE(CACHE_MODE_1,
 		   _MASKED_BIT_ENABLE(PIXEL_SUBSPAN_COLLECT_OPT_DISABLE));
 
 	/*
-	 * WaDisableVLVClockGating_VBIIssue
+	 * WaDisableVLVClockGating_VBIIssue:vlv
 	 * Disable clock gating on th GCFG unit to prevent a delay
 	 * in the reporting of vblank events.
 	 */
@@ -4048,6 +4935,8 @@ static void g4x_init_clock_gating(struct drm_device *dev)
 	/* WaDisableRenderCachePipelinedFlush */
 	I915_WRITE(CACHE_MODE_0,
 		   _MASKED_BIT_ENABLE(CM0_PIPELINED_RENDER_FLUSH_DISABLE));
+
+	g4x_disable_trickle_feed(dev);
 }
 
 static void crestline_init_clock_gating(struct drm_device *dev)
@@ -4059,6 +4948,8 @@ static void crestline_init_clock_gating(struct drm_device *dev)
 	I915_WRITE(DSPCLK_GATE_D, 0);
 	I915_WRITE(RAMCLK_GATE_D, 0);
 	I915_WRITE16(DEUC, 0);
+	I915_WRITE(MI_ARB_STATE,
+		   _MASKED_BIT_ENABLE(MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE));
 }
 
 static void broadwater_init_clock_gating(struct drm_device *dev)
@@ -4071,6 +4962,8 @@ static void broadwater_init_clock_gating(struct drm_device *dev)
 		   I965_ISC_CLOCK_GATE_DISABLE |
 		   I965_FBC_CLOCK_GATE_DISABLE);
 	I915_WRITE(RENCLK_GATE_D2, 0);
+	I915_WRITE(MI_ARB_STATE,
+		   _MASKED_BIT_ENABLE(MI_ARB_DISPLAY_TRICKLE_FEED_DISABLE));
 }
 
 static void gen3_init_clock_gating(struct drm_device *dev)
@@ -4110,34 +5003,50 @@ void intel_init_clock_gating(struct drm_device *dev)
 	dev_priv->display.init_clock_gating(dev);
 }
 
+void intel_suspend_hw(struct drm_device *dev)
+{
+	if (HAS_PCH_LPT(dev))
+		lpt_suspend_hw(dev);
+}
+
 /**
  * We should only use the power well if we explicitly asked the hardware to
  * enable it, so check if it's enabled and also check if we've requested it to
  * be enabled.
  */
-bool intel_using_power_well(struct drm_device *dev)
+bool intel_display_power_enabled(struct drm_device *dev,
+				 enum intel_display_power_domain domain)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (IS_HASWELL(dev))
+	if (!HAS_POWER_WELL(dev))
+		return true;
+
+	switch (domain) {
+	case POWER_DOMAIN_PIPE_A:
+	case POWER_DOMAIN_TRANSCODER_EDP:
+		return true;
+	case POWER_DOMAIN_PIPE_B:
+	case POWER_DOMAIN_PIPE_C:
+	case POWER_DOMAIN_PIPE_A_PANEL_FITTER:
+	case POWER_DOMAIN_PIPE_B_PANEL_FITTER:
+	case POWER_DOMAIN_PIPE_C_PANEL_FITTER:
+	case POWER_DOMAIN_TRANSCODER_A:
+	case POWER_DOMAIN_TRANSCODER_B:
+	case POWER_DOMAIN_TRANSCODER_C:
 		return I915_READ(HSW_PWR_WELL_DRIVER) ==
 		       (HSW_PWR_WELL_ENABLE | HSW_PWR_WELL_STATE);
-	else
-		return true;
+	default:
+		BUG();
+	}
 }
 
-void intel_set_power_well(struct drm_device *dev, bool enable)
+static void __intel_set_power_well(struct drm_device *dev, bool enable)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	bool is_enabled, enable_requested;
 	uint32_t tmp;
 
-	if (!HAS_POWER_WELL(dev))
-		return;
-
-	if (!i915_disable_power_well && !enable)
-		return;
-
 	tmp = I915_READ(HSW_PWR_WELL_DRIVER);
 	is_enabled = tmp & HSW_PWR_WELL_STATE;
 	enable_requested = tmp & HSW_PWR_WELL_ENABLE;
@@ -4160,6 +5069,79 @@ void intel_set_power_well(struct drm_device *dev, bool enable)
 	}
 }
 
+static struct i915_power_well *hsw_pwr;
+
+/* Display audio driver power well request */
+void i915_request_power_well(void)
+{
+	if (WARN_ON(!hsw_pwr))
+		return;
+
+	spin_lock_irq(&hsw_pwr->lock);
+	if (!hsw_pwr->count++ &&
+			!hsw_pwr->i915_request)
+		__intel_set_power_well(hsw_pwr->device, true);
+	spin_unlock_irq(&hsw_pwr->lock);
+}
+EXPORT_SYMBOL_GPL(i915_request_power_well);
+
+/* Display audio driver power well release */
+void i915_release_power_well(void)
+{
+	if (WARN_ON(!hsw_pwr))
+		return;
+
+	spin_lock_irq(&hsw_pwr->lock);
+	WARN_ON(!hsw_pwr->count);
+	if (!--hsw_pwr->count &&
+		       !hsw_pwr->i915_request)
+		__intel_set_power_well(hsw_pwr->device, false);
+	spin_unlock_irq(&hsw_pwr->lock);
+}
+EXPORT_SYMBOL_GPL(i915_release_power_well);
+
+int i915_init_power_well(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	hsw_pwr = &dev_priv->power_well;
+
+	hsw_pwr->device = dev;
+	spin_lock_init(&hsw_pwr->lock);
+	hsw_pwr->count = 0;
+
+	return 0;
+}
+
+void i915_remove_power_well(struct drm_device *dev)
+{
+	hsw_pwr = NULL;
+}
+
+void intel_set_power_well(struct drm_device *dev, bool enable)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct i915_power_well *power_well = &dev_priv->power_well;
+
+	if (!HAS_POWER_WELL(dev))
+		return;
+
+	if (!i915_disable_power_well && !enable)
+		return;
+
+	spin_lock_irq(&power_well->lock);
+	power_well->i915_request = enable;
+
+	/* only reject "disable" power well request */
+	if (power_well->count && !enable) {
+		spin_unlock_irq(&power_well->lock);
+		return;
+	}
+
+	__intel_set_power_well(dev, enable);
+	spin_unlock_irq(&power_well->lock);
+}
+
 /*
  * Starting with Haswell, we have a "Power Down Well" that can be turned off
  * when not needed anymore. We have 4 registers that can request the power well
@@ -4190,7 +5172,12 @@ void intel_init_pm(struct drm_device *dev)
 	if (I915_HAS_FBC(dev)) {
 		if (HAS_PCH_SPLIT(dev)) {
 			dev_priv->display.fbc_enabled = ironlake_fbc_enabled;
-			dev_priv->display.enable_fbc = ironlake_enable_fbc;
+			if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev))
+				dev_priv->display.enable_fbc =
+					gen7_enable_fbc;
+			else
+				dev_priv->display.enable_fbc =
+					ironlake_enable_fbc;
 			dev_priv->display.disable_fbc = ironlake_disable_fbc;
 		} else if (IS_GM45(dev)) {
 			dev_priv->display.fbc_enabled = g4x_fbc_enabled;
@@ -4242,10 +5229,10 @@ void intel_init_pm(struct drm_device *dev)
 			}
 			dev_priv->display.init_clock_gating = ivybridge_init_clock_gating;
 		} else if (IS_HASWELL(dev)) {
-			if (SNB_READ_WM0_LATENCY()) {
-				dev_priv->display.update_wm = sandybridge_update_wm;
-				dev_priv->display.update_sprite_wm = sandybridge_update_sprite_wm;
-				dev_priv->display.update_linetime_wm = haswell_update_linetime_wm;
+			if (I915_READ64(MCH_SSKPD)) {
+				dev_priv->display.update_wm = haswell_update_wm;
+				dev_priv->display.update_sprite_wm =
+					haswell_update_sprite_wm;
 			} else {
 				DRM_DEBUG_KMS("Failed to read display plane latency. "
 					      "Disable CxSR\n");
@@ -4340,6 +5327,7 @@ static void __gen6_gt_force_wake_get(struct drm_i915_private *dev_priv)
 			    FORCEWAKE_ACK_TIMEOUT_MS))
 		DRM_ERROR("Timed out waiting for forcewake to ack request.\n");
 
+	/* WaRsForcewakeWaitTC0:snb */
 	__gen6_gt_wait_for_thread_c0(dev_priv);
 }
 
@@ -4371,6 +5359,7 @@ static void __gen6_gt_force_wake_mt_get(struct drm_i915_private *dev_priv)
 			    FORCEWAKE_ACK_TIMEOUT_MS))
 		DRM_ERROR("Timed out waiting for forcewake to ack request.\n");
 
+	/* WaRsForcewakeWaitTC0:ivb,hsw */
 	__gen6_gt_wait_for_thread_c0(dev_priv);
 }
 
@@ -4474,6 +5463,7 @@ static void vlv_force_wake_get(struct drm_i915_private *dev_priv)
 			    FORCEWAKE_ACK_TIMEOUT_MS))
 		DRM_ERROR("Timed out waiting for media to ack forcewake request.\n");
 
+	/* WaRsForcewakeWaitTC0:vlv */
 	__gen6_gt_wait_for_thread_c0(dev_priv);
 }
 
@@ -4486,7 +5476,7 @@ static void vlv_force_wake_put(struct drm_i915_private *dev_priv)
 	gen6_gt_check_fifodbg(dev_priv);
 }
 
-void intel_gt_reset(struct drm_device *dev)
+void intel_gt_sanitize(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
@@ -4497,6 +5487,10 @@ void intel_gt_reset(struct drm_device *dev)
 		if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev))
 			__gen6_gt_force_wake_mt_reset(dev_priv);
 	}
+
+	/* BIOS often leaves RC6 enabled, but disable it for hw init */
+	if (INTEL_INFO(dev)->gen >= 6)
+		intel_disable_gt_powersave(dev);
 }
 
 void intel_gt_init(struct drm_device *dev)
@@ -4505,14 +5499,41 @@ void intel_gt_init(struct drm_device *dev)
 
 	spin_lock_init(&dev_priv->gt_lock);
 
-	intel_gt_reset(dev);
-
 	if (IS_VALLEYVIEW(dev)) {
 		dev_priv->gt.force_wake_get = vlv_force_wake_get;
 		dev_priv->gt.force_wake_put = vlv_force_wake_put;
-	} else if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev)) {
+	} else if (IS_HASWELL(dev)) {
 		dev_priv->gt.force_wake_get = __gen6_gt_force_wake_mt_get;
 		dev_priv->gt.force_wake_put = __gen6_gt_force_wake_mt_put;
+	} else if (IS_IVYBRIDGE(dev)) {
+		u32 ecobus;
+
+		/* IVB configs may use multi-threaded forcewake */
+
+		/* A small trick here - if the bios hasn't configured
+		 * MT forcewake, and if the device is in RC6, then
+		 * force_wake_mt_get will not wake the device and the
+		 * ECOBUS read will return zero. Which will be
+		 * (correctly) interpreted by the test below as MT
+		 * forcewake being disabled.
+		 */
+		mutex_lock(&dev->struct_mutex);
+		__gen6_gt_force_wake_mt_get(dev_priv);
+		ecobus = I915_READ_NOTRACE(ECOBUS);
+		__gen6_gt_force_wake_mt_put(dev_priv);
+		mutex_unlock(&dev->struct_mutex);
+
+		if (ecobus & FORCEWAKE_MT_ENABLE) {
+			dev_priv->gt.force_wake_get =
+						__gen6_gt_force_wake_mt_get;
+			dev_priv->gt.force_wake_put =
+						__gen6_gt_force_wake_mt_put;
+		} else {
+			DRM_INFO("No MT forcewake available on Ivybridge, this can result in issues\n");
+			DRM_INFO("when using vblank-synced partial screen updates.\n");
+			dev_priv->gt.force_wake_get = __gen6_gt_force_wake_get;
+			dev_priv->gt.force_wake_put = __gen6_gt_force_wake_put;
+		}
 	} else if (IS_GEN6(dev)) {
 		dev_priv->gt.force_wake_get = __gen6_gt_force_wake_get;
 		dev_priv->gt.force_wake_put = __gen6_gt_force_wake_put;
@@ -4568,55 +5589,58 @@ int sandybridge_pcode_write(struct drm_i915_private *dev_priv, u8 mbox, u32 val)
 	return 0;
 }
 
-static int vlv_punit_rw(struct drm_i915_private *dev_priv, u8 opcode,
-			u8 addr, u32 *val)
+int vlv_gpu_freq(int ddr_freq, int val)
 {
-	u32 cmd, devfn, port, be, bar;
-
-	bar = 0;
-	be = 0xf;
-	port = IOSF_PORT_PUNIT;
-	devfn = PCI_DEVFN(2, 0);
+	int mult, base;
 
-	cmd = (devfn << IOSF_DEVFN_SHIFT) | (opcode << IOSF_OPCODE_SHIFT) |
-		(port << IOSF_PORT_SHIFT) | (be << IOSF_BYTE_ENABLES_SHIFT) |
-		(bar << IOSF_BAR_SHIFT);
-
-	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
-
-	if (I915_READ(VLV_IOSF_DOORBELL_REQ) & IOSF_SB_BUSY) {
-		DRM_DEBUG_DRIVER("warning: pcode (%s) mailbox access failed\n",
-				 opcode == PUNIT_OPCODE_REG_READ ?
-				 "read" : "write");
-		return -EAGAIN;
+	switch (ddr_freq) {
+	case 800:
+		mult = 20;
+		base = 120;
+		break;
+	case 1066:
+		mult = 22;
+		base = 133;
+		break;
+	case 1333:
+		mult = 21;
+		base = 125;
+		break;
+	default:
+		return -1;
 	}
 
-	I915_WRITE(VLV_IOSF_ADDR, addr);
-	if (opcode == PUNIT_OPCODE_REG_WRITE)
-		I915_WRITE(VLV_IOSF_DATA, *val);
-	I915_WRITE(VLV_IOSF_DOORBELL_REQ, cmd);
+	return ((val - 0xbd) * mult) + base;
+}
 
-	if (wait_for((I915_READ(VLV_IOSF_DOORBELL_REQ) & IOSF_SB_BUSY) == 0,
-		     500)) {
-		DRM_ERROR("timeout waiting for pcode %s (%d) to finish\n",
-			  opcode == PUNIT_OPCODE_REG_READ ? "read" : "write",
-			  addr);
-		return -ETIMEDOUT;
+int vlv_freq_opcode(int ddr_freq, int val)
+{
+	int mult, base;
+
+	switch (ddr_freq) {
+	case 800:
+		mult = 20;
+		base = 120;
+		break;
+	case 1066:
+		mult = 22;
+		base = 133;
+		break;
+	case 1333:
+		mult = 21;
+		base = 125;
+		break;
+	default:
+		return -1;
 	}
 
-	if (opcode == PUNIT_OPCODE_REG_READ)
-		*val = I915_READ(VLV_IOSF_DATA);
-	I915_WRITE(VLV_IOSF_DATA, 0);
+	val /= mult;
+	val -= base / mult;
+	val += 0xbd;
 
-	return 0;
-}
+	if (val > 0xea)
+		val = 0xea;
 
-int valleyview_punit_read(struct drm_i915_private *dev_priv, u8 addr, u32 *val)
-{
-	return vlv_punit_rw(dev_priv, PUNIT_OPCODE_REG_READ, addr, val);
+	return val;
 }
 
-int valleyview_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val)
-{
-	return vlv_punit_rw(dev_priv, PUNIT_OPCODE_REG_WRITE, addr, &val);
-}
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 1d5d613eb6be..664118d8c1d6 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -280,6 +280,27 @@ gen7_render_ring_cs_stall_wa(struct intel_ring_buffer *ring)
 	return 0;
 }
 
+static int gen7_ring_fbc_flush(struct intel_ring_buffer *ring, u32 value)
+{
+	int ret;
+
+	if (!ring->fbc_dirty)
+		return 0;
+
+	ret = intel_ring_begin(ring, 4);
+	if (ret)
+		return ret;
+	intel_ring_emit(ring, MI_NOOP);
+	/* WaFbcNukeOn3DBlt:ivb/hsw */
+	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
+	intel_ring_emit(ring, MSG_FBC_REND_STATE);
+	intel_ring_emit(ring, value);
+	intel_ring_advance(ring);
+
+	ring->fbc_dirty = false;
+	return 0;
+}
+
 static int
 gen7_render_ring_flush(struct intel_ring_buffer *ring,
 		       u32 invalidate_domains, u32 flush_domains)
@@ -336,6 +357,9 @@ gen7_render_ring_flush(struct intel_ring_buffer *ring,
 	intel_ring_emit(ring, 0);
 	intel_ring_advance(ring);
 
+	if (flush_domains)
+		return gen7_ring_fbc_flush(ring, FBC_REND_NUKE);
+
 	return 0;
 }
 
@@ -355,6 +379,17 @@ u32 intel_ring_get_active_head(struct intel_ring_buffer *ring)
 	return I915_READ(acthd_reg);
 }
 
+static void ring_setup_phys_status_page(struct intel_ring_buffer *ring)
+{
+	struct drm_i915_private *dev_priv = ring->dev->dev_private;
+	u32 addr;
+
+	addr = dev_priv->status_page_dmah->busaddr;
+	if (INTEL_INFO(ring->dev)->gen >= 4)
+		addr |= (dev_priv->status_page_dmah->busaddr >> 28) & 0xf0;
+	I915_WRITE(HWS_PGA, addr);
+}
+
 static int init_ring_common(struct intel_ring_buffer *ring)
 {
 	struct drm_device *dev = ring->dev;
@@ -366,6 +401,11 @@ static int init_ring_common(struct intel_ring_buffer *ring)
 	if (HAS_FORCE_WAKE(dev))
 		gen6_gt_force_wake_get(dev_priv);
 
+	if (I915_NEED_GFX_HWS(dev))
+		intel_ring_setup_status_page(ring);
+	else
+		ring_setup_phys_status_page(ring);
+
 	/* Stop the ring if it's running. */
 	I915_WRITE_CTL(ring, 0);
 	I915_WRITE_HEAD(ring, 0);
@@ -429,6 +469,8 @@ static int init_ring_common(struct intel_ring_buffer *ring)
 		ring->last_retired_head = -1;
 	}
 
+	memset(&ring->hangcheck, 0, sizeof(ring->hangcheck));
+
 out:
 	if (HAS_FORCE_WAKE(dev))
 		gen6_gt_force_wake_put(dev_priv);
@@ -464,9 +506,11 @@ init_pipe_control(struct intel_ring_buffer *ring)
 		goto err_unref;
 
 	pc->gtt_offset = obj->gtt_offset;
-	pc->cpu_page =  kmap(sg_page(obj->pages->sgl));
-	if (pc->cpu_page == NULL)
+	pc->cpu_page = kmap(sg_page(obj->pages->sgl));
+	if (pc->cpu_page == NULL) {
+		ret = -ENOMEM;
 		goto err_unpin;
+	}
 
 	DRM_DEBUG_DRIVER("%s pipe control offset: 0x%08x\n",
 			 ring->name, pc->gtt_offset);
@@ -490,9 +534,6 @@ cleanup_pipe_control(struct intel_ring_buffer *ring)
 	struct pipe_control *pc = ring->private;
 	struct drm_i915_gem_object *obj;
 
-	if (!ring->private)
-		return;
-
 	obj = pc->obj;
 
 	kunmap(sg_page(obj->pages->sgl));
@@ -500,7 +541,6 @@ cleanup_pipe_control(struct intel_ring_buffer *ring)
 	drm_gem_object_unreference(&obj->base);
 
 	kfree(pc);
-	ring->private = NULL;
 }
 
 static int init_render_ring(struct intel_ring_buffer *ring)
@@ -515,6 +555,8 @@ static int init_render_ring(struct intel_ring_buffer *ring)
 	/* We need to disable the AsyncFlip performance optimisations in order
 	 * to use MI_WAIT_FOR_EVENT within the CS. It should already be
 	 * programmed to '1' on all products.
+	 *
+	 * WaDisableAsyncFlipPerfMode:snb,ivb,hsw,vlv
 	 */
 	if (INTEL_INFO(dev)->gen >= 6)
 		I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(ASYNC_FLIP_PERF_DISABLE));
@@ -556,7 +598,7 @@ static int init_render_ring(struct intel_ring_buffer *ring)
 		I915_WRITE(INSTPM, _MASKED_BIT_ENABLE(INSTPM_FORCE_ORDERING));
 
 	if (HAS_L3_GPU_CACHE(dev))
-		I915_WRITE_IMR(ring, ~GEN6_RENDER_L3_PARITY_ERROR);
+		I915_WRITE_IMR(ring, ~GT_RENDER_L3_PARITY_ERROR_INTERRUPT);
 
 	return ret;
 }
@@ -571,16 +613,26 @@ static void render_ring_cleanup(struct intel_ring_buffer *ring)
 	if (HAS_BROKEN_CS_TLB(dev))
 		drm_gem_object_unreference(to_gem_object(ring->private));
 
-	cleanup_pipe_control(ring);
+	if (INTEL_INFO(dev)->gen >= 5)
+		cleanup_pipe_control(ring);
+
+	ring->private = NULL;
 }
 
 static void
 update_mboxes(struct intel_ring_buffer *ring,
 	      u32 mmio_offset)
 {
+/* NB: In order to be able to do semaphore MBOX updates for varying number
+ * of rings, it's easiest if we round up each individual update to a
+ * multiple of 2 (since ring updates must always be a multiple of 2)
+ * even though the actual update only requires 3 dwords.
+ */
+#define MBOX_UPDATE_DWORDS 4
 	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
 	intel_ring_emit(ring, mmio_offset);
 	intel_ring_emit(ring, ring->outstanding_lazy_request);
+	intel_ring_emit(ring, MI_NOOP);
 }
 
 /**
@@ -595,19 +647,24 @@ update_mboxes(struct intel_ring_buffer *ring,
 static int
 gen6_add_request(struct intel_ring_buffer *ring)
 {
-	u32 mbox1_reg;
-	u32 mbox2_reg;
-	int ret;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_ring_buffer *useless;
+	int i, ret;
 
-	ret = intel_ring_begin(ring, 10);
+	ret = intel_ring_begin(ring, ((I915_NUM_RINGS-1) *
+				      MBOX_UPDATE_DWORDS) +
+				      4);
 	if (ret)
 		return ret;
+#undef MBOX_UPDATE_DWORDS
 
-	mbox1_reg = ring->signal_mbox[0];
-	mbox2_reg = ring->signal_mbox[1];
+	for_each_ring(useless, dev_priv, i) {
+		u32 mbox_reg = ring->signal_mbox[i];
+		if (mbox_reg != GEN6_NOSYNC)
+			update_mboxes(ring, mbox_reg);
+	}
 
-	update_mboxes(ring, mbox1_reg);
-	update_mboxes(ring, mbox2_reg);
 	intel_ring_emit(ring, MI_STORE_DWORD_INDEX);
 	intel_ring_emit(ring, I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT);
 	intel_ring_emit(ring, ring->outstanding_lazy_request);
@@ -779,7 +836,7 @@ gen5_ring_get_irq(struct intel_ring_buffer *ring)
 		return false;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (ring->irq_refcount++ == 0) {
+	if (ring->irq_refcount.gt++ == 0) {
 		dev_priv->gt_irq_mask &= ~ring->irq_enable_mask;
 		I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 		POSTING_READ(GTIMR);
@@ -797,7 +854,7 @@ gen5_ring_put_irq(struct intel_ring_buffer *ring)
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--ring->irq_refcount == 0) {
+	if (--ring->irq_refcount.gt == 0) {
 		dev_priv->gt_irq_mask |= ring->irq_enable_mask;
 		I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
 		POSTING_READ(GTIMR);
@@ -816,7 +873,7 @@ i9xx_ring_get_irq(struct intel_ring_buffer *ring)
 		return false;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (ring->irq_refcount++ == 0) {
+	if (ring->irq_refcount.gt++ == 0) {
 		dev_priv->irq_mask &= ~ring->irq_enable_mask;
 		I915_WRITE(IMR, dev_priv->irq_mask);
 		POSTING_READ(IMR);
@@ -834,7 +891,7 @@ i9xx_ring_put_irq(struct intel_ring_buffer *ring)
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--ring->irq_refcount == 0) {
+	if (--ring->irq_refcount.gt == 0) {
 		dev_priv->irq_mask |= ring->irq_enable_mask;
 		I915_WRITE(IMR, dev_priv->irq_mask);
 		POSTING_READ(IMR);
@@ -853,7 +910,7 @@ i8xx_ring_get_irq(struct intel_ring_buffer *ring)
 		return false;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (ring->irq_refcount++ == 0) {
+	if (ring->irq_refcount.gt++ == 0) {
 		dev_priv->irq_mask &= ~ring->irq_enable_mask;
 		I915_WRITE16(IMR, dev_priv->irq_mask);
 		POSTING_READ16(IMR);
@@ -871,7 +928,7 @@ i8xx_ring_put_irq(struct intel_ring_buffer *ring)
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--ring->irq_refcount == 0) {
+	if (--ring->irq_refcount.gt == 0) {
 		dev_priv->irq_mask |= ring->irq_enable_mask;
 		I915_WRITE16(IMR, dev_priv->irq_mask);
 		POSTING_READ16(IMR);
@@ -899,6 +956,9 @@ void intel_ring_setup_status_page(struct intel_ring_buffer *ring)
 		case VCS:
 			mmio = BSD_HWS_PGA_GEN7;
 			break;
+		case VECS:
+			mmio = VEBOX_HWS_PGA_GEN7;
+			break;
 		}
 	} else if (IS_GEN6(ring->dev)) {
 		mmio = RING_HWS_PGA_GEN6(ring->mmio_base);
@@ -961,10 +1021,11 @@ gen6_ring_get_irq(struct intel_ring_buffer *ring)
 	gen6_gt_force_wake_get(dev_priv);
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (ring->irq_refcount++ == 0) {
+	if (ring->irq_refcount.gt++ == 0) {
 		if (HAS_L3_GPU_CACHE(dev) && ring->id == RCS)
-			I915_WRITE_IMR(ring, ~(ring->irq_enable_mask |
-						GEN6_RENDER_L3_PARITY_ERROR));
+			I915_WRITE_IMR(ring,
+				       ~(ring->irq_enable_mask |
+					 GT_RENDER_L3_PARITY_ERROR_INTERRUPT));
 		else
 			I915_WRITE_IMR(ring, ~ring->irq_enable_mask);
 		dev_priv->gt_irq_mask &= ~ring->irq_enable_mask;
@@ -984,9 +1045,10 @@ gen6_ring_put_irq(struct intel_ring_buffer *ring)
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev_priv->irq_lock, flags);
-	if (--ring->irq_refcount == 0) {
+	if (--ring->irq_refcount.gt == 0) {
 		if (HAS_L3_GPU_CACHE(dev) && ring->id == RCS)
-			I915_WRITE_IMR(ring, ~GEN6_RENDER_L3_PARITY_ERROR);
+			I915_WRITE_IMR(ring,
+				       ~GT_RENDER_L3_PARITY_ERROR_INTERRUPT);
 		else
 			I915_WRITE_IMR(ring, ~0);
 		dev_priv->gt_irq_mask |= ring->irq_enable_mask;
@@ -998,6 +1060,48 @@ gen6_ring_put_irq(struct intel_ring_buffer *ring)
 	gen6_gt_force_wake_put(dev_priv);
 }
 
+static bool
+hsw_vebox_get_irq(struct intel_ring_buffer *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long flags;
+
+	if (!dev->irq_enabled)
+		return false;
+
+	spin_lock_irqsave(&dev_priv->rps.lock, flags);
+	if (ring->irq_refcount.pm++ == 0) {
+		u32 pm_imr = I915_READ(GEN6_PMIMR);
+		I915_WRITE_IMR(ring, ~ring->irq_enable_mask);
+		I915_WRITE(GEN6_PMIMR, pm_imr & ~ring->irq_enable_mask);
+		POSTING_READ(GEN6_PMIMR);
+	}
+	spin_unlock_irqrestore(&dev_priv->rps.lock, flags);
+
+	return true;
+}
+
+static void
+hsw_vebox_put_irq(struct intel_ring_buffer *ring)
+{
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long flags;
+
+	if (!dev->irq_enabled)
+		return;
+
+	spin_lock_irqsave(&dev_priv->rps.lock, flags);
+	if (--ring->irq_refcount.pm == 0) {
+		u32 pm_imr = I915_READ(GEN6_PMIMR);
+		I915_WRITE_IMR(ring, ~0);
+		I915_WRITE(GEN6_PMIMR, pm_imr | ring->irq_enable_mask);
+		POSTING_READ(GEN6_PMIMR);
+	}
+	spin_unlock_irqrestore(&dev_priv->rps.lock, flags);
+}
+
 static int
 i965_dispatch_execbuffer(struct intel_ring_buffer *ring,
 			 u32 offset, u32 length,
@@ -1134,7 +1238,6 @@ static int init_status_page(struct intel_ring_buffer *ring)
 	ring->status_page.obj = obj;
 	memset(ring->status_page.page_addr, 0, PAGE_SIZE);
 
-	intel_ring_setup_status_page(ring);
 	DRM_DEBUG_DRIVER("%s hws offset: 0x%08x\n",
 			ring->name, ring->status_page.gfx_addr);
 
@@ -1148,10 +1251,9 @@ err:
 	return ret;
 }
 
-static int init_phys_hws_pga(struct intel_ring_buffer *ring)
+static int init_phys_status_page(struct intel_ring_buffer *ring)
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
-	u32 addr;
 
 	if (!dev_priv->status_page_dmah) {
 		dev_priv->status_page_dmah =
@@ -1160,11 +1262,6 @@ static int init_phys_hws_pga(struct intel_ring_buffer *ring)
 			return -ENOMEM;
 	}
 
-	addr = dev_priv->status_page_dmah->busaddr;
-	if (INTEL_INFO(ring->dev)->gen >= 4)
-		addr |= (dev_priv->status_page_dmah->busaddr >> 28) & 0xf0;
-	I915_WRITE(HWS_PGA, addr);
-
 	ring->status_page.page_addr = dev_priv->status_page_dmah->vaddr;
 	memset(ring->status_page.page_addr, 0, PAGE_SIZE);
 
@@ -1192,7 +1289,7 @@ static int intel_init_ring_buffer(struct drm_device *dev,
 			return ret;
 	} else {
 		BUG_ON(ring->id != RCS);
-		ret = init_phys_hws_pga(ring);
+		ret = init_phys_status_page(ring);
 		if (ret)
 			return ret;
 	}
@@ -1423,7 +1520,7 @@ int intel_ring_idle(struct intel_ring_buffer *ring)
 
 	/* We need to add any requests required to flush the objects and ring */
 	if (ring->outstanding_lazy_request) {
-		ret = i915_add_request(ring, NULL, NULL);
+		ret = i915_add_request(ring, NULL);
 		if (ret)
 			return ret;
 	}
@@ -1500,6 +1597,7 @@ void intel_ring_init_seqno(struct intel_ring_buffer *ring, u32 seqno)
 	}
 
 	ring->set_seqno(ring, seqno);
+	ring->hangcheck.seqno = seqno;
 }
 
 void intel_ring_advance(struct intel_ring_buffer *ring)
@@ -1546,8 +1644,8 @@ static void gen6_bsd_ring_write_tail(struct intel_ring_buffer *ring,
 		   _MASKED_BIT_DISABLE(GEN6_BSD_SLEEP_MSG_DISABLE));
 }
 
-static int gen6_ring_flush(struct intel_ring_buffer *ring,
-			   u32 invalidate, u32 flush)
+static int gen6_bsd_ring_flush(struct intel_ring_buffer *ring,
+			       u32 invalidate, u32 flush)
 {
 	uint32_t cmd;
 	int ret;
@@ -1618,9 +1716,10 @@ gen6_ring_dispatch_execbuffer(struct intel_ring_buffer *ring,
 
 /* Blitter support (SandyBridge+) */
 
-static int blt_ring_flush(struct intel_ring_buffer *ring,
-			  u32 invalidate, u32 flush)
+static int gen6_ring_flush(struct intel_ring_buffer *ring,
+			   u32 invalidate, u32 flush)
 {
+	struct drm_device *dev = ring->dev;
 	uint32_t cmd;
 	int ret;
 
@@ -1643,6 +1742,10 @@ static int blt_ring_flush(struct intel_ring_buffer *ring,
 	intel_ring_emit(ring, 0);
 	intel_ring_emit(ring, MI_NOOP);
 	intel_ring_advance(ring);
+
+	if (IS_GEN7(dev) && flush)
+		return gen7_ring_fbc_flush(ring, FBC_REND_CACHE_CLEAN);
+
 	return 0;
 }
 
@@ -1662,15 +1765,18 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 			ring->flush = gen6_render_ring_flush;
 		ring->irq_get = gen6_ring_get_irq;
 		ring->irq_put = gen6_ring_put_irq;
-		ring->irq_enable_mask = GT_USER_INTERRUPT;
+		ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT;
 		ring->get_seqno = gen6_ring_get_seqno;
 		ring->set_seqno = ring_set_seqno;
 		ring->sync_to = gen6_ring_sync;
-		ring->semaphore_register[0] = MI_SEMAPHORE_SYNC_INVALID;
-		ring->semaphore_register[1] = MI_SEMAPHORE_SYNC_RV;
-		ring->semaphore_register[2] = MI_SEMAPHORE_SYNC_RB;
-		ring->signal_mbox[0] = GEN6_VRSYNC;
-		ring->signal_mbox[1] = GEN6_BRSYNC;
+		ring->semaphore_register[RCS] = MI_SEMAPHORE_SYNC_INVALID;
+		ring->semaphore_register[VCS] = MI_SEMAPHORE_SYNC_RV;
+		ring->semaphore_register[BCS] = MI_SEMAPHORE_SYNC_RB;
+		ring->semaphore_register[VECS] = MI_SEMAPHORE_SYNC_RVE;
+		ring->signal_mbox[RCS] = GEN6_NOSYNC;
+		ring->signal_mbox[VCS] = GEN6_VRSYNC;
+		ring->signal_mbox[BCS] = GEN6_BRSYNC;
+		ring->signal_mbox[VECS] = GEN6_VERSYNC;
 	} else if (IS_GEN5(dev)) {
 		ring->add_request = pc_render_add_request;
 		ring->flush = gen4_render_ring_flush;
@@ -1678,7 +1784,8 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 		ring->set_seqno = pc_render_set_seqno;
 		ring->irq_get = gen5_ring_get_irq;
 		ring->irq_put = gen5_ring_put_irq;
-		ring->irq_enable_mask = GT_USER_INTERRUPT | GT_PIPE_NOTIFY;
+		ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT |
+					GT_RENDER_PIPECTL_NOTIFY_INTERRUPT;
 	} else {
 		ring->add_request = i9xx_add_request;
 		if (INTEL_INFO(dev)->gen < 4)
@@ -1794,7 +1901,7 @@ int intel_render_ring_init_dri(struct drm_device *dev, u64 start, u32 size)
 	}
 
 	if (!I915_NEED_GFX_HWS(dev)) {
-		ret = init_phys_hws_pga(ring);
+		ret = init_phys_status_page(ring);
 		if (ret)
 			return ret;
 	}
@@ -1816,20 +1923,23 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev)
 		/* gen6 bsd needs a special wa for tail updates */
 		if (IS_GEN6(dev))
 			ring->write_tail = gen6_bsd_ring_write_tail;
-		ring->flush = gen6_ring_flush;
+		ring->flush = gen6_bsd_ring_flush;
 		ring->add_request = gen6_add_request;
 		ring->get_seqno = gen6_ring_get_seqno;
 		ring->set_seqno = ring_set_seqno;
-		ring->irq_enable_mask = GEN6_BSD_USER_INTERRUPT;
+		ring->irq_enable_mask = GT_BSD_USER_INTERRUPT;
 		ring->irq_get = gen6_ring_get_irq;
 		ring->irq_put = gen6_ring_put_irq;
 		ring->dispatch_execbuffer = gen6_ring_dispatch_execbuffer;
 		ring->sync_to = gen6_ring_sync;
-		ring->semaphore_register[0] = MI_SEMAPHORE_SYNC_VR;
-		ring->semaphore_register[1] = MI_SEMAPHORE_SYNC_INVALID;
-		ring->semaphore_register[2] = MI_SEMAPHORE_SYNC_VB;
-		ring->signal_mbox[0] = GEN6_RVSYNC;
-		ring->signal_mbox[1] = GEN6_BVSYNC;
+		ring->semaphore_register[RCS] = MI_SEMAPHORE_SYNC_VR;
+		ring->semaphore_register[VCS] = MI_SEMAPHORE_SYNC_INVALID;
+		ring->semaphore_register[BCS] = MI_SEMAPHORE_SYNC_VB;
+		ring->semaphore_register[VECS] = MI_SEMAPHORE_SYNC_VVE;
+		ring->signal_mbox[RCS] = GEN6_RVSYNC;
+		ring->signal_mbox[VCS] = GEN6_NOSYNC;
+		ring->signal_mbox[BCS] = GEN6_BVSYNC;
+		ring->signal_mbox[VECS] = GEN6_VEVSYNC;
 	} else {
 		ring->mmio_base = BSD_RING_BASE;
 		ring->flush = bsd_ring_flush;
@@ -1837,7 +1947,7 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev)
 		ring->get_seqno = ring_get_seqno;
 		ring->set_seqno = ring_set_seqno;
 		if (IS_GEN5(dev)) {
-			ring->irq_enable_mask = GT_BSD_USER_INTERRUPT;
+			ring->irq_enable_mask = ILK_BSD_USER_INTERRUPT;
 			ring->irq_get = gen5_ring_get_irq;
 			ring->irq_put = gen5_ring_put_irq;
 		} else {
@@ -1862,20 +1972,56 @@ int intel_init_blt_ring_buffer(struct drm_device *dev)
 
 	ring->mmio_base = BLT_RING_BASE;
 	ring->write_tail = ring_write_tail;
-	ring->flush = blt_ring_flush;
+	ring->flush = gen6_ring_flush;
 	ring->add_request = gen6_add_request;
 	ring->get_seqno = gen6_ring_get_seqno;
 	ring->set_seqno = ring_set_seqno;
-	ring->irq_enable_mask = GEN6_BLITTER_USER_INTERRUPT;
+	ring->irq_enable_mask = GT_BLT_USER_INTERRUPT;
 	ring->irq_get = gen6_ring_get_irq;
 	ring->irq_put = gen6_ring_put_irq;
 	ring->dispatch_execbuffer = gen6_ring_dispatch_execbuffer;
 	ring->sync_to = gen6_ring_sync;
-	ring->semaphore_register[0] = MI_SEMAPHORE_SYNC_BR;
-	ring->semaphore_register[1] = MI_SEMAPHORE_SYNC_BV;
-	ring->semaphore_register[2] = MI_SEMAPHORE_SYNC_INVALID;
-	ring->signal_mbox[0] = GEN6_RBSYNC;
-	ring->signal_mbox[1] = GEN6_VBSYNC;
+	ring->semaphore_register[RCS] = MI_SEMAPHORE_SYNC_BR;
+	ring->semaphore_register[VCS] = MI_SEMAPHORE_SYNC_BV;
+	ring->semaphore_register[BCS] = MI_SEMAPHORE_SYNC_INVALID;
+	ring->semaphore_register[VECS] = MI_SEMAPHORE_SYNC_BVE;
+	ring->signal_mbox[RCS] = GEN6_RBSYNC;
+	ring->signal_mbox[VCS] = GEN6_VBSYNC;
+	ring->signal_mbox[BCS] = GEN6_NOSYNC;
+	ring->signal_mbox[VECS] = GEN6_VEBSYNC;
+	ring->init = init_ring_common;
+
+	return intel_init_ring_buffer(dev, ring);
+}
+
+int intel_init_vebox_ring_buffer(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct intel_ring_buffer *ring = &dev_priv->ring[VECS];
+
+	ring->name = "video enhancement ring";
+	ring->id = VECS;
+
+	ring->mmio_base = VEBOX_RING_BASE;
+	ring->write_tail = ring_write_tail;
+	ring->flush = gen6_ring_flush;
+	ring->add_request = gen6_add_request;
+	ring->get_seqno = gen6_ring_get_seqno;
+	ring->set_seqno = ring_set_seqno;
+	ring->irq_enable_mask = PM_VEBOX_USER_INTERRUPT |
+		PM_VEBOX_CS_ERROR_INTERRUPT;
+	ring->irq_get = hsw_vebox_get_irq;
+	ring->irq_put = hsw_vebox_put_irq;
+	ring->dispatch_execbuffer = gen6_ring_dispatch_execbuffer;
+	ring->sync_to = gen6_ring_sync;
+	ring->semaphore_register[RCS] = MI_SEMAPHORE_SYNC_VER;
+	ring->semaphore_register[VCS] = MI_SEMAPHORE_SYNC_VEV;
+	ring->semaphore_register[BCS] = MI_SEMAPHORE_SYNC_VEB;
+	ring->semaphore_register[VECS] = MI_SEMAPHORE_SYNC_INVALID;
+	ring->signal_mbox[RCS] = GEN6_RVESYNC;
+	ring->signal_mbox[VCS] = GEN6_VVESYNC;
+	ring->signal_mbox[BCS] = GEN6_BVESYNC;
+	ring->signal_mbox[VECS] = GEN6_NOSYNC;
 	ring->init = init_ring_common;
 
 	return intel_init_ring_buffer(dev, ring);
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index d66208c2c48b..799f04c9da45 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -37,14 +37,25 @@ struct  intel_hw_status_page {
 #define I915_READ_SYNC_0(ring) I915_READ(RING_SYNC_0((ring)->mmio_base))
 #define I915_READ_SYNC_1(ring) I915_READ(RING_SYNC_1((ring)->mmio_base))
 
+enum intel_ring_hangcheck_action { wait, active, kick, hung };
+
+struct intel_ring_hangcheck {
+	bool deadlock;
+	u32 seqno;
+	u32 acthd;
+	int score;
+	enum intel_ring_hangcheck_action action;
+};
+
 struct  intel_ring_buffer {
 	const char	*name;
 	enum intel_ring_id {
 		RCS = 0x0,
 		VCS,
 		BCS,
+		VECS,
 	} id;
-#define I915_NUM_RINGS 3
+#define I915_NUM_RINGS 4
 	u32		mmio_base;
 	void		__iomem *virtual_start;
 	struct		drm_device *dev;
@@ -67,7 +78,10 @@ struct  intel_ring_buffer {
 	 */
 	u32		last_retired_head;
 
-	u32		irq_refcount;		/* protected by dev_priv->irq_lock */
+	struct {
+		u32	gt; /*  protected by dev_priv->irq_lock */
+		u32	pm; /*  protected by dev_priv->rps.lock (sucks) */
+	} irq_refcount;
 	u32		irq_enable_mask;	/* bitmask to enable ring interrupt */
 	u32		trace_irq_seqno;
 	u32		sync_seqno[I915_NUM_RINGS-1];
@@ -102,8 +116,11 @@ struct  intel_ring_buffer {
 				   struct intel_ring_buffer *to,
 				   u32 seqno);
 
-	u32		semaphore_register[3]; /*our mbox written by others */
-	u32		signal_mbox[2]; /* mboxes this ring signals to */
+	/* our mbox written by others */
+	u32		semaphore_register[I915_NUM_RINGS];
+	/* mboxes this ring signals to */
+	u32		signal_mbox[I915_NUM_RINGS];
+
 	/**
 	 * List of objects currently involved in rendering from the
 	 * ringbuffer.
@@ -127,6 +144,7 @@ struct  intel_ring_buffer {
 	 */
 	u32 outstanding_lazy_request;
 	bool gpu_caches_dirty;
+	bool fbc_dirty;
 
 	wait_queue_head_t irq_queue;
 
@@ -135,7 +153,9 @@ struct  intel_ring_buffer {
 	 */
 	bool itlb_before_ctx_switch;
 	struct i915_hw_context *default_context;
-	struct drm_i915_gem_object *last_context_obj;
+	struct i915_hw_context *last_context;
+
+	struct intel_ring_hangcheck hangcheck;
 
 	void *private;
 };
@@ -224,6 +244,7 @@ int intel_ring_invalidate_all_caches(struct intel_ring_buffer *ring);
 int intel_init_render_ring_buffer(struct drm_device *dev);
 int intel_init_bsd_ring_buffer(struct drm_device *dev);
 int intel_init_blt_ring_buffer(struct drm_device *dev);
+int intel_init_vebox_ring_buffer(struct drm_device *dev);
 
 u32 intel_ring_get_active_head(struct intel_ring_buffer *ring);
 void intel_ring_setup_status_page(struct intel_ring_buffer *ring);
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index d4ea6c265ce1..2628d5622449 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -80,7 +80,7 @@ struct intel_sdvo {
 
 	/*
 	 * Capabilities of the SDVO device returned by
-	 * i830_sdvo_get_capabilities()
+	 * intel_sdvo_get_capabilities()
 	 */
 	struct intel_sdvo_caps caps;
 
@@ -712,6 +712,13 @@ static bool intel_sdvo_set_timing(struct intel_sdvo *intel_sdvo, u8 cmd,
 		intel_sdvo_set_value(intel_sdvo, cmd + 1, &dtd->part2, sizeof(dtd->part2));
 }
 
+static bool intel_sdvo_get_timing(struct intel_sdvo *intel_sdvo, u8 cmd,
+				  struct intel_sdvo_dtd *dtd)
+{
+	return intel_sdvo_get_value(intel_sdvo, cmd, &dtd->part1, sizeof(dtd->part1)) &&
+		intel_sdvo_get_value(intel_sdvo, cmd + 1, &dtd->part2, sizeof(dtd->part2));
+}
+
 static bool intel_sdvo_set_input_timing(struct intel_sdvo *intel_sdvo,
 					 struct intel_sdvo_dtd *dtd)
 {
@@ -726,6 +733,13 @@ static bool intel_sdvo_set_output_timing(struct intel_sdvo *intel_sdvo,
 				     SDVO_CMD_SET_OUTPUT_TIMINGS_PART1, dtd);
 }
 
+static bool intel_sdvo_get_input_timing(struct intel_sdvo *intel_sdvo,
+					struct intel_sdvo_dtd *dtd)
+{
+	return intel_sdvo_get_timing(intel_sdvo,
+				     SDVO_CMD_GET_INPUT_TIMINGS_PART1, dtd);
+}
+
 static bool
 intel_sdvo_create_preferred_input_timing(struct intel_sdvo *intel_sdvo,
 					 uint16_t clock,
@@ -1041,6 +1055,32 @@ intel_sdvo_get_preferred_input_mode(struct intel_sdvo *intel_sdvo,
 	return true;
 }
 
+static void i9xx_adjust_sdvo_tv_clock(struct intel_crtc_config *pipe_config)
+{
+	unsigned dotclock = pipe_config->adjusted_mode.clock;
+	struct dpll *clock = &pipe_config->dpll;
+
+	/* SDVO TV has fixed PLL values depend on its clock range,
+	   this mirrors vbios setting. */
+	if (dotclock >= 100000 && dotclock < 140500) {
+		clock->p1 = 2;
+		clock->p2 = 10;
+		clock->n = 3;
+		clock->m1 = 16;
+		clock->m2 = 8;
+	} else if (dotclock >= 140500 && dotclock <= 200000) {
+		clock->p1 = 1;
+		clock->p2 = 10;
+		clock->n = 6;
+		clock->m1 = 12;
+		clock->m2 = 8;
+	} else {
+		WARN(1, "SDVO TV clock out of range: %i\n", dotclock);
+	}
+
+	pipe_config->clock_set = true;
+}
+
 static bool intel_sdvo_compute_config(struct intel_encoder *encoder,
 				      struct intel_crtc_config *pipe_config)
 {
@@ -1066,6 +1106,7 @@ static bool intel_sdvo_compute_config(struct intel_encoder *encoder,
 		(void) intel_sdvo_get_preferred_input_mode(intel_sdvo,
 							   mode,
 							   adjusted_mode);
+		pipe_config->sdvo_tv_clock = true;
 	} else if (intel_sdvo->is_lvds) {
 		if (!intel_sdvo_set_output_timings_from_mode(intel_sdvo,
 							     intel_sdvo->sdvo_lvds_fixed_mode))
@@ -1097,6 +1138,10 @@ static bool intel_sdvo_compute_config(struct intel_encoder *encoder,
 	if (intel_sdvo->color_range)
 		pipe_config->limited_color_range = true;
 
+	/* Clock computation needs to happen after pixel multiplier. */
+	if (intel_sdvo->is_tv)
+		i9xx_adjust_sdvo_tv_clock(pipe_config);
+
 	return true;
 }
 
@@ -1174,6 +1219,7 @@ static void intel_sdvo_mode_set(struct intel_encoder *intel_encoder)
 
 	switch (intel_crtc->config.pixel_multiplier) {
 	default:
+		WARN(1, "unknown pixel mutlipler specified\n");
 	case 1: rate = SDVO_CLOCK_RATE_MULT_1X; break;
 	case 2: rate = SDVO_CLOCK_RATE_MULT_2X; break;
 	case 4: rate = SDVO_CLOCK_RATE_MULT_4X; break;
@@ -1231,7 +1277,7 @@ static bool intel_sdvo_connector_get_hw_state(struct intel_connector *connector)
 	struct intel_sdvo_connector *intel_sdvo_connector =
 		to_intel_sdvo_connector(&connector->base);
 	struct intel_sdvo *intel_sdvo = intel_attached_sdvo(&connector->base);
-	u16 active_outputs;
+	u16 active_outputs = 0;
 
 	intel_sdvo_get_active_outputs(intel_sdvo, &active_outputs);
 
@@ -1247,7 +1293,7 @@ static bool intel_sdvo_get_hw_state(struct intel_encoder *encoder,
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_sdvo *intel_sdvo = to_intel_sdvo(&encoder->base);
-	u16 active_outputs;
+	u16 active_outputs = 0;
 	u32 tmp;
 
 	tmp = I915_READ(intel_sdvo->sdvo_reg);
@@ -1264,6 +1310,74 @@ static bool intel_sdvo_get_hw_state(struct intel_encoder *encoder,
 	return true;
 }
 
+static void intel_sdvo_get_config(struct intel_encoder *encoder,
+				  struct intel_crtc_config *pipe_config)
+{
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_sdvo *intel_sdvo = to_intel_sdvo(&encoder->base);
+	struct intel_sdvo_dtd dtd;
+	int encoder_pixel_multiplier = 0;
+	u32 flags = 0, sdvox;
+	u8 val;
+	bool ret;
+
+	ret = intel_sdvo_get_input_timing(intel_sdvo, &dtd);
+	if (!ret) {
+		/* Some sdvo encoders are not spec compliant and don't
+		 * implement the mandatory get_timings function. */
+		DRM_DEBUG_DRIVER("failed to retrieve SDVO DTD\n");
+		pipe_config->quirks |= PIPE_CONFIG_QUIRK_MODE_SYNC_FLAGS;
+	} else {
+		if (dtd.part2.dtd_flags & DTD_FLAG_HSYNC_POSITIVE)
+			flags |= DRM_MODE_FLAG_PHSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NHSYNC;
+
+		if (dtd.part2.dtd_flags & DTD_FLAG_VSYNC_POSITIVE)
+			flags |= DRM_MODE_FLAG_PVSYNC;
+		else
+			flags |= DRM_MODE_FLAG_NVSYNC;
+	}
+
+	pipe_config->adjusted_mode.flags |= flags;
+
+	/*
+	 * pixel multiplier readout is tricky: Only on i915g/gm it is stored in
+	 * the sdvo port register, on all other platforms it is part of the dpll
+	 * state. Since the general pipe state readout happens before the
+	 * encoder->get_config we so already have a valid pixel multplier on all
+	 * other platfroms.
+	 */
+	if (IS_I915G(dev) || IS_I915GM(dev)) {
+		sdvox = I915_READ(intel_sdvo->sdvo_reg);
+		pipe_config->pixel_multiplier =
+			((sdvox & SDVO_PORT_MULTIPLY_MASK)
+			 >> SDVO_PORT_MULTIPLY_SHIFT) + 1;
+	}
+
+	/* Cross check the port pixel multiplier with the sdvo encoder state. */
+	intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_CLOCK_RATE_MULT, &val, 1);
+	switch (val) {
+	case SDVO_CLOCK_RATE_MULT_1X:
+		encoder_pixel_multiplier = 1;
+		break;
+	case SDVO_CLOCK_RATE_MULT_2X:
+		encoder_pixel_multiplier = 2;
+		break;
+	case SDVO_CLOCK_RATE_MULT_4X:
+		encoder_pixel_multiplier = 4;
+		break;
+	}
+
+	if(HAS_PCH_SPLIT(dev))
+		return; /* no pixel multiplier readout support yet */
+
+	WARN(encoder_pixel_multiplier != pipe_config->pixel_multiplier,
+	     "SDVO pixel multiplier mismatch, port: %i, encoder: %i\n",
+	     pipe_config->pixel_multiplier, encoder_pixel_multiplier);
+}
+
 static void intel_disable_sdvo(struct intel_encoder *encoder)
 {
 	struct drm_i915_private *dev_priv = encoder->base.dev->dev_private;
@@ -1344,6 +1458,7 @@ static void intel_enable_sdvo(struct intel_encoder *encoder)
 	intel_sdvo_set_active_outputs(intel_sdvo, intel_sdvo->attached_output);
 }
 
+/* Special dpms function to support cloning between dvo/sdvo/crt. */
 static void intel_sdvo_dpms(struct drm_connector *connector, int mode)
 {
 	struct drm_crtc *crtc;
@@ -1365,6 +1480,8 @@ static void intel_sdvo_dpms(struct drm_connector *connector, int mode)
 		return;
 	}
 
+	/* We set active outputs manually below in case pipe dpms doesn't change
+	 * due to cloning. */
 	if (mode != DRM_MODE_DPMS_ON) {
 		intel_sdvo_set_active_outputs(intel_sdvo, 0);
 		if (0)
@@ -1495,7 +1612,7 @@ intel_sdvo_get_analog_edid(struct drm_connector *connector)
 
 	return drm_get_edid(connector,
 			    intel_gmbus_get_adapter(dev_priv,
-						    dev_priv->crt_ddc_pin));
+						    dev_priv->vbt.crt_ddc_pin));
 }
 
 static enum drm_connector_status
@@ -1625,12 +1742,9 @@ intel_sdvo_detect(struct drm_connector *connector, bool force)
 	if (ret == connector_status_connected) {
 		intel_sdvo->is_tv = false;
 		intel_sdvo->is_lvds = false;
-		intel_sdvo->base.needs_tv_clock = false;
 
-		if (response & SDVO_TV_MASK) {
+		if (response & SDVO_TV_MASK)
 			intel_sdvo->is_tv = true;
-			intel_sdvo->base.needs_tv_clock = true;
-		}
 		if (response & SDVO_LVDS_MASK)
 			intel_sdvo->is_lvds = intel_sdvo->sdvo_lvds_fixed_mode != NULL;
 	}
@@ -1772,21 +1886,12 @@ static void intel_sdvo_get_lvds_modes(struct drm_connector *connector)
 	struct drm_display_mode *newmode;
 
 	/*
-	 * Attempt to get the mode list from DDC.
-	 * Assume that the preferred modes are
-	 * arranged in priority order.
-	 */
-	intel_ddc_get_modes(connector, &intel_sdvo->ddc);
-
-	/*
 	 * Fetch modes from VBT. For SDVO prefer the VBT mode since some
-	 * SDVO->LVDS transcoders can't cope with the EDID mode. Since
-	 * drm_mode_probed_add adds the mode at the head of the list we add it
-	 * last.
+	 * SDVO->LVDS transcoders can't cope with the EDID mode.
 	 */
-	if (dev_priv->sdvo_lvds_vbt_mode != NULL) {
+	if (dev_priv->vbt.sdvo_lvds_vbt_mode != NULL) {
 		newmode = drm_mode_duplicate(connector->dev,
-					     dev_priv->sdvo_lvds_vbt_mode);
+					     dev_priv->vbt.sdvo_lvds_vbt_mode);
 		if (newmode != NULL) {
 			/* Guarantee the mode is preferred */
 			newmode->type = (DRM_MODE_TYPE_PREFERRED |
@@ -1795,6 +1900,13 @@ static void intel_sdvo_get_lvds_modes(struct drm_connector *connector)
 		}
 	}
 
+	/*
+	 * Attempt to get the mode list from DDC.
+	 * Assume that the preferred modes are
+	 * arranged in priority order.
+	 */
+	intel_ddc_get_modes(connector, &intel_sdvo->ddc);
+
 	list_for_each_entry(newmode, &connector->probed_modes, head) {
 		if (newmode->type & DRM_MODE_TYPE_PREFERRED) {
 			intel_sdvo->sdvo_lvds_fixed_mode =
@@ -2329,7 +2441,6 @@ intel_sdvo_tv_init(struct intel_sdvo *intel_sdvo, int type)
 	intel_sdvo_connector->output_flag = type;
 
 	intel_sdvo->is_tv = true;
-	intel_sdvo->base.needs_tv_clock = true;
 
 	intel_sdvo_connector_init(intel_sdvo_connector, intel_sdvo);
 
@@ -2417,7 +2528,6 @@ static bool
 intel_sdvo_output_setup(struct intel_sdvo *intel_sdvo, uint16_t flags)
 {
 	intel_sdvo->is_tv = false;
-	intel_sdvo->base.needs_tv_clock = false;
 	intel_sdvo->is_lvds = false;
 
 	/* SDVO requires XXX1 function may not exist unless it has XXX0 function.*/
@@ -2751,7 +2861,6 @@ bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg, bool is_sdvob)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_encoder *intel_encoder;
 	struct intel_sdvo *intel_sdvo;
-	u32 hotplug_mask;
 	int i;
 	intel_sdvo = kzalloc(sizeof(struct intel_sdvo), GFP_KERNEL);
 	if (!intel_sdvo)
@@ -2780,23 +2889,12 @@ bool intel_sdvo_init(struct drm_device *dev, uint32_t sdvo_reg, bool is_sdvob)
 		}
 	}
 
-	hotplug_mask = 0;
-	if (IS_G4X(dev)) {
-		hotplug_mask = intel_sdvo->is_sdvob ?
-			SDVOB_HOTPLUG_INT_STATUS_G4X : SDVOC_HOTPLUG_INT_STATUS_G4X;
-	} else if (IS_GEN4(dev)) {
-		hotplug_mask = intel_sdvo->is_sdvob ?
-			SDVOB_HOTPLUG_INT_STATUS_I965 : SDVOC_HOTPLUG_INT_STATUS_I965;
-	} else {
-		hotplug_mask = intel_sdvo->is_sdvob ?
-			SDVOB_HOTPLUG_INT_STATUS_I915 : SDVOC_HOTPLUG_INT_STATUS_I915;
-	}
-
 	intel_encoder->compute_config = intel_sdvo_compute_config;
 	intel_encoder->disable = intel_disable_sdvo;
 	intel_encoder->mode_set = intel_sdvo_mode_set;
 	intel_encoder->enable = intel_enable_sdvo;
 	intel_encoder->get_hw_state = intel_sdvo_get_hw_state;
+	intel_encoder->get_config = intel_sdvo_get_config;
 
 	/* In default case sdvo lvds is false */
 	if (!intel_sdvo_get_capabilities(intel_sdvo, &intel_sdvo->caps))
diff --git a/drivers/gpu/drm/i915/intel_sideband.c b/drivers/gpu/drm/i915/intel_sideband.c
new file mode 100644
index 000000000000..9a0e6c5ea540
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_sideband.c
@@ -0,0 +1,177 @@
+/*
+ * Copyright © 2013 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "i915_drv.h"
+#include "intel_drv.h"
+
+/* IOSF sideband */
+static int vlv_sideband_rw(struct drm_i915_private *dev_priv, u32 devfn,
+			   u32 port, u32 opcode, u32 addr, u32 *val)
+{
+	u32 cmd, be = 0xf, bar = 0;
+	bool is_read = (opcode == PUNIT_OPCODE_REG_READ ||
+			opcode == DPIO_OPCODE_REG_READ);
+
+	cmd = (devfn << IOSF_DEVFN_SHIFT) | (opcode << IOSF_OPCODE_SHIFT) |
+		(port << IOSF_PORT_SHIFT) | (be << IOSF_BYTE_ENABLES_SHIFT) |
+		(bar << IOSF_BAR_SHIFT);
+
+	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
+
+	if (wait_for((I915_READ(VLV_IOSF_DOORBELL_REQ) & IOSF_SB_BUSY) == 0, 5)) {
+		DRM_DEBUG_DRIVER("IOSF sideband idle wait (%s) timed out\n",
+				 is_read ? "read" : "write");
+		return -EAGAIN;
+	}
+
+	I915_WRITE(VLV_IOSF_ADDR, addr);
+	if (!is_read)
+		I915_WRITE(VLV_IOSF_DATA, *val);
+	I915_WRITE(VLV_IOSF_DOORBELL_REQ, cmd);
+
+	if (wait_for((I915_READ(VLV_IOSF_DOORBELL_REQ) & IOSF_SB_BUSY) == 0, 5)) {
+		DRM_DEBUG_DRIVER("IOSF sideband finish wait (%s) timed out\n",
+				 is_read ? "read" : "write");
+		return -ETIMEDOUT;
+	}
+
+	if (is_read)
+		*val = I915_READ(VLV_IOSF_DATA);
+	I915_WRITE(VLV_IOSF_DATA, 0);
+
+	return 0;
+}
+
+u32 vlv_punit_read(struct drm_i915_private *dev_priv, u8 addr)
+{
+	u32 val = 0;
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+	mutex_lock(&dev_priv->dpio_lock);
+	vlv_sideband_rw(dev_priv, PCI_DEVFN(2, 0), IOSF_PORT_PUNIT,
+			PUNIT_OPCODE_REG_READ, addr, &val);
+	mutex_unlock(&dev_priv->dpio_lock);
+
+	return val;
+}
+
+void vlv_punit_write(struct drm_i915_private *dev_priv, u8 addr, u32 val)
+{
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+	mutex_lock(&dev_priv->dpio_lock);
+	vlv_sideband_rw(dev_priv, PCI_DEVFN(2, 0), IOSF_PORT_PUNIT,
+			PUNIT_OPCODE_REG_WRITE, addr, &val);
+	mutex_unlock(&dev_priv->dpio_lock);
+}
+
+u32 vlv_nc_read(struct drm_i915_private *dev_priv, u8 addr)
+{
+	u32 val = 0;
+
+	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
+
+	mutex_lock(&dev_priv->dpio_lock);
+	vlv_sideband_rw(dev_priv, PCI_DEVFN(2, 0), IOSF_PORT_NC,
+			PUNIT_OPCODE_REG_READ, addr, &val);
+	mutex_unlock(&dev_priv->dpio_lock);
+
+	return val;
+}
+
+u32 vlv_dpio_read(struct drm_i915_private *dev_priv, int reg)
+{
+	u32 val = 0;
+
+	vlv_sideband_rw(dev_priv, DPIO_DEVFN, IOSF_PORT_DPIO,
+			DPIO_OPCODE_REG_READ, reg, &val);
+
+	return val;
+}
+
+void vlv_dpio_write(struct drm_i915_private *dev_priv, int reg, u32 val)
+{
+	vlv_sideband_rw(dev_priv, DPIO_DEVFN, IOSF_PORT_DPIO,
+			DPIO_OPCODE_REG_WRITE, reg, &val);
+}
+
+/* SBI access */
+u32 intel_sbi_read(struct drm_i915_private *dev_priv, u16 reg,
+		   enum intel_sbi_destination destination)
+{
+	u32 value = 0;
+	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
+
+	if (wait_for((I915_READ(SBI_CTL_STAT) & SBI_BUSY) == 0,
+				100)) {
+		DRM_ERROR("timeout waiting for SBI to become ready\n");
+		return 0;
+	}
+
+	I915_WRITE(SBI_ADDR, (reg << 16));
+
+	if (destination == SBI_ICLK)
+		value = SBI_CTL_DEST_ICLK | SBI_CTL_OP_CRRD;
+	else
+		value = SBI_CTL_DEST_MPHY | SBI_CTL_OP_IORD;
+	I915_WRITE(SBI_CTL_STAT, value | SBI_BUSY);
+
+	if (wait_for((I915_READ(SBI_CTL_STAT) & (SBI_BUSY | SBI_RESPONSE_FAIL)) == 0,
+				100)) {
+		DRM_ERROR("timeout waiting for SBI to complete read transaction\n");
+		return 0;
+	}
+
+	return I915_READ(SBI_DATA);
+}
+
+void intel_sbi_write(struct drm_i915_private *dev_priv, u16 reg, u32 value,
+		     enum intel_sbi_destination destination)
+{
+	u32 tmp;
+
+	WARN_ON(!mutex_is_locked(&dev_priv->dpio_lock));
+
+	if (wait_for((I915_READ(SBI_CTL_STAT) & SBI_BUSY) == 0,
+				100)) {
+		DRM_ERROR("timeout waiting for SBI to become ready\n");
+		return;
+	}
+
+	I915_WRITE(SBI_ADDR, (reg << 16));
+	I915_WRITE(SBI_DATA, value);
+
+	if (destination == SBI_ICLK)
+		tmp = SBI_CTL_DEST_ICLK | SBI_CTL_OP_CRWR;
+	else
+		tmp = SBI_CTL_DEST_MPHY | SBI_CTL_OP_IOWR;
+	I915_WRITE(SBI_CTL_STAT, SBI_BUSY | tmp);
+
+	if (wait_for((I915_READ(SBI_CTL_STAT) & (SBI_BUSY | SBI_RESPONSE_FAIL)) == 0,
+				100)) {
+		DRM_ERROR("timeout waiting for SBI to complete write transaction\n");
+		return;
+	}
+}
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index c7d25c5dd4e6..1fa5612a4572 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -32,6 +32,7 @@
 #include <drm/drmP.h>
 #include <drm/drm_crtc.h>
 #include <drm/drm_fourcc.h>
+#include <drm/drm_rect.h>
 #include "intel_drv.h"
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
@@ -113,7 +114,7 @@ vlv_update_plane(struct drm_plane *dplane, struct drm_framebuffer *fb,
 	crtc_w--;
 	crtc_h--;
 
-	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size);
+	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size, true);
 
 	I915_WRITE(SPSTRIDE(pipe, plane), fb->pitches[0]);
 	I915_WRITE(SPPOS(pipe, plane), (crtc_y << 16) | crtc_x);
@@ -267,7 +268,7 @@ ivb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
 	crtc_w--;
 	crtc_h--;
 
-	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size);
+	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size, true);
 
 	/*
 	 * IVB workaround: must disable low power watermarks for at least
@@ -334,6 +335,8 @@ ivb_disable_plane(struct drm_plane *plane)
 
 	dev_priv->sprite_scaling_enabled &= ~(1 << pipe);
 
+	intel_update_sprite_watermarks(dev, pipe, 0, 0, false);
+
 	/* potentially re-enable LP watermarks */
 	if (scaling_was_enabled && !dev_priv->sprite_scaling_enabled)
 		intel_update_watermarks(dev);
@@ -452,7 +455,7 @@ ilk_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
 	crtc_w--;
 	crtc_h--;
 
-	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size);
+	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size, true);
 
 	dvsscale = 0;
 	if (IS_GEN5(dev) || crtc_w != src_w || crtc_h != src_h)
@@ -583,6 +586,20 @@ ilk_get_colorkey(struct drm_plane *plane, struct drm_intel_sprite_colorkey *key)
 		key->flags = I915_SET_COLORKEY_NONE;
 }
 
+static bool
+format_is_yuv(uint32_t format)
+{
+	switch (format) {
+	case DRM_FORMAT_YUYV:
+	case DRM_FORMAT_UYVY:
+	case DRM_FORMAT_VYUY:
+	case DRM_FORMAT_YVYU:
+		return true;
+	default:
+		return false;
+	}
+}
+
 static int
 intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		   struct drm_framebuffer *fb, int crtc_x, int crtc_y,
@@ -600,9 +617,29 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	enum transcoder cpu_transcoder = intel_pipe_to_cpu_transcoder(dev_priv,
 								      pipe);
 	int ret = 0;
-	int x = src_x >> 16, y = src_y >> 16;
-	int primary_w = crtc->mode.hdisplay, primary_h = crtc->mode.vdisplay;
 	bool disable_primary = false;
+	bool visible;
+	int hscale, vscale;
+	int max_scale, min_scale;
+	int pixel_size = drm_format_plane_cpp(fb->pixel_format, 0);
+	struct drm_rect src = {
+		/* sample coordinates in 16.16 fixed point */
+		.x1 = src_x,
+		.x2 = src_x + src_w,
+		.y1 = src_y,
+		.y2 = src_y + src_h,
+	};
+	struct drm_rect dst = {
+		/* integer pixels */
+		.x1 = crtc_x,
+		.x2 = crtc_x + crtc_w,
+		.y1 = crtc_y,
+		.y2 = crtc_y + crtc_h,
+	};
+	const struct drm_rect clip = {
+		.x2 = crtc->mode.hdisplay,
+		.y2 = crtc->mode.vdisplay,
+	};
 
 	intel_fb = to_intel_framebuffer(fb);
 	obj = intel_fb->obj;
@@ -618,19 +655,23 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	intel_plane->src_w = src_w;
 	intel_plane->src_h = src_h;
 
-	src_w = src_w >> 16;
-	src_h = src_h >> 16;
-
 	/* Pipe must be running... */
-	if (!(I915_READ(PIPECONF(cpu_transcoder)) & PIPECONF_ENABLE))
+	if (!(I915_READ(PIPECONF(cpu_transcoder)) & PIPECONF_ENABLE)) {
+		DRM_DEBUG_KMS("Pipe disabled\n");
 		return -EINVAL;
+	}
 
-	if (crtc_x >= primary_w || crtc_y >= primary_h)
+	/* Don't modify another pipe's plane */
+	if (intel_plane->pipe != intel_crtc->pipe) {
+		DRM_DEBUG_KMS("Wrong plane <-> crtc mapping\n");
 		return -EINVAL;
+	}
 
-	/* Don't modify another pipe's plane */
-	if (intel_plane->pipe != intel_crtc->pipe)
+	/* FIXME check all gen limits */
+	if (fb->width < 3 || fb->height < 3 || fb->pitches[0] > 16384) {
+		DRM_DEBUG_KMS("Unsuitable framebuffer for plane\n");
 		return -EINVAL;
+	}
 
 	/* Sprite planes can be linear or x-tiled surfaces */
 	switch (obj->tiling_mode) {
@@ -638,55 +679,123 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 		case I915_TILING_X:
 			break;
 		default:
+			DRM_DEBUG_KMS("Unsupported tiling mode\n");
 			return -EINVAL;
 	}
 
 	/*
-	 * Clamp the width & height into the visible area.  Note we don't
-	 * try to scale the source if part of the visible region is offscreen.
-	 * The caller must handle that by adjusting source offset and size.
+	 * FIXME the following code does a bunch of fuzzy adjustments to the
+	 * coordinates and sizes. We probably need some way to decide whether
+	 * more strict checking should be done instead.
 	 */
-	if ((crtc_x < 0) && ((crtc_x + crtc_w) > 0)) {
-		crtc_w += crtc_x;
-		crtc_x = 0;
-	}
-	if ((crtc_x + crtc_w) <= 0) /* Nothing to display */
-		goto out;
-	if ((crtc_x + crtc_w) > primary_w)
-		crtc_w = primary_w - crtc_x;
+	max_scale = intel_plane->max_downscale << 16;
+	min_scale = intel_plane->can_scale ? 1 : (1 << 16);
+
+	hscale = drm_rect_calc_hscale_relaxed(&src, &dst, min_scale, max_scale);
+	BUG_ON(hscale < 0);
+
+	vscale = drm_rect_calc_vscale_relaxed(&src, &dst, min_scale, max_scale);
+	BUG_ON(vscale < 0);
+
+	visible = drm_rect_clip_scaled(&src, &dst, &clip, hscale, vscale);
+
+	crtc_x = dst.x1;
+	crtc_y = dst.y1;
+	crtc_w = drm_rect_width(&dst);
+	crtc_h = drm_rect_height(&dst);
+
+	if (visible) {
+		/* check again in case clipping clamped the results */
+		hscale = drm_rect_calc_hscale(&src, &dst, min_scale, max_scale);
+		if (hscale < 0) {
+			DRM_DEBUG_KMS("Horizontal scaling factor out of limits\n");
+			drm_rect_debug_print(&src, true);
+			drm_rect_debug_print(&dst, false);
+
+			return hscale;
+		}
 
-	if ((crtc_y < 0) && ((crtc_y + crtc_h) > 0)) {
-		crtc_h += crtc_y;
-		crtc_y = 0;
+		vscale = drm_rect_calc_vscale(&src, &dst, min_scale, max_scale);
+		if (vscale < 0) {
+			DRM_DEBUG_KMS("Vertical scaling factor out of limits\n");
+			drm_rect_debug_print(&src, true);
+			drm_rect_debug_print(&dst, false);
+
+			return vscale;
+		}
+
+		/* Make the source viewport size an exact multiple of the scaling factors. */
+		drm_rect_adjust_size(&src,
+				     drm_rect_width(&dst) * hscale - drm_rect_width(&src),
+				     drm_rect_height(&dst) * vscale - drm_rect_height(&src));
+
+		/* sanity check to make sure the src viewport wasn't enlarged */
+		WARN_ON(src.x1 < (int) src_x ||
+			src.y1 < (int) src_y ||
+			src.x2 > (int) (src_x + src_w) ||
+			src.y2 > (int) (src_y + src_h));
+
+		/*
+		 * Hardware doesn't handle subpixel coordinates.
+		 * Adjust to (macro)pixel boundary, but be careful not to
+		 * increase the source viewport size, because that could
+		 * push the downscaling factor out of bounds.
+		 */
+		src_x = src.x1 >> 16;
+		src_w = drm_rect_width(&src) >> 16;
+		src_y = src.y1 >> 16;
+		src_h = drm_rect_height(&src) >> 16;
+
+		if (format_is_yuv(fb->pixel_format)) {
+			src_x &= ~1;
+			src_w &= ~1;
+
+			/*
+			 * Must keep src and dst the
+			 * same if we can't scale.
+			 */
+			if (!intel_plane->can_scale)
+				crtc_w &= ~1;
+
+			if (crtc_w == 0)
+				visible = false;
+		}
 	}
-	if ((crtc_y + crtc_h) <= 0) /* Nothing to display */
-		goto out;
-	if (crtc_y + crtc_h > primary_h)
-		crtc_h = primary_h - crtc_y;
 
-	if (!crtc_w || !crtc_h) /* Again, nothing to display */
-		goto out;
+	/* Check size restrictions when scaling */
+	if (visible && (src_w != crtc_w || src_h != crtc_h)) {
+		unsigned int width_bytes;
 
-	/*
-	 * We may not have a scaler, eg. HSW does not have it any more
-	 */
-	if (!intel_plane->can_scale && (crtc_w != src_w || crtc_h != src_h))
-		return -EINVAL;
+		WARN_ON(!intel_plane->can_scale);
 
-	/*
-	 * We can take a larger source and scale it down, but
-	 * only so much...  16x is the max on SNB.
-	 */
-	if (((src_w * src_h) / (crtc_w * crtc_h)) > intel_plane->max_downscale)
-		return -EINVAL;
+		/* FIXME interlacing min height is 6 */
+
+		if (crtc_w < 3 || crtc_h < 3)
+			visible = false;
+
+		if (src_w < 3 || src_h < 3)
+			visible = false;
+
+		width_bytes = ((src_x * pixel_size) & 63) + src_w * pixel_size;
+
+		if (src_w > 2048 || src_h > 2048 ||
+		    width_bytes > 4096 || fb->pitches[0] > 4096) {
+			DRM_DEBUG_KMS("Source dimensions exceed hardware limits\n");
+			return -EINVAL;
+		}
+	}
+
+	dst.x1 = crtc_x;
+	dst.x2 = crtc_x + crtc_w;
+	dst.y1 = crtc_y;
+	dst.y2 = crtc_y + crtc_h;
 
 	/*
 	 * If the sprite is completely covering the primary plane,
 	 * we can disable the primary and save power.
 	 */
-	if ((crtc_x == 0) && (crtc_y == 0) &&
-	    (crtc_w == primary_w) && (crtc_h == primary_h))
-		disable_primary = true;
+	disable_primary = drm_rect_equals(&dst, &clip);
+	WARN_ON(disable_primary && !visible);
 
 	mutex_lock(&dev->struct_mutex);
 
@@ -708,8 +817,12 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 	if (!disable_primary)
 		intel_enable_primary(crtc);
 
-	intel_plane->update_plane(plane, fb, obj, crtc_x, crtc_y,
-				  crtc_w, crtc_h, x, y, src_w, src_h);
+	if (visible)
+		intel_plane->update_plane(plane, fb, obj,
+					  crtc_x, crtc_y, crtc_w, crtc_h,
+					  src_x, src_y, src_w, src_h);
+	else
+		intel_plane->disable_plane(plane);
 
 	if (disable_primary)
 		intel_disable_primary(crtc);
@@ -732,7 +845,6 @@ intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
 
 out_unlock:
 	mutex_unlock(&dev->struct_mutex);
-out:
 	return ret;
 }
 
@@ -845,6 +957,14 @@ void intel_plane_restore(struct drm_plane *plane)
 			   intel_plane->src_w, intel_plane->src_h);
 }
 
+void intel_plane_disable(struct drm_plane *plane)
+{
+	if (!plane->crtc || !plane->fb)
+		return;
+
+	intel_disable_plane(plane);
+}
+
 static const struct drm_plane_funcs intel_plane_funcs = {
 	.update_plane = intel_update_plane,
 	.disable_plane = intel_disable_plane,
@@ -918,13 +1038,15 @@ intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane)
 		break;
 
 	case 7:
-		if (IS_HASWELL(dev) || IS_VALLEYVIEW(dev))
-			intel_plane->can_scale = false;
-		else
+		if (IS_IVYBRIDGE(dev)) {
 			intel_plane->can_scale = true;
+			intel_plane->max_downscale = 2;
+		} else {
+			intel_plane->can_scale = false;
+			intel_plane->max_downscale = 1;
+		}
 
 		if (IS_VALLEYVIEW(dev)) {
-			intel_plane->max_downscale = 1;
 			intel_plane->update_plane = vlv_update_plane;
 			intel_plane->disable_plane = vlv_disable_plane;
 			intel_plane->update_colorkey = vlv_update_colorkey;
@@ -933,7 +1055,6 @@ intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane)
 			plane_formats = vlv_plane_formats;
 			num_plane_formats = ARRAY_SIZE(vlv_plane_formats);
 		} else {
-			intel_plane->max_downscale = 2;
 			intel_plane->update_plane = ivb_update_plane;
 			intel_plane->disable_plane = ivb_disable_plane;
 			intel_plane->update_colorkey = ivb_update_colorkey;
diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c
index b945bc54207a..39debd80d190 100644
--- a/drivers/gpu/drm/i915/intel_tv.c
+++ b/drivers/gpu/drm/i915/intel_tv.c
@@ -914,9 +914,6 @@ intel_tv_compute_config(struct intel_encoder *encoder,
 	if (!tv_mode)
 		return false;
 
-	if (intel_encoder_check_is_cloned(&intel_tv->base))
-		return false;
-
 	pipe_config->adjusted_mode.clock = tv_mode->clock;
 	DRM_DEBUG_KMS("forcing bpc to 8 for TV\n");
 	pipe_config->pipe_bpp = 8*3;
@@ -1521,12 +1518,12 @@ static int tv_is_present_in_vbt(struct drm_device *dev)
 	struct child_device_config *p_child;
 	int i, ret;
 
-	if (!dev_priv->child_dev_num)
+	if (!dev_priv->vbt.child_dev_num)
 		return 1;
 
 	ret = 0;
-	for (i = 0; i < dev_priv->child_dev_num; i++) {
-		p_child = dev_priv->child_dev + i;
+	for (i = 0; i < dev_priv->vbt.child_dev_num; i++) {
+		p_child = dev_priv->vbt.child_dev + i;
 		/*
 		 * If the device type is not TV, continue.
 		 */
@@ -1564,7 +1561,7 @@ intel_tv_init(struct drm_device *dev)
 		return;
 	}
 	/* Even if we have an encoder we may not have a connector */
-	if (!dev_priv->int_tv_support)
+	if (!dev_priv->vbt.int_tv_support)
 		return;
 
 	/*
diff --git a/drivers/gpu/drm/mgag200/Makefile b/drivers/gpu/drm/mgag200/Makefile
index 7db592eedbf1..a9a0300f09fc 100644
--- a/drivers/gpu/drm/mgag200/Makefile
+++ b/drivers/gpu/drm/mgag200/Makefile
@@ -1,5 +1,5 @@
 ccflags-y := -Iinclude/drm
-mgag200-y   := mgag200_main.o mgag200_mode.o \
+mgag200-y   := mgag200_main.o mgag200_mode.o mgag200_cursor.o \
 	mgag200_drv.o mgag200_fb.o mgag200_i2c.o mgag200_ttm.o
 
 obj-$(CONFIG_DRM_MGAG200) += mgag200.o
diff --git a/drivers/gpu/drm/mgag200/mgag200_cursor.c b/drivers/gpu/drm/mgag200/mgag200_cursor.c
new file mode 100644
index 000000000000..801731aeab61
--- /dev/null
+++ b/drivers/gpu/drm/mgag200/mgag200_cursor.c
@@ -0,0 +1,275 @@
+/*
+ * Copyright 2013 Matrox Graphics
+ *
+ * This file is subject to the terms and conditions of the GNU General
+ * Public License version 2. See the file COPYING in the main
+ * directory of this archive for more details.
+ *
+ * Author: Christopher Harvey <charvey@matrox.com>
+ */
+
+#include <drm/drmP.h>
+#include "mgag200_drv.h"
+
+static bool warn_transparent = true;
+static bool warn_palette = true;
+
+/*
+  Hide the cursor off screen. We can't disable the cursor hardware because it
+  takes too long to re-activate and causes momentary corruption
+*/
+static void mga_hide_cursor(struct mga_device *mdev)
+{
+	WREG8(MGA_CURPOSXL, 0);
+	WREG8(MGA_CURPOSXH, 0);
+	mgag200_bo_unpin(mdev->cursor.pixels_1);
+	mgag200_bo_unpin(mdev->cursor.pixels_2);
+}
+
+int mga_crtc_cursor_set(struct drm_crtc *crtc,
+			struct drm_file *file_priv,
+			uint32_t handle,
+			uint32_t width,
+			uint32_t height)
+{
+	struct drm_device *dev = (struct drm_device *)file_priv->minor->dev;
+	struct mga_device *mdev = (struct mga_device *)dev->dev_private;
+	struct mgag200_bo *pixels_1 = mdev->cursor.pixels_1;
+	struct mgag200_bo *pixels_2 = mdev->cursor.pixels_2;
+	struct mgag200_bo *pixels_current = mdev->cursor.pixels_current;
+	struct mgag200_bo *pixels_prev = mdev->cursor.pixels_prev;
+	struct drm_gem_object *obj;
+	struct mgag200_bo *bo = NULL;
+	int ret = 0;
+	unsigned int i, row, col;
+	uint32_t colour_set[16];
+	uint32_t *next_space = &colour_set[0];
+	uint32_t *palette_iter;
+	uint32_t this_colour;
+	bool found = false;
+	int colour_count = 0;
+	u64 gpu_addr;
+	u8 reg_index;
+	u8 this_row[48];
+
+	if (!pixels_1 || !pixels_2) {
+		WREG8(MGA_CURPOSXL, 0);
+		WREG8(MGA_CURPOSXH, 0);
+		return -ENOTSUPP; /* Didn't allocate space for cursors */
+	}
+
+	if ((width != 64 || height != 64) && handle) {
+		WREG8(MGA_CURPOSXL, 0);
+		WREG8(MGA_CURPOSXH, 0);
+		return -EINVAL;
+	}
+
+	BUG_ON(pixels_1 != pixels_current && pixels_1 != pixels_prev);
+	BUG_ON(pixels_2 != pixels_current && pixels_2 != pixels_prev);
+	BUG_ON(pixels_current == pixels_prev);
+
+	ret = mgag200_bo_reserve(pixels_1, true);
+	if (ret) {
+		WREG8(MGA_CURPOSXL, 0);
+		WREG8(MGA_CURPOSXH, 0);
+		return ret;
+	}
+	ret = mgag200_bo_reserve(pixels_2, true);
+	if (ret) {
+		WREG8(MGA_CURPOSXL, 0);
+		WREG8(MGA_CURPOSXH, 0);
+		mgag200_bo_unreserve(pixels_1);
+		return ret;
+	}
+
+	if (!handle) {
+		mga_hide_cursor(mdev);
+		ret = 0;
+		goto out1;
+	}
+
+	/* Move cursor buffers into VRAM if they aren't already */
+	if (!pixels_1->pin_count) {
+		ret = mgag200_bo_pin(pixels_1, TTM_PL_FLAG_VRAM,
+				     &mdev->cursor.pixels_1_gpu_addr);
+		if (ret)
+			goto out1;
+	}
+	if (!pixels_2->pin_count) {
+		ret = mgag200_bo_pin(pixels_2, TTM_PL_FLAG_VRAM,
+				     &mdev->cursor.pixels_2_gpu_addr);
+		if (ret) {
+			mgag200_bo_unpin(pixels_1);
+			goto out1;
+		}
+	}
+
+	mutex_lock(&dev->struct_mutex);
+	obj = drm_gem_object_lookup(dev, file_priv, handle);
+	if (!obj) {
+		mutex_unlock(&dev->struct_mutex);
+		ret = -ENOENT;
+		goto out1;
+	}
+	drm_gem_object_unreference(obj);
+	mutex_unlock(&dev->struct_mutex);
+
+	bo = gem_to_mga_bo(obj);
+	ret = mgag200_bo_reserve(bo, true);
+	if (ret) {
+		dev_err(&dev->pdev->dev, "failed to reserve user bo\n");
+		goto out1;
+	}
+	if (!bo->kmap.virtual) {
+		ret = ttm_bo_kmap(&bo->bo, 0, bo->bo.num_pages, &bo->kmap);
+		if (ret) {
+			dev_err(&dev->pdev->dev, "failed to kmap user buffer updates\n");
+			goto out2;
+		}
+	}
+
+	memset(&colour_set[0], 0, sizeof(uint32_t)*16);
+	/* width*height*4 = 16384 */
+	for (i = 0; i < 16384; i += 4) {
+		this_colour = ioread32(bo->kmap.virtual + i);
+		/* No transparency */
+		if (this_colour>>24 != 0xff &&
+			this_colour>>24 != 0x0) {
+			if (warn_transparent) {
+				dev_info(&dev->pdev->dev, "Video card doesn't support cursors with partial transparency.\n");
+				dev_info(&dev->pdev->dev, "Not enabling hardware cursor.\n");
+				warn_transparent = false; /* Only tell the user once. */
+			}
+			ret = -EINVAL;
+			goto out3;
+		}
+		/* Don't need to store transparent pixels as colours */
+		if (this_colour>>24 == 0x0)
+			continue;
+		found = false;
+		for (palette_iter = &colour_set[0]; palette_iter != next_space; palette_iter++) {
+			if (*palette_iter == this_colour) {
+				found = true;
+				break;
+			}
+		}
+		if (found)
+			continue;
+		/* We only support 4bit paletted cursors */
+		if (colour_count >= 16) {
+			if (warn_palette) {
+				dev_info(&dev->pdev->dev, "Video card only supports cursors with up to 16 colours.\n");
+				dev_info(&dev->pdev->dev, "Not enabling hardware cursor.\n");
+				warn_palette = false; /* Only tell the user once. */
+			}
+			ret = -EINVAL;
+			goto out3;
+		}
+		*next_space = this_colour;
+		next_space++;
+		colour_count++;
+	}
+
+	/* Program colours from cursor icon into palette */
+	for (i = 0; i < colour_count; i++) {
+		if (i <= 2)
+			reg_index = 0x8 + i*0x4;
+		else
+			reg_index = 0x60 + i*0x3;
+		WREG_DAC(reg_index, colour_set[i] & 0xff);
+		WREG_DAC(reg_index+1, colour_set[i]>>8 & 0xff);
+		WREG_DAC(reg_index+2, colour_set[i]>>16 & 0xff);
+		BUG_ON((colour_set[i]>>24 & 0xff) != 0xff);
+	}
+
+	/* Map up-coming buffer to write colour indices */
+	if (!pixels_prev->kmap.virtual) {
+		ret = ttm_bo_kmap(&pixels_prev->bo, 0,
+				  pixels_prev->bo.num_pages,
+				  &pixels_prev->kmap);
+		if (ret) {
+			dev_err(&dev->pdev->dev, "failed to kmap cursor updates\n");
+			goto out3;
+		}
+	}
+
+	/* now write colour indices into hardware cursor buffer */
+	for (row = 0; row < 64; row++) {
+		memset(&this_row[0], 0, 48);
+		for (col = 0; col < 64; col++) {
+			this_colour = ioread32(bo->kmap.virtual + 4*(col + 64*row));
+			/* write transparent pixels */
+			if (this_colour>>24 == 0x0) {
+				this_row[47 - col/8] |= 0x80>>(col%8);
+				continue;
+			}
+
+			/* write colour index here */
+			for (i = 0; i < colour_count; i++) {
+				if (colour_set[i] == this_colour) {
+					if (col % 2)
+						this_row[col/2] |= i<<4;
+					else
+						this_row[col/2] |= i;
+					break;
+				}
+			}
+		}
+		memcpy_toio(pixels_prev->kmap.virtual + row*48, &this_row[0], 48);
+	}
+
+	/* Program gpu address of cursor buffer */
+	if (pixels_prev == pixels_1)
+		gpu_addr = mdev->cursor.pixels_1_gpu_addr;
+	else
+		gpu_addr = mdev->cursor.pixels_2_gpu_addr;
+	WREG_DAC(MGA1064_CURSOR_BASE_ADR_LOW, (u8)((gpu_addr>>10) & 0xff));
+	WREG_DAC(MGA1064_CURSOR_BASE_ADR_HI, (u8)((gpu_addr>>18) & 0x3f));
+
+	/* Adjust cursor control register to turn on the cursor */
+	WREG_DAC(MGA1064_CURSOR_CTL, 4); /* 16-colour palletized cursor mode */
+
+	/* Now swap internal buffer pointers */
+	if (mdev->cursor.pixels_1 == mdev->cursor.pixels_prev) {
+		mdev->cursor.pixels_prev = mdev->cursor.pixels_2;
+		mdev->cursor.pixels_current = mdev->cursor.pixels_1;
+	} else if (mdev->cursor.pixels_1 == mdev->cursor.pixels_current) {
+		mdev->cursor.pixels_prev = mdev->cursor.pixels_1;
+		mdev->cursor.pixels_current = mdev->cursor.pixels_2;
+	} else {
+		BUG();
+	}
+	ret = 0;
+
+	ttm_bo_kunmap(&pixels_prev->kmap);
+ out3:
+	ttm_bo_kunmap(&bo->kmap);
+ out2:
+	mgag200_bo_unreserve(bo);
+ out1:
+	if (ret)
+		mga_hide_cursor(mdev);
+	mgag200_bo_unreserve(pixels_1);
+	mgag200_bo_unreserve(pixels_2);
+	return ret;
+}
+
+int mga_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
+{
+	struct mga_device *mdev = (struct mga_device *)crtc->dev->dev_private;
+	/* Our origin is at (64,64) */
+	x += 64;
+	y += 64;
+
+	BUG_ON(x <= 0);
+	BUG_ON(y <= 0);
+	BUG_ON(x & ~0xffff);
+	BUG_ON(y & ~0xffff);
+
+	WREG8(MGA_CURPOSXL, x & 0xff);
+	WREG8(MGA_CURPOSXH, (x>>8) & 0xff);
+
+	WREG8(MGA_CURPOSYL, y & 0xff);
+	WREG8(MGA_CURPOSYH, (y>>8) & 0xff);
+	return 0;
+}
diff --git a/drivers/gpu/drm/mgag200/mgag200_drv.h b/drivers/gpu/drm/mgag200/mgag200_drv.h
index bf29b2f4d68d..12e2499d9352 100644
--- a/drivers/gpu/drm/mgag200/mgag200_drv.h
+++ b/drivers/gpu/drm/mgag200/mgag200_drv.h
@@ -149,6 +149,21 @@ struct mga_connector {
 	struct mga_i2c_chan *i2c;
 };
 
+struct mga_cursor {
+	/*
+	   We have to have 2 buffers for the cursor to avoid occasional
+	   corruption while switching cursor icons.
+	   If either of these is NULL, then don't do hardware cursors, and
+	   fall back to software.
+	*/
+	struct mgag200_bo *pixels_1;
+	struct mgag200_bo *pixels_2;
+	u64 pixels_1_gpu_addr, pixels_2_gpu_addr;
+	/* The currently displayed icon, this points to one of pixels_1, or pixels_2 */
+	struct mgag200_bo *pixels_current;
+	/* The previously displayed icon */
+	struct mgag200_bo *pixels_prev;
+};
 
 struct mga_mc {
 	resource_size_t			vram_size;
@@ -181,6 +196,7 @@ struct mga_device {
 	struct mga_mode_info		mode_info;
 
 	struct mga_fbdev *mfbdev;
+	struct mga_cursor cursor;
 
 	bool				suspended;
 	int				num_crtc;
@@ -198,7 +214,8 @@ struct mga_device {
 		struct ttm_bo_device bdev;
 	} ttm;
 
-	u32 reg_1e24; /* SE model number */
+	/* SE model number stored in reg 0x1e24 */
+	u32 unique_rev_id;
 };
 
 
@@ -263,8 +280,24 @@ void mgag200_i2c_destroy(struct mga_i2c_chan *i2c);
 #define DRM_FILE_PAGE_OFFSET (0x100000000ULL >> PAGE_SHIFT)
 void mgag200_ttm_placement(struct mgag200_bo *bo, int domain);
 
-int mgag200_bo_reserve(struct mgag200_bo *bo, bool no_wait);
-void mgag200_bo_unreserve(struct mgag200_bo *bo);
+static inline int mgag200_bo_reserve(struct mgag200_bo *bo, bool no_wait)
+{
+	int ret;
+
+	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
+	if (ret) {
+		if (ret != -ERESTARTSYS && ret != -EBUSY)
+			DRM_ERROR("reserve failed %p\n", bo);
+		return ret;
+	}
+	return 0;
+}
+
+static inline void mgag200_bo_unreserve(struct mgag200_bo *bo)
+{
+	ttm_bo_unreserve(&bo->bo);
+}
+
 int mgag200_bo_create(struct drm_device *dev, int size, int align,
 		      uint32_t flags, struct mgag200_bo **pastbo);
 int mgag200_mm_init(struct mga_device *mdev);
@@ -273,4 +306,9 @@ int mgag200_mmap(struct file *filp, struct vm_area_struct *vma);
 int mgag200_bo_pin(struct mgag200_bo *bo, u32 pl_flag, u64 *gpu_addr);
 int mgag200_bo_unpin(struct mgag200_bo *bo);
 int mgag200_bo_push_sysram(struct mgag200_bo *bo);
+			   /* mgag200_cursor.c */
+int mga_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
+						uint32_t handle, uint32_t width, uint32_t height);
+int mga_crtc_cursor_move(struct drm_crtc *crtc, int x, int y);
+
 #endif				/* __MGAG200_DRV_H__ */
diff --git a/drivers/gpu/drm/mgag200/mgag200_fb.c b/drivers/gpu/drm/mgag200/mgag200_fb.c
index 5da824ce9ba1..964f58cee5ea 100644
--- a/drivers/gpu/drm/mgag200/mgag200_fb.c
+++ b/drivers/gpu/drm/mgag200/mgag200_fb.c
@@ -27,7 +27,7 @@ static void mga_dirty_update(struct mga_fbdev *mfbdev,
 	struct mgag200_bo *bo;
 	int src_offset, dst_offset;
 	int bpp = (mfbdev->mfb.base.bits_per_pixel + 7)/8;
-	int ret;
+	int ret = -EBUSY;
 	bool unmap = false;
 	bool store_for_later = false;
 	int x2, y2;
@@ -41,7 +41,8 @@ static void mga_dirty_update(struct mga_fbdev *mfbdev,
 	 * then the BO is being moved and we should
 	 * store up the damage until later.
 	 */
-	ret = mgag200_bo_reserve(bo, true);
+	if (!in_interrupt())
+		ret = mgag200_bo_reserve(bo, true);
 	if (ret) {
 		if (ret != -EBUSY)
 			return;
diff --git a/drivers/gpu/drm/mgag200/mgag200_main.c b/drivers/gpu/drm/mgag200/mgag200_main.c
index 99059237da38..9fa5685baee0 100644
--- a/drivers/gpu/drm/mgag200/mgag200_main.c
+++ b/drivers/gpu/drm/mgag200/mgag200_main.c
@@ -176,7 +176,7 @@ static int mgag200_device_init(struct drm_device *dev,
 
 	/* stash G200 SE model number for later use */
 	if (IS_G200_SE(mdev))
-		mdev->reg_1e24 = RREG32(0x1e24);
+		mdev->unique_rev_id = RREG32(0x1e24);
 
 	ret = mga_vram_init(mdev);
 	if (ret)
@@ -209,7 +209,7 @@ int mgag200_driver_load(struct drm_device *dev, unsigned long flags)
 	r = mgag200_device_init(dev, flags);
 	if (r) {
 		dev_err(&dev->pdev->dev, "Fatal error during GPU init: %d\n", r);
-		goto out;
+		return r;
 	}
 	r = mgag200_mm_init(mdev);
 	if (r)
@@ -221,8 +221,27 @@ int mgag200_driver_load(struct drm_device *dev, unsigned long flags)
 	dev->mode_config.prefer_shadow = 1;
 
 	r = mgag200_modeset_init(mdev);
-	if (r)
+	if (r) {
 		dev_err(&dev->pdev->dev, "Fatal error during modeset init: %d\n", r);
+		goto out;
+	}
+
+	/* Make small buffers to store a hardware cursor (double buffered icon updates) */
+	mgag200_bo_create(dev, roundup(48*64, PAGE_SIZE), 0, 0,
+					  &mdev->cursor.pixels_1);
+	mgag200_bo_create(dev, roundup(48*64, PAGE_SIZE), 0, 0,
+					  &mdev->cursor.pixels_2);
+	if (!mdev->cursor.pixels_2 || !mdev->cursor.pixels_1)
+		goto cursor_nospace;
+	mdev->cursor.pixels_current = mdev->cursor.pixels_1;
+	mdev->cursor.pixels_prev = mdev->cursor.pixels_2;
+	goto cursor_done;
+ cursor_nospace:
+	mdev->cursor.pixels_1 = NULL;
+	mdev->cursor.pixels_2 = NULL;
+	dev_warn(&dev->pdev->dev, "Could not allocate space for cursors. Not doing hardware cursors.\n");
+ cursor_done:
+
 out:
 	if (r)
 		mgag200_driver_unload(dev);
diff --git a/drivers/gpu/drm/mgag200/mgag200_mode.c b/drivers/gpu/drm/mgag200/mgag200_mode.c
index ee66badc8bb6..251784aa2225 100644
--- a/drivers/gpu/drm/mgag200/mgag200_mode.c
+++ b/drivers/gpu/drm/mgag200/mgag200_mode.c
@@ -1008,7 +1008,7 @@ static int mga_crtc_mode_set(struct drm_crtc *crtc,
 
 
 	if (IS_G200_SE(mdev)) {
-		if (mdev->reg_1e24 >= 0x02) {
+		if (mdev->unique_rev_id >= 0x02) {
 			u8 hi_pri_lvl;
 			u32 bpp;
 			u32 mb;
@@ -1038,7 +1038,7 @@ static int mga_crtc_mode_set(struct drm_crtc *crtc,
 			WREG8(MGAREG_CRTCEXT_DATA, hi_pri_lvl);
 		} else {
 			WREG8(MGAREG_CRTCEXT_INDEX, 0x06);
-			if (mdev->reg_1e24 >= 0x01)
+			if (mdev->unique_rev_id >= 0x01)
 				WREG8(MGAREG_CRTCEXT_DATA, 0x03);
 			else
 				WREG8(MGAREG_CRTCEXT_DATA, 0x04);
@@ -1253,6 +1253,8 @@ static void mga_crtc_destroy(struct drm_crtc *crtc)
 
 /* These provide the minimum set of functions required to handle a CRTC */
 static const struct drm_crtc_funcs mga_crtc_funcs = {
+	.cursor_set = mga_crtc_cursor_set,
+	.cursor_move = mga_crtc_cursor_move,
 	.gamma_set = mga_crtc_gamma_set,
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = mga_crtc_destroy,
@@ -1410,6 +1412,32 @@ static int mga_vga_get_modes(struct drm_connector *connector)
 	return ret;
 }
 
+static uint32_t mga_vga_calculate_mode_bandwidth(struct drm_display_mode *mode,
+							int bits_per_pixel)
+{
+	uint32_t total_area, divisor;
+	int64_t active_area, pixels_per_second, bandwidth;
+	uint64_t bytes_per_pixel = (bits_per_pixel + 7) / 8;
+
+	divisor = 1024;
+
+	if (!mode->htotal || !mode->vtotal || !mode->clock)
+		return 0;
+
+	active_area = mode->hdisplay * mode->vdisplay;
+	total_area = mode->htotal * mode->vtotal;
+
+	pixels_per_second = active_area * mode->clock * 1000;
+	do_div(pixels_per_second, total_area);
+
+	bandwidth = pixels_per_second * bytes_per_pixel * 100;
+	do_div(bandwidth, divisor);
+
+	return (uint32_t)(bandwidth);
+}
+
+#define MODE_BANDWIDTH	MODE_BAD
+
 static int mga_vga_mode_valid(struct drm_connector *connector,
 				 struct drm_display_mode *mode)
 {
@@ -1421,7 +1449,45 @@ static int mga_vga_mode_valid(struct drm_connector *connector,
 	int bpp = 32;
 	int i = 0;
 
-	/* FIXME: Add bandwidth and g200se limitations */
+	if (IS_G200_SE(mdev)) {
+		if (mdev->unique_rev_id == 0x01) {
+			if (mode->hdisplay > 1600)
+				return MODE_VIRTUAL_X;
+			if (mode->vdisplay > 1200)
+				return MODE_VIRTUAL_Y;
+			if (mga_vga_calculate_mode_bandwidth(mode, bpp)
+				> (24400 * 1024))
+				return MODE_BANDWIDTH;
+		} else if (mdev->unique_rev_id >= 0x02) {
+			if (mode->hdisplay > 1920)
+				return MODE_VIRTUAL_X;
+			if (mode->vdisplay > 1200)
+				return MODE_VIRTUAL_Y;
+			if (mga_vga_calculate_mode_bandwidth(mode, bpp)
+				> (30100 * 1024))
+				return MODE_BANDWIDTH;
+		}
+	} else if (mdev->type == G200_WB) {
+		if (mode->hdisplay > 1280)
+			return MODE_VIRTUAL_X;
+		if (mode->vdisplay > 1024)
+			return MODE_VIRTUAL_Y;
+		if (mga_vga_calculate_mode_bandwidth(mode,
+			bpp > (31877 * 1024)))
+			return MODE_BANDWIDTH;
+	} else if (mdev->type == G200_EV &&
+		(mga_vga_calculate_mode_bandwidth(mode, bpp)
+			> (32700 * 1024))) {
+		return MODE_BANDWIDTH;
+	} else if (mode->type == G200_EH &&
+		(mga_vga_calculate_mode_bandwidth(mode, bpp)
+			> (37500 * 1024))) {
+		return MODE_BANDWIDTH;
+	} else if (mode->type == G200_ER &&
+		(mga_vga_calculate_mode_bandwidth(mode,
+			bpp) > (55000 * 1024))) {
+		return MODE_BANDWIDTH;
+	}
 
 	if (mode->crtc_hdisplay > 2048 || mode->crtc_hsync_start > 4096 ||
 	    mode->crtc_hsync_end > 4096 || mode->crtc_htotal > 4096 ||
diff --git a/drivers/gpu/drm/mgag200/mgag200_reg.h b/drivers/gpu/drm/mgag200/mgag200_reg.h
index fb24d8655feb..3ae442a64bd6 100644
--- a/drivers/gpu/drm/mgag200/mgag200_reg.h
+++ b/drivers/gpu/drm/mgag200/mgag200_reg.h
@@ -235,7 +235,11 @@
 #define MGAREG_CRTCEXT_INDEX	0x1fde
 #define MGAREG_CRTCEXT_DATA	0x1fdf
 
-
+/* Cursor X and Y position */
+#define MGA_CURPOSXL 0x3c0c
+#define MGA_CURPOSXH 0x3c0d
+#define MGA_CURPOSYL 0x3c0e
+#define MGA_CURPOSYH 0x3c0f
 
 /* MGA bits for registers PCI_OPTION_REG */
 #define MGA1064_OPT_SYS_CLK_PCI   		( 0x00 << 0 )
diff --git a/drivers/gpu/drm/mgag200/mgag200_ttm.c b/drivers/gpu/drm/mgag200/mgag200_ttm.c
index 401c9891d3a8..3acb2b044c7b 100644
--- a/drivers/gpu/drm/mgag200/mgag200_ttm.c
+++ b/drivers/gpu/drm/mgag200/mgag200_ttm.c
@@ -270,26 +270,20 @@ int mgag200_mm_init(struct mga_device *mdev)
 		return ret;
 	}
 
-	mdev->fb_mtrr = drm_mtrr_add(pci_resource_start(dev->pdev, 0),
-				    pci_resource_len(dev->pdev, 0),
-				    DRM_MTRR_WC);
+	mdev->fb_mtrr = arch_phys_wc_add(pci_resource_start(dev->pdev, 0),
+					 pci_resource_len(dev->pdev, 0));
 
 	return 0;
 }
 
 void mgag200_mm_fini(struct mga_device *mdev)
 {
-	struct drm_device *dev = mdev->dev;
 	ttm_bo_device_release(&mdev->ttm.bdev);
 
 	mgag200_ttm_global_release(mdev);
 
-	if (mdev->fb_mtrr >= 0) {
-		drm_mtrr_del(mdev->fb_mtrr,
-			     pci_resource_start(dev->pdev, 0),
-			     pci_resource_len(dev->pdev, 0), DRM_MTRR_WC);
-		mdev->fb_mtrr = -1;
-	}
+	arch_phys_wc_del(mdev->fb_mtrr);
+	mdev->fb_mtrr = 0;
 }
 
 void mgag200_ttm_placement(struct mgag200_bo *bo, int domain)
@@ -309,24 +303,6 @@ void mgag200_ttm_placement(struct mgag200_bo *bo, int domain)
 	bo->placement.num_busy_placement = c;
 }
 
-int mgag200_bo_reserve(struct mgag200_bo *bo, bool no_wait)
-{
-	int ret;
-
-	ret = ttm_bo_reserve(&bo->bo, true, no_wait, false, 0);
-	if (ret) {
-		if (ret != -ERESTARTSYS && ret != -EBUSY)
-			DRM_ERROR("reserve failed %p %d\n", bo, ret);
-		return ret;
-	}
-	return 0;
-}
-
-void mgag200_bo_unreserve(struct mgag200_bo *bo)
-{
-	ttm_bo_unreserve(&bo->bo);
-}
-
 int mgag200_bo_create(struct drm_device *dev, int size, int align,
 		  uint32_t flags, struct mgag200_bo **pmgabo)
 {
diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig
index a7ff6d5a34b9..ff80f12480ea 100644
--- a/drivers/gpu/drm/nouveau/Kconfig
+++ b/drivers/gpu/drm/nouveau/Kconfig
@@ -15,6 +15,13 @@ config DRM_NOUVEAU
 	select ACPI_WMI if ACPI && X86
 	select MXM_WMI if ACPI && X86
 	select POWER_SUPPLY
+	# Similar to i915, we need to select ACPI_VIDEO and it's dependencies
+	select BACKLIGHT_LCD_SUPPORT if ACPI && X86
+	select BACKLIGHT_CLASS_DEVICE if ACPI && X86
+	select VIDEO_OUTPUT_CONTROL if ACPI && X86
+	select INPUT if ACPI && X86
+	select THERMAL if ACPI && X86
+	select ACPI_VIDEO if ACPI && X86
 	help
 	  Choose this option for open-source nVidia support.
 
diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 998e8b4444f3..d939a1da3203 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -12,7 +12,6 @@ nouveau-y += core/core/engctx.o
 nouveau-y += core/core/engine.o
 nouveau-y += core/core/enum.o
 nouveau-y += core/core/event.o
-nouveau-y += core/core/falcon.o
 nouveau-y += core/core/gpuobj.o
 nouveau-y += core/core/handle.o
 nouveau-y += core/core/mm.o
@@ -60,6 +59,8 @@ nouveau-y += core/subdev/devinit/nv10.o
 nouveau-y += core/subdev/devinit/nv1a.o
 nouveau-y += core/subdev/devinit/nv20.o
 nouveau-y += core/subdev/devinit/nv50.o
+nouveau-y += core/subdev/devinit/nva3.o
+nouveau-y += core/subdev/devinit/nvc0.o
 nouveau-y += core/subdev/fb/base.o
 nouveau-y += core/subdev/fb/nv04.o
 nouveau-y += core/subdev/fb/nv10.o
@@ -78,6 +79,17 @@ nouveau-y += core/subdev/fb/nv49.o
 nouveau-y += core/subdev/fb/nv4e.o
 nouveau-y += core/subdev/fb/nv50.o
 nouveau-y += core/subdev/fb/nvc0.o
+nouveau-y += core/subdev/fb/ramnv04.o
+nouveau-y += core/subdev/fb/ramnv10.o
+nouveau-y += core/subdev/fb/ramnv1a.o
+nouveau-y += core/subdev/fb/ramnv20.o
+nouveau-y += core/subdev/fb/ramnv40.o
+nouveau-y += core/subdev/fb/ramnv41.o
+nouveau-y += core/subdev/fb/ramnv44.o
+nouveau-y += core/subdev/fb/ramnv49.o
+nouveau-y += core/subdev/fb/ramnv4e.o
+nouveau-y += core/subdev/fb/ramnv50.o
+nouveau-y += core/subdev/fb/ramnvc0.o
 nouveau-y += core/subdev/gpio/base.o
 nouveau-y += core/subdev/gpio/nv10.o
 nouveau-y += core/subdev/gpio/nv50.o
@@ -129,12 +141,15 @@ nouveau-y += core/subdev/vm/nv44.o
 nouveau-y += core/subdev/vm/nv50.o
 nouveau-y += core/subdev/vm/nvc0.o
 
+nouveau-y += core/engine/falcon.o
+nouveau-y += core/engine/xtensa.o
 nouveau-y += core/engine/dmaobj/base.o
 nouveau-y += core/engine/dmaobj/nv04.o
 nouveau-y += core/engine/dmaobj/nv50.o
 nouveau-y += core/engine/dmaobj/nvc0.o
 nouveau-y += core/engine/dmaobj/nvd0.o
 nouveau-y += core/engine/bsp/nv84.o
+nouveau-y += core/engine/bsp/nv98.o
 nouveau-y += core/engine/bsp/nvc0.o
 nouveau-y += core/engine/bsp/nve0.o
 nouveau-y += core/engine/copy/nva3.o
@@ -185,7 +200,13 @@ nouveau-y += core/engine/fifo/nve0.o
 nouveau-y += core/engine/graph/ctxnv40.o
 nouveau-y += core/engine/graph/ctxnv50.o
 nouveau-y += core/engine/graph/ctxnvc0.o
-nouveau-y += core/engine/graph/ctxnve0.o
+nouveau-y += core/engine/graph/ctxnvc1.o
+nouveau-y += core/engine/graph/ctxnvc3.o
+nouveau-y += core/engine/graph/ctxnvc8.o
+nouveau-y += core/engine/graph/ctxnvd7.o
+nouveau-y += core/engine/graph/ctxnvd9.o
+nouveau-y += core/engine/graph/ctxnve4.o
+nouveau-y += core/engine/graph/ctxnvf0.o
 nouveau-y += core/engine/graph/nv04.o
 nouveau-y += core/engine/graph/nv10.o
 nouveau-y += core/engine/graph/nv20.o
@@ -197,7 +218,13 @@ nouveau-y += core/engine/graph/nv35.o
 nouveau-y += core/engine/graph/nv40.o
 nouveau-y += core/engine/graph/nv50.o
 nouveau-y += core/engine/graph/nvc0.o
-nouveau-y += core/engine/graph/nve0.o
+nouveau-y += core/engine/graph/nvc1.o
+nouveau-y += core/engine/graph/nvc3.o
+nouveau-y += core/engine/graph/nvc8.o
+nouveau-y += core/engine/graph/nvd7.o
+nouveau-y += core/engine/graph/nvd9.o
+nouveau-y += core/engine/graph/nve4.o
+nouveau-y += core/engine/graph/nvf0.o
 nouveau-y += core/engine/mpeg/nv31.o
 nouveau-y += core/engine/mpeg/nv40.o
 nouveau-y += core/engine/mpeg/nv50.o
@@ -209,6 +236,7 @@ nouveau-y += core/engine/software/nv10.o
 nouveau-y += core/engine/software/nv50.o
 nouveau-y += core/engine/software/nvc0.o
 nouveau-y += core/engine/vp/nv84.o
+nouveau-y += core/engine/vp/nv98.o
 nouveau-y += core/engine/vp/nvc0.o
 nouveau-y += core/engine/vp/nve0.o
 
diff --git a/drivers/gpu/drm/nouveau/core/core/mm.c b/drivers/gpu/drm/nouveau/core/core/mm.c
index 0261a11b2ae0..d8291724dbd4 100644
--- a/drivers/gpu/drm/nouveau/core/core/mm.c
+++ b/drivers/gpu/drm/nouveau/core/core/mm.c
@@ -208,7 +208,6 @@ nouveau_mm_init(struct nouveau_mm *mm, u32 offset, u32 length, u32 block)
 	struct nouveau_mm_node *node;
 
 	if (block) {
-		mutex_init(&mm->mutex);
 		INIT_LIST_HEAD(&mm->nodes);
 		INIT_LIST_HEAD(&mm->free);
 		mm->block_size = block;
diff --git a/drivers/gpu/drm/nouveau/core/engine/bsp/nv84.c b/drivers/gpu/drm/nouveau/core/engine/bsp/nv84.c
index 1d9f614cb97d..1e8e75c0684a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/bsp/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/bsp/nv84.c
@@ -19,24 +19,19 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
- * Authors: Ben Skeggs
+ * Authors: Ben Skeggs, Ilia Mirkin
  */
 
-#include <core/engctx.h>
-#include <core/class.h>
-
+#include <engine/xtensa.h>
 #include <engine/bsp.h>
 
-struct nv84_bsp_priv {
-	struct nouveau_engine base;
-};
-
 /*******************************************************************************
  * BSP object classes
  ******************************************************************************/
 
 static struct nouveau_oclass
 nv84_bsp_sclass[] = {
+	{ 0x74b0, &nouveau_object_ofuncs },
 	{},
 };
 
@@ -48,7 +43,7 @@ static struct nouveau_oclass
 nv84_bsp_cclass = {
 	.handle = NV_ENGCTX(BSP, 0x84),
 	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = _nouveau_engctx_ctor,
+		.ctor = _nouveau_xtensa_engctx_ctor,
 		.dtor = _nouveau_engctx_dtor,
 		.init = _nouveau_engctx_init,
 		.fini = _nouveau_engctx_fini,
@@ -66,10 +61,10 @@ nv84_bsp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	      struct nouveau_oclass *oclass, void *data, u32 size,
 	      struct nouveau_object **pobject)
 {
-	struct nv84_bsp_priv *priv;
+	struct nouveau_xtensa *priv;
 	int ret;
 
-	ret = nouveau_engine_create(parent, engine, oclass, true,
+	ret = nouveau_xtensa_create(parent, engine, oclass, 0x103000, true,
 				    "PBSP", "bsp", &priv);
 	*pobject = nv_object(priv);
 	if (ret)
@@ -78,6 +73,8 @@ nv84_bsp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->unit = 0x04008000;
 	nv_engine(priv)->cclass = &nv84_bsp_cclass;
 	nv_engine(priv)->sclass = nv84_bsp_sclass;
+	priv->fifo_val = 0x1111;
+	priv->unkd28 = 0x90044;
 	return 0;
 }
 
@@ -86,8 +83,10 @@ nv84_bsp_oclass = {
 	.handle = NV_ENGINE(BSP, 0x84),
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv84_bsp_ctor,
-		.dtor = _nouveau_engine_dtor,
-		.init = _nouveau_engine_init,
-		.fini = _nouveau_engine_fini,
+		.dtor = _nouveau_xtensa_dtor,
+		.init = _nouveau_xtensa_init,
+		.fini = _nouveau_xtensa_fini,
+		.rd32 = _nouveau_xtensa_rd32,
+		.wr32 = _nouveau_xtensa_wr32,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/engine/bsp/nv98.c b/drivers/gpu/drm/nouveau/core/engine/bsp/nv98.c
new file mode 100644
index 000000000000..8bf92b0e6d82
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/bsp/nv98.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <core/engctx.h>
+#include <core/class.h>
+
+#include <engine/bsp.h>
+
+struct nv98_bsp_priv {
+	struct nouveau_engine base;
+};
+
+/*******************************************************************************
+ * BSP object classes
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nv98_bsp_sclass[] = {
+	{},
+};
+
+/*******************************************************************************
+ * BSP context
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nv98_bsp_cclass = {
+	.handle = NV_ENGCTX(BSP, 0x98),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nouveau_engctx_ctor,
+		.dtor = _nouveau_engctx_dtor,
+		.init = _nouveau_engctx_init,
+		.fini = _nouveau_engctx_fini,
+		.rd32 = _nouveau_engctx_rd32,
+		.wr32 = _nouveau_engctx_wr32,
+	},
+};
+
+/*******************************************************************************
+ * BSP engine/subdev functions
+ ******************************************************************************/
+
+static int
+nv98_bsp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	      struct nouveau_oclass *oclass, void *data, u32 size,
+	      struct nouveau_object **pobject)
+{
+	struct nv98_bsp_priv *priv;
+	int ret;
+
+	ret = nouveau_engine_create(parent, engine, oclass, true,
+				    "PBSP", "bsp", &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	nv_subdev(priv)->unit = 0x04008000;
+	nv_engine(priv)->cclass = &nv98_bsp_cclass;
+	nv_engine(priv)->sclass = nv98_bsp_sclass;
+	return 0;
+}
+
+struct nouveau_oclass
+nv98_bsp_oclass = {
+	.handle = NV_ENGINE(BSP, 0x98),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv98_bsp_ctor,
+		.dtor = _nouveau_engine_dtor,
+		.init = _nouveau_engine_init,
+		.fini = _nouveau_engine_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/bsp/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/bsp/nvc0.c
index 0a5aa6bb0870..ce860de43e61 100644
--- a/drivers/gpu/drm/nouveau/core/engine/bsp/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/bsp/nvc0.c
@@ -22,8 +22,7 @@
  * Authors: Maarten Lankhorst
  */
 
-#include <core/falcon.h>
-
+#include <engine/falcon.h>
 #include <engine/bsp.h>
 
 struct nvc0_bsp_priv {
@@ -91,6 +90,7 @@ nvc0_bsp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	nv_subdev(priv)->unit = 0x00008000;
+	nv_subdev(priv)->intr = nouveau_falcon_intr;
 	nv_engine(priv)->cclass = &nvc0_bsp_cclass;
 	nv_engine(priv)->sclass = nvc0_bsp_sclass;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/engine/bsp/nve0.c b/drivers/gpu/drm/nouveau/core/engine/bsp/nve0.c
index d4f23bbd75b4..ba6aeca0285e 100644
--- a/drivers/gpu/drm/nouveau/core/engine/bsp/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/bsp/nve0.c
@@ -22,8 +22,7 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/falcon.h>
-
+#include <engine/falcon.h>
 #include <engine/bsp.h>
 
 struct nve0_bsp_priv {
@@ -91,6 +90,7 @@ nve0_bsp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	nv_subdev(priv)->unit = 0x00008000;
+	nv_subdev(priv)->intr = nouveau_falcon_intr;
 	nv_engine(priv)->cclass = &nve0_bsp_cclass;
 	nv_engine(priv)->sclass = nve0_bsp_sclass;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nva3.fuc.h b/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nva3.fuc.h
index c92520f3ed46..241b27201206 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nva3.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nva3.fuc.h
@@ -1,4 +1,4 @@
-static u32 nva3_pcopy_data[] = {
+uint32_t nva3_pcopy_data[] = {
 /* 0x0000: ctx_object */
 	0x00000000,
 /* 0x0004: ctx_dma */
@@ -183,7 +183,7 @@ static u32 nva3_pcopy_data[] = {
 	0x00000800,
 };
 
-static u32 nva3_pcopy_code[] = {
+uint32_t nva3_pcopy_code[] = {
 /* 0x0000: main */
 	0x04fe04bd,
 	0x3517f000,
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nvc0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nvc0.fuc.h
index 0d98c6c0958d..98cc4216a372 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nvc0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/fuc/nvc0.fuc.h
@@ -1,4 +1,4 @@
-static u32 nvc0_pcopy_data[] = {
+uint32_t nvc0_pcopy_data[] = {
 /* 0x0000: ctx_object */
 	0x00000000,
 /* 0x0004: ctx_query_address_high */
@@ -171,7 +171,7 @@ static u32 nvc0_pcopy_data[] = {
 	0x00000800,
 };
 
-static u32 nvc0_pcopy_code[] = {
+uint32_t nvc0_pcopy_code[] = {
 /* 0x0000: main */
 	0x04fe04bd,
 	0x3517f000,
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c b/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
index d6dc2a65ccd1..f31527733e00 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/nva3.c
@@ -22,16 +22,17 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/client.h>
-#include <core/falcon.h>
-#include <core/class.h>
-#include <core/enum.h>
+#include <engine/falcon.h>
+#include <engine/fifo.h>
+#include <engine/copy.h>
 
 #include <subdev/fb.h>
 #include <subdev/vm.h>
 
-#include <engine/fifo.h>
-#include <engine/copy.h>
+#include <core/client.h>
+#include <core/class.h>
+#include <core/enum.h>
+
 
 #include "fuc/nva3.fuc.h"
 
@@ -117,13 +118,6 @@ nva3_copy_intr(struct nouveau_subdev *subdev)
 }
 
 static int
-nva3_copy_tlb_flush(struct nouveau_engine *engine)
-{
-	nv50_vm_flush_engine(&engine->base, 0x0d);
-	return 0;
-}
-
-static int
 nva3_copy_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
 	       struct nouveau_object **pobject)
@@ -142,7 +136,6 @@ nva3_copy_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->intr = nva3_copy_intr;
 	nv_engine(priv)->cclass = &nva3_copy_cclass;
 	nv_engine(priv)->sclass = nva3_copy_sclass;
-	nv_engine(priv)->tlb_flush = nva3_copy_tlb_flush;
 	nv_falcon(priv)->code.data = nva3_pcopy_code;
 	nv_falcon(priv)->code.size = sizeof(nva3_pcopy_code);
 	nv_falcon(priv)->data.data = nva3_pcopy_data;
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/copy/nvc0.c
index b3ed2737e21f..993df09ad643 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/nvc0.c
@@ -22,13 +22,15 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/falcon.h>
-#include <core/class.h>
-#include <core/enum.h>
-
+#include <engine/falcon.h>
 #include <engine/fifo.h>
 #include <engine/copy.h>
 
+#include <core/class.h>
+#include <core/enum.h>
+#include <core/class.h>
+#include <core/enum.h>
+
 #include "fuc/nvc0.fuc.h"
 
 struct nvc0_copy_priv {
diff --git a/drivers/gpu/drm/nouveau/core/engine/copy/nve0.c b/drivers/gpu/drm/nouveau/core/engine/copy/nve0.c
index dbbe9e8998fe..30f1ef1edcc5 100644
--- a/drivers/gpu/drm/nouveau/core/engine/copy/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/copy/nve0.c
@@ -67,6 +67,19 @@ nve0_copy_cclass = {
  * PCOPY engine/subdev functions
  ******************************************************************************/
 
+static void
+nve0_copy_intr(struct nouveau_subdev *subdev)
+{
+	const int ce = nv_subidx(nv_object(subdev)) - NVDEV_ENGINE_COPY0;
+	struct nve0_copy_priv *priv = (void *)subdev;
+	u32 stat = nv_rd32(priv, 0x104908 + (ce * 0x1000));
+
+	if (stat) {
+		nv_warn(priv, "unhandled intr 0x%08x\n", stat);
+		nv_wr32(priv, 0x104908 + (ce * 0x1000), stat);
+	}
+}
+
 static int
 nve0_copy0_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		struct nouveau_oclass *oclass, void *data, u32 size,
@@ -85,6 +98,7 @@ nve0_copy0_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	nv_subdev(priv)->unit = 0x00000040;
+	nv_subdev(priv)->intr = nve0_copy_intr;
 	nv_engine(priv)->cclass = &nve0_copy_cclass;
 	nv_engine(priv)->sclass = nve0_copy_sclass;
 	return 0;
@@ -108,6 +122,28 @@ nve0_copy1_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	nv_subdev(priv)->unit = 0x00000080;
+	nv_subdev(priv)->intr = nve0_copy_intr;
+	nv_engine(priv)->cclass = &nve0_copy_cclass;
+	nv_engine(priv)->sclass = nve0_copy_sclass;
+	return 0;
+}
+
+static int
+nve0_copy2_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nve0_copy_priv *priv;
+	int ret;
+
+	ret = nouveau_engine_create(parent, engine, oclass, true,
+				    "PCE2", "copy2", &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	nv_subdev(priv)->unit = 0x00200000;
+	nv_subdev(priv)->intr = nve0_copy_intr;
 	nv_engine(priv)->cclass = &nve0_copy_cclass;
 	nv_engine(priv)->sclass = nve0_copy_sclass;
 	return 0;
@@ -134,3 +170,14 @@ nve0_copy1_oclass = {
 		.fini = _nouveau_engine_fini,
 	},
 };
+
+struct nouveau_oclass
+nve0_copy2_oclass = {
+	.handle = NV_ENGINE(COPY2, 0xe0),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nve0_copy2_ctor,
+		.dtor = _nouveau_engine_dtor,
+		.init = _nouveau_engine_init,
+		.fini = _nouveau_engine_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/fuc/nv98.fuc.h b/drivers/gpu/drm/nouveau/core/engine/crypt/fuc/nv98.fuc.h
index 09962e4210e9..38676c74e6e0 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/fuc/nv98.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/fuc/nv98.fuc.h
@@ -1,4 +1,4 @@
-static uint32_t nv98_pcrypt_data[] = {
+uint32_t nv98_pcrypt_data[] = {
 /* 0x0000: ctx_dma */
 /* 0x0000: ctx_dma_query */
 	0x00000000,
@@ -150,7 +150,7 @@ static uint32_t nv98_pcrypt_data[] = {
 	0x00000000,
 };
 
-static uint32_t nv98_pcrypt_code[] = {
+uint32_t nv98_pcrypt_code[] = {
 	0x17f004bd,
 	0x0010fe35,
 	0xf10004fe,
diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c b/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
index 5bc021f471f9..2551dafbec73 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/nv84.c
@@ -141,13 +141,6 @@ nv84_crypt_intr(struct nouveau_subdev *subdev)
 }
 
 static int
-nv84_crypt_tlb_flush(struct nouveau_engine *engine)
-{
-	nv50_vm_flush_engine(&engine->base, 0x0a);
-	return 0;
-}
-
-static int
 nv84_crypt_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
 	       struct nouveau_object **pobject)
@@ -165,7 +158,6 @@ nv84_crypt_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->intr = nv84_crypt_intr;
 	nv_engine(priv)->cclass = &nv84_crypt_cclass;
 	nv_engine(priv)->sclass = nv84_crypt_sclass;
-	nv_engine(priv)->tlb_flush = nv84_crypt_tlb_flush;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c b/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
index 8bf8955051d4..c7082377ec76 100644
--- a/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
+++ b/drivers/gpu/drm/nouveau/core/engine/crypt/nv98.c
@@ -27,11 +27,11 @@
 #include <core/enum.h>
 #include <core/class.h>
 #include <core/engctx.h>
-#include <core/falcon.h>
 
 #include <subdev/timer.h>
 #include <subdev/fb.h>
 
+#include <engine/falcon.h>
 #include <engine/fifo.h>
 #include <engine/crypt.h>
 
@@ -119,13 +119,6 @@ nv98_crypt_intr(struct nouveau_subdev *subdev)
 }
 
 static int
-nv98_crypt_tlb_flush(struct nouveau_engine *engine)
-{
-	nv50_vm_flush_engine(&engine->base, 0x0a);
-	return 0;
-}
-
-static int
 nv98_crypt_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_oclass *oclass, void *data, u32 size,
 	       struct nouveau_object **pobject)
@@ -143,7 +136,6 @@ nv98_crypt_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->intr = nv98_crypt_intr;
 	nv_engine(priv)->cclass = &nv98_crypt_cclass;
 	nv_engine(priv)->sclass = nv98_crypt_sclass;
-	nv_engine(priv)->tlb_flush = nv98_crypt_tlb_flush;
 	nv_falcon(priv)->code.data = nv98_pcrypt_code;
 	nv_falcon(priv)->code.size = sizeof(nv98_pcrypt_code);
 	nv_falcon(priv)->data.data = nv98_pcrypt_data;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nv50.c b/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
index 5e8c3de75593..ffc18b80c5d9 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nv50.c
@@ -227,9 +227,9 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
 		device->oclass[NVDEV_ENGINE_CRYPT  ] = &nv98_crypt_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nv94_disp_oclass;
 		break;
@@ -279,9 +279,9 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
 		device->oclass[NVDEV_ENGINE_CRYPT  ] = &nv98_crypt_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nv94_disp_oclass;
 		break;
@@ -305,9 +305,9 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
 		device->oclass[NVDEV_ENGINE_CRYPT  ] = &nv98_crypt_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nv94_disp_oclass;
 		break;
@@ -319,7 +319,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nva3_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -332,8 +332,8 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
 		device->oclass[NVDEV_ENGINE_MPEG   ] = &nv84_mpeg_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nva3_copy_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nva3_disp_oclass;
@@ -346,7 +346,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nva3_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -358,8 +358,8 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nva3_copy_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nva3_disp_oclass;
@@ -372,7 +372,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nva3_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -384,8 +384,8 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nva3_copy_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nva3_disp_oclass;
@@ -398,7 +398,7 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nva3_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nva3_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nv98_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nv50_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -410,8 +410,8 @@ nv50_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nv84_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nv50_software_oclass;
 		device->oclass[NVDEV_ENGINE_GR     ] = &nv50_graph_oclass;
-		device->oclass[NVDEV_ENGINE_VP     ] = &nv84_vp_oclass;
-		device->oclass[NVDEV_ENGINE_BSP    ] = &nv84_bsp_oclass;
+		device->oclass[NVDEV_ENGINE_VP     ] = &nv98_vp_oclass;
+		device->oclass[NVDEV_ENGINE_BSP    ] = &nv98_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nv98_ppp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nva3_copy_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nva3_disp_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
index a36e64e98ef3..418f51f50d7a 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nvc0.c
@@ -62,7 +62,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -75,7 +75,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc0_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -91,7 +91,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -104,7 +104,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc3_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -120,7 +120,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -133,7 +133,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc3_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -148,7 +148,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -161,7 +161,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc3_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -177,7 +177,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -190,7 +190,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc3_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -206,7 +206,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -219,7 +219,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc1_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -234,7 +234,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nva3_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -247,7 +247,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvc0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvc8_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -263,7 +263,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -276,7 +276,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvd9_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -291,7 +291,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -304,7 +304,7 @@ nvc0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nvc0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nvc0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvd7_graph_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nvc0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nvc0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/device/nve0.c b/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
index a354e409cdff..7aca1877add4 100644
--- a/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/device/nve0.c
@@ -62,7 +62,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -75,10 +75,11 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nve0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nve4_graph_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nve0_disp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nve0_copy0_oclass;
 		device->oclass[NVDEV_ENGINE_COPY1  ] = &nve0_copy1_oclass;
+		device->oclass[NVDEV_ENGINE_COPY2  ] = &nve0_copy2_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nve0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nve0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -91,7 +92,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -104,10 +105,11 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nve0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nve4_graph_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nve0_disp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nve0_copy0_oclass;
 		device->oclass[NVDEV_ENGINE_COPY1  ] = &nve0_copy1_oclass;
+		device->oclass[NVDEV_ENGINE_COPY2  ] = &nve0_copy2_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nve0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nve0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -120,7 +122,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -133,10 +135,11 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nve0_graph_oclass;
+		device->oclass[NVDEV_ENGINE_GR     ] =  nve4_graph_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nve0_disp_oclass;
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nve0_copy0_oclass;
 		device->oclass[NVDEV_ENGINE_COPY1  ] = &nve0_copy1_oclass;
+		device->oclass[NVDEV_ENGINE_COPY2  ] = &nve0_copy2_oclass;
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nve0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nve0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
@@ -149,7 +152,7 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_CLOCK  ] = &nvc0_clock_oclass;
 		device->oclass[NVDEV_SUBDEV_THERM  ] = &nvd0_therm_oclass;
 		device->oclass[NVDEV_SUBDEV_MXM    ] = &nv50_mxm_oclass;
-		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nv50_devinit_oclass;
+		device->oclass[NVDEV_SUBDEV_DEVINIT] = &nvc0_devinit_oclass;
 		device->oclass[NVDEV_SUBDEV_MC     ] = &nvc0_mc_oclass;
 		device->oclass[NVDEV_SUBDEV_BUS    ] = &nvc0_bus_oclass;
 		device->oclass[NVDEV_SUBDEV_TIMER  ] = &nv04_timer_oclass;
@@ -160,16 +163,14 @@ nve0_identify(struct nouveau_device *device)
 		device->oclass[NVDEV_SUBDEV_VM     ] = &nvc0_vmmgr_oclass;
 		device->oclass[NVDEV_SUBDEV_BAR    ] = &nvc0_bar_oclass;
 		device->oclass[NVDEV_ENGINE_DMAOBJ ] = &nvd0_dmaeng_oclass;
-#if 0
 		device->oclass[NVDEV_ENGINE_FIFO   ] = &nve0_fifo_oclass;
 		device->oclass[NVDEV_ENGINE_SW     ] = &nvc0_software_oclass;
-		device->oclass[NVDEV_ENGINE_GR     ] = &nve0_graph_oclass;
-#endif
+		device->oclass[NVDEV_ENGINE_GR     ] =  nvf0_graph_oclass;
 		device->oclass[NVDEV_ENGINE_DISP   ] = &nvf0_disp_oclass;
-#if 0
 		device->oclass[NVDEV_ENGINE_COPY0  ] = &nve0_copy0_oclass;
 		device->oclass[NVDEV_ENGINE_COPY1  ] = &nve0_copy1_oclass;
 		device->oclass[NVDEV_ENGINE_COPY2  ] = &nve0_copy2_oclass;
+#if 0
 		device->oclass[NVDEV_ENGINE_BSP    ] = &nve0_bsp_oclass;
 		device->oclass[NVDEV_ENGINE_VP     ] = &nve0_vp_oclass;
 		device->oclass[NVDEV_ENGINE_PPP    ] = &nvc0_ppp_oclass;
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/hdminva3.c b/drivers/gpu/drm/nouveau/core/engine/disp/hdminva3.c
index f065fc248adf..db8c6fd46278 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/hdminva3.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/hdminva3.c
@@ -55,6 +55,10 @@ nva3_hdmi_ctrl(struct nv50_disp_priv *priv, int head, int or, u32 data)
 	nv_wr32(priv, 0x61c510 + soff, 0x00000000);
 	nv_mask(priv, 0x61c500 + soff, 0x00000001, 0x00000001);
 
+	nv_mask(priv, 0x61c5d0 + soff, 0x00070001, 0x00010001); /* SPARE, HW_CTS */
+	nv_mask(priv, 0x61c568 + soff, 0x00010101, 0x00000000); /* ACR_CTRL, ?? */
+	nv_mask(priv, 0x61c578 + soff, 0x80000000, 0x80000000); /* ACR_0441_ENABLE */
+
 	/* ??? */
 	nv_mask(priv, 0x61733c, 0x00100000, 0x00100000); /* RESETF */
 	nv_mask(priv, 0x61733c, 0x10000000, 0x10000000); /* LOOKUP_EN */
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
index 6a38402fa56c..7ffe2f309f12 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nv50.c
@@ -34,9 +34,9 @@
 #include <subdev/bios/disp.h>
 #include <subdev/bios/init.h>
 #include <subdev/bios/pll.h>
+#include <subdev/devinit.h>
 #include <subdev/timer.h>
 #include <subdev/fb.h>
-#include <subdev/clock.h>
 
 #include "nv50.h"
 
@@ -987,10 +987,10 @@ nv50_disp_intr_unk20_0(struct nv50_disp_priv *priv, int head)
 static void
 nv50_disp_intr_unk20_1(struct nv50_disp_priv *priv, int head)
 {
-	struct nouveau_clock *clk = nouveau_clock(priv);
+	struct nouveau_devinit *devinit = nouveau_devinit(priv);
 	u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
 	if (pclk)
-		clk->pll_set(clk, PLL_VPLL0 + head, pclk);
+		devinit->pll_set(devinit, PLL_VPLL0 + head, pclk);
 }
 
 static void
@@ -1107,6 +1107,7 @@ nv50_disp_intr_unk20_2(struct nv50_disp_priv *priv, int head)
 	u32 pclk = nv_rd32(priv, 0x610ad0 + (head * 0x540)) & 0x3fffff;
 	u32 hval, hreg = 0x614200 + (head * 0x800);
 	u32 oval, oreg;
+	u32 mask;
 	u32 conf = exec_clkcmp(priv, head, 0xff, pclk, &outp);
 	if (conf != ~0) {
 		if (outp.location == 0 && outp.type == DCB_OUTPUT_DP) {
@@ -1133,6 +1134,7 @@ nv50_disp_intr_unk20_2(struct nv50_disp_priv *priv, int head)
 			oreg = 0x614280 + (ffs(outp.or) - 1) * 0x800;
 			oval = 0x00000000;
 			hval = 0x00000000;
+			mask = 0xffffffff;
 		} else
 		if (!outp.location) {
 			if (outp.type == DCB_OUTPUT_DP)
@@ -1140,14 +1142,16 @@ nv50_disp_intr_unk20_2(struct nv50_disp_priv *priv, int head)
 			oreg = 0x614300 + (ffs(outp.or) - 1) * 0x800;
 			oval = (conf & 0x0100) ? 0x00000101 : 0x00000000;
 			hval = 0x00000000;
+			mask = 0x00000707;
 		} else {
 			oreg = 0x614380 + (ffs(outp.or) - 1) * 0x800;
 			oval = 0x00000001;
 			hval = 0x00000001;
+			mask = 0x00000707;
 		}
 
 		nv_mask(priv, hreg, 0x0000000f, hval);
-		nv_mask(priv, oreg, 0x00000707, oval);
+		nv_mask(priv, oreg, mask, oval);
 	}
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
index 019eacd8a68f..52dd7a1db729 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvd0.c
@@ -29,15 +29,14 @@
 
 #include <engine/disp.h>
 
-#include <subdev/timer.h>
-#include <subdev/fb.h>
-#include <subdev/clock.h>
-
 #include <subdev/bios.h>
 #include <subdev/bios/dcb.h>
 #include <subdev/bios/disp.h>
 #include <subdev/bios/init.h>
 #include <subdev/bios/pll.h>
+#include <subdev/devinit.h>
+#include <subdev/fb.h>
+#include <subdev/timer.h>
 
 #include "nv50.h"
 
@@ -738,10 +737,10 @@ nvd0_disp_intr_unk2_0(struct nv50_disp_priv *priv, int head)
 static void
 nvd0_disp_intr_unk2_1(struct nv50_disp_priv *priv, int head)
 {
-	struct nouveau_clock *clk = nouveau_clock(priv);
+	struct nouveau_devinit *devinit = nouveau_devinit(priv);
 	u32 pclk = nv_rd32(priv, 0x660450 + (head * 0x300)) / 1000;
 	if (pclk)
-		clk->pll_set(clk, PLL_VPLL0 + head, pclk);
+		devinit->pll_set(devinit, PLL_VPLL0 + head, pclk);
 	nv_wr32(priv, 0x612200 + (head * 0x800), 0x00000000);
 }
 
@@ -959,6 +958,9 @@ nvd0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	int heads = nv_rd32(parent, 0x022448);
 	int ret;
 
+	if (nv_rd32(parent, 0x022500) & 0x00000001)
+		return -ENODEV;
+
 	ret = nouveau_disp_create(parent, engine, oclass, heads,
 				  "PDISP", "display", &priv);
 	*pobject = nv_object(priv);
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
index 20725b363d58..fb1fe6ae5e74 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nve0.c
@@ -54,6 +54,9 @@ nve0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	int heads = nv_rd32(parent, 0x022448);
 	int ret;
 
+	if (nv_rd32(parent, 0x022500) & 0x00000001)
+		return -ENODEV;
+
 	ret = nouveau_disp_create(parent, engine, oclass, heads,
 				  "PDISP", "display", &priv);
 	*pobject = nv_object(priv);
diff --git a/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c b/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
index a488c36e40f9..42aa6b97dbea 100644
--- a/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/disp/nvf0.c
@@ -54,6 +54,9 @@ nvf0_disp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	int heads = nv_rd32(parent, 0x022448);
 	int ret;
 
+	if (nv_rd32(parent, 0x022500) & 0x00000001)
+		return -ENODEV;
+
 	ret = nouveau_disp_create(parent, engine, oclass, heads,
 				  "PDISP", "display", &priv);
 	*pobject = nv_object(priv);
diff --git a/drivers/gpu/drm/nouveau/core/core/falcon.c b/drivers/gpu/drm/nouveau/core/engine/falcon.c
index e05c15777588..e03fc8e4dc1d 100644
--- a/drivers/gpu/drm/nouveau/core/core/falcon.c
+++ b/drivers/gpu/drm/nouveau/core/engine/falcon.c
@@ -20,10 +20,28 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
-#include <core/falcon.h>
-
+#include <engine/falcon.h>
 #include <subdev/timer.h>
 
+void
+nouveau_falcon_intr(struct nouveau_subdev *subdev)
+{
+	struct nouveau_falcon *falcon = (void *)subdev;
+	u32 dispatch = nv_ro32(falcon, 0x01c);
+	u32 intr = nv_ro32(falcon, 0x008) & dispatch & ~(dispatch >> 16);
+
+	if (intr & 0x00000010) {
+		nv_debug(falcon, "ucode halted\n");
+		nv_wo32(falcon, 0x004, 0x00000010);
+		intr &= ~0x00000010;
+	}
+
+	if (intr)  {
+		nv_error(falcon, "unhandled intr 0x%08x\n", intr);
+		nv_wo32(falcon, 0x004, intr);
+	}
+}
+
 u32
 _nouveau_falcon_rd32(struct nouveau_object *object, u64 addr)
 {
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv40.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv40.c
index 2b1f91721225..5c7433d5069f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv40.c
@@ -320,7 +320,7 @@ nv40_fifo_init(struct nouveau_object *object)
 		break;
 	default:
 		nv_wr32(priv, 0x002230, 0x00000000);
-		nv_wr32(priv, 0x002220, ((pfb->ram.size - 512 * 1024 +
+		nv_wr32(priv, 0x002220, ((pfb->ram->size - 512 * 1024 +
 					 priv->ramfc->addr) >> 16) |
 					0x00030000);
 		break;
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
index 35b94bd18808..7f53196cff52 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nv84.c
@@ -56,7 +56,9 @@ nv84_fifo_context_attach(struct nouveau_object *parent,
 	switch (nv_engidx(object->engine)) {
 	case NVDEV_ENGINE_SW   : return 0;
 	case NVDEV_ENGINE_GR   : addr = 0x0020; break;
+	case NVDEV_ENGINE_VP   : addr = 0x0040; break;
 	case NVDEV_ENGINE_MPEG : addr = 0x0060; break;
+	case NVDEV_ENGINE_BSP  : addr = 0x0080; break;
 	case NVDEV_ENGINE_CRYPT: addr = 0x00a0; break;
 	case NVDEV_ENGINE_COPY0: addr = 0x00c0; break;
 	default:
@@ -89,7 +91,9 @@ nv84_fifo_context_detach(struct nouveau_object *parent, bool suspend,
 	switch (nv_engidx(object->engine)) {
 	case NVDEV_ENGINE_SW   : return 0;
 	case NVDEV_ENGINE_GR   : engn = 0; addr = 0x0020; break;
+	case NVDEV_ENGINE_VP   : engn = 3; addr = 0x0040; break;
 	case NVDEV_ENGINE_MPEG : engn = 1; addr = 0x0060; break;
+	case NVDEV_ENGINE_BSP  : engn = 5; addr = 0x0080; break;
 	case NVDEV_ENGINE_CRYPT: engn = 4; addr = 0x00a0; break;
 	case NVDEV_ENGINE_COPY0: engn = 2; addr = 0x00c0; break;
 	default:
diff --git a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
index 56192a7242ae..09644fa9602c 100644
--- a/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/fifo/nve0.c
@@ -44,7 +44,8 @@ static const struct {
 	u64 subdev;
 	u64 mask;
 } fifo_engine[] = {
-	_(NVDEV_ENGINE_GR      , (1ULL << NVDEV_ENGINE_SW)),
+	_(NVDEV_ENGINE_GR      , (1ULL << NVDEV_ENGINE_SW) |
+				 (1ULL << NVDEV_ENGINE_COPY2)),
 	_(NVDEV_ENGINE_VP      , 0),
 	_(NVDEV_ENGINE_PPP     , 0),
 	_(NVDEV_ENGINE_BSP     , 0),
@@ -96,18 +97,6 @@ nve0_fifo_playlist_update(struct nve0_fifo_priv *priv, u32 engine)
 
 	mutex_lock(&nv_subdev(priv)->mutex);
 	cur = engn->playlist[engn->cur_playlist];
-	if (unlikely(cur == NULL)) {
-		int ret = nouveau_gpuobj_new(nv_object(priv), NULL,
-					     0x8000, 0x1000, 0, &cur);
-		if (ret) {
-			mutex_unlock(&nv_subdev(priv)->mutex);
-			nv_error(priv, "playlist alloc failed\n");
-			return;
-		}
-
-		engn->playlist[engn->cur_playlist] = cur;
-	}
-
 	engn->cur_playlist = !engn->cur_playlist;
 
 	for (i = 0, p = 0; i < priv->base.max; i++) {
@@ -138,10 +127,12 @@ nve0_fifo_context_attach(struct nouveau_object *parent,
 	int ret;
 
 	switch (nv_engidx(object->engine)) {
-	case NVDEV_ENGINE_SW   : return 0;
-	case NVDEV_ENGINE_GR   :
+	case NVDEV_ENGINE_SW   :
 	case NVDEV_ENGINE_COPY0:
-	case NVDEV_ENGINE_COPY1: addr = 0x0210; break;
+	case NVDEV_ENGINE_COPY1:
+	case NVDEV_ENGINE_COPY2:
+		return 0;
+	case NVDEV_ENGINE_GR   : addr = 0x0210; break;
 	case NVDEV_ENGINE_BSP  : addr = 0x0270; break;
 	case NVDEV_ENGINE_VP   : addr = 0x0250; break;
 	case NVDEV_ENGINE_PPP  : addr = 0x0260; break;
@@ -176,9 +167,10 @@ nve0_fifo_context_detach(struct nouveau_object *parent, bool suspend,
 
 	switch (nv_engidx(object->engine)) {
 	case NVDEV_ENGINE_SW   : return 0;
-	case NVDEV_ENGINE_GR   :
 	case NVDEV_ENGINE_COPY0:
-	case NVDEV_ENGINE_COPY1: addr = 0x0210; break;
+	case NVDEV_ENGINE_COPY1:
+	case NVDEV_ENGINE_COPY2: addr = 0x0000; break;
+	case NVDEV_ENGINE_GR   : addr = 0x0210; break;
 	case NVDEV_ENGINE_BSP  : addr = 0x0270; break;
 	case NVDEV_ENGINE_VP   : addr = 0x0250; break;
 	case NVDEV_ENGINE_PPP  : addr = 0x0260; break;
@@ -194,9 +186,12 @@ nve0_fifo_context_detach(struct nouveau_object *parent, bool suspend,
 			return -EBUSY;
 	}
 
-	nv_wo32(base, addr + 0x00, 0x00000000);
-	nv_wo32(base, addr + 0x04, 0x00000000);
-	bar->flush(bar);
+	if (addr) {
+		nv_wo32(base, addr + 0x00, 0x00000000);
+		nv_wo32(base, addr + 0x04, 0x00000000);
+		bar->flush(bar);
+	}
+
 	return 0;
 }
 
@@ -226,8 +221,10 @@ nve0_fifo_chan_ctor(struct nouveau_object *parent,
 		}
 	}
 
-	if (i == FIFO_ENGINE_NR)
+	if (i == FIFO_ENGINE_NR) {
+		nv_error(priv, "unsupported engines 0x%08x\n", args->engine);
 		return -ENODEV;
+	}
 
 	ret = nouveau_fifo_channel_create(parent, engine, oclass, 1,
 					  priv->user.bar.offset, 0x200,
@@ -592,13 +589,25 @@ nve0_fifo_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	       struct nouveau_object **pobject)
 {
 	struct nve0_fifo_priv *priv;
-	int ret;
+	int ret, i;
 
 	ret = nouveau_fifo_create(parent, engine, oclass, 0, 4095, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
+	for (i = 0; i < FIFO_ENGINE_NR; i++) {
+		ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x8000, 0x1000,
+					 0, &priv->engine[i].playlist[0]);
+		if (ret)
+			return ret;
+
+		ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x8000, 0x1000,
+					 0, &priv->engine[i].playlist[1]);
+		if (ret)
+			return ret;
+	}
+
 	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 4096 * 0x200, 0x1000,
 				 NVOBJ_FLAG_ZERO_ALLOC, &priv->user.mem);
 	if (ret)
@@ -629,7 +638,7 @@ nve0_fifo_dtor(struct nouveau_object *object)
 	nouveau_gpuobj_unmap(&priv->user.bar);
 	nouveau_gpuobj_ref(NULL, &priv->user.mem);
 
-	for (i = 0; i < ARRAY_SIZE(priv->engine); i++) {
+	for (i = 0; i < FIFO_ENGINE_NR; i++) {
 		nouveau_gpuobj_ref(NULL, &priv->engine[i].playlist[1]);
 		nouveau_gpuobj_ref(NULL, &priv->engine[i].playlist[0]);
 	}
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c
index 4cc6269d4077..64dca260912f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc0.c
@@ -24,3015 +24,1220 @@
 
 #include "nvc0.h"
 
-void
-nv_icmd(struct nvc0_graph_priv *priv, u32 icmd, u32 data)
-{
-	nv_wr32(priv, 0x400204, data);
-	nv_wr32(priv, 0x400200, icmd);
-	while (nv_rd32(priv, 0x400700) & 2) {}
-}
+struct nvc0_graph_init
+nvc0_grctx_init_icmd[] = {
+	{ 0x001000,   1, 0x01, 0x00000004 },
+	{ 0x0000a9,   1, 0x01, 0x0000ffff },
+	{ 0x000038,   1, 0x01, 0x0fac6881 },
+	{ 0x00003d,   1, 0x01, 0x00000001 },
+	{ 0x0000e8,   8, 0x01, 0x00000400 },
+	{ 0x000078,   8, 0x01, 0x00000300 },
+	{ 0x000050,   1, 0x01, 0x00000011 },
+	{ 0x000058,   8, 0x01, 0x00000008 },
+	{ 0x000208,   8, 0x01, 0x00000001 },
+	{ 0x000081,   1, 0x01, 0x00000001 },
+	{ 0x000085,   1, 0x01, 0x00000004 },
+	{ 0x000088,   1, 0x01, 0x00000400 },
+	{ 0x000090,   1, 0x01, 0x00000300 },
+	{ 0x000098,   1, 0x01, 0x00001001 },
+	{ 0x0000e3,   1, 0x01, 0x00000001 },
+	{ 0x0000da,   1, 0x01, 0x00000001 },
+	{ 0x0000f8,   1, 0x01, 0x00000003 },
+	{ 0x0000fa,   1, 0x01, 0x00000001 },
+	{ 0x00009f,   4, 0x01, 0x0000ffff },
+	{ 0x0000b1,   1, 0x01, 0x00000001 },
+	{ 0x0000b2,  40, 0x01, 0x00000000 },
+	{ 0x000210,   8, 0x01, 0x00000040 },
+	{ 0x000218,   8, 0x01, 0x0000c080 },
+	{ 0x0000ad,   1, 0x01, 0x0000013e },
+	{ 0x0000e1,   1, 0x01, 0x00000010 },
+	{ 0x000290,  16, 0x01, 0x00000000 },
+	{ 0x0003b0,  16, 0x01, 0x00000000 },
+	{ 0x0002a0,  16, 0x01, 0x00000000 },
+	{ 0x000420,  16, 0x01, 0x00000000 },
+	{ 0x0002b0,  16, 0x01, 0x00000000 },
+	{ 0x000430,  16, 0x01, 0x00000000 },
+	{ 0x0002c0,  16, 0x01, 0x00000000 },
+	{ 0x0004d0,  16, 0x01, 0x00000000 },
+	{ 0x000720,  16, 0x01, 0x00000000 },
+	{ 0x0008c0,  16, 0x01, 0x00000000 },
+	{ 0x000890,  16, 0x01, 0x00000000 },
+	{ 0x0008e0,  16, 0x01, 0x00000000 },
+	{ 0x0008a0,  16, 0x01, 0x00000000 },
+	{ 0x0008f0,  16, 0x01, 0x00000000 },
+	{ 0x00094c,   1, 0x01, 0x000000ff },
+	{ 0x00094d,   1, 0x01, 0xffffffff },
+	{ 0x00094e,   1, 0x01, 0x00000002 },
+	{ 0x0002ec,   1, 0x01, 0x00000001 },
+	{ 0x000303,   1, 0x01, 0x00000001 },
+	{ 0x0002e6,   1, 0x01, 0x00000001 },
+	{ 0x000466,   1, 0x01, 0x00000052 },
+	{ 0x000301,   1, 0x01, 0x3f800000 },
+	{ 0x000304,   1, 0x01, 0x30201000 },
+	{ 0x000305,   1, 0x01, 0x70605040 },
+	{ 0x000306,   1, 0x01, 0xb8a89888 },
+	{ 0x000307,   1, 0x01, 0xf8e8d8c8 },
+	{ 0x00030a,   1, 0x01, 0x00ffff00 },
+	{ 0x00030b,   1, 0x01, 0x0000001a },
+	{ 0x00030c,   1, 0x01, 0x00000001 },
+	{ 0x000318,   1, 0x01, 0x00000001 },
+	{ 0x000340,   1, 0x01, 0x00000000 },
+	{ 0x000375,   1, 0x01, 0x00000001 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x00037d,   1, 0x01, 0x00000006 },
+	{ 0x0003a0,   1, 0x01, 0x00000002 },
+	{ 0x0003aa,   1, 0x01, 0x00000001 },
+	{ 0x0003a9,   1, 0x01, 0x00000001 },
+	{ 0x000380,   1, 0x01, 0x00000001 },
+	{ 0x000360,   1, 0x01, 0x00000040 },
+	{ 0x000366,   2, 0x01, 0x00000000 },
+	{ 0x000368,   1, 0x01, 0x00001fff },
+	{ 0x000370,   2, 0x01, 0x00000000 },
+	{ 0x000372,   1, 0x01, 0x003fffff },
+	{ 0x00037a,   1, 0x01, 0x00000012 },
+	{ 0x0005e0,   5, 0x01, 0x00000022 },
+	{ 0x000619,   1, 0x01, 0x00000003 },
+	{ 0x000811,   1, 0x01, 0x00000003 },
+	{ 0x000812,   1, 0x01, 0x00000004 },
+	{ 0x000813,   1, 0x01, 0x00000006 },
+	{ 0x000814,   1, 0x01, 0x00000008 },
+	{ 0x000815,   1, 0x01, 0x0000000b },
+	{ 0x000800,   6, 0x01, 0x00000001 },
+	{ 0x000632,   1, 0x01, 0x00000001 },
+	{ 0x000633,   1, 0x01, 0x00000002 },
+	{ 0x000634,   1, 0x01, 0x00000003 },
+	{ 0x000635,   1, 0x01, 0x00000004 },
+	{ 0x000654,   1, 0x01, 0x3f800000 },
+	{ 0x000657,   1, 0x01, 0x3f800000 },
+	{ 0x000655,   2, 0x01, 0x3f800000 },
+	{ 0x0006cd,   1, 0x01, 0x3f800000 },
+	{ 0x0007f5,   1, 0x01, 0x3f800000 },
+	{ 0x0007dc,   1, 0x01, 0x39291909 },
+	{ 0x0007dd,   1, 0x01, 0x79695949 },
+	{ 0x0007de,   1, 0x01, 0xb9a99989 },
+	{ 0x0007df,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007e8,   1, 0x01, 0x00003210 },
+	{ 0x0007e9,   1, 0x01, 0x00007654 },
+	{ 0x0007ea,   1, 0x01, 0x00000098 },
+	{ 0x0007ec,   1, 0x01, 0x39291909 },
+	{ 0x0007ed,   1, 0x01, 0x79695949 },
+	{ 0x0007ee,   1, 0x01, 0xb9a99989 },
+	{ 0x0007ef,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007f0,   1, 0x01, 0x00003210 },
+	{ 0x0007f1,   1, 0x01, 0x00007654 },
+	{ 0x0007f2,   1, 0x01, 0x00000098 },
+	{ 0x0005a5,   1, 0x01, 0x00000001 },
+	{ 0x000980, 128, 0x01, 0x00000000 },
+	{ 0x000468,   1, 0x01, 0x00000004 },
+	{ 0x00046c,   1, 0x01, 0x00000001 },
+	{ 0x000470,  96, 0x01, 0x00000000 },
+	{ 0x000510,  16, 0x01, 0x3f800000 },
+	{ 0x000520,   1, 0x01, 0x000002b6 },
+	{ 0x000529,   1, 0x01, 0x00000001 },
+	{ 0x000530,  16, 0x01, 0xffff0000 },
+	{ 0x000585,   1, 0x01, 0x0000003f },
+	{ 0x000576,   1, 0x01, 0x00000003 },
+	{ 0x000586,   1, 0x01, 0x00000040 },
+	{ 0x000582,   2, 0x01, 0x00000080 },
+	{ 0x0005c2,   1, 0x01, 0x00000001 },
+	{ 0x000638,   1, 0x01, 0x00000001 },
+	{ 0x000639,   1, 0x01, 0x00000001 },
+	{ 0x00063a,   1, 0x01, 0x00000002 },
+	{ 0x00063b,   2, 0x01, 0x00000001 },
+	{ 0x00063d,   1, 0x01, 0x00000002 },
+	{ 0x00063e,   1, 0x01, 0x00000001 },
+	{ 0x0008b8,   8, 0x01, 0x00000001 },
+	{ 0x000900,   8, 0x01, 0x00000001 },
+	{ 0x000908,   8, 0x01, 0x00000002 },
+	{ 0x000910,  16, 0x01, 0x00000001 },
+	{ 0x000920,   8, 0x01, 0x00000002 },
+	{ 0x000928,   8, 0x01, 0x00000001 },
+	{ 0x000648,   9, 0x01, 0x00000001 },
+	{ 0x000658,   1, 0x01, 0x0000000f },
+	{ 0x0007ff,   1, 0x01, 0x0000000a },
+	{ 0x00066a,   1, 0x01, 0x40000000 },
+	{ 0x00066b,   1, 0x01, 0x10000000 },
+	{ 0x00066c,   2, 0x01, 0xffff0000 },
+	{ 0x0007af,   2, 0x01, 0x00000008 },
+	{ 0x0007f6,   1, 0x01, 0x00000001 },
+	{ 0x0006b2,   1, 0x01, 0x00000055 },
+	{ 0x0007ad,   1, 0x01, 0x00000003 },
+	{ 0x000937,   1, 0x01, 0x00000001 },
+	{ 0x000971,   1, 0x01, 0x00000008 },
+	{ 0x000972,   1, 0x01, 0x00000040 },
+	{ 0x000973,   1, 0x01, 0x0000012c },
+	{ 0x00097c,   1, 0x01, 0x00000040 },
+	{ 0x000979,   1, 0x01, 0x00000003 },
+	{ 0x000975,   1, 0x01, 0x00000020 },
+	{ 0x000976,   1, 0x01, 0x00000001 },
+	{ 0x000977,   1, 0x01, 0x00000020 },
+	{ 0x000978,   1, 0x01, 0x00000001 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095e,   1, 0x01, 0x20164010 },
+	{ 0x00095f,   1, 0x01, 0x00000020 },
+	{ 0x000683,   1, 0x01, 0x00000006 },
+	{ 0x000685,   1, 0x01, 0x003fffff },
+	{ 0x000687,   1, 0x01, 0x00000c48 },
+	{ 0x0006a0,   1, 0x01, 0x00000005 },
+	{ 0x000840,   1, 0x01, 0x00300008 },
+	{ 0x000841,   1, 0x01, 0x04000080 },
+	{ 0x000842,   1, 0x01, 0x00300008 },
+	{ 0x000843,   1, 0x01, 0x04000080 },
+	{ 0x000818,   8, 0x01, 0x00000000 },
+	{ 0x000848,  16, 0x01, 0x00000000 },
+	{ 0x000738,   1, 0x01, 0x00000000 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ab,   1, 0x01, 0x00000002 },
+	{ 0x0006ac,   1, 0x01, 0x00000080 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x0006bb,   1, 0x01, 0x000000cf },
+	{ 0x0006ce,   1, 0x01, 0x2a712488 },
+	{ 0x000739,   1, 0x01, 0x4085c000 },
+	{ 0x00073a,   1, 0x01, 0x00000080 },
+	{ 0x000786,   1, 0x01, 0x80000100 },
+	{ 0x00073c,   1, 0x01, 0x00010100 },
+	{ 0x00073d,   1, 0x01, 0x02800000 },
+	{ 0x000787,   1, 0x01, 0x000000cf },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x000836,   1, 0x01, 0x00000001 },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x000944,   1, 0x01, 0x00000022 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x00c1b0,   8, 0x01, 0x0000000f },
+	{ 0x00c1b8,   1, 0x01, 0x0fac6881 },
+	{ 0x00c1b9,   1, 0x01, 0x00fac688 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000002 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000014 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000001 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_9097[] = {
+	{ 0x000800,   8, 0x40, 0x00000000 },
+	{ 0x000804,   8, 0x40, 0x00000000 },
+	{ 0x000808,   8, 0x40, 0x00000400 },
+	{ 0x00080c,   8, 0x40, 0x00000300 },
+	{ 0x000810,   1, 0x04, 0x000000cf },
+	{ 0x000850,   7, 0x40, 0x00000000 },
+	{ 0x000814,   8, 0x40, 0x00000040 },
+	{ 0x000818,   8, 0x40, 0x00000001 },
+	{ 0x00081c,   8, 0x40, 0x00000000 },
+	{ 0x000820,   8, 0x40, 0x00000000 },
+	{ 0x002700,   8, 0x20, 0x00000000 },
+	{ 0x002704,   8, 0x20, 0x00000000 },
+	{ 0x002708,   8, 0x20, 0x00000000 },
+	{ 0x00270c,   8, 0x20, 0x00000000 },
+	{ 0x002710,   8, 0x20, 0x00014000 },
+	{ 0x002714,   8, 0x20, 0x00000040 },
+	{ 0x001c00,  16, 0x10, 0x00000000 },
+	{ 0x001c04,  16, 0x10, 0x00000000 },
+	{ 0x001c08,  16, 0x10, 0x00000000 },
+	{ 0x001c0c,  16, 0x10, 0x00000000 },
+	{ 0x001d00,  16, 0x10, 0x00000000 },
+	{ 0x001d04,  16, 0x10, 0x00000000 },
+	{ 0x001d08,  16, 0x10, 0x00000000 },
+	{ 0x001d0c,  16, 0x10, 0x00000000 },
+	{ 0x001f00,  16, 0x08, 0x00000000 },
+	{ 0x001f04,  16, 0x08, 0x00000000 },
+	{ 0x001f80,  16, 0x08, 0x00000000 },
+	{ 0x001f84,  16, 0x08, 0x00000000 },
+	{ 0x002200,   5, 0x10, 0x00000022 },
+	{ 0x002000,   1, 0x04, 0x00000000 },
+	{ 0x002040,   1, 0x04, 0x00000011 },
+	{ 0x002080,   1, 0x04, 0x00000020 },
+	{ 0x0020c0,   1, 0x04, 0x00000030 },
+	{ 0x002100,   1, 0x04, 0x00000040 },
+	{ 0x002140,   1, 0x04, 0x00000051 },
+	{ 0x00200c,   6, 0x40, 0x00000001 },
+	{ 0x002010,   1, 0x04, 0x00000000 },
+	{ 0x002050,   1, 0x04, 0x00000000 },
+	{ 0x002090,   1, 0x04, 0x00000001 },
+	{ 0x0020d0,   1, 0x04, 0x00000002 },
+	{ 0x002110,   1, 0x04, 0x00000003 },
+	{ 0x002150,   1, 0x04, 0x00000004 },
+	{ 0x000380,   4, 0x20, 0x00000000 },
+	{ 0x000384,   4, 0x20, 0x00000000 },
+	{ 0x000388,   4, 0x20, 0x00000000 },
+	{ 0x00038c,   4, 0x20, 0x00000000 },
+	{ 0x000700,   4, 0x10, 0x00000000 },
+	{ 0x000704,   4, 0x10, 0x00000000 },
+	{ 0x000708,   4, 0x10, 0x00000000 },
+	{ 0x002800, 128, 0x04, 0x00000000 },
+	{ 0x000a00,  16, 0x20, 0x00000000 },
+	{ 0x000a04,  16, 0x20, 0x00000000 },
+	{ 0x000a08,  16, 0x20, 0x00000000 },
+	{ 0x000a0c,  16, 0x20, 0x00000000 },
+	{ 0x000a10,  16, 0x20, 0x00000000 },
+	{ 0x000a14,  16, 0x20, 0x00000000 },
+	{ 0x000c00,  16, 0x10, 0x00000000 },
+	{ 0x000c04,  16, 0x10, 0x00000000 },
+	{ 0x000c08,  16, 0x10, 0x00000000 },
+	{ 0x000c0c,  16, 0x10, 0x3f800000 },
+	{ 0x000d00,   8, 0x08, 0xffff0000 },
+	{ 0x000d04,   8, 0x08, 0xffff0000 },
+	{ 0x000e00,  16, 0x10, 0x00000000 },
+	{ 0x000e04,  16, 0x10, 0xffff0000 },
+	{ 0x000e08,  16, 0x10, 0xffff0000 },
+	{ 0x000d40,   4, 0x08, 0x00000000 },
+	{ 0x000d44,   4, 0x08, 0x00000000 },
+	{ 0x001e00,   8, 0x20, 0x00000001 },
+	{ 0x001e04,   8, 0x20, 0x00000001 },
+	{ 0x001e08,   8, 0x20, 0x00000002 },
+	{ 0x001e0c,   8, 0x20, 0x00000001 },
+	{ 0x001e10,   8, 0x20, 0x00000001 },
+	{ 0x001e14,   8, 0x20, 0x00000002 },
+	{ 0x001e18,   8, 0x20, 0x00000001 },
+	{ 0x003400, 128, 0x04, 0x00000000 },
+	{ 0x00030c,   1, 0x04, 0x00000001 },
+	{ 0x001944,   1, 0x04, 0x00000000 },
+	{ 0x001514,   1, 0x04, 0x00000000 },
+	{ 0x000d68,   1, 0x04, 0x0000ffff },
+	{ 0x00121c,   1, 0x04, 0x0fac6881 },
+	{ 0x000fac,   1, 0x04, 0x00000001 },
+	{ 0x001538,   1, 0x04, 0x00000001 },
+	{ 0x000fe0,   2, 0x04, 0x00000000 },
+	{ 0x000fe8,   1, 0x04, 0x00000014 },
+	{ 0x000fec,   1, 0x04, 0x00000040 },
+	{ 0x000ff0,   1, 0x04, 0x00000000 },
+	{ 0x00179c,   1, 0x04, 0x00000000 },
+	{ 0x001228,   1, 0x04, 0x00000400 },
+	{ 0x00122c,   1, 0x04, 0x00000300 },
+	{ 0x001230,   1, 0x04, 0x00010001 },
+	{ 0x0007f8,   1, 0x04, 0x00000000 },
+	{ 0x0015b4,   1, 0x04, 0x00000001 },
+	{ 0x0015cc,   1, 0x04, 0x00000000 },
+	{ 0x001534,   1, 0x04, 0x00000000 },
+	{ 0x000fb0,   1, 0x04, 0x00000000 },
+	{ 0x0015d0,   1, 0x04, 0x00000000 },
+	{ 0x00153c,   1, 0x04, 0x00000000 },
+	{ 0x0016b4,   1, 0x04, 0x00000003 },
+	{ 0x000fbc,   4, 0x04, 0x0000ffff },
+	{ 0x000df8,   2, 0x04, 0x00000000 },
+	{ 0x001948,   1, 0x04, 0x00000000 },
+	{ 0x001970,   1, 0x04, 0x00000001 },
+	{ 0x00161c,   1, 0x04, 0x000009f0 },
+	{ 0x000dcc,   1, 0x04, 0x00000010 },
+	{ 0x00163c,   1, 0x04, 0x00000000 },
+	{ 0x0015e4,   1, 0x04, 0x00000000 },
+	{ 0x001160,  32, 0x04, 0x25e00040 },
+	{ 0x001880,  32, 0x04, 0x00000000 },
+	{ 0x000f84,   2, 0x04, 0x00000000 },
+	{ 0x0017c8,   2, 0x04, 0x00000000 },
+	{ 0x0017d0,   1, 0x04, 0x000000ff },
+	{ 0x0017d4,   1, 0x04, 0xffffffff },
+	{ 0x0017d8,   1, 0x04, 0x00000002 },
+	{ 0x0017dc,   1, 0x04, 0x00000000 },
+	{ 0x0015f4,   2, 0x04, 0x00000000 },
+	{ 0x001434,   2, 0x04, 0x00000000 },
+	{ 0x000d74,   1, 0x04, 0x00000000 },
+	{ 0x000dec,   1, 0x04, 0x00000001 },
+	{ 0x0013a4,   1, 0x04, 0x00000000 },
+	{ 0x001318,   1, 0x04, 0x00000001 },
+	{ 0x001644,   1, 0x04, 0x00000000 },
+	{ 0x000748,   1, 0x04, 0x00000000 },
+	{ 0x000de8,   1, 0x04, 0x00000000 },
+	{ 0x001648,   1, 0x04, 0x00000000 },
+	{ 0x0012a4,   1, 0x04, 0x00000000 },
+	{ 0x001120,   4, 0x04, 0x00000000 },
+	{ 0x001118,   1, 0x04, 0x00000000 },
+	{ 0x00164c,   1, 0x04, 0x00000000 },
+	{ 0x001658,   1, 0x04, 0x00000000 },
+	{ 0x001910,   1, 0x04, 0x00000290 },
+	{ 0x001518,   1, 0x04, 0x00000000 },
+	{ 0x00165c,   1, 0x04, 0x00000001 },
+	{ 0x001520,   1, 0x04, 0x00000000 },
+	{ 0x001604,   1, 0x04, 0x00000000 },
+	{ 0x001570,   1, 0x04, 0x00000000 },
+	{ 0x0013b0,   2, 0x04, 0x3f800000 },
+	{ 0x00020c,   1, 0x04, 0x00000000 },
+	{ 0x001670,   1, 0x04, 0x30201000 },
+	{ 0x001674,   1, 0x04, 0x70605040 },
+	{ 0x001678,   1, 0x04, 0xb8a89888 },
+	{ 0x00167c,   1, 0x04, 0xf8e8d8c8 },
+	{ 0x00166c,   1, 0x04, 0x00000000 },
+	{ 0x001680,   1, 0x04, 0x00ffff00 },
+	{ 0x0012d0,   1, 0x04, 0x00000003 },
+	{ 0x0012d4,   1, 0x04, 0x00000002 },
+	{ 0x001684,   2, 0x04, 0x00000000 },
+	{ 0x000dac,   2, 0x04, 0x00001b02 },
+	{ 0x000db4,   1, 0x04, 0x00000000 },
+	{ 0x00168c,   1, 0x04, 0x00000000 },
+	{ 0x0015bc,   1, 0x04, 0x00000000 },
+	{ 0x00156c,   1, 0x04, 0x00000000 },
+	{ 0x00187c,   1, 0x04, 0x00000000 },
+	{ 0x001110,   1, 0x04, 0x00000001 },
+	{ 0x000dc0,   3, 0x04, 0x00000000 },
+	{ 0x001234,   1, 0x04, 0x00000000 },
+	{ 0x001690,   1, 0x04, 0x00000000 },
+	{ 0x0012ac,   1, 0x04, 0x00000001 },
+	{ 0x0002c4,   1, 0x04, 0x00000000 },
+	{ 0x000790,   5, 0x04, 0x00000000 },
+	{ 0x00077c,   1, 0x04, 0x00000000 },
+	{ 0x001000,   1, 0x04, 0x00000010 },
+	{ 0x0010fc,   1, 0x04, 0x00000000 },
+	{ 0x001290,   1, 0x04, 0x00000000 },
+	{ 0x000218,   1, 0x04, 0x00000010 },
+	{ 0x0012d8,   1, 0x04, 0x00000000 },
+	{ 0x0012dc,   1, 0x04, 0x00000010 },
+	{ 0x000d94,   1, 0x04, 0x00000001 },
+	{ 0x00155c,   2, 0x04, 0x00000000 },
+	{ 0x001564,   1, 0x04, 0x00001fff },
+	{ 0x001574,   2, 0x04, 0x00000000 },
+	{ 0x00157c,   1, 0x04, 0x003fffff },
+	{ 0x001354,   1, 0x04, 0x00000000 },
+	{ 0x001664,   1, 0x04, 0x00000000 },
+	{ 0x001610,   1, 0x04, 0x00000012 },
+	{ 0x001608,   2, 0x04, 0x00000000 },
+	{ 0x00162c,   1, 0x04, 0x00000003 },
+	{ 0x000210,   1, 0x04, 0x00000000 },
+	{ 0x000320,   1, 0x04, 0x00000000 },
+	{ 0x000324,   6, 0x04, 0x3f800000 },
+	{ 0x000750,   1, 0x04, 0x00000000 },
+	{ 0x000760,   1, 0x04, 0x39291909 },
+	{ 0x000764,   1, 0x04, 0x79695949 },
+	{ 0x000768,   1, 0x04, 0xb9a99989 },
+	{ 0x00076c,   1, 0x04, 0xf9e9d9c9 },
+	{ 0x000770,   1, 0x04, 0x30201000 },
+	{ 0x000774,   1, 0x04, 0x70605040 },
+	{ 0x000778,   1, 0x04, 0x00009080 },
+	{ 0x000780,   1, 0x04, 0x39291909 },
+	{ 0x000784,   1, 0x04, 0x79695949 },
+	{ 0x000788,   1, 0x04, 0xb9a99989 },
+	{ 0x00078c,   1, 0x04, 0xf9e9d9c9 },
+	{ 0x0007d0,   1, 0x04, 0x30201000 },
+	{ 0x0007d4,   1, 0x04, 0x70605040 },
+	{ 0x0007d8,   1, 0x04, 0x00009080 },
+	{ 0x00037c,   1, 0x04, 0x00000001 },
+	{ 0x000740,   2, 0x04, 0x00000000 },
+	{ 0x002600,   1, 0x04, 0x00000000 },
+	{ 0x001918,   1, 0x04, 0x00000000 },
+	{ 0x00191c,   1, 0x04, 0x00000900 },
+	{ 0x001920,   1, 0x04, 0x00000405 },
+	{ 0x001308,   1, 0x04, 0x00000001 },
+	{ 0x001924,   1, 0x04, 0x00000000 },
+	{ 0x0013ac,   1, 0x04, 0x00000000 },
+	{ 0x00192c,   1, 0x04, 0x00000001 },
+	{ 0x00193c,   1, 0x04, 0x00002c1c },
+	{ 0x000d7c,   1, 0x04, 0x00000000 },
+	{ 0x000f8c,   1, 0x04, 0x00000000 },
+	{ 0x0002c0,   1, 0x04, 0x00000001 },
+	{ 0x001510,   1, 0x04, 0x00000000 },
+	{ 0x001940,   1, 0x04, 0x00000000 },
+	{ 0x000ff4,   2, 0x04, 0x00000000 },
+	{ 0x00194c,   2, 0x04, 0x00000000 },
+	{ 0x001968,   1, 0x04, 0x00000000 },
+	{ 0x001590,   1, 0x04, 0x0000003f },
+	{ 0x0007e8,   4, 0x04, 0x00000000 },
+	{ 0x00196c,   1, 0x04, 0x00000011 },
+	{ 0x00197c,   1, 0x04, 0x00000000 },
+	{ 0x000fcc,   2, 0x04, 0x00000000 },
+	{ 0x0002d8,   1, 0x04, 0x00000040 },
+	{ 0x001980,   1, 0x04, 0x00000080 },
+	{ 0x001504,   1, 0x04, 0x00000080 },
+	{ 0x001984,   1, 0x04, 0x00000000 },
+	{ 0x000300,   1, 0x04, 0x00000001 },
+	{ 0x0013a8,   1, 0x04, 0x00000000 },
+	{ 0x0012ec,   1, 0x04, 0x00000000 },
+	{ 0x001310,   1, 0x04, 0x00000000 },
+	{ 0x001314,   1, 0x04, 0x00000001 },
+	{ 0x001380,   1, 0x04, 0x00000000 },
+	{ 0x001384,   4, 0x04, 0x00000001 },
+	{ 0x001394,   1, 0x04, 0x00000000 },
+	{ 0x00139c,   1, 0x04, 0x00000000 },
+	{ 0x001398,   1, 0x04, 0x00000000 },
+	{ 0x001594,   1, 0x04, 0x00000000 },
+	{ 0x001598,   4, 0x04, 0x00000001 },
+	{ 0x000f54,   3, 0x04, 0x00000000 },
+	{ 0x0019bc,   1, 0x04, 0x00000000 },
+	{ 0x000f9c,   2, 0x04, 0x00000000 },
+	{ 0x0012cc,   1, 0x04, 0x00000000 },
+	{ 0x0012e8,   1, 0x04, 0x00000000 },
+	{ 0x00130c,   1, 0x04, 0x00000001 },
+	{ 0x001360,   8, 0x04, 0x00000000 },
+	{ 0x00133c,   2, 0x04, 0x00000001 },
+	{ 0x001344,   1, 0x04, 0x00000002 },
+	{ 0x001348,   2, 0x04, 0x00000001 },
+	{ 0x001350,   1, 0x04, 0x00000002 },
+	{ 0x001358,   1, 0x04, 0x00000001 },
+	{ 0x0012e4,   1, 0x04, 0x00000000 },
+	{ 0x00131c,   1, 0x04, 0x00000000 },
+	{ 0x001320,   3, 0x04, 0x00000000 },
+	{ 0x0019c0,   1, 0x04, 0x00000000 },
+	{ 0x001140,   1, 0x04, 0x00000000 },
+	{ 0x0019c4,   1, 0x04, 0x00000000 },
+	{ 0x0019c8,   1, 0x04, 0x00001500 },
+	{ 0x00135c,   1, 0x04, 0x00000000 },
+	{ 0x000f90,   1, 0x04, 0x00000000 },
+	{ 0x0019e0,   8, 0x04, 0x00000001 },
+	{ 0x0019cc,   1, 0x04, 0x00000001 },
+	{ 0x0015b8,   1, 0x04, 0x00000000 },
+	{ 0x001a00,   1, 0x04, 0x00001111 },
+	{ 0x001a04,   7, 0x04, 0x00000000 },
+	{ 0x000d6c,   2, 0x04, 0xffff0000 },
+	{ 0x0010f8,   1, 0x04, 0x00001010 },
+	{ 0x000d80,   5, 0x04, 0x00000000 },
+	{ 0x000da0,   1, 0x04, 0x00000000 },
+	{ 0x001508,   1, 0x04, 0x80000000 },
+	{ 0x00150c,   1, 0x04, 0x40000000 },
+	{ 0x001668,   1, 0x04, 0x00000000 },
+	{ 0x000318,   2, 0x04, 0x00000008 },
+	{ 0x000d9c,   1, 0x04, 0x00000001 },
+	{ 0x0007dc,   1, 0x04, 0x00000000 },
+	{ 0x00074c,   1, 0x04, 0x00000055 },
+	{ 0x001420,   1, 0x04, 0x00000003 },
+	{ 0x0017bc,   2, 0x04, 0x00000000 },
+	{ 0x0017c4,   1, 0x04, 0x00000001 },
+	{ 0x001008,   1, 0x04, 0x00000008 },
+	{ 0x00100c,   1, 0x04, 0x00000040 },
+	{ 0x001010,   1, 0x04, 0x0000012c },
+	{ 0x000d60,   1, 0x04, 0x00000040 },
+	{ 0x00075c,   1, 0x04, 0x00000003 },
+	{ 0x001018,   1, 0x04, 0x00000020 },
+	{ 0x00101c,   1, 0x04, 0x00000001 },
+	{ 0x001020,   1, 0x04, 0x00000020 },
+	{ 0x001024,   1, 0x04, 0x00000001 },
+	{ 0x001444,   3, 0x04, 0x00000000 },
+	{ 0x000360,   1, 0x04, 0x20164010 },
+	{ 0x000364,   1, 0x04, 0x00000020 },
+	{ 0x000368,   1, 0x04, 0x00000000 },
+	{ 0x000de4,   1, 0x04, 0x00000000 },
+	{ 0x000204,   1, 0x04, 0x00000006 },
+	{ 0x000208,   1, 0x04, 0x00000000 },
+	{ 0x0002cc,   1, 0x04, 0x003fffff },
+	{ 0x0002d0,   1, 0x04, 0x00000c48 },
+	{ 0x001220,   1, 0x04, 0x00000005 },
+	{ 0x000fdc,   1, 0x04, 0x00000000 },
+	{ 0x000f98,   1, 0x04, 0x00300008 },
+	{ 0x001284,   1, 0x04, 0x04000080 },
+	{ 0x001450,   1, 0x04, 0x00300008 },
+	{ 0x001454,   1, 0x04, 0x04000080 },
+	{ 0x000214,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_902d[] = {
+	{ 0x000200,   1, 0x04, 0x000000cf },
+	{ 0x000204,   1, 0x04, 0x00000001 },
+	{ 0x000208,   1, 0x04, 0x00000020 },
+	{ 0x00020c,   1, 0x04, 0x00000001 },
+	{ 0x000210,   1, 0x04, 0x00000000 },
+	{ 0x000214,   1, 0x04, 0x00000080 },
+	{ 0x000218,   2, 0x04, 0x00000100 },
+	{ 0x000220,   2, 0x04, 0x00000000 },
+	{ 0x000230,   1, 0x04, 0x000000cf },
+	{ 0x000234,   1, 0x04, 0x00000001 },
+	{ 0x000238,   1, 0x04, 0x00000020 },
+	{ 0x00023c,   1, 0x04, 0x00000001 },
+	{ 0x000244,   1, 0x04, 0x00000080 },
+	{ 0x000248,   2, 0x04, 0x00000100 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_9039[] = {
+	{ 0x00030c,   3, 0x04, 0x00000000 },
+	{ 0x000320,   1, 0x04, 0x00000000 },
+	{ 0x000238,   2, 0x04, 0x00000000 },
+	{ 0x000318,   2, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_90c0[] = {
+	{ 0x00270c,   8, 0x20, 0x00000000 },
+	{ 0x00030c,   1, 0x04, 0x00000001 },
+	{ 0x001944,   1, 0x04, 0x00000000 },
+	{ 0x000758,   1, 0x04, 0x00000100 },
+	{ 0x0002c4,   1, 0x04, 0x00000000 },
+	{ 0x000790,   5, 0x04, 0x00000000 },
+	{ 0x00077c,   1, 0x04, 0x00000000 },
+	{ 0x000204,   3, 0x04, 0x00000000 },
+	{ 0x000214,   1, 0x04, 0x00000000 },
+	{ 0x00024c,   1, 0x04, 0x00000000 },
+	{ 0x000d94,   1, 0x04, 0x00000001 },
+	{ 0x001608,   2, 0x04, 0x00000000 },
+	{ 0x001664,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_base[] = {
+	{ 0x400204,   2, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk40xx[] = {
+	{ 0x404004,  10, 0x04, 0x00000000 },
+	{ 0x404044,   1, 0x04, 0x00000000 },
+	{ 0x404094,   1, 0x04, 0x00000000 },
+	{ 0x404098,  12, 0x04, 0x00000000 },
+	{ 0x4040c8,   1, 0x04, 0xf0000087 },
+	{ 0x4040d0,   6, 0x04, 0x00000000 },
+	{ 0x4040e8,   1, 0x04, 0x00001000 },
+	{ 0x4040f8,   1, 0x04, 0x00000000 },
+	{ 0x404130,   1, 0x04, 0x00000000 },
+	{ 0x404134,   1, 0x04, 0x00000000 },
+	{ 0x404138,   1, 0x04, 0x20000040 },
+	{ 0x404150,   1, 0x04, 0x0000002e },
+	{ 0x404154,   1, 0x04, 0x00000400 },
+	{ 0x404158,   1, 0x04, 0x00000200 },
+	{ 0x404164,   1, 0x04, 0x00000055 },
+	{ 0x404168,   1, 0x04, 0x00000000 },
+	{ 0x404174,   1, 0x04, 0x00000000 },
+	{ 0x404178,   2, 0x04, 0x00000000 },
+	{ 0x404200,   8, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk44xx[] = {
+	{ 0x404404,  14, 0x04, 0x00000000 },
+	{ 0x404460,   2, 0x04, 0x00000000 },
+	{ 0x404468,   1, 0x04, 0x00ffffff },
+	{ 0x40446c,   1, 0x04, 0x00000000 },
+	{ 0x404480,   1, 0x04, 0x00000001 },
+	{ 0x404498,   1, 0x04, 0x00000001 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk46xx[] = {
+	{ 0x404604,   1, 0x04, 0x00000015 },
+	{ 0x404608,   1, 0x04, 0x00000000 },
+	{ 0x40460c,   1, 0x04, 0x00002e00 },
+	{ 0x404610,   1, 0x04, 0x00000100 },
+	{ 0x404618,   8, 0x04, 0x00000000 },
+	{ 0x404638,   1, 0x04, 0x00000004 },
+	{ 0x40463c,   8, 0x04, 0x00000000 },
+	{ 0x40465c,   1, 0x04, 0x007f0100 },
+	{ 0x404660,   7, 0x04, 0x00000000 },
+	{ 0x40467c,   1, 0x04, 0x00000002 },
+	{ 0x404680,   8, 0x04, 0x00000000 },
+	{ 0x4046a0,   1, 0x04, 0x007f0080 },
+	{ 0x4046a4,  18, 0x04, 0x00000000 },
+	{ 0x4046f0,   2, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk47xx[] = {
+	{ 0x404700,  13, 0x04, 0x00000000 },
+	{ 0x404734,   1, 0x04, 0x00000100 },
+	{ 0x404738,   8, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk58xx[] = {
+	{ 0x405800,   1, 0x04, 0x078000bf },
+	{ 0x405830,   1, 0x04, 0x02180000 },
+	{ 0x405834,   2, 0x04, 0x00000000 },
+	{ 0x405854,   1, 0x04, 0x00000000 },
+	{ 0x405870,   4, 0x04, 0x00000001 },
+	{ 0x405a00,   2, 0x04, 0x00000000 },
+	{ 0x405a18,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk60xx[] = {
+	{ 0x406020,   1, 0x04, 0x000103c1 },
+	{ 0x406028,   4, 0x04, 0x00000001 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk64xx[] = {
+	{ 0x4064a8,   1, 0x04, 0x00000000 },
+	{ 0x4064ac,   1, 0x04, 0x00003fff },
+	{ 0x4064b4,   2, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk78xx[] = {
+	{ 0x407804,   1, 0x04, 0x00000023 },
+	{ 0x40780c,   1, 0x04, 0x0a418820 },
+	{ 0x407810,   1, 0x04, 0x062080e6 },
+	{ 0x407814,   1, 0x04, 0x020398a4 },
+	{ 0x407818,   1, 0x04, 0x0e629062 },
+	{ 0x40781c,   1, 0x04, 0x0a418820 },
+	{ 0x407820,   1, 0x04, 0x000000e6 },
+	{ 0x4078bc,   1, 0x04, 0x00000103 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_unk80xx[] = {
+	{ 0x408000,   2, 0x04, 0x00000000 },
+	{ 0x408008,   1, 0x04, 0x00000018 },
+	{ 0x40800c,   2, 0x04, 0x00000000 },
+	{ 0x408014,   1, 0x04, 0x00000069 },
+	{ 0x408018,   1, 0x04, 0xe100e100 },
+	{ 0x408064,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_rop[] = {
+	{ 0x408800,   1, 0x04, 0x02802a3c },
+	{ 0x408804,   1, 0x04, 0x00000040 },
+	{ 0x408808,   1, 0x04, 0x0003e00d },
+	{ 0x408900,   1, 0x04, 0x3080b801 },
+	{ 0x408904,   1, 0x04, 0x02000001 },
+	{ 0x408908,   1, 0x04, 0x00c80929 },
+	{ 0x408980,   1, 0x04, 0x0000011d },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_gpc_0[] = {
+	{ 0x418380,   1, 0x04, 0x00000016 },
+	{ 0x418400,   1, 0x04, 0x38004e00 },
+	{ 0x418404,   1, 0x04, 0x71e0ffff },
+	{ 0x418408,   1, 0x04, 0x00000000 },
+	{ 0x41840c,   1, 0x04, 0x00001008 },
+	{ 0x418410,   1, 0x04, 0x0fff0fff },
+	{ 0x418414,   1, 0x04, 0x00200fff },
+	{ 0x418450,   6, 0x04, 0x00000000 },
+	{ 0x418468,   1, 0x04, 0x00000001 },
+	{ 0x41846c,   2, 0x04, 0x00000000 },
+	{ 0x418600,   1, 0x04, 0x0000001f },
+	{ 0x418684,   1, 0x04, 0x0000000f },
+	{ 0x418700,   1, 0x04, 0x00000002 },
+	{ 0x418704,   1, 0x04, 0x00000080 },
+	{ 0x418708,   1, 0x04, 0x00000000 },
+	{ 0x41870c,   1, 0x04, 0x07c80000 },
+	{ 0x418710,   1, 0x04, 0x00000000 },
+	{ 0x418800,   1, 0x04, 0x0006860a },
+	{ 0x418808,   3, 0x04, 0x00000000 },
+	{ 0x418828,   1, 0x04, 0x00008442 },
+	{ 0x418830,   1, 0x04, 0x00000001 },
+	{ 0x4188d8,   1, 0x04, 0x00000008 },
+	{ 0x4188e0,   1, 0x04, 0x01000000 },
+	{ 0x4188e8,   5, 0x04, 0x00000000 },
+	{ 0x4188fc,   1, 0x04, 0x00100000 },
+	{ 0x41891c,   1, 0x04, 0x00ff00ff },
+	{ 0x418924,   1, 0x04, 0x00000000 },
+	{ 0x418928,   1, 0x04, 0x00ffff00 },
+	{ 0x41892c,   1, 0x04, 0x0000ff00 },
+	{ 0x418b00,   1, 0x04, 0x00000000 },
+	{ 0x418b08,   1, 0x04, 0x0a418820 },
+	{ 0x418b0c,   1, 0x04, 0x062080e6 },
+	{ 0x418b10,   1, 0x04, 0x020398a4 },
+	{ 0x418b14,   1, 0x04, 0x0e629062 },
+	{ 0x418b18,   1, 0x04, 0x0a418820 },
+	{ 0x418b1c,   1, 0x04, 0x000000e6 },
+	{ 0x418bb8,   1, 0x04, 0x00000103 },
+	{ 0x418c08,   1, 0x04, 0x00000001 },
+	{ 0x418c10,   8, 0x04, 0x00000000 },
+	{ 0x418c80,   1, 0x04, 0x20200004 },
+	{ 0x418c8c,   1, 0x04, 0x00000001 },
+	{ 0x419000,   1, 0x04, 0x00000780 },
+	{ 0x419004,   2, 0x04, 0x00000000 },
+	{ 0x419014,   1, 0x04, 0x00000004 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_gpc_1[] = {
+	{ 0x418a00,   3, 0x04, 0x00000000 },
+	{ 0x418a0c,   1, 0x04, 0x00010000 },
+	{ 0x418a10,   3, 0x04, 0x00000000 },
+	{ 0x418a20,   3, 0x04, 0x00000000 },
+	{ 0x418a2c,   1, 0x04, 0x00010000 },
+	{ 0x418a30,   3, 0x04, 0x00000000 },
+	{ 0x418a40,   3, 0x04, 0x00000000 },
+	{ 0x418a4c,   1, 0x04, 0x00010000 },
+	{ 0x418a50,   3, 0x04, 0x00000000 },
+	{ 0x418a60,   3, 0x04, 0x00000000 },
+	{ 0x418a6c,   1, 0x04, 0x00010000 },
+	{ 0x418a70,   3, 0x04, 0x00000000 },
+	{ 0x418a80,   3, 0x04, 0x00000000 },
+	{ 0x418a8c,   1, 0x04, 0x00010000 },
+	{ 0x418a90,   3, 0x04, 0x00000000 },
+	{ 0x418aa0,   3, 0x04, 0x00000000 },
+	{ 0x418aac,   1, 0x04, 0x00010000 },
+	{ 0x418ab0,   3, 0x04, 0x00000000 },
+	{ 0x418ac0,   3, 0x04, 0x00000000 },
+	{ 0x418acc,   1, 0x04, 0x00010000 },
+	{ 0x418ad0,   3, 0x04, 0x00000000 },
+	{ 0x418ae0,   3, 0x04, 0x00000000 },
+	{ 0x418aec,   1, 0x04, 0x00010000 },
+	{ 0x418af0,   3, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_tpc[] = {
+	{ 0x419818,   1, 0x04, 0x00000000 },
+	{ 0x41983c,   1, 0x04, 0x00038bc7 },
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x0000012a },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000001f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000023 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419b00,   1, 0x04, 0x0a418820 },
+	{ 0x419b04,   1, 0x04, 0x062080e6 },
+	{ 0x419b08,   1, 0x04, 0x020398a4 },
+	{ 0x419b0c,   1, 0x04, 0x0e629062 },
+	{ 0x419b10,   1, 0x04, 0x0a418820 },
+	{ 0x419b14,   1, 0x04, 0x000000e6 },
+	{ 0x419bd0,   1, 0x04, 0x00900103 },
+	{ 0x419be0,   1, 0x04, 0x00000001 },
+	{ 0x419be4,   1, 0x04, 0x00000000 },
+	{ 0x419c00,   1, 0x04, 0x00000002 },
+	{ 0x419c04,   1, 0x04, 0x00000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x00060048 },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000183 },
+	{ 0x419d20,   1, 0x04, 0x02180000 },
+	{ 0x419d24,   1, 0x04, 0x00001fff },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000002 },
+	{ 0x419e44,   1, 0x04, 0x001beff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000000f },
+	{ 0x419e50,  17, 0x04, 0x00000000 },
+	{ 0x419e98,   1, 0x04, 0x00000000 },
+	{ 0x419f50,   2, 0x04, 0x00000000 },
+	{}
+};
 
-int
-nvc0_grctx_init(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+void
+nvc0_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 {
-	struct nouveau_bar *bar = nouveau_bar(priv);
-	struct nouveau_gpuobj *chan;
-	u32 size = (0x80000 + priv->size + 4095) & ~4095;
-	int ret, i;
-
-	/* allocate memory to for a "channel", which we'll use to generate
-	 * the default context values
-	 */
-	ret = nouveau_gpuobj_new(nv_object(priv), NULL, size, 0x1000,
-				 NVOBJ_FLAG_ZERO_ALLOC, &info->chan);
-	chan = info->chan;
-	if (ret) {
-		nv_error(priv, "failed to allocate channel memory, %d\n", ret);
-		return ret;
-	}
-
-	/* PGD pointer */
-	nv_wo32(chan, 0x0200, lower_32_bits(chan->addr + 0x1000));
-	nv_wo32(chan, 0x0204, upper_32_bits(chan->addr + 0x1000));
-	nv_wo32(chan, 0x0208, 0xffffffff);
-	nv_wo32(chan, 0x020c, 0x000000ff);
-
-	/* PGT[0] pointer */
-	nv_wo32(chan, 0x1000, 0x00000000);
-	nv_wo32(chan, 0x1004, 0x00000001 | (chan->addr + 0x2000) >> 8);
-
-	/* identity-map the whole "channel" into its own vm */
-	for (i = 0; i < size / 4096; i++) {
-		u64 addr = ((chan->addr + (i * 4096)) >> 8) | 1;
-		nv_wo32(chan, 0x2000 + (i * 8), lower_32_bits(addr));
-		nv_wo32(chan, 0x2004 + (i * 8), upper_32_bits(addr));
-	}
-
-	/* context pointer (virt) */
-	nv_wo32(chan, 0x0210, 0x00080004);
-	nv_wo32(chan, 0x0214, 0x00000000);
+	int gpc, tpc;
+	u32 offset;
 
-	bar->flush(bar);
-
-	nv_wr32(priv, 0x100cb8, (chan->addr + 0x1000) >> 8);
-	nv_wr32(priv, 0x100cbc, 0x80000001);
-	nv_wait(priv, 0x100c80, 0x00008000, 0x00008000);
-
-	/* setup default state for mmio list construction */
-	info->data = priv->mmio_data;
-	info->mmio = priv->mmio_list;
-	info->addr = 0x2000 + (i * 8);
-	info->priv = priv;
-	info->buffer_nr = 0;
+	mmio_data(0x002000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
 
-	if (priv->firmware) {
-		nv_wr32(priv, 0x409840, 0x00000030);
-		nv_wr32(priv, 0x409500, 0x80000000 | chan->addr >> 12);
-		nv_wr32(priv, 0x409504, 0x00000003);
-		if (!nv_wait(priv, 0x409800, 0x00000010, 0x00000010))
-			nv_error(priv, "load_ctx timeout\n");
+	mmio_list(0x408004, 0x00000000,  8, 0);
+	mmio_list(0x408008, 0x80000018,  0, 0);
+	mmio_list(0x40800c, 0x00000000,  8, 1);
+	mmio_list(0x408010, 0x80000000,  0, 0);
+	mmio_list(0x418810, 0x80000000, 12, 2);
+	mmio_list(0x419848, 0x10000000, 12, 2);
+	mmio_list(0x419004, 0x00000000,  8, 1);
+	mmio_list(0x419008, 0x00000000,  0, 0);
+	mmio_list(0x418808, 0x00000000,  8, 0);
+	mmio_list(0x41880c, 0x80000018,  0, 0);
 
-		nv_wo32(chan, 0x8001c, 1);
-		nv_wo32(chan, 0x80020, 0);
-		nv_wo32(chan, 0x80028, 0);
-		nv_wo32(chan, 0x8002c, 0);
-		bar->flush(bar);
-		return 0;
-	}
+	mmio_list(0x405830, 0x02180000, 0, 0);
 
-	/* HUB_FUC(SET_CHAN) */
-	nv_wr32(priv, 0x409840, 0x80000000);
-	nv_wr32(priv, 0x409500, 0x80000000 | chan->addr >> 12);
-	nv_wr32(priv, 0x409504, 0x00000001);
-	if (!nv_wait(priv, 0x409800, 0x80000000, 0x80000000)) {
-		nv_error(priv, "HUB_SET_CHAN timeout\n");
-		nvc0_graph_ctxctl_debug(priv);
-		nouveau_gpuobj_ref(NULL, &info->chan);
-		return -EBUSY;
+	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
+		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
+			u32 addr = TPC_UNIT(gpc, tpc, 0x0520);
+			mmio_list(addr, 0x02180000 | offset, 0, 0);
+			offset += 0x0324;
+		}
 	}
-
-	return 0;
 }
 
 void
-nvc0_grctx_data(struct nvc0_grctx *info, u32 size, u32 align, u32 access)
+nvc0_grctx_generate_unkn(struct nvc0_graph_priv *priv)
 {
-	info->buffer[info->buffer_nr]  = info->addr;
-	info->buffer[info->buffer_nr] +=  (align - 1);
-	info->buffer[info->buffer_nr] &= ~(align - 1);
-	info->addr = info->buffer[info->buffer_nr++] + size;
-
-	info->data->size = size;
-	info->data->align = align;
-	info->data->access = access;
-	info->data++;
 }
 
 void
-nvc0_grctx_mmio(struct nvc0_grctx *info, u32 addr, u32 data, u32 shift, u32 buf)
+nvc0_grctx_generate_tpcid(struct nvc0_graph_priv *priv)
 {
-	struct nvc0_graph_priv *priv = info->priv;
-
-	info->mmio->addr = addr;
-	info->mmio->data = data;
-	info->mmio->shift = shift;
-	info->mmio->buffer = buf;
-	info->mmio++;
+	int gpc, tpc, id;
 
-	if (shift)
-		data |= info->buffer[buf] >> shift;
-	nv_wr32(priv, addr, data);
-}
-
-int
-nvc0_grctx_fini(struct nvc0_grctx *info)
-{
-	struct nvc0_graph_priv *priv = info->priv;
-	int i;
-
-	/* trigger a context unload by unsetting the "next channel valid" bit
-	 * and faking a context switch interrupt
-	 */
-	nv_mask(priv, 0x409b04, 0x80000000, 0x00000000);
-	nv_wr32(priv, 0x409000, 0x00000100);
-	if (!nv_wait(priv, 0x409b00, 0x80000000, 0x00000000)) {
-		nv_error(priv, "grctx template channel unload timeout\n");
-		return -EBUSY;
-	}
-
-	priv->data = kmalloc(priv->size, GFP_KERNEL);
-	if (priv->data) {
-		for (i = 0; i < priv->size; i += 4)
-			priv->data[i / 4] = nv_ro32(info->chan, 0x80000 + i);
-	}
-
-	nouveau_gpuobj_ref(NULL, &info->chan);
-	return priv->data ? 0 : -ENOMEM;
-}
+	for (tpc = 0, id = 0; tpc < 4; tpc++) {
+		for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+			if (tpc < priv->tpc_nr[gpc]) {
+				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x698), id);
+				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x4e8), id);
+				nv_wr32(priv, GPC_UNIT(gpc, 0x0c10 + tpc * 4), id);
+				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x088), id);
+				id++;
+			}
 
-static void
-nvc0_grctx_generate_9097(struct nvc0_graph_priv *priv)
-{
-	u32 fermi = nvc0_graph_class(priv);
-	u32 mthd;
-
-	nv_mthd(priv, 0x9097, 0x0800, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0840, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0880, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x08c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0900, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0940, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0980, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x09c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0804, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0844, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0884, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x08c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0904, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0944, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0984, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x09c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0808, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x0848, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x0888, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x08c8, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x0908, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x0948, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x0988, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x09c8, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x080c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x084c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x088c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x08cc, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x090c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x094c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x098c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x09cc, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x0810, 0x000000cf);
-	nv_mthd(priv, 0x9097, 0x0850, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0890, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x08d0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0910, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0950, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0990, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x09d0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0814, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0854, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0894, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x08d4, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0914, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0954, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0994, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x09d4, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0818, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0858, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0898, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x08d8, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0918, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0958, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0998, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x09d8, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x081c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x085c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x089c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x08dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x091c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x095c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x099c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x09dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0820, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0860, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x08a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x08e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0920, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0960, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x09a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x09e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2700, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2720, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2740, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2760, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2780, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2704, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2724, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2744, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2764, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2784, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2708, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2728, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2748, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2768, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2788, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27a8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27c8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x270c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x272c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x274c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x276c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x278c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27ac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x27ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2710, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x2730, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x2750, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x2770, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x2790, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x27b0, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x27d0, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x27f0, 0x00014000);
-	nv_mthd(priv, 0x9097, 0x2714, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x2734, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x2754, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x2774, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x2794, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x27b4, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x27d4, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x27f4, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x1c00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ca0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cb0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cc0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cd0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ce0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cf0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c24, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c34, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c64, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c94, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ca4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cb4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cc4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cd4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ce4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cf4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c18, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c28, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c38, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c58, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c68, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c78, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c98, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ca8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cb8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cc8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cd8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ce8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cf8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c0c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c1c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c2c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c3c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c4c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c5c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c6c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c7c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1c9c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cbc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ccc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cdc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1cfc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1da0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1db0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dc0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dd0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1de0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1df0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d24, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d34, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d64, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d94, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1da4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1db4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dc4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dd4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1de4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1df4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d18, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d28, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d38, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d58, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d68, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d78, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d98, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1da8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1db8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dc8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dd8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1de8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1df8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d0c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d1c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d2c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d3c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d4c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d5c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d6c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d7c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1d9c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dbc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dcc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ddc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1dfc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f18, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f28, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f38, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f58, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f68, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f78, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f0c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f1c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f24, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f2c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f34, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f3c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f4c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f5c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f64, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f6c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f7c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f98, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fa0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fa8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fb0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fb8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fc0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fc8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fd0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fd8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fe0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fe8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ff0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ff8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f94, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1f9c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fa4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fb4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fbc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fc4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fcc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fd4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fdc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fe4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1fec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ff4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1ffc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2200, 0x00000022);
-	nv_mthd(priv, 0x9097, 0x2210, 0x00000022);
-	nv_mthd(priv, 0x9097, 0x2220, 0x00000022);
-	nv_mthd(priv, 0x9097, 0x2230, 0x00000022);
-	nv_mthd(priv, 0x9097, 0x2240, 0x00000022);
-	nv_mthd(priv, 0x9097, 0x2000, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2040, 0x00000011);
-	nv_mthd(priv, 0x9097, 0x2080, 0x00000020);
-	nv_mthd(priv, 0x9097, 0x20c0, 0x00000030);
-	nv_mthd(priv, 0x9097, 0x2100, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x2140, 0x00000051);
-	nv_mthd(priv, 0x9097, 0x200c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x204c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x208c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x20cc, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x210c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x214c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x2010, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2050, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2090, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x20d0, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x2110, 0x00000003);
-	nv_mthd(priv, 0x9097, 0x2150, 0x00000004);
-	nv_mthd(priv, 0x9097, 0x0380, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0384, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0388, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03a8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03c8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x038c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03ac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x03ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0700, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0710, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0720, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0730, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0704, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0714, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0724, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0734, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0708, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0718, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0728, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0738, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2800, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2804, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2808, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x280c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2810, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2814, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2818, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x281c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2820, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2824, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2828, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x282c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2830, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2834, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2838, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x283c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2840, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2844, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2848, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x284c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2850, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2854, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2858, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x285c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2860, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2864, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2868, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x286c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2870, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2874, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2878, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x287c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2880, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2884, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2888, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x288c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2890, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2894, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2898, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x289c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28a8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28ac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28b0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28b4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28b8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28bc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28c8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28d0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28d4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28d8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28f0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28f4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28f8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x28fc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2900, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2904, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2908, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x290c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2910, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2914, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2918, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x291c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2920, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2924, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2928, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x292c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2930, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2934, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2938, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x293c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2940, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2944, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2948, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x294c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2950, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2954, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2958, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x295c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2960, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2964, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2968, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x296c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2970, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2974, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2978, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x297c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2980, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2984, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2988, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x298c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2990, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2994, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2998, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x299c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29a8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29ac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29b0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29b4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29b8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29bc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29c8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29d0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29d4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29d8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29f0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29f4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29f8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x29fc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0aa0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ac0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ae0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ba0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bc0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0be0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a24, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a64, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0aa4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ac4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ae4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b24, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b64, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ba4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bc4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0be4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a28, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a68, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0aa8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ac8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ae8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b28, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b68, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ba8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bc8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0be8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a0c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a2c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a4c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a6c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0aac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0acc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0aec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b0c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b2c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b4c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b6c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bcc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ab0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ad0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0af0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bb0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bd0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bf0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a34, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0a94, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ab4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ad4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0af4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b34, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0b94, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bb4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bd4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0bf4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ca0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cb0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cc0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cd0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ce0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cf0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c24, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c34, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c64, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c94, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ca4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cb4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cc4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cd4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ce4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cf4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c18, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c28, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c38, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c58, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c68, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c78, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c98, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ca8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cb8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cc8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cd8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ce8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0cf8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0c0c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c1c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c2c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c3c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c4c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c5c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c6c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c7c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c8c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0c9c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0cac, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0cbc, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0ccc, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0cdc, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0cec, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0cfc, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0d00, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d08, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d10, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d18, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d20, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d28, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d30, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d38, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d04, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d0c, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d14, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d1c, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d24, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d2c, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d34, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d3c, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e00, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e20, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e30, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e60, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e70, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ea0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0eb0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ec0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ed0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ee0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ef0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0e04, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e14, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e24, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e34, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e44, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e54, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e64, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e74, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e84, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e94, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ea4, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0eb4, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ec4, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ed4, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ee4, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ef4, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e08, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e18, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e28, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e38, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e48, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e58, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e68, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e78, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e88, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0e98, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ea8, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0eb8, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ec8, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ed8, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ee8, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0ef8, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d40, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d48, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d50, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d58, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d44, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d4c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d5c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1e00, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e20, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e40, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e60, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e80, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ea0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ec0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ee0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e04, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e24, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e44, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e64, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e84, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ea4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ec4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ee4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e08, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e28, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e48, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e68, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e88, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1ea8, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1ec8, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1ee8, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e0c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e2c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e4c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e6c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e8c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1eac, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ecc, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1eec, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e10, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e30, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e50, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e70, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e90, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1eb0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ed0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ef0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e14, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e34, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e54, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e74, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e94, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1eb4, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1ed4, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1ef4, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1e18, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e38, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e58, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e78, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1e98, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1eb8, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ed8, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1ef8, 0x00000001);
-	if (fermi == 0x9097) {
-		for (mthd = 0x3400; mthd <= 0x35fc; mthd += 4)
-			nv_mthd(priv, 0x9097, mthd, 0x00000000);
+			nv_wr32(priv, GPC_UNIT(gpc, 0x0c08), priv->tpc_nr[gpc]);
+			nv_wr32(priv, GPC_UNIT(gpc, 0x0c8c), priv->tpc_nr[gpc]);
+		}
 	}
-	nv_mthd(priv, 0x9097, 0x030c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1944, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1514, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d68, 0x0000ffff);
-	nv_mthd(priv, 0x9097, 0x121c, 0x0fac6881);
-	nv_mthd(priv, 0x9097, 0x0fac, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1538, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0fe0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0fe4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0fe8, 0x00000014);
-	nv_mthd(priv, 0x9097, 0x0fec, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x0ff0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x179c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1228, 0x00000400);
-	nv_mthd(priv, 0x9097, 0x122c, 0x00000300);
-	nv_mthd(priv, 0x9097, 0x1230, 0x00010001);
-	nv_mthd(priv, 0x9097, 0x07f8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x15b4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x15cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1534, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0fb0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x15d0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x153c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x16b4, 0x00000003);
-	nv_mthd(priv, 0x9097, 0x0fbc, 0x0000ffff);
-	nv_mthd(priv, 0x9097, 0x0fc0, 0x0000ffff);
-	nv_mthd(priv, 0x9097, 0x0fc4, 0x0000ffff);
-	nv_mthd(priv, 0x9097, 0x0fc8, 0x0000ffff);
-	nv_mthd(priv, 0x9097, 0x0df8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0dfc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1948, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1970, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x161c, 0x000009f0);
-	nv_mthd(priv, 0x9097, 0x0dcc, 0x00000010);
-	nv_mthd(priv, 0x9097, 0x163c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x15e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1160, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1164, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1168, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x116c, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1170, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1174, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1178, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x117c, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1180, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1184, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1188, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x118c, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1190, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1194, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1198, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x119c, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11a0, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11a4, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11a8, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11ac, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11b0, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11b4, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11b8, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11bc, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11c0, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11c4, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11c8, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11cc, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11d0, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11d4, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11d8, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x11dc, 0x25e00040);
-	nv_mthd(priv, 0x9097, 0x1880, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1884, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1888, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x188c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1890, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1894, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1898, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x189c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18a8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18ac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18b0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18b4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18b8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18bc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18c8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18d0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18d4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18d8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18e0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18f0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18f4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18f8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x18fc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x17c8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x17cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x17d0, 0x000000ff);
-	nv_mthd(priv, 0x9097, 0x17d4, 0xffffffff);
-	nv_mthd(priv, 0x9097, 0x17d8, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x17dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x15f4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x15f8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1434, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1438, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d74, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0dec, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x13a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1318, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1644, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0748, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0de8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1648, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x12a4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1120, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1124, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1128, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x112c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1118, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x164c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1658, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1910, 0x00000290);
-	nv_mthd(priv, 0x9097, 0x1518, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x165c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1520, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1604, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1570, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x13b0, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x13b4, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x020c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1670, 0x30201000);
-	nv_mthd(priv, 0x9097, 0x1674, 0x70605040);
-	nv_mthd(priv, 0x9097, 0x1678, 0xb8a89888);
-	nv_mthd(priv, 0x9097, 0x167c, 0xf8e8d8c8);
-	nv_mthd(priv, 0x9097, 0x166c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1680, 0x00ffff00);
-	nv_mthd(priv, 0x9097, 0x12d0, 0x00000003);
-	nv_mthd(priv, 0x9097, 0x12d4, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1684, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1688, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0dac, 0x00001b02);
-	nv_mthd(priv, 0x9097, 0x0db0, 0x00001b02);
-	nv_mthd(priv, 0x9097, 0x0db4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x168c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x15bc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x156c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x187c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1110, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0dc0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0dc4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0dc8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1234, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1690, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x12ac, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x02c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0790, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0794, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0798, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x079c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x07a0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x077c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1000, 0x00000010);
-	nv_mthd(priv, 0x9097, 0x10fc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1290, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0218, 0x00000010);
-	nv_mthd(priv, 0x9097, 0x12d8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x12dc, 0x00000010);
-	nv_mthd(priv, 0x9097, 0x0d94, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x155c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1560, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1564, 0x00001fff);
-	nv_mthd(priv, 0x9097, 0x1574, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1578, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x157c, 0x003fffff);
-	nv_mthd(priv, 0x9097, 0x1354, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1664, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1610, 0x00000012);
-	nv_mthd(priv, 0x9097, 0x1608, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x160c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x162c, 0x00000003);
-	nv_mthd(priv, 0x9097, 0x0210, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0320, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0324, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0328, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x032c, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0330, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0334, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0338, 0x3f800000);
-	nv_mthd(priv, 0x9097, 0x0750, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0760, 0x39291909);
-	nv_mthd(priv, 0x9097, 0x0764, 0x79695949);
-	nv_mthd(priv, 0x9097, 0x0768, 0xb9a99989);
-	nv_mthd(priv, 0x9097, 0x076c, 0xf9e9d9c9);
-	nv_mthd(priv, 0x9097, 0x0770, 0x30201000);
-	nv_mthd(priv, 0x9097, 0x0774, 0x70605040);
-	nv_mthd(priv, 0x9097, 0x0778, 0x00009080);
-	nv_mthd(priv, 0x9097, 0x0780, 0x39291909);
-	nv_mthd(priv, 0x9097, 0x0784, 0x79695949);
-	nv_mthd(priv, 0x9097, 0x0788, 0xb9a99989);
-	nv_mthd(priv, 0x9097, 0x078c, 0xf9e9d9c9);
-	nv_mthd(priv, 0x9097, 0x07d0, 0x30201000);
-	nv_mthd(priv, 0x9097, 0x07d4, 0x70605040);
-	nv_mthd(priv, 0x9097, 0x07d8, 0x00009080);
-	nv_mthd(priv, 0x9097, 0x037c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0740, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0744, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x2600, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1918, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x191c, 0x00000900);
-	nv_mthd(priv, 0x9097, 0x1920, 0x00000405);
-	nv_mthd(priv, 0x9097, 0x1308, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1924, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x13ac, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x192c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x193c, 0x00002c1c);
-	nv_mthd(priv, 0x9097, 0x0d7c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x02c0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1510, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1940, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ff4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0ff8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x194c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1950, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1968, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1590, 0x0000003f);
-	nv_mthd(priv, 0x9097, 0x07e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x07ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x07f0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x07f4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x196c, 0x00000011);
-	nv_mthd(priv, 0x9097, 0x197c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0fcc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0fd0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x02d8, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x1980, 0x00000080);
-	nv_mthd(priv, 0x9097, 0x1504, 0x00000080);
-	nv_mthd(priv, 0x9097, 0x1984, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0300, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x13a8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x12ec, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1310, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1314, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1380, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1384, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1388, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x138c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1390, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1394, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x139c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1398, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1594, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1598, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x159c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x15a0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x15a4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x0f54, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f58, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f5c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x19bc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f9c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0fa0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x12cc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x12e8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x130c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1360, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1364, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1368, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x136c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1370, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1374, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1378, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x137c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x133c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1340, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1344, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1348, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x134c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1350, 0x00000002);
-	nv_mthd(priv, 0x9097, 0x1358, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x12e4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x131c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1320, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1324, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1328, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x19c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1140, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x19c4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x19c8, 0x00001500);
-	nv_mthd(priv, 0x9097, 0x135c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x19e0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19e4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19e8, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19ec, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19f0, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19f4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19f8, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19fc, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x19cc, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x15b8, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a00, 0x00001111);
-	nv_mthd(priv, 0x9097, 0x1a04, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a08, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a0c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a10, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a14, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a18, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1a1c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d6c, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x0d70, 0xffff0000);
-	nv_mthd(priv, 0x9097, 0x10f8, 0x00001010);
-	nv_mthd(priv, 0x9097, 0x0d80, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d84, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d88, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d8c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0d90, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0da0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1508, 0x80000000);
-	nv_mthd(priv, 0x9097, 0x150c, 0x40000000);
-	nv_mthd(priv, 0x9097, 0x1668, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0318, 0x00000008);
-	nv_mthd(priv, 0x9097, 0x031c, 0x00000008);
-	nv_mthd(priv, 0x9097, 0x0d9c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x07dc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x074c, 0x00000055);
-	nv_mthd(priv, 0x9097, 0x1420, 0x00000003);
-	nv_mthd(priv, 0x9097, 0x17bc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x17c0, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x17c4, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1008, 0x00000008);
-	nv_mthd(priv, 0x9097, 0x100c, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x1010, 0x0000012c);
-	nv_mthd(priv, 0x9097, 0x0d60, 0x00000040);
-	nv_mthd(priv, 0x9097, 0x075c, 0x00000003);
-	nv_mthd(priv, 0x9097, 0x1018, 0x00000020);
-	nv_mthd(priv, 0x9097, 0x101c, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1020, 0x00000020);
-	nv_mthd(priv, 0x9097, 0x1024, 0x00000001);
-	nv_mthd(priv, 0x9097, 0x1444, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x1448, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x144c, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0360, 0x20164010);
-	nv_mthd(priv, 0x9097, 0x0364, 0x00000020);
-	nv_mthd(priv, 0x9097, 0x0368, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0de4, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0204, 0x00000006);
-	nv_mthd(priv, 0x9097, 0x0208, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x02cc, 0x003fffff);
-	nv_mthd(priv, 0x9097, 0x02d0, 0x00000c48);
-	nv_mthd(priv, 0x9097, 0x1220, 0x00000005);
-	nv_mthd(priv, 0x9097, 0x0fdc, 0x00000000);
-	nv_mthd(priv, 0x9097, 0x0f98, 0x00300008);
-	nv_mthd(priv, 0x9097, 0x1284, 0x04000080);
-	nv_mthd(priv, 0x9097, 0x1450, 0x00300008);
-	nv_mthd(priv, 0x9097, 0x1454, 0x04000080);
-	nv_mthd(priv, 0x9097, 0x0214, 0x00000000);
-	/* in trace, right after 0x90c0, not here */
-	nv_mthd(priv, 0x9097, 0x3410, 0x80002006);
 }
 
-static void
-nvc0_grctx_generate_9197(struct nvc0_graph_priv *priv)
+void
+nvc0_grctx_generate_r406028(struct nvc0_graph_priv *priv)
 {
-	u32 fermi = nvc0_graph_class(priv);
-	u32 mthd;
-
-	if (fermi == 0x9197) {
-		for (mthd = 0x3400; mthd <= 0x35fc; mthd += 4)
-			nv_mthd(priv, 0x9197, mthd, 0x00000000);
+	u32 tmp[GPC_MAX / 8] = {}, i = 0;
+	for (i = 0; i < priv->gpc_nr; i++)
+		tmp[i / 8] |= priv->tpc_nr[i] << ((i % 8) * 4);
+	for (i = 0; i < 4; i++) {
+		nv_wr32(priv, 0x406028 + (i * 4), tmp[i]);
+		nv_wr32(priv, 0x405870 + (i * 4), tmp[i]);
 	}
-	nv_mthd(priv, 0x9197, 0x02e4, 0x0000b001);
 }
 
-static void
-nvc0_grctx_generate_9297(struct nvc0_graph_priv *priv)
+void
+nvc0_grctx_generate_r4060a8(struct nvc0_graph_priv *priv)
 {
-	u32 fermi = nvc0_graph_class(priv);
-	u32 mthd;
-
-	if (fermi == 0x9297) {
-		for (mthd = 0x3400; mthd <= 0x35fc; mthd += 4)
-			nv_mthd(priv, 0x9297, mthd, 0x00000000);
+	u8  tpcnr[GPC_MAX], data[TPC_MAX];
+	int gpc, tpc, i;
+
+	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+	memset(data, 0x1f, sizeof(data));
+
+	gpc = -1;
+	for (tpc = 0; tpc < priv->tpc_total; tpc++) {
+		do {
+			gpc = (gpc + 1) % priv->gpc_nr;
+		} while (!tpcnr[gpc]);
+		tpcnr[gpc]--;
+		data[tpc] = gpc;
 	}
-	nv_mthd(priv, 0x9297, 0x036c, 0x00000000);
-	nv_mthd(priv, 0x9297, 0x0370, 0x00000000);
-	nv_mthd(priv, 0x9297, 0x07a4, 0x00000000);
-	nv_mthd(priv, 0x9297, 0x07a8, 0x00000000);
-	nv_mthd(priv, 0x9297, 0x0374, 0x00000000);
-	nv_mthd(priv, 0x9297, 0x0378, 0x00000020);
-}
 
-static void
-nvc0_grctx_generate_902d(struct nvc0_graph_priv *priv)
-{
-	nv_mthd(priv, 0x902d, 0x0200, 0x000000cf);
-	nv_mthd(priv, 0x902d, 0x0204, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0208, 0x00000020);
-	nv_mthd(priv, 0x902d, 0x020c, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0210, 0x00000000);
-	nv_mthd(priv, 0x902d, 0x0214, 0x00000080);
-	nv_mthd(priv, 0x902d, 0x0218, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x021c, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x0220, 0x00000000);
-	nv_mthd(priv, 0x902d, 0x0224, 0x00000000);
-	nv_mthd(priv, 0x902d, 0x0230, 0x000000cf);
-	nv_mthd(priv, 0x902d, 0x0234, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0238, 0x00000020);
-	nv_mthd(priv, 0x902d, 0x023c, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0244, 0x00000080);
-	nv_mthd(priv, 0x902d, 0x0248, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x024c, 0x00000100);
-}
-
-static void
-nvc0_grctx_generate_9039(struct nvc0_graph_priv *priv)
-{
-	nv_mthd(priv, 0x9039, 0x030c, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x0310, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x0314, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x0320, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x0238, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x023c, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x0318, 0x00000000);
-	nv_mthd(priv, 0x9039, 0x031c, 0x00000000);
+	for (i = 0; i < 4; i++)
+		nv_wr32(priv, 0x4060a8 + (i * 4), ((u32 *)data)[i]);
 }
 
-static void
-nvc0_grctx_generate_90c0(struct nvc0_graph_priv *priv)
+void
+nvc0_grctx_generate_r418bb8(struct nvc0_graph_priv *priv)
 {
-	int i;
-
-	for (i = 0; nv_device(priv)->chipset >= 0xd0 && i < 4; i++) {
-		nv_mthd(priv, 0x90c0, 0x2700 + (i * 0x40), 0x00000000);
-		nv_mthd(priv, 0x90c0, 0x2720 + (i * 0x40), 0x00000000);
-		nv_mthd(priv, 0x90c0, 0x2704 + (i * 0x40), 0x00000000);
-		nv_mthd(priv, 0x90c0, 0x2724 + (i * 0x40), 0x00000000);
-		nv_mthd(priv, 0x90c0, 0x2708 + (i * 0x40), 0x00000000);
-		nv_mthd(priv, 0x90c0, 0x2728 + (i * 0x40), 0x00000000);
+	u32 data[6] = {}, data2[2] = {};
+	u8  tpcnr[GPC_MAX];
+	u8  shift, ntpcv;
+	int gpc, tpc, i;
+
+	/* calculate first set of magics */
+	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+
+	gpc = -1;
+	for (tpc = 0; tpc < priv->tpc_total; tpc++) {
+		do {
+			gpc = (gpc + 1) % priv->gpc_nr;
+		} while (!tpcnr[gpc]);
+		tpcnr[gpc]--;
+
+		data[tpc / 6] |= gpc << ((tpc % 6) * 5);
 	}
-	nv_mthd(priv, 0x90c0, 0x270c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x272c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x274c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x276c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x278c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x27ac, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x27cc, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x27ec, 0x00000000);
-	for (i = 0; nv_device(priv)->chipset >= 0xd0 && i < 4; i++) {
-		nv_mthd(priv, 0x90c0, 0x2710 + (i * 0x40), 0x00014000);
-		nv_mthd(priv, 0x90c0, 0x2730 + (i * 0x40), 0x00014000);
-		nv_mthd(priv, 0x90c0, 0x2714 + (i * 0x40), 0x00000040);
-		nv_mthd(priv, 0x90c0, 0x2734 + (i * 0x40), 0x00000040);
-	}
-	nv_mthd(priv, 0x90c0, 0x030c, 0x00000001);
-	nv_mthd(priv, 0x90c0, 0x1944, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0758, 0x00000100);
-	nv_mthd(priv, 0x90c0, 0x02c4, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0790, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0794, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0798, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x079c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x07a0, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x077c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0204, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0208, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x020c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0214, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x024c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x0d94, 0x00000001);
-	nv_mthd(priv, 0x90c0, 0x1608, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x160c, 0x00000000);
-	nv_mthd(priv, 0x90c0, 0x1664, 0x00000000);
-}
 
-static void
-nvc0_grctx_generate_dispatch(struct nvc0_graph_priv *priv)
-{
-	int i;
+	for (; tpc < 32; tpc++)
+		data[tpc / 6] |= 7 << ((tpc % 6) * 5);
 
-	nv_wr32(priv, 0x404004, 0x00000000);
-	nv_wr32(priv, 0x404008, 0x00000000);
-	nv_wr32(priv, 0x40400c, 0x00000000);
-	nv_wr32(priv, 0x404010, 0x00000000);
-	nv_wr32(priv, 0x404014, 0x00000000);
-	nv_wr32(priv, 0x404018, 0x00000000);
-	nv_wr32(priv, 0x40401c, 0x00000000);
-	nv_wr32(priv, 0x404020, 0x00000000);
-	nv_wr32(priv, 0x404024, 0x00000000);
-	nv_wr32(priv, 0x404028, 0x00000000);
-	nv_wr32(priv, 0x40402c, 0x00000000);
-	nv_wr32(priv, 0x404044, 0x00000000);
-	nv_wr32(priv, 0x404094, 0x00000000);
-	nv_wr32(priv, 0x404098, 0x00000000);
-	nv_wr32(priv, 0x40409c, 0x00000000);
-	nv_wr32(priv, 0x4040a0, 0x00000000);
-	nv_wr32(priv, 0x4040a4, 0x00000000);
-	nv_wr32(priv, 0x4040a8, 0x00000000);
-	nv_wr32(priv, 0x4040ac, 0x00000000);
-	nv_wr32(priv, 0x4040b0, 0x00000000);
-	nv_wr32(priv, 0x4040b4, 0x00000000);
-	nv_wr32(priv, 0x4040b8, 0x00000000);
-	nv_wr32(priv, 0x4040bc, 0x00000000);
-	nv_wr32(priv, 0x4040c0, 0x00000000);
-	nv_wr32(priv, 0x4040c4, 0x00000000);
-	nv_wr32(priv, 0x4040c8, 0xf0000087);
-	nv_wr32(priv, 0x4040d4, 0x00000000);
-	nv_wr32(priv, 0x4040d8, 0x00000000);
-	nv_wr32(priv, 0x4040dc, 0x00000000);
-	nv_wr32(priv, 0x4040e0, 0x00000000);
-	nv_wr32(priv, 0x4040e4, 0x00000000);
-	nv_wr32(priv, 0x4040e8, 0x00001000);
-	nv_wr32(priv, 0x4040f8, 0x00000000);
-	nv_wr32(priv, 0x404130, 0x00000000);
-	nv_wr32(priv, 0x404134, 0x00000000);
-	nv_wr32(priv, 0x404138, 0x20000040);
-	nv_wr32(priv, 0x404150, 0x0000002e);
-	nv_wr32(priv, 0x404154, 0x00000400);
-	nv_wr32(priv, 0x404158, 0x00000200);
-	nv_wr32(priv, 0x404164, 0x00000055);
-	nv_wr32(priv, 0x404168, 0x00000000);
-	nv_wr32(priv, 0x404174, 0x00000000);
-	nv_wr32(priv, 0x404178, 0x00000000);
-	nv_wr32(priv, 0x40417c, 0x00000000);
-	for (i = 0; i < 8; i++)
-		nv_wr32(priv, 0x404200 + (i * 4), 0x00000000); /* subc */
-}
-
-static void
-nvc0_grctx_generate_macro(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404404, 0x00000000);
-	nv_wr32(priv, 0x404408, 0x00000000);
-	nv_wr32(priv, 0x40440c, 0x00000000);
-	nv_wr32(priv, 0x404410, 0x00000000);
-	nv_wr32(priv, 0x404414, 0x00000000);
-	nv_wr32(priv, 0x404418, 0x00000000);
-	nv_wr32(priv, 0x40441c, 0x00000000);
-	nv_wr32(priv, 0x404420, 0x00000000);
-	nv_wr32(priv, 0x404424, 0x00000000);
-	nv_wr32(priv, 0x404428, 0x00000000);
-	nv_wr32(priv, 0x40442c, 0x00000000);
-	nv_wr32(priv, 0x404430, 0x00000000);
-	nv_wr32(priv, 0x404434, 0x00000000);
-	nv_wr32(priv, 0x404438, 0x00000000);
-	nv_wr32(priv, 0x404460, 0x00000000);
-	nv_wr32(priv, 0x404464, 0x00000000);
-	nv_wr32(priv, 0x404468, 0x00ffffff);
-	nv_wr32(priv, 0x40446c, 0x00000000);
-	nv_wr32(priv, 0x404480, 0x00000001);
-	nv_wr32(priv, 0x404498, 0x00000001);
-}
-
-static void
-nvc0_grctx_generate_m2mf(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404604, 0x00000015);
-	nv_wr32(priv, 0x404608, 0x00000000);
-	nv_wr32(priv, 0x40460c, 0x00002e00);
-	nv_wr32(priv, 0x404610, 0x00000100);
-	nv_wr32(priv, 0x404618, 0x00000000);
-	nv_wr32(priv, 0x40461c, 0x00000000);
-	nv_wr32(priv, 0x404620, 0x00000000);
-	nv_wr32(priv, 0x404624, 0x00000000);
-	nv_wr32(priv, 0x404628, 0x00000000);
-	nv_wr32(priv, 0x40462c, 0x00000000);
-	nv_wr32(priv, 0x404630, 0x00000000);
-	nv_wr32(priv, 0x404634, 0x00000000);
-	nv_wr32(priv, 0x404638, 0x00000004);
-	nv_wr32(priv, 0x40463c, 0x00000000);
-	nv_wr32(priv, 0x404640, 0x00000000);
-	nv_wr32(priv, 0x404644, 0x00000000);
-	nv_wr32(priv, 0x404648, 0x00000000);
-	nv_wr32(priv, 0x40464c, 0x00000000);
-	nv_wr32(priv, 0x404650, 0x00000000);
-	nv_wr32(priv, 0x404654, 0x00000000);
-	nv_wr32(priv, 0x404658, 0x00000000);
-	nv_wr32(priv, 0x40465c, 0x007f0100);
-	nv_wr32(priv, 0x404660, 0x00000000);
-	nv_wr32(priv, 0x404664, 0x00000000);
-	nv_wr32(priv, 0x404668, 0x00000000);
-	nv_wr32(priv, 0x40466c, 0x00000000);
-	nv_wr32(priv, 0x404670, 0x00000000);
-	nv_wr32(priv, 0x404674, 0x00000000);
-	nv_wr32(priv, 0x404678, 0x00000000);
-	nv_wr32(priv, 0x40467c, 0x00000002);
-	nv_wr32(priv, 0x404680, 0x00000000);
-	nv_wr32(priv, 0x404684, 0x00000000);
-	nv_wr32(priv, 0x404688, 0x00000000);
-	nv_wr32(priv, 0x40468c, 0x00000000);
-	nv_wr32(priv, 0x404690, 0x00000000);
-	nv_wr32(priv, 0x404694, 0x00000000);
-	nv_wr32(priv, 0x404698, 0x00000000);
-	nv_wr32(priv, 0x40469c, 0x00000000);
-	nv_wr32(priv, 0x4046a0, 0x007f0080);
-	nv_wr32(priv, 0x4046a4, 0x00000000);
-	nv_wr32(priv, 0x4046a8, 0x00000000);
-	nv_wr32(priv, 0x4046ac, 0x00000000);
-	nv_wr32(priv, 0x4046b0, 0x00000000);
-	nv_wr32(priv, 0x4046b4, 0x00000000);
-	nv_wr32(priv, 0x4046b8, 0x00000000);
-	nv_wr32(priv, 0x4046bc, 0x00000000);
-	nv_wr32(priv, 0x4046c0, 0x00000000);
-	nv_wr32(priv, 0x4046c4, 0x00000000);
-	nv_wr32(priv, 0x4046c8, 0x00000000);
-	nv_wr32(priv, 0x4046cc, 0x00000000);
-	nv_wr32(priv, 0x4046d0, 0x00000000);
-	nv_wr32(priv, 0x4046d4, 0x00000000);
-	nv_wr32(priv, 0x4046d8, 0x00000000);
-	nv_wr32(priv, 0x4046dc, 0x00000000);
-	nv_wr32(priv, 0x4046e0, 0x00000000);
-	nv_wr32(priv, 0x4046e4, 0x00000000);
-	nv_wr32(priv, 0x4046e8, 0x00000000);
-	nv_wr32(priv, 0x4046f0, 0x00000000);
-	nv_wr32(priv, 0x4046f4, 0x00000000);
-}
+	/* and the second... */
+	shift = 0;
+	ntpcv = priv->tpc_total;
+	while (!(ntpcv & (1 << 4))) {
+		ntpcv <<= 1;
+		shift++;
+	}
 
-static void
-nvc0_grctx_generate_unk47xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404700, 0x00000000);
-	nv_wr32(priv, 0x404704, 0x00000000);
-	nv_wr32(priv, 0x404708, 0x00000000);
-	nv_wr32(priv, 0x40470c, 0x00000000);
-	nv_wr32(priv, 0x404710, 0x00000000);
-	nv_wr32(priv, 0x404714, 0x00000000);
-	nv_wr32(priv, 0x404718, 0x00000000);
-	nv_wr32(priv, 0x40471c, 0x00000000);
-	nv_wr32(priv, 0x404720, 0x00000000);
-	nv_wr32(priv, 0x404724, 0x00000000);
-	nv_wr32(priv, 0x404728, 0x00000000);
-	nv_wr32(priv, 0x40472c, 0x00000000);
-	nv_wr32(priv, 0x404730, 0x00000000);
-	nv_wr32(priv, 0x404734, 0x00000100);
-	nv_wr32(priv, 0x404738, 0x00000000);
-	nv_wr32(priv, 0x40473c, 0x00000000);
-	nv_wr32(priv, 0x404740, 0x00000000);
-	nv_wr32(priv, 0x404744, 0x00000000);
-	nv_wr32(priv, 0x404748, 0x00000000);
-	nv_wr32(priv, 0x40474c, 0x00000000);
-	nv_wr32(priv, 0x404750, 0x00000000);
-	nv_wr32(priv, 0x404754, 0x00000000);
-}
+	data2[0]  = (ntpcv << 16);
+	data2[0] |= (shift << 21);
+	data2[0] |= (((1 << (0 + 5)) % ntpcv) << 24);
+	for (i = 1; i < 7; i++)
+		data2[1] |= ((1 << (i + 5)) % ntpcv) << ((i - 1) * 5);
 
-static void
-nvc0_grctx_generate_shaders(struct nvc0_graph_priv *priv)
-{
+	/* GPC_BROADCAST */
+	nv_wr32(priv, 0x418bb8, (priv->tpc_total << 8) |
+				 priv->magic_not_rop_nr);
+	for (i = 0; i < 6; i++)
+		nv_wr32(priv, 0x418b08 + (i * 4), data[i]);
 
-	if (nv_device(priv)->chipset >= 0xd0) {
-		nv_wr32(priv, 0x405800, 0x0f8000bf);
-		nv_wr32(priv, 0x405830, 0x02180218);
-		nv_wr32(priv, 0x405834, 0x08000000);
-	} else
-	if (nv_device(priv)->chipset == 0xc1) {
-		nv_wr32(priv, 0x405800, 0x0f8000bf);
-		nv_wr32(priv, 0x405830, 0x02180218);
-		nv_wr32(priv, 0x405834, 0x00000000);
-	} else {
-		nv_wr32(priv, 0x405800, 0x078000bf);
-		nv_wr32(priv, 0x405830, 0x02180000);
-		nv_wr32(priv, 0x405834, 0x00000000);
-	}
-	nv_wr32(priv, 0x405838, 0x00000000);
-	nv_wr32(priv, 0x405854, 0x00000000);
-	nv_wr32(priv, 0x405870, 0x00000001);
-	nv_wr32(priv, 0x405874, 0x00000001);
-	nv_wr32(priv, 0x405878, 0x00000001);
-	nv_wr32(priv, 0x40587c, 0x00000001);
-	nv_wr32(priv, 0x405a00, 0x00000000);
-	nv_wr32(priv, 0x405a04, 0x00000000);
-	nv_wr32(priv, 0x405a18, 0x00000000);
+	/* GPC_BROADCAST.TP_BROADCAST */
+	nv_wr32(priv, 0x419bd0, (priv->tpc_total << 8) |
+				 priv->magic_not_rop_nr | data2[0]);
+	nv_wr32(priv, 0x419be4, data2[1]);
+	for (i = 0; i < 6; i++)
+		nv_wr32(priv, 0x419b00 + (i * 4), data[i]);
+
+	/* UNK78xx */
+	nv_wr32(priv, 0x4078bc, (priv->tpc_total << 8) |
+				 priv->magic_not_rop_nr);
+	for (i = 0; i < 6; i++)
+		nv_wr32(priv, 0x40780c + (i * 4), data[i]);
 }
 
-static void
-nvc0_grctx_generate_unk60xx(struct nvc0_graph_priv *priv)
+void
+nvc0_grctx_generate_r406800(struct nvc0_graph_priv *priv)
 {
-	nv_wr32(priv, 0x406020, 0x000103c1);
-	nv_wr32(priv, 0x406028, 0x00000001);
-	nv_wr32(priv, 0x40602c, 0x00000001);
-	nv_wr32(priv, 0x406030, 0x00000001);
-	nv_wr32(priv, 0x406034, 0x00000001);
-}
+	u64 tpc_mask = 0, tpc_set = 0;
+	u8  tpcnr[GPC_MAX];
+	int gpc, tpc;
+	int i, a, b;
+
+	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++)
+		tpc_mask |= ((1ULL << priv->tpc_nr[gpc]) - 1) << (gpc * 8);
+
+	for (i = 0, gpc = -1, b = -1; i < 32; i++) {
+		a = (i * (priv->tpc_total - 1)) / 32;
+		if (a != b) {
+			b = a;
+			do {
+				gpc = (gpc + 1) % priv->gpc_nr;
+			} while (!tpcnr[gpc]);
+			tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
 
-static void
-nvc0_grctx_generate_unk64xx(struct nvc0_graph_priv *priv)
-{
+			tpc_set |= 1 << ((gpc * 8) + tpc);
+		}
 
-	nv_wr32(priv, 0x4064a8, 0x00000000);
-	nv_wr32(priv, 0x4064ac, 0x00003fff);
-	nv_wr32(priv, 0x4064b4, 0x00000000);
-	nv_wr32(priv, 0x4064b8, 0x00000000);
-	if (nv_device(priv)->chipset >= 0xd0)
-		nv_wr32(priv, 0x4064bc, 0x00000000);
-	if (nv_device(priv)->chipset == 0xc1 ||
-	    nv_device(priv)->chipset >= 0xd0) {
-		nv_wr32(priv, 0x4064c0, 0x80140078);
-		nv_wr32(priv, 0x4064c4, 0x0086ffff);
+		nv_wr32(priv, 0x406800 + (i * 0x20), lower_32_bits(tpc_set));
+		nv_wr32(priv, 0x406c00 + (i * 0x20), lower_32_bits(tpc_set ^ tpc_mask));
+		if (priv->gpc_nr > 4) {
+			nv_wr32(priv, 0x406804 + (i * 0x20), upper_32_bits(tpc_set));
+			nv_wr32(priv, 0x406c04 + (i * 0x20), upper_32_bits(tpc_set ^ tpc_mask));
+		}
 	}
 }
 
-static void
-nvc0_grctx_generate_tpbus(struct nvc0_graph_priv *priv)
+void
+nvc0_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
 {
-	nv_wr32(priv, 0x407804, 0x00000023);
-	nv_wr32(priv, 0x40780c, 0x0a418820);
-	nv_wr32(priv, 0x407810, 0x062080e6);
-	nv_wr32(priv, 0x407814, 0x020398a4);
-	nv_wr32(priv, 0x407818, 0x0e629062);
-	nv_wr32(priv, 0x40781c, 0x0a418820);
-	nv_wr32(priv, 0x407820, 0x000000e6);
-	nv_wr32(priv, 0x4078bc, 0x00000103);
-}
+	struct nvc0_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass;
+	int i;
 
-static void
-nvc0_grctx_generate_ccache(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x408000, 0x00000000);
-	nv_wr32(priv, 0x408004, 0x00000000);
-	nv_wr32(priv, 0x408008, 0x00000018);
-	nv_wr32(priv, 0x40800c, 0x00000000);
-	nv_wr32(priv, 0x408010, 0x00000000);
-	nv_wr32(priv, 0x408014, 0x00000069);
-	nv_wr32(priv, 0x408018, 0xe100e100);
-	nv_wr32(priv, 0x408064, 0x00000000);
-}
+	nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
 
-static void
-nvc0_grctx_generate_rop(struct nvc0_graph_priv *priv)
-{
-	int chipset = nv_device(priv)->chipset;
-
-	/* ROPC_BROADCAST */
-	nv_wr32(priv, 0x408800, 0x02802a3c);
-	nv_wr32(priv, 0x408804, 0x00000040);
-	if (chipset >= 0xd0) {
-		nv_wr32(priv, 0x408808, 0x1043e005);
-		nv_wr32(priv, 0x408900, 0x3080b801);
-		nv_wr32(priv, 0x408904, 0x1043e005);
-		nv_wr32(priv, 0x408908, 0x00c8102f);
-	} else
-	if (chipset == 0xc1) {
-		nv_wr32(priv, 0x408808, 0x1003e005);
-		nv_wr32(priv, 0x408900, 0x3080b801);
-		nv_wr32(priv, 0x408904, 0x62000001);
-		nv_wr32(priv, 0x408908, 0x00c80929);
-	} else {
-		nv_wr32(priv, 0x408808, 0x0003e00d);
-		nv_wr32(priv, 0x408900, 0x3080b801);
-		nv_wr32(priv, 0x408904, 0x02000001);
-		nv_wr32(priv, 0x408908, 0x00c80929);
-	}
-	nv_wr32(priv, 0x40890c, 0x00000000);
-	nv_wr32(priv, 0x408980, 0x0000011d);
-}
+	for (i = 0; oclass->hub[i]; i++)
+		nvc0_graph_mmio(priv, oclass->hub[i]);
+	for (i = 0; oclass->gpc[i]; i++)
+		nvc0_graph_mmio(priv, oclass->gpc[i]);
 
-static void
-nvc0_grctx_generate_gpc(struct nvc0_graph_priv *priv)
-{
-	int chipset = nv_device(priv)->chipset;
-	int i;
+	nv_wr32(priv, 0x404154, 0x00000000);
 
-	/* GPC_BROADCAST */
-	nv_wr32(priv, 0x418380, 0x00000016);
-	nv_wr32(priv, 0x418400, 0x38004e00);
-	nv_wr32(priv, 0x418404, 0x71e0ffff);
-	nv_wr32(priv, 0x418408, 0x00000000);
-	nv_wr32(priv, 0x41840c, 0x00001008);
-	nv_wr32(priv, 0x418410, 0x0fff0fff);
-	nv_wr32(priv, 0x418414, chipset < 0xd0 ? 0x00200fff : 0x02200fff);
-	nv_wr32(priv, 0x418450, 0x00000000);
-	nv_wr32(priv, 0x418454, 0x00000000);
-	nv_wr32(priv, 0x418458, 0x00000000);
-	nv_wr32(priv, 0x41845c, 0x00000000);
-	nv_wr32(priv, 0x418460, 0x00000000);
-	nv_wr32(priv, 0x418464, 0x00000000);
-	nv_wr32(priv, 0x418468, 0x00000001);
-	nv_wr32(priv, 0x41846c, 0x00000000);
-	nv_wr32(priv, 0x418470, 0x00000000);
-	nv_wr32(priv, 0x418600, 0x0000001f);
-	nv_wr32(priv, 0x418684, 0x0000000f);
-	nv_wr32(priv, 0x418700, 0x00000002);
-	nv_wr32(priv, 0x418704, 0x00000080);
-	nv_wr32(priv, 0x418708, 0x00000000);
-	nv_wr32(priv, 0x41870c, chipset < 0xd0 ? 0x07c80000 : 0x00000000);
-	nv_wr32(priv, 0x418710, 0x00000000);
-	nv_wr32(priv, 0x418800, chipset < 0xd0 ? 0x0006860a : 0x7006860a);
-	nv_wr32(priv, 0x418808, 0x00000000);
-	nv_wr32(priv, 0x41880c, 0x00000000);
-	nv_wr32(priv, 0x418810, 0x00000000);
-	nv_wr32(priv, 0x418828, 0x00008442);
-	if (chipset == 0xc1 || chipset >= 0xd0)
-		nv_wr32(priv, 0x418830, 0x10000001);
-	else
-		nv_wr32(priv, 0x418830, 0x00000001);
-	nv_wr32(priv, 0x4188d8, 0x00000008);
-	nv_wr32(priv, 0x4188e0, 0x01000000);
-	nv_wr32(priv, 0x4188e8, 0x00000000);
-	nv_wr32(priv, 0x4188ec, 0x00000000);
-	nv_wr32(priv, 0x4188f0, 0x00000000);
-	nv_wr32(priv, 0x4188f4, 0x00000000);
-	nv_wr32(priv, 0x4188f8, 0x00000000);
-	if (chipset >= 0xd0)
-		nv_wr32(priv, 0x4188fc, 0x20100008);
-	else if (chipset == 0xc1)
-		nv_wr32(priv, 0x4188fc, 0x00100018);
-	else
-		nv_wr32(priv, 0x4188fc, 0x00100000);
-	nv_wr32(priv, 0x41891c, 0x00ff00ff);
-	nv_wr32(priv, 0x418924, 0x00000000);
-	nv_wr32(priv, 0x418928, 0x00ffff00);
-	nv_wr32(priv, 0x41892c, 0x0000ff00);
-	for (i = 0; i < 8; i++) {
-		nv_wr32(priv, 0x418a00 + (i * 0x20), 0x00000000);
-		nv_wr32(priv, 0x418a04 + (i * 0x20), 0x00000000);
-		nv_wr32(priv, 0x418a08 + (i * 0x20), 0x00000000);
-		nv_wr32(priv, 0x418a0c + (i * 0x20), 0x00010000);
-		nv_wr32(priv, 0x418a10 + (i * 0x20), 0x00000000);
-		nv_wr32(priv, 0x418a14 + (i * 0x20), 0x00000000);
-		nv_wr32(priv, 0x418a18 + (i * 0x20), 0x00000000);
-	}
-	nv_wr32(priv, 0x418b00, chipset < 0xd0 ? 0x00000000 : 0x00000006);
-	nv_wr32(priv, 0x418b08, 0x0a418820);
-	nv_wr32(priv, 0x418b0c, 0x062080e6);
-	nv_wr32(priv, 0x418b10, 0x020398a4);
-	nv_wr32(priv, 0x418b14, 0x0e629062);
-	nv_wr32(priv, 0x418b18, 0x0a418820);
-	nv_wr32(priv, 0x418b1c, 0x000000e6);
-	nv_wr32(priv, 0x418bb8, 0x00000103);
-	nv_wr32(priv, 0x418c08, 0x00000001);
-	nv_wr32(priv, 0x418c10, 0x00000000);
-	nv_wr32(priv, 0x418c14, 0x00000000);
-	nv_wr32(priv, 0x418c18, 0x00000000);
-	nv_wr32(priv, 0x418c1c, 0x00000000);
-	nv_wr32(priv, 0x418c20, 0x00000000);
-	nv_wr32(priv, 0x418c24, 0x00000000);
-	nv_wr32(priv, 0x418c28, 0x00000000);
-	nv_wr32(priv, 0x418c2c, 0x00000000);
-	if (chipset == 0xc1 || chipset >= 0xd0)
-		nv_wr32(priv, 0x418c6c, 0x00000001);
-	nv_wr32(priv, 0x418c80, 0x20200004);
-	nv_wr32(priv, 0x418c8c, 0x00000001);
-	nv_wr32(priv, 0x419000, 0x00000780);
-	nv_wr32(priv, 0x419004, 0x00000000);
-	nv_wr32(priv, 0x419008, 0x00000000);
-	nv_wr32(priv, 0x419014, 0x00000004);
-}
+	oclass->mods(priv, info);
+	oclass->unkn(priv);
 
-static void
-nvc0_grctx_generate_tp(struct nvc0_graph_priv *priv)
-{
-	int chipset = nv_device(priv)->chipset;
+	nvc0_grctx_generate_tpcid(priv);
+	nvc0_grctx_generate_r406028(priv);
+	nvc0_grctx_generate_r4060a8(priv);
+	nvc0_grctx_generate_r418bb8(priv);
+	nvc0_grctx_generate_r406800(priv);
 
-	/* GPC_BROADCAST.TP_BROADCAST */
-	nv_wr32(priv, 0x419818, 0x00000000);
-	nv_wr32(priv, 0x41983c, 0x00038bc7);
-	nv_wr32(priv, 0x419848, 0x00000000);
-	if (chipset == 0xc1 || chipset >= 0xd0)
-		nv_wr32(priv, 0x419864, 0x00000129);
-	else
-		nv_wr32(priv, 0x419864, 0x0000012a);
-	nv_wr32(priv, 0x419888, 0x00000000);
-	nv_wr32(priv, 0x419a00, 0x000001f0);
-	nv_wr32(priv, 0x419a04, 0x00000001);
-	nv_wr32(priv, 0x419a08, 0x00000023);
-	nv_wr32(priv, 0x419a0c, 0x00020000);
-	nv_wr32(priv, 0x419a10, 0x00000000);
-	nv_wr32(priv, 0x419a14, 0x00000200);
-	nv_wr32(priv, 0x419a1c, 0x00000000);
-	nv_wr32(priv, 0x419a20, 0x00000800);
-	if (chipset >= 0xd0)
-		nv_wr32(priv, 0x00419ac4, 0x0017f440);
-	else if (chipset != 0xc0 && chipset != 0xc8)
-		nv_wr32(priv, 0x00419ac4, 0x0007f440);
-	nv_wr32(priv, 0x419b00, 0x0a418820);
-	nv_wr32(priv, 0x419b04, 0x062080e6);
-	nv_wr32(priv, 0x419b08, 0x020398a4);
-	nv_wr32(priv, 0x419b0c, 0x0e629062);
-	nv_wr32(priv, 0x419b10, 0x0a418820);
-	nv_wr32(priv, 0x419b14, 0x000000e6);
-	nv_wr32(priv, 0x419bd0, 0x00900103);
-	if (chipset == 0xc1 || chipset >= 0xd0)
-		nv_wr32(priv, 0x419be0, 0x00400001);
-	else
-		nv_wr32(priv, 0x419be0, 0x00000001);
-	nv_wr32(priv, 0x419be4, 0x00000000);
-	nv_wr32(priv, 0x419c00, chipset < 0xd0 ? 0x00000002 : 0x0000000a);
-	nv_wr32(priv, 0x419c04, 0x00000006);
-	nv_wr32(priv, 0x419c08, 0x00000002);
-	nv_wr32(priv, 0x419c20, 0x00000000);
-	if (nv_device(priv)->chipset >= 0xd0) {
-		nv_wr32(priv, 0x419c24, 0x00084210);
-		nv_wr32(priv, 0x419c28, 0x3cf3cf3c);
-		nv_wr32(priv, 0x419cb0, 0x00020048);
-	} else
-	if (chipset == 0xce || chipset == 0xcf) {
-		nv_wr32(priv, 0x419cb0, 0x00020048);
-	} else {
-		nv_wr32(priv, 0x419cb0, 0x00060048);
-	}
-	nv_wr32(priv, 0x419ce8, 0x00000000);
-	nv_wr32(priv, 0x419cf4, 0x00000183);
-	if (chipset == 0xc1 || chipset >= 0xd0)
-		nv_wr32(priv, 0x419d20, 0x12180000);
-	else
-		nv_wr32(priv, 0x419d20, 0x02180000);
-	nv_wr32(priv, 0x419d24, 0x00001fff);
-	if (chipset == 0xc1 || chipset >= 0xd0)
-		nv_wr32(priv, 0x419d44, 0x02180218);
-	nv_wr32(priv, 0x419e04, 0x00000000);
-	nv_wr32(priv, 0x419e08, 0x00000000);
-	nv_wr32(priv, 0x419e0c, 0x00000000);
-	nv_wr32(priv, 0x419e10, 0x00000002);
-	nv_wr32(priv, 0x419e44, 0x001beff2);
-	nv_wr32(priv, 0x419e48, 0x00000000);
-	nv_wr32(priv, 0x419e4c, 0x0000000f);
-	nv_wr32(priv, 0x419e50, 0x00000000);
-	nv_wr32(priv, 0x419e54, 0x00000000);
-	nv_wr32(priv, 0x419e58, 0x00000000);
-	nv_wr32(priv, 0x419e5c, 0x00000000);
-	nv_wr32(priv, 0x419e60, 0x00000000);
-	nv_wr32(priv, 0x419e64, 0x00000000);
-	nv_wr32(priv, 0x419e68, 0x00000000);
-	nv_wr32(priv, 0x419e6c, 0x00000000);
-	nv_wr32(priv, 0x419e70, 0x00000000);
-	nv_wr32(priv, 0x419e74, 0x00000000);
-	nv_wr32(priv, 0x419e78, 0x00000000);
-	nv_wr32(priv, 0x419e7c, 0x00000000);
-	nv_wr32(priv, 0x419e80, 0x00000000);
-	nv_wr32(priv, 0x419e84, 0x00000000);
-	nv_wr32(priv, 0x419e88, 0x00000000);
-	nv_wr32(priv, 0x419e8c, 0x00000000);
-	nv_wr32(priv, 0x419e90, 0x00000000);
-	nv_wr32(priv, 0x419e98, 0x00000000);
-	if (chipset != 0xc0 && chipset != 0xc8)
-		nv_wr32(priv, 0x419ee0, 0x00011110);
-	nv_wr32(priv, 0x419f50, 0x00000000);
-	nv_wr32(priv, 0x419f54, 0x00000000);
-	if (chipset != 0xc0 && chipset != 0xc8)
-		nv_wr32(priv, 0x419f58, 0x00000000);
+	nvc0_graph_icmd(priv, oclass->icmd);
+	nv_wr32(priv, 0x404154, 0x00000400);
+	nvc0_graph_mthd(priv, oclass->mthd);
+	nv_mask(priv, 0x000260, 0x00000001, 0x00000001);
 }
 
 int
 nvc0_grctx_generate(struct nvc0_graph_priv *priv)
 {
+	struct nvc0_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass;
+	struct nouveau_bar *bar = nouveau_bar(priv);
+	struct nouveau_gpuobj *chan;
 	struct nvc0_grctx info;
-	int ret, i, gpc, tpc, id;
-	u32 fermi = nvc0_graph_class(priv);
-	u32 r000260, tmp;
+	int ret, i;
 
-	ret = nvc0_grctx_init(priv, &info);
-	if (ret)
+	/* allocate memory to for a "channel", which we'll use to generate
+	 * the default context values
+	 */
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x80000 + priv->size,
+				 0x1000, NVOBJ_FLAG_ZERO_ALLOC, &chan);
+	if (ret) {
+		nv_error(priv, "failed to allocate channel memory, %d\n", ret);
 		return ret;
-
-	r000260 = nv_rd32(priv, 0x000260);
-	nv_wr32(priv, 0x000260, r000260 & ~1);
-	nv_wr32(priv, 0x400208, 0x00000000);
-
-	nvc0_grctx_generate_dispatch(priv);
-	nvc0_grctx_generate_macro(priv);
-	nvc0_grctx_generate_m2mf(priv);
-	nvc0_grctx_generate_unk47xx(priv);
-	nvc0_grctx_generate_shaders(priv);
-	nvc0_grctx_generate_unk60xx(priv);
-	nvc0_grctx_generate_unk64xx(priv);
-	nvc0_grctx_generate_tpbus(priv);
-	nvc0_grctx_generate_ccache(priv);
-	nvc0_grctx_generate_rop(priv);
-	nvc0_grctx_generate_gpc(priv);
-	nvc0_grctx_generate_tp(priv);
-
-	nv_wr32(priv, 0x404154, 0x00000000);
-
-	/* generate per-context mmio list data */
-	mmio_data(0x002000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
-	mmio_list(0x408004, 0x00000000,  8, 0);
-	mmio_list(0x408008, 0x80000018,  0, 0);
-	mmio_list(0x40800c, 0x00000000,  8, 1);
-	mmio_list(0x408010, 0x80000000,  0, 0);
-	mmio_list(0x418810, 0x80000000, 12, 2);
-	mmio_list(0x419848, 0x10000000, 12, 2);
-	mmio_list(0x419004, 0x00000000,  8, 1);
-	mmio_list(0x419008, 0x00000000,  0, 0);
-	mmio_list(0x418808, 0x00000000,  8, 0);
-	mmio_list(0x41880c, 0x80000018,  0, 0);
-	if (nv_device(priv)->chipset != 0xc1) {
-		tmp = 0x02180000;
-		mmio_list(0x405830, tmp, 0, 0);
-		for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-			for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-				u32 reg = TPC_UNIT(gpc, tpc, 0x0520);
-				mmio_list(reg, tmp, 0, 0);
-				tmp += 0x0324;
-			}
-		}
-	} else {
-		tmp = 0x02180000;
-		mmio_list(0x405830, 0x00000218 | tmp, 0, 0);
-		mmio_list(0x4064c4, 0x0086ffff, 0, 0);
-		for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-			for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-				u32 reg = TPC_UNIT(gpc, tpc, 0x0520);
-				mmio_list(reg, 0x10000000 | tmp, 0, 0);
-				tmp += 0x0324;
-			}
-			for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-				u32 reg = TPC_UNIT(gpc, tpc, 0x0544);
-				mmio_list(reg, tmp, 0, 0);
-				tmp += 0x0324;
-			}
-		}
 	}
 
-	for (tpc = 0, id = 0; tpc < 4; tpc++) {
-		for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-			if (tpc < priv->tpc_nr[gpc]) {
-				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x698), id);
-				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x4e8), id);
-				nv_wr32(priv, GPC_UNIT(gpc, 0x0c10 + tpc * 4), id);
-				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x088), id);
-				id++;
-			}
-
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0c08), priv->tpc_nr[gpc]);
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0c8c), priv->tpc_nr[gpc]);
-		}
-	}
-
-	tmp = 0;
-	for (i = 0; i < priv->gpc_nr; i++)
-		tmp |= priv->tpc_nr[i] << (i * 4);
-	nv_wr32(priv, 0x406028, tmp);
-	nv_wr32(priv, 0x405870, tmp);
-
-	nv_wr32(priv, 0x40602c, 0x00000000);
-	nv_wr32(priv, 0x405874, 0x00000000);
-	nv_wr32(priv, 0x406030, 0x00000000);
-	nv_wr32(priv, 0x405878, 0x00000000);
-	nv_wr32(priv, 0x406034, 0x00000000);
-	nv_wr32(priv, 0x40587c, 0x00000000);
-
-	if (1) {
-		u8 tpcnr[GPC_MAX], data[TPC_MAX];
-
-		memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
-		memset(data, 0x1f, sizeof(data));
+	/* PGD pointer */
+	nv_wo32(chan, 0x0200, lower_32_bits(chan->addr + 0x1000));
+	nv_wo32(chan, 0x0204, upper_32_bits(chan->addr + 0x1000));
+	nv_wo32(chan, 0x0208, 0xffffffff);
+	nv_wo32(chan, 0x020c, 0x000000ff);
 
-		gpc = -1;
-		for (tpc = 0; tpc < priv->tpc_total; tpc++) {
-			do {
-				gpc = (gpc + 1) % priv->gpc_nr;
-			} while (!tpcnr[gpc]);
-			tpcnr[gpc]--;
-			data[tpc] = gpc;
-		}
+	/* PGT[0] pointer */
+	nv_wo32(chan, 0x1000, 0x00000000);
+	nv_wo32(chan, 0x1004, 0x00000001 | (chan->addr + 0x2000) >> 8);
 
-		for (i = 0; i < 4; i++)
-			nv_wr32(priv, 0x4060a8 + (i * 4), ((u32 *)data)[i]);
+	/* identity-map the whole "channel" into its own vm */
+	for (i = 0; i < chan->size / 4096; i++) {
+		u64 addr = ((chan->addr + (i * 4096)) >> 8) | 1;
+		nv_wo32(chan, 0x2000 + (i * 8), lower_32_bits(addr));
+		nv_wo32(chan, 0x2004 + (i * 8), upper_32_bits(addr));
 	}
 
-	if (1) {
-		u32 data[6] = {}, data2[2] = {};
-		u8 tpcnr[GPC_MAX];
-		u8 shift, ntpcv;
-
-		/* calculate first set of magics */
-		memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+	/* context pointer (virt) */
+	nv_wo32(chan, 0x0210, 0x00080004);
+	nv_wo32(chan, 0x0214, 0x00000000);
 
-		gpc = -1;
-		for (tpc = 0; tpc < priv->tpc_total; tpc++) {
-			do {
-				gpc = (gpc + 1) % priv->gpc_nr;
-			} while (!tpcnr[gpc]);
-			tpcnr[gpc]--;
+	bar->flush(bar);
 
-			data[tpc / 6] |= gpc << ((tpc % 6) * 5);
-		}
+	nv_wr32(priv, 0x100cb8, (chan->addr + 0x1000) >> 8);
+	nv_wr32(priv, 0x100cbc, 0x80000001);
+	nv_wait(priv, 0x100c80, 0x00008000, 0x00008000);
 
-		for (; tpc < 32; tpc++)
-			data[tpc / 6] |= 7 << ((tpc % 6) * 5);
+	/* setup default state for mmio list construction */
+	info.priv = priv;
+	info.data = priv->mmio_data;
+	info.mmio = priv->mmio_list;
+	info.addr = 0x2000 + (i * 8);
+	info.buffer_nr = 0;
 
-		/* and the second... */
-		shift = 0;
-		ntpcv = priv->tpc_total;
-		while (!(ntpcv & (1 << 4))) {
-			ntpcv <<= 1;
-			shift++;
-		}
+	/* make channel current */
+	if (priv->firmware) {
+		nv_wr32(priv, 0x409840, 0x00000030);
+		nv_wr32(priv, 0x409500, 0x80000000 | chan->addr >> 12);
+		nv_wr32(priv, 0x409504, 0x00000003);
+		if (!nv_wait(priv, 0x409800, 0x00000010, 0x00000010))
+			nv_error(priv, "load_ctx timeout\n");
 
-		data2[0]  = (ntpcv << 16);
-		data2[0] |= (shift << 21);
-		data2[0] |= (((1 << (0 + 5)) % ntpcv) << 24);
-		for (i = 1; i < 7; i++)
-			data2[1] |= ((1 << (i + 5)) % ntpcv) << ((i - 1) * 5);
-
-		/* GPC_BROADCAST */
-		nv_wr32(priv, 0x418bb8, (priv->tpc_total << 8) |
-					priv->magic_not_rop_nr);
-		for (i = 0; i < 6; i++)
-			nv_wr32(priv, 0x418b08 + (i * 4), data[i]);
-
-		/* GPC_BROADCAST.TP_BROADCAST */
-		nv_wr32(priv, 0x419bd0, (priv->tpc_total << 8) |
-				       priv->magic_not_rop_nr |
-				       data2[0]);
-		nv_wr32(priv, 0x419be4, data2[1]);
-		for (i = 0; i < 6; i++)
-			nv_wr32(priv, 0x419b00 + (i * 4), data[i]);
-
-		/* UNK78xx */
-		nv_wr32(priv, 0x4078bc, (priv->tpc_total << 8) |
-					priv->magic_not_rop_nr);
-		for (i = 0; i < 6; i++)
-			nv_wr32(priv, 0x40780c + (i * 4), data[i]);
+		nv_wo32(chan, 0x8001c, 1);
+		nv_wo32(chan, 0x80020, 0);
+		nv_wo32(chan, 0x80028, 0);
+		nv_wo32(chan, 0x8002c, 0);
+		bar->flush(bar);
+	} else {
+		nv_wr32(priv, 0x409840, 0x80000000);
+		nv_wr32(priv, 0x409500, 0x80000000 | chan->addr >> 12);
+		nv_wr32(priv, 0x409504, 0x00000001);
+		if (!nv_wait(priv, 0x409800, 0x80000000, 0x80000000))
+			nv_error(priv, "HUB_SET_CHAN timeout\n");
 	}
 
-	if (1) {
-		u32 tpc_mask = 0, tpc_set = 0;
-		u8  tpcnr[GPC_MAX], a, b;
-
-		memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
-		for (gpc = 0; gpc < priv->gpc_nr; gpc++)
-			tpc_mask |= ((1 << priv->tpc_nr[gpc]) - 1) << (gpc * 8);
-
-		for (i = 0, gpc = -1, b = -1; i < 32; i++) {
-			a = (i * (priv->tpc_total - 1)) / 32;
-			if (a != b) {
-				b = a;
-				do {
-					gpc = (gpc + 1) % priv->gpc_nr;
-				} while (!tpcnr[gpc]);
-				tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
-
-				tpc_set |= 1 << ((gpc * 8) + tpc);
-			}
+	oclass->main(priv, &info);
 
-			nv_wr32(priv, 0x406800 + (i * 0x20), tpc_set);
-			nv_wr32(priv, 0x406c00 + (i * 0x20), tpc_set ^ tpc_mask);
-		}
+	/* trigger a context unload by unsetting the "next channel valid" bit
+	 * and faking a context switch interrupt
+	 */
+	nv_mask(priv, 0x409b04, 0x80000000, 0x00000000);
+	nv_wr32(priv, 0x409000, 0x00000100);
+	if (!nv_wait(priv, 0x409b00, 0x80000000, 0x00000000)) {
+		nv_error(priv, "grctx template channel unload timeout\n");
+		ret = -EBUSY;
+		goto done;
 	}
 
-	nv_wr32(priv, 0x400208, 0x80000000);
-
-	nv_icmd(priv, 0x00001000, 0x00000004);
-	nv_icmd(priv, 0x000000a9, 0x0000ffff);
-	nv_icmd(priv, 0x00000038, 0x0fac6881);
-	nv_icmd(priv, 0x0000003d, 0x00000001);
-	nv_icmd(priv, 0x000000e8, 0x00000400);
-	nv_icmd(priv, 0x000000e9, 0x00000400);
-	nv_icmd(priv, 0x000000ea, 0x00000400);
-	nv_icmd(priv, 0x000000eb, 0x00000400);
-	nv_icmd(priv, 0x000000ec, 0x00000400);
-	nv_icmd(priv, 0x000000ed, 0x00000400);
-	nv_icmd(priv, 0x000000ee, 0x00000400);
-	nv_icmd(priv, 0x000000ef, 0x00000400);
-	nv_icmd(priv, 0x00000078, 0x00000300);
-	nv_icmd(priv, 0x00000079, 0x00000300);
-	nv_icmd(priv, 0x0000007a, 0x00000300);
-	nv_icmd(priv, 0x0000007b, 0x00000300);
-	nv_icmd(priv, 0x0000007c, 0x00000300);
-	nv_icmd(priv, 0x0000007d, 0x00000300);
-	nv_icmd(priv, 0x0000007e, 0x00000300);
-	nv_icmd(priv, 0x0000007f, 0x00000300);
-	nv_icmd(priv, 0x00000050, 0x00000011);
-	nv_icmd(priv, 0x00000058, 0x00000008);
-	nv_icmd(priv, 0x00000059, 0x00000008);
-	nv_icmd(priv, 0x0000005a, 0x00000008);
-	nv_icmd(priv, 0x0000005b, 0x00000008);
-	nv_icmd(priv, 0x0000005c, 0x00000008);
-	nv_icmd(priv, 0x0000005d, 0x00000008);
-	nv_icmd(priv, 0x0000005e, 0x00000008);
-	nv_icmd(priv, 0x0000005f, 0x00000008);
-	nv_icmd(priv, 0x00000208, 0x00000001);
-	nv_icmd(priv, 0x00000209, 0x00000001);
-	nv_icmd(priv, 0x0000020a, 0x00000001);
-	nv_icmd(priv, 0x0000020b, 0x00000001);
-	nv_icmd(priv, 0x0000020c, 0x00000001);
-	nv_icmd(priv, 0x0000020d, 0x00000001);
-	nv_icmd(priv, 0x0000020e, 0x00000001);
-	nv_icmd(priv, 0x0000020f, 0x00000001);
-	nv_icmd(priv, 0x00000081, 0x00000001);
-	nv_icmd(priv, 0x00000085, 0x00000004);
-	nv_icmd(priv, 0x00000088, 0x00000400);
-	nv_icmd(priv, 0x00000090, 0x00000300);
-	nv_icmd(priv, 0x00000098, 0x00001001);
-	nv_icmd(priv, 0x000000e3, 0x00000001);
-	nv_icmd(priv, 0x000000da, 0x00000001);
-	nv_icmd(priv, 0x000000f8, 0x00000003);
-	nv_icmd(priv, 0x000000fa, 0x00000001);
-	nv_icmd(priv, 0x0000009f, 0x0000ffff);
-	nv_icmd(priv, 0x000000a0, 0x0000ffff);
-	nv_icmd(priv, 0x000000a1, 0x0000ffff);
-	nv_icmd(priv, 0x000000a2, 0x0000ffff);
-	nv_icmd(priv, 0x000000b1, 0x00000001);
-	nv_icmd(priv, 0x000000b2, 0x00000000);
-	nv_icmd(priv, 0x000000b3, 0x00000000);
-	nv_icmd(priv, 0x000000b4, 0x00000000);
-	nv_icmd(priv, 0x000000b5, 0x00000000);
-	nv_icmd(priv, 0x000000b6, 0x00000000);
-	nv_icmd(priv, 0x000000b7, 0x00000000);
-	nv_icmd(priv, 0x000000b8, 0x00000000);
-	nv_icmd(priv, 0x000000b9, 0x00000000);
-	nv_icmd(priv, 0x000000ba, 0x00000000);
-	nv_icmd(priv, 0x000000bb, 0x00000000);
-	nv_icmd(priv, 0x000000bc, 0x00000000);
-	nv_icmd(priv, 0x000000bd, 0x00000000);
-	nv_icmd(priv, 0x000000be, 0x00000000);
-	nv_icmd(priv, 0x000000bf, 0x00000000);
-	nv_icmd(priv, 0x000000c0, 0x00000000);
-	nv_icmd(priv, 0x000000c1, 0x00000000);
-	nv_icmd(priv, 0x000000c2, 0x00000000);
-	nv_icmd(priv, 0x000000c3, 0x00000000);
-	nv_icmd(priv, 0x000000c4, 0x00000000);
-	nv_icmd(priv, 0x000000c5, 0x00000000);
-	nv_icmd(priv, 0x000000c6, 0x00000000);
-	nv_icmd(priv, 0x000000c7, 0x00000000);
-	nv_icmd(priv, 0x000000c8, 0x00000000);
-	nv_icmd(priv, 0x000000c9, 0x00000000);
-	nv_icmd(priv, 0x000000ca, 0x00000000);
-	nv_icmd(priv, 0x000000cb, 0x00000000);
-	nv_icmd(priv, 0x000000cc, 0x00000000);
-	nv_icmd(priv, 0x000000cd, 0x00000000);
-	nv_icmd(priv, 0x000000ce, 0x00000000);
-	nv_icmd(priv, 0x000000cf, 0x00000000);
-	nv_icmd(priv, 0x000000d0, 0x00000000);
-	nv_icmd(priv, 0x000000d1, 0x00000000);
-	nv_icmd(priv, 0x000000d2, 0x00000000);
-	nv_icmd(priv, 0x000000d3, 0x00000000);
-	nv_icmd(priv, 0x000000d4, 0x00000000);
-	nv_icmd(priv, 0x000000d5, 0x00000000);
-	nv_icmd(priv, 0x000000d6, 0x00000000);
-	nv_icmd(priv, 0x000000d7, 0x00000000);
-	nv_icmd(priv, 0x000000d8, 0x00000000);
-	nv_icmd(priv, 0x000000d9, 0x00000000);
-	nv_icmd(priv, 0x00000210, 0x00000040);
-	nv_icmd(priv, 0x00000211, 0x00000040);
-	nv_icmd(priv, 0x00000212, 0x00000040);
-	nv_icmd(priv, 0x00000213, 0x00000040);
-	nv_icmd(priv, 0x00000214, 0x00000040);
-	nv_icmd(priv, 0x00000215, 0x00000040);
-	nv_icmd(priv, 0x00000216, 0x00000040);
-	nv_icmd(priv, 0x00000217, 0x00000040);
-	if (nv_device(priv)->chipset >= 0xd0) {
-		for (i = 0x0400; i <= 0x0417; i++)
-			nv_icmd(priv, i, 0x00000040);
-	}
-	nv_icmd(priv, 0x00000218, 0x0000c080);
-	nv_icmd(priv, 0x00000219, 0x0000c080);
-	nv_icmd(priv, 0x0000021a, 0x0000c080);
-	nv_icmd(priv, 0x0000021b, 0x0000c080);
-	nv_icmd(priv, 0x0000021c, 0x0000c080);
-	nv_icmd(priv, 0x0000021d, 0x0000c080);
-	nv_icmd(priv, 0x0000021e, 0x0000c080);
-	nv_icmd(priv, 0x0000021f, 0x0000c080);
-	if (nv_device(priv)->chipset >= 0xd0) {
-		for (i = 0x0440; i <= 0x0457; i++)
-			nv_icmd(priv, i, 0x0000c080);
+	priv->data = kmalloc(priv->size, GFP_KERNEL);
+	if (priv->data) {
+		for (i = 0; i < priv->size; i += 4)
+			priv->data[i / 4] = nv_ro32(chan, 0x80000 + i);
+		ret = 0;
+	} else {
+		ret = -ENOMEM;
 	}
-	nv_icmd(priv, 0x000000ad, 0x0000013e);
-	nv_icmd(priv, 0x000000e1, 0x00000010);
-	nv_icmd(priv, 0x00000290, 0x00000000);
-	nv_icmd(priv, 0x00000291, 0x00000000);
-	nv_icmd(priv, 0x00000292, 0x00000000);
-	nv_icmd(priv, 0x00000293, 0x00000000);
-	nv_icmd(priv, 0x00000294, 0x00000000);
-	nv_icmd(priv, 0x00000295, 0x00000000);
-	nv_icmd(priv, 0x00000296, 0x00000000);
-	nv_icmd(priv, 0x00000297, 0x00000000);
-	nv_icmd(priv, 0x00000298, 0x00000000);
-	nv_icmd(priv, 0x00000299, 0x00000000);
-	nv_icmd(priv, 0x0000029a, 0x00000000);
-	nv_icmd(priv, 0x0000029b, 0x00000000);
-	nv_icmd(priv, 0x0000029c, 0x00000000);
-	nv_icmd(priv, 0x0000029d, 0x00000000);
-	nv_icmd(priv, 0x0000029e, 0x00000000);
-	nv_icmd(priv, 0x0000029f, 0x00000000);
-	nv_icmd(priv, 0x000003b0, 0x00000000);
-	nv_icmd(priv, 0x000003b1, 0x00000000);
-	nv_icmd(priv, 0x000003b2, 0x00000000);
-	nv_icmd(priv, 0x000003b3, 0x00000000);
-	nv_icmd(priv, 0x000003b4, 0x00000000);
-	nv_icmd(priv, 0x000003b5, 0x00000000);
-	nv_icmd(priv, 0x000003b6, 0x00000000);
-	nv_icmd(priv, 0x000003b7, 0x00000000);
-	nv_icmd(priv, 0x000003b8, 0x00000000);
-	nv_icmd(priv, 0x000003b9, 0x00000000);
-	nv_icmd(priv, 0x000003ba, 0x00000000);
-	nv_icmd(priv, 0x000003bb, 0x00000000);
-	nv_icmd(priv, 0x000003bc, 0x00000000);
-	nv_icmd(priv, 0x000003bd, 0x00000000);
-	nv_icmd(priv, 0x000003be, 0x00000000);
-	nv_icmd(priv, 0x000003bf, 0x00000000);
-	nv_icmd(priv, 0x000002a0, 0x00000000);
-	nv_icmd(priv, 0x000002a1, 0x00000000);
-	nv_icmd(priv, 0x000002a2, 0x00000000);
-	nv_icmd(priv, 0x000002a3, 0x00000000);
-	nv_icmd(priv, 0x000002a4, 0x00000000);
-	nv_icmd(priv, 0x000002a5, 0x00000000);
-	nv_icmd(priv, 0x000002a6, 0x00000000);
-	nv_icmd(priv, 0x000002a7, 0x00000000);
-	nv_icmd(priv, 0x000002a8, 0x00000000);
-	nv_icmd(priv, 0x000002a9, 0x00000000);
-	nv_icmd(priv, 0x000002aa, 0x00000000);
-	nv_icmd(priv, 0x000002ab, 0x00000000);
-	nv_icmd(priv, 0x000002ac, 0x00000000);
-	nv_icmd(priv, 0x000002ad, 0x00000000);
-	nv_icmd(priv, 0x000002ae, 0x00000000);
-	nv_icmd(priv, 0x000002af, 0x00000000);
-	nv_icmd(priv, 0x00000420, 0x00000000);
-	nv_icmd(priv, 0x00000421, 0x00000000);
-	nv_icmd(priv, 0x00000422, 0x00000000);
-	nv_icmd(priv, 0x00000423, 0x00000000);
-	nv_icmd(priv, 0x00000424, 0x00000000);
-	nv_icmd(priv, 0x00000425, 0x00000000);
-	nv_icmd(priv, 0x00000426, 0x00000000);
-	nv_icmd(priv, 0x00000427, 0x00000000);
-	nv_icmd(priv, 0x00000428, 0x00000000);
-	nv_icmd(priv, 0x00000429, 0x00000000);
-	nv_icmd(priv, 0x0000042a, 0x00000000);
-	nv_icmd(priv, 0x0000042b, 0x00000000);
-	nv_icmd(priv, 0x0000042c, 0x00000000);
-	nv_icmd(priv, 0x0000042d, 0x00000000);
-	nv_icmd(priv, 0x0000042e, 0x00000000);
-	nv_icmd(priv, 0x0000042f, 0x00000000);
-	nv_icmd(priv, 0x000002b0, 0x00000000);
-	nv_icmd(priv, 0x000002b1, 0x00000000);
-	nv_icmd(priv, 0x000002b2, 0x00000000);
-	nv_icmd(priv, 0x000002b3, 0x00000000);
-	nv_icmd(priv, 0x000002b4, 0x00000000);
-	nv_icmd(priv, 0x000002b5, 0x00000000);
-	nv_icmd(priv, 0x000002b6, 0x00000000);
-	nv_icmd(priv, 0x000002b7, 0x00000000);
-	nv_icmd(priv, 0x000002b8, 0x00000000);
-	nv_icmd(priv, 0x000002b9, 0x00000000);
-	nv_icmd(priv, 0x000002ba, 0x00000000);
-	nv_icmd(priv, 0x000002bb, 0x00000000);
-	nv_icmd(priv, 0x000002bc, 0x00000000);
-	nv_icmd(priv, 0x000002bd, 0x00000000);
-	nv_icmd(priv, 0x000002be, 0x00000000);
-	nv_icmd(priv, 0x000002bf, 0x00000000);
-	nv_icmd(priv, 0x00000430, 0x00000000);
-	nv_icmd(priv, 0x00000431, 0x00000000);
-	nv_icmd(priv, 0x00000432, 0x00000000);
-	nv_icmd(priv, 0x00000433, 0x00000000);
-	nv_icmd(priv, 0x00000434, 0x00000000);
-	nv_icmd(priv, 0x00000435, 0x00000000);
-	nv_icmd(priv, 0x00000436, 0x00000000);
-	nv_icmd(priv, 0x00000437, 0x00000000);
-	nv_icmd(priv, 0x00000438, 0x00000000);
-	nv_icmd(priv, 0x00000439, 0x00000000);
-	nv_icmd(priv, 0x0000043a, 0x00000000);
-	nv_icmd(priv, 0x0000043b, 0x00000000);
-	nv_icmd(priv, 0x0000043c, 0x00000000);
-	nv_icmd(priv, 0x0000043d, 0x00000000);
-	nv_icmd(priv, 0x0000043e, 0x00000000);
-	nv_icmd(priv, 0x0000043f, 0x00000000);
-	nv_icmd(priv, 0x000002c0, 0x00000000);
-	nv_icmd(priv, 0x000002c1, 0x00000000);
-	nv_icmd(priv, 0x000002c2, 0x00000000);
-	nv_icmd(priv, 0x000002c3, 0x00000000);
-	nv_icmd(priv, 0x000002c4, 0x00000000);
-	nv_icmd(priv, 0x000002c5, 0x00000000);
-	nv_icmd(priv, 0x000002c6, 0x00000000);
-	nv_icmd(priv, 0x000002c7, 0x00000000);
-	nv_icmd(priv, 0x000002c8, 0x00000000);
-	nv_icmd(priv, 0x000002c9, 0x00000000);
-	nv_icmd(priv, 0x000002ca, 0x00000000);
-	nv_icmd(priv, 0x000002cb, 0x00000000);
-	nv_icmd(priv, 0x000002cc, 0x00000000);
-	nv_icmd(priv, 0x000002cd, 0x00000000);
-	nv_icmd(priv, 0x000002ce, 0x00000000);
-	nv_icmd(priv, 0x000002cf, 0x00000000);
-	nv_icmd(priv, 0x000004d0, 0x00000000);
-	nv_icmd(priv, 0x000004d1, 0x00000000);
-	nv_icmd(priv, 0x000004d2, 0x00000000);
-	nv_icmd(priv, 0x000004d3, 0x00000000);
-	nv_icmd(priv, 0x000004d4, 0x00000000);
-	nv_icmd(priv, 0x000004d5, 0x00000000);
-	nv_icmd(priv, 0x000004d6, 0x00000000);
-	nv_icmd(priv, 0x000004d7, 0x00000000);
-	nv_icmd(priv, 0x000004d8, 0x00000000);
-	nv_icmd(priv, 0x000004d9, 0x00000000);
-	nv_icmd(priv, 0x000004da, 0x00000000);
-	nv_icmd(priv, 0x000004db, 0x00000000);
-	nv_icmd(priv, 0x000004dc, 0x00000000);
-	nv_icmd(priv, 0x000004dd, 0x00000000);
-	nv_icmd(priv, 0x000004de, 0x00000000);
-	nv_icmd(priv, 0x000004df, 0x00000000);
-	nv_icmd(priv, 0x00000720, 0x00000000);
-	nv_icmd(priv, 0x00000721, 0x00000000);
-	nv_icmd(priv, 0x00000722, 0x00000000);
-	nv_icmd(priv, 0x00000723, 0x00000000);
-	nv_icmd(priv, 0x00000724, 0x00000000);
-	nv_icmd(priv, 0x00000725, 0x00000000);
-	nv_icmd(priv, 0x00000726, 0x00000000);
-	nv_icmd(priv, 0x00000727, 0x00000000);
-	nv_icmd(priv, 0x00000728, 0x00000000);
-	nv_icmd(priv, 0x00000729, 0x00000000);
-	nv_icmd(priv, 0x0000072a, 0x00000000);
-	nv_icmd(priv, 0x0000072b, 0x00000000);
-	nv_icmd(priv, 0x0000072c, 0x00000000);
-	nv_icmd(priv, 0x0000072d, 0x00000000);
-	nv_icmd(priv, 0x0000072e, 0x00000000);
-	nv_icmd(priv, 0x0000072f, 0x00000000);
-	nv_icmd(priv, 0x000008c0, 0x00000000);
-	nv_icmd(priv, 0x000008c1, 0x00000000);
-	nv_icmd(priv, 0x000008c2, 0x00000000);
-	nv_icmd(priv, 0x000008c3, 0x00000000);
-	nv_icmd(priv, 0x000008c4, 0x00000000);
-	nv_icmd(priv, 0x000008c5, 0x00000000);
-	nv_icmd(priv, 0x000008c6, 0x00000000);
-	nv_icmd(priv, 0x000008c7, 0x00000000);
-	nv_icmd(priv, 0x000008c8, 0x00000000);
-	nv_icmd(priv, 0x000008c9, 0x00000000);
-	nv_icmd(priv, 0x000008ca, 0x00000000);
-	nv_icmd(priv, 0x000008cb, 0x00000000);
-	nv_icmd(priv, 0x000008cc, 0x00000000);
-	nv_icmd(priv, 0x000008cd, 0x00000000);
-	nv_icmd(priv, 0x000008ce, 0x00000000);
-	nv_icmd(priv, 0x000008cf, 0x00000000);
-	nv_icmd(priv, 0x00000890, 0x00000000);
-	nv_icmd(priv, 0x00000891, 0x00000000);
-	nv_icmd(priv, 0x00000892, 0x00000000);
-	nv_icmd(priv, 0x00000893, 0x00000000);
-	nv_icmd(priv, 0x00000894, 0x00000000);
-	nv_icmd(priv, 0x00000895, 0x00000000);
-	nv_icmd(priv, 0x00000896, 0x00000000);
-	nv_icmd(priv, 0x00000897, 0x00000000);
-	nv_icmd(priv, 0x00000898, 0x00000000);
-	nv_icmd(priv, 0x00000899, 0x00000000);
-	nv_icmd(priv, 0x0000089a, 0x00000000);
-	nv_icmd(priv, 0x0000089b, 0x00000000);
-	nv_icmd(priv, 0x0000089c, 0x00000000);
-	nv_icmd(priv, 0x0000089d, 0x00000000);
-	nv_icmd(priv, 0x0000089e, 0x00000000);
-	nv_icmd(priv, 0x0000089f, 0x00000000);
-	nv_icmd(priv, 0x000008e0, 0x00000000);
-	nv_icmd(priv, 0x000008e1, 0x00000000);
-	nv_icmd(priv, 0x000008e2, 0x00000000);
-	nv_icmd(priv, 0x000008e3, 0x00000000);
-	nv_icmd(priv, 0x000008e4, 0x00000000);
-	nv_icmd(priv, 0x000008e5, 0x00000000);
-	nv_icmd(priv, 0x000008e6, 0x00000000);
-	nv_icmd(priv, 0x000008e7, 0x00000000);
-	nv_icmd(priv, 0x000008e8, 0x00000000);
-	nv_icmd(priv, 0x000008e9, 0x00000000);
-	nv_icmd(priv, 0x000008ea, 0x00000000);
-	nv_icmd(priv, 0x000008eb, 0x00000000);
-	nv_icmd(priv, 0x000008ec, 0x00000000);
-	nv_icmd(priv, 0x000008ed, 0x00000000);
-	nv_icmd(priv, 0x000008ee, 0x00000000);
-	nv_icmd(priv, 0x000008ef, 0x00000000);
-	nv_icmd(priv, 0x000008a0, 0x00000000);
-	nv_icmd(priv, 0x000008a1, 0x00000000);
-	nv_icmd(priv, 0x000008a2, 0x00000000);
-	nv_icmd(priv, 0x000008a3, 0x00000000);
-	nv_icmd(priv, 0x000008a4, 0x00000000);
-	nv_icmd(priv, 0x000008a5, 0x00000000);
-	nv_icmd(priv, 0x000008a6, 0x00000000);
-	nv_icmd(priv, 0x000008a7, 0x00000000);
-	nv_icmd(priv, 0x000008a8, 0x00000000);
-	nv_icmd(priv, 0x000008a9, 0x00000000);
-	nv_icmd(priv, 0x000008aa, 0x00000000);
-	nv_icmd(priv, 0x000008ab, 0x00000000);
-	nv_icmd(priv, 0x000008ac, 0x00000000);
-	nv_icmd(priv, 0x000008ad, 0x00000000);
-	nv_icmd(priv, 0x000008ae, 0x00000000);
-	nv_icmd(priv, 0x000008af, 0x00000000);
-	nv_icmd(priv, 0x000008f0, 0x00000000);
-	nv_icmd(priv, 0x000008f1, 0x00000000);
-	nv_icmd(priv, 0x000008f2, 0x00000000);
-	nv_icmd(priv, 0x000008f3, 0x00000000);
-	nv_icmd(priv, 0x000008f4, 0x00000000);
-	nv_icmd(priv, 0x000008f5, 0x00000000);
-	nv_icmd(priv, 0x000008f6, 0x00000000);
-	nv_icmd(priv, 0x000008f7, 0x00000000);
-	nv_icmd(priv, 0x000008f8, 0x00000000);
-	nv_icmd(priv, 0x000008f9, 0x00000000);
-	nv_icmd(priv, 0x000008fa, 0x00000000);
-	nv_icmd(priv, 0x000008fb, 0x00000000);
-	nv_icmd(priv, 0x000008fc, 0x00000000);
-	nv_icmd(priv, 0x000008fd, 0x00000000);
-	nv_icmd(priv, 0x000008fe, 0x00000000);
-	nv_icmd(priv, 0x000008ff, 0x00000000);
-	nv_icmd(priv, 0x0000094c, 0x000000ff);
-	nv_icmd(priv, 0x0000094d, 0xffffffff);
-	nv_icmd(priv, 0x0000094e, 0x00000002);
-	nv_icmd(priv, 0x000002ec, 0x00000001);
-	nv_icmd(priv, 0x00000303, 0x00000001);
-	nv_icmd(priv, 0x000002e6, 0x00000001);
-	nv_icmd(priv, 0x00000466, 0x00000052);
-	nv_icmd(priv, 0x00000301, 0x3f800000);
-	nv_icmd(priv, 0x00000304, 0x30201000);
-	nv_icmd(priv, 0x00000305, 0x70605040);
-	nv_icmd(priv, 0x00000306, 0xb8a89888);
-	nv_icmd(priv, 0x00000307, 0xf8e8d8c8);
-	nv_icmd(priv, 0x0000030a, 0x00ffff00);
-	nv_icmd(priv, 0x0000030b, 0x0000001a);
-	nv_icmd(priv, 0x0000030c, 0x00000001);
-	nv_icmd(priv, 0x00000318, 0x00000001);
-	nv_icmd(priv, 0x00000340, 0x00000000);
-	nv_icmd(priv, 0x00000375, 0x00000001);
-	nv_icmd(priv, 0x00000351, 0x00000100);
-	nv_icmd(priv, 0x0000037d, 0x00000006);
-	nv_icmd(priv, 0x000003a0, 0x00000002);
-	nv_icmd(priv, 0x000003aa, 0x00000001);
-	nv_icmd(priv, 0x000003a9, 0x00000001);
-	nv_icmd(priv, 0x00000380, 0x00000001);
-	nv_icmd(priv, 0x00000360, 0x00000040);
-	nv_icmd(priv, 0x00000366, 0x00000000);
-	nv_icmd(priv, 0x00000367, 0x00000000);
-	nv_icmd(priv, 0x00000368, 0x00001fff);
-	nv_icmd(priv, 0x00000370, 0x00000000);
-	nv_icmd(priv, 0x00000371, 0x00000000);
-	nv_icmd(priv, 0x00000372, 0x003fffff);
-	nv_icmd(priv, 0x0000037a, 0x00000012);
-	nv_icmd(priv, 0x000005e0, 0x00000022);
-	nv_icmd(priv, 0x000005e1, 0x00000022);
-	nv_icmd(priv, 0x000005e2, 0x00000022);
-	nv_icmd(priv, 0x000005e3, 0x00000022);
-	nv_icmd(priv, 0x000005e4, 0x00000022);
-	nv_icmd(priv, 0x00000619, 0x00000003);
-	nv_icmd(priv, 0x00000811, 0x00000003);
-	nv_icmd(priv, 0x00000812, 0x00000004);
-	nv_icmd(priv, 0x00000813, 0x00000006);
-	nv_icmd(priv, 0x00000814, 0x00000008);
-	nv_icmd(priv, 0x00000815, 0x0000000b);
-	nv_icmd(priv, 0x00000800, 0x00000001);
-	nv_icmd(priv, 0x00000801, 0x00000001);
-	nv_icmd(priv, 0x00000802, 0x00000001);
-	nv_icmd(priv, 0x00000803, 0x00000001);
-	nv_icmd(priv, 0x00000804, 0x00000001);
-	nv_icmd(priv, 0x00000805, 0x00000001);
-	nv_icmd(priv, 0x00000632, 0x00000001);
-	nv_icmd(priv, 0x00000633, 0x00000002);
-	nv_icmd(priv, 0x00000634, 0x00000003);
-	nv_icmd(priv, 0x00000635, 0x00000004);
-	nv_icmd(priv, 0x00000654, 0x3f800000);
-	nv_icmd(priv, 0x00000657, 0x3f800000);
-	nv_icmd(priv, 0x00000655, 0x3f800000);
-	nv_icmd(priv, 0x00000656, 0x3f800000);
-	nv_icmd(priv, 0x000006cd, 0x3f800000);
-	nv_icmd(priv, 0x000007f5, 0x3f800000);
-	nv_icmd(priv, 0x000007dc, 0x39291909);
-	nv_icmd(priv, 0x000007dd, 0x79695949);
-	nv_icmd(priv, 0x000007de, 0xb9a99989);
-	nv_icmd(priv, 0x000007df, 0xf9e9d9c9);
-	nv_icmd(priv, 0x000007e8, 0x00003210);
-	nv_icmd(priv, 0x000007e9, 0x00007654);
-	nv_icmd(priv, 0x000007ea, 0x00000098);
-	nv_icmd(priv, 0x000007ec, 0x39291909);
-	nv_icmd(priv, 0x000007ed, 0x79695949);
-	nv_icmd(priv, 0x000007ee, 0xb9a99989);
-	nv_icmd(priv, 0x000007ef, 0xf9e9d9c9);
-	nv_icmd(priv, 0x000007f0, 0x00003210);
-	nv_icmd(priv, 0x000007f1, 0x00007654);
-	nv_icmd(priv, 0x000007f2, 0x00000098);
-	nv_icmd(priv, 0x000005a5, 0x00000001);
-	nv_icmd(priv, 0x00000980, 0x00000000);
-	nv_icmd(priv, 0x00000981, 0x00000000);
-	nv_icmd(priv, 0x00000982, 0x00000000);
-	nv_icmd(priv, 0x00000983, 0x00000000);
-	nv_icmd(priv, 0x00000984, 0x00000000);
-	nv_icmd(priv, 0x00000985, 0x00000000);
-	nv_icmd(priv, 0x00000986, 0x00000000);
-	nv_icmd(priv, 0x00000987, 0x00000000);
-	nv_icmd(priv, 0x00000988, 0x00000000);
-	nv_icmd(priv, 0x00000989, 0x00000000);
-	nv_icmd(priv, 0x0000098a, 0x00000000);
-	nv_icmd(priv, 0x0000098b, 0x00000000);
-	nv_icmd(priv, 0x0000098c, 0x00000000);
-	nv_icmd(priv, 0x0000098d, 0x00000000);
-	nv_icmd(priv, 0x0000098e, 0x00000000);
-	nv_icmd(priv, 0x0000098f, 0x00000000);
-	nv_icmd(priv, 0x00000990, 0x00000000);
-	nv_icmd(priv, 0x00000991, 0x00000000);
-	nv_icmd(priv, 0x00000992, 0x00000000);
-	nv_icmd(priv, 0x00000993, 0x00000000);
-	nv_icmd(priv, 0x00000994, 0x00000000);
-	nv_icmd(priv, 0x00000995, 0x00000000);
-	nv_icmd(priv, 0x00000996, 0x00000000);
-	nv_icmd(priv, 0x00000997, 0x00000000);
-	nv_icmd(priv, 0x00000998, 0x00000000);
-	nv_icmd(priv, 0x00000999, 0x00000000);
-	nv_icmd(priv, 0x0000099a, 0x00000000);
-	nv_icmd(priv, 0x0000099b, 0x00000000);
-	nv_icmd(priv, 0x0000099c, 0x00000000);
-	nv_icmd(priv, 0x0000099d, 0x00000000);
-	nv_icmd(priv, 0x0000099e, 0x00000000);
-	nv_icmd(priv, 0x0000099f, 0x00000000);
-	nv_icmd(priv, 0x000009a0, 0x00000000);
-	nv_icmd(priv, 0x000009a1, 0x00000000);
-	nv_icmd(priv, 0x000009a2, 0x00000000);
-	nv_icmd(priv, 0x000009a3, 0x00000000);
-	nv_icmd(priv, 0x000009a4, 0x00000000);
-	nv_icmd(priv, 0x000009a5, 0x00000000);
-	nv_icmd(priv, 0x000009a6, 0x00000000);
-	nv_icmd(priv, 0x000009a7, 0x00000000);
-	nv_icmd(priv, 0x000009a8, 0x00000000);
-	nv_icmd(priv, 0x000009a9, 0x00000000);
-	nv_icmd(priv, 0x000009aa, 0x00000000);
-	nv_icmd(priv, 0x000009ab, 0x00000000);
-	nv_icmd(priv, 0x000009ac, 0x00000000);
-	nv_icmd(priv, 0x000009ad, 0x00000000);
-	nv_icmd(priv, 0x000009ae, 0x00000000);
-	nv_icmd(priv, 0x000009af, 0x00000000);
-	nv_icmd(priv, 0x000009b0, 0x00000000);
-	nv_icmd(priv, 0x000009b1, 0x00000000);
-	nv_icmd(priv, 0x000009b2, 0x00000000);
-	nv_icmd(priv, 0x000009b3, 0x00000000);
-	nv_icmd(priv, 0x000009b4, 0x00000000);
-	nv_icmd(priv, 0x000009b5, 0x00000000);
-	nv_icmd(priv, 0x000009b6, 0x00000000);
-	nv_icmd(priv, 0x000009b7, 0x00000000);
-	nv_icmd(priv, 0x000009b8, 0x00000000);
-	nv_icmd(priv, 0x000009b9, 0x00000000);
-	nv_icmd(priv, 0x000009ba, 0x00000000);
-	nv_icmd(priv, 0x000009bb, 0x00000000);
-	nv_icmd(priv, 0x000009bc, 0x00000000);
-	nv_icmd(priv, 0x000009bd, 0x00000000);
-	nv_icmd(priv, 0x000009be, 0x00000000);
-	nv_icmd(priv, 0x000009bf, 0x00000000);
-	nv_icmd(priv, 0x000009c0, 0x00000000);
-	nv_icmd(priv, 0x000009c1, 0x00000000);
-	nv_icmd(priv, 0x000009c2, 0x00000000);
-	nv_icmd(priv, 0x000009c3, 0x00000000);
-	nv_icmd(priv, 0x000009c4, 0x00000000);
-	nv_icmd(priv, 0x000009c5, 0x00000000);
-	nv_icmd(priv, 0x000009c6, 0x00000000);
-	nv_icmd(priv, 0x000009c7, 0x00000000);
-	nv_icmd(priv, 0x000009c8, 0x00000000);
-	nv_icmd(priv, 0x000009c9, 0x00000000);
-	nv_icmd(priv, 0x000009ca, 0x00000000);
-	nv_icmd(priv, 0x000009cb, 0x00000000);
-	nv_icmd(priv, 0x000009cc, 0x00000000);
-	nv_icmd(priv, 0x000009cd, 0x00000000);
-	nv_icmd(priv, 0x000009ce, 0x00000000);
-	nv_icmd(priv, 0x000009cf, 0x00000000);
-	nv_icmd(priv, 0x000009d0, 0x00000000);
-	nv_icmd(priv, 0x000009d1, 0x00000000);
-	nv_icmd(priv, 0x000009d2, 0x00000000);
-	nv_icmd(priv, 0x000009d3, 0x00000000);
-	nv_icmd(priv, 0x000009d4, 0x00000000);
-	nv_icmd(priv, 0x000009d5, 0x00000000);
-	nv_icmd(priv, 0x000009d6, 0x00000000);
-	nv_icmd(priv, 0x000009d7, 0x00000000);
-	nv_icmd(priv, 0x000009d8, 0x00000000);
-	nv_icmd(priv, 0x000009d9, 0x00000000);
-	nv_icmd(priv, 0x000009da, 0x00000000);
-	nv_icmd(priv, 0x000009db, 0x00000000);
-	nv_icmd(priv, 0x000009dc, 0x00000000);
-	nv_icmd(priv, 0x000009dd, 0x00000000);
-	nv_icmd(priv, 0x000009de, 0x00000000);
-	nv_icmd(priv, 0x000009df, 0x00000000);
-	nv_icmd(priv, 0x000009e0, 0x00000000);
-	nv_icmd(priv, 0x000009e1, 0x00000000);
-	nv_icmd(priv, 0x000009e2, 0x00000000);
-	nv_icmd(priv, 0x000009e3, 0x00000000);
-	nv_icmd(priv, 0x000009e4, 0x00000000);
-	nv_icmd(priv, 0x000009e5, 0x00000000);
-	nv_icmd(priv, 0x000009e6, 0x00000000);
-	nv_icmd(priv, 0x000009e7, 0x00000000);
-	nv_icmd(priv, 0x000009e8, 0x00000000);
-	nv_icmd(priv, 0x000009e9, 0x00000000);
-	nv_icmd(priv, 0x000009ea, 0x00000000);
-	nv_icmd(priv, 0x000009eb, 0x00000000);
-	nv_icmd(priv, 0x000009ec, 0x00000000);
-	nv_icmd(priv, 0x000009ed, 0x00000000);
-	nv_icmd(priv, 0x000009ee, 0x00000000);
-	nv_icmd(priv, 0x000009ef, 0x00000000);
-	nv_icmd(priv, 0x000009f0, 0x00000000);
-	nv_icmd(priv, 0x000009f1, 0x00000000);
-	nv_icmd(priv, 0x000009f2, 0x00000000);
-	nv_icmd(priv, 0x000009f3, 0x00000000);
-	nv_icmd(priv, 0x000009f4, 0x00000000);
-	nv_icmd(priv, 0x000009f5, 0x00000000);
-	nv_icmd(priv, 0x000009f6, 0x00000000);
-	nv_icmd(priv, 0x000009f7, 0x00000000);
-	nv_icmd(priv, 0x000009f8, 0x00000000);
-	nv_icmd(priv, 0x000009f9, 0x00000000);
-	nv_icmd(priv, 0x000009fa, 0x00000000);
-	nv_icmd(priv, 0x000009fb, 0x00000000);
-	nv_icmd(priv, 0x000009fc, 0x00000000);
-	nv_icmd(priv, 0x000009fd, 0x00000000);
-	nv_icmd(priv, 0x000009fe, 0x00000000);
-	nv_icmd(priv, 0x000009ff, 0x00000000);
-	nv_icmd(priv, 0x00000468, 0x00000004);
-	nv_icmd(priv, 0x0000046c, 0x00000001);
-	nv_icmd(priv, 0x00000470, 0x00000000);
-	nv_icmd(priv, 0x00000471, 0x00000000);
-	nv_icmd(priv, 0x00000472, 0x00000000);
-	nv_icmd(priv, 0x00000473, 0x00000000);
-	nv_icmd(priv, 0x00000474, 0x00000000);
-	nv_icmd(priv, 0x00000475, 0x00000000);
-	nv_icmd(priv, 0x00000476, 0x00000000);
-	nv_icmd(priv, 0x00000477, 0x00000000);
-	nv_icmd(priv, 0x00000478, 0x00000000);
-	nv_icmd(priv, 0x00000479, 0x00000000);
-	nv_icmd(priv, 0x0000047a, 0x00000000);
-	nv_icmd(priv, 0x0000047b, 0x00000000);
-	nv_icmd(priv, 0x0000047c, 0x00000000);
-	nv_icmd(priv, 0x0000047d, 0x00000000);
-	nv_icmd(priv, 0x0000047e, 0x00000000);
-	nv_icmd(priv, 0x0000047f, 0x00000000);
-	nv_icmd(priv, 0x00000480, 0x00000000);
-	nv_icmd(priv, 0x00000481, 0x00000000);
-	nv_icmd(priv, 0x00000482, 0x00000000);
-	nv_icmd(priv, 0x00000483, 0x00000000);
-	nv_icmd(priv, 0x00000484, 0x00000000);
-	nv_icmd(priv, 0x00000485, 0x00000000);
-	nv_icmd(priv, 0x00000486, 0x00000000);
-	nv_icmd(priv, 0x00000487, 0x00000000);
-	nv_icmd(priv, 0x00000488, 0x00000000);
-	nv_icmd(priv, 0x00000489, 0x00000000);
-	nv_icmd(priv, 0x0000048a, 0x00000000);
-	nv_icmd(priv, 0x0000048b, 0x00000000);
-	nv_icmd(priv, 0x0000048c, 0x00000000);
-	nv_icmd(priv, 0x0000048d, 0x00000000);
-	nv_icmd(priv, 0x0000048e, 0x00000000);
-	nv_icmd(priv, 0x0000048f, 0x00000000);
-	nv_icmd(priv, 0x00000490, 0x00000000);
-	nv_icmd(priv, 0x00000491, 0x00000000);
-	nv_icmd(priv, 0x00000492, 0x00000000);
-	nv_icmd(priv, 0x00000493, 0x00000000);
-	nv_icmd(priv, 0x00000494, 0x00000000);
-	nv_icmd(priv, 0x00000495, 0x00000000);
-	nv_icmd(priv, 0x00000496, 0x00000000);
-	nv_icmd(priv, 0x00000497, 0x00000000);
-	nv_icmd(priv, 0x00000498, 0x00000000);
-	nv_icmd(priv, 0x00000499, 0x00000000);
-	nv_icmd(priv, 0x0000049a, 0x00000000);
-	nv_icmd(priv, 0x0000049b, 0x00000000);
-	nv_icmd(priv, 0x0000049c, 0x00000000);
-	nv_icmd(priv, 0x0000049d, 0x00000000);
-	nv_icmd(priv, 0x0000049e, 0x00000000);
-	nv_icmd(priv, 0x0000049f, 0x00000000);
-	nv_icmd(priv, 0x000004a0, 0x00000000);
-	nv_icmd(priv, 0x000004a1, 0x00000000);
-	nv_icmd(priv, 0x000004a2, 0x00000000);
-	nv_icmd(priv, 0x000004a3, 0x00000000);
-	nv_icmd(priv, 0x000004a4, 0x00000000);
-	nv_icmd(priv, 0x000004a5, 0x00000000);
-	nv_icmd(priv, 0x000004a6, 0x00000000);
-	nv_icmd(priv, 0x000004a7, 0x00000000);
-	nv_icmd(priv, 0x000004a8, 0x00000000);
-	nv_icmd(priv, 0x000004a9, 0x00000000);
-	nv_icmd(priv, 0x000004aa, 0x00000000);
-	nv_icmd(priv, 0x000004ab, 0x00000000);
-	nv_icmd(priv, 0x000004ac, 0x00000000);
-	nv_icmd(priv, 0x000004ad, 0x00000000);
-	nv_icmd(priv, 0x000004ae, 0x00000000);
-	nv_icmd(priv, 0x000004af, 0x00000000);
-	nv_icmd(priv, 0x000004b0, 0x00000000);
-	nv_icmd(priv, 0x000004b1, 0x00000000);
-	nv_icmd(priv, 0x000004b2, 0x00000000);
-	nv_icmd(priv, 0x000004b3, 0x00000000);
-	nv_icmd(priv, 0x000004b4, 0x00000000);
-	nv_icmd(priv, 0x000004b5, 0x00000000);
-	nv_icmd(priv, 0x000004b6, 0x00000000);
-	nv_icmd(priv, 0x000004b7, 0x00000000);
-	nv_icmd(priv, 0x000004b8, 0x00000000);
-	nv_icmd(priv, 0x000004b9, 0x00000000);
-	nv_icmd(priv, 0x000004ba, 0x00000000);
-	nv_icmd(priv, 0x000004bb, 0x00000000);
-	nv_icmd(priv, 0x000004bc, 0x00000000);
-	nv_icmd(priv, 0x000004bd, 0x00000000);
-	nv_icmd(priv, 0x000004be, 0x00000000);
-	nv_icmd(priv, 0x000004bf, 0x00000000);
-	nv_icmd(priv, 0x000004c0, 0x00000000);
-	nv_icmd(priv, 0x000004c1, 0x00000000);
-	nv_icmd(priv, 0x000004c2, 0x00000000);
-	nv_icmd(priv, 0x000004c3, 0x00000000);
-	nv_icmd(priv, 0x000004c4, 0x00000000);
-	nv_icmd(priv, 0x000004c5, 0x00000000);
-	nv_icmd(priv, 0x000004c6, 0x00000000);
-	nv_icmd(priv, 0x000004c7, 0x00000000);
-	nv_icmd(priv, 0x000004c8, 0x00000000);
-	nv_icmd(priv, 0x000004c9, 0x00000000);
-	nv_icmd(priv, 0x000004ca, 0x00000000);
-	nv_icmd(priv, 0x000004cb, 0x00000000);
-	nv_icmd(priv, 0x000004cc, 0x00000000);
-	nv_icmd(priv, 0x000004cd, 0x00000000);
-	nv_icmd(priv, 0x000004ce, 0x00000000);
-	nv_icmd(priv, 0x000004cf, 0x00000000);
-	nv_icmd(priv, 0x00000510, 0x3f800000);
-	nv_icmd(priv, 0x00000511, 0x3f800000);
-	nv_icmd(priv, 0x00000512, 0x3f800000);
-	nv_icmd(priv, 0x00000513, 0x3f800000);
-	nv_icmd(priv, 0x00000514, 0x3f800000);
-	nv_icmd(priv, 0x00000515, 0x3f800000);
-	nv_icmd(priv, 0x00000516, 0x3f800000);
-	nv_icmd(priv, 0x00000517, 0x3f800000);
-	nv_icmd(priv, 0x00000518, 0x3f800000);
-	nv_icmd(priv, 0x00000519, 0x3f800000);
-	nv_icmd(priv, 0x0000051a, 0x3f800000);
-	nv_icmd(priv, 0x0000051b, 0x3f800000);
-	nv_icmd(priv, 0x0000051c, 0x3f800000);
-	nv_icmd(priv, 0x0000051d, 0x3f800000);
-	nv_icmd(priv, 0x0000051e, 0x3f800000);
-	nv_icmd(priv, 0x0000051f, 0x3f800000);
-	nv_icmd(priv, 0x00000520, 0x000002b6);
-	nv_icmd(priv, 0x00000529, 0x00000001);
-	nv_icmd(priv, 0x00000530, 0xffff0000);
-	nv_icmd(priv, 0x00000531, 0xffff0000);
-	nv_icmd(priv, 0x00000532, 0xffff0000);
-	nv_icmd(priv, 0x00000533, 0xffff0000);
-	nv_icmd(priv, 0x00000534, 0xffff0000);
-	nv_icmd(priv, 0x00000535, 0xffff0000);
-	nv_icmd(priv, 0x00000536, 0xffff0000);
-	nv_icmd(priv, 0x00000537, 0xffff0000);
-	nv_icmd(priv, 0x00000538, 0xffff0000);
-	nv_icmd(priv, 0x00000539, 0xffff0000);
-	nv_icmd(priv, 0x0000053a, 0xffff0000);
-	nv_icmd(priv, 0x0000053b, 0xffff0000);
-	nv_icmd(priv, 0x0000053c, 0xffff0000);
-	nv_icmd(priv, 0x0000053d, 0xffff0000);
-	nv_icmd(priv, 0x0000053e, 0xffff0000);
-	nv_icmd(priv, 0x0000053f, 0xffff0000);
-	nv_icmd(priv, 0x00000585, 0x0000003f);
-	nv_icmd(priv, 0x00000576, 0x00000003);
-	if (nv_device(priv)->chipset == 0xc1 ||
-	    nv_device(priv)->chipset >= 0xd0)
-		nv_icmd(priv, 0x0000057b, 0x00000059);
-	nv_icmd(priv, 0x00000586, 0x00000040);
-	nv_icmd(priv, 0x00000582, 0x00000080);
-	nv_icmd(priv, 0x00000583, 0x00000080);
-	nv_icmd(priv, 0x000005c2, 0x00000001);
-	nv_icmd(priv, 0x00000638, 0x00000001);
-	nv_icmd(priv, 0x00000639, 0x00000001);
-	nv_icmd(priv, 0x0000063a, 0x00000002);
-	nv_icmd(priv, 0x0000063b, 0x00000001);
-	nv_icmd(priv, 0x0000063c, 0x00000001);
-	nv_icmd(priv, 0x0000063d, 0x00000002);
-	nv_icmd(priv, 0x0000063e, 0x00000001);
-	nv_icmd(priv, 0x000008b8, 0x00000001);
-	nv_icmd(priv, 0x000008b9, 0x00000001);
-	nv_icmd(priv, 0x000008ba, 0x00000001);
-	nv_icmd(priv, 0x000008bb, 0x00000001);
-	nv_icmd(priv, 0x000008bc, 0x00000001);
-	nv_icmd(priv, 0x000008bd, 0x00000001);
-	nv_icmd(priv, 0x000008be, 0x00000001);
-	nv_icmd(priv, 0x000008bf, 0x00000001);
-	nv_icmd(priv, 0x00000900, 0x00000001);
-	nv_icmd(priv, 0x00000901, 0x00000001);
-	nv_icmd(priv, 0x00000902, 0x00000001);
-	nv_icmd(priv, 0x00000903, 0x00000001);
-	nv_icmd(priv, 0x00000904, 0x00000001);
-	nv_icmd(priv, 0x00000905, 0x00000001);
-	nv_icmd(priv, 0x00000906, 0x00000001);
-	nv_icmd(priv, 0x00000907, 0x00000001);
-	nv_icmd(priv, 0x00000908, 0x00000002);
-	nv_icmd(priv, 0x00000909, 0x00000002);
-	nv_icmd(priv, 0x0000090a, 0x00000002);
-	nv_icmd(priv, 0x0000090b, 0x00000002);
-	nv_icmd(priv, 0x0000090c, 0x00000002);
-	nv_icmd(priv, 0x0000090d, 0x00000002);
-	nv_icmd(priv, 0x0000090e, 0x00000002);
-	nv_icmd(priv, 0x0000090f, 0x00000002);
-	nv_icmd(priv, 0x00000910, 0x00000001);
-	nv_icmd(priv, 0x00000911, 0x00000001);
-	nv_icmd(priv, 0x00000912, 0x00000001);
-	nv_icmd(priv, 0x00000913, 0x00000001);
-	nv_icmd(priv, 0x00000914, 0x00000001);
-	nv_icmd(priv, 0x00000915, 0x00000001);
-	nv_icmd(priv, 0x00000916, 0x00000001);
-	nv_icmd(priv, 0x00000917, 0x00000001);
-	nv_icmd(priv, 0x00000918, 0x00000001);
-	nv_icmd(priv, 0x00000919, 0x00000001);
-	nv_icmd(priv, 0x0000091a, 0x00000001);
-	nv_icmd(priv, 0x0000091b, 0x00000001);
-	nv_icmd(priv, 0x0000091c, 0x00000001);
-	nv_icmd(priv, 0x0000091d, 0x00000001);
-	nv_icmd(priv, 0x0000091e, 0x00000001);
-	nv_icmd(priv, 0x0000091f, 0x00000001);
-	nv_icmd(priv, 0x00000920, 0x00000002);
-	nv_icmd(priv, 0x00000921, 0x00000002);
-	nv_icmd(priv, 0x00000922, 0x00000002);
-	nv_icmd(priv, 0x00000923, 0x00000002);
-	nv_icmd(priv, 0x00000924, 0x00000002);
-	nv_icmd(priv, 0x00000925, 0x00000002);
-	nv_icmd(priv, 0x00000926, 0x00000002);
-	nv_icmd(priv, 0x00000927, 0x00000002);
-	nv_icmd(priv, 0x00000928, 0x00000001);
-	nv_icmd(priv, 0x00000929, 0x00000001);
-	nv_icmd(priv, 0x0000092a, 0x00000001);
-	nv_icmd(priv, 0x0000092b, 0x00000001);
-	nv_icmd(priv, 0x0000092c, 0x00000001);
-	nv_icmd(priv, 0x0000092d, 0x00000001);
-	nv_icmd(priv, 0x0000092e, 0x00000001);
-	nv_icmd(priv, 0x0000092f, 0x00000001);
-	nv_icmd(priv, 0x00000648, 0x00000001);
-	nv_icmd(priv, 0x00000649, 0x00000001);
-	nv_icmd(priv, 0x0000064a, 0x00000001);
-	nv_icmd(priv, 0x0000064b, 0x00000001);
-	nv_icmd(priv, 0x0000064c, 0x00000001);
-	nv_icmd(priv, 0x0000064d, 0x00000001);
-	nv_icmd(priv, 0x0000064e, 0x00000001);
-	nv_icmd(priv, 0x0000064f, 0x00000001);
-	nv_icmd(priv, 0x00000650, 0x00000001);
-	nv_icmd(priv, 0x00000658, 0x0000000f);
-	nv_icmd(priv, 0x000007ff, 0x0000000a);
-	nv_icmd(priv, 0x0000066a, 0x40000000);
-	nv_icmd(priv, 0x0000066b, 0x10000000);
-	nv_icmd(priv, 0x0000066c, 0xffff0000);
-	nv_icmd(priv, 0x0000066d, 0xffff0000);
-	nv_icmd(priv, 0x000007af, 0x00000008);
-	nv_icmd(priv, 0x000007b0, 0x00000008);
-	nv_icmd(priv, 0x000007f6, 0x00000001);
-	nv_icmd(priv, 0x000006b2, 0x00000055);
-	nv_icmd(priv, 0x000007ad, 0x00000003);
-	nv_icmd(priv, 0x00000937, 0x00000001);
-	nv_icmd(priv, 0x00000971, 0x00000008);
-	nv_icmd(priv, 0x00000972, 0x00000040);
-	nv_icmd(priv, 0x00000973, 0x0000012c);
-	nv_icmd(priv, 0x0000097c, 0x00000040);
-	nv_icmd(priv, 0x00000979, 0x00000003);
-	nv_icmd(priv, 0x00000975, 0x00000020);
-	nv_icmd(priv, 0x00000976, 0x00000001);
-	nv_icmd(priv, 0x00000977, 0x00000020);
-	nv_icmd(priv, 0x00000978, 0x00000001);
-	nv_icmd(priv, 0x00000957, 0x00000003);
-	nv_icmd(priv, 0x0000095e, 0x20164010);
-	nv_icmd(priv, 0x0000095f, 0x00000020);
-	if (nv_device(priv)->chipset >= 0xd0)
-		nv_icmd(priv, 0x0000097d, 0x00000020);
-	nv_icmd(priv, 0x00000683, 0x00000006);
-	nv_icmd(priv, 0x00000685, 0x003fffff);
-	nv_icmd(priv, 0x00000687, 0x00000c48);
-	nv_icmd(priv, 0x000006a0, 0x00000005);
-	nv_icmd(priv, 0x00000840, 0x00300008);
-	nv_icmd(priv, 0x00000841, 0x04000080);
-	nv_icmd(priv, 0x00000842, 0x00300008);
-	nv_icmd(priv, 0x00000843, 0x04000080);
-	nv_icmd(priv, 0x00000818, 0x00000000);
-	nv_icmd(priv, 0x00000819, 0x00000000);
-	nv_icmd(priv, 0x0000081a, 0x00000000);
-	nv_icmd(priv, 0x0000081b, 0x00000000);
-	nv_icmd(priv, 0x0000081c, 0x00000000);
-	nv_icmd(priv, 0x0000081d, 0x00000000);
-	nv_icmd(priv, 0x0000081e, 0x00000000);
-	nv_icmd(priv, 0x0000081f, 0x00000000);
-	nv_icmd(priv, 0x00000848, 0x00000000);
-	nv_icmd(priv, 0x00000849, 0x00000000);
-	nv_icmd(priv, 0x0000084a, 0x00000000);
-	nv_icmd(priv, 0x0000084b, 0x00000000);
-	nv_icmd(priv, 0x0000084c, 0x00000000);
-	nv_icmd(priv, 0x0000084d, 0x00000000);
-	nv_icmd(priv, 0x0000084e, 0x00000000);
-	nv_icmd(priv, 0x0000084f, 0x00000000);
-	nv_icmd(priv, 0x00000850, 0x00000000);
-	nv_icmd(priv, 0x00000851, 0x00000000);
-	nv_icmd(priv, 0x00000852, 0x00000000);
-	nv_icmd(priv, 0x00000853, 0x00000000);
-	nv_icmd(priv, 0x00000854, 0x00000000);
-	nv_icmd(priv, 0x00000855, 0x00000000);
-	nv_icmd(priv, 0x00000856, 0x00000000);
-	nv_icmd(priv, 0x00000857, 0x00000000);
-	nv_icmd(priv, 0x00000738, 0x00000000);
-	nv_icmd(priv, 0x000006aa, 0x00000001);
-	nv_icmd(priv, 0x000006ab, 0x00000002);
-	nv_icmd(priv, 0x000006ac, 0x00000080);
-	nv_icmd(priv, 0x000006ad, 0x00000100);
-	nv_icmd(priv, 0x000006ae, 0x00000100);
-	nv_icmd(priv, 0x000006b1, 0x00000011);
-	nv_icmd(priv, 0x000006bb, 0x000000cf);
-	nv_icmd(priv, 0x000006ce, 0x2a712488);
-	nv_icmd(priv, 0x00000739, 0x4085c000);
-	nv_icmd(priv, 0x0000073a, 0x00000080);
-	nv_icmd(priv, 0x00000786, 0x80000100);
-	nv_icmd(priv, 0x0000073c, 0x00010100);
-	nv_icmd(priv, 0x0000073d, 0x02800000);
-	nv_icmd(priv, 0x00000787, 0x000000cf);
-	nv_icmd(priv, 0x0000078c, 0x00000008);
-	nv_icmd(priv, 0x00000792, 0x00000001);
-	nv_icmd(priv, 0x00000794, 0x00000001);
-	nv_icmd(priv, 0x00000795, 0x00000001);
-	nv_icmd(priv, 0x00000796, 0x00000001);
-	nv_icmd(priv, 0x00000797, 0x000000cf);
-	nv_icmd(priv, 0x00000836, 0x00000001);
-	nv_icmd(priv, 0x0000079a, 0x00000002);
-	nv_icmd(priv, 0x00000833, 0x04444480);
-	nv_icmd(priv, 0x000007a1, 0x00000001);
-	nv_icmd(priv, 0x000007a3, 0x00000001);
-	nv_icmd(priv, 0x000007a4, 0x00000001);
-	nv_icmd(priv, 0x000007a5, 0x00000001);
-	nv_icmd(priv, 0x00000831, 0x00000004);
-	nv_icmd(priv, 0x0000080c, 0x00000002);
-	nv_icmd(priv, 0x0000080d, 0x00000100);
-	nv_icmd(priv, 0x0000080e, 0x00000100);
-	nv_icmd(priv, 0x0000080f, 0x00000001);
-	nv_icmd(priv, 0x00000823, 0x00000002);
-	nv_icmd(priv, 0x00000824, 0x00000100);
-	nv_icmd(priv, 0x00000825, 0x00000100);
-	nv_icmd(priv, 0x00000826, 0x00000001);
-	nv_icmd(priv, 0x0000095d, 0x00000001);
-	nv_icmd(priv, 0x0000082b, 0x00000004);
-	nv_icmd(priv, 0x00000942, 0x00010001);
-	nv_icmd(priv, 0x00000943, 0x00000001);
-	nv_icmd(priv, 0x00000944, 0x00000022);
-	nv_icmd(priv, 0x000007c5, 0x00010001);
-	nv_icmd(priv, 0x00000834, 0x00000001);
-	nv_icmd(priv, 0x000007c7, 0x00000001);
-	nv_icmd(priv, 0x0000c1b0, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b1, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b2, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b3, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b4, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b5, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b6, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b7, 0x0000000f);
-	nv_icmd(priv, 0x0000c1b8, 0x0fac6881);
-	nv_icmd(priv, 0x0000c1b9, 0x00fac688);
-	nv_icmd(priv, 0x0001e100, 0x00000001);
-	nv_icmd(priv, 0x00001000, 0x00000002);
-	nv_icmd(priv, 0x000006aa, 0x00000001);
-	nv_icmd(priv, 0x000006ad, 0x00000100);
-	nv_icmd(priv, 0x000006ae, 0x00000100);
-	nv_icmd(priv, 0x000006b1, 0x00000011);
-	nv_icmd(priv, 0x0000078c, 0x00000008);
-	nv_icmd(priv, 0x00000792, 0x00000001);
-	nv_icmd(priv, 0x00000794, 0x00000001);
-	nv_icmd(priv, 0x00000795, 0x00000001);
-	nv_icmd(priv, 0x00000796, 0x00000001);
-	nv_icmd(priv, 0x00000797, 0x000000cf);
-	nv_icmd(priv, 0x0000079a, 0x00000002);
-	nv_icmd(priv, 0x00000833, 0x04444480);
-	nv_icmd(priv, 0x000007a1, 0x00000001);
-	nv_icmd(priv, 0x000007a3, 0x00000001);
-	nv_icmd(priv, 0x000007a4, 0x00000001);
-	nv_icmd(priv, 0x000007a5, 0x00000001);
-	nv_icmd(priv, 0x00000831, 0x00000004);
-	nv_icmd(priv, 0x0001e100, 0x00000001);
-	nv_icmd(priv, 0x00001000, 0x00000014);
-	nv_icmd(priv, 0x00000351, 0x00000100);
-	nv_icmd(priv, 0x00000957, 0x00000003);
-	nv_icmd(priv, 0x0000095d, 0x00000001);
-	nv_icmd(priv, 0x0000082b, 0x00000004);
-	nv_icmd(priv, 0x00000942, 0x00010001);
-	nv_icmd(priv, 0x00000943, 0x00000001);
-	nv_icmd(priv, 0x000007c5, 0x00010001);
-	nv_icmd(priv, 0x00000834, 0x00000001);
-	nv_icmd(priv, 0x000007c7, 0x00000001);
-	nv_icmd(priv, 0x0001e100, 0x00000001);
-	nv_icmd(priv, 0x00001000, 0x00000001);
-	nv_icmd(priv, 0x0000080c, 0x00000002);
-	nv_icmd(priv, 0x0000080d, 0x00000100);
-	nv_icmd(priv, 0x0000080e, 0x00000100);
-	nv_icmd(priv, 0x0000080f, 0x00000001);
-	nv_icmd(priv, 0x00000823, 0x00000002);
-	nv_icmd(priv, 0x00000824, 0x00000100);
-	nv_icmd(priv, 0x00000825, 0x00000100);
-	nv_icmd(priv, 0x00000826, 0x00000001);
-	nv_icmd(priv, 0x0001e100, 0x00000001);
-	nv_wr32(priv, 0x400208, 0x00000000);
-	nv_wr32(priv, 0x404154, 0x00000400);
-
-	nvc0_grctx_generate_9097(priv);
-	if (fermi >= 0x9197)
-		nvc0_grctx_generate_9197(priv);
-	if (fermi >= 0x9297)
-		nvc0_grctx_generate_9297(priv);
-	nvc0_grctx_generate_902d(priv);
-	nvc0_grctx_generate_9039(priv);
-	nvc0_grctx_generate_90c0(priv);
 
-	nv_wr32(priv, 0x000260, r000260);
-
-	return nvc0_grctx_fini(&info);
+done:
+	nouveau_gpuobj_ref(NULL, &chan);
+	return ret;
 }
+
+struct nvc0_graph_init *
+nvc0_grctx_init_hub[] = {
+	nvc0_grctx_init_base,
+	nvc0_grctx_init_unk40xx,
+	nvc0_grctx_init_unk44xx,
+	nvc0_grctx_init_unk46xx,
+	nvc0_grctx_init_unk47xx,
+	nvc0_grctx_init_unk58xx,
+	nvc0_grctx_init_unk60xx,
+	nvc0_grctx_init_unk64xx,
+	nvc0_grctx_init_unk78xx,
+	nvc0_grctx_init_unk80xx,
+	nvc0_grctx_init_rop,
+	NULL
+};
+
+static struct nvc0_graph_init *
+nvc0_grctx_init_gpc[] = {
+	nvc0_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvc0_grctx_init_tpc,
+	NULL
+};
+
+struct nvc0_graph_init
+nvc0_grctx_init_mthd_magic[] = {
+	{ 0x3410, 1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_mthd
+nvc0_grctx_init_mthd[] = {
+	{ 0x9097, nvc0_grctx_init_9097, },
+	{ 0x902d, nvc0_grctx_init_902d, },
+	{ 0x9039, nvc0_grctx_init_9039, },
+	{ 0x90c0, nvc0_grctx_init_90c0, },
+	{ 0x902d, nvc0_grctx_init_mthd_magic, },
+	{}
+};
+
+struct nouveau_oclass *
+nvc0_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xc0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nvc0_grctx_generate_main,
+	.mods = nvc0_grctx_generate_mods,
+	.unkn = nvc0_grctx_generate_unkn,
+	.hub  = nvc0_grctx_init_hub,
+	.gpc  = nvc0_grctx_init_gpc,
+	.icmd = nvc0_grctx_init_icmd,
+	.mthd = nvc0_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc1.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc1.c
new file mode 100644
index 000000000000..e5be3ee7f172
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc1.c
@@ -0,0 +1,823 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+static struct nvc0_graph_init
+nvc1_grctx_init_icmd[] = {
+	{ 0x001000,   1, 0x01, 0x00000004 },
+	{ 0x0000a9,   1, 0x01, 0x0000ffff },
+	{ 0x000038,   1, 0x01, 0x0fac6881 },
+	{ 0x00003d,   1, 0x01, 0x00000001 },
+	{ 0x0000e8,   8, 0x01, 0x00000400 },
+	{ 0x000078,   8, 0x01, 0x00000300 },
+	{ 0x000050,   1, 0x01, 0x00000011 },
+	{ 0x000058,   8, 0x01, 0x00000008 },
+	{ 0x000208,   8, 0x01, 0x00000001 },
+	{ 0x000081,   1, 0x01, 0x00000001 },
+	{ 0x000085,   1, 0x01, 0x00000004 },
+	{ 0x000088,   1, 0x01, 0x00000400 },
+	{ 0x000090,   1, 0x01, 0x00000300 },
+	{ 0x000098,   1, 0x01, 0x00001001 },
+	{ 0x0000e3,   1, 0x01, 0x00000001 },
+	{ 0x0000da,   1, 0x01, 0x00000001 },
+	{ 0x0000f8,   1, 0x01, 0x00000003 },
+	{ 0x0000fa,   1, 0x01, 0x00000001 },
+	{ 0x00009f,   4, 0x01, 0x0000ffff },
+	{ 0x0000b1,   1, 0x01, 0x00000001 },
+	{ 0x0000b2,  40, 0x01, 0x00000000 },
+	{ 0x000210,   8, 0x01, 0x00000040 },
+	{ 0x000218,   8, 0x01, 0x0000c080 },
+	{ 0x0000ad,   1, 0x01, 0x0000013e },
+	{ 0x0000e1,   1, 0x01, 0x00000010 },
+	{ 0x000290,  16, 0x01, 0x00000000 },
+	{ 0x0003b0,  16, 0x01, 0x00000000 },
+	{ 0x0002a0,  16, 0x01, 0x00000000 },
+	{ 0x000420,  16, 0x01, 0x00000000 },
+	{ 0x0002b0,  16, 0x01, 0x00000000 },
+	{ 0x000430,  16, 0x01, 0x00000000 },
+	{ 0x0002c0,  16, 0x01, 0x00000000 },
+	{ 0x0004d0,  16, 0x01, 0x00000000 },
+	{ 0x000720,  16, 0x01, 0x00000000 },
+	{ 0x0008c0,  16, 0x01, 0x00000000 },
+	{ 0x000890,  16, 0x01, 0x00000000 },
+	{ 0x0008e0,  16, 0x01, 0x00000000 },
+	{ 0x0008a0,  16, 0x01, 0x00000000 },
+	{ 0x0008f0,  16, 0x01, 0x00000000 },
+	{ 0x00094c,   1, 0x01, 0x000000ff },
+	{ 0x00094d,   1, 0x01, 0xffffffff },
+	{ 0x00094e,   1, 0x01, 0x00000002 },
+	{ 0x0002ec,   1, 0x01, 0x00000001 },
+	{ 0x000303,   1, 0x01, 0x00000001 },
+	{ 0x0002e6,   1, 0x01, 0x00000001 },
+	{ 0x000466,   1, 0x01, 0x00000052 },
+	{ 0x000301,   1, 0x01, 0x3f800000 },
+	{ 0x000304,   1, 0x01, 0x30201000 },
+	{ 0x000305,   1, 0x01, 0x70605040 },
+	{ 0x000306,   1, 0x01, 0xb8a89888 },
+	{ 0x000307,   1, 0x01, 0xf8e8d8c8 },
+	{ 0x00030a,   1, 0x01, 0x00ffff00 },
+	{ 0x00030b,   1, 0x01, 0x0000001a },
+	{ 0x00030c,   1, 0x01, 0x00000001 },
+	{ 0x000318,   1, 0x01, 0x00000001 },
+	{ 0x000340,   1, 0x01, 0x00000000 },
+	{ 0x000375,   1, 0x01, 0x00000001 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x00037d,   1, 0x01, 0x00000006 },
+	{ 0x0003a0,   1, 0x01, 0x00000002 },
+	{ 0x0003aa,   1, 0x01, 0x00000001 },
+	{ 0x0003a9,   1, 0x01, 0x00000001 },
+	{ 0x000380,   1, 0x01, 0x00000001 },
+	{ 0x000360,   1, 0x01, 0x00000040 },
+	{ 0x000366,   2, 0x01, 0x00000000 },
+	{ 0x000368,   1, 0x01, 0x00001fff },
+	{ 0x000370,   2, 0x01, 0x00000000 },
+	{ 0x000372,   1, 0x01, 0x003fffff },
+	{ 0x00037a,   1, 0x01, 0x00000012 },
+	{ 0x0005e0,   5, 0x01, 0x00000022 },
+	{ 0x000619,   1, 0x01, 0x00000003 },
+	{ 0x000811,   1, 0x01, 0x00000003 },
+	{ 0x000812,   1, 0x01, 0x00000004 },
+	{ 0x000813,   1, 0x01, 0x00000006 },
+	{ 0x000814,   1, 0x01, 0x00000008 },
+	{ 0x000815,   1, 0x01, 0x0000000b },
+	{ 0x000800,   6, 0x01, 0x00000001 },
+	{ 0x000632,   1, 0x01, 0x00000001 },
+	{ 0x000633,   1, 0x01, 0x00000002 },
+	{ 0x000634,   1, 0x01, 0x00000003 },
+	{ 0x000635,   1, 0x01, 0x00000004 },
+	{ 0x000654,   1, 0x01, 0x3f800000 },
+	{ 0x000657,   1, 0x01, 0x3f800000 },
+	{ 0x000655,   2, 0x01, 0x3f800000 },
+	{ 0x0006cd,   1, 0x01, 0x3f800000 },
+	{ 0x0007f5,   1, 0x01, 0x3f800000 },
+	{ 0x0007dc,   1, 0x01, 0x39291909 },
+	{ 0x0007dd,   1, 0x01, 0x79695949 },
+	{ 0x0007de,   1, 0x01, 0xb9a99989 },
+	{ 0x0007df,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007e8,   1, 0x01, 0x00003210 },
+	{ 0x0007e9,   1, 0x01, 0x00007654 },
+	{ 0x0007ea,   1, 0x01, 0x00000098 },
+	{ 0x0007ec,   1, 0x01, 0x39291909 },
+	{ 0x0007ed,   1, 0x01, 0x79695949 },
+	{ 0x0007ee,   1, 0x01, 0xb9a99989 },
+	{ 0x0007ef,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007f0,   1, 0x01, 0x00003210 },
+	{ 0x0007f1,   1, 0x01, 0x00007654 },
+	{ 0x0007f2,   1, 0x01, 0x00000098 },
+	{ 0x0005a5,   1, 0x01, 0x00000001 },
+	{ 0x000980, 128, 0x01, 0x00000000 },
+	{ 0x000468,   1, 0x01, 0x00000004 },
+	{ 0x00046c,   1, 0x01, 0x00000001 },
+	{ 0x000470,  96, 0x01, 0x00000000 },
+	{ 0x000510,  16, 0x01, 0x3f800000 },
+	{ 0x000520,   1, 0x01, 0x000002b6 },
+	{ 0x000529,   1, 0x01, 0x00000001 },
+	{ 0x000530,  16, 0x01, 0xffff0000 },
+	{ 0x000585,   1, 0x01, 0x0000003f },
+	{ 0x000576,   1, 0x01, 0x00000003 },
+	{ 0x00057b,   1, 0x01, 0x00000059 },
+	{ 0x000586,   1, 0x01, 0x00000040 },
+	{ 0x000582,   2, 0x01, 0x00000080 },
+	{ 0x0005c2,   1, 0x01, 0x00000001 },
+	{ 0x000638,   1, 0x01, 0x00000001 },
+	{ 0x000639,   1, 0x01, 0x00000001 },
+	{ 0x00063a,   1, 0x01, 0x00000002 },
+	{ 0x00063b,   2, 0x01, 0x00000001 },
+	{ 0x00063d,   1, 0x01, 0x00000002 },
+	{ 0x00063e,   1, 0x01, 0x00000001 },
+	{ 0x0008b8,   8, 0x01, 0x00000001 },
+	{ 0x000900,   8, 0x01, 0x00000001 },
+	{ 0x000908,   8, 0x01, 0x00000002 },
+	{ 0x000910,  16, 0x01, 0x00000001 },
+	{ 0x000920,   8, 0x01, 0x00000002 },
+	{ 0x000928,   8, 0x01, 0x00000001 },
+	{ 0x000648,   9, 0x01, 0x00000001 },
+	{ 0x000658,   1, 0x01, 0x0000000f },
+	{ 0x0007ff,   1, 0x01, 0x0000000a },
+	{ 0x00066a,   1, 0x01, 0x40000000 },
+	{ 0x00066b,   1, 0x01, 0x10000000 },
+	{ 0x00066c,   2, 0x01, 0xffff0000 },
+	{ 0x0007af,   2, 0x01, 0x00000008 },
+	{ 0x0007f6,   1, 0x01, 0x00000001 },
+	{ 0x0006b2,   1, 0x01, 0x00000055 },
+	{ 0x0007ad,   1, 0x01, 0x00000003 },
+	{ 0x000937,   1, 0x01, 0x00000001 },
+	{ 0x000971,   1, 0x01, 0x00000008 },
+	{ 0x000972,   1, 0x01, 0x00000040 },
+	{ 0x000973,   1, 0x01, 0x0000012c },
+	{ 0x00097c,   1, 0x01, 0x00000040 },
+	{ 0x000979,   1, 0x01, 0x00000003 },
+	{ 0x000975,   1, 0x01, 0x00000020 },
+	{ 0x000976,   1, 0x01, 0x00000001 },
+	{ 0x000977,   1, 0x01, 0x00000020 },
+	{ 0x000978,   1, 0x01, 0x00000001 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095e,   1, 0x01, 0x20164010 },
+	{ 0x00095f,   1, 0x01, 0x00000020 },
+	{ 0x000683,   1, 0x01, 0x00000006 },
+	{ 0x000685,   1, 0x01, 0x003fffff },
+	{ 0x000687,   1, 0x01, 0x00000c48 },
+	{ 0x0006a0,   1, 0x01, 0x00000005 },
+	{ 0x000840,   1, 0x01, 0x00300008 },
+	{ 0x000841,   1, 0x01, 0x04000080 },
+	{ 0x000842,   1, 0x01, 0x00300008 },
+	{ 0x000843,   1, 0x01, 0x04000080 },
+	{ 0x000818,   8, 0x01, 0x00000000 },
+	{ 0x000848,  16, 0x01, 0x00000000 },
+	{ 0x000738,   1, 0x01, 0x00000000 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ab,   1, 0x01, 0x00000002 },
+	{ 0x0006ac,   1, 0x01, 0x00000080 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x0006bb,   1, 0x01, 0x000000cf },
+	{ 0x0006ce,   1, 0x01, 0x2a712488 },
+	{ 0x000739,   1, 0x01, 0x4085c000 },
+	{ 0x00073a,   1, 0x01, 0x00000080 },
+	{ 0x000786,   1, 0x01, 0x80000100 },
+	{ 0x00073c,   1, 0x01, 0x00010100 },
+	{ 0x00073d,   1, 0x01, 0x02800000 },
+	{ 0x000787,   1, 0x01, 0x000000cf },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x000836,   1, 0x01, 0x00000001 },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x000944,   1, 0x01, 0x00000022 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x00c1b0,   8, 0x01, 0x0000000f },
+	{ 0x00c1b8,   1, 0x01, 0x0fac6881 },
+	{ 0x00c1b9,   1, 0x01, 0x00fac688 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000002 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000014 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000001 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc1_grctx_init_9097[] = {
+	{ 0x000800,   8, 0x40, 0x00000000 },
+	{ 0x000804,   8, 0x40, 0x00000000 },
+	{ 0x000808,   8, 0x40, 0x00000400 },
+	{ 0x00080c,   8, 0x40, 0x00000300 },
+	{ 0x000810,   1, 0x04, 0x000000cf },
+	{ 0x000850,   7, 0x40, 0x00000000 },
+	{ 0x000814,   8, 0x40, 0x00000040 },
+	{ 0x000818,   8, 0x40, 0x00000001 },
+	{ 0x00081c,   8, 0x40, 0x00000000 },
+	{ 0x000820,   8, 0x40, 0x00000000 },
+	{ 0x002700,   8, 0x20, 0x00000000 },
+	{ 0x002704,   8, 0x20, 0x00000000 },
+	{ 0x002708,   8, 0x20, 0x00000000 },
+	{ 0x00270c,   8, 0x20, 0x00000000 },
+	{ 0x002710,   8, 0x20, 0x00014000 },
+	{ 0x002714,   8, 0x20, 0x00000040 },
+	{ 0x001c00,  16, 0x10, 0x00000000 },
+	{ 0x001c04,  16, 0x10, 0x00000000 },
+	{ 0x001c08,  16, 0x10, 0x00000000 },
+	{ 0x001c0c,  16, 0x10, 0x00000000 },
+	{ 0x001d00,  16, 0x10, 0x00000000 },
+	{ 0x001d04,  16, 0x10, 0x00000000 },
+	{ 0x001d08,  16, 0x10, 0x00000000 },
+	{ 0x001d0c,  16, 0x10, 0x00000000 },
+	{ 0x001f00,  16, 0x08, 0x00000000 },
+	{ 0x001f04,  16, 0x08, 0x00000000 },
+	{ 0x001f80,  16, 0x08, 0x00000000 },
+	{ 0x001f84,  16, 0x08, 0x00000000 },
+	{ 0x002200,   5, 0x10, 0x00000022 },
+	{ 0x002000,   1, 0x04, 0x00000000 },
+	{ 0x002040,   1, 0x04, 0x00000011 },
+	{ 0x002080,   1, 0x04, 0x00000020 },
+	{ 0x0020c0,   1, 0x04, 0x00000030 },
+	{ 0x002100,   1, 0x04, 0x00000040 },
+	{ 0x002140,   1, 0x04, 0x00000051 },
+	{ 0x00200c,   6, 0x40, 0x00000001 },
+	{ 0x002010,   1, 0x04, 0x00000000 },
+	{ 0x002050,   1, 0x04, 0x00000000 },
+	{ 0x002090,   1, 0x04, 0x00000001 },
+	{ 0x0020d0,   1, 0x04, 0x00000002 },
+	{ 0x002110,   1, 0x04, 0x00000003 },
+	{ 0x002150,   1, 0x04, 0x00000004 },
+	{ 0x000380,   4, 0x20, 0x00000000 },
+	{ 0x000384,   4, 0x20, 0x00000000 },
+	{ 0x000388,   4, 0x20, 0x00000000 },
+	{ 0x00038c,   4, 0x20, 0x00000000 },
+	{ 0x000700,   4, 0x10, 0x00000000 },
+	{ 0x000704,   4, 0x10, 0x00000000 },
+	{ 0x000708,   4, 0x10, 0x00000000 },
+	{ 0x002800, 128, 0x04, 0x00000000 },
+	{ 0x000a00,  16, 0x20, 0x00000000 },
+	{ 0x000a04,  16, 0x20, 0x00000000 },
+	{ 0x000a08,  16, 0x20, 0x00000000 },
+	{ 0x000a0c,  16, 0x20, 0x00000000 },
+	{ 0x000a10,  16, 0x20, 0x00000000 },
+	{ 0x000a14,  16, 0x20, 0x00000000 },
+	{ 0x000c00,  16, 0x10, 0x00000000 },
+	{ 0x000c04,  16, 0x10, 0x00000000 },
+	{ 0x000c08,  16, 0x10, 0x00000000 },
+	{ 0x000c0c,  16, 0x10, 0x3f800000 },
+	{ 0x000d00,   8, 0x08, 0xffff0000 },
+	{ 0x000d04,   8, 0x08, 0xffff0000 },
+	{ 0x000e00,  16, 0x10, 0x00000000 },
+	{ 0x000e04,  16, 0x10, 0xffff0000 },
+	{ 0x000e08,  16, 0x10, 0xffff0000 },
+	{ 0x000d40,   4, 0x08, 0x00000000 },
+	{ 0x000d44,   4, 0x08, 0x00000000 },
+	{ 0x001e00,   8, 0x20, 0x00000001 },
+	{ 0x001e04,   8, 0x20, 0x00000001 },
+	{ 0x001e08,   8, 0x20, 0x00000002 },
+	{ 0x001e0c,   8, 0x20, 0x00000001 },
+	{ 0x001e10,   8, 0x20, 0x00000001 },
+	{ 0x001e14,   8, 0x20, 0x00000002 },
+	{ 0x001e18,   8, 0x20, 0x00000001 },
+	{ 0x00030c,   1, 0x04, 0x00000001 },
+	{ 0x001944,   1, 0x04, 0x00000000 },
+	{ 0x001514,   1, 0x04, 0x00000000 },
+	{ 0x000d68,   1, 0x04, 0x0000ffff },
+	{ 0x00121c,   1, 0x04, 0x0fac6881 },
+	{ 0x000fac,   1, 0x04, 0x00000001 },
+	{ 0x001538,   1, 0x04, 0x00000001 },
+	{ 0x000fe0,   2, 0x04, 0x00000000 },
+	{ 0x000fe8,   1, 0x04, 0x00000014 },
+	{ 0x000fec,   1, 0x04, 0x00000040 },
+	{ 0x000ff0,   1, 0x04, 0x00000000 },
+	{ 0x00179c,   1, 0x04, 0x00000000 },
+	{ 0x001228,   1, 0x04, 0x00000400 },
+	{ 0x00122c,   1, 0x04, 0x00000300 },
+	{ 0x001230,   1, 0x04, 0x00010001 },
+	{ 0x0007f8,   1, 0x04, 0x00000000 },
+	{ 0x0015b4,   1, 0x04, 0x00000001 },
+	{ 0x0015cc,   1, 0x04, 0x00000000 },
+	{ 0x001534,   1, 0x04, 0x00000000 },
+	{ 0x000fb0,   1, 0x04, 0x00000000 },
+	{ 0x0015d0,   1, 0x04, 0x00000000 },
+	{ 0x00153c,   1, 0x04, 0x00000000 },
+	{ 0x0016b4,   1, 0x04, 0x00000003 },
+	{ 0x000fbc,   4, 0x04, 0x0000ffff },
+	{ 0x000df8,   2, 0x04, 0x00000000 },
+	{ 0x001948,   1, 0x04, 0x00000000 },
+	{ 0x001970,   1, 0x04, 0x00000001 },
+	{ 0x00161c,   1, 0x04, 0x000009f0 },
+	{ 0x000dcc,   1, 0x04, 0x00000010 },
+	{ 0x00163c,   1, 0x04, 0x00000000 },
+	{ 0x0015e4,   1, 0x04, 0x00000000 },
+	{ 0x001160,  32, 0x04, 0x25e00040 },
+	{ 0x001880,  32, 0x04, 0x00000000 },
+	{ 0x000f84,   2, 0x04, 0x00000000 },
+	{ 0x0017c8,   2, 0x04, 0x00000000 },
+	{ 0x0017d0,   1, 0x04, 0x000000ff },
+	{ 0x0017d4,   1, 0x04, 0xffffffff },
+	{ 0x0017d8,   1, 0x04, 0x00000002 },
+	{ 0x0017dc,   1, 0x04, 0x00000000 },
+	{ 0x0015f4,   2, 0x04, 0x00000000 },
+	{ 0x001434,   2, 0x04, 0x00000000 },
+	{ 0x000d74,   1, 0x04, 0x00000000 },
+	{ 0x000dec,   1, 0x04, 0x00000001 },
+	{ 0x0013a4,   1, 0x04, 0x00000000 },
+	{ 0x001318,   1, 0x04, 0x00000001 },
+	{ 0x001644,   1, 0x04, 0x00000000 },
+	{ 0x000748,   1, 0x04, 0x00000000 },
+	{ 0x000de8,   1, 0x04, 0x00000000 },
+	{ 0x001648,   1, 0x04, 0x00000000 },
+	{ 0x0012a4,   1, 0x04, 0x00000000 },
+	{ 0x001120,   4, 0x04, 0x00000000 },
+	{ 0x001118,   1, 0x04, 0x00000000 },
+	{ 0x00164c,   1, 0x04, 0x00000000 },
+	{ 0x001658,   1, 0x04, 0x00000000 },
+	{ 0x001910,   1, 0x04, 0x00000290 },
+	{ 0x001518,   1, 0x04, 0x00000000 },
+	{ 0x00165c,   1, 0x04, 0x00000001 },
+	{ 0x001520,   1, 0x04, 0x00000000 },
+	{ 0x001604,   1, 0x04, 0x00000000 },
+	{ 0x001570,   1, 0x04, 0x00000000 },
+	{ 0x0013b0,   2, 0x04, 0x3f800000 },
+	{ 0x00020c,   1, 0x04, 0x00000000 },
+	{ 0x001670,   1, 0x04, 0x30201000 },
+	{ 0x001674,   1, 0x04, 0x70605040 },
+	{ 0x001678,   1, 0x04, 0xb8a89888 },
+	{ 0x00167c,   1, 0x04, 0xf8e8d8c8 },
+	{ 0x00166c,   1, 0x04, 0x00000000 },
+	{ 0x001680,   1, 0x04, 0x00ffff00 },
+	{ 0x0012d0,   1, 0x04, 0x00000003 },
+	{ 0x0012d4,   1, 0x04, 0x00000002 },
+	{ 0x001684,   2, 0x04, 0x00000000 },
+	{ 0x000dac,   2, 0x04, 0x00001b02 },
+	{ 0x000db4,   1, 0x04, 0x00000000 },
+	{ 0x00168c,   1, 0x04, 0x00000000 },
+	{ 0x0015bc,   1, 0x04, 0x00000000 },
+	{ 0x00156c,   1, 0x04, 0x00000000 },
+	{ 0x00187c,   1, 0x04, 0x00000000 },
+	{ 0x001110,   1, 0x04, 0x00000001 },
+	{ 0x000dc0,   3, 0x04, 0x00000000 },
+	{ 0x001234,   1, 0x04, 0x00000000 },
+	{ 0x001690,   1, 0x04, 0x00000000 },
+	{ 0x0012ac,   1, 0x04, 0x00000001 },
+	{ 0x0002c4,   1, 0x04, 0x00000000 },
+	{ 0x000790,   5, 0x04, 0x00000000 },
+	{ 0x00077c,   1, 0x04, 0x00000000 },
+	{ 0x001000,   1, 0x04, 0x00000010 },
+	{ 0x0010fc,   1, 0x04, 0x00000000 },
+	{ 0x001290,   1, 0x04, 0x00000000 },
+	{ 0x000218,   1, 0x04, 0x00000010 },
+	{ 0x0012d8,   1, 0x04, 0x00000000 },
+	{ 0x0012dc,   1, 0x04, 0x00000010 },
+	{ 0x000d94,   1, 0x04, 0x00000001 },
+	{ 0x00155c,   2, 0x04, 0x00000000 },
+	{ 0x001564,   1, 0x04, 0x00001fff },
+	{ 0x001574,   2, 0x04, 0x00000000 },
+	{ 0x00157c,   1, 0x04, 0x003fffff },
+	{ 0x001354,   1, 0x04, 0x00000000 },
+	{ 0x001664,   1, 0x04, 0x00000000 },
+	{ 0x001610,   1, 0x04, 0x00000012 },
+	{ 0x001608,   2, 0x04, 0x00000000 },
+	{ 0x00162c,   1, 0x04, 0x00000003 },
+	{ 0x000210,   1, 0x04, 0x00000000 },
+	{ 0x000320,   1, 0x04, 0x00000000 },
+	{ 0x000324,   6, 0x04, 0x3f800000 },
+	{ 0x000750,   1, 0x04, 0x00000000 },
+	{ 0x000760,   1, 0x04, 0x39291909 },
+	{ 0x000764,   1, 0x04, 0x79695949 },
+	{ 0x000768,   1, 0x04, 0xb9a99989 },
+	{ 0x00076c,   1, 0x04, 0xf9e9d9c9 },
+	{ 0x000770,   1, 0x04, 0x30201000 },
+	{ 0x000774,   1, 0x04, 0x70605040 },
+	{ 0x000778,   1, 0x04, 0x00009080 },
+	{ 0x000780,   1, 0x04, 0x39291909 },
+	{ 0x000784,   1, 0x04, 0x79695949 },
+	{ 0x000788,   1, 0x04, 0xb9a99989 },
+	{ 0x00078c,   1, 0x04, 0xf9e9d9c9 },
+	{ 0x0007d0,   1, 0x04, 0x30201000 },
+	{ 0x0007d4,   1, 0x04, 0x70605040 },
+	{ 0x0007d8,   1, 0x04, 0x00009080 },
+	{ 0x00037c,   1, 0x04, 0x00000001 },
+	{ 0x000740,   2, 0x04, 0x00000000 },
+	{ 0x002600,   1, 0x04, 0x00000000 },
+	{ 0x001918,   1, 0x04, 0x00000000 },
+	{ 0x00191c,   1, 0x04, 0x00000900 },
+	{ 0x001920,   1, 0x04, 0x00000405 },
+	{ 0x001308,   1, 0x04, 0x00000001 },
+	{ 0x001924,   1, 0x04, 0x00000000 },
+	{ 0x0013ac,   1, 0x04, 0x00000000 },
+	{ 0x00192c,   1, 0x04, 0x00000001 },
+	{ 0x00193c,   1, 0x04, 0x00002c1c },
+	{ 0x000d7c,   1, 0x04, 0x00000000 },
+	{ 0x000f8c,   1, 0x04, 0x00000000 },
+	{ 0x0002c0,   1, 0x04, 0x00000001 },
+	{ 0x001510,   1, 0x04, 0x00000000 },
+	{ 0x001940,   1, 0x04, 0x00000000 },
+	{ 0x000ff4,   2, 0x04, 0x00000000 },
+	{ 0x00194c,   2, 0x04, 0x00000000 },
+	{ 0x001968,   1, 0x04, 0x00000000 },
+	{ 0x001590,   1, 0x04, 0x0000003f },
+	{ 0x0007e8,   4, 0x04, 0x00000000 },
+	{ 0x00196c,   1, 0x04, 0x00000011 },
+	{ 0x00197c,   1, 0x04, 0x00000000 },
+	{ 0x000fcc,   2, 0x04, 0x00000000 },
+	{ 0x0002d8,   1, 0x04, 0x00000040 },
+	{ 0x001980,   1, 0x04, 0x00000080 },
+	{ 0x001504,   1, 0x04, 0x00000080 },
+	{ 0x001984,   1, 0x04, 0x00000000 },
+	{ 0x000300,   1, 0x04, 0x00000001 },
+	{ 0x0013a8,   1, 0x04, 0x00000000 },
+	{ 0x0012ec,   1, 0x04, 0x00000000 },
+	{ 0x001310,   1, 0x04, 0x00000000 },
+	{ 0x001314,   1, 0x04, 0x00000001 },
+	{ 0x001380,   1, 0x04, 0x00000000 },
+	{ 0x001384,   4, 0x04, 0x00000001 },
+	{ 0x001394,   1, 0x04, 0x00000000 },
+	{ 0x00139c,   1, 0x04, 0x00000000 },
+	{ 0x001398,   1, 0x04, 0x00000000 },
+	{ 0x001594,   1, 0x04, 0x00000000 },
+	{ 0x001598,   4, 0x04, 0x00000001 },
+	{ 0x000f54,   3, 0x04, 0x00000000 },
+	{ 0x0019bc,   1, 0x04, 0x00000000 },
+	{ 0x000f9c,   2, 0x04, 0x00000000 },
+	{ 0x0012cc,   1, 0x04, 0x00000000 },
+	{ 0x0012e8,   1, 0x04, 0x00000000 },
+	{ 0x00130c,   1, 0x04, 0x00000001 },
+	{ 0x001360,   8, 0x04, 0x00000000 },
+	{ 0x00133c,   2, 0x04, 0x00000001 },
+	{ 0x001344,   1, 0x04, 0x00000002 },
+	{ 0x001348,   2, 0x04, 0x00000001 },
+	{ 0x001350,   1, 0x04, 0x00000002 },
+	{ 0x001358,   1, 0x04, 0x00000001 },
+	{ 0x0012e4,   1, 0x04, 0x00000000 },
+	{ 0x00131c,   1, 0x04, 0x00000000 },
+	{ 0x001320,   3, 0x04, 0x00000000 },
+	{ 0x0019c0,   1, 0x04, 0x00000000 },
+	{ 0x001140,   1, 0x04, 0x00000000 },
+	{ 0x0019c4,   1, 0x04, 0x00000000 },
+	{ 0x0019c8,   1, 0x04, 0x00001500 },
+	{ 0x00135c,   1, 0x04, 0x00000000 },
+	{ 0x000f90,   1, 0x04, 0x00000000 },
+	{ 0x0019e0,   8, 0x04, 0x00000001 },
+	{ 0x0019cc,   1, 0x04, 0x00000001 },
+	{ 0x0015b8,   1, 0x04, 0x00000000 },
+	{ 0x001a00,   1, 0x04, 0x00001111 },
+	{ 0x001a04,   7, 0x04, 0x00000000 },
+	{ 0x000d6c,   2, 0x04, 0xffff0000 },
+	{ 0x0010f8,   1, 0x04, 0x00001010 },
+	{ 0x000d80,   5, 0x04, 0x00000000 },
+	{ 0x000da0,   1, 0x04, 0x00000000 },
+	{ 0x001508,   1, 0x04, 0x80000000 },
+	{ 0x00150c,   1, 0x04, 0x40000000 },
+	{ 0x001668,   1, 0x04, 0x00000000 },
+	{ 0x000318,   2, 0x04, 0x00000008 },
+	{ 0x000d9c,   1, 0x04, 0x00000001 },
+	{ 0x0007dc,   1, 0x04, 0x00000000 },
+	{ 0x00074c,   1, 0x04, 0x00000055 },
+	{ 0x001420,   1, 0x04, 0x00000003 },
+	{ 0x0017bc,   2, 0x04, 0x00000000 },
+	{ 0x0017c4,   1, 0x04, 0x00000001 },
+	{ 0x001008,   1, 0x04, 0x00000008 },
+	{ 0x00100c,   1, 0x04, 0x00000040 },
+	{ 0x001010,   1, 0x04, 0x0000012c },
+	{ 0x000d60,   1, 0x04, 0x00000040 },
+	{ 0x00075c,   1, 0x04, 0x00000003 },
+	{ 0x001018,   1, 0x04, 0x00000020 },
+	{ 0x00101c,   1, 0x04, 0x00000001 },
+	{ 0x001020,   1, 0x04, 0x00000020 },
+	{ 0x001024,   1, 0x04, 0x00000001 },
+	{ 0x001444,   3, 0x04, 0x00000000 },
+	{ 0x000360,   1, 0x04, 0x20164010 },
+	{ 0x000364,   1, 0x04, 0x00000020 },
+	{ 0x000368,   1, 0x04, 0x00000000 },
+	{ 0x000de4,   1, 0x04, 0x00000000 },
+	{ 0x000204,   1, 0x04, 0x00000006 },
+	{ 0x000208,   1, 0x04, 0x00000000 },
+	{ 0x0002cc,   1, 0x04, 0x003fffff },
+	{ 0x0002d0,   1, 0x04, 0x00000c48 },
+	{ 0x001220,   1, 0x04, 0x00000005 },
+	{ 0x000fdc,   1, 0x04, 0x00000000 },
+	{ 0x000f98,   1, 0x04, 0x00300008 },
+	{ 0x001284,   1, 0x04, 0x04000080 },
+	{ 0x001450,   1, 0x04, 0x00300008 },
+	{ 0x001454,   1, 0x04, 0x04000080 },
+	{ 0x000214,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc1_grctx_init_9197[] = {
+	{ 0x003400, 128, 0x04, 0x00000000 },
+	{ 0x0002e4,   1, 0x04, 0x0000b001 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc1_grctx_init_unk58xx[] = {
+	{ 0x405800,   1, 0x04, 0x0f8000bf },
+	{ 0x405830,   1, 0x04, 0x02180218 },
+	{ 0x405834,   2, 0x04, 0x00000000 },
+	{ 0x405854,   1, 0x04, 0x00000000 },
+	{ 0x405870,   4, 0x04, 0x00000001 },
+	{ 0x405a00,   2, 0x04, 0x00000000 },
+	{ 0x405a18,   1, 0x04, 0x00000000 },
+};
+
+static struct nvc0_graph_init
+nvc1_grctx_init_rop[] = {
+	{ 0x408800,   1, 0x04, 0x02802a3c },
+	{ 0x408804,   1, 0x04, 0x00000040 },
+	{ 0x408808,   1, 0x04, 0x1003e005 },
+	{ 0x408900,   1, 0x04, 0x3080b801 },
+	{ 0x408904,   1, 0x04, 0x62000001 },
+	{ 0x408908,   1, 0x04, 0x00c80929 },
+	{ 0x408980,   1, 0x04, 0x0000011d },
+};
+
+static struct nvc0_graph_init
+nvc1_grctx_init_gpc_0[] = {
+	{ 0x418380,   1, 0x04, 0x00000016 },
+	{ 0x418400,   1, 0x04, 0x38004e00 },
+	{ 0x418404,   1, 0x04, 0x71e0ffff },
+	{ 0x418408,   1, 0x04, 0x00000000 },
+	{ 0x41840c,   1, 0x04, 0x00001008 },
+	{ 0x418410,   1, 0x04, 0x0fff0fff },
+	{ 0x418414,   1, 0x04, 0x00200fff },
+	{ 0x418450,   6, 0x04, 0x00000000 },
+	{ 0x418468,   1, 0x04, 0x00000001 },
+	{ 0x41846c,   2, 0x04, 0x00000000 },
+	{ 0x418600,   1, 0x04, 0x0000001f },
+	{ 0x418684,   1, 0x04, 0x0000000f },
+	{ 0x418700,   1, 0x04, 0x00000002 },
+	{ 0x418704,   1, 0x04, 0x00000080 },
+	{ 0x418708,   1, 0x04, 0x00000000 },
+	{ 0x41870c,   1, 0x04, 0x07c80000 },
+	{ 0x418710,   1, 0x04, 0x00000000 },
+	{ 0x418800,   1, 0x04, 0x0006860a },
+	{ 0x418808,   3, 0x04, 0x00000000 },
+	{ 0x418828,   1, 0x04, 0x00008442 },
+	{ 0x418830,   1, 0x04, 0x10000001 },
+	{ 0x4188d8,   1, 0x04, 0x00000008 },
+	{ 0x4188e0,   1, 0x04, 0x01000000 },
+	{ 0x4188e8,   5, 0x04, 0x00000000 },
+	{ 0x4188fc,   1, 0x04, 0x00100018 },
+	{ 0x41891c,   1, 0x04, 0x00ff00ff },
+	{ 0x418924,   1, 0x04, 0x00000000 },
+	{ 0x418928,   1, 0x04, 0x00ffff00 },
+	{ 0x41892c,   1, 0x04, 0x0000ff00 },
+	{ 0x418a00,   3, 0x04, 0x00000000 },
+	{ 0x418a0c,   1, 0x04, 0x00010000 },
+	{ 0x418a10,   3, 0x04, 0x00000000 },
+	{ 0x418a20,   3, 0x04, 0x00000000 },
+	{ 0x418a2c,   1, 0x04, 0x00010000 },
+	{ 0x418a30,   3, 0x04, 0x00000000 },
+	{ 0x418a40,   3, 0x04, 0x00000000 },
+	{ 0x418a4c,   1, 0x04, 0x00010000 },
+	{ 0x418a50,   3, 0x04, 0x00000000 },
+	{ 0x418a60,   3, 0x04, 0x00000000 },
+	{ 0x418a6c,   1, 0x04, 0x00010000 },
+	{ 0x418a70,   3, 0x04, 0x00000000 },
+	{ 0x418a80,   3, 0x04, 0x00000000 },
+	{ 0x418a8c,   1, 0x04, 0x00010000 },
+	{ 0x418a90,   3, 0x04, 0x00000000 },
+	{ 0x418aa0,   3, 0x04, 0x00000000 },
+	{ 0x418aac,   1, 0x04, 0x00010000 },
+	{ 0x418ab0,   3, 0x04, 0x00000000 },
+	{ 0x418ac0,   3, 0x04, 0x00000000 },
+	{ 0x418acc,   1, 0x04, 0x00010000 },
+	{ 0x418ad0,   3, 0x04, 0x00000000 },
+	{ 0x418ae0,   3, 0x04, 0x00000000 },
+	{ 0x418aec,   1, 0x04, 0x00010000 },
+	{ 0x418af0,   3, 0x04, 0x00000000 },
+	{ 0x418b00,   1, 0x04, 0x00000000 },
+	{ 0x418b08,   1, 0x04, 0x0a418820 },
+	{ 0x418b0c,   1, 0x04, 0x062080e6 },
+	{ 0x418b10,   1, 0x04, 0x020398a4 },
+	{ 0x418b14,   1, 0x04, 0x0e629062 },
+	{ 0x418b18,   1, 0x04, 0x0a418820 },
+	{ 0x418b1c,   1, 0x04, 0x000000e6 },
+	{ 0x418bb8,   1, 0x04, 0x00000103 },
+	{ 0x418c08,   1, 0x04, 0x00000001 },
+	{ 0x418c10,   8, 0x04, 0x00000000 },
+	{ 0x418c6c,   1, 0x04, 0x00000001 },
+	{ 0x418c80,   1, 0x04, 0x20200004 },
+	{ 0x418c8c,   1, 0x04, 0x00000001 },
+	{ 0x419000,   1, 0x04, 0x00000780 },
+	{ 0x419004,   2, 0x04, 0x00000000 },
+	{ 0x419014,   1, 0x04, 0x00000004 },
+};
+
+static struct nvc0_graph_init
+nvc1_grctx_init_tpc[] = {
+	{ 0x419818,   1, 0x04, 0x00000000 },
+	{ 0x41983c,   1, 0x04, 0x00038bc7 },
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x00000129 },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000001f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000023 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x00000000 },
+	{ 0x419a20,   1, 0x04, 0x00000800 },
+	{ 0x419ac4,   1, 0x04, 0x0007f440 },
+	{ 0x419b00,   1, 0x04, 0x0a418820 },
+	{ 0x419b04,   1, 0x04, 0x062080e6 },
+	{ 0x419b08,   1, 0x04, 0x020398a4 },
+	{ 0x419b0c,   1, 0x04, 0x0e629062 },
+	{ 0x419b10,   1, 0x04, 0x0a418820 },
+	{ 0x419b14,   1, 0x04, 0x000000e6 },
+	{ 0x419bd0,   1, 0x04, 0x00900103 },
+	{ 0x419be0,   1, 0x04, 0x00400001 },
+	{ 0x419be4,   1, 0x04, 0x00000000 },
+	{ 0x419c00,   1, 0x04, 0x00000002 },
+	{ 0x419c04,   1, 0x04, 0x00000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x00020048 },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000183 },
+	{ 0x419d20,   1, 0x04, 0x12180000 },
+	{ 0x419d24,   1, 0x04, 0x00001fff },
+	{ 0x419d44,   1, 0x04, 0x02180218 },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000002 },
+	{ 0x419e44,   1, 0x04, 0x001beff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000000f },
+	{ 0x419e50,  17, 0x04, 0x00000000 },
+	{ 0x419e98,   1, 0x04, 0x00000000 },
+	{ 0x419ee0,   1, 0x04, 0x00011110 },
+	{ 0x419f30,  11, 0x04, 0x00000000 },
+};
+
+void
+nvc1_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+{
+	int gpc, tpc;
+	u32 offset;
+
+	mmio_data(0x002000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
+	mmio_list(0x408004, 0x00000000,  8, 0);
+	mmio_list(0x408008, 0x80000018,  0, 0);
+	mmio_list(0x40800c, 0x00000000,  8, 1);
+	mmio_list(0x408010, 0x80000000,  0, 0);
+	mmio_list(0x418810, 0x80000000, 12, 2);
+	mmio_list(0x419848, 0x10000000, 12, 2);
+	mmio_list(0x419004, 0x00000000,  8, 1);
+	mmio_list(0x419008, 0x00000000,  0, 0);
+	mmio_list(0x418808, 0x00000000,  8, 0);
+	mmio_list(0x41880c, 0x80000018,  0, 0);
+
+	mmio_list(0x405830, 0x02180218, 0, 0);
+	mmio_list(0x4064c4, 0x0086ffff, 0, 0);
+
+	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
+		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
+			u32 addr = TPC_UNIT(gpc, tpc, 0x0520);
+			mmio_list(addr, 0x12180000 | offset, 0, 0);
+			offset += 0x0324;
+		}
+		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
+			u32 addr = TPC_UNIT(gpc, tpc, 0x0544);
+			mmio_list(addr, 0x02180000 | offset, 0, 0);
+			offset += 0x0324;
+		}
+	}
+}
+
+void
+nvc1_grctx_generate_unkn(struct nvc0_graph_priv *priv)
+{
+	nv_mask(priv, 0x418c6c, 0x00000001, 0x00000001);
+	nv_mask(priv, 0x41980c, 0x00000010, 0x00000010);
+	nv_mask(priv, 0x419814, 0x00000004, 0x00000004);
+	nv_mask(priv, 0x4064c0, 0x80000000, 0x80000000);
+	nv_mask(priv, 0x405800, 0x08000000, 0x08000000);
+	nv_mask(priv, 0x419c00, 0x00000008, 0x00000008);
+}
+
+static struct nvc0_graph_init *
+nvc1_grctx_init_hub[] = {
+	nvc0_grctx_init_base,
+	nvc0_grctx_init_unk40xx,
+	nvc0_grctx_init_unk44xx,
+	nvc0_grctx_init_unk46xx,
+	nvc0_grctx_init_unk47xx,
+	nvc1_grctx_init_unk58xx,
+	nvc0_grctx_init_unk60xx,
+	nvc0_grctx_init_unk64xx,
+	nvc0_grctx_init_unk78xx,
+	nvc0_grctx_init_unk80xx,
+	nvc1_grctx_init_rop,
+	NULL
+};
+
+struct nvc0_graph_init *
+nvc1_grctx_init_gpc[] = {
+	nvc1_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvc1_grctx_init_tpc,
+	NULL
+};
+
+static struct nvc0_graph_mthd
+nvc1_grctx_init_mthd[] = {
+	{ 0x9097, nvc1_grctx_init_9097, },
+	{ 0x9197, nvc1_grctx_init_9197, },
+	{ 0x902d, nvc0_grctx_init_902d, },
+	{ 0x9039, nvc0_grctx_init_9039, },
+	{ 0x90c0, nvc0_grctx_init_90c0, },
+	{ 0x902d, nvc0_grctx_init_mthd_magic, },
+	{}
+};
+
+struct nouveau_oclass *
+nvc1_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xc1),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nvc0_grctx_generate_main,
+	.mods = nvc1_grctx_generate_mods,
+	.unkn = nvc1_grctx_generate_unkn,
+	.hub  = nvc1_grctx_init_hub,
+	.gpc  = nvc1_grctx_init_gpc,
+	.icmd = nvc1_grctx_init_icmd,
+	.mthd = nvc1_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc3.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc3.c
new file mode 100644
index 000000000000..8f237b3bd8c6
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc3.c
@@ -0,0 +1,99 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+static struct nvc0_graph_init
+nvc3_grctx_init_tpc[] = {
+	{ 0x419818,   1, 0x04, 0x00000000 },
+	{ 0x41983c,   1, 0x04, 0x00038bc7 },
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x0000012a },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000001f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000023 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x00000000 },
+	{ 0x419a20,   1, 0x04, 0x00000800 },
+	{ 0x419ac4,   1, 0x04, 0x0007f440 },
+	{ 0x419b00,   1, 0x04, 0x0a418820 },
+	{ 0x419b04,   1, 0x04, 0x062080e6 },
+	{ 0x419b08,   1, 0x04, 0x020398a4 },
+	{ 0x419b0c,   1, 0x04, 0x0e629062 },
+	{ 0x419b10,   1, 0x04, 0x0a418820 },
+	{ 0x419b14,   1, 0x04, 0x000000e6 },
+	{ 0x419bd0,   1, 0x04, 0x00900103 },
+	{ 0x419be0,   1, 0x04, 0x00000001 },
+	{ 0x419be4,   1, 0x04, 0x00000000 },
+	{ 0x419c00,   1, 0x04, 0x00000002 },
+	{ 0x419c04,   1, 0x04, 0x00000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x00020048 },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000183 },
+	{ 0x419d20,   1, 0x04, 0x02180000 },
+	{ 0x419d24,   1, 0x04, 0x00001fff },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000002 },
+	{ 0x419e44,   1, 0x04, 0x001beff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000000f },
+	{ 0x419e50,  17, 0x04, 0x00000000 },
+	{ 0x419e98,   1, 0x04, 0x00000000 },
+	{ 0x419ee0,   1, 0x04, 0x00011110 },
+	{ 0x419f30,  11, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init *
+nvc3_grctx_init_gpc[] = {
+	nvc0_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvc3_grctx_init_tpc,
+	NULL
+};
+
+struct nouveau_oclass *
+nvc3_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xc3),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nvc0_grctx_generate_main,
+	.mods = nvc0_grctx_generate_mods,
+	.unkn = nvc0_grctx_generate_unkn,
+	.hub  = nvc0_grctx_init_hub,
+	.gpc  = nvc3_grctx_init_gpc,
+	.icmd = nvc0_grctx_init_icmd,
+	.mthd = nvc0_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc8.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc8.c
new file mode 100644
index 000000000000..d0d4ce3c4892
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvc8.c
@@ -0,0 +1,370 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+static struct nvc0_graph_init
+nvc8_grctx_init_icmd[] = {
+	{ 0x001000,   1, 0x01, 0x00000004 },
+	{ 0x0000a9,   1, 0x01, 0x0000ffff },
+	{ 0x000038,   1, 0x01, 0x0fac6881 },
+	{ 0x00003d,   1, 0x01, 0x00000001 },
+	{ 0x0000e8,   8, 0x01, 0x00000400 },
+	{ 0x000078,   8, 0x01, 0x00000300 },
+	{ 0x000050,   1, 0x01, 0x00000011 },
+	{ 0x000058,   8, 0x01, 0x00000008 },
+	{ 0x000208,   8, 0x01, 0x00000001 },
+	{ 0x000081,   1, 0x01, 0x00000001 },
+	{ 0x000085,   1, 0x01, 0x00000004 },
+	{ 0x000088,   1, 0x01, 0x00000400 },
+	{ 0x000090,   1, 0x01, 0x00000300 },
+	{ 0x000098,   1, 0x01, 0x00001001 },
+	{ 0x0000e3,   1, 0x01, 0x00000001 },
+	{ 0x0000da,   1, 0x01, 0x00000001 },
+	{ 0x0000f8,   1, 0x01, 0x00000003 },
+	{ 0x0000fa,   1, 0x01, 0x00000001 },
+	{ 0x00009f,   4, 0x01, 0x0000ffff },
+	{ 0x0000b1,   1, 0x01, 0x00000001 },
+	{ 0x0000b2,  40, 0x01, 0x00000000 },
+	{ 0x000210,   8, 0x01, 0x00000040 },
+	{ 0x000218,   8, 0x01, 0x0000c080 },
+	{ 0x0000ad,   1, 0x01, 0x0000013e },
+	{ 0x0000e1,   1, 0x01, 0x00000010 },
+	{ 0x000290,  16, 0x01, 0x00000000 },
+	{ 0x0003b0,  16, 0x01, 0x00000000 },
+	{ 0x0002a0,  16, 0x01, 0x00000000 },
+	{ 0x000420,  16, 0x01, 0x00000000 },
+	{ 0x0002b0,  16, 0x01, 0x00000000 },
+	{ 0x000430,  16, 0x01, 0x00000000 },
+	{ 0x0002c0,  16, 0x01, 0x00000000 },
+	{ 0x0004d0,  16, 0x01, 0x00000000 },
+	{ 0x000720,  16, 0x01, 0x00000000 },
+	{ 0x0008c0,  16, 0x01, 0x00000000 },
+	{ 0x000890,  16, 0x01, 0x00000000 },
+	{ 0x0008e0,  16, 0x01, 0x00000000 },
+	{ 0x0008a0,  16, 0x01, 0x00000000 },
+	{ 0x0008f0,  16, 0x01, 0x00000000 },
+	{ 0x00094c,   1, 0x01, 0x000000ff },
+	{ 0x00094d,   1, 0x01, 0xffffffff },
+	{ 0x00094e,   1, 0x01, 0x00000002 },
+	{ 0x0002ec,   1, 0x01, 0x00000001 },
+	{ 0x000303,   1, 0x01, 0x00000001 },
+	{ 0x0002e6,   1, 0x01, 0x00000001 },
+	{ 0x000466,   1, 0x01, 0x00000052 },
+	{ 0x000301,   1, 0x01, 0x3f800000 },
+	{ 0x000304,   1, 0x01, 0x30201000 },
+	{ 0x000305,   1, 0x01, 0x70605040 },
+	{ 0x000306,   1, 0x01, 0xb8a89888 },
+	{ 0x000307,   1, 0x01, 0xf8e8d8c8 },
+	{ 0x00030a,   1, 0x01, 0x00ffff00 },
+	{ 0x00030b,   1, 0x01, 0x0000001a },
+	{ 0x00030c,   1, 0x01, 0x00000001 },
+	{ 0x000318,   1, 0x01, 0x00000001 },
+	{ 0x000340,   1, 0x01, 0x00000000 },
+	{ 0x000375,   1, 0x01, 0x00000001 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x00037d,   1, 0x01, 0x00000006 },
+	{ 0x0003a0,   1, 0x01, 0x00000002 },
+	{ 0x0003aa,   1, 0x01, 0x00000001 },
+	{ 0x0003a9,   1, 0x01, 0x00000001 },
+	{ 0x000380,   1, 0x01, 0x00000001 },
+	{ 0x000360,   1, 0x01, 0x00000040 },
+	{ 0x000366,   2, 0x01, 0x00000000 },
+	{ 0x000368,   1, 0x01, 0x00001fff },
+	{ 0x000370,   2, 0x01, 0x00000000 },
+	{ 0x000372,   1, 0x01, 0x003fffff },
+	{ 0x00037a,   1, 0x01, 0x00000012 },
+	{ 0x0005e0,   5, 0x01, 0x00000022 },
+	{ 0x000619,   1, 0x01, 0x00000003 },
+	{ 0x000811,   1, 0x01, 0x00000003 },
+	{ 0x000812,   1, 0x01, 0x00000004 },
+	{ 0x000813,   1, 0x01, 0x00000006 },
+	{ 0x000814,   1, 0x01, 0x00000008 },
+	{ 0x000815,   1, 0x01, 0x0000000b },
+	{ 0x000800,   6, 0x01, 0x00000001 },
+	{ 0x000632,   1, 0x01, 0x00000001 },
+	{ 0x000633,   1, 0x01, 0x00000002 },
+	{ 0x000634,   1, 0x01, 0x00000003 },
+	{ 0x000635,   1, 0x01, 0x00000004 },
+	{ 0x000654,   1, 0x01, 0x3f800000 },
+	{ 0x000657,   1, 0x01, 0x3f800000 },
+	{ 0x000655,   2, 0x01, 0x3f800000 },
+	{ 0x0006cd,   1, 0x01, 0x3f800000 },
+	{ 0x0007f5,   1, 0x01, 0x3f800000 },
+	{ 0x0007dc,   1, 0x01, 0x39291909 },
+	{ 0x0007dd,   1, 0x01, 0x79695949 },
+	{ 0x0007de,   1, 0x01, 0xb9a99989 },
+	{ 0x0007df,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007e8,   1, 0x01, 0x00003210 },
+	{ 0x0007e9,   1, 0x01, 0x00007654 },
+	{ 0x0007ea,   1, 0x01, 0x00000098 },
+	{ 0x0007ec,   1, 0x01, 0x39291909 },
+	{ 0x0007ed,   1, 0x01, 0x79695949 },
+	{ 0x0007ee,   1, 0x01, 0xb9a99989 },
+	{ 0x0007ef,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007f0,   1, 0x01, 0x00003210 },
+	{ 0x0007f1,   1, 0x01, 0x00007654 },
+	{ 0x0007f2,   1, 0x01, 0x00000098 },
+	{ 0x0005a5,   1, 0x01, 0x00000001 },
+	{ 0x000980, 128, 0x01, 0x00000000 },
+	{ 0x000468,   1, 0x01, 0x00000004 },
+	{ 0x00046c,   1, 0x01, 0x00000001 },
+	{ 0x000470,  96, 0x01, 0x00000000 },
+	{ 0x000510,  16, 0x01, 0x3f800000 },
+	{ 0x000520,   1, 0x01, 0x000002b6 },
+	{ 0x000529,   1, 0x01, 0x00000001 },
+	{ 0x000530,  16, 0x01, 0xffff0000 },
+	{ 0x000585,   1, 0x01, 0x0000003f },
+	{ 0x000576,   1, 0x01, 0x00000003 },
+	{ 0x00057b,   1, 0x01, 0x00000059 },
+	{ 0x000586,   1, 0x01, 0x00000040 },
+	{ 0x000582,   2, 0x01, 0x00000080 },
+	{ 0x0005c2,   1, 0x01, 0x00000001 },
+	{ 0x000638,   1, 0x01, 0x00000001 },
+	{ 0x000639,   1, 0x01, 0x00000001 },
+	{ 0x00063a,   1, 0x01, 0x00000002 },
+	{ 0x00063b,   2, 0x01, 0x00000001 },
+	{ 0x00063d,   1, 0x01, 0x00000002 },
+	{ 0x00063e,   1, 0x01, 0x00000001 },
+	{ 0x0008b8,   8, 0x01, 0x00000001 },
+	{ 0x000900,   8, 0x01, 0x00000001 },
+	{ 0x000908,   8, 0x01, 0x00000002 },
+	{ 0x000910,  16, 0x01, 0x00000001 },
+	{ 0x000920,   8, 0x01, 0x00000002 },
+	{ 0x000928,   8, 0x01, 0x00000001 },
+	{ 0x000648,   9, 0x01, 0x00000001 },
+	{ 0x000658,   1, 0x01, 0x0000000f },
+	{ 0x0007ff,   1, 0x01, 0x0000000a },
+	{ 0x00066a,   1, 0x01, 0x40000000 },
+	{ 0x00066b,   1, 0x01, 0x10000000 },
+	{ 0x00066c,   2, 0x01, 0xffff0000 },
+	{ 0x0007af,   2, 0x01, 0x00000008 },
+	{ 0x0007f6,   1, 0x01, 0x00000001 },
+	{ 0x0006b2,   1, 0x01, 0x00000055 },
+	{ 0x0007ad,   1, 0x01, 0x00000003 },
+	{ 0x000937,   1, 0x01, 0x00000001 },
+	{ 0x000971,   1, 0x01, 0x00000008 },
+	{ 0x000972,   1, 0x01, 0x00000040 },
+	{ 0x000973,   1, 0x01, 0x0000012c },
+	{ 0x00097c,   1, 0x01, 0x00000040 },
+	{ 0x000979,   1, 0x01, 0x00000003 },
+	{ 0x000975,   1, 0x01, 0x00000020 },
+	{ 0x000976,   1, 0x01, 0x00000001 },
+	{ 0x000977,   1, 0x01, 0x00000020 },
+	{ 0x000978,   1, 0x01, 0x00000001 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095e,   1, 0x01, 0x20164010 },
+	{ 0x00095f,   1, 0x01, 0x00000020 },
+	{ 0x00097d,   1, 0x01, 0x00000020 },
+	{ 0x000683,   1, 0x01, 0x00000006 },
+	{ 0x000685,   1, 0x01, 0x003fffff },
+	{ 0x000687,   1, 0x01, 0x00000c48 },
+	{ 0x0006a0,   1, 0x01, 0x00000005 },
+	{ 0x000840,   1, 0x01, 0x00300008 },
+	{ 0x000841,   1, 0x01, 0x04000080 },
+	{ 0x000842,   1, 0x01, 0x00300008 },
+	{ 0x000843,   1, 0x01, 0x04000080 },
+	{ 0x000818,   8, 0x01, 0x00000000 },
+	{ 0x000848,  16, 0x01, 0x00000000 },
+	{ 0x000738,   1, 0x01, 0x00000000 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ab,   1, 0x01, 0x00000002 },
+	{ 0x0006ac,   1, 0x01, 0x00000080 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x0006bb,   1, 0x01, 0x000000cf },
+	{ 0x0006ce,   1, 0x01, 0x2a712488 },
+	{ 0x000739,   1, 0x01, 0x4085c000 },
+	{ 0x00073a,   1, 0x01, 0x00000080 },
+	{ 0x000786,   1, 0x01, 0x80000100 },
+	{ 0x00073c,   1, 0x01, 0x00010100 },
+	{ 0x00073d,   1, 0x01, 0x02800000 },
+	{ 0x000787,   1, 0x01, 0x000000cf },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x000836,   1, 0x01, 0x00000001 },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x000944,   1, 0x01, 0x00000022 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x00c1b0,   8, 0x01, 0x0000000f },
+	{ 0x00c1b8,   1, 0x01, 0x0fac6881 },
+	{ 0x00c1b9,   1, 0x01, 0x00fac688 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000002 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000014 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000001 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc8_grctx_init_tpc[] = {
+	{ 0x419818,   1, 0x04, 0x00000000 },
+	{ 0x41983c,   1, 0x04, 0x00038bc7 },
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x0000012a },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000001f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000023 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x00000000 },
+	{ 0x419a20,   1, 0x04, 0x00000800 },
+	{ 0x419b00,   1, 0x04, 0x0a418820 },
+	{ 0x419b04,   1, 0x04, 0x062080e6 },
+	{ 0x419b08,   1, 0x04, 0x020398a4 },
+	{ 0x419b0c,   1, 0x04, 0x0e629062 },
+	{ 0x419b10,   1, 0x04, 0x0a418820 },
+	{ 0x419b14,   1, 0x04, 0x000000e6 },
+	{ 0x419bd0,   1, 0x04, 0x00900103 },
+	{ 0x419be0,   1, 0x04, 0x00000001 },
+	{ 0x419be4,   1, 0x04, 0x00000000 },
+	{ 0x419c00,   1, 0x04, 0x00000002 },
+	{ 0x419c04,   1, 0x04, 0x00000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x00060048 },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000183 },
+	{ 0x419d20,   1, 0x04, 0x02180000 },
+	{ 0x419d24,   1, 0x04, 0x00001fff },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000002 },
+	{ 0x419e44,   1, 0x04, 0x001beff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000000f },
+	{ 0x419e50,  17, 0x04, 0x00000000 },
+	{ 0x419e98,   1, 0x04, 0x00000000 },
+	{ 0x419f50,   2, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc8_grctx_init_9197[] = {
+	{ 0x0002e4,   1, 0x04, 0x0000b001 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc8_grctx_init_9297[] = {
+	{ 0x003400, 128, 0x04, 0x00000000 },
+	{ 0x00036c,   2, 0x04, 0x00000000 },
+	{ 0x0007a4,   2, 0x04, 0x00000000 },
+	{ 0x000374,   1, 0x04, 0x00000000 },
+	{ 0x000378,   1, 0x04, 0x00000020 },
+	{}
+};
+
+static struct nvc0_graph_mthd
+nvc8_grctx_init_mthd[] = {
+	{ 0x9097, nvc1_grctx_init_9097, },
+	{ 0x9197, nvc8_grctx_init_9197, },
+	{ 0x9297, nvc8_grctx_init_9297, },
+	{ 0x902d, nvc0_grctx_init_902d, },
+	{ 0x9039, nvc0_grctx_init_9039, },
+	{ 0x90c0, nvc0_grctx_init_90c0, },
+	{ 0x902d, nvc0_grctx_init_mthd_magic, },
+	{}
+};
+
+static struct nvc0_graph_init *
+nvc8_grctx_init_gpc[] = {
+	nvc0_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvc8_grctx_init_tpc,
+	NULL
+};
+
+struct nouveau_oclass *
+nvc8_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xc8),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nvc0_grctx_generate_main,
+	.mods = nvc0_grctx_generate_mods,
+	.unkn = nvc0_grctx_generate_unkn,
+	.hub  = nvc0_grctx_init_hub,
+	.gpc  = nvc8_grctx_init_gpc,
+	.icmd = nvc8_grctx_init_icmd,
+	.mthd = nvc8_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd7.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd7.c
new file mode 100644
index 000000000000..438e78410808
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd7.c
@@ -0,0 +1,290 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+struct nvc0_graph_init
+nvd7_grctx_init_unk40xx[] = {
+	{ 0x404004,  10, 0x04, 0x00000000 },
+	{ 0x404044,   1, 0x04, 0x00000000 },
+	{ 0x404094,   1, 0x04, 0x00000000 },
+	{ 0x404098,  12, 0x04, 0x00000000 },
+	{ 0x4040c8,   1, 0x04, 0xf0000087 },
+	{ 0x4040d0,   6, 0x04, 0x00000000 },
+	{ 0x4040e8,   1, 0x04, 0x00001000 },
+	{ 0x4040f8,   1, 0x04, 0x00000000 },
+	{ 0x404130,   1, 0x04, 0x00000000 },
+	{ 0x404134,   1, 0x04, 0x00000000 },
+	{ 0x404138,   1, 0x04, 0x20000040 },
+	{ 0x404150,   1, 0x04, 0x0000002e },
+	{ 0x404154,   1, 0x04, 0x00000400 },
+	{ 0x404158,   1, 0x04, 0x00000200 },
+	{ 0x404164,   1, 0x04, 0x00000055 },
+	{ 0x404168,   1, 0x04, 0x00000000 },
+	{ 0x404178,   2, 0x04, 0x00000000 },
+	{ 0x404200,   8, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_grctx_init_unk58xx[] = {
+	{ 0x405800,   1, 0x04, 0x0f8000bf },
+	{ 0x405830,   1, 0x04, 0x02180324 },
+	{ 0x405834,   1, 0x04, 0x08000000 },
+	{ 0x405838,   1, 0x04, 0x00000000 },
+	{ 0x405854,   1, 0x04, 0x00000000 },
+	{ 0x405870,   4, 0x04, 0x00000001 },
+	{ 0x405a00,   2, 0x04, 0x00000000 },
+	{ 0x405a18,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_grctx_init_unk64xx[] = {
+	{ 0x4064a8,   1, 0x04, 0x00000000 },
+	{ 0x4064ac,   1, 0x04, 0x00003fff },
+	{ 0x4064b4,   3, 0x04, 0x00000000 },
+	{ 0x4064c0,   1, 0x04, 0x801a0078 },
+	{ 0x4064c4,   1, 0x04, 0x00c9ffff },
+	{ 0x4064d0,   8, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_grctx_init_gpc_0[] = {
+	{ 0x418380,   1, 0x04, 0x00000016 },
+	{ 0x418400,   1, 0x04, 0x38004e00 },
+	{ 0x418404,   1, 0x04, 0x71e0ffff },
+	{ 0x41840c,   1, 0x04, 0x00001008 },
+	{ 0x418410,   1, 0x04, 0x0fff0fff },
+	{ 0x418414,   1, 0x04, 0x02200fff },
+	{ 0x418450,   6, 0x04, 0x00000000 },
+	{ 0x418468,   1, 0x04, 0x00000001 },
+	{ 0x41846c,   2, 0x04, 0x00000000 },
+	{ 0x418600,   1, 0x04, 0x0000001f },
+	{ 0x418684,   1, 0x04, 0x0000000f },
+	{ 0x418700,   1, 0x04, 0x00000002 },
+	{ 0x418704,   1, 0x04, 0x00000080 },
+	{ 0x418708,   3, 0x04, 0x00000000 },
+	{ 0x418800,   1, 0x04, 0x7006860a },
+	{ 0x418808,   3, 0x04, 0x00000000 },
+	{ 0x418828,   1, 0x04, 0x00008442 },
+	{ 0x418830,   1, 0x04, 0x10000001 },
+	{ 0x4188d8,   1, 0x04, 0x00000008 },
+	{ 0x4188e0,   1, 0x04, 0x01000000 },
+	{ 0x4188e8,   5, 0x04, 0x00000000 },
+	{ 0x4188fc,   1, 0x04, 0x20100018 },
+	{ 0x41891c,   1, 0x04, 0x00ff00ff },
+	{ 0x418924,   1, 0x04, 0x00000000 },
+	{ 0x418928,   1, 0x04, 0x00ffff00 },
+	{ 0x41892c,   1, 0x04, 0x0000ff00 },
+	{ 0x418b00,   1, 0x04, 0x00000006 },
+	{ 0x418b08,   1, 0x04, 0x0a418820 },
+	{ 0x418b0c,   1, 0x04, 0x062080e6 },
+	{ 0x418b10,   1, 0x04, 0x020398a4 },
+	{ 0x418b14,   1, 0x04, 0x0e629062 },
+	{ 0x418b18,   1, 0x04, 0x0a418820 },
+	{ 0x418b1c,   1, 0x04, 0x000000e6 },
+	{ 0x418bb8,   1, 0x04, 0x00000103 },
+	{ 0x418c08,   1, 0x04, 0x00000001 },
+	{ 0x418c10,   8, 0x04, 0x00000000 },
+	{ 0x418c6c,   1, 0x04, 0x00000001 },
+	{ 0x418c80,   1, 0x04, 0x20200004 },
+	{ 0x418c8c,   1, 0x04, 0x00000001 },
+	{ 0x419000,   1, 0x04, 0x00000780 },
+	{ 0x419004,   2, 0x04, 0x00000000 },
+	{ 0x419014,   1, 0x04, 0x00000004 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_grctx_init_tpc[] = {
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x00000129 },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000001f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000023 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x00008000 },
+	{ 0x419a20,   1, 0x04, 0x00000800 },
+	{ 0x419ac4,   1, 0x04, 0x0017f440 },
+	{ 0x419c00,   1, 0x04, 0x0000000a },
+	{ 0x419c04,   1, 0x04, 0x00000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419c24,   1, 0x04, 0x00084210 },
+	{ 0x419c28,   1, 0x04, 0x3efbefbe },
+	{ 0x419cb0,   1, 0x04, 0x00020048 },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000183 },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000002 },
+	{ 0x419e44,   1, 0x04, 0x001beff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000000f },
+	{ 0x419e50,  17, 0x04, 0x00000000 },
+	{ 0x419e98,   1, 0x04, 0x00000000 },
+	{ 0x419ee0,   1, 0x04, 0x00010110 },
+	{ 0x419f30,  11, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_grctx_init_unk[] = {
+	{ 0x41be24,   1, 0x04, 0x00000002 },
+	{ 0x41bec0,   1, 0x04, 0x12180000 },
+	{ 0x41bec4,   1, 0x04, 0x00003fff },
+	{ 0x41bee4,   1, 0x04, 0x03240218 },
+	{ 0x41bf00,   1, 0x04, 0x0a418820 },
+	{ 0x41bf04,   1, 0x04, 0x062080e6 },
+	{ 0x41bf08,   1, 0x04, 0x020398a4 },
+	{ 0x41bf0c,   1, 0x04, 0x0e629062 },
+	{ 0x41bf10,   1, 0x04, 0x0a418820 },
+	{ 0x41bf14,   1, 0x04, 0x000000e6 },
+	{ 0x41bfd0,   1, 0x04, 0x00900103 },
+	{ 0x41bfe0,   1, 0x04, 0x00400001 },
+	{ 0x41bfe4,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static void
+nvd7_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+{
+	u32 magic[GPC_MAX][2];
+	u32 offset;
+	int gpc;
+
+	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
+	mmio_list(0x40800c, 0x00000000,  8, 1);
+	mmio_list(0x408010, 0x80000000,  0, 0);
+	mmio_list(0x419004, 0x00000000,  8, 1);
+	mmio_list(0x419008, 0x00000000,  0, 0);
+	mmio_list(0x408004, 0x00000000,  8, 0);
+	mmio_list(0x408008, 0x80000018,  0, 0);
+	mmio_list(0x418808, 0x00000000,  8, 0);
+	mmio_list(0x41880c, 0x80000018,  0, 0);
+	mmio_list(0x418810, 0x80000000, 12, 2);
+	mmio_list(0x419848, 0x10000000, 12, 2);
+
+	mmio_list(0x405830, 0x02180324,  0, 0);
+	mmio_list(0x4064c4, 0x00c9ffff,  0, 0);
+
+	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
+		u16 magic0 = 0x0218 * priv->tpc_nr[gpc];
+		u16 magic1 = 0x0324 * priv->tpc_nr[gpc];
+		magic[gpc][0]  = 0x10000000 | (magic0 << 16) | offset;
+		magic[gpc][1]  = 0x00000000 | (magic1 << 16);
+		offset += 0x0324 * priv->tpc_nr[gpc];
+	}
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		mmio_list(GPC_UNIT(gpc, 0x30c0), magic[gpc][0], 0, 0);
+		mmio_list(GPC_UNIT(gpc, 0x30e4), magic[gpc][1] | offset, 0, 0);
+		offset += 0x07ff * priv->tpc_nr[gpc];
+	}
+	mmio_list(0x17e91c, 0x03060609, 0, 0); /* different from kepler */
+}
+
+void
+nvd7_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+{
+	struct nvc0_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass;
+	int i;
+
+	nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
+
+	for (i = 0; oclass->hub[i]; i++)
+		nvc0_graph_mmio(priv, oclass->hub[i]);
+	for (i = 0; oclass->gpc[i]; i++)
+		nvc0_graph_mmio(priv, oclass->gpc[i]);
+
+	nv_wr32(priv, 0x404154, 0x00000000);
+
+	oclass->mods(priv, info);
+	oclass->unkn(priv);
+
+	nvc0_grctx_generate_tpcid(priv);
+	nvc0_grctx_generate_r406028(priv);
+	nvc0_grctx_generate_r4060a8(priv);
+	nve4_grctx_generate_r418bb8(priv);
+	nvc0_grctx_generate_r406800(priv);
+
+	for (i = 0; i < 8; i++)
+		nv_wr32(priv, 0x4064d0 + (i * 0x04), 0x00000000);
+
+	nvc0_graph_icmd(priv, oclass->icmd);
+	nv_wr32(priv, 0x404154, 0x00000400);
+	nvc0_graph_mthd(priv, oclass->mthd);
+	nv_mask(priv, 0x000260, 0x00000001, 0x00000001);
+}
+
+
+static struct nvc0_graph_init *
+nvd7_grctx_init_hub[] = {
+	nvc0_grctx_init_base,
+	nvd7_grctx_init_unk40xx,
+	nvc0_grctx_init_unk44xx,
+	nvc0_grctx_init_unk46xx,
+	nvc0_grctx_init_unk47xx,
+	nvd7_grctx_init_unk58xx,
+	nvc0_grctx_init_unk60xx,
+	nvd7_grctx_init_unk64xx,
+	nvc0_grctx_init_unk78xx,
+	nvc0_grctx_init_unk80xx,
+	nvd9_grctx_init_rop,
+};
+
+struct nvc0_graph_init *
+nvd7_grctx_init_gpc[] = {
+	nvd7_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvd7_grctx_init_tpc,
+	nvd7_grctx_init_unk,
+	NULL
+};
+
+struct nouveau_oclass *
+nvd7_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xd7),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nvd7_grctx_generate_main,
+	.mods = nvd7_grctx_generate_mods,
+	.unkn = nve4_grctx_generate_unkn,
+	.hub  = nvd7_grctx_init_hub,
+	.gpc  = nvd7_grctx_init_gpc,
+	.icmd = nvd9_grctx_init_icmd,
+	.mthd = nvd9_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd9.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd9.c
new file mode 100644
index 000000000000..818a4751df46
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvd9.c
@@ -0,0 +1,515 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+struct nvc0_graph_init
+nvd9_grctx_init_90c0[] = {
+	{ 0x002700,   4, 0x40, 0x00000000 },
+	{ 0x002720,   4, 0x40, 0x00000000 },
+	{ 0x002704,   4, 0x40, 0x00000000 },
+	{ 0x002724,   4, 0x40, 0x00000000 },
+	{ 0x002708,   4, 0x40, 0x00000000 },
+	{ 0x002728,   4, 0x40, 0x00000000 },
+	{ 0x00270c,   8, 0x20, 0x00000000 },
+	{ 0x002710,   4, 0x40, 0x00014000 },
+	{ 0x002730,   4, 0x40, 0x00014000 },
+	{ 0x002714,   4, 0x40, 0x00000040 },
+	{ 0x002734,   4, 0x40, 0x00000040 },
+	{ 0x00030c,   1, 0x04, 0x00000001 },
+	{ 0x001944,   1, 0x04, 0x00000000 },
+	{ 0x000758,   1, 0x04, 0x00000100 },
+	{ 0x0002c4,   1, 0x04, 0x00000000 },
+	{ 0x000790,   5, 0x04, 0x00000000 },
+	{ 0x00077c,   1, 0x04, 0x00000000 },
+	{ 0x000204,   3, 0x04, 0x00000000 },
+	{ 0x000214,   1, 0x04, 0x00000000 },
+	{ 0x00024c,   1, 0x04, 0x00000000 },
+	{ 0x000d94,   1, 0x04, 0x00000001 },
+	{ 0x001608,   2, 0x04, 0x00000000 },
+	{ 0x001664,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvd9_grctx_init_icmd[] = {
+	{ 0x001000,   1, 0x01, 0x00000004 },
+	{ 0x0000a9,   1, 0x01, 0x0000ffff },
+	{ 0x000038,   1, 0x01, 0x0fac6881 },
+	{ 0x00003d,   1, 0x01, 0x00000001 },
+	{ 0x0000e8,   8, 0x01, 0x00000400 },
+	{ 0x000078,   8, 0x01, 0x00000300 },
+	{ 0x000050,   1, 0x01, 0x00000011 },
+	{ 0x000058,   8, 0x01, 0x00000008 },
+	{ 0x000208,   8, 0x01, 0x00000001 },
+	{ 0x000081,   1, 0x01, 0x00000001 },
+	{ 0x000085,   1, 0x01, 0x00000004 },
+	{ 0x000088,   1, 0x01, 0x00000400 },
+	{ 0x000090,   1, 0x01, 0x00000300 },
+	{ 0x000098,   1, 0x01, 0x00001001 },
+	{ 0x0000e3,   1, 0x01, 0x00000001 },
+	{ 0x0000da,   1, 0x01, 0x00000001 },
+	{ 0x0000f8,   1, 0x01, 0x00000003 },
+	{ 0x0000fa,   1, 0x01, 0x00000001 },
+	{ 0x00009f,   4, 0x01, 0x0000ffff },
+	{ 0x0000b1,   1, 0x01, 0x00000001 },
+	{ 0x0000b2,  40, 0x01, 0x00000000 },
+	{ 0x000210,   8, 0x01, 0x00000040 },
+	{ 0x000400,  24, 0x01, 0x00000040 },
+	{ 0x000218,   8, 0x01, 0x0000c080 },
+	{ 0x000440,  24, 0x01, 0x0000c080 },
+	{ 0x0000ad,   1, 0x01, 0x0000013e },
+	{ 0x0000e1,   1, 0x01, 0x00000010 },
+	{ 0x000290,  16, 0x01, 0x00000000 },
+	{ 0x0003b0,  16, 0x01, 0x00000000 },
+	{ 0x0002a0,  16, 0x01, 0x00000000 },
+	{ 0x000420,  16, 0x01, 0x00000000 },
+	{ 0x0002b0,  16, 0x01, 0x00000000 },
+	{ 0x000430,  16, 0x01, 0x00000000 },
+	{ 0x0002c0,  16, 0x01, 0x00000000 },
+	{ 0x0004d0,  16, 0x01, 0x00000000 },
+	{ 0x000720,  16, 0x01, 0x00000000 },
+	{ 0x0008c0,  16, 0x01, 0x00000000 },
+	{ 0x000890,  16, 0x01, 0x00000000 },
+	{ 0x0008e0,  16, 0x01, 0x00000000 },
+	{ 0x0008a0,  16, 0x01, 0x00000000 },
+	{ 0x0008f0,  16, 0x01, 0x00000000 },
+	{ 0x00094c,   1, 0x01, 0x000000ff },
+	{ 0x00094d,   1, 0x01, 0xffffffff },
+	{ 0x00094e,   1, 0x01, 0x00000002 },
+	{ 0x0002ec,   1, 0x01, 0x00000001 },
+	{ 0x000303,   1, 0x01, 0x00000001 },
+	{ 0x0002e6,   1, 0x01, 0x00000001 },
+	{ 0x000466,   1, 0x01, 0x00000052 },
+	{ 0x000301,   1, 0x01, 0x3f800000 },
+	{ 0x000304,   1, 0x01, 0x30201000 },
+	{ 0x000305,   1, 0x01, 0x70605040 },
+	{ 0x000306,   1, 0x01, 0xb8a89888 },
+	{ 0x000307,   1, 0x01, 0xf8e8d8c8 },
+	{ 0x00030a,   1, 0x01, 0x00ffff00 },
+	{ 0x00030b,   1, 0x01, 0x0000001a },
+	{ 0x00030c,   1, 0x01, 0x00000001 },
+	{ 0x000318,   1, 0x01, 0x00000001 },
+	{ 0x000340,   1, 0x01, 0x00000000 },
+	{ 0x000375,   1, 0x01, 0x00000001 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x00037d,   1, 0x01, 0x00000006 },
+	{ 0x0003a0,   1, 0x01, 0x00000002 },
+	{ 0x0003aa,   1, 0x01, 0x00000001 },
+	{ 0x0003a9,   1, 0x01, 0x00000001 },
+	{ 0x000380,   1, 0x01, 0x00000001 },
+	{ 0x000360,   1, 0x01, 0x00000040 },
+	{ 0x000366,   2, 0x01, 0x00000000 },
+	{ 0x000368,   1, 0x01, 0x00001fff },
+	{ 0x000370,   2, 0x01, 0x00000000 },
+	{ 0x000372,   1, 0x01, 0x003fffff },
+	{ 0x00037a,   1, 0x01, 0x00000012 },
+	{ 0x0005e0,   5, 0x01, 0x00000022 },
+	{ 0x000619,   1, 0x01, 0x00000003 },
+	{ 0x000811,   1, 0x01, 0x00000003 },
+	{ 0x000812,   1, 0x01, 0x00000004 },
+	{ 0x000813,   1, 0x01, 0x00000006 },
+	{ 0x000814,   1, 0x01, 0x00000008 },
+	{ 0x000815,   1, 0x01, 0x0000000b },
+	{ 0x000800,   6, 0x01, 0x00000001 },
+	{ 0x000632,   1, 0x01, 0x00000001 },
+	{ 0x000633,   1, 0x01, 0x00000002 },
+	{ 0x000634,   1, 0x01, 0x00000003 },
+	{ 0x000635,   1, 0x01, 0x00000004 },
+	{ 0x000654,   1, 0x01, 0x3f800000 },
+	{ 0x000657,   1, 0x01, 0x3f800000 },
+	{ 0x000655,   2, 0x01, 0x3f800000 },
+	{ 0x0006cd,   1, 0x01, 0x3f800000 },
+	{ 0x0007f5,   1, 0x01, 0x3f800000 },
+	{ 0x0007dc,   1, 0x01, 0x39291909 },
+	{ 0x0007dd,   1, 0x01, 0x79695949 },
+	{ 0x0007de,   1, 0x01, 0xb9a99989 },
+	{ 0x0007df,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007e8,   1, 0x01, 0x00003210 },
+	{ 0x0007e9,   1, 0x01, 0x00007654 },
+	{ 0x0007ea,   1, 0x01, 0x00000098 },
+	{ 0x0007ec,   1, 0x01, 0x39291909 },
+	{ 0x0007ed,   1, 0x01, 0x79695949 },
+	{ 0x0007ee,   1, 0x01, 0xb9a99989 },
+	{ 0x0007ef,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007f0,   1, 0x01, 0x00003210 },
+	{ 0x0007f1,   1, 0x01, 0x00007654 },
+	{ 0x0007f2,   1, 0x01, 0x00000098 },
+	{ 0x0005a5,   1, 0x01, 0x00000001 },
+	{ 0x000980, 128, 0x01, 0x00000000 },
+	{ 0x000468,   1, 0x01, 0x00000004 },
+	{ 0x00046c,   1, 0x01, 0x00000001 },
+	{ 0x000470,  96, 0x01, 0x00000000 },
+	{ 0x000510,  16, 0x01, 0x3f800000 },
+	{ 0x000520,   1, 0x01, 0x000002b6 },
+	{ 0x000529,   1, 0x01, 0x00000001 },
+	{ 0x000530,  16, 0x01, 0xffff0000 },
+	{ 0x000585,   1, 0x01, 0x0000003f },
+	{ 0x000576,   1, 0x01, 0x00000003 },
+	{ 0x00057b,   1, 0x01, 0x00000059 },
+	{ 0x000586,   1, 0x01, 0x00000040 },
+	{ 0x000582,   2, 0x01, 0x00000080 },
+	{ 0x0005c2,   1, 0x01, 0x00000001 },
+	{ 0x000638,   1, 0x01, 0x00000001 },
+	{ 0x000639,   1, 0x01, 0x00000001 },
+	{ 0x00063a,   1, 0x01, 0x00000002 },
+	{ 0x00063b,   2, 0x01, 0x00000001 },
+	{ 0x00063d,   1, 0x01, 0x00000002 },
+	{ 0x00063e,   1, 0x01, 0x00000001 },
+	{ 0x0008b8,   8, 0x01, 0x00000001 },
+	{ 0x000900,   8, 0x01, 0x00000001 },
+	{ 0x000908,   8, 0x01, 0x00000002 },
+	{ 0x000910,  16, 0x01, 0x00000001 },
+	{ 0x000920,   8, 0x01, 0x00000002 },
+	{ 0x000928,   8, 0x01, 0x00000001 },
+	{ 0x000648,   9, 0x01, 0x00000001 },
+	{ 0x000658,   1, 0x01, 0x0000000f },
+	{ 0x0007ff,   1, 0x01, 0x0000000a },
+	{ 0x00066a,   1, 0x01, 0x40000000 },
+	{ 0x00066b,   1, 0x01, 0x10000000 },
+	{ 0x00066c,   2, 0x01, 0xffff0000 },
+	{ 0x0007af,   2, 0x01, 0x00000008 },
+	{ 0x0007f6,   1, 0x01, 0x00000001 },
+	{ 0x0006b2,   1, 0x01, 0x00000055 },
+	{ 0x0007ad,   1, 0x01, 0x00000003 },
+	{ 0x000937,   1, 0x01, 0x00000001 },
+	{ 0x000971,   1, 0x01, 0x00000008 },
+	{ 0x000972,   1, 0x01, 0x00000040 },
+	{ 0x000973,   1, 0x01, 0x0000012c },
+	{ 0x00097c,   1, 0x01, 0x00000040 },
+	{ 0x000979,   1, 0x01, 0x00000003 },
+	{ 0x000975,   1, 0x01, 0x00000020 },
+	{ 0x000976,   1, 0x01, 0x00000001 },
+	{ 0x000977,   1, 0x01, 0x00000020 },
+	{ 0x000978,   1, 0x01, 0x00000001 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095e,   1, 0x01, 0x20164010 },
+	{ 0x00095f,   1, 0x01, 0x00000020 },
+	{ 0x00097d,   1, 0x01, 0x00000020 },
+	{ 0x000683,   1, 0x01, 0x00000006 },
+	{ 0x000685,   1, 0x01, 0x003fffff },
+	{ 0x000687,   1, 0x01, 0x00000c48 },
+	{ 0x0006a0,   1, 0x01, 0x00000005 },
+	{ 0x000840,   1, 0x01, 0x00300008 },
+	{ 0x000841,   1, 0x01, 0x04000080 },
+	{ 0x000842,   1, 0x01, 0x00300008 },
+	{ 0x000843,   1, 0x01, 0x04000080 },
+	{ 0x000818,   8, 0x01, 0x00000000 },
+	{ 0x000848,  16, 0x01, 0x00000000 },
+	{ 0x000738,   1, 0x01, 0x00000000 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ab,   1, 0x01, 0x00000002 },
+	{ 0x0006ac,   1, 0x01, 0x00000080 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x0006bb,   1, 0x01, 0x000000cf },
+	{ 0x0006ce,   1, 0x01, 0x2a712488 },
+	{ 0x000739,   1, 0x01, 0x4085c000 },
+	{ 0x00073a,   1, 0x01, 0x00000080 },
+	{ 0x000786,   1, 0x01, 0x80000100 },
+	{ 0x00073c,   1, 0x01, 0x00010100 },
+	{ 0x00073d,   1, 0x01, 0x02800000 },
+	{ 0x000787,   1, 0x01, 0x000000cf },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x000836,   1, 0x01, 0x00000001 },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x000944,   1, 0x01, 0x00000022 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x00c1b0,   8, 0x01, 0x0000000f },
+	{ 0x00c1b8,   1, 0x01, 0x0fac6881 },
+	{ 0x00c1b9,   1, 0x01, 0x00fac688 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000002 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000014 },
+	{ 0x000351,   1, 0x01, 0x00000100 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095d,   1, 0x01, 0x00000001 },
+	{ 0x00082b,   1, 0x01, 0x00000004 },
+	{ 0x000942,   1, 0x01, 0x00010001 },
+	{ 0x000943,   1, 0x01, 0x00000001 },
+	{ 0x0007c5,   1, 0x01, 0x00010001 },
+	{ 0x000834,   1, 0x01, 0x00000001 },
+	{ 0x0007c7,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000001 },
+	{ 0x00080c,   1, 0x01, 0x00000002 },
+	{ 0x00080d,   2, 0x01, 0x00000100 },
+	{ 0x00080f,   1, 0x01, 0x00000001 },
+	{ 0x000823,   1, 0x01, 0x00000002 },
+	{ 0x000824,   2, 0x01, 0x00000100 },
+	{ 0x000826,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{}
+};
+
+struct nvc0_graph_init
+nvd9_grctx_init_unk40xx[] = {
+	{ 0x404004,  11, 0x04, 0x00000000 },
+	{ 0x404044,   1, 0x04, 0x00000000 },
+	{ 0x404094,   1, 0x04, 0x00000000 },
+	{ 0x404098,  12, 0x04, 0x00000000 },
+	{ 0x4040c8,   1, 0x04, 0xf0000087 },
+	{ 0x4040d0,   6, 0x04, 0x00000000 },
+	{ 0x4040e8,   1, 0x04, 0x00001000 },
+	{ 0x4040f8,   1, 0x04, 0x00000000 },
+	{ 0x404130,   1, 0x04, 0x00000000 },
+	{ 0x404134,   1, 0x04, 0x00000000 },
+	{ 0x404138,   1, 0x04, 0x20000040 },
+	{ 0x404150,   1, 0x04, 0x0000002e },
+	{ 0x404154,   1, 0x04, 0x00000400 },
+	{ 0x404158,   1, 0x04, 0x00000200 },
+	{ 0x404164,   1, 0x04, 0x00000055 },
+	{ 0x404168,   1, 0x04, 0x00000000 },
+	{ 0x404178,   2, 0x04, 0x00000000 },
+	{ 0x404200,   8, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd9_grctx_init_unk58xx[] = {
+	{ 0x405800,   1, 0x04, 0x0f8000bf },
+	{ 0x405830,   1, 0x04, 0x02180218 },
+	{ 0x405834,   1, 0x04, 0x08000000 },
+	{ 0x405838,   1, 0x04, 0x00000000 },
+	{ 0x405854,   1, 0x04, 0x00000000 },
+	{ 0x405870,   4, 0x04, 0x00000001 },
+	{ 0x405a00,   2, 0x04, 0x00000000 },
+	{ 0x405a18,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd9_grctx_init_unk64xx[] = {
+	{ 0x4064a8,   1, 0x04, 0x00000000 },
+	{ 0x4064ac,   1, 0x04, 0x00003fff },
+	{ 0x4064b4,   3, 0x04, 0x00000000 },
+	{ 0x4064c0,   1, 0x04, 0x80140078 },
+	{ 0x4064c4,   1, 0x04, 0x0086ffff },
+	{}
+};
+
+struct nvc0_graph_init
+nvd9_grctx_init_rop[] = {
+	{ 0x408800,   1, 0x04, 0x02802a3c },
+	{ 0x408804,   1, 0x04, 0x00000040 },
+	{ 0x408808,   1, 0x04, 0x1043e005 },
+	{ 0x408900,   1, 0x04, 0x3080b801 },
+	{ 0x408904,   1, 0x04, 0x1043e005 },
+	{ 0x408908,   1, 0x04, 0x00c8102f },
+	{ 0x408980,   1, 0x04, 0x0000011d },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd9_grctx_init_gpc_0[] = {
+	{ 0x418380,   1, 0x04, 0x00000016 },
+	{ 0x418400,   1, 0x04, 0x38004e00 },
+	{ 0x418404,   1, 0x04, 0x71e0ffff },
+	{ 0x41840c,   1, 0x04, 0x00001008 },
+	{ 0x418410,   1, 0x04, 0x0fff0fff },
+	{ 0x418414,   1, 0x04, 0x02200fff },
+	{ 0x418450,   6, 0x04, 0x00000000 },
+	{ 0x418468,   1, 0x04, 0x00000001 },
+	{ 0x41846c,   2, 0x04, 0x00000000 },
+	{ 0x418600,   1, 0x04, 0x0000001f },
+	{ 0x418684,   1, 0x04, 0x0000000f },
+	{ 0x418700,   1, 0x04, 0x00000002 },
+	{ 0x418704,   1, 0x04, 0x00000080 },
+	{ 0x418708,   3, 0x04, 0x00000000 },
+	{ 0x418800,   1, 0x04, 0x7006860a },
+	{ 0x418808,   3, 0x04, 0x00000000 },
+	{ 0x418828,   1, 0x04, 0x00008442 },
+	{ 0x418830,   1, 0x04, 0x10000001 },
+	{ 0x4188d8,   1, 0x04, 0x00000008 },
+	{ 0x4188e0,   1, 0x04, 0x01000000 },
+	{ 0x4188e8,   5, 0x04, 0x00000000 },
+	{ 0x4188fc,   1, 0x04, 0x20100008 },
+	{ 0x41891c,   1, 0x04, 0x00ff00ff },
+	{ 0x418924,   1, 0x04, 0x00000000 },
+	{ 0x418928,   1, 0x04, 0x00ffff00 },
+	{ 0x41892c,   1, 0x04, 0x0000ff00 },
+	{ 0x418b00,   1, 0x04, 0x00000006 },
+	{ 0x418b08,   1, 0x04, 0x0a418820 },
+	{ 0x418b0c,   1, 0x04, 0x062080e6 },
+	{ 0x418b10,   1, 0x04, 0x020398a4 },
+	{ 0x418b14,   1, 0x04, 0x0e629062 },
+	{ 0x418b18,   1, 0x04, 0x0a418820 },
+	{ 0x418b1c,   1, 0x04, 0x000000e6 },
+	{ 0x418bb8,   1, 0x04, 0x00000103 },
+	{ 0x418c08,   1, 0x04, 0x00000001 },
+	{ 0x418c10,   8, 0x04, 0x00000000 },
+	{ 0x418c6c,   1, 0x04, 0x00000001 },
+	{ 0x418c80,   1, 0x04, 0x20200004 },
+	{ 0x418c8c,   1, 0x04, 0x00000001 },
+	{ 0x419000,   1, 0x04, 0x00000780 },
+	{ 0x419004,   2, 0x04, 0x00000000 },
+	{ 0x419014,   1, 0x04, 0x00000004 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd9_grctx_init_tpc[] = {
+	{ 0x419818,   1, 0x04, 0x00000000 },
+	{ 0x41983c,   1, 0x04, 0x00038bc7 },
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x00000129 },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000001f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000023 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x00000000 },
+	{ 0x419a20,   1, 0x04, 0x00000800 },
+	{ 0x419ac4,   1, 0x04, 0x0017f440 },
+	{ 0x419b00,   1, 0x04, 0x0a418820 },
+	{ 0x419b04,   1, 0x04, 0x062080e6 },
+	{ 0x419b08,   1, 0x04, 0x020398a4 },
+	{ 0x419b0c,   1, 0x04, 0x0e629062 },
+	{ 0x419b10,   1, 0x04, 0x0a418820 },
+	{ 0x419b14,   1, 0x04, 0x000000e6 },
+	{ 0x419bd0,   1, 0x04, 0x00900103 },
+	{ 0x419be0,   1, 0x04, 0x00400001 },
+	{ 0x419be4,   1, 0x04, 0x00000000 },
+	{ 0x419c00,   1, 0x04, 0x0000000a },
+	{ 0x419c04,   1, 0x04, 0x00000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419c24,   1, 0x04, 0x00084210 },
+	{ 0x419c28,   1, 0x04, 0x3cf3cf3c },
+	{ 0x419cb0,   1, 0x04, 0x00020048 },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000183 },
+	{ 0x419d20,   1, 0x04, 0x12180000 },
+	{ 0x419d24,   1, 0x04, 0x00001fff },
+	{ 0x419d44,   1, 0x04, 0x02180218 },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000002 },
+	{ 0x419e44,   1, 0x04, 0x001beff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000000f },
+	{ 0x419e50,  17, 0x04, 0x00000000 },
+	{ 0x419e98,   1, 0x04, 0x00000000 },
+	{ 0x419ee0,   1, 0x04, 0x00010110 },
+	{ 0x419f30,  11, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init *
+nvd9_grctx_init_hub[] = {
+	nvc0_grctx_init_base,
+	nvd9_grctx_init_unk40xx,
+	nvc0_grctx_init_unk44xx,
+	nvc0_grctx_init_unk46xx,
+	nvc0_grctx_init_unk47xx,
+	nvd9_grctx_init_unk58xx,
+	nvc0_grctx_init_unk60xx,
+	nvd9_grctx_init_unk64xx,
+	nvc0_grctx_init_unk78xx,
+	nvc0_grctx_init_unk80xx,
+	nvd9_grctx_init_rop,
+};
+
+struct nvc0_graph_init *
+nvd9_grctx_init_gpc[] = {
+	nvd9_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvd9_grctx_init_tpc,
+	NULL
+};
+
+struct nvc0_graph_init
+nvd9_grctx_init_mthd_magic[] = {
+	{ 0x3410, 1, 0x04, 0x80002006 },
+	{}
+};
+
+struct nvc0_graph_mthd
+nvd9_grctx_init_mthd[] = {
+	{ 0x9097, nvc1_grctx_init_9097, },
+	{ 0x9197, nvc8_grctx_init_9197, },
+	{ 0x9297, nvc8_grctx_init_9297, },
+	{ 0x902d, nvc0_grctx_init_902d, },
+	{ 0x9039, nvc0_grctx_init_9039, },
+	{ 0x90c0, nvd9_grctx_init_90c0, },
+	{ 0x902d, nvd9_grctx_init_mthd_magic, },
+	{}
+};
+
+struct nouveau_oclass *
+nvd9_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xd9),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nvc0_grctx_generate_main,
+	.mods = nvc1_grctx_generate_mods,
+	.unkn = nvc1_grctx_generate_unkn,
+	.hub  = nvd9_grctx_init_hub,
+	.gpc  = nvd9_grctx_init_gpc,
+	.icmd = nvd9_grctx_init_icmd,
+	.mthd = nvd9_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve0.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve0.c
deleted file mode 100644
index ae27dae3fe38..000000000000
--- a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve0.c
+++ /dev/null
@@ -1,2793 +0,0 @@
-/*
- * Copyright 2010 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "nvc0.h"
-
-static void
-nve0_grctx_generate_icmd(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x400208, 0x80000000);
-	nv_icmd(priv, 0x001000, 0x00000004);
-	nv_icmd(priv, 0x000039, 0x00000000);
-	nv_icmd(priv, 0x00003a, 0x00000000);
-	nv_icmd(priv, 0x00003b, 0x00000000);
-	nv_icmd(priv, 0x0000a9, 0x0000ffff);
-	nv_icmd(priv, 0x000038, 0x0fac6881);
-	nv_icmd(priv, 0x00003d, 0x00000001);
-	nv_icmd(priv, 0x0000e8, 0x00000400);
-	nv_icmd(priv, 0x0000e9, 0x00000400);
-	nv_icmd(priv, 0x0000ea, 0x00000400);
-	nv_icmd(priv, 0x0000eb, 0x00000400);
-	nv_icmd(priv, 0x0000ec, 0x00000400);
-	nv_icmd(priv, 0x0000ed, 0x00000400);
-	nv_icmd(priv, 0x0000ee, 0x00000400);
-	nv_icmd(priv, 0x0000ef, 0x00000400);
-	nv_icmd(priv, 0x000078, 0x00000300);
-	nv_icmd(priv, 0x000079, 0x00000300);
-	nv_icmd(priv, 0x00007a, 0x00000300);
-	nv_icmd(priv, 0x00007b, 0x00000300);
-	nv_icmd(priv, 0x00007c, 0x00000300);
-	nv_icmd(priv, 0x00007d, 0x00000300);
-	nv_icmd(priv, 0x00007e, 0x00000300);
-	nv_icmd(priv, 0x00007f, 0x00000300);
-	nv_icmd(priv, 0x000050, 0x00000011);
-	nv_icmd(priv, 0x000058, 0x00000008);
-	nv_icmd(priv, 0x000059, 0x00000008);
-	nv_icmd(priv, 0x00005a, 0x00000008);
-	nv_icmd(priv, 0x00005b, 0x00000008);
-	nv_icmd(priv, 0x00005c, 0x00000008);
-	nv_icmd(priv, 0x00005d, 0x00000008);
-	nv_icmd(priv, 0x00005e, 0x00000008);
-	nv_icmd(priv, 0x00005f, 0x00000008);
-	nv_icmd(priv, 0x000208, 0x00000001);
-	nv_icmd(priv, 0x000209, 0x00000001);
-	nv_icmd(priv, 0x00020a, 0x00000001);
-	nv_icmd(priv, 0x00020b, 0x00000001);
-	nv_icmd(priv, 0x00020c, 0x00000001);
-	nv_icmd(priv, 0x00020d, 0x00000001);
-	nv_icmd(priv, 0x00020e, 0x00000001);
-	nv_icmd(priv, 0x00020f, 0x00000001);
-	nv_icmd(priv, 0x000081, 0x00000001);
-	nv_icmd(priv, 0x000085, 0x00000004);
-	nv_icmd(priv, 0x000088, 0x00000400);
-	nv_icmd(priv, 0x000090, 0x00000300);
-	nv_icmd(priv, 0x000098, 0x00001001);
-	nv_icmd(priv, 0x0000e3, 0x00000001);
-	nv_icmd(priv, 0x0000da, 0x00000001);
-	nv_icmd(priv, 0x0000f8, 0x00000003);
-	nv_icmd(priv, 0x0000fa, 0x00000001);
-	nv_icmd(priv, 0x00009f, 0x0000ffff);
-	nv_icmd(priv, 0x0000a0, 0x0000ffff);
-	nv_icmd(priv, 0x0000a1, 0x0000ffff);
-	nv_icmd(priv, 0x0000a2, 0x0000ffff);
-	nv_icmd(priv, 0x0000b1, 0x00000001);
-	nv_icmd(priv, 0x0000ad, 0x0000013e);
-	nv_icmd(priv, 0x0000e1, 0x00000010);
-	nv_icmd(priv, 0x000290, 0x00000000);
-	nv_icmd(priv, 0x000291, 0x00000000);
-	nv_icmd(priv, 0x000292, 0x00000000);
-	nv_icmd(priv, 0x000293, 0x00000000);
-	nv_icmd(priv, 0x000294, 0x00000000);
-	nv_icmd(priv, 0x000295, 0x00000000);
-	nv_icmd(priv, 0x000296, 0x00000000);
-	nv_icmd(priv, 0x000297, 0x00000000);
-	nv_icmd(priv, 0x000298, 0x00000000);
-	nv_icmd(priv, 0x000299, 0x00000000);
-	nv_icmd(priv, 0x00029a, 0x00000000);
-	nv_icmd(priv, 0x00029b, 0x00000000);
-	nv_icmd(priv, 0x00029c, 0x00000000);
-	nv_icmd(priv, 0x00029d, 0x00000000);
-	nv_icmd(priv, 0x00029e, 0x00000000);
-	nv_icmd(priv, 0x00029f, 0x00000000);
-	nv_icmd(priv, 0x0003b0, 0x00000000);
-	nv_icmd(priv, 0x0003b1, 0x00000000);
-	nv_icmd(priv, 0x0003b2, 0x00000000);
-	nv_icmd(priv, 0x0003b3, 0x00000000);
-	nv_icmd(priv, 0x0003b4, 0x00000000);
-	nv_icmd(priv, 0x0003b5, 0x00000000);
-	nv_icmd(priv, 0x0003b6, 0x00000000);
-	nv_icmd(priv, 0x0003b7, 0x00000000);
-	nv_icmd(priv, 0x0003b8, 0x00000000);
-	nv_icmd(priv, 0x0003b9, 0x00000000);
-	nv_icmd(priv, 0x0003ba, 0x00000000);
-	nv_icmd(priv, 0x0003bb, 0x00000000);
-	nv_icmd(priv, 0x0003bc, 0x00000000);
-	nv_icmd(priv, 0x0003bd, 0x00000000);
-	nv_icmd(priv, 0x0003be, 0x00000000);
-	nv_icmd(priv, 0x0003bf, 0x00000000);
-	nv_icmd(priv, 0x0002a0, 0x00000000);
-	nv_icmd(priv, 0x0002a1, 0x00000000);
-	nv_icmd(priv, 0x0002a2, 0x00000000);
-	nv_icmd(priv, 0x0002a3, 0x00000000);
-	nv_icmd(priv, 0x0002a4, 0x00000000);
-	nv_icmd(priv, 0x0002a5, 0x00000000);
-	nv_icmd(priv, 0x0002a6, 0x00000000);
-	nv_icmd(priv, 0x0002a7, 0x00000000);
-	nv_icmd(priv, 0x0002a8, 0x00000000);
-	nv_icmd(priv, 0x0002a9, 0x00000000);
-	nv_icmd(priv, 0x0002aa, 0x00000000);
-	nv_icmd(priv, 0x0002ab, 0x00000000);
-	nv_icmd(priv, 0x0002ac, 0x00000000);
-	nv_icmd(priv, 0x0002ad, 0x00000000);
-	nv_icmd(priv, 0x0002ae, 0x00000000);
-	nv_icmd(priv, 0x0002af, 0x00000000);
-	nv_icmd(priv, 0x000420, 0x00000000);
-	nv_icmd(priv, 0x000421, 0x00000000);
-	nv_icmd(priv, 0x000422, 0x00000000);
-	nv_icmd(priv, 0x000423, 0x00000000);
-	nv_icmd(priv, 0x000424, 0x00000000);
-	nv_icmd(priv, 0x000425, 0x00000000);
-	nv_icmd(priv, 0x000426, 0x00000000);
-	nv_icmd(priv, 0x000427, 0x00000000);
-	nv_icmd(priv, 0x000428, 0x00000000);
-	nv_icmd(priv, 0x000429, 0x00000000);
-	nv_icmd(priv, 0x00042a, 0x00000000);
-	nv_icmd(priv, 0x00042b, 0x00000000);
-	nv_icmd(priv, 0x00042c, 0x00000000);
-	nv_icmd(priv, 0x00042d, 0x00000000);
-	nv_icmd(priv, 0x00042e, 0x00000000);
-	nv_icmd(priv, 0x00042f, 0x00000000);
-	nv_icmd(priv, 0x0002b0, 0x00000000);
-	nv_icmd(priv, 0x0002b1, 0x00000000);
-	nv_icmd(priv, 0x0002b2, 0x00000000);
-	nv_icmd(priv, 0x0002b3, 0x00000000);
-	nv_icmd(priv, 0x0002b4, 0x00000000);
-	nv_icmd(priv, 0x0002b5, 0x00000000);
-	nv_icmd(priv, 0x0002b6, 0x00000000);
-	nv_icmd(priv, 0x0002b7, 0x00000000);
-	nv_icmd(priv, 0x0002b8, 0x00000000);
-	nv_icmd(priv, 0x0002b9, 0x00000000);
-	nv_icmd(priv, 0x0002ba, 0x00000000);
-	nv_icmd(priv, 0x0002bb, 0x00000000);
-	nv_icmd(priv, 0x0002bc, 0x00000000);
-	nv_icmd(priv, 0x0002bd, 0x00000000);
-	nv_icmd(priv, 0x0002be, 0x00000000);
-	nv_icmd(priv, 0x0002bf, 0x00000000);
-	nv_icmd(priv, 0x000430, 0x00000000);
-	nv_icmd(priv, 0x000431, 0x00000000);
-	nv_icmd(priv, 0x000432, 0x00000000);
-	nv_icmd(priv, 0x000433, 0x00000000);
-	nv_icmd(priv, 0x000434, 0x00000000);
-	nv_icmd(priv, 0x000435, 0x00000000);
-	nv_icmd(priv, 0x000436, 0x00000000);
-	nv_icmd(priv, 0x000437, 0x00000000);
-	nv_icmd(priv, 0x000438, 0x00000000);
-	nv_icmd(priv, 0x000439, 0x00000000);
-	nv_icmd(priv, 0x00043a, 0x00000000);
-	nv_icmd(priv, 0x00043b, 0x00000000);
-	nv_icmd(priv, 0x00043c, 0x00000000);
-	nv_icmd(priv, 0x00043d, 0x00000000);
-	nv_icmd(priv, 0x00043e, 0x00000000);
-	nv_icmd(priv, 0x00043f, 0x00000000);
-	nv_icmd(priv, 0x0002c0, 0x00000000);
-	nv_icmd(priv, 0x0002c1, 0x00000000);
-	nv_icmd(priv, 0x0002c2, 0x00000000);
-	nv_icmd(priv, 0x0002c3, 0x00000000);
-	nv_icmd(priv, 0x0002c4, 0x00000000);
-	nv_icmd(priv, 0x0002c5, 0x00000000);
-	nv_icmd(priv, 0x0002c6, 0x00000000);
-	nv_icmd(priv, 0x0002c7, 0x00000000);
-	nv_icmd(priv, 0x0002c8, 0x00000000);
-	nv_icmd(priv, 0x0002c9, 0x00000000);
-	nv_icmd(priv, 0x0002ca, 0x00000000);
-	nv_icmd(priv, 0x0002cb, 0x00000000);
-	nv_icmd(priv, 0x0002cc, 0x00000000);
-	nv_icmd(priv, 0x0002cd, 0x00000000);
-	nv_icmd(priv, 0x0002ce, 0x00000000);
-	nv_icmd(priv, 0x0002cf, 0x00000000);
-	nv_icmd(priv, 0x0004d0, 0x00000000);
-	nv_icmd(priv, 0x0004d1, 0x00000000);
-	nv_icmd(priv, 0x0004d2, 0x00000000);
-	nv_icmd(priv, 0x0004d3, 0x00000000);
-	nv_icmd(priv, 0x0004d4, 0x00000000);
-	nv_icmd(priv, 0x0004d5, 0x00000000);
-	nv_icmd(priv, 0x0004d6, 0x00000000);
-	nv_icmd(priv, 0x0004d7, 0x00000000);
-	nv_icmd(priv, 0x0004d8, 0x00000000);
-	nv_icmd(priv, 0x0004d9, 0x00000000);
-	nv_icmd(priv, 0x0004da, 0x00000000);
-	nv_icmd(priv, 0x0004db, 0x00000000);
-	nv_icmd(priv, 0x0004dc, 0x00000000);
-	nv_icmd(priv, 0x0004dd, 0x00000000);
-	nv_icmd(priv, 0x0004de, 0x00000000);
-	nv_icmd(priv, 0x0004df, 0x00000000);
-	nv_icmd(priv, 0x000720, 0x00000000);
-	nv_icmd(priv, 0x000721, 0x00000000);
-	nv_icmd(priv, 0x000722, 0x00000000);
-	nv_icmd(priv, 0x000723, 0x00000000);
-	nv_icmd(priv, 0x000724, 0x00000000);
-	nv_icmd(priv, 0x000725, 0x00000000);
-	nv_icmd(priv, 0x000726, 0x00000000);
-	nv_icmd(priv, 0x000727, 0x00000000);
-	nv_icmd(priv, 0x000728, 0x00000000);
-	nv_icmd(priv, 0x000729, 0x00000000);
-	nv_icmd(priv, 0x00072a, 0x00000000);
-	nv_icmd(priv, 0x00072b, 0x00000000);
-	nv_icmd(priv, 0x00072c, 0x00000000);
-	nv_icmd(priv, 0x00072d, 0x00000000);
-	nv_icmd(priv, 0x00072e, 0x00000000);
-	nv_icmd(priv, 0x00072f, 0x00000000);
-	nv_icmd(priv, 0x0008c0, 0x00000000);
-	nv_icmd(priv, 0x0008c1, 0x00000000);
-	nv_icmd(priv, 0x0008c2, 0x00000000);
-	nv_icmd(priv, 0x0008c3, 0x00000000);
-	nv_icmd(priv, 0x0008c4, 0x00000000);
-	nv_icmd(priv, 0x0008c5, 0x00000000);
-	nv_icmd(priv, 0x0008c6, 0x00000000);
-	nv_icmd(priv, 0x0008c7, 0x00000000);
-	nv_icmd(priv, 0x0008c8, 0x00000000);
-	nv_icmd(priv, 0x0008c9, 0x00000000);
-	nv_icmd(priv, 0x0008ca, 0x00000000);
-	nv_icmd(priv, 0x0008cb, 0x00000000);
-	nv_icmd(priv, 0x0008cc, 0x00000000);
-	nv_icmd(priv, 0x0008cd, 0x00000000);
-	nv_icmd(priv, 0x0008ce, 0x00000000);
-	nv_icmd(priv, 0x0008cf, 0x00000000);
-	nv_icmd(priv, 0x000890, 0x00000000);
-	nv_icmd(priv, 0x000891, 0x00000000);
-	nv_icmd(priv, 0x000892, 0x00000000);
-	nv_icmd(priv, 0x000893, 0x00000000);
-	nv_icmd(priv, 0x000894, 0x00000000);
-	nv_icmd(priv, 0x000895, 0x00000000);
-	nv_icmd(priv, 0x000896, 0x00000000);
-	nv_icmd(priv, 0x000897, 0x00000000);
-	nv_icmd(priv, 0x000898, 0x00000000);
-	nv_icmd(priv, 0x000899, 0x00000000);
-	nv_icmd(priv, 0x00089a, 0x00000000);
-	nv_icmd(priv, 0x00089b, 0x00000000);
-	nv_icmd(priv, 0x00089c, 0x00000000);
-	nv_icmd(priv, 0x00089d, 0x00000000);
-	nv_icmd(priv, 0x00089e, 0x00000000);
-	nv_icmd(priv, 0x00089f, 0x00000000);
-	nv_icmd(priv, 0x0008e0, 0x00000000);
-	nv_icmd(priv, 0x0008e1, 0x00000000);
-	nv_icmd(priv, 0x0008e2, 0x00000000);
-	nv_icmd(priv, 0x0008e3, 0x00000000);
-	nv_icmd(priv, 0x0008e4, 0x00000000);
-	nv_icmd(priv, 0x0008e5, 0x00000000);
-	nv_icmd(priv, 0x0008e6, 0x00000000);
-	nv_icmd(priv, 0x0008e7, 0x00000000);
-	nv_icmd(priv, 0x0008e8, 0x00000000);
-	nv_icmd(priv, 0x0008e9, 0x00000000);
-	nv_icmd(priv, 0x0008ea, 0x00000000);
-	nv_icmd(priv, 0x0008eb, 0x00000000);
-	nv_icmd(priv, 0x0008ec, 0x00000000);
-	nv_icmd(priv, 0x0008ed, 0x00000000);
-	nv_icmd(priv, 0x0008ee, 0x00000000);
-	nv_icmd(priv, 0x0008ef, 0x00000000);
-	nv_icmd(priv, 0x0008a0, 0x00000000);
-	nv_icmd(priv, 0x0008a1, 0x00000000);
-	nv_icmd(priv, 0x0008a2, 0x00000000);
-	nv_icmd(priv, 0x0008a3, 0x00000000);
-	nv_icmd(priv, 0x0008a4, 0x00000000);
-	nv_icmd(priv, 0x0008a5, 0x00000000);
-	nv_icmd(priv, 0x0008a6, 0x00000000);
-	nv_icmd(priv, 0x0008a7, 0x00000000);
-	nv_icmd(priv, 0x0008a8, 0x00000000);
-	nv_icmd(priv, 0x0008a9, 0x00000000);
-	nv_icmd(priv, 0x0008aa, 0x00000000);
-	nv_icmd(priv, 0x0008ab, 0x00000000);
-	nv_icmd(priv, 0x0008ac, 0x00000000);
-	nv_icmd(priv, 0x0008ad, 0x00000000);
-	nv_icmd(priv, 0x0008ae, 0x00000000);
-	nv_icmd(priv, 0x0008af, 0x00000000);
-	nv_icmd(priv, 0x0008f0, 0x00000000);
-	nv_icmd(priv, 0x0008f1, 0x00000000);
-	nv_icmd(priv, 0x0008f2, 0x00000000);
-	nv_icmd(priv, 0x0008f3, 0x00000000);
-	nv_icmd(priv, 0x0008f4, 0x00000000);
-	nv_icmd(priv, 0x0008f5, 0x00000000);
-	nv_icmd(priv, 0x0008f6, 0x00000000);
-	nv_icmd(priv, 0x0008f7, 0x00000000);
-	nv_icmd(priv, 0x0008f8, 0x00000000);
-	nv_icmd(priv, 0x0008f9, 0x00000000);
-	nv_icmd(priv, 0x0008fa, 0x00000000);
-	nv_icmd(priv, 0x0008fb, 0x00000000);
-	nv_icmd(priv, 0x0008fc, 0x00000000);
-	nv_icmd(priv, 0x0008fd, 0x00000000);
-	nv_icmd(priv, 0x0008fe, 0x00000000);
-	nv_icmd(priv, 0x0008ff, 0x00000000);
-	nv_icmd(priv, 0x00094c, 0x000000ff);
-	nv_icmd(priv, 0x00094d, 0xffffffff);
-	nv_icmd(priv, 0x00094e, 0x00000002);
-	nv_icmd(priv, 0x0002ec, 0x00000001);
-	nv_icmd(priv, 0x000303, 0x00000001);
-	nv_icmd(priv, 0x0002e6, 0x00000001);
-	nv_icmd(priv, 0x000466, 0x00000052);
-	nv_icmd(priv, 0x000301, 0x3f800000);
-	nv_icmd(priv, 0x000304, 0x30201000);
-	nv_icmd(priv, 0x000305, 0x70605040);
-	nv_icmd(priv, 0x000306, 0xb8a89888);
-	nv_icmd(priv, 0x000307, 0xf8e8d8c8);
-	nv_icmd(priv, 0x00030a, 0x00ffff00);
-	nv_icmd(priv, 0x00030b, 0x0000001a);
-	nv_icmd(priv, 0x00030c, 0x00000001);
-	nv_icmd(priv, 0x000318, 0x00000001);
-	nv_icmd(priv, 0x000340, 0x00000000);
-	nv_icmd(priv, 0x000375, 0x00000001);
-	nv_icmd(priv, 0x00037d, 0x00000006);
-	nv_icmd(priv, 0x0003a0, 0x00000002);
-	nv_icmd(priv, 0x0003aa, 0x00000001);
-	nv_icmd(priv, 0x0003a9, 0x00000001);
-	nv_icmd(priv, 0x000380, 0x00000001);
-	nv_icmd(priv, 0x000383, 0x00000011);
-	nv_icmd(priv, 0x000360, 0x00000040);
-	nv_icmd(priv, 0x000366, 0x00000000);
-	nv_icmd(priv, 0x000367, 0x00000000);
-	nv_icmd(priv, 0x000368, 0x00000fff);
-	nv_icmd(priv, 0x000370, 0x00000000);
-	nv_icmd(priv, 0x000371, 0x00000000);
-	nv_icmd(priv, 0x000372, 0x000fffff);
-	nv_icmd(priv, 0x00037a, 0x00000012);
-	nv_icmd(priv, 0x000619, 0x00000003);
-	nv_icmd(priv, 0x000811, 0x00000003);
-	nv_icmd(priv, 0x000812, 0x00000004);
-	nv_icmd(priv, 0x000813, 0x00000006);
-	nv_icmd(priv, 0x000814, 0x00000008);
-	nv_icmd(priv, 0x000815, 0x0000000b);
-	nv_icmd(priv, 0x000800, 0x00000001);
-	nv_icmd(priv, 0x000801, 0x00000001);
-	nv_icmd(priv, 0x000802, 0x00000001);
-	nv_icmd(priv, 0x000803, 0x00000001);
-	nv_icmd(priv, 0x000804, 0x00000001);
-	nv_icmd(priv, 0x000805, 0x00000001);
-	nv_icmd(priv, 0x000632, 0x00000001);
-	nv_icmd(priv, 0x000633, 0x00000002);
-	nv_icmd(priv, 0x000634, 0x00000003);
-	nv_icmd(priv, 0x000635, 0x00000004);
-	nv_icmd(priv, 0x000654, 0x3f800000);
-	nv_icmd(priv, 0x000657, 0x3f800000);
-	nv_icmd(priv, 0x000655, 0x3f800000);
-	nv_icmd(priv, 0x000656, 0x3f800000);
-	nv_icmd(priv, 0x0006cd, 0x3f800000);
-	nv_icmd(priv, 0x0007f5, 0x3f800000);
-	nv_icmd(priv, 0x0007dc, 0x39291909);
-	nv_icmd(priv, 0x0007dd, 0x79695949);
-	nv_icmd(priv, 0x0007de, 0xb9a99989);
-	nv_icmd(priv, 0x0007df, 0xf9e9d9c9);
-	nv_icmd(priv, 0x0007e8, 0x00003210);
-	nv_icmd(priv, 0x0007e9, 0x00007654);
-	nv_icmd(priv, 0x0007ea, 0x00000098);
-	nv_icmd(priv, 0x0007ec, 0x39291909);
-	nv_icmd(priv, 0x0007ed, 0x79695949);
-	nv_icmd(priv, 0x0007ee, 0xb9a99989);
-	nv_icmd(priv, 0x0007ef, 0xf9e9d9c9);
-	nv_icmd(priv, 0x0007f0, 0x00003210);
-	nv_icmd(priv, 0x0007f1, 0x00007654);
-	nv_icmd(priv, 0x0007f2, 0x00000098);
-	nv_icmd(priv, 0x0005a5, 0x00000001);
-	nv_icmd(priv, 0x000980, 0x00000000);
-	nv_icmd(priv, 0x000981, 0x00000000);
-	nv_icmd(priv, 0x000982, 0x00000000);
-	nv_icmd(priv, 0x000983, 0x00000000);
-	nv_icmd(priv, 0x000984, 0x00000000);
-	nv_icmd(priv, 0x000985, 0x00000000);
-	nv_icmd(priv, 0x000986, 0x00000000);
-	nv_icmd(priv, 0x000987, 0x00000000);
-	nv_icmd(priv, 0x000988, 0x00000000);
-	nv_icmd(priv, 0x000989, 0x00000000);
-	nv_icmd(priv, 0x00098a, 0x00000000);
-	nv_icmd(priv, 0x00098b, 0x00000000);
-	nv_icmd(priv, 0x00098c, 0x00000000);
-	nv_icmd(priv, 0x00098d, 0x00000000);
-	nv_icmd(priv, 0x00098e, 0x00000000);
-	nv_icmd(priv, 0x00098f, 0x00000000);
-	nv_icmd(priv, 0x000990, 0x00000000);
-	nv_icmd(priv, 0x000991, 0x00000000);
-	nv_icmd(priv, 0x000992, 0x00000000);
-	nv_icmd(priv, 0x000993, 0x00000000);
-	nv_icmd(priv, 0x000994, 0x00000000);
-	nv_icmd(priv, 0x000995, 0x00000000);
-	nv_icmd(priv, 0x000996, 0x00000000);
-	nv_icmd(priv, 0x000997, 0x00000000);
-	nv_icmd(priv, 0x000998, 0x00000000);
-	nv_icmd(priv, 0x000999, 0x00000000);
-	nv_icmd(priv, 0x00099a, 0x00000000);
-	nv_icmd(priv, 0x00099b, 0x00000000);
-	nv_icmd(priv, 0x00099c, 0x00000000);
-	nv_icmd(priv, 0x00099d, 0x00000000);
-	nv_icmd(priv, 0x00099e, 0x00000000);
-	nv_icmd(priv, 0x00099f, 0x00000000);
-	nv_icmd(priv, 0x0009a0, 0x00000000);
-	nv_icmd(priv, 0x0009a1, 0x00000000);
-	nv_icmd(priv, 0x0009a2, 0x00000000);
-	nv_icmd(priv, 0x0009a3, 0x00000000);
-	nv_icmd(priv, 0x0009a4, 0x00000000);
-	nv_icmd(priv, 0x0009a5, 0x00000000);
-	nv_icmd(priv, 0x0009a6, 0x00000000);
-	nv_icmd(priv, 0x0009a7, 0x00000000);
-	nv_icmd(priv, 0x0009a8, 0x00000000);
-	nv_icmd(priv, 0x0009a9, 0x00000000);
-	nv_icmd(priv, 0x0009aa, 0x00000000);
-	nv_icmd(priv, 0x0009ab, 0x00000000);
-	nv_icmd(priv, 0x0009ac, 0x00000000);
-	nv_icmd(priv, 0x0009ad, 0x00000000);
-	nv_icmd(priv, 0x0009ae, 0x00000000);
-	nv_icmd(priv, 0x0009af, 0x00000000);
-	nv_icmd(priv, 0x0009b0, 0x00000000);
-	nv_icmd(priv, 0x0009b1, 0x00000000);
-	nv_icmd(priv, 0x0009b2, 0x00000000);
-	nv_icmd(priv, 0x0009b3, 0x00000000);
-	nv_icmd(priv, 0x0009b4, 0x00000000);
-	nv_icmd(priv, 0x0009b5, 0x00000000);
-	nv_icmd(priv, 0x0009b6, 0x00000000);
-	nv_icmd(priv, 0x0009b7, 0x00000000);
-	nv_icmd(priv, 0x0009b8, 0x00000000);
-	nv_icmd(priv, 0x0009b9, 0x00000000);
-	nv_icmd(priv, 0x0009ba, 0x00000000);
-	nv_icmd(priv, 0x0009bb, 0x00000000);
-	nv_icmd(priv, 0x0009bc, 0x00000000);
-	nv_icmd(priv, 0x0009bd, 0x00000000);
-	nv_icmd(priv, 0x0009be, 0x00000000);
-	nv_icmd(priv, 0x0009bf, 0x00000000);
-	nv_icmd(priv, 0x0009c0, 0x00000000);
-	nv_icmd(priv, 0x0009c1, 0x00000000);
-	nv_icmd(priv, 0x0009c2, 0x00000000);
-	nv_icmd(priv, 0x0009c3, 0x00000000);
-	nv_icmd(priv, 0x0009c4, 0x00000000);
-	nv_icmd(priv, 0x0009c5, 0x00000000);
-	nv_icmd(priv, 0x0009c6, 0x00000000);
-	nv_icmd(priv, 0x0009c7, 0x00000000);
-	nv_icmd(priv, 0x0009c8, 0x00000000);
-	nv_icmd(priv, 0x0009c9, 0x00000000);
-	nv_icmd(priv, 0x0009ca, 0x00000000);
-	nv_icmd(priv, 0x0009cb, 0x00000000);
-	nv_icmd(priv, 0x0009cc, 0x00000000);
-	nv_icmd(priv, 0x0009cd, 0x00000000);
-	nv_icmd(priv, 0x0009ce, 0x00000000);
-	nv_icmd(priv, 0x0009cf, 0x00000000);
-	nv_icmd(priv, 0x0009d0, 0x00000000);
-	nv_icmd(priv, 0x0009d1, 0x00000000);
-	nv_icmd(priv, 0x0009d2, 0x00000000);
-	nv_icmd(priv, 0x0009d3, 0x00000000);
-	nv_icmd(priv, 0x0009d4, 0x00000000);
-	nv_icmd(priv, 0x0009d5, 0x00000000);
-	nv_icmd(priv, 0x0009d6, 0x00000000);
-	nv_icmd(priv, 0x0009d7, 0x00000000);
-	nv_icmd(priv, 0x0009d8, 0x00000000);
-	nv_icmd(priv, 0x0009d9, 0x00000000);
-	nv_icmd(priv, 0x0009da, 0x00000000);
-	nv_icmd(priv, 0x0009db, 0x00000000);
-	nv_icmd(priv, 0x0009dc, 0x00000000);
-	nv_icmd(priv, 0x0009dd, 0x00000000);
-	nv_icmd(priv, 0x0009de, 0x00000000);
-	nv_icmd(priv, 0x0009df, 0x00000000);
-	nv_icmd(priv, 0x0009e0, 0x00000000);
-	nv_icmd(priv, 0x0009e1, 0x00000000);
-	nv_icmd(priv, 0x0009e2, 0x00000000);
-	nv_icmd(priv, 0x0009e3, 0x00000000);
-	nv_icmd(priv, 0x0009e4, 0x00000000);
-	nv_icmd(priv, 0x0009e5, 0x00000000);
-	nv_icmd(priv, 0x0009e6, 0x00000000);
-	nv_icmd(priv, 0x0009e7, 0x00000000);
-	nv_icmd(priv, 0x0009e8, 0x00000000);
-	nv_icmd(priv, 0x0009e9, 0x00000000);
-	nv_icmd(priv, 0x0009ea, 0x00000000);
-	nv_icmd(priv, 0x0009eb, 0x00000000);
-	nv_icmd(priv, 0x0009ec, 0x00000000);
-	nv_icmd(priv, 0x0009ed, 0x00000000);
-	nv_icmd(priv, 0x0009ee, 0x00000000);
-	nv_icmd(priv, 0x0009ef, 0x00000000);
-	nv_icmd(priv, 0x0009f0, 0x00000000);
-	nv_icmd(priv, 0x0009f1, 0x00000000);
-	nv_icmd(priv, 0x0009f2, 0x00000000);
-	nv_icmd(priv, 0x0009f3, 0x00000000);
-	nv_icmd(priv, 0x0009f4, 0x00000000);
-	nv_icmd(priv, 0x0009f5, 0x00000000);
-	nv_icmd(priv, 0x0009f6, 0x00000000);
-	nv_icmd(priv, 0x0009f7, 0x00000000);
-	nv_icmd(priv, 0x0009f8, 0x00000000);
-	nv_icmd(priv, 0x0009f9, 0x00000000);
-	nv_icmd(priv, 0x0009fa, 0x00000000);
-	nv_icmd(priv, 0x0009fb, 0x00000000);
-	nv_icmd(priv, 0x0009fc, 0x00000000);
-	nv_icmd(priv, 0x0009fd, 0x00000000);
-	nv_icmd(priv, 0x0009fe, 0x00000000);
-	nv_icmd(priv, 0x0009ff, 0x00000000);
-	nv_icmd(priv, 0x000468, 0x00000004);
-	nv_icmd(priv, 0x00046c, 0x00000001);
-	nv_icmd(priv, 0x000470, 0x00000000);
-	nv_icmd(priv, 0x000471, 0x00000000);
-	nv_icmd(priv, 0x000472, 0x00000000);
-	nv_icmd(priv, 0x000473, 0x00000000);
-	nv_icmd(priv, 0x000474, 0x00000000);
-	nv_icmd(priv, 0x000475, 0x00000000);
-	nv_icmd(priv, 0x000476, 0x00000000);
-	nv_icmd(priv, 0x000477, 0x00000000);
-	nv_icmd(priv, 0x000478, 0x00000000);
-	nv_icmd(priv, 0x000479, 0x00000000);
-	nv_icmd(priv, 0x00047a, 0x00000000);
-	nv_icmd(priv, 0x00047b, 0x00000000);
-	nv_icmd(priv, 0x00047c, 0x00000000);
-	nv_icmd(priv, 0x00047d, 0x00000000);
-	nv_icmd(priv, 0x00047e, 0x00000000);
-	nv_icmd(priv, 0x00047f, 0x00000000);
-	nv_icmd(priv, 0x000480, 0x00000000);
-	nv_icmd(priv, 0x000481, 0x00000000);
-	nv_icmd(priv, 0x000482, 0x00000000);
-	nv_icmd(priv, 0x000483, 0x00000000);
-	nv_icmd(priv, 0x000484, 0x00000000);
-	nv_icmd(priv, 0x000485, 0x00000000);
-	nv_icmd(priv, 0x000486, 0x00000000);
-	nv_icmd(priv, 0x000487, 0x00000000);
-	nv_icmd(priv, 0x000488, 0x00000000);
-	nv_icmd(priv, 0x000489, 0x00000000);
-	nv_icmd(priv, 0x00048a, 0x00000000);
-	nv_icmd(priv, 0x00048b, 0x00000000);
-	nv_icmd(priv, 0x00048c, 0x00000000);
-	nv_icmd(priv, 0x00048d, 0x00000000);
-	nv_icmd(priv, 0x00048e, 0x00000000);
-	nv_icmd(priv, 0x00048f, 0x00000000);
-	nv_icmd(priv, 0x000490, 0x00000000);
-	nv_icmd(priv, 0x000491, 0x00000000);
-	nv_icmd(priv, 0x000492, 0x00000000);
-	nv_icmd(priv, 0x000493, 0x00000000);
-	nv_icmd(priv, 0x000494, 0x00000000);
-	nv_icmd(priv, 0x000495, 0x00000000);
-	nv_icmd(priv, 0x000496, 0x00000000);
-	nv_icmd(priv, 0x000497, 0x00000000);
-	nv_icmd(priv, 0x000498, 0x00000000);
-	nv_icmd(priv, 0x000499, 0x00000000);
-	nv_icmd(priv, 0x00049a, 0x00000000);
-	nv_icmd(priv, 0x00049b, 0x00000000);
-	nv_icmd(priv, 0x00049c, 0x00000000);
-	nv_icmd(priv, 0x00049d, 0x00000000);
-	nv_icmd(priv, 0x00049e, 0x00000000);
-	nv_icmd(priv, 0x00049f, 0x00000000);
-	nv_icmd(priv, 0x0004a0, 0x00000000);
-	nv_icmd(priv, 0x0004a1, 0x00000000);
-	nv_icmd(priv, 0x0004a2, 0x00000000);
-	nv_icmd(priv, 0x0004a3, 0x00000000);
-	nv_icmd(priv, 0x0004a4, 0x00000000);
-	nv_icmd(priv, 0x0004a5, 0x00000000);
-	nv_icmd(priv, 0x0004a6, 0x00000000);
-	nv_icmd(priv, 0x0004a7, 0x00000000);
-	nv_icmd(priv, 0x0004a8, 0x00000000);
-	nv_icmd(priv, 0x0004a9, 0x00000000);
-	nv_icmd(priv, 0x0004aa, 0x00000000);
-	nv_icmd(priv, 0x0004ab, 0x00000000);
-	nv_icmd(priv, 0x0004ac, 0x00000000);
-	nv_icmd(priv, 0x0004ad, 0x00000000);
-	nv_icmd(priv, 0x0004ae, 0x00000000);
-	nv_icmd(priv, 0x0004af, 0x00000000);
-	nv_icmd(priv, 0x0004b0, 0x00000000);
-	nv_icmd(priv, 0x0004b1, 0x00000000);
-	nv_icmd(priv, 0x0004b2, 0x00000000);
-	nv_icmd(priv, 0x0004b3, 0x00000000);
-	nv_icmd(priv, 0x0004b4, 0x00000000);
-	nv_icmd(priv, 0x0004b5, 0x00000000);
-	nv_icmd(priv, 0x0004b6, 0x00000000);
-	nv_icmd(priv, 0x0004b7, 0x00000000);
-	nv_icmd(priv, 0x0004b8, 0x00000000);
-	nv_icmd(priv, 0x0004b9, 0x00000000);
-	nv_icmd(priv, 0x0004ba, 0x00000000);
-	nv_icmd(priv, 0x0004bb, 0x00000000);
-	nv_icmd(priv, 0x0004bc, 0x00000000);
-	nv_icmd(priv, 0x0004bd, 0x00000000);
-	nv_icmd(priv, 0x0004be, 0x00000000);
-	nv_icmd(priv, 0x0004bf, 0x00000000);
-	nv_icmd(priv, 0x0004c0, 0x00000000);
-	nv_icmd(priv, 0x0004c1, 0x00000000);
-	nv_icmd(priv, 0x0004c2, 0x00000000);
-	nv_icmd(priv, 0x0004c3, 0x00000000);
-	nv_icmd(priv, 0x0004c4, 0x00000000);
-	nv_icmd(priv, 0x0004c5, 0x00000000);
-	nv_icmd(priv, 0x0004c6, 0x00000000);
-	nv_icmd(priv, 0x0004c7, 0x00000000);
-	nv_icmd(priv, 0x0004c8, 0x00000000);
-	nv_icmd(priv, 0x0004c9, 0x00000000);
-	nv_icmd(priv, 0x0004ca, 0x00000000);
-	nv_icmd(priv, 0x0004cb, 0x00000000);
-	nv_icmd(priv, 0x0004cc, 0x00000000);
-	nv_icmd(priv, 0x0004cd, 0x00000000);
-	nv_icmd(priv, 0x0004ce, 0x00000000);
-	nv_icmd(priv, 0x0004cf, 0x00000000);
-	nv_icmd(priv, 0x000510, 0x3f800000);
-	nv_icmd(priv, 0x000511, 0x3f800000);
-	nv_icmd(priv, 0x000512, 0x3f800000);
-	nv_icmd(priv, 0x000513, 0x3f800000);
-	nv_icmd(priv, 0x000514, 0x3f800000);
-	nv_icmd(priv, 0x000515, 0x3f800000);
-	nv_icmd(priv, 0x000516, 0x3f800000);
-	nv_icmd(priv, 0x000517, 0x3f800000);
-	nv_icmd(priv, 0x000518, 0x3f800000);
-	nv_icmd(priv, 0x000519, 0x3f800000);
-	nv_icmd(priv, 0x00051a, 0x3f800000);
-	nv_icmd(priv, 0x00051b, 0x3f800000);
-	nv_icmd(priv, 0x00051c, 0x3f800000);
-	nv_icmd(priv, 0x00051d, 0x3f800000);
-	nv_icmd(priv, 0x00051e, 0x3f800000);
-	nv_icmd(priv, 0x00051f, 0x3f800000);
-	nv_icmd(priv, 0x000520, 0x000002b6);
-	nv_icmd(priv, 0x000529, 0x00000001);
-	nv_icmd(priv, 0x000530, 0xffff0000);
-	nv_icmd(priv, 0x000531, 0xffff0000);
-	nv_icmd(priv, 0x000532, 0xffff0000);
-	nv_icmd(priv, 0x000533, 0xffff0000);
-	nv_icmd(priv, 0x000534, 0xffff0000);
-	nv_icmd(priv, 0x000535, 0xffff0000);
-	nv_icmd(priv, 0x000536, 0xffff0000);
-	nv_icmd(priv, 0x000537, 0xffff0000);
-	nv_icmd(priv, 0x000538, 0xffff0000);
-	nv_icmd(priv, 0x000539, 0xffff0000);
-	nv_icmd(priv, 0x00053a, 0xffff0000);
-	nv_icmd(priv, 0x00053b, 0xffff0000);
-	nv_icmd(priv, 0x00053c, 0xffff0000);
-	nv_icmd(priv, 0x00053d, 0xffff0000);
-	nv_icmd(priv, 0x00053e, 0xffff0000);
-	nv_icmd(priv, 0x00053f, 0xffff0000);
-	nv_icmd(priv, 0x000585, 0x0000003f);
-	nv_icmd(priv, 0x000576, 0x00000003);
-	nv_icmd(priv, 0x00057b, 0x00000059);
-	nv_icmd(priv, 0x000586, 0x00000040);
-	nv_icmd(priv, 0x000582, 0x00000080);
-	nv_icmd(priv, 0x000583, 0x00000080);
-	nv_icmd(priv, 0x0005c2, 0x00000001);
-	nv_icmd(priv, 0x000638, 0x00000001);
-	nv_icmd(priv, 0x000639, 0x00000001);
-	nv_icmd(priv, 0x00063a, 0x00000002);
-	nv_icmd(priv, 0x00063b, 0x00000001);
-	nv_icmd(priv, 0x00063c, 0x00000001);
-	nv_icmd(priv, 0x00063d, 0x00000002);
-	nv_icmd(priv, 0x00063e, 0x00000001);
-	nv_icmd(priv, 0x0008b8, 0x00000001);
-	nv_icmd(priv, 0x0008b9, 0x00000001);
-	nv_icmd(priv, 0x0008ba, 0x00000001);
-	nv_icmd(priv, 0x0008bb, 0x00000001);
-	nv_icmd(priv, 0x0008bc, 0x00000001);
-	nv_icmd(priv, 0x0008bd, 0x00000001);
-	nv_icmd(priv, 0x0008be, 0x00000001);
-	nv_icmd(priv, 0x0008bf, 0x00000001);
-	nv_icmd(priv, 0x000900, 0x00000001);
-	nv_icmd(priv, 0x000901, 0x00000001);
-	nv_icmd(priv, 0x000902, 0x00000001);
-	nv_icmd(priv, 0x000903, 0x00000001);
-	nv_icmd(priv, 0x000904, 0x00000001);
-	nv_icmd(priv, 0x000905, 0x00000001);
-	nv_icmd(priv, 0x000906, 0x00000001);
-	nv_icmd(priv, 0x000907, 0x00000001);
-	nv_icmd(priv, 0x000908, 0x00000002);
-	nv_icmd(priv, 0x000909, 0x00000002);
-	nv_icmd(priv, 0x00090a, 0x00000002);
-	nv_icmd(priv, 0x00090b, 0x00000002);
-	nv_icmd(priv, 0x00090c, 0x00000002);
-	nv_icmd(priv, 0x00090d, 0x00000002);
-	nv_icmd(priv, 0x00090e, 0x00000002);
-	nv_icmd(priv, 0x00090f, 0x00000002);
-	nv_icmd(priv, 0x000910, 0x00000001);
-	nv_icmd(priv, 0x000911, 0x00000001);
-	nv_icmd(priv, 0x000912, 0x00000001);
-	nv_icmd(priv, 0x000913, 0x00000001);
-	nv_icmd(priv, 0x000914, 0x00000001);
-	nv_icmd(priv, 0x000915, 0x00000001);
-	nv_icmd(priv, 0x000916, 0x00000001);
-	nv_icmd(priv, 0x000917, 0x00000001);
-	nv_icmd(priv, 0x000918, 0x00000001);
-	nv_icmd(priv, 0x000919, 0x00000001);
-	nv_icmd(priv, 0x00091a, 0x00000001);
-	nv_icmd(priv, 0x00091b, 0x00000001);
-	nv_icmd(priv, 0x00091c, 0x00000001);
-	nv_icmd(priv, 0x00091d, 0x00000001);
-	nv_icmd(priv, 0x00091e, 0x00000001);
-	nv_icmd(priv, 0x00091f, 0x00000001);
-	nv_icmd(priv, 0x000920, 0x00000002);
-	nv_icmd(priv, 0x000921, 0x00000002);
-	nv_icmd(priv, 0x000922, 0x00000002);
-	nv_icmd(priv, 0x000923, 0x00000002);
-	nv_icmd(priv, 0x000924, 0x00000002);
-	nv_icmd(priv, 0x000925, 0x00000002);
-	nv_icmd(priv, 0x000926, 0x00000002);
-	nv_icmd(priv, 0x000927, 0x00000002);
-	nv_icmd(priv, 0x000928, 0x00000001);
-	nv_icmd(priv, 0x000929, 0x00000001);
-	nv_icmd(priv, 0x00092a, 0x00000001);
-	nv_icmd(priv, 0x00092b, 0x00000001);
-	nv_icmd(priv, 0x00092c, 0x00000001);
-	nv_icmd(priv, 0x00092d, 0x00000001);
-	nv_icmd(priv, 0x00092e, 0x00000001);
-	nv_icmd(priv, 0x00092f, 0x00000001);
-	nv_icmd(priv, 0x000648, 0x00000001);
-	nv_icmd(priv, 0x000649, 0x00000001);
-	nv_icmd(priv, 0x00064a, 0x00000001);
-	nv_icmd(priv, 0x00064b, 0x00000001);
-	nv_icmd(priv, 0x00064c, 0x00000001);
-	nv_icmd(priv, 0x00064d, 0x00000001);
-	nv_icmd(priv, 0x00064e, 0x00000001);
-	nv_icmd(priv, 0x00064f, 0x00000001);
-	nv_icmd(priv, 0x000650, 0x00000001);
-	nv_icmd(priv, 0x000658, 0x0000000f);
-	nv_icmd(priv, 0x0007ff, 0x0000000a);
-	nv_icmd(priv, 0x00066a, 0x40000000);
-	nv_icmd(priv, 0x00066b, 0x10000000);
-	nv_icmd(priv, 0x00066c, 0xffff0000);
-	nv_icmd(priv, 0x00066d, 0xffff0000);
-	nv_icmd(priv, 0x0007af, 0x00000008);
-	nv_icmd(priv, 0x0007b0, 0x00000008);
-	nv_icmd(priv, 0x0007f6, 0x00000001);
-	nv_icmd(priv, 0x0006b2, 0x00000055);
-	nv_icmd(priv, 0x0007ad, 0x00000003);
-	nv_icmd(priv, 0x000937, 0x00000001);
-	nv_icmd(priv, 0x000971, 0x00000008);
-	nv_icmd(priv, 0x000972, 0x00000040);
-	nv_icmd(priv, 0x000973, 0x0000012c);
-	nv_icmd(priv, 0x00097c, 0x00000040);
-	nv_icmd(priv, 0x000979, 0x00000003);
-	nv_icmd(priv, 0x000975, 0x00000020);
-	nv_icmd(priv, 0x000976, 0x00000001);
-	nv_icmd(priv, 0x000977, 0x00000020);
-	nv_icmd(priv, 0x000978, 0x00000001);
-	nv_icmd(priv, 0x000957, 0x00000003);
-	nv_icmd(priv, 0x00095e, 0x20164010);
-	nv_icmd(priv, 0x00095f, 0x00000020);
-	nv_icmd(priv, 0x00097d, 0x00000020);
-	nv_icmd(priv, 0x000683, 0x00000006);
-	nv_icmd(priv, 0x000685, 0x003fffff);
-	nv_icmd(priv, 0x000687, 0x003fffff);
-	nv_icmd(priv, 0x0006a0, 0x00000005);
-	nv_icmd(priv, 0x000840, 0x00400008);
-	nv_icmd(priv, 0x000841, 0x08000080);
-	nv_icmd(priv, 0x000842, 0x00400008);
-	nv_icmd(priv, 0x000843, 0x08000080);
-	nv_icmd(priv, 0x000818, 0x00000000);
-	nv_icmd(priv, 0x000819, 0x00000000);
-	nv_icmd(priv, 0x00081a, 0x00000000);
-	nv_icmd(priv, 0x00081b, 0x00000000);
-	nv_icmd(priv, 0x00081c, 0x00000000);
-	nv_icmd(priv, 0x00081d, 0x00000000);
-	nv_icmd(priv, 0x00081e, 0x00000000);
-	nv_icmd(priv, 0x00081f, 0x00000000);
-	nv_icmd(priv, 0x000848, 0x00000000);
-	nv_icmd(priv, 0x000849, 0x00000000);
-	nv_icmd(priv, 0x00084a, 0x00000000);
-	nv_icmd(priv, 0x00084b, 0x00000000);
-	nv_icmd(priv, 0x00084c, 0x00000000);
-	nv_icmd(priv, 0x00084d, 0x00000000);
-	nv_icmd(priv, 0x00084e, 0x00000000);
-	nv_icmd(priv, 0x00084f, 0x00000000);
-	nv_icmd(priv, 0x000850, 0x00000000);
-	nv_icmd(priv, 0x000851, 0x00000000);
-	nv_icmd(priv, 0x000852, 0x00000000);
-	nv_icmd(priv, 0x000853, 0x00000000);
-	nv_icmd(priv, 0x000854, 0x00000000);
-	nv_icmd(priv, 0x000855, 0x00000000);
-	nv_icmd(priv, 0x000856, 0x00000000);
-	nv_icmd(priv, 0x000857, 0x00000000);
-	nv_icmd(priv, 0x000738, 0x00000000);
-	nv_icmd(priv, 0x0006aa, 0x00000001);
-	nv_icmd(priv, 0x0006ab, 0x00000002);
-	nv_icmd(priv, 0x0006ac, 0x00000080);
-	nv_icmd(priv, 0x0006ad, 0x00000100);
-	nv_icmd(priv, 0x0006ae, 0x00000100);
-	nv_icmd(priv, 0x0006b1, 0x00000011);
-	nv_icmd(priv, 0x0006bb, 0x000000cf);
-	nv_icmd(priv, 0x0006ce, 0x2a712488);
-	nv_icmd(priv, 0x000739, 0x4085c000);
-	nv_icmd(priv, 0x00073a, 0x00000080);
-	nv_icmd(priv, 0x000786, 0x80000100);
-	nv_icmd(priv, 0x00073c, 0x00010100);
-	nv_icmd(priv, 0x00073d, 0x02800000);
-	nv_icmd(priv, 0x000787, 0x000000cf);
-	nv_icmd(priv, 0x00078c, 0x00000008);
-	nv_icmd(priv, 0x000792, 0x00000001);
-	nv_icmd(priv, 0x000794, 0x00000001);
-	nv_icmd(priv, 0x000795, 0x00000001);
-	nv_icmd(priv, 0x000796, 0x00000001);
-	nv_icmd(priv, 0x000797, 0x000000cf);
-	nv_icmd(priv, 0x000836, 0x00000001);
-	nv_icmd(priv, 0x00079a, 0x00000002);
-	nv_icmd(priv, 0x000833, 0x04444480);
-	nv_icmd(priv, 0x0007a1, 0x00000001);
-	nv_icmd(priv, 0x0007a3, 0x00000001);
-	nv_icmd(priv, 0x0007a4, 0x00000001);
-	nv_icmd(priv, 0x0007a5, 0x00000001);
-	nv_icmd(priv, 0x000831, 0x00000004);
-	nv_icmd(priv, 0x000b07, 0x00000002);
-	nv_icmd(priv, 0x000b08, 0x00000100);
-	nv_icmd(priv, 0x000b09, 0x00000100);
-	nv_icmd(priv, 0x000b0a, 0x00000001);
-	nv_icmd(priv, 0x000a04, 0x000000ff);
-	nv_icmd(priv, 0x000a0b, 0x00000040);
-	nv_icmd(priv, 0x00097f, 0x00000100);
-	nv_icmd(priv, 0x000a02, 0x00000001);
-	nv_icmd(priv, 0x000809, 0x00000007);
-	nv_icmd(priv, 0x00c221, 0x00000040);
-	nv_icmd(priv, 0x00c1b0, 0x0000000f);
-	nv_icmd(priv, 0x00c1b1, 0x0000000f);
-	nv_icmd(priv, 0x00c1b2, 0x0000000f);
-	nv_icmd(priv, 0x00c1b3, 0x0000000f);
-	nv_icmd(priv, 0x00c1b4, 0x0000000f);
-	nv_icmd(priv, 0x00c1b5, 0x0000000f);
-	nv_icmd(priv, 0x00c1b6, 0x0000000f);
-	nv_icmd(priv, 0x00c1b7, 0x0000000f);
-	nv_icmd(priv, 0x00c1b8, 0x0fac6881);
-	nv_icmd(priv, 0x00c1b9, 0x00fac688);
-	nv_icmd(priv, 0x00c401, 0x00000001);
-	nv_icmd(priv, 0x00c402, 0x00010001);
-	nv_icmd(priv, 0x00c403, 0x00000001);
-	nv_icmd(priv, 0x00c404, 0x00000001);
-	nv_icmd(priv, 0x00c40e, 0x00000020);
-	nv_icmd(priv, 0x00c500, 0x00000003);
-	nv_icmd(priv, 0x01e100, 0x00000001);
-	nv_icmd(priv, 0x001000, 0x00000002);
-	nv_icmd(priv, 0x0006aa, 0x00000001);
-	nv_icmd(priv, 0x0006ad, 0x00000100);
-	nv_icmd(priv, 0x0006ae, 0x00000100);
-	nv_icmd(priv, 0x0006b1, 0x00000011);
-	nv_icmd(priv, 0x00078c, 0x00000008);
-	nv_icmd(priv, 0x000792, 0x00000001);
-	nv_icmd(priv, 0x000794, 0x00000001);
-	nv_icmd(priv, 0x000795, 0x00000001);
-	nv_icmd(priv, 0x000796, 0x00000001);
-	nv_icmd(priv, 0x000797, 0x000000cf);
-	nv_icmd(priv, 0x00079a, 0x00000002);
-	nv_icmd(priv, 0x000833, 0x04444480);
-	nv_icmd(priv, 0x0007a1, 0x00000001);
-	nv_icmd(priv, 0x0007a3, 0x00000001);
-	nv_icmd(priv, 0x0007a4, 0x00000001);
-	nv_icmd(priv, 0x0007a5, 0x00000001);
-	nv_icmd(priv, 0x000831, 0x00000004);
-	nv_icmd(priv, 0x01e100, 0x00000001);
-	nv_icmd(priv, 0x001000, 0x00000008);
-	nv_icmd(priv, 0x000039, 0x00000000);
-	nv_icmd(priv, 0x00003a, 0x00000000);
-	nv_icmd(priv, 0x00003b, 0x00000000);
-	nv_icmd(priv, 0x000380, 0x00000001);
-	nv_icmd(priv, 0x000366, 0x00000000);
-	nv_icmd(priv, 0x000367, 0x00000000);
-	nv_icmd(priv, 0x000368, 0x00000fff);
-	nv_icmd(priv, 0x000370, 0x00000000);
-	nv_icmd(priv, 0x000371, 0x00000000);
-	nv_icmd(priv, 0x000372, 0x000fffff);
-	nv_icmd(priv, 0x000813, 0x00000006);
-	nv_icmd(priv, 0x000814, 0x00000008);
-	nv_icmd(priv, 0x000957, 0x00000003);
-	nv_icmd(priv, 0x000818, 0x00000000);
-	nv_icmd(priv, 0x000819, 0x00000000);
-	nv_icmd(priv, 0x00081a, 0x00000000);
-	nv_icmd(priv, 0x00081b, 0x00000000);
-	nv_icmd(priv, 0x00081c, 0x00000000);
-	nv_icmd(priv, 0x00081d, 0x00000000);
-	nv_icmd(priv, 0x00081e, 0x00000000);
-	nv_icmd(priv, 0x00081f, 0x00000000);
-	nv_icmd(priv, 0x000848, 0x00000000);
-	nv_icmd(priv, 0x000849, 0x00000000);
-	nv_icmd(priv, 0x00084a, 0x00000000);
-	nv_icmd(priv, 0x00084b, 0x00000000);
-	nv_icmd(priv, 0x00084c, 0x00000000);
-	nv_icmd(priv, 0x00084d, 0x00000000);
-	nv_icmd(priv, 0x00084e, 0x00000000);
-	nv_icmd(priv, 0x00084f, 0x00000000);
-	nv_icmd(priv, 0x000850, 0x00000000);
-	nv_icmd(priv, 0x000851, 0x00000000);
-	nv_icmd(priv, 0x000852, 0x00000000);
-	nv_icmd(priv, 0x000853, 0x00000000);
-	nv_icmd(priv, 0x000854, 0x00000000);
-	nv_icmd(priv, 0x000855, 0x00000000);
-	nv_icmd(priv, 0x000856, 0x00000000);
-	nv_icmd(priv, 0x000857, 0x00000000);
-	nv_icmd(priv, 0x000738, 0x00000000);
-	nv_icmd(priv, 0x000b07, 0x00000002);
-	nv_icmd(priv, 0x000b08, 0x00000100);
-	nv_icmd(priv, 0x000b09, 0x00000100);
-	nv_icmd(priv, 0x000b0a, 0x00000001);
-	nv_icmd(priv, 0x000a04, 0x000000ff);
-	nv_icmd(priv, 0x00097f, 0x00000100);
-	nv_icmd(priv, 0x000a02, 0x00000001);
-	nv_icmd(priv, 0x000809, 0x00000007);
-	nv_icmd(priv, 0x00c221, 0x00000040);
-	nv_icmd(priv, 0x00c401, 0x00000001);
-	nv_icmd(priv, 0x00c402, 0x00010001);
-	nv_icmd(priv, 0x00c403, 0x00000001);
-	nv_icmd(priv, 0x00c404, 0x00000001);
-	nv_icmd(priv, 0x00c40e, 0x00000020);
-	nv_icmd(priv, 0x00c500, 0x00000003);
-	nv_icmd(priv, 0x01e100, 0x00000001);
-	nv_icmd(priv, 0x001000, 0x00000001);
-	nv_icmd(priv, 0x000b07, 0x00000002);
-	nv_icmd(priv, 0x000b08, 0x00000100);
-	nv_icmd(priv, 0x000b09, 0x00000100);
-	nv_icmd(priv, 0x000b0a, 0x00000001);
-	nv_icmd(priv, 0x01e100, 0x00000001);
-	nv_wr32(priv, 0x400208, 0x00000000);
-}
-
-static void
-nve0_grctx_generate_a097(struct nvc0_graph_priv *priv)
-{
-	nv_mthd(priv, 0xa097, 0x0800, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0840, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0880, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x08c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0900, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0940, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0980, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x09c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0804, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0844, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0884, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x08c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0904, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0944, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0984, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x09c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0808, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x0848, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x0888, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x08c8, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x0908, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x0948, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x0988, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x09c8, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x080c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x084c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x088c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x08cc, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x090c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x094c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x098c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x09cc, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x0810, 0x000000cf);
-	nv_mthd(priv, 0xa097, 0x0850, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0890, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x08d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0910, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0950, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0990, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x09d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0814, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0854, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0894, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x08d4, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0914, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0954, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0994, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x09d4, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0818, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0858, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0898, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x08d8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0918, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0958, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0998, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x09d8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x081c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x085c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x089c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x08dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x091c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x095c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x099c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x09dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0820, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0860, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x08a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x08e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0920, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0960, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x09a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x09e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ca0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cb0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cc0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cd0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ce0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cf0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c24, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c34, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c64, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c94, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ca4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cb4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cc4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cd4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ce4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cf4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c18, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c28, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c38, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c58, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c68, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c78, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c98, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ca8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cb8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cc8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cd8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ce8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cf8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c0c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c1c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c2c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c3c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c4c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c5c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c6c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c7c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1c9c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cbc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ccc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cdc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1cfc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1da0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1db0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dc0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dd0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1de0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1df0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d24, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d34, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d64, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d94, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1da4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1db4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dc4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dd4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1de4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1df4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d18, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d28, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d38, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d58, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d68, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d78, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d98, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1da8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1db8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dc8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dd8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1de8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1df8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d0c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d1c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d2c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d3c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d4c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d5c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d6c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d7c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1d9c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dbc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dcc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ddc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1dfc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f18, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f28, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f38, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f58, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f68, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f78, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f0c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f1c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f24, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f2c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f34, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f3c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f4c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f5c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f64, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f6c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f7c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f98, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fa0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fa8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fb0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fb8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fc0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fc8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fd0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fd8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fe0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fe8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ff0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ff8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f94, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1f9c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fa4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fb4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fbc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fc4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fcc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fd4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fdc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fe4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1fec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ff4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1ffc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2000, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2040, 0x00000011);
-	nv_mthd(priv, 0xa097, 0x2080, 0x00000020);
-	nv_mthd(priv, 0xa097, 0x20c0, 0x00000030);
-	nv_mthd(priv, 0xa097, 0x2100, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x2140, 0x00000051);
-	nv_mthd(priv, 0xa097, 0x200c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x204c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x208c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x20cc, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x210c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x214c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x2010, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2050, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2090, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x20d0, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x2110, 0x00000003);
-	nv_mthd(priv, 0xa097, 0x2150, 0x00000004);
-	nv_mthd(priv, 0xa097, 0x0380, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0384, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0388, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x038c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x03ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0700, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0710, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0720, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0730, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0704, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0714, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0724, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0734, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0708, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0718, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0728, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0738, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2800, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2804, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2808, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x280c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2810, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2814, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2818, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x281c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2820, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2824, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2828, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x282c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2830, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2834, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2838, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x283c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2840, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2844, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2848, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x284c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2850, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2854, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2858, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x285c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2860, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2864, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2868, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x286c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2870, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2874, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2878, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x287c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2880, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2884, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2888, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x288c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2890, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2894, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2898, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x289c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28b0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28b4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28b8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28d4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28d8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28f0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x28fc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2900, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2904, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2908, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x290c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2910, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2914, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2918, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x291c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2920, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2924, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2928, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x292c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2930, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2934, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2938, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x293c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2940, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2944, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2948, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x294c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2950, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2954, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2958, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x295c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2960, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2964, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2968, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x296c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2970, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2974, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2978, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x297c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2980, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2984, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2988, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x298c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2990, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2994, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2998, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x299c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29b0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29b4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29b8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29d4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29d8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29f0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x29fc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0aa0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ac0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ae0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ba0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bc0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0be0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a24, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a64, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0aa4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ac4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ae4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b24, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b64, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ba4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bc4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0be4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a28, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a68, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0aa8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ac8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ae8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b28, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b68, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ba8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bc8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0be8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a0c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a2c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a4c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a6c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0aac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0acc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0aec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b0c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b2c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b4c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b6c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bcc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ab0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ad0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0af0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bb0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bd0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bf0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a34, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0a94, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ab4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ad4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0af4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b34, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0b94, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bb4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bd4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0bf4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ca0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cb0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cc0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cd0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ce0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cf0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c24, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c34, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c64, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c94, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ca4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cb4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cc4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cd4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ce4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cf4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c18, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c28, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c38, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c58, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c68, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c78, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c98, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ca8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cb8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cc8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cd8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ce8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0cf8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0c0c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c1c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c2c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c3c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c4c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c5c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c6c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c7c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c8c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0c9c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0cac, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0cbc, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0ccc, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0cdc, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0cec, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0cfc, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0d00, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d08, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d10, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d18, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d20, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d28, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d30, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d38, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d04, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d0c, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d14, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d1c, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d24, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d2c, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d34, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d3c, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e00, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e20, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e30, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e60, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e70, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ea0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0eb0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ec0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ed0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ee0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ef0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0e04, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e14, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e24, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e34, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e44, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e54, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e64, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e74, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e84, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e94, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ea4, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0eb4, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ec4, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ed4, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ee4, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ef4, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e08, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e18, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e28, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e38, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e48, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e58, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e68, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e78, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e88, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0e98, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ea8, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0eb8, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ec8, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ed8, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ee8, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0ef8, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d40, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d48, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d50, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d58, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d44, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d4c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d5c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1e00, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e20, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e40, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e60, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e80, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ea0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ec0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ee0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e04, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e24, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e44, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e64, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e84, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ea4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ec4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ee4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e08, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e28, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e48, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e68, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e88, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1ea8, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1ec8, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1ee8, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e0c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e2c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e4c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e6c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e8c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1eac, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ecc, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1eec, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e10, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e30, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e50, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e70, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e90, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1eb0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ed0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ef0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e14, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e34, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e54, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e74, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e94, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1eb4, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1ed4, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1ef4, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1e18, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e38, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e58, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e78, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1e98, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1eb8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ed8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1ef8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x3400, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3404, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3408, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x340c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3410, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3414, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3418, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x341c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3420, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3424, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3428, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x342c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3430, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3434, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3438, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x343c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3440, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3444, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3448, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x344c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3450, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3454, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3458, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x345c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3460, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3464, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3468, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x346c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3470, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3474, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3478, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x347c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3480, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3484, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3488, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x348c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3490, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3494, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3498, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x349c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34b0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34b4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34b8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34d4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34d8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34f0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x34fc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3500, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3504, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3508, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x350c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3510, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3514, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3518, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x351c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3520, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3524, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3528, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x352c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3530, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3534, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3538, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x353c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3540, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3544, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3548, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x354c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3550, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3554, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3558, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x355c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3560, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3564, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3568, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x356c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3570, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3574, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3578, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x357c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3580, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3584, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3588, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x358c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3590, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3594, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x3598, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x359c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35b0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35b4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35b8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35d4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35d8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35f0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x35fc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x030c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1944, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1514, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d68, 0x0000ffff);
-	nv_mthd(priv, 0xa097, 0x121c, 0x0fac6881);
-	nv_mthd(priv, 0xa097, 0x0fac, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1538, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0fe0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0fe4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0fe8, 0x00000014);
-	nv_mthd(priv, 0xa097, 0x0fec, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x0ff0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x179c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1228, 0x00000400);
-	nv_mthd(priv, 0xa097, 0x122c, 0x00000300);
-	nv_mthd(priv, 0xa097, 0x1230, 0x00010001);
-	nv_mthd(priv, 0xa097, 0x07f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x15b4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x15cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1534, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0fb0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x15d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x153c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x16b4, 0x00000003);
-	nv_mthd(priv, 0xa097, 0x0fbc, 0x0000ffff);
-	nv_mthd(priv, 0xa097, 0x0fc0, 0x0000ffff);
-	nv_mthd(priv, 0xa097, 0x0fc4, 0x0000ffff);
-	nv_mthd(priv, 0xa097, 0x0fc8, 0x0000ffff);
-	nv_mthd(priv, 0xa097, 0x0df8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0dfc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1948, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1970, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x161c, 0x000009f0);
-	nv_mthd(priv, 0xa097, 0x0dcc, 0x00000010);
-	nv_mthd(priv, 0xa097, 0x163c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x15e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1160, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1164, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1168, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x116c, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1170, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1174, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1178, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x117c, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1180, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1184, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1188, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x118c, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1190, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1194, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1198, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x119c, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11a0, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11a4, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11a8, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11ac, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11b0, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11b4, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11b8, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11bc, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11c0, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11c4, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11c8, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11cc, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11d0, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11d4, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11d8, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x11dc, 0x25e00040);
-	nv_mthd(priv, 0xa097, 0x1880, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1884, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1888, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x188c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1890, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1894, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1898, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x189c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18b0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18b4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18b8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18d0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18d4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18d8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18e0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18f0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x18fc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x17c8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x17cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x17d0, 0x000000ff);
-	nv_mthd(priv, 0xa097, 0x17d4, 0xffffffff);
-	nv_mthd(priv, 0xa097, 0x17d8, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x17dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x15f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x15f8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1434, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1438, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d74, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0dec, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x13a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1318, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1644, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0748, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0de8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1648, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x12a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1120, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1124, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1128, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x112c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1118, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x164c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1658, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1910, 0x00000290);
-	nv_mthd(priv, 0xa097, 0x1518, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x165c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1520, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1604, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1570, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x13b0, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x13b4, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x020c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1670, 0x30201000);
-	nv_mthd(priv, 0xa097, 0x1674, 0x70605040);
-	nv_mthd(priv, 0xa097, 0x1678, 0xb8a89888);
-	nv_mthd(priv, 0xa097, 0x167c, 0xf8e8d8c8);
-	nv_mthd(priv, 0xa097, 0x166c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1680, 0x00ffff00);
-	nv_mthd(priv, 0xa097, 0x12d0, 0x00000003);
-	nv_mthd(priv, 0xa097, 0x12d4, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1684, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1688, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0dac, 0x00001b02);
-	nv_mthd(priv, 0xa097, 0x0db0, 0x00001b02);
-	nv_mthd(priv, 0xa097, 0x0db4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x168c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x15bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x156c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x187c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1110, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0dc0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0dc4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0dc8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1234, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1690, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x12ac, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0790, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0794, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0798, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x079c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07a0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x077c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1000, 0x00000010);
-	nv_mthd(priv, 0xa097, 0x10fc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1290, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0218, 0x00000010);
-	nv_mthd(priv, 0xa097, 0x12d8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x12dc, 0x00000010);
-	nv_mthd(priv, 0xa097, 0x0d94, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x155c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1560, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1564, 0x00000fff);
-	nv_mthd(priv, 0xa097, 0x1574, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1578, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x157c, 0x000fffff);
-	nv_mthd(priv, 0xa097, 0x1354, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1610, 0x00000012);
-	nv_mthd(priv, 0xa097, 0x1608, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x160c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x260c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x162c, 0x00000003);
-	nv_mthd(priv, 0xa097, 0x0210, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0320, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0324, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0328, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x032c, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0330, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0334, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0338, 0x3f800000);
-	nv_mthd(priv, 0xa097, 0x0750, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0760, 0x39291909);
-	nv_mthd(priv, 0xa097, 0x0764, 0x79695949);
-	nv_mthd(priv, 0xa097, 0x0768, 0xb9a99989);
-	nv_mthd(priv, 0xa097, 0x076c, 0xf9e9d9c9);
-	nv_mthd(priv, 0xa097, 0x0770, 0x30201000);
-	nv_mthd(priv, 0xa097, 0x0774, 0x70605040);
-	nv_mthd(priv, 0xa097, 0x0778, 0x00009080);
-	nv_mthd(priv, 0xa097, 0x0780, 0x39291909);
-	nv_mthd(priv, 0xa097, 0x0784, 0x79695949);
-	nv_mthd(priv, 0xa097, 0x0788, 0xb9a99989);
-	nv_mthd(priv, 0xa097, 0x078c, 0xf9e9d9c9);
-	nv_mthd(priv, 0xa097, 0x07d0, 0x30201000);
-	nv_mthd(priv, 0xa097, 0x07d4, 0x70605040);
-	nv_mthd(priv, 0xa097, 0x07d8, 0x00009080);
-	nv_mthd(priv, 0xa097, 0x037c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0740, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0744, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x2600, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1918, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x191c, 0x00000900);
-	nv_mthd(priv, 0xa097, 0x1920, 0x00000405);
-	nv_mthd(priv, 0xa097, 0x1308, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1924, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x13ac, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x192c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x193c, 0x00002c1c);
-	nv_mthd(priv, 0xa097, 0x0d7c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x02c0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1510, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1940, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ff4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0ff8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x194c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1950, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1968, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1590, 0x0000003f);
-	nv_mthd(priv, 0xa097, 0x07e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07f0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07f4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x196c, 0x00000011);
-	nv_mthd(priv, 0xa097, 0x02e4, 0x0000b001);
-	nv_mthd(priv, 0xa097, 0x036c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0370, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x197c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0fcc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0fd0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x02d8, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x1980, 0x00000080);
-	nv_mthd(priv, 0xa097, 0x1504, 0x00000080);
-	nv_mthd(priv, 0xa097, 0x1984, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0300, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x13a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x12ec, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1310, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1314, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1380, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1384, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1388, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x138c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1390, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1394, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x139c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1398, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1594, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1598, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x159c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x15a0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x15a4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0f54, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f58, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f5c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x19bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f9c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0fa0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x12cc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x12e8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x130c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1360, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1364, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1368, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x136c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1370, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1374, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1378, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x137c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x133c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1340, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1344, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1348, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x134c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1350, 0x00000002);
-	nv_mthd(priv, 0xa097, 0x1358, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x12e4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x131c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1320, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1324, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1328, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x19c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1140, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x19c4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x19c8, 0x00001500);
-	nv_mthd(priv, 0xa097, 0x135c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x19e0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19e4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19e8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19ec, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19f0, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19f4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19f8, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19fc, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x19cc, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x15b8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a00, 0x00001111);
-	nv_mthd(priv, 0xa097, 0x1a04, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a08, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a0c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a10, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a14, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a18, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1a1c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d6c, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x0d70, 0xffff0000);
-	nv_mthd(priv, 0xa097, 0x10f8, 0x00001010);
-	nv_mthd(priv, 0xa097, 0x0d80, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d84, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d88, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d8c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0d90, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0da0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07a4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x07a8, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1508, 0x80000000);
-	nv_mthd(priv, 0xa097, 0x150c, 0x40000000);
-	nv_mthd(priv, 0xa097, 0x1668, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0318, 0x00000008);
-	nv_mthd(priv, 0xa097, 0x031c, 0x00000008);
-	nv_mthd(priv, 0xa097, 0x0d9c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x0374, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0378, 0x00000020);
-	nv_mthd(priv, 0xa097, 0x07dc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x074c, 0x00000055);
-	nv_mthd(priv, 0xa097, 0x1420, 0x00000003);
-	nv_mthd(priv, 0xa097, 0x17bc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x17c0, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x17c4, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1008, 0x00000008);
-	nv_mthd(priv, 0xa097, 0x100c, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x1010, 0x0000012c);
-	nv_mthd(priv, 0xa097, 0x0d60, 0x00000040);
-	nv_mthd(priv, 0xa097, 0x075c, 0x00000003);
-	nv_mthd(priv, 0xa097, 0x1018, 0x00000020);
-	nv_mthd(priv, 0xa097, 0x101c, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1020, 0x00000020);
-	nv_mthd(priv, 0xa097, 0x1024, 0x00000001);
-	nv_mthd(priv, 0xa097, 0x1444, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x1448, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x144c, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0360, 0x20164010);
-	nv_mthd(priv, 0xa097, 0x0364, 0x00000020);
-	nv_mthd(priv, 0xa097, 0x0368, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0de4, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0204, 0x00000006);
-	nv_mthd(priv, 0xa097, 0x0208, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x02cc, 0x003fffff);
-	nv_mthd(priv, 0xa097, 0x02d0, 0x003fffff);
-	nv_mthd(priv, 0xa097, 0x1220, 0x00000005);
-	nv_mthd(priv, 0xa097, 0x0fdc, 0x00000000);
-	nv_mthd(priv, 0xa097, 0x0f98, 0x00400008);
-	nv_mthd(priv, 0xa097, 0x1284, 0x08000080);
-	nv_mthd(priv, 0xa097, 0x1450, 0x00400008);
-	nv_mthd(priv, 0xa097, 0x1454, 0x08000080);
-	nv_mthd(priv, 0xa097, 0x0214, 0x00000000);
-}
-
-static void
-nve0_grctx_generate_902d(struct nvc0_graph_priv *priv)
-{
-	nv_mthd(priv, 0x902d, 0x0200, 0x000000cf);
-	nv_mthd(priv, 0x902d, 0x0204, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0208, 0x00000020);
-	nv_mthd(priv, 0x902d, 0x020c, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0210, 0x00000000);
-	nv_mthd(priv, 0x902d, 0x0214, 0x00000080);
-	nv_mthd(priv, 0x902d, 0x0218, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x021c, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x0220, 0x00000000);
-	nv_mthd(priv, 0x902d, 0x0224, 0x00000000);
-	nv_mthd(priv, 0x902d, 0x0230, 0x000000cf);
-	nv_mthd(priv, 0x902d, 0x0234, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0238, 0x00000020);
-	nv_mthd(priv, 0x902d, 0x023c, 0x00000001);
-	nv_mthd(priv, 0x902d, 0x0244, 0x00000080);
-	nv_mthd(priv, 0x902d, 0x0248, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x024c, 0x00000100);
-	nv_mthd(priv, 0x902d, 0x3410, 0x00000000);
-}
-
-static void
-nve0_graph_generate_unk40xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404010, 0x0);
-	nv_wr32(priv, 0x404014, 0x0);
-	nv_wr32(priv, 0x404018, 0x0);
-	nv_wr32(priv, 0x40401c, 0x0);
-	nv_wr32(priv, 0x404020, 0x0);
-	nv_wr32(priv, 0x404024, 0xe000);
-	nv_wr32(priv, 0x404028, 0x0);
-	nv_wr32(priv, 0x4040a8, 0x0);
-	nv_wr32(priv, 0x4040ac, 0x0);
-	nv_wr32(priv, 0x4040b0, 0x0);
-	nv_wr32(priv, 0x4040b4, 0x0);
-	nv_wr32(priv, 0x4040b8, 0x0);
-	nv_wr32(priv, 0x4040bc, 0x0);
-	nv_wr32(priv, 0x4040c0, 0x0);
-	nv_wr32(priv, 0x4040c4, 0x0);
-	nv_wr32(priv, 0x4040c8, 0xf800008f);
-	nv_wr32(priv, 0x4040d0, 0x0);
-	nv_wr32(priv, 0x4040d4, 0x0);
-	nv_wr32(priv, 0x4040d8, 0x0);
-	nv_wr32(priv, 0x4040dc, 0x0);
-	nv_wr32(priv, 0x4040e0, 0x0);
-	nv_wr32(priv, 0x4040e4, 0x0);
-	nv_wr32(priv, 0x4040e8, 0x1000);
-	nv_wr32(priv, 0x4040f8, 0x0);
-	nv_wr32(priv, 0x404130, 0x0);
-	nv_wr32(priv, 0x404134, 0x0);
-	nv_wr32(priv, 0x404138, 0x20000040);
-	nv_wr32(priv, 0x404150, 0x2e);
-	nv_wr32(priv, 0x404154, 0x400);
-	nv_wr32(priv, 0x404158, 0x200);
-	nv_wr32(priv, 0x404164, 0x55);
-	nv_wr32(priv, 0x4041a0, 0x0);
-	nv_wr32(priv, 0x4041a4, 0x0);
-	nv_wr32(priv, 0x4041a8, 0x0);
-	nv_wr32(priv, 0x4041ac, 0x0);
-	nv_wr32(priv, 0x404200, 0x0);
-	nv_wr32(priv, 0x404204, 0x0);
-	nv_wr32(priv, 0x404208, 0x0);
-	nv_wr32(priv, 0x40420c, 0x0);
-}
-
-static void
-nve0_graph_generate_unk44xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404404, 0x0);
-	nv_wr32(priv, 0x404408, 0x0);
-	nv_wr32(priv, 0x40440c, 0x0);
-	nv_wr32(priv, 0x404410, 0x0);
-	nv_wr32(priv, 0x404414, 0x0);
-	nv_wr32(priv, 0x404418, 0x0);
-	nv_wr32(priv, 0x40441c, 0x0);
-	nv_wr32(priv, 0x404420, 0x0);
-	nv_wr32(priv, 0x404424, 0x0);
-	nv_wr32(priv, 0x404428, 0x0);
-	nv_wr32(priv, 0x40442c, 0x0);
-	nv_wr32(priv, 0x404430, 0x0);
-	nv_wr32(priv, 0x404434, 0x0);
-	nv_wr32(priv, 0x404438, 0x0);
-	nv_wr32(priv, 0x404460, 0x0);
-	nv_wr32(priv, 0x404464, 0x0);
-	nv_wr32(priv, 0x404468, 0xffffff);
-	nv_wr32(priv, 0x40446c, 0x0);
-	nv_wr32(priv, 0x404480, 0x1);
-	nv_wr32(priv, 0x404498, 0x1);
-}
-
-static void
-nve0_graph_generate_unk46xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404604, 0x14);
-	nv_wr32(priv, 0x404608, 0x0);
-	nv_wr32(priv, 0x40460c, 0x3fff);
-	nv_wr32(priv, 0x404610, 0x100);
-	nv_wr32(priv, 0x404618, 0x0);
-	nv_wr32(priv, 0x40461c, 0x0);
-	nv_wr32(priv, 0x404620, 0x0);
-	nv_wr32(priv, 0x404624, 0x0);
-	nv_wr32(priv, 0x40462c, 0x0);
-	nv_wr32(priv, 0x404630, 0x0);
-	nv_wr32(priv, 0x404640, 0x0);
-	nv_wr32(priv, 0x404654, 0x0);
-	nv_wr32(priv, 0x404660, 0x0);
-	nv_wr32(priv, 0x404678, 0x0);
-	nv_wr32(priv, 0x40467c, 0x2);
-	nv_wr32(priv, 0x404680, 0x0);
-	nv_wr32(priv, 0x404684, 0x0);
-	nv_wr32(priv, 0x404688, 0x0);
-	nv_wr32(priv, 0x40468c, 0x0);
-	nv_wr32(priv, 0x404690, 0x0);
-	nv_wr32(priv, 0x404694, 0x0);
-	nv_wr32(priv, 0x404698, 0x0);
-	nv_wr32(priv, 0x40469c, 0x0);
-	nv_wr32(priv, 0x4046a0, 0x7f0080);
-	nv_wr32(priv, 0x4046a4, 0x0);
-	nv_wr32(priv, 0x4046a8, 0x0);
-	nv_wr32(priv, 0x4046ac, 0x0);
-	nv_wr32(priv, 0x4046b0, 0x0);
-	nv_wr32(priv, 0x4046b4, 0x0);
-	nv_wr32(priv, 0x4046b8, 0x0);
-	nv_wr32(priv, 0x4046bc, 0x0);
-	nv_wr32(priv, 0x4046c0, 0x0);
-	nv_wr32(priv, 0x4046c8, 0x0);
-	nv_wr32(priv, 0x4046cc, 0x0);
-	nv_wr32(priv, 0x4046d0, 0x0);
-}
-
-static void
-nve0_graph_generate_unk47xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x404700, 0x0);
-	nv_wr32(priv, 0x404704, 0x0);
-	nv_wr32(priv, 0x404708, 0x0);
-	nv_wr32(priv, 0x404718, 0x0);
-	nv_wr32(priv, 0x40471c, 0x0);
-	nv_wr32(priv, 0x404720, 0x0);
-	nv_wr32(priv, 0x404724, 0x0);
-	nv_wr32(priv, 0x404728, 0x0);
-	nv_wr32(priv, 0x40472c, 0x0);
-	nv_wr32(priv, 0x404730, 0x0);
-	nv_wr32(priv, 0x404734, 0x100);
-	nv_wr32(priv, 0x404738, 0x0);
-	nv_wr32(priv, 0x40473c, 0x0);
-	nv_wr32(priv, 0x404744, 0x0);
-	nv_wr32(priv, 0x404748, 0x0);
-	nv_wr32(priv, 0x404754, 0x0);
-}
-
-static void
-nve0_graph_generate_unk58xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x405800, 0xf8000bf);
-	nv_wr32(priv, 0x405830, 0x2180648);
-	nv_wr32(priv, 0x405834, 0x8000000);
-	nv_wr32(priv, 0x405838, 0x0);
-	nv_wr32(priv, 0x405854, 0x0);
-	nv_wr32(priv, 0x405870, 0x1);
-	nv_wr32(priv, 0x405874, 0x1);
-	nv_wr32(priv, 0x405878, 0x1);
-	nv_wr32(priv, 0x40587c, 0x1);
-	nv_wr32(priv, 0x405a00, 0x0);
-	nv_wr32(priv, 0x405a04, 0x0);
-	nv_wr32(priv, 0x405a18, 0x0);
-	nv_wr32(priv, 0x405b00, 0x0);
-	nv_wr32(priv, 0x405b10, 0x1000);
-}
-
-static void
-nve0_graph_generate_unk60xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x406020, 0x4103c1);
-	nv_wr32(priv, 0x406028, 0x1);
-	nv_wr32(priv, 0x40602c, 0x1);
-	nv_wr32(priv, 0x406030, 0x1);
-	nv_wr32(priv, 0x406034, 0x1);
-}
-
-static void
-nve0_graph_generate_unk64xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x4064a8, 0x0);
-	nv_wr32(priv, 0x4064ac, 0x3fff);
-	nv_wr32(priv, 0x4064b4, 0x0);
-	nv_wr32(priv, 0x4064b8, 0x0);
-	nv_wr32(priv, 0x4064c0, 0x801a00f0);
-	nv_wr32(priv, 0x4064c4, 0x192ffff);
-	nv_wr32(priv, 0x4064c8, 0x1800600);
-	nv_wr32(priv, 0x4064cc, 0x0);
-	nv_wr32(priv, 0x4064d0, 0x0);
-	nv_wr32(priv, 0x4064d4, 0x0);
-	nv_wr32(priv, 0x4064d8, 0x0);
-	nv_wr32(priv, 0x4064dc, 0x0);
-	nv_wr32(priv, 0x4064e0, 0x0);
-	nv_wr32(priv, 0x4064e4, 0x0);
-	nv_wr32(priv, 0x4064e8, 0x0);
-	nv_wr32(priv, 0x4064ec, 0x0);
-	nv_wr32(priv, 0x4064fc, 0x22a);
-}
-
-static void
-nve0_graph_generate_unk70xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x407040, 0x0);
-}
-
-static void
-nve0_graph_generate_unk78xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x407804, 0x23);
-	nv_wr32(priv, 0x40780c, 0xa418820);
-	nv_wr32(priv, 0x407810, 0x62080e6);
-	nv_wr32(priv, 0x407814, 0x20398a4);
-	nv_wr32(priv, 0x407818, 0xe629062);
-	nv_wr32(priv, 0x40781c, 0xa418820);
-	nv_wr32(priv, 0x407820, 0xe6);
-	nv_wr32(priv, 0x4078bc, 0x103);
-}
-
-static void
-nve0_graph_generate_unk80xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x408000, 0x0);
-	nv_wr32(priv, 0x408004, 0x0);
-	nv_wr32(priv, 0x408008, 0x30);
-	nv_wr32(priv, 0x40800c, 0x0);
-	nv_wr32(priv, 0x408010, 0x0);
-	nv_wr32(priv, 0x408014, 0x69);
-	nv_wr32(priv, 0x408018, 0xe100e100);
-	nv_wr32(priv, 0x408064, 0x0);
-}
-
-static void
-nve0_graph_generate_unk88xx(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x408800, 0x2802a3c);
-	nv_wr32(priv, 0x408804, 0x40);
-	nv_wr32(priv, 0x408808, 0x1043e005);
-	nv_wr32(priv, 0x408840, 0xb);
-	nv_wr32(priv, 0x408900, 0x3080b801);
-	nv_wr32(priv, 0x408904, 0x62000001);
-	nv_wr32(priv, 0x408908, 0xc8102f);
-	nv_wr32(priv, 0x408980, 0x11d);
-}
-
-static void
-nve0_graph_generate_gpc(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x418380, 0x16);
-	nv_wr32(priv, 0x418400, 0x38004e00);
-	nv_wr32(priv, 0x418404, 0x71e0ffff);
-	nv_wr32(priv, 0x41840c, 0x1008);
-	nv_wr32(priv, 0x418410, 0xfff0fff);
-	nv_wr32(priv, 0x418414, 0x2200fff);
-	nv_wr32(priv, 0x418450, 0x0);
-	nv_wr32(priv, 0x418454, 0x0);
-	nv_wr32(priv, 0x418458, 0x0);
-	nv_wr32(priv, 0x41845c, 0x0);
-	nv_wr32(priv, 0x418460, 0x0);
-	nv_wr32(priv, 0x418464, 0x0);
-	nv_wr32(priv, 0x418468, 0x1);
-	nv_wr32(priv, 0x41846c, 0x0);
-	nv_wr32(priv, 0x418470, 0x0);
-	nv_wr32(priv, 0x418600, 0x1f);
-	nv_wr32(priv, 0x418684, 0xf);
-	nv_wr32(priv, 0x418700, 0x2);
-	nv_wr32(priv, 0x418704, 0x80);
-	nv_wr32(priv, 0x418708, 0x0);
-	nv_wr32(priv, 0x41870c, 0x0);
-	nv_wr32(priv, 0x418710, 0x0);
-	nv_wr32(priv, 0x418800, 0x7006860a);
-	nv_wr32(priv, 0x418808, 0x0);
-	nv_wr32(priv, 0x41880c, 0x0);
-	nv_wr32(priv, 0x418810, 0x0);
-	nv_wr32(priv, 0x418828, 0x44);
-	nv_wr32(priv, 0x418830, 0x10000001);
-	nv_wr32(priv, 0x4188d8, 0x8);
-	nv_wr32(priv, 0x4188e0, 0x1000000);
-	nv_wr32(priv, 0x4188e8, 0x0);
-	nv_wr32(priv, 0x4188ec, 0x0);
-	nv_wr32(priv, 0x4188f0, 0x0);
-	nv_wr32(priv, 0x4188f4, 0x0);
-	nv_wr32(priv, 0x4188f8, 0x0);
-	nv_wr32(priv, 0x4188fc, 0x20100018);
-	nv_wr32(priv, 0x41891c, 0xff00ff);
-	nv_wr32(priv, 0x418924, 0x0);
-	nv_wr32(priv, 0x418928, 0xffff00);
-	nv_wr32(priv, 0x41892c, 0xff00);
-	nv_wr32(priv, 0x418a00, 0x0);
-	nv_wr32(priv, 0x418a04, 0x0);
-	nv_wr32(priv, 0x418a08, 0x0);
-	nv_wr32(priv, 0x418a0c, 0x10000);
-	nv_wr32(priv, 0x418a10, 0x0);
-	nv_wr32(priv, 0x418a14, 0x0);
-	nv_wr32(priv, 0x418a18, 0x0);
-	nv_wr32(priv, 0x418a20, 0x0);
-	nv_wr32(priv, 0x418a24, 0x0);
-	nv_wr32(priv, 0x418a28, 0x0);
-	nv_wr32(priv, 0x418a2c, 0x10000);
-	nv_wr32(priv, 0x418a30, 0x0);
-	nv_wr32(priv, 0x418a34, 0x0);
-	nv_wr32(priv, 0x418a38, 0x0);
-	nv_wr32(priv, 0x418a40, 0x0);
-	nv_wr32(priv, 0x418a44, 0x0);
-	nv_wr32(priv, 0x418a48, 0x0);
-	nv_wr32(priv, 0x418a4c, 0x10000);
-	nv_wr32(priv, 0x418a50, 0x0);
-	nv_wr32(priv, 0x418a54, 0x0);
-	nv_wr32(priv, 0x418a58, 0x0);
-	nv_wr32(priv, 0x418a60, 0x0);
-	nv_wr32(priv, 0x418a64, 0x0);
-	nv_wr32(priv, 0x418a68, 0x0);
-	nv_wr32(priv, 0x418a6c, 0x10000);
-	nv_wr32(priv, 0x418a70, 0x0);
-	nv_wr32(priv, 0x418a74, 0x0);
-	nv_wr32(priv, 0x418a78, 0x0);
-	nv_wr32(priv, 0x418a80, 0x0);
-	nv_wr32(priv, 0x418a84, 0x0);
-	nv_wr32(priv, 0x418a88, 0x0);
-	nv_wr32(priv, 0x418a8c, 0x10000);
-	nv_wr32(priv, 0x418a90, 0x0);
-	nv_wr32(priv, 0x418a94, 0x0);
-	nv_wr32(priv, 0x418a98, 0x0);
-	nv_wr32(priv, 0x418aa0, 0x0);
-	nv_wr32(priv, 0x418aa4, 0x0);
-	nv_wr32(priv, 0x418aa8, 0x0);
-	nv_wr32(priv, 0x418aac, 0x10000);
-	nv_wr32(priv, 0x418ab0, 0x0);
-	nv_wr32(priv, 0x418ab4, 0x0);
-	nv_wr32(priv, 0x418ab8, 0x0);
-	nv_wr32(priv, 0x418ac0, 0x0);
-	nv_wr32(priv, 0x418ac4, 0x0);
-	nv_wr32(priv, 0x418ac8, 0x0);
-	nv_wr32(priv, 0x418acc, 0x10000);
-	nv_wr32(priv, 0x418ad0, 0x0);
-	nv_wr32(priv, 0x418ad4, 0x0);
-	nv_wr32(priv, 0x418ad8, 0x0);
-	nv_wr32(priv, 0x418ae0, 0x0);
-	nv_wr32(priv, 0x418ae4, 0x0);
-	nv_wr32(priv, 0x418ae8, 0x0);
-	nv_wr32(priv, 0x418aec, 0x10000);
-	nv_wr32(priv, 0x418af0, 0x0);
-	nv_wr32(priv, 0x418af4, 0x0);
-	nv_wr32(priv, 0x418af8, 0x0);
-	nv_wr32(priv, 0x418b00, 0x6);
-	nv_wr32(priv, 0x418b08, 0xa418820);
-	nv_wr32(priv, 0x418b0c, 0x62080e6);
-	nv_wr32(priv, 0x418b10, 0x20398a4);
-	nv_wr32(priv, 0x418b14, 0xe629062);
-	nv_wr32(priv, 0x418b18, 0xa418820);
-	nv_wr32(priv, 0x418b1c, 0xe6);
-	nv_wr32(priv, 0x418bb8, 0x103);
-	nv_wr32(priv, 0x418c08, 0x1);
-	nv_wr32(priv, 0x418c10, 0x0);
-	nv_wr32(priv, 0x418c14, 0x0);
-	nv_wr32(priv, 0x418c18, 0x0);
-	nv_wr32(priv, 0x418c1c, 0x0);
-	nv_wr32(priv, 0x418c20, 0x0);
-	nv_wr32(priv, 0x418c24, 0x0);
-	nv_wr32(priv, 0x418c28, 0x0);
-	nv_wr32(priv, 0x418c2c, 0x0);
-	nv_wr32(priv, 0x418c40, 0xffffffff);
-	nv_wr32(priv, 0x418c6c, 0x1);
-	nv_wr32(priv, 0x418c80, 0x20200004);
-	nv_wr32(priv, 0x418c8c, 0x1);
-	nv_wr32(priv, 0x419000, 0x780);
-	nv_wr32(priv, 0x419004, 0x0);
-	nv_wr32(priv, 0x419008, 0x0);
-	nv_wr32(priv, 0x419014, 0x4);
-}
-
-static void
-nve0_graph_generate_tpc(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x419848, 0x0);
-	nv_wr32(priv, 0x419864, 0x129);
-	nv_wr32(priv, 0x419888, 0x0);
-	nv_wr32(priv, 0x419a00, 0xf0);
-	nv_wr32(priv, 0x419a04, 0x1);
-	nv_wr32(priv, 0x419a08, 0x21);
-	nv_wr32(priv, 0x419a0c, 0x20000);
-	nv_wr32(priv, 0x419a10, 0x0);
-	nv_wr32(priv, 0x419a14, 0x200);
-	nv_wr32(priv, 0x419a1c, 0xc000);
-	nv_wr32(priv, 0x419a20, 0x800);
-	nv_wr32(priv, 0x419a30, 0x1);
-	nv_wr32(priv, 0x419ac4, 0x37f440);
-	nv_wr32(priv, 0x419c00, 0xa);
-	nv_wr32(priv, 0x419c04, 0x80000006);
-	nv_wr32(priv, 0x419c08, 0x2);
-	nv_wr32(priv, 0x419c20, 0x0);
-	nv_wr32(priv, 0x419c24, 0x84210);
-	nv_wr32(priv, 0x419c28, 0x3efbefbe);
-	nv_wr32(priv, 0x419ce8, 0x0);
-	nv_wr32(priv, 0x419cf4, 0x3203);
-	nv_wr32(priv, 0x419e04, 0x0);
-	nv_wr32(priv, 0x419e08, 0x0);
-	nv_wr32(priv, 0x419e0c, 0x0);
-	nv_wr32(priv, 0x419e10, 0x402);
-	nv_wr32(priv, 0x419e44, 0x13eff2);
-	nv_wr32(priv, 0x419e48, 0x0);
-	nv_wr32(priv, 0x419e4c, 0x7f);
-	nv_wr32(priv, 0x419e50, 0x0);
-	nv_wr32(priv, 0x419e54, 0x0);
-	nv_wr32(priv, 0x419e58, 0x0);
-	nv_wr32(priv, 0x419e5c, 0x0);
-	nv_wr32(priv, 0x419e60, 0x0);
-	nv_wr32(priv, 0x419e64, 0x0);
-	nv_wr32(priv, 0x419e68, 0x0);
-	nv_wr32(priv, 0x419e6c, 0x0);
-	nv_wr32(priv, 0x419e70, 0x0);
-	nv_wr32(priv, 0x419e74, 0x0);
-	nv_wr32(priv, 0x419e78, 0x0);
-	nv_wr32(priv, 0x419e7c, 0x0);
-	nv_wr32(priv, 0x419e80, 0x0);
-	nv_wr32(priv, 0x419e84, 0x0);
-	nv_wr32(priv, 0x419e88, 0x0);
-	nv_wr32(priv, 0x419e8c, 0x0);
-	nv_wr32(priv, 0x419e90, 0x0);
-	nv_wr32(priv, 0x419e94, 0x0);
-	nv_wr32(priv, 0x419e98, 0x0);
-	nv_wr32(priv, 0x419eac, 0x1fcf);
-	nv_wr32(priv, 0x419eb0, 0xd3f);
-	nv_wr32(priv, 0x419ec8, 0x1304f);
-	nv_wr32(priv, 0x419f30, 0x0);
-	nv_wr32(priv, 0x419f34, 0x0);
-	nv_wr32(priv, 0x419f38, 0x0);
-	nv_wr32(priv, 0x419f3c, 0x0);
-	nv_wr32(priv, 0x419f40, 0x0);
-	nv_wr32(priv, 0x419f44, 0x0);
-	nv_wr32(priv, 0x419f48, 0x0);
-	nv_wr32(priv, 0x419f4c, 0x0);
-	nv_wr32(priv, 0x419f58, 0x0);
-	nv_wr32(priv, 0x419f78, 0xb);
-}
-
-static void
-nve0_graph_generate_tpcunk(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x41be24, 0x6);
-	nv_wr32(priv, 0x41bec0, 0x12180000);
-	nv_wr32(priv, 0x41bec4, 0x37f7f);
-	nv_wr32(priv, 0x41bee4, 0x6480430);
-	nv_wr32(priv, 0x41bf00, 0xa418820);
-	nv_wr32(priv, 0x41bf04, 0x62080e6);
-	nv_wr32(priv, 0x41bf08, 0x20398a4);
-	nv_wr32(priv, 0x41bf0c, 0xe629062);
-	nv_wr32(priv, 0x41bf10, 0xa418820);
-	nv_wr32(priv, 0x41bf14, 0xe6);
-	nv_wr32(priv, 0x41bfd0, 0x900103);
-	nv_wr32(priv, 0x41bfe0, 0x400001);
-	nv_wr32(priv, 0x41bfe4, 0x0);
-}
-
-int
-nve0_grctx_generate(struct nvc0_graph_priv *priv)
-{
-	struct nvc0_grctx info;
-	int ret, i, gpc, tpc, id;
-	u32 data[6] = {}, data2[2] = {}, tmp;
-	u32 tpc_set = 0, tpc_mask = 0;
-	u32 magic[GPC_MAX][2], offset;
-	u8 tpcnr[GPC_MAX], a, b;
-	u8 shift, ntpcv;
-
-	ret = nvc0_grctx_init(priv, &info);
-	if (ret)
-		return ret;
-
-	nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
-	nv_wr32(priv, 0x400204, 0x00000000);
-	nv_wr32(priv, 0x400208, 0x00000000);
-
-	nve0_graph_generate_unk40xx(priv);
-	nve0_graph_generate_unk44xx(priv);
-	nve0_graph_generate_unk46xx(priv);
-	nve0_graph_generate_unk47xx(priv);
-	nve0_graph_generate_unk58xx(priv);
-	nve0_graph_generate_unk60xx(priv);
-	nve0_graph_generate_unk64xx(priv);
-	nve0_graph_generate_unk70xx(priv);
-	nve0_graph_generate_unk78xx(priv);
-	nve0_graph_generate_unk80xx(priv);
-	nve0_graph_generate_unk88xx(priv);
-	nve0_graph_generate_gpc(priv);
-	nve0_graph_generate_tpc(priv);
-	nve0_graph_generate_tpcunk(priv);
-
-	nv_wr32(priv, 0x404154, 0x0);
-
-	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
-	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
-	mmio_list(0x40800c, 0x00000000,  8, 1);
-	mmio_list(0x408010, 0x80000000,  0, 0);
-	mmio_list(0x419004, 0x00000000,  8, 1);
-	mmio_list(0x419008, 0x00000000,  0, 0);
-	mmio_list(0x4064cc, 0x80000000,  0, 0);
-	mmio_list(0x408004, 0x00000000,  8, 0);
-	mmio_list(0x408008, 0x80000030,  0, 0);
-	mmio_list(0x418808, 0x00000000,  8, 0);
-	mmio_list(0x41880c, 0x80000030,  0, 0);
-	mmio_list(0x4064c8, 0x01800600,  0, 0);
-	mmio_list(0x418810, 0x80000000, 12, 2);
-	mmio_list(0x419848, 0x10000000, 12, 2);
-	mmio_list(0x405830, 0x02180648,  0, 0);
-	mmio_list(0x4064c4, 0x0192ffff,  0, 0);
-	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
-		u16 magic0 = 0x0218 * priv->tpc_nr[gpc];
-		u16 magic1 = 0x0648 * priv->tpc_nr[gpc];
-		magic[gpc][0]  = 0x10000000 | (magic0 << 16) | offset;
-		magic[gpc][1]  = 0x00000000 | (magic1 << 16);
-		offset += 0x0324 * priv->tpc_nr[gpc];
-	}
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		mmio_list(GPC_UNIT(gpc, 0x30c0), magic[gpc][0], 0, 0);
-		mmio_list(GPC_UNIT(gpc, 0x30e4), magic[gpc][1] | offset, 0, 0);
-		offset += 0x07ff * priv->tpc_nr[gpc];
-	}
-	mmio_list(0x17e91c, 0x06060609, 0, 0);
-	mmio_list(0x17e920, 0x00090a05, 0, 0);
-
-	nv_wr32(priv, 0x418c6c, 0x1);
-	nv_wr32(priv, 0x41980c, 0x10);
-	nv_wr32(priv, 0x41be08, 0x4);
-	nv_wr32(priv, 0x4064c0, 0x801a00f0);
-	nv_wr32(priv, 0x405800, 0xf8000bf);
-	nv_wr32(priv, 0x419c00, 0xa);
-
-	for (tpc = 0, id = 0; tpc < 4; tpc++) {
-		for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-			if (tpc < priv->tpc_nr[gpc]) {
-				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0698), id);
-				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x04e8), id);
-				nv_wr32(priv, GPC_UNIT(gpc, 0x0c10 + tpc * 4), id);
-				nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0088), id++);
-			}
-
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0c08), priv->tpc_nr[gpc]);
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0c8c), priv->tpc_nr[gpc]);
-		}
-	}
-
-	tmp = 0;
-	for (i = 0; i < priv->gpc_nr; i++)
-		tmp |= priv->tpc_nr[i] << (i * 4);
-	nv_wr32(priv, 0x406028, tmp);
-	nv_wr32(priv, 0x405870, tmp);
-
-	nv_wr32(priv, 0x40602c, 0x0);
-	nv_wr32(priv, 0x405874, 0x0);
-	nv_wr32(priv, 0x406030, 0x0);
-	nv_wr32(priv, 0x405878, 0x0);
-	nv_wr32(priv, 0x406034, 0x0);
-	nv_wr32(priv, 0x40587c, 0x0);
-
-	/* calculate first set of magics */
-	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
-
-	gpc = -1;
-	for (tpc = 0; tpc < priv->tpc_total; tpc++) {
-		do {
-			gpc = (gpc + 1) % priv->gpc_nr;
-		} while (!tpcnr[gpc]);
-		tpcnr[gpc]--;
-
-		data[tpc / 6] |= gpc << ((tpc % 6) * 5);
-	}
-
-	for (; tpc < 32; tpc++)
-		data[tpc / 6] |= 7 << ((tpc % 6) * 5);
-
-	/* and the second... */
-	shift = 0;
-	ntpcv = priv->tpc_total;
-	while (!(ntpcv & (1 << 4))) {
-		ntpcv <<= 1;
-		shift++;
-	}
-
-	data2[0]  = ntpcv << 16;
-	data2[0] |= shift << 21;
-	data2[0] |= (((1 << (0 + 5)) % ntpcv) << 24);
-	data2[0] |= priv->tpc_total << 8;
-	data2[0] |= priv->magic_not_rop_nr;
-	for (i = 1; i < 7; i++)
-		data2[1] |= ((1 << (i + 5)) % ntpcv) << ((i - 1) * 5);
-
-	/* and write it all the various parts of PGRAPH */
-	nv_wr32(priv, 0x418bb8, (priv->tpc_total << 8) | priv->magic_not_rop_nr);
-	for (i = 0; i < 6; i++)
-		nv_wr32(priv, 0x418b08 + (i * 4), data[i]);
-
-	nv_wr32(priv, 0x41bfd0, data2[0]);
-	nv_wr32(priv, 0x41bfe4, data2[1]);
-	for (i = 0; i < 6; i++)
-		nv_wr32(priv, 0x41bf00 + (i * 4), data[i]);
-
-	nv_wr32(priv, 0x4078bc, (priv->tpc_total << 8) | priv->magic_not_rop_nr);
-	for (i = 0; i < 6; i++)
-		nv_wr32(priv, 0x40780c + (i * 4), data[i]);
-
-
-	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++)
-		tpc_mask |= ((1 << priv->tpc_nr[gpc]) - 1) << (gpc * 8);
-
-	for (i = 0, gpc = -1, b = -1; i < 32; i++) {
-		a = (i * (priv->tpc_total - 1)) / 32;
-		if (a != b) {
-			b = a;
-			do {
-				gpc = (gpc + 1) % priv->gpc_nr;
-			} while (!tpcnr[gpc]);
-			tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
-
-			tpc_set |= 1 << ((gpc * 8) + tpc);
-		}
-
-		nv_wr32(priv, 0x406800 + (i * 0x20), tpc_set);
-		nv_wr32(priv, 0x406c00 + (i * 0x20), tpc_set ^ tpc_mask);
-	}
-
-	for (i = 0; i < 8; i++)
-		nv_wr32(priv, 0x4064d0 + (i * 0x04), 0x00000000);
-
-	nv_wr32(priv, 0x405b00, (priv->tpc_total << 8) | priv->gpc_nr);
-	if (priv->gpc_nr == 1) {
-		nv_mask(priv, 0x408850, 0x0000000f, priv->tpc_nr[0]);
-		nv_mask(priv, 0x408958, 0x0000000f, priv->tpc_nr[0]);
-	} else {
-		nv_mask(priv, 0x408850, 0x0000000f, priv->gpc_nr);
-		nv_mask(priv, 0x408958, 0x0000000f, priv->gpc_nr);
-	}
-	nv_mask(priv, 0x419f78, 0x00000001, 0x00000000);
-
-	nve0_grctx_generate_icmd(priv);
-	nve0_grctx_generate_a097(priv);
-	nve0_grctx_generate_902d(priv);
-
-	nv_mask(priv, 0x000260, 0x00000001, 0x00000001);
-	nv_wr32(priv, 0x418800, 0x7026860a); //XXX
-	nv_wr32(priv, 0x41be10, 0x00bb8bc7); //XXX
-	return nvc0_grctx_fini(&info);
-}
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve4.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve4.c
new file mode 100644
index 000000000000..e2de73ee5eee
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnve4.c
@@ -0,0 +1,1018 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+struct nvc0_graph_init
+nve4_grctx_init_icmd[] = {
+	{ 0x001000,   1, 0x01, 0x00000004 },
+	{ 0x000039,   3, 0x01, 0x00000000 },
+	{ 0x0000a9,   1, 0x01, 0x0000ffff },
+	{ 0x000038,   1, 0x01, 0x0fac6881 },
+	{ 0x00003d,   1, 0x01, 0x00000001 },
+	{ 0x0000e8,   8, 0x01, 0x00000400 },
+	{ 0x000078,   8, 0x01, 0x00000300 },
+	{ 0x000050,   1, 0x01, 0x00000011 },
+	{ 0x000058,   8, 0x01, 0x00000008 },
+	{ 0x000208,   8, 0x01, 0x00000001 },
+	{ 0x000081,   1, 0x01, 0x00000001 },
+	{ 0x000085,   1, 0x01, 0x00000004 },
+	{ 0x000088,   1, 0x01, 0x00000400 },
+	{ 0x000090,   1, 0x01, 0x00000300 },
+	{ 0x000098,   1, 0x01, 0x00001001 },
+	{ 0x0000e3,   1, 0x01, 0x00000001 },
+	{ 0x0000da,   1, 0x01, 0x00000001 },
+	{ 0x0000f8,   1, 0x01, 0x00000003 },
+	{ 0x0000fa,   1, 0x01, 0x00000001 },
+	{ 0x00009f,   4, 0x01, 0x0000ffff },
+	{ 0x0000b1,   1, 0x01, 0x00000001 },
+	{ 0x0000ad,   1, 0x01, 0x0000013e },
+	{ 0x0000e1,   1, 0x01, 0x00000010 },
+	{ 0x000290,  16, 0x01, 0x00000000 },
+	{ 0x0003b0,  16, 0x01, 0x00000000 },
+	{ 0x0002a0,  16, 0x01, 0x00000000 },
+	{ 0x000420,  16, 0x01, 0x00000000 },
+	{ 0x0002b0,  16, 0x01, 0x00000000 },
+	{ 0x000430,  16, 0x01, 0x00000000 },
+	{ 0x0002c0,  16, 0x01, 0x00000000 },
+	{ 0x0004d0,  16, 0x01, 0x00000000 },
+	{ 0x000720,  16, 0x01, 0x00000000 },
+	{ 0x0008c0,  16, 0x01, 0x00000000 },
+	{ 0x000890,  16, 0x01, 0x00000000 },
+	{ 0x0008e0,  16, 0x01, 0x00000000 },
+	{ 0x0008a0,  16, 0x01, 0x00000000 },
+	{ 0x0008f0,  16, 0x01, 0x00000000 },
+	{ 0x00094c,   1, 0x01, 0x000000ff },
+	{ 0x00094d,   1, 0x01, 0xffffffff },
+	{ 0x00094e,   1, 0x01, 0x00000002 },
+	{ 0x0002ec,   1, 0x01, 0x00000001 },
+	{ 0x000303,   1, 0x01, 0x00000001 },
+	{ 0x0002e6,   1, 0x01, 0x00000001 },
+	{ 0x000466,   1, 0x01, 0x00000052 },
+	{ 0x000301,   1, 0x01, 0x3f800000 },
+	{ 0x000304,   1, 0x01, 0x30201000 },
+	{ 0x000305,   1, 0x01, 0x70605040 },
+	{ 0x000306,   1, 0x01, 0xb8a89888 },
+	{ 0x000307,   1, 0x01, 0xf8e8d8c8 },
+	{ 0x00030a,   1, 0x01, 0x00ffff00 },
+	{ 0x00030b,   1, 0x01, 0x0000001a },
+	{ 0x00030c,   1, 0x01, 0x00000001 },
+	{ 0x000318,   1, 0x01, 0x00000001 },
+	{ 0x000340,   1, 0x01, 0x00000000 },
+	{ 0x000375,   1, 0x01, 0x00000001 },
+	{ 0x00037d,   1, 0x01, 0x00000006 },
+	{ 0x0003a0,   1, 0x01, 0x00000002 },
+	{ 0x0003aa,   1, 0x01, 0x00000001 },
+	{ 0x0003a9,   1, 0x01, 0x00000001 },
+	{ 0x000380,   1, 0x01, 0x00000001 },
+	{ 0x000383,   1, 0x01, 0x00000011 },
+	{ 0x000360,   1, 0x01, 0x00000040 },
+	{ 0x000366,   2, 0x01, 0x00000000 },
+	{ 0x000368,   1, 0x01, 0x00000fff },
+	{ 0x000370,   2, 0x01, 0x00000000 },
+	{ 0x000372,   1, 0x01, 0x000fffff },
+	{ 0x00037a,   1, 0x01, 0x00000012 },
+	{ 0x000619,   1, 0x01, 0x00000003 },
+	{ 0x000811,   1, 0x01, 0x00000003 },
+	{ 0x000812,   1, 0x01, 0x00000004 },
+	{ 0x000813,   1, 0x01, 0x00000006 },
+	{ 0x000814,   1, 0x01, 0x00000008 },
+	{ 0x000815,   1, 0x01, 0x0000000b },
+	{ 0x000800,   6, 0x01, 0x00000001 },
+	{ 0x000632,   1, 0x01, 0x00000001 },
+	{ 0x000633,   1, 0x01, 0x00000002 },
+	{ 0x000634,   1, 0x01, 0x00000003 },
+	{ 0x000635,   1, 0x01, 0x00000004 },
+	{ 0x000654,   1, 0x01, 0x3f800000 },
+	{ 0x000657,   1, 0x01, 0x3f800000 },
+	{ 0x000655,   2, 0x01, 0x3f800000 },
+	{ 0x0006cd,   1, 0x01, 0x3f800000 },
+	{ 0x0007f5,   1, 0x01, 0x3f800000 },
+	{ 0x0007dc,   1, 0x01, 0x39291909 },
+	{ 0x0007dd,   1, 0x01, 0x79695949 },
+	{ 0x0007de,   1, 0x01, 0xb9a99989 },
+	{ 0x0007df,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007e8,   1, 0x01, 0x00003210 },
+	{ 0x0007e9,   1, 0x01, 0x00007654 },
+	{ 0x0007ea,   1, 0x01, 0x00000098 },
+	{ 0x0007ec,   1, 0x01, 0x39291909 },
+	{ 0x0007ed,   1, 0x01, 0x79695949 },
+	{ 0x0007ee,   1, 0x01, 0xb9a99989 },
+	{ 0x0007ef,   1, 0x01, 0xf9e9d9c9 },
+	{ 0x0007f0,   1, 0x01, 0x00003210 },
+	{ 0x0007f1,   1, 0x01, 0x00007654 },
+	{ 0x0007f2,   1, 0x01, 0x00000098 },
+	{ 0x0005a5,   1, 0x01, 0x00000001 },
+	{ 0x000980, 128, 0x01, 0x00000000 },
+	{ 0x000468,   1, 0x01, 0x00000004 },
+	{ 0x00046c,   1, 0x01, 0x00000001 },
+	{ 0x000470,  96, 0x01, 0x00000000 },
+	{ 0x000510,  16, 0x01, 0x3f800000 },
+	{ 0x000520,   1, 0x01, 0x000002b6 },
+	{ 0x000529,   1, 0x01, 0x00000001 },
+	{ 0x000530,  16, 0x01, 0xffff0000 },
+	{ 0x000585,   1, 0x01, 0x0000003f },
+	{ 0x000576,   1, 0x01, 0x00000003 },
+	{ 0x00057b,   1, 0x01, 0x00000059 },
+	{ 0x000586,   1, 0x01, 0x00000040 },
+	{ 0x000582,   2, 0x01, 0x00000080 },
+	{ 0x0005c2,   1, 0x01, 0x00000001 },
+	{ 0x000638,   1, 0x01, 0x00000001 },
+	{ 0x000639,   1, 0x01, 0x00000001 },
+	{ 0x00063a,   1, 0x01, 0x00000002 },
+	{ 0x00063b,   2, 0x01, 0x00000001 },
+	{ 0x00063d,   1, 0x01, 0x00000002 },
+	{ 0x00063e,   1, 0x01, 0x00000001 },
+	{ 0x0008b8,   8, 0x01, 0x00000001 },
+	{ 0x000900,   8, 0x01, 0x00000001 },
+	{ 0x000908,   8, 0x01, 0x00000002 },
+	{ 0x000910,  16, 0x01, 0x00000001 },
+	{ 0x000920,   8, 0x01, 0x00000002 },
+	{ 0x000928,   8, 0x01, 0x00000001 },
+	{ 0x000648,   9, 0x01, 0x00000001 },
+	{ 0x000658,   1, 0x01, 0x0000000f },
+	{ 0x0007ff,   1, 0x01, 0x0000000a },
+	{ 0x00066a,   1, 0x01, 0x40000000 },
+	{ 0x00066b,   1, 0x01, 0x10000000 },
+	{ 0x00066c,   2, 0x01, 0xffff0000 },
+	{ 0x0007af,   2, 0x01, 0x00000008 },
+	{ 0x0007f6,   1, 0x01, 0x00000001 },
+	{ 0x0006b2,   1, 0x01, 0x00000055 },
+	{ 0x0007ad,   1, 0x01, 0x00000003 },
+	{ 0x000937,   1, 0x01, 0x00000001 },
+	{ 0x000971,   1, 0x01, 0x00000008 },
+	{ 0x000972,   1, 0x01, 0x00000040 },
+	{ 0x000973,   1, 0x01, 0x0000012c },
+	{ 0x00097c,   1, 0x01, 0x00000040 },
+	{ 0x000979,   1, 0x01, 0x00000003 },
+	{ 0x000975,   1, 0x01, 0x00000020 },
+	{ 0x000976,   1, 0x01, 0x00000001 },
+	{ 0x000977,   1, 0x01, 0x00000020 },
+	{ 0x000978,   1, 0x01, 0x00000001 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x00095e,   1, 0x01, 0x20164010 },
+	{ 0x00095f,   1, 0x01, 0x00000020 },
+	{ 0x00097d,   1, 0x01, 0x00000020 },
+	{ 0x000683,   1, 0x01, 0x00000006 },
+	{ 0x000685,   1, 0x01, 0x003fffff },
+	{ 0x000687,   1, 0x01, 0x003fffff },
+	{ 0x0006a0,   1, 0x01, 0x00000005 },
+	{ 0x000840,   1, 0x01, 0x00400008 },
+	{ 0x000841,   1, 0x01, 0x08000080 },
+	{ 0x000842,   1, 0x01, 0x00400008 },
+	{ 0x000843,   1, 0x01, 0x08000080 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ab,   1, 0x01, 0x00000002 },
+	{ 0x0006ac,   1, 0x01, 0x00000080 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x0006bb,   1, 0x01, 0x000000cf },
+	{ 0x0006ce,   1, 0x01, 0x2a712488 },
+	{ 0x000739,   1, 0x01, 0x4085c000 },
+	{ 0x00073a,   1, 0x01, 0x00000080 },
+	{ 0x000786,   1, 0x01, 0x80000100 },
+	{ 0x00073c,   1, 0x01, 0x00010100 },
+	{ 0x00073d,   1, 0x01, 0x02800000 },
+	{ 0x000787,   1, 0x01, 0x000000cf },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x000836,   1, 0x01, 0x00000001 },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x000b07,   1, 0x01, 0x00000002 },
+	{ 0x000b08,   2, 0x01, 0x00000100 },
+	{ 0x000b0a,   1, 0x01, 0x00000001 },
+	{ 0x000a04,   1, 0x01, 0x000000ff },
+	{ 0x000a0b,   1, 0x01, 0x00000040 },
+	{ 0x00097f,   1, 0x01, 0x00000100 },
+	{ 0x000a02,   1, 0x01, 0x00000001 },
+	{ 0x000809,   1, 0x01, 0x00000007 },
+	{ 0x00c221,   1, 0x01, 0x00000040 },
+	{ 0x00c1b0,   8, 0x01, 0x0000000f },
+	{ 0x00c1b8,   1, 0x01, 0x0fac6881 },
+	{ 0x00c1b9,   1, 0x01, 0x00fac688 },
+	{ 0x00c401,   1, 0x01, 0x00000001 },
+	{ 0x00c402,   1, 0x01, 0x00010001 },
+	{ 0x00c403,   2, 0x01, 0x00000001 },
+	{ 0x00c40e,   1, 0x01, 0x00000020 },
+	{ 0x00c500,   1, 0x01, 0x00000003 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000002 },
+	{ 0x0006aa,   1, 0x01, 0x00000001 },
+	{ 0x0006ad,   2, 0x01, 0x00000100 },
+	{ 0x0006b1,   1, 0x01, 0x00000011 },
+	{ 0x00078c,   1, 0x01, 0x00000008 },
+	{ 0x000792,   1, 0x01, 0x00000001 },
+	{ 0x000794,   1, 0x01, 0x00000001 },
+	{ 0x000795,   2, 0x01, 0x00000001 },
+	{ 0x000797,   1, 0x01, 0x000000cf },
+	{ 0x00079a,   1, 0x01, 0x00000002 },
+	{ 0x000833,   1, 0x01, 0x04444480 },
+	{ 0x0007a1,   1, 0x01, 0x00000001 },
+	{ 0x0007a3,   1, 0x01, 0x00000001 },
+	{ 0x0007a4,   2, 0x01, 0x00000001 },
+	{ 0x000831,   1, 0x01, 0x00000004 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000008 },
+	{ 0x000039,   3, 0x01, 0x00000000 },
+	{ 0x000380,   1, 0x01, 0x00000001 },
+	{ 0x000366,   2, 0x01, 0x00000000 },
+	{ 0x000368,   1, 0x01, 0x00000fff },
+	{ 0x000370,   2, 0x01, 0x00000000 },
+	{ 0x000372,   1, 0x01, 0x000fffff },
+	{ 0x000813,   1, 0x01, 0x00000006 },
+	{ 0x000814,   1, 0x01, 0x00000008 },
+	{ 0x000957,   1, 0x01, 0x00000003 },
+	{ 0x000b07,   1, 0x01, 0x00000002 },
+	{ 0x000b08,   2, 0x01, 0x00000100 },
+	{ 0x000b0a,   1, 0x01, 0x00000001 },
+	{ 0x000a04,   1, 0x01, 0x000000ff },
+	{ 0x00097f,   1, 0x01, 0x00000100 },
+	{ 0x000a02,   1, 0x01, 0x00000001 },
+	{ 0x000809,   1, 0x01, 0x00000007 },
+	{ 0x00c221,   1, 0x01, 0x00000040 },
+	{ 0x00c401,   1, 0x01, 0x00000001 },
+	{ 0x00c402,   1, 0x01, 0x00010001 },
+	{ 0x00c403,   2, 0x01, 0x00000001 },
+	{ 0x00c40e,   1, 0x01, 0x00000020 },
+	{ 0x00c500,   1, 0x01, 0x00000003 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{ 0x001000,   1, 0x01, 0x00000001 },
+	{ 0x000b07,   1, 0x01, 0x00000002 },
+	{ 0x000b08,   2, 0x01, 0x00000100 },
+	{ 0x000b0a,   1, 0x01, 0x00000001 },
+	{ 0x01e100,   1, 0x01, 0x00000001 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_grctx_init_a097[] = {
+	{ 0x000800,   8, 0x40, 0x00000000 },
+	{ 0x000804,   8, 0x40, 0x00000000 },
+	{ 0x000808,   8, 0x40, 0x00000400 },
+	{ 0x00080c,   8, 0x40, 0x00000300 },
+	{ 0x000810,   1, 0x04, 0x000000cf },
+	{ 0x000850,   7, 0x40, 0x00000000 },
+	{ 0x000814,   8, 0x40, 0x00000040 },
+	{ 0x000818,   8, 0x40, 0x00000001 },
+	{ 0x00081c,   8, 0x40, 0x00000000 },
+	{ 0x000820,   8, 0x40, 0x00000000 },
+	{ 0x001c00,  16, 0x10, 0x00000000 },
+	{ 0x001c04,  16, 0x10, 0x00000000 },
+	{ 0x001c08,  16, 0x10, 0x00000000 },
+	{ 0x001c0c,  16, 0x10, 0x00000000 },
+	{ 0x001d00,  16, 0x10, 0x00000000 },
+	{ 0x001d04,  16, 0x10, 0x00000000 },
+	{ 0x001d08,  16, 0x10, 0x00000000 },
+	{ 0x001d0c,  16, 0x10, 0x00000000 },
+	{ 0x001f00,  16, 0x08, 0x00000000 },
+	{ 0x001f04,  16, 0x08, 0x00000000 },
+	{ 0x001f80,  16, 0x08, 0x00000000 },
+	{ 0x001f84,  16, 0x08, 0x00000000 },
+	{ 0x002000,   1, 0x04, 0x00000000 },
+	{ 0x002040,   1, 0x04, 0x00000011 },
+	{ 0x002080,   1, 0x04, 0x00000020 },
+	{ 0x0020c0,   1, 0x04, 0x00000030 },
+	{ 0x002100,   1, 0x04, 0x00000040 },
+	{ 0x002140,   1, 0x04, 0x00000051 },
+	{ 0x00200c,   6, 0x40, 0x00000001 },
+	{ 0x002010,   1, 0x04, 0x00000000 },
+	{ 0x002050,   1, 0x04, 0x00000000 },
+	{ 0x002090,   1, 0x04, 0x00000001 },
+	{ 0x0020d0,   1, 0x04, 0x00000002 },
+	{ 0x002110,   1, 0x04, 0x00000003 },
+	{ 0x002150,   1, 0x04, 0x00000004 },
+	{ 0x000380,   4, 0x20, 0x00000000 },
+	{ 0x000384,   4, 0x20, 0x00000000 },
+	{ 0x000388,   4, 0x20, 0x00000000 },
+	{ 0x00038c,   4, 0x20, 0x00000000 },
+	{ 0x000700,   4, 0x10, 0x00000000 },
+	{ 0x000704,   4, 0x10, 0x00000000 },
+	{ 0x000708,   4, 0x10, 0x00000000 },
+	{ 0x002800, 128, 0x04, 0x00000000 },
+	{ 0x000a00,  16, 0x20, 0x00000000 },
+	{ 0x000a04,  16, 0x20, 0x00000000 },
+	{ 0x000a08,  16, 0x20, 0x00000000 },
+	{ 0x000a0c,  16, 0x20, 0x00000000 },
+	{ 0x000a10,  16, 0x20, 0x00000000 },
+	{ 0x000a14,  16, 0x20, 0x00000000 },
+	{ 0x000c00,  16, 0x10, 0x00000000 },
+	{ 0x000c04,  16, 0x10, 0x00000000 },
+	{ 0x000c08,  16, 0x10, 0x00000000 },
+	{ 0x000c0c,  16, 0x10, 0x3f800000 },
+	{ 0x000d00,   8, 0x08, 0xffff0000 },
+	{ 0x000d04,   8, 0x08, 0xffff0000 },
+	{ 0x000e00,  16, 0x10, 0x00000000 },
+	{ 0x000e04,  16, 0x10, 0xffff0000 },
+	{ 0x000e08,  16, 0x10, 0xffff0000 },
+	{ 0x000d40,   4, 0x08, 0x00000000 },
+	{ 0x000d44,   4, 0x08, 0x00000000 },
+	{ 0x001e00,   8, 0x20, 0x00000001 },
+	{ 0x001e04,   8, 0x20, 0x00000001 },
+	{ 0x001e08,   8, 0x20, 0x00000002 },
+	{ 0x001e0c,   8, 0x20, 0x00000001 },
+	{ 0x001e10,   8, 0x20, 0x00000001 },
+	{ 0x001e14,   8, 0x20, 0x00000002 },
+	{ 0x001e18,   8, 0x20, 0x00000001 },
+	{ 0x003400, 128, 0x04, 0x00000000 },
+	{ 0x00030c,   1, 0x04, 0x00000001 },
+	{ 0x001944,   1, 0x04, 0x00000000 },
+	{ 0x001514,   1, 0x04, 0x00000000 },
+	{ 0x000d68,   1, 0x04, 0x0000ffff },
+	{ 0x00121c,   1, 0x04, 0x0fac6881 },
+	{ 0x000fac,   1, 0x04, 0x00000001 },
+	{ 0x001538,   1, 0x04, 0x00000001 },
+	{ 0x000fe0,   2, 0x04, 0x00000000 },
+	{ 0x000fe8,   1, 0x04, 0x00000014 },
+	{ 0x000fec,   1, 0x04, 0x00000040 },
+	{ 0x000ff0,   1, 0x04, 0x00000000 },
+	{ 0x00179c,   1, 0x04, 0x00000000 },
+	{ 0x001228,   1, 0x04, 0x00000400 },
+	{ 0x00122c,   1, 0x04, 0x00000300 },
+	{ 0x001230,   1, 0x04, 0x00010001 },
+	{ 0x0007f8,   1, 0x04, 0x00000000 },
+	{ 0x0015b4,   1, 0x04, 0x00000001 },
+	{ 0x0015cc,   1, 0x04, 0x00000000 },
+	{ 0x001534,   1, 0x04, 0x00000000 },
+	{ 0x000fb0,   1, 0x04, 0x00000000 },
+	{ 0x0015d0,   1, 0x04, 0x00000000 },
+	{ 0x00153c,   1, 0x04, 0x00000000 },
+	{ 0x0016b4,   1, 0x04, 0x00000003 },
+	{ 0x000fbc,   4, 0x04, 0x0000ffff },
+	{ 0x000df8,   2, 0x04, 0x00000000 },
+	{ 0x001948,   1, 0x04, 0x00000000 },
+	{ 0x001970,   1, 0x04, 0x00000001 },
+	{ 0x00161c,   1, 0x04, 0x000009f0 },
+	{ 0x000dcc,   1, 0x04, 0x00000010 },
+	{ 0x00163c,   1, 0x04, 0x00000000 },
+	{ 0x0015e4,   1, 0x04, 0x00000000 },
+	{ 0x001160,  32, 0x04, 0x25e00040 },
+	{ 0x001880,  32, 0x04, 0x00000000 },
+	{ 0x000f84,   2, 0x04, 0x00000000 },
+	{ 0x0017c8,   2, 0x04, 0x00000000 },
+	{ 0x0017d0,   1, 0x04, 0x000000ff },
+	{ 0x0017d4,   1, 0x04, 0xffffffff },
+	{ 0x0017d8,   1, 0x04, 0x00000002 },
+	{ 0x0017dc,   1, 0x04, 0x00000000 },
+	{ 0x0015f4,   2, 0x04, 0x00000000 },
+	{ 0x001434,   2, 0x04, 0x00000000 },
+	{ 0x000d74,   1, 0x04, 0x00000000 },
+	{ 0x000dec,   1, 0x04, 0x00000001 },
+	{ 0x0013a4,   1, 0x04, 0x00000000 },
+	{ 0x001318,   1, 0x04, 0x00000001 },
+	{ 0x001644,   1, 0x04, 0x00000000 },
+	{ 0x000748,   1, 0x04, 0x00000000 },
+	{ 0x000de8,   1, 0x04, 0x00000000 },
+	{ 0x001648,   1, 0x04, 0x00000000 },
+	{ 0x0012a4,   1, 0x04, 0x00000000 },
+	{ 0x001120,   4, 0x04, 0x00000000 },
+	{ 0x001118,   1, 0x04, 0x00000000 },
+	{ 0x00164c,   1, 0x04, 0x00000000 },
+	{ 0x001658,   1, 0x04, 0x00000000 },
+	{ 0x001910,   1, 0x04, 0x00000290 },
+	{ 0x001518,   1, 0x04, 0x00000000 },
+	{ 0x00165c,   1, 0x04, 0x00000001 },
+	{ 0x001520,   1, 0x04, 0x00000000 },
+	{ 0x001604,   1, 0x04, 0x00000000 },
+	{ 0x001570,   1, 0x04, 0x00000000 },
+	{ 0x0013b0,   2, 0x04, 0x3f800000 },
+	{ 0x00020c,   1, 0x04, 0x00000000 },
+	{ 0x001670,   1, 0x04, 0x30201000 },
+	{ 0x001674,   1, 0x04, 0x70605040 },
+	{ 0x001678,   1, 0x04, 0xb8a89888 },
+	{ 0x00167c,   1, 0x04, 0xf8e8d8c8 },
+	{ 0x00166c,   1, 0x04, 0x00000000 },
+	{ 0x001680,   1, 0x04, 0x00ffff00 },
+	{ 0x0012d0,   1, 0x04, 0x00000003 },
+	{ 0x0012d4,   1, 0x04, 0x00000002 },
+	{ 0x001684,   2, 0x04, 0x00000000 },
+	{ 0x000dac,   2, 0x04, 0x00001b02 },
+	{ 0x000db4,   1, 0x04, 0x00000000 },
+	{ 0x00168c,   1, 0x04, 0x00000000 },
+	{ 0x0015bc,   1, 0x04, 0x00000000 },
+	{ 0x00156c,   1, 0x04, 0x00000000 },
+	{ 0x00187c,   1, 0x04, 0x00000000 },
+	{ 0x001110,   1, 0x04, 0x00000001 },
+	{ 0x000dc0,   3, 0x04, 0x00000000 },
+	{ 0x001234,   1, 0x04, 0x00000000 },
+	{ 0x001690,   1, 0x04, 0x00000000 },
+	{ 0x0012ac,   1, 0x04, 0x00000001 },
+	{ 0x000790,   5, 0x04, 0x00000000 },
+	{ 0x00077c,   1, 0x04, 0x00000000 },
+	{ 0x001000,   1, 0x04, 0x00000010 },
+	{ 0x0010fc,   1, 0x04, 0x00000000 },
+	{ 0x001290,   1, 0x04, 0x00000000 },
+	{ 0x000218,   1, 0x04, 0x00000010 },
+	{ 0x0012d8,   1, 0x04, 0x00000000 },
+	{ 0x0012dc,   1, 0x04, 0x00000010 },
+	{ 0x000d94,   1, 0x04, 0x00000001 },
+	{ 0x00155c,   2, 0x04, 0x00000000 },
+	{ 0x001564,   1, 0x04, 0x00000fff },
+	{ 0x001574,   2, 0x04, 0x00000000 },
+	{ 0x00157c,   1, 0x04, 0x000fffff },
+	{ 0x001354,   1, 0x04, 0x00000000 },
+	{ 0x001610,   1, 0x04, 0x00000012 },
+	{ 0x001608,   2, 0x04, 0x00000000 },
+	{ 0x00260c,   1, 0x04, 0x00000000 },
+	{ 0x0007ac,   1, 0x04, 0x00000000 },
+	{ 0x00162c,   1, 0x04, 0x00000003 },
+	{ 0x000210,   1, 0x04, 0x00000000 },
+	{ 0x000320,   1, 0x04, 0x00000000 },
+	{ 0x000324,   6, 0x04, 0x3f800000 },
+	{ 0x000750,   1, 0x04, 0x00000000 },
+	{ 0x000760,   1, 0x04, 0x39291909 },
+	{ 0x000764,   1, 0x04, 0x79695949 },
+	{ 0x000768,   1, 0x04, 0xb9a99989 },
+	{ 0x00076c,   1, 0x04, 0xf9e9d9c9 },
+	{ 0x000770,   1, 0x04, 0x30201000 },
+	{ 0x000774,   1, 0x04, 0x70605040 },
+	{ 0x000778,   1, 0x04, 0x00009080 },
+	{ 0x000780,   1, 0x04, 0x39291909 },
+	{ 0x000784,   1, 0x04, 0x79695949 },
+	{ 0x000788,   1, 0x04, 0xb9a99989 },
+	{ 0x00078c,   1, 0x04, 0xf9e9d9c9 },
+	{ 0x0007d0,   1, 0x04, 0x30201000 },
+	{ 0x0007d4,   1, 0x04, 0x70605040 },
+	{ 0x0007d8,   1, 0x04, 0x00009080 },
+	{ 0x00037c,   1, 0x04, 0x00000001 },
+	{ 0x000740,   2, 0x04, 0x00000000 },
+	{ 0x002600,   1, 0x04, 0x00000000 },
+	{ 0x001918,   1, 0x04, 0x00000000 },
+	{ 0x00191c,   1, 0x04, 0x00000900 },
+	{ 0x001920,   1, 0x04, 0x00000405 },
+	{ 0x001308,   1, 0x04, 0x00000001 },
+	{ 0x001924,   1, 0x04, 0x00000000 },
+	{ 0x0013ac,   1, 0x04, 0x00000000 },
+	{ 0x00192c,   1, 0x04, 0x00000001 },
+	{ 0x00193c,   1, 0x04, 0x00002c1c },
+	{ 0x000d7c,   1, 0x04, 0x00000000 },
+	{ 0x000f8c,   1, 0x04, 0x00000000 },
+	{ 0x0002c0,   1, 0x04, 0x00000001 },
+	{ 0x001510,   1, 0x04, 0x00000000 },
+	{ 0x001940,   1, 0x04, 0x00000000 },
+	{ 0x000ff4,   2, 0x04, 0x00000000 },
+	{ 0x00194c,   2, 0x04, 0x00000000 },
+	{ 0x001968,   1, 0x04, 0x00000000 },
+	{ 0x001590,   1, 0x04, 0x0000003f },
+	{ 0x0007e8,   4, 0x04, 0x00000000 },
+	{ 0x00196c,   1, 0x04, 0x00000011 },
+	{ 0x0002e4,   1, 0x04, 0x0000b001 },
+	{ 0x00036c,   2, 0x04, 0x00000000 },
+	{ 0x00197c,   1, 0x04, 0x00000000 },
+	{ 0x000fcc,   2, 0x04, 0x00000000 },
+	{ 0x0002d8,   1, 0x04, 0x00000040 },
+	{ 0x001980,   1, 0x04, 0x00000080 },
+	{ 0x001504,   1, 0x04, 0x00000080 },
+	{ 0x001984,   1, 0x04, 0x00000000 },
+	{ 0x000300,   1, 0x04, 0x00000001 },
+	{ 0x0013a8,   1, 0x04, 0x00000000 },
+	{ 0x0012ec,   1, 0x04, 0x00000000 },
+	{ 0x001310,   1, 0x04, 0x00000000 },
+	{ 0x001314,   1, 0x04, 0x00000001 },
+	{ 0x001380,   1, 0x04, 0x00000000 },
+	{ 0x001384,   4, 0x04, 0x00000001 },
+	{ 0x001394,   1, 0x04, 0x00000000 },
+	{ 0x00139c,   1, 0x04, 0x00000000 },
+	{ 0x001398,   1, 0x04, 0x00000000 },
+	{ 0x001594,   1, 0x04, 0x00000000 },
+	{ 0x001598,   4, 0x04, 0x00000001 },
+	{ 0x000f54,   3, 0x04, 0x00000000 },
+	{ 0x0019bc,   1, 0x04, 0x00000000 },
+	{ 0x000f9c,   2, 0x04, 0x00000000 },
+	{ 0x0012cc,   1, 0x04, 0x00000000 },
+	{ 0x0012e8,   1, 0x04, 0x00000000 },
+	{ 0x00130c,   1, 0x04, 0x00000001 },
+	{ 0x001360,   8, 0x04, 0x00000000 },
+	{ 0x00133c,   2, 0x04, 0x00000001 },
+	{ 0x001344,   1, 0x04, 0x00000002 },
+	{ 0x001348,   2, 0x04, 0x00000001 },
+	{ 0x001350,   1, 0x04, 0x00000002 },
+	{ 0x001358,   1, 0x04, 0x00000001 },
+	{ 0x0012e4,   1, 0x04, 0x00000000 },
+	{ 0x00131c,   1, 0x04, 0x00000000 },
+	{ 0x001320,   3, 0x04, 0x00000000 },
+	{ 0x0019c0,   1, 0x04, 0x00000000 },
+	{ 0x001140,   1, 0x04, 0x00000000 },
+	{ 0x0019c4,   1, 0x04, 0x00000000 },
+	{ 0x0019c8,   1, 0x04, 0x00001500 },
+	{ 0x00135c,   1, 0x04, 0x00000000 },
+	{ 0x000f90,   1, 0x04, 0x00000000 },
+	{ 0x0019e0,   8, 0x04, 0x00000001 },
+	{ 0x0019cc,   1, 0x04, 0x00000001 },
+	{ 0x0015b8,   1, 0x04, 0x00000000 },
+	{ 0x001a00,   1, 0x04, 0x00001111 },
+	{ 0x001a04,   7, 0x04, 0x00000000 },
+	{ 0x000d6c,   2, 0x04, 0xffff0000 },
+	{ 0x0010f8,   1, 0x04, 0x00001010 },
+	{ 0x000d80,   5, 0x04, 0x00000000 },
+	{ 0x000da0,   1, 0x04, 0x00000000 },
+	{ 0x0007a4,   2, 0x04, 0x00000000 },
+	{ 0x001508,   1, 0x04, 0x80000000 },
+	{ 0x00150c,   1, 0x04, 0x40000000 },
+	{ 0x001668,   1, 0x04, 0x00000000 },
+	{ 0x000318,   2, 0x04, 0x00000008 },
+	{ 0x000d9c,   1, 0x04, 0x00000001 },
+	{ 0x000374,   1, 0x04, 0x00000000 },
+	{ 0x000378,   1, 0x04, 0x00000020 },
+	{ 0x0007dc,   1, 0x04, 0x00000000 },
+	{ 0x00074c,   1, 0x04, 0x00000055 },
+	{ 0x001420,   1, 0x04, 0x00000003 },
+	{ 0x0017bc,   2, 0x04, 0x00000000 },
+	{ 0x0017c4,   1, 0x04, 0x00000001 },
+	{ 0x001008,   1, 0x04, 0x00000008 },
+	{ 0x00100c,   1, 0x04, 0x00000040 },
+	{ 0x001010,   1, 0x04, 0x0000012c },
+	{ 0x000d60,   1, 0x04, 0x00000040 },
+	{ 0x00075c,   1, 0x04, 0x00000003 },
+	{ 0x001018,   1, 0x04, 0x00000020 },
+	{ 0x00101c,   1, 0x04, 0x00000001 },
+	{ 0x001020,   1, 0x04, 0x00000020 },
+	{ 0x001024,   1, 0x04, 0x00000001 },
+	{ 0x001444,   3, 0x04, 0x00000000 },
+	{ 0x000360,   1, 0x04, 0x20164010 },
+	{ 0x000364,   1, 0x04, 0x00000020 },
+	{ 0x000368,   1, 0x04, 0x00000000 },
+	{ 0x000de4,   1, 0x04, 0x00000000 },
+	{ 0x000204,   1, 0x04, 0x00000006 },
+	{ 0x000208,   1, 0x04, 0x00000000 },
+	{ 0x0002cc,   2, 0x04, 0x003fffff },
+	{ 0x001220,   1, 0x04, 0x00000005 },
+	{ 0x000fdc,   1, 0x04, 0x00000000 },
+	{ 0x000f98,   1, 0x04, 0x00400008 },
+	{ 0x001284,   1, 0x04, 0x08000080 },
+	{ 0x001450,   1, 0x04, 0x00400008 },
+	{ 0x001454,   1, 0x04, 0x08000080 },
+	{ 0x000214,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_unk40xx[] = {
+	{ 0x404010,   5, 0x04, 0x00000000 },
+	{ 0x404024,   1, 0x04, 0x0000e000 },
+	{ 0x404028,   1, 0x04, 0x00000000 },
+	{ 0x4040a8,   1, 0x04, 0x00000000 },
+	{ 0x4040ac,   7, 0x04, 0x00000000 },
+	{ 0x4040c8,   1, 0x04, 0xf800008f },
+	{ 0x4040d0,   6, 0x04, 0x00000000 },
+	{ 0x4040e8,   1, 0x04, 0x00001000 },
+	{ 0x4040f8,   1, 0x04, 0x00000000 },
+	{ 0x404130,   1, 0x04, 0x00000000 },
+	{ 0x404134,   1, 0x04, 0x00000000 },
+	{ 0x404138,   1, 0x04, 0x20000040 },
+	{ 0x404150,   1, 0x04, 0x0000002e },
+	{ 0x404154,   1, 0x04, 0x00000400 },
+	{ 0x404158,   1, 0x04, 0x00000200 },
+	{ 0x404164,   1, 0x04, 0x00000055 },
+	{ 0x4041a0,   4, 0x04, 0x00000000 },
+	{ 0x404200,   4, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_grctx_init_unk46xx[] = {
+	{ 0x404604,   1, 0x04, 0x00000014 },
+	{ 0x404608,   1, 0x04, 0x00000000 },
+	{ 0x40460c,   1, 0x04, 0x00003fff },
+	{ 0x404610,   1, 0x04, 0x00000100 },
+	{ 0x404618,   4, 0x04, 0x00000000 },
+	{ 0x40462c,   2, 0x04, 0x00000000 },
+	{ 0x404640,   1, 0x04, 0x00000000 },
+	{ 0x404654,   1, 0x04, 0x00000000 },
+	{ 0x404660,   1, 0x04, 0x00000000 },
+	{ 0x404678,   1, 0x04, 0x00000000 },
+	{ 0x40467c,   1, 0x04, 0x00000002 },
+	{ 0x404680,   8, 0x04, 0x00000000 },
+	{ 0x4046a0,   1, 0x04, 0x007f0080 },
+	{ 0x4046a4,   8, 0x04, 0x00000000 },
+	{ 0x4046c8,   3, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_grctx_init_unk47xx[] = {
+	{ 0x404700,   3, 0x04, 0x00000000 },
+	{ 0x404718,   7, 0x04, 0x00000000 },
+	{ 0x404734,   1, 0x04, 0x00000100 },
+	{ 0x404738,   2, 0x04, 0x00000000 },
+	{ 0x404744,   2, 0x04, 0x00000000 },
+	{ 0x404754,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_grctx_init_unk58xx[] = {
+	{ 0x405800,   1, 0x04, 0x0f8000bf },
+	{ 0x405830,   1, 0x04, 0x02180648 },
+	{ 0x405834,   1, 0x04, 0x08000000 },
+	{ 0x405838,   1, 0x04, 0x00000000 },
+	{ 0x405854,   1, 0x04, 0x00000000 },
+	{ 0x405870,   4, 0x04, 0x00000001 },
+	{ 0x405a00,   2, 0x04, 0x00000000 },
+	{ 0x405a18,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_unk5bxx[] = {
+	{ 0x405b00,   1, 0x04, 0x00000000 },
+	{ 0x405b10,   1, 0x04, 0x00001000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_unk60xx[] = {
+	{ 0x406020,   1, 0x04, 0x004103c1 },
+	{ 0x406028,   4, 0x04, 0x00000001 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_unk64xx[] = {
+	{ 0x4064a8,   1, 0x04, 0x00000000 },
+	{ 0x4064ac,   1, 0x04, 0x00003fff },
+	{ 0x4064b4,   2, 0x04, 0x00000000 },
+	{ 0x4064c0,   1, 0x04, 0x801a00f0 },
+	{ 0x4064c4,   1, 0x04, 0x0192ffff },
+	{ 0x4064c8,   1, 0x04, 0x01800600 },
+	{ 0x4064cc,   9, 0x04, 0x00000000 },
+	{ 0x4064fc,   1, 0x04, 0x0000022a },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_unk70xx[] = {
+	{ 0x407040,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_grctx_init_unk80xx[] = {
+	{ 0x408000,   2, 0x04, 0x00000000 },
+	{ 0x408008,   1, 0x04, 0x00000030 },
+	{ 0x40800c,   2, 0x04, 0x00000000 },
+	{ 0x408014,   1, 0x04, 0x00000069 },
+	{ 0x408018,   1, 0x04, 0xe100e100 },
+	{ 0x408064,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_rop[] = {
+	{ 0x408800,   1, 0x04, 0x02802a3c },
+	{ 0x408804,   1, 0x04, 0x00000040 },
+	{ 0x408808,   1, 0x04, 0x1043e005 },
+	{ 0x408840,   1, 0x04, 0x0000000b },
+	{ 0x408900,   1, 0x04, 0x3080b801 },
+	{ 0x408904,   1, 0x04, 0x62000001 },
+	{ 0x408908,   1, 0x04, 0x00c8102f },
+	{ 0x408980,   1, 0x04, 0x0000011d },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_gpc_0[] = {
+	{ 0x418380,   1, 0x04, 0x00000016 },
+	{ 0x418400,   1, 0x04, 0x38004e00 },
+	{ 0x418404,   1, 0x04, 0x71e0ffff },
+	{ 0x41840c,   1, 0x04, 0x00001008 },
+	{ 0x418410,   1, 0x04, 0x0fff0fff },
+	{ 0x418414,   1, 0x04, 0x02200fff },
+	{ 0x418450,   6, 0x04, 0x00000000 },
+	{ 0x418468,   1, 0x04, 0x00000001 },
+	{ 0x41846c,   2, 0x04, 0x00000000 },
+	{ 0x418600,   1, 0x04, 0x0000001f },
+	{ 0x418684,   1, 0x04, 0x0000000f },
+	{ 0x418700,   1, 0x04, 0x00000002 },
+	{ 0x418704,   1, 0x04, 0x00000080 },
+	{ 0x418708,   3, 0x04, 0x00000000 },
+	{ 0x418800,   1, 0x04, 0x7006860a },
+	{ 0x418808,   3, 0x04, 0x00000000 },
+	{ 0x418828,   1, 0x04, 0x00000044 },
+	{ 0x418830,   1, 0x04, 0x10000001 },
+	{ 0x4188d8,   1, 0x04, 0x00000008 },
+	{ 0x4188e0,   1, 0x04, 0x01000000 },
+	{ 0x4188e8,   5, 0x04, 0x00000000 },
+	{ 0x4188fc,   1, 0x04, 0x20100018 },
+	{ 0x41891c,   1, 0x04, 0x00ff00ff },
+	{ 0x418924,   1, 0x04, 0x00000000 },
+	{ 0x418928,   1, 0x04, 0x00ffff00 },
+	{ 0x41892c,   1, 0x04, 0x0000ff00 },
+	{ 0x418b00,   1, 0x04, 0x00000006 },
+	{ 0x418b08,   1, 0x04, 0x0a418820 },
+	{ 0x418b0c,   1, 0x04, 0x062080e6 },
+	{ 0x418b10,   1, 0x04, 0x020398a4 },
+	{ 0x418b14,   1, 0x04, 0x0e629062 },
+	{ 0x418b18,   1, 0x04, 0x0a418820 },
+	{ 0x418b1c,   1, 0x04, 0x000000e6 },
+	{ 0x418bb8,   1, 0x04, 0x00000103 },
+	{ 0x418c08,   1, 0x04, 0x00000001 },
+	{ 0x418c10,   8, 0x04, 0x00000000 },
+	{ 0x418c40,   1, 0x04, 0xffffffff },
+	{ 0x418c6c,   1, 0x04, 0x00000001 },
+	{ 0x418c80,   1, 0x04, 0x20200004 },
+	{ 0x418c8c,   1, 0x04, 0x00000001 },
+	{ 0x419000,   1, 0x04, 0x00000780 },
+	{ 0x419004,   2, 0x04, 0x00000000 },
+	{ 0x419014,   1, 0x04, 0x00000004 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_tpc[] = {
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x00000129 },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000000f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000021 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x0000c000 },
+	{ 0x419a20,   1, 0x04, 0x00000800 },
+	{ 0x419a30,   1, 0x04, 0x00000001 },
+	{ 0x419ac4,   1, 0x04, 0x0037f440 },
+	{ 0x419c00,   1, 0x04, 0x0000000a },
+	{ 0x419c04,   1, 0x04, 0x80000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419c24,   1, 0x04, 0x00084210 },
+	{ 0x419c28,   1, 0x04, 0x3efbefbe },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00003203 },
+	{ 0x419e04,   3, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00000402 },
+	{ 0x419e44,   1, 0x04, 0x0013eff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000007f },
+	{ 0x419e50,  19, 0x04, 0x00000000 },
+	{ 0x419eac,   1, 0x04, 0x00001f8f },
+	{ 0x419eb0,   1, 0x04, 0x00000d3f },
+	{ 0x419ec8,   1, 0x04, 0x0001304f },
+	{ 0x419f30,   8, 0x04, 0x00000000 },
+	{ 0x419f58,   1, 0x04, 0x00000000 },
+	{ 0x419f70,   1, 0x04, 0x00000000 },
+	{ 0x419f78,   1, 0x04, 0x0000000b },
+	{ 0x419f7c,   1, 0x04, 0x0000027a },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_grctx_init_unk[] = {
+	{ 0x41be24,   1, 0x04, 0x00000006 },
+	{ 0x41bec0,   1, 0x04, 0x12180000 },
+	{ 0x41bec4,   1, 0x04, 0x00037f7f },
+	{ 0x41bee4,   1, 0x04, 0x06480430 },
+	{ 0x41bf00,   1, 0x04, 0x0a418820 },
+	{ 0x41bf04,   1, 0x04, 0x062080e6 },
+	{ 0x41bf08,   1, 0x04, 0x020398a4 },
+	{ 0x41bf0c,   1, 0x04, 0x0e629062 },
+	{ 0x41bf10,   1, 0x04, 0x0a418820 },
+	{ 0x41bf14,   1, 0x04, 0x000000e6 },
+	{ 0x41bfd0,   1, 0x04, 0x00900103 },
+	{ 0x41bfe0,   1, 0x04, 0x00400001 },
+	{ 0x41bfe4,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static void
+nve4_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+{
+	u32 magic[GPC_MAX][2];
+	u32 offset;
+	int gpc;
+
+	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
+	mmio_list(0x40800c, 0x00000000,  8, 1);
+	mmio_list(0x408010, 0x80000000,  0, 0);
+	mmio_list(0x419004, 0x00000000,  8, 1);
+	mmio_list(0x419008, 0x00000000,  0, 0);
+	mmio_list(0x4064cc, 0x80000000,  0, 0);
+	mmio_list(0x408004, 0x00000000,  8, 0);
+	mmio_list(0x408008, 0x80000030,  0, 0);
+	mmio_list(0x418808, 0x00000000,  8, 0);
+	mmio_list(0x41880c, 0x80000030,  0, 0);
+	mmio_list(0x4064c8, 0x01800600,  0, 0);
+	mmio_list(0x418810, 0x80000000, 12, 2);
+	mmio_list(0x419848, 0x10000000, 12, 2);
+
+	mmio_list(0x405830, 0x02180648,  0, 0);
+	mmio_list(0x4064c4, 0x0192ffff,  0, 0);
+
+	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
+		u16 magic0 = 0x0218 * priv->tpc_nr[gpc];
+		u16 magic1 = 0x0648 * priv->tpc_nr[gpc];
+		magic[gpc][0]  = 0x10000000 | (magic0 << 16) | offset;
+		magic[gpc][1]  = 0x00000000 | (magic1 << 16);
+		offset += 0x0324 * priv->tpc_nr[gpc];
+	}
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		mmio_list(GPC_UNIT(gpc, 0x30c0), magic[gpc][0], 0, 0);
+		mmio_list(GPC_UNIT(gpc, 0x30e4), magic[gpc][1] | offset, 0, 0);
+		offset += 0x07ff * priv->tpc_nr[gpc];
+	}
+
+	mmio_list(0x17e91c, 0x06060609, 0, 0);
+	mmio_list(0x17e920, 0x00090a05, 0, 0);
+}
+
+void
+nve4_grctx_generate_unkn(struct nvc0_graph_priv *priv)
+{
+	nv_mask(priv, 0x418c6c, 0x00000001, 0x00000001);
+	nv_mask(priv, 0x41980c, 0x00000010, 0x00000010);
+	nv_mask(priv, 0x41be08, 0x00000004, 0x00000004);
+	nv_mask(priv, 0x4064c0, 0x80000000, 0x80000000);
+	nv_mask(priv, 0x405800, 0x08000000, 0x08000000);
+	nv_mask(priv, 0x419c00, 0x00000008, 0x00000008);
+}
+
+void
+nve4_grctx_generate_r418bb8(struct nvc0_graph_priv *priv)
+{
+	u32 data[6] = {}, data2[2] = {};
+	u8  tpcnr[GPC_MAX];
+	u8  shift, ntpcv;
+	int gpc, tpc, i;
+
+	/* calculate first set of magics */
+	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+
+	gpc = -1;
+	for (tpc = 0; tpc < priv->tpc_total; tpc++) {
+		do {
+			gpc = (gpc + 1) % priv->gpc_nr;
+		} while (!tpcnr[gpc]);
+		tpcnr[gpc]--;
+
+		data[tpc / 6] |= gpc << ((tpc % 6) * 5);
+	}
+
+	for (; tpc < 32; tpc++)
+		data[tpc / 6] |= 7 << ((tpc % 6) * 5);
+
+	/* and the second... */
+	shift = 0;
+	ntpcv = priv->tpc_total;
+	while (!(ntpcv & (1 << 4))) {
+		ntpcv <<= 1;
+		shift++;
+	}
+
+	data2[0]  = (ntpcv << 16);
+	data2[0] |= (shift << 21);
+	data2[0] |= (((1 << (0 + 5)) % ntpcv) << 24);
+	for (i = 1; i < 7; i++)
+		data2[1] |= ((1 << (i + 5)) % ntpcv) << ((i - 1) * 5);
+
+	/* GPC_BROADCAST */
+	nv_wr32(priv, 0x418bb8, (priv->tpc_total << 8) |
+				 priv->magic_not_rop_nr);
+	for (i = 0; i < 6; i++)
+		nv_wr32(priv, 0x418b08 + (i * 4), data[i]);
+
+	/* GPC_BROADCAST.TP_BROADCAST */
+	nv_wr32(priv, 0x41bfd0, (priv->tpc_total << 8) |
+				 priv->magic_not_rop_nr | data2[0]);
+	nv_wr32(priv, 0x41bfe4, data2[1]);
+	for (i = 0; i < 6; i++)
+		nv_wr32(priv, 0x41bf00 + (i * 4), data[i]);
+
+	/* UNK78xx */
+	nv_wr32(priv, 0x4078bc, (priv->tpc_total << 8) |
+				 priv->magic_not_rop_nr);
+	for (i = 0; i < 6; i++)
+		nv_wr32(priv, 0x40780c + (i * 4), data[i]);
+}
+
+void
+nve4_grctx_generate_main(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+{
+	struct nvc0_grctx_oclass *oclass = (void *)nv_engine(priv)->cclass;
+	int i;
+
+	nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
+
+	for (i = 0; oclass->hub[i]; i++)
+		nvc0_graph_mmio(priv, oclass->hub[i]);
+	for (i = 0; oclass->gpc[i]; i++)
+		nvc0_graph_mmio(priv, oclass->gpc[i]);
+
+	nv_wr32(priv, 0x404154, 0x00000000);
+
+	oclass->mods(priv, info);
+	oclass->unkn(priv);
+
+	nvc0_grctx_generate_tpcid(priv);
+	nvc0_grctx_generate_r406028(priv);
+	nve4_grctx_generate_r418bb8(priv);
+	nvc0_grctx_generate_r406800(priv);
+
+	for (i = 0; i < 8; i++)
+		nv_wr32(priv, 0x4064d0 + (i * 0x04), 0x00000000);
+
+	nv_wr32(priv, 0x405b00, (priv->tpc_total << 8) | priv->gpc_nr);
+	if (priv->gpc_nr == 1) {
+		nv_mask(priv, 0x408850, 0x0000000f, priv->tpc_nr[0]);
+		nv_mask(priv, 0x408958, 0x0000000f, priv->tpc_nr[0]);
+	} else {
+		nv_mask(priv, 0x408850, 0x0000000f, priv->gpc_nr);
+		nv_mask(priv, 0x408958, 0x0000000f, priv->gpc_nr);
+	}
+	nv_mask(priv, 0x419f78, 0x00000001, 0x00000000);
+
+	nvc0_graph_icmd(priv, oclass->icmd);
+	nv_wr32(priv, 0x404154, 0x00000400);
+	nvc0_graph_mthd(priv, oclass->mthd);
+	nv_mask(priv, 0x000260, 0x00000001, 0x00000001);
+
+	nv_mask(priv, 0x418800, 0x00200000, 0x00200000);
+	nv_mask(priv, 0x41be10, 0x00800000, 0x00800000);
+}
+
+static struct nvc0_graph_init *
+nve4_grctx_init_hub[] = {
+	nvc0_grctx_init_base,
+	nve4_grctx_init_unk40xx,
+	nvc0_grctx_init_unk44xx,
+	nve4_grctx_init_unk46xx,
+	nve4_grctx_init_unk47xx,
+	nve4_grctx_init_unk58xx,
+	nve4_grctx_init_unk5bxx,
+	nve4_grctx_init_unk60xx,
+	nve4_grctx_init_unk64xx,
+	nve4_grctx_init_unk70xx,
+	nvc0_grctx_init_unk78xx,
+	nve4_grctx_init_unk80xx,
+	nve4_grctx_init_rop,
+	NULL
+};
+
+struct nvc0_graph_init *
+nve4_grctx_init_gpc[] = {
+	nve4_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nve4_grctx_init_tpc,
+	nve4_grctx_init_unk,
+	NULL
+};
+
+static struct nvc0_graph_mthd
+nve4_grctx_init_mthd[] = {
+	{ 0xa097, nve4_grctx_init_a097, },
+	{ 0x902d, nvc0_grctx_init_902d, },
+	{ 0x902d, nvc0_grctx_init_mthd_magic, },
+	{}
+};
+
+struct nouveau_oclass *
+nve4_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xe4),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nve4_grctx_generate_main,
+	.mods = nve4_grctx_generate_mods,
+	.unkn = nve4_grctx_generate_unkn,
+	.hub  = nve4_grctx_init_hub,
+	.gpc  = nve4_grctx_init_gpc,
+	.icmd = nve4_grctx_init_icmd,
+	.mthd = nve4_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvf0.c b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvf0.c
new file mode 100644
index 000000000000..dcb2ebb8c29d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/ctxnvf0.c
@@ -0,0 +1,328 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk40xx[] = {
+	{ 0x404004,   8, 0x04, 0x00000000 },
+	{ 0x404024,   1, 0x04, 0x0000e000 },
+	{ 0x404028,   8, 0x04, 0x00000000 },
+	{ 0x4040a8,   8, 0x04, 0x00000000 },
+	{ 0x4040c8,   1, 0x04, 0xf800008f },
+	{ 0x4040d0,   6, 0x04, 0x00000000 },
+	{ 0x4040e8,   1, 0x04, 0x00001000 },
+	{ 0x4040f8,   1, 0x04, 0x00000000 },
+	{ 0x404100,  10, 0x04, 0x00000000 },
+	{ 0x404130,   2, 0x04, 0x00000000 },
+	{ 0x404138,   1, 0x04, 0x20000040 },
+	{ 0x404150,   1, 0x04, 0x0000002e },
+	{ 0x404154,   1, 0x04, 0x00000400 },
+	{ 0x404158,   1, 0x04, 0x00000200 },
+	{ 0x404164,   1, 0x04, 0x00000055 },
+	{ 0x40417c,   2, 0x04, 0x00000000 },
+	{ 0x4041a0,   4, 0x04, 0x00000000 },
+	{ 0x404200,   1, 0x04, 0x0000a197 },
+	{ 0x404204,   1, 0x04, 0x0000a1c0 },
+	{ 0x404208,   1, 0x04, 0x0000a140 },
+	{ 0x40420c,   1, 0x04, 0x0000902d },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk44xx[] = {
+	{ 0x404404,  12, 0x04, 0x00000000 },
+	{ 0x404438,   1, 0x04, 0x00000000 },
+	{ 0x404460,   2, 0x04, 0x00000000 },
+	{ 0x404468,   1, 0x04, 0x00ffffff },
+	{ 0x40446c,   1, 0x04, 0x00000000 },
+	{ 0x404480,   1, 0x04, 0x00000001 },
+	{ 0x404498,   1, 0x04, 0x00000001 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk5bxx[] = {
+	{ 0x405b00,   1, 0x04, 0x00000000 },
+	{ 0x405b10,   1, 0x04, 0x00001000 },
+	{ 0x405b20,   1, 0x04, 0x04000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk60xx[] = {
+	{ 0x406020,   1, 0x04, 0x034103c1 },
+	{ 0x406028,   4, 0x04, 0x00000001 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk64xx[] = {
+	{ 0x4064a8,   1, 0x04, 0x00000000 },
+	{ 0x4064ac,   1, 0x04, 0x00003fff },
+	{ 0x4064b0,   3, 0x04, 0x00000000 },
+	{ 0x4064c0,   1, 0x04, 0x802000f0 },
+	{ 0x4064c4,   1, 0x04, 0x0192ffff },
+	{ 0x4064c8,   1, 0x04, 0x018007c0 },
+	{ 0x4064cc,   9, 0x04, 0x00000000 },
+	{ 0x4064fc,   1, 0x04, 0x0000022a },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk88xx[] = {
+	{ 0x408800,   1, 0x04, 0x12802a3c },
+	{ 0x408804,   1, 0x04, 0x00000040 },
+	{ 0x408808,   1, 0x04, 0x1003e005 },
+	{ 0x408840,   1, 0x04, 0x0000000b },
+	{ 0x408900,   1, 0x04, 0x3080b801 },
+	{ 0x408904,   1, 0x04, 0x62000001 },
+	{ 0x408908,   1, 0x04, 0x00c8102f },
+	{ 0x408980,   1, 0x04, 0x0000011d },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_gpc_0[] = {
+	{ 0x418380,   1, 0x04, 0x00000016 },
+	{ 0x418400,   1, 0x04, 0x38004e00 },
+	{ 0x418404,   1, 0x04, 0x71e0ffff },
+	{ 0x41840c,   1, 0x04, 0x00001008 },
+	{ 0x418410,   1, 0x04, 0x0fff0fff },
+	{ 0x418414,   1, 0x04, 0x02200fff },
+	{ 0x418450,   6, 0x04, 0x00000000 },
+	{ 0x418468,   1, 0x04, 0x00000001 },
+	{ 0x41846c,   2, 0x04, 0x00000000 },
+	{ 0x418600,   1, 0x04, 0x0000001f },
+	{ 0x418684,   1, 0x04, 0x0000000f },
+	{ 0x418700,   1, 0x04, 0x00000002 },
+	{ 0x418704,   1, 0x04, 0x00000080 },
+	{ 0x418708,   3, 0x04, 0x00000000 },
+	{ 0x418800,   1, 0x04, 0x7006860a },
+	{ 0x418808,   1, 0x04, 0x00000000 },
+	{ 0x41880c,   1, 0x04, 0x00000030 },
+	{ 0x418810,   1, 0x04, 0x00000000 },
+	{ 0x418828,   1, 0x04, 0x00000044 },
+	{ 0x418830,   1, 0x04, 0x10000001 },
+	{ 0x4188d8,   1, 0x04, 0x00000008 },
+	{ 0x4188e0,   1, 0x04, 0x01000000 },
+	{ 0x4188e8,   5, 0x04, 0x00000000 },
+	{ 0x4188fc,   1, 0x04, 0x20100018 },
+	{ 0x41891c,   1, 0x04, 0x00ff00ff },
+	{ 0x418924,   1, 0x04, 0x00000000 },
+	{ 0x418928,   1, 0x04, 0x00ffff00 },
+	{ 0x41892c,   1, 0x04, 0x0000ff00 },
+	{ 0x418b00,   1, 0x04, 0x00000006 },
+	{ 0x418b08,   1, 0x04, 0x0a418820 },
+	{ 0x418b0c,   1, 0x04, 0x062080e6 },
+	{ 0x418b10,   1, 0x04, 0x020398a4 },
+	{ 0x418b14,   1, 0x04, 0x0e629062 },
+	{ 0x418b18,   1, 0x04, 0x0a418820 },
+	{ 0x418b1c,   1, 0x04, 0x000000e6 },
+	{ 0x418bb8,   1, 0x04, 0x00000103 },
+	{ 0x418c08,   1, 0x04, 0x00000001 },
+	{ 0x418c10,   8, 0x04, 0x00000000 },
+	{ 0x418c40,   1, 0x04, 0xffffffff },
+	{ 0x418c6c,   1, 0x04, 0x00000001 },
+	{ 0x418c80,   1, 0x04, 0x20200004 },
+	{ 0x418c8c,   1, 0x04, 0x00000001 },
+	{ 0x418d24,   1, 0x04, 0x00000000 },
+	{ 0x419000,   1, 0x04, 0x00000780 },
+	{ 0x419004,   2, 0x04, 0x00000000 },
+	{ 0x419014,   1, 0x04, 0x00000004 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_tpc[] = {
+	{ 0x419848,   1, 0x04, 0x00000000 },
+	{ 0x419864,   1, 0x04, 0x00000129 },
+	{ 0x419888,   1, 0x04, 0x00000000 },
+	{ 0x419a00,   1, 0x04, 0x000000f0 },
+	{ 0x419a04,   1, 0x04, 0x00000001 },
+	{ 0x419a08,   1, 0x04, 0x00000021 },
+	{ 0x419a0c,   1, 0x04, 0x00020000 },
+	{ 0x419a10,   1, 0x04, 0x00000000 },
+	{ 0x419a14,   1, 0x04, 0x00000200 },
+	{ 0x419a1c,   1, 0x04, 0x0000c000 },
+	{ 0x419a20,   1, 0x04, 0x00020800 },
+	{ 0x419a30,   1, 0x04, 0x00000001 },
+	{ 0x419ac4,   1, 0x04, 0x0037f440 },
+	{ 0x419c00,   1, 0x04, 0x0000001a },
+	{ 0x419c04,   1, 0x04, 0x80000006 },
+	{ 0x419c08,   1, 0x04, 0x00000002 },
+	{ 0x419c20,   1, 0x04, 0x00000000 },
+	{ 0x419c24,   1, 0x04, 0x00084210 },
+	{ 0x419c28,   1, 0x04, 0x3efbefbe },
+	{ 0x419ce8,   1, 0x04, 0x00000000 },
+	{ 0x419cf4,   1, 0x04, 0x00000203 },
+	{ 0x419e04,   1, 0x04, 0x00000000 },
+	{ 0x419e08,   1, 0x04, 0x0000001d },
+	{ 0x419e0c,   1, 0x04, 0x00000000 },
+	{ 0x419e10,   1, 0x04, 0x00001c02 },
+	{ 0x419e44,   1, 0x04, 0x0013eff2 },
+	{ 0x419e48,   1, 0x04, 0x00000000 },
+	{ 0x419e4c,   1, 0x04, 0x0000007f },
+	{ 0x419e50,   2, 0x04, 0x00000000 },
+	{ 0x419e58,   1, 0x04, 0x00000001 },
+	{ 0x419e5c,   3, 0x04, 0x00000000 },
+	{ 0x419e68,   1, 0x04, 0x00000002 },
+	{ 0x419e6c,  12, 0x04, 0x00000000 },
+	{ 0x419eac,   1, 0x04, 0x00001fcf },
+	{ 0x419eb0,   1, 0x04, 0x0db00da0 },
+	{ 0x419eb8,   1, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x0001304f },
+	{ 0x419f30,   4, 0x04, 0x00000000 },
+	{ 0x419f40,   1, 0x04, 0x00000018 },
+	{ 0x419f44,   3, 0x04, 0x00000000 },
+	{ 0x419f58,   1, 0x04, 0x00000000 },
+	{ 0x419f70,   1, 0x04, 0x00007300 },
+	{ 0x419f78,   1, 0x04, 0x000000eb },
+	{ 0x419f7c,   1, 0x04, 0x00000404 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_grctx_init_unk[] = {
+	{ 0x41be24,   1, 0x04, 0x00000006 },
+	{ 0x41bec0,   1, 0x04, 0x10000000 },
+	{ 0x41bec4,   1, 0x04, 0x00037f7f },
+	{ 0x41bee4,   1, 0x04, 0x00000000 },
+	{ 0x41bf00,   1, 0x04, 0x0a418820 },
+	{ 0x41bf04,   1, 0x04, 0x062080e6 },
+	{ 0x41bf08,   1, 0x04, 0x020398a4 },
+	{ 0x41bf0c,   1, 0x04, 0x0e629062 },
+	{ 0x41bf10,   1, 0x04, 0x0a418820 },
+	{ 0x41bf14,   1, 0x04, 0x000000e6 },
+	{ 0x41bfd0,   1, 0x04, 0x00900103 },
+	{ 0x41bfe0,   1, 0x04, 0x00400001 },
+	{ 0x41bfe4,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static void
+nvf0_grctx_generate_mods(struct nvc0_graph_priv *priv, struct nvc0_grctx *info)
+{
+	u32 magic[GPC_MAX][4];
+	u32 offset;
+	int gpc;
+
+	mmio_data(0x003000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x008000, 0x0100, NV_MEM_ACCESS_RW | NV_MEM_ACCESS_SYS);
+	mmio_data(0x060000, 0x1000, NV_MEM_ACCESS_RW);
+	mmio_list(0x40800c, 0x00000000,  8, 1);
+	mmio_list(0x408010, 0x80000000,  0, 0);
+	mmio_list(0x419004, 0x00000000,  8, 1);
+	mmio_list(0x419008, 0x00000000,  0, 0);
+	mmio_list(0x4064cc, 0x80000000,  0, 0);
+	mmio_list(0x408004, 0x00000000,  8, 0);
+	mmio_list(0x408008, 0x80000030,  0, 0);
+	mmio_list(0x418808, 0x00000000,  8, 0);
+	mmio_list(0x41880c, 0x80000030,  0, 0);
+	mmio_list(0x4064c8, 0x01800600,  0, 0);
+	mmio_list(0x418810, 0x80000000, 12, 2);
+	mmio_list(0x419848, 0x10000000, 12, 2);
+
+	mmio_list(0x405830, 0x02180648,  0, 0);
+	mmio_list(0x4064c4, 0x0192ffff,  0, 0);
+
+	for (gpc = 0, offset = 0; gpc < priv->gpc_nr; gpc++) {
+		u16 magic0 = 0x0218 * (priv->tpc_nr[gpc] - 1);
+		u16 magic1 = 0x0648 * (priv->tpc_nr[gpc] - 1);
+		u16 magic2 = 0x0218;
+		u16 magic3 = 0x0648;
+		magic[gpc][0]  = 0x10000000 | (magic0 << 16) | offset;
+		magic[gpc][1]  = 0x00000000 | (magic1 << 16);
+		offset += 0x0324 * (priv->tpc_nr[gpc] - 1);;
+		magic[gpc][2]  = 0x10000000 | (magic2 << 16) | offset;
+		magic[gpc][3]  = 0x00000000 | (magic3 << 16);
+		offset += 0x0324;
+	}
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		mmio_list(GPC_UNIT(gpc, 0x30c0), magic[gpc][0], 0, 0);
+		mmio_list(GPC_UNIT(gpc, 0x30e4), magic[gpc][1] | offset, 0, 0);
+		offset += 0x07ff * (priv->tpc_nr[gpc] - 1);
+		mmio_list(GPC_UNIT(gpc, 0x32c0), magic[gpc][2], 0, 0);
+		mmio_list(GPC_UNIT(gpc, 0x32e4), magic[gpc][3] | offset, 0, 0);
+		offset += 0x07ff;
+	}
+
+	mmio_list(0x17e91c, 0x06060609, 0, 0);
+	mmio_list(0x17e920, 0x00090a05, 0, 0);
+}
+
+static struct nvc0_graph_init *
+nvf0_grctx_init_hub[] = {
+	nvc0_grctx_init_base,
+	nvf0_grctx_init_unk40xx,
+	nvf0_grctx_init_unk44xx,
+	nve4_grctx_init_unk46xx,
+	nve4_grctx_init_unk47xx,
+	nve4_grctx_init_unk58xx,
+	nvf0_grctx_init_unk5bxx,
+	nvf0_grctx_init_unk60xx,
+	nvf0_grctx_init_unk64xx,
+	nve4_grctx_init_unk80xx,
+	nvf0_grctx_init_unk88xx,
+	nvd9_grctx_init_rop,
+	NULL
+};
+
+struct nvc0_graph_init *
+nvf0_grctx_init_gpc[] = {
+	nvf0_grctx_init_gpc_0,
+	nvc0_grctx_init_gpc_1,
+	nvf0_grctx_init_tpc,
+	nvf0_grctx_init_unk,
+	NULL
+};
+
+static struct nvc0_graph_mthd
+nvf0_grctx_init_mthd[] = {
+	{ 0xa197, nvc1_grctx_init_9097, },
+	{ 0x902d, nvc0_grctx_init_902d, },
+	{ 0x902d, nvc0_grctx_init_mthd_magic, },
+	{}
+};
+
+struct nouveau_oclass *
+nvf0_grctx_oclass = &(struct nvc0_grctx_oclass) {
+	.base.handle = NV_ENGCTX(GR, 0xf0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_context_ctor,
+		.dtor = nvc0_graph_context_dtor,
+		.init = _nouveau_graph_context_init,
+		.fini = _nouveau_graph_context_fini,
+		.rd32 = _nouveau_graph_context_rd32,
+		.wr32 = _nouveau_graph_context_wr32,
+	},
+	.main = nve4_grctx_generate_main,
+	.mods = nvf0_grctx_generate_mods,
+	.unkn = nve4_grctx_generate_unkn,
+	.hub  = nvf0_grctx_init_hub,
+	.gpc  = nvf0_grctx_init_gpc,
+	.icmd = nvc0_grctx_init_icmd,
+	.mthd = nvf0_grctx_init_mthd,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/nvc0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/com.fuc
index e6b228844a32..5d24b6de16cc 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/nvc0.fuc
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/com.fuc
@@ -23,42 +23,7 @@
  * Authors: Ben Skeggs
  */
 
-define(`mmctx_data', `.b32 eval((($2 - 1) << 26) | $1)')
-define(`queue_init', `.skip eval((2 * 4) + ((8 * 4) * 2))')
-
-ifdef(`include_code', `
-// Error codes
-define(`E_BAD_COMMAND', 0x01)
-define(`E_CMD_OVERFLOW', 0x02)
-
-// Util macros to help with debugging ucode hangs etc
-define(`T_WAIT', 0)
-define(`T_MMCTX', 1)
-define(`T_STRWAIT', 2)
-define(`T_STRINIT', 3)
-define(`T_AUTO', 4)
-define(`T_CHAN', 5)
-define(`T_LOAD', 6)
-define(`T_SAVE', 7)
-define(`T_LCHAN', 8)
-define(`T_LCTXH', 9)
-
-define(`trace_set', `
-	mov $r8 0x83c
-	shl b32 $r8 6
-	clear b32 $r9
-	bset $r9 $1
-	iowr I[$r8 + 0x000] $r9		// CC_SCRATCH[7]
-')
-
-define(`trace_clr', `
-	mov $r8 0x85c
-	shl b32 $r8 6
-	clear b32 $r9
-	bset $r9 $1
-	iowr I[$r8 + 0x000] $r9		// CC_SCRATCH[7]
-')
-
+#ifdef INCLUDE_CODE
 // queue_put - add request to queue
 //
 // In : $r13 queue pointer
@@ -178,27 +143,37 @@ watchdog_clear:
 	iowr I[$r8 + 0x000] $r0
 	ret
 
-// wait_done{z,o} - wait on FUC_DONE bit to become clear/set
+// wait_donez - wait on FUC_DONE bit to become clear
+//
+// In : $r10 bit to wait on
+//
+wait_donez:
+	trace_set(T_WAIT);
+	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_VAL(6), 0, $r10)
+	wait_donez_ne:
+		nv_iord($r8, NV_PGRAPH_FECS_SIGNAL, 0)
+		xbit $r8 $r8 $r10
+		bra ne #wait_donez_ne
+	trace_clr(T_WAIT)
+	ret
+
+// wait_doneo - wait on FUC_DONE bit to become set
 //
 // In : $r10 bit to wait on
 //
-define(`wait_done', `
-$1:
+wait_doneo:
 	trace_set(T_WAIT);
 	mov $r8 0x818
 	shl b32 $r8 6
-	iowr I[$r8 + 0x000] $r10	// CC_SCRATCH[6] = wait bit
-	wait_done_$1:
+	iowr I[$r8 + 0x000] $r10
+	wait_doneo_e:
 		mov $r8 0x400
 		shl b32 $r8 6
-		iord $r8 I[$r8 + 0x000]	// DONE
+		iord $r8 I[$r8 + 0x000]
 		xbit $r8 $r8 $r10
-		bra $2 #wait_done_$1
+		bra e #wait_doneo_e
 	trace_clr(T_WAIT)
 	ret
-')
-wait_done(wait_donez, ne)
-wait_done(wait_doneo, e)
 
 // mmctx_size - determine size of a mmio list transfer
 //
@@ -397,4 +372,4 @@ strand_ctx_init:
 	sub b32 $r15 $r14 $r15
 	trace_clr(T_STRINIT)
 	ret
-')
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpc.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpc.fuc
new file mode 100644
index 000000000000..5547c1b3f4f2
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpc.fuc
@@ -0,0 +1,404 @@
+/* fuc microcode for nvc0 PGRAPH/GPC
+ *
+ * Copyright 2011 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+/* TODO
+ * - bracket certain functions with scratch writes, useful for debugging
+ * - watchdog timer around ctx operations
+ */
+
+#ifdef INCLUDE_DATA
+gpc_mmio_list_head:	.b32 #mmio_list_base
+gpc_mmio_list_tail:
+tpc_mmio_list_head:	.b32 #mmio_list_base
+tpc_mmio_list_tail:
+unk_mmio_list_head:	.b32 #mmio_list_base
+unk_mmio_list_tail:	.b32 #mmio_list_base
+
+gpc_id:			.b32 0
+
+tpc_count:		.b32 0
+tpc_mask:		.b32 0
+
+#if NV_PGRAPH_GPCX_UNK__SIZE > 0
+unk_count:		.b32 0
+unk_mask:		.b32 0
+#endif
+
+cmd_queue:		queue_init
+
+mmio_list_base:
+#endif
+
+#ifdef INCLUDE_CODE
+// reports an exception to the host
+//
+// In: $r15 error code (see nvc0.fuc)
+//
+error:
+	push $r14
+	mov $r14 -0x67ec 	// 0x9814
+	sethi $r14 0x400000
+	call #nv_wr32		// HUB_CTXCTL_CC_SCRATCH[5] = error code
+	add b32 $r14 0x41c
+	mov $r15 1
+	call #nv_wr32		// HUB_CTXCTL_INTR_UP_SET
+	pop $r14
+	ret
+
+// GPC fuc initialisation, executed by triggering ucode start, will
+// fall through to main loop after completion.
+//
+// Input:
+//   CC_SCRATCH[1]: context base
+//
+// Output:
+//   CC_SCRATCH[0]:
+//	     31:31: set to signal completion
+//   CC_SCRATCH[1]:
+//	      31:0: GPC context size
+//
+init:
+	clear b32 $r0
+	mov $sp $r0
+
+	// enable fifo access
+	mov $r1 0x1200
+	mov $r2 2
+	iowr I[$r1 + 0x000] $r2		// FIFO_ENABLE
+
+	// setup i0 handler, and route all interrupts to it
+	mov $r1 #ih
+	mov $iv0 $r1
+	mov $r1 0x400
+	iowr I[$r1 + 0x300] $r0		// INTR_DISPATCH
+
+	// enable fifo interrupt
+	mov $r2 4
+	iowr I[$r1 + 0x000] $r2		// INTR_EN_SET
+
+	// enable interrupts
+	bset $flags ie0
+
+	// figure out which GPC we are, and how many TPCs we have
+	mov $r1 0x608
+	shl b32 $r1 6
+	iord $r2 I[$r1 + 0x000]		// UNITS
+	mov $r3 1
+	and $r2 0x1f
+	shl b32 $r3 $r2
+	sub b32 $r3 1
+	st b32 D[$r0 + #tpc_count] $r2
+	st b32 D[$r0 + #tpc_mask] $r3
+	add b32 $r1 0x400
+	iord $r2 I[$r1 + 0x000]		// MYINDEX
+	st b32 D[$r0 + #gpc_id] $r2
+
+#if NV_PGRAPH_GPCX_UNK__SIZE > 0
+	// figure out which, and how many, UNKs are actually present
+	mov $r14 0x0c30
+	sethi $r14 0x500000
+	clear b32 $r2
+	clear b32 $r3
+	clear b32 $r4
+	init_unk_loop:
+		call #nv_rd32
+		cmp b32 $r15 0
+		bra z #init_unk_next
+			mov $r15 1
+			shl b32 $r15 $r2
+			or $r4 $r15
+			add b32 $r3 1
+		init_unk_next:
+		add b32 $r2 1
+		add b32 $r14 4
+		cmp b32 $r2 NV_PGRAPH_GPCX_UNK__SIZE
+		bra ne #init_unk_loop
+	init_unk_done:
+	st b32 D[$r0 + #unk_count] $r3
+	st b32 D[$r0 + #unk_mask] $r4
+#endif
+
+	// initialise context base, and size tracking
+	nv_iord($r2, NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_VAL(1), 0)
+	clear b32 $r3		// track GPC context size here
+
+	// set mmctx base addresses now so we don't have to do it later,
+	// they don't currently ever change
+	mov $r4 0x700
+	shl b32 $r4 6
+	shr b32 $r5 $r2 8
+	iowr I[$r4 + 0x000] $r5		// MMCTX_SAVE_SWBASE
+	iowr I[$r4 + 0x100] $r5		// MMCTX_LOAD_SWBASE
+
+	// calculate GPC mmio context size
+	ld b32 $r14 D[$r0 + #gpc_mmio_list_head]
+	ld b32 $r15 D[$r0 + #gpc_mmio_list_tail]
+	call #mmctx_size
+	add b32 $r2 $r15
+	add b32 $r3 $r15
+
+	// calculate per-TPC mmio context size
+	ld b32 $r14 D[$r0 + #tpc_mmio_list_head]
+	ld b32 $r15 D[$r0 + #tpc_mmio_list_tail]
+	call #mmctx_size
+	ld b32 $r14 D[$r0 + #tpc_count]
+	mulu $r14 $r15
+	add b32 $r2 $r14
+	add b32 $r3 $r14
+
+#if NV_PGRAPH_GPCX_UNK__SIZE > 0
+	// calculate per-UNK mmio context size
+	ld b32 $r14 D[$r0 + #unk_mmio_list_head]
+	ld b32 $r15 D[$r0 + #unk_mmio_list_tail]
+	call #mmctx_size
+	ld b32 $r14 D[$r0 + #unk_count]
+	mulu $r14 $r15
+	add b32 $r2 $r14
+	add b32 $r3 $r14
+#endif
+
+	// round up base/size to 256 byte boundary (for strand SWBASE)
+	add b32 $r4 0x1300
+	shr b32 $r3 2
+	iowr I[$r4 + 0x000] $r3		// MMCTX_LOAD_COUNT, wtf for?!?
+	shr b32 $r2 8
+	shr b32 $r3 6
+	add b32 $r2 1
+	add b32 $r3 1
+	shl b32 $r2 8
+	shl b32 $r3 8
+
+	// calculate size of strand context data
+	mov b32 $r15 $r2
+	call #strand_ctx_init
+	add b32 $r3 $r15
+
+	// save context size, and tell HUB we're done
+	nv_iowr(NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_VAL(1), 0, $r3)
+	clear b32 $r2
+	bset $r2 31
+	nv_iowr(NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_SET(0), 0, $r2)
+
+// Main program loop, very simple, sleeps until woken up by the interrupt
+// handler, pulls a command from the queue and executes its handler
+//
+main:
+	bset $flags $p0
+	sleep $p0
+	mov $r13 #cmd_queue
+	call #queue_get
+	bra $p1 #main
+
+	// 0x0000-0x0003 are all context transfers
+	cmpu b32 $r14 0x04
+	bra nc #main_not_ctx_xfer
+		// fetch $flags and mask off $p1/$p2
+		mov $r1 $flags
+		mov $r2 0x0006
+		not b32 $r2
+		and $r1 $r2
+		// set $p1/$p2 according to transfer type
+		shl b32 $r14 1
+		or $r1 $r14
+		mov $flags $r1
+		// transfer context data
+		call #ctx_xfer
+		bra #main
+
+	main_not_ctx_xfer:
+	shl b32 $r15 $r14 16
+	or $r15 E_BAD_COMMAND
+	call #error
+	bra #main
+
+// interrupt handler
+ih:
+	push $r8
+	mov $r8 $flags
+	push $r8
+	push $r9
+	push $r10
+	push $r11
+	push $r13
+	push $r14
+	push $r15
+	clear b32 $r0
+
+	// incoming fifo command?
+	iord $r10 I[$r0 + 0x200]	// INTR
+	and $r11 $r10 0x00000004
+	bra e #ih_no_fifo
+		// queue incoming fifo command for later processing
+		mov $r11 0x1900
+		mov $r13 #cmd_queue
+		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
+		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
+		call #queue_put
+		add b32 $r11 0x400
+		mov $r14 1
+		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
+
+	// ack, and wake up main()
+	ih_no_fifo:
+	iowr I[$r0 + 0x100] $r10	// INTR_ACK
+
+	pop $r15
+	pop $r14
+	pop $r13
+	pop $r11
+	pop $r10
+	pop $r9
+	pop $r8
+	mov $flags $r8
+	pop $r8
+	bclr $flags $p0
+	iret
+
+// Set this GPC's bit in HUB_BAR, used to signal completion of various
+// activities to the HUB fuc
+//
+hub_barrier_done:
+	mov $r15 1
+	ld b32 $r14 D[$r0 + #gpc_id]
+	shl b32 $r15 $r14
+	mov $r14 -0x6be8 	// 0x409418 - HUB_BAR_SET
+	sethi $r14 0x400000
+	call #nv_wr32
+	ret
+
+// Disables various things, waits a bit, and re-enables them..
+//
+// Not sure how exactly this helps, perhaps "ENABLE" is not such a
+// good description for the bits we turn off?  Anyways, without this,
+// funny things happen.
+//
+ctx_redswitch:
+	mov $r14 0x614
+	shl b32 $r14 6
+	mov $r15 0x020
+	iowr I[$r14] $r15	// GPC_RED_SWITCH = POWER
+	mov $r15 8
+	ctx_redswitch_delay:
+		sub b32 $r15 1
+		bra ne #ctx_redswitch_delay
+	mov $r15 0xa20
+	iowr I[$r14] $r15	// GPC_RED_SWITCH = UNK11, ENABLE, POWER
+	ret
+
+// Transfer GPC context data between GPU and storage area
+//
+// In: $r15 context base address
+//     $p1 clear on save, set on load
+//     $p2 set if opposite direction done/will be done, so:
+//		on save it means: "a load will follow this save"
+//		on load it means: "a save preceeded this load"
+//
+ctx_xfer:
+	// set context base address
+	mov $r1 0xa04
+	shl b32 $r1 6
+	iowr I[$r1 + 0x000] $r15// MEM_BASE
+	bra not $p1 #ctx_xfer_not_load
+		call #ctx_redswitch
+	ctx_xfer_not_load:
+
+	// strands
+	mov $r1 0x4afc
+	sethi $r1 0x20000
+	mov $r2 0xc
+	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
+	call #strand_wait
+	mov $r2 0x47fc
+	sethi $r2 0x20000
+	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
+	xbit $r2 $flags $p1
+	add b32 $r2 3
+	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
+
+	// mmio context
+	xbit $r10 $flags $p1	// direction
+	or $r10 2		// first
+	mov $r11 0x0000
+	sethi $r11 0x500000
+	ld b32 $r12 D[$r0 + #gpc_id]
+	shl b32 $r12 15
+	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn
+	ld b32 $r12 D[$r0 + #gpc_mmio_list_head]
+	ld b32 $r13 D[$r0 + #gpc_mmio_list_tail]
+	mov $r14 0		// not multi
+	call #mmctx_xfer
+
+	// per-TPC mmio context
+	xbit $r10 $flags $p1	// direction
+#if !NV_PGRAPH_GPCX_UNK__SIZE
+	or $r10 4		// last
+#endif
+	mov $r11 0x4000
+	sethi $r11 0x500000	// base = NV_PGRAPH_GPC0_TPC0
+	ld b32 $r12 D[$r0 + #gpc_id]
+	shl b32 $r12 15
+	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn_TPC0
+	ld b32 $r12 D[$r0 + #tpc_mmio_list_head]
+	ld b32 $r13 D[$r0 + #tpc_mmio_list_tail]
+	ld b32 $r15 D[$r0 + #tpc_mask]
+	mov $r14 0x800		// stride = 0x800
+	call #mmctx_xfer
+
+#if NV_PGRAPH_GPCX_UNK__SIZE > 0
+	// per-UNK mmio context
+	xbit $r10 $flags $p1	// direction
+	or $r10 4		// last
+	mov $r11 0x3000
+	sethi $r11 0x500000	// base = NV_PGRAPH_GPC0_UNK0
+	ld b32 $r12 D[$r0 + #gpc_id]
+	shl b32 $r12 15
+	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn_UNK0
+	ld b32 $r12 D[$r0 + #unk_mmio_list_head]
+	ld b32 $r13 D[$r0 + #unk_mmio_list_tail]
+	ld b32 $r15 D[$r0 + #unk_mask]
+	mov $r14 0x200		// stride = 0x200
+	call #mmctx_xfer
+#endif
+
+	// wait for strands to finish
+	call #strand_wait
+
+	// if load, or a save without a load following, do some
+	// unknown stuff that's done after finishing a block of
+	// strand commands
+	bra $p1 #ctx_xfer_post
+	bra not $p2 #ctx_xfer_done
+	ctx_xfer_post:
+		mov $r1 0x4afc
+		sethi $r1 0x20000
+		mov $r2 0xd
+		iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0d
+		call #strand_wait
+
+	// mark completion in HUB's barrier
+	ctx_xfer_done:
+	call #hub_barrier_done
+	ret
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc
index f7055af0f2a6..5ae06a2d64c9 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc
@@ -1,6 +1,5 @@
-/* fuc microcode for nvc0 PGRAPH/GPC
- *
- * Copyright 2011 Red Hat Inc.
+/*
+ * Copyright 2013 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -20,525 +19,24 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
- * Authors: Ben Skeggs
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
-/* To build:
- *    m4 gpcnvc0.fuc | envyas -a -w -m fuc -V fuc3 -o gpcnvc0.fuc.h
- */
+#define NV_PGRAPH_GPCX_UNK__SIZE                                     0x00000000
 
-/* TODO
- * - bracket certain functions with scratch writes, useful for debugging
- * - watchdog timer around ctx operations
- */
+#define CHIPSET GF100
+#include "macros.fuc"
 
 .section #nvc0_grgpc_data
-include(`nvc0.fuc')
-gpc_id:			.b32 0
-gpc_mmio_list_head:	.b32 0
-gpc_mmio_list_tail:	.b32 0
-
-tpc_count:		.b32 0
-tpc_mask:		.b32 0
-tpc_mmio_list_head:	.b32 0
-tpc_mmio_list_tail:	.b32 0
-
-cmd_queue:		queue_init
-
-// chipset descriptions
-chipsets:
-.b8  0xc0 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc0_tpc_mmio_tail
-.b8  0xc1 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc1_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc1_tpc_mmio_tail
-.b8  0xc3 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc3_tpc_mmio_tail
-.b8  0xc4 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc3_tpc_mmio_tail
-.b8  0xc8 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc0_tpc_mmio_tail
-.b8  0xce 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc3_tpc_mmio_tail
-.b8  0xcf 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvcf_tpc_mmio_tail
-.b8  0xd9 0 0 0
-.b16 #nvd9_gpc_mmio_head
-.b16 #nvd9_gpc_mmio_tail
-.b16 #nvd9_tpc_mmio_head
-.b16 #nvd9_tpc_mmio_tail
-.b8  0xd7 0 0 0
-.b16 #nvd9_gpc_mmio_head
-.b16 #nvd9_gpc_mmio_tail
-.b16 #nvd9_tpc_mmio_head
-.b16 #nvd9_tpc_mmio_tail
-.b8  0 0 0 0
-
-// GPC mmio lists
-nvc0_gpc_mmio_head:
-mmctx_data(0x000380, 1)
-mmctx_data(0x000400, 6)
-mmctx_data(0x000450, 9)
-mmctx_data(0x000600, 1)
-mmctx_data(0x000684, 1)
-mmctx_data(0x000700, 5)
-mmctx_data(0x000800, 1)
-mmctx_data(0x000808, 3)
-mmctx_data(0x000828, 1)
-mmctx_data(0x000830, 1)
-mmctx_data(0x0008d8, 1)
-mmctx_data(0x0008e0, 1)
-mmctx_data(0x0008e8, 6)
-mmctx_data(0x00091c, 1)
-mmctx_data(0x000924, 3)
-mmctx_data(0x000b00, 1)
-mmctx_data(0x000b08, 6)
-mmctx_data(0x000bb8, 1)
-mmctx_data(0x000c08, 1)
-mmctx_data(0x000c10, 8)
-mmctx_data(0x000c80, 1)
-mmctx_data(0x000c8c, 1)
-mmctx_data(0x001000, 3)
-mmctx_data(0x001014, 1)
-nvc0_gpc_mmio_tail:
-mmctx_data(0x000c6c, 1);
-nvc1_gpc_mmio_tail:
-
-nvd9_gpc_mmio_head:
-mmctx_data(0x000380, 1)
-mmctx_data(0x000400, 2)
-mmctx_data(0x00040c, 3)
-mmctx_data(0x000450, 9)
-mmctx_data(0x000600, 1)
-mmctx_data(0x000684, 1)
-mmctx_data(0x000700, 5)
-mmctx_data(0x000800, 1)
-mmctx_data(0x000808, 3)
-mmctx_data(0x000828, 1)
-mmctx_data(0x000830, 1)
-mmctx_data(0x0008d8, 1)
-mmctx_data(0x0008e0, 1)
-mmctx_data(0x0008e8, 6)
-mmctx_data(0x00091c, 1)
-mmctx_data(0x000924, 3)
-mmctx_data(0x000b00, 1)
-mmctx_data(0x000b08, 6)
-mmctx_data(0x000bb8, 1)
-mmctx_data(0x000c08, 1)
-mmctx_data(0x000c10, 8)
-mmctx_data(0x000c6c, 1)
-mmctx_data(0x000c80, 1)
-mmctx_data(0x000c8c, 1)
-mmctx_data(0x001000, 3)
-mmctx_data(0x001014, 1)
-nvd9_gpc_mmio_tail:
-
-// TPC mmio lists
-nvc0_tpc_mmio_head:
-mmctx_data(0x000018, 1)
-mmctx_data(0x00003c, 1)
-mmctx_data(0x000048, 1)
-mmctx_data(0x000064, 1)
-mmctx_data(0x000088, 1)
-mmctx_data(0x000200, 6)
-mmctx_data(0x00021c, 2)
-mmctx_data(0x000300, 6)
-mmctx_data(0x0003d0, 1)
-mmctx_data(0x0003e0, 2)
-mmctx_data(0x000400, 3)
-mmctx_data(0x000420, 1)
-mmctx_data(0x0004b0, 1)
-mmctx_data(0x0004e8, 1)
-mmctx_data(0x0004f4, 1)
-mmctx_data(0x000520, 2)
-mmctx_data(0x000604, 4)
-mmctx_data(0x000644, 20)
-mmctx_data(0x000698, 1)
-mmctx_data(0x000750, 2)
-nvc0_tpc_mmio_tail:
-mmctx_data(0x000758, 1)
-mmctx_data(0x0002c4, 1)
-mmctx_data(0x0006e0, 1)
-nvcf_tpc_mmio_tail:
-mmctx_data(0x0004bc, 1)
-nvc3_tpc_mmio_tail:
-mmctx_data(0x000544, 1)
-nvc1_tpc_mmio_tail:
-
-nvd9_tpc_mmio_head:
-mmctx_data(0x000018, 1)
-mmctx_data(0x00003c, 1)
-mmctx_data(0x000048, 1)
-mmctx_data(0x000064, 1)
-mmctx_data(0x000088, 1)
-mmctx_data(0x000200, 6)
-mmctx_data(0x00021c, 2)
-mmctx_data(0x0002c4, 1)
-mmctx_data(0x000300, 6)
-mmctx_data(0x0003d0, 1)
-mmctx_data(0x0003e0, 2)
-mmctx_data(0x000400, 3)
-mmctx_data(0x000420, 3)
-mmctx_data(0x0004b0, 1)
-mmctx_data(0x0004e8, 1)
-mmctx_data(0x0004f4, 1)
-mmctx_data(0x000520, 2)
-mmctx_data(0x000544, 1)
-mmctx_data(0x000604, 4)
-mmctx_data(0x000644, 20)
-mmctx_data(0x000698, 1)
-mmctx_data(0x0006e0, 1)
-mmctx_data(0x000750, 3)
-nvd9_tpc_mmio_tail:
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "gpc.fuc"
+#undef INCLUDE_DATA
 
 .section #nvc0_grgpc_code
+#define INCLUDE_CODE
 bra #init
-define(`include_code')
-include(`nvc0.fuc')
-
-// reports an exception to the host
-//
-// In: $r15 error code (see nvc0.fuc)
-//
-error:
-	push $r14
-	mov $r14 -0x67ec 	// 0x9814
-	sethi $r14 0x400000
-	call #nv_wr32		// HUB_CTXCTL_CC_SCRATCH[5] = error code
-	add b32 $r14 0x41c
-	mov $r15 1
-	call #nv_wr32		// HUB_CTXCTL_INTR_UP_SET
-	pop $r14
-	ret
-
-// GPC fuc initialisation, executed by triggering ucode start, will
-// fall through to main loop after completion.
-//
-// Input:
-//   CC_SCRATCH[0]: chipset (PMC_BOOT_0 read returns 0x0bad0bad... sigh)
-//   CC_SCRATCH[1]: context base
-//
-// Output:
-//   CC_SCRATCH[0]:
-//	     31:31: set to signal completion
-//   CC_SCRATCH[1]:
-//	      31:0: GPC context size
-//
-init:
-	clear b32 $r0
-	mov $sp $r0
-
-	// enable fifo access
-	mov $r1 0x1200
-	mov $r2 2
-	iowr I[$r1 + 0x000] $r2		// FIFO_ENABLE
-
-	// setup i0 handler, and route all interrupts to it
-	mov $r1 #ih
-	mov $iv0 $r1
-	mov $r1 0x400
-	iowr I[$r1 + 0x300] $r0		// INTR_DISPATCH
-
-	// enable fifo interrupt
-	mov $r2 4
-	iowr I[$r1 + 0x000] $r2		// INTR_EN_SET
-
-	// enable interrupts
-	bset $flags ie0
-
-	// figure out which GPC we are, and how many TPCs we have
-	mov $r1 0x608
-	shl b32 $r1 6
-	iord $r2 I[$r1 + 0x000]		// UNITS
-	mov $r3 1
-	and $r2 0x1f
-	shl b32 $r3 $r2
-	sub b32 $r3 1
-	st b32 D[$r0 + #tpc_count] $r2
-	st b32 D[$r0 + #tpc_mask] $r3
-	add b32 $r1 0x400
-	iord $r2 I[$r1 + 0x000]		// MYINDEX
-	st b32 D[$r0 + #gpc_id] $r2
-
-	// find context data for this chipset
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x000]		// CC_SCRATCH[0]
-	mov $r1 #chipsets - 12
-	init_find_chipset:
-		add b32 $r1 12
-		ld b32 $r3 D[$r1 + 0x00]
-		cmpu b32 $r3 $r2
-		bra e #init_context
-		cmpu b32 $r3 0
-		bra ne #init_find_chipset
-		// unknown chipset
-		ret
-
-	// initialise context base, and size tracking
-	init_context:
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x100]	// CC_SCRATCH[1], initial base
-	clear b32 $r3		// track GPC context size here
-
-	// set mmctx base addresses now so we don't have to do it later,
-	// they don't currently ever change
-	mov $r4 0x700
-	shl b32 $r4 6
-	shr b32 $r5 $r2 8
-	iowr I[$r4 + 0x000] $r5		// MMCTX_SAVE_SWBASE
-	iowr I[$r4 + 0x100] $r5		// MMCTX_LOAD_SWBASE
-
-	// calculate GPC mmio context size, store the chipset-specific
-	// mmio list pointers somewhere we can get at them later without
-	// re-parsing the chipset list
-	clear b32 $r14
-	clear b32 $r15
-	ld b16 $r14 D[$r1 + 4]
-	ld b16 $r15 D[$r1 + 6]
-	st b16 D[$r0 + #gpc_mmio_list_head] $r14
-	st b16 D[$r0 + #gpc_mmio_list_tail] $r15
-	call #mmctx_size
-	add b32 $r2 $r15
-	add b32 $r3 $r15
-
-	// calculate per-TPC mmio context size, store the list pointers
-	ld b16 $r14 D[$r1 + 8]
-	ld b16 $r15 D[$r1 + 10]
-	st b16 D[$r0 + #tpc_mmio_list_head] $r14
-	st b16 D[$r0 + #tpc_mmio_list_tail] $r15
-	call #mmctx_size
-	ld b32 $r14 D[$r0 + #tpc_count]
-	mulu $r14 $r15
-	add b32 $r2 $r14
-	add b32 $r3 $r14
-
-	// round up base/size to 256 byte boundary (for strand SWBASE)
-	add b32 $r4 0x1300
-	shr b32 $r3 2
-	iowr I[$r4 + 0x000] $r3		// MMCTX_LOAD_COUNT, wtf for?!?
-	shr b32 $r2 8
-	shr b32 $r3 6
-	add b32 $r2 1
-	add b32 $r3 1
-	shl b32 $r2 8
-	shl b32 $r3 8
-
-	// calculate size of strand context data
-	mov b32 $r15 $r2
-	call #strand_ctx_init
-	add b32 $r3 $r15
-
-	// save context size, and tell HUB we're done
-	mov $r1 0x800
-	shl b32 $r1 6
-	iowr I[$r1 + 0x100] $r3		// CC_SCRATCH[1]  = context size
-	add b32 $r1 0x800
-	clear b32 $r2
-	bset $r2 31
-	iowr I[$r1 + 0x000] $r2		// CC_SCRATCH[0] |= 0x80000000
-
-// Main program loop, very simple, sleeps until woken up by the interrupt
-// handler, pulls a command from the queue and executes its handler
-//
-main:
-	bset $flags $p0
-	sleep $p0
-	mov $r13 #cmd_queue
-	call #queue_get
-	bra $p1 #main
-
-	// 0x0000-0x0003 are all context transfers
-	cmpu b32 $r14 0x04
-	bra nc #main_not_ctx_xfer
-		// fetch $flags and mask off $p1/$p2
-		mov $r1 $flags
-		mov $r2 0x0006
-		not b32 $r2
-		and $r1 $r2
-		// set $p1/$p2 according to transfer type
-		shl b32 $r14 1
-		or $r1 $r14
-		mov $flags $r1
-		// transfer context data
-		call #ctx_xfer
-		bra #main
-
-	main_not_ctx_xfer:
-	shl b32 $r15 $r14 16
-	or $r15 E_BAD_COMMAND
-	call #error
-	bra #main
-
-// interrupt handler
-ih:
-	push $r8
-	mov $r8 $flags
-	push $r8
-	push $r9
-	push $r10
-	push $r11
-	push $r13
-	push $r14
-	push $r15
-
-	// incoming fifo command?
-	iord $r10 I[$r0 + 0x200]	// INTR
-	and $r11 $r10 0x00000004
-	bra e #ih_no_fifo
-		// queue incoming fifo command for later processing
-		mov $r11 0x1900
-		mov $r13 #cmd_queue
-		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
-		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
-		call #queue_put
-		add b32 $r11 0x400
-		mov $r14 1
-		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
-
-	// ack, and wake up main()
-	ih_no_fifo:
-	iowr I[$r0 + 0x100] $r10	// INTR_ACK
-
-	pop $r15
-	pop $r14
-	pop $r13
-	pop $r11
-	pop $r10
-	pop $r9
-	pop $r8
-	mov $flags $r8
-	pop $r8
-	bclr $flags $p0
-	iret
-
-// Set this GPC's bit in HUB_BAR, used to signal completion of various
-// activities to the HUB fuc
-//
-hub_barrier_done:
-	mov $r15 1
-	ld b32 $r14 D[$r0 + #gpc_id]
-	shl b32 $r15 $r14
-	mov $r14 -0x6be8 	// 0x409418 - HUB_BAR_SET
-	sethi $r14 0x400000
-	call #nv_wr32
-	ret
-
-// Disables various things, waits a bit, and re-enables them..
-//
-// Not sure how exactly this helps, perhaps "ENABLE" is not such a
-// good description for the bits we turn off?  Anyways, without this,
-// funny things happen.
-//
-ctx_redswitch:
-	mov $r14 0x614
-	shl b32 $r14 6
-	mov $r15 0x020
-	iowr I[$r14] $r15	// GPC_RED_SWITCH = POWER
-	mov $r15 8
-	ctx_redswitch_delay:
-		sub b32 $r15 1
-		bra ne #ctx_redswitch_delay
-	mov $r15 0xa20
-	iowr I[$r14] $r15	// GPC_RED_SWITCH = UNK11, ENABLE, POWER
-	ret
-
-// Transfer GPC context data between GPU and storage area
-//
-// In: $r15 context base address
-//     $p1 clear on save, set on load
-//     $p2 set if opposite direction done/will be done, so:
-//		on save it means: "a load will follow this save"
-//		on load it means: "a save preceeded this load"
-//
-ctx_xfer:
-	// set context base address
-	mov $r1 0xa04
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r15// MEM_BASE
-	bra not $p1 #ctx_xfer_not_load
-		call #ctx_redswitch
-	ctx_xfer_not_load:
-
-	// strands
-	mov $r1 0x4afc
-	sethi $r1 0x20000
-	mov $r2 0xc
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
-	call #strand_wait
-	mov $r2 0x47fc
-	sethi $r2 0x20000
-	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
-	xbit $r2 $flags $p1
-	add b32 $r2 3
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
-
-	// mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 2		// first
-	mov $r11 0x0000
-	sethi $r11 0x500000
-	ld b32 $r12 D[$r0 + #gpc_id]
-	shl b32 $r12 15
-	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn
-	ld b32 $r12 D[$r0 + #gpc_mmio_list_head]
-	ld b32 $r13 D[$r0 + #gpc_mmio_list_tail]
-	mov $r14 0		// not multi
-	call #mmctx_xfer
-
-	// per-TPC mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 4		// last
-	mov $r11 0x4000
-	sethi $r11 0x500000	// base = NV_PGRAPH_GPC0_TPC0
-	ld b32 $r12 D[$r0 + #gpc_id]
-	shl b32 $r12 15
-	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn_TPC0
-	ld b32 $r12 D[$r0 + #tpc_mmio_list_head]
-	ld b32 $r13 D[$r0 + #tpc_mmio_list_tail]
-	ld b32 $r15 D[$r0 + #tpc_mask]
-	mov $r14 0x800		// stride = 0x800
-	call #mmctx_xfer
-
-	// wait for strands to finish
-	call #strand_wait
-
-	// if load, or a save without a load following, do some
-	// unknown stuff that's done after finishing a block of
-	// strand commands
-	bra $p1 #ctx_xfer_post
-	bra not $p2 #ctx_xfer_done
-	ctx_xfer_post:
-		mov $r1 0x4afc
-		sethi $r1 0x20000
-		mov $r2 0xd
-		iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0d
-		call #strand_wait
-
-	// mark completion in HUB's barrier
-	ctx_xfer_done:
-	call #hub_barrier_done
-	ret
-
+#include "com.fuc"
+#include "gpc.fuc"
 .align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc.h
index 96050ddb22ca..f2b0dea80116 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvc0.fuc.h
@@ -1,17 +1,19 @@
 uint32_t nvc0_grgpc_data[] = {
-/* 0x0000: gpc_id */
-	0x00000000,
-/* 0x0004: gpc_mmio_list_head */
-	0x00000000,
-/* 0x0008: gpc_mmio_list_tail */
-	0x00000000,
-/* 0x000c: tpc_count */
-	0x00000000,
-/* 0x0010: tpc_mask */
+/* 0x0000: gpc_mmio_list_head */
+	0x00000064,
+/* 0x0004: gpc_mmio_list_tail */
+/* 0x0004: tpc_mmio_list_head */
+	0x00000064,
+/* 0x0008: tpc_mmio_list_tail */
+/* 0x0008: unk_mmio_list_head */
+	0x00000064,
+/* 0x000c: unk_mmio_list_tail */
+	0x00000064,
+/* 0x0010: gpc_id */
 	0x00000000,
-/* 0x0014: tpc_mmio_list_head */
+/* 0x0014: tpc_count */
 	0x00000000,
-/* 0x0018: tpc_mmio_list_tail */
+/* 0x0018: tpc_mask */
 	0x00000000,
 /* 0x001c: cmd_queue */
 	0x00000000,
@@ -32,153 +34,17 @@ uint32_t nvc0_grgpc_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0064: chipsets */
-	0x000000c0,
-	0x012800c8,
-	0x01e40194,
-	0x000000c1,
-	0x012c00c8,
-	0x01f80194,
-	0x000000c3,
-	0x012800c8,
-	0x01f40194,
-	0x000000c4,
-	0x012800c8,
-	0x01f40194,
-	0x000000c8,
-	0x012800c8,
-	0x01e40194,
-	0x000000ce,
-	0x012800c8,
-	0x01f40194,
-	0x000000cf,
-	0x012800c8,
-	0x01f00194,
-	0x000000d9,
-	0x0194012c,
-	0x025401f8,
-	0x00000000,
-/* 0x00c8: nvc0_gpc_mmio_head */
-	0x00000380,
-	0x14000400,
-	0x20000450,
-	0x00000600,
-	0x00000684,
-	0x10000700,
-	0x00000800,
-	0x08000808,
-	0x00000828,
-	0x00000830,
-	0x000008d8,
-	0x000008e0,
-	0x140008e8,
-	0x0000091c,
-	0x08000924,
-	0x00000b00,
-	0x14000b08,
-	0x00000bb8,
-	0x00000c08,
-	0x1c000c10,
-	0x00000c80,
-	0x00000c8c,
-	0x08001000,
-	0x00001014,
-/* 0x0128: nvc0_gpc_mmio_tail */
-	0x00000c6c,
-/* 0x012c: nvc1_gpc_mmio_tail */
-/* 0x012c: nvd9_gpc_mmio_head */
-	0x00000380,
-	0x04000400,
-	0x0800040c,
-	0x20000450,
-	0x00000600,
-	0x00000684,
-	0x10000700,
-	0x00000800,
-	0x08000808,
-	0x00000828,
-	0x00000830,
-	0x000008d8,
-	0x000008e0,
-	0x140008e8,
-	0x0000091c,
-	0x08000924,
-	0x00000b00,
-	0x14000b08,
-	0x00000bb8,
-	0x00000c08,
-	0x1c000c10,
-	0x00000c6c,
-	0x00000c80,
-	0x00000c8c,
-	0x08001000,
-	0x00001014,
-/* 0x0194: nvd9_gpc_mmio_tail */
-/* 0x0194: nvc0_tpc_mmio_head */
-	0x00000018,
-	0x0000003c,
-	0x00000048,
-	0x00000064,
-	0x00000088,
-	0x14000200,
-	0x0400021c,
-	0x14000300,
-	0x000003d0,
-	0x040003e0,
-	0x08000400,
-	0x00000420,
-	0x000004b0,
-	0x000004e8,
-	0x000004f4,
-	0x04000520,
-	0x0c000604,
-	0x4c000644,
-	0x00000698,
-	0x04000750,
-/* 0x01e4: nvc0_tpc_mmio_tail */
-	0x00000758,
-	0x000002c4,
-	0x000006e0,
-/* 0x01f0: nvcf_tpc_mmio_tail */
-	0x000004bc,
-/* 0x01f4: nvc3_tpc_mmio_tail */
-	0x00000544,
-/* 0x01f8: nvc1_tpc_mmio_tail */
-/* 0x01f8: nvd9_tpc_mmio_head */
-	0x00000018,
-	0x0000003c,
-	0x00000048,
-	0x00000064,
-	0x00000088,
-	0x14000200,
-	0x0400021c,
-	0x000002c4,
-	0x14000300,
-	0x000003d0,
-	0x040003e0,
-	0x08000400,
-	0x08000420,
-	0x000004b0,
-	0x000004e8,
-	0x000004f4,
-	0x04000520,
-	0x00000544,
-	0x0c000604,
-	0x4c000644,
-	0x00000698,
-	0x000006e0,
-	0x08000750,
 };
 
 uint32_t nvc0_grgpc_code[] = {
-	0x03060ef5,
+	0x03180ef5,
 /* 0x0004: queue_put */
 	0x9800d898,
 	0x86f001d9,
 	0x0489b808,
 	0xf00c1bf4,
 	0x21f502f7,
-	0x00f802ec,
+	0x00f802fe,
 /* 0x001c: queue_put_next */
 	0xb60798c4,
 	0x8dbb0384,
@@ -210,7 +76,7 @@ uint32_t nvc0_grgpc_code[] = {
 	0xc800bccf,
 	0x1bf41fcc,
 	0x06a7f0fa,
-	0x010321f5,
+	0x010921f5,
 	0xf840bfcf,
 /* 0x008d: nv_wr32 */
 	0x28b7f100,
@@ -232,63 +98,66 @@ uint32_t nvc0_grgpc_code[] = {
 	0x0684b604,
 	0xf80080d0,
 /* 0x00c9: wait_donez */
-	0x3c87f100,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d000,
-	0x081887f1,
-	0xd00684b6,
-/* 0x00e2: wait_done_wait_donez */
-	0x87f1008a,
-	0x84b60400,
-	0x0088cf06,
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
 	0xf4888aff,
-	0x87f1f31b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00099,
-/* 0x0103: wait_doneo */
-	0xf100f800,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00099f0,
-	0x87f10089,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x87f104bd,
 	0x84b60818,
 	0x008ad006,
-/* 0x011c: wait_done_wait_doneo */
+/* 0x0124: wait_doneo_e */
 	0x040087f1,
 	0xcf0684b6,
 	0x8aff0088,
 	0xf30bf488,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0099f094,
-	0xf80089d0,
-/* 0x013d: mmctx_size */
-/* 0x013f: nv_mmctx_size_loop */
-	0x9894bd00,
-	0x85b600e8,
-	0x0180b61a,
-	0xbb0284b6,
-	0xe0b60098,
-	0x04efb804,
-	0xb9eb1bf4,
-	0x00f8029f,
-/* 0x015c: mmctx_xfer */
-	0x083c87f1,
-	0xbd0684b6,
-	0x0199f094,
-	0xf10089d0,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf104bd00,
 	0xb6071087,
 	0x94bd0684,
 	0xf405bbfd,
 	0x8bd0090b,
 	0x0099f000,
-/* 0x0180: mmctx_base_disabled */
+/* 0x018c: mmctx_base_disabled */
 	0xf405eefd,
 	0x8ed00c0b,
 	0xc08fd080,
-/* 0x018f: mmctx_multi_disabled */
+/* 0x019b: mmctx_multi_disabled */
 	0xb70199f0,
 	0xc8010080,
 	0xb4b600ab,
@@ -296,8 +165,8 @@ uint32_t nvc0_grgpc_code[] = {
 	0xb601aec8,
 	0xbefd11e4,
 	0x008bd005,
-/* 0x01a8: mmctx_exec_loop */
-/* 0x01a8: mmctx_wait_free */
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
 	0xf0008ecf,
 	0x0bf41fe4,
 	0x00ce98fa,
@@ -306,76 +175,77 @@ uint32_t nvc0_grgpc_code[] = {
 	0x04cdb804,
 	0xc8e81bf4,
 	0x1bf402ab,
-/* 0x01c9: mmctx_fini_wait */
+/* 0x01d5: mmctx_fini_wait */
 	0x008bcf18,
 	0xb01fb4f0,
 	0x1bf410b4,
 	0x02a7f0f7,
 	0xf4c921f4,
-/* 0x01de: mmctx_stop */
+/* 0x01ea: mmctx_stop */
 	0xabc81b0e,
 	0x10b4b600,
 	0xf00cb9f0,
 	0x8bd012b9,
-/* 0x01ed: mmctx_stop_wait */
+/* 0x01f9: mmctx_stop_wait */
 	0x008bcf00,
 	0xf412bbc8,
-/* 0x01f6: mmctx_done */
-	0x87f1fa1b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00199,
-/* 0x0207: strand_wait */
-	0xf900f800,
-	0x02a7f0a0,
-	0xfcc921f4,
-/* 0x0213: strand_pre */
-	0xf100f8a0,
-	0xf04afc87,
-	0x97f00283,
-	0x0089d00c,
-	0x020721f5,
-/* 0x0226: strand_post */
-	0x87f100f8,
-	0x83f04afc,
-	0x0d97f002,
-	0xf50089d0,
-	0xf8020721,
-/* 0x0239: strand_set */
-	0xfca7f100,
-	0x02a3f04f,
-	0x0500aba2,
-	0xd00fc7f0,
-	0xc7f000ac,
-	0x00bcd00b,
-	0x020721f5,
-	0xf000aed0,
-	0xbcd00ac7,
-	0x0721f500,
-/* 0x0263: strand_ctx_init */
-	0xf100f802,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00399f0,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
 	0x21f50089,
-	0xe7f00213,
-	0x3921f503,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f00f00,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
 	0xfca7f102,
 	0x02a3f046,
 	0x0400aba0,
 	0xf040a0d0,
 	0xbcd001c7,
-	0x0721f500,
+	0x1521f500,
 	0x010c9202,
 	0xf000acd0,
 	0xbcd002c7,
-	0x0721f500,
-	0x2621f502,
+	0x1521f500,
+	0x3421f502,
 	0x8087f102,
 	0x0684b608,
 	0xb70089cf,
 	0x95220080,
-/* 0x02ba: ctx_init_strand_loop */
+/* 0x02ca: ctx_init_strand_loop */
 	0x8ed008fe,
 	0x408ed000,
 	0xb6808acf,
@@ -384,86 +254,74 @@ uint32_t nvc0_grgpc_code[] = {
 	0xb60480b6,
 	0x1bf40192,
 	0x08e4b6e8,
-	0xf1f2efbc,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00399f0,
-	0x00f80089,
-/* 0x02ec: error */
-	0xe7f1e0f9,
-	0xe3f09814,
-	0x8d21f440,
-	0x041ce0b7,
-	0xf401f7f0,
-	0xe0fc8d21,
-/* 0x0306: init */
-	0x04bd00f8,
-	0xf10004fe,
-	0xf0120017,
-	0x12d00227,
-	0x3e17f100,
-	0x0010fe04,
-	0x040017f1,
-	0xf0c010d0,
-	0x12d00427,
-	0x1031f400,
-	0x060817f1,
-	0xcf0614b6,
-	0x37f00012,
-	0x1f24f001,
-	0xb60432bb,
-	0x02800132,
-	0x04038003,
-	0x040010b7,
-	0x800012cf,
-	0x27f10002,
-	0x24b60800,
-	0x0022cf06,
-/* 0x035f: init_find_chipset */
-	0xb65817f0,
-	0x13980c10,
-	0x0432b800,
-	0xb00b0bf4,
-	0x1bf40034,
-/* 0x0373: init_context */
-	0xf100f8f1,
-	0xb6080027,
-	0x22cf0624,
-	0xf134bd40,
-	0xb6070047,
-	0x25950644,
-	0x0045d008,
-	0xbd4045d0,
-	0x58f4bde4,
-	0x1f58021e,
-	0x020e4003,
-	0xf5040f40,
-	0xbb013d21,
-	0x3fbb002f,
-	0x041e5800,
-	0x40051f58,
-	0x0f400a0e,
-	0x3d21f50c,
-	0x030e9801,
-	0xbb00effd,
-	0x3ebb002e,
-	0x0040b700,
-	0x0235b613,
-	0xb60043d0,
-	0x35b60825,
-	0x0120b606,
-	0xb60130b6,
-	0x34b60824,
-	0x022fb908,
-	0x026321f5,
-	0xf1003fbb,
-	0xb6080017,
-	0x13d00614,
-	0x0010b740,
-	0xf024bd08,
-	0x12d01f29,
-/* 0x0401: main */
-	0x0031f400,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0xe0f900f8,
+	0x9814e7f1,
+	0xf440e3f0,
+	0xe0b78d21,
+	0xf7f0041c,
+	0x8d21f401,
+	0x00f8e0fc,
+/* 0x0318: init */
+	0x04fe04bd,
+	0x0017f100,
+	0x0227f012,
+	0xf10012d0,
+	0xfe042617,
+	0x17f10010,
+	0x10d00400,
+	0x0427f0c0,
+	0xf40012d0,
+	0x17f11031,
+	0x14b60608,
+	0x0012cf06,
+	0xf00137f0,
+	0x32bb1f24,
+	0x0132b604,
+	0x80050280,
+	0x10b70603,
+	0x12cf0400,
+	0x04028000,
+	0x010027f1,
+	0xcf0223f0,
+	0x34bd0022,
+	0x070047f1,
+	0x950644b6,
+	0x45d00825,
+	0x4045d000,
+	0x98000e98,
+	0x21f5010f,
+	0x2fbb0147,
+	0x003fbb00,
+	0x98010e98,
+	0x21f5020f,
+	0x0e980147,
+	0x00effd05,
+	0xbb002ebb,
+	0x40b7003e,
+	0x35b61300,
+	0x0043d002,
+	0xb60825b6,
+	0x20b60635,
+	0x0130b601,
+	0xb60824b6,
+	0x2fb90834,
+	0x7121f502,
+	0x003fbb02,
+	0x010007f1,
+	0xd00203f0,
+	0x04bd0003,
+	0x29f024bd,
+	0x0007f11f,
+	0x0203f008,
+	0xbd0002d0,
+/* 0x03e9: main */
+	0x0031f404,
 	0xf00028f4,
 	0x21f41cd7,
 	0xf401f439,
@@ -474,94 +332,100 @@ uint32_t nvc0_grgpc_code[] = {
 	0x01e4b604,
 	0xfe051efd,
 	0x21f50018,
-	0x0ef404c3,
-/* 0x0431: main_not_ctx_xfer */
+	0x0ef404ad,
+/* 0x0419: main_not_ctx_xfer */
 	0x10ef94d3,
 	0xf501f5f0,
-	0xf402ec21,
-/* 0x043e: ih */
+	0xf402fe21,
+/* 0x0426: ih */
 	0x80f9c60e,
 	0xf90188fe,
 	0xf990f980,
 	0xf9b0f9a0,
 	0xf9e0f9d0,
-	0x800acff0,
-	0xf404abc4,
-	0xb7f11d0b,
-	0xd7f01900,
-	0x40becf1c,
-	0xf400bfcf,
-	0xb0b70421,
-	0xe7f00400,
-	0x00bed001,
-/* 0x0474: ih_no_fifo */
-	0xfc400ad0,
-	0xfce0fcf0,
-	0xfcb0fcd0,
-	0xfc90fca0,
-	0x0088fe80,
-	0x32f480fc,
-/* 0x048f: hub_barrier_done */
-	0xf001f800,
-	0x0e9801f7,
-	0x04febb00,
-	0x9418e7f1,
-	0xf440e3f0,
-	0x00f88d21,
-/* 0x04a4: ctx_redswitch */
-	0x0614e7f1,
-	0xf006e4b6,
-	0xefd020f7,
-	0x08f7f000,
-/* 0x04b4: ctx_redswitch_delay */
-	0xf401f2b6,
-	0xf7f1fd1b,
-	0xefd00a20,
-/* 0x04c3: ctx_xfer */
-	0xf100f800,
-	0xb60a0417,
-	0x1fd00614,
-	0x0711f400,
-	0x04a421f5,
-/* 0x04d4: ctx_xfer_not_load */
-	0x4afc17f1,
-	0xf00213f0,
-	0x12d00c27,
-	0x0721f500,
-	0xfc27f102,
-	0x0223f047,
-	0xf00020d0,
-	0x20b6012c,
-	0x0012d003,
+	0xcf04bdf0,
+	0xabc4800a,
+	0x1d0bf404,
+	0x1900b7f1,
+	0xcf1cd7f0,
+	0xbfcf40be,
+	0x0421f400,
+	0x0400b0b7,
+	0xd001e7f0,
+/* 0x045e: ih_no_fifo */
+	0x0ad000be,
+	0xfcf0fc40,
+	0xfcd0fce0,
+	0xfca0fcb0,
+	0xfe80fc90,
+	0x80fc0088,
+	0xf80032f4,
+/* 0x0479: hub_barrier_done */
+	0x01f7f001,
+	0xbb040e98,
+	0xe7f104fe,
+	0xe3f09418,
+	0x8d21f440,
+/* 0x048e: ctx_redswitch */
+	0xe7f100f8,
+	0xe4b60614,
+	0x20f7f006,
+	0xf000efd0,
+/* 0x049e: ctx_redswitch_delay */
+	0xf2b608f7,
+	0xfd1bf401,
+	0x0a20f7f1,
+	0xf800efd0,
+/* 0x04ad: ctx_xfer */
+	0x0417f100,
+	0x0614b60a,
+	0xf4001fd0,
+	0x21f50711,
+/* 0x04be: ctx_xfer_not_load */
+	0x17f1048e,
+	0x13f04afc,
+	0x0c27f002,
+	0xf50012d0,
+	0xf1021521,
+	0xf047fc27,
+	0x20d00223,
+	0x012cf000,
+	0xd00320b6,
+	0xacf00012,
+	0x02a5f001,
+	0xf000b7f0,
+	0x0c9850b3,
+	0x0fc4b604,
+	0x9800bcbb,
+	0x0d98000c,
+	0x00e7f001,
+	0x016621f5,
 	0xf001acf0,
-	0xb7f002a5,
-	0x50b3f000,
-	0xb6000c98,
-	0xbcbb0fc4,
-	0x010c9800,
-	0xf0020d98,
-	0x21f500e7,
-	0xacf0015c,
-	0x04a5f001,
-	0x4000b7f1,
-	0x9850b3f0,
-	0xc4b6000c,
-	0x00bcbb0f,
-	0x98050c98,
-	0x0f98060d,
-	0x00e7f104,
-	0x5c21f508,
-	0x0721f501,
-	0x0601f402,
-/* 0x054b: ctx_xfer_post */
-	0xf11412f4,
-	0xf04afc17,
-	0x27f00213,
-	0x0012d00d,
-	0x020721f5,
-/* 0x055c: ctx_xfer_done */
-	0x048f21f5,
-	0x000000f8,
+	0xb7f104a5,
+	0xb3f04000,
+	0x040c9850,
+	0xbb0fc4b6,
+	0x0c9800bc,
+	0x020d9801,
+	0xf1060f98,
+	0xf50800e7,
+	0xf5016621,
+	0xf4021521,
+	0x12f40601,
+/* 0x0535: ctx_xfer_post */
+	0xfc17f114,
+	0x0213f04a,
+	0xd00d27f0,
+	0x21f50012,
+/* 0x0546: ctx_xfer_done */
+	0x21f50215,
+	0x00f80479,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvd7.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvd7.fuc
new file mode 100644
index 000000000000..c2f754edbd7d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvd7.fuc
@@ -0,0 +1,42 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#define NV_PGRAPH_GPCX_UNK__SIZE                                     0x00000001
+
+#define CHIPSET GF117
+#include "macros.fuc"
+
+.section #nvd7_grgpc_data
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "gpc.fuc"
+#undef INCLUDE_DATA
+
+.section #nvd7_grgpc_code
+#define INCLUDE_CODE
+bra #init
+#include "com.fuc"
+#include "gpc.fuc"
+.align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvd7.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvd7.fuc.h
new file mode 100644
index 000000000000..dd346c2a1624
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvd7.fuc.h
@@ -0,0 +1,475 @@
+uint32_t nvd7_grgpc_data[] = {
+/* 0x0000: gpc_mmio_list_head */
+	0x0000006c,
+/* 0x0004: gpc_mmio_list_tail */
+/* 0x0004: tpc_mmio_list_head */
+	0x0000006c,
+/* 0x0008: tpc_mmio_list_tail */
+/* 0x0008: unk_mmio_list_head */
+	0x0000006c,
+/* 0x000c: unk_mmio_list_tail */
+	0x0000006c,
+/* 0x0010: gpc_id */
+	0x00000000,
+/* 0x0014: tpc_count */
+	0x00000000,
+/* 0x0018: tpc_mask */
+	0x00000000,
+/* 0x001c: unk_count */
+	0x00000000,
+/* 0x0020: unk_mask */
+	0x00000000,
+/* 0x0024: cmd_queue */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+};
+
+uint32_t nvd7_grgpc_code[] = {
+	0x03180ef5,
+/* 0x0004: queue_put */
+	0x9800d898,
+	0x86f001d9,
+	0x0489b808,
+	0xf00c1bf4,
+	0x21f502f7,
+	0x00f802fe,
+/* 0x001c: queue_put_next */
+	0xb60798c4,
+	0x8dbb0384,
+	0x0880b600,
+	0x80008e80,
+	0x90b6018f,
+	0x0f94f001,
+	0xf801d980,
+/* 0x0039: queue_get */
+	0x0131f400,
+	0x9800d898,
+	0x89b801d9,
+	0x210bf404,
+	0xb60789c4,
+	0x9dbb0394,
+	0x0890b600,
+	0x98009e98,
+	0x80b6019f,
+	0x0f84f001,
+	0xf400d880,
+/* 0x0066: queue_get_done */
+	0x00f80132,
+/* 0x0068: nv_rd32 */
+	0x0728b7f1,
+	0xb906b4b6,
+	0xc9f002ec,
+	0x00bcd01f,
+/* 0x0078: nv_rd32_wait */
+	0xc800bccf,
+	0x1bf41fcc,
+	0x06a7f0fa,
+	0x010921f5,
+	0xf840bfcf,
+/* 0x008d: nv_wr32 */
+	0x28b7f100,
+	0x06b4b607,
+	0xb980bfd0,
+	0xc9f002ec,
+	0x1ec9f01f,
+/* 0x00a3: nv_wr32_wait */
+	0xcf00bcd0,
+	0xccc800bc,
+	0xfa1bf41f,
+/* 0x00ae: watchdog_reset */
+	0x87f100f8,
+	0x84b60430,
+	0x1ff9f006,
+	0xf8008fd0,
+/* 0x00bd: watchdog_clear */
+	0x3087f100,
+	0x0684b604,
+	0xf80080d0,
+/* 0x00c9: wait_donez */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
+	0xf4888aff,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x87f104bd,
+	0x84b60818,
+	0x008ad006,
+/* 0x0124: wait_doneo_e */
+	0x040087f1,
+	0xcf0684b6,
+	0x8aff0088,
+	0xf30bf488,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf104bd00,
+	0xb6071087,
+	0x94bd0684,
+	0xf405bbfd,
+	0x8bd0090b,
+	0x0099f000,
+/* 0x018c: mmctx_base_disabled */
+	0xf405eefd,
+	0x8ed00c0b,
+	0xc08fd080,
+/* 0x019b: mmctx_multi_disabled */
+	0xb70199f0,
+	0xc8010080,
+	0xb4b600ab,
+	0x0cb9f010,
+	0xb601aec8,
+	0xbefd11e4,
+	0x008bd005,
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
+	0xf0008ecf,
+	0x0bf41fe4,
+	0x00ce98fa,
+	0xd005e9fd,
+	0xc0b6c08e,
+	0x04cdb804,
+	0xc8e81bf4,
+	0x1bf402ab,
+/* 0x01d5: mmctx_fini_wait */
+	0x008bcf18,
+	0xb01fb4f0,
+	0x1bf410b4,
+	0x02a7f0f7,
+	0xf4c921f4,
+/* 0x01ea: mmctx_stop */
+	0xabc81b0e,
+	0x10b4b600,
+	0xf00cb9f0,
+	0x8bd012b9,
+/* 0x01f9: mmctx_stop_wait */
+	0x008bcf00,
+	0xf412bbc8,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
+	0x21f50089,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f00f00,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
+	0xfca7f102,
+	0x02a3f046,
+	0x0400aba0,
+	0xf040a0d0,
+	0xbcd001c7,
+	0x1521f500,
+	0x010c9202,
+	0xf000acd0,
+	0xbcd002c7,
+	0x1521f500,
+	0x3421f502,
+	0x8087f102,
+	0x0684b608,
+	0xb70089cf,
+	0x95220080,
+/* 0x02ca: ctx_init_strand_loop */
+	0x8ed008fe,
+	0x408ed000,
+	0xb6808acf,
+	0xa0b606a5,
+	0x00eabb01,
+	0xb60480b6,
+	0x1bf40192,
+	0x08e4b6e8,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0xe0f900f8,
+	0x9814e7f1,
+	0xf440e3f0,
+	0xe0b78d21,
+	0xf7f0041c,
+	0x8d21f401,
+	0x00f8e0fc,
+/* 0x0318: init */
+	0x04fe04bd,
+	0x0017f100,
+	0x0227f012,
+	0xf10012d0,
+	0xfe047017,
+	0x17f10010,
+	0x10d00400,
+	0x0427f0c0,
+	0xf40012d0,
+	0x17f11031,
+	0x14b60608,
+	0x0012cf06,
+	0xf00137f0,
+	0x32bb1f24,
+	0x0132b604,
+	0x80050280,
+	0x10b70603,
+	0x12cf0400,
+	0x04028000,
+	0x0c30e7f1,
+	0xbd50e3f0,
+	0xbd34bd24,
+/* 0x0371: init_unk_loop */
+	0x6821f444,
+	0xf400f6b0,
+	0xf7f00f0b,
+	0x04f2bb01,
+	0xb6054ffd,
+/* 0x0386: init_unk_next */
+	0x20b60130,
+	0x04e0b601,
+	0xf40126b0,
+/* 0x0392: init_unk_done */
+	0x0380e21b,
+	0x08048007,
+	0x010027f1,
+	0xcf0223f0,
+	0x34bd0022,
+	0x070047f1,
+	0x950644b6,
+	0x45d00825,
+	0x4045d000,
+	0x98000e98,
+	0x21f5010f,
+	0x2fbb0147,
+	0x003fbb00,
+	0x98010e98,
+	0x21f5020f,
+	0x0e980147,
+	0x00effd05,
+	0xbb002ebb,
+	0x0e98003e,
+	0x030f9802,
+	0x014721f5,
+	0xfd070e98,
+	0x2ebb00ef,
+	0x003ebb00,
+	0x130040b7,
+	0xd00235b6,
+	0x25b60043,
+	0x0635b608,
+	0xb60120b6,
+	0x24b60130,
+	0x0834b608,
+	0xf5022fb9,
+	0xbb027121,
+	0x07f1003f,
+	0x03f00100,
+	0x0003d002,
+	0x24bd04bd,
+	0xf11f29f0,
+	0xf0080007,
+	0x02d00203,
+/* 0x0433: main */
+	0xf404bd00,
+	0x28f40031,
+	0x24d7f000,
+	0xf43921f4,
+	0xe4b0f401,
+	0x1e18f404,
+	0xf00181fe,
+	0x20bd0627,
+	0xb60412fd,
+	0x1efd01e4,
+	0x0018fe05,
+	0x04f721f5,
+/* 0x0463: main_not_ctx_xfer */
+	0x94d30ef4,
+	0xf5f010ef,
+	0xfe21f501,
+	0xc60ef402,
+/* 0x0470: ih */
+	0x88fe80f9,
+	0xf980f901,
+	0xf9a0f990,
+	0xf9d0f9b0,
+	0xbdf0f9e0,
+	0x800acf04,
+	0xf404abc4,
+	0xb7f11d0b,
+	0xd7f01900,
+	0x40becf24,
+	0xf400bfcf,
+	0xb0b70421,
+	0xe7f00400,
+	0x00bed001,
+/* 0x04a8: ih_no_fifo */
+	0xfc400ad0,
+	0xfce0fcf0,
+	0xfcb0fcd0,
+	0xfc90fca0,
+	0x0088fe80,
+	0x32f480fc,
+/* 0x04c3: hub_barrier_done */
+	0xf001f800,
+	0x0e9801f7,
+	0x04febb04,
+	0x9418e7f1,
+	0xf440e3f0,
+	0x00f88d21,
+/* 0x04d8: ctx_redswitch */
+	0x0614e7f1,
+	0xf006e4b6,
+	0xefd020f7,
+	0x08f7f000,
+/* 0x04e8: ctx_redswitch_delay */
+	0xf401f2b6,
+	0xf7f1fd1b,
+	0xefd00a20,
+/* 0x04f7: ctx_xfer */
+	0xf100f800,
+	0xb60a0417,
+	0x1fd00614,
+	0x0711f400,
+	0x04d821f5,
+/* 0x0508: ctx_xfer_not_load */
+	0x4afc17f1,
+	0xf00213f0,
+	0x12d00c27,
+	0x1521f500,
+	0xfc27f102,
+	0x0223f047,
+	0xf00020d0,
+	0x20b6012c,
+	0x0012d003,
+	0xf001acf0,
+	0xb7f002a5,
+	0x50b3f000,
+	0xb6040c98,
+	0xbcbb0fc4,
+	0x000c9800,
+	0xf0010d98,
+	0x21f500e7,
+	0xacf00166,
+	0x00b7f101,
+	0x50b3f040,
+	0xb6040c98,
+	0xbcbb0fc4,
+	0x010c9800,
+	0x98020d98,
+	0xe7f1060f,
+	0x21f50800,
+	0xacf00166,
+	0x04a5f001,
+	0x3000b7f1,
+	0x9850b3f0,
+	0xc4b6040c,
+	0x00bcbb0f,
+	0x98020c98,
+	0x0f98030d,
+	0x00e7f108,
+	0x6621f502,
+	0x1521f501,
+	0x0601f402,
+/* 0x05a3: ctx_xfer_post */
+	0xf11412f4,
+	0xf04afc17,
+	0x27f00213,
+	0x0012d00d,
+	0x021521f5,
+/* 0x05b4: ctx_xfer_done */
+	0x04c321f5,
+	0x000000f8,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc
index 62ab231cd6b6..6b906cd2a31f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc
@@ -1,6 +1,5 @@
-/* fuc microcode for nve0 PGRAPH/GPC
- *
- * Copyright 2011 Red Hat Inc.
+/*
+ * Copyright 2013 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -20,437 +19,24 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
- * Authors: Ben Skeggs
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
-/* To build:
- *    m4 nve0_grgpc.fuc | envyas -a -w -m fuc -V nva3 -o nve0_grgpc.fuc.h
- */
+#define NV_PGRAPH_GPCX_UNK__SIZE                                     0x00000001
 
-/* TODO
- * - bracket certain functions with scratch writes, useful for debugging
- * - watchdog timer around ctx operations
- */
+#define CHIPSET GK100
+#include "macros.fuc"
 
 .section #nve0_grgpc_data
-include(`nve0.fuc')
-gpc_id:			.b32 0
-gpc_mmio_list_head:	.b32 0
-gpc_mmio_list_tail:	.b32 0
-
-tpc_count:		.b32 0
-tpc_mask:		.b32 0
-tpc_mmio_list_head:	.b32 0
-tpc_mmio_list_tail:	.b32 0
-
-cmd_queue:		queue_init
-
-// chipset descriptions
-chipsets:
-.b8  0xe4 0 0 0
-.b16 #nve4_gpc_mmio_head
-.b16 #nve4_gpc_mmio_tail
-.b16 #nve4_tpc_mmio_head
-.b16 #nve4_tpc_mmio_tail
-.b8  0xe7 0 0 0
-.b16 #nve4_gpc_mmio_head
-.b16 #nve4_gpc_mmio_tail
-.b16 #nve4_tpc_mmio_head
-.b16 #nve4_tpc_mmio_tail
-.b8  0xe6 0 0 0
-.b16 #nve4_gpc_mmio_head
-.b16 #nve4_gpc_mmio_tail
-.b16 #nve4_tpc_mmio_head
-.b16 #nve4_tpc_mmio_tail
-.b8  0 0 0 0
-
-// GPC mmio lists
-nve4_gpc_mmio_head:
-mmctx_data(0x000380, 1)
-mmctx_data(0x000400, 2)
-mmctx_data(0x00040c, 3)
-mmctx_data(0x000450, 9)
-mmctx_data(0x000600, 1)
-mmctx_data(0x000684, 1)
-mmctx_data(0x000700, 5)
-mmctx_data(0x000800, 1)
-mmctx_data(0x000808, 3)
-mmctx_data(0x000828, 1)
-mmctx_data(0x000830, 1)
-mmctx_data(0x0008d8, 1)
-mmctx_data(0x0008e0, 1)
-mmctx_data(0x0008e8, 6)
-mmctx_data(0x00091c, 1)
-mmctx_data(0x000924, 3)
-mmctx_data(0x000b00, 1)
-mmctx_data(0x000b08, 6)
-mmctx_data(0x000bb8, 1)
-mmctx_data(0x000c08, 1)
-mmctx_data(0x000c10, 8)
-mmctx_data(0x000c40, 1)
-mmctx_data(0x000c6c, 1)
-mmctx_data(0x000c80, 1)
-mmctx_data(0x000c8c, 1)
-mmctx_data(0x001000, 3)
-mmctx_data(0x001014, 1)
-mmctx_data(0x003024, 1)
-mmctx_data(0x0030c0, 2)
-mmctx_data(0x0030e4, 1)
-mmctx_data(0x003100, 6)
-mmctx_data(0x0031d0, 1)
-mmctx_data(0x0031e0, 2)
-nve4_gpc_mmio_tail:
-
-// TPC mmio lists
-nve4_tpc_mmio_head:
-mmctx_data(0x000048, 1)
-mmctx_data(0x000064, 1)
-mmctx_data(0x000088, 1)
-mmctx_data(0x000200, 6)
-mmctx_data(0x00021c, 2)
-mmctx_data(0x000230, 1)
-mmctx_data(0x0002c4, 1)
-mmctx_data(0x000400, 3)
-mmctx_data(0x000420, 3)
-mmctx_data(0x0004e8, 1)
-mmctx_data(0x0004f4, 1)
-mmctx_data(0x000604, 4)
-mmctx_data(0x000644, 22)
-mmctx_data(0x0006ac, 2)
-mmctx_data(0x0006c8, 1)
-mmctx_data(0x000730, 8)
-mmctx_data(0x000758, 1)
-mmctx_data(0x000778, 1)
-nve4_tpc_mmio_tail:
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "gpc.fuc"
+#undef INCLUDE_DATA
 
 .section #nve0_grgpc_code
+#define INCLUDE_CODE
 bra #init
-define(`include_code')
-include(`nve0.fuc')
-
-// reports an exception to the host
-//
-// In: $r15 error code (see nve0.fuc)
-//
-error:
-	push $r14
-	mov $r14 -0x67ec 	// 0x9814
-	sethi $r14 0x400000
-	call #nv_wr32		// HUB_CTXCTL_CC_SCRATCH[5] = error code
-	add b32 $r14 0x41c
-	mov $r15 1
-	call #nv_wr32		// HUB_CTXCTL_INTR_UP_SET
-	pop $r14
-	ret
-
-// GPC fuc initialisation, executed by triggering ucode start, will
-// fall through to main loop after completion.
-//
-// Input:
-//   CC_SCRATCH[0]: chipset (PMC_BOOT_0 read returns 0x0bad0bad... sigh)
-//   CC_SCRATCH[1]: context base
-//
-// Output:
-//   CC_SCRATCH[0]:
-//	     31:31: set to signal completion
-//   CC_SCRATCH[1]:
-//	      31:0: GPC context size
-//
-init:
-	clear b32 $r0
-	mov $sp $r0
-
-	// enable fifo access
-	mov $r1 0x1200
-	mov $r2 2
-	iowr I[$r1 + 0x000] $r2		// FIFO_ENABLE
-
-	// setup i0 handler, and route all interrupts to it
-	mov $r1 #ih
-	mov $iv0 $r1
-	mov $r1 0x400
-	iowr I[$r1 + 0x300] $r0		// INTR_DISPATCH
-
-	// enable fifo interrupt
-	mov $r2 4
-	iowr I[$r1 + 0x000] $r2		// INTR_EN_SET
-
-	// enable interrupts
-	bset $flags ie0
-
-	// figure out which GPC we are, and how many TPCs we have
-	mov $r1 0x608
-	shl b32 $r1 6
-	iord $r2 I[$r1 + 0x000]		// UNITS
-	mov $r3 1
-	and $r2 0x1f
-	shl b32 $r3 $r2
-	sub b32 $r3 1
-	st b32 D[$r0 + #tpc_count] $r2
-	st b32 D[$r0 + #tpc_mask] $r3
-	add b32 $r1 0x400
-	iord $r2 I[$r1 + 0x000]		// MYINDEX
-	st b32 D[$r0 + #gpc_id] $r2
-
-	// find context data for this chipset
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x000]		// CC_SCRATCH[0]
-	mov $r1 #chipsets - 12
-	init_find_chipset:
-		add b32 $r1 12
-		ld b32 $r3 D[$r1 + 0x00]
-		cmpu b32 $r3 $r2
-		bra e #init_context
-		cmpu b32 $r3 0
-		bra ne #init_find_chipset
-		// unknown chipset
-		ret
-
-	// initialise context base, and size tracking
-	init_context:
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x100]	// CC_SCRATCH[1], initial base
-	clear b32 $r3		// track GPC context size here
-
-	// set mmctx base addresses now so we don't have to do it later,
-	// they don't currently ever change
-	mov $r4 0x700
-	shl b32 $r4 6
-	shr b32 $r5 $r2 8
-	iowr I[$r4 + 0x000] $r5		// MMCTX_SAVE_SWBASE
-	iowr I[$r4 + 0x100] $r5		// MMCTX_LOAD_SWBASE
-
-	// calculate GPC mmio context size, store the chipset-specific
-	// mmio list pointers somewhere we can get at them later without
-	// re-parsing the chipset list
-	clear b32 $r14
-	clear b32 $r15
-	ld b16 $r14 D[$r1 + 4]
-	ld b16 $r15 D[$r1 + 6]
-	st b16 D[$r0 + #gpc_mmio_list_head] $r14
-	st b16 D[$r0 + #gpc_mmio_list_tail] $r15
-	call #mmctx_size
-	add b32 $r2 $r15
-	add b32 $r3 $r15
-
-	// calculate per-TPC mmio context size, store the list pointers
-	ld b16 $r14 D[$r1 + 8]
-	ld b16 $r15 D[$r1 + 10]
-	st b16 D[$r0 + #tpc_mmio_list_head] $r14
-	st b16 D[$r0 + #tpc_mmio_list_tail] $r15
-	call #mmctx_size
-	ld b32 $r14 D[$r0 + #tpc_count]
-	mulu $r14 $r15
-	add b32 $r2 $r14
-	add b32 $r3 $r14
-
-	// round up base/size to 256 byte boundary (for strand SWBASE)
-	add b32 $r4 0x1300
-	shr b32 $r3 2
-	iowr I[$r4 + 0x000] $r3		// MMCTX_LOAD_COUNT, wtf for?!?
-	shr b32 $r2 8
-	shr b32 $r3 6
-	add b32 $r2 1
-	add b32 $r3 1
-	shl b32 $r2 8
-	shl b32 $r3 8
-
-	// calculate size of strand context data
-	mov b32 $r15 $r2
-	call #strand_ctx_init
-	add b32 $r3 $r15
-
-	// save context size, and tell HUB we're done
-	mov $r1 0x800
-	shl b32 $r1 6
-	iowr I[$r1 + 0x100] $r3		// CC_SCRATCH[1]  = context size
-	add b32 $r1 0x800
-	clear b32 $r2
-	bset $r2 31
-	iowr I[$r1 + 0x000] $r2		// CC_SCRATCH[0] |= 0x80000000
-
-// Main program loop, very simple, sleeps until woken up by the interrupt
-// handler, pulls a command from the queue and executes its handler
-//
-main:
-	bset $flags $p0
-	sleep $p0
-	mov $r13 #cmd_queue
-	call #queue_get
-	bra $p1 #main
-
-	// 0x0000-0x0003 are all context transfers
-	cmpu b32 $r14 0x04
-	bra nc #main_not_ctx_xfer
-		// fetch $flags and mask off $p1/$p2
-		mov $r1 $flags
-		mov $r2 0x0006
-		not b32 $r2
-		and $r1 $r2
-		// set $p1/$p2 according to transfer type
-		shl b32 $r14 1
-		or $r1 $r14
-		mov $flags $r1
-		// transfer context data
-		call #ctx_xfer
-		bra #main
-
-	main_not_ctx_xfer:
-	shl b32 $r15 $r14 16
-	or $r15 E_BAD_COMMAND
-	call #error
-	bra #main
-
-// interrupt handler
-ih:
-	push $r8
-	mov $r8 $flags
-	push $r8
-	push $r9
-	push $r10
-	push $r11
-	push $r13
-	push $r14
-	push $r15
-
-	// incoming fifo command?
-	iord $r10 I[$r0 + 0x200]	// INTR
-	and $r11 $r10 0x00000004
-	bra e #ih_no_fifo
-		// queue incoming fifo command for later processing
-		mov $r11 0x1900
-		mov $r13 #cmd_queue
-		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
-		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
-		call #queue_put
-		add b32 $r11 0x400
-		mov $r14 1
-		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
-
-	// ack, and wake up main()
-	ih_no_fifo:
-	iowr I[$r0 + 0x100] $r10	// INTR_ACK
-
-	pop $r15
-	pop $r14
-	pop $r13
-	pop $r11
-	pop $r10
-	pop $r9
-	pop $r8
-	mov $flags $r8
-	pop $r8
-	bclr $flags $p0
-	iret
-
-// Set this GPC's bit in HUB_BAR, used to signal completion of various
-// activities to the HUB fuc
-//
-hub_barrier_done:
-	mov $r15 1
-	ld b32 $r14 D[$r0 + #gpc_id]
-	shl b32 $r15 $r14
-	mov $r14 -0x6be8 	// 0x409418 - HUB_BAR_SET
-	sethi $r14 0x400000
-	call #nv_wr32
-	ret
-
-// Disables various things, waits a bit, and re-enables them..
-//
-// Not sure how exactly this helps, perhaps "ENABLE" is not such a
-// good description for the bits we turn off?  Anyways, without this,
-// funny things happen.
-//
-ctx_redswitch:
-	mov $r14 0x614
-	shl b32 $r14 6
-	mov $r15 0x020
-	iowr I[$r14] $r15	// GPC_RED_SWITCH = POWER
-	mov $r15 8
-	ctx_redswitch_delay:
-		sub b32 $r15 1
-		bra ne #ctx_redswitch_delay
-	mov $r15 0xa20
-	iowr I[$r14] $r15	// GPC_RED_SWITCH = UNK11, ENABLE, POWER
-	ret
-
-// Transfer GPC context data between GPU and storage area
-//
-// In: $r15 context base address
-//     $p1 clear on save, set on load
-//     $p2 set if opposite direction done/will be done, so:
-//		on save it means: "a load will follow this save"
-//		on load it means: "a save preceeded this load"
-//
-ctx_xfer:
-	// set context base address
-	mov $r1 0xa04
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r15// MEM_BASE
-	bra not $p1 #ctx_xfer_not_load
-		call #ctx_redswitch
-	ctx_xfer_not_load:
-
-	// strands
-	mov $r1 0x4afc
-	sethi $r1 0x20000
-	mov $r2 0xc
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
-	call #strand_wait
-	mov $r2 0x47fc
-	sethi $r2 0x20000
-	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
-	xbit $r2 $flags $p1
-	add b32 $r2 3
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
-
-	// mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 2		// first
-	mov $r11 0x0000
-	sethi $r11 0x500000
-	ld b32 $r12 D[$r0 + #gpc_id]
-	shl b32 $r12 15
-	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn
-	ld b32 $r12 D[$r0 + #gpc_mmio_list_head]
-	ld b32 $r13 D[$r0 + #gpc_mmio_list_tail]
-	mov $r14 0		// not multi
-	call #mmctx_xfer
-
-	// per-TPC mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 4		// last
-	mov $r11 0x4000
-	sethi $r11 0x500000	// base = NV_PGRAPH_GPC0_TPC0
-	ld b32 $r12 D[$r0 + #gpc_id]
-	shl b32 $r12 15
-	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn_TPC0
-	ld b32 $r12 D[$r0 + #tpc_mmio_list_head]
-	ld b32 $r13 D[$r0 + #tpc_mmio_list_tail]
-	ld b32 $r15 D[$r0 + #tpc_mask]
-	mov $r14 0x800		// stride = 0x800
-	call #mmctx_xfer
-
-	// wait for strands to finish
-	call #strand_wait
-
-	// if load, or a save without a load following, do some
-	// unknown stuff that's done after finishing a block of
-	// strand commands
-	bra $p1 #ctx_xfer_post
-	bra not $p2 #ctx_xfer_done
-	ctx_xfer_post:
-		mov $r1 0x4afc
-		sethi $r1 0x20000
-		mov $r2 0xd
-		iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0d
-		call #strand_wait
-
-	// mark completion in HUB's barrier
-	ctx_xfer_done:
-	call #hub_barrier_done
-	ret
-
+#include "com.fuc"
+#include "gpc.fuc"
 .align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc.h
index 09ee4702c8b2..7ff5ef6b0804 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnve0.fuc.h
@@ -1,19 +1,27 @@
 uint32_t nve0_grgpc_data[] = {
-/* 0x0000: gpc_id */
+/* 0x0000: gpc_mmio_list_head */
+	0x0000006c,
+/* 0x0004: gpc_mmio_list_tail */
+/* 0x0004: tpc_mmio_list_head */
+	0x0000006c,
+/* 0x0008: tpc_mmio_list_tail */
+/* 0x0008: unk_mmio_list_head */
+	0x0000006c,
+/* 0x000c: unk_mmio_list_tail */
+	0x0000006c,
+/* 0x0010: gpc_id */
 	0x00000000,
-/* 0x0004: gpc_mmio_list_head */
+/* 0x0014: tpc_count */
 	0x00000000,
-/* 0x0008: gpc_mmio_list_tail */
+/* 0x0018: tpc_mask */
 	0x00000000,
-/* 0x000c: tpc_count */
+/* 0x001c: unk_count */
 	0x00000000,
-/* 0x0010: tpc_mask */
+/* 0x0020: unk_mask */
 	0x00000000,
-/* 0x0014: tpc_mmio_list_head */
+/* 0x0024: cmd_queue */
 	0x00000000,
-/* 0x0018: tpc_mmio_list_tail */
 	0x00000000,
-/* 0x001c: cmd_queue */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -30,84 +38,17 @@ uint32_t nve0_grgpc_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x00000000,
-	0x00000000,
-/* 0x0064: chipsets */
-	0x000000e4,
-	0x0110008c,
-	0x01580110,
-	0x000000e7,
-	0x0110008c,
-	0x01580110,
-	0x000000e6,
-	0x0110008c,
-	0x01580110,
-	0x00000000,
-/* 0x008c: nve4_gpc_mmio_head */
-	0x00000380,
-	0x04000400,
-	0x0800040c,
-	0x20000450,
-	0x00000600,
-	0x00000684,
-	0x10000700,
-	0x00000800,
-	0x08000808,
-	0x00000828,
-	0x00000830,
-	0x000008d8,
-	0x000008e0,
-	0x140008e8,
-	0x0000091c,
-	0x08000924,
-	0x00000b00,
-	0x14000b08,
-	0x00000bb8,
-	0x00000c08,
-	0x1c000c10,
-	0x00000c40,
-	0x00000c6c,
-	0x00000c80,
-	0x00000c8c,
-	0x08001000,
-	0x00001014,
-	0x00003024,
-	0x040030c0,
-	0x000030e4,
-	0x14003100,
-	0x000031d0,
-	0x040031e0,
-/* 0x0110: nve4_gpc_mmio_tail */
-/* 0x0110: nve4_tpc_mmio_head */
-	0x00000048,
-	0x00000064,
-	0x00000088,
-	0x14000200,
-	0x0400021c,
-	0x00000230,
-	0x000002c4,
-	0x08000400,
-	0x08000420,
-	0x000004e8,
-	0x000004f4,
-	0x0c000604,
-	0x54000644,
-	0x040006ac,
-	0x000006c8,
-	0x1c000730,
-	0x00000758,
-	0x00000778,
 };
 
 uint32_t nve0_grgpc_code[] = {
-	0x03060ef5,
+	0x03180ef5,
 /* 0x0004: queue_put */
 	0x9800d898,
 	0x86f001d9,
 	0x0489b808,
 	0xf00c1bf4,
 	0x21f502f7,
-	0x00f802ec,
+	0x00f802fe,
 /* 0x001c: queue_put_next */
 	0xb60798c4,
 	0x8dbb0384,
@@ -139,7 +80,7 @@ uint32_t nve0_grgpc_code[] = {
 	0xc800bccf,
 	0x1bf41fcc,
 	0x06a7f0fa,
-	0x010321f5,
+	0x010921f5,
 	0xf840bfcf,
 /* 0x008d: nv_wr32 */
 	0x28b7f100,
@@ -161,63 +102,66 @@ uint32_t nve0_grgpc_code[] = {
 	0x0684b604,
 	0xf80080d0,
 /* 0x00c9: wait_donez */
-	0x3c87f100,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d000,
-	0x081887f1,
-	0xd00684b6,
-/* 0x00e2: wait_done_wait_donez */
-	0x87f1008a,
-	0x84b60400,
-	0x0088cf06,
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
 	0xf4888aff,
-	0x87f1f31b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00099,
-/* 0x0103: wait_doneo */
-	0xf100f800,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00099f0,
-	0x87f10089,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x87f104bd,
 	0x84b60818,
 	0x008ad006,
-/* 0x011c: wait_done_wait_doneo */
+/* 0x0124: wait_doneo_e */
 	0x040087f1,
 	0xcf0684b6,
 	0x8aff0088,
 	0xf30bf488,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0099f094,
-	0xf80089d0,
-/* 0x013d: mmctx_size */
-/* 0x013f: nv_mmctx_size_loop */
-	0x9894bd00,
-	0x85b600e8,
-	0x0180b61a,
-	0xbb0284b6,
-	0xe0b60098,
-	0x04efb804,
-	0xb9eb1bf4,
-	0x00f8029f,
-/* 0x015c: mmctx_xfer */
-	0x083c87f1,
-	0xbd0684b6,
-	0x0199f094,
-	0xf10089d0,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf104bd00,
 	0xb6071087,
 	0x94bd0684,
 	0xf405bbfd,
 	0x8bd0090b,
 	0x0099f000,
-/* 0x0180: mmctx_base_disabled */
+/* 0x018c: mmctx_base_disabled */
 	0xf405eefd,
 	0x8ed00c0b,
 	0xc08fd080,
-/* 0x018f: mmctx_multi_disabled */
+/* 0x019b: mmctx_multi_disabled */
 	0xb70199f0,
 	0xc8010080,
 	0xb4b600ab,
@@ -225,8 +169,8 @@ uint32_t nve0_grgpc_code[] = {
 	0xb601aec8,
 	0xbefd11e4,
 	0x008bd005,
-/* 0x01a8: mmctx_exec_loop */
-/* 0x01a8: mmctx_wait_free */
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
 	0xf0008ecf,
 	0x0bf41fe4,
 	0x00ce98fa,
@@ -235,76 +179,77 @@ uint32_t nve0_grgpc_code[] = {
 	0x04cdb804,
 	0xc8e81bf4,
 	0x1bf402ab,
-/* 0x01c9: mmctx_fini_wait */
+/* 0x01d5: mmctx_fini_wait */
 	0x008bcf18,
 	0xb01fb4f0,
 	0x1bf410b4,
 	0x02a7f0f7,
 	0xf4c921f4,
-/* 0x01de: mmctx_stop */
+/* 0x01ea: mmctx_stop */
 	0xabc81b0e,
 	0x10b4b600,
 	0xf00cb9f0,
 	0x8bd012b9,
-/* 0x01ed: mmctx_stop_wait */
+/* 0x01f9: mmctx_stop_wait */
 	0x008bcf00,
 	0xf412bbc8,
-/* 0x01f6: mmctx_done */
-	0x87f1fa1b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00199,
-/* 0x0207: strand_wait */
-	0xf900f800,
-	0x02a7f0a0,
-	0xfcc921f4,
-/* 0x0213: strand_pre */
-	0xf100f8a0,
-	0xf04afc87,
-	0x97f00283,
-	0x0089d00c,
-	0x020721f5,
-/* 0x0226: strand_post */
-	0x87f100f8,
-	0x83f04afc,
-	0x0d97f002,
-	0xf50089d0,
-	0xf8020721,
-/* 0x0239: strand_set */
-	0xfca7f100,
-	0x02a3f04f,
-	0x0500aba2,
-	0xd00fc7f0,
-	0xc7f000ac,
-	0x00bcd00b,
-	0x020721f5,
-	0xf000aed0,
-	0xbcd00ac7,
-	0x0721f500,
-/* 0x0263: strand_ctx_init */
-	0xf100f802,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00399f0,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
 	0x21f50089,
-	0xe7f00213,
-	0x3921f503,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f00f00,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
 	0xfca7f102,
 	0x02a3f046,
 	0x0400aba0,
 	0xf040a0d0,
 	0xbcd001c7,
-	0x0721f500,
+	0x1521f500,
 	0x010c9202,
 	0xf000acd0,
 	0xbcd002c7,
-	0x0721f500,
-	0x2621f502,
+	0x1521f500,
+	0x3421f502,
 	0x8087f102,
 	0x0684b608,
 	0xb70089cf,
 	0x95220080,
-/* 0x02ba: ctx_init_strand_loop */
+/* 0x02ca: ctx_init_strand_loop */
 	0x8ed008fe,
 	0x408ed000,
 	0xb6808acf,
@@ -313,150 +258,160 @@ uint32_t nve0_grgpc_code[] = {
 	0xb60480b6,
 	0x1bf40192,
 	0x08e4b6e8,
-	0xf1f2efbc,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00399f0,
-	0x00f80089,
-/* 0x02ec: error */
-	0xe7f1e0f9,
-	0xe3f09814,
-	0x8d21f440,
-	0x041ce0b7,
-	0xf401f7f0,
-	0xe0fc8d21,
-/* 0x0306: init */
-	0x04bd00f8,
-	0xf10004fe,
-	0xf0120017,
-	0x12d00227,
-	0x3e17f100,
-	0x0010fe04,
-	0x040017f1,
-	0xf0c010d0,
-	0x12d00427,
-	0x1031f400,
-	0x060817f1,
-	0xcf0614b6,
-	0x37f00012,
-	0x1f24f001,
-	0xb60432bb,
-	0x02800132,
-	0x04038003,
-	0x040010b7,
-	0x800012cf,
-	0x27f10002,
-	0x24b60800,
-	0x0022cf06,
-/* 0x035f: init_find_chipset */
-	0xb65817f0,
-	0x13980c10,
-	0x0432b800,
-	0xb00b0bf4,
-	0x1bf40034,
-/* 0x0373: init_context */
-	0xf100f8f1,
-	0xb6080027,
-	0x22cf0624,
-	0xf134bd40,
-	0xb6070047,
-	0x25950644,
-	0x0045d008,
-	0xbd4045d0,
-	0x58f4bde4,
-	0x1f58021e,
-	0x020e4003,
-	0xf5040f40,
-	0xbb013d21,
-	0x3fbb002f,
-	0x041e5800,
-	0x40051f58,
-	0x0f400a0e,
-	0x3d21f50c,
-	0x030e9801,
-	0xbb00effd,
-	0x3ebb002e,
-	0x0040b700,
-	0x0235b613,
-	0xb60043d0,
-	0x35b60825,
-	0x0120b606,
-	0xb60130b6,
-	0x34b60824,
-	0x022fb908,
-	0x026321f5,
-	0xf1003fbb,
-	0xb6080017,
-	0x13d00614,
-	0x0010b740,
-	0xf024bd08,
-	0x12d01f29,
-/* 0x0401: main */
-	0x0031f400,
-	0xf00028f4,
-	0x21f41cd7,
-	0xf401f439,
-	0xf404e4b0,
-	0x81fe1e18,
-	0x0627f001,
-	0x12fd20bd,
-	0x01e4b604,
-	0xfe051efd,
-	0x21f50018,
-	0x0ef404c3,
-/* 0x0431: main_not_ctx_xfer */
-	0x10ef94d3,
-	0xf501f5f0,
-	0xf402ec21,
-/* 0x043e: ih */
-	0x80f9c60e,
-	0xf90188fe,
-	0xf990f980,
-	0xf9b0f9a0,
-	0xf9e0f9d0,
-	0x800acff0,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0xe0f900f8,
+	0x9814e7f1,
+	0xf440e3f0,
+	0xe0b78d21,
+	0xf7f0041c,
+	0x8d21f401,
+	0x00f8e0fc,
+/* 0x0318: init */
+	0x04fe04bd,
+	0x0017f100,
+	0x0227f012,
+	0xf10012d0,
+	0xfe047017,
+	0x17f10010,
+	0x10d00400,
+	0x0427f0c0,
+	0xf40012d0,
+	0x17f11031,
+	0x14b60608,
+	0x0012cf06,
+	0xf00137f0,
+	0x32bb1f24,
+	0x0132b604,
+	0x80050280,
+	0x10b70603,
+	0x12cf0400,
+	0x04028000,
+	0x0c30e7f1,
+	0xbd50e3f0,
+	0xbd34bd24,
+/* 0x0371: init_unk_loop */
+	0x6821f444,
+	0xf400f6b0,
+	0xf7f00f0b,
+	0x04f2bb01,
+	0xb6054ffd,
+/* 0x0386: init_unk_next */
+	0x20b60130,
+	0x04e0b601,
+	0xf40126b0,
+/* 0x0392: init_unk_done */
+	0x0380e21b,
+	0x08048007,
+	0x010027f1,
+	0xcf0223f0,
+	0x34bd0022,
+	0x070047f1,
+	0x950644b6,
+	0x45d00825,
+	0x4045d000,
+	0x98000e98,
+	0x21f5010f,
+	0x2fbb0147,
+	0x003fbb00,
+	0x98010e98,
+	0x21f5020f,
+	0x0e980147,
+	0x00effd05,
+	0xbb002ebb,
+	0x0e98003e,
+	0x030f9802,
+	0x014721f5,
+	0xfd070e98,
+	0x2ebb00ef,
+	0x003ebb00,
+	0x130040b7,
+	0xd00235b6,
+	0x25b60043,
+	0x0635b608,
+	0xb60120b6,
+	0x24b60130,
+	0x0834b608,
+	0xf5022fb9,
+	0xbb027121,
+	0x07f1003f,
+	0x03f00100,
+	0x0003d002,
+	0x24bd04bd,
+	0xf11f29f0,
+	0xf0080007,
+	0x02d00203,
+/* 0x0433: main */
+	0xf404bd00,
+	0x28f40031,
+	0x24d7f000,
+	0xf43921f4,
+	0xe4b0f401,
+	0x1e18f404,
+	0xf00181fe,
+	0x20bd0627,
+	0xb60412fd,
+	0x1efd01e4,
+	0x0018fe05,
+	0x04f721f5,
+/* 0x0463: main_not_ctx_xfer */
+	0x94d30ef4,
+	0xf5f010ef,
+	0xfe21f501,
+	0xc60ef402,
+/* 0x0470: ih */
+	0x88fe80f9,
+	0xf980f901,
+	0xf9a0f990,
+	0xf9d0f9b0,
+	0xbdf0f9e0,
+	0x800acf04,
 	0xf404abc4,
 	0xb7f11d0b,
 	0xd7f01900,
-	0x40becf1c,
+	0x40becf24,
 	0xf400bfcf,
 	0xb0b70421,
 	0xe7f00400,
 	0x00bed001,
-/* 0x0474: ih_no_fifo */
+/* 0x04a8: ih_no_fifo */
 	0xfc400ad0,
 	0xfce0fcf0,
 	0xfcb0fcd0,
 	0xfc90fca0,
 	0x0088fe80,
 	0x32f480fc,
-/* 0x048f: hub_barrier_done */
+/* 0x04c3: hub_barrier_done */
 	0xf001f800,
 	0x0e9801f7,
-	0x04febb00,
+	0x04febb04,
 	0x9418e7f1,
 	0xf440e3f0,
 	0x00f88d21,
-/* 0x04a4: ctx_redswitch */
+/* 0x04d8: ctx_redswitch */
 	0x0614e7f1,
 	0xf006e4b6,
 	0xefd020f7,
 	0x08f7f000,
-/* 0x04b4: ctx_redswitch_delay */
+/* 0x04e8: ctx_redswitch_delay */
 	0xf401f2b6,
 	0xf7f1fd1b,
 	0xefd00a20,
-/* 0x04c3: ctx_xfer */
+/* 0x04f7: ctx_xfer */
 	0xf100f800,
 	0xb60a0417,
 	0x1fd00614,
 	0x0711f400,
-	0x04a421f5,
-/* 0x04d4: ctx_xfer_not_load */
+	0x04d821f5,
+/* 0x0508: ctx_xfer_not_load */
 	0x4afc17f1,
 	0xf00213f0,
 	0x12d00c27,
-	0x0721f500,
+	0x1521f500,
 	0xfc27f102,
 	0x0223f047,
 	0xf00020d0,
@@ -465,31 +420,40 @@ uint32_t nve0_grgpc_code[] = {
 	0xf001acf0,
 	0xb7f002a5,
 	0x50b3f000,
-	0xb6000c98,
+	0xb6040c98,
 	0xbcbb0fc4,
-	0x010c9800,
-	0xf0020d98,
+	0x000c9800,
+	0xf0010d98,
 	0x21f500e7,
-	0xacf0015c,
+	0xacf00166,
+	0x00b7f101,
+	0x50b3f040,
+	0xb6040c98,
+	0xbcbb0fc4,
+	0x010c9800,
+	0x98020d98,
+	0xe7f1060f,
+	0x21f50800,
+	0xacf00166,
 	0x04a5f001,
-	0x4000b7f1,
+	0x3000b7f1,
 	0x9850b3f0,
-	0xc4b6000c,
+	0xc4b6040c,
 	0x00bcbb0f,
-	0x98050c98,
-	0x0f98060d,
-	0x00e7f104,
-	0x5c21f508,
-	0x0721f501,
+	0x98020c98,
+	0x0f98030d,
+	0x00e7f108,
+	0x6621f502,
+	0x1521f501,
 	0x0601f402,
-/* 0x054b: ctx_xfer_post */
+/* 0x05a3: ctx_xfer_post */
 	0xf11412f4,
 	0xf04afc17,
 	0x27f00213,
 	0x0012d00d,
-	0x020721f5,
-/* 0x055c: ctx_xfer_done */
-	0x048f21f5,
+	0x021521f5,
+/* 0x05b4: ctx_xfer_done */
+	0x04c321f5,
 	0x000000f8,
 	0x00000000,
 	0x00000000,
@@ -508,26 +472,4 @@ uint32_t nve0_grgpc_code[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
 };
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvf0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvf0.fuc
new file mode 100644
index 000000000000..90bbe525b626
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvf0.fuc
@@ -0,0 +1,42 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#define NV_PGRAPH_GPCX_UNK__SIZE                                     0x00000002
+
+#define CHIPSET GK110
+#include "macros.fuc"
+
+.section #nvf0_grgpc_data
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "gpc.fuc"
+#undef INCLUDE_DATA
+
+.section #nvf0_grgpc_code
+#define INCLUDE_CODE
+bra #init
+#include "com.fuc"
+#include "gpc.fuc"
+.align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvf0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvf0.fuc.h
new file mode 100644
index 000000000000..f870507be880
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/gpcnvf0.fuc.h
@@ -0,0 +1,475 @@
+uint32_t nvf0_grgpc_data[] = {
+/* 0x0000: gpc_mmio_list_head */
+	0x0000006c,
+/* 0x0004: gpc_mmio_list_tail */
+/* 0x0004: tpc_mmio_list_head */
+	0x0000006c,
+/* 0x0008: tpc_mmio_list_tail */
+/* 0x0008: unk_mmio_list_head */
+	0x0000006c,
+/* 0x000c: unk_mmio_list_tail */
+	0x0000006c,
+/* 0x0010: gpc_id */
+	0x00000000,
+/* 0x0014: tpc_count */
+	0x00000000,
+/* 0x0018: tpc_mask */
+	0x00000000,
+/* 0x001c: unk_count */
+	0x00000000,
+/* 0x0020: unk_mask */
+	0x00000000,
+/* 0x0024: cmd_queue */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+};
+
+uint32_t nvf0_grgpc_code[] = {
+	0x03180ef5,
+/* 0x0004: queue_put */
+	0x9800d898,
+	0x86f001d9,
+	0x0489b808,
+	0xf00c1bf4,
+	0x21f502f7,
+	0x00f802fe,
+/* 0x001c: queue_put_next */
+	0xb60798c4,
+	0x8dbb0384,
+	0x0880b600,
+	0x80008e80,
+	0x90b6018f,
+	0x0f94f001,
+	0xf801d980,
+/* 0x0039: queue_get */
+	0x0131f400,
+	0x9800d898,
+	0x89b801d9,
+	0x210bf404,
+	0xb60789c4,
+	0x9dbb0394,
+	0x0890b600,
+	0x98009e98,
+	0x80b6019f,
+	0x0f84f001,
+	0xf400d880,
+/* 0x0066: queue_get_done */
+	0x00f80132,
+/* 0x0068: nv_rd32 */
+	0x0728b7f1,
+	0xb906b4b6,
+	0xc9f002ec,
+	0x00bcd01f,
+/* 0x0078: nv_rd32_wait */
+	0xc800bccf,
+	0x1bf41fcc,
+	0x06a7f0fa,
+	0x010921f5,
+	0xf840bfcf,
+/* 0x008d: nv_wr32 */
+	0x28b7f100,
+	0x06b4b607,
+	0xb980bfd0,
+	0xc9f002ec,
+	0x1ec9f01f,
+/* 0x00a3: nv_wr32_wait */
+	0xcf00bcd0,
+	0xccc800bc,
+	0xfa1bf41f,
+/* 0x00ae: watchdog_reset */
+	0x87f100f8,
+	0x84b60430,
+	0x1ff9f006,
+	0xf8008fd0,
+/* 0x00bd: watchdog_clear */
+	0x3087f100,
+	0x0684b604,
+	0xf80080d0,
+/* 0x00c9: wait_donez */
+	0xf094bd00,
+	0x07f10099,
+	0x03f03700,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
+	0xf4888aff,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f03700,
+	0x0009d002,
+	0x87f104bd,
+	0x84b60818,
+	0x008ad006,
+/* 0x0124: wait_doneo_e */
+	0x040087f1,
+	0xcf0684b6,
+	0x8aff0088,
+	0xf30bf488,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf0370007,
+	0x09d00203,
+	0xf104bd00,
+	0xb6071087,
+	0x94bd0684,
+	0xf405bbfd,
+	0x8bd0090b,
+	0x0099f000,
+/* 0x018c: mmctx_base_disabled */
+	0xf405eefd,
+	0x8ed00c0b,
+	0xc08fd080,
+/* 0x019b: mmctx_multi_disabled */
+	0xb70199f0,
+	0xc8010080,
+	0xb4b600ab,
+	0x0cb9f010,
+	0xb601aec8,
+	0xbefd11e4,
+	0x008bd005,
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
+	0xf0008ecf,
+	0x0bf41fe4,
+	0x00ce98fa,
+	0xd005e9fd,
+	0xc0b6c08e,
+	0x04cdb804,
+	0xc8e81bf4,
+	0x1bf402ab,
+/* 0x01d5: mmctx_fini_wait */
+	0x008bcf18,
+	0xb01fb4f0,
+	0x1bf410b4,
+	0x02a7f0f7,
+	0xf4c921f4,
+/* 0x01ea: mmctx_stop */
+	0xabc81b0e,
+	0x10b4b600,
+	0xf00cb9f0,
+	0x8bd012b9,
+/* 0x01f9: mmctx_stop_wait */
+	0x008bcf00,
+	0xf412bbc8,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
+	0x21f50089,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f03700,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
+	0xfca7f102,
+	0x02a3f046,
+	0x0400aba0,
+	0xf040a0d0,
+	0xbcd001c7,
+	0x1521f500,
+	0x010c9202,
+	0xf000acd0,
+	0xbcd002c7,
+	0x1521f500,
+	0x3421f502,
+	0x8087f102,
+	0x0684b608,
+	0xb70089cf,
+	0x95220080,
+/* 0x02ca: ctx_init_strand_loop */
+	0x8ed008fe,
+	0x408ed000,
+	0xb6808acf,
+	0xa0b606a5,
+	0x00eabb01,
+	0xb60480b6,
+	0x1bf40192,
+	0x08e4b6e8,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0xe0f900f8,
+	0x9814e7f1,
+	0xf440e3f0,
+	0xe0b78d21,
+	0xf7f0041c,
+	0x8d21f401,
+	0x00f8e0fc,
+/* 0x0318: init */
+	0x04fe04bd,
+	0x0017f100,
+	0x0227f012,
+	0xf10012d0,
+	0xfe047017,
+	0x17f10010,
+	0x10d00400,
+	0x0427f0c0,
+	0xf40012d0,
+	0x17f11031,
+	0x14b60608,
+	0x0012cf06,
+	0xf00137f0,
+	0x32bb1f24,
+	0x0132b604,
+	0x80050280,
+	0x10b70603,
+	0x12cf0400,
+	0x04028000,
+	0x0c30e7f1,
+	0xbd50e3f0,
+	0xbd34bd24,
+/* 0x0371: init_unk_loop */
+	0x6821f444,
+	0xf400f6b0,
+	0xf7f00f0b,
+	0x04f2bb01,
+	0xb6054ffd,
+/* 0x0386: init_unk_next */
+	0x20b60130,
+	0x04e0b601,
+	0xf40226b0,
+/* 0x0392: init_unk_done */
+	0x0380e21b,
+	0x08048007,
+	0x010027f1,
+	0xcf0223f0,
+	0x34bd0022,
+	0x070047f1,
+	0x950644b6,
+	0x45d00825,
+	0x4045d000,
+	0x98000e98,
+	0x21f5010f,
+	0x2fbb0147,
+	0x003fbb00,
+	0x98010e98,
+	0x21f5020f,
+	0x0e980147,
+	0x00effd05,
+	0xbb002ebb,
+	0x0e98003e,
+	0x030f9802,
+	0x014721f5,
+	0xfd070e98,
+	0x2ebb00ef,
+	0x003ebb00,
+	0x130040b7,
+	0xd00235b6,
+	0x25b60043,
+	0x0635b608,
+	0xb60120b6,
+	0x24b60130,
+	0x0834b608,
+	0xf5022fb9,
+	0xbb027121,
+	0x07f1003f,
+	0x03f00100,
+	0x0003d002,
+	0x24bd04bd,
+	0xf11f29f0,
+	0xf0300007,
+	0x02d00203,
+/* 0x0433: main */
+	0xf404bd00,
+	0x28f40031,
+	0x24d7f000,
+	0xf43921f4,
+	0xe4b0f401,
+	0x1e18f404,
+	0xf00181fe,
+	0x20bd0627,
+	0xb60412fd,
+	0x1efd01e4,
+	0x0018fe05,
+	0x04f721f5,
+/* 0x0463: main_not_ctx_xfer */
+	0x94d30ef4,
+	0xf5f010ef,
+	0xfe21f501,
+	0xc60ef402,
+/* 0x0470: ih */
+	0x88fe80f9,
+	0xf980f901,
+	0xf9a0f990,
+	0xf9d0f9b0,
+	0xbdf0f9e0,
+	0x800acf04,
+	0xf404abc4,
+	0xb7f11d0b,
+	0xd7f01900,
+	0x40becf24,
+	0xf400bfcf,
+	0xb0b70421,
+	0xe7f00400,
+	0x00bed001,
+/* 0x04a8: ih_no_fifo */
+	0xfc400ad0,
+	0xfce0fcf0,
+	0xfcb0fcd0,
+	0xfc90fca0,
+	0x0088fe80,
+	0x32f480fc,
+/* 0x04c3: hub_barrier_done */
+	0xf001f800,
+	0x0e9801f7,
+	0x04febb04,
+	0x9418e7f1,
+	0xf440e3f0,
+	0x00f88d21,
+/* 0x04d8: ctx_redswitch */
+	0x0614e7f1,
+	0xf006e4b6,
+	0xefd020f7,
+	0x08f7f000,
+/* 0x04e8: ctx_redswitch_delay */
+	0xf401f2b6,
+	0xf7f1fd1b,
+	0xefd00a20,
+/* 0x04f7: ctx_xfer */
+	0xf100f800,
+	0xb60a0417,
+	0x1fd00614,
+	0x0711f400,
+	0x04d821f5,
+/* 0x0508: ctx_xfer_not_load */
+	0x4afc17f1,
+	0xf00213f0,
+	0x12d00c27,
+	0x1521f500,
+	0xfc27f102,
+	0x0223f047,
+	0xf00020d0,
+	0x20b6012c,
+	0x0012d003,
+	0xf001acf0,
+	0xb7f002a5,
+	0x50b3f000,
+	0xb6040c98,
+	0xbcbb0fc4,
+	0x000c9800,
+	0xf0010d98,
+	0x21f500e7,
+	0xacf00166,
+	0x00b7f101,
+	0x50b3f040,
+	0xb6040c98,
+	0xbcbb0fc4,
+	0x010c9800,
+	0x98020d98,
+	0xe7f1060f,
+	0x21f50800,
+	0xacf00166,
+	0x04a5f001,
+	0x3000b7f1,
+	0x9850b3f0,
+	0xc4b6040c,
+	0x00bcbb0f,
+	0x98020c98,
+	0x0f98030d,
+	0x00e7f108,
+	0x6621f502,
+	0x1521f501,
+	0x0601f402,
+/* 0x05a3: ctx_xfer_post */
+	0xf11412f4,
+	0xf04afc17,
+	0x27f00213,
+	0x0012d00d,
+	0x021521f5,
+/* 0x05b4: ctx_xfer_done */
+	0x04c321f5,
+	0x000000f8,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hub.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hub.fuc
new file mode 100644
index 000000000000..b82d2ae89917
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hub.fuc
@@ -0,0 +1,724 @@
+/* fuc microcode for nvc0 PGRAPH/HUB
+ *
+ * Copyright 2011 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#ifdef INCLUDE_DATA
+hub_mmio_list_head:	.b32 #hub_mmio_list_base
+hub_mmio_list_tail:	.b32 #hub_mmio_list_next
+
+gpc_count:		.b32 0
+rop_count:		.b32 0
+cmd_queue:		queue_init
+
+ctx_current:		.b32 0
+
+.align 256
+chan_data:
+chan_mmio_count:	.b32 0
+chan_mmio_address:	.b32 0
+
+.align 256
+xfer_data: 		.skip 256
+
+hub_mmio_list_base:
+.b32 0x0417e91c // 0x17e91c, 2
+hub_mmio_list_next:
+#endif
+
+#ifdef INCLUDE_CODE
+// reports an exception to the host
+//
+// In: $r15 error code (see nvc0.fuc)
+//
+error:
+	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_VAL(5), 0, $r15)
+	mov $r15 1
+	nv_iowr(NV_PGRAPH_FECS_INTR_UP_SET, 0, $r15)
+	ret
+
+// HUB fuc initialisation, executed by triggering ucode start, will
+// fall through to main loop after completion.
+//
+// Output:
+//   CC_SCRATCH[0]:
+//	     31:31: set to signal completion
+//   CC_SCRATCH[1]:
+//	      31:0: total PGRAPH context size
+//
+init:
+	clear b32 $r0
+	mov $sp $r0
+	mov $xdbase $r0
+
+	// enable fifo access
+	mov $r1 0x1200
+	mov $r2 2
+	iowr I[$r1 + 0x000] $r2	// FIFO_ENABLE
+
+	// setup i0 handler, and route all interrupts to it
+	mov $r1 #ih
+	mov $iv0 $r1
+	mov $r1 0x400
+	iowr I[$r1 + 0x300] $r0	// INTR_DISPATCH
+
+	// route HUB_CHANNEL_SWITCH to fuc interrupt 8
+	mov $r3 0x404
+	shl b32 $r3 6
+	mov $r2 0x2003		// { HUB_CHANNEL_SWITCH, ZERO } -> intr 8
+	iowr I[$r3 + 0x000] $r2
+
+	// not sure what these are, route them because NVIDIA does, and
+	// the IRQ handler will signal the host if we ever get one.. we
+	// may find out if/why we need to handle these if so..
+	//
+	mov $r2 0x2004
+	iowr I[$r3 + 0x004] $r2 // { 0x04, ZERO } -> intr 9
+	mov $r2 0x200b
+	iowr I[$r3 + 0x008] $r2 // { 0x0b, ZERO } -> intr 10
+	mov $r2 0x200c
+	iowr I[$r3 + 0x01c] $r2 // { 0x0c, ZERO } -> intr 15
+
+	// enable all INTR_UP interrupts
+	mov $r2 0xc24
+	shl b32 $r2 6
+	not b32 $r3 $r0
+	iowr I[$r2] $r3
+
+	// enable fifo, ctxsw, 9, 10, 15 interrupts
+	mov $r2 -0x78fc		// 0x8704
+	sethi $r2 0
+	iowr I[$r1 + 0x000] $r2	// INTR_EN_SET
+
+	// fifo level triggered, rest edge
+	sub b32 $r1 0x100
+	mov $r2 4
+	iowr I[$r1] $r2
+
+	// enable interrupts
+	bset $flags ie0
+
+	// fetch enabled GPC/ROP counts
+	mov $r14 -0x69fc	// 0x409604
+	sethi $r14 0x400000
+	call #nv_rd32
+	extr $r1 $r15 16:20
+	st b32 D[$r0 + #rop_count] $r1
+	and $r15 0x1f
+	st b32 D[$r0 + #gpc_count] $r15
+
+	// set BAR_REQMASK to GPC mask
+	mov $r1 1
+	shl b32 $r1 $r15
+	sub b32 $r1 1
+	mov $r2 0x40c
+	shl b32 $r2 6
+	iowr I[$r2 + 0x000] $r1
+	iowr I[$r2 + 0x100] $r1
+
+	// context size calculation, reserve first 256 bytes for use by fuc
+	mov $r1 256
+
+	// calculate size of mmio context data
+	ld b32 $r14 D[$r0 + #hub_mmio_list_head]
+	ld b32 $r15 D[$r0 + #hub_mmio_list_tail]
+	call #mmctx_size
+
+	// set mmctx base addresses now so we don't have to do it later,
+	// they don't (currently) ever change
+	mov $r3 0x700
+	shl b32 $r3 6
+	shr b32 $r4 $r1 8
+	iowr I[$r3 + 0x000] $r4		// MMCTX_SAVE_SWBASE
+	iowr I[$r3 + 0x100] $r4		// MMCTX_LOAD_SWBASE
+	add b32 $r3 0x1300
+	add b32 $r1 $r15
+	shr b32 $r15 2
+	iowr I[$r3 + 0x000] $r15	// MMCTX_LOAD_COUNT, wtf for?!?
+
+	// strands, base offset needs to be aligned to 256 bytes
+	shr b32 $r1 8
+	add b32 $r1 1
+	shl b32 $r1 8
+	mov b32 $r15 $r1
+	call #strand_ctx_init
+	add b32 $r1 $r15
+
+	// initialise each GPC in sequence by passing in the offset of its
+	// context data in GPCn_CC_SCRATCH[1], and starting its FUC (which
+	// has previously been uploaded by the host) running.
+	//
+	// the GPC fuc init sequence will set GPCn_CC_SCRATCH[0] bit 31
+	// when it has completed, and return the size of its context data
+	// in GPCn_CC_SCRATCH[1]
+	//
+	ld b32 $r3 D[$r0 + #gpc_count]
+	mov $r4 0x2000
+	sethi $r4 0x500000
+	init_gpc:
+		// setup, and start GPC ucode running
+		add b32 $r14 $r4 0x804
+		mov b32 $r15 $r1
+		call #nv_wr32			// CC_SCRATCH[1] = ctx offset
+		add b32 $r14 $r4 0x10c
+		clear b32 $r15
+		call #nv_wr32
+		add b32 $r14 $r4 0x104
+		call #nv_wr32			// ENTRY
+		add b32 $r14 $r4 0x100
+		mov $r15 2			// CTRL_START_TRIGGER
+		call #nv_wr32			// CTRL
+
+		// wait for it to complete, and adjust context size
+		add b32 $r14 $r4 0x800
+		init_gpc_wait:
+			call #nv_rd32
+			xbit $r15 $r15 31
+			bra e #init_gpc_wait
+		add b32 $r14 $r4 0x804
+		call #nv_rd32
+		add b32 $r1 $r15
+
+		// next!
+		add b32 $r4 0x8000
+		sub b32 $r3 1
+		bra ne #init_gpc
+
+	// save context size, and tell host we're ready
+	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_VAL(1), 0, $r1)
+	clear b32 $r1
+	bset $r1 31
+	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_SET(0), 0, $r1)
+
+// Main program loop, very simple, sleeps until woken up by the interrupt
+// handler, pulls a command from the queue and executes its handler
+//
+main:
+	// sleep until we have something to do
+	bset $flags $p0
+	sleep $p0
+	mov $r13 #cmd_queue
+	call #queue_get
+	bra $p1 #main
+
+	// context switch, requested by GPU?
+	cmpu b32 $r14 0x4001
+	bra ne #main_not_ctx_switch
+		trace_set(T_AUTO)
+		mov $r1 0xb00
+		shl b32 $r1 6
+		iord $r2 I[$r1 + 0x100]		// CHAN_NEXT
+		iord $r1 I[$r1 + 0x000]		// CHAN_CUR
+
+		xbit $r3 $r1 31
+		bra e #chsw_no_prev
+			xbit $r3 $r2 31
+			bra e #chsw_prev_no_next
+				push $r2
+				mov b32 $r2 $r1
+				trace_set(T_SAVE)
+				bclr $flags $p1
+				bset $flags $p2
+				call #ctx_xfer
+				trace_clr(T_SAVE);
+				pop $r2
+				trace_set(T_LOAD);
+				bset $flags $p1
+				call #ctx_xfer
+				trace_clr(T_LOAD);
+				bra #chsw_done
+			chsw_prev_no_next:
+				push $r2
+				mov b32 $r2 $r1
+				bclr $flags $p1
+				bclr $flags $p2
+				call #ctx_xfer
+				pop $r2
+				mov $r1 0xb00
+				shl b32 $r1 6
+				iowr I[$r1] $r2
+				bra #chsw_done
+		chsw_no_prev:
+			xbit $r3 $r2 31
+			bra e #chsw_done
+				bset $flags $p1
+				bclr $flags $p2
+				call #ctx_xfer
+
+		// ack the context switch request
+		chsw_done:
+		mov $r1 0xb0c
+		shl b32 $r1 6
+		mov $r2 1
+		iowr I[$r1 + 0x000] $r2		// 0x409b0c
+		trace_clr(T_AUTO)
+		bra #main
+
+	// request to set current channel? (*not* a context switch)
+	main_not_ctx_switch:
+	cmpu b32 $r14 0x0001
+	bra ne #main_not_ctx_chan
+		mov b32 $r2 $r15
+		call #ctx_chan
+		bra #main_done
+
+	// request to store current channel context?
+	main_not_ctx_chan:
+	cmpu b32 $r14 0x0002
+	bra ne #main_not_ctx_save
+		trace_set(T_SAVE)
+		bclr $flags $p1
+		bclr $flags $p2
+		call #ctx_xfer
+		trace_clr(T_SAVE)
+		bra #main_done
+
+	main_not_ctx_save:
+		shl b32 $r15 $r14 16
+		or $r15 E_BAD_COMMAND
+		call #error
+		bra #main
+
+	main_done:
+	clear b32 $r2
+	bset $r2 31
+	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_SET(0), 0, $r2)
+	bra #main
+
+// interrupt handler
+ih:
+	push $r8
+	mov $r8 $flags
+	push $r8
+	push $r9
+	push $r10
+	push $r11
+	push $r13
+	push $r14
+	push $r15
+	clear b32 $r0
+
+	// incoming fifo command?
+	iord $r10 I[$r0 + 0x200]	// INTR
+	and $r11 $r10 0x00000004
+	bra e #ih_no_fifo
+		// queue incoming fifo command for later processing
+		mov $r11 0x1900
+		mov $r13 #cmd_queue
+		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
+		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
+		call #queue_put
+		add b32 $r11 0x400
+		mov $r14 1
+		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
+
+	// context switch request?
+	ih_no_fifo:
+	and $r11 $r10 0x00000100
+	bra e #ih_no_ctxsw
+		// enqueue a context switch for later processing
+		mov $r13 #cmd_queue
+		mov $r14 0x4001
+		call #queue_put
+
+	// anything we didn't handle, bring it to the host's attention
+	ih_no_ctxsw:
+	mov $r11 0x104
+	not b32 $r11
+	and $r11 $r10 $r11
+	bra e #ih_no_other
+		mov $r10 0xc1c
+		shl b32 $r10 6
+		iowr I[$r10] $r11	// INTR_UP_SET
+
+	// ack, and wake up main()
+	ih_no_other:
+	iowr I[$r0 + 0x100] $r10	// INTR_ACK
+
+	pop $r15
+	pop $r14
+	pop $r13
+	pop $r11
+	pop $r10
+	pop $r9
+	pop $r8
+	mov $flags $r8
+	pop $r8
+	bclr $flags $p0
+	iret
+
+#if CHIPSET < GK100
+// Not real sure, but, MEM_CMD 7 will hang forever if this isn't done
+ctx_4160s:
+	mov $r14 0x4160
+	sethi $r14 0x400000
+	mov $r15 1
+	call #nv_wr32
+	ctx_4160s_wait:
+		call #nv_rd32
+		xbit $r15 $r15 4
+		bra e #ctx_4160s_wait
+	ret
+
+// Without clearing again at end of xfer, some things cause PGRAPH
+// to hang with STATUS=0x00000007 until it's cleared.. fbcon can
+// still function with it set however...
+ctx_4160c:
+	mov $r14 0x4160
+	sethi $r14 0x400000
+	clear b32 $r15
+	call #nv_wr32
+	ret
+#endif
+
+// Again, not real sure
+//
+// In: $r15 value to set 0x404170 to
+//
+ctx_4170s:
+	mov $r14 0x4170
+	sethi $r14 0x400000
+	or $r15 0x10
+	call #nv_wr32
+	ret
+
+// Waits for a ctx_4170s() call to complete
+//
+ctx_4170w:
+	mov $r14 0x4170
+	sethi $r14 0x400000
+	call #nv_rd32
+	and $r15 0x10
+	bra ne #ctx_4170w
+	ret
+
+// Disables various things, waits a bit, and re-enables them..
+//
+// Not sure how exactly this helps, perhaps "ENABLE" is not such a
+// good description for the bits we turn off?  Anyways, without this,
+// funny things happen.
+//
+ctx_redswitch:
+	mov $r14 0x614
+	shl b32 $r14 6
+	mov $r15 0x270
+	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_GPC, POWER_ALL
+	mov $r15 8
+	ctx_redswitch_delay:
+		sub b32 $r15 1
+		bra ne #ctx_redswitch_delay
+	mov $r15 0x770
+	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_ALL, POWER_ALL
+	ret
+
+// Not a clue what this is for, except that unless the value is 0x10, the
+// strand context is saved (and presumably restored) incorrectly..
+//
+// In: $r15 value to set to (0x00/0x10 are used)
+//
+ctx_86c:
+	mov $r14 0x86c
+	shl b32 $r14 6
+	iowr I[$r14] $r15	// HUB(0x86c) = val
+	mov $r14 -0x75ec
+	sethi $r14 0x400000
+	call #nv_wr32		// ROP(0xa14) = val
+	mov $r14 -0x5794
+	sethi $r14 0x410000
+	call #nv_wr32		// GPC(0x86c) = val
+	ret
+
+// ctx_load - load's a channel's ctxctl data, and selects its vm
+//
+// In: $r2 channel address
+//
+ctx_load:
+	trace_set(T_CHAN)
+
+	// switch to channel, somewhat magic in parts..
+	mov $r10 12		// DONE_UNK12
+	call #wait_donez
+	mov $r1 0xa24
+	shl b32 $r1 6
+	iowr I[$r1 + 0x000] $r0	// 0x409a24
+	mov $r3 0xb00
+	shl b32 $r3 6
+	iowr I[$r3 + 0x100] $r2	// CHAN_NEXT
+	mov $r1 0xa0c
+	shl b32 $r1 6
+	mov $r4 7
+	iowr I[$r1 + 0x000] $r2 // MEM_CHAN
+	iowr I[$r1 + 0x100] $r4	// MEM_CMD
+	ctx_chan_wait_0:
+		iord $r4 I[$r1 + 0x100]
+		and $r4 0x1f
+		bra ne #ctx_chan_wait_0
+	iowr I[$r3 + 0x000] $r2	// CHAN_CUR
+
+	// load channel header, fetch PGRAPH context pointer
+	mov $xtargets $r0
+	bclr $r2 31
+	shl b32 $r2 4
+	add b32 $r2 2
+
+	trace_set(T_LCHAN)
+	mov $r1 0xa04
+	shl b32 $r1 6
+	iowr I[$r1 + 0x000] $r2		// MEM_BASE
+	mov $r1 0xa20
+	shl b32 $r1 6
+	mov $r2 0x0002
+	sethi $r2 0x80000000
+	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vram
+	mov $r1 0x10			// chan + 0x0210
+	mov $r2 #xfer_data
+	sethi $r2 0x00020000		// 16 bytes
+	xdld $r1 $r2
+	xdwait
+	trace_clr(T_LCHAN)
+
+	// update current context
+	ld b32 $r1 D[$r0 + #xfer_data + 4]
+	shl b32 $r1 24
+	ld b32 $r2 D[$r0 + #xfer_data + 0]
+	shr b32 $r2 8
+	or $r1 $r2
+	st b32 D[$r0 + #ctx_current] $r1
+
+	// set transfer base to start of context, and fetch context header
+	trace_set(T_LCTXH)
+	mov $r2 0xa04
+	shl b32 $r2 6
+	iowr I[$r2 + 0x000] $r1		// MEM_BASE
+	mov $r2 1
+	mov $r1 0xa20
+	shl b32 $r1 6
+	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vm
+	mov $r1 #chan_data
+	sethi $r1 0x00060000		// 256 bytes
+	xdld $r0 $r1
+	xdwait
+	trace_clr(T_LCTXH)
+
+	trace_clr(T_CHAN)
+	ret
+
+// ctx_chan - handler for HUB_SET_CHAN command, will set a channel as
+//            the active channel for ctxctl, but not actually transfer
+//            any context data.  intended for use only during initial
+//            context construction.
+//
+// In: $r2 channel address
+//
+ctx_chan:
+#if CHIPSET < GK100
+	call #ctx_4160s
+#endif
+	call #ctx_load
+	mov $r10 12			// DONE_UNK12
+	call #wait_donez
+	mov $r1 0xa10
+	shl b32 $r1 6
+	mov $r2 5
+	iowr I[$r1 + 0x000] $r2		// MEM_CMD = 5 (???)
+	ctx_chan_wait:
+		iord $r2 I[$r1 + 0x000]
+		or $r2 $r2
+		bra ne #ctx_chan_wait
+#if CHIPSET < GK100
+	call #ctx_4160c
+#endif
+	ret
+
+// Execute per-context state overrides list
+//
+// Only executed on the first load of a channel.  Might want to look into
+// removing this and having the host directly modify the channel's context
+// to change this state...  The nouveau DRM already builds this list as
+// it's definitely needed for NVIDIA's, so we may as well use it for now
+//
+// Input: $r1 mmio list length
+//
+ctx_mmio_exec:
+	// set transfer base to be the mmio list
+	ld b32 $r3 D[$r0 + #chan_mmio_address]
+	mov $r2 0xa04
+	shl b32 $r2 6
+	iowr I[$r2 + 0x000] $r3		// MEM_BASE
+
+	clear b32 $r3
+	ctx_mmio_loop:
+		// fetch next 256 bytes of mmio list if necessary
+		and $r4 $r3 0xff
+		bra ne #ctx_mmio_pull
+			mov $r5 #xfer_data
+			sethi $r5 0x00060000	// 256 bytes
+			xdld $r3 $r5
+			xdwait
+
+		// execute a single list entry
+		ctx_mmio_pull:
+		ld b32 $r14 D[$r4 + #xfer_data + 0x00]
+		ld b32 $r15 D[$r4 + #xfer_data + 0x04]
+		call #nv_wr32
+
+		// next!
+		add b32 $r3 8
+		sub b32 $r1 1
+		bra ne #ctx_mmio_loop
+
+	// set transfer base back to the current context
+	ctx_mmio_done:
+	ld b32 $r3 D[$r0 + #ctx_current]
+	iowr I[$r2 + 0x000] $r3		// MEM_BASE
+
+	// disable the mmio list now, we don't need/want to execute it again
+	st b32 D[$r0 + #chan_mmio_count] $r0
+	mov $r1 #chan_data
+	sethi $r1 0x00060000		// 256 bytes
+	xdst $r0 $r1
+	xdwait
+	ret
+
+// Transfer HUB context data between GPU and storage area
+//
+// In: $r2 channel address
+//     $p1 clear on save, set on load
+//     $p2 set if opposite direction done/will be done, so:
+//		on save it means: "a load will follow this save"
+//		on load it means: "a save preceeded this load"
+//
+ctx_xfer:
+	// according to mwk, some kind of wait for idle
+	mov $r15 0xc00
+	shl b32 $r15 6
+	mov $r14 4
+	iowr I[$r15 + 0x200] $r14
+	ctx_xfer_idle:
+		iord $r14 I[$r15 + 0x000]
+		and $r14 0x2000
+		bra ne #ctx_xfer_idle
+
+	bra not $p1 #ctx_xfer_pre
+	bra $p2 #ctx_xfer_pre_load
+	ctx_xfer_pre:
+		mov $r15 0x10
+		call #ctx_86c
+#if CHIPSET < GK100
+		call #ctx_4160s
+#endif
+		bra not $p1 #ctx_xfer_exec
+
+	ctx_xfer_pre_load:
+		mov $r15 2
+		call #ctx_4170s
+		call #ctx_4170w
+		call #ctx_redswitch
+		clear b32 $r15
+		call #ctx_4170s
+		call #ctx_load
+
+	// fetch context pointer, and initiate xfer on all GPCs
+	ctx_xfer_exec:
+	ld b32 $r1 D[$r0 + #ctx_current]
+	mov $r2 0x414
+	shl b32 $r2 6
+	iowr I[$r2 + 0x000] $r0	// BAR_STATUS = reset
+	mov $r14 -0x5b00
+	sethi $r14 0x410000
+	mov b32 $r15 $r1
+	call #nv_wr32		// GPC_BCAST_WRCMD_DATA = ctx pointer
+	add b32 $r14 4
+	xbit $r15 $flags $p1
+	xbit $r2 $flags $p2
+	shl b32 $r2 1
+	or $r15 $r2
+	call #nv_wr32		// GPC_BCAST_WRCMD_CMD = GPC_XFER(type)
+
+	// strands
+	mov $r1 0x4afc
+	sethi $r1 0x20000
+	mov $r2 0xc
+	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
+	call #strand_wait
+	mov $r2 0x47fc
+	sethi $r2 0x20000
+	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
+	xbit $r2 $flags $p1
+	add b32 $r2 3
+	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
+
+	// mmio context
+	xbit $r10 $flags $p1	// direction
+	or $r10 6		// first, last
+	mov $r11 0		// base = 0
+	ld b32 $r12 D[$r0 + #hub_mmio_list_head]
+	ld b32 $r13 D[$r0 + #hub_mmio_list_tail]
+	mov $r14 0		// not multi
+	call #mmctx_xfer
+
+	// wait for GPCs to all complete
+	mov $r10 8		// DONE_BAR
+	call #wait_doneo
+
+	// wait for strand xfer to complete
+	call #strand_wait
+
+	// post-op
+	bra $p1 #ctx_xfer_post
+		mov $r10 12		// DONE_UNK12
+		call #wait_donez
+		mov $r1 0xa10
+		shl b32 $r1 6
+		mov $r2 5
+		iowr I[$r1] $r2		// MEM_CMD
+		ctx_xfer_post_save_wait:
+			iord $r2 I[$r1]
+			or $r2 $r2
+			bra ne #ctx_xfer_post_save_wait
+
+	bra $p2 #ctx_xfer_done
+	ctx_xfer_post:
+		mov $r15 2
+		call #ctx_4170s
+		clear b32 $r15
+		call #ctx_86c
+		call #strand_post
+		call #ctx_4170w
+		clear b32 $r15
+		call #ctx_4170s
+
+		bra not $p1 #ctx_xfer_no_post_mmio
+		ld b32 $r1 D[$r0 + #chan_mmio_count]
+		or $r1 $r1
+		bra e #ctx_xfer_no_post_mmio
+			call #ctx_mmio_exec
+
+		ctx_xfer_no_post_mmio:
+#if CHIPSET < GK100
+		call #ctx_4160c
+#endif
+
+	ctx_xfer_done:
+	ret
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc
index 7fbdebb2bafb..3ff52badf932 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc
@@ -1,6 +1,5 @@
-/* fuc microcode for nvc0 PGRAPH/HUB
- *
- * Copyright 2011 Red Hat Inc.
+/*
+ * Copyright 2013 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -20,850 +19,22 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
- * Authors: Ben Skeggs
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
-/* To build:
- *    m4 hubnvc0.fuc | envyas -a -w -m fuc -V fuc3 -o hubnvc0.fuc.h
- */
+#define CHIPSET GF100
+#include "macros.fuc"
 
 .section #nvc0_grhub_data
-include(`nvc0.fuc')
-gpc_count:		.b32 0
-rop_count:		.b32 0
-cmd_queue:		queue_init
-hub_mmio_list_head:	.b32 0
-hub_mmio_list_tail:	.b32 0
-
-ctx_current:		.b32 0
-
-chipsets:
-.b8  0xc0 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xc1 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc1_hub_mmio_tail
-.b8  0xc3 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xc4 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xc8 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xce 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xcf 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xd9 0 0 0
-.b16 #nvd9_hub_mmio_head
-.b16 #nvd9_hub_mmio_tail
-.b8  0xd7 0 0 0
-.b16 #nvd9_hub_mmio_head
-.b16 #nvd9_hub_mmio_tail
-.b8  0 0 0 0
-
-nvc0_hub_mmio_head:
-mmctx_data(0x17e91c, 2)
-mmctx_data(0x400204, 2)
-mmctx_data(0x404004, 11)
-mmctx_data(0x404044, 1)
-mmctx_data(0x404094, 14)
-mmctx_data(0x4040d0, 7)
-mmctx_data(0x4040f8, 1)
-mmctx_data(0x404130, 3)
-mmctx_data(0x404150, 3)
-mmctx_data(0x404164, 2)
-mmctx_data(0x404174, 3)
-mmctx_data(0x404200, 8)
-mmctx_data(0x404404, 14)
-mmctx_data(0x404460, 4)
-mmctx_data(0x404480, 1)
-mmctx_data(0x404498, 1)
-mmctx_data(0x404604, 4)
-mmctx_data(0x404618, 32)
-mmctx_data(0x404698, 21)
-mmctx_data(0x4046f0, 2)
-mmctx_data(0x404700, 22)
-mmctx_data(0x405800, 1)
-mmctx_data(0x405830, 3)
-mmctx_data(0x405854, 1)
-mmctx_data(0x405870, 4)
-mmctx_data(0x405a00, 2)
-mmctx_data(0x405a18, 1)
-mmctx_data(0x406020, 1)
-mmctx_data(0x406028, 4)
-mmctx_data(0x4064a8, 2)
-mmctx_data(0x4064b4, 2)
-mmctx_data(0x407804, 1)
-mmctx_data(0x40780c, 6)
-mmctx_data(0x4078bc, 1)
-mmctx_data(0x408000, 7)
-mmctx_data(0x408064, 1)
-mmctx_data(0x408800, 3)
-mmctx_data(0x408900, 4)
-mmctx_data(0x408980, 1)
-nvc0_hub_mmio_tail:
-mmctx_data(0x4064c0, 2)
-nvc1_hub_mmio_tail:
-
-nvd9_hub_mmio_head:
-mmctx_data(0x17e91c, 2)
-mmctx_data(0x400204, 2)
-mmctx_data(0x404004, 10)
-mmctx_data(0x404044, 1)
-mmctx_data(0x404094, 14)
-mmctx_data(0x4040d0, 7)
-mmctx_data(0x4040f8, 1)
-mmctx_data(0x404130, 3)
-mmctx_data(0x404150, 3)
-mmctx_data(0x404164, 2)
-mmctx_data(0x404178, 2)
-mmctx_data(0x404200, 8)
-mmctx_data(0x404404, 14)
-mmctx_data(0x404460, 4)
-mmctx_data(0x404480, 1)
-mmctx_data(0x404498, 1)
-mmctx_data(0x404604, 4)
-mmctx_data(0x404618, 32)
-mmctx_data(0x404698, 21)
-mmctx_data(0x4046f0, 2)
-mmctx_data(0x404700, 22)
-mmctx_data(0x405800, 1)
-mmctx_data(0x405830, 3)
-mmctx_data(0x405854, 1)
-mmctx_data(0x405870, 4)
-mmctx_data(0x405a00, 2)
-mmctx_data(0x405a18, 1)
-mmctx_data(0x406020, 1)
-mmctx_data(0x406028, 4)
-mmctx_data(0x4064a8, 2)
-mmctx_data(0x4064b4, 5)
-mmctx_data(0x407804, 1)
-mmctx_data(0x40780c, 6)
-mmctx_data(0x4078bc, 1)
-mmctx_data(0x408000, 7)
-mmctx_data(0x408064, 1)
-mmctx_data(0x408800, 3)
-mmctx_data(0x408900, 4)
-mmctx_data(0x408980, 1)
-nvd9_hub_mmio_tail:
-
-.align 256
-chan_data:
-chan_mmio_count:	.b32 0
-chan_mmio_address:	.b32 0
-
-.align 256
-xfer_data: 		.b32 0
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "hub.fuc"
+#undef INCLUDE_DATA
 
 .section #nvc0_grhub_code
+#define INCLUDE_CODE
 bra #init
-define(`include_code')
-include(`nvc0.fuc')
-
-// reports an exception to the host
-//
-// In: $r15 error code (see nvc0.fuc)
-//
-error:
-	push $r14
-	mov $r14 0x814
-	shl b32 $r14 6
-	iowr I[$r14 + 0x000] $r15	// CC_SCRATCH[5] = error code
-	mov $r14 0xc1c
-	shl b32 $r14 6
-	mov $r15 1
-	iowr I[$r14 + 0x000] $r15	// INTR_UP_SET
-	pop $r14
-	ret
-
-// HUB fuc initialisation, executed by triggering ucode start, will
-// fall through to main loop after completion.
-//
-// Input:
-//   CC_SCRATCH[0]: chipset (PMC_BOOT_0 read returns 0x0bad0bad... sigh)
-//
-// Output:
-//   CC_SCRATCH[0]:
-//	     31:31: set to signal completion
-//   CC_SCRATCH[1]:
-//	      31:0: total PGRAPH context size
-//
-init:
-	clear b32 $r0
-	mov $sp $r0
-	mov $xdbase $r0
-
-	// enable fifo access
-	mov $r1 0x1200
-	mov $r2 2
-	iowr I[$r1 + 0x000] $r2	// FIFO_ENABLE
-
-	// setup i0 handler, and route all interrupts to it
-	mov $r1 #ih
-	mov $iv0 $r1
-	mov $r1 0x400
-	iowr I[$r1 + 0x300] $r0	// INTR_DISPATCH
-
-	// route HUB_CHANNEL_SWITCH to fuc interrupt 8
-	mov $r3 0x404
-	shl b32 $r3 6
-	mov $r2 0x2003		// { HUB_CHANNEL_SWITCH, ZERO } -> intr 8
-	iowr I[$r3 + 0x000] $r2
-
-	// not sure what these are, route them because NVIDIA does, and
-	// the IRQ handler will signal the host if we ever get one.. we
-	// may find out if/why we need to handle these if so..
-	//
-	mov $r2 0x2004
-	iowr I[$r3 + 0x004] $r2 // { 0x04, ZERO } -> intr 9
-	mov $r2 0x200b
-	iowr I[$r3 + 0x008] $r2 // { 0x0b, ZERO } -> intr 10
-	mov $r2 0x200c
-	iowr I[$r3 + 0x01c] $r2 // { 0x0c, ZERO } -> intr 15
-
-	// enable all INTR_UP interrupts
-	mov $r2 0xc24
-	shl b32 $r2 6
-	not b32 $r3 $r0
-	iowr I[$r2] $r3
-
-	// enable fifo, ctxsw, 9, 10, 15 interrupts
-	mov $r2 -0x78fc		// 0x8704
-	sethi $r2 0
-	iowr I[$r1 + 0x000] $r2	// INTR_EN_SET
-
-	// fifo level triggered, rest edge
-	sub b32 $r1 0x100
-	mov $r2 4
-	iowr I[$r1] $r2
-
-	// enable interrupts
-	bset $flags ie0
-
-	// fetch enabled GPC/ROP counts
-	mov $r14 -0x69fc	// 0x409604
-	sethi $r14 0x400000
-	call #nv_rd32
-	extr $r1 $r15 16:20
-	st b32 D[$r0 + #rop_count] $r1
-	and $r15 0x1f
-	st b32 D[$r0 + #gpc_count] $r15
-
-	// set BAR_REQMASK to GPC mask
-	mov $r1 1
-	shl b32 $r1 $r15
-	sub b32 $r1 1
-	mov $r2 0x40c
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r1
-	iowr I[$r2 + 0x100] $r1
-
-	// find context data for this chipset
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x000]		// CC_SCRATCH[0]
-	mov $r15 #chipsets - 8
-	init_find_chipset:
-		add b32 $r15 8
-		ld b32 $r3 D[$r15 + 0x00]
-		cmpu b32 $r3 $r2
-		bra e #init_context
-		cmpu b32 $r3 0
-		bra ne #init_find_chipset
-		// unknown chipset
-		ret
-
-	// context size calculation, reserve first 256 bytes for use by fuc
-	init_context:
-	mov $r1 256
-
-	// calculate size of mmio context data
-	ld b16 $r14 D[$r15 + 4]
-	ld b16 $r15 D[$r15 + 6]
-	sethi $r14 0
-	st b32 D[$r0 + #hub_mmio_list_head] $r14
-	st b32 D[$r0 + #hub_mmio_list_tail] $r15
-	call #mmctx_size
-
-	// set mmctx base addresses now so we don't have to do it later,
-	// they don't (currently) ever change
-	mov $r3 0x700
-	shl b32 $r3 6
-	shr b32 $r4 $r1 8
-	iowr I[$r3 + 0x000] $r4		// MMCTX_SAVE_SWBASE
-	iowr I[$r3 + 0x100] $r4		// MMCTX_LOAD_SWBASE
-	add b32 $r3 0x1300
-	add b32 $r1 $r15
-	shr b32 $r15 2
-	iowr I[$r3 + 0x000] $r15	// MMCTX_LOAD_COUNT, wtf for?!?
-
-	// strands, base offset needs to be aligned to 256 bytes
-	shr b32 $r1 8
-	add b32 $r1 1
-	shl b32 $r1 8
-	mov b32 $r15 $r1
-	call #strand_ctx_init
-	add b32 $r1 $r15
-
-	// initialise each GPC in sequence by passing in the offset of its
-	// context data in GPCn_CC_SCRATCH[1], and starting its FUC (which
-	// has previously been uploaded by the host) running.
-	//
-	// the GPC fuc init sequence will set GPCn_CC_SCRATCH[0] bit 31
-	// when it has completed, and return the size of its context data
-	// in GPCn_CC_SCRATCH[1]
-	//
-	ld b32 $r3 D[$r0 + #gpc_count]
-	mov $r4 0x2000
-	sethi $r4 0x500000
-	init_gpc:
-		// setup, and start GPC ucode running
-		add b32 $r14 $r4 0x804
-		mov b32 $r15 $r1
-		call #nv_wr32			// CC_SCRATCH[1] = ctx offset
-		add b32 $r14 $r4 0x800
-		mov b32 $r15 $r2
-		call #nv_wr32			// CC_SCRATCH[0] = chipset
-		add b32 $r14 $r4 0x10c
-		clear b32 $r15
-		call #nv_wr32
-		add b32 $r14 $r4 0x104
-		call #nv_wr32			// ENTRY
-		add b32 $r14 $r4 0x100
-		mov $r15 2			// CTRL_START_TRIGGER
-		call #nv_wr32			// CTRL
-
-		// wait for it to complete, and adjust context size
-		add b32 $r14 $r4 0x800
-		init_gpc_wait:
-			call #nv_rd32
-			xbit $r15 $r15 31
-			bra e #init_gpc_wait
-		add b32 $r14 $r4 0x804
-		call #nv_rd32
-		add b32 $r1 $r15
-
-		// next!
-		add b32 $r4 0x8000
-		sub b32 $r3 1
-		bra ne #init_gpc
-
-	// save context size, and tell host we're ready
-	mov $r2 0x800
-	shl b32 $r2 6
-	iowr I[$r2 + 0x100] $r1		// CC_SCRATCH[1]  = context size
-	add b32 $r2 0x800
-	clear b32 $r1
-	bset $r1 31
-	iowr I[$r2 + 0x000] $r1		// CC_SCRATCH[0] |= 0x80000000
-
-// Main program loop, very simple, sleeps until woken up by the interrupt
-// handler, pulls a command from the queue and executes its handler
-//
-main:
-	// sleep until we have something to do
-	bset $flags $p0
-	sleep $p0
-	mov $r13 #cmd_queue
-	call #queue_get
-	bra $p1 #main
-
-	// context switch, requested by GPU?
-	cmpu b32 $r14 0x4001
-	bra ne #main_not_ctx_switch
-		trace_set(T_AUTO)
-		mov $r1 0xb00
-		shl b32 $r1 6
-		iord $r2 I[$r1 + 0x100]		// CHAN_NEXT
-		iord $r1 I[$r1 + 0x000]		// CHAN_CUR
-
-		xbit $r3 $r1 31
-		bra e #chsw_no_prev
-			xbit $r3 $r2 31
-			bra e #chsw_prev_no_next
-				push $r2
-				mov b32 $r2 $r1
-				trace_set(T_SAVE)
-				bclr $flags $p1
-				bset $flags $p2
-				call #ctx_xfer
-				trace_clr(T_SAVE);
-				pop $r2
-				trace_set(T_LOAD);
-				bset $flags $p1
-				call #ctx_xfer
-				trace_clr(T_LOAD);
-				bra #chsw_done
-			chsw_prev_no_next:
-				push $r2
-				mov b32 $r2 $r1
-				bclr $flags $p1
-				bclr $flags $p2
-				call #ctx_xfer
-				pop $r2
-				mov $r1 0xb00
-				shl b32 $r1 6
-				iowr I[$r1] $r2
-				bra #chsw_done
-		chsw_no_prev:
-			xbit $r3 $r2 31
-			bra e #chsw_done
-				bset $flags $p1
-				bclr $flags $p2
-				call #ctx_xfer
-
-		// ack the context switch request
-		chsw_done:
-		mov $r1 0xb0c
-		shl b32 $r1 6
-		mov $r2 1
-		iowr I[$r1 + 0x000] $r2		// 0x409b0c
-		trace_clr(T_AUTO)
-		bra #main
-
-	// request to set current channel? (*not* a context switch)
-	main_not_ctx_switch:
-	cmpu b32 $r14 0x0001
-	bra ne #main_not_ctx_chan
-		mov b32 $r2 $r15
-		call #ctx_chan
-		bra #main_done
-
-	// request to store current channel context?
-	main_not_ctx_chan:
-	cmpu b32 $r14 0x0002
-	bra ne #main_not_ctx_save
-		trace_set(T_SAVE)
-		bclr $flags $p1
-		bclr $flags $p2
-		call #ctx_xfer
-		trace_clr(T_SAVE)
-		bra #main_done
-
-	main_not_ctx_save:
-		shl b32 $r15 $r14 16
-		or $r15 E_BAD_COMMAND
-		call #error
-		bra #main
-
-	main_done:
-	mov $r1 0x820
-	shl b32 $r1 6
-	clear b32 $r2
-	bset $r2 31
-	iowr I[$r1 + 0x000] $r2		// CC_SCRATCH[0] |= 0x80000000
-	bra #main
-
-// interrupt handler
-ih:
-	push $r8
-	mov $r8 $flags
-	push $r8
-	push $r9
-	push $r10
-	push $r11
-	push $r13
-	push $r14
-	push $r15
-
-	// incoming fifo command?
-	iord $r10 I[$r0 + 0x200]	// INTR
-	and $r11 $r10 0x00000004
-	bra e #ih_no_fifo
-		// queue incoming fifo command for later processing
-		mov $r11 0x1900
-		mov $r13 #cmd_queue
-		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
-		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
-		call #queue_put
-		add b32 $r11 0x400
-		mov $r14 1
-		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
-
-	// context switch request?
-	ih_no_fifo:
-	and $r11 $r10 0x00000100
-	bra e #ih_no_ctxsw
-		// enqueue a context switch for later processing
-		mov $r13 #cmd_queue
-		mov $r14 0x4001
-		call #queue_put
-
-	// anything we didn't handle, bring it to the host's attention
-	ih_no_ctxsw:
-	mov $r11 0x104
-	not b32 $r11
-	and $r11 $r10 $r11
-	bra e #ih_no_other
-		mov $r10 0xc1c
-		shl b32 $r10 6
-		iowr I[$r10] $r11	// INTR_UP_SET
-
-	// ack, and wake up main()
-	ih_no_other:
-	iowr I[$r0 + 0x100] $r10	// INTR_ACK
-
-	pop $r15
-	pop $r14
-	pop $r13
-	pop $r11
-	pop $r10
-	pop $r9
-	pop $r8
-	mov $flags $r8
-	pop $r8
-	bclr $flags $p0
-	iret
-
-// Not real sure, but, MEM_CMD 7 will hang forever if this isn't done
-ctx_4160s:
-	mov $r14 0x4160
-	sethi $r14 0x400000
-	mov $r15 1
-	call #nv_wr32
-	ctx_4160s_wait:
-		call #nv_rd32
-		xbit $r15 $r15 4
-		bra e #ctx_4160s_wait
-	ret
-
-// Without clearing again at end of xfer, some things cause PGRAPH
-// to hang with STATUS=0x00000007 until it's cleared.. fbcon can
-// still function with it set however...
-ctx_4160c:
-	mov $r14 0x4160
-	sethi $r14 0x400000
-	clear b32 $r15
-	call #nv_wr32
-	ret
-
-// Again, not real sure
-//
-// In: $r15 value to set 0x404170 to
-//
-ctx_4170s:
-	mov $r14 0x4170
-	sethi $r14 0x400000
-	or $r15 0x10
-	call #nv_wr32
-	ret
-
-// Waits for a ctx_4170s() call to complete
-//
-ctx_4170w:
-	mov $r14 0x4170
-	sethi $r14 0x400000
-	call #nv_rd32
-	and $r15 0x10
-	bra ne #ctx_4170w
-	ret
-
-// Disables various things, waits a bit, and re-enables them..
-//
-// Not sure how exactly this helps, perhaps "ENABLE" is not such a
-// good description for the bits we turn off?  Anyways, without this,
-// funny things happen.
-//
-ctx_redswitch:
-	mov $r14 0x614
-	shl b32 $r14 6
-	mov $r15 0x270
-	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_GPC, POWER_ALL
-	mov $r15 8
-	ctx_redswitch_delay:
-		sub b32 $r15 1
-		bra ne #ctx_redswitch_delay
-	mov $r15 0x770
-	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_ALL, POWER_ALL
-	ret
-
-// Not a clue what this is for, except that unless the value is 0x10, the
-// strand context is saved (and presumably restored) incorrectly..
-//
-// In: $r15 value to set to (0x00/0x10 are used)
-//
-ctx_86c:
-	mov $r14 0x86c
-	shl b32 $r14 6
-	iowr I[$r14] $r15	// HUB(0x86c) = val
-	mov $r14 -0x75ec
-	sethi $r14 0x400000
-	call #nv_wr32		// ROP(0xa14) = val
-	mov $r14 -0x5794
-	sethi $r14 0x410000
-	call #nv_wr32		// GPC(0x86c) = val
-	ret
-
-// ctx_load - load's a channel's ctxctl data, and selects its vm
-//
-// In: $r2 channel address
-//
-ctx_load:
-	trace_set(T_CHAN)
-
-	// switch to channel, somewhat magic in parts..
-	mov $r10 12		// DONE_UNK12
-	call #wait_donez
-	mov $r1 0xa24
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r0	// 0x409a24
-	mov $r3 0xb00
-	shl b32 $r3 6
-	iowr I[$r3 + 0x100] $r2	// CHAN_NEXT
-	mov $r1 0xa0c
-	shl b32 $r1 6
-	mov $r4 7
-	iowr I[$r1 + 0x000] $r2 // MEM_CHAN
-	iowr I[$r1 + 0x100] $r4	// MEM_CMD
-	ctx_chan_wait_0:
-		iord $r4 I[$r1 + 0x100]
-		and $r4 0x1f
-		bra ne #ctx_chan_wait_0
-	iowr I[$r3 + 0x000] $r2	// CHAN_CUR
-
-	// load channel header, fetch PGRAPH context pointer
-	mov $xtargets $r0
-	bclr $r2 31
-	shl b32 $r2 4
-	add b32 $r2 2
-
-	trace_set(T_LCHAN)
-	mov $r1 0xa04
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r2		// MEM_BASE
-	mov $r1 0xa20
-	shl b32 $r1 6
-	mov $r2 0x0002
-	sethi $r2 0x80000000
-	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vram
-	mov $r1 0x10			// chan + 0x0210
-	mov $r2 #xfer_data
-	sethi $r2 0x00020000		// 16 bytes
-	xdld $r1 $r2
-	xdwait
-	trace_clr(T_LCHAN)
-
-	// update current context
-	ld b32 $r1 D[$r0 + #xfer_data + 4]
-	shl b32 $r1 24
-	ld b32 $r2 D[$r0 + #xfer_data + 0]
-	shr b32 $r2 8
-	or $r1 $r2
-	st b32 D[$r0 + #ctx_current] $r1
-
-	// set transfer base to start of context, and fetch context header
-	trace_set(T_LCTXH)
-	mov $r2 0xa04
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r1		// MEM_BASE
-	mov $r2 1
-	mov $r1 0xa20
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vm
-	mov $r1 #chan_data
-	sethi $r1 0x00060000		// 256 bytes
-	xdld $r0 $r1
-	xdwait
-	trace_clr(T_LCTXH)
-
-	trace_clr(T_CHAN)
-	ret
-
-// ctx_chan - handler for HUB_SET_CHAN command, will set a channel as
-//            the active channel for ctxctl, but not actually transfer
-//            any context data.  intended for use only during initial
-//            context construction.
-//
-// In: $r2 channel address
-//
-ctx_chan:
-	call #ctx_4160s
-	call #ctx_load
-	mov $r10 12			// DONE_UNK12
-	call #wait_donez
-	mov $r1 0xa10
-	shl b32 $r1 6
-	mov $r2 5
-	iowr I[$r1 + 0x000] $r2		// MEM_CMD = 5 (???)
-	ctx_chan_wait:
-		iord $r2 I[$r1 + 0x000]
-		or $r2 $r2
-		bra ne #ctx_chan_wait
-	call #ctx_4160c
-	ret
-
-// Execute per-context state overrides list
-//
-// Only executed on the first load of a channel.  Might want to look into
-// removing this and having the host directly modify the channel's context
-// to change this state...  The nouveau DRM already builds this list as
-// it's definitely needed for NVIDIA's, so we may as well use it for now
-//
-// Input: $r1 mmio list length
-//
-ctx_mmio_exec:
-	// set transfer base to be the mmio list
-	ld b32 $r3 D[$r0 + #chan_mmio_address]
-	mov $r2 0xa04
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r3		// MEM_BASE
-
-	clear b32 $r3
-	ctx_mmio_loop:
-		// fetch next 256 bytes of mmio list if necessary
-		and $r4 $r3 0xff
-		bra ne #ctx_mmio_pull
-			mov $r5 #xfer_data
-			sethi $r5 0x00060000	// 256 bytes
-			xdld $r3 $r5
-			xdwait
-
-		// execute a single list entry
-		ctx_mmio_pull:
-		ld b32 $r14 D[$r4 + #xfer_data + 0x00]
-		ld b32 $r15 D[$r4 + #xfer_data + 0x04]
-		call #nv_wr32
-
-		// next!
-		add b32 $r3 8
-		sub b32 $r1 1
-		bra ne #ctx_mmio_loop
-
-	// set transfer base back to the current context
-	ctx_mmio_done:
-	ld b32 $r3 D[$r0 + #ctx_current]
-	iowr I[$r2 + 0x000] $r3		// MEM_BASE
-
-	// disable the mmio list now, we don't need/want to execute it again
-	st b32 D[$r0 + #chan_mmio_count] $r0
-	mov $r1 #chan_data
-	sethi $r1 0x00060000		// 256 bytes
-	xdst $r0 $r1
-	xdwait
-	ret
-
-// Transfer HUB context data between GPU and storage area
-//
-// In: $r2 channel address
-//     $p1 clear on save, set on load
-//     $p2 set if opposite direction done/will be done, so:
-//		on save it means: "a load will follow this save"
-//		on load it means: "a save preceeded this load"
-//
-ctx_xfer:
-	// according to mwk, some kind of wait for idle
-	mov $r15 0xc00
-	shl b32 $r15 6
-	mov $r14 4
-	iowr I[$r15 + 0x200] $r14
-	ctx_xfer_idle:
-		iord $r14 I[$r15 + 0x000]
-		and $r14 0x2000
-		bra ne #ctx_xfer_idle
-
-	bra not $p1 #ctx_xfer_pre
-	bra $p2 #ctx_xfer_pre_load
-	ctx_xfer_pre:
-		mov $r15 0x10
-		call #ctx_86c
-		call #ctx_4160s
-		bra not $p1 #ctx_xfer_exec
-
-	ctx_xfer_pre_load:
-		mov $r15 2
-		call #ctx_4170s
-		call #ctx_4170w
-		call #ctx_redswitch
-		clear b32 $r15
-		call #ctx_4170s
-		call #ctx_load
-
-	// fetch context pointer, and initiate xfer on all GPCs
-	ctx_xfer_exec:
-	ld b32 $r1 D[$r0 + #ctx_current]
-	mov $r2 0x414
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r0	// BAR_STATUS = reset
-	mov $r14 -0x5b00
-	sethi $r14 0x410000
-	mov b32 $r15 $r1
-	call #nv_wr32		// GPC_BCAST_WRCMD_DATA = ctx pointer
-	add b32 $r14 4
-	xbit $r15 $flags $p1
-	xbit $r2 $flags $p2
-	shl b32 $r2 1
-	or $r15 $r2
-	call #nv_wr32		// GPC_BCAST_WRCMD_CMD = GPC_XFER(type)
-
-	// strands
-	mov $r1 0x4afc
-	sethi $r1 0x20000
-	mov $r2 0xc
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
-	call #strand_wait
-	mov $r2 0x47fc
-	sethi $r2 0x20000
-	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
-	xbit $r2 $flags $p1
-	add b32 $r2 3
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
-
-	// mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 6		// first, last
-	mov $r11 0		// base = 0
-	ld b32 $r12 D[$r0 + #hub_mmio_list_head]
-	ld b32 $r13 D[$r0 + #hub_mmio_list_tail]
-	mov $r14 0		// not multi
-	call #mmctx_xfer
-
-	// wait for GPCs to all complete
-	mov $r10 8		// DONE_BAR
-	call #wait_doneo
-
-	// wait for strand xfer to complete
-	call #strand_wait
-
-	// post-op
-	bra $p1 #ctx_xfer_post
-		mov $r10 12		// DONE_UNK12
-		call #wait_donez
-		mov $r1 0xa10
-		shl b32 $r1 6
-		mov $r2 5
-		iowr I[$r1] $r2		// MEM_CMD
-		ctx_xfer_post_save_wait:
-			iord $r2 I[$r1]
-			or $r2 $r2
-			bra ne #ctx_xfer_post_save_wait
-
-	bra $p2 #ctx_xfer_done
-	ctx_xfer_post:
-		mov $r15 2
-		call #ctx_4170s
-		clear b32 $r15
-		call #ctx_86c
-		call #strand_post
-		call #ctx_4170w
-		clear b32 $r15
-		call #ctx_4170s
-
-		bra not $p1 #ctx_xfer_no_post_mmio
-		ld b32 $r1 D[$r0 + #chan_mmio_count]
-		or $r1 $r1
-		bra e #ctx_xfer_no_post_mmio
-			call #ctx_mmio_exec
-
-		ctx_xfer_no_post_mmio:
-		call #ctx_4160c
-
-	ctx_xfer_done:
-	ret
-
+#include "com.fuc"
+#include "hub.fuc"
 .align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc.h
index bb03d2a1d57b..b59f694c0423 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvc0.fuc.h
@@ -1,9 +1,90 @@
 uint32_t nvc0_grhub_data[] = {
-/* 0x0000: gpc_count */
+/* 0x0000: hub_mmio_list_head */
+	0x00000300,
+/* 0x0004: hub_mmio_list_tail */
+	0x00000304,
+/* 0x0008: gpc_count */
+	0x00000000,
+/* 0x000c: rop_count */
+	0x00000000,
+/* 0x0010: cmd_queue */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0058: ctx_current */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0100: chan_data */
+/* 0x0100: chan_mmio_count */
+	0x00000000,
+/* 0x0104: chan_mmio_address */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
-/* 0x0004: rop_count */
 	0x00000000,
-/* 0x0008: cmd_queue */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -22,114 +103,9 @@ uint32_t nvc0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0050: hub_mmio_list_head */
 	0x00000000,
-/* 0x0054: hub_mmio_list_tail */
 	0x00000000,
-/* 0x0058: ctx_current */
 	0x00000000,
-/* 0x005c: chipsets */
-	0x000000c0,
-	0x013c00a0,
-	0x000000c1,
-	0x014000a0,
-	0x000000c3,
-	0x013c00a0,
-	0x000000c4,
-	0x013c00a0,
-	0x000000c8,
-	0x013c00a0,
-	0x000000ce,
-	0x013c00a0,
-	0x000000cf,
-	0x013c00a0,
-	0x000000d9,
-	0x01dc0140,
-	0x00000000,
-/* 0x00a0: nvc0_hub_mmio_head */
-	0x0417e91c,
-	0x04400204,
-	0x28404004,
-	0x00404044,
-	0x34404094,
-	0x184040d0,
-	0x004040f8,
-	0x08404130,
-	0x08404150,
-	0x04404164,
-	0x08404174,
-	0x1c404200,
-	0x34404404,
-	0x0c404460,
-	0x00404480,
-	0x00404498,
-	0x0c404604,
-	0x7c404618,
-	0x50404698,
-	0x044046f0,
-	0x54404700,
-	0x00405800,
-	0x08405830,
-	0x00405854,
-	0x0c405870,
-	0x04405a00,
-	0x00405a18,
-	0x00406020,
-	0x0c406028,
-	0x044064a8,
-	0x044064b4,
-	0x00407804,
-	0x1440780c,
-	0x004078bc,
-	0x18408000,
-	0x00408064,
-	0x08408800,
-	0x0c408900,
-	0x00408980,
-/* 0x013c: nvc0_hub_mmio_tail */
-	0x044064c0,
-/* 0x0140: nvc1_hub_mmio_tail */
-/* 0x0140: nvd9_hub_mmio_head */
-	0x0417e91c,
-	0x04400204,
-	0x24404004,
-	0x00404044,
-	0x34404094,
-	0x184040d0,
-	0x004040f8,
-	0x08404130,
-	0x08404150,
-	0x04404164,
-	0x04404178,
-	0x1c404200,
-	0x34404404,
-	0x0c404460,
-	0x00404480,
-	0x00404498,
-	0x0c404604,
-	0x7c404618,
-	0x50404698,
-	0x044046f0,
-	0x54404700,
-	0x00405800,
-	0x08405830,
-	0x00405854,
-	0x0c405870,
-	0x04405a00,
-	0x00405a18,
-	0x00406020,
-	0x0c406028,
-	0x044064a8,
-	0x104064b4,
-	0x00407804,
-	0x1440780c,
-	0x004078bc,
-	0x18408000,
-	0x00408064,
-	0x08408800,
-	0x0c408900,
-	0x00408980,
-/* 0x01dc: nvd9_hub_mmio_tail */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -139,10 +115,7 @@ uint32_t nvc0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0200: chan_data */
-/* 0x0200: chan_mmio_count */
 	0x00000000,
-/* 0x0204: chan_mmio_address */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -163,6 +136,7 @@ uint32_t nvc0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+/* 0x0200: xfer_data */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -206,19 +180,40 @@ uint32_t nvc0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0300: xfer_data */
 	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0300: hub_mmio_list_base */
+	0x0417e91c,
 };
 
 uint32_t nvc0_grhub_code[] = {
-	0x03090ef5,
+	0x031b0ef5,
 /* 0x0004: queue_put */
 	0x9800d898,
 	0x86f001d9,
 	0x0489b808,
 	0xf00c1bf4,
 	0x21f502f7,
-	0x00f802ec,
+	0x00f802fe,
 /* 0x001c: queue_put_next */
 	0xb60798c4,
 	0x8dbb0384,
@@ -250,7 +245,7 @@ uint32_t nvc0_grhub_code[] = {
 	0xc800bccf,
 	0x1bf41fcc,
 	0x06a7f0fa,
-	0x010321f5,
+	0x010921f5,
 	0xf840bfcf,
 /* 0x008d: nv_wr32 */
 	0x28b7f100,
@@ -272,63 +267,66 @@ uint32_t nvc0_grhub_code[] = {
 	0x0684b604,
 	0xf80080d0,
 /* 0x00c9: wait_donez */
-	0x3c87f100,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d000,
-	0x081887f1,
-	0xd00684b6,
-/* 0x00e2: wait_done_wait_donez */
-	0x87f1008a,
-	0x84b60400,
-	0x0088cf06,
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
 	0xf4888aff,
-	0x87f1f31b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00099,
-/* 0x0103: wait_doneo */
-	0xf100f800,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00099f0,
-	0x87f10089,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x87f104bd,
 	0x84b60818,
 	0x008ad006,
-/* 0x011c: wait_done_wait_doneo */
+/* 0x0124: wait_doneo_e */
 	0x040087f1,
 	0xcf0684b6,
 	0x8aff0088,
 	0xf30bf488,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0099f094,
-	0xf80089d0,
-/* 0x013d: mmctx_size */
-/* 0x013f: nv_mmctx_size_loop */
-	0x9894bd00,
-	0x85b600e8,
-	0x0180b61a,
-	0xbb0284b6,
-	0xe0b60098,
-	0x04efb804,
-	0xb9eb1bf4,
-	0x00f8029f,
-/* 0x015c: mmctx_xfer */
-	0x083c87f1,
-	0xbd0684b6,
-	0x0199f094,
-	0xf10089d0,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf104bd00,
 	0xb6071087,
 	0x94bd0684,
 	0xf405bbfd,
 	0x8bd0090b,
 	0x0099f000,
-/* 0x0180: mmctx_base_disabled */
+/* 0x018c: mmctx_base_disabled */
 	0xf405eefd,
 	0x8ed00c0b,
 	0xc08fd080,
-/* 0x018f: mmctx_multi_disabled */
+/* 0x019b: mmctx_multi_disabled */
 	0xb70199f0,
 	0xc8010080,
 	0xb4b600ab,
@@ -336,8 +334,8 @@ uint32_t nvc0_grhub_code[] = {
 	0xb601aec8,
 	0xbefd11e4,
 	0x008bd005,
-/* 0x01a8: mmctx_exec_loop */
-/* 0x01a8: mmctx_wait_free */
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
 	0xf0008ecf,
 	0x0bf41fe4,
 	0x00ce98fa,
@@ -346,76 +344,77 @@ uint32_t nvc0_grhub_code[] = {
 	0x04cdb804,
 	0xc8e81bf4,
 	0x1bf402ab,
-/* 0x01c9: mmctx_fini_wait */
+/* 0x01d5: mmctx_fini_wait */
 	0x008bcf18,
 	0xb01fb4f0,
 	0x1bf410b4,
 	0x02a7f0f7,
 	0xf4c921f4,
-/* 0x01de: mmctx_stop */
+/* 0x01ea: mmctx_stop */
 	0xabc81b0e,
 	0x10b4b600,
 	0xf00cb9f0,
 	0x8bd012b9,
-/* 0x01ed: mmctx_stop_wait */
+/* 0x01f9: mmctx_stop_wait */
 	0x008bcf00,
 	0xf412bbc8,
-/* 0x01f6: mmctx_done */
-	0x87f1fa1b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00199,
-/* 0x0207: strand_wait */
-	0xf900f800,
-	0x02a7f0a0,
-	0xfcc921f4,
-/* 0x0213: strand_pre */
-	0xf100f8a0,
-	0xf04afc87,
-	0x97f00283,
-	0x0089d00c,
-	0x020721f5,
-/* 0x0226: strand_post */
-	0x87f100f8,
-	0x83f04afc,
-	0x0d97f002,
-	0xf50089d0,
-	0xf8020721,
-/* 0x0239: strand_set */
-	0xfca7f100,
-	0x02a3f04f,
-	0x0500aba2,
-	0xd00fc7f0,
-	0xc7f000ac,
-	0x00bcd00b,
-	0x020721f5,
-	0xf000aed0,
-	0xbcd00ac7,
-	0x0721f500,
-/* 0x0263: strand_ctx_init */
-	0xf100f802,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00399f0,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
 	0x21f50089,
-	0xe7f00213,
-	0x3921f503,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f00f00,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
 	0xfca7f102,
 	0x02a3f046,
 	0x0400aba0,
 	0xf040a0d0,
 	0xbcd001c7,
-	0x0721f500,
+	0x1521f500,
 	0x010c9202,
 	0xf000acd0,
 	0xbcd002c7,
-	0x0721f500,
-	0x2621f502,
+	0x1521f500,
+	0x3421f502,
 	0x8087f102,
 	0x0684b608,
 	0xb70089cf,
 	0x95220080,
-/* 0x02ba: ctx_init_strand_loop */
+/* 0x02ca: ctx_init_strand_loop */
 	0x8ed008fe,
 	0x408ed000,
 	0xb6808acf,
@@ -424,73 +423,61 @@ uint32_t nvc0_grhub_code[] = {
 	0xb60480b6,
 	0x1bf40192,
 	0x08e4b6e8,
-	0xf1f2efbc,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00399f0,
-	0x00f80089,
-/* 0x02ec: error */
-	0xe7f1e0f9,
-	0xe4b60814,
-	0x00efd006,
-	0x0c1ce7f1,
-	0xf006e4b6,
-	0xefd001f7,
-	0xf8e0fc00,
-/* 0x0309: init */
-	0xfe04bd00,
-	0x07fe0004,
-	0x0017f100,
-	0x0227f012,
-	0xf10012d0,
-	0xfe05b917,
-	0x17f10010,
-	0x10d00400,
-	0x0437f1c0,
-	0x0634b604,
-	0x200327f1,
-	0xf10032d0,
-	0xd0200427,
-	0x27f10132,
-	0x32d0200b,
-	0x0c27f102,
-	0x0732d020,
-	0x0c2427f1,
-	0xb90624b6,
-	0x23d00003,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0x07f100f8,
+	0x03f00500,
+	0x000fd002,
+	0xf7f004bd,
+	0x0007f101,
+	0x0303f007,
+	0xbd000fd0,
+/* 0x031b: init */
+	0xbd00f804,
+	0x0004fe04,
+	0xf10007fe,
+	0xf0120017,
+	0x12d00227,
+	0xb117f100,
+	0x0010fe05,
+	0x040017f1,
+	0xf1c010d0,
+	0xb6040437,
+	0x27f10634,
+	0x32d02003,
 	0x0427f100,
-	0x0023f087,
-	0xb70012d0,
-	0xf0010012,
-	0x12d00427,
-	0x1031f400,
-	0x9604e7f1,
-	0xf440e3f0,
-	0xf1c76821,
-	0x01018090,
-	0x801ff4f0,
-	0x17f0000f,
-	0x041fbb01,
-	0xf10112b6,
-	0xb6040c27,
-	0x21d00624,
-	0x4021d000,
-	0x080027f1,
-	0xcf0624b6,
-	0xf7f00022,
-/* 0x03a9: init_find_chipset */
-	0x08f0b654,
-	0xb800f398,
-	0x0bf40432,
-	0x0034b00b,
-	0xf8f11bf4,
-/* 0x03bd: init_context */
-	0x0017f100,
-	0x02fe5801,
-	0xf003ff58,
-	0x0e8000e3,
-	0x150f8014,
-	0x013d21f5,
+	0x0132d020,
+	0x200b27f1,
+	0xf10232d0,
+	0xd0200c27,
+	0x27f10732,
+	0x24b60c24,
+	0x0003b906,
+	0xf10023d0,
+	0xf0870427,
+	0x12d00023,
+	0x0012b700,
+	0x0427f001,
+	0xf40012d0,
+	0xe7f11031,
+	0xe3f09604,
+	0x6821f440,
+	0x8090f1c7,
+	0xf4f00301,
+	0x020f801f,
+	0xbb0117f0,
+	0x12b6041f,
+	0x0c27f101,
+	0x0624b604,
+	0xd00021d0,
+	0x17f14021,
+	0x0e980100,
+	0x010f9800,
+	0x014721f5,
 	0x070037f1,
 	0x950634b6,
 	0x34d00814,
@@ -501,208 +488,213 @@ uint32_t nvc0_grhub_code[] = {
 	0x0815b600,
 	0xb60110b6,
 	0x1fb90814,
-	0x6321f502,
+	0x7121f502,
 	0x001fbb02,
-	0xf1000398,
+	0xf1020398,
 	0xf0200047,
-/* 0x040e: init_gpc */
+/* 0x03f6: init_gpc */
 	0x4ea05043,
 	0x1fb90804,
 	0x8d21f402,
-	0x08004ea0,
-	0xf4022fb9,
-	0x4ea08d21,
-	0xf4bd010c,
-	0xa08d21f4,
-	0xf401044e,
+	0x010c4ea0,
+	0x21f4f4bd,
+	0x044ea08d,
+	0x8d21f401,
+	0x01004ea0,
+	0xf402f7f0,
 	0x4ea08d21,
-	0xf7f00100,
-	0x8d21f402,
-	0x08004ea0,
-/* 0x0440: init_gpc_wait */
-	0xc86821f4,
-	0x0bf41fff,
-	0x044ea0fa,
-	0x6821f408,
-	0xb7001fbb,
-	0xb6800040,
-	0x1bf40132,
-	0x0027f1b4,
-	0x0624b608,
-	0xb74021d0,
-	0xbd080020,
+/* 0x041e: init_gpc_wait */
+	0x21f40800,
+	0x1fffc868,
+	0xa0fa0bf4,
+	0xf408044e,
+	0x1fbb6821,
+	0x0040b700,
+	0x0132b680,
+	0xf1be1bf4,
+	0xf0010007,
+	0x01d00203,
+	0xbd04bd00,
 	0x1f19f014,
-/* 0x0473: main */
-	0xf40021d0,
-	0x28f40031,
-	0x08d7f000,
-	0xf43921f4,
-	0xe4b1f401,
-	0x1bf54001,
-	0x87f100d1,
-	0x84b6083c,
-	0xf094bd06,
-	0x89d00499,
-	0x0017f100,
-	0x0614b60b,
-	0xcf4012cf,
-	0x13c80011,
-	0x7e0bf41f,
+	0x080007f1,
+	0xd00203f0,
+	0x04bd0001,
+/* 0x0458: main */
+	0xf40031f4,
+	0xd7f00028,
+	0x3921f410,
+	0xb1f401f4,
+	0xf54001e4,
+	0xbd00de1b,
+	0x0499f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0b0017f1,
+	0xcf0614b6,
+	0x11cf4012,
+	0x1f13c800,
+	0x00870bf5,
 	0xf41f23c8,
-	0x20f95a0b,
-	0xf10212b9,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00799f0,
-	0x32f40089,
-	0x0231f401,
-	0x082921f5,
-	0x085c87f1,
-	0xbd0684b6,
+	0x20f9620b,
+	0xbd0212b9,
 	0x0799f094,
-	0xfc0089d0,
-	0x3c87f120,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d006,
-	0xf50131f4,
-	0xf1082921,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00699f0,
-	0x0ef40089,
-/* 0x0509: chsw_prev_no_next */
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0xf40132f4,
+	0x21f50231,
+	0x94bd082f,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xfc04bd00,
+	0xf094bd20,
+	0x07f10699,
+	0x03f00f00,
+	0x0009d002,
+	0x31f404bd,
+	0x2f21f501,
+	0xf094bd08,
+	0x07f10699,
+	0x03f01700,
+	0x0009d002,
+	0x0ef404bd,
+/* 0x04f9: chsw_prev_no_next */
 	0xb920f931,
 	0x32f40212,
 	0x0232f401,
-	0x082921f5,
+	0x082f21f5,
 	0x17f120fc,
 	0x14b60b00,
 	0x0012d006,
-/* 0x0527: chsw_no_prev */
+/* 0x0517: chsw_no_prev */
 	0xc8130ef4,
 	0x0bf41f23,
 	0x0131f40d,
 	0xf50232f4,
-/* 0x0537: chsw_done */
-	0xf1082921,
+/* 0x0527: chsw_done */
+	0xf1082f21,
 	0xb60b0c17,
 	0x27f00614,
 	0x0012d001,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0499f094,
-	0xf50089d0,
-/* 0x0557: main_not_ctx_switch */
-	0xb0ff200e,
-	0x1bf401e4,
-	0x02f2b90d,
-	0x07b521f5,
-/* 0x0567: main_not_ctx_chan */
-	0xb0420ef4,
-	0x1bf402e4,
-	0x3c87f12e,
-	0x0684b608,
 	0x99f094bd,
-	0x0089d007,
+	0x0007f104,
+	0x0203f017,
+	0xbd0009d0,
+	0x130ef504,
+/* 0x0549: main_not_ctx_switch */
+	0x01e4b0ff,
+	0xb90d1bf4,
+	0x21f502f2,
+	0x0ef407bb,
+/* 0x0559: main_not_ctx_chan */
+	0x02e4b046,
+	0xbd321bf4,
+	0x0799f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
 	0xf40132f4,
 	0x21f50232,
-	0x87f10829,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00799,
-	0x110ef400,
-/* 0x0598: main_not_ctx_save */
-	0xf010ef94,
-	0x21f501f5,
-	0x0ef502ec,
-/* 0x05a6: main_done */
-	0x17f1fed1,
-	0x14b60820,
-	0xf024bd06,
-	0x12d01f29,
-	0xbe0ef500,
-/* 0x05b9: ih */
+	0x94bd082f,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xf404bd00,
+/* 0x058e: main_not_ctx_save */
+	0xef94110e,
+	0x01f5f010,
+	0x02fe21f5,
+	0xfec00ef5,
+/* 0x059c: main_done */
+	0x29f024bd,
+	0x0007f11f,
+	0x0203f008,
+	0xbd0002d0,
+	0xab0ef504,
+/* 0x05b1: ih */
 	0xfe80f9fe,
 	0x80f90188,
 	0xa0f990f9,
 	0xd0f9b0f9,
 	0xf0f9e0f9,
-	0xc4800acf,
-	0x0bf404ab,
-	0x00b7f11d,
-	0x08d7f019,
-	0xcf40becf,
-	0x21f400bf,
-	0x00b0b704,
-	0x01e7f004,
-/* 0x05ef: ih_no_fifo */
-	0xe400bed0,
-	0xf40100ab,
-	0xd7f00d0b,
-	0x01e7f108,
-	0x0421f440,
-/* 0x0600: ih_no_ctxsw */
-	0x0104b7f1,
-	0xabffb0bd,
-	0x0d0bf4b4,
-	0x0c1ca7f1,
-	0xd006a4b6,
-/* 0x0616: ih_no_other */
-	0x0ad000ab,
-	0xfcf0fc40,
-	0xfcd0fce0,
-	0xfca0fcb0,
-	0xfe80fc90,
-	0x80fc0088,
-	0xf80032f4,
-/* 0x0631: ctx_4160s */
-	0x60e7f101,
-	0x40e3f041,
-	0xf401f7f0,
-/* 0x063e: ctx_4160s_wait */
-	0x21f48d21,
-	0x04ffc868,
-	0xf8fa0bf4,
-/* 0x0649: ctx_4160c */
-	0x60e7f100,
+	0x0acf04bd,
+	0x04abc480,
+	0xf11d0bf4,
+	0xf01900b7,
+	0xbecf10d7,
+	0x00bfcf40,
+	0xb70421f4,
+	0xf00400b0,
+	0xbed001e7,
+/* 0x05e9: ih_no_fifo */
+	0x00abe400,
+	0x0d0bf401,
+	0xf110d7f0,
+	0xf44001e7,
+/* 0x05fa: ih_no_ctxsw */
+	0xb7f10421,
+	0xb0bd0104,
+	0xf4b4abff,
+	0xa7f10d0b,
+	0xa4b60c1c,
+	0x00abd006,
+/* 0x0610: ih_no_other */
+	0xfc400ad0,
+	0xfce0fcf0,
+	0xfcb0fcd0,
+	0xfc90fca0,
+	0x0088fe80,
+	0x32f480fc,
+/* 0x062b: ctx_4160s */
+	0xf101f800,
+	0xf04160e7,
+	0xf7f040e3,
+	0x8d21f401,
+/* 0x0638: ctx_4160s_wait */
+	0xc86821f4,
+	0x0bf404ff,
+/* 0x0643: ctx_4160c */
+	0xf100f8fa,
+	0xf04160e7,
+	0xf4bd40e3,
+	0xf88d21f4,
+/* 0x0651: ctx_4170s */
+	0x70e7f100,
 	0x40e3f041,
-	0x21f4f4bd,
-/* 0x0657: ctx_4170s */
-	0xf100f88d,
-	0xf04170e7,
-	0xf5f040e3,
-	0x8d21f410,
-/* 0x0666: ctx_4170w */
-	0xe7f100f8,
-	0xe3f04170,
-	0x6821f440,
-	0xf410f4f0,
-	0x00f8f31b,
-/* 0x0678: ctx_redswitch */
-	0x0614e7f1,
-	0xf106e4b6,
-	0xd00270f7,
-	0xf7f000ef,
-/* 0x0689: ctx_redswitch_delay */
-	0x01f2b608,
-	0xf1fd1bf4,
-	0xd00770f7,
-	0x00f800ef,
-/* 0x0698: ctx_86c */
-	0x086ce7f1,
-	0xd006e4b6,
-	0xe7f100ef,
-	0xe3f08a14,
-	0x8d21f440,
-	0xa86ce7f1,
-	0xf441e3f0,
+	0xf410f5f0,
 	0x00f88d21,
-/* 0x06b8: ctx_load */
-	0x083c87f1,
-	0xbd0684b6,
-	0x0599f094,
-	0xf00089d0,
+/* 0x0660: ctx_4170w */
+	0x4170e7f1,
+	0xf440e3f0,
+	0xf4f06821,
+	0xf31bf410,
+/* 0x0672: ctx_redswitch */
+	0xe7f100f8,
+	0xe4b60614,
+	0x70f7f106,
+	0x00efd002,
+/* 0x0683: ctx_redswitch_delay */
+	0xb608f7f0,
+	0x1bf401f2,
+	0x70f7f1fd,
+	0x00efd007,
+/* 0x0692: ctx_86c */
+	0xe7f100f8,
+	0xe4b6086c,
+	0x00efd006,
+	0x8a14e7f1,
+	0xf440e3f0,
+	0xe7f18d21,
+	0xe3f0a86c,
+	0x8d21f441,
+/* 0x06b2: ctx_load */
+	0x94bd00f8,
+	0xf10599f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf004bd00,
 	0x21f40ca7,
 	0x2417f1c9,
 	0x0614b60a,
@@ -713,168 +705,169 @@ uint32_t nvc0_grhub_code[] = {
 	0x0614b60a,
 	0xd00747f0,
 	0x14d00012,
-/* 0x06f1: ctx_chan_wait_0 */
+/* 0x06ed: ctx_chan_wait_0 */
 	0x4014cf40,
 	0xf41f44f0,
 	0x32d0fa1b,
 	0x000bfe00,
 	0xb61f2af0,
 	0x20b60424,
-	0x3c87f102,
-	0x0684b608,
+	0xf094bd02,
+	0x07f10899,
+	0x03f00f00,
+	0x0009d002,
+	0x17f104bd,
+	0x14b60a04,
+	0x0012d006,
+	0x0a2017f1,
+	0xf00614b6,
+	0x23f10227,
+	0x12d08000,
+	0x1017f000,
+	0x020027f1,
+	0xfa0223f0,
+	0x03f80512,
 	0x99f094bd,
-	0x0089d008,
-	0x0a0417f1,
-	0xd00614b6,
-	0x17f10012,
-	0x14b60a20,
-	0x0227f006,
-	0x800023f1,
-	0xf00012d0,
-	0x27f11017,
-	0x23f00300,
-	0x0512fa02,
-	0x87f103f8,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00899,
-	0xc1019800,
+	0x0007f108,
+	0x0203f017,
+	0xbd0009d0,
+	0x81019804,
 	0x981814b6,
-	0x25b6c002,
+	0x25b68002,
 	0x0512fd08,
-	0xf1160180,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00999f0,
-	0x27f10089,
-	0x24b60a04,
-	0x0021d006,
-	0xf10127f0,
-	0xb60a2017,
-	0x12d00614,
-	0x0017f100,
-	0x0613f002,
-	0xf80501fa,
-	0x5c87f103,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d009,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0599f094,
-	0xf80089d0,
-/* 0x07b5: ctx_chan */
-	0x3121f500,
-	0xb821f506,
-	0x0ca7f006,
-	0xf1c921f4,
-	0xb60a1017,
-	0x27f00614,
-	0x0012d005,
-/* 0x07d0: ctx_chan_wait */
-	0xfd0012cf,
-	0x1bf40522,
-	0x4921f5fa,
-/* 0x07df: ctx_mmio_exec */
-	0x9800f806,
-	0x27f18103,
-	0x24b60a04,
-	0x0023d006,
-/* 0x07ee: ctx_mmio_loop */
-	0x34c434bd,
-	0x0f1bf4ff,
-	0x030057f1,
-	0xfa0653f0,
-	0x03f80535,
-/* 0x0800: ctx_mmio_pull */
-	0x98c04e98,
-	0x21f4c14f,
-	0x0830b68d,
-	0xf40112b6,
-/* 0x0812: ctx_mmio_done */
-	0x0398df1b,
-	0x0023d016,
-	0xf1800080,
-	0xf0020017,
+	0xbd160180,
+	0x0999f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0a0427f1,
+	0xd00624b6,
+	0x27f00021,
+	0x2017f101,
+	0x0614b60a,
+	0xf10012d0,
+	0xf0010017,
 	0x01fa0613,
-	0xf803f806,
-/* 0x0829: ctx_xfer */
-	0x00f7f100,
-	0x06f4b60c,
-	0xd004e7f0,
-/* 0x0836: ctx_xfer_idle */
-	0xfecf80fe,
-	0x00e4f100,
-	0xf91bf420,
-	0xf40611f4,
-/* 0x0846: ctx_xfer_pre */
-	0xf7f01102,
-	0x9821f510,
-	0x3121f506,
-	0x1c11f406,
-/* 0x0854: ctx_xfer_pre_load */
-	0xf502f7f0,
-	0xf5065721,
-	0xf5066621,
-	0xbd067821,
-	0x5721f5f4,
-	0xb821f506,
-/* 0x086d: ctx_xfer_exec */
-	0x16019806,
-	0x041427f1,
+	0xbd03f805,
+	0x0999f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x99f094bd,
+	0x0007f105,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x07bb: ctx_chan */
+	0xf500f804,
+	0xf5062b21,
+	0xf006b221,
+	0x21f40ca7,
+	0x1017f1c9,
+	0x0614b60a,
+	0xd00527f0,
+/* 0x07d6: ctx_chan_wait */
+	0x12cf0012,
+	0x0522fd00,
+	0xf5fa1bf4,
+	0xf8064321,
+/* 0x07e5: ctx_mmio_exec */
+	0x41039800,
+	0x0a0427f1,
 	0xd00624b6,
-	0xe7f10020,
-	0xe3f0a500,
-	0x021fb941,
+	0x34bd0023,
+/* 0x07f4: ctx_mmio_loop */
+	0xf4ff34c4,
+	0x57f10f1b,
+	0x53f00200,
+	0x0535fa06,
+/* 0x0806: ctx_mmio_pull */
+	0x4e9803f8,
+	0x814f9880,
 	0xb68d21f4,
-	0xfcf004e0,
-	0x022cf001,
-	0xfd0124b6,
-	0x21f405f2,
-	0xfc17f18d,
-	0x0213f04a,
-	0xd00c27f0,
-	0x21f50012,
-	0x27f10207,
-	0x23f047fc,
-	0x0020d002,
-	0xb6012cf0,
-	0x12d00320,
-	0x01acf000,
-	0xf006a5f0,
-	0x0c9800b7,
-	0x150d9814,
-	0xf500e7f0,
-	0xf0015c21,
-	0x21f508a7,
-	0x21f50103,
-	0x01f40207,
-	0x0ca7f022,
-	0xf1c921f4,
-	0xb60a1017,
-	0x27f00614,
-	0x0012d005,
-/* 0x08f4: ctx_xfer_post_save_wait */
-	0xfd0012cf,
-	0x1bf40522,
-	0x3202f4fa,
-/* 0x0900: ctx_xfer_post */
-	0xf502f7f0,
-	0xbd065721,
-	0x9821f5f4,
-	0x2621f506,
-	0x6621f502,
+	0x12b60830,
+	0xdf1bf401,
+/* 0x0818: ctx_mmio_done */
+	0xd0160398,
+	0x00800023,
+	0x0017f140,
+	0x0613f001,
+	0xf80601fa,
+/* 0x082f: ctx_xfer */
+	0xf100f803,
+	0xb60c00f7,
+	0xe7f006f4,
+	0x80fed004,
+/* 0x083c: ctx_xfer_idle */
+	0xf100fecf,
+	0xf42000e4,
+	0x11f4f91b,
+	0x1102f406,
+/* 0x084c: ctx_xfer_pre */
+	0xf510f7f0,
+	0xf5069221,
+	0xf4062b21,
+/* 0x085a: ctx_xfer_pre_load */
+	0xf7f01c11,
+	0x5121f502,
+	0x6021f506,
+	0x7221f506,
 	0xf5f4bd06,
-	0xf4065721,
-	0x01981011,
-	0x0511fd80,
-	0xf5070bf4,
-/* 0x092b: ctx_xfer_no_post_mmio */
-	0xf507df21,
-/* 0x092f: ctx_xfer_done */
-	0xf8064921,
-	0x00000000,
-	0x00000000,
+	0xf5065121,
+/* 0x0873: ctx_xfer_exec */
+	0x9806b221,
+	0x27f11601,
+	0x24b60414,
+	0x0020d006,
+	0xa500e7f1,
+	0xb941e3f0,
+	0x21f4021f,
+	0x04e0b68d,
+	0xf001fcf0,
+	0x24b6022c,
+	0x05f2fd01,
+	0xf18d21f4,
+	0xf04afc17,
+	0x27f00213,
+	0x0012d00c,
+	0x021521f5,
+	0x47fc27f1,
+	0xd00223f0,
+	0x2cf00020,
+	0x0320b601,
+	0xf00012d0,
+	0xa5f001ac,
+	0x00b7f006,
+	0x98000c98,
+	0xe7f0010d,
+	0x6621f500,
+	0x08a7f001,
+	0x010921f5,
+	0x021521f5,
+	0xf02201f4,
+	0x21f40ca7,
+	0x1017f1c9,
+	0x0614b60a,
+	0xd00527f0,
+/* 0x08fa: ctx_xfer_post_save_wait */
+	0x12cf0012,
+	0x0522fd00,
+	0xf4fa1bf4,
+/* 0x0906: ctx_xfer_post */
+	0xf7f03202,
+	0x5121f502,
+	0xf5f4bd06,
+	0xf5069221,
+	0xf5023421,
+	0xbd066021,
+	0x5121f5f4,
+	0x1011f406,
+	0xfd400198,
+	0x0bf40511,
+	0xe521f507,
+/* 0x0931: ctx_xfer_no_post_mmio */
+	0x4321f507,
+/* 0x0935: ctx_xfer_done */
+	0x0000f806,
 	0x00000000,
 	0x00000000,
 	0x00000000,
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvd7.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvd7.fuc
new file mode 100644
index 000000000000..afbe03ac9077
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvd7.fuc
@@ -0,0 +1,40 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#define CHIPSET GF117
+#include "macros.fuc"
+
+.section #nvd7_grhub_data
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "hub.fuc"
+#undef INCLUDE_DATA
+
+.section #nvd7_grhub_code
+#define INCLUDE_CODE
+bra #init
+#include "com.fuc"
+#include "hub.fuc"
+.align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvd7.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvd7.fuc.h
new file mode 100644
index 000000000000..a1b9f763996a
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvd7.fuc.h
@@ -0,0 +1,921 @@
+uint32_t nvd7_grhub_data[] = {
+/* 0x0000: hub_mmio_list_head */
+	0x00000300,
+/* 0x0004: hub_mmio_list_tail */
+	0x00000304,
+/* 0x0008: gpc_count */
+	0x00000000,
+/* 0x000c: rop_count */
+	0x00000000,
+/* 0x0010: cmd_queue */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0058: ctx_current */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0100: chan_data */
+/* 0x0100: chan_mmio_count */
+	0x00000000,
+/* 0x0104: chan_mmio_address */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0200: xfer_data */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0300: hub_mmio_list_base */
+	0x0417e91c,
+};
+
+uint32_t nvd7_grhub_code[] = {
+	0x031b0ef5,
+/* 0x0004: queue_put */
+	0x9800d898,
+	0x86f001d9,
+	0x0489b808,
+	0xf00c1bf4,
+	0x21f502f7,
+	0x00f802fe,
+/* 0x001c: queue_put_next */
+	0xb60798c4,
+	0x8dbb0384,
+	0x0880b600,
+	0x80008e80,
+	0x90b6018f,
+	0x0f94f001,
+	0xf801d980,
+/* 0x0039: queue_get */
+	0x0131f400,
+	0x9800d898,
+	0x89b801d9,
+	0x210bf404,
+	0xb60789c4,
+	0x9dbb0394,
+	0x0890b600,
+	0x98009e98,
+	0x80b6019f,
+	0x0f84f001,
+	0xf400d880,
+/* 0x0066: queue_get_done */
+	0x00f80132,
+/* 0x0068: nv_rd32 */
+	0x0728b7f1,
+	0xb906b4b6,
+	0xc9f002ec,
+	0x00bcd01f,
+/* 0x0078: nv_rd32_wait */
+	0xc800bccf,
+	0x1bf41fcc,
+	0x06a7f0fa,
+	0x010921f5,
+	0xf840bfcf,
+/* 0x008d: nv_wr32 */
+	0x28b7f100,
+	0x06b4b607,
+	0xb980bfd0,
+	0xc9f002ec,
+	0x1ec9f01f,
+/* 0x00a3: nv_wr32_wait */
+	0xcf00bcd0,
+	0xccc800bc,
+	0xfa1bf41f,
+/* 0x00ae: watchdog_reset */
+	0x87f100f8,
+	0x84b60430,
+	0x1ff9f006,
+	0xf8008fd0,
+/* 0x00bd: watchdog_clear */
+	0x3087f100,
+	0x0684b604,
+	0xf80080d0,
+/* 0x00c9: wait_donez */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
+	0xf4888aff,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x87f104bd,
+	0x84b60818,
+	0x008ad006,
+/* 0x0124: wait_doneo_e */
+	0x040087f1,
+	0xcf0684b6,
+	0x8aff0088,
+	0xf30bf488,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf104bd00,
+	0xb6071087,
+	0x94bd0684,
+	0xf405bbfd,
+	0x8bd0090b,
+	0x0099f000,
+/* 0x018c: mmctx_base_disabled */
+	0xf405eefd,
+	0x8ed00c0b,
+	0xc08fd080,
+/* 0x019b: mmctx_multi_disabled */
+	0xb70199f0,
+	0xc8010080,
+	0xb4b600ab,
+	0x0cb9f010,
+	0xb601aec8,
+	0xbefd11e4,
+	0x008bd005,
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
+	0xf0008ecf,
+	0x0bf41fe4,
+	0x00ce98fa,
+	0xd005e9fd,
+	0xc0b6c08e,
+	0x04cdb804,
+	0xc8e81bf4,
+	0x1bf402ab,
+/* 0x01d5: mmctx_fini_wait */
+	0x008bcf18,
+	0xb01fb4f0,
+	0x1bf410b4,
+	0x02a7f0f7,
+	0xf4c921f4,
+/* 0x01ea: mmctx_stop */
+	0xabc81b0e,
+	0x10b4b600,
+	0xf00cb9f0,
+	0x8bd012b9,
+/* 0x01f9: mmctx_stop_wait */
+	0x008bcf00,
+	0xf412bbc8,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
+	0x21f50089,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f00f00,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
+	0xfca7f102,
+	0x02a3f046,
+	0x0400aba0,
+	0xf040a0d0,
+	0xbcd001c7,
+	0x1521f500,
+	0x010c9202,
+	0xf000acd0,
+	0xbcd002c7,
+	0x1521f500,
+	0x3421f502,
+	0x8087f102,
+	0x0684b608,
+	0xb70089cf,
+	0x95220080,
+/* 0x02ca: ctx_init_strand_loop */
+	0x8ed008fe,
+	0x408ed000,
+	0xb6808acf,
+	0xa0b606a5,
+	0x00eabb01,
+	0xb60480b6,
+	0x1bf40192,
+	0x08e4b6e8,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0x07f100f8,
+	0x03f00500,
+	0x000fd002,
+	0xf7f004bd,
+	0x0007f101,
+	0x0303f007,
+	0xbd000fd0,
+/* 0x031b: init */
+	0xbd00f804,
+	0x0004fe04,
+	0xf10007fe,
+	0xf0120017,
+	0x12d00227,
+	0xb117f100,
+	0x0010fe05,
+	0x040017f1,
+	0xf1c010d0,
+	0xb6040437,
+	0x27f10634,
+	0x32d02003,
+	0x0427f100,
+	0x0132d020,
+	0x200b27f1,
+	0xf10232d0,
+	0xd0200c27,
+	0x27f10732,
+	0x24b60c24,
+	0x0003b906,
+	0xf10023d0,
+	0xf0870427,
+	0x12d00023,
+	0x0012b700,
+	0x0427f001,
+	0xf40012d0,
+	0xe7f11031,
+	0xe3f09604,
+	0x6821f440,
+	0x8090f1c7,
+	0xf4f00301,
+	0x020f801f,
+	0xbb0117f0,
+	0x12b6041f,
+	0x0c27f101,
+	0x0624b604,
+	0xd00021d0,
+	0x17f14021,
+	0x0e980100,
+	0x010f9800,
+	0x014721f5,
+	0x070037f1,
+	0x950634b6,
+	0x34d00814,
+	0x4034d000,
+	0x130030b7,
+	0xb6001fbb,
+	0x3fd002f5,
+	0x0815b600,
+	0xb60110b6,
+	0x1fb90814,
+	0x7121f502,
+	0x001fbb02,
+	0xf1020398,
+	0xf0200047,
+/* 0x03f6: init_gpc */
+	0x4ea05043,
+	0x1fb90804,
+	0x8d21f402,
+	0x010c4ea0,
+	0x21f4f4bd,
+	0x044ea08d,
+	0x8d21f401,
+	0x01004ea0,
+	0xf402f7f0,
+	0x4ea08d21,
+/* 0x041e: init_gpc_wait */
+	0x21f40800,
+	0x1fffc868,
+	0xa0fa0bf4,
+	0xf408044e,
+	0x1fbb6821,
+	0x0040b700,
+	0x0132b680,
+	0xf1be1bf4,
+	0xf0010007,
+	0x01d00203,
+	0xbd04bd00,
+	0x1f19f014,
+	0x080007f1,
+	0xd00203f0,
+	0x04bd0001,
+/* 0x0458: main */
+	0xf40031f4,
+	0xd7f00028,
+	0x3921f410,
+	0xb1f401f4,
+	0xf54001e4,
+	0xbd00de1b,
+	0x0499f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0b0017f1,
+	0xcf0614b6,
+	0x11cf4012,
+	0x1f13c800,
+	0x00870bf5,
+	0xf41f23c8,
+	0x20f9620b,
+	0xbd0212b9,
+	0x0799f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0xf40132f4,
+	0x21f50231,
+	0x94bd082f,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xfc04bd00,
+	0xf094bd20,
+	0x07f10699,
+	0x03f00f00,
+	0x0009d002,
+	0x31f404bd,
+	0x2f21f501,
+	0xf094bd08,
+	0x07f10699,
+	0x03f01700,
+	0x0009d002,
+	0x0ef404bd,
+/* 0x04f9: chsw_prev_no_next */
+	0xb920f931,
+	0x32f40212,
+	0x0232f401,
+	0x082f21f5,
+	0x17f120fc,
+	0x14b60b00,
+	0x0012d006,
+/* 0x0517: chsw_no_prev */
+	0xc8130ef4,
+	0x0bf41f23,
+	0x0131f40d,
+	0xf50232f4,
+/* 0x0527: chsw_done */
+	0xf1082f21,
+	0xb60b0c17,
+	0x27f00614,
+	0x0012d001,
+	0x99f094bd,
+	0x0007f104,
+	0x0203f017,
+	0xbd0009d0,
+	0x130ef504,
+/* 0x0549: main_not_ctx_switch */
+	0x01e4b0ff,
+	0xb90d1bf4,
+	0x21f502f2,
+	0x0ef407bb,
+/* 0x0559: main_not_ctx_chan */
+	0x02e4b046,
+	0xbd321bf4,
+	0x0799f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0xf40132f4,
+	0x21f50232,
+	0x94bd082f,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xf404bd00,
+/* 0x058e: main_not_ctx_save */
+	0xef94110e,
+	0x01f5f010,
+	0x02fe21f5,
+	0xfec00ef5,
+/* 0x059c: main_done */
+	0x29f024bd,
+	0x0007f11f,
+	0x0203f008,
+	0xbd0002d0,
+	0xab0ef504,
+/* 0x05b1: ih */
+	0xfe80f9fe,
+	0x80f90188,
+	0xa0f990f9,
+	0xd0f9b0f9,
+	0xf0f9e0f9,
+	0x0acf04bd,
+	0x04abc480,
+	0xf11d0bf4,
+	0xf01900b7,
+	0xbecf10d7,
+	0x00bfcf40,
+	0xb70421f4,
+	0xf00400b0,
+	0xbed001e7,
+/* 0x05e9: ih_no_fifo */
+	0x00abe400,
+	0x0d0bf401,
+	0xf110d7f0,
+	0xf44001e7,
+/* 0x05fa: ih_no_ctxsw */
+	0xb7f10421,
+	0xb0bd0104,
+	0xf4b4abff,
+	0xa7f10d0b,
+	0xa4b60c1c,
+	0x00abd006,
+/* 0x0610: ih_no_other */
+	0xfc400ad0,
+	0xfce0fcf0,
+	0xfcb0fcd0,
+	0xfc90fca0,
+	0x0088fe80,
+	0x32f480fc,
+/* 0x062b: ctx_4160s */
+	0xf101f800,
+	0xf04160e7,
+	0xf7f040e3,
+	0x8d21f401,
+/* 0x0638: ctx_4160s_wait */
+	0xc86821f4,
+	0x0bf404ff,
+/* 0x0643: ctx_4160c */
+	0xf100f8fa,
+	0xf04160e7,
+	0xf4bd40e3,
+	0xf88d21f4,
+/* 0x0651: ctx_4170s */
+	0x70e7f100,
+	0x40e3f041,
+	0xf410f5f0,
+	0x00f88d21,
+/* 0x0660: ctx_4170w */
+	0x4170e7f1,
+	0xf440e3f0,
+	0xf4f06821,
+	0xf31bf410,
+/* 0x0672: ctx_redswitch */
+	0xe7f100f8,
+	0xe4b60614,
+	0x70f7f106,
+	0x00efd002,
+/* 0x0683: ctx_redswitch_delay */
+	0xb608f7f0,
+	0x1bf401f2,
+	0x70f7f1fd,
+	0x00efd007,
+/* 0x0692: ctx_86c */
+	0xe7f100f8,
+	0xe4b6086c,
+	0x00efd006,
+	0x8a14e7f1,
+	0xf440e3f0,
+	0xe7f18d21,
+	0xe3f0a86c,
+	0x8d21f441,
+/* 0x06b2: ctx_load */
+	0x94bd00f8,
+	0xf10599f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf004bd00,
+	0x21f40ca7,
+	0x2417f1c9,
+	0x0614b60a,
+	0xf10010d0,
+	0xb60b0037,
+	0x32d00634,
+	0x0c17f140,
+	0x0614b60a,
+	0xd00747f0,
+	0x14d00012,
+/* 0x06ed: ctx_chan_wait_0 */
+	0x4014cf40,
+	0xf41f44f0,
+	0x32d0fa1b,
+	0x000bfe00,
+	0xb61f2af0,
+	0x20b60424,
+	0xf094bd02,
+	0x07f10899,
+	0x03f00f00,
+	0x0009d002,
+	0x17f104bd,
+	0x14b60a04,
+	0x0012d006,
+	0x0a2017f1,
+	0xf00614b6,
+	0x23f10227,
+	0x12d08000,
+	0x1017f000,
+	0x020027f1,
+	0xfa0223f0,
+	0x03f80512,
+	0x99f094bd,
+	0x0007f108,
+	0x0203f017,
+	0xbd0009d0,
+	0x81019804,
+	0x981814b6,
+	0x25b68002,
+	0x0512fd08,
+	0xbd160180,
+	0x0999f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0a0427f1,
+	0xd00624b6,
+	0x27f00021,
+	0x2017f101,
+	0x0614b60a,
+	0xf10012d0,
+	0xf0010017,
+	0x01fa0613,
+	0xbd03f805,
+	0x0999f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x99f094bd,
+	0x0007f105,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x07bb: ctx_chan */
+	0xf500f804,
+	0xf5062b21,
+	0xf006b221,
+	0x21f40ca7,
+	0x1017f1c9,
+	0x0614b60a,
+	0xd00527f0,
+/* 0x07d6: ctx_chan_wait */
+	0x12cf0012,
+	0x0522fd00,
+	0xf5fa1bf4,
+	0xf8064321,
+/* 0x07e5: ctx_mmio_exec */
+	0x41039800,
+	0x0a0427f1,
+	0xd00624b6,
+	0x34bd0023,
+/* 0x07f4: ctx_mmio_loop */
+	0xf4ff34c4,
+	0x57f10f1b,
+	0x53f00200,
+	0x0535fa06,
+/* 0x0806: ctx_mmio_pull */
+	0x4e9803f8,
+	0x814f9880,
+	0xb68d21f4,
+	0x12b60830,
+	0xdf1bf401,
+/* 0x0818: ctx_mmio_done */
+	0xd0160398,
+	0x00800023,
+	0x0017f140,
+	0x0613f001,
+	0xf80601fa,
+/* 0x082f: ctx_xfer */
+	0xf100f803,
+	0xb60c00f7,
+	0xe7f006f4,
+	0x80fed004,
+/* 0x083c: ctx_xfer_idle */
+	0xf100fecf,
+	0xf42000e4,
+	0x11f4f91b,
+	0x1102f406,
+/* 0x084c: ctx_xfer_pre */
+	0xf510f7f0,
+	0xf5069221,
+	0xf4062b21,
+/* 0x085a: ctx_xfer_pre_load */
+	0xf7f01c11,
+	0x5121f502,
+	0x6021f506,
+	0x7221f506,
+	0xf5f4bd06,
+	0xf5065121,
+/* 0x0873: ctx_xfer_exec */
+	0x9806b221,
+	0x27f11601,
+	0x24b60414,
+	0x0020d006,
+	0xa500e7f1,
+	0xb941e3f0,
+	0x21f4021f,
+	0x04e0b68d,
+	0xf001fcf0,
+	0x24b6022c,
+	0x05f2fd01,
+	0xf18d21f4,
+	0xf04afc17,
+	0x27f00213,
+	0x0012d00c,
+	0x021521f5,
+	0x47fc27f1,
+	0xd00223f0,
+	0x2cf00020,
+	0x0320b601,
+	0xf00012d0,
+	0xa5f001ac,
+	0x00b7f006,
+	0x98000c98,
+	0xe7f0010d,
+	0x6621f500,
+	0x08a7f001,
+	0x010921f5,
+	0x021521f5,
+	0xf02201f4,
+	0x21f40ca7,
+	0x1017f1c9,
+	0x0614b60a,
+	0xd00527f0,
+/* 0x08fa: ctx_xfer_post_save_wait */
+	0x12cf0012,
+	0x0522fd00,
+	0xf4fa1bf4,
+/* 0x0906: ctx_xfer_post */
+	0xf7f03202,
+	0x5121f502,
+	0xf5f4bd06,
+	0xf5069221,
+	0xf5023421,
+	0xbd066021,
+	0x5121f5f4,
+	0x1011f406,
+	0xfd400198,
+	0x0bf40511,
+	0xe521f507,
+/* 0x0931: ctx_xfer_no_post_mmio */
+	0x4321f507,
+/* 0x0935: ctx_xfer_done */
+	0x0000f806,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc
index 7fe9d7cf486b..d4840f1879fd 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc
@@ -1,6 +1,5 @@
-/* fuc microcode for nve0 PGRAPH/HUB
- *
- * Copyright 2011 Red Hat Inc.
+/*
+ * Copyright 2013 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -20,774 +19,22 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
- * Authors: Ben Skeggs
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
  */
 
-/* To build:
- *    m4 nve0_grhub.fuc | envyas -a -w -m fuc -V nva3 -o nve0_grhub.fuc.h
- */
+#define CHIPSET GK100
+#include "macros.fuc"
 
 .section #nve0_grhub_data
-include(`nve0.fuc')
-gpc_count:		.b32 0
-rop_count:		.b32 0
-cmd_queue:		queue_init
-hub_mmio_list_head:	.b32 0
-hub_mmio_list_tail:	.b32 0
-
-ctx_current:		.b32 0
-
-chipsets:
-.b8  0xe4 0 0 0
-.b16 #nve4_hub_mmio_head
-.b16 #nve4_hub_mmio_tail
-.b8  0xe7 0 0 0
-.b16 #nve4_hub_mmio_head
-.b16 #nve4_hub_mmio_tail
-.b8  0xe6 0 0 0
-.b16 #nve4_hub_mmio_head
-.b16 #nve4_hub_mmio_tail
-.b8  0 0 0 0
-
-nve4_hub_mmio_head:
-mmctx_data(0x17e91c, 2)
-mmctx_data(0x400204, 2)
-mmctx_data(0x404010, 7)
-mmctx_data(0x4040a8, 9)
-mmctx_data(0x4040d0, 7)
-mmctx_data(0x4040f8, 1)
-mmctx_data(0x404130, 3)
-mmctx_data(0x404150, 3)
-mmctx_data(0x404164, 1)
-mmctx_data(0x4041a0, 4)
-mmctx_data(0x404200, 4)
-mmctx_data(0x404404, 14)
-mmctx_data(0x404460, 4)
-mmctx_data(0x404480, 1)
-mmctx_data(0x404498, 1)
-mmctx_data(0x404604, 4)
-mmctx_data(0x404618, 4)
-mmctx_data(0x40462c, 2)
-mmctx_data(0x404640, 1)
-mmctx_data(0x404654, 1)
-mmctx_data(0x404660, 1)
-mmctx_data(0x404678, 19)
-mmctx_data(0x4046c8, 3)
-mmctx_data(0x404700, 3)
-mmctx_data(0x404718, 10)
-mmctx_data(0x404744, 2)
-mmctx_data(0x404754, 1)
-mmctx_data(0x405800, 1)
-mmctx_data(0x405830, 3)
-mmctx_data(0x405854, 1)
-mmctx_data(0x405870, 4)
-mmctx_data(0x405a00, 2)
-mmctx_data(0x405a18, 1)
-mmctx_data(0x405b00, 1)
-mmctx_data(0x405b10, 1)
-mmctx_data(0x406020, 1)
-mmctx_data(0x406028, 4)
-mmctx_data(0x4064a8, 2)
-mmctx_data(0x4064b4, 2)
-mmctx_data(0x4064c0, 12)
-mmctx_data(0x4064fc, 1)
-mmctx_data(0x407040, 1)
-mmctx_data(0x407804, 1)
-mmctx_data(0x40780c, 6)
-mmctx_data(0x4078bc, 1)
-mmctx_data(0x408000, 7)
-mmctx_data(0x408064, 1)
-mmctx_data(0x408800, 3)
-mmctx_data(0x408840, 1)
-mmctx_data(0x408900, 3)
-mmctx_data(0x408980, 1)
-nve4_hub_mmio_tail:
-
-.align 256
-chan_data:
-chan_mmio_count:	.b32 0
-chan_mmio_address:	.b32 0
-
-.align 256
-xfer_data: 		.b32 0
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "hub.fuc"
+#undef INCLUDE_DATA
 
 .section #nve0_grhub_code
+#define INCLUDE_CODE
 bra #init
-define(`include_code')
-include(`nve0.fuc')
-
-// reports an exception to the host
-//
-// In: $r15 error code (see nve0.fuc)
-//
-error:
-	push $r14
-	mov $r14 0x814
-	shl b32 $r14 6
-	iowr I[$r14 + 0x000] $r15	// CC_SCRATCH[5] = error code
-	mov $r14 0xc1c
-	shl b32 $r14 6
-	mov $r15 1
-	iowr I[$r14 + 0x000] $r15	// INTR_UP_SET
-	pop $r14
-	ret
-
-// HUB fuc initialisation, executed by triggering ucode start, will
-// fall through to main loop after completion.
-//
-// Input:
-//   CC_SCRATCH[0]: chipset (PMC_BOOT_0 read returns 0x0bad0bad... sigh)
-//
-// Output:
-//   CC_SCRATCH[0]:
-//	     31:31: set to signal completion
-//   CC_SCRATCH[1]:
-//	      31:0: total PGRAPH context size
-//
-init:
-	clear b32 $r0
-	mov $sp $r0
-	mov $xdbase $r0
-
-	// enable fifo access
-	mov $r1 0x1200
-	mov $r2 2
-	iowr I[$r1 + 0x000] $r2	// FIFO_ENABLE
-
-	// setup i0 handler, and route all interrupts to it
-	mov $r1 #ih
-	mov $iv0 $r1
-	mov $r1 0x400
-	iowr I[$r1 + 0x300] $r0	// INTR_DISPATCH
-
-	// route HUB_CHANNEL_SWITCH to fuc interrupt 8
-	mov $r3 0x404
-	shl b32 $r3 6
-	mov $r2 0x2003		// { HUB_CHANNEL_SWITCH, ZERO } -> intr 8
-	iowr I[$r3 + 0x000] $r2
-
-	// not sure what these are, route them because NVIDIA does, and
-	// the IRQ handler will signal the host if we ever get one.. we
-	// may find out if/why we need to handle these if so..
-	//
-	mov $r2 0x2004
-	iowr I[$r3 + 0x004] $r2 // { 0x04, ZERO } -> intr 9
-	mov $r2 0x200b
-	iowr I[$r3 + 0x008] $r2 // { 0x0b, ZERO } -> intr 10
-	mov $r2 0x200c
-	iowr I[$r3 + 0x01c] $r2 // { 0x0c, ZERO } -> intr 15
-
-	// enable all INTR_UP interrupts
-	mov $r2 0xc24
-	shl b32 $r2 6
-	not b32 $r3 $r0
-	iowr I[$r2] $r3
-
-	// enable fifo, ctxsw, 9, 10, 15 interrupts
-	mov $r2 -0x78fc		// 0x8704
-	sethi $r2 0
-	iowr I[$r1 + 0x000] $r2	// INTR_EN_SET
-
-	// fifo level triggered, rest edge
-	sub b32 $r1 0x100
-	mov $r2 4
-	iowr I[$r1] $r2
-
-	// enable interrupts
-	bset $flags ie0
-
-	// fetch enabled GPC/ROP counts
-	mov $r14 -0x69fc	// 0x409604
-	sethi $r14 0x400000
-	call #nv_rd32
-	extr $r1 $r15 16:20
-	st b32 D[$r0 + #rop_count] $r1
-	and $r15 0x1f
-	st b32 D[$r0 + #gpc_count] $r15
-
-	// set BAR_REQMASK to GPC mask
-	mov $r1 1
-	shl b32 $r1 $r15
-	sub b32 $r1 1
-	mov $r2 0x40c
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r1
-	iowr I[$r2 + 0x100] $r1
-
-	// find context data for this chipset
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x000]		// CC_SCRATCH[0]
-	mov $r15 #chipsets - 8
-	init_find_chipset:
-		add b32 $r15 8
-		ld b32 $r3 D[$r15 + 0x00]
-		cmpu b32 $r3 $r2
-		bra e #init_context
-		cmpu b32 $r3 0
-		bra ne #init_find_chipset
-		// unknown chipset
-		ret
-
-	// context size calculation, reserve first 256 bytes for use by fuc
-	init_context:
-	mov $r1 256
-
-	// calculate size of mmio context data
-	ld b16 $r14 D[$r15 + 4]
-	ld b16 $r15 D[$r15 + 6]
-	sethi $r14 0
-	st b32 D[$r0 + #hub_mmio_list_head] $r14
-	st b32 D[$r0 + #hub_mmio_list_tail] $r15
-	call #mmctx_size
-
-	// set mmctx base addresses now so we don't have to do it later,
-	// they don't (currently) ever change
-	mov $r3 0x700
-	shl b32 $r3 6
-	shr b32 $r4 $r1 8
-	iowr I[$r3 + 0x000] $r4		// MMCTX_SAVE_SWBASE
-	iowr I[$r3 + 0x100] $r4		// MMCTX_LOAD_SWBASE
-	add b32 $r3 0x1300
-	add b32 $r1 $r15
-	shr b32 $r15 2
-	iowr I[$r3 + 0x000] $r15	// MMCTX_LOAD_COUNT, wtf for?!?
-
-	// strands, base offset needs to be aligned to 256 bytes
-	shr b32 $r1 8
-	add b32 $r1 1
-	shl b32 $r1 8
-	mov b32 $r15 $r1
-	call #strand_ctx_init
-	add b32 $r1 $r15
-
-	// initialise each GPC in sequence by passing in the offset of its
-	// context data in GPCn_CC_SCRATCH[1], and starting its FUC (which
-	// has previously been uploaded by the host) running.
-	//
-	// the GPC fuc init sequence will set GPCn_CC_SCRATCH[0] bit 31
-	// when it has completed, and return the size of its context data
-	// in GPCn_CC_SCRATCH[1]
-	//
-	ld b32 $r3 D[$r0 + #gpc_count]
-	mov $r4 0x2000
-	sethi $r4 0x500000
-	init_gpc:
-		// setup, and start GPC ucode running
-		add b32 $r14 $r4 0x804
-		mov b32 $r15 $r1
-		call #nv_wr32			// CC_SCRATCH[1] = ctx offset
-		add b32 $r14 $r4 0x800
-		mov b32 $r15 $r2
-		call #nv_wr32			// CC_SCRATCH[0] = chipset
-		add b32 $r14 $r4 0x10c
-		clear b32 $r15
-		call #nv_wr32
-		add b32 $r14 $r4 0x104
-		call #nv_wr32			// ENTRY
-		add b32 $r14 $r4 0x100
-		mov $r15 2			// CTRL_START_TRIGGER
-		call #nv_wr32			// CTRL
-
-		// wait for it to complete, and adjust context size
-		add b32 $r14 $r4 0x800
-		init_gpc_wait:
-			call #nv_rd32
-			xbit $r15 $r15 31
-			bra e #init_gpc_wait
-		add b32 $r14 $r4 0x804
-		call #nv_rd32
-		add b32 $r1 $r15
-
-		// next!
-		add b32 $r4 0x8000
-		sub b32 $r3 1
-		bra ne #init_gpc
-
-	// save context size, and tell host we're ready
-	mov $r2 0x800
-	shl b32 $r2 6
-	iowr I[$r2 + 0x100] $r1		// CC_SCRATCH[1]  = context size
-	add b32 $r2 0x800
-	clear b32 $r1
-	bset $r1 31
-	iowr I[$r2 + 0x000] $r1		// CC_SCRATCH[0] |= 0x80000000
-
-// Main program loop, very simple, sleeps until woken up by the interrupt
-// handler, pulls a command from the queue and executes its handler
-//
-main:
-	// sleep until we have something to do
-	bset $flags $p0
-	sleep $p0
-	mov $r13 #cmd_queue
-	call #queue_get
-	bra $p1 #main
-
-	// context switch, requested by GPU?
-	cmpu b32 $r14 0x4001
-	bra ne #main_not_ctx_switch
-		trace_set(T_AUTO)
-		mov $r1 0xb00
-		shl b32 $r1 6
-		iord $r2 I[$r1 + 0x100]		// CHAN_NEXT
-		iord $r1 I[$r1 + 0x000]		// CHAN_CUR
-
-		xbit $r3 $r1 31
-		bra e #chsw_no_prev
-			xbit $r3 $r2 31
-			bra e #chsw_prev_no_next
-				push $r2
-				mov b32 $r2 $r1
-				trace_set(T_SAVE)
-				bclr $flags $p1
-				bset $flags $p2
-				call #ctx_xfer
-				trace_clr(T_SAVE);
-				pop $r2
-				trace_set(T_LOAD);
-				bset $flags $p1
-				call #ctx_xfer
-				trace_clr(T_LOAD);
-				bra #chsw_done
-			chsw_prev_no_next:
-				push $r2
-				mov b32 $r2 $r1
-				bclr $flags $p1
-				bclr $flags $p2
-				call #ctx_xfer
-				pop $r2
-				mov $r1 0xb00
-				shl b32 $r1 6
-				iowr I[$r1] $r2
-				bra #chsw_done
-		chsw_no_prev:
-			xbit $r3 $r2 31
-			bra e #chsw_done
-				bset $flags $p1
-				bclr $flags $p2
-				call #ctx_xfer
-
-		// ack the context switch request
-		chsw_done:
-		mov $r1 0xb0c
-		shl b32 $r1 6
-		mov $r2 1
-		iowr I[$r1 + 0x000] $r2		// 0x409b0c
-		trace_clr(T_AUTO)
-		bra #main
-
-	// request to set current channel? (*not* a context switch)
-	main_not_ctx_switch:
-	cmpu b32 $r14 0x0001
-	bra ne #main_not_ctx_chan
-		mov b32 $r2 $r15
-		call #ctx_chan
-		bra #main_done
-
-	// request to store current channel context?
-	main_not_ctx_chan:
-	cmpu b32 $r14 0x0002
-	bra ne #main_not_ctx_save
-		trace_set(T_SAVE)
-		bclr $flags $p1
-		bclr $flags $p2
-		call #ctx_xfer
-		trace_clr(T_SAVE)
-		bra #main_done
-
-	main_not_ctx_save:
-		shl b32 $r15 $r14 16
-		or $r15 E_BAD_COMMAND
-		call #error
-		bra #main
-
-	main_done:
-	mov $r1 0x820
-	shl b32 $r1 6
-	clear b32 $r2
-	bset $r2 31
-	iowr I[$r1 + 0x000] $r2		// CC_SCRATCH[0] |= 0x80000000
-	bra #main
-
-// interrupt handler
-ih:
-	push $r8
-	mov $r8 $flags
-	push $r8
-	push $r9
-	push $r10
-	push $r11
-	push $r13
-	push $r14
-	push $r15
-
-	// incoming fifo command?
-	iord $r10 I[$r0 + 0x200]	// INTR
-	and $r11 $r10 0x00000004
-	bra e #ih_no_fifo
-		// queue incoming fifo command for later processing
-		mov $r11 0x1900
-		mov $r13 #cmd_queue
-		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
-		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
-		call #queue_put
-		add b32 $r11 0x400
-		mov $r14 1
-		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
-
-	// context switch request?
-	ih_no_fifo:
-	and $r11 $r10 0x00000100
-	bra e #ih_no_ctxsw
-		// enqueue a context switch for later processing
-		mov $r13 #cmd_queue
-		mov $r14 0x4001
-		call #queue_put
-
-	// anything we didn't handle, bring it to the host's attention
-	ih_no_ctxsw:
-	mov $r11 0x104
-	not b32 $r11
-	and $r11 $r10 $r11
-	bra e #ih_no_other
-		mov $r10 0xc1c
-		shl b32 $r10 6
-		iowr I[$r10] $r11	// INTR_UP_SET
-
-	// ack, and wake up main()
-	ih_no_other:
-	iowr I[$r0 + 0x100] $r10	// INTR_ACK
-
-	pop $r15
-	pop $r14
-	pop $r13
-	pop $r11
-	pop $r10
-	pop $r9
-	pop $r8
-	mov $flags $r8
-	pop $r8
-	bclr $flags $p0
-	iret
-
-// Again, not real sure
-//
-// In: $r15 value to set 0x404170 to
-//
-ctx_4170s:
-	mov $r14 0x4170
-	sethi $r14 0x400000
-	or $r15 0x10
-	call #nv_wr32
-	ret
-
-// Waits for a ctx_4170s() call to complete
-//
-ctx_4170w:
-	mov $r14 0x4170
-	sethi $r14 0x400000
-	call #nv_rd32
-	and $r15 0x10
-	bra ne #ctx_4170w
-	ret
-
-// Disables various things, waits a bit, and re-enables them..
-//
-// Not sure how exactly this helps, perhaps "ENABLE" is not such a
-// good description for the bits we turn off?  Anyways, without this,
-// funny things happen.
-//
-ctx_redswitch:
-	mov $r14 0x614
-	shl b32 $r14 6
-	mov $r15 0x270
-	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_GPC, POWER_ALL
-	mov $r15 8
-	ctx_redswitch_delay:
-		sub b32 $r15 1
-		bra ne #ctx_redswitch_delay
-	mov $r15 0x770
-	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_ALL, POWER_ALL
-	ret
-
-// Not a clue what this is for, except that unless the value is 0x10, the
-// strand context is saved (and presumably restored) incorrectly..
-//
-// In: $r15 value to set to (0x00/0x10 are used)
-//
-ctx_86c:
-	mov $r14 0x86c
-	shl b32 $r14 6
-	iowr I[$r14] $r15	// HUB(0x86c) = val
-	mov $r14 -0x75ec
-	sethi $r14 0x400000
-	call #nv_wr32		// ROP(0xa14) = val
-	mov $r14 -0x5794
-	sethi $r14 0x410000
-	call #nv_wr32		// GPC(0x86c) = val
-	ret
-
-// ctx_load - load's a channel's ctxctl data, and selects its vm
-//
-// In: $r2 channel address
-//
-ctx_load:
-	trace_set(T_CHAN)
-
-	// switch to channel, somewhat magic in parts..
-	mov $r10 12		// DONE_UNK12
-	call #wait_donez
-	mov $r1 0xa24
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r0	// 0x409a24
-	mov $r3 0xb00
-	shl b32 $r3 6
-	iowr I[$r3 + 0x100] $r2	// CHAN_NEXT
-	mov $r1 0xa0c
-	shl b32 $r1 6
-	mov $r4 7
-	iowr I[$r1 + 0x000] $r2 // MEM_CHAN
-	iowr I[$r1 + 0x100] $r4	// MEM_CMD
-	ctx_chan_wait_0:
-		iord $r4 I[$r1 + 0x100]
-		and $r4 0x1f
-		bra ne #ctx_chan_wait_0
-	iowr I[$r3 + 0x000] $r2	// CHAN_CUR
-
-	// load channel header, fetch PGRAPH context pointer
-	mov $xtargets $r0
-	bclr $r2 31
-	shl b32 $r2 4
-	add b32 $r2 2
-
-	trace_set(T_LCHAN)
-	mov $r1 0xa04
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r2		// MEM_BASE
-	mov $r1 0xa20
-	shl b32 $r1 6
-	mov $r2 0x0002
-	sethi $r2 0x80000000
-	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vram
-	mov $r1 0x10			// chan + 0x0210
-	mov $r2 #xfer_data
-	sethi $r2 0x00020000		// 16 bytes
-	xdld $r1 $r2
-	xdwait
-	trace_clr(T_LCHAN)
-
-	// update current context
-	ld b32 $r1 D[$r0 + #xfer_data + 4]
-	shl b32 $r1 24
-	ld b32 $r2 D[$r0 + #xfer_data + 0]
-	shr b32 $r2 8
-	or $r1 $r2
-	st b32 D[$r0 + #ctx_current] $r1
-
-	// set transfer base to start of context, and fetch context header
-	trace_set(T_LCTXH)
-	mov $r2 0xa04
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r1		// MEM_BASE
-	mov $r2 1
-	mov $r1 0xa20
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vm
-	mov $r1 #chan_data
-	sethi $r1 0x00060000		// 256 bytes
-	xdld $r0 $r1
-	xdwait
-	trace_clr(T_LCTXH)
-
-	trace_clr(T_CHAN)
-	ret
-
-// ctx_chan - handler for HUB_SET_CHAN command, will set a channel as
-//            the active channel for ctxctl, but not actually transfer
-//            any context data.  intended for use only during initial
-//            context construction.
-//
-// In: $r2 channel address
-//
-ctx_chan:
-	call #ctx_load
-	mov $r10 12			// DONE_UNK12
-	call #wait_donez
-	mov $r1 0xa10
-	shl b32 $r1 6
-	mov $r2 5
-	iowr I[$r1 + 0x000] $r2		// MEM_CMD = 5 (???)
-	ctx_chan_wait:
-		iord $r2 I[$r1 + 0x000]
-		or $r2 $r2
-		bra ne #ctx_chan_wait
-	ret
-
-// Execute per-context state overrides list
-//
-// Only executed on the first load of a channel.  Might want to look into
-// removing this and having the host directly modify the channel's context
-// to change this state...  The nouveau DRM already builds this list as
-// it's definitely needed for NVIDIA's, so we may as well use it for now
-//
-// Input: $r1 mmio list length
-//
-ctx_mmio_exec:
-	// set transfer base to be the mmio list
-	ld b32 $r3 D[$r0 + #chan_mmio_address]
-	mov $r2 0xa04
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r3		// MEM_BASE
-
-	clear b32 $r3
-	ctx_mmio_loop:
-		// fetch next 256 bytes of mmio list if necessary
-		and $r4 $r3 0xff
-		bra ne #ctx_mmio_pull
-			mov $r5 #xfer_data
-			sethi $r5 0x00060000	// 256 bytes
-			xdld $r3 $r5
-			xdwait
-
-		// execute a single list entry
-		ctx_mmio_pull:
-		ld b32 $r14 D[$r4 + #xfer_data + 0x00]
-		ld b32 $r15 D[$r4 + #xfer_data + 0x04]
-		call #nv_wr32
-
-		// next!
-		add b32 $r3 8
-		sub b32 $r1 1
-		bra ne #ctx_mmio_loop
-
-	// set transfer base back to the current context
-	ctx_mmio_done:
-	ld b32 $r3 D[$r0 + #ctx_current]
-	iowr I[$r2 + 0x000] $r3		// MEM_BASE
-
-	// disable the mmio list now, we don't need/want to execute it again
-	st b32 D[$r0 + #chan_mmio_count] $r0
-	mov $r1 #chan_data
-	sethi $r1 0x00060000		// 256 bytes
-	xdst $r0 $r1
-	xdwait
-	ret
-
-// Transfer HUB context data between GPU and storage area
-//
-// In: $r2 channel address
-//     $p1 clear on save, set on load
-//     $p2 set if opposite direction done/will be done, so:
-//		on save it means: "a load will follow this save"
-//		on load it means: "a save preceeded this load"
-//
-ctx_xfer:
-	// according to mwk, some kind of wait for idle
-	mov $r15 0xc00
-	shl b32 $r15 6
-	mov $r14 4
-	iowr I[$r15 + 0x200] $r14
-	ctx_xfer_idle:
-		iord $r14 I[$r15 + 0x000]
-		and $r14 0x2000
-		bra ne #ctx_xfer_idle
-
-	bra not $p1 #ctx_xfer_pre
-	bra $p2 #ctx_xfer_pre_load
-	ctx_xfer_pre:
-		mov $r15 0x10
-		call #ctx_86c
-		bra not $p1 #ctx_xfer_exec
-
-	ctx_xfer_pre_load:
-		mov $r15 2
-		call #ctx_4170s
-		call #ctx_4170w
-		call #ctx_redswitch
-		clear b32 $r15
-		call #ctx_4170s
-		call #ctx_load
-
-	// fetch context pointer, and initiate xfer on all GPCs
-	ctx_xfer_exec:
-	ld b32 $r1 D[$r0 + #ctx_current]
-	mov $r2 0x414
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r0	// BAR_STATUS = reset
-	mov $r14 -0x5b00
-	sethi $r14 0x410000
-	mov b32 $r15 $r1
-	call #nv_wr32		// GPC_BCAST_WRCMD_DATA = ctx pointer
-	add b32 $r14 4
-	xbit $r15 $flags $p1
-	xbit $r2 $flags $p2
-	shl b32 $r2 1
-	or $r15 $r2
-	call #nv_wr32		// GPC_BCAST_WRCMD_CMD = GPC_XFER(type)
-
-	// strands
-	mov $r1 0x4afc
-	sethi $r1 0x20000
-	mov $r2 0xc
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
-	call #strand_wait
-	mov $r2 0x47fc
-	sethi $r2 0x20000
-	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
-	xbit $r2 $flags $p1
-	add b32 $r2 3
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
-
-	// mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 6		// first, last
-	mov $r11 0		// base = 0
-	ld b32 $r12 D[$r0 + #hub_mmio_list_head]
-	ld b32 $r13 D[$r0 + #hub_mmio_list_tail]
-	mov $r14 0		// not multi
-	call #mmctx_xfer
-
-	// wait for GPCs to all complete
-	mov $r10 8		// DONE_BAR
-	call #wait_doneo
-
-	// wait for strand xfer to complete
-	call #strand_wait
-
-	// post-op
-	bra $p1 #ctx_xfer_post
-		mov $r10 12		// DONE_UNK12
-		call #wait_donez
-		mov $r1 0xa10
-		shl b32 $r1 6
-		mov $r2 5
-		iowr I[$r1] $r2		// MEM_CMD
-		ctx_xfer_post_save_wait:
-			iord $r2 I[$r1]
-			or $r2 $r2
-			bra ne #ctx_xfer_post_save_wait
-
-	bra $p2 #ctx_xfer_done
-	ctx_xfer_post:
-		mov $r15 2
-		call #ctx_4170s
-		clear b32 $r15
-		call #ctx_86c
-		call #strand_post
-		call #ctx_4170w
-		clear b32 $r15
-		call #ctx_4170s
-
-		bra not $p1 #ctx_xfer_no_post_mmio
-		ld b32 $r1 D[$r0 + #chan_mmio_count]
-		or $r1 $r1
-		bra e #ctx_xfer_no_post_mmio
-			call #ctx_mmio_exec
-
-		ctx_xfer_no_post_mmio:
-
-	ctx_xfer_done:
-	ret
-
+#include "com.fuc"
+#include "hub.fuc"
 .align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc.h
index e3421af68ab9..eb7bc0e9576e 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnve0.fuc.h
@@ -1,9 +1,13 @@
 uint32_t nve0_grhub_data[] = {
-/* 0x0000: gpc_count */
+/* 0x0000: hub_mmio_list_head */
+	0x00000300,
+/* 0x0004: hub_mmio_list_tail */
+	0x00000304,
+/* 0x0008: gpc_count */
 	0x00000000,
-/* 0x0004: rop_count */
+/* 0x000c: rop_count */
 	0x00000000,
-/* 0x0008: cmd_queue */
+/* 0x0010: cmd_queue */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -22,73 +26,11 @@ uint32_t nve0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0050: hub_mmio_list_head */
+/* 0x0058: ctx_current */
 	0x00000000,
-/* 0x0054: hub_mmio_list_tail */
 	0x00000000,
-/* 0x0058: ctx_current */
 	0x00000000,
-/* 0x005c: chipsets */
-	0x000000e4,
-	0x01440078,
-	0x000000e7,
-	0x01440078,
-	0x000000e6,
-	0x01440078,
 	0x00000000,
-/* 0x0078: nve4_hub_mmio_head */
-	0x0417e91c,
-	0x04400204,
-	0x18404010,
-	0x204040a8,
-	0x184040d0,
-	0x004040f8,
-	0x08404130,
-	0x08404150,
-	0x00404164,
-	0x0c4041a0,
-	0x0c404200,
-	0x34404404,
-	0x0c404460,
-	0x00404480,
-	0x00404498,
-	0x0c404604,
-	0x0c404618,
-	0x0440462c,
-	0x00404640,
-	0x00404654,
-	0x00404660,
-	0x48404678,
-	0x084046c8,
-	0x08404700,
-	0x24404718,
-	0x04404744,
-	0x00404754,
-	0x00405800,
-	0x08405830,
-	0x00405854,
-	0x0c405870,
-	0x04405a00,
-	0x00405a18,
-	0x00405b00,
-	0x00405b10,
-	0x00406020,
-	0x0c406028,
-	0x044064a8,
-	0x044064b4,
-	0x2c4064c0,
-	0x004064fc,
-	0x00407040,
-	0x00407804,
-	0x1440780c,
-	0x004078bc,
-	0x18408000,
-	0x00408064,
-	0x08408800,
-	0x00408840,
-	0x08408900,
-	0x00408980,
-/* 0x0144: nve4_hub_mmio_tail */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -127,6 +69,47 @@ uint32_t nve0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+/* 0x0100: chan_data */
+/* 0x0100: chan_mmio_count */
+	0x00000000,
+/* 0x0104: chan_mmio_address */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -136,10 +119,7 @@ uint32_t nve0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0200: chan_data */
-/* 0x0200: chan_mmio_count */
 	0x00000000,
-/* 0x0204: chan_mmio_address */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -156,6 +136,7 @@ uint32_t nve0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
+/* 0x0200: xfer_data */
 	0x00000000,
 	0x00000000,
 	0x00000000,
@@ -203,19 +184,36 @@ uint32_t nve0_grhub_data[] = {
 	0x00000000,
 	0x00000000,
 	0x00000000,
-/* 0x0300: xfer_data */
 	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0300: hub_mmio_list_base */
+	0x0417e91c,
 };
 
 uint32_t nve0_grhub_code[] = {
-	0x03090ef5,
+	0x031b0ef5,
 /* 0x0004: queue_put */
 	0x9800d898,
 	0x86f001d9,
 	0x0489b808,
 	0xf00c1bf4,
 	0x21f502f7,
-	0x00f802ec,
+	0x00f802fe,
 /* 0x001c: queue_put_next */
 	0xb60798c4,
 	0x8dbb0384,
@@ -247,7 +245,7 @@ uint32_t nve0_grhub_code[] = {
 	0xc800bccf,
 	0x1bf41fcc,
 	0x06a7f0fa,
-	0x010321f5,
+	0x010921f5,
 	0xf840bfcf,
 /* 0x008d: nv_wr32 */
 	0x28b7f100,
@@ -269,63 +267,66 @@ uint32_t nve0_grhub_code[] = {
 	0x0684b604,
 	0xf80080d0,
 /* 0x00c9: wait_donez */
-	0x3c87f100,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d000,
-	0x081887f1,
-	0xd00684b6,
-/* 0x00e2: wait_done_wait_donez */
-	0x87f1008a,
-	0x84b60400,
-	0x0088cf06,
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
 	0xf4888aff,
-	0x87f1f31b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00099,
-/* 0x0103: wait_doneo */
-	0xf100f800,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00099f0,
-	0x87f10089,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f00f00,
+	0x0009d002,
+	0x87f104bd,
 	0x84b60818,
 	0x008ad006,
-/* 0x011c: wait_done_wait_doneo */
+/* 0x0124: wait_doneo_e */
 	0x040087f1,
 	0xcf0684b6,
 	0x8aff0088,
 	0xf30bf488,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0099f094,
-	0xf80089d0,
-/* 0x013d: mmctx_size */
-/* 0x013f: nv_mmctx_size_loop */
-	0x9894bd00,
-	0x85b600e8,
-	0x0180b61a,
-	0xbb0284b6,
-	0xe0b60098,
-	0x04efb804,
-	0xb9eb1bf4,
-	0x00f8029f,
-/* 0x015c: mmctx_xfer */
-	0x083c87f1,
-	0xbd0684b6,
-	0x0199f094,
-	0xf10089d0,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf00f0007,
+	0x09d00203,
+	0xf104bd00,
 	0xb6071087,
 	0x94bd0684,
 	0xf405bbfd,
 	0x8bd0090b,
 	0x0099f000,
-/* 0x0180: mmctx_base_disabled */
+/* 0x018c: mmctx_base_disabled */
 	0xf405eefd,
 	0x8ed00c0b,
 	0xc08fd080,
-/* 0x018f: mmctx_multi_disabled */
+/* 0x019b: mmctx_multi_disabled */
 	0xb70199f0,
 	0xc8010080,
 	0xb4b600ab,
@@ -333,8 +334,8 @@ uint32_t nve0_grhub_code[] = {
 	0xb601aec8,
 	0xbefd11e4,
 	0x008bd005,
-/* 0x01a8: mmctx_exec_loop */
-/* 0x01a8: mmctx_wait_free */
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
 	0xf0008ecf,
 	0x0bf41fe4,
 	0x00ce98fa,
@@ -343,76 +344,77 @@ uint32_t nve0_grhub_code[] = {
 	0x04cdb804,
 	0xc8e81bf4,
 	0x1bf402ab,
-/* 0x01c9: mmctx_fini_wait */
+/* 0x01d5: mmctx_fini_wait */
 	0x008bcf18,
 	0xb01fb4f0,
 	0x1bf410b4,
 	0x02a7f0f7,
 	0xf4c921f4,
-/* 0x01de: mmctx_stop */
+/* 0x01ea: mmctx_stop */
 	0xabc81b0e,
 	0x10b4b600,
 	0xf00cb9f0,
 	0x8bd012b9,
-/* 0x01ed: mmctx_stop_wait */
+/* 0x01f9: mmctx_stop_wait */
 	0x008bcf00,
 	0xf412bbc8,
-/* 0x01f6: mmctx_done */
-	0x87f1fa1b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00199,
-/* 0x0207: strand_wait */
-	0xf900f800,
-	0x02a7f0a0,
-	0xfcc921f4,
-/* 0x0213: strand_pre */
-	0xf100f8a0,
-	0xf04afc87,
-	0x97f00283,
-	0x0089d00c,
-	0x020721f5,
-/* 0x0226: strand_post */
-	0x87f100f8,
-	0x83f04afc,
-	0x0d97f002,
-	0xf50089d0,
-	0xf8020721,
-/* 0x0239: strand_set */
-	0xfca7f100,
-	0x02a3f04f,
-	0x0500aba2,
-	0xd00fc7f0,
-	0xc7f000ac,
-	0x00bcd00b,
-	0x020721f5,
-	0xf000aed0,
-	0xbcd00ac7,
-	0x0721f500,
-/* 0x0263: strand_ctx_init */
-	0xf100f802,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00399f0,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
 	0x21f50089,
-	0xe7f00213,
-	0x3921f503,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f00f00,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
 	0xfca7f102,
 	0x02a3f046,
 	0x0400aba0,
 	0xf040a0d0,
 	0xbcd001c7,
-	0x0721f500,
+	0x1521f500,
 	0x010c9202,
 	0xf000acd0,
 	0xbcd002c7,
-	0x0721f500,
-	0x2621f502,
+	0x1521f500,
+	0x3421f502,
 	0x8087f102,
 	0x0684b608,
 	0xb70089cf,
 	0x95220080,
-/* 0x02ba: ctx_init_strand_loop */
+/* 0x02ca: ctx_init_strand_loop */
 	0x8ed008fe,
 	0x408ed000,
 	0xb6808acf,
@@ -421,73 +423,61 @@ uint32_t nve0_grhub_code[] = {
 	0xb60480b6,
 	0x1bf40192,
 	0x08e4b6e8,
-	0xf1f2efbc,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00399f0,
-	0x00f80089,
-/* 0x02ec: error */
-	0xe7f1e0f9,
-	0xe4b60814,
-	0x00efd006,
-	0x0c1ce7f1,
-	0xf006e4b6,
-	0xefd001f7,
-	0xf8e0fc00,
-/* 0x0309: init */
-	0xfe04bd00,
-	0x07fe0004,
-	0x0017f100,
-	0x0227f012,
-	0xf10012d0,
-	0xfe05b917,
-	0x17f10010,
-	0x10d00400,
-	0x0437f1c0,
-	0x0634b604,
-	0x200327f1,
-	0xf10032d0,
-	0xd0200427,
-	0x27f10132,
-	0x32d0200b,
-	0x0c27f102,
-	0x0732d020,
-	0x0c2427f1,
-	0xb90624b6,
-	0x23d00003,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0x07f100f8,
+	0x03f00500,
+	0x000fd002,
+	0xf7f004bd,
+	0x0007f101,
+	0x0303f007,
+	0xbd000fd0,
+/* 0x031b: init */
+	0xbd00f804,
+	0x0004fe04,
+	0xf10007fe,
+	0xf0120017,
+	0x12d00227,
+	0xb117f100,
+	0x0010fe05,
+	0x040017f1,
+	0xf1c010d0,
+	0xb6040437,
+	0x27f10634,
+	0x32d02003,
 	0x0427f100,
-	0x0023f087,
-	0xb70012d0,
-	0xf0010012,
-	0x12d00427,
-	0x1031f400,
-	0x9604e7f1,
-	0xf440e3f0,
-	0xf1c76821,
-	0x01018090,
-	0x801ff4f0,
-	0x17f0000f,
-	0x041fbb01,
-	0xf10112b6,
-	0xb6040c27,
-	0x21d00624,
-	0x4021d000,
-	0x080027f1,
-	0xcf0624b6,
-	0xf7f00022,
-/* 0x03a9: init_find_chipset */
-	0x08f0b654,
-	0xb800f398,
-	0x0bf40432,
-	0x0034b00b,
-	0xf8f11bf4,
-/* 0x03bd: init_context */
-	0x0017f100,
-	0x02fe5801,
-	0xf003ff58,
-	0x0e8000e3,
-	0x150f8014,
-	0x013d21f5,
+	0x0132d020,
+	0x200b27f1,
+	0xf10232d0,
+	0xd0200c27,
+	0x27f10732,
+	0x24b60c24,
+	0x0003b906,
+	0xf10023d0,
+	0xf0870427,
+	0x12d00023,
+	0x0012b700,
+	0x0427f001,
+	0xf40012d0,
+	0xe7f11031,
+	0xe3f09604,
+	0x6821f440,
+	0x8090f1c7,
+	0xf4f00301,
+	0x020f801f,
+	0xbb0117f0,
+	0x12b6041f,
+	0x0c27f101,
+	0x0624b604,
+	0xd00021d0,
+	0x17f14021,
+	0x0e980100,
+	0x010f9800,
+	0x014721f5,
 	0x070037f1,
 	0x950634b6,
 	0x34d00814,
@@ -498,196 +488,201 @@ uint32_t nve0_grhub_code[] = {
 	0x0815b600,
 	0xb60110b6,
 	0x1fb90814,
-	0x6321f502,
+	0x7121f502,
 	0x001fbb02,
-	0xf1000398,
+	0xf1020398,
 	0xf0200047,
-/* 0x040e: init_gpc */
+/* 0x03f6: init_gpc */
 	0x4ea05043,
 	0x1fb90804,
 	0x8d21f402,
-	0x08004ea0,
-	0xf4022fb9,
-	0x4ea08d21,
-	0xf4bd010c,
-	0xa08d21f4,
-	0xf401044e,
+	0x010c4ea0,
+	0x21f4f4bd,
+	0x044ea08d,
+	0x8d21f401,
+	0x01004ea0,
+	0xf402f7f0,
 	0x4ea08d21,
-	0xf7f00100,
-	0x8d21f402,
-	0x08004ea0,
-/* 0x0440: init_gpc_wait */
-	0xc86821f4,
-	0x0bf41fff,
-	0x044ea0fa,
-	0x6821f408,
-	0xb7001fbb,
-	0xb6800040,
-	0x1bf40132,
-	0x0027f1b4,
-	0x0624b608,
-	0xb74021d0,
-	0xbd080020,
+/* 0x041e: init_gpc_wait */
+	0x21f40800,
+	0x1fffc868,
+	0xa0fa0bf4,
+	0xf408044e,
+	0x1fbb6821,
+	0x0040b700,
+	0x0132b680,
+	0xf1be1bf4,
+	0xf0010007,
+	0x01d00203,
+	0xbd04bd00,
 	0x1f19f014,
-/* 0x0473: main */
-	0xf40021d0,
-	0x28f40031,
-	0x08d7f000,
-	0xf43921f4,
-	0xe4b1f401,
-	0x1bf54001,
-	0x87f100d1,
-	0x84b6083c,
-	0xf094bd06,
-	0x89d00499,
-	0x0017f100,
-	0x0614b60b,
-	0xcf4012cf,
-	0x13c80011,
-	0x7e0bf41f,
+	0x080007f1,
+	0xd00203f0,
+	0x04bd0001,
+/* 0x0458: main */
+	0xf40031f4,
+	0xd7f00028,
+	0x3921f410,
+	0xb1f401f4,
+	0xf54001e4,
+	0xbd00de1b,
+	0x0499f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0b0017f1,
+	0xcf0614b6,
+	0x11cf4012,
+	0x1f13c800,
+	0x00870bf5,
 	0xf41f23c8,
-	0x20f95a0b,
-	0xf10212b9,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00799f0,
-	0x32f40089,
-	0x0231f401,
-	0x07fb21f5,
-	0x085c87f1,
-	0xbd0684b6,
+	0x20f9620b,
+	0xbd0212b9,
 	0x0799f094,
-	0xfc0089d0,
-	0x3c87f120,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d006,
-	0xf50131f4,
-	0xf107fb21,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00699f0,
-	0x0ef40089,
-/* 0x0509: chsw_prev_no_next */
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0xf40132f4,
+	0x21f50231,
+	0x94bd0801,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xfc04bd00,
+	0xf094bd20,
+	0x07f10699,
+	0x03f00f00,
+	0x0009d002,
+	0x31f404bd,
+	0x0121f501,
+	0xf094bd08,
+	0x07f10699,
+	0x03f01700,
+	0x0009d002,
+	0x0ef404bd,
+/* 0x04f9: chsw_prev_no_next */
 	0xb920f931,
 	0x32f40212,
 	0x0232f401,
-	0x07fb21f5,
+	0x080121f5,
 	0x17f120fc,
 	0x14b60b00,
 	0x0012d006,
-/* 0x0527: chsw_no_prev */
+/* 0x0517: chsw_no_prev */
 	0xc8130ef4,
 	0x0bf41f23,
 	0x0131f40d,
 	0xf50232f4,
-/* 0x0537: chsw_done */
-	0xf107fb21,
+/* 0x0527: chsw_done */
+	0xf1080121,
 	0xb60b0c17,
 	0x27f00614,
 	0x0012d001,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0499f094,
-	0xf50089d0,
-/* 0x0557: main_not_ctx_switch */
-	0xb0ff200e,
-	0x1bf401e4,
-	0x02f2b90d,
-	0x078f21f5,
-/* 0x0567: main_not_ctx_chan */
-	0xb0420ef4,
-	0x1bf402e4,
-	0x3c87f12e,
-	0x0684b608,
 	0x99f094bd,
-	0x0089d007,
+	0x0007f104,
+	0x0203f017,
+	0xbd0009d0,
+	0x130ef504,
+/* 0x0549: main_not_ctx_switch */
+	0x01e4b0ff,
+	0xb90d1bf4,
+	0x21f502f2,
+	0x0ef40795,
+/* 0x0559: main_not_ctx_chan */
+	0x02e4b046,
+	0xbd321bf4,
+	0x0799f094,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
 	0xf40132f4,
 	0x21f50232,
-	0x87f107fb,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00799,
-	0x110ef400,
-/* 0x0598: main_not_ctx_save */
-	0xf010ef94,
-	0x21f501f5,
-	0x0ef502ec,
-/* 0x05a6: main_done */
-	0x17f1fed1,
-	0x14b60820,
-	0xf024bd06,
-	0x12d01f29,
-	0xbe0ef500,
-/* 0x05b9: ih */
+	0x94bd0801,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xf404bd00,
+/* 0x058e: main_not_ctx_save */
+	0xef94110e,
+	0x01f5f010,
+	0x02fe21f5,
+	0xfec00ef5,
+/* 0x059c: main_done */
+	0x29f024bd,
+	0x0007f11f,
+	0x0203f008,
+	0xbd0002d0,
+	0xab0ef504,
+/* 0x05b1: ih */
 	0xfe80f9fe,
 	0x80f90188,
 	0xa0f990f9,
 	0xd0f9b0f9,
 	0xf0f9e0f9,
-	0xc4800acf,
-	0x0bf404ab,
-	0x00b7f11d,
-	0x08d7f019,
-	0xcf40becf,
-	0x21f400bf,
-	0x00b0b704,
-	0x01e7f004,
-/* 0x05ef: ih_no_fifo */
-	0xe400bed0,
-	0xf40100ab,
-	0xd7f00d0b,
-	0x01e7f108,
-	0x0421f440,
-/* 0x0600: ih_no_ctxsw */
-	0x0104b7f1,
-	0xabffb0bd,
-	0x0d0bf4b4,
-	0x0c1ca7f1,
-	0xd006a4b6,
-/* 0x0616: ih_no_other */
-	0x0ad000ab,
-	0xfcf0fc40,
-	0xfcd0fce0,
-	0xfca0fcb0,
-	0xfe80fc90,
-	0x80fc0088,
-	0xf80032f4,
-/* 0x0631: ctx_4170s */
-	0x70e7f101,
-	0x40e3f041,
-	0xf410f5f0,
-	0x00f88d21,
-/* 0x0640: ctx_4170w */
-	0x4170e7f1,
-	0xf440e3f0,
-	0xf4f06821,
-	0xf31bf410,
-/* 0x0652: ctx_redswitch */
+	0x0acf04bd,
+	0x04abc480,
+	0xf11d0bf4,
+	0xf01900b7,
+	0xbecf10d7,
+	0x00bfcf40,
+	0xb70421f4,
+	0xf00400b0,
+	0xbed001e7,
+/* 0x05e9: ih_no_fifo */
+	0x00abe400,
+	0x0d0bf401,
+	0xf110d7f0,
+	0xf44001e7,
+/* 0x05fa: ih_no_ctxsw */
+	0xb7f10421,
+	0xb0bd0104,
+	0xf4b4abff,
+	0xa7f10d0b,
+	0xa4b60c1c,
+	0x00abd006,
+/* 0x0610: ih_no_other */
+	0xfc400ad0,
+	0xfce0fcf0,
+	0xfcb0fcd0,
+	0xfc90fca0,
+	0x0088fe80,
+	0x32f480fc,
+/* 0x062b: ctx_4170s */
+	0xf101f800,
+	0xf04170e7,
+	0xf5f040e3,
+	0x8d21f410,
+/* 0x063a: ctx_4170w */
 	0xe7f100f8,
-	0xe4b60614,
-	0x70f7f106,
-	0x00efd002,
-/* 0x0663: ctx_redswitch_delay */
-	0xb608f7f0,
-	0x1bf401f2,
-	0x70f7f1fd,
-	0x00efd007,
-/* 0x0672: ctx_86c */
-	0xe7f100f8,
-	0xe4b6086c,
-	0x00efd006,
-	0x8a14e7f1,
-	0xf440e3f0,
-	0xe7f18d21,
-	0xe3f0a86c,
-	0x8d21f441,
-/* 0x0692: ctx_load */
-	0x87f100f8,
-	0x84b6083c,
-	0xf094bd06,
-	0x89d00599,
-	0x0ca7f000,
+	0xe3f04170,
+	0x6821f440,
+	0xf410f4f0,
+	0x00f8f31b,
+/* 0x064c: ctx_redswitch */
+	0x0614e7f1,
+	0xf106e4b6,
+	0xd00270f7,
+	0xf7f000ef,
+/* 0x065d: ctx_redswitch_delay */
+	0x01f2b608,
+	0xf1fd1bf4,
+	0xd00770f7,
+	0x00f800ef,
+/* 0x066c: ctx_86c */
+	0x086ce7f1,
+	0xd006e4b6,
+	0xe7f100ef,
+	0xe3f08a14,
+	0x8d21f440,
+	0xa86ce7f1,
+	0xf441e3f0,
+	0x00f88d21,
+/* 0x068c: ctx_load */
+	0x99f094bd,
+	0x0007f105,
+	0x0203f00f,
+	0xbd0009d0,
+	0x0ca7f004,
 	0xf1c921f4,
 	0xb60a2417,
 	0x10d00614,
@@ -697,162 +692,227 @@ uint32_t nve0_grhub_code[] = {
 	0xb60a0c17,
 	0x47f00614,
 	0x0012d007,
-/* 0x06cb: ctx_chan_wait_0 */
+/* 0x06c7: ctx_chan_wait_0 */
 	0xcf4014d0,
 	0x44f04014,
 	0xfa1bf41f,
 	0xfe0032d0,
 	0x2af0000b,
 	0x0424b61f,
-	0xf10220b6,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00899f0,
-	0x17f10089,
-	0x14b60a04,
-	0x0012d006,
-	0x0a2017f1,
-	0xf00614b6,
-	0x23f10227,
-	0x12d08000,
-	0x1017f000,
-	0x030027f1,
-	0xfa0223f0,
-	0x03f80512,
-	0x085c87f1,
-	0xbd0684b6,
+	0xbd0220b6,
 	0x0899f094,
-	0x980089d0,
-	0x14b6c101,
-	0xc0029818,
+	0x0f0007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0a0417f1,
+	0xd00614b6,
+	0x17f10012,
+	0x14b60a20,
+	0x0227f006,
+	0x800023f1,
+	0xf00012d0,
+	0x27f11017,
+	0x23f00200,
+	0x0512fa02,
+	0x94bd03f8,
+	0xf10899f0,
+	0xf0170007,
+	0x09d00203,
+	0x9804bd00,
+	0x14b68101,
+	0x80029818,
 	0xfd0825b6,
 	0x01800512,
-	0x3c87f116,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d009,
-	0x0a0427f1,
-	0xd00624b6,
-	0x27f00021,
-	0x2017f101,
-	0x0614b60a,
-	0xf10012d0,
-	0xf0020017,
+	0xf094bd16,
+	0x07f10999,
+	0x03f00f00,
+	0x0009d002,
+	0x27f104bd,
+	0x24b60a04,
+	0x0021d006,
+	0xf10127f0,
+	0xb60a2017,
+	0x12d00614,
+	0x0017f100,
+	0x0613f001,
+	0xf80501fa,
+	0xf094bd03,
+	0x07f10999,
+	0x03f01700,
+	0x0009d002,
+	0x94bd04bd,
+	0xf10599f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0795: ctx_chan */
+	0x8c21f500,
+	0x0ca7f006,
+	0xf1c921f4,
+	0xb60a1017,
+	0x27f00614,
+	0x0012d005,
+/* 0x07ac: ctx_chan_wait */
+	0xfd0012cf,
+	0x1bf40522,
+/* 0x07b7: ctx_mmio_exec */
+	0x9800f8fa,
+	0x27f14103,
+	0x24b60a04,
+	0x0023d006,
+/* 0x07c6: ctx_mmio_loop */
+	0x34c434bd,
+	0x0f1bf4ff,
+	0x020057f1,
+	0xfa0653f0,
+	0x03f80535,
+/* 0x07d8: ctx_mmio_pull */
+	0x98804e98,
+	0x21f4814f,
+	0x0830b68d,
+	0xf40112b6,
+/* 0x07ea: ctx_mmio_done */
+	0x0398df1b,
+	0x0023d016,
+	0xf1400080,
+	0xf0010017,
 	0x01fa0613,
-	0xf103f805,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00999f0,
-	0x87f10089,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00599,
-/* 0x078f: ctx_chan */
-	0xf500f800,
-	0xf0069221,
-	0x21f40ca7,
-	0x1017f1c9,
-	0x0614b60a,
-	0xd00527f0,
-/* 0x07a6: ctx_chan_wait */
-	0x12cf0012,
-	0x0522fd00,
-	0xf8fa1bf4,
-/* 0x07b1: ctx_mmio_exec */
-	0x81039800,
-	0x0a0427f1,
+	0xf803f806,
+/* 0x0801: ctx_xfer */
+	0x00f7f100,
+	0x06f4b60c,
+	0xd004e7f0,
+/* 0x080e: ctx_xfer_idle */
+	0xfecf80fe,
+	0x00e4f100,
+	0xf91bf420,
+	0xf40611f4,
+/* 0x081e: ctx_xfer_pre */
+	0xf7f00d02,
+	0x6c21f510,
+	0x1c11f406,
+/* 0x0828: ctx_xfer_pre_load */
+	0xf502f7f0,
+	0xf5062b21,
+	0xf5063a21,
+	0xbd064c21,
+	0x2b21f5f4,
+	0x8c21f506,
+/* 0x0841: ctx_xfer_exec */
+	0x16019806,
+	0x041427f1,
 	0xd00624b6,
-	0x34bd0023,
-/* 0x07c0: ctx_mmio_loop */
-	0xf4ff34c4,
-	0x57f10f1b,
-	0x53f00300,
-	0x0535fa06,
-/* 0x07d2: ctx_mmio_pull */
-	0x4e9803f8,
-	0xc14f98c0,
+	0xe7f10020,
+	0xe3f0a500,
+	0x021fb941,
 	0xb68d21f4,
-	0x12b60830,
-	0xdf1bf401,
-/* 0x07e4: ctx_mmio_done */
-	0xd0160398,
-	0x00800023,
-	0x0017f180,
-	0x0613f002,
-	0xf80601fa,
-/* 0x07fb: ctx_xfer */
-	0xf100f803,
-	0xb60c00f7,
-	0xe7f006f4,
-	0x80fed004,
-/* 0x0808: ctx_xfer_idle */
-	0xf100fecf,
-	0xf42000e4,
-	0x11f4f91b,
-	0x0d02f406,
-/* 0x0818: ctx_xfer_pre */
-	0xf510f7f0,
-	0xf4067221,
-/* 0x0822: ctx_xfer_pre_load */
-	0xf7f01c11,
-	0x3121f502,
-	0x4021f506,
-	0x5221f506,
-	0xf5f4bd06,
-	0xf5063121,
-/* 0x083b: ctx_xfer_exec */
-	0x98069221,
-	0x27f11601,
-	0x24b60414,
-	0x0020d006,
-	0xa500e7f1,
-	0xb941e3f0,
-	0x21f4021f,
-	0x04e0b68d,
-	0xf001fcf0,
-	0x24b6022c,
-	0x05f2fd01,
-	0xf18d21f4,
-	0xf04afc17,
-	0x27f00213,
-	0x0012d00c,
-	0x020721f5,
-	0x47fc27f1,
-	0xd00223f0,
-	0x2cf00020,
-	0x0320b601,
-	0xf00012d0,
-	0xa5f001ac,
-	0x00b7f006,
-	0x98140c98,
-	0xe7f0150d,
-	0x5c21f500,
-	0x08a7f001,
-	0x010321f5,
-	0x020721f5,
-	0xf02201f4,
-	0x21f40ca7,
-	0x1017f1c9,
-	0x0614b60a,
-	0xd00527f0,
-/* 0x08c2: ctx_xfer_post_save_wait */
-	0x12cf0012,
-	0x0522fd00,
-	0xf4fa1bf4,
-/* 0x08ce: ctx_xfer_post */
-	0xf7f02e02,
-	0x3121f502,
+	0xfcf004e0,
+	0x022cf001,
+	0xfd0124b6,
+	0x21f405f2,
+	0xfc17f18d,
+	0x0213f04a,
+	0xd00c27f0,
+	0x21f50012,
+	0x27f10215,
+	0x23f047fc,
+	0x0020d002,
+	0xb6012cf0,
+	0x12d00320,
+	0x01acf000,
+	0xf006a5f0,
+	0x0c9800b7,
+	0x010d9800,
+	0xf500e7f0,
+	0xf0016621,
+	0x21f508a7,
+	0x21f50109,
+	0x01f40215,
+	0x0ca7f022,
+	0xf1c921f4,
+	0xb60a1017,
+	0x27f00614,
+	0x0012d005,
+/* 0x08c8: ctx_xfer_post_save_wait */
+	0xfd0012cf,
+	0x1bf40522,
+	0x2e02f4fa,
+/* 0x08d4: ctx_xfer_post */
+	0xf502f7f0,
+	0xbd062b21,
+	0x6c21f5f4,
+	0x3421f506,
+	0x3a21f502,
 	0xf5f4bd06,
-	0xf5067221,
-	0xf5022621,
-	0xbd064021,
-	0x3121f5f4,
-	0x1011f406,
-	0xfd800198,
-	0x0bf40511,
-	0xb121f507,
-/* 0x08f9: ctx_xfer_no_post_mmio */
-/* 0x08f9: ctx_xfer_done */
-	0x0000f807,
+	0xf4062b21,
+	0x01981011,
+	0x0511fd40,
+	0xf5070bf4,
+/* 0x08ff: ctx_xfer_no_post_mmio */
+/* 0x08ff: ctx_xfer_done */
+	0xf807b721,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
 	0x00000000,
 };
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvf0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvf0.fuc
new file mode 100644
index 000000000000..ec42ed29b50d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvf0.fuc
@@ -0,0 +1,40 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#define CHIPSET GK110
+#include "macros.fuc"
+
+.section #nvf0_grhub_data
+#define INCLUDE_DATA
+#include "com.fuc"
+#include "hub.fuc"
+#undef INCLUDE_DATA
+
+.section #nvf0_grhub_code
+#define INCLUDE_CODE
+bra #init
+#include "com.fuc"
+#include "hub.fuc"
+.align 256
+#undef INCLUDE_CODE
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvf0.fuc.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvf0.fuc.h
new file mode 100644
index 000000000000..438506d14749
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/hubnvf0.fuc.h
@@ -0,0 +1,918 @@
+uint32_t nvf0_grhub_data[] = {
+/* 0x0000: hub_mmio_list_head */
+	0x00000300,
+/* 0x0004: hub_mmio_list_tail */
+	0x00000304,
+/* 0x0008: gpc_count */
+	0x00000000,
+/* 0x000c: rop_count */
+	0x00000000,
+/* 0x0010: cmd_queue */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0058: ctx_current */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0100: chan_data */
+/* 0x0100: chan_mmio_count */
+	0x00000000,
+/* 0x0104: chan_mmio_address */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0200: xfer_data */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+/* 0x0300: hub_mmio_list_base */
+	0x0417e91c,
+};
+
+uint32_t nvf0_grhub_code[] = {
+	0x031b0ef5,
+/* 0x0004: queue_put */
+	0x9800d898,
+	0x86f001d9,
+	0x0489b808,
+	0xf00c1bf4,
+	0x21f502f7,
+	0x00f802fe,
+/* 0x001c: queue_put_next */
+	0xb60798c4,
+	0x8dbb0384,
+	0x0880b600,
+	0x80008e80,
+	0x90b6018f,
+	0x0f94f001,
+	0xf801d980,
+/* 0x0039: queue_get */
+	0x0131f400,
+	0x9800d898,
+	0x89b801d9,
+	0x210bf404,
+	0xb60789c4,
+	0x9dbb0394,
+	0x0890b600,
+	0x98009e98,
+	0x80b6019f,
+	0x0f84f001,
+	0xf400d880,
+/* 0x0066: queue_get_done */
+	0x00f80132,
+/* 0x0068: nv_rd32 */
+	0x0728b7f1,
+	0xb906b4b6,
+	0xc9f002ec,
+	0x00bcd01f,
+/* 0x0078: nv_rd32_wait */
+	0xc800bccf,
+	0x1bf41fcc,
+	0x06a7f0fa,
+	0x010921f5,
+	0xf840bfcf,
+/* 0x008d: nv_wr32 */
+	0x28b7f100,
+	0x06b4b607,
+	0xb980bfd0,
+	0xc9f002ec,
+	0x1ec9f01f,
+/* 0x00a3: nv_wr32_wait */
+	0xcf00bcd0,
+	0xccc800bc,
+	0xfa1bf41f,
+/* 0x00ae: watchdog_reset */
+	0x87f100f8,
+	0x84b60430,
+	0x1ff9f006,
+	0xf8008fd0,
+/* 0x00bd: watchdog_clear */
+	0x3087f100,
+	0x0684b604,
+	0xf80080d0,
+/* 0x00c9: wait_donez */
+	0xf094bd00,
+	0x07f10099,
+	0x03f03700,
+	0x0009d002,
+	0x07f104bd,
+	0x03f00600,
+	0x000ad002,
+/* 0x00e6: wait_donez_ne */
+	0x87f104bd,
+	0x83f00000,
+	0x0088cf01,
+	0xf4888aff,
+	0x94bdf31b,
+	0xf10099f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0109: wait_doneo */
+	0xf094bd00,
+	0x07f10099,
+	0x03f03700,
+	0x0009d002,
+	0x87f104bd,
+	0x84b60818,
+	0x008ad006,
+/* 0x0124: wait_doneo_e */
+	0x040087f1,
+	0xcf0684b6,
+	0x8aff0088,
+	0xf30bf488,
+	0x99f094bd,
+	0x0007f100,
+	0x0203f017,
+	0xbd0009d0,
+/* 0x0147: mmctx_size */
+	0xbd00f804,
+/* 0x0149: nv_mmctx_size_loop */
+	0x00e89894,
+	0xb61a85b6,
+	0x84b60180,
+	0x0098bb02,
+	0xb804e0b6,
+	0x1bf404ef,
+	0x029fb9eb,
+/* 0x0166: mmctx_xfer */
+	0x94bd00f8,
+	0xf10199f0,
+	0xf0370007,
+	0x09d00203,
+	0xf104bd00,
+	0xb6071087,
+	0x94bd0684,
+	0xf405bbfd,
+	0x8bd0090b,
+	0x0099f000,
+/* 0x018c: mmctx_base_disabled */
+	0xf405eefd,
+	0x8ed00c0b,
+	0xc08fd080,
+/* 0x019b: mmctx_multi_disabled */
+	0xb70199f0,
+	0xc8010080,
+	0xb4b600ab,
+	0x0cb9f010,
+	0xb601aec8,
+	0xbefd11e4,
+	0x008bd005,
+/* 0x01b4: mmctx_exec_loop */
+/* 0x01b4: mmctx_wait_free */
+	0xf0008ecf,
+	0x0bf41fe4,
+	0x00ce98fa,
+	0xd005e9fd,
+	0xc0b6c08e,
+	0x04cdb804,
+	0xc8e81bf4,
+	0x1bf402ab,
+/* 0x01d5: mmctx_fini_wait */
+	0x008bcf18,
+	0xb01fb4f0,
+	0x1bf410b4,
+	0x02a7f0f7,
+	0xf4c921f4,
+/* 0x01ea: mmctx_stop */
+	0xabc81b0e,
+	0x10b4b600,
+	0xf00cb9f0,
+	0x8bd012b9,
+/* 0x01f9: mmctx_stop_wait */
+	0x008bcf00,
+	0xf412bbc8,
+/* 0x0202: mmctx_done */
+	0x94bdfa1b,
+	0xf10199f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0215: strand_wait */
+	0xf0a0f900,
+	0x21f402a7,
+	0xf8a0fcc9,
+/* 0x0221: strand_pre */
+	0xfc87f100,
+	0x0283f04a,
+	0xd00c97f0,
+	0x21f50089,
+	0x00f80215,
+/* 0x0234: strand_post */
+	0x4afc87f1,
+	0xf00283f0,
+	0x89d00d97,
+	0x1521f500,
+/* 0x0247: strand_set */
+	0xf100f802,
+	0xf04ffca7,
+	0xaba202a3,
+	0xc7f00500,
+	0x00acd00f,
+	0xd00bc7f0,
+	0x21f500bc,
+	0xaed00215,
+	0x0ac7f000,
+	0xf500bcd0,
+	0xf8021521,
+/* 0x0271: strand_ctx_init */
+	0xf094bd00,
+	0x07f10399,
+	0x03f03700,
+	0x0009d002,
+	0x21f504bd,
+	0xe7f00221,
+	0x4721f503,
+	0xfca7f102,
+	0x02a3f046,
+	0x0400aba0,
+	0xf040a0d0,
+	0xbcd001c7,
+	0x1521f500,
+	0x010c9202,
+	0xf000acd0,
+	0xbcd002c7,
+	0x1521f500,
+	0x3421f502,
+	0x8087f102,
+	0x0684b608,
+	0xb70089cf,
+	0x95220080,
+/* 0x02ca: ctx_init_strand_loop */
+	0x8ed008fe,
+	0x408ed000,
+	0xb6808acf,
+	0xa0b606a5,
+	0x00eabb01,
+	0xb60480b6,
+	0x1bf40192,
+	0x08e4b6e8,
+	0xbdf2efbc,
+	0x0399f094,
+	0x170007f1,
+	0xd00203f0,
+	0x04bd0009,
+/* 0x02fe: error */
+	0x07f100f8,
+	0x03f00500,
+	0x000fd002,
+	0xf7f004bd,
+	0x0007f101,
+	0x0303f007,
+	0xbd000fd0,
+/* 0x031b: init */
+	0xbd00f804,
+	0x0004fe04,
+	0xf10007fe,
+	0xf0120017,
+	0x12d00227,
+	0xb117f100,
+	0x0010fe05,
+	0x040017f1,
+	0xf1c010d0,
+	0xb6040437,
+	0x27f10634,
+	0x32d02003,
+	0x0427f100,
+	0x0132d020,
+	0x200b27f1,
+	0xf10232d0,
+	0xd0200c27,
+	0x27f10732,
+	0x24b60c24,
+	0x0003b906,
+	0xf10023d0,
+	0xf0870427,
+	0x12d00023,
+	0x0012b700,
+	0x0427f001,
+	0xf40012d0,
+	0xe7f11031,
+	0xe3f09604,
+	0x6821f440,
+	0x8090f1c7,
+	0xf4f00301,
+	0x020f801f,
+	0xbb0117f0,
+	0x12b6041f,
+	0x0c27f101,
+	0x0624b604,
+	0xd00021d0,
+	0x17f14021,
+	0x0e980100,
+	0x010f9800,
+	0x014721f5,
+	0x070037f1,
+	0x950634b6,
+	0x34d00814,
+	0x4034d000,
+	0x130030b7,
+	0xb6001fbb,
+	0x3fd002f5,
+	0x0815b600,
+	0xb60110b6,
+	0x1fb90814,
+	0x7121f502,
+	0x001fbb02,
+	0xf1020398,
+	0xf0200047,
+/* 0x03f6: init_gpc */
+	0x4ea05043,
+	0x1fb90804,
+	0x8d21f402,
+	0x010c4ea0,
+	0x21f4f4bd,
+	0x044ea08d,
+	0x8d21f401,
+	0x01004ea0,
+	0xf402f7f0,
+	0x4ea08d21,
+/* 0x041e: init_gpc_wait */
+	0x21f40800,
+	0x1fffc868,
+	0xa0fa0bf4,
+	0xf408044e,
+	0x1fbb6821,
+	0x0040b700,
+	0x0132b680,
+	0xf1be1bf4,
+	0xf0010007,
+	0x01d00203,
+	0xbd04bd00,
+	0x1f19f014,
+	0x300007f1,
+	0xd00203f0,
+	0x04bd0001,
+/* 0x0458: main */
+	0xf40031f4,
+	0xd7f00028,
+	0x3921f410,
+	0xb1f401f4,
+	0xf54001e4,
+	0xbd00de1b,
+	0x0499f094,
+	0x370007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0b0017f1,
+	0xcf0614b6,
+	0x11cf4012,
+	0x1f13c800,
+	0x00870bf5,
+	0xf41f23c8,
+	0x20f9620b,
+	0xbd0212b9,
+	0x0799f094,
+	0x370007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0xf40132f4,
+	0x21f50231,
+	0x94bd0801,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xfc04bd00,
+	0xf094bd20,
+	0x07f10699,
+	0x03f03700,
+	0x0009d002,
+	0x31f404bd,
+	0x0121f501,
+	0xf094bd08,
+	0x07f10699,
+	0x03f01700,
+	0x0009d002,
+	0x0ef404bd,
+/* 0x04f9: chsw_prev_no_next */
+	0xb920f931,
+	0x32f40212,
+	0x0232f401,
+	0x080121f5,
+	0x17f120fc,
+	0x14b60b00,
+	0x0012d006,
+/* 0x0517: chsw_no_prev */
+	0xc8130ef4,
+	0x0bf41f23,
+	0x0131f40d,
+	0xf50232f4,
+/* 0x0527: chsw_done */
+	0xf1080121,
+	0xb60b0c17,
+	0x27f00614,
+	0x0012d001,
+	0x99f094bd,
+	0x0007f104,
+	0x0203f017,
+	0xbd0009d0,
+	0x130ef504,
+/* 0x0549: main_not_ctx_switch */
+	0x01e4b0ff,
+	0xb90d1bf4,
+	0x21f502f2,
+	0x0ef40795,
+/* 0x0559: main_not_ctx_chan */
+	0x02e4b046,
+	0xbd321bf4,
+	0x0799f094,
+	0x370007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0xf40132f4,
+	0x21f50232,
+	0x94bd0801,
+	0xf10799f0,
+	0xf0170007,
+	0x09d00203,
+	0xf404bd00,
+/* 0x058e: main_not_ctx_save */
+	0xef94110e,
+	0x01f5f010,
+	0x02fe21f5,
+	0xfec00ef5,
+/* 0x059c: main_done */
+	0x29f024bd,
+	0x0007f11f,
+	0x0203f030,
+	0xbd0002d0,
+	0xab0ef504,
+/* 0x05b1: ih */
+	0xfe80f9fe,
+	0x80f90188,
+	0xa0f990f9,
+	0xd0f9b0f9,
+	0xf0f9e0f9,
+	0x0acf04bd,
+	0x04abc480,
+	0xf11d0bf4,
+	0xf01900b7,
+	0xbecf10d7,
+	0x00bfcf40,
+	0xb70421f4,
+	0xf00400b0,
+	0xbed001e7,
+/* 0x05e9: ih_no_fifo */
+	0x00abe400,
+	0x0d0bf401,
+	0xf110d7f0,
+	0xf44001e7,
+/* 0x05fa: ih_no_ctxsw */
+	0xb7f10421,
+	0xb0bd0104,
+	0xf4b4abff,
+	0xa7f10d0b,
+	0xa4b60c1c,
+	0x00abd006,
+/* 0x0610: ih_no_other */
+	0xfc400ad0,
+	0xfce0fcf0,
+	0xfcb0fcd0,
+	0xfc90fca0,
+	0x0088fe80,
+	0x32f480fc,
+/* 0x062b: ctx_4170s */
+	0xf101f800,
+	0xf04170e7,
+	0xf5f040e3,
+	0x8d21f410,
+/* 0x063a: ctx_4170w */
+	0xe7f100f8,
+	0xe3f04170,
+	0x6821f440,
+	0xf410f4f0,
+	0x00f8f31b,
+/* 0x064c: ctx_redswitch */
+	0x0614e7f1,
+	0xf106e4b6,
+	0xd00270f7,
+	0xf7f000ef,
+/* 0x065d: ctx_redswitch_delay */
+	0x01f2b608,
+	0xf1fd1bf4,
+	0xd00770f7,
+	0x00f800ef,
+/* 0x066c: ctx_86c */
+	0x086ce7f1,
+	0xd006e4b6,
+	0xe7f100ef,
+	0xe3f08a14,
+	0x8d21f440,
+	0xa86ce7f1,
+	0xf441e3f0,
+	0x00f88d21,
+/* 0x068c: ctx_load */
+	0x99f094bd,
+	0x0007f105,
+	0x0203f037,
+	0xbd0009d0,
+	0x0ca7f004,
+	0xf1c921f4,
+	0xb60a2417,
+	0x10d00614,
+	0x0037f100,
+	0x0634b60b,
+	0xf14032d0,
+	0xb60a0c17,
+	0x47f00614,
+	0x0012d007,
+/* 0x06c7: ctx_chan_wait_0 */
+	0xcf4014d0,
+	0x44f04014,
+	0xfa1bf41f,
+	0xfe0032d0,
+	0x2af0000b,
+	0x0424b61f,
+	0xbd0220b6,
+	0x0899f094,
+	0x370007f1,
+	0xd00203f0,
+	0x04bd0009,
+	0x0a0417f1,
+	0xd00614b6,
+	0x17f10012,
+	0x14b60a20,
+	0x0227f006,
+	0x800023f1,
+	0xf00012d0,
+	0x27f11017,
+	0x23f00200,
+	0x0512fa02,
+	0x94bd03f8,
+	0xf10899f0,
+	0xf0170007,
+	0x09d00203,
+	0x9804bd00,
+	0x14b68101,
+	0x80029818,
+	0xfd0825b6,
+	0x01800512,
+	0xf094bd16,
+	0x07f10999,
+	0x03f03700,
+	0x0009d002,
+	0x27f104bd,
+	0x24b60a04,
+	0x0021d006,
+	0xf10127f0,
+	0xb60a2017,
+	0x12d00614,
+	0x0017f100,
+	0x0613f001,
+	0xf80501fa,
+	0xf094bd03,
+	0x07f10999,
+	0x03f01700,
+	0x0009d002,
+	0x94bd04bd,
+	0xf10599f0,
+	0xf0170007,
+	0x09d00203,
+	0xf804bd00,
+/* 0x0795: ctx_chan */
+	0x8c21f500,
+	0x0ca7f006,
+	0xf1c921f4,
+	0xb60a1017,
+	0x27f00614,
+	0x0012d005,
+/* 0x07ac: ctx_chan_wait */
+	0xfd0012cf,
+	0x1bf40522,
+/* 0x07b7: ctx_mmio_exec */
+	0x9800f8fa,
+	0x27f14103,
+	0x24b60a04,
+	0x0023d006,
+/* 0x07c6: ctx_mmio_loop */
+	0x34c434bd,
+	0x0f1bf4ff,
+	0x020057f1,
+	0xfa0653f0,
+	0x03f80535,
+/* 0x07d8: ctx_mmio_pull */
+	0x98804e98,
+	0x21f4814f,
+	0x0830b68d,
+	0xf40112b6,
+/* 0x07ea: ctx_mmio_done */
+	0x0398df1b,
+	0x0023d016,
+	0xf1400080,
+	0xf0010017,
+	0x01fa0613,
+	0xf803f806,
+/* 0x0801: ctx_xfer */
+	0x00f7f100,
+	0x06f4b60c,
+	0xd004e7f0,
+/* 0x080e: ctx_xfer_idle */
+	0xfecf80fe,
+	0x00e4f100,
+	0xf91bf420,
+	0xf40611f4,
+/* 0x081e: ctx_xfer_pre */
+	0xf7f00d02,
+	0x6c21f510,
+	0x1c11f406,
+/* 0x0828: ctx_xfer_pre_load */
+	0xf502f7f0,
+	0xf5062b21,
+	0xf5063a21,
+	0xbd064c21,
+	0x2b21f5f4,
+	0x8c21f506,
+/* 0x0841: ctx_xfer_exec */
+	0x16019806,
+	0x041427f1,
+	0xd00624b6,
+	0xe7f10020,
+	0xe3f0a500,
+	0x021fb941,
+	0xb68d21f4,
+	0xfcf004e0,
+	0x022cf001,
+	0xfd0124b6,
+	0x21f405f2,
+	0xfc17f18d,
+	0x0213f04a,
+	0xd00c27f0,
+	0x21f50012,
+	0x27f10215,
+	0x23f047fc,
+	0x0020d002,
+	0xb6012cf0,
+	0x12d00320,
+	0x01acf000,
+	0xf006a5f0,
+	0x0c9800b7,
+	0x010d9800,
+	0xf500e7f0,
+	0xf0016621,
+	0x21f508a7,
+	0x21f50109,
+	0x01f40215,
+	0x0ca7f022,
+	0xf1c921f4,
+	0xb60a1017,
+	0x27f00614,
+	0x0012d005,
+/* 0x08c8: ctx_xfer_post_save_wait */
+	0xfd0012cf,
+	0x1bf40522,
+	0x2e02f4fa,
+/* 0x08d4: ctx_xfer_post */
+	0xf502f7f0,
+	0xbd062b21,
+	0x6c21f5f4,
+	0x3421f506,
+	0x3a21f502,
+	0xf5f4bd06,
+	0xf4062b21,
+	0x01981011,
+	0x0511fd40,
+	0xf5070bf4,
+/* 0x08ff: ctx_xfer_no_post_mmio */
+/* 0x08ff: ctx_xfer_done */
+	0xf807b721,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/macros.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/macros.fuc
new file mode 100644
index 000000000000..33a5a82eccbd
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/macros.fuc
@@ -0,0 +1,89 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "os.h"
+
+#define GF100 0xc0
+#define GF117 0xd7
+#define GK100 0xe0
+#define GK110 0xf0
+
+#define NV_PGRAPH_FECS_SIGNAL                                          0x409400
+#if CHIPSET < GK110
+#define NV_PGRAPH_FECS_CC_SCRATCH_VAL(n)                    ((n) * 4 + 0x409800)
+#define NV_PGRAPH_FECS_CC_SCRATCH_SET(n)                    ((n) * 4 + 0x409820)
+#define NV_PGRAPH_FECS_CC_SCRATCH_CLR(n)                    ((n) * 4 + 0x409840)
+#else
+#define NV_PGRAPH_FECS_CC_SCRATCH_VAL(n)                    ((n) * 4 + 0x409800)
+#define NV_PGRAPH_FECS_CC_SCRATCH_CLR(n)                    ((n) * 4 + 0x409840)
+#define NV_PGRAPH_FECS_CC_SCRATCH_SET(n)                    ((n) * 4 + 0x4098c0)
+#endif
+#define NV_PGRAPH_FECS_INTR_UP_SET                                     0x409c1c
+
+#if CHIPSET < GK110
+#define NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_VAL(n)              ((n) * 4 + 0x41a800)
+#define NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_SET(n)              ((n) * 4 + 0x41a820)
+#define NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_CLR(n)              ((n) * 4 + 0x41a840)
+#else
+#define NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_VAL(n)              ((n) * 4 + 0x41a800)
+#define NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_CLR(n)              ((n) * 4 + 0x41a840)
+#define NV_PGRAPH_GPCX_GPCCS_CC_SCRATCH_SET(n)              ((n) * 4 + 0x41a8c0)
+#endif
+
+#define mmctx_data(r,c) .b32 (((c - 1) << 26) | r)
+#define queue_init      .skip 72 // (2 * 4) + ((8 * 4) * 2)
+
+#define T_WAIT    0
+#define T_MMCTX   1
+#define T_STRWAIT 2
+#define T_STRINIT 3
+#define T_AUTO    4
+#define T_CHAN    5
+#define T_LOAD    6
+#define T_SAVE    7
+#define T_LCHAN   8
+#define T_LCTXH   9
+
+#define nv_mkmm(rv,r) /*
+*/	movw rv  ((r) & 0x0000fffc) /*
+*/	sethi rv ((r) & 0x00ff0000)
+#define nv_mkio(rv,r,i) /*
+*/	nv_mkmm(rv, (((r) & 0xffc) << 6) | ((i) << 2))
+
+#define nv_iord(rv,r,i) /*
+*/	nv_mkio(rv,r,i) /*
+*/	iord rv I[rv]
+#define nv_iowr(r,i,rv) /*
+*/	nv_mkio($r0,r,i) /*
+*/	iowr I[$r0] rv /*
+*/	clear b32 $r0
+
+#define trace_set(bit) /*
+*/	clear b32 $r9 /*
+*/	bset $r9 bit /*
+*/	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_SET(7), 0, $r9)
+#define trace_clr(bit) /*
+*/	clear b32 $r9 /*
+*/	bset $r9 bit /*
+*/	nv_iowr(NV_PGRAPH_FECS_CC_SCRATCH_CLR(7), 0, $r9)
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/nve0.fuc b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/nve0.fuc
deleted file mode 100644
index f16a5d53319d..000000000000
--- a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/nve0.fuc
+++ /dev/null
@@ -1,400 +0,0 @@
-/* fuc microcode util functions for nve0 PGRAPH
- *
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-define(`mmctx_data', `.b32 eval((($2 - 1) << 26) | $1)')
-define(`queue_init', `.skip eval((2 * 4) + ((8 * 4) * 2))')
-
-ifdef(`include_code', `
-// Error codes
-define(`E_BAD_COMMAND', 0x01)
-define(`E_CMD_OVERFLOW', 0x02)
-
-// Util macros to help with debugging ucode hangs etc
-define(`T_WAIT', 0)
-define(`T_MMCTX', 1)
-define(`T_STRWAIT', 2)
-define(`T_STRINIT', 3)
-define(`T_AUTO', 4)
-define(`T_CHAN', 5)
-define(`T_LOAD', 6)
-define(`T_SAVE', 7)
-define(`T_LCHAN', 8)
-define(`T_LCTXH', 9)
-
-define(`trace_set', `
-	mov $r8 0x83c
-	shl b32 $r8 6
-	clear b32 $r9
-	bset $r9 $1
-	iowr I[$r8 + 0x000] $r9		// CC_SCRATCH[7]
-')
-
-define(`trace_clr', `
-	mov $r8 0x85c
-	shl b32 $r8 6
-	clear b32 $r9
-	bset $r9 $1
-	iowr I[$r8 + 0x000] $r9		// CC_SCRATCH[7]
-')
-
-// queue_put - add request to queue
-//
-// In : $r13 queue pointer
-//	$r14 command
-//	$r15 data
-//
-queue_put:
-	// make sure we have space..
-	ld b32 $r8 D[$r13 + 0x0]	// GET
-	ld b32 $r9 D[$r13 + 0x4]	// PUT
-	xor $r8 8
-	cmpu b32 $r8 $r9
-	bra ne #queue_put_next
-		mov $r15 E_CMD_OVERFLOW
-		call #error
-		ret
-
-	// store cmd/data on queue
-	queue_put_next:
-	and $r8 $r9 7
-	shl b32 $r8 3
-	add b32 $r8 $r13
-	add b32 $r8 8
-	st b32 D[$r8 + 0x0] $r14
-	st b32 D[$r8 + 0x4] $r15
-
-	// update PUT
-	add b32 $r9 1
-	and $r9 0xf
-	st b32 D[$r13 + 0x4] $r9
-	ret
-
-// queue_get - fetch request from queue
-//
-// In : $r13 queue pointer
-//
-// Out:	$p1  clear on success (data available)
-//	$r14 command
-// 	$r15 data
-//
-queue_get:
-	bset $flags $p1
-	ld b32 $r8 D[$r13 + 0x0]	// GET
-	ld b32 $r9 D[$r13 + 0x4]	// PUT
-	cmpu b32 $r8 $r9
-	bra e #queue_get_done
-		// fetch first cmd/data pair
-		and $r9 $r8 7
-		shl b32 $r9 3
-		add b32 $r9 $r13
-		add b32 $r9 8
-		ld b32 $r14 D[$r9 + 0x0]
-		ld b32 $r15 D[$r9 + 0x4]
-
-		// update GET
-		add b32 $r8 1
-		and $r8 0xf
-		st b32 D[$r13 + 0x0] $r8
-		bclr $flags $p1
-queue_get_done:
-	ret
-
-// nv_rd32 - read 32-bit value from nv register
-//
-// In : $r14 register
-// Out: $r15 value
-//
-nv_rd32:
-	mov $r11 0x728
-	shl b32 $r11 6
-	mov b32 $r12 $r14
-	bset $r12 31			// MMIO_CTRL_PENDING
-	iowr I[$r11 + 0x000] $r12	// MMIO_CTRL
-	nv_rd32_wait:
-		iord $r12 I[$r11 + 0x000]
-		xbit $r12 $r12 31
-		bra ne #nv_rd32_wait
-	mov $r10 6			// DONE_MMIO_RD
-	call #wait_doneo
-	iord $r15 I[$r11 + 0x100]	// MMIO_RDVAL
-	ret
-
-// nv_wr32 - write 32-bit value to nv register
-//
-// In : $r14 register
-//      $r15 value
-//
-nv_wr32:
-	mov $r11 0x728
-	shl b32 $r11 6
-	iowr I[$r11 + 0x200] $r15	// MMIO_WRVAL
-	mov b32 $r12 $r14
-	bset $r12 31			// MMIO_CTRL_PENDING
-	bset $r12 30			// MMIO_CTRL_WRITE
-	iowr I[$r11 + 0x000] $r12	// MMIO_CTRL
-	nv_wr32_wait:
-		iord $r12 I[$r11 + 0x000]
-		xbit $r12 $r12 31
-		bra ne #nv_wr32_wait
-	ret
-
-// (re)set watchdog timer
-//
-// In : $r15 timeout
-//
-watchdog_reset:
-	mov $r8 0x430
-	shl b32 $r8 6
-	bset $r15 31
-	iowr I[$r8 + 0x000] $r15
-	ret
-
-// clear watchdog timer
-watchdog_clear:
-	mov $r8 0x430
-	shl b32 $r8 6
-	iowr I[$r8 + 0x000] $r0
-	ret
-
-// wait_done{z,o} - wait on FUC_DONE bit to become clear/set
-//
-// In : $r10 bit to wait on
-//
-define(`wait_done', `
-$1:
-	trace_set(T_WAIT);
-	mov $r8 0x818
-	shl b32 $r8 6
-	iowr I[$r8 + 0x000] $r10	// CC_SCRATCH[6] = wait bit
-	wait_done_$1:
-		mov $r8 0x400
-		shl b32 $r8 6
-		iord $r8 I[$r8 + 0x000]	// DONE
-		xbit $r8 $r8 $r10
-		bra $2 #wait_done_$1
-	trace_clr(T_WAIT)
-	ret
-')
-wait_done(wait_donez, ne)
-wait_done(wait_doneo, e)
-
-// mmctx_size - determine size of a mmio list transfer
-//
-// In : $r14 mmio list head
-//      $r15 mmio list tail
-// Out: $r15 transfer size (in bytes)
-//
-mmctx_size:
-	clear b32 $r9
-	nv_mmctx_size_loop:
-		ld b32 $r8 D[$r14]
-		shr b32 $r8 26
-		add b32 $r8 1
-		shl b32 $r8 2
-		add b32 $r9 $r8
-		add b32 $r14 4
-		cmpu b32 $r14 $r15
-		bra ne #nv_mmctx_size_loop
-	mov b32 $r15 $r9
-	ret
-
-// mmctx_xfer - execute a list of mmio transfers
-//
-// In : $r10 flags
-//		bit 0: direction (0 = save, 1 = load)
-//		bit 1: set if first transfer
-//		bit 2: set if last transfer
-//	$r11 base
-//	$r12 mmio list head
-//	$r13 mmio list tail
-//	$r14 multi_stride
-//	$r15 multi_mask
-//
-mmctx_xfer:
-	trace_set(T_MMCTX)
-	mov $r8 0x710
-	shl b32 $r8 6
-	clear b32 $r9
-	or $r11 $r11
-	bra e #mmctx_base_disabled
-		iowr I[$r8 + 0x000] $r11	// MMCTX_BASE
-		bset $r9 0			// BASE_EN
-	mmctx_base_disabled:
-	or $r14 $r14
-	bra e #mmctx_multi_disabled
-		iowr I[$r8 + 0x200] $r14 	// MMCTX_MULTI_STRIDE
-		iowr I[$r8 + 0x300] $r15 	// MMCTX_MULTI_MASK
-		bset $r9 1			// MULTI_EN
-	mmctx_multi_disabled:
-	add b32 $r8 0x100
-
-	xbit $r11 $r10 0
-	shl b32 $r11 16			// DIR
-	bset $r11 12			// QLIMIT = 0x10
-	xbit $r14 $r10 1
-	shl b32 $r14 17
-	or $r11 $r14			// START_TRIGGER
-	iowr I[$r8 + 0x000] $r11	// MMCTX_CTRL
-
-	// loop over the mmio list, and send requests to the hw
-	mmctx_exec_loop:
-		// wait for space in mmctx queue
-		mmctx_wait_free:
-			iord $r14 I[$r8 + 0x000] // MMCTX_CTRL
-			and $r14 0x1f
-			bra e #mmctx_wait_free
-
-		// queue up an entry
-		ld b32 $r14 D[$r12]
-		or $r14 $r9
-		iowr I[$r8 + 0x300] $r14
-		add b32 $r12 4
-		cmpu b32 $r12 $r13
-		bra ne #mmctx_exec_loop
-
-	xbit $r11 $r10 2
-	bra ne #mmctx_stop
-		// wait for queue to empty
-		mmctx_fini_wait:
-			iord $r11 I[$r8 + 0x000]	// MMCTX_CTRL
-			and $r11 0x1f
-			cmpu b32 $r11 0x10
-			bra ne #mmctx_fini_wait
-		mov $r10 2				// DONE_MMCTX
-		call #wait_donez
-		bra #mmctx_done
-	mmctx_stop:
-		xbit $r11 $r10 0
-		shl b32 $r11 16			// DIR
-		bset $r11 12			// QLIMIT = 0x10
-		bset $r11 18			// STOP_TRIGGER
-		iowr I[$r8 + 0x000] $r11	// MMCTX_CTRL
-		mmctx_stop_wait:
-			// wait for STOP_TRIGGER to clear
-			iord $r11 I[$r8 + 0x000] // MMCTX_CTRL
-			xbit $r11 $r11 18
-			bra ne #mmctx_stop_wait
-	mmctx_done:
-	trace_clr(T_MMCTX)
-	ret
-
-// Wait for DONE_STRAND
-//
-strand_wait:
-	push $r10
-	mov $r10 2
-	call #wait_donez
-	pop $r10
-	ret
-
-// unknown - call before issuing strand commands
-//
-strand_pre:
-	mov $r8 0x4afc
-	sethi $r8 0x20000
-	mov $r9 0xc
-	iowr I[$r8] $r9
-	call #strand_wait
-	ret
-
-// unknown - call after issuing strand commands
-//
-strand_post:
-	mov $r8 0x4afc
-	sethi $r8 0x20000
-	mov $r9 0xd
-	iowr I[$r8] $r9
-	call #strand_wait
-	ret
-
-// Selects strand set?!
-//
-// In: $r14 id
-//
-strand_set:
-	mov $r10 0x4ffc
-	sethi $r10 0x20000
-	sub b32 $r11 $r10 0x500
-	mov $r12 0xf
-	iowr I[$r10 + 0x000] $r12		// 0x93c = 0xf
-	mov $r12 0xb
-	iowr I[$r11 + 0x000] $r12		// 0x928 = 0xb
-	call #strand_wait
-	iowr I[$r10 + 0x000] $r14		// 0x93c = <id>
-	mov $r12 0xa
-	iowr I[$r11 + 0x000] $r12		// 0x928 = 0xa
-	call #strand_wait
-	ret
-
-// Initialise strand context data
-//
-// In : $r15 context base
-// Out: $r15 context size (in bytes)
-//
-// Strandset(?) 3 hardcoded currently
-//
-strand_ctx_init:
-	trace_set(T_STRINIT)
-	call #strand_pre
-	mov $r14 3
-	call #strand_set
-	mov $r10 0x46fc
-	sethi $r10 0x20000
-	add b32 $r11 $r10 0x400
-	iowr I[$r10 + 0x100] $r0	// STRAND_FIRST_GENE = 0
-	mov $r12 1
-	iowr I[$r11 + 0x000] $r12	// STRAND_CMD = LATCH_FIRST_GENE
-	call #strand_wait
-	sub b32 $r12 $r0 1
-	iowr I[$r10 + 0x000] $r12	// STRAND_GENE_CNT = 0xffffffff
-	mov $r12 2
-	iowr I[$r11 + 0x000] $r12	// STRAND_CMD = LATCH_GENE_CNT
-	call #strand_wait
-	call #strand_post
-
-	// read the size of each strand, poke the context offset of
-	// each into STRAND_{SAVE,LOAD}_SWBASE now, no need to worry
-	// about it later then.
-	mov $r8 0x880
-	shl b32 $r8 6
-	iord $r9 I[$r8 + 0x000]		// STRANDS
-	add b32 $r8 0x2200
-	shr b32 $r14 $r15 8
-	ctx_init_strand_loop:
-		iowr I[$r8 + 0x000] $r14	// STRAND_SAVE_SWBASE
-		iowr I[$r8 + 0x100] $r14	// STRAND_LOAD_SWBASE
-		iord $r10 I[$r8 + 0x200]	// STRAND_SIZE
-		shr b32 $r10 6
-		add b32 $r10 1
-		add b32 $r14 $r10
-		add b32 $r8 4
-		sub b32 $r9 1
-		bra ne #ctx_init_strand_loop
-
-	shl b32 $r14 8
-	sub b32 $r15 $r14 $r15
-	trace_clr(T_STRINIT)
-	ret
-')
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/fuc/os.h b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/os.h
new file mode 100644
index 000000000000..fd1d380de094
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/fuc/os.h
@@ -0,0 +1,7 @@
+#ifndef __NVKM_GRAPH_OS_H__
+#define __NVKM_GRAPH_OS_H__
+
+#define E_BAD_COMMAND  0x00000001
+#define E_CMD_OVERFLOW 0x00000002
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
index 1ac36110ca19..03de5175dd9f 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nv50.c
@@ -186,13 +186,6 @@ nv50_graph_cclass = {
  * PGRAPH engine/subdev functions
  ******************************************************************************/
 
-static int
-nv50_graph_tlb_flush(struct nouveau_engine *engine)
-{
-	nv50_vm_flush_engine(&engine->base, 0x00);
-	return 0;
-}
-
 static const struct nouveau_bitfield nv50_pgraph_status[] = {
 	{ 0x00000001, "BUSY" }, /* set when any bit is set */
 	{ 0x00000002, "DISPATCH" },
@@ -302,8 +295,10 @@ nv84_graph_tlb_flush(struct nouveau_engine *engine)
 				nv_rd32(priv, 0x400388));
 	}
 
-	nv50_vm_flush_engine(&engine->base, 0x00);
 
+	nv_wr32(priv, 0x100c80, 0x00000001);
+	if (!nv_wait(priv, 0x100c80, 0x00000001, 0x00000000))
+		nv_error(priv, "vm flush timeout\n");
 	nv_mask(priv, 0x400500, 0x00000001, 0x00000001);
 	spin_unlock_irqrestore(&priv->lock, flags);
 	return timeout ? -EBUSY : 0;
@@ -857,10 +852,9 @@ nv50_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 
 	};
 
-	if (nv_device(priv)->chipset == 0x50 ||
-	    nv_device(priv)->chipset == 0xac)
-		nv_engine(priv)->tlb_flush = nv50_graph_tlb_flush;
-	else
+	/* unfortunate hw bug workaround... */
+	if (nv_device(priv)->chipset != 0x50 &&
+	    nv_device(priv)->chipset != 0xac)
 		nv_engine(priv)->tlb_flush = nv84_graph_tlb_flush;
 
 	spin_lock_init(&priv->lock);
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
index f9b9d82c287f..3f4f35cc3848 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.c
@@ -23,14 +23,12 @@
  */
 
 #include "nvc0.h"
-#include "fuc/hubnvc0.fuc.h"
-#include "fuc/gpcnvc0.fuc.h"
 
 /*******************************************************************************
  * Graphics object classes
  ******************************************************************************/
 
-static struct nouveau_oclass
+struct nouveau_oclass
 nvc0_graph_sclass[] = {
 	{ 0x902d, &nouveau_object_ofuncs },
 	{ 0x9039, &nouveau_object_ofuncs },
@@ -39,40 +37,6 @@ nvc0_graph_sclass[] = {
 	{}
 };
 
-static struct nouveau_oclass
-nvc1_graph_sclass[] = {
-	{ 0x902d, &nouveau_object_ofuncs },
-	{ 0x9039, &nouveau_object_ofuncs },
-	{ 0x9097, &nouveau_object_ofuncs },
-	{ 0x90c0, &nouveau_object_ofuncs },
-	{ 0x9197, &nouveau_object_ofuncs },
-	{}
-};
-
-static struct nouveau_oclass
-nvc8_graph_sclass[] = {
-	{ 0x902d, &nouveau_object_ofuncs },
-	{ 0x9039, &nouveau_object_ofuncs },
-	{ 0x9097, &nouveau_object_ofuncs },
-	{ 0x90c0, &nouveau_object_ofuncs },
-	{ 0x9197, &nouveau_object_ofuncs },
-	{ 0x9297, &nouveau_object_ofuncs },
-	{}
-};
-
-u64
-nvc0_graph_units(struct nouveau_graph *graph)
-{
-	struct nvc0_graph_priv *priv = (void *)graph;
-	u64 cfg;
-
-	cfg  = (u32)priv->gpc_nr;
-	cfg |= (u32)priv->tpc_total << 8;
-	cfg |= (u64)priv->rop_nr << 32;
-
-	return cfg;
-}
-
 /*******************************************************************************
  * PGRAPH context
  ******************************************************************************/
@@ -181,60 +145,308 @@ nvc0_graph_context_dtor(struct nouveau_object *object)
 	nouveau_graph_context_destroy(&chan->base);
 }
 
-static struct nouveau_oclass
-nvc0_graph_cclass = {
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nvc0_graph_context_ctor,
-		.dtor = nvc0_graph_context_dtor,
-		.init = _nouveau_graph_context_init,
-		.fini = _nouveau_graph_context_fini,
-		.rd32 = _nouveau_graph_context_rd32,
-		.wr32 = _nouveau_graph_context_wr32,
-	},
-};
-
 /*******************************************************************************
  * PGRAPH engine/subdev functions
  ******************************************************************************/
 
-static void
-nvc0_graph_ctxctl_debug_unit(struct nvc0_graph_priv *priv, u32 base)
+struct nvc0_graph_init
+nvc0_graph_init_regs[] = {
+	{ 0x400080,   1, 0x04, 0x003083c2 },
+	{ 0x400088,   1, 0x04, 0x00006fe7 },
+	{ 0x40008c,   1, 0x04, 0x00000000 },
+	{ 0x400090,   1, 0x04, 0x00000030 },
+	{ 0x40013c,   1, 0x04, 0x013901f7 },
+	{ 0x400140,   1, 0x04, 0x00000100 },
+	{ 0x400144,   1, 0x04, 0x00000000 },
+	{ 0x400148,   1, 0x04, 0x00000110 },
+	{ 0x400138,   1, 0x04, 0x00000000 },
+	{ 0x400130,   2, 0x04, 0x00000000 },
+	{ 0x400124,   1, 0x04, 0x00000002 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk40xx[] = {
+	{ 0x40415c,   1, 0x04, 0x00000000 },
+	{ 0x404170,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk44xx[] = {
+	{ 0x404488,   2, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk78xx[] = {
+	{ 0x407808,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk60xx[] = {
+	{ 0x406024,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk58xx[] = {
+	{ 0x405844,   1, 0x04, 0x00ffffff },
+	{ 0x405850,   1, 0x04, 0x00000000 },
+	{ 0x405908,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk80xx[] = {
+	{ 0x40803c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_gpc[] = {
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x80000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   1, 0x04, 0x80000000 },
+	{ 0x4188cc,   1, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000050 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc0_graph_init_tpc[] = {
+	{ 0x419d08,   2, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x41980c,   3, 0x04, 0x00000000 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x41984c,   1, 0x04, 0x00005bc5 },
+	{ 0x419850,   4, 0x04, 0x00000000 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x80000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00008bf4 },
+	{ 0x419cbc,   1, 0x04, 0x28137606 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419bd4,   1, 0x04, 0x00800000 },
+	{ 0x419bdc,   1, 0x04, 0x00000000 },
+	{ 0x419d2c,   1, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00001100 },
+	{ 0x419eac,   1, 0x04, 0x11100702 },
+	{ 0x419eb0,   1, 0x04, 0x00000003 },
+	{ 0x419eb4,   4, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x06060618 },
+	{ 0x419ed0,   1, 0x04, 0x0eff0e38 },
+	{ 0x419ed4,   1, 0x04, 0x011104f1 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f2c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_init_unk88xx[] = {
+	{ 0x40880c,   1, 0x04, 0x00000000 },
+	{ 0x408910,   9, 0x04, 0x00000000 },
+	{ 0x408950,   1, 0x04, 0x00000000 },
+	{ 0x408954,   1, 0x04, 0x0000ffff },
+	{ 0x408984,   1, 0x04, 0x00000000 },
+	{ 0x408988,   1, 0x04, 0x08040201 },
+	{ 0x40898c,   1, 0x04, 0x80402010 },
+	{}
+};
+
+struct nvc0_graph_init
+nvc0_graph_tpc_0[] = {
+	{ 0x50405c,   1, 0x04, 0x00000001 },
+	{}
+};
+
+void
+nvc0_graph_mmio(struct nvc0_graph_priv *priv, struct nvc0_graph_init *init)
 {
-	nv_error(priv, "%06x - done 0x%08x\n", base,
-		 nv_rd32(priv, base + 0x400));
-	nv_error(priv, "%06x - stat 0x%08x 0x%08x 0x%08x 0x%08x\n", base,
-		 nv_rd32(priv, base + 0x800), nv_rd32(priv, base + 0x804),
-		 nv_rd32(priv, base + 0x808), nv_rd32(priv, base + 0x80c));
-	nv_error(priv, "%06x - stat 0x%08x 0x%08x 0x%08x 0x%08x\n", base,
-		 nv_rd32(priv, base + 0x810), nv_rd32(priv, base + 0x814),
-		 nv_rd32(priv, base + 0x818), nv_rd32(priv, base + 0x81c));
+	for (; init && init->count; init++) {
+		u32 addr = init->addr, i;
+		for (i = 0; i < init->count; i++) {
+			nv_wr32(priv, addr, init->data);
+			addr += init->pitch;
+		}
+	}
 }
 
 void
-nvc0_graph_ctxctl_debug(struct nvc0_graph_priv *priv)
+nvc0_graph_icmd(struct nvc0_graph_priv *priv, struct nvc0_graph_init *init)
 {
-	u32 gpcnr = nv_rd32(priv, 0x409604) & 0xffff;
-	u32 gpc;
+	u32 addr, data;
+	int i, j;
+
+	nv_wr32(priv, 0x400208, 0x80000000);
+	for (i = 0; init->count; init++, i++) {
+		if (!i || data != init->data) {
+			nv_wr32(priv, 0x400204, init->data);
+			data = init->data;
+		}
 
-	nvc0_graph_ctxctl_debug_unit(priv, 0x409000);
-	for (gpc = 0; gpc < gpcnr; gpc++)
-		nvc0_graph_ctxctl_debug_unit(priv, 0x502000 + (gpc * 0x8000));
+		addr = init->addr;
+		for (j = 0; j < init->count; j++) {
+			nv_wr32(priv, 0x400200, addr);
+			addr += init->pitch;
+			while (nv_rd32(priv, 0x400700) & 0x00000002) {}
+		}
+	}
+	nv_wr32(priv, 0x400208, 0x00000000);
+}
+
+void
+nvc0_graph_mthd(struct nvc0_graph_priv *priv, struct nvc0_graph_mthd *mthds)
+{
+	struct nvc0_graph_mthd *mthd;
+	struct nvc0_graph_init *init;
+	int i = 0, j;
+	u32 data;
+
+	while ((mthd = &mthds[i++]) && (init = mthd->init)) {
+		u32  addr = 0x80000000 | mthd->oclass;
+		for (data = 0; init->count; init++) {
+			if (data != init->data) {
+				nv_wr32(priv, 0x40448c, init->data);
+				data = init->data;
+			}
+
+			addr = (addr & 0x8000ffff) | (init->addr << 14);
+			for (j = 0; j < init->count; j++) {
+				nv_wr32(priv, 0x404488, addr);
+				addr += init->pitch << 14;
+			}
+		}
+	}
+}
+
+u64
+nvc0_graph_units(struct nouveau_graph *graph)
+{
+	struct nvc0_graph_priv *priv = (void *)graph;
+	u64 cfg;
+
+	cfg  = (u32)priv->gpc_nr;
+	cfg |= (u32)priv->tpc_total << 8;
+	cfg |= (u64)priv->rop_nr << 32;
+
+	return cfg;
 }
 
+static const struct nouveau_enum nve0_sked_error[] = {
+	{ 7, "CONSTANT_BUFFER_SIZE" },
+	{ 9, "LOCAL_MEMORY_SIZE_POS" },
+	{ 10, "LOCAL_MEMORY_SIZE_NEG" },
+	{ 11, "WARP_CSTACK_SIZE" },
+	{ 12, "TOTAL_TEMP_SIZE" },
+	{ 13, "REGISTER_COUNT" },
+	{ 18, "TOTAL_THREADS" },
+	{ 20, "PROGRAM_OFFSET" },
+	{ 21, "SHARED_MEMORY_SIZE" },
+	{ 25, "SHARED_CONFIG_TOO_SMALL" },
+	{ 26, "TOTAL_REGISTER_COUNT" },
+	{}
+};
+
+static const struct nouveau_enum nvc0_gpc_rop_error[] = {
+	{ 1, "RT_PITCH_OVERRUN" },
+	{ 4, "RT_WIDTH_OVERRUN" },
+	{ 5, "RT_HEIGHT_OVERRUN" },
+	{ 7, "ZETA_STORAGE_TYPE_MISMATCH" },
+	{ 8, "RT_STORAGE_TYPE_MISMATCH" },
+	{ 10, "RT_LINEAR_MISMATCH" },
+	{}
+};
+
 static void
-nvc0_graph_ctxctl_isr(struct nvc0_graph_priv *priv)
+nvc0_graph_trap_gpc_rop(struct nvc0_graph_priv *priv, int gpc)
 {
-	u32 ustat = nv_rd32(priv, 0x409c18);
+	u32 trap[4];
+	int i;
 
-	if (ustat & 0x00000001)
-		nv_error(priv, "CTXCTRL ucode error\n");
-	if (ustat & 0x00080000)
-		nv_error(priv, "CTXCTRL watchdog timeout\n");
-	if (ustat & ~0x00080001)
-		nv_error(priv, "CTXCTRL 0x%08x\n", ustat);
+	trap[0] = nv_rd32(priv, GPC_UNIT(gpc, 0x0420));
+	trap[1] = nv_rd32(priv, GPC_UNIT(gpc, 0x0434));
+	trap[2] = nv_rd32(priv, GPC_UNIT(gpc, 0x0438));
+	trap[3] = nv_rd32(priv, GPC_UNIT(gpc, 0x043c));
+
+	nv_error(priv, "GPC%d/PROP trap:", gpc);
+	for (i = 0; i <= 29; ++i) {
+		if (!(trap[0] & (1 << i)))
+			continue;
+		pr_cont(" ");
+		nouveau_enum_print(nvc0_gpc_rop_error, i);
+	}
+	pr_cont("\n");
 
-	nvc0_graph_ctxctl_debug(priv);
-	nv_wr32(priv, 0x409c20, ustat);
+	nv_error(priv, "x = %u, y = %u, format = %x, storage type = %x\n",
+		 trap[1] & 0xffff, trap[1] >> 16, (trap[2] >> 8) & 0x3f,
+		 trap[3] & 0xff);
+	nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
+}
+
+static const struct nouveau_enum nvc0_mp_warp_error[] = {
+	{ 0x00, "NO_ERROR" },
+	{ 0x01, "STACK_MISMATCH" },
+	{ 0x05, "MISALIGNED_PC" },
+	{ 0x08, "MISALIGNED_GPR" },
+	{ 0x09, "INVALID_OPCODE" },
+	{ 0x0d, "GPR_OUT_OF_BOUNDS" },
+	{ 0x0e, "MEM_OUT_OF_BOUNDS" },
+	{ 0x0f, "UNALIGNED_MEM_ACCESS" },
+	{ 0x11, "INVALID_PARAM" },
+	{}
+};
+
+static const struct nouveau_bitfield nvc0_mp_global_error[] = {
+	{ 0x00000004, "MULTIPLE_WARP_ERRORS" },
+	{ 0x00000008, "OUT_OF_STACK_SPACE" },
+	{}
+};
+
+static void
+nvc0_graph_trap_mp(struct nvc0_graph_priv *priv, int gpc, int tpc)
+{
+	u32 werr = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x648));
+	u32 gerr = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x650));
+
+	nv_error(priv, "GPC%i/TPC%i/MP trap:", gpc, tpc);
+	nouveau_bitfield_print(nvc0_mp_global_error, gerr);
+	if (werr) {
+		pr_cont(" ");
+		nouveau_enum_print(nvc0_mp_warp_error, werr & 0xffff);
+	}
+	pr_cont("\n");
+
+	nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x648), 0x00000000);
+	nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x650), gerr);
 }
 
 static void
@@ -246,18 +458,11 @@ nvc0_graph_trap_tpc(struct nvc0_graph_priv *priv, int gpc, int tpc)
 		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x0224));
 		nv_error(priv, "GPC%d/TPC%d/TEX: 0x%08x\n", gpc, tpc, trap);
 		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0224), 0xc0000000);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0508), 0x00000001);
 		stat &= ~0x00000001;
 	}
 
 	if (stat & 0x00000002) {
-		u32 trap0 = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x0644));
-		u32 trap1 = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x064c));
-		nv_error(priv, "GPC%d/TPC%d/MP: 0x%08x 0x%08x\n",
-			       gpc, tpc, trap0, trap1);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0644), 0x001ffffe);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x064c), 0x0000000f);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0508), 0x00000002);
+		nvc0_graph_trap_mp(priv, gpc, tpc);
 		stat &= ~0x00000002;
 	}
 
@@ -265,7 +470,6 @@ nvc0_graph_trap_tpc(struct nvc0_graph_priv *priv, int gpc, int tpc)
 		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x0084));
 		nv_error(priv, "GPC%d/TPC%d/POLY: 0x%08x\n", gpc, tpc, trap);
 		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0084), 0xc0000000);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0508), 0x00000004);
 		stat &= ~0x00000004;
 	}
 
@@ -273,13 +477,11 @@ nvc0_graph_trap_tpc(struct nvc0_graph_priv *priv, int gpc, int tpc)
 		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tpc, 0x048c));
 		nv_error(priv, "GPC%d/TPC%d/L1C: 0x%08x\n", gpc, tpc, trap);
 		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x048c), 0xc0000000);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0508), 0x00000008);
 		stat &= ~0x00000008;
 	}
 
 	if (stat) {
 		nv_error(priv, "GPC%d/TPC%d/0x%08x: unknown\n", gpc, tpc, stat);
-		nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x0508), stat);
 	}
 }
 
@@ -290,10 +492,7 @@ nvc0_graph_trap_gpc(struct nvc0_graph_priv *priv, int gpc)
 	int tpc;
 
 	if (stat & 0x00000001) {
-		u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x0420));
-		nv_error(priv, "GPC%d/PROP: 0x%08x\n", gpc, trap);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0x00000001);
+		nvc0_graph_trap_gpc_rop(priv, gpc);
 		stat &= ~0x00000001;
 	}
 
@@ -301,7 +500,6 @@ nvc0_graph_trap_gpc(struct nvc0_graph_priv *priv, int gpc)
 		u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x0900));
 		nv_error(priv, "GPC%d/ZCULL: 0x%08x\n", gpc, trap);
 		nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0x00000002);
 		stat &= ~0x00000002;
 	}
 
@@ -309,7 +507,6 @@ nvc0_graph_trap_gpc(struct nvc0_graph_priv *priv, int gpc)
 		u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x1028));
 		nv_error(priv, "GPC%d/CCACHE: 0x%08x\n", gpc, trap);
 		nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0x00000004);
 		stat &= ~0x00000004;
 	}
 
@@ -317,7 +514,6 @@ nvc0_graph_trap_gpc(struct nvc0_graph_priv *priv, int gpc)
 		u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x0824));
 		nv_error(priv, "GPC%d/ESETUP: 0x%08x\n", gpc, trap);
 		nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0x00000008);
 		stat &= ~0x00000009;
 	}
 
@@ -332,7 +528,6 @@ nvc0_graph_trap_gpc(struct nvc0_graph_priv *priv, int gpc)
 
 	if (stat) {
 		nv_error(priv, "GPC%d/0x%08x: unknown\n", gpc, stat);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), stat);
 	}
 }
 
@@ -340,7 +535,7 @@ static void
 nvc0_graph_trap_intr(struct nvc0_graph_priv *priv)
 {
 	u32 trap = nv_rd32(priv, 0x400108);
-	int rop, gpc;
+	int rop, gpc, i;
 
 	if (trap & 0x00000001) {
 		u32 stat = nv_rd32(priv, 0x404000);
@@ -390,6 +585,24 @@ nvc0_graph_trap_intr(struct nvc0_graph_priv *priv)
 		trap &= ~0x00000080;
 	}
 
+	if (trap & 0x00000100) {
+		u32 stat = nv_rd32(priv, 0x407020);
+
+		nv_error(priv, "SKED:");
+		for (i = 0; i <= 29; ++i) {
+			if (!(stat & (1 << i)))
+				continue;
+			pr_cont(" ");
+			nouveau_enum_print(nve0_sked_error, i);
+		}
+		pr_cont("\n");
+
+		if (stat & 0x3fffffff)
+			nv_wr32(priv, 0x407020, 0x40000000);
+		nv_wr32(priv, 0x400108, 0x00000100);
+		trap &= ~0x00000100;
+	}
+
 	if (trap & 0x01000000) {
 		u32 stat = nv_rd32(priv, 0x400118);
 		for (gpc = 0; stat && gpc < priv->gpc_nr; gpc++) {
@@ -424,6 +637,46 @@ nvc0_graph_trap_intr(struct nvc0_graph_priv *priv)
 }
 
 static void
+nvc0_graph_ctxctl_debug_unit(struct nvc0_graph_priv *priv, u32 base)
+{
+	nv_error(priv, "%06x - done 0x%08x\n", base,
+		 nv_rd32(priv, base + 0x400));
+	nv_error(priv, "%06x - stat 0x%08x 0x%08x 0x%08x 0x%08x\n", base,
+		 nv_rd32(priv, base + 0x800), nv_rd32(priv, base + 0x804),
+		 nv_rd32(priv, base + 0x808), nv_rd32(priv, base + 0x80c));
+	nv_error(priv, "%06x - stat 0x%08x 0x%08x 0x%08x 0x%08x\n", base,
+		 nv_rd32(priv, base + 0x810), nv_rd32(priv, base + 0x814),
+		 nv_rd32(priv, base + 0x818), nv_rd32(priv, base + 0x81c));
+}
+
+void
+nvc0_graph_ctxctl_debug(struct nvc0_graph_priv *priv)
+{
+	u32 gpcnr = nv_rd32(priv, 0x409604) & 0xffff;
+	u32 gpc;
+
+	nvc0_graph_ctxctl_debug_unit(priv, 0x409000);
+	for (gpc = 0; gpc < gpcnr; gpc++)
+		nvc0_graph_ctxctl_debug_unit(priv, 0x502000 + (gpc * 0x8000));
+}
+
+static void
+nvc0_graph_ctxctl_isr(struct nvc0_graph_priv *priv)
+{
+	u32 ustat = nv_rd32(priv, 0x409c18);
+
+	if (ustat & 0x00000001)
+		nv_error(priv, "CTXCTL ucode error\n");
+	if (ustat & 0x00080000)
+		nv_error(priv, "CTXCTL watchdog timeout\n");
+	if (ustat & ~0x00080001)
+		nv_error(priv, "CTXCTL 0x%08x\n", ustat);
+
+	nvc0_graph_ctxctl_debug(priv);
+	nv_wr32(priv, 0x409c20, ustat);
+}
+
+static void
 nvc0_graph_intr(struct nouveau_subdev *subdev)
 {
 	struct nouveau_fifo *pfifo = nouveau_fifo(subdev);
@@ -499,290 +752,6 @@ nvc0_graph_intr(struct nouveau_subdev *subdev)
 	nouveau_engctx_put(engctx);
 }
 
-int
-nvc0_graph_ctor_fw(struct nvc0_graph_priv *priv, const char *fwname,
-		   struct nvc0_graph_fuc *fuc)
-{
-	struct nouveau_device *device = nv_device(priv);
-	const struct firmware *fw;
-	char f[32];
-	int ret;
-
-	snprintf(f, sizeof(f), "nouveau/nv%02x_%s", device->chipset, fwname);
-	ret = request_firmware(&fw, f, &device->pdev->dev);
-	if (ret) {
-		snprintf(f, sizeof(f), "nouveau/%s", fwname);
-		ret = request_firmware(&fw, f, &device->pdev->dev);
-		if (ret) {
-			nv_error(priv, "failed to load %s\n", fwname);
-			return ret;
-		}
-	}
-
-	fuc->size = fw->size;
-	fuc->data = kmemdup(fw->data, fuc->size, GFP_KERNEL);
-	release_firmware(fw);
-	return (fuc->data != NULL) ? 0 : -ENOMEM;
-}
-
-static int
-nvc0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-		struct nouveau_oclass *oclass, void *data, u32 size,
-		struct nouveau_object **pobject)
-{
-	struct nouveau_device *device = nv_device(parent);
-	struct nvc0_graph_priv *priv;
-	bool enable = device->chipset != 0xd7;
-	int ret, i;
-
-	ret = nouveau_graph_create(parent, engine, oclass, enable, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
-		return ret;
-
-	nv_subdev(priv)->unit = 0x18001000;
-	nv_subdev(priv)->intr = nvc0_graph_intr;
-	nv_engine(priv)->cclass = &nvc0_graph_cclass;
-
-	priv->base.units = nvc0_graph_units;
-
-	if (nouveau_boolopt(device->cfgopt, "NvGrUseFW", false)) {
-		nv_info(priv, "using external firmware\n");
-		if (nvc0_graph_ctor_fw(priv, "fuc409c", &priv->fuc409c) ||
-		    nvc0_graph_ctor_fw(priv, "fuc409d", &priv->fuc409d) ||
-		    nvc0_graph_ctor_fw(priv, "fuc41ac", &priv->fuc41ac) ||
-		    nvc0_graph_ctor_fw(priv, "fuc41ad", &priv->fuc41ad))
-			return -EINVAL;
-		priv->firmware = true;
-	}
-
-	switch (nvc0_graph_class(priv)) {
-	case 0x9097:
-		nv_engine(priv)->sclass = nvc0_graph_sclass;
-		break;
-	case 0x9197:
-		nv_engine(priv)->sclass = nvc1_graph_sclass;
-		break;
-	case 0x9297:
-		nv_engine(priv)->sclass = nvc8_graph_sclass;
-		break;
-	}
-
-	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
-				&priv->unk4188b4);
-	if (ret)
-		return ret;
-
-	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
-				&priv->unk4188b8);
-	if (ret)
-		return ret;
-
-	for (i = 0; i < 0x1000; i += 4) {
-		nv_wo32(priv->unk4188b4, i, 0x00000010);
-		nv_wo32(priv->unk4188b8, i, 0x00000010);
-	}
-
-	priv->rop_nr = (nv_rd32(priv, 0x409604) & 0x001f0000) >> 16;
-	priv->gpc_nr =  nv_rd32(priv, 0x409604) & 0x0000001f;
-	for (i = 0; i < priv->gpc_nr; i++) {
-		priv->tpc_nr[i]  = nv_rd32(priv, GPC_UNIT(i, 0x2608));
-		priv->tpc_total += priv->tpc_nr[i];
-	}
-
-	/*XXX: these need figuring out... though it might not even matter */
-	switch (nv_device(priv)->chipset) {
-	case 0xc0:
-		if (priv->tpc_total == 11) { /* 465, 3/4/4/0, 4 */
-			priv->magic_not_rop_nr = 0x07;
-		} else
-		if (priv->tpc_total == 14) { /* 470, 3/3/4/4, 5 */
-			priv->magic_not_rop_nr = 0x05;
-		} else
-		if (priv->tpc_total == 15) { /* 480, 3/4/4/4, 6 */
-			priv->magic_not_rop_nr = 0x06;
-		}
-		break;
-	case 0xc3: /* 450, 4/0/0/0, 2 */
-		priv->magic_not_rop_nr = 0x03;
-		break;
-	case 0xc4: /* 460, 3/4/0/0, 4 */
-		priv->magic_not_rop_nr = 0x01;
-		break;
-	case 0xc1: /* 2/0/0/0, 1 */
-		priv->magic_not_rop_nr = 0x01;
-		break;
-	case 0xc8: /* 4/4/3/4, 5 */
-		priv->magic_not_rop_nr = 0x06;
-		break;
-	case 0xce: /* 4/4/0/0, 4 */
-		priv->magic_not_rop_nr = 0x03;
-		break;
-	case 0xcf: /* 4/0/0/0, 3 */
-		priv->magic_not_rop_nr = 0x03;
-		break;
-	case 0xd9: /* 1/0/0/0, 1 */
-		priv->magic_not_rop_nr = 0x01;
-		break;
-	}
-
-	return 0;
-}
-
-static void
-nvc0_graph_dtor_fw(struct nvc0_graph_fuc *fuc)
-{
-	kfree(fuc->data);
-	fuc->data = NULL;
-}
-
-void
-nvc0_graph_dtor(struct nouveau_object *object)
-{
-	struct nvc0_graph_priv *priv = (void *)object;
-
-	kfree(priv->data);
-
-	nvc0_graph_dtor_fw(&priv->fuc409c);
-	nvc0_graph_dtor_fw(&priv->fuc409d);
-	nvc0_graph_dtor_fw(&priv->fuc41ac);
-	nvc0_graph_dtor_fw(&priv->fuc41ad);
-
-	nouveau_gpuobj_ref(NULL, &priv->unk4188b8);
-	nouveau_gpuobj_ref(NULL, &priv->unk4188b4);
-
-	nouveau_graph_destroy(&priv->base);
-}
-
-static void
-nvc0_graph_init_obj418880(struct nvc0_graph_priv *priv)
-{
-	int i;
-
-	nv_wr32(priv, GPC_BCAST(0x0880), 0x00000000);
-	nv_wr32(priv, GPC_BCAST(0x08a4), 0x00000000);
-	for (i = 0; i < 4; i++)
-		nv_wr32(priv, GPC_BCAST(0x0888) + (i * 4), 0x00000000);
-	nv_wr32(priv, GPC_BCAST(0x08b4), priv->unk4188b4->addr >> 8);
-	nv_wr32(priv, GPC_BCAST(0x08b8), priv->unk4188b8->addr >> 8);
-}
-
-static void
-nvc0_graph_init_regs(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x400080, 0x003083c2);
-	nv_wr32(priv, 0x400088, 0x00006fe7);
-	nv_wr32(priv, 0x40008c, 0x00000000);
-	nv_wr32(priv, 0x400090, 0x00000030);
-	nv_wr32(priv, 0x40013c, 0x013901f7);
-	nv_wr32(priv, 0x400140, 0x00000100);
-	nv_wr32(priv, 0x400144, 0x00000000);
-	nv_wr32(priv, 0x400148, 0x00000110);
-	nv_wr32(priv, 0x400138, 0x00000000);
-	nv_wr32(priv, 0x400130, 0x00000000);
-	nv_wr32(priv, 0x400134, 0x00000000);
-	nv_wr32(priv, 0x400124, 0x00000002);
-}
-
-static void
-nvc0_graph_init_gpc_0(struct nvc0_graph_priv *priv)
-{
-	const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, priv->tpc_total);
-	u32 data[TPC_MAX / 8];
-	u8  tpcnr[GPC_MAX];
-	int i, gpc, tpc;
-
-	nv_wr32(priv, TPC_UNIT(0, 0, 0x5c), 1); /* affects TFB offset queries */
-
-	/*
-	 *      TP      ROP UNKVAL(magic_not_rop_nr)
-	 * 450: 4/0/0/0 2        3
-	 * 460: 3/4/0/0 4        1
-	 * 465: 3/4/4/0 4        7
-	 * 470: 3/3/4/4 5        5
-	 * 480: 3/4/4/4 6        6
-	 */
-
-	memset(data, 0x00, sizeof(data));
-	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
-	for (i = 0, gpc = -1; i < priv->tpc_total; i++) {
-		do {
-			gpc = (gpc + 1) % priv->gpc_nr;
-		} while (!tpcnr[gpc]);
-		tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
-
-		data[i / 8] |= tpc << ((i % 8) * 4);
-	}
-
-	nv_wr32(priv, GPC_BCAST(0x0980), data[0]);
-	nv_wr32(priv, GPC_BCAST(0x0984), data[1]);
-	nv_wr32(priv, GPC_BCAST(0x0988), data[2]);
-	nv_wr32(priv, GPC_BCAST(0x098c), data[3]);
-
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0914), priv->magic_not_rop_nr << 8 |
-						  priv->tpc_nr[gpc]);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0910), 0x00040000 | priv->tpc_total);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0918), magicgpc918);
-	}
-
-	nv_wr32(priv, GPC_BCAST(0x1bd4), magicgpc918);
-	nv_wr32(priv, GPC_BCAST(0x08ac), nv_rd32(priv, 0x100800));
-}
-
-static void
-nvc0_graph_init_units(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x409c24, 0x000f0000);
-	nv_wr32(priv, 0x404000, 0xc0000000); /* DISPATCH */
-	nv_wr32(priv, 0x404600, 0xc0000000); /* M2MF */
-	nv_wr32(priv, 0x408030, 0xc0000000);
-	nv_wr32(priv, 0x40601c, 0xc0000000);
-	nv_wr32(priv, 0x404490, 0xc0000000); /* MACRO */
-	nv_wr32(priv, 0x406018, 0xc0000000);
-	nv_wr32(priv, 0x405840, 0xc0000000);
-	nv_wr32(priv, 0x405844, 0x00ffffff);
-	nv_mask(priv, 0x419cc0, 0x00000008, 0x00000008);
-	nv_mask(priv, 0x419eb4, 0x00001000, 0x00001000);
-}
-
-static void
-nvc0_graph_init_gpc_1(struct nvc0_graph_priv *priv)
-{
-	int gpc, tpc;
-
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
-		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x508), 0xffffffff);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x50c), 0xffffffff);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x224), 0xc0000000);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x48c), 0xc0000000);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x084), 0xc0000000);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x644), 0x001ffffe);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x64c), 0x0000000f);
-		}
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0xffffffff);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c94), 0xffffffff);
-	}
-}
-
-static void
-nvc0_graph_init_rop(struct nvc0_graph_priv *priv)
-{
-	int rop;
-
-	for (rop = 0; rop < priv->rop_nr; rop++) {
-		nv_wr32(priv, ROP_UNIT(rop, 0x144), 0xc0000000);
-		nv_wr32(priv, ROP_UNIT(rop, 0x070), 0xc0000000);
-		nv_wr32(priv, ROP_UNIT(rop, 0x204), 0xffffffff);
-		nv_wr32(priv, ROP_UNIT(rop, 0x208), 0xffffffff);
-	}
-}
-
 void
 nvc0_graph_init_fw(struct nvc0_graph_priv *priv, u32 fuc_base,
 		   struct nvc0_graph_fuc *code, struct nvc0_graph_fuc *data)
@@ -801,9 +770,46 @@ nvc0_graph_init_fw(struct nvc0_graph_priv *priv, u32 fuc_base,
 	}
 }
 
-static int
+static void
+nvc0_graph_init_csdata(struct nvc0_graph_priv *priv,
+		       struct nvc0_graph_init *init,
+		       u32 falcon, u32 starstar, u32 base)
+{
+	u32 addr = init->addr;
+	u32 next = addr;
+	u32 star, temp;
+
+	nv_wr32(priv, falcon + 0x01c0, 0x02000000 + starstar);
+	star = nv_rd32(priv, falcon + 0x01c4);
+	temp = nv_rd32(priv, falcon + 0x01c4);
+	if (temp > star)
+		star = temp;
+	nv_wr32(priv, falcon + 0x01c0, 0x01000000 + star);
+
+	do {
+		if (init->addr != next) {
+			while (addr < next) {
+				u32 nr = min((int)(next - addr) / 4, 32);
+				nv_wr32(priv, falcon + 0x01c4,
+					((nr - 1) << 26) | (addr - base));
+				addr += nr * 4;
+				star += 4;
+			}
+			addr = next = init->addr;
+		}
+		next += init->count * 4;
+	} while ((init++)->count);
+
+	nv_wr32(priv, falcon + 0x01c0, 0x01000004 + starstar);
+	nv_wr32(priv, falcon + 0x01c4, star);
+}
+
+int
 nvc0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
 {
+	struct nvc0_graph_oclass *oclass = (void *)nv_object(priv)->oclass;
+	struct nvc0_grctx_oclass *cclass = (void *)nv_engine(priv)->cclass;
+	struct nvc0_graph_init *init;
 	u32 r000260;
 	int i;
 
@@ -854,6 +860,38 @@ nvc0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
 			return -EBUSY;
 		}
 
+		if (nv_device(priv)->chipset >= 0xe0) {
+			nv_wr32(priv, 0x409800, 0x00000000);
+			nv_wr32(priv, 0x409500, 0x00000001);
+			nv_wr32(priv, 0x409504, 0x00000030);
+			if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
+				nv_error(priv, "fuc09 req 0x30 timeout\n");
+				return -EBUSY;
+			}
+
+			nv_wr32(priv, 0x409810, 0xb00095c8);
+			nv_wr32(priv, 0x409800, 0x00000000);
+			nv_wr32(priv, 0x409500, 0x00000001);
+			nv_wr32(priv, 0x409504, 0x00000031);
+			if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
+				nv_error(priv, "fuc09 req 0x31 timeout\n");
+				return -EBUSY;
+			}
+
+			nv_wr32(priv, 0x409810, 0x00080420);
+			nv_wr32(priv, 0x409800, 0x00000000);
+			nv_wr32(priv, 0x409500, 0x00000001);
+			nv_wr32(priv, 0x409504, 0x00000032);
+			if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
+				nv_error(priv, "fuc09 req 0x32 timeout\n");
+				return -EBUSY;
+			}
+
+			nv_wr32(priv, 0x409614, 0x00000070);
+			nv_wr32(priv, 0x409614, 0x00000770);
+			nv_wr32(priv, 0x40802c, 0x00000001);
+		}
+
 		if (priv->data == NULL) {
 			int ret = nvc0_grctx_generate(priv);
 			if (ret) {
@@ -868,31 +906,41 @@ nvc0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
 	/* load HUB microcode */
 	r000260 = nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
 	nv_wr32(priv, 0x4091c0, 0x01000000);
-	for (i = 0; i < sizeof(nvc0_grhub_data) / 4; i++)
-		nv_wr32(priv, 0x4091c4, nvc0_grhub_data[i]);
+	for (i = 0; i < oclass->fecs.ucode->data.size / 4; i++)
+		nv_wr32(priv, 0x4091c4, oclass->fecs.ucode->data.data[i]);
 
 	nv_wr32(priv, 0x409180, 0x01000000);
-	for (i = 0; i < sizeof(nvc0_grhub_code) / 4; i++) {
+	for (i = 0; i < oclass->fecs.ucode->code.size / 4; i++) {
 		if ((i & 0x3f) == 0)
 			nv_wr32(priv, 0x409188, i >> 6);
-		nv_wr32(priv, 0x409184, nvc0_grhub_code[i]);
+		nv_wr32(priv, 0x409184, oclass->fecs.ucode->code.data[i]);
+	}
+
+	for (i = 0; (init = cclass->hub[i]); i++) {
+		nvc0_graph_init_csdata(priv, init, 0x409000, 0x000, 0x000000);
 	}
 
 	/* load GPC microcode */
 	nv_wr32(priv, 0x41a1c0, 0x01000000);
-	for (i = 0; i < sizeof(nvc0_grgpc_data) / 4; i++)
-		nv_wr32(priv, 0x41a1c4, nvc0_grgpc_data[i]);
+	for (i = 0; i < oclass->gpccs.ucode->data.size / 4; i++)
+		nv_wr32(priv, 0x41a1c4, oclass->gpccs.ucode->data.data[i]);
 
 	nv_wr32(priv, 0x41a180, 0x01000000);
-	for (i = 0; i < sizeof(nvc0_grgpc_code) / 4; i++) {
+	for (i = 0; i < oclass->gpccs.ucode->code.size / 4; i++) {
 		if ((i & 0x3f) == 0)
 			nv_wr32(priv, 0x41a188, i >> 6);
-		nv_wr32(priv, 0x41a184, nvc0_grgpc_code[i]);
+		nv_wr32(priv, 0x41a184, oclass->gpccs.ucode->code.data[i]);
 	}
 	nv_wr32(priv, 0x000260, r000260);
 
+	if ((init = cclass->gpc[0]))
+		nvc0_graph_init_csdata(priv, init, 0x41a000, 0x000, 0x418000);
+	if ((init = cclass->gpc[2]))
+		nvc0_graph_init_csdata(priv, init, 0x41a000, 0x004, 0x419800);
+	if ((init = cclass->gpc[3]))
+		nvc0_graph_init_csdata(priv, init, 0x41a000, 0x008, 0x41be00);
+
 	/* start HUB ucode running, it'll init the GPCs */
-	nv_wr32(priv, 0x409800, nv_device(priv)->chipset);
 	nv_wr32(priv, 0x40910c, 0x00000000);
 	nv_wr32(priv, 0x409100, 0x00000002);
 	if (!nv_wait(priv, 0x409800, 0x80000000, 0x80000000)) {
@@ -913,29 +961,104 @@ nvc0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
 	return 0;
 }
 
-static int
+int
 nvc0_graph_init(struct nouveau_object *object)
 {
+	struct nvc0_graph_oclass *oclass = (void *)object->oclass;
 	struct nvc0_graph_priv *priv = (void *)object;
-	int ret;
+	const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, priv->tpc_total);
+	u32 data[TPC_MAX / 8] = {};
+	u8  tpcnr[GPC_MAX];
+	int gpc, tpc, rop;
+	int ret, i;
 
 	ret = nouveau_graph_init(&priv->base);
 	if (ret)
 		return ret;
 
-	nvc0_graph_init_obj418880(priv);
-	nvc0_graph_init_regs(priv);
-	/*nvc0_graph_init_unitplemented_magics(priv);*/
-	nvc0_graph_init_gpc_0(priv);
-	/*nvc0_graph_init_unitplemented_c242(priv);*/
+	nv_wr32(priv, GPC_BCAST(0x0880), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x08a4), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x0888), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x088c), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x0890), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x0894), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x08b4), priv->unk4188b4->addr >> 8);
+	nv_wr32(priv, GPC_BCAST(0x08b8), priv->unk4188b8->addr >> 8);
+
+	for (i = 0; oclass->mmio[i]; i++)
+		nvc0_graph_mmio(priv, oclass->mmio[i]);
+
+	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+	for (i = 0, gpc = -1; i < priv->tpc_total; i++) {
+		do {
+			gpc = (gpc + 1) % priv->gpc_nr;
+		} while (!tpcnr[gpc]);
+		tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
+
+		data[i / 8] |= tpc << ((i % 8) * 4);
+	}
+
+	nv_wr32(priv, GPC_BCAST(0x0980), data[0]);
+	nv_wr32(priv, GPC_BCAST(0x0984), data[1]);
+	nv_wr32(priv, GPC_BCAST(0x0988), data[2]);
+	nv_wr32(priv, GPC_BCAST(0x098c), data[3]);
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0914),
+			priv->magic_not_rop_nr << 8 | priv->tpc_nr[gpc]);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0910), 0x00040000 |
+			priv->tpc_total);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0918), magicgpc918);
+	}
+
+	if (nv_device(priv)->chipset != 0xd7)
+		nv_wr32(priv, GPC_BCAST(0x1bd4), magicgpc918);
+	else
+		nv_wr32(priv, GPC_BCAST(0x3fd4), magicgpc918);
+
+	nv_wr32(priv, GPC_BCAST(0x08ac), nv_rd32(priv, 0x100800));
 
 	nv_wr32(priv, 0x400500, 0x00010001);
+
 	nv_wr32(priv, 0x400100, 0xffffffff);
 	nv_wr32(priv, 0x40013c, 0xffffffff);
 
-	nvc0_graph_init_units(priv);
-	nvc0_graph_init_gpc_1(priv);
-	nvc0_graph_init_rop(priv);
+	nv_wr32(priv, 0x409c24, 0x000f0000);
+	nv_wr32(priv, 0x404000, 0xc0000000);
+	nv_wr32(priv, 0x404600, 0xc0000000);
+	nv_wr32(priv, 0x408030, 0xc0000000);
+	nv_wr32(priv, 0x40601c, 0xc0000000);
+	nv_wr32(priv, 0x404490, 0xc0000000);
+	nv_wr32(priv, 0x406018, 0xc0000000);
+	nv_wr32(priv, 0x405840, 0xc0000000);
+	nv_wr32(priv, 0x405844, 0x00ffffff);
+	nv_mask(priv, 0x419cc0, 0x00000008, 0x00000008);
+	nv_mask(priv, 0x419eb4, 0x00001000, 0x00001000);
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
+		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x508), 0xffffffff);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x50c), 0xffffffff);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x224), 0xc0000000);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x48c), 0xc0000000);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x084), 0xc0000000);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x644), 0x001ffffe);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x64c), 0x0000000f);
+		}
+		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0xffffffff);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x2c94), 0xffffffff);
+	}
+
+	for (rop = 0; rop < priv->rop_nr; rop++) {
+		nv_wr32(priv, ROP_UNIT(rop, 0x144), 0xc0000000);
+		nv_wr32(priv, ROP_UNIT(rop, 0x070), 0xc0000000);
+		nv_wr32(priv, ROP_UNIT(rop, 0x204), 0xffffffff);
+		nv_wr32(priv, ROP_UNIT(rop, 0x208), 0xffffffff);
+	}
 
 	nv_wr32(priv, 0x400108, 0xffffffff);
 	nv_wr32(priv, 0x400138, 0xffffffff);
@@ -943,22 +1066,205 @@ nvc0_graph_init(struct nouveau_object *object)
 	nv_wr32(priv, 0x400130, 0xffffffff);
 	nv_wr32(priv, 0x40011c, 0xffffffff);
 	nv_wr32(priv, 0x400134, 0xffffffff);
+
 	nv_wr32(priv, 0x400054, 0x34ce3464);
+	return nvc0_graph_init_ctxctl(priv);
+}
 
-	ret = nvc0_graph_init_ctxctl(priv);
+static void
+nvc0_graph_dtor_fw(struct nvc0_graph_fuc *fuc)
+{
+	kfree(fuc->data);
+	fuc->data = NULL;
+}
+
+int
+nvc0_graph_ctor_fw(struct nvc0_graph_priv *priv, const char *fwname,
+		   struct nvc0_graph_fuc *fuc)
+{
+	struct nouveau_device *device = nv_device(priv);
+	const struct firmware *fw;
+	char f[32];
+	int ret;
+
+	snprintf(f, sizeof(f), "nouveau/nv%02x_%s", device->chipset, fwname);
+	ret = request_firmware(&fw, f, &device->pdev->dev);
+	if (ret) {
+		snprintf(f, sizeof(f), "nouveau/%s", fwname);
+		ret = request_firmware(&fw, f, &device->pdev->dev);
+		if (ret) {
+			nv_error(priv, "failed to load %s\n", fwname);
+			return ret;
+		}
+	}
+
+	fuc->size = fw->size;
+	fuc->data = kmemdup(fw->data, fuc->size, GFP_KERNEL);
+	release_firmware(fw);
+	return (fuc->data != NULL) ? 0 : -ENOMEM;
+}
+
+void
+nvc0_graph_dtor(struct nouveau_object *object)
+{
+	struct nvc0_graph_priv *priv = (void *)object;
+
+	kfree(priv->data);
+
+	nvc0_graph_dtor_fw(&priv->fuc409c);
+	nvc0_graph_dtor_fw(&priv->fuc409d);
+	nvc0_graph_dtor_fw(&priv->fuc41ac);
+	nvc0_graph_dtor_fw(&priv->fuc41ad);
+
+	nouveau_gpuobj_ref(NULL, &priv->unk4188b8);
+	nouveau_gpuobj_ref(NULL, &priv->unk4188b4);
+
+	nouveau_graph_destroy(&priv->base);
+}
+
+int
+nvc0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *bclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nvc0_graph_oclass *oclass = (void *)bclass;
+	struct nouveau_device *device = nv_device(parent);
+	struct nvc0_graph_priv *priv;
+	int ret, i;
+
+	ret = nouveau_graph_create(parent, engine, bclass,
+				   (oclass->fecs.ucode != NULL), &priv);
+	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
+	nv_subdev(priv)->unit = 0x18001000;
+	nv_subdev(priv)->intr = nvc0_graph_intr;
+
+	priv->base.units = nvc0_graph_units;
+
+	if (nouveau_boolopt(device->cfgopt, "NvGrUseFW", false)) {
+		nv_info(priv, "using external firmware\n");
+		if (nvc0_graph_ctor_fw(priv, "fuc409c", &priv->fuc409c) ||
+		    nvc0_graph_ctor_fw(priv, "fuc409d", &priv->fuc409d) ||
+		    nvc0_graph_ctor_fw(priv, "fuc41ac", &priv->fuc41ac) ||
+		    nvc0_graph_ctor_fw(priv, "fuc41ad", &priv->fuc41ad))
+			return -EINVAL;
+		priv->firmware = true;
+	}
+
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
+				&priv->unk4188b4);
+	if (ret)
+		return ret;
+
+	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
+				&priv->unk4188b8);
+	if (ret)
+		return ret;
+
+	for (i = 0; i < 0x1000; i += 4) {
+		nv_wo32(priv->unk4188b4, i, 0x00000010);
+		nv_wo32(priv->unk4188b8, i, 0x00000010);
+	}
+
+	priv->rop_nr = (nv_rd32(priv, 0x409604) & 0x001f0000) >> 16;
+	priv->gpc_nr =  nv_rd32(priv, 0x409604) & 0x0000001f;
+	for (i = 0; i < priv->gpc_nr; i++) {
+		priv->tpc_nr[i]  = nv_rd32(priv, GPC_UNIT(i, 0x2608));
+		priv->tpc_total += priv->tpc_nr[i];
+	}
+
+	/*XXX: these need figuring out... though it might not even matter */
+	switch (nv_device(priv)->chipset) {
+	case 0xc0:
+		if (priv->tpc_total == 11) { /* 465, 3/4/4/0, 4 */
+			priv->magic_not_rop_nr = 0x07;
+		} else
+		if (priv->tpc_total == 14) { /* 470, 3/3/4/4, 5 */
+			priv->magic_not_rop_nr = 0x05;
+		} else
+		if (priv->tpc_total == 15) { /* 480, 3/4/4/4, 6 */
+			priv->magic_not_rop_nr = 0x06;
+		}
+		break;
+	case 0xc3: /* 450, 4/0/0/0, 2 */
+		priv->magic_not_rop_nr = 0x03;
+		break;
+	case 0xc4: /* 460, 3/4/0/0, 4 */
+		priv->magic_not_rop_nr = 0x01;
+		break;
+	case 0xc1: /* 2/0/0/0, 1 */
+		priv->magic_not_rop_nr = 0x01;
+		break;
+	case 0xc8: /* 4/4/3/4, 5 */
+		priv->magic_not_rop_nr = 0x06;
+		break;
+	case 0xce: /* 4/4/0/0, 4 */
+		priv->magic_not_rop_nr = 0x03;
+		break;
+	case 0xcf: /* 4/0/0/0, 3 */
+		priv->magic_not_rop_nr = 0x03;
+		break;
+	case 0xd7:
+	case 0xd9: /* 1/0/0/0, 1 */
+		priv->magic_not_rop_nr = 0x01;
+		break;
+	}
+
+	nv_engine(priv)->cclass = *oclass->cclass;
+	nv_engine(priv)->sclass =  oclass->sclass;
 	return 0;
 }
 
-struct nouveau_oclass
-nvc0_graph_oclass = {
-	.handle = NV_ENGINE(GR, 0xc0),
-	.ofuncs = &(struct nouveau_ofuncs) {
+struct nvc0_graph_init *
+nvc0_graph_init_mmio[] = {
+	nvc0_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvc0_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvc0_graph_init_gpc,
+	nvc0_graph_init_tpc,
+	nvc0_graph_init_unk88xx,
+	nvc0_graph_tpc_0,
+	NULL
+};
+
+#include "fuc/hubnvc0.fuc.h"
+
+struct nvc0_graph_ucode
+nvc0_graph_fecs_ucode = {
+	.code.data = nvc0_grhub_code,
+	.code.size = sizeof(nvc0_grhub_code),
+	.data.data = nvc0_grhub_data,
+	.data.size = sizeof(nvc0_grhub_data),
+};
+
+#include "fuc/gpcnvc0.fuc.h"
+
+struct nvc0_graph_ucode
+nvc0_graph_gpccs_ucode = {
+	.code.data = nvc0_grgpc_code,
+	.code.size = sizeof(nvc0_grgpc_code),
+	.data.data = nvc0_grgpc_data,
+	.data.size = sizeof(nvc0_grgpc_data),
+};
+
+struct nouveau_oclass *
+nvc0_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xc0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nvc0_graph_ctor,
 		.dtor = nvc0_graph_dtor,
 		.init = nvc0_graph_init,
 		.fini = _nouveau_graph_fini,
 	},
-};
+	.cclass = &nvc0_grctx_oclass,
+	.sclass =  nvc0_graph_sclass,
+	.mmio = nvc0_graph_init_mmio,
+	.fecs.ucode = &nvc0_graph_fecs_ucode,
+	.gpccs.ucode = &nvc0_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h
index c870dad0f670..ea17a80ad7fc 100644
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc0.h
@@ -38,8 +38,8 @@
 #include <engine/fifo.h>
 #include <engine/graph.h>
 
-#define GPC_MAX 4
-#define TPC_MAX 32
+#define GPC_MAX 32
+#define TPC_MAX (GPC_MAX * 8)
 
 #define ROP_BCAST(r)      (0x408800 + (r))
 #define ROP_UNIT(u, r)    (0x410000 + (u) * 0x400 + (r))
@@ -102,74 +102,187 @@ struct nvc0_graph_chan {
 	} data[4];
 };
 
-static inline u32
-nvc0_graph_class(void *obj)
-{
-	struct nouveau_device *device = nv_device(obj);
-
-	switch (device->chipset) {
-	case 0xc0:
-	case 0xc3:
-	case 0xc4:
-	case 0xce: /* guess, mmio trace shows only 0x9097 state */
-	case 0xcf: /* guess, mmio trace shows only 0x9097 state */
-		return 0x9097;
-	case 0xc1:
-		return 0x9197;
-	case 0xc8:
-	case 0xd9:
-	case 0xd7:
-		return 0x9297;
-	case 0xe4:
-	case 0xe7:
-	case 0xe6:
-		return 0xa097;
-	default:
-		return 0;
-	}
-}
-
-void nv_icmd(struct nvc0_graph_priv *priv, u32 icmd, u32 data);
-
-static inline void
-nv_mthd(struct nvc0_graph_priv *priv, u32 class, u32 mthd, u32 data)
-{
-	nv_wr32(priv, 0x40448c, data);
-	nv_wr32(priv, 0x404488, 0x80000000 | (mthd << 14) | class);
-}
+int  nvc0_grctx_generate(struct nvc0_graph_priv *);
+
+int  nvc0_graph_context_ctor(struct nouveau_object *, struct nouveau_object *,
+			     struct nouveau_oclass *, void *, u32,
+			     struct nouveau_object **);
+void nvc0_graph_context_dtor(struct nouveau_object *);
+
+void nvc0_graph_ctxctl_debug(struct nvc0_graph_priv *);
+
+u64  nvc0_graph_units(struct nouveau_graph *);
+int  nvc0_graph_ctor(struct nouveau_object *, struct nouveau_object *,
+		     struct nouveau_oclass *, void *data, u32 size,
+		     struct nouveau_object **);
+void nvc0_graph_dtor(struct nouveau_object *);
+int  nvc0_graph_init(struct nouveau_object *);
+int  nve4_graph_init(struct nouveau_object *);
+
+extern struct nouveau_oclass nvc0_graph_sclass[];
+
+extern struct nouveau_oclass nvc8_graph_sclass[];
+
+struct nvc0_graph_init {
+	u32 addr;
+	u8  count;
+	u8  pitch;
+	u32 data;
+};
+
+struct nvc0_graph_mthd {
+	u16 oclass;
+	struct nvc0_graph_init *init;
+};
 
 struct nvc0_grctx {
 	struct nvc0_graph_priv *priv;
 	struct nvc0_graph_data *data;
 	struct nvc0_graph_mmio *mmio;
-	struct nouveau_gpuobj *chan;
 	int buffer_nr;
 	u64 buffer[4];
 	u64 addr;
 };
 
+struct nvc0_grctx_oclass {
+	struct nouveau_oclass base;
+	/* main context generation function */
+	void  (*main)(struct nvc0_graph_priv *, struct nvc0_grctx *);
+	/* context-specific modify-on-first-load list generation function */
+	void  (*mods)(struct nvc0_graph_priv *, struct nvc0_grctx *);
+	void  (*unkn)(struct nvc0_graph_priv *);
+	/* mmio context data */
+	struct nvc0_graph_init **hub;
+	struct nvc0_graph_init **gpc;
+	/* indirect context data, generated with icmds/mthds */
+	struct nvc0_graph_init *icmd;
+	struct nvc0_graph_mthd *mthd;
+};
+
+struct nvc0_graph_ucode {
+	struct nvc0_graph_fuc code;
+	struct nvc0_graph_fuc data;
+};
+
+extern struct nvc0_graph_ucode nvc0_graph_fecs_ucode;
+extern struct nvc0_graph_ucode nvc0_graph_gpccs_ucode;
+
+struct nvc0_graph_oclass {
+	struct nouveau_oclass base;
+	struct nouveau_oclass **cclass;
+	struct nouveau_oclass *sclass;
+	struct nvc0_graph_init **mmio;
+	struct {
+		struct nvc0_graph_ucode *ucode;
+	} fecs;
+	struct {
+		struct nvc0_graph_ucode *ucode;
+	} gpccs;
+};
+
+void nvc0_graph_mmio(struct nvc0_graph_priv *, struct nvc0_graph_init *);
+void nvc0_graph_icmd(struct nvc0_graph_priv *, struct nvc0_graph_init *);
+void nvc0_graph_mthd(struct nvc0_graph_priv *, struct nvc0_graph_mthd *);
+int  nvc0_graph_init_ctxctl(struct nvc0_graph_priv *);
+
+extern struct nvc0_graph_init nvc0_graph_init_regs[];
+extern struct nvc0_graph_init nvc0_graph_init_unk40xx[];
+extern struct nvc0_graph_init nvc0_graph_init_unk44xx[];
+extern struct nvc0_graph_init nvc0_graph_init_unk78xx[];
+extern struct nvc0_graph_init nvc0_graph_init_unk60xx[];
+extern struct nvc0_graph_init nvc0_graph_init_unk58xx[];
+extern struct nvc0_graph_init nvc0_graph_init_unk80xx[];
+extern struct nvc0_graph_init nvc0_graph_init_gpc[];
+extern struct nvc0_graph_init nvc0_graph_init_unk88xx[];
+extern struct nvc0_graph_init nvc0_graph_tpc_0[];
+
+extern struct nvc0_graph_init nvc3_graph_init_unk58xx[];
+
+extern struct nvc0_graph_init nvd9_graph_init_unk58xx[];
+extern struct nvc0_graph_init nvd9_graph_init_unk64xx[];
+
+extern struct nvc0_graph_init nve4_graph_init_regs[];
+extern struct nvc0_graph_init nve4_graph_init_unk[];
+extern struct nvc0_graph_init nve4_graph_init_unk88xx[];
+
 int  nvc0_grctx_generate(struct nvc0_graph_priv *);
-int  nvc0_grctx_init(struct nvc0_graph_priv *, struct nvc0_grctx *);
-void nvc0_grctx_data(struct nvc0_grctx *, u32, u32, u32);
-void nvc0_grctx_mmio(struct nvc0_grctx *, u32, u32, u32, u32);
-int  nvc0_grctx_fini(struct nvc0_grctx *);
+void nvc0_grctx_generate_main(struct nvc0_graph_priv *, struct nvc0_grctx *);
+void nvc0_grctx_generate_mods(struct nvc0_graph_priv *, struct nvc0_grctx *);
+void nvc0_grctx_generate_unkn(struct nvc0_graph_priv *);
+void nvc0_grctx_generate_tpcid(struct nvc0_graph_priv *);
+void nvc0_grctx_generate_r406028(struct nvc0_graph_priv *);
+void nvc0_grctx_generate_r4060a8(struct nvc0_graph_priv *);
+void nvc0_grctx_generate_r418bb8(struct nvc0_graph_priv *);
+void nve4_grctx_generate_r418bb8(struct nvc0_graph_priv *);
+void nvc0_grctx_generate_r406800(struct nvc0_graph_priv *);
 
-int  nve0_grctx_generate(struct nvc0_graph_priv *);
+extern struct nouveau_oclass *nvc0_grctx_oclass;
+extern struct nvc0_graph_init *nvc0_grctx_init_hub[];
+extern struct nvc0_graph_init nvc0_grctx_init_base[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk40xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk44xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk46xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk47xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk60xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk64xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk78xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_unk80xx[];
+extern struct nvc0_graph_init nvc0_grctx_init_gpc_0[];
+extern struct nvc0_graph_init nvc0_grctx_init_gpc_1[];
+extern struct nvc0_graph_init nvc0_grctx_init_tpc[];
+extern struct nvc0_graph_init nvc0_grctx_init_icmd[];
+extern struct nvc0_graph_init nvd9_grctx_init_icmd[]; //
 
-#define mmio_data(s,a,p) nvc0_grctx_data(&info, (s), (a), (p))
-#define mmio_list(r,d,s,b) nvc0_grctx_mmio(&info, (r), (d), (s), (b))
+extern struct nvc0_graph_mthd nvc0_grctx_init_mthd[];
+extern struct nvc0_graph_init nvc0_grctx_init_902d[];
+extern struct nvc0_graph_init nvc0_grctx_init_9039[];
+extern struct nvc0_graph_init nvc0_grctx_init_90c0[];
+extern struct nvc0_graph_init nvc0_grctx_init_mthd_magic[];
 
-void nvc0_graph_ctxctl_debug(struct nvc0_graph_priv *);
-int  nvc0_graph_ctor_fw(struct nvc0_graph_priv *, const char *,
-			struct nvc0_graph_fuc *);
-void nvc0_graph_dtor(struct nouveau_object *);
-void nvc0_graph_init_fw(struct nvc0_graph_priv *, u32 base,
-			struct nvc0_graph_fuc *, struct nvc0_graph_fuc *);
-int  nvc0_graph_context_ctor(struct nouveau_object *, struct nouveau_object *,
-			     struct nouveau_oclass *, void *, u32,
-			     struct nouveau_object **);
-void nvc0_graph_context_dtor(struct nouveau_object *);
+void nvc1_grctx_generate_mods(struct nvc0_graph_priv *, struct nvc0_grctx *);
+void nvc1_grctx_generate_unkn(struct nvc0_graph_priv *);
+extern struct nouveau_oclass *nvc1_grctx_oclass;
+extern struct nvc0_graph_init nvc1_grctx_init_9097[];
+
+extern struct nouveau_oclass *nvc3_grctx_oclass;
+
+extern struct nouveau_oclass *nvc8_grctx_oclass;
+extern struct nvc0_graph_init nvc8_grctx_init_9197[];
+extern struct nvc0_graph_init nvc8_grctx_init_9297[];
+
+extern struct nouveau_oclass *nvd7_grctx_oclass;
+
+extern struct nouveau_oclass *nvd9_grctx_oclass;
+extern struct nvc0_graph_init nvd9_grctx_init_rop[];
+extern struct nvc0_graph_mthd nvd9_grctx_init_mthd[];
+
+void nve4_grctx_generate_main(struct nvc0_graph_priv *, struct nvc0_grctx *);
+void nve4_grctx_generate_unkn(struct nvc0_graph_priv *);
+extern struct nouveau_oclass *nve4_grctx_oclass;
+extern struct nvc0_graph_init nve4_grctx_init_unk46xx[];
+extern struct nvc0_graph_init nve4_grctx_init_unk47xx[];
+extern struct nvc0_graph_init nve4_grctx_init_unk58xx[];
+extern struct nvc0_graph_init nve4_grctx_init_unk80xx[];
+extern struct nvc0_graph_init nve4_grctx_init_unk90xx[];
+
+extern struct nouveau_oclass *nvf0_grctx_oclass;
+
+#define mmio_data(s,a,p) do {                                                  \
+	info->buffer[info->buffer_nr] = round_up(info->addr, (a));             \
+	info->addr = info->buffer[info->buffer_nr++] + (s);                    \
+	info->data->size = (s);                                                \
+	info->data->align = (a);                                               \
+	info->data->access = (p);                                              \
+	info->data++;                                                          \
+} while(0)
 
-u64 nvc0_graph_units(struct nouveau_graph *);
+#define mmio_list(r,d,s,b) do {                                                \
+	info->mmio->addr = (r);                                                \
+	info->mmio->data = (d);                                                \
+	info->mmio->shift = (s);                                               \
+	info->mmio->buffer = (b);                                              \
+	info->mmio++;                                                          \
+	nv_wr32(priv, (r), (d) | ((s) ? (info->buffer[(b)] >> (s)) : 0));      \
+} while(0)
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc1.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc1.c
new file mode 100644
index 000000000000..bc4a469b86cb
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc1.c
@@ -0,0 +1,144 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * Graphics object classes
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nvc1_graph_sclass[] = {
+	{ 0x902d, &nouveau_object_ofuncs },
+	{ 0x9039, &nouveau_object_ofuncs },
+	{ 0x9097, &nouveau_object_ofuncs },
+	{ 0x90c0, &nouveau_object_ofuncs },
+	{ 0x9197, &nouveau_object_ofuncs },
+	{}
+};
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+static struct nvc0_graph_init
+nvc1_graph_init_gpc[] = {
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x00000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   2, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000003 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc1_graph_init_tpc[] = {
+	{ 0x419d08,   2, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ac8,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x41980c,   2, 0x04, 0x00000000 },
+	{ 0x419814,   1, 0x04, 0x00000004 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x41984c,   1, 0x04, 0x00005bc5 },
+	{ 0x419850,   4, 0x04, 0x00000000 },
+	{ 0x419880,   1, 0x04, 0x00000002 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x80000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00008bf4 },
+	{ 0x419cbc,   1, 0x04, 0x28137606 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419bd4,   1, 0x04, 0x00800000 },
+	{ 0x419bdc,   1, 0x04, 0x00000000 },
+	{ 0x419d2c,   1, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00001100 },
+	{ 0x419eac,   1, 0x04, 0x11100702 },
+	{ 0x419eb0,   1, 0x04, 0x00000003 },
+	{ 0x419eb4,   4, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x0e063818 },
+	{ 0x419ecc,   1, 0x04, 0x0e060e06 },
+	{ 0x419ed0,   1, 0x04, 0x00003818 },
+	{ 0x419ed4,   1, 0x04, 0x011104f1 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f2c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init *
+nvc1_graph_init_mmio[] = {
+	nvc0_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvc3_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvc1_graph_init_gpc,
+	nvc1_graph_init_tpc,
+	nvc0_graph_init_unk88xx,
+	nvc0_graph_tpc_0,
+	NULL
+};
+
+struct nouveau_oclass *
+nvc1_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xc1),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nvc0_graph_init,
+		.fini = _nouveau_graph_fini,
+	},
+	.cclass = &nvc1_grctx_oclass,
+	.sclass = nvc1_graph_sclass,
+	.mmio = nvc1_graph_init_mmio,
+	.fecs.ucode = &nvc0_graph_fecs_ucode,
+	.gpccs.ucode = &nvc0_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc3.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc3.c
new file mode 100644
index 000000000000..d44b3b3ee800
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc3.c
@@ -0,0 +1,110 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+struct nvc0_graph_init
+nvc3_graph_init_unk58xx[] = {
+	{ 0x405844,   1, 0x04, 0x00ffffff },
+	{ 0x405850,   1, 0x04, 0x00000000 },
+	{ 0x405900,   1, 0x04, 0x00002834 },
+	{ 0x405908,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc3_graph_init_tpc[] = {
+	{ 0x419d08,   2, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ac8,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x41980c,   3, 0x04, 0x00000000 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x41984c,   1, 0x04, 0x00005bc5 },
+	{ 0x419850,   4, 0x04, 0x00000000 },
+	{ 0x419880,   1, 0x04, 0x00000002 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x80000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00008bf4 },
+	{ 0x419cbc,   1, 0x04, 0x28137606 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419bd4,   1, 0x04, 0x00800000 },
+	{ 0x419bdc,   1, 0x04, 0x00000000 },
+	{ 0x419d2c,   1, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00001100 },
+	{ 0x419eac,   1, 0x04, 0x11100702 },
+	{ 0x419eb0,   1, 0x04, 0x00000003 },
+	{ 0x419eb4,   4, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x0e063818 },
+	{ 0x419ecc,   1, 0x04, 0x0e060e06 },
+	{ 0x419ed0,   1, 0x04, 0x00003818 },
+	{ 0x419ed4,   1, 0x04, 0x011104f1 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f2c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init *
+nvc3_graph_init_mmio[] = {
+	nvc0_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvc3_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvc0_graph_init_gpc,
+	nvc3_graph_init_tpc,
+	nvc0_graph_init_unk88xx,
+	nvc0_graph_tpc_0,
+	NULL
+};
+
+struct nouveau_oclass *
+nvc3_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xc3),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nvc0_graph_init,
+		.fini = _nouveau_graph_fini,
+	},
+	.cclass = &nvc3_grctx_oclass,
+	.sclass = nvc0_graph_sclass,
+	.mmio = nvc3_graph_init_mmio,
+	.fecs.ucode = &nvc0_graph_fecs_ucode,
+	.gpccs.ucode = &nvc0_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvc8.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvc8.c
new file mode 100644
index 000000000000..02845e567314
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvc8.c
@@ -0,0 +1,141 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * Graphics object classes
+ ******************************************************************************/
+
+struct nouveau_oclass
+nvc8_graph_sclass[] = {
+	{ 0x902d, &nouveau_object_ofuncs },
+	{ 0x9039, &nouveau_object_ofuncs },
+	{ 0x9097, &nouveau_object_ofuncs },
+	{ 0x90c0, &nouveau_object_ofuncs },
+	{ 0x9197, &nouveau_object_ofuncs },
+	{ 0x9297, &nouveau_object_ofuncs },
+	{}
+};
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+static struct nvc0_graph_init
+nvc8_graph_init_gpc[] = {
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x80000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   2, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000050 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvc8_graph_init_tpc[] = {
+	{ 0x419d08,   2, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x41980c,   3, 0x04, 0x00000000 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x41984c,   1, 0x04, 0x00005bc5 },
+	{ 0x419850,   4, 0x04, 0x00000000 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x80000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00008bf4 },
+	{ 0x419cbc,   1, 0x04, 0x28137606 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419bd4,   1, 0x04, 0x00800000 },
+	{ 0x419bdc,   1, 0x04, 0x00000000 },
+	{ 0x419d2c,   1, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00001100 },
+	{ 0x419eac,   1, 0x04, 0x11100f02 },
+	{ 0x419eb0,   1, 0x04, 0x00000003 },
+	{ 0x419eb4,   4, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x06060618 },
+	{ 0x419ed0,   1, 0x04, 0x0eff0e38 },
+	{ 0x419ed4,   1, 0x04, 0x011104f1 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f2c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init *
+nvc8_graph_init_mmio[] = {
+	nvc0_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvc0_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvc8_graph_init_gpc,
+	nvc8_graph_init_tpc,
+	nvc0_graph_init_unk88xx,
+	nvc0_graph_tpc_0,
+	NULL
+};
+
+struct nouveau_oclass *
+nvc8_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xc8),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nvc0_graph_init,
+		.fini = _nouveau_graph_fini,
+	},
+	.cclass = &nvc8_grctx_oclass,
+	.sclass = nvc8_graph_sclass,
+	.mmio = nvc8_graph_init_mmio,
+	.fecs.ucode = &nvc0_graph_fecs_ucode,
+	.gpccs.ucode = &nvc0_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvd7.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvd7.c
new file mode 100644
index 000000000000..5052d7ab4d72
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvd7.c
@@ -0,0 +1,167 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+#include "fuc/hubnvd7.fuc.h"
+
+struct nvc0_graph_ucode
+nvd7_graph_fecs_ucode = {
+	.code.data = nvd7_grhub_code,
+	.code.size = sizeof(nvd7_grhub_code),
+	.data.data = nvd7_grhub_data,
+	.data.size = sizeof(nvd7_grhub_data),
+};
+
+#include "fuc/gpcnvd7.fuc.h"
+
+struct nvc0_graph_ucode
+nvd7_graph_gpccs_ucode = {
+	.code.data = nvd7_grgpc_code,
+	.code.size = sizeof(nvd7_grgpc_code),
+	.data.data = nvd7_grgpc_data,
+	.data.size = sizeof(nvd7_grgpc_data),
+};
+
+static struct nvc0_graph_init
+nvd7_graph_init_gpc[] = {
+	{ 0x418408,   1, 0x04, 0x00000000 },
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x4184a4,   2, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x00000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   2, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c64,   1, 0x04, 0x00000000 },
+	{ 0x418c68,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418cb4,   2, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418d28,   1, 0x04, 0x00000000 },
+	{ 0x418f00,   1, 0x04, 0x00000000 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418f20,   2, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000003 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x418e1c,   1, 0x04, 0x00000000 },
+	{ 0x418e20,   1, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_graph_init_tpc[] = {
+	{ 0x419d08,   2, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ac8,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x419ab4,   1, 0x04, 0x00000000 },
+	{ 0x41980c,   1, 0x04, 0x00000010 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x41984c,   1, 0x04, 0x00005bc8 },
+	{ 0x419850,   2, 0x04, 0x00000000 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x80000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00008bf4 },
+	{ 0x419cbc,   1, 0x04, 0x28137606 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x02001100 },
+	{ 0x419eac,   1, 0x04, 0x11100702 },
+	{ 0x419eb0,   1, 0x04, 0x00000003 },
+	{ 0x419eb4,   4, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x0e063818 },
+	{ 0x419ecc,   1, 0x04, 0x0e060e06 },
+	{ 0x419ed0,   1, 0x04, 0x00003818 },
+	{ 0x419ed4,   1, 0x04, 0x011104f1 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f2c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd7_graph_init_tpc_0[] = {
+	{ 0x40402c,   1, 0x04, 0x00000000 },
+	{ 0x4040f0,   1, 0x04, 0x00000000 },
+	{ 0x404174,   1, 0x04, 0x00000000 },
+	{ 0x503018,   1, 0x04, 0x00000001 },
+	{}
+};
+
+static struct nvc0_graph_init *
+nvd7_graph_init_mmio[] = {
+	nvc0_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvd9_graph_init_unk64xx,
+	nvd9_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvd7_graph_init_gpc,
+	nvd7_graph_init_tpc,
+	nve4_graph_init_unk,
+	nvc0_graph_init_unk88xx,
+	nvd7_graph_init_tpc_0,
+	NULL
+};
+
+struct nouveau_oclass *
+nvd7_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xd7),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nvc0_graph_init,
+		.fini = _nouveau_graph_fini,
+	},
+	.cclass = &nvd7_grctx_oclass,
+	.sclass = nvc8_graph_sclass,
+	.mmio = nvd7_graph_init_mmio,
+	.fecs.ucode = &nvd7_graph_fecs_ucode,
+	.gpccs.ucode = &nvd7_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvd9.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvd9.c
new file mode 100644
index 000000000000..652098e0df3f
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvd9.c
@@ -0,0 +1,165 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+struct nvc0_graph_init
+nvd9_graph_init_unk64xx[] = {
+	{ 0x4064f0,   3, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nvd9_graph_init_unk58xx[] = {
+	{ 0x405844,   1, 0x04, 0x00ffffff },
+	{ 0x405850,   1, 0x04, 0x00000000 },
+	{ 0x405900,   1, 0x04, 0x00002834 },
+	{ 0x405908,   1, 0x04, 0x00000000 },
+	{ 0x405928,   1, 0x04, 0x00000000 },
+	{ 0x40592c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd9_graph_init_gpc[] = {
+	{ 0x418408,   1, 0x04, 0x00000000 },
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x4184a4,   2, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x00000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   2, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c64,   1, 0x04, 0x00000000 },
+	{ 0x418c68,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418cb4,   2, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418d28,   1, 0x04, 0x00000000 },
+	{ 0x418d2c,   1, 0x04, 0x00000000 },
+	{ 0x418f00,   1, 0x04, 0x00000000 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418f20,   2, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000003 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x418e1c,   1, 0x04, 0x00000000 },
+	{ 0x418e20,   1, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvd9_graph_init_tpc[] = {
+	{ 0x419d08,   2, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ac8,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x419ab4,   1, 0x04, 0x00000000 },
+	{ 0x41980c,   1, 0x04, 0x00000010 },
+	{ 0x419810,   1, 0x04, 0x00000000 },
+	{ 0x419814,   1, 0x04, 0x00000004 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x41984c,   1, 0x04, 0x0000a918 },
+	{ 0x419850,   4, 0x04, 0x00000000 },
+	{ 0x419880,   1, 0x04, 0x00000002 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x80000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00008bf4 },
+	{ 0x419cbc,   1, 0x04, 0x28137606 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419bd4,   1, 0x04, 0x00800000 },
+	{ 0x419bdc,   1, 0x04, 0x00000000 },
+	{ 0x419bf8,   1, 0x04, 0x00000000 },
+	{ 0x419bfc,   1, 0x04, 0x00000000 },
+	{ 0x419d2c,   1, 0x04, 0x00000000 },
+	{ 0x419d48,   1, 0x04, 0x00000000 },
+	{ 0x419d4c,   1, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x02001100 },
+	{ 0x419eac,   1, 0x04, 0x11100702 },
+	{ 0x419eb0,   1, 0x04, 0x00000003 },
+	{ 0x419eb4,   4, 0x04, 0x00000000 },
+	{ 0x419ec8,   1, 0x04, 0x0e063818 },
+	{ 0x419ecc,   1, 0x04, 0x0e060e06 },
+	{ 0x419ed0,   1, 0x04, 0x00003818 },
+	{ 0x419ed4,   1, 0x04, 0x011104f1 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f2c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init *
+nvd9_graph_init_mmio[] = {
+	nvc0_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvd9_graph_init_unk64xx,
+	nvd9_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvd9_graph_init_gpc,
+	nvd9_graph_init_tpc,
+	nvc0_graph_init_unk88xx,
+	nvc0_graph_tpc_0,
+	NULL
+};
+
+struct nouveau_oclass *
+nvd9_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xd9),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nvc0_graph_init,
+		.fini = _nouveau_graph_fini,
+	},
+	.cclass = &nvd9_grctx_oclass,
+	.sclass = nvc8_graph_sclass,
+	.mmio = nvd9_graph_init_mmio,
+	.fecs.ucode = &nvc0_graph_fecs_ucode,
+	.gpccs.ucode = &nvc0_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
deleted file mode 100644
index 678c16f63055..000000000000
--- a/drivers/gpu/drm/nouveau/core/engine/graph/nve0.c
+++ /dev/null
@@ -1,807 +0,0 @@
-/*
- * Copyright 2012 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "nvc0.h"
-#include "fuc/hubnve0.fuc.h"
-#include "fuc/gpcnve0.fuc.h"
-
-/*******************************************************************************
- * Graphics object classes
- ******************************************************************************/
-
-static struct nouveau_oclass
-nve0_graph_sclass[] = {
-	{ 0x902d, &nouveau_object_ofuncs },
-	{ 0xa040, &nouveau_object_ofuncs },
-	{ 0xa097, &nouveau_object_ofuncs },
-	{ 0xa0c0, &nouveau_object_ofuncs },
-	{ 0xa0b5, &nouveau_object_ofuncs },
-	{}
-};
-
-/*******************************************************************************
- * PGRAPH context
- ******************************************************************************/
-
-static struct nouveau_oclass
-nve0_graph_cclass = {
-	.handle = NV_ENGCTX(GR, 0xe0),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nvc0_graph_context_ctor,
-		.dtor = nvc0_graph_context_dtor,
-		.init = _nouveau_graph_context_init,
-		.fini = _nouveau_graph_context_fini,
-		.rd32 = _nouveau_graph_context_rd32,
-		.wr32 = _nouveau_graph_context_wr32,
-	},
-};
-
-/*******************************************************************************
- * PGRAPH engine/subdev functions
- ******************************************************************************/
-
-static void
-nve0_graph_ctxctl_isr(struct nvc0_graph_priv *priv)
-{
-	u32 ustat = nv_rd32(priv, 0x409c18);
-
-	if (ustat & 0x00000001)
-		nv_error(priv, "CTXCTRL ucode error\n");
-	if (ustat & 0x00080000)
-		nv_error(priv, "CTXCTRL watchdog timeout\n");
-	if (ustat & ~0x00080001)
-		nv_error(priv, "CTXCTRL 0x%08x\n", ustat);
-
-	nvc0_graph_ctxctl_debug(priv);
-	nv_wr32(priv, 0x409c20, ustat);
-}
-
-static const struct nouveau_enum nve0_mp_warp_error[] = {
-	{ 0x00, "NO_ERROR" },
-	{ 0x01, "STACK_MISMATCH" },
-	{ 0x05, "MISALIGNED_PC" },
-	{ 0x08, "MISALIGNED_GPR" },
-	{ 0x09, "INVALID_OPCODE" },
-	{ 0x0d, "GPR_OUT_OF_BOUNDS" },
-	{ 0x0e, "MEM_OUT_OF_BOUNDS" },
-	{ 0x0f, "UNALIGNED_MEM_ACCESS" },
-	{ 0x11, "INVALID_PARAM" },
-	{}
-};
-
-static const struct nouveau_enum nve0_mp_global_error[] = {
-	{ 2, "MULTIPLE_WARP_ERRORS" },
-	{ 3, "OUT_OF_STACK_SPACE" },
-	{}
-};
-
-static const struct nouveau_enum nve0_gpc_rop_error[] = {
-	{ 1, "RT_PITCH_OVERRUN" },
-	{ 4, "RT_WIDTH_OVERRUN" },
-	{ 5, "RT_HEIGHT_OVERRUN" },
-	{ 7, "ZETA_STORAGE_TYPE_MISMATCH" },
-	{ 8, "RT_STORAGE_TYPE_MISMATCH" },
-	{ 10, "RT_LINEAR_MISMATCH" },
-	{}
-};
-
-static const struct nouveau_enum nve0_sked_error[] = {
-	{ 7, "CONSTANT_BUFFER_SIZE" },
-	{ 9, "LOCAL_MEMORY_SIZE_POS" },
-	{ 10, "LOCAL_MEMORY_SIZE_NEG" },
-	{ 11, "WARP_CSTACK_SIZE" },
-	{ 12, "TOTAL_TEMP_SIZE" },
-	{ 13, "REGISTER_COUNT" },
-	{ 18, "TOTAL_THREADS" },
-	{ 20, "PROGRAM_OFFSET" },
-	{ 21, "SHARED_MEMORY_SIZE" },
-	{ 25, "SHARED_CONFIG_TOO_SMALL" },
-	{ 26, "TOTAL_REGISTER_COUNT" },
-	{}
-};
-
-static void
-nve0_graph_mp_trap(struct nvc0_graph_priv *priv, int gpc, int tp)
-{
-	int i;
-	u32 werr = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x648));
-	u32 gerr = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x650));
-
-	nv_error(priv, "GPC%i/TP%i/MP trap:", gpc, tp);
-
-	for (i = 0; i <= 31; ++i) {
-		if (!(gerr & (1 << i)))
-			continue;
-		pr_cont(" ");
-		nouveau_enum_print(nve0_mp_global_error, i);
-	}
-	if (werr) {
-		pr_cont(" ");
-		nouveau_enum_print(nve0_mp_warp_error, werr & 0xffff);
-	}
-	pr_cont("\n");
-
-	/* disable MP trap to avoid spam */
-	nv_mask(priv, TPC_UNIT(gpc, tp, 0x50c), 0x2, 0x0);
-
-	/* TODO: figure out how to resume after an MP trap */
-}
-
-static void
-nve0_graph_tp_trap(struct nvc0_graph_priv *priv, int gpc, int tp)
-{
-	u32 stat = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x508));
-
-	if (stat & 0x1) {
-		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x224));
-		nv_error(priv, "GPC%i/TP%i/TEX trap: %08x\n",
-			 gpc, tp, trap);
-
-		nv_wr32(priv, TPC_UNIT(gpc, tp, 0x224), 0xc0000000);
-		stat &= ~0x1;
-	}
-
-	if (stat & 0x2) {
-		nve0_graph_mp_trap(priv, gpc, tp);
-		stat &= ~0x2;
-	}
-
-	if (stat & 0x4) {
-		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x084));
-		nv_error(priv, "GPC%i/TP%i/POLY trap: %08x\n",
-			 gpc, tp, trap);
-
-		nv_wr32(priv, TPC_UNIT(gpc, tp, 0x084), 0xc0000000);
-		stat &= ~0x4;
-	}
-
-	if (stat & 0x8) {
-		u32 trap = nv_rd32(priv, TPC_UNIT(gpc, tp, 0x48c));
-		nv_error(priv, "GPC%i/TP%i/L1C trap: %08x\n",
-			 gpc, tp, trap);
-
-		nv_wr32(priv, TPC_UNIT(gpc, tp, 0x48c), 0xc0000000);
-		stat &= ~0x8;
-	}
-
-	if (stat) {
-		nv_error(priv, "GPC%i/TP%i: unknown stat %08x\n",
-			 gpc, tp, stat);
-	}
-}
-
-static void
-nve0_graph_gpc_trap(struct nvc0_graph_priv *priv)
-{
-	const u32 mask = nv_rd32(priv, 0x400118);
-	int gpc;
-
-	for (gpc = 0; gpc < 4; ++gpc) {
-		u32 stat;
-		int tp;
-
-		if (!(mask & (1 << gpc)))
-			continue;
-		stat = nv_rd32(priv, GPC_UNIT(gpc, 0x2c90));
-
-		if (stat & 0x0001) {
-			u32 trap[4];
-			int i;
-
-			trap[0] = nv_rd32(priv, GPC_UNIT(gpc, 0x0420));
-			trap[1] = nv_rd32(priv, GPC_UNIT(gpc, 0x0434));
-			trap[2] = nv_rd32(priv, GPC_UNIT(gpc, 0x0438));
-			trap[3] = nv_rd32(priv, GPC_UNIT(gpc, 0x043c));
-
-			nv_error(priv, "GPC%i/PROP trap:", gpc);
-			for (i = 0; i <= 29; ++i) {
-				if (!(trap[0] & (1 << i)))
-					continue;
-				pr_cont(" ");
-				nouveau_enum_print(nve0_gpc_rop_error, i);
-			}
-			pr_cont("\n");
-
-			nv_error(priv, "x = %u, y = %u, "
-				 "format = %x, storage type = %x\n",
-				 trap[1] & 0xffff,
-				 trap[1] >> 16,
-				 (trap[2] >> 8) & 0x3f,
-				 trap[3] & 0xff);
-
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
-			stat &= ~0x0001;
-		}
-
-		if (stat & 0x0002) {
-			u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x0900));
-			nv_error(priv, "GPC%i/ZCULL trap: %08x\n", gpc,
-				 trap);
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
-			stat &= ~0x0002;
-		}
-
-		if (stat & 0x0004) {
-			u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x1028));
-			nv_error(priv, "GPC%i/CCACHE trap: %08x\n", gpc,
-				 trap);
-			nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
-			stat &= ~0x0004;
-		}
-
-		if (stat & 0x0008) {
-			u32 trap = nv_rd32(priv, GPC_UNIT(gpc, 0x0824));
-			nv_error(priv, "GPC%i/ESETUP trap %08x\n", gpc,
-				 trap);
-			nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
-			stat &= ~0x0008;
-		}
-
-		for (tp = 0; tp < 8; ++tp) {
-			if (stat & (1 << (16 + tp)))
-				nve0_graph_tp_trap(priv, gpc, tp);
-		}
-		stat &= ~0xff0000;
-
-		if (stat) {
-			nv_error(priv, "GPC%i: unknown stat %08x\n",
-				 gpc, stat);
-		}
-	}
-}
-
-
-static void
-nve0_graph_trap_isr(struct nvc0_graph_priv *priv, int chid, u64 inst,
-		struct nouveau_object *engctx)
-{
-	u32 trap = nv_rd32(priv, 0x400108);
-	int i;
-	int rop;
-
-	if (trap & 0x00000001) {
-		u32 stat = nv_rd32(priv, 0x404000);
-		nv_error(priv, "DISPATCH ch %d [0x%010llx %s] 0x%08x\n",
-			 chid, inst, nouveau_client_name(engctx), stat);
-		nv_wr32(priv, 0x404000, 0xc0000000);
-		nv_wr32(priv, 0x400108, 0x00000001);
-		trap &= ~0x00000001;
-	}
-
-	if (trap & 0x00000010) {
-		u32 stat = nv_rd32(priv, 0x405840);
-		nv_error(priv, "SHADER ch %d [0x%010llx %s] 0x%08x\n",
-			 chid, inst, nouveau_client_name(engctx), stat);
-		nv_wr32(priv, 0x405840, 0xc0000000);
-		nv_wr32(priv, 0x400108, 0x00000010);
-		trap &= ~0x00000010;
-	}
-
-	if (trap & 0x00000100) {
-		u32 stat = nv_rd32(priv, 0x407020);
-		nv_error(priv, "SKED ch %d [0x%010llx %s]:",
-			 chid, inst, nouveau_client_name(engctx));
-
-		for (i = 0; i <= 29; ++i) {
-			if (!(stat & (1 << i)))
-				continue;
-			pr_cont(" ");
-			nouveau_enum_print(nve0_sked_error, i);
-		}
-		pr_cont("\n");
-
-		if (stat & 0x3fffffff)
-			nv_wr32(priv, 0x407020, 0x40000000);
-		nv_wr32(priv, 0x400108, 0x00000100);
-		trap &= ~0x00000100;
-	}
-
-	if (trap & 0x01000000) {
-		nv_error(priv, "GPC ch %d [0x%010llx %s]:\n",
-			 chid, inst, nouveau_client_name(engctx));
-		nve0_graph_gpc_trap(priv);
-		trap &= ~0x01000000;
-	}
-
-	if (trap & 0x02000000) {
-		for (rop = 0; rop < priv->rop_nr; rop++) {
-			u32 statz = nv_rd32(priv, ROP_UNIT(rop, 0x070));
-			u32 statc = nv_rd32(priv, ROP_UNIT(rop, 0x144));
-			nv_error(priv,
-				 "ROP%d ch %d [0x%010llx %s] 0x%08x 0x%08x\n",
-				 rop, chid, inst, nouveau_client_name(engctx),
-				 statz, statc);
-			nv_wr32(priv, ROP_UNIT(rop, 0x070), 0xc0000000);
-			nv_wr32(priv, ROP_UNIT(rop, 0x144), 0xc0000000);
-		}
-		nv_wr32(priv, 0x400108, 0x02000000);
-		trap &= ~0x02000000;
-	}
-
-	if (trap) {
-		nv_error(priv, "TRAP ch %d [0x%010llx %s] 0x%08x\n",
-			 chid, inst, nouveau_client_name(engctx), trap);
-		nv_wr32(priv, 0x400108, trap);
-	}
-}
-
-static void
-nve0_graph_intr(struct nouveau_subdev *subdev)
-{
-	struct nouveau_fifo *pfifo = nouveau_fifo(subdev);
-	struct nouveau_engine *engine = nv_engine(subdev);
-	struct nouveau_object *engctx;
-	struct nouveau_handle *handle;
-	struct nvc0_graph_priv *priv = (void *)subdev;
-	u64 inst = nv_rd32(priv, 0x409b00) & 0x0fffffff;
-	u32 stat = nv_rd32(priv, 0x400100);
-	u32 addr = nv_rd32(priv, 0x400704);
-	u32 mthd = (addr & 0x00003ffc);
-	u32 subc = (addr & 0x00070000) >> 16;
-	u32 data = nv_rd32(priv, 0x400708);
-	u32 code = nv_rd32(priv, 0x400110);
-	u32 class = nv_rd32(priv, 0x404200 + (subc * 4));
-	int chid;
-
-	engctx = nouveau_engctx_get(engine, inst);
-	chid   = pfifo->chid(pfifo, engctx);
-
-	if (stat & 0x00000010) {
-		handle = nouveau_handle_get_class(engctx, class);
-		if (!handle || nv_call(handle->object, mthd, data)) {
-			nv_error(priv,
-				 "ILLEGAL_MTHD ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
-				 chid, inst, nouveau_client_name(engctx), subc,
-				 class, mthd, data);
-		}
-		nouveau_handle_put(handle);
-		nv_wr32(priv, 0x400100, 0x00000010);
-		stat &= ~0x00000010;
-	}
-
-	if (stat & 0x00000020) {
-		nv_error(priv,
-			 "ILLEGAL_CLASS ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
-			 chid, inst, nouveau_client_name(engctx), subc, class,
-			 mthd, data);
-		nv_wr32(priv, 0x400100, 0x00000020);
-		stat &= ~0x00000020;
-	}
-
-	if (stat & 0x00100000) {
-		nv_error(priv, "DATA_ERROR [");
-		nouveau_enum_print(nv50_data_error_names, code);
-		pr_cont("] ch %d [0x%010llx %s] subc %d class 0x%04x mthd 0x%04x data 0x%08x\n",
-			chid, inst, nouveau_client_name(engctx), subc, class,
-			mthd, data);
-		nv_wr32(priv, 0x400100, 0x00100000);
-		stat &= ~0x00100000;
-	}
-
-	if (stat & 0x00200000) {
-		nve0_graph_trap_isr(priv, chid, inst, engctx);
-		nv_wr32(priv, 0x400100, 0x00200000);
-		stat &= ~0x00200000;
-	}
-
-	if (stat & 0x00080000) {
-		nve0_graph_ctxctl_isr(priv);
-		nv_wr32(priv, 0x400100, 0x00080000);
-		stat &= ~0x00080000;
-	}
-
-	if (stat) {
-		nv_error(priv, "unknown stat 0x%08x\n", stat);
-		nv_wr32(priv, 0x400100, stat);
-	}
-
-	nv_wr32(priv, 0x400500, 0x00010001);
-	nouveau_engctx_put(engctx);
-}
-
-static int
-nve0_graph_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-	       struct nouveau_oclass *oclass, void *data, u32 size,
-	       struct nouveau_object **pobject)
-{
-	struct nouveau_device *device = nv_device(parent);
-	struct nvc0_graph_priv *priv;
-	int ret, i;
-
-	ret = nouveau_graph_create(parent, engine, oclass, true, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
-		return ret;
-
-	nv_subdev(priv)->unit = 0x18001000;
-	nv_subdev(priv)->intr = nve0_graph_intr;
-	nv_engine(priv)->cclass = &nve0_graph_cclass;
-	nv_engine(priv)->sclass = nve0_graph_sclass;
-
-	priv->base.units = nvc0_graph_units;
-
-	if (nouveau_boolopt(device->cfgopt, "NvGrUseFW", false)) {
-		nv_info(priv, "using external firmware\n");
-		if (nvc0_graph_ctor_fw(priv, "fuc409c", &priv->fuc409c) ||
-		    nvc0_graph_ctor_fw(priv, "fuc409d", &priv->fuc409d) ||
-		    nvc0_graph_ctor_fw(priv, "fuc41ac", &priv->fuc41ac) ||
-		    nvc0_graph_ctor_fw(priv, "fuc41ad", &priv->fuc41ad))
-			return -EINVAL;
-		priv->firmware = true;
-	}
-
-	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
-				&priv->unk4188b4);
-	if (ret)
-		return ret;
-
-	ret = nouveau_gpuobj_new(nv_object(priv), NULL, 0x1000, 256, 0,
-				&priv->unk4188b8);
-	if (ret)
-		return ret;
-
-	for (i = 0; i < 0x1000; i += 4) {
-		nv_wo32(priv->unk4188b4, i, 0x00000010);
-		nv_wo32(priv->unk4188b8, i, 0x00000010);
-	}
-
-	priv->gpc_nr =  nv_rd32(priv, 0x409604) & 0x0000001f;
-	priv->rop_nr = (nv_rd32(priv, 0x409604) & 0x001f0000) >> 16;
-	for (i = 0; i < priv->gpc_nr; i++) {
-		priv->tpc_nr[i] = nv_rd32(priv, GPC_UNIT(i, 0x2608));
-		priv->tpc_total += priv->tpc_nr[i];
-	}
-
-	switch (nv_device(priv)->chipset) {
-	case 0xe4:
-		if (priv->tpc_total == 8)
-			priv->magic_not_rop_nr = 3;
-		else
-		if (priv->tpc_total == 7)
-			priv->magic_not_rop_nr = 1;
-		break;
-	case 0xe7:
-	case 0xe6:
-		priv->magic_not_rop_nr = 1;
-		break;
-	default:
-		break;
-	}
-
-	return 0;
-}
-
-static void
-nve0_graph_init_obj418880(struct nvc0_graph_priv *priv)
-{
-	int i;
-
-	nv_wr32(priv, GPC_BCAST(0x0880), 0x00000000);
-	nv_wr32(priv, GPC_BCAST(0x08a4), 0x00000000);
-	for (i = 0; i < 4; i++)
-		nv_wr32(priv, GPC_BCAST(0x0888) + (i * 4), 0x00000000);
-	nv_wr32(priv, GPC_BCAST(0x08b4), priv->unk4188b4->addr >> 8);
-	nv_wr32(priv, GPC_BCAST(0x08b8), priv->unk4188b8->addr >> 8);
-}
-
-static void
-nve0_graph_init_regs(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x400080, 0x003083c2);
-	nv_wr32(priv, 0x400088, 0x0001ffe7);
-	nv_wr32(priv, 0x40008c, 0x00000000);
-	nv_wr32(priv, 0x400090, 0x00000030);
-	nv_wr32(priv, 0x40013c, 0x003901f7);
-	nv_wr32(priv, 0x400140, 0x00000100);
-	nv_wr32(priv, 0x400144, 0x00000000);
-	nv_wr32(priv, 0x400148, 0x00000110);
-	nv_wr32(priv, 0x400138, 0x00000000);
-	nv_wr32(priv, 0x400130, 0x00000000);
-	nv_wr32(priv, 0x400134, 0x00000000);
-	nv_wr32(priv, 0x400124, 0x00000002);
-}
-
-static void
-nve0_graph_init_units(struct nvc0_graph_priv *priv)
-{
-	nv_wr32(priv, 0x409ffc, 0x00000000);
-	nv_wr32(priv, 0x409c14, 0x00003e3e);
-	nv_wr32(priv, 0x409c24, 0x000f0000);
-
-	nv_wr32(priv, 0x404000, 0xc0000000);
-	nv_wr32(priv, 0x404600, 0xc0000000);
-	nv_wr32(priv, 0x408030, 0xc0000000);
-	nv_wr32(priv, 0x404490, 0xc0000000);
-	nv_wr32(priv, 0x406018, 0xc0000000);
-	nv_wr32(priv, 0x407020, 0xc0000000);
-	nv_wr32(priv, 0x405840, 0xc0000000);
-	nv_wr32(priv, 0x405844, 0x00ffffff);
-
-	nv_mask(priv, 0x419cc0, 0x00000008, 0x00000008);
-	nv_mask(priv, 0x419eb4, 0x00001000, 0x00001000);
-
-}
-
-static void
-nve0_graph_init_gpc_0(struct nvc0_graph_priv *priv)
-{
-	const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, priv->tpc_total);
-	u32 data[TPC_MAX / 8];
-	u8  tpcnr[GPC_MAX];
-	int i, gpc, tpc;
-
-	nv_wr32(priv, GPC_UNIT(0, 0x3018), 0x00000001);
-
-	memset(data, 0x00, sizeof(data));
-	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
-	for (i = 0, gpc = -1; i < priv->tpc_total; i++) {
-		do {
-			gpc = (gpc + 1) % priv->gpc_nr;
-		} while (!tpcnr[gpc]);
-		tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
-
-		data[i / 8] |= tpc << ((i % 8) * 4);
-	}
-
-	nv_wr32(priv, GPC_BCAST(0x0980), data[0]);
-	nv_wr32(priv, GPC_BCAST(0x0984), data[1]);
-	nv_wr32(priv, GPC_BCAST(0x0988), data[2]);
-	nv_wr32(priv, GPC_BCAST(0x098c), data[3]);
-
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0914), priv->magic_not_rop_nr << 8 |
-						  priv->tpc_nr[gpc]);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0910), 0x00040000 | priv->tpc_total);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0918), magicgpc918);
-	}
-
-	nv_wr32(priv, GPC_BCAST(0x3fd4), magicgpc918);
-	nv_wr32(priv, GPC_BCAST(0x08ac), nv_rd32(priv, 0x100800));
-}
-
-static void
-nve0_graph_init_gpc_1(struct nvc0_graph_priv *priv)
-{
-	int gpc, tpc;
-
-	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-		nv_wr32(priv, GPC_UNIT(gpc, 0x3038), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
-		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x508), 0xffffffff);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x50c), 0xffffffff);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x224), 0xc0000000);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x48c), 0xc0000000);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x084), 0xc0000000);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x644), 0x001ffffe);
-			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x64c), 0x0000000f);
-		}
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0xffffffff);
-		nv_wr32(priv, GPC_UNIT(gpc, 0x2c94), 0xffffffff);
-	}
-}
-
-static void
-nve0_graph_init_rop(struct nvc0_graph_priv *priv)
-{
-	int rop;
-
-	for (rop = 0; rop < priv->rop_nr; rop++) {
-		nv_wr32(priv, ROP_UNIT(rop, 0x144), 0xc0000000);
-		nv_wr32(priv, ROP_UNIT(rop, 0x070), 0xc0000000);
-		nv_wr32(priv, ROP_UNIT(rop, 0x204), 0xffffffff);
-		nv_wr32(priv, ROP_UNIT(rop, 0x208), 0xffffffff);
-	}
-}
-
-static int
-nve0_graph_init_ctxctl(struct nvc0_graph_priv *priv)
-{
-	u32 r000260;
-	int i;
-
-	if (priv->firmware) {
-		/* load fuc microcode */
-		r000260 = nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
-		nvc0_graph_init_fw(priv, 0x409000, &priv->fuc409c, &priv->fuc409d);
-		nvc0_graph_init_fw(priv, 0x41a000, &priv->fuc41ac, &priv->fuc41ad);
-		nv_wr32(priv, 0x000260, r000260);
-
-		/* start both of them running */
-		nv_wr32(priv, 0x409840, 0xffffffff);
-		nv_wr32(priv, 0x41a10c, 0x00000000);
-		nv_wr32(priv, 0x40910c, 0x00000000);
-		nv_wr32(priv, 0x41a100, 0x00000002);
-		nv_wr32(priv, 0x409100, 0x00000002);
-		if (!nv_wait(priv, 0x409800, 0x00000001, 0x00000001))
-			nv_error(priv, "0x409800 wait failed\n");
-
-		nv_wr32(priv, 0x409840, 0xffffffff);
-		nv_wr32(priv, 0x409500, 0x7fffffff);
-		nv_wr32(priv, 0x409504, 0x00000021);
-
-		nv_wr32(priv, 0x409840, 0xffffffff);
-		nv_wr32(priv, 0x409500, 0x00000000);
-		nv_wr32(priv, 0x409504, 0x00000010);
-		if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
-			nv_error(priv, "fuc09 req 0x10 timeout\n");
-			return -EBUSY;
-		}
-		priv->size = nv_rd32(priv, 0x409800);
-
-		nv_wr32(priv, 0x409840, 0xffffffff);
-		nv_wr32(priv, 0x409500, 0x00000000);
-		nv_wr32(priv, 0x409504, 0x00000016);
-		if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
-			nv_error(priv, "fuc09 req 0x16 timeout\n");
-			return -EBUSY;
-		}
-
-		nv_wr32(priv, 0x409840, 0xffffffff);
-		nv_wr32(priv, 0x409500, 0x00000000);
-		nv_wr32(priv, 0x409504, 0x00000025);
-		if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
-			nv_error(priv, "fuc09 req 0x25 timeout\n");
-			return -EBUSY;
-		}
-
-		nv_wr32(priv, 0x409800, 0x00000000);
-		nv_wr32(priv, 0x409500, 0x00000001);
-		nv_wr32(priv, 0x409504, 0x00000030);
-		if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
-			nv_error(priv, "fuc09 req 0x30 timeout\n");
-			return -EBUSY;
-		}
-
-		nv_wr32(priv, 0x409810, 0xb00095c8);
-		nv_wr32(priv, 0x409800, 0x00000000);
-		nv_wr32(priv, 0x409500, 0x00000001);
-		nv_wr32(priv, 0x409504, 0x00000031);
-		if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
-			nv_error(priv, "fuc09 req 0x31 timeout\n");
-			return -EBUSY;
-		}
-
-		nv_wr32(priv, 0x409810, 0x00080420);
-		nv_wr32(priv, 0x409800, 0x00000000);
-		nv_wr32(priv, 0x409500, 0x00000001);
-		nv_wr32(priv, 0x409504, 0x00000032);
-		if (!nv_wait_ne(priv, 0x409800, 0xffffffff, 0x00000000)) {
-			nv_error(priv, "fuc09 req 0x32 timeout\n");
-			return -EBUSY;
-		}
-
-		nv_wr32(priv, 0x409614, 0x00000070);
-		nv_wr32(priv, 0x409614, 0x00000770);
-		nv_wr32(priv, 0x40802c, 0x00000001);
-
-		if (priv->data == NULL) {
-			int ret = nve0_grctx_generate(priv);
-			if (ret) {
-				nv_error(priv, "failed to construct context\n");
-				return ret;
-			}
-		}
-
-		return 0;
-	}
-
-	/* load HUB microcode */
-	r000260 = nv_mask(priv, 0x000260, 0x00000001, 0x00000000);
-	nv_wr32(priv, 0x4091c0, 0x01000000);
-	for (i = 0; i < sizeof(nve0_grhub_data) / 4; i++)
-		nv_wr32(priv, 0x4091c4, nve0_grhub_data[i]);
-
-	nv_wr32(priv, 0x409180, 0x01000000);
-	for (i = 0; i < sizeof(nve0_grhub_code) / 4; i++) {
-		if ((i & 0x3f) == 0)
-			nv_wr32(priv, 0x409188, i >> 6);
-		nv_wr32(priv, 0x409184, nve0_grhub_code[i]);
-	}
-
-	/* load GPC microcode */
-	nv_wr32(priv, 0x41a1c0, 0x01000000);
-	for (i = 0; i < sizeof(nve0_grgpc_data) / 4; i++)
-		nv_wr32(priv, 0x41a1c4, nve0_grgpc_data[i]);
-
-	nv_wr32(priv, 0x41a180, 0x01000000);
-	for (i = 0; i < sizeof(nve0_grgpc_code) / 4; i++) {
-		if ((i & 0x3f) == 0)
-			nv_wr32(priv, 0x41a188, i >> 6);
-		nv_wr32(priv, 0x41a184, nve0_grgpc_code[i]);
-	}
-	nv_wr32(priv, 0x000260, r000260);
-
-	/* start HUB ucode running, it'll init the GPCs */
-	nv_wr32(priv, 0x409800, nv_device(priv)->chipset);
-	nv_wr32(priv, 0x40910c, 0x00000000);
-	nv_wr32(priv, 0x409100, 0x00000002);
-	if (!nv_wait(priv, 0x409800, 0x80000000, 0x80000000)) {
-		nv_error(priv, "HUB_INIT timed out\n");
-		nvc0_graph_ctxctl_debug(priv);
-		return -EBUSY;
-	}
-
-	priv->size = nv_rd32(priv, 0x409804);
-	if (priv->data == NULL) {
-		int ret = nve0_grctx_generate(priv);
-		if (ret) {
-			nv_error(priv, "failed to construct context\n");
-			return ret;
-		}
-	}
-
-	return 0;
-}
-
-static int
-nve0_graph_init(struct nouveau_object *object)
-{
-	struct nvc0_graph_priv *priv = (void *)object;
-	int ret;
-
-	ret = nouveau_graph_init(&priv->base);
-	if (ret)
-		return ret;
-
-	nve0_graph_init_obj418880(priv);
-	nve0_graph_init_regs(priv);
-	nve0_graph_init_gpc_0(priv);
-
-	nv_wr32(priv, 0x400500, 0x00010001);
-	nv_wr32(priv, 0x400100, 0xffffffff);
-	nv_wr32(priv, 0x40013c, 0xffffffff);
-
-	nve0_graph_init_units(priv);
-	nve0_graph_init_gpc_1(priv);
-	nve0_graph_init_rop(priv);
-
-	nv_wr32(priv, 0x400108, 0xffffffff);
-	nv_wr32(priv, 0x400138, 0xffffffff);
-	nv_wr32(priv, 0x400118, 0xffffffff);
-	nv_wr32(priv, 0x400130, 0xffffffff);
-	nv_wr32(priv, 0x40011c, 0xffffffff);
-	nv_wr32(priv, 0x400134, 0xffffffff);
-	nv_wr32(priv, 0x400054, 0x34ce3464);
-
-	ret = nve0_graph_init_ctxctl(priv);
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
-struct nouveau_oclass
-nve0_graph_oclass = {
-	.handle = NV_ENGINE(GR, 0xe0),
-	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = nve0_graph_ctor,
-		.dtor = nvc0_graph_dtor,
-		.init = nve0_graph_init,
-		.fini = _nouveau_graph_fini,
-	},
-};
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nve4.c b/drivers/gpu/drm/nouveau/core/engine/graph/nve4.c
new file mode 100644
index 000000000000..05ec09c88517
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nve4.c
@@ -0,0 +1,354 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * Graphics object classes
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nve4_graph_sclass[] = {
+	{ 0x902d, &nouveau_object_ofuncs },
+	{ 0xa040, &nouveau_object_ofuncs },
+	{ 0xa097, &nouveau_object_ofuncs },
+	{ 0xa0c0, &nouveau_object_ofuncs },
+	{}
+};
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+struct nvc0_graph_init
+nve4_graph_init_regs[] = {
+	{ 0x400080,   1, 0x04, 0x003083c2 },
+	{ 0x400088,   1, 0x04, 0x0001ffe7 },
+	{ 0x40008c,   1, 0x04, 0x00000000 },
+	{ 0x400090,   1, 0x04, 0x00000030 },
+	{ 0x40013c,   1, 0x04, 0x003901f7 },
+	{ 0x400140,   1, 0x04, 0x00000100 },
+	{ 0x400144,   1, 0x04, 0x00000000 },
+	{ 0x400148,   1, 0x04, 0x00000110 },
+	{ 0x400138,   1, 0x04, 0x00000000 },
+	{ 0x400130,   2, 0x04, 0x00000000 },
+	{ 0x400124,   1, 0x04, 0x00000002 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_graph_init_unk58xx[] = {
+	{ 0x405844,   1, 0x04, 0x00ffffff },
+	{ 0x405850,   1, 0x04, 0x00000000 },
+	{ 0x405900,   1, 0x04, 0x0000ff34 },
+	{ 0x405908,   1, 0x04, 0x00000000 },
+	{ 0x405928,   1, 0x04, 0x00000000 },
+	{ 0x40592c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_graph_init_unk70xx[] = {
+	{ 0x407010,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_graph_init_unk5bxx[] = {
+	{ 0x405b50,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_graph_init_gpc[] = {
+	{ 0x418408,   1, 0x04, 0x00000000 },
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x4184a4,   2, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x00000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   2, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c64,   1, 0x04, 0x00000000 },
+	{ 0x418c68,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418cb4,   2, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418d28,   1, 0x04, 0x00000000 },
+	{ 0x418d2c,   1, 0x04, 0x00000000 },
+	{ 0x418f00,   1, 0x04, 0x00000000 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418f20,   2, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000060 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x418e1c,   1, 0x04, 0x00000000 },
+	{ 0x418e20,   1, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nve4_graph_init_tpc[] = {
+	{ 0x419d0c,   1, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ac8,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x419ab4,   1, 0x04, 0x00000000 },
+	{ 0x41980c,   1, 0x04, 0x00000010 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x419850,   1, 0x04, 0x00000004 },
+	{ 0x419854,   2, 0x04, 0x00000000 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x01000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00b08bea },
+	{ 0x419c84,   1, 0x04, 0x00010384 },
+	{ 0x419cbc,   1, 0x04, 0x28137646 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419c80,   1, 0x04, 0x00020232 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000000 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ee4,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00000000 },
+	{ 0x419eb4,   1, 0x04, 0x00000000 },
+	{ 0x419eb8,   3, 0x04, 0x00000000 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419f74,   1, 0x04, 0x00000555 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_graph_init_unk[] = {
+	{ 0x41be04,   1, 0x04, 0x00000000 },
+	{ 0x41be08,   1, 0x04, 0x00000004 },
+	{ 0x41be0c,   1, 0x04, 0x00000000 },
+	{ 0x41be10,   1, 0x04, 0x003b8bc7 },
+	{ 0x41be14,   2, 0x04, 0x00000000 },
+	{ 0x41bfd4,   1, 0x04, 0x00800000 },
+	{ 0x41bfdc,   1, 0x04, 0x00000000 },
+	{ 0x41bff8,   1, 0x04, 0x00000000 },
+	{ 0x41bffc,   1, 0x04, 0x00000000 },
+	{ 0x41becc,   1, 0x04, 0x00000000 },
+	{ 0x41bee8,   1, 0x04, 0x00000000 },
+	{ 0x41beec,   1, 0x04, 0x00000000 },
+	{}
+};
+
+struct nvc0_graph_init
+nve4_graph_init_unk88xx[] = {
+	{ 0x40880c,   1, 0x04, 0x00000000 },
+	{ 0x408850,   1, 0x04, 0x00000004 },
+	{ 0x408910,   9, 0x04, 0x00000000 },
+	{ 0x408950,   1, 0x04, 0x00000000 },
+	{ 0x408954,   1, 0x04, 0x0000ffff },
+	{ 0x408958,   1, 0x04, 0x00000034 },
+	{ 0x408984,   1, 0x04, 0x00000000 },
+	{ 0x408988,   1, 0x04, 0x08040201 },
+	{ 0x40898c,   1, 0x04, 0x80402010 },
+	{}
+};
+
+int
+nve4_graph_init(struct nouveau_object *object)
+{
+	struct nvc0_graph_oclass *oclass = (void *)object->oclass;
+	struct nvc0_graph_priv *priv = (void *)object;
+	const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, priv->tpc_total);
+	u32 data[TPC_MAX / 8] = {};
+	u8  tpcnr[GPC_MAX];
+	int gpc, tpc, rop;
+	int ret, i;
+
+	ret = nouveau_graph_init(&priv->base);
+	if (ret)
+		return ret;
+
+	nv_wr32(priv, GPC_BCAST(0x0880), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x08a4), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x0888), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x088c), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x0890), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x0894), 0x00000000);
+	nv_wr32(priv, GPC_BCAST(0x08b4), priv->unk4188b4->addr >> 8);
+	nv_wr32(priv, GPC_BCAST(0x08b8), priv->unk4188b8->addr >> 8);
+
+	for (i = 0; oclass->mmio[i]; i++)
+		nvc0_graph_mmio(priv, oclass->mmio[i]);
+
+	nv_wr32(priv, GPC_UNIT(0, 0x3018), 0x00000001);
+
+	memset(data, 0x00, sizeof(data));
+	memcpy(tpcnr, priv->tpc_nr, sizeof(priv->tpc_nr));
+	for (i = 0, gpc = -1; i < priv->tpc_total; i++) {
+		do {
+			gpc = (gpc + 1) % priv->gpc_nr;
+		} while (!tpcnr[gpc]);
+		tpc = priv->tpc_nr[gpc] - tpcnr[gpc]--;
+
+		data[i / 8] |= tpc << ((i % 8) * 4);
+	}
+
+	nv_wr32(priv, GPC_BCAST(0x0980), data[0]);
+	nv_wr32(priv, GPC_BCAST(0x0984), data[1]);
+	nv_wr32(priv, GPC_BCAST(0x0988), data[2]);
+	nv_wr32(priv, GPC_BCAST(0x098c), data[3]);
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0914),
+			priv->magic_not_rop_nr << 8 | priv->tpc_nr[gpc]);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0910), 0x00040000 |
+			priv->tpc_total);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0918), magicgpc918);
+	}
+
+	nv_wr32(priv, GPC_BCAST(0x3fd4), magicgpc918);
+	nv_wr32(priv, GPC_BCAST(0x08ac), nv_rd32(priv, 0x100800));
+
+	nv_wr32(priv, 0x400500, 0x00010001);
+
+	nv_wr32(priv, 0x400100, 0xffffffff);
+	nv_wr32(priv, 0x40013c, 0xffffffff);
+
+	nv_wr32(priv, 0x409ffc, 0x00000000);
+	nv_wr32(priv, 0x409c14, 0x00003e3e);
+	nv_wr32(priv, 0x409c24, 0x000f0001);
+	nv_wr32(priv, 0x404000, 0xc0000000);
+	nv_wr32(priv, 0x404600, 0xc0000000);
+	nv_wr32(priv, 0x408030, 0xc0000000);
+	nv_wr32(priv, 0x404490, 0xc0000000);
+	nv_wr32(priv, 0x406018, 0xc0000000);
+	nv_wr32(priv, 0x407020, 0x40000000);
+	nv_wr32(priv, 0x405840, 0xc0000000);
+	nv_wr32(priv, 0x405844, 0x00ffffff);
+	nv_mask(priv, 0x419cc0, 0x00000008, 0x00000008);
+	nv_mask(priv, 0x419eb4, 0x00001000, 0x00001000);
+
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		nv_wr32(priv, GPC_UNIT(gpc, 0x3038), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0420), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0900), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x1028), 0xc0000000);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x0824), 0xc0000000);
+		for (tpc = 0; tpc < priv->tpc_nr[gpc]; tpc++) {
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x508), 0xffffffff);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x50c), 0xffffffff);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x224), 0xc0000000);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x48c), 0xc0000000);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x084), 0xc0000000);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x644), 0x001ffffe);
+			nv_wr32(priv, TPC_UNIT(gpc, tpc, 0x64c), 0x0000000f);
+		}
+		nv_wr32(priv, GPC_UNIT(gpc, 0x2c90), 0xffffffff);
+		nv_wr32(priv, GPC_UNIT(gpc, 0x2c94), 0xffffffff);
+	}
+
+	for (rop = 0; rop < priv->rop_nr; rop++) {
+		nv_wr32(priv, ROP_UNIT(rop, 0x144), 0xc0000000);
+		nv_wr32(priv, ROP_UNIT(rop, 0x070), 0xc0000000);
+		nv_wr32(priv, ROP_UNIT(rop, 0x204), 0xffffffff);
+		nv_wr32(priv, ROP_UNIT(rop, 0x208), 0xffffffff);
+	}
+
+	nv_wr32(priv, 0x400108, 0xffffffff);
+	nv_wr32(priv, 0x400138, 0xffffffff);
+	nv_wr32(priv, 0x400118, 0xffffffff);
+	nv_wr32(priv, 0x400130, 0xffffffff);
+	nv_wr32(priv, 0x40011c, 0xffffffff);
+	nv_wr32(priv, 0x400134, 0xffffffff);
+
+	nv_wr32(priv, 0x400054, 0x34ce3464);
+	return nvc0_graph_init_ctxctl(priv);
+}
+
+static struct nvc0_graph_init *
+nve4_graph_init_mmio[] = {
+	nve4_graph_init_regs,
+	nvc0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvd9_graph_init_unk64xx,
+	nve4_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nve4_graph_init_unk70xx,
+	nve4_graph_init_unk5bxx,
+	nve4_graph_init_gpc,
+	nve4_graph_init_tpc,
+	nve4_graph_init_unk,
+	nve4_graph_init_unk88xx,
+	NULL
+};
+
+#include "fuc/hubnve0.fuc.h"
+
+static struct nvc0_graph_ucode
+nve4_graph_fecs_ucode = {
+	.code.data = nve0_grhub_code,
+	.code.size = sizeof(nve0_grhub_code),
+	.data.data = nve0_grhub_data,
+	.data.size = sizeof(nve0_grhub_data),
+};
+
+#include "fuc/gpcnve0.fuc.h"
+
+static struct nvc0_graph_ucode
+nve4_graph_gpccs_ucode = {
+	.code.data = nve0_grgpc_code,
+	.code.size = sizeof(nve0_grgpc_code),
+	.data.data = nve0_grgpc_data,
+	.data.size = sizeof(nve0_grgpc_data),
+};
+
+struct nouveau_oclass *
+nve4_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xe4),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nve4_graph_init,
+		.fini = _nouveau_graph_fini,
+	},
+	.cclass = &nve4_grctx_oclass,
+	.sclass = nve4_graph_sclass,
+	.mmio = nve4_graph_init_mmio,
+	.fecs.ucode = &nve4_graph_fecs_ucode,
+	.gpccs.ucode = &nve4_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/graph/nvf0.c b/drivers/gpu/drm/nouveau/core/engine/graph/nvf0.c
new file mode 100644
index 000000000000..2f0ac7832234
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/graph/nvf0.c
@@ -0,0 +1,248 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs <bskeggs@redhat.com>
+ */
+
+#include "nvc0.h"
+
+/*******************************************************************************
+ * Graphics object classes
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nvf0_graph_sclass[] = {
+	{ 0x902d, &nouveau_object_ofuncs },
+	{ 0xa140, &nouveau_object_ofuncs },
+	{ 0xa197, &nouveau_object_ofuncs },
+	{ 0xa1c0, &nouveau_object_ofuncs },
+	{}
+};
+
+/*******************************************************************************
+ * PGRAPH engine/subdev functions
+ ******************************************************************************/
+
+static struct nvc0_graph_init
+nvf0_graph_init_unk40xx[] = {
+	{ 0x40415c,   1, 0x04, 0x00000000 },
+	{ 0x404170,   1, 0x04, 0x00000000 },
+	{ 0x4041b4,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_graph_init_unk58xx[] = {
+	{ 0x405844,   1, 0x04, 0x00ffffff },
+	{ 0x405850,   1, 0x04, 0x00000000 },
+	{ 0x405900,   1, 0x04, 0x0000ff00 },
+	{ 0x405908,   1, 0x04, 0x00000000 },
+	{ 0x405928,   1, 0x04, 0x00000000 },
+	{ 0x40592c,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_graph_init_unk70xx[] = {
+	{ 0x407010,   1, 0x04, 0x00000000 },
+	{ 0x407040,   1, 0x04, 0x80440424 },
+	{ 0x407048,   1, 0x04, 0x0000000a },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_graph_init_unk5bxx[] = {
+	{ 0x405b44,   1, 0x04, 0x00000000 },
+	{ 0x405b50,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_graph_init_gpc[] = {
+	{ 0x418408,   1, 0x04, 0x00000000 },
+	{ 0x4184a0,   1, 0x04, 0x00000000 },
+	{ 0x4184a4,   2, 0x04, 0x00000000 },
+	{ 0x418604,   1, 0x04, 0x00000000 },
+	{ 0x418680,   1, 0x04, 0x00000000 },
+	{ 0x418714,   1, 0x04, 0x00000000 },
+	{ 0x418384,   1, 0x04, 0x00000000 },
+	{ 0x418814,   3, 0x04, 0x00000000 },
+	{ 0x418b04,   1, 0x04, 0x00000000 },
+	{ 0x4188c8,   2, 0x04, 0x00000000 },
+	{ 0x4188d0,   1, 0x04, 0x00010000 },
+	{ 0x4188d4,   1, 0x04, 0x00000001 },
+	{ 0x418910,   1, 0x04, 0x00010001 },
+	{ 0x418914,   1, 0x04, 0x00000301 },
+	{ 0x418918,   1, 0x04, 0x00800000 },
+	{ 0x418980,   1, 0x04, 0x77777770 },
+	{ 0x418984,   3, 0x04, 0x77777777 },
+	{ 0x418c04,   1, 0x04, 0x00000000 },
+	{ 0x418c64,   1, 0x04, 0x00000000 },
+	{ 0x418c68,   1, 0x04, 0x00000000 },
+	{ 0x418c88,   1, 0x04, 0x00000000 },
+	{ 0x418cb4,   2, 0x04, 0x00000000 },
+	{ 0x418d00,   1, 0x04, 0x00000000 },
+	{ 0x418d28,   1, 0x04, 0x00000000 },
+	{ 0x418d2c,   1, 0x04, 0x00000000 },
+	{ 0x418f00,   1, 0x04, 0x00000400 },
+	{ 0x418f08,   1, 0x04, 0x00000000 },
+	{ 0x418f20,   1, 0x04, 0x00000000 },
+	{ 0x418f24,   1, 0x04, 0x00000000 },
+	{ 0x418e00,   1, 0x04, 0x00000000 },
+	{ 0x418e08,   1, 0x04, 0x00000000 },
+	{ 0x418e1c,   2, 0x04, 0x00000000 },
+	{ 0x41900c,   1, 0x04, 0x00000000 },
+	{ 0x419018,   1, 0x04, 0x00000000 },
+	{}
+};
+
+static struct nvc0_graph_init
+nvf0_graph_init_tpc[] = {
+	{ 0x419d0c,   1, 0x04, 0x00000000 },
+	{ 0x419d10,   1, 0x04, 0x00000014 },
+	{ 0x419ab0,   1, 0x04, 0x00000000 },
+	{ 0x419ac8,   1, 0x04, 0x00000000 },
+	{ 0x419ab8,   1, 0x04, 0x000000e7 },
+	{ 0x419aec,   1, 0x04, 0x00000000 },
+	{ 0x419abc,   2, 0x04, 0x00000000 },
+	{ 0x419ab4,   1, 0x04, 0x00000000 },
+	{ 0x419aa8,   2, 0x04, 0x00000000 },
+	{ 0x41980c,   1, 0x04, 0x00000010 },
+	{ 0x419844,   1, 0x04, 0x00000000 },
+	{ 0x419850,   1, 0x04, 0x00000004 },
+	{ 0x419854,   2, 0x04, 0x00000000 },
+	{ 0x419c98,   1, 0x04, 0x00000000 },
+	{ 0x419ca8,   1, 0x04, 0x00000000 },
+	{ 0x419cb0,   1, 0x04, 0x01000000 },
+	{ 0x419cb4,   1, 0x04, 0x00000000 },
+	{ 0x419cb8,   1, 0x04, 0x00b08bea },
+	{ 0x419c84,   1, 0x04, 0x00010384 },
+	{ 0x419cbc,   1, 0x04, 0x281b3646 },
+	{ 0x419cc0,   2, 0x04, 0x00000000 },
+	{ 0x419c80,   1, 0x04, 0x00020230 },
+	{ 0x419ccc,   2, 0x04, 0x00000000 },
+	{ 0x419c0c,   1, 0x04, 0x00000000 },
+	{ 0x419e00,   1, 0x04, 0x00000080 },
+	{ 0x419ea0,   1, 0x04, 0x00000000 },
+	{ 0x419ee4,   1, 0x04, 0x00000000 },
+	{ 0x419ea4,   1, 0x04, 0x00000100 },
+	{ 0x419ea8,   1, 0x04, 0x00000000 },
+	{ 0x419eb4,   1, 0x04, 0x00000000 },
+	{ 0x419ebc,   2, 0x04, 0x00000000 },
+	{ 0x419edc,   1, 0x04, 0x00000000 },
+	{ 0x419f00,   1, 0x04, 0x00000000 },
+	{ 0x419ed0,   1, 0x04, 0x00003234 },
+	{ 0x419f74,   1, 0x04, 0x00015555 },
+	{ 0x419f80,   4, 0x04, 0x00000000 },
+	{}
+};
+
+static int
+nvf0_graph_fini(struct nouveau_object *object, bool suspend)
+{
+	struct nvc0_graph_priv *priv = (void *)object;
+	static const struct {
+		u32 addr;
+		u32 data;
+	} magic[] = {
+		{ 0x020520, 0xfffffffc },
+		{ 0x020524, 0xfffffffe },
+		{ 0x020524, 0xfffffffc },
+		{ 0x020524, 0xfffffff8 },
+		{ 0x020524, 0xffffffe0 },
+		{ 0x020530, 0xfffffffe },
+		{ 0x02052c, 0xfffffffa },
+		{ 0x02052c, 0xfffffff0 },
+		{ 0x02052c, 0xffffffc0 },
+		{ 0x02052c, 0xffffff00 },
+		{ 0x02052c, 0xfffffc00 },
+		{ 0x02052c, 0xfffcfc00 },
+		{ 0x02052c, 0xfff0fc00 },
+		{ 0x02052c, 0xff80fc00 },
+		{ 0x020528, 0xfffffffe },
+		{ 0x020528, 0xfffffffc },
+	};
+	int i;
+
+	nv_mask(priv, 0x000200, 0x08001000, 0x00000000);
+	nv_mask(priv, 0x0206b4, 0x00000000, 0x00000000);
+	for (i = 0; i < ARRAY_SIZE(magic); i++) {
+		nv_wr32(priv, magic[i].addr, magic[i].data);
+		nv_wait(priv, magic[i].addr, 0x80000000, 0x00000000);
+	}
+
+	return nouveau_graph_fini(&priv->base, suspend);
+}
+
+static struct nvc0_graph_init *
+nvf0_graph_init_mmio[] = {
+	nve4_graph_init_regs,
+	nvf0_graph_init_unk40xx,
+	nvc0_graph_init_unk44xx,
+	nvc0_graph_init_unk78xx,
+	nvc0_graph_init_unk60xx,
+	nvd9_graph_init_unk64xx,
+	nvf0_graph_init_unk58xx,
+	nvc0_graph_init_unk80xx,
+	nvf0_graph_init_unk70xx,
+	nvf0_graph_init_unk5bxx,
+	nvf0_graph_init_gpc,
+	nvf0_graph_init_tpc,
+	nve4_graph_init_unk,
+	nve4_graph_init_unk88xx,
+	NULL
+};
+
+#include "fuc/hubnvf0.fuc.h"
+
+static struct nvc0_graph_ucode
+nvf0_graph_fecs_ucode = {
+	.code.data = nvf0_grhub_code,
+	.code.size = sizeof(nvf0_grhub_code),
+	.data.data = nvf0_grhub_data,
+	.data.size = sizeof(nvf0_grhub_data),
+};
+
+#include "fuc/gpcnvf0.fuc.h"
+
+static struct nvc0_graph_ucode
+nvf0_graph_gpccs_ucode = {
+	.code.data = nvf0_grgpc_code,
+	.code.size = sizeof(nvf0_grgpc_code),
+	.data.data = nvf0_grgpc_data,
+	.data.size = sizeof(nvf0_grgpc_data),
+};
+
+struct nouveau_oclass *
+nvf0_graph_oclass = &(struct nvc0_graph_oclass) {
+	.base.handle = NV_ENGINE(GR, 0xf0),
+	.base.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_graph_ctor,
+		.dtor = nvc0_graph_dtor,
+		.init = nve4_graph_init,
+		.fini = nvf0_graph_fini,
+	},
+	.cclass = &nvf0_grctx_oclass,
+	.sclass =  nvf0_graph_sclass,
+	.mmio = nvf0_graph_init_mmio,
+	.fecs.ucode = 0 ? &nvf0_graph_fecs_ucode : NULL,
+	.gpccs.ucode = &nvf0_graph_gpccs_ucode,
+}.base;
diff --git a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv50.c b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv50.c
index bc7d12b30fc1..37a2bd9e8078 100644
--- a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv50.c
@@ -125,13 +125,6 @@ nv50_mpeg_cclass = {
  * PMPEG engine/subdev functions
  ******************************************************************************/
 
-int
-nv50_mpeg_tlb_flush(struct nouveau_engine *engine)
-{
-	nv50_vm_flush_engine(&engine->base, 0x08);
-	return 0;
-}
-
 void
 nv50_mpeg_intr(struct nouveau_subdev *subdev)
 {
@@ -191,7 +184,6 @@ nv50_mpeg_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->intr = nv50_vpe_intr;
 	nv_engine(priv)->cclass = &nv50_mpeg_cclass;
 	nv_engine(priv)->sclass = nv50_mpeg_sclass;
-	nv_engine(priv)->tlb_flush = nv50_mpeg_tlb_flush;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv84.c b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv84.c
index 8f805b44d59e..96f5aa92677b 100644
--- a/drivers/gpu/drm/nouveau/core/engine/mpeg/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/mpeg/nv84.c
@@ -88,7 +88,6 @@ nv84_mpeg_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->intr = nv50_mpeg_intr;
 	nv_engine(priv)->cclass = &nv84_mpeg_cclass;
 	nv_engine(priv)->sclass = nv84_mpeg_sclass;
-	nv_engine(priv)->tlb_flush = nv50_mpeg_tlb_flush;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/engine/ppp/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/ppp/nvc0.c
index ebf0d860e2dd..73719aaa62d6 100644
--- a/drivers/gpu/drm/nouveau/core/engine/ppp/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/ppp/nvc0.c
@@ -22,8 +22,7 @@
  * Authors: Maarten Lankhorst
  */
 
-#include <core/falcon.h>
-
+#include <engine/falcon.h>
 #include <engine/ppp.h>
 
 struct nvc0_ppp_priv {
@@ -91,6 +90,7 @@ nvc0_ppp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	nv_subdev(priv)->unit = 0x00000002;
+	nv_subdev(priv)->intr = nouveau_falcon_intr;
 	nv_engine(priv)->cclass = &nvc0_ppp_cclass;
 	nv_engine(priv)->sclass = nvc0_ppp_sclass;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/engine/vp/nv84.c b/drivers/gpu/drm/nouveau/core/engine/vp/nv84.c
index 261cd96e6951..fd6272b8cdb2 100644
--- a/drivers/gpu/drm/nouveau/core/engine/vp/nv84.c
+++ b/drivers/gpu/drm/nouveau/core/engine/vp/nv84.c
@@ -19,24 +19,19 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
- * Authors: Ben Skeggs
+ * Authors: Ben Skeggs, Ilia Mirkin
  */
 
-#include <core/engctx.h>
-#include <core/class.h>
-
+#include <engine/xtensa.h>
 #include <engine/vp.h>
 
-struct nv84_vp_priv {
-	struct nouveau_engine base;
-};
-
 /*******************************************************************************
  * VP object classes
  ******************************************************************************/
 
 static struct nouveau_oclass
 nv84_vp_sclass[] = {
+	{ 0x7476, &nouveau_object_ofuncs },
 	{},
 };
 
@@ -48,7 +43,7 @@ static struct nouveau_oclass
 nv84_vp_cclass = {
 	.handle = NV_ENGCTX(VP, 0x84),
 	.ofuncs = &(struct nouveau_ofuncs) {
-		.ctor = _nouveau_engctx_ctor,
+		.ctor = _nouveau_xtensa_engctx_ctor,
 		.dtor = _nouveau_engctx_dtor,
 		.init = _nouveau_engctx_init,
 		.fini = _nouveau_engctx_fini,
@@ -66,10 +61,10 @@ nv84_vp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	     struct nouveau_oclass *oclass, void *data, u32 size,
 	     struct nouveau_object **pobject)
 {
-	struct nv84_vp_priv *priv;
+	struct nouveau_xtensa *priv;
 	int ret;
 
-	ret = nouveau_engine_create(parent, engine, oclass, true,
+	ret = nouveau_xtensa_create(parent, engine, oclass, 0xf000, true,
 				    "PVP", "vp", &priv);
 	*pobject = nv_object(priv);
 	if (ret)
@@ -78,6 +73,8 @@ nv84_vp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	nv_subdev(priv)->unit = 0x01020000;
 	nv_engine(priv)->cclass = &nv84_vp_cclass;
 	nv_engine(priv)->sclass = nv84_vp_sclass;
+	priv->fifo_val = 0x111;
+	priv->unkd28 = 0x9c544;
 	return 0;
 }
 
@@ -86,8 +83,10 @@ nv84_vp_oclass = {
 	.handle = NV_ENGINE(VP, 0x84),
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv84_vp_ctor,
-		.dtor = _nouveau_engine_dtor,
-		.init = _nouveau_engine_init,
-		.fini = _nouveau_engine_fini,
+		.dtor = _nouveau_xtensa_dtor,
+		.init = _nouveau_xtensa_init,
+		.fini = _nouveau_xtensa_fini,
+		.rd32 = _nouveau_xtensa_rd32,
+		.wr32 = _nouveau_xtensa_wr32,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/engine/vp/nv98.c b/drivers/gpu/drm/nouveau/core/engine/vp/nv98.c
new file mode 100644
index 000000000000..8a8236bc84de
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/vp/nv98.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright 2012 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <core/engctx.h>
+#include <core/class.h>
+
+#include <engine/vp.h>
+
+struct nv98_vp_priv {
+	struct nouveau_engine base;
+};
+
+/*******************************************************************************
+ * VP object classes
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nv98_vp_sclass[] = {
+	{},
+};
+
+/*******************************************************************************
+ * PVP context
+ ******************************************************************************/
+
+static struct nouveau_oclass
+nv98_vp_cclass = {
+	.handle = NV_ENGCTX(VP, 0x98),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = _nouveau_engctx_ctor,
+		.dtor = _nouveau_engctx_dtor,
+		.init = _nouveau_engctx_init,
+		.fini = _nouveau_engctx_fini,
+		.rd32 = _nouveau_engctx_rd32,
+		.wr32 = _nouveau_engctx_wr32,
+	},
+};
+
+/*******************************************************************************
+ * PVP engine/subdev functions
+ ******************************************************************************/
+
+static int
+nv98_vp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+	     struct nouveau_oclass *oclass, void *data, u32 size,
+	     struct nouveau_object **pobject)
+{
+	struct nv98_vp_priv *priv;
+	int ret;
+
+	ret = nouveau_engine_create(parent, engine, oclass, true,
+				    "PVP", "vp", &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	nv_subdev(priv)->unit = 0x01020000;
+	nv_engine(priv)->cclass = &nv98_vp_cclass;
+	nv_engine(priv)->sclass = nv98_vp_sclass;
+	return 0;
+}
+
+struct nouveau_oclass
+nv98_vp_oclass = {
+	.handle = NV_ENGINE(VP, 0x98),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv98_vp_ctor,
+		.dtor = _nouveau_engine_dtor,
+		.init = _nouveau_engine_init,
+		.fini = _nouveau_engine_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/engine/vp/nvc0.c b/drivers/gpu/drm/nouveau/core/engine/vp/nvc0.c
index f761949d7039..ac1f62aace72 100644
--- a/drivers/gpu/drm/nouveau/core/engine/vp/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/vp/nvc0.c
@@ -22,8 +22,7 @@
  * Authors: Maarten Lankhorst
  */
 
-#include <core/falcon.h>
-
+#include <engine/falcon.h>
 #include <engine/vp.h>
 
 struct nvc0_vp_priv {
@@ -91,6 +90,7 @@ nvc0_vp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	nv_subdev(priv)->unit = 0x00020000;
+	nv_subdev(priv)->intr = nouveau_falcon_intr;
 	nv_engine(priv)->cclass = &nvc0_vp_cclass;
 	nv_engine(priv)->sclass = nvc0_vp_sclass;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/engine/vp/nve0.c b/drivers/gpu/drm/nouveau/core/engine/vp/nve0.c
index 2384ce5dbe16..d4c3108479c9 100644
--- a/drivers/gpu/drm/nouveau/core/engine/vp/nve0.c
+++ b/drivers/gpu/drm/nouveau/core/engine/vp/nve0.c
@@ -22,8 +22,7 @@
  * Authors: Ben Skeggs
  */
 
-#include <core/falcon.h>
-
+#include <engine/falcon.h>
 #include <engine/vp.h>
 
 struct nve0_vp_priv {
@@ -91,6 +90,7 @@ nve0_vp_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	nv_subdev(priv)->unit = 0x00020000;
+	nv_subdev(priv)->intr = nouveau_falcon_intr;
 	nv_engine(priv)->cclass = &nve0_vp_cclass;
 	nv_engine(priv)->sclass = nve0_vp_sclass;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/engine/xtensa.c b/drivers/gpu/drm/nouveau/core/engine/xtensa.c
new file mode 100644
index 000000000000..0639bc59d0a5
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/engine/xtensa.c
@@ -0,0 +1,170 @@
+/*
+ * Copyright 2013 Ilia Mirkin
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include <engine/xtensa.h>
+
+u32
+_nouveau_xtensa_rd32(struct nouveau_object *object, u64 addr)
+{
+	struct nouveau_xtensa *xtensa = (void *)object;
+	return nv_rd32(xtensa, xtensa->addr + addr);
+}
+
+void
+_nouveau_xtensa_wr32(struct nouveau_object *object, u64 addr, u32 data)
+{
+	struct nouveau_xtensa *xtensa = (void *)object;
+	nv_wr32(xtensa, xtensa->addr + addr, data);
+}
+
+int
+_nouveau_xtensa_engctx_ctor(struct nouveau_object *parent,
+			    struct nouveau_object *engine,
+			    struct nouveau_oclass *oclass, void *data, u32 size,
+			    struct nouveau_object **pobject)
+{
+	struct nouveau_engctx *engctx;
+	int ret;
+
+	ret = nouveau_engctx_create(parent, engine, oclass, NULL,
+				    0x10000, 0x1000,
+				    NVOBJ_FLAG_ZERO_ALLOC, &engctx);
+	*pobject = nv_object(engctx);
+	return ret;
+}
+
+void
+_nouveau_xtensa_intr(struct nouveau_subdev *subdev)
+{
+	struct nouveau_xtensa *xtensa = (void *)subdev;
+	u32 unk104 = nv_ro32(xtensa, 0xd04);
+	u32 intr = nv_ro32(xtensa, 0xc20);
+	u32 chan = nv_ro32(xtensa, 0xc28);
+	u32 unk10c = nv_ro32(xtensa, 0xd0c);
+
+	if (intr & 0x10)
+		nv_warn(xtensa, "Watchdog interrupt, engine hung.\n");
+	nv_wo32(xtensa, 0xc20, intr);
+	intr = nv_ro32(xtensa, 0xc20);
+	if (unk104 == 0x10001 && unk10c == 0x200 && chan && !intr) {
+		nv_debug(xtensa, "Enabling FIFO_CTRL\n");
+		nv_mask(xtensa, xtensa->addr + 0xd94, 0, xtensa->fifo_val);
+	}
+}
+
+int
+nouveau_xtensa_create_(struct nouveau_object *parent,
+		       struct nouveau_object *engine,
+		       struct nouveau_oclass *oclass, u32 addr, bool enable,
+		       const char *iname, const char *fname,
+		       int length, void **pobject)
+{
+	struct nouveau_xtensa *xtensa;
+	int ret;
+
+	ret = nouveau_engine_create_(parent, engine, oclass, enable, iname,
+				     fname, length, pobject);
+	xtensa = *pobject;
+	if (ret)
+		return ret;
+
+	nv_subdev(xtensa)->intr = _nouveau_xtensa_intr;
+
+	xtensa->addr = addr;
+
+	return 0;
+}
+
+int
+_nouveau_xtensa_init(struct nouveau_object *object)
+{
+	struct nouveau_device *device = nv_device(object);
+	struct nouveau_xtensa *xtensa = (void *)object;
+	const struct firmware *fw;
+	char name[32];
+	int i, ret;
+	u32 tmp;
+
+	ret = nouveau_engine_init(&xtensa->base);
+	if (ret)
+		return ret;
+
+	if (!xtensa->gpu_fw) {
+		snprintf(name, sizeof(name), "nouveau/nv84_xuc%03x",
+			 xtensa->addr >> 12);
+
+		ret = request_firmware(&fw, name, &device->pdev->dev);
+		if (ret) {
+			nv_warn(xtensa, "unable to load firmware %s\n", name);
+			return ret;
+		}
+
+		ret = nouveau_gpuobj_new(object, NULL, fw->size, 0x1000, 0,
+					 &xtensa->gpu_fw);
+		if (ret) {
+			release_firmware(fw);
+			return ret;
+		}
+
+		nv_debug(xtensa, "Loading firmware to address: 0x%llx\n",
+			 xtensa->gpu_fw->addr);
+
+		for (i = 0; i < fw->size / 4; i++)
+			nv_wo32(xtensa->gpu_fw, i * 4, *((u32 *)fw->data + i));
+		release_firmware(fw);
+	}
+
+	nv_wo32(xtensa, 0xd10, 0x1fffffff); /* ?? */
+	nv_wo32(xtensa, 0xd08, 0x0fffffff); /* ?? */
+
+	nv_wo32(xtensa, 0xd28, xtensa->unkd28); /* ?? */
+	nv_wo32(xtensa, 0xc20, 0x3f); /* INTR */
+	nv_wo32(xtensa, 0xd84, 0x3f); /* INTR_EN */
+
+	nv_wo32(xtensa, 0xcc0, xtensa->gpu_fw->addr >> 8); /* XT_REGION_BASE */
+	nv_wo32(xtensa, 0xcc4, 0x1c); /* XT_REGION_SETUP */
+	nv_wo32(xtensa, 0xcc8, xtensa->gpu_fw->size >> 8); /* XT_REGION_LIMIT */
+
+	tmp = nv_rd32(xtensa, 0x0);
+	nv_wo32(xtensa, 0xde0, tmp); /* SCRATCH_H2X */
+
+	nv_wo32(xtensa, 0xce8, 0xf); /* XT_REGION_SETUP */
+
+	nv_wo32(xtensa, 0xc20, 0x3f); /* INTR */
+	nv_wo32(xtensa, 0xd84, 0x3f); /* INTR_EN */
+
+	return 0;
+}
+
+int
+_nouveau_xtensa_fini(struct nouveau_object *object, bool suspend)
+{
+	struct nouveau_xtensa *xtensa = (void *)object;
+
+	nv_wo32(xtensa, 0xd84, 0); /* INTR_EN */
+	nv_wo32(xtensa, 0xd94, 0); /* FIFO_CTRL */
+
+	if (!suspend)
+		nouveau_gpuobj_ref(NULL, &xtensa->gpu_fw);
+
+	return nouveau_engine_fini(&xtensa->base, suspend);
+}
diff --git a/drivers/gpu/drm/nouveau/core/include/core/device.h b/drivers/gpu/drm/nouveau/core/include/core/device.h
index 05840f3eee98..99b6600fe80a 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/device.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/device.h
@@ -17,8 +17,7 @@ enum nv_subdev_type {
 	NVDEV_SUBDEV_DEVINIT,
 	NVDEV_SUBDEV_GPIO,
 	NVDEV_SUBDEV_I2C,
-	NVDEV_SUBDEV_CLOCK,
-	NVDEV_SUBDEV_DEVINIT_LAST = NVDEV_SUBDEV_CLOCK,
+	NVDEV_SUBDEV_DEVINIT_LAST = NVDEV_SUBDEV_I2C,
 
 	/* This grouping of subdevs are initialised right after they've
 	 * been created, and are allowed to assume any subdevs in the
@@ -35,6 +34,7 @@ enum nv_subdev_type {
 	NVDEV_SUBDEV_VM,
 	NVDEV_SUBDEV_BAR,
 	NVDEV_SUBDEV_VOLT,
+	NVDEV_SUBDEV_CLOCK,
 	NVDEV_SUBDEV_THERM,
 
 	NVDEV_ENGINE_DMAOBJ,
@@ -49,6 +49,7 @@ enum nv_subdev_type {
 	NVDEV_ENGINE_PPP,
 	NVDEV_ENGINE_COPY0,
 	NVDEV_ENGINE_COPY1,
+	NVDEV_ENGINE_COPY2,
 	NVDEV_ENGINE_UNK1C1,
 	NVDEV_ENGINE_VENC,
 	NVDEV_ENGINE_DISP,
diff --git a/drivers/gpu/drm/nouveau/core/include/core/mm.h b/drivers/gpu/drm/nouveau/core/include/core/mm.h
index 2514e81ade02..2bf7d0e32261 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/mm.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/mm.h
@@ -15,8 +15,6 @@ struct nouveau_mm {
 	struct list_head nodes;
 	struct list_head free;
 
-	struct mutex mutex;
-
 	u32 block_size;
 	int heap_nodes;
 };
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/bsp.h b/drivers/gpu/drm/nouveau/core/include/engine/bsp.h
index 13ccdf54dfad..67662e2c4547 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/bsp.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/bsp.h
@@ -2,6 +2,7 @@
 #define __NOUVEAU_BSP_H__
 
 extern struct nouveau_oclass nv84_bsp_oclass;
+extern struct nouveau_oclass nv98_bsp_oclass;
 extern struct nouveau_oclass nvc0_bsp_oclass;
 extern struct nouveau_oclass nve0_bsp_oclass;
 
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/copy.h b/drivers/gpu/drm/nouveau/core/include/engine/copy.h
index 8cad2cf28cef..316a28ae5f5c 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/copy.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/copy.h
@@ -8,5 +8,6 @@ extern struct nouveau_oclass nvc0_copy0_oclass;
 extern struct nouveau_oclass nvc0_copy1_oclass;
 extern struct nouveau_oclass nve0_copy0_oclass;
 extern struct nouveau_oclass nve0_copy1_oclass;
+extern struct nouveau_oclass nve0_copy2_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/core/falcon.h b/drivers/gpu/drm/nouveau/core/include/engine/falcon.h
index 1edec386ab36..181aa7da524d 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/falcon.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/falcon.h
@@ -72,6 +72,8 @@ int nouveau_falcon_create_(struct nouveau_object *, struct nouveau_object *,
 			   struct nouveau_oclass *, u32, bool, const char *,
 			   const char *, int, void **);
 
+void nouveau_falcon_intr(struct nouveau_subdev *subdev);
+
 #define _nouveau_falcon_dtor _nouveau_engine_dtor
 int  _nouveau_falcon_init(struct nouveau_object *);
 int  _nouveau_falcon_fini(struct nouveau_object *, bool);
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/graph.h b/drivers/gpu/drm/nouveau/core/include/engine/graph.h
index 5d392439f2ac..8e1b52312ddc 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/graph.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/graph.h
@@ -61,8 +61,14 @@ extern struct nouveau_oclass nv34_graph_oclass;
 extern struct nouveau_oclass nv35_graph_oclass;
 extern struct nouveau_oclass nv40_graph_oclass;
 extern struct nouveau_oclass nv50_graph_oclass;
-extern struct nouveau_oclass nvc0_graph_oclass;
-extern struct nouveau_oclass nve0_graph_oclass;
+extern struct nouveau_oclass *nvc0_graph_oclass;
+extern struct nouveau_oclass *nvc1_graph_oclass;
+extern struct nouveau_oclass *nvc3_graph_oclass;
+extern struct nouveau_oclass *nvc8_graph_oclass;
+extern struct nouveau_oclass *nvd7_graph_oclass;
+extern struct nouveau_oclass *nvd9_graph_oclass;
+extern struct nouveau_oclass *nve4_graph_oclass;
+extern struct nouveau_oclass *nvf0_graph_oclass;
 
 extern const struct nouveau_bitfield nv04_graph_nsource[];
 extern struct nouveau_ofuncs nv04_graph_ofuncs;
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/mpeg.h b/drivers/gpu/drm/nouveau/core/include/engine/mpeg.h
index bbf0d4a5bbd7..1d1a89a06ee4 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/mpeg.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/mpeg.h
@@ -54,7 +54,6 @@ extern struct nouveau_ofuncs nv50_mpeg_ofuncs;
 int  nv50_mpeg_context_ctor(struct nouveau_object *, struct nouveau_object *,
 			    struct nouveau_oclass *, void *, u32,
 			    struct nouveau_object **);
-int  nv50_mpeg_tlb_flush(struct nouveau_engine *);
 void nv50_mpeg_intr(struct nouveau_subdev *);
 int  nv50_mpeg_init(struct nouveau_object *);
 
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/vp.h b/drivers/gpu/drm/nouveau/core/include/engine/vp.h
index d7b287b115bf..39baebec7fbb 100644
--- a/drivers/gpu/drm/nouveau/core/include/engine/vp.h
+++ b/drivers/gpu/drm/nouveau/core/include/engine/vp.h
@@ -2,6 +2,7 @@
 #define __NOUVEAU_VP_H__
 
 extern struct nouveau_oclass nv84_vp_oclass;
+extern struct nouveau_oclass nv98_vp_oclass;
 extern struct nouveau_oclass nvc0_vp_oclass;
 extern struct nouveau_oclass nve0_vp_oclass;
 
diff --git a/drivers/gpu/drm/nouveau/core/include/engine/xtensa.h b/drivers/gpu/drm/nouveau/core/include/engine/xtensa.h
new file mode 100644
index 000000000000..306100f31f02
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/include/engine/xtensa.h
@@ -0,0 +1,38 @@
+#ifndef __NOUVEAU_XTENSA_H__
+#define __NOUVEAU_XTENSA_H__
+
+#include <core/engine.h>
+#include <core/engctx.h>
+#include <core/gpuobj.h>
+
+struct nouveau_xtensa {
+	struct nouveau_engine base;
+
+	u32 addr;
+	struct nouveau_gpuobj *gpu_fw;
+	u32 fifo_val;
+	u32 unkd28;
+};
+
+#define nouveau_xtensa_create(p,e,c,b,d,i,f,r)				\
+	nouveau_xtensa_create_((p), (e), (c), (b), (d), (i), (f),	\
+			       sizeof(**r),(void **)r)
+
+int _nouveau_xtensa_engctx_ctor(struct nouveau_object *,
+				struct nouveau_object *,
+				struct nouveau_oclass *, void *, u32,
+				struct nouveau_object **);
+
+void _nouveau_xtensa_intr(struct nouveau_subdev *);
+int nouveau_xtensa_create_(struct nouveau_object *,
+			   struct nouveau_object *,
+			   struct nouveau_oclass *, u32, bool,
+			   const char *, const char *,
+			   int, void **);
+#define _nouveau_xtensa_dtor _nouveau_engine_dtor
+int _nouveau_xtensa_init(struct nouveau_object *);
+int _nouveau_xtensa_fini(struct nouveau_object *, bool);
+u32  _nouveau_xtensa_rd32(struct nouveau_object *, u64);
+void _nouveau_xtensa_wr32(struct nouveau_object *, u64, u32);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/clock.h b/drivers/gpu/drm/nouveau/core/include/subdev/clock.h
index 41b7a6a76f19..89ee289097a6 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/clock.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/clock.h
@@ -10,8 +10,6 @@ struct nvbios_pll;
 struct nouveau_clock {
 	struct nouveau_subdev base;
 
-	int (*pll_set)(struct nouveau_clock *, u32 type, u32 freq);
-
 	/*XXX: die, these are here *only* to support the completely
 	 *     bat-shit insane what-was-nouveau_hw.c code
 	 */
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/devinit.h b/drivers/gpu/drm/nouveau/core/include/subdev/devinit.h
index 29e4cc1f6cc0..685c9b12ee4c 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/devinit.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/devinit.h
@@ -8,6 +8,8 @@ struct nouveau_devinit {
 	struct nouveau_subdev base;
 	bool post;
 	void (*meminit)(struct nouveau_devinit *);
+	int  (*pll_set)(struct nouveau_devinit *, u32 type, u32 freq);
+
 };
 
 static inline struct nouveau_devinit *
@@ -20,11 +22,20 @@ nouveau_devinit(void *obj)
 	nouveau_devinit_create_((p), (e), (o), sizeof(**d), (void **)d)
 #define nouveau_devinit_destroy(p)                                             \
 	nouveau_subdev_destroy(&(p)->base)
+#define nouveau_devinit_init(p) ({                                             \
+	struct nouveau_devinit *d = (p);                                       \
+	_nouveau_devinit_init(nv_object(d));                                   \
+})
+#define nouveau_devinit_fini(p,s) ({                                           \
+	struct nouveau_devinit *d = (p);                                       \
+	_nouveau_devinit_fini(nv_object(d), (s));                              \
+})
 
 int nouveau_devinit_create_(struct nouveau_object *, struct nouveau_object *,
 			    struct nouveau_oclass *, int, void **);
-int nouveau_devinit_init(struct nouveau_devinit *);
-int nouveau_devinit_fini(struct nouveau_devinit *, bool suspend);
+#define _nouveau_devinit_dtor _nouveau_subdev_dtor
+int _nouveau_devinit_init(struct nouveau_object *);
+int _nouveau_devinit_fini(struct nouveau_object *, bool suspend);
 
 extern struct nouveau_oclass nv04_devinit_oclass;
 extern struct nouveau_oclass nv05_devinit_oclass;
@@ -32,9 +43,7 @@ extern struct nouveau_oclass nv10_devinit_oclass;
 extern struct nouveau_oclass nv1a_devinit_oclass;
 extern struct nouveau_oclass nv20_devinit_oclass;
 extern struct nouveau_oclass nv50_devinit_oclass;
-
-void nv04_devinit_dtor(struct nouveau_object *);
-int  nv04_devinit_init(struct nouveau_object *);
-int  nv04_devinit_fini(struct nouveau_object *, bool);
+extern struct nouveau_oclass nva3_devinit_oclass;
+extern struct nouveau_oclass nvc0_devinit_oclass;
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/fb.h b/drivers/gpu/drm/nouveau/core/include/subdev/fb.h
index da470e6851b1..2e7405084261 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/fb.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/fb.h
@@ -53,31 +53,7 @@ struct nouveau_fb {
 
 	bool (*memtype_valid)(struct nouveau_fb *, u32 memtype);
 
-	struct {
-		enum {
-			NV_MEM_TYPE_UNKNOWN = 0,
-			NV_MEM_TYPE_STOLEN,
-			NV_MEM_TYPE_SGRAM,
-			NV_MEM_TYPE_SDRAM,
-			NV_MEM_TYPE_DDR1,
-			NV_MEM_TYPE_DDR2,
-			NV_MEM_TYPE_DDR3,
-			NV_MEM_TYPE_GDDR2,
-			NV_MEM_TYPE_GDDR3,
-			NV_MEM_TYPE_GDDR4,
-			NV_MEM_TYPE_GDDR5
-		} type;
-		u64 stolen;
-		u64 size;
-
-		int ranks;
-		int parts;
-
-		int  (*init)(struct nouveau_fb *);
-		int  (*get)(struct nouveau_fb *, u64 size, u32 align,
-			    u32 size_nc, u32 type, struct nouveau_mem **);
-		void (*put)(struct nouveau_fb *, struct nouveau_mem **);
-	} ram;
+	struct nouveau_ram *ram;
 
 	struct nouveau_mm vram;
 	struct nouveau_mm tags;
@@ -102,18 +78,6 @@ nouveau_fb(void *obj)
 	return (void *)nv_device(obj)->subdev[NVDEV_SUBDEV_FB];
 }
 
-#define nouveau_fb_create(p,e,c,d)                                             \
-	nouveau_subdev_create((p), (e), (c), 0, "PFB", "fb", (d))
-int  nouveau_fb_preinit(struct nouveau_fb *);
-void nouveau_fb_destroy(struct nouveau_fb *);
-int  nouveau_fb_init(struct nouveau_fb *);
-#define nouveau_fb_fini(p,s)                                                   \
-	nouveau_subdev_fini(&(p)->base, (s))
-
-void _nouveau_fb_dtor(struct nouveau_object *);
-int  _nouveau_fb_init(struct nouveau_object *);
-#define _nouveau_fb_fini _nouveau_subdev_fini
-
 extern struct nouveau_oclass nv04_fb_oclass;
 extern struct nouveau_oclass nv10_fb_oclass;
 extern struct nouveau_oclass nv1a_fb_oclass;
@@ -132,40 +96,31 @@ extern struct nouveau_oclass nv4e_fb_oclass;
 extern struct nouveau_oclass nv50_fb_oclass;
 extern struct nouveau_oclass nvc0_fb_oclass;
 
-struct nouveau_bios;
-int  nouveau_fb_bios_memtype(struct nouveau_bios *);
-
-bool nv04_fb_memtype_valid(struct nouveau_fb *, u32 memtype);
-
-void nv10_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
-		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
-void nv10_fb_tile_fini(struct nouveau_fb *, int i, struct nouveau_fb_tile *);
-void nv10_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
-
-int  nv20_fb_vram_init(struct nouveau_fb *);
-void nv20_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
-		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
-void nv20_fb_tile_fini(struct nouveau_fb *, int i, struct nouveau_fb_tile *);
-void nv20_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
-
-int  nv30_fb_init(struct nouveau_object *);
-void nv30_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
-		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
-
-void nv40_fb_tile_comp(struct nouveau_fb *, int i, u32 size, u32 flags,
-		       struct nouveau_fb_tile *);
-
-int  nv41_fb_vram_init(struct nouveau_fb *);
-int  nv41_fb_init(struct nouveau_object *);
-void nv41_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
-
-int  nv44_fb_vram_init(struct nouveau_fb *);
-int  nv44_fb_init(struct nouveau_object *);
-void nv44_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
+struct nouveau_ram {
+	struct nouveau_object base;
+	enum {
+		NV_MEM_TYPE_UNKNOWN = 0,
+		NV_MEM_TYPE_STOLEN,
+		NV_MEM_TYPE_SGRAM,
+		NV_MEM_TYPE_SDRAM,
+		NV_MEM_TYPE_DDR1,
+		NV_MEM_TYPE_DDR2,
+		NV_MEM_TYPE_DDR3,
+		NV_MEM_TYPE_GDDR2,
+		NV_MEM_TYPE_GDDR3,
+		NV_MEM_TYPE_GDDR4,
+		NV_MEM_TYPE_GDDR5
+	} type;
+	u64 stolen;
+	u64 size;
+	u32 tags;
 
-void nv46_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
-		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
+	int ranks;
+	int parts;
 
-void nv50_fb_vram_del(struct nouveau_fb *, struct nouveau_mem **);
+	int  (*get)(struct nouveau_fb *, u64 size, u32 align,
+		    u32 size_nc, u32 type, struct nouveau_mem **);
+	void (*put)(struct nouveau_fb *, struct nouveau_mem **);
+};
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/include/subdev/vm.h b/drivers/gpu/drm/nouveau/core/include/subdev/vm.h
index 9d595efe667a..f2e87b105666 100644
--- a/drivers/gpu/drm/nouveau/core/include/subdev/vm.h
+++ b/drivers/gpu/drm/nouveau/core/include/subdev/vm.h
@@ -58,7 +58,7 @@ struct nouveau_vm {
 	int refcount;
 
 	struct list_head pgd_list;
-	atomic_t engref[64]; //NVDEV_SUBDEV_NR];
+	atomic_t engref[NVDEV_SUBDEV_NR];
 
 	struct nouveau_vm_pgt *pgt;
 	u32 fpde;
@@ -117,9 +117,6 @@ int  nv04_vm_create(struct nouveau_vmmgr *, u64, u64, u64,
 		    struct nouveau_vm **);
 void nv04_vmmgr_dtor(struct nouveau_object *);
 
-void nv50_vm_flush_engine(struct nouveau_subdev *, int engine);
-void nvc0_vm_flush_engine(struct nouveau_subdev *, u64 addr, int type);
-
 /* nouveau_vm.c */
 int  nouveau_vm_create(struct nouveau_vmmgr *, u64 offset, u64 length,
 		       u64 mm_offset, u32 block, struct nouveau_vm **);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/bar/nv50.c
index 649f1ced1fe0..160d27f3c7b4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bar/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/nv50.c
@@ -53,7 +53,6 @@ nv50_bar_kmap(struct nouveau_bar *bar, struct nouveau_mem *mem,
 		return ret;
 
 	nouveau_vm_map(vma, mem);
-	nv50_vm_flush_engine(nv_subdev(bar), 6);
 	return 0;
 }
 
@@ -69,7 +68,6 @@ nv50_bar_umap(struct nouveau_bar *bar, struct nouveau_mem *mem,
 		return ret;
 
 	nouveau_vm_map(vma, mem);
-	nv50_vm_flush_engine(nv_subdev(bar), 6);
 	return 0;
 }
 
@@ -77,7 +75,6 @@ static void
 nv50_bar_unmap(struct nouveau_bar *bar, struct nouveau_vma *vma)
 {
 	nouveau_vm_unmap(vma);
-	nv50_vm_flush_engine(nv_subdev(bar), 6);
 	nouveau_vm_put(vma);
 }
 
@@ -147,6 +144,8 @@ nv50_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	atomic_inc(&vm->engref[NVDEV_SUBDEV_BAR]);
+
 	ret = nouveau_gpuobj_new(nv_object(priv), heap,
 				 ((limit-- - start) >> 12) * 8, 0x1000,
 				 NVOBJ_FLAG_ZERO_ALLOC, &vm->pgt[0].obj[0]);
@@ -179,6 +178,8 @@ nv50_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	atomic_inc(&vm->engref[NVDEV_SUBDEV_BAR]);
+
 	ret = nouveau_vm_ref(vm, &priv->bar1_vm, priv->pgd);
 	nouveau_vm_ref(NULL, &vm, NULL);
 	if (ret)
@@ -237,7 +238,11 @@ nv50_bar_init(struct nouveau_object *object)
 
 	nv_mask(priv, 0x000200, 0x00000100, 0x00000000);
 	nv_mask(priv, 0x000200, 0x00000100, 0x00000100);
-	nv50_vm_flush_engine(nv_subdev(priv), 6);
+	nv_wr32(priv, 0x100c80, 0x00060001);
+	if (!nv_wait(priv, 0x100c80, 0x00000001, 0x00000000)) {
+		nv_error(priv, "vm flush timeout\n");
+		return -EBUSY;
+	}
 
 	nv_wr32(priv, 0x001704, 0x00000000 | priv->mem->addr >> 12);
 	nv_wr32(priv, 0x001704, 0x40000000 | priv->mem->addr >> 12);
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
index f8a44956dec1..b2ec7411eb2e 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bar/nvc0.c
@@ -51,7 +51,6 @@ nvc0_bar_kmap(struct nouveau_bar *bar, struct nouveau_mem *mem,
 		return ret;
 
 	nouveau_vm_map(vma, mem);
-	nvc0_vm_flush_engine(nv_subdev(bar), priv->bar[0].pgd->addr, 5);
 	return 0;
 }
 
@@ -68,18 +67,13 @@ nvc0_bar_umap(struct nouveau_bar *bar, struct nouveau_mem *mem,
 		return ret;
 
 	nouveau_vm_map(vma, mem);
-	nvc0_vm_flush_engine(nv_subdev(bar), priv->bar[1].pgd->addr, 5);
 	return 0;
 }
 
 static void
 nvc0_bar_unmap(struct nouveau_bar *bar, struct nouveau_vma *vma)
 {
-	struct nvc0_bar_priv *priv = (void *)bar;
-	int i = !(vma->vm == priv->bar[0].vm);
-
 	nouveau_vm_unmap(vma);
-	nvc0_vm_flush_engine(nv_subdev(bar), priv->bar[i].pgd->addr, 5);
 	nouveau_vm_put(vma);
 }
 
@@ -116,6 +110,8 @@ nvc0_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	atomic_inc(&vm->engref[NVDEV_SUBDEV_BAR]);
+
 	ret = nouveau_gpuobj_new(nv_object(priv), NULL,
 				 (pci_resource_len(pdev, 3) >> 12) * 8,
 				 0x1000, NVOBJ_FLAG_ZERO_ALLOC,
@@ -150,6 +146,8 @@ nvc0_bar_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	atomic_inc(&vm->engref[NVDEV_SUBDEV_BAR]);
+
 	ret = nouveau_vm_ref(vm, &priv->bar[1].vm, priv->bar[1].pgd);
 	nouveau_vm_ref(NULL, &vm, NULL);
 	if (ret)
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
index 0e2c1a4f1659..aa0fbbec7f08 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/base.c
@@ -85,11 +85,15 @@ static void
 nouveau_bios_shadow_pramin(struct nouveau_bios *bios)
 {
 	struct nouveau_device *device = nv_device(bios);
+	u64 addr = 0;
 	u32 bar0 = 0;
 	int i;
 
 	if (device->card_type >= NV_50) {
-		u64 addr = (u64)(nv_rd32(bios, 0x619f04) & 0xffffff00) << 8;
+		if (  device->card_type < NV_C0 ||
+		    !(nv_rd32(bios, 0x022500) & 0x00000001))
+			addr = (u64)(nv_rd32(bios, 0x619f04) & 0xffffff00) << 8;
+
 		if (!addr) {
 			addr  = (u64)nv_rd32(bios, 0x001700) << 16;
 			addr += 0xf0000;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
index c434d398d16f..0687e6481438 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/bios/init.c
@@ -10,7 +10,6 @@
 #include <subdev/bios/gpio.h>
 #include <subdev/bios/init.h>
 #include <subdev/devinit.h>
-#include <subdev/clock.h>
 #include <subdev/i2c.h>
 #include <subdev/vga.h>
 #include <subdev/gpio.h>
@@ -300,9 +299,9 @@ init_wrauxr(struct nvbios_init *init, u32 addr, u8 data)
 static void
 init_prog_pll(struct nvbios_init *init, u32 id, u32 freq)
 {
-	struct nouveau_clock *clk = nouveau_clock(init->bios);
-	if (clk && clk->pll_set && init_exec(init)) {
-		int ret = clk->pll_set(clk, id, freq);
+	struct nouveau_devinit *devinit = nouveau_devinit(init->bios);
+	if (devinit->pll_set && init_exec(init)) {
+		int ret = devinit->pll_set(devinit, id, freq);
 		if (ret)
 			warn("failed to prog pll 0x%08x to %dkHz\n", id, freq);
 	}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nv04.c
index b7fd1151166e..a14277586595 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nv04.c
@@ -22,9 +22,10 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/clock.h>
 #include <subdev/bios.h>
 #include <subdev/bios/pll.h>
+#include <subdev/clock.h>
+#include <subdev/devinit/priv.h>
 
 #include "pll.h"
 
@@ -32,272 +33,12 @@ struct nv04_clock_priv {
 	struct nouveau_clock base;
 };
 
-static int
-powerctrl_1_shift(int chip_version, int reg)
-{
-	int shift = -4;
-
-	if (chip_version < 0x17 || chip_version == 0x1a || chip_version == 0x20)
-		return shift;
-
-	switch (reg) {
-	case 0x680520:
-		shift += 4;
-	case 0x680508:
-		shift += 4;
-	case 0x680504:
-		shift += 4;
-	case 0x680500:
-		shift += 4;
-	}
-
-	/*
-	 * the shift for vpll regs is only used for nv3x chips with a single
-	 * stage pll
-	 */
-	if (shift > 4 && (chip_version < 0x32 || chip_version == 0x35 ||
-			  chip_version == 0x36 || chip_version >= 0x40))
-		shift = -4;
-
-	return shift;
-}
-
-static void
-setPLL_single(struct nv04_clock_priv *priv, u32 reg,
-	      struct nouveau_pll_vals *pv)
-{
-	int chip_version = nouveau_bios(priv)->version.chip;
-	uint32_t oldpll = nv_rd32(priv, reg);
-	int oldN = (oldpll >> 8) & 0xff, oldM = oldpll & 0xff;
-	uint32_t pll = (oldpll & 0xfff80000) | pv->log2P << 16 | pv->NM1;
-	uint32_t saved_powerctrl_1 = 0;
-	int shift_powerctrl_1 = powerctrl_1_shift(chip_version, reg);
-
-	if (oldpll == pll)
-		return;	/* already set */
-
-	if (shift_powerctrl_1 >= 0) {
-		saved_powerctrl_1 = nv_rd32(priv, 0x001584);
-		nv_wr32(priv, 0x001584,
-			(saved_powerctrl_1 & ~(0xf << shift_powerctrl_1)) |
-			1 << shift_powerctrl_1);
-	}
-
-	if (oldM && pv->M1 && (oldN / oldM < pv->N1 / pv->M1))
-		/* upclock -- write new post divider first */
-		nv_wr32(priv, reg, pv->log2P << 16 | (oldpll & 0xffff));
-	else
-		/* downclock -- write new NM first */
-		nv_wr32(priv, reg, (oldpll & 0xffff0000) | pv->NM1);
-
-	if (chip_version < 0x17 && chip_version != 0x11)
-		/* wait a bit on older chips */
-		msleep(64);
-	nv_rd32(priv, reg);
-
-	/* then write the other half as well */
-	nv_wr32(priv, reg, pll);
-
-	if (shift_powerctrl_1 >= 0)
-		nv_wr32(priv, 0x001584, saved_powerctrl_1);
-}
-
-static uint32_t
-new_ramdac580(uint32_t reg1, bool ss, uint32_t ramdac580)
-{
-	bool head_a = (reg1 == 0x680508);
-
-	if (ss)	/* single stage pll mode */
-		ramdac580 |= head_a ? 0x00000100 : 0x10000000;
-	else
-		ramdac580 &= head_a ? 0xfffffeff : 0xefffffff;
-
-	return ramdac580;
-}
-
-static void
-setPLL_double_highregs(struct nv04_clock_priv *priv, u32 reg1,
-		       struct nouveau_pll_vals *pv)
-{
-	int chip_version = nouveau_bios(priv)->version.chip;
-	bool nv3035 = chip_version == 0x30 || chip_version == 0x35;
-	uint32_t reg2 = reg1 + ((reg1 == 0x680520) ? 0x5c : 0x70);
-	uint32_t oldpll1 = nv_rd32(priv, reg1);
-	uint32_t oldpll2 = !nv3035 ? nv_rd32(priv, reg2) : 0;
-	uint32_t pll1 = (oldpll1 & 0xfff80000) | pv->log2P << 16 | pv->NM1;
-	uint32_t pll2 = (oldpll2 & 0x7fff0000) | 1 << 31 | pv->NM2;
-	uint32_t oldramdac580 = 0, ramdac580 = 0;
-	bool single_stage = !pv->NM2 || pv->N2 == pv->M2;	/* nv41+ only */
-	uint32_t saved_powerctrl_1 = 0, savedc040 = 0;
-	int shift_powerctrl_1 = powerctrl_1_shift(chip_version, reg1);
-
-	/* model specific additions to generic pll1 and pll2 set up above */
-	if (nv3035) {
-		pll1 = (pll1 & 0xfcc7ffff) | (pv->N2 & 0x18) << 21 |
-		       (pv->N2 & 0x7) << 19 | 8 << 4 | (pv->M2 & 7) << 4;
-		pll2 = 0;
-	}
-	if (chip_version > 0x40 && reg1 >= 0x680508) { /* !nv40 */
-		oldramdac580 = nv_rd32(priv, 0x680580);
-		ramdac580 = new_ramdac580(reg1, single_stage, oldramdac580);
-		if (oldramdac580 != ramdac580)
-			oldpll1 = ~0;	/* force mismatch */
-		if (single_stage)
-			/* magic value used by nvidia in single stage mode */
-			pll2 |= 0x011f;
-	}
-	if (chip_version > 0x70)
-		/* magic bits set by the blob (but not the bios) on g71-73 */
-		pll1 = (pll1 & 0x7fffffff) | (single_stage ? 0x4 : 0xc) << 28;
-
-	if (oldpll1 == pll1 && oldpll2 == pll2)
-		return;	/* already set */
-
-	if (shift_powerctrl_1 >= 0) {
-		saved_powerctrl_1 = nv_rd32(priv, 0x001584);
-		nv_wr32(priv, 0x001584,
-			(saved_powerctrl_1 & ~(0xf << shift_powerctrl_1)) |
-			1 << shift_powerctrl_1);
-	}
-
-	if (chip_version >= 0x40) {
-		int shift_c040 = 14;
-
-		switch (reg1) {
-		case 0x680504:
-			shift_c040 += 2;
-		case 0x680500:
-			shift_c040 += 2;
-		case 0x680520:
-			shift_c040 += 2;
-		case 0x680508:
-			shift_c040 += 2;
-		}
-
-		savedc040 = nv_rd32(priv, 0xc040);
-		if (shift_c040 != 14)
-			nv_wr32(priv, 0xc040, savedc040 & ~(3 << shift_c040));
-	}
-
-	if (oldramdac580 != ramdac580)
-		nv_wr32(priv, 0x680580, ramdac580);
-
-	if (!nv3035)
-		nv_wr32(priv, reg2, pll2);
-	nv_wr32(priv, reg1, pll1);
-
-	if (shift_powerctrl_1 >= 0)
-		nv_wr32(priv, 0x001584, saved_powerctrl_1);
-	if (chip_version >= 0x40)
-		nv_wr32(priv, 0xc040, savedc040);
-}
-
-static void
-setPLL_double_lowregs(struct nv04_clock_priv *priv, u32 NMNMreg,
-		      struct nouveau_pll_vals *pv)
-{
-	/* When setting PLLs, there is a merry game of disabling and enabling
-	 * various bits of hardware during the process. This function is a
-	 * synthesis of six nv4x traces, nearly each card doing a subtly
-	 * different thing. With luck all the necessary bits for each card are
-	 * combined herein. Without luck it deviates from each card's formula
-	 * so as to not work on any :)
-	 */
-
-	uint32_t Preg = NMNMreg - 4;
-	bool mpll = Preg == 0x4020;
-	uint32_t oldPval = nv_rd32(priv, Preg);
-	uint32_t NMNM = pv->NM2 << 16 | pv->NM1;
-	uint32_t Pval = (oldPval & (mpll ? ~(0x77 << 16) : ~(7 << 16))) |
-			0xc << 28 | pv->log2P << 16;
-	uint32_t saved4600 = 0;
-	/* some cards have different maskc040s */
-	uint32_t maskc040 = ~(3 << 14), savedc040;
-	bool single_stage = !pv->NM2 || pv->N2 == pv->M2;
-
-	if (nv_rd32(priv, NMNMreg) == NMNM && (oldPval & 0xc0070000) == Pval)
-		return;
-
-	if (Preg == 0x4000)
-		maskc040 = ~0x333;
-	if (Preg == 0x4058)
-		maskc040 = ~(0xc << 24);
-
-	if (mpll) {
-		struct nvbios_pll info;
-		uint8_t Pval2;
-
-		if (nvbios_pll_parse(nouveau_bios(priv), Preg, &info))
-			return;
-
-		Pval2 = pv->log2P + info.bias_p;
-		if (Pval2 > info.max_p)
-			Pval2 = info.max_p;
-		Pval |= 1 << 28 | Pval2 << 20;
-
-		saved4600 = nv_rd32(priv, 0x4600);
-		nv_wr32(priv, 0x4600, saved4600 | 8 << 28);
-	}
-	if (single_stage)
-		Pval |= mpll ? 1 << 12 : 1 << 8;
-
-	nv_wr32(priv, Preg, oldPval | 1 << 28);
-	nv_wr32(priv, Preg, Pval & ~(4 << 28));
-	if (mpll) {
-		Pval |= 8 << 20;
-		nv_wr32(priv, 0x4020, Pval & ~(0xc << 28));
-		nv_wr32(priv, 0x4038, Pval & ~(0xc << 28));
-	}
-
-	savedc040 = nv_rd32(priv, 0xc040);
-	nv_wr32(priv, 0xc040, savedc040 & maskc040);
-
-	nv_wr32(priv, NMNMreg, NMNM);
-	if (NMNMreg == 0x4024)
-		nv_wr32(priv, 0x403c, NMNM);
-
-	nv_wr32(priv, Preg, Pval);
-	if (mpll) {
-		Pval &= ~(8 << 20);
-		nv_wr32(priv, 0x4020, Pval);
-		nv_wr32(priv, 0x4038, Pval);
-		nv_wr32(priv, 0x4600, saved4600);
-	}
-
-	nv_wr32(priv, 0xc040, savedc040);
-
-	if (mpll) {
-		nv_wr32(priv, 0x4020, Pval & ~(1 << 28));
-		nv_wr32(priv, 0x4038, Pval & ~(1 << 28));
-	}
-}
-
-int
-nv04_clock_pll_set(struct nouveau_clock *clk, u32 type, u32 freq)
-{
-	struct nv04_clock_priv *priv = (void *)clk;
-	struct nouveau_pll_vals pv;
-	struct nvbios_pll info;
-	int ret;
-
-	ret = nvbios_pll_parse(nouveau_bios(priv), type > 0x405c ?
-			       type : type - 4, &info);
-	if (ret)
-		return ret;
-
-	ret = clk->pll_calc(clk, &info, freq, &pv);
-	if (!ret)
-		return ret;
-
-	return clk->pll_prog(clk, type, &pv);
-}
-
 int
 nv04_clock_pll_calc(struct nouveau_clock *clock, struct nvbios_pll *info,
 		    int clk, struct nouveau_pll_vals *pv)
 {
 	int N1, M1, N2, M2, P;
-	int ret = nv04_pll_calc(clock, info, clk, &N1, &M1, &N2, &M2, &P);
+	int ret = nv04_pll_calc(nv_subdev(clock), info, clk, &N1, &M1, &N2, &M2, &P);
 	if (ret) {
 		pv->refclk = info->refclk;
 		pv->N1 = N1;
@@ -313,17 +54,17 @@ int
 nv04_clock_pll_prog(struct nouveau_clock *clk, u32 reg1,
 		    struct nouveau_pll_vals *pv)
 {
-	struct nv04_clock_priv *priv = (void *)clk;
+	struct nouveau_devinit *devinit = nouveau_devinit(clk);
 	int cv = nouveau_bios(clk)->version.chip;
 
 	if (cv == 0x30 || cv == 0x31 || cv == 0x35 || cv == 0x36 ||
 	    cv >= 0x40) {
 		if (reg1 > 0x405c)
-			setPLL_double_highregs(priv, reg1, pv);
+			setPLL_double_highregs(devinit, reg1, pv);
 		else
-			setPLL_double_lowregs(priv, reg1, pv);
+			setPLL_double_lowregs(devinit, reg1, pv);
 	} else
-		setPLL_single(priv, reg1, pv);
+		setPLL_single(devinit, reg1, pv);
 
 	return 0;
 }
@@ -341,7 +82,6 @@ nv04_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.pll_set = nv04_clock_pll_set;
 	priv->base.pll_calc = nv04_clock_pll_calc;
 	priv->base.pll_prog = nv04_clock_pll_prog;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nv40.c
index a4b2b7ebf9af..0db5dbfd91b5 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nv40.c
@@ -41,7 +41,6 @@ nv40_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.pll_set = nv04_clock_pll_set;
 	priv->base.pll_calc = nv04_clock_pll_calc;
 	priv->base.pll_prog = nv04_clock_pll_prog;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nv50.c
index f4147f67eda6..d09d3e78040c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nv50.c
@@ -33,50 +33,6 @@ struct nv50_clock_priv {
 };
 
 static int
-nv50_clock_pll_set(struct nouveau_clock *clk, u32 type, u32 freq)
-{
-	struct nv50_clock_priv *priv = (void *)clk;
-	struct nouveau_bios *bios = nouveau_bios(priv);
-	struct nvbios_pll info;
-	int N1, M1, N2, M2, P;
-	int ret;
-
-	ret = nvbios_pll_parse(bios, type, &info);
-	if (ret) {
-		nv_error(clk, "failed to retrieve pll data, %d\n", ret);
-		return ret;
-	}
-
-	ret = nv04_pll_calc(clk, &info, freq, &N1, &M1, &N2, &M2, &P);
-	if (!ret) {
-		nv_error(clk, "failed pll calculation\n");
-		return ret;
-	}
-
-	switch (info.type) {
-	case PLL_VPLL0:
-	case PLL_VPLL1:
-		nv_wr32(priv, info.reg + 0, 0x10000611);
-		nv_mask(priv, info.reg + 4, 0x00ff00ff, (M1 << 16) | N1);
-		nv_mask(priv, info.reg + 8, 0x7fff00ff, (P  << 28) |
-							(M2 << 16) | N2);
-		break;
-	case PLL_MEMORY:
-		nv_mask(priv, info.reg + 0, 0x01ff0000, (P << 22) |
-						        (info.bias_p << 19) |
-							(P << 16));
-		nv_wr32(priv, info.reg + 4, (N1 << 8) | M1);
-		break;
-	default:
-		nv_mask(priv, info.reg + 0, 0x00070000, (P << 16));
-		nv_wr32(priv, info.reg + 4, (N1 << 8) | M1);
-		break;
-	}
-
-	return 0;
-}
-
-static int
 nv50_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		struct nouveau_oclass *oclass, void *data, u32 size,
 		struct nouveau_object **pobject)
@@ -89,7 +45,6 @@ nv50_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.pll_set = nv50_clock_pll_set;
 	priv->base.pll_calc = nv04_clock_pll_calc;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c
index 9068c98b96f6..f074cd20bc9c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nva3.c
@@ -32,47 +32,13 @@ struct nva3_clock_priv {
 	struct nouveau_clock base;
 };
 
-static int
-nva3_clock_pll_set(struct nouveau_clock *clk, u32 type, u32 freq)
-{
-	struct nva3_clock_priv *priv = (void *)clk;
-	struct nouveau_bios *bios = nouveau_bios(priv);
-	struct nvbios_pll info;
-	int N, fN, M, P;
-	int ret;
-
-	ret = nvbios_pll_parse(bios, type, &info);
-	if (ret)
-		return ret;
-
-	ret = nva3_pll_calc(clk, &info, freq, &N, &fN, &M, &P);
-	if (ret < 0)
-		return ret;
-
-	switch (info.type) {
-	case PLL_VPLL0:
-	case PLL_VPLL1:
-		nv_wr32(priv, info.reg + 0, 0x50000610);
-		nv_mask(priv, info.reg + 4, 0x003fffff,
-					    (P << 16) | (M << 8) | N);
-		nv_wr32(priv, info.reg + 8, fN);
-		break;
-	default:
-		nv_warn(priv, "0x%08x/%dKhz unimplemented\n", type, freq);
-		ret = -EINVAL;
-		break;
-	}
-
-	return ret;
-}
-
 int
 nva3_clock_pll_calc(struct nouveau_clock *clock, struct nvbios_pll *info,
 		    int clk, struct nouveau_pll_vals *pv)
 {
 	int ret, N, M, P;
 
-	ret = nva3_pll_calc(clock, info, clk, &N, NULL, &M, &P);
+	ret = nva3_pll_calc(nv_subdev(clock), info, clk, &N, NULL, &M, &P);
 
 	if (ret > 0) {
 		pv->refclk = info->refclk;
@@ -97,7 +63,6 @@ nva3_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.pll_set = nva3_clock_pll_set;
 	priv->base.pll_calc = nva3_clock_pll_calc;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/clock/nvc0.c
index 7c9626258a46..439d81c26130 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/nvc0.c
@@ -33,41 +33,6 @@ struct nvc0_clock_priv {
 };
 
 static int
-nvc0_clock_pll_set(struct nouveau_clock *clk, u32 type, u32 freq)
-{
-	struct nvc0_clock_priv *priv = (void *)clk;
-	struct nouveau_bios *bios = nouveau_bios(priv);
-	struct nvbios_pll info;
-	int N, fN, M, P;
-	int ret;
-
-	ret = nvbios_pll_parse(bios, type, &info);
-	if (ret)
-		return ret;
-
-	ret = nva3_pll_calc(clk, &info, freq, &N, &fN, &M, &P);
-	if (ret < 0)
-		return ret;
-
-	switch (info.type) {
-	case PLL_VPLL0:
-	case PLL_VPLL1:
-	case PLL_VPLL2:
-	case PLL_VPLL3:
-		nv_mask(priv, info.reg + 0x0c, 0x00000000, 0x00000100);
-		nv_wr32(priv, info.reg + 0x04, (P << 16) | (N << 8) | M);
-		nv_wr32(priv, info.reg + 0x10, fN << 16);
-		break;
-	default:
-		nv_warn(priv, "0x%08x/%dKhz unimplemented\n", type, freq);
-		ret = -EINVAL;
-		break;
-	}
-
-	return ret;
-}
-
-static int
 nvc0_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		struct nouveau_oclass *oclass, void *data, u32 size,
 		struct nouveau_object **pobject)
@@ -80,7 +45,6 @@ nvc0_clock_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	priv->base.pll_set = nvc0_clock_pll_set;
 	priv->base.pll_calc = nva3_clock_pll_calc;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/pll.h b/drivers/gpu/drm/nouveau/core/subdev/clock/pll.h
index ef2c0078f337..445b14c33a98 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/pll.h
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/pll.h
@@ -1,9 +1,9 @@
 #ifndef __NOUVEAU_PLL_H__
 #define __NOUVEAU_PLL_H__
 
-int nv04_pll_calc(struct nouveau_clock *, struct nvbios_pll *, u32 freq,
+int nv04_pll_calc(struct nouveau_subdev *, struct nvbios_pll *, u32 freq,
 		  int *N1, int *M1, int *N2, int *M2, int *P);
-int nva3_pll_calc(struct nouveau_clock *, struct nvbios_pll *, u32 freq,
+int nva3_pll_calc(struct nouveau_subdev *, struct nvbios_pll *, u32 freq,
 		  int *N, int *fN, int *M, int *P);
 
 #endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/pllnv04.c b/drivers/gpu/drm/nouveau/core/subdev/clock/pllnv04.c
index a2ab6d051ba8..cf1ed0dc9bc9 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/pllnv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/pllnv04.c
@@ -21,14 +21,13 @@
  * SOFTWARE.
  */
 
-#include <subdev/clock.h>
 #include <subdev/bios.h>
 #include <subdev/bios/pll.h>
 
 #include "pll.h"
 
 static int
-getMNP_single(struct nouveau_clock *clock, struct nvbios_pll *info, int clk,
+getMNP_single(struct nouveau_subdev *subdev, struct nvbios_pll *info, int clk,
 	      int *pN, int *pM, int *pP)
 {
 	/* Find M, N and P for a single stage PLL
@@ -39,7 +38,7 @@ getMNP_single(struct nouveau_clock *clock, struct nvbios_pll *info, int clk,
 	 * "clk" parameter in kHz
 	 * returns calculated clock
 	 */
-	int cv = nouveau_bios(clock)->version.chip;
+	int cv = nouveau_bios(subdev)->version.chip;
 	int minvco = info->vco1.min_freq, maxvco = info->vco1.max_freq;
 	int minM = info->vco1.min_m, maxM = info->vco1.max_m;
 	int minN = info->vco1.min_n, maxN = info->vco1.max_n;
@@ -124,7 +123,7 @@ getMNP_single(struct nouveau_clock *clock, struct nvbios_pll *info, int clk,
 }
 
 static int
-getMNP_double(struct nouveau_clock *clock, struct nvbios_pll *info, int clk,
+getMNP_double(struct nouveau_subdev *subdev, struct nvbios_pll *info, int clk,
 	      int *pN1, int *pM1, int *pN2, int *pM2, int *pP)
 {
 	/* Find M, N and P for a two stage PLL
@@ -135,7 +134,7 @@ getMNP_double(struct nouveau_clock *clock, struct nvbios_pll *info, int clk,
 	 * "clk" parameter in kHz
 	 * returns calculated clock
 	 */
-	int chip_version = nouveau_bios(clock)->version.chip;
+	int chip_version = nouveau_bios(subdev)->version.chip;
 	int minvco1 = info->vco1.min_freq, maxvco1 = info->vco1.max_freq;
 	int minvco2 = info->vco2.min_freq, maxvco2 = info->vco2.max_freq;
 	int minU1 = info->vco1.min_inputfreq, minU2 = info->vco2.min_inputfreq;
@@ -223,20 +222,20 @@ getMNP_double(struct nouveau_clock *clock, struct nvbios_pll *info, int clk,
 }
 
 int
-nv04_pll_calc(struct nouveau_clock *clk, struct nvbios_pll *info, u32 freq,
+nv04_pll_calc(struct nouveau_subdev *subdev, struct nvbios_pll *info, u32 freq,
 	      int *N1, int *M1, int *N2, int *M2, int *P)
 {
 	int ret;
 
 	if (!info->vco2.max_freq) {
-		ret = getMNP_single(clk, info, freq, N1, M1, P);
+		ret = getMNP_single(subdev, info, freq, N1, M1, P);
 		*N2 = 1;
 		*M2 = 1;
 	} else {
-		ret = getMNP_double(clk, info, freq, N1, M1, N2, M2, P);
+		ret = getMNP_double(subdev, info, freq, N1, M1, N2, M2, P);
 	}
 
 	if (!ret)
-		nv_error(clk, "unable to compute acceptable pll values\n");
+		nv_error(subdev, "unable to compute acceptable pll values\n");
 	return ret;
 }
diff --git a/drivers/gpu/drm/nouveau/core/subdev/clock/pllnva3.c b/drivers/gpu/drm/nouveau/core/subdev/clock/pllnva3.c
index eed5c16cf610..2fe1f712eefa 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/clock/pllnva3.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/clock/pllnva3.c
@@ -29,7 +29,7 @@
 #include "pll.h"
 
 int
-nva3_pll_calc(struct nouveau_clock *clock, struct nvbios_pll *info,
+nva3_pll_calc(struct nouveau_subdev *subdev, struct nvbios_pll *info,
 	      u32 freq, int *pN, int *pfN, int *pM, int *P)
 {
 	u32 best_err = ~0, err;
@@ -50,8 +50,15 @@ nva3_pll_calc(struct nouveau_clock *clock, struct nvbios_pll *info,
 		u32 tmp = freq * *P * M;
 		N  = tmp / info->refclk;
 		fN = tmp % info->refclk;
-		if (!pfN && fN >= info->refclk / 2)
-			N++;
+
+		if (!pfN) {
+			if (fN >= info->refclk / 2)
+				N++;
+		} else {
+			if (fN <  info->refclk / 2)
+				N--;
+			fN = tmp - (N * info->refclk);
+		}
 
 		if (N < info->vco1.min_n)
 			continue;
@@ -66,13 +73,14 @@ nva3_pll_calc(struct nouveau_clock *clock, struct nvbios_pll *info,
 		}
 
 		if (pfN) {
-			*pfN = (((fN << 13) / info->refclk) - 4096) & 0xffff;
+			*pfN = ((fN << 13) + info->refclk / 2) / info->refclk;
+			*pfN = (*pfN - 4096) & 0xffff;
 			return freq;
 		}
 	}
 
 	if (unlikely(best_err == ~0)) {
-		nv_error(clock, "unable to find matching pll values\n");
+		nv_error(subdev, "unable to find matching pll values\n");
 		return -EINVAL;
 	}
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/base.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/base.c
index 5a07a39c1735..79c81d3d9bac 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/base.c
@@ -29,18 +29,10 @@
 #include <subdev/bios/init.h>
 
 int
-nouveau_devinit_init(struct nouveau_devinit *devinit)
+_nouveau_devinit_fini(struct nouveau_object *object, bool suspend)
 {
-	int ret = nouveau_subdev_init(&devinit->base);
-	if (ret)
-		return ret;
+	struct nouveau_devinit *devinit = (void *)object;
 
-	return nvbios_init(&devinit->base, devinit->post);
-}
-
-int
-nouveau_devinit_fini(struct nouveau_devinit *devinit, bool suspend)
-{
 	/* force full reinit on resume */
 	if (suspend)
 		devinit->post = true;
@@ -49,6 +41,17 @@ nouveau_devinit_fini(struct nouveau_devinit *devinit, bool suspend)
 }
 
 int
+_nouveau_devinit_init(struct nouveau_object *object)
+{
+	struct nouveau_devinit *devinit = (void *)object;
+	int ret = nouveau_subdev_init(&devinit->base);
+	if (ret)
+		return ret;
+
+	return nvbios_init(&devinit->base, devinit->post);
+}
+
+int
 nouveau_devinit_create_(struct nouveau_object *parent,
 			struct nouveau_object *engine,
 			struct nouveau_oclass *oclass,
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv04.c
index 7a72d9394340..b22357d9b821 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv04.c
@@ -24,10 +24,10 @@
  *
  */
 
-#include <subdev/devinit.h>
 #include <subdev/vga.h>
 
 #include "fbmem.h"
+#include "priv.h"
 
 struct nv04_devinit_priv {
 	struct nouveau_devinit base;
@@ -111,33 +111,298 @@ nv04_devinit_meminit(struct nouveau_devinit *devinit)
 }
 
 static int
-nv04_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-		  struct nouveau_oclass *oclass, void *data, u32 size,
-		  struct nouveau_object **pobject)
+powerctrl_1_shift(int chip_version, int reg)
 {
-	struct nv04_devinit_priv *priv;
+	int shift = -4;
+
+	if (chip_version < 0x17 || chip_version == 0x1a || chip_version == 0x20)
+		return shift;
+
+	switch (reg) {
+	case 0x680520:
+		shift += 4;
+	case 0x680508:
+		shift += 4;
+	case 0x680504:
+		shift += 4;
+	case 0x680500:
+		shift += 4;
+	}
+
+	/*
+	 * the shift for vpll regs is only used for nv3x chips with a single
+	 * stage pll
+	 */
+	if (shift > 4 && (chip_version < 0x32 || chip_version == 0x35 ||
+			  chip_version == 0x36 || chip_version >= 0x40))
+		shift = -4;
+
+	return shift;
+}
+
+void
+setPLL_single(struct nouveau_devinit *devinit, u32 reg,
+	      struct nouveau_pll_vals *pv)
+{
+	int chip_version = nouveau_bios(devinit)->version.chip;
+	uint32_t oldpll = nv_rd32(devinit, reg);
+	int oldN = (oldpll >> 8) & 0xff, oldM = oldpll & 0xff;
+	uint32_t pll = (oldpll & 0xfff80000) | pv->log2P << 16 | pv->NM1;
+	uint32_t saved_powerctrl_1 = 0;
+	int shift_powerctrl_1 = powerctrl_1_shift(chip_version, reg);
+
+	if (oldpll == pll)
+		return;	/* already set */
+
+	if (shift_powerctrl_1 >= 0) {
+		saved_powerctrl_1 = nv_rd32(devinit, 0x001584);
+		nv_wr32(devinit, 0x001584,
+			(saved_powerctrl_1 & ~(0xf << shift_powerctrl_1)) |
+			1 << shift_powerctrl_1);
+	}
+
+	if (oldM && pv->M1 && (oldN / oldM < pv->N1 / pv->M1))
+		/* upclock -- write new post divider first */
+		nv_wr32(devinit, reg, pv->log2P << 16 | (oldpll & 0xffff));
+	else
+		/* downclock -- write new NM first */
+		nv_wr32(devinit, reg, (oldpll & 0xffff0000) | pv->NM1);
+
+	if (chip_version < 0x17 && chip_version != 0x11)
+		/* wait a bit on older chips */
+		msleep(64);
+	nv_rd32(devinit, reg);
+
+	/* then write the other half as well */
+	nv_wr32(devinit, reg, pll);
+
+	if (shift_powerctrl_1 >= 0)
+		nv_wr32(devinit, 0x001584, saved_powerctrl_1);
+}
+
+static uint32_t
+new_ramdac580(uint32_t reg1, bool ss, uint32_t ramdac580)
+{
+	bool head_a = (reg1 == 0x680508);
+
+	if (ss)	/* single stage pll mode */
+		ramdac580 |= head_a ? 0x00000100 : 0x10000000;
+	else
+		ramdac580 &= head_a ? 0xfffffeff : 0xefffffff;
+
+	return ramdac580;
+}
+
+void
+setPLL_double_highregs(struct nouveau_devinit *devinit, u32 reg1,
+		       struct nouveau_pll_vals *pv)
+{
+	int chip_version = nouveau_bios(devinit)->version.chip;
+	bool nv3035 = chip_version == 0x30 || chip_version == 0x35;
+	uint32_t reg2 = reg1 + ((reg1 == 0x680520) ? 0x5c : 0x70);
+	uint32_t oldpll1 = nv_rd32(devinit, reg1);
+	uint32_t oldpll2 = !nv3035 ? nv_rd32(devinit, reg2) : 0;
+	uint32_t pll1 = (oldpll1 & 0xfff80000) | pv->log2P << 16 | pv->NM1;
+	uint32_t pll2 = (oldpll2 & 0x7fff0000) | 1 << 31 | pv->NM2;
+	uint32_t oldramdac580 = 0, ramdac580 = 0;
+	bool single_stage = !pv->NM2 || pv->N2 == pv->M2;	/* nv41+ only */
+	uint32_t saved_powerctrl_1 = 0, savedc040 = 0;
+	int shift_powerctrl_1 = powerctrl_1_shift(chip_version, reg1);
+
+	/* model specific additions to generic pll1 and pll2 set up above */
+	if (nv3035) {
+		pll1 = (pll1 & 0xfcc7ffff) | (pv->N2 & 0x18) << 21 |
+		       (pv->N2 & 0x7) << 19 | 8 << 4 | (pv->M2 & 7) << 4;
+		pll2 = 0;
+	}
+	if (chip_version > 0x40 && reg1 >= 0x680508) { /* !nv40 */
+		oldramdac580 = nv_rd32(devinit, 0x680580);
+		ramdac580 = new_ramdac580(reg1, single_stage, oldramdac580);
+		if (oldramdac580 != ramdac580)
+			oldpll1 = ~0;	/* force mismatch */
+		if (single_stage)
+			/* magic value used by nvidia in single stage mode */
+			pll2 |= 0x011f;
+	}
+	if (chip_version > 0x70)
+		/* magic bits set by the blob (but not the bios) on g71-73 */
+		pll1 = (pll1 & 0x7fffffff) | (single_stage ? 0x4 : 0xc) << 28;
+
+	if (oldpll1 == pll1 && oldpll2 == pll2)
+		return;	/* already set */
+
+	if (shift_powerctrl_1 >= 0) {
+		saved_powerctrl_1 = nv_rd32(devinit, 0x001584);
+		nv_wr32(devinit, 0x001584,
+			(saved_powerctrl_1 & ~(0xf << shift_powerctrl_1)) |
+			1 << shift_powerctrl_1);
+	}
+
+	if (chip_version >= 0x40) {
+		int shift_c040 = 14;
+
+		switch (reg1) {
+		case 0x680504:
+			shift_c040 += 2;
+		case 0x680500:
+			shift_c040 += 2;
+		case 0x680520:
+			shift_c040 += 2;
+		case 0x680508:
+			shift_c040 += 2;
+		}
+
+		savedc040 = nv_rd32(devinit, 0xc040);
+		if (shift_c040 != 14)
+			nv_wr32(devinit, 0xc040, savedc040 & ~(3 << shift_c040));
+	}
+
+	if (oldramdac580 != ramdac580)
+		nv_wr32(devinit, 0x680580, ramdac580);
+
+	if (!nv3035)
+		nv_wr32(devinit, reg2, pll2);
+	nv_wr32(devinit, reg1, pll1);
+
+	if (shift_powerctrl_1 >= 0)
+		nv_wr32(devinit, 0x001584, saved_powerctrl_1);
+	if (chip_version >= 0x40)
+		nv_wr32(devinit, 0xc040, savedc040);
+}
+
+void
+setPLL_double_lowregs(struct nouveau_devinit *devinit, u32 NMNMreg,
+		      struct nouveau_pll_vals *pv)
+{
+	/* When setting PLLs, there is a merry game of disabling and enabling
+	 * various bits of hardware during the process. This function is a
+	 * synthesis of six nv4x traces, nearly each card doing a subtly
+	 * different thing. With luck all the necessary bits for each card are
+	 * combined herein. Without luck it deviates from each card's formula
+	 * so as to not work on any :)
+	 */
+
+	uint32_t Preg = NMNMreg - 4;
+	bool mpll = Preg == 0x4020;
+	uint32_t oldPval = nv_rd32(devinit, Preg);
+	uint32_t NMNM = pv->NM2 << 16 | pv->NM1;
+	uint32_t Pval = (oldPval & (mpll ? ~(0x77 << 16) : ~(7 << 16))) |
+			0xc << 28 | pv->log2P << 16;
+	uint32_t saved4600 = 0;
+	/* some cards have different maskc040s */
+	uint32_t maskc040 = ~(3 << 14), savedc040;
+	bool single_stage = !pv->NM2 || pv->N2 == pv->M2;
+
+	if (nv_rd32(devinit, NMNMreg) == NMNM && (oldPval & 0xc0070000) == Pval)
+		return;
+
+	if (Preg == 0x4000)
+		maskc040 = ~0x333;
+	if (Preg == 0x4058)
+		maskc040 = ~(0xc << 24);
+
+	if (mpll) {
+		struct nvbios_pll info;
+		uint8_t Pval2;
+
+		if (nvbios_pll_parse(nouveau_bios(devinit), Preg, &info))
+			return;
+
+		Pval2 = pv->log2P + info.bias_p;
+		if (Pval2 > info.max_p)
+			Pval2 = info.max_p;
+		Pval |= 1 << 28 | Pval2 << 20;
+
+		saved4600 = nv_rd32(devinit, 0x4600);
+		nv_wr32(devinit, 0x4600, saved4600 | 8 << 28);
+	}
+	if (single_stage)
+		Pval |= mpll ? 1 << 12 : 1 << 8;
+
+	nv_wr32(devinit, Preg, oldPval | 1 << 28);
+	nv_wr32(devinit, Preg, Pval & ~(4 << 28));
+	if (mpll) {
+		Pval |= 8 << 20;
+		nv_wr32(devinit, 0x4020, Pval & ~(0xc << 28));
+		nv_wr32(devinit, 0x4038, Pval & ~(0xc << 28));
+	}
+
+	savedc040 = nv_rd32(devinit, 0xc040);
+	nv_wr32(devinit, 0xc040, savedc040 & maskc040);
+
+	nv_wr32(devinit, NMNMreg, NMNM);
+	if (NMNMreg == 0x4024)
+		nv_wr32(devinit, 0x403c, NMNM);
+
+	nv_wr32(devinit, Preg, Pval);
+	if (mpll) {
+		Pval &= ~(8 << 20);
+		nv_wr32(devinit, 0x4020, Pval);
+		nv_wr32(devinit, 0x4038, Pval);
+		nv_wr32(devinit, 0x4600, saved4600);
+	}
+
+	nv_wr32(devinit, 0xc040, savedc040);
+
+	if (mpll) {
+		nv_wr32(devinit, 0x4020, Pval & ~(1 << 28));
+		nv_wr32(devinit, 0x4038, Pval & ~(1 << 28));
+	}
+}
+
+int
+nv04_devinit_pll_set(struct nouveau_devinit *devinit, u32 type, u32 freq)
+{
+	struct nouveau_bios *bios = nouveau_bios(devinit);
+	struct nouveau_pll_vals pv;
+	struct nvbios_pll info;
+	int cv = bios->version.chip;
+	int N1, M1, N2, M2, P;
 	int ret;
 
-	ret = nouveau_devinit_create(parent, engine, oclass, &priv);
-	*pobject = nv_object(priv);
+	ret = nvbios_pll_parse(bios, type > 0x405c ? type : type - 4, &info);
 	if (ret)
 		return ret;
 
-	priv->base.meminit = nv04_devinit_meminit;
-	priv->owner = -1;
+	ret = nv04_pll_calc(nv_subdev(devinit), &info, freq,
+			   &N1, &M1, &N2, &M2, &P);
+	if (!ret)
+		return -EINVAL;
+
+	pv.refclk = info.refclk;
+	pv.N1 = N1;
+	pv.M1 = M1;
+	pv.N2 = N2;
+	pv.M2 = M2;
+	pv.log2P = P;
+
+	if (cv == 0x30 || cv == 0x31 || cv == 0x35 || cv == 0x36 ||
+	    cv >= 0x40) {
+		if (type > 0x405c)
+			setPLL_double_highregs(devinit, type, &pv);
+		else
+			setPLL_double_lowregs(devinit, type, &pv);
+	} else
+		setPLL_single(devinit, type, &pv);
+
 	return 0;
 }
 
-void
-nv04_devinit_dtor(struct nouveau_object *object)
+int
+nv04_devinit_fini(struct nouveau_object *object, bool suspend)
 {
 	struct nv04_devinit_priv *priv = (void *)object;
 
-	/* restore vga owner saved at first init, and lock crtc regs  */
-	nv_wrvgaowner(priv, priv->owner);
-	nv_lockvgac(priv, true);
+	/* make i2c busses accessible */
+	nv_mask(priv, 0x000200, 0x00000001, 0x00000001);
 
-	nouveau_devinit_destroy(&priv->base);
+	/* unlock extended vga crtc regs, and unslave crtcs */
+	nv_lockvgac(priv, false);
+	if (priv->owner < 0)
+		priv->owner = nv_rdvgaowner(priv);
+	nv_wrvgaowner(priv, 0);
+
+	return nouveau_devinit_fini(&priv->base, suspend);
 }
 
 int
@@ -160,21 +425,35 @@ nv04_devinit_init(struct nouveau_object *object)
 	return nouveau_devinit_init(&priv->base);
 }
 
-int
-nv04_devinit_fini(struct nouveau_object *object, bool suspend)
+void
+nv04_devinit_dtor(struct nouveau_object *object)
 {
 	struct nv04_devinit_priv *priv = (void *)object;
 
-	/* make i2c busses accessible */
-	nv_mask(priv, 0x000200, 0x00000001, 0x00000001);
+	/* restore vga owner saved at first init, and lock crtc regs  */
+	nv_wrvgaowner(priv, priv->owner);
+	nv_lockvgac(priv, true);
 
-	/* unlock extended vga crtc regs, and unslave crtcs */
-	nv_lockvgac(priv, false);
-	if (priv->owner < 0)
-		priv->owner = nv_rdvgaowner(priv);
-	nv_wrvgaowner(priv, 0);
+	nouveau_devinit_destroy(&priv->base);
+}
 
-	return nouveau_devinit_fini(&priv->base, suspend);
+static int
+nv04_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		  struct nouveau_oclass *oclass, void *data, u32 size,
+		  struct nouveau_object **pobject)
+{
+	struct nv04_devinit_priv *priv;
+	int ret;
+
+	ret = nouveau_devinit_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.meminit = nv04_devinit_meminit;
+	priv->base.pll_set = nv04_devinit_pll_set;
+	priv->owner = -1;
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv05.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv05.c
index 191447d0d252..b1912a8a8942 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv05.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv05.c
@@ -24,12 +24,12 @@
  *
  */
 
-#include <subdev/devinit.h>
 #include <subdev/bios.h>
 #include <subdev/bios/bmp.h>
 #include <subdev/vga.h>
 
 #include "fbmem.h"
+#include "priv.h"
 
 struct nv05_devinit_priv {
 	struct nouveau_devinit base;
@@ -144,6 +144,7 @@ nv05_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	priv->base.meminit = nv05_devinit_meminit;
+	priv->base.pll_set = nv04_devinit_pll_set;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv10.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv10.c
index eb76ffab6b0c..463b08fa0968 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv10.c
@@ -24,10 +24,10 @@
  *
  */
 
-#include <subdev/devinit.h>
 #include <subdev/vga.h>
 
 #include "fbmem.h"
+#include "priv.h"
 
 struct nv10_devinit_priv {
 	struct nouveau_devinit base;
@@ -109,6 +109,7 @@ nv10_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	priv->base.meminit = nv10_devinit_meminit;
+	priv->base.pll_set = nv04_devinit_pll_set;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv1a.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv1a.c
index 5b2ba630d913..e9743cdabe75 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv1a.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv1a.c
@@ -22,8 +22,7 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/devinit.h>
-#include <subdev/vga.h>
+#include "priv.h"
 
 struct nv1a_devinit_priv {
 	struct nouveau_devinit base;
@@ -43,6 +42,7 @@ nv1a_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
+	priv->base.pll_set = nv04_devinit_pll_set;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv20.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv20.c
index eb32e99005e4..6cc6080d3bc0 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv20.c
@@ -24,9 +24,7 @@
  *
  */
 
-#include <subdev/devinit.h>
-#include <subdev/vga.h>
-
+#include "priv.h"
 #include "fbmem.h"
 
 struct nv20_devinit_priv {
@@ -81,6 +79,7 @@ nv20_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 		return ret;
 
 	priv->base.meminit = nv20_devinit_meminit;
+	priv->base.pll_set = nv04_devinit_pll_set;
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c
index 4a8577838417..6df72247c477 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nv50.c
@@ -1,5 +1,5 @@
 /*
- * Copyright 2012 Red Hat Inc.
+ * Copyright 2013 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -26,37 +26,55 @@
 #include <subdev/bios/dcb.h>
 #include <subdev/bios/disp.h>
 #include <subdev/bios/init.h>
-#include <subdev/devinit.h>
 #include <subdev/vga.h>
 
-struct nv50_devinit_priv {
-	struct nouveau_devinit base;
-};
+#include "priv.h"
 
 static int
-nv50_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
-		  struct nouveau_oclass *oclass, void *data, u32 size,
-		  struct nouveau_object **pobject)
+nv50_devinit_pll_set(struct nouveau_devinit *devinit, u32 type, u32 freq)
 {
-	struct nv50_devinit_priv *priv;
+	struct nv50_devinit_priv *priv = (void *)devinit;
+	struct nouveau_bios *bios = nouveau_bios(priv);
+	struct nvbios_pll info;
+	int N1, M1, N2, M2, P;
 	int ret;
 
-	ret = nouveau_devinit_create(parent, engine, oclass, &priv);
-	*pobject = nv_object(priv);
-	if (ret)
+	ret = nvbios_pll_parse(bios, type, &info);
+	if (ret) {
+		nv_error(devinit, "failed to retrieve pll data, %d\n", ret);
 		return ret;
+	}
 
-	return 0;
-}
+	ret = nv04_pll_calc(nv_subdev(devinit), &info, freq, &N1, &M1, &N2, &M2, &P);
+	if (!ret) {
+		nv_error(devinit, "failed pll calculation\n");
+		return ret;
+	}
 
-static void
-nv50_devinit_dtor(struct nouveau_object *object)
-{
-	struct nv50_devinit_priv *priv = (void *)object;
-	nouveau_devinit_destroy(&priv->base);
+	switch (info.type) {
+	case PLL_VPLL0:
+	case PLL_VPLL1:
+		nv_wr32(priv, info.reg + 0, 0x10000611);
+		nv_mask(priv, info.reg + 4, 0x00ff00ff, (M1 << 16) | N1);
+		nv_mask(priv, info.reg + 8, 0x7fff00ff, (P  << 28) |
+							(M2 << 16) | N2);
+		break;
+	case PLL_MEMORY:
+		nv_mask(priv, info.reg + 0, 0x01ff0000, (P << 22) |
+						        (info.bias_p << 19) |
+							(P << 16));
+		nv_wr32(priv, info.reg + 4, (N1 << 8) | M1);
+		break;
+	default:
+		nv_mask(priv, info.reg + 0, 0x00070000, (P << 16));
+		nv_wr32(priv, info.reg + 4, (N1 << 8) | M1);
+		break;
+	}
+
+	return 0;
 }
 
-static int
+int
 nv50_devinit_init(struct nouveau_object *object)
 {
 	struct nouveau_bios *bios = nouveau_bios(object);
@@ -103,10 +121,20 @@ nv50_devinit_init(struct nouveau_object *object)
 }
 
 static int
-nv50_devinit_fini(struct nouveau_object *object, bool suspend)
+nv50_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		  struct nouveau_oclass *oclass, void *data, u32 size,
+		  struct nouveau_object **pobject)
 {
-	struct nv50_devinit_priv *priv = (void *)object;
-	return nouveau_devinit_fini(&priv->base, suspend);
+	struct nv50_devinit_priv *priv;
+	int ret;
+
+	ret = nouveau_devinit_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.pll_set = nv50_devinit_pll_set;
+	return 0;
 }
 
 struct nouveau_oclass
@@ -114,8 +142,8 @@ nv50_devinit_oclass = {
 	.handle = NV_SUBDEV(DEVINIT, 0x50),
 	.ofuncs = &(struct nouveau_ofuncs) {
 		.ctor = nv50_devinit_ctor,
-		.dtor = nv50_devinit_dtor,
+		.dtor = _nouveau_devinit_dtor,
 		.init = nv50_devinit_init,
-		.fini = nv50_devinit_fini,
+		.fini = _nouveau_devinit_fini,
 	},
 };
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nva3.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nva3.c
new file mode 100644
index 000000000000..76a68b290141
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nva3.c
@@ -0,0 +1,87 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nva3_devinit_pll_set(struct nouveau_devinit *devinit, u32 type, u32 freq)
+{
+	struct nva3_devinit_priv *priv = (void *)devinit;
+	struct nouveau_bios *bios = nouveau_bios(priv);
+	struct nvbios_pll info;
+	int N, fN, M, P;
+	int ret;
+
+	ret = nvbios_pll_parse(bios, type, &info);
+	if (ret)
+		return ret;
+
+	ret = nva3_pll_calc(nv_subdev(devinit), &info, freq, &N, &fN, &M, &P);
+	if (ret < 0)
+		return ret;
+
+	switch (info.type) {
+	case PLL_VPLL0:
+	case PLL_VPLL1:
+		nv_wr32(priv, info.reg + 0, 0x50000610);
+		nv_mask(priv, info.reg + 4, 0x003fffff,
+					    (P << 16) | (M << 8) | N);
+		nv_wr32(priv, info.reg + 8, fN);
+		break;
+	default:
+		nv_warn(priv, "0x%08x/%dKhz unimplemented\n", type, freq);
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+static int
+nva3_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		  struct nouveau_oclass *oclass, void *data, u32 size,
+		  struct nouveau_object **pobject)
+{
+	struct nv50_devinit_priv *priv;
+	int ret;
+
+	ret = nouveau_devinit_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.pll_set = nva3_devinit_pll_set;
+	return 0;
+}
+
+struct nouveau_oclass
+nva3_devinit_oclass = {
+	.handle = NV_SUBDEV(DEVINIT, 0xa3),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nva3_devinit_ctor,
+		.dtor = _nouveau_devinit_dtor,
+		.init = nv50_devinit_init,
+		.fini = _nouveau_devinit_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/devinit/nvc0.c
new file mode 100644
index 000000000000..19e265bf4574
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/nvc0.c
@@ -0,0 +1,90 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nvc0_devinit_pll_set(struct nouveau_devinit *devinit, u32 type, u32 freq)
+{
+	struct nvc0_devinit_priv *priv = (void *)devinit;
+	struct nouveau_bios *bios = nouveau_bios(priv);
+	struct nvbios_pll info;
+	int N, fN, M, P;
+	int ret;
+
+	ret = nvbios_pll_parse(bios, type, &info);
+	if (ret)
+		return ret;
+
+	ret = nva3_pll_calc(nv_subdev(devinit), &info, freq, &N, &fN, &M, &P);
+	if (ret < 0)
+		return ret;
+
+	switch (info.type) {
+	case PLL_VPLL0:
+	case PLL_VPLL1:
+	case PLL_VPLL2:
+	case PLL_VPLL3:
+		nv_mask(priv, info.reg + 0x0c, 0x00000000, 0x00000100);
+		nv_wr32(priv, info.reg + 0x04, (P << 16) | (N << 8) | M);
+		nv_wr32(priv, info.reg + 0x10, fN << 16);
+		break;
+	default:
+		nv_warn(priv, "0x%08x/%dKhz unimplemented\n", type, freq);
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+static int
+nvc0_devinit_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
+		  struct nouveau_oclass *oclass, void *data, u32 size,
+		  struct nouveau_object **pobject)
+{
+	struct nv50_devinit_priv *priv;
+	int ret;
+
+	ret = nouveau_devinit_create(parent, engine, oclass, &priv);
+	*pobject = nv_object(priv);
+	if (ret)
+		return ret;
+
+	priv->base.pll_set = nvc0_devinit_pll_set;
+	if (nv_rd32(priv, 0x022500) & 0x00000001)
+		priv->base.post = true;
+	return 0;
+}
+
+struct nouveau_oclass
+nvc0_devinit_oclass = {
+	.handle = NV_SUBDEV(DEVINIT, 0xc0),
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_devinit_ctor,
+		.dtor = _nouveau_devinit_dtor,
+		.init = nv50_devinit_init,
+		.fini = _nouveau_devinit_fini,
+	},
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/devinit/priv.h b/drivers/gpu/drm/nouveau/core/subdev/devinit/priv.h
new file mode 100644
index 000000000000..7d622e2b0171
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/devinit/priv.h
@@ -0,0 +1,25 @@
+#ifndef __NVKM_DEVINIT_PRIV_H__
+#define __NVKM_DEVINIT_PRIV_H__
+
+#include <subdev/bios.h>
+#include <subdev/bios/pll.h>
+#include <subdev/clock/pll.h>
+#include <subdev/devinit.h>
+
+void nv04_devinit_dtor(struct nouveau_object *);
+int  nv04_devinit_init(struct nouveau_object *);
+int  nv04_devinit_fini(struct nouveau_object *, bool);
+int  nv04_devinit_pll_set(struct nouveau_devinit *, u32, u32);
+
+void setPLL_single(struct nouveau_devinit *, u32, struct nouveau_pll_vals *);
+void setPLL_double_highregs(struct nouveau_devinit *, u32, struct nouveau_pll_vals *);
+void setPLL_double_lowregs(struct nouveau_devinit *, u32, struct nouveau_pll_vals *);
+
+
+struct nv50_devinit_priv {
+	struct nouveau_devinit base;
+};
+
+int  nv50_devinit_init(struct nouveau_object *);
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/base.c b/drivers/gpu/drm/nouveau/core/subdev/fb/base.c
index d62045f454b2..821cd75b86a3 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/base.c
@@ -57,7 +57,57 @@ nouveau_fb_bios_memtype(struct nouveau_bios *bios)
 }
 
 int
-nouveau_fb_preinit(struct nouveau_fb *pfb)
+_nouveau_fb_fini(struct nouveau_object *object, bool suspend)
+{
+	struct nouveau_fb *pfb = (void *)object;
+	int ret;
+
+	ret = nv_ofuncs(pfb->ram)->fini(nv_object(pfb->ram), suspend);
+	if (ret && suspend)
+		return ret;
+
+	return nouveau_subdev_fini(&pfb->base, suspend);
+}
+
+int
+_nouveau_fb_init(struct nouveau_object *object)
+{
+	struct nouveau_fb *pfb = (void *)object;
+	int ret, i;
+
+	ret = nouveau_subdev_init(&pfb->base);
+	if (ret)
+		return ret;
+
+	ret = nv_ofuncs(pfb->ram)->init(nv_object(pfb->ram));
+	if (ret)
+		return ret;
+
+	for (i = 0; i < pfb->tile.regions; i++)
+		pfb->tile.prog(pfb, i, &pfb->tile.region[i]);
+
+	return 0;
+}
+
+void
+_nouveau_fb_dtor(struct nouveau_object *object)
+{
+	struct nouveau_fb *pfb = (void *)object;
+	int i;
+
+	for (i = 0; i < pfb->tile.regions; i++)
+		pfb->tile.fini(pfb, i, &pfb->tile.region[i]);
+	nouveau_mm_fini(&pfb->tags);
+	nouveau_mm_fini(&pfb->vram);
+
+	nouveau_object_ref(NULL, (struct nouveau_object **)&pfb->ram);
+	nouveau_subdev_destroy(&pfb->base);
+}
+
+int
+nouveau_fb_create_(struct nouveau_object *parent, struct nouveau_object *engine,
+		   struct nouveau_oclass *oclass, struct nouveau_oclass *ramcls,
+		   int length, void **pobject)
 {
 	static const char *name[] = {
 		[NV_MEM_TYPE_UNKNOWN] = "unknown",
@@ -72,69 +122,42 @@ nouveau_fb_preinit(struct nouveau_fb *pfb)
 		[NV_MEM_TYPE_GDDR4  ] = "GDDR4",
 		[NV_MEM_TYPE_GDDR5  ] = "GDDR5",
 	};
-	int ret, tags;
+	struct nouveau_object *ram;
+	struct nouveau_fb *pfb;
+	int ret;
 
-	tags = pfb->ram.init(pfb);
-	if (tags < 0 || !pfb->ram.size) {
+	ret = nouveau_subdev_create_(parent, engine, oclass, 0, "PFB", "fb",
+				     length, pobject);
+	pfb = *pobject;
+	if (ret)
+		return ret;
+
+	ret = nouveau_object_ctor(nv_object(pfb), nv_object(pfb),
+				  ramcls, NULL, 0, &ram);
+	if (ret) {
 		nv_fatal(pfb, "error detecting memory configuration!!\n");
-		return (tags < 0) ? tags : -ERANGE;
+		return ret;
 	}
 
+	atomic_dec(&ram->parent->refcount);
+	atomic_dec(&ram->engine->refcount);
+	pfb->ram = (void *)ram;
+
 	if (!nouveau_mm_initialised(&pfb->vram)) {
-		ret = nouveau_mm_init(&pfb->vram, 0, pfb->ram.size >> 12, 1);
+		ret = nouveau_mm_init(&pfb->vram, 0, pfb->ram->size >> 12, 1);
 		if (ret)
 			return ret;
 	}
 
 	if (!nouveau_mm_initialised(&pfb->tags)) {
-		ret = nouveau_mm_init(&pfb->tags, 0, tags ? ++tags : 0, 1);
+		ret = nouveau_mm_init(&pfb->tags, 0, pfb->ram->tags ?
+				     ++pfb->ram->tags : 0, 1);
 		if (ret)
 			return ret;
 	}
 
-	nv_info(pfb, "RAM type: %s\n", name[pfb->ram.type]);
-	nv_info(pfb, "RAM size: %d MiB\n", (int)(pfb->ram.size >> 20));
-	nv_info(pfb, "   ZCOMP: %d tags\n", tags);
+	nv_info(pfb, "RAM type: %s\n", name[pfb->ram->type]);
+	nv_info(pfb, "RAM size: %d MiB\n", (int)(pfb->ram->size >> 20));
+	nv_info(pfb, "   ZCOMP: %d tags\n", pfb->ram->tags);
 	return 0;
 }
-
-void
-nouveau_fb_destroy(struct nouveau_fb *pfb)
-{
-	int i;
-
-	for (i = 0; i < pfb->tile.regions; i++)
-		pfb->tile.fini(pfb, i, &pfb->tile.region[i]);
-	nouveau_mm_fini(&pfb->tags);
-	nouveau_mm_fini(&pfb->vram);
-
-	nouveau_subdev_destroy(&pfb->base);
-}
-
-void
-_nouveau_fb_dtor(struct nouveau_object *object)
-{
-	struct nouveau_fb *pfb = (void *)object;
-	nouveau_fb_destroy(pfb);
-}
-int
-nouveau_fb_init(struct nouveau_fb *pfb)
-{
-	int ret, i;
-
-	ret = nouveau_subdev_init(&pfb->base);
-	if (ret)
-		return ret;
-
-	for (i = 0; i < pfb->tile.regions; i++)
-		pfb->tile.prog(pfb, i, &pfb->tile.region[i]);
-
-	return 0;
-}
-
-int
-_nouveau_fb_init(struct nouveau_object *object)
-{
-	struct nouveau_fb *pfb = (void *)object;
-	return nouveau_fb_init(pfb);
-}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv04.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv04.c
index 6e369f85361e..1f103c7b89fa 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv04.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv04.c
@@ -22,24 +22,8 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
-#define NV04_PFB_BOOT_0						0x00100000
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT			0x00000003
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_32MB			0x00000000
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_4MB			0x00000001
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_8MB			0x00000002
-#	define NV04_PFB_BOOT_0_RAM_AMOUNT_16MB			0x00000003
-#	define NV04_PFB_BOOT_0_RAM_WIDTH_128			0x00000004
-#	define NV04_PFB_BOOT_0_RAM_TYPE				0x00000028
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_8MBIT		0x00000000
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT		0x00000008
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT_4BANK	0x00000010
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_16MBIT		0x00000018
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBIT		0x00000020
-#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBITX16		0x00000028
-#	define NV04_PFB_BOOT_0_UMA_ENABLE			0x00000100
-#	define NV04_PFB_BOOT_0_UMA_SIZE				0x0000f000
 #define NV04_PFB_CFG0						0x00100200
 
 struct nv04_fb_priv {
@@ -56,37 +40,6 @@ nv04_fb_memtype_valid(struct nouveau_fb *pfb, u32 tile_flags)
 }
 
 static int
-nv04_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 boot0 = nv_rd32(pfb, NV04_PFB_BOOT_0);
-	if (boot0 & 0x00000100) {
-		pfb->ram.size  = ((boot0 >> 12) & 0xf) * 2 + 2;
-		pfb->ram.size *= 1024 * 1024;
-	} else {
-		switch (boot0 & NV04_PFB_BOOT_0_RAM_AMOUNT) {
-		case NV04_PFB_BOOT_0_RAM_AMOUNT_32MB:
-			pfb->ram.size = 32 * 1024 * 1024;
-			break;
-		case NV04_PFB_BOOT_0_RAM_AMOUNT_16MB:
-			pfb->ram.size = 16 * 1024 * 1024;
-			break;
-		case NV04_PFB_BOOT_0_RAM_AMOUNT_8MB:
-			pfb->ram.size = 8 * 1024 * 1024;
-			break;
-		case NV04_PFB_BOOT_0_RAM_AMOUNT_4MB:
-			pfb->ram.size = 4 * 1024 * 1024;
-			break;
-		}
-	}
-
-	if ((boot0 & 0x00000038) <= 0x10)
-		pfb->ram.type = NV_MEM_TYPE_SGRAM;
-	else
-		pfb->ram.type = NV_MEM_TYPE_SDRAM;
-	return 0;
-}
-
-static int
 nv04_fb_init(struct nouveau_object *object)
 {
 	struct nv04_fb_priv *priv = (void *)object;
@@ -112,14 +65,13 @@ nv04_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv04_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv04_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv04_fb_vram_init;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv10.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv10.c
index edbbe26e858d..be069b5306b6 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv10.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv10.c
@@ -24,25 +24,12 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv10_fb_priv {
 	struct nouveau_fb base;
 };
 
-static int
-nv10_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 cfg0 = nv_rd32(pfb, 0x100200);
-	if (cfg0 & 0x00000001)
-		pfb->ram.type = NV_MEM_TYPE_DDR1;
-	else
-		pfb->ram.type = NV_MEM_TYPE_SDRAM;
-
-	pfb->ram.size = nv_rd32(pfb, 0x10020c) & 0xff000000;
-	return 0;
-}
-
 void
 nv10_fb_tile_init(struct nouveau_fb *pfb, int i, u32 addr, u32 size, u32 pitch,
 		  u32 flags, struct nouveau_fb_tile *tile)
@@ -78,18 +65,17 @@ nv10_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv10_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv10_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv10_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv10_fb_tile_init;
 	priv->base.tile.fini = nv10_fb_tile_fini;
 	priv->base.tile.prog = nv10_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv1a.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv1a.c
index 48366841db4a..57a2af0079b3 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv1a.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv1a.c
@@ -24,38 +24,13 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv1a_fb_priv {
 	struct nouveau_fb base;
 };
 
 static int
-nv1a_fb_vram_init(struct nouveau_fb *pfb)
-{
-	struct pci_dev *bridge;
-	u32 mem, mib;
-
-	bridge = pci_get_bus_and_slot(0, PCI_DEVFN(0, 1));
-	if (!bridge) {
-		nv_fatal(pfb, "no bridge device\n");
-		return -ENODEV;
-	}
-
-	if (nv_device(pfb)->chipset == 0x1a) {
-		pci_read_config_dword(bridge, 0x7c, &mem);
-		mib = ((mem >> 6) & 31) + 1;
-	} else {
-		pci_read_config_dword(bridge, 0x84, &mem);
-		mib = ((mem >> 4) & 127) + 1;
-	}
-
-	pfb->ram.type = NV_MEM_TYPE_STOLEN;
-	pfb->ram.size = mib * 1024 * 1024;
-	return 0;
-}
-
-static int
 nv1a_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	     struct nouveau_oclass *oclass, void *data, u32 size,
 	     struct nouveau_object **pobject)
@@ -63,18 +38,17 @@ nv1a_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv1a_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv1a_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv1a_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv10_fb_tile_init;
 	priv->base.tile.fini = nv10_fb_tile_fini;
 	priv->base.tile.prog = nv10_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c
index 5d14612a2c8e..b18c4e63bb47 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv20.c
@@ -24,29 +24,12 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv20_fb_priv {
 	struct nouveau_fb base;
 };
 
-int
-nv20_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 pbus1218 = nv_rd32(pfb, 0x001218);
-
-	switch (pbus1218 & 0x00000300) {
-	case 0x00000000: pfb->ram.type = NV_MEM_TYPE_SDRAM; break;
-	case 0x00000100: pfb->ram.type = NV_MEM_TYPE_DDR1; break;
-	case 0x00000200: pfb->ram.type = NV_MEM_TYPE_GDDR3; break;
-	case 0x00000300: pfb->ram.type = NV_MEM_TYPE_GDDR2; break;
-	}
-	pfb->ram.size  = (nv_rd32(pfb, 0x10020c) & 0xff000000);
-	pfb->ram.parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
-
-	return nv_rd32(pfb, 0x100320);
-}
-
 void
 nv20_fb_tile_init(struct nouveau_fb *pfb, int i, u32 addr, u32 size, u32 pitch,
 		  u32 flags, struct nouveau_fb_tile *tile)
@@ -65,7 +48,7 @@ nv20_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 		  struct nouveau_fb_tile *tile)
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
-	u32 tags  = round_up(tiles / pfb->ram.parts, 0x40);
+	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
 	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
 		if (!(flags & 2)) tile->zcomp = 0x00000000; /* Z16 */
 		else              tile->zcomp = 0x04000000; /* Z24S8 */
@@ -105,19 +88,18 @@ nv20_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv20_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv20_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv20_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv20_fb_tile_init;
 	priv->base.tile.comp = nv20_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv20_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c
index 0042ace6bef9..32ccabf10c45 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv25.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv25_fb_priv {
 	struct nouveau_fb base;
@@ -35,7 +35,7 @@ nv25_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 		  struct nouveau_fb_tile *tile)
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
-	u32 tags  = round_up(tiles / pfb->ram.parts, 0x40);
+	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
 	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
 		if (!(flags & 2)) tile->zcomp = 0x00100000; /* Z16 */
 		else              tile->zcomp = 0x00200000; /* Z24S8 */
@@ -54,19 +54,18 @@ nv25_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv25_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv20_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv20_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv20_fb_tile_init;
 	priv->base.tile.comp = nv25_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv20_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c
index a7ba0d048aec..bef756d43d33 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv30.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv30_fb_priv {
 	struct nouveau_fb base;
@@ -54,7 +54,7 @@ nv30_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 		  struct nouveau_fb_tile *tile)
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
-	u32 tags  = round_up(tiles / pfb->ram.parts, 0x40);
+	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
 	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x01000000; /* Z16 */
 		else           tile->zcomp |= 0x02000000; /* Z24S8 */
@@ -132,19 +132,18 @@ nv30_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv30_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv20_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv20_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv30_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv20_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c
index 092f6f4f3521..097d8e3824f2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv35.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv35_fb_priv {
 	struct nouveau_fb base;
@@ -35,7 +35,7 @@ nv35_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 		  struct nouveau_fb_tile *tile)
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
-	u32 tags  = round_up(tiles / pfb->ram.parts, 0x40);
+	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
 	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x04000000; /* Z16 */
 		else           tile->zcomp |= 0x08000000; /* Z24S8 */
@@ -55,19 +55,18 @@ nv35_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv35_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv20_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv20_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv35_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv20_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c
index 797ab3b821b9..9d6d9df896d9 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv36.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv36_fb_priv {
 	struct nouveau_fb base;
@@ -35,7 +35,7 @@ nv36_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 		  struct nouveau_fb_tile *tile)
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x40);
-	u32 tags  = round_up(tiles / pfb->ram.parts, 0x40);
+	u32 tags  = round_up(tiles / pfb->ram->parts, 0x40);
 	if (!nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
 		if (flags & 2) tile->zcomp |= 0x10000000; /* Z16 */
 		else           tile->zcomp |= 0x20000000; /* Z24S8 */
@@ -55,19 +55,18 @@ nv36_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv36_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv20_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv20_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv36_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv20_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c
index 65e131b90f37..33b4393a7829 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv40.c
@@ -24,34 +24,18 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv40_fb_priv {
 	struct nouveau_fb base;
 };
 
-static int
-nv40_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 pbus1218 = nv_rd32(pfb, 0x001218);
-	switch (pbus1218 & 0x00000300) {
-	case 0x00000000: pfb->ram.type = NV_MEM_TYPE_SDRAM; break;
-	case 0x00000100: pfb->ram.type = NV_MEM_TYPE_DDR1; break;
-	case 0x00000200: pfb->ram.type = NV_MEM_TYPE_GDDR3; break;
-	case 0x00000300: pfb->ram.type = NV_MEM_TYPE_DDR2; break;
-	}
-
-	pfb->ram.size  =  nv_rd32(pfb, 0x10020c) & 0xff000000;
-	pfb->ram.parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
-	return nv_rd32(pfb, 0x100320);
-}
-
 void
 nv40_fb_tile_comp(struct nouveau_fb *pfb, int i, u32 size, u32 flags,
 		  struct nouveau_fb_tile *tile)
 {
 	u32 tiles = DIV_ROUND_UP(size, 0x80);
-	u32 tags  = round_up(tiles / pfb->ram.parts, 0x100);
+	u32 tags  = round_up(tiles / pfb->ram->parts, 0x100);
 	if ( (flags & 2) &&
 	    !nouveau_mm_head(&pfb->tags, 1, tags, tags, 1, &tile->tag)) {
 		tile->zcomp  = 0x28000000; /* Z24S8_SPLIT_GRAD */
@@ -85,19 +69,18 @@ nv40_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv40_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv40_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv40_fb_vram_init;
 	priv->base.tile.regions = 8;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv40_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv20_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv41.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv41.c
index e9e5a08c41a1..02cd83789cd4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv41.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv41.c
@@ -24,28 +24,12 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv41_fb_priv {
 	struct nouveau_fb base;
 };
 
-int
-nv41_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 pfb474 = nv_rd32(pfb, 0x100474);
-	if (pfb474 & 0x00000004)
-		pfb->ram.type = NV_MEM_TYPE_GDDR3;
-	if (pfb474 & 0x00000002)
-		pfb->ram.type = NV_MEM_TYPE_DDR2;
-	if (pfb474 & 0x00000001)
-		pfb->ram.type = NV_MEM_TYPE_DDR1;
-
-	pfb->ram.size =   nv_rd32(pfb, 0x10020c) & 0xff000000;
-	pfb->ram.parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
-	return nv_rd32(pfb, 0x100320);
-}
-
 void
 nv41_fb_tile_prog(struct nouveau_fb *pfb, int i, struct nouveau_fb_tile *tile)
 {
@@ -78,19 +62,18 @@ nv41_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv41_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv41_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv41_fb_vram_init;
 	priv->base.tile.regions = 12;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv40_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv41_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv44.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv44.c
index ae89b5006f7a..c5246c29f293 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv44.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv44.c
@@ -24,27 +24,12 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv44_fb_priv {
 	struct nouveau_fb base;
 };
 
-int
-nv44_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 pfb474 = nv_rd32(pfb, 0x100474);
-	if (pfb474 & 0x00000004)
-		pfb->ram.type = NV_MEM_TYPE_GDDR3;
-	if (pfb474 & 0x00000002)
-		pfb->ram.type = NV_MEM_TYPE_DDR2;
-	if (pfb474 & 0x00000001)
-		pfb->ram.type = NV_MEM_TYPE_DDR1;
-
-	pfb->ram.size = nv_rd32(pfb, 0x10020c) & 0xff000000;
-	return 0;
-}
-
 static void
 nv44_fb_tile_init(struct nouveau_fb *pfb, int i, u32 addr, u32 size, u32 pitch,
 		  u32 flags, struct nouveau_fb_tile *tile)
@@ -87,18 +72,17 @@ nv44_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv44_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv44_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv44_fb_vram_init;
 	priv->base.tile.regions = 12;
 	priv->base.tile.init = nv44_fb_tile_init;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv44_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv46.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv46.c
index 589b93ea2994..e2b57909bfca 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv46.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv46.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv46_fb_priv {
 	struct nouveau_fb base;
@@ -52,18 +52,17 @@ nv46_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv46_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv44_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv44_fb_vram_init;
 	priv->base.tile.regions = 15;
 	priv->base.tile.init = nv46_fb_tile_init;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv44_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv47.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv47.c
index 818bba35b368..fe6a2278621d 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv47.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv47.c
@@ -24,7 +24,7 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv47_fb_priv {
 	struct nouveau_fb base;
@@ -38,19 +38,18 @@ nv47_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv47_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv41_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv41_fb_vram_init;
 	priv->base.tile.regions = 15;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv40_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv41_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv49.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv49.c
index 84a31af16ab4..5eca99b8c7e2 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv49.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv49.c
@@ -24,30 +24,13 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv49_fb_priv {
 	struct nouveau_fb base;
 };
 
 static int
-nv49_fb_vram_init(struct nouveau_fb *pfb)
-{
-	u32 pfb914 = nv_rd32(pfb, 0x100914);
-
-	switch (pfb914 & 0x00000003) {
-	case 0x00000000: pfb->ram.type = NV_MEM_TYPE_DDR1; break;
-	case 0x00000001: pfb->ram.type = NV_MEM_TYPE_DDR2; break;
-	case 0x00000002: pfb->ram.type = NV_MEM_TYPE_GDDR3; break;
-	case 0x00000003: break;
-	}
-
-	pfb->ram.size =   nv_rd32(pfb, 0x10020c) & 0xff000000;
-	pfb->ram.parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
-	return nv_rd32(pfb, 0x100320);
-}
-
-static int
 nv49_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	     struct nouveau_oclass *oclass, void *data, u32 size,
 	     struct nouveau_object **pobject)
@@ -55,20 +38,18 @@ nv49_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv49_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv49_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv49_fb_vram_init;
 	priv->base.tile.regions = 15;
 	priv->base.tile.init = nv30_fb_tile_init;
 	priv->base.tile.comp = nv40_fb_tile_comp;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv41_fb_tile_prog;
-
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv4e.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv4e.c
index 797fd558170b..1190b78a1e91 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv4e.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv4e.c
@@ -24,21 +24,13 @@
  *
  */
 
-#include <subdev/fb.h>
+#include "priv.h"
 
 struct nv4e_fb_priv {
 	struct nouveau_fb base;
 };
 
 static int
-nv4e_fb_vram_init(struct nouveau_fb *pfb)
-{
-	pfb->ram.size = nv_rd32(pfb, 0x10020c) & 0xff000000;
-	pfb->ram.type = NV_MEM_TYPE_STOLEN;
-	return 0;
-}
-
-static int
 nv4e_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	     struct nouveau_oclass *oclass, void *data, u32 size,
 	     struct nouveau_object **pobject)
@@ -46,18 +38,17 @@ nv4e_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv4e_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv4e_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nv04_fb_memtype_valid;
-	priv->base.ram.init = nv4e_fb_vram_init;
 	priv->base.tile.regions = 12;
 	priv->base.tile.init = nv46_fb_tile_init;
 	priv->base.tile.fini = nv20_fb_tile_fini;
 	priv->base.tile.prog = nv44_fb_tile_prog;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 struct nouveau_oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
index 0772ec978165..da614ec5564b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
@@ -27,7 +27,7 @@
 #include <core/engctx.h>
 #include <core/object.h>
 
-#include <subdev/fb.h>
+#include "priv.h"
 #include <subdev/bios.h>
 
 struct nv50_fb_priv {
@@ -36,7 +36,8 @@ struct nv50_fb_priv {
 	dma_addr_t r100c08;
 };
 
-static int types[0x80] = {
+int
+nv50_fb_memtype[0x80] = {
 	1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
 	1, 1, 1, 1, 0, 0, 0, 0, 2, 2, 2, 2, 0, 0, 0, 0,
 	1, 1, 1, 1, 1, 1, 1, 0, 2, 2, 2, 2, 2, 2, 2, 0,
@@ -50,192 +51,7 @@ static int types[0x80] = {
 static bool
 nv50_fb_memtype_valid(struct nouveau_fb *pfb, u32 memtype)
 {
-	return types[(memtype & 0xff00) >> 8] != 0;
-}
-
-static u32
-nv50_fb_vram_rblock(struct nouveau_fb *pfb)
-{
-	int i, parts, colbits, rowbitsa, rowbitsb, banks;
-	u64 rowsize, predicted;
-	u32 r0, r4, rt, ru, rblock_size;
-
-	r0 = nv_rd32(pfb, 0x100200);
-	r4 = nv_rd32(pfb, 0x100204);
-	rt = nv_rd32(pfb, 0x100250);
-	ru = nv_rd32(pfb, 0x001540);
-	nv_debug(pfb, "memcfg 0x%08x 0x%08x 0x%08x 0x%08x\n", r0, r4, rt, ru);
-
-	for (i = 0, parts = 0; i < 8; i++) {
-		if (ru & (0x00010000 << i))
-			parts++;
-	}
-
-	colbits  =  (r4 & 0x0000f000) >> 12;
-	rowbitsa = ((r4 & 0x000f0000) >> 16) + 8;
-	rowbitsb = ((r4 & 0x00f00000) >> 20) + 8;
-	banks    = 1 << (((r4 & 0x03000000) >> 24) + 2);
-
-	rowsize = parts * banks * (1 << colbits) * 8;
-	predicted = rowsize << rowbitsa;
-	if (r0 & 0x00000004)
-		predicted += rowsize << rowbitsb;
-
-	if (predicted != pfb->ram.size) {
-		nv_warn(pfb, "memory controller reports %d MiB VRAM\n",
-			(u32)(pfb->ram.size >> 20));
-	}
-
-	rblock_size = rowsize;
-	if (rt & 1)
-		rblock_size *= 3;
-
-	nv_debug(pfb, "rblock %d bytes\n", rblock_size);
-	return rblock_size;
-}
-
-static int
-nv50_fb_vram_init(struct nouveau_fb *pfb)
-{
-	struct nouveau_device *device = nv_device(pfb);
-	struct nouveau_bios *bios = nouveau_bios(device);
-	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
-	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
-	u32 size, tags = 0;
-	int ret;
-
-	pfb->ram.size = nv_rd32(pfb, 0x10020c);
-	pfb->ram.size = (pfb->ram.size & 0xffffff00) |
-		       ((pfb->ram.size & 0x000000ff) << 32);
-
-	size = (pfb->ram.size >> 12) - rsvd_head - rsvd_tail;
-	switch (device->chipset) {
-	case 0xaa:
-	case 0xac:
-	case 0xaf: /* IGPs, no reordering, no real VRAM */
-		ret = nouveau_mm_init(&pfb->vram, rsvd_head, size, 1);
-		if (ret)
-			return ret;
-
-		pfb->ram.type   = NV_MEM_TYPE_STOLEN;
-		pfb->ram.stolen = (u64)nv_rd32(pfb, 0x100e10) << 12;
-		break;
-	default:
-		switch (nv_rd32(pfb, 0x100714) & 0x00000007) {
-		case 0: pfb->ram.type = NV_MEM_TYPE_DDR1; break;
-		case 1:
-			if (nouveau_fb_bios_memtype(bios) == NV_MEM_TYPE_DDR3)
-				pfb->ram.type = NV_MEM_TYPE_DDR3;
-			else
-				pfb->ram.type = NV_MEM_TYPE_DDR2;
-			break;
-		case 2: pfb->ram.type = NV_MEM_TYPE_GDDR3; break;
-		case 3: pfb->ram.type = NV_MEM_TYPE_GDDR4; break;
-		case 4: pfb->ram.type = NV_MEM_TYPE_GDDR5; break;
-		default:
-			break;
-		}
-
-		ret = nouveau_mm_init(&pfb->vram, rsvd_head, size,
-				      nv50_fb_vram_rblock(pfb) >> 12);
-		if (ret)
-			return ret;
-
-		pfb->ram.ranks = (nv_rd32(pfb, 0x100200) & 0x4) ? 2 : 1;
-		tags = nv_rd32(pfb, 0x100320);
-		break;
-	}
-
-	return tags;
-}
-
-static int
-nv50_fb_vram_new(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
-		 u32 memtype, struct nouveau_mem **pmem)
-{
-	struct nv50_fb_priv *priv = (void *)pfb;
-	struct nouveau_mm *heap = &priv->base.vram;
-	struct nouveau_mm *tags = &priv->base.tags;
-	struct nouveau_mm_node *r;
-	struct nouveau_mem *mem;
-	int comp = (memtype & 0x300) >> 8;
-	int type = (memtype & 0x07f);
-	int back = (memtype & 0x800);
-	int min, max, ret;
-
-	max = (size >> 12);
-	min = ncmin ? (ncmin >> 12) : max;
-	align >>= 12;
-
-	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
-	if (!mem)
-		return -ENOMEM;
-
-	mutex_lock(&pfb->base.mutex);
-	if (comp) {
-		if (align == 16) {
-			int n = (max >> 4) * comp;
-
-			ret = nouveau_mm_head(tags, 1, n, n, 1, &mem->tag);
-			if (ret)
-				mem->tag = NULL;
-		}
-
-		if (unlikely(!mem->tag))
-			comp = 0;
-	}
-
-	INIT_LIST_HEAD(&mem->regions);
-	mem->memtype = (comp << 7) | type;
-	mem->size = max;
-
-	type = types[type];
-	do {
-		if (back)
-			ret = nouveau_mm_tail(heap, type, max, min, align, &r);
-		else
-			ret = nouveau_mm_head(heap, type, max, min, align, &r);
-		if (ret) {
-			mutex_unlock(&pfb->base.mutex);
-			pfb->ram.put(pfb, &mem);
-			return ret;
-		}
-
-		list_add_tail(&r->rl_entry, &mem->regions);
-		max -= r->length;
-	} while (max);
-	mutex_unlock(&pfb->base.mutex);
-
-	r = list_first_entry(&mem->regions, struct nouveau_mm_node, rl_entry);
-	mem->offset = (u64)r->offset << 12;
-	*pmem = mem;
-	return 0;
-}
-
-void
-nv50_fb_vram_del(struct nouveau_fb *pfb, struct nouveau_mem **pmem)
-{
-	struct nv50_fb_priv *priv = (void *)pfb;
-	struct nouveau_mm_node *this;
-	struct nouveau_mem *mem;
-
-	mem = *pmem;
-	*pmem = NULL;
-	if (unlikely(mem == NULL))
-		return;
-
-	mutex_lock(&pfb->base.mutex);
-	while (!list_empty(&mem->regions)) {
-		this = list_first_entry(&mem->regions, typeof(*this), rl_entry);
-
-		list_del(&this->rl_entry);
-		nouveau_mm_free(&priv->base.vram, &this);
-	}
-
-	nouveau_mm_free(&priv->base.tags, &mem->tag);
-	mutex_unlock(&pfb->base.mutex);
-
-	kfree(mem);
+	return nv50_fb_memtype[(memtype & 0xff00) >> 8] != 0;
 }
 
 static const struct nouveau_enum vm_dispatch_subclients[] = {
@@ -432,7 +248,7 @@ nv50_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nv50_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nv50_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
@@ -449,11 +265,8 @@ nv50_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	}
 
 	priv->base.memtype_valid = nv50_fb_memtype_valid;
-	priv->base.ram.init = nv50_fb_vram_init;
-	priv->base.ram.get = nv50_fb_vram_new;
-	priv->base.ram.put = nv50_fb_vram_del;
 	nv_subdev(priv)->intr = nv50_fb_intr;
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 static void
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c
index 86ad59203c8b..f35d76fd746d 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nvc0.c
@@ -22,9 +22,7 @@
  * Authors: Ben Skeggs
  */
 
-#include <subdev/fb.h>
-#include <subdev/ltcg.h>
-#include <subdev/bios.h>
+#include "priv.h"
 
 struct nvc0_fb_priv {
 	struct nouveau_fb base;
@@ -34,7 +32,6 @@ struct nvc0_fb_priv {
 
 extern const u8 nvc0_pte_storage_type_map[256];
 
-
 static bool
 nvc0_fb_memtype_valid(struct nouveau_fb *pfb, u32 tile_flags)
 {
@@ -43,137 +40,6 @@ nvc0_fb_memtype_valid(struct nouveau_fb *pfb, u32 tile_flags)
 }
 
 static int
-nvc0_fb_vram_init(struct nouveau_fb *pfb)
-{
-	struct nouveau_bios *bios = nouveau_bios(pfb);
-	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
-	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
-	u32 parts = nv_rd32(pfb, 0x022438);
-	u32 pmask = nv_rd32(pfb, 0x022554);
-	u32 bsize = nv_rd32(pfb, 0x10f20c);
-	u32 offset, length;
-	bool uniform = true;
-	int ret, part;
-
-	nv_debug(pfb, "0x100800: 0x%08x\n", nv_rd32(pfb, 0x100800));
-	nv_debug(pfb, "parts 0x%08x mask 0x%08x\n", parts, pmask);
-
-	pfb->ram.type = nouveau_fb_bios_memtype(bios);
-	pfb->ram.ranks = (nv_rd32(pfb, 0x10f200) & 0x00000004) ? 2 : 1;
-
-	/* read amount of vram attached to each memory controller */
-	for (part = 0; part < parts; part++) {
-		if (!(pmask & (1 << part))) {
-			u32 psize = nv_rd32(pfb, 0x11020c + (part * 0x1000));
-			if (psize != bsize) {
-				if (psize < bsize)
-					bsize = psize;
-				uniform = false;
-			}
-
-			nv_debug(pfb, "%d: mem_amount 0x%08x\n", part, psize);
-			pfb->ram.size += (u64)psize << 20;
-		}
-	}
-
-	/* if all controllers have the same amount attached, there's no holes */
-	if (uniform) {
-		offset = rsvd_head;
-		length = (pfb->ram.size >> 12) - rsvd_head - rsvd_tail;
-		return nouveau_mm_init(&pfb->vram, offset, length, 1);
-	}
-
-	/* otherwise, address lowest common amount from 0GiB */
-	ret = nouveau_mm_init(&pfb->vram, rsvd_head, (bsize << 8) * parts, 1);
-	if (ret)
-		return ret;
-
-	/* and the rest starting from (8GiB + common_size) */
-	offset = (0x0200000000ULL >> 12) + (bsize << 8);
-	length = (pfb->ram.size >> 12) - (bsize << 8) - rsvd_tail;
-
-	ret = nouveau_mm_init(&pfb->vram, offset, length, 0);
-	if (ret) {
-		nouveau_mm_fini(&pfb->vram);
-		return ret;
-	}
-
-	return 0;
-}
-
-static int
-nvc0_fb_vram_new(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
-		 u32 memtype, struct nouveau_mem **pmem)
-{
-	struct nouveau_mm *mm = &pfb->vram;
-	struct nouveau_mm_node *r;
-	struct nouveau_mem *mem;
-	int type = (memtype & 0x0ff);
-	int back = (memtype & 0x800);
-	int ret;
-	const bool comp = nvc0_pte_storage_type_map[type] != type;
-
-	size  >>= 12;
-	align >>= 12;
-	ncmin >>= 12;
-	if (!ncmin)
-		ncmin = size;
-
-	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
-	if (!mem)
-		return -ENOMEM;
-
-	INIT_LIST_HEAD(&mem->regions);
-	mem->size = size;
-
-	mutex_lock(&pfb->base.mutex);
-	if (comp) {
-		struct nouveau_ltcg *ltcg = nouveau_ltcg(pfb->base.base.parent);
-
-		/* compression only works with lpages */
-		if (align == (1 << (17 - 12))) {
-			int n = size >> 5;
-			ltcg->tags_alloc(ltcg, n, &mem->tag);
-		}
-		if (unlikely(!mem->tag))
-			type = nvc0_pte_storage_type_map[type];
-	}
-	mem->memtype = type;
-
-	do {
-		if (back)
-			ret = nouveau_mm_tail(mm, 1, size, ncmin, align, &r);
-		else
-			ret = nouveau_mm_head(mm, 1, size, ncmin, align, &r);
-		if (ret) {
-			mutex_unlock(&pfb->base.mutex);
-			pfb->ram.put(pfb, &mem);
-			return ret;
-		}
-
-		list_add_tail(&r->rl_entry, &mem->regions);
-		size -= r->length;
-	} while (size);
-	mutex_unlock(&pfb->base.mutex);
-
-	r = list_first_entry(&mem->regions, struct nouveau_mm_node, rl_entry);
-	mem->offset = (u64)r->offset << 12;
-	*pmem = mem;
-	return 0;
-}
-
-static void
-nvc0_fb_vram_del(struct nouveau_fb *pfb, struct nouveau_mem **pmem)
-{
-	struct nouveau_ltcg *ltcg = nouveau_ltcg(pfb->base.base.parent);
-
-	if ((*pmem)->tag)
-		ltcg->tags_free(ltcg, &(*pmem)->tag);
-
-	nv50_fb_vram_del(pfb, pmem);
-}
-
-static int
 nvc0_fb_init(struct nouveau_object *object)
 {
 	struct nvc0_fb_priv *priv = (void *)object;
@@ -212,15 +78,12 @@ nvc0_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	struct nvc0_fb_priv *priv;
 	int ret;
 
-	ret = nouveau_fb_create(parent, engine, oclass, &priv);
+	ret = nouveau_fb_create(parent, engine, oclass, &nvc0_ram_oclass, &priv);
 	*pobject = nv_object(priv);
 	if (ret)
 		return ret;
 
 	priv->base.memtype_valid = nvc0_fb_memtype_valid;
-	priv->base.ram.init = nvc0_fb_vram_init;
-	priv->base.ram.get = nvc0_fb_vram_new;
-	priv->base.ram.put = nvc0_fb_vram_del;
 
 	priv->r100c10_page = alloc_page(GFP_KERNEL | __GFP_ZERO);
 	if (priv->r100c10_page) {
@@ -231,7 +94,7 @@ nvc0_fb_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 			return -EFAULT;
 	}
 
-	return nouveau_fb_preinit(&priv->base);
+	return 0;
 }
 
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h b/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h
new file mode 100644
index 000000000000..6c974dd83e8b
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/priv.h
@@ -0,0 +1,87 @@
+#ifndef __NVKM_FB_PRIV_H__
+#define __NVKM_FB_PRIV_H__
+
+#include <subdev/fb.h>
+
+#define nouveau_ram_create(p,e,o,d)                                            \
+	nouveau_object_create_((p), (e), (o), 0, sizeof(**d), (void **)d)
+#define nouveau_ram_destroy(p)                                                 \
+	nouveau_object_destroy(&(p)->base)
+#define nouveau_ram_init(p)                                                    \
+	nouveau_object_init(&(p)->base)
+#define nouveau_ram_fini(p,s)                                                  \
+	nouveau_object_fini(&(p)->base, (s))
+
+#define _nouveau_ram_dtor nouveau_object_destroy
+#define _nouveau_ram_init nouveau_object_init
+#define _nouveau_ram_fini nouveau_object_fini
+
+extern struct nouveau_oclass nv04_ram_oclass;
+extern struct nouveau_oclass nv10_ram_oclass;
+extern struct nouveau_oclass nv1a_ram_oclass;
+extern struct nouveau_oclass nv20_ram_oclass;
+extern struct nouveau_oclass nv40_ram_oclass;
+extern struct nouveau_oclass nv41_ram_oclass;
+extern struct nouveau_oclass nv44_ram_oclass;
+extern struct nouveau_oclass nv49_ram_oclass;
+extern struct nouveau_oclass nv4e_ram_oclass;
+extern struct nouveau_oclass nv50_ram_oclass;
+extern struct nouveau_oclass nvc0_ram_oclass;
+
+#define nouveau_fb_create(p,e,c,r,d)                                           \
+	nouveau_fb_create_((p), (e), (c), (r), sizeof(**d), (void **)d)
+#define nouveau_fb_destroy(p) ({                                               \
+	struct nouveau_fb *pfb = (p);                                          \
+	_nouveau_fb_dtor(nv_object(pfb));                                      \
+})
+#define nouveau_fb_init(p) ({                                                  \
+	struct nouveau_fb *pfb = (p);                                          \
+	_nouveau_fb_init(nv_object(pfb));                                      \
+})
+#define nouveau_fb_fini(p,s) ({                                                \
+	struct nouveau_fb *pfb = (p);                                          \
+	_nouveau_fb_fini(nv_object(pfb), (s));                                 \
+})
+
+int nouveau_fb_create_(struct nouveau_object *, struct nouveau_object *,
+		       struct nouveau_oclass *, struct nouveau_oclass *,
+		       int length, void **pobject);
+void _nouveau_fb_dtor(struct nouveau_object *);
+int  _nouveau_fb_init(struct nouveau_object *);
+int  _nouveau_fb_fini(struct nouveau_object *, bool);
+
+struct nouveau_bios;
+int  nouveau_fb_bios_memtype(struct nouveau_bios *);
+
+bool nv04_fb_memtype_valid(struct nouveau_fb *, u32 memtype);
+
+void nv10_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
+		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
+void nv10_fb_tile_fini(struct nouveau_fb *, int i, struct nouveau_fb_tile *);
+void nv10_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
+
+void nv20_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
+		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
+void nv20_fb_tile_fini(struct nouveau_fb *, int i, struct nouveau_fb_tile *);
+void nv20_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
+
+int  nv30_fb_init(struct nouveau_object *);
+void nv30_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
+		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
+
+void nv40_fb_tile_comp(struct nouveau_fb *, int i, u32 size, u32 flags,
+		       struct nouveau_fb_tile *);
+
+int  nv41_fb_init(struct nouveau_object *);
+void nv41_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
+
+int  nv44_fb_init(struct nouveau_object *);
+void nv44_fb_tile_prog(struct nouveau_fb *, int, struct nouveau_fb_tile *);
+
+void nv46_fb_tile_init(struct nouveau_fb *, int i, u32 addr, u32 size,
+		       u32 pitch, u32 flags, struct nouveau_fb_tile *);
+
+void nv50_ram_put(struct nouveau_fb *, struct nouveau_mem **);
+extern int nv50_fb_memtype[0x80];
+
+#endif
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c
new file mode 100644
index 000000000000..e781080d3327
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv04.c
@@ -0,0 +1,95 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#define NV04_PFB_BOOT_0						0x00100000
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT			0x00000003
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_32MB			0x00000000
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_4MB			0x00000001
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_8MB			0x00000002
+#	define NV04_PFB_BOOT_0_RAM_AMOUNT_16MB			0x00000003
+#	define NV04_PFB_BOOT_0_RAM_WIDTH_128			0x00000004
+#	define NV04_PFB_BOOT_0_RAM_TYPE				0x00000028
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_8MBIT		0x00000000
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT		0x00000008
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SGRAM_16MBIT_4BANK	0x00000010
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_16MBIT		0x00000018
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBIT		0x00000020
+#	define NV04_PFB_BOOT_0_RAM_TYPE_SDRAM_64MBITX16		0x00000028
+#	define NV04_PFB_BOOT_0_UMA_ENABLE			0x00000100
+#	define NV04_PFB_BOOT_0_UMA_SIZE				0x0000f000
+
+#include "priv.h"
+
+static int
+nv04_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 boot0 = nv_rd32(pfb, NV04_PFB_BOOT_0);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	if (boot0 & 0x00000100) {
+		ram->size  = ((boot0 >> 12) & 0xf) * 2 + 2;
+		ram->size *= 1024 * 1024;
+	} else {
+		switch (boot0 & NV04_PFB_BOOT_0_RAM_AMOUNT) {
+		case NV04_PFB_BOOT_0_RAM_AMOUNT_32MB:
+			ram->size = 32 * 1024 * 1024;
+			break;
+		case NV04_PFB_BOOT_0_RAM_AMOUNT_16MB:
+			ram->size = 16 * 1024 * 1024;
+			break;
+		case NV04_PFB_BOOT_0_RAM_AMOUNT_8MB:
+			ram->size = 8 * 1024 * 1024;
+			break;
+		case NV04_PFB_BOOT_0_RAM_AMOUNT_4MB:
+			ram->size = 4 * 1024 * 1024;
+			break;
+		}
+	}
+
+	if ((boot0 & 0x00000038) <= 0x10)
+		ram->type = NV_MEM_TYPE_SGRAM;
+	else
+		ram->type = NV_MEM_TYPE_SDRAM;
+	return 0;
+}
+
+struct nouveau_oclass
+nv04_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv04_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv10.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv10.c
new file mode 100644
index 000000000000..8311f3774edf
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv10.c
@@ -0,0 +1,61 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv10_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 cfg0 = nv_rd32(pfb, 0x100200);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	if (cfg0 & 0x00000001)
+		ram->type = NV_MEM_TYPE_DDR1;
+	else
+		ram->type = NV_MEM_TYPE_SDRAM;
+
+	ram->size = nv_rd32(pfb, 0x10020c) & 0xff000000;
+	return 0;
+}
+
+
+struct nouveau_oclass
+nv10_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv10_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv1a.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv1a.c
new file mode 100644
index 000000000000..d0caddfb9db0
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv1a.c
@@ -0,0 +1,71 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv1a_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	struct pci_dev *bridge;
+	u32 mem, mib;
+	int ret;
+
+	bridge = pci_get_bus_and_slot(0, PCI_DEVFN(0, 1));
+	if (!bridge) {
+		nv_fatal(pfb, "no bridge device\n");
+		return -ENODEV;
+	}
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	if (nv_device(pfb)->chipset == 0x1a) {
+		pci_read_config_dword(bridge, 0x7c, &mem);
+		mib = ((mem >> 6) & 31) + 1;
+	} else {
+		pci_read_config_dword(bridge, 0x84, &mem);
+		mib = ((mem >> 4) & 127) + 1;
+	}
+
+	ram->type = NV_MEM_TYPE_STOLEN;
+	ram->size = mib * 1024 * 1024;
+	return 0;
+}
+
+struct nouveau_oclass
+nv1a_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv1a_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv20.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv20.c
new file mode 100644
index 000000000000..fdc11bba226d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv20.c
@@ -0,0 +1,63 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv20_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 pbus1218 = nv_rd32(pfb, 0x001218);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	switch (pbus1218 & 0x00000300) {
+	case 0x00000000: ram->type = NV_MEM_TYPE_SDRAM; break;
+	case 0x00000100: ram->type = NV_MEM_TYPE_DDR1; break;
+	case 0x00000200: ram->type = NV_MEM_TYPE_GDDR3; break;
+	case 0x00000300: ram->type = NV_MEM_TYPE_GDDR2; break;
+	}
+	ram->size  = (nv_rd32(pfb, 0x10020c) & 0xff000000);
+	ram->parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
+	ram->tags  = nv_rd32(pfb, 0x100320);
+	return 0;
+}
+
+struct nouveau_oclass
+nv20_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv20_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv40.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv40.c
new file mode 100644
index 000000000000..ee49ac4dbdb6
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv40.c
@@ -0,0 +1,65 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv40_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 pbus1218 = nv_rd32(pfb, 0x001218);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	switch (pbus1218 & 0x00000300) {
+	case 0x00000000: ram->type = NV_MEM_TYPE_SDRAM; break;
+	case 0x00000100: ram->type = NV_MEM_TYPE_DDR1; break;
+	case 0x00000200: ram->type = NV_MEM_TYPE_GDDR3; break;
+	case 0x00000300: ram->type = NV_MEM_TYPE_DDR2; break;
+	}
+
+	ram->size  =  nv_rd32(pfb, 0x10020c) & 0xff000000;
+	ram->parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
+	ram->tags  =  nv_rd32(pfb, 0x100320);
+	return 0;
+}
+
+
+struct nouveau_oclass
+nv40_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv40_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv41.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv41.c
new file mode 100644
index 000000000000..1dab7e12abab
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv41.c
@@ -0,0 +1,64 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv41_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 pfb474 = nv_rd32(pfb, 0x100474);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	if (pfb474 & 0x00000004)
+		ram->type = NV_MEM_TYPE_GDDR3;
+	if (pfb474 & 0x00000002)
+		ram->type = NV_MEM_TYPE_DDR2;
+	if (pfb474 & 0x00000001)
+		ram->type = NV_MEM_TYPE_DDR1;
+
+	ram->size  =  nv_rd32(pfb, 0x10020c) & 0xff000000;
+	ram->parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
+	ram->tags  =  nv_rd32(pfb, 0x100320);
+	return 0;
+}
+
+struct nouveau_oclass
+nv41_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv41_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv44.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv44.c
new file mode 100644
index 000000000000..25fff842e5c1
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv44.c
@@ -0,0 +1,62 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv44_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 pfb474 = nv_rd32(pfb, 0x100474);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	if (pfb474 & 0x00000004)
+		ram->type = NV_MEM_TYPE_GDDR3;
+	if (pfb474 & 0x00000002)
+		ram->type = NV_MEM_TYPE_DDR2;
+	if (pfb474 & 0x00000001)
+		ram->type = NV_MEM_TYPE_DDR1;
+
+	ram->size = nv_rd32(pfb, 0x10020c) & 0xff000000;
+	return 0;
+}
+
+struct nouveau_oclass
+nv44_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv44_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv49.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv49.c
new file mode 100644
index 000000000000..19e3a9a63a02
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv49.c
@@ -0,0 +1,64 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv49_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	u32 pfb914 = nv_rd32(pfb, 0x100914);
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	switch (pfb914 & 0x00000003) {
+	case 0x00000000: pfb->ram->type = NV_MEM_TYPE_DDR1; break;
+	case 0x00000001: pfb->ram->type = NV_MEM_TYPE_DDR2; break;
+	case 0x00000002: pfb->ram->type = NV_MEM_TYPE_GDDR3; break;
+	case 0x00000003: break;
+	}
+
+	pfb->ram->size  =  nv_rd32(pfb, 0x10020c) & 0xff000000;
+	pfb->ram->parts = (nv_rd32(pfb, 0x100200) & 0x00000003) + 1;
+	pfb->ram->tags  =  nv_rd32(pfb, 0x100320);
+	return 0;
+}
+
+struct nouveau_oclass
+nv49_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv49_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv4e.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv4e.c
new file mode 100644
index 000000000000..7192aa6e5577
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv4e.c
@@ -0,0 +1,55 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include "priv.h"
+
+static int
+nv4e_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_ram *ram;
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	pfb->ram->size = nv_rd32(pfb, 0x10020c) & 0xff000000;
+	pfb->ram->type = NV_MEM_TYPE_STOLEN;
+	return 0;
+}
+
+struct nouveau_oclass
+nv4e_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv4e_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c
new file mode 100644
index 000000000000..af5aa7ee8ad9
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnv50.c
@@ -0,0 +1,232 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/bios.h>
+#include <core/mm.h>
+#include "priv.h"
+
+void
+nv50_ram_put(struct nouveau_fb *pfb, struct nouveau_mem **pmem)
+{
+	struct nouveau_mm_node *this;
+	struct nouveau_mem *mem;
+
+	mem = *pmem;
+	*pmem = NULL;
+	if (unlikely(mem == NULL))
+		return;
+
+	mutex_lock(&pfb->base.mutex);
+	while (!list_empty(&mem->regions)) {
+		this = list_first_entry(&mem->regions, typeof(*this), rl_entry);
+
+		list_del(&this->rl_entry);
+		nouveau_mm_free(&pfb->vram, &this);
+	}
+
+	nouveau_mm_free(&pfb->tags, &mem->tag);
+	mutex_unlock(&pfb->base.mutex);
+
+	kfree(mem);
+}
+
+static int
+nv50_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
+	     u32 memtype, struct nouveau_mem **pmem)
+{
+	struct nouveau_mm *heap = &pfb->vram;
+	struct nouveau_mm *tags = &pfb->tags;
+	struct nouveau_mm_node *r;
+	struct nouveau_mem *mem;
+	int comp = (memtype & 0x300) >> 8;
+	int type = (memtype & 0x07f);
+	int back = (memtype & 0x800);
+	int min, max, ret;
+
+	max = (size >> 12);
+	min = ncmin ? (ncmin >> 12) : max;
+	align >>= 12;
+
+	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
+	if (!mem)
+		return -ENOMEM;
+
+	mutex_lock(&pfb->base.mutex);
+	if (comp) {
+		if (align == 16) {
+			int n = (max >> 4) * comp;
+
+			ret = nouveau_mm_head(tags, 1, n, n, 1, &mem->tag);
+			if (ret)
+				mem->tag = NULL;
+		}
+
+		if (unlikely(!mem->tag))
+			comp = 0;
+	}
+
+	INIT_LIST_HEAD(&mem->regions);
+	mem->memtype = (comp << 7) | type;
+	mem->size = max;
+
+	type = nv50_fb_memtype[type];
+	do {
+		if (back)
+			ret = nouveau_mm_tail(heap, type, max, min, align, &r);
+		else
+			ret = nouveau_mm_head(heap, type, max, min, align, &r);
+		if (ret) {
+			mutex_unlock(&pfb->base.mutex);
+			pfb->ram->put(pfb, &mem);
+			return ret;
+		}
+
+		list_add_tail(&r->rl_entry, &mem->regions);
+		max -= r->length;
+	} while (max);
+	mutex_unlock(&pfb->base.mutex);
+
+	r = list_first_entry(&mem->regions, struct nouveau_mm_node, rl_entry);
+	mem->offset = (u64)r->offset << 12;
+	*pmem = mem;
+	return 0;
+}
+
+static u32
+nv50_fb_vram_rblock(struct nouveau_fb *pfb, struct nouveau_ram *ram)
+{
+	int i, parts, colbits, rowbitsa, rowbitsb, banks;
+	u64 rowsize, predicted;
+	u32 r0, r4, rt, ru, rblock_size;
+
+	r0 = nv_rd32(pfb, 0x100200);
+	r4 = nv_rd32(pfb, 0x100204);
+	rt = nv_rd32(pfb, 0x100250);
+	ru = nv_rd32(pfb, 0x001540);
+	nv_debug(pfb, "memcfg 0x%08x 0x%08x 0x%08x 0x%08x\n", r0, r4, rt, ru);
+
+	for (i = 0, parts = 0; i < 8; i++) {
+		if (ru & (0x00010000 << i))
+			parts++;
+	}
+
+	colbits  =  (r4 & 0x0000f000) >> 12;
+	rowbitsa = ((r4 & 0x000f0000) >> 16) + 8;
+	rowbitsb = ((r4 & 0x00f00000) >> 20) + 8;
+	banks    = 1 << (((r4 & 0x03000000) >> 24) + 2);
+
+	rowsize = parts * banks * (1 << colbits) * 8;
+	predicted = rowsize << rowbitsa;
+	if (r0 & 0x00000004)
+		predicted += rowsize << rowbitsb;
+
+	if (predicted != ram->size) {
+		nv_warn(pfb, "memory controller reports %d MiB VRAM\n",
+			(u32)(ram->size >> 20));
+	}
+
+	rblock_size = rowsize;
+	if (rt & 1)
+		rblock_size *= 3;
+
+	nv_debug(pfb, "rblock %d bytes\n", rblock_size);
+	return rblock_size;
+}
+
+static int
+nv50_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 datasize,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_device *device = nv_device(pfb);
+	struct nouveau_bios *bios = nouveau_bios(device);
+	struct nouveau_ram *ram;
+	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
+	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
+	u32 size;
+	int ret;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	ram->size = nv_rd32(pfb, 0x10020c);
+	ram->size = (ram->size & 0xffffff00) |
+		       ((ram->size & 0x000000ff) << 32);
+
+	size = (ram->size >> 12) - rsvd_head - rsvd_tail;
+	switch (device->chipset) {
+	case 0xaa:
+	case 0xac:
+	case 0xaf: /* IGPs, no reordering, no real VRAM */
+		ret = nouveau_mm_init(&pfb->vram, rsvd_head, size, 1);
+		if (ret)
+			return ret;
+
+		ram->type   = NV_MEM_TYPE_STOLEN;
+		ram->stolen = (u64)nv_rd32(pfb, 0x100e10) << 12;
+		break;
+	default:
+		switch (nv_rd32(pfb, 0x100714) & 0x00000007) {
+		case 0: ram->type = NV_MEM_TYPE_DDR1; break;
+		case 1:
+			if (nouveau_fb_bios_memtype(bios) == NV_MEM_TYPE_DDR3)
+				ram->type = NV_MEM_TYPE_DDR3;
+			else
+				ram->type = NV_MEM_TYPE_DDR2;
+			break;
+		case 2: ram->type = NV_MEM_TYPE_GDDR3; break;
+		case 3: ram->type = NV_MEM_TYPE_GDDR4; break;
+		case 4: ram->type = NV_MEM_TYPE_GDDR5; break;
+		default:
+			break;
+		}
+
+		ret = nouveau_mm_init(&pfb->vram, rsvd_head, size,
+				      nv50_fb_vram_rblock(pfb, ram) >> 12);
+		if (ret)
+			return ret;
+
+		ram->ranks = (nv_rd32(pfb, 0x100200) & 0x4) ? 2 : 1;
+		ram->tags  =  nv_rd32(pfb, 0x100320);
+		break;
+	}
+
+	ram->get = nv50_ram_get;
+	ram->put = nv50_ram_put;
+	return 0;
+}
+
+struct nouveau_oclass
+nv50_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nv50_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c
new file mode 100644
index 000000000000..9c3634acbb9d
--- /dev/null
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/ramnvc0.c
@@ -0,0 +1,186 @@
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#include <subdev/bios.h>
+#include <subdev/ltcg.h>
+
+#include "priv.h"
+
+extern const u8 nvc0_pte_storage_type_map[256];
+
+void
+nvc0_ram_put(struct nouveau_fb *pfb, struct nouveau_mem **pmem)
+{
+	struct nouveau_ltcg *ltcg = nouveau_ltcg(pfb);
+
+	if ((*pmem)->tag)
+		ltcg->tags_free(ltcg, &(*pmem)->tag);
+
+	nv50_ram_put(pfb, pmem);
+}
+
+int
+nvc0_ram_get(struct nouveau_fb *pfb, u64 size, u32 align, u32 ncmin,
+	     u32 memtype, struct nouveau_mem **pmem)
+{
+	struct nouveau_mm *mm = &pfb->vram;
+	struct nouveau_mm_node *r;
+	struct nouveau_mem *mem;
+	int type = (memtype & 0x0ff);
+	int back = (memtype & 0x800);
+	const bool comp = nvc0_pte_storage_type_map[type] != type;
+	int ret;
+
+	size  >>= 12;
+	align >>= 12;
+	ncmin >>= 12;
+	if (!ncmin)
+		ncmin = size;
+
+	mem = kzalloc(sizeof(*mem), GFP_KERNEL);
+	if (!mem)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&mem->regions);
+	mem->size = size;
+
+	mutex_lock(&pfb->base.mutex);
+	if (comp) {
+		struct nouveau_ltcg *ltcg = nouveau_ltcg(pfb);
+
+		/* compression only works with lpages */
+		if (align == (1 << (17 - 12))) {
+			int n = size >> 5;
+			ltcg->tags_alloc(ltcg, n, &mem->tag);
+		}
+
+		if (unlikely(!mem->tag))
+			type = nvc0_pte_storage_type_map[type];
+	}
+	mem->memtype = type;
+
+	do {
+		if (back)
+			ret = nouveau_mm_tail(mm, 1, size, ncmin, align, &r);
+		else
+			ret = nouveau_mm_head(mm, 1, size, ncmin, align, &r);
+		if (ret) {
+			mutex_unlock(&pfb->base.mutex);
+			pfb->ram->put(pfb, &mem);
+			return ret;
+		}
+
+		list_add_tail(&r->rl_entry, &mem->regions);
+		size -= r->length;
+	} while (size);
+	mutex_unlock(&pfb->base.mutex);
+
+	r = list_first_entry(&mem->regions, struct nouveau_mm_node, rl_entry);
+	mem->offset = (u64)r->offset << 12;
+	*pmem = mem;
+	return 0;
+}
+
+static int
+nvc0_ram_create(struct nouveau_object *parent, struct nouveau_object *engine,
+		struct nouveau_oclass *oclass, void *data, u32 size,
+		struct nouveau_object **pobject)
+{
+	struct nouveau_fb *pfb = nouveau_fb(parent);
+	struct nouveau_bios *bios = nouveau_bios(pfb);
+	struct nouveau_ram *ram;
+	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
+	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
+	u32 parts = nv_rd32(pfb, 0x022438);
+	u32 pmask = nv_rd32(pfb, 0x022554);
+	u32 bsize = nv_rd32(pfb, 0x10f20c);
+	u32 offset, length;
+	bool uniform = true;
+	int ret, part;
+
+	ret = nouveau_ram_create(parent, engine, oclass, &ram);
+	*pobject = nv_object(ram);
+	if (ret)
+		return ret;
+
+	nv_debug(pfb, "0x100800: 0x%08x\n", nv_rd32(pfb, 0x100800));
+	nv_debug(pfb, "parts 0x%08x mask 0x%08x\n", parts, pmask);
+
+	ram->type = nouveau_fb_bios_memtype(bios);
+	ram->ranks = (nv_rd32(pfb, 0x10f200) & 0x00000004) ? 2 : 1;
+
+	/* read amount of vram attached to each memory controller */
+	for (part = 0; part < parts; part++) {
+		if (!(pmask & (1 << part))) {
+			u32 psize = nv_rd32(pfb, 0x11020c + (part * 0x1000));
+			if (psize != bsize) {
+				if (psize < bsize)
+					bsize = psize;
+				uniform = false;
+			}
+
+			nv_debug(pfb, "%d: mem_amount 0x%08x\n", part, psize);
+			ram->size += (u64)psize << 20;
+		}
+	}
+
+	/* if all controllers have the same amount attached, there's no holes */
+	if (uniform) {
+		offset = rsvd_head;
+		length = (ram->size >> 12) - rsvd_head - rsvd_tail;
+		ret = nouveau_mm_init(&pfb->vram, offset, length, 1);
+	} else {
+		/* otherwise, address lowest common amount from 0GiB */
+		ret = nouveau_mm_init(&pfb->vram, rsvd_head,
+				      (bsize << 8) * parts, 1);
+		if (ret)
+			return ret;
+
+		/* and the rest starting from (8GiB + common_size) */
+		offset = (0x0200000000ULL >> 12) + (bsize << 8);
+		length = (ram->size >> 12) - (bsize << 8) - rsvd_tail;
+
+		ret = nouveau_mm_init(&pfb->vram, offset, length, 0);
+		if (ret)
+			nouveau_mm_fini(&pfb->vram);
+	}
+
+	if (ret)
+		return ret;
+
+	ram->get = nvc0_ram_get;
+	ram->put = nvc0_ram_put;
+	return 0;
+}
+
+struct nouveau_oclass
+nvc0_ram_oclass = {
+	.handle = 0,
+	.ofuncs = &(struct nouveau_ofuncs) {
+		.ctor = nvc0_ram_create,
+		.dtor = _nouveau_ram_dtor,
+		.init = _nouveau_ram_init,
+		.fini = _nouveau_ram_fini,
+	}
+};
diff --git a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv50.c
index cfc7e31461de..97bc5dff93e7 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/instmem/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/instmem/nv50.c
@@ -56,7 +56,7 @@ nv50_instobj_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	if (ret)
 		return ret;
 
-	ret = pfb->ram.get(pfb, size, align, 0, 0x800, &node->mem);
+	ret = pfb->ram->get(pfb, size, align, 0, 0x800, &node->mem);
 	if (ret)
 		return ret;
 
@@ -71,7 +71,7 @@ nv50_instobj_dtor(struct nouveau_object *object)
 {
 	struct nv50_instobj_priv *node = (void *)object;
 	struct nouveau_fb *pfb = nouveau_fb(object);
-	pfb->ram.put(pfb, &node->mem);
+	pfb->ram->put(pfb, &node->mem);
 	nouveau_instobj_destroy(&node->base);
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/ltcg/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/ltcg/nvc0.c
index fb794e997fbc..bcca883018f4 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/ltcg/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/ltcg/nvc0.c
@@ -122,7 +122,7 @@ nvc0_ltcg_init_tag_ram(struct nouveau_fb *pfb, struct nvc0_ltcg_priv *priv)
 		nv_wr32(priv, 0x17e000, priv->part_nr);
 
 	/* tags for 1/4 of VRAM should be enough (8192/4 per GiB of VRAM) */
-	priv->num_tags = (pfb->ram.size >> 17) / 4;
+	priv->num_tags = (pfb->ram->size >> 17) / 4;
 	if (priv->num_tags > (1 << 17))
 		priv->num_tags = 1 << 17; /* we have 17 bits in PTE */
 	priv->num_tags = (priv->num_tags + 63) & ~63; /* round up to 64 */
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c
index d796924f9930..0cb322a5e72c 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nv50.c
@@ -35,6 +35,7 @@ nv50_mc_intr[] = {
 	{ 0x00001000, NVDEV_ENGINE_GR },
 	{ 0x00004000, NVDEV_ENGINE_CRYPT },	/* NV84- */
 	{ 0x00008000, NVDEV_ENGINE_BSP },	/* NV84- */
+	{ 0x00020000, NVDEV_ENGINE_VP },	/* NV84- */
 	{ 0x00100000, NVDEV_SUBDEV_TIMER },
 	{ 0x00200000, NVDEV_SUBDEV_GPIO },
 	{ 0x04000000, NVDEV_ENGINE_DISP },
diff --git a/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
index 737bd4b682e1..c5da3babbc62 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/mc/nvc0.c
@@ -33,6 +33,7 @@ nvc0_mc_intr[] = {
 	{ 0x00000001, NVDEV_ENGINE_PPP },
 	{ 0x00000020, NVDEV_ENGINE_COPY0 },
 	{ 0x00000040, NVDEV_ENGINE_COPY1 },
+	{ 0x00000080, NVDEV_ENGINE_COPY2 },
 	{ 0x00000100, NVDEV_ENGINE_FIFO },
 	{ 0x00001000, NVDEV_ENGINE_GR },
 	{ 0x00008000, NVDEV_ENGINE_BSP },
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/base.c b/drivers/gpu/drm/nouveau/core/subdev/vm/base.c
index 77c67fc970e6..67fcb6c852ac 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/base.c
@@ -236,9 +236,9 @@ nouveau_vm_unmap_pgt(struct nouveau_vm *vm, int big, u32 fpde, u32 lpde)
 			vmm->map_pgt(vpgd->obj, pde, vpgt->obj);
 		}
 
-		mutex_unlock(&vm->mm.mutex);
+		mutex_unlock(&nv_subdev(vmm)->mutex);
 		nouveau_gpuobj_ref(NULL, &pgt);
-		mutex_lock(&vm->mm.mutex);
+		mutex_lock(&nv_subdev(vmm)->mutex);
 	}
 }
 
@@ -256,18 +256,18 @@ nouveau_vm_map_pgt(struct nouveau_vm *vm, u32 pde, u32 type)
 	pgt_size  = (1 << (vmm->pgt_bits + 12)) >> type;
 	pgt_size *= 8;
 
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&nv_subdev(vmm)->mutex);
 	ret = nouveau_gpuobj_new(nv_object(vm->vmm), NULL, pgt_size, 0x1000,
 				 NVOBJ_FLAG_ZERO_ALLOC, &pgt);
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&nv_subdev(vmm)->mutex);
 	if (unlikely(ret))
 		return ret;
 
 	/* someone beat us to filling the PDE while we didn't have the lock */
 	if (unlikely(vpgt->refcount[big]++)) {
-		mutex_unlock(&vm->mm.mutex);
+		mutex_unlock(&nv_subdev(vmm)->mutex);
 		nouveau_gpuobj_ref(NULL, &pgt);
-		mutex_lock(&vm->mm.mutex);
+		mutex_lock(&nv_subdev(vmm)->mutex);
 		return 0;
 	}
 
@@ -289,11 +289,11 @@ nouveau_vm_get(struct nouveau_vm *vm, u64 size, u32 page_shift,
 	u32 fpde, lpde, pde;
 	int ret;
 
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&nv_subdev(vmm)->mutex);
 	ret = nouveau_mm_head(&vm->mm, page_shift, msize, msize, align,
 			     &vma->node);
 	if (unlikely(ret != 0)) {
-		mutex_unlock(&vm->mm.mutex);
+		mutex_unlock(&nv_subdev(vmm)->mutex);
 		return ret;
 	}
 
@@ -314,13 +314,14 @@ nouveau_vm_get(struct nouveau_vm *vm, u64 size, u32 page_shift,
 			if (pde != fpde)
 				nouveau_vm_unmap_pgt(vm, big, fpde, pde - 1);
 			nouveau_mm_free(&vm->mm, &vma->node);
-			mutex_unlock(&vm->mm.mutex);
+			mutex_unlock(&nv_subdev(vmm)->mutex);
 			return ret;
 		}
 	}
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&nv_subdev(vmm)->mutex);
 
-	vma->vm     = vm;
+	vma->vm = NULL;
+	nouveau_vm_ref(vm, &vma->vm, NULL);
 	vma->offset = (u64)vma->node->offset << 12;
 	vma->access = access;
 	return 0;
@@ -338,10 +339,12 @@ nouveau_vm_put(struct nouveau_vma *vma)
 	fpde = (vma->node->offset >> vmm->pgt_bits);
 	lpde = (vma->node->offset + vma->node->length - 1) >> vmm->pgt_bits;
 
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&nv_subdev(vmm)->mutex);
 	nouveau_vm_unmap_pgt(vm, vma->node->type != vmm->spg_shift, fpde, lpde);
 	nouveau_mm_free(&vm->mm, &vma->node);
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&nv_subdev(vmm)->mutex);
+
+	nouveau_vm_ref(NULL, &vma->vm, NULL);
 }
 
 int
@@ -362,7 +365,7 @@ nouveau_vm_create(struct nouveau_vmmgr *vmm, u64 offset, u64 length,
 	vm->fpde = offset >> (vmm->pgt_bits + 12);
 	vm->lpde = (offset + length - 1) >> (vmm->pgt_bits + 12);
 
-	vm->pgt  = kcalloc(vm->lpde - vm->fpde + 1, sizeof(*vm->pgt), GFP_KERNEL);
+	vm->pgt  = vzalloc((vm->lpde - vm->fpde + 1) * sizeof(*vm->pgt));
 	if (!vm->pgt) {
 		kfree(vm);
 		return -ENOMEM;
@@ -371,7 +374,7 @@ nouveau_vm_create(struct nouveau_vmmgr *vmm, u64 offset, u64 length,
 	ret = nouveau_mm_init(&vm->mm, mm_offset >> 12, mm_length >> 12,
 			      block >> 12);
 	if (ret) {
-		kfree(vm->pgt);
+		vfree(vm->pgt);
 		kfree(vm);
 		return ret;
 	}
@@ -405,24 +408,25 @@ nouveau_vm_link(struct nouveau_vm *vm, struct nouveau_gpuobj *pgd)
 
 	nouveau_gpuobj_ref(pgd, &vpgd->obj);
 
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&nv_subdev(vmm)->mutex);
 	for (i = vm->fpde; i <= vm->lpde; i++)
 		vmm->map_pgt(pgd, i, vm->pgt[i - vm->fpde].obj);
 	list_add(&vpgd->head, &vm->pgd_list);
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&nv_subdev(vmm)->mutex);
 	return 0;
 }
 
 static void
 nouveau_vm_unlink(struct nouveau_vm *vm, struct nouveau_gpuobj *mpgd)
 {
+	struct nouveau_vmmgr *vmm = vm->vmm;
 	struct nouveau_vm_pgd *vpgd, *tmp;
 	struct nouveau_gpuobj *pgd = NULL;
 
 	if (!mpgd)
 		return;
 
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&nv_subdev(vmm)->mutex);
 	list_for_each_entry_safe(vpgd, tmp, &vm->pgd_list, head) {
 		if (vpgd->obj == mpgd) {
 			pgd = vpgd->obj;
@@ -431,7 +435,7 @@ nouveau_vm_unlink(struct nouveau_vm *vm, struct nouveau_gpuobj *mpgd)
 			break;
 		}
 	}
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&nv_subdev(vmm)->mutex);
 
 	nouveau_gpuobj_ref(NULL, &pgd);
 }
@@ -446,7 +450,7 @@ nouveau_vm_del(struct nouveau_vm *vm)
 	}
 
 	nouveau_mm_fini(&vm->mm);
-	kfree(vm->pgt);
+	vfree(vm->pgt);
 	kfree(vm);
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/vm/nv50.c
index e067f81c97b3..07dd1fe2d6fb 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/nv50.c
@@ -27,11 +27,11 @@
 
 #include <subdev/timer.h>
 #include <subdev/fb.h>
+#include <subdev/bar.h>
 #include <subdev/vm.h>
 
 struct nv50_vmmgr_priv {
 	struct nouveau_vmmgr base;
-	spinlock_t lock;
 };
 
 static void
@@ -86,8 +86,8 @@ nv50_vm_map(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt,
 
 	/* IGPs don't have real VRAM, re-target to stolen system memory */
 	target = 0;
-	if (nouveau_fb(vma->vm->vmm)->ram.stolen) {
-		phys += nouveau_fb(vma->vm->vmm)->ram.stolen;
+	if (nouveau_fb(vma->vm->vmm)->ram->stolen) {
+		phys += nouveau_fb(vma->vm->vmm)->ram->stolen;
 		target = 3;
 	}
 
@@ -151,29 +151,42 @@ nv50_vm_unmap(struct nouveau_gpuobj *pgt, u32 pte, u32 cnt)
 static void
 nv50_vm_flush(struct nouveau_vm *vm)
 {
+	struct nv50_vmmgr_priv *priv = (void *)vm->vmm;
+	struct nouveau_bar *bar = nouveau_bar(priv);
 	struct nouveau_engine *engine;
-	int i;
+	int i, vme;
+
+	bar->flush(bar);
 
+	mutex_lock(&nv_subdev(priv)->mutex);
 	for (i = 0; i < NVDEV_SUBDEV_NR; i++) {
-		if (atomic_read(&vm->engref[i])) {
-			engine = nouveau_engine(vm->vmm, i);
-			if (engine && engine->tlb_flush)
-				engine->tlb_flush(engine);
+		if (!atomic_read(&vm->engref[i]))
+			continue;
+
+		/* unfortunate hw bug workaround... */
+		engine = nouveau_engine(priv, i);
+		if (engine && engine->tlb_flush) {
+			engine->tlb_flush(engine);
+			continue;
 		}
-	}
-}
 
-void
-nv50_vm_flush_engine(struct nouveau_subdev *subdev, int engine)
-{
-	struct nv50_vmmgr_priv *priv = (void *)nouveau_vmmgr(subdev);
-	unsigned long flags;
-
-	spin_lock_irqsave(&priv->lock, flags);
-	nv_wr32(subdev, 0x100c80, (engine << 16) | 1);
-	if (!nv_wait(subdev, 0x100c80, 0x00000001, 0x00000000))
-		nv_error(subdev, "vm flush timeout: engine %d\n", engine);
-	spin_unlock_irqrestore(&priv->lock, flags);
+		switch (i) {
+		case NVDEV_ENGINE_GR   : vme = 0x00; break;
+		case NVDEV_ENGINE_VP   : vme = 0x01; break;
+		case NVDEV_SUBDEV_BAR  : vme = 0x06; break;
+		case NVDEV_ENGINE_MPEG : vme = 0x08; break;
+		case NVDEV_ENGINE_BSP  : vme = 0x09; break;
+		case NVDEV_ENGINE_CRYPT: vme = 0x0a; break;
+		case NVDEV_ENGINE_COPY0: vme = 0x0d; break;
+		default:
+			continue;
+		}
+
+		nv_wr32(priv, 0x100c80, (vme << 16) | 1);
+		if (!nv_wait(priv, 0x100c80, 0x00000001, 0x00000000))
+			nv_error(priv, "vm flush timeout: engine %d\n", vme);
+	}
+	mutex_unlock(&nv_subdev(priv)->mutex);
 }
 
 static int
@@ -211,7 +224,6 @@ nv50_vmmgr_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->base.map_sg = nv50_vm_map_sg;
 	priv->base.unmap = nv50_vm_unmap;
 	priv->base.flush = nv50_vm_flush;
-	spin_lock_init(&priv->lock);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c b/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c
index 4c3b0a23b9d6..668cf964e4a9 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/vm/nvc0.c
@@ -29,10 +29,10 @@
 #include <subdev/fb.h>
 #include <subdev/vm.h>
 #include <subdev/ltcg.h>
+#include <subdev/bar.h>
 
 struct nvc0_vmmgr_priv {
 	struct nouveau_vmmgr base;
-	spinlock_t lock;
 };
 
 
@@ -160,40 +160,40 @@ nvc0_vm_unmap(struct nouveau_gpuobj *pgt, u32 pte, u32 cnt)
 	}
 }
 
-void
-nvc0_vm_flush_engine(struct nouveau_subdev *subdev, u64 addr, int type)
-{
-	struct nvc0_vmmgr_priv *priv = (void *)nouveau_vmmgr(subdev);
-	unsigned long flags;
-
-	/* looks like maybe a "free flush slots" counter, the
-	 * faster you write to 0x100cbc to more it decreases
-	 */
-	spin_lock_irqsave(&priv->lock, flags);
-	if (!nv_wait_ne(subdev, 0x100c80, 0x00ff0000, 0x00000000)) {
-		nv_error(subdev, "vm timeout 0: 0x%08x %d\n",
-			 nv_rd32(subdev, 0x100c80), type);
-	}
-
-	nv_wr32(subdev, 0x100cb8, addr >> 8);
-	nv_wr32(subdev, 0x100cbc, 0x80000000 | type);
-
-	/* wait for flush to be queued? */
-	if (!nv_wait(subdev, 0x100c80, 0x00008000, 0x00008000)) {
-		nv_error(subdev, "vm timeout 1: 0x%08x %d\n",
-			 nv_rd32(subdev, 0x100c80), type);
-	}
-	spin_unlock_irqrestore(&priv->lock, flags);
-}
-
 static void
 nvc0_vm_flush(struct nouveau_vm *vm)
 {
+	struct nvc0_vmmgr_priv *priv = (void *)vm->vmm;
+	struct nouveau_bar *bar = nouveau_bar(priv);
 	struct nouveau_vm_pgd *vpgd;
+	u32 type;
+
+	bar->flush(bar);
+
+	type = 0x00000001; /* PAGE_ALL */
+	if (atomic_read(&vm->engref[NVDEV_SUBDEV_BAR]))
+		type |= 0x00000004; /* HUB_ONLY */
 
+	mutex_lock(&nv_subdev(priv)->mutex);
 	list_for_each_entry(vpgd, &vm->pgd_list, head) {
-		nvc0_vm_flush_engine(nv_subdev(vm->vmm), vpgd->obj->addr, 1);
+		/* looks like maybe a "free flush slots" counter, the
+		 * faster you write to 0x100cbc to more it decreases
+		 */
+		if (!nv_wait_ne(priv, 0x100c80, 0x00ff0000, 0x00000000)) {
+			nv_error(priv, "vm timeout 0: 0x%08x %d\n",
+				 nv_rd32(priv, 0x100c80), type);
+		}
+
+		nv_wr32(priv, 0x100cb8, vpgd->obj->addr >> 8);
+		nv_wr32(priv, 0x100cbc, 0x80000000 | type);
+
+		/* wait for flush to be queued? */
+		if (!nv_wait(priv, 0x100c80, 0x00008000, 0x00008000)) {
+			nv_error(priv, "vm timeout 1: 0x%08x %d\n",
+				 nv_rd32(priv, 0x100c80), type);
+		}
 	}
+	mutex_unlock(&nv_subdev(priv)->mutex);
 }
 
 static int
@@ -227,7 +227,6 @@ nvc0_vmmgr_ctor(struct nouveau_object *parent, struct nouveau_object *engine,
 	priv->base.map_sg = nvc0_vm_map_sg;
 	priv->base.unmap = nvc0_vm_unmap;
 	priv->base.flush = nvc0_vm_flush;
-	spin_lock_init(&priv->lock);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_abi16.c b/drivers/gpu/drm/nouveau/nouveau_abi16.c
index 1c4c6c9161ac..8f467e7bfd19 100644
--- a/drivers/gpu/drm/nouveau/nouveau_abi16.c
+++ b/drivers/gpu/drm/nouveau/nouveau_abi16.c
@@ -129,6 +129,7 @@ nouveau_abi16_chan_fini(struct nouveau_abi16 *abi16,
 
 	if (chan->ntfy) {
 		nouveau_bo_vma_del(chan->ntfy, &chan->ntfy_vma);
+		nouveau_bo_unpin(chan->ntfy);
 		drm_gem_object_unreference_unlocked(chan->ntfy->gem);
 	}
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index 6aa2137e093a..3e7287675ecf 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -1878,9 +1878,6 @@ parse_dcb_table(struct drm_device *dev, struct nvbios *bios)
 	if (dcb->version < 0x21)
 		merge_like_dcb_entries(dev, dcb);
 
-	if (!dcb->entries)
-		return -ENXIO;
-
 	/* dump connector table entries to log, if any exist */
 	idx = -1;
 	while ((conn = olddcb_conn(dev, ++idx))) {
@@ -2054,19 +2051,14 @@ nouveau_bios_posted(struct drm_device *dev)
 	struct nouveau_drm *drm = nouveau_drm(dev);
 	unsigned htotal;
 
-	if (nv_device(drm->device)->card_type >= NV_50) {
-		if (NVReadVgaCrtc(dev, 0, 0x00) == 0 &&
-		    NVReadVgaCrtc(dev, 0, 0x1a) == 0)
-			return false;
+	if (nv_device(drm->device)->card_type >= NV_50)
 		return true;
-	}
 
 	htotal  = NVReadVgaCrtc(dev, 0, 0x06);
 	htotal |= (NVReadVgaCrtc(dev, 0, 0x07) & 0x01) << 8;
 	htotal |= (NVReadVgaCrtc(dev, 0, 0x07) & 0x20) << 4;
 	htotal |= (NVReadVgaCrtc(dev, 0, 0x25) & 0x01) << 10;
 	htotal |= (NVReadVgaCrtc(dev, 0, 0x41) & 0x01) << 11;
-
 	return (htotal != 0);
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index 7ff10711a4d0..4e7ee5f4155c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -148,6 +148,7 @@ nouveau_bo_del_ttm(struct ttm_buffer_object *bo)
 
 	if (unlikely(nvbo->gem))
 		DRM_ERROR("bo %p still attached to GEM object\n", bo);
+	WARN_ON(nvbo->pin_refcnt > 0);
 	nv10_bo_put_tile_region(dev, nvbo->tile, NULL);
 	kfree(nvbo);
 }
@@ -197,6 +198,12 @@ nouveau_bo_new(struct drm_device *dev, int size, int align,
 	size_t acc_size;
 	int ret;
 	int type = ttm_bo_type_device;
+	int max_size = INT_MAX & ~((1 << drm->client.base.vm->vmm->lpg_shift) - 1);
+
+	if (size <= 0 || size > max_size) {
+		nv_warn(drm, "skipped size %x\n", (u32)size);
+		return -EINVAL;
+	}
 
 	if (sg)
 		type = ttm_bo_type_sg;
@@ -255,7 +262,7 @@ set_placement_range(struct nouveau_bo *nvbo, uint32_t type)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	struct nouveau_fb *pfb = nouveau_fb(drm->device);
-	u32 vram_pages = pfb->ram.size >> PAGE_SHIFT;
+	u32 vram_pages = pfb->ram->size >> PAGE_SHIFT;
 
 	if (nv_device(drm->device)->card_type == NV_10 &&
 	    nvbo->tile_mode && (type & TTM_PL_FLAG_VRAM) &&
@@ -340,13 +347,15 @@ nouveau_bo_unpin(struct nouveau_bo *nvbo)
 {
 	struct nouveau_drm *drm = nouveau_bdev(nvbo->bo.bdev);
 	struct ttm_buffer_object *bo = &nvbo->bo;
-	int ret;
+	int ret, ref;
 
 	ret = ttm_bo_reserve(bo, false, false, false, 0);
 	if (ret)
 		return ret;
 
-	if (--nvbo->pin_refcnt)
+	ref = --nvbo->pin_refcnt;
+	WARN_ON_ONCE(ref < 0);
+	if (ref)
 		goto out;
 
 	nouveau_bo_placement_set(nvbo, bo->mem.placement, 0);
@@ -578,7 +587,7 @@ nve0_bo_move_init(struct nouveau_channel *chan, u32 handle)
 	int ret = RING_SPACE(chan, 2);
 	if (ret == 0) {
 		BEGIN_NVC0(chan, NvSubCopy, 0x0000, 1);
-		OUT_RING  (chan, handle);
+		OUT_RING  (chan, handle & 0x0000ffff);
 		FIRE_RING (chan);
 	}
 	return ret;
@@ -968,12 +977,12 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
 		     bool no_wait_gpu, struct ttm_mem_reg *new_mem)
 {
 	struct nouveau_drm *drm = nouveau_bdev(bo->bdev);
-	struct nouveau_channel *chan = chan = drm->channel;
+	struct nouveau_channel *chan = chan = drm->ttm.chan;
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
 	struct ttm_mem_reg *old_mem = &bo->mem;
 	int ret;
 
-	mutex_lock(&chan->cli->mutex);
+	mutex_lock_nested(&chan->cli->mutex, SINGLE_DEPTH_NESTING);
 
 	/* create temporary vmas for the transfer and attach them to the
 	 * old nouveau_mem node, these will get cleaned up after ttm has
@@ -1014,7 +1023,7 @@ nouveau_bo_move_init(struct nouveau_drm *drm)
 			    struct ttm_mem_reg *, struct ttm_mem_reg *);
 		int (*init)(struct nouveau_channel *, u32 handle);
 	} _methods[] = {
-		{  "COPY", 0, 0xa0b5, nve0_bo_move_copy, nve0_bo_move_init },
+		{  "COPY", 4, 0xa0b5, nve0_bo_move_copy, nve0_bo_move_init },
 		{  "GRCE", 0, 0xa0b5, nve0_bo_move_copy, nvc0_bo_move_init },
 		{ "COPY1", 5, 0x90b8, nvc0_bo_move_copy, nvc0_bo_move_init },
 		{ "COPY0", 4, 0x90b5, nvc0_bo_move_copy, nvc0_bo_move_init },
@@ -1034,7 +1043,7 @@ nouveau_bo_move_init(struct nouveau_drm *drm)
 		struct nouveau_channel *chan;
 		u32 handle = (mthd->engine << 16) | mthd->oclass;
 
-		if (mthd->init == nve0_bo_move_init)
+		if (mthd->engine)
 			chan = drm->cechan;
 		else
 			chan = drm->channel;
@@ -1052,6 +1061,7 @@ nouveau_bo_move_init(struct nouveau_drm *drm)
 			}
 
 			drm->ttm.move = mthd->exec;
+			drm->ttm.chan = chan;
 			name = mthd->name;
 			break;
 		}
@@ -1550,13 +1560,8 @@ void
 nouveau_bo_vma_del(struct nouveau_bo *nvbo, struct nouveau_vma *vma)
 {
 	if (vma->node) {
-		if (nvbo->bo.mem.mem_type != TTM_PL_SYSTEM) {
-			spin_lock(&nvbo->bo.bdev->fence_lock);
-			ttm_bo_wait(&nvbo->bo, false, false, false);
-			spin_unlock(&nvbo->bo.bdev->fence_lock);
+		if (nvbo->bo.mem.mem_type != TTM_PL_SYSTEM)
 			nouveau_vm_unmap(vma);
-		}
-
 		nouveau_vm_put(vma);
 		list_del(&vma->head);
 	}
diff --git a/drivers/gpu/drm/nouveau/nouveau_chan.c b/drivers/gpu/drm/nouveau/nouveau_chan.c
index eaa80a2b81ee..e84f4c32331b 100644
--- a/drivers/gpu/drm/nouveau/nouveau_chan.c
+++ b/drivers/gpu/drm/nouveau/nouveau_chan.c
@@ -147,7 +147,7 @@ nouveau_channel_prep(struct nouveau_drm *drm, struct nouveau_cli *cli,
 		args.limit = client->vm->vmm->limit - 1;
 	} else
 	if (chan->push.buffer->bo.mem.mem_type == TTM_PL_VRAM) {
-		u64 limit = pfb->ram.size - imem->reserved - 1;
+		u64 limit = pfb->ram->size - imem->reserved - 1;
 		if (device->card_type == NV_04) {
 			/* nv04 vram pushbuf hack, retarget to its location in
 			 * the framebuffer bar rather than direct vram access..
@@ -282,7 +282,7 @@ nouveau_channel_init(struct nouveau_channel *chan, u32 vram, u32 gart)
 		} else {
 			args.flags = NV_DMA_TARGET_VRAM | NV_DMA_ACCESS_RDWR;
 			args.start = 0;
-			args.limit = pfb->ram.size - imem->reserved - 1;
+			args.limit = pfb->ram->size - imem->reserved - 1;
 		}
 
 		ret = nouveau_object_new(nv_object(client), chan->handle, vram,
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index f17dc2ab03ec..907d20ef6d4d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -26,6 +26,7 @@
 
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/ttm/ttm_execbuf_util.h>
 
 #include "nouveau_fbcon.h"
 #include "dispnv04/hw.h"
@@ -137,7 +138,7 @@ nouveau_user_framebuffer_create(struct drm_device *dev,
 {
 	struct nouveau_framebuffer *nouveau_fb;
 	struct drm_gem_object *gem;
-	int ret;
+	int ret = -ENOMEM;
 
 	gem = drm_gem_object_lookup(dev, file_priv, mode_cmd->handles[0]);
 	if (!gem)
@@ -145,15 +146,19 @@ nouveau_user_framebuffer_create(struct drm_device *dev,
 
 	nouveau_fb = kzalloc(sizeof(struct nouveau_framebuffer), GFP_KERNEL);
 	if (!nouveau_fb)
-		return ERR_PTR(-ENOMEM);
+		goto err_unref;
 
 	ret = nouveau_framebuffer_init(dev, nouveau_fb, mode_cmd, nouveau_gem_object(gem));
-	if (ret) {
-		drm_gem_object_unreference(gem);
-		return ERR_PTR(ret);
-	}
+	if (ret)
+		goto err;
 
 	return &nouveau_fb->base;
+
+err:
+	kfree(nouveau_fb);
+err_unref:
+	drm_gem_object_unreference(gem);
+	return ERR_PTR(ret);
 }
 
 static const struct drm_mode_config_funcs nouveau_mode_config_funcs = {
@@ -332,10 +337,15 @@ nouveau_display_create(struct drm_device *dev)
 
 	if (nouveau_modeset == 1 ||
 	    (nouveau_modeset < 0 && pclass == PCI_CLASS_DISPLAY_VGA)) {
-		if (nv_device(drm->device)->card_type < NV_50)
-			ret = nv04_display_create(dev);
-		else
-			ret = nv50_display_create(dev);
+		if (drm->vbios.dcb.entries) {
+			if (nv_device(drm->device)->card_type < NV_50)
+				ret = nv04_display_create(dev);
+			else
+				ret = nv50_display_create(dev);
+		} else {
+			ret = 0;
+		}
+
 		if (ret)
 			goto disp_create_err;
 
@@ -457,51 +467,6 @@ nouveau_display_resume(struct drm_device *dev)
 }
 
 static int
-nouveau_page_flip_reserve(struct nouveau_bo *old_bo,
-			  struct nouveau_bo *new_bo)
-{
-	int ret;
-
-	ret = nouveau_bo_pin(new_bo, TTM_PL_FLAG_VRAM);
-	if (ret)
-		return ret;
-
-	ret = ttm_bo_reserve(&new_bo->bo, false, false, false, 0);
-	if (ret)
-		goto fail;
-
-	if (likely(old_bo != new_bo)) {
-		ret = ttm_bo_reserve(&old_bo->bo, false, false, false, 0);
-		if (ret)
-			goto fail_unreserve;
-	}
-
-	return 0;
-
-fail_unreserve:
-	ttm_bo_unreserve(&new_bo->bo);
-fail:
-	nouveau_bo_unpin(new_bo);
-	return ret;
-}
-
-static void
-nouveau_page_flip_unreserve(struct nouveau_bo *old_bo,
-			    struct nouveau_bo *new_bo,
-			    struct nouveau_fence *fence)
-{
-	nouveau_bo_fence(new_bo, fence);
-	ttm_bo_unreserve(&new_bo->bo);
-
-	if (likely(old_bo != new_bo)) {
-		nouveau_bo_fence(old_bo, fence);
-		ttm_bo_unreserve(&old_bo->bo);
-	}
-
-	nouveau_bo_unpin(old_bo);
-}
-
-static int
 nouveau_page_flip_emit(struct nouveau_channel *chan,
 		       struct nouveau_bo *old_bo,
 		       struct nouveau_bo *new_bo,
@@ -563,6 +528,12 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	struct nouveau_page_flip_state *s;
 	struct nouveau_channel *chan = NULL;
 	struct nouveau_fence *fence;
+	struct ttm_validate_buffer resv[2] = {
+		{ .bo = &old_bo->bo },
+		{ .bo = &new_bo->bo },
+	};
+	struct ww_acquire_ctx ticket;
+	LIST_HEAD(res);
 	int ret;
 
 	if (!drm->channel)
@@ -572,32 +543,40 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	if (!s)
 		return -ENOMEM;
 
-	/* Don't let the buffers go away while we flip */
-	ret = nouveau_page_flip_reserve(old_bo, new_bo);
-	if (ret)
-		goto fail_free;
-
-	/* Initialize a page flip struct */
-	*s = (struct nouveau_page_flip_state)
-		{ { }, event, nouveau_crtc(crtc)->index,
-		  fb->bits_per_pixel, fb->pitches[0], crtc->x, crtc->y,
-		  new_bo->bo.offset };
-
 	/* Choose the channel the flip will be handled in */
+	spin_lock(&old_bo->bo.bdev->fence_lock);
 	fence = new_bo->bo.sync_obj;
 	if (fence)
 		chan = fence->channel;
 	if (!chan)
 		chan = drm->channel;
+	spin_unlock(&old_bo->bo.bdev->fence_lock);
+
+	if (new_bo != old_bo) {
+		ret = nouveau_bo_pin(new_bo, TTM_PL_FLAG_VRAM);
+		if (ret)
+			goto fail_free;
+
+		list_add(&resv[1].head, &res);
+	}
+	list_add(&resv[0].head, &res);
+
 	mutex_lock(&chan->cli->mutex);
+	ret = ttm_eu_reserve_buffers(&ticket, &res);
+	if (ret)
+		goto fail_unpin;
+
+	/* Initialize a page flip struct */
+	*s = (struct nouveau_page_flip_state)
+		{ { }, event, nouveau_crtc(crtc)->index,
+		  fb->bits_per_pixel, fb->pitches[0], crtc->x, crtc->y,
+		  new_bo->bo.offset };
 
 	/* Emit a page flip */
 	if (nv_device(drm->device)->card_type >= NV_50) {
 		ret = nv50_display_flip_next(crtc, fb, chan, 0);
-		if (ret) {
-			mutex_unlock(&chan->cli->mutex);
+		if (ret)
 			goto fail_unreserve;
-		}
 	}
 
 	ret = nouveau_page_flip_emit(chan, old_bo, new_bo, s, &fence);
@@ -608,12 +587,18 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	/* Update the crtc struct and cleanup */
 	crtc->fb = fb;
 
-	nouveau_page_flip_unreserve(old_bo, new_bo, fence);
+	ttm_eu_fence_buffer_objects(&ticket, &res, fence);
+	if (old_bo != new_bo)
+		nouveau_bo_unpin(old_bo);
 	nouveau_fence_unref(&fence);
 	return 0;
 
 fail_unreserve:
-	nouveau_page_flip_unreserve(old_bo, new_bo, NULL);
+	ttm_eu_backoff_reservation(&ticket, &res);
+fail_unpin:
+	mutex_unlock(&chan->cli->mutex);
+	if (old_bo != new_bo)
+		nouveau_bo_unpin(new_bo);
 fail_free:
 	kfree(s);
 	return ret;
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index 383f4e6ea9d1..61972668fd05 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -192,6 +192,18 @@ nouveau_accel_init(struct nouveau_drm *drm)
 
 		arg0 = NVE0_CHANNEL_IND_ENGINE_GR;
 		arg1 = 1;
+	} else
+	if (device->chipset >= 0xa3 &&
+	    device->chipset != 0xaa &&
+	    device->chipset != 0xac) {
+		ret = nouveau_channel_new(drm, &drm->client, NVDRM_DEVICE,
+					  NVDRM_CHAN + 1, NvDmaFB, NvDmaTT,
+					  &drm->cechan);
+		if (ret)
+			NV_ERROR(drm, "failed to create ce channel, %d\n", ret);
+
+		arg0 = NvDmaFB;
+		arg1 = NvDmaTT;
 	} else {
 		arg0 = NvDmaFB;
 		arg1 = NvDmaTT;
@@ -284,8 +296,6 @@ static int nouveau_drm_probe(struct pci_dev *pdev,
 	return 0;
 }
 
-static struct lock_class_key drm_client_lock_class_key;
-
 static int
 nouveau_drm_load(struct drm_device *dev, unsigned long flags)
 {
@@ -297,7 +307,6 @@ nouveau_drm_load(struct drm_device *dev, unsigned long flags)
 	ret = nouveau_cli_create(pdev, "DRM", sizeof(*drm), (void**)&drm);
 	if (ret)
 		return ret;
-	lockdep_set_class(&drm->client.mutex, &drm_client_lock_class_key);
 
 	dev->dev_private = drm;
 	drm->dev = dev;
@@ -702,6 +711,7 @@ driver = {
 	.gem_prime_export = drm_gem_prime_export,
 	.gem_prime_import = drm_gem_prime_import,
 	.gem_prime_pin = nouveau_gem_prime_pin,
+	.gem_prime_unpin = nouveau_gem_prime_unpin,
 	.gem_prime_get_sg_table = nouveau_gem_prime_get_sg_table,
 	.gem_prime_import_sg_table = nouveau_gem_prime_import_sg_table,
 	.gem_prime_vmap = nouveau_gem_prime_vmap,
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.h b/drivers/gpu/drm/nouveau/nouveau_drm.h
index f2b30f89dee0..41ff7e0d403a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.h
@@ -96,6 +96,7 @@ struct nouveau_drm {
 		int (*move)(struct nouveau_channel *,
 			    struct ttm_buffer_object *,
 			    struct ttm_mem_reg *, struct ttm_mem_reg *);
+		struct nouveau_channel *chan;
 		int mtrr;
 	} ttm;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
index b03531781580..4c1bc061fae2 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
@@ -289,16 +289,13 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_VRAM);
 	if (ret) {
 		NV_ERROR(drm, "failed to pin fb: %d\n", ret);
-		nouveau_bo_ref(NULL, &nvbo);
-		goto out;
+		goto out_unref;
 	}
 
 	ret = nouveau_bo_map(nvbo);
 	if (ret) {
 		NV_ERROR(drm, "failed to map fb: %d\n", ret);
-		nouveau_bo_unpin(nvbo);
-		nouveau_bo_ref(NULL, &nvbo);
-		goto out;
+		goto out_unpin;
 	}
 
 	chan = nouveau_nofbaccel ? NULL : drm->channel;
@@ -316,13 +313,14 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 	info = framebuffer_alloc(0, &pdev->dev);
 	if (!info) {
 		ret = -ENOMEM;
-		goto out_unref;
+		goto out_unlock;
 	}
 
 	ret = fb_alloc_cmap(&info->cmap, 256, 0);
 	if (ret) {
 		ret = -ENOMEM;
-		goto out_unref;
+		framebuffer_release(info);
+		goto out_unlock;
 	}
 
 	info->par = fbcon;
@@ -337,7 +335,7 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 	fbcon->helper.fbdev = info;
 
 	strcpy(info->fix.id, "nouveaufb");
-	if (nouveau_nofbaccel)
+	if (!chan)
 		info->flags = FBINFO_DEFAULT | FBINFO_HWACCEL_DISABLED;
 	else
 		info->flags = FBINFO_DEFAULT | FBINFO_HWACCEL_COPYAREA |
@@ -383,8 +381,15 @@ nouveau_fbcon_create(struct drm_fb_helper *helper,
 	vga_switcheroo_client_fb_set(dev->pdev, info);
 	return 0;
 
-out_unref:
+out_unlock:
 	mutex_unlock(&dev->struct_mutex);
+	if (chan)
+		nouveau_bo_vma_del(nvbo, &fbcon->nouveau_fb.vma);
+	nouveau_bo_unmap(nvbo);
+out_unpin:
+	nouveau_bo_unpin(nvbo);
+out_unref:
+	nouveau_bo_ref(NULL, &nvbo);
 out:
 	return ret;
 }
@@ -413,6 +418,7 @@ nouveau_fbcon_destroy(struct drm_device *dev, struct nouveau_fbdev *fbcon)
 	if (nouveau_fb->nvbo) {
 		nouveau_bo_unmap(nouveau_fb->nvbo);
 		nouveau_bo_vma_del(nouveau_fb->nvbo, &nouveau_fb->vma);
+		nouveau_bo_unpin(nouveau_fb->nvbo);
 		drm_gem_object_unreference_unlocked(nouveau_fb->nvbo->gem);
 		nouveau_fb->nvbo = NULL;
 	}
@@ -467,10 +473,10 @@ nouveau_fbcon_init(struct drm_device *dev)
 
 	drm_fb_helper_single_add_all_connectors(&fbcon->helper);
 
-	if (pfb->ram.size <= 32 * 1024 * 1024)
+	if (pfb->ram->size <= 32 * 1024 * 1024)
 		preferred_bpp = 8;
 	else
-	if (pfb->ram.size <= 64 * 1024 * 1024)
+	if (pfb->ram->size <= 64 * 1024 * 1024)
 		preferred_bpp = 16;
 	else
 		preferred_bpp = 32;
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c
index 6c946837a0aa..be3149932c2d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.c
@@ -35,15 +35,34 @@
 
 #include <engine/fifo.h>
 
+struct fence_work {
+	struct work_struct base;
+	struct list_head head;
+	void (*func)(void *);
+	void *data;
+};
+
+static void
+nouveau_fence_signal(struct nouveau_fence *fence)
+{
+	struct fence_work *work, *temp;
+
+	list_for_each_entry_safe(work, temp, &fence->work, head) {
+		schedule_work(&work->base);
+		list_del(&work->head);
+	}
+
+	fence->channel = NULL;
+	list_del(&fence->head);
+}
+
 void
 nouveau_fence_context_del(struct nouveau_fence_chan *fctx)
 {
 	struct nouveau_fence *fence, *fnext;
 	spin_lock(&fctx->lock);
 	list_for_each_entry_safe(fence, fnext, &fctx->pending, head) {
-		fence->channel = NULL;
-		list_del(&fence->head);
-		nouveau_fence_unref(&fence);
+		nouveau_fence_signal(fence);
 	}
 	spin_unlock(&fctx->lock);
 }
@@ -57,6 +76,50 @@ nouveau_fence_context_new(struct nouveau_fence_chan *fctx)
 }
 
 static void
+nouveau_fence_work_handler(struct work_struct *kwork)
+{
+	struct fence_work *work = container_of(kwork, typeof(*work), base);
+	work->func(work->data);
+	kfree(work);
+}
+
+void
+nouveau_fence_work(struct nouveau_fence *fence,
+		   void (*func)(void *), void *data)
+{
+	struct nouveau_channel *chan = fence->channel;
+	struct nouveau_fence_chan *fctx;
+	struct fence_work *work = NULL;
+
+	if (nouveau_fence_done(fence)) {
+		func(data);
+		return;
+	}
+
+	fctx = chan->fence;
+	work = kmalloc(sizeof(*work), GFP_KERNEL);
+	if (!work) {
+		WARN_ON(nouveau_fence_wait(fence, false, false));
+		func(data);
+		return;
+	}
+
+	spin_lock(&fctx->lock);
+	if (!fence->channel) {
+		spin_unlock(&fctx->lock);
+		kfree(work);
+		func(data);
+		return;
+	}
+
+	INIT_WORK(&work->base, nouveau_fence_work_handler);
+	work->func = func;
+	work->data = data;
+	list_add(&work->head, &fence->work);
+	spin_unlock(&fctx->lock);
+}
+
+static void
 nouveau_fence_update(struct nouveau_channel *chan)
 {
 	struct nouveau_fence_chan *fctx = chan->fence;
@@ -67,8 +130,7 @@ nouveau_fence_update(struct nouveau_channel *chan)
 		if (fctx->read(chan) < fence->sequence)
 			break;
 
-		fence->channel = NULL;
-		list_del(&fence->head);
+		nouveau_fence_signal(fence);
 		nouveau_fence_unref(&fence);
 	}
 	spin_unlock(&fctx->lock);
@@ -81,7 +143,7 @@ nouveau_fence_emit(struct nouveau_fence *fence, struct nouveau_channel *chan)
 	int ret;
 
 	fence->channel  = chan;
-	fence->timeout  = jiffies + (3 * DRM_HZ);
+	fence->timeout  = jiffies + (15 * DRM_HZ);
 	fence->sequence = ++fctx->sequence;
 
 	ret = fctx->emit(fence);
@@ -265,6 +327,7 @@ nouveau_fence_new(struct nouveau_channel *chan, bool sysmem,
 	if (!fence)
 		return -ENOMEM;
 
+	INIT_LIST_HEAD(&fence->work);
 	fence->sysmem = sysmem;
 	kref_init(&fence->kref);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.h b/drivers/gpu/drm/nouveau/nouveau_fence.h
index c89943407b52..c57bb61da58c 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.h
@@ -5,6 +5,7 @@ struct nouveau_drm;
 
 struct nouveau_fence {
 	struct list_head head;
+	struct list_head work;
 	struct kref kref;
 
 	bool sysmem;
@@ -22,6 +23,7 @@ void nouveau_fence_unref(struct nouveau_fence **);
 
 int  nouveau_fence_emit(struct nouveau_fence *, struct nouveau_channel *);
 bool nouveau_fence_done(struct nouveau_fence *);
+void nouveau_fence_work(struct nouveau_fence *, void (*)(void *), void *);
 int  nouveau_fence_wait(struct nouveau_fence *, bool lazy, bool intr);
 int  nouveau_fence_sync(struct nouveau_fence *, struct nouveau_channel *);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index b4b4d0c1f4af..830cb7bad922 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -50,11 +50,6 @@ nouveau_gem_object_del(struct drm_gem_object *gem)
 		return;
 	nvbo->gem = NULL;
 
-	if (unlikely(nvbo->pin_refcnt)) {
-		nvbo->pin_refcnt = 1;
-		nouveau_bo_unpin(nvbo);
-	}
-
 	if (gem->import_attach)
 		drm_prime_gem_destroy(gem, nvbo->bo.sg);
 
@@ -101,6 +96,41 @@ out:
 	return ret;
 }
 
+static void
+nouveau_gem_object_delete(void *data)
+{
+	struct nouveau_vma *vma = data;
+	nouveau_vm_unmap(vma);
+	nouveau_vm_put(vma);
+	kfree(vma);
+}
+
+static void
+nouveau_gem_object_unmap(struct nouveau_bo *nvbo, struct nouveau_vma *vma)
+{
+	const bool mapped = nvbo->bo.mem.mem_type != TTM_PL_SYSTEM;
+	struct nouveau_fence *fence = NULL;
+
+	list_del(&vma->head);
+
+	if (mapped) {
+		spin_lock(&nvbo->bo.bdev->fence_lock);
+		if (nvbo->bo.sync_obj)
+			fence = nouveau_fence_ref(nvbo->bo.sync_obj);
+		spin_unlock(&nvbo->bo.bdev->fence_lock);
+	}
+
+	if (fence) {
+		nouveau_fence_work(fence, nouveau_gem_object_delete, vma);
+	} else {
+		if (mapped)
+			nouveau_vm_unmap(vma);
+		nouveau_vm_put(vma);
+		kfree(vma);
+	}
+	nouveau_fence_unref(&fence);
+}
+
 void
 nouveau_gem_object_close(struct drm_gem_object *gem, struct drm_file *file_priv)
 {
@@ -118,10 +148,8 @@ nouveau_gem_object_close(struct drm_gem_object *gem, struct drm_file *file_priv)
 
 	vma = nouveau_bo_vma_find(nvbo, cli->base.vm);
 	if (vma) {
-		if (--vma->refcount == 0) {
-			nouveau_bo_vma_del(nvbo, vma);
-			kfree(vma);
-		}
+		if (--vma->refcount == 0)
+			nouveau_gem_object_unmap(nvbo, vma);
 	}
 	ttm_bo_unreserve(&nvbo->bo);
 }
@@ -276,10 +304,12 @@ struct validate_op {
 	struct list_head vram_list;
 	struct list_head gart_list;
 	struct list_head both_list;
+	struct ww_acquire_ctx ticket;
 };
 
 static void
-validate_fini_list(struct list_head *list, struct nouveau_fence *fence)
+validate_fini_list(struct list_head *list, struct nouveau_fence *fence,
+		   struct ww_acquire_ctx *ticket)
 {
 	struct list_head *entry, *tmp;
 	struct nouveau_bo *nvbo;
@@ -296,17 +326,24 @@ validate_fini_list(struct list_head *list, struct nouveau_fence *fence)
 
 		list_del(&nvbo->entry);
 		nvbo->reserved_by = NULL;
-		ttm_bo_unreserve(&nvbo->bo);
+		ttm_bo_unreserve_ticket(&nvbo->bo, ticket);
 		drm_gem_object_unreference_unlocked(nvbo->gem);
 	}
 }
 
 static void
-validate_fini(struct validate_op *op, struct nouveau_fence* fence)
+validate_fini_no_ticket(struct validate_op *op, struct nouveau_fence *fence)
 {
-	validate_fini_list(&op->vram_list, fence);
-	validate_fini_list(&op->gart_list, fence);
-	validate_fini_list(&op->both_list, fence);
+	validate_fini_list(&op->vram_list, fence, &op->ticket);
+	validate_fini_list(&op->gart_list, fence, &op->ticket);
+	validate_fini_list(&op->both_list, fence, &op->ticket);
+}
+
+static void
+validate_fini(struct validate_op *op, struct nouveau_fence *fence)
+{
+	validate_fini_no_ticket(op, fence);
+	ww_acquire_fini(&op->ticket);
 }
 
 static int
@@ -316,13 +353,11 @@ validate_init(struct nouveau_channel *chan, struct drm_file *file_priv,
 {
 	struct nouveau_cli *cli = nouveau_cli(file_priv);
 	struct drm_device *dev = chan->drm->dev;
-	struct nouveau_drm *drm = nouveau_drm(dev);
-	uint32_t sequence;
 	int trycnt = 0;
 	int ret, i;
 	struct nouveau_bo *res_bo = NULL;
 
-	sequence = atomic_add_return(1, &drm->ttm.validate_sequence);
+	ww_acquire_init(&op->ticket, &reservation_ww_class);
 retry:
 	if (++trycnt > 100000) {
 		NV_ERROR(cli, "%s failed and gave up.\n", __func__);
@@ -337,6 +372,7 @@ retry:
 		gem = drm_gem_object_lookup(dev, file_priv, b->handle);
 		if (!gem) {
 			NV_ERROR(cli, "Unknown handle 0x%08x\n", b->handle);
+			ww_acquire_done(&op->ticket);
 			validate_fini(op, NULL);
 			return -ENOENT;
 		}
@@ -351,21 +387,23 @@ retry:
 			NV_ERROR(cli, "multiple instances of buffer %d on "
 				      "validation list\n", b->handle);
 			drm_gem_object_unreference_unlocked(gem);
+			ww_acquire_done(&op->ticket);
 			validate_fini(op, NULL);
 			return -EINVAL;
 		}
 
-		ret = ttm_bo_reserve(&nvbo->bo, true, false, true, sequence);
+		ret = ttm_bo_reserve(&nvbo->bo, true, false, true, &op->ticket);
 		if (ret) {
-			validate_fini(op, NULL);
-			if (unlikely(ret == -EAGAIN)) {
-				sequence = atomic_add_return(1, &drm->ttm.validate_sequence);
+			validate_fini_no_ticket(op, NULL);
+			if (unlikely(ret == -EDEADLK)) {
 				ret = ttm_bo_reserve_slowpath(&nvbo->bo, true,
-							      sequence);
+							      &op->ticket);
 				if (!ret)
 					res_bo = nvbo;
 			}
 			if (unlikely(ret)) {
+				ww_acquire_done(&op->ticket);
+				ww_acquire_fini(&op->ticket);
 				drm_gem_object_unreference_unlocked(gem);
 				if (ret != -ERESTARTSYS)
 					NV_ERROR(cli, "fail reserve\n");
@@ -389,6 +427,7 @@ retry:
 			NV_ERROR(cli, "invalid valid domains: 0x%08x\n",
 				 b->valid_domains);
 			list_add_tail(&nvbo->entry, &op->both_list);
+			ww_acquire_done(&op->ticket);
 			validate_fini(op, NULL);
 			return -EINVAL;
 		}
@@ -396,6 +435,7 @@ retry:
 			goto retry;
 	}
 
+	ww_acquire_done(&op->ticket);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.h b/drivers/gpu/drm/nouveau/nouveau_gem.h
index 8d7a3f0aeb86..502e4290aa8f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.h
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.h
@@ -36,6 +36,7 @@ extern int nouveau_gem_ioctl_info(struct drm_device *, void *,
 				  struct drm_file *);
 
 extern int nouveau_gem_prime_pin(struct drm_gem_object *);
+extern void nouveau_gem_prime_unpin(struct drm_gem_object *);
 extern struct sg_table *nouveau_gem_prime_get_sg_table(struct drm_gem_object *);
 extern struct drm_gem_object *nouveau_gem_prime_import_sg_table(
 	struct drm_device *, size_t size, struct sg_table *);
diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c
index 7e0ff10a2759..4f6a572f2258 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_mem.c
@@ -125,7 +125,7 @@ nv50_mem_timing_calc(struct drm_device *dev, u32 freq,
 		t->reg[7] = 0x4000202 | (e->tCL - 1) << 16;
 
 		/* XXX: P.version == 1 only has DDR2 and GDDR3? */
-		if (pfb->ram.type == NV_MEM_TYPE_DDR2) {
+		if (pfb->ram->type == NV_MEM_TYPE_DDR2) {
 			t->reg[5] |= (e->tCL + 3) << 8;
 			t->reg[6] |= (t->tCWL - 2) << 8;
 			t->reg[8] |= (e->tCL - 4);
@@ -428,7 +428,7 @@ nouveau_mem_timing_calc(struct drm_device *dev, u32 freq,
 		break;
 	}
 
-	switch (pfb->ram.type * !ret) {
+	switch (pfb->ram->type * !ret) {
 	case NV_MEM_TYPE_GDDR3:
 		ret = nouveau_mem_gddr3_mr(dev, freq, e, len, boot, t);
 		break;
@@ -455,7 +455,7 @@ nouveau_mem_timing_calc(struct drm_device *dev, u32 freq,
 		else
 			dll_off = !!(ramcfg[2] & 0x40);
 
-		switch (pfb->ram.type) {
+		switch (pfb->ram->type) {
 		case NV_MEM_TYPE_GDDR3:
 			t->mr[1] &= ~0x00000040;
 			t->mr[1] |=  0x00000040 * dll_off;
@@ -522,7 +522,7 @@ nouveau_mem_timing_read(struct drm_device *dev, struct nouveau_pm_memtiming *t)
 	t->odt = 0;
 	t->drive_strength = 0;
 
-	switch (pfb->ram.type) {
+	switch (pfb->ram->type) {
 	case NV_MEM_TYPE_DDR3:
 		t->odt |= (t->mr[1] & 0x200) >> 7;
 	case NV_MEM_TYPE_DDR2:
@@ -551,7 +551,7 @@ nouveau_mem_exec(struct nouveau_mem_exec_func *exec,
 	u32 mr[3] = { info->mr[0], info->mr[1], info->mr[2] };
 	u32 mr1_dlloff;
 
-	switch (pfb->ram.type) {
+	switch (pfb->ram->type) {
 	case NV_MEM_TYPE_DDR2:
 		tDLLK = 2000;
 		mr1_dlloff = 0x00000001;
@@ -572,7 +572,7 @@ nouveau_mem_exec(struct nouveau_mem_exec_func *exec,
 	}
 
 	/* fetch current MRs */
-	switch (pfb->ram.type) {
+	switch (pfb->ram->type) {
 	case NV_MEM_TYPE_GDDR3:
 	case NV_MEM_TYPE_DDR3:
 		mr[2] = exec->mrg(exec, 2);
@@ -639,7 +639,7 @@ nouveau_mem_exec(struct nouveau_mem_exec_func *exec,
 		exec->mrs (exec, 0, info->mr[0] | 0x00000000);
 		exec->wait(exec, tMRD);
 		exec->wait(exec, tDLLK);
-		if (pfb->ram.type == NV_MEM_TYPE_GDDR3)
+		if (pfb->ram->type == NV_MEM_TYPE_GDDR3)
 			exec->precharge(exec);
 	}
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_prime.c b/drivers/gpu/drm/nouveau/nouveau_prime.c
index f53e10874cae..e90468d5e5c0 100644
--- a/drivers/gpu/drm/nouveau/nouveau_prime.c
+++ b/drivers/gpu/drm/nouveau/nouveau_prime.c
@@ -84,7 +84,7 @@ struct drm_gem_object *nouveau_gem_prime_import_sg_table(struct drm_device *dev,
 int nouveau_gem_prime_pin(struct drm_gem_object *obj)
 {
 	struct nouveau_bo *nvbo = nouveau_gem_object(obj);
-	int ret = 0;
+	int ret;
 
 	/* pin buffer into GTT */
 	ret = nouveau_bo_pin(nvbo, TTM_PL_FLAG_TT);
@@ -93,3 +93,10 @@ int nouveau_gem_prime_pin(struct drm_gem_object *obj)
 
 	return 0;
 }
+
+void nouveau_gem_prime_unpin(struct drm_gem_object *obj)
+{
+	struct nouveau_bo *nvbo = nouveau_gem_object(obj);
+
+	nouveau_bo_unpin(nvbo);
+}
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c
index f19a15a3bc03..19e3757291fb 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c
@@ -69,7 +69,7 @@ nouveau_vram_manager_del(struct ttm_mem_type_manager *man,
 	struct nouveau_drm *drm = nouveau_bdev(man->bdev);
 	struct nouveau_fb *pfb = nouveau_fb(drm->device);
 	nouveau_mem_node_cleanup(mem->mm_node);
-	pfb->ram.put(pfb, (struct nouveau_mem **)&mem->mm_node);
+	pfb->ram->put(pfb, (struct nouveau_mem **)&mem->mm_node);
 }
 
 static int
@@ -88,7 +88,7 @@ nouveau_vram_manager_new(struct ttm_mem_type_manager *man,
 	if (nvbo->tile_flags & NOUVEAU_GEM_TILE_NONCONTIG)
 		size_nc = 1 << nvbo->page_shift;
 
-	ret = pfb->ram.get(pfb, mem->num_pages << PAGE_SHIFT,
+	ret = pfb->ram->get(pfb, mem->num_pages << PAGE_SHIFT,
 			   mem->page_alignment << PAGE_SHIFT, size_nc,
 			   (nvbo->tile_flags >> 8) & 0x3ff, &node);
 	if (ret) {
@@ -111,7 +111,7 @@ nouveau_vram_manager_debug(struct ttm_mem_type_manager *man, const char *prefix)
 	struct nouveau_mm_node *r;
 	u32 total = 0, free = 0;
 
-	mutex_lock(&mm->mutex);
+	mutex_lock(&nv_subdev(pfb)->mutex);
 	list_for_each_entry(r, &mm->nodes, nl_entry) {
 		printk(KERN_DEBUG "%s %d: 0x%010llx 0x%010llx\n",
 		       prefix, r->type, ((u64)r->offset << 12),
@@ -121,7 +121,7 @@ nouveau_vram_manager_debug(struct ttm_mem_type_manager *man, const char *prefix)
 		if (!r->type)
 			free += r->length;
 	}
-	mutex_unlock(&mm->mutex);
+	mutex_unlock(&nv_subdev(pfb)->mutex);
 
 	printk(KERN_DEBUG "%s  total: 0x%010llx free: 0x%010llx\n",
 	       prefix, (u64)total << 12, (u64)free << 12);
@@ -168,9 +168,6 @@ nouveau_gart_manager_new(struct ttm_mem_type_manager *man,
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
 	struct nouveau_mem *node;
 
-	if (unlikely((mem->num_pages << PAGE_SHIFT) >= 512 * 1024 * 1024))
-		return -ENOMEM;
-
 	node = kzalloc(sizeof(*node), GFP_KERNEL);
 	if (!node)
 		return -ENOMEM;
@@ -386,7 +383,7 @@ nouveau_ttm_init(struct nouveau_drm *drm)
 	}
 
 	/* VRAM init */
-	drm->gem.vram_available  = nouveau_fb(drm->device)->ram.size;
+	drm->gem.vram_available  = nouveau_fb(drm->device)->ram->size;
 	drm->gem.vram_available -= nouveau_instmem(drm->device)->reserved;
 
 	ret = ttm_bo_init_mm(&drm->ttm.bdev, TTM_PL_VRAM,
@@ -396,15 +393,12 @@ nouveau_ttm_init(struct nouveau_drm *drm)
 		return ret;
 	}
 
-	drm->ttm.mtrr = drm_mtrr_add(pci_resource_start(dev->pdev, 1),
-				     pci_resource_len(dev->pdev, 1),
-				     DRM_MTRR_WC);
+	drm->ttm.mtrr = arch_phys_wc_add(pci_resource_start(dev->pdev, 1),
+					 pci_resource_len(dev->pdev, 1));
 
 	/* GART init */
 	if (drm->agp.stat != ENABLED) {
 		drm->gem.gart_available = nouveau_vmmgr(drm->device)->limit;
-		if (drm->gem.gart_available > 512 * 1024 * 1024)
-			drm->gem.gart_available = 512 * 1024 * 1024;
 	} else {
 		drm->gem.gart_available = drm->agp.size;
 	}
@@ -433,10 +427,6 @@ nouveau_ttm_fini(struct nouveau_drm *drm)
 
 	nouveau_ttm_global_release(drm);
 
-	if (drm->ttm.mtrr >= 0) {
-		drm_mtrr_del(drm->ttm.mtrr,
-			     pci_resource_start(drm->dev->pdev, 1),
-			     pci_resource_len(drm->dev->pdev, 1), DRM_MTRR_WC);
-		drm->ttm.mtrr = -1;
-	}
+	arch_phys_wc_del(drm->ttm.mtrr);
+	drm->ttm.mtrr = 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index dd5e01f89f28..8b40a36c1b57 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -159,7 +159,7 @@ nv50_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NV50_DMA_CONF0_ENABLE |
 					         NV50_DMA_CONF0_PART_256,
 				     }, sizeof(struct nv_dma_class), &object);
@@ -172,7 +172,7 @@ nv50_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NV50_DMA_CONF0_ENABLE | 0x70 |
 					         NV50_DMA_CONF0_PART_256,
 				 }, sizeof(struct nv_dma_class), &object);
@@ -185,7 +185,7 @@ nv50_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NV50_DMA_CONF0_ENABLE | 0x7a |
 					         NV50_DMA_CONF0_PART_256,
 				 }, sizeof(struct nv_dma_class), &object);
@@ -204,7 +204,7 @@ nvc0_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NVC0_DMA_CONF0_ENABLE,
 				     }, sizeof(struct nv_dma_class), &object);
 	if (ret)
@@ -216,7 +216,7 @@ nvc0_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NVC0_DMA_CONF0_ENABLE | 0xfe,
 				 }, sizeof(struct nv_dma_class), &object);
 	if (ret)
@@ -228,7 +228,7 @@ nvc0_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NVC0_DMA_CONF0_ENABLE | 0xfe,
 				 }, sizeof(struct nv_dma_class), &object);
 	return ret;
@@ -246,7 +246,7 @@ nvd0_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NVD0_DMA_CONF0_ENABLE |
 						 NVD0_DMA_CONF0_PAGE_LP,
 				     }, sizeof(struct nv_dma_class), &object);
@@ -259,7 +259,7 @@ nvd0_dmac_create_fbdma(struct nouveau_object *core, u32 parent)
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 					.conf0 = NVD0_DMA_CONF0_ENABLE | 0xfe |
 						 NVD0_DMA_CONF0_PAGE_LP,
 				 }, sizeof(struct nv_dma_class), &object);
@@ -316,7 +316,7 @@ nv50_dmac_create(struct nouveau_object *core, u32 bclass, u8 head,
 					.flags = NV_DMA_TARGET_VRAM |
 						 NV_DMA_ACCESS_RDWR,
 					.start = 0,
-					.limit = pfb->ram.size - 1,
+					.limit = pfb->ram->size - 1,
 				 }, sizeof(struct nv_dma_class), &object);
 	if (ret)
 		return ret;
@@ -355,6 +355,7 @@ struct nv50_oimm {
 
 struct nv50_head {
 	struct nouveau_crtc base;
+	struct nouveau_bo *image;
 	struct nv50_curs curs;
 	struct nv50_sync sync;
 	struct nv50_ovly ovly;
@@ -517,9 +518,10 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 {
 	struct nouveau_framebuffer *nv_fb = nouveau_framebuffer(fb);
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
+	struct nv50_head *head = nv50_head(crtc);
 	struct nv50_sync *sync = nv50_sync(crtc);
-	int head = nv_crtc->index, ret;
 	u32 *push;
+	int ret;
 
 	swap_interval <<= 4;
 	if (swap_interval == 0)
@@ -537,7 +539,7 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			return ret;
 
 		BEGIN_NV04(chan, 0, NV11_SUBCHAN_DMA_SEMAPHORE, 2);
-		OUT_RING  (chan, NvEvoSema0 + head);
+		OUT_RING  (chan, NvEvoSema0 + nv_crtc->index);
 		OUT_RING  (chan, sync->addr ^ 0x10);
 		BEGIN_NV04(chan, 0, NV11_SUBCHAN_SEMAPHORE_RELEASE, 1);
 		OUT_RING  (chan, sync->data + 1);
@@ -546,7 +548,7 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		OUT_RING  (chan, sync->data);
 	} else
 	if (chan && nv_mclass(chan->object) < NVC0_CHANNEL_IND_CLASS) {
-		u64 addr = nv84_fence_crtc(chan, head) + sync->addr;
+		u64 addr = nv84_fence_crtc(chan, nv_crtc->index) + sync->addr;
 		ret = RING_SPACE(chan, 12);
 		if (ret)
 			return ret;
@@ -565,7 +567,7 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		OUT_RING  (chan, NV84_SUBCHAN_SEMAPHORE_TRIGGER_ACQUIRE_EQUAL);
 	} else
 	if (chan) {
-		u64 addr = nv84_fence_crtc(chan, head) + sync->addr;
+		u64 addr = nv84_fence_crtc(chan, nv_crtc->index) + sync->addr;
 		ret = RING_SPACE(chan, 10);
 		if (ret)
 			return ret;
@@ -630,6 +632,8 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	evo_mthd(push, 0x0080, 1);
 	evo_data(push, 0x00000000);
 	evo_kick(push, sync);
+
+	nouveau_bo_ref(nv_fb->nvbo, &head->image);
 	return 0;
 }
 
@@ -1038,18 +1042,17 @@ static int
 nv50_crtc_swap_fbs(struct drm_crtc *crtc, struct drm_framebuffer *old_fb)
 {
 	struct nouveau_framebuffer *nvfb = nouveau_framebuffer(crtc->fb);
+	struct nv50_head *head = nv50_head(crtc);
 	int ret;
 
 	ret = nouveau_bo_pin(nvfb->nvbo, TTM_PL_FLAG_VRAM);
-	if (ret)
-		return ret;
-
-	if (old_fb) {
-		nvfb = nouveau_framebuffer(old_fb);
-		nouveau_bo_unpin(nvfb->nvbo);
+	if (ret == 0) {
+		if (head->image)
+			nouveau_bo_unpin(head->image);
+		nouveau_bo_ref(nvfb->nvbo, &head->image);
 	}
 
-	return 0;
+	return ret;
 }
 
 static int
@@ -1198,6 +1201,15 @@ nv50_crtc_lut_load(struct drm_crtc *crtc)
 	}
 }
 
+static void
+nv50_crtc_disable(struct drm_crtc *crtc)
+{
+	struct nv50_head *head = nv50_head(crtc);
+	if (head->image)
+		nouveau_bo_unpin(head->image);
+	nouveau_bo_ref(NULL, &head->image);
+}
+
 static int
 nv50_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 		     uint32_t handle, uint32_t width, uint32_t height)
@@ -1271,18 +1283,29 @@ nv50_crtc_destroy(struct drm_crtc *crtc)
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
 	struct nv50_disp *disp = nv50_disp(crtc->dev);
 	struct nv50_head *head = nv50_head(crtc);
+
 	nv50_dmac_destroy(disp->core, &head->ovly.base);
 	nv50_pioc_destroy(disp->core, &head->oimm.base);
 	nv50_dmac_destroy(disp->core, &head->sync.base);
 	nv50_pioc_destroy(disp->core, &head->curs.base);
+
+	/*XXX: this shouldn't be necessary, but the core doesn't call
+	 *     disconnect() during the cleanup paths
+	 */
+	if (head->image)
+		nouveau_bo_unpin(head->image);
+	nouveau_bo_ref(NULL, &head->image);
+
 	nouveau_bo_unmap(nv_crtc->cursor.nvbo);
 	if (nv_crtc->cursor.nvbo)
 		nouveau_bo_unpin(nv_crtc->cursor.nvbo);
 	nouveau_bo_ref(NULL, &nv_crtc->cursor.nvbo);
+
 	nouveau_bo_unmap(nv_crtc->lut.nvbo);
 	if (nv_crtc->lut.nvbo)
 		nouveau_bo_unpin(nv_crtc->lut.nvbo);
 	nouveau_bo_ref(NULL, &nv_crtc->lut.nvbo);
+
 	drm_crtc_cleanup(crtc);
 	kfree(crtc);
 }
@@ -1296,6 +1319,7 @@ static const struct drm_crtc_helper_funcs nv50_crtc_hfunc = {
 	.mode_set_base = nv50_crtc_mode_set_base,
 	.mode_set_base_atomic = nv50_crtc_mode_set_base_atomic,
 	.load_lut = nv50_crtc_lut_load,
+	.disable = nv50_crtc_disable,
 };
 
 static const struct drm_crtc_funcs nv50_crtc_func = {
diff --git a/drivers/gpu/drm/nouveau/nv50_pm.c b/drivers/gpu/drm/nouveau/nv50_pm.c
index 69620e39c90c..4efc33fa73fc 100644
--- a/drivers/gpu/drm/nouveau/nv50_pm.c
+++ b/drivers/gpu/drm/nouveau/nv50_pm.c
@@ -493,12 +493,12 @@ mclk_mrs(struct nouveau_mem_exec_func *exec, int mr, u32 data)
 	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
 
 	if (mr <= 1) {
-		if (pfb->ram.ranks > 1)
+		if (pfb->ram->ranks > 1)
 			hwsq_wr32(hwsq, 0x1002c8 + ((mr - 0) * 4), data);
 		hwsq_wr32(hwsq, 0x1002c0 + ((mr - 0) * 4), data);
 	} else
 	if (mr <= 3) {
-		if (pfb->ram.ranks > 1)
+		if (pfb->ram->ranks > 1)
 			hwsq_wr32(hwsq, 0x1002e8 + ((mr - 2) * 4), data);
 		hwsq_wr32(hwsq, 0x1002e0 + ((mr - 2) * 4), data);
 	}
diff --git a/drivers/gpu/drm/nouveau/nva3_pm.c b/drivers/gpu/drm/nouveau/nva3_pm.c
index 863f010fafeb..0d0ed597fea8 100644
--- a/drivers/gpu/drm/nouveau/nva3_pm.c
+++ b/drivers/gpu/drm/nouveau/nva3_pm.c
@@ -389,12 +389,12 @@ mclk_mrs(struct nouveau_mem_exec_func *exec, int mr, u32 data)
 	struct nouveau_device *device = nouveau_dev(exec->dev);
 	struct nouveau_fb *pfb = nouveau_fb(device);
 	if (mr <= 1) {
-		if (pfb->ram.ranks > 1)
+		if (pfb->ram->ranks > 1)
 			nv_wr32(device, 0x1002c8 + ((mr - 0) * 4), data);
 		nv_wr32(device, 0x1002c0 + ((mr - 0) * 4), data);
 	} else
 	if (mr <= 3) {
-		if (pfb->ram.ranks > 1)
+		if (pfb->ram->ranks > 1)
 			nv_wr32(device, 0x1002e8 + ((mr - 2) * 4), data);
 		nv_wr32(device, 0x1002e0 + ((mr - 2) * 4), data);
 	}
diff --git a/drivers/gpu/drm/nouveau/nvc0_pm.c b/drivers/gpu/drm/nouveau/nvc0_pm.c
index 0d34eb581179..3b7041cb013f 100644
--- a/drivers/gpu/drm/nouveau/nvc0_pm.c
+++ b/drivers/gpu/drm/nouveau/nvc0_pm.c
@@ -477,7 +477,7 @@ mclk_mrg(struct nouveau_mem_exec_func *exec, int mr)
 {
 	struct nouveau_device *device = nouveau_dev(exec->dev);
 	struct nouveau_fb *pfb = nouveau_fb(device);
-	if (pfb->ram.type != NV_MEM_TYPE_GDDR5) {
+	if (pfb->ram->type != NV_MEM_TYPE_GDDR5) {
 		if (mr <= 1)
 			return nv_rd32(device, 0x10f300 + ((mr - 0) * 4));
 		return nv_rd32(device, 0x10f320 + ((mr - 2) * 4));
@@ -496,15 +496,15 @@ mclk_mrs(struct nouveau_mem_exec_func *exec, int mr, u32 data)
 {
 	struct nouveau_device *device = nouveau_dev(exec->dev);
 	struct nouveau_fb *pfb = nouveau_fb(device);
-	if (pfb->ram.type != NV_MEM_TYPE_GDDR5) {
+	if (pfb->ram->type != NV_MEM_TYPE_GDDR5) {
 		if (mr <= 1) {
 			nv_wr32(device, 0x10f300 + ((mr - 0) * 4), data);
-			if (pfb->ram.ranks > 1)
+			if (pfb->ram->ranks > 1)
 				nv_wr32(device, 0x10f308 + ((mr - 0) * 4), data);
 		} else
 		if (mr <= 3) {
 			nv_wr32(device, 0x10f320 + ((mr - 2) * 4), data);
-			if (pfb->ram.ranks > 1)
+			if (pfb->ram->ranks > 1)
 				nv_wr32(device, 0x10f328 + ((mr - 2) * 4), data);
 		}
 	} else {
diff --git a/drivers/gpu/drm/omapdrm/Kconfig b/drivers/gpu/drm/omapdrm/Kconfig
index 09f65dc3d2c8..20c41e73d448 100644
--- a/drivers/gpu/drm/omapdrm/Kconfig
+++ b/drivers/gpu/drm/omapdrm/Kconfig
@@ -1,7 +1,7 @@
 
 config DRM_OMAP
 	tristate "OMAP DRM"
-	depends on DRM && !CONFIG_FB_OMAP2
+	depends on DRM
 	depends on ARCH_OMAP2PLUS || ARCH_MULTIPLATFORM
 	depends on OMAP2_DSS
 	select DRM_KMS_HELPER
diff --git a/drivers/gpu/drm/omapdrm/omap_crtc.c b/drivers/gpu/drm/omapdrm/omap_crtc.c
index 79b200aee18a..11a5263a5e9f 100644
--- a/drivers/gpu/drm/omapdrm/omap_crtc.c
+++ b/drivers/gpu/drm/omapdrm/omap_crtc.c
@@ -40,7 +40,7 @@ struct omap_crtc {
 	 * mgr->id.)  Eventually this will be replaced w/ something
 	 * more common-panel-framework-y
 	 */
-	struct omap_overlay_manager mgr;
+	struct omap_overlay_manager *mgr;
 
 	struct omap_video_timings timings;
 	bool enabled;
@@ -90,7 +90,32 @@ uint32_t pipe2vbl(struct drm_crtc *crtc)
  * job of sequencing the setup of the video pipe in the proper order
  */
 
+/* ovl-mgr-id -> crtc */
+static struct omap_crtc *omap_crtcs[8];
+
 /* we can probably ignore these until we support command-mode panels: */
+static int omap_crtc_connect(struct omap_overlay_manager *mgr,
+		struct omap_dss_device *dst)
+{
+	if (mgr->output)
+		return -EINVAL;
+
+	if ((mgr->supported_outputs & dst->id) == 0)
+		return -EINVAL;
+
+	dst->manager = mgr;
+	mgr->output = dst;
+
+	return 0;
+}
+
+static void omap_crtc_disconnect(struct omap_overlay_manager *mgr,
+		struct omap_dss_device *dst)
+{
+	mgr->output->manager = NULL;
+	mgr->output = NULL;
+}
+
 static void omap_crtc_start_update(struct omap_overlay_manager *mgr)
 {
 }
@@ -107,7 +132,7 @@ static void omap_crtc_disable(struct omap_overlay_manager *mgr)
 static void omap_crtc_set_timings(struct omap_overlay_manager *mgr,
 		const struct omap_video_timings *timings)
 {
-	struct omap_crtc *omap_crtc = container_of(mgr, struct omap_crtc, mgr);
+	struct omap_crtc *omap_crtc = omap_crtcs[mgr->id];
 	DBG("%s", omap_crtc->name);
 	omap_crtc->timings = *timings;
 	omap_crtc->full_update = true;
@@ -116,7 +141,7 @@ static void omap_crtc_set_timings(struct omap_overlay_manager *mgr,
 static void omap_crtc_set_lcd_config(struct omap_overlay_manager *mgr,
 		const struct dss_lcd_mgr_config *config)
 {
-	struct omap_crtc *omap_crtc = container_of(mgr, struct omap_crtc, mgr);
+	struct omap_crtc *omap_crtc = omap_crtcs[mgr->id];
 	DBG("%s", omap_crtc->name);
 	dispc_mgr_set_lcd_config(omap_crtc->channel, config);
 }
@@ -135,6 +160,8 @@ static void omap_crtc_unregister_framedone_handler(
 }
 
 static const struct dss_mgr_ops mgr_ops = {
+		.connect = omap_crtc_connect,
+		.disconnect = omap_crtc_disconnect,
 		.start_update = omap_crtc_start_update,
 		.enable = omap_crtc_enable,
 		.disable = omap_crtc_disable,
@@ -253,10 +280,6 @@ static int omap_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 			NULL, NULL);
 }
 
-static void omap_crtc_load_lut(struct drm_crtc *crtc)
-{
-}
-
 static void vblank_cb(void *arg)
 {
 	struct drm_crtc *crtc = arg;
@@ -366,7 +389,6 @@ static const struct drm_crtc_helper_funcs omap_crtc_helper_funcs = {
 	.prepare = omap_crtc_prepare,
 	.commit = omap_crtc_commit,
 	.mode_set_base = omap_crtc_mode_set_base,
-	.load_lut = omap_crtc_load_lut,
 };
 
 const struct omap_video_timings *omap_crtc_timings(struct drm_crtc *crtc)
@@ -569,7 +591,7 @@ static void omap_crtc_pre_apply(struct omap_drm_apply *apply)
 	} else {
 		if (encoder) {
 			omap_encoder_set_enabled(encoder, false);
-			omap_encoder_update(encoder, &omap_crtc->mgr,
+			omap_encoder_update(encoder, omap_crtc->mgr,
 					&omap_crtc->timings);
 			omap_encoder_set_enabled(encoder, true);
 			omap_crtc->full_update = false;
@@ -595,6 +617,11 @@ static const char *channel_names[] = {
 		[OMAP_DSS_CHANNEL_LCD2] = "lcd2",
 };
 
+void omap_crtc_pre_init(void)
+{
+	dss_install_mgr_ops(&mgr_ops);
+}
+
 /* initialize crtc */
 struct drm_crtc *omap_crtc_init(struct drm_device *dev,
 		struct drm_plane *plane, enum omap_channel channel, int id)
@@ -635,9 +662,7 @@ struct drm_crtc *omap_crtc_init(struct drm_device *dev,
 	omap_irq_register(dev, &omap_crtc->error_irq);
 
 	/* temporary: */
-	omap_crtc->mgr.id = channel;
-
-	dss_install_mgr_ops(&mgr_ops);
+	omap_crtc->mgr = omap_dss_get_overlay_manager(channel);
 
 	/* TODO: fix hard-coded setup.. add properties! */
 	info = &omap_crtc->info;
@@ -651,6 +676,8 @@ struct drm_crtc *omap_crtc_init(struct drm_device *dev,
 
 	omap_plane_install_properties(omap_crtc->plane, &crtc->base);
 
+	omap_crtcs[channel] = omap_crtc;
+
 	return crtc;
 
 fail:
diff --git a/drivers/gpu/drm/omapdrm/omap_drv.c b/drivers/gpu/drm/omapdrm/omap_drv.c
index 826586ffbe83..a3004f12b9a3 100644
--- a/drivers/gpu/drm/omapdrm/omap_drv.c
+++ b/drivers/gpu/drm/omapdrm/omap_drv.c
@@ -65,10 +65,8 @@ static int get_connector_type(struct omap_dss_device *dssdev)
 	switch (dssdev->type) {
 	case OMAP_DISPLAY_TYPE_HDMI:
 		return DRM_MODE_CONNECTOR_HDMIA;
-	case OMAP_DISPLAY_TYPE_DPI:
-		if (!strcmp(dssdev->name, "dvi"))
-			return DRM_MODE_CONNECTOR_DVID;
-		/* fallthrough */
+	case OMAP_DISPLAY_TYPE_DVI:
+		return DRM_MODE_CONNECTOR_DVID;
 	default:
 		return DRM_MODE_CONNECTOR_Unknown;
 	}
@@ -97,6 +95,9 @@ static int omap_modeset_init(struct drm_device *dev)
 	int num_mgrs = dss_feat_get_num_mgrs();
 	int num_crtcs;
 	int i, id = 0;
+	int r;
+
+	omap_crtc_pre_init();
 
 	drm_mode_config_init(dev);
 
@@ -116,6 +117,7 @@ static int omap_modeset_init(struct drm_device *dev)
 		struct drm_connector *connector;
 		struct drm_encoder *encoder;
 		enum omap_channel channel;
+		struct omap_overlay_manager *mgr;
 
 		if (!dssdev->driver) {
 			dev_warn(dev->dev, "%s has no driver.. skipping it\n",
@@ -131,6 +133,13 @@ static int omap_modeset_init(struct drm_device *dev)
 			continue;
 		}
 
+		r = dssdev->driver->connect(dssdev);
+		if (r) {
+			dev_err(dev->dev, "could not connect display: %s\n",
+					dssdev->name);
+			continue;
+		}
+
 		encoder = omap_encoder_init(dev, dssdev);
 
 		if (!encoder) {
@@ -172,8 +181,9 @@ static int omap_modeset_init(struct drm_device *dev)
 		 * other possible channels to which the encoder can connect are
 		 * not considered.
 		 */
-		channel = dssdev->output->dispc_channel;
 
+		mgr = omapdss_find_mgr_from_display(dssdev);
+		channel = mgr->id;
 		/*
 		 * if this channel hasn't already been taken by a previously
 		 * allocated crtc, we create a new crtc for it
@@ -247,6 +257,9 @@ static int omap_modeset_init(struct drm_device *dev)
 		struct drm_encoder *encoder = priv->encoders[i];
 		struct omap_dss_device *dssdev =
 					omap_encoder_get_dssdev(encoder);
+		struct omap_dss_device *output;
+
+		output = omapdss_find_output_from_display(dssdev);
 
 		/* figure out which crtc's we can connect the encoder to: */
 		encoder->possible_crtcs = 0;
@@ -259,9 +272,11 @@ static int omap_modeset_init(struct drm_device *dev)
 			supported_outputs =
 				dss_feat_get_supported_outputs(crtc_channel);
 
-			if (supported_outputs & dssdev->output->id)
+			if (supported_outputs & output->id)
 				encoder->possible_crtcs |= (1 << id);
 		}
+
+		omap_dss_put_device(output);
 	}
 
 	DBG("registered %d planes, %d crtcs, %d encoders and %d connectors\n",
diff --git a/drivers/gpu/drm/omapdrm/omap_drv.h b/drivers/gpu/drm/omapdrm/omap_drv.h
index 215a20dd340c..14f17da2ce25 100644
--- a/drivers/gpu/drm/omapdrm/omap_drv.h
+++ b/drivers/gpu/drm/omapdrm/omap_drv.h
@@ -157,6 +157,7 @@ const struct omap_video_timings *omap_crtc_timings(struct drm_crtc *crtc);
 enum omap_channel omap_crtc_channel(struct drm_crtc *crtc);
 int omap_crtc_apply(struct drm_crtc *crtc,
 		struct omap_drm_apply *apply);
+void omap_crtc_pre_init(void);
 struct drm_crtc *omap_crtc_init(struct drm_device *dev,
 		struct drm_plane *plane, enum omap_channel channel, int id);
 
diff --git a/drivers/gpu/drm/omapdrm/omap_fbdev.c b/drivers/gpu/drm/omapdrm/omap_fbdev.c
index b11ce609fcc2..002988d09021 100644
--- a/drivers/gpu/drm/omapdrm/omap_fbdev.c
+++ b/drivers/gpu/drm/omapdrm/omap_fbdev.c
@@ -281,21 +281,7 @@ fail:
 	return ret;
 }
 
-static void omap_crtc_fb_gamma_set(struct drm_crtc *crtc,
-		u16 red, u16 green, u16 blue, int regno)
-{
-	DBG("fbdev: set gamma");
-}
-
-static void omap_crtc_fb_gamma_get(struct drm_crtc *crtc,
-		u16 *red, u16 *green, u16 *blue, int regno)
-{
-	DBG("fbdev: get gamma");
-}
-
 static struct drm_fb_helper_funcs omap_fb_helper_funcs = {
-	.gamma_set = omap_crtc_fb_gamma_set,
-	.gamma_get = omap_crtc_fb_gamma_get,
 	.fb_probe = omap_fbdev_create,
 };
 
diff --git a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c
index be7cd97a0db0..4fcca8d42796 100644
--- a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c
+++ b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c
@@ -136,44 +136,21 @@ static void omap_gem_dmabuf_kunmap(struct dma_buf *buffer,
 	kunmap(pages[page_num]);
 }
 
-/*
- * TODO maybe we can split up drm_gem_mmap to avoid duplicating
- * some here.. or at least have a drm_dmabuf_mmap helper.
- */
 static int omap_gem_dmabuf_mmap(struct dma_buf *buffer,
 		struct vm_area_struct *vma)
 {
 	struct drm_gem_object *obj = buffer->priv;
+	struct drm_device *dev = obj->dev;
 	int ret = 0;
 
 	if (WARN_ON(!obj->filp))
 		return -EINVAL;
 
-	/* Check for valid size. */
-	if (omap_gem_mmap_size(obj) < vma->vm_end - vma->vm_start) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	if (!obj->dev->driver->gem_vm_ops) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP;
-	vma->vm_ops = obj->dev->driver->gem_vm_ops;
-	vma->vm_private_data = obj;
-	vma->vm_page_prot =  pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
-
-	/* Take a ref for this mapping of the object, so that the fault
-	 * handler can dereference the mmap offset's pointer to the object.
-	 * This reference is cleaned up by the corresponding vm_close
-	 * (which should happen whether the vma was created by this call, or
-	 * by a vm_open due to mremap or partial unmap or whatever).
-	 */
-	vma->vm_ops->open(vma);
-
-out_unlock:
+	mutex_lock(&dev->struct_mutex);
+	ret = drm_gem_mmap_obj(obj, omap_gem_mmap_size(obj), vma);
+	mutex_unlock(&dev->struct_mutex);
+	if (ret < 0)
+		return ret;
 
 	return omap_gem_mmap_obj(obj, vma);
 }
diff --git a/drivers/gpu/drm/qxl/qxl_cmd.c b/drivers/gpu/drm/qxl/qxl_cmd.c
index f86771481317..93c2f2cceb51 100644
--- a/drivers/gpu/drm/qxl/qxl_cmd.c
+++ b/drivers/gpu/drm/qxl/qxl_cmd.c
@@ -49,6 +49,11 @@ void qxl_ring_free(struct qxl_ring *ring)
 	kfree(ring);
 }
 
+void qxl_ring_init_hdr(struct qxl_ring *ring)
+{
+	ring->ring->header.notify_on_prod = ring->n_elements;
+}
+
 struct qxl_ring *
 qxl_ring_create(struct qxl_ring_header *header,
 		int element_size,
@@ -69,7 +74,7 @@ qxl_ring_create(struct qxl_ring_header *header,
 	ring->prod_notify = prod_notify;
 	ring->push_event = push_event;
 	if (set_prod_notify)
-		header->notify_on_prod = ring->n_elements;
+		qxl_ring_init_hdr(ring);
 	spin_lock_init(&ring->lock);
 	return ring;
 }
@@ -87,7 +92,7 @@ static int qxl_check_header(struct qxl_ring *ring)
 	return ret;
 }
 
-static int qxl_check_idle(struct qxl_ring *ring)
+int qxl_check_idle(struct qxl_ring *ring)
 {
 	int ret;
 	struct qxl_ring_header *header = &(ring->ring->header);
@@ -375,8 +380,8 @@ void qxl_io_destroy_primary(struct qxl_device *qdev)
 	wait_for_io_cmd(qdev, 0, QXL_IO_DESTROY_PRIMARY_ASYNC);
 }
 
-void qxl_io_create_primary(struct qxl_device *qdev, unsigned width,
-			   unsigned height, unsigned offset, struct qxl_bo *bo)
+void qxl_io_create_primary(struct qxl_device *qdev,
+			   unsigned offset, struct qxl_bo *bo)
 {
 	struct qxl_surface_create *create;
 
@@ -384,8 +389,8 @@ void qxl_io_create_primary(struct qxl_device *qdev, unsigned width,
 		 qdev->ram_header);
 	create = &qdev->ram_header->create_surface;
 	create->format = bo->surf.format;
-	create->width = width;
-	create->height = height;
+	create->width = bo->surf.width;
+	create->height = bo->surf.height;
 	create->stride = bo->surf.stride;
 	create->mem = qxl_bo_physical_address(qdev, bo, offset);
 
diff --git a/drivers/gpu/drm/qxl/qxl_display.c b/drivers/gpu/drm/qxl/qxl_display.c
index 823d29e926ec..f76f5dd7bfc4 100644
--- a/drivers/gpu/drm/qxl/qxl_display.c
+++ b/drivers/gpu/drm/qxl/qxl_display.c
@@ -30,53 +30,9 @@
 #include "qxl_object.h"
 #include "drm_crtc_helper.h"
 
-static void qxl_crtc_set_to_mode(struct qxl_device *qdev,
-				 struct drm_connector *connector,
-				 struct qxl_head *head)
+static bool qxl_head_enabled(struct qxl_head *head)
 {
-	struct drm_device *dev = connector->dev;
-	struct drm_display_mode *mode, *t;
-	int width = head->width;
-	int height = head->height;
-
-	if (width < 320 || height < 240) {
-		qxl_io_log(qdev, "%s: bad head: %dx%d", width, height);
-		width = 1024;
-		height = 768;
-	}
-	if (width * height * 4 > 16*1024*1024) {
-		width = 1024;
-		height = 768;
-	}
-	/* TODO: go over regular modes and removed preferred? */
-	list_for_each_entry_safe(mode, t, &connector->probed_modes, head)
-		drm_mode_remove(connector, mode);
-	mode = drm_cvt_mode(dev, width, height, 60, false, false, false);
-	mode->type |= DRM_MODE_TYPE_PREFERRED;
-	mode->status = MODE_OK;
-	drm_mode_probed_add(connector, mode);
-	qxl_io_log(qdev, "%s: %d x %d\n", __func__, width, height);
-}
-
-void qxl_crtc_set_from_monitors_config(struct qxl_device *qdev)
-{
-	struct drm_connector *connector;
-	int i;
-	struct drm_device *dev = qdev->ddev;
-
-	i = 0;
-	qxl_io_log(qdev, "%s: %d, %d\n", __func__,
-		   dev->mode_config.num_connector,
-		   qdev->monitors_config->count);
-	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
-		if (i > qdev->monitors_config->count) {
-			/* crtc will be reported as disabled */
-			continue;
-		}
-		qxl_crtc_set_to_mode(qdev, connector,
-				     &qdev->monitors_config->heads[i]);
-		++i;
-	}
+	return head->width && head->height;
 }
 
 void qxl_alloc_client_monitors_config(struct qxl_device *qdev, unsigned count)
@@ -106,7 +62,6 @@ static int qxl_display_copy_rom_client_monitors_config(struct qxl_device *qdev)
 	int num_monitors;
 	uint32_t crc;
 
-	BUG_ON(!qdev->monitors_config);
 	num_monitors = qdev->rom->client_monitors_config.count;
 	crc = crc32(0, (const uint8_t *)&qdev->rom->client_monitors_config,
 		  sizeof(qdev->rom->client_monitors_config));
@@ -117,8 +72,8 @@ static int qxl_display_copy_rom_client_monitors_config(struct qxl_device *qdev)
 		return 1;
 	}
 	if (num_monitors > qdev->monitors_config->max_allowed) {
-		DRM_INFO("client monitors list will be truncated: %d < %d\n",
-			 qdev->monitors_config->max_allowed, num_monitors);
+		DRM_DEBUG_KMS("client monitors list will be truncated: %d < %d\n",
+			      qdev->monitors_config->max_allowed, num_monitors);
 		num_monitors = qdev->monitors_config->max_allowed;
 	} else {
 		num_monitors = qdev->rom->client_monitors_config.count;
@@ -132,18 +87,15 @@ static int qxl_display_copy_rom_client_monitors_config(struct qxl_device *qdev)
 			&qdev->rom->client_monitors_config.heads[i];
 		struct qxl_head *client_head =
 			&qdev->client_monitors_config->heads[i];
-		struct qxl_head *head = &qdev->monitors_config->heads[i];
-		client_head->x = head->x = c_rect->left;
-		client_head->y = head->y = c_rect->top;
-		client_head->width = head->width =
-						c_rect->right - c_rect->left;
-		client_head->height = head->height =
-						c_rect->bottom - c_rect->top;
-		client_head->surface_id = head->surface_id = 0;
-		client_head->id = head->id = i;
-		client_head->flags = head->flags = 0;
-		QXL_DEBUG(qdev, "read %dx%d+%d+%d\n", head->width, head->height,
-			  head->x, head->y);
+		client_head->x = c_rect->left;
+		client_head->y = c_rect->top;
+		client_head->width = c_rect->right - c_rect->left;
+		client_head->height = c_rect->bottom - c_rect->top;
+		client_head->surface_id = 0;
+		client_head->id = i;
+		client_head->flags = 0;
+		DRM_DEBUG_KMS("read %dx%d+%d+%d\n", client_head->width, client_head->height,
+			  client_head->x, client_head->y);
 	}
 	return 0;
 }
@@ -155,10 +107,7 @@ void qxl_display_read_client_monitors_config(struct qxl_device *qdev)
 		qxl_io_log(qdev, "failed crc check for client_monitors_config,"
 				 " retrying\n");
 	}
-	qxl_crtc_set_from_monitors_config(qdev);
-	/* fire off a uevent and let userspace tell us what to do */
-	qxl_io_log(qdev, "calling drm_sysfs_hotplug_event\n");
-	drm_sysfs_hotplug_event(qdev->ddev);
+	drm_helper_hpd_irq_event(qdev->ddev);
 }
 
 static int qxl_add_monitors_config_modes(struct drm_connector *connector)
@@ -170,9 +119,9 @@ static int qxl_add_monitors_config_modes(struct drm_connector *connector)
 	struct drm_display_mode *mode = NULL;
 	struct qxl_head *head;
 
-	if (!qdev->monitors_config)
+	if (!qdev->client_monitors_config)
 		return 0;
-	head = &qdev->monitors_config->heads[h];
+	head = &qdev->client_monitors_config->heads[h];
 
 	mode = drm_cvt_mode(dev, head->width, head->height, 60, false, false,
 			    false);
@@ -222,12 +171,6 @@ static int qxl_add_common_modes(struct drm_connector *connector)
 	return i - 1;
 }
 
-static void qxl_crtc_gamma_set(struct drm_crtc *crtc, u16 *red, u16 *green,
-			       u16 *blue, uint32_t start, uint32_t size)
-{
-	/* TODO */
-}
-
 static void qxl_crtc_destroy(struct drm_crtc *crtc)
 {
 	struct qxl_crtc *qxl_crtc = to_qxl_crtc(crtc);
@@ -255,11 +198,11 @@ qxl_hide_cursor(struct qxl_device *qdev)
 	qxl_release_unreserve(qdev, release);
 }
 
-static int qxl_crtc_cursor_set(struct drm_crtc *crtc,
-			       struct drm_file *file_priv,
-			       uint32_t handle,
-			       uint32_t width,
-			       uint32_t height)
+static int qxl_crtc_cursor_set2(struct drm_crtc *crtc,
+				struct drm_file *file_priv,
+				uint32_t handle,
+				uint32_t width,
+				uint32_t height, int32_t hot_x, int32_t hot_y)
 {
 	struct drm_device *dev = crtc->dev;
 	struct qxl_device *qdev = dev->dev_private;
@@ -315,8 +258,8 @@ static int qxl_crtc_cursor_set(struct drm_crtc *crtc,
 	cursor->header.type = SPICE_CURSOR_TYPE_ALPHA;
 	cursor->header.width = 64;
 	cursor->header.height = 64;
-	cursor->header.hot_spot_x = 0;
-	cursor->header.hot_spot_y = 0;
+	cursor->header.hot_spot_x = hot_x;
+	cursor->header.hot_spot_y = hot_y;
 	cursor->data_size = size;
 	cursor->chunk.next_chunk = 0;
 	cursor->chunk.prev_chunk = 0;
@@ -397,9 +340,8 @@ static int qxl_crtc_cursor_move(struct drm_crtc *crtc,
 
 
 static const struct drm_crtc_funcs qxl_crtc_funcs = {
-	.cursor_set = qxl_crtc_cursor_set,
+	.cursor_set2 = qxl_crtc_cursor_set2,
 	.cursor_move = qxl_crtc_cursor_move,
-	.gamma_set = qxl_crtc_gamma_set,
 	.set_config = drm_crtc_helper_set_config,
 	.destroy = qxl_crtc_destroy,
 };
@@ -506,7 +448,7 @@ qxl_send_monitors_config(struct qxl_device *qdev)
 	for (i = 0 ; i < qdev->monitors_config->count ; ++i) {
 		struct qxl_head *head = &qdev->monitors_config->heads[i];
 
-		if (head->y > 8192 || head->y < head->x ||
+		if (head->y > 8192 || head->x > 8192 ||
 		    head->width > 8192 || head->height > 8192) {
 			DRM_ERROR("head %d wrong: %dx%d+%d+%d\n",
 				  i, head->width, head->height,
@@ -517,16 +459,19 @@ qxl_send_monitors_config(struct qxl_device *qdev)
 	qxl_io_monitors_config(qdev);
 }
 
-static void qxl_monitors_config_set_single(struct qxl_device *qdev,
-					   unsigned x, unsigned y,
-					   unsigned width, unsigned height)
+static void qxl_monitors_config_set(struct qxl_device *qdev,
+				    int index,
+				    unsigned x, unsigned y,
+				    unsigned width, unsigned height,
+				    unsigned surf_id)
 {
-	DRM_DEBUG("%dx%d+%d+%d\n", width, height, x, y);
-	qdev->monitors_config->count = 1;
-	qdev->monitors_config->heads[0].x = x;
-	qdev->monitors_config->heads[0].y = y;
-	qdev->monitors_config->heads[0].width = width;
-	qdev->monitors_config->heads[0].height = height;
+	DRM_DEBUG_KMS("%d:%dx%d+%d+%d\n", index, width, height, x, y);
+	qdev->monitors_config->heads[index].x = x;
+	qdev->monitors_config->heads[index].y = y;
+	qdev->monitors_config->heads[index].width = width;
+	qdev->monitors_config->heads[index].height = height;
+	qdev->monitors_config->heads[index].surface_id = surf_id;
+
 }
 
 static int qxl_crtc_mode_set(struct drm_crtc *crtc,
@@ -540,10 +485,11 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 	struct qxl_mode *m = (void *)mode->private;
 	struct qxl_framebuffer *qfb;
 	struct qxl_bo *bo, *old_bo = NULL;
+	struct qxl_crtc *qcrtc = to_qxl_crtc(crtc);
 	uint32_t width, height, base_offset;
 	bool recreate_primary = false;
 	int ret;
-
+	int surf_id;
 	if (!crtc->fb) {
 		DRM_DEBUG_KMS("No FB bound\n");
 		return 0;
@@ -567,7 +513,8 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 		  adjusted_mode->hdisplay,
 		  adjusted_mode->vdisplay);
 
-	recreate_primary = true;
+	if (qcrtc->index == 0)
+		recreate_primary = true;
 
 	width = mode->hdisplay;
 	height = mode->vdisplay;
@@ -588,8 +535,11 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 			   "recreate primary: %dx%d (was %dx%d,%d,%d)\n",
 			   width, height, bo->surf.width,
 			   bo->surf.height, bo->surf.stride, bo->surf.format);
-		qxl_io_create_primary(qdev, width, height, base_offset, bo);
+		qxl_io_create_primary(qdev, base_offset, bo);
 		bo->is_primary = true;
+		surf_id = 0;
+	} else {
+		surf_id = bo->surface_id;
 	}
 
 	if (old_bo && old_bo != bo) {
@@ -599,11 +549,9 @@ static int qxl_crtc_mode_set(struct drm_crtc *crtc,
 		qxl_bo_unreserve(old_bo);
 	}
 
-	if (qdev->monitors_config->count == 0) {
-		qxl_monitors_config_set_single(qdev, x, y,
-					       mode->hdisplay,
-					       mode->vdisplay);
-	}
+	qxl_monitors_config_set(qdev, qcrtc->index, x, y,
+				mode->hdisplay,
+				mode->vdisplay, surf_id);
 	return 0;
 }
 
@@ -619,21 +567,36 @@ static void qxl_crtc_commit(struct drm_crtc *crtc)
 	DRM_DEBUG("\n");
 }
 
-static void qxl_crtc_load_lut(struct drm_crtc *crtc)
+static void qxl_crtc_disable(struct drm_crtc *crtc)
 {
-	DRM_DEBUG("\n");
+	struct qxl_crtc *qcrtc = to_qxl_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct qxl_device *qdev = dev->dev_private;
+	if (crtc->fb) {
+		struct qxl_framebuffer *qfb = to_qxl_framebuffer(crtc->fb);
+		struct qxl_bo *bo = gem_to_qxl_bo(qfb->obj);
+		int ret;
+		ret = qxl_bo_reserve(bo, false);
+		qxl_bo_unpin(bo);
+		qxl_bo_unreserve(bo);
+		crtc->fb = NULL;
+	}
+
+	qxl_monitors_config_set(qdev, qcrtc->index, 0, 0, 0, 0, 0);
+
+	qxl_send_monitors_config(qdev);
 }
 
 static const struct drm_crtc_helper_funcs qxl_crtc_helper_funcs = {
 	.dpms = qxl_crtc_dpms,
+	.disable = qxl_crtc_disable,
 	.mode_fixup = qxl_crtc_mode_fixup,
 	.mode_set = qxl_crtc_mode_set,
 	.prepare = qxl_crtc_prepare,
 	.commit = qxl_crtc_commit,
-	.load_lut = qxl_crtc_load_lut,
 };
 
-static int qdev_crtc_init(struct drm_device *dev, int num_crtc)
+static int qdev_crtc_init(struct drm_device *dev, int crtc_id)
 {
 	struct qxl_crtc *qxl_crtc;
 
@@ -642,7 +605,7 @@ static int qdev_crtc_init(struct drm_device *dev, int num_crtc)
 		return -ENOMEM;
 
 	drm_crtc_init(dev, &qxl_crtc->base, &qxl_crtc_funcs);
-
+	qxl_crtc->index = crtc_id;
 	drm_mode_crtc_set_gamma_size(&qxl_crtc->base, 256);
 	drm_crtc_helper_add(&qxl_crtc->base, &qxl_crtc_helper_funcs);
 	return 0;
@@ -670,18 +633,13 @@ static void qxl_write_monitors_config_for_encoder(struct qxl_device *qdev,
 		struct drm_encoder *encoder)
 {
 	int i;
+	struct qxl_output *output = drm_encoder_to_qxl_output(encoder);
 	struct qxl_head *head;
 	struct drm_display_mode *mode;
 
 	BUG_ON(!encoder);
 	/* TODO: ugly, do better */
-	for (i = 0 ; (encoder->possible_crtcs != (1 << i)) && i < 32; ++i)
-		;
-	if (encoder->possible_crtcs != (1 << i)) {
-		DRM_ERROR("encoder has wrong possible_crtcs: %x\n",
-			  encoder->possible_crtcs);
-		return;
-	}
+	i = output->index;
 	if (!qdev->monitors_config ||
 	    qdev->monitors_config->max_allowed <= i) {
 		DRM_ERROR(
@@ -699,7 +657,6 @@ static void qxl_write_monitors_config_for_encoder(struct qxl_device *qdev,
 		DRM_DEBUG("missing for multiple monitors: no head holes\n");
 	head = &qdev->monitors_config->heads[i];
 	head->id = i;
-	head->surface_id = 0;
 	if (encoder->crtc->enabled) {
 		mode = &encoder->crtc->mode;
 		head->width = mode->hdisplay;
@@ -714,8 +671,8 @@ static void qxl_write_monitors_config_for_encoder(struct qxl_device *qdev,
 		head->x = 0;
 		head->y = 0;
 	}
-	DRM_DEBUG("setting head %d to +%d+%d %dx%d\n",
-		  i, head->x, head->y, head->width, head->height);
+	DRM_DEBUG_KMS("setting head %d to +%d+%d %dx%d out of %d\n",
+		      i, head->x, head->y, head->width, head->height, qdev->monitors_config->count);
 	head->flags = 0;
 	/* TODO - somewhere else to call this for multiple monitors
 	 * (config_commit?) */
@@ -810,8 +767,9 @@ static enum drm_connector_status qxl_conn_detect(
 
 	/* The first monitor is always connected */
 	connected = (output->index == 0) ||
-		    (qdev->monitors_config &&
-		     qdev->monitors_config->count > output->index);
+		    (qdev->client_monitors_config &&
+		     qdev->client_monitors_config->count > output->index &&
+		     qxl_head_enabled(&qdev->client_monitors_config->heads[output->index]));
 
 	DRM_DEBUG("\n");
 	return connected ? connector_status_connected
@@ -875,6 +833,8 @@ static int qdev_output_init(struct drm_device *dev, int num_output)
 	drm_encoder_init(dev, &qxl_output->enc, &qxl_enc_funcs,
 			 DRM_MODE_ENCODER_VIRTUAL);
 
+	/* we get HPD via client monitors config */
+	connector->polled = DRM_CONNECTOR_POLL_HPD;
 	encoder->possible_crtcs = 1 << num_output;
 	drm_mode_connector_attach_encoder(&qxl_output->base,
 					  &qxl_output->enc);
@@ -914,16 +874,14 @@ static const struct drm_mode_config_funcs qxl_mode_funcs = {
 	.fb_create = qxl_user_framebuffer_create,
 };
 
-int qxl_modeset_init(struct qxl_device *qdev)
+int qxl_create_monitors_object(struct qxl_device *qdev)
 {
-	int i;
 	int ret;
 	struct drm_gem_object *gobj;
-	int max_allowed = QXL_NUM_OUTPUTS;
+	int max_allowed = qxl_num_crtc;
 	int monitors_config_size = sizeof(struct qxl_monitors_config) +
-				   max_allowed * sizeof(struct qxl_head);
+		max_allowed * sizeof(struct qxl_head);
 
-	drm_mode_config_init(qdev->ddev);
 	ret = qxl_gem_object_create(qdev, monitors_config_size, 0,
 				    QXL_GEM_DOMAIN_VRAM,
 				    false, false, NULL, &gobj);
@@ -932,13 +890,59 @@ int qxl_modeset_init(struct qxl_device *qdev)
 		return -ENOMEM;
 	}
 	qdev->monitors_config_bo = gem_to_qxl_bo(gobj);
+
+	ret = qxl_bo_reserve(qdev->monitors_config_bo, false);
+	if (ret)
+		return ret;
+
+	ret = qxl_bo_pin(qdev->monitors_config_bo, QXL_GEM_DOMAIN_VRAM, NULL);
+	if (ret) {
+		qxl_bo_unreserve(qdev->monitors_config_bo);
+		return ret;
+	}
+
+	qxl_bo_unreserve(qdev->monitors_config_bo);
+
 	qxl_bo_kmap(qdev->monitors_config_bo, NULL);
+
 	qdev->monitors_config = qdev->monitors_config_bo->kptr;
 	qdev->ram_header->monitors_config =
 		qxl_bo_physical_address(qdev, qdev->monitors_config_bo, 0);
 
 	memset(qdev->monitors_config, 0, monitors_config_size);
 	qdev->monitors_config->max_allowed = max_allowed;
+	return 0;
+}
+
+int qxl_destroy_monitors_object(struct qxl_device *qdev)
+{
+	int ret;
+
+	qdev->monitors_config = NULL;
+	qdev->ram_header->monitors_config = 0;
+
+	qxl_bo_kunmap(qdev->monitors_config_bo);
+	ret = qxl_bo_reserve(qdev->monitors_config_bo, false);
+	if (ret)
+		return ret;
+
+	qxl_bo_unpin(qdev->monitors_config_bo);
+	qxl_bo_unreserve(qdev->monitors_config_bo);
+
+	qxl_bo_unref(&qdev->monitors_config_bo);
+	return 0;
+}
+
+int qxl_modeset_init(struct qxl_device *qdev)
+{
+	int i;
+	int ret;
+
+	drm_mode_config_init(qdev->ddev);
+
+	ret = qxl_create_monitors_object(qdev);
+	if (ret)
+		return ret;
 
 	qdev->ddev->mode_config.funcs = (void *)&qxl_mode_funcs;
 
@@ -949,7 +953,7 @@ int qxl_modeset_init(struct qxl_device *qdev)
 	qdev->ddev->mode_config.max_height = 8192;
 
 	qdev->ddev->mode_config.fb_base = qdev->vram_base;
-	for (i = 0 ; i < QXL_NUM_OUTPUTS; ++i) {
+	for (i = 0 ; i < qxl_num_crtc; ++i) {
 		qdev_crtc_init(qdev->ddev, i);
 		qdev_output_init(qdev->ddev, i);
 	}
@@ -966,6 +970,8 @@ int qxl_modeset_init(struct qxl_device *qdev)
 void qxl_modeset_fini(struct qxl_device *qdev)
 {
 	qxl_fbdev_fini(qdev);
+
+	qxl_destroy_monitors_object(qdev);
 	if (qdev->mode_info.mode_config_initialized) {
 		drm_mode_config_cleanup(qdev->ddev);
 		qdev->mode_info.mode_config_initialized = false;
diff --git a/drivers/gpu/drm/qxl/qxl_drv.c b/drivers/gpu/drm/qxl/qxl_drv.c
index aa291d8a98a2..df0b577a6608 100644
--- a/drivers/gpu/drm/qxl/qxl_drv.c
+++ b/drivers/gpu/drm/qxl/qxl_drv.c
@@ -33,8 +33,9 @@
 
 #include "drmP.h"
 #include "drm/drm.h"
-
+#include "drm_crtc_helper.h"
 #include "qxl_drv.h"
+#include "qxl_object.h"
 
 extern int qxl_max_ioctls;
 static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
@@ -47,10 +48,14 @@ static DEFINE_PCI_DEVICE_TABLE(pciidlist) = {
 MODULE_DEVICE_TABLE(pci, pciidlist);
 
 static int qxl_modeset = -1;
+int qxl_num_crtc = 4;
 
 MODULE_PARM_DESC(modeset, "Disable/Enable modesetting");
 module_param_named(modeset, qxl_modeset, int, 0400);
 
+MODULE_PARM_DESC(num_heads, "Number of virtual crtcs to expose (default 4)");
+module_param_named(num_heads, qxl_num_crtc, int, 0400);
+
 static struct drm_driver qxl_driver;
 static struct pci_driver qxl_pci_driver;
 
@@ -73,13 +78,6 @@ qxl_pci_remove(struct pci_dev *pdev)
 	drm_put_dev(dev);
 }
 
-static struct pci_driver qxl_pci_driver = {
-	 .name = DRIVER_NAME,
-	 .id_table = pciidlist,
-	 .probe = qxl_pci_probe,
-	 .remove = qxl_pci_remove,
-};
-
 static const struct file_operations qxl_fops = {
 	.owner = THIS_MODULE,
 	.open = drm_open,
@@ -90,6 +88,130 @@ static const struct file_operations qxl_fops = {
 	.mmap = qxl_mmap,
 };
 
+static int qxl_drm_freeze(struct drm_device *dev)
+{
+	struct pci_dev *pdev = dev->pdev;
+	struct qxl_device *qdev = dev->dev_private;
+	struct drm_crtc *crtc;
+
+	drm_kms_helper_poll_disable(dev);
+
+	console_lock();
+	qxl_fbdev_set_suspend(qdev, 1);
+	console_unlock();
+
+	/* unpin the front buffers */
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private;
+		if (crtc->enabled)
+			(*crtc_funcs->disable)(crtc);
+	}
+
+	qxl_destroy_monitors_object(qdev);
+	qxl_surf_evict(qdev);
+	qxl_vram_evict(qdev);
+
+	while (!qxl_check_idle(qdev->command_ring));
+	while (!qxl_check_idle(qdev->release_ring))
+		qxl_queue_garbage_collect(qdev, 1);
+
+	pci_save_state(pdev);
+
+	return 0;
+}
+
+static int qxl_drm_resume(struct drm_device *dev, bool thaw)
+{
+	struct qxl_device *qdev = dev->dev_private;
+
+	qdev->ram_header->int_mask = QXL_INTERRUPT_MASK;
+	if (!thaw) {
+		qxl_reinit_memslots(qdev);
+		qxl_ring_init_hdr(qdev->release_ring);
+	}
+
+	qxl_create_monitors_object(qdev);
+	drm_helper_resume_force_mode(dev);
+
+	console_lock();
+	qxl_fbdev_set_suspend(qdev, 0);
+	console_unlock();
+
+	drm_kms_helper_poll_enable(dev);
+	return 0;
+}
+
+static int qxl_pm_suspend(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	int error;
+
+	error = qxl_drm_freeze(drm_dev);
+	if (error)
+		return error;
+
+	pci_disable_device(pdev);
+	pci_set_power_state(pdev, PCI_D3hot);
+	return 0;
+}
+
+static int qxl_pm_resume(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+
+	pci_set_power_state(pdev, PCI_D0);
+	pci_restore_state(pdev);
+	if (pci_enable_device(pdev)) {
+		return -EIO;
+	}
+
+	return qxl_drm_resume(drm_dev, false);
+}
+
+static int qxl_pm_thaw(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+
+	return qxl_drm_resume(drm_dev, true);
+}
+
+static int qxl_pm_freeze(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+
+	return qxl_drm_freeze(drm_dev);
+}
+
+static int qxl_pm_restore(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	struct qxl_device *qdev = drm_dev->dev_private;
+
+	qxl_io_reset(qdev);
+	return qxl_drm_resume(drm_dev, false);
+}
+
+static const struct dev_pm_ops qxl_pm_ops = {
+	.suspend = qxl_pm_suspend,
+	.resume = qxl_pm_resume,
+	.freeze = qxl_pm_freeze,
+	.thaw = qxl_pm_thaw,
+	.poweroff = qxl_pm_freeze,
+	.restore = qxl_pm_restore,
+};
+static struct pci_driver qxl_pci_driver = {
+	 .name = DRIVER_NAME,
+	 .id_table = pciidlist,
+	 .probe = qxl_pci_probe,
+	 .remove = qxl_pci_remove,
+	 .driver.pm = &qxl_pm_ops,
+};
+
 static struct drm_driver qxl_driver = {
 	.driver_features = DRIVER_GEM | DRIVER_MODESET |
 			   DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED,
diff --git a/drivers/gpu/drm/qxl/qxl_drv.h b/drivers/gpu/drm/qxl/qxl_drv.h
index 43d06ab28a21..aacb791464a3 100644
--- a/drivers/gpu/drm/qxl/qxl_drv.h
+++ b/drivers/gpu/drm/qxl/qxl_drv.h
@@ -55,11 +55,10 @@
 #define DRIVER_MINOR 1
 #define DRIVER_PATCHLEVEL 0
 
-#define QXL_NUM_OUTPUTS 1
-
 #define QXL_DEBUGFS_MAX_COMPONENTS		32
 
 extern int qxl_log_level;
+extern int qxl_num_crtc;
 
 enum {
 	QXL_INFO_LEVEL = 1,
@@ -139,6 +138,7 @@ struct qxl_reloc_list {
 
 struct qxl_crtc {
 	struct drm_crtc base;
+	int index;
 	int cur_x;
 	int cur_y;
 };
@@ -156,7 +156,7 @@ struct qxl_framebuffer {
 
 #define to_qxl_crtc(x) container_of(x, struct qxl_crtc, base)
 #define drm_connector_to_qxl_output(x) container_of(x, struct qxl_output, base)
-#define drm_encoder_to_qxl_output(x) container_of(x, struct qxl_output, base)
+#define drm_encoder_to_qxl_output(x) container_of(x, struct qxl_output, enc)
 #define to_qxl_framebuffer(x) container_of(x, struct qxl_framebuffer, base)
 
 struct qxl_mman {
@@ -331,6 +331,10 @@ void qxl_modeset_fini(struct qxl_device *qdev);
 int qxl_bo_init(struct qxl_device *qdev);
 void qxl_bo_fini(struct qxl_device *qdev);
 
+void qxl_reinit_memslots(struct qxl_device *qdev);
+int qxl_surf_evict(struct qxl_device *qdev);
+int qxl_vram_evict(struct qxl_device *qdev);
+
 struct qxl_ring *qxl_ring_create(struct qxl_ring_header *header,
 				 int element_size,
 				 int n_elements,
@@ -338,6 +342,8 @@ struct qxl_ring *qxl_ring_create(struct qxl_ring_header *header,
 				 bool set_prod_notify,
 				 wait_queue_head_t *push_event);
 void qxl_ring_free(struct qxl_ring *ring);
+void qxl_ring_init_hdr(struct qxl_ring *ring);
+int qxl_check_idle(struct qxl_ring *ring);
 
 static inline void *
 qxl_fb_virtual_address(struct qxl_device *qdev, unsigned long physical)
@@ -365,6 +371,7 @@ void qxl_fbdev_fini(struct qxl_device *qdev);
 int qxl_get_handle_for_primary_fb(struct qxl_device *qdev,
 				  struct drm_file *file_priv,
 				  uint32_t *handle);
+void qxl_fbdev_set_suspend(struct qxl_device *qdev, int state);
 
 /* qxl_display.c */
 int
@@ -374,6 +381,8 @@ qxl_framebuffer_init(struct drm_device *dev,
 		     struct drm_gem_object *obj);
 void qxl_display_read_client_monitors_config(struct qxl_device *qdev);
 void qxl_send_monitors_config(struct qxl_device *qdev);
+int qxl_create_monitors_object(struct qxl_device *qdev);
+int qxl_destroy_monitors_object(struct qxl_device *qdev);
 
 /* used by qxl_debugfs only */
 void qxl_crtc_set_from_monitors_config(struct qxl_device *qdev);
@@ -435,7 +444,7 @@ void qxl_update_screen(struct qxl_device *qxl);
 /* qxl io operations (qxl_cmd.c) */
 
 void qxl_io_create_primary(struct qxl_device *qdev,
-			   unsigned width, unsigned height, unsigned offset,
+			   unsigned offset,
 			   struct qxl_bo *bo);
 void qxl_io_destroy_primary(struct qxl_device *qdev);
 void qxl_io_memslot_add(struct qxl_device *qdev, uint8_t id);
@@ -528,6 +537,7 @@ irqreturn_t qxl_irq_handler(DRM_IRQ_ARGS);
 
 /* qxl_fb.c */
 int qxl_fb_init(struct qxl_device *qdev);
+bool qxl_fbdev_qobj_is_fb(struct qxl_device *qdev, struct qxl_bo *qobj);
 
 int qxl_debugfs_add_files(struct qxl_device *qdev,
 			  struct drm_info_list *files,
diff --git a/drivers/gpu/drm/qxl/qxl_fb.c b/drivers/gpu/drm/qxl/qxl_fb.c
index b3c51275df5c..76f39d88d684 100644
--- a/drivers/gpu/drm/qxl/qxl_fb.c
+++ b/drivers/gpu/drm/qxl/qxl_fb.c
@@ -520,10 +520,6 @@ static int qxl_fbdev_destroy(struct drm_device *dev, struct qxl_fbdev *qfbdev)
 }
 
 static struct drm_fb_helper_funcs qxl_fb_helper_funcs = {
-	/* TODO
-	.gamma_set = qxl_crtc_fb_gamma_set,
-	.gamma_get = qxl_crtc_fb_gamma_get,
-	*/
 	.fb_probe = qxl_fb_find_or_create_single,
 };
 
@@ -542,7 +538,7 @@ int qxl_fbdev_init(struct qxl_device *qdev)
 	qfbdev->helper.funcs = &qxl_fb_helper_funcs;
 
 	ret = drm_fb_helper_init(qdev->ddev, &qfbdev->helper,
-				 1 /* num_crtc - QXL supports just 1 */,
+				 qxl_num_crtc /* num_crtc - QXL supports just 1 */,
 				 QXLFB_CONN_LIMIT);
 	if (ret) {
 		kfree(qfbdev);
@@ -564,4 +560,14 @@ void qxl_fbdev_fini(struct qxl_device *qdev)
 	qdev->mode_info.qfbdev = NULL;
 }
 
+void qxl_fbdev_set_suspend(struct qxl_device *qdev, int state)
+{
+	fb_set_suspend(qdev->mode_info.qfbdev->helper.fbdev, state);
+}
 
+bool qxl_fbdev_qobj_is_fb(struct qxl_device *qdev, struct qxl_bo *qobj)
+{
+	if (qobj == gem_to_qxl_bo(qdev->mode_info.qfbdev->qfb.obj))
+		return true;
+	return false;
+}
diff --git a/drivers/gpu/drm/qxl/qxl_ioctl.c b/drivers/gpu/drm/qxl/qxl_ioctl.c
index a30f29425c21..27f45e49250d 100644
--- a/drivers/gpu/drm/qxl/qxl_ioctl.c
+++ b/drivers/gpu/drm/qxl/qxl_ioctl.c
@@ -188,6 +188,12 @@ static int qxl_execbuffer_ioctl(struct drm_device *dev, void *data,
 		/* TODO copy slow path code from i915 */
 		fb_cmd = qxl_bo_kmap_atomic_page(qdev, cmd_bo, (release->release_offset & PAGE_SIZE));
 		unwritten = __copy_from_user_inatomic_nocache(fb_cmd + sizeof(union qxl_release_info) + (release->release_offset & ~PAGE_SIZE), (void *)(unsigned long)user_cmd.command, user_cmd.command_size);
+
+		{
+			struct qxl_drawable *draw = fb_cmd;
+
+			draw->mm_time = qdev->rom->mm_clock;
+		}
 		qxl_bo_kunmap_atomic_page(qdev, cmd_bo, fb_cmd);
 		if (unwritten) {
 			DRM_ERROR("got unwritten %d\n", unwritten);
diff --git a/drivers/gpu/drm/qxl/qxl_kms.c b/drivers/gpu/drm/qxl/qxl_kms.c
index e27ce2a907cf..9e8da9ee9731 100644
--- a/drivers/gpu/drm/qxl/qxl_kms.c
+++ b/drivers/gpu/drm/qxl/qxl_kms.c
@@ -26,6 +26,7 @@
 #include "qxl_drv.h"
 #include "qxl_object.h"
 
+#include <drm/drm_crtc_helper.h>
 #include <linux/io-mapping.h>
 
 int qxl_log_level;
@@ -72,21 +73,28 @@ static bool qxl_check_device(struct qxl_device *qdev)
 	return true;
 }
 
+static void setup_hw_slot(struct qxl_device *qdev, int slot_index,
+			  struct qxl_memslot *slot)
+{
+	qdev->ram_header->mem_slot.mem_start = slot->start_phys_addr;
+	qdev->ram_header->mem_slot.mem_end = slot->end_phys_addr;
+	qxl_io_memslot_add(qdev, slot_index);
+}
+
 static uint8_t setup_slot(struct qxl_device *qdev, uint8_t slot_index_offset,
 	unsigned long start_phys_addr, unsigned long end_phys_addr)
 {
 	uint64_t high_bits;
 	struct qxl_memslot *slot;
 	uint8_t slot_index;
-	struct qxl_ram_header *ram_header = qdev->ram_header;
 
 	slot_index = qdev->rom->slots_start + slot_index_offset;
 	slot = &qdev->mem_slots[slot_index];
 	slot->start_phys_addr = start_phys_addr;
 	slot->end_phys_addr = end_phys_addr;
-	ram_header->mem_slot.mem_start = slot->start_phys_addr;
-	ram_header->mem_slot.mem_end = slot->end_phys_addr;
-	qxl_io_memslot_add(qdev, slot_index);
+
+	setup_hw_slot(qdev, slot_index, slot);
+
 	slot->generation = qdev->rom->slot_generation;
 	high_bits = slot_index << qdev->slot_gen_bits;
 	high_bits |= slot->generation;
@@ -95,6 +103,12 @@ static uint8_t setup_slot(struct qxl_device *qdev, uint8_t slot_index_offset,
 	return slot_index;
 }
 
+void qxl_reinit_memslots(struct qxl_device *qdev)
+{
+	setup_hw_slot(qdev, qdev->main_mem_slot, &qdev->mem_slots[qdev->main_mem_slot]);
+	setup_hw_slot(qdev, qdev->surfaces_mem_slot, &qdev->mem_slots[qdev->surfaces_mem_slot]);
+}
+
 static void qxl_gc_work(struct work_struct *work)
 {
 	struct qxl_device *qdev = container_of(work, struct qxl_device, gc_work);
@@ -294,6 +308,8 @@ int qxl_driver_load(struct drm_device *dev, unsigned long flags)
 		goto out;
 	}
 
+	drm_kms_helper_poll_init(qdev->ddev);
+
 	return 0;
 out:
 	kfree(qdev);
diff --git a/drivers/gpu/drm/qxl/qxl_object.c b/drivers/gpu/drm/qxl/qxl_object.c
index d9b12e7bc6e1..1191fe7788c9 100644
--- a/drivers/gpu/drm/qxl/qxl_object.c
+++ b/drivers/gpu/drm/qxl/qxl_object.c
@@ -363,3 +363,13 @@ int qxl_bo_list_add(struct qxl_reloc_list *reloc_list, struct qxl_bo *bo)
 		return ret;
 	return 0;
 }
+
+int qxl_surf_evict(struct qxl_device *qdev)
+{
+	return ttm_bo_evict_mm(&qdev->mman.bdev, TTM_PL_PRIV0);
+}
+
+int qxl_vram_evict(struct qxl_device *qdev)
+{
+	return ttm_bo_evict_mm(&qdev->mman.bdev, TTM_PL_VRAM);
+}
diff --git a/drivers/gpu/drm/qxl/qxl_object.h b/drivers/gpu/drm/qxl/qxl_object.h
index b4fd89fbd8b7..ee7ad79ce781 100644
--- a/drivers/gpu/drm/qxl/qxl_object.h
+++ b/drivers/gpu/drm/qxl/qxl_object.h
@@ -57,11 +57,6 @@ static inline unsigned long qxl_bo_size(struct qxl_bo *bo)
 	return bo->tbo.num_pages << PAGE_SHIFT;
 }
 
-static inline bool qxl_bo_is_reserved(struct qxl_bo *bo)
-{
-	return !!atomic_read(&bo->tbo.reserved);
-}
-
 static inline u64 qxl_bo_mmap_offset(struct qxl_bo *bo)
 {
 	return bo->tbo.addr_space_offset;
diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile
index 86c5e3611892..c3df52c1a60c 100644
--- a/drivers/gpu/drm/radeon/Makefile
+++ b/drivers/gpu/drm/radeon/Makefile
@@ -76,7 +76,10 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
 	evergreen.o evergreen_cs.o evergreen_blit_shaders.o evergreen_blit_kms.o \
 	evergreen_hdmi.o radeon_trace_points.o ni.o cayman_blit_shaders.o \
 	atombios_encoders.o radeon_semaphore.o radeon_sa.o atombios_i2c.o si.o \
-	si_blit_shaders.o radeon_prime.o radeon_uvd.o
+	si_blit_shaders.o radeon_prime.o radeon_uvd.o cik.o cik_blit_shaders.o \
+	r600_dpm.o rs780_dpm.o rv6xx_dpm.o rv770_dpm.o rv730_dpm.o rv740_dpm.o \
+	rv770_smc.o cypress_dpm.o btc_dpm.o sumo_dpm.o sumo_smc.o trinity_dpm.o \
+	trinity_smc.o ni_dpm.o si_smc.o si_dpm.o
 
 radeon-$(CONFIG_COMPAT) += radeon_ioc32.o
 radeon-$(CONFIG_VGA_SWITCHEROO) += radeon_atpx_handler.o
diff --git a/drivers/gpu/drm/radeon/ObjectID.h b/drivers/gpu/drm/radeon/ObjectID.h
index ca4b038050d2..06192698bd96 100644
--- a/drivers/gpu/drm/radeon/ObjectID.h
+++ b/drivers/gpu/drm/radeon/ObjectID.h
@@ -69,6 +69,8 @@
 #define ENCODER_OBJECT_ID_ALMOND                  0x22
 #define ENCODER_OBJECT_ID_TRAVIS                  0x23
 #define ENCODER_OBJECT_ID_NUTMEG                  0x22
+#define ENCODER_OBJECT_ID_HDMI_ANX9805            0x26
+
 /* Kaleidoscope (KLDSCP) Class Display Hardware (internal) */
 #define ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1   0x13
 #define ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1    0x14
@@ -86,6 +88,8 @@
 #define ENCODER_OBJECT_ID_INTERNAL_UNIPHY1        0x20
 #define ENCODER_OBJECT_ID_INTERNAL_UNIPHY2        0x21
 #define ENCODER_OBJECT_ID_INTERNAL_VCE            0x24
+#define ENCODER_OBJECT_ID_INTERNAL_UNIPHY3        0x25
+#define ENCODER_OBJECT_ID_INTERNAL_AMCLK          0x27
 
 #define ENCODER_OBJECT_ID_GENERAL_EXTERNAL_DVO    0xFF
 
@@ -364,6 +368,14 @@
                                                  GRAPH_OBJECT_ENUM_ID2 << ENUM_ID_SHIFT |\
                                                  ENCODER_OBJECT_ID_INTERNAL_UNIPHY2 << OBJECT_ID_SHIFT)
 
+#define ENCODER_INTERNAL_UNIPHY3_ENUM_ID1         ( GRAPH_OBJECT_TYPE_ENCODER << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
+                                                 ENCODER_OBJECT_ID_INTERNAL_UNIPHY3 << OBJECT_ID_SHIFT)
+
+#define ENCODER_INTERNAL_UNIPHY3_ENUM_ID2         ( GRAPH_OBJECT_TYPE_ENCODER << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID2 << ENUM_ID_SHIFT |\
+                                                 ENCODER_OBJECT_ID_INTERNAL_UNIPHY3 << OBJECT_ID_SHIFT)
+
 #define ENCODER_GENERAL_EXTERNAL_DVO_ENUM_ID1    ( GRAPH_OBJECT_TYPE_ENCODER << OBJECT_TYPE_SHIFT |\
                                                   GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
                                                   ENCODER_OBJECT_ID_GENERAL_EXTERNAL_DVO << OBJECT_ID_SHIFT)
@@ -392,6 +404,10 @@
                                                   GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
                                                   ENCODER_OBJECT_ID_INTERNAL_VCE << OBJECT_ID_SHIFT)
 
+#define ENCODER_HDMI_ANX9805_ENUM_ID1            ( GRAPH_OBJECT_TYPE_ENCODER << OBJECT_TYPE_SHIFT |\
+                                                  GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
+                                                  ENCODER_OBJECT_ID_HDMI_ANX9805 << OBJECT_ID_SHIFT)
+
 /****************************************************/
 /* Connector Object ID definition - Shared with BIOS */
 /****************************************************/
@@ -461,6 +477,14 @@
                                                  GRAPH_OBJECT_ENUM_ID4 << ENUM_ID_SHIFT |\
                                                  CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_D << OBJECT_ID_SHIFT)
 
+#define CONNECTOR_SINGLE_LINK_DVI_D_ENUM_ID5   ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID5 << ENUM_ID_SHIFT |\
+                                                 CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_D << OBJECT_ID_SHIFT)
+
+#define CONNECTOR_SINGLE_LINK_DVI_D_ENUM_ID6   ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID6 << ENUM_ID_SHIFT |\
+                                                 CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_D << OBJECT_ID_SHIFT)
+
 #define CONNECTOR_DUAL_LINK_DVI_D_ENUM_ID1     ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
                                                  GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
                                                  CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D << OBJECT_ID_SHIFT)
@@ -473,6 +497,10 @@
                                                  GRAPH_OBJECT_ENUM_ID3 << ENUM_ID_SHIFT |\
                                                  CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D << OBJECT_ID_SHIFT)
 
+#define CONNECTOR_DUAL_LINK_DVI_D_ENUM_ID4     ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID4 << ENUM_ID_SHIFT |\
+                                                 CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_D << OBJECT_ID_SHIFT)
+
 #define CONNECTOR_VGA_ENUM_ID1                 ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
                                                  GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
                                                  CONNECTOR_OBJECT_ID_VGA << OBJECT_ID_SHIFT)
@@ -541,6 +569,18 @@
                                                  GRAPH_OBJECT_ENUM_ID3 << ENUM_ID_SHIFT |\
                                                  CONNECTOR_OBJECT_ID_HDMI_TYPE_A << OBJECT_ID_SHIFT)
 
+#define CONNECTOR_HDMI_TYPE_A_ENUM_ID4         ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID4 << ENUM_ID_SHIFT |\
+                                                 CONNECTOR_OBJECT_ID_HDMI_TYPE_A << OBJECT_ID_SHIFT)
+
+#define CONNECTOR_HDMI_TYPE_A_ENUM_ID5         ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID5 << ENUM_ID_SHIFT |\
+                                                 CONNECTOR_OBJECT_ID_HDMI_TYPE_A << OBJECT_ID_SHIFT)
+
+#define CONNECTOR_HDMI_TYPE_A_ENUM_ID6         ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
+                                                 GRAPH_OBJECT_ENUM_ID6 << ENUM_ID_SHIFT |\
+                                                 CONNECTOR_OBJECT_ID_HDMI_TYPE_A << OBJECT_ID_SHIFT)
+
 #define CONNECTOR_HDMI_TYPE_B_ENUM_ID1         ( GRAPH_OBJECT_TYPE_CONNECTOR << OBJECT_TYPE_SHIFT |\
                                                  GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
                                                  CONNECTOR_OBJECT_ID_HDMI_TYPE_B << OBJECT_ID_SHIFT)
diff --git a/drivers/gpu/drm/radeon/atombios.h b/drivers/gpu/drm/radeon/atombios.h
index 0ee573743de9..16b120c3f144 100644
--- a/drivers/gpu/drm/radeon/atombios.h
+++ b/drivers/gpu/drm/radeon/atombios.h
@@ -74,6 +74,8 @@
 #define ATOM_PPLL2            1
 #define ATOM_DCPLL            2
 #define ATOM_PPLL0            2
+#define ATOM_PPLL3            3
+
 #define ATOM_EXT_PLL1         8
 #define ATOM_EXT_PLL2         9
 #define ATOM_EXT_CLOCK        10
@@ -259,7 +261,7 @@ typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES{
   USHORT AdjustDisplayPll;											 //Atomic Table,  used by various SW componentes. 
   USHORT AdjustMemoryController;                 //Atomic Table,  indirectly used by various SW components,called from SetMemoryClock                
   USHORT EnableASIC_StaticPwrMgt;                //Atomic Table,  only used by Bios
-  USHORT ASIC_StaticPwrMgtStatusChange;          //Obsolete ,     only used by Bios   
+  USHORT SetUniphyInstance;                      //Atomic Table,  only used by Bios   
   USHORT DAC_LoadDetection;                      //Atomic Table,  directly used by various SW components,latest version 1.2  
   USHORT LVTMAEncoderControl;                    //Atomic Table,directly used by various SW components,latest version 1.3
   USHORT HW_Misc_Operation;                      //Atomic Table,  directly used by various SW components,latest version 1.1 
@@ -271,7 +273,7 @@ typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES{
   USHORT TVEncoderControl;                       //Function Table,directly used by various SW components,latest version 1.1
   USHORT PatchMCSetting;                         //only used by BIOS
   USHORT MC_SEQ_Control;                         //only used by BIOS
-  USHORT TV1OutputControl;                       //Atomic Table,  Obsolete from Ry6xx, use DAC2 Output instead
+  USHORT Gfx_Harvesting;                         //Atomic Table,  Obsolete from Ry6xx, Now only used by BIOS for GFX harvesting
   USHORT EnableScaler;                           //Atomic Table,  used only by Bios
   USHORT BlankCRTC;                              //Atomic Table,  directly used by various SW components,latest version 1.1 
   USHORT EnableCRTC;                             //Atomic Table,  directly used by various SW components,latest version 1.1 
@@ -328,7 +330,7 @@ typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES{
 #define UNIPHYTransmitterControl			     DIG1TransmitterControl
 #define LVTMATransmitterControl				     DIG2TransmitterControl
 #define SetCRTC_DPM_State                        GetConditionalGoldenSetting
-#define SetUniphyInstance                        ASIC_StaticPwrMgtStatusChange
+#define ASIC_StaticPwrMgtStatusChange            SetUniphyInstance 
 #define HPDInterruptService                      ReadHWAssistedI2CStatus
 #define EnableVGA_Access                         GetSCLKOverMCLKRatio
 #define EnableYUV                                GetDispObjectInfo                         
@@ -338,7 +340,7 @@ typedef struct _ATOM_MASTER_LIST_OF_COMMAND_TABLES{
 #define TMDSAEncoderControl                      PatchMCSetting
 #define LVDSEncoderControl                       MC_SEQ_Control
 #define LCD1OutputControl                        HW_Misc_Operation
-
+#define TV1OutputControl                         Gfx_Harvesting
 
 typedef struct _ATOM_MASTER_COMMAND_TABLE
 {
@@ -478,11 +480,11 @@ typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V3
 typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V4
 {
 #if ATOM_BIG_ENDIAN
-  ULONG  ucPostDiv;          //return parameter: post divider which is used to program to register directly
+  ULONG  ucPostDiv:8;        //return parameter: post divider which is used to program to register directly
   ULONG  ulClock:24;         //Input= target clock, output = actual clock 
 #else
   ULONG  ulClock:24;         //Input= target clock, output = actual clock 
-  ULONG  ucPostDiv;          //return parameter: post divider which is used to program to register directly
+  ULONG  ucPostDiv:8;        //return parameter: post divider which is used to program to register directly
 #endif
 }COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V4;
 
@@ -504,6 +506,32 @@ typedef struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V5
   UCHAR   ucReserved;                       
 }COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V5;
 
+
+typedef struct _COMPUTE_GPU_CLOCK_INPUT_PARAMETERS_V1_6
+{
+  ATOM_COMPUTE_CLOCK_FREQ  ulClock;         //Input Parameter
+  ULONG   ulReserved[2];
+}COMPUTE_GPU_CLOCK_INPUT_PARAMETERS_V1_6;
+
+//ATOM_COMPUTE_CLOCK_FREQ.ulComputeClockFlag
+#define COMPUTE_GPUCLK_INPUT_FLAG_CLK_TYPE_MASK            0x0f
+#define COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK           0x00
+#define COMPUTE_GPUCLK_INPUT_FLAG_SCLK                     0x01
+
+typedef struct _COMPUTE_GPU_CLOCK_OUTPUT_PARAMETERS_V1_6
+{
+  COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V4  ulClock;         //Output Parameter: ucPostDiv=DFS divider
+  ATOM_S_MPLL_FB_DIVIDER   ulFbDiv;         //Output Parameter: PLL FB divider
+  UCHAR   ucPllRefDiv;                      //Output Parameter: PLL ref divider      
+  UCHAR   ucPllPostDiv;                     //Output Parameter: PLL post divider      
+  UCHAR   ucPllCntlFlag;                    //Output Flags: control flag
+  UCHAR   ucReserved;                       
+}COMPUTE_GPU_CLOCK_OUTPUT_PARAMETERS_V1_6;
+
+//ucPllCntlFlag
+#define SPLL_CNTL_FLAG_VCO_MODE_MASK            0x03 
+
+
 // ucInputFlag
 #define ATOM_PLL_INPUT_FLAG_PLL_STROBE_MODE_EN  1   // 1-StrobeMode, 0-PerformanceMode
 
@@ -1686,6 +1714,7 @@ typedef struct _PIXEL_CLOCK_PARAMETERS_V6
 #define PIXEL_CLOCK_V6_MISC_HDMI_30BPP              0x08
 #define PIXEL_CLOCK_V6_MISC_HDMI_48BPP              0x0c
 #define PIXEL_CLOCK_V6_MISC_REF_DIV_SRC             0x10
+#define PIXEL_CLOCK_V6_MISC_GEN_DPREFCLK            0x40
 
 typedef struct _GET_DISP_PLL_STATUS_INPUT_PARAMETERS_V2
 {
@@ -2102,6 +2131,17 @@ typedef struct _DVO_ENCODER_CONTROL_PARAMETERS_V3
 }DVO_ENCODER_CONTROL_PARAMETERS_V3;
 #define DVO_ENCODER_CONTROL_PS_ALLOCATION_V3	DVO_ENCODER_CONTROL_PARAMETERS_V3
 
+typedef struct _DVO_ENCODER_CONTROL_PARAMETERS_V1_4
+{
+  USHORT usPixelClock; 
+  UCHAR  ucDVOConfig;
+  UCHAR  ucAction;														//ATOM_ENABLE/ATOM_DISABLE/ATOM_HPD_INIT
+  UCHAR  ucBitPerColor;                       //please refer to definition of PANEL_xBIT_PER_COLOR
+  UCHAR  ucReseved[3];
+}DVO_ENCODER_CONTROL_PARAMETERS_V1_4;
+#define DVO_ENCODER_CONTROL_PS_ALLOCATION_V1_4	DVO_ENCODER_CONTROL_PARAMETERS_V1_4
+
+
 //ucTableFormatRevision=1
 //ucTableContentRevision=3 structure is not changed but usMisc add bit 1 as another input for 
 // bit1=0: non-coherent mode
@@ -2165,7 +2205,7 @@ typedef struct _DVO_ENCODER_CONTROL_PARAMETERS_V3
 #define SET_ASIC_VOLTAGE_MODE_SOURCE_B         0x4
 
 #define	SET_ASIC_VOLTAGE_MODE_SET_VOLTAGE      0x0
-#define	SET_ASIC_VOLTAGE_MODE_GET_GPIOVAL      0x1	
+#define	SET_ASIC_VOLTAGE_MODE_GET_GPIOVAL      0x1
 #define	SET_ASIC_VOLTAGE_MODE_GET_GPIOMASK     0x2
 
 typedef struct	_SET_VOLTAGE_PARAMETERS
@@ -2200,15 +2240,20 @@ typedef struct	_SET_VOLTAGE_PARAMETERS_V1_3
 //SET_VOLTAGE_PARAMETERS_V3.ucVoltageMode
 #define ATOM_SET_VOLTAGE                     0        //Set voltage Level
 #define ATOM_INIT_VOLTAGE_REGULATOR          3        //Init Regulator
-#define ATOM_SET_VOLTAGE_PHASE               4        //Set Vregulator Phase
-#define ATOM_GET_MAX_VOLTAGE                 6        //Get Max Voltage, not used in SetVoltageTable v1.3
-#define ATOM_GET_VOLTAGE_LEVEL               6        //Get Voltage level from vitual voltage ID
+#define ATOM_SET_VOLTAGE_PHASE               4        //Set Vregulator Phase, only for SVID/PVID regulator
+#define ATOM_GET_MAX_VOLTAGE                 6        //Get Max Voltage, not used from SetVoltageTable v1.3
+#define ATOM_GET_VOLTAGE_LEVEL               6        //Get Voltage level from vitual voltage ID, not used for SetVoltage v1.4
+#define ATOM_GET_LEAKAGE_ID                  8        //Get Leakage Voltage Id ( starting from SMU7x IP ), SetVoltage v1.4 
 
 // define vitual voltage id in usVoltageLevel
 #define ATOM_VIRTUAL_VOLTAGE_ID0             0xff01
 #define ATOM_VIRTUAL_VOLTAGE_ID1             0xff02
 #define ATOM_VIRTUAL_VOLTAGE_ID2             0xff03
 #define ATOM_VIRTUAL_VOLTAGE_ID3             0xff04
+#define ATOM_VIRTUAL_VOLTAGE_ID4             0xff05
+#define ATOM_VIRTUAL_VOLTAGE_ID5             0xff06
+#define ATOM_VIRTUAL_VOLTAGE_ID6             0xff07
+#define ATOM_VIRTUAL_VOLTAGE_ID7             0xff08
 
 typedef struct _SET_VOLTAGE_PS_ALLOCATION
 {
@@ -2628,7 +2673,8 @@ typedef struct _ATOM_FIRMWARE_INFO_V2_2
   ULONG                           ulFirmwareRevision;
   ULONG                           ulDefaultEngineClock;       //In 10Khz unit
   ULONG                           ulDefaultMemoryClock;       //In 10Khz unit
-  ULONG                           ulReserved[2];
+  ULONG                           ulSPLL_OutputFreq;          //In 10Khz unit  
+  ULONG                           ulGPUPLL_OutputFreq;        //In 10Khz unit
   ULONG                           ulReserved1;                //Was ulMaxEngineClockPLL_Output; //In 10Khz unit*
   ULONG                           ulReserved2;                //Was ulMaxMemoryClockPLL_Output; //In 10Khz unit*
   ULONG                           ulMaxPixelClockPLL_Output;  //In 10Khz unit
@@ -3813,6 +3859,12 @@ typedef struct _ATOM_GPIO_PIN_ASSIGNMENT
   UCHAR                    ucGPIO_ID;
 }ATOM_GPIO_PIN_ASSIGNMENT;
 
+//ucGPIO_ID pre-define id for multiple usage
+//from SMU7.x, if ucGPIO_ID=PP_AC_DC_SWITCH_GPIO_PINID in GPIO_LUTTable, AC/DC swithing feature is enable
+#define PP_AC_DC_SWITCH_GPIO_PINID          60
+//from SMU7.x, if ucGPIO_ID=VDDC_REGULATOR_VRHOT_GPIO_PINID in GPIO_LUTable, VRHot feature is enable
+#define VDDC_VRHOT_GPIO_PINID               61
+
 typedef struct _ATOM_GPIO_PIN_LUT
 {
   ATOM_COMMON_TABLE_HEADER  sHeader;
@@ -4074,17 +4126,19 @@ typedef struct _EXT_DISPLAY_PATH
 
 //usCaps
 #define  EXT_DISPLAY_PATH_CAPS__HBR2_DISABLE          0x01
+#define  EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN        0x02
 
 typedef  struct _ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO
 {
   ATOM_COMMON_TABLE_HEADER sHeader;
   UCHAR                    ucGuid [NUMBER_OF_UCHAR_FOR_GUID];     // a GUID is a 16 byte long string
   EXT_DISPLAY_PATH         sPath[MAX_NUMBER_OF_EXT_DISPLAY_PATH]; // total of fixed 7 entries.
-  UCHAR                    ucChecksum;                            // a  simple Checksum of the sum of whole structure equal to 0x0. 
+  UCHAR                    ucChecksum;                            // a simple Checksum of the sum of whole structure equal to 0x0. 
   UCHAR                    uc3DStereoPinId;                       // use for eDP panel
   UCHAR                    ucRemoteDisplayConfig;
   UCHAR                    uceDPToLVDSRxId;
-  UCHAR                    Reserved[4];                           // for potential expansion
+  UCHAR                    ucFixDPVoltageSwing;                   // usCaps[1]=1, this indicate DP_LANE_SET value
+  UCHAR                    Reserved[3];                           // for potential expansion
 }ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO;
 
 //Related definitions, all records are different but they have a commond header
@@ -4416,6 +4470,13 @@ typedef struct _ATOM_VOLTAGE_CONTROL
 #define	VOLTAGE_CONTROL_ID_CHL822x						0x08									
 #define	VOLTAGE_CONTROL_ID_VT1586M						0x09
 #define VOLTAGE_CONTROL_ID_UP1637 						0x0A
+#define	VOLTAGE_CONTROL_ID_CHL8214            0x0B
+#define	VOLTAGE_CONTROL_ID_UP1801             0x0C
+#define	VOLTAGE_CONTROL_ID_ST6788A            0x0D
+#define VOLTAGE_CONTROL_ID_CHLIR3564SVI2      0x0E
+#define VOLTAGE_CONTROL_ID_AD527x      	      0x0F
+#define VOLTAGE_CONTROL_ID_NCP81022    	      0x10
+#define VOLTAGE_CONTROL_ID_LTC2635			  0x11
 
 typedef struct  _ATOM_VOLTAGE_OBJECT
 {
@@ -4458,6 +4519,15 @@ typedef struct _ATOM_VOLTAGE_OBJECT_HEADER_V3{
 	 USHORT		usSize;													//Size of Object	
 }ATOM_VOLTAGE_OBJECT_HEADER_V3;
 
+// ATOM_VOLTAGE_OBJECT_HEADER_V3.ucVoltageMode
+#define VOLTAGE_OBJ_GPIO_LUT                 0        //VOLTAGE and GPIO Lookup table ->ATOM_GPIO_VOLTAGE_OBJECT_V3
+#define VOLTAGE_OBJ_VR_I2C_INIT_SEQ          3        //VOLTAGE REGULATOR INIT sequece through I2C -> ATOM_I2C_VOLTAGE_OBJECT_V3
+#define VOLTAGE_OBJ_PHASE_LUT                4        //Set Vregulator Phase lookup table ->ATOM_GPIO_VOLTAGE_OBJECT_V3
+#define VOLTAGE_OBJ_SVID2                    7        //Indicate voltage control by SVID2 ->ATOM_SVID2_VOLTAGE_OBJECT_V3
+#define	VOLTAGE_OBJ_PWRBOOST_LEAKAGE_LUT     0x10     //Powerboost Voltage and LeakageId lookup table->ATOM_LEAKAGE_VOLTAGE_OBJECT_V3
+#define	VOLTAGE_OBJ_HIGH_STATE_LEAKAGE_LUT   0x11     //High voltage state Voltage and LeakageId lookup table->ATOM_LEAKAGE_VOLTAGE_OBJECT_V3
+#define VOLTAGE_OBJ_HIGH1_STATE_LEAKAGE_LUT  0x12     //High1 voltage state Voltage and LeakageId lookup table->ATOM_LEAKAGE_VOLTAGE_OBJECT_V3
+
 typedef struct  _VOLTAGE_LUT_ENTRY_V2
 {
 	 ULONG		ulVoltageId;									  // The Voltage ID which is used to program GPIO register
@@ -4473,7 +4543,7 @@ typedef struct  _LEAKAGE_VOLTAGE_LUT_ENTRY_V2
 
 typedef struct  _ATOM_I2C_VOLTAGE_OBJECT_V3
 {
-   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;
+   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;    // voltage mode = VOLTAGE_OBJ_VR_I2C_INIT_SEQ
    UCHAR	ucVoltageRegulatorId;					  //Indicate Voltage Regulator Id
    UCHAR    ucVoltageControlI2cLine;
    UCHAR    ucVoltageControlAddress;
@@ -4484,7 +4554,7 @@ typedef struct  _ATOM_I2C_VOLTAGE_OBJECT_V3
 
 typedef struct  _ATOM_GPIO_VOLTAGE_OBJECT_V3
 {
-   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;   
+   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;   // voltage mode = VOLTAGE_OBJ_GPIO_LUT or VOLTAGE_OBJ_PHASE_LUT
    UCHAR    ucVoltageGpioCntlId;         // default is 0 which indicate control through CG VID mode 
    UCHAR    ucGpioEntryNum;              // indiate the entry numbers of Votlage/Gpio value Look up table
    UCHAR    ucPhaseDelay;                // phase delay in unit of micro second
@@ -4495,7 +4565,7 @@ typedef struct  _ATOM_GPIO_VOLTAGE_OBJECT_V3
 
 typedef struct  _ATOM_LEAKAGE_VOLTAGE_OBJECT_V3
 {
-   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;
+   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;    // voltage mode = 0x10/0x11/0x12
    UCHAR    ucLeakageCntlId;             // default is 0
    UCHAR    ucLeakageEntryNum;           // indicate the entry number of LeakageId/Voltage Lut table
    UCHAR    ucReserved[2];               
@@ -4503,10 +4573,26 @@ typedef struct  _ATOM_LEAKAGE_VOLTAGE_OBJECT_V3
    LEAKAGE_VOLTAGE_LUT_ENTRY_V2 asLeakageIdLut[1];   
 }ATOM_LEAKAGE_VOLTAGE_OBJECT_V3;
 
+
+typedef struct  _ATOM_SVID2_VOLTAGE_OBJECT_V3
+{
+   ATOM_VOLTAGE_OBJECT_HEADER_V3 sHeader;    // voltage mode = VOLTAGE_OBJ_SVID2
+// 14:7 – PSI0_VID
+// 6 – PSI0_EN
+// 5 – PSI1
+// 4:2 – load line slope trim. 
+// 1:0 – offset trim, 
+   USHORT   usLoadLine_PSI;    
+// GPU GPIO pin Id to SVID2 regulator VRHot pin. possible value 0~31. 0 means GPIO0, 31 means GPIO31
+   UCHAR    ucReserved[2];
+   ULONG    ulReserved;
+}ATOM_SVID2_VOLTAGE_OBJECT_V3;
+
 typedef union _ATOM_VOLTAGE_OBJECT_V3{
   ATOM_GPIO_VOLTAGE_OBJECT_V3 asGpioVoltageObj;
   ATOM_I2C_VOLTAGE_OBJECT_V3 asI2cVoltageObj;
   ATOM_LEAKAGE_VOLTAGE_OBJECT_V3 asLeakageObj;
+  ATOM_SVID2_VOLTAGE_OBJECT_V3 asSVID2Obj;
 }ATOM_VOLTAGE_OBJECT_V3;
 
 typedef struct  _ATOM_VOLTAGE_OBJECT_INFO_V3_1
@@ -4536,6 +4622,21 @@ typedef struct  _ATOM_ASIC_PROFILING_INFO
 	ATOM_ASIC_PROFILE_VOLTAGE			asVoltage;
 }ATOM_ASIC_PROFILING_INFO;
 
+typedef struct  _ATOM_ASIC_PROFILING_INFO_V2_1
+{
+  ATOM_COMMON_TABLE_HEADER			asHeader; 
+  UCHAR  ucLeakageBinNum;                // indicate the entry number of LeakageId/Voltage Lut table
+  USHORT usLeakageBinArrayOffset;        // offset of USHORT Leakage Bin list array ( from lower LeakageId to higher) 
+
+  UCHAR  ucElbVDDC_Num;               
+  USHORT usElbVDDC_IdArrayOffset;        // offset of USHORT virtual VDDC voltage id ( 0xff01~0xff08 )
+  USHORT usElbVDDC_LevelArrayOffset;     // offset of 2 dimension voltage level USHORT array
+
+  UCHAR  ucElbVDDCI_Num;
+  USHORT usElbVDDCI_IdArrayOffset;       // offset of USHORT virtual VDDCI voltage id ( 0xff01~0xff08 )
+  USHORT usElbVDDCI_LevelArrayOffset;    // offset of 2 dimension voltage level USHORT array
+}ATOM_ASIC_PROFILING_INFO_V2_1;
+
 typedef struct _ATOM_POWER_SOURCE_OBJECT
 {
 	UCHAR	ucPwrSrcId;													// Power source
@@ -4652,6 +4753,8 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V6
 #define SYS_INFO_LVDSMISC__888_BPC                                                   0x04
 #define SYS_INFO_LVDSMISC__OVERRIDE_EN                                               0x08
 #define SYS_INFO_LVDSMISC__BLON_ACTIVE_LOW                                           0x10
+// new since Trinity
+#define SYS_INFO_LVDSMISC__TRAVIS_LVDS_VOL_OVERRIDE_EN                               0x20
 
 // not used any more
 #define SYS_INFO_LVDSMISC__VSYNC_ACTIVE_LOW                                          0x04
@@ -4752,6 +4855,29 @@ typedef struct _ATOM_FUSION_SYSTEM_INFO_V1
   ATOM_INTEGRATED_SYSTEM_INFO_V6    sIntegratedSysInfo;   
   ULONG  ulPowerplayTable[128];  
 }ATOM_FUSION_SYSTEM_INFO_V1; 
+
+
+typedef struct _ATOM_TDP_CONFIG_BITS
+{
+#if ATOM_BIG_ENDIAN
+  ULONG   uReserved:2;
+  ULONG   uTDP_Value:14;  // Original TDP value in tens of milli watts
+  ULONG   uCTDP_Value:14; // Override value in tens of milli watts
+  ULONG   uCTDP_Enable:2; // = (uCTDP_Value > uTDP_Value? 2: (uCTDP_Value < uTDP_Value))
+#else
+  ULONG   uCTDP_Enable:2; // = (uCTDP_Value > uTDP_Value? 2: (uCTDP_Value < uTDP_Value))
+  ULONG   uCTDP_Value:14; // Override value in tens of milli watts
+  ULONG   uTDP_Value:14;  // Original TDP value in tens of milli watts
+  ULONG   uReserved:2;
+#endif
+}ATOM_TDP_CONFIG_BITS;
+
+typedef union _ATOM_TDP_CONFIG
+{
+  ATOM_TDP_CONFIG_BITS TDP_config;
+  ULONG            TDP_config_all;
+}ATOM_TDP_CONFIG;
+
 /**********************************************************************************************************************
   ATOM_FUSION_SYSTEM_INFO_V1 Description
 sIntegratedSysInfo:               refer to ATOM_INTEGRATED_SYSTEM_INFO_V6 definition.
@@ -4784,7 +4910,8 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7
   UCHAR  ucMemoryType;  
   UCHAR  ucUMAChannelNumber;
   UCHAR  strVBIOSMsg[40];
-  ULONG  ulReserved[20];
+  ATOM_TDP_CONFIG  asTdpConfig;
+  ULONG  ulReserved[19];
   ATOM_AVAILABLE_SCLK_LIST   sAvail_SCLK[5];
   ULONG  ulGMCRestoreResetTime;
   ULONG  ulMinimumNClk;
@@ -4809,7 +4936,7 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7
   USHORT GnbTdpLimit;
   USHORT usMaxLVDSPclkFreqInSingleLink;
   UCHAR  ucLvdsMisc;
-  UCHAR  ucLVDSReserved;
+  UCHAR  ucTravisLVDSVolAdjust;
   UCHAR  ucLVDSPwrOnSeqDIGONtoDE_in4Ms;
   UCHAR  ucLVDSPwrOnSeqDEtoVARY_BL_in4Ms;
   UCHAR  ucLVDSPwrOffSeqVARY_BLtoDE_in4Ms;
@@ -4817,7 +4944,7 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7
   UCHAR  ucLVDSOffToOnDelay_in4Ms;
   UCHAR  ucLVDSPwrOnSeqVARY_BLtoBLON_in4Ms;
   UCHAR  ucLVDSPwrOffSeqBLONtoVARY_BL_in4Ms;
-  UCHAR  ucLVDSReserved1;
+  UCHAR  ucMinAllowedBL_Level;
   ULONG  ulLCDBitDepthControlVal;
   ULONG  ulNbpStateMemclkFreq[4];
   USHORT usNBP2Voltage;               
@@ -4846,6 +4973,7 @@ typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7
 #define SYS_INFO_GPUCAPS__TMDSHDMI_COHERENT_SINGLEPLL_MODE                0x01
 #define SYS_INFO_GPUCAPS__DP_SINGLEPLL_MODE                               0x02
 #define SYS_INFO_GPUCAPS__DISABLE_AUX_MODE_DETECT                         0x08
+#define SYS_INFO_GPUCAPS__ENABEL_DFS_BYPASS                               0x10
 
 /**********************************************************************************************************************
   ATOM_INTEGRATED_SYSTEM_INFO_V1_7 Description
@@ -4945,6 +5073,9 @@ ucLVDSMisc:                       [bit0] LVDS 888bit panel mode =0: LVDS 888 pan
                                   [bit2] LVDS 888bit per color mode  =0: 666 bit per color =1:888 bit per color
                                   [bit3] LVDS parameter override enable  =0: ucLvdsMisc parameter are not used =1: ucLvdsMisc parameter should be used
                                   [bit4] Polarity of signal sent to digital BLON output pin. =0: not inverted(active high) =1: inverted ( active low )
+                                  [bit5] Travid LVDS output voltage override enable, when =1, use ucTravisLVDSVolAdjust value to overwrite Traivs register LVDS_CTRL_4
+ucTravisLVDSVolAdjust             When ucLVDSMisc[5]=1,it means platform SBIOS want to overwrite TravisLVDSVoltage. Then VBIOS will use ucTravisLVDSVolAdjust 
+                                  value to program Travis register LVDS_CTRL_4
 ucLVDSPwrOnSeqDIGONtoDE_in4Ms:    LVDS power up sequence time in unit of 4ms, time delay from DIGON signal active to data enable signal active( DE ).
                                   =0 mean use VBIOS default which is 8 ( 32ms ). The LVDS power up sequence is as following: DIGON->DE->VARY_BL->BLON. 
                                   This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
@@ -4964,18 +5095,241 @@ ucLVDSOffToOnDelay_in4Ms:         LVDS power down sequence time in unit of 4ms.
                                   =0 means to use VBIOS default delay which is 125 ( 500ms ).
                                   This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
 
-ucLVDSPwrOnVARY_BLtoBLON_in4Ms:   LVDS power up sequence time in unit of 4ms. Time delay from VARY_BL signal on to DLON signal active. 
+ucLVDSPwrOnSeqVARY_BLtoBLON_in4Ms:
+                                  LVDS power up sequence time in unit of 4ms. Time delay from VARY_BL signal on to DLON signal active. 
                                   =0 means to use VBIOS default delay which is 0 ( 0ms ).
                                   This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
 
-ucLVDSPwrOffBLONtoVARY_BL_in4Ms:  LVDS power down sequence time in unit of 4ms. Time delay from BLON signal off to VARY_BL signal off. 
+ucLVDSPwrOffSeqBLONtoVARY_BL_in4Ms:  
+                                  LVDS power down sequence time in unit of 4ms. Time delay from BLON signal off to VARY_BL signal off. 
                                   =0 means to use VBIOS default delay which is 0 ( 0ms ).
                                   This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
 
+ucMinAllowedBL_Level:             Lowest LCD backlight PWM level. This is customer platform specific parameters. By default it is 0. 
+
 ulNbpStateMemclkFreq[4]:          system memory clock frequncey in unit of 10Khz in different NB pstate. 
 
 **********************************************************************************************************************/
 
+// this IntegrateSystemInfoTable is used for Kaveri & Kabini APU
+typedef struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_8
+{
+  ATOM_COMMON_TABLE_HEADER   sHeader;
+  ULONG  ulBootUpEngineClock;
+  ULONG  ulDentistVCOFreq;
+  ULONG  ulBootUpUMAClock;
+  ATOM_CLK_VOLT_CAPABILITY   sDISPCLK_Voltage[4];
+  ULONG  ulBootUpReqDisplayVector;
+  ULONG  ulVBIOSMisc;
+  ULONG  ulGPUCapInfo;
+  ULONG  ulDISP_CLK2Freq;
+  USHORT usRequestedPWMFreqInHz;
+  UCHAR  ucHtcTmpLmt;
+  UCHAR  ucHtcHystLmt;
+  ULONG  ulReserved2;
+  ULONG  ulSystemConfig;            
+  ULONG  ulCPUCapInfo;
+  ULONG  ulReserved3;
+  USHORT usGPUReservedSysMemSize;
+  USHORT usExtDispConnInfoOffset;
+  USHORT usPanelRefreshRateRange;     
+  UCHAR  ucMemoryType;  
+  UCHAR  ucUMAChannelNumber;
+  UCHAR  strVBIOSMsg[40];
+  ATOM_TDP_CONFIG  asTdpConfig;
+  ULONG  ulReserved[19];
+  ATOM_AVAILABLE_SCLK_LIST   sAvail_SCLK[5];
+  ULONG  ulGMCRestoreResetTime;
+  ULONG  ulReserved4;
+  ULONG  ulIdleNClk;
+  ULONG  ulDDR_DLL_PowerUpTime;
+  ULONG  ulDDR_PLL_PowerUpTime;
+  USHORT usPCIEClkSSPercentage;
+  USHORT usPCIEClkSSType;
+  USHORT usLvdsSSPercentage;
+  USHORT usLvdsSSpreadRateIn10Hz;
+  USHORT usHDMISSPercentage;
+  USHORT usHDMISSpreadRateIn10Hz;
+  USHORT usDVISSPercentage;
+  USHORT usDVISSpreadRateIn10Hz;
+  ULONG  ulGPUReservedSysMemBaseAddrLo;
+  ULONG  ulGPUReservedSysMemBaseAddrHi;
+  ULONG  ulReserved5[3];
+  USHORT usMaxLVDSPclkFreqInSingleLink;
+  UCHAR  ucLvdsMisc;
+  UCHAR  ucTravisLVDSVolAdjust;
+  UCHAR  ucLVDSPwrOnSeqDIGONtoDE_in4Ms;
+  UCHAR  ucLVDSPwrOnSeqDEtoVARY_BL_in4Ms;
+  UCHAR  ucLVDSPwrOffSeqVARY_BLtoDE_in4Ms;
+  UCHAR  ucLVDSPwrOffSeqDEtoDIGON_in4Ms;
+  UCHAR  ucLVDSOffToOnDelay_in4Ms;
+  UCHAR  ucLVDSPwrOnSeqVARY_BLtoBLON_in4Ms;
+  UCHAR  ucLVDSPwrOffSeqBLONtoVARY_BL_in4Ms;
+  UCHAR  ucMinAllowedBL_Level;
+  ULONG  ulLCDBitDepthControlVal;
+  ULONG  ulNbpStateMemclkFreq[4];
+  ULONG  ulReserved6;               
+  ULONG  ulNbpStateNClkFreq[4];
+  USHORT usNBPStateVoltage[4];            
+  USHORT usBootUpNBVoltage;   
+  USHORT usReserved2; 
+  ATOM_EXTERNAL_DISPLAY_CONNECTION_INFO sExtDispConnInfo;
+}ATOM_INTEGRATED_SYSTEM_INFO_V1_8;
+
+/**********************************************************************************************************************
+  ATOM_INTEGRATED_SYSTEM_INFO_V1_8 Description
+ulBootUpEngineClock:              VBIOS bootup Engine clock frequency, in 10kHz unit. if it is equal 0, then VBIOS use pre-defined bootup engine clock
+ulDentistVCOFreq:                 Dentist VCO clock in 10kHz unit. 
+ulBootUpUMAClock:                 System memory boot up clock frequency in 10Khz unit. 
+sDISPCLK_Voltage:                 Report Display clock frequency requirement on GNB voltage(up to 4 voltage levels).
+ 
+ulBootUpReqDisplayVector:         VBIOS boot up display IDs, following are supported devices in Trinity projects:
+                                  ATOM_DEVICE_CRT1_SUPPORT                  0x0001
+                                  ATOM_DEVICE_DFP1_SUPPORT                  0x0008
+                                  ATOM_DEVICE_DFP6_SUPPORT                  0x0040
+                                  ATOM_DEVICE_DFP2_SUPPORT                  0x0080
+                                  ATOM_DEVICE_DFP3_SUPPORT                  0x0200
+                                  ATOM_DEVICE_DFP4_SUPPORT                  0x0400
+                                  ATOM_DEVICE_DFP5_SUPPORT                  0x0800
+                                  ATOM_DEVICE_LCD1_SUPPORT                  0x0002
+
+ulVBIOSMisc:      	              Miscellenous flags for VBIOS requirement and interface 
+                                  bit[0]=0: INT15 callback function Get LCD EDID ( ax=4e08, bl=1b ) is not supported by SBIOS. 
+                                        =1: INT15 callback function Get LCD EDID ( ax=4e08, bl=1b ) is supported by SBIOS.
+                                  bit[1]=0: INT15 callback function Get boot display( ax=4e08, bl=01h) is not supported by SBIOS
+                                        =1: INT15 callback function Get boot display( ax=4e08, bl=01h) is supported by SBIOS
+                                  bit[2]=0: INT15 callback function Get panel Expansion ( ax=4e08, bl=02h) is not supported by SBIOS
+                                        =1: INT15 callback function Get panel Expansion ( ax=4e08, bl=02h) is supported by SBIOS
+                                  bit[3]=0: VBIOS fast boot is disable
+                                        =1: VBIOS fast boot is enable. ( VBIOS skip display device detection in every set mode if LCD panel is connect and LID is open)
+
+ulGPUCapInfo:                     bit[0~2]= Reserved
+                                  bit[3]=0: Enable AUX HW mode detection logic
+                                        =1: Disable AUX HW mode detection logic
+                                  bit[4]=0: Disable DFS bypass feature
+                                        =1: Enable DFS bypass feature
+
+usRequestedPWMFreqInHz:           When it's set to 0x0 by SBIOS: the LCD BackLight is not controlled by GPU(SW). 
+                                  Any attempt to change BL using VBIOS function or enable VariBri from PP table is not effective since ATOM_BIOS_INFO_BL_CONTROLLED_BY_GPU==0;
+                                  
+                                  When it's set to a non-zero frequency, the BackLight is controlled by GPU (SW) in one of two ways below:
+                                  1. SW uses the GPU BL PWM output to control the BL, in chis case, this non-zero frequency determines what freq GPU should use;
+                                  VBIOS will set up proper PWM frequency and ATOM_BIOS_INFO_BL_CONTROLLED_BY_GPU==1,as the result,
+                                  Changing BL using VBIOS function is functional in both driver and non-driver present environment; 
+                                  and enabling VariBri under the driver environment from PP table is optional.
+
+                                  2. SW uses other means to control BL (like DPCD),this non-zero frequency serves as a flag only indicating
+                                  that BL control from GPU is expected.
+                                  VBIOS will NOT set up PWM frequency but make ATOM_BIOS_INFO_BL_CONTROLLED_BY_GPU==1
+                                  Changing BL using VBIOS function could be functional in both driver and non-driver present environment,but
+                                  it's per platform 
+                                  and enabling VariBri under the driver environment from PP table is optional.
+
+ucHtcTmpLmt:                      Refer to D18F3x64 bit[22:16], HtcTmpLmt. Threshold on value to enter HTC_active state.
+ucHtcHystLmt:                     Refer to D18F3x64 bit[27:24], HtcHystLmt. 
+                                  To calculate threshold off value to exit HTC_active state, which is Threshold on vlaue minus ucHtcHystLmt.
+
+ulSystemConfig:                   Bit[0]=0: PCIE Power Gating Disabled 
+                                        =1: PCIE Power Gating Enabled
+                                  Bit[1]=0: DDR-DLL shut-down feature disabled.
+                                         1: DDR-DLL shut-down feature enabled.
+                                  Bit[2]=0: DDR-PLL Power down feature disabled.
+                                         1: DDR-PLL Power down feature enabled. 
+                                  Bit[3]=0: GNB DPM is disabled
+                                        =1: GNB DPM is enabled                                
+ulCPUCapInfo:                     TBD
+
+usExtDispConnInfoOffset:          Offset to sExtDispConnInfo inside the structure
+usPanelRefreshRateRange:          Bit vector for LCD supported refresh rate range. If DRR is requestd by the platform, at least two bits need to be set
+                                  to indicate a range.
+                                  SUPPORTED_LCD_REFRESHRATE_30Hz          0x0004
+                                  SUPPORTED_LCD_REFRESHRATE_40Hz          0x0008
+                                  SUPPORTED_LCD_REFRESHRATE_50Hz          0x0010
+                                  SUPPORTED_LCD_REFRESHRATE_60Hz          0x0020
+
+ucMemoryType:                     [3:0]=1:DDR1;=2:DDR2;=3:DDR3;=5:GDDR5; [7:4] is reserved.
+ucUMAChannelNumber:      	        System memory channel numbers. 
+
+strVBIOSMsg[40]:                  VBIOS boot up customized message string 
+
+sAvail_SCLK[5]:                   Arrays to provide availabe list of SLCK and corresponding voltage, order from low to high  
+
+ulGMCRestoreResetTime:            GMC power restore and GMC reset time to calculate data reconnection latency. Unit in ns. 
+ulIdleNClk:                       NCLK speed while memory runs in self-refresh state, used to calculate self-refresh latency. Unit in 10kHz.
+ulDDR_DLL_PowerUpTime:            DDR PHY DLL power up time. Unit in ns.
+ulDDR_PLL_PowerUpTime:            DDR PHY PLL power up time. Unit in ns.
+
+usPCIEClkSSPercentage:            PCIE Clock Spread Spectrum Percentage in unit 0.01%; 100 mean 1%.
+usPCIEClkSSType:                  PCIE Clock Spread Spectrum Type. 0 for Down spread(default); 1 for Center spread.
+usLvdsSSPercentage:               LVDS panel ( not include eDP ) Spread Spectrum Percentage in unit of 0.01%, =0, use VBIOS default setting. 
+usLvdsSSpreadRateIn10Hz:          LVDS panel ( not include eDP ) Spread Spectrum frequency in unit of 10Hz, =0, use VBIOS default setting. 
+usHDMISSPercentage:               HDMI Spread Spectrum Percentage in unit 0.01%; 100 mean 1%,  =0, use VBIOS default setting. 
+usHDMISSpreadRateIn10Hz:          HDMI Spread Spectrum frequency in unit of 10Hz,  =0, use VBIOS default setting. 
+usDVISSPercentage:                DVI Spread Spectrum Percentage in unit 0.01%; 100 mean 1%,  =0, use VBIOS default setting. 
+usDVISSpreadRateIn10Hz:           DVI Spread Spectrum frequency in unit of 10Hz,  =0, use VBIOS default setting. 
+
+usGPUReservedSysMemSize:          Reserved system memory size for ACP engine in APU GNB, units in MB. 0/2/4MB based on CMOS options, current default could be 0MB. KV only, not on KB.
+ulGPUReservedSysMemBaseAddrLo:    Low 32 bits base address to the reserved system memory. 
+ulGPUReservedSysMemBaseAddrHi:    High 32 bits base address to the reserved system memory. 
+
+usMaxLVDSPclkFreqInSingleLink:    Max pixel clock LVDS panel single link, if=0 means VBIOS use default threhold, right now it is 85Mhz
+ucLVDSMisc:                       [bit0] LVDS 888bit panel mode =0: LVDS 888 panel in LDI mode, =1: LVDS 888 panel in FPDI mode
+                                  [bit1] LVDS panel lower and upper link mapping =0: lower link and upper link not swap, =1: lower link and upper link are swapped
+                                  [bit2] LVDS 888bit per color mode  =0: 666 bit per color =1:888 bit per color
+                                  [bit3] LVDS parameter override enable  =0: ucLvdsMisc parameter are not used =1: ucLvdsMisc parameter should be used
+                                  [bit4] Polarity of signal sent to digital BLON output pin. =0: not inverted(active high) =1: inverted ( active low )
+                                  [bit5] Travid LVDS output voltage override enable, when =1, use ucTravisLVDSVolAdjust value to overwrite Traivs register LVDS_CTRL_4
+ucTravisLVDSVolAdjust             When ucLVDSMisc[5]=1,it means platform SBIOS want to overwrite TravisLVDSVoltage. Then VBIOS will use ucTravisLVDSVolAdjust 
+                                  value to program Travis register LVDS_CTRL_4
+ucLVDSPwrOnSeqDIGONtoDE_in4Ms:    
+                                  LVDS power up sequence time in unit of 4ms, time delay from DIGON signal active to data enable signal active( DE ).
+                                  =0 mean use VBIOS default which is 8 ( 32ms ). The LVDS power up sequence is as following: DIGON->DE->VARY_BL->BLON. 
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucLVDSPwrOnDEtoVARY_BL_in4Ms:     
+                                  LVDS power up sequence time in unit of 4ms., time delay from DE( data enable ) active to Vary Brightness enable signal active( VARY_BL ).  
+                                  =0 mean use VBIOS default which is 90 ( 360ms ). The LVDS power up sequence is as following: DIGON->DE->VARY_BL->BLON. 
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucLVDSPwrOffVARY_BLtoDE_in4Ms:    
+                                  LVDS power down sequence time in unit of 4ms, time delay from data enable ( DE ) signal off to LCDVCC (DIGON) off. 
+                                  =0 mean use VBIOS default delay which is 8 ( 32ms ). The LVDS power down sequence is as following: BLON->VARY_BL->DE->DIGON
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucLVDSPwrOffDEtoDIGON_in4Ms:      
+                                   LVDS power down sequence time in unit of 4ms, time delay from vary brightness enable signal( VARY_BL) off to data enable ( DE ) signal off. 
+                                  =0 mean use VBIOS default which is 90 ( 360ms ). The LVDS power down sequence is as following: BLON->VARY_BL->DE->DIGON
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucLVDSOffToOnDelay_in4Ms:         
+                                  LVDS power down sequence time in unit of 4ms. Time delay from DIGON signal off to DIGON signal active. 
+                                  =0 means to use VBIOS default delay which is 125 ( 500ms ).
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucLVDSPwrOnSeqVARY_BLtoBLON_in4Ms:
+                                  LVDS power up sequence time in unit of 4ms. Time delay from VARY_BL signal on to DLON signal active. 
+                                  =0 means to use VBIOS default delay which is 0 ( 0ms ).
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+
+ucLVDSPwrOffSeqBLONtoVARY_BL_in4Ms:  
+                                  LVDS power down sequence time in unit of 4ms. Time delay from BLON signal off to VARY_BL signal off. 
+                                  =0 means to use VBIOS default delay which is 0 ( 0ms ).
+                                  This parameter is used by VBIOS only. VBIOS will patch LVDS_InfoTable.
+ucMinAllowedBL_Level:             Lowest LCD backlight PWM level. This is customer platform specific parameters. By default it is 0. 
+
+ulLCDBitDepthControlVal:          GPU display control encoder bit dither control setting, used to program register mmFMT_BIT_DEPTH_CONTROL
+
+ulNbpStateMemclkFreq[4]:          system memory clock frequncey in unit of 10Khz in different NB P-State(P0, P1, P2 & P3).
+ulNbpStateNClkFreq[4]:            NB P-State NClk frequency in different NB P-State
+usNBPStateVoltage[4]:             NB P-State (P0/P1 & P2/P3) voltage; NBP3 refers to lowes voltage
+usBootUpNBVoltage:                NB P-State voltage during boot up before driver loaded 
+sExtDispConnInfo:                 Display connector information table provided to VBIOS
+
+**********************************************************************************************************************/
+
+// this Table is used for Kaveri/Kabini APU
+typedef struct _ATOM_FUSION_SYSTEM_INFO_V2
+{
+  ATOM_INTEGRATED_SYSTEM_INFO_V1_8    sIntegratedSysInfo;       // refer to ATOM_INTEGRATED_SYSTEM_INFO_V1_8 definition
+  ULONG                               ulPowerplayTable[128];    // Update comments here to link new powerplay table definition structure
+}ATOM_FUSION_SYSTEM_INFO_V2; 
+
+
 /**************************************************************************/
 // This portion is only used when ext thermal chip or engine/memory clock SS chip is populated on a design
 //Memory SS Info Table
@@ -5026,22 +5380,24 @@ typedef struct _ATOM_ASIC_SS_ASSIGNMENT
 
 //Define ucClockIndication, SW uses the IDs below to search if the SS is required/enabled on a clock branch/signal type.
 //SS is not required or enabled if a match is not found.
-#define ASIC_INTERNAL_MEMORY_SS			1
-#define ASIC_INTERNAL_ENGINE_SS			2
-#define ASIC_INTERNAL_UVD_SS        3
-#define ASIC_INTERNAL_SS_ON_TMDS    4
-#define ASIC_INTERNAL_SS_ON_HDMI    5
-#define ASIC_INTERNAL_SS_ON_LVDS    6
-#define ASIC_INTERNAL_SS_ON_DP      7
-#define ASIC_INTERNAL_SS_ON_DCPLL   8
-#define ASIC_EXTERNAL_SS_ON_DP_CLOCK 9
-#define ASIC_INTERNAL_VCE_SS        10
+#define ASIC_INTERNAL_MEMORY_SS	         1
+#define ASIC_INTERNAL_ENGINE_SS	         2
+#define ASIC_INTERNAL_UVD_SS             3
+#define ASIC_INTERNAL_SS_ON_TMDS         4
+#define ASIC_INTERNAL_SS_ON_HDMI         5
+#define ASIC_INTERNAL_SS_ON_LVDS         6
+#define ASIC_INTERNAL_SS_ON_DP           7
+#define ASIC_INTERNAL_SS_ON_DCPLL        8
+#define ASIC_EXTERNAL_SS_ON_DP_CLOCK     9
+#define ASIC_INTERNAL_VCE_SS             10
+#define ASIC_INTERNAL_GPUPLL_SS          11
+
 
 typedef struct _ATOM_ASIC_SS_ASSIGNMENT_V2
 {
 	ULONG								ulTargetClockRange;						//For mem/engine/uvd, Clock Out frequence (VCO ), in unit of 10Khz
                                                     //For TMDS/HDMI/LVDS, it is pixel clock , for DP, it is link clock ( 27000 or 16200 )
-  USHORT              usSpreadSpectrumPercentage;		//in unit of 0.01%
+  USHORT              usSpreadSpectrumPercentage;		//in unit of 0.01% or 0.001%, decided by ucSpreadSpectrumMode bit4
 	USHORT							usSpreadRateIn10Hz;						//in unit of 10Hz, modulation freq
   UCHAR               ucClockIndication;					  //Indicate which clock source needs SS
 	UCHAR								ucSpreadSpectrumMode;					//Bit0=0 Down Spread,=1 Center Spread, bit1=0: internal SS bit1=1: external SS
@@ -5079,6 +5435,11 @@ typedef struct _ATOM_ASIC_SS_ASSIGNMENT_V3
 	UCHAR								ucReserved[2];
 }ATOM_ASIC_SS_ASSIGNMENT_V3;
 
+//ATOM_ASIC_SS_ASSIGNMENT_V3.ucSpreadSpectrumMode
+#define SS_MODE_V3_CENTRE_SPREAD_MASK             0x01
+#define SS_MODE_V3_EXTERNAL_SS_MASK               0x02
+#define SS_MODE_V3_PERCENTAGE_DIV_BY_1000_MASK    0x10
+
 typedef struct _ATOM_ASIC_INTERNAL_SS_INFO_V3
 {
   ATOM_COMMON_TABLE_HEADER	      sHeader; 
@@ -5719,6 +6080,7 @@ typedef struct _INDIRECT_IO_ACCESS
 #define INDIRECT_IO_PCIE           3
 #define INDIRECT_IO_PCIEP          4
 #define INDIRECT_IO_NBMISC         5
+#define INDIRECT_IO_SMU            5
 
 #define INDIRECT_IO_PLL_READ       INDIRECT_IO_PLL   | INDIRECT_READ
 #define INDIRECT_IO_PLL_WRITE      INDIRECT_IO_PLL   | INDIRECT_WRITE
@@ -5730,6 +6092,8 @@ typedef struct _INDIRECT_IO_ACCESS
 #define INDIRECT_IO_PCIEP_WRITE    INDIRECT_IO_PCIEP | INDIRECT_WRITE
 #define INDIRECT_IO_NBMISC_READ    INDIRECT_IO_NBMISC | INDIRECT_READ
 #define INDIRECT_IO_NBMISC_WRITE   INDIRECT_IO_NBMISC | INDIRECT_WRITE
+#define INDIRECT_IO_SMU_READ       INDIRECT_IO_SMU | INDIRECT_READ
+#define INDIRECT_IO_SMU_WRITE      INDIRECT_IO_SMU | INDIRECT_WRITE
 
 typedef struct _ATOM_OEM_INFO
 { 
@@ -5875,6 +6239,7 @@ typedef struct _ATOM_MC_INIT_PARAM_TABLE
 #define _64Mx32             0x43
 #define _128Mx8             0x51
 #define _128Mx16            0x52
+#define _128Mx32            0x53
 #define _256Mx8             0x61
 #define _256Mx16            0x62
 
@@ -5893,6 +6258,8 @@ typedef struct _ATOM_MC_INIT_PARAM_TABLE
 #define PROMOS              MOSEL
 #define KRETON              INFINEON
 #define ELIXIR              NANYA
+#define MEZZA               ELPIDA
+
 
 /////////////Support for GDDR5 MC uCode to reside in upper 64K of ROM/////////////
 
@@ -6625,6 +6992,10 @@ typedef struct _ATOM_DISP_OUT_INFO_V3
 	ASIC_TRANSMITTER_INFO_V2  asTransmitterInfo[1];     // for alligment only
 }ATOM_DISP_OUT_INFO_V3;
 
+//ucDispCaps
+#define DISPLAY_CAPS__DP_PCLK_FROM_PPLL        0x01
+#define DISPLAY_CAPS__FORCE_DISPDEV_CONNECTED  0x02
+
 typedef enum CORE_REF_CLK_SOURCE{
   CLOCK_SRC_XTALIN=0,
   CLOCK_SRC_XO_IN=1,
@@ -6829,6 +7200,17 @@ typedef struct _DIG_TRANSMITTER_INFO_HEADER_V3_1{
   USHORT usPhyPllSettingOffset;          // offset of CLOCK_CONDITION_SETTING_ENTRY* with Phy Pll Settings
 }DIG_TRANSMITTER_INFO_HEADER_V3_1;
 
+typedef struct _DIG_TRANSMITTER_INFO_HEADER_V3_2{  
+  ATOM_COMMON_TABLE_HEADER sHeader;  
+  USHORT usDPVsPreEmphSettingOffset;     // offset of PHY_ANALOG_SETTING_INFO * with DP Voltage Swing and Pre-Emphasis for each Link clock 
+  USHORT usPhyAnalogRegListOffset;       // offset of CLOCK_CONDITION_REGESTER_INFO* with None-DP mode Analog Setting's register Info 
+  USHORT usPhyAnalogSettingOffset;       // offset of CLOCK_CONDITION_SETTING_ENTRY* with None-DP mode Analog Setting for each link clock range
+  USHORT usPhyPllRegListOffset;          // offset of CLOCK_CONDITION_REGESTER_INFO* with Phy Pll register Info 
+  USHORT usPhyPllSettingOffset;          // offset of CLOCK_CONDITION_SETTING_ENTRY* with Phy Pll Settings
+  USHORT usDPSSRegListOffset;            // offset of CLOCK_CONDITION_REGESTER_INFO* with Phy SS Pll register Info 
+  USHORT usDPSSSettingOffset;            // offset of CLOCK_CONDITION_SETTING_ENTRY* with Phy SS Pll Settings
+}DIG_TRANSMITTER_INFO_HEADER_V3_2;
+
 typedef struct _CLOCK_CONDITION_REGESTER_INFO{
   USHORT usRegisterIndex;
   UCHAR  ucStartBit;
@@ -6852,12 +7234,24 @@ typedef struct _PHY_CONDITION_REG_VAL{
   ULONG  ulRegVal;
 }PHY_CONDITION_REG_VAL;
 
+typedef struct _PHY_CONDITION_REG_VAL_V2{
+  ULONG  ulCondition;
+  UCHAR  ucCondition2;
+  ULONG  ulRegVal;
+}PHY_CONDITION_REG_VAL_V2;
+
 typedef struct _PHY_CONDITION_REG_INFO{
   USHORT usRegIndex;
   USHORT usSize;
   PHY_CONDITION_REG_VAL asRegVal[1];
 }PHY_CONDITION_REG_INFO;
 
+typedef struct _PHY_CONDITION_REG_INFO_V2{
+  USHORT usRegIndex;
+  USHORT usSize;
+  PHY_CONDITION_REG_VAL_V2 asRegVal[1];
+}PHY_CONDITION_REG_INFO_V2;
+
 typedef struct _PHY_ANALOG_SETTING_INFO{
   UCHAR  ucEncodeMode;
   UCHAR  ucPhySel;
@@ -6865,6 +7259,25 @@ typedef struct _PHY_ANALOG_SETTING_INFO{
   PHY_CONDITION_REG_INFO  asAnalogSetting[1];
 }PHY_ANALOG_SETTING_INFO;
 
+typedef struct _PHY_ANALOG_SETTING_INFO_V2{
+  UCHAR  ucEncodeMode;
+  UCHAR  ucPhySel;
+  USHORT usSize;
+  PHY_CONDITION_REG_INFO_V2  asAnalogSetting[1];
+}PHY_ANALOG_SETTING_INFO_V2;
+
+typedef struct _GFX_HAVESTING_PARAMETERS {
+  UCHAR ucGfxBlkId;                        //GFX blk id to be harvested, like CU, RB or PRIM
+  UCHAR ucReserved;                        //reserved 
+  UCHAR ucActiveUnitNumPerSH;              //requested active CU/RB/PRIM number per shader array
+  UCHAR ucMaxUnitNumPerSH;                 //max CU/RB/PRIM number per shader array   
+} GFX_HAVESTING_PARAMETERS;
+
+//ucGfxBlkId
+#define GFX_HARVESTING_CU_ID               0
+#define GFX_HARVESTING_RB_ID               1
+#define GFX_HARVESTING_PRIM_ID             2
+
 /****************************************************************************/	
 //Portion VI: Definitinos for vbios MC scratch registers that driver used
 /****************************************************************************/
@@ -6875,8 +7288,17 @@ typedef struct _PHY_ANALOG_SETTING_INFO{
 #define MC_MISC0__MEMORY_TYPE__GDDR3  0x30000000
 #define MC_MISC0__MEMORY_TYPE__GDDR4  0x40000000
 #define MC_MISC0__MEMORY_TYPE__GDDR5  0x50000000
+#define MC_MISC0__MEMORY_TYPE__HBM    0x60000000
 #define MC_MISC0__MEMORY_TYPE__DDR3   0xB0000000
 
+#define ATOM_MEM_TYPE_DDR_STRING      "DDR"
+#define ATOM_MEM_TYPE_DDR2_STRING     "DDR2"
+#define ATOM_MEM_TYPE_GDDR3_STRING    "GDDR3"
+#define ATOM_MEM_TYPE_GDDR4_STRING    "GDDR4"
+#define ATOM_MEM_TYPE_GDDR5_STRING    "GDDR5"
+#define ATOM_MEM_TYPE_HBM_STRING      "HBM"
+#define ATOM_MEM_TYPE_DDR3_STRING     "DDR3"
+
 /****************************************************************************/	
 //Portion VI: Definitinos being oboselete
 /****************************************************************************/
@@ -7274,6 +7696,7 @@ typedef struct _ATOM_PPLIB_THERMALCONTROLLER
 #define ATOM_PP_THERMALCONTROLLER_NISLANDS  15
 #define ATOM_PP_THERMALCONTROLLER_SISLANDS  16
 #define ATOM_PP_THERMALCONTROLLER_LM96163   17
+#define ATOM_PP_THERMALCONTROLLER_CISLANDS  18
 
 // Thermal controller 'combo type' to use an external controller for Fan control and an internal controller for thermal.
 // We probably should reserve the bit 0x80 for this use.
@@ -7316,6 +7739,8 @@ typedef struct _ATOM_PPLIB_EXTENDEDHEADER
     // Add extra system parameters here, always adjust size to include all fields.
     USHORT  usVCETableOffset; //points to ATOM_PPLIB_VCE_Table
     USHORT  usUVDTableOffset;   //points to ATOM_PPLIB_UVD_Table
+    USHORT  usSAMUTableOffset;  //points to ATOM_PPLIB_SAMU_Table
+    USHORT  usPPMTableOffset;   //points to ATOM_PPLIB_PPM_Table
 } ATOM_PPLIB_EXTENDEDHEADER;
 
 //// ATOM_PPLIB_POWERPLAYTABLE::ulPlatformCaps
@@ -7337,7 +7762,10 @@ typedef struct _ATOM_PPLIB_EXTENDEDHEADER
 #define ATOM_PP_PLATFORM_CAP_VDDCI_CONTROL 0x8000                   // Does the driver control VDDCI independently from VDDC.
 #define ATOM_PP_PLATFORM_CAP_REGULATOR_HOT 0x00010000               // Enable the 'regulator hot' feature.
 #define ATOM_PP_PLATFORM_CAP_BACO          0x00020000               // Does the driver supports BACO state.
-
+#define ATOM_PP_PLATFORM_CAP_NEW_CAC_VOLTAGE   0x00040000           // Does the driver supports new CAC voltage table.
+#define ATOM_PP_PLATFORM_CAP_REVERT_GPIO5_POLARITY   0x00080000     // Does the driver supports revert GPIO5 polarity.
+#define ATOM_PP_PLATFORM_CAP_OUTPUT_THERMAL2GPIO17   0x00100000     // Does the driver supports thermal2GPIO17.
+#define ATOM_PP_PLATFORM_CAP_VRHOT_GPIO_CONFIGURABLE   0x00200000   // Does the driver supports VR HOT GPIO Configurable.
 
 typedef struct _ATOM_PPLIB_POWERPLAYTABLE
 {
@@ -7398,7 +7826,7 @@ typedef struct _ATOM_PPLIB_POWERPLAYTABLE4
     USHORT                     usVddcDependencyOnMCLKOffset;
     USHORT                     usMaxClockVoltageOnDCOffset;
     USHORT                     usVddcPhaseShedLimitsTableOffset;    // Points to ATOM_PPLIB_PhaseSheddingLimits_Table
-    USHORT                     usReserved;  
+    USHORT                     usMvddDependencyOnMCLKOffset;  
 } ATOM_PPLIB_POWERPLAYTABLE4, *LPATOM_PPLIB_POWERPLAYTABLE4;
 
 typedef struct _ATOM_PPLIB_POWERPLAYTABLE5
@@ -7563,6 +7991,17 @@ typedef struct _ATOM_PPLIB_SI_CLOCK_INFO
 
 } ATOM_PPLIB_SI_CLOCK_INFO;
 
+typedef struct _ATOM_PPLIB_CI_CLOCK_INFO
+{
+      USHORT usEngineClockLow;
+      UCHAR  ucEngineClockHigh;
+
+      USHORT usMemoryClockLow;
+      UCHAR  ucMemoryClockHigh;
+      
+      UCHAR  ucPCIEGen;
+      USHORT usPCIELane;
+} ATOM_PPLIB_CI_CLOCK_INFO;
 
 typedef struct _ATOM_PPLIB_RS780_CLOCK_INFO
 
@@ -7680,8 +8119,8 @@ typedef struct _ATOM_PPLIB_Clock_Voltage_Limit_Table
 
 typedef struct _ATOM_PPLIB_CAC_Leakage_Record
 {
-    USHORT usVddc;  // We use this field for the "fake" standardized VDDC for power calculations                                                  
-    ULONG  ulLeakageValue;
+    USHORT usVddc;  // We use this field for the "fake" standardized VDDC for power calculations; For CI and newer, we use this as the real VDDC value.
+    ULONG  ulLeakageValue;  // For CI and newer we use this as the "fake" standar VDDC value.
 }ATOM_PPLIB_CAC_Leakage_Record;
 
 typedef struct _ATOM_PPLIB_CAC_Leakage_Table
@@ -7796,6 +8235,42 @@ typedef struct _ATOM_PPLIB_UVD_Table
 //    ATOM_PPLIB_UVD_State_Table states;
 }ATOM_PPLIB_UVD_Table;
 
+
+typedef struct _ATOM_PPLIB_SAMClk_Voltage_Limit_Record
+{
+      USHORT usVoltage;
+      USHORT usSAMClockLow;
+      UCHAR  ucSAMClockHigh;
+}ATOM_PPLIB_SAMClk_Voltage_Limit_Record;
+
+typedef struct _ATOM_PPLIB_SAMClk_Voltage_Limit_Table{
+    UCHAR numEntries;
+    ATOM_PPLIB_SAMClk_Voltage_Limit_Record entries[1];
+}ATOM_PPLIB_SAMClk_Voltage_Limit_Table;
+
+typedef struct _ATOM_PPLIB_SAMU_Table
+{
+      UCHAR revid;
+      ATOM_PPLIB_SAMClk_Voltage_Limit_Table limits;
+}ATOM_PPLIB_SAMU_Table;
+
+#define ATOM_PPM_A_A    1
+#define ATOM_PPM_A_I    2
+typedef struct _ATOM_PPLIB_PPM_Table
+{
+      UCHAR  ucRevId;
+      UCHAR  ucPpmDesign;          //A+I or A+A
+      USHORT usCpuCoreNumber;
+      ULONG  ulPlatformTDP;
+      ULONG  ulSmallACPlatformTDP;
+      ULONG  ulPlatformTDC;
+      ULONG  ulSmallACPlatformTDC;
+      ULONG  ulApuTDP;
+      ULONG  ulDGpuTDP;  
+      ULONG  ulDGpuUlvPower;
+      ULONG  ulTjmax;
+} ATOM_PPLIB_PPM_Table;
+
 /**************************************************************************/
 
 
diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c
index d5df8fd10217..b9d3b43f19c0 100644
--- a/drivers/gpu/drm/radeon/atombios_crtc.c
+++ b/drivers/gpu/drm/radeon/atombios_crtc.c
@@ -555,7 +555,7 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc,
 		if (rdev->family < CHIP_RV770)
 			radeon_crtc->pll_flags |= RADEON_PLL_PREFER_MINM_OVER_MAXP;
 		/* use frac fb div on APUs */
-		if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev))
+		if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE61(rdev) || ASIC_IS_DCE8(rdev))
 			radeon_crtc->pll_flags |= RADEON_PLL_USE_FRAC_FB_DIV;
 		/* use frac fb div on RS780/RS880 */
 		if ((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880))
@@ -743,7 +743,7 @@ static void atombios_crtc_set_disp_eng_pll(struct radeon_device *rdev,
 			 * SetPixelClock provides the dividers
 			 */
 			args.v6.ulDispEngClkFreq = cpu_to_le32(dispclk);
-			if (ASIC_IS_DCE61(rdev))
+			if (ASIC_IS_DCE61(rdev) || ASIC_IS_DCE8(rdev))
 				args.v6.ucPpll = ATOM_EXT_PLL1;
 			else if (ASIC_IS_DCE6(rdev))
 				args.v6.ucPpll = ATOM_PPLL0;
@@ -1143,7 +1143,9 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc,
 	}
 
 	if (tiling_flags & RADEON_TILING_MACRO) {
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			tmp = rdev->config.cik.tile_config;
+		else if (rdev->family >= CHIP_TAHITI)
 			tmp = rdev->config.si.tile_config;
 		else if (rdev->family >= CHIP_CAYMAN)
 			tmp = rdev->config.cayman.tile_config;
@@ -1170,11 +1172,29 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc,
 		fb_format |= EVERGREEN_GRPH_BANK_WIDTH(bankw);
 		fb_format |= EVERGREEN_GRPH_BANK_HEIGHT(bankh);
 		fb_format |= EVERGREEN_GRPH_MACRO_TILE_ASPECT(mtaspect);
+		if (rdev->family >= CHIP_BONAIRE) {
+			/* XXX need to know more about the surface tiling mode */
+			fb_format |= CIK_GRPH_MICRO_TILE_MODE(CIK_DISPLAY_MICRO_TILING);
+		}
 	} else if (tiling_flags & RADEON_TILING_MICRO)
 		fb_format |= EVERGREEN_GRPH_ARRAY_MODE(EVERGREEN_GRPH_ARRAY_1D_TILED_THIN1);
 
-	if ((rdev->family == CHIP_TAHITI) ||
-	    (rdev->family == CHIP_PITCAIRN))
+	if (rdev->family >= CHIP_BONAIRE) {
+		u32 num_pipe_configs = rdev->config.cik.max_tile_pipes;
+		u32 num_rb = rdev->config.cik.max_backends_per_se;
+		if (num_pipe_configs > 8)
+			num_pipe_configs = 8;
+		if (num_pipe_configs == 8)
+			fb_format |= CIK_GRPH_PIPE_CONFIG(CIK_ADDR_SURF_P8_32x32_16x16);
+		else if (num_pipe_configs == 4) {
+			if (num_rb == 4)
+				fb_format |= CIK_GRPH_PIPE_CONFIG(CIK_ADDR_SURF_P4_16x16);
+			else if (num_rb < 4)
+				fb_format |= CIK_GRPH_PIPE_CONFIG(CIK_ADDR_SURF_P4_8x16);
+		} else if (num_pipe_configs == 2)
+			fb_format |= CIK_GRPH_PIPE_CONFIG(CIK_ADDR_SURF_P2);
+	} else if ((rdev->family == CHIP_TAHITI) ||
+		   (rdev->family == CHIP_PITCAIRN))
 		fb_format |= SI_GRPH_PIPE_CONFIG(SI_ADDR_SURF_P8_32x32_8x16);
 	else if (rdev->family == CHIP_VERDE)
 		fb_format |= SI_GRPH_PIPE_CONFIG(SI_ADDR_SURF_P4_8x16);
@@ -1224,8 +1244,12 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc,
 	WREG32(EVERGREEN_GRPH_PITCH + radeon_crtc->crtc_offset, fb_pitch_pixels);
 	WREG32(EVERGREEN_GRPH_ENABLE + radeon_crtc->crtc_offset, 1);
 
-	WREG32(EVERGREEN_DESKTOP_HEIGHT + radeon_crtc->crtc_offset,
-	       target_fb->height);
+	if (rdev->family >= CHIP_BONAIRE)
+		WREG32(CIK_LB_DESKTOP_HEIGHT + radeon_crtc->crtc_offset,
+		       target_fb->height);
+	else
+		WREG32(EVERGREEN_DESKTOP_HEIGHT + radeon_crtc->crtc_offset,
+		       target_fb->height);
 	x &= ~3;
 	y &= ~1;
 	WREG32(EVERGREEN_VIEWPORT_START + radeon_crtc->crtc_offset,
@@ -1597,6 +1621,12 @@ static int radeon_get_shared_nondp_ppll(struct drm_crtc *crtc)
  *
  * Asic specific PLL information
  *
+ * DCE 8.x
+ * KB/KV
+ * - PPLL1, PPLL2 are available for all UNIPHY (both DP and non-DP)
+ * CI
+ * - PPLL0, PPLL1, PPLL2 are available for all UNIPHY (both DP and non-DP) and DAC
+ *
  * DCE 6.1
  * - PPLL2 is only available to UNIPHYA (both DP and non-DP)
  * - PPLL0, PPLL1 are available for UNIPHYB/C/D/E/F (both DP and non-DP)
@@ -1623,7 +1653,47 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc)
 	u32 pll_in_use;
 	int pll;
 
-	if (ASIC_IS_DCE61(rdev)) {
+	if (ASIC_IS_DCE8(rdev)) {
+		if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(radeon_crtc->encoder))) {
+			if (rdev->clock.dp_extclk)
+				/* skip PPLL programming if using ext clock */
+				return ATOM_PPLL_INVALID;
+			else {
+				/* use the same PPLL for all DP monitors */
+				pll = radeon_get_shared_dp_ppll(crtc);
+				if (pll != ATOM_PPLL_INVALID)
+					return pll;
+			}
+		} else {
+			/* use the same PPLL for all monitors with the same clock */
+			pll = radeon_get_shared_nondp_ppll(crtc);
+			if (pll != ATOM_PPLL_INVALID)
+				return pll;
+		}
+		/* otherwise, pick one of the plls */
+		if ((rdev->family == CHIP_KAVERI) ||
+		    (rdev->family == CHIP_KABINI)) {
+			/* KB/KV has PPLL1 and PPLL2 */
+			pll_in_use = radeon_get_pll_use_mask(crtc);
+			if (!(pll_in_use & (1 << ATOM_PPLL2)))
+				return ATOM_PPLL2;
+			if (!(pll_in_use & (1 << ATOM_PPLL1)))
+				return ATOM_PPLL1;
+			DRM_ERROR("unable to allocate a PPLL\n");
+			return ATOM_PPLL_INVALID;
+		} else {
+			/* CI has PPLL0, PPLL1, and PPLL2 */
+			pll_in_use = radeon_get_pll_use_mask(crtc);
+			if (!(pll_in_use & (1 << ATOM_PPLL2)))
+				return ATOM_PPLL2;
+			if (!(pll_in_use & (1 << ATOM_PPLL1)))
+				return ATOM_PPLL1;
+			if (!(pll_in_use & (1 << ATOM_PPLL0)))
+				return ATOM_PPLL0;
+			DRM_ERROR("unable to allocate a PPLL\n");
+			return ATOM_PPLL_INVALID;
+		}
+	} else if (ASIC_IS_DCE61(rdev)) {
 		struct radeon_encoder_atom_dig *dig =
 			radeon_encoder->enc_priv;
 
@@ -1771,6 +1841,9 @@ int atombios_crtc_mode_set(struct drm_crtc *crtc,
 	atombios_crtc_set_base(crtc, x, y, old_fb);
 	atombios_overscan_setup(crtc, mode, adjusted_mode);
 	atombios_scaler_setup(crtc);
+	/* update the hw version fpr dpm */
+	radeon_crtc->hw_mode = *adjusted_mode;
+
 	return 0;
 }
 
@@ -1861,7 +1934,7 @@ static void atombios_crtc_disable(struct drm_crtc *crtc)
 		break;
 	case ATOM_PPLL0:
 		/* disable the ppll */
-		if (ASIC_IS_DCE61(rdev))
+		if ((rdev->family == CHIP_ARUBA) || (rdev->family == CHIP_BONAIRE))
 			atombios_crtc_program_pll(crtc, radeon_crtc->crtc_id, radeon_crtc->pll_id,
 						  0, 0, ATOM_DISABLE, 0, 0, 0, 0, 0, false, &ss);
 		break;
diff --git a/drivers/gpu/drm/radeon/atombios_encoders.c b/drivers/gpu/drm/radeon/atombios_encoders.c
index 8406c8251fbf..092275d53d4a 100644
--- a/drivers/gpu/drm/radeon/atombios_encoders.c
+++ b/drivers/gpu/drm/radeon/atombios_encoders.c
@@ -186,6 +186,13 @@ void radeon_atom_backlight_init(struct radeon_encoder *radeon_encoder,
 	u8 backlight_level;
 	char bl_name[16];
 
+	/* Mac laptops with multiple GPUs use the gmux driver for backlight
+	 * so don't register a backlight device
+	 */
+	if ((rdev->pdev->subsystem_vendor == PCI_VENDOR_ID_APPLE) &&
+	    (rdev->pdev->device == 0x6741))
+		return;
+
 	if (!radeon_encoder->enc_priv)
 		return;
 
@@ -296,6 +303,7 @@ static inline bool radeon_encoder_is_digital(struct drm_encoder *encoder)
 	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 		return true;
 	default:
 		return false;
@@ -479,11 +487,11 @@ static u8 radeon_atom_get_bpc(struct drm_encoder *encoder)
 	}
 }
 
-
 union dvo_encoder_control {
 	ENABLE_EXTERNAL_TMDS_ENCODER_PS_ALLOCATION ext_tmds;
 	DVO_ENCODER_CONTROL_PS_ALLOCATION dvo;
 	DVO_ENCODER_CONTROL_PS_ALLOCATION_V3 dvo_v3;
+	DVO_ENCODER_CONTROL_PS_ALLOCATION_V1_4 dvo_v4;
 };
 
 void
@@ -533,6 +541,13 @@ atombios_dvo_setup(struct drm_encoder *encoder, int action)
 			args.dvo_v3.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
 			args.dvo_v3.ucDVOConfig = 0; /* XXX */
 			break;
+		case 4:
+			/* DCE8 */
+			args.dvo_v4.ucAction = action;
+			args.dvo_v4.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+			args.dvo_v4.ucDVOConfig = 0; /* XXX */
+			args.dvo_v4.ucBitPerColor = radeon_atom_get_bpc(encoder);
+			break;
 		default:
 			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
 			break;
@@ -915,10 +930,14 @@ atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mo
 				args.v4.ucLaneNum = 4;
 
 			if (ENCODER_MODE_IS_DP(args.v4.ucEncoderMode)) {
-				if (dp_clock == 270000)
-					args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_2_70GHZ;
-				else if (dp_clock == 540000)
+				if (dp_clock == 540000)
 					args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_5_40GHZ;
+				else if (dp_clock == 324000)
+					args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_3_24GHZ;
+				else if (dp_clock == 270000)
+					args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_2_70GHZ;
+				else
+					args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_1_62GHZ;
 			}
 			args.v4.acConfig.ucDigSel = dig->dig_encoder;
 			args.v4.ucBitPerColor = radeon_atom_get_bpc(encoder);
@@ -1012,6 +1031,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 		index = GetIndexIntoMasterTable(COMMAND, UNIPHYTransmitterControl);
 		break;
 	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
@@ -1271,6 +1291,9 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t
 				else
 					args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYE;
 				break;
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
+				args.v5.ucPhyId = ATOM_PHY_ID_UNIPHYG;
+				break;
 			}
 			if (is_dp)
 				args.v5.ucLaneNum = dp_lane_count;
@@ -1735,6 +1758,7 @@ radeon_atom_encoder_dpms(struct drm_encoder *encoder, int mode)
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 		radeon_atom_encoder_dpms_dig(encoder, mode);
 		break;
@@ -1872,6 +1896,7 @@ atombios_set_encoder_crtc_source(struct drm_encoder *encoder)
 			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 				dig = radeon_encoder->enc_priv;
 				switch (dig->dig_encoder) {
@@ -1893,6 +1918,9 @@ atombios_set_encoder_crtc_source(struct drm_encoder *encoder)
 				case 5:
 					args.v2.ucEncoderID = ASIC_INT_DIG6_ENCODER_ID;
 					break;
+				case 6:
+					args.v2.ucEncoderID = ASIC_INT_DIG7_ENCODER_ID;
+					break;
 				}
 				break;
 			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
@@ -1955,7 +1983,13 @@ atombios_apply_encoder_quirks(struct drm_encoder *encoder,
 	/* set scaler clears this on some chips */
 	if (ASIC_IS_AVIVO(rdev) &&
 	    (!(radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)))) {
-		if (ASIC_IS_DCE4(rdev)) {
+		if (ASIC_IS_DCE8(rdev)) {
+			if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+				WREG32(CIK_LB_DATA_FORMAT + radeon_crtc->crtc_offset,
+				       CIK_INTERLEAVE_EN);
+			else
+				WREG32(CIK_LB_DATA_FORMAT + radeon_crtc->crtc_offset, 0);
+		} else if (ASIC_IS_DCE4(rdev)) {
 			if (mode->flags & DRM_MODE_FLAG_INTERLACE)
 				WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset,
 				       EVERGREEN_INTERLEAVE_EN);
@@ -2002,6 +2036,9 @@ static int radeon_atom_pick_dig_encoder(struct drm_encoder *encoder)
 			else
 				return 4;
 			break;
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
+			return 6;
+			break;
 		}
 	} else if (ASIC_IS_DCE4(rdev)) {
 		/* DCE4/5 */
@@ -2086,6 +2123,7 @@ radeon_atom_encoder_init(struct radeon_device *rdev)
 		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_INIT, 0, 0);
 			break;
@@ -2130,6 +2168,7 @@ radeon_atom_encoder_mode_set(struct drm_encoder *encoder,
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 		/* handled in dpms */
 		break;
@@ -2395,6 +2434,7 @@ static void radeon_atom_encoder_disable(struct drm_encoder *encoder)
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 		/* handled in dpms */
 		break;
@@ -2626,6 +2666,7 @@ radeon_add_atom_encoder(struct drm_device *dev,
 	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
 	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
 		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
 			radeon_encoder->rmx_type = RMX_FULL;
 			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_LVDS);
diff --git a/drivers/gpu/drm/radeon/btc_dpm.c b/drivers/gpu/drm/radeon/btc_dpm.c
new file mode 100644
index 000000000000..0bfd55e08820
--- /dev/null
+++ b/drivers/gpu/drm/radeon/btc_dpm.c
@@ -0,0 +1,2751 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "btcd.h"
+#include "r600_dpm.h"
+#include "cypress_dpm.h"
+#include "btc_dpm.h"
+#include "atom.h"
+
+#define MC_CG_ARB_FREQ_F0           0x0a
+#define MC_CG_ARB_FREQ_F1           0x0b
+#define MC_CG_ARB_FREQ_F2           0x0c
+#define MC_CG_ARB_FREQ_F3           0x0d
+
+#define MC_CG_SEQ_DRAMCONF_S0       0x05
+#define MC_CG_SEQ_DRAMCONF_S1       0x06
+#define MC_CG_SEQ_YCLK_SUSPEND      0x04
+#define MC_CG_SEQ_YCLK_RESUME       0x0a
+
+#define SMC_RAM_END 0x8000
+
+#ifndef BTC_MGCG_SEQUENCE
+#define BTC_MGCG_SEQUENCE  300
+
+struct rv7xx_ps *rv770_get_ps(struct radeon_ps *rps);
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev);
+struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev);
+
+
+//********* BARTS **************//
+static const u32 barts_cgcg_cgls_default[] =
+{
+	/* Register,   Value,     Mask bits */
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define BARTS_CGCG_CGLS_DEFAULT_LENGTH sizeof(barts_cgcg_cgls_default) / (3 * sizeof(u32))
+
+static const u32 barts_cgcg_cgls_disable[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00000644, 0x000f7912, 0x001f4180,
+	0x00000644, 0x000f3812, 0x001f4180
+};
+#define BARTS_CGCG_CGLS_DISABLE_LENGTH sizeof(barts_cgcg_cgls_disable) / (3 * sizeof(u32))
+
+static const u32 barts_cgcg_cgls_enable[] =
+{
+	/* 0x0000c124, 0x84180000, 0x00180000, */
+	0x00000644, 0x000f7892, 0x001f4080,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff
+};
+#define BARTS_CGCG_CGLS_ENABLE_LENGTH sizeof(barts_cgcg_cgls_enable) / (3 * sizeof(u32))
+
+static const u32 barts_mgcg_default[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00005448, 0x00000100, 0xffffffff,
+	0x000055e4, 0x00600100, 0xffffffff,
+	0x0000160c, 0x00000100, 0xffffffff,
+	0x0000c164, 0x00000100, 0xffffffff,
+	0x00008a18, 0x00000100, 0xffffffff,
+	0x0000897c, 0x06000100, 0xffffffff,
+	0x00008b28, 0x00000100, 0xffffffff,
+	0x00009144, 0x00000100, 0xffffffff,
+	0x00009a60, 0x00000100, 0xffffffff,
+	0x00009868, 0x00000100, 0xffffffff,
+	0x00008d58, 0x00000100, 0xffffffff,
+	0x00009510, 0x00000100, 0xffffffff,
+	0x0000949c, 0x00000100, 0xffffffff,
+	0x00009654, 0x00000100, 0xffffffff,
+	0x00009030, 0x00000100, 0xffffffff,
+	0x00009034, 0x00000100, 0xffffffff,
+	0x00009038, 0x00000100, 0xffffffff,
+	0x0000903c, 0x00000100, 0xffffffff,
+	0x00009040, 0x00000100, 0xffffffff,
+	0x0000a200, 0x00000100, 0xffffffff,
+	0x0000a204, 0x00000100, 0xffffffff,
+	0x0000a208, 0x00000100, 0xffffffff,
+	0x0000a20c, 0x00000100, 0xffffffff,
+	0x0000977c, 0x00000100, 0xffffffff,
+	0x00003f80, 0x00000100, 0xffffffff,
+	0x0000a210, 0x00000100, 0xffffffff,
+	0x0000a214, 0x00000100, 0xffffffff,
+	0x000004d8, 0x00000100, 0xffffffff,
+	0x00009784, 0x00000100, 0xffffffff,
+	0x00009698, 0x00000100, 0xffffffff,
+	0x000004d4, 0x00000200, 0xffffffff,
+	0x000004d0, 0x00000000, 0xffffffff,
+	0x000030cc, 0x00000100, 0xffffffff,
+	0x0000d0c0, 0xff000100, 0xffffffff,
+	0x0000802c, 0x40000000, 0xffffffff,
+	0x0000915c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091b0, 0x00070000, 0xffffffff,
+	0x000091b4, 0x00030002, 0xffffffff,
+	0x000091b8, 0x00050004, 0xffffffff,
+	0x000091c4, 0x00010006, 0xffffffff,
+	0x000091c8, 0x00090008, 0xffffffff,
+	0x000091cc, 0x00070000, 0xffffffff,
+	0x000091d0, 0x00030002, 0xffffffff,
+	0x000091d4, 0x00050004, 0xffffffff,
+	0x000091e0, 0x00010006, 0xffffffff,
+	0x000091e4, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x000091ec, 0x00070000, 0xffffffff,
+	0x000091f0, 0x00030002, 0xffffffff,
+	0x000091f4, 0x00050004, 0xffffffff,
+	0x00009200, 0x00010006, 0xffffffff,
+	0x00009204, 0x00090008, 0xffffffff,
+	0x00009208, 0x00070000, 0xffffffff,
+	0x0000920c, 0x00030002, 0xffffffff,
+	0x00009210, 0x00050004, 0xffffffff,
+	0x0000921c, 0x00010006, 0xffffffff,
+	0x00009220, 0x00090008, 0xffffffff,
+	0x00009224, 0x00070000, 0xffffffff,
+	0x00009228, 0x00030002, 0xffffffff,
+	0x0000922c, 0x00050004, 0xffffffff,
+	0x00009238, 0x00010006, 0xffffffff,
+	0x0000923c, 0x00090008, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff,
+	0x0000802c, 0x40010000, 0xffffffff,
+	0x0000915c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091b0, 0x00070000, 0xffffffff,
+	0x000091b4, 0x00030002, 0xffffffff,
+	0x000091b8, 0x00050004, 0xffffffff,
+	0x000091c4, 0x00010006, 0xffffffff,
+	0x000091c8, 0x00090008, 0xffffffff,
+	0x000091cc, 0x00070000, 0xffffffff,
+	0x000091d0, 0x00030002, 0xffffffff,
+	0x000091d4, 0x00050004, 0xffffffff,
+	0x000091e0, 0x00010006, 0xffffffff,
+	0x000091e4, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x000091ec, 0x00070000, 0xffffffff,
+	0x000091f0, 0x00030002, 0xffffffff,
+	0x000091f4, 0x00050004, 0xffffffff,
+	0x00009200, 0x00010006, 0xffffffff,
+	0x00009204, 0x00090008, 0xffffffff,
+	0x00009208, 0x00070000, 0xffffffff,
+	0x0000920c, 0x00030002, 0xffffffff,
+	0x00009210, 0x00050004, 0xffffffff,
+	0x0000921c, 0x00010006, 0xffffffff,
+	0x00009220, 0x00090008, 0xffffffff,
+	0x00009224, 0x00070000, 0xffffffff,
+	0x00009228, 0x00030002, 0xffffffff,
+	0x0000922c, 0x00050004, 0xffffffff,
+	0x00009238, 0x00010006, 0xffffffff,
+	0x0000923c, 0x00090008, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff,
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define BARTS_MGCG_DEFAULT_LENGTH sizeof(barts_mgcg_default) / (3 * sizeof(u32))
+
+static const u32 barts_mgcg_disable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x00009150, 0x00600000, 0xffffffff
+};
+#define BARTS_MGCG_DISABLE_LENGTH sizeof(barts_mgcg_disable) / (3 * sizeof(u32))
+
+static const u32 barts_mgcg_enable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00009150, 0x81944000, 0xffffffff
+};
+#define BARTS_MGCG_ENABLE_LENGTH sizeof(barts_mgcg_enable) / (3 * sizeof(u32))
+
+//********* CAICOS **************//
+static const u32 caicos_cgcg_cgls_default[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define CAICOS_CGCG_CGLS_DEFAULT_LENGTH sizeof(caicos_cgcg_cgls_default) / (3 * sizeof(u32))
+
+static const u32 caicos_cgcg_cgls_disable[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00000644, 0x000f7912, 0x001f4180,
+	0x00000644, 0x000f3812, 0x001f4180
+};
+#define CAICOS_CGCG_CGLS_DISABLE_LENGTH sizeof(caicos_cgcg_cgls_disable) / (3 * sizeof(u32))
+
+static const u32 caicos_cgcg_cgls_enable[] =
+{
+	/* 0x0000c124, 0x84180000, 0x00180000, */
+	0x00000644, 0x000f7892, 0x001f4080,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff
+};
+#define CAICOS_CGCG_CGLS_ENABLE_LENGTH sizeof(caicos_cgcg_cgls_enable) / (3 * sizeof(u32))
+
+static const u32 caicos_mgcg_default[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00005448, 0x00000100, 0xffffffff,
+	0x000055e4, 0x00600100, 0xffffffff,
+	0x0000160c, 0x00000100, 0xffffffff,
+	0x0000c164, 0x00000100, 0xffffffff,
+	0x00008a18, 0x00000100, 0xffffffff,
+	0x0000897c, 0x06000100, 0xffffffff,
+	0x00008b28, 0x00000100, 0xffffffff,
+	0x00009144, 0x00000100, 0xffffffff,
+	0x00009a60, 0x00000100, 0xffffffff,
+	0x00009868, 0x00000100, 0xffffffff,
+	0x00008d58, 0x00000100, 0xffffffff,
+	0x00009510, 0x00000100, 0xffffffff,
+	0x0000949c, 0x00000100, 0xffffffff,
+	0x00009654, 0x00000100, 0xffffffff,
+	0x00009030, 0x00000100, 0xffffffff,
+	0x00009034, 0x00000100, 0xffffffff,
+	0x00009038, 0x00000100, 0xffffffff,
+	0x0000903c, 0x00000100, 0xffffffff,
+	0x00009040, 0x00000100, 0xffffffff,
+	0x0000a200, 0x00000100, 0xffffffff,
+	0x0000a204, 0x00000100, 0xffffffff,
+	0x0000a208, 0x00000100, 0xffffffff,
+	0x0000a20c, 0x00000100, 0xffffffff,
+	0x0000977c, 0x00000100, 0xffffffff,
+	0x00003f80, 0x00000100, 0xffffffff,
+	0x0000a210, 0x00000100, 0xffffffff,
+	0x0000a214, 0x00000100, 0xffffffff,
+	0x000004d8, 0x00000100, 0xffffffff,
+	0x00009784, 0x00000100, 0xffffffff,
+	0x00009698, 0x00000100, 0xffffffff,
+	0x000004d4, 0x00000200, 0xffffffff,
+	0x000004d0, 0x00000000, 0xffffffff,
+	0x000030cc, 0x00000100, 0xffffffff,
+	0x0000d0c0, 0xff000100, 0xffffffff,
+	0x0000915c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define CAICOS_MGCG_DEFAULT_LENGTH sizeof(caicos_mgcg_default) / (3 * sizeof(u32))
+
+static const u32 caicos_mgcg_disable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x00009150, 0x00600000, 0xffffffff
+};
+#define CAICOS_MGCG_DISABLE_LENGTH sizeof(caicos_mgcg_disable) / (3 * sizeof(u32))
+
+static const u32 caicos_mgcg_enable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00009150, 0x46944040, 0xffffffff
+};
+#define CAICOS_MGCG_ENABLE_LENGTH sizeof(caicos_mgcg_enable) / (3 * sizeof(u32))
+
+//********* TURKS **************//
+static const u32 turks_cgcg_cgls_default[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define TURKS_CGCG_CGLS_DEFAULT_LENGTH  sizeof(turks_cgcg_cgls_default) / (3 * sizeof(u32))
+
+static const u32 turks_cgcg_cgls_disable[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00000644, 0x000f7912, 0x001f4180,
+	0x00000644, 0x000f3812, 0x001f4180
+};
+#define TURKS_CGCG_CGLS_DISABLE_LENGTH sizeof(turks_cgcg_cgls_disable) / (3 * sizeof(u32))
+
+static const u32 turks_cgcg_cgls_enable[] =
+{
+	/* 0x0000c124, 0x84180000, 0x00180000, */
+	0x00000644, 0x000f7892, 0x001f4080,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff
+};
+#define TURKS_CGCG_CGLS_ENABLE_LENGTH sizeof(turks_cgcg_cgls_enable) / (3 * sizeof(u32))
+
+// These are the sequences for turks_mgcg_shls
+static const u32 turks_mgcg_default[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00005448, 0x00000100, 0xffffffff,
+	0x000055e4, 0x00600100, 0xffffffff,
+	0x0000160c, 0x00000100, 0xffffffff,
+	0x0000c164, 0x00000100, 0xffffffff,
+	0x00008a18, 0x00000100, 0xffffffff,
+	0x0000897c, 0x06000100, 0xffffffff,
+	0x00008b28, 0x00000100, 0xffffffff,
+	0x00009144, 0x00000100, 0xffffffff,
+	0x00009a60, 0x00000100, 0xffffffff,
+	0x00009868, 0x00000100, 0xffffffff,
+	0x00008d58, 0x00000100, 0xffffffff,
+	0x00009510, 0x00000100, 0xffffffff,
+	0x0000949c, 0x00000100, 0xffffffff,
+	0x00009654, 0x00000100, 0xffffffff,
+	0x00009030, 0x00000100, 0xffffffff,
+	0x00009034, 0x00000100, 0xffffffff,
+	0x00009038, 0x00000100, 0xffffffff,
+	0x0000903c, 0x00000100, 0xffffffff,
+	0x00009040, 0x00000100, 0xffffffff,
+	0x0000a200, 0x00000100, 0xffffffff,
+	0x0000a204, 0x00000100, 0xffffffff,
+	0x0000a208, 0x00000100, 0xffffffff,
+	0x0000a20c, 0x00000100, 0xffffffff,
+	0x0000977c, 0x00000100, 0xffffffff,
+	0x00003f80, 0x00000100, 0xffffffff,
+	0x0000a210, 0x00000100, 0xffffffff,
+	0x0000a214, 0x00000100, 0xffffffff,
+	0x000004d8, 0x00000100, 0xffffffff,
+	0x00009784, 0x00000100, 0xffffffff,
+	0x00009698, 0x00000100, 0xffffffff,
+	0x000004d4, 0x00000200, 0xffffffff,
+	0x000004d0, 0x00000000, 0xffffffff,
+	0x000030cc, 0x00000100, 0xffffffff,
+	0x0000d0c0, 0x00000100, 0xffffffff,
+	0x0000915c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091b0, 0x00070000, 0xffffffff,
+	0x000091b4, 0x00030002, 0xffffffff,
+	0x000091b8, 0x00050004, 0xffffffff,
+	0x000091c4, 0x00010006, 0xffffffff,
+	0x000091c8, 0x00090008, 0xffffffff,
+	0x000091cc, 0x00070000, 0xffffffff,
+	0x000091d0, 0x00030002, 0xffffffff,
+	0x000091d4, 0x00050004, 0xffffffff,
+	0x000091e0, 0x00010006, 0xffffffff,
+	0x000091e4, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x000091ec, 0x00070000, 0xffffffff,
+	0x000091f0, 0x00030002, 0xffffffff,
+	0x000091f4, 0x00050004, 0xffffffff,
+	0x00009200, 0x00010006, 0xffffffff,
+	0x00009204, 0x00090008, 0xffffffff,
+	0x00009208, 0x00070000, 0xffffffff,
+	0x0000920c, 0x00030002, 0xffffffff,
+	0x00009210, 0x00050004, 0xffffffff,
+	0x0000921c, 0x00010006, 0xffffffff,
+	0x00009220, 0x00090008, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define TURKS_MGCG_DEFAULT_LENGTH sizeof(turks_mgcg_default) / (3 * sizeof(u32))
+
+static const u32 turks_mgcg_disable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x00009150, 0x00600000, 0xffffffff
+};
+#define TURKS_MGCG_DISABLE_LENGTH sizeof(turks_mgcg_disable) / (3 * sizeof(u32))
+
+static const u32 turks_mgcg_enable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00009150, 0x6e944000, 0xffffffff
+};
+#define TURKS_MGCG_ENABLE_LENGTH sizeof(turks_mgcg_enable) / (3 * sizeof(u32))
+
+#endif
+
+#ifndef BTC_SYSLS_SEQUENCE
+#define BTC_SYSLS_SEQUENCE  100
+
+
+//********* BARTS **************//
+static const u32 barts_sysls_default[] =
+{
+	/* Register,   Value,     Mask bits */
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+#define BARTS_SYSLS_DEFAULT_LENGTH sizeof(barts_sysls_default) / (3 * sizeof(u32))
+
+static const u32 barts_sysls_disable[] =
+{
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x00041401, 0xffffffff,
+	0x0000264c, 0x00040400, 0xffffffff,
+	0x00002648, 0x00040400, 0xffffffff,
+	0x00002650, 0x00040400, 0xffffffff,
+	0x000020b8, 0x00040400, 0xffffffff,
+	0x000020bc, 0x00040400, 0xffffffff,
+	0x000020c0, 0x00040c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680000, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00007ffd, 0xffffffff,
+	0x00000c7c, 0x0000ff00, 0xffffffff,
+	0x00006dfc, 0x0000007f, 0xffffffff
+};
+#define BARTS_SYSLS_DISABLE_LENGTH sizeof(barts_sysls_disable) / (3 * sizeof(u32))
+
+static const u32 barts_sysls_enable[] =
+{
+	0x000055e8, 0x00000001, 0xffffffff,
+	0x0000d0bc, 0x00000100, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x000004c8, 0x00000000, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+#define BARTS_SYSLS_ENABLE_LENGTH sizeof(barts_sysls_enable) / (3 * sizeof(u32))
+
+//********* CAICOS **************//
+static const u32 caicos_sysls_default[] =
+{
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+#define CAICOS_SYSLS_DEFAULT_LENGTH sizeof(caicos_sysls_default) / (3 * sizeof(u32))
+
+static const u32 caicos_sysls_disable[] =
+{
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x00041401, 0xffffffff,
+	0x0000264c, 0x00040400, 0xffffffff,
+	0x00002648, 0x00040400, 0xffffffff,
+	0x00002650, 0x00040400, 0xffffffff,
+	0x000020b8, 0x00040400, 0xffffffff,
+	0x000020bc, 0x00040400, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680000, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00007ffd, 0xffffffff,
+	0x00000c7c, 0x0000ff00, 0xffffffff,
+	0x00006dfc, 0x0000007f, 0xffffffff
+};
+#define CAICOS_SYSLS_DISABLE_LENGTH sizeof(caicos_sysls_disable) / (3 * sizeof(u32))
+
+static const u32 caicos_sysls_enable[] =
+{
+	0x000055e8, 0x00000001, 0xffffffff,
+	0x0000d0bc, 0x00000100, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff,
+	0x000004c8, 0x00000000, 0xffffffff
+};
+#define CAICOS_SYSLS_ENABLE_LENGTH sizeof(caicos_sysls_enable) / (3 * sizeof(u32))
+
+//********* TURKS **************//
+static const u32 turks_sysls_default[] =
+{
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+#define TURKS_SYSLS_DEFAULT_LENGTH sizeof(turks_sysls_default) / (3 * sizeof(u32))
+
+static const u32 turks_sysls_disable[] =
+{
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x00041401, 0xffffffff,
+	0x0000264c, 0x00040400, 0xffffffff,
+	0x00002648, 0x00040400, 0xffffffff,
+	0x00002650, 0x00040400, 0xffffffff,
+	0x000020b8, 0x00040400, 0xffffffff,
+	0x000020bc, 0x00040400, 0xffffffff,
+	0x000020c0, 0x00040c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680000, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00007ffd, 0xffffffff,
+	0x00000c7c, 0x0000ff00, 0xffffffff,
+	0x00006dfc, 0x0000007f, 0xffffffff
+};
+#define TURKS_SYSLS_DISABLE_LENGTH sizeof(turks_sysls_disable) / (3 * sizeof(u32))
+
+static const u32 turks_sysls_enable[] =
+{
+	0x000055e8, 0x00000001, 0xffffffff,
+	0x0000d0bc, 0x00000100, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x000004c8, 0x00000000, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+#define TURKS_SYSLS_ENABLE_LENGTH sizeof(turks_sysls_enable) / (3 * sizeof(u32))
+
+#endif
+
+u32 btc_valid_sclk[40] =
+{
+	5000,   10000,  15000,  20000,  25000,  30000,  35000,  40000,  45000,  50000,
+	55000,  60000,  65000,  70000,  75000,  80000,  85000,  90000,  95000,  100000,
+	105000, 110000, 11500,  120000, 125000, 130000, 135000, 140000, 145000, 150000,
+	155000, 160000, 165000, 170000, 175000, 180000, 185000, 190000, 195000, 200000
+};
+
+static const struct radeon_blacklist_clocks btc_blacklist_clocks[] =
+{
+        { 10000, 30000, RADEON_SCLK_UP },
+        { 15000, 30000, RADEON_SCLK_UP },
+        { 20000, 30000, RADEON_SCLK_UP },
+        { 25000, 30000, RADEON_SCLK_UP }
+};
+
+void btc_apply_voltage_dependency_rules(struct radeon_clock_voltage_dependency_table *table,
+					u32 clock, u16 max_voltage, u16 *voltage)
+{
+	u32 i;
+
+	if ((table == NULL) || (table->count == 0))
+		return;
+
+	for (i= 0; i < table->count; i++) {
+		if (clock <= table->entries[i].clk) {
+			if (*voltage < table->entries[i].v)
+				*voltage = (u16)((table->entries[i].v < max_voltage) ?
+						  table->entries[i].v : max_voltage);
+			return;
+		}
+	}
+
+	*voltage = (*voltage > max_voltage) ? *voltage : max_voltage;
+}
+
+static u32 btc_find_valid_clock(struct radeon_clock_array *clocks,
+				u32 max_clock, u32 requested_clock)
+{
+	unsigned int i;
+
+	if ((clocks == NULL) || (clocks->count == 0))
+		return (requested_clock < max_clock) ? requested_clock : max_clock;
+
+	for (i = 0; i < clocks->count; i++) {
+		if (clocks->values[i] >= requested_clock)
+			return (clocks->values[i] < max_clock) ? clocks->values[i] : max_clock;
+	}
+
+	return (clocks->values[clocks->count - 1] < max_clock) ?
+		clocks->values[clocks->count - 1] : max_clock;
+}
+
+static u32 btc_get_valid_mclk(struct radeon_device *rdev,
+			      u32 max_mclk, u32 requested_mclk)
+{
+	return btc_find_valid_clock(&rdev->pm.dpm.dyn_state.valid_mclk_values,
+				    max_mclk, requested_mclk);
+}
+
+static u32 btc_get_valid_sclk(struct radeon_device *rdev,
+			      u32 max_sclk, u32 requested_sclk)
+{
+	return btc_find_valid_clock(&rdev->pm.dpm.dyn_state.valid_sclk_values,
+				    max_sclk, requested_sclk);
+}
+
+void btc_skip_blacklist_clocks(struct radeon_device *rdev,
+			       const u32 max_sclk, const u32 max_mclk,
+			       u32 *sclk, u32 *mclk)
+{
+	int i, num_blacklist_clocks;
+
+	if ((sclk == NULL) || (mclk == NULL))
+		return;
+
+	num_blacklist_clocks = ARRAY_SIZE(btc_blacklist_clocks);
+
+	for (i = 0; i < num_blacklist_clocks; i++) {
+		if ((btc_blacklist_clocks[i].sclk == *sclk) &&
+		    (btc_blacklist_clocks[i].mclk == *mclk))
+			break;
+	}
+
+	if (i < num_blacklist_clocks) {
+		if (btc_blacklist_clocks[i].action == RADEON_SCLK_UP) {
+			*sclk = btc_get_valid_sclk(rdev, max_sclk, *sclk + 1);
+
+			if (*sclk < max_sclk)
+				btc_skip_blacklist_clocks(rdev, max_sclk, max_mclk, sclk, mclk);
+		}
+	}
+}
+
+void btc_adjust_clock_combinations(struct radeon_device *rdev,
+				   const struct radeon_clock_and_voltage_limits *max_limits,
+				   struct rv7xx_pl *pl)
+{
+
+	if ((pl->mclk == 0) || (pl->sclk == 0))
+		return;
+
+	if (pl->mclk == pl->sclk)
+		return;
+
+	if (pl->mclk > pl->sclk) {
+		if (((pl->mclk + (pl->sclk - 1)) / pl->sclk) > rdev->pm.dpm.dyn_state.mclk_sclk_ratio)
+			pl->sclk = btc_get_valid_sclk(rdev,
+						      max_limits->sclk,
+						      (pl->mclk +
+						       (rdev->pm.dpm.dyn_state.mclk_sclk_ratio - 1)) /
+						      rdev->pm.dpm.dyn_state.mclk_sclk_ratio);
+	} else {
+		if ((pl->sclk - pl->mclk) > rdev->pm.dpm.dyn_state.sclk_mclk_delta)
+			pl->mclk = btc_get_valid_mclk(rdev,
+						      max_limits->mclk,
+						      pl->sclk -
+						      rdev->pm.dpm.dyn_state.sclk_mclk_delta);
+	}
+}
+
+static u16 btc_find_voltage(struct atom_voltage_table *table, u16 voltage)
+{
+	unsigned int i;
+
+	for (i = 0; i < table->count; i++) {
+		if (voltage <= table->entries[i].value)
+			return table->entries[i].value;
+	}
+
+	return table->entries[table->count - 1].value;
+}
+
+void btc_apply_voltage_delta_rules(struct radeon_device *rdev,
+				   u16 max_vddc, u16 max_vddci,
+				   u16 *vddc, u16 *vddci)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u16 new_voltage;
+
+	if ((0 == *vddc) || (0 == *vddci))
+		return;
+
+	if (*vddc > *vddci) {
+		if ((*vddc - *vddci) > rdev->pm.dpm.dyn_state.vddc_vddci_delta) {
+			new_voltage = btc_find_voltage(&eg_pi->vddci_voltage_table,
+						       (*vddc - rdev->pm.dpm.dyn_state.vddc_vddci_delta));
+			*vddci = (new_voltage < max_vddci) ? new_voltage : max_vddci;
+		}
+	} else {
+		if ((*vddci - *vddc) > rdev->pm.dpm.dyn_state.vddc_vddci_delta) {
+			new_voltage = btc_find_voltage(&eg_pi->vddc_voltage_table,
+						       (*vddci - rdev->pm.dpm.dyn_state.vddc_vddci_delta));
+			*vddc = (new_voltage < max_vddc) ? new_voltage : max_vddc;
+		}
+	}
+}
+
+static void btc_enable_bif_dynamic_pcie_gen2(struct radeon_device *rdev,
+					     bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp, bif;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	if (enable) {
+		if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+		    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
+			if (!pi->boot_in_gen2) {
+				bif = RREG32(CG_BIF_REQ_AND_RSP) & ~CG_CLIENT_REQ_MASK;
+				bif |= CG_CLIENT_REQ(0xd);
+				WREG32(CG_BIF_REQ_AND_RSP, bif);
+
+				tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+				tmp |= LC_HW_VOLTAGE_IF_CONTROL(1);
+				tmp |= LC_GEN2_EN_STRAP;
+
+				tmp |= LC_CLR_FAILED_SPD_CHANGE_CNT;
+				WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+				udelay(10);
+				tmp &= ~LC_CLR_FAILED_SPD_CHANGE_CNT;
+				WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+			}
+		}
+	} else {
+		if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) ||
+		    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
+			if (!pi->boot_in_gen2) {
+				bif = RREG32(CG_BIF_REQ_AND_RSP) & ~CG_CLIENT_REQ_MASK;
+				bif |= CG_CLIENT_REQ(0xd);
+				WREG32(CG_BIF_REQ_AND_RSP, bif);
+
+				tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+				tmp &= ~LC_GEN2_EN_STRAP;
+			}
+			WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+		}
+	}
+}
+
+static void btc_enable_dynamic_pcie_gen2(struct radeon_device *rdev,
+					 bool enable)
+{
+	btc_enable_bif_dynamic_pcie_gen2(rdev, enable);
+
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, ENABLE_GEN2PCIE, ~ENABLE_GEN2PCIE);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~ENABLE_GEN2PCIE);
+}
+
+static int btc_disable_ulv(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (eg_pi->ulv.supported) {
+		if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_DisableULV) != PPSMC_Result_OK)
+			return -EINVAL;
+	}
+	return 0;
+}
+
+static int btc_populate_ulv_state(struct radeon_device *rdev,
+				  RV770_SMC_STATETABLE *table)
+{
+	int ret = -EINVAL;
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_pl *ulv_pl = eg_pi->ulv.pl;
+
+	if (ulv_pl->vddc) {
+		ret = cypress_convert_power_level_to_smc(rdev,
+							 ulv_pl,
+							 &table->ULVState.levels[0],
+							 PPSMC_DISPLAY_WATERMARK_LOW);
+		if (ret == 0) {
+			table->ULVState.levels[0].arbValue = MC_CG_ARB_FREQ_F0;
+			table->ULVState.levels[0].ACIndex = 1;
+
+			table->ULVState.levels[1] = table->ULVState.levels[0];
+			table->ULVState.levels[2] = table->ULVState.levels[0];
+
+			table->ULVState.flags |= PPSMC_SWSTATE_FLAG_DC;
+
+			WREG32(CG_ULV_CONTROL, BTC_CGULVCONTROL_DFLT);
+			WREG32(CG_ULV_PARAMETER, BTC_CGULVPARAMETER_DFLT);
+		}
+	}
+
+	return ret;
+}
+
+static int btc_populate_smc_acpi_state(struct radeon_device *rdev,
+				       RV770_SMC_STATETABLE *table)
+{
+	int ret = cypress_populate_smc_acpi_state(rdev, table);
+
+	if (ret == 0) {
+		table->ACPIState.levels[0].ACIndex = 0;
+		table->ACPIState.levels[1].ACIndex = 0;
+		table->ACPIState.levels[2].ACIndex = 0;
+	}
+
+	return ret;
+}
+
+void btc_program_mgcg_hw_sequence(struct radeon_device *rdev,
+				  const u32 *sequence, u32 count)
+{
+	u32 i, length = count * 3;
+	u32 tmp;
+
+	for (i = 0; i < length; i+=3) {
+		tmp = RREG32(sequence[i]);
+		tmp &= ~sequence[i+2];
+		tmp |= sequence[i+1] & sequence[i+2];
+		WREG32(sequence[i], tmp);
+	}
+}
+
+static void btc_cg_clock_gating_default(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *p = NULL;
+
+	if (rdev->family == CHIP_BARTS) {
+		p = (const u32 *)&barts_cgcg_cgls_default;
+		count = BARTS_CGCG_CGLS_DEFAULT_LENGTH;
+	} else if (rdev->family == CHIP_TURKS) {
+		p = (const u32 *)&turks_cgcg_cgls_default;
+		count = TURKS_CGCG_CGLS_DEFAULT_LENGTH;
+	} else if (rdev->family == CHIP_CAICOS) {
+		p = (const u32 *)&caicos_cgcg_cgls_default;
+		count = CAICOS_CGCG_CGLS_DEFAULT_LENGTH;
+	} else
+		return;
+
+	btc_program_mgcg_hw_sequence(rdev, p, count);
+}
+
+static void btc_cg_clock_gating_enable(struct radeon_device *rdev,
+				       bool enable)
+{
+	u32 count;
+	const u32 *p = NULL;
+
+	if (enable) {
+		if (rdev->family == CHIP_BARTS) {
+			p = (const u32 *)&barts_cgcg_cgls_enable;
+			count = BARTS_CGCG_CGLS_ENABLE_LENGTH;
+		} else if (rdev->family == CHIP_TURKS) {
+			p = (const u32 *)&turks_cgcg_cgls_enable;
+			count = TURKS_CGCG_CGLS_ENABLE_LENGTH;
+		} else if (rdev->family == CHIP_CAICOS) {
+			p = (const u32 *)&caicos_cgcg_cgls_enable;
+			count = CAICOS_CGCG_CGLS_ENABLE_LENGTH;
+		} else
+			return;
+	} else {
+		if (rdev->family == CHIP_BARTS) {
+			p = (const u32 *)&barts_cgcg_cgls_disable;
+			count = BARTS_CGCG_CGLS_DISABLE_LENGTH;
+		} else if (rdev->family == CHIP_TURKS) {
+			p = (const u32 *)&turks_cgcg_cgls_disable;
+			count = TURKS_CGCG_CGLS_DISABLE_LENGTH;
+		} else if (rdev->family == CHIP_CAICOS) {
+			p = (const u32 *)&caicos_cgcg_cgls_disable;
+			count = CAICOS_CGCG_CGLS_DISABLE_LENGTH;
+		} else
+			return;
+	}
+
+	btc_program_mgcg_hw_sequence(rdev, p, count);
+}
+
+static void btc_mg_clock_gating_default(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *p = NULL;
+
+	if (rdev->family == CHIP_BARTS) {
+		p = (const u32 *)&barts_mgcg_default;
+		count = BARTS_MGCG_DEFAULT_LENGTH;
+	} else if (rdev->family == CHIP_TURKS) {
+		p = (const u32 *)&turks_mgcg_default;
+		count = TURKS_MGCG_DEFAULT_LENGTH;
+	} else if (rdev->family == CHIP_CAICOS) {
+		p = (const u32 *)&caicos_mgcg_default;
+		count = CAICOS_MGCG_DEFAULT_LENGTH;
+	} else
+		return;
+
+	btc_program_mgcg_hw_sequence(rdev, p, count);
+}
+
+static void btc_mg_clock_gating_enable(struct radeon_device *rdev,
+				       bool enable)
+{
+	u32 count;
+	const u32 *p = NULL;
+
+	if (enable) {
+		if (rdev->family == CHIP_BARTS) {
+			p = (const u32 *)&barts_mgcg_enable;
+			count = BARTS_MGCG_ENABLE_LENGTH;
+		} else if (rdev->family == CHIP_TURKS) {
+			p = (const u32 *)&turks_mgcg_enable;
+			count = TURKS_MGCG_ENABLE_LENGTH;
+		} else if (rdev->family == CHIP_CAICOS) {
+			p = (const u32 *)&caicos_mgcg_enable;
+			count = CAICOS_MGCG_ENABLE_LENGTH;
+		} else
+			return;
+	} else {
+		if (rdev->family == CHIP_BARTS) {
+			p = (const u32 *)&barts_mgcg_disable[0];
+			count = BARTS_MGCG_DISABLE_LENGTH;
+		} else if (rdev->family == CHIP_TURKS) {
+			p = (const u32 *)&turks_mgcg_disable[0];
+			count = TURKS_MGCG_DISABLE_LENGTH;
+		} else if (rdev->family == CHIP_CAICOS) {
+			p = (const u32 *)&caicos_mgcg_disable[0];
+			count = CAICOS_MGCG_DISABLE_LENGTH;
+		} else
+			return;
+	}
+
+	btc_program_mgcg_hw_sequence(rdev, p, count);
+}
+
+static void btc_ls_clock_gating_default(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *p = NULL;
+
+	if (rdev->family == CHIP_BARTS) {
+		p = (const u32 *)&barts_sysls_default;
+		count = BARTS_SYSLS_DEFAULT_LENGTH;
+	} else if (rdev->family == CHIP_TURKS) {
+		p = (const u32 *)&turks_sysls_default;
+		count = TURKS_SYSLS_DEFAULT_LENGTH;
+	} else if (rdev->family == CHIP_CAICOS) {
+		p = (const u32 *)&caicos_sysls_default;
+		count = CAICOS_SYSLS_DEFAULT_LENGTH;
+	} else
+		return;
+
+	btc_program_mgcg_hw_sequence(rdev, p, count);
+}
+
+static void btc_ls_clock_gating_enable(struct radeon_device *rdev,
+				       bool enable)
+{
+	u32 count;
+	const u32 *p = NULL;
+
+	if (enable) {
+		if (rdev->family == CHIP_BARTS) {
+			p = (const u32 *)&barts_sysls_enable;
+			count = BARTS_SYSLS_ENABLE_LENGTH;
+		} else if (rdev->family == CHIP_TURKS) {
+			p = (const u32 *)&turks_sysls_enable;
+			count = TURKS_SYSLS_ENABLE_LENGTH;
+		} else if (rdev->family == CHIP_CAICOS) {
+			p = (const u32 *)&caicos_sysls_enable;
+			count = CAICOS_SYSLS_ENABLE_LENGTH;
+		} else
+			return;
+	} else {
+		if (rdev->family == CHIP_BARTS) {
+			p = (const u32 *)&barts_sysls_disable;
+			count = BARTS_SYSLS_DISABLE_LENGTH;
+		} else if (rdev->family == CHIP_TURKS) {
+			p = (const u32 *)&turks_sysls_disable;
+			count = TURKS_SYSLS_DISABLE_LENGTH;
+		} else if (rdev->family == CHIP_CAICOS) {
+			p = (const u32 *)&caicos_sysls_disable;
+			count = CAICOS_SYSLS_DISABLE_LENGTH;
+		} else
+			return;
+	}
+
+	btc_program_mgcg_hw_sequence(rdev, p, count);
+}
+
+bool btc_dpm_enabled(struct radeon_device *rdev)
+{
+	if (rv770_is_smc_running(rdev))
+		return true;
+	else
+		return false;
+}
+
+static int btc_init_smc_table(struct radeon_device *rdev,
+			      struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	RV770_SMC_STATETABLE *table = &pi->smc_statetable;
+	int ret;
+
+	memset(table, 0, sizeof(RV770_SMC_STATETABLE));
+
+	cypress_populate_smc_voltage_tables(rdev, table);
+
+	switch (rdev->pm.int_thermal_type) {
+        case THERMAL_TYPE_EVERGREEN:
+        case THERMAL_TYPE_EMC2103_WITH_INTERNAL:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_INTERNAL;
+		break;
+        case THERMAL_TYPE_NONE:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_NONE;
+		break;
+        default:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_EXTERNAL;
+		break;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_HARDWAREDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GPIO_DC;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REGULATOR_HOT)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_REGULATOR_HOT;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_STEPVDDC;
+
+	if (pi->mem_gddr5)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GDDR5;
+
+	ret = cypress_populate_smc_initial_state(rdev, radeon_boot_state, table);
+	if (ret)
+		return ret;
+
+	if (eg_pi->sclk_deep_sleep)
+		WREG32_P(SCLK_PSKIP_CNTL, PSKIP_ON_ALLOW_STOP_HI(32),
+			 ~PSKIP_ON_ALLOW_STOP_HI_MASK);
+
+	ret = btc_populate_smc_acpi_state(rdev, table);
+	if (ret)
+		return ret;
+
+	if (eg_pi->ulv.supported) {
+		ret = btc_populate_ulv_state(rdev, table);
+		if (ret)
+			eg_pi->ulv.supported = false;
+	}
+
+	table->driverState = table->initialState;
+
+	return rv770_copy_bytes_to_smc(rdev,
+				       pi->state_table_start,
+				       (u8 *)table,
+				       sizeof(RV770_SMC_STATETABLE),
+				       pi->sram_end);
+}
+
+static void btc_set_at_for_uvd(struct radeon_device *rdev,
+			       struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	int idx = 0;
+
+	if (r600_is_uvd_state(radeon_new_state->class, radeon_new_state->class2))
+		idx = 1;
+
+	if ((idx == 1) && !eg_pi->smu_uvd_hs) {
+		pi->rlp = 10;
+		pi->rmp = 100;
+		pi->lhp = 100;
+		pi->lmp = 10;
+	} else {
+		pi->rlp = eg_pi->ats[idx].rlp;
+		pi->rmp = eg_pi->ats[idx].rmp;
+		pi->lhp = eg_pi->ats[idx].lhp;
+		pi->lmp = eg_pi->ats[idx].lmp;
+	}
+
+}
+
+void btc_notify_uvd_to_smc(struct radeon_device *rdev,
+			   struct radeon_ps *radeon_new_state)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (r600_is_uvd_state(radeon_new_state->class, radeon_new_state->class2)) {
+		rv770_write_smc_soft_register(rdev,
+					      RV770_SMC_SOFT_REGISTER_uvd_enabled, 1);
+		eg_pi->uvd_enabled = true;
+	} else {
+		rv770_write_smc_soft_register(rdev,
+					      RV770_SMC_SOFT_REGISTER_uvd_enabled, 0);
+		eg_pi->uvd_enabled = false;
+	}
+}
+
+int btc_reset_to_default(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_ResetToDefaults) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return 0;
+}
+
+static void btc_stop_smc(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (((RREG32(LB_SYNC_RESET_SEL) & LB_SYNC_RESET_SEL_MASK) >> LB_SYNC_RESET_SEL_SHIFT) != 1)
+			break;
+		udelay(1);
+	}
+	udelay(100);
+
+	r7xx_stop_smc(rdev);
+}
+
+void btc_read_arb_registers(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct evergreen_arb_registers *arb_registers =
+		&eg_pi->bootup_arb_registers;
+
+	arb_registers->mc_arb_dram_timing = RREG32(MC_ARB_DRAM_TIMING);
+	arb_registers->mc_arb_dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+	arb_registers->mc_arb_rfsh_rate = RREG32(MC_ARB_RFSH_RATE);
+	arb_registers->mc_arb_burst_time = RREG32(MC_ARB_BURST_TIME);
+}
+
+
+static void btc_set_arb0_registers(struct radeon_device *rdev,
+				   struct evergreen_arb_registers *arb_registers)
+{
+	u32 val;
+
+	WREG32(MC_ARB_DRAM_TIMING,  arb_registers->mc_arb_dram_timing);
+	WREG32(MC_ARB_DRAM_TIMING2, arb_registers->mc_arb_dram_timing2);
+
+	val = (arb_registers->mc_arb_rfsh_rate & POWERMODE0_MASK) >>
+		POWERMODE0_SHIFT;
+	WREG32_P(MC_ARB_RFSH_RATE, POWERMODE0(val), ~POWERMODE0_MASK);
+
+	val = (arb_registers->mc_arb_burst_time & STATE0_MASK) >>
+		STATE0_SHIFT;
+	WREG32_P(MC_ARB_BURST_TIME, STATE0(val), ~STATE0_MASK);
+}
+
+static void btc_set_boot_state_timing(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (eg_pi->ulv.supported)
+		btc_set_arb0_registers(rdev, &eg_pi->bootup_arb_registers);
+}
+
+static bool btc_is_state_ulv_compatible(struct radeon_device *rdev,
+					struct radeon_ps *radeon_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_pl *ulv_pl = eg_pi->ulv.pl;
+
+	if (state->low.mclk != ulv_pl->mclk)
+		return false;
+
+	if (state->low.vddci != ulv_pl->vddci)
+		return false;
+
+	/* XXX check minclocks, etc. */
+
+	return true;
+}
+
+
+static int btc_set_ulv_dram_timing(struct radeon_device *rdev)
+{
+	u32 val;
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_pl *ulv_pl = eg_pi->ulv.pl;
+
+	radeon_atom_set_engine_dram_timings(rdev,
+					    ulv_pl->sclk,
+					    ulv_pl->mclk);
+
+	val = rv770_calculate_memory_refresh_rate(rdev, ulv_pl->sclk);
+	WREG32_P(MC_ARB_RFSH_RATE, POWERMODE0(val), ~POWERMODE0_MASK);
+
+	val = cypress_calculate_burst_time(rdev, ulv_pl->sclk, ulv_pl->mclk);
+	WREG32_P(MC_ARB_BURST_TIME, STATE0(val), ~STATE0_MASK);
+
+	return 0;
+}
+
+static int btc_enable_ulv(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableULV) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int btc_set_power_state_conditionally_enable_ulv(struct radeon_device *rdev,
+							struct radeon_ps *radeon_new_state)
+{
+	int ret = 0;
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (eg_pi->ulv.supported) {
+		if (btc_is_state_ulv_compatible(rdev, radeon_new_state)) {
+			// Set ARB[0] to reflect the DRAM timing needed for ULV.
+			ret = btc_set_ulv_dram_timing(rdev);
+			if (ret == 0)
+				ret = btc_enable_ulv(rdev);
+		}
+	}
+
+	return ret;
+}
+
+static bool btc_check_s0_mc_reg_index(u16 in_reg, u16 *out_reg)
+{
+	bool result = true;
+
+	switch (in_reg) {
+	case MC_SEQ_RAS_TIMING >> 2:
+		*out_reg = MC_SEQ_RAS_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_CAS_TIMING >> 2:
+		*out_reg = MC_SEQ_CAS_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_MISC_TIMING >> 2:
+		*out_reg = MC_SEQ_MISC_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_MISC_TIMING2 >> 2:
+		*out_reg = MC_SEQ_MISC_TIMING2_LP >> 2;
+		break;
+        case MC_SEQ_RD_CTL_D0 >> 2:
+		*out_reg = MC_SEQ_RD_CTL_D0_LP >> 2;
+		break;
+        case MC_SEQ_RD_CTL_D1 >> 2:
+		*out_reg = MC_SEQ_RD_CTL_D1_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_D0 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_D0_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_D1 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_D1_LP >> 2;
+		break;
+        case MC_PMG_CMD_EMRS >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS1 >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+		break;
+        default:
+		result = false;
+		break;
+	}
+
+	return result;
+}
+
+static void btc_set_valid_flag(struct evergreen_mc_reg_table *table)
+{
+	u8 i, j;
+
+	for (i = 0; i < table->last; i++) {
+		for (j = 1; j < table->num_entries; j++) {
+			if (table->mc_reg_table_entry[j-1].mc_data[i] !=
+			    table->mc_reg_table_entry[j].mc_data[i]) {
+				table->valid_flag |= (1 << i);
+				break;
+			}
+		}
+	}
+}
+
+static int btc_set_mc_special_registers(struct radeon_device *rdev,
+					struct evergreen_mc_reg_table *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 i, j, k;
+	u32 tmp;
+
+	for (i = 0, j = table->last; i < table->last; i++) {
+		switch (table->mc_reg_address[i].s1) {
+		case MC_SEQ_MISC1 >> 2:
+			tmp = RREG32(MC_PMG_CMD_EMRS);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_EMRS >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+			for (k = 0; k < table->num_entries; k++) {
+				table->mc_reg_table_entry[k].mc_data[j] =
+					((tmp & 0xffff0000)) |
+					((table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16);
+			}
+			j++;
+
+			if (j > SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+
+			tmp = RREG32(MC_PMG_CMD_MRS);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+			for (k = 0; k < table->num_entries; k++) {
+				table->mc_reg_table_entry[k].mc_data[j] =
+					(tmp & 0xffff0000) |
+					(table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff);
+				if (!pi->mem_gddr5)
+					table->mc_reg_table_entry[k].mc_data[j] |= 0x100;
+			}
+			j++;
+
+			if (j > SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			break;
+		case MC_SEQ_RESERVE_M >> 2:
+			tmp = RREG32(MC_PMG_CMD_MRS1);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS1 >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+			for (k = 0; k < table->num_entries; k++) {
+				table->mc_reg_table_entry[k].mc_data[j] =
+					(tmp & 0xffff0000) |
+					(table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff);
+			}
+			j++;
+
+			if (j > SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			break;
+		default:
+			break;
+		}
+	}
+
+	table->last = j;
+
+	return 0;
+}
+
+static void btc_set_s0_mc_reg_index(struct evergreen_mc_reg_table *table)
+{
+	u32 i;
+	u16 address;
+
+	for (i = 0; i < table->last; i++) {
+		table->mc_reg_address[i].s0 =
+			btc_check_s0_mc_reg_index(table->mc_reg_address[i].s1, &address) ?
+			address : table->mc_reg_address[i].s1;
+	}
+}
+
+static int btc_copy_vbios_mc_reg_table(struct atom_mc_reg_table *table,
+				       struct evergreen_mc_reg_table *eg_table)
+{
+	u8 i, j;
+
+	if (table->last > SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE)
+		return -EINVAL;
+
+	if (table->num_entries > MAX_AC_TIMING_ENTRIES)
+		return -EINVAL;
+
+	for (i = 0; i < table->last; i++)
+		eg_table->mc_reg_address[i].s1 = table->mc_reg_address[i].s1;
+	eg_table->last = table->last;
+
+	for (i = 0; i < table->num_entries; i++) {
+		eg_table->mc_reg_table_entry[i].mclk_max =
+			table->mc_reg_table_entry[i].mclk_max;
+		for(j = 0; j < table->last; j++)
+			eg_table->mc_reg_table_entry[i].mc_data[j] =
+				table->mc_reg_table_entry[i].mc_data[j];
+	}
+	eg_table->num_entries = table->num_entries;
+
+	return 0;
+}
+
+static int btc_initialize_mc_reg_table(struct radeon_device *rdev)
+{
+	int ret;
+	struct atom_mc_reg_table *table;
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct evergreen_mc_reg_table *eg_table = &eg_pi->mc_reg_table;
+	u8 module_index = rv770_get_memory_module_index(rdev);
+
+	table = kzalloc(sizeof(struct atom_mc_reg_table), GFP_KERNEL);
+	if (!table)
+		return -ENOMEM;
+
+	/* Program additional LP registers that are no longer programmed by VBIOS */
+	WREG32(MC_SEQ_RAS_TIMING_LP, RREG32(MC_SEQ_RAS_TIMING));
+	WREG32(MC_SEQ_CAS_TIMING_LP, RREG32(MC_SEQ_CAS_TIMING));
+	WREG32(MC_SEQ_MISC_TIMING_LP, RREG32(MC_SEQ_MISC_TIMING));
+	WREG32(MC_SEQ_MISC_TIMING2_LP, RREG32(MC_SEQ_MISC_TIMING2));
+	WREG32(MC_SEQ_RD_CTL_D0_LP, RREG32(MC_SEQ_RD_CTL_D0));
+	WREG32(MC_SEQ_RD_CTL_D1_LP, RREG32(MC_SEQ_RD_CTL_D1));
+	WREG32(MC_SEQ_WR_CTL_D0_LP, RREG32(MC_SEQ_WR_CTL_D0));
+	WREG32(MC_SEQ_WR_CTL_D1_LP, RREG32(MC_SEQ_WR_CTL_D1));
+	WREG32(MC_SEQ_PMG_CMD_EMRS_LP, RREG32(MC_PMG_CMD_EMRS));
+	WREG32(MC_SEQ_PMG_CMD_MRS_LP, RREG32(MC_PMG_CMD_MRS));
+	WREG32(MC_SEQ_PMG_CMD_MRS1_LP, RREG32(MC_PMG_CMD_MRS1));
+
+	ret = radeon_atom_init_mc_reg_table(rdev, module_index, table);
+
+	if (ret)
+		goto init_mc_done;
+
+	ret = btc_copy_vbios_mc_reg_table(table, eg_table);
+
+	if (ret)
+		goto init_mc_done;
+
+	btc_set_s0_mc_reg_index(eg_table);
+	ret = btc_set_mc_special_registers(rdev, eg_table);
+
+	if (ret)
+		goto init_mc_done;
+
+	btc_set_valid_flag(eg_table);
+
+init_mc_done:
+	kfree(table);
+
+	return ret;
+}
+
+static void btc_init_stutter_mode(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp;
+
+	if (pi->mclk_stutter_mode_threshold) {
+		if (pi->mem_gddr5) {
+			tmp = RREG32(MC_PMG_AUTO_CFG);
+			if ((0x200 & tmp) == 0) {
+				tmp = (tmp & 0xfffffc0b) | 0x204;
+				WREG32(MC_PMG_AUTO_CFG, tmp);
+			}
+		}
+	}
+}
+
+bool btc_dpm_vblank_too_short(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 vblank_time = r600_dpm_get_vblank_time(rdev);
+	u32 switch_limit = pi->mem_gddr5 ? 450 : 100;
+
+	if (vblank_time < switch_limit)
+		return true;
+	else
+		return false;
+
+}
+
+static void btc_apply_state_adjust_rules(struct radeon_device *rdev,
+					 struct radeon_ps *rps)
+{
+	struct rv7xx_ps *ps = rv770_get_ps(rps);
+	struct radeon_clock_and_voltage_limits *max_limits;
+	bool disable_mclk_switching;
+	u32 mclk, sclk;
+	u16 vddc, vddci;
+
+	if ((rdev->pm.dpm.new_active_crtc_count > 1) ||
+	    btc_dpm_vblank_too_short(rdev))
+		disable_mclk_switching = true;
+	else
+		disable_mclk_switching = false;
+
+	if (rdev->pm.dpm.ac_power)
+		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac;
+	else
+		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc;
+
+	if (rdev->pm.dpm.ac_power == false) {
+		if (ps->high.mclk > max_limits->mclk)
+			ps->high.mclk = max_limits->mclk;
+		if (ps->high.sclk > max_limits->sclk)
+			ps->high.sclk = max_limits->sclk;
+		if (ps->high.vddc > max_limits->vddc)
+			ps->high.vddc = max_limits->vddc;
+		if (ps->high.vddci > max_limits->vddci)
+			ps->high.vddci = max_limits->vddci;
+
+		if (ps->medium.mclk > max_limits->mclk)
+			ps->medium.mclk = max_limits->mclk;
+		if (ps->medium.sclk > max_limits->sclk)
+			ps->medium.sclk = max_limits->sclk;
+		if (ps->medium.vddc > max_limits->vddc)
+			ps->medium.vddc = max_limits->vddc;
+		if (ps->medium.vddci > max_limits->vddci)
+			ps->medium.vddci = max_limits->vddci;
+
+		if (ps->low.mclk > max_limits->mclk)
+			ps->low.mclk = max_limits->mclk;
+		if (ps->low.sclk > max_limits->sclk)
+			ps->low.sclk = max_limits->sclk;
+		if (ps->low.vddc > max_limits->vddc)
+			ps->low.vddc = max_limits->vddc;
+		if (ps->low.vddci > max_limits->vddci)
+			ps->low.vddci = max_limits->vddci;
+	}
+
+	/* XXX validate the min clocks required for display */
+
+	if (disable_mclk_switching) {
+		sclk = ps->low.sclk;
+		mclk = ps->high.mclk;
+		vddc = ps->low.vddc;
+		vddci = ps->high.vddci;
+	} else {
+		sclk = ps->low.sclk;
+		mclk = ps->low.mclk;
+		vddc = ps->low.vddc;
+		vddci = ps->low.vddci;
+	}
+
+	/* adjusted low state */
+	ps->low.sclk = sclk;
+	ps->low.mclk = mclk;
+	ps->low.vddc = vddc;
+	ps->low.vddci = vddci;
+
+	btc_skip_blacklist_clocks(rdev, max_limits->sclk, max_limits->mclk,
+				  &ps->low.sclk, &ps->low.mclk);
+
+	/* adjusted medium, high states */
+	if (ps->medium.sclk < ps->low.sclk)
+		ps->medium.sclk = ps->low.sclk;
+	if (ps->medium.vddc < ps->low.vddc)
+		ps->medium.vddc = ps->low.vddc;
+	if (ps->high.sclk < ps->medium.sclk)
+		ps->high.sclk = ps->medium.sclk;
+	if (ps->high.vddc < ps->medium.vddc)
+		ps->high.vddc = ps->medium.vddc;
+
+	if (disable_mclk_switching) {
+		mclk = ps->low.mclk;
+		if (mclk < ps->medium.mclk)
+			mclk = ps->medium.mclk;
+		if (mclk < ps->high.mclk)
+			mclk = ps->high.mclk;
+		ps->low.mclk = mclk;
+		ps->low.vddci = vddci;
+		ps->medium.mclk = mclk;
+		ps->medium.vddci = vddci;
+		ps->high.mclk = mclk;
+		ps->high.vddci = vddci;
+	} else {
+		if (ps->medium.mclk < ps->low.mclk)
+			ps->medium.mclk = ps->low.mclk;
+		if (ps->medium.vddci < ps->low.vddci)
+			ps->medium.vddci = ps->low.vddci;
+		if (ps->high.mclk < ps->medium.mclk)
+			ps->high.mclk = ps->medium.mclk;
+		if (ps->high.vddci < ps->medium.vddci)
+			ps->high.vddci = ps->medium.vddci;
+	}
+
+	btc_skip_blacklist_clocks(rdev, max_limits->sclk, max_limits->mclk,
+				  &ps->medium.sclk, &ps->medium.mclk);
+	btc_skip_blacklist_clocks(rdev, max_limits->sclk, max_limits->mclk,
+				  &ps->high.sclk, &ps->high.mclk);
+
+	btc_adjust_clock_combinations(rdev, max_limits, &ps->low);
+	btc_adjust_clock_combinations(rdev, max_limits, &ps->medium);
+	btc_adjust_clock_combinations(rdev, max_limits, &ps->high);
+
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
+					   ps->low.sclk, max_limits->vddc, &ps->low.vddc);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+					   ps->low.mclk, max_limits->vddci, &ps->low.vddci);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+					   ps->low.mclk, max_limits->vddc, &ps->low.vddc);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk,
+					   rdev->clock.current_dispclk, max_limits->vddc, &ps->low.vddc);
+
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
+					   ps->medium.sclk, max_limits->vddc, &ps->medium.vddc);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+					   ps->medium.mclk, max_limits->vddci, &ps->medium.vddci);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+					   ps->medium.mclk, max_limits->vddc, &ps->medium.vddc);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk,
+					   rdev->clock.current_dispclk, max_limits->vddc, &ps->medium.vddc);
+
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
+					   ps->high.sclk, max_limits->vddc, &ps->high.vddc);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+					   ps->high.mclk, max_limits->vddci, &ps->high.vddci);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+					   ps->high.mclk, max_limits->vddc, &ps->high.vddc);
+	btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk,
+					   rdev->clock.current_dispclk, max_limits->vddc, &ps->high.vddc);
+
+	btc_apply_voltage_delta_rules(rdev, max_limits->vddc, max_limits->vddci,
+				      &ps->low.vddc, &ps->low.vddci);
+	btc_apply_voltage_delta_rules(rdev, max_limits->vddc, max_limits->vddci,
+				      &ps->medium.vddc, &ps->medium.vddci);
+	btc_apply_voltage_delta_rules(rdev, max_limits->vddc, max_limits->vddci,
+				      &ps->high.vddc, &ps->high.vddci);
+
+	if ((ps->high.vddc <= rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddc) &&
+	    (ps->medium.vddc <= rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddc) &&
+	    (ps->low.vddc <= rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddc))
+		ps->dc_compatible = true;
+	else
+		ps->dc_compatible = false;
+
+	if (ps->low.vddc < rdev->pm.dpm.dyn_state.min_vddc_for_pcie_gen2)
+		ps->low.flags &= ~ATOM_PPLIB_R600_FLAGS_PCIEGEN2;
+	if (ps->medium.vddc < rdev->pm.dpm.dyn_state.min_vddc_for_pcie_gen2)
+		ps->medium.flags &= ~ATOM_PPLIB_R600_FLAGS_PCIEGEN2;
+	if (ps->high.vddc < rdev->pm.dpm.dyn_state.min_vddc_for_pcie_gen2)
+		ps->high.flags &= ~ATOM_PPLIB_R600_FLAGS_PCIEGEN2;
+}
+
+static void btc_update_current_ps(struct radeon_device *rdev,
+				  struct radeon_ps *rps)
+{
+	struct rv7xx_ps *new_ps = rv770_get_ps(rps);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	eg_pi->current_rps = *rps;
+	eg_pi->current_ps = *new_ps;
+	eg_pi->current_rps.ps_priv = &eg_pi->current_ps;
+}
+
+static void btc_update_requested_ps(struct radeon_device *rdev,
+				    struct radeon_ps *rps)
+{
+	struct rv7xx_ps *new_ps = rv770_get_ps(rps);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	eg_pi->requested_rps = *rps;
+	eg_pi->requested_ps = *new_ps;
+	eg_pi->requested_rps.ps_priv = &eg_pi->requested_ps;
+}
+
+void btc_dpm_reset_asic(struct radeon_device *rdev)
+{
+	rv770_restrict_performance_levels_before_switch(rdev);
+	btc_disable_ulv(rdev);
+	btc_set_boot_state_timing(rdev);
+	rv770_set_boot_state(rdev);
+}
+
+int btc_dpm_pre_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps requested_ps = *rdev->pm.dpm.requested_ps;
+	struct radeon_ps *new_ps = &requested_ps;
+
+	btc_update_requested_ps(rdev, new_ps);
+
+	btc_apply_state_adjust_rules(rdev, &eg_pi->requested_rps);
+
+	return 0;
+}
+
+int btc_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = &eg_pi->requested_rps;
+	struct radeon_ps *old_ps = &eg_pi->current_rps;
+	int ret;
+
+	ret = btc_disable_ulv(rdev);
+	btc_set_boot_state_timing(rdev);
+	ret = rv770_restrict_performance_levels_before_switch(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_restrict_performance_levels_before_switch failed\n");
+		return ret;
+	}
+	if (eg_pi->pcie_performance_request)
+		cypress_notify_link_speed_change_before_state_change(rdev, new_ps, old_ps);
+
+	rv770_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+	ret = rv770_halt_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_halt_smc failed\n");
+		return ret;
+	}
+	btc_set_at_for_uvd(rdev, new_ps);
+	if (eg_pi->smu_uvd_hs)
+		btc_notify_uvd_to_smc(rdev, new_ps);
+	ret = cypress_upload_sw_state(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("cypress_upload_sw_state failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = cypress_upload_mc_reg_table(rdev, new_ps);
+		if (ret) {
+			DRM_ERROR("cypress_upload_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+
+	cypress_program_memory_timing_parameters(rdev, new_ps);
+
+	ret = rv770_resume_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_resume_smc failed\n");
+		return ret;
+	}
+	ret = rv770_set_sw_state(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_set_sw_state failed\n");
+		return ret;
+	}
+	rv770_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+
+	if (eg_pi->pcie_performance_request)
+		cypress_notify_link_speed_change_after_state_change(rdev, new_ps, old_ps);
+
+	ret = btc_set_power_state_conditionally_enable_ulv(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("btc_set_power_state_conditionally_enable_ulv failed\n");
+		return ret;
+	}
+
+	ret = rv770_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO);
+	if (ret) {
+		DRM_ERROR("rv770_dpm_force_performance_level failed\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+void btc_dpm_post_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = &eg_pi->requested_rps;
+
+	btc_update_current_ps(rdev, new_ps);
+}
+
+int btc_dpm_enable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	if (pi->gfx_clock_gating)
+		btc_cg_clock_gating_default(rdev);
+
+	if (btc_dpm_enabled(rdev))
+		return -EINVAL;
+
+	if (pi->mg_clock_gating)
+		btc_mg_clock_gating_default(rdev);
+
+	if (eg_pi->ls_clock_gating)
+		btc_ls_clock_gating_default(rdev);
+
+	if (pi->voltage_control) {
+		rv770_enable_voltage_control(rdev, true);
+		ret = cypress_construct_voltage_tables(rdev);
+		if (ret) {
+			DRM_ERROR("cypress_construct_voltage_tables failed\n");
+			return ret;
+		}
+	}
+
+	if (pi->mvdd_control) {
+		ret = cypress_get_mvdd_configuration(rdev);
+		if (ret) {
+			DRM_ERROR("cypress_get_mvdd_configuration failed\n");
+			return ret;
+		}
+	}
+
+	if (eg_pi->dynamic_ac_timing) {
+		ret = btc_initialize_mc_reg_table(rdev);
+		if (ret)
+			eg_pi->dynamic_ac_timing = false;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv770_enable_backbias(rdev, true);
+
+	if (pi->dynamic_ss)
+		cypress_enable_spread_spectrum(rdev, true);
+
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, true);
+
+	rv770_setup_bsp(rdev);
+	rv770_program_git(rdev);
+	rv770_program_tp(rdev);
+	rv770_program_tpp(rdev);
+	rv770_program_sstp(rdev);
+	rv770_program_engine_speed_parameters(rdev);
+	cypress_enable_display_gap(rdev);
+	rv770_program_vc(rdev);
+
+	if (pi->dynamic_pcie_gen2)
+		btc_enable_dynamic_pcie_gen2(rdev, true);
+
+	ret = rv770_upload_firmware(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_upload_firmware failed\n");
+		return ret;
+	}
+	ret = cypress_get_table_locations(rdev);
+	if (ret) {
+		DRM_ERROR("cypress_get_table_locations failed\n");
+		return ret;
+	}
+	ret = btc_init_smc_table(rdev, boot_ps);
+	if (ret)
+		return ret;
+
+	if (eg_pi->dynamic_ac_timing) {
+		ret = cypress_populate_mc_reg_table(rdev, boot_ps);
+		if (ret) {
+			DRM_ERROR("cypress_populate_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+
+	cypress_program_response_times(rdev);
+	r7xx_start_smc(rdev);
+	ret = cypress_notify_smc_display_change(rdev, false);
+	if (ret) {
+		DRM_ERROR("cypress_notify_smc_display_change failed\n");
+		return ret;
+	}
+	cypress_enable_sclk_control(rdev, true);
+
+	if (eg_pi->memory_transition)
+		cypress_enable_mclk_control(rdev, true);
+
+	cypress_start_dpm(rdev);
+
+	if (pi->gfx_clock_gating)
+		btc_cg_clock_gating_enable(rdev, true);
+
+	if (pi->mg_clock_gating)
+		btc_mg_clock_gating_enable(rdev, true);
+
+	if (eg_pi->ls_clock_gating)
+		btc_ls_clock_gating_enable(rdev, true);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		PPSMC_Result result;
+
+		ret = rv770_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+		result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+
+		if (result != PPSMC_Result_OK)
+			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
+	}
+
+	rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
+
+	btc_init_stutter_mode(rdev);
+
+	btc_update_current_ps(rdev, rdev->pm.dpm.boot_ps);
+
+	return 0;
+};
+
+void btc_dpm_disable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (!btc_dpm_enabled(rdev))
+		return;
+
+	rv770_clear_vc(rdev);
+
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, false);
+
+	if (pi->dynamic_pcie_gen2)
+		btc_enable_dynamic_pcie_gen2(rdev, false);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	if (pi->gfx_clock_gating)
+		btc_cg_clock_gating_enable(rdev, false);
+
+	if (pi->mg_clock_gating)
+		btc_mg_clock_gating_enable(rdev, false);
+
+	if (eg_pi->ls_clock_gating)
+		btc_ls_clock_gating_enable(rdev, false);
+
+	rv770_stop_dpm(rdev);
+	btc_reset_to_default(rdev);
+	btc_stop_smc(rdev);
+	cypress_enable_spread_spectrum(rdev, false);
+
+	btc_update_current_ps(rdev, rdev->pm.dpm.boot_ps);
+}
+
+void btc_dpm_setup_asic(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	rv770_get_memory_type(rdev);
+	rv740_read_clock_registers(rdev);
+	btc_read_arb_registers(rdev);
+	rv770_read_voltage_smio_registers(rdev);
+
+	if (eg_pi->pcie_performance_request)
+		cypress_advertise_gen2_capability(rdev);
+
+	rv770_get_pcie_gen2_status(rdev);
+	rv770_enable_acpi_pm(rdev);
+}
+
+int btc_dpm_init(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi;
+	struct evergreen_power_info *eg_pi;
+	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	u16 data_offset, size;
+	u8 frev, crev;
+	struct atom_clock_dividers dividers;
+	int ret;
+
+	eg_pi = kzalloc(sizeof(struct evergreen_power_info), GFP_KERNEL);
+	if (eg_pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = eg_pi;
+	pi = &eg_pi->rv7xx;
+
+	rv770_get_max_vddc(rdev);
+
+	eg_pi->ulv.supported = false;
+	pi->acpi_vddc = 0;
+	eg_pi->acpi_vddci = 0;
+	pi->min_vddc_in_table = 0;
+	pi->max_vddc_in_table = 0;
+
+	ret = rv7xx_parse_power_table(rdev);
+	if (ret)
+		return ret;
+	ret = r600_parse_extended_power_table(rdev);
+	if (ret)
+		return ret;
+
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries =
+		kzalloc(4 * sizeof(struct radeon_clock_voltage_dependency_entry), GFP_KERNEL);
+	if (!rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries) {
+		r600_free_extended_power_table(rdev);
+		return -ENOMEM;
+	}
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.count = 4;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].clk = 0;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].v = 0;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].clk = 36000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].v = 800;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].clk = 54000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].v = 800;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].clk = 72000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].v = 800;
+
+	if (rdev->pm.dpm.voltage_response_time == 0)
+		rdev->pm.dpm.voltage_response_time = R600_VOLTAGERESPONSETIME_DFLT;
+	if (rdev->pm.dpm.backbias_response_time == 0)
+		rdev->pm.dpm.backbias_response_time = R600_BACKBIASRESPONSETIME_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->ref_div = dividers.ref_div + 1;
+	else
+		pi->ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	pi->mclk_strobe_mode_threshold = 40000;
+	pi->mclk_edc_enable_threshold = 40000;
+	eg_pi->mclk_edc_wr_enable_threshold = 40000;
+
+	pi->rlp = RV770_RLP_DFLT;
+	pi->rmp = RV770_RMP_DFLT;
+	pi->lhp = RV770_LHP_DFLT;
+	pi->lmp = RV770_LMP_DFLT;
+
+	eg_pi->ats[0].rlp = RV770_RLP_DFLT;
+	eg_pi->ats[0].rmp = RV770_RMP_DFLT;
+	eg_pi->ats[0].lhp = RV770_LHP_DFLT;
+	eg_pi->ats[0].lmp = RV770_LMP_DFLT;
+
+	eg_pi->ats[1].rlp = BTC_RLP_UVD_DFLT;
+	eg_pi->ats[1].rmp = BTC_RMP_UVD_DFLT;
+	eg_pi->ats[1].lhp = BTC_LHP_UVD_DFLT;
+	eg_pi->ats[1].lmp = BTC_LMP_UVD_DFLT;
+
+	eg_pi->smu_uvd_hs = true;
+
+	pi->voltage_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, 0);
+
+	pi->mvdd_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, 0);
+
+	eg_pi->vddci_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0);
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+                                   &frev, &crev, &data_offset)) {
+		pi->sclk_ss = true;
+		pi->mclk_ss = true;
+		pi->dynamic_ss = true;
+	} else {
+		pi->sclk_ss = false;
+		pi->mclk_ss = false;
+		pi->dynamic_ss = true;
+	}
+
+	pi->asi = RV770_ASI_DFLT;
+	pi->pasi = CYPRESS_HASI_DFLT;
+	pi->vrc = CYPRESS_VRC_DFLT;
+
+	pi->power_gating = false;
+
+	pi->gfx_clock_gating = true;
+
+	pi->mg_clock_gating = true;
+	pi->mgcgtssm = true;
+	eg_pi->ls_clock_gating = false;
+	eg_pi->sclk_deep_sleep = false;
+
+	pi->dynamic_pcie_gen2 = true;
+
+	if (pi->gfx_clock_gating &&
+	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+		pi->thermal_protection = true;
+	else
+		pi->thermal_protection = false;
+
+	pi->display_gap = true;
+
+	if (rdev->flags & RADEON_IS_MOBILITY)
+		pi->dcodt = true;
+	else
+		pi->dcodt = false;
+
+	pi->ulps = true;
+
+	eg_pi->dynamic_ac_timing = true;
+	eg_pi->abm = true;
+	eg_pi->mcls = true;
+	eg_pi->light_sleep = true;
+	eg_pi->memory_transition = true;
+#if defined(CONFIG_ACPI)
+	eg_pi->pcie_performance_request =
+		radeon_acpi_is_pcie_performance_request_supported(rdev);
+#else
+	eg_pi->pcie_performance_request = false;
+#endif
+
+	if (rdev->family == CHIP_BARTS)
+		eg_pi->dll_default_on = true;
+	else
+		eg_pi->dll_default_on = false;
+
+	eg_pi->sclk_deep_sleep = false;
+	if (ASIC_IS_LOMBOK(rdev))
+		pi->mclk_stutter_mode_threshold = 30000;
+	else
+		pi->mclk_stutter_mode_threshold = 0;
+
+	pi->sram_end = SMC_RAM_END;
+
+	rdev->pm.dpm.dyn_state.mclk_sclk_ratio = 4;
+	rdev->pm.dpm.dyn_state.vddc_vddci_delta = 200;
+	rdev->pm.dpm.dyn_state.min_vddc_for_pcie_gen2 = 900;
+	rdev->pm.dpm.dyn_state.valid_sclk_values.count = ARRAY_SIZE(btc_valid_sclk);
+	rdev->pm.dpm.dyn_state.valid_sclk_values.values = btc_valid_sclk;
+	rdev->pm.dpm.dyn_state.valid_mclk_values.count = 0;
+	rdev->pm.dpm.dyn_state.valid_mclk_values.values = NULL;
+
+	if (rdev->family == CHIP_TURKS)
+		rdev->pm.dpm.dyn_state.sclk_mclk_delta = 15000;
+	else
+		rdev->pm.dpm.dyn_state.sclk_mclk_delta = 10000;
+
+	return 0;
+}
+
+void btc_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+	kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries);
+	r600_free_extended_power_table(rdev);
+}
+
+u32 btc_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_ps *requested_state = rv770_get_ps(&eg_pi->requested_rps);
+
+	if (low)
+		return requested_state->low.sclk;
+	else
+		return requested_state->high.sclk;
+}
+
+u32 btc_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_ps *requested_state = rv770_get_ps(&eg_pi->requested_rps);
+
+	if (low)
+		return requested_state->low.mclk;
+	else
+		return requested_state->high.mclk;
+}
diff --git a/drivers/gpu/drm/radeon/btc_dpm.h b/drivers/gpu/drm/radeon/btc_dpm.h
new file mode 100644
index 000000000000..1a15e0e41950
--- /dev/null
+++ b/drivers/gpu/drm/radeon/btc_dpm.h
@@ -0,0 +1,57 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __BTC_DPM_H__
+#define __BTC_DPM_H__
+
+#define BTC_RLP_UVD_DFLT                              20
+#define BTC_RMP_UVD_DFLT                              50
+#define BTC_LHP_UVD_DFLT                              50
+#define BTC_LMP_UVD_DFLT                              20
+#define BARTS_MGCGCGTSSMCTRL_DFLT                     0x81944000
+#define TURKS_MGCGCGTSSMCTRL_DFLT                     0x6e944000
+#define CAICOS_MGCGCGTSSMCTRL_DFLT                    0x46944040
+#define BTC_CGULVPARAMETER_DFLT                       0x00040035
+#define BTC_CGULVCONTROL_DFLT                         0x00001450
+
+extern u32 btc_valid_sclk[40];
+
+void btc_read_arb_registers(struct radeon_device *rdev);
+void btc_program_mgcg_hw_sequence(struct radeon_device *rdev,
+				  const u32 *sequence, u32 count);
+void btc_skip_blacklist_clocks(struct radeon_device *rdev,
+			       const u32 max_sclk, const u32 max_mclk,
+			       u32 *sclk, u32 *mclk);
+void btc_adjust_clock_combinations(struct radeon_device *rdev,
+				   const struct radeon_clock_and_voltage_limits *max_limits,
+				   struct rv7xx_pl *pl);
+void btc_apply_voltage_dependency_rules(struct radeon_clock_voltage_dependency_table *table,
+					u32 clock, u16 max_voltage, u16 *voltage);
+void btc_apply_voltage_delta_rules(struct radeon_device *rdev,
+				   u16 max_vddc, u16 max_vddci,
+				   u16 *vddc, u16 *vddci);
+bool btc_dpm_enabled(struct radeon_device *rdev);
+int btc_reset_to_default(struct radeon_device *rdev);
+void btc_notify_uvd_to_smc(struct radeon_device *rdev,
+			   struct radeon_ps *radeon_new_state);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/btcd.h b/drivers/gpu/drm/radeon/btcd.h
new file mode 100644
index 000000000000..29e32de7e025
--- /dev/null
+++ b/drivers/gpu/drm/radeon/btcd.h
@@ -0,0 +1,181 @@
+/*
+ * Copyright 2010 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+#ifndef _BTCD_H_
+#define _BTCD_H_
+
+/* pm registers */
+
+#define GENERAL_PWRMGT                                  0x63c
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+#       define STATIC_PM_EN                             (1 << 1)
+#       define THERMAL_PROTECTION_DIS                   (1 << 2)
+#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
+#       define ENABLE_GEN2PCIE                          (1 << 4)
+#       define ENABLE_GEN2XSP                           (1 << 5)
+#       define SW_SMIO_INDEX(x)                         ((x) << 6)
+#       define SW_SMIO_INDEX_MASK                       (3 << 6)
+#       define SW_SMIO_INDEX_SHIFT                      6
+#       define LOW_VOLT_D2_ACPI                         (1 << 8)
+#       define LOW_VOLT_D3_ACPI                         (1 << 9)
+#       define VOLT_PWRMGT_EN                           (1 << 10)
+#       define BACKBIAS_PAD_EN                          (1 << 18)
+#       define BACKBIAS_VALUE                           (1 << 19)
+#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
+#       define AC_DC_SW                                 (1 << 24)
+
+#define	CG_BIF_REQ_AND_RSP				0x7f4
+#define		CG_CLIENT_REQ(x)			((x) << 0)
+#define		CG_CLIENT_REQ_MASK			(0xff << 0)
+#define		CG_CLIENT_REQ_SHIFT			0
+#define		CG_CLIENT_RESP(x)			((x) << 8)
+#define		CG_CLIENT_RESP_MASK			(0xff << 8)
+#define		CG_CLIENT_RESP_SHIFT			8
+#define		CLIENT_CG_REQ(x)			((x) << 16)
+#define		CLIENT_CG_REQ_MASK			(0xff << 16)
+#define		CLIENT_CG_REQ_SHIFT			16
+#define		CLIENT_CG_RESP(x)			((x) << 24)
+#define		CLIENT_CG_RESP_MASK			(0xff << 24)
+#define		CLIENT_CG_RESP_SHIFT			24
+
+#define	SCLK_PSKIP_CNTL					0x8c0
+#define		PSKIP_ON_ALLOW_STOP_HI(x)		((x) << 16)
+#define		PSKIP_ON_ALLOW_STOP_HI_MASK		(0xff << 16)
+#define		PSKIP_ON_ALLOW_STOP_HI_SHIFT		16
+
+#define	CG_ULV_CONTROL					0x8c8
+#define	CG_ULV_PARAMETER				0x8cc
+
+#define	MC_ARB_DRAM_TIMING				0x2774
+#define	MC_ARB_DRAM_TIMING2				0x2778
+
+#define	MC_ARB_RFSH_RATE				0x27b0
+#define		POWERMODE0(x)				((x) << 0)
+#define		POWERMODE0_MASK				(0xff << 0)
+#define		POWERMODE0_SHIFT			0
+#define		POWERMODE1(x)				((x) << 8)
+#define		POWERMODE1_MASK				(0xff << 8)
+#define		POWERMODE1_SHIFT			8
+#define		POWERMODE2(x)				((x) << 16)
+#define		POWERMODE2_MASK				(0xff << 16)
+#define		POWERMODE2_SHIFT			16
+#define		POWERMODE3(x)				((x) << 24)
+#define		POWERMODE3_MASK				(0xff << 24)
+#define		POWERMODE3_SHIFT			24
+
+#define MC_ARB_BURST_TIME                               0x2808
+#define		STATE0(x)				((x) << 0)
+#define		STATE0_MASK				(0x1f << 0)
+#define		STATE0_SHIFT				0
+#define		STATE1(x)				((x) << 5)
+#define		STATE1_MASK				(0x1f << 5)
+#define		STATE1_SHIFT				5
+#define		STATE2(x)				((x) << 10)
+#define		STATE2_MASK				(0x1f << 10)
+#define		STATE2_SHIFT				10
+#define		STATE3(x)				((x) << 15)
+#define		STATE3_MASK				(0x1f << 15)
+#define		STATE3_SHIFT				15
+
+#define MC_SEQ_RAS_TIMING                               0x28a0
+#define MC_SEQ_CAS_TIMING                               0x28a4
+#define MC_SEQ_MISC_TIMING                              0x28a8
+#define MC_SEQ_MISC_TIMING2                             0x28ac
+
+#define MC_SEQ_RD_CTL_D0                                0x28b4
+#define MC_SEQ_RD_CTL_D1                                0x28b8
+#define MC_SEQ_WR_CTL_D0                                0x28bc
+#define MC_SEQ_WR_CTL_D1                                0x28c0
+
+#define MC_PMG_AUTO_CFG                                 0x28d4
+
+#define MC_SEQ_STATUS_M                                 0x29f4
+#       define PMG_PWRSTATE                             (1 << 16)
+
+#define MC_SEQ_MISC0                                    0x2a00
+#define         MC_SEQ_MISC0_GDDR5_SHIFT                28
+#define         MC_SEQ_MISC0_GDDR5_MASK                 0xf0000000
+#define         MC_SEQ_MISC0_GDDR5_VALUE                5
+#define MC_SEQ_MISC1                                    0x2a04
+#define MC_SEQ_RESERVE_M                                0x2a08
+#define MC_PMG_CMD_EMRS                                 0x2a0c
+
+#define MC_SEQ_MISC3                                    0x2a2c
+
+#define MC_SEQ_MISC5                                    0x2a54
+#define MC_SEQ_MISC6                                    0x2a58
+
+#define MC_SEQ_MISC7                                    0x2a64
+
+#define MC_SEQ_CG                                       0x2a68
+#define		CG_SEQ_REQ(x)				((x) << 0)
+#define		CG_SEQ_REQ_MASK				(0xff << 0)
+#define		CG_SEQ_REQ_SHIFT			0
+#define		CG_SEQ_RESP(x)				((x) << 8)
+#define		CG_SEQ_RESP_MASK			(0xff << 8)
+#define		CG_SEQ_RESP_SHIFT			8
+#define		SEQ_CG_REQ(x)				((x) << 16)
+#define		SEQ_CG_REQ_MASK				(0xff << 16)
+#define		SEQ_CG_REQ_SHIFT			16
+#define		SEQ_CG_RESP(x)				((x) << 24)
+#define		SEQ_CG_RESP_MASK			(0xff << 24)
+#define		SEQ_CG_RESP_SHIFT			24
+#define MC_SEQ_RAS_TIMING_LP                            0x2a6c
+#define MC_SEQ_CAS_TIMING_LP                            0x2a70
+#define MC_SEQ_MISC_TIMING_LP                           0x2a74
+#define MC_SEQ_MISC_TIMING2_LP                          0x2a78
+#define MC_SEQ_WR_CTL_D0_LP                             0x2a7c
+#define MC_SEQ_WR_CTL_D1_LP                             0x2a80
+#define MC_SEQ_PMG_CMD_EMRS_LP                          0x2a84
+#define MC_SEQ_PMG_CMD_MRS_LP                           0x2a88
+
+#define MC_PMG_CMD_MRS                                  0x2aac
+
+#define MC_SEQ_RD_CTL_D0_LP                             0x2b1c
+#define MC_SEQ_RD_CTL_D1_LP                             0x2b20
+
+#define MC_PMG_CMD_MRS1                                 0x2b44
+#define MC_SEQ_PMG_CMD_MRS1_LP                          0x2b48
+
+#define	LB_SYNC_RESET_SEL				0x6b28
+#define		LB_SYNC_RESET_SEL_MASK			(3 << 0)
+#define		LB_SYNC_RESET_SEL_SHIFT			0
+
+/* PCIE link stuff */
+#define PCIE_LC_SPEED_CNTL                                0xa4 /* PCIE_P */
+#       define LC_GEN2_EN_STRAP                           (1 << 0)
+#       define LC_TARGET_LINK_SPEED_OVERRIDE_EN           (1 << 1)
+#       define LC_FORCE_EN_HW_SPEED_CHANGE                (1 << 5)
+#       define LC_FORCE_DIS_HW_SPEED_CHANGE               (1 << 6)
+#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK      (0x3 << 8)
+#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT     3
+#       define LC_CURRENT_DATA_RATE                       (1 << 11)
+#       define LC_HW_VOLTAGE_IF_CONTROL(x)                ((x) << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_MASK              (3 << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_SHIFT             12
+#       define LC_VOLTAGE_TIMER_SEL_MASK                  (0xf << 14)
+#       define LC_CLR_FAILED_SPD_CHANGE_CNT               (1 << 21)
+#       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 23)
+#       define LC_OTHER_SIDE_SUPPORTS_GEN2                (1 << 24)
+
+#endif
diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c
new file mode 100644
index 000000000000..6dacec4e2090
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cik.c
@@ -0,0 +1,7004 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+#include <linux/firmware.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include "drmP.h"
+#include "radeon.h"
+#include "radeon_asic.h"
+#include "cikd.h"
+#include "atom.h"
+#include "cik_blit_shaders.h"
+
+/* GFX */
+#define CIK_PFP_UCODE_SIZE 2144
+#define CIK_ME_UCODE_SIZE 2144
+#define CIK_CE_UCODE_SIZE 2144
+/* compute */
+#define CIK_MEC_UCODE_SIZE 4192
+/* interrupts */
+#define BONAIRE_RLC_UCODE_SIZE 2048
+#define KB_RLC_UCODE_SIZE 2560
+#define KV_RLC_UCODE_SIZE 2560
+/* gddr controller */
+#define CIK_MC_UCODE_SIZE 7866
+/* sdma */
+#define CIK_SDMA_UCODE_SIZE 1050
+#define CIK_SDMA_UCODE_VERSION 64
+
+MODULE_FIRMWARE("radeon/BONAIRE_pfp.bin");
+MODULE_FIRMWARE("radeon/BONAIRE_me.bin");
+MODULE_FIRMWARE("radeon/BONAIRE_ce.bin");
+MODULE_FIRMWARE("radeon/BONAIRE_mec.bin");
+MODULE_FIRMWARE("radeon/BONAIRE_mc.bin");
+MODULE_FIRMWARE("radeon/BONAIRE_rlc.bin");
+MODULE_FIRMWARE("radeon/BONAIRE_sdma.bin");
+MODULE_FIRMWARE("radeon/KAVERI_pfp.bin");
+MODULE_FIRMWARE("radeon/KAVERI_me.bin");
+MODULE_FIRMWARE("radeon/KAVERI_ce.bin");
+MODULE_FIRMWARE("radeon/KAVERI_mec.bin");
+MODULE_FIRMWARE("radeon/KAVERI_rlc.bin");
+MODULE_FIRMWARE("radeon/KAVERI_sdma.bin");
+MODULE_FIRMWARE("radeon/KABINI_pfp.bin");
+MODULE_FIRMWARE("radeon/KABINI_me.bin");
+MODULE_FIRMWARE("radeon/KABINI_ce.bin");
+MODULE_FIRMWARE("radeon/KABINI_mec.bin");
+MODULE_FIRMWARE("radeon/KABINI_rlc.bin");
+MODULE_FIRMWARE("radeon/KABINI_sdma.bin");
+
+extern int r600_ih_ring_alloc(struct radeon_device *rdev);
+extern void r600_ih_ring_fini(struct radeon_device *rdev);
+extern void evergreen_mc_stop(struct radeon_device *rdev, struct evergreen_mc_save *save);
+extern void evergreen_mc_resume(struct radeon_device *rdev, struct evergreen_mc_save *save);
+extern bool evergreen_is_display_hung(struct radeon_device *rdev);
+extern void si_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc);
+extern void si_rlc_fini(struct radeon_device *rdev);
+extern int si_rlc_init(struct radeon_device *rdev);
+static void cik_rlc_stop(struct radeon_device *rdev);
+
+/*
+ * Indirect registers accessor
+ */
+u32 cik_pciep_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(PCIE_INDEX, reg);
+	(void)RREG32(PCIE_INDEX);
+	r = RREG32(PCIE_DATA);
+	return r;
+}
+
+void cik_pciep_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(PCIE_INDEX, reg);
+	(void)RREG32(PCIE_INDEX);
+	WREG32(PCIE_DATA, v);
+	(void)RREG32(PCIE_DATA);
+}
+
+static const u32 bonaire_golden_spm_registers[] =
+{
+	0x30800, 0xe0ffffff, 0xe0000000
+};
+
+static const u32 bonaire_golden_common_registers[] =
+{
+	0xc770, 0xffffffff, 0x00000800,
+	0xc774, 0xffffffff, 0x00000800,
+	0xc798, 0xffffffff, 0x00007fbf,
+	0xc79c, 0xffffffff, 0x00007faf
+};
+
+static const u32 bonaire_golden_registers[] =
+{
+	0x3354, 0x00000333, 0x00000333,
+	0x3350, 0x000c0fc0, 0x00040200,
+	0x9a10, 0x00010000, 0x00058208,
+	0x3c000, 0xffff1fff, 0x00140000,
+	0x3c200, 0xfdfc0fff, 0x00000100,
+	0x3c234, 0x40000000, 0x40000200,
+	0x9830, 0xffffffff, 0x00000000,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x9838, 0x0002021c, 0x00020200,
+	0xc78, 0x00000080, 0x00000000,
+	0x5bb0, 0x000000f0, 0x00000070,
+	0x5bc0, 0xf0311fff, 0x80300000,
+	0x98f8, 0x73773777, 0x12010001,
+	0x350c, 0x00810000, 0x408af000,
+	0x7030, 0x31000111, 0x00000011,
+	0x2f48, 0x73773777, 0x12010001,
+	0x220c, 0x00007fb6, 0x0021a1b1,
+	0x2210, 0x00007fb6, 0x002021b1,
+	0x2180, 0x00007fb6, 0x00002191,
+	0x2218, 0x00007fb6, 0x002121b1,
+	0x221c, 0x00007fb6, 0x002021b1,
+	0x21dc, 0x00007fb6, 0x00002191,
+	0x21e0, 0x00007fb6, 0x00002191,
+	0x3628, 0x0000003f, 0x0000000a,
+	0x362c, 0x0000003f, 0x0000000a,
+	0x2ae4, 0x00073ffe, 0x000022a2,
+	0x240c, 0x000007ff, 0x00000000,
+	0x8a14, 0xf000003f, 0x00000007,
+	0x8bf0, 0x00002001, 0x00000001,
+	0x8b24, 0xffffffff, 0x00ffffff,
+	0x30a04, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x06000000,
+	0x4d8, 0x00000fff, 0x00000100,
+	0x3e78, 0x00000001, 0x00000002,
+	0x9100, 0x03000000, 0x0362c688,
+	0x8c00, 0x000000ff, 0x00000001,
+	0xe40, 0x00001fff, 0x00001fff,
+	0x9060, 0x0000007f, 0x00000020,
+	0x9508, 0x00010000, 0x00010000,
+	0xac14, 0x000003ff, 0x000000f3,
+	0xac0c, 0xffffffff, 0x00001032
+};
+
+static const u32 bonaire_mgcg_cgcg_init[] =
+{
+	0xc420, 0xffffffff, 0xfffffffc,
+	0x30800, 0xffffffff, 0xe0000000,
+	0x3c2a0, 0xffffffff, 0x00000100,
+	0x3c208, 0xffffffff, 0x00000100,
+	0x3c2c0, 0xffffffff, 0xc0000100,
+	0x3c2c8, 0xffffffff, 0xc0000100,
+	0x3c2c4, 0xffffffff, 0xc0000100,
+	0x55e4, 0xffffffff, 0x00600100,
+	0x3c280, 0xffffffff, 0x00000100,
+	0x3c214, 0xffffffff, 0x06000100,
+	0x3c220, 0xffffffff, 0x00000100,
+	0x3c218, 0xffffffff, 0x06000100,
+	0x3c204, 0xffffffff, 0x00000100,
+	0x3c2e0, 0xffffffff, 0x00000100,
+	0x3c224, 0xffffffff, 0x00000100,
+	0x3c200, 0xffffffff, 0x00000100,
+	0x3c230, 0xffffffff, 0x00000100,
+	0x3c234, 0xffffffff, 0x00000100,
+	0x3c250, 0xffffffff, 0x00000100,
+	0x3c254, 0xffffffff, 0x00000100,
+	0x3c258, 0xffffffff, 0x00000100,
+	0x3c25c, 0xffffffff, 0x00000100,
+	0x3c260, 0xffffffff, 0x00000100,
+	0x3c27c, 0xffffffff, 0x00000100,
+	0x3c278, 0xffffffff, 0x00000100,
+	0x3c210, 0xffffffff, 0x06000100,
+	0x3c290, 0xffffffff, 0x00000100,
+	0x3c274, 0xffffffff, 0x00000100,
+	0x3c2b4, 0xffffffff, 0x00000100,
+	0x3c2b0, 0xffffffff, 0x00000100,
+	0x3c270, 0xffffffff, 0x00000100,
+	0x30800, 0xffffffff, 0xe0000000,
+	0x3c020, 0xffffffff, 0x00010000,
+	0x3c024, 0xffffffff, 0x00030002,
+	0x3c028, 0xffffffff, 0x00040007,
+	0x3c02c, 0xffffffff, 0x00060005,
+	0x3c030, 0xffffffff, 0x00090008,
+	0x3c034, 0xffffffff, 0x00010000,
+	0x3c038, 0xffffffff, 0x00030002,
+	0x3c03c, 0xffffffff, 0x00040007,
+	0x3c040, 0xffffffff, 0x00060005,
+	0x3c044, 0xffffffff, 0x00090008,
+	0x3c048, 0xffffffff, 0x00010000,
+	0x3c04c, 0xffffffff, 0x00030002,
+	0x3c050, 0xffffffff, 0x00040007,
+	0x3c054, 0xffffffff, 0x00060005,
+	0x3c058, 0xffffffff, 0x00090008,
+	0x3c05c, 0xffffffff, 0x00010000,
+	0x3c060, 0xffffffff, 0x00030002,
+	0x3c064, 0xffffffff, 0x00040007,
+	0x3c068, 0xffffffff, 0x00060005,
+	0x3c06c, 0xffffffff, 0x00090008,
+	0x3c070, 0xffffffff, 0x00010000,
+	0x3c074, 0xffffffff, 0x00030002,
+	0x3c078, 0xffffffff, 0x00040007,
+	0x3c07c, 0xffffffff, 0x00060005,
+	0x3c080, 0xffffffff, 0x00090008,
+	0x3c084, 0xffffffff, 0x00010000,
+	0x3c088, 0xffffffff, 0x00030002,
+	0x3c08c, 0xffffffff, 0x00040007,
+	0x3c090, 0xffffffff, 0x00060005,
+	0x3c094, 0xffffffff, 0x00090008,
+	0x3c098, 0xffffffff, 0x00010000,
+	0x3c09c, 0xffffffff, 0x00030002,
+	0x3c0a0, 0xffffffff, 0x00040007,
+	0x3c0a4, 0xffffffff, 0x00060005,
+	0x3c0a8, 0xffffffff, 0x00090008,
+	0x3c000, 0xffffffff, 0x96e00200,
+	0x8708, 0xffffffff, 0x00900100,
+	0xc424, 0xffffffff, 0x0020003f,
+	0x38, 0xffffffff, 0x0140001c,
+	0x3c, 0x000f0000, 0x000f0000,
+	0x220, 0xffffffff, 0xC060000C,
+	0x224, 0xc0000fff, 0x00000100,
+	0xf90, 0xffffffff, 0x00000100,
+	0xf98, 0x00000101, 0x00000000,
+	0x20a8, 0xffffffff, 0x00000104,
+	0x55e4, 0xff000fff, 0x00000100,
+	0x30cc, 0xc0000fff, 0x00000104,
+	0xc1e4, 0x00000001, 0x00000001,
+	0xd00c, 0xff000ff0, 0x00000100,
+	0xd80c, 0xff000ff0, 0x00000100
+};
+
+static const u32 spectre_golden_spm_registers[] =
+{
+	0x30800, 0xe0ffffff, 0xe0000000
+};
+
+static const u32 spectre_golden_common_registers[] =
+{
+	0xc770, 0xffffffff, 0x00000800,
+	0xc774, 0xffffffff, 0x00000800,
+	0xc798, 0xffffffff, 0x00007fbf,
+	0xc79c, 0xffffffff, 0x00007faf
+};
+
+static const u32 spectre_golden_registers[] =
+{
+	0x3c000, 0xffff1fff, 0x96940200,
+	0x3c00c, 0xffff0001, 0xff000000,
+	0x3c200, 0xfffc0fff, 0x00000100,
+	0x6ed8, 0x00010101, 0x00010000,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x9838, 0xfffffffc, 0x00020200,
+	0x5bb0, 0x000000f0, 0x00000070,
+	0x5bc0, 0xf0311fff, 0x80300000,
+	0x98f8, 0x73773777, 0x12010001,
+	0x9b7c, 0x00ff0000, 0x00fc0000,
+	0x2f48, 0x73773777, 0x12010001,
+	0x8a14, 0xf000003f, 0x00000007,
+	0x8b24, 0xffffffff, 0x00ffffff,
+	0x28350, 0x3f3f3fff, 0x00000082,
+	0x28355, 0x0000003f, 0x00000000,
+	0x3e78, 0x00000001, 0x00000002,
+	0x913c, 0xffff03df, 0x00000004,
+	0xc768, 0x00000008, 0x00000008,
+	0x8c00, 0x000008ff, 0x00000800,
+	0x9508, 0x00010000, 0x00010000,
+	0xac0c, 0xffffffff, 0x54763210,
+	0x214f8, 0x01ff01ff, 0x00000002,
+	0x21498, 0x007ff800, 0x00200000,
+	0x2015c, 0xffffffff, 0x00000f40,
+	0x30934, 0xffffffff, 0x00000001
+};
+
+static const u32 spectre_mgcg_cgcg_init[] =
+{
+	0xc420, 0xffffffff, 0xfffffffc,
+	0x30800, 0xffffffff, 0xe0000000,
+	0x3c2a0, 0xffffffff, 0x00000100,
+	0x3c208, 0xffffffff, 0x00000100,
+	0x3c2c0, 0xffffffff, 0x00000100,
+	0x3c2c8, 0xffffffff, 0x00000100,
+	0x3c2c4, 0xffffffff, 0x00000100,
+	0x55e4, 0xffffffff, 0x00600100,
+	0x3c280, 0xffffffff, 0x00000100,
+	0x3c214, 0xffffffff, 0x06000100,
+	0x3c220, 0xffffffff, 0x00000100,
+	0x3c218, 0xffffffff, 0x06000100,
+	0x3c204, 0xffffffff, 0x00000100,
+	0x3c2e0, 0xffffffff, 0x00000100,
+	0x3c224, 0xffffffff, 0x00000100,
+	0x3c200, 0xffffffff, 0x00000100,
+	0x3c230, 0xffffffff, 0x00000100,
+	0x3c234, 0xffffffff, 0x00000100,
+	0x3c250, 0xffffffff, 0x00000100,
+	0x3c254, 0xffffffff, 0x00000100,
+	0x3c258, 0xffffffff, 0x00000100,
+	0x3c25c, 0xffffffff, 0x00000100,
+	0x3c260, 0xffffffff, 0x00000100,
+	0x3c27c, 0xffffffff, 0x00000100,
+	0x3c278, 0xffffffff, 0x00000100,
+	0x3c210, 0xffffffff, 0x06000100,
+	0x3c290, 0xffffffff, 0x00000100,
+	0x3c274, 0xffffffff, 0x00000100,
+	0x3c2b4, 0xffffffff, 0x00000100,
+	0x3c2b0, 0xffffffff, 0x00000100,
+	0x3c270, 0xffffffff, 0x00000100,
+	0x30800, 0xffffffff, 0xe0000000,
+	0x3c020, 0xffffffff, 0x00010000,
+	0x3c024, 0xffffffff, 0x00030002,
+	0x3c028, 0xffffffff, 0x00040007,
+	0x3c02c, 0xffffffff, 0x00060005,
+	0x3c030, 0xffffffff, 0x00090008,
+	0x3c034, 0xffffffff, 0x00010000,
+	0x3c038, 0xffffffff, 0x00030002,
+	0x3c03c, 0xffffffff, 0x00040007,
+	0x3c040, 0xffffffff, 0x00060005,
+	0x3c044, 0xffffffff, 0x00090008,
+	0x3c048, 0xffffffff, 0x00010000,
+	0x3c04c, 0xffffffff, 0x00030002,
+	0x3c050, 0xffffffff, 0x00040007,
+	0x3c054, 0xffffffff, 0x00060005,
+	0x3c058, 0xffffffff, 0x00090008,
+	0x3c05c, 0xffffffff, 0x00010000,
+	0x3c060, 0xffffffff, 0x00030002,
+	0x3c064, 0xffffffff, 0x00040007,
+	0x3c068, 0xffffffff, 0x00060005,
+	0x3c06c, 0xffffffff, 0x00090008,
+	0x3c070, 0xffffffff, 0x00010000,
+	0x3c074, 0xffffffff, 0x00030002,
+	0x3c078, 0xffffffff, 0x00040007,
+	0x3c07c, 0xffffffff, 0x00060005,
+	0x3c080, 0xffffffff, 0x00090008,
+	0x3c084, 0xffffffff, 0x00010000,
+	0x3c088, 0xffffffff, 0x00030002,
+	0x3c08c, 0xffffffff, 0x00040007,
+	0x3c090, 0xffffffff, 0x00060005,
+	0x3c094, 0xffffffff, 0x00090008,
+	0x3c098, 0xffffffff, 0x00010000,
+	0x3c09c, 0xffffffff, 0x00030002,
+	0x3c0a0, 0xffffffff, 0x00040007,
+	0x3c0a4, 0xffffffff, 0x00060005,
+	0x3c0a8, 0xffffffff, 0x00090008,
+	0x3c0ac, 0xffffffff, 0x00010000,
+	0x3c0b0, 0xffffffff, 0x00030002,
+	0x3c0b4, 0xffffffff, 0x00040007,
+	0x3c0b8, 0xffffffff, 0x00060005,
+	0x3c0bc, 0xffffffff, 0x00090008,
+	0x3c000, 0xffffffff, 0x96e00200,
+	0x8708, 0xffffffff, 0x00900100,
+	0xc424, 0xffffffff, 0x0020003f,
+	0x38, 0xffffffff, 0x0140001c,
+	0x3c, 0x000f0000, 0x000f0000,
+	0x220, 0xffffffff, 0xC060000C,
+	0x224, 0xc0000fff, 0x00000100,
+	0xf90, 0xffffffff, 0x00000100,
+	0xf98, 0x00000101, 0x00000000,
+	0x20a8, 0xffffffff, 0x00000104,
+	0x55e4, 0xff000fff, 0x00000100,
+	0x30cc, 0xc0000fff, 0x00000104,
+	0xc1e4, 0x00000001, 0x00000001,
+	0xd00c, 0xff000ff0, 0x00000100,
+	0xd80c, 0xff000ff0, 0x00000100
+};
+
+static const u32 kalindi_golden_spm_registers[] =
+{
+	0x30800, 0xe0ffffff, 0xe0000000
+};
+
+static const u32 kalindi_golden_common_registers[] =
+{
+	0xc770, 0xffffffff, 0x00000800,
+	0xc774, 0xffffffff, 0x00000800,
+	0xc798, 0xffffffff, 0x00007fbf,
+	0xc79c, 0xffffffff, 0x00007faf
+};
+
+static const u32 kalindi_golden_registers[] =
+{
+	0x3c000, 0xffffdfff, 0x6e944040,
+	0x55e4, 0xff607fff, 0xfc000100,
+	0x3c220, 0xff000fff, 0x00000100,
+	0x3c224, 0xff000fff, 0x00000100,
+	0x3c200, 0xfffc0fff, 0x00000100,
+	0x6ed8, 0x00010101, 0x00010000,
+	0x9830, 0xffffffff, 0x00000000,
+	0x9834, 0xf00fffff, 0x00000400,
+	0x5bb0, 0x000000f0, 0x00000070,
+	0x5bc0, 0xf0311fff, 0x80300000,
+	0x98f8, 0x73773777, 0x12010001,
+	0x98fc, 0xffffffff, 0x00000010,
+	0x9b7c, 0x00ff0000, 0x00fc0000,
+	0x8030, 0x00001f0f, 0x0000100a,
+	0x2f48, 0x73773777, 0x12010001,
+	0x2408, 0x000fffff, 0x000c007f,
+	0x8a14, 0xf000003f, 0x00000007,
+	0x8b24, 0x3fff3fff, 0x00ffcfff,
+	0x30a04, 0x0000ff0f, 0x00000000,
+	0x28a4c, 0x07ffffff, 0x06000000,
+	0x4d8, 0x00000fff, 0x00000100,
+	0x3e78, 0x00000001, 0x00000002,
+	0xc768, 0x00000008, 0x00000008,
+	0x8c00, 0x000000ff, 0x00000003,
+	0x214f8, 0x01ff01ff, 0x00000002,
+	0x21498, 0x007ff800, 0x00200000,
+	0x2015c, 0xffffffff, 0x00000f40,
+	0x88c4, 0x001f3ae3, 0x00000082,
+	0x88d4, 0x0000001f, 0x00000010,
+	0x30934, 0xffffffff, 0x00000000
+};
+
+static const u32 kalindi_mgcg_cgcg_init[] =
+{
+	0xc420, 0xffffffff, 0xfffffffc,
+	0x30800, 0xffffffff, 0xe0000000,
+	0x3c2a0, 0xffffffff, 0x00000100,
+	0x3c208, 0xffffffff, 0x00000100,
+	0x3c2c0, 0xffffffff, 0x00000100,
+	0x3c2c8, 0xffffffff, 0x00000100,
+	0x3c2c4, 0xffffffff, 0x00000100,
+	0x55e4, 0xffffffff, 0x00600100,
+	0x3c280, 0xffffffff, 0x00000100,
+	0x3c214, 0xffffffff, 0x06000100,
+	0x3c220, 0xffffffff, 0x00000100,
+	0x3c218, 0xffffffff, 0x06000100,
+	0x3c204, 0xffffffff, 0x00000100,
+	0x3c2e0, 0xffffffff, 0x00000100,
+	0x3c224, 0xffffffff, 0x00000100,
+	0x3c200, 0xffffffff, 0x00000100,
+	0x3c230, 0xffffffff, 0x00000100,
+	0x3c234, 0xffffffff, 0x00000100,
+	0x3c250, 0xffffffff, 0x00000100,
+	0x3c254, 0xffffffff, 0x00000100,
+	0x3c258, 0xffffffff, 0x00000100,
+	0x3c25c, 0xffffffff, 0x00000100,
+	0x3c260, 0xffffffff, 0x00000100,
+	0x3c27c, 0xffffffff, 0x00000100,
+	0x3c278, 0xffffffff, 0x00000100,
+	0x3c210, 0xffffffff, 0x06000100,
+	0x3c290, 0xffffffff, 0x00000100,
+	0x3c274, 0xffffffff, 0x00000100,
+	0x3c2b4, 0xffffffff, 0x00000100,
+	0x3c2b0, 0xffffffff, 0x00000100,
+	0x3c270, 0xffffffff, 0x00000100,
+	0x30800, 0xffffffff, 0xe0000000,
+	0x3c020, 0xffffffff, 0x00010000,
+	0x3c024, 0xffffffff, 0x00030002,
+	0x3c028, 0xffffffff, 0x00040007,
+	0x3c02c, 0xffffffff, 0x00060005,
+	0x3c030, 0xffffffff, 0x00090008,
+	0x3c034, 0xffffffff, 0x00010000,
+	0x3c038, 0xffffffff, 0x00030002,
+	0x3c03c, 0xffffffff, 0x00040007,
+	0x3c040, 0xffffffff, 0x00060005,
+	0x3c044, 0xffffffff, 0x00090008,
+	0x3c000, 0xffffffff, 0x96e00200,
+	0x8708, 0xffffffff, 0x00900100,
+	0xc424, 0xffffffff, 0x0020003f,
+	0x38, 0xffffffff, 0x0140001c,
+	0x3c, 0x000f0000, 0x000f0000,
+	0x220, 0xffffffff, 0xC060000C,
+	0x224, 0xc0000fff, 0x00000100,
+	0x20a8, 0xffffffff, 0x00000104,
+	0x55e4, 0xff000fff, 0x00000100,
+	0x30cc, 0xc0000fff, 0x00000104,
+	0xc1e4, 0x00000001, 0x00000001,
+	0xd00c, 0xff000ff0, 0x00000100,
+	0xd80c, 0xff000ff0, 0x00000100
+};
+
+static void cik_init_golden_registers(struct radeon_device *rdev)
+{
+	switch (rdev->family) {
+	case CHIP_BONAIRE:
+		radeon_program_register_sequence(rdev,
+						 bonaire_mgcg_cgcg_init,
+						 (const u32)ARRAY_SIZE(bonaire_mgcg_cgcg_init));
+		radeon_program_register_sequence(rdev,
+						 bonaire_golden_registers,
+						 (const u32)ARRAY_SIZE(bonaire_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 bonaire_golden_common_registers,
+						 (const u32)ARRAY_SIZE(bonaire_golden_common_registers));
+		radeon_program_register_sequence(rdev,
+						 bonaire_golden_spm_registers,
+						 (const u32)ARRAY_SIZE(bonaire_golden_spm_registers));
+		break;
+	case CHIP_KABINI:
+		radeon_program_register_sequence(rdev,
+						 kalindi_mgcg_cgcg_init,
+						 (const u32)ARRAY_SIZE(kalindi_mgcg_cgcg_init));
+		radeon_program_register_sequence(rdev,
+						 kalindi_golden_registers,
+						 (const u32)ARRAY_SIZE(kalindi_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 kalindi_golden_common_registers,
+						 (const u32)ARRAY_SIZE(kalindi_golden_common_registers));
+		radeon_program_register_sequence(rdev,
+						 kalindi_golden_spm_registers,
+						 (const u32)ARRAY_SIZE(kalindi_golden_spm_registers));
+		break;
+	case CHIP_KAVERI:
+		radeon_program_register_sequence(rdev,
+						 spectre_mgcg_cgcg_init,
+						 (const u32)ARRAY_SIZE(spectre_mgcg_cgcg_init));
+		radeon_program_register_sequence(rdev,
+						 spectre_golden_registers,
+						 (const u32)ARRAY_SIZE(spectre_golden_registers));
+		radeon_program_register_sequence(rdev,
+						 spectre_golden_common_registers,
+						 (const u32)ARRAY_SIZE(spectre_golden_common_registers));
+		radeon_program_register_sequence(rdev,
+						 spectre_golden_spm_registers,
+						 (const u32)ARRAY_SIZE(spectre_golden_spm_registers));
+		break;
+	default:
+		break;
+	}
+}
+
+/**
+ * cik_get_xclk - get the xclk
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Returns the reference clock used by the gfx engine
+ * (CIK).
+ */
+u32 cik_get_xclk(struct radeon_device *rdev)
+{
+        u32 reference_clock = rdev->clock.spll.reference_freq;
+
+	if (rdev->flags & RADEON_IS_IGP) {
+		if (RREG32_SMC(GENERAL_PWRMGT) & GPU_COUNTER_CLK)
+			return reference_clock / 2;
+	} else {
+		if (RREG32_SMC(CG_CLKPIN_CNTL) & XTALIN_DIVIDE)
+			return reference_clock / 4;
+	}
+	return reference_clock;
+}
+
+/**
+ * cik_mm_rdoorbell - read a doorbell dword
+ *
+ * @rdev: radeon_device pointer
+ * @offset: byte offset into the aperture
+ *
+ * Returns the value in the doorbell aperture at the
+ * requested offset (CIK).
+ */
+u32 cik_mm_rdoorbell(struct radeon_device *rdev, u32 offset)
+{
+	if (offset < rdev->doorbell.size) {
+		return readl(((void __iomem *)rdev->doorbell.ptr) + offset);
+	} else {
+		DRM_ERROR("reading beyond doorbell aperture: 0x%08x!\n", offset);
+		return 0;
+	}
+}
+
+/**
+ * cik_mm_wdoorbell - write a doorbell dword
+ *
+ * @rdev: radeon_device pointer
+ * @offset: byte offset into the aperture
+ * @v: value to write
+ *
+ * Writes @v to the doorbell aperture at the
+ * requested offset (CIK).
+ */
+void cik_mm_wdoorbell(struct radeon_device *rdev, u32 offset, u32 v)
+{
+	if (offset < rdev->doorbell.size) {
+		writel(v, ((void __iomem *)rdev->doorbell.ptr) + offset);
+	} else {
+		DRM_ERROR("writing beyond doorbell aperture: 0x%08x!\n", offset);
+	}
+}
+
+#define BONAIRE_IO_MC_REGS_SIZE 36
+
+static const u32 bonaire_io_mc_regs[BONAIRE_IO_MC_REGS_SIZE][2] =
+{
+	{0x00000070, 0x04400000},
+	{0x00000071, 0x80c01803},
+	{0x00000072, 0x00004004},
+	{0x00000073, 0x00000100},
+	{0x00000074, 0x00ff0000},
+	{0x00000075, 0x34000000},
+	{0x00000076, 0x08000014},
+	{0x00000077, 0x00cc08ec},
+	{0x00000078, 0x00000400},
+	{0x00000079, 0x00000000},
+	{0x0000007a, 0x04090000},
+	{0x0000007c, 0x00000000},
+	{0x0000007e, 0x4408a8e8},
+	{0x0000007f, 0x00000304},
+	{0x00000080, 0x00000000},
+	{0x00000082, 0x00000001},
+	{0x00000083, 0x00000002},
+	{0x00000084, 0xf3e4f400},
+	{0x00000085, 0x052024e3},
+	{0x00000087, 0x00000000},
+	{0x00000088, 0x01000000},
+	{0x0000008a, 0x1c0a0000},
+	{0x0000008b, 0xff010000},
+	{0x0000008d, 0xffffefff},
+	{0x0000008e, 0xfff3efff},
+	{0x0000008f, 0xfff3efbf},
+	{0x00000092, 0xf7ffffff},
+	{0x00000093, 0xffffff7f},
+	{0x00000095, 0x00101101},
+	{0x00000096, 0x00000fff},
+	{0x00000097, 0x00116fff},
+	{0x00000098, 0x60010000},
+	{0x00000099, 0x10010000},
+	{0x0000009a, 0x00006000},
+	{0x0000009b, 0x00001000},
+	{0x0000009f, 0x00b48000}
+};
+
+/**
+ * cik_srbm_select - select specific register instances
+ *
+ * @rdev: radeon_device pointer
+ * @me: selected ME (micro engine)
+ * @pipe: pipe
+ * @queue: queue
+ * @vmid: VMID
+ *
+ * Switches the currently active registers instances.  Some
+ * registers are instanced per VMID, others are instanced per
+ * me/pipe/queue combination.
+ */
+static void cik_srbm_select(struct radeon_device *rdev,
+			    u32 me, u32 pipe, u32 queue, u32 vmid)
+{
+	u32 srbm_gfx_cntl = (PIPEID(pipe & 0x3) |
+			     MEID(me & 0x3) |
+			     VMID(vmid & 0xf) |
+			     QUEUEID(queue & 0x7));
+	WREG32(SRBM_GFX_CNTL, srbm_gfx_cntl);
+}
+
+/* ucode loading */
+/**
+ * ci_mc_load_microcode - load MC ucode into the hw
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Load the GDDR MC ucode into the hw (CIK).
+ * Returns 0 on success, error on failure.
+ */
+static int ci_mc_load_microcode(struct radeon_device *rdev)
+{
+	const __be32 *fw_data;
+	u32 running, blackout = 0;
+	u32 *io_mc_regs;
+	int i, ucode_size, regs_size;
+
+	if (!rdev->mc_fw)
+		return -EINVAL;
+
+	switch (rdev->family) {
+	case CHIP_BONAIRE:
+	default:
+		io_mc_regs = (u32 *)&bonaire_io_mc_regs;
+		ucode_size = CIK_MC_UCODE_SIZE;
+		regs_size = BONAIRE_IO_MC_REGS_SIZE;
+		break;
+	}
+
+	running = RREG32(MC_SEQ_SUP_CNTL) & RUN_MASK;
+
+	if (running == 0) {
+		if (running) {
+			blackout = RREG32(MC_SHARED_BLACKOUT_CNTL);
+			WREG32(MC_SHARED_BLACKOUT_CNTL, blackout | 1);
+		}
+
+		/* reset the engine and set to writable */
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000008);
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000010);
+
+		/* load mc io regs */
+		for (i = 0; i < regs_size; i++) {
+			WREG32(MC_SEQ_IO_DEBUG_INDEX, io_mc_regs[(i << 1)]);
+			WREG32(MC_SEQ_IO_DEBUG_DATA, io_mc_regs[(i << 1) + 1]);
+		}
+		/* load the MC ucode */
+		fw_data = (const __be32 *)rdev->mc_fw->data;
+		for (i = 0; i < ucode_size; i++)
+			WREG32(MC_SEQ_SUP_PGM, be32_to_cpup(fw_data++));
+
+		/* put the engine back into the active state */
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000008);
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000004);
+		WREG32(MC_SEQ_SUP_CNTL, 0x00000001);
+
+		/* wait for training to complete */
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if (RREG32(MC_SEQ_TRAIN_WAKEUP_CNTL) & TRAIN_DONE_D0)
+				break;
+			udelay(1);
+		}
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if (RREG32(MC_SEQ_TRAIN_WAKEUP_CNTL) & TRAIN_DONE_D1)
+				break;
+			udelay(1);
+		}
+
+		if (running)
+			WREG32(MC_SHARED_BLACKOUT_CNTL, blackout);
+	}
+
+	return 0;
+}
+
+/**
+ * cik_init_microcode - load ucode images from disk
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Use the firmware interface to load the ucode images into
+ * the driver (not loaded into hw).
+ * Returns 0 on success, error on failure.
+ */
+static int cik_init_microcode(struct radeon_device *rdev)
+{
+	const char *chip_name;
+	size_t pfp_req_size, me_req_size, ce_req_size,
+		mec_req_size, rlc_req_size, mc_req_size,
+		sdma_req_size;
+	char fw_name[30];
+	int err;
+
+	DRM_DEBUG("\n");
+
+	switch (rdev->family) {
+	case CHIP_BONAIRE:
+		chip_name = "BONAIRE";
+		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
+		me_req_size = CIK_ME_UCODE_SIZE * 4;
+		ce_req_size = CIK_CE_UCODE_SIZE * 4;
+		mec_req_size = CIK_MEC_UCODE_SIZE * 4;
+		rlc_req_size = BONAIRE_RLC_UCODE_SIZE * 4;
+		mc_req_size = CIK_MC_UCODE_SIZE * 4;
+		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
+		break;
+	case CHIP_KAVERI:
+		chip_name = "KAVERI";
+		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
+		me_req_size = CIK_ME_UCODE_SIZE * 4;
+		ce_req_size = CIK_CE_UCODE_SIZE * 4;
+		mec_req_size = CIK_MEC_UCODE_SIZE * 4;
+		rlc_req_size = KV_RLC_UCODE_SIZE * 4;
+		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
+		break;
+	case CHIP_KABINI:
+		chip_name = "KABINI";
+		pfp_req_size = CIK_PFP_UCODE_SIZE * 4;
+		me_req_size = CIK_ME_UCODE_SIZE * 4;
+		ce_req_size = CIK_CE_UCODE_SIZE * 4;
+		mec_req_size = CIK_MEC_UCODE_SIZE * 4;
+		rlc_req_size = KB_RLC_UCODE_SIZE * 4;
+		sdma_req_size = CIK_SDMA_UCODE_SIZE * 4;
+		break;
+	default: BUG();
+	}
+
+	DRM_INFO("Loading %s Microcode\n", chip_name);
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
+	err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev);
+	if (err)
+		goto out;
+	if (rdev->pfp_fw->size != pfp_req_size) {
+		printk(KERN_ERR
+		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->pfp_fw->size, fw_name);
+		err = -EINVAL;
+		goto out;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
+	err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
+	if (err)
+		goto out;
+	if (rdev->me_fw->size != me_req_size) {
+		printk(KERN_ERR
+		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->me_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
+	err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev);
+	if (err)
+		goto out;
+	if (rdev->ce_fw->size != ce_req_size) {
+		printk(KERN_ERR
+		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->ce_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mec.bin", chip_name);
+	err = request_firmware(&rdev->mec_fw, fw_name, rdev->dev);
+	if (err)
+		goto out;
+	if (rdev->mec_fw->size != mec_req_size) {
+		printk(KERN_ERR
+		       "cik_cp: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->mec_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", chip_name);
+	err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev);
+	if (err)
+		goto out;
+	if (rdev->rlc_fw->size != rlc_req_size) {
+		printk(KERN_ERR
+		       "cik_rlc: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->rlc_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_sdma.bin", chip_name);
+	err = request_firmware(&rdev->sdma_fw, fw_name, rdev->dev);
+	if (err)
+		goto out;
+	if (rdev->sdma_fw->size != sdma_req_size) {
+		printk(KERN_ERR
+		       "cik_sdma: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->sdma_fw->size, fw_name);
+		err = -EINVAL;
+	}
+
+	/* No MC ucode on APUs */
+	if (!(rdev->flags & RADEON_IS_IGP)) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
+		err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->mc_fw->size != mc_req_size) {
+			printk(KERN_ERR
+			       "cik_mc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->mc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	}
+
+out:
+	if (err) {
+		if (err != -EINVAL)
+			printk(KERN_ERR
+			       "cik_cp: Failed to load firmware \"%s\"\n",
+			       fw_name);
+		release_firmware(rdev->pfp_fw);
+		rdev->pfp_fw = NULL;
+		release_firmware(rdev->me_fw);
+		rdev->me_fw = NULL;
+		release_firmware(rdev->ce_fw);
+		rdev->ce_fw = NULL;
+		release_firmware(rdev->rlc_fw);
+		rdev->rlc_fw = NULL;
+		release_firmware(rdev->mc_fw);
+		rdev->mc_fw = NULL;
+	}
+	return err;
+}
+
+/*
+ * Core functions
+ */
+/**
+ * cik_tiling_mode_table_init - init the hw tiling table
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Starting with SI, the tiling setup is done globally in a
+ * set of 32 tiling modes.  Rather than selecting each set of
+ * parameters per surface as on older asics, we just select
+ * which index in the tiling table we want to use, and the
+ * surface uses those parameters (CIK).
+ */
+static void cik_tiling_mode_table_init(struct radeon_device *rdev)
+{
+	const u32 num_tile_mode_states = 32;
+	const u32 num_secondary_tile_mode_states = 16;
+	u32 reg_offset, gb_tile_moden, split_equal_to_row_size;
+	u32 num_pipe_configs;
+	u32 num_rbs = rdev->config.cik.max_backends_per_se *
+		rdev->config.cik.max_shader_engines;
+
+	switch (rdev->config.cik.mem_row_size_in_kb) {
+	case 1:
+		split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_1KB;
+		break;
+	case 2:
+	default:
+		split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_2KB;
+		break;
+	case 4:
+		split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_4KB;
+		break;
+	}
+
+	num_pipe_configs = rdev->config.cik.max_tile_pipes;
+	if (num_pipe_configs > 8)
+		num_pipe_configs = 8; /* ??? */
+
+	if (num_pipe_configs == 8) {
+		for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B));
+				break;
+			case 1:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B));
+				break;
+			case 2:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+				break;
+			case 3:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B));
+				break;
+			case 4:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(split_equal_to_row_size));
+				break;
+			case 5:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING));
+				break;
+			case 6:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+				break;
+			case 7:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 TILE_SPLIT(split_equal_to_row_size));
+				break;
+			case 8:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16));
+				break;
+			case 9:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING));
+				break;
+			case 10:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 11:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 12:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 13:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING));
+				break;
+			case 14:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 16:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 17:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 27:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING));
+				break;
+			case 28:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 29:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 30:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_16x16) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			rdev->config.cik.tile_mode_array[reg_offset] = gb_tile_moden;
+			WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+		}
+		for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 1:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 2:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 3:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 4:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_8_BANK));
+				break;
+			case 5:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_4_BANK));
+				break;
+			case 6:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_2_BANK));
+				break;
+			case 8:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 9:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 10:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 11:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 12:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_8_BANK));
+				break;
+			case 13:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_4_BANK));
+				break;
+			case 14:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_2_BANK));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			WREG32(GB_MACROTILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+		}
+	} else if (num_pipe_configs == 4) {
+		if (num_rbs == 4) {
+			for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
+				switch (reg_offset) {
+				case 0:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B));
+					break;
+				case 1:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B));
+					break;
+				case 2:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+					break;
+				case 3:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B));
+					break;
+				case 4:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(split_equal_to_row_size));
+					break;
+				case 5:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING));
+					break;
+				case 6:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+					break;
+				case 7:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 TILE_SPLIT(split_equal_to_row_size));
+					break;
+				case 8:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16));
+					break;
+				case 9:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING));
+					break;
+				case 10:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 11:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 12:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 13:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING));
+					break;
+				case 14:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 16:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 17:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 27:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING));
+					break;
+				case 28:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 29:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 30:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_16x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				default:
+					gb_tile_moden = 0;
+					break;
+				}
+				rdev->config.cik.tile_mode_array[reg_offset] = gb_tile_moden;
+				WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+			}
+		} else if (num_rbs < 4) {
+			for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
+				switch (reg_offset) {
+				case 0:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B));
+					break;
+				case 1:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B));
+					break;
+				case 2:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+					break;
+				case 3:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B));
+					break;
+				case 4:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(split_equal_to_row_size));
+					break;
+				case 5:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING));
+					break;
+				case 6:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+					break;
+				case 7:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 TILE_SPLIT(split_equal_to_row_size));
+					break;
+				case 8:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16));
+					break;
+				case 9:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING));
+					break;
+				case 10:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 11:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 12:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 13:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING));
+					break;
+				case 14:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 16:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 17:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 27:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING));
+					break;
+				case 28:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 29:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				case 30:
+					gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+							 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+							 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+							 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+					break;
+				default:
+					gb_tile_moden = 0;
+					break;
+				}
+				rdev->config.cik.tile_mode_array[reg_offset] = gb_tile_moden;
+				WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+			}
+		}
+		for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 1:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 2:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 3:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 4:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 5:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_8_BANK));
+				break;
+			case 6:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_4_BANK));
+				break;
+			case 8:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 9:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 10:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 11:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 12:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 13:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_8_BANK));
+				break;
+			case 14:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) |
+						 NUM_BANKS(ADDR_SURF_4_BANK));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			WREG32(GB_MACROTILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+		}
+	} else if (num_pipe_configs == 2) {
+		for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B));
+				break;
+			case 1:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B));
+				break;
+			case 2:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+				break;
+			case 3:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B));
+				break;
+			case 4:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(split_equal_to_row_size));
+				break;
+			case 5:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING));
+				break;
+			case 6:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B));
+				break;
+			case 7:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 TILE_SPLIT(split_equal_to_row_size));
+				break;
+			case 8:
+				gb_tile_moden = ARRAY_MODE(ARRAY_LINEAR_ALIGNED);
+				break;
+			case 9:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING));
+				break;
+			case 10:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 11:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 12:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 13:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING));
+				break;
+			case 14:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 16:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 17:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 27:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING));
+				break;
+			case 28:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 29:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			case 30:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P2) |
+						 SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			rdev->config.cik.tile_mode_array[reg_offset] = gb_tile_moden;
+			WREG32(GB_TILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+		}
+		for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 1:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 2:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 3:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 4:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 5:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 6:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_8_BANK));
+				break;
+			case 8:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_4) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 9:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_4) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 10:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 11:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 12:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 13:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) |
+						 NUM_BANKS(ADDR_SURF_16_BANK));
+				break;
+			case 14:
+				gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) |
+						 NUM_BANKS(ADDR_SURF_8_BANK));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			WREG32(GB_MACROTILE_MODE0 + (reg_offset * 4), gb_tile_moden);
+		}
+	} else
+		DRM_ERROR("unknown num pipe config: 0x%x\n", num_pipe_configs);
+}
+
+/**
+ * cik_select_se_sh - select which SE, SH to address
+ *
+ * @rdev: radeon_device pointer
+ * @se_num: shader engine to address
+ * @sh_num: sh block to address
+ *
+ * Select which SE, SH combinations to address. Certain
+ * registers are instanced per SE or SH.  0xffffffff means
+ * broadcast to all SEs or SHs (CIK).
+ */
+static void cik_select_se_sh(struct radeon_device *rdev,
+			     u32 se_num, u32 sh_num)
+{
+	u32 data = INSTANCE_BROADCAST_WRITES;
+
+	if ((se_num == 0xffffffff) && (sh_num == 0xffffffff))
+		data |= SH_BROADCAST_WRITES | SE_BROADCAST_WRITES;
+	else if (se_num == 0xffffffff)
+		data |= SE_BROADCAST_WRITES | SH_INDEX(sh_num);
+	else if (sh_num == 0xffffffff)
+		data |= SH_BROADCAST_WRITES | SE_INDEX(se_num);
+	else
+		data |= SH_INDEX(sh_num) | SE_INDEX(se_num);
+	WREG32(GRBM_GFX_INDEX, data);
+}
+
+/**
+ * cik_create_bitmask - create a bitmask
+ *
+ * @bit_width: length of the mask
+ *
+ * create a variable length bit mask (CIK).
+ * Returns the bitmask.
+ */
+static u32 cik_create_bitmask(u32 bit_width)
+{
+	u32 i, mask = 0;
+
+	for (i = 0; i < bit_width; i++) {
+		mask <<= 1;
+		mask |= 1;
+	}
+	return mask;
+}
+
+/**
+ * cik_select_se_sh - select which SE, SH to address
+ *
+ * @rdev: radeon_device pointer
+ * @max_rb_num: max RBs (render backends) for the asic
+ * @se_num: number of SEs (shader engines) for the asic
+ * @sh_per_se: number of SH blocks per SE for the asic
+ *
+ * Calculates the bitmask of disabled RBs (CIK).
+ * Returns the disabled RB bitmask.
+ */
+static u32 cik_get_rb_disabled(struct radeon_device *rdev,
+			      u32 max_rb_num, u32 se_num,
+			      u32 sh_per_se)
+{
+	u32 data, mask;
+
+	data = RREG32(CC_RB_BACKEND_DISABLE);
+	if (data & 1)
+		data &= BACKEND_DISABLE_MASK;
+	else
+		data = 0;
+	data |= RREG32(GC_USER_RB_BACKEND_DISABLE);
+
+	data >>= BACKEND_DISABLE_SHIFT;
+
+	mask = cik_create_bitmask(max_rb_num / se_num / sh_per_se);
+
+	return data & mask;
+}
+
+/**
+ * cik_setup_rb - setup the RBs on the asic
+ *
+ * @rdev: radeon_device pointer
+ * @se_num: number of SEs (shader engines) for the asic
+ * @sh_per_se: number of SH blocks per SE for the asic
+ * @max_rb_num: max RBs (render backends) for the asic
+ *
+ * Configures per-SE/SH RB registers (CIK).
+ */
+static void cik_setup_rb(struct radeon_device *rdev,
+			 u32 se_num, u32 sh_per_se,
+			 u32 max_rb_num)
+{
+	int i, j;
+	u32 data, mask;
+	u32 disabled_rbs = 0;
+	u32 enabled_rbs = 0;
+
+	for (i = 0; i < se_num; i++) {
+		for (j = 0; j < sh_per_se; j++) {
+			cik_select_se_sh(rdev, i, j);
+			data = cik_get_rb_disabled(rdev, max_rb_num, se_num, sh_per_se);
+			disabled_rbs |= data << ((i * sh_per_se + j) * CIK_RB_BITMAP_WIDTH_PER_SH);
+		}
+	}
+	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+
+	mask = 1;
+	for (i = 0; i < max_rb_num; i++) {
+		if (!(disabled_rbs & mask))
+			enabled_rbs |= mask;
+		mask <<= 1;
+	}
+
+	for (i = 0; i < se_num; i++) {
+		cik_select_se_sh(rdev, i, 0xffffffff);
+		data = 0;
+		for (j = 0; j < sh_per_se; j++) {
+			switch (enabled_rbs & 3) {
+			case 1:
+				data |= (RASTER_CONFIG_RB_MAP_0 << (i * sh_per_se + j) * 2);
+				break;
+			case 2:
+				data |= (RASTER_CONFIG_RB_MAP_3 << (i * sh_per_se + j) * 2);
+				break;
+			case 3:
+			default:
+				data |= (RASTER_CONFIG_RB_MAP_2 << (i * sh_per_se + j) * 2);
+				break;
+			}
+			enabled_rbs >>= 2;
+		}
+		WREG32(PA_SC_RASTER_CONFIG, data);
+	}
+	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+}
+
+/**
+ * cik_gpu_init - setup the 3D engine
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Configures the 3D engine and tiling configuration
+ * registers so that the 3D engine is usable.
+ */
+static void cik_gpu_init(struct radeon_device *rdev)
+{
+	u32 gb_addr_config = RREG32(GB_ADDR_CONFIG);
+	u32 mc_shared_chmap, mc_arb_ramcfg;
+	u32 hdp_host_path_cntl;
+	u32 tmp;
+	int i, j;
+
+	switch (rdev->family) {
+	case CHIP_BONAIRE:
+		rdev->config.cik.max_shader_engines = 2;
+		rdev->config.cik.max_tile_pipes = 4;
+		rdev->config.cik.max_cu_per_sh = 7;
+		rdev->config.cik.max_sh_per_se = 1;
+		rdev->config.cik.max_backends_per_se = 2;
+		rdev->config.cik.max_texture_channel_caches = 4;
+		rdev->config.cik.max_gprs = 256;
+		rdev->config.cik.max_gs_threads = 32;
+		rdev->config.cik.max_hw_contexts = 8;
+
+		rdev->config.cik.sc_prim_fifo_size_frontend = 0x20;
+		rdev->config.cik.sc_prim_fifo_size_backend = 0x100;
+		rdev->config.cik.sc_hiz_tile_fifo_size = 0x30;
+		rdev->config.cik.sc_earlyz_tile_fifo_size = 0x130;
+		gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN;
+		break;
+	case CHIP_KAVERI:
+		/* TODO */
+		break;
+	case CHIP_KABINI:
+	default:
+		rdev->config.cik.max_shader_engines = 1;
+		rdev->config.cik.max_tile_pipes = 2;
+		rdev->config.cik.max_cu_per_sh = 2;
+		rdev->config.cik.max_sh_per_se = 1;
+		rdev->config.cik.max_backends_per_se = 1;
+		rdev->config.cik.max_texture_channel_caches = 2;
+		rdev->config.cik.max_gprs = 256;
+		rdev->config.cik.max_gs_threads = 16;
+		rdev->config.cik.max_hw_contexts = 8;
+
+		rdev->config.cik.sc_prim_fifo_size_frontend = 0x20;
+		rdev->config.cik.sc_prim_fifo_size_backend = 0x100;
+		rdev->config.cik.sc_hiz_tile_fifo_size = 0x30;
+		rdev->config.cik.sc_earlyz_tile_fifo_size = 0x130;
+		gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN;
+		break;
+	}
+
+	/* Initialize HDP */
+	for (i = 0, j = 0; i < 32; i++, j += 0x18) {
+		WREG32((0x2c14 + j), 0x00000000);
+		WREG32((0x2c18 + j), 0x00000000);
+		WREG32((0x2c1c + j), 0x00000000);
+		WREG32((0x2c20 + j), 0x00000000);
+		WREG32((0x2c24 + j), 0x00000000);
+	}
+
+	WREG32(GRBM_CNTL, GRBM_READ_TIMEOUT(0xff));
+
+	WREG32(BIF_FB_EN, FB_READ_EN | FB_WRITE_EN);
+
+	mc_shared_chmap = RREG32(MC_SHARED_CHMAP);
+	mc_arb_ramcfg = RREG32(MC_ARB_RAMCFG);
+
+	rdev->config.cik.num_tile_pipes = rdev->config.cik.max_tile_pipes;
+	rdev->config.cik.mem_max_burst_length_bytes = 256;
+	tmp = (mc_arb_ramcfg & NOOFCOLS_MASK) >> NOOFCOLS_SHIFT;
+	rdev->config.cik.mem_row_size_in_kb = (4 * (1 << (8 + tmp))) / 1024;
+	if (rdev->config.cik.mem_row_size_in_kb > 4)
+		rdev->config.cik.mem_row_size_in_kb = 4;
+	/* XXX use MC settings? */
+	rdev->config.cik.shader_engine_tile_size = 32;
+	rdev->config.cik.num_gpus = 1;
+	rdev->config.cik.multi_gpu_tile_size = 64;
+
+	/* fix up row size */
+	gb_addr_config &= ~ROW_SIZE_MASK;
+	switch (rdev->config.cik.mem_row_size_in_kb) {
+	case 1:
+	default:
+		gb_addr_config |= ROW_SIZE(0);
+		break;
+	case 2:
+		gb_addr_config |= ROW_SIZE(1);
+		break;
+	case 4:
+		gb_addr_config |= ROW_SIZE(2);
+		break;
+	}
+
+	/* setup tiling info dword.  gb_addr_config is not adequate since it does
+	 * not have bank info, so create a custom tiling dword.
+	 * bits 3:0   num_pipes
+	 * bits 7:4   num_banks
+	 * bits 11:8  group_size
+	 * bits 15:12 row_size
+	 */
+	rdev->config.cik.tile_config = 0;
+	switch (rdev->config.cik.num_tile_pipes) {
+	case 1:
+		rdev->config.cik.tile_config |= (0 << 0);
+		break;
+	case 2:
+		rdev->config.cik.tile_config |= (1 << 0);
+		break;
+	case 4:
+		rdev->config.cik.tile_config |= (2 << 0);
+		break;
+	case 8:
+	default:
+		/* XXX what about 12? */
+		rdev->config.cik.tile_config |= (3 << 0);
+		break;
+	}
+	if ((mc_arb_ramcfg & NOOFBANK_MASK) >> NOOFBANK_SHIFT)
+		rdev->config.cik.tile_config |= 1 << 4;
+	else
+		rdev->config.cik.tile_config |= 0 << 4;
+	rdev->config.cik.tile_config |=
+		((gb_addr_config & PIPE_INTERLEAVE_SIZE_MASK) >> PIPE_INTERLEAVE_SIZE_SHIFT) << 8;
+	rdev->config.cik.tile_config |=
+		((gb_addr_config & ROW_SIZE_MASK) >> ROW_SIZE_SHIFT) << 12;
+
+	WREG32(GB_ADDR_CONFIG, gb_addr_config);
+	WREG32(HDP_ADDR_CONFIG, gb_addr_config);
+	WREG32(DMIF_ADDR_CALC, gb_addr_config);
+	WREG32(SDMA0_TILING_CONFIG + SDMA0_REGISTER_OFFSET, gb_addr_config & 0x70);
+	WREG32(SDMA0_TILING_CONFIG + SDMA1_REGISTER_OFFSET, gb_addr_config & 0x70);
+	WREG32(UVD_UDEC_ADDR_CONFIG, gb_addr_config);
+	WREG32(UVD_UDEC_DB_ADDR_CONFIG, gb_addr_config);
+	WREG32(UVD_UDEC_DBW_ADDR_CONFIG, gb_addr_config);
+
+	cik_tiling_mode_table_init(rdev);
+
+	cik_setup_rb(rdev, rdev->config.cik.max_shader_engines,
+		     rdev->config.cik.max_sh_per_se,
+		     rdev->config.cik.max_backends_per_se);
+
+	/* set HW defaults for 3D engine */
+	WREG32(CP_MEQ_THRESHOLDS, MEQ1_START(0x30) | MEQ2_START(0x60));
+
+	WREG32(SX_DEBUG_1, 0x20);
+
+	WREG32(TA_CNTL_AUX, 0x00010000);
+
+	tmp = RREG32(SPI_CONFIG_CNTL);
+	tmp |= 0x03000000;
+	WREG32(SPI_CONFIG_CNTL, tmp);
+
+	WREG32(SQ_CONFIG, 1);
+
+	WREG32(DB_DEBUG, 0);
+
+	tmp = RREG32(DB_DEBUG2) & ~0xf00fffff;
+	tmp |= 0x00000400;
+	WREG32(DB_DEBUG2, tmp);
+
+	tmp = RREG32(DB_DEBUG3) & ~0x0002021c;
+	tmp |= 0x00020200;
+	WREG32(DB_DEBUG3, tmp);
+
+	tmp = RREG32(CB_HW_CONTROL) & ~0x00010000;
+	tmp |= 0x00018208;
+	WREG32(CB_HW_CONTROL, tmp);
+
+	WREG32(SPI_CONFIG_CNTL_1, VTX_DONE_DELAY(4));
+
+	WREG32(PA_SC_FIFO_SIZE, (SC_FRONTEND_PRIM_FIFO_SIZE(rdev->config.cik.sc_prim_fifo_size_frontend) |
+				 SC_BACKEND_PRIM_FIFO_SIZE(rdev->config.cik.sc_prim_fifo_size_backend) |
+				 SC_HIZ_TILE_FIFO_SIZE(rdev->config.cik.sc_hiz_tile_fifo_size) |
+				 SC_EARLYZ_TILE_FIFO_SIZE(rdev->config.cik.sc_earlyz_tile_fifo_size)));
+
+	WREG32(VGT_NUM_INSTANCES, 1);
+
+	WREG32(CP_PERFMON_CNTL, 0);
+
+	WREG32(SQ_CONFIG, 0);
+
+	WREG32(PA_SC_FORCE_EOV_MAX_CNTS, (FORCE_EOV_MAX_CLK_CNT(4095) |
+					  FORCE_EOV_MAX_REZ_CNT(255)));
+
+	WREG32(VGT_CACHE_INVALIDATION, CACHE_INVALIDATION(VC_AND_TC) |
+	       AUTO_INVLD_EN(ES_AND_GS_AUTO));
+
+	WREG32(VGT_GS_VERTEX_REUSE, 16);
+	WREG32(PA_SC_LINE_STIPPLE_STATE, 0);
+
+	tmp = RREG32(HDP_MISC_CNTL);
+	tmp |= HDP_FLUSH_INVALIDATE_CACHE;
+	WREG32(HDP_MISC_CNTL, tmp);
+
+	hdp_host_path_cntl = RREG32(HDP_HOST_PATH_CNTL);
+	WREG32(HDP_HOST_PATH_CNTL, hdp_host_path_cntl);
+
+	WREG32(PA_CL_ENHANCE, CLIP_VTX_REORDER_ENA | NUM_CLIP_SEQ(3));
+	WREG32(PA_SC_ENHANCE, ENABLE_PA_SC_OUT_OF_ORDER);
+
+	udelay(50);
+}
+
+/*
+ * GPU scratch registers helpers function.
+ */
+/**
+ * cik_scratch_init - setup driver info for CP scratch regs
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Set up the number and offset of the CP scratch registers.
+ * NOTE: use of CP scratch registers is a legacy inferface and
+ * is not used by default on newer asics (r6xx+).  On newer asics,
+ * memory buffers are used for fences rather than scratch regs.
+ */
+static void cik_scratch_init(struct radeon_device *rdev)
+{
+	int i;
+
+	rdev->scratch.num_reg = 7;
+	rdev->scratch.reg_base = SCRATCH_REG0;
+	for (i = 0; i < rdev->scratch.num_reg; i++) {
+		rdev->scratch.free[i] = true;
+		rdev->scratch.reg[i] = rdev->scratch.reg_base + (i * 4);
+	}
+}
+
+/**
+ * cik_ring_test - basic gfx ring test
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Allocate a scratch register and write to it using the gfx ring (CIK).
+ * Provides a basic gfx ring test to verify that the ring is working.
+ * Used by cik_cp_gfx_resume();
+ * Returns 0 on success, error on failure.
+ */
+int cik_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	uint32_t scratch;
+	uint32_t tmp = 0;
+	unsigned i;
+	int r;
+
+	r = radeon_scratch_get(rdev, &scratch);
+	if (r) {
+		DRM_ERROR("radeon: cp failed to get scratch reg (%d).\n", r);
+		return r;
+	}
+	WREG32(scratch, 0xCAFEDEAD);
+	r = radeon_ring_lock(rdev, ring, 3);
+	if (r) {
+		DRM_ERROR("radeon: cp failed to lock ring %d (%d).\n", ring->idx, r);
+		radeon_scratch_free(rdev, scratch);
+		return r;
+	}
+	radeon_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
+	radeon_ring_write(ring, ((scratch - PACKET3_SET_UCONFIG_REG_START) >> 2));
+	radeon_ring_write(ring, 0xDEADBEEF);
+	radeon_ring_unlock_commit(rdev, ring);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32(scratch);
+		if (tmp == 0xDEADBEEF)
+			break;
+		DRM_UDELAY(1);
+	}
+	if (i < rdev->usec_timeout) {
+		DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i);
+	} else {
+		DRM_ERROR("radeon: ring %d test failed (scratch(0x%04X)=0x%08X)\n",
+			  ring->idx, scratch, tmp);
+		r = -EINVAL;
+	}
+	radeon_scratch_free(rdev, scratch);
+	return r;
+}
+
+/**
+ * cik_fence_gfx_ring_emit - emit a fence on the gfx ring
+ *
+ * @rdev: radeon_device pointer
+ * @fence: radeon fence object
+ *
+ * Emits a fence sequnce number on the gfx ring and flushes
+ * GPU caches.
+ */
+void cik_fence_gfx_ring_emit(struct radeon_device *rdev,
+			     struct radeon_fence *fence)
+{
+	struct radeon_ring *ring = &rdev->ring[fence->ring];
+	u64 addr = rdev->fence_drv[fence->ring].gpu_addr;
+
+	/* EVENT_WRITE_EOP - flush caches, send int */
+	radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4));
+	radeon_ring_write(ring, (EOP_TCL1_ACTION_EN |
+				 EOP_TC_ACTION_EN |
+				 EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) |
+				 EVENT_INDEX(5)));
+	radeon_ring_write(ring, addr & 0xfffffffc);
+	radeon_ring_write(ring, (upper_32_bits(addr) & 0xffff) | DATA_SEL(1) | INT_SEL(2));
+	radeon_ring_write(ring, fence->seq);
+	radeon_ring_write(ring, 0);
+	/* HDP flush */
+	/* We should be using the new WAIT_REG_MEM special op packet here
+	 * but it causes the CP to hang
+	 */
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, 0);
+}
+
+/**
+ * cik_fence_compute_ring_emit - emit a fence on the compute ring
+ *
+ * @rdev: radeon_device pointer
+ * @fence: radeon fence object
+ *
+ * Emits a fence sequnce number on the compute ring and flushes
+ * GPU caches.
+ */
+void cik_fence_compute_ring_emit(struct radeon_device *rdev,
+				 struct radeon_fence *fence)
+{
+	struct radeon_ring *ring = &rdev->ring[fence->ring];
+	u64 addr = rdev->fence_drv[fence->ring].gpu_addr;
+
+	/* RELEASE_MEM - flush caches, send int */
+	radeon_ring_write(ring, PACKET3(PACKET3_RELEASE_MEM, 5));
+	radeon_ring_write(ring, (EOP_TCL1_ACTION_EN |
+				 EOP_TC_ACTION_EN |
+				 EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) |
+				 EVENT_INDEX(5)));
+	radeon_ring_write(ring, DATA_SEL(1) | INT_SEL(2));
+	radeon_ring_write(ring, addr & 0xfffffffc);
+	radeon_ring_write(ring, upper_32_bits(addr));
+	radeon_ring_write(ring, fence->seq);
+	radeon_ring_write(ring, 0);
+	/* HDP flush */
+	/* We should be using the new WAIT_REG_MEM special op packet here
+	 * but it causes the CP to hang
+	 */
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, 0);
+}
+
+void cik_semaphore_ring_emit(struct radeon_device *rdev,
+			     struct radeon_ring *ring,
+			     struct radeon_semaphore *semaphore,
+			     bool emit_wait)
+{
+	uint64_t addr = semaphore->gpu_addr;
+	unsigned sel = emit_wait ? PACKET3_SEM_SEL_WAIT : PACKET3_SEM_SEL_SIGNAL;
+
+	radeon_ring_write(ring, PACKET3(PACKET3_MEM_SEMAPHORE, 1));
+	radeon_ring_write(ring, addr & 0xffffffff);
+	radeon_ring_write(ring, (upper_32_bits(addr) & 0xffff) | sel);
+}
+
+/*
+ * IB stuff
+ */
+/**
+ * cik_ring_ib_execute - emit an IB (Indirect Buffer) on the gfx ring
+ *
+ * @rdev: radeon_device pointer
+ * @ib: radeon indirect buffer object
+ *
+ * Emits an DE (drawing engine) or CE (constant engine) IB
+ * on the gfx ring.  IBs are usually generated by userspace
+ * acceleration drivers and submitted to the kernel for
+ * sheduling on the ring.  This function schedules the IB
+ * on the gfx ring for execution by the GPU.
+ */
+void cik_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
+{
+	struct radeon_ring *ring = &rdev->ring[ib->ring];
+	u32 header, control = INDIRECT_BUFFER_VALID;
+
+	if (ib->is_const_ib) {
+		/* set switch buffer packet before const IB */
+		radeon_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
+		radeon_ring_write(ring, 0);
+
+		header = PACKET3(PACKET3_INDIRECT_BUFFER_CONST, 2);
+	} else {
+		u32 next_rptr;
+		if (ring->rptr_save_reg) {
+			next_rptr = ring->wptr + 3 + 4;
+			radeon_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1));
+			radeon_ring_write(ring, ((ring->rptr_save_reg -
+						  PACKET3_SET_UCONFIG_REG_START) >> 2));
+			radeon_ring_write(ring, next_rptr);
+		} else if (rdev->wb.enabled) {
+			next_rptr = ring->wptr + 5 + 4;
+			radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+			radeon_ring_write(ring, WRITE_DATA_DST_SEL(1));
+			radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc);
+			radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff);
+			radeon_ring_write(ring, next_rptr);
+		}
+
+		header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
+	}
+
+	control |= ib->length_dw |
+		(ib->vm ? (ib->vm->id << 24) : 0);
+
+	radeon_ring_write(ring, header);
+	radeon_ring_write(ring,
+#ifdef __BIG_ENDIAN
+			  (2 << 0) |
+#endif
+			  (ib->gpu_addr & 0xFFFFFFFC));
+	radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFFFF);
+	radeon_ring_write(ring, control);
+}
+
+/**
+ * cik_ib_test - basic gfx ring IB test
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Allocate an IB and execute it on the gfx ring (CIK).
+ * Provides a basic gfx ring test to verify that IBs are working.
+ * Returns 0 on success, error on failure.
+ */
+int cik_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	struct radeon_ib ib;
+	uint32_t scratch;
+	uint32_t tmp = 0;
+	unsigned i;
+	int r;
+
+	r = radeon_scratch_get(rdev, &scratch);
+	if (r) {
+		DRM_ERROR("radeon: failed to get scratch reg (%d).\n", r);
+		return r;
+	}
+	WREG32(scratch, 0xCAFEDEAD);
+	r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256);
+	if (r) {
+		DRM_ERROR("radeon: failed to get ib (%d).\n", r);
+		return r;
+	}
+	ib.ptr[0] = PACKET3(PACKET3_SET_UCONFIG_REG, 1);
+	ib.ptr[1] = ((scratch - PACKET3_SET_UCONFIG_REG_START) >> 2);
+	ib.ptr[2] = 0xDEADBEEF;
+	ib.length_dw = 3;
+	r = radeon_ib_schedule(rdev, &ib, NULL);
+	if (r) {
+		radeon_scratch_free(rdev, scratch);
+		radeon_ib_free(rdev, &ib);
+		DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
+		return r;
+	}
+	r = radeon_fence_wait(ib.fence, false);
+	if (r) {
+		DRM_ERROR("radeon: fence wait failed (%d).\n", r);
+		return r;
+	}
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32(scratch);
+		if (tmp == 0xDEADBEEF)
+			break;
+		DRM_UDELAY(1);
+	}
+	if (i < rdev->usec_timeout) {
+		DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib.fence->ring, i);
+	} else {
+		DRM_ERROR("radeon: ib test failed (scratch(0x%04X)=0x%08X)\n",
+			  scratch, tmp);
+		r = -EINVAL;
+	}
+	radeon_scratch_free(rdev, scratch);
+	radeon_ib_free(rdev, &ib);
+	return r;
+}
+
+/*
+ * CP.
+ * On CIK, gfx and compute now have independant command processors.
+ *
+ * GFX
+ * Gfx consists of a single ring and can process both gfx jobs and
+ * compute jobs.  The gfx CP consists of three microengines (ME):
+ * PFP - Pre-Fetch Parser
+ * ME - Micro Engine
+ * CE - Constant Engine
+ * The PFP and ME make up what is considered the Drawing Engine (DE).
+ * The CE is an asynchronous engine used for updating buffer desciptors
+ * used by the DE so that they can be loaded into cache in parallel
+ * while the DE is processing state update packets.
+ *
+ * Compute
+ * The compute CP consists of two microengines (ME):
+ * MEC1 - Compute MicroEngine 1
+ * MEC2 - Compute MicroEngine 2
+ * Each MEC supports 4 compute pipes and each pipe supports 8 queues.
+ * The queues are exposed to userspace and are programmed directly
+ * by the compute runtime.
+ */
+/**
+ * cik_cp_gfx_enable - enable/disable the gfx CP MEs
+ *
+ * @rdev: radeon_device pointer
+ * @enable: enable or disable the MEs
+ *
+ * Halts or unhalts the gfx MEs.
+ */
+static void cik_cp_gfx_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32(CP_ME_CNTL, 0);
+	else {
+		WREG32(CP_ME_CNTL, (CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT));
+		rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
+	}
+	udelay(50);
+}
+
+/**
+ * cik_cp_gfx_load_microcode - load the gfx CP ME ucode
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Loads the gfx PFP, ME, and CE ucode.
+ * Returns 0 for success, -EINVAL if the ucode is not available.
+ */
+static int cik_cp_gfx_load_microcode(struct radeon_device *rdev)
+{
+	const __be32 *fw_data;
+	int i;
+
+	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw)
+		return -EINVAL;
+
+	cik_cp_gfx_enable(rdev, false);
+
+	/* PFP */
+	fw_data = (const __be32 *)rdev->pfp_fw->data;
+	WREG32(CP_PFP_UCODE_ADDR, 0);
+	for (i = 0; i < CIK_PFP_UCODE_SIZE; i++)
+		WREG32(CP_PFP_UCODE_DATA, be32_to_cpup(fw_data++));
+	WREG32(CP_PFP_UCODE_ADDR, 0);
+
+	/* CE */
+	fw_data = (const __be32 *)rdev->ce_fw->data;
+	WREG32(CP_CE_UCODE_ADDR, 0);
+	for (i = 0; i < CIK_CE_UCODE_SIZE; i++)
+		WREG32(CP_CE_UCODE_DATA, be32_to_cpup(fw_data++));
+	WREG32(CP_CE_UCODE_ADDR, 0);
+
+	/* ME */
+	fw_data = (const __be32 *)rdev->me_fw->data;
+	WREG32(CP_ME_RAM_WADDR, 0);
+	for (i = 0; i < CIK_ME_UCODE_SIZE; i++)
+		WREG32(CP_ME_RAM_DATA, be32_to_cpup(fw_data++));
+	WREG32(CP_ME_RAM_WADDR, 0);
+
+	WREG32(CP_PFP_UCODE_ADDR, 0);
+	WREG32(CP_CE_UCODE_ADDR, 0);
+	WREG32(CP_ME_RAM_WADDR, 0);
+	WREG32(CP_ME_RAM_RADDR, 0);
+	return 0;
+}
+
+/**
+ * cik_cp_gfx_start - start the gfx ring
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Enables the ring and loads the clear state context and other
+ * packets required to init the ring.
+ * Returns 0 for success, error for failure.
+ */
+static int cik_cp_gfx_start(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	int r, i;
+
+	/* init the CP */
+	WREG32(CP_MAX_CONTEXT, rdev->config.cik.max_hw_contexts - 1);
+	WREG32(CP_ENDIAN_SWAP, 0);
+	WREG32(CP_DEVICE_ID, 1);
+
+	cik_cp_gfx_enable(rdev, true);
+
+	r = radeon_ring_lock(rdev, ring, cik_default_size + 17);
+	if (r) {
+		DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r);
+		return r;
+	}
+
+	/* init the CE partitions.  CE only used for gfx on CIK */
+	radeon_ring_write(ring, PACKET3(PACKET3_SET_BASE, 2));
+	radeon_ring_write(ring, PACKET3_BASE_INDEX(CE_PARTITION_BASE));
+	radeon_ring_write(ring, 0xc000);
+	radeon_ring_write(ring, 0xc000);
+
+	/* setup clear context state */
+	radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	radeon_ring_write(ring, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
+
+	radeon_ring_write(ring, PACKET3(PACKET3_CONTEXT_CONTROL, 1));
+	radeon_ring_write(ring, 0x80000000);
+	radeon_ring_write(ring, 0x80000000);
+
+	for (i = 0; i < cik_default_size; i++)
+		radeon_ring_write(ring, cik_default_state[i]);
+
+	radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	radeon_ring_write(ring, PACKET3_PREAMBLE_END_CLEAR_STATE);
+
+	/* set clear context state */
+	radeon_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0));
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(ring, 0x00000316);
+	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
+	radeon_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */
+
+	radeon_ring_unlock_commit(rdev, ring);
+
+	return 0;
+}
+
+/**
+ * cik_cp_gfx_fini - stop the gfx ring
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Stop the gfx ring and tear down the driver ring
+ * info.
+ */
+static void cik_cp_gfx_fini(struct radeon_device *rdev)
+{
+	cik_cp_gfx_enable(rdev, false);
+	radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+}
+
+/**
+ * cik_cp_gfx_resume - setup the gfx ring buffer registers
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Program the location and size of the gfx ring buffer
+ * and test it to make sure it's working.
+ * Returns 0 for success, error for failure.
+ */
+static int cik_cp_gfx_resume(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring;
+	u32 tmp;
+	u32 rb_bufsz;
+	u64 rb_addr;
+	int r;
+
+	WREG32(CP_SEM_WAIT_TIMER, 0x0);
+	WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0);
+
+	/* Set the write pointer delay */
+	WREG32(CP_RB_WPTR_DELAY, 0);
+
+	/* set the RB to use vmid 0 */
+	WREG32(CP_RB_VMID, 0);
+
+	WREG32(SCRATCH_ADDR, ((rdev->wb.gpu_addr + RADEON_WB_SCRATCH_OFFSET) >> 8) & 0xFFFFFFFF);
+
+	/* ring 0 - compute and gfx */
+	/* Set ring buffer size */
+	ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	rb_bufsz = drm_order(ring->ring_size / 8);
+	tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
+#ifdef __BIG_ENDIAN
+	tmp |= BUF_SWAP_32BIT;
+#endif
+	WREG32(CP_RB0_CNTL, tmp);
+
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32(CP_RB0_CNTL, tmp | RB_RPTR_WR_ENA);
+	ring->wptr = 0;
+	WREG32(CP_RB0_WPTR, ring->wptr);
+
+	/* set the wb address wether it's enabled or not */
+	WREG32(CP_RB0_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC);
+	WREG32(CP_RB0_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFF);
+
+	/* scratch register shadowing is no longer supported */
+	WREG32(SCRATCH_UMSK, 0);
+
+	if (!rdev->wb.enabled)
+		tmp |= RB_NO_UPDATE;
+
+	mdelay(1);
+	WREG32(CP_RB0_CNTL, tmp);
+
+	rb_addr = ring->gpu_addr >> 8;
+	WREG32(CP_RB0_BASE, rb_addr);
+	WREG32(CP_RB0_BASE_HI, upper_32_bits(rb_addr));
+
+	ring->rptr = RREG32(CP_RB0_RPTR);
+
+	/* start the ring */
+	cik_cp_gfx_start(rdev);
+	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = true;
+	r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	if (r) {
+		rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
+		return r;
+	}
+	return 0;
+}
+
+u32 cik_compute_ring_get_rptr(struct radeon_device *rdev,
+			      struct radeon_ring *ring)
+{
+	u32 rptr;
+
+
+
+	if (rdev->wb.enabled) {
+		rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]);
+	} else {
+		cik_srbm_select(rdev, ring->me, ring->pipe, ring->queue, 0);
+		rptr = RREG32(CP_HQD_PQ_RPTR);
+		cik_srbm_select(rdev, 0, 0, 0, 0);
+	}
+	rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+
+	return rptr;
+}
+
+u32 cik_compute_ring_get_wptr(struct radeon_device *rdev,
+			      struct radeon_ring *ring)
+{
+	u32 wptr;
+
+	if (rdev->wb.enabled) {
+		wptr = le32_to_cpu(rdev->wb.wb[ring->wptr_offs/4]);
+	} else {
+		cik_srbm_select(rdev, ring->me, ring->pipe, ring->queue, 0);
+		wptr = RREG32(CP_HQD_PQ_WPTR);
+		cik_srbm_select(rdev, 0, 0, 0, 0);
+	}
+	wptr = (wptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+
+	return wptr;
+}
+
+void cik_compute_ring_set_wptr(struct radeon_device *rdev,
+			       struct radeon_ring *ring)
+{
+	u32 wptr = (ring->wptr << ring->ptr_reg_shift) & ring->ptr_reg_mask;
+
+	rdev->wb.wb[ring->wptr_offs/4] = cpu_to_le32(wptr);
+	WDOORBELL32(ring->doorbell_offset, wptr);
+}
+
+/**
+ * cik_cp_compute_enable - enable/disable the compute CP MEs
+ *
+ * @rdev: radeon_device pointer
+ * @enable: enable or disable the MEs
+ *
+ * Halts or unhalts the compute MEs.
+ */
+static void cik_cp_compute_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32(CP_MEC_CNTL, 0);
+	else
+		WREG32(CP_MEC_CNTL, (MEC_ME1_HALT | MEC_ME2_HALT));
+	udelay(50);
+}
+
+/**
+ * cik_cp_compute_load_microcode - load the compute CP ME ucode
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Loads the compute MEC1&2 ucode.
+ * Returns 0 for success, -EINVAL if the ucode is not available.
+ */
+static int cik_cp_compute_load_microcode(struct radeon_device *rdev)
+{
+	const __be32 *fw_data;
+	int i;
+
+	if (!rdev->mec_fw)
+		return -EINVAL;
+
+	cik_cp_compute_enable(rdev, false);
+
+	/* MEC1 */
+	fw_data = (const __be32 *)rdev->mec_fw->data;
+	WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
+	for (i = 0; i < CIK_MEC_UCODE_SIZE; i++)
+		WREG32(CP_MEC_ME1_UCODE_DATA, be32_to_cpup(fw_data++));
+	WREG32(CP_MEC_ME1_UCODE_ADDR, 0);
+
+	if (rdev->family == CHIP_KAVERI) {
+		/* MEC2 */
+		fw_data = (const __be32 *)rdev->mec_fw->data;
+		WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+		for (i = 0; i < CIK_MEC_UCODE_SIZE; i++)
+			WREG32(CP_MEC_ME2_UCODE_DATA, be32_to_cpup(fw_data++));
+		WREG32(CP_MEC_ME2_UCODE_ADDR, 0);
+	}
+
+	return 0;
+}
+
+/**
+ * cik_cp_compute_start - start the compute queues
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Enable the compute queues.
+ * Returns 0 for success, error for failure.
+ */
+static int cik_cp_compute_start(struct radeon_device *rdev)
+{
+	cik_cp_compute_enable(rdev, true);
+
+	return 0;
+}
+
+/**
+ * cik_cp_compute_fini - stop the compute queues
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Stop the compute queues and tear down the driver queue
+ * info.
+ */
+static void cik_cp_compute_fini(struct radeon_device *rdev)
+{
+	int i, idx, r;
+
+	cik_cp_compute_enable(rdev, false);
+
+	for (i = 0; i < 2; i++) {
+		if (i == 0)
+			idx = CAYMAN_RING_TYPE_CP1_INDEX;
+		else
+			idx = CAYMAN_RING_TYPE_CP2_INDEX;
+
+		if (rdev->ring[idx].mqd_obj) {
+			r = radeon_bo_reserve(rdev->ring[idx].mqd_obj, false);
+			if (unlikely(r != 0))
+				dev_warn(rdev->dev, "(%d) reserve MQD bo failed\n", r);
+
+			radeon_bo_unpin(rdev->ring[idx].mqd_obj);
+			radeon_bo_unreserve(rdev->ring[idx].mqd_obj);
+
+			radeon_bo_unref(&rdev->ring[idx].mqd_obj);
+			rdev->ring[idx].mqd_obj = NULL;
+		}
+	}
+}
+
+static void cik_mec_fini(struct radeon_device *rdev)
+{
+	int r;
+
+	if (rdev->mec.hpd_eop_obj) {
+		r = radeon_bo_reserve(rdev->mec.hpd_eop_obj, false);
+		if (unlikely(r != 0))
+			dev_warn(rdev->dev, "(%d) reserve HPD EOP bo failed\n", r);
+		radeon_bo_unpin(rdev->mec.hpd_eop_obj);
+		radeon_bo_unreserve(rdev->mec.hpd_eop_obj);
+
+		radeon_bo_unref(&rdev->mec.hpd_eop_obj);
+		rdev->mec.hpd_eop_obj = NULL;
+	}
+}
+
+#define MEC_HPD_SIZE 2048
+
+static int cik_mec_init(struct radeon_device *rdev)
+{
+	int r;
+	u32 *hpd;
+
+	/*
+	 * KV:    2 MEC, 4 Pipes/MEC, 8 Queues/Pipe - 64 Queues total
+	 * CI/KB: 1 MEC, 4 Pipes/MEC, 8 Queues/Pipe - 32 Queues total
+	 */
+	if (rdev->family == CHIP_KAVERI)
+		rdev->mec.num_mec = 2;
+	else
+		rdev->mec.num_mec = 1;
+	rdev->mec.num_pipe = 4;
+	rdev->mec.num_queue = rdev->mec.num_mec * rdev->mec.num_pipe * 8;
+
+	if (rdev->mec.hpd_eop_obj == NULL) {
+		r = radeon_bo_create(rdev,
+				     rdev->mec.num_mec *rdev->mec.num_pipe * MEC_HPD_SIZE * 2,
+				     PAGE_SIZE, true,
+				     RADEON_GEM_DOMAIN_GTT, NULL,
+				     &rdev->mec.hpd_eop_obj);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) create HDP EOP bo failed\n", r);
+			return r;
+		}
+	}
+
+	r = radeon_bo_reserve(rdev->mec.hpd_eop_obj, false);
+	if (unlikely(r != 0)) {
+		cik_mec_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_pin(rdev->mec.hpd_eop_obj, RADEON_GEM_DOMAIN_GTT,
+			  &rdev->mec.hpd_eop_gpu_addr);
+	if (r) {
+		dev_warn(rdev->dev, "(%d) pin HDP EOP bo failed\n", r);
+		cik_mec_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_kmap(rdev->mec.hpd_eop_obj, (void **)&hpd);
+	if (r) {
+		dev_warn(rdev->dev, "(%d) map HDP EOP bo failed\n", r);
+		cik_mec_fini(rdev);
+		return r;
+	}
+
+	/* clear memory.  Not sure if this is required or not */
+	memset(hpd, 0, rdev->mec.num_mec *rdev->mec.num_pipe * MEC_HPD_SIZE * 2);
+
+	radeon_bo_kunmap(rdev->mec.hpd_eop_obj);
+	radeon_bo_unreserve(rdev->mec.hpd_eop_obj);
+
+	return 0;
+}
+
+struct hqd_registers
+{
+	u32 cp_mqd_base_addr;
+	u32 cp_mqd_base_addr_hi;
+	u32 cp_hqd_active;
+	u32 cp_hqd_vmid;
+	u32 cp_hqd_persistent_state;
+	u32 cp_hqd_pipe_priority;
+	u32 cp_hqd_queue_priority;
+	u32 cp_hqd_quantum;
+	u32 cp_hqd_pq_base;
+	u32 cp_hqd_pq_base_hi;
+	u32 cp_hqd_pq_rptr;
+	u32 cp_hqd_pq_rptr_report_addr;
+	u32 cp_hqd_pq_rptr_report_addr_hi;
+	u32 cp_hqd_pq_wptr_poll_addr;
+	u32 cp_hqd_pq_wptr_poll_addr_hi;
+	u32 cp_hqd_pq_doorbell_control;
+	u32 cp_hqd_pq_wptr;
+	u32 cp_hqd_pq_control;
+	u32 cp_hqd_ib_base_addr;
+	u32 cp_hqd_ib_base_addr_hi;
+	u32 cp_hqd_ib_rptr;
+	u32 cp_hqd_ib_control;
+	u32 cp_hqd_iq_timer;
+	u32 cp_hqd_iq_rptr;
+	u32 cp_hqd_dequeue_request;
+	u32 cp_hqd_dma_offload;
+	u32 cp_hqd_sema_cmd;
+	u32 cp_hqd_msg_type;
+	u32 cp_hqd_atomic0_preop_lo;
+	u32 cp_hqd_atomic0_preop_hi;
+	u32 cp_hqd_atomic1_preop_lo;
+	u32 cp_hqd_atomic1_preop_hi;
+	u32 cp_hqd_hq_scheduler0;
+	u32 cp_hqd_hq_scheduler1;
+	u32 cp_mqd_control;
+};
+
+struct bonaire_mqd
+{
+	u32 header;
+	u32 dispatch_initiator;
+	u32 dimensions[3];
+	u32 start_idx[3];
+	u32 num_threads[3];
+	u32 pipeline_stat_enable;
+	u32 perf_counter_enable;
+	u32 pgm[2];
+	u32 tba[2];
+	u32 tma[2];
+	u32 pgm_rsrc[2];
+	u32 vmid;
+	u32 resource_limits;
+	u32 static_thread_mgmt01[2];
+	u32 tmp_ring_size;
+	u32 static_thread_mgmt23[2];
+	u32 restart[3];
+	u32 thread_trace_enable;
+	u32 reserved1;
+	u32 user_data[16];
+	u32 vgtcs_invoke_count[2];
+	struct hqd_registers queue_state;
+	u32 dequeue_cntr;
+	u32 interrupt_queue[64];
+};
+
+/**
+ * cik_cp_compute_resume - setup the compute queue registers
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Program the compute queues and test them to make sure they
+ * are working.
+ * Returns 0 for success, error for failure.
+ */
+static int cik_cp_compute_resume(struct radeon_device *rdev)
+{
+	int r, i, idx;
+	u32 tmp;
+	bool use_doorbell = true;
+	u64 hqd_gpu_addr;
+	u64 mqd_gpu_addr;
+	u64 eop_gpu_addr;
+	u64 wb_gpu_addr;
+	u32 *buf;
+	struct bonaire_mqd *mqd;
+
+	r = cik_cp_compute_start(rdev);
+	if (r)
+		return r;
+
+	/* fix up chicken bits */
+	tmp = RREG32(CP_CPF_DEBUG);
+	tmp |= (1 << 23);
+	WREG32(CP_CPF_DEBUG, tmp);
+
+	/* init the pipes */
+	for (i = 0; i < (rdev->mec.num_pipe * rdev->mec.num_mec); i++) {
+		int me = (i < 4) ? 1 : 2;
+		int pipe = (i < 4) ? i : (i - 4);
+
+		eop_gpu_addr = rdev->mec.hpd_eop_gpu_addr + (i * MEC_HPD_SIZE * 2);
+
+		cik_srbm_select(rdev, me, pipe, 0, 0);
+
+		/* write the EOP addr */
+		WREG32(CP_HPD_EOP_BASE_ADDR, eop_gpu_addr >> 8);
+		WREG32(CP_HPD_EOP_BASE_ADDR_HI, upper_32_bits(eop_gpu_addr) >> 8);
+
+		/* set the VMID assigned */
+		WREG32(CP_HPD_EOP_VMID, 0);
+
+		/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
+		tmp = RREG32(CP_HPD_EOP_CONTROL);
+		tmp &= ~EOP_SIZE_MASK;
+		tmp |= drm_order(MEC_HPD_SIZE / 8);
+		WREG32(CP_HPD_EOP_CONTROL, tmp);
+	}
+	cik_srbm_select(rdev, 0, 0, 0, 0);
+
+	/* init the queues.  Just two for now. */
+	for (i = 0; i < 2; i++) {
+		if (i == 0)
+			idx = CAYMAN_RING_TYPE_CP1_INDEX;
+		else
+			idx = CAYMAN_RING_TYPE_CP2_INDEX;
+
+		if (rdev->ring[idx].mqd_obj == NULL) {
+			r = radeon_bo_create(rdev,
+					     sizeof(struct bonaire_mqd),
+					     PAGE_SIZE, true,
+					     RADEON_GEM_DOMAIN_GTT, NULL,
+					     &rdev->ring[idx].mqd_obj);
+			if (r) {
+				dev_warn(rdev->dev, "(%d) create MQD bo failed\n", r);
+				return r;
+			}
+		}
+
+		r = radeon_bo_reserve(rdev->ring[idx].mqd_obj, false);
+		if (unlikely(r != 0)) {
+			cik_cp_compute_fini(rdev);
+			return r;
+		}
+		r = radeon_bo_pin(rdev->ring[idx].mqd_obj, RADEON_GEM_DOMAIN_GTT,
+				  &mqd_gpu_addr);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) pin MQD bo failed\n", r);
+			cik_cp_compute_fini(rdev);
+			return r;
+		}
+		r = radeon_bo_kmap(rdev->ring[idx].mqd_obj, (void **)&buf);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) map MQD bo failed\n", r);
+			cik_cp_compute_fini(rdev);
+			return r;
+		}
+
+		/* doorbell offset */
+		rdev->ring[idx].doorbell_offset =
+			(rdev->ring[idx].doorbell_page_num * PAGE_SIZE) + 0;
+
+		/* init the mqd struct */
+		memset(buf, 0, sizeof(struct bonaire_mqd));
+
+		mqd = (struct bonaire_mqd *)buf;
+		mqd->header = 0xC0310800;
+		mqd->static_thread_mgmt01[0] = 0xffffffff;
+		mqd->static_thread_mgmt01[1] = 0xffffffff;
+		mqd->static_thread_mgmt23[0] = 0xffffffff;
+		mqd->static_thread_mgmt23[1] = 0xffffffff;
+
+		cik_srbm_select(rdev, rdev->ring[idx].me,
+				rdev->ring[idx].pipe,
+				rdev->ring[idx].queue, 0);
+
+		/* disable wptr polling */
+		tmp = RREG32(CP_PQ_WPTR_POLL_CNTL);
+		tmp &= ~WPTR_POLL_EN;
+		WREG32(CP_PQ_WPTR_POLL_CNTL, tmp);
+
+		/* enable doorbell? */
+		mqd->queue_state.cp_hqd_pq_doorbell_control =
+			RREG32(CP_HQD_PQ_DOORBELL_CONTROL);
+		if (use_doorbell)
+			mqd->queue_state.cp_hqd_pq_doorbell_control |= DOORBELL_EN;
+		else
+			mqd->queue_state.cp_hqd_pq_doorbell_control &= ~DOORBELL_EN;
+		WREG32(CP_HQD_PQ_DOORBELL_CONTROL,
+		       mqd->queue_state.cp_hqd_pq_doorbell_control);
+
+		/* disable the queue if it's active */
+		mqd->queue_state.cp_hqd_dequeue_request = 0;
+		mqd->queue_state.cp_hqd_pq_rptr = 0;
+		mqd->queue_state.cp_hqd_pq_wptr= 0;
+		if (RREG32(CP_HQD_ACTIVE) & 1) {
+			WREG32(CP_HQD_DEQUEUE_REQUEST, 1);
+			for (i = 0; i < rdev->usec_timeout; i++) {
+				if (!(RREG32(CP_HQD_ACTIVE) & 1))
+					break;
+				udelay(1);
+			}
+			WREG32(CP_HQD_DEQUEUE_REQUEST, mqd->queue_state.cp_hqd_dequeue_request);
+			WREG32(CP_HQD_PQ_RPTR, mqd->queue_state.cp_hqd_pq_rptr);
+			WREG32(CP_HQD_PQ_WPTR, mqd->queue_state.cp_hqd_pq_wptr);
+		}
+
+		/* set the pointer to the MQD */
+		mqd->queue_state.cp_mqd_base_addr = mqd_gpu_addr & 0xfffffffc;
+		mqd->queue_state.cp_mqd_base_addr_hi = upper_32_bits(mqd_gpu_addr);
+		WREG32(CP_MQD_BASE_ADDR, mqd->queue_state.cp_mqd_base_addr);
+		WREG32(CP_MQD_BASE_ADDR_HI, mqd->queue_state.cp_mqd_base_addr_hi);
+		/* set MQD vmid to 0 */
+		mqd->queue_state.cp_mqd_control = RREG32(CP_MQD_CONTROL);
+		mqd->queue_state.cp_mqd_control &= ~MQD_VMID_MASK;
+		WREG32(CP_MQD_CONTROL, mqd->queue_state.cp_mqd_control);
+
+		/* set the pointer to the HQD, this is similar CP_RB0_BASE/_HI */
+		hqd_gpu_addr = rdev->ring[idx].gpu_addr >> 8;
+		mqd->queue_state.cp_hqd_pq_base = hqd_gpu_addr;
+		mqd->queue_state.cp_hqd_pq_base_hi = upper_32_bits(hqd_gpu_addr);
+		WREG32(CP_HQD_PQ_BASE, mqd->queue_state.cp_hqd_pq_base);
+		WREG32(CP_HQD_PQ_BASE_HI, mqd->queue_state.cp_hqd_pq_base_hi);
+
+		/* set up the HQD, this is similar to CP_RB0_CNTL */
+		mqd->queue_state.cp_hqd_pq_control = RREG32(CP_HQD_PQ_CONTROL);
+		mqd->queue_state.cp_hqd_pq_control &=
+			~(QUEUE_SIZE_MASK | RPTR_BLOCK_SIZE_MASK);
+
+		mqd->queue_state.cp_hqd_pq_control |=
+			drm_order(rdev->ring[idx].ring_size / 8);
+		mqd->queue_state.cp_hqd_pq_control |=
+			(drm_order(RADEON_GPU_PAGE_SIZE/8) << 8);
+#ifdef __BIG_ENDIAN
+		mqd->queue_state.cp_hqd_pq_control |= BUF_SWAP_32BIT;
+#endif
+		mqd->queue_state.cp_hqd_pq_control &=
+			~(UNORD_DISPATCH | ROQ_PQ_IB_FLIP | PQ_VOLATILE);
+		mqd->queue_state.cp_hqd_pq_control |=
+			PRIV_STATE | KMD_QUEUE; /* assuming kernel queue control */
+		WREG32(CP_HQD_PQ_CONTROL, mqd->queue_state.cp_hqd_pq_control);
+
+		/* only used if CP_PQ_WPTR_POLL_CNTL.WPTR_POLL_EN=1 */
+		if (i == 0)
+			wb_gpu_addr = rdev->wb.gpu_addr + CIK_WB_CP1_WPTR_OFFSET;
+		else
+			wb_gpu_addr = rdev->wb.gpu_addr + CIK_WB_CP2_WPTR_OFFSET;
+		mqd->queue_state.cp_hqd_pq_wptr_poll_addr = wb_gpu_addr & 0xfffffffc;
+		mqd->queue_state.cp_hqd_pq_wptr_poll_addr_hi = upper_32_bits(wb_gpu_addr) & 0xffff;
+		WREG32(CP_HQD_PQ_WPTR_POLL_ADDR, mqd->queue_state.cp_hqd_pq_wptr_poll_addr);
+		WREG32(CP_HQD_PQ_WPTR_POLL_ADDR_HI,
+		       mqd->queue_state.cp_hqd_pq_wptr_poll_addr_hi);
+
+		/* set the wb address wether it's enabled or not */
+		if (i == 0)
+			wb_gpu_addr = rdev->wb.gpu_addr + RADEON_WB_CP1_RPTR_OFFSET;
+		else
+			wb_gpu_addr = rdev->wb.gpu_addr + RADEON_WB_CP2_RPTR_OFFSET;
+		mqd->queue_state.cp_hqd_pq_rptr_report_addr = wb_gpu_addr & 0xfffffffc;
+		mqd->queue_state.cp_hqd_pq_rptr_report_addr_hi =
+			upper_32_bits(wb_gpu_addr) & 0xffff;
+		WREG32(CP_HQD_PQ_RPTR_REPORT_ADDR,
+		       mqd->queue_state.cp_hqd_pq_rptr_report_addr);
+		WREG32(CP_HQD_PQ_RPTR_REPORT_ADDR_HI,
+		       mqd->queue_state.cp_hqd_pq_rptr_report_addr_hi);
+
+		/* enable the doorbell if requested */
+		if (use_doorbell) {
+			mqd->queue_state.cp_hqd_pq_doorbell_control =
+				RREG32(CP_HQD_PQ_DOORBELL_CONTROL);
+			mqd->queue_state.cp_hqd_pq_doorbell_control &= ~DOORBELL_OFFSET_MASK;
+			mqd->queue_state.cp_hqd_pq_doorbell_control |=
+				DOORBELL_OFFSET(rdev->ring[idx].doorbell_offset / 4);
+			mqd->queue_state.cp_hqd_pq_doorbell_control |= DOORBELL_EN;
+			mqd->queue_state.cp_hqd_pq_doorbell_control &=
+				~(DOORBELL_SOURCE | DOORBELL_HIT);
+
+		} else {
+			mqd->queue_state.cp_hqd_pq_doorbell_control = 0;
+		}
+		WREG32(CP_HQD_PQ_DOORBELL_CONTROL,
+		       mqd->queue_state.cp_hqd_pq_doorbell_control);
+
+		/* read and write pointers, similar to CP_RB0_WPTR/_RPTR */
+		rdev->ring[idx].wptr = 0;
+		mqd->queue_state.cp_hqd_pq_wptr = rdev->ring[idx].wptr;
+		WREG32(CP_HQD_PQ_WPTR, mqd->queue_state.cp_hqd_pq_wptr);
+		rdev->ring[idx].rptr = RREG32(CP_HQD_PQ_RPTR);
+		mqd->queue_state.cp_hqd_pq_rptr = rdev->ring[idx].rptr;
+
+		/* set the vmid for the queue */
+		mqd->queue_state.cp_hqd_vmid = 0;
+		WREG32(CP_HQD_VMID, mqd->queue_state.cp_hqd_vmid);
+
+		/* activate the queue */
+		mqd->queue_state.cp_hqd_active = 1;
+		WREG32(CP_HQD_ACTIVE, mqd->queue_state.cp_hqd_active);
+
+		cik_srbm_select(rdev, 0, 0, 0, 0);
+
+		radeon_bo_kunmap(rdev->ring[idx].mqd_obj);
+		radeon_bo_unreserve(rdev->ring[idx].mqd_obj);
+
+		rdev->ring[idx].ready = true;
+		r = radeon_ring_test(rdev, idx, &rdev->ring[idx]);
+		if (r)
+			rdev->ring[idx].ready = false;
+	}
+
+	return 0;
+}
+
+static void cik_cp_enable(struct radeon_device *rdev, bool enable)
+{
+	cik_cp_gfx_enable(rdev, enable);
+	cik_cp_compute_enable(rdev, enable);
+}
+
+static int cik_cp_load_microcode(struct radeon_device *rdev)
+{
+	int r;
+
+	r = cik_cp_gfx_load_microcode(rdev);
+	if (r)
+		return r;
+	r = cik_cp_compute_load_microcode(rdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+static void cik_cp_fini(struct radeon_device *rdev)
+{
+	cik_cp_gfx_fini(rdev);
+	cik_cp_compute_fini(rdev);
+}
+
+static int cik_cp_resume(struct radeon_device *rdev)
+{
+	int r;
+
+	/* Reset all cp blocks */
+	WREG32(GRBM_SOFT_RESET, SOFT_RESET_CP);
+	RREG32(GRBM_SOFT_RESET);
+	mdelay(15);
+	WREG32(GRBM_SOFT_RESET, 0);
+	RREG32(GRBM_SOFT_RESET);
+
+	r = cik_cp_load_microcode(rdev);
+	if (r)
+		return r;
+
+	r = cik_cp_gfx_resume(rdev);
+	if (r)
+		return r;
+	r = cik_cp_compute_resume(rdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+/*
+ * sDMA - System DMA
+ * Starting with CIK, the GPU has new asynchronous
+ * DMA engines.  These engines are used for compute
+ * and gfx.  There are two DMA engines (SDMA0, SDMA1)
+ * and each one supports 1 ring buffer used for gfx
+ * and 2 queues used for compute.
+ *
+ * The programming model is very similar to the CP
+ * (ring buffer, IBs, etc.), but sDMA has it's own
+ * packet format that is different from the PM4 format
+ * used by the CP. sDMA supports copying data, writing
+ * embedded data, solid fills, and a number of other
+ * things.  It also has support for tiling/detiling of
+ * buffers.
+ */
+/**
+ * cik_sdma_ring_ib_execute - Schedule an IB on the DMA engine
+ *
+ * @rdev: radeon_device pointer
+ * @ib: IB object to schedule
+ *
+ * Schedule an IB in the DMA ring (CIK).
+ */
+void cik_sdma_ring_ib_execute(struct radeon_device *rdev,
+			      struct radeon_ib *ib)
+{
+	struct radeon_ring *ring = &rdev->ring[ib->ring];
+	u32 extra_bits = (ib->vm ? ib->vm->id : 0) & 0xf;
+
+	if (rdev->wb.enabled) {
+		u32 next_rptr = ring->wptr + 5;
+		while ((next_rptr & 7) != 4)
+			next_rptr++;
+		next_rptr += 4;
+		radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0));
+		radeon_ring_write(ring, ring->next_rptr_gpu_addr & 0xfffffffc);
+		radeon_ring_write(ring, upper_32_bits(ring->next_rptr_gpu_addr) & 0xffffffff);
+		radeon_ring_write(ring, 1); /* number of DWs to follow */
+		radeon_ring_write(ring, next_rptr);
+	}
+
+	/* IB packet must end on a 8 DW boundary */
+	while ((ring->wptr & 7) != 4)
+		radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0));
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_INDIRECT_BUFFER, 0, extra_bits));
+	radeon_ring_write(ring, ib->gpu_addr & 0xffffffe0); /* base must be 32 byte aligned */
+	radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xffffffff);
+	radeon_ring_write(ring, ib->length_dw);
+
+}
+
+/**
+ * cik_sdma_fence_ring_emit - emit a fence on the DMA ring
+ *
+ * @rdev: radeon_device pointer
+ * @fence: radeon fence object
+ *
+ * Add a DMA fence packet to the ring to write
+ * the fence seq number and DMA trap packet to generate
+ * an interrupt if needed (CIK).
+ */
+void cik_sdma_fence_ring_emit(struct radeon_device *rdev,
+			      struct radeon_fence *fence)
+{
+	struct radeon_ring *ring = &rdev->ring[fence->ring];
+	u64 addr = rdev->fence_drv[fence->ring].gpu_addr;
+	u32 extra_bits = (SDMA_POLL_REG_MEM_EXTRA_OP(1) |
+			  SDMA_POLL_REG_MEM_EXTRA_FUNC(3)); /* == */
+	u32 ref_and_mask;
+
+	if (fence->ring == R600_RING_TYPE_DMA_INDEX)
+		ref_and_mask = SDMA0;
+	else
+		ref_and_mask = SDMA1;
+
+	/* write the fence */
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_FENCE, 0, 0));
+	radeon_ring_write(ring, addr & 0xffffffff);
+	radeon_ring_write(ring, upper_32_bits(addr) & 0xffffffff);
+	radeon_ring_write(ring, fence->seq);
+	/* generate an interrupt */
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_TRAP, 0, 0));
+	/* flush HDP */
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_POLL_REG_MEM, 0, extra_bits));
+	radeon_ring_write(ring, GPU_HDP_FLUSH_DONE);
+	radeon_ring_write(ring, GPU_HDP_FLUSH_REQ);
+	radeon_ring_write(ring, ref_and_mask); /* REFERENCE */
+	radeon_ring_write(ring, ref_and_mask); /* MASK */
+	radeon_ring_write(ring, (4 << 16) | 10); /* RETRY_COUNT, POLL_INTERVAL */
+}
+
+/**
+ * cik_sdma_semaphore_ring_emit - emit a semaphore on the dma ring
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ * @semaphore: radeon semaphore object
+ * @emit_wait: wait or signal semaphore
+ *
+ * Add a DMA semaphore packet to the ring wait on or signal
+ * other rings (CIK).
+ */
+void cik_sdma_semaphore_ring_emit(struct radeon_device *rdev,
+				  struct radeon_ring *ring,
+				  struct radeon_semaphore *semaphore,
+				  bool emit_wait)
+{
+	u64 addr = semaphore->gpu_addr;
+	u32 extra_bits = emit_wait ? 0 : SDMA_SEMAPHORE_EXTRA_S;
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SEMAPHORE, 0, extra_bits));
+	radeon_ring_write(ring, addr & 0xfffffff8);
+	radeon_ring_write(ring, upper_32_bits(addr) & 0xffffffff);
+}
+
+/**
+ * cik_sdma_gfx_stop - stop the gfx async dma engines
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Stop the gfx async dma ring buffers (CIK).
+ */
+static void cik_sdma_gfx_stop(struct radeon_device *rdev)
+{
+	u32 rb_cntl, reg_offset;
+	int i;
+
+	radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size);
+
+	for (i = 0; i < 2; i++) {
+		if (i == 0)
+			reg_offset = SDMA0_REGISTER_OFFSET;
+		else
+			reg_offset = SDMA1_REGISTER_OFFSET;
+		rb_cntl = RREG32(SDMA0_GFX_RB_CNTL + reg_offset);
+		rb_cntl &= ~SDMA_RB_ENABLE;
+		WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl);
+		WREG32(SDMA0_GFX_IB_CNTL + reg_offset, 0);
+	}
+}
+
+/**
+ * cik_sdma_rlc_stop - stop the compute async dma engines
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Stop the compute async dma queues (CIK).
+ */
+static void cik_sdma_rlc_stop(struct radeon_device *rdev)
+{
+	/* XXX todo */
+}
+
+/**
+ * cik_sdma_enable - stop the async dma engines
+ *
+ * @rdev: radeon_device pointer
+ * @enable: enable/disable the DMA MEs.
+ *
+ * Halt or unhalt the async dma engines (CIK).
+ */
+static void cik_sdma_enable(struct radeon_device *rdev, bool enable)
+{
+	u32 me_cntl, reg_offset;
+	int i;
+
+	for (i = 0; i < 2; i++) {
+		if (i == 0)
+			reg_offset = SDMA0_REGISTER_OFFSET;
+		else
+			reg_offset = SDMA1_REGISTER_OFFSET;
+		me_cntl = RREG32(SDMA0_ME_CNTL + reg_offset);
+		if (enable)
+			me_cntl &= ~SDMA_HALT;
+		else
+			me_cntl |= SDMA_HALT;
+		WREG32(SDMA0_ME_CNTL + reg_offset, me_cntl);
+	}
+}
+
+/**
+ * cik_sdma_gfx_resume - setup and start the async dma engines
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Set up the gfx DMA ring buffers and enable them (CIK).
+ * Returns 0 for success, error for failure.
+ */
+static int cik_sdma_gfx_resume(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring;
+	u32 rb_cntl, ib_cntl;
+	u32 rb_bufsz;
+	u32 reg_offset, wb_offset;
+	int i, r;
+
+	for (i = 0; i < 2; i++) {
+		if (i == 0) {
+			ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX];
+			reg_offset = SDMA0_REGISTER_OFFSET;
+			wb_offset = R600_WB_DMA_RPTR_OFFSET;
+		} else {
+			ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX];
+			reg_offset = SDMA1_REGISTER_OFFSET;
+			wb_offset = CAYMAN_WB_DMA1_RPTR_OFFSET;
+		}
+
+		WREG32(SDMA0_SEM_INCOMPLETE_TIMER_CNTL + reg_offset, 0);
+		WREG32(SDMA0_SEM_WAIT_FAIL_TIMER_CNTL + reg_offset, 0);
+
+		/* Set ring buffer size in dwords */
+		rb_bufsz = drm_order(ring->ring_size / 4);
+		rb_cntl = rb_bufsz << 1;
+#ifdef __BIG_ENDIAN
+		rb_cntl |= SDMA_RB_SWAP_ENABLE | SDMA_RPTR_WRITEBACK_SWAP_ENABLE;
+#endif
+		WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl);
+
+		/* Initialize the ring buffer's read and write pointers */
+		WREG32(SDMA0_GFX_RB_RPTR + reg_offset, 0);
+		WREG32(SDMA0_GFX_RB_WPTR + reg_offset, 0);
+
+		/* set the wb address whether it's enabled or not */
+		WREG32(SDMA0_GFX_RB_RPTR_ADDR_HI + reg_offset,
+		       upper_32_bits(rdev->wb.gpu_addr + wb_offset) & 0xFFFFFFFF);
+		WREG32(SDMA0_GFX_RB_RPTR_ADDR_LO + reg_offset,
+		       ((rdev->wb.gpu_addr + wb_offset) & 0xFFFFFFFC));
+
+		if (rdev->wb.enabled)
+			rb_cntl |= SDMA_RPTR_WRITEBACK_ENABLE;
+
+		WREG32(SDMA0_GFX_RB_BASE + reg_offset, ring->gpu_addr >> 8);
+		WREG32(SDMA0_GFX_RB_BASE_HI + reg_offset, ring->gpu_addr >> 40);
+
+		ring->wptr = 0;
+		WREG32(SDMA0_GFX_RB_WPTR + reg_offset, ring->wptr << 2);
+
+		ring->rptr = RREG32(SDMA0_GFX_RB_RPTR + reg_offset) >> 2;
+
+		/* enable DMA RB */
+		WREG32(SDMA0_GFX_RB_CNTL + reg_offset, rb_cntl | SDMA_RB_ENABLE);
+
+		ib_cntl = SDMA_IB_ENABLE;
+#ifdef __BIG_ENDIAN
+		ib_cntl |= SDMA_IB_SWAP_ENABLE;
+#endif
+		/* enable DMA IBs */
+		WREG32(SDMA0_GFX_IB_CNTL + reg_offset, ib_cntl);
+
+		ring->ready = true;
+
+		r = radeon_ring_test(rdev, ring->idx, ring);
+		if (r) {
+			ring->ready = false;
+			return r;
+		}
+	}
+
+	radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size);
+
+	return 0;
+}
+
+/**
+ * cik_sdma_rlc_resume - setup and start the async dma engines
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Set up the compute DMA queues and enable them (CIK).
+ * Returns 0 for success, error for failure.
+ */
+static int cik_sdma_rlc_resume(struct radeon_device *rdev)
+{
+	/* XXX todo */
+	return 0;
+}
+
+/**
+ * cik_sdma_load_microcode - load the sDMA ME ucode
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Loads the sDMA0/1 ucode.
+ * Returns 0 for success, -EINVAL if the ucode is not available.
+ */
+static int cik_sdma_load_microcode(struct radeon_device *rdev)
+{
+	const __be32 *fw_data;
+	int i;
+
+	if (!rdev->sdma_fw)
+		return -EINVAL;
+
+	/* stop the gfx rings and rlc compute queues */
+	cik_sdma_gfx_stop(rdev);
+	cik_sdma_rlc_stop(rdev);
+
+	/* halt the MEs */
+	cik_sdma_enable(rdev, false);
+
+	/* sdma0 */
+	fw_data = (const __be32 *)rdev->sdma_fw->data;
+	WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0);
+	for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++)
+		WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, be32_to_cpup(fw_data++));
+	WREG32(SDMA0_UCODE_DATA + SDMA0_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+
+	/* sdma1 */
+	fw_data = (const __be32 *)rdev->sdma_fw->data;
+	WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0);
+	for (i = 0; i < CIK_SDMA_UCODE_SIZE; i++)
+		WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, be32_to_cpup(fw_data++));
+	WREG32(SDMA0_UCODE_DATA + SDMA1_REGISTER_OFFSET, CIK_SDMA_UCODE_VERSION);
+
+	WREG32(SDMA0_UCODE_ADDR + SDMA0_REGISTER_OFFSET, 0);
+	WREG32(SDMA0_UCODE_ADDR + SDMA1_REGISTER_OFFSET, 0);
+	return 0;
+}
+
+/**
+ * cik_sdma_resume - setup and start the async dma engines
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Set up the DMA engines and enable them (CIK).
+ * Returns 0 for success, error for failure.
+ */
+static int cik_sdma_resume(struct radeon_device *rdev)
+{
+	int r;
+
+	/* Reset dma */
+	WREG32(SRBM_SOFT_RESET, SOFT_RESET_SDMA | SOFT_RESET_SDMA1);
+	RREG32(SRBM_SOFT_RESET);
+	udelay(50);
+	WREG32(SRBM_SOFT_RESET, 0);
+	RREG32(SRBM_SOFT_RESET);
+
+	r = cik_sdma_load_microcode(rdev);
+	if (r)
+		return r;
+
+	/* unhalt the MEs */
+	cik_sdma_enable(rdev, true);
+
+	/* start the gfx rings and rlc compute queues */
+	r = cik_sdma_gfx_resume(rdev);
+	if (r)
+		return r;
+	r = cik_sdma_rlc_resume(rdev);
+	if (r)
+		return r;
+
+	return 0;
+}
+
+/**
+ * cik_sdma_fini - tear down the async dma engines
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Stop the async dma engines and free the rings (CIK).
+ */
+static void cik_sdma_fini(struct radeon_device *rdev)
+{
+	/* stop the gfx rings and rlc compute queues */
+	cik_sdma_gfx_stop(rdev);
+	cik_sdma_rlc_stop(rdev);
+	/* halt the MEs */
+	cik_sdma_enable(rdev, false);
+	radeon_ring_fini(rdev, &rdev->ring[R600_RING_TYPE_DMA_INDEX]);
+	radeon_ring_fini(rdev, &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX]);
+	/* XXX - compute dma queue tear down */
+}
+
+/**
+ * cik_copy_dma - copy pages using the DMA engine
+ *
+ * @rdev: radeon_device pointer
+ * @src_offset: src GPU address
+ * @dst_offset: dst GPU address
+ * @num_gpu_pages: number of GPU pages to xfer
+ * @fence: radeon fence object
+ *
+ * Copy GPU paging using the DMA engine (CIK).
+ * Used by the radeon ttm implementation to move pages if
+ * registered as the asic copy callback.
+ */
+int cik_copy_dma(struct radeon_device *rdev,
+		 uint64_t src_offset, uint64_t dst_offset,
+		 unsigned num_gpu_pages,
+		 struct radeon_fence **fence)
+{
+	struct radeon_semaphore *sem = NULL;
+	int ring_index = rdev->asic->copy.dma_ring_index;
+	struct radeon_ring *ring = &rdev->ring[ring_index];
+	u32 size_in_bytes, cur_size_in_bytes;
+	int i, num_loops;
+	int r = 0;
+
+	r = radeon_semaphore_create(rdev, &sem);
+	if (r) {
+		DRM_ERROR("radeon: moving bo (%d).\n", r);
+		return r;
+	}
+
+	size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
+	num_loops = DIV_ROUND_UP(size_in_bytes, 0x1fffff);
+	r = radeon_ring_lock(rdev, ring, num_loops * 7 + 14);
+	if (r) {
+		DRM_ERROR("radeon: moving bo (%d).\n", r);
+		radeon_semaphore_free(rdev, &sem, NULL);
+		return r;
+	}
+
+	if (radeon_fence_need_sync(*fence, ring->idx)) {
+		radeon_semaphore_sync_rings(rdev, sem, (*fence)->ring,
+					    ring->idx);
+		radeon_fence_note_sync(*fence, ring->idx);
+	} else {
+		radeon_semaphore_free(rdev, &sem, NULL);
+	}
+
+	for (i = 0; i < num_loops; i++) {
+		cur_size_in_bytes = size_in_bytes;
+		if (cur_size_in_bytes > 0x1fffff)
+			cur_size_in_bytes = 0x1fffff;
+		size_in_bytes -= cur_size_in_bytes;
+		radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_COPY, SDMA_COPY_SUB_OPCODE_LINEAR, 0));
+		radeon_ring_write(ring, cur_size_in_bytes);
+		radeon_ring_write(ring, 0); /* src/dst endian swap */
+		radeon_ring_write(ring, src_offset & 0xffffffff);
+		radeon_ring_write(ring, upper_32_bits(src_offset) & 0xffffffff);
+		radeon_ring_write(ring, dst_offset & 0xfffffffc);
+		radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xffffffff);
+		src_offset += cur_size_in_bytes;
+		dst_offset += cur_size_in_bytes;
+	}
+
+	r = radeon_fence_emit(rdev, fence, ring->idx);
+	if (r) {
+		radeon_ring_unlock_undo(rdev, ring);
+		return r;
+	}
+
+	radeon_ring_unlock_commit(rdev, ring);
+	radeon_semaphore_free(rdev, &sem, *fence);
+
+	return r;
+}
+
+/**
+ * cik_sdma_ring_test - simple async dma engine test
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Test the DMA engine by writing using it to write an
+ * value to memory. (CIK).
+ * Returns 0 for success, error for failure.
+ */
+int cik_sdma_ring_test(struct radeon_device *rdev,
+		       struct radeon_ring *ring)
+{
+	unsigned i;
+	int r;
+	void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
+	u32 tmp;
+
+	if (!ptr) {
+		DRM_ERROR("invalid vram scratch pointer\n");
+		return -EINVAL;
+	}
+
+	tmp = 0xCAFEDEAD;
+	writel(tmp, ptr);
+
+	r = radeon_ring_lock(rdev, ring, 4);
+	if (r) {
+		DRM_ERROR("radeon: dma failed to lock ring %d (%d).\n", ring->idx, r);
+		return r;
+	}
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0));
+	radeon_ring_write(ring, rdev->vram_scratch.gpu_addr & 0xfffffffc);
+	radeon_ring_write(ring, upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xffffffff);
+	radeon_ring_write(ring, 1); /* number of DWs to follow */
+	radeon_ring_write(ring, 0xDEADBEEF);
+	radeon_ring_unlock_commit(rdev, ring);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = readl(ptr);
+		if (tmp == 0xDEADBEEF)
+			break;
+		DRM_UDELAY(1);
+	}
+
+	if (i < rdev->usec_timeout) {
+		DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i);
+	} else {
+		DRM_ERROR("radeon: ring %d test failed (0x%08X)\n",
+			  ring->idx, tmp);
+		r = -EINVAL;
+	}
+	return r;
+}
+
+/**
+ * cik_sdma_ib_test - test an IB on the DMA engine
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Test a simple IB in the DMA ring (CIK).
+ * Returns 0 on success, error on failure.
+ */
+int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	struct radeon_ib ib;
+	unsigned i;
+	int r;
+	void __iomem *ptr = (void *)rdev->vram_scratch.ptr;
+	u32 tmp = 0;
+
+	if (!ptr) {
+		DRM_ERROR("invalid vram scratch pointer\n");
+		return -EINVAL;
+	}
+
+	tmp = 0xCAFEDEAD;
+	writel(tmp, ptr);
+
+	r = radeon_ib_get(rdev, ring->idx, &ib, NULL, 256);
+	if (r) {
+		DRM_ERROR("radeon: failed to get ib (%d).\n", r);
+		return r;
+	}
+
+	ib.ptr[0] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
+	ib.ptr[1] = rdev->vram_scratch.gpu_addr & 0xfffffffc;
+	ib.ptr[2] = upper_32_bits(rdev->vram_scratch.gpu_addr) & 0xffffffff;
+	ib.ptr[3] = 1;
+	ib.ptr[4] = 0xDEADBEEF;
+	ib.length_dw = 5;
+
+	r = radeon_ib_schedule(rdev, &ib, NULL);
+	if (r) {
+		radeon_ib_free(rdev, &ib);
+		DRM_ERROR("radeon: failed to schedule ib (%d).\n", r);
+		return r;
+	}
+	r = radeon_fence_wait(ib.fence, false);
+	if (r) {
+		DRM_ERROR("radeon: fence wait failed (%d).\n", r);
+		return r;
+	}
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = readl(ptr);
+		if (tmp == 0xDEADBEEF)
+			break;
+		DRM_UDELAY(1);
+	}
+	if (i < rdev->usec_timeout) {
+		DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib.fence->ring, i);
+	} else {
+		DRM_ERROR("radeon: ib test failed (0x%08X)\n", tmp);
+		r = -EINVAL;
+	}
+	radeon_ib_free(rdev, &ib);
+	return r;
+}
+
+
+static void cik_print_gpu_status_regs(struct radeon_device *rdev)
+{
+	dev_info(rdev->dev, "  GRBM_STATUS=0x%08X\n",
+		RREG32(GRBM_STATUS));
+	dev_info(rdev->dev, "  GRBM_STATUS2=0x%08X\n",
+		RREG32(GRBM_STATUS2));
+	dev_info(rdev->dev, "  GRBM_STATUS_SE0=0x%08X\n",
+		RREG32(GRBM_STATUS_SE0));
+	dev_info(rdev->dev, "  GRBM_STATUS_SE1=0x%08X\n",
+		RREG32(GRBM_STATUS_SE1));
+	dev_info(rdev->dev, "  GRBM_STATUS_SE2=0x%08X\n",
+		RREG32(GRBM_STATUS_SE2));
+	dev_info(rdev->dev, "  GRBM_STATUS_SE3=0x%08X\n",
+		RREG32(GRBM_STATUS_SE3));
+	dev_info(rdev->dev, "  SRBM_STATUS=0x%08X\n",
+		RREG32(SRBM_STATUS));
+	dev_info(rdev->dev, "  SRBM_STATUS2=0x%08X\n",
+		RREG32(SRBM_STATUS2));
+	dev_info(rdev->dev, "  SDMA0_STATUS_REG   = 0x%08X\n",
+		RREG32(SDMA0_STATUS_REG + SDMA0_REGISTER_OFFSET));
+	dev_info(rdev->dev, "  SDMA1_STATUS_REG   = 0x%08X\n",
+		 RREG32(SDMA0_STATUS_REG + SDMA1_REGISTER_OFFSET));
+	dev_info(rdev->dev, "  CP_STAT = 0x%08x\n", RREG32(CP_STAT));
+	dev_info(rdev->dev, "  CP_STALLED_STAT1 = 0x%08x\n",
+		 RREG32(CP_STALLED_STAT1));
+	dev_info(rdev->dev, "  CP_STALLED_STAT2 = 0x%08x\n",
+		 RREG32(CP_STALLED_STAT2));
+	dev_info(rdev->dev, "  CP_STALLED_STAT3 = 0x%08x\n",
+		 RREG32(CP_STALLED_STAT3));
+	dev_info(rdev->dev, "  CP_CPF_BUSY_STAT = 0x%08x\n",
+		 RREG32(CP_CPF_BUSY_STAT));
+	dev_info(rdev->dev, "  CP_CPF_STALLED_STAT1 = 0x%08x\n",
+		 RREG32(CP_CPF_STALLED_STAT1));
+	dev_info(rdev->dev, "  CP_CPF_STATUS = 0x%08x\n", RREG32(CP_CPF_STATUS));
+	dev_info(rdev->dev, "  CP_CPC_BUSY_STAT = 0x%08x\n", RREG32(CP_CPC_BUSY_STAT));
+	dev_info(rdev->dev, "  CP_CPC_STALLED_STAT1 = 0x%08x\n",
+		 RREG32(CP_CPC_STALLED_STAT1));
+	dev_info(rdev->dev, "  CP_CPC_STATUS = 0x%08x\n", RREG32(CP_CPC_STATUS));
+}
+
+/**
+ * cik_gpu_check_soft_reset - check which blocks are busy
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Check which blocks are busy and return the relevant reset
+ * mask to be used by cik_gpu_soft_reset().
+ * Returns a mask of the blocks to be reset.
+ */
+static u32 cik_gpu_check_soft_reset(struct radeon_device *rdev)
+{
+	u32 reset_mask = 0;
+	u32 tmp;
+
+	/* GRBM_STATUS */
+	tmp = RREG32(GRBM_STATUS);
+	if (tmp & (PA_BUSY | SC_BUSY |
+		   BCI_BUSY | SX_BUSY |
+		   TA_BUSY | VGT_BUSY |
+		   DB_BUSY | CB_BUSY |
+		   GDS_BUSY | SPI_BUSY |
+		   IA_BUSY | IA_BUSY_NO_DMA))
+		reset_mask |= RADEON_RESET_GFX;
+
+	if (tmp & (CP_BUSY | CP_COHERENCY_BUSY))
+		reset_mask |= RADEON_RESET_CP;
+
+	/* GRBM_STATUS2 */
+	tmp = RREG32(GRBM_STATUS2);
+	if (tmp & RLC_BUSY)
+		reset_mask |= RADEON_RESET_RLC;
+
+	/* SDMA0_STATUS_REG */
+	tmp = RREG32(SDMA0_STATUS_REG + SDMA0_REGISTER_OFFSET);
+	if (!(tmp & SDMA_IDLE))
+		reset_mask |= RADEON_RESET_DMA;
+
+	/* SDMA1_STATUS_REG */
+	tmp = RREG32(SDMA0_STATUS_REG + SDMA1_REGISTER_OFFSET);
+	if (!(tmp & SDMA_IDLE))
+		reset_mask |= RADEON_RESET_DMA1;
+
+	/* SRBM_STATUS2 */
+	tmp = RREG32(SRBM_STATUS2);
+	if (tmp & SDMA_BUSY)
+		reset_mask |= RADEON_RESET_DMA;
+
+	if (tmp & SDMA1_BUSY)
+		reset_mask |= RADEON_RESET_DMA1;
+
+	/* SRBM_STATUS */
+	tmp = RREG32(SRBM_STATUS);
+
+	if (tmp & IH_BUSY)
+		reset_mask |= RADEON_RESET_IH;
+
+	if (tmp & SEM_BUSY)
+		reset_mask |= RADEON_RESET_SEM;
+
+	if (tmp & GRBM_RQ_PENDING)
+		reset_mask |= RADEON_RESET_GRBM;
+
+	if (tmp & VMC_BUSY)
+		reset_mask |= RADEON_RESET_VMC;
+
+	if (tmp & (MCB_BUSY | MCB_NON_DISPLAY_BUSY |
+		   MCC_BUSY | MCD_BUSY))
+		reset_mask |= RADEON_RESET_MC;
+
+	if (evergreen_is_display_hung(rdev))
+		reset_mask |= RADEON_RESET_DISPLAY;
+
+	/* Skip MC reset as it's mostly likely not hung, just busy */
+	if (reset_mask & RADEON_RESET_MC) {
+		DRM_DEBUG("MC busy: 0x%08X, clearing.\n", reset_mask);
+		reset_mask &= ~RADEON_RESET_MC;
+	}
+
+	return reset_mask;
+}
+
+/**
+ * cik_gpu_soft_reset - soft reset GPU
+ *
+ * @rdev: radeon_device pointer
+ * @reset_mask: mask of which blocks to reset
+ *
+ * Soft reset the blocks specified in @reset_mask.
+ */
+static void cik_gpu_soft_reset(struct radeon_device *rdev, u32 reset_mask)
+{
+	struct evergreen_mc_save save;
+	u32 grbm_soft_reset = 0, srbm_soft_reset = 0;
+	u32 tmp;
+
+	if (reset_mask == 0)
+		return;
+
+	dev_info(rdev->dev, "GPU softreset: 0x%08X\n", reset_mask);
+
+	cik_print_gpu_status_regs(rdev);
+	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x%08X\n",
+		 RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR));
+	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
+		 RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS));
+
+	/* stop the rlc */
+	cik_rlc_stop(rdev);
+
+	/* Disable GFX parsing/prefetching */
+	WREG32(CP_ME_CNTL, CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT);
+
+	/* Disable MEC parsing/prefetching */
+	WREG32(CP_MEC_CNTL, MEC_ME1_HALT | MEC_ME2_HALT);
+
+	if (reset_mask & RADEON_RESET_DMA) {
+		/* sdma0 */
+		tmp = RREG32(SDMA0_ME_CNTL + SDMA0_REGISTER_OFFSET);
+		tmp |= SDMA_HALT;
+		WREG32(SDMA0_ME_CNTL + SDMA0_REGISTER_OFFSET, tmp);
+	}
+	if (reset_mask & RADEON_RESET_DMA1) {
+		/* sdma1 */
+		tmp = RREG32(SDMA0_ME_CNTL + SDMA1_REGISTER_OFFSET);
+		tmp |= SDMA_HALT;
+		WREG32(SDMA0_ME_CNTL + SDMA1_REGISTER_OFFSET, tmp);
+	}
+
+	evergreen_mc_stop(rdev, &save);
+	if (evergreen_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+
+	if (reset_mask & (RADEON_RESET_GFX | RADEON_RESET_COMPUTE | RADEON_RESET_CP))
+		grbm_soft_reset = SOFT_RESET_CP | SOFT_RESET_GFX;
+
+	if (reset_mask & RADEON_RESET_CP) {
+		grbm_soft_reset |= SOFT_RESET_CP;
+
+		srbm_soft_reset |= SOFT_RESET_GRBM;
+	}
+
+	if (reset_mask & RADEON_RESET_DMA)
+		srbm_soft_reset |= SOFT_RESET_SDMA;
+
+	if (reset_mask & RADEON_RESET_DMA1)
+		srbm_soft_reset |= SOFT_RESET_SDMA1;
+
+	if (reset_mask & RADEON_RESET_DISPLAY)
+		srbm_soft_reset |= SOFT_RESET_DC;
+
+	if (reset_mask & RADEON_RESET_RLC)
+		grbm_soft_reset |= SOFT_RESET_RLC;
+
+	if (reset_mask & RADEON_RESET_SEM)
+		srbm_soft_reset |= SOFT_RESET_SEM;
+
+	if (reset_mask & RADEON_RESET_IH)
+		srbm_soft_reset |= SOFT_RESET_IH;
+
+	if (reset_mask & RADEON_RESET_GRBM)
+		srbm_soft_reset |= SOFT_RESET_GRBM;
+
+	if (reset_mask & RADEON_RESET_VMC)
+		srbm_soft_reset |= SOFT_RESET_VMC;
+
+	if (!(rdev->flags & RADEON_IS_IGP)) {
+		if (reset_mask & RADEON_RESET_MC)
+			srbm_soft_reset |= SOFT_RESET_MC;
+	}
+
+	if (grbm_soft_reset) {
+		tmp = RREG32(GRBM_SOFT_RESET);
+		tmp |= grbm_soft_reset;
+		dev_info(rdev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(GRBM_SOFT_RESET, tmp);
+		tmp = RREG32(GRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~grbm_soft_reset;
+		WREG32(GRBM_SOFT_RESET, tmp);
+		tmp = RREG32(GRBM_SOFT_RESET);
+	}
+
+	if (srbm_soft_reset) {
+		tmp = RREG32(SRBM_SOFT_RESET);
+		tmp |= srbm_soft_reset;
+		dev_info(rdev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(SRBM_SOFT_RESET, tmp);
+		tmp = RREG32(SRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~srbm_soft_reset;
+		WREG32(SRBM_SOFT_RESET, tmp);
+		tmp = RREG32(SRBM_SOFT_RESET);
+	}
+
+	/* Wait a little for things to settle down */
+	udelay(50);
+
+	evergreen_mc_resume(rdev, &save);
+	udelay(50);
+
+	cik_print_gpu_status_regs(rdev);
+}
+
+/**
+ * cik_asic_reset - soft reset GPU
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Look up which blocks are hung and attempt
+ * to reset them.
+ * Returns 0 for success.
+ */
+int cik_asic_reset(struct radeon_device *rdev)
+{
+	u32 reset_mask;
+
+	reset_mask = cik_gpu_check_soft_reset(rdev);
+
+	if (reset_mask)
+		r600_set_bios_scratch_engine_hung(rdev, true);
+
+	cik_gpu_soft_reset(rdev, reset_mask);
+
+	reset_mask = cik_gpu_check_soft_reset(rdev);
+
+	if (!reset_mask)
+		r600_set_bios_scratch_engine_hung(rdev, false);
+
+	return 0;
+}
+
+/**
+ * cik_gfx_is_lockup - check if the 3D engine is locked up
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Check if the 3D engine is locked up (CIK).
+ * Returns true if the engine is locked, false if not.
+ */
+bool cik_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	u32 reset_mask = cik_gpu_check_soft_reset(rdev);
+
+	if (!(reset_mask & (RADEON_RESET_GFX |
+			    RADEON_RESET_COMPUTE |
+			    RADEON_RESET_CP))) {
+		radeon_ring_lockup_update(ring);
+		return false;
+	}
+	/* force CP activities */
+	radeon_ring_force_activity(rdev, ring);
+	return radeon_ring_test_lockup(rdev, ring);
+}
+
+/**
+ * cik_sdma_is_lockup - Check if the DMA engine is locked up
+ *
+ * @rdev: radeon_device pointer
+ * @ring: radeon_ring structure holding ring information
+ *
+ * Check if the async DMA engine is locked up (CIK).
+ * Returns true if the engine appears to be locked up, false if not.
+ */
+bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+{
+	u32 reset_mask = cik_gpu_check_soft_reset(rdev);
+	u32 mask;
+
+	if (ring->idx == R600_RING_TYPE_DMA_INDEX)
+		mask = RADEON_RESET_DMA;
+	else
+		mask = RADEON_RESET_DMA1;
+
+	if (!(reset_mask & mask)) {
+		radeon_ring_lockup_update(ring);
+		return false;
+	}
+	/* force ring activities */
+	radeon_ring_force_activity(rdev, ring);
+	return radeon_ring_test_lockup(rdev, ring);
+}
+
+/* MC */
+/**
+ * cik_mc_program - program the GPU memory controller
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Set the location of vram, gart, and AGP in the GPU's
+ * physical address space (CIK).
+ */
+static void cik_mc_program(struct radeon_device *rdev)
+{
+	struct evergreen_mc_save save;
+	u32 tmp;
+	int i, j;
+
+	/* Initialize HDP */
+	for (i = 0, j = 0; i < 32; i++, j += 0x18) {
+		WREG32((0x2c14 + j), 0x00000000);
+		WREG32((0x2c18 + j), 0x00000000);
+		WREG32((0x2c1c + j), 0x00000000);
+		WREG32((0x2c20 + j), 0x00000000);
+		WREG32((0x2c24 + j), 0x00000000);
+	}
+	WREG32(HDP_REG_COHERENCY_FLUSH_CNTL, 0);
+
+	evergreen_mc_stop(rdev, &save);
+	if (radeon_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+	/* Lockout access through VGA aperture*/
+	WREG32(VGA_HDP_CONTROL, VGA_MEMORY_DISABLE);
+	/* Update configuration */
+	WREG32(MC_VM_SYSTEM_APERTURE_LOW_ADDR,
+	       rdev->mc.vram_start >> 12);
+	WREG32(MC_VM_SYSTEM_APERTURE_HIGH_ADDR,
+	       rdev->mc.vram_end >> 12);
+	WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR,
+	       rdev->vram_scratch.gpu_addr >> 12);
+	tmp = ((rdev->mc.vram_end >> 24) & 0xFFFF) << 16;
+	tmp |= ((rdev->mc.vram_start >> 24) & 0xFFFF);
+	WREG32(MC_VM_FB_LOCATION, tmp);
+	/* XXX double check these! */
+	WREG32(HDP_NONSURFACE_BASE, (rdev->mc.vram_start >> 8));
+	WREG32(HDP_NONSURFACE_INFO, (2 << 7) | (1 << 30));
+	WREG32(HDP_NONSURFACE_SIZE, 0x3FFFFFFF);
+	WREG32(MC_VM_AGP_BASE, 0);
+	WREG32(MC_VM_AGP_TOP, 0x0FFFFFFF);
+	WREG32(MC_VM_AGP_BOT, 0x0FFFFFFF);
+	if (radeon_mc_wait_for_idle(rdev)) {
+		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
+	}
+	evergreen_mc_resume(rdev, &save);
+	/* we need to own VRAM, so turn off the VGA renderer here
+	 * to stop it overwriting our objects */
+	rv515_vga_render_disable(rdev);
+}
+
+/**
+ * cik_mc_init - initialize the memory controller driver params
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Look up the amount of vram, vram width, and decide how to place
+ * vram and gart within the GPU's physical address space (CIK).
+ * Returns 0 for success.
+ */
+static int cik_mc_init(struct radeon_device *rdev)
+{
+	u32 tmp;
+	int chansize, numchan;
+
+	/* Get VRAM informations */
+	rdev->mc.vram_is_ddr = true;
+	tmp = RREG32(MC_ARB_RAMCFG);
+	if (tmp & CHANSIZE_MASK) {
+		chansize = 64;
+	} else {
+		chansize = 32;
+	}
+	tmp = RREG32(MC_SHARED_CHMAP);
+	switch ((tmp & NOOFCHAN_MASK) >> NOOFCHAN_SHIFT) {
+	case 0:
+	default:
+		numchan = 1;
+		break;
+	case 1:
+		numchan = 2;
+		break;
+	case 2:
+		numchan = 4;
+		break;
+	case 3:
+		numchan = 8;
+		break;
+	case 4:
+		numchan = 3;
+		break;
+	case 5:
+		numchan = 6;
+		break;
+	case 6:
+		numchan = 10;
+		break;
+	case 7:
+		numchan = 12;
+		break;
+	case 8:
+		numchan = 16;
+		break;
+	}
+	rdev->mc.vram_width = numchan * chansize;
+	/* Could aper size report 0 ? */
+	rdev->mc.aper_base = pci_resource_start(rdev->pdev, 0);
+	rdev->mc.aper_size = pci_resource_len(rdev->pdev, 0);
+	/* size in MB on si */
+	rdev->mc.mc_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024;
+	rdev->mc.real_vram_size = RREG32(CONFIG_MEMSIZE) * 1024 * 1024;
+	rdev->mc.visible_vram_size = rdev->mc.aper_size;
+	si_vram_gtt_location(rdev, &rdev->mc);
+	radeon_update_bandwidth_info(rdev);
+
+	return 0;
+}
+
+/*
+ * GART
+ * VMID 0 is the physical GPU addresses as used by the kernel.
+ * VMIDs 1-15 are used for userspace clients and are handled
+ * by the radeon vm/hsa code.
+ */
+/**
+ * cik_pcie_gart_tlb_flush - gart tlb flush callback
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Flush the TLB for the VMID 0 page table (CIK).
+ */
+void cik_pcie_gart_tlb_flush(struct radeon_device *rdev)
+{
+	/* flush hdp cache */
+	WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0);
+
+	/* bits 0-15 are the VM contexts0-15 */
+	WREG32(VM_INVALIDATE_REQUEST, 0x1);
+}
+
+/**
+ * cik_pcie_gart_enable - gart enable
+ *
+ * @rdev: radeon_device pointer
+ *
+ * This sets up the TLBs, programs the page tables for VMID0,
+ * sets up the hw for VMIDs 1-15 which are allocated on
+ * demand, and sets up the global locations for the LDS, GDS,
+ * and GPUVM for FSA64 clients (CIK).
+ * Returns 0 for success, errors for failure.
+ */
+static int cik_pcie_gart_enable(struct radeon_device *rdev)
+{
+	int r, i;
+
+	if (rdev->gart.robj == NULL) {
+		dev_err(rdev->dev, "No VRAM object for PCIE GART.\n");
+		return -EINVAL;
+	}
+	r = radeon_gart_table_vram_pin(rdev);
+	if (r)
+		return r;
+	radeon_gart_restore(rdev);
+	/* Setup TLB control */
+	WREG32(MC_VM_MX_L1_TLB_CNTL,
+	       (0xA << 7) |
+	       ENABLE_L1_TLB |
+	       SYSTEM_ACCESS_MODE_NOT_IN_SYS |
+	       ENABLE_ADVANCED_DRIVER_MODEL |
+	       SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU);
+	/* Setup L2 cache */
+	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE |
+	       ENABLE_L2_FRAGMENT_PROCESSING |
+	       ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |
+	       ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE |
+	       EFFECTIVE_L2_QUEUE_SIZE(7) |
+	       CONTEXT1_IDENTITY_ACCESS_MODE(1));
+	WREG32(VM_L2_CNTL2, INVALIDATE_ALL_L1_TLBS | INVALIDATE_L2_CACHE);
+	WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY |
+	       L2_CACHE_BIGK_FRAGMENT_SIZE(6));
+	/* setup context0 */
+	WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR, rdev->mc.gtt_start >> 12);
+	WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR, rdev->mc.gtt_end >> 12);
+	WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR, rdev->gart.table_addr >> 12);
+	WREG32(VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR,
+			(u32)(rdev->dummy_page.addr >> 12));
+	WREG32(VM_CONTEXT0_CNTL2, 0);
+	WREG32(VM_CONTEXT0_CNTL, (ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) |
+				  RANGE_PROTECTION_FAULT_ENABLE_DEFAULT));
+
+	WREG32(0x15D4, 0);
+	WREG32(0x15D8, 0);
+	WREG32(0x15DC, 0);
+
+	/* empty context1-15 */
+	/* FIXME start with 4G, once using 2 level pt switch to full
+	 * vm size space
+	 */
+	/* set vm size, must be a multiple of 4 */
+	WREG32(VM_CONTEXT1_PAGE_TABLE_START_ADDR, 0);
+	WREG32(VM_CONTEXT1_PAGE_TABLE_END_ADDR, rdev->vm_manager.max_pfn);
+	for (i = 1; i < 16; i++) {
+		if (i < 8)
+			WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2),
+			       rdev->gart.table_addr >> 12);
+		else
+			WREG32(VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((i - 8) << 2),
+			       rdev->gart.table_addr >> 12);
+	}
+
+	/* enable context1-15 */
+	WREG32(VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR,
+	       (u32)(rdev->dummy_page.addr >> 12));
+	WREG32(VM_CONTEXT1_CNTL2, 4);
+	WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(1) |
+				RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT |
+				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT |
+				DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT |
+				DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT |
+				PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT |
+				PDE0_PROTECTION_FAULT_ENABLE_DEFAULT |
+				VALID_PROTECTION_FAULT_ENABLE_INTERRUPT |
+				VALID_PROTECTION_FAULT_ENABLE_DEFAULT |
+				READ_PROTECTION_FAULT_ENABLE_INTERRUPT |
+				READ_PROTECTION_FAULT_ENABLE_DEFAULT |
+				WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT |
+				WRITE_PROTECTION_FAULT_ENABLE_DEFAULT);
+
+	/* TC cache setup ??? */
+	WREG32(TC_CFG_L1_LOAD_POLICY0, 0);
+	WREG32(TC_CFG_L1_LOAD_POLICY1, 0);
+	WREG32(TC_CFG_L1_STORE_POLICY, 0);
+
+	WREG32(TC_CFG_L2_LOAD_POLICY0, 0);
+	WREG32(TC_CFG_L2_LOAD_POLICY1, 0);
+	WREG32(TC_CFG_L2_STORE_POLICY0, 0);
+	WREG32(TC_CFG_L2_STORE_POLICY1, 0);
+	WREG32(TC_CFG_L2_ATOMIC_POLICY, 0);
+
+	WREG32(TC_CFG_L1_VOLATILE, 0);
+	WREG32(TC_CFG_L2_VOLATILE, 0);
+
+	if (rdev->family == CHIP_KAVERI) {
+		u32 tmp = RREG32(CHUB_CONTROL);
+		tmp &= ~BYPASS_VM;
+		WREG32(CHUB_CONTROL, tmp);
+	}
+
+	/* XXX SH_MEM regs */
+	/* where to put LDS, scratch, GPUVM in FSA64 space */
+	for (i = 0; i < 16; i++) {
+		cik_srbm_select(rdev, 0, 0, 0, i);
+		/* CP and shaders */
+		WREG32(SH_MEM_CONFIG, 0);
+		WREG32(SH_MEM_APE1_BASE, 1);
+		WREG32(SH_MEM_APE1_LIMIT, 0);
+		WREG32(SH_MEM_BASES, 0);
+		/* SDMA GFX */
+		WREG32(SDMA0_GFX_VIRTUAL_ADDR + SDMA0_REGISTER_OFFSET, 0);
+		WREG32(SDMA0_GFX_APE1_CNTL + SDMA0_REGISTER_OFFSET, 0);
+		WREG32(SDMA0_GFX_VIRTUAL_ADDR + SDMA1_REGISTER_OFFSET, 0);
+		WREG32(SDMA0_GFX_APE1_CNTL + SDMA1_REGISTER_OFFSET, 0);
+		/* XXX SDMA RLC - todo */
+	}
+	cik_srbm_select(rdev, 0, 0, 0, 0);
+
+	cik_pcie_gart_tlb_flush(rdev);
+	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
+		 (unsigned)(rdev->mc.gtt_size >> 20),
+		 (unsigned long long)rdev->gart.table_addr);
+	rdev->gart.ready = true;
+	return 0;
+}
+
+/**
+ * cik_pcie_gart_disable - gart disable
+ *
+ * @rdev: radeon_device pointer
+ *
+ * This disables all VM page table (CIK).
+ */
+static void cik_pcie_gart_disable(struct radeon_device *rdev)
+{
+	/* Disable all tables */
+	WREG32(VM_CONTEXT0_CNTL, 0);
+	WREG32(VM_CONTEXT1_CNTL, 0);
+	/* Setup TLB control */
+	WREG32(MC_VM_MX_L1_TLB_CNTL, SYSTEM_ACCESS_MODE_NOT_IN_SYS |
+	       SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU);
+	/* Setup L2 cache */
+	WREG32(VM_L2_CNTL,
+	       ENABLE_L2_FRAGMENT_PROCESSING |
+	       ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE |
+	       ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE |
+	       EFFECTIVE_L2_QUEUE_SIZE(7) |
+	       CONTEXT1_IDENTITY_ACCESS_MODE(1));
+	WREG32(VM_L2_CNTL2, 0);
+	WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY |
+	       L2_CACHE_BIGK_FRAGMENT_SIZE(6));
+	radeon_gart_table_vram_unpin(rdev);
+}
+
+/**
+ * cik_pcie_gart_fini - vm fini callback
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Tears down the driver GART/VM setup (CIK).
+ */
+static void cik_pcie_gart_fini(struct radeon_device *rdev)
+{
+	cik_pcie_gart_disable(rdev);
+	radeon_gart_table_vram_free(rdev);
+	radeon_gart_fini(rdev);
+}
+
+/* vm parser */
+/**
+ * cik_ib_parse - vm ib_parse callback
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer pointer
+ *
+ * CIK uses hw IB checking so this is a nop (CIK).
+ */
+int cik_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
+{
+	return 0;
+}
+
+/*
+ * vm
+ * VMID 0 is the physical GPU addresses as used by the kernel.
+ * VMIDs 1-15 are used for userspace clients and are handled
+ * by the radeon vm/hsa code.
+ */
+/**
+ * cik_vm_init - cik vm init callback
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Inits cik specific vm parameters (number of VMs, base of vram for
+ * VMIDs 1-15) (CIK).
+ * Returns 0 for success.
+ */
+int cik_vm_init(struct radeon_device *rdev)
+{
+	/* number of VMs */
+	rdev->vm_manager.nvm = 16;
+	/* base offset of vram pages */
+	if (rdev->flags & RADEON_IS_IGP) {
+		u64 tmp = RREG32(MC_VM_FB_OFFSET);
+		tmp <<= 22;
+		rdev->vm_manager.vram_base_offset = tmp;
+	} else
+		rdev->vm_manager.vram_base_offset = 0;
+
+	return 0;
+}
+
+/**
+ * cik_vm_fini - cik vm fini callback
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Tear down any asic specific VM setup (CIK).
+ */
+void cik_vm_fini(struct radeon_device *rdev)
+{
+}
+
+/**
+ * cik_vm_decode_fault - print human readable fault info
+ *
+ * @rdev: radeon_device pointer
+ * @status: VM_CONTEXT1_PROTECTION_FAULT_STATUS register value
+ * @addr: VM_CONTEXT1_PROTECTION_FAULT_ADDR register value
+ *
+ * Print human readable fault information (CIK).
+ */
+static void cik_vm_decode_fault(struct radeon_device *rdev,
+				u32 status, u32 addr, u32 mc_client)
+{
+	u32 mc_id = (status & MEMORY_CLIENT_ID_MASK) >> MEMORY_CLIENT_ID_SHIFT;
+	u32 vmid = (status & FAULT_VMID_MASK) >> FAULT_VMID_SHIFT;
+	u32 protections = (status & PROTECTIONS_MASK) >> PROTECTIONS_SHIFT;
+	char *block = (char *)&mc_client;
+
+	printk("VM fault (0x%02x, vmid %d) at page %u, %s from %s (%d)\n",
+	       protections, vmid, addr,
+	       (status & MEMORY_CLIENT_RW_MASK) ? "write" : "read",
+	       block, mc_id);
+}
+
+/**
+ * cik_vm_flush - cik vm flush using the CP
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Update the page table base and flush the VM TLB
+ * using the CP (CIK).
+ */
+void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
+{
+	struct radeon_ring *ring = &rdev->ring[ridx];
+
+	if (vm == NULL)
+		return;
+
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	if (vm->id < 8) {
+		radeon_ring_write(ring,
+				  (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2);
+	} else {
+		radeon_ring_write(ring,
+				  (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2);
+	}
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, vm->pd_gpu_addr >> 12);
+
+	/* update SH_MEM_* regs */
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, VMID(vm->id));
+
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 6));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, SH_MEM_BASES >> 2);
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, 0); /* SH_MEM_BASES */
+	radeon_ring_write(ring, 0); /* SH_MEM_CONFIG */
+	radeon_ring_write(ring, 1); /* SH_MEM_APE1_BASE */
+	radeon_ring_write(ring, 0); /* SH_MEM_APE1_LIMIT */
+
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, VMID(0));
+
+	/* HDP flush */
+	/* We should be using the WAIT_REG_MEM packet here like in
+	 * cik_fence_ring_emit(), but it causes the CP to hang in this
+	 * context...
+	 */
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, 0);
+
+	/* bits 0-15 are the VM contexts0-15 */
+	radeon_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	radeon_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) |
+				 WRITE_DATA_DST_SEL(0)));
+	radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
+	radeon_ring_write(ring, 0);
+	radeon_ring_write(ring, 1 << vm->id);
+
+	/* compute doesn't have PFP */
+	if (ridx == RADEON_RING_TYPE_GFX_INDEX) {
+		/* sync PFP to ME, otherwise we might get invalid PFP reads */
+		radeon_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
+		radeon_ring_write(ring, 0x0);
+	}
+}
+
+/**
+ * cik_vm_set_page - update the page tables using sDMA
+ *
+ * @rdev: radeon_device pointer
+ * @ib: indirect buffer to fill with commands
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: access flags
+ *
+ * Update the page tables using CP or sDMA (CIK).
+ */
+void cik_vm_set_page(struct radeon_device *rdev,
+		     struct radeon_ib *ib,
+		     uint64_t pe,
+		     uint64_t addr, unsigned count,
+		     uint32_t incr, uint32_t flags)
+{
+	uint32_t r600_flags = cayman_vm_page_flags(rdev, flags);
+	uint64_t value;
+	unsigned ndw;
+
+	if (rdev->asic->vm.pt_ring_index == RADEON_RING_TYPE_GFX_INDEX) {
+		/* CP */
+		while (count) {
+			ndw = 2 + count * 2;
+			if (ndw > 0x3FFE)
+				ndw = 0x3FFE;
+
+			ib->ptr[ib->length_dw++] = PACKET3(PACKET3_WRITE_DATA, ndw);
+			ib->ptr[ib->length_dw++] = (WRITE_DATA_ENGINE_SEL(0) |
+						    WRITE_DATA_DST_SEL(1));
+			ib->ptr[ib->length_dw++] = pe;
+			ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+			for (; ndw > 2; ndw -= 2, --count, pe += 8) {
+				if (flags & RADEON_VM_PAGE_SYSTEM) {
+					value = radeon_vm_map_gart(rdev, addr);
+					value &= 0xFFFFFFFFFFFFF000ULL;
+				} else if (flags & RADEON_VM_PAGE_VALID) {
+					value = addr;
+				} else {
+					value = 0;
+				}
+				addr += incr;
+				value |= r600_flags;
+				ib->ptr[ib->length_dw++] = value;
+				ib->ptr[ib->length_dw++] = upper_32_bits(value);
+			}
+		}
+	} else {
+		/* DMA */
+		if (flags & RADEON_VM_PAGE_SYSTEM) {
+			while (count) {
+				ndw = count * 2;
+				if (ndw > 0xFFFFE)
+					ndw = 0xFFFFE;
+
+				/* for non-physically contiguous pages (system) */
+				ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_WRITE, SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
+				ib->ptr[ib->length_dw++] = pe;
+				ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+				ib->ptr[ib->length_dw++] = ndw;
+				for (; ndw > 0; ndw -= 2, --count, pe += 8) {
+					if (flags & RADEON_VM_PAGE_SYSTEM) {
+						value = radeon_vm_map_gart(rdev, addr);
+						value &= 0xFFFFFFFFFFFFF000ULL;
+					} else if (flags & RADEON_VM_PAGE_VALID) {
+						value = addr;
+					} else {
+						value = 0;
+					}
+					addr += incr;
+					value |= r600_flags;
+					ib->ptr[ib->length_dw++] = value;
+					ib->ptr[ib->length_dw++] = upper_32_bits(value);
+				}
+			}
+		} else {
+			while (count) {
+				ndw = count;
+				if (ndw > 0x7FFFF)
+					ndw = 0x7FFFF;
+
+				if (flags & RADEON_VM_PAGE_VALID)
+					value = addr;
+				else
+					value = 0;
+				/* for physically contiguous pages (vram) */
+				ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_GENERATE_PTE_PDE, 0, 0);
+				ib->ptr[ib->length_dw++] = pe; /* dst addr */
+				ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+				ib->ptr[ib->length_dw++] = r600_flags; /* mask */
+				ib->ptr[ib->length_dw++] = 0;
+				ib->ptr[ib->length_dw++] = value; /* value */
+				ib->ptr[ib->length_dw++] = upper_32_bits(value);
+				ib->ptr[ib->length_dw++] = incr; /* increment size */
+				ib->ptr[ib->length_dw++] = 0;
+				ib->ptr[ib->length_dw++] = ndw; /* number of entries */
+				pe += ndw * 8;
+				addr += ndw * incr;
+				count -= ndw;
+			}
+		}
+		while (ib->length_dw & 0x7)
+			ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0);
+	}
+}
+
+/**
+ * cik_dma_vm_flush - cik vm flush using sDMA
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Update the page table base and flush the VM TLB
+ * using sDMA (CIK).
+ */
+void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
+{
+	struct radeon_ring *ring = &rdev->ring[ridx];
+	u32 extra_bits = (SDMA_POLL_REG_MEM_EXTRA_OP(1) |
+			  SDMA_POLL_REG_MEM_EXTRA_FUNC(3)); /* == */
+	u32 ref_and_mask;
+
+	if (vm == NULL)
+		return;
+
+	if (ridx == R600_RING_TYPE_DMA_INDEX)
+		ref_and_mask = SDMA0;
+	else
+		ref_and_mask = SDMA1;
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	if (vm->id < 8) {
+		radeon_ring_write(ring, (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2)) >> 2);
+	} else {
+		radeon_ring_write(ring, (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + ((vm->id - 8) << 2)) >> 2);
+	}
+	radeon_ring_write(ring, vm->pd_gpu_addr >> 12);
+
+	/* update SH_MEM_* regs */
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
+	radeon_ring_write(ring, VMID(vm->id));
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, SH_MEM_BASES >> 2);
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, SH_MEM_CONFIG >> 2);
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, SH_MEM_APE1_BASE >> 2);
+	radeon_ring_write(ring, 1);
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, SH_MEM_APE1_LIMIT >> 2);
+	radeon_ring_write(ring, 0);
+
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, SRBM_GFX_CNTL >> 2);
+	radeon_ring_write(ring, VMID(0));
+
+	/* flush HDP */
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_POLL_REG_MEM, 0, extra_bits));
+	radeon_ring_write(ring, GPU_HDP_FLUSH_DONE);
+	radeon_ring_write(ring, GPU_HDP_FLUSH_REQ);
+	radeon_ring_write(ring, ref_and_mask); /* REFERENCE */
+	radeon_ring_write(ring, ref_and_mask); /* MASK */
+	radeon_ring_write(ring, (4 << 16) | 10); /* RETRY_COUNT, POLL_INTERVAL */
+
+	/* flush TLB */
+	radeon_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SRBM_WRITE, 0, 0xf000));
+	radeon_ring_write(ring, VM_INVALIDATE_REQUEST >> 2);
+	radeon_ring_write(ring, 1 << vm->id);
+}
+
+/*
+ * RLC
+ * The RLC is a multi-purpose microengine that handles a
+ * variety of functions, the most important of which is
+ * the interrupt controller.
+ */
+/**
+ * cik_rlc_stop - stop the RLC ME
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Halt the RLC ME (MicroEngine) (CIK).
+ */
+static void cik_rlc_stop(struct radeon_device *rdev)
+{
+	int i, j, k;
+	u32 mask, tmp;
+
+	tmp = RREG32(CP_INT_CNTL_RING0);
+	tmp &= ~(CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	WREG32(CP_INT_CNTL_RING0, tmp);
+
+	RREG32(CB_CGTT_SCLK_CTRL);
+	RREG32(CB_CGTT_SCLK_CTRL);
+	RREG32(CB_CGTT_SCLK_CTRL);
+	RREG32(CB_CGTT_SCLK_CTRL);
+
+	tmp = RREG32(RLC_CGCG_CGLS_CTRL) & 0xfffffffc;
+	WREG32(RLC_CGCG_CGLS_CTRL, tmp);
+
+	WREG32(RLC_CNTL, 0);
+
+	for (i = 0; i < rdev->config.cik.max_shader_engines; i++) {
+		for (j = 0; j < rdev->config.cik.max_sh_per_se; j++) {
+			cik_select_se_sh(rdev, i, j);
+			for (k = 0; k < rdev->usec_timeout; k++) {
+				if (RREG32(RLC_SERDES_CU_MASTER_BUSY) == 0)
+					break;
+				udelay(1);
+			}
+		}
+	}
+	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+
+	mask = SE_MASTER_BUSY_MASK | GC_MASTER_BUSY | TC0_MASTER_BUSY | TC1_MASTER_BUSY;
+	for (k = 0; k < rdev->usec_timeout; k++) {
+		if ((RREG32(RLC_SERDES_NONCU_MASTER_BUSY) & mask) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+/**
+ * cik_rlc_start - start the RLC ME
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Unhalt the RLC ME (MicroEngine) (CIK).
+ */
+static void cik_rlc_start(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	WREG32(RLC_CNTL, RLC_ENABLE);
+
+	tmp = RREG32(CP_INT_CNTL_RING0);
+	tmp |= (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	WREG32(CP_INT_CNTL_RING0, tmp);
+
+	udelay(50);
+}
+
+/**
+ * cik_rlc_resume - setup the RLC hw
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Initialize the RLC registers, load the ucode,
+ * and start the RLC (CIK).
+ * Returns 0 for success, -EINVAL if the ucode is not available.
+ */
+static int cik_rlc_resume(struct radeon_device *rdev)
+{
+	u32 i, size;
+	u32 clear_state_info[3];
+	const __be32 *fw_data;
+
+	if (!rdev->rlc_fw)
+		return -EINVAL;
+
+	switch (rdev->family) {
+	case CHIP_BONAIRE:
+	default:
+		size = BONAIRE_RLC_UCODE_SIZE;
+		break;
+	case CHIP_KAVERI:
+		size = KV_RLC_UCODE_SIZE;
+		break;
+	case CHIP_KABINI:
+		size = KB_RLC_UCODE_SIZE;
+		break;
+	}
+
+	cik_rlc_stop(rdev);
+
+	WREG32(GRBM_SOFT_RESET, SOFT_RESET_RLC);
+	RREG32(GRBM_SOFT_RESET);
+	udelay(50);
+	WREG32(GRBM_SOFT_RESET, 0);
+	RREG32(GRBM_SOFT_RESET);
+	udelay(50);
+
+	WREG32(RLC_LB_CNTR_INIT, 0);
+	WREG32(RLC_LB_CNTR_MAX, 0x00008000);
+
+	cik_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+	WREG32(RLC_LB_INIT_CU_MASK, 0xffffffff);
+	WREG32(RLC_LB_PARAMS, 0x00600408);
+	WREG32(RLC_LB_CNTL, 0x80000004);
+
+	WREG32(RLC_MC_CNTL, 0);
+	WREG32(RLC_UCODE_CNTL, 0);
+
+	fw_data = (const __be32 *)rdev->rlc_fw->data;
+		WREG32(RLC_GPM_UCODE_ADDR, 0);
+	for (i = 0; i < size; i++)
+		WREG32(RLC_GPM_UCODE_DATA, be32_to_cpup(fw_data++));
+	WREG32(RLC_GPM_UCODE_ADDR, 0);
+
+	/* XXX */
+	clear_state_info[0] = 0;//upper_32_bits(rdev->rlc.save_restore_gpu_addr);
+	clear_state_info[1] = 0;//rdev->rlc.save_restore_gpu_addr;
+	clear_state_info[2] = 0;//cik_default_size;
+	WREG32(RLC_GPM_SCRATCH_ADDR, 0x3d);
+	for (i = 0; i < 3; i++)
+		WREG32(RLC_GPM_SCRATCH_DATA, clear_state_info[i]);
+	WREG32(RLC_DRIVER_DMA_STATUS, 0);
+
+	cik_rlc_start(rdev);
+
+	return 0;
+}
+
+/*
+ * Interrupts
+ * Starting with r6xx, interrupts are handled via a ring buffer.
+ * Ring buffers are areas of GPU accessible memory that the GPU
+ * writes interrupt vectors into and the host reads vectors out of.
+ * There is a rptr (read pointer) that determines where the
+ * host is currently reading, and a wptr (write pointer)
+ * which determines where the GPU has written.  When the
+ * pointers are equal, the ring is idle.  When the GPU
+ * writes vectors to the ring buffer, it increments the
+ * wptr.  When there is an interrupt, the host then starts
+ * fetching commands and processing them until the pointers are
+ * equal again at which point it updates the rptr.
+ */
+
+/**
+ * cik_enable_interrupts - Enable the interrupt ring buffer
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Enable the interrupt ring buffer (CIK).
+ */
+static void cik_enable_interrupts(struct radeon_device *rdev)
+{
+	u32 ih_cntl = RREG32(IH_CNTL);
+	u32 ih_rb_cntl = RREG32(IH_RB_CNTL);
+
+	ih_cntl |= ENABLE_INTR;
+	ih_rb_cntl |= IH_RB_ENABLE;
+	WREG32(IH_CNTL, ih_cntl);
+	WREG32(IH_RB_CNTL, ih_rb_cntl);
+	rdev->ih.enabled = true;
+}
+
+/**
+ * cik_disable_interrupts - Disable the interrupt ring buffer
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Disable the interrupt ring buffer (CIK).
+ */
+static void cik_disable_interrupts(struct radeon_device *rdev)
+{
+	u32 ih_rb_cntl = RREG32(IH_RB_CNTL);
+	u32 ih_cntl = RREG32(IH_CNTL);
+
+	ih_rb_cntl &= ~IH_RB_ENABLE;
+	ih_cntl &= ~ENABLE_INTR;
+	WREG32(IH_RB_CNTL, ih_rb_cntl);
+	WREG32(IH_CNTL, ih_cntl);
+	/* set rptr, wptr to 0 */
+	WREG32(IH_RB_RPTR, 0);
+	WREG32(IH_RB_WPTR, 0);
+	rdev->ih.enabled = false;
+	rdev->ih.rptr = 0;
+}
+
+/**
+ * cik_disable_interrupt_state - Disable all interrupt sources
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Clear all interrupt enable bits used by the driver (CIK).
+ */
+static void cik_disable_interrupt_state(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	/* gfx ring */
+	WREG32(CP_INT_CNTL_RING0, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	/* sdma */
+	tmp = RREG32(SDMA0_CNTL + SDMA0_REGISTER_OFFSET) & ~TRAP_ENABLE;
+	WREG32(SDMA0_CNTL + SDMA0_REGISTER_OFFSET, tmp);
+	tmp = RREG32(SDMA0_CNTL + SDMA1_REGISTER_OFFSET) & ~TRAP_ENABLE;
+	WREG32(SDMA0_CNTL + SDMA1_REGISTER_OFFSET, tmp);
+	/* compute queues */
+	WREG32(CP_ME1_PIPE0_INT_CNTL, 0);
+	WREG32(CP_ME1_PIPE1_INT_CNTL, 0);
+	WREG32(CP_ME1_PIPE2_INT_CNTL, 0);
+	WREG32(CP_ME1_PIPE3_INT_CNTL, 0);
+	WREG32(CP_ME2_PIPE0_INT_CNTL, 0);
+	WREG32(CP_ME2_PIPE1_INT_CNTL, 0);
+	WREG32(CP_ME2_PIPE2_INT_CNTL, 0);
+	WREG32(CP_ME2_PIPE3_INT_CNTL, 0);
+	/* grbm */
+	WREG32(GRBM_INT_CNTL, 0);
+	/* vline/vblank, etc. */
+	WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, 0);
+	WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, 0);
+	if (rdev->num_crtc >= 4) {
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC2_REGISTER_OFFSET, 0);
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC3_REGISTER_OFFSET, 0);
+	}
+	if (rdev->num_crtc >= 6) {
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, 0);
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, 0);
+	}
+
+	/* dac hotplug */
+	WREG32(DAC_AUTODETECT_INT_CONTROL, 0);
+
+	/* digital hotplug */
+	tmp = RREG32(DC_HPD1_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD1_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD2_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD2_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD3_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD3_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD4_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD4_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD5_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD5_INT_CONTROL, tmp);
+	tmp = RREG32(DC_HPD6_INT_CONTROL) & DC_HPDx_INT_POLARITY;
+	WREG32(DC_HPD6_INT_CONTROL, tmp);
+
+}
+
+/**
+ * cik_irq_init - init and enable the interrupt ring
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Allocate a ring buffer for the interrupt controller,
+ * enable the RLC, disable interrupts, enable the IH
+ * ring buffer and enable it (CIK).
+ * Called at device load and reume.
+ * Returns 0 for success, errors for failure.
+ */
+static int cik_irq_init(struct radeon_device *rdev)
+{
+	int ret = 0;
+	int rb_bufsz;
+	u32 interrupt_cntl, ih_cntl, ih_rb_cntl;
+
+	/* allocate ring */
+	ret = r600_ih_ring_alloc(rdev);
+	if (ret)
+		return ret;
+
+	/* disable irqs */
+	cik_disable_interrupts(rdev);
+
+	/* init rlc */
+	ret = cik_rlc_resume(rdev);
+	if (ret) {
+		r600_ih_ring_fini(rdev);
+		return ret;
+	}
+
+	/* setup interrupt control */
+	/* XXX this should actually be a bus address, not an MC address. same on older asics */
+	WREG32(INTERRUPT_CNTL2, rdev->ih.gpu_addr >> 8);
+	interrupt_cntl = RREG32(INTERRUPT_CNTL);
+	/* IH_DUMMY_RD_OVERRIDE=0 - dummy read disabled with msi, enabled without msi
+	 * IH_DUMMY_RD_OVERRIDE=1 - dummy read controlled by IH_DUMMY_RD_EN
+	 */
+	interrupt_cntl &= ~IH_DUMMY_RD_OVERRIDE;
+	/* IH_REQ_NONSNOOP_EN=1 if ring is in non-cacheable memory, e.g., vram */
+	interrupt_cntl &= ~IH_REQ_NONSNOOP_EN;
+	WREG32(INTERRUPT_CNTL, interrupt_cntl);
+
+	WREG32(IH_RB_BASE, rdev->ih.gpu_addr >> 8);
+	rb_bufsz = drm_order(rdev->ih.ring_size / 4);
+
+	ih_rb_cntl = (IH_WPTR_OVERFLOW_ENABLE |
+		      IH_WPTR_OVERFLOW_CLEAR |
+		      (rb_bufsz << 1));
+
+	if (rdev->wb.enabled)
+		ih_rb_cntl |= IH_WPTR_WRITEBACK_ENABLE;
+
+	/* set the writeback address whether it's enabled or not */
+	WREG32(IH_RB_WPTR_ADDR_LO, (rdev->wb.gpu_addr + R600_WB_IH_WPTR_OFFSET) & 0xFFFFFFFC);
+	WREG32(IH_RB_WPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + R600_WB_IH_WPTR_OFFSET) & 0xFF);
+
+	WREG32(IH_RB_CNTL, ih_rb_cntl);
+
+	/* set rptr, wptr to 0 */
+	WREG32(IH_RB_RPTR, 0);
+	WREG32(IH_RB_WPTR, 0);
+
+	/* Default settings for IH_CNTL (disabled at first) */
+	ih_cntl = MC_WRREQ_CREDIT(0x10) | MC_WR_CLEAN_CNT(0x10) | MC_VMID(0);
+	/* RPTR_REARM only works if msi's are enabled */
+	if (rdev->msi_enabled)
+		ih_cntl |= RPTR_REARM;
+	WREG32(IH_CNTL, ih_cntl);
+
+	/* force the active interrupt state to all disabled */
+	cik_disable_interrupt_state(rdev);
+
+	pci_set_master(rdev->pdev);
+
+	/* enable irqs */
+	cik_enable_interrupts(rdev);
+
+	return ret;
+}
+
+/**
+ * cik_irq_set - enable/disable interrupt sources
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Enable interrupt sources on the GPU (vblanks, hpd,
+ * etc.) (CIK).
+ * Returns 0 for success, errors for failure.
+ */
+int cik_irq_set(struct radeon_device *rdev)
+{
+	u32 cp_int_cntl = CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE |
+		PRIV_INSTR_INT_ENABLE | PRIV_REG_INT_ENABLE;
+	u32 cp_m1p0, cp_m1p1, cp_m1p2, cp_m1p3;
+	u32 cp_m2p0, cp_m2p1, cp_m2p2, cp_m2p3;
+	u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0;
+	u32 hpd1, hpd2, hpd3, hpd4, hpd5, hpd6;
+	u32 grbm_int_cntl = 0;
+	u32 dma_cntl, dma_cntl1;
+
+	if (!rdev->irq.installed) {
+		WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
+		return -EINVAL;
+	}
+	/* don't enable anything if the ih is disabled */
+	if (!rdev->ih.enabled) {
+		cik_disable_interrupts(rdev);
+		/* force the active interrupt state to all disabled */
+		cik_disable_interrupt_state(rdev);
+		return 0;
+	}
+
+	hpd1 = RREG32(DC_HPD1_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd2 = RREG32(DC_HPD2_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd3 = RREG32(DC_HPD3_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd4 = RREG32(DC_HPD4_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd5 = RREG32(DC_HPD5_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	hpd6 = RREG32(DC_HPD6_INT_CONTROL) & ~DC_HPDx_INT_EN;
+
+	dma_cntl = RREG32(SDMA0_CNTL + SDMA0_REGISTER_OFFSET) & ~TRAP_ENABLE;
+	dma_cntl1 = RREG32(SDMA0_CNTL + SDMA1_REGISTER_OFFSET) & ~TRAP_ENABLE;
+
+	cp_m1p0 = RREG32(CP_ME1_PIPE0_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m1p1 = RREG32(CP_ME1_PIPE1_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m1p2 = RREG32(CP_ME1_PIPE2_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m1p3 = RREG32(CP_ME1_PIPE3_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m2p0 = RREG32(CP_ME2_PIPE0_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m2p1 = RREG32(CP_ME2_PIPE1_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m2p2 = RREG32(CP_ME2_PIPE2_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+	cp_m2p3 = RREG32(CP_ME2_PIPE3_INT_CNTL) & ~TIME_STAMP_INT_ENABLE;
+
+	/* enable CP interrupts on all rings */
+	if (atomic_read(&rdev->irq.ring_int[RADEON_RING_TYPE_GFX_INDEX])) {
+		DRM_DEBUG("cik_irq_set: sw int gfx\n");
+		cp_int_cntl |= TIME_STAMP_INT_ENABLE;
+	}
+	if (atomic_read(&rdev->irq.ring_int[CAYMAN_RING_TYPE_CP1_INDEX])) {
+		struct radeon_ring *ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
+		DRM_DEBUG("si_irq_set: sw int cp1\n");
+		if (ring->me == 1) {
+			switch (ring->pipe) {
+			case 0:
+				cp_m1p0 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 1:
+				cp_m1p1 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 2:
+				cp_m1p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 3:
+				cp_m1p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			default:
+				DRM_DEBUG("si_irq_set: sw int cp1 invalid pipe %d\n", ring->pipe);
+				break;
+			}
+		} else if (ring->me == 2) {
+			switch (ring->pipe) {
+			case 0:
+				cp_m2p0 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 1:
+				cp_m2p1 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 2:
+				cp_m2p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 3:
+				cp_m2p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			default:
+				DRM_DEBUG("si_irq_set: sw int cp1 invalid pipe %d\n", ring->pipe);
+				break;
+			}
+		} else {
+			DRM_DEBUG("si_irq_set: sw int cp1 invalid me %d\n", ring->me);
+		}
+	}
+	if (atomic_read(&rdev->irq.ring_int[CAYMAN_RING_TYPE_CP2_INDEX])) {
+		struct radeon_ring *ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
+		DRM_DEBUG("si_irq_set: sw int cp2\n");
+		if (ring->me == 1) {
+			switch (ring->pipe) {
+			case 0:
+				cp_m1p0 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 1:
+				cp_m1p1 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 2:
+				cp_m1p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 3:
+				cp_m1p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			default:
+				DRM_DEBUG("si_irq_set: sw int cp2 invalid pipe %d\n", ring->pipe);
+				break;
+			}
+		} else if (ring->me == 2) {
+			switch (ring->pipe) {
+			case 0:
+				cp_m2p0 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 1:
+				cp_m2p1 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 2:
+				cp_m2p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			case 3:
+				cp_m2p2 |= TIME_STAMP_INT_ENABLE;
+				break;
+			default:
+				DRM_DEBUG("si_irq_set: sw int cp2 invalid pipe %d\n", ring->pipe);
+				break;
+			}
+		} else {
+			DRM_DEBUG("si_irq_set: sw int cp2 invalid me %d\n", ring->me);
+		}
+	}
+
+	if (atomic_read(&rdev->irq.ring_int[R600_RING_TYPE_DMA_INDEX])) {
+		DRM_DEBUG("cik_irq_set: sw int dma\n");
+		dma_cntl |= TRAP_ENABLE;
+	}
+
+	if (atomic_read(&rdev->irq.ring_int[CAYMAN_RING_TYPE_DMA1_INDEX])) {
+		DRM_DEBUG("cik_irq_set: sw int dma1\n");
+		dma_cntl1 |= TRAP_ENABLE;
+	}
+
+	if (rdev->irq.crtc_vblank_int[0] ||
+	    atomic_read(&rdev->irq.pflip[0])) {
+		DRM_DEBUG("cik_irq_set: vblank 0\n");
+		crtc1 |= VBLANK_INTERRUPT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[1] ||
+	    atomic_read(&rdev->irq.pflip[1])) {
+		DRM_DEBUG("cik_irq_set: vblank 1\n");
+		crtc2 |= VBLANK_INTERRUPT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[2] ||
+	    atomic_read(&rdev->irq.pflip[2])) {
+		DRM_DEBUG("cik_irq_set: vblank 2\n");
+		crtc3 |= VBLANK_INTERRUPT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[3] ||
+	    atomic_read(&rdev->irq.pflip[3])) {
+		DRM_DEBUG("cik_irq_set: vblank 3\n");
+		crtc4 |= VBLANK_INTERRUPT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[4] ||
+	    atomic_read(&rdev->irq.pflip[4])) {
+		DRM_DEBUG("cik_irq_set: vblank 4\n");
+		crtc5 |= VBLANK_INTERRUPT_MASK;
+	}
+	if (rdev->irq.crtc_vblank_int[5] ||
+	    atomic_read(&rdev->irq.pflip[5])) {
+		DRM_DEBUG("cik_irq_set: vblank 5\n");
+		crtc6 |= VBLANK_INTERRUPT_MASK;
+	}
+	if (rdev->irq.hpd[0]) {
+		DRM_DEBUG("cik_irq_set: hpd 1\n");
+		hpd1 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[1]) {
+		DRM_DEBUG("cik_irq_set: hpd 2\n");
+		hpd2 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[2]) {
+		DRM_DEBUG("cik_irq_set: hpd 3\n");
+		hpd3 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[3]) {
+		DRM_DEBUG("cik_irq_set: hpd 4\n");
+		hpd4 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[4]) {
+		DRM_DEBUG("cik_irq_set: hpd 5\n");
+		hpd5 |= DC_HPDx_INT_EN;
+	}
+	if (rdev->irq.hpd[5]) {
+		DRM_DEBUG("cik_irq_set: hpd 6\n");
+		hpd6 |= DC_HPDx_INT_EN;
+	}
+
+	WREG32(CP_INT_CNTL_RING0, cp_int_cntl);
+
+	WREG32(SDMA0_CNTL + SDMA0_REGISTER_OFFSET, dma_cntl);
+	WREG32(SDMA0_CNTL + SDMA1_REGISTER_OFFSET, dma_cntl1);
+
+	WREG32(CP_ME1_PIPE0_INT_CNTL, cp_m1p0);
+	WREG32(CP_ME1_PIPE1_INT_CNTL, cp_m1p1);
+	WREG32(CP_ME1_PIPE2_INT_CNTL, cp_m1p2);
+	WREG32(CP_ME1_PIPE3_INT_CNTL, cp_m1p3);
+	WREG32(CP_ME2_PIPE0_INT_CNTL, cp_m2p0);
+	WREG32(CP_ME2_PIPE1_INT_CNTL, cp_m2p1);
+	WREG32(CP_ME2_PIPE2_INT_CNTL, cp_m2p2);
+	WREG32(CP_ME2_PIPE3_INT_CNTL, cp_m2p3);
+
+	WREG32(GRBM_INT_CNTL, grbm_int_cntl);
+
+	WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, crtc1);
+	WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, crtc2);
+	if (rdev->num_crtc >= 4) {
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC2_REGISTER_OFFSET, crtc3);
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC3_REGISTER_OFFSET, crtc4);
+	}
+	if (rdev->num_crtc >= 6) {
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC4_REGISTER_OFFSET, crtc5);
+		WREG32(LB_INTERRUPT_MASK + EVERGREEN_CRTC5_REGISTER_OFFSET, crtc6);
+	}
+
+	WREG32(DC_HPD1_INT_CONTROL, hpd1);
+	WREG32(DC_HPD2_INT_CONTROL, hpd2);
+	WREG32(DC_HPD3_INT_CONTROL, hpd3);
+	WREG32(DC_HPD4_INT_CONTROL, hpd4);
+	WREG32(DC_HPD5_INT_CONTROL, hpd5);
+	WREG32(DC_HPD6_INT_CONTROL, hpd6);
+
+	return 0;
+}
+
+/**
+ * cik_irq_ack - ack interrupt sources
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Ack interrupt sources on the GPU (vblanks, hpd,
+ * etc.) (CIK).  Certain interrupts sources are sw
+ * generated and do not require an explicit ack.
+ */
+static inline void cik_irq_ack(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	rdev->irq.stat_regs.cik.disp_int = RREG32(DISP_INTERRUPT_STATUS);
+	rdev->irq.stat_regs.cik.disp_int_cont = RREG32(DISP_INTERRUPT_STATUS_CONTINUE);
+	rdev->irq.stat_regs.cik.disp_int_cont2 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE2);
+	rdev->irq.stat_regs.cik.disp_int_cont3 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE3);
+	rdev->irq.stat_regs.cik.disp_int_cont4 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE4);
+	rdev->irq.stat_regs.cik.disp_int_cont5 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE5);
+	rdev->irq.stat_regs.cik.disp_int_cont6 = RREG32(DISP_INTERRUPT_STATUS_CONTINUE6);
+
+	if (rdev->irq.stat_regs.cik.disp_int & LB_D1_VBLANK_INTERRUPT)
+		WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VBLANK_ACK);
+	if (rdev->irq.stat_regs.cik.disp_int & LB_D1_VLINE_INTERRUPT)
+		WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC0_REGISTER_OFFSET, VLINE_ACK);
+	if (rdev->irq.stat_regs.cik.disp_int_cont & LB_D2_VBLANK_INTERRUPT)
+		WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VBLANK_ACK);
+	if (rdev->irq.stat_regs.cik.disp_int_cont & LB_D2_VLINE_INTERRUPT)
+		WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC1_REGISTER_OFFSET, VLINE_ACK);
+
+	if (rdev->num_crtc >= 4) {
+		if (rdev->irq.stat_regs.cik.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT)
+			WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.cik.disp_int_cont2 & LB_D3_VLINE_INTERRUPT)
+			WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC2_REGISTER_OFFSET, VLINE_ACK);
+		if (rdev->irq.stat_regs.cik.disp_int_cont3 & LB_D4_VBLANK_INTERRUPT)
+			WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.cik.disp_int_cont3 & LB_D4_VLINE_INTERRUPT)
+			WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC3_REGISTER_OFFSET, VLINE_ACK);
+	}
+
+	if (rdev->num_crtc >= 6) {
+		if (rdev->irq.stat_regs.cik.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT)
+			WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.cik.disp_int_cont4 & LB_D5_VLINE_INTERRUPT)
+			WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC4_REGISTER_OFFSET, VLINE_ACK);
+		if (rdev->irq.stat_regs.cik.disp_int_cont5 & LB_D6_VBLANK_INTERRUPT)
+			WREG32(LB_VBLANK_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, VBLANK_ACK);
+		if (rdev->irq.stat_regs.cik.disp_int_cont5 & LB_D6_VLINE_INTERRUPT)
+			WREG32(LB_VLINE_STATUS + EVERGREEN_CRTC5_REGISTER_OFFSET, VLINE_ACK);
+	}
+
+	if (rdev->irq.stat_regs.cik.disp_int & DC_HPD1_INTERRUPT) {
+		tmp = RREG32(DC_HPD1_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD1_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.cik.disp_int_cont & DC_HPD2_INTERRUPT) {
+		tmp = RREG32(DC_HPD2_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD2_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.cik.disp_int_cont2 & DC_HPD3_INTERRUPT) {
+		tmp = RREG32(DC_HPD3_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD3_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.cik.disp_int_cont3 & DC_HPD4_INTERRUPT) {
+		tmp = RREG32(DC_HPD4_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD4_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.cik.disp_int_cont4 & DC_HPD5_INTERRUPT) {
+		tmp = RREG32(DC_HPD5_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD5_INT_CONTROL, tmp);
+	}
+	if (rdev->irq.stat_regs.cik.disp_int_cont5 & DC_HPD6_INTERRUPT) {
+		tmp = RREG32(DC_HPD5_INT_CONTROL);
+		tmp |= DC_HPDx_INT_ACK;
+		WREG32(DC_HPD6_INT_CONTROL, tmp);
+	}
+}
+
+/**
+ * cik_irq_disable - disable interrupts
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Disable interrupts on the hw (CIK).
+ */
+static void cik_irq_disable(struct radeon_device *rdev)
+{
+	cik_disable_interrupts(rdev);
+	/* Wait and acknowledge irq */
+	mdelay(1);
+	cik_irq_ack(rdev);
+	cik_disable_interrupt_state(rdev);
+}
+
+/**
+ * cik_irq_disable - disable interrupts for suspend
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Disable interrupts and stop the RLC (CIK).
+ * Used for suspend.
+ */
+static void cik_irq_suspend(struct radeon_device *rdev)
+{
+	cik_irq_disable(rdev);
+	cik_rlc_stop(rdev);
+}
+
+/**
+ * cik_irq_fini - tear down interrupt support
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Disable interrupts on the hw and free the IH ring
+ * buffer (CIK).
+ * Used for driver unload.
+ */
+static void cik_irq_fini(struct radeon_device *rdev)
+{
+	cik_irq_suspend(rdev);
+	r600_ih_ring_fini(rdev);
+}
+
+/**
+ * cik_get_ih_wptr - get the IH ring buffer wptr
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Get the IH ring buffer wptr from either the register
+ * or the writeback memory buffer (CIK).  Also check for
+ * ring buffer overflow and deal with it.
+ * Used by cik_irq_process().
+ * Returns the value of the wptr.
+ */
+static inline u32 cik_get_ih_wptr(struct radeon_device *rdev)
+{
+	u32 wptr, tmp;
+
+	if (rdev->wb.enabled)
+		wptr = le32_to_cpu(rdev->wb.wb[R600_WB_IH_WPTR_OFFSET/4]);
+	else
+		wptr = RREG32(IH_RB_WPTR);
+
+	if (wptr & RB_OVERFLOW) {
+		/* When a ring buffer overflow happen start parsing interrupt
+		 * from the last not overwritten vector (wptr + 16). Hopefully
+		 * this should allow us to catchup.
+		 */
+		dev_warn(rdev->dev, "IH ring buffer overflow (0x%08X, %d, %d)\n",
+			wptr, rdev->ih.rptr, (wptr + 16) + rdev->ih.ptr_mask);
+		rdev->ih.rptr = (wptr + 16) & rdev->ih.ptr_mask;
+		tmp = RREG32(IH_RB_CNTL);
+		tmp |= IH_WPTR_OVERFLOW_CLEAR;
+		WREG32(IH_RB_CNTL, tmp);
+	}
+	return (wptr & rdev->ih.ptr_mask);
+}
+
+/*        CIK IV Ring
+ * Each IV ring entry is 128 bits:
+ * [7:0]    - interrupt source id
+ * [31:8]   - reserved
+ * [59:32]  - interrupt source data
+ * [63:60]  - reserved
+ * [71:64]  - RINGID
+ *            CP:
+ *            ME_ID [1:0], PIPE_ID[1:0], QUEUE_ID[2:0]
+ *            QUEUE_ID - for compute, which of the 8 queues owned by the dispatcher
+ *                     - for gfx, hw shader state (0=PS...5=LS, 6=CS)
+ *            ME_ID - 0 = gfx, 1 = first 4 CS pipes, 2 = second 4 CS pipes
+ *            PIPE_ID - ME0 0=3D
+ *                    - ME1&2 compute dispatcher (4 pipes each)
+ *            SDMA:
+ *            INSTANCE_ID [1:0], QUEUE_ID[1:0]
+ *            INSTANCE_ID - 0 = sdma0, 1 = sdma1
+ *            QUEUE_ID - 0 = gfx, 1 = rlc0, 2 = rlc1
+ * [79:72]  - VMID
+ * [95:80]  - PASID
+ * [127:96] - reserved
+ */
+/**
+ * cik_irq_process - interrupt handler
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Interrupt hander (CIK).  Walk the IH ring,
+ * ack interrupts and schedule work to handle
+ * interrupt events.
+ * Returns irq process return code.
+ */
+int cik_irq_process(struct radeon_device *rdev)
+{
+	struct radeon_ring *cp1_ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
+	struct radeon_ring *cp2_ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
+	u32 wptr;
+	u32 rptr;
+	u32 src_id, src_data, ring_id;
+	u8 me_id, pipe_id, queue_id;
+	u32 ring_index;
+	bool queue_hotplug = false;
+	bool queue_reset = false;
+	u32 addr, status, mc_client;
+
+	if (!rdev->ih.enabled || rdev->shutdown)
+		return IRQ_NONE;
+
+	wptr = cik_get_ih_wptr(rdev);
+
+restart_ih:
+	/* is somebody else already processing irqs? */
+	if (atomic_xchg(&rdev->ih.lock, 1))
+		return IRQ_NONE;
+
+	rptr = rdev->ih.rptr;
+	DRM_DEBUG("cik_irq_process start: rptr %d, wptr %d\n", rptr, wptr);
+
+	/* Order reading of wptr vs. reading of IH ring data */
+	rmb();
+
+	/* display interrupts */
+	cik_irq_ack(rdev);
+
+	while (rptr != wptr) {
+		/* wptr/rptr are in bytes! */
+		ring_index = rptr / 4;
+		src_id =  le32_to_cpu(rdev->ih.ring[ring_index]) & 0xff;
+		src_data = le32_to_cpu(rdev->ih.ring[ring_index + 1]) & 0xfffffff;
+		ring_id = le32_to_cpu(rdev->ih.ring[ring_index + 2]) & 0xff;
+
+		switch (src_id) {
+		case 1: /* D1 vblank/vline */
+			switch (src_data) {
+			case 0: /* D1 vblank */
+				if (rdev->irq.stat_regs.cik.disp_int & LB_D1_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[0]) {
+						drm_handle_vblank(rdev->ddev, 0);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (atomic_read(&rdev->irq.pflip[0]))
+						radeon_crtc_handle_flip(rdev, 0);
+					rdev->irq.stat_regs.cik.disp_int &= ~LB_D1_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D1 vblank\n");
+				}
+				break;
+			case 1: /* D1 vline */
+				if (rdev->irq.stat_regs.cik.disp_int & LB_D1_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int &= ~LB_D1_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D1 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 2: /* D2 vblank/vline */
+			switch (src_data) {
+			case 0: /* D2 vblank */
+				if (rdev->irq.stat_regs.cik.disp_int_cont & LB_D2_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[1]) {
+						drm_handle_vblank(rdev->ddev, 1);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (atomic_read(&rdev->irq.pflip[1]))
+						radeon_crtc_handle_flip(rdev, 1);
+					rdev->irq.stat_regs.cik.disp_int_cont &= ~LB_D2_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D2 vblank\n");
+				}
+				break;
+			case 1: /* D2 vline */
+				if (rdev->irq.stat_regs.cik.disp_int_cont & LB_D2_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont &= ~LB_D2_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D2 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 3: /* D3 vblank/vline */
+			switch (src_data) {
+			case 0: /* D3 vblank */
+				if (rdev->irq.stat_regs.cik.disp_int_cont2 & LB_D3_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[2]) {
+						drm_handle_vblank(rdev->ddev, 2);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (atomic_read(&rdev->irq.pflip[2]))
+						radeon_crtc_handle_flip(rdev, 2);
+					rdev->irq.stat_regs.cik.disp_int_cont2 &= ~LB_D3_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D3 vblank\n");
+				}
+				break;
+			case 1: /* D3 vline */
+				if (rdev->irq.stat_regs.cik.disp_int_cont2 & LB_D3_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont2 &= ~LB_D3_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D3 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 4: /* D4 vblank/vline */
+			switch (src_data) {
+			case 0: /* D4 vblank */
+				if (rdev->irq.stat_regs.cik.disp_int_cont3 & LB_D4_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[3]) {
+						drm_handle_vblank(rdev->ddev, 3);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (atomic_read(&rdev->irq.pflip[3]))
+						radeon_crtc_handle_flip(rdev, 3);
+					rdev->irq.stat_regs.cik.disp_int_cont3 &= ~LB_D4_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D4 vblank\n");
+				}
+				break;
+			case 1: /* D4 vline */
+				if (rdev->irq.stat_regs.cik.disp_int_cont3 & LB_D4_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont3 &= ~LB_D4_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D4 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 5: /* D5 vblank/vline */
+			switch (src_data) {
+			case 0: /* D5 vblank */
+				if (rdev->irq.stat_regs.cik.disp_int_cont4 & LB_D5_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[4]) {
+						drm_handle_vblank(rdev->ddev, 4);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (atomic_read(&rdev->irq.pflip[4]))
+						radeon_crtc_handle_flip(rdev, 4);
+					rdev->irq.stat_regs.cik.disp_int_cont4 &= ~LB_D5_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D5 vblank\n");
+				}
+				break;
+			case 1: /* D5 vline */
+				if (rdev->irq.stat_regs.cik.disp_int_cont4 & LB_D5_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont4 &= ~LB_D5_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D5 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 6: /* D6 vblank/vline */
+			switch (src_data) {
+			case 0: /* D6 vblank */
+				if (rdev->irq.stat_regs.cik.disp_int_cont5 & LB_D6_VBLANK_INTERRUPT) {
+					if (rdev->irq.crtc_vblank_int[5]) {
+						drm_handle_vblank(rdev->ddev, 5);
+						rdev->pm.vblank_sync = true;
+						wake_up(&rdev->irq.vblank_queue);
+					}
+					if (atomic_read(&rdev->irq.pflip[5]))
+						radeon_crtc_handle_flip(rdev, 5);
+					rdev->irq.stat_regs.cik.disp_int_cont5 &= ~LB_D6_VBLANK_INTERRUPT;
+					DRM_DEBUG("IH: D6 vblank\n");
+				}
+				break;
+			case 1: /* D6 vline */
+				if (rdev->irq.stat_regs.cik.disp_int_cont5 & LB_D6_VLINE_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont5 &= ~LB_D6_VLINE_INTERRUPT;
+					DRM_DEBUG("IH: D6 vline\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 42: /* HPD hotplug */
+			switch (src_data) {
+			case 0:
+				if (rdev->irq.stat_regs.cik.disp_int & DC_HPD1_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int &= ~DC_HPD1_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD1\n");
+				}
+				break;
+			case 1:
+				if (rdev->irq.stat_regs.cik.disp_int_cont & DC_HPD2_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont &= ~DC_HPD2_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD2\n");
+				}
+				break;
+			case 2:
+				if (rdev->irq.stat_regs.cik.disp_int_cont2 & DC_HPD3_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont2 &= ~DC_HPD3_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD3\n");
+				}
+				break;
+			case 3:
+				if (rdev->irq.stat_regs.cik.disp_int_cont3 & DC_HPD4_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont3 &= ~DC_HPD4_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD4\n");
+				}
+				break;
+			case 4:
+				if (rdev->irq.stat_regs.cik.disp_int_cont4 & DC_HPD5_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont4 &= ~DC_HPD5_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD5\n");
+				}
+				break;
+			case 5:
+				if (rdev->irq.stat_regs.cik.disp_int_cont5 & DC_HPD6_INTERRUPT) {
+					rdev->irq.stat_regs.cik.disp_int_cont5 &= ~DC_HPD6_INTERRUPT;
+					queue_hotplug = true;
+					DRM_DEBUG("IH: HPD6\n");
+				}
+				break;
+			default:
+				DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+				break;
+			}
+			break;
+		case 146:
+		case 147:
+			addr = RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR);
+			status = RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS);
+			mc_client = RREG32(VM_CONTEXT1_PROTECTION_FAULT_MCCLIENT);
+			dev_err(rdev->dev, "GPU fault detected: %d 0x%08x\n", src_id, src_data);
+			dev_err(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x%08X\n",
+				addr);
+			dev_err(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
+				status);
+			cik_vm_decode_fault(rdev, status, addr, mc_client);
+			/* reset addr and status */
+			WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1);
+			break;
+		case 176: /* GFX RB CP_INT */
+		case 177: /* GFX IB CP_INT */
+			radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
+			break;
+		case 181: /* CP EOP event */
+			DRM_DEBUG("IH: CP EOP\n");
+			/* XXX check the bitfield order! */
+			me_id = (ring_id & 0x60) >> 5;
+			pipe_id = (ring_id & 0x18) >> 3;
+			queue_id = (ring_id & 0x7) >> 0;
+			switch (me_id) {
+			case 0:
+				radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
+				break;
+			case 1:
+			case 2:
+				if ((cp1_ring->me == me_id) & (cp1_ring->pipe == pipe_id))
+					radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP1_INDEX);
+				if ((cp2_ring->me == me_id) & (cp2_ring->pipe == pipe_id))
+					radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP2_INDEX);
+				break;
+			}
+			break;
+		case 184: /* CP Privileged reg access */
+			DRM_ERROR("Illegal register access in command stream\n");
+			/* XXX check the bitfield order! */
+			me_id = (ring_id & 0x60) >> 5;
+			pipe_id = (ring_id & 0x18) >> 3;
+			queue_id = (ring_id & 0x7) >> 0;
+			switch (me_id) {
+			case 0:
+				/* This results in a full GPU reset, but all we need to do is soft
+				 * reset the CP for gfx
+				 */
+				queue_reset = true;
+				break;
+			case 1:
+				/* XXX compute */
+				queue_reset = true;
+				break;
+			case 2:
+				/* XXX compute */
+				queue_reset = true;
+				break;
+			}
+			break;
+		case 185: /* CP Privileged inst */
+			DRM_ERROR("Illegal instruction in command stream\n");
+			/* XXX check the bitfield order! */
+			me_id = (ring_id & 0x60) >> 5;
+			pipe_id = (ring_id & 0x18) >> 3;
+			queue_id = (ring_id & 0x7) >> 0;
+			switch (me_id) {
+			case 0:
+				/* This results in a full GPU reset, but all we need to do is soft
+				 * reset the CP for gfx
+				 */
+				queue_reset = true;
+				break;
+			case 1:
+				/* XXX compute */
+				queue_reset = true;
+				break;
+			case 2:
+				/* XXX compute */
+				queue_reset = true;
+				break;
+			}
+			break;
+		case 224: /* SDMA trap event */
+			/* XXX check the bitfield order! */
+			me_id = (ring_id & 0x3) >> 0;
+			queue_id = (ring_id & 0xc) >> 2;
+			DRM_DEBUG("IH: SDMA trap\n");
+			switch (me_id) {
+			case 0:
+				switch (queue_id) {
+				case 0:
+					radeon_fence_process(rdev, R600_RING_TYPE_DMA_INDEX);
+					break;
+				case 1:
+					/* XXX compute */
+					break;
+				case 2:
+					/* XXX compute */
+					break;
+				}
+				break;
+			case 1:
+				switch (queue_id) {
+				case 0:
+					radeon_fence_process(rdev, CAYMAN_RING_TYPE_DMA1_INDEX);
+					break;
+				case 1:
+					/* XXX compute */
+					break;
+				case 2:
+					/* XXX compute */
+					break;
+				}
+				break;
+			}
+			break;
+		case 241: /* SDMA Privileged inst */
+		case 247: /* SDMA Privileged inst */
+			DRM_ERROR("Illegal instruction in SDMA command stream\n");
+			/* XXX check the bitfield order! */
+			me_id = (ring_id & 0x3) >> 0;
+			queue_id = (ring_id & 0xc) >> 2;
+			switch (me_id) {
+			case 0:
+				switch (queue_id) {
+				case 0:
+					queue_reset = true;
+					break;
+				case 1:
+					/* XXX compute */
+					queue_reset = true;
+					break;
+				case 2:
+					/* XXX compute */
+					queue_reset = true;
+					break;
+				}
+				break;
+			case 1:
+				switch (queue_id) {
+				case 0:
+					queue_reset = true;
+					break;
+				case 1:
+					/* XXX compute */
+					queue_reset = true;
+					break;
+				case 2:
+					/* XXX compute */
+					queue_reset = true;
+					break;
+				}
+				break;
+			}
+			break;
+		case 233: /* GUI IDLE */
+			DRM_DEBUG("IH: GUI idle\n");
+			break;
+		default:
+			DRM_DEBUG("Unhandled interrupt: %d %d\n", src_id, src_data);
+			break;
+		}
+
+		/* wptr/rptr are in bytes! */
+		rptr += 16;
+		rptr &= rdev->ih.ptr_mask;
+	}
+	if (queue_hotplug)
+		schedule_work(&rdev->hotplug_work);
+	if (queue_reset)
+		schedule_work(&rdev->reset_work);
+	rdev->ih.rptr = rptr;
+	WREG32(IH_RB_RPTR, rdev->ih.rptr);
+	atomic_set(&rdev->ih.lock, 0);
+
+	/* make sure wptr hasn't changed while processing */
+	wptr = cik_get_ih_wptr(rdev);
+	if (wptr != rptr)
+		goto restart_ih;
+
+	return IRQ_HANDLED;
+}
+
+/*
+ * startup/shutdown callbacks
+ */
+/**
+ * cik_startup - program the asic to a functional state
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Programs the asic to a functional state (CIK).
+ * Called by cik_init() and cik_resume().
+ * Returns 0 for success, error for failure.
+ */
+static int cik_startup(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring;
+	int r;
+
+	if (rdev->flags & RADEON_IS_IGP) {
+		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw ||
+		    !rdev->mec_fw || !rdev->sdma_fw || !rdev->rlc_fw) {
+			r = cik_init_microcode(rdev);
+			if (r) {
+				DRM_ERROR("Failed to load firmware!\n");
+				return r;
+			}
+		}
+	} else {
+		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw ||
+		    !rdev->mec_fw || !rdev->sdma_fw || !rdev->rlc_fw ||
+		    !rdev->mc_fw) {
+			r = cik_init_microcode(rdev);
+			if (r) {
+				DRM_ERROR("Failed to load firmware!\n");
+				return r;
+			}
+		}
+
+		r = ci_mc_load_microcode(rdev);
+		if (r) {
+			DRM_ERROR("Failed to load MC firmware!\n");
+			return r;
+		}
+	}
+
+	r = r600_vram_scratch_init(rdev);
+	if (r)
+		return r;
+
+	cik_mc_program(rdev);
+	r = cik_pcie_gart_enable(rdev);
+	if (r)
+		return r;
+	cik_gpu_init(rdev);
+
+	/* allocate rlc buffers */
+	r = si_rlc_init(rdev);
+	if (r) {
+		DRM_ERROR("Failed to init rlc BOs!\n");
+		return r;
+	}
+
+	/* allocate wb buffer */
+	r = radeon_wb_init(rdev);
+	if (r)
+		return r;
+
+	/* allocate mec buffers */
+	r = cik_mec_init(rdev);
+	if (r) {
+		DRM_ERROR("Failed to init MEC BOs!\n");
+		return r;
+	}
+
+	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
+		return r;
+	}
+
+	r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP1_INDEX);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
+		return r;
+	}
+
+	r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP2_INDEX);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
+		return r;
+	}
+
+	r = radeon_fence_driver_start_ring(rdev, R600_RING_TYPE_DMA_INDEX);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing DMA fences (%d).\n", r);
+		return r;
+	}
+
+	r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_DMA1_INDEX);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing DMA fences (%d).\n", r);
+		return r;
+	}
+
+	r = cik_uvd_resume(rdev);
+	if (!r) {
+		r = radeon_fence_driver_start_ring(rdev,
+						   R600_RING_TYPE_UVD_INDEX);
+		if (r)
+			dev_err(rdev->dev, "UVD fences init error (%d).\n", r);
+	}
+	if (r)
+		rdev->ring[R600_RING_TYPE_UVD_INDEX].ring_size = 0;
+
+	/* Enable IRQ */
+	if (!rdev->irq.installed) {
+		r = radeon_irq_kms_init(rdev);
+		if (r)
+			return r;
+	}
+
+	r = cik_irq_init(rdev);
+	if (r) {
+		DRM_ERROR("radeon: IH init failed (%d).\n", r);
+		radeon_irq_kms_fini(rdev);
+		return r;
+	}
+	cik_irq_set(rdev);
+
+	ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET,
+			     CP_RB0_RPTR, CP_RB0_WPTR,
+			     0, 0xfffff, RADEON_CP_PACKET2);
+	if (r)
+		return r;
+
+	/* set up the compute queues */
+	/* type-2 packets are deprecated on MEC, use type-3 instead */
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
+	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP1_RPTR_OFFSET,
+			     CP_HQD_PQ_RPTR, CP_HQD_PQ_WPTR,
+			     0, 0xfffff, PACKET3(PACKET3_NOP, 0x3FFF));
+	if (r)
+		return r;
+	ring->me = 1; /* first MEC */
+	ring->pipe = 0; /* first pipe */
+	ring->queue = 0; /* first queue */
+	ring->wptr_offs = CIK_WB_CP1_WPTR_OFFSET;
+
+	/* type-2 packets are deprecated on MEC, use type-3 instead */
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
+	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP2_RPTR_OFFSET,
+			     CP_HQD_PQ_RPTR, CP_HQD_PQ_WPTR,
+			     0, 0xffffffff, PACKET3(PACKET3_NOP, 0x3FFF));
+	if (r)
+		return r;
+	/* dGPU only have 1 MEC */
+	ring->me = 1; /* first MEC */
+	ring->pipe = 0; /* first pipe */
+	ring->queue = 1; /* second queue */
+	ring->wptr_offs = CIK_WB_CP2_WPTR_OFFSET;
+
+	ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX];
+	r = radeon_ring_init(rdev, ring, ring->ring_size, R600_WB_DMA_RPTR_OFFSET,
+			     SDMA0_GFX_RB_RPTR + SDMA0_REGISTER_OFFSET,
+			     SDMA0_GFX_RB_WPTR + SDMA0_REGISTER_OFFSET,
+			     2, 0xfffffffc, SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0));
+	if (r)
+		return r;
+
+	ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX];
+	r = radeon_ring_init(rdev, ring, ring->ring_size, CAYMAN_WB_DMA1_RPTR_OFFSET,
+			     SDMA0_GFX_RB_RPTR + SDMA1_REGISTER_OFFSET,
+			     SDMA0_GFX_RB_WPTR + SDMA1_REGISTER_OFFSET,
+			     2, 0xfffffffc, SDMA_PACKET(SDMA_OPCODE_NOP, 0, 0));
+	if (r)
+		return r;
+
+	r = cik_cp_resume(rdev);
+	if (r)
+		return r;
+
+	r = cik_sdma_resume(rdev);
+	if (r)
+		return r;
+
+	ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+	if (ring->ring_size) {
+		r = radeon_ring_init(rdev, ring, ring->ring_size,
+				     R600_WB_UVD_RPTR_OFFSET,
+				     UVD_RBC_RB_RPTR, UVD_RBC_RB_WPTR,
+				     0, 0xfffff, RADEON_CP_PACKET2);
+		if (!r)
+			r = r600_uvd_init(rdev);
+		if (r)
+			DRM_ERROR("radeon: failed initializing UVD (%d).\n", r);
+	}
+
+	r = radeon_ib_pool_init(rdev);
+	if (r) {
+		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
+		return r;
+	}
+
+	r = radeon_vm_manager_init(rdev);
+	if (r) {
+		dev_err(rdev->dev, "vm manager initialization failed (%d).\n", r);
+		return r;
+	}
+
+	return 0;
+}
+
+/**
+ * cik_resume - resume the asic to a functional state
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Programs the asic to a functional state (CIK).
+ * Called at resume.
+ * Returns 0 for success, error for failure.
+ */
+int cik_resume(struct radeon_device *rdev)
+{
+	int r;
+
+	/* post card */
+	atom_asic_init(rdev->mode_info.atom_context);
+
+	/* init golden registers */
+	cik_init_golden_registers(rdev);
+
+	rdev->accel_working = true;
+	r = cik_startup(rdev);
+	if (r) {
+		DRM_ERROR("cik startup failed on resume\n");
+		rdev->accel_working = false;
+		return r;
+	}
+
+	return r;
+
+}
+
+/**
+ * cik_suspend - suspend the asic
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Bring the chip into a state suitable for suspend (CIK).
+ * Called at suspend.
+ * Returns 0 for success.
+ */
+int cik_suspend(struct radeon_device *rdev)
+{
+	radeon_vm_manager_fini(rdev);
+	cik_cp_enable(rdev, false);
+	cik_sdma_enable(rdev, false);
+	r600_uvd_rbc_stop(rdev);
+	radeon_uvd_suspend(rdev);
+	cik_irq_suspend(rdev);
+	radeon_wb_disable(rdev);
+	cik_pcie_gart_disable(rdev);
+	return 0;
+}
+
+/* Plan is to move initialization in that function and use
+ * helper function so that radeon_device_init pretty much
+ * do nothing more than calling asic specific function. This
+ * should also allow to remove a bunch of callback function
+ * like vram_info.
+ */
+/**
+ * cik_init - asic specific driver and hw init
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Setup asic specific driver variables and program the hw
+ * to a functional state (CIK).
+ * Called at driver startup.
+ * Returns 0 for success, errors for failure.
+ */
+int cik_init(struct radeon_device *rdev)
+{
+	struct radeon_ring *ring;
+	int r;
+
+	/* Read BIOS */
+	if (!radeon_get_bios(rdev)) {
+		if (ASIC_IS_AVIVO(rdev))
+			return -EINVAL;
+	}
+	/* Must be an ATOMBIOS */
+	if (!rdev->is_atom_bios) {
+		dev_err(rdev->dev, "Expecting atombios for cayman GPU\n");
+		return -EINVAL;
+	}
+	r = radeon_atombios_init(rdev);
+	if (r)
+		return r;
+
+	/* Post card if necessary */
+	if (!radeon_card_posted(rdev)) {
+		if (!rdev->bios) {
+			dev_err(rdev->dev, "Card not posted and no BIOS - ignoring\n");
+			return -EINVAL;
+		}
+		DRM_INFO("GPU not posted. posting now...\n");
+		atom_asic_init(rdev->mode_info.atom_context);
+	}
+	/* init golden registers */
+	cik_init_golden_registers(rdev);
+	/* Initialize scratch registers */
+	cik_scratch_init(rdev);
+	/* Initialize surface registers */
+	radeon_surface_init(rdev);
+	/* Initialize clocks */
+	radeon_get_clock_info(rdev->ddev);
+
+	/* Fence driver */
+	r = radeon_fence_driver_init(rdev);
+	if (r)
+		return r;
+
+	/* initialize memory controller */
+	r = cik_mc_init(rdev);
+	if (r)
+		return r;
+	/* Memory manager */
+	r = radeon_bo_init(rdev);
+	if (r)
+		return r;
+
+	ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 1024 * 1024);
+
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 1024 * 1024);
+	r = radeon_doorbell_get(rdev, &ring->doorbell_page_num);
+	if (r)
+		return r;
+
+	ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 1024 * 1024);
+	r = radeon_doorbell_get(rdev, &ring->doorbell_page_num);
+	if (r)
+		return r;
+
+	ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 256 * 1024);
+
+	ring = &rdev->ring[CAYMAN_RING_TYPE_DMA1_INDEX];
+	ring->ring_obj = NULL;
+	r600_ring_init(rdev, ring, 256 * 1024);
+
+	r = radeon_uvd_init(rdev);
+	if (!r) {
+		ring = &rdev->ring[R600_RING_TYPE_UVD_INDEX];
+		ring->ring_obj = NULL;
+		r600_ring_init(rdev, ring, 4096);
+	}
+
+	rdev->ih.ring_obj = NULL;
+	r600_ih_ring_init(rdev, 64 * 1024);
+
+	r = r600_pcie_gart_init(rdev);
+	if (r)
+		return r;
+
+	rdev->accel_working = true;
+	r = cik_startup(rdev);
+	if (r) {
+		dev_err(rdev->dev, "disabling GPU acceleration\n");
+		cik_cp_fini(rdev);
+		cik_sdma_fini(rdev);
+		cik_irq_fini(rdev);
+		si_rlc_fini(rdev);
+		cik_mec_fini(rdev);
+		radeon_wb_fini(rdev);
+		radeon_ib_pool_fini(rdev);
+		radeon_vm_manager_fini(rdev);
+		radeon_irq_kms_fini(rdev);
+		cik_pcie_gart_fini(rdev);
+		rdev->accel_working = false;
+	}
+
+	/* Don't start up if the MC ucode is missing.
+	 * The default clocks and voltages before the MC ucode
+	 * is loaded are not suffient for advanced operations.
+	 */
+	if (!rdev->mc_fw && !(rdev->flags & RADEON_IS_IGP)) {
+		DRM_ERROR("radeon: MC ucode required for NI+.\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+/**
+ * cik_fini - asic specific driver and hw fini
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Tear down the asic specific driver variables and program the hw
+ * to an idle state (CIK).
+ * Called at driver unload.
+ */
+void cik_fini(struct radeon_device *rdev)
+{
+	cik_cp_fini(rdev);
+	cik_sdma_fini(rdev);
+	cik_irq_fini(rdev);
+	si_rlc_fini(rdev);
+	cik_mec_fini(rdev);
+	radeon_wb_fini(rdev);
+	radeon_vm_manager_fini(rdev);
+	radeon_ib_pool_fini(rdev);
+	radeon_irq_kms_fini(rdev);
+	radeon_uvd_fini(rdev);
+	cik_pcie_gart_fini(rdev);
+	r600_vram_scratch_fini(rdev);
+	radeon_gem_fini(rdev);
+	radeon_fence_driver_fini(rdev);
+	radeon_bo_fini(rdev);
+	radeon_atombios_fini(rdev);
+	kfree(rdev->bios);
+	rdev->bios = NULL;
+}
+
+/* display watermark setup */
+/**
+ * dce8_line_buffer_adjust - Set up the line buffer
+ *
+ * @rdev: radeon_device pointer
+ * @radeon_crtc: the selected display controller
+ * @mode: the current display mode on the selected display
+ * controller
+ *
+ * Setup up the line buffer allocation for
+ * the selected display controller (CIK).
+ * Returns the line buffer size in pixels.
+ */
+static u32 dce8_line_buffer_adjust(struct radeon_device *rdev,
+				   struct radeon_crtc *radeon_crtc,
+				   struct drm_display_mode *mode)
+{
+	u32 tmp;
+
+	/*
+	 * Line Buffer Setup
+	 * There are 6 line buffers, one for each display controllers.
+	 * There are 3 partitions per LB. Select the number of partitions
+	 * to enable based on the display width.  For display widths larger
+	 * than 4096, you need use to use 2 display controllers and combine
+	 * them using the stereo blender.
+	 */
+	if (radeon_crtc->base.enabled && mode) {
+		if (mode->crtc_hdisplay < 1920)
+			tmp = 1;
+		else if (mode->crtc_hdisplay < 2560)
+			tmp = 2;
+		else if (mode->crtc_hdisplay < 4096)
+			tmp = 0;
+		else {
+			DRM_DEBUG_KMS("Mode too big for LB!\n");
+			tmp = 0;
+		}
+	} else
+		tmp = 1;
+
+	WREG32(LB_MEMORY_CTRL + radeon_crtc->crtc_offset,
+	       LB_MEMORY_CONFIG(tmp) | LB_MEMORY_SIZE(0x6B0));
+
+	if (radeon_crtc->base.enabled && mode) {
+		switch (tmp) {
+		case 0:
+		default:
+			return 4096 * 2;
+		case 1:
+			return 1920 * 2;
+		case 2:
+			return 2560 * 2;
+		}
+	}
+
+	/* controller not enabled, so no lb used */
+	return 0;
+}
+
+/**
+ * cik_get_number_of_dram_channels - get the number of dram channels
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Look up the number of video ram channels (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the number of dram channels
+ */
+static u32 cik_get_number_of_dram_channels(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(MC_SHARED_CHMAP);
+
+	switch ((tmp & NOOFCHAN_MASK) >> NOOFCHAN_SHIFT) {
+	case 0:
+	default:
+		return 1;
+	case 1:
+		return 2;
+	case 2:
+		return 4;
+	case 3:
+		return 8;
+	case 4:
+		return 3;
+	case 5:
+		return 6;
+	case 6:
+		return 10;
+	case 7:
+		return 12;
+	case 8:
+		return 16;
+	}
+}
+
+struct dce8_wm_params {
+	u32 dram_channels; /* number of dram channels */
+	u32 yclk;          /* bandwidth per dram data pin in kHz */
+	u32 sclk;          /* engine clock in kHz */
+	u32 disp_clk;      /* display clock in kHz */
+	u32 src_width;     /* viewport width */
+	u32 active_time;   /* active display time in ns */
+	u32 blank_time;    /* blank time in ns */
+	bool interlaced;    /* mode is interlaced */
+	fixed20_12 vsc;    /* vertical scale ratio */
+	u32 num_heads;     /* number of active crtcs */
+	u32 bytes_per_pixel; /* bytes per pixel display + overlay */
+	u32 lb_size;       /* line buffer allocated to pipe */
+	u32 vtaps;         /* vertical scaler taps */
+};
+
+/**
+ * dce8_dram_bandwidth - get the dram bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the raw dram bandwidth (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the dram bandwidth in MBytes/s
+ */
+static u32 dce8_dram_bandwidth(struct dce8_wm_params *wm)
+{
+	/* Calculate raw DRAM Bandwidth */
+	fixed20_12 dram_efficiency; /* 0.7 */
+	fixed20_12 yclk, dram_channels, bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	yclk.full = dfixed_const(wm->yclk);
+	yclk.full = dfixed_div(yclk, a);
+	dram_channels.full = dfixed_const(wm->dram_channels * 4);
+	a.full = dfixed_const(10);
+	dram_efficiency.full = dfixed_const(7);
+	dram_efficiency.full = dfixed_div(dram_efficiency, a);
+	bandwidth.full = dfixed_mul(dram_channels, yclk);
+	bandwidth.full = dfixed_mul(bandwidth, dram_efficiency);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce8_dram_bandwidth_for_display - get the dram bandwidth for display
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the dram bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the dram bandwidth for display in MBytes/s
+ */
+static u32 dce8_dram_bandwidth_for_display(struct dce8_wm_params *wm)
+{
+	/* Calculate DRAM Bandwidth and the part allocated to display. */
+	fixed20_12 disp_dram_allocation; /* 0.3 to 0.7 */
+	fixed20_12 yclk, dram_channels, bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	yclk.full = dfixed_const(wm->yclk);
+	yclk.full = dfixed_div(yclk, a);
+	dram_channels.full = dfixed_const(wm->dram_channels * 4);
+	a.full = dfixed_const(10);
+	disp_dram_allocation.full = dfixed_const(3); /* XXX worse case value 0.3 */
+	disp_dram_allocation.full = dfixed_div(disp_dram_allocation, a);
+	bandwidth.full = dfixed_mul(dram_channels, yclk);
+	bandwidth.full = dfixed_mul(bandwidth, disp_dram_allocation);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce8_data_return_bandwidth - get the data return bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the data return bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the data return bandwidth in MBytes/s
+ */
+static u32 dce8_data_return_bandwidth(struct dce8_wm_params *wm)
+{
+	/* Calculate the display Data return Bandwidth */
+	fixed20_12 return_efficiency; /* 0.8 */
+	fixed20_12 sclk, bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	sclk.full = dfixed_const(wm->sclk);
+	sclk.full = dfixed_div(sclk, a);
+	a.full = dfixed_const(10);
+	return_efficiency.full = dfixed_const(8);
+	return_efficiency.full = dfixed_div(return_efficiency, a);
+	a.full = dfixed_const(32);
+	bandwidth.full = dfixed_mul(a, sclk);
+	bandwidth.full = dfixed_mul(bandwidth, return_efficiency);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce8_dmif_request_bandwidth - get the dmif bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the dmif bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the dmif bandwidth in MBytes/s
+ */
+static u32 dce8_dmif_request_bandwidth(struct dce8_wm_params *wm)
+{
+	/* Calculate the DMIF Request Bandwidth */
+	fixed20_12 disp_clk_request_efficiency; /* 0.8 */
+	fixed20_12 disp_clk, bandwidth;
+	fixed20_12 a, b;
+
+	a.full = dfixed_const(1000);
+	disp_clk.full = dfixed_const(wm->disp_clk);
+	disp_clk.full = dfixed_div(disp_clk, a);
+	a.full = dfixed_const(32);
+	b.full = dfixed_mul(a, disp_clk);
+
+	a.full = dfixed_const(10);
+	disp_clk_request_efficiency.full = dfixed_const(8);
+	disp_clk_request_efficiency.full = dfixed_div(disp_clk_request_efficiency, a);
+
+	bandwidth.full = dfixed_mul(b, disp_clk_request_efficiency);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce8_available_bandwidth - get the min available bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the min available bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the min available bandwidth in MBytes/s
+ */
+static u32 dce8_available_bandwidth(struct dce8_wm_params *wm)
+{
+	/* Calculate the Available bandwidth. Display can use this temporarily but not in average. */
+	u32 dram_bandwidth = dce8_dram_bandwidth(wm);
+	u32 data_return_bandwidth = dce8_data_return_bandwidth(wm);
+	u32 dmif_req_bandwidth = dce8_dmif_request_bandwidth(wm);
+
+	return min(dram_bandwidth, min(data_return_bandwidth, dmif_req_bandwidth));
+}
+
+/**
+ * dce8_average_bandwidth - get the average available bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the average available bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the average available bandwidth in MBytes/s
+ */
+static u32 dce8_average_bandwidth(struct dce8_wm_params *wm)
+{
+	/* Calculate the display mode Average Bandwidth
+	 * DisplayMode should contain the source and destination dimensions,
+	 * timing, etc.
+	 */
+	fixed20_12 bpp;
+	fixed20_12 line_time;
+	fixed20_12 src_width;
+	fixed20_12 bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	line_time.full = dfixed_const(wm->active_time + wm->blank_time);
+	line_time.full = dfixed_div(line_time, a);
+	bpp.full = dfixed_const(wm->bytes_per_pixel);
+	src_width.full = dfixed_const(wm->src_width);
+	bandwidth.full = dfixed_mul(src_width, bpp);
+	bandwidth.full = dfixed_mul(bandwidth, wm->vsc);
+	bandwidth.full = dfixed_div(bandwidth, line_time);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce8_latency_watermark - get the latency watermark
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the latency watermark (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the latency watermark in ns
+ */
+static u32 dce8_latency_watermark(struct dce8_wm_params *wm)
+{
+	/* First calculate the latency in ns */
+	u32 mc_latency = 2000; /* 2000 ns. */
+	u32 available_bandwidth = dce8_available_bandwidth(wm);
+	u32 worst_chunk_return_time = (512 * 8 * 1000) / available_bandwidth;
+	u32 cursor_line_pair_return_time = (128 * 4 * 1000) / available_bandwidth;
+	u32 dc_latency = 40000000 / wm->disp_clk; /* dc pipe latency */
+	u32 other_heads_data_return_time = ((wm->num_heads + 1) * worst_chunk_return_time) +
+		(wm->num_heads * cursor_line_pair_return_time);
+	u32 latency = mc_latency + other_heads_data_return_time + dc_latency;
+	u32 max_src_lines_per_dst_line, lb_fill_bw, line_fill_time;
+	u32 tmp, dmif_size = 12288;
+	fixed20_12 a, b, c;
+
+	if (wm->num_heads == 0)
+		return 0;
+
+	a.full = dfixed_const(2);
+	b.full = dfixed_const(1);
+	if ((wm->vsc.full > a.full) ||
+	    ((wm->vsc.full > b.full) && (wm->vtaps >= 3)) ||
+	    (wm->vtaps >= 5) ||
+	    ((wm->vsc.full >= a.full) && wm->interlaced))
+		max_src_lines_per_dst_line = 4;
+	else
+		max_src_lines_per_dst_line = 2;
+
+	a.full = dfixed_const(available_bandwidth);
+	b.full = dfixed_const(wm->num_heads);
+	a.full = dfixed_div(a, b);
+
+	b.full = dfixed_const(mc_latency + 512);
+	c.full = dfixed_const(wm->disp_clk);
+	b.full = dfixed_div(b, c);
+
+	c.full = dfixed_const(dmif_size);
+	b.full = dfixed_div(c, b);
+
+	tmp = min(dfixed_trunc(a), dfixed_trunc(b));
+
+	b.full = dfixed_const(1000);
+	c.full = dfixed_const(wm->disp_clk);
+	b.full = dfixed_div(c, b);
+	c.full = dfixed_const(wm->bytes_per_pixel);
+	b.full = dfixed_mul(b, c);
+
+	lb_fill_bw = min(tmp, dfixed_trunc(b));
+
+	a.full = dfixed_const(max_src_lines_per_dst_line * wm->src_width * wm->bytes_per_pixel);
+	b.full = dfixed_const(1000);
+	c.full = dfixed_const(lb_fill_bw);
+	b.full = dfixed_div(c, b);
+	a.full = dfixed_div(a, b);
+	line_fill_time = dfixed_trunc(a);
+
+	if (line_fill_time < wm->active_time)
+		return latency;
+	else
+		return latency + (line_fill_time - wm->active_time);
+
+}
+
+/**
+ * dce8_average_bandwidth_vs_dram_bandwidth_for_display - check
+ * average and available dram bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Check if the display average bandwidth fits in the display
+ * dram bandwidth (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns true if the display fits, false if not.
+ */
+static bool dce8_average_bandwidth_vs_dram_bandwidth_for_display(struct dce8_wm_params *wm)
+{
+	if (dce8_average_bandwidth(wm) <=
+	    (dce8_dram_bandwidth_for_display(wm) / wm->num_heads))
+		return true;
+	else
+		return false;
+}
+
+/**
+ * dce8_average_bandwidth_vs_available_bandwidth - check
+ * average and available bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Check if the display average bandwidth fits in the display
+ * available bandwidth (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns true if the display fits, false if not.
+ */
+static bool dce8_average_bandwidth_vs_available_bandwidth(struct dce8_wm_params *wm)
+{
+	if (dce8_average_bandwidth(wm) <=
+	    (dce8_available_bandwidth(wm) / wm->num_heads))
+		return true;
+	else
+		return false;
+}
+
+/**
+ * dce8_check_latency_hiding - check latency hiding
+ *
+ * @wm: watermark calculation data
+ *
+ * Check latency hiding (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns true if the display fits, false if not.
+ */
+static bool dce8_check_latency_hiding(struct dce8_wm_params *wm)
+{
+	u32 lb_partitions = wm->lb_size / wm->src_width;
+	u32 line_time = wm->active_time + wm->blank_time;
+	u32 latency_tolerant_lines;
+	u32 latency_hiding;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1);
+	if (wm->vsc.full > a.full)
+		latency_tolerant_lines = 1;
+	else {
+		if (lb_partitions <= (wm->vtaps + 1))
+			latency_tolerant_lines = 1;
+		else
+			latency_tolerant_lines = 2;
+	}
+
+	latency_hiding = (latency_tolerant_lines * line_time + wm->blank_time);
+
+	if (dce8_latency_watermark(wm) <= latency_hiding)
+		return true;
+	else
+		return false;
+}
+
+/**
+ * dce8_program_watermarks - program display watermarks
+ *
+ * @rdev: radeon_device pointer
+ * @radeon_crtc: the selected display controller
+ * @lb_size: line buffer size
+ * @num_heads: number of display controllers in use
+ *
+ * Calculate and program the display watermarks for the
+ * selected display controller (CIK).
+ */
+static void dce8_program_watermarks(struct radeon_device *rdev,
+				    struct radeon_crtc *radeon_crtc,
+				    u32 lb_size, u32 num_heads)
+{
+	struct drm_display_mode *mode = &radeon_crtc->base.mode;
+	struct dce8_wm_params wm;
+	u32 pixel_period;
+	u32 line_time = 0;
+	u32 latency_watermark_a = 0, latency_watermark_b = 0;
+	u32 tmp, wm_mask;
+
+	if (radeon_crtc->base.enabled && num_heads && mode) {
+		pixel_period = 1000000 / (u32)mode->clock;
+		line_time = min((u32)mode->crtc_htotal * pixel_period, (u32)65535);
+
+		wm.yclk = rdev->pm.current_mclk * 10;
+		wm.sclk = rdev->pm.current_sclk * 10;
+		wm.disp_clk = mode->clock;
+		wm.src_width = mode->crtc_hdisplay;
+		wm.active_time = mode->crtc_hdisplay * pixel_period;
+		wm.blank_time = line_time - wm.active_time;
+		wm.interlaced = false;
+		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+			wm.interlaced = true;
+		wm.vsc = radeon_crtc->vsc;
+		wm.vtaps = 1;
+		if (radeon_crtc->rmx_type != RMX_OFF)
+			wm.vtaps = 2;
+		wm.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm.lb_size = lb_size;
+		wm.dram_channels = cik_get_number_of_dram_channels(rdev);
+		wm.num_heads = num_heads;
+
+		/* set for high clocks */
+		latency_watermark_a = min(dce8_latency_watermark(&wm), (u32)65535);
+		/* set for low clocks */
+		/* wm.yclk = low clk; wm.sclk = low clk */
+		latency_watermark_b = min(dce8_latency_watermark(&wm), (u32)65535);
+
+		/* possibly force display priority to high */
+		/* should really do this at mode validation time... */
+		if (!dce8_average_bandwidth_vs_dram_bandwidth_for_display(&wm) ||
+		    !dce8_average_bandwidth_vs_available_bandwidth(&wm) ||
+		    !dce8_check_latency_hiding(&wm) ||
+		    (rdev->disp_priority == 2)) {
+			DRM_DEBUG_KMS("force priority to high\n");
+		}
+	}
+
+	/* select wm A */
+	wm_mask = RREG32(DPG_WATERMARK_MASK_CONTROL + radeon_crtc->crtc_offset);
+	tmp = wm_mask;
+	tmp &= ~LATENCY_WATERMARK_MASK(3);
+	tmp |= LATENCY_WATERMARK_MASK(1);
+	WREG32(DPG_WATERMARK_MASK_CONTROL + radeon_crtc->crtc_offset, tmp);
+	WREG32(DPG_PIPE_LATENCY_CONTROL + radeon_crtc->crtc_offset,
+	       (LATENCY_LOW_WATERMARK(latency_watermark_a) |
+		LATENCY_HIGH_WATERMARK(line_time)));
+	/* select wm B */
+	tmp = RREG32(DPG_WATERMARK_MASK_CONTROL + radeon_crtc->crtc_offset);
+	tmp &= ~LATENCY_WATERMARK_MASK(3);
+	tmp |= LATENCY_WATERMARK_MASK(2);
+	WREG32(DPG_WATERMARK_MASK_CONTROL + radeon_crtc->crtc_offset, tmp);
+	WREG32(DPG_PIPE_LATENCY_CONTROL + radeon_crtc->crtc_offset,
+	       (LATENCY_LOW_WATERMARK(latency_watermark_b) |
+		LATENCY_HIGH_WATERMARK(line_time)));
+	/* restore original selection */
+	WREG32(DPG_WATERMARK_MASK_CONTROL + radeon_crtc->crtc_offset, wm_mask);
+}
+
+/**
+ * dce8_bandwidth_update - program display watermarks
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Calculate and program the display watermarks and line
+ * buffer allocation (CIK).
+ */
+void dce8_bandwidth_update(struct radeon_device *rdev)
+{
+	struct drm_display_mode *mode = NULL;
+	u32 num_heads = 0, lb_size;
+	int i;
+
+	radeon_update_display_priority(rdev);
+
+	for (i = 0; i < rdev->num_crtc; i++) {
+		if (rdev->mode_info.crtcs[i]->base.enabled)
+			num_heads++;
+	}
+	for (i = 0; i < rdev->num_crtc; i++) {
+		mode = &rdev->mode_info.crtcs[i]->base.mode;
+		lb_size = dce8_line_buffer_adjust(rdev, rdev->mode_info.crtcs[i], mode);
+		dce8_program_watermarks(rdev, rdev->mode_info.crtcs[i], lb_size, num_heads);
+	}
+}
+
+/**
+ * cik_get_gpu_clock_counter - return GPU clock counter snapshot
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Fetches a GPU clock counter snapshot (SI).
+ * Returns the 64 bit clock counter snapshot.
+ */
+uint64_t cik_get_gpu_clock_counter(struct radeon_device *rdev)
+{
+	uint64_t clock;
+
+	mutex_lock(&rdev->gpu_clock_mutex);
+	WREG32(RLC_CAPTURE_GPU_CLOCK_COUNT, 1);
+	clock = (uint64_t)RREG32(RLC_GPU_CLOCK_COUNT_LSB) |
+	        ((uint64_t)RREG32(RLC_GPU_CLOCK_COUNT_MSB) << 32ULL);
+	mutex_unlock(&rdev->gpu_clock_mutex);
+	return clock;
+}
+
+static int cik_set_uvd_clock(struct radeon_device *rdev, u32 clock,
+                              u32 cntl_reg, u32 status_reg)
+{
+	int r, i;
+	struct atom_clock_dividers dividers;
+	uint32_t tmp;
+
+	r = radeon_atom_get_clock_dividers(rdev, COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK,
+					   clock, false, &dividers);
+	if (r)
+		return r;
+
+	tmp = RREG32_SMC(cntl_reg);
+	tmp &= ~(DCLK_DIR_CNTL_EN|DCLK_DIVIDER_MASK);
+	tmp |= dividers.post_divider;
+	WREG32_SMC(cntl_reg, tmp);
+
+	for (i = 0; i < 100; i++) {
+		if (RREG32_SMC(status_reg) & DCLK_STATUS)
+			break;
+		mdelay(10);
+	}
+	if (i == 100)
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
+int cik_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk)
+{
+	int r = 0;
+
+	r = cik_set_uvd_clock(rdev, vclk, CG_VCLK_CNTL, CG_VCLK_STATUS);
+	if (r)
+		return r;
+
+	r = cik_set_uvd_clock(rdev, dclk, CG_DCLK_CNTL, CG_DCLK_STATUS);
+	return r;
+}
+
+int cik_uvd_resume(struct radeon_device *rdev)
+{
+	uint64_t addr;
+	uint32_t size;
+	int r;
+
+	r = radeon_uvd_resume(rdev);
+	if (r)
+		return r;
+
+	/* programm the VCPU memory controller bits 0-27 */
+	addr = rdev->uvd.gpu_addr >> 3;
+	size = RADEON_GPU_PAGE_ALIGN(rdev->uvd.fw_size + 4) >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET0, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE0, size);
+
+	addr += size;
+	size = RADEON_UVD_STACK_SIZE >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET1, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE1, size);
+
+	addr += size;
+	size = RADEON_UVD_HEAP_SIZE >> 3;
+	WREG32(UVD_VCPU_CACHE_OFFSET2, addr);
+	WREG32(UVD_VCPU_CACHE_SIZE2, size);
+
+	/* bits 28-31 */
+	addr = (rdev->uvd.gpu_addr >> 28) & 0xF;
+	WREG32(UVD_LMI_ADDR_EXT, (addr << 12) | (addr << 0));
+
+	/* bits 32-39 */
+	addr = (rdev->uvd.gpu_addr >> 32) & 0xFF;
+	WREG32(UVD_LMI_EXT40_ADDR, addr | (0x9 << 16) | (0x1 << 31));
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/radeon/cik_blit_shaders.c b/drivers/gpu/drm/radeon/cik_blit_shaders.c
new file mode 100644
index 000000000000..ff1311806e91
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cik_blit_shaders.c
@@ -0,0 +1,246 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ * Authors:
+ *     Alex Deucher <alexander.deucher@amd.com>
+ */
+
+#include <linux/types.h>
+#include <linux/bug.h>
+#include <linux/kernel.h>
+
+const u32 cik_default_state[] =
+{
+	0xc0066900,
+	0x00000000,
+	0x00000060, /* DB_RENDER_CONTROL */
+	0x00000000, /* DB_COUNT_CONTROL */
+	0x00000000, /* DB_DEPTH_VIEW */
+	0x0000002a, /* DB_RENDER_OVERRIDE */
+	0x00000000, /* DB_RENDER_OVERRIDE2 */
+	0x00000000, /* DB_HTILE_DATA_BASE */
+
+	0xc0046900,
+	0x00000008,
+	0x00000000, /* DB_DEPTH_BOUNDS_MIN */
+	0x00000000, /* DB_DEPTH_BOUNDS_MAX */
+	0x00000000, /* DB_STENCIL_CLEAR */
+	0x00000000, /* DB_DEPTH_CLEAR */
+
+	0xc0036900,
+	0x0000000f,
+	0x00000000, /* DB_DEPTH_INFO */
+	0x00000000, /* DB_Z_INFO */
+	0x00000000, /* DB_STENCIL_INFO */
+
+	0xc0016900,
+	0x00000080,
+	0x00000000, /* PA_SC_WINDOW_OFFSET */
+
+	0xc00d6900,
+	0x00000083,
+	0x0000ffff, /* PA_SC_CLIPRECT_RULE */
+	0x00000000, /* PA_SC_CLIPRECT_0_TL */
+	0x20002000, /* PA_SC_CLIPRECT_0_BR */
+	0x00000000,
+	0x20002000,
+	0x00000000,
+	0x20002000,
+	0x00000000,
+	0x20002000,
+	0xaaaaaaaa, /* PA_SC_EDGERULE */
+	0x00000000, /* PA_SU_HARDWARE_SCREEN_OFFSET */
+	0x0000000f, /* CB_TARGET_MASK */
+	0x0000000f, /* CB_SHADER_MASK */
+
+	0xc0226900,
+	0x00000094,
+	0x80000000, /* PA_SC_VPORT_SCISSOR_0_TL */
+	0x20002000, /* PA_SC_VPORT_SCISSOR_0_BR */
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x80000000,
+	0x20002000,
+	0x00000000, /* PA_SC_VPORT_ZMIN_0 */
+	0x3f800000, /* PA_SC_VPORT_ZMAX_0 */
+
+	0xc0046900,
+	0x00000100,
+	0xffffffff, /* VGT_MAX_VTX_INDX */
+	0x00000000, /* VGT_MIN_VTX_INDX */
+	0x00000000, /* VGT_INDX_OFFSET */
+	0x00000000, /* VGT_MULTI_PRIM_IB_RESET_INDX */
+
+	0xc0046900,
+	0x00000105,
+	0x00000000, /* CB_BLEND_RED */
+	0x00000000, /* CB_BLEND_GREEN */
+	0x00000000, /* CB_BLEND_BLUE */
+	0x00000000, /* CB_BLEND_ALPHA */
+
+	0xc0016900,
+	0x000001e0,
+	0x00000000, /* CB_BLEND0_CONTROL */
+
+	0xc00c6900,
+	0x00000200,
+	0x00000000, /* DB_DEPTH_CONTROL */
+	0x00000000, /* DB_EQAA */
+	0x00cc0010, /* CB_COLOR_CONTROL */
+	0x00000210, /* DB_SHADER_CONTROL */
+	0x00010000, /* PA_CL_CLIP_CNTL */
+	0x00000004, /* PA_SU_SC_MODE_CNTL */
+	0x00000100, /* PA_CL_VTE_CNTL */
+	0x00000000, /* PA_CL_VS_OUT_CNTL */
+	0x00000000, /* PA_CL_NANINF_CNTL */
+	0x00000000, /* PA_SU_LINE_STIPPLE_CNTL */
+	0x00000000, /* PA_SU_LINE_STIPPLE_SCALE */
+	0x00000000, /* PA_SU_PRIM_FILTER_CNTL */
+
+	0xc0116900,
+	0x00000280,
+	0x00000000, /* PA_SU_POINT_SIZE */
+	0x00000000, /* PA_SU_POINT_MINMAX */
+	0x00000008, /* PA_SU_LINE_CNTL */
+	0x00000000, /* PA_SC_LINE_STIPPLE */
+	0x00000000, /* VGT_OUTPUT_PATH_CNTL */
+	0x00000000, /* VGT_HOS_CNTL */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000, /* VGT_GS_MODE */
+
+	0xc0026900,
+	0x00000292,
+	0x00000000, /* PA_SC_MODE_CNTL_0 */
+	0x00000000, /* PA_SC_MODE_CNTL_1 */
+
+	0xc0016900,
+	0x000002a1,
+	0x00000000, /* VGT_PRIMITIVEID_EN */
+
+	0xc0016900,
+	0x000002a5,
+	0x00000000, /* VGT_MULTI_PRIM_IB_RESET_EN */
+
+	0xc0026900,
+	0x000002a8,
+	0x00000000, /* VGT_INSTANCE_STEP_RATE_0 */
+	0x00000000,
+
+	0xc0026900,
+	0x000002ad,
+	0x00000000, /* VGT_REUSE_OFF */
+	0x00000000,
+
+	0xc0016900,
+	0x000002d5,
+	0x00000000, /* VGT_SHADER_STAGES_EN */
+
+	0xc0016900,
+	0x000002dc,
+	0x0000aa00, /* DB_ALPHA_TO_MASK */
+
+	0xc0066900,
+	0x000002de,
+	0x00000000, /* PA_SU_POLY_OFFSET_DB_FMT_CNTL */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+
+	0xc0026900,
+	0x000002e5,
+	0x00000000, /* VGT_STRMOUT_CONFIG */
+	0x00000000,
+
+	0xc01b6900,
+	0x000002f5,
+	0x76543210, /* PA_SC_CENTROID_PRIORITY_0 */
+	0xfedcba98, /* PA_SC_CENTROID_PRIORITY_1 */
+	0x00000000, /* PA_SC_LINE_CNTL */
+	0x00000000, /* PA_SC_AA_CONFIG */
+	0x00000005, /* PA_SU_VTX_CNTL */
+	0x3f800000, /* PA_CL_GB_VERT_CLIP_ADJ */
+	0x3f800000, /* PA_CL_GB_VERT_DISC_ADJ */
+	0x3f800000, /* PA_CL_GB_HORZ_CLIP_ADJ */
+	0x3f800000, /* PA_CL_GB_HORZ_DISC_ADJ */
+	0x00000000, /* PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0 */
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0x00000000,
+	0xffffffff, /* PA_SC_AA_MASK_X0Y0_X1Y0 */
+	0xffffffff,
+
+	0xc0026900,
+	0x00000316,
+	0x0000000e, /* VGT_VERTEX_REUSE_BLOCK_CNTL */
+	0x00000010, /*  */
+};
+
+const u32 cik_default_size = ARRAY_SIZE(cik_default_state);
diff --git a/drivers/gpu/drm/radeon/cik_blit_shaders.h b/drivers/gpu/drm/radeon/cik_blit_shaders.h
new file mode 100644
index 000000000000..dfe7314f9ff4
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cik_blit_shaders.h
@@ -0,0 +1,32 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef CIK_BLIT_SHADERS_H
+#define CIK_BLIT_SHADERS_H
+
+extern const u32 cik_default_state[];
+
+extern const u32 cik_default_size;
+
+#endif
diff --git a/drivers/gpu/drm/radeon/cik_reg.h b/drivers/gpu/drm/radeon/cik_reg.h
new file mode 100644
index 000000000000..d71e46d571f5
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cik_reg.h
@@ -0,0 +1,147 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+#ifndef __CIK_REG_H__
+#define __CIK_REG_H__
+
+#define CIK_DC_GPIO_HPD_MASK                      0x65b0
+#define CIK_DC_GPIO_HPD_A                         0x65b4
+#define CIK_DC_GPIO_HPD_EN                        0x65b8
+#define CIK_DC_GPIO_HPD_Y                         0x65bc
+
+#define CIK_GRPH_CONTROL                          0x6804
+#       define CIK_GRPH_DEPTH(x)                  (((x) & 0x3) << 0)
+#       define CIK_GRPH_DEPTH_8BPP                0
+#       define CIK_GRPH_DEPTH_16BPP               1
+#       define CIK_GRPH_DEPTH_32BPP               2
+#       define CIK_GRPH_NUM_BANKS(x)              (((x) & 0x3) << 2)
+#       define CIK_ADDR_SURF_2_BANK               0
+#       define CIK_ADDR_SURF_4_BANK               1
+#       define CIK_ADDR_SURF_8_BANK               2
+#       define CIK_ADDR_SURF_16_BANK              3
+#       define CIK_GRPH_Z(x)                      (((x) & 0x3) << 4)
+#       define CIK_GRPH_BANK_WIDTH(x)             (((x) & 0x3) << 6)
+#       define CIK_ADDR_SURF_BANK_WIDTH_1         0
+#       define CIK_ADDR_SURF_BANK_WIDTH_2         1
+#       define CIK_ADDR_SURF_BANK_WIDTH_4         2
+#       define CIK_ADDR_SURF_BANK_WIDTH_8         3
+#       define CIK_GRPH_FORMAT(x)                 (((x) & 0x7) << 8)
+/* 8 BPP */
+#       define CIK_GRPH_FORMAT_INDEXED            0
+/* 16 BPP */
+#       define CIK_GRPH_FORMAT_ARGB1555           0
+#       define CIK_GRPH_FORMAT_ARGB565            1
+#       define CIK_GRPH_FORMAT_ARGB4444           2
+#       define CIK_GRPH_FORMAT_AI88               3
+#       define CIK_GRPH_FORMAT_MONO16             4
+#       define CIK_GRPH_FORMAT_BGRA5551           5
+/* 32 BPP */
+#       define CIK_GRPH_FORMAT_ARGB8888           0
+#       define CIK_GRPH_FORMAT_ARGB2101010        1
+#       define CIK_GRPH_FORMAT_32BPP_DIG          2
+#       define CIK_GRPH_FORMAT_8B_ARGB2101010     3
+#       define CIK_GRPH_FORMAT_BGRA1010102        4
+#       define CIK_GRPH_FORMAT_8B_BGRA1010102     5
+#       define CIK_GRPH_FORMAT_RGB111110          6
+#       define CIK_GRPH_FORMAT_BGR101111          7
+#       define CIK_GRPH_BANK_HEIGHT(x)            (((x) & 0x3) << 11)
+#       define CIK_ADDR_SURF_BANK_HEIGHT_1        0
+#       define CIK_ADDR_SURF_BANK_HEIGHT_2        1
+#       define CIK_ADDR_SURF_BANK_HEIGHT_4        2
+#       define CIK_ADDR_SURF_BANK_HEIGHT_8        3
+#       define CIK_GRPH_TILE_SPLIT(x)             (((x) & 0x7) << 13)
+#       define CIK_ADDR_SURF_TILE_SPLIT_64B       0
+#       define CIK_ADDR_SURF_TILE_SPLIT_128B      1
+#       define CIK_ADDR_SURF_TILE_SPLIT_256B      2
+#       define CIK_ADDR_SURF_TILE_SPLIT_512B      3
+#       define CIK_ADDR_SURF_TILE_SPLIT_1KB       4
+#       define CIK_ADDR_SURF_TILE_SPLIT_2KB       5
+#       define CIK_ADDR_SURF_TILE_SPLIT_4KB       6
+#       define CIK_GRPH_MACRO_TILE_ASPECT(x)      (((x) & 0x3) << 18)
+#       define CIK_ADDR_SURF_MACRO_TILE_ASPECT_1  0
+#       define CIK_ADDR_SURF_MACRO_TILE_ASPECT_2  1
+#       define CIK_ADDR_SURF_MACRO_TILE_ASPECT_4  2
+#       define CIK_ADDR_SURF_MACRO_TILE_ASPECT_8  3
+#       define CIK_GRPH_ARRAY_MODE(x)             (((x) & 0x7) << 20)
+#       define CIK_GRPH_ARRAY_LINEAR_GENERAL      0
+#       define CIK_GRPH_ARRAY_LINEAR_ALIGNED      1
+#       define CIK_GRPH_ARRAY_1D_TILED_THIN1      2
+#       define CIK_GRPH_ARRAY_2D_TILED_THIN1      4
+#       define CIK_GRPH_PIPE_CONFIG(x)		 (((x) & 0x1f) << 24)
+#       define CIK_ADDR_SURF_P2			 0
+#       define CIK_ADDR_SURF_P4_8x16		 4
+#       define CIK_ADDR_SURF_P4_16x16		 5
+#       define CIK_ADDR_SURF_P4_16x32		 6
+#       define CIK_ADDR_SURF_P4_32x32		 7
+#       define CIK_ADDR_SURF_P8_16x16_8x16	 8
+#       define CIK_ADDR_SURF_P8_16x32_8x16	 9
+#       define CIK_ADDR_SURF_P8_32x32_8x16	 10
+#       define CIK_ADDR_SURF_P8_16x32_16x16	 11
+#       define CIK_ADDR_SURF_P8_32x32_16x16	 12
+#       define CIK_ADDR_SURF_P8_32x32_16x32	 13
+#       define CIK_ADDR_SURF_P8_32x64_32x32	 14
+#       define CIK_GRPH_MICRO_TILE_MODE(x)       (((x) & 0x7) << 29)
+#       define CIK_DISPLAY_MICRO_TILING          0
+#       define CIK_THIN_MICRO_TILING             1
+#       define CIK_DEPTH_MICRO_TILING            2
+#       define CIK_ROTATED_MICRO_TILING          4
+
+/* CUR blocks at 0x6998, 0x7598, 0x10198, 0x10d98, 0x11998, 0x12598 */
+#define CIK_CUR_CONTROL                           0x6998
+#       define CIK_CURSOR_EN                      (1 << 0)
+#       define CIK_CURSOR_MODE(x)                 (((x) & 0x3) << 8)
+#       define CIK_CURSOR_MONO                    0
+#       define CIK_CURSOR_24_1                    1
+#       define CIK_CURSOR_24_8_PRE_MULT           2
+#       define CIK_CURSOR_24_8_UNPRE_MULT         3
+#       define CIK_CURSOR_2X_MAGNIFY              (1 << 16)
+#       define CIK_CURSOR_FORCE_MC_ON             (1 << 20)
+#       define CIK_CURSOR_URGENT_CONTROL(x)       (((x) & 0x7) << 24)
+#       define CIK_CURSOR_URGENT_ALWAYS           0
+#       define CIK_CURSOR_URGENT_1_8              1
+#       define CIK_CURSOR_URGENT_1_4              2
+#       define CIK_CURSOR_URGENT_3_8              3
+#       define CIK_CURSOR_URGENT_1_2              4
+#define CIK_CUR_SURFACE_ADDRESS                   0x699c
+#       define CIK_CUR_SURFACE_ADDRESS_MASK       0xfffff000
+#define CIK_CUR_SIZE                              0x69a0
+#define CIK_CUR_SURFACE_ADDRESS_HIGH              0x69a4
+#define CIK_CUR_POSITION                          0x69a8
+#define CIK_CUR_HOT_SPOT                          0x69ac
+#define CIK_CUR_COLOR1                            0x69b0
+#define CIK_CUR_COLOR2                            0x69b4
+#define CIK_CUR_UPDATE                            0x69b8
+#       define CIK_CURSOR_UPDATE_PENDING          (1 << 0)
+#       define CIK_CURSOR_UPDATE_TAKEN            (1 << 1)
+#       define CIK_CURSOR_UPDATE_LOCK             (1 << 16)
+#       define CIK_CURSOR_DISABLE_MULTIPLE_UPDATE (1 << 24)
+
+#define CIK_ALPHA_CONTROL                         0x6af0
+#       define CIK_CURSOR_ALPHA_BLND_ENA          (1 << 1)
+
+#define CIK_LB_DATA_FORMAT                        0x6b00
+#       define CIK_INTERLEAVE_EN                  (1 << 3)
+
+#define CIK_LB_DESKTOP_HEIGHT                     0x6b0c
+
+#endif
diff --git a/drivers/gpu/drm/radeon/cikd.h b/drivers/gpu/drm/radeon/cikd.h
new file mode 100644
index 000000000000..7e9275eaef80
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cikd.h
@@ -0,0 +1,1313 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+#ifndef CIK_H
+#define CIK_H
+
+#define BONAIRE_GB_ADDR_CONFIG_GOLDEN        0x12010001
+
+#define CIK_RB_BITMAP_WIDTH_PER_SH  2
+
+/* SMC IND registers */
+#define GENERAL_PWRMGT                                    0xC0200000
+#       define GPU_COUNTER_CLK                            (1 << 15)
+
+#define CG_CLKPIN_CNTL                                    0xC05001A0
+#       define XTALIN_DIVIDE                              (1 << 1)
+
+#define PCIE_INDEX  					0x38
+#define PCIE_DATA  					0x3C
+
+#define VGA_HDP_CONTROL  				0x328
+#define		VGA_MEMORY_DISABLE				(1 << 4)
+
+#define DMIF_ADDR_CALC  				0xC00
+
+#define	SRBM_GFX_CNTL				        0xE44
+#define		PIPEID(x)					((x) << 0)
+#define		MEID(x)						((x) << 2)
+#define		VMID(x)						((x) << 4)
+#define		QUEUEID(x)					((x) << 8)
+
+#define	SRBM_STATUS2				        0xE4C
+#define		SDMA_BUSY 				(1 << 5)
+#define		SDMA1_BUSY 				(1 << 6)
+#define	SRBM_STATUS				        0xE50
+#define		UVD_RQ_PENDING 				(1 << 1)
+#define		GRBM_RQ_PENDING 			(1 << 5)
+#define		VMC_BUSY 				(1 << 8)
+#define		MCB_BUSY 				(1 << 9)
+#define		MCB_NON_DISPLAY_BUSY 			(1 << 10)
+#define		MCC_BUSY 				(1 << 11)
+#define		MCD_BUSY 				(1 << 12)
+#define		SEM_BUSY 				(1 << 14)
+#define		IH_BUSY 				(1 << 17)
+#define		UVD_BUSY 				(1 << 19)
+
+#define	SRBM_SOFT_RESET				        0xE60
+#define		SOFT_RESET_BIF				(1 << 1)
+#define		SOFT_RESET_R0PLL			(1 << 4)
+#define		SOFT_RESET_DC				(1 << 5)
+#define		SOFT_RESET_SDMA1			(1 << 6)
+#define		SOFT_RESET_GRBM				(1 << 8)
+#define		SOFT_RESET_HDP				(1 << 9)
+#define		SOFT_RESET_IH				(1 << 10)
+#define		SOFT_RESET_MC				(1 << 11)
+#define		SOFT_RESET_ROM				(1 << 14)
+#define		SOFT_RESET_SEM				(1 << 15)
+#define		SOFT_RESET_VMC				(1 << 17)
+#define		SOFT_RESET_SDMA				(1 << 20)
+#define		SOFT_RESET_TST				(1 << 21)
+#define		SOFT_RESET_REGBB		       	(1 << 22)
+#define		SOFT_RESET_ORB				(1 << 23)
+#define		SOFT_RESET_VCE				(1 << 24)
+
+#define VM_L2_CNTL					0x1400
+#define		ENABLE_L2_CACHE					(1 << 0)
+#define		ENABLE_L2_FRAGMENT_PROCESSING			(1 << 1)
+#define		L2_CACHE_PTE_ENDIAN_SWAP_MODE(x)		((x) << 2)
+#define		L2_CACHE_PDE_ENDIAN_SWAP_MODE(x)		((x) << 4)
+#define		ENABLE_L2_PTE_CACHE_LRU_UPDATE_BY_WRITE		(1 << 9)
+#define		ENABLE_L2_PDE0_CACHE_LRU_UPDATE_BY_WRITE	(1 << 10)
+#define		EFFECTIVE_L2_QUEUE_SIZE(x)			(((x) & 7) << 15)
+#define		CONTEXT1_IDENTITY_ACCESS_MODE(x)		(((x) & 3) << 19)
+#define VM_L2_CNTL2					0x1404
+#define		INVALIDATE_ALL_L1_TLBS				(1 << 0)
+#define		INVALIDATE_L2_CACHE				(1 << 1)
+#define		INVALIDATE_CACHE_MODE(x)			((x) << 26)
+#define			INVALIDATE_PTE_AND_PDE_CACHES		0
+#define			INVALIDATE_ONLY_PTE_CACHES		1
+#define			INVALIDATE_ONLY_PDE_CACHES		2
+#define VM_L2_CNTL3					0x1408
+#define		BANK_SELECT(x)					((x) << 0)
+#define		L2_CACHE_UPDATE_MODE(x)				((x) << 6)
+#define		L2_CACHE_BIGK_FRAGMENT_SIZE(x)			((x) << 15)
+#define		L2_CACHE_BIGK_ASSOCIATIVITY			(1 << 20)
+#define	VM_L2_STATUS					0x140C
+#define		L2_BUSY						(1 << 0)
+#define VM_CONTEXT0_CNTL				0x1410
+#define		ENABLE_CONTEXT					(1 << 0)
+#define		PAGE_TABLE_DEPTH(x)				(((x) & 3) << 1)
+#define		RANGE_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 3)
+#define		RANGE_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 4)
+#define		DUMMY_PAGE_PROTECTION_FAULT_ENABLE_INTERRUPT	(1 << 6)
+#define		DUMMY_PAGE_PROTECTION_FAULT_ENABLE_DEFAULT	(1 << 7)
+#define		PDE0_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 9)
+#define		PDE0_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 10)
+#define		VALID_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 12)
+#define		VALID_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 13)
+#define		READ_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 15)
+#define		READ_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 16)
+#define		WRITE_PROTECTION_FAULT_ENABLE_INTERRUPT		(1 << 18)
+#define		WRITE_PROTECTION_FAULT_ENABLE_DEFAULT		(1 << 19)
+#define VM_CONTEXT1_CNTL				0x1414
+#define VM_CONTEXT0_CNTL2				0x1430
+#define VM_CONTEXT1_CNTL2				0x1434
+#define	VM_CONTEXT8_PAGE_TABLE_BASE_ADDR		0x1438
+#define	VM_CONTEXT9_PAGE_TABLE_BASE_ADDR		0x143c
+#define	VM_CONTEXT10_PAGE_TABLE_BASE_ADDR		0x1440
+#define	VM_CONTEXT11_PAGE_TABLE_BASE_ADDR		0x1444
+#define	VM_CONTEXT12_PAGE_TABLE_BASE_ADDR		0x1448
+#define	VM_CONTEXT13_PAGE_TABLE_BASE_ADDR		0x144c
+#define	VM_CONTEXT14_PAGE_TABLE_BASE_ADDR		0x1450
+#define	VM_CONTEXT15_PAGE_TABLE_BASE_ADDR		0x1454
+
+#define VM_INVALIDATE_REQUEST				0x1478
+#define VM_INVALIDATE_RESPONSE				0x147c
+
+#define	VM_CONTEXT1_PROTECTION_FAULT_STATUS		0x14DC
+#define		PROTECTIONS_MASK			(0xf << 0)
+#define		PROTECTIONS_SHIFT			0
+		/* bit 0: range
+		 * bit 1: pde0
+		 * bit 2: valid
+		 * bit 3: read
+		 * bit 4: write
+		 */
+#define		MEMORY_CLIENT_ID_MASK			(0xff << 12)
+#define		MEMORY_CLIENT_ID_SHIFT			12
+#define		MEMORY_CLIENT_RW_MASK			(1 << 24)
+#define		MEMORY_CLIENT_RW_SHIFT			24
+#define		FAULT_VMID_MASK				(0xf << 25)
+#define		FAULT_VMID_SHIFT			25
+
+#define	VM_CONTEXT1_PROTECTION_FAULT_MCCLIENT		0x14E4
+
+#define	VM_CONTEXT1_PROTECTION_FAULT_ADDR		0x14FC
+
+#define VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR	0x1518
+#define VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR	0x151c
+
+#define	VM_CONTEXT0_PAGE_TABLE_BASE_ADDR		0x153c
+#define	VM_CONTEXT1_PAGE_TABLE_BASE_ADDR		0x1540
+#define	VM_CONTEXT2_PAGE_TABLE_BASE_ADDR		0x1544
+#define	VM_CONTEXT3_PAGE_TABLE_BASE_ADDR		0x1548
+#define	VM_CONTEXT4_PAGE_TABLE_BASE_ADDR		0x154c
+#define	VM_CONTEXT5_PAGE_TABLE_BASE_ADDR		0x1550
+#define	VM_CONTEXT6_PAGE_TABLE_BASE_ADDR		0x1554
+#define	VM_CONTEXT7_PAGE_TABLE_BASE_ADDR		0x1558
+#define	VM_CONTEXT0_PAGE_TABLE_START_ADDR		0x155c
+#define	VM_CONTEXT1_PAGE_TABLE_START_ADDR		0x1560
+
+#define	VM_CONTEXT0_PAGE_TABLE_END_ADDR			0x157C
+#define	VM_CONTEXT1_PAGE_TABLE_END_ADDR			0x1580
+
+#define MC_SHARED_CHMAP						0x2004
+#define		NOOFCHAN_SHIFT					12
+#define		NOOFCHAN_MASK					0x0000f000
+#define MC_SHARED_CHREMAP					0x2008
+
+#define CHUB_CONTROL					0x1864
+#define		BYPASS_VM					(1 << 0)
+
+#define	MC_VM_FB_LOCATION				0x2024
+#define	MC_VM_AGP_TOP					0x2028
+#define	MC_VM_AGP_BOT					0x202C
+#define	MC_VM_AGP_BASE					0x2030
+#define	MC_VM_SYSTEM_APERTURE_LOW_ADDR			0x2034
+#define	MC_VM_SYSTEM_APERTURE_HIGH_ADDR			0x2038
+#define	MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR		0x203C
+
+#define	MC_VM_MX_L1_TLB_CNTL				0x2064
+#define		ENABLE_L1_TLB					(1 << 0)
+#define		ENABLE_L1_FRAGMENT_PROCESSING			(1 << 1)
+#define		SYSTEM_ACCESS_MODE_PA_ONLY			(0 << 3)
+#define		SYSTEM_ACCESS_MODE_USE_SYS_MAP			(1 << 3)
+#define		SYSTEM_ACCESS_MODE_IN_SYS			(2 << 3)
+#define		SYSTEM_ACCESS_MODE_NOT_IN_SYS			(3 << 3)
+#define		SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU	(0 << 5)
+#define		ENABLE_ADVANCED_DRIVER_MODEL			(1 << 6)
+#define	MC_VM_FB_OFFSET					0x2068
+
+#define MC_SHARED_BLACKOUT_CNTL           		0x20ac
+
+#define	MC_ARB_RAMCFG					0x2760
+#define		NOOFBANK_SHIFT					0
+#define		NOOFBANK_MASK					0x00000003
+#define		NOOFRANK_SHIFT					2
+#define		NOOFRANK_MASK					0x00000004
+#define		NOOFROWS_SHIFT					3
+#define		NOOFROWS_MASK					0x00000038
+#define		NOOFCOLS_SHIFT					6
+#define		NOOFCOLS_MASK					0x000000C0
+#define		CHANSIZE_SHIFT					8
+#define		CHANSIZE_MASK					0x00000100
+#define		NOOFGROUPS_SHIFT				12
+#define		NOOFGROUPS_MASK					0x00001000
+
+#define MC_SEQ_SUP_CNTL           			0x28c8
+#define		RUN_MASK      				(1 << 0)
+#define MC_SEQ_SUP_PGM           			0x28cc
+
+#define	MC_SEQ_TRAIN_WAKEUP_CNTL			0x28e8
+#define		TRAIN_DONE_D0      			(1 << 30)
+#define		TRAIN_DONE_D1      			(1 << 31)
+
+#define MC_IO_PAD_CNTL_D0           			0x29d0
+#define		MEM_FALL_OUT_CMD      			(1 << 8)
+
+#define MC_SEQ_IO_DEBUG_INDEX           		0x2a44
+#define MC_SEQ_IO_DEBUG_DATA           			0x2a48
+
+#define	HDP_HOST_PATH_CNTL				0x2C00
+#define	HDP_NONSURFACE_BASE				0x2C04
+#define	HDP_NONSURFACE_INFO				0x2C08
+#define	HDP_NONSURFACE_SIZE				0x2C0C
+
+#define HDP_ADDR_CONFIG  				0x2F48
+#define HDP_MISC_CNTL					0x2F4C
+#define 	HDP_FLUSH_INVALIDATE_CACHE			(1 << 0)
+
+#define IH_RB_CNTL                                        0x3e00
+#       define IH_RB_ENABLE                               (1 << 0)
+#       define IH_RB_SIZE(x)                              ((x) << 1) /* log2 */
+#       define IH_RB_FULL_DRAIN_ENABLE                    (1 << 6)
+#       define IH_WPTR_WRITEBACK_ENABLE                   (1 << 8)
+#       define IH_WPTR_WRITEBACK_TIMER(x)                 ((x) << 9) /* log2 */
+#       define IH_WPTR_OVERFLOW_ENABLE                    (1 << 16)
+#       define IH_WPTR_OVERFLOW_CLEAR                     (1 << 31)
+#define IH_RB_BASE                                        0x3e04
+#define IH_RB_RPTR                                        0x3e08
+#define IH_RB_WPTR                                        0x3e0c
+#       define RB_OVERFLOW                                (1 << 0)
+#       define WPTR_OFFSET_MASK                           0x3fffc
+#define IH_RB_WPTR_ADDR_HI                                0x3e10
+#define IH_RB_WPTR_ADDR_LO                                0x3e14
+#define IH_CNTL                                           0x3e18
+#       define ENABLE_INTR                                (1 << 0)
+#       define IH_MC_SWAP(x)                              ((x) << 1)
+#       define IH_MC_SWAP_NONE                            0
+#       define IH_MC_SWAP_16BIT                           1
+#       define IH_MC_SWAP_32BIT                           2
+#       define IH_MC_SWAP_64BIT                           3
+#       define RPTR_REARM                                 (1 << 4)
+#       define MC_WRREQ_CREDIT(x)                         ((x) << 15)
+#       define MC_WR_CLEAN_CNT(x)                         ((x) << 20)
+#       define MC_VMID(x)                                 ((x) << 25)
+
+#define	CONFIG_MEMSIZE					0x5428
+
+#define INTERRUPT_CNTL                                    0x5468
+#       define IH_DUMMY_RD_OVERRIDE                       (1 << 0)
+#       define IH_DUMMY_RD_EN                             (1 << 1)
+#       define IH_REQ_NONSNOOP_EN                         (1 << 3)
+#       define GEN_IH_INT_EN                              (1 << 8)
+#define INTERRUPT_CNTL2                                   0x546c
+
+#define HDP_MEM_COHERENCY_FLUSH_CNTL			0x5480
+
+#define	BIF_FB_EN						0x5490
+#define		FB_READ_EN					(1 << 0)
+#define		FB_WRITE_EN					(1 << 1)
+
+#define HDP_REG_COHERENCY_FLUSH_CNTL			0x54A0
+
+#define GPU_HDP_FLUSH_REQ				0x54DC
+#define GPU_HDP_FLUSH_DONE				0x54E0
+#define		CP0					(1 << 0)
+#define		CP1					(1 << 1)
+#define		CP2					(1 << 2)
+#define		CP3					(1 << 3)
+#define		CP4					(1 << 4)
+#define		CP5					(1 << 5)
+#define		CP6					(1 << 6)
+#define		CP7					(1 << 7)
+#define		CP8					(1 << 8)
+#define		CP9					(1 << 9)
+#define		SDMA0					(1 << 10)
+#define		SDMA1					(1 << 11)
+
+/* 0x6b04, 0x7704, 0x10304, 0x10f04, 0x11b04, 0x12704 */
+#define	LB_MEMORY_CTRL					0x6b04
+#define		LB_MEMORY_SIZE(x)			((x) << 0)
+#define		LB_MEMORY_CONFIG(x)			((x) << 20)
+
+#define	DPG_WATERMARK_MASK_CONTROL			0x6cc8
+#       define LATENCY_WATERMARK_MASK(x)		((x) << 8)
+#define	DPG_PIPE_LATENCY_CONTROL			0x6ccc
+#       define LATENCY_LOW_WATERMARK(x)			((x) << 0)
+#       define LATENCY_HIGH_WATERMARK(x)		((x) << 16)
+
+/* 0x6b24, 0x7724, 0x10324, 0x10f24, 0x11b24, 0x12724 */
+#define LB_VLINE_STATUS                                 0x6b24
+#       define VLINE_OCCURRED                           (1 << 0)
+#       define VLINE_ACK                                (1 << 4)
+#       define VLINE_STAT                               (1 << 12)
+#       define VLINE_INTERRUPT                          (1 << 16)
+#       define VLINE_INTERRUPT_TYPE                     (1 << 17)
+/* 0x6b2c, 0x772c, 0x1032c, 0x10f2c, 0x11b2c, 0x1272c */
+#define LB_VBLANK_STATUS                                0x6b2c
+#       define VBLANK_OCCURRED                          (1 << 0)
+#       define VBLANK_ACK                               (1 << 4)
+#       define VBLANK_STAT                              (1 << 12)
+#       define VBLANK_INTERRUPT                         (1 << 16)
+#       define VBLANK_INTERRUPT_TYPE                    (1 << 17)
+
+/* 0x6b20, 0x7720, 0x10320, 0x10f20, 0x11b20, 0x12720 */
+#define LB_INTERRUPT_MASK                               0x6b20
+#       define VBLANK_INTERRUPT_MASK                    (1 << 0)
+#       define VLINE_INTERRUPT_MASK                     (1 << 4)
+#       define VLINE2_INTERRUPT_MASK                    (1 << 8)
+
+#define DISP_INTERRUPT_STATUS                           0x60f4
+#       define LB_D1_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D1_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD1_INTERRUPT                        (1 << 17)
+#       define DC_HPD1_RX_INTERRUPT                     (1 << 18)
+#       define DACA_AUTODETECT_INTERRUPT                (1 << 22)
+#       define DACB_AUTODETECT_INTERRUPT                (1 << 23)
+#       define DC_I2C_SW_DONE_INTERRUPT                 (1 << 24)
+#       define DC_I2C_HW_DONE_INTERRUPT                 (1 << 25)
+#define DISP_INTERRUPT_STATUS_CONTINUE                  0x60f8
+#       define LB_D2_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D2_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD2_INTERRUPT                        (1 << 17)
+#       define DC_HPD2_RX_INTERRUPT                     (1 << 18)
+#       define DISP_TIMER_INTERRUPT                     (1 << 24)
+#define DISP_INTERRUPT_STATUS_CONTINUE2                 0x60fc
+#       define LB_D3_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D3_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD3_INTERRUPT                        (1 << 17)
+#       define DC_HPD3_RX_INTERRUPT                     (1 << 18)
+#define DISP_INTERRUPT_STATUS_CONTINUE3                 0x6100
+#       define LB_D4_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D4_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD4_INTERRUPT                        (1 << 17)
+#       define DC_HPD4_RX_INTERRUPT                     (1 << 18)
+#define DISP_INTERRUPT_STATUS_CONTINUE4                 0x614c
+#       define LB_D5_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D5_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD5_INTERRUPT                        (1 << 17)
+#       define DC_HPD5_RX_INTERRUPT                     (1 << 18)
+#define DISP_INTERRUPT_STATUS_CONTINUE5                 0x6150
+#       define LB_D6_VLINE_INTERRUPT                    (1 << 2)
+#       define LB_D6_VBLANK_INTERRUPT                   (1 << 3)
+#       define DC_HPD6_INTERRUPT                        (1 << 17)
+#       define DC_HPD6_RX_INTERRUPT                     (1 << 18)
+#define DISP_INTERRUPT_STATUS_CONTINUE6                 0x6780
+
+#define	DAC_AUTODETECT_INT_CONTROL			0x67c8
+
+#define DC_HPD1_INT_STATUS                              0x601c
+#define DC_HPD2_INT_STATUS                              0x6028
+#define DC_HPD3_INT_STATUS                              0x6034
+#define DC_HPD4_INT_STATUS                              0x6040
+#define DC_HPD5_INT_STATUS                              0x604c
+#define DC_HPD6_INT_STATUS                              0x6058
+#       define DC_HPDx_INT_STATUS                       (1 << 0)
+#       define DC_HPDx_SENSE                            (1 << 1)
+#       define DC_HPDx_SENSE_DELAYED                    (1 << 4)
+#       define DC_HPDx_RX_INT_STATUS                    (1 << 8)
+
+#define DC_HPD1_INT_CONTROL                             0x6020
+#define DC_HPD2_INT_CONTROL                             0x602c
+#define DC_HPD3_INT_CONTROL                             0x6038
+#define DC_HPD4_INT_CONTROL                             0x6044
+#define DC_HPD5_INT_CONTROL                             0x6050
+#define DC_HPD6_INT_CONTROL                             0x605c
+#       define DC_HPDx_INT_ACK                          (1 << 0)
+#       define DC_HPDx_INT_POLARITY                     (1 << 8)
+#       define DC_HPDx_INT_EN                           (1 << 16)
+#       define DC_HPDx_RX_INT_ACK                       (1 << 20)
+#       define DC_HPDx_RX_INT_EN                        (1 << 24)
+
+#define DC_HPD1_CONTROL                                   0x6024
+#define DC_HPD2_CONTROL                                   0x6030
+#define DC_HPD3_CONTROL                                   0x603c
+#define DC_HPD4_CONTROL                                   0x6048
+#define DC_HPD5_CONTROL                                   0x6054
+#define DC_HPD6_CONTROL                                   0x6060
+#       define DC_HPDx_CONNECTION_TIMER(x)                ((x) << 0)
+#       define DC_HPDx_RX_INT_TIMER(x)                    ((x) << 16)
+#       define DC_HPDx_EN                                 (1 << 28)
+
+#define	GRBM_CNTL					0x8000
+#define		GRBM_READ_TIMEOUT(x)				((x) << 0)
+
+#define	GRBM_STATUS2					0x8008
+#define		ME0PIPE1_CMDFIFO_AVAIL_MASK			0x0000000F
+#define		ME0PIPE1_CF_RQ_PENDING				(1 << 4)
+#define		ME0PIPE1_PF_RQ_PENDING				(1 << 5)
+#define		ME1PIPE0_RQ_PENDING				(1 << 6)
+#define		ME1PIPE1_RQ_PENDING				(1 << 7)
+#define		ME1PIPE2_RQ_PENDING				(1 << 8)
+#define		ME1PIPE3_RQ_PENDING				(1 << 9)
+#define		ME2PIPE0_RQ_PENDING				(1 << 10)
+#define		ME2PIPE1_RQ_PENDING				(1 << 11)
+#define		ME2PIPE2_RQ_PENDING				(1 << 12)
+#define		ME2PIPE3_RQ_PENDING				(1 << 13)
+#define		RLC_RQ_PENDING 					(1 << 14)
+#define		RLC_BUSY 					(1 << 24)
+#define		TC_BUSY 					(1 << 25)
+#define		CPF_BUSY 					(1 << 28)
+#define		CPC_BUSY 					(1 << 29)
+#define		CPG_BUSY 					(1 << 30)
+
+#define	GRBM_STATUS					0x8010
+#define		ME0PIPE0_CMDFIFO_AVAIL_MASK			0x0000000F
+#define		SRBM_RQ_PENDING					(1 << 5)
+#define		ME0PIPE0_CF_RQ_PENDING				(1 << 7)
+#define		ME0PIPE0_PF_RQ_PENDING				(1 << 8)
+#define		GDS_DMA_RQ_PENDING				(1 << 9)
+#define		DB_CLEAN					(1 << 12)
+#define		CB_CLEAN					(1 << 13)
+#define		TA_BUSY 					(1 << 14)
+#define		GDS_BUSY 					(1 << 15)
+#define		WD_BUSY_NO_DMA 					(1 << 16)
+#define		VGT_BUSY					(1 << 17)
+#define		IA_BUSY_NO_DMA					(1 << 18)
+#define		IA_BUSY						(1 << 19)
+#define		SX_BUSY 					(1 << 20)
+#define		WD_BUSY 					(1 << 21)
+#define		SPI_BUSY					(1 << 22)
+#define		BCI_BUSY					(1 << 23)
+#define		SC_BUSY 					(1 << 24)
+#define		PA_BUSY 					(1 << 25)
+#define		DB_BUSY 					(1 << 26)
+#define		CP_COHERENCY_BUSY      				(1 << 28)
+#define		CP_BUSY 					(1 << 29)
+#define		CB_BUSY 					(1 << 30)
+#define		GUI_ACTIVE					(1 << 31)
+#define	GRBM_STATUS_SE0					0x8014
+#define	GRBM_STATUS_SE1					0x8018
+#define	GRBM_STATUS_SE2					0x8038
+#define	GRBM_STATUS_SE3					0x803C
+#define		SE_DB_CLEAN					(1 << 1)
+#define		SE_CB_CLEAN					(1 << 2)
+#define		SE_BCI_BUSY					(1 << 22)
+#define		SE_VGT_BUSY					(1 << 23)
+#define		SE_PA_BUSY					(1 << 24)
+#define		SE_TA_BUSY					(1 << 25)
+#define		SE_SX_BUSY					(1 << 26)
+#define		SE_SPI_BUSY					(1 << 27)
+#define		SE_SC_BUSY					(1 << 29)
+#define		SE_DB_BUSY					(1 << 30)
+#define		SE_CB_BUSY					(1 << 31)
+
+#define	GRBM_SOFT_RESET					0x8020
+#define		SOFT_RESET_CP					(1 << 0)  /* All CP blocks */
+#define		SOFT_RESET_RLC					(1 << 2)  /* RLC */
+#define		SOFT_RESET_GFX					(1 << 16) /* GFX */
+#define		SOFT_RESET_CPF					(1 << 17) /* CP fetcher shared by gfx and compute */
+#define		SOFT_RESET_CPC					(1 << 18) /* CP Compute (MEC1/2) */
+#define		SOFT_RESET_CPG					(1 << 19) /* CP GFX (PFP, ME, CE) */
+
+#define GRBM_INT_CNTL                                   0x8060
+#       define RDERR_INT_ENABLE                         (1 << 0)
+#       define GUI_IDLE_INT_ENABLE                      (1 << 19)
+
+#define CP_CPC_STATUS					0x8210
+#define CP_CPC_BUSY_STAT				0x8214
+#define CP_CPC_STALLED_STAT1				0x8218
+#define CP_CPF_STATUS					0x821c
+#define CP_CPF_BUSY_STAT				0x8220
+#define CP_CPF_STALLED_STAT1				0x8224
+
+#define CP_MEC_CNTL					0x8234
+#define		MEC_ME2_HALT					(1 << 28)
+#define		MEC_ME1_HALT					(1 << 30)
+
+#define CP_MEC_CNTL					0x8234
+#define		MEC_ME2_HALT					(1 << 28)
+#define		MEC_ME1_HALT					(1 << 30)
+
+#define CP_STALLED_STAT3				0x8670
+#define CP_STALLED_STAT1				0x8674
+#define CP_STALLED_STAT2				0x8678
+
+#define CP_STAT						0x8680
+
+#define CP_ME_CNTL					0x86D8
+#define		CP_CE_HALT					(1 << 24)
+#define		CP_PFP_HALT					(1 << 26)
+#define		CP_ME_HALT					(1 << 28)
+
+#define	CP_RB0_RPTR					0x8700
+#define	CP_RB_WPTR_DELAY				0x8704
+
+#define CP_MEQ_THRESHOLDS				0x8764
+#define		MEQ1_START(x)				((x) << 0)
+#define		MEQ2_START(x)				((x) << 8)
+
+#define	VGT_VTX_VECT_EJECT_REG				0x88B0
+
+#define	VGT_CACHE_INVALIDATION				0x88C4
+#define		CACHE_INVALIDATION(x)				((x) << 0)
+#define			VC_ONLY						0
+#define			TC_ONLY						1
+#define			VC_AND_TC					2
+#define		AUTO_INVLD_EN(x)				((x) << 6)
+#define			NO_AUTO						0
+#define			ES_AUTO						1
+#define			GS_AUTO						2
+#define			ES_AND_GS_AUTO					3
+
+#define	VGT_GS_VERTEX_REUSE				0x88D4
+
+#define CC_GC_SHADER_ARRAY_CONFIG			0x89bc
+#define		INACTIVE_CUS_MASK			0xFFFF0000
+#define		INACTIVE_CUS_SHIFT			16
+#define GC_USER_SHADER_ARRAY_CONFIG			0x89c0
+
+#define	PA_CL_ENHANCE					0x8A14
+#define		CLIP_VTX_REORDER_ENA				(1 << 0)
+#define		NUM_CLIP_SEQ(x)					((x) << 1)
+
+#define	PA_SC_FORCE_EOV_MAX_CNTS			0x8B24
+#define		FORCE_EOV_MAX_CLK_CNT(x)			((x) << 0)
+#define		FORCE_EOV_MAX_REZ_CNT(x)			((x) << 16)
+
+#define	PA_SC_FIFO_SIZE					0x8BCC
+#define		SC_FRONTEND_PRIM_FIFO_SIZE(x)			((x) << 0)
+#define		SC_BACKEND_PRIM_FIFO_SIZE(x)			((x) << 6)
+#define		SC_HIZ_TILE_FIFO_SIZE(x)			((x) << 15)
+#define		SC_EARLYZ_TILE_FIFO_SIZE(x)			((x) << 23)
+
+#define	PA_SC_ENHANCE					0x8BF0
+#define		ENABLE_PA_SC_OUT_OF_ORDER			(1 << 0)
+#define		DISABLE_PA_SC_GUIDANCE				(1 << 13)
+
+#define	SQ_CONFIG					0x8C00
+
+#define	SH_MEM_BASES					0x8C28
+/* if PTR32, these are the bases for scratch and lds */
+#define		PRIVATE_BASE(x)					((x) << 0) /* scratch */
+#define		SHARED_BASE(x)					((x) << 16) /* LDS */
+#define	SH_MEM_APE1_BASE				0x8C2C
+/* if PTR32, this is the base location of GPUVM */
+#define	SH_MEM_APE1_LIMIT				0x8C30
+/* if PTR32, this is the upper limit of GPUVM */
+#define	SH_MEM_CONFIG					0x8C34
+#define		PTR32						(1 << 0)
+#define		ALIGNMENT_MODE(x)				((x) << 2)
+#define			SH_MEM_ALIGNMENT_MODE_DWORD			0
+#define			SH_MEM_ALIGNMENT_MODE_DWORD_STRICT		1
+#define			SH_MEM_ALIGNMENT_MODE_STRICT			2
+#define			SH_MEM_ALIGNMENT_MODE_UNALIGNED			3
+#define		DEFAULT_MTYPE(x)				((x) << 4)
+#define		APE1_MTYPE(x)					((x) << 7)
+
+#define	SX_DEBUG_1					0x9060
+
+#define	SPI_CONFIG_CNTL					0x9100
+
+#define	SPI_CONFIG_CNTL_1				0x913C
+#define		VTX_DONE_DELAY(x)				((x) << 0)
+#define		INTERP_ONE_PRIM_PER_ROW				(1 << 4)
+
+#define	TA_CNTL_AUX					0x9508
+
+#define DB_DEBUG					0x9830
+#define DB_DEBUG2					0x9834
+#define DB_DEBUG3					0x9838
+
+#define CC_RB_BACKEND_DISABLE				0x98F4
+#define		BACKEND_DISABLE(x)     			((x) << 16)
+#define GB_ADDR_CONFIG  				0x98F8
+#define		NUM_PIPES(x)				((x) << 0)
+#define		NUM_PIPES_MASK				0x00000007
+#define		NUM_PIPES_SHIFT				0
+#define		PIPE_INTERLEAVE_SIZE(x)			((x) << 4)
+#define		PIPE_INTERLEAVE_SIZE_MASK		0x00000070
+#define		PIPE_INTERLEAVE_SIZE_SHIFT		4
+#define		NUM_SHADER_ENGINES(x)			((x) << 12)
+#define		NUM_SHADER_ENGINES_MASK			0x00003000
+#define		NUM_SHADER_ENGINES_SHIFT		12
+#define		SHADER_ENGINE_TILE_SIZE(x)     		((x) << 16)
+#define		SHADER_ENGINE_TILE_SIZE_MASK		0x00070000
+#define		SHADER_ENGINE_TILE_SIZE_SHIFT		16
+#define		ROW_SIZE(x)             		((x) << 28)
+#define		ROW_SIZE_MASK				0x30000000
+#define		ROW_SIZE_SHIFT				28
+
+#define	GB_TILE_MODE0					0x9910
+#       define ARRAY_MODE(x)					((x) << 2)
+#              define	ARRAY_LINEAR_GENERAL			0
+#              define	ARRAY_LINEAR_ALIGNED			1
+#              define	ARRAY_1D_TILED_THIN1			2
+#              define	ARRAY_2D_TILED_THIN1			4
+#              define	ARRAY_PRT_TILED_THIN1			5
+#              define	ARRAY_PRT_2D_TILED_THIN1		6
+#       define PIPE_CONFIG(x)					((x) << 6)
+#              define	ADDR_SURF_P2				0
+#              define	ADDR_SURF_P4_8x16			4
+#              define	ADDR_SURF_P4_16x16			5
+#              define	ADDR_SURF_P4_16x32			6
+#              define	ADDR_SURF_P4_32x32			7
+#              define	ADDR_SURF_P8_16x16_8x16			8
+#              define	ADDR_SURF_P8_16x32_8x16			9
+#              define	ADDR_SURF_P8_32x32_8x16			10
+#              define	ADDR_SURF_P8_16x32_16x16		11
+#              define	ADDR_SURF_P8_32x32_16x16		12
+#              define	ADDR_SURF_P8_32x32_16x32		13
+#              define	ADDR_SURF_P8_32x64_32x32		14
+#       define TILE_SPLIT(x)					((x) << 11)
+#              define	ADDR_SURF_TILE_SPLIT_64B		0
+#              define	ADDR_SURF_TILE_SPLIT_128B		1
+#              define	ADDR_SURF_TILE_SPLIT_256B		2
+#              define	ADDR_SURF_TILE_SPLIT_512B		3
+#              define	ADDR_SURF_TILE_SPLIT_1KB		4
+#              define	ADDR_SURF_TILE_SPLIT_2KB		5
+#              define	ADDR_SURF_TILE_SPLIT_4KB		6
+#       define MICRO_TILE_MODE_NEW(x)				((x) << 22)
+#              define	ADDR_SURF_DISPLAY_MICRO_TILING		0
+#              define	ADDR_SURF_THIN_MICRO_TILING		1
+#              define	ADDR_SURF_DEPTH_MICRO_TILING		2
+#              define	ADDR_SURF_ROTATED_MICRO_TILING		3
+#       define SAMPLE_SPLIT(x)					((x) << 25)
+#              define	ADDR_SURF_SAMPLE_SPLIT_1		0
+#              define	ADDR_SURF_SAMPLE_SPLIT_2		1
+#              define	ADDR_SURF_SAMPLE_SPLIT_4		2
+#              define	ADDR_SURF_SAMPLE_SPLIT_8		3
+
+#define	GB_MACROTILE_MODE0					0x9990
+#       define BANK_WIDTH(x)					((x) << 0)
+#              define	ADDR_SURF_BANK_WIDTH_1			0
+#              define	ADDR_SURF_BANK_WIDTH_2			1
+#              define	ADDR_SURF_BANK_WIDTH_4			2
+#              define	ADDR_SURF_BANK_WIDTH_8			3
+#       define BANK_HEIGHT(x)					((x) << 2)
+#              define	ADDR_SURF_BANK_HEIGHT_1			0
+#              define	ADDR_SURF_BANK_HEIGHT_2			1
+#              define	ADDR_SURF_BANK_HEIGHT_4			2
+#              define	ADDR_SURF_BANK_HEIGHT_8			3
+#       define MACRO_TILE_ASPECT(x)				((x) << 4)
+#              define	ADDR_SURF_MACRO_ASPECT_1		0
+#              define	ADDR_SURF_MACRO_ASPECT_2		1
+#              define	ADDR_SURF_MACRO_ASPECT_4		2
+#              define	ADDR_SURF_MACRO_ASPECT_8		3
+#       define NUM_BANKS(x)					((x) << 6)
+#              define	ADDR_SURF_2_BANK			0
+#              define	ADDR_SURF_4_BANK			1
+#              define	ADDR_SURF_8_BANK			2
+#              define	ADDR_SURF_16_BANK			3
+
+#define	CB_HW_CONTROL					0x9A10
+
+#define	GC_USER_RB_BACKEND_DISABLE			0x9B7C
+#define		BACKEND_DISABLE_MASK			0x00FF0000
+#define		BACKEND_DISABLE_SHIFT			16
+
+#define	TCP_CHAN_STEER_LO				0xac0c
+#define	TCP_CHAN_STEER_HI				0xac10
+
+#define	TC_CFG_L1_LOAD_POLICY0				0xAC68
+#define	TC_CFG_L1_LOAD_POLICY1				0xAC6C
+#define	TC_CFG_L1_STORE_POLICY				0xAC70
+#define	TC_CFG_L2_LOAD_POLICY0				0xAC74
+#define	TC_CFG_L2_LOAD_POLICY1				0xAC78
+#define	TC_CFG_L2_STORE_POLICY0				0xAC7C
+#define	TC_CFG_L2_STORE_POLICY1				0xAC80
+#define	TC_CFG_L2_ATOMIC_POLICY				0xAC84
+#define	TC_CFG_L1_VOLATILE				0xAC88
+#define	TC_CFG_L2_VOLATILE				0xAC8C
+
+#define	CP_RB0_BASE					0xC100
+#define	CP_RB0_CNTL					0xC104
+#define		RB_BUFSZ(x)					((x) << 0)
+#define		RB_BLKSZ(x)					((x) << 8)
+#define		BUF_SWAP_32BIT					(2 << 16)
+#define		RB_NO_UPDATE					(1 << 27)
+#define		RB_RPTR_WR_ENA					(1 << 31)
+
+#define	CP_RB0_RPTR_ADDR				0xC10C
+#define		RB_RPTR_SWAP_32BIT				(2 << 0)
+#define	CP_RB0_RPTR_ADDR_HI				0xC110
+#define	CP_RB0_WPTR					0xC114
+
+#define	CP_DEVICE_ID					0xC12C
+#define	CP_ENDIAN_SWAP					0xC140
+#define	CP_RB_VMID					0xC144
+
+#define	CP_PFP_UCODE_ADDR				0xC150
+#define	CP_PFP_UCODE_DATA				0xC154
+#define	CP_ME_RAM_RADDR					0xC158
+#define	CP_ME_RAM_WADDR					0xC15C
+#define	CP_ME_RAM_DATA					0xC160
+
+#define	CP_CE_UCODE_ADDR				0xC168
+#define	CP_CE_UCODE_DATA				0xC16C
+#define	CP_MEC_ME1_UCODE_ADDR				0xC170
+#define	CP_MEC_ME1_UCODE_DATA				0xC174
+#define	CP_MEC_ME2_UCODE_ADDR				0xC178
+#define	CP_MEC_ME2_UCODE_DATA				0xC17C
+
+#define CP_INT_CNTL_RING0                               0xC1A8
+#       define CNTX_BUSY_INT_ENABLE                     (1 << 19)
+#       define CNTX_EMPTY_INT_ENABLE                    (1 << 20)
+#       define PRIV_INSTR_INT_ENABLE                    (1 << 22)
+#       define PRIV_REG_INT_ENABLE                      (1 << 23)
+#       define TIME_STAMP_INT_ENABLE                    (1 << 26)
+#       define CP_RINGID2_INT_ENABLE                    (1 << 29)
+#       define CP_RINGID1_INT_ENABLE                    (1 << 30)
+#       define CP_RINGID0_INT_ENABLE                    (1 << 31)
+
+#define CP_INT_STATUS_RING0                             0xC1B4
+#       define PRIV_INSTR_INT_STAT                      (1 << 22)
+#       define PRIV_REG_INT_STAT                        (1 << 23)
+#       define TIME_STAMP_INT_STAT                      (1 << 26)
+#       define CP_RINGID2_INT_STAT                      (1 << 29)
+#       define CP_RINGID1_INT_STAT                      (1 << 30)
+#       define CP_RINGID0_INT_STAT                      (1 << 31)
+
+#define CP_CPF_DEBUG                                    0xC200
+
+#define CP_PQ_WPTR_POLL_CNTL                            0xC20C
+#define		WPTR_POLL_EN      			(1 << 31)
+
+#define CP_ME1_PIPE0_INT_CNTL                           0xC214
+#define CP_ME1_PIPE1_INT_CNTL                           0xC218
+#define CP_ME1_PIPE2_INT_CNTL                           0xC21C
+#define CP_ME1_PIPE3_INT_CNTL                           0xC220
+#define CP_ME2_PIPE0_INT_CNTL                           0xC224
+#define CP_ME2_PIPE1_INT_CNTL                           0xC228
+#define CP_ME2_PIPE2_INT_CNTL                           0xC22C
+#define CP_ME2_PIPE3_INT_CNTL                           0xC230
+#       define DEQUEUE_REQUEST_INT_ENABLE               (1 << 13)
+#       define WRM_POLL_TIMEOUT_INT_ENABLE              (1 << 17)
+#       define PRIV_REG_INT_ENABLE                      (1 << 23)
+#       define TIME_STAMP_INT_ENABLE                    (1 << 26)
+#       define GENERIC2_INT_ENABLE                      (1 << 29)
+#       define GENERIC1_INT_ENABLE                      (1 << 30)
+#       define GENERIC0_INT_ENABLE                      (1 << 31)
+#define CP_ME1_PIPE0_INT_STATUS                         0xC214
+#define CP_ME1_PIPE1_INT_STATUS                         0xC218
+#define CP_ME1_PIPE2_INT_STATUS                         0xC21C
+#define CP_ME1_PIPE3_INT_STATUS                         0xC220
+#define CP_ME2_PIPE0_INT_STATUS                         0xC224
+#define CP_ME2_PIPE1_INT_STATUS                         0xC228
+#define CP_ME2_PIPE2_INT_STATUS                         0xC22C
+#define CP_ME2_PIPE3_INT_STATUS                         0xC230
+#       define DEQUEUE_REQUEST_INT_STATUS               (1 << 13)
+#       define WRM_POLL_TIMEOUT_INT_STATUS              (1 << 17)
+#       define PRIV_REG_INT_STATUS                      (1 << 23)
+#       define TIME_STAMP_INT_STATUS                    (1 << 26)
+#       define GENERIC2_INT_STATUS                      (1 << 29)
+#       define GENERIC1_INT_STATUS                      (1 << 30)
+#       define GENERIC0_INT_STATUS                      (1 << 31)
+
+#define	CP_MAX_CONTEXT					0xC2B8
+
+#define	CP_RB0_BASE_HI					0xC2C4
+
+#define RLC_CNTL                                          0xC300
+#       define RLC_ENABLE                                 (1 << 0)
+
+#define RLC_MC_CNTL                                       0xC30C
+
+#define RLC_LB_CNTR_MAX                                   0xC348
+
+#define RLC_LB_CNTL                                       0xC364
+
+#define RLC_LB_CNTR_INIT                                  0xC36C
+
+#define RLC_SAVE_AND_RESTORE_BASE                         0xC374
+#define RLC_DRIVER_DMA_STATUS                             0xC378
+
+#define RLC_GPM_UCODE_ADDR                                0xC388
+#define RLC_GPM_UCODE_DATA                                0xC38C
+#define RLC_GPU_CLOCK_COUNT_LSB                           0xC390
+#define RLC_GPU_CLOCK_COUNT_MSB                           0xC394
+#define RLC_CAPTURE_GPU_CLOCK_COUNT                       0xC398
+#define RLC_UCODE_CNTL                                    0xC39C
+
+#define RLC_CGCG_CGLS_CTRL                                0xC424
+
+#define RLC_LB_INIT_CU_MASK                               0xC43C
+
+#define RLC_LB_PARAMS                                     0xC444
+
+#define RLC_SERDES_CU_MASTER_BUSY                         0xC484
+#define RLC_SERDES_NONCU_MASTER_BUSY                      0xC488
+#       define SE_MASTER_BUSY_MASK                        0x0000ffff
+#       define GC_MASTER_BUSY                             (1 << 16)
+#       define TC0_MASTER_BUSY                            (1 << 17)
+#       define TC1_MASTER_BUSY                            (1 << 18)
+
+#define RLC_GPM_SCRATCH_ADDR                              0xC4B0
+#define RLC_GPM_SCRATCH_DATA                              0xC4B4
+
+#define CP_HPD_EOP_BASE_ADDR                              0xC904
+#define CP_HPD_EOP_BASE_ADDR_HI                           0xC908
+#define CP_HPD_EOP_VMID                                   0xC90C
+#define CP_HPD_EOP_CONTROL                                0xC910
+#define		EOP_SIZE(x)				((x) << 0)
+#define		EOP_SIZE_MASK      			(0x3f << 0)
+#define CP_MQD_BASE_ADDR                                  0xC914
+#define CP_MQD_BASE_ADDR_HI                               0xC918
+#define CP_HQD_ACTIVE                                     0xC91C
+#define CP_HQD_VMID                                       0xC920
+
+#define CP_HQD_PQ_BASE                                    0xC934
+#define CP_HQD_PQ_BASE_HI                                 0xC938
+#define CP_HQD_PQ_RPTR                                    0xC93C
+#define CP_HQD_PQ_RPTR_REPORT_ADDR                        0xC940
+#define CP_HQD_PQ_RPTR_REPORT_ADDR_HI                     0xC944
+#define CP_HQD_PQ_WPTR_POLL_ADDR                          0xC948
+#define CP_HQD_PQ_WPTR_POLL_ADDR_HI                       0xC94C
+#define CP_HQD_PQ_DOORBELL_CONTROL                        0xC950
+#define		DOORBELL_OFFSET(x)			((x) << 2)
+#define		DOORBELL_OFFSET_MASK			(0x1fffff << 2)
+#define		DOORBELL_SOURCE      			(1 << 28)
+#define		DOORBELL_SCHD_HIT      			(1 << 29)
+#define		DOORBELL_EN      			(1 << 30)
+#define		DOORBELL_HIT      			(1 << 31)
+#define CP_HQD_PQ_WPTR                                    0xC954
+#define CP_HQD_PQ_CONTROL                                 0xC958
+#define		QUEUE_SIZE(x)				((x) << 0)
+#define		QUEUE_SIZE_MASK      			(0x3f << 0)
+#define		RPTR_BLOCK_SIZE(x)			((x) << 8)
+#define		RPTR_BLOCK_SIZE_MASK			(0x3f << 8)
+#define		PQ_VOLATILE      			(1 << 26)
+#define		NO_UPDATE_RPTR      			(1 << 27)
+#define		UNORD_DISPATCH      			(1 << 28)
+#define		ROQ_PQ_IB_FLIP      			(1 << 29)
+#define		PRIV_STATE      			(1 << 30)
+#define		KMD_QUEUE      				(1 << 31)
+
+#define CP_HQD_DEQUEUE_REQUEST                          0xC974
+
+#define CP_MQD_CONTROL                                  0xC99C
+#define		MQD_VMID(x)				((x) << 0)
+#define		MQD_VMID_MASK      			(0xf << 0)
+
+#define PA_SC_RASTER_CONFIG                             0x28350
+#       define RASTER_CONFIG_RB_MAP_0                   0
+#       define RASTER_CONFIG_RB_MAP_1                   1
+#       define RASTER_CONFIG_RB_MAP_2                   2
+#       define RASTER_CONFIG_RB_MAP_3                   3
+
+#define VGT_EVENT_INITIATOR                             0x28a90
+#       define SAMPLE_STREAMOUTSTATS1                   (1 << 0)
+#       define SAMPLE_STREAMOUTSTATS2                   (2 << 0)
+#       define SAMPLE_STREAMOUTSTATS3                   (3 << 0)
+#       define CACHE_FLUSH_TS                           (4 << 0)
+#       define CACHE_FLUSH                              (6 << 0)
+#       define CS_PARTIAL_FLUSH                         (7 << 0)
+#       define VGT_STREAMOUT_RESET                      (10 << 0)
+#       define END_OF_PIPE_INCR_DE                      (11 << 0)
+#       define END_OF_PIPE_IB_END                       (12 << 0)
+#       define RST_PIX_CNT                              (13 << 0)
+#       define VS_PARTIAL_FLUSH                         (15 << 0)
+#       define PS_PARTIAL_FLUSH                         (16 << 0)
+#       define CACHE_FLUSH_AND_INV_TS_EVENT             (20 << 0)
+#       define ZPASS_DONE                               (21 << 0)
+#       define CACHE_FLUSH_AND_INV_EVENT                (22 << 0)
+#       define PERFCOUNTER_START                        (23 << 0)
+#       define PERFCOUNTER_STOP                         (24 << 0)
+#       define PIPELINESTAT_START                       (25 << 0)
+#       define PIPELINESTAT_STOP                        (26 << 0)
+#       define PERFCOUNTER_SAMPLE                       (27 << 0)
+#       define SAMPLE_PIPELINESTAT                      (30 << 0)
+#       define SO_VGT_STREAMOUT_FLUSH                   (31 << 0)
+#       define SAMPLE_STREAMOUTSTATS                    (32 << 0)
+#       define RESET_VTX_CNT                            (33 << 0)
+#       define VGT_FLUSH                                (36 << 0)
+#       define BOTTOM_OF_PIPE_TS                        (40 << 0)
+#       define DB_CACHE_FLUSH_AND_INV                   (42 << 0)
+#       define FLUSH_AND_INV_DB_DATA_TS                 (43 << 0)
+#       define FLUSH_AND_INV_DB_META                    (44 << 0)
+#       define FLUSH_AND_INV_CB_DATA_TS                 (45 << 0)
+#       define FLUSH_AND_INV_CB_META                    (46 << 0)
+#       define CS_DONE                                  (47 << 0)
+#       define PS_DONE                                  (48 << 0)
+#       define FLUSH_AND_INV_CB_PIXEL_DATA              (49 << 0)
+#       define THREAD_TRACE_START                       (51 << 0)
+#       define THREAD_TRACE_STOP                        (52 << 0)
+#       define THREAD_TRACE_FLUSH                       (54 << 0)
+#       define THREAD_TRACE_FINISH                      (55 << 0)
+#       define PIXEL_PIPE_STAT_CONTROL                  (56 << 0)
+#       define PIXEL_PIPE_STAT_DUMP                     (57 << 0)
+#       define PIXEL_PIPE_STAT_RESET                    (58 << 0)
+
+#define	SCRATCH_REG0					0x30100
+#define	SCRATCH_REG1					0x30104
+#define	SCRATCH_REG2					0x30108
+#define	SCRATCH_REG3					0x3010C
+#define	SCRATCH_REG4					0x30110
+#define	SCRATCH_REG5					0x30114
+#define	SCRATCH_REG6					0x30118
+#define	SCRATCH_REG7					0x3011C
+
+#define	SCRATCH_UMSK					0x30140
+#define	SCRATCH_ADDR					0x30144
+
+#define	CP_SEM_WAIT_TIMER				0x301BC
+
+#define	CP_SEM_INCOMPLETE_TIMER_CNTL			0x301C8
+
+#define	CP_WAIT_REG_MEM_TIMEOUT				0x301D0
+
+#define GRBM_GFX_INDEX          			0x30800
+#define		INSTANCE_INDEX(x)			((x) << 0)
+#define		SH_INDEX(x)     			((x) << 8)
+#define		SE_INDEX(x)     			((x) << 16)
+#define		SH_BROADCAST_WRITES      		(1 << 29)
+#define		INSTANCE_BROADCAST_WRITES      		(1 << 30)
+#define		SE_BROADCAST_WRITES      		(1 << 31)
+
+#define	VGT_ESGS_RING_SIZE				0x30900
+#define	VGT_GSVS_RING_SIZE				0x30904
+#define	VGT_PRIMITIVE_TYPE				0x30908
+#define	VGT_INDEX_TYPE					0x3090C
+
+#define	VGT_NUM_INDICES					0x30930
+#define	VGT_NUM_INSTANCES				0x30934
+#define	VGT_TF_RING_SIZE				0x30938
+#define	VGT_HS_OFFCHIP_PARAM				0x3093C
+#define	VGT_TF_MEMORY_BASE				0x30940
+
+#define	PA_SU_LINE_STIPPLE_VALUE			0x30a00
+#define	PA_SC_LINE_STIPPLE_STATE			0x30a04
+
+#define	SQC_CACHES					0x30d20
+
+#define	CP_PERFMON_CNTL					0x36020
+
+#define	CGTS_TCC_DISABLE				0x3c00c
+#define	CGTS_USER_TCC_DISABLE				0x3c010
+#define		TCC_DISABLE_MASK				0xFFFF0000
+#define		TCC_DISABLE_SHIFT				16
+
+#define	CB_CGTT_SCLK_CTRL				0x3c2a0
+
+/*
+ * PM4
+ */
+#define	PACKET_TYPE0	0
+#define	PACKET_TYPE1	1
+#define	PACKET_TYPE2	2
+#define	PACKET_TYPE3	3
+
+#define CP_PACKET_GET_TYPE(h) (((h) >> 30) & 3)
+#define CP_PACKET_GET_COUNT(h) (((h) >> 16) & 0x3FFF)
+#define CP_PACKET0_GET_REG(h) (((h) & 0xFFFF) << 2)
+#define CP_PACKET3_GET_OPCODE(h) (((h) >> 8) & 0xFF)
+#define PACKET0(reg, n)	((PACKET_TYPE0 << 30) |				\
+			 (((reg) >> 2) & 0xFFFF) |			\
+			 ((n) & 0x3FFF) << 16)
+#define CP_PACKET2			0x80000000
+#define		PACKET2_PAD_SHIFT		0
+#define		PACKET2_PAD_MASK		(0x3fffffff << 0)
+
+#define PACKET2(v)	(CP_PACKET2 | REG_SET(PACKET2_PAD, (v)))
+
+#define PACKET3(op, n)	((PACKET_TYPE3 << 30) |				\
+			 (((op) & 0xFF) << 8) |				\
+			 ((n) & 0x3FFF) << 16)
+
+#define PACKET3_COMPUTE(op, n) (PACKET3(op, n) | 1 << 1)
+
+/* Packet 3 types */
+#define	PACKET3_NOP					0x10
+#define	PACKET3_SET_BASE				0x11
+#define		PACKET3_BASE_INDEX(x)                  ((x) << 0)
+#define			CE_PARTITION_BASE		3
+#define	PACKET3_CLEAR_STATE				0x12
+#define	PACKET3_INDEX_BUFFER_SIZE			0x13
+#define	PACKET3_DISPATCH_DIRECT				0x15
+#define	PACKET3_DISPATCH_INDIRECT			0x16
+#define	PACKET3_ATOMIC_GDS				0x1D
+#define	PACKET3_ATOMIC_MEM				0x1E
+#define	PACKET3_OCCLUSION_QUERY				0x1F
+#define	PACKET3_SET_PREDICATION				0x20
+#define	PACKET3_REG_RMW					0x21
+#define	PACKET3_COND_EXEC				0x22
+#define	PACKET3_PRED_EXEC				0x23
+#define	PACKET3_DRAW_INDIRECT				0x24
+#define	PACKET3_DRAW_INDEX_INDIRECT			0x25
+#define	PACKET3_INDEX_BASE				0x26
+#define	PACKET3_DRAW_INDEX_2				0x27
+#define	PACKET3_CONTEXT_CONTROL				0x28
+#define	PACKET3_INDEX_TYPE				0x2A
+#define	PACKET3_DRAW_INDIRECT_MULTI			0x2C
+#define	PACKET3_DRAW_INDEX_AUTO				0x2D
+#define	PACKET3_NUM_INSTANCES				0x2F
+#define	PACKET3_DRAW_INDEX_MULTI_AUTO			0x30
+#define	PACKET3_INDIRECT_BUFFER_CONST			0x33
+#define	PACKET3_STRMOUT_BUFFER_UPDATE			0x34
+#define	PACKET3_DRAW_INDEX_OFFSET_2			0x35
+#define	PACKET3_DRAW_PREAMBLE				0x36
+#define	PACKET3_WRITE_DATA				0x37
+#define		WRITE_DATA_DST_SEL(x)                   ((x) << 8)
+                /* 0 - register
+		 * 1 - memory (sync - via GRBM)
+		 * 2 - gl2
+		 * 3 - gds
+		 * 4 - reserved
+		 * 5 - memory (async - direct)
+		 */
+#define		WR_ONE_ADDR                             (1 << 16)
+#define		WR_CONFIRM                              (1 << 20)
+#define		WRITE_DATA_CACHE_POLICY(x)              ((x) << 25)
+                /* 0 - LRU
+		 * 1 - Stream
+		 */
+#define		WRITE_DATA_ENGINE_SEL(x)                ((x) << 30)
+                /* 0 - me
+		 * 1 - pfp
+		 * 2 - ce
+		 */
+#define	PACKET3_DRAW_INDEX_INDIRECT_MULTI		0x38
+#define	PACKET3_MEM_SEMAPHORE				0x39
+#              define PACKET3_SEM_USE_MAILBOX       (0x1 << 16)
+#              define PACKET3_SEM_SEL_SIGNAL_TYPE   (0x1 << 20) /* 0 = increment, 1 = write 1 */
+#              define PACKET3_SEM_CLIENT_CODE	    ((x) << 24) /* 0 = CP, 1 = CB, 2 = DB */
+#              define PACKET3_SEM_SEL_SIGNAL	    (0x6 << 29)
+#              define PACKET3_SEM_SEL_WAIT	    (0x7 << 29)
+#define	PACKET3_COPY_DW					0x3B
+#define	PACKET3_WAIT_REG_MEM				0x3C
+#define		WAIT_REG_MEM_FUNCTION(x)                ((x) << 0)
+                /* 0 - always
+		 * 1 - <
+		 * 2 - <=
+		 * 3 - ==
+		 * 4 - !=
+		 * 5 - >=
+		 * 6 - >
+		 */
+#define		WAIT_REG_MEM_MEM_SPACE(x)               ((x) << 4)
+                /* 0 - reg
+		 * 1 - mem
+		 */
+#define		WAIT_REG_MEM_OPERATION(x)               ((x) << 6)
+                /* 0 - wait_reg_mem
+		 * 1 - wr_wait_wr_reg
+		 */
+#define		WAIT_REG_MEM_ENGINE(x)                  ((x) << 8)
+                /* 0 - me
+		 * 1 - pfp
+		 */
+#define	PACKET3_INDIRECT_BUFFER				0x3F
+#define		INDIRECT_BUFFER_TCL2_VOLATILE           (1 << 22)
+#define		INDIRECT_BUFFER_VALID                   (1 << 23)
+#define		INDIRECT_BUFFER_CACHE_POLICY(x)         ((x) << 28)
+                /* 0 - LRU
+		 * 1 - Stream
+		 * 2 - Bypass
+		 */
+#define	PACKET3_COPY_DATA				0x40
+#define	PACKET3_PFP_SYNC_ME				0x42
+#define	PACKET3_SURFACE_SYNC				0x43
+#              define PACKET3_DEST_BASE_0_ENA      (1 << 0)
+#              define PACKET3_DEST_BASE_1_ENA      (1 << 1)
+#              define PACKET3_CB0_DEST_BASE_ENA    (1 << 6)
+#              define PACKET3_CB1_DEST_BASE_ENA    (1 << 7)
+#              define PACKET3_CB2_DEST_BASE_ENA    (1 << 8)
+#              define PACKET3_CB3_DEST_BASE_ENA    (1 << 9)
+#              define PACKET3_CB4_DEST_BASE_ENA    (1 << 10)
+#              define PACKET3_CB5_DEST_BASE_ENA    (1 << 11)
+#              define PACKET3_CB6_DEST_BASE_ENA    (1 << 12)
+#              define PACKET3_CB7_DEST_BASE_ENA    (1 << 13)
+#              define PACKET3_DB_DEST_BASE_ENA     (1 << 14)
+#              define PACKET3_TCL1_VOL_ACTION_ENA  (1 << 15)
+#              define PACKET3_TC_VOL_ACTION_ENA    (1 << 16) /* L2 */
+#              define PACKET3_TC_WB_ACTION_ENA     (1 << 18) /* L2 */
+#              define PACKET3_DEST_BASE_2_ENA      (1 << 19)
+#              define PACKET3_DEST_BASE_3_ENA      (1 << 21)
+#              define PACKET3_TCL1_ACTION_ENA      (1 << 22)
+#              define PACKET3_TC_ACTION_ENA        (1 << 23) /* L2 */
+#              define PACKET3_CB_ACTION_ENA        (1 << 25)
+#              define PACKET3_DB_ACTION_ENA        (1 << 26)
+#              define PACKET3_SH_KCACHE_ACTION_ENA (1 << 27)
+#              define PACKET3_SH_KCACHE_VOL_ACTION_ENA (1 << 28)
+#              define PACKET3_SH_ICACHE_ACTION_ENA (1 << 29)
+#define	PACKET3_COND_WRITE				0x45
+#define	PACKET3_EVENT_WRITE				0x46
+#define		EVENT_TYPE(x)                           ((x) << 0)
+#define		EVENT_INDEX(x)                          ((x) << 8)
+                /* 0 - any non-TS event
+		 * 1 - ZPASS_DONE, PIXEL_PIPE_STAT_*
+		 * 2 - SAMPLE_PIPELINESTAT
+		 * 3 - SAMPLE_STREAMOUTSTAT*
+		 * 4 - *S_PARTIAL_FLUSH
+		 * 5 - EOP events
+		 * 6 - EOS events
+		 */
+#define	PACKET3_EVENT_WRITE_EOP				0x47
+#define		EOP_TCL1_VOL_ACTION_EN                  (1 << 12)
+#define		EOP_TC_VOL_ACTION_EN                    (1 << 13) /* L2 */
+#define		EOP_TC_WB_ACTION_EN                     (1 << 15) /* L2 */
+#define		EOP_TCL1_ACTION_EN                      (1 << 16)
+#define		EOP_TC_ACTION_EN                        (1 << 17) /* L2 */
+#define		EOP_CACHE_POLICY(x)                     ((x) << 25)
+                /* 0 - LRU
+		 * 1 - Stream
+		 * 2 - Bypass
+		 */
+#define		EOP_TCL2_VOLATILE                       (1 << 27)
+#define		DATA_SEL(x)                             ((x) << 29)
+                /* 0 - discard
+		 * 1 - send low 32bit data
+		 * 2 - send 64bit data
+		 * 3 - send 64bit GPU counter value
+		 * 4 - send 64bit sys counter value
+		 */
+#define		INT_SEL(x)                              ((x) << 24)
+                /* 0 - none
+		 * 1 - interrupt only (DATA_SEL = 0)
+		 * 2 - interrupt when data write is confirmed
+		 */
+#define		DST_SEL(x)                              ((x) << 16)
+                /* 0 - MC
+		 * 1 - TC/L2
+		 */
+#define	PACKET3_EVENT_WRITE_EOS				0x48
+#define	PACKET3_RELEASE_MEM				0x49
+#define	PACKET3_PREAMBLE_CNTL				0x4A
+#              define PACKET3_PREAMBLE_BEGIN_CLEAR_STATE     (2 << 28)
+#              define PACKET3_PREAMBLE_END_CLEAR_STATE       (3 << 28)
+#define	PACKET3_DMA_DATA				0x50
+#define	PACKET3_AQUIRE_MEM				0x58
+#define	PACKET3_REWIND					0x59
+#define	PACKET3_LOAD_UCONFIG_REG			0x5E
+#define	PACKET3_LOAD_SH_REG				0x5F
+#define	PACKET3_LOAD_CONFIG_REG				0x60
+#define	PACKET3_LOAD_CONTEXT_REG			0x61
+#define	PACKET3_SET_CONFIG_REG				0x68
+#define		PACKET3_SET_CONFIG_REG_START			0x00008000
+#define		PACKET3_SET_CONFIG_REG_END			0x0000b000
+#define	PACKET3_SET_CONTEXT_REG				0x69
+#define		PACKET3_SET_CONTEXT_REG_START			0x00028000
+#define		PACKET3_SET_CONTEXT_REG_END			0x00029000
+#define	PACKET3_SET_CONTEXT_REG_INDIRECT		0x73
+#define	PACKET3_SET_SH_REG				0x76
+#define		PACKET3_SET_SH_REG_START			0x0000b000
+#define		PACKET3_SET_SH_REG_END				0x0000c000
+#define	PACKET3_SET_SH_REG_OFFSET			0x77
+#define	PACKET3_SET_QUEUE_REG				0x78
+#define	PACKET3_SET_UCONFIG_REG				0x79
+#define		PACKET3_SET_UCONFIG_REG_START			0x00030000
+#define		PACKET3_SET_UCONFIG_REG_END			0x00031000
+#define	PACKET3_SCRATCH_RAM_WRITE			0x7D
+#define	PACKET3_SCRATCH_RAM_READ			0x7E
+#define	PACKET3_LOAD_CONST_RAM				0x80
+#define	PACKET3_WRITE_CONST_RAM				0x81
+#define	PACKET3_DUMP_CONST_RAM				0x83
+#define	PACKET3_INCREMENT_CE_COUNTER			0x84
+#define	PACKET3_INCREMENT_DE_COUNTER			0x85
+#define	PACKET3_WAIT_ON_CE_COUNTER			0x86
+#define	PACKET3_WAIT_ON_DE_COUNTER_DIFF			0x88
+#define	PACKET3_SWITCH_BUFFER				0x8B
+
+/* SDMA - first instance at 0xd000, second at 0xd800 */
+#define SDMA0_REGISTER_OFFSET                             0x0 /* not a register */
+#define SDMA1_REGISTER_OFFSET                             0x800 /* not a register */
+
+#define	SDMA0_UCODE_ADDR                                  0xD000
+#define	SDMA0_UCODE_DATA                                  0xD004
+
+#define SDMA0_CNTL                                        0xD010
+#       define TRAP_ENABLE                                (1 << 0)
+#       define SEM_INCOMPLETE_INT_ENABLE                  (1 << 1)
+#       define SEM_WAIT_INT_ENABLE                        (1 << 2)
+#       define DATA_SWAP_ENABLE                           (1 << 3)
+#       define FENCE_SWAP_ENABLE                          (1 << 4)
+#       define AUTO_CTXSW_ENABLE                          (1 << 18)
+#       define CTXEMPTY_INT_ENABLE                        (1 << 28)
+
+#define SDMA0_TILING_CONFIG  				  0xD018
+
+#define SDMA0_SEM_INCOMPLETE_TIMER_CNTL                   0xD020
+#define SDMA0_SEM_WAIT_FAIL_TIMER_CNTL                    0xD024
+
+#define SDMA0_STATUS_REG                                  0xd034
+#       define SDMA_IDLE                                  (1 << 0)
+
+#define SDMA0_ME_CNTL                                     0xD048
+#       define SDMA_HALT                                  (1 << 0)
+
+#define SDMA0_GFX_RB_CNTL                                 0xD200
+#       define SDMA_RB_ENABLE                             (1 << 0)
+#       define SDMA_RB_SIZE(x)                            ((x) << 1) /* log2 */
+#       define SDMA_RB_SWAP_ENABLE                        (1 << 9) /* 8IN32 */
+#       define SDMA_RPTR_WRITEBACK_ENABLE                 (1 << 12)
+#       define SDMA_RPTR_WRITEBACK_SWAP_ENABLE            (1 << 13)  /* 8IN32 */
+#       define SDMA_RPTR_WRITEBACK_TIMER(x)               ((x) << 16) /* log2 */
+#define SDMA0_GFX_RB_BASE                                 0xD204
+#define SDMA0_GFX_RB_BASE_HI                              0xD208
+#define SDMA0_GFX_RB_RPTR                                 0xD20C
+#define SDMA0_GFX_RB_WPTR                                 0xD210
+
+#define SDMA0_GFX_RB_RPTR_ADDR_HI                         0xD220
+#define SDMA0_GFX_RB_RPTR_ADDR_LO                         0xD224
+#define SDMA0_GFX_IB_CNTL                                 0xD228
+#       define SDMA_IB_ENABLE                             (1 << 0)
+#       define SDMA_IB_SWAP_ENABLE                        (1 << 4)
+#       define SDMA_SWITCH_INSIDE_IB                      (1 << 8)
+#       define SDMA_CMD_VMID(x)                           ((x) << 16)
+
+#define SDMA0_GFX_VIRTUAL_ADDR                            0xD29C
+#define SDMA0_GFX_APE1_CNTL                               0xD2A0
+
+#define SDMA_PACKET(op, sub_op, e)	((((e) & 0xFFFF) << 16) |	\
+					 (((sub_op) & 0xFF) << 8) |	\
+					 (((op) & 0xFF) << 0))
+/* sDMA opcodes */
+#define	SDMA_OPCODE_NOP					  0
+#define	SDMA_OPCODE_COPY				  1
+#       define SDMA_COPY_SUB_OPCODE_LINEAR                0
+#       define SDMA_COPY_SUB_OPCODE_TILED                 1
+#       define SDMA_COPY_SUB_OPCODE_SOA                   3
+#       define SDMA_COPY_SUB_OPCODE_LINEAR_SUB_WINDOW     4
+#       define SDMA_COPY_SUB_OPCODE_TILED_SUB_WINDOW      5
+#       define SDMA_COPY_SUB_OPCODE_T2T_SUB_WINDOW        6
+#define	SDMA_OPCODE_WRITE				  2
+#       define SDMA_WRITE_SUB_OPCODE_LINEAR               0
+#       define SDMA_WRTIE_SUB_OPCODE_TILED                1
+#define	SDMA_OPCODE_INDIRECT_BUFFER			  4
+#define	SDMA_OPCODE_FENCE				  5
+#define	SDMA_OPCODE_TRAP				  6
+#define	SDMA_OPCODE_SEMAPHORE				  7
+#       define SDMA_SEMAPHORE_EXTRA_O                     (1 << 13)
+                /* 0 - increment
+		 * 1 - write 1
+		 */
+#       define SDMA_SEMAPHORE_EXTRA_S                     (1 << 14)
+                /* 0 - wait
+		 * 1 - signal
+		 */
+#       define SDMA_SEMAPHORE_EXTRA_M                     (1 << 15)
+                /* mailbox */
+#define	SDMA_OPCODE_POLL_REG_MEM			  8
+#       define SDMA_POLL_REG_MEM_EXTRA_OP(x)              ((x) << 10)
+                /* 0 - wait_reg_mem
+		 * 1 - wr_wait_wr_reg
+		 */
+#       define SDMA_POLL_REG_MEM_EXTRA_FUNC(x)            ((x) << 12)
+                /* 0 - always
+		 * 1 - <
+		 * 2 - <=
+		 * 3 - ==
+		 * 4 - !=
+		 * 5 - >=
+		 * 6 - >
+		 */
+#       define SDMA_POLL_REG_MEM_EXTRA_M                  (1 << 15)
+                /* 0 = register
+		 * 1 = memory
+		 */
+#define	SDMA_OPCODE_COND_EXEC				  9
+#define	SDMA_OPCODE_CONSTANT_FILL			  11
+#       define SDMA_CONSTANT_FILL_EXTRA_SIZE(x)           ((x) << 14)
+                /* 0 = byte fill
+		 * 2 = DW fill
+		 */
+#define	SDMA_OPCODE_GENERATE_PTE_PDE			  12
+#define	SDMA_OPCODE_TIMESTAMP				  13
+#       define SDMA_TIMESTAMP_SUB_OPCODE_SET_LOCAL        0
+#       define SDMA_TIMESTAMP_SUB_OPCODE_GET_LOCAL        1
+#       define SDMA_TIMESTAMP_SUB_OPCODE_GET_GLOBAL       2
+#define	SDMA_OPCODE_SRBM_WRITE				  14
+#       define SDMA_SRBM_WRITE_EXTRA_BYTE_ENABLE(x)       ((x) << 12)
+                /* byte mask */
+
+/* UVD */
+
+#define UVD_UDEC_ADDR_CONFIG		0xef4c
+#define UVD_UDEC_DB_ADDR_CONFIG		0xef50
+#define UVD_UDEC_DBW_ADDR_CONFIG	0xef54
+
+#define UVD_LMI_EXT40_ADDR		0xf498
+#define UVD_LMI_ADDR_EXT		0xf594
+#define UVD_VCPU_CACHE_OFFSET0		0xf608
+#define UVD_VCPU_CACHE_SIZE0		0xf60c
+#define UVD_VCPU_CACHE_OFFSET1		0xf610
+#define UVD_VCPU_CACHE_SIZE1		0xf614
+#define UVD_VCPU_CACHE_OFFSET2		0xf618
+#define UVD_VCPU_CACHE_SIZE2		0xf61c
+
+#define UVD_RBC_RB_RPTR			0xf690
+#define UVD_RBC_RB_WPTR			0xf694
+
+/* UVD clocks */
+
+#define CG_DCLK_CNTL			0xC050009C
+#	define DCLK_DIVIDER_MASK	0x7f
+#	define DCLK_DIR_CNTL_EN		(1 << 8)
+#define CG_DCLK_STATUS			0xC05000A0
+#	define DCLK_STATUS		(1 << 0)
+#define CG_VCLK_CNTL			0xC05000A4
+#define CG_VCLK_STATUS			0xC05000A8
+
+#endif
diff --git a/drivers/gpu/drm/radeon/clearstate_cayman.h b/drivers/gpu/drm/radeon/clearstate_cayman.h
new file mode 100644
index 000000000000..c00339440c5e
--- /dev/null
+++ b/drivers/gpu/drm/radeon/clearstate_cayman.h
@@ -0,0 +1,1081 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+static const u32 SECT_CONTEXT_def_1[] =
+{
+    0x00000000, // DB_RENDER_CONTROL
+    0x00000000, // DB_COUNT_CONTROL
+    0x00000000, // DB_DEPTH_VIEW
+    0x00000000, // DB_RENDER_OVERRIDE
+    0x00000000, // DB_RENDER_OVERRIDE2
+    0x00000000, // DB_HTILE_DATA_BASE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_STENCIL_CLEAR
+    0x00000000, // DB_DEPTH_CLEAR
+    0x00000000, // PA_SC_SCREEN_SCISSOR_TL
+    0x40004000, // PA_SC_SCREEN_SCISSOR_BR
+    0, // HOLE
+    0x00000000, // DB_DEPTH_INFO
+    0x00000000, // DB_Z_INFO
+    0x00000000, // DB_STENCIL_INFO
+    0x00000000, // DB_Z_READ_BASE
+    0x00000000, // DB_STENCIL_READ_BASE
+    0x00000000, // DB_Z_WRITE_BASE
+    0x00000000, // DB_STENCIL_WRITE_BASE
+    0x00000000, // DB_DEPTH_SIZE
+    0x00000000, // DB_DEPTH_SLICE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_15
+    0x00000000, // PA_SC_WINDOW_OFFSET
+    0x80000000, // PA_SC_WINDOW_SCISSOR_TL
+    0x40004000, // PA_SC_WINDOW_SCISSOR_BR
+    0x0000ffff, // PA_SC_CLIPRECT_RULE
+    0x00000000, // PA_SC_CLIPRECT_0_TL
+    0x40004000, // PA_SC_CLIPRECT_0_BR
+    0x00000000, // PA_SC_CLIPRECT_1_TL
+    0x40004000, // PA_SC_CLIPRECT_1_BR
+    0x00000000, // PA_SC_CLIPRECT_2_TL
+    0x40004000, // PA_SC_CLIPRECT_2_BR
+    0x00000000, // PA_SC_CLIPRECT_3_TL
+    0x40004000, // PA_SC_CLIPRECT_3_BR
+    0xaa99aaaa, // PA_SC_EDGERULE
+    0x00000000, // PA_SU_HARDWARE_SCREEN_OFFSET
+    0xffffffff, // CB_TARGET_MASK
+    0xffffffff, // CB_SHADER_MASK
+    0x80000000, // PA_SC_GENERIC_SCISSOR_TL
+    0x40004000, // PA_SC_GENERIC_SCISSOR_BR
+    0x00000000, // COHER_DEST_BASE_0
+    0x00000000, // COHER_DEST_BASE_1
+    0x80000000, // PA_SC_VPORT_SCISSOR_0_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_0_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_1_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_1_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_2_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_2_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_3_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_3_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_4_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_4_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_5_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_5_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_6_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_6_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_7_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_7_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_8_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_8_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_9_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_9_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_10_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_10_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_11_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_11_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_12_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_12_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_13_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_13_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_14_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_14_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_15_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_15_BR
+    0x00000000, // PA_SC_VPORT_ZMIN_0
+    0x3f800000, // PA_SC_VPORT_ZMAX_0
+    0x00000000, // PA_SC_VPORT_ZMIN_1
+    0x3f800000, // PA_SC_VPORT_ZMAX_1
+    0x00000000, // PA_SC_VPORT_ZMIN_2
+    0x3f800000, // PA_SC_VPORT_ZMAX_2
+    0x00000000, // PA_SC_VPORT_ZMIN_3
+    0x3f800000, // PA_SC_VPORT_ZMAX_3
+    0x00000000, // PA_SC_VPORT_ZMIN_4
+    0x3f800000, // PA_SC_VPORT_ZMAX_4
+    0x00000000, // PA_SC_VPORT_ZMIN_5
+    0x3f800000, // PA_SC_VPORT_ZMAX_5
+    0x00000000, // PA_SC_VPORT_ZMIN_6
+    0x3f800000, // PA_SC_VPORT_ZMAX_6
+    0x00000000, // PA_SC_VPORT_ZMIN_7
+    0x3f800000, // PA_SC_VPORT_ZMAX_7
+    0x00000000, // PA_SC_VPORT_ZMIN_8
+    0x3f800000, // PA_SC_VPORT_ZMAX_8
+    0x00000000, // PA_SC_VPORT_ZMIN_9
+    0x3f800000, // PA_SC_VPORT_ZMAX_9
+    0x00000000, // PA_SC_VPORT_ZMIN_10
+    0x3f800000, // PA_SC_VPORT_ZMAX_10
+    0x00000000, // PA_SC_VPORT_ZMIN_11
+    0x3f800000, // PA_SC_VPORT_ZMAX_11
+    0x00000000, // PA_SC_VPORT_ZMIN_12
+    0x3f800000, // PA_SC_VPORT_ZMAX_12
+    0x00000000, // PA_SC_VPORT_ZMIN_13
+    0x3f800000, // PA_SC_VPORT_ZMAX_13
+    0x00000000, // PA_SC_VPORT_ZMIN_14
+    0x3f800000, // PA_SC_VPORT_ZMAX_14
+    0x00000000, // PA_SC_VPORT_ZMIN_15
+    0x3f800000, // PA_SC_VPORT_ZMAX_15
+    0x00000000, // SX_MISC
+    0x00000000, // SX_SURFACE_SYNC
+    0x00000000, // SX_SCATTER_EXPORT_BASE
+    0x00000000, // SX_SCATTER_EXPORT_SIZE
+    0x00000000, // CP_PERFMON_CNTX_CNTL
+    0x00000000, // CP_RINGID
+    0x00000000, // CP_VMID
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_VTX_SEMANTIC_0
+    0x00000000, // SQ_VTX_SEMANTIC_1
+    0x00000000, // SQ_VTX_SEMANTIC_2
+    0x00000000, // SQ_VTX_SEMANTIC_3
+    0x00000000, // SQ_VTX_SEMANTIC_4
+    0x00000000, // SQ_VTX_SEMANTIC_5
+    0x00000000, // SQ_VTX_SEMANTIC_6
+    0x00000000, // SQ_VTX_SEMANTIC_7
+    0x00000000, // SQ_VTX_SEMANTIC_8
+    0x00000000, // SQ_VTX_SEMANTIC_9
+    0x00000000, // SQ_VTX_SEMANTIC_10
+    0x00000000, // SQ_VTX_SEMANTIC_11
+    0x00000000, // SQ_VTX_SEMANTIC_12
+    0x00000000, // SQ_VTX_SEMANTIC_13
+    0x00000000, // SQ_VTX_SEMANTIC_14
+    0x00000000, // SQ_VTX_SEMANTIC_15
+    0x00000000, // SQ_VTX_SEMANTIC_16
+    0x00000000, // SQ_VTX_SEMANTIC_17
+    0x00000000, // SQ_VTX_SEMANTIC_18
+    0x00000000, // SQ_VTX_SEMANTIC_19
+    0x00000000, // SQ_VTX_SEMANTIC_20
+    0x00000000, // SQ_VTX_SEMANTIC_21
+    0x00000000, // SQ_VTX_SEMANTIC_22
+    0x00000000, // SQ_VTX_SEMANTIC_23
+    0x00000000, // SQ_VTX_SEMANTIC_24
+    0x00000000, // SQ_VTX_SEMANTIC_25
+    0x00000000, // SQ_VTX_SEMANTIC_26
+    0x00000000, // SQ_VTX_SEMANTIC_27
+    0x00000000, // SQ_VTX_SEMANTIC_28
+    0x00000000, // SQ_VTX_SEMANTIC_29
+    0x00000000, // SQ_VTX_SEMANTIC_30
+    0x00000000, // SQ_VTX_SEMANTIC_31
+    0xffffffff, // VGT_MAX_VTX_INDX
+    0x00000000, // VGT_MIN_VTX_INDX
+    0x00000000, // VGT_INDX_OFFSET
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_INDX
+    0x00000000, // SX_ALPHA_TEST_CONTROL
+    0x00000000, // CB_BLEND_RED
+    0x00000000, // CB_BLEND_GREEN
+    0x00000000, // CB_BLEND_BLUE
+    0x00000000, // CB_BLEND_ALPHA
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_STENCILREFMASK
+    0x00000000, // DB_STENCILREFMASK_BF
+    0x00000000, // SX_ALPHA_REF
+    0x00000000, // PA_CL_VPORT_XSCALE
+    0x00000000, // PA_CL_VPORT_XOFFSET
+    0x00000000, // PA_CL_VPORT_YSCALE
+    0x00000000, // PA_CL_VPORT_YOFFSET
+    0x00000000, // PA_CL_VPORT_ZSCALE
+    0x00000000, // PA_CL_VPORT_ZOFFSET
+    0x00000000, // PA_CL_VPORT_XSCALE_1
+    0x00000000, // PA_CL_VPORT_XOFFSET_1
+    0x00000000, // PA_CL_VPORT_YSCALE_1
+    0x00000000, // PA_CL_VPORT_YOFFSET_1
+    0x00000000, // PA_CL_VPORT_ZSCALE_1
+    0x00000000, // PA_CL_VPORT_ZOFFSET_1
+    0x00000000, // PA_CL_VPORT_XSCALE_2
+    0x00000000, // PA_CL_VPORT_XOFFSET_2
+    0x00000000, // PA_CL_VPORT_YSCALE_2
+    0x00000000, // PA_CL_VPORT_YOFFSET_2
+    0x00000000, // PA_CL_VPORT_ZSCALE_2
+    0x00000000, // PA_CL_VPORT_ZOFFSET_2
+    0x00000000, // PA_CL_VPORT_XSCALE_3
+    0x00000000, // PA_CL_VPORT_XOFFSET_3
+    0x00000000, // PA_CL_VPORT_YSCALE_3
+    0x00000000, // PA_CL_VPORT_YOFFSET_3
+    0x00000000, // PA_CL_VPORT_ZSCALE_3
+    0x00000000, // PA_CL_VPORT_ZOFFSET_3
+    0x00000000, // PA_CL_VPORT_XSCALE_4
+    0x00000000, // PA_CL_VPORT_XOFFSET_4
+    0x00000000, // PA_CL_VPORT_YSCALE_4
+    0x00000000, // PA_CL_VPORT_YOFFSET_4
+    0x00000000, // PA_CL_VPORT_ZSCALE_4
+    0x00000000, // PA_CL_VPORT_ZOFFSET_4
+    0x00000000, // PA_CL_VPORT_XSCALE_5
+    0x00000000, // PA_CL_VPORT_XOFFSET_5
+    0x00000000, // PA_CL_VPORT_YSCALE_5
+    0x00000000, // PA_CL_VPORT_YOFFSET_5
+    0x00000000, // PA_CL_VPORT_ZSCALE_5
+    0x00000000, // PA_CL_VPORT_ZOFFSET_5
+    0x00000000, // PA_CL_VPORT_XSCALE_6
+    0x00000000, // PA_CL_VPORT_XOFFSET_6
+    0x00000000, // PA_CL_VPORT_YSCALE_6
+    0x00000000, // PA_CL_VPORT_YOFFSET_6
+    0x00000000, // PA_CL_VPORT_ZSCALE_6
+    0x00000000, // PA_CL_VPORT_ZOFFSET_6
+    0x00000000, // PA_CL_VPORT_XSCALE_7
+    0x00000000, // PA_CL_VPORT_XOFFSET_7
+    0x00000000, // PA_CL_VPORT_YSCALE_7
+    0x00000000, // PA_CL_VPORT_YOFFSET_7
+    0x00000000, // PA_CL_VPORT_ZSCALE_7
+    0x00000000, // PA_CL_VPORT_ZOFFSET_7
+    0x00000000, // PA_CL_VPORT_XSCALE_8
+    0x00000000, // PA_CL_VPORT_XOFFSET_8
+    0x00000000, // PA_CL_VPORT_YSCALE_8
+    0x00000000, // PA_CL_VPORT_YOFFSET_8
+    0x00000000, // PA_CL_VPORT_ZSCALE_8
+    0x00000000, // PA_CL_VPORT_ZOFFSET_8
+    0x00000000, // PA_CL_VPORT_XSCALE_9
+    0x00000000, // PA_CL_VPORT_XOFFSET_9
+    0x00000000, // PA_CL_VPORT_YSCALE_9
+    0x00000000, // PA_CL_VPORT_YOFFSET_9
+    0x00000000, // PA_CL_VPORT_ZSCALE_9
+    0x00000000, // PA_CL_VPORT_ZOFFSET_9
+    0x00000000, // PA_CL_VPORT_XSCALE_10
+    0x00000000, // PA_CL_VPORT_XOFFSET_10
+    0x00000000, // PA_CL_VPORT_YSCALE_10
+    0x00000000, // PA_CL_VPORT_YOFFSET_10
+    0x00000000, // PA_CL_VPORT_ZSCALE_10
+    0x00000000, // PA_CL_VPORT_ZOFFSET_10
+    0x00000000, // PA_CL_VPORT_XSCALE_11
+    0x00000000, // PA_CL_VPORT_XOFFSET_11
+    0x00000000, // PA_CL_VPORT_YSCALE_11
+    0x00000000, // PA_CL_VPORT_YOFFSET_11
+    0x00000000, // PA_CL_VPORT_ZSCALE_11
+    0x00000000, // PA_CL_VPORT_ZOFFSET_11
+    0x00000000, // PA_CL_VPORT_XSCALE_12
+    0x00000000, // PA_CL_VPORT_XOFFSET_12
+    0x00000000, // PA_CL_VPORT_YSCALE_12
+    0x00000000, // PA_CL_VPORT_YOFFSET_12
+    0x00000000, // PA_CL_VPORT_ZSCALE_12
+    0x00000000, // PA_CL_VPORT_ZOFFSET_12
+    0x00000000, // PA_CL_VPORT_XSCALE_13
+    0x00000000, // PA_CL_VPORT_XOFFSET_13
+    0x00000000, // PA_CL_VPORT_YSCALE_13
+    0x00000000, // PA_CL_VPORT_YOFFSET_13
+    0x00000000, // PA_CL_VPORT_ZSCALE_13
+    0x00000000, // PA_CL_VPORT_ZOFFSET_13
+    0x00000000, // PA_CL_VPORT_XSCALE_14
+    0x00000000, // PA_CL_VPORT_XOFFSET_14
+    0x00000000, // PA_CL_VPORT_YSCALE_14
+    0x00000000, // PA_CL_VPORT_YOFFSET_14
+    0x00000000, // PA_CL_VPORT_ZSCALE_14
+    0x00000000, // PA_CL_VPORT_ZOFFSET_14
+    0x00000000, // PA_CL_VPORT_XSCALE_15
+    0x00000000, // PA_CL_VPORT_XOFFSET_15
+    0x00000000, // PA_CL_VPORT_YSCALE_15
+    0x00000000, // PA_CL_VPORT_YOFFSET_15
+    0x00000000, // PA_CL_VPORT_ZSCALE_15
+    0x00000000, // PA_CL_VPORT_ZOFFSET_15
+    0x00000000, // PA_CL_UCP_0_X
+    0x00000000, // PA_CL_UCP_0_Y
+    0x00000000, // PA_CL_UCP_0_Z
+    0x00000000, // PA_CL_UCP_0_W
+    0x00000000, // PA_CL_UCP_1_X
+    0x00000000, // PA_CL_UCP_1_Y
+    0x00000000, // PA_CL_UCP_1_Z
+    0x00000000, // PA_CL_UCP_1_W
+    0x00000000, // PA_CL_UCP_2_X
+    0x00000000, // PA_CL_UCP_2_Y
+    0x00000000, // PA_CL_UCP_2_Z
+    0x00000000, // PA_CL_UCP_2_W
+    0x00000000, // PA_CL_UCP_3_X
+    0x00000000, // PA_CL_UCP_3_Y
+    0x00000000, // PA_CL_UCP_3_Z
+    0x00000000, // PA_CL_UCP_3_W
+    0x00000000, // PA_CL_UCP_4_X
+    0x00000000, // PA_CL_UCP_4_Y
+    0x00000000, // PA_CL_UCP_4_Z
+    0x00000000, // PA_CL_UCP_4_W
+    0x00000000, // PA_CL_UCP_5_X
+    0x00000000, // PA_CL_UCP_5_Y
+    0x00000000, // PA_CL_UCP_5_Z
+    0x00000000, // PA_CL_UCP_5_W
+    0x00000000, // SPI_VS_OUT_ID_0
+    0x00000000, // SPI_VS_OUT_ID_1
+    0x00000000, // SPI_VS_OUT_ID_2
+    0x00000000, // SPI_VS_OUT_ID_3
+    0x00000000, // SPI_VS_OUT_ID_4
+    0x00000000, // SPI_VS_OUT_ID_5
+    0x00000000, // SPI_VS_OUT_ID_6
+    0x00000000, // SPI_VS_OUT_ID_7
+    0x00000000, // SPI_VS_OUT_ID_8
+    0x00000000, // SPI_VS_OUT_ID_9
+    0x00000000, // SPI_PS_INPUT_CNTL_0
+    0x00000000, // SPI_PS_INPUT_CNTL_1
+    0x00000000, // SPI_PS_INPUT_CNTL_2
+    0x00000000, // SPI_PS_INPUT_CNTL_3
+    0x00000000, // SPI_PS_INPUT_CNTL_4
+    0x00000000, // SPI_PS_INPUT_CNTL_5
+    0x00000000, // SPI_PS_INPUT_CNTL_6
+    0x00000000, // SPI_PS_INPUT_CNTL_7
+    0x00000000, // SPI_PS_INPUT_CNTL_8
+    0x00000000, // SPI_PS_INPUT_CNTL_9
+    0x00000000, // SPI_PS_INPUT_CNTL_10
+    0x00000000, // SPI_PS_INPUT_CNTL_11
+    0x00000000, // SPI_PS_INPUT_CNTL_12
+    0x00000000, // SPI_PS_INPUT_CNTL_13
+    0x00000000, // SPI_PS_INPUT_CNTL_14
+    0x00000000, // SPI_PS_INPUT_CNTL_15
+    0x00000000, // SPI_PS_INPUT_CNTL_16
+    0x00000000, // SPI_PS_INPUT_CNTL_17
+    0x00000000, // SPI_PS_INPUT_CNTL_18
+    0x00000000, // SPI_PS_INPUT_CNTL_19
+    0x00000000, // SPI_PS_INPUT_CNTL_20
+    0x00000000, // SPI_PS_INPUT_CNTL_21
+    0x00000000, // SPI_PS_INPUT_CNTL_22
+    0x00000000, // SPI_PS_INPUT_CNTL_23
+    0x00000000, // SPI_PS_INPUT_CNTL_24
+    0x00000000, // SPI_PS_INPUT_CNTL_25
+    0x00000000, // SPI_PS_INPUT_CNTL_26
+    0x00000000, // SPI_PS_INPUT_CNTL_27
+    0x00000000, // SPI_PS_INPUT_CNTL_28
+    0x00000000, // SPI_PS_INPUT_CNTL_29
+    0x00000000, // SPI_PS_INPUT_CNTL_30
+    0x00000000, // SPI_PS_INPUT_CNTL_31
+    0x00000000, // SPI_VS_OUT_CONFIG
+    0x00000001, // SPI_THREAD_GROUPING
+    0x00000002, // SPI_PS_IN_CONTROL_0
+    0x00000000, // SPI_PS_IN_CONTROL_1
+    0x00000000, // SPI_INTERP_CONTROL_0
+    0x00000000, // SPI_INPUT_Z
+    0x00000000, // SPI_FOG_CNTL
+    0x00000000, // SPI_BARYC_CNTL
+    0x00000000, // SPI_PS_IN_CONTROL_2
+    0x00000000, // SPI_COMPUTE_INPUT_CNTL
+    0x00000000, // SPI_COMPUTE_NUM_THREAD_X
+    0x00000000, // SPI_COMPUTE_NUM_THREAD_Y
+    0x00000000, // SPI_COMPUTE_NUM_THREAD_Z
+    0x00000000, // SPI_GPR_MGMT
+    0x00000000, // SPI_LDS_MGMT
+    0x00000000, // SPI_STACK_MGMT
+    0x00000000, // SPI_WAVE_MGMT_1
+    0x00000000, // SPI_WAVE_MGMT_2
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // GDS_ADDR_BASE
+    0x00003fff, // GDS_ADDR_SIZE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // GDS_ORDERED_COUNT
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // GDS_APPEND_CONSUME_UAV0
+    0x00000000, // GDS_APPEND_CONSUME_UAV1
+    0x00000000, // GDS_APPEND_CONSUME_UAV2
+    0x00000000, // GDS_APPEND_CONSUME_UAV3
+    0x00000000, // GDS_APPEND_CONSUME_UAV4
+    0x00000000, // GDS_APPEND_CONSUME_UAV5
+    0x00000000, // GDS_APPEND_CONSUME_UAV6
+    0x00000000, // GDS_APPEND_CONSUME_UAV7
+    0x00000000, // GDS_APPEND_CONSUME_UAV8
+    0x00000000, // GDS_APPEND_CONSUME_UAV9
+    0x00000000, // GDS_APPEND_CONSUME_UAV10
+    0x00000000, // GDS_APPEND_CONSUME_UAV11
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_BLEND0_CONTROL
+    0x00000000, // CB_BLEND1_CONTROL
+    0x00000000, // CB_BLEND2_CONTROL
+    0x00000000, // CB_BLEND3_CONTROL
+    0x00000000, // CB_BLEND4_CONTROL
+    0x00000000, // CB_BLEND5_CONTROL
+    0x00000000, // CB_BLEND6_CONTROL
+    0x00000000, // CB_BLEND7_CONTROL
+};
+static const u32 SECT_CONTEXT_def_2[] =
+{
+    0x00000000, // PA_CL_POINT_X_RAD
+    0x00000000, // PA_CL_POINT_Y_RAD
+    0x00000000, // PA_CL_POINT_SIZE
+    0x00000000, // PA_CL_POINT_CULL_RAD
+    0x00000000, // VGT_DMA_BASE_HI
+    0x00000000, // VGT_DMA_BASE
+};
+static const u32 SECT_CONTEXT_def_3[] =
+{
+    0x00000000, // DB_DEPTH_CONTROL
+    0x00000000, // DB_EQAA
+    0x00000000, // CB_COLOR_CONTROL
+    0x00000200, // DB_SHADER_CONTROL
+    0x00000000, // PA_CL_CLIP_CNTL
+    0x00000000, // PA_SU_SC_MODE_CNTL
+    0x00000000, // PA_CL_VTE_CNTL
+    0x00000000, // PA_CL_VS_OUT_CNTL
+    0x00000000, // PA_CL_NANINF_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_SCALE
+    0x00000000, // PA_SU_PRIM_FILTER_CNTL
+    0x00000000, // SQ_LSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_HSTMP_RING_ITEMSIZE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_PS
+    0x00000000, // SQ_PGM_RESOURCES_PS
+    0x00000000, // SQ_PGM_RESOURCES_2_PS
+    0x00000000, // SQ_PGM_EXPORTS_PS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_VS
+    0x00000000, // SQ_PGM_RESOURCES_VS
+    0x00000000, // SQ_PGM_RESOURCES_2_VS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_GS
+    0x00000000, // SQ_PGM_RESOURCES_GS
+    0x00000000, // SQ_PGM_RESOURCES_2_GS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_ES
+    0x00000000, // SQ_PGM_RESOURCES_ES
+    0x00000000, // SQ_PGM_RESOURCES_2_ES
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_FS
+    0x00000000, // SQ_PGM_RESOURCES_FS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_HS
+    0x00000000, // SQ_PGM_RESOURCES_HS
+    0x00000000, // SQ_PGM_RESOURCES_2_HS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_LS
+    0x00000000, // SQ_PGM_RESOURCES_LS
+    0x00000000, // SQ_PGM_RESOURCES_2_LS
+};
+static const u32 SECT_CONTEXT_def_4[] =
+{
+    0x00000000, // SQ_LDS_ALLOC
+    0x00000000, // SQ_LDS_ALLOC_PS
+    0x00000000, // SQ_VTX_SEMANTIC_CLEAR
+    0, // HOLE
+    0x00000000, // SQ_THREAD_TRACE_CTRL
+    0, // HOLE
+    0x00000000, // SQ_ESGS_RING_ITEMSIZE
+    0x00000000, // SQ_GSVS_RING_ITEMSIZE
+    0x00000000, // SQ_ESTMP_RING_ITEMSIZE
+    0x00000000, // SQ_GSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_VSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_PSTMP_RING_ITEMSIZE
+    0, // HOLE
+    0x00000000, // SQ_GS_VERT_ITEMSIZE
+    0x00000000, // SQ_GS_VERT_ITEMSIZE_1
+    0x00000000, // SQ_GS_VERT_ITEMSIZE_2
+    0x00000000, // SQ_GS_VERT_ITEMSIZE_3
+    0x00000000, // SQ_GSVS_RING_OFFSET_1
+    0x00000000, // SQ_GSVS_RING_OFFSET_2
+    0x00000000, // SQ_GSVS_RING_OFFSET_3
+    0x00000000, // SQ_GWS_RING_OFFSET
+    0, // HOLE
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_15
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_15
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_15
+    0x00000000, // PA_SU_POINT_SIZE
+    0x00000000, // PA_SU_POINT_MINMAX
+    0x00000000, // PA_SU_LINE_CNTL
+    0x00000000, // PA_SC_LINE_STIPPLE
+    0x00000000, // VGT_OUTPUT_PATH_CNTL
+    0x00000000, // VGT_HOS_CNTL
+    0x00000000, // VGT_HOS_MAX_TESS_LEVEL
+    0x00000000, // VGT_HOS_MIN_TESS_LEVEL
+    0x00000000, // VGT_HOS_REUSE_DEPTH
+    0x00000000, // VGT_GROUP_PRIM_TYPE
+    0x00000000, // VGT_GROUP_FIRST_DECR
+    0x00000000, // VGT_GROUP_DECR
+    0x00000000, // VGT_GROUP_VECT_0_CNTL
+    0x00000000, // VGT_GROUP_VECT_1_CNTL
+    0x00000000, // VGT_GROUP_VECT_0_FMT_CNTL
+    0x00000000, // VGT_GROUP_VECT_1_FMT_CNTL
+    0x00000000, // VGT_GS_MODE
+    0, // HOLE
+    0x00000000, // PA_SC_MODE_CNTL_0
+    0x00000000, // PA_SC_MODE_CNTL_1
+    0x00000000, // VGT_ENHANCE
+    0x00000100, // VGT_GS_PER_ES
+    0x00000080, // VGT_ES_PER_GS
+    0x00000002, // VGT_GS_PER_VS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_GS_OUT_PRIM_TYPE
+    0x00000000, // IA_ENHANCE
+};
+static const u32 SECT_CONTEXT_def_5[] =
+{
+    0x00000000, // VGT_DMA_MAX_SIZE
+    0x00000000, // VGT_DMA_INDEX_TYPE
+    0, // HOLE
+    0x00000000, // VGT_PRIMITIVEID_EN
+    0x00000000, // VGT_DMA_NUM_INSTANCES
+};
+static const u32 SECT_CONTEXT_def_6[] =
+{
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_EN
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_INSTANCE_STEP_RATE_0
+    0x00000000, // VGT_INSTANCE_STEP_RATE_1
+    0x000000ff, // IA_MULTI_VGT_PARAM
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_REUSE_OFF
+    0x00000000, // VGT_VTX_CNT_EN
+    0x00000000, // DB_HTILE_SURFACE
+    0x00000000, // DB_SRESULTS_COMPARE_STATE0
+    0x00000000, // DB_SRESULTS_COMPARE_STATE1
+    0x00000000, // DB_PRELOAD_CONTROL
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_0
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_0
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_0
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_0
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_1
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_1
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_1
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_1
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_2
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_2
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_2
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_2
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_3
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_3
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_3
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_3
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_0
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_1
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_2
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_3
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_OFFSET
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
+    0, // HOLE
+    0x00000000, // VGT_GS_MAX_VERT_OUT
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_0
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_1
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_2
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_3
+    0x00000000, // VGT_SHADER_STAGES_EN
+    0x00000000, // VGT_LS_HS_CONFIG
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_TF_PARAM
+    0x00000000, // DB_ALPHA_TO_MASK
+};
+static const u32 SECT_CONTEXT_def_7[] =
+{
+    0x00000000, // PA_SU_POLY_OFFSET_DB_FMT_CNTL
+    0x00000000, // PA_SU_POLY_OFFSET_CLAMP
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_OFFSET
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_OFFSET
+    0x00000000, // VGT_GS_INSTANCE_CNT
+    0x00000000, // VGT_STRMOUT_CONFIG
+    0x00000000, // VGT_STRMOUT_BUFFER_CONFIG
+    0x00000000, // CB_IMMED0_BASE
+    0x00000000, // CB_IMMED1_BASE
+    0x00000000, // CB_IMMED2_BASE
+    0x00000000, // CB_IMMED3_BASE
+    0x00000000, // CB_IMMED4_BASE
+    0x00000000, // CB_IMMED5_BASE
+    0x00000000, // CB_IMMED6_BASE
+    0x00000000, // CB_IMMED7_BASE
+    0x00000000, // CB_IMMED8_BASE
+    0x00000000, // CB_IMMED9_BASE
+    0x00000000, // CB_IMMED10_BASE
+    0x00000000, // CB_IMMED11_BASE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // PA_SC_CENTROID_PRIORITY_0
+    0x00000000, // PA_SC_CENTROID_PRIORITY_1
+    0x00001000, // PA_SC_LINE_CNTL
+    0x00000000, // PA_SC_AA_CONFIG
+    0x00000005, // PA_SU_VTX_CNTL
+    0x3f800000, // PA_CL_GB_VERT_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_VERT_DISC_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_DISC_ADJ
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_3
+    0xffffffff, // PA_SC_AA_MASK_X0Y0_X1Y0
+    0xffffffff, // PA_SC_AA_MASK_X0Y1_X1Y1
+    0x00000000, // CB_CLRCMP_CONTROL
+    0x00000000, // CB_CLRCMP_SRC
+    0x00000000, // CB_CLRCMP_DST
+    0x00000000, // CB_CLRCMP_MSK
+    0, // HOLE
+    0, // HOLE
+    0x0000000e, // VGT_VERTEX_REUSE_BLOCK_CNTL
+    0x00000010, // VGT_OUT_DEALLOC_CNTL
+    0x00000000, // CB_COLOR0_BASE
+    0x00000000, // CB_COLOR0_PITCH
+    0x00000000, // CB_COLOR0_SLICE
+    0x00000000, // CB_COLOR0_VIEW
+    0x00000000, // CB_COLOR0_INFO
+    0x00000000, // CB_COLOR0_ATTRIB
+    0x00000000, // CB_COLOR0_DIM
+    0x00000000, // CB_COLOR0_CMASK
+    0x00000000, // CB_COLOR0_CMASK_SLICE
+    0x00000000, // CB_COLOR0_FMASK
+    0x00000000, // CB_COLOR0_FMASK_SLICE
+    0x00000000, // CB_COLOR0_CLEAR_WORD0
+    0x00000000, // CB_COLOR0_CLEAR_WORD1
+    0x00000000, // CB_COLOR0_CLEAR_WORD2
+    0x00000000, // CB_COLOR0_CLEAR_WORD3
+    0x00000000, // CB_COLOR1_BASE
+    0x00000000, // CB_COLOR1_PITCH
+    0x00000000, // CB_COLOR1_SLICE
+    0x00000000, // CB_COLOR1_VIEW
+    0x00000000, // CB_COLOR1_INFO
+    0x00000000, // CB_COLOR1_ATTRIB
+    0x00000000, // CB_COLOR1_DIM
+    0x00000000, // CB_COLOR1_CMASK
+    0x00000000, // CB_COLOR1_CMASK_SLICE
+    0x00000000, // CB_COLOR1_FMASK
+    0x00000000, // CB_COLOR1_FMASK_SLICE
+    0x00000000, // CB_COLOR1_CLEAR_WORD0
+    0x00000000, // CB_COLOR1_CLEAR_WORD1
+    0x00000000, // CB_COLOR1_CLEAR_WORD2
+    0x00000000, // CB_COLOR1_CLEAR_WORD3
+    0x00000000, // CB_COLOR2_BASE
+    0x00000000, // CB_COLOR2_PITCH
+    0x00000000, // CB_COLOR2_SLICE
+    0x00000000, // CB_COLOR2_VIEW
+    0x00000000, // CB_COLOR2_INFO
+    0x00000000, // CB_COLOR2_ATTRIB
+    0x00000000, // CB_COLOR2_DIM
+    0x00000000, // CB_COLOR2_CMASK
+    0x00000000, // CB_COLOR2_CMASK_SLICE
+    0x00000000, // CB_COLOR2_FMASK
+    0x00000000, // CB_COLOR2_FMASK_SLICE
+    0x00000000, // CB_COLOR2_CLEAR_WORD0
+    0x00000000, // CB_COLOR2_CLEAR_WORD1
+    0x00000000, // CB_COLOR2_CLEAR_WORD2
+    0x00000000, // CB_COLOR2_CLEAR_WORD3
+    0x00000000, // CB_COLOR3_BASE
+    0x00000000, // CB_COLOR3_PITCH
+    0x00000000, // CB_COLOR3_SLICE
+    0x00000000, // CB_COLOR3_VIEW
+    0x00000000, // CB_COLOR3_INFO
+    0x00000000, // CB_COLOR3_ATTRIB
+    0x00000000, // CB_COLOR3_DIM
+    0x00000000, // CB_COLOR3_CMASK
+    0x00000000, // CB_COLOR3_CMASK_SLICE
+    0x00000000, // CB_COLOR3_FMASK
+    0x00000000, // CB_COLOR3_FMASK_SLICE
+    0x00000000, // CB_COLOR3_CLEAR_WORD0
+    0x00000000, // CB_COLOR3_CLEAR_WORD1
+    0x00000000, // CB_COLOR3_CLEAR_WORD2
+    0x00000000, // CB_COLOR3_CLEAR_WORD3
+    0x00000000, // CB_COLOR4_BASE
+    0x00000000, // CB_COLOR4_PITCH
+    0x00000000, // CB_COLOR4_SLICE
+    0x00000000, // CB_COLOR4_VIEW
+    0x00000000, // CB_COLOR4_INFO
+    0x00000000, // CB_COLOR4_ATTRIB
+    0x00000000, // CB_COLOR4_DIM
+    0x00000000, // CB_COLOR4_CMASK
+    0x00000000, // CB_COLOR4_CMASK_SLICE
+    0x00000000, // CB_COLOR4_FMASK
+    0x00000000, // CB_COLOR4_FMASK_SLICE
+    0x00000000, // CB_COLOR4_CLEAR_WORD0
+    0x00000000, // CB_COLOR4_CLEAR_WORD1
+    0x00000000, // CB_COLOR4_CLEAR_WORD2
+    0x00000000, // CB_COLOR4_CLEAR_WORD3
+    0x00000000, // CB_COLOR5_BASE
+    0x00000000, // CB_COLOR5_PITCH
+    0x00000000, // CB_COLOR5_SLICE
+    0x00000000, // CB_COLOR5_VIEW
+    0x00000000, // CB_COLOR5_INFO
+    0x00000000, // CB_COLOR5_ATTRIB
+    0x00000000, // CB_COLOR5_DIM
+    0x00000000, // CB_COLOR5_CMASK
+    0x00000000, // CB_COLOR5_CMASK_SLICE
+    0x00000000, // CB_COLOR5_FMASK
+    0x00000000, // CB_COLOR5_FMASK_SLICE
+    0x00000000, // CB_COLOR5_CLEAR_WORD0
+    0x00000000, // CB_COLOR5_CLEAR_WORD1
+    0x00000000, // CB_COLOR5_CLEAR_WORD2
+    0x00000000, // CB_COLOR5_CLEAR_WORD3
+    0x00000000, // CB_COLOR6_BASE
+    0x00000000, // CB_COLOR6_PITCH
+    0x00000000, // CB_COLOR6_SLICE
+    0x00000000, // CB_COLOR6_VIEW
+    0x00000000, // CB_COLOR6_INFO
+    0x00000000, // CB_COLOR6_ATTRIB
+    0x00000000, // CB_COLOR6_DIM
+    0x00000000, // CB_COLOR6_CMASK
+    0x00000000, // CB_COLOR6_CMASK_SLICE
+    0x00000000, // CB_COLOR6_FMASK
+    0x00000000, // CB_COLOR6_FMASK_SLICE
+    0x00000000, // CB_COLOR6_CLEAR_WORD0
+    0x00000000, // CB_COLOR6_CLEAR_WORD1
+    0x00000000, // CB_COLOR6_CLEAR_WORD2
+    0x00000000, // CB_COLOR6_CLEAR_WORD3
+    0x00000000, // CB_COLOR7_BASE
+    0x00000000, // CB_COLOR7_PITCH
+    0x00000000, // CB_COLOR7_SLICE
+    0x00000000, // CB_COLOR7_VIEW
+    0x00000000, // CB_COLOR7_INFO
+    0x00000000, // CB_COLOR7_ATTRIB
+    0x00000000, // CB_COLOR7_DIM
+    0x00000000, // CB_COLOR7_CMASK
+    0x00000000, // CB_COLOR7_CMASK_SLICE
+    0x00000000, // CB_COLOR7_FMASK
+    0x00000000, // CB_COLOR7_FMASK_SLICE
+    0x00000000, // CB_COLOR7_CLEAR_WORD0
+    0x00000000, // CB_COLOR7_CLEAR_WORD1
+    0x00000000, // CB_COLOR7_CLEAR_WORD2
+    0x00000000, // CB_COLOR7_CLEAR_WORD3
+    0x00000000, // CB_COLOR8_BASE
+    0x00000000, // CB_COLOR8_PITCH
+    0x00000000, // CB_COLOR8_SLICE
+    0x00000000, // CB_COLOR8_VIEW
+    0x00000000, // CB_COLOR8_INFO
+    0x00000000, // CB_COLOR8_ATTRIB
+    0x00000000, // CB_COLOR8_DIM
+    0x00000000, // CB_COLOR9_BASE
+    0x00000000, // CB_COLOR9_PITCH
+    0x00000000, // CB_COLOR9_SLICE
+    0x00000000, // CB_COLOR9_VIEW
+    0x00000000, // CB_COLOR9_INFO
+    0x00000000, // CB_COLOR9_ATTRIB
+    0x00000000, // CB_COLOR9_DIM
+    0x00000000, // CB_COLOR10_BASE
+    0x00000000, // CB_COLOR10_PITCH
+    0x00000000, // CB_COLOR10_SLICE
+    0x00000000, // CB_COLOR10_VIEW
+    0x00000000, // CB_COLOR10_INFO
+    0x00000000, // CB_COLOR10_ATTRIB
+    0x00000000, // CB_COLOR10_DIM
+    0x00000000, // CB_COLOR11_BASE
+    0x00000000, // CB_COLOR11_PITCH
+    0x00000000, // CB_COLOR11_SLICE
+    0x00000000, // CB_COLOR11_VIEW
+    0x00000000, // CB_COLOR11_INFO
+    0x00000000, // CB_COLOR11_ATTRIB
+    0x00000000, // CB_COLOR11_DIM
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_15
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_15
+};
+static const struct cs_extent_def SECT_CONTEXT_defs[] =
+{
+    {SECT_CONTEXT_def_1, 0x0000a000, 488 },
+    {SECT_CONTEXT_def_2, 0x0000a1f5, 6 },
+    {SECT_CONTEXT_def_3, 0x0000a200, 55 },
+    {SECT_CONTEXT_def_4, 0x0000a23a, 99 },
+    {SECT_CONTEXT_def_5, 0x0000a29e, 5 },
+    {SECT_CONTEXT_def_6, 0x0000a2a5, 56 },
+    {SECT_CONTEXT_def_7, 0x0000a2de, 290 },
+    { 0, 0, 0 }
+};
+static const u32 SECT_CLEAR_def_1[] =
+{
+    0xffffffff, // SQ_TEX_SAMPLER_CLEAR
+    0xffffffff, // SQ_TEX_RESOURCE_CLEAR
+    0xffffffff, // SQ_LOOP_BOOL_CLEAR
+};
+static const struct cs_extent_def SECT_CLEAR_defs[] =
+{
+    {SECT_CLEAR_def_1, 0x0000ffc0, 3 },
+    { 0, 0, 0 }
+};
+static const u32 SECT_CTRLCONST_def_1[] =
+{
+    0x00000000, // SQ_VTX_BASE_VTX_LOC
+    0x00000000, // SQ_VTX_START_INST_LOC
+};
+static const struct cs_extent_def SECT_CTRLCONST_defs[] =
+{
+    {SECT_CTRLCONST_def_1, 0x0000f3fc, 2 },
+    { 0, 0, 0 }
+};
+struct cs_section_def cayman_cs_data[] = {
+    { SECT_CONTEXT_defs, SECT_CONTEXT },
+    { SECT_CLEAR_defs, SECT_CLEAR },
+    { SECT_CTRLCONST_defs, SECT_CTRLCONST },
+    { 0, SECT_NONE }
+};
diff --git a/drivers/gpu/drm/radeon/clearstate_defs.h b/drivers/gpu/drm/radeon/clearstate_defs.h
new file mode 100644
index 000000000000..3eda707d7388
--- /dev/null
+++ b/drivers/gpu/drm/radeon/clearstate_defs.h
@@ -0,0 +1,44 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef CLEARSTATE_DEFS_H
+#define CLEARSTATE_DEFS_H
+
+enum section_id {
+    SECT_NONE,
+    SECT_CONTEXT,
+    SECT_CLEAR,
+    SECT_CTRLCONST
+};
+
+struct cs_extent_def {
+    const unsigned int *extent;
+    const unsigned int reg_index;
+    const unsigned int reg_count;
+};
+
+struct cs_section_def {
+    const struct cs_extent_def *section;
+    const enum section_id id;
+};
+
+#endif
diff --git a/drivers/gpu/drm/radeon/clearstate_evergreen.h b/drivers/gpu/drm/radeon/clearstate_evergreen.h
new file mode 100644
index 000000000000..4791d856b7fd
--- /dev/null
+++ b/drivers/gpu/drm/radeon/clearstate_evergreen.h
@@ -0,0 +1,1080 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+static const u32 SECT_CONTEXT_def_1[] =
+{
+    0x00000000, // DB_RENDER_CONTROL
+    0x00000000, // DB_COUNT_CONTROL
+    0x00000000, // DB_DEPTH_VIEW
+    0x00000000, // DB_RENDER_OVERRIDE
+    0x00000000, // DB_RENDER_OVERRIDE2
+    0x00000000, // DB_HTILE_DATA_BASE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_STENCIL_CLEAR
+    0x00000000, // DB_DEPTH_CLEAR
+    0x00000000, // PA_SC_SCREEN_SCISSOR_TL
+    0x40004000, // PA_SC_SCREEN_SCISSOR_BR
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_Z_INFO
+    0x00000000, // DB_STENCIL_INFO
+    0x00000000, // DB_Z_READ_BASE
+    0x00000000, // DB_STENCIL_READ_BASE
+    0x00000000, // DB_Z_WRITE_BASE
+    0x00000000, // DB_STENCIL_WRITE_BASE
+    0x00000000, // DB_DEPTH_SIZE
+    0x00000000, // DB_DEPTH_SLICE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_PS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_VS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_GS_15
+    0x00000000, // PA_SC_WINDOW_OFFSET
+    0x80000000, // PA_SC_WINDOW_SCISSOR_TL
+    0x40004000, // PA_SC_WINDOW_SCISSOR_BR
+    0x0000ffff, // PA_SC_CLIPRECT_RULE
+    0x00000000, // PA_SC_CLIPRECT_0_TL
+    0x40004000, // PA_SC_CLIPRECT_0_BR
+    0x00000000, // PA_SC_CLIPRECT_1_TL
+    0x40004000, // PA_SC_CLIPRECT_1_BR
+    0x00000000, // PA_SC_CLIPRECT_2_TL
+    0x40004000, // PA_SC_CLIPRECT_2_BR
+    0x00000000, // PA_SC_CLIPRECT_3_TL
+    0x40004000, // PA_SC_CLIPRECT_3_BR
+    0xaa99aaaa, // PA_SC_EDGERULE
+    0x00000000, // PA_SU_HARDWARE_SCREEN_OFFSET
+    0xffffffff, // CB_TARGET_MASK
+    0xffffffff, // CB_SHADER_MASK
+    0x80000000, // PA_SC_GENERIC_SCISSOR_TL
+    0x40004000, // PA_SC_GENERIC_SCISSOR_BR
+    0x00000000, // COHER_DEST_BASE_0
+    0x00000000, // COHER_DEST_BASE_1
+    0x80000000, // PA_SC_VPORT_SCISSOR_0_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_0_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_1_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_1_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_2_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_2_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_3_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_3_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_4_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_4_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_5_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_5_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_6_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_6_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_7_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_7_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_8_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_8_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_9_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_9_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_10_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_10_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_11_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_11_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_12_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_12_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_13_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_13_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_14_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_14_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_15_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_15_BR
+    0x00000000, // PA_SC_VPORT_ZMIN_0
+    0x3f800000, // PA_SC_VPORT_ZMAX_0
+    0x00000000, // PA_SC_VPORT_ZMIN_1
+    0x3f800000, // PA_SC_VPORT_ZMAX_1
+    0x00000000, // PA_SC_VPORT_ZMIN_2
+    0x3f800000, // PA_SC_VPORT_ZMAX_2
+    0x00000000, // PA_SC_VPORT_ZMIN_3
+    0x3f800000, // PA_SC_VPORT_ZMAX_3
+    0x00000000, // PA_SC_VPORT_ZMIN_4
+    0x3f800000, // PA_SC_VPORT_ZMAX_4
+    0x00000000, // PA_SC_VPORT_ZMIN_5
+    0x3f800000, // PA_SC_VPORT_ZMAX_5
+    0x00000000, // PA_SC_VPORT_ZMIN_6
+    0x3f800000, // PA_SC_VPORT_ZMAX_6
+    0x00000000, // PA_SC_VPORT_ZMIN_7
+    0x3f800000, // PA_SC_VPORT_ZMAX_7
+    0x00000000, // PA_SC_VPORT_ZMIN_8
+    0x3f800000, // PA_SC_VPORT_ZMAX_8
+    0x00000000, // PA_SC_VPORT_ZMIN_9
+    0x3f800000, // PA_SC_VPORT_ZMAX_9
+    0x00000000, // PA_SC_VPORT_ZMIN_10
+    0x3f800000, // PA_SC_VPORT_ZMAX_10
+    0x00000000, // PA_SC_VPORT_ZMIN_11
+    0x3f800000, // PA_SC_VPORT_ZMAX_11
+    0x00000000, // PA_SC_VPORT_ZMIN_12
+    0x3f800000, // PA_SC_VPORT_ZMAX_12
+    0x00000000, // PA_SC_VPORT_ZMIN_13
+    0x3f800000, // PA_SC_VPORT_ZMAX_13
+    0x00000000, // PA_SC_VPORT_ZMIN_14
+    0x3f800000, // PA_SC_VPORT_ZMAX_14
+    0x00000000, // PA_SC_VPORT_ZMIN_15
+    0x3f800000, // PA_SC_VPORT_ZMAX_15
+    0x00000000, // SX_MISC
+    0x00000000, // SX_SURFACE_SYNC
+    0x00000000, // CP_PERFMON_CNTX_CNTL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_VTX_SEMANTIC_0
+    0x00000000, // SQ_VTX_SEMANTIC_1
+    0x00000000, // SQ_VTX_SEMANTIC_2
+    0x00000000, // SQ_VTX_SEMANTIC_3
+    0x00000000, // SQ_VTX_SEMANTIC_4
+    0x00000000, // SQ_VTX_SEMANTIC_5
+    0x00000000, // SQ_VTX_SEMANTIC_6
+    0x00000000, // SQ_VTX_SEMANTIC_7
+    0x00000000, // SQ_VTX_SEMANTIC_8
+    0x00000000, // SQ_VTX_SEMANTIC_9
+    0x00000000, // SQ_VTX_SEMANTIC_10
+    0x00000000, // SQ_VTX_SEMANTIC_11
+    0x00000000, // SQ_VTX_SEMANTIC_12
+    0x00000000, // SQ_VTX_SEMANTIC_13
+    0x00000000, // SQ_VTX_SEMANTIC_14
+    0x00000000, // SQ_VTX_SEMANTIC_15
+    0x00000000, // SQ_VTX_SEMANTIC_16
+    0x00000000, // SQ_VTX_SEMANTIC_17
+    0x00000000, // SQ_VTX_SEMANTIC_18
+    0x00000000, // SQ_VTX_SEMANTIC_19
+    0x00000000, // SQ_VTX_SEMANTIC_20
+    0x00000000, // SQ_VTX_SEMANTIC_21
+    0x00000000, // SQ_VTX_SEMANTIC_22
+    0x00000000, // SQ_VTX_SEMANTIC_23
+    0x00000000, // SQ_VTX_SEMANTIC_24
+    0x00000000, // SQ_VTX_SEMANTIC_25
+    0x00000000, // SQ_VTX_SEMANTIC_26
+    0x00000000, // SQ_VTX_SEMANTIC_27
+    0x00000000, // SQ_VTX_SEMANTIC_28
+    0x00000000, // SQ_VTX_SEMANTIC_29
+    0x00000000, // SQ_VTX_SEMANTIC_30
+    0x00000000, // SQ_VTX_SEMANTIC_31
+    0xffffffff, // VGT_MAX_VTX_INDX
+    0x00000000, // VGT_MIN_VTX_INDX
+    0x00000000, // VGT_INDX_OFFSET
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_INDX
+    0x00000000, // SX_ALPHA_TEST_CONTROL
+    0x00000000, // CB_BLEND_RED
+    0x00000000, // CB_BLEND_GREEN
+    0x00000000, // CB_BLEND_BLUE
+    0x00000000, // CB_BLEND_ALPHA
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_STENCILREFMASK
+    0x00000000, // DB_STENCILREFMASK_BF
+    0x00000000, // SX_ALPHA_REF
+    0x00000000, // PA_CL_VPORT_XSCALE
+    0x00000000, // PA_CL_VPORT_XOFFSET
+    0x00000000, // PA_CL_VPORT_YSCALE
+    0x00000000, // PA_CL_VPORT_YOFFSET
+    0x00000000, // PA_CL_VPORT_ZSCALE
+    0x00000000, // PA_CL_VPORT_ZOFFSET
+    0x00000000, // PA_CL_VPORT_XSCALE_1
+    0x00000000, // PA_CL_VPORT_XOFFSET_1
+    0x00000000, // PA_CL_VPORT_YSCALE_1
+    0x00000000, // PA_CL_VPORT_YOFFSET_1
+    0x00000000, // PA_CL_VPORT_ZSCALE_1
+    0x00000000, // PA_CL_VPORT_ZOFFSET_1
+    0x00000000, // PA_CL_VPORT_XSCALE_2
+    0x00000000, // PA_CL_VPORT_XOFFSET_2
+    0x00000000, // PA_CL_VPORT_YSCALE_2
+    0x00000000, // PA_CL_VPORT_YOFFSET_2
+    0x00000000, // PA_CL_VPORT_ZSCALE_2
+    0x00000000, // PA_CL_VPORT_ZOFFSET_2
+    0x00000000, // PA_CL_VPORT_XSCALE_3
+    0x00000000, // PA_CL_VPORT_XOFFSET_3
+    0x00000000, // PA_CL_VPORT_YSCALE_3
+    0x00000000, // PA_CL_VPORT_YOFFSET_3
+    0x00000000, // PA_CL_VPORT_ZSCALE_3
+    0x00000000, // PA_CL_VPORT_ZOFFSET_3
+    0x00000000, // PA_CL_VPORT_XSCALE_4
+    0x00000000, // PA_CL_VPORT_XOFFSET_4
+    0x00000000, // PA_CL_VPORT_YSCALE_4
+    0x00000000, // PA_CL_VPORT_YOFFSET_4
+    0x00000000, // PA_CL_VPORT_ZSCALE_4
+    0x00000000, // PA_CL_VPORT_ZOFFSET_4
+    0x00000000, // PA_CL_VPORT_XSCALE_5
+    0x00000000, // PA_CL_VPORT_XOFFSET_5
+    0x00000000, // PA_CL_VPORT_YSCALE_5
+    0x00000000, // PA_CL_VPORT_YOFFSET_5
+    0x00000000, // PA_CL_VPORT_ZSCALE_5
+    0x00000000, // PA_CL_VPORT_ZOFFSET_5
+    0x00000000, // PA_CL_VPORT_XSCALE_6
+    0x00000000, // PA_CL_VPORT_XOFFSET_6
+    0x00000000, // PA_CL_VPORT_YSCALE_6
+    0x00000000, // PA_CL_VPORT_YOFFSET_6
+    0x00000000, // PA_CL_VPORT_ZSCALE_6
+    0x00000000, // PA_CL_VPORT_ZOFFSET_6
+    0x00000000, // PA_CL_VPORT_XSCALE_7
+    0x00000000, // PA_CL_VPORT_XOFFSET_7
+    0x00000000, // PA_CL_VPORT_YSCALE_7
+    0x00000000, // PA_CL_VPORT_YOFFSET_7
+    0x00000000, // PA_CL_VPORT_ZSCALE_7
+    0x00000000, // PA_CL_VPORT_ZOFFSET_7
+    0x00000000, // PA_CL_VPORT_XSCALE_8
+    0x00000000, // PA_CL_VPORT_XOFFSET_8
+    0x00000000, // PA_CL_VPORT_YSCALE_8
+    0x00000000, // PA_CL_VPORT_YOFFSET_8
+    0x00000000, // PA_CL_VPORT_ZSCALE_8
+    0x00000000, // PA_CL_VPORT_ZOFFSET_8
+    0x00000000, // PA_CL_VPORT_XSCALE_9
+    0x00000000, // PA_CL_VPORT_XOFFSET_9
+    0x00000000, // PA_CL_VPORT_YSCALE_9
+    0x00000000, // PA_CL_VPORT_YOFFSET_9
+    0x00000000, // PA_CL_VPORT_ZSCALE_9
+    0x00000000, // PA_CL_VPORT_ZOFFSET_9
+    0x00000000, // PA_CL_VPORT_XSCALE_10
+    0x00000000, // PA_CL_VPORT_XOFFSET_10
+    0x00000000, // PA_CL_VPORT_YSCALE_10
+    0x00000000, // PA_CL_VPORT_YOFFSET_10
+    0x00000000, // PA_CL_VPORT_ZSCALE_10
+    0x00000000, // PA_CL_VPORT_ZOFFSET_10
+    0x00000000, // PA_CL_VPORT_XSCALE_11
+    0x00000000, // PA_CL_VPORT_XOFFSET_11
+    0x00000000, // PA_CL_VPORT_YSCALE_11
+    0x00000000, // PA_CL_VPORT_YOFFSET_11
+    0x00000000, // PA_CL_VPORT_ZSCALE_11
+    0x00000000, // PA_CL_VPORT_ZOFFSET_11
+    0x00000000, // PA_CL_VPORT_XSCALE_12
+    0x00000000, // PA_CL_VPORT_XOFFSET_12
+    0x00000000, // PA_CL_VPORT_YSCALE_12
+    0x00000000, // PA_CL_VPORT_YOFFSET_12
+    0x00000000, // PA_CL_VPORT_ZSCALE_12
+    0x00000000, // PA_CL_VPORT_ZOFFSET_12
+    0x00000000, // PA_CL_VPORT_XSCALE_13
+    0x00000000, // PA_CL_VPORT_XOFFSET_13
+    0x00000000, // PA_CL_VPORT_YSCALE_13
+    0x00000000, // PA_CL_VPORT_YOFFSET_13
+    0x00000000, // PA_CL_VPORT_ZSCALE_13
+    0x00000000, // PA_CL_VPORT_ZOFFSET_13
+    0x00000000, // PA_CL_VPORT_XSCALE_14
+    0x00000000, // PA_CL_VPORT_XOFFSET_14
+    0x00000000, // PA_CL_VPORT_YSCALE_14
+    0x00000000, // PA_CL_VPORT_YOFFSET_14
+    0x00000000, // PA_CL_VPORT_ZSCALE_14
+    0x00000000, // PA_CL_VPORT_ZOFFSET_14
+    0x00000000, // PA_CL_VPORT_XSCALE_15
+    0x00000000, // PA_CL_VPORT_XOFFSET_15
+    0x00000000, // PA_CL_VPORT_YSCALE_15
+    0x00000000, // PA_CL_VPORT_YOFFSET_15
+    0x00000000, // PA_CL_VPORT_ZSCALE_15
+    0x00000000, // PA_CL_VPORT_ZOFFSET_15
+    0x00000000, // PA_CL_UCP_0_X
+    0x00000000, // PA_CL_UCP_0_Y
+    0x00000000, // PA_CL_UCP_0_Z
+    0x00000000, // PA_CL_UCP_0_W
+    0x00000000, // PA_CL_UCP_1_X
+    0x00000000, // PA_CL_UCP_1_Y
+    0x00000000, // PA_CL_UCP_1_Z
+    0x00000000, // PA_CL_UCP_1_W
+    0x00000000, // PA_CL_UCP_2_X
+    0x00000000, // PA_CL_UCP_2_Y
+    0x00000000, // PA_CL_UCP_2_Z
+    0x00000000, // PA_CL_UCP_2_W
+    0x00000000, // PA_CL_UCP_3_X
+    0x00000000, // PA_CL_UCP_3_Y
+    0x00000000, // PA_CL_UCP_3_Z
+    0x00000000, // PA_CL_UCP_3_W
+    0x00000000, // PA_CL_UCP_4_X
+    0x00000000, // PA_CL_UCP_4_Y
+    0x00000000, // PA_CL_UCP_4_Z
+    0x00000000, // PA_CL_UCP_4_W
+    0x00000000, // PA_CL_UCP_5_X
+    0x00000000, // PA_CL_UCP_5_Y
+    0x00000000, // PA_CL_UCP_5_Z
+    0x00000000, // PA_CL_UCP_5_W
+    0x00000000, // SPI_VS_OUT_ID_0
+    0x00000000, // SPI_VS_OUT_ID_1
+    0x00000000, // SPI_VS_OUT_ID_2
+    0x00000000, // SPI_VS_OUT_ID_3
+    0x00000000, // SPI_VS_OUT_ID_4
+    0x00000000, // SPI_VS_OUT_ID_5
+    0x00000000, // SPI_VS_OUT_ID_6
+    0x00000000, // SPI_VS_OUT_ID_7
+    0x00000000, // SPI_VS_OUT_ID_8
+    0x00000000, // SPI_VS_OUT_ID_9
+    0x00000000, // SPI_PS_INPUT_CNTL_0
+    0x00000000, // SPI_PS_INPUT_CNTL_1
+    0x00000000, // SPI_PS_INPUT_CNTL_2
+    0x00000000, // SPI_PS_INPUT_CNTL_3
+    0x00000000, // SPI_PS_INPUT_CNTL_4
+    0x00000000, // SPI_PS_INPUT_CNTL_5
+    0x00000000, // SPI_PS_INPUT_CNTL_6
+    0x00000000, // SPI_PS_INPUT_CNTL_7
+    0x00000000, // SPI_PS_INPUT_CNTL_8
+    0x00000000, // SPI_PS_INPUT_CNTL_9
+    0x00000000, // SPI_PS_INPUT_CNTL_10
+    0x00000000, // SPI_PS_INPUT_CNTL_11
+    0x00000000, // SPI_PS_INPUT_CNTL_12
+    0x00000000, // SPI_PS_INPUT_CNTL_13
+    0x00000000, // SPI_PS_INPUT_CNTL_14
+    0x00000000, // SPI_PS_INPUT_CNTL_15
+    0x00000000, // SPI_PS_INPUT_CNTL_16
+    0x00000000, // SPI_PS_INPUT_CNTL_17
+    0x00000000, // SPI_PS_INPUT_CNTL_18
+    0x00000000, // SPI_PS_INPUT_CNTL_19
+    0x00000000, // SPI_PS_INPUT_CNTL_20
+    0x00000000, // SPI_PS_INPUT_CNTL_21
+    0x00000000, // SPI_PS_INPUT_CNTL_22
+    0x00000000, // SPI_PS_INPUT_CNTL_23
+    0x00000000, // SPI_PS_INPUT_CNTL_24
+    0x00000000, // SPI_PS_INPUT_CNTL_25
+    0x00000000, // SPI_PS_INPUT_CNTL_26
+    0x00000000, // SPI_PS_INPUT_CNTL_27
+    0x00000000, // SPI_PS_INPUT_CNTL_28
+    0x00000000, // SPI_PS_INPUT_CNTL_29
+    0x00000000, // SPI_PS_INPUT_CNTL_30
+    0x00000000, // SPI_PS_INPUT_CNTL_31
+    0x00000000, // SPI_VS_OUT_CONFIG
+    0x00000001, // SPI_THREAD_GROUPING
+    0x00000000, // SPI_PS_IN_CONTROL_0
+    0x00000000, // SPI_PS_IN_CONTROL_1
+    0x00000000, // SPI_INTERP_CONTROL_0
+    0x00000000, // SPI_INPUT_Z
+    0x00000000, // SPI_FOG_CNTL
+    0x00000000, // SPI_BARYC_CNTL
+    0x00000000, // SPI_PS_IN_CONTROL_2
+    0x00000000, // SPI_COMPUTE_INPUT_CNTL
+    0x00000000, // SPI_COMPUTE_NUM_THREAD_X
+    0x00000000, // SPI_COMPUTE_NUM_THREAD_Y
+    0x00000000, // SPI_COMPUTE_NUM_THREAD_Z
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // GDS_ADDR_BASE
+    0x00003fff, // GDS_ADDR_SIZE
+    0x00000001, // GDS_ORDERED_WAVE_PER_SE
+    0x00000000, // GDS_APPEND_CONSUME_UAV0
+    0x00000000, // GDS_APPEND_CONSUME_UAV1
+    0x00000000, // GDS_APPEND_CONSUME_UAV2
+    0x00000000, // GDS_APPEND_CONSUME_UAV3
+    0x00000000, // GDS_APPEND_CONSUME_UAV4
+    0x00000000, // GDS_APPEND_CONSUME_UAV5
+    0x00000000, // GDS_APPEND_CONSUME_UAV6
+    0x00000000, // GDS_APPEND_CONSUME_UAV7
+    0x00000000, // GDS_APPEND_CONSUME_UAV8
+    0x00000000, // GDS_APPEND_CONSUME_UAV9
+    0x00000000, // GDS_APPEND_CONSUME_UAV10
+    0x00000000, // GDS_APPEND_CONSUME_UAV11
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_BLEND0_CONTROL
+    0x00000000, // CB_BLEND1_CONTROL
+    0x00000000, // CB_BLEND2_CONTROL
+    0x00000000, // CB_BLEND3_CONTROL
+    0x00000000, // CB_BLEND4_CONTROL
+    0x00000000, // CB_BLEND5_CONTROL
+    0x00000000, // CB_BLEND6_CONTROL
+    0x00000000, // CB_BLEND7_CONTROL
+};
+static const u32 SECT_CONTEXT_def_2[] =
+{
+    0x00000000, // PA_CL_POINT_X_RAD
+    0x00000000, // PA_CL_POINT_Y_RAD
+    0x00000000, // PA_CL_POINT_SIZE
+    0x00000000, // PA_CL_POINT_CULL_RAD
+    0x00000000, // VGT_DMA_BASE_HI
+    0x00000000, // VGT_DMA_BASE
+};
+static const u32 SECT_CONTEXT_def_3[] =
+{
+    0x00000000, // DB_DEPTH_CONTROL
+    0, // HOLE
+    0x00000000, // CB_COLOR_CONTROL
+    0x00000200, // DB_SHADER_CONTROL
+    0x00000000, // PA_CL_CLIP_CNTL
+    0x00000000, // PA_SU_SC_MODE_CNTL
+    0x00000000, // PA_CL_VTE_CNTL
+    0x00000000, // PA_CL_VS_OUT_CNTL
+    0x00000000, // PA_CL_NANINF_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_SCALE
+    0x00000000, // PA_SU_PRIM_FILTER_CNTL
+    0x00000000, // SQ_LSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_HSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_DYN_GPR_RESOURCE_LIMIT_1
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_PS
+    0x00000000, // SQ_PGM_RESOURCES_PS
+    0x00000000, // SQ_PGM_RESOURCES_2_PS
+    0x00000000, // SQ_PGM_EXPORTS_PS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_VS
+    0x00000000, // SQ_PGM_RESOURCES_VS
+    0x00000000, // SQ_PGM_RESOURCES_2_VS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_GS
+    0x00000000, // SQ_PGM_RESOURCES_GS
+    0x00000000, // SQ_PGM_RESOURCES_2_GS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_ES
+    0x00000000, // SQ_PGM_RESOURCES_ES
+    0x00000000, // SQ_PGM_RESOURCES_2_ES
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_FS
+    0x00000000, // SQ_PGM_RESOURCES_FS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_HS
+    0x00000000, // SQ_PGM_RESOURCES_HS
+    0x00000000, // SQ_PGM_RESOURCES_2_HS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_PGM_START_LS
+    0x00000000, // SQ_PGM_RESOURCES_LS
+    0x00000000, // SQ_PGM_RESOURCES_2_LS
+};
+static const u32 SECT_CONTEXT_def_4[] =
+{
+    0x00000000, // SQ_LDS_ALLOC
+    0x00000000, // SQ_LDS_ALLOC_PS
+    0x00000000, // SQ_VTX_SEMANTIC_CLEAR
+    0, // HOLE
+    0x00000000, // SQ_THREAD_TRACE_CTRL
+    0, // HOLE
+    0x00000000, // SQ_ESGS_RING_ITEMSIZE
+    0x00000000, // SQ_GSVS_RING_ITEMSIZE
+    0x00000000, // SQ_ESTMP_RING_ITEMSIZE
+    0x00000000, // SQ_GSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_VSTMP_RING_ITEMSIZE
+    0x00000000, // SQ_PSTMP_RING_ITEMSIZE
+    0, // HOLE
+    0x00000000, // SQ_GS_VERT_ITEMSIZE
+    0x00000000, // SQ_GS_VERT_ITEMSIZE_1
+    0x00000000, // SQ_GS_VERT_ITEMSIZE_2
+    0x00000000, // SQ_GS_VERT_ITEMSIZE_3
+    0x00000000, // SQ_GSVS_RING_OFFSET_1
+    0x00000000, // SQ_GSVS_RING_OFFSET_2
+    0x00000000, // SQ_GSVS_RING_OFFSET_3
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_PS_15
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_VS_15
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_GS_15
+    0x00000000, // PA_SU_POINT_SIZE
+    0x00000000, // PA_SU_POINT_MINMAX
+    0x00000000, // PA_SU_LINE_CNTL
+    0x00000000, // PA_SC_LINE_STIPPLE
+    0x00000000, // VGT_OUTPUT_PATH_CNTL
+    0x00000000, // VGT_HOS_CNTL
+    0x00000000, // VGT_HOS_MAX_TESS_LEVEL
+    0x00000000, // VGT_HOS_MIN_TESS_LEVEL
+    0x00000000, // VGT_HOS_REUSE_DEPTH
+    0x00000000, // VGT_GROUP_PRIM_TYPE
+    0x00000000, // VGT_GROUP_FIRST_DECR
+    0x00000000, // VGT_GROUP_DECR
+    0x00000000, // VGT_GROUP_VECT_0_CNTL
+    0x00000000, // VGT_GROUP_VECT_1_CNTL
+    0x00000000, // VGT_GROUP_VECT_0_FMT_CNTL
+    0x00000000, // VGT_GROUP_VECT_1_FMT_CNTL
+    0x00000000, // VGT_GS_MODE
+    0, // HOLE
+    0x00000000, // PA_SC_MODE_CNTL_0
+    0x00000000, // PA_SC_MODE_CNTL_1
+    0x00000000, // VGT_ENHANCE
+    0x00000000, // VGT_GS_PER_ES
+    0x00000000, // VGT_ES_PER_GS
+    0x00000000, // VGT_GS_PER_VS
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_GS_OUT_PRIM_TYPE
+};
+static const u32 SECT_CONTEXT_def_5[] =
+{
+    0x00000000, // VGT_DMA_MAX_SIZE
+    0x00000000, // VGT_DMA_INDEX_TYPE
+    0, // HOLE
+    0x00000000, // VGT_PRIMITIVEID_EN
+    0x00000000, // VGT_DMA_NUM_INSTANCES
+};
+static const u32 SECT_CONTEXT_def_6[] =
+{
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_EN
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_INSTANCE_STEP_RATE_0
+    0x00000000, // VGT_INSTANCE_STEP_RATE_1
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_REUSE_OFF
+    0x00000000, // VGT_VTX_CNT_EN
+    0x00000000, // DB_HTILE_SURFACE
+    0x00000000, // DB_SRESULTS_COMPARE_STATE0
+    0x00000000, // DB_SRESULTS_COMPARE_STATE1
+    0x00000000, // DB_PRELOAD_CONTROL
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_0
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_0
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_0
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_0
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_1
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_1
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_1
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_1
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_2
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_2
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_2
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_2
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_3
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_3
+    0x00000000, // VGT_STRMOUT_BUFFER_BASE_3
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_3
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_0
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_1
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_2
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_3
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_OFFSET
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
+    0, // HOLE
+    0x00000000, // VGT_GS_MAX_VERT_OUT
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_0
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_1
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_2
+    0x00000000, // VGT_STRMOUT_BASE_OFFSET_HI_3
+    0x00000000, // VGT_SHADER_STAGES_EN
+    0x00000000, // VGT_LS_HS_CONFIG
+    0x00000000, // VGT_LS_SIZE
+    0x00000000, // VGT_HS_SIZE
+    0x00000000, // VGT_LS_HS_ALLOC
+    0x00000000, // VGT_HS_PATCH_CONST
+    0x00000000, // VGT_TF_PARAM
+    0x00000000, // DB_ALPHA_TO_MASK
+};
+static const u32 SECT_CONTEXT_def_7[] =
+{
+    0x00000000, // PA_SU_POLY_OFFSET_DB_FMT_CNTL
+    0x00000000, // PA_SU_POLY_OFFSET_CLAMP
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_OFFSET
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_OFFSET
+    0x00000000, // VGT_GS_INSTANCE_CNT
+    0x00000000, // VGT_STRMOUT_CONFIG
+    0x00000000, // VGT_STRMOUT_BUFFER_CONFIG
+    0x00000000, // CB_IMMED0_BASE
+    0x00000000, // CB_IMMED1_BASE
+    0x00000000, // CB_IMMED2_BASE
+    0x00000000, // CB_IMMED3_BASE
+    0x00000000, // CB_IMMED4_BASE
+    0x00000000, // CB_IMMED5_BASE
+    0x00000000, // CB_IMMED6_BASE
+    0x00000000, // CB_IMMED7_BASE
+    0x00000000, // CB_IMMED8_BASE
+    0x00000000, // CB_IMMED9_BASE
+    0x00000000, // CB_IMMED10_BASE
+    0x00000000, // CB_IMMED11_BASE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00001000, // PA_SC_LINE_CNTL
+    0x00000000, // PA_SC_AA_CONFIG
+    0x00000005, // PA_SU_VTX_CNTL
+    0x3f800000, // PA_CL_GB_VERT_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_VERT_DISC_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_DISC_ADJ
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_4
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_5
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_6
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_7
+    0xffffffff, // PA_SC_AA_MASK
+    0x00000000, // CB_CLRCMP_CONTROL
+    0x00000000, // CB_CLRCMP_SRC
+    0x00000000, // CB_CLRCMP_DST
+    0x00000000, // CB_CLRCMP_MSK
+    0, // HOLE
+    0, // HOLE
+    0x0000000e, // VGT_VERTEX_REUSE_BLOCK_CNTL
+    0x00000010, // VGT_OUT_DEALLOC_CNTL
+    0x00000000, // CB_COLOR0_BASE
+    0x00000000, // CB_COLOR0_PITCH
+    0x00000000, // CB_COLOR0_SLICE
+    0x00000000, // CB_COLOR0_VIEW
+    0x00000000, // CB_COLOR0_INFO
+    0x00000000, // CB_COLOR0_ATTRIB
+    0x00000000, // CB_COLOR0_DIM
+    0x00000000, // CB_COLOR0_CMASK
+    0x00000000, // CB_COLOR0_CMASK_SLICE
+    0x00000000, // CB_COLOR0_FMASK
+    0x00000000, // CB_COLOR0_FMASK_SLICE
+    0x00000000, // CB_COLOR0_CLEAR_WORD0
+    0x00000000, // CB_COLOR0_CLEAR_WORD1
+    0x00000000, // CB_COLOR0_CLEAR_WORD2
+    0x00000000, // CB_COLOR0_CLEAR_WORD3
+    0x00000000, // CB_COLOR1_BASE
+    0x00000000, // CB_COLOR1_PITCH
+    0x00000000, // CB_COLOR1_SLICE
+    0x00000000, // CB_COLOR1_VIEW
+    0x00000000, // CB_COLOR1_INFO
+    0x00000000, // CB_COLOR1_ATTRIB
+    0x00000000, // CB_COLOR1_DIM
+    0x00000000, // CB_COLOR1_CMASK
+    0x00000000, // CB_COLOR1_CMASK_SLICE
+    0x00000000, // CB_COLOR1_FMASK
+    0x00000000, // CB_COLOR1_FMASK_SLICE
+    0x00000000, // CB_COLOR1_CLEAR_WORD0
+    0x00000000, // CB_COLOR1_CLEAR_WORD1
+    0x00000000, // CB_COLOR1_CLEAR_WORD2
+    0x00000000, // CB_COLOR1_CLEAR_WORD3
+    0x00000000, // CB_COLOR2_BASE
+    0x00000000, // CB_COLOR2_PITCH
+    0x00000000, // CB_COLOR2_SLICE
+    0x00000000, // CB_COLOR2_VIEW
+    0x00000000, // CB_COLOR2_INFO
+    0x00000000, // CB_COLOR2_ATTRIB
+    0x00000000, // CB_COLOR2_DIM
+    0x00000000, // CB_COLOR2_CMASK
+    0x00000000, // CB_COLOR2_CMASK_SLICE
+    0x00000000, // CB_COLOR2_FMASK
+    0x00000000, // CB_COLOR2_FMASK_SLICE
+    0x00000000, // CB_COLOR2_CLEAR_WORD0
+    0x00000000, // CB_COLOR2_CLEAR_WORD1
+    0x00000000, // CB_COLOR2_CLEAR_WORD2
+    0x00000000, // CB_COLOR2_CLEAR_WORD3
+    0x00000000, // CB_COLOR3_BASE
+    0x00000000, // CB_COLOR3_PITCH
+    0x00000000, // CB_COLOR3_SLICE
+    0x00000000, // CB_COLOR3_VIEW
+    0x00000000, // CB_COLOR3_INFO
+    0x00000000, // CB_COLOR3_ATTRIB
+    0x00000000, // CB_COLOR3_DIM
+    0x00000000, // CB_COLOR3_CMASK
+    0x00000000, // CB_COLOR3_CMASK_SLICE
+    0x00000000, // CB_COLOR3_FMASK
+    0x00000000, // CB_COLOR3_FMASK_SLICE
+    0x00000000, // CB_COLOR3_CLEAR_WORD0
+    0x00000000, // CB_COLOR3_CLEAR_WORD1
+    0x00000000, // CB_COLOR3_CLEAR_WORD2
+    0x00000000, // CB_COLOR3_CLEAR_WORD3
+    0x00000000, // CB_COLOR4_BASE
+    0x00000000, // CB_COLOR4_PITCH
+    0x00000000, // CB_COLOR4_SLICE
+    0x00000000, // CB_COLOR4_VIEW
+    0x00000000, // CB_COLOR4_INFO
+    0x00000000, // CB_COLOR4_ATTRIB
+    0x00000000, // CB_COLOR4_DIM
+    0x00000000, // CB_COLOR4_CMASK
+    0x00000000, // CB_COLOR4_CMASK_SLICE
+    0x00000000, // CB_COLOR4_FMASK
+    0x00000000, // CB_COLOR4_FMASK_SLICE
+    0x00000000, // CB_COLOR4_CLEAR_WORD0
+    0x00000000, // CB_COLOR4_CLEAR_WORD1
+    0x00000000, // CB_COLOR4_CLEAR_WORD2
+    0x00000000, // CB_COLOR4_CLEAR_WORD3
+    0x00000000, // CB_COLOR5_BASE
+    0x00000000, // CB_COLOR5_PITCH
+    0x00000000, // CB_COLOR5_SLICE
+    0x00000000, // CB_COLOR5_VIEW
+    0x00000000, // CB_COLOR5_INFO
+    0x00000000, // CB_COLOR5_ATTRIB
+    0x00000000, // CB_COLOR5_DIM
+    0x00000000, // CB_COLOR5_CMASK
+    0x00000000, // CB_COLOR5_CMASK_SLICE
+    0x00000000, // CB_COLOR5_FMASK
+    0x00000000, // CB_COLOR5_FMASK_SLICE
+    0x00000000, // CB_COLOR5_CLEAR_WORD0
+    0x00000000, // CB_COLOR5_CLEAR_WORD1
+    0x00000000, // CB_COLOR5_CLEAR_WORD2
+    0x00000000, // CB_COLOR5_CLEAR_WORD3
+    0x00000000, // CB_COLOR6_BASE
+    0x00000000, // CB_COLOR6_PITCH
+    0x00000000, // CB_COLOR6_SLICE
+    0x00000000, // CB_COLOR6_VIEW
+    0x00000000, // CB_COLOR6_INFO
+    0x00000000, // CB_COLOR6_ATTRIB
+    0x00000000, // CB_COLOR6_DIM
+    0x00000000, // CB_COLOR6_CMASK
+    0x00000000, // CB_COLOR6_CMASK_SLICE
+    0x00000000, // CB_COLOR6_FMASK
+    0x00000000, // CB_COLOR6_FMASK_SLICE
+    0x00000000, // CB_COLOR6_CLEAR_WORD0
+    0x00000000, // CB_COLOR6_CLEAR_WORD1
+    0x00000000, // CB_COLOR6_CLEAR_WORD2
+    0x00000000, // CB_COLOR6_CLEAR_WORD3
+    0x00000000, // CB_COLOR7_BASE
+    0x00000000, // CB_COLOR7_PITCH
+    0x00000000, // CB_COLOR7_SLICE
+    0x00000000, // CB_COLOR7_VIEW
+    0x00000000, // CB_COLOR7_INFO
+    0x00000000, // CB_COLOR7_ATTRIB
+    0x00000000, // CB_COLOR7_DIM
+    0x00000000, // CB_COLOR7_CMASK
+    0x00000000, // CB_COLOR7_CMASK_SLICE
+    0x00000000, // CB_COLOR7_FMASK
+    0x00000000, // CB_COLOR7_FMASK_SLICE
+    0x00000000, // CB_COLOR7_CLEAR_WORD0
+    0x00000000, // CB_COLOR7_CLEAR_WORD1
+    0x00000000, // CB_COLOR7_CLEAR_WORD2
+    0x00000000, // CB_COLOR7_CLEAR_WORD3
+    0x00000000, // CB_COLOR8_BASE
+    0x00000000, // CB_COLOR8_PITCH
+    0x00000000, // CB_COLOR8_SLICE
+    0x00000000, // CB_COLOR8_VIEW
+    0x00000000, // CB_COLOR8_INFO
+    0x00000000, // CB_COLOR8_ATTRIB
+    0x00000000, // CB_COLOR8_DIM
+    0x00000000, // CB_COLOR9_BASE
+    0x00000000, // CB_COLOR9_PITCH
+    0x00000000, // CB_COLOR9_SLICE
+    0x00000000, // CB_COLOR9_VIEW
+    0x00000000, // CB_COLOR9_INFO
+    0x00000000, // CB_COLOR9_ATTRIB
+    0x00000000, // CB_COLOR9_DIM
+    0x00000000, // CB_COLOR10_BASE
+    0x00000000, // CB_COLOR10_PITCH
+    0x00000000, // CB_COLOR10_SLICE
+    0x00000000, // CB_COLOR10_VIEW
+    0x00000000, // CB_COLOR10_INFO
+    0x00000000, // CB_COLOR10_ATTRIB
+    0x00000000, // CB_COLOR10_DIM
+    0x00000000, // CB_COLOR11_BASE
+    0x00000000, // CB_COLOR11_PITCH
+    0x00000000, // CB_COLOR11_SLICE
+    0x00000000, // CB_COLOR11_VIEW
+    0x00000000, // CB_COLOR11_INFO
+    0x00000000, // CB_COLOR11_ATTRIB
+    0x00000000, // CB_COLOR11_DIM
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_HS_15
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_0
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_1
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_2
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_3
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_4
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_5
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_6
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_7
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_8
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_9
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_10
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_11
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_12
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_13
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_14
+    0x00000000, // SQ_ALU_CONST_CACHE_LS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_HS_15
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_0
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_1
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_2
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_3
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_4
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_5
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_6
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_7
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_8
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_9
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_10
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_11
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_12
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_13
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_14
+    0x00000000, // SQ_ALU_CONST_BUFFER_SIZE_LS_15
+};
+static const struct cs_extent_def SECT_CONTEXT_defs[] =
+{
+    {SECT_CONTEXT_def_1, 0x0000a000, 488 },
+    {SECT_CONTEXT_def_2, 0x0000a1f5, 6 },
+    {SECT_CONTEXT_def_3, 0x0000a200, 55 },
+    {SECT_CONTEXT_def_4, 0x0000a23a, 98 },
+    {SECT_CONTEXT_def_5, 0x0000a29e, 5 },
+    {SECT_CONTEXT_def_6, 0x0000a2a5, 56 },
+    {SECT_CONTEXT_def_7, 0x0000a2de, 290 },
+    { 0, 0, 0 }
+};
+static const u32 SECT_CLEAR_def_1[] =
+{
+    0xffffffff, // SQ_TEX_SAMPLER_CLEAR
+    0xffffffff, // SQ_TEX_RESOURCE_CLEAR
+    0xffffffff, // SQ_LOOP_BOOL_CLEAR
+};
+static const struct cs_extent_def SECT_CLEAR_defs[] =
+{
+    {SECT_CLEAR_def_1, 0x0000ffc0, 3 },
+    { 0, 0, 0 }
+};
+static const u32 SECT_CTRLCONST_def_1[] =
+{
+    0x00000000, // SQ_VTX_BASE_VTX_LOC
+    0x00000000, // SQ_VTX_START_INST_LOC
+};
+static const struct cs_extent_def SECT_CTRLCONST_defs[] =
+{
+    {SECT_CTRLCONST_def_1, 0x0000f3fc, 2 },
+    { 0, 0, 0 }
+};
+struct cs_section_def evergreen_cs_data[] = {
+    { SECT_CONTEXT_defs, SECT_CONTEXT },
+    { SECT_CLEAR_defs, SECT_CLEAR },
+    { SECT_CTRLCONST_defs, SECT_CTRLCONST },
+    { 0, SECT_NONE }
+};
diff --git a/drivers/gpu/drm/radeon/clearstate_si.h b/drivers/gpu/drm/radeon/clearstate_si.h
new file mode 100644
index 000000000000..b994cb2a35a0
--- /dev/null
+++ b/drivers/gpu/drm/radeon/clearstate_si.h
@@ -0,0 +1,941 @@
+/*
+ * Copyright 2013 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+static const u32 si_SECT_CONTEXT_def_1[] =
+{
+    0x00000000, // DB_RENDER_CONTROL
+    0x00000000, // DB_COUNT_CONTROL
+    0x00000000, // DB_DEPTH_VIEW
+    0x00000000, // DB_RENDER_OVERRIDE
+    0x00000000, // DB_RENDER_OVERRIDE2
+    0x00000000, // DB_HTILE_DATA_BASE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_DEPTH_BOUNDS_MIN
+    0x00000000, // DB_DEPTH_BOUNDS_MAX
+    0x00000000, // DB_STENCIL_CLEAR
+    0x00000000, // DB_DEPTH_CLEAR
+    0x00000000, // PA_SC_SCREEN_SCISSOR_TL
+    0x40004000, // PA_SC_SCREEN_SCISSOR_BR
+    0, // HOLE
+    0x00000000, // DB_DEPTH_INFO
+    0x00000000, // DB_Z_INFO
+    0x00000000, // DB_STENCIL_INFO
+    0x00000000, // DB_Z_READ_BASE
+    0x00000000, // DB_STENCIL_READ_BASE
+    0x00000000, // DB_Z_WRITE_BASE
+    0x00000000, // DB_STENCIL_WRITE_BASE
+    0x00000000, // DB_DEPTH_SIZE
+    0x00000000, // DB_DEPTH_SLICE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // TA_BC_BASE_ADDR
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // COHER_DEST_BASE_2
+    0x00000000, // COHER_DEST_BASE_3
+    0x00000000, // PA_SC_WINDOW_OFFSET
+    0x80000000, // PA_SC_WINDOW_SCISSOR_TL
+    0x40004000, // PA_SC_WINDOW_SCISSOR_BR
+    0x0000ffff, // PA_SC_CLIPRECT_RULE
+    0x00000000, // PA_SC_CLIPRECT_0_TL
+    0x40004000, // PA_SC_CLIPRECT_0_BR
+    0x00000000, // PA_SC_CLIPRECT_1_TL
+    0x40004000, // PA_SC_CLIPRECT_1_BR
+    0x00000000, // PA_SC_CLIPRECT_2_TL
+    0x40004000, // PA_SC_CLIPRECT_2_BR
+    0x00000000, // PA_SC_CLIPRECT_3_TL
+    0x40004000, // PA_SC_CLIPRECT_3_BR
+    0xaa99aaaa, // PA_SC_EDGERULE
+    0x00000000, // PA_SU_HARDWARE_SCREEN_OFFSET
+    0xffffffff, // CB_TARGET_MASK
+    0xffffffff, // CB_SHADER_MASK
+    0x80000000, // PA_SC_GENERIC_SCISSOR_TL
+    0x40004000, // PA_SC_GENERIC_SCISSOR_BR
+    0x00000000, // COHER_DEST_BASE_0
+    0x00000000, // COHER_DEST_BASE_1
+    0x80000000, // PA_SC_VPORT_SCISSOR_0_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_0_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_1_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_1_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_2_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_2_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_3_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_3_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_4_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_4_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_5_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_5_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_6_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_6_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_7_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_7_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_8_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_8_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_9_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_9_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_10_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_10_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_11_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_11_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_12_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_12_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_13_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_13_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_14_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_14_BR
+    0x80000000, // PA_SC_VPORT_SCISSOR_15_TL
+    0x40004000, // PA_SC_VPORT_SCISSOR_15_BR
+    0x00000000, // PA_SC_VPORT_ZMIN_0
+    0x3f800000, // PA_SC_VPORT_ZMAX_0
+    0x00000000, // PA_SC_VPORT_ZMIN_1
+    0x3f800000, // PA_SC_VPORT_ZMAX_1
+    0x00000000, // PA_SC_VPORT_ZMIN_2
+    0x3f800000, // PA_SC_VPORT_ZMAX_2
+    0x00000000, // PA_SC_VPORT_ZMIN_3
+    0x3f800000, // PA_SC_VPORT_ZMAX_3
+    0x00000000, // PA_SC_VPORT_ZMIN_4
+    0x3f800000, // PA_SC_VPORT_ZMAX_4
+    0x00000000, // PA_SC_VPORT_ZMIN_5
+    0x3f800000, // PA_SC_VPORT_ZMAX_5
+    0x00000000, // PA_SC_VPORT_ZMIN_6
+    0x3f800000, // PA_SC_VPORT_ZMAX_6
+    0x00000000, // PA_SC_VPORT_ZMIN_7
+    0x3f800000, // PA_SC_VPORT_ZMAX_7
+    0x00000000, // PA_SC_VPORT_ZMIN_8
+    0x3f800000, // PA_SC_VPORT_ZMAX_8
+    0x00000000, // PA_SC_VPORT_ZMIN_9
+    0x3f800000, // PA_SC_VPORT_ZMAX_9
+    0x00000000, // PA_SC_VPORT_ZMIN_10
+    0x3f800000, // PA_SC_VPORT_ZMAX_10
+    0x00000000, // PA_SC_VPORT_ZMIN_11
+    0x3f800000, // PA_SC_VPORT_ZMAX_11
+    0x00000000, // PA_SC_VPORT_ZMIN_12
+    0x3f800000, // PA_SC_VPORT_ZMAX_12
+    0x00000000, // PA_SC_VPORT_ZMIN_13
+    0x3f800000, // PA_SC_VPORT_ZMAX_13
+    0x00000000, // PA_SC_VPORT_ZMIN_14
+    0x3f800000, // PA_SC_VPORT_ZMAX_14
+    0x00000000, // PA_SC_VPORT_ZMIN_15
+    0x3f800000, // PA_SC_VPORT_ZMAX_15
+};
+static const u32 si_SECT_CONTEXT_def_2[] =
+{
+    0x00000000, // CP_PERFMON_CNTX_CNTL
+    0x00000000, // CP_RINGID
+    0x00000000, // CP_VMID
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0xffffffff, // VGT_MAX_VTX_INDX
+    0x00000000, // VGT_MIN_VTX_INDX
+    0x00000000, // VGT_INDX_OFFSET
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_INDX
+    0, // HOLE
+    0x00000000, // CB_BLEND_RED
+    0x00000000, // CB_BLEND_GREEN
+    0x00000000, // CB_BLEND_BLUE
+    0x00000000, // CB_BLEND_ALPHA
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // DB_STENCIL_CONTROL
+    0x00000000, // DB_STENCILREFMASK
+    0x00000000, // DB_STENCILREFMASK_BF
+    0, // HOLE
+    0x00000000, // PA_CL_VPORT_XSCALE
+    0x00000000, // PA_CL_VPORT_XOFFSET
+    0x00000000, // PA_CL_VPORT_YSCALE
+    0x00000000, // PA_CL_VPORT_YOFFSET
+    0x00000000, // PA_CL_VPORT_ZSCALE
+    0x00000000, // PA_CL_VPORT_ZOFFSET
+    0x00000000, // PA_CL_VPORT_XSCALE_1
+    0x00000000, // PA_CL_VPORT_XOFFSET_1
+    0x00000000, // PA_CL_VPORT_YSCALE_1
+    0x00000000, // PA_CL_VPORT_YOFFSET_1
+    0x00000000, // PA_CL_VPORT_ZSCALE_1
+    0x00000000, // PA_CL_VPORT_ZOFFSET_1
+    0x00000000, // PA_CL_VPORT_XSCALE_2
+    0x00000000, // PA_CL_VPORT_XOFFSET_2
+    0x00000000, // PA_CL_VPORT_YSCALE_2
+    0x00000000, // PA_CL_VPORT_YOFFSET_2
+    0x00000000, // PA_CL_VPORT_ZSCALE_2
+    0x00000000, // PA_CL_VPORT_ZOFFSET_2
+    0x00000000, // PA_CL_VPORT_XSCALE_3
+    0x00000000, // PA_CL_VPORT_XOFFSET_3
+    0x00000000, // PA_CL_VPORT_YSCALE_3
+    0x00000000, // PA_CL_VPORT_YOFFSET_3
+    0x00000000, // PA_CL_VPORT_ZSCALE_3
+    0x00000000, // PA_CL_VPORT_ZOFFSET_3
+    0x00000000, // PA_CL_VPORT_XSCALE_4
+    0x00000000, // PA_CL_VPORT_XOFFSET_4
+    0x00000000, // PA_CL_VPORT_YSCALE_4
+    0x00000000, // PA_CL_VPORT_YOFFSET_4
+    0x00000000, // PA_CL_VPORT_ZSCALE_4
+    0x00000000, // PA_CL_VPORT_ZOFFSET_4
+    0x00000000, // PA_CL_VPORT_XSCALE_5
+    0x00000000, // PA_CL_VPORT_XOFFSET_5
+    0x00000000, // PA_CL_VPORT_YSCALE_5
+    0x00000000, // PA_CL_VPORT_YOFFSET_5
+    0x00000000, // PA_CL_VPORT_ZSCALE_5
+    0x00000000, // PA_CL_VPORT_ZOFFSET_5
+    0x00000000, // PA_CL_VPORT_XSCALE_6
+    0x00000000, // PA_CL_VPORT_XOFFSET_6
+    0x00000000, // PA_CL_VPORT_YSCALE_6
+    0x00000000, // PA_CL_VPORT_YOFFSET_6
+    0x00000000, // PA_CL_VPORT_ZSCALE_6
+    0x00000000, // PA_CL_VPORT_ZOFFSET_6
+    0x00000000, // PA_CL_VPORT_XSCALE_7
+    0x00000000, // PA_CL_VPORT_XOFFSET_7
+    0x00000000, // PA_CL_VPORT_YSCALE_7
+    0x00000000, // PA_CL_VPORT_YOFFSET_7
+    0x00000000, // PA_CL_VPORT_ZSCALE_7
+    0x00000000, // PA_CL_VPORT_ZOFFSET_7
+    0x00000000, // PA_CL_VPORT_XSCALE_8
+    0x00000000, // PA_CL_VPORT_XOFFSET_8
+    0x00000000, // PA_CL_VPORT_YSCALE_8
+    0x00000000, // PA_CL_VPORT_YOFFSET_8
+    0x00000000, // PA_CL_VPORT_ZSCALE_8
+    0x00000000, // PA_CL_VPORT_ZOFFSET_8
+    0x00000000, // PA_CL_VPORT_XSCALE_9
+    0x00000000, // PA_CL_VPORT_XOFFSET_9
+    0x00000000, // PA_CL_VPORT_YSCALE_9
+    0x00000000, // PA_CL_VPORT_YOFFSET_9
+    0x00000000, // PA_CL_VPORT_ZSCALE_9
+    0x00000000, // PA_CL_VPORT_ZOFFSET_9
+    0x00000000, // PA_CL_VPORT_XSCALE_10
+    0x00000000, // PA_CL_VPORT_XOFFSET_10
+    0x00000000, // PA_CL_VPORT_YSCALE_10
+    0x00000000, // PA_CL_VPORT_YOFFSET_10
+    0x00000000, // PA_CL_VPORT_ZSCALE_10
+    0x00000000, // PA_CL_VPORT_ZOFFSET_10
+    0x00000000, // PA_CL_VPORT_XSCALE_11
+    0x00000000, // PA_CL_VPORT_XOFFSET_11
+    0x00000000, // PA_CL_VPORT_YSCALE_11
+    0x00000000, // PA_CL_VPORT_YOFFSET_11
+    0x00000000, // PA_CL_VPORT_ZSCALE_11
+    0x00000000, // PA_CL_VPORT_ZOFFSET_11
+    0x00000000, // PA_CL_VPORT_XSCALE_12
+    0x00000000, // PA_CL_VPORT_XOFFSET_12
+    0x00000000, // PA_CL_VPORT_YSCALE_12
+    0x00000000, // PA_CL_VPORT_YOFFSET_12
+    0x00000000, // PA_CL_VPORT_ZSCALE_12
+    0x00000000, // PA_CL_VPORT_ZOFFSET_12
+    0x00000000, // PA_CL_VPORT_XSCALE_13
+    0x00000000, // PA_CL_VPORT_XOFFSET_13
+    0x00000000, // PA_CL_VPORT_YSCALE_13
+    0x00000000, // PA_CL_VPORT_YOFFSET_13
+    0x00000000, // PA_CL_VPORT_ZSCALE_13
+    0x00000000, // PA_CL_VPORT_ZOFFSET_13
+    0x00000000, // PA_CL_VPORT_XSCALE_14
+    0x00000000, // PA_CL_VPORT_XOFFSET_14
+    0x00000000, // PA_CL_VPORT_YSCALE_14
+    0x00000000, // PA_CL_VPORT_YOFFSET_14
+    0x00000000, // PA_CL_VPORT_ZSCALE_14
+    0x00000000, // PA_CL_VPORT_ZOFFSET_14
+    0x00000000, // PA_CL_VPORT_XSCALE_15
+    0x00000000, // PA_CL_VPORT_XOFFSET_15
+    0x00000000, // PA_CL_VPORT_YSCALE_15
+    0x00000000, // PA_CL_VPORT_YOFFSET_15
+    0x00000000, // PA_CL_VPORT_ZSCALE_15
+    0x00000000, // PA_CL_VPORT_ZOFFSET_15
+    0x00000000, // PA_CL_UCP_0_X
+    0x00000000, // PA_CL_UCP_0_Y
+    0x00000000, // PA_CL_UCP_0_Z
+    0x00000000, // PA_CL_UCP_0_W
+    0x00000000, // PA_CL_UCP_1_X
+    0x00000000, // PA_CL_UCP_1_Y
+    0x00000000, // PA_CL_UCP_1_Z
+    0x00000000, // PA_CL_UCP_1_W
+    0x00000000, // PA_CL_UCP_2_X
+    0x00000000, // PA_CL_UCP_2_Y
+    0x00000000, // PA_CL_UCP_2_Z
+    0x00000000, // PA_CL_UCP_2_W
+    0x00000000, // PA_CL_UCP_3_X
+    0x00000000, // PA_CL_UCP_3_Y
+    0x00000000, // PA_CL_UCP_3_Z
+    0x00000000, // PA_CL_UCP_3_W
+    0x00000000, // PA_CL_UCP_4_X
+    0x00000000, // PA_CL_UCP_4_Y
+    0x00000000, // PA_CL_UCP_4_Z
+    0x00000000, // PA_CL_UCP_4_W
+    0x00000000, // PA_CL_UCP_5_X
+    0x00000000, // PA_CL_UCP_5_Y
+    0x00000000, // PA_CL_UCP_5_Z
+    0x00000000, // PA_CL_UCP_5_W
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SPI_PS_INPUT_CNTL_0
+    0x00000000, // SPI_PS_INPUT_CNTL_1
+    0x00000000, // SPI_PS_INPUT_CNTL_2
+    0x00000000, // SPI_PS_INPUT_CNTL_3
+    0x00000000, // SPI_PS_INPUT_CNTL_4
+    0x00000000, // SPI_PS_INPUT_CNTL_5
+    0x00000000, // SPI_PS_INPUT_CNTL_6
+    0x00000000, // SPI_PS_INPUT_CNTL_7
+    0x00000000, // SPI_PS_INPUT_CNTL_8
+    0x00000000, // SPI_PS_INPUT_CNTL_9
+    0x00000000, // SPI_PS_INPUT_CNTL_10
+    0x00000000, // SPI_PS_INPUT_CNTL_11
+    0x00000000, // SPI_PS_INPUT_CNTL_12
+    0x00000000, // SPI_PS_INPUT_CNTL_13
+    0x00000000, // SPI_PS_INPUT_CNTL_14
+    0x00000000, // SPI_PS_INPUT_CNTL_15
+    0x00000000, // SPI_PS_INPUT_CNTL_16
+    0x00000000, // SPI_PS_INPUT_CNTL_17
+    0x00000000, // SPI_PS_INPUT_CNTL_18
+    0x00000000, // SPI_PS_INPUT_CNTL_19
+    0x00000000, // SPI_PS_INPUT_CNTL_20
+    0x00000000, // SPI_PS_INPUT_CNTL_21
+    0x00000000, // SPI_PS_INPUT_CNTL_22
+    0x00000000, // SPI_PS_INPUT_CNTL_23
+    0x00000000, // SPI_PS_INPUT_CNTL_24
+    0x00000000, // SPI_PS_INPUT_CNTL_25
+    0x00000000, // SPI_PS_INPUT_CNTL_26
+    0x00000000, // SPI_PS_INPUT_CNTL_27
+    0x00000000, // SPI_PS_INPUT_CNTL_28
+    0x00000000, // SPI_PS_INPUT_CNTL_29
+    0x00000000, // SPI_PS_INPUT_CNTL_30
+    0x00000000, // SPI_PS_INPUT_CNTL_31
+    0x00000000, // SPI_VS_OUT_CONFIG
+    0, // HOLE
+    0x00000000, // SPI_PS_INPUT_ENA
+    0x00000000, // SPI_PS_INPUT_ADDR
+    0x00000000, // SPI_INTERP_CONTROL_0
+    0x00000002, // SPI_PS_IN_CONTROL
+    0, // HOLE
+    0x00000000, // SPI_BARYC_CNTL
+    0, // HOLE
+    0x00000000, // SPI_TMPRING_SIZE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // SPI_WAVE_MGMT_1
+    0x00000000, // SPI_WAVE_MGMT_2
+    0x00000000, // SPI_SHADER_POS_FORMAT
+    0x00000000, // SPI_SHADER_Z_FORMAT
+    0x00000000, // SPI_SHADER_COL_FORMAT
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_BLEND0_CONTROL
+    0x00000000, // CB_BLEND1_CONTROL
+    0x00000000, // CB_BLEND2_CONTROL
+    0x00000000, // CB_BLEND3_CONTROL
+    0x00000000, // CB_BLEND4_CONTROL
+    0x00000000, // CB_BLEND5_CONTROL
+    0x00000000, // CB_BLEND6_CONTROL
+    0x00000000, // CB_BLEND7_CONTROL
+};
+static const u32 si_SECT_CONTEXT_def_3[] =
+{
+    0x00000000, // PA_CL_POINT_X_RAD
+    0x00000000, // PA_CL_POINT_Y_RAD
+    0x00000000, // PA_CL_POINT_SIZE
+    0x00000000, // PA_CL_POINT_CULL_RAD
+    0x00000000, // VGT_DMA_BASE_HI
+    0x00000000, // VGT_DMA_BASE
+};
+static const u32 si_SECT_CONTEXT_def_4[] =
+{
+    0x00000000, // DB_DEPTH_CONTROL
+    0x00000000, // DB_EQAA
+    0x00000000, // CB_COLOR_CONTROL
+    0x00000000, // DB_SHADER_CONTROL
+    0x00090000, // PA_CL_CLIP_CNTL
+    0x00000004, // PA_SU_SC_MODE_CNTL
+    0x00000000, // PA_CL_VTE_CNTL
+    0x00000000, // PA_CL_VS_OUT_CNTL
+    0x00000000, // PA_CL_NANINF_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_CNTL
+    0x00000000, // PA_SU_LINE_STIPPLE_SCALE
+    0x00000000, // PA_SU_PRIM_FILTER_CNTL
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // PA_SU_POINT_SIZE
+    0x00000000, // PA_SU_POINT_MINMAX
+    0x00000000, // PA_SU_LINE_CNTL
+    0x00000000, // PA_SC_LINE_STIPPLE
+    0x00000000, // VGT_OUTPUT_PATH_CNTL
+    0x00000000, // VGT_HOS_CNTL
+    0x00000000, // VGT_HOS_MAX_TESS_LEVEL
+    0x00000000, // VGT_HOS_MIN_TESS_LEVEL
+    0x00000000, // VGT_HOS_REUSE_DEPTH
+    0x00000000, // VGT_GROUP_PRIM_TYPE
+    0x00000000, // VGT_GROUP_FIRST_DECR
+    0x00000000, // VGT_GROUP_DECR
+    0x00000000, // VGT_GROUP_VECT_0_CNTL
+    0x00000000, // VGT_GROUP_VECT_1_CNTL
+    0x00000000, // VGT_GROUP_VECT_0_FMT_CNTL
+    0x00000000, // VGT_GROUP_VECT_1_FMT_CNTL
+    0x00000000, // VGT_GS_MODE
+    0, // HOLE
+    0x00000000, // PA_SC_MODE_CNTL_0
+    0x00000000, // PA_SC_MODE_CNTL_1
+    0x00000000, // VGT_ENHANCE
+    0x00000100, // VGT_GS_PER_ES
+    0x00000080, // VGT_ES_PER_GS
+    0x00000002, // VGT_GS_PER_VS
+    0x00000000, // VGT_GSVS_RING_OFFSET_1
+    0x00000000, // VGT_GSVS_RING_OFFSET_2
+    0x00000000, // VGT_GSVS_RING_OFFSET_3
+    0x00000000, // VGT_GS_OUT_PRIM_TYPE
+    0x00000000, // IA_ENHANCE
+};
+static const u32 si_SECT_CONTEXT_def_5[] =
+{
+    0x00000000, // VGT_PRIMITIVEID_EN
+};
+static const u32 si_SECT_CONTEXT_def_6[] =
+{
+    0x00000000, // VGT_PRIMITIVEID_RESET
+};
+static const u32 si_SECT_CONTEXT_def_7[] =
+{
+    0x00000000, // VGT_MULTI_PRIM_IB_RESET_EN
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_INSTANCE_STEP_RATE_0
+    0x00000000, // VGT_INSTANCE_STEP_RATE_1
+    0x000000ff, // IA_MULTI_VGT_PARAM
+    0x00000000, // VGT_ESGS_RING_ITEMSIZE
+    0x00000000, // VGT_GSVS_RING_ITEMSIZE
+    0x00000000, // VGT_REUSE_OFF
+    0x00000000, // VGT_VTX_CNT_EN
+    0x00000000, // DB_HTILE_SURFACE
+    0x00000000, // DB_SRESULTS_COMPARE_STATE0
+    0x00000000, // DB_SRESULTS_COMPARE_STATE1
+    0x00000000, // DB_PRELOAD_CONTROL
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_0
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_0
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_0
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_1
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_1
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_1
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_2
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_2
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_2
+    0x00000000, // VGT_STRMOUT_BUFFER_SIZE_3
+    0x00000000, // VGT_STRMOUT_VTX_STRIDE_3
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_BUFFER_OFFSET_3
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_OFFSET
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
+    0x00000000, // VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
+    0, // HOLE
+    0x00000000, // VGT_GS_MAX_VERT_OUT
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // VGT_SHADER_STAGES_EN
+    0x00000000, // VGT_LS_HS_CONFIG
+    0x00000000, // VGT_GS_VERT_ITEMSIZE
+    0x00000000, // VGT_GS_VERT_ITEMSIZE_1
+    0x00000000, // VGT_GS_VERT_ITEMSIZE_2
+    0x00000000, // VGT_GS_VERT_ITEMSIZE_3
+    0x00000000, // VGT_TF_PARAM
+    0x00000000, // DB_ALPHA_TO_MASK
+    0, // HOLE
+    0x00000000, // PA_SU_POLY_OFFSET_DB_FMT_CNTL
+    0x00000000, // PA_SU_POLY_OFFSET_CLAMP
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_FRONT_OFFSET
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_SCALE
+    0x00000000, // PA_SU_POLY_OFFSET_BACK_OFFSET
+    0x00000000, // VGT_GS_INSTANCE_CNT
+    0x00000000, // VGT_STRMOUT_CONFIG
+    0x00000000, // VGT_STRMOUT_BUFFER_CONFIG
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // PA_SC_CENTROID_PRIORITY_0
+    0x00000000, // PA_SC_CENTROID_PRIORITY_1
+    0x00001000, // PA_SC_LINE_CNTL
+    0x00000000, // PA_SC_AA_CONFIG
+    0x00000005, // PA_SU_VTX_CNTL
+    0x3f800000, // PA_CL_GB_VERT_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_VERT_DISC_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_CLIP_ADJ
+    0x3f800000, // PA_CL_GB_HORZ_DISC_ADJ
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y0_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y0_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X0Y1_3
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_0
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_1
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_2
+    0x00000000, // PA_SC_AA_SAMPLE_LOCS_PIXEL_X1Y1_3
+    0xffffffff, // PA_SC_AA_MASK_X0Y0_X1Y0
+    0xffffffff, // PA_SC_AA_MASK_X0Y1_X1Y1
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0, // HOLE
+    0x0000000e, // VGT_VERTEX_REUSE_BLOCK_CNTL
+    0x00000010, // VGT_OUT_DEALLOC_CNTL
+    0x00000000, // CB_COLOR0_BASE
+    0x00000000, // CB_COLOR0_PITCH
+    0x00000000, // CB_COLOR0_SLICE
+    0x00000000, // CB_COLOR0_VIEW
+    0x00000000, // CB_COLOR0_INFO
+    0x00000000, // CB_COLOR0_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR0_CMASK
+    0x00000000, // CB_COLOR0_CMASK_SLICE
+    0x00000000, // CB_COLOR0_FMASK
+    0x00000000, // CB_COLOR0_FMASK_SLICE
+    0x00000000, // CB_COLOR0_CLEAR_WORD0
+    0x00000000, // CB_COLOR0_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR1_BASE
+    0x00000000, // CB_COLOR1_PITCH
+    0x00000000, // CB_COLOR1_SLICE
+    0x00000000, // CB_COLOR1_VIEW
+    0x00000000, // CB_COLOR1_INFO
+    0x00000000, // CB_COLOR1_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR1_CMASK
+    0x00000000, // CB_COLOR1_CMASK_SLICE
+    0x00000000, // CB_COLOR1_FMASK
+    0x00000000, // CB_COLOR1_FMASK_SLICE
+    0x00000000, // CB_COLOR1_CLEAR_WORD0
+    0x00000000, // CB_COLOR1_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR2_BASE
+    0x00000000, // CB_COLOR2_PITCH
+    0x00000000, // CB_COLOR2_SLICE
+    0x00000000, // CB_COLOR2_VIEW
+    0x00000000, // CB_COLOR2_INFO
+    0x00000000, // CB_COLOR2_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR2_CMASK
+    0x00000000, // CB_COLOR2_CMASK_SLICE
+    0x00000000, // CB_COLOR2_FMASK
+    0x00000000, // CB_COLOR2_FMASK_SLICE
+    0x00000000, // CB_COLOR2_CLEAR_WORD0
+    0x00000000, // CB_COLOR2_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR3_BASE
+    0x00000000, // CB_COLOR3_PITCH
+    0x00000000, // CB_COLOR3_SLICE
+    0x00000000, // CB_COLOR3_VIEW
+    0x00000000, // CB_COLOR3_INFO
+    0x00000000, // CB_COLOR3_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR3_CMASK
+    0x00000000, // CB_COLOR3_CMASK_SLICE
+    0x00000000, // CB_COLOR3_FMASK
+    0x00000000, // CB_COLOR3_FMASK_SLICE
+    0x00000000, // CB_COLOR3_CLEAR_WORD0
+    0x00000000, // CB_COLOR3_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR4_BASE
+    0x00000000, // CB_COLOR4_PITCH
+    0x00000000, // CB_COLOR4_SLICE
+    0x00000000, // CB_COLOR4_VIEW
+    0x00000000, // CB_COLOR4_INFO
+    0x00000000, // CB_COLOR4_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR4_CMASK
+    0x00000000, // CB_COLOR4_CMASK_SLICE
+    0x00000000, // CB_COLOR4_FMASK
+    0x00000000, // CB_COLOR4_FMASK_SLICE
+    0x00000000, // CB_COLOR4_CLEAR_WORD0
+    0x00000000, // CB_COLOR4_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR5_BASE
+    0x00000000, // CB_COLOR5_PITCH
+    0x00000000, // CB_COLOR5_SLICE
+    0x00000000, // CB_COLOR5_VIEW
+    0x00000000, // CB_COLOR5_INFO
+    0x00000000, // CB_COLOR5_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR5_CMASK
+    0x00000000, // CB_COLOR5_CMASK_SLICE
+    0x00000000, // CB_COLOR5_FMASK
+    0x00000000, // CB_COLOR5_FMASK_SLICE
+    0x00000000, // CB_COLOR5_CLEAR_WORD0
+    0x00000000, // CB_COLOR5_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR6_BASE
+    0x00000000, // CB_COLOR6_PITCH
+    0x00000000, // CB_COLOR6_SLICE
+    0x00000000, // CB_COLOR6_VIEW
+    0x00000000, // CB_COLOR6_INFO
+    0x00000000, // CB_COLOR6_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR6_CMASK
+    0x00000000, // CB_COLOR6_CMASK_SLICE
+    0x00000000, // CB_COLOR6_FMASK
+    0x00000000, // CB_COLOR6_FMASK_SLICE
+    0x00000000, // CB_COLOR6_CLEAR_WORD0
+    0x00000000, // CB_COLOR6_CLEAR_WORD1
+    0, // HOLE
+    0, // HOLE
+    0x00000000, // CB_COLOR7_BASE
+    0x00000000, // CB_COLOR7_PITCH
+    0x00000000, // CB_COLOR7_SLICE
+    0x00000000, // CB_COLOR7_VIEW
+    0x00000000, // CB_COLOR7_INFO
+    0x00000000, // CB_COLOR7_ATTRIB
+    0, // HOLE
+    0x00000000, // CB_COLOR7_CMASK
+    0x00000000, // CB_COLOR7_CMASK_SLICE
+    0x00000000, // CB_COLOR7_FMASK
+    0x00000000, // CB_COLOR7_FMASK_SLICE
+    0x00000000, // CB_COLOR7_CLEAR_WORD0
+    0x00000000, // CB_COLOR7_CLEAR_WORD1
+};
+static const struct cs_extent_def si_SECT_CONTEXT_defs[] =
+{
+    {si_SECT_CONTEXT_def_1, 0x0000a000, 212 },
+    {si_SECT_CONTEXT_def_2, 0x0000a0d8, 272 },
+    {si_SECT_CONTEXT_def_3, 0x0000a1f5, 6 },
+    {si_SECT_CONTEXT_def_4, 0x0000a200, 157 },
+    {si_SECT_CONTEXT_def_5, 0x0000a2a1, 1 },
+    {si_SECT_CONTEXT_def_6, 0x0000a2a3, 1 },
+    {si_SECT_CONTEXT_def_7, 0x0000a2a5, 233 },
+    { 0, 0, 0 }
+};
+static const struct cs_section_def si_cs_data[] = {
+    { si_SECT_CONTEXT_defs, SECT_CONTEXT },
+    { 0, SECT_NONE }
+};
diff --git a/drivers/gpu/drm/radeon/cypress_dpm.c b/drivers/gpu/drm/radeon/cypress_dpm.c
new file mode 100644
index 000000000000..9bcdd174780f
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cypress_dpm.c
@@ -0,0 +1,2189 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "evergreend.h"
+#include "r600_dpm.h"
+#include "cypress_dpm.h"
+#include "atom.h"
+
+#define SMC_RAM_END 0x8000
+
+#define MC_CG_ARB_FREQ_F0           0x0a
+#define MC_CG_ARB_FREQ_F1           0x0b
+#define MC_CG_ARB_FREQ_F2           0x0c
+#define MC_CG_ARB_FREQ_F3           0x0d
+
+#define MC_CG_SEQ_DRAMCONF_S0       0x05
+#define MC_CG_SEQ_DRAMCONF_S1       0x06
+#define MC_CG_SEQ_YCLK_SUSPEND      0x04
+#define MC_CG_SEQ_YCLK_RESUME       0x0a
+
+struct rv7xx_ps *rv770_get_ps(struct radeon_ps *rps);
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev);
+struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev);
+
+static void cypress_enable_bif_dynamic_pcie_gen2(struct radeon_device *rdev,
+						 bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp, bif;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	if (enable) {
+		if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+		    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
+			if (!pi->boot_in_gen2) {
+				bif = RREG32(CG_BIF_REQ_AND_RSP) & ~CG_CLIENT_REQ_MASK;
+				bif |= CG_CLIENT_REQ(0xd);
+				WREG32(CG_BIF_REQ_AND_RSP, bif);
+
+				tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+				tmp |= LC_HW_VOLTAGE_IF_CONTROL(1);
+				tmp |= LC_GEN2_EN_STRAP;
+
+				tmp |= LC_CLR_FAILED_SPD_CHANGE_CNT;
+				WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+				udelay(10);
+				tmp &= ~LC_CLR_FAILED_SPD_CHANGE_CNT;
+				WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+			}
+		}
+	} else {
+		if (!pi->boot_in_gen2) {
+			tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+			tmp &= ~LC_GEN2_EN_STRAP;
+		}
+		if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) ||
+		    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2))
+			WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+	}
+}
+
+static void cypress_enable_dynamic_pcie_gen2(struct radeon_device *rdev,
+					     bool enable)
+{
+	cypress_enable_bif_dynamic_pcie_gen2(rdev, enable);
+
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, ENABLE_GEN2PCIE, ~ENABLE_GEN2PCIE);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~ENABLE_GEN2PCIE);
+}
+
+#if 0
+static int cypress_enter_ulp_state(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (pi->gfx_clock_gating) {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+
+		RREG32(GB_ADDR_CONFIG);
+	}
+
+	WREG32_P(SMC_MSG, HOST_SMC_MSG(PPSMC_MSG_SwitchToMinimumPower),
+		 ~HOST_SMC_MSG_MASK);
+
+	udelay(7000);
+
+	return 0;
+}
+#endif
+
+static void cypress_gfx_clock_gating_enable(struct radeon_device *rdev,
+					    bool enable)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (enable) {
+		if (eg_pi->light_sleep) {
+			WREG32(GRBM_GFX_INDEX, 0xC0000000);
+
+			WREG32_CG(CG_CGLS_TILE_0, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_1, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_2, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_3, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_4, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_5, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_6, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_7, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_8, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_9, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_10, 0xFFFFFFFF);
+			WREG32_CG(CG_CGLS_TILE_11, 0xFFFFFFFF);
+
+			WREG32_P(SCLK_PWRMGT_CNTL, DYN_LIGHT_SLEEP_EN, ~DYN_LIGHT_SLEEP_EN);
+		}
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+	} else {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+		RREG32(GB_ADDR_CONFIG);
+
+		if (eg_pi->light_sleep) {
+			WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_LIGHT_SLEEP_EN);
+
+			WREG32(GRBM_GFX_INDEX, 0xC0000000);
+
+			WREG32_CG(CG_CGLS_TILE_0, 0);
+			WREG32_CG(CG_CGLS_TILE_1, 0);
+			WREG32_CG(CG_CGLS_TILE_2, 0);
+			WREG32_CG(CG_CGLS_TILE_3, 0);
+			WREG32_CG(CG_CGLS_TILE_4, 0);
+			WREG32_CG(CG_CGLS_TILE_5, 0);
+			WREG32_CG(CG_CGLS_TILE_6, 0);
+			WREG32_CG(CG_CGLS_TILE_7, 0);
+			WREG32_CG(CG_CGLS_TILE_8, 0);
+			WREG32_CG(CG_CGLS_TILE_9, 0);
+			WREG32_CG(CG_CGLS_TILE_10, 0);
+			WREG32_CG(CG_CGLS_TILE_11, 0);
+		}
+	}
+}
+
+static void cypress_mg_clock_gating_enable(struct radeon_device *rdev,
+					   bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (enable) {
+		u32 cgts_sm_ctrl_reg;
+
+		if (rdev->family == CHIP_CEDAR)
+			cgts_sm_ctrl_reg = CEDAR_MGCGCGTSSMCTRL_DFLT;
+		else if (rdev->family == CHIP_REDWOOD)
+			cgts_sm_ctrl_reg = REDWOOD_MGCGCGTSSMCTRL_DFLT;
+		else
+			cgts_sm_ctrl_reg = CYPRESS_MGCGCGTSSMCTRL_DFLT;
+
+		WREG32(GRBM_GFX_INDEX, 0xC0000000);
+
+		WREG32_CG(CG_CGTT_LOCAL_0, CYPRESS_MGCGTTLOCAL0_DFLT);
+		WREG32_CG(CG_CGTT_LOCAL_1, CYPRESS_MGCGTTLOCAL1_DFLT & 0xFFFFCFFF);
+		WREG32_CG(CG_CGTT_LOCAL_2, CYPRESS_MGCGTTLOCAL2_DFLT);
+		WREG32_CG(CG_CGTT_LOCAL_3, CYPRESS_MGCGTTLOCAL3_DFLT);
+
+		if (pi->mgcgtssm)
+			WREG32(CGTS_SM_CTRL_REG, cgts_sm_ctrl_reg);
+
+		if (eg_pi->mcls) {
+			WREG32_P(MC_CITF_MISC_RD_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(MC_CITF_MISC_WR_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(MC_CITF_MISC_VM_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(MC_HUB_MISC_HUB_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(MC_HUB_MISC_VM_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(MC_HUB_MISC_SIP_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(MC_XPB_CLK_GAT, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+			WREG32_P(VM_L2_CG, MEM_LS_ENABLE, ~MEM_LS_ENABLE);
+		}
+	} else {
+		WREG32(GRBM_GFX_INDEX, 0xC0000000);
+
+		WREG32_CG(CG_CGTT_LOCAL_0, 0xFFFFFFFF);
+		WREG32_CG(CG_CGTT_LOCAL_1, 0xFFFFFFFF);
+		WREG32_CG(CG_CGTT_LOCAL_2, 0xFFFFFFFF);
+		WREG32_CG(CG_CGTT_LOCAL_3, 0xFFFFFFFF);
+
+		if (pi->mgcgtssm)
+			WREG32(CGTS_SM_CTRL_REG, 0x81f44bc0);
+	}
+}
+
+void cypress_enable_spread_spectrum(struct radeon_device *rdev,
+				    bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (enable) {
+		if (pi->sclk_ss)
+			WREG32_P(GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, ~DYN_SPREAD_SPECTRUM_EN);
+
+		if (pi->mclk_ss)
+			WREG32_P(MPLL_CNTL_MODE, SS_SSEN, ~SS_SSEN);
+	} else {
+		WREG32_P(CG_SPLL_SPREAD_SPECTRUM, 0, ~SSEN);
+		WREG32_P(GENERAL_PWRMGT, 0, ~DYN_SPREAD_SPECTRUM_EN);
+		WREG32_P(MPLL_CNTL_MODE, 0, ~SS_SSEN);
+		WREG32_P(MPLL_CNTL_MODE, 0, ~SS_DSMODE_EN);
+	}
+}
+
+void cypress_start_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN);
+}
+
+void cypress_enable_sclk_control(struct radeon_device *rdev,
+				 bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_OFF);
+	else
+		WREG32_P(SCLK_PWRMGT_CNTL, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF);
+}
+
+void cypress_enable_mclk_control(struct radeon_device *rdev,
+				 bool enable)
+{
+	if (enable)
+		WREG32_P(MCLK_PWRMGT_CNTL, 0, ~MPLL_PWRMGT_OFF);
+	else
+		WREG32_P(MCLK_PWRMGT_CNTL, MPLL_PWRMGT_OFF, ~MPLL_PWRMGT_OFF);
+}
+
+int cypress_notify_smc_display_change(struct radeon_device *rdev,
+				      bool has_display)
+{
+	PPSMC_Msg msg = has_display ?
+		(PPSMC_Msg)PPSMC_MSG_HasDisplay : (PPSMC_Msg)PPSMC_MSG_NoDisplay;
+
+	if (rv770_send_msg_to_smc(rdev, msg) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return 0;
+}
+
+void cypress_program_response_times(struct radeon_device *rdev)
+{
+	u32 reference_clock;
+	u32 mclk_switch_limit;
+
+	reference_clock = radeon_get_xclk(rdev);
+	mclk_switch_limit = (460 * reference_clock) / 100;
+
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_mclk_switch_lim,
+				      mclk_switch_limit);
+
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_mvdd_chg_time, 1);
+
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_mc_block_delay, 0xAA);
+
+	rv770_program_response_times(rdev);
+
+	if (ASIC_IS_LOMBOK(rdev))
+		rv770_write_smc_soft_register(rdev,
+					      RV770_SMC_SOFT_REGISTER_is_asic_lombok, 1);
+
+}
+
+static int cypress_pcie_performance_request(struct radeon_device *rdev,
+					    u8 perf_req, bool advertise)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 tmp;
+
+	udelay(10);
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	if ((perf_req == PCIE_PERF_REQ_PECI_GEN1) && (tmp & LC_CURRENT_DATA_RATE))
+		return 0;
+
+#if defined(CONFIG_ACPI)
+	if ((perf_req == PCIE_PERF_REQ_PECI_GEN1) ||
+	    (perf_req == PCIE_PERF_REQ_PECI_GEN2)) {
+		eg_pi->pcie_performance_request_registered = true;
+		return radeon_acpi_pcie_performance_request(rdev, perf_req, advertise);
+	} else if ((perf_req == PCIE_PERF_REQ_REMOVE_REGISTRY) &&
+		   eg_pi->pcie_performance_request_registered) {
+		eg_pi->pcie_performance_request_registered = false;
+		return radeon_acpi_pcie_performance_request(rdev, perf_req, advertise);
+	}
+#endif
+
+	return 0;
+}
+
+void cypress_advertise_gen2_capability(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp;
+
+#if defined(CONFIG_ACPI)
+	radeon_acpi_pcie_notify_device_ready(rdev);
+#endif
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+
+	if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+	    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2))
+		pi->pcie_gen2 = true;
+	else
+		pi->pcie_gen2 = false;
+
+	if (!pi->pcie_gen2)
+		cypress_pcie_performance_request(rdev, PCIE_PERF_REQ_PECI_GEN2, true);
+
+}
+
+static enum radeon_pcie_gen cypress_get_maximum_link_speed(struct radeon_ps *radeon_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+
+	if (state->high.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2)
+		return 1;
+	return 0;
+}
+
+void cypress_notify_link_speed_change_after_state_change(struct radeon_device *rdev,
+							 struct radeon_ps *radeon_new_state,
+							 struct radeon_ps *radeon_current_state)
+{
+	enum radeon_pcie_gen pcie_link_speed_target =
+		cypress_get_maximum_link_speed(radeon_new_state);
+	enum radeon_pcie_gen pcie_link_speed_current =
+		cypress_get_maximum_link_speed(radeon_current_state);
+	u8 request;
+
+	if (pcie_link_speed_target < pcie_link_speed_current) {
+		if (pcie_link_speed_target == RADEON_PCIE_GEN1)
+			request = PCIE_PERF_REQ_PECI_GEN1;
+		else if (pcie_link_speed_target == RADEON_PCIE_GEN2)
+			request = PCIE_PERF_REQ_PECI_GEN2;
+		else
+			request = PCIE_PERF_REQ_PECI_GEN3;
+
+		cypress_pcie_performance_request(rdev, request, false);
+	}
+}
+
+void cypress_notify_link_speed_change_before_state_change(struct radeon_device *rdev,
+							  struct radeon_ps *radeon_new_state,
+							  struct radeon_ps *radeon_current_state)
+{
+	enum radeon_pcie_gen pcie_link_speed_target =
+		cypress_get_maximum_link_speed(radeon_new_state);
+	enum radeon_pcie_gen pcie_link_speed_current =
+		cypress_get_maximum_link_speed(radeon_current_state);
+	u8 request;
+
+	if (pcie_link_speed_target > pcie_link_speed_current) {
+		if (pcie_link_speed_target == RADEON_PCIE_GEN1)
+			request = PCIE_PERF_REQ_PECI_GEN1;
+		else if (pcie_link_speed_target == RADEON_PCIE_GEN2)
+			request = PCIE_PERF_REQ_PECI_GEN2;
+		else
+			request = PCIE_PERF_REQ_PECI_GEN3;
+
+		cypress_pcie_performance_request(rdev, request, false);
+	}
+}
+
+static int cypress_populate_voltage_value(struct radeon_device *rdev,
+					  struct atom_voltage_table *table,
+					  u16 value, RV770_SMC_VOLTAGE_VALUE *voltage)
+{
+	unsigned int i;
+
+	for (i = 0; i < table->count; i++) {
+		if (value <= table->entries[i].value) {
+			voltage->index = (u8)i;
+			voltage->value = cpu_to_be16(table->entries[i].value);
+			break;
+		}
+	}
+
+	if (i == table->count)
+		return -EINVAL;
+
+	return 0;
+}
+
+u8 cypress_get_strobe_mode_settings(struct radeon_device *rdev, u32 mclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 result = 0;
+	bool strobe_mode = false;
+
+	if (pi->mem_gddr5) {
+		if (mclk <= pi->mclk_strobe_mode_threshold)
+			strobe_mode = true;
+		result = cypress_get_mclk_frequency_ratio(rdev, mclk, strobe_mode);
+
+		if (strobe_mode)
+			result |= SMC_STROBE_ENABLE;
+	}
+
+	return result;
+}
+
+u32 cypress_map_clkf_to_ibias(struct radeon_device *rdev, u32 clkf)
+{
+	u32 ref_clk = rdev->clock.mpll.reference_freq;
+	u32 vco = clkf * ref_clk;
+
+	/* 100 Mhz ref clk */
+	if (ref_clk == 10000) {
+		if (vco > 500000)
+			return 0xC6;
+		if (vco > 400000)
+			return 0x9D;
+		if (vco > 330000)
+			return 0x6C;
+		if (vco > 250000)
+			return 0x2B;
+		if (vco >  160000)
+			return 0x5B;
+		if (vco > 120000)
+			return 0x0A;
+		return 0x4B;
+	}
+
+	/* 27 Mhz ref clk */
+	if (vco > 250000)
+		return 0x8B;
+	if (vco > 200000)
+		return 0xCC;
+	if (vco > 150000)
+		return 0x9B;
+	return 0x6B;
+}
+
+static int cypress_populate_mclk_value(struct radeon_device *rdev,
+				       u32 engine_clock, u32 memory_clock,
+				       RV7XX_SMC_MCLK_VALUE *mclk,
+				       bool strobe_mode, bool dll_state_on)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	u32 mpll_ad_func_cntl =
+		pi->clk_regs.rv770.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl =
+		pi->clk_regs.rv770.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_dq_func_cntl_2;
+	u32 mclk_pwrmgt_cntl =
+		pi->clk_regs.rv770.mclk_pwrmgt_cntl;
+	u32 dll_cntl =
+		pi->clk_regs.rv770.dll_cntl;
+	u32 mpll_ss1 = pi->clk_regs.rv770.mpll_ss1;
+	u32 mpll_ss2 = pi->clk_regs.rv770.mpll_ss2;
+	struct atom_clock_dividers dividers;
+	u32 ibias;
+	u32 dll_speed;
+	int ret;
+	u32 mc_seq_misc7;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					     memory_clock, strobe_mode, &dividers);
+	if (ret)
+		return ret;
+
+	if (!strobe_mode) {
+		mc_seq_misc7 = RREG32(MC_SEQ_MISC7);
+
+		if(mc_seq_misc7 & 0x8000000)
+			dividers.post_div = 1;
+	}
+
+	ibias = cypress_map_clkf_to_ibias(rdev, dividers.whole_fb_div);
+
+	mpll_ad_func_cntl &= ~(CLKR_MASK |
+			       YCLK_POST_DIV_MASK |
+			       CLKF_MASK |
+			       CLKFRAC_MASK |
+			       IBIAS_MASK);
+	mpll_ad_func_cntl |= CLKR(dividers.ref_div);
+	mpll_ad_func_cntl |= YCLK_POST_DIV(dividers.post_div);
+	mpll_ad_func_cntl |= CLKF(dividers.whole_fb_div);
+	mpll_ad_func_cntl |= CLKFRAC(dividers.frac_fb_div);
+	mpll_ad_func_cntl |= IBIAS(ibias);
+
+	if (dividers.vco_mode)
+		mpll_ad_func_cntl_2 |= VCO_MODE;
+	else
+		mpll_ad_func_cntl_2 &= ~VCO_MODE;
+
+	if (pi->mem_gddr5) {
+		mpll_dq_func_cntl &= ~(CLKR_MASK |
+				       YCLK_POST_DIV_MASK |
+				       CLKF_MASK |
+				       CLKFRAC_MASK |
+				       IBIAS_MASK);
+		mpll_dq_func_cntl |= CLKR(dividers.ref_div);
+		mpll_dq_func_cntl |= YCLK_POST_DIV(dividers.post_div);
+		mpll_dq_func_cntl |= CLKF(dividers.whole_fb_div);
+		mpll_dq_func_cntl |= CLKFRAC(dividers.frac_fb_div);
+		mpll_dq_func_cntl |= IBIAS(ibias);
+
+		if (strobe_mode)
+			mpll_dq_func_cntl &= ~PDNB;
+		else
+			mpll_dq_func_cntl |= PDNB;
+
+		if (dividers.vco_mode)
+			mpll_dq_func_cntl_2 |= VCO_MODE;
+		else
+			mpll_dq_func_cntl_2 &= ~VCO_MODE;
+	}
+
+	if (pi->mclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = memory_clock * dividers.post_div;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_MEMORY_SS, vco_freq)) {
+			u32 reference_clock = rdev->clock.mpll.reference_freq;
+			u32 decoded_ref = rv740_get_decoded_reference_divider(dividers.ref_div);
+			u32 clk_s = reference_clock * 5 / (decoded_ref * ss.rate);
+			u32 clk_v = ss.percentage *
+				(0x4000 * dividers.whole_fb_div + 0x800 * dividers.frac_fb_div) / (clk_s * 625);
+
+			mpll_ss1 &= ~CLKV_MASK;
+			mpll_ss1 |= CLKV(clk_v);
+
+			mpll_ss2 &= ~CLKS_MASK;
+			mpll_ss2 |= CLKS(clk_s);
+		}
+	}
+
+	dll_speed = rv740_get_dll_speed(pi->mem_gddr5,
+					memory_clock);
+
+	mclk_pwrmgt_cntl &= ~DLL_SPEED_MASK;
+	mclk_pwrmgt_cntl |= DLL_SPEED(dll_speed);
+	if (dll_state_on)
+		mclk_pwrmgt_cntl |= (MRDCKA0_PDNB |
+				     MRDCKA1_PDNB |
+				     MRDCKB0_PDNB |
+				     MRDCKB1_PDNB |
+				     MRDCKC0_PDNB |
+				     MRDCKC1_PDNB |
+				     MRDCKD0_PDNB |
+				     MRDCKD1_PDNB);
+	else
+		mclk_pwrmgt_cntl &= ~(MRDCKA0_PDNB |
+				      MRDCKA1_PDNB |
+				      MRDCKB0_PDNB |
+				      MRDCKB1_PDNB |
+				      MRDCKC0_PDNB |
+				      MRDCKC1_PDNB |
+				      MRDCKD0_PDNB |
+				      MRDCKD1_PDNB);
+
+	mclk->mclk770.mclk_value = cpu_to_be32(memory_clock);
+	mclk->mclk770.vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	mclk->mclk770.vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	mclk->mclk770.vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	mclk->mclk770.vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+	mclk->mclk770.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	mclk->mclk770.vDLL_CNTL = cpu_to_be32(dll_cntl);
+	mclk->mclk770.vMPLL_SS = cpu_to_be32(mpll_ss1);
+	mclk->mclk770.vMPLL_SS2 = cpu_to_be32(mpll_ss2);
+
+	return 0;
+}
+
+u8 cypress_get_mclk_frequency_ratio(struct radeon_device *rdev,
+				    u32 memory_clock, bool strobe_mode)
+{
+	u8 mc_para_index;
+
+	if (rdev->family >= CHIP_BARTS) {
+		if (strobe_mode) {
+			if (memory_clock < 10000)
+				mc_para_index = 0x00;
+			else if (memory_clock > 47500)
+				mc_para_index = 0x0f;
+			else
+				mc_para_index = (u8)((memory_clock - 10000) / 2500);
+		} else {
+			if (memory_clock < 65000)
+				mc_para_index = 0x00;
+			else if (memory_clock > 135000)
+				mc_para_index = 0x0f;
+			else
+				mc_para_index = (u8)((memory_clock - 60000) / 5000);
+		}
+	} else {
+		if (strobe_mode) {
+			if (memory_clock < 10000)
+				mc_para_index = 0x00;
+			else if (memory_clock > 47500)
+				mc_para_index = 0x0f;
+			else
+				mc_para_index = (u8)((memory_clock - 10000) / 2500);
+		} else {
+			if (memory_clock < 40000)
+				mc_para_index = 0x00;
+			else if (memory_clock > 115000)
+				mc_para_index = 0x0f;
+			else
+				mc_para_index = (u8)((memory_clock - 40000) / 5000);
+		}
+	}
+	return mc_para_index;
+}
+
+static int cypress_populate_mvdd_value(struct radeon_device *rdev,
+				       u32 mclk,
+				       RV770_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (!pi->mvdd_control) {
+		voltage->index = eg_pi->mvdd_high_index;
+		voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+		return 0;
+	}
+
+	if (mclk <= pi->mvdd_split_frequency) {
+		voltage->index = eg_pi->mvdd_low_index;
+		voltage->value = cpu_to_be16(MVDD_LOW_VALUE);
+	} else {
+		voltage->index = eg_pi->mvdd_high_index;
+		voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+	}
+
+	return 0;
+}
+
+int cypress_convert_power_level_to_smc(struct radeon_device *rdev,
+				       struct rv7xx_pl *pl,
+				       RV770_SMC_HW_PERFORMANCE_LEVEL *level,
+				       u8 watermark_level)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	int ret;
+	bool dll_state_on;
+
+	level->gen2PCIE = pi->pcie_gen2 ?
+		((pl->flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? 1 : 0) : 0;
+	level->gen2XSP  = (pl->flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? 1 : 0;
+	level->backbias = (pl->flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ? 1 : 0;
+	level->displayWatermark = watermark_level;
+
+	ret = rv740_populate_sclk_value(rdev, pl->sclk, &level->sclk);
+	if (ret)
+		return ret;
+
+	level->mcFlags =  0;
+	if (pi->mclk_stutter_mode_threshold &&
+	    (pl->mclk <= pi->mclk_stutter_mode_threshold) &&
+	    !eg_pi->uvd_enabled) {
+		level->mcFlags |= SMC_MC_STUTTER_EN;
+		if (eg_pi->sclk_deep_sleep)
+			level->stateFlags |= PPSMC_STATEFLAG_AUTO_PULSE_SKIP;
+		else
+			level->stateFlags &= ~PPSMC_STATEFLAG_AUTO_PULSE_SKIP;
+	}
+
+	if (pi->mem_gddr5) {
+		if (pl->mclk > pi->mclk_edc_enable_threshold)
+			level->mcFlags |= SMC_MC_EDC_RD_FLAG;
+
+		if (pl->mclk > eg_pi->mclk_edc_wr_enable_threshold)
+			level->mcFlags |= SMC_MC_EDC_WR_FLAG;
+
+		level->strobeMode = cypress_get_strobe_mode_settings(rdev, pl->mclk);
+
+		if (level->strobeMode & SMC_STROBE_ENABLE) {
+			if (cypress_get_mclk_frequency_ratio(rdev, pl->mclk, true) >=
+			    ((RREG32(MC_SEQ_MISC7) >> 16) & 0xf))
+				dll_state_on = ((RREG32(MC_SEQ_MISC5) >> 1) & 0x1) ? true : false;
+			else
+				dll_state_on = ((RREG32(MC_SEQ_MISC6) >> 1) & 0x1) ? true : false;
+		} else
+			dll_state_on = eg_pi->dll_default_on;
+
+		ret = cypress_populate_mclk_value(rdev,
+						  pl->sclk,
+						  pl->mclk,
+						  &level->mclk,
+						  (level->strobeMode & SMC_STROBE_ENABLE) != 0,
+						  dll_state_on);
+	} else {
+		ret = cypress_populate_mclk_value(rdev,
+						  pl->sclk,
+						  pl->mclk,
+						  &level->mclk,
+						  true,
+						  true);
+	}
+	if (ret)
+		return ret;
+
+	ret = cypress_populate_voltage_value(rdev,
+					     &eg_pi->vddc_voltage_table,
+					     pl->vddc,
+					     &level->vddc);
+	if (ret)
+		return ret;
+
+	if (eg_pi->vddci_control) {
+		ret = cypress_populate_voltage_value(rdev,
+						     &eg_pi->vddci_voltage_table,
+						     pl->vddci,
+						     &level->vddci);
+		if (ret)
+			return ret;
+	}
+
+	ret = cypress_populate_mvdd_value(rdev, pl->mclk, &level->mvdd);
+
+	return ret;
+}
+
+static int cypress_convert_power_state_to_smc(struct radeon_device *rdev,
+					      struct radeon_ps *radeon_state,
+					      RV770_SMC_SWSTATE *smc_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	int ret;
+
+	if (!(radeon_state->caps & ATOM_PPLIB_DISALLOW_ON_DC))
+		smc_state->flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	ret = cypress_convert_power_level_to_smc(rdev,
+						 &state->low,
+						 &smc_state->levels[0],
+						 PPSMC_DISPLAY_WATERMARK_LOW);
+	if (ret)
+		return ret;
+
+	ret = cypress_convert_power_level_to_smc(rdev,
+						 &state->medium,
+						 &smc_state->levels[1],
+						 PPSMC_DISPLAY_WATERMARK_LOW);
+	if (ret)
+		return ret;
+
+	ret = cypress_convert_power_level_to_smc(rdev,
+						 &state->high,
+						 &smc_state->levels[2],
+						 PPSMC_DISPLAY_WATERMARK_HIGH);
+	if (ret)
+		return ret;
+
+	smc_state->levels[0].arbValue = MC_CG_ARB_FREQ_F1;
+	smc_state->levels[1].arbValue = MC_CG_ARB_FREQ_F2;
+	smc_state->levels[2].arbValue = MC_CG_ARB_FREQ_F3;
+
+	if (eg_pi->dynamic_ac_timing) {
+		smc_state->levels[0].ACIndex = 2;
+		smc_state->levels[1].ACIndex = 3;
+		smc_state->levels[2].ACIndex = 4;
+	} else {
+		smc_state->levels[0].ACIndex = 0;
+		smc_state->levels[1].ACIndex = 0;
+		smc_state->levels[2].ACIndex = 0;
+	}
+
+	rv770_populate_smc_sp(rdev, radeon_state, smc_state);
+
+	return rv770_populate_smc_t(rdev, radeon_state, smc_state);
+}
+
+static void cypress_convert_mc_registers(struct evergreen_mc_reg_entry *entry,
+					 SMC_Evergreen_MCRegisterSet *data,
+					 u32 num_entries, u32 valid_flag)
+{
+	u32 i, j;
+
+	for (i = 0, j = 0; j < num_entries; j++) {
+		if (valid_flag & (1 << j)) {
+			data->value[i] = cpu_to_be32(entry->mc_data[j]);
+			i++;
+		}
+	}
+}
+
+static void cypress_convert_mc_reg_table_entry_to_smc(struct radeon_device *rdev,
+						      struct rv7xx_pl *pl,
+						      SMC_Evergreen_MCRegisterSet *mc_reg_table_data)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 i = 0;
+
+	for (i = 0; i < eg_pi->mc_reg_table.num_entries; i++) {
+		if (pl->mclk <=
+		    eg_pi->mc_reg_table.mc_reg_table_entry[i].mclk_max)
+			break;
+	}
+
+	if ((i == eg_pi->mc_reg_table.num_entries) && (i > 0))
+		--i;
+
+	cypress_convert_mc_registers(&eg_pi->mc_reg_table.mc_reg_table_entry[i],
+				     mc_reg_table_data,
+				     eg_pi->mc_reg_table.last,
+				     eg_pi->mc_reg_table.valid_flag);
+}
+
+static void cypress_convert_mc_reg_table_to_smc(struct radeon_device *rdev,
+						struct radeon_ps *radeon_state,
+						SMC_Evergreen_MCRegisters *mc_reg_table)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+
+	cypress_convert_mc_reg_table_entry_to_smc(rdev,
+						  &state->low,
+						  &mc_reg_table->data[2]);
+	cypress_convert_mc_reg_table_entry_to_smc(rdev,
+						  &state->medium,
+						  &mc_reg_table->data[3]);
+	cypress_convert_mc_reg_table_entry_to_smc(rdev,
+						  &state->high,
+						  &mc_reg_table->data[4]);
+}
+
+int cypress_upload_sw_state(struct radeon_device *rdev,
+			    struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u16 address = pi->state_table_start +
+		offsetof(RV770_SMC_STATETABLE, driverState);
+	RV770_SMC_SWSTATE state = { 0 };
+	int ret;
+
+	ret = cypress_convert_power_state_to_smc(rdev, radeon_new_state, &state);
+	if (ret)
+		return ret;
+
+	return rv770_copy_bytes_to_smc(rdev, address, (u8 *)&state,
+				    sizeof(RV770_SMC_SWSTATE),
+				    pi->sram_end);
+}
+
+int cypress_upload_mc_reg_table(struct radeon_device *rdev,
+				struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	SMC_Evergreen_MCRegisters mc_reg_table = { 0 };
+	u16 address;
+
+	cypress_convert_mc_reg_table_to_smc(rdev, radeon_new_state, &mc_reg_table);
+
+	address = eg_pi->mc_reg_table_start +
+		(u16)offsetof(SMC_Evergreen_MCRegisters, data[2]);
+
+	return rv770_copy_bytes_to_smc(rdev, address,
+				       (u8 *)&mc_reg_table.data[2],
+				       sizeof(SMC_Evergreen_MCRegisterSet) * 3,
+				       pi->sram_end);
+}
+
+u32 cypress_calculate_burst_time(struct radeon_device *rdev,
+				 u32 engine_clock, u32 memory_clock)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 multiplier = pi->mem_gddr5 ? 1 : 2;
+	u32 result = (4 * multiplier * engine_clock) / (memory_clock / 2);
+	u32 burst_time;
+
+	if (result <= 4)
+		burst_time = 0;
+	else if (result < 8)
+		burst_time = result - 4;
+	else {
+		burst_time = result / 2 ;
+		if (burst_time > 18)
+			burst_time = 18;
+	}
+
+	return burst_time;
+}
+
+void cypress_program_memory_timing_parameters(struct radeon_device *rdev,
+					      struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_ps *new_state = rv770_get_ps(radeon_new_state);
+	u32 mc_arb_burst_time = RREG32(MC_ARB_BURST_TIME);
+
+	mc_arb_burst_time &= ~(STATE1_MASK | STATE2_MASK | STATE3_MASK);
+
+	mc_arb_burst_time |= STATE1(cypress_calculate_burst_time(rdev,
+								 new_state->low.sclk,
+								 new_state->low.mclk));
+	mc_arb_burst_time |= STATE2(cypress_calculate_burst_time(rdev,
+								 new_state->medium.sclk,
+								 new_state->medium.mclk));
+	mc_arb_burst_time |= STATE3(cypress_calculate_burst_time(rdev,
+								 new_state->high.sclk,
+								 new_state->high.mclk));
+
+	rv730_program_memory_timing_parameters(rdev, radeon_new_state);
+
+	WREG32(MC_ARB_BURST_TIME, mc_arb_burst_time);
+}
+
+static void cypress_populate_mc_reg_addresses(struct radeon_device *rdev,
+					      SMC_Evergreen_MCRegisters *mc_reg_table)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 i, j;
+
+	for (i = 0, j = 0; j < eg_pi->mc_reg_table.last; j++) {
+		if (eg_pi->mc_reg_table.valid_flag & (1 << j)) {
+			mc_reg_table->address[i].s0 =
+				cpu_to_be16(eg_pi->mc_reg_table.mc_reg_address[j].s0);
+			mc_reg_table->address[i].s1 =
+				cpu_to_be16(eg_pi->mc_reg_table.mc_reg_address[j].s1);
+			i++;
+		}
+	}
+
+	mc_reg_table->last = (u8)i;
+}
+
+static void cypress_set_mc_reg_address_table(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 i = 0;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_RAS_TIMING_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_RAS_TIMING >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_CAS_TIMING_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_CAS_TIMING >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_MISC_TIMING_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_MISC_TIMING >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_MISC_TIMING2_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_MISC_TIMING2 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_RD_CTL_D0_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_RD_CTL_D0 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_RD_CTL_D1_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_RD_CTL_D1 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_WR_CTL_D0_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_WR_CTL_D0 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_WR_CTL_D1_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_WR_CTL_D1 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_PMG_CMD_EMRS >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_PMG_CMD_MRS >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_PMG_CMD_MRS1 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_MISC1 >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_MISC1 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_RESERVE_M >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_RESERVE_M >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.mc_reg_address[i].s0 = MC_SEQ_MISC3 >> 2;
+	eg_pi->mc_reg_table.mc_reg_address[i].s1 = MC_SEQ_MISC3 >> 2;
+	i++;
+
+	eg_pi->mc_reg_table.last = (u8)i;
+}
+
+static void cypress_retrieve_ac_timing_for_one_entry(struct radeon_device *rdev,
+						     struct evergreen_mc_reg_entry *entry)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 i;
+
+	for (i = 0; i < eg_pi->mc_reg_table.last; i++)
+		entry->mc_data[i] =
+			RREG32(eg_pi->mc_reg_table.mc_reg_address[i].s1 << 2);
+
+}
+
+static void cypress_retrieve_ac_timing_for_all_ranges(struct radeon_device *rdev,
+						      struct atom_memory_clock_range_table *range_table)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 i, j;
+
+	for (i = 0; i < range_table->num_entries; i++) {
+		eg_pi->mc_reg_table.mc_reg_table_entry[i].mclk_max =
+			range_table->mclk[i];
+		radeon_atom_set_ac_timing(rdev, range_table->mclk[i]);
+		cypress_retrieve_ac_timing_for_one_entry(rdev,
+							 &eg_pi->mc_reg_table.mc_reg_table_entry[i]);
+	}
+
+	eg_pi->mc_reg_table.num_entries = range_table->num_entries;
+	eg_pi->mc_reg_table.valid_flag = 0;
+
+	for (i = 0; i < eg_pi->mc_reg_table.last; i++) {
+		for (j = 1; j < range_table->num_entries; j++) {
+			if (eg_pi->mc_reg_table.mc_reg_table_entry[j-1].mc_data[i] !=
+			    eg_pi->mc_reg_table.mc_reg_table_entry[j].mc_data[i]) {
+				eg_pi->mc_reg_table.valid_flag |= (1 << i);
+				break;
+			}
+		}
+	}
+}
+
+static int cypress_initialize_mc_reg_table(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 module_index = rv770_get_memory_module_index(rdev);
+	struct atom_memory_clock_range_table range_table = { 0 };
+	int ret;
+
+	ret = radeon_atom_get_mclk_range_table(rdev,
+					       pi->mem_gddr5,
+					       module_index, &range_table);
+	if (ret)
+		return ret;
+
+	cypress_retrieve_ac_timing_for_all_ranges(rdev, &range_table);
+
+	return 0;
+}
+
+static void cypress_wait_for_mc_sequencer(struct radeon_device *rdev, u8 value)
+{
+	u32 i, j;
+	u32 channels = 2;
+
+	if ((rdev->family == CHIP_CYPRESS) ||
+	    (rdev->family == CHIP_HEMLOCK))
+		channels = 4;
+	else if (rdev->family == CHIP_CEDAR)
+		channels = 1;
+
+	for (i = 0; i < channels; i++) {
+		if ((rdev->family == CHIP_CYPRESS) ||
+		    (rdev->family == CHIP_HEMLOCK)) {
+			WREG32_P(MC_CONFIG_MCD, MC_RD_ENABLE_MCD(i), ~MC_RD_ENABLE_MCD_MASK);
+			WREG32_P(MC_CG_CONFIG_MCD, MC_RD_ENABLE_MCD(i), ~MC_RD_ENABLE_MCD_MASK);
+		} else {
+			WREG32_P(MC_CONFIG, MC_RD_ENABLE(i), ~MC_RD_ENABLE_MASK);
+			WREG32_P(MC_CG_CONFIG, MC_RD_ENABLE(i), ~MC_RD_ENABLE_MASK);
+		}
+		for (j = 0; j < rdev->usec_timeout; j++) {
+			if (((RREG32(MC_SEQ_CG) & CG_SEQ_RESP_MASK) >> CG_SEQ_RESP_SHIFT) == value)
+				break;
+			udelay(1);
+		}
+	}
+}
+
+static void cypress_force_mc_use_s1(struct radeon_device *rdev,
+				    struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_ps *boot_state = rv770_get_ps(radeon_boot_state);
+	u32 strobe_mode;
+	u32 mc_seq_cg;
+	int i;
+
+	if (RREG32(MC_SEQ_STATUS_M) & PMG_PWRSTATE)
+		return;
+
+	radeon_atom_set_ac_timing(rdev, boot_state->low.mclk);
+	radeon_mc_wait_for_idle(rdev);
+
+	if ((rdev->family == CHIP_CYPRESS) ||
+	    (rdev->family == CHIP_HEMLOCK)) {
+		WREG32(MC_CONFIG_MCD, 0xf);
+		WREG32(MC_CG_CONFIG_MCD, 0xf);
+	} else {
+		WREG32(MC_CONFIG, 0xf);
+		WREG32(MC_CG_CONFIG, 0xf);
+	}
+
+	for (i = 0; i < rdev->num_crtc; i++)
+		radeon_wait_for_vblank(rdev, i);
+
+	WREG32(MC_SEQ_CG, MC_CG_SEQ_YCLK_SUSPEND);
+	cypress_wait_for_mc_sequencer(rdev, MC_CG_SEQ_YCLK_SUSPEND);
+
+	strobe_mode = cypress_get_strobe_mode_settings(rdev,
+						       boot_state->low.mclk);
+
+	mc_seq_cg = CG_SEQ_REQ(MC_CG_SEQ_DRAMCONF_S1);
+	mc_seq_cg |= SEQ_CG_RESP(strobe_mode);
+	WREG32(MC_SEQ_CG, mc_seq_cg);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(MC_SEQ_STATUS_M) & PMG_PWRSTATE)
+			break;
+		udelay(1);
+	}
+
+	mc_seq_cg &= ~CG_SEQ_REQ_MASK;
+	mc_seq_cg |= CG_SEQ_REQ(MC_CG_SEQ_YCLK_RESUME);
+	WREG32(MC_SEQ_CG, mc_seq_cg);
+
+	cypress_wait_for_mc_sequencer(rdev, MC_CG_SEQ_YCLK_RESUME);
+}
+
+static void cypress_copy_ac_timing_from_s1_to_s0(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 value;
+	u32 i;
+
+	for (i = 0; i < eg_pi->mc_reg_table.last; i++) {
+		value = RREG32(eg_pi->mc_reg_table.mc_reg_address[i].s1 << 2);
+		WREG32(eg_pi->mc_reg_table.mc_reg_address[i].s0 << 2, value);
+	}
+}
+
+static void cypress_force_mc_use_s0(struct radeon_device *rdev,
+				    struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_ps *boot_state = rv770_get_ps(radeon_boot_state);
+	u32 strobe_mode;
+	u32 mc_seq_cg;
+	int i;
+
+	cypress_copy_ac_timing_from_s1_to_s0(rdev);
+	radeon_mc_wait_for_idle(rdev);
+
+	if ((rdev->family == CHIP_CYPRESS) ||
+	    (rdev->family == CHIP_HEMLOCK)) {
+		WREG32(MC_CONFIG_MCD, 0xf);
+		WREG32(MC_CG_CONFIG_MCD, 0xf);
+	} else {
+		WREG32(MC_CONFIG, 0xf);
+		WREG32(MC_CG_CONFIG, 0xf);
+	}
+
+	for (i = 0; i < rdev->num_crtc; i++)
+		radeon_wait_for_vblank(rdev, i);
+
+	WREG32(MC_SEQ_CG, MC_CG_SEQ_YCLK_SUSPEND);
+	cypress_wait_for_mc_sequencer(rdev, MC_CG_SEQ_YCLK_SUSPEND);
+
+	strobe_mode = cypress_get_strobe_mode_settings(rdev,
+						       boot_state->low.mclk);
+
+	mc_seq_cg = CG_SEQ_REQ(MC_CG_SEQ_DRAMCONF_S0);
+	mc_seq_cg |= SEQ_CG_RESP(strobe_mode);
+	WREG32(MC_SEQ_CG, mc_seq_cg);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (!(RREG32(MC_SEQ_STATUS_M) & PMG_PWRSTATE))
+			break;
+		udelay(1);
+	}
+
+	mc_seq_cg &= ~CG_SEQ_REQ_MASK;
+	mc_seq_cg |= CG_SEQ_REQ(MC_CG_SEQ_YCLK_RESUME);
+	WREG32(MC_SEQ_CG, mc_seq_cg);
+
+	cypress_wait_for_mc_sequencer(rdev, MC_CG_SEQ_YCLK_RESUME);
+}
+
+static int cypress_populate_initial_mvdd_value(struct radeon_device *rdev,
+					       RV770_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	voltage->index = eg_pi->mvdd_high_index;
+	voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+
+	return 0;
+}
+
+int cypress_populate_smc_initial_state(struct radeon_device *rdev,
+				       struct radeon_ps *radeon_initial_state,
+				       RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_ps *initial_state = rv770_get_ps(radeon_initial_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 a_t;
+
+	table->initialState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ad_func_cntl);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ad_func_cntl_2);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_dq_func_cntl);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_dq_func_cntl_2);
+	table->initialState.levels[0].mclk.mclk770.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.mclk_pwrmgt_cntl);
+	table->initialState.levels[0].mclk.mclk770.vDLL_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.dll_cntl);
+
+	table->initialState.levels[0].mclk.mclk770.vMPLL_SS =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ss1);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_SS2 =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ss2);
+
+	table->initialState.levels[0].mclk.mclk770.mclk_value =
+		cpu_to_be32(initial_state->low.mclk);
+
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_func_cntl);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_func_cntl_2);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_func_cntl_3);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_spread_spectrum);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM_2 =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_spread_spectrum_2);
+
+	table->initialState.levels[0].sclk.sclk_value =
+		cpu_to_be32(initial_state->low.sclk);
+
+	table->initialState.levels[0].arbValue = MC_CG_ARB_FREQ_F0;
+
+	table->initialState.levels[0].ACIndex = 0;
+
+	cypress_populate_voltage_value(rdev,
+				       &eg_pi->vddc_voltage_table,
+				       initial_state->low.vddc,
+				       &table->initialState.levels[0].vddc);
+
+	if (eg_pi->vddci_control)
+		cypress_populate_voltage_value(rdev,
+					       &eg_pi->vddci_voltage_table,
+					       initial_state->low.vddci,
+					       &table->initialState.levels[0].vddci);
+
+	cypress_populate_initial_mvdd_value(rdev,
+					    &table->initialState.levels[0].mvdd);
+
+	a_t = CG_R(0xffff) | CG_L(0);
+	table->initialState.levels[0].aT = cpu_to_be32(a_t);
+
+	table->initialState.levels[0].bSP = cpu_to_be32(pi->dsp);
+
+
+	if (pi->boot_in_gen2)
+		table->initialState.levels[0].gen2PCIE = 1;
+	else
+		table->initialState.levels[0].gen2PCIE = 0;
+	if (initial_state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2)
+		table->initialState.levels[0].gen2XSP = 1;
+	else
+		table->initialState.levels[0].gen2XSP = 0;
+
+	if (pi->mem_gddr5) {
+		table->initialState.levels[0].strobeMode =
+			cypress_get_strobe_mode_settings(rdev,
+							 initial_state->low.mclk);
+
+		if (initial_state->low.mclk > pi->mclk_edc_enable_threshold)
+			table->initialState.levels[0].mcFlags = SMC_MC_EDC_RD_FLAG | SMC_MC_EDC_WR_FLAG;
+		else
+			table->initialState.levels[0].mcFlags =  0;
+	}
+
+	table->initialState.levels[1] = table->initialState.levels[0];
+	table->initialState.levels[2] = table->initialState.levels[0];
+
+	table->initialState.flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	return 0;
+}
+
+int cypress_populate_smc_acpi_state(struct radeon_device *rdev,
+				    RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 mpll_ad_func_cntl =
+		pi->clk_regs.rv770.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl =
+		pi->clk_regs.rv770.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_dq_func_cntl_2;
+	u32 spll_func_cntl =
+		pi->clk_regs.rv770.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 =
+		pi->clk_regs.rv770.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 =
+		pi->clk_regs.rv770.cg_spll_func_cntl_3;
+	u32 mclk_pwrmgt_cntl =
+		pi->clk_regs.rv770.mclk_pwrmgt_cntl;
+	u32 dll_cntl =
+		pi->clk_regs.rv770.dll_cntl;
+
+	table->ACPIState = table->initialState;
+
+	table->ACPIState.flags &= ~PPSMC_SWSTATE_FLAG_DC;
+
+	if (pi->acpi_vddc) {
+		cypress_populate_voltage_value(rdev,
+					       &eg_pi->vddc_voltage_table,
+					       pi->acpi_vddc,
+					       &table->ACPIState.levels[0].vddc);
+		if (pi->pcie_gen2) {
+			if (pi->acpi_pcie_gen2)
+				table->ACPIState.levels[0].gen2PCIE = 1;
+			else
+				table->ACPIState.levels[0].gen2PCIE = 0;
+		} else
+			table->ACPIState.levels[0].gen2PCIE = 0;
+		if (pi->acpi_pcie_gen2)
+			table->ACPIState.levels[0].gen2XSP = 1;
+		else
+			table->ACPIState.levels[0].gen2XSP = 0;
+	} else {
+		cypress_populate_voltage_value(rdev,
+					       &eg_pi->vddc_voltage_table,
+					       pi->min_vddc_in_table,
+					       &table->ACPIState.levels[0].vddc);
+		table->ACPIState.levels[0].gen2PCIE = 0;
+	}
+
+	if (eg_pi->acpi_vddci) {
+		if (eg_pi->vddci_control) {
+			cypress_populate_voltage_value(rdev,
+						       &eg_pi->vddci_voltage_table,
+						       eg_pi->acpi_vddci,
+						       &table->ACPIState.levels[0].vddci);
+		}
+	}
+
+	mpll_ad_func_cntl &= ~PDNB;
+
+	mpll_ad_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN;
+
+	if (pi->mem_gddr5)
+		mpll_dq_func_cntl &= ~PDNB;
+	mpll_dq_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN | BYPASS;
+
+	mclk_pwrmgt_cntl |= (MRDCKA0_RESET |
+			     MRDCKA1_RESET |
+			     MRDCKB0_RESET |
+			     MRDCKB1_RESET |
+			     MRDCKC0_RESET |
+			     MRDCKC1_RESET |
+			     MRDCKD0_RESET |
+			     MRDCKD1_RESET);
+
+	mclk_pwrmgt_cntl &= ~(MRDCKA0_PDNB |
+			      MRDCKA1_PDNB |
+			      MRDCKB0_PDNB |
+			      MRDCKB1_PDNB |
+			      MRDCKC0_PDNB |
+			      MRDCKC1_PDNB |
+			      MRDCKD0_PDNB |
+			      MRDCKD1_PDNB);
+
+	dll_cntl |= (MRDCKA0_BYPASS |
+		     MRDCKA1_BYPASS |
+		     MRDCKB0_BYPASS |
+		     MRDCKB1_BYPASS |
+		     MRDCKC0_BYPASS |
+		     MRDCKC1_BYPASS |
+		     MRDCKD0_BYPASS |
+		     MRDCKD1_BYPASS);
+
+	/* evergreen only */
+	if (rdev->family <= CHIP_HEMLOCK)
+		spll_func_cntl |= SPLL_RESET | SPLL_SLEEP | SPLL_BYPASS_EN;
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(4);
+
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL =
+		cpu_to_be32(mpll_ad_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL_2 =
+		cpu_to_be32(mpll_ad_func_cntl_2);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL =
+		cpu_to_be32(mpll_dq_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL_2 =
+		cpu_to_be32(mpll_dq_func_cntl_2);
+	table->ACPIState.levels[0].mclk.mclk770.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(mclk_pwrmgt_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vDLL_CNTL = cpu_to_be32(dll_cntl);
+
+	table->ACPIState.levels[0].mclk.mclk770.mclk_value = 0;
+
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(spll_func_cntl);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(spll_func_cntl_2);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(spll_func_cntl_3);
+
+	table->ACPIState.levels[0].sclk.sclk_value = 0;
+
+	cypress_populate_mvdd_value(rdev, 0, &table->ACPIState.levels[0].mvdd);
+
+	if (eg_pi->dynamic_ac_timing)
+		table->ACPIState.levels[0].ACIndex = 1;
+
+	table->ACPIState.levels[1] = table->ACPIState.levels[0];
+	table->ACPIState.levels[2] = table->ACPIState.levels[0];
+
+	return 0;
+}
+
+static void cypress_trim_voltage_table_to_fit_state_table(struct radeon_device *rdev,
+							  struct atom_voltage_table *voltage_table)
+{
+	unsigned int i, diff;
+
+	if (voltage_table->count <= MAX_NO_VREG_STEPS)
+		return;
+
+	diff = voltage_table->count - MAX_NO_VREG_STEPS;
+
+	for (i= 0; i < MAX_NO_VREG_STEPS; i++)
+		voltage_table->entries[i] = voltage_table->entries[i + diff];
+
+	voltage_table->count = MAX_NO_VREG_STEPS;
+}
+
+int cypress_construct_voltage_tables(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	int ret;
+
+	ret = radeon_atom_get_voltage_table(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, 0,
+					    &eg_pi->vddc_voltage_table);
+	if (ret)
+		return ret;
+
+	if (eg_pi->vddc_voltage_table.count > MAX_NO_VREG_STEPS)
+		cypress_trim_voltage_table_to_fit_state_table(rdev,
+							      &eg_pi->vddc_voltage_table);
+
+	if (eg_pi->vddci_control) {
+		ret = radeon_atom_get_voltage_table(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0,
+						    &eg_pi->vddci_voltage_table);
+		if (ret)
+			return ret;
+
+		if (eg_pi->vddci_voltage_table.count > MAX_NO_VREG_STEPS)
+			cypress_trim_voltage_table_to_fit_state_table(rdev,
+								      &eg_pi->vddci_voltage_table);
+	}
+
+	return 0;
+}
+
+static void cypress_populate_smc_voltage_table(struct radeon_device *rdev,
+					       struct atom_voltage_table *voltage_table,
+					       RV770_SMC_STATETABLE *table)
+{
+	unsigned int i;
+
+	for (i = 0; i < voltage_table->count; i++) {
+		table->highSMIO[i] = 0;
+		table->lowSMIO[i] |= cpu_to_be32(voltage_table->entries[i].smio_low);
+	}
+}
+
+int cypress_populate_smc_voltage_tables(struct radeon_device *rdev,
+					RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	unsigned char i;
+
+	if (eg_pi->vddc_voltage_table.count) {
+		cypress_populate_smc_voltage_table(rdev,
+						   &eg_pi->vddc_voltage_table,
+						   table);
+
+		table->voltageMaskTable.highMask[RV770_SMC_VOLTAGEMASK_VDDC] = 0;
+		table->voltageMaskTable.lowMask[RV770_SMC_VOLTAGEMASK_VDDC] =
+			cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
+
+		for (i = 0; i < eg_pi->vddc_voltage_table.count; i++) {
+			if (pi->max_vddc_in_table <=
+			    eg_pi->vddc_voltage_table.entries[i].value) {
+				table->maxVDDCIndexInPPTable = i;
+				break;
+			}
+		}
+	}
+
+	if (eg_pi->vddci_voltage_table.count) {
+		cypress_populate_smc_voltage_table(rdev,
+						   &eg_pi->vddci_voltage_table,
+						   table);
+
+		table->voltageMaskTable.highMask[RV770_SMC_VOLTAGEMASK_VDDCI] = 0;
+		table->voltageMaskTable.lowMask[RV770_SMC_VOLTAGEMASK_VDDCI] =
+			cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
+	}
+
+	return 0;
+}
+
+static u32 cypress_get_mclk_split_point(struct atom_memory_info *memory_info)
+{
+	if ((memory_info->mem_type == MEM_TYPE_GDDR3) ||
+	    (memory_info->mem_type == MEM_TYPE_DDR3))
+		return 30000;
+
+	return 0;
+}
+
+int cypress_get_mvdd_configuration(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u8 module_index;
+	struct atom_memory_info memory_info;
+	u32 tmp = RREG32(GENERAL_PWRMGT);
+
+	if (!(tmp & BACKBIAS_PAD_EN)) {
+		eg_pi->mvdd_high_index = 0;
+		eg_pi->mvdd_low_index = 1;
+		pi->mvdd_control = false;
+		return 0;
+	}
+
+	if (tmp & BACKBIAS_VALUE)
+		eg_pi->mvdd_high_index = 1;
+	else
+		eg_pi->mvdd_high_index = 0;
+
+	eg_pi->mvdd_low_index =
+		(eg_pi->mvdd_high_index == 0) ? 1 : 0;
+
+	module_index = rv770_get_memory_module_index(rdev);
+
+	if (radeon_atom_get_memory_info(rdev, module_index, &memory_info)) {
+		pi->mvdd_control = false;
+		return 0;
+	}
+
+	pi->mvdd_split_frequency =
+		cypress_get_mclk_split_point(&memory_info);
+
+	if (pi->mvdd_split_frequency == 0) {
+		pi->mvdd_control = false;
+		return 0;
+	}
+
+	return 0;
+}
+
+static int cypress_init_smc_table(struct radeon_device *rdev,
+				  struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	RV770_SMC_STATETABLE *table = &pi->smc_statetable;
+	int ret;
+
+	memset(table, 0, sizeof(RV770_SMC_STATETABLE));
+
+	cypress_populate_smc_voltage_tables(rdev, table);
+
+	switch (rdev->pm.int_thermal_type) {
+        case THERMAL_TYPE_EVERGREEN:
+        case THERMAL_TYPE_EMC2103_WITH_INTERNAL:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_INTERNAL;
+		break;
+        case THERMAL_TYPE_NONE:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_NONE;
+		break;
+        default:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_EXTERNAL;
+		break;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_HARDWAREDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GPIO_DC;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REGULATOR_HOT)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_REGULATOR_HOT;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_STEPVDDC;
+
+	if (pi->mem_gddr5)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GDDR5;
+
+	ret = cypress_populate_smc_initial_state(rdev, radeon_boot_state, table);
+	if (ret)
+		return ret;
+
+	ret = cypress_populate_smc_acpi_state(rdev, table);
+	if (ret)
+		return ret;
+
+	table->driverState = table->initialState;
+
+	return rv770_copy_bytes_to_smc(rdev,
+				       pi->state_table_start,
+				       (u8 *)table, sizeof(RV770_SMC_STATETABLE),
+				       pi->sram_end);
+}
+
+int cypress_populate_mc_reg_table(struct radeon_device *rdev,
+				  struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_ps *boot_state = rv770_get_ps(radeon_boot_state);
+	SMC_Evergreen_MCRegisters mc_reg_table = { 0 };
+
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_seq_index, 1);
+
+	cypress_populate_mc_reg_addresses(rdev, &mc_reg_table);
+
+	cypress_convert_mc_reg_table_entry_to_smc(rdev,
+						  &boot_state->low,
+						  &mc_reg_table.data[0]);
+
+	cypress_convert_mc_registers(&eg_pi->mc_reg_table.mc_reg_table_entry[0],
+				     &mc_reg_table.data[1], eg_pi->mc_reg_table.last,
+				     eg_pi->mc_reg_table.valid_flag);
+
+	cypress_convert_mc_reg_table_to_smc(rdev, radeon_boot_state, &mc_reg_table);
+
+	return rv770_copy_bytes_to_smc(rdev, eg_pi->mc_reg_table_start,
+				       (u8 *)&mc_reg_table, sizeof(SMC_Evergreen_MCRegisters),
+				       pi->sram_end);
+}
+
+int cypress_get_table_locations(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					EVERGREEN_SMC_FIRMWARE_HEADER_LOCATION +
+					EVERGREEN_SMC_FIRMWARE_HEADER_stateTable,
+					&tmp, pi->sram_end);
+	if (ret)
+		return ret;
+
+	pi->state_table_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					EVERGREEN_SMC_FIRMWARE_HEADER_LOCATION +
+					EVERGREEN_SMC_FIRMWARE_HEADER_softRegisters,
+					&tmp, pi->sram_end);
+	if (ret)
+		return ret;
+
+	pi->soft_regs_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					EVERGREEN_SMC_FIRMWARE_HEADER_LOCATION +
+					EVERGREEN_SMC_FIRMWARE_HEADER_mcRegisterTable,
+					&tmp, pi->sram_end);
+	if (ret)
+		return ret;
+
+	eg_pi->mc_reg_table_start = (u16)tmp;
+
+	return 0;
+}
+
+void cypress_enable_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL);
+
+	tmp &= ~(DISP1_GAP_MASK | DISP2_GAP_MASK);
+	tmp |= (DISP1_GAP(R600_PM_DISPLAY_GAP_IGNORE) |
+		DISP2_GAP(R600_PM_DISPLAY_GAP_IGNORE));
+
+	tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK);
+	tmp |= (DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK) |
+		DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE));
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+}
+
+static void cypress_program_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp, pipe;
+	int i;
+
+	tmp = RREG32(CG_DISPLAY_GAP_CNTL) & ~(DISP1_GAP_MASK | DISP2_GAP_MASK);
+	if (rdev->pm.dpm.new_active_crtc_count > 0)
+		tmp |= DISP1_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM);
+	else
+		tmp |= DISP1_GAP(R600_PM_DISPLAY_GAP_IGNORE);
+
+	if (rdev->pm.dpm.new_active_crtc_count > 1)
+		tmp |= DISP2_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM);
+	else
+		tmp |= DISP2_GAP(R600_PM_DISPLAY_GAP_IGNORE);
+
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+
+	tmp = RREG32(DCCG_DISP_SLOW_SELECT_REG);
+	pipe = (tmp & DCCG_DISP1_SLOW_SELECT_MASK) >> DCCG_DISP1_SLOW_SELECT_SHIFT;
+
+	if ((rdev->pm.dpm.new_active_crtc_count > 0) &&
+	    (!(rdev->pm.dpm.new_active_crtcs & (1 << pipe)))) {
+		/* find the first active crtc */
+		for (i = 0; i < rdev->num_crtc; i++) {
+			if (rdev->pm.dpm.new_active_crtcs & (1 << i))
+				break;
+		}
+		if (i == rdev->num_crtc)
+			pipe = 0;
+		else
+			pipe = i;
+
+		tmp &= ~DCCG_DISP1_SLOW_SELECT_MASK;
+		tmp |= DCCG_DISP1_SLOW_SELECT(pipe);
+		WREG32(DCCG_DISP_SLOW_SELECT_REG, tmp);
+	}
+
+	cypress_notify_smc_display_change(rdev, rdev->pm.dpm.new_active_crtc_count > 0);
+}
+
+void cypress_dpm_setup_asic(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	rv740_read_clock_registers(rdev);
+	rv770_read_voltage_smio_registers(rdev);
+	rv770_get_max_vddc(rdev);
+	rv770_get_memory_type(rdev);
+
+	if (eg_pi->pcie_performance_request)
+		eg_pi->pcie_performance_request_registered = false;
+
+	if (eg_pi->pcie_performance_request)
+		cypress_advertise_gen2_capability(rdev);
+
+	rv770_get_pcie_gen2_status(rdev);
+
+	rv770_enable_acpi_pm(rdev);
+}
+
+int cypress_dpm_enable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	if (pi->gfx_clock_gating)
+		rv770_restore_cgcg(rdev);
+
+	if (rv770_dpm_enabled(rdev))
+		return -EINVAL;
+
+	if (pi->voltage_control) {
+		rv770_enable_voltage_control(rdev, true);
+		ret = cypress_construct_voltage_tables(rdev);
+		if (ret) {
+			DRM_ERROR("cypress_construct_voltage_tables failed\n");
+			return ret;
+		}
+	}
+
+	if (pi->mvdd_control) {
+		ret = cypress_get_mvdd_configuration(rdev);
+		if (ret) {
+			DRM_ERROR("cypress_get_mvdd_configuration failed\n");
+			return ret;
+		}
+	}
+
+	if (eg_pi->dynamic_ac_timing) {
+		cypress_set_mc_reg_address_table(rdev);
+		cypress_force_mc_use_s0(rdev, boot_ps);
+		ret = cypress_initialize_mc_reg_table(rdev);
+		if (ret)
+			eg_pi->dynamic_ac_timing = false;
+		cypress_force_mc_use_s1(rdev, boot_ps);
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv770_enable_backbias(rdev, true);
+
+	if (pi->dynamic_ss)
+		cypress_enable_spread_spectrum(rdev, true);
+
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, true);
+
+	rv770_setup_bsp(rdev);
+	rv770_program_git(rdev);
+	rv770_program_tp(rdev);
+	rv770_program_tpp(rdev);
+	rv770_program_sstp(rdev);
+	rv770_program_engine_speed_parameters(rdev);
+	cypress_enable_display_gap(rdev);
+	rv770_program_vc(rdev);
+
+	if (pi->dynamic_pcie_gen2)
+		cypress_enable_dynamic_pcie_gen2(rdev, true);
+
+	ret = rv770_upload_firmware(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_upload_firmware failed\n");
+		return ret;
+	}
+
+	ret = cypress_get_table_locations(rdev);
+	if (ret) {
+		DRM_ERROR("cypress_get_table_locations failed\n");
+		return ret;
+	}
+	ret = cypress_init_smc_table(rdev, boot_ps);
+	if (ret) {
+		DRM_ERROR("cypress_init_smc_table failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = cypress_populate_mc_reg_table(rdev, boot_ps);
+		if (ret) {
+			DRM_ERROR("cypress_populate_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+
+	cypress_program_response_times(rdev);
+
+	r7xx_start_smc(rdev);
+
+	ret = cypress_notify_smc_display_change(rdev, false);
+	if (ret) {
+		DRM_ERROR("cypress_notify_smc_display_change failed\n");
+		return ret;
+	}
+	cypress_enable_sclk_control(rdev, true);
+
+	if (eg_pi->memory_transition)
+		cypress_enable_mclk_control(rdev, true);
+
+	cypress_start_dpm(rdev);
+
+	if (pi->gfx_clock_gating)
+		cypress_gfx_clock_gating_enable(rdev, true);
+
+	if (pi->mg_clock_gating)
+		cypress_mg_clock_gating_enable(rdev, true);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		PPSMC_Result result;
+
+		ret = rv770_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+		result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+
+		if (result != PPSMC_Result_OK)
+			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
+	}
+
+	rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
+
+	return 0;
+}
+
+void cypress_dpm_disable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+
+	if (!rv770_dpm_enabled(rdev))
+		return;
+
+	rv770_clear_vc(rdev);
+
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, false);
+
+	if (pi->dynamic_pcie_gen2)
+		cypress_enable_dynamic_pcie_gen2(rdev, false);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	if (pi->gfx_clock_gating)
+		cypress_gfx_clock_gating_enable(rdev, false);
+
+	if (pi->mg_clock_gating)
+		cypress_mg_clock_gating_enable(rdev, false);
+
+	rv770_stop_dpm(rdev);
+	r7xx_stop_smc(rdev);
+
+	cypress_enable_spread_spectrum(rdev, false);
+
+	if (eg_pi->dynamic_ac_timing)
+		cypress_force_mc_use_s1(rdev, boot_ps);
+
+	rv770_reset_smio_status(rdev);
+}
+
+int cypress_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = rdev->pm.dpm.requested_ps;
+	struct radeon_ps *old_ps = rdev->pm.dpm.current_ps;
+	int ret;
+
+	ret = rv770_restrict_performance_levels_before_switch(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_restrict_performance_levels_before_switch failed\n");
+		return ret;
+	}
+	if (eg_pi->pcie_performance_request)
+		cypress_notify_link_speed_change_before_state_change(rdev, new_ps, old_ps);
+
+	rv770_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+	ret = rv770_halt_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_halt_smc failed\n");
+		return ret;
+	}
+	ret = cypress_upload_sw_state(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("cypress_upload_sw_state failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = cypress_upload_mc_reg_table(rdev, new_ps);
+		if (ret) {
+			DRM_ERROR("cypress_upload_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+
+	cypress_program_memory_timing_parameters(rdev, new_ps);
+
+	ret = rv770_resume_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_resume_smc failed\n");
+		return ret;
+	}
+	ret = rv770_set_sw_state(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_set_sw_state failed\n");
+		return ret;
+	}
+	rv770_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+
+	if (eg_pi->pcie_performance_request)
+		cypress_notify_link_speed_change_after_state_change(rdev, new_ps, old_ps);
+
+	ret = rv770_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO);
+	if (ret) {
+		DRM_ERROR("rv770_dpm_force_performance_level failed\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+void cypress_dpm_reset_asic(struct radeon_device *rdev)
+{
+	rv770_restrict_performance_levels_before_switch(rdev);
+	rv770_set_boot_state(rdev);
+}
+
+void cypress_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+	cypress_program_display_gap(rdev);
+}
+
+int cypress_dpm_init(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi;
+	struct evergreen_power_info *eg_pi;
+	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	uint16_t data_offset, size;
+	uint8_t frev, crev;
+	struct atom_clock_dividers dividers;
+	int ret;
+
+	eg_pi = kzalloc(sizeof(struct evergreen_power_info), GFP_KERNEL);
+	if (eg_pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = eg_pi;
+	pi = &eg_pi->rv7xx;
+
+	rv770_get_max_vddc(rdev);
+
+	eg_pi->ulv.supported = false;
+	pi->acpi_vddc = 0;
+	eg_pi->acpi_vddci = 0;
+	pi->min_vddc_in_table = 0;
+	pi->max_vddc_in_table = 0;
+
+	ret = rv7xx_parse_power_table(rdev);
+	if (ret)
+		return ret;
+
+	if (rdev->pm.dpm.voltage_response_time == 0)
+		rdev->pm.dpm.voltage_response_time = R600_VOLTAGERESPONSETIME_DFLT;
+	if (rdev->pm.dpm.backbias_response_time == 0)
+		rdev->pm.dpm.backbias_response_time = R600_BACKBIASRESPONSETIME_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->ref_div = dividers.ref_div + 1;
+	else
+		pi->ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	pi->mclk_strobe_mode_threshold = 40000;
+	pi->mclk_edc_enable_threshold = 40000;
+	eg_pi->mclk_edc_wr_enable_threshold = 40000;
+
+	pi->rlp = RV770_RLP_DFLT;
+	pi->rmp = RV770_RMP_DFLT;
+	pi->lhp = RV770_LHP_DFLT;
+	pi->lmp = RV770_LMP_DFLT;
+
+	pi->voltage_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, 0);
+
+	pi->mvdd_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, 0);
+
+	eg_pi->vddci_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0);
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+                                   &frev, &crev, &data_offset)) {
+		pi->sclk_ss = true;
+		pi->mclk_ss = true;
+		pi->dynamic_ss = true;
+	} else {
+		pi->sclk_ss = false;
+		pi->mclk_ss = false;
+		pi->dynamic_ss = true;
+	}
+
+	pi->asi = RV770_ASI_DFLT;
+	pi->pasi = CYPRESS_HASI_DFLT;
+	pi->vrc = CYPRESS_VRC_DFLT;
+
+	pi->power_gating = false;
+
+	if ((rdev->family == CHIP_CYPRESS) ||
+	    (rdev->family == CHIP_HEMLOCK))
+		pi->gfx_clock_gating = false;
+	else
+		pi->gfx_clock_gating = true;
+
+	pi->mg_clock_gating = true;
+	pi->mgcgtssm = true;
+	eg_pi->ls_clock_gating = false;
+	eg_pi->sclk_deep_sleep = false;
+
+	pi->dynamic_pcie_gen2 = true;
+
+	if (pi->gfx_clock_gating &&
+	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+		pi->thermal_protection = true;
+	else
+		pi->thermal_protection = false;
+
+	pi->display_gap = true;
+
+	if (rdev->flags & RADEON_IS_MOBILITY)
+		pi->dcodt = true;
+	else
+		pi->dcodt = false;
+
+	pi->ulps = true;
+
+	eg_pi->dynamic_ac_timing = true;
+	eg_pi->abm = true;
+	eg_pi->mcls = true;
+	eg_pi->light_sleep = true;
+	eg_pi->memory_transition = true;
+#if defined(CONFIG_ACPI)
+	eg_pi->pcie_performance_request =
+		radeon_acpi_is_pcie_performance_request_supported(rdev);
+#else
+	eg_pi->pcie_performance_request = false;
+#endif
+
+	if ((rdev->family == CHIP_CYPRESS) ||
+	    (rdev->family == CHIP_HEMLOCK) ||
+	    (rdev->family == CHIP_JUNIPER))
+		eg_pi->dll_default_on = true;
+	else
+		eg_pi->dll_default_on = false;
+
+	eg_pi->sclk_deep_sleep = false;
+	pi->mclk_stutter_mode_threshold = 0;
+
+	pi->sram_end = SMC_RAM_END;
+
+	return 0;
+}
+
+void cypress_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+}
+
+bool cypress_dpm_vblank_too_short(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 vblank_time = r600_dpm_get_vblank_time(rdev);
+	u32 switch_limit = pi->mem_gddr5 ? 450 : 300;
+
+	if (vblank_time < switch_limit)
+		return true;
+	else
+		return false;
+
+}
diff --git a/drivers/gpu/drm/radeon/cypress_dpm.h b/drivers/gpu/drm/radeon/cypress_dpm.h
new file mode 100644
index 000000000000..4c3f18c69f4f
--- /dev/null
+++ b/drivers/gpu/drm/radeon/cypress_dpm.h
@@ -0,0 +1,160 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __CYPRESS_DPM_H__
+#define __CYPRESS_DPM_H__
+
+#include "rv770_dpm.h"
+#include "evergreen_smc.h"
+
+struct evergreen_mc_reg_entry {
+	u32 mclk_max;
+	u32 mc_data[SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE];
+};
+
+struct evergreen_mc_reg_table {
+	u8 last;
+	u8 num_entries;
+	u16 valid_flag;
+	struct evergreen_mc_reg_entry mc_reg_table_entry[MAX_AC_TIMING_ENTRIES];
+	SMC_Evergreen_MCRegisterAddress mc_reg_address[SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE];
+};
+
+struct evergreen_ulv_param {
+	bool supported;
+	struct rv7xx_pl *pl;
+};
+
+struct evergreen_arb_registers {
+	u32 mc_arb_dram_timing;
+	u32 mc_arb_dram_timing2;
+	u32 mc_arb_rfsh_rate;
+	u32 mc_arb_burst_time;
+};
+
+struct at {
+	u32 rlp;
+	u32 rmp;
+	u32 lhp;
+	u32 lmp;
+};
+
+struct evergreen_power_info {
+	/* must be first! */
+	struct rv7xx_power_info rv7xx;
+	/* flags */
+	bool vddci_control;
+	bool dynamic_ac_timing;
+	bool abm;
+	bool mcls;
+	bool light_sleep;
+	bool memory_transition;
+	bool pcie_performance_request;
+	bool pcie_performance_request_registered;
+	bool sclk_deep_sleep;
+	bool dll_default_on;
+	bool ls_clock_gating;
+	bool smu_uvd_hs;
+	bool uvd_enabled;
+	/* stored values */
+	u16 acpi_vddci;
+	u8 mvdd_high_index;
+	u8 mvdd_low_index;
+	u32 mclk_edc_wr_enable_threshold;
+	struct evergreen_mc_reg_table mc_reg_table;
+	struct atom_voltage_table vddc_voltage_table;
+	struct atom_voltage_table vddci_voltage_table;
+	struct evergreen_arb_registers bootup_arb_registers;
+	struct evergreen_ulv_param ulv;
+	struct at ats[2];
+	/* smc offsets */
+	u16 mc_reg_table_start;
+	struct radeon_ps current_rps;
+	struct rv7xx_ps current_ps;
+	struct radeon_ps requested_rps;
+	struct rv7xx_ps requested_ps;
+};
+
+#define CYPRESS_HASI_DFLT                               400000
+#define CYPRESS_MGCGTTLOCAL0_DFLT                       0x00000000
+#define CYPRESS_MGCGTTLOCAL1_DFLT                       0x00000000
+#define CYPRESS_MGCGTTLOCAL2_DFLT                       0x00000000
+#define CYPRESS_MGCGTTLOCAL3_DFLT                       0x00000000
+#define CYPRESS_MGCGCGTSSMCTRL_DFLT                     0x81944bc0
+#define REDWOOD_MGCGCGTSSMCTRL_DFLT                     0x6e944040
+#define CEDAR_MGCGCGTSSMCTRL_DFLT                       0x46944040
+#define CYPRESS_VRC_DFLT                                0xC00033
+
+#define PCIE_PERF_REQ_REMOVE_REGISTRY   0
+#define PCIE_PERF_REQ_FORCE_LOWPOWER    1
+#define PCIE_PERF_REQ_PECI_GEN1         2
+#define PCIE_PERF_REQ_PECI_GEN2         3
+#define PCIE_PERF_REQ_PECI_GEN3         4
+
+int cypress_convert_power_level_to_smc(struct radeon_device *rdev,
+				       struct rv7xx_pl *pl,
+				       RV770_SMC_HW_PERFORMANCE_LEVEL *level,
+				       u8 watermark_level);
+int cypress_populate_smc_acpi_state(struct radeon_device *rdev,
+				    RV770_SMC_STATETABLE *table);
+int cypress_populate_smc_voltage_tables(struct radeon_device *rdev,
+					RV770_SMC_STATETABLE *table);
+int cypress_populate_smc_initial_state(struct radeon_device *rdev,
+				       struct radeon_ps *radeon_initial_state,
+				       RV770_SMC_STATETABLE *table);
+u32 cypress_calculate_burst_time(struct radeon_device *rdev,
+				 u32 engine_clock, u32 memory_clock);
+void cypress_notify_link_speed_change_before_state_change(struct radeon_device *rdev,
+							  struct radeon_ps *radeon_new_state,
+							  struct radeon_ps *radeon_current_state);
+int cypress_upload_sw_state(struct radeon_device *rdev,
+			    struct radeon_ps *radeon_new_state);
+int cypress_upload_mc_reg_table(struct radeon_device *rdev,
+				struct radeon_ps *radeon_new_state);
+void cypress_program_memory_timing_parameters(struct radeon_device *rdev,
+					      struct radeon_ps *radeon_new_state);
+void cypress_notify_link_speed_change_after_state_change(struct radeon_device *rdev,
+							 struct radeon_ps *radeon_new_state,
+							 struct radeon_ps *radeon_current_state);
+int cypress_construct_voltage_tables(struct radeon_device *rdev);
+int cypress_get_mvdd_configuration(struct radeon_device *rdev);
+void cypress_enable_spread_spectrum(struct radeon_device *rdev,
+				    bool enable);
+void cypress_enable_display_gap(struct radeon_device *rdev);
+int cypress_get_table_locations(struct radeon_device *rdev);
+int cypress_populate_mc_reg_table(struct radeon_device *rdev,
+				  struct radeon_ps *radeon_boot_state);
+void cypress_program_response_times(struct radeon_device *rdev);
+int cypress_notify_smc_display_change(struct radeon_device *rdev,
+				      bool has_display);
+void cypress_enable_sclk_control(struct radeon_device *rdev,
+				 bool enable);
+void cypress_enable_mclk_control(struct radeon_device *rdev,
+				 bool enable);
+void cypress_start_dpm(struct radeon_device *rdev);
+void cypress_advertise_gen2_capability(struct radeon_device *rdev);
+u32 cypress_map_clkf_to_ibias(struct radeon_device *rdev, u32 clkf);
+u8 cypress_get_mclk_frequency_ratio(struct radeon_device *rdev,
+				    u32 memory_clock, bool strobe_mode);
+u8 cypress_get_strobe_mode_settings(struct radeon_device *rdev, u32 mclk);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 0f89ce3d02b9..038dcac7670c 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -33,9 +33,7 @@
 #include "avivod.h"
 #include "evergreen_reg.h"
 #include "evergreen_blit_shaders.h"
-
-#define EVERGREEN_PFP_UCODE_SIZE 1120
-#define EVERGREEN_PM4_UCODE_SIZE 1376
+#include "radeon_ucode.h"
 
 static const u32 crtc_offsets[6] =
 {
@@ -47,11 +45,102 @@ static const u32 crtc_offsets[6] =
 	EVERGREEN_CRTC5_REGISTER_OFFSET
 };
 
+#include "clearstate_evergreen.h"
+
+static u32 sumo_rlc_save_restore_register_list[] =
+{
+	0x98fc,
+	0x9830,
+	0x9834,
+	0x9838,
+	0x9870,
+	0x9874,
+	0x8a14,
+	0x8b24,
+	0x8bcc,
+	0x8b10,
+	0x8d00,
+	0x8d04,
+	0x8c00,
+	0x8c04,
+	0x8c08,
+	0x8c0c,
+	0x8d8c,
+	0x8c20,
+	0x8c24,
+	0x8c28,
+	0x8c18,
+	0x8c1c,
+	0x8cf0,
+	0x8e2c,
+	0x8e38,
+	0x8c30,
+	0x9508,
+	0x9688,
+	0x9608,
+	0x960c,
+	0x9610,
+	0x9614,
+	0x88c4,
+	0x88d4,
+	0xa008,
+	0x900c,
+	0x9100,
+	0x913c,
+	0x98f8,
+	0x98f4,
+	0x9b7c,
+	0x3f8c,
+	0x8950,
+	0x8954,
+	0x8a18,
+	0x8b28,
+	0x9144,
+	0x9148,
+	0x914c,
+	0x3f90,
+	0x3f94,
+	0x915c,
+	0x9160,
+	0x9178,
+	0x917c,
+	0x9180,
+	0x918c,
+	0x9190,
+	0x9194,
+	0x9198,
+	0x919c,
+	0x91a8,
+	0x91ac,
+	0x91b0,
+	0x91b4,
+	0x91b8,
+	0x91c4,
+	0x91c8,
+	0x91cc,
+	0x91d0,
+	0x91d4,
+	0x91e0,
+	0x91e4,
+	0x91ec,
+	0x91f0,
+	0x91f4,
+	0x9200,
+	0x9204,
+	0x929c,
+	0x9150,
+	0x802c,
+};
+static u32 sumo_rlc_save_restore_register_list_size = ARRAY_SIZE(sumo_rlc_save_restore_register_list);
+
 static void evergreen_gpu_init(struct radeon_device *rdev);
 void evergreen_fini(struct radeon_device *rdev);
 void evergreen_pcie_gen2_enable(struct radeon_device *rdev);
+void evergreen_program_aspm(struct radeon_device *rdev);
 extern void cayman_cp_int_cntl_setup(struct radeon_device *rdev,
 				     int ring, u32 cp_int_cntl);
+extern void cayman_vm_decode_fault(struct radeon_device *rdev,
+				   u32 status, u32 addr);
 
 static const u32 evergreen_golden_registers[] =
 {
@@ -1417,8 +1506,8 @@ void evergreen_pm_misc(struct radeon_device *rdev)
 	struct radeon_voltage *voltage = &ps->clock_info[req_cm_idx].voltage;
 
 	if (voltage->type == VOLTAGE_SW) {
-		/* 0xff01 is a flag rather then an actual voltage */
-		if (voltage->voltage == 0xff01)
+		/* 0xff0x are flags rather then an actual voltage */
+		if ((voltage->voltage & 0xff00) == 0xff00)
 			return;
 		if (voltage->voltage && (voltage->voltage != rdev->pm.current_vddc)) {
 			radeon_atom_set_voltage(rdev, voltage->voltage, SET_VOLTAGE_TYPE_ASIC_VDDC);
@@ -1438,8 +1527,8 @@ void evergreen_pm_misc(struct radeon_device *rdev)
 			voltage = &rdev->pm.power_state[req_ps_idx].
 				clock_info[rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_cm_idx].voltage;
 
-		/* 0xff01 is a flag rather then an actual voltage */
-		if (voltage->vddci == 0xff01)
+		/* 0xff0x are flags rather then an actual voltage */
+		if ((voltage->vddci & 0xff00) == 0xff00)
 			return;
 		if (voltage->vddci && (voltage->vddci != rdev->pm.current_vddci)) {
 			radeon_atom_set_voltage(rdev, voltage->vddci, SET_VOLTAGE_TYPE_ASIC_VDDCI);
@@ -2036,7 +2125,8 @@ static void evergreen_program_watermarks(struct radeon_device *rdev,
 					 u32 lb_size, u32 num_heads)
 {
 	struct drm_display_mode *mode = &radeon_crtc->base.mode;
-	struct evergreen_wm_params wm;
+	struct evergreen_wm_params wm_low, wm_high;
+	u32 dram_channels;
 	u32 pixel_period;
 	u32 line_time = 0;
 	u32 latency_watermark_a = 0, latency_watermark_b = 0;
@@ -2052,39 +2142,81 @@ static void evergreen_program_watermarks(struct radeon_device *rdev,
 		line_time = min((u32)mode->crtc_htotal * pixel_period, (u32)65535);
 		priority_a_cnt = 0;
 		priority_b_cnt = 0;
+		dram_channels = evergreen_get_number_of_dram_channels(rdev);
+
+		/* watermark for high clocks */
+		if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+			wm_high.yclk =
+				radeon_dpm_get_mclk(rdev, false) * 10;
+			wm_high.sclk =
+				radeon_dpm_get_sclk(rdev, false) * 10;
+		} else {
+			wm_high.yclk = rdev->pm.current_mclk * 10;
+			wm_high.sclk = rdev->pm.current_sclk * 10;
+		}
 
-		wm.yclk = rdev->pm.current_mclk * 10;
-		wm.sclk = rdev->pm.current_sclk * 10;
-		wm.disp_clk = mode->clock;
-		wm.src_width = mode->crtc_hdisplay;
-		wm.active_time = mode->crtc_hdisplay * pixel_period;
-		wm.blank_time = line_time - wm.active_time;
-		wm.interlaced = false;
+		wm_high.disp_clk = mode->clock;
+		wm_high.src_width = mode->crtc_hdisplay;
+		wm_high.active_time = mode->crtc_hdisplay * pixel_period;
+		wm_high.blank_time = line_time - wm_high.active_time;
+		wm_high.interlaced = false;
 		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-			wm.interlaced = true;
-		wm.vsc = radeon_crtc->vsc;
-		wm.vtaps = 1;
+			wm_high.interlaced = true;
+		wm_high.vsc = radeon_crtc->vsc;
+		wm_high.vtaps = 1;
 		if (radeon_crtc->rmx_type != RMX_OFF)
-			wm.vtaps = 2;
-		wm.bytes_per_pixel = 4; /* XXX: get this from fb config */
-		wm.lb_size = lb_size;
-		wm.dram_channels = evergreen_get_number_of_dram_channels(rdev);
-		wm.num_heads = num_heads;
+			wm_high.vtaps = 2;
+		wm_high.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm_high.lb_size = lb_size;
+		wm_high.dram_channels = dram_channels;
+		wm_high.num_heads = num_heads;
+
+		/* watermark for low clocks */
+		if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+			wm_low.yclk =
+				radeon_dpm_get_mclk(rdev, true) * 10;
+			wm_low.sclk =
+				radeon_dpm_get_sclk(rdev, true) * 10;
+		} else {
+			wm_low.yclk = rdev->pm.current_mclk * 10;
+			wm_low.sclk = rdev->pm.current_sclk * 10;
+		}
+
+		wm_low.disp_clk = mode->clock;
+		wm_low.src_width = mode->crtc_hdisplay;
+		wm_low.active_time = mode->crtc_hdisplay * pixel_period;
+		wm_low.blank_time = line_time - wm_low.active_time;
+		wm_low.interlaced = false;
+		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+			wm_low.interlaced = true;
+		wm_low.vsc = radeon_crtc->vsc;
+		wm_low.vtaps = 1;
+		if (radeon_crtc->rmx_type != RMX_OFF)
+			wm_low.vtaps = 2;
+		wm_low.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm_low.lb_size = lb_size;
+		wm_low.dram_channels = dram_channels;
+		wm_low.num_heads = num_heads;
 
 		/* set for high clocks */
-		latency_watermark_a = min(evergreen_latency_watermark(&wm), (u32)65535);
+		latency_watermark_a = min(evergreen_latency_watermark(&wm_high), (u32)65535);
 		/* set for low clocks */
-		/* wm.yclk = low clk; wm.sclk = low clk */
-		latency_watermark_b = min(evergreen_latency_watermark(&wm), (u32)65535);
+		latency_watermark_b = min(evergreen_latency_watermark(&wm_low), (u32)65535);
 
 		/* possibly force display priority to high */
 		/* should really do this at mode validation time... */
-		if (!evergreen_average_bandwidth_vs_dram_bandwidth_for_display(&wm) ||
-		    !evergreen_average_bandwidth_vs_available_bandwidth(&wm) ||
-		    !evergreen_check_latency_hiding(&wm) ||
+		if (!evergreen_average_bandwidth_vs_dram_bandwidth_for_display(&wm_high) ||
+		    !evergreen_average_bandwidth_vs_available_bandwidth(&wm_high) ||
+		    !evergreen_check_latency_hiding(&wm_high) ||
 		    (rdev->disp_priority == 2)) {
-			DRM_DEBUG_KMS("force priority to high\n");
+			DRM_DEBUG_KMS("force priority a to high\n");
 			priority_a_cnt |= PRIORITY_ALWAYS_ON;
+		}
+		if (!evergreen_average_bandwidth_vs_dram_bandwidth_for_display(&wm_low) ||
+		    !evergreen_average_bandwidth_vs_available_bandwidth(&wm_low) ||
+		    !evergreen_check_latency_hiding(&wm_low) ||
+		    (rdev->disp_priority == 2)) {
+			DRM_DEBUG_KMS("force priority b to high\n");
 			priority_b_cnt |= PRIORITY_ALWAYS_ON;
 		}
 
@@ -2137,6 +2269,10 @@ static void evergreen_program_watermarks(struct radeon_device *rdev,
 	WREG32(PRIORITY_A_CNT + radeon_crtc->crtc_offset, priority_a_cnt);
 	WREG32(PRIORITY_B_CNT + radeon_crtc->crtc_offset, priority_b_cnt);
 
+	/* save values for DPM */
+	radeon_crtc->line_time = line_time;
+	radeon_crtc->wm_high = latency_watermark_a;
+	radeon_crtc->wm_low = latency_watermark_b;
 }
 
 /**
@@ -3120,10 +3256,8 @@ static void evergreen_gpu_init(struct radeon_device *rdev)
 		u32 efuse_straps_4;
 		u32 efuse_straps_3;
 
-		WREG32(RCU_IND_INDEX, 0x204);
-		efuse_straps_4 = RREG32(RCU_IND_DATA);
-		WREG32(RCU_IND_INDEX, 0x203);
-		efuse_straps_3 = RREG32(RCU_IND_DATA);
+		efuse_straps_4 = RREG32_RCU(0x204);
+		efuse_straps_3 = RREG32_RCU(0x203);
 		tmp = (((efuse_straps_4 & 0xf) << 4) |
 		      ((efuse_straps_3 & 0xf0000000) >> 28));
 	} else {
@@ -3727,6 +3861,262 @@ bool evergreen_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *rin
 	return radeon_ring_test_lockup(rdev, ring);
 }
 
+/*
+ * RLC
+ */
+#define RLC_SAVE_RESTORE_LIST_END_MARKER    0x00000000
+#define RLC_CLEAR_STATE_END_MARKER          0x00000001
+
+void sumo_rlc_fini(struct radeon_device *rdev)
+{
+	int r;
+
+	/* save restore block */
+	if (rdev->rlc.save_restore_obj) {
+		r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false);
+		if (unlikely(r != 0))
+			dev_warn(rdev->dev, "(%d) reserve RLC sr bo failed\n", r);
+		radeon_bo_unpin(rdev->rlc.save_restore_obj);
+		radeon_bo_unreserve(rdev->rlc.save_restore_obj);
+
+		radeon_bo_unref(&rdev->rlc.save_restore_obj);
+		rdev->rlc.save_restore_obj = NULL;
+	}
+
+	/* clear state block */
+	if (rdev->rlc.clear_state_obj) {
+		r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false);
+		if (unlikely(r != 0))
+			dev_warn(rdev->dev, "(%d) reserve RLC c bo failed\n", r);
+		radeon_bo_unpin(rdev->rlc.clear_state_obj);
+		radeon_bo_unreserve(rdev->rlc.clear_state_obj);
+
+		radeon_bo_unref(&rdev->rlc.clear_state_obj);
+		rdev->rlc.clear_state_obj = NULL;
+	}
+}
+
+int sumo_rlc_init(struct radeon_device *rdev)
+{
+	u32 *src_ptr;
+	volatile u32 *dst_ptr;
+	u32 dws, data, i, j, k, reg_num;
+	u32 reg_list_num, reg_list_hdr_blk_index, reg_list_blk_index;
+	u64 reg_list_mc_addr;
+	struct cs_section_def *cs_data;
+	int r;
+
+	src_ptr = rdev->rlc.reg_list;
+	dws = rdev->rlc.reg_list_size;
+	cs_data = rdev->rlc.cs_data;
+
+	/* save restore block */
+	if (rdev->rlc.save_restore_obj == NULL) {
+		r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
+				     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.save_restore_obj);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) create RLC sr bo failed\n", r);
+			return r;
+		}
+	}
+
+	r = radeon_bo_reserve(rdev->rlc.save_restore_obj, false);
+	if (unlikely(r != 0)) {
+		sumo_rlc_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_pin(rdev->rlc.save_restore_obj, RADEON_GEM_DOMAIN_VRAM,
+			  &rdev->rlc.save_restore_gpu_addr);
+	if (r) {
+		radeon_bo_unreserve(rdev->rlc.save_restore_obj);
+		dev_warn(rdev->dev, "(%d) pin RLC sr bo failed\n", r);
+		sumo_rlc_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_kmap(rdev->rlc.save_restore_obj, (void **)&rdev->rlc.sr_ptr);
+	if (r) {
+		dev_warn(rdev->dev, "(%d) map RLC sr bo failed\n", r);
+		sumo_rlc_fini(rdev);
+		return r;
+	}
+	/* write the sr buffer */
+	dst_ptr = rdev->rlc.sr_ptr;
+	/* format:
+	 * dw0: (reg2 << 16) | reg1
+	 * dw1: reg1 save space
+	 * dw2: reg2 save space
+	 */
+	for (i = 0; i < dws; i++) {
+		data = src_ptr[i] >> 2;
+		i++;
+		if (i < dws)
+			data |= (src_ptr[i] >> 2) << 16;
+		j = (((i - 1) * 3) / 2);
+		dst_ptr[j] = data;
+	}
+	j = ((i * 3) / 2);
+	dst_ptr[j] = RLC_SAVE_RESTORE_LIST_END_MARKER;
+
+	radeon_bo_kunmap(rdev->rlc.save_restore_obj);
+	radeon_bo_unreserve(rdev->rlc.save_restore_obj);
+
+	/* clear state block */
+	reg_list_num = 0;
+	dws = 0;
+	for (i = 0; cs_data[i].section != NULL; i++) {
+		for (j = 0; cs_data[i].section[j].extent != NULL; j++) {
+			reg_list_num++;
+			dws += cs_data[i].section[j].reg_count;
+		}
+	}
+	reg_list_blk_index = (3 * reg_list_num + 2);
+	dws += reg_list_blk_index;
+
+	if (rdev->rlc.clear_state_obj == NULL) {
+		r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
+				     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.clear_state_obj);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r);
+			sumo_rlc_fini(rdev);
+			return r;
+		}
+	}
+	r = radeon_bo_reserve(rdev->rlc.clear_state_obj, false);
+	if (unlikely(r != 0)) {
+		sumo_rlc_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_pin(rdev->rlc.clear_state_obj, RADEON_GEM_DOMAIN_VRAM,
+			  &rdev->rlc.clear_state_gpu_addr);
+	if (r) {
+
+		radeon_bo_unreserve(rdev->rlc.clear_state_obj);
+		dev_warn(rdev->dev, "(%d) pin RLC c bo failed\n", r);
+		sumo_rlc_fini(rdev);
+		return r;
+	}
+	r = radeon_bo_kmap(rdev->rlc.clear_state_obj, (void **)&rdev->rlc.cs_ptr);
+	if (r) {
+		dev_warn(rdev->dev, "(%d) map RLC c bo failed\n", r);
+		sumo_rlc_fini(rdev);
+		return r;
+	}
+	/* set up the cs buffer */
+	dst_ptr = rdev->rlc.cs_ptr;
+	reg_list_hdr_blk_index = 0;
+	reg_list_mc_addr = rdev->rlc.clear_state_gpu_addr + (reg_list_blk_index * 4);
+	data = upper_32_bits(reg_list_mc_addr);
+	dst_ptr[reg_list_hdr_blk_index] = data;
+	reg_list_hdr_blk_index++;
+	for (i = 0; cs_data[i].section != NULL; i++) {
+		for (j = 0; cs_data[i].section[j].extent != NULL; j++) {
+			reg_num = cs_data[i].section[j].reg_count;
+			data = reg_list_mc_addr & 0xffffffff;
+			dst_ptr[reg_list_hdr_blk_index] = data;
+			reg_list_hdr_blk_index++;
+
+			data = (cs_data[i].section[j].reg_index * 4) & 0xffffffff;
+			dst_ptr[reg_list_hdr_blk_index] = data;
+			reg_list_hdr_blk_index++;
+
+			data = 0x08000000 | (reg_num * 4);
+			dst_ptr[reg_list_hdr_blk_index] = data;
+			reg_list_hdr_blk_index++;
+
+			for (k = 0; k < reg_num; k++) {
+				data = cs_data[i].section[j].extent[k];
+				dst_ptr[reg_list_blk_index + k] = data;
+			}
+			reg_list_mc_addr += reg_num * 4;
+			reg_list_blk_index += reg_num;
+		}
+	}
+	dst_ptr[reg_list_hdr_blk_index] = RLC_CLEAR_STATE_END_MARKER;
+
+	radeon_bo_kunmap(rdev->rlc.clear_state_obj);
+	radeon_bo_unreserve(rdev->rlc.clear_state_obj);
+
+	return 0;
+}
+
+static void evergreen_rlc_start(struct radeon_device *rdev)
+{
+	u32 mask = RLC_ENABLE;
+
+	if (rdev->flags & RADEON_IS_IGP) {
+		mask |= GFX_POWER_GATING_ENABLE | GFX_POWER_GATING_SRC;
+	}
+
+	WREG32(RLC_CNTL, mask);
+}
+
+int evergreen_rlc_resume(struct radeon_device *rdev)
+{
+	u32 i;
+	const __be32 *fw_data;
+
+	if (!rdev->rlc_fw)
+		return -EINVAL;
+
+	r600_rlc_stop(rdev);
+
+	WREG32(RLC_HB_CNTL, 0);
+
+	if (rdev->flags & RADEON_IS_IGP) {
+		if (rdev->family == CHIP_ARUBA) {
+			u32 always_on_bitmap =
+				3 | (3 << (16 * rdev->config.cayman.max_shader_engines));
+			/* find out the number of active simds */
+			u32 tmp = (RREG32(CC_GC_SHADER_PIPE_CONFIG) & 0xffff0000) >> 16;
+			tmp |= 0xffffffff << rdev->config.cayman.max_simds_per_se;
+			tmp = hweight32(~tmp);
+			if (tmp == rdev->config.cayman.max_simds_per_se) {
+				WREG32(TN_RLC_LB_ALWAYS_ACTIVE_SIMD_MASK, always_on_bitmap);
+				WREG32(TN_RLC_LB_PARAMS, 0x00601004);
+				WREG32(TN_RLC_LB_INIT_SIMD_MASK, 0xffffffff);
+				WREG32(TN_RLC_LB_CNTR_INIT, 0x00000000);
+				WREG32(TN_RLC_LB_CNTR_MAX, 0x00002000);
+			}
+		} else {
+			WREG32(RLC_HB_WPTR_LSB_ADDR, 0);
+			WREG32(RLC_HB_WPTR_MSB_ADDR, 0);
+		}
+		WREG32(TN_RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
+		WREG32(TN_RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+	} else {
+		WREG32(RLC_HB_BASE, 0);
+		WREG32(RLC_HB_RPTR, 0);
+		WREG32(RLC_HB_WPTR, 0);
+		WREG32(RLC_HB_WPTR_LSB_ADDR, 0);
+		WREG32(RLC_HB_WPTR_MSB_ADDR, 0);
+	}
+	WREG32(RLC_MC_CNTL, 0);
+	WREG32(RLC_UCODE_CNTL, 0);
+
+	fw_data = (const __be32 *)rdev->rlc_fw->data;
+	if (rdev->family >= CHIP_ARUBA) {
+		for (i = 0; i < ARUBA_RLC_UCODE_SIZE; i++) {
+			WREG32(RLC_UCODE_ADDR, i);
+			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
+		}
+	} else if (rdev->family >= CHIP_CAYMAN) {
+		for (i = 0; i < CAYMAN_RLC_UCODE_SIZE; i++) {
+			WREG32(RLC_UCODE_ADDR, i);
+			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
+		}
+	} else {
+		for (i = 0; i < EVERGREEN_RLC_UCODE_SIZE; i++) {
+			WREG32(RLC_UCODE_ADDR, i);
+			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
+		}
+	}
+	WREG32(RLC_UCODE_ADDR, 0);
+
+	evergreen_rlc_start(rdev);
+
+	return 0;
+}
+
 /* Interrupts */
 
 u32 evergreen_get_vblank_counter(struct radeon_device *rdev, int crtc)
@@ -3805,6 +4195,7 @@ int evergreen_irq_set(struct radeon_device *rdev)
 	u32 grph1 = 0, grph2 = 0, grph3 = 0, grph4 = 0, grph5 = 0, grph6 = 0;
 	u32 afmt1 = 0, afmt2 = 0, afmt3 = 0, afmt4 = 0, afmt5 = 0, afmt6 = 0;
 	u32 dma_cntl, dma_cntl1 = 0;
+	u32 thermal_int = 0;
 
 	if (!rdev->irq.installed) {
 		WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
@@ -3824,6 +4215,12 @@ int evergreen_irq_set(struct radeon_device *rdev)
 	hpd4 = RREG32(DC_HPD4_INT_CONTROL) & ~DC_HPDx_INT_EN;
 	hpd5 = RREG32(DC_HPD5_INT_CONTROL) & ~DC_HPDx_INT_EN;
 	hpd6 = RREG32(DC_HPD6_INT_CONTROL) & ~DC_HPDx_INT_EN;
+	if (rdev->family == CHIP_ARUBA)
+		thermal_int = RREG32(TN_CG_THERMAL_INT_CTRL) &
+			~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW);
+	else
+		thermal_int = RREG32(CG_THERMAL_INT) &
+			~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW);
 
 	afmt1 = RREG32(AFMT_AUDIO_PACKET_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET) & ~AFMT_AZ_FORMAT_WTRIG_MASK;
 	afmt2 = RREG32(AFMT_AUDIO_PACKET_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET) & ~AFMT_AZ_FORMAT_WTRIG_MASK;
@@ -3869,6 +4266,11 @@ int evergreen_irq_set(struct radeon_device *rdev)
 		}
 	}
 
+	if (rdev->irq.dpm_thermal) {
+		DRM_DEBUG("dpm thermal\n");
+		thermal_int |= THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW;
+	}
+
 	if (rdev->irq.crtc_vblank_int[0] ||
 	    atomic_read(&rdev->irq.pflip[0])) {
 		DRM_DEBUG("evergreen_irq_set: vblank 0\n");
@@ -3990,6 +4392,10 @@ int evergreen_irq_set(struct radeon_device *rdev)
 	WREG32(DC_HPD4_INT_CONTROL, hpd4);
 	WREG32(DC_HPD5_INT_CONTROL, hpd5);
 	WREG32(DC_HPD6_INT_CONTROL, hpd6);
+	if (rdev->family == CHIP_ARUBA)
+		WREG32(TN_CG_THERMAL_INT_CTRL, thermal_int);
+	else
+		WREG32(CG_THERMAL_INT, thermal_int);
 
 	WREG32(AFMT_AUDIO_PACKET_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET, afmt1);
 	WREG32(AFMT_AUDIO_PACKET_CONTROL + EVERGREEN_CRTC1_REGISTER_OFFSET, afmt2);
@@ -4181,6 +4587,8 @@ int evergreen_irq_process(struct radeon_device *rdev)
 	u32 ring_index;
 	bool queue_hotplug = false;
 	bool queue_hdmi = false;
+	bool queue_thermal = false;
+	u32 status, addr;
 
 	if (!rdev->ih.enabled || rdev->shutdown)
 		return IRQ_NONE;
@@ -4467,11 +4875,14 @@ restart_ih:
 			break;
 		case 146:
 		case 147:
+			addr = RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR);
+			status = RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS);
 			dev_err(rdev->dev, "GPU fault detected: %d 0x%08x\n", src_id, src_data);
 			dev_err(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x%08X\n",
-				RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR));
+				addr);
 			dev_err(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
-				RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS));
+				status);
+			cayman_vm_decode_fault(rdev, status, addr);
 			/* reset addr and status */
 			WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1);
 			break;
@@ -4502,6 +4913,16 @@ restart_ih:
 			DRM_DEBUG("IH: DMA trap\n");
 			radeon_fence_process(rdev, R600_RING_TYPE_DMA_INDEX);
 			break;
+		case 230: /* thermal low to high */
+			DRM_DEBUG("IH: thermal low to high\n");
+			rdev->pm.dpm.thermal.high_to_low = false;
+			queue_thermal = true;
+			break;
+		case 231: /* thermal high to low */
+			DRM_DEBUG("IH: thermal high to low\n");
+			rdev->pm.dpm.thermal.high_to_low = true;
+			queue_thermal = true;
+			break;
 		case 233: /* GUI IDLE */
 			DRM_DEBUG("IH: GUI idle\n");
 			break;
@@ -4524,6 +4945,8 @@ restart_ih:
 		schedule_work(&rdev->hotplug_work);
 	if (queue_hdmi)
 		schedule_work(&rdev->audio_work);
+	if (queue_thermal && rdev->pm.dpm_enabled)
+		schedule_work(&rdev->pm.dpm.thermal.work);
 	rdev->ih.rptr = rptr;
 	WREG32(IH_RB_RPTR, rdev->ih.rptr);
 	atomic_set(&rdev->ih.lock, 0);
@@ -4680,6 +5103,8 @@ static int evergreen_startup(struct radeon_device *rdev)
 
 	/* enable pcie gen2 link */
 	evergreen_pcie_gen2_enable(rdev);
+	/* enable aspm */
+	evergreen_program_aspm(rdev);
 
 	if (ASIC_IS_DCE5(rdev)) {
 		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) {
@@ -4725,6 +5150,18 @@ static int evergreen_startup(struct radeon_device *rdev)
 		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
 	}
 
+	/* allocate rlc buffers */
+	if (rdev->flags & RADEON_IS_IGP) {
+		rdev->rlc.reg_list = sumo_rlc_save_restore_register_list;
+		rdev->rlc.reg_list_size = sumo_rlc_save_restore_register_list_size;
+		rdev->rlc.cs_data = evergreen_cs_data;
+		r = sumo_rlc_init(rdev);
+		if (r) {
+			DRM_ERROR("Failed to init rlc BOs!\n");
+			return r;
+		}
+	}
+
 	/* allocate wb buffer */
 	r = radeon_wb_init(rdev);
 	if (r)
@@ -4956,6 +5393,8 @@ int evergreen_init(struct radeon_device *rdev)
 		r700_cp_fini(rdev);
 		r600_dma_fini(rdev);
 		r600_irq_fini(rdev);
+		if (rdev->flags & RADEON_IS_IGP)
+			sumo_rlc_fini(rdev);
 		radeon_wb_fini(rdev);
 		radeon_ib_pool_fini(rdev);
 		radeon_irq_kms_fini(rdev);
@@ -4984,6 +5423,8 @@ void evergreen_fini(struct radeon_device *rdev)
 	r700_cp_fini(rdev);
 	r600_dma_fini(rdev);
 	r600_irq_fini(rdev);
+	if (rdev->flags & RADEON_IS_IGP)
+		sumo_rlc_fini(rdev);
 	radeon_wb_fini(rdev);
 	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
@@ -5061,3 +5502,153 @@ void evergreen_pcie_gen2_enable(struct radeon_device *rdev)
 		WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, link_width_cntl);
 	}
 }
+
+void evergreen_program_aspm(struct radeon_device *rdev)
+{
+	u32 data, orig;
+	u32 pcie_lc_cntl, pcie_lc_cntl_old;
+	bool disable_l0s, disable_l1 = false, disable_plloff_in_l1 = false;
+	/* fusion_platform = true
+	 * if the system is a fusion system
+	 * (APU or DGPU in a fusion system).
+	 * todo: check if the system is a fusion platform.
+	 */
+	bool fusion_platform = false;
+
+	if (radeon_aspm == 0)
+		return;
+
+	if (!(rdev->flags & RADEON_IS_PCIE))
+		return;
+
+	switch (rdev->family) {
+	case CHIP_CYPRESS:
+	case CHIP_HEMLOCK:
+	case CHIP_JUNIPER:
+	case CHIP_REDWOOD:
+	case CHIP_CEDAR:
+	case CHIP_SUMO:
+	case CHIP_SUMO2:
+	case CHIP_PALM:
+	case CHIP_ARUBA:
+		disable_l0s = true;
+		break;
+	default:
+		disable_l0s = false;
+		break;
+	}
+
+	if (rdev->flags & RADEON_IS_IGP)
+		fusion_platform = true; /* XXX also dGPUs in a fusion system */
+
+	data = orig = RREG32_PIF_PHY0(PB0_PIF_PAIRING);
+	if (fusion_platform)
+		data &= ~MULTI_PIF;
+	else
+		data |= MULTI_PIF;
+	if (data != orig)
+		WREG32_PIF_PHY0(PB0_PIF_PAIRING, data);
+
+	data = orig = RREG32_PIF_PHY1(PB1_PIF_PAIRING);
+	if (fusion_platform)
+		data &= ~MULTI_PIF;
+	else
+		data |= MULTI_PIF;
+	if (data != orig)
+		WREG32_PIF_PHY1(PB1_PIF_PAIRING, data);
+
+	pcie_lc_cntl = pcie_lc_cntl_old = RREG32_PCIE_PORT(PCIE_LC_CNTL);
+	pcie_lc_cntl &= ~(LC_L0S_INACTIVITY_MASK | LC_L1_INACTIVITY_MASK);
+	if (!disable_l0s) {
+		if (rdev->family >= CHIP_BARTS)
+			pcie_lc_cntl |= LC_L0S_INACTIVITY(7);
+		else
+			pcie_lc_cntl |= LC_L0S_INACTIVITY(3);
+	}
+
+	if (!disable_l1) {
+		if (rdev->family >= CHIP_BARTS)
+			pcie_lc_cntl |= LC_L1_INACTIVITY(7);
+		else
+			pcie_lc_cntl |= LC_L1_INACTIVITY(8);
+
+		if (!disable_plloff_in_l1) {
+			data = orig = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0);
+			data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7);
+			if (data != orig)
+				WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0, data);
+
+			data = orig = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1);
+			data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7);
+			if (data != orig)
+				WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1, data);
+
+			data = orig = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0);
+			data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7);
+			if (data != orig)
+				WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0, data);
+
+			data = orig = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1);
+			data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7);
+			if (data != orig)
+				WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1, data);
+
+			if (rdev->family >= CHIP_BARTS) {
+				data = orig = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0);
+				data &= ~PLL_RAMP_UP_TIME_0_MASK;
+				data |= PLL_RAMP_UP_TIME_0(4);
+				if (data != orig)
+					WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0, data);
+
+				data = orig = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1);
+				data &= ~PLL_RAMP_UP_TIME_1_MASK;
+				data |= PLL_RAMP_UP_TIME_1(4);
+				if (data != orig)
+					WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1, data);
+
+				data = orig = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0);
+				data &= ~PLL_RAMP_UP_TIME_0_MASK;
+				data |= PLL_RAMP_UP_TIME_0(4);
+				if (data != orig)
+					WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0, data);
+
+				data = orig = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1);
+				data &= ~PLL_RAMP_UP_TIME_1_MASK;
+				data |= PLL_RAMP_UP_TIME_1(4);
+				if (data != orig)
+					WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1, data);
+			}
+
+			data = orig = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
+			data &= ~LC_DYN_LANES_PWR_STATE_MASK;
+			data |= LC_DYN_LANES_PWR_STATE(3);
+			if (data != orig)
+				WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, data);
+
+			if (rdev->family >= CHIP_BARTS) {
+				data = orig = RREG32_PIF_PHY0(PB0_PIF_CNTL);
+				data &= ~LS2_EXIT_TIME_MASK;
+				data |= LS2_EXIT_TIME(1);
+				if (data != orig)
+					WREG32_PIF_PHY0(PB0_PIF_CNTL, data);
+
+				data = orig = RREG32_PIF_PHY1(PB1_PIF_CNTL);
+				data &= ~LS2_EXIT_TIME_MASK;
+				data |= LS2_EXIT_TIME(1);
+				if (data != orig)
+					WREG32_PIF_PHY1(PB1_PIF_CNTL, data);
+			}
+		}
+	}
+
+	/* evergreen parts only */
+	if (rdev->family < CHIP_BARTS)
+		pcie_lc_cntl |= LC_PMI_TO_L1_DIS;
+
+	if (pcie_lc_cntl != pcie_lc_cntl_old)
+		WREG32_PCIE_PORT(PCIE_LC_CNTL, pcie_lc_cntl);
+}
diff --git a/drivers/gpu/drm/radeon/evergreen_hdmi.c b/drivers/gpu/drm/radeon/evergreen_hdmi.c
index ed7c8a768092..b0d3fb341417 100644
--- a/drivers/gpu/drm/radeon/evergreen_hdmi.c
+++ b/drivers/gpu/drm/radeon/evergreen_hdmi.c
@@ -128,14 +128,7 @@ static void evergreen_hdmi_update_avi_infoframe(struct drm_encoder *encoder,
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 	uint32_t offset = dig->afmt->offset;
 	uint8_t *frame = buffer + 3;
-
-	/* Our header values (type, version, length) should be alright, Intel
-	 * is using the same. Checksum function also seems to be OK, it works
-	 * fine for audio infoframe. However calculated value is always lower
-	 * by 2 in comparison to fglrx. It breaks displaying anything in case
-	 * of TVs that strictly check the checksum. Hack it manually here to
-	 * workaround this issue. */
-	frame[0x0] += 2;
+	uint8_t *header = buffer;
 
 	WREG32(AFMT_AVI_INFO0 + offset,
 		frame[0x0] | (frame[0x1] << 8) | (frame[0x2] << 16) | (frame[0x3] << 24));
@@ -144,7 +137,7 @@ static void evergreen_hdmi_update_avi_infoframe(struct drm_encoder *encoder,
 	WREG32(AFMT_AVI_INFO2 + offset,
 		frame[0x8] | (frame[0x9] << 8) | (frame[0xA] << 16) | (frame[0xB] << 24));
 	WREG32(AFMT_AVI_INFO3 + offset,
-		frame[0xC] | (frame[0xD] << 8));
+		frame[0xC] | (frame[0xD] << 8) | (header[1] << 24));
 }
 
 static void evergreen_audio_set_dto(struct drm_encoder *encoder, u32 clock)
@@ -184,6 +177,9 @@ void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode
 	uint32_t offset;
 	ssize_t err;
 
+	if (!dig || !dig->afmt)
+		return;
+
 	/* Silent, r600_hdmi_enable will raise WARN for us */
 	if (!dig->afmt->enabled)
 		return;
@@ -287,6 +283,9 @@ void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable)
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 
+	if (!dig || !dig->afmt)
+		return;
+
 	/* Silent, r600_hdmi_enable will raise WARN for us */
 	if (enable && dig->afmt->enabled)
 		return;
diff --git a/drivers/gpu/drm/radeon/evergreen_reg.h b/drivers/gpu/drm/radeon/evergreen_reg.h
index 881aba23c477..8a4e641f0e3c 100644
--- a/drivers/gpu/drm/radeon/evergreen_reg.h
+++ b/drivers/gpu/drm/radeon/evergreen_reg.h
@@ -24,7 +24,16 @@
 #ifndef __EVERGREEN_REG_H__
 #define __EVERGREEN_REG_H__
 
+/* trinity */
+#define TN_SMC_IND_INDEX_0                              0x200
+#define TN_SMC_IND_DATA_0                               0x204
+
 /* evergreen */
+#define EVERGREEN_PIF_PHY0_INDEX                        0x8
+#define EVERGREEN_PIF_PHY0_DATA                         0xc
+#define EVERGREEN_PIF_PHY1_INDEX                        0x10
+#define EVERGREEN_PIF_PHY1_DATA                         0x14
+
 #define EVERGREEN_VGA_MEMORY_BASE_ADDRESS               0x310
 #define EVERGREEN_VGA_MEMORY_BASE_ADDRESS_HIGH          0x324
 #define EVERGREEN_D3VGA_CONTROL                         0x3e0
@@ -40,6 +49,9 @@
 #define EVERGREEN_AUDIO_PLL1_DIV			0x5b4
 #define EVERGREEN_AUDIO_PLL1_UNK			0x5bc
 
+#define EVERGREEN_CG_IND_ADDR                           0x8f8
+#define EVERGREEN_CG_IND_DATA                           0x8fc
+
 #define EVERGREEN_AUDIO_ENABLE				0x5e78
 #define EVERGREEN_AUDIO_VENDOR_ID			0x5ec0
 
diff --git a/drivers/gpu/drm/radeon/evergreen_smc.h b/drivers/gpu/drm/radeon/evergreen_smc.h
new file mode 100644
index 000000000000..76ada8cfe902
--- /dev/null
+++ b/drivers/gpu/drm/radeon/evergreen_smc.h
@@ -0,0 +1,67 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __EVERGREEN_SMC_H__
+#define __EVERGREEN_SMC_H__
+
+#include "rv770_smc.h"
+
+#pragma pack(push, 1)
+
+#define SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE 16
+
+struct SMC_Evergreen_MCRegisterAddress
+{
+    uint16_t s0;
+    uint16_t s1;
+};
+
+typedef struct SMC_Evergreen_MCRegisterAddress SMC_Evergreen_MCRegisterAddress;
+
+
+struct SMC_Evergreen_MCRegisterSet
+{
+    uint32_t value[SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE];
+};
+
+typedef struct SMC_Evergreen_MCRegisterSet SMC_Evergreen_MCRegisterSet;
+
+struct SMC_Evergreen_MCRegisters
+{
+    uint8_t                             last;
+    uint8_t                             reserved[3];
+    SMC_Evergreen_MCRegisterAddress     address[SMC_EVERGREEN_MC_REGISTER_ARRAY_SIZE];
+    SMC_Evergreen_MCRegisterSet         data[5];
+};
+
+typedef struct SMC_Evergreen_MCRegisters SMC_Evergreen_MCRegisters;
+
+#define EVERGREEN_SMC_FIRMWARE_HEADER_LOCATION 0x100
+
+#define EVERGREEN_SMC_FIRMWARE_HEADER_softRegisters   0x0
+#define EVERGREEN_SMC_FIRMWARE_HEADER_stateTable      0xC
+#define EVERGREEN_SMC_FIRMWARE_HEADER_mcRegisterTable 0x20
+
+
+#pragma pack(pop)
+
+#endif
diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h
index 75c05631146d..a7baf67aef6c 100644
--- a/drivers/gpu/drm/radeon/evergreend.h
+++ b/drivers/gpu/drm/radeon/evergreend.h
@@ -48,6 +48,293 @@
 #define SUMO_GB_ADDR_CONFIG_GOLDEN           0x02010002
 #define SUMO2_GB_ADDR_CONFIG_GOLDEN          0x02010002
 
+/* pm registers */
+#define	SMC_MSG						0x20c
+#define		HOST_SMC_MSG(x)				((x) << 0)
+#define		HOST_SMC_MSG_MASK			(0xff << 0)
+#define		HOST_SMC_MSG_SHIFT			0
+#define		HOST_SMC_RESP(x)			((x) << 8)
+#define		HOST_SMC_RESP_MASK			(0xff << 8)
+#define		HOST_SMC_RESP_SHIFT			8
+#define		SMC_HOST_MSG(x)				((x) << 16)
+#define		SMC_HOST_MSG_MASK			(0xff << 16)
+#define		SMC_HOST_MSG_SHIFT			16
+#define		SMC_HOST_RESP(x)			((x) << 24)
+#define		SMC_HOST_RESP_MASK			(0xff << 24)
+#define		SMC_HOST_RESP_SHIFT			24
+
+#define DCCG_DISP_SLOW_SELECT_REG                       0x4fc
+#define		DCCG_DISP1_SLOW_SELECT(x)		((x) << 0)
+#define		DCCG_DISP1_SLOW_SELECT_MASK		(7 << 0)
+#define		DCCG_DISP1_SLOW_SELECT_SHIFT		0
+#define		DCCG_DISP2_SLOW_SELECT(x)		((x) << 4)
+#define		DCCG_DISP2_SLOW_SELECT_MASK		(7 << 4)
+#define		DCCG_DISP2_SLOW_SELECT_SHIFT		4
+
+#define	CG_SPLL_FUNC_CNTL				0x600
+#define		SPLL_RESET				(1 << 0)
+#define		SPLL_SLEEP				(1 << 1)
+#define		SPLL_BYPASS_EN				(1 << 3)
+#define		SPLL_REF_DIV(x)				((x) << 4)
+#define		SPLL_REF_DIV_MASK			(0x3f << 4)
+#define		SPLL_PDIV_A(x)				((x) << 20)
+#define		SPLL_PDIV_A_MASK			(0x7f << 20)
+#define	CG_SPLL_FUNC_CNTL_2				0x604
+#define		SCLK_MUX_SEL(x)				((x) << 0)
+#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_SPLL_FUNC_CNTL_3				0x608
+#define		SPLL_FB_DIV(x)				((x) << 0)
+#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		SPLL_DITHEN				(1 << 28)
+
+#define MPLL_CNTL_MODE                                  0x61c
+#       define SS_SSEN                                  (1 << 24)
+#       define SS_DSMODE_EN                             (1 << 25)
+
+#define	MPLL_AD_FUNC_CNTL				0x624
+#define		CLKF(x)					((x) << 0)
+#define		CLKF_MASK				(0x7f << 0)
+#define		CLKR(x)					((x) << 7)
+#define		CLKR_MASK				(0x1f << 7)
+#define		CLKFRAC(x)				((x) << 12)
+#define		CLKFRAC_MASK				(0x1f << 12)
+#define		YCLK_POST_DIV(x)			((x) << 17)
+#define		YCLK_POST_DIV_MASK			(3 << 17)
+#define		IBIAS(x)				((x) << 20)
+#define		IBIAS_MASK				(0x3ff << 20)
+#define		RESET					(1 << 30)
+#define		PDNB					(1 << 31)
+#define	MPLL_AD_FUNC_CNTL_2				0x628
+#define		BYPASS					(1 << 19)
+#define		BIAS_GEN_PDNB				(1 << 24)
+#define		RESET_EN				(1 << 25)
+#define		VCO_MODE				(1 << 29)
+#define	MPLL_DQ_FUNC_CNTL				0x62c
+#define	MPLL_DQ_FUNC_CNTL_2				0x630
+
+#define GENERAL_PWRMGT                                  0x63c
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+#       define STATIC_PM_EN                             (1 << 1)
+#       define THERMAL_PROTECTION_DIS                   (1 << 2)
+#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
+#       define ENABLE_GEN2PCIE                          (1 << 4)
+#       define ENABLE_GEN2XSP                           (1 << 5)
+#       define SW_SMIO_INDEX(x)                         ((x) << 6)
+#       define SW_SMIO_INDEX_MASK                       (3 << 6)
+#       define SW_SMIO_INDEX_SHIFT                      6
+#       define LOW_VOLT_D2_ACPI                         (1 << 8)
+#       define LOW_VOLT_D3_ACPI                         (1 << 9)
+#       define VOLT_PWRMGT_EN                           (1 << 10)
+#       define BACKBIAS_PAD_EN                          (1 << 18)
+#       define BACKBIAS_VALUE                           (1 << 19)
+#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
+#       define AC_DC_SW                                 (1 << 24)
+
+#define SCLK_PWRMGT_CNTL                                  0x644
+#       define SCLK_PWRMGT_OFF                            (1 << 0)
+#       define SCLK_LOW_D1                                (1 << 1)
+#       define FIR_RESET                                  (1 << 4)
+#       define FIR_FORCE_TREND_SEL                        (1 << 5)
+#       define FIR_TREND_MODE                             (1 << 6)
+#       define DYN_GFX_CLK_OFF_EN                         (1 << 7)
+#       define GFX_CLK_FORCE_ON                           (1 << 8)
+#       define GFX_CLK_REQUEST_OFF                        (1 << 9)
+#       define GFX_CLK_FORCE_OFF                          (1 << 10)
+#       define GFX_CLK_OFF_ACPI_D1                        (1 << 11)
+#       define GFX_CLK_OFF_ACPI_D2                        (1 << 12)
+#       define GFX_CLK_OFF_ACPI_D3                        (1 << 13)
+#       define DYN_LIGHT_SLEEP_EN                         (1 << 14)
+#define	MCLK_PWRMGT_CNTL				0x648
+#       define DLL_SPEED(x)				((x) << 0)
+#       define DLL_SPEED_MASK				(0x1f << 0)
+#       define MPLL_PWRMGT_OFF                          (1 << 5)
+#       define DLL_READY                                (1 << 6)
+#       define MC_INT_CNTL                              (1 << 7)
+#       define MRDCKA0_PDNB                             (1 << 8)
+#       define MRDCKA1_PDNB                             (1 << 9)
+#       define MRDCKB0_PDNB                             (1 << 10)
+#       define MRDCKB1_PDNB                             (1 << 11)
+#       define MRDCKC0_PDNB                             (1 << 12)
+#       define MRDCKC1_PDNB                             (1 << 13)
+#       define MRDCKD0_PDNB                             (1 << 14)
+#       define MRDCKD1_PDNB                             (1 << 15)
+#       define MRDCKA0_RESET                            (1 << 16)
+#       define MRDCKA1_RESET                            (1 << 17)
+#       define MRDCKB0_RESET                            (1 << 18)
+#       define MRDCKB1_RESET                            (1 << 19)
+#       define MRDCKC0_RESET                            (1 << 20)
+#       define MRDCKC1_RESET                            (1 << 21)
+#       define MRDCKD0_RESET                            (1 << 22)
+#       define MRDCKD1_RESET                            (1 << 23)
+#       define DLL_READY_READ                           (1 << 24)
+#       define USE_DISPLAY_GAP                          (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                (1 << 26)
+#       define MPLL_TURNOFF_D2                          (1 << 28)
+#define	DLL_CNTL					0x64c
+#       define MRDCKA0_BYPASS                           (1 << 24)
+#       define MRDCKA1_BYPASS                           (1 << 25)
+#       define MRDCKB0_BYPASS                           (1 << 26)
+#       define MRDCKB1_BYPASS                           (1 << 27)
+#       define MRDCKC0_BYPASS                           (1 << 28)
+#       define MRDCKC1_BYPASS                           (1 << 29)
+#       define MRDCKD0_BYPASS                           (1 << 30)
+#       define MRDCKD1_BYPASS                           (1 << 31)
+
+#define CG_AT                                           0x6d4
+#       define CG_R(x)					((x) << 0)
+#       define CG_R_MASK				(0xffff << 0)
+#       define CG_L(x)					((x) << 16)
+#       define CG_L_MASK				(0xffff << 16)
+
+#define CG_DISPLAY_GAP_CNTL                               0x714
+#       define DISP1_GAP(x)                               ((x) << 0)
+#       define DISP1_GAP_MASK                             (3 << 0)
+#       define DISP2_GAP(x)                               ((x) << 2)
+#       define DISP2_GAP_MASK                             (3 << 2)
+#       define VBI_TIMER_COUNT(x)                         ((x) << 4)
+#       define VBI_TIMER_COUNT_MASK                       (0x3fff << 4)
+#       define VBI_TIMER_UNIT(x)                          ((x) << 20)
+#       define VBI_TIMER_UNIT_MASK                        (7 << 20)
+#       define DISP1_GAP_MCHG(x)                          ((x) << 24)
+#       define DISP1_GAP_MCHG_MASK                        (3 << 24)
+#       define DISP2_GAP_MCHG(x)                          ((x) << 26)
+#       define DISP2_GAP_MCHG_MASK                        (3 << 26)
+
+#define	CG_BIF_REQ_AND_RSP				0x7f4
+#define		CG_CLIENT_REQ(x)			((x) << 0)
+#define		CG_CLIENT_REQ_MASK			(0xff << 0)
+#define		CG_CLIENT_REQ_SHIFT			0
+#define		CG_CLIENT_RESP(x)			((x) << 8)
+#define		CG_CLIENT_RESP_MASK			(0xff << 8)
+#define		CG_CLIENT_RESP_SHIFT			8
+#define		CLIENT_CG_REQ(x)			((x) << 16)
+#define		CLIENT_CG_REQ_MASK			(0xff << 16)
+#define		CLIENT_CG_REQ_SHIFT			16
+#define		CLIENT_CG_RESP(x)			((x) << 24)
+#define		CLIENT_CG_RESP_MASK			(0xff << 24)
+#define		CLIENT_CG_RESP_SHIFT			24
+
+#define	CG_SPLL_SPREAD_SPECTRUM				0x790
+#define		SSEN					(1 << 0)
+#define	CG_SPLL_SPREAD_SPECTRUM_2			0x794
+
+#define	MPLL_SS1					0x85c
+#define		CLKV(x)					((x) << 0)
+#define		CLKV_MASK				(0x3ffffff << 0)
+#define	MPLL_SS2					0x860
+#define		CLKS(x)					((x) << 0)
+#define		CLKS_MASK				(0xfff << 0)
+
+#define	CG_IND_ADDR					0x8f8
+#define	CG_IND_DATA					0x8fc
+/* CGIND regs */
+#define	CG_CGTT_LOCAL_0					0x00
+#define	CG_CGTT_LOCAL_1					0x01
+#define	CG_CGTT_LOCAL_2					0x02
+#define	CG_CGTT_LOCAL_3					0x03
+#define	CG_CGLS_TILE_0					0x20
+#define	CG_CGLS_TILE_1					0x21
+#define	CG_CGLS_TILE_2					0x22
+#define	CG_CGLS_TILE_3					0x23
+#define	CG_CGLS_TILE_4					0x24
+#define	CG_CGLS_TILE_5					0x25
+#define	CG_CGLS_TILE_6					0x26
+#define	CG_CGLS_TILE_7					0x27
+#define	CG_CGLS_TILE_8					0x28
+#define	CG_CGLS_TILE_9					0x29
+#define	CG_CGLS_TILE_10					0x2a
+#define	CG_CGLS_TILE_11					0x2b
+
+#define VM_L2_CG                                        0x15c0
+
+#define MC_CONFIG                                       0x2000
+
+#define MC_CONFIG_MCD                                   0x20a0
+#define MC_CG_CONFIG_MCD                                0x20a4
+#define		MC_RD_ENABLE_MCD(x)			((x) << 8)
+#define		MC_RD_ENABLE_MCD_MASK			(7 << 8)
+
+#define MC_HUB_MISC_HUB_CG                              0x20b8
+#define MC_HUB_MISC_VM_CG                               0x20bc
+#define MC_HUB_MISC_SIP_CG                              0x20c0
+
+#define MC_XPB_CLK_GAT                                  0x2478
+
+#define MC_CG_CONFIG                                    0x25bc
+#define		MC_RD_ENABLE(x)				((x) << 4)
+#define		MC_RD_ENABLE_MASK			(3 << 4)
+
+#define MC_CITF_MISC_RD_CG                              0x2648
+#define MC_CITF_MISC_WR_CG                              0x264c
+#define MC_CITF_MISC_VM_CG                              0x2650
+#       define MEM_LS_ENABLE                            (1 << 19)
+
+#define MC_ARB_BURST_TIME                               0x2808
+#define		STATE0(x)				((x) << 0)
+#define		STATE0_MASK				(0x1f << 0)
+#define		STATE1(x)				((x) << 5)
+#define		STATE1_MASK				(0x1f << 5)
+#define		STATE2(x)				((x) << 10)
+#define		STATE2_MASK				(0x1f << 10)
+#define		STATE3(x)				((x) << 15)
+#define		STATE3_MASK				(0x1f << 15)
+
+#define MC_SEQ_RAS_TIMING                               0x28a0
+#define MC_SEQ_CAS_TIMING                               0x28a4
+#define MC_SEQ_MISC_TIMING                              0x28a8
+#define MC_SEQ_MISC_TIMING2                             0x28ac
+
+#define MC_SEQ_RD_CTL_D0                                0x28b4
+#define MC_SEQ_RD_CTL_D1                                0x28b8
+#define MC_SEQ_WR_CTL_D0                                0x28bc
+#define MC_SEQ_WR_CTL_D1                                0x28c0
+
+#define MC_SEQ_STATUS_M                                 0x29f4
+#       define PMG_PWRSTATE                             (1 << 16)
+
+#define MC_SEQ_MISC1                                    0x2a04
+#define MC_SEQ_RESERVE_M                                0x2a08
+#define MC_PMG_CMD_EMRS                                 0x2a0c
+
+#define MC_SEQ_MISC3                                    0x2a2c
+
+#define MC_SEQ_MISC5                                    0x2a54
+#define MC_SEQ_MISC6                                    0x2a58
+
+#define MC_SEQ_MISC7                                    0x2a64
+
+#define MC_SEQ_CG                                       0x2a68
+#define		CG_SEQ_REQ(x)				((x) << 0)
+#define		CG_SEQ_REQ_MASK				(0xff << 0)
+#define		CG_SEQ_REQ_SHIFT			0
+#define		CG_SEQ_RESP(x)				((x) << 8)
+#define		CG_SEQ_RESP_MASK			(0xff << 8)
+#define		CG_SEQ_RESP_SHIFT			8
+#define		SEQ_CG_REQ(x)				((x) << 16)
+#define		SEQ_CG_REQ_MASK				(0xff << 16)
+#define		SEQ_CG_REQ_SHIFT			16
+#define		SEQ_CG_RESP(x)				((x) << 24)
+#define		SEQ_CG_RESP_MASK			(0xff << 24)
+#define		SEQ_CG_RESP_SHIFT			24
+#define MC_SEQ_RAS_TIMING_LP                            0x2a6c
+#define MC_SEQ_CAS_TIMING_LP                            0x2a70
+#define MC_SEQ_MISC_TIMING_LP                           0x2a74
+#define MC_SEQ_MISC_TIMING2_LP                          0x2a78
+#define MC_SEQ_WR_CTL_D0_LP                             0x2a7c
+#define MC_SEQ_WR_CTL_D1_LP                             0x2a80
+#define MC_SEQ_PMG_CMD_EMRS_LP                          0x2a84
+#define MC_SEQ_PMG_CMD_MRS_LP                           0x2a88
+
+#define MC_PMG_CMD_MRS                                  0x2aac
+
+#define MC_SEQ_RD_CTL_D0_LP                             0x2b1c
+#define MC_SEQ_RD_CTL_D1_LP                             0x2b20
+
+#define MC_PMG_CMD_MRS1                                 0x2b44
+#define MC_SEQ_PMG_CMD_MRS1_LP                          0x2b48
+
+#define CGTS_SM_CTRL_REG                                0x9150
+
 /* Registers */
 
 #define RCU_IND_INDEX           			0x100
@@ -90,6 +377,34 @@
 #define CG_VCLK_STATUS                                  0x61c
 #define	CG_SCRATCH1					0x820
 
+#define RLC_CNTL                                        0x3f00
+#       define RLC_ENABLE                               (1 << 0)
+#       define GFX_POWER_GATING_ENABLE                  (1 << 7)
+#       define GFX_POWER_GATING_SRC                     (1 << 8)
+#       define DYN_PER_SIMD_PG_ENABLE                   (1 << 27)
+#       define LB_CNT_SPIM_ACTIVE                       (1 << 30)
+#       define LOAD_BALANCE_ENABLE                      (1 << 31)
+
+#define RLC_HB_BASE                                       0x3f10
+#define RLC_HB_CNTL                                       0x3f0c
+#define RLC_HB_RPTR                                       0x3f20
+#define RLC_HB_WPTR                                       0x3f1c
+#define RLC_HB_WPTR_LSB_ADDR                              0x3f14
+#define RLC_HB_WPTR_MSB_ADDR                              0x3f18
+#define RLC_MC_CNTL                                       0x3f44
+#define RLC_UCODE_CNTL                                    0x3f48
+#define RLC_UCODE_ADDR                                    0x3f2c
+#define RLC_UCODE_DATA                                    0x3f30
+
+/* new for TN */
+#define TN_RLC_SAVE_AND_RESTORE_BASE                      0x3f10
+#define TN_RLC_LB_CNTR_MAX                                0x3f14
+#define TN_RLC_LB_CNTR_INIT                               0x3f18
+#define TN_RLC_CLEAR_STATE_RESTORE_BASE                   0x3f20
+#define TN_RLC_LB_INIT_SIMD_MASK                          0x3fe4
+#define TN_RLC_LB_ALWAYS_ACTIVE_SIMD_MASK                 0x3fe8
+#define TN_RLC_LB_PARAMS                                  0x3fec
+
 #define GRBM_GFX_INDEX          			0x802C
 #define		INSTANCE_INDEX(x)			((x) << 0)
 #define		SE_INDEX(x)     			((x) << 16)
@@ -503,6 +818,30 @@
 #define	CG_THERMAL_CTRL					0x72c
 #define		TOFFSET_MASK			        0x00003FE0
 #define		TOFFSET_SHIFT			        5
+#define		DIG_THERM_DPM(x)			((x) << 14)
+#define		DIG_THERM_DPM_MASK			0x003FC000
+#define		DIG_THERM_DPM_SHIFT			14
+
+#define	CG_THERMAL_INT					0x734
+#define		DIG_THERM_INTH(x)			((x) << 8)
+#define		DIG_THERM_INTH_MASK			0x0000FF00
+#define		DIG_THERM_INTH_SHIFT			8
+#define		DIG_THERM_INTL(x)			((x) << 16)
+#define		DIG_THERM_INTL_MASK			0x00FF0000
+#define		DIG_THERM_INTL_SHIFT			16
+#define 	THERM_INT_MASK_HIGH			(1 << 24)
+#define 	THERM_INT_MASK_LOW			(1 << 25)
+
+#define	TN_CG_THERMAL_INT_CTRL				0x738
+#define		TN_DIG_THERM_INTH(x)			((x) << 0)
+#define		TN_DIG_THERM_INTH_MASK			0x000000FF
+#define		TN_DIG_THERM_INTH_SHIFT			0
+#define		TN_DIG_THERM_INTL(x)			((x) << 8)
+#define		TN_DIG_THERM_INTL_MASK			0x0000FF00
+#define		TN_DIG_THERM_INTL_SHIFT			8
+#define 	TN_THERM_INT_MASK_HIGH			(1 << 24)
+#define 	TN_THERM_INT_MASK_LOW			(1 << 25)
+
 #define	CG_MULT_THERMAL_STATUS				0x740
 #define		ASIC_T(x)			        ((x) << 16)
 #define		ASIC_T_MASK			        0x07FF0000
@@ -510,6 +849,7 @@
 #define	CG_TS0_STATUS					0x760
 #define		TS0_ADC_DOUT_MASK			0x000003FF
 #define		TS0_ADC_DOUT_SHIFT			0
+
 /* APU */
 #define	CG_THERMAL_STATUS			        0x678
 
@@ -992,7 +1332,48 @@
 #define	DMA_PACKET_CONSTANT_FILL                0xd
 #define	DMA_PACKET_NOP                          0xf
 
-/* PCIE link stuff */
+/* PIF PHY0 indirect regs */
+#define PB0_PIF_CNTL                                      0x10
+#       define LS2_EXIT_TIME(x)                           ((x) << 17)
+#       define LS2_EXIT_TIME_MASK                         (0x7 << 17)
+#       define LS2_EXIT_TIME_SHIFT                        17
+#define PB0_PIF_PAIRING                                   0x11
+#       define MULTI_PIF                                  (1 << 25)
+#define PB0_PIF_PWRDOWN_0                                 0x12
+#       define PLL_POWER_STATE_IN_TXS2_0(x)               ((x) << 7)
+#       define PLL_POWER_STATE_IN_TXS2_0_MASK             (0x7 << 7)
+#       define PLL_POWER_STATE_IN_TXS2_0_SHIFT            7
+#       define PLL_POWER_STATE_IN_OFF_0(x)                ((x) << 10)
+#       define PLL_POWER_STATE_IN_OFF_0_MASK              (0x7 << 10)
+#       define PLL_POWER_STATE_IN_OFF_0_SHIFT             10
+#       define PLL_RAMP_UP_TIME_0(x)                      ((x) << 24)
+#       define PLL_RAMP_UP_TIME_0_MASK                    (0x7 << 24)
+#       define PLL_RAMP_UP_TIME_0_SHIFT                   24
+#define PB0_PIF_PWRDOWN_1                                 0x13
+#       define PLL_POWER_STATE_IN_TXS2_1(x)               ((x) << 7)
+#       define PLL_POWER_STATE_IN_TXS2_1_MASK             (0x7 << 7)
+#       define PLL_POWER_STATE_IN_TXS2_1_SHIFT            7
+#       define PLL_POWER_STATE_IN_OFF_1(x)                ((x) << 10)
+#       define PLL_POWER_STATE_IN_OFF_1_MASK              (0x7 << 10)
+#       define PLL_POWER_STATE_IN_OFF_1_SHIFT             10
+#       define PLL_RAMP_UP_TIME_1(x)                      ((x) << 24)
+#       define PLL_RAMP_UP_TIME_1_MASK                    (0x7 << 24)
+#       define PLL_RAMP_UP_TIME_1_SHIFT                   24
+/* PIF PHY1 indirect regs */
+#define PB1_PIF_CNTL                                      0x10
+#define PB1_PIF_PAIRING                                   0x11
+#define PB1_PIF_PWRDOWN_0                                 0x12
+#define PB1_PIF_PWRDOWN_1                                 0x13
+/* PCIE PORT indirect regs */
+#define PCIE_LC_CNTL                                      0xa0
+#       define LC_L0S_INACTIVITY(x)                       ((x) << 8)
+#       define LC_L0S_INACTIVITY_MASK                     (0xf << 8)
+#       define LC_L0S_INACTIVITY_SHIFT                    8
+#       define LC_L1_INACTIVITY(x)                        ((x) << 12)
+#       define LC_L1_INACTIVITY_MASK                      (0xf << 12)
+#       define LC_L1_INACTIVITY_SHIFT                     12
+#       define LC_PMI_TO_L1_DIS                           (1 << 16)
+#       define LC_ASPM_TO_L1_DIS                          (1 << 24)
 #define PCIE_LC_TRAINING_CNTL                             0xa1 /* PCIE_P */
 #define PCIE_LC_LINK_WIDTH_CNTL                           0xa2 /* PCIE_P */
 #       define LC_LINK_WIDTH_SHIFT                        0
@@ -1012,6 +1393,9 @@
 #       define LC_SHORT_RECONFIG_EN                       (1 << 11)
 #       define LC_UPCONFIGURE_SUPPORT                     (1 << 12)
 #       define LC_UPCONFIGURE_DIS                         (1 << 13)
+#       define LC_DYN_LANES_PWR_STATE(x)                  ((x) << 21)
+#       define LC_DYN_LANES_PWR_STATE_MASK                (0x3 << 21)
+#       define LC_DYN_LANES_PWR_STATE_SHIFT               21
 #define PCIE_LC_SPEED_CNTL                                0xa4 /* PCIE_P */
 #       define LC_GEN2_EN_STRAP                           (1 << 0)
 #       define LC_TARGET_LINK_SPEED_OVERRIDE_EN           (1 << 1)
@@ -1020,6 +1404,9 @@
 #       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK      (0x3 << 8)
 #       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT     3
 #       define LC_CURRENT_DATA_RATE                       (1 << 11)
+#       define LC_HW_VOLTAGE_IF_CONTROL(x)                ((x) << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_MASK              (3 << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_SHIFT             12
 #       define LC_VOLTAGE_TIMER_SEL_MASK                  (0xf << 14)
 #       define LC_CLR_FAILED_SPD_CHANGE_CNT               (1 << 21)
 #       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 23)
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 84583302b081..56bd4f3be4fe 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -22,7 +22,6 @@
  * Authors: Alex Deucher
  */
 #include <linux/firmware.h>
-#include <linux/platform_device.h>
 #include <linux/slab.h>
 #include <linux/module.h>
 #include <drm/drmP.h>
@@ -33,6 +32,135 @@
 #include "atom.h"
 #include "ni_reg.h"
 #include "cayman_blit_shaders.h"
+#include "radeon_ucode.h"
+#include "clearstate_cayman.h"
+
+static u32 tn_rlc_save_restore_register_list[] =
+{
+	0x98fc,
+	0x98f0,
+	0x9834,
+	0x9838,
+	0x9870,
+	0x9874,
+	0x8a14,
+	0x8b24,
+	0x8bcc,
+	0x8b10,
+	0x8c30,
+	0x8d00,
+	0x8d04,
+	0x8c00,
+	0x8c04,
+	0x8c10,
+	0x8c14,
+	0x8d8c,
+	0x8cf0,
+	0x8e38,
+	0x9508,
+	0x9688,
+	0x9608,
+	0x960c,
+	0x9610,
+	0x9614,
+	0x88c4,
+	0x8978,
+	0x88d4,
+	0x900c,
+	0x9100,
+	0x913c,
+	0x90e8,
+	0x9354,
+	0xa008,
+	0x98f8,
+	0x9148,
+	0x914c,
+	0x3f94,
+	0x98f4,
+	0x9b7c,
+	0x3f8c,
+	0x8950,
+	0x8954,
+	0x8a18,
+	0x8b28,
+	0x9144,
+	0x3f90,
+	0x915c,
+	0x9160,
+	0x9178,
+	0x917c,
+	0x9180,
+	0x918c,
+	0x9190,
+	0x9194,
+	0x9198,
+	0x919c,
+	0x91a8,
+	0x91ac,
+	0x91b0,
+	0x91b4,
+	0x91b8,
+	0x91c4,
+	0x91c8,
+	0x91cc,
+	0x91d0,
+	0x91d4,
+	0x91e0,
+	0x91e4,
+	0x91ec,
+	0x91f0,
+	0x91f4,
+	0x9200,
+	0x9204,
+	0x929c,
+	0x8030,
+	0x9150,
+	0x9a60,
+	0x920c,
+	0x9210,
+	0x9228,
+	0x922c,
+	0x9244,
+	0x9248,
+	0x91e8,
+	0x9294,
+	0x9208,
+	0x9224,
+	0x9240,
+	0x9220,
+	0x923c,
+	0x9258,
+	0x9744,
+	0xa200,
+	0xa204,
+	0xa208,
+	0xa20c,
+	0x8d58,
+	0x9030,
+	0x9034,
+	0x9038,
+	0x903c,
+	0x9040,
+	0x9654,
+	0x897c,
+	0xa210,
+	0xa214,
+	0x9868,
+	0xa02c,
+	0x9664,
+	0x9698,
+	0x949c,
+	0x8e10,
+	0x8e18,
+	0x8c50,
+	0x8c58,
+	0x8c60,
+	0x8c68,
+	0x89b4,
+	0x9830,
+	0x802c,
+};
+static u32 tn_rlc_save_restore_register_list_size = ARRAY_SIZE(tn_rlc_save_restore_register_list);
 
 extern bool evergreen_is_display_hung(struct radeon_device *rdev);
 extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev);
@@ -44,36 +172,29 @@ extern void evergreen_irq_suspend(struct radeon_device *rdev);
 extern int evergreen_mc_init(struct radeon_device *rdev);
 extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev);
 extern void evergreen_pcie_gen2_enable(struct radeon_device *rdev);
-extern void si_rlc_fini(struct radeon_device *rdev);
-extern int si_rlc_init(struct radeon_device *rdev);
-
-#define EVERGREEN_PFP_UCODE_SIZE 1120
-#define EVERGREEN_PM4_UCODE_SIZE 1376
-#define EVERGREEN_RLC_UCODE_SIZE 768
-#define BTC_MC_UCODE_SIZE 6024
-
-#define CAYMAN_PFP_UCODE_SIZE 2176
-#define CAYMAN_PM4_UCODE_SIZE 2176
-#define CAYMAN_RLC_UCODE_SIZE 1024
-#define CAYMAN_MC_UCODE_SIZE 6037
-
-#define ARUBA_RLC_UCODE_SIZE 1536
+extern void evergreen_program_aspm(struct radeon_device *rdev);
+extern void sumo_rlc_fini(struct radeon_device *rdev);
+extern int sumo_rlc_init(struct radeon_device *rdev);
 
 /* Firmware Names */
 MODULE_FIRMWARE("radeon/BARTS_pfp.bin");
 MODULE_FIRMWARE("radeon/BARTS_me.bin");
 MODULE_FIRMWARE("radeon/BARTS_mc.bin");
+MODULE_FIRMWARE("radeon/BARTS_smc.bin");
 MODULE_FIRMWARE("radeon/BTC_rlc.bin");
 MODULE_FIRMWARE("radeon/TURKS_pfp.bin");
 MODULE_FIRMWARE("radeon/TURKS_me.bin");
 MODULE_FIRMWARE("radeon/TURKS_mc.bin");
+MODULE_FIRMWARE("radeon/TURKS_smc.bin");
 MODULE_FIRMWARE("radeon/CAICOS_pfp.bin");
 MODULE_FIRMWARE("radeon/CAICOS_me.bin");
 MODULE_FIRMWARE("radeon/CAICOS_mc.bin");
+MODULE_FIRMWARE("radeon/CAICOS_smc.bin");
 MODULE_FIRMWARE("radeon/CAYMAN_pfp.bin");
 MODULE_FIRMWARE("radeon/CAYMAN_me.bin");
 MODULE_FIRMWARE("radeon/CAYMAN_mc.bin");
 MODULE_FIRMWARE("radeon/CAYMAN_rlc.bin");
+MODULE_FIRMWARE("radeon/CAYMAN_smc.bin");
 MODULE_FIRMWARE("radeon/ARUBA_pfp.bin");
 MODULE_FIRMWARE("radeon/ARUBA_me.bin");
 MODULE_FIRMWARE("radeon/ARUBA_rlc.bin");
@@ -562,22 +683,15 @@ int ni_mc_load_microcode(struct radeon_device *rdev)
 
 int ni_init_microcode(struct radeon_device *rdev)
 {
-	struct platform_device *pdev;
 	const char *chip_name;
 	const char *rlc_chip_name;
 	size_t pfp_req_size, me_req_size, rlc_req_size, mc_req_size;
+	size_t smc_req_size = 0;
 	char fw_name[30];
 	int err;
 
 	DRM_DEBUG("\n");
 
-	pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0);
-	err = IS_ERR(pdev);
-	if (err) {
-		printk(KERN_ERR "radeon_cp: Failed to register firmware\n");
-		return -EINVAL;
-	}
-
 	switch (rdev->family) {
 	case CHIP_BARTS:
 		chip_name = "BARTS";
@@ -586,6 +700,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 		me_req_size = EVERGREEN_PM4_UCODE_SIZE * 4;
 		rlc_req_size = EVERGREEN_RLC_UCODE_SIZE * 4;
 		mc_req_size = BTC_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(BARTS_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_TURKS:
 		chip_name = "TURKS";
@@ -594,6 +709,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 		me_req_size = EVERGREEN_PM4_UCODE_SIZE * 4;
 		rlc_req_size = EVERGREEN_RLC_UCODE_SIZE * 4;
 		mc_req_size = BTC_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(TURKS_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_CAICOS:
 		chip_name = "CAICOS";
@@ -602,6 +718,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 		me_req_size = EVERGREEN_PM4_UCODE_SIZE * 4;
 		rlc_req_size = EVERGREEN_RLC_UCODE_SIZE * 4;
 		mc_req_size = BTC_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(CAICOS_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_CAYMAN:
 		chip_name = "CAYMAN";
@@ -610,6 +727,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 		me_req_size = CAYMAN_PM4_UCODE_SIZE * 4;
 		rlc_req_size = CAYMAN_RLC_UCODE_SIZE * 4;
 		mc_req_size = CAYMAN_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(CAYMAN_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_ARUBA:
 		chip_name = "ARUBA";
@@ -626,7 +744,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 	DRM_INFO("Loading %s Microcode\n", chip_name);
 
 	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
-	err = request_firmware(&rdev->pfp_fw, fw_name, &pdev->dev);
+	err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev);
 	if (err)
 		goto out;
 	if (rdev->pfp_fw->size != pfp_req_size) {
@@ -638,7 +756,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 	}
 
 	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
-	err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev);
+	err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
 	if (err)
 		goto out;
 	if (rdev->me_fw->size != me_req_size) {
@@ -649,7 +767,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 	}
 
 	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", rlc_chip_name);
-	err = request_firmware(&rdev->rlc_fw, fw_name, &pdev->dev);
+	err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev);
 	if (err)
 		goto out;
 	if (rdev->rlc_fw->size != rlc_req_size) {
@@ -662,7 +780,7 @@ int ni_init_microcode(struct radeon_device *rdev)
 	/* no MC ucode on TN */
 	if (!(rdev->flags & RADEON_IS_IGP)) {
 		snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
-		err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev);
+		err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
 		if (err)
 			goto out;
 		if (rdev->mc_fw->size != mc_req_size) {
@@ -672,9 +790,21 @@ int ni_init_microcode(struct radeon_device *rdev)
 			err = -EINVAL;
 		}
 	}
-out:
-	platform_device_unregister(pdev);
 
+	if ((rdev->family >= CHIP_BARTS) && (rdev->family <= CHIP_CAYMAN)) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
+		err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->smc_fw->size != smc_req_size) {
+			printk(KERN_ERR
+			       "ni_mc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->mc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	}
+
+out:
 	if (err) {
 		if (err != -EINVAL)
 			printk(KERN_ERR
@@ -692,6 +822,14 @@ out:
 	return err;
 }
 
+int tn_get_temp(struct radeon_device *rdev)
+{
+	u32 temp = RREG32_SMC(TN_CURRENT_GNB_TEMP) & 0x7ff;
+	int actual_temp = (temp / 8) - 49;
+
+	return actual_temp * 1000;
+}
+
 /*
  * Core functions
  */
@@ -1027,6 +1165,16 @@ static void cayman_gpu_init(struct radeon_device *rdev)
 	WREG32(PA_CL_ENHANCE, CLIP_VTX_REORDER_ENA | NUM_CLIP_SEQ(3));
 
 	udelay(50);
+
+	/* set clockgating golden values on TN */
+	if (rdev->family == CHIP_ARUBA) {
+		tmp = RREG32_CG(CG_CGTT_LOCAL_0);
+		tmp &= ~0x00380000;
+		WREG32_CG(CG_CGTT_LOCAL_0, tmp);
+                tmp = RREG32_CG(CG_CGTT_LOCAL_1);
+		tmp &= ~0x0e000000;
+		WREG32_CG(CG_CGTT_LOCAL_1, tmp);
+	}
 }
 
 /*
@@ -1928,6 +2076,8 @@ static int cayman_startup(struct radeon_device *rdev)
 
 	/* enable pcie gen2 link */
 	evergreen_pcie_gen2_enable(rdev);
+	/* enable aspm */
+	evergreen_program_aspm(rdev);
 
 	if (rdev->flags & RADEON_IS_IGP) {
 		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw) {
@@ -1972,7 +2122,10 @@ static int cayman_startup(struct radeon_device *rdev)
 
 	/* allocate rlc buffers */
 	if (rdev->flags & RADEON_IS_IGP) {
-		r = si_rlc_init(rdev);
+		rdev->rlc.reg_list = tn_rlc_save_restore_register_list;
+		rdev->rlc.reg_list_size = tn_rlc_save_restore_register_list_size;
+		rdev->rlc.cs_data = cayman_cs_data;
+		r = sumo_rlc_init(rdev);
 		if (r) {
 			DRM_ERROR("Failed to init rlc BOs!\n");
 			return r;
@@ -2229,7 +2382,7 @@ int cayman_init(struct radeon_device *rdev)
 		cayman_dma_fini(rdev);
 		r600_irq_fini(rdev);
 		if (rdev->flags & RADEON_IS_IGP)
-			si_rlc_fini(rdev);
+			sumo_rlc_fini(rdev);
 		radeon_wb_fini(rdev);
 		radeon_ib_pool_fini(rdev);
 		radeon_vm_manager_fini(rdev);
@@ -2260,7 +2413,7 @@ void cayman_fini(struct radeon_device *rdev)
 	cayman_dma_fini(rdev);
 	r600_irq_fini(rdev);
 	if (rdev->flags & RADEON_IS_IGP)
-		si_rlc_fini(rdev);
+		sumo_rlc_fini(rdev);
 	radeon_wb_fini(rdev);
 	radeon_vm_manager_fini(rdev);
 	radeon_ib_pool_fini(rdev);
@@ -2297,6 +2450,167 @@ void cayman_vm_fini(struct radeon_device *rdev)
 {
 }
 
+/**
+ * cayman_vm_decode_fault - print human readable fault info
+ *
+ * @rdev: radeon_device pointer
+ * @status: VM_CONTEXT1_PROTECTION_FAULT_STATUS register value
+ * @addr: VM_CONTEXT1_PROTECTION_FAULT_ADDR register value
+ *
+ * Print human readable fault information (cayman/TN).
+ */
+void cayman_vm_decode_fault(struct radeon_device *rdev,
+			    u32 status, u32 addr)
+{
+	u32 mc_id = (status & MEMORY_CLIENT_ID_MASK) >> MEMORY_CLIENT_ID_SHIFT;
+	u32 vmid = (status & FAULT_VMID_MASK) >> FAULT_VMID_SHIFT;
+	u32 protections = (status & PROTECTIONS_MASK) >> PROTECTIONS_SHIFT;
+	char *block;
+
+	switch (mc_id) {
+	case 32:
+	case 16:
+	case 96:
+	case 80:
+	case 160:
+	case 144:
+	case 224:
+	case 208:
+		block = "CB";
+		break;
+	case 33:
+	case 17:
+	case 97:
+	case 81:
+	case 161:
+	case 145:
+	case 225:
+	case 209:
+		block = "CB_FMASK";
+		break;
+	case 34:
+	case 18:
+	case 98:
+	case 82:
+	case 162:
+	case 146:
+	case 226:
+	case 210:
+		block = "CB_CMASK";
+		break;
+	case 35:
+	case 19:
+	case 99:
+	case 83:
+	case 163:
+	case 147:
+	case 227:
+	case 211:
+		block = "CB_IMMED";
+		break;
+	case 36:
+	case 20:
+	case 100:
+	case 84:
+	case 164:
+	case 148:
+	case 228:
+	case 212:
+		block = "DB";
+		break;
+	case 37:
+	case 21:
+	case 101:
+	case 85:
+	case 165:
+	case 149:
+	case 229:
+	case 213:
+		block = "DB_HTILE";
+		break;
+	case 38:
+	case 22:
+	case 102:
+	case 86:
+	case 166:
+	case 150:
+	case 230:
+	case 214:
+		block = "SX";
+		break;
+	case 39:
+	case 23:
+	case 103:
+	case 87:
+	case 167:
+	case 151:
+	case 231:
+	case 215:
+		block = "DB_STEN";
+		break;
+	case 40:
+	case 24:
+	case 104:
+	case 88:
+	case 232:
+	case 216:
+	case 168:
+	case 152:
+		block = "TC_TFETCH";
+		break;
+	case 41:
+	case 25:
+	case 105:
+	case 89:
+	case 233:
+	case 217:
+	case 169:
+	case 153:
+		block = "TC_VFETCH";
+		break;
+	case 42:
+	case 26:
+	case 106:
+	case 90:
+	case 234:
+	case 218:
+	case 170:
+	case 154:
+		block = "VC";
+		break;
+	case 112:
+		block = "CP";
+		break;
+	case 113:
+	case 114:
+		block = "SH";
+		break;
+	case 115:
+		block = "VGT";
+		break;
+	case 178:
+		block = "IH";
+		break;
+	case 51:
+		block = "RLC";
+		break;
+	case 55:
+		block = "DMA";
+		break;
+	case 56:
+		block = "HDP";
+		break;
+	default:
+		block = "unknown";
+		break;
+	}
+
+	printk("VM fault (0x%02x, vmid %d) at page %u, %s from %s (%d)\n",
+	       protections, vmid, addr,
+	       (status & MEMORY_CLIENT_RW_MASK) ? "write" : "read",
+	       block, mc_id);
+}
+
 #define R600_ENTRY_VALID   (1 << 0)
 #define R600_PTE_SYSTEM    (1 << 1)
 #define R600_PTE_SNOOPED   (1 << 2)
diff --git a/drivers/gpu/drm/radeon/ni_dpm.c b/drivers/gpu/drm/radeon/ni_dpm.c
new file mode 100644
index 000000000000..559cf24d51af
--- /dev/null
+++ b/drivers/gpu/drm/radeon/ni_dpm.c
@@ -0,0 +1,4370 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "nid.h"
+#include "r600_dpm.h"
+#include "ni_dpm.h"
+#include "atom.h"
+#include <linux/math64.h>
+#include <linux/seq_file.h>
+
+#define MC_CG_ARB_FREQ_F0           0x0a
+#define MC_CG_ARB_FREQ_F1           0x0b
+#define MC_CG_ARB_FREQ_F2           0x0c
+#define MC_CG_ARB_FREQ_F3           0x0d
+
+#define SMC_RAM_END 0xC000
+
+static const struct ni_cac_weights cac_weights_cayman_xt =
+{
+	0x15,
+	0x2,
+	0x19,
+	0x2,
+	0x8,
+	0x14,
+	0x2,
+	0x16,
+	0xE,
+	0x17,
+	0x13,
+	0x2B,
+	0x10,
+	0x7,
+	0x5,
+	0x5,
+	0x5,
+	0x2,
+	0x3,
+	0x9,
+	0x10,
+	0x10,
+	0x2B,
+	0xA,
+	0x9,
+	0x4,
+	0xD,
+	0xD,
+	0x3E,
+	0x18,
+	0x14,
+	0,
+	0x3,
+	0x3,
+	0x5,
+	0,
+	0x2,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0x1CC,
+	0,
+	0x164,
+	1,
+	1,
+	1,
+	1,
+	12,
+	12,
+	12,
+	0x12,
+	0x1F,
+	132,
+	5,
+	7,
+	0,
+	{ 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0 },
+	true
+};
+
+static const struct ni_cac_weights cac_weights_cayman_pro =
+{
+	0x16,
+	0x4,
+	0x10,
+	0x2,
+	0xA,
+	0x16,
+	0x2,
+	0x18,
+	0x10,
+	0x1A,
+	0x16,
+	0x2D,
+	0x12,
+	0xA,
+	0x6,
+	0x6,
+	0x6,
+	0x2,
+	0x4,
+	0xB,
+	0x11,
+	0x11,
+	0x2D,
+	0xC,
+	0xC,
+	0x7,
+	0x10,
+	0x10,
+	0x3F,
+	0x1A,
+	0x16,
+	0,
+	0x7,
+	0x4,
+	0x6,
+	1,
+	0x2,
+	0x1,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0x30,
+	0,
+	0x1CF,
+	0,
+	0x166,
+	1,
+	1,
+	1,
+	1,
+	12,
+	12,
+	12,
+	0x15,
+	0x1F,
+	132,
+	6,
+	6,
+	0,
+	{ 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0 },
+	true
+};
+
+static const struct ni_cac_weights cac_weights_cayman_le =
+{
+	0x7,
+	0xE,
+	0x1,
+	0xA,
+	0x1,
+	0x3F,
+	0x2,
+	0x18,
+	0x10,
+	0x1A,
+	0x1,
+	0x3F,
+	0x1,
+	0xE,
+	0x6,
+	0x6,
+	0x6,
+	0x2,
+	0x4,
+	0x9,
+	0x1A,
+	0x1A,
+	0x2C,
+	0xA,
+	0x11,
+	0x8,
+	0x19,
+	0x19,
+	0x1,
+	0x1,
+	0x1A,
+	0,
+	0x8,
+	0x5,
+	0x8,
+	0x1,
+	0x3,
+	0x1,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0x38,
+	0x38,
+	0x239,
+	0x3,
+	0x18A,
+	1,
+	1,
+	1,
+	1,
+	12,
+	12,
+	12,
+	0x15,
+	0x22,
+	132,
+	6,
+	6,
+	0,
+	{ 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0 },
+	true
+};
+
+#define NISLANDS_MGCG_SEQUENCE  300
+
+static const u32 cayman_cgcg_cgls_default[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define CAYMAN_CGCG_CGLS_DEFAULT_LENGTH sizeof(cayman_cgcg_cgls_default) / (3 * sizeof(u32))
+
+static const u32 cayman_cgcg_cgls_disable[] =
+{
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00000644, 0x000f7902, 0x001f4180,
+	0x00000644, 0x000f3802, 0x001f4180
+};
+#define CAYMAN_CGCG_CGLS_DISABLE_LENGTH sizeof(cayman_cgcg_cgls_disable) / (3 * sizeof(u32))
+
+static const u32 cayman_cgcg_cgls_enable[] =
+{
+	0x00000644, 0x000f7882, 0x001f4080,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000020, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000021, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000022, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000023, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000024, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000025, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000026, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000027, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000028, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000029, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002a, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x0000002b, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff
+};
+#define CAYMAN_CGCG_CGLS_ENABLE_LENGTH  sizeof(cayman_cgcg_cgls_enable) / (3 * sizeof(u32))
+
+static const u32 cayman_mgcg_default[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00003fc4, 0xc0000000, 0xffffffff,
+	0x00005448, 0x00000100, 0xffffffff,
+	0x000055e4, 0x00000100, 0xffffffff,
+	0x0000160c, 0x00000100, 0xffffffff,
+	0x00008984, 0x06000100, 0xffffffff,
+	0x0000c164, 0x00000100, 0xffffffff,
+	0x00008a18, 0x00000100, 0xffffffff,
+	0x0000897c, 0x06000100, 0xffffffff,
+	0x00008b28, 0x00000100, 0xffffffff,
+	0x00009144, 0x00800200, 0xffffffff,
+	0x00009a60, 0x00000100, 0xffffffff,
+	0x00009868, 0x00000100, 0xffffffff,
+	0x00008d58, 0x00000100, 0xffffffff,
+	0x00009510, 0x00000100, 0xffffffff,
+	0x0000949c, 0x00000100, 0xffffffff,
+	0x00009654, 0x00000100, 0xffffffff,
+	0x00009030, 0x00000100, 0xffffffff,
+	0x00009034, 0x00000100, 0xffffffff,
+	0x00009038, 0x00000100, 0xffffffff,
+	0x0000903c, 0x00000100, 0xffffffff,
+	0x00009040, 0x00000100, 0xffffffff,
+	0x0000a200, 0x00000100, 0xffffffff,
+	0x0000a204, 0x00000100, 0xffffffff,
+	0x0000a208, 0x00000100, 0xffffffff,
+	0x0000a20c, 0x00000100, 0xffffffff,
+	0x00009744, 0x00000100, 0xffffffff,
+	0x00003f80, 0x00000100, 0xffffffff,
+	0x0000a210, 0x00000100, 0xffffffff,
+	0x0000a214, 0x00000100, 0xffffffff,
+	0x000004d8, 0x00000100, 0xffffffff,
+	0x00009664, 0x00000100, 0xffffffff,
+	0x00009698, 0x00000100, 0xffffffff,
+	0x000004d4, 0x00000200, 0xffffffff,
+	0x000004d0, 0x00000000, 0xffffffff,
+	0x000030cc, 0x00000104, 0xffffffff,
+	0x0000d0c0, 0x00000100, 0xffffffff,
+	0x0000d8c0, 0x00000100, 0xffffffff,
+	0x0000802c, 0x40000000, 0xffffffff,
+	0x00003fc4, 0x40000000, 0xffffffff,
+	0x0000915c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091b0, 0x00070000, 0xffffffff,
+	0x000091b4, 0x00030002, 0xffffffff,
+	0x000091b8, 0x00050004, 0xffffffff,
+	0x000091c4, 0x00010006, 0xffffffff,
+	0x000091c8, 0x00090008, 0xffffffff,
+	0x000091cc, 0x00070000, 0xffffffff,
+	0x000091d0, 0x00030002, 0xffffffff,
+	0x000091d4, 0x00050004, 0xffffffff,
+	0x000091e0, 0x00010006, 0xffffffff,
+	0x000091e4, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x000091ec, 0x00070000, 0xffffffff,
+	0x000091f0, 0x00030002, 0xffffffff,
+	0x000091f4, 0x00050004, 0xffffffff,
+	0x00009200, 0x00010006, 0xffffffff,
+	0x00009204, 0x00090008, 0xffffffff,
+	0x00009208, 0x00070000, 0xffffffff,
+	0x0000920c, 0x00030002, 0xffffffff,
+	0x00009210, 0x00050004, 0xffffffff,
+	0x0000921c, 0x00010006, 0xffffffff,
+	0x00009220, 0x00090008, 0xffffffff,
+	0x00009224, 0x00070000, 0xffffffff,
+	0x00009228, 0x00030002, 0xffffffff,
+	0x0000922c, 0x00050004, 0xffffffff,
+	0x00009238, 0x00010006, 0xffffffff,
+	0x0000923c, 0x00090008, 0xffffffff,
+	0x00009240, 0x00070000, 0xffffffff,
+	0x00009244, 0x00030002, 0xffffffff,
+	0x00009248, 0x00050004, 0xffffffff,
+	0x00009254, 0x00010006, 0xffffffff,
+	0x00009258, 0x00090008, 0xffffffff,
+	0x0000925c, 0x00070000, 0xffffffff,
+	0x00009260, 0x00030002, 0xffffffff,
+	0x00009264, 0x00050004, 0xffffffff,
+	0x00009270, 0x00010006, 0xffffffff,
+	0x00009274, 0x00090008, 0xffffffff,
+	0x00009278, 0x00070000, 0xffffffff,
+	0x0000927c, 0x00030002, 0xffffffff,
+	0x00009280, 0x00050004, 0xffffffff,
+	0x0000928c, 0x00010006, 0xffffffff,
+	0x00009290, 0x00090008, 0xffffffff,
+	0x000092a8, 0x00070000, 0xffffffff,
+	0x000092ac, 0x00030002, 0xffffffff,
+	0x000092b0, 0x00050004, 0xffffffff,
+	0x000092bc, 0x00010006, 0xffffffff,
+	0x000092c0, 0x00090008, 0xffffffff,
+	0x000092c4, 0x00070000, 0xffffffff,
+	0x000092c8, 0x00030002, 0xffffffff,
+	0x000092cc, 0x00050004, 0xffffffff,
+	0x000092d8, 0x00010006, 0xffffffff,
+	0x000092dc, 0x00090008, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff,
+	0x0000802c, 0x40010000, 0xffffffff,
+	0x00003fc4, 0x40010000, 0xffffffff,
+	0x0000915c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091b0, 0x00070000, 0xffffffff,
+	0x000091b4, 0x00030002, 0xffffffff,
+	0x000091b8, 0x00050004, 0xffffffff,
+	0x000091c4, 0x00010006, 0xffffffff,
+	0x000091c8, 0x00090008, 0xffffffff,
+	0x000091cc, 0x00070000, 0xffffffff,
+	0x000091d0, 0x00030002, 0xffffffff,
+	0x000091d4, 0x00050004, 0xffffffff,
+	0x000091e0, 0x00010006, 0xffffffff,
+	0x000091e4, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x000091ec, 0x00070000, 0xffffffff,
+	0x000091f0, 0x00030002, 0xffffffff,
+	0x000091f4, 0x00050004, 0xffffffff,
+	0x00009200, 0x00010006, 0xffffffff,
+	0x00009204, 0x00090008, 0xffffffff,
+	0x00009208, 0x00070000, 0xffffffff,
+	0x0000920c, 0x00030002, 0xffffffff,
+	0x00009210, 0x00050004, 0xffffffff,
+	0x0000921c, 0x00010006, 0xffffffff,
+	0x00009220, 0x00090008, 0xffffffff,
+	0x00009224, 0x00070000, 0xffffffff,
+	0x00009228, 0x00030002, 0xffffffff,
+	0x0000922c, 0x00050004, 0xffffffff,
+	0x00009238, 0x00010006, 0xffffffff,
+	0x0000923c, 0x00090008, 0xffffffff,
+	0x00009240, 0x00070000, 0xffffffff,
+	0x00009244, 0x00030002, 0xffffffff,
+	0x00009248, 0x00050004, 0xffffffff,
+	0x00009254, 0x00010006, 0xffffffff,
+	0x00009258, 0x00090008, 0xffffffff,
+	0x0000925c, 0x00070000, 0xffffffff,
+	0x00009260, 0x00030002, 0xffffffff,
+	0x00009264, 0x00050004, 0xffffffff,
+	0x00009270, 0x00010006, 0xffffffff,
+	0x00009274, 0x00090008, 0xffffffff,
+	0x00009278, 0x00070000, 0xffffffff,
+	0x0000927c, 0x00030002, 0xffffffff,
+	0x00009280, 0x00050004, 0xffffffff,
+	0x0000928c, 0x00010006, 0xffffffff,
+	0x00009290, 0x00090008, 0xffffffff,
+	0x000092a8, 0x00070000, 0xffffffff,
+	0x000092ac, 0x00030002, 0xffffffff,
+	0x000092b0, 0x00050004, 0xffffffff,
+	0x000092bc, 0x00010006, 0xffffffff,
+	0x000092c0, 0x00090008, 0xffffffff,
+	0x000092c4, 0x00070000, 0xffffffff,
+	0x000092c8, 0x00030002, 0xffffffff,
+	0x000092cc, 0x00050004, 0xffffffff,
+	0x000092d8, 0x00010006, 0xffffffff,
+	0x000092dc, 0x00090008, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff,
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00003fc4, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000010, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000011, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000012, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000013, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000014, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000015, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000016, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000017, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000018, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000019, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001a, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x0000001b, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff
+};
+#define CAYMAN_MGCG_DEFAULT_LENGTH sizeof(cayman_mgcg_default) / (3 * sizeof(u32))
+
+static const u32 cayman_mgcg_disable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xffffffff,
+	0x00009150, 0x00600000, 0xffffffff
+};
+#define CAYMAN_MGCG_DISABLE_LENGTH   sizeof(cayman_mgcg_disable) / (3 * sizeof(u32))
+
+static const u32 cayman_mgcg_enable[] =
+{
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x000008f8, 0x00000002, 0xffffffff,
+	0x000008fc, 0x00600000, 0xffffffff,
+	0x000008f8, 0x00000003, 0xffffffff,
+	0x000008fc, 0x00000000, 0xffffffff,
+	0x00009150, 0x96944200, 0xffffffff
+};
+
+#define CAYMAN_MGCG_ENABLE_LENGTH   sizeof(cayman_mgcg_enable) / (3 * sizeof(u32))
+
+#define NISLANDS_SYSLS_SEQUENCE  100
+
+static const u32 cayman_sysls_default[] =
+{
+	/* Register,   Value,     Mask bits */
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x0000d8bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x00002f50, 0x00000404, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00008dfc, 0x00000000, 0xffffffff
+};
+#define CAYMAN_SYSLS_DEFAULT_LENGTH sizeof(cayman_sysls_default) / (3 * sizeof(u32))
+
+static const u32 cayman_sysls_disable[] =
+{
+	/* Register,   Value,     Mask bits */
+	0x0000d0c0, 0x00000000, 0xffffffff,
+	0x0000d8c0, 0x00000000, 0xffffffff,
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x0000d8bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x00041401, 0xffffffff,
+	0x0000264c, 0x00040400, 0xffffffff,
+	0x00002648, 0x00040400, 0xffffffff,
+	0x00002650, 0x00040400, 0xffffffff,
+	0x000020b8, 0x00040400, 0xffffffff,
+	0x000020bc, 0x00040400, 0xffffffff,
+	0x000020c0, 0x00040c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680000, 0xffffffff,
+	0x00002f50, 0x00000404, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x000064ec, 0x00007ffd, 0xffffffff,
+	0x00000c7c, 0x0000ff00, 0xffffffff,
+	0x00008dfc, 0x0000007f, 0xffffffff
+};
+#define CAYMAN_SYSLS_DISABLE_LENGTH sizeof(cayman_sysls_disable) / (3 * sizeof(u32))
+
+static const u32 cayman_sysls_enable[] =
+{
+	/* Register,   Value,     Mask bits */
+	0x000055e8, 0x00000001, 0xffffffff,
+	0x0000d0bc, 0x00000100, 0xffffffff,
+	0x0000d8bc, 0x00000100, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x00002f50, 0x00000903, 0xffffffff,
+	0x000004c8, 0x00000000, 0xffffffff,
+	0x000064ec, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00008dfc, 0x00000000, 0xffffffff
+};
+#define CAYMAN_SYSLS_ENABLE_LENGTH sizeof(cayman_sysls_enable) / (3 * sizeof(u32))
+
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev);
+struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev);
+
+struct ni_power_info *ni_get_pi(struct radeon_device *rdev)
+{
+        struct ni_power_info *pi = rdev->pm.dpm.priv;
+
+        return pi;
+}
+
+struct ni_ps *ni_get_ps(struct radeon_ps *rps)
+{
+	struct ni_ps *ps = rps->ps_priv;
+
+	return ps;
+}
+
+static void ni_calculate_leakage_for_v_and_t_formula(const struct ni_leakage_coeffients *coeff,
+						     u16 v, s32 t,
+						     u32 ileakage,
+						     u32 *leakage)
+{
+	s64 kt, kv, leakage_w, i_leakage, vddc, temperature;
+
+	i_leakage = div64_s64(drm_int2fixp(ileakage), 1000);
+	vddc = div64_s64(drm_int2fixp(v), 1000);
+	temperature = div64_s64(drm_int2fixp(t), 1000);
+
+	kt = drm_fixp_mul(div64_s64(drm_int2fixp(coeff->at), 1000),
+			  drm_fixp_exp(drm_fixp_mul(div64_s64(drm_int2fixp(coeff->bt), 1000), temperature)));
+	kv = drm_fixp_mul(div64_s64(drm_int2fixp(coeff->av), 1000),
+			  drm_fixp_exp(drm_fixp_mul(div64_s64(drm_int2fixp(coeff->bv), 1000), vddc)));
+
+	leakage_w = drm_fixp_mul(drm_fixp_mul(drm_fixp_mul(i_leakage, kt), kv), vddc);
+
+	*leakage = drm_fixp2int(leakage_w * 1000);
+}
+
+static void ni_calculate_leakage_for_v_and_t(struct radeon_device *rdev,
+					     const struct ni_leakage_coeffients *coeff,
+					     u16 v,
+					     s32 t,
+					     u32 i_leakage,
+					     u32 *leakage)
+{
+	ni_calculate_leakage_for_v_and_t_formula(coeff, v, t, i_leakage, leakage);
+}
+
+bool ni_dpm_vblank_too_short(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 vblank_time = r600_dpm_get_vblank_time(rdev);
+	u32 switch_limit = pi->mem_gddr5 ? 450 : 300;
+
+	if (vblank_time < switch_limit)
+		return true;
+	else
+		return false;
+
+}
+
+static void ni_apply_state_adjust_rules(struct radeon_device *rdev,
+					struct radeon_ps *rps)
+{
+	struct ni_ps *ps = ni_get_ps(rps);
+	struct radeon_clock_and_voltage_limits *max_limits;
+	bool disable_mclk_switching;
+	u32 mclk, sclk;
+	u16 vddc, vddci;
+	int i;
+
+	if ((rdev->pm.dpm.new_active_crtc_count > 1) ||
+	    ni_dpm_vblank_too_short(rdev))
+		disable_mclk_switching = true;
+	else
+		disable_mclk_switching = false;
+
+	if (rdev->pm.dpm.ac_power)
+		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac;
+	else
+		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc;
+
+	if (rdev->pm.dpm.ac_power == false) {
+		for (i = 0; i < ps->performance_level_count; i++) {
+			if (ps->performance_levels[i].mclk > max_limits->mclk)
+				ps->performance_levels[i].mclk = max_limits->mclk;
+			if (ps->performance_levels[i].sclk > max_limits->sclk)
+				ps->performance_levels[i].sclk = max_limits->sclk;
+			if (ps->performance_levels[i].vddc > max_limits->vddc)
+				ps->performance_levels[i].vddc = max_limits->vddc;
+			if (ps->performance_levels[i].vddci > max_limits->vddci)
+				ps->performance_levels[i].vddci = max_limits->vddci;
+		}
+	}
+
+	/* XXX validate the min clocks required for display */
+
+	if (disable_mclk_switching) {
+		mclk  = ps->performance_levels[ps->performance_level_count - 1].mclk;
+		sclk = ps->performance_levels[0].sclk;
+		vddc = ps->performance_levels[0].vddc;
+		vddci = ps->performance_levels[ps->performance_level_count - 1].vddci;
+	} else {
+		sclk = ps->performance_levels[0].sclk;
+		mclk = ps->performance_levels[0].mclk;
+		vddc = ps->performance_levels[0].vddc;
+		vddci = ps->performance_levels[0].vddci;
+	}
+
+	/* adjusted low state */
+	ps->performance_levels[0].sclk = sclk;
+	ps->performance_levels[0].mclk = mclk;
+	ps->performance_levels[0].vddc = vddc;
+	ps->performance_levels[0].vddci = vddci;
+
+	btc_skip_blacklist_clocks(rdev, max_limits->sclk, max_limits->mclk,
+				  &ps->performance_levels[0].sclk,
+				  &ps->performance_levels[0].mclk);
+
+	for (i = 1; i < ps->performance_level_count; i++) {
+		if (ps->performance_levels[i].sclk < ps->performance_levels[i - 1].sclk)
+			ps->performance_levels[i].sclk = ps->performance_levels[i - 1].sclk;
+		if (ps->performance_levels[i].vddc < ps->performance_levels[i - 1].vddc)
+			ps->performance_levels[i].vddc = ps->performance_levels[i - 1].vddc;
+	}
+
+	if (disable_mclk_switching) {
+		mclk = ps->performance_levels[0].mclk;
+		for (i = 1; i < ps->performance_level_count; i++) {
+			if (mclk < ps->performance_levels[i].mclk)
+				mclk = ps->performance_levels[i].mclk;
+		}
+		for (i = 0; i < ps->performance_level_count; i++) {
+			ps->performance_levels[i].mclk = mclk;
+			ps->performance_levels[i].vddci = vddci;
+		}
+	} else {
+		for (i = 1; i < ps->performance_level_count; i++) {
+			if (ps->performance_levels[i].mclk < ps->performance_levels[i - 1].mclk)
+				ps->performance_levels[i].mclk = ps->performance_levels[i - 1].mclk;
+			if (ps->performance_levels[i].vddci < ps->performance_levels[i - 1].vddci)
+				ps->performance_levels[i].vddci = ps->performance_levels[i - 1].vddci;
+		}
+	}
+
+	for (i = 1; i < ps->performance_level_count; i++)
+		btc_skip_blacklist_clocks(rdev, max_limits->sclk, max_limits->mclk,
+					  &ps->performance_levels[i].sclk,
+					  &ps->performance_levels[i].mclk);
+
+	for (i = 0; i < ps->performance_level_count; i++)
+		btc_adjust_clock_combinations(rdev, max_limits,
+					      &ps->performance_levels[i]);
+
+	for (i = 0; i < ps->performance_level_count; i++) {
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
+						   ps->performance_levels[i].sclk,
+						   max_limits->vddc,  &ps->performance_levels[i].vddc);
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+						   ps->performance_levels[i].mclk,
+						   max_limits->vddci, &ps->performance_levels[i].vddci);
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+						   ps->performance_levels[i].mclk,
+						   max_limits->vddc,  &ps->performance_levels[i].vddc);
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk,
+						   rdev->clock.current_dispclk,
+						   max_limits->vddc,  &ps->performance_levels[i].vddc);
+	}
+
+	for (i = 0; i < ps->performance_level_count; i++) {
+		btc_apply_voltage_delta_rules(rdev,
+					      max_limits->vddc, max_limits->vddci,
+					      &ps->performance_levels[i].vddc,
+					      &ps->performance_levels[i].vddci);
+	}
+
+	ps->dc_compatible = true;
+	for (i = 0; i < ps->performance_level_count; i++) {
+		if (ps->performance_levels[i].vddc > rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddc)
+			ps->dc_compatible = false;
+
+		if (ps->performance_levels[i].vddc < rdev->pm.dpm.dyn_state.min_vddc_for_pcie_gen2)
+			ps->performance_levels[i].flags &= ~ATOM_PPLIB_R600_FLAGS_PCIEGEN2;
+	}
+}
+
+static void ni_cg_clockgating_default(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *ps = NULL;
+
+	ps = (const u32 *)&cayman_cgcg_cgls_default;
+	count = CAYMAN_CGCG_CGLS_DEFAULT_LENGTH;
+
+	btc_program_mgcg_hw_sequence(rdev, ps, count);
+}
+
+static void ni_gfx_clockgating_enable(struct radeon_device *rdev,
+				      bool enable)
+{
+	u32 count;
+	const u32 *ps = NULL;
+
+	if (enable) {
+		ps = (const u32 *)&cayman_cgcg_cgls_enable;
+		count = CAYMAN_CGCG_CGLS_ENABLE_LENGTH;
+	} else {
+		ps = (const u32 *)&cayman_cgcg_cgls_disable;
+		count = CAYMAN_CGCG_CGLS_DISABLE_LENGTH;
+	}
+
+	btc_program_mgcg_hw_sequence(rdev, ps, count);
+}
+
+static void ni_mg_clockgating_default(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *ps = NULL;
+
+	ps = (const u32 *)&cayman_mgcg_default;
+	count = CAYMAN_MGCG_DEFAULT_LENGTH;
+
+	btc_program_mgcg_hw_sequence(rdev, ps, count);
+}
+
+static void ni_mg_clockgating_enable(struct radeon_device *rdev,
+				     bool enable)
+{
+	u32 count;
+	const u32 *ps = NULL;
+
+	if (enable) {
+		ps = (const u32 *)&cayman_mgcg_enable;
+		count = CAYMAN_MGCG_ENABLE_LENGTH;
+	} else {
+		ps = (const u32 *)&cayman_mgcg_disable;
+		count = CAYMAN_MGCG_DISABLE_LENGTH;
+	}
+
+	btc_program_mgcg_hw_sequence(rdev, ps, count);
+}
+
+static void ni_ls_clockgating_default(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *ps = NULL;
+
+	ps = (const u32 *)&cayman_sysls_default;
+	count = CAYMAN_SYSLS_DEFAULT_LENGTH;
+
+	btc_program_mgcg_hw_sequence(rdev, ps, count);
+}
+
+static void ni_ls_clockgating_enable(struct radeon_device *rdev,
+				     bool enable)
+{
+	u32 count;
+	const u32 *ps = NULL;
+
+	if (enable) {
+		ps = (const u32 *)&cayman_sysls_enable;
+		count = CAYMAN_SYSLS_ENABLE_LENGTH;
+	} else {
+		ps = (const u32 *)&cayman_sysls_disable;
+		count = CAYMAN_SYSLS_DISABLE_LENGTH;
+	}
+
+	btc_program_mgcg_hw_sequence(rdev, ps, count);
+
+}
+
+static int ni_patch_single_dependency_table_based_on_leakage(struct radeon_device *rdev,
+							     struct radeon_clock_voltage_dependency_table *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 i;
+
+	if (table) {
+		for (i = 0; i < table->count; i++) {
+			if (0xff01 == table->entries[i].v) {
+				if (pi->max_vddc == 0)
+					return -EINVAL;
+				table->entries[i].v = pi->max_vddc;
+			}
+		}
+	}
+	return 0;
+}
+
+static int ni_patch_dependency_tables_based_on_leakage(struct radeon_device *rdev)
+{
+	int ret = 0;
+
+	ret = ni_patch_single_dependency_table_based_on_leakage(rdev,
+								&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk);
+
+	ret = ni_patch_single_dependency_table_based_on_leakage(rdev,
+								&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk);
+	return ret;
+}
+
+static void ni_stop_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, 0, ~GLOBAL_PWRMGT_EN);
+}
+
+#if 0
+static int ni_notify_hw_of_power_source(struct radeon_device *rdev,
+					bool ac_power)
+{
+	if (ac_power)
+		return (rv770_send_msg_to_smc(rdev, PPSMC_MSG_RunningOnAC) == PPSMC_Result_OK) ?
+			0 : -EINVAL;
+
+	return 0;
+}
+#endif
+
+static PPSMC_Result ni_send_msg_to_smc_with_parameter(struct radeon_device *rdev,
+						      PPSMC_Msg msg, u32 parameter)
+{
+	WREG32(SMC_SCRATCH0, parameter);
+	return rv770_send_msg_to_smc(rdev, msg);
+}
+
+static int ni_restrict_performance_levels_before_switch(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_NoForcedLevel) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 1) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+int ni_dpm_force_performance_level(struct radeon_device *rdev,
+				   enum radeon_dpm_forced_level level)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct ni_ps *ps = ni_get_ps(rps);
+	u32 levels;
+
+	if (level == RADEON_DPM_FORCED_LEVEL_HIGH) {
+		if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+
+		if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 1) != PPSMC_Result_OK)
+			return -EINVAL;
+	} else if (level == RADEON_DPM_FORCED_LEVEL_LOW) {
+		if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+
+		levels = ps->performance_level_count - 1;
+		if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, levels) != PPSMC_Result_OK)
+			return -EINVAL;
+	} else if (level == RADEON_DPM_FORCED_LEVEL_AUTO) {
+		if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+
+		if (ni_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+	}
+
+	rdev->pm.dpm.forced_level = level;
+
+	return 0;
+}
+
+static void ni_stop_smc(struct radeon_device *rdev)
+{
+	u32 tmp;
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32(LB_SYNC_RESET_SEL) & LB_SYNC_RESET_SEL_MASK;
+		if (tmp != 1)
+			break;
+		udelay(1);
+	}
+
+	udelay(100);
+
+	r7xx_stop_smc(rdev);
+}
+
+static int ni_process_firmware_header(struct radeon_device *rdev)
+{
+        struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+        struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_stateTable,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	pi->state_table_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_softRegisters,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	pi->soft_regs_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_mcRegisterTable,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	eg_pi->mc_reg_table_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_fanTable,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	ni_pi->fan_table_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_mcArbDramAutoRefreshTable,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	ni_pi->arb_table_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_cacTable,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	ni_pi->cac_table_start = (u16)tmp;
+
+	ret = rv770_read_smc_sram_dword(rdev,
+					NISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+					NISLANDS_SMC_FIRMWARE_HEADER_spllTable,
+					&tmp, pi->sram_end);
+
+	if (ret)
+		return ret;
+
+	ni_pi->spll_table_start = (u16)tmp;
+
+
+	return ret;
+}
+
+static void ni_read_clock_registers(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+
+	ni_pi->clock_registers.cg_spll_func_cntl = RREG32(CG_SPLL_FUNC_CNTL);
+	ni_pi->clock_registers.cg_spll_func_cntl_2 = RREG32(CG_SPLL_FUNC_CNTL_2);
+	ni_pi->clock_registers.cg_spll_func_cntl_3 = RREG32(CG_SPLL_FUNC_CNTL_3);
+	ni_pi->clock_registers.cg_spll_func_cntl_4 = RREG32(CG_SPLL_FUNC_CNTL_4);
+	ni_pi->clock_registers.cg_spll_spread_spectrum = RREG32(CG_SPLL_SPREAD_SPECTRUM);
+	ni_pi->clock_registers.cg_spll_spread_spectrum_2 = RREG32(CG_SPLL_SPREAD_SPECTRUM_2);
+	ni_pi->clock_registers.mpll_ad_func_cntl = RREG32(MPLL_AD_FUNC_CNTL);
+	ni_pi->clock_registers.mpll_ad_func_cntl_2 = RREG32(MPLL_AD_FUNC_CNTL_2);
+	ni_pi->clock_registers.mpll_dq_func_cntl = RREG32(MPLL_DQ_FUNC_CNTL);
+	ni_pi->clock_registers.mpll_dq_func_cntl_2 = RREG32(MPLL_DQ_FUNC_CNTL_2);
+	ni_pi->clock_registers.mclk_pwrmgt_cntl = RREG32(MCLK_PWRMGT_CNTL);
+	ni_pi->clock_registers.dll_cntl = RREG32(DLL_CNTL);
+	ni_pi->clock_registers.mpll_ss1 = RREG32(MPLL_SS1);
+	ni_pi->clock_registers.mpll_ss2 = RREG32(MPLL_SS2);
+}
+
+#if 0
+static int ni_enter_ulp_state(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (pi->gfx_clock_gating) {
+                WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+                WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+		RREG32(GB_ADDR_CONFIG);
+        }
+
+	WREG32_P(SMC_MSG, HOST_SMC_MSG(PPSMC_MSG_SwitchToMinimumPower),
+                 ~HOST_SMC_MSG_MASK);
+
+	udelay(25000);
+
+	return 0;
+}
+#endif
+
+static void ni_program_response_times(struct radeon_device *rdev)
+{
+	u32 voltage_response_time, backbias_response_time, acpi_delay_time, vbi_time_out;
+	u32 vddc_dly, bb_dly, acpi_dly, vbi_dly, mclk_switch_limit;
+	u32 reference_clock;
+
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_mvdd_chg_time, 1);
+
+	voltage_response_time = (u32)rdev->pm.dpm.voltage_response_time;
+	backbias_response_time = (u32)rdev->pm.dpm.backbias_response_time;
+
+	if (voltage_response_time == 0)
+		voltage_response_time = 1000;
+
+	if (backbias_response_time == 0)
+		backbias_response_time = 1000;
+
+	acpi_delay_time = 15000;
+	vbi_time_out = 100000;
+
+	reference_clock = radeon_get_xclk(rdev);
+
+	vddc_dly = (voltage_response_time  * reference_clock) / 1600;
+	bb_dly   = (backbias_response_time * reference_clock) / 1600;
+	acpi_dly = (acpi_delay_time * reference_clock) / 1600;
+	vbi_dly  = (vbi_time_out * reference_clock) / 1600;
+
+	mclk_switch_limit = (460 * reference_clock) / 100;
+
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_delay_vreg,  vddc_dly);
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_delay_bbias, bb_dly);
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_delay_acpi,  acpi_dly);
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_mclk_chg_timeout, vbi_dly);
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_mc_block_delay, 0xAA);
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_mclk_switch_lim, mclk_switch_limit);
+}
+
+static void ni_populate_smc_voltage_table(struct radeon_device *rdev,
+					  struct atom_voltage_table *voltage_table,
+					  NISLANDS_SMC_STATETABLE *table)
+{
+	unsigned int i;
+
+	for (i = 0; i < voltage_table->count; i++) {
+		table->highSMIO[i] = 0;
+		table->lowSMIO[i] |= cpu_to_be32(voltage_table->entries[i].smio_low);
+	}
+}
+
+static void ni_populate_smc_voltage_tables(struct radeon_device *rdev,
+					   NISLANDS_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	unsigned char i;
+
+	if (eg_pi->vddc_voltage_table.count) {
+		ni_populate_smc_voltage_table(rdev, &eg_pi->vddc_voltage_table, table);
+		table->voltageMaskTable.highMask[NISLANDS_SMC_VOLTAGEMASK_VDDC] = 0;
+		table->voltageMaskTable.lowMask[NISLANDS_SMC_VOLTAGEMASK_VDDC] =
+			cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
+
+		for (i = 0; i < eg_pi->vddc_voltage_table.count; i++) {
+			if (pi->max_vddc_in_table <= eg_pi->vddc_voltage_table.entries[i].value) {
+				table->maxVDDCIndexInPPTable = i;
+				break;
+			}
+		}
+	}
+
+	if (eg_pi->vddci_voltage_table.count) {
+		ni_populate_smc_voltage_table(rdev, &eg_pi->vddci_voltage_table, table);
+
+		table->voltageMaskTable.highMask[NISLANDS_SMC_VOLTAGEMASK_VDDCI] = 0;
+		table->voltageMaskTable.lowMask[NISLANDS_SMC_VOLTAGEMASK_VDDCI] =
+			cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
+	}
+}
+
+static int ni_populate_voltage_value(struct radeon_device *rdev,
+				     struct atom_voltage_table *table,
+				     u16 value,
+				     NISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	unsigned int i;
+
+	for (i = 0; i < table->count; i++) {
+		if (value <= table->entries[i].value) {
+			voltage->index = (u8)i;
+			voltage->value = cpu_to_be16(table->entries[i].value);
+			break;
+		}
+	}
+
+	if (i >= table->count)
+		return -EINVAL;
+
+	return 0;
+}
+
+static void ni_populate_mvdd_value(struct radeon_device *rdev,
+				   u32 mclk,
+				   NISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+        struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (!pi->mvdd_control) {
+		voltage->index = eg_pi->mvdd_high_index;
+                voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+		return;
+	}
+
+	if (mclk <= pi->mvdd_split_frequency) {
+		voltage->index = eg_pi->mvdd_low_index;
+		voltage->value = cpu_to_be16(MVDD_LOW_VALUE);
+	} else {
+		voltage->index = eg_pi->mvdd_high_index;
+		voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+	}
+}
+
+static int ni_get_std_voltage_value(struct radeon_device *rdev,
+				    NISLANDS_SMC_VOLTAGE_VALUE *voltage,
+				    u16 *std_voltage)
+{
+	if (rdev->pm.dpm.dyn_state.cac_leakage_table.entries &&
+	    ((u32)voltage->index < rdev->pm.dpm.dyn_state.cac_leakage_table.count))
+		*std_voltage = rdev->pm.dpm.dyn_state.cac_leakage_table.entries[voltage->index].vddc;
+	else
+		*std_voltage = be16_to_cpu(voltage->value);
+
+	return 0;
+}
+
+static void ni_populate_std_voltage_value(struct radeon_device *rdev,
+					  u16 value, u8 index,
+					  NISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	voltage->index = index;
+	voltage->value = cpu_to_be16(value);
+}
+
+static u32 ni_get_smc_power_scaling_factor(struct radeon_device *rdev)
+{
+	u32 xclk_period;
+	u32 xclk = radeon_get_xclk(rdev);
+	u32 tmp = RREG32(CG_CAC_CTRL) & TID_CNT_MASK;
+
+	xclk_period = (1000000000UL / xclk);
+	xclk_period /= 10000UL;
+
+	return tmp * xclk_period;
+}
+
+static u32 ni_scale_power_for_smc(u32 power_in_watts, u32 scaling_factor)
+{
+	return (power_in_watts * scaling_factor) << 2;
+}
+
+static u32 ni_calculate_power_boost_limit(struct radeon_device *rdev,
+					  struct radeon_ps *radeon_state,
+					  u32 near_tdp_limit)
+{
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 power_boost_limit = 0;
+	int ret;
+
+	if (ni_pi->enable_power_containment &&
+	    ni_pi->use_power_boost_limit) {
+		NISLANDS_SMC_VOLTAGE_VALUE vddc;
+		u16 std_vddc_med;
+		u16 std_vddc_high;
+		u64 tmp, n, d;
+
+		if (state->performance_level_count < 3)
+			return 0;
+
+		ret = ni_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+						state->performance_levels[state->performance_level_count - 2].vddc,
+						&vddc);
+		if (ret)
+			return 0;
+
+		ret = ni_get_std_voltage_value(rdev, &vddc, &std_vddc_med);
+		if (ret)
+			return 0;
+
+		ret = ni_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+						state->performance_levels[state->performance_level_count - 1].vddc,
+						&vddc);
+		if (ret)
+			return 0;
+
+		ret = ni_get_std_voltage_value(rdev, &vddc, &std_vddc_high);
+		if (ret)
+			return 0;
+
+		n = ((u64)near_tdp_limit * ((u64)std_vddc_med * (u64)std_vddc_med) * 90);
+		d = ((u64)std_vddc_high * (u64)std_vddc_high * 100);
+		tmp = div64_u64(n, d);
+
+		if (tmp >> 32)
+			return 0;
+		power_boost_limit = (u32)tmp;
+	}
+
+	return power_boost_limit;
+}
+
+static int ni_calculate_adjusted_tdp_limits(struct radeon_device *rdev,
+					    bool adjust_polarity,
+					    u32 tdp_adjustment,
+					    u32 *tdp_limit,
+					    u32 *near_tdp_limit)
+{
+	if (tdp_adjustment > (u32)rdev->pm.dpm.tdp_od_limit)
+		return -EINVAL;
+
+	if (adjust_polarity) {
+		*tdp_limit = ((100 + tdp_adjustment) * rdev->pm.dpm.tdp_limit) / 100;
+		*near_tdp_limit = rdev->pm.dpm.near_tdp_limit + (*tdp_limit - rdev->pm.dpm.tdp_limit);
+	} else {
+		*tdp_limit = ((100 - tdp_adjustment) * rdev->pm.dpm.tdp_limit) / 100;
+		*near_tdp_limit = rdev->pm.dpm.near_tdp_limit - (rdev->pm.dpm.tdp_limit - *tdp_limit);
+	}
+
+	return 0;
+}
+
+static int ni_populate_smc_tdp_limits(struct radeon_device *rdev,
+				      struct radeon_ps *radeon_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+
+	if (ni_pi->enable_power_containment) {
+		NISLANDS_SMC_STATETABLE *smc_table = &ni_pi->smc_statetable;
+		u32 scaling_factor = ni_get_smc_power_scaling_factor(rdev);
+		u32 tdp_limit;
+		u32 near_tdp_limit;
+		u32 power_boost_limit;
+		int ret;
+
+		if (scaling_factor == 0)
+			return -EINVAL;
+
+		memset(smc_table, 0, sizeof(NISLANDS_SMC_STATETABLE));
+
+		ret = ni_calculate_adjusted_tdp_limits(rdev,
+						       false, /* ??? */
+						       rdev->pm.dpm.tdp_adjustment,
+						       &tdp_limit,
+						       &near_tdp_limit);
+		if (ret)
+			return ret;
+
+		power_boost_limit = ni_calculate_power_boost_limit(rdev, radeon_state,
+								   near_tdp_limit);
+
+		smc_table->dpm2Params.TDPLimit =
+			cpu_to_be32(ni_scale_power_for_smc(tdp_limit, scaling_factor));
+		smc_table->dpm2Params.NearTDPLimit =
+			cpu_to_be32(ni_scale_power_for_smc(near_tdp_limit, scaling_factor));
+		smc_table->dpm2Params.SafePowerLimit =
+			cpu_to_be32(ni_scale_power_for_smc((near_tdp_limit * NISLANDS_DPM2_TDP_SAFE_LIMIT_PERCENT) / 100,
+							   scaling_factor));
+		smc_table->dpm2Params.PowerBoostLimit =
+			cpu_to_be32(ni_scale_power_for_smc(power_boost_limit, scaling_factor));
+
+		ret = rv770_copy_bytes_to_smc(rdev,
+					      (u16)(pi->state_table_start + offsetof(NISLANDS_SMC_STATETABLE, dpm2Params) +
+						    offsetof(PP_NIslands_DPM2Parameters, TDPLimit)),
+					      (u8 *)(&smc_table->dpm2Params.TDPLimit),
+					      sizeof(u32) * 4, pi->sram_end);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+int ni_copy_and_switch_arb_sets(struct radeon_device *rdev,
+				u32 arb_freq_src, u32 arb_freq_dest)
+{
+	u32 mc_arb_dram_timing;
+	u32 mc_arb_dram_timing2;
+	u32 burst_time;
+	u32 mc_cg_config;
+
+	switch (arb_freq_src) {
+        case MC_CG_ARB_FREQ_F0:
+		mc_arb_dram_timing  = RREG32(MC_ARB_DRAM_TIMING);
+		mc_arb_dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+		burst_time = (RREG32(MC_ARB_BURST_TIME) & STATE0_MASK) >> STATE0_SHIFT;
+		break;
+        case MC_CG_ARB_FREQ_F1:
+		mc_arb_dram_timing  = RREG32(MC_ARB_DRAM_TIMING_1);
+		mc_arb_dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2_1);
+		burst_time = (RREG32(MC_ARB_BURST_TIME) & STATE1_MASK) >> STATE1_SHIFT;
+		break;
+        case MC_CG_ARB_FREQ_F2:
+		mc_arb_dram_timing  = RREG32(MC_ARB_DRAM_TIMING_2);
+		mc_arb_dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2_2);
+		burst_time = (RREG32(MC_ARB_BURST_TIME) & STATE2_MASK) >> STATE2_SHIFT;
+		break;
+        case MC_CG_ARB_FREQ_F3:
+		mc_arb_dram_timing  = RREG32(MC_ARB_DRAM_TIMING_3);
+		mc_arb_dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2_3);
+		burst_time = (RREG32(MC_ARB_BURST_TIME) & STATE3_MASK) >> STATE3_SHIFT;
+		break;
+        default:
+		return -EINVAL;
+	}
+
+	switch (arb_freq_dest) {
+        case MC_CG_ARB_FREQ_F0:
+		WREG32(MC_ARB_DRAM_TIMING, mc_arb_dram_timing);
+		WREG32(MC_ARB_DRAM_TIMING2, mc_arb_dram_timing2);
+		WREG32_P(MC_ARB_BURST_TIME, STATE0(burst_time), ~STATE0_MASK);
+		break;
+        case MC_CG_ARB_FREQ_F1:
+		WREG32(MC_ARB_DRAM_TIMING_1, mc_arb_dram_timing);
+		WREG32(MC_ARB_DRAM_TIMING2_1, mc_arb_dram_timing2);
+		WREG32_P(MC_ARB_BURST_TIME, STATE1(burst_time), ~STATE1_MASK);
+		break;
+        case MC_CG_ARB_FREQ_F2:
+		WREG32(MC_ARB_DRAM_TIMING_2, mc_arb_dram_timing);
+		WREG32(MC_ARB_DRAM_TIMING2_2, mc_arb_dram_timing2);
+		WREG32_P(MC_ARB_BURST_TIME, STATE2(burst_time), ~STATE2_MASK);
+		break;
+        case MC_CG_ARB_FREQ_F3:
+		WREG32(MC_ARB_DRAM_TIMING_3, mc_arb_dram_timing);
+		WREG32(MC_ARB_DRAM_TIMING2_3, mc_arb_dram_timing2);
+		WREG32_P(MC_ARB_BURST_TIME, STATE3(burst_time), ~STATE3_MASK);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	mc_cg_config = RREG32(MC_CG_CONFIG) | 0x0000000F;
+	WREG32(MC_CG_CONFIG, mc_cg_config);
+	WREG32_P(MC_ARB_CG, CG_ARB_REQ(arb_freq_dest), ~CG_ARB_REQ_MASK);
+
+	return 0;
+}
+
+static int ni_init_arb_table_index(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = rv770_read_smc_sram_dword(rdev, ni_pi->arb_table_start,
+					&tmp, pi->sram_end);
+	if (ret)
+		return ret;
+
+	tmp &= 0x00FFFFFF;
+	tmp |= ((u32)MC_CG_ARB_FREQ_F1) << 24;
+
+	return rv770_write_smc_sram_dword(rdev, ni_pi->arb_table_start,
+					  tmp, pi->sram_end);
+}
+
+static int ni_initial_switch_from_arb_f0_to_f1(struct radeon_device *rdev)
+{
+	return ni_copy_and_switch_arb_sets(rdev, MC_CG_ARB_FREQ_F0, MC_CG_ARB_FREQ_F1);
+}
+
+static int ni_force_switch_to_arb_f0(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = rv770_read_smc_sram_dword(rdev, ni_pi->arb_table_start,
+					&tmp, pi->sram_end);
+	if (ret)
+		return ret;
+
+	tmp = (tmp >> 24) & 0xff;
+
+	if (tmp == MC_CG_ARB_FREQ_F0)
+		return 0;
+
+	return ni_copy_and_switch_arb_sets(rdev, tmp, MC_CG_ARB_FREQ_F0);
+}
+
+static int ni_populate_memory_timing_parameters(struct radeon_device *rdev,
+						struct rv7xx_pl *pl,
+						SMC_NIslands_MCArbDramTimingRegisterSet *arb_regs)
+{
+	u32 dram_timing;
+	u32 dram_timing2;
+
+	arb_regs->mc_arb_rfsh_rate =
+		(u8)rv770_calculate_memory_refresh_rate(rdev, pl->sclk);
+
+
+	radeon_atom_set_engine_dram_timings(rdev,
+                                            pl->sclk,
+                                            pl->mclk);
+
+	dram_timing = RREG32(MC_ARB_DRAM_TIMING);
+	dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+
+	arb_regs->mc_arb_dram_timing  = cpu_to_be32(dram_timing);
+	arb_regs->mc_arb_dram_timing2 = cpu_to_be32(dram_timing2);
+
+	return 0;
+}
+
+static int ni_do_program_memory_timing_parameters(struct radeon_device *rdev,
+						  struct radeon_ps *radeon_state,
+						  unsigned int first_arb_set)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	SMC_NIslands_MCArbDramTimingRegisterSet arb_regs = { 0 };
+	int i, ret = 0;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		ret = ni_populate_memory_timing_parameters(rdev, &state->performance_levels[i], &arb_regs);
+		if (ret)
+			break;
+
+		ret = rv770_copy_bytes_to_smc(rdev,
+					      (u16)(ni_pi->arb_table_start +
+						    offsetof(SMC_NIslands_MCArbDramTimingRegisters, data) +
+						    sizeof(SMC_NIslands_MCArbDramTimingRegisterSet) * (first_arb_set + i)),
+					      (u8 *)&arb_regs,
+					      (u16)sizeof(SMC_NIslands_MCArbDramTimingRegisterSet),
+					      pi->sram_end);
+		if (ret)
+			break;
+	}
+	return ret;
+}
+
+static int ni_program_memory_timing_parameters(struct radeon_device *rdev,
+					       struct radeon_ps *radeon_new_state)
+{
+	return ni_do_program_memory_timing_parameters(rdev, radeon_new_state,
+						      NISLANDS_DRIVER_STATE_ARB_INDEX);
+}
+
+static void ni_populate_initial_mvdd_value(struct radeon_device *rdev,
+					   struct NISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	voltage->index = eg_pi->mvdd_high_index;
+	voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+}
+
+static int ni_populate_smc_initial_state(struct radeon_device *rdev,
+					 struct radeon_ps *radeon_initial_state,
+					 NISLANDS_SMC_STATETABLE *table)
+{
+	struct ni_ps *initial_state = ni_get_ps(radeon_initial_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 reg;
+	int ret;
+
+	table->initialState.levels[0].mclk.vMPLL_AD_FUNC_CNTL =
+		cpu_to_be32(ni_pi->clock_registers.mpll_ad_func_cntl);
+	table->initialState.levels[0].mclk.vMPLL_AD_FUNC_CNTL_2 =
+		cpu_to_be32(ni_pi->clock_registers.mpll_ad_func_cntl_2);
+	table->initialState.levels[0].mclk.vMPLL_DQ_FUNC_CNTL =
+		cpu_to_be32(ni_pi->clock_registers.mpll_dq_func_cntl);
+	table->initialState.levels[0].mclk.vMPLL_DQ_FUNC_CNTL_2 =
+		cpu_to_be32(ni_pi->clock_registers.mpll_dq_func_cntl_2);
+	table->initialState.levels[0].mclk.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(ni_pi->clock_registers.mclk_pwrmgt_cntl);
+	table->initialState.levels[0].mclk.vDLL_CNTL =
+		cpu_to_be32(ni_pi->clock_registers.dll_cntl);
+	table->initialState.levels[0].mclk.vMPLL_SS =
+		cpu_to_be32(ni_pi->clock_registers.mpll_ss1);
+	table->initialState.levels[0].mclk.vMPLL_SS2 =
+		cpu_to_be32(ni_pi->clock_registers.mpll_ss2);
+	table->initialState.levels[0].mclk.mclk_value =
+		cpu_to_be32(initial_state->performance_levels[0].mclk);
+
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(ni_pi->clock_registers.cg_spll_func_cntl);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(ni_pi->clock_registers.cg_spll_func_cntl_2);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(ni_pi->clock_registers.cg_spll_func_cntl_3);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_4 =
+		cpu_to_be32(ni_pi->clock_registers.cg_spll_func_cntl_4);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM =
+		cpu_to_be32(ni_pi->clock_registers.cg_spll_spread_spectrum);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM_2 =
+		cpu_to_be32(ni_pi->clock_registers.cg_spll_spread_spectrum_2);
+	table->initialState.levels[0].sclk.sclk_value =
+		cpu_to_be32(initial_state->performance_levels[0].sclk);
+	table->initialState.levels[0].arbRefreshState =
+		NISLANDS_INITIAL_STATE_ARB_INDEX;
+
+	table->initialState.levels[0].ACIndex = 0;
+
+	ret = ni_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+					initial_state->performance_levels[0].vddc,
+					&table->initialState.levels[0].vddc);
+	if (!ret) {
+		u16 std_vddc;
+
+		ret = ni_get_std_voltage_value(rdev,
+					       &table->initialState.levels[0].vddc,
+					       &std_vddc);
+		if (!ret)
+			ni_populate_std_voltage_value(rdev, std_vddc,
+						      table->initialState.levels[0].vddc.index,
+						      &table->initialState.levels[0].std_vddc);
+	}
+
+	if (eg_pi->vddci_control)
+		ni_populate_voltage_value(rdev,
+					  &eg_pi->vddci_voltage_table,
+					  initial_state->performance_levels[0].vddci,
+					  &table->initialState.levels[0].vddci);
+
+	ni_populate_initial_mvdd_value(rdev, &table->initialState.levels[0].mvdd);
+
+	reg = CG_R(0xffff) | CG_L(0);
+	table->initialState.levels[0].aT = cpu_to_be32(reg);
+
+	table->initialState.levels[0].bSP = cpu_to_be32(pi->dsp);
+
+	if (pi->boot_in_gen2)
+		table->initialState.levels[0].gen2PCIE = 1;
+	else
+		table->initialState.levels[0].gen2PCIE = 0;
+
+	if (pi->mem_gddr5) {
+		table->initialState.levels[0].strobeMode =
+			cypress_get_strobe_mode_settings(rdev,
+							 initial_state->performance_levels[0].mclk);
+
+		if (initial_state->performance_levels[0].mclk > pi->mclk_edc_enable_threshold)
+			table->initialState.levels[0].mcFlags = NISLANDS_SMC_MC_EDC_RD_FLAG | NISLANDS_SMC_MC_EDC_WR_FLAG;
+		else
+			table->initialState.levels[0].mcFlags =  0;
+	}
+
+	table->initialState.levelCount = 1;
+
+	table->initialState.flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	table->initialState.levels[0].dpm2.MaxPS = 0;
+	table->initialState.levels[0].dpm2.NearTDPDec = 0;
+	table->initialState.levels[0].dpm2.AboveSafeInc = 0;
+	table->initialState.levels[0].dpm2.BelowSafeInc = 0;
+
+	reg = MIN_POWER_MASK | MAX_POWER_MASK;
+	table->initialState.levels[0].SQPowerThrottle = cpu_to_be32(reg);
+
+	reg = MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK;
+	table->initialState.levels[0].SQPowerThrottle_2 = cpu_to_be32(reg);
+
+	return 0;
+}
+
+static int ni_populate_smc_acpi_state(struct radeon_device *rdev,
+				      NISLANDS_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 mpll_ad_func_cntl   = ni_pi->clock_registers.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 = ni_pi->clock_registers.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl   = ni_pi->clock_registers.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 = ni_pi->clock_registers.mpll_dq_func_cntl_2;
+	u32 spll_func_cntl      = ni_pi->clock_registers.cg_spll_func_cntl;
+	u32 spll_func_cntl_2    = ni_pi->clock_registers.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3    = ni_pi->clock_registers.cg_spll_func_cntl_3;
+	u32 spll_func_cntl_4    = ni_pi->clock_registers.cg_spll_func_cntl_4;
+	u32 mclk_pwrmgt_cntl    = ni_pi->clock_registers.mclk_pwrmgt_cntl;
+	u32 dll_cntl            = ni_pi->clock_registers.dll_cntl;
+	u32 reg;
+	int ret;
+
+	table->ACPIState = table->initialState;
+
+	table->ACPIState.flags &= ~PPSMC_SWSTATE_FLAG_DC;
+
+	if (pi->acpi_vddc) {
+		ret = ni_populate_voltage_value(rdev,
+						&eg_pi->vddc_voltage_table,
+						pi->acpi_vddc, &table->ACPIState.levels[0].vddc);
+		if (!ret) {
+			u16 std_vddc;
+
+			ret = ni_get_std_voltage_value(rdev,
+						       &table->ACPIState.levels[0].vddc, &std_vddc);
+			if (!ret)
+				ni_populate_std_voltage_value(rdev, std_vddc,
+							      table->ACPIState.levels[0].vddc.index,
+							      &table->ACPIState.levels[0].std_vddc);
+		}
+
+		if (pi->pcie_gen2) {
+			if (pi->acpi_pcie_gen2)
+				table->ACPIState.levels[0].gen2PCIE = 1;
+			else
+				table->ACPIState.levels[0].gen2PCIE = 0;
+		} else {
+			table->ACPIState.levels[0].gen2PCIE = 0;
+		}
+	} else {
+		ret = ni_populate_voltage_value(rdev,
+						&eg_pi->vddc_voltage_table,
+						pi->min_vddc_in_table,
+						&table->ACPIState.levels[0].vddc);
+		if (!ret) {
+			u16 std_vddc;
+
+			ret = ni_get_std_voltage_value(rdev,
+						       &table->ACPIState.levels[0].vddc,
+						       &std_vddc);
+			if (!ret)
+				ni_populate_std_voltage_value(rdev, std_vddc,
+							      table->ACPIState.levels[0].vddc.index,
+							      &table->ACPIState.levels[0].std_vddc);
+		}
+		table->ACPIState.levels[0].gen2PCIE = 0;
+	}
+
+	if (eg_pi->acpi_vddci) {
+		if (eg_pi->vddci_control)
+			ni_populate_voltage_value(rdev,
+						  &eg_pi->vddci_voltage_table,
+						  eg_pi->acpi_vddci,
+						  &table->ACPIState.levels[0].vddci);
+	}
+
+
+	mpll_ad_func_cntl &= ~PDNB;
+
+	mpll_ad_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN;
+
+        if (pi->mem_gddr5)
+                mpll_dq_func_cntl &= ~PDNB;
+        mpll_dq_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN | BYPASS;
+
+
+	mclk_pwrmgt_cntl |= (MRDCKA0_RESET |
+			     MRDCKA1_RESET |
+			     MRDCKB0_RESET |
+			     MRDCKB1_RESET |
+			     MRDCKC0_RESET |
+			     MRDCKC1_RESET |
+			     MRDCKD0_RESET |
+			     MRDCKD1_RESET);
+
+	mclk_pwrmgt_cntl &= ~(MRDCKA0_PDNB |
+			      MRDCKA1_PDNB |
+			      MRDCKB0_PDNB |
+			      MRDCKB1_PDNB |
+			      MRDCKC0_PDNB |
+			      MRDCKC1_PDNB |
+			      MRDCKD0_PDNB |
+			      MRDCKD1_PDNB);
+
+	dll_cntl |= (MRDCKA0_BYPASS |
+                     MRDCKA1_BYPASS |
+                     MRDCKB0_BYPASS |
+                     MRDCKB1_BYPASS |
+                     MRDCKC0_BYPASS |
+                     MRDCKC1_BYPASS |
+                     MRDCKD0_BYPASS |
+                     MRDCKD1_BYPASS);
+
+        spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(4);
+
+	table->ACPIState.levels[0].mclk.vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	table->ACPIState.levels[0].mclk.vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	table->ACPIState.levels[0].mclk.vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	table->ACPIState.levels[0].mclk.vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+	table->ACPIState.levels[0].mclk.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	table->ACPIState.levels[0].mclk.vDLL_CNTL = cpu_to_be32(dll_cntl);
+
+	table->ACPIState.levels[0].mclk.mclk_value = 0;
+
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_4 = cpu_to_be32(spll_func_cntl_4);
+
+	table->ACPIState.levels[0].sclk.sclk_value = 0;
+
+	ni_populate_mvdd_value(rdev, 0, &table->ACPIState.levels[0].mvdd);
+
+	if (eg_pi->dynamic_ac_timing)
+		table->ACPIState.levels[0].ACIndex = 1;
+
+	table->ACPIState.levels[0].dpm2.MaxPS = 0;
+	table->ACPIState.levels[0].dpm2.NearTDPDec = 0;
+	table->ACPIState.levels[0].dpm2.AboveSafeInc = 0;
+	table->ACPIState.levels[0].dpm2.BelowSafeInc = 0;
+
+	reg = MIN_POWER_MASK | MAX_POWER_MASK;
+	table->ACPIState.levels[0].SQPowerThrottle = cpu_to_be32(reg);
+
+	reg = MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK;
+	table->ACPIState.levels[0].SQPowerThrottle_2 = cpu_to_be32(reg);
+
+	return 0;
+}
+
+static int ni_init_smc_table(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	int ret;
+	struct radeon_ps *radeon_boot_state = rdev->pm.dpm.boot_ps;
+	NISLANDS_SMC_STATETABLE *table = &ni_pi->smc_statetable;
+
+	memset(table, 0, sizeof(NISLANDS_SMC_STATETABLE));
+
+	ni_populate_smc_voltage_tables(rdev, table);
+
+	switch (rdev->pm.int_thermal_type) {
+	case THERMAL_TYPE_NI:
+	case THERMAL_TYPE_EMC2103_WITH_INTERNAL:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_INTERNAL;
+		break;
+	case THERMAL_TYPE_NONE:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_NONE;
+		break;
+	default:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_EXTERNAL;
+		break;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_HARDWAREDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GPIO_DC;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REGULATOR_HOT)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_REGULATOR_HOT;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_STEPVDDC;
+
+	if (pi->mem_gddr5)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GDDR5;
+
+	ret = ni_populate_smc_initial_state(rdev, radeon_boot_state, table);
+	if (ret)
+		return ret;
+
+	ret = ni_populate_smc_acpi_state(rdev, table);
+	if (ret)
+		return ret;
+
+	table->driverState = table->initialState;
+
+	table->ULVState = table->initialState;
+
+	ret = ni_do_program_memory_timing_parameters(rdev, radeon_boot_state,
+						     NISLANDS_INITIAL_STATE_ARB_INDEX);
+	if (ret)
+		return ret;
+
+	return rv770_copy_bytes_to_smc(rdev, pi->state_table_start, (u8 *)table,
+				       sizeof(NISLANDS_SMC_STATETABLE), pi->sram_end);
+}
+
+static int ni_calculate_sclk_params(struct radeon_device *rdev,
+				    u32 engine_clock,
+				    NISLANDS_SMC_SCLK_VALUE *sclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	u32 spll_func_cntl = ni_pi->clock_registers.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 = ni_pi->clock_registers.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 = ni_pi->clock_registers.cg_spll_func_cntl_3;
+	u32 spll_func_cntl_4 = ni_pi->clock_registers.cg_spll_func_cntl_4;
+	u32 cg_spll_spread_spectrum = ni_pi->clock_registers.cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2 = ni_pi->clock_registers.cg_spll_spread_spectrum_2;
+	u64 tmp;
+	u32 reference_clock = rdev->clock.spll.reference_freq;
+	u32 reference_divider;
+	u32 fbdiv;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     engine_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	reference_divider = 1 + dividers.ref_div;
+
+
+	tmp = (u64) engine_clock * reference_divider * dividers.post_div * 16834;
+	do_div(tmp, reference_clock);
+	fbdiv = (u32) tmp;
+
+	spll_func_cntl &= ~(SPLL_PDIV_A_MASK | SPLL_REF_DIV_MASK);
+	spll_func_cntl |= SPLL_REF_DIV(dividers.ref_div);
+	spll_func_cntl |= SPLL_PDIV_A(dividers.post_div);
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(2);
+
+	spll_func_cntl_3 &= ~SPLL_FB_DIV_MASK;
+	spll_func_cntl_3 |= SPLL_FB_DIV(fbdiv);
+	spll_func_cntl_3 |= SPLL_DITHEN;
+
+	if (pi->sclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = engine_clock * dividers.post_div;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_ENGINE_SS, vco_freq)) {
+			u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate);
+			u32 clk_v = 4 * ss.percentage * fbdiv / (clk_s * 10000);
+
+			cg_spll_spread_spectrum &= ~CLK_S_MASK;
+			cg_spll_spread_spectrum |= CLK_S(clk_s);
+			cg_spll_spread_spectrum |= SSEN;
+
+			cg_spll_spread_spectrum_2 &= ~CLK_V_MASK;
+			cg_spll_spread_spectrum_2 |= CLK_V(clk_v);
+		}
+	}
+
+	sclk->sclk_value = engine_clock;
+	sclk->vCG_SPLL_FUNC_CNTL = spll_func_cntl;
+	sclk->vCG_SPLL_FUNC_CNTL_2 = spll_func_cntl_2;
+	sclk->vCG_SPLL_FUNC_CNTL_3 = spll_func_cntl_3;
+	sclk->vCG_SPLL_FUNC_CNTL_4 = spll_func_cntl_4;
+	sclk->vCG_SPLL_SPREAD_SPECTRUM = cg_spll_spread_spectrum;
+	sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cg_spll_spread_spectrum_2;
+
+	return 0;
+}
+
+static int ni_populate_sclk_value(struct radeon_device *rdev,
+				  u32 engine_clock,
+				  NISLANDS_SMC_SCLK_VALUE *sclk)
+{
+	NISLANDS_SMC_SCLK_VALUE sclk_tmp;
+	int ret;
+
+	ret = ni_calculate_sclk_params(rdev, engine_clock, &sclk_tmp);
+	if (!ret) {
+		sclk->sclk_value = cpu_to_be32(sclk_tmp.sclk_value);
+		sclk->vCG_SPLL_FUNC_CNTL = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL);
+		sclk->vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL_2);
+		sclk->vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL_3);
+		sclk->vCG_SPLL_FUNC_CNTL_4 = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL_4);
+		sclk->vCG_SPLL_SPREAD_SPECTRUM = cpu_to_be32(sclk_tmp.vCG_SPLL_SPREAD_SPECTRUM);
+		sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cpu_to_be32(sclk_tmp.vCG_SPLL_SPREAD_SPECTRUM_2);
+	}
+
+	return ret;
+}
+
+static int ni_init_smc_spll_table(struct radeon_device *rdev)
+{
+        struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	SMC_NISLANDS_SPLL_DIV_TABLE *spll_table;
+	NISLANDS_SMC_SCLK_VALUE sclk_params;
+	u32 fb_div;
+	u32 p_div;
+	u32 clk_s;
+	u32 clk_v;
+	u32 sclk = 0;
+	int i, ret;
+	u32 tmp;
+
+	if (ni_pi->spll_table_start == 0)
+		return -EINVAL;
+
+	spll_table = kzalloc(sizeof(SMC_NISLANDS_SPLL_DIV_TABLE), GFP_KERNEL);
+	if (spll_table == NULL)
+		return -ENOMEM;
+
+	for (i = 0; i < 256; i++) {
+		ret = ni_calculate_sclk_params(rdev, sclk, &sclk_params);
+		if (ret)
+			break;
+
+		p_div = (sclk_params.vCG_SPLL_FUNC_CNTL & SPLL_PDIV_A_MASK) >> SPLL_PDIV_A_SHIFT;
+		fb_div = (sclk_params.vCG_SPLL_FUNC_CNTL_3 & SPLL_FB_DIV_MASK) >> SPLL_FB_DIV_SHIFT;
+		clk_s = (sclk_params.vCG_SPLL_SPREAD_SPECTRUM & CLK_S_MASK) >> CLK_S_SHIFT;
+		clk_v = (sclk_params.vCG_SPLL_SPREAD_SPECTRUM_2 & CLK_V_MASK) >> CLK_V_SHIFT;
+
+		fb_div &= ~0x00001FFF;
+		fb_div >>= 1;
+		clk_v >>= 6;
+
+		if (p_div & ~(SMC_NISLANDS_SPLL_DIV_TABLE_PDIV_MASK >> SMC_NISLANDS_SPLL_DIV_TABLE_PDIV_SHIFT))
+			ret = -EINVAL;
+
+		if (clk_s & ~(SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_MASK >> SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT))
+			ret = -EINVAL;
+
+		if (clk_s & ~(SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_MASK >> SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT))
+			ret = -EINVAL;
+
+		if (clk_v & ~(SMC_NISLANDS_SPLL_DIV_TABLE_CLKV_MASK >> SMC_NISLANDS_SPLL_DIV_TABLE_CLKV_SHIFT))
+			ret = -EINVAL;
+
+		if (ret)
+			break;
+
+		tmp = ((fb_div << SMC_NISLANDS_SPLL_DIV_TABLE_FBDIV_SHIFT) & SMC_NISLANDS_SPLL_DIV_TABLE_FBDIV_MASK) |
+			((p_div << SMC_NISLANDS_SPLL_DIV_TABLE_PDIV_SHIFT) & SMC_NISLANDS_SPLL_DIV_TABLE_PDIV_MASK);
+		spll_table->freq[i] = cpu_to_be32(tmp);
+
+		tmp = ((clk_v << SMC_NISLANDS_SPLL_DIV_TABLE_CLKV_SHIFT) & SMC_NISLANDS_SPLL_DIV_TABLE_CLKV_MASK) |
+			((clk_s << SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT) & SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_MASK);
+		spll_table->ss[i] = cpu_to_be32(tmp);
+
+		sclk += 512;
+	}
+
+	if (!ret)
+		ret = rv770_copy_bytes_to_smc(rdev, ni_pi->spll_table_start, (u8 *)spll_table,
+					      sizeof(SMC_NISLANDS_SPLL_DIV_TABLE), pi->sram_end);
+
+	kfree(spll_table);
+
+	return ret;
+}
+
+static int ni_populate_mclk_value(struct radeon_device *rdev,
+				  u32 engine_clock,
+				  u32 memory_clock,
+				  NISLANDS_SMC_MCLK_VALUE *mclk,
+				  bool strobe_mode,
+				  bool dll_state_on)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 mpll_ad_func_cntl = ni_pi->clock_registers.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 = ni_pi->clock_registers.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl = ni_pi->clock_registers.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 = ni_pi->clock_registers.mpll_dq_func_cntl_2;
+	u32 mclk_pwrmgt_cntl = ni_pi->clock_registers.mclk_pwrmgt_cntl;
+	u32 dll_cntl = ni_pi->clock_registers.dll_cntl;
+	u32 mpll_ss1 = ni_pi->clock_registers.mpll_ss1;
+	u32 mpll_ss2 = ni_pi->clock_registers.mpll_ss2;
+	struct atom_clock_dividers dividers;
+	u32 ibias;
+	u32 dll_speed;
+	int ret;
+	u32 mc_seq_misc7;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					     memory_clock, strobe_mode, &dividers);
+	if (ret)
+		return ret;
+
+	if (!strobe_mode) {
+		mc_seq_misc7 = RREG32(MC_SEQ_MISC7);
+
+		if (mc_seq_misc7 & 0x8000000)
+			dividers.post_div = 1;
+	}
+
+	ibias = cypress_map_clkf_to_ibias(rdev, dividers.whole_fb_div);
+
+	mpll_ad_func_cntl &= ~(CLKR_MASK |
+			       YCLK_POST_DIV_MASK |
+			       CLKF_MASK |
+			       CLKFRAC_MASK |
+			       IBIAS_MASK);
+	mpll_ad_func_cntl |= CLKR(dividers.ref_div);
+	mpll_ad_func_cntl |= YCLK_POST_DIV(dividers.post_div);
+	mpll_ad_func_cntl |= CLKF(dividers.whole_fb_div);
+	mpll_ad_func_cntl |= CLKFRAC(dividers.frac_fb_div);
+	mpll_ad_func_cntl |= IBIAS(ibias);
+
+	if (dividers.vco_mode)
+		mpll_ad_func_cntl_2 |= VCO_MODE;
+	else
+		mpll_ad_func_cntl_2 &= ~VCO_MODE;
+
+	if (pi->mem_gddr5) {
+		mpll_dq_func_cntl &= ~(CLKR_MASK |
+				       YCLK_POST_DIV_MASK |
+				       CLKF_MASK |
+				       CLKFRAC_MASK |
+				       IBIAS_MASK);
+		mpll_dq_func_cntl |= CLKR(dividers.ref_div);
+		mpll_dq_func_cntl |= YCLK_POST_DIV(dividers.post_div);
+		mpll_dq_func_cntl |= CLKF(dividers.whole_fb_div);
+		mpll_dq_func_cntl |= CLKFRAC(dividers.frac_fb_div);
+		mpll_dq_func_cntl |= IBIAS(ibias);
+
+		if (strobe_mode)
+			mpll_dq_func_cntl &= ~PDNB;
+		else
+			mpll_dq_func_cntl |= PDNB;
+
+		if (dividers.vco_mode)
+			mpll_dq_func_cntl_2 |= VCO_MODE;
+		else
+			mpll_dq_func_cntl_2 &= ~VCO_MODE;
+	}
+
+	if (pi->mclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = memory_clock * dividers.post_div;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_MEMORY_SS, vco_freq)) {
+			u32 reference_clock = rdev->clock.mpll.reference_freq;
+			u32 decoded_ref = rv740_get_decoded_reference_divider(dividers.ref_div);
+			u32 clk_s = reference_clock * 5 / (decoded_ref * ss.rate);
+			u32 clk_v = ss.percentage *
+				(0x4000 * dividers.whole_fb_div + 0x800 * dividers.frac_fb_div) / (clk_s * 625);
+
+			mpll_ss1 &= ~CLKV_MASK;
+			mpll_ss1 |= CLKV(clk_v);
+
+			mpll_ss2 &= ~CLKS_MASK;
+			mpll_ss2 |= CLKS(clk_s);
+		}
+	}
+
+	dll_speed = rv740_get_dll_speed(pi->mem_gddr5,
+					memory_clock);
+
+	mclk_pwrmgt_cntl &= ~DLL_SPEED_MASK;
+	mclk_pwrmgt_cntl |= DLL_SPEED(dll_speed);
+	if (dll_state_on)
+		mclk_pwrmgt_cntl |= (MRDCKA0_PDNB |
+				     MRDCKA1_PDNB |
+				     MRDCKB0_PDNB |
+				     MRDCKB1_PDNB |
+				     MRDCKC0_PDNB |
+				     MRDCKC1_PDNB |
+				     MRDCKD0_PDNB |
+				     MRDCKD1_PDNB);
+	else
+		mclk_pwrmgt_cntl &= ~(MRDCKA0_PDNB |
+				      MRDCKA1_PDNB |
+				      MRDCKB0_PDNB |
+				      MRDCKB1_PDNB |
+				      MRDCKC0_PDNB |
+				      MRDCKC1_PDNB |
+				      MRDCKD0_PDNB |
+				      MRDCKD1_PDNB);
+
+
+	mclk->mclk_value = cpu_to_be32(memory_clock);
+	mclk->vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	mclk->vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	mclk->vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	mclk->vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+	mclk->vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	mclk->vDLL_CNTL = cpu_to_be32(dll_cntl);
+	mclk->vMPLL_SS = cpu_to_be32(mpll_ss1);
+	mclk->vMPLL_SS2 = cpu_to_be32(mpll_ss2);
+
+	return 0;
+}
+
+static void ni_populate_smc_sp(struct radeon_device *rdev,
+			       struct radeon_ps *radeon_state,
+			       NISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct ni_ps *ps = ni_get_ps(radeon_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+
+	for (i = 0; i < ps->performance_level_count - 1; i++)
+		smc_state->levels[i].bSP = cpu_to_be32(pi->dsp);
+
+	smc_state->levels[ps->performance_level_count - 1].bSP =
+		cpu_to_be32(pi->psp);
+}
+
+static int ni_convert_power_level_to_smc(struct radeon_device *rdev,
+					 struct rv7xx_pl *pl,
+					 NISLANDS_SMC_HW_PERFORMANCE_LEVEL *level)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+        struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	int ret;
+	bool dll_state_on;
+	u16 std_vddc;
+	u32 tmp = RREG32(DC_STUTTER_CNTL);
+
+	level->gen2PCIE = pi->pcie_gen2 ?
+		((pl->flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? 1 : 0) : 0;
+
+	ret = ni_populate_sclk_value(rdev, pl->sclk, &level->sclk);
+	if (ret)
+		return ret;
+
+	level->mcFlags =  0;
+	if (pi->mclk_stutter_mode_threshold &&
+	    (pl->mclk <= pi->mclk_stutter_mode_threshold) &&
+	    !eg_pi->uvd_enabled &&
+	    (tmp & DC_STUTTER_ENABLE_A) &&
+	    (tmp & DC_STUTTER_ENABLE_B))
+		level->mcFlags |= NISLANDS_SMC_MC_STUTTER_EN;
+
+	if (pi->mem_gddr5) {
+		if (pl->mclk > pi->mclk_edc_enable_threshold)
+			level->mcFlags |= NISLANDS_SMC_MC_EDC_RD_FLAG;
+		if (pl->mclk > eg_pi->mclk_edc_wr_enable_threshold)
+			level->mcFlags |= NISLANDS_SMC_MC_EDC_WR_FLAG;
+
+		level->strobeMode = cypress_get_strobe_mode_settings(rdev, pl->mclk);
+
+		if (level->strobeMode & NISLANDS_SMC_STROBE_ENABLE) {
+			if (cypress_get_mclk_frequency_ratio(rdev, pl->mclk, true) >=
+			    ((RREG32(MC_SEQ_MISC7) >> 16) & 0xf))
+				dll_state_on = ((RREG32(MC_SEQ_MISC5) >> 1) & 0x1) ? true : false;
+			else
+				dll_state_on = ((RREG32(MC_SEQ_MISC6) >> 1) & 0x1) ? true : false;
+		} else {
+			dll_state_on = false;
+			if (pl->mclk > ni_pi->mclk_rtt_mode_threshold)
+				level->mcFlags |= NISLANDS_SMC_MC_RTT_ENABLE;
+		}
+
+		ret = ni_populate_mclk_value(rdev, pl->sclk, pl->mclk,
+					     &level->mclk,
+					     (level->strobeMode & NISLANDS_SMC_STROBE_ENABLE) != 0,
+					     dll_state_on);
+	} else
+		ret = ni_populate_mclk_value(rdev, pl->sclk, pl->mclk, &level->mclk, 1, 1);
+
+	if (ret)
+		return ret;
+
+	ret = ni_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+					pl->vddc, &level->vddc);
+	if (ret)
+		return ret;
+
+	ret = ni_get_std_voltage_value(rdev, &level->vddc, &std_vddc);
+	if (ret)
+		return ret;
+
+	ni_populate_std_voltage_value(rdev, std_vddc,
+				      level->vddc.index, &level->std_vddc);
+
+	if (eg_pi->vddci_control) {
+		ret = ni_populate_voltage_value(rdev, &eg_pi->vddci_voltage_table,
+						pl->vddci, &level->vddci);
+		if (ret)
+			return ret;
+	}
+
+	ni_populate_mvdd_value(rdev, pl->mclk, &level->mvdd);
+
+	return ret;
+}
+
+static int ni_populate_smc_t(struct radeon_device *rdev,
+			     struct radeon_ps *radeon_state,
+			     NISLANDS_SMC_SWSTATE *smc_state)
+{
+        struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+        struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	u32 a_t;
+	u32 t_l, t_h;
+	u32 high_bsp;
+	int i, ret;
+
+	if (state->performance_level_count >= 9)
+		return -EINVAL;
+
+	if (state->performance_level_count < 2) {
+		a_t = CG_R(0xffff) | CG_L(0);
+		smc_state->levels[0].aT = cpu_to_be32(a_t);
+		return 0;
+	}
+
+	smc_state->levels[0].aT = cpu_to_be32(0);
+
+	for (i = 0; i <= state->performance_level_count - 2; i++) {
+		if (eg_pi->uvd_enabled)
+			ret = r600_calculate_at(
+				1000 * (i * (eg_pi->smu_uvd_hs ? 2 : 8) + 2),
+				100 * R600_AH_DFLT,
+				state->performance_levels[i + 1].sclk,
+				state->performance_levels[i].sclk,
+				&t_l,
+				&t_h);
+		else
+			ret = r600_calculate_at(
+				1000 * (i + 1),
+				100 * R600_AH_DFLT,
+				state->performance_levels[i + 1].sclk,
+				state->performance_levels[i].sclk,
+				&t_l,
+				&t_h);
+
+		if (ret) {
+			t_h = (i + 1) * 1000 - 50 * R600_AH_DFLT;
+			t_l = (i + 1) * 1000 + 50 * R600_AH_DFLT;
+		}
+
+		a_t = be32_to_cpu(smc_state->levels[i].aT) & ~CG_R_MASK;
+		a_t |= CG_R(t_l * pi->bsp / 20000);
+		smc_state->levels[i].aT = cpu_to_be32(a_t);
+
+		high_bsp = (i == state->performance_level_count - 2) ?
+			pi->pbsp : pi->bsp;
+
+		a_t = CG_R(0xffff) | CG_L(t_h * high_bsp / 20000);
+		smc_state->levels[i + 1].aT = cpu_to_be32(a_t);
+	}
+
+	return 0;
+}
+
+static int ni_populate_power_containment_values(struct radeon_device *rdev,
+						struct radeon_ps *radeon_state,
+						NISLANDS_SMC_SWSTATE *smc_state)
+{
+        struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+        struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	u32 prev_sclk;
+	u32 max_sclk;
+	u32 min_sclk;
+	int i, ret;
+	u32 tdp_limit;
+	u32 near_tdp_limit;
+	u32 power_boost_limit;
+	u8 max_ps_percent;
+
+	if (ni_pi->enable_power_containment == false)
+		return 0;
+
+	if (state->performance_level_count == 0)
+		return -EINVAL;
+
+	if (smc_state->levelCount != state->performance_level_count)
+		return -EINVAL;
+
+	ret = ni_calculate_adjusted_tdp_limits(rdev,
+					       false, /* ??? */
+					       rdev->pm.dpm.tdp_adjustment,
+					       &tdp_limit,
+					       &near_tdp_limit);
+	if (ret)
+		return ret;
+
+	power_boost_limit = ni_calculate_power_boost_limit(rdev, radeon_state, near_tdp_limit);
+
+	ret = rv770_write_smc_sram_dword(rdev,
+					 pi->state_table_start +
+					 offsetof(NISLANDS_SMC_STATETABLE, dpm2Params) +
+					 offsetof(PP_NIslands_DPM2Parameters, PowerBoostLimit),
+					 ni_scale_power_for_smc(power_boost_limit, ni_get_smc_power_scaling_factor(rdev)),
+					 pi->sram_end);
+	if (ret)
+		power_boost_limit = 0;
+
+	smc_state->levels[0].dpm2.MaxPS = 0;
+	smc_state->levels[0].dpm2.NearTDPDec = 0;
+	smc_state->levels[0].dpm2.AboveSafeInc = 0;
+	smc_state->levels[0].dpm2.BelowSafeInc = 0;
+	smc_state->levels[0].stateFlags |= power_boost_limit ? PPSMC_STATEFLAG_POWERBOOST : 0;
+
+	for (i = 1; i < state->performance_level_count; i++) {
+		prev_sclk = state->performance_levels[i-1].sclk;
+		max_sclk  = state->performance_levels[i].sclk;
+		max_ps_percent = (i != (state->performance_level_count - 1)) ?
+			NISLANDS_DPM2_MAXPS_PERCENT_M : NISLANDS_DPM2_MAXPS_PERCENT_H;
+
+		if (max_sclk < prev_sclk)
+			return -EINVAL;
+
+		if ((max_ps_percent == 0) || (prev_sclk == max_sclk) || eg_pi->uvd_enabled)
+			min_sclk = max_sclk;
+		else if (1 == i)
+			min_sclk = prev_sclk;
+		else
+			min_sclk = (prev_sclk * (u32)max_ps_percent) / 100;
+
+		if (min_sclk < state->performance_levels[0].sclk)
+			min_sclk = state->performance_levels[0].sclk;
+
+		if (min_sclk == 0)
+			return -EINVAL;
+
+		smc_state->levels[i].dpm2.MaxPS =
+			(u8)((NISLANDS_DPM2_MAX_PULSE_SKIP * (max_sclk - min_sclk)) / max_sclk);
+		smc_state->levels[i].dpm2.NearTDPDec = NISLANDS_DPM2_NEAR_TDP_DEC;
+		smc_state->levels[i].dpm2.AboveSafeInc = NISLANDS_DPM2_ABOVE_SAFE_INC;
+		smc_state->levels[i].dpm2.BelowSafeInc = NISLANDS_DPM2_BELOW_SAFE_INC;
+		smc_state->levels[i].stateFlags |=
+			((i != (state->performance_level_count - 1)) && power_boost_limit) ?
+			PPSMC_STATEFLAG_POWERBOOST : 0;
+	}
+
+	return 0;
+}
+
+static int ni_populate_sq_ramping_values(struct radeon_device *rdev,
+					 struct radeon_ps *radeon_state,
+					 NISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	u32 sq_power_throttle;
+	u32 sq_power_throttle2;
+	bool enable_sq_ramping = ni_pi->enable_sq_ramping;
+	int i;
+
+	if (state->performance_level_count == 0)
+		return -EINVAL;
+
+	if (smc_state->levelCount != state->performance_level_count)
+		return -EINVAL;
+
+	if (rdev->pm.dpm.sq_ramping_threshold == 0)
+		return -EINVAL;
+
+	if (NISLANDS_DPM2_SQ_RAMP_MAX_POWER > (MAX_POWER_MASK >> MAX_POWER_SHIFT))
+		enable_sq_ramping = false;
+
+	if (NISLANDS_DPM2_SQ_RAMP_MIN_POWER > (MIN_POWER_MASK >> MIN_POWER_SHIFT))
+		enable_sq_ramping = false;
+
+	if (NISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA > (MAX_POWER_DELTA_MASK >> MAX_POWER_DELTA_SHIFT))
+		enable_sq_ramping = false;
+
+	if (NISLANDS_DPM2_SQ_RAMP_STI_SIZE > (STI_SIZE_MASK >> STI_SIZE_SHIFT))
+		enable_sq_ramping = false;
+
+	if (NISLANDS_DPM2_SQ_RAMP_LTI_RATIO <= (LTI_RATIO_MASK >> LTI_RATIO_SHIFT))
+		enable_sq_ramping = false;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		sq_power_throttle  = 0;
+		sq_power_throttle2 = 0;
+
+		if ((state->performance_levels[i].sclk >= rdev->pm.dpm.sq_ramping_threshold) &&
+		    enable_sq_ramping) {
+			sq_power_throttle |= MAX_POWER(NISLANDS_DPM2_SQ_RAMP_MAX_POWER);
+			sq_power_throttle |= MIN_POWER(NISLANDS_DPM2_SQ_RAMP_MIN_POWER);
+			sq_power_throttle2 |= MAX_POWER_DELTA(NISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA);
+			sq_power_throttle2 |= STI_SIZE(NISLANDS_DPM2_SQ_RAMP_STI_SIZE);
+			sq_power_throttle2 |= LTI_RATIO(NISLANDS_DPM2_SQ_RAMP_LTI_RATIO);
+		} else {
+			sq_power_throttle |= MAX_POWER_MASK | MIN_POWER_MASK;
+			sq_power_throttle2 |= MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK;
+		}
+
+		smc_state->levels[i].SQPowerThrottle   = cpu_to_be32(sq_power_throttle);
+		smc_state->levels[i].SQPowerThrottle_2 = cpu_to_be32(sq_power_throttle2);
+	}
+
+	return 0;
+}
+
+static int ni_enable_power_containment(struct radeon_device *rdev,
+				       struct radeon_ps *radeon_new_state,
+				       bool enable)
+{
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	PPSMC_Result smc_result;
+	int ret = 0;
+
+	if (ni_pi->enable_power_containment) {
+		if (enable) {
+			if (!r600_is_uvd_state(radeon_new_state->class, radeon_new_state->class2)) {
+				smc_result = rv770_send_msg_to_smc(rdev, PPSMC_TDPClampingActive);
+				if (smc_result != PPSMC_Result_OK) {
+					ret = -EINVAL;
+					ni_pi->pc_enabled = false;
+				} else {
+					ni_pi->pc_enabled = true;
+				}
+			}
+		} else {
+			smc_result = rv770_send_msg_to_smc(rdev, PPSMC_TDPClampingInactive);
+			if (smc_result != PPSMC_Result_OK)
+				ret = -EINVAL;
+			ni_pi->pc_enabled = false;
+		}
+	}
+
+	return ret;
+}
+
+static int ni_convert_power_state_to_smc(struct radeon_device *rdev,
+					 struct radeon_ps *radeon_state,
+					 NISLANDS_SMC_SWSTATE *smc_state)
+{
+        struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	int i, ret;
+	u32 threshold = state->performance_levels[state->performance_level_count - 1].sclk * 100 / 100;
+
+	if (!(radeon_state->caps & ATOM_PPLIB_DISALLOW_ON_DC))
+		smc_state->flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	smc_state->levelCount = 0;
+
+	if (state->performance_level_count > NISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE)
+		return -EINVAL;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		ret = ni_convert_power_level_to_smc(rdev, &state->performance_levels[i],
+						    &smc_state->levels[i]);
+		smc_state->levels[i].arbRefreshState =
+			(u8)(NISLANDS_DRIVER_STATE_ARB_INDEX + i);
+
+		if (ret)
+			return ret;
+
+		if (ni_pi->enable_power_containment)
+			smc_state->levels[i].displayWatermark =
+				(state->performance_levels[i].sclk < threshold) ?
+				PPSMC_DISPLAY_WATERMARK_LOW : PPSMC_DISPLAY_WATERMARK_HIGH;
+		else
+			smc_state->levels[i].displayWatermark = (i < 2) ?
+				PPSMC_DISPLAY_WATERMARK_LOW : PPSMC_DISPLAY_WATERMARK_HIGH;
+
+		if (eg_pi->dynamic_ac_timing)
+			smc_state->levels[i].ACIndex = NISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT + i;
+		else
+			smc_state->levels[i].ACIndex = 0;
+
+		smc_state->levelCount++;
+	}
+
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_watermark_threshold,
+				      cpu_to_be32(threshold / 512));
+
+	ni_populate_smc_sp(rdev, radeon_state, smc_state);
+
+	ret = ni_populate_power_containment_values(rdev, radeon_state, smc_state);
+	if (ret)
+		ni_pi->enable_power_containment = false;
+
+	ret = ni_populate_sq_ramping_values(rdev, radeon_state, smc_state);
+	if (ret)
+		ni_pi->enable_sq_ramping = false;
+
+	return ni_populate_smc_t(rdev, radeon_state, smc_state);
+}
+
+static int ni_upload_sw_state(struct radeon_device *rdev,
+			      struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u16 address = pi->state_table_start +
+		offsetof(NISLANDS_SMC_STATETABLE, driverState);
+	u16 state_size = sizeof(NISLANDS_SMC_SWSTATE) +
+		((NISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1) * sizeof(NISLANDS_SMC_HW_PERFORMANCE_LEVEL));
+	int ret;
+	NISLANDS_SMC_SWSTATE *smc_state = kzalloc(state_size, GFP_KERNEL);
+
+	if (smc_state == NULL)
+		return -ENOMEM;
+
+	ret = ni_convert_power_state_to_smc(rdev, radeon_new_state, smc_state);
+	if (ret)
+		goto done;
+
+	ret = rv770_copy_bytes_to_smc(rdev, address, (u8 *)smc_state, state_size, pi->sram_end);
+
+done:
+	kfree(smc_state);
+
+	return ret;
+}
+
+static int ni_set_mc_special_registers(struct radeon_device *rdev,
+				       struct ni_mc_reg_table *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 i, j, k;
+	u32 temp_reg;
+
+	for (i = 0, j = table->last; i < table->last; i++) {
+		switch (table->mc_reg_address[i].s1) {
+		case MC_SEQ_MISC1 >> 2:
+			if (j >= SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			temp_reg = RREG32(MC_PMG_CMD_EMRS);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_EMRS >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+			for (k = 0; k < table->num_entries; k++)
+				table->mc_reg_table_entry[k].mc_data[j] =
+					((temp_reg & 0xffff0000)) |
+					((table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16);
+			j++;
+			if (j >= SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+
+			temp_reg = RREG32(MC_PMG_CMD_MRS);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+			for(k = 0; k < table->num_entries; k++) {
+				table->mc_reg_table_entry[k].mc_data[j] =
+					(temp_reg & 0xffff0000) |
+					(table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff);
+				if (!pi->mem_gddr5)
+					table->mc_reg_table_entry[k].mc_data[j] |= 0x100;
+			}
+			j++;
+			if (j > SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			break;
+		case MC_SEQ_RESERVE_M >> 2:
+			temp_reg = RREG32(MC_PMG_CMD_MRS1);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS1 >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+			for (k = 0; k < table->num_entries; k++)
+				table->mc_reg_table_entry[k].mc_data[j] =
+					(temp_reg & 0xffff0000) |
+					(table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff);
+			j++;
+			if (j > SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			break;
+		default:
+			break;
+		}
+	}
+
+	table->last = j;
+
+	return 0;
+}
+
+static bool ni_check_s0_mc_reg_index(u16 in_reg, u16 *out_reg)
+{
+	bool result = true;
+
+	switch (in_reg) {
+        case  MC_SEQ_RAS_TIMING >> 2:
+		*out_reg = MC_SEQ_RAS_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_CAS_TIMING >> 2:
+		*out_reg = MC_SEQ_CAS_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_MISC_TIMING >> 2:
+		*out_reg = MC_SEQ_MISC_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_MISC_TIMING2 >> 2:
+		*out_reg = MC_SEQ_MISC_TIMING2_LP >> 2;
+		break;
+        case MC_SEQ_RD_CTL_D0 >> 2:
+		*out_reg = MC_SEQ_RD_CTL_D0_LP >> 2;
+		break;
+        case MC_SEQ_RD_CTL_D1 >> 2:
+		*out_reg = MC_SEQ_RD_CTL_D1_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_D0 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_D0_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_D1 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_D1_LP >> 2;
+		break;
+        case MC_PMG_CMD_EMRS >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS1 >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+		break;
+        case MC_SEQ_PMG_TIMING >> 2:
+		*out_reg = MC_SEQ_PMG_TIMING_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS2 >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS2_LP >> 2;
+		break;
+        default:
+		result = false;
+		break;
+	}
+
+	return result;
+}
+
+static void ni_set_valid_flag(struct ni_mc_reg_table *table)
+{
+	u8 i, j;
+
+	for (i = 0; i < table->last; i++) {
+		for (j = 1; j < table->num_entries; j++) {
+			if (table->mc_reg_table_entry[j-1].mc_data[i] != table->mc_reg_table_entry[j].mc_data[i]) {
+				table->valid_flag |= 1 << i;
+				break;
+			}
+		}
+	}
+}
+
+static void ni_set_s0_mc_reg_index(struct ni_mc_reg_table *table)
+{
+	u32 i;
+	u16 address;
+
+	for (i = 0; i < table->last; i++)
+		table->mc_reg_address[i].s0 =
+			ni_check_s0_mc_reg_index(table->mc_reg_address[i].s1, &address) ?
+			address : table->mc_reg_address[i].s1;
+}
+
+static int ni_copy_vbios_mc_reg_table(struct atom_mc_reg_table *table,
+				      struct ni_mc_reg_table *ni_table)
+{
+	u8 i, j;
+
+	if (table->last > SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+		return -EINVAL;
+	if (table->num_entries > MAX_AC_TIMING_ENTRIES)
+		return -EINVAL;
+
+	for (i = 0; i < table->last; i++)
+		ni_table->mc_reg_address[i].s1 = table->mc_reg_address[i].s1;
+	ni_table->last = table->last;
+
+	for (i = 0; i < table->num_entries; i++) {
+		ni_table->mc_reg_table_entry[i].mclk_max =
+			table->mc_reg_table_entry[i].mclk_max;
+		for (j = 0; j < table->last; j++)
+			ni_table->mc_reg_table_entry[i].mc_data[j] =
+				table->mc_reg_table_entry[i].mc_data[j];
+	}
+	ni_table->num_entries = table->num_entries;
+
+	return 0;
+}
+
+static int ni_initialize_mc_reg_table(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	int ret;
+	struct atom_mc_reg_table *table;
+	struct ni_mc_reg_table *ni_table = &ni_pi->mc_reg_table;
+	u8 module_index = rv770_get_memory_module_index(rdev);
+
+        table = kzalloc(sizeof(struct atom_mc_reg_table), GFP_KERNEL);
+        if (!table)
+                return -ENOMEM;
+
+	WREG32(MC_SEQ_RAS_TIMING_LP, RREG32(MC_SEQ_RAS_TIMING));
+	WREG32(MC_SEQ_CAS_TIMING_LP, RREG32(MC_SEQ_CAS_TIMING));
+	WREG32(MC_SEQ_MISC_TIMING_LP, RREG32(MC_SEQ_MISC_TIMING));
+	WREG32(MC_SEQ_MISC_TIMING2_LP, RREG32(MC_SEQ_MISC_TIMING2));
+	WREG32(MC_SEQ_PMG_CMD_EMRS_LP, RREG32(MC_PMG_CMD_EMRS));
+	WREG32(MC_SEQ_PMG_CMD_MRS_LP, RREG32(MC_PMG_CMD_MRS));
+	WREG32(MC_SEQ_PMG_CMD_MRS1_LP, RREG32(MC_PMG_CMD_MRS1));
+	WREG32(MC_SEQ_WR_CTL_D0_LP, RREG32(MC_SEQ_WR_CTL_D0));
+	WREG32(MC_SEQ_WR_CTL_D1_LP, RREG32(MC_SEQ_WR_CTL_D1));
+	WREG32(MC_SEQ_RD_CTL_D0_LP, RREG32(MC_SEQ_RD_CTL_D0));
+	WREG32(MC_SEQ_RD_CTL_D1_LP, RREG32(MC_SEQ_RD_CTL_D1));
+	WREG32(MC_SEQ_PMG_TIMING_LP, RREG32(MC_SEQ_PMG_TIMING));
+	WREG32(MC_SEQ_PMG_CMD_MRS2_LP, RREG32(MC_PMG_CMD_MRS2));
+
+	ret = radeon_atom_init_mc_reg_table(rdev, module_index, table);
+
+        if (ret)
+                goto init_mc_done;
+
+	ret = ni_copy_vbios_mc_reg_table(table, ni_table);
+
+        if (ret)
+                goto init_mc_done;
+
+	ni_set_s0_mc_reg_index(ni_table);
+
+	ret = ni_set_mc_special_registers(rdev, ni_table);
+
+        if (ret)
+                goto init_mc_done;
+
+	ni_set_valid_flag(ni_table);
+
+init_mc_done:
+        kfree(table);
+
+	return ret;
+}
+
+static void ni_populate_mc_reg_addresses(struct radeon_device *rdev,
+					 SMC_NIslands_MCRegisters *mc_reg_table)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 i, j;
+
+	for (i = 0, j = 0; j < ni_pi->mc_reg_table.last; j++) {
+		if (ni_pi->mc_reg_table.valid_flag & (1 << j)) {
+			if (i >= SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+				break;
+			mc_reg_table->address[i].s0 =
+				cpu_to_be16(ni_pi->mc_reg_table.mc_reg_address[j].s0);
+			mc_reg_table->address[i].s1 =
+				cpu_to_be16(ni_pi->mc_reg_table.mc_reg_address[j].s1);
+			i++;
+		}
+	}
+	mc_reg_table->last = (u8)i;
+}
+
+
+static void ni_convert_mc_registers(struct ni_mc_reg_entry *entry,
+				    SMC_NIslands_MCRegisterSet *data,
+				    u32 num_entries, u32 valid_flag)
+{
+	u32 i, j;
+
+	for (i = 0, j = 0; j < num_entries; j++) {
+		if (valid_flag & (1 << j)) {
+			data->value[i] = cpu_to_be32(entry->mc_data[j]);
+			i++;
+		}
+	}
+}
+
+static void ni_convert_mc_reg_table_entry_to_smc(struct radeon_device *rdev,
+						 struct rv7xx_pl *pl,
+						 SMC_NIslands_MCRegisterSet *mc_reg_table_data)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 i = 0;
+
+	for (i = 0; i < ni_pi->mc_reg_table.num_entries; i++) {
+		if (pl->mclk <= ni_pi->mc_reg_table.mc_reg_table_entry[i].mclk_max)
+			break;
+	}
+
+	if ((i == ni_pi->mc_reg_table.num_entries) && (i > 0))
+		--i;
+
+	ni_convert_mc_registers(&ni_pi->mc_reg_table.mc_reg_table_entry[i],
+				mc_reg_table_data,
+				ni_pi->mc_reg_table.last,
+				ni_pi->mc_reg_table.valid_flag);
+}
+
+static void ni_convert_mc_reg_table_to_smc(struct radeon_device *rdev,
+					   struct radeon_ps *radeon_state,
+					   SMC_NIslands_MCRegisters *mc_reg_table)
+{
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	int i;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		ni_convert_mc_reg_table_entry_to_smc(rdev,
+						     &state->performance_levels[i],
+						     &mc_reg_table->data[NISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT + i]);
+	}
+}
+
+static int ni_populate_mc_reg_table(struct radeon_device *rdev,
+				    struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *boot_state = ni_get_ps(radeon_boot_state);
+	SMC_NIslands_MCRegisters *mc_reg_table = &ni_pi->smc_mc_reg_table;
+
+	memset(mc_reg_table, 0, sizeof(SMC_NIslands_MCRegisters));
+
+	rv770_write_smc_soft_register(rdev, NI_SMC_SOFT_REGISTER_seq_index, 1);
+
+	ni_populate_mc_reg_addresses(rdev, mc_reg_table);
+
+	ni_convert_mc_reg_table_entry_to_smc(rdev, &boot_state->performance_levels[0],
+					     &mc_reg_table->data[0]);
+
+	ni_convert_mc_registers(&ni_pi->mc_reg_table.mc_reg_table_entry[0],
+				&mc_reg_table->data[1],
+				ni_pi->mc_reg_table.last,
+				ni_pi->mc_reg_table.valid_flag);
+
+	ni_convert_mc_reg_table_to_smc(rdev, radeon_boot_state, mc_reg_table);
+
+	return rv770_copy_bytes_to_smc(rdev, eg_pi->mc_reg_table_start,
+				       (u8 *)mc_reg_table,
+				       sizeof(SMC_NIslands_MCRegisters),
+				       pi->sram_end);
+}
+
+static int ni_upload_mc_reg_table(struct radeon_device *rdev,
+				  struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *ni_new_state = ni_get_ps(radeon_new_state);
+	SMC_NIslands_MCRegisters *mc_reg_table = &ni_pi->smc_mc_reg_table;
+	u16 address;
+
+	memset(mc_reg_table, 0, sizeof(SMC_NIslands_MCRegisters));
+
+	ni_convert_mc_reg_table_to_smc(rdev, radeon_new_state, mc_reg_table);
+
+	address = eg_pi->mc_reg_table_start +
+		(u16)offsetof(SMC_NIslands_MCRegisters, data[NISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT]);
+
+	return rv770_copy_bytes_to_smc(rdev, address,
+				       (u8 *)&mc_reg_table->data[NISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT],
+				       sizeof(SMC_NIslands_MCRegisterSet) * ni_new_state->performance_level_count,
+				       pi->sram_end);
+}
+
+static int ni_init_driver_calculated_leakage_table(struct radeon_device *rdev,
+						   PP_NIslands_CACTABLES *cac_tables)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 leakage = 0;
+	unsigned int i, j, table_size;
+	s32 t;
+	u32 smc_leakage, max_leakage = 0;
+	u32 scaling_factor;
+
+	table_size = eg_pi->vddc_voltage_table.count;
+
+	if (SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES < table_size)
+		table_size = SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES;
+
+	scaling_factor = ni_get_smc_power_scaling_factor(rdev);
+
+	for (i = 0; i < SMC_NISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES; i++) {
+		for (j = 0; j < table_size; j++) {
+			t = (1000 * ((i + 1) * 8));
+
+			if (t < ni_pi->cac_data.leakage_minimum_temperature)
+				t = ni_pi->cac_data.leakage_minimum_temperature;
+
+			ni_calculate_leakage_for_v_and_t(rdev,
+							 &ni_pi->cac_data.leakage_coefficients,
+							 eg_pi->vddc_voltage_table.entries[j].value,
+							 t,
+							 ni_pi->cac_data.i_leakage,
+							 &leakage);
+
+			smc_leakage = ni_scale_power_for_smc(leakage, scaling_factor) / 1000;
+			if (smc_leakage > max_leakage)
+				max_leakage = smc_leakage;
+
+			cac_tables->cac_lkge_lut[i][j] = cpu_to_be32(smc_leakage);
+		}
+	}
+
+	for (j = table_size; j < SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES; j++) {
+		for (i = 0; i < SMC_NISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES; i++)
+			cac_tables->cac_lkge_lut[i][j] = cpu_to_be32(max_leakage);
+	}
+	return 0;
+}
+
+static int ni_init_simplified_leakage_table(struct radeon_device *rdev,
+					    PP_NIslands_CACTABLES *cac_tables)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_cac_leakage_table *leakage_table =
+		&rdev->pm.dpm.dyn_state.cac_leakage_table;
+	u32 i, j, table_size;
+	u32 smc_leakage, max_leakage = 0;
+	u32 scaling_factor;
+
+	if (!leakage_table)
+		return -EINVAL;
+
+	table_size = leakage_table->count;
+
+	if (eg_pi->vddc_voltage_table.count != table_size)
+		table_size = (eg_pi->vddc_voltage_table.count < leakage_table->count) ?
+			eg_pi->vddc_voltage_table.count : leakage_table->count;
+
+	if (SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES < table_size)
+		table_size = SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES;
+
+	if (table_size == 0)
+		return -EINVAL;
+
+	scaling_factor = ni_get_smc_power_scaling_factor(rdev);
+
+	for (j = 0; j < table_size; j++) {
+		smc_leakage = leakage_table->entries[j].leakage;
+
+		if (smc_leakage > max_leakage)
+			max_leakage = smc_leakage;
+
+		for (i = 0; i < SMC_NISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES; i++)
+			cac_tables->cac_lkge_lut[i][j] =
+				cpu_to_be32(ni_scale_power_for_smc(smc_leakage, scaling_factor));
+	}
+
+	for (j = table_size; j < SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES; j++) {
+		for (i = 0; i < SMC_NISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES; i++)
+			cac_tables->cac_lkge_lut[i][j] =
+				cpu_to_be32(ni_scale_power_for_smc(max_leakage, scaling_factor));
+	}
+	return 0;
+}
+
+static int ni_initialize_smc_cac_tables(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	PP_NIslands_CACTABLES *cac_tables = NULL;
+	int i, ret;
+        u32 reg;
+
+	if (ni_pi->enable_cac == false)
+		return 0;
+
+	cac_tables = kzalloc(sizeof(PP_NIslands_CACTABLES), GFP_KERNEL);
+	if (!cac_tables)
+		return -ENOMEM;
+
+	reg = RREG32(CG_CAC_CTRL) & ~(TID_CNT_MASK | TID_UNIT_MASK);
+	reg |= (TID_CNT(ni_pi->cac_weights->tid_cnt) |
+		TID_UNIT(ni_pi->cac_weights->tid_unit));
+	WREG32(CG_CAC_CTRL, reg);
+
+	for (i = 0; i < NISLANDS_DCCAC_MAX_LEVELS; i++)
+		ni_pi->dc_cac_table[i] = ni_pi->cac_weights->dc_cac[i];
+
+	for (i = 0; i < SMC_NISLANDS_BIF_LUT_NUM_OF_ENTRIES; i++)
+		cac_tables->cac_bif_lut[i] = ni_pi->cac_weights->pcie_cac[i];
+
+	ni_pi->cac_data.i_leakage = rdev->pm.dpm.cac_leakage;
+	ni_pi->cac_data.pwr_const = 0;
+	ni_pi->cac_data.dc_cac_value = ni_pi->dc_cac_table[NISLANDS_DCCAC_LEVEL_0];
+	ni_pi->cac_data.bif_cac_value = 0;
+	ni_pi->cac_data.mc_wr_weight = ni_pi->cac_weights->mc_write_weight;
+	ni_pi->cac_data.mc_rd_weight = ni_pi->cac_weights->mc_read_weight;
+	ni_pi->cac_data.allow_ovrflw = 0;
+	ni_pi->cac_data.l2num_win_tdp = ni_pi->lta_window_size;
+	ni_pi->cac_data.num_win_tdp = 0;
+	ni_pi->cac_data.lts_truncate_n = ni_pi->lts_truncate;
+
+	if (ni_pi->driver_calculate_cac_leakage)
+		ret = ni_init_driver_calculated_leakage_table(rdev, cac_tables);
+	else
+		ret = ni_init_simplified_leakage_table(rdev, cac_tables);
+
+	if (ret)
+		goto done_free;
+
+	cac_tables->pwr_const      = cpu_to_be32(ni_pi->cac_data.pwr_const);
+	cac_tables->dc_cacValue    = cpu_to_be32(ni_pi->cac_data.dc_cac_value);
+	cac_tables->bif_cacValue   = cpu_to_be32(ni_pi->cac_data.bif_cac_value);
+	cac_tables->AllowOvrflw    = ni_pi->cac_data.allow_ovrflw;
+	cac_tables->MCWrWeight     = ni_pi->cac_data.mc_wr_weight;
+	cac_tables->MCRdWeight     = ni_pi->cac_data.mc_rd_weight;
+	cac_tables->numWin_TDP     = ni_pi->cac_data.num_win_tdp;
+	cac_tables->l2numWin_TDP   = ni_pi->cac_data.l2num_win_tdp;
+	cac_tables->lts_truncate_n = ni_pi->cac_data.lts_truncate_n;
+
+	ret = rv770_copy_bytes_to_smc(rdev, ni_pi->cac_table_start, (u8 *)cac_tables,
+				      sizeof(PP_NIslands_CACTABLES), pi->sram_end);
+
+done_free:
+	if (ret) {
+		ni_pi->enable_cac = false;
+		ni_pi->enable_power_containment = false;
+	}
+
+	kfree(cac_tables);
+
+	return 0;
+}
+
+static int ni_initialize_hardware_cac_manager(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	u32 reg;
+
+	if (!ni_pi->enable_cac ||
+	    !ni_pi->cac_configuration_required)
+		return 0;
+
+	if (ni_pi->cac_weights == NULL)
+		return -EINVAL;
+
+	reg = RREG32_CG(CG_CAC_REGION_1_WEIGHT_0) & ~(WEIGHT_TCP_SIG0_MASK |
+						      WEIGHT_TCP_SIG1_MASK |
+						      WEIGHT_TA_SIG_MASK);
+	reg |= (WEIGHT_TCP_SIG0(ni_pi->cac_weights->weight_tcp_sig0) |
+		WEIGHT_TCP_SIG1(ni_pi->cac_weights->weight_tcp_sig1) |
+		WEIGHT_TA_SIG(ni_pi->cac_weights->weight_ta_sig));
+	WREG32_CG(CG_CAC_REGION_1_WEIGHT_0, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_1_WEIGHT_1) & ~(WEIGHT_TCC_EN0_MASK |
+						      WEIGHT_TCC_EN1_MASK |
+						      WEIGHT_TCC_EN2_MASK);
+	reg |= (WEIGHT_TCC_EN0(ni_pi->cac_weights->weight_tcc_en0) |
+		WEIGHT_TCC_EN1(ni_pi->cac_weights->weight_tcc_en1) |
+		WEIGHT_TCC_EN2(ni_pi->cac_weights->weight_tcc_en2));
+	WREG32_CG(CG_CAC_REGION_1_WEIGHT_1, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_2_WEIGHT_0) & ~(WEIGHT_CB_EN0_MASK |
+						      WEIGHT_CB_EN1_MASK |
+						      WEIGHT_CB_EN2_MASK |
+						      WEIGHT_CB_EN3_MASK);
+	reg |= (WEIGHT_CB_EN0(ni_pi->cac_weights->weight_cb_en0) |
+		WEIGHT_CB_EN1(ni_pi->cac_weights->weight_cb_en1) |
+		WEIGHT_CB_EN2(ni_pi->cac_weights->weight_cb_en2) |
+		WEIGHT_CB_EN3(ni_pi->cac_weights->weight_cb_en3));
+	WREG32_CG(CG_CAC_REGION_2_WEIGHT_0, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_2_WEIGHT_1) & ~(WEIGHT_DB_SIG0_MASK |
+						      WEIGHT_DB_SIG1_MASK |
+						      WEIGHT_DB_SIG2_MASK |
+						      WEIGHT_DB_SIG3_MASK);
+	reg |= (WEIGHT_DB_SIG0(ni_pi->cac_weights->weight_db_sig0) |
+		WEIGHT_DB_SIG1(ni_pi->cac_weights->weight_db_sig1) |
+		WEIGHT_DB_SIG2(ni_pi->cac_weights->weight_db_sig2) |
+		WEIGHT_DB_SIG3(ni_pi->cac_weights->weight_db_sig3));
+	WREG32_CG(CG_CAC_REGION_2_WEIGHT_1, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_2_WEIGHT_2) & ~(WEIGHT_SXM_SIG0_MASK |
+						      WEIGHT_SXM_SIG1_MASK |
+						      WEIGHT_SXM_SIG2_MASK |
+						      WEIGHT_SXS_SIG0_MASK |
+						      WEIGHT_SXS_SIG1_MASK);
+	reg |= (WEIGHT_SXM_SIG0(ni_pi->cac_weights->weight_sxm_sig0) |
+		WEIGHT_SXM_SIG1(ni_pi->cac_weights->weight_sxm_sig1) |
+		WEIGHT_SXM_SIG2(ni_pi->cac_weights->weight_sxm_sig2) |
+		WEIGHT_SXS_SIG0(ni_pi->cac_weights->weight_sxs_sig0) |
+		WEIGHT_SXS_SIG1(ni_pi->cac_weights->weight_sxs_sig1));
+	WREG32_CG(CG_CAC_REGION_2_WEIGHT_2, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_3_WEIGHT_0) & ~(WEIGHT_XBR_0_MASK |
+						      WEIGHT_XBR_1_MASK |
+						      WEIGHT_XBR_2_MASK |
+						      WEIGHT_SPI_SIG0_MASK);
+	reg |= (WEIGHT_XBR_0(ni_pi->cac_weights->weight_xbr_0) |
+		WEIGHT_XBR_1(ni_pi->cac_weights->weight_xbr_1) |
+		WEIGHT_XBR_2(ni_pi->cac_weights->weight_xbr_2) |
+		WEIGHT_SPI_SIG0(ni_pi->cac_weights->weight_spi_sig0));
+	WREG32_CG(CG_CAC_REGION_3_WEIGHT_0, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_3_WEIGHT_1) & ~(WEIGHT_SPI_SIG1_MASK |
+						      WEIGHT_SPI_SIG2_MASK |
+						      WEIGHT_SPI_SIG3_MASK |
+						      WEIGHT_SPI_SIG4_MASK |
+						      WEIGHT_SPI_SIG5_MASK);
+	reg |= (WEIGHT_SPI_SIG1(ni_pi->cac_weights->weight_spi_sig1) |
+		WEIGHT_SPI_SIG2(ni_pi->cac_weights->weight_spi_sig2) |
+		WEIGHT_SPI_SIG3(ni_pi->cac_weights->weight_spi_sig3) |
+		WEIGHT_SPI_SIG4(ni_pi->cac_weights->weight_spi_sig4) |
+		WEIGHT_SPI_SIG5(ni_pi->cac_weights->weight_spi_sig5));
+	WREG32_CG(CG_CAC_REGION_3_WEIGHT_1, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_4_WEIGHT_0) & ~(WEIGHT_LDS_SIG0_MASK |
+						      WEIGHT_LDS_SIG1_MASK |
+						      WEIGHT_SC_MASK);
+	reg |= (WEIGHT_LDS_SIG0(ni_pi->cac_weights->weight_lds_sig0) |
+		WEIGHT_LDS_SIG1(ni_pi->cac_weights->weight_lds_sig1) |
+		WEIGHT_SC(ni_pi->cac_weights->weight_sc));
+	WREG32_CG(CG_CAC_REGION_4_WEIGHT_0, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_4_WEIGHT_1) & ~(WEIGHT_BIF_MASK |
+						      WEIGHT_CP_MASK |
+						      WEIGHT_PA_SIG0_MASK |
+						      WEIGHT_PA_SIG1_MASK |
+						      WEIGHT_VGT_SIG0_MASK);
+	reg |= (WEIGHT_BIF(ni_pi->cac_weights->weight_bif) |
+		WEIGHT_CP(ni_pi->cac_weights->weight_cp) |
+		WEIGHT_PA_SIG0(ni_pi->cac_weights->weight_pa_sig0) |
+		WEIGHT_PA_SIG1(ni_pi->cac_weights->weight_pa_sig1) |
+		WEIGHT_VGT_SIG0(ni_pi->cac_weights->weight_vgt_sig0));
+	WREG32_CG(CG_CAC_REGION_4_WEIGHT_1, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_4_WEIGHT_2) & ~(WEIGHT_VGT_SIG1_MASK |
+						      WEIGHT_VGT_SIG2_MASK |
+						      WEIGHT_DC_SIG0_MASK |
+						      WEIGHT_DC_SIG1_MASK |
+						      WEIGHT_DC_SIG2_MASK);
+	reg |= (WEIGHT_VGT_SIG1(ni_pi->cac_weights->weight_vgt_sig1) |
+		WEIGHT_VGT_SIG2(ni_pi->cac_weights->weight_vgt_sig2) |
+		WEIGHT_DC_SIG0(ni_pi->cac_weights->weight_dc_sig0) |
+		WEIGHT_DC_SIG1(ni_pi->cac_weights->weight_dc_sig1) |
+		WEIGHT_DC_SIG2(ni_pi->cac_weights->weight_dc_sig2));
+	WREG32_CG(CG_CAC_REGION_4_WEIGHT_2, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_4_WEIGHT_3) & ~(WEIGHT_DC_SIG3_MASK |
+						      WEIGHT_UVD_SIG0_MASK |
+						      WEIGHT_UVD_SIG1_MASK |
+						      WEIGHT_SPARE0_MASK |
+						      WEIGHT_SPARE1_MASK);
+	reg |= (WEIGHT_DC_SIG3(ni_pi->cac_weights->weight_dc_sig3) |
+		WEIGHT_UVD_SIG0(ni_pi->cac_weights->weight_uvd_sig0) |
+		WEIGHT_UVD_SIG1(ni_pi->cac_weights->weight_uvd_sig1) |
+		WEIGHT_SPARE0(ni_pi->cac_weights->weight_spare0) |
+		WEIGHT_SPARE1(ni_pi->cac_weights->weight_spare1));
+	WREG32_CG(CG_CAC_REGION_4_WEIGHT_3, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_5_WEIGHT_0) & ~(WEIGHT_SQ_VSP_MASK |
+						      WEIGHT_SQ_VSP0_MASK);
+	reg |= (WEIGHT_SQ_VSP(ni_pi->cac_weights->weight_sq_vsp) |
+		WEIGHT_SQ_VSP0(ni_pi->cac_weights->weight_sq_vsp0));
+	WREG32_CG(CG_CAC_REGION_5_WEIGHT_0, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_5_WEIGHT_1) & ~(WEIGHT_SQ_GPR_MASK);
+	reg |= WEIGHT_SQ_GPR(ni_pi->cac_weights->weight_sq_gpr);
+	WREG32_CG(CG_CAC_REGION_5_WEIGHT_1, reg);
+
+	reg = RREG32_CG(CG_CAC_REGION_4_OVERRIDE_4) & ~(OVR_MODE_SPARE_0_MASK |
+							OVR_VAL_SPARE_0_MASK |
+							OVR_MODE_SPARE_1_MASK |
+							OVR_VAL_SPARE_1_MASK);
+	reg |= (OVR_MODE_SPARE_0(ni_pi->cac_weights->ovr_mode_spare_0) |
+		OVR_VAL_SPARE_0(ni_pi->cac_weights->ovr_val_spare_0) |
+		OVR_MODE_SPARE_1(ni_pi->cac_weights->ovr_mode_spare_1) |
+		OVR_VAL_SPARE_1(ni_pi->cac_weights->ovr_val_spare_1));
+	WREG32_CG(CG_CAC_REGION_4_OVERRIDE_4, reg);
+
+	reg = RREG32(SQ_CAC_THRESHOLD) & ~(VSP_MASK |
+					   VSP0_MASK |
+					   GPR_MASK);
+	reg |= (VSP(ni_pi->cac_weights->vsp) |
+		VSP0(ni_pi->cac_weights->vsp0) |
+		GPR(ni_pi->cac_weights->gpr));
+	WREG32(SQ_CAC_THRESHOLD, reg);
+
+	reg = (MCDW_WR_ENABLE |
+	       MCDX_WR_ENABLE |
+	       MCDY_WR_ENABLE |
+	       MCDZ_WR_ENABLE |
+	       INDEX(0x09D4));
+	WREG32(MC_CG_CONFIG, reg);
+
+	reg = (READ_WEIGHT(ni_pi->cac_weights->mc_read_weight) |
+	       WRITE_WEIGHT(ni_pi->cac_weights->mc_write_weight) |
+	       ALLOW_OVERFLOW);
+	WREG32(MC_CG_DATAPORT, reg);
+
+	return 0;
+}
+
+static int ni_enable_smc_cac(struct radeon_device *rdev,
+			     struct radeon_ps *radeon_new_state,
+			     bool enable)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	int ret = 0;
+	PPSMC_Result smc_result;
+
+	if (ni_pi->enable_cac) {
+		if (enable) {
+			if (!r600_is_uvd_state(radeon_new_state->class, radeon_new_state->class2)) {
+				smc_result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_CollectCAC_PowerCorreln);
+
+				if (ni_pi->support_cac_long_term_average) {
+					smc_result = rv770_send_msg_to_smc(rdev, PPSMC_CACLongTermAvgEnable);
+					if (PPSMC_Result_OK != smc_result)
+						ni_pi->support_cac_long_term_average = false;
+				}
+
+				smc_result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableCac);
+				if (PPSMC_Result_OK != smc_result)
+					ret = -EINVAL;
+
+				ni_pi->cac_enabled = (PPSMC_Result_OK == smc_result) ? true : false;
+			}
+		} else if (ni_pi->cac_enabled) {
+			smc_result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_DisableCac);
+
+			ni_pi->cac_enabled = false;
+
+			if (ni_pi->support_cac_long_term_average) {
+				smc_result = rv770_send_msg_to_smc(rdev, PPSMC_CACLongTermAvgDisable);
+				if (PPSMC_Result_OK != smc_result)
+					ni_pi->support_cac_long_term_average = false;
+			}
+		}
+	}
+
+	return ret;
+}
+
+static int ni_pcie_performance_request(struct radeon_device *rdev,
+				       u8 perf_req, bool advertise)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+#if defined(CONFIG_ACPI)
+	if ((perf_req == PCIE_PERF_REQ_PECI_GEN1) ||
+            (perf_req == PCIE_PERF_REQ_PECI_GEN2)) {
+		if (eg_pi->pcie_performance_request_registered == false)
+			radeon_acpi_pcie_notify_device_ready(rdev);
+		eg_pi->pcie_performance_request_registered = true;
+		return radeon_acpi_pcie_performance_request(rdev, perf_req, advertise);
+	} else if ((perf_req == PCIE_PERF_REQ_REMOVE_REGISTRY) &&
+                   eg_pi->pcie_performance_request_registered) {
+		eg_pi->pcie_performance_request_registered = false;
+		return radeon_acpi_pcie_performance_request(rdev, perf_req, advertise);
+	}
+#endif
+	return 0;
+}
+
+static int ni_advertise_gen2_capability(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp;
+
+        tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+
+        if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+            (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2))
+                pi->pcie_gen2 = true;
+        else
+		pi->pcie_gen2 = false;
+
+	if (!pi->pcie_gen2)
+		ni_pcie_performance_request(rdev, PCIE_PERF_REQ_PECI_GEN2, true);
+
+	return 0;
+}
+
+static void ni_enable_bif_dynamic_pcie_gen2(struct radeon_device *rdev,
+					    bool enable)
+{
+        struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+        u32 tmp, bif;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+
+	if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+	    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
+		if (enable) {
+			if (!pi->boot_in_gen2) {
+				bif = RREG32(CG_BIF_REQ_AND_RSP) & ~CG_CLIENT_REQ_MASK;
+				bif |= CG_CLIENT_REQ(0xd);
+				WREG32(CG_BIF_REQ_AND_RSP, bif);
+			}
+			tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+			tmp |= LC_HW_VOLTAGE_IF_CONTROL(1);
+			tmp |= LC_GEN2_EN_STRAP;
+
+			tmp |= LC_CLR_FAILED_SPD_CHANGE_CNT;
+			WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+			udelay(10);
+			tmp &= ~LC_CLR_FAILED_SPD_CHANGE_CNT;
+			WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+		} else {
+			if (!pi->boot_in_gen2) {
+				bif = RREG32(CG_BIF_REQ_AND_RSP) & ~CG_CLIENT_REQ_MASK;
+				bif |= CG_CLIENT_REQ(0xd);
+				WREG32(CG_BIF_REQ_AND_RSP, bif);
+
+				tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+				tmp &= ~LC_GEN2_EN_STRAP;
+			}
+			WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+		}
+	}
+}
+
+static void ni_enable_dynamic_pcie_gen2(struct radeon_device *rdev,
+					bool enable)
+{
+	ni_enable_bif_dynamic_pcie_gen2(rdev, enable);
+
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, ENABLE_GEN2PCIE, ~ENABLE_GEN2PCIE);
+	else
+                WREG32_P(GENERAL_PWRMGT, 0, ~ENABLE_GEN2PCIE);
+}
+
+void ni_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+					   struct radeon_ps *new_ps,
+					   struct radeon_ps *old_ps)
+{
+	struct ni_ps *new_state = ni_get_ps(new_ps);
+	struct ni_ps *current_state = ni_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->performance_levels[new_state->performance_level_count - 1].sclk >=
+	    current_state->performance_levels[current_state->performance_level_count - 1].sclk)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+void ni_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+					  struct radeon_ps *new_ps,
+					  struct radeon_ps *old_ps)
+{
+	struct ni_ps *new_state = ni_get_ps(new_ps);
+	struct ni_ps *current_state = ni_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->performance_levels[new_state->performance_level_count - 1].sclk <
+	    current_state->performance_levels[current_state->performance_level_count - 1].sclk)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+void ni_dpm_setup_asic(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	ni_read_clock_registers(rdev);
+	btc_read_arb_registers(rdev);
+	rv770_get_memory_type(rdev);
+	if (eg_pi->pcie_performance_request)
+		ni_advertise_gen2_capability(rdev);
+	rv770_get_pcie_gen2_status(rdev);
+	rv770_enable_acpi_pm(rdev);
+}
+
+void ni_update_current_ps(struct radeon_device *rdev,
+			  struct radeon_ps *rps)
+{
+	struct ni_ps *new_ps = ni_get_ps(rps);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+
+	eg_pi->current_rps = *rps;
+	ni_pi->current_ps = *new_ps;
+	eg_pi->current_rps.ps_priv = &ni_pi->current_ps;
+}
+
+void ni_update_requested_ps(struct radeon_device *rdev,
+			    struct radeon_ps *rps)
+{
+	struct ni_ps *new_ps = ni_get_ps(rps);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+        struct ni_power_info *ni_pi = ni_get_pi(rdev);
+
+	eg_pi->requested_rps = *rps;
+	ni_pi->requested_ps = *new_ps;
+	eg_pi->requested_rps.ps_priv = &ni_pi->requested_ps;
+}
+
+int ni_dpm_enable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	if (pi->gfx_clock_gating)
+		ni_cg_clockgating_default(rdev);
+        if (btc_dpm_enabled(rdev))
+                return -EINVAL;
+	if (pi->mg_clock_gating)
+		ni_mg_clockgating_default(rdev);
+	if (eg_pi->ls_clock_gating)
+		ni_ls_clockgating_default(rdev);
+	if (pi->voltage_control) {
+		rv770_enable_voltage_control(rdev, true);
+		ret = cypress_construct_voltage_tables(rdev);
+		if (ret) {
+			DRM_ERROR("cypress_construct_voltage_tables failed\n");
+			return ret;
+		}
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = ni_initialize_mc_reg_table(rdev);
+		if (ret)
+			eg_pi->dynamic_ac_timing = false;
+	}
+	if (pi->dynamic_ss)
+		cypress_enable_spread_spectrum(rdev, true);
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, true);
+	rv770_setup_bsp(rdev);
+	rv770_program_git(rdev);
+	rv770_program_tp(rdev);
+	rv770_program_tpp(rdev);
+	rv770_program_sstp(rdev);
+	cypress_enable_display_gap(rdev);
+	rv770_program_vc(rdev);
+	if (pi->dynamic_pcie_gen2)
+		ni_enable_dynamic_pcie_gen2(rdev, true);
+	ret = rv770_upload_firmware(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_upload_firmware failed\n");
+		return ret;
+	}
+	ret = ni_process_firmware_header(rdev);
+	if (ret) {
+		DRM_ERROR("ni_process_firmware_header failed\n");
+		return ret;
+	}
+	ret = ni_initial_switch_from_arb_f0_to_f1(rdev);
+	if (ret) {
+		DRM_ERROR("ni_initial_switch_from_arb_f0_to_f1 failed\n");
+		return ret;
+	}
+	ret = ni_init_smc_table(rdev);
+	if (ret) {
+		DRM_ERROR("ni_init_smc_table failed\n");
+		return ret;
+	}
+	ret = ni_init_smc_spll_table(rdev);
+	if (ret) {
+		DRM_ERROR("ni_init_smc_spll_table failed\n");
+		return ret;
+	}
+	ret = ni_init_arb_table_index(rdev);
+	if (ret) {
+		DRM_ERROR("ni_init_arb_table_index failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = ni_populate_mc_reg_table(rdev, boot_ps);
+		if (ret) {
+			DRM_ERROR("ni_populate_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+	ret = ni_initialize_smc_cac_tables(rdev);
+	if (ret) {
+		DRM_ERROR("ni_initialize_smc_cac_tables failed\n");
+		return ret;
+	}
+	ret = ni_initialize_hardware_cac_manager(rdev);
+	if (ret) {
+		DRM_ERROR("ni_initialize_hardware_cac_manager failed\n");
+		return ret;
+	}
+	ret = ni_populate_smc_tdp_limits(rdev, boot_ps);
+	if (ret) {
+		DRM_ERROR("ni_populate_smc_tdp_limits failed\n");
+		return ret;
+	}
+	ni_program_response_times(rdev);
+	r7xx_start_smc(rdev);
+	ret = cypress_notify_smc_display_change(rdev, false);
+	if (ret) {
+		DRM_ERROR("cypress_notify_smc_display_change failed\n");
+		return ret;
+	}
+	cypress_enable_sclk_control(rdev, true);
+	if (eg_pi->memory_transition)
+		cypress_enable_mclk_control(rdev, true);
+	cypress_start_dpm(rdev);
+	if (pi->gfx_clock_gating)
+		ni_gfx_clockgating_enable(rdev, true);
+	if (pi->mg_clock_gating)
+		ni_mg_clockgating_enable(rdev, true);
+	if (eg_pi->ls_clock_gating)
+		ni_ls_clockgating_enable(rdev, true);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		PPSMC_Result result;
+
+		ret = rv770_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, 0xff * 1000);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+		result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+
+		if (result != PPSMC_Result_OK)
+			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
+	}
+
+	rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
+
+	ni_update_current_ps(rdev, boot_ps);
+
+	return 0;
+}
+
+void ni_dpm_disable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+
+	if (!btc_dpm_enabled(rdev))
+		return;
+	rv770_clear_vc(rdev);
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, false);
+	ni_enable_power_containment(rdev, boot_ps, false);
+	ni_enable_smc_cac(rdev, boot_ps, false);
+	cypress_enable_spread_spectrum(rdev, false);
+	rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, false);
+	if (pi->dynamic_pcie_gen2)
+		ni_enable_dynamic_pcie_gen2(rdev, false);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	if (pi->gfx_clock_gating)
+		ni_gfx_clockgating_enable(rdev, false);
+	if (pi->mg_clock_gating)
+		ni_mg_clockgating_enable(rdev, false);
+	if (eg_pi->ls_clock_gating)
+		ni_ls_clockgating_enable(rdev, false);
+	ni_stop_dpm(rdev);
+	btc_reset_to_default(rdev);
+	ni_stop_smc(rdev);
+	ni_force_switch_to_arb_f0(rdev);
+
+	ni_update_current_ps(rdev, boot_ps);
+}
+
+static int ni_power_control_set_level(struct radeon_device *rdev)
+{
+	struct radeon_ps *new_ps = rdev->pm.dpm.requested_ps;
+	int ret;
+
+	ret = ni_restrict_performance_levels_before_switch(rdev);
+	if (ret)
+		return ret;
+	ret = rv770_halt_smc(rdev);
+	if (ret)
+		return ret;
+	ret = ni_populate_smc_tdp_limits(rdev, new_ps);
+	if (ret)
+		return ret;
+	ret = rv770_resume_smc(rdev);
+	if (ret)
+		return ret;
+	ret = rv770_set_sw_state(rdev);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+int ni_dpm_pre_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps requested_ps = *rdev->pm.dpm.requested_ps;
+	struct radeon_ps *new_ps = &requested_ps;
+
+	ni_update_requested_ps(rdev, new_ps);
+
+	ni_apply_state_adjust_rules(rdev, &eg_pi->requested_rps);
+
+	return 0;
+}
+
+int ni_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = &eg_pi->requested_rps;
+	struct radeon_ps *old_ps = &eg_pi->current_rps;
+	int ret;
+
+	ret = ni_restrict_performance_levels_before_switch(rdev);
+	if (ret) {
+		DRM_ERROR("ni_restrict_performance_levels_before_switch failed\n");
+		return ret;
+	}
+	ni_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+	ret = ni_enable_power_containment(rdev, new_ps, false);
+	if (ret) {
+		DRM_ERROR("ni_enable_power_containment failed\n");
+		return ret;
+	}
+	ret = ni_enable_smc_cac(rdev, new_ps, false);
+	if (ret) {
+		DRM_ERROR("ni_enable_smc_cac failed\n");
+		return ret;
+	}
+	ret = rv770_halt_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_halt_smc failed\n");
+		return ret;
+	}
+	if (eg_pi->smu_uvd_hs)
+		btc_notify_uvd_to_smc(rdev, new_ps);
+	ret = ni_upload_sw_state(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("ni_upload_sw_state failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = ni_upload_mc_reg_table(rdev, new_ps);
+		if (ret) {
+			DRM_ERROR("ni_upload_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+	ret = ni_program_memory_timing_parameters(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("ni_program_memory_timing_parameters failed\n");
+		return ret;
+	}
+	ret = rv770_resume_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_resume_smc failed\n");
+		return ret;
+	}
+	ret = rv770_set_sw_state(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_set_sw_state failed\n");
+		return ret;
+	}
+	ni_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+	ret = ni_enable_smc_cac(rdev, new_ps, true);
+	if (ret) {
+		DRM_ERROR("ni_enable_smc_cac failed\n");
+		return ret;
+	}
+	ret = ni_enable_power_containment(rdev, new_ps, true);
+	if (ret) {
+		DRM_ERROR("ni_enable_power_containment failed\n");
+		return ret;
+	}
+
+	/* update tdp */
+	ret = ni_power_control_set_level(rdev);
+	if (ret) {
+		DRM_ERROR("ni_power_control_set_level failed\n");
+		return ret;
+	}
+
+	ret = ni_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO);
+	if (ret) {
+		DRM_ERROR("ni_dpm_force_performance_level failed\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+void ni_dpm_post_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = &eg_pi->requested_rps;
+
+	ni_update_current_ps(rdev, new_ps);
+}
+
+void ni_dpm_reset_asic(struct radeon_device *rdev)
+{
+	ni_restrict_performance_levels_before_switch(rdev);
+	rv770_set_boot_state(rdev);
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void ni_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					  struct radeon_ps *rps,
+					  struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info,
+					  u8 table_rev)
+{
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) {
+		rps->vclk = le32_to_cpu(non_clock_info->ulVCLK);
+		rps->dclk = le32_to_cpu(non_clock_info->ulDCLK);
+	} else if (r600_is_uvd_state(rps->class, rps->class2)) {
+		rps->vclk = RV770_DEFAULT_VCLK_FREQ;
+		rps->dclk = RV770_DEFAULT_DCLK_FREQ;
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT)
+		rdev->pm.dpm.boot_ps = rps;
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void ni_parse_pplib_clock_info(struct radeon_device *rdev,
+				      struct radeon_ps *rps, int index,
+				      union pplib_clock_info *clock_info)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_ps *ps = ni_get_ps(rps);
+	u16 vddc;
+	struct rv7xx_pl *pl = &ps->performance_levels[index];
+
+	ps->performance_level_count = index + 1;
+
+	pl->sclk = le16_to_cpu(clock_info->evergreen.usEngineClockLow);
+	pl->sclk |= clock_info->evergreen.ucEngineClockHigh << 16;
+	pl->mclk = le16_to_cpu(clock_info->evergreen.usMemoryClockLow);
+	pl->mclk |= clock_info->evergreen.ucMemoryClockHigh << 16;
+
+	pl->vddc = le16_to_cpu(clock_info->evergreen.usVDDC);
+	pl->vddci = le16_to_cpu(clock_info->evergreen.usVDDCI);
+	pl->flags = le32_to_cpu(clock_info->evergreen.ulFlags);
+
+	/* patch up vddc if necessary */
+	if (pl->vddc == 0xff01) {
+		if (radeon_atom_get_max_vddc(rdev, 0, 0, &vddc) == 0)
+			pl->vddc = vddc;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_ACPI) {
+		pi->acpi_vddc = pl->vddc;
+		eg_pi->acpi_vddci = pl->vddci;
+		if (ps->performance_levels[0].flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2)
+			pi->acpi_pcie_gen2 = true;
+		else
+			pi->acpi_pcie_gen2 = false;
+	}
+
+	if (rps->class2 & ATOM_PPLIB_CLASSIFICATION2_ULV) {
+		eg_pi->ulv.supported = true;
+		eg_pi->ulv.pl = pl;
+	}
+
+	if (pi->min_vddc_in_table > pl->vddc)
+		pi->min_vddc_in_table = pl->vddc;
+
+	if (pi->max_vddc_in_table < pl->vddc)
+		pi->max_vddc_in_table = pl->vddc;
+
+	/* patch up boot state */
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		u16 vddc, vddci, mvdd;
+		radeon_atombios_get_default_voltages(rdev, &vddc, &vddci, &mvdd);
+		pl->mclk = rdev->clock.default_mclk;
+		pl->sclk = rdev->clock.default_sclk;
+		pl->vddc = vddc;
+		pl->vddci = vddci;
+	}
+
+	if ((rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) ==
+	    ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE) {
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk = pl->sclk;
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.mclk = pl->mclk;
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddc = pl->vddc;
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddci = pl->vddci;
+	}
+}
+
+static int ni_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i, j;
+	union pplib_clock_info *clock_info;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	struct ni_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  power_info->pplib.ucNumStates, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+
+	for (i = 0; i < power_info->pplib.ucNumStates; i++) {
+		power_state = (union pplib_power_state *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usStateArrayOffset) +
+			 i * power_info->pplib.ucStateEntrySize);
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset) +
+			 (power_state->v1.ucNonClockStateIndex *
+			  power_info->pplib.ucNonClockSize));
+		if (power_info->pplib.ucStateEntrySize - 1) {
+			ps = kzalloc(sizeof(struct ni_ps), GFP_KERNEL);
+			if (ps == NULL) {
+				kfree(rdev->pm.dpm.ps);
+				return -ENOMEM;
+			}
+			rdev->pm.dpm.ps[i].ps_priv = ps;
+			ni_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+							 non_clock_info,
+							 power_info->pplib.ucNonClockSize);
+			for (j = 0; j < (power_info->pplib.ucStateEntrySize - 1); j++) {
+				clock_info = (union pplib_clock_info *)
+					(mode_info->atom_context->bios + data_offset +
+					 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) +
+					 (power_state->v1.ucClockStateIndices[j] *
+					  power_info->pplib.ucClockInfoSize));
+				ni_parse_pplib_clock_info(rdev,
+							  &rdev->pm.dpm.ps[i], j,
+							  clock_info);
+			}
+		}
+	}
+	rdev->pm.dpm.num_ps = power_info->pplib.ucNumStates;
+	return 0;
+}
+
+int ni_dpm_init(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi;
+	struct evergreen_power_info *eg_pi;
+	struct ni_power_info *ni_pi;
+	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	u16 data_offset, size;
+	u8 frev, crev;
+	struct atom_clock_dividers dividers;
+	int ret;
+
+	ni_pi = kzalloc(sizeof(struct ni_power_info), GFP_KERNEL);
+	if (ni_pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = ni_pi;
+	eg_pi = &ni_pi->eg;
+	pi = &eg_pi->rv7xx;
+
+	rv770_get_max_vddc(rdev);
+
+	eg_pi->ulv.supported = false;
+	pi->acpi_vddc = 0;
+	eg_pi->acpi_vddci = 0;
+	pi->min_vddc_in_table = 0;
+	pi->max_vddc_in_table = 0;
+
+	ret = ni_parse_power_table(rdev);
+	if (ret)
+		return ret;
+	ret = r600_parse_extended_power_table(rdev);
+	if (ret)
+		return ret;
+
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries =
+		kzalloc(4 * sizeof(struct radeon_clock_voltage_dependency_entry), GFP_KERNEL);
+	if (!rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries) {
+		r600_free_extended_power_table(rdev);
+		return -ENOMEM;
+	}
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.count = 4;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].clk = 0;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].v = 0;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].clk = 36000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].v = 720;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].clk = 54000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].v = 810;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].clk = 72000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].v = 900;
+
+	ni_patch_dependency_tables_based_on_leakage(rdev);
+
+	if (rdev->pm.dpm.voltage_response_time == 0)
+		rdev->pm.dpm.voltage_response_time = R600_VOLTAGERESPONSETIME_DFLT;
+	if (rdev->pm.dpm.backbias_response_time == 0)
+		rdev->pm.dpm.backbias_response_time = R600_BACKBIASRESPONSETIME_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->ref_div = dividers.ref_div + 1;
+	else
+		pi->ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	pi->rlp = RV770_RLP_DFLT;
+	pi->rmp = RV770_RMP_DFLT;
+	pi->lhp = RV770_LHP_DFLT;
+	pi->lmp = RV770_LMP_DFLT;
+
+	eg_pi->ats[0].rlp = RV770_RLP_DFLT;
+	eg_pi->ats[0].rmp = RV770_RMP_DFLT;
+	eg_pi->ats[0].lhp = RV770_LHP_DFLT;
+	eg_pi->ats[0].lmp = RV770_LMP_DFLT;
+
+	eg_pi->ats[1].rlp = BTC_RLP_UVD_DFLT;
+	eg_pi->ats[1].rmp = BTC_RMP_UVD_DFLT;
+	eg_pi->ats[1].lhp = BTC_LHP_UVD_DFLT;
+	eg_pi->ats[1].lmp = BTC_LMP_UVD_DFLT;
+
+	eg_pi->smu_uvd_hs = true;
+
+	if (rdev->pdev->device == 0x6707) {
+		pi->mclk_strobe_mode_threshold = 55000;
+		pi->mclk_edc_enable_threshold = 55000;
+		eg_pi->mclk_edc_wr_enable_threshold = 55000;
+	} else {
+		pi->mclk_strobe_mode_threshold = 40000;
+		pi->mclk_edc_enable_threshold = 40000;
+		eg_pi->mclk_edc_wr_enable_threshold = 40000;
+	}
+	ni_pi->mclk_rtt_mode_threshold = eg_pi->mclk_edc_wr_enable_threshold;
+
+	pi->voltage_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, 0);
+
+	pi->mvdd_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, 0);
+
+	eg_pi->vddci_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, 0);
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+                                   &frev, &crev, &data_offset)) {
+		pi->sclk_ss = true;
+		pi->mclk_ss = true;
+		pi->dynamic_ss = true;
+	} else {
+		pi->sclk_ss = false;
+		pi->mclk_ss = false;
+		pi->dynamic_ss = true;
+	}
+
+	pi->asi = RV770_ASI_DFLT;
+	pi->pasi = CYPRESS_HASI_DFLT;
+	pi->vrc = CYPRESS_VRC_DFLT;
+
+	pi->power_gating = false;
+
+	pi->gfx_clock_gating = true;
+
+	pi->mg_clock_gating = true;
+	pi->mgcgtssm = true;
+	eg_pi->ls_clock_gating = false;
+	eg_pi->sclk_deep_sleep = false;
+
+	pi->dynamic_pcie_gen2 = true;
+
+	if (pi->gfx_clock_gating &&
+	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+		pi->thermal_protection = true;
+	else
+		pi->thermal_protection = false;
+
+	pi->display_gap = true;
+
+	pi->dcodt = true;
+
+	pi->ulps = true;
+
+	eg_pi->dynamic_ac_timing = true;
+	eg_pi->abm = true;
+	eg_pi->mcls = true;
+	eg_pi->light_sleep = true;
+	eg_pi->memory_transition = true;
+#if defined(CONFIG_ACPI)
+	eg_pi->pcie_performance_request =
+		radeon_acpi_is_pcie_performance_request_supported(rdev);
+#else
+	eg_pi->pcie_performance_request = false;
+#endif
+
+	eg_pi->dll_default_on = false;
+
+	eg_pi->sclk_deep_sleep = false;
+
+	pi->mclk_stutter_mode_threshold = 0;
+
+	pi->sram_end = SMC_RAM_END;
+
+	rdev->pm.dpm.dyn_state.mclk_sclk_ratio = 3;
+	rdev->pm.dpm.dyn_state.vddc_vddci_delta = 200;
+	rdev->pm.dpm.dyn_state.min_vddc_for_pcie_gen2 = 900;
+	rdev->pm.dpm.dyn_state.valid_sclk_values.count = ARRAY_SIZE(btc_valid_sclk);
+	rdev->pm.dpm.dyn_state.valid_sclk_values.values = btc_valid_sclk;
+	rdev->pm.dpm.dyn_state.valid_mclk_values.count = 0;
+	rdev->pm.dpm.dyn_state.valid_mclk_values.values = NULL;
+	rdev->pm.dpm.dyn_state.sclk_mclk_delta = 12500;
+
+	ni_pi->cac_data.leakage_coefficients.at = 516;
+	ni_pi->cac_data.leakage_coefficients.bt = 18;
+	ni_pi->cac_data.leakage_coefficients.av = 51;
+	ni_pi->cac_data.leakage_coefficients.bv = 2957;
+
+	switch (rdev->pdev->device) {
+	case 0x6700:
+	case 0x6701:
+	case 0x6702:
+	case 0x6703:
+	case 0x6718:
+		ni_pi->cac_weights = &cac_weights_cayman_xt;
+		break;
+	case 0x6705:
+	case 0x6719:
+	case 0x671D:
+	case 0x671C:
+	default:
+		ni_pi->cac_weights = &cac_weights_cayman_pro;
+		break;
+	case 0x6704:
+	case 0x6706:
+	case 0x6707:
+	case 0x6708:
+	case 0x6709:
+		ni_pi->cac_weights = &cac_weights_cayman_le;
+		break;
+	}
+
+	if (ni_pi->cac_weights->enable_power_containment_by_default) {
+		ni_pi->enable_power_containment = true;
+		ni_pi->enable_cac = true;
+		ni_pi->enable_sq_ramping = true;
+	} else {
+		ni_pi->enable_power_containment = false;
+		ni_pi->enable_cac = false;
+		ni_pi->enable_sq_ramping = false;
+	}
+
+	ni_pi->driver_calculate_cac_leakage = false;
+	ni_pi->cac_configuration_required = true;
+
+	if (ni_pi->cac_configuration_required) {
+		ni_pi->support_cac_long_term_average = true;
+		ni_pi->lta_window_size = ni_pi->cac_weights->l2_lta_window_size;
+		ni_pi->lts_truncate = ni_pi->cac_weights->lts_truncate;
+	} else {
+		ni_pi->support_cac_long_term_average = false;
+		ni_pi->lta_window_size = 0;
+		ni_pi->lts_truncate = 0;
+	}
+
+	ni_pi->use_power_boost_limit = true;
+
+	return 0;
+}
+
+void ni_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+	kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries);
+	r600_free_extended_power_table(rdev);
+}
+
+void ni_dpm_print_power_state(struct radeon_device *rdev,
+			      struct radeon_ps *rps)
+{
+	struct ni_ps *ps = ni_get_ps(rps);
+	struct rv7xx_pl *pl;
+	int i;
+
+	r600_dpm_print_class_info(rps->class, rps->class2);
+	r600_dpm_print_cap_info(rps->caps);
+	printk("\tuvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+	for (i = 0; i < ps->performance_level_count; i++) {
+		pl = &ps->performance_levels[i];
+		if (rdev->family >= CHIP_TAHITI)
+			printk("\t\tpower level %d    sclk: %u mclk: %u vddc: %u vddci: %u pcie gen: %u\n",
+			       i, pl->sclk, pl->mclk, pl->vddc, pl->vddci, pl->pcie_gen + 1);
+		else
+			printk("\t\tpower level %d    sclk: %u mclk: %u vddc: %u vddci: %u\n",
+			       i, pl->sclk, pl->mclk, pl->vddc, pl->vddci);
+	}
+	r600_dpm_print_ps_status(rdev, rps);
+}
+
+void ni_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						    struct seq_file *m)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct ni_ps *ps = ni_get_ps(rps);
+	struct rv7xx_pl *pl;
+	u32 current_index =
+		(RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_STATE_INDEX_MASK) >>
+		CURRENT_STATE_INDEX_SHIFT;
+
+	if (current_index >= ps->performance_level_count) {
+		seq_printf(m, "invalid dpm profile %d\n", current_index);
+	} else {
+		pl = &ps->performance_levels[current_index];
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		seq_printf(m, "power level %d    sclk: %u mclk: %u vddc: %u vddci: %u\n",
+			   current_index, pl->sclk, pl->mclk, pl->vddc, pl->vddci);
+	}
+}
+
+u32 ni_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_ps *requested_state = ni_get_ps(&eg_pi->requested_rps);
+
+	if (low)
+		return requested_state->performance_levels[0].sclk;
+	else
+		return requested_state->performance_levels[requested_state->performance_level_count - 1].sclk;
+}
+
+u32 ni_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_ps *requested_state = ni_get_ps(&eg_pi->requested_rps);
+
+	if (low)
+		return requested_state->performance_levels[0].mclk;
+	else
+		return requested_state->performance_levels[requested_state->performance_level_count - 1].mclk;
+}
+
diff --git a/drivers/gpu/drm/radeon/ni_dpm.h b/drivers/gpu/drm/radeon/ni_dpm.h
new file mode 100644
index 000000000000..6bbee9180909
--- /dev/null
+++ b/drivers/gpu/drm/radeon/ni_dpm.h
@@ -0,0 +1,250 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __NI_DPM_H__
+#define __NI_DPM_H__
+
+#include "cypress_dpm.h"
+#include "btc_dpm.h"
+#include "nislands_smc.h"
+
+struct ni_clock_registers {
+	u32 cg_spll_func_cntl;
+	u32 cg_spll_func_cntl_2;
+	u32 cg_spll_func_cntl_3;
+	u32 cg_spll_func_cntl_4;
+	u32 cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2;
+	u32 mclk_pwrmgt_cntl;
+	u32 dll_cntl;
+	u32 mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2;
+	u32 mpll_ss1;
+	u32 mpll_ss2;
+};
+
+struct ni_mc_reg_entry {
+	u32 mclk_max;
+	u32 mc_data[SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE];
+};
+
+struct ni_mc_reg_table {
+	u8 last;
+	u8 num_entries;
+	u16 valid_flag;
+	struct ni_mc_reg_entry mc_reg_table_entry[MAX_AC_TIMING_ENTRIES];
+	SMC_NIslands_MCRegisterAddress mc_reg_address[SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE];
+};
+
+#define NISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT 2
+
+enum ni_dc_cac_level
+{
+	NISLANDS_DCCAC_LEVEL_0 = 0,
+	NISLANDS_DCCAC_LEVEL_1,
+	NISLANDS_DCCAC_LEVEL_2,
+	NISLANDS_DCCAC_LEVEL_3,
+	NISLANDS_DCCAC_LEVEL_4,
+	NISLANDS_DCCAC_LEVEL_5,
+	NISLANDS_DCCAC_LEVEL_6,
+	NISLANDS_DCCAC_LEVEL_7,
+	NISLANDS_DCCAC_MAX_LEVELS
+};
+
+struct ni_leakage_coeffients
+{
+	u32 at;
+	u32 bt;
+	u32 av;
+	u32 bv;
+	s32 t_slope;
+	s32 t_intercept;
+	u32 t_ref;
+};
+
+struct ni_cac_data
+{
+	struct ni_leakage_coeffients leakage_coefficients;
+	u32 i_leakage;
+	s32 leakage_minimum_temperature;
+	u32 pwr_const;
+	u32 dc_cac_value;
+	u32 bif_cac_value;
+	u32 lkge_pwr;
+	u8 mc_wr_weight;
+	u8 mc_rd_weight;
+	u8 allow_ovrflw;
+	u8 num_win_tdp;
+	u8 l2num_win_tdp;
+	u8 lts_truncate_n;
+};
+
+struct ni_cac_weights
+{
+	u32 weight_tcp_sig0;
+	u32 weight_tcp_sig1;
+	u32 weight_ta_sig;
+	u32 weight_tcc_en0;
+	u32 weight_tcc_en1;
+	u32 weight_tcc_en2;
+	u32 weight_cb_en0;
+	u32 weight_cb_en1;
+	u32 weight_cb_en2;
+	u32 weight_cb_en3;
+	u32 weight_db_sig0;
+	u32 weight_db_sig1;
+	u32 weight_db_sig2;
+	u32 weight_db_sig3;
+	u32 weight_sxm_sig0;
+	u32 weight_sxm_sig1;
+	u32 weight_sxm_sig2;
+	u32 weight_sxs_sig0;
+	u32 weight_sxs_sig1;
+	u32 weight_xbr_0;
+	u32 weight_xbr_1;
+	u32 weight_xbr_2;
+	u32 weight_spi_sig0;
+	u32 weight_spi_sig1;
+	u32 weight_spi_sig2;
+	u32 weight_spi_sig3;
+	u32 weight_spi_sig4;
+	u32 weight_spi_sig5;
+	u32 weight_lds_sig0;
+	u32 weight_lds_sig1;
+	u32 weight_sc;
+	u32 weight_bif;
+	u32 weight_cp;
+	u32 weight_pa_sig0;
+	u32 weight_pa_sig1;
+	u32 weight_vgt_sig0;
+	u32 weight_vgt_sig1;
+	u32 weight_vgt_sig2;
+	u32 weight_dc_sig0;
+	u32 weight_dc_sig1;
+	u32 weight_dc_sig2;
+	u32 weight_dc_sig3;
+	u32 weight_uvd_sig0;
+	u32 weight_uvd_sig1;
+	u32 weight_spare0;
+	u32 weight_spare1;
+	u32 weight_sq_vsp;
+	u32 weight_sq_vsp0;
+	u32 weight_sq_gpr;
+	u32 ovr_mode_spare_0;
+	u32 ovr_val_spare_0;
+	u32 ovr_mode_spare_1;
+	u32 ovr_val_spare_1;
+	u32 vsp;
+	u32 vsp0;
+	u32 gpr;
+	u8 mc_read_weight;
+	u8 mc_write_weight;
+	u32 tid_cnt;
+	u32 tid_unit;
+	u32 l2_lta_window_size;
+	u32 lts_truncate;
+	u32 dc_cac[NISLANDS_DCCAC_MAX_LEVELS];
+	u32 pcie_cac[SMC_NISLANDS_BIF_LUT_NUM_OF_ENTRIES];
+	bool enable_power_containment_by_default;
+};
+
+struct ni_ps {
+	u16 performance_level_count;
+	bool dc_compatible;
+	struct rv7xx_pl performance_levels[NISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE];
+};
+
+struct ni_power_info {
+	/* must be first! */
+	struct evergreen_power_info eg;
+	struct ni_clock_registers clock_registers;
+	struct ni_mc_reg_table mc_reg_table;
+	u32 mclk_rtt_mode_threshold;
+	/* flags */
+	bool use_power_boost_limit;
+	bool support_cac_long_term_average;
+	bool cac_enabled;
+	bool cac_configuration_required;
+	bool driver_calculate_cac_leakage;
+	bool pc_enabled;
+	bool enable_power_containment;
+	bool enable_cac;
+	bool enable_sq_ramping;
+	/* smc offsets */
+	u16 arb_table_start;
+	u16 fan_table_start;
+	u16 cac_table_start;
+	u16 spll_table_start;
+	/* CAC stuff */
+	struct ni_cac_data cac_data;
+	u32 dc_cac_table[NISLANDS_DCCAC_MAX_LEVELS];
+	const struct ni_cac_weights *cac_weights;
+	u8 lta_window_size;
+	u8 lts_truncate;
+	struct ni_ps current_ps;
+	struct ni_ps requested_ps;
+	/* scratch structs */
+	SMC_NIslands_MCRegisters smc_mc_reg_table;
+	NISLANDS_SMC_STATETABLE smc_statetable;
+};
+
+#define NISLANDS_INITIAL_STATE_ARB_INDEX    0
+#define NISLANDS_ACPI_STATE_ARB_INDEX       1
+#define NISLANDS_ULV_STATE_ARB_INDEX        2
+#define NISLANDS_DRIVER_STATE_ARB_INDEX     3
+
+#define NISLANDS_DPM2_MAX_PULSE_SKIP        256
+
+#define NISLANDS_DPM2_NEAR_TDP_DEC          10
+#define NISLANDS_DPM2_ABOVE_SAFE_INC        5
+#define NISLANDS_DPM2_BELOW_SAFE_INC        20
+
+#define NISLANDS_DPM2_TDP_SAFE_LIMIT_PERCENT            80
+
+#define NISLANDS_DPM2_MAXPS_PERCENT_H                   90
+#define NISLANDS_DPM2_MAXPS_PERCENT_M                   0
+
+#define NISLANDS_DPM2_SQ_RAMP_MAX_POWER                 0x3FFF
+#define NISLANDS_DPM2_SQ_RAMP_MIN_POWER                 0x12
+#define NISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA           0x15
+#define NISLANDS_DPM2_SQ_RAMP_STI_SIZE                  0x1E
+#define NISLANDS_DPM2_SQ_RAMP_LTI_RATIO                 0xF
+
+int ni_copy_and_switch_arb_sets(struct radeon_device *rdev,
+				u32 arb_freq_src, u32 arb_freq_dest);
+void ni_update_current_ps(struct radeon_device *rdev,
+			  struct radeon_ps *rps);
+void ni_update_requested_ps(struct radeon_device *rdev,
+			    struct radeon_ps *rps);
+
+void ni_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+					   struct radeon_ps *new_ps,
+					   struct radeon_ps *old_ps);
+void ni_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+					  struct radeon_ps *new_ps,
+					  struct radeon_ps *old_ps);
+
+bool ni_dpm_vblank_too_short(struct radeon_device *rdev);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/nid.h b/drivers/gpu/drm/radeon/nid.h
index e226faf16fea..22421bc80c0d 100644
--- a/drivers/gpu/drm/radeon/nid.h
+++ b/drivers/gpu/drm/radeon/nid.h
@@ -133,6 +133,22 @@
 #define VM_CONTEXT1_CNTL2				0x1434
 #define VM_INVALIDATE_REQUEST				0x1478
 #define VM_INVALIDATE_RESPONSE				0x147c
+#define	VM_CONTEXT1_PROTECTION_FAULT_ADDR		0x14FC
+#define	VM_CONTEXT1_PROTECTION_FAULT_STATUS		0x14DC
+#define		PROTECTIONS_MASK			(0xf << 0)
+#define		PROTECTIONS_SHIFT			0
+		/* bit 0: range
+		 * bit 2: pde0
+		 * bit 3: valid
+		 * bit 4: read
+		 * bit 5: write
+		 */
+#define		MEMORY_CLIENT_ID_MASK			(0xff << 12)
+#define		MEMORY_CLIENT_ID_SHIFT			12
+#define		MEMORY_CLIENT_RW_MASK			(1 << 24)
+#define		MEMORY_CLIENT_RW_SHIFT			24
+#define		FAULT_VMID_MASK				(0x7 << 25)
+#define		FAULT_VMID_SHIFT			25
 #define VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR	0x1518
 #define VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR	0x151c
 #define	VM_CONTEXT0_PAGE_TABLE_BASE_ADDR		0x153C
@@ -489,6 +505,571 @@
 #       define CACHE_FLUSH_AND_INV_EVENT_TS                     (0x14 << 0)
 #       define CACHE_FLUSH_AND_INV_EVENT                        (0x16 << 0)
 
+/* TN SMU registers */
+#define	TN_CURRENT_GNB_TEMP				0x1F390
+
+/* pm registers */
+#define	SMC_MSG						0x20c
+#define		HOST_SMC_MSG(x)				((x) << 0)
+#define		HOST_SMC_MSG_MASK			(0xff << 0)
+#define		HOST_SMC_MSG_SHIFT			0
+#define		HOST_SMC_RESP(x)			((x) << 8)
+#define		HOST_SMC_RESP_MASK			(0xff << 8)
+#define		HOST_SMC_RESP_SHIFT			8
+#define		SMC_HOST_MSG(x)				((x) << 16)
+#define		SMC_HOST_MSG_MASK			(0xff << 16)
+#define		SMC_HOST_MSG_SHIFT			16
+#define		SMC_HOST_RESP(x)			((x) << 24)
+#define		SMC_HOST_RESP_MASK			(0xff << 24)
+#define		SMC_HOST_RESP_SHIFT			24
+
+#define	CG_SPLL_FUNC_CNTL				0x600
+#define		SPLL_RESET				(1 << 0)
+#define		SPLL_SLEEP				(1 << 1)
+#define		SPLL_BYPASS_EN				(1 << 3)
+#define		SPLL_REF_DIV(x)				((x) << 4)
+#define		SPLL_REF_DIV_MASK			(0x3f << 4)
+#define		SPLL_PDIV_A(x)				((x) << 20)
+#define		SPLL_PDIV_A_MASK			(0x7f << 20)
+#define		SPLL_PDIV_A_SHIFT			20
+#define	CG_SPLL_FUNC_CNTL_2				0x604
+#define		SCLK_MUX_SEL(x)				((x) << 0)
+#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_SPLL_FUNC_CNTL_3				0x608
+#define		SPLL_FB_DIV(x)				((x) << 0)
+#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		SPLL_FB_DIV_SHIFT			0
+#define		SPLL_DITHEN				(1 << 28)
+
+#define MPLL_CNTL_MODE                                  0x61c
+#       define SS_SSEN                                  (1 << 24)
+#       define SS_DSMODE_EN                             (1 << 25)
+
+#define	MPLL_AD_FUNC_CNTL				0x624
+#define		CLKF(x)					((x) << 0)
+#define		CLKF_MASK				(0x7f << 0)
+#define		CLKR(x)					((x) << 7)
+#define		CLKR_MASK				(0x1f << 7)
+#define		CLKFRAC(x)				((x) << 12)
+#define		CLKFRAC_MASK				(0x1f << 12)
+#define		YCLK_POST_DIV(x)			((x) << 17)
+#define		YCLK_POST_DIV_MASK			(3 << 17)
+#define		IBIAS(x)				((x) << 20)
+#define		IBIAS_MASK				(0x3ff << 20)
+#define		RESET					(1 << 30)
+#define		PDNB					(1 << 31)
+#define	MPLL_AD_FUNC_CNTL_2				0x628
+#define		BYPASS					(1 << 19)
+#define		BIAS_GEN_PDNB				(1 << 24)
+#define		RESET_EN				(1 << 25)
+#define		VCO_MODE				(1 << 29)
+#define	MPLL_DQ_FUNC_CNTL				0x62c
+#define	MPLL_DQ_FUNC_CNTL_2				0x630
+
+#define GENERAL_PWRMGT                                  0x63c
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+#       define STATIC_PM_EN                             (1 << 1)
+#       define THERMAL_PROTECTION_DIS                   (1 << 2)
+#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
+#       define ENABLE_GEN2PCIE                          (1 << 4)
+#       define ENABLE_GEN2XSP                           (1 << 5)
+#       define SW_SMIO_INDEX(x)                         ((x) << 6)
+#       define SW_SMIO_INDEX_MASK                       (3 << 6)
+#       define SW_SMIO_INDEX_SHIFT                      6
+#       define LOW_VOLT_D2_ACPI                         (1 << 8)
+#       define LOW_VOLT_D3_ACPI                         (1 << 9)
+#       define VOLT_PWRMGT_EN                           (1 << 10)
+#       define BACKBIAS_PAD_EN                          (1 << 18)
+#       define BACKBIAS_VALUE                           (1 << 19)
+#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
+#       define AC_DC_SW                                 (1 << 24)
+
+#define SCLK_PWRMGT_CNTL                                  0x644
+#       define SCLK_PWRMGT_OFF                            (1 << 0)
+#       define SCLK_LOW_D1                                (1 << 1)
+#       define FIR_RESET                                  (1 << 4)
+#       define FIR_FORCE_TREND_SEL                        (1 << 5)
+#       define FIR_TREND_MODE                             (1 << 6)
+#       define DYN_GFX_CLK_OFF_EN                         (1 << 7)
+#       define GFX_CLK_FORCE_ON                           (1 << 8)
+#       define GFX_CLK_REQUEST_OFF                        (1 << 9)
+#       define GFX_CLK_FORCE_OFF                          (1 << 10)
+#       define GFX_CLK_OFF_ACPI_D1                        (1 << 11)
+#       define GFX_CLK_OFF_ACPI_D2                        (1 << 12)
+#       define GFX_CLK_OFF_ACPI_D3                        (1 << 13)
+#       define DYN_LIGHT_SLEEP_EN                         (1 << 14)
+#define	MCLK_PWRMGT_CNTL				0x648
+#       define DLL_SPEED(x)				((x) << 0)
+#       define DLL_SPEED_MASK				(0x1f << 0)
+#       define MPLL_PWRMGT_OFF                          (1 << 5)
+#       define DLL_READY                                (1 << 6)
+#       define MC_INT_CNTL                              (1 << 7)
+#       define MRDCKA0_PDNB                             (1 << 8)
+#       define MRDCKA1_PDNB                             (1 << 9)
+#       define MRDCKB0_PDNB                             (1 << 10)
+#       define MRDCKB1_PDNB                             (1 << 11)
+#       define MRDCKC0_PDNB                             (1 << 12)
+#       define MRDCKC1_PDNB                             (1 << 13)
+#       define MRDCKD0_PDNB                             (1 << 14)
+#       define MRDCKD1_PDNB                             (1 << 15)
+#       define MRDCKA0_RESET                            (1 << 16)
+#       define MRDCKA1_RESET                            (1 << 17)
+#       define MRDCKB0_RESET                            (1 << 18)
+#       define MRDCKB1_RESET                            (1 << 19)
+#       define MRDCKC0_RESET                            (1 << 20)
+#       define MRDCKC1_RESET                            (1 << 21)
+#       define MRDCKD0_RESET                            (1 << 22)
+#       define MRDCKD1_RESET                            (1 << 23)
+#       define DLL_READY_READ                           (1 << 24)
+#       define USE_DISPLAY_GAP                          (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                (1 << 26)
+#       define MPLL_TURNOFF_D2                          (1 << 28)
+#define	DLL_CNTL					0x64c
+#       define MRDCKA0_BYPASS                           (1 << 24)
+#       define MRDCKA1_BYPASS                           (1 << 25)
+#       define MRDCKB0_BYPASS                           (1 << 26)
+#       define MRDCKB1_BYPASS                           (1 << 27)
+#       define MRDCKC0_BYPASS                           (1 << 28)
+#       define MRDCKC1_BYPASS                           (1 << 29)
+#       define MRDCKD0_BYPASS                           (1 << 30)
+#       define MRDCKD1_BYPASS                           (1 << 31)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                  0x66c
+#       define CURRENT_STATE_INDEX_MASK                   (0xf << 4)
+#       define CURRENT_STATE_INDEX_SHIFT                  4
+
+#define CG_AT                                           0x6d4
+#       define CG_R(x)					((x) << 0)
+#       define CG_R_MASK				(0xffff << 0)
+#       define CG_L(x)					((x) << 16)
+#       define CG_L_MASK				(0xffff << 16)
+
+#define	CG_BIF_REQ_AND_RSP				0x7f4
+#define		CG_CLIENT_REQ(x)			((x) << 0)
+#define		CG_CLIENT_REQ_MASK			(0xff << 0)
+#define		CG_CLIENT_REQ_SHIFT			0
+#define		CG_CLIENT_RESP(x)			((x) << 8)
+#define		CG_CLIENT_RESP_MASK			(0xff << 8)
+#define		CG_CLIENT_RESP_SHIFT			8
+#define		CLIENT_CG_REQ(x)			((x) << 16)
+#define		CLIENT_CG_REQ_MASK			(0xff << 16)
+#define		CLIENT_CG_REQ_SHIFT			16
+#define		CLIENT_CG_RESP(x)			((x) << 24)
+#define		CLIENT_CG_RESP_MASK			(0xff << 24)
+#define		CLIENT_CG_RESP_SHIFT			24
+
+#define	CG_SPLL_SPREAD_SPECTRUM				0x790
+#define		SSEN					(1 << 0)
+#define		CLK_S(x)				((x) << 4)
+#define		CLK_S_MASK				(0xfff << 4)
+#define		CLK_S_SHIFT				4
+#define	CG_SPLL_SPREAD_SPECTRUM_2			0x794
+#define		CLK_V(x)				((x) << 0)
+#define		CLK_V_MASK				(0x3ffffff << 0)
+#define		CLK_V_SHIFT				0
+
+#define SMC_SCRATCH0                                    0x81c
+
+#define	CG_SPLL_FUNC_CNTL_4				0x850
+
+#define	MPLL_SS1					0x85c
+#define		CLKV(x)					((x) << 0)
+#define		CLKV_MASK				(0x3ffffff << 0)
+#define	MPLL_SS2					0x860
+#define		CLKS(x)					((x) << 0)
+#define		CLKS_MASK				(0xfff << 0)
+
+#define	CG_CAC_CTRL					0x88c
+#define		TID_CNT(x)				((x) << 0)
+#define		TID_CNT_MASK				(0x3fff << 0)
+#define		TID_UNIT(x)				((x) << 14)
+#define		TID_UNIT_MASK				(0xf << 14)
+
+#define	CG_IND_ADDR					0x8f8
+#define	CG_IND_DATA					0x8fc
+/* CGIND regs */
+#define	CG_CGTT_LOCAL_0					0x00
+#define	CG_CGTT_LOCAL_1					0x01
+
+#define MC_CG_CONFIG                                    0x25bc
+#define         MCDW_WR_ENABLE                          (1 << 0)
+#define         MCDX_WR_ENABLE                          (1 << 1)
+#define         MCDY_WR_ENABLE                          (1 << 2)
+#define         MCDZ_WR_ENABLE                          (1 << 3)
+#define		MC_RD_ENABLE(x)				((x) << 4)
+#define		MC_RD_ENABLE_MASK			(3 << 4)
+#define		INDEX(x)				((x) << 6)
+#define		INDEX_MASK				(0xfff << 6)
+#define		INDEX_SHIFT				6
+
+#define	MC_ARB_CAC_CNTL					0x2750
+#define         ENABLE                                  (1 << 0)
+#define		READ_WEIGHT(x)				((x) << 1)
+#define		READ_WEIGHT_MASK			(0x3f << 1)
+#define		READ_WEIGHT_SHIFT			1
+#define		WRITE_WEIGHT(x)				((x) << 7)
+#define		WRITE_WEIGHT_MASK			(0x3f << 7)
+#define		WRITE_WEIGHT_SHIFT			7
+#define         ALLOW_OVERFLOW                          (1 << 13)
+
+#define	MC_ARB_DRAM_TIMING				0x2774
+#define	MC_ARB_DRAM_TIMING2				0x2778
+
+#define	MC_ARB_RFSH_RATE				0x27b0
+#define		POWERMODE0(x)				((x) << 0)
+#define		POWERMODE0_MASK				(0xff << 0)
+#define		POWERMODE0_SHIFT			0
+#define		POWERMODE1(x)				((x) << 8)
+#define		POWERMODE1_MASK				(0xff << 8)
+#define		POWERMODE1_SHIFT			8
+#define		POWERMODE2(x)				((x) << 16)
+#define		POWERMODE2_MASK				(0xff << 16)
+#define		POWERMODE2_SHIFT			16
+#define		POWERMODE3(x)				((x) << 24)
+#define		POWERMODE3_MASK				(0xff << 24)
+#define		POWERMODE3_SHIFT			24
+
+#define MC_ARB_CG                                       0x27e8
+#define		CG_ARB_REQ(x)				((x) << 0)
+#define		CG_ARB_REQ_MASK				(0xff << 0)
+#define		CG_ARB_REQ_SHIFT			0
+#define		CG_ARB_RESP(x)				((x) << 8)
+#define		CG_ARB_RESP_MASK			(0xff << 8)
+#define		CG_ARB_RESP_SHIFT			8
+#define		ARB_CG_REQ(x)				((x) << 16)
+#define		ARB_CG_REQ_MASK				(0xff << 16)
+#define		ARB_CG_REQ_SHIFT			16
+#define		ARB_CG_RESP(x)				((x) << 24)
+#define		ARB_CG_RESP_MASK			(0xff << 24)
+#define		ARB_CG_RESP_SHIFT			24
+
+#define	MC_ARB_DRAM_TIMING_1				0x27f0
+#define	MC_ARB_DRAM_TIMING_2				0x27f4
+#define	MC_ARB_DRAM_TIMING_3				0x27f8
+#define	MC_ARB_DRAM_TIMING2_1				0x27fc
+#define	MC_ARB_DRAM_TIMING2_2				0x2800
+#define	MC_ARB_DRAM_TIMING2_3				0x2804
+#define MC_ARB_BURST_TIME                               0x2808
+#define		STATE0(x)				((x) << 0)
+#define		STATE0_MASK				(0x1f << 0)
+#define		STATE0_SHIFT				0
+#define		STATE1(x)				((x) << 5)
+#define		STATE1_MASK				(0x1f << 5)
+#define		STATE1_SHIFT				5
+#define		STATE2(x)				((x) << 10)
+#define		STATE2_MASK				(0x1f << 10)
+#define		STATE2_SHIFT				10
+#define		STATE3(x)				((x) << 15)
+#define		STATE3_MASK				(0x1f << 15)
+#define		STATE3_SHIFT				15
+
+#define MC_CG_DATAPORT                                  0x2884
+
+#define MC_SEQ_RAS_TIMING                               0x28a0
+#define MC_SEQ_CAS_TIMING                               0x28a4
+#define MC_SEQ_MISC_TIMING                              0x28a8
+#define MC_SEQ_MISC_TIMING2                             0x28ac
+#define MC_SEQ_PMG_TIMING                               0x28b0
+#define MC_SEQ_RD_CTL_D0                                0x28b4
+#define MC_SEQ_RD_CTL_D1                                0x28b8
+#define MC_SEQ_WR_CTL_D0                                0x28bc
+#define MC_SEQ_WR_CTL_D1                                0x28c0
+
+#define MC_SEQ_MISC0                                    0x2a00
+#define         MC_SEQ_MISC0_GDDR5_SHIFT                28
+#define         MC_SEQ_MISC0_GDDR5_MASK                 0xf0000000
+#define         MC_SEQ_MISC0_GDDR5_VALUE                5
+#define MC_SEQ_MISC1                                    0x2a04
+#define MC_SEQ_RESERVE_M                                0x2a08
+#define MC_PMG_CMD_EMRS                                 0x2a0c
+
+#define MC_SEQ_MISC3                                    0x2a2c
+
+#define MC_SEQ_MISC5                                    0x2a54
+#define MC_SEQ_MISC6                                    0x2a58
+
+#define MC_SEQ_MISC7                                    0x2a64
+
+#define MC_SEQ_RAS_TIMING_LP                            0x2a6c
+#define MC_SEQ_CAS_TIMING_LP                            0x2a70
+#define MC_SEQ_MISC_TIMING_LP                           0x2a74
+#define MC_SEQ_MISC_TIMING2_LP                          0x2a78
+#define MC_SEQ_WR_CTL_D0_LP                             0x2a7c
+#define MC_SEQ_WR_CTL_D1_LP                             0x2a80
+#define MC_SEQ_PMG_CMD_EMRS_LP                          0x2a84
+#define MC_SEQ_PMG_CMD_MRS_LP                           0x2a88
+
+#define MC_PMG_CMD_MRS                                  0x2aac
+
+#define MC_SEQ_RD_CTL_D0_LP                             0x2b1c
+#define MC_SEQ_RD_CTL_D1_LP                             0x2b20
+
+#define MC_PMG_CMD_MRS1                                 0x2b44
+#define MC_SEQ_PMG_CMD_MRS1_LP                          0x2b48
+#define MC_SEQ_PMG_TIMING_LP                            0x2b4c
+
+#define MC_PMG_CMD_MRS2                                 0x2b5c
+#define MC_SEQ_PMG_CMD_MRS2_LP                          0x2b60
+
+#define	LB_SYNC_RESET_SEL				0x6b28
+#define		LB_SYNC_RESET_SEL_MASK			(3 << 0)
+#define		LB_SYNC_RESET_SEL_SHIFT			0
+
+#define	DC_STUTTER_CNTL					0x6b30
+#define		DC_STUTTER_ENABLE_A			(1 << 0)
+#define		DC_STUTTER_ENABLE_B			(1 << 1)
+
+#define SQ_CAC_THRESHOLD                                0x8e4c
+#define		VSP(x)					((x) << 0)
+#define		VSP_MASK				(0xff << 0)
+#define		VSP_SHIFT				0
+#define		VSP0(x)					((x) << 8)
+#define		VSP0_MASK				(0xff << 8)
+#define		VSP0_SHIFT				8
+#define		GPR(x)					((x) << 16)
+#define		GPR_MASK				(0xff << 16)
+#define		GPR_SHIFT				16
+
+#define SQ_POWER_THROTTLE                               0x8e58
+#define		MIN_POWER(x)				((x) << 0)
+#define		MIN_POWER_MASK				(0x3fff << 0)
+#define		MIN_POWER_SHIFT				0
+#define		MAX_POWER(x)				((x) << 16)
+#define		MAX_POWER_MASK				(0x3fff << 16)
+#define		MAX_POWER_SHIFT				0
+#define SQ_POWER_THROTTLE2                              0x8e5c
+#define		MAX_POWER_DELTA(x)			((x) << 0)
+#define		MAX_POWER_DELTA_MASK			(0x3fff << 0)
+#define		MAX_POWER_DELTA_SHIFT			0
+#define		STI_SIZE(x)				((x) << 16)
+#define		STI_SIZE_MASK				(0x3ff << 16)
+#define		STI_SIZE_SHIFT				16
+#define		LTI_RATIO(x)				((x) << 27)
+#define		LTI_RATIO_MASK				(0xf << 27)
+#define		LTI_RATIO_SHIFT				27
+
+/* CG indirect registers */
+#define CG_CAC_REGION_1_WEIGHT_0                        0x83
+#define		WEIGHT_TCP_SIG0(x)			((x) << 0)
+#define		WEIGHT_TCP_SIG0_MASK			(0x3f << 0)
+#define		WEIGHT_TCP_SIG0_SHIFT			0
+#define		WEIGHT_TCP_SIG1(x)			((x) << 6)
+#define		WEIGHT_TCP_SIG1_MASK			(0x3f << 6)
+#define		WEIGHT_TCP_SIG1_SHIFT			6
+#define		WEIGHT_TA_SIG(x)			((x) << 12)
+#define		WEIGHT_TA_SIG_MASK			(0x3f << 12)
+#define		WEIGHT_TA_SIG_SHIFT			12
+#define CG_CAC_REGION_1_WEIGHT_1                        0x84
+#define		WEIGHT_TCC_EN0(x)			((x) << 0)
+#define		WEIGHT_TCC_EN0_MASK			(0x3f << 0)
+#define		WEIGHT_TCC_EN0_SHIFT			0
+#define		WEIGHT_TCC_EN1(x)			((x) << 6)
+#define		WEIGHT_TCC_EN1_MASK			(0x3f << 6)
+#define		WEIGHT_TCC_EN1_SHIFT			6
+#define		WEIGHT_TCC_EN2(x)			((x) << 12)
+#define		WEIGHT_TCC_EN2_MASK			(0x3f << 12)
+#define		WEIGHT_TCC_EN2_SHIFT			12
+#define		WEIGHT_TCC_EN3(x)			((x) << 18)
+#define		WEIGHT_TCC_EN3_MASK			(0x3f << 18)
+#define		WEIGHT_TCC_EN3_SHIFT			18
+#define CG_CAC_REGION_2_WEIGHT_0                        0x85
+#define		WEIGHT_CB_EN0(x)			((x) << 0)
+#define		WEIGHT_CB_EN0_MASK			(0x3f << 0)
+#define		WEIGHT_CB_EN0_SHIFT			0
+#define		WEIGHT_CB_EN1(x)			((x) << 6)
+#define		WEIGHT_CB_EN1_MASK			(0x3f << 6)
+#define		WEIGHT_CB_EN1_SHIFT			6
+#define		WEIGHT_CB_EN2(x)			((x) << 12)
+#define		WEIGHT_CB_EN2_MASK			(0x3f << 12)
+#define		WEIGHT_CB_EN2_SHIFT			12
+#define		WEIGHT_CB_EN3(x)			((x) << 18)
+#define		WEIGHT_CB_EN3_MASK			(0x3f << 18)
+#define		WEIGHT_CB_EN3_SHIFT			18
+#define CG_CAC_REGION_2_WEIGHT_1                        0x86
+#define		WEIGHT_DB_SIG0(x)			((x) << 0)
+#define		WEIGHT_DB_SIG0_MASK			(0x3f << 0)
+#define		WEIGHT_DB_SIG0_SHIFT			0
+#define		WEIGHT_DB_SIG1(x)			((x) << 6)
+#define		WEIGHT_DB_SIG1_MASK			(0x3f << 6)
+#define		WEIGHT_DB_SIG1_SHIFT			6
+#define		WEIGHT_DB_SIG2(x)			((x) << 12)
+#define		WEIGHT_DB_SIG2_MASK			(0x3f << 12)
+#define		WEIGHT_DB_SIG2_SHIFT			12
+#define		WEIGHT_DB_SIG3(x)			((x) << 18)
+#define		WEIGHT_DB_SIG3_MASK			(0x3f << 18)
+#define		WEIGHT_DB_SIG3_SHIFT			18
+#define CG_CAC_REGION_2_WEIGHT_2                        0x87
+#define		WEIGHT_SXM_SIG0(x)			((x) << 0)
+#define		WEIGHT_SXM_SIG0_MASK			(0x3f << 0)
+#define		WEIGHT_SXM_SIG0_SHIFT			0
+#define		WEIGHT_SXM_SIG1(x)			((x) << 6)
+#define		WEIGHT_SXM_SIG1_MASK			(0x3f << 6)
+#define		WEIGHT_SXM_SIG1_SHIFT			6
+#define		WEIGHT_SXM_SIG2(x)			((x) << 12)
+#define		WEIGHT_SXM_SIG2_MASK			(0x3f << 12)
+#define		WEIGHT_SXM_SIG2_SHIFT			12
+#define		WEIGHT_SXS_SIG0(x)			((x) << 18)
+#define		WEIGHT_SXS_SIG0_MASK			(0x3f << 18)
+#define		WEIGHT_SXS_SIG0_SHIFT			18
+#define		WEIGHT_SXS_SIG1(x)			((x) << 24)
+#define		WEIGHT_SXS_SIG1_MASK			(0x3f << 24)
+#define		WEIGHT_SXS_SIG1_SHIFT			24
+#define CG_CAC_REGION_3_WEIGHT_0                        0x88
+#define		WEIGHT_XBR_0(x)				((x) << 0)
+#define		WEIGHT_XBR_0_MASK			(0x3f << 0)
+#define		WEIGHT_XBR_0_SHIFT			0
+#define		WEIGHT_XBR_1(x)				((x) << 6)
+#define		WEIGHT_XBR_1_MASK			(0x3f << 6)
+#define		WEIGHT_XBR_1_SHIFT			6
+#define		WEIGHT_XBR_2(x)				((x) << 12)
+#define		WEIGHT_XBR_2_MASK			(0x3f << 12)
+#define		WEIGHT_XBR_2_SHIFT			12
+#define		WEIGHT_SPI_SIG0(x)			((x) << 18)
+#define		WEIGHT_SPI_SIG0_MASK			(0x3f << 18)
+#define		WEIGHT_SPI_SIG0_SHIFT			18
+#define CG_CAC_REGION_3_WEIGHT_1                        0x89
+#define		WEIGHT_SPI_SIG1(x)			((x) << 0)
+#define		WEIGHT_SPI_SIG1_MASK			(0x3f << 0)
+#define		WEIGHT_SPI_SIG1_SHIFT			0
+#define		WEIGHT_SPI_SIG2(x)			((x) << 6)
+#define		WEIGHT_SPI_SIG2_MASK			(0x3f << 6)
+#define		WEIGHT_SPI_SIG2_SHIFT			6
+#define		WEIGHT_SPI_SIG3(x)			((x) << 12)
+#define		WEIGHT_SPI_SIG3_MASK			(0x3f << 12)
+#define		WEIGHT_SPI_SIG3_SHIFT			12
+#define		WEIGHT_SPI_SIG4(x)			((x) << 18)
+#define		WEIGHT_SPI_SIG4_MASK			(0x3f << 18)
+#define		WEIGHT_SPI_SIG4_SHIFT			18
+#define		WEIGHT_SPI_SIG5(x)			((x) << 24)
+#define		WEIGHT_SPI_SIG5_MASK			(0x3f << 24)
+#define		WEIGHT_SPI_SIG5_SHIFT			24
+#define CG_CAC_REGION_4_WEIGHT_0                        0x8a
+#define		WEIGHT_LDS_SIG0(x)			((x) << 0)
+#define		WEIGHT_LDS_SIG0_MASK			(0x3f << 0)
+#define		WEIGHT_LDS_SIG0_SHIFT			0
+#define		WEIGHT_LDS_SIG1(x)			((x) << 6)
+#define		WEIGHT_LDS_SIG1_MASK			(0x3f << 6)
+#define		WEIGHT_LDS_SIG1_SHIFT			6
+#define		WEIGHT_SC(x)				((x) << 24)
+#define		WEIGHT_SC_MASK				(0x3f << 24)
+#define		WEIGHT_SC_SHIFT				24
+#define CG_CAC_REGION_4_WEIGHT_1                        0x8b
+#define		WEIGHT_BIF(x)				((x) << 0)
+#define		WEIGHT_BIF_MASK				(0x3f << 0)
+#define		WEIGHT_BIF_SHIFT			0
+#define		WEIGHT_CP(x)				((x) << 6)
+#define		WEIGHT_CP_MASK				(0x3f << 6)
+#define		WEIGHT_CP_SHIFT				6
+#define		WEIGHT_PA_SIG0(x)			((x) << 12)
+#define		WEIGHT_PA_SIG0_MASK			(0x3f << 12)
+#define		WEIGHT_PA_SIG0_SHIFT			12
+#define		WEIGHT_PA_SIG1(x)			((x) << 18)
+#define		WEIGHT_PA_SIG1_MASK			(0x3f << 18)
+#define		WEIGHT_PA_SIG1_SHIFT			18
+#define		WEIGHT_VGT_SIG0(x)			((x) << 24)
+#define		WEIGHT_VGT_SIG0_MASK			(0x3f << 24)
+#define		WEIGHT_VGT_SIG0_SHIFT			24
+#define CG_CAC_REGION_4_WEIGHT_2                        0x8c
+#define		WEIGHT_VGT_SIG1(x)			((x) << 0)
+#define		WEIGHT_VGT_SIG1_MASK			(0x3f << 0)
+#define		WEIGHT_VGT_SIG1_SHIFT			0
+#define		WEIGHT_VGT_SIG2(x)			((x) << 6)
+#define		WEIGHT_VGT_SIG2_MASK			(0x3f << 6)
+#define		WEIGHT_VGT_SIG2_SHIFT			6
+#define		WEIGHT_DC_SIG0(x)			((x) << 12)
+#define		WEIGHT_DC_SIG0_MASK			(0x3f << 12)
+#define		WEIGHT_DC_SIG0_SHIFT			12
+#define		WEIGHT_DC_SIG1(x)			((x) << 18)
+#define		WEIGHT_DC_SIG1_MASK			(0x3f << 18)
+#define		WEIGHT_DC_SIG1_SHIFT			18
+#define		WEIGHT_DC_SIG2(x)			((x) << 24)
+#define		WEIGHT_DC_SIG2_MASK			(0x3f << 24)
+#define		WEIGHT_DC_SIG2_SHIFT			24
+#define CG_CAC_REGION_4_WEIGHT_3                        0x8d
+#define		WEIGHT_DC_SIG3(x)			((x) << 0)
+#define		WEIGHT_DC_SIG3_MASK			(0x3f << 0)
+#define		WEIGHT_DC_SIG3_SHIFT			0
+#define		WEIGHT_UVD_SIG0(x)			((x) << 6)
+#define		WEIGHT_UVD_SIG0_MASK			(0x3f << 6)
+#define		WEIGHT_UVD_SIG0_SHIFT			6
+#define		WEIGHT_UVD_SIG1(x)			((x) << 12)
+#define		WEIGHT_UVD_SIG1_MASK			(0x3f << 12)
+#define		WEIGHT_UVD_SIG1_SHIFT			12
+#define		WEIGHT_SPARE0(x)			((x) << 18)
+#define		WEIGHT_SPARE0_MASK			(0x3f << 18)
+#define		WEIGHT_SPARE0_SHIFT			18
+#define		WEIGHT_SPARE1(x)			((x) << 24)
+#define		WEIGHT_SPARE1_MASK			(0x3f << 24)
+#define		WEIGHT_SPARE1_SHIFT			24
+#define CG_CAC_REGION_5_WEIGHT_0                        0x8e
+#define		WEIGHT_SQ_VSP(x)			((x) << 0)
+#define		WEIGHT_SQ_VSP_MASK			(0x3fff << 0)
+#define		WEIGHT_SQ_VSP_SHIFT			0
+#define		WEIGHT_SQ_VSP0(x)			((x) << 14)
+#define		WEIGHT_SQ_VSP0_MASK			(0x3fff << 14)
+#define		WEIGHT_SQ_VSP0_SHIFT			14
+#define CG_CAC_REGION_4_OVERRIDE_4                      0xab
+#define		OVR_MODE_SPARE_0(x)			((x) << 16)
+#define		OVR_MODE_SPARE_0_MASK			(0x1 << 16)
+#define		OVR_MODE_SPARE_0_SHIFT			16
+#define		OVR_VAL_SPARE_0(x)			((x) << 17)
+#define		OVR_VAL_SPARE_0_MASK			(0x1 << 17)
+#define		OVR_VAL_SPARE_0_SHIFT			17
+#define		OVR_MODE_SPARE_1(x)			((x) << 18)
+#define		OVR_MODE_SPARE_1_MASK			(0x3f << 18)
+#define		OVR_MODE_SPARE_1_SHIFT			18
+#define		OVR_VAL_SPARE_1(x)			((x) << 19)
+#define		OVR_VAL_SPARE_1_MASK			(0x3f << 19)
+#define		OVR_VAL_SPARE_1_SHIFT			19
+#define CG_CAC_REGION_5_WEIGHT_1                        0xb7
+#define		WEIGHT_SQ_GPR(x)			((x) << 0)
+#define		WEIGHT_SQ_GPR_MASK			(0x3fff << 0)
+#define		WEIGHT_SQ_GPR_SHIFT			0
+#define		WEIGHT_SQ_LDS(x)			((x) << 14)
+#define		WEIGHT_SQ_LDS_MASK			(0x3fff << 14)
+#define		WEIGHT_SQ_LDS_SHIFT			14
+
+/* PCIE link stuff */
+#define PCIE_LC_TRAINING_CNTL                             0xa1 /* PCIE_P */
+#define PCIE_LC_LINK_WIDTH_CNTL                           0xa2 /* PCIE_P */
+#       define LC_LINK_WIDTH_SHIFT                        0
+#       define LC_LINK_WIDTH_MASK                         0x7
+#       define LC_LINK_WIDTH_X0                           0
+#       define LC_LINK_WIDTH_X1                           1
+#       define LC_LINK_WIDTH_X2                           2
+#       define LC_LINK_WIDTH_X4                           3
+#       define LC_LINK_WIDTH_X8                           4
+#       define LC_LINK_WIDTH_X16                          6
+#       define LC_LINK_WIDTH_RD_SHIFT                     4
+#       define LC_LINK_WIDTH_RD_MASK                      0x70
+#       define LC_RECONFIG_ARC_MISSING_ESCAPE             (1 << 7)
+#       define LC_RECONFIG_NOW                            (1 << 8)
+#       define LC_RENEGOTIATION_SUPPORT                   (1 << 9)
+#       define LC_RENEGOTIATE_EN                          (1 << 10)
+#       define LC_SHORT_RECONFIG_EN                       (1 << 11)
+#       define LC_UPCONFIGURE_SUPPORT                     (1 << 12)
+#       define LC_UPCONFIGURE_DIS                         (1 << 13)
+#define PCIE_LC_SPEED_CNTL                                0xa4 /* PCIE_P */
+#       define LC_GEN2_EN_STRAP                           (1 << 0)
+#       define LC_TARGET_LINK_SPEED_OVERRIDE_EN           (1 << 1)
+#       define LC_FORCE_EN_HW_SPEED_CHANGE                (1 << 5)
+#       define LC_FORCE_DIS_HW_SPEED_CHANGE               (1 << 6)
+#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK      (0x3 << 8)
+#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT     3
+#       define LC_CURRENT_DATA_RATE                       (1 << 11)
+#       define LC_HW_VOLTAGE_IF_CONTROL(x)                ((x) << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_MASK              (3 << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_SHIFT             12
+#       define LC_VOLTAGE_TIMER_SEL_MASK                  (0xf << 14)
+#       define LC_CLR_FAILED_SPD_CHANGE_CNT               (1 << 21)
+#       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 23)
+#       define LC_OTHER_SIDE_SUPPORTS_GEN2                (1 << 24)
+#define MM_CFGREGS_CNTL                                   0x544c
+#       define MM_WR_TO_CFG_EN                            (1 << 3)
+#define LINK_CNTL2                                        0x88 /* F0 */
+#       define TARGET_LINK_SPEED_MASK                     (0xf << 0)
+#       define SELECTABLE_DEEMPHASIS                      (1 << 6)
+
 /*
  * UVD
  */
diff --git a/drivers/gpu/drm/radeon/nislands_smc.h b/drivers/gpu/drm/radeon/nislands_smc.h
new file mode 100644
index 000000000000..3cf8fc0d83f4
--- /dev/null
+++ b/drivers/gpu/drm/radeon/nislands_smc.h
@@ -0,0 +1,329 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __NISLANDS_SMC_H__
+#define __NISLANDS_SMC_H__
+
+#pragma pack(push, 1)
+
+#define NISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE 16
+
+struct PP_NIslands_Dpm2PerfLevel
+{
+    uint8_t     MaxPS;
+    uint8_t     TgtAct;
+    uint8_t     MaxPS_StepInc;
+    uint8_t     MaxPS_StepDec;
+    uint8_t     PSST;
+    uint8_t     NearTDPDec;
+    uint8_t     AboveSafeInc;
+    uint8_t     BelowSafeInc;
+    uint8_t     PSDeltaLimit;
+    uint8_t     PSDeltaWin;
+    uint8_t     Reserved[6];
+};
+
+typedef struct PP_NIslands_Dpm2PerfLevel PP_NIslands_Dpm2PerfLevel;
+
+struct PP_NIslands_DPM2Parameters
+{
+    uint32_t    TDPLimit;
+    uint32_t    NearTDPLimit;
+    uint32_t    SafePowerLimit;
+    uint32_t    PowerBoostLimit;
+};
+typedef struct PP_NIslands_DPM2Parameters PP_NIslands_DPM2Parameters;
+
+struct NISLANDS_SMC_SCLK_VALUE
+{
+    uint32_t        vCG_SPLL_FUNC_CNTL;
+    uint32_t        vCG_SPLL_FUNC_CNTL_2;
+    uint32_t        vCG_SPLL_FUNC_CNTL_3;
+    uint32_t        vCG_SPLL_FUNC_CNTL_4;
+    uint32_t        vCG_SPLL_SPREAD_SPECTRUM;
+    uint32_t        vCG_SPLL_SPREAD_SPECTRUM_2;
+    uint32_t        sclk_value;
+};
+
+typedef struct NISLANDS_SMC_SCLK_VALUE NISLANDS_SMC_SCLK_VALUE;
+
+struct NISLANDS_SMC_MCLK_VALUE
+{
+    uint32_t        vMPLL_FUNC_CNTL;
+    uint32_t        vMPLL_FUNC_CNTL_1;
+    uint32_t        vMPLL_FUNC_CNTL_2;
+    uint32_t        vMPLL_AD_FUNC_CNTL;
+    uint32_t        vMPLL_AD_FUNC_CNTL_2;
+    uint32_t        vMPLL_DQ_FUNC_CNTL;
+    uint32_t        vMPLL_DQ_FUNC_CNTL_2;
+    uint32_t        vMCLK_PWRMGT_CNTL;
+    uint32_t        vDLL_CNTL;
+    uint32_t        vMPLL_SS;
+    uint32_t        vMPLL_SS2;
+    uint32_t        mclk_value;
+};
+
+typedef struct NISLANDS_SMC_MCLK_VALUE NISLANDS_SMC_MCLK_VALUE;
+
+struct NISLANDS_SMC_VOLTAGE_VALUE
+{
+    uint16_t             value;
+    uint8_t              index;
+    uint8_t              padding;
+};
+
+typedef struct NISLANDS_SMC_VOLTAGE_VALUE NISLANDS_SMC_VOLTAGE_VALUE;
+
+struct NISLANDS_SMC_HW_PERFORMANCE_LEVEL
+{
+    uint8_t                     arbValue;
+    uint8_t                     ACIndex;
+    uint8_t                     displayWatermark;
+    uint8_t                     gen2PCIE;
+    uint8_t                     reserved1;
+    uint8_t                     reserved2;
+    uint8_t                     strobeMode;
+    uint8_t                     mcFlags;
+    uint32_t                    aT;
+    uint32_t                    bSP;
+    NISLANDS_SMC_SCLK_VALUE     sclk;
+    NISLANDS_SMC_MCLK_VALUE     mclk;
+    NISLANDS_SMC_VOLTAGE_VALUE  vddc;
+    NISLANDS_SMC_VOLTAGE_VALUE  mvdd;
+    NISLANDS_SMC_VOLTAGE_VALUE  vddci;
+    NISLANDS_SMC_VOLTAGE_VALUE  std_vddc;
+    uint32_t                    powergate_en;
+    uint8_t                     hUp;
+    uint8_t                     hDown;
+    uint8_t                     stateFlags;
+    uint8_t                     arbRefreshState;
+    uint32_t                    SQPowerThrottle;
+    uint32_t                    SQPowerThrottle_2;
+    uint32_t                    reserved[2];
+    PP_NIslands_Dpm2PerfLevel   dpm2;
+};
+
+#define NISLANDS_SMC_STROBE_RATIO    0x0F
+#define NISLANDS_SMC_STROBE_ENABLE   0x10
+
+#define NISLANDS_SMC_MC_EDC_RD_FLAG  0x01
+#define NISLANDS_SMC_MC_EDC_WR_FLAG  0x02
+#define NISLANDS_SMC_MC_RTT_ENABLE   0x04
+#define NISLANDS_SMC_MC_STUTTER_EN   0x08
+
+typedef struct NISLANDS_SMC_HW_PERFORMANCE_LEVEL NISLANDS_SMC_HW_PERFORMANCE_LEVEL;
+
+struct NISLANDS_SMC_SWSTATE
+{
+    uint8_t                             flags;
+    uint8_t                             levelCount;
+    uint8_t                             padding2;
+    uint8_t                             padding3;
+    NISLANDS_SMC_HW_PERFORMANCE_LEVEL   levels[1];
+};
+
+typedef struct NISLANDS_SMC_SWSTATE NISLANDS_SMC_SWSTATE;
+
+#define NISLANDS_SMC_VOLTAGEMASK_VDDC  0
+#define NISLANDS_SMC_VOLTAGEMASK_MVDD  1
+#define NISLANDS_SMC_VOLTAGEMASK_VDDCI 2
+#define NISLANDS_SMC_VOLTAGEMASK_MAX   4
+
+struct NISLANDS_SMC_VOLTAGEMASKTABLE
+{
+    uint8_t  highMask[NISLANDS_SMC_VOLTAGEMASK_MAX];
+    uint32_t lowMask[NISLANDS_SMC_VOLTAGEMASK_MAX];
+};
+
+typedef struct NISLANDS_SMC_VOLTAGEMASKTABLE NISLANDS_SMC_VOLTAGEMASKTABLE;
+
+#define NISLANDS_MAX_NO_VREG_STEPS 32
+
+struct NISLANDS_SMC_STATETABLE
+{
+    uint8_t                             thermalProtectType;
+    uint8_t                             systemFlags;
+    uint8_t                             maxVDDCIndexInPPTable;
+    uint8_t                             extraFlags;
+    uint8_t                             highSMIO[NISLANDS_MAX_NO_VREG_STEPS];
+    uint32_t                            lowSMIO[NISLANDS_MAX_NO_VREG_STEPS];
+    NISLANDS_SMC_VOLTAGEMASKTABLE       voltageMaskTable;
+    PP_NIslands_DPM2Parameters          dpm2Params;
+    NISLANDS_SMC_SWSTATE                initialState;
+    NISLANDS_SMC_SWSTATE                ACPIState;
+    NISLANDS_SMC_SWSTATE                ULVState;
+    NISLANDS_SMC_SWSTATE                driverState;
+    NISLANDS_SMC_HW_PERFORMANCE_LEVEL   dpmLevels[NISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1];
+};
+
+typedef struct NISLANDS_SMC_STATETABLE NISLANDS_SMC_STATETABLE;
+
+#define NI_SMC_SOFT_REGISTERS_START        0x108
+
+#define NI_SMC_SOFT_REGISTER_mclk_chg_timeout        0x0
+#define NI_SMC_SOFT_REGISTER_delay_bbias             0xC
+#define NI_SMC_SOFT_REGISTER_delay_vreg              0x10
+#define NI_SMC_SOFT_REGISTER_delay_acpi              0x2C
+#define NI_SMC_SOFT_REGISTER_seq_index               0x64
+#define NI_SMC_SOFT_REGISTER_mvdd_chg_time           0x68
+#define NI_SMC_SOFT_REGISTER_mclk_switch_lim         0x78
+#define NI_SMC_SOFT_REGISTER_watermark_threshold     0x80
+#define NI_SMC_SOFT_REGISTER_mc_block_delay          0x84
+#define NI_SMC_SOFT_REGISTER_uvd_enabled             0x98
+
+#define SMC_NISLANDS_MC_TPP_CAC_NUM_OF_ENTRIES 16
+#define SMC_NISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES 16
+#define SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES 16
+#define SMC_NISLANDS_BIF_LUT_NUM_OF_ENTRIES 4
+
+struct SMC_NISLANDS_MC_TPP_CAC_TABLE
+{
+    uint32_t    tpp[SMC_NISLANDS_MC_TPP_CAC_NUM_OF_ENTRIES];
+    uint32_t    cacValue[SMC_NISLANDS_MC_TPP_CAC_NUM_OF_ENTRIES];
+};
+
+typedef struct SMC_NISLANDS_MC_TPP_CAC_TABLE SMC_NISLANDS_MC_TPP_CAC_TABLE;
+
+
+struct PP_NIslands_CACTABLES
+{
+    uint32_t                cac_bif_lut[SMC_NISLANDS_BIF_LUT_NUM_OF_ENTRIES];
+    uint32_t                cac_lkge_lut[SMC_NISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES][SMC_NISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES];
+
+    uint32_t                pwr_const;
+
+    uint32_t                dc_cacValue;
+    uint32_t                bif_cacValue;
+    uint32_t                lkge_pwr;
+
+    uint8_t                 cac_width;
+    uint8_t                 window_size_p2;
+
+    uint8_t                 num_drop_lsb;
+    uint8_t                 padding_0;
+
+    uint32_t                last_power;
+
+    uint8_t                 AllowOvrflw;
+    uint8_t                 MCWrWeight;
+    uint8_t                 MCRdWeight;
+    uint8_t                 padding_1[9];
+
+    uint8_t                 enableWinAvg;
+    uint8_t                 numWin_TDP;
+    uint8_t                 l2numWin_TDP;
+    uint8_t                 WinIndex;
+
+    uint32_t                dynPwr_TDP[4];
+    uint32_t                lkgePwr_TDP[4];
+    uint32_t                power_TDP[4];
+    uint32_t                avg_dynPwr_TDP;
+    uint32_t                avg_lkgePwr_TDP;
+    uint32_t                avg_power_TDP;
+    uint32_t                lts_power_TDP;
+    uint8_t                 lts_truncate_n;
+    uint8_t                 padding_2[7];
+};
+
+typedef struct PP_NIslands_CACTABLES PP_NIslands_CACTABLES;
+
+#define SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE 32
+#define SMC_NISLANDS_MC_REGISTER_ARRAY_SET_COUNT 20
+
+struct SMC_NIslands_MCRegisterAddress
+{
+    uint16_t s0;
+    uint16_t s1;
+};
+
+typedef struct SMC_NIslands_MCRegisterAddress SMC_NIslands_MCRegisterAddress;
+
+
+struct SMC_NIslands_MCRegisterSet
+{
+    uint32_t value[SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE];
+};
+
+typedef struct SMC_NIslands_MCRegisterSet SMC_NIslands_MCRegisterSet;
+
+struct SMC_NIslands_MCRegisters
+{
+    uint8_t                             last;
+    uint8_t                             reserved[3];
+    SMC_NIslands_MCRegisterAddress      address[SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE];
+    SMC_NIslands_MCRegisterSet          data[SMC_NISLANDS_MC_REGISTER_ARRAY_SET_COUNT];
+};
+
+typedef struct SMC_NIslands_MCRegisters SMC_NIslands_MCRegisters;
+
+struct SMC_NIslands_MCArbDramTimingRegisterSet
+{
+    uint32_t mc_arb_dram_timing;
+    uint32_t mc_arb_dram_timing2;
+    uint8_t  mc_arb_rfsh_rate;
+    uint8_t  padding[3];
+};
+
+typedef struct SMC_NIslands_MCArbDramTimingRegisterSet SMC_NIslands_MCArbDramTimingRegisterSet;
+
+struct SMC_NIslands_MCArbDramTimingRegisters
+{
+    uint8_t                                     arb_current;
+    uint8_t                                     reserved[3];
+    SMC_NIslands_MCArbDramTimingRegisterSet     data[20];
+};
+
+typedef struct SMC_NIslands_MCArbDramTimingRegisters SMC_NIslands_MCArbDramTimingRegisters;
+
+struct SMC_NISLANDS_SPLL_DIV_TABLE
+{
+    uint32_t    freq[256];
+    uint32_t    ss[256];
+};
+
+#define SMC_NISLANDS_SPLL_DIV_TABLE_FBDIV_MASK  0x01ffffff
+#define SMC_NISLANDS_SPLL_DIV_TABLE_FBDIV_SHIFT 0
+#define SMC_NISLANDS_SPLL_DIV_TABLE_PDIV_MASK   0xfe000000
+#define SMC_NISLANDS_SPLL_DIV_TABLE_PDIV_SHIFT  25
+#define SMC_NISLANDS_SPLL_DIV_TABLE_CLKV_MASK   0x000fffff
+#define SMC_NISLANDS_SPLL_DIV_TABLE_CLKV_SHIFT  0
+#define SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_MASK   0xfff00000
+#define SMC_NISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT  20
+
+typedef struct SMC_NISLANDS_SPLL_DIV_TABLE SMC_NISLANDS_SPLL_DIV_TABLE;
+
+#define NISLANDS_SMC_FIRMWARE_HEADER_LOCATION 0x100
+
+#define NISLANDS_SMC_FIRMWARE_HEADER_version                   0x0
+#define NISLANDS_SMC_FIRMWARE_HEADER_flags                     0x4
+#define NISLANDS_SMC_FIRMWARE_HEADER_softRegisters             0x8
+#define NISLANDS_SMC_FIRMWARE_HEADER_stateTable                0xC
+#define NISLANDS_SMC_FIRMWARE_HEADER_fanTable                  0x10
+#define NISLANDS_SMC_FIRMWARE_HEADER_cacTable                  0x14
+#define NISLANDS_SMC_FIRMWARE_HEADER_mcRegisterTable           0x20
+#define NISLANDS_SMC_FIRMWARE_HEADER_mcArbDramAutoRefreshTable 0x2C
+#define NISLANDS_SMC_FIRMWARE_HEADER_spllTable                 0x30
+
+#pragma pack(pop)
+
+#endif
+
diff --git a/drivers/gpu/drm/radeon/ppsmc.h b/drivers/gpu/drm/radeon/ppsmc.h
new file mode 100644
index 000000000000..b5564a3645d2
--- /dev/null
+++ b/drivers/gpu/drm/radeon/ppsmc.h
@@ -0,0 +1,116 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef PP_SMC_H
+#define PP_SMC_H
+
+#pragma pack(push, 1)
+
+#define PPSMC_SWSTATE_FLAG_DC                           0x01
+#define PPSMC_SWSTATE_FLAG_UVD                          0x02
+#define PPSMC_SWSTATE_FLAG_VCE                          0x04
+#define PPSMC_SWSTATE_FLAG_PCIE_X1                      0x08
+
+#define PPSMC_THERMAL_PROTECT_TYPE_INTERNAL             0x00
+#define PPSMC_THERMAL_PROTECT_TYPE_EXTERNAL             0x01
+#define PPSMC_THERMAL_PROTECT_TYPE_NONE                 0xff
+
+#define PPSMC_SYSTEMFLAG_GPIO_DC                        0x01
+#define PPSMC_SYSTEMFLAG_STEPVDDC                       0x02
+#define PPSMC_SYSTEMFLAG_GDDR5                          0x04
+#define PPSMC_SYSTEMFLAG_DISABLE_BABYSTEP               0x08
+#define PPSMC_SYSTEMFLAG_REGULATOR_HOT                  0x10
+#define PPSMC_SYSTEMFLAG_REGULATOR_HOT_ANALOG           0x20
+#define PPSMC_SYSTEMFLAG_REGULATOR_HOT_PROG_GPIO        0x40
+
+#define PPSMC_EXTRAFLAGS_AC2DC_ACTION_MASK              0x07
+#define PPSMC_EXTRAFLAGS_AC2DC_DONT_WAIT_FOR_VBLANK     0x08
+#define PPSMC_EXTRAFLAGS_AC2DC_ACTION_GOTODPMLOWSTATE   0x00
+#define PPSMC_EXTRAFLAGS_AC2DC_ACTION_GOTOINITIALSTATE  0x01
+#define PPSMC_EXTRAFLAGS_AC2DC_GPIO5_POLARITY_HIGH      0x02
+
+#define PPSMC_DISPLAY_WATERMARK_LOW                     0
+#define PPSMC_DISPLAY_WATERMARK_HIGH                    1
+
+#define PPSMC_STATEFLAG_AUTO_PULSE_SKIP    0x01
+#define PPSMC_STATEFLAG_POWERBOOST         0x02
+#define PPSMC_STATEFLAG_DEEPSLEEP_THROTTLE 0x20
+#define PPSMC_STATEFLAG_DEEPSLEEP_BYPASS   0x40
+
+#define PPSMC_Result_OK             ((uint8_t)0x01)
+#define PPSMC_Result_Failed         ((uint8_t)0xFF)
+
+typedef uint8_t PPSMC_Result;
+
+#define PPSMC_MSG_Halt                      ((uint8_t)0x10)
+#define PPSMC_MSG_Resume                    ((uint8_t)0x11)
+#define PPSMC_MSG_ZeroLevelsDisabled        ((uint8_t)0x13)
+#define PPSMC_MSG_OneLevelsDisabled         ((uint8_t)0x14)
+#define PPSMC_MSG_TwoLevelsDisabled         ((uint8_t)0x15)
+#define PPSMC_MSG_EnableThermalInterrupt    ((uint8_t)0x16)
+#define PPSMC_MSG_RunningOnAC               ((uint8_t)0x17)
+#define PPSMC_MSG_SwitchToSwState           ((uint8_t)0x20)
+#define PPSMC_MSG_SwitchToInitialState      ((uint8_t)0x40)
+#define PPSMC_MSG_NoForcedLevel             ((uint8_t)0x41)
+#define PPSMC_MSG_ForceHigh                 ((uint8_t)0x42)
+#define PPSMC_MSG_ForceMediumOrHigh         ((uint8_t)0x43)
+#define PPSMC_MSG_SwitchToMinimumPower      ((uint8_t)0x51)
+#define PPSMC_MSG_ResumeFromMinimumPower    ((uint8_t)0x52)
+#define PPSMC_MSG_EnableCac                 ((uint8_t)0x53)
+#define PPSMC_MSG_DisableCac                ((uint8_t)0x54)
+#define PPSMC_TDPClampingActive             ((uint8_t)0x59)
+#define PPSMC_TDPClampingInactive           ((uint8_t)0x5A)
+#define PPSMC_MSG_NoDisplay                 ((uint8_t)0x5D)
+#define PPSMC_MSG_HasDisplay                ((uint8_t)0x5E)
+#define PPSMC_MSG_UVDPowerOFF               ((uint8_t)0x60)
+#define PPSMC_MSG_UVDPowerON                ((uint8_t)0x61)
+#define PPSMC_MSG_EnableULV                 ((uint8_t)0x62)
+#define PPSMC_MSG_DisableULV                ((uint8_t)0x63)
+#define PPSMC_MSG_EnterULV                  ((uint8_t)0x64)
+#define PPSMC_MSG_ExitULV                   ((uint8_t)0x65)
+#define PPSMC_CACLongTermAvgEnable          ((uint8_t)0x6E)
+#define PPSMC_CACLongTermAvgDisable         ((uint8_t)0x6F)
+#define PPSMC_MSG_CollectCAC_PowerCorreln   ((uint8_t)0x7A)
+#define PPSMC_FlushDataCache                ((uint8_t)0x80)
+#define PPSMC_MSG_SetEnabledLevels          ((uint8_t)0x82)
+#define PPSMC_MSG_SetForcedLevels           ((uint8_t)0x83)
+#define PPSMC_MSG_ResetToDefaults           ((uint8_t)0x84)
+#define PPSMC_MSG_EnableDTE                 ((uint8_t)0x87)
+#define PPSMC_MSG_DisableDTE                ((uint8_t)0x88)
+#define PPSMC_MSG_ThrottleOVRDSCLKDS        ((uint8_t)0x96)
+#define PPSMC_MSG_CancelThrottleOVRDSCLKDS  ((uint8_t)0x97)
+
+/* TN */
+#define PPSMC_MSG_DPM_Config                ((uint32_t) 0x102)
+#define PPSMC_MSG_DPM_ForceState            ((uint32_t) 0x104)
+#define PPSMC_MSG_PG_SIMD_Config            ((uint32_t) 0x108)
+#define PPSMC_MSG_DPM_N_LevelsDisabled      ((uint32_t) 0x112)
+#define PPSMC_MSG_DCE_RemoveVoltageAdjustment   ((uint32_t) 0x11d)
+#define PPSMC_MSG_DCE_AllowVoltageAdjustment    ((uint32_t) 0x11e)
+#define PPSMC_MSG_UVD_DPM_Config            ((uint32_t) 0x124)
+
+
+typedef uint16_t PPSMC_Msg;
+
+#pragma pack(pop)
+
+#endif
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index d0314ecbd7c1..75349cdaa84b 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -39,7 +39,6 @@
 #include "atom.h"
 
 #include <linux/firmware.h>
-#include <linux/platform_device.h>
 #include <linux/module.h>
 
 #include "r100_reg_safe.h"
@@ -989,18 +988,11 @@ void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 /* Load the microcode for the CP */
 static int r100_cp_init_microcode(struct radeon_device *rdev)
 {
-	struct platform_device *pdev;
 	const char *fw_name = NULL;
 	int err;
 
 	DRM_DEBUG_KMS("\n");
 
-	pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0);
-	err = IS_ERR(pdev);
-	if (err) {
-		printk(KERN_ERR "radeon_cp: Failed to register firmware\n");
-		return -EINVAL;
-	}
 	if ((rdev->family == CHIP_R100) || (rdev->family == CHIP_RV100) ||
 	    (rdev->family == CHIP_RV200) || (rdev->family == CHIP_RS100) ||
 	    (rdev->family == CHIP_RS200)) {
@@ -1042,8 +1034,7 @@ static int r100_cp_init_microcode(struct radeon_device *rdev)
 		fw_name = FIRMWARE_R520;
 	}
 
-	err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev);
-	platform_device_unregister(pdev);
+	err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
 	if (err) {
 		printk(KERN_ERR "radeon_cp: Failed to load firmware \"%s\"\n",
 		       fw_name);
@@ -3077,6 +3068,10 @@ int r100_set_surface_reg(struct radeon_device *rdev, int reg,
 			flags |= RADEON_SURF_TILE_COLOR_BOTH;
 		if (tiling_flags & RADEON_TILING_MACRO)
 			flags |= RADEON_SURF_TILE_COLOR_MACRO;
+		/* setting pitch to 0 disables tiling */
+		if ((tiling_flags & (RADEON_TILING_MACRO|RADEON_TILING_MICRO))
+				== 0)
+			pitch = 0;
 	} else if (rdev->family <= CHIP_RV280) {
 		if (tiling_flags & (RADEON_TILING_MACRO))
 			flags |= R200_SURF_TILE_COLOR_MACRO;
@@ -3094,13 +3089,6 @@ int r100_set_surface_reg(struct radeon_device *rdev, int reg,
 	if (tiling_flags & RADEON_TILING_SWAP_32BIT)
 		flags |= RADEON_SURF_AP0_SWP_32BPP | RADEON_SURF_AP1_SWP_32BPP;
 
-	/* when we aren't tiling the pitch seems to needs to be furtherdivided down. - tested on power5 + rn50 server */
-	if (tiling_flags & (RADEON_TILING_SWAP_16BIT | RADEON_TILING_SWAP_32BIT)) {
-		if (!(tiling_flags & (RADEON_TILING_MACRO | RADEON_TILING_MICRO)))
-			if (ASIC_IS_RN50(rdev))
-				pitch /= 16;
-	}
-
 	/* r100/r200 divide by 16 */
 	if (rdev->family < CHIP_R300)
 		flags |= pitch / 16;
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index 6948eb88c2b7..393880a09412 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -28,7 +28,6 @@
 #include <linux/slab.h>
 #include <linux/seq_file.h>
 #include <linux/firmware.h>
-#include <linux/platform_device.h>
 #include <linux/module.h>
 #include <drm/drmP.h>
 #include <drm/radeon_drm.h>
@@ -38,18 +37,7 @@
 #include "r600d.h"
 #include "atom.h"
 #include "avivod.h"
-
-#define PFP_UCODE_SIZE 576
-#define PM4_UCODE_SIZE 1792
-#define RLC_UCODE_SIZE 768
-#define R700_PFP_UCODE_SIZE 848
-#define R700_PM4_UCODE_SIZE 1360
-#define R700_RLC_UCODE_SIZE 1024
-#define EVERGREEN_PFP_UCODE_SIZE 1120
-#define EVERGREEN_PM4_UCODE_SIZE 1376
-#define EVERGREEN_RLC_UCODE_SIZE 768
-#define CAYMAN_RLC_UCODE_SIZE 1024
-#define ARUBA_RLC_UCODE_SIZE 1536
+#include "radeon_ucode.h"
 
 /* Firmware Names */
 MODULE_FIRMWARE("radeon/R600_pfp.bin");
@@ -68,24 +56,32 @@ MODULE_FIRMWARE("radeon/RS780_pfp.bin");
 MODULE_FIRMWARE("radeon/RS780_me.bin");
 MODULE_FIRMWARE("radeon/RV770_pfp.bin");
 MODULE_FIRMWARE("radeon/RV770_me.bin");
+MODULE_FIRMWARE("radeon/RV770_smc.bin");
 MODULE_FIRMWARE("radeon/RV730_pfp.bin");
 MODULE_FIRMWARE("radeon/RV730_me.bin");
+MODULE_FIRMWARE("radeon/RV730_smc.bin");
+MODULE_FIRMWARE("radeon/RV740_smc.bin");
 MODULE_FIRMWARE("radeon/RV710_pfp.bin");
 MODULE_FIRMWARE("radeon/RV710_me.bin");
+MODULE_FIRMWARE("radeon/RV710_smc.bin");
 MODULE_FIRMWARE("radeon/R600_rlc.bin");
 MODULE_FIRMWARE("radeon/R700_rlc.bin");
 MODULE_FIRMWARE("radeon/CEDAR_pfp.bin");
 MODULE_FIRMWARE("radeon/CEDAR_me.bin");
 MODULE_FIRMWARE("radeon/CEDAR_rlc.bin");
+MODULE_FIRMWARE("radeon/CEDAR_smc.bin");
 MODULE_FIRMWARE("radeon/REDWOOD_pfp.bin");
 MODULE_FIRMWARE("radeon/REDWOOD_me.bin");
 MODULE_FIRMWARE("radeon/REDWOOD_rlc.bin");
+MODULE_FIRMWARE("radeon/REDWOOD_smc.bin");
 MODULE_FIRMWARE("radeon/JUNIPER_pfp.bin");
 MODULE_FIRMWARE("radeon/JUNIPER_me.bin");
 MODULE_FIRMWARE("radeon/JUNIPER_rlc.bin");
+MODULE_FIRMWARE("radeon/JUNIPER_smc.bin");
 MODULE_FIRMWARE("radeon/CYPRESS_pfp.bin");
 MODULE_FIRMWARE("radeon/CYPRESS_me.bin");
 MODULE_FIRMWARE("radeon/CYPRESS_rlc.bin");
+MODULE_FIRMWARE("radeon/CYPRESS_smc.bin");
 MODULE_FIRMWARE("radeon/PALM_pfp.bin");
 MODULE_FIRMWARE("radeon/PALM_me.bin");
 MODULE_FIRMWARE("radeon/SUMO_rlc.bin");
@@ -108,6 +104,7 @@ static void r600_gpu_init(struct radeon_device *rdev);
 void r600_fini(struct radeon_device *rdev);
 void r600_irq_disable(struct radeon_device *rdev);
 static void r600_pcie_gen2_enable(struct radeon_device *rdev);
+extern int evergreen_rlc_resume(struct radeon_device *rdev);
 
 /**
  * r600_get_xclk - get the xclk
@@ -2146,22 +2143,15 @@ void r600_cp_stop(struct radeon_device *rdev)
 
 int r600_init_microcode(struct radeon_device *rdev)
 {
-	struct platform_device *pdev;
 	const char *chip_name;
 	const char *rlc_chip_name;
-	size_t pfp_req_size, me_req_size, rlc_req_size;
+	const char *smc_chip_name = "RV770";
+	size_t pfp_req_size, me_req_size, rlc_req_size, smc_req_size = 0;
 	char fw_name[30];
 	int err;
 
 	DRM_DEBUG("\n");
 
-	pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0);
-	err = IS_ERR(pdev);
-	if (err) {
-		printk(KERN_ERR "radeon_cp: Failed to register firmware\n");
-		return -EINVAL;
-	}
-
 	switch (rdev->family) {
 	case CHIP_R600:
 		chip_name = "R600";
@@ -2195,32 +2185,51 @@ int r600_init_microcode(struct radeon_device *rdev)
 	case CHIP_RV770:
 		chip_name = "RV770";
 		rlc_chip_name = "R700";
+		smc_chip_name = "RV770";
+		smc_req_size = ALIGN(RV770_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_RV730:
-	case CHIP_RV740:
 		chip_name = "RV730";
 		rlc_chip_name = "R700";
+		smc_chip_name = "RV730";
+		smc_req_size = ALIGN(RV730_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_RV710:
 		chip_name = "RV710";
 		rlc_chip_name = "R700";
+		smc_chip_name = "RV710";
+		smc_req_size = ALIGN(RV710_SMC_UCODE_SIZE, 4);
+		break;
+	case CHIP_RV740:
+		chip_name = "RV730";
+		rlc_chip_name = "R700";
+		smc_chip_name = "RV740";
+		smc_req_size = ALIGN(RV740_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_CEDAR:
 		chip_name = "CEDAR";
 		rlc_chip_name = "CEDAR";
+		smc_chip_name = "CEDAR";
+		smc_req_size = ALIGN(CEDAR_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_REDWOOD:
 		chip_name = "REDWOOD";
 		rlc_chip_name = "REDWOOD";
+		smc_chip_name = "REDWOOD";
+		smc_req_size = ALIGN(REDWOOD_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_JUNIPER:
 		chip_name = "JUNIPER";
 		rlc_chip_name = "JUNIPER";
+		smc_chip_name = "JUNIPER";
+		smc_req_size = ALIGN(JUNIPER_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_CYPRESS:
 	case CHIP_HEMLOCK:
 		chip_name = "CYPRESS";
 		rlc_chip_name = "CYPRESS";
+		smc_chip_name = "CYPRESS";
+		smc_req_size = ALIGN(CYPRESS_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_PALM:
 		chip_name = "PALM";
@@ -2246,15 +2255,15 @@ int r600_init_microcode(struct radeon_device *rdev)
 		me_req_size = R700_PM4_UCODE_SIZE * 4;
 		rlc_req_size = R700_RLC_UCODE_SIZE * 4;
 	} else {
-		pfp_req_size = PFP_UCODE_SIZE * 4;
-		me_req_size = PM4_UCODE_SIZE * 12;
-		rlc_req_size = RLC_UCODE_SIZE * 4;
+		pfp_req_size = R600_PFP_UCODE_SIZE * 4;
+		me_req_size = R600_PM4_UCODE_SIZE * 12;
+		rlc_req_size = R600_RLC_UCODE_SIZE * 4;
 	}
 
 	DRM_INFO("Loading %s Microcode\n", chip_name);
 
 	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
-	err = request_firmware(&rdev->pfp_fw, fw_name, &pdev->dev);
+	err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev);
 	if (err)
 		goto out;
 	if (rdev->pfp_fw->size != pfp_req_size) {
@@ -2266,7 +2275,7 @@ int r600_init_microcode(struct radeon_device *rdev)
 	}
 
 	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
-	err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev);
+	err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
 	if (err)
 		goto out;
 	if (rdev->me_fw->size != me_req_size) {
@@ -2277,7 +2286,7 @@ int r600_init_microcode(struct radeon_device *rdev)
 	}
 
 	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", rlc_chip_name);
-	err = request_firmware(&rdev->rlc_fw, fw_name, &pdev->dev);
+	err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev);
 	if (err)
 		goto out;
 	if (rdev->rlc_fw->size != rlc_req_size) {
@@ -2287,9 +2296,20 @@ int r600_init_microcode(struct radeon_device *rdev)
 		err = -EINVAL;
 	}
 
-out:
-	platform_device_unregister(pdev);
+	if ((rdev->family >= CHIP_RV770) && (rdev->family <= CHIP_HEMLOCK)) {
+		snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", smc_chip_name);
+		err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
+		if (err)
+			goto out;
+		if (rdev->smc_fw->size != smc_req_size) {
+			printk(KERN_ERR
+			       "smc: Bogus length %zu in firmware \"%s\"\n",
+			       rdev->smc_fw->size, fw_name);
+			err = -EINVAL;
+		}
+	}
 
+out:
 	if (err) {
 		if (err != -EINVAL)
 			printk(KERN_ERR
@@ -2301,6 +2321,8 @@ out:
 		rdev->me_fw = NULL;
 		release_firmware(rdev->rlc_fw);
 		rdev->rlc_fw = NULL;
+		release_firmware(rdev->smc_fw);
+		rdev->smc_fw = NULL;
 	}
 	return err;
 }
@@ -2331,13 +2353,13 @@ static int r600_cp_load_microcode(struct radeon_device *rdev)
 
 	fw_data = (const __be32 *)rdev->me_fw->data;
 	WREG32(CP_ME_RAM_WADDR, 0);
-	for (i = 0; i < PM4_UCODE_SIZE * 3; i++)
+	for (i = 0; i < R600_PM4_UCODE_SIZE * 3; i++)
 		WREG32(CP_ME_RAM_DATA,
 		       be32_to_cpup(fw_data++));
 
 	fw_data = (const __be32 *)rdev->pfp_fw->data;
 	WREG32(CP_PFP_UCODE_ADDR, 0);
-	for (i = 0; i < PFP_UCODE_SIZE; i++)
+	for (i = 0; i < R600_PFP_UCODE_SIZE; i++)
 		WREG32(CP_PFP_UCODE_DATA,
 		       be32_to_cpup(fw_data++));
 
@@ -2986,7 +3008,7 @@ void r600_uvd_fence_emit(struct radeon_device *rdev,
 			 struct radeon_fence *fence)
 {
 	struct radeon_ring *ring = &rdev->ring[fence->ring];
-	uint32_t addr = rdev->fence_drv[fence->ring].gpu_addr;
+	uint64_t addr = rdev->fence_drv[fence->ring].gpu_addr;
 
 	radeon_ring_write(ring, PACKET0(UVD_CONTEXT_ID, 0));
 	radeon_ring_write(ring, fence->seq);
@@ -3112,6 +3134,87 @@ int r600_copy_blit(struct radeon_device *rdev,
 }
 
 /**
+ * r600_copy_cpdma - copy pages using the CP DMA engine
+ *
+ * @rdev: radeon_device pointer
+ * @src_offset: src GPU address
+ * @dst_offset: dst GPU address
+ * @num_gpu_pages: number of GPU pages to xfer
+ * @fence: radeon fence object
+ *
+ * Copy GPU paging using the CP DMA engine (r6xx+).
+ * Used by the radeon ttm implementation to move pages if
+ * registered as the asic copy callback.
+ */
+int r600_copy_cpdma(struct radeon_device *rdev,
+		    uint64_t src_offset, uint64_t dst_offset,
+		    unsigned num_gpu_pages,
+		    struct radeon_fence **fence)
+{
+	struct radeon_semaphore *sem = NULL;
+	int ring_index = rdev->asic->copy.blit_ring_index;
+	struct radeon_ring *ring = &rdev->ring[ring_index];
+	u32 size_in_bytes, cur_size_in_bytes, tmp;
+	int i, num_loops;
+	int r = 0;
+
+	r = radeon_semaphore_create(rdev, &sem);
+	if (r) {
+		DRM_ERROR("radeon: moving bo (%d).\n", r);
+		return r;
+	}
+
+	size_in_bytes = (num_gpu_pages << RADEON_GPU_PAGE_SHIFT);
+	num_loops = DIV_ROUND_UP(size_in_bytes, 0x1fffff);
+	r = radeon_ring_lock(rdev, ring, num_loops * 6 + 21);
+	if (r) {
+		DRM_ERROR("radeon: moving bo (%d).\n", r);
+		radeon_semaphore_free(rdev, &sem, NULL);
+		return r;
+	}
+
+	if (radeon_fence_need_sync(*fence, ring->idx)) {
+		radeon_semaphore_sync_rings(rdev, sem, (*fence)->ring,
+					    ring->idx);
+		radeon_fence_note_sync(*fence, ring->idx);
+	} else {
+		radeon_semaphore_free(rdev, &sem, NULL);
+	}
+
+	for (i = 0; i < num_loops; i++) {
+		cur_size_in_bytes = size_in_bytes;
+		if (cur_size_in_bytes > 0x1fffff)
+			cur_size_in_bytes = 0x1fffff;
+		size_in_bytes -= cur_size_in_bytes;
+		tmp = upper_32_bits(src_offset) & 0xff;
+		if (size_in_bytes == 0)
+			tmp |= PACKET3_CP_DMA_CP_SYNC;
+		radeon_ring_write(ring, PACKET3(PACKET3_CP_DMA, 4));
+		radeon_ring_write(ring, src_offset & 0xffffffff);
+		radeon_ring_write(ring, tmp);
+		radeon_ring_write(ring, dst_offset & 0xffffffff);
+		radeon_ring_write(ring, upper_32_bits(dst_offset) & 0xff);
+		radeon_ring_write(ring, cur_size_in_bytes);
+		src_offset += cur_size_in_bytes;
+		dst_offset += cur_size_in_bytes;
+	}
+	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+	radeon_ring_write(ring, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
+	radeon_ring_write(ring, WAIT_CP_DMA_IDLE_bit);
+
+	r = radeon_fence_emit(rdev, fence, ring->idx);
+	if (r) {
+		radeon_ring_unlock_undo(rdev, ring);
+		return r;
+	}
+
+	radeon_ring_unlock_commit(rdev, ring);
+	radeon_semaphore_free(rdev, &sem, *fence);
+
+	return r;
+}
+
+/**
  * r600_copy_dma - copy pages using the DMA engine
  *
  * @rdev: radeon_device pointer
@@ -3789,7 +3892,7 @@ static void r600_rlc_start(struct radeon_device *rdev)
 	WREG32(RLC_CNTL, RLC_ENABLE);
 }
 
-static int r600_rlc_init(struct radeon_device *rdev)
+static int r600_rlc_resume(struct radeon_device *rdev)
 {
 	u32 i;
 	const __be32 *fw_data;
@@ -3801,45 +3904,22 @@ static int r600_rlc_init(struct radeon_device *rdev)
 
 	WREG32(RLC_HB_CNTL, 0);
 
-	if (rdev->family == CHIP_ARUBA) {
-		WREG32(TN_RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
-		WREG32(TN_RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
-	}
-	if (rdev->family <= CHIP_CAYMAN) {
-		WREG32(RLC_HB_BASE, 0);
-		WREG32(RLC_HB_RPTR, 0);
-		WREG32(RLC_HB_WPTR, 0);
-	}
-	if (rdev->family <= CHIP_CAICOS) {
-		WREG32(RLC_HB_WPTR_LSB_ADDR, 0);
-		WREG32(RLC_HB_WPTR_MSB_ADDR, 0);
-	}
+	WREG32(RLC_HB_BASE, 0);
+	WREG32(RLC_HB_RPTR, 0);
+	WREG32(RLC_HB_WPTR, 0);
+	WREG32(RLC_HB_WPTR_LSB_ADDR, 0);
+	WREG32(RLC_HB_WPTR_MSB_ADDR, 0);
 	WREG32(RLC_MC_CNTL, 0);
 	WREG32(RLC_UCODE_CNTL, 0);
 
 	fw_data = (const __be32 *)rdev->rlc_fw->data;
-	if (rdev->family >= CHIP_ARUBA) {
-		for (i = 0; i < ARUBA_RLC_UCODE_SIZE; i++) {
-			WREG32(RLC_UCODE_ADDR, i);
-			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
-		}
-	} else if (rdev->family >= CHIP_CAYMAN) {
-		for (i = 0; i < CAYMAN_RLC_UCODE_SIZE; i++) {
-			WREG32(RLC_UCODE_ADDR, i);
-			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
-		}
-	} else if (rdev->family >= CHIP_CEDAR) {
-		for (i = 0; i < EVERGREEN_RLC_UCODE_SIZE; i++) {
-			WREG32(RLC_UCODE_ADDR, i);
-			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
-		}
-	} else if (rdev->family >= CHIP_RV770) {
+	if (rdev->family >= CHIP_RV770) {
 		for (i = 0; i < R700_RLC_UCODE_SIZE; i++) {
 			WREG32(RLC_UCODE_ADDR, i);
 			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
 		}
 	} else {
-		for (i = 0; i < RLC_UCODE_SIZE; i++) {
+		for (i = 0; i < R600_RLC_UCODE_SIZE; i++) {
 			WREG32(RLC_UCODE_ADDR, i);
 			WREG32(RLC_UCODE_DATA, be32_to_cpup(fw_data++));
 		}
@@ -3947,7 +4027,10 @@ int r600_irq_init(struct radeon_device *rdev)
 	r600_disable_interrupts(rdev);
 
 	/* init rlc */
-	ret = r600_rlc_init(rdev);
+	if (rdev->family >= CHIP_CEDAR)
+		ret = evergreen_rlc_resume(rdev);
+	else
+		ret = r600_rlc_resume(rdev);
 	if (ret) {
 		r600_ih_ring_fini(rdev);
 		return ret;
@@ -4028,6 +4111,7 @@ int r600_irq_set(struct radeon_device *rdev)
 	u32 hdmi0, hdmi1;
 	u32 d1grph = 0, d2grph = 0;
 	u32 dma_cntl;
+	u32 thermal_int = 0;
 
 	if (!rdev->irq.installed) {
 		WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
@@ -4062,8 +4146,21 @@ int r600_irq_set(struct radeon_device *rdev)
 		hdmi0 = RREG32(HDMI0_AUDIO_PACKET_CONTROL) & ~HDMI0_AZ_FORMAT_WTRIG_MASK;
 		hdmi1 = RREG32(HDMI1_AUDIO_PACKET_CONTROL) & ~HDMI0_AZ_FORMAT_WTRIG_MASK;
 	}
+
 	dma_cntl = RREG32(DMA_CNTL) & ~TRAP_ENABLE;
 
+	if ((rdev->family > CHIP_R600) && (rdev->family < CHIP_RV770)) {
+		thermal_int = RREG32(CG_THERMAL_INT) &
+			~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW);
+	} else if (rdev->family >= CHIP_RV770) {
+		thermal_int = RREG32(RV770_CG_THERMAL_INT) &
+			~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW);
+	}
+	if (rdev->irq.dpm_thermal) {
+		DRM_DEBUG("dpm thermal\n");
+		thermal_int |= THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW;
+	}
+
 	if (atomic_read(&rdev->irq.ring_int[RADEON_RING_TYPE_GFX_INDEX])) {
 		DRM_DEBUG("r600_irq_set: sw int\n");
 		cp_int_cntl |= RB_INT_ENABLE;
@@ -4145,6 +4242,11 @@ int r600_irq_set(struct radeon_device *rdev)
 		WREG32(HDMI0_AUDIO_PACKET_CONTROL, hdmi0);
 		WREG32(HDMI1_AUDIO_PACKET_CONTROL, hdmi1);
 	}
+	if ((rdev->family > CHIP_R600) && (rdev->family < CHIP_RV770)) {
+		WREG32(CG_THERMAL_INT, thermal_int);
+	} else if (rdev->family >= CHIP_RV770) {
+		WREG32(RV770_CG_THERMAL_INT, thermal_int);
+	}
 
 	return 0;
 }
@@ -4336,6 +4438,7 @@ int r600_irq_process(struct radeon_device *rdev)
 	u32 ring_index;
 	bool queue_hotplug = false;
 	bool queue_hdmi = false;
+	bool queue_thermal = false;
 
 	if (!rdev->ih.enabled || rdev->shutdown)
 		return IRQ_NONE;
@@ -4503,6 +4606,16 @@ restart_ih:
 			DRM_DEBUG("IH: DMA trap\n");
 			radeon_fence_process(rdev, R600_RING_TYPE_DMA_INDEX);
 			break;
+		case 230: /* thermal low to high */
+			DRM_DEBUG("IH: thermal low to high\n");
+			rdev->pm.dpm.thermal.high_to_low = false;
+			queue_thermal = true;
+			break;
+		case 231: /* thermal high to low */
+			DRM_DEBUG("IH: thermal high to low\n");
+			rdev->pm.dpm.thermal.high_to_low = true;
+			queue_thermal = true;
+			break;
 		case 233: /* GUI IDLE */
 			DRM_DEBUG("IH: GUI idle\n");
 			break;
@@ -4519,6 +4632,8 @@ restart_ih:
 		schedule_work(&rdev->hotplug_work);
 	if (queue_hdmi)
 		schedule_work(&rdev->audio_work);
+	if (queue_thermal && rdev->pm.dpm_enabled)
+		schedule_work(&rdev->pm.dpm.thermal.work);
 	rdev->ih.rptr = rptr;
 	WREG32(IH_RB_RPTR, rdev->ih.rptr);
 	atomic_set(&rdev->ih.lock, 0);
diff --git a/drivers/gpu/drm/radeon/r600_dpm.c b/drivers/gpu/drm/radeon/r600_dpm.c
new file mode 100644
index 000000000000..b88f54b134ab
--- /dev/null
+++ b/drivers/gpu/drm/radeon/r600_dpm.c
@@ -0,0 +1,1048 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "r600d.h"
+#include "r600_dpm.h"
+#include "atom.h"
+
+const u32 r600_utc[R600_PM_NUMBER_OF_TC] =
+{
+	R600_UTC_DFLT_00,
+	R600_UTC_DFLT_01,
+	R600_UTC_DFLT_02,
+	R600_UTC_DFLT_03,
+	R600_UTC_DFLT_04,
+	R600_UTC_DFLT_05,
+	R600_UTC_DFLT_06,
+	R600_UTC_DFLT_07,
+	R600_UTC_DFLT_08,
+	R600_UTC_DFLT_09,
+	R600_UTC_DFLT_10,
+	R600_UTC_DFLT_11,
+	R600_UTC_DFLT_12,
+	R600_UTC_DFLT_13,
+	R600_UTC_DFLT_14,
+};
+
+const u32 r600_dtc[R600_PM_NUMBER_OF_TC] =
+{
+	R600_DTC_DFLT_00,
+	R600_DTC_DFLT_01,
+	R600_DTC_DFLT_02,
+	R600_DTC_DFLT_03,
+	R600_DTC_DFLT_04,
+	R600_DTC_DFLT_05,
+	R600_DTC_DFLT_06,
+	R600_DTC_DFLT_07,
+	R600_DTC_DFLT_08,
+	R600_DTC_DFLT_09,
+	R600_DTC_DFLT_10,
+	R600_DTC_DFLT_11,
+	R600_DTC_DFLT_12,
+	R600_DTC_DFLT_13,
+	R600_DTC_DFLT_14,
+};
+
+void r600_dpm_print_class_info(u32 class, u32 class2)
+{
+	printk("\tui class: ");
+	switch (class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) {
+	case ATOM_PPLIB_CLASSIFICATION_UI_NONE:
+	default:
+		printk("none\n");
+		break;
+	case ATOM_PPLIB_CLASSIFICATION_UI_BATTERY:
+		printk("battery\n");
+		break;
+	case ATOM_PPLIB_CLASSIFICATION_UI_BALANCED:
+		printk("balanced\n");
+		break;
+	case ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE:
+		printk("performance\n");
+		break;
+	}
+	printk("\tinternal class: ");
+	if (((class & ~ATOM_PPLIB_CLASSIFICATION_UI_MASK) == 0) &&
+	    (class2 == 0))
+		printk("none");
+	else {
+		if (class & ATOM_PPLIB_CLASSIFICATION_BOOT)
+			printk("boot ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_THERMAL)
+			printk("thermal ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_LIMITEDPOWERSOURCE)
+			printk("limited_pwr ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_REST)
+			printk("rest ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_FORCED)
+			printk("forced ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_3DPERFORMANCE)
+			printk("3d_perf ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_OVERDRIVETEMPLATE)
+			printk("ovrdrv ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+			printk("uvd ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_3DLOW)
+			printk("3d_low ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_ACPI)
+			printk("acpi ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_HD2STATE)
+			printk("uvd_hd2 ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_HDSTATE)
+			printk("uvd_hd ");
+		if (class & ATOM_PPLIB_CLASSIFICATION_SDSTATE)
+			printk("uvd_sd ");
+		if (class2 & ATOM_PPLIB_CLASSIFICATION2_LIMITEDPOWERSOURCE_2)
+			printk("limited_pwr2 ");
+		if (class2 & ATOM_PPLIB_CLASSIFICATION2_ULV)
+			printk("ulv ");
+		if (class2 & ATOM_PPLIB_CLASSIFICATION2_MVC)
+			printk("uvd_mvc ");
+	}
+	printk("\n");
+}
+
+void r600_dpm_print_cap_info(u32 caps)
+{
+	printk("\tcaps: ");
+	if (caps & ATOM_PPLIB_SINGLE_DISPLAY_ONLY)
+		printk("single_disp ");
+	if (caps & ATOM_PPLIB_SUPPORTS_VIDEO_PLAYBACK)
+		printk("video ");
+	if (caps & ATOM_PPLIB_DISALLOW_ON_DC)
+		printk("no_dc ");
+	printk("\n");
+}
+
+void r600_dpm_print_ps_status(struct radeon_device *rdev,
+			      struct radeon_ps *rps)
+{
+	printk("\tstatus: ");
+	if (rps == rdev->pm.dpm.current_ps)
+		printk("c ");
+	if (rps == rdev->pm.dpm.requested_ps)
+		printk("r ");
+	if (rps == rdev->pm.dpm.boot_ps)
+		printk("b ");
+	printk("\n");
+}
+
+u32 r600_dpm_get_vblank_time(struct radeon_device *rdev)
+{
+	struct drm_device *dev = rdev->ddev;
+	struct drm_crtc *crtc;
+	struct radeon_crtc *radeon_crtc;
+	u32 line_time_us, vblank_lines;
+	u32 vblank_time_us = 0xffffffff; /* if the displays are off, vblank time is max */
+
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		radeon_crtc = to_radeon_crtc(crtc);
+		if (crtc->enabled && radeon_crtc->enabled && radeon_crtc->hw_mode.clock) {
+			line_time_us = (radeon_crtc->hw_mode.crtc_htotal * 1000) /
+				radeon_crtc->hw_mode.clock;
+			vblank_lines = radeon_crtc->hw_mode.crtc_vblank_end -
+				radeon_crtc->hw_mode.crtc_vdisplay +
+				(radeon_crtc->v_border * 2);
+			vblank_time_us = vblank_lines * line_time_us;
+			break;
+		}
+	}
+
+	return vblank_time_us;
+}
+
+void r600_calculate_u_and_p(u32 i, u32 r_c, u32 p_b,
+			    u32 *p, u32 *u)
+{
+	u32 b_c = 0;
+	u32 i_c;
+	u32 tmp;
+
+	i_c = (i * r_c) / 100;
+	tmp = i_c >> p_b;
+
+	while (tmp) {
+		b_c++;
+		tmp >>= 1;
+	}
+
+	*u = (b_c + 1) / 2;
+	*p = i_c / (1 << (2 * (*u)));
+}
+
+int r600_calculate_at(u32 t, u32 h, u32 fh, u32 fl, u32 *tl, u32 *th)
+{
+	u32 k, a, ah, al;
+	u32 t1;
+
+	if ((fl == 0) || (fh == 0) || (fl > fh))
+		return -EINVAL;
+
+	k = (100 * fh) / fl;
+	t1 = (t * (k - 100));
+	a = (1000 * (100 * h + t1)) / (10000 + (t1 / 100));
+	a = (a + 5) / 10;
+	ah = ((a * t) + 5000) / 10000;
+	al = a - ah;
+
+	*th = t - ah;
+	*tl = t + al;
+
+	return 0;
+}
+
+void r600_gfx_clockgating_enable(struct radeon_device *rdev, bool enable)
+{
+	int i;
+
+	if (enable) {
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+	} else {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+
+		WREG32(CG_RLC_REQ_AND_RSP, 0x2);
+
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if (((RREG32(CG_RLC_REQ_AND_RSP) & CG_RLC_RSP_TYPE_MASK) >> CG_RLC_RSP_TYPE_SHIFT) == 1)
+				break;
+			udelay(1);
+		}
+
+		WREG32(CG_RLC_REQ_AND_RSP, 0x0);
+
+		WREG32(GRBM_PWR_CNTL, 0x1);
+		RREG32(GRBM_PWR_CNTL);
+	}
+}
+
+void r600_dynamicpm_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~GLOBAL_PWRMGT_EN);
+}
+
+void r600_enable_thermal_protection(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS);
+	else
+		WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS);
+}
+
+void r600_enable_acpi_pm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, STATIC_PM_EN, ~STATIC_PM_EN);
+}
+
+void r600_enable_dynamic_pcie_gen2(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, ENABLE_GEN2PCIE, ~ENABLE_GEN2PCIE);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~ENABLE_GEN2PCIE);
+}
+
+bool r600_dynamicpm_enabled(struct radeon_device *rdev)
+{
+	if (RREG32(GENERAL_PWRMGT) & GLOBAL_PWRMGT_EN)
+		return true;
+	else
+		return false;
+}
+
+void r600_enable_sclk_control(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, 0, ~SCLK_PWRMGT_OFF);
+	else
+		WREG32_P(GENERAL_PWRMGT, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF);
+}
+
+void r600_enable_mclk_control(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(MCLK_PWRMGT_CNTL, 0, ~MPLL_PWRMGT_OFF);
+	else
+		WREG32_P(MCLK_PWRMGT_CNTL, MPLL_PWRMGT_OFF, ~MPLL_PWRMGT_OFF);
+}
+
+void r600_enable_spll_bypass(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(CG_SPLL_FUNC_CNTL, SPLL_BYPASS_EN, ~SPLL_BYPASS_EN);
+	else
+		WREG32_P(CG_SPLL_FUNC_CNTL, 0, ~SPLL_BYPASS_EN);
+}
+
+void r600_wait_for_spll_change(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(CG_SPLL_FUNC_CNTL) & SPLL_CHG_STATUS)
+			break;
+		udelay(1);
+	}
+}
+
+void r600_set_bsp(struct radeon_device *rdev, u32 u, u32 p)
+{
+	WREG32(CG_BSP, BSP(p) | BSU(u));
+}
+
+void r600_set_at(struct radeon_device *rdev,
+		 u32 l_to_m, u32 m_to_h,
+		 u32 h_to_m, u32 m_to_l)
+{
+	WREG32(CG_RT, FLS(l_to_m) | FMS(m_to_h));
+	WREG32(CG_LT, FHS(h_to_m) | FMS(m_to_l));
+}
+
+void r600_set_tc(struct radeon_device *rdev,
+		 u32 index, u32 u_t, u32 d_t)
+{
+	WREG32(CG_FFCT_0 + (index * 4), UTC_0(u_t) | DTC_0(d_t));
+}
+
+void r600_select_td(struct radeon_device *rdev,
+		    enum r600_td td)
+{
+	if (td == R600_TD_AUTO)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_FORCE_TREND_SEL);
+	else
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_FORCE_TREND_SEL, ~FIR_FORCE_TREND_SEL);
+	if (td == R600_TD_UP)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_TREND_MODE);
+	if (td == R600_TD_DOWN)
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_TREND_MODE, ~FIR_TREND_MODE);
+}
+
+void r600_set_vrc(struct radeon_device *rdev, u32 vrv)
+{
+	WREG32(CG_FTV, vrv);
+}
+
+void r600_set_tpu(struct radeon_device *rdev, u32 u)
+{
+	WREG32_P(CG_TPC, TPU(u), ~TPU_MASK);
+}
+
+void r600_set_tpc(struct radeon_device *rdev, u32 c)
+{
+	WREG32_P(CG_TPC, TPCC(c), ~TPCC_MASK);
+}
+
+void r600_set_sstu(struct radeon_device *rdev, u32 u)
+{
+	WREG32_P(CG_SSP, CG_SSTU(u), ~CG_SSTU_MASK);
+}
+
+void r600_set_sst(struct radeon_device *rdev, u32 t)
+{
+	WREG32_P(CG_SSP, CG_SST(t), ~CG_SST_MASK);
+}
+
+void r600_set_git(struct radeon_device *rdev, u32 t)
+{
+	WREG32_P(CG_GIT, CG_GICST(t), ~CG_GICST_MASK);
+}
+
+void r600_set_fctu(struct radeon_device *rdev, u32 u)
+{
+	WREG32_P(CG_FC_T, FC_TU(u), ~FC_TU_MASK);
+}
+
+void r600_set_fct(struct radeon_device *rdev, u32 t)
+{
+	WREG32_P(CG_FC_T, FC_T(t), ~FC_T_MASK);
+}
+
+void r600_set_ctxcgtt3d_rphc(struct radeon_device *rdev, u32 p)
+{
+	WREG32_P(CG_CTX_CGTT3D_R, PHC(p), ~PHC_MASK);
+}
+
+void r600_set_ctxcgtt3d_rsdc(struct radeon_device *rdev, u32 s)
+{
+	WREG32_P(CG_CTX_CGTT3D_R, SDC(s), ~SDC_MASK);
+}
+
+void r600_set_vddc3d_oorsu(struct radeon_device *rdev, u32 u)
+{
+	WREG32_P(CG_VDDC3D_OOR, SU(u), ~SU_MASK);
+}
+
+void r600_set_vddc3d_oorphc(struct radeon_device *rdev, u32 p)
+{
+	WREG32_P(CG_VDDC3D_OOR, PHC(p), ~PHC_MASK);
+}
+
+void r600_set_vddc3d_oorsdc(struct radeon_device *rdev, u32 s)
+{
+	WREG32_P(CG_VDDC3D_OOR, SDC(s), ~SDC_MASK);
+}
+
+void r600_set_mpll_lock_time(struct radeon_device *rdev, u32 lock_time)
+{
+	WREG32_P(MPLL_TIME, MPLL_LOCK_TIME(lock_time), ~MPLL_LOCK_TIME_MASK);
+}
+
+void r600_set_mpll_reset_time(struct radeon_device *rdev, u32 reset_time)
+{
+	WREG32_P(MPLL_TIME, MPLL_RESET_TIME(reset_time), ~MPLL_RESET_TIME_MASK);
+}
+
+void r600_engine_clock_entry_enable(struct radeon_device *rdev,
+				    u32 index, bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART2 + (index * 4 * 2),
+			 STEP_0_SPLL_ENTRY_VALID, ~STEP_0_SPLL_ENTRY_VALID);
+	else
+		WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART2 + (index * 4 * 2),
+			 0, ~STEP_0_SPLL_ENTRY_VALID);
+}
+
+void r600_engine_clock_entry_enable_pulse_skipping(struct radeon_device *rdev,
+						   u32 index, bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART2 + (index * 4 * 2),
+			 STEP_0_SPLL_STEP_ENABLE, ~STEP_0_SPLL_STEP_ENABLE);
+	else
+		WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART2 + (index * 4 * 2),
+			 0, ~STEP_0_SPLL_STEP_ENABLE);
+}
+
+void r600_engine_clock_entry_enable_post_divider(struct radeon_device *rdev,
+						 u32 index, bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART2 + (index * 4 * 2),
+			 STEP_0_POST_DIV_EN, ~STEP_0_POST_DIV_EN);
+	else
+		WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART2 + (index * 4 * 2),
+			 0, ~STEP_0_POST_DIV_EN);
+}
+
+void r600_engine_clock_entry_set_post_divider(struct radeon_device *rdev,
+					      u32 index, u32 divider)
+{
+	WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART1 + (index * 4 * 2),
+		 STEP_0_SPLL_POST_DIV(divider), ~STEP_0_SPLL_POST_DIV_MASK);
+}
+
+void r600_engine_clock_entry_set_reference_divider(struct radeon_device *rdev,
+						   u32 index, u32 divider)
+{
+	WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART1 + (index * 4 * 2),
+		 STEP_0_SPLL_REF_DIV(divider), ~STEP_0_SPLL_REF_DIV_MASK);
+}
+
+void r600_engine_clock_entry_set_feedback_divider(struct radeon_device *rdev,
+						  u32 index, u32 divider)
+{
+	WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART1 + (index * 4 * 2),
+		 STEP_0_SPLL_FB_DIV(divider), ~STEP_0_SPLL_FB_DIV_MASK);
+}
+
+void r600_engine_clock_entry_set_step_time(struct radeon_device *rdev,
+					   u32 index, u32 step_time)
+{
+	WREG32_P(SCLK_FREQ_SETTING_STEP_0_PART1 + (index * 4 * 2),
+		 STEP_0_SPLL_STEP_TIME(step_time), ~STEP_0_SPLL_STEP_TIME_MASK);
+}
+
+void r600_vid_rt_set_ssu(struct radeon_device *rdev, u32 u)
+{
+	WREG32_P(VID_RT, SSTU(u), ~SSTU_MASK);
+}
+
+void r600_vid_rt_set_vru(struct radeon_device *rdev, u32 u)
+{
+	WREG32_P(VID_RT, VID_CRTU(u), ~VID_CRTU_MASK);
+}
+
+void r600_vid_rt_set_vrt(struct radeon_device *rdev, u32 rt)
+{
+	WREG32_P(VID_RT, VID_CRT(rt), ~VID_CRT_MASK);
+}
+
+void r600_voltage_control_enable_pins(struct radeon_device *rdev,
+				      u64 mask)
+{
+	WREG32(LOWER_GPIO_ENABLE, mask & 0xffffffff);
+	WREG32(UPPER_GPIO_ENABLE, upper_32_bits(mask));
+}
+
+
+void r600_voltage_control_program_voltages(struct radeon_device *rdev,
+					   enum r600_power_level index, u64 pins)
+{
+	u32 tmp, mask;
+	u32 ix = 3 - (3 & index);
+
+	WREG32(CTXSW_VID_LOWER_GPIO_CNTL + (ix * 4), pins & 0xffffffff);
+
+	mask = 7 << (3 * ix);
+	tmp = RREG32(VID_UPPER_GPIO_CNTL);
+	tmp = (tmp & ~mask) | ((pins >> (32 - (3 * ix))) & mask);
+	WREG32(VID_UPPER_GPIO_CNTL, tmp);
+}
+
+void r600_voltage_control_deactivate_static_control(struct radeon_device *rdev,
+						    u64 mask)
+{
+	u32 gpio;
+
+	gpio = RREG32(GPIOPAD_MASK);
+	gpio &= ~mask;
+	WREG32(GPIOPAD_MASK, gpio);
+
+	gpio = RREG32(GPIOPAD_EN);
+	gpio &= ~mask;
+	WREG32(GPIOPAD_EN, gpio);
+
+	gpio = RREG32(GPIOPAD_A);
+	gpio &= ~mask;
+	WREG32(GPIOPAD_A, gpio);
+}
+
+void r600_power_level_enable(struct radeon_device *rdev,
+			     enum r600_power_level index, bool enable)
+{
+	u32 ix = 3 - (3 & index);
+
+	if (enable)
+		WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4), CTXSW_FREQ_STATE_ENABLE,
+			 ~CTXSW_FREQ_STATE_ENABLE);
+	else
+		WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4), 0,
+			 ~CTXSW_FREQ_STATE_ENABLE);
+}
+
+void r600_power_level_set_voltage_index(struct radeon_device *rdev,
+					enum r600_power_level index, u32 voltage_index)
+{
+	u32 ix = 3 - (3 & index);
+
+	WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4),
+		 CTXSW_FREQ_VIDS_CFG_INDEX(voltage_index), ~CTXSW_FREQ_VIDS_CFG_INDEX_MASK);
+}
+
+void r600_power_level_set_mem_clock_index(struct radeon_device *rdev,
+					  enum r600_power_level index, u32 mem_clock_index)
+{
+	u32 ix = 3 - (3 & index);
+
+	WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4),
+		 CTXSW_FREQ_MCLK_CFG_INDEX(mem_clock_index), ~CTXSW_FREQ_MCLK_CFG_INDEX_MASK);
+}
+
+void r600_power_level_set_eng_clock_index(struct radeon_device *rdev,
+					  enum r600_power_level index, u32 eng_clock_index)
+{
+	u32 ix = 3 - (3 & index);
+
+	WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4),
+		 CTXSW_FREQ_SCLK_CFG_INDEX(eng_clock_index), ~CTXSW_FREQ_SCLK_CFG_INDEX_MASK);
+}
+
+void r600_power_level_set_watermark_id(struct radeon_device *rdev,
+				       enum r600_power_level index,
+				       enum r600_display_watermark watermark_id)
+{
+	u32 ix = 3 - (3 & index);
+	u32 tmp = 0;
+
+	if (watermark_id == R600_DISPLAY_WATERMARK_HIGH)
+		tmp = CTXSW_FREQ_DISPLAY_WATERMARK;
+	WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4), tmp, ~CTXSW_FREQ_DISPLAY_WATERMARK);
+}
+
+void r600_power_level_set_pcie_gen2(struct radeon_device *rdev,
+				    enum r600_power_level index, bool compatible)
+{
+	u32 ix = 3 - (3 & index);
+	u32 tmp = 0;
+
+	if (compatible)
+		tmp = CTXSW_FREQ_GEN2PCIE_VOLT;
+	WREG32_P(CTXSW_PROFILE_INDEX + (ix * 4), tmp, ~CTXSW_FREQ_GEN2PCIE_VOLT);
+}
+
+enum r600_power_level r600_power_level_get_current_index(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_PROFILE_INDEX_MASK;
+	tmp >>= CURRENT_PROFILE_INDEX_SHIFT;
+	return tmp;
+}
+
+enum r600_power_level r600_power_level_get_target_index(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & TARGET_PROFILE_INDEX_MASK;
+	tmp >>= TARGET_PROFILE_INDEX_SHIFT;
+	return tmp;
+}
+
+void r600_power_level_set_enter_index(struct radeon_device *rdev,
+				      enum r600_power_level index)
+{
+	WREG32_P(TARGET_AND_CURRENT_PROFILE_INDEX, DYN_PWR_ENTER_INDEX(index),
+		 ~DYN_PWR_ENTER_INDEX_MASK);
+}
+
+void r600_wait_for_power_level_unequal(struct radeon_device *rdev,
+				       enum r600_power_level index)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (r600_power_level_get_target_index(rdev) != index)
+			break;
+		udelay(1);
+	}
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (r600_power_level_get_current_index(rdev) != index)
+			break;
+		udelay(1);
+	}
+}
+
+void r600_wait_for_power_level(struct radeon_device *rdev,
+			       enum r600_power_level index)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (r600_power_level_get_target_index(rdev) == index)
+			break;
+		udelay(1);
+	}
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (r600_power_level_get_current_index(rdev) == index)
+			break;
+		udelay(1);
+	}
+}
+
+void r600_start_dpm(struct radeon_device *rdev)
+{
+	r600_enable_sclk_control(rdev, false);
+	r600_enable_mclk_control(rdev, false);
+
+	r600_dynamicpm_enable(rdev, true);
+
+	radeon_wait_for_vblank(rdev, 0);
+	radeon_wait_for_vblank(rdev, 1);
+
+	r600_enable_spll_bypass(rdev, true);
+	r600_wait_for_spll_change(rdev);
+	r600_enable_spll_bypass(rdev, false);
+	r600_wait_for_spll_change(rdev);
+
+	r600_enable_spll_bypass(rdev, true);
+	r600_wait_for_spll_change(rdev);
+	r600_enable_spll_bypass(rdev, false);
+	r600_wait_for_spll_change(rdev);
+
+	r600_enable_sclk_control(rdev, true);
+	r600_enable_mclk_control(rdev, true);
+}
+
+void r600_stop_dpm(struct radeon_device *rdev)
+{
+	r600_dynamicpm_enable(rdev, false);
+}
+
+int r600_dpm_pre_set_power_state(struct radeon_device *rdev)
+{
+	return 0;
+}
+
+void r600_dpm_post_set_power_state(struct radeon_device *rdev)
+{
+
+}
+
+bool r600_is_uvd_state(u32 class, u32 class2)
+{
+	if (class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		return true;
+	if (class & ATOM_PPLIB_CLASSIFICATION_HD2STATE)
+		return true;
+	if (class & ATOM_PPLIB_CLASSIFICATION_HDSTATE)
+		return true;
+	if (class & ATOM_PPLIB_CLASSIFICATION_SDSTATE)
+		return true;
+	if (class2 & ATOM_PPLIB_CLASSIFICATION2_MVC)
+		return true;
+	return false;
+}
+
+int r600_set_thermal_temperature_range(struct radeon_device *rdev,
+				       int min_temp, int max_temp)
+{
+	int low_temp = 0 * 1000;
+	int high_temp = 255 * 1000;
+
+	if (low_temp < min_temp)
+		low_temp = min_temp;
+	if (high_temp > max_temp)
+		high_temp = max_temp;
+	if (high_temp < low_temp) {
+		DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp);
+		return -EINVAL;
+	}
+
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTH(high_temp / 1000), ~DIG_THERM_INTH_MASK);
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTL(low_temp / 1000), ~DIG_THERM_INTL_MASK);
+	WREG32_P(CG_THERMAL_CTRL, DIG_THERM_DPM(high_temp / 1000), ~DIG_THERM_DPM_MASK);
+
+	rdev->pm.dpm.thermal.min_temp = low_temp;
+	rdev->pm.dpm.thermal.max_temp = high_temp;
+
+	return 0;
+}
+
+bool r600_is_internal_thermal_sensor(enum radeon_int_thermal_type sensor)
+{
+	switch (sensor) {
+	case THERMAL_TYPE_RV6XX:
+	case THERMAL_TYPE_RV770:
+	case THERMAL_TYPE_EVERGREEN:
+	case THERMAL_TYPE_SUMO:
+	case THERMAL_TYPE_NI:
+	case THERMAL_TYPE_SI:
+		return true;
+	case THERMAL_TYPE_ADT7473_WITH_INTERNAL:
+	case THERMAL_TYPE_EMC2103_WITH_INTERNAL:
+		return false; /* need special handling */
+	case THERMAL_TYPE_NONE:
+	case THERMAL_TYPE_EXTERNAL:
+	case THERMAL_TYPE_EXTERNAL_GPIO:
+	default:
+		return false;
+	}
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE4 pplib4;
+	struct _ATOM_PPLIB_POWERPLAYTABLE5 pplib5;
+};
+
+union fan_info {
+	struct _ATOM_PPLIB_FANTABLE fan;
+	struct _ATOM_PPLIB_FANTABLE2 fan2;
+};
+
+static int r600_parse_clk_voltage_dep_table(struct radeon_clock_voltage_dependency_table *radeon_table,
+					    ATOM_PPLIB_Clock_Voltage_Dependency_Table *atom_table)
+{
+	u32 size = atom_table->ucNumEntries *
+		sizeof(struct radeon_clock_voltage_dependency_entry);
+	int i;
+
+	radeon_table->entries = kzalloc(size, GFP_KERNEL);
+	if (!radeon_table->entries)
+		return -ENOMEM;
+
+	for (i = 0; i < atom_table->ucNumEntries; i++) {
+		radeon_table->entries[i].clk = le16_to_cpu(atom_table->entries[i].usClockLow) |
+			(atom_table->entries[i].ucClockHigh << 16);
+		radeon_table->entries[i].v = le16_to_cpu(atom_table->entries[i].usVoltage);
+	}
+	radeon_table->count = atom_table->ucNumEntries;
+
+	return 0;
+}
+
+/* sizeof(ATOM_PPLIB_EXTENDEDHEADER) */
+#define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V2 12
+#define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V3 14
+#define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V4 16
+#define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V5 18
+#define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V6 20
+#define SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V7 22
+
+int r600_parse_extended_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	union power_info *power_info;
+	union fan_info *fan_info;
+	ATOM_PPLIB_Clock_Voltage_Dependency_Table *dep_table;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	int ret, i;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	/* fan table */
+	if (le16_to_cpu(power_info->pplib.usTableSize) >=
+	    sizeof(struct _ATOM_PPLIB_POWERPLAYTABLE3)) {
+		if (power_info->pplib3.usFanTableOffset) {
+			fan_info = (union fan_info *)(mode_info->atom_context->bios + data_offset +
+						      le16_to_cpu(power_info->pplib3.usFanTableOffset));
+			rdev->pm.dpm.fan.t_hyst = fan_info->fan.ucTHyst;
+			rdev->pm.dpm.fan.t_min = le16_to_cpu(fan_info->fan.usTMin);
+			rdev->pm.dpm.fan.t_med = le16_to_cpu(fan_info->fan.usTMed);
+			rdev->pm.dpm.fan.t_high = le16_to_cpu(fan_info->fan.usTHigh);
+			rdev->pm.dpm.fan.pwm_min = le16_to_cpu(fan_info->fan.usPWMMin);
+			rdev->pm.dpm.fan.pwm_med = le16_to_cpu(fan_info->fan.usPWMMed);
+			rdev->pm.dpm.fan.pwm_high = le16_to_cpu(fan_info->fan.usPWMHigh);
+			if (fan_info->fan.ucFanTableFormat >= 2)
+				rdev->pm.dpm.fan.t_max = le16_to_cpu(fan_info->fan2.usTMax);
+			else
+				rdev->pm.dpm.fan.t_max = 10900;
+			rdev->pm.dpm.fan.cycle_delay = 100000;
+			rdev->pm.dpm.fan.ucode_fan_control = true;
+		}
+	}
+
+	/* clock dependancy tables, shedding tables */
+	if (le16_to_cpu(power_info->pplib.usTableSize) >=
+	    sizeof(struct _ATOM_PPLIB_POWERPLAYTABLE4)) {
+		if (power_info->pplib4.usVddcDependencyOnSCLKOffset) {
+			dep_table = (ATOM_PPLIB_Clock_Voltage_Dependency_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib4.usVddcDependencyOnSCLKOffset));
+			ret = r600_parse_clk_voltage_dep_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
+							       dep_table);
+			if (ret)
+				return ret;
+		}
+		if (power_info->pplib4.usVddciDependencyOnMCLKOffset) {
+			dep_table = (ATOM_PPLIB_Clock_Voltage_Dependency_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib4.usVddciDependencyOnMCLKOffset));
+			ret = r600_parse_clk_voltage_dep_table(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+							       dep_table);
+			if (ret) {
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries);
+				return ret;
+			}
+		}
+		if (power_info->pplib4.usVddcDependencyOnMCLKOffset) {
+			dep_table = (ATOM_PPLIB_Clock_Voltage_Dependency_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib4.usVddcDependencyOnMCLKOffset));
+			ret = r600_parse_clk_voltage_dep_table(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+							       dep_table);
+			if (ret) {
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries);
+				return ret;
+			}
+		}
+		if (power_info->pplib4.usMaxClockVoltageOnDCOffset) {
+			ATOM_PPLIB_Clock_Voltage_Limit_Table *clk_v =
+				(ATOM_PPLIB_Clock_Voltage_Limit_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib4.usMaxClockVoltageOnDCOffset));
+			if (clk_v->ucNumEntries) {
+				rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.sclk =
+					le16_to_cpu(clk_v->entries[0].usSclkLow) |
+					(clk_v->entries[0].ucSclkHigh << 16);
+				rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.mclk =
+					le16_to_cpu(clk_v->entries[0].usMclkLow) |
+					(clk_v->entries[0].ucMclkHigh << 16);
+				rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddc =
+					le16_to_cpu(clk_v->entries[0].usVddc);
+				rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddci =
+					le16_to_cpu(clk_v->entries[0].usVddci);
+			}
+		}
+		if (power_info->pplib4.usVddcPhaseShedLimitsTableOffset) {
+			ATOM_PPLIB_PhaseSheddingLimits_Table *psl =
+				(ATOM_PPLIB_PhaseSheddingLimits_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib4.usVddcPhaseShedLimitsTableOffset));
+
+			rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries =
+				kzalloc(psl->ucNumEntries *
+					sizeof(struct radeon_phase_shedding_limits_entry),
+					GFP_KERNEL);
+			if (!rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries) {
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries);
+				return -ENOMEM;
+			}
+
+			for (i = 0; i < psl->ucNumEntries; i++) {
+				rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries[i].sclk =
+					le16_to_cpu(psl->entries[i].usSclkLow) |
+					(psl->entries[i].ucSclkHigh << 16);
+				rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries[i].mclk =
+					le16_to_cpu(psl->entries[i].usMclkLow) |
+					(psl->entries[i].ucMclkHigh << 16);
+				rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries[i].voltage =
+					le16_to_cpu(psl->entries[i].usVoltage);
+			}
+			rdev->pm.dpm.dyn_state.phase_shedding_limits_table.count =
+				psl->ucNumEntries;
+		}
+	}
+
+	/* cac data */
+	if (le16_to_cpu(power_info->pplib.usTableSize) >=
+	    sizeof(struct _ATOM_PPLIB_POWERPLAYTABLE5)) {
+		rdev->pm.dpm.tdp_limit = le32_to_cpu(power_info->pplib5.ulTDPLimit);
+		rdev->pm.dpm.near_tdp_limit = le32_to_cpu(power_info->pplib5.ulNearTDPLimit);
+		rdev->pm.dpm.near_tdp_limit_adjusted = rdev->pm.dpm.near_tdp_limit;
+		rdev->pm.dpm.tdp_od_limit = le16_to_cpu(power_info->pplib5.usTDPODLimit);
+		if (rdev->pm.dpm.tdp_od_limit)
+			rdev->pm.dpm.power_control = true;
+		else
+			rdev->pm.dpm.power_control = false;
+		rdev->pm.dpm.tdp_adjustment = 0;
+		rdev->pm.dpm.sq_ramping_threshold = le32_to_cpu(power_info->pplib5.ulSQRampingThreshold);
+		rdev->pm.dpm.cac_leakage = le32_to_cpu(power_info->pplib5.ulCACLeakage);
+		rdev->pm.dpm.load_line_slope = le16_to_cpu(power_info->pplib5.usLoadLineSlope);
+		if (power_info->pplib5.usCACLeakageTableOffset) {
+			ATOM_PPLIB_CAC_Leakage_Table *cac_table =
+				(ATOM_PPLIB_CAC_Leakage_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib5.usCACLeakageTableOffset));
+			u32 size = cac_table->ucNumEntries * sizeof(struct radeon_cac_leakage_table);
+			rdev->pm.dpm.dyn_state.cac_leakage_table.entries = kzalloc(size, GFP_KERNEL);
+			if (!rdev->pm.dpm.dyn_state.cac_leakage_table.entries) {
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries);
+				return -ENOMEM;
+			}
+			for (i = 0; i < cac_table->ucNumEntries; i++) {
+				rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].vddc =
+					le16_to_cpu(cac_table->entries[i].usVddc);
+				rdev->pm.dpm.dyn_state.cac_leakage_table.entries[i].leakage =
+					le32_to_cpu(cac_table->entries[i].ulLeakageValue);
+			}
+			rdev->pm.dpm.dyn_state.cac_leakage_table.count = cac_table->ucNumEntries;
+		}
+	}
+
+	/* ppm table */
+	if (le16_to_cpu(power_info->pplib.usTableSize) >=
+	    sizeof(struct _ATOM_PPLIB_POWERPLAYTABLE3)) {
+		ATOM_PPLIB_EXTENDEDHEADER *ext_hdr = (ATOM_PPLIB_EXTENDEDHEADER *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib3.usExtendendedHeaderOffset));
+		if ((le16_to_cpu(ext_hdr->usSize) >= SIZE_OF_ATOM_PPLIB_EXTENDEDHEADER_V5) &&
+		    ext_hdr->usPPMTableOffset) {
+			ATOM_PPLIB_PPM_Table *ppm = (ATOM_PPLIB_PPM_Table *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(ext_hdr->usPPMTableOffset));
+			rdev->pm.dpm.dyn_state.ppm_table =
+				kzalloc(sizeof(struct radeon_ppm_table), GFP_KERNEL);
+			if (!rdev->pm.dpm.dyn_state.ppm_table) {
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries);
+				kfree(rdev->pm.dpm.dyn_state.cac_leakage_table.entries);
+				return -ENOMEM;
+			}
+			rdev->pm.dpm.dyn_state.ppm_table->ppm_design = ppm->ucPpmDesign;
+			rdev->pm.dpm.dyn_state.ppm_table->cpu_core_number =
+				le16_to_cpu(ppm->usCpuCoreNumber);
+			rdev->pm.dpm.dyn_state.ppm_table->platform_tdp =
+				le32_to_cpu(ppm->ulPlatformTDP);
+			rdev->pm.dpm.dyn_state.ppm_table->small_ac_platform_tdp =
+				le32_to_cpu(ppm->ulSmallACPlatformTDP);
+			rdev->pm.dpm.dyn_state.ppm_table->platform_tdc =
+				le32_to_cpu(ppm->ulPlatformTDC);
+			rdev->pm.dpm.dyn_state.ppm_table->small_ac_platform_tdc =
+				le32_to_cpu(ppm->ulSmallACPlatformTDC);
+			rdev->pm.dpm.dyn_state.ppm_table->apu_tdp =
+				le32_to_cpu(ppm->ulApuTDP);
+			rdev->pm.dpm.dyn_state.ppm_table->dgpu_tdp =
+				le32_to_cpu(ppm->ulDGpuTDP);
+			rdev->pm.dpm.dyn_state.ppm_table->dgpu_ulv_power =
+				le32_to_cpu(ppm->ulDGpuUlvPower);
+			rdev->pm.dpm.dyn_state.ppm_table->tj_max =
+				le32_to_cpu(ppm->ulTjmax);
+		}
+	}
+
+	return 0;
+}
+
+void r600_free_extended_power_table(struct radeon_device *rdev)
+{
+	if (rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries)
+		kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries);
+	if (rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries)
+		kfree(rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk.entries);
+	if (rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries)
+		kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk.entries);
+	if (rdev->pm.dpm.dyn_state.cac_leakage_table.entries)
+		kfree(rdev->pm.dpm.dyn_state.cac_leakage_table.entries);
+	if (rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries)
+		kfree(rdev->pm.dpm.dyn_state.phase_shedding_limits_table.entries);
+	if (rdev->pm.dpm.dyn_state.ppm_table)
+		kfree(rdev->pm.dpm.dyn_state.ppm_table);
+}
+
+enum radeon_pcie_gen r600_get_pcie_gen_support(struct radeon_device *rdev,
+					       u32 sys_mask,
+					       enum radeon_pcie_gen asic_gen,
+					       enum radeon_pcie_gen default_gen)
+{
+	switch (asic_gen) {
+	case RADEON_PCIE_GEN1:
+		return RADEON_PCIE_GEN1;
+	case RADEON_PCIE_GEN2:
+		return RADEON_PCIE_GEN2;
+	case RADEON_PCIE_GEN3:
+		return RADEON_PCIE_GEN3;
+	default:
+		if ((sys_mask & DRM_PCIE_SPEED_80) && (default_gen == RADEON_PCIE_GEN3))
+			return RADEON_PCIE_GEN3;
+		else if ((sys_mask & DRM_PCIE_SPEED_50) && (default_gen == RADEON_PCIE_GEN2))
+			return RADEON_PCIE_GEN2;
+		else
+			return RADEON_PCIE_GEN1;
+	}
+	return RADEON_PCIE_GEN1;
+}
diff --git a/drivers/gpu/drm/radeon/r600_dpm.h b/drivers/gpu/drm/radeon/r600_dpm.h
new file mode 100644
index 000000000000..7c822d9ae53d
--- /dev/null
+++ b/drivers/gpu/drm/radeon/r600_dpm.h
@@ -0,0 +1,227 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __R600_DPM_H__
+#define __R600_DPM_H__
+
+#define R600_ASI_DFLT                                10000
+#define R600_BSP_DFLT                                0x41EB
+#define R600_BSU_DFLT                                0x2
+#define R600_AH_DFLT                                 5
+#define R600_RLP_DFLT                                25
+#define R600_RMP_DFLT                                65
+#define R600_LHP_DFLT                                40
+#define R600_LMP_DFLT                                15
+#define R600_TD_DFLT                                 0
+#define R600_UTC_DFLT_00                             0x24
+#define R600_UTC_DFLT_01                             0x22
+#define R600_UTC_DFLT_02                             0x22
+#define R600_UTC_DFLT_03                             0x22
+#define R600_UTC_DFLT_04                             0x22
+#define R600_UTC_DFLT_05                             0x22
+#define R600_UTC_DFLT_06                             0x22
+#define R600_UTC_DFLT_07                             0x22
+#define R600_UTC_DFLT_08                             0x22
+#define R600_UTC_DFLT_09                             0x22
+#define R600_UTC_DFLT_10                             0x22
+#define R600_UTC_DFLT_11                             0x22
+#define R600_UTC_DFLT_12                             0x22
+#define R600_UTC_DFLT_13                             0x22
+#define R600_UTC_DFLT_14                             0x22
+#define R600_DTC_DFLT_00                             0x24
+#define R600_DTC_DFLT_01                             0x22
+#define R600_DTC_DFLT_02                             0x22
+#define R600_DTC_DFLT_03                             0x22
+#define R600_DTC_DFLT_04                             0x22
+#define R600_DTC_DFLT_05                             0x22
+#define R600_DTC_DFLT_06                             0x22
+#define R600_DTC_DFLT_07                             0x22
+#define R600_DTC_DFLT_08                             0x22
+#define R600_DTC_DFLT_09                             0x22
+#define R600_DTC_DFLT_10                             0x22
+#define R600_DTC_DFLT_11                             0x22
+#define R600_DTC_DFLT_12                             0x22
+#define R600_DTC_DFLT_13                             0x22
+#define R600_DTC_DFLT_14                             0x22
+#define R600_VRC_DFLT                                0x0000C003
+#define R600_VOLTAGERESPONSETIME_DFLT                1000
+#define R600_BACKBIASRESPONSETIME_DFLT               1000
+#define R600_VRU_DFLT                                0x3
+#define R600_SPLLSTEPTIME_DFLT                       0x1000
+#define R600_SPLLSTEPUNIT_DFLT                       0x3
+#define R600_TPU_DFLT                                0
+#define R600_TPC_DFLT                                0x200
+#define R600_SSTU_DFLT                               0
+#define R600_SST_DFLT                                0x00C8
+#define R600_GICST_DFLT                              0x200
+#define R600_FCT_DFLT                                0x0400
+#define R600_FCTU_DFLT                               0
+#define R600_CTXCGTT3DRPHC_DFLT                      0x20
+#define R600_CTXCGTT3DRSDC_DFLT                      0x40
+#define R600_VDDC3DOORPHC_DFLT                       0x100
+#define R600_VDDC3DOORSDC_DFLT                       0x7
+#define R600_VDDC3DOORSU_DFLT                        0
+#define R600_MPLLLOCKTIME_DFLT                       100
+#define R600_MPLLRESETTIME_DFLT                      150
+#define R600_VCOSTEPPCT_DFLT                          20
+#define R600_ENDINGVCOSTEPPCT_DFLT                    5
+#define R600_REFERENCEDIVIDER_DFLT                    4
+
+#define R600_PM_NUMBER_OF_TC 15
+#define R600_PM_NUMBER_OF_SCLKS 20
+#define R600_PM_NUMBER_OF_MCLKS 4
+#define R600_PM_NUMBER_OF_VOLTAGE_LEVELS 4
+#define R600_PM_NUMBER_OF_ACTIVITY_LEVELS 3
+
+/* XXX are these ok? */
+#define R600_TEMP_RANGE_MIN (90 * 1000)
+#define R600_TEMP_RANGE_MAX (120 * 1000)
+
+enum r600_power_level {
+	R600_POWER_LEVEL_LOW = 0,
+	R600_POWER_LEVEL_MEDIUM = 1,
+	R600_POWER_LEVEL_HIGH = 2,
+	R600_POWER_LEVEL_CTXSW = 3,
+};
+
+enum r600_td {
+	R600_TD_AUTO,
+	R600_TD_UP,
+	R600_TD_DOWN,
+};
+
+enum r600_display_watermark {
+	R600_DISPLAY_WATERMARK_LOW = 0,
+	R600_DISPLAY_WATERMARK_HIGH = 1,
+};
+
+enum r600_display_gap
+{
+    R600_PM_DISPLAY_GAP_VBLANK_OR_WM = 0,
+    R600_PM_DISPLAY_GAP_VBLANK       = 1,
+    R600_PM_DISPLAY_GAP_WATERMARK    = 2,
+    R600_PM_DISPLAY_GAP_IGNORE       = 3,
+};
+
+extern const u32 r600_utc[R600_PM_NUMBER_OF_TC];
+extern const u32 r600_dtc[R600_PM_NUMBER_OF_TC];
+
+void r600_dpm_print_class_info(u32 class, u32 class2);
+void r600_dpm_print_cap_info(u32 caps);
+void r600_dpm_print_ps_status(struct radeon_device *rdev,
+			      struct radeon_ps *rps);
+u32 r600_dpm_get_vblank_time(struct radeon_device *rdev);
+bool r600_is_uvd_state(u32 class, u32 class2);
+void r600_calculate_u_and_p(u32 i, u32 r_c, u32 p_b,
+			    u32 *p, u32 *u);
+int r600_calculate_at(u32 t, u32 h, u32 fh, u32 fl, u32 *tl, u32 *th);
+void r600_gfx_clockgating_enable(struct radeon_device *rdev, bool enable);
+void r600_dynamicpm_enable(struct radeon_device *rdev, bool enable);
+void r600_enable_thermal_protection(struct radeon_device *rdev, bool enable);
+void r600_enable_acpi_pm(struct radeon_device *rdev);
+void r600_enable_dynamic_pcie_gen2(struct radeon_device *rdev, bool enable);
+bool r600_dynamicpm_enabled(struct radeon_device *rdev);
+void r600_enable_sclk_control(struct radeon_device *rdev, bool enable);
+void r600_enable_mclk_control(struct radeon_device *rdev, bool enable);
+void r600_enable_spll_bypass(struct radeon_device *rdev, bool enable);
+void r600_wait_for_spll_change(struct radeon_device *rdev);
+void r600_set_bsp(struct radeon_device *rdev, u32 u, u32 p);
+void r600_set_at(struct radeon_device *rdev,
+		 u32 l_to_m, u32 m_to_h,
+		 u32 h_to_m, u32 m_to_l);
+void r600_set_tc(struct radeon_device *rdev, u32 index, u32 u_t, u32 d_t);
+void r600_select_td(struct radeon_device *rdev, enum r600_td td);
+void r600_set_vrc(struct radeon_device *rdev, u32 vrv);
+void r600_set_tpu(struct radeon_device *rdev, u32 u);
+void r600_set_tpc(struct radeon_device *rdev, u32 c);
+void r600_set_sstu(struct radeon_device *rdev, u32 u);
+void r600_set_sst(struct radeon_device *rdev, u32 t);
+void r600_set_git(struct radeon_device *rdev, u32 t);
+void r600_set_fctu(struct radeon_device *rdev, u32 u);
+void r600_set_fct(struct radeon_device *rdev, u32 t);
+void r600_set_ctxcgtt3d_rphc(struct radeon_device *rdev, u32 p);
+void r600_set_ctxcgtt3d_rsdc(struct radeon_device *rdev, u32 s);
+void r600_set_vddc3d_oorsu(struct radeon_device *rdev, u32 u);
+void r600_set_vddc3d_oorphc(struct radeon_device *rdev, u32 p);
+void r600_set_vddc3d_oorsdc(struct radeon_device *rdev, u32 s);
+void r600_set_mpll_lock_time(struct radeon_device *rdev, u32 lock_time);
+void r600_set_mpll_reset_time(struct radeon_device *rdev, u32 reset_time);
+void r600_engine_clock_entry_enable(struct radeon_device *rdev,
+				    u32 index, bool enable);
+void r600_engine_clock_entry_enable_pulse_skipping(struct radeon_device *rdev,
+						   u32 index, bool enable);
+void r600_engine_clock_entry_enable_post_divider(struct radeon_device *rdev,
+						 u32 index, bool enable);
+void r600_engine_clock_entry_set_post_divider(struct radeon_device *rdev,
+					      u32 index, u32 divider);
+void r600_engine_clock_entry_set_reference_divider(struct radeon_device *rdev,
+						   u32 index, u32 divider);
+void r600_engine_clock_entry_set_feedback_divider(struct radeon_device *rdev,
+						  u32 index, u32 divider);
+void r600_engine_clock_entry_set_step_time(struct radeon_device *rdev,
+					   u32 index, u32 step_time);
+void r600_vid_rt_set_ssu(struct radeon_device *rdev, u32 u);
+void r600_vid_rt_set_vru(struct radeon_device *rdev, u32 u);
+void r600_vid_rt_set_vrt(struct radeon_device *rdev, u32 rt);
+void r600_voltage_control_enable_pins(struct radeon_device *rdev,
+				      u64 mask);
+void r600_voltage_control_program_voltages(struct radeon_device *rdev,
+					   enum r600_power_level index, u64 pins);
+void r600_voltage_control_deactivate_static_control(struct radeon_device *rdev,
+						    u64 mask);
+void r600_power_level_enable(struct radeon_device *rdev,
+			     enum r600_power_level index, bool enable);
+void r600_power_level_set_voltage_index(struct radeon_device *rdev,
+					enum r600_power_level index, u32 voltage_index);
+void r600_power_level_set_mem_clock_index(struct radeon_device *rdev,
+					  enum r600_power_level index, u32 mem_clock_index);
+void r600_power_level_set_eng_clock_index(struct radeon_device *rdev,
+					  enum r600_power_level index, u32 eng_clock_index);
+void r600_power_level_set_watermark_id(struct radeon_device *rdev,
+				       enum r600_power_level index,
+				       enum r600_display_watermark watermark_id);
+void r600_power_level_set_pcie_gen2(struct radeon_device *rdev,
+				    enum r600_power_level index, bool compatible);
+enum r600_power_level r600_power_level_get_current_index(struct radeon_device *rdev);
+enum r600_power_level r600_power_level_get_target_index(struct radeon_device *rdev);
+void r600_power_level_set_enter_index(struct radeon_device *rdev,
+				      enum r600_power_level index);
+void r600_wait_for_power_level_unequal(struct radeon_device *rdev,
+				       enum r600_power_level index);
+void r600_wait_for_power_level(struct radeon_device *rdev,
+			       enum r600_power_level index);
+void r600_start_dpm(struct radeon_device *rdev);
+void r600_stop_dpm(struct radeon_device *rdev);
+
+int r600_set_thermal_temperature_range(struct radeon_device *rdev,
+				       int min_temp, int max_temp);
+bool r600_is_internal_thermal_sensor(enum radeon_int_thermal_type sensor);
+
+int r600_parse_extended_power_table(struct radeon_device *rdev);
+void r600_free_extended_power_table(struct radeon_device *rdev);
+
+enum radeon_pcie_gen r600_get_pcie_gen_support(struct radeon_device *rdev,
+					       u32 sys_mask,
+					       enum radeon_pcie_gen asic_gen,
+					       enum radeon_pcie_gen default_gen);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/r600_hdmi.c b/drivers/gpu/drm/radeon/r600_hdmi.c
index 456750a0daa5..f48240bb8c56 100644
--- a/drivers/gpu/drm/radeon/r600_hdmi.c
+++ b/drivers/gpu/drm/radeon/r600_hdmi.c
@@ -133,14 +133,7 @@ static void r600_hdmi_update_avi_infoframe(struct drm_encoder *encoder,
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 	uint32_t offset = dig->afmt->offset;
 	uint8_t *frame = buffer + 3;
-
-	/* Our header values (type, version, length) should be alright, Intel
-	 * is using the same. Checksum function also seems to be OK, it works
-	 * fine for audio infoframe. However calculated value is always lower
-	 * by 2 in comparison to fglrx. It breaks displaying anything in case
-	 * of TVs that strictly check the checksum. Hack it manually here to
-	 * workaround this issue. */
-	frame[0x0] += 2;
+	uint8_t *header = buffer;
 
 	WREG32(HDMI0_AVI_INFO0 + offset,
 		frame[0x0] | (frame[0x1] << 8) | (frame[0x2] << 16) | (frame[0x3] << 24));
@@ -149,7 +142,7 @@ static void r600_hdmi_update_avi_infoframe(struct drm_encoder *encoder,
 	WREG32(HDMI0_AVI_INFO2 + offset,
 		frame[0x8] | (frame[0x9] << 8) | (frame[0xA] << 16) | (frame[0xB] << 24));
 	WREG32(HDMI0_AVI_INFO3 + offset,
-		frame[0xC] | (frame[0xD] << 8));
+		frame[0xC] | (frame[0xD] << 8) | (header[1] << 24));
 }
 
 /*
@@ -273,6 +266,9 @@ void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mod
 	uint32_t offset;
 	ssize_t err;
 
+	if (!dig || !dig->afmt)
+		return;
+
 	/* Silent, r600_hdmi_enable will raise WARN for us */
 	if (!dig->afmt->enabled)
 		return;
@@ -455,6 +451,9 @@ void r600_hdmi_enable(struct drm_encoder *encoder, bool enable)
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 	u32 hdmi = HDMI0_ERROR_ACK;
 
+	if (!dig || !dig->afmt)
+		return;
+
 	/* Silent, r600_hdmi_enable will raise WARN for us */
 	if (enable && dig->afmt->enabled)
 		return;
diff --git a/drivers/gpu/drm/radeon/r600_reg.h b/drivers/gpu/drm/radeon/r600_reg.h
index 909219b1bf80..3ef202629e7e 100644
--- a/drivers/gpu/drm/radeon/r600_reg.h
+++ b/drivers/gpu/drm/radeon/r600_reg.h
@@ -31,6 +31,12 @@
 #define R600_PCIE_PORT_INDEX                0x0038
 #define R600_PCIE_PORT_DATA                 0x003c
 
+#define R600_RCU_INDEX                      0x0100
+#define R600_RCU_DATA                       0x0104
+
+#define R600_UVD_CTX_INDEX                  0xf4a0
+#define R600_UVD_CTX_DATA                   0xf4a4
+
 #define R600_MC_VM_FB_LOCATION			0x2180
 #define		R600_MC_FB_BASE_MASK			0x0000FFFF
 #define		R600_MC_FB_BASE_SHIFT			0
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index 79df558f8c40..8e3fe815edab 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -302,10 +302,25 @@
 #define	GRBM_SOFT_RESET					0x8020
 #define		SOFT_RESET_CP					(1<<0)
 
+#define	CG_THERMAL_CTRL					0x7F0
+#define		DIG_THERM_DPM(x)			((x) << 12)
+#define		DIG_THERM_DPM_MASK			0x000FF000
+#define		DIG_THERM_DPM_SHIFT			12
 #define	CG_THERMAL_STATUS				0x7F4
 #define		ASIC_T(x)			        ((x) << 0)
 #define		ASIC_T_MASK			        0x1FF
 #define		ASIC_T_SHIFT			        0
+#define	CG_THERMAL_INT					0x7F8
+#define		DIG_THERM_INTH(x)			((x) << 8)
+#define		DIG_THERM_INTH_MASK			0x0000FF00
+#define		DIG_THERM_INTH_SHIFT			8
+#define		DIG_THERM_INTL(x)			((x) << 16)
+#define		DIG_THERM_INTL_MASK			0x00FF0000
+#define		DIG_THERM_INTL_SHIFT			16
+#define 	THERM_INT_MASK_HIGH			(1 << 24)
+#define 	THERM_INT_MASK_LOW			(1 << 25)
+
+#define	RV770_CG_THERMAL_INT				0x734
 
 #define	HDP_HOST_PATH_CNTL				0x2C00
 #define	HDP_NONSURFACE_BASE				0x2C04
@@ -587,6 +602,7 @@
 #define		L2_BUSY						(1 << 0)
 
 #define	WAIT_UNTIL					0x8040
+#define         WAIT_CP_DMA_IDLE_bit                            (1 << 8)
 #define         WAIT_2D_IDLE_bit                                (1 << 14)
 #define         WAIT_3D_IDLE_bit                                (1 << 15)
 #define         WAIT_2D_IDLECLEAN_bit                           (1 << 16)
@@ -684,10 +700,6 @@
 #define RLC_UCODE_ADDR                                    0x3f2c
 #define RLC_UCODE_DATA                                    0x3f30
 
-/* new for TN */
-#define TN_RLC_SAVE_AND_RESTORE_BASE                      0x3f10
-#define TN_RLC_CLEAR_STATE_RESTORE_BASE                   0x3f20
-
 #define SRBM_SOFT_RESET                                   0xe60
 #       define SOFT_RESET_DMA                             (1 << 12)
 #       define SOFT_RESET_RLC                             (1 << 13)
@@ -1148,6 +1160,219 @@
 #       define AFMT_AZ_FORMAT_WTRIG_ACK      (1 << 29)
 #       define AFMT_AZ_AUDIO_ENABLE_CHG_ACK  (1 << 30)
 
+/* Power management */
+#define CG_SPLL_FUNC_CNTL                                 0x600
+#       define SPLL_RESET                                (1 << 0)
+#       define SPLL_SLEEP                                (1 << 1)
+#       define SPLL_REF_DIV(x)                           ((x) << 2)
+#       define SPLL_REF_DIV_MASK                         (7 << 2)
+#       define SPLL_FB_DIV(x)                            ((x) << 5)
+#       define SPLL_FB_DIV_MASK                          (0xff << 5)
+#       define SPLL_PULSEEN                              (1 << 13)
+#       define SPLL_PULSENUM(x)                          ((x) << 14)
+#       define SPLL_PULSENUM_MASK                        (3 << 14)
+#       define SPLL_SW_HILEN(x)                          ((x) << 16)
+#       define SPLL_SW_HILEN_MASK                        (0xf << 16)
+#       define SPLL_SW_LOLEN(x)                          ((x) << 20)
+#       define SPLL_SW_LOLEN_MASK                        (0xf << 20)
+#       define SPLL_DIVEN                                (1 << 24)
+#       define SPLL_BYPASS_EN                            (1 << 25)
+#       define SPLL_CHG_STATUS                           (1 << 29)
+#       define SPLL_CTLREQ                               (1 << 30)
+#       define SPLL_CTLACK                               (1 << 31)
+
+#define GENERAL_PWRMGT                                    0x618
+#       define GLOBAL_PWRMGT_EN                           (1 << 0)
+#       define STATIC_PM_EN                               (1 << 1)
+#       define MOBILE_SU                                  (1 << 2)
+#       define THERMAL_PROTECTION_DIS                     (1 << 3)
+#       define THERMAL_PROTECTION_TYPE                    (1 << 4)
+#       define ENABLE_GEN2PCIE                            (1 << 5)
+#       define SW_GPIO_INDEX(x)                           ((x) << 6)
+#       define SW_GPIO_INDEX_MASK                         (3 << 6)
+#       define LOW_VOLT_D2_ACPI                           (1 << 8)
+#       define LOW_VOLT_D3_ACPI                           (1 << 9)
+#       define VOLT_PWRMGT_EN                             (1 << 10)
+#define CG_TPC                                            0x61c
+#       define TPCC(x)                                    ((x) << 0)
+#       define TPCC_MASK                                  (0x7fffff << 0)
+#       define TPU(x)                                     ((x) << 23)
+#       define TPU_MASK                                   (0x1f << 23)
+#define SCLK_PWRMGT_CNTL                                  0x620
+#       define SCLK_PWRMGT_OFF                            (1 << 0)
+#       define SCLK_TURNOFF                               (1 << 1)
+#       define SPLL_TURNOFF                               (1 << 2)
+#       define SU_SCLK_USE_BCLK                           (1 << 3)
+#       define DYNAMIC_GFX_ISLAND_PWR_DOWN                (1 << 4)
+#       define DYNAMIC_GFX_ISLAND_PWR_LP                  (1 << 5)
+#       define CLK_TURN_ON_STAGGER                        (1 << 6)
+#       define CLK_TURN_OFF_STAGGER                       (1 << 7)
+#       define FIR_FORCE_TREND_SEL                        (1 << 8)
+#       define FIR_TREND_MODE                             (1 << 9)
+#       define DYN_GFX_CLK_OFF_EN                         (1 << 10)
+#       define VDDC3D_TURNOFF_D1                          (1 << 11)
+#       define VDDC3D_TURNOFF_D2                          (1 << 12)
+#       define VDDC3D_TURNOFF_D3                          (1 << 13)
+#       define SPLL_TURNOFF_D2                            (1 << 14)
+#       define SCLK_LOW_D1                                (1 << 15)
+#       define DYN_GFX_CLK_OFF_MC_EN                      (1 << 16)
+#define MCLK_PWRMGT_CNTL                                  0x624
+#       define MPLL_PWRMGT_OFF                            (1 << 0)
+#       define YCLK_TURNOFF                               (1 << 1)
+#       define MPLL_TURNOFF                               (1 << 2)
+#       define SU_MCLK_USE_BCLK                           (1 << 3)
+#       define DLL_READY                                  (1 << 4)
+#       define MC_BUSY                                    (1 << 5)
+#       define MC_INT_CNTL                                (1 << 7)
+#       define MRDCKA_SLEEP                               (1 << 8)
+#       define MRDCKB_SLEEP                               (1 << 9)
+#       define MRDCKC_SLEEP                               (1 << 10)
+#       define MRDCKD_SLEEP                               (1 << 11)
+#       define MRDCKE_SLEEP                               (1 << 12)
+#       define MRDCKF_SLEEP                               (1 << 13)
+#       define MRDCKG_SLEEP                               (1 << 14)
+#       define MRDCKH_SLEEP                               (1 << 15)
+#       define MRDCKA_RESET                               (1 << 16)
+#       define MRDCKB_RESET                               (1 << 17)
+#       define MRDCKC_RESET                               (1 << 18)
+#       define MRDCKD_RESET                               (1 << 19)
+#       define MRDCKE_RESET                               (1 << 20)
+#       define MRDCKF_RESET                               (1 << 21)
+#       define MRDCKG_RESET                               (1 << 22)
+#       define MRDCKH_RESET                               (1 << 23)
+#       define DLL_READY_READ                             (1 << 24)
+#       define USE_DISPLAY_GAP                            (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                  (1 << 26)
+#       define USE_DISPLAY_GAP_CTXSW                      (1 << 27)
+#       define MPLL_TURNOFF_D2                            (1 << 28)
+#       define USE_DISPLAY_URGENT_CTXSW                   (1 << 29)
+
+#define MPLL_TIME                                         0x634
+#       define MPLL_LOCK_TIME(x)                          ((x) << 0)
+#       define MPLL_LOCK_TIME_MASK                        (0xffff << 0)
+#       define MPLL_RESET_TIME(x)                         ((x) << 16)
+#       define MPLL_RESET_TIME_MASK                       (0xffff << 16)
+
+#define SCLK_FREQ_SETTING_STEP_0_PART1                    0x648
+#       define STEP_0_SPLL_POST_DIV(x)                    ((x) << 0)
+#       define STEP_0_SPLL_POST_DIV_MASK                  (0xff << 0)
+#       define STEP_0_SPLL_FB_DIV(x)                      ((x) << 8)
+#       define STEP_0_SPLL_FB_DIV_MASK                    (0xff << 8)
+#       define STEP_0_SPLL_REF_DIV(x)                     ((x) << 16)
+#       define STEP_0_SPLL_REF_DIV_MASK                   (7 << 16)
+#       define STEP_0_SPLL_STEP_TIME(x)                   ((x) << 19)
+#       define STEP_0_SPLL_STEP_TIME_MASK                 (0x1fff << 19)
+#define SCLK_FREQ_SETTING_STEP_0_PART2                    0x64c
+#       define STEP_0_PULSE_HIGH_CNT(x)                   ((x) << 0)
+#       define STEP_0_PULSE_HIGH_CNT_MASK                 (0x1ff << 0)
+#       define STEP_0_POST_DIV_EN                         (1 << 9)
+#       define STEP_0_SPLL_STEP_ENABLE                    (1 << 30)
+#       define STEP_0_SPLL_ENTRY_VALID                    (1 << 31)
+
+#define VID_RT                                            0x6f8
+#       define VID_CRT(x)                                 ((x) << 0)
+#       define VID_CRT_MASK                               (0x1fff << 0)
+#       define VID_CRTU(x)                                ((x) << 13)
+#       define VID_CRTU_MASK                              (7 << 13)
+#       define SSTU(x)                                    ((x) << 16)
+#       define SSTU_MASK                                  (7 << 16)
+#define CTXSW_PROFILE_INDEX                               0x6fc
+#       define CTXSW_FREQ_VIDS_CFG_INDEX(x)               ((x) << 0)
+#       define CTXSW_FREQ_VIDS_CFG_INDEX_MASK             (3 << 0)
+#       define CTXSW_FREQ_VIDS_CFG_INDEX_SHIFT            0
+#       define CTXSW_FREQ_MCLK_CFG_INDEX(x)               ((x) << 2)
+#       define CTXSW_FREQ_MCLK_CFG_INDEX_MASK             (3 << 2)
+#       define CTXSW_FREQ_MCLK_CFG_INDEX_SHIFT            2
+#       define CTXSW_FREQ_SCLK_CFG_INDEX(x)               ((x) << 4)
+#       define CTXSW_FREQ_SCLK_CFG_INDEX_MASK             (0x1f << 4)
+#       define CTXSW_FREQ_SCLK_CFG_INDEX_SHIFT            4
+#       define CTXSW_FREQ_STATE_SPLL_RESET_EN             (1 << 9)
+#       define CTXSW_FREQ_STATE_ENABLE                    (1 << 10)
+#       define CTXSW_FREQ_DISPLAY_WATERMARK               (1 << 11)
+#       define CTXSW_FREQ_GEN2PCIE_VOLT                   (1 << 12)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                  0x70c
+#       define TARGET_PROFILE_INDEX_MASK                  (3 << 0)
+#       define TARGET_PROFILE_INDEX_SHIFT                 0
+#       define CURRENT_PROFILE_INDEX_MASK                 (3 << 2)
+#       define CURRENT_PROFILE_INDEX_SHIFT                2
+#       define DYN_PWR_ENTER_INDEX(x)                     ((x) << 4)
+#       define DYN_PWR_ENTER_INDEX_MASK                   (3 << 4)
+#       define DYN_PWR_ENTER_INDEX_SHIFT                  4
+#       define CURR_MCLK_INDEX_MASK                       (3 << 6)
+#       define CURR_MCLK_INDEX_SHIFT                      6
+#       define CURR_SCLK_INDEX_MASK                       (0x1f << 8)
+#       define CURR_SCLK_INDEX_SHIFT                      8
+#       define CURR_VID_INDEX_MASK                        (3 << 13)
+#       define CURR_VID_INDEX_SHIFT                       13
+
+#define LOWER_GPIO_ENABLE                                 0x710
+#define UPPER_GPIO_ENABLE                                 0x714
+#define CTXSW_VID_LOWER_GPIO_CNTL                         0x718
+
+#define VID_UPPER_GPIO_CNTL                               0x740
+#define CG_CTX_CGTT3D_R                                   0x744
+#       define PHC(x)                                     ((x) << 0)
+#       define PHC_MASK                                   (0x1ff << 0)
+#       define SDC(x)                                     ((x) << 9)
+#       define SDC_MASK                                   (0x3fff << 9)
+#define CG_VDDC3D_OOR                                     0x748
+#       define SU(x)                                      ((x) << 23)
+#       define SU_MASK                                    (0xf << 23)
+#define CG_FTV                                            0x74c
+#define CG_FFCT_0                                         0x750
+#       define UTC_0(x)                                   ((x) << 0)
+#       define UTC_0_MASK                                 (0x3ff << 0)
+#       define DTC_0(x)                                   ((x) << 10)
+#       define DTC_0_MASK                                 (0x3ff << 10)
+
+#define CG_BSP                                            0x78c
+#       define BSP(x)                                     ((x) << 0)
+#       define BSP_MASK                                   (0xffff << 0)
+#       define BSU(x)                                     ((x) << 16)
+#       define BSU_MASK                                   (0xf << 16)
+#define CG_RT                                             0x790
+#       define FLS(x)                                     ((x) << 0)
+#       define FLS_MASK                                   (0xffff << 0)
+#       define FMS(x)                                     ((x) << 16)
+#       define FMS_MASK                                   (0xffff << 16)
+#define CG_LT                                             0x794
+#       define FHS(x)                                     ((x) << 0)
+#       define FHS_MASK                                   (0xffff << 0)
+#define CG_GIT                                            0x798
+#       define CG_GICST(x)                                ((x) << 0)
+#       define CG_GICST_MASK                              (0xffff << 0)
+#       define CG_GIPOT(x)                                ((x) << 16)
+#       define CG_GIPOT_MASK                              (0xffff << 16)
+
+#define CG_SSP                                            0x7a8
+#       define CG_SST(x)                                  ((x) << 0)
+#       define CG_SST_MASK                                (0xffff << 0)
+#       define CG_SSTU(x)                                 ((x) << 16)
+#       define CG_SSTU_MASK                               (0xf << 16)
+
+#define CG_RLC_REQ_AND_RSP                                0x7c4
+#       define RLC_CG_REQ_TYPE_MASK                       0xf
+#       define RLC_CG_REQ_TYPE_SHIFT                      0
+#       define CG_RLC_RSP_TYPE_MASK                       0xf0
+#       define CG_RLC_RSP_TYPE_SHIFT                      4
+
+#define CG_FC_T                                           0x7cc
+#       define FC_T(x)                                    ((x) << 0)
+#       define FC_T_MASK                                  (0xffff << 0)
+#       define FC_TU(x)                                   ((x) << 16)
+#       define FC_TU_MASK                                 (0x1f << 16)
+
+#define GPIOPAD_MASK                                      0x1798
+#define GPIOPAD_A                                         0x179c
+#define GPIOPAD_EN                                        0x17a0
+
+#define GRBM_PWR_CNTL                                     0x800c
+#       define REQ_TYPE_MASK                              0xf
+#       define REQ_TYPE_SHIFT                             0
+#       define RSP_TYPE_MASK                              0xf0
+#       define RSP_TYPE_SHIFT                             4
+
 /*
  * UVD
  */
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 142ce6cc69f5..2f08219c39b6 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -96,6 +96,8 @@ extern int radeon_pcie_gen2;
 extern int radeon_msi;
 extern int radeon_lockup_timeout;
 extern int radeon_fastfb;
+extern int radeon_dpm;
+extern int radeon_aspm;
 
 /*
  * Copy from radeon_drv.h so we don't have to include both and have conflicting
@@ -150,6 +152,13 @@ extern int radeon_fastfb;
 #define RADEON_RESET_MC				(1 << 10)
 #define RADEON_RESET_DISPLAY			(1 << 11)
 
+/* max cursor sizes (in pixels) */
+#define CURSOR_WIDTH 64
+#define CURSOR_HEIGHT 64
+
+#define CIK_CURSOR_WIDTH 128
+#define CIK_CURSOR_HEIGHT 128
+
 /*
  * Errata workarounds.
  */
@@ -192,6 +201,7 @@ struct radeon_clock {
 	uint32_t default_mclk;
 	uint32_t default_sclk;
 	uint32_t default_dispclk;
+	uint32_t current_dispclk;
 	uint32_t dp_extclk;
 	uint32_t max_pixel_clock;
 };
@@ -211,13 +221,51 @@ int radeon_atom_get_clock_dividers(struct radeon_device *rdev,
 				   u32 clock,
 				   bool strobe_mode,
 				   struct atom_clock_dividers *dividers);
+int radeon_atom_get_memory_pll_dividers(struct radeon_device *rdev,
+					u32 clock,
+					bool strobe_mode,
+					struct atom_mpll_param *mpll_param);
 void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 voltage_type);
+int radeon_atom_get_voltage_gpio_settings(struct radeon_device *rdev,
+					  u16 voltage_level, u8 voltage_type,
+					  u32 *gpio_value, u32 *gpio_mask);
+void radeon_atom_set_engine_dram_timings(struct radeon_device *rdev,
+					 u32 eng_clock, u32 mem_clock);
+int radeon_atom_get_voltage_step(struct radeon_device *rdev,
+				 u8 voltage_type, u16 *voltage_step);
+int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
+			     u16 voltage_id, u16 *voltage);
+int radeon_atom_get_leakage_vddc_based_on_leakage_idx(struct radeon_device *rdev,
+						      u16 *voltage,
+						      u16 leakage_idx);
+int radeon_atom_round_to_true_voltage(struct radeon_device *rdev,
+				      u8 voltage_type,
+				      u16 nominal_voltage,
+				      u16 *true_voltage);
+int radeon_atom_get_min_voltage(struct radeon_device *rdev,
+				u8 voltage_type, u16 *min_voltage);
+int radeon_atom_get_max_voltage(struct radeon_device *rdev,
+				u8 voltage_type, u16 *max_voltage);
+int radeon_atom_get_voltage_table(struct radeon_device *rdev,
+				  u8 voltage_type, u8 voltage_mode,
+				  struct atom_voltage_table *voltage_table);
+bool radeon_atom_is_voltage_gpio(struct radeon_device *rdev,
+				 u8 voltage_type, u8 voltage_mode);
+void radeon_atom_update_memory_dll(struct radeon_device *rdev,
+				   u32 mem_clock);
+void radeon_atom_set_ac_timing(struct radeon_device *rdev,
+			       u32 mem_clock);
+int radeon_atom_init_mc_reg_table(struct radeon_device *rdev,
+				  u8 module_index,
+				  struct atom_mc_reg_table *reg_table);
+int radeon_atom_get_memory_info(struct radeon_device *rdev,
+				u8 module_index, struct atom_memory_info *mem_info);
+int radeon_atom_get_mclk_range_table(struct radeon_device *rdev,
+				     bool gddr5, u8 module_index,
+				     struct atom_memory_clock_range_table *mclk_range_table);
+int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
+			     u16 voltage_id, u16 *voltage);
 void rs690_pm_info(struct radeon_device *rdev);
-extern int rv6xx_get_temp(struct radeon_device *rdev);
-extern int rv770_get_temp(struct radeon_device *rdev);
-extern int evergreen_get_temp(struct radeon_device *rdev);
-extern int sumo_get_temp(struct radeon_device *rdev);
-extern int si_get_temp(struct radeon_device *rdev);
 extern void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw,
 				    unsigned *bankh, unsigned *mtaspect,
 				    unsigned *tile_split);
@@ -408,6 +456,7 @@ struct radeon_sa_manager {
 	uint64_t		gpu_addr;
 	void			*cpu_ptr;
 	uint32_t		domain;
+	uint32_t		align;
 };
 
 struct radeon_sa_bo;
@@ -549,6 +598,20 @@ struct radeon_scratch {
 int radeon_scratch_get(struct radeon_device *rdev, uint32_t *reg);
 void radeon_scratch_free(struct radeon_device *rdev, uint32_t reg);
 
+/*
+ * GPU doorbell structures, functions & helpers
+ */
+struct radeon_doorbell {
+	u32			num_pages;
+	bool			free[1024];
+	/* doorbell mmio */
+	resource_size_t			base;
+	resource_size_t			size;
+	void __iomem			*ptr;
+};
+
+int radeon_doorbell_get(struct radeon_device *rdev, u32 *page);
+void radeon_doorbell_free(struct radeon_device *rdev, u32 doorbell);
 
 /*
  * IRQS.
@@ -600,10 +663,21 @@ struct evergreen_irq_stat_regs {
 	u32 afmt_status6;
 };
 
+struct cik_irq_stat_regs {
+	u32 disp_int;
+	u32 disp_int_cont;
+	u32 disp_int_cont2;
+	u32 disp_int_cont3;
+	u32 disp_int_cont4;
+	u32 disp_int_cont5;
+	u32 disp_int_cont6;
+};
+
 union radeon_irq_stat_regs {
 	struct r500_irq_stat_regs r500;
 	struct r600_irq_stat_regs r600;
 	struct evergreen_irq_stat_regs evergreen;
+	struct cik_irq_stat_regs cik;
 };
 
 #define RADEON_MAX_HPD_PINS 6
@@ -620,6 +694,7 @@ struct radeon_irq {
 	bool				hpd[RADEON_MAX_HPD_PINS];
 	bool				afmt[RADEON_MAX_AFMT_BLOCKS];
 	union radeon_irq_stat_regs	stat_regs;
+	bool				dpm_thermal;
 };
 
 int radeon_irq_kms_init(struct radeon_device *rdev);
@@ -677,6 +752,22 @@ struct radeon_ring {
 	u32			idx;
 	u64			last_semaphore_signal_addr;
 	u64			last_semaphore_wait_addr;
+	/* for CIK queues */
+	u32 me;
+	u32 pipe;
+	u32 queue;
+	struct radeon_bo	*mqd_obj;
+	u32 doorbell_page_num;
+	u32 doorbell_offset;
+	unsigned		wptr_offs;
+};
+
+struct radeon_mec {
+	struct radeon_bo	*hpd_eop_obj;
+	u64			hpd_eop_gpu_addr;
+	u32 num_pipe;
+	u32 num_mec;
+	u32 num_queue;
 };
 
 /*
@@ -694,6 +785,11 @@ struct radeon_ring {
 /* number of entries in page table */
 #define RADEON_VM_PTE_COUNT (1 << RADEON_VM_BLOCK_SIZE)
 
+/* PTBs (Page Table Blocks) need to be aligned to 32K */
+#define RADEON_VM_PTB_ALIGN_SIZE   32768
+#define RADEON_VM_PTB_ALIGN_MASK (RADEON_VM_PTB_ALIGN_SIZE - 1)
+#define RADEON_VM_PTB_ALIGN(a) (((a) + RADEON_VM_PTB_ALIGN_MASK) & ~RADEON_VM_PTB_ALIGN_MASK)
+
 struct radeon_vm {
 	struct list_head		list;
 	struct list_head		va;
@@ -778,15 +874,22 @@ struct r600_blit {
 };
 
 /*
- * SI RLC stuff
+ * RLC stuff
  */
-struct si_rlc {
+#include "clearstate_defs.h"
+
+struct radeon_rlc {
 	/* for power gating */
 	struct radeon_bo	*save_restore_obj;
 	uint64_t		save_restore_gpu_addr;
+	volatile uint32_t	*sr_ptr;
+	u32                     *reg_list;
+	u32                     reg_list_size;
 	/* for clear state */
 	struct radeon_bo	*clear_state_obj;
 	uint64_t		clear_state_gpu_addr;
+	volatile uint32_t	*cs_ptr;
+	struct cs_section_def   *cs_data;
 };
 
 int radeon_ib_get(struct radeon_device *rdev, int ring,
@@ -883,6 +986,7 @@ struct radeon_cs_parser {
 	u32			cs_flags;
 	u32			ring;
 	s32			priority;
+	struct ww_acquire_ctx	ticket;
 };
 
 extern int radeon_cs_finish_pages(struct radeon_cs_parser *p);
@@ -934,6 +1038,8 @@ struct radeon_wb {
 #define CAYMAN_WB_DMA1_RPTR_OFFSET   2304
 #define R600_WB_UVD_RPTR_OFFSET  2560
 #define R600_WB_EVENT_OFFSET     3072
+#define CIK_WB_CP1_WPTR_OFFSET     3328
+#define CIK_WB_CP2_WPTR_OFFSET     3584
 
 /**
  * struct radeon_pm - power management datas
@@ -958,6 +1064,7 @@ struct radeon_wb {
 enum radeon_pm_method {
 	PM_METHOD_PROFILE,
 	PM_METHOD_DYNPM,
+	PM_METHOD_DPM,
 };
 
 enum radeon_dynpm_state {
@@ -983,11 +1090,24 @@ enum radeon_voltage_type {
 };
 
 enum radeon_pm_state_type {
+	/* not used for dpm */
 	POWER_STATE_TYPE_DEFAULT,
 	POWER_STATE_TYPE_POWERSAVE,
+	/* user selectable states */
 	POWER_STATE_TYPE_BATTERY,
 	POWER_STATE_TYPE_BALANCED,
 	POWER_STATE_TYPE_PERFORMANCE,
+	/* internal states */
+	POWER_STATE_TYPE_INTERNAL_UVD,
+	POWER_STATE_TYPE_INTERNAL_UVD_SD,
+	POWER_STATE_TYPE_INTERNAL_UVD_HD,
+	POWER_STATE_TYPE_INTERNAL_UVD_HD2,
+	POWER_STATE_TYPE_INTERNAL_UVD_MVC,
+	POWER_STATE_TYPE_INTERNAL_BOOT,
+	POWER_STATE_TYPE_INTERNAL_THERMAL,
+	POWER_STATE_TYPE_INTERNAL_ACPI,
+	POWER_STATE_TYPE_INTERNAL_ULV,
+	POWER_STATE_TYPE_INTERNAL_3DPERF,
 };
 
 enum radeon_pm_profile_type {
@@ -1016,12 +1136,17 @@ struct radeon_pm_profile {
 
 enum radeon_int_thermal_type {
 	THERMAL_TYPE_NONE,
+	THERMAL_TYPE_EXTERNAL,
+	THERMAL_TYPE_EXTERNAL_GPIO,
 	THERMAL_TYPE_RV6XX,
 	THERMAL_TYPE_RV770,
+	THERMAL_TYPE_ADT7473_WITH_INTERNAL,
 	THERMAL_TYPE_EVERGREEN,
 	THERMAL_TYPE_SUMO,
 	THERMAL_TYPE_NI,
 	THERMAL_TYPE_SI,
+	THERMAL_TYPE_EMC2103_WITH_INTERNAL,
+	THERMAL_TYPE_CI,
 };
 
 struct radeon_voltage {
@@ -1075,6 +1200,201 @@ struct radeon_power_state {
  */
 #define RADEON_MODE_OVERCLOCK_MARGIN 500 /* 5 MHz */
 
+enum radeon_dpm_auto_throttle_src {
+	RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL,
+	RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL
+};
+
+enum radeon_dpm_event_src {
+	RADEON_DPM_EVENT_SRC_ANALOG = 0,
+	RADEON_DPM_EVENT_SRC_EXTERNAL = 1,
+	RADEON_DPM_EVENT_SRC_DIGITAL = 2,
+	RADEON_DPM_EVENT_SRC_ANALOG_OR_EXTERNAL = 3,
+	RADEON_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL = 4
+};
+
+struct radeon_ps {
+	u32 caps; /* vbios flags */
+	u32 class; /* vbios flags */
+	u32 class2; /* vbios flags */
+	/* UVD clocks */
+	u32 vclk;
+	u32 dclk;
+	/* asic priv */
+	void *ps_priv;
+};
+
+struct radeon_dpm_thermal {
+	/* thermal interrupt work */
+	struct work_struct work;
+	/* low temperature threshold */
+	int                min_temp;
+	/* high temperature threshold */
+	int                max_temp;
+	/* was interrupt low to high or high to low */
+	bool               high_to_low;
+};
+
+enum radeon_clk_action
+{
+	RADEON_SCLK_UP = 1,
+	RADEON_SCLK_DOWN
+};
+
+struct radeon_blacklist_clocks
+{
+	u32 sclk;
+	u32 mclk;
+	enum radeon_clk_action action;
+};
+
+struct radeon_clock_and_voltage_limits {
+	u32 sclk;
+	u32 mclk;
+	u32 vddc;
+	u32 vddci;
+};
+
+struct radeon_clock_array {
+	u32 count;
+	u32 *values;
+};
+
+struct radeon_clock_voltage_dependency_entry {
+	u32 clk;
+	u16 v;
+};
+
+struct radeon_clock_voltage_dependency_table {
+	u32 count;
+	struct radeon_clock_voltage_dependency_entry *entries;
+};
+
+struct radeon_cac_leakage_entry {
+	u16 vddc;
+	u32 leakage;
+};
+
+struct radeon_cac_leakage_table {
+	u32 count;
+	struct radeon_cac_leakage_entry *entries;
+};
+
+struct radeon_phase_shedding_limits_entry {
+	u16 voltage;
+	u32 sclk;
+	u32 mclk;
+};
+
+struct radeon_phase_shedding_limits_table {
+	u32 count;
+	struct radeon_phase_shedding_limits_entry *entries;
+};
+
+struct radeon_ppm_table {
+	u8 ppm_design;
+	u16 cpu_core_number;
+	u32 platform_tdp;
+	u32 small_ac_platform_tdp;
+	u32 platform_tdc;
+	u32 small_ac_platform_tdc;
+	u32 apu_tdp;
+	u32 dgpu_tdp;
+	u32 dgpu_ulv_power;
+	u32 tj_max;
+};
+
+struct radeon_dpm_dynamic_state {
+	struct radeon_clock_voltage_dependency_table vddc_dependency_on_sclk;
+	struct radeon_clock_voltage_dependency_table vddci_dependency_on_mclk;
+	struct radeon_clock_voltage_dependency_table vddc_dependency_on_mclk;
+	struct radeon_clock_voltage_dependency_table vddc_dependency_on_dispclk;
+	struct radeon_clock_array valid_sclk_values;
+	struct radeon_clock_array valid_mclk_values;
+	struct radeon_clock_and_voltage_limits max_clock_voltage_on_dc;
+	struct radeon_clock_and_voltage_limits max_clock_voltage_on_ac;
+	u32 mclk_sclk_ratio;
+	u32 sclk_mclk_delta;
+	u16 vddc_vddci_delta;
+	u16 min_vddc_for_pcie_gen2;
+	struct radeon_cac_leakage_table cac_leakage_table;
+	struct radeon_phase_shedding_limits_table phase_shedding_limits_table;
+	struct radeon_ppm_table *ppm_table;
+};
+
+struct radeon_dpm_fan {
+	u16 t_min;
+	u16 t_med;
+	u16 t_high;
+	u16 pwm_min;
+	u16 pwm_med;
+	u16 pwm_high;
+	u8 t_hyst;
+	u32 cycle_delay;
+	u16 t_max;
+	bool ucode_fan_control;
+};
+
+enum radeon_pcie_gen {
+	RADEON_PCIE_GEN1 = 0,
+	RADEON_PCIE_GEN2 = 1,
+	RADEON_PCIE_GEN3 = 2,
+	RADEON_PCIE_GEN_INVALID = 0xffff
+};
+
+enum radeon_dpm_forced_level {
+	RADEON_DPM_FORCED_LEVEL_AUTO = 0,
+	RADEON_DPM_FORCED_LEVEL_LOW = 1,
+	RADEON_DPM_FORCED_LEVEL_HIGH = 2,
+};
+
+struct radeon_dpm {
+	struct radeon_ps        *ps;
+	/* number of valid power states */
+	int                     num_ps;
+	/* current power state that is active */
+	struct radeon_ps        *current_ps;
+	/* requested power state */
+	struct radeon_ps        *requested_ps;
+	/* boot up power state */
+	struct radeon_ps        *boot_ps;
+	/* default uvd power state */
+	struct radeon_ps        *uvd_ps;
+	enum radeon_pm_state_type state;
+	enum radeon_pm_state_type user_state;
+	u32                     platform_caps;
+	u32                     voltage_response_time;
+	u32                     backbias_response_time;
+	void                    *priv;
+	u32			new_active_crtcs;
+	int			new_active_crtc_count;
+	u32			current_active_crtcs;
+	int			current_active_crtc_count;
+	struct radeon_dpm_dynamic_state dyn_state;
+	struct radeon_dpm_fan fan;
+	u32 tdp_limit;
+	u32 near_tdp_limit;
+	u32 near_tdp_limit_adjusted;
+	u32 sq_ramping_threshold;
+	u32 cac_leakage;
+	u16 tdp_od_limit;
+	u32 tdp_adjustment;
+	u16 load_line_slope;
+	bool power_control;
+	bool ac_power;
+	/* special states active */
+	bool                    thermal_active;
+	bool                    uvd_active;
+	/* thermal handling */
+	struct radeon_dpm_thermal thermal;
+	/* forced levels */
+	enum radeon_dpm_forced_level forced_level;
+};
+
+void radeon_dpm_enable_power_state(struct radeon_device *rdev,
+				    enum radeon_pm_state_type dpm_state);
+
+
 struct radeon_pm {
 	struct mutex		mutex;
 	/* write locked while reprogramming mclk */
@@ -1128,6 +1448,9 @@ struct radeon_pm {
 	/* internal thermal controller on rv6xx+ */
 	enum radeon_int_thermal_type int_thermal_type;
 	struct device	        *int_hwmon_dev;
+	/* dpm */
+	bool                    dpm_enabled;
+	struct radeon_dpm       dpm;
 };
 
 int radeon_pm_get_type_index(struct radeon_device *rdev,
@@ -1144,6 +1467,8 @@ struct radeon_uvd {
 	struct radeon_bo	*vcpu_bo;
 	void			*cpu_addr;
 	uint64_t		gpu_addr;
+	void			*saved_bo;
+	unsigned		fw_size;
 	atomic_t		handles[RADEON_MAX_UVD_HANDLES];
 	struct drm_file		*filp[RADEON_MAX_UVD_HANDLES];
 	struct delayed_work	idle_work;
@@ -1266,6 +1591,10 @@ struct radeon_asic {
 		int (*ib_test)(struct radeon_device *rdev, struct radeon_ring *cp);
 		bool (*is_lockup)(struct radeon_device *rdev, struct radeon_ring *cp);
 		void (*vm_flush)(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+
+		u32 (*get_rptr)(struct radeon_device *rdev, struct radeon_ring *ring);
+		u32 (*get_wptr)(struct radeon_device *rdev, struct radeon_ring *ring);
+		void (*set_wptr)(struct radeon_device *rdev, struct radeon_ring *ring);
 	} ring[RADEON_NUM_RINGS];
 	/* irqs */
 	struct {
@@ -1325,7 +1654,7 @@ struct radeon_asic {
 		bool (*sense)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
 		void (*set_polarity)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
 	} hpd;
-	/* power management */
+	/* static power management */
 	struct {
 		void (*misc)(struct radeon_device *rdev);
 		void (*prepare)(struct radeon_device *rdev);
@@ -1340,7 +1669,26 @@ struct radeon_asic {
 		void (*set_pcie_lanes)(struct radeon_device *rdev, int lanes);
 		void (*set_clock_gating)(struct radeon_device *rdev, int enable);
 		int (*set_uvd_clocks)(struct radeon_device *rdev, u32 vclk, u32 dclk);
+		int (*get_temperature)(struct radeon_device *rdev);
 	} pm;
+	/* dynamic power management */
+	struct {
+		int (*init)(struct radeon_device *rdev);
+		void (*setup_asic)(struct radeon_device *rdev);
+		int (*enable)(struct radeon_device *rdev);
+		void (*disable)(struct radeon_device *rdev);
+		int (*pre_set_power_state)(struct radeon_device *rdev);
+		int (*set_power_state)(struct radeon_device *rdev);
+		void (*post_set_power_state)(struct radeon_device *rdev);
+		void (*display_configuration_changed)(struct radeon_device *rdev);
+		void (*fini)(struct radeon_device *rdev);
+		u32 (*get_sclk)(struct radeon_device *rdev, bool low);
+		u32 (*get_mclk)(struct radeon_device *rdev, bool low);
+		void (*print_power_state)(struct radeon_device *rdev, struct radeon_ps *ps);
+		void (*debugfs_print_current_performance_level)(struct radeon_device *rdev, struct seq_file *m);
+		int (*force_performance_level)(struct radeon_device *rdev, enum radeon_dpm_forced_level level);
+		bool (*vblank_too_short)(struct radeon_device *rdev);
+	} dpm;
 	/* pageflipping */
 	struct {
 		void (*pre_page_flip)(struct radeon_device *rdev, int crtc);
@@ -1505,6 +1853,36 @@ struct si_asic {
 	uint32_t tile_mode_array[32];
 };
 
+struct cik_asic {
+	unsigned max_shader_engines;
+	unsigned max_tile_pipes;
+	unsigned max_cu_per_sh;
+	unsigned max_sh_per_se;
+	unsigned max_backends_per_se;
+	unsigned max_texture_channel_caches;
+	unsigned max_gprs;
+	unsigned max_gs_threads;
+	unsigned max_hw_contexts;
+	unsigned sc_prim_fifo_size_frontend;
+	unsigned sc_prim_fifo_size_backend;
+	unsigned sc_hiz_tile_fifo_size;
+	unsigned sc_earlyz_tile_fifo_size;
+
+	unsigned num_tile_pipes;
+	unsigned num_backends_per_se;
+	unsigned backend_disable_mask_per_asic;
+	unsigned backend_map;
+	unsigned num_texture_channel_caches;
+	unsigned mem_max_burst_length_bytes;
+	unsigned mem_row_size_in_kb;
+	unsigned shader_engine_tile_size;
+	unsigned num_gpus;
+	unsigned multi_gpu_tile_size;
+
+	unsigned tile_config;
+	uint32_t tile_mode_array[32];
+};
+
 union radeon_asic_config {
 	struct r300_asic	r300;
 	struct r100_asic	r100;
@@ -1513,6 +1891,7 @@ union radeon_asic_config {
 	struct evergreen_asic	evergreen;
 	struct cayman_asic	cayman;
 	struct si_asic		si;
+	struct cik_asic		cik;
 };
 
 /*
@@ -1657,6 +2036,7 @@ struct radeon_device {
 	struct radeon_gart		gart;
 	struct radeon_mode_info		mode_info;
 	struct radeon_scratch		scratch;
+	struct radeon_doorbell		doorbell;
 	struct radeon_mman		mman;
 	struct radeon_fence_driver	fence_drv[RADEON_NUM_RINGS];
 	wait_queue_head_t		fence_queue;
@@ -1683,14 +2063,18 @@ struct radeon_device {
 	const struct firmware *rlc_fw;	/* r6/700 RLC firmware */
 	const struct firmware *mc_fw;	/* NI MC firmware */
 	const struct firmware *ce_fw;	/* SI CE firmware */
-	const struct firmware *uvd_fw;	/* UVD firmware */
+	const struct firmware *mec_fw;	/* CIK MEC firmware */
+	const struct firmware *sdma_fw;	/* CIK SDMA firmware */
+	const struct firmware *smc_fw;	/* SMC firmware */
 	struct r600_blit r600_blit;
 	struct r600_vram_scratch vram_scratch;
 	int msi_enabled; /* msi enabled */
 	struct r600_ih ih; /* r6/700 interrupt ring */
-	struct si_rlc rlc;
+	struct radeon_rlc rlc;
+	struct radeon_mec mec;
 	struct work_struct hotplug_work;
 	struct work_struct audio_work;
+	struct work_struct reset_work;
 	int num_crtc; /* number of crtcs */
 	struct mutex dc_hw_i2c_mutex; /* display controller hw i2c mutex */
 	bool audio_enabled;
@@ -1727,6 +2111,9 @@ void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v,
 u32 r100_io_rreg(struct radeon_device *rdev, u32 reg);
 void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 
+u32 cik_mm_rdoorbell(struct radeon_device *rdev, u32 offset);
+void cik_mm_wdoorbell(struct radeon_device *rdev, u32 offset, u32 v);
+
 /*
  * Cast helper
  */
@@ -1754,6 +2141,18 @@ void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 #define WREG32_PCIE(reg, v) rv370_pcie_wreg(rdev, (reg), (v))
 #define RREG32_PCIE_PORT(reg) rdev->pciep_rreg(rdev, (reg))
 #define WREG32_PCIE_PORT(reg, v) rdev->pciep_wreg(rdev, (reg), (v))
+#define RREG32_SMC(reg) tn_smc_rreg(rdev, (reg))
+#define WREG32_SMC(reg, v) tn_smc_wreg(rdev, (reg), (v))
+#define RREG32_RCU(reg) r600_rcu_rreg(rdev, (reg))
+#define WREG32_RCU(reg, v) r600_rcu_wreg(rdev, (reg), (v))
+#define RREG32_CG(reg) eg_cg_rreg(rdev, (reg))
+#define WREG32_CG(reg, v) eg_cg_wreg(rdev, (reg), (v))
+#define RREG32_PIF_PHY0(reg) eg_pif_phy0_rreg(rdev, (reg))
+#define WREG32_PIF_PHY0(reg, v) eg_pif_phy0_wreg(rdev, (reg), (v))
+#define RREG32_PIF_PHY1(reg) eg_pif_phy1_rreg(rdev, (reg))
+#define WREG32_PIF_PHY1(reg, v) eg_pif_phy1_wreg(rdev, (reg), (v))
+#define RREG32_UVD_CTX(reg) r600_uvd_ctx_rreg(rdev, (reg))
+#define WREG32_UVD_CTX(reg, v) r600_uvd_ctx_wreg(rdev, (reg), (v))
 #define WREG32_P(reg, val, mask)				\
 	do {							\
 		uint32_t tmp_ = RREG32(reg);			\
@@ -1774,6 +2173,9 @@ void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 #define RREG32_IO(reg) r100_io_rreg(rdev, (reg))
 #define WREG32_IO(reg, v) r100_io_wreg(rdev, (reg), (v))
 
+#define RDOORBELL32(offset) cik_mm_rdoorbell(rdev, (offset))
+#define WDOORBELL32(offset, v) cik_mm_wdoorbell(rdev, (offset), (v))
+
 /*
  * Indirect registers accessor
  */
@@ -1792,6 +2194,96 @@ static inline void rv370_pcie_wreg(struct radeon_device *rdev, uint32_t reg, uin
 	WREG32(RADEON_PCIE_DATA, (v));
 }
 
+static inline u32 tn_smc_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(TN_SMC_IND_INDEX_0, (reg));
+	r = RREG32(TN_SMC_IND_DATA_0);
+	return r;
+}
+
+static inline void tn_smc_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(TN_SMC_IND_INDEX_0, (reg));
+	WREG32(TN_SMC_IND_DATA_0, (v));
+}
+
+static inline u32 r600_rcu_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(R600_RCU_INDEX, ((reg) & 0x1fff));
+	r = RREG32(R600_RCU_DATA);
+	return r;
+}
+
+static inline void r600_rcu_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(R600_RCU_INDEX, ((reg) & 0x1fff));
+	WREG32(R600_RCU_DATA, (v));
+}
+
+static inline u32 eg_cg_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(EVERGREEN_CG_IND_ADDR, ((reg) & 0xffff));
+	r = RREG32(EVERGREEN_CG_IND_DATA);
+	return r;
+}
+
+static inline void eg_cg_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(EVERGREEN_CG_IND_ADDR, ((reg) & 0xffff));
+	WREG32(EVERGREEN_CG_IND_DATA, (v));
+}
+
+static inline u32 eg_pif_phy0_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(EVERGREEN_PIF_PHY0_INDEX, ((reg) & 0xffff));
+	r = RREG32(EVERGREEN_PIF_PHY0_DATA);
+	return r;
+}
+
+static inline void eg_pif_phy0_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(EVERGREEN_PIF_PHY0_INDEX, ((reg) & 0xffff));
+	WREG32(EVERGREEN_PIF_PHY0_DATA, (v));
+}
+
+static inline u32 eg_pif_phy1_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(EVERGREEN_PIF_PHY1_INDEX, ((reg) & 0xffff));
+	r = RREG32(EVERGREEN_PIF_PHY1_DATA);
+	return r;
+}
+
+static inline void eg_pif_phy1_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(EVERGREEN_PIF_PHY1_INDEX, ((reg) & 0xffff));
+	WREG32(EVERGREEN_PIF_PHY1_DATA, (v));
+}
+
+static inline u32 r600_uvd_ctx_rreg(struct radeon_device *rdev, u32 reg)
+{
+	u32 r;
+
+	WREG32(R600_UVD_CTX_INDEX, ((reg) & 0x1ff));
+	r = RREG32(R600_UVD_CTX_DATA);
+	return r;
+}
+
+static inline void r600_uvd_ctx_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	WREG32(R600_UVD_CTX_INDEX, ((reg) & 0x1ff));
+	WREG32(R600_UVD_CTX_DATA, (v));
+}
+
 void r100_pll_errata_after_index(struct radeon_device *rdev);
 
 
@@ -1840,6 +2332,16 @@ void r100_pll_errata_after_index(struct radeon_device *rdev);
 			     (rdev->flags & RADEON_IS_IGP))
 #define ASIC_IS_DCE64(rdev) ((rdev->family == CHIP_OLAND))
 #define ASIC_IS_NODCE(rdev) ((rdev->family == CHIP_HAINAN))
+#define ASIC_IS_DCE8(rdev) ((rdev->family >= CHIP_BONAIRE))
+
+#define ASIC_IS_LOMBOK(rdev) ((rdev->ddev->pdev->device == 0x6849) || \
+			      (rdev->ddev->pdev->device == 0x6850) || \
+			      (rdev->ddev->pdev->device == 0x6858) || \
+			      (rdev->ddev->pdev->device == 0x6859) || \
+			      (rdev->ddev->pdev->device == 0x6840) || \
+			      (rdev->ddev->pdev->device == 0x6841) || \
+			      (rdev->ddev->pdev->device == 0x6842) || \
+			      (rdev->ddev->pdev->device == 0x6843))
 
 /*
  * BIOS helpers.
@@ -1892,6 +2394,9 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_ring_ib_parse(rdev, r, ib) (rdev)->asic->ring[(r)].ib_parse((rdev), (ib))
 #define radeon_ring_is_lockup(rdev, r, cp) (rdev)->asic->ring[(r)].is_lockup((rdev), (cp))
 #define radeon_ring_vm_flush(rdev, r, vm) (rdev)->asic->ring[(r)].vm_flush((rdev), (r), (vm))
+#define radeon_ring_get_rptr(rdev, r) (rdev)->asic->ring[(r)->idx].get_rptr((rdev), (r))
+#define radeon_ring_get_wptr(rdev, r) (rdev)->asic->ring[(r)->idx].get_wptr((rdev), (r))
+#define radeon_ring_set_wptr(rdev, r) (rdev)->asic->ring[(r)->idx].set_wptr((rdev), (r))
 #define radeon_irq_set(rdev) (rdev)->asic->irq.set((rdev))
 #define radeon_irq_process(rdev) (rdev)->asic->irq.process((rdev))
 #define radeon_get_vblank_counter(rdev, crtc) (rdev)->asic->display.get_vblank_counter((rdev), (crtc))
@@ -1915,6 +2420,7 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_set_pcie_lanes(rdev, l) (rdev)->asic->pm.set_pcie_lanes((rdev), (l))
 #define radeon_set_clock_gating(rdev, e) (rdev)->asic->pm.set_clock_gating((rdev), (e))
 #define radeon_set_uvd_clocks(rdev, v, d) (rdev)->asic->pm.set_uvd_clocks((rdev), (v), (d))
+#define radeon_get_temperature(rdev) (rdev)->asic->pm.get_temperature((rdev))
 #define radeon_set_surface_reg(rdev, r, f, p, o, s) ((rdev)->asic->surface.set_reg((rdev), (r), (f), (p), (o), (s)))
 #define radeon_clear_surface_reg(rdev, r) ((rdev)->asic->surface.clear_reg((rdev), (r)))
 #define radeon_bandwidth_update(rdev) (rdev)->asic->display.bandwidth_update((rdev))
@@ -1935,6 +2441,21 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_mc_wait_for_idle(rdev) (rdev)->asic->mc_wait_for_idle((rdev))
 #define radeon_get_xclk(rdev) (rdev)->asic->get_xclk((rdev))
 #define radeon_get_gpu_clock_counter(rdev) (rdev)->asic->get_gpu_clock_counter((rdev))
+#define radeon_dpm_init(rdev) rdev->asic->dpm.init((rdev))
+#define radeon_dpm_setup_asic(rdev) rdev->asic->dpm.setup_asic((rdev))
+#define radeon_dpm_enable(rdev) rdev->asic->dpm.enable((rdev))
+#define radeon_dpm_disable(rdev) rdev->asic->dpm.disable((rdev))
+#define radeon_dpm_pre_set_power_state(rdev) rdev->asic->dpm.pre_set_power_state((rdev))
+#define radeon_dpm_set_power_state(rdev) rdev->asic->dpm.set_power_state((rdev))
+#define radeon_dpm_post_set_power_state(rdev) rdev->asic->dpm.post_set_power_state((rdev))
+#define radeon_dpm_display_configuration_changed(rdev) rdev->asic->dpm.display_configuration_changed((rdev))
+#define radeon_dpm_fini(rdev) rdev->asic->dpm.fini((rdev))
+#define radeon_dpm_get_sclk(rdev, l) rdev->asic->dpm.get_sclk((rdev), (l))
+#define radeon_dpm_get_mclk(rdev, l) rdev->asic->dpm.get_mclk((rdev), (l))
+#define radeon_dpm_print_power_state(rdev, ps) rdev->asic->dpm.print_power_state((rdev), (ps))
+#define radeon_dpm_debugfs_print_current_performance_level(rdev, m) rdev->asic->dpm.debugfs_print_current_performance_level((rdev), (m))
+#define radeon_dpm_force_performance_level(rdev, l) rdev->asic->dpm.force_performance_level((rdev), (l))
+#define radeon_dpm_vblank_too_short(rdev) rdev->asic->dpm.vblank_too_short((rdev))
 
 /* Common functions */
 /* AGP */
@@ -2054,6 +2575,10 @@ extern int ni_mc_load_microcode(struct radeon_device *rdev);
 #if defined(CONFIG_ACPI)
 extern int radeon_acpi_init(struct radeon_device *rdev);
 extern void radeon_acpi_fini(struct radeon_device *rdev);
+extern bool radeon_acpi_is_pcie_performance_request_supported(struct radeon_device *rdev);
+extern int radeon_acpi_pcie_performance_request(struct radeon_device *rdev,
+						u8 perf_req, bool advertise);
+extern int radeon_acpi_pcie_notify_device_ready(struct radeon_device *rdev);
 #else
 static inline int radeon_acpi_init(struct radeon_device *rdev) { return 0; }
 static inline void radeon_acpi_fini(struct radeon_device *rdev) { }
diff --git a/drivers/gpu/drm/radeon/radeon_acpi.c b/drivers/gpu/drm/radeon/radeon_acpi.c
index 196d28d99570..10f98c7742d8 100644
--- a/drivers/gpu/drm/radeon/radeon_acpi.c
+++ b/drivers/gpu/drm/radeon/radeon_acpi.c
@@ -78,6 +78,22 @@ struct atcs_verify_interface {
 	u32 function_bits;	/* supported functions bit vector */
 } __packed;
 
+#define ATCS_VALID_FLAGS_MASK	0x3
+
+struct atcs_pref_req_input {
+	u16 size;		/* structure size in bytes (includes size field) */
+	u16 client_id;		/* client id (bit 2-0: func num, 7-3: dev num, 15-8: bus num) */
+	u16 valid_flags_mask;	/* valid flags mask */
+	u16 flags;		/* flags */
+	u8 req_type;		/* request type */
+	u8 perf_req;		/* performance request */
+} __packed;
+
+struct atcs_pref_req_output {
+	u16 size;		/* structure size in bytes (includes size field) */
+	u8 ret_val;		/* return value */
+} __packed;
+
 /* Call the ATIF method
  */
 /**
@@ -506,6 +522,135 @@ out:
 }
 
 /**
+ * radeon_acpi_is_pcie_performance_request_supported
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Check if the ATCS pcie_perf_req and pcie_dev_rdy methods
+ * are supported (all asics).
+ * returns true if supported, false if not.
+ */
+bool radeon_acpi_is_pcie_performance_request_supported(struct radeon_device *rdev)
+{
+	struct radeon_atcs *atcs = &rdev->atcs;
+
+	if (atcs->functions.pcie_perf_req && atcs->functions.pcie_dev_rdy)
+		return true;
+
+	return false;
+}
+
+/**
+ * radeon_acpi_pcie_notify_device_ready
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Executes the PCIE_DEVICE_READY_NOTIFICATION method
+ * (all asics).
+ * returns 0 on success, error on failure.
+ */
+int radeon_acpi_pcie_notify_device_ready(struct radeon_device *rdev)
+{
+	acpi_handle handle;
+	union acpi_object *info;
+	struct radeon_atcs *atcs = &rdev->atcs;
+
+	/* Get the device handle */
+	handle = DEVICE_ACPI_HANDLE(&rdev->pdev->dev);
+	if (!handle)
+		return -EINVAL;
+
+	if (!atcs->functions.pcie_dev_rdy)
+		return -EINVAL;
+
+	info = radeon_atcs_call(handle, ATCS_FUNCTION_PCIE_DEVICE_READY_NOTIFICATION, NULL);
+	if (!info)
+		return -EIO;
+
+	kfree(info);
+
+	return 0;
+}
+
+/**
+ * radeon_acpi_pcie_performance_request
+ *
+ * @rdev: radeon_device pointer
+ * @perf_req: requested perf level (pcie gen speed)
+ * @advertise: set advertise caps flag if set
+ *
+ * Executes the PCIE_PERFORMANCE_REQUEST method to
+ * change the pcie gen speed (all asics).
+ * returns 0 on success, error on failure.
+ */
+int radeon_acpi_pcie_performance_request(struct radeon_device *rdev,
+					 u8 perf_req, bool advertise)
+{
+	acpi_handle handle;
+	union acpi_object *info;
+	struct radeon_atcs *atcs = &rdev->atcs;
+	struct atcs_pref_req_input atcs_input;
+	struct atcs_pref_req_output atcs_output;
+	struct acpi_buffer params;
+	size_t size;
+	u32 retry = 3;
+
+	/* Get the device handle */
+	handle = DEVICE_ACPI_HANDLE(&rdev->pdev->dev);
+	if (!handle)
+		return -EINVAL;
+
+	if (!atcs->functions.pcie_perf_req)
+		return -EINVAL;
+
+	atcs_input.size = sizeof(struct atcs_pref_req_input);
+	/* client id (bit 2-0: func num, 7-3: dev num, 15-8: bus num) */
+	atcs_input.client_id = rdev->pdev->devfn | (rdev->pdev->bus->number << 8);
+	atcs_input.valid_flags_mask = ATCS_VALID_FLAGS_MASK;
+	atcs_input.flags = ATCS_WAIT_FOR_COMPLETION;
+	if (advertise)
+		atcs_input.flags |= ATCS_ADVERTISE_CAPS;
+	atcs_input.req_type = ATCS_PCIE_LINK_SPEED;
+	atcs_input.perf_req = perf_req;
+
+	params.length = sizeof(struct atcs_pref_req_input);
+	params.pointer = &atcs_input;
+
+	while (retry--) {
+		info = radeon_atcs_call(handle, ATCS_FUNCTION_PCIE_PERFORMANCE_REQUEST, &params);
+		if (!info)
+			return -EIO;
+
+		memset(&atcs_output, 0, sizeof(atcs_output));
+
+		size = *(u16 *) info->buffer.pointer;
+		if (size < 3) {
+			DRM_INFO("ATCS buffer is too small: %zu\n", size);
+			kfree(info);
+			return -EINVAL;
+		}
+		size = min(sizeof(atcs_output), size);
+
+		memcpy(&atcs_output, info->buffer.pointer, size);
+
+		kfree(info);
+
+		switch (atcs_output.ret_val) {
+		case ATCS_REQUEST_REFUSED:
+		default:
+			return -EINVAL;
+		case ATCS_REQUEST_COMPLETE:
+			return 0;
+		case ATCS_REQUEST_IN_PROGRESS:
+			udelay(10);
+			break;
+		}
+	}
+
+	return 0;
+}
+
+/**
  * radeon_acpi_event - handle notify events
  *
  * @nb: notifier block
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index a2802b47ee95..78bec1a58ed1 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -126,7 +126,11 @@ static void radeon_register_accessor_init(struct radeon_device *rdev)
 		rdev->mc_rreg = &rs780_mc_rreg;
 		rdev->mc_wreg = &rs780_mc_wreg;
 	}
-	if (rdev->family >= CHIP_R600) {
+
+	if (rdev->family >= CHIP_BONAIRE) {
+		rdev->pciep_rreg = &cik_pciep_rreg;
+		rdev->pciep_wreg = &cik_pciep_wreg;
+	} else if (rdev->family >= CHIP_R600) {
 		rdev->pciep_rreg = &r600_pciep_rreg;
 		rdev->pciep_wreg = &r600_pciep_wreg;
 	}
@@ -192,6 +196,9 @@ static struct radeon_asic r100_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -268,6 +275,9 @@ static struct radeon_asic r200_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -344,6 +354,9 @@ static struct radeon_asic r300_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -420,6 +433,9 @@ static struct radeon_asic r300_asic_pcie = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -496,6 +512,9 @@ static struct radeon_asic r420_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -572,6 +591,9 @@ static struct radeon_asic rs400_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -648,6 +670,9 @@ static struct radeon_asic rs600_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -726,6 +751,9 @@ static struct radeon_asic rs690_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -804,6 +832,9 @@ static struct radeon_asic rv515_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -880,6 +911,9 @@ static struct radeon_asic r520_asic = {
 			.ring_test = &r100_ring_test,
 			.ib_test = &r100_ib_test,
 			.is_lockup = &r100_gpu_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -957,6 +991,9 @@ static struct radeon_asic r600_asic = {
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
 			.is_lockup = &r600_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &r600_dma_ring_ib_execute,
@@ -966,6 +1003,9 @@ static struct radeon_asic r600_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &r600_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -986,8 +1026,8 @@ static struct radeon_asic r600_asic = {
 		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
 		.dma = &r600_copy_dma,
 		.dma_ring_index = R600_RING_TYPE_DMA_INDEX,
-		.copy = &r600_copy_dma,
-		.copy_ring_index = R600_RING_TYPE_DMA_INDEX,
+		.copy = &r600_copy_cpdma,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
 	},
 	.surface = {
 		.set_reg = r600_set_surface_reg,
@@ -1012,6 +1052,115 @@ static struct radeon_asic r600_asic = {
 		.get_pcie_lanes = &r600_get_pcie_lanes,
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
+		.get_temperature = &rv6xx_get_temp,
+	},
+	.pflip = {
+		.pre_page_flip = &rs600_pre_page_flip,
+		.page_flip = &rs600_page_flip,
+		.post_page_flip = &rs600_post_page_flip,
+	},
+};
+
+static struct radeon_asic rv6xx_asic = {
+	.init = &r600_init,
+	.fini = &r600_fini,
+	.suspend = &r600_suspend,
+	.resume = &r600_resume,
+	.vga_set_state = &r600_vga_set_state,
+	.asic_reset = &r600_asic_reset,
+	.ioctl_wait_idle = r600_ioctl_wait_idle,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &r600_mc_wait_for_idle,
+	.get_xclk = &r600_get_xclk,
+	.get_gpu_clock_counter = &r600_get_gpu_clock_counter,
+	.gart = {
+		.tlb_flush = &r600_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
+	.ring = {
+		[RADEON_RING_TYPE_GFX_INDEX] = {
+			.ib_execute = &r600_ring_ib_execute,
+			.emit_fence = &r600_fence_ring_emit,
+			.emit_semaphore = &r600_semaphore_ring_emit,
+			.cs_parse = &r600_cs_parse,
+			.ring_test = &r600_ring_test,
+			.ib_test = &r600_ib_test,
+			.is_lockup = &r600_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[R600_RING_TYPE_DMA_INDEX] = {
+			.ib_execute = &r600_dma_ring_ib_execute,
+			.emit_fence = &r600_dma_fence_ring_emit,
+			.emit_semaphore = &r600_dma_semaphore_ring_emit,
+			.cs_parse = &r600_dma_cs_parse,
+			.ring_test = &r600_dma_ring_test,
+			.ib_test = &r600_dma_ib_test,
+			.is_lockup = &r600_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		}
+	},
+	.irq = {
+		.set = &r600_irq_set,
+		.process = &r600_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &rv515_bandwidth_update,
+		.get_vblank_counter = &rs600_get_vblank_counter,
+		.wait_for_vblank = &avivo_wait_for_vblank,
+		.set_backlight_level = &atombios_set_backlight_level,
+		.get_backlight_level = &atombios_get_backlight_level,
+	},
+	.copy = {
+		.blit = &r600_copy_blit,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &r600_copy_dma,
+		.dma_ring_index = R600_RING_TYPE_DMA_INDEX,
+		.copy = &r600_copy_cpdma,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &r600_hpd_init,
+		.fini = &r600_hpd_fini,
+		.sense = &r600_hpd_sense,
+		.set_polarity = &r600_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &r600_pm_misc,
+		.prepare = &rs600_pm_prepare,
+		.finish = &rs600_pm_finish,
+		.init_profile = &r600_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.set_engine_clock = &radeon_atom_set_engine_clock,
+		.get_memory_clock = &radeon_atom_get_memory_clock,
+		.set_memory_clock = &radeon_atom_set_memory_clock,
+		.get_pcie_lanes = &r600_get_pcie_lanes,
+		.set_pcie_lanes = &r600_set_pcie_lanes,
+		.set_clock_gating = NULL,
+		.get_temperature = &rv6xx_get_temp,
+	},
+	.dpm = {
+		.init = &rv6xx_dpm_init,
+		.setup_asic = &rv6xx_setup_asic,
+		.enable = &rv6xx_dpm_enable,
+		.disable = &rv6xx_dpm_disable,
+		.pre_set_power_state = &r600_dpm_pre_set_power_state,
+		.set_power_state = &rv6xx_dpm_set_power_state,
+		.post_set_power_state = &r600_dpm_post_set_power_state,
+		.display_configuration_changed = &rv6xx_dpm_display_configuration_changed,
+		.fini = &rv6xx_dpm_fini,
+		.get_sclk = &rv6xx_dpm_get_sclk,
+		.get_mclk = &rv6xx_dpm_get_mclk,
+		.print_power_state = &rv6xx_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &rv6xx_dpm_debugfs_print_current_performance_level,
 	},
 	.pflip = {
 		.pre_page_flip = &rs600_pre_page_flip,
@@ -1045,6 +1194,9 @@ static struct radeon_asic rs780_asic = {
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
 			.is_lockup = &r600_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &r600_dma_ring_ib_execute,
@@ -1054,6 +1206,9 @@ static struct radeon_asic rs780_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &r600_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1074,8 +1229,8 @@ static struct radeon_asic rs780_asic = {
 		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
 		.dma = &r600_copy_dma,
 		.dma_ring_index = R600_RING_TYPE_DMA_INDEX,
-		.copy = &r600_copy_dma,
-		.copy_ring_index = R600_RING_TYPE_DMA_INDEX,
+		.copy = &r600_copy_cpdma,
+		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
 	},
 	.surface = {
 		.set_reg = r600_set_surface_reg,
@@ -1100,6 +1255,22 @@ static struct radeon_asic rs780_asic = {
 		.get_pcie_lanes = NULL,
 		.set_pcie_lanes = NULL,
 		.set_clock_gating = NULL,
+		.get_temperature = &rv6xx_get_temp,
+	},
+	.dpm = {
+		.init = &rs780_dpm_init,
+		.setup_asic = &rs780_dpm_setup_asic,
+		.enable = &rs780_dpm_enable,
+		.disable = &rs780_dpm_disable,
+		.pre_set_power_state = &r600_dpm_pre_set_power_state,
+		.set_power_state = &rs780_dpm_set_power_state,
+		.post_set_power_state = &r600_dpm_post_set_power_state,
+		.display_configuration_changed = &rs780_dpm_display_configuration_changed,
+		.fini = &rs780_dpm_fini,
+		.get_sclk = &rs780_dpm_get_sclk,
+		.get_mclk = &rs780_dpm_get_mclk,
+		.print_power_state = &rs780_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &rs780_dpm_debugfs_print_current_performance_level,
 	},
 	.pflip = {
 		.pre_page_flip = &rs600_pre_page_flip,
@@ -1133,6 +1304,9 @@ static struct radeon_asic rv770_asic = {
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
 			.is_lockup = &r600_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &r600_dma_ring_ib_execute,
@@ -1142,6 +1316,9 @@ static struct radeon_asic rv770_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &r600_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1151,6 +1328,9 @@ static struct radeon_asic rv770_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1198,6 +1378,24 @@ static struct radeon_asic rv770_asic = {
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = &radeon_atom_set_clock_gating,
 		.set_uvd_clocks = &rv770_set_uvd_clocks,
+		.get_temperature = &rv770_get_temp,
+	},
+	.dpm = {
+		.init = &rv770_dpm_init,
+		.setup_asic = &rv770_dpm_setup_asic,
+		.enable = &rv770_dpm_enable,
+		.disable = &rv770_dpm_disable,
+		.pre_set_power_state = &r600_dpm_pre_set_power_state,
+		.set_power_state = &rv770_dpm_set_power_state,
+		.post_set_power_state = &r600_dpm_post_set_power_state,
+		.display_configuration_changed = &rv770_dpm_display_configuration_changed,
+		.fini = &rv770_dpm_fini,
+		.get_sclk = &rv770_dpm_get_sclk,
+		.get_mclk = &rv770_dpm_get_mclk,
+		.print_power_state = &rv770_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &rv770_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &rv770_dpm_force_performance_level,
+		.vblank_too_short = &rv770_dpm_vblank_too_short,
 	},
 	.pflip = {
 		.pre_page_flip = &rs600_pre_page_flip,
@@ -1231,6 +1429,9 @@ static struct radeon_asic evergreen_asic = {
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
 			.is_lockup = &evergreen_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &evergreen_dma_ring_ib_execute,
@@ -1240,6 +1441,9 @@ static struct radeon_asic evergreen_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &evergreen_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1249,6 +1453,9 @@ static struct radeon_asic evergreen_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1296,6 +1503,24 @@ static struct radeon_asic evergreen_asic = {
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
 		.set_uvd_clocks = &evergreen_set_uvd_clocks,
+		.get_temperature = &evergreen_get_temp,
+	},
+	.dpm = {
+		.init = &cypress_dpm_init,
+		.setup_asic = &cypress_dpm_setup_asic,
+		.enable = &cypress_dpm_enable,
+		.disable = &cypress_dpm_disable,
+		.pre_set_power_state = &r600_dpm_pre_set_power_state,
+		.set_power_state = &cypress_dpm_set_power_state,
+		.post_set_power_state = &r600_dpm_post_set_power_state,
+		.display_configuration_changed = &cypress_dpm_display_configuration_changed,
+		.fini = &cypress_dpm_fini,
+		.get_sclk = &rv770_dpm_get_sclk,
+		.get_mclk = &rv770_dpm_get_mclk,
+		.print_power_state = &rv770_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &rv770_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &rv770_dpm_force_performance_level,
+		.vblank_too_short = &cypress_dpm_vblank_too_short,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1329,6 +1554,9 @@ static struct radeon_asic sumo_asic = {
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
 			.is_lockup = &evergreen_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &evergreen_dma_ring_ib_execute,
@@ -1338,6 +1566,9 @@ static struct radeon_asic sumo_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &evergreen_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1347,6 +1578,9 @@ static struct radeon_asic sumo_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1394,6 +1628,23 @@ static struct radeon_asic sumo_asic = {
 		.set_pcie_lanes = NULL,
 		.set_clock_gating = NULL,
 		.set_uvd_clocks = &sumo_set_uvd_clocks,
+		.get_temperature = &sumo_get_temp,
+	},
+	.dpm = {
+		.init = &sumo_dpm_init,
+		.setup_asic = &sumo_dpm_setup_asic,
+		.enable = &sumo_dpm_enable,
+		.disable = &sumo_dpm_disable,
+		.pre_set_power_state = &sumo_dpm_pre_set_power_state,
+		.set_power_state = &sumo_dpm_set_power_state,
+		.post_set_power_state = &sumo_dpm_post_set_power_state,
+		.display_configuration_changed = &sumo_dpm_display_configuration_changed,
+		.fini = &sumo_dpm_fini,
+		.get_sclk = &sumo_dpm_get_sclk,
+		.get_mclk = &sumo_dpm_get_mclk,
+		.print_power_state = &sumo_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &sumo_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &sumo_dpm_force_performance_level,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1427,6 +1678,9 @@ static struct radeon_asic btc_asic = {
 			.ring_test = &r600_ring_test,
 			.ib_test = &r600_ib_test,
 			.is_lockup = &evergreen_gfx_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &evergreen_dma_ring_ib_execute,
@@ -1436,6 +1690,9 @@ static struct radeon_asic btc_asic = {
 			.ring_test = &r600_dma_ring_test,
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &evergreen_dma_is_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1445,6 +1702,9 @@ static struct radeon_asic btc_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1492,6 +1752,24 @@ static struct radeon_asic btc_asic = {
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
 		.set_uvd_clocks = &evergreen_set_uvd_clocks,
+		.get_temperature = &evergreen_get_temp,
+	},
+	.dpm = {
+		.init = &btc_dpm_init,
+		.setup_asic = &btc_dpm_setup_asic,
+		.enable = &btc_dpm_enable,
+		.disable = &btc_dpm_disable,
+		.pre_set_power_state = &btc_dpm_pre_set_power_state,
+		.set_power_state = &btc_dpm_set_power_state,
+		.post_set_power_state = &btc_dpm_post_set_power_state,
+		.display_configuration_changed = &cypress_dpm_display_configuration_changed,
+		.fini = &btc_dpm_fini,
+		.get_sclk = &btc_dpm_get_sclk,
+		.get_mclk = &btc_dpm_get_mclk,
+		.print_power_state = &rv770_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &rv770_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &rv770_dpm_force_performance_level,
+		.vblank_too_short = &btc_dpm_vblank_too_short,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1533,6 +1811,9 @@ static struct radeon_asic cayman_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_CP1_INDEX] = {
 			.ib_execute = &cayman_ring_ib_execute,
@@ -1544,6 +1825,9 @@ static struct radeon_asic cayman_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_CP2_INDEX] = {
 			.ib_execute = &cayman_ring_ib_execute,
@@ -1555,6 +1839,9 @@ static struct radeon_asic cayman_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &cayman_dma_ring_ib_execute,
@@ -1566,6 +1853,9 @@ static struct radeon_asic cayman_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &cayman_dma_is_lockup,
 			.vm_flush = &cayman_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_DMA1_INDEX] = {
 			.ib_execute = &cayman_dma_ring_ib_execute,
@@ -1577,6 +1867,9 @@ static struct radeon_asic cayman_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &cayman_dma_is_lockup,
 			.vm_flush = &cayman_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1586,6 +1879,9 @@ static struct radeon_asic cayman_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1633,6 +1929,24 @@ static struct radeon_asic cayman_asic = {
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
 		.set_uvd_clocks = &evergreen_set_uvd_clocks,
+		.get_temperature = &evergreen_get_temp,
+	},
+	.dpm = {
+		.init = &ni_dpm_init,
+		.setup_asic = &ni_dpm_setup_asic,
+		.enable = &ni_dpm_enable,
+		.disable = &ni_dpm_disable,
+		.pre_set_power_state = &ni_dpm_pre_set_power_state,
+		.set_power_state = &ni_dpm_set_power_state,
+		.post_set_power_state = &ni_dpm_post_set_power_state,
+		.display_configuration_changed = &cypress_dpm_display_configuration_changed,
+		.fini = &ni_dpm_fini,
+		.get_sclk = &ni_dpm_get_sclk,
+		.get_mclk = &ni_dpm_get_mclk,
+		.print_power_state = &ni_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &ni_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &ni_dpm_force_performance_level,
+		.vblank_too_short = &ni_dpm_vblank_too_short,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1674,6 +1988,9 @@ static struct radeon_asic trinity_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_CP1_INDEX] = {
 			.ib_execute = &cayman_ring_ib_execute,
@@ -1685,6 +2002,9 @@ static struct radeon_asic trinity_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_CP2_INDEX] = {
 			.ib_execute = &cayman_ring_ib_execute,
@@ -1696,6 +2016,9 @@ static struct radeon_asic trinity_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &cayman_gfx_is_lockup,
 			.vm_flush = &cayman_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &cayman_dma_ring_ib_execute,
@@ -1707,6 +2030,9 @@ static struct radeon_asic trinity_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &cayman_dma_is_lockup,
 			.vm_flush = &cayman_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_DMA1_INDEX] = {
 			.ib_execute = &cayman_dma_ring_ib_execute,
@@ -1718,6 +2044,9 @@ static struct radeon_asic trinity_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &cayman_dma_is_lockup,
 			.vm_flush = &cayman_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1727,6 +2056,9 @@ static struct radeon_asic trinity_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1772,6 +2104,23 @@ static struct radeon_asic trinity_asic = {
 		.set_pcie_lanes = NULL,
 		.set_clock_gating = NULL,
 		.set_uvd_clocks = &sumo_set_uvd_clocks,
+		.get_temperature = &tn_get_temp,
+	},
+	.dpm = {
+		.init = &trinity_dpm_init,
+		.setup_asic = &trinity_dpm_setup_asic,
+		.enable = &trinity_dpm_enable,
+		.disable = &trinity_dpm_disable,
+		.pre_set_power_state = &trinity_dpm_pre_set_power_state,
+		.set_power_state = &trinity_dpm_set_power_state,
+		.post_set_power_state = &trinity_dpm_post_set_power_state,
+		.display_configuration_changed = &trinity_dpm_display_configuration_changed,
+		.fini = &trinity_dpm_fini,
+		.get_sclk = &trinity_dpm_get_sclk,
+		.get_mclk = &trinity_dpm_get_mclk,
+		.print_power_state = &trinity_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &trinity_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &trinity_dpm_force_performance_level,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1813,6 +2162,9 @@ static struct radeon_asic si_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &si_gfx_is_lockup,
 			.vm_flush = &si_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_CP1_INDEX] = {
 			.ib_execute = &si_ring_ib_execute,
@@ -1824,6 +2176,9 @@ static struct radeon_asic si_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &si_gfx_is_lockup,
 			.vm_flush = &si_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_CP2_INDEX] = {
 			.ib_execute = &si_ring_ib_execute,
@@ -1835,6 +2190,9 @@ static struct radeon_asic si_asic = {
 			.ib_test = &r600_ib_test,
 			.is_lockup = &si_gfx_is_lockup,
 			.vm_flush = &si_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_DMA_INDEX] = {
 			.ib_execute = &cayman_dma_ring_ib_execute,
@@ -1846,6 +2204,9 @@ static struct radeon_asic si_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &si_dma_is_lockup,
 			.vm_flush = &si_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[CAYMAN_RING_TYPE_DMA1_INDEX] = {
 			.ib_execute = &cayman_dma_ring_ib_execute,
@@ -1857,6 +2218,9 @@ static struct radeon_asic si_asic = {
 			.ib_test = &r600_dma_ib_test,
 			.is_lockup = &si_dma_is_lockup,
 			.vm_flush = &si_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		},
 		[R600_RING_TYPE_UVD_INDEX] = {
 			.ib_execute = &r600_uvd_ib_execute,
@@ -1866,6 +2230,9 @@ static struct radeon_asic si_asic = {
 			.ring_test = &r600_uvd_ring_test,
 			.ib_test = &r600_uvd_ib_test,
 			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
 		}
 	},
 	.irq = {
@@ -1911,6 +2278,334 @@ static struct radeon_asic si_asic = {
 		.set_pcie_lanes = &r600_set_pcie_lanes,
 		.set_clock_gating = NULL,
 		.set_uvd_clocks = &si_set_uvd_clocks,
+		.get_temperature = &si_get_temp,
+	},
+	.dpm = {
+		.init = &si_dpm_init,
+		.setup_asic = &si_dpm_setup_asic,
+		.enable = &si_dpm_enable,
+		.disable = &si_dpm_disable,
+		.pre_set_power_state = &si_dpm_pre_set_power_state,
+		.set_power_state = &si_dpm_set_power_state,
+		.post_set_power_state = &si_dpm_post_set_power_state,
+		.display_configuration_changed = &si_dpm_display_configuration_changed,
+		.fini = &si_dpm_fini,
+		.get_sclk = &ni_dpm_get_sclk,
+		.get_mclk = &ni_dpm_get_mclk,
+		.print_power_state = &ni_dpm_print_power_state,
+		.debugfs_print_current_performance_level = &si_dpm_debugfs_print_current_performance_level,
+		.force_performance_level = &si_dpm_force_performance_level,
+		.vblank_too_short = &ni_dpm_vblank_too_short,
+	},
+	.pflip = {
+		.pre_page_flip = &evergreen_pre_page_flip,
+		.page_flip = &evergreen_page_flip,
+		.post_page_flip = &evergreen_post_page_flip,
+	},
+};
+
+static struct radeon_asic ci_asic = {
+	.init = &cik_init,
+	.fini = &cik_fini,
+	.suspend = &cik_suspend,
+	.resume = &cik_resume,
+	.asic_reset = &cik_asic_reset,
+	.vga_set_state = &r600_vga_set_state,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.get_xclk = &cik_get_xclk,
+	.get_gpu_clock_counter = &cik_get_gpu_clock_counter,
+	.gart = {
+		.tlb_flush = &cik_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
+	.vm = {
+		.init = &cik_vm_init,
+		.fini = &cik_vm_fini,
+		.pt_ring_index = R600_RING_TYPE_DMA_INDEX,
+		.set_page = &cik_vm_set_page,
+	},
+	.ring = {
+		[RADEON_RING_TYPE_GFX_INDEX] = {
+			.ib_execute = &cik_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_fence_gfx_ring_emit,
+			.emit_semaphore = &cik_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_ring_test,
+			.ib_test = &cik_ib_test,
+			.is_lockup = &cik_gfx_is_lockup,
+			.vm_flush = &cik_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[CAYMAN_RING_TYPE_CP1_INDEX] = {
+			.ib_execute = &cik_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_fence_compute_ring_emit,
+			.emit_semaphore = &cik_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_ring_test,
+			.ib_test = &cik_ib_test,
+			.is_lockup = &cik_gfx_is_lockup,
+			.vm_flush = &cik_vm_flush,
+			.get_rptr = &cik_compute_ring_get_rptr,
+			.get_wptr = &cik_compute_ring_get_wptr,
+			.set_wptr = &cik_compute_ring_set_wptr,
+		},
+		[CAYMAN_RING_TYPE_CP2_INDEX] = {
+			.ib_execute = &cik_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_fence_compute_ring_emit,
+			.emit_semaphore = &cik_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_ring_test,
+			.ib_test = &cik_ib_test,
+			.is_lockup = &cik_gfx_is_lockup,
+			.vm_flush = &cik_vm_flush,
+			.get_rptr = &cik_compute_ring_get_rptr,
+			.get_wptr = &cik_compute_ring_get_wptr,
+			.set_wptr = &cik_compute_ring_set_wptr,
+		},
+		[R600_RING_TYPE_DMA_INDEX] = {
+			.ib_execute = &cik_sdma_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_sdma_fence_ring_emit,
+			.emit_semaphore = &cik_sdma_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_sdma_ring_test,
+			.ib_test = &cik_sdma_ib_test,
+			.is_lockup = &cik_sdma_is_lockup,
+			.vm_flush = &cik_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[CAYMAN_RING_TYPE_DMA1_INDEX] = {
+			.ib_execute = &cik_sdma_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_sdma_fence_ring_emit,
+			.emit_semaphore = &cik_sdma_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_sdma_ring_test,
+			.ib_test = &cik_sdma_ib_test,
+			.is_lockup = &cik_sdma_is_lockup,
+			.vm_flush = &cik_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[R600_RING_TYPE_UVD_INDEX] = {
+			.ib_execute = &r600_uvd_ib_execute,
+			.emit_fence = &r600_uvd_fence_emit,
+			.emit_semaphore = &cayman_uvd_semaphore_emit,
+			.cs_parse = &radeon_uvd_cs_parse,
+			.ring_test = &r600_uvd_ring_test,
+			.ib_test = &r600_uvd_ib_test,
+			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		}
+	},
+	.irq = {
+		.set = &cik_irq_set,
+		.process = &cik_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &dce8_bandwidth_update,
+		.get_vblank_counter = &evergreen_get_vblank_counter,
+		.wait_for_vblank = &dce4_wait_for_vblank,
+	},
+	.copy = {
+		.blit = NULL,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &cik_copy_dma,
+		.dma_ring_index = R600_RING_TYPE_DMA_INDEX,
+		.copy = &cik_copy_dma,
+		.copy_ring_index = R600_RING_TYPE_DMA_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &evergreen_hpd_init,
+		.fini = &evergreen_hpd_fini,
+		.sense = &evergreen_hpd_sense,
+		.set_polarity = &evergreen_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &evergreen_pm_misc,
+		.prepare = &evergreen_pm_prepare,
+		.finish = &evergreen_pm_finish,
+		.init_profile = &sumo_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.set_engine_clock = &radeon_atom_set_engine_clock,
+		.get_memory_clock = &radeon_atom_get_memory_clock,
+		.set_memory_clock = &radeon_atom_set_memory_clock,
+		.get_pcie_lanes = NULL,
+		.set_pcie_lanes = NULL,
+		.set_clock_gating = NULL,
+		.set_uvd_clocks = &cik_set_uvd_clocks,
+	},
+	.pflip = {
+		.pre_page_flip = &evergreen_pre_page_flip,
+		.page_flip = &evergreen_page_flip,
+		.post_page_flip = &evergreen_post_page_flip,
+	},
+};
+
+static struct radeon_asic kv_asic = {
+	.init = &cik_init,
+	.fini = &cik_fini,
+	.suspend = &cik_suspend,
+	.resume = &cik_resume,
+	.asic_reset = &cik_asic_reset,
+	.vga_set_state = &r600_vga_set_state,
+	.ioctl_wait_idle = NULL,
+	.gui_idle = &r600_gui_idle,
+	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
+	.get_xclk = &cik_get_xclk,
+	.get_gpu_clock_counter = &cik_get_gpu_clock_counter,
+	.gart = {
+		.tlb_flush = &cik_pcie_gart_tlb_flush,
+		.set_page = &rs600_gart_set_page,
+	},
+	.vm = {
+		.init = &cik_vm_init,
+		.fini = &cik_vm_fini,
+		.pt_ring_index = R600_RING_TYPE_DMA_INDEX,
+		.set_page = &cik_vm_set_page,
+	},
+	.ring = {
+		[RADEON_RING_TYPE_GFX_INDEX] = {
+			.ib_execute = &cik_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_fence_gfx_ring_emit,
+			.emit_semaphore = &cik_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_ring_test,
+			.ib_test = &cik_ib_test,
+			.is_lockup = &cik_gfx_is_lockup,
+			.vm_flush = &cik_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[CAYMAN_RING_TYPE_CP1_INDEX] = {
+			.ib_execute = &cik_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_fence_compute_ring_emit,
+			.emit_semaphore = &cik_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_ring_test,
+			.ib_test = &cik_ib_test,
+			.is_lockup = &cik_gfx_is_lockup,
+			.vm_flush = &cik_vm_flush,
+			.get_rptr = &cik_compute_ring_get_rptr,
+			.get_wptr = &cik_compute_ring_get_wptr,
+			.set_wptr = &cik_compute_ring_set_wptr,
+		},
+		[CAYMAN_RING_TYPE_CP2_INDEX] = {
+			.ib_execute = &cik_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_fence_compute_ring_emit,
+			.emit_semaphore = &cik_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_ring_test,
+			.ib_test = &cik_ib_test,
+			.is_lockup = &cik_gfx_is_lockup,
+			.vm_flush = &cik_vm_flush,
+			.get_rptr = &cik_compute_ring_get_rptr,
+			.get_wptr = &cik_compute_ring_get_wptr,
+			.set_wptr = &cik_compute_ring_set_wptr,
+		},
+		[R600_RING_TYPE_DMA_INDEX] = {
+			.ib_execute = &cik_sdma_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_sdma_fence_ring_emit,
+			.emit_semaphore = &cik_sdma_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_sdma_ring_test,
+			.ib_test = &cik_sdma_ib_test,
+			.is_lockup = &cik_sdma_is_lockup,
+			.vm_flush = &cik_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[CAYMAN_RING_TYPE_DMA1_INDEX] = {
+			.ib_execute = &cik_sdma_ring_ib_execute,
+			.ib_parse = &cik_ib_parse,
+			.emit_fence = &cik_sdma_fence_ring_emit,
+			.emit_semaphore = &cik_sdma_semaphore_ring_emit,
+			.cs_parse = NULL,
+			.ring_test = &cik_sdma_ring_test,
+			.ib_test = &cik_sdma_ib_test,
+			.is_lockup = &cik_sdma_is_lockup,
+			.vm_flush = &cik_dma_vm_flush,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		},
+		[R600_RING_TYPE_UVD_INDEX] = {
+			.ib_execute = &r600_uvd_ib_execute,
+			.emit_fence = &r600_uvd_fence_emit,
+			.emit_semaphore = &cayman_uvd_semaphore_emit,
+			.cs_parse = &radeon_uvd_cs_parse,
+			.ring_test = &r600_uvd_ring_test,
+			.ib_test = &r600_uvd_ib_test,
+			.is_lockup = &radeon_ring_test_lockup,
+			.get_rptr = &radeon_ring_generic_get_rptr,
+			.get_wptr = &radeon_ring_generic_get_wptr,
+			.set_wptr = &radeon_ring_generic_set_wptr,
+		}
+	},
+	.irq = {
+		.set = &cik_irq_set,
+		.process = &cik_irq_process,
+	},
+	.display = {
+		.bandwidth_update = &dce8_bandwidth_update,
+		.get_vblank_counter = &evergreen_get_vblank_counter,
+		.wait_for_vblank = &dce4_wait_for_vblank,
+	},
+	.copy = {
+		.blit = NULL,
+		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
+		.dma = &cik_copy_dma,
+		.dma_ring_index = R600_RING_TYPE_DMA_INDEX,
+		.copy = &cik_copy_dma,
+		.copy_ring_index = R600_RING_TYPE_DMA_INDEX,
+	},
+	.surface = {
+		.set_reg = r600_set_surface_reg,
+		.clear_reg = r600_clear_surface_reg,
+	},
+	.hpd = {
+		.init = &evergreen_hpd_init,
+		.fini = &evergreen_hpd_fini,
+		.sense = &evergreen_hpd_sense,
+		.set_polarity = &evergreen_hpd_set_polarity,
+	},
+	.pm = {
+		.misc = &evergreen_pm_misc,
+		.prepare = &evergreen_pm_prepare,
+		.finish = &evergreen_pm_finish,
+		.init_profile = &sumo_pm_init_profile,
+		.get_dynpm_state = &r600_pm_get_dynpm_state,
+		.get_engine_clock = &radeon_atom_get_engine_clock,
+		.set_engine_clock = &radeon_atom_set_engine_clock,
+		.get_memory_clock = &radeon_atom_get_memory_clock,
+		.set_memory_clock = &radeon_atom_set_memory_clock,
+		.get_pcie_lanes = NULL,
+		.set_pcie_lanes = NULL,
+		.set_clock_gating = NULL,
+		.set_uvd_clocks = &cik_set_uvd_clocks,
 	},
 	.pflip = {
 		.pre_page_flip = &evergreen_pre_page_flip,
@@ -1999,16 +2694,15 @@ int radeon_asic_init(struct radeon_device *rdev)
 		rdev->asic = &r520_asic;
 		break;
 	case CHIP_R600:
+		rdev->asic = &r600_asic;
+		break;
 	case CHIP_RV610:
 	case CHIP_RV630:
 	case CHIP_RV620:
 	case CHIP_RV635:
 	case CHIP_RV670:
-		rdev->asic = &r600_asic;
-		if (rdev->family == CHIP_R600)
-			rdev->has_uvd = false;
-		else
-			rdev->has_uvd = true;
+		rdev->asic = &rv6xx_asic;
+		rdev->has_uvd = true;
 		break;
 	case CHIP_RS780:
 	case CHIP_RS880:
@@ -2082,6 +2776,19 @@ int radeon_asic_init(struct radeon_device *rdev)
 		else
 			rdev->has_uvd = true;
 		break;
+	case CHIP_BONAIRE:
+		rdev->asic = &ci_asic;
+		rdev->num_crtc = 6;
+		break;
+	case CHIP_KAVERI:
+	case CHIP_KABINI:
+		rdev->asic = &kv_asic;
+		/* set num crtcs */
+		if (rdev->family == CHIP_KAVERI)
+			rdev->num_crtc = 4;
+		else
+			rdev->num_crtc = 2;
+		break;
 	default:
 		/* FIXME: not supported yet */
 		return -EINVAL;
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index a72759ede753..ca1895709908 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -47,6 +47,12 @@ u8 atombios_get_backlight_level(struct radeon_encoder *radeon_encoder);
 void radeon_legacy_set_backlight_level(struct radeon_encoder *radeon_encoder, u8 level);
 u8 radeon_legacy_get_backlight_level(struct radeon_encoder *radeon_encoder);
 
+u32 radeon_ring_generic_get_rptr(struct radeon_device *rdev,
+				 struct radeon_ring *ring);
+u32 radeon_ring_generic_get_wptr(struct radeon_device *rdev,
+				 struct radeon_ring *ring);
+void radeon_ring_generic_set_wptr(struct radeon_device *rdev,
+				  struct radeon_ring *ring);
 
 /*
  * r100,rv100,rs100,rv200,rs200
@@ -334,6 +340,9 @@ int r600_uvd_ring_test(struct radeon_device *rdev, struct radeon_ring *ring);
 int r600_copy_blit(struct radeon_device *rdev,
 		   uint64_t src_offset, uint64_t dst_offset,
 		   unsigned num_gpu_pages, struct radeon_fence **fence);
+int r600_copy_cpdma(struct radeon_device *rdev,
+		    uint64_t src_offset, uint64_t dst_offset,
+		    unsigned num_gpu_pages, struct radeon_fence **fence);
 int r600_copy_dma(struct radeon_device *rdev,
 		  uint64_t src_offset, uint64_t dst_offset,
 		  unsigned num_gpu_pages, struct radeon_fence **fence);
@@ -395,6 +404,37 @@ void r600_kms_blit_copy(struct radeon_device *rdev,
 int r600_mc_wait_for_idle(struct radeon_device *rdev);
 u32 r600_get_xclk(struct radeon_device *rdev);
 uint64_t r600_get_gpu_clock_counter(struct radeon_device *rdev);
+int rv6xx_get_temp(struct radeon_device *rdev);
+int r600_dpm_pre_set_power_state(struct radeon_device *rdev);
+void r600_dpm_post_set_power_state(struct radeon_device *rdev);
+/* rv6xx dpm */
+int rv6xx_dpm_init(struct radeon_device *rdev);
+int rv6xx_dpm_enable(struct radeon_device *rdev);
+void rv6xx_dpm_disable(struct radeon_device *rdev);
+int rv6xx_dpm_set_power_state(struct radeon_device *rdev);
+void rv6xx_setup_asic(struct radeon_device *rdev);
+void rv6xx_dpm_display_configuration_changed(struct radeon_device *rdev);
+void rv6xx_dpm_fini(struct radeon_device *rdev);
+u32 rv6xx_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 rv6xx_dpm_get_mclk(struct radeon_device *rdev, bool low);
+void rv6xx_dpm_print_power_state(struct radeon_device *rdev,
+				 struct radeon_ps *ps);
+void rv6xx_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						       struct seq_file *m);
+/* rs780 dpm */
+int rs780_dpm_init(struct radeon_device *rdev);
+int rs780_dpm_enable(struct radeon_device *rdev);
+void rs780_dpm_disable(struct radeon_device *rdev);
+int rs780_dpm_set_power_state(struct radeon_device *rdev);
+void rs780_dpm_setup_asic(struct radeon_device *rdev);
+void rs780_dpm_display_configuration_changed(struct radeon_device *rdev);
+void rs780_dpm_fini(struct radeon_device *rdev);
+u32 rs780_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 rs780_dpm_get_mclk(struct radeon_device *rdev, bool low);
+void rs780_dpm_print_power_state(struct radeon_device *rdev,
+				 struct radeon_ps *ps);
+void rs780_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						       struct seq_file *m);
 
 /* uvd */
 int r600_uvd_init(struct radeon_device *rdev);
@@ -428,6 +468,24 @@ int rv770_copy_dma(struct radeon_device *rdev,
 u32 rv770_get_xclk(struct radeon_device *rdev);
 int rv770_uvd_resume(struct radeon_device *rdev);
 int rv770_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
+int rv770_get_temp(struct radeon_device *rdev);
+/* rv7xx pm */
+int rv770_dpm_init(struct radeon_device *rdev);
+int rv770_dpm_enable(struct radeon_device *rdev);
+void rv770_dpm_disable(struct radeon_device *rdev);
+int rv770_dpm_set_power_state(struct radeon_device *rdev);
+void rv770_dpm_setup_asic(struct radeon_device *rdev);
+void rv770_dpm_display_configuration_changed(struct radeon_device *rdev);
+void rv770_dpm_fini(struct radeon_device *rdev);
+u32 rv770_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 rv770_dpm_get_mclk(struct radeon_device *rdev, bool low);
+void rv770_dpm_print_power_state(struct radeon_device *rdev,
+				 struct radeon_ps *ps);
+void rv770_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						       struct seq_file *m);
+int rv770_dpm_force_performance_level(struct radeon_device *rdev,
+				      enum radeon_dpm_forced_level level);
+bool rv770_dpm_vblank_too_short(struct radeon_device *rdev);
 
 /*
  * evergreen
@@ -482,6 +540,45 @@ int evergreen_copy_dma(struct radeon_device *rdev,
 		       struct radeon_fence **fence);
 void evergreen_hdmi_enable(struct drm_encoder *encoder, bool enable);
 void evergreen_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mode);
+int evergreen_get_temp(struct radeon_device *rdev);
+int sumo_get_temp(struct radeon_device *rdev);
+int tn_get_temp(struct radeon_device *rdev);
+int cypress_dpm_init(struct radeon_device *rdev);
+void cypress_dpm_setup_asic(struct radeon_device *rdev);
+int cypress_dpm_enable(struct radeon_device *rdev);
+void cypress_dpm_disable(struct radeon_device *rdev);
+int cypress_dpm_set_power_state(struct radeon_device *rdev);
+void cypress_dpm_display_configuration_changed(struct radeon_device *rdev);
+void cypress_dpm_fini(struct radeon_device *rdev);
+bool cypress_dpm_vblank_too_short(struct radeon_device *rdev);
+int btc_dpm_init(struct radeon_device *rdev);
+void btc_dpm_setup_asic(struct radeon_device *rdev);
+int btc_dpm_enable(struct radeon_device *rdev);
+void btc_dpm_disable(struct radeon_device *rdev);
+int btc_dpm_pre_set_power_state(struct radeon_device *rdev);
+int btc_dpm_set_power_state(struct radeon_device *rdev);
+void btc_dpm_post_set_power_state(struct radeon_device *rdev);
+void btc_dpm_fini(struct radeon_device *rdev);
+u32 btc_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 btc_dpm_get_mclk(struct radeon_device *rdev, bool low);
+bool btc_dpm_vblank_too_short(struct radeon_device *rdev);
+int sumo_dpm_init(struct radeon_device *rdev);
+int sumo_dpm_enable(struct radeon_device *rdev);
+void sumo_dpm_disable(struct radeon_device *rdev);
+int sumo_dpm_pre_set_power_state(struct radeon_device *rdev);
+int sumo_dpm_set_power_state(struct radeon_device *rdev);
+void sumo_dpm_post_set_power_state(struct radeon_device *rdev);
+void sumo_dpm_setup_asic(struct radeon_device *rdev);
+void sumo_dpm_display_configuration_changed(struct radeon_device *rdev);
+void sumo_dpm_fini(struct radeon_device *rdev);
+u32 sumo_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 sumo_dpm_get_mclk(struct radeon_device *rdev, bool low);
+void sumo_dpm_print_power_state(struct radeon_device *rdev,
+				struct radeon_ps *ps);
+void sumo_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						      struct seq_file *m);
+int sumo_dpm_force_performance_level(struct radeon_device *rdev,
+				     enum radeon_dpm_forced_level level);
 
 /*
  * cayman
@@ -516,6 +613,41 @@ bool cayman_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
 bool cayman_dma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
 void cayman_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
 
+int ni_dpm_init(struct radeon_device *rdev);
+void ni_dpm_setup_asic(struct radeon_device *rdev);
+int ni_dpm_enable(struct radeon_device *rdev);
+void ni_dpm_disable(struct radeon_device *rdev);
+int ni_dpm_pre_set_power_state(struct radeon_device *rdev);
+int ni_dpm_set_power_state(struct radeon_device *rdev);
+void ni_dpm_post_set_power_state(struct radeon_device *rdev);
+void ni_dpm_fini(struct radeon_device *rdev);
+u32 ni_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 ni_dpm_get_mclk(struct radeon_device *rdev, bool low);
+void ni_dpm_print_power_state(struct radeon_device *rdev,
+			      struct radeon_ps *ps);
+void ni_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						    struct seq_file *m);
+int ni_dpm_force_performance_level(struct radeon_device *rdev,
+				   enum radeon_dpm_forced_level level);
+bool ni_dpm_vblank_too_short(struct radeon_device *rdev);
+int trinity_dpm_init(struct radeon_device *rdev);
+int trinity_dpm_enable(struct radeon_device *rdev);
+void trinity_dpm_disable(struct radeon_device *rdev);
+int trinity_dpm_pre_set_power_state(struct radeon_device *rdev);
+int trinity_dpm_set_power_state(struct radeon_device *rdev);
+void trinity_dpm_post_set_power_state(struct radeon_device *rdev);
+void trinity_dpm_setup_asic(struct radeon_device *rdev);
+void trinity_dpm_display_configuration_changed(struct radeon_device *rdev);
+void trinity_dpm_fini(struct radeon_device *rdev);
+u32 trinity_dpm_get_sclk(struct radeon_device *rdev, bool low);
+u32 trinity_dpm_get_mclk(struct radeon_device *rdev, bool low);
+void trinity_dpm_print_power_state(struct radeon_device *rdev,
+				   struct radeon_ps *ps);
+void trinity_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+							 struct seq_file *m);
+int trinity_dpm_force_performance_level(struct radeon_device *rdev,
+					enum radeon_dpm_forced_level level);
+
 /* DCE6 - SI */
 void dce6_bandwidth_update(struct radeon_device *rdev);
 
@@ -552,5 +684,82 @@ void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 u32 si_get_xclk(struct radeon_device *rdev);
 uint64_t si_get_gpu_clock_counter(struct radeon_device *rdev);
 int si_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
+int si_get_temp(struct radeon_device *rdev);
+int si_dpm_init(struct radeon_device *rdev);
+void si_dpm_setup_asic(struct radeon_device *rdev);
+int si_dpm_enable(struct radeon_device *rdev);
+void si_dpm_disable(struct radeon_device *rdev);
+int si_dpm_pre_set_power_state(struct radeon_device *rdev);
+int si_dpm_set_power_state(struct radeon_device *rdev);
+void si_dpm_post_set_power_state(struct radeon_device *rdev);
+void si_dpm_fini(struct radeon_device *rdev);
+void si_dpm_display_configuration_changed(struct radeon_device *rdev);
+void si_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						    struct seq_file *m);
+int si_dpm_force_performance_level(struct radeon_device *rdev,
+				   enum radeon_dpm_forced_level level);
+
+/* DCE8 - CIK */
+void dce8_bandwidth_update(struct radeon_device *rdev);
+
+/*
+ * cik
+ */
+uint64_t cik_get_gpu_clock_counter(struct radeon_device *rdev);
+u32 cik_get_xclk(struct radeon_device *rdev);
+uint32_t cik_pciep_rreg(struct radeon_device *rdev, uint32_t reg);
+void cik_pciep_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
+int cik_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk);
+int cik_uvd_resume(struct radeon_device *rdev);
+void cik_sdma_fence_ring_emit(struct radeon_device *rdev,
+			      struct radeon_fence *fence);
+void cik_sdma_semaphore_ring_emit(struct radeon_device *rdev,
+				  struct radeon_ring *ring,
+				  struct radeon_semaphore *semaphore,
+				  bool emit_wait);
+void cik_sdma_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
+int cik_copy_dma(struct radeon_device *rdev,
+		 uint64_t src_offset, uint64_t dst_offset,
+		 unsigned num_gpu_pages,
+		 struct radeon_fence **fence);
+int cik_sdma_ring_test(struct radeon_device *rdev, struct radeon_ring *ring);
+int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
+bool cik_sdma_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring);
+void cik_fence_gfx_ring_emit(struct radeon_device *rdev,
+			     struct radeon_fence *fence);
+void cik_fence_compute_ring_emit(struct radeon_device *rdev,
+				 struct radeon_fence *fence);
+void cik_semaphore_ring_emit(struct radeon_device *rdev,
+			     struct radeon_ring *cp,
+			     struct radeon_semaphore *semaphore,
+			     bool emit_wait);
+void cik_pcie_gart_tlb_flush(struct radeon_device *rdev);
+int cik_init(struct radeon_device *rdev);
+void cik_fini(struct radeon_device *rdev);
+int cik_suspend(struct radeon_device *rdev);
+int cik_resume(struct radeon_device *rdev);
+bool cik_gfx_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+int cik_asic_reset(struct radeon_device *rdev);
+void cik_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
+int cik_ring_test(struct radeon_device *rdev, struct radeon_ring *ring);
+int cik_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
+int cik_irq_set(struct radeon_device *rdev);
+int cik_irq_process(struct radeon_device *rdev);
+int cik_vm_init(struct radeon_device *rdev);
+void cik_vm_fini(struct radeon_device *rdev);
+void cik_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+void cik_vm_set_page(struct radeon_device *rdev,
+		     struct radeon_ib *ib,
+		     uint64_t pe,
+		     uint64_t addr, unsigned count,
+		     uint32_t incr, uint32_t flags);
+void cik_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm);
+int cik_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
+u32 cik_compute_ring_get_rptr(struct radeon_device *rdev,
+			      struct radeon_ring *ring);
+u32 cik_compute_ring_get_wptr(struct radeon_device *rdev,
+			      struct radeon_ring *ring);
+void cik_compute_ring_set_wptr(struct radeon_device *rdev,
+			       struct radeon_ring *ring);
 
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c
index dea6f63c9724..e3f3e8841789 100644
--- a/drivers/gpu/drm/radeon/radeon_atombios.c
+++ b/drivers/gpu/drm/radeon/radeon_atombios.c
@@ -56,10 +56,6 @@ extern void
 radeon_add_legacy_encoder(struct drm_device *dev, uint32_t encoder_enum,
 			  uint32_t supported_device);
 
-/* local */
-static int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
-				    u16 voltage_id, u16 *voltage);
-
 union atom_supported_devices {
 	struct _ATOM_SUPPORTED_DEVICES_INFO info;
 	struct _ATOM_SUPPORTED_DEVICES_INFO_2 info_2;
@@ -1247,6 +1243,7 @@ bool radeon_atom_get_clock_info(struct drm_device *dev)
 			}
 			rdev->clock.dp_extclk =
 				le16_to_cpu(firmware_info->info_21.usUniphyDPModeExtClkFreq);
+			rdev->clock.current_dispclk = rdev->clock.default_dispclk;
 		}
 		*dcpll = *p1pll;
 
@@ -1269,6 +1266,7 @@ union igp_info {
 	struct _ATOM_INTEGRATED_SYSTEM_INFO_V2 info_2;
 	struct _ATOM_INTEGRATED_SYSTEM_INFO_V6 info_6;
 	struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7 info_7;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_8 info_8;
 };
 
 bool radeon_atombios_sideport_present(struct radeon_device *rdev)
@@ -1438,6 +1436,22 @@ static void radeon_atombios_get_igp_ss_overrides(struct radeon_device *rdev,
 				break;
 			}
 			break;
+		case 8:
+			switch (id) {
+			case ASIC_INTERNAL_SS_ON_TMDS:
+				percentage = le16_to_cpu(igp_info->info_8.usDVISSPercentage);
+				rate = le16_to_cpu(igp_info->info_8.usDVISSpreadRateIn10Hz);
+				break;
+			case ASIC_INTERNAL_SS_ON_HDMI:
+				percentage = le16_to_cpu(igp_info->info_8.usHDMISSPercentage);
+				rate = le16_to_cpu(igp_info->info_8.usHDMISSpreadRateIn10Hz);
+				break;
+			case ASIC_INTERNAL_SS_ON_LVDS:
+				percentage = le16_to_cpu(igp_info->info_8.usLvdsSSPercentage);
+				rate = le16_to_cpu(igp_info->info_8.usLvdsSSpreadRateIn10Hz);
+				break;
+			}
+			break;
 		default:
 			DRM_ERROR("Unsupported IGP table: %d %d\n", frev, crev);
 			break;
@@ -1499,6 +1513,10 @@ bool radeon_atombios_get_asic_ss_info(struct radeon_device *rdev,
 						le16_to_cpu(ss_info->info_2.asSpreadSpectrum[i].usSpreadSpectrumPercentage);
 					ss->type = ss_info->info_2.asSpreadSpectrum[i].ucSpreadSpectrumMode;
 					ss->rate = le16_to_cpu(ss_info->info_2.asSpreadSpectrum[i].usSpreadRateIn10Hz);
+					if ((crev == 2) &&
+					    ((id == ASIC_INTERNAL_ENGINE_SS) ||
+					     (id == ASIC_INTERNAL_MEMORY_SS)))
+						ss->rate /= 100;
 					return true;
 				}
 			}
@@ -1513,6 +1531,9 @@ bool radeon_atombios_get_asic_ss_info(struct radeon_device *rdev,
 						le16_to_cpu(ss_info->info_3.asSpreadSpectrum[i].usSpreadSpectrumPercentage);
 					ss->type = ss_info->info_3.asSpreadSpectrum[i].ucSpreadSpectrumMode;
 					ss->rate = le16_to_cpu(ss_info->info_3.asSpreadSpectrum[i].usSpreadRateIn10Hz);
+					if ((id == ASIC_INTERNAL_ENGINE_SS) ||
+					    (id == ASIC_INTERNAL_MEMORY_SS))
+						ss->rate /= 100;
 					if (rdev->flags & RADEON_IS_IGP)
 						radeon_atombios_get_igp_ss_overrides(rdev, ss, id);
 					return true;
@@ -1927,6 +1948,7 @@ static const char *pp_lib_thermal_controller_names[] = {
 	"Northern Islands",
 	"Southern Islands",
 	"lm96163",
+	"Sea Islands",
 };
 
 union power_info {
@@ -1944,6 +1966,7 @@ union pplib_clock_info {
 	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
 	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
 	struct _ATOM_PPLIB_SI_CLOCK_INFO si;
+	struct _ATOM_PPLIB_CI_CLOCK_INFO ci;
 };
 
 union pplib_power_state {
@@ -2209,6 +2232,11 @@ static void radeon_atombios_add_pplib_thermal_controller(struct radeon_device *r
 				 (controller->ucFanParameters &
 				  ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with");
 			rdev->pm.int_thermal_type = THERMAL_TYPE_SI;
+		} else if (controller->ucType == ATOM_PP_THERMALCONTROLLER_CISLANDS) {
+			DRM_INFO("Internal thermal controller %s fan control\n",
+				 (controller->ucFanParameters &
+				  ATOM_PP_FANPARAMETERS_NOFAN) ? "without" : "with");
+			rdev->pm.int_thermal_type = THERMAL_TYPE_CI;
 		} else if ((controller->ucType ==
 			    ATOM_PP_THERMALCONTROLLER_EXTERNAL_GPIO) ||
 			   (controller->ucType ==
@@ -2241,8 +2269,8 @@ static void radeon_atombios_add_pplib_thermal_controller(struct radeon_device *r
 	}
 }
 
-static void radeon_atombios_get_default_voltages(struct radeon_device *rdev,
-						 u16 *vddc, u16 *vddci)
+void radeon_atombios_get_default_voltages(struct radeon_device *rdev,
+					  u16 *vddc, u16 *vddci, u16 *mvdd)
 {
 	struct radeon_mode_info *mode_info = &rdev->mode_info;
 	int index = GetIndexIntoMasterTable(DATA, FirmwareInfo);
@@ -2252,6 +2280,7 @@ static void radeon_atombios_get_default_voltages(struct radeon_device *rdev,
 
 	*vddc = 0;
 	*vddci = 0;
+	*mvdd = 0;
 
 	if (atom_parse_data_header(mode_info->atom_context, index, NULL,
 				   &frev, &crev, &data_offset)) {
@@ -2259,8 +2288,10 @@ static void radeon_atombios_get_default_voltages(struct radeon_device *rdev,
 			(union firmware_info *)(mode_info->atom_context->bios +
 						data_offset);
 		*vddc = le16_to_cpu(firmware_info->info_14.usBootUpVDDCVoltage);
-		if ((frev == 2) && (crev >= 2))
+		if ((frev == 2) && (crev >= 2)) {
 			*vddci = le16_to_cpu(firmware_info->info_22.usBootUpVDDCIVoltage);
+			*mvdd = le16_to_cpu(firmware_info->info_22.usBootUpMVDDCVoltage);
+		}
 	}
 }
 
@@ -2271,9 +2302,9 @@ static void radeon_atombios_parse_pplib_non_clock_info(struct radeon_device *rde
 	int j;
 	u32 misc = le32_to_cpu(non_clock_info->ulCapsAndSettings);
 	u32 misc2 = le16_to_cpu(non_clock_info->usClassification);
-	u16 vddc, vddci;
+	u16 vddc, vddci, mvdd;
 
-	radeon_atombios_get_default_voltages(rdev, &vddc, &vddci);
+	radeon_atombios_get_default_voltages(rdev, &vddc, &vddci, &mvdd);
 
 	rdev->pm.power_state[state_index].misc = misc;
 	rdev->pm.power_state[state_index].misc2 = misc2;
@@ -2316,7 +2347,13 @@ static void radeon_atombios_parse_pplib_non_clock_info(struct radeon_device *rde
 			rdev->pm.default_vddc = rdev->pm.power_state[state_index].clock_info[0].voltage.voltage;
 			rdev->pm.default_vddci = rdev->pm.power_state[state_index].clock_info[0].voltage.vddci;
 		} else {
-			/* patch the table values with the default slck/mclk from firmware info */
+			u16 max_vddci = 0;
+
+			if (ASIC_IS_DCE4(rdev))
+				radeon_atom_get_max_voltage(rdev,
+							    SET_VOLTAGE_TYPE_ASIC_VDDCI,
+							    &max_vddci);
+			/* patch the table values with the default sclk/mclk from firmware info */
 			for (j = 0; j < mode_index; j++) {
 				rdev->pm.power_state[state_index].clock_info[j].mclk =
 					rdev->clock.default_mclk;
@@ -2325,6 +2362,9 @@ static void radeon_atombios_parse_pplib_non_clock_info(struct radeon_device *rde
 				if (vddc)
 					rdev->pm.power_state[state_index].clock_info[j].voltage.voltage =
 						vddc;
+				if (max_vddci)
+					rdev->pm.power_state[state_index].clock_info[j].voltage.vddci =
+						max_vddci;
 			}
 		}
 	}
@@ -2347,6 +2387,15 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev,
 			sclk |= clock_info->rs780.ucLowEngineClockHigh << 16;
 			rdev->pm.power_state[state_index].clock_info[mode_index].sclk = sclk;
 		}
+	} else if (rdev->family >= CHIP_BONAIRE) {
+		sclk = le16_to_cpu(clock_info->ci.usEngineClockLow);
+		sclk |= clock_info->ci.ucEngineClockHigh << 16;
+		mclk = le16_to_cpu(clock_info->ci.usMemoryClockLow);
+		mclk |= clock_info->ci.ucMemoryClockHigh << 16;
+		rdev->pm.power_state[state_index].clock_info[mode_index].mclk = mclk;
+		rdev->pm.power_state[state_index].clock_info[mode_index].sclk = sclk;
+		rdev->pm.power_state[state_index].clock_info[mode_index].voltage.type =
+			VOLTAGE_NONE;
 	} else if (rdev->family >= CHIP_TAHITI) {
 		sclk = le16_to_cpu(clock_info->si.usEngineClockLow);
 		sclk |= clock_info->si.ucEngineClockHigh << 16;
@@ -2392,6 +2441,10 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev,
 	case ATOM_VIRTUAL_VOLTAGE_ID1:
 	case ATOM_VIRTUAL_VOLTAGE_ID2:
 	case ATOM_VIRTUAL_VOLTAGE_ID3:
+	case ATOM_VIRTUAL_VOLTAGE_ID4:
+	case ATOM_VIRTUAL_VOLTAGE_ID5:
+	case ATOM_VIRTUAL_VOLTAGE_ID6:
+	case ATOM_VIRTUAL_VOLTAGE_ID7:
 		if (radeon_atom_get_max_vddc(rdev, VOLTAGE_TYPE_VDDC,
 					     rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage,
 					     &vddc) == 0)
@@ -2667,6 +2720,8 @@ union get_clock_dividers {
 	struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V3 v3;
 	struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V4 v4;
 	struct _COMPUTE_MEMORY_ENGINE_PLL_PARAMETERS_V5 v5;
+	struct _COMPUTE_GPU_CLOCK_INPUT_PARAMETERS_V1_6 v6_in;
+	struct _COMPUTE_GPU_CLOCK_OUTPUT_PARAMETERS_V1_6 v6_out;
 };
 
 int radeon_atom_get_clock_dividers(struct radeon_device *rdev,
@@ -2699,7 +2754,8 @@ int radeon_atom_get_clock_dividers(struct radeon_device *rdev,
 		break;
 	case 2:
 	case 3:
-		/* r6xx, r7xx, evergreen, ni */
+	case 5:
+		/* r6xx, r7xx, evergreen, ni, si */
 		if (rdev->family <= CHIP_RV770) {
 			args.v2.ucAction = clock_type;
 			args.v2.ulClock = cpu_to_le32(clock);	/* 10 khz */
@@ -2732,6 +2788,9 @@ int radeon_atom_get_clock_dividers(struct radeon_device *rdev,
 				dividers->vco_mode = (args.v3.ucCntlFlag &
 						      ATOM_PLL_CNTL_FLAG_MPLL_VCO_MODE) ? 1 : 0;
 			} else {
+				/* for SI we use ComputeMemoryClockParam for memory plls */
+				if (rdev->family >= CHIP_TAHITI)
+					return -EINVAL;
 				args.v5.ulClockParams = cpu_to_le32((clock_type << 24) | clock);
 				if (strobe_mode)
 					args.v5.ucInputFlag = ATOM_PLL_INPUT_FLAG_PLL_STROBE_MODE_EN;
@@ -2757,9 +2816,76 @@ int radeon_atom_get_clock_dividers(struct radeon_device *rdev,
 
 		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 
-		dividers->post_div = args.v4.ucPostDiv;
+		dividers->post_divider = dividers->post_div = args.v4.ucPostDiv;
 		dividers->real_clock = le32_to_cpu(args.v4.ulClock);
 		break;
+	case 6:
+		/* CI */
+		/* COMPUTE_GPUCLK_INPUT_FLAG_DEFAULT_GPUCLK, COMPUTE_GPUCLK_INPUT_FLAG_SCLK */
+		args.v6_in.ulClock.ulComputeClockFlag = clock_type;
+		args.v6_in.ulClock.ulClockFreq = cpu_to_le32(clock);	/* 10 khz */
+
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+		dividers->whole_fb_div = le16_to_cpu(args.v6_out.ulFbDiv.usFbDiv);
+		dividers->frac_fb_div = le16_to_cpu(args.v6_out.ulFbDiv.usFbDivFrac);
+		dividers->ref_div = args.v6_out.ucPllRefDiv;
+		dividers->post_div = args.v6_out.ucPllPostDiv;
+		dividers->flags = args.v6_out.ucPllCntlFlag;
+		dividers->real_clock = le32_to_cpu(args.v6_out.ulClock.ulClock);
+		dividers->post_divider = args.v6_out.ulClock.ucPostDiv;
+		break;
+	default:
+		return -EINVAL;
+	}
+	return 0;
+}
+
+int radeon_atom_get_memory_pll_dividers(struct radeon_device *rdev,
+					u32 clock,
+					bool strobe_mode,
+					struct atom_mpll_param *mpll_param)
+{
+	COMPUTE_MEMORY_CLOCK_PARAM_PARAMETERS_V2_1 args;
+	int index = GetIndexIntoMasterTable(COMMAND, ComputeMemoryClockParam);
+	u8 frev, crev;
+
+	memset(&args, 0, sizeof(args));
+	memset(mpll_param, 0, sizeof(struct atom_mpll_param));
+
+	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
+		return -EINVAL;
+
+	switch (frev) {
+	case 2:
+		switch (crev) {
+		case 1:
+			/* SI */
+			args.ulClock = cpu_to_le32(clock);	/* 10 khz */
+			args.ucInputFlag = 0;
+			if (strobe_mode)
+				args.ucInputFlag |= MPLL_INPUT_FLAG_STROBE_MODE_EN;
+
+			atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+			mpll_param->clkfrac = le16_to_cpu(args.ulFbDiv.usFbDivFrac);
+			mpll_param->clkf = le16_to_cpu(args.ulFbDiv.usFbDiv);
+			mpll_param->post_div = args.ucPostDiv;
+			mpll_param->dll_speed = args.ucDllSpeed;
+			mpll_param->bwcntl = args.ucBWCntl;
+			mpll_param->vco_mode =
+				(args.ucPllCntlFlag & MPLL_CNTL_FLAG_VCO_MODE_MASK) ? 1 : 0;
+			mpll_param->yclk_sel =
+				(args.ucPllCntlFlag & MPLL_CNTL_FLAG_BYPASS_DQ_PLL) ? 1 : 0;
+			mpll_param->qdr =
+				(args.ucPllCntlFlag & MPLL_CNTL_FLAG_QDR_ENABLE) ? 1 : 0;
+			mpll_param->half_rate =
+				(args.ucPllCntlFlag & MPLL_CNTL_FLAG_AD_HALF_RATE) ? 1 : 0;
+			break;
+		default:
+			return -EINVAL;
+		}
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -2819,6 +2945,48 @@ void radeon_atom_set_memory_clock(struct radeon_device *rdev,
 	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 }
 
+void radeon_atom_set_engine_dram_timings(struct radeon_device *rdev,
+					 u32 eng_clock, u32 mem_clock)
+{
+	SET_ENGINE_CLOCK_PS_ALLOCATION args;
+	int index = GetIndexIntoMasterTable(COMMAND, DynamicMemorySettings);
+	u32 tmp;
+
+	memset(&args, 0, sizeof(args));
+
+	tmp = eng_clock & SET_CLOCK_FREQ_MASK;
+	tmp |= (COMPUTE_ENGINE_PLL_PARAM << 24);
+
+	args.ulTargetEngineClock = cpu_to_le32(tmp);
+	if (mem_clock)
+		args.sReserved.ulClock = cpu_to_le32(mem_clock & SET_CLOCK_FREQ_MASK);
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+}
+
+void radeon_atom_update_memory_dll(struct radeon_device *rdev,
+				   u32 mem_clock)
+{
+	u32 args;
+	int index = GetIndexIntoMasterTable(COMMAND, DynamicMemorySettings);
+
+	args = cpu_to_le32(mem_clock);	/* 10 khz */
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+}
+
+void radeon_atom_set_ac_timing(struct radeon_device *rdev,
+			       u32 mem_clock)
+{
+	SET_MEMORY_CLOCK_PS_ALLOCATION args;
+	int index = GetIndexIntoMasterTable(COMMAND, DynamicMemorySettings);
+	u32 tmp = mem_clock | (COMPUTE_MEMORY_PLL_PARAM << 24);
+
+	args.ulTargetMemoryClock = cpu_to_le32(tmp);	/* 10 khz */
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+}
+
 union set_voltage {
 	struct _SET_VOLTAGE_PS_ALLOCATION alloc;
 	struct _SET_VOLTAGE_PARAMETERS v1;
@@ -2863,8 +3031,8 @@ void radeon_atom_set_voltage(struct radeon_device *rdev, u16 voltage_level, u8 v
 	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 }
 
-static int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
-				    u16 voltage_id, u16 *voltage)
+int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
+			     u16 voltage_id, u16 *voltage)
 {
 	union set_voltage args;
 	int index = GetIndexIntoMasterTable(COMMAND, SetVoltage);
@@ -2902,6 +3070,693 @@ static int radeon_atom_get_max_vddc(struct radeon_device *rdev, u8 voltage_type,
 	return 0;
 }
 
+int radeon_atom_get_leakage_vddc_based_on_leakage_idx(struct radeon_device *rdev,
+						      u16 *voltage,
+						      u16 leakage_idx)
+{
+	return radeon_atom_get_max_vddc(rdev, VOLTAGE_TYPE_VDDC, leakage_idx, voltage);
+}
+
+int radeon_atom_get_voltage_gpio_settings(struct radeon_device *rdev,
+					  u16 voltage_level, u8 voltage_type,
+					  u32 *gpio_value, u32 *gpio_mask)
+{
+	union set_voltage args;
+	int index = GetIndexIntoMasterTable(COMMAND, SetVoltage);
+	u8 frev, crev;
+
+	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
+		return -EINVAL;
+
+	switch (crev) {
+	case 1:
+		return -EINVAL;
+	case 2:
+		args.v2.ucVoltageType = voltage_type;
+		args.v2.ucVoltageMode = SET_ASIC_VOLTAGE_MODE_GET_GPIOMASK;
+		args.v2.usVoltageLevel = cpu_to_le16(voltage_level);
+
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+		*gpio_mask = le32_to_cpu(*(u32 *)&args.v2);
+
+		args.v2.ucVoltageType = voltage_type;
+		args.v2.ucVoltageMode = SET_ASIC_VOLTAGE_MODE_GET_GPIOVAL;
+		args.v2.usVoltageLevel = cpu_to_le16(voltage_level);
+
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+		*gpio_value = le32_to_cpu(*(u32 *)&args.v2);
+		break;
+	default:
+		DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+union voltage_object_info {
+	struct _ATOM_VOLTAGE_OBJECT_INFO v1;
+	struct _ATOM_VOLTAGE_OBJECT_INFO_V2 v2;
+	struct _ATOM_VOLTAGE_OBJECT_INFO_V3_1 v3;
+};
+
+union voltage_object {
+	struct _ATOM_VOLTAGE_OBJECT v1;
+	struct _ATOM_VOLTAGE_OBJECT_V2 v2;
+	union _ATOM_VOLTAGE_OBJECT_V3 v3;
+};
+
+static ATOM_VOLTAGE_OBJECT *atom_lookup_voltage_object_v1(ATOM_VOLTAGE_OBJECT_INFO *v1,
+							  u8 voltage_type)
+{
+	u32 size = le16_to_cpu(v1->sHeader.usStructureSize);
+	u32 offset = offsetof(ATOM_VOLTAGE_OBJECT_INFO, asVoltageObj[0]);
+	u8 *start = (u8 *)v1;
+
+	while (offset < size) {
+		ATOM_VOLTAGE_OBJECT *vo = (ATOM_VOLTAGE_OBJECT *)(start + offset);
+		if (vo->ucVoltageType == voltage_type)
+			return vo;
+		offset += offsetof(ATOM_VOLTAGE_OBJECT, asFormula.ucVIDAdjustEntries) +
+			vo->asFormula.ucNumOfVoltageEntries;
+	}
+	return NULL;
+}
+
+static ATOM_VOLTAGE_OBJECT_V2 *atom_lookup_voltage_object_v2(ATOM_VOLTAGE_OBJECT_INFO_V2 *v2,
+							     u8 voltage_type)
+{
+	u32 size = le16_to_cpu(v2->sHeader.usStructureSize);
+	u32 offset = offsetof(ATOM_VOLTAGE_OBJECT_INFO_V2, asVoltageObj[0]);
+	u8 *start = (u8*)v2;
+
+	while (offset < size) {
+		ATOM_VOLTAGE_OBJECT_V2 *vo = (ATOM_VOLTAGE_OBJECT_V2 *)(start + offset);
+		if (vo->ucVoltageType == voltage_type)
+			return vo;
+		offset += offsetof(ATOM_VOLTAGE_OBJECT_V2, asFormula.asVIDAdjustEntries) +
+			(vo->asFormula.ucNumOfVoltageEntries * sizeof(VOLTAGE_LUT_ENTRY));
+	}
+	return NULL;
+}
+
+static ATOM_VOLTAGE_OBJECT_V3 *atom_lookup_voltage_object_v3(ATOM_VOLTAGE_OBJECT_INFO_V3_1 *v3,
+							     u8 voltage_type, u8 voltage_mode)
+{
+	u32 size = le16_to_cpu(v3->sHeader.usStructureSize);
+	u32 offset = offsetof(ATOM_VOLTAGE_OBJECT_INFO_V3_1, asVoltageObj[0]);
+	u8 *start = (u8*)v3;
+
+	while (offset < size) {
+		ATOM_VOLTAGE_OBJECT_V3 *vo = (ATOM_VOLTAGE_OBJECT_V3 *)(start + offset);
+		if ((vo->asGpioVoltageObj.sHeader.ucVoltageType == voltage_type) &&
+		    (vo->asGpioVoltageObj.sHeader.ucVoltageMode == voltage_mode))
+			return vo;
+		offset += le16_to_cpu(vo->asGpioVoltageObj.sHeader.usSize);
+	}
+	return NULL;
+}
+
+bool
+radeon_atom_is_voltage_gpio(struct radeon_device *rdev,
+			    u8 voltage_type, u8 voltage_mode)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+
+		switch (frev) {
+		case 1:
+		case 2:
+			switch (crev) {
+			case 1:
+				voltage_object = (union voltage_object *)
+					atom_lookup_voltage_object_v1(&voltage_info->v1, voltage_type);
+				if (voltage_object &&
+				    (voltage_object->v1.asControl.ucVoltageControlId == VOLTAGE_CONTROLLED_BY_GPIO))
+					return true;
+				break;
+			case 2:
+				voltage_object = (union voltage_object *)
+					atom_lookup_voltage_object_v2(&voltage_info->v2, voltage_type);
+				if (voltage_object &&
+				    (voltage_object->v2.asControl.ucVoltageControlId == VOLTAGE_CONTROLLED_BY_GPIO))
+					return true;
+				break;
+			default:
+				DRM_ERROR("unknown voltage object table\n");
+				return false;
+			}
+			break;
+		case 3:
+			switch (crev) {
+			case 1:
+				if (atom_lookup_voltage_object_v3(&voltage_info->v3,
+								  voltage_type, voltage_mode))
+					return true;
+				break;
+			default:
+				DRM_ERROR("unknown voltage object table\n");
+				return false;
+			}
+			break;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return false;
+		}
+
+	}
+	return false;
+}
+
+int radeon_atom_get_max_voltage(struct radeon_device *rdev,
+				u8 voltage_type, u16 *max_voltage)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+
+		switch (crev) {
+		case 1:
+			voltage_object = (union voltage_object *)
+				atom_lookup_voltage_object_v1(&voltage_info->v1, voltage_type);
+			if (voltage_object) {
+				ATOM_VOLTAGE_FORMULA *formula =
+					&voltage_object->v1.asFormula;
+				if (formula->ucFlag & 1)
+					*max_voltage =
+						le16_to_cpu(formula->usVoltageBaseLevel) +
+						formula->ucNumOfVoltageEntries / 2 *
+						le16_to_cpu(formula->usVoltageStep);
+				else
+					*max_voltage =
+						le16_to_cpu(formula->usVoltageBaseLevel) +
+						(formula->ucNumOfVoltageEntries - 1) *
+						le16_to_cpu(formula->usVoltageStep);
+				return 0;
+			}
+			break;
+		case 2:
+			voltage_object = (union voltage_object *)
+				atom_lookup_voltage_object_v2(&voltage_info->v2, voltage_type);
+			if (voltage_object) {
+				ATOM_VOLTAGE_FORMULA_V2 *formula =
+					&voltage_object->v2.asFormula;
+				if (formula->ucNumOfVoltageEntries) {
+					*max_voltage =
+						le16_to_cpu(formula->asVIDAdjustEntries[
+								    formula->ucNumOfVoltageEntries - 1
+								    ].usVoltageValue);
+					return 0;
+				}
+			}
+			break;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return -EINVAL;
+		}
+
+	}
+	return -EINVAL;
+}
+
+int radeon_atom_get_min_voltage(struct radeon_device *rdev,
+				u8 voltage_type, u16 *min_voltage)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+
+		switch (crev) {
+		case 1:
+			voltage_object = (union voltage_object *)
+				atom_lookup_voltage_object_v1(&voltage_info->v1, voltage_type);
+			if (voltage_object) {
+				ATOM_VOLTAGE_FORMULA *formula =
+					&voltage_object->v1.asFormula;
+				*min_voltage =
+					le16_to_cpu(formula->usVoltageBaseLevel);
+				return 0;
+			}
+			break;
+		case 2:
+			voltage_object = (union voltage_object *)
+				atom_lookup_voltage_object_v2(&voltage_info->v2, voltage_type);
+			if (voltage_object) {
+				ATOM_VOLTAGE_FORMULA_V2 *formula =
+					&voltage_object->v2.asFormula;
+				if (formula->ucNumOfVoltageEntries) {
+					*min_voltage =
+						le16_to_cpu(formula->asVIDAdjustEntries[
+								    0
+								    ].usVoltageValue);
+					return 0;
+				}
+			}
+			break;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return -EINVAL;
+		}
+
+	}
+	return -EINVAL;
+}
+
+int radeon_atom_get_voltage_step(struct radeon_device *rdev,
+				 u8 voltage_type, u16 *voltage_step)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+
+		switch (crev) {
+		case 1:
+			voltage_object = (union voltage_object *)
+				atom_lookup_voltage_object_v1(&voltage_info->v1, voltage_type);
+			if (voltage_object) {
+				ATOM_VOLTAGE_FORMULA *formula =
+					&voltage_object->v1.asFormula;
+				if (formula->ucFlag & 1)
+					*voltage_step =
+						(le16_to_cpu(formula->usVoltageStep) + 1) / 2;
+				else
+					*voltage_step =
+						le16_to_cpu(formula->usVoltageStep);
+				return 0;
+			}
+			break;
+		case 2:
+			return -EINVAL;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return -EINVAL;
+		}
+
+	}
+	return -EINVAL;
+}
+
+int radeon_atom_round_to_true_voltage(struct radeon_device *rdev,
+				      u8 voltage_type,
+				      u16 nominal_voltage,
+				      u16 *true_voltage)
+{
+	u16 min_voltage, max_voltage, voltage_step;
+
+	if (radeon_atom_get_max_voltage(rdev, voltage_type, &max_voltage))
+		return -EINVAL;
+	if (radeon_atom_get_min_voltage(rdev, voltage_type, &min_voltage))
+		return -EINVAL;
+	if (radeon_atom_get_voltage_step(rdev, voltage_type, &voltage_step))
+		return -EINVAL;
+
+	if (nominal_voltage <= min_voltage)
+		*true_voltage = min_voltage;
+	else if (nominal_voltage >= max_voltage)
+		*true_voltage = max_voltage;
+	else
+		*true_voltage = min_voltage +
+			((nominal_voltage - min_voltage) / voltage_step) *
+			voltage_step;
+
+	return 0;
+}
+
+int radeon_atom_get_voltage_table(struct radeon_device *rdev,
+				  u8 voltage_type, u8 voltage_mode,
+				  struct atom_voltage_table *voltage_table)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	int i, ret;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+
+		switch (frev) {
+		case 1:
+		case 2:
+			switch (crev) {
+			case 1:
+				DRM_ERROR("old table version %d, %d\n", frev, crev);
+				return -EINVAL;
+			case 2:
+				voltage_object = (union voltage_object *)
+					atom_lookup_voltage_object_v2(&voltage_info->v2, voltage_type);
+				if (voltage_object) {
+					ATOM_VOLTAGE_FORMULA_V2 *formula =
+						&voltage_object->v2.asFormula;
+					if (formula->ucNumOfVoltageEntries > MAX_VOLTAGE_ENTRIES)
+						return -EINVAL;
+					for (i = 0; i < formula->ucNumOfVoltageEntries; i++) {
+						voltage_table->entries[i].value =
+							le16_to_cpu(formula->asVIDAdjustEntries[i].usVoltageValue);
+						ret = radeon_atom_get_voltage_gpio_settings(rdev,
+											    voltage_table->entries[i].value,
+											    voltage_type,
+											    &voltage_table->entries[i].smio_low,
+											    &voltage_table->mask_low);
+						if (ret)
+							return ret;
+					}
+					voltage_table->count = formula->ucNumOfVoltageEntries;
+					return 0;
+				}
+				break;
+			default:
+				DRM_ERROR("unknown voltage object table\n");
+				return -EINVAL;
+			}
+			break;
+		case 3:
+			switch (crev) {
+			case 1:
+				voltage_object = (union voltage_object *)
+					atom_lookup_voltage_object_v3(&voltage_info->v3,
+								      voltage_type, voltage_mode);
+				if (voltage_object) {
+					ATOM_GPIO_VOLTAGE_OBJECT_V3 *gpio =
+						&voltage_object->v3.asGpioVoltageObj;
+					if (gpio->ucGpioEntryNum > MAX_VOLTAGE_ENTRIES)
+						return -EINVAL;
+					for (i = 0; i < gpio->ucGpioEntryNum; i++) {
+						voltage_table->entries[i].value =
+							le16_to_cpu(gpio->asVolGpioLut[i].usVoltageValue);
+						voltage_table->entries[i].smio_low =
+							le32_to_cpu(gpio->asVolGpioLut[i].ulVoltageId);
+					}
+					voltage_table->mask_low = le32_to_cpu(gpio->ulGpioMaskVal);
+					voltage_table->count = gpio->ucGpioEntryNum;
+					voltage_table->phase_delay = gpio->ucPhaseDelay;
+					return 0;
+				}
+				break;
+			default:
+				DRM_ERROR("unknown voltage object table\n");
+				return -EINVAL;
+			}
+			break;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return -EINVAL;
+		}
+	}
+	return -EINVAL;
+}
+
+union vram_info {
+	struct _ATOM_VRAM_INFO_V3 v1_3;
+	struct _ATOM_VRAM_INFO_V4 v1_4;
+	struct _ATOM_VRAM_INFO_HEADER_V2_1 v2_1;
+};
+
+int radeon_atom_get_memory_info(struct radeon_device *rdev,
+				u8 module_index, struct atom_memory_info *mem_info)
+{
+	int index = GetIndexIntoMasterTable(DATA, VRAM_Info);
+	u8 frev, crev, i;
+	u16 data_offset, size;
+	union vram_info *vram_info;
+
+	memset(mem_info, 0, sizeof(struct atom_memory_info));
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		vram_info = (union vram_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+		switch (frev) {
+		case 1:
+			switch (crev) {
+			case 3:
+				/* r6xx */
+				if (module_index < vram_info->v1_3.ucNumOfVRAMModule) {
+					ATOM_VRAM_MODULE_V3 *vram_module =
+						(ATOM_VRAM_MODULE_V3 *)vram_info->v1_3.aVramInfo;
+
+					for (i = 0; i < module_index; i++) {
+						if (le16_to_cpu(vram_module->usSize) == 0)
+							return -EINVAL;
+						vram_module = (ATOM_VRAM_MODULE_V3 *)
+							((u8 *)vram_module + le16_to_cpu(vram_module->usSize));
+					}
+					mem_info->mem_vendor = vram_module->asMemory.ucMemoryVenderID & 0xf;
+					mem_info->mem_type = vram_module->asMemory.ucMemoryType & 0xf0;
+				} else
+					return -EINVAL;
+				break;
+			case 4:
+				/* r7xx, evergreen */
+				if (module_index < vram_info->v1_4.ucNumOfVRAMModule) {
+					ATOM_VRAM_MODULE_V4 *vram_module =
+						(ATOM_VRAM_MODULE_V4 *)vram_info->v1_4.aVramInfo;
+
+					for (i = 0; i < module_index; i++) {
+						if (le16_to_cpu(vram_module->usModuleSize) == 0)
+							return -EINVAL;
+						vram_module = (ATOM_VRAM_MODULE_V4 *)
+							((u8 *)vram_module + le16_to_cpu(vram_module->usModuleSize));
+					}
+					mem_info->mem_vendor = vram_module->ucMemoryVenderID & 0xf;
+					mem_info->mem_type = vram_module->ucMemoryType & 0xf0;
+				} else
+					return -EINVAL;
+				break;
+			default:
+				DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+				return -EINVAL;
+			}
+			break;
+		case 2:
+			switch (crev) {
+			case 1:
+				/* ni */
+				if (module_index < vram_info->v2_1.ucNumOfVRAMModule) {
+					ATOM_VRAM_MODULE_V7 *vram_module =
+						(ATOM_VRAM_MODULE_V7 *)vram_info->v2_1.aVramInfo;
+
+					for (i = 0; i < module_index; i++) {
+						if (le16_to_cpu(vram_module->usModuleSize) == 0)
+							return -EINVAL;
+						vram_module = (ATOM_VRAM_MODULE_V7 *)
+							((u8 *)vram_module + le16_to_cpu(vram_module->usModuleSize));
+					}
+					mem_info->mem_vendor = vram_module->ucMemoryVenderID & 0xf;
+					mem_info->mem_type = vram_module->ucMemoryType & 0xf0;
+				} else
+					return -EINVAL;
+				break;
+			default:
+				DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+				return -EINVAL;
+			}
+			break;
+		default:
+			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+			return -EINVAL;
+		}
+		return 0;
+	}
+	return -EINVAL;
+}
+
+int radeon_atom_get_mclk_range_table(struct radeon_device *rdev,
+				     bool gddr5, u8 module_index,
+				     struct atom_memory_clock_range_table *mclk_range_table)
+{
+	int index = GetIndexIntoMasterTable(DATA, VRAM_Info);
+	u8 frev, crev, i;
+	u16 data_offset, size;
+	union vram_info *vram_info;
+	u32 mem_timing_size = gddr5 ?
+		sizeof(ATOM_MEMORY_TIMING_FORMAT_V2) : sizeof(ATOM_MEMORY_TIMING_FORMAT);
+	u8 *p;
+
+	memset(mclk_range_table, 0, sizeof(struct atom_memory_clock_range_table));
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		vram_info = (union vram_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+		switch (frev) {
+		case 1:
+			switch (crev) {
+			case 3:
+				DRM_ERROR("old table version %d, %d\n", frev, crev);
+				return -EINVAL;
+			case 4:
+				/* r7xx, evergreen */
+				if (module_index < vram_info->v1_4.ucNumOfVRAMModule) {
+					ATOM_VRAM_MODULE_V4 *vram_module =
+						(ATOM_VRAM_MODULE_V4 *)vram_info->v1_4.aVramInfo;
+
+					for (i = 0; i < module_index; i++) {
+						if (le16_to_cpu(vram_module->usModuleSize) == 0)
+							return -EINVAL;
+						vram_module = (ATOM_VRAM_MODULE_V4 *)
+							((u8 *)vram_module + le16_to_cpu(vram_module->usModuleSize));
+					}
+					mclk_range_table->num_entries = (u8)
+						((le16_to_cpu(vram_module->usModuleSize) - offsetof(ATOM_VRAM_MODULE_V4, asMemTiming)) /
+						 mem_timing_size);
+					p = (u8 *)&vram_module->asMemTiming[0];
+					for (i = 0; i < mclk_range_table->num_entries; i++) {
+						ATOM_MEMORY_TIMING_FORMAT *format = (ATOM_MEMORY_TIMING_FORMAT *)p;
+						mclk_range_table->mclk[i] = le32_to_cpu(format->ulClkRange);
+						p += mem_timing_size;
+					}
+				} else
+					return -EINVAL;
+				break;
+			default:
+				DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+				return -EINVAL;
+			}
+			break;
+		case 2:
+			DRM_ERROR("new table version %d, %d\n", frev, crev);
+			return -EINVAL;
+		default:
+			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+			return -EINVAL;
+		}
+		return 0;
+	}
+	return -EINVAL;
+}
+
+#define MEM_ID_MASK           0xff000000
+#define MEM_ID_SHIFT          24
+#define CLOCK_RANGE_MASK      0x00ffffff
+#define CLOCK_RANGE_SHIFT     0
+#define LOW_NIBBLE_MASK       0xf
+#define DATA_EQU_PREV         0
+#define DATA_FROM_TABLE       4
+
+int radeon_atom_init_mc_reg_table(struct radeon_device *rdev,
+				  u8 module_index,
+				  struct atom_mc_reg_table *reg_table)
+{
+	int index = GetIndexIntoMasterTable(DATA, VRAM_Info);
+	u8 frev, crev, num_entries, t_mem_id, num_ranges = 0;
+	u32 i = 0, j;
+	u16 data_offset, size;
+	union vram_info *vram_info;
+
+	memset(reg_table, 0, sizeof(struct atom_mc_reg_table));
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		vram_info = (union vram_info *)
+			(rdev->mode_info.atom_context->bios + data_offset);
+		switch (frev) {
+		case 1:
+			DRM_ERROR("old table version %d, %d\n", frev, crev);
+			return -EINVAL;
+		case 2:
+			switch (crev) {
+			case 1:
+				if (module_index < vram_info->v2_1.ucNumOfVRAMModule) {
+					ATOM_INIT_REG_BLOCK *reg_block =
+						(ATOM_INIT_REG_BLOCK *)
+						((u8 *)vram_info + le16_to_cpu(vram_info->v2_1.usMemClkPatchTblOffset));
+					ATOM_MEMORY_SETTING_DATA_BLOCK *reg_data =
+						(ATOM_MEMORY_SETTING_DATA_BLOCK *)
+						((u8 *)reg_block + (2 * sizeof(u16)) +
+						 le16_to_cpu(reg_block->usRegIndexTblSize));
+					ATOM_INIT_REG_INDEX_FORMAT *format = &reg_block->asRegIndexBuf[0];
+					num_entries = (u8)((le16_to_cpu(reg_block->usRegIndexTblSize)) /
+							   sizeof(ATOM_INIT_REG_INDEX_FORMAT)) - 1;
+					if (num_entries > VBIOS_MC_REGISTER_ARRAY_SIZE)
+						return -EINVAL;
+					while (i < num_entries) {
+						if (format->ucPreRegDataLength & ACCESS_PLACEHOLDER)
+							break;
+						reg_table->mc_reg_address[i].s1 =
+							(u16)(le16_to_cpu(format->usRegIndex));
+						reg_table->mc_reg_address[i].pre_reg_data =
+							(u8)(format->ucPreRegDataLength);
+						i++;
+						format = (ATOM_INIT_REG_INDEX_FORMAT *)
+							((u8 *)format + sizeof(ATOM_INIT_REG_INDEX_FORMAT));
+					}
+					reg_table->last = i;
+					while ((*(u32 *)reg_data != END_OF_REG_DATA_BLOCK) &&
+					       (num_ranges < VBIOS_MAX_AC_TIMING_ENTRIES)) {
+						t_mem_id = (u8)((*(u32 *)reg_data & MEM_ID_MASK) >> MEM_ID_SHIFT);
+						if (module_index == t_mem_id) {
+							reg_table->mc_reg_table_entry[num_ranges].mclk_max =
+								(u32)((*(u32 *)reg_data & CLOCK_RANGE_MASK) >> CLOCK_RANGE_SHIFT);
+							for (i = 0, j = 1; i < reg_table->last; i++) {
+								if ((reg_table->mc_reg_address[i].pre_reg_data & LOW_NIBBLE_MASK) == DATA_FROM_TABLE) {
+									reg_table->mc_reg_table_entry[num_ranges].mc_data[i] =
+										(u32)*((u32 *)reg_data + j);
+									j++;
+								} else if ((reg_table->mc_reg_address[i].pre_reg_data & LOW_NIBBLE_MASK) == DATA_EQU_PREV) {
+									reg_table->mc_reg_table_entry[num_ranges].mc_data[i] =
+										reg_table->mc_reg_table_entry[num_ranges].mc_data[i - 1];
+								}
+							}
+							num_ranges++;
+						}
+						reg_data = (ATOM_MEMORY_SETTING_DATA_BLOCK *)
+							((u8 *)reg_data + le16_to_cpu(reg_block->usRegDataBlkSize));
+					}
+					if (*(u32 *)reg_data != END_OF_REG_DATA_BLOCK)
+						return -EINVAL;
+					reg_table->num_entries = num_ranges;
+				} else
+					return -EINVAL;
+				break;
+			default:
+				DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+				return -EINVAL;
+			}
+			break;
+		default:
+			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+			return -EINVAL;
+		}
+		return 0;
+	}
+	return -EINVAL;
+}
+
 void radeon_atom_initialize_bios_scratch_regs(struct drm_device *dev)
 {
 	struct radeon_device *rdev = dev->dev_private;
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index 7e265a58141f..13a130fb3517 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -106,7 +106,7 @@ static int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 		radeon_bo_list_add_object(&p->relocs[i].lobj,
 					  &p->validated);
 	}
-	return radeon_bo_list_validate(&p->validated, p->ring);
+	return radeon_bo_list_validate(&p->ticket, &p->validated, p->ring);
 }
 
 static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority)
@@ -314,15 +314,17 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
  * If error is set than unvalidate buffer, otherwise just free memory
  * used by parsing context.
  **/
-static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error)
+static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error, bool backoff)
 {
 	unsigned i;
 
 	if (!error) {
-		ttm_eu_fence_buffer_objects(&parser->validated,
+		ttm_eu_fence_buffer_objects(&parser->ticket,
+					    &parser->validated,
 					    parser->ib.fence);
-	} else {
-		ttm_eu_backoff_reservation(&parser->validated);
+	} else if (backoff) {
+		ttm_eu_backoff_reservation(&parser->ticket,
+					   &parser->validated);
 	}
 
 	if (parser->relocs != NULL) {
@@ -535,7 +537,7 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	r = radeon_cs_parser_init(&parser, data);
 	if (r) {
 		DRM_ERROR("Failed to initialize parser !\n");
-		radeon_cs_parser_fini(&parser, r);
+		radeon_cs_parser_fini(&parser, r, false);
 		up_read(&rdev->exclusive_lock);
 		r = radeon_cs_handle_lockup(rdev, r);
 		return r;
@@ -544,12 +546,13 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	if (r) {
 		if (r != -ERESTARTSYS)
 			DRM_ERROR("Failed to parse relocation %d!\n", r);
-		radeon_cs_parser_fini(&parser, r);
+		radeon_cs_parser_fini(&parser, r, false);
 		up_read(&rdev->exclusive_lock);
 		r = radeon_cs_handle_lockup(rdev, r);
 		return r;
 	}
 
+	/* XXX pick SD/HD/MVC */
 	if (parser.ring == R600_RING_TYPE_UVD_INDEX)
 		radeon_uvd_note_usage(rdev);
 
@@ -562,7 +565,7 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		goto out;
 	}
 out:
-	radeon_cs_parser_fini(&parser, r);
+	radeon_cs_parser_fini(&parser, r, true);
 	up_read(&rdev->exclusive_lock);
 	r = radeon_cs_handle_lockup(rdev, r);
 	return r;
diff --git a/drivers/gpu/drm/radeon/radeon_cursor.c b/drivers/gpu/drm/radeon/radeon_cursor.c
index b097d5b4ff39..9630e8d95fb4 100644
--- a/drivers/gpu/drm/radeon/radeon_cursor.c
+++ b/drivers/gpu/drm/radeon/radeon_cursor.c
@@ -27,9 +27,6 @@
 #include <drm/radeon_drm.h>
 #include "radeon.h"
 
-#define CURSOR_WIDTH 64
-#define CURSOR_HEIGHT 64
-
 static void radeon_lock_cursor(struct drm_crtc *crtc, bool lock)
 {
 	struct radeon_device *rdev = crtc->dev->dev_private;
@@ -167,7 +164,8 @@ int radeon_crtc_cursor_set(struct drm_crtc *crtc,
 		goto unpin;
 	}
 
-	if ((width > CURSOR_WIDTH) || (height > CURSOR_HEIGHT)) {
+	if ((width > radeon_crtc->max_cursor_width) ||
+	    (height > radeon_crtc->max_cursor_height)) {
 		DRM_ERROR("bad cursor width or height %d x %d\n", width, height);
 		return -EINVAL;
 	}
@@ -233,11 +231,11 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc,
 	DRM_DEBUG("x %d y %d c->x %d c->y %d\n", x, y, crtc->x, crtc->y);
 
 	if (x < 0) {
-		xorigin = min(-x, CURSOR_WIDTH - 1);
+		xorigin = min(-x, radeon_crtc->max_cursor_width - 1);
 		x = 0;
 	}
 	if (y < 0) {
-		yorigin = min(-y, CURSOR_HEIGHT - 1);
+		yorigin = min(-y, radeon_crtc->max_cursor_height - 1);
 		y = 0;
 	}
 
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index b0dc0b6cb4e0..82335e38ec4f 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -95,6 +95,9 @@ static const char radeon_family_name[][16] = {
 	"VERDE",
 	"OLAND",
 	"HAINAN",
+	"BONAIRE",
+	"KAVERI",
+	"KABINI",
 	"LAST",
 };
 
@@ -229,6 +232,94 @@ void radeon_scratch_free(struct radeon_device *rdev, uint32_t reg)
 }
 
 /*
+ * GPU doorbell aperture helpers function.
+ */
+/**
+ * radeon_doorbell_init - Init doorbell driver information.
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Init doorbell driver information (CIK)
+ * Returns 0 on success, error on failure.
+ */
+int radeon_doorbell_init(struct radeon_device *rdev)
+{
+	int i;
+
+	/* doorbell bar mapping */
+	rdev->doorbell.base = pci_resource_start(rdev->pdev, 2);
+	rdev->doorbell.size = pci_resource_len(rdev->pdev, 2);
+
+	/* limit to 4 MB for now */
+	if (rdev->doorbell.size > (4 * 1024 * 1024))
+		rdev->doorbell.size = 4 * 1024 * 1024;
+
+	rdev->doorbell.ptr = ioremap(rdev->doorbell.base, rdev->doorbell.size);
+	if (rdev->doorbell.ptr == NULL) {
+		return -ENOMEM;
+	}
+	DRM_INFO("doorbell mmio base: 0x%08X\n", (uint32_t)rdev->doorbell.base);
+	DRM_INFO("doorbell mmio size: %u\n", (unsigned)rdev->doorbell.size);
+
+	rdev->doorbell.num_pages = rdev->doorbell.size / PAGE_SIZE;
+
+	for (i = 0; i < rdev->doorbell.num_pages; i++) {
+		rdev->doorbell.free[i] = true;
+	}
+	return 0;
+}
+
+/**
+ * radeon_doorbell_fini - Tear down doorbell driver information.
+ *
+ * @rdev: radeon_device pointer
+ *
+ * Tear down doorbell driver information (CIK)
+ */
+void radeon_doorbell_fini(struct radeon_device *rdev)
+{
+	iounmap(rdev->doorbell.ptr);
+	rdev->doorbell.ptr = NULL;
+}
+
+/**
+ * radeon_doorbell_get - Allocate a doorbell page
+ *
+ * @rdev: radeon_device pointer
+ * @doorbell: doorbell page number
+ *
+ * Allocate a doorbell page for use by the driver (all asics).
+ * Returns 0 on success or -EINVAL on failure.
+ */
+int radeon_doorbell_get(struct radeon_device *rdev, u32 *doorbell)
+{
+	int i;
+
+	for (i = 0; i < rdev->doorbell.num_pages; i++) {
+		if (rdev->doorbell.free[i]) {
+			rdev->doorbell.free[i] = false;
+			*doorbell = i;
+			return 0;
+		}
+	}
+	return -EINVAL;
+}
+
+/**
+ * radeon_doorbell_free - Free a doorbell page
+ *
+ * @rdev: radeon_device pointer
+ * @doorbell: doorbell page number
+ *
+ * Free a doorbell page allocated for use by the driver (all asics)
+ */
+void radeon_doorbell_free(struct radeon_device *rdev, u32 doorbell)
+{
+	if (doorbell < rdev->doorbell.num_pages)
+		rdev->doorbell.free[doorbell] = true;
+}
+
+/*
  * radeon_wb_*()
  * Writeback is the the method by which the the GPU updates special pages
  * in memory with the status of certain GPU events (fences, ring pointers,
@@ -1145,8 +1236,13 @@ int radeon_device_init(struct radeon_device *rdev,
 	/* Registers mapping */
 	/* TODO: block userspace mapping of io register */
 	spin_lock_init(&rdev->mmio_idx_lock);
-	rdev->rmmio_base = pci_resource_start(rdev->pdev, 2);
-	rdev->rmmio_size = pci_resource_len(rdev->pdev, 2);
+	if (rdev->family >= CHIP_BONAIRE) {
+		rdev->rmmio_base = pci_resource_start(rdev->pdev, 5);
+		rdev->rmmio_size = pci_resource_len(rdev->pdev, 5);
+	} else {
+		rdev->rmmio_base = pci_resource_start(rdev->pdev, 2);
+		rdev->rmmio_size = pci_resource_len(rdev->pdev, 2);
+	}
 	rdev->rmmio = ioremap(rdev->rmmio_base, rdev->rmmio_size);
 	if (rdev->rmmio == NULL) {
 		return -ENOMEM;
@@ -1154,6 +1250,10 @@ int radeon_device_init(struct radeon_device *rdev,
 	DRM_INFO("register mmio base: 0x%08X\n", (uint32_t)rdev->rmmio_base);
 	DRM_INFO("register mmio size: %u\n", (unsigned)rdev->rmmio_size);
 
+	/* doorbell bar mapping */
+	if (rdev->family >= CHIP_BONAIRE)
+		radeon_doorbell_init(rdev);
+
 	/* io port mapping */
 	for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
 		if (pci_resource_flags(rdev->pdev, i) & IORESOURCE_IO) {
@@ -1231,6 +1331,8 @@ void radeon_device_fini(struct radeon_device *rdev)
 	rdev->rio_mem = NULL;
 	iounmap(rdev->rmmio);
 	rdev->rmmio = NULL;
+	if (rdev->family >= CHIP_BONAIRE)
+		radeon_doorbell_fini(rdev);
 	radeon_debugfs_remove_files(rdev);
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c
index eb18bb7af1cc..c2b67b4e1ac2 100644
--- a/drivers/gpu/drm/radeon/radeon_display.c
+++ b/drivers/gpu/drm/radeon/radeon_display.c
@@ -153,7 +153,13 @@ static void dce5_crtc_load_lut(struct drm_crtc *crtc)
 		NI_OUTPUT_CSC_OVL_MODE(NI_OUTPUT_CSC_BYPASS)));
 	/* XXX match this to the depth of the crtc fmt block, move to modeset? */
 	WREG32(0x6940 + radeon_crtc->crtc_offset, 0);
-
+	if (ASIC_IS_DCE8(rdev)) {
+		/* XXX this only needs to be programmed once per crtc at startup,
+		 * not sure where the best place for it is
+		 */
+		WREG32(CIK_ALPHA_CONTROL + radeon_crtc->crtc_offset,
+		       CIK_CURSOR_ALPHA_BLND_ENA);
+	}
 }
 
 static void legacy_crtc_load_lut(struct drm_crtc *crtc)
@@ -512,6 +518,14 @@ static void radeon_crtc_init(struct drm_device *dev, int index)
 	radeon_crtc->crtc_id = index;
 	rdev->mode_info.crtcs[index] = radeon_crtc;
 
+	if (rdev->family >= CHIP_BONAIRE) {
+		radeon_crtc->max_cursor_width = CIK_CURSOR_WIDTH;
+		radeon_crtc->max_cursor_height = CIK_CURSOR_HEIGHT;
+	} else {
+		radeon_crtc->max_cursor_width = CURSOR_WIDTH;
+		radeon_crtc->max_cursor_height = CURSOR_HEIGHT;
+	}
+
 #if 0
 	radeon_crtc->mode_set.crtc = &radeon_crtc->base;
 	radeon_crtc->mode_set.connectors = (struct drm_connector **)(radeon_crtc + 1);
@@ -530,7 +544,7 @@ static void radeon_crtc_init(struct drm_device *dev, int index)
 		radeon_legacy_init_crtc(dev, radeon_crtc);
 }
 
-static const char *encoder_names[37] = {
+static const char *encoder_names[38] = {
 	"NONE",
 	"INTERNAL_LVDS",
 	"INTERNAL_TMDS1",
@@ -567,7 +581,8 @@ static const char *encoder_names[37] = {
 	"INTERNAL_UNIPHY2",
 	"NUTMEG",
 	"TRAVIS",
-	"INTERNAL_VCE"
+	"INTERNAL_VCE",
+	"INTERNAL_UNIPHY3",
 };
 
 static const char *hpd_names[6] = {
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index 094e7e5ea39e..29876b1be8ec 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -74,9 +74,10 @@
  *   2.31.0 - Add fastfb support for rs690
  *   2.32.0 - new info request for rings working
  *   2.33.0 - Add SI tiling mode array query
+ *   2.34.0 - Add CIK tiling mode array query
  */
 #define KMS_DRIVER_MAJOR	2
-#define KMS_DRIVER_MINOR	33
+#define KMS_DRIVER_MINOR	34
 #define KMS_DRIVER_PATCHLEVEL	0
 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags);
 int radeon_driver_unload_kms(struct drm_device *dev);
@@ -127,6 +128,7 @@ struct drm_gem_object *radeon_gem_prime_import_sg_table(struct drm_device *dev,
 							size_t size,
 							struct sg_table *sg);
 int radeon_gem_prime_pin(struct drm_gem_object *obj);
+void radeon_gem_prime_unpin(struct drm_gem_object *obj);
 void *radeon_gem_prime_vmap(struct drm_gem_object *obj);
 void radeon_gem_prime_vunmap(struct drm_gem_object *obj, void *vaddr);
 extern long radeon_kms_compat_ioctl(struct file *filp, unsigned int cmd,
@@ -164,6 +166,8 @@ int radeon_pcie_gen2 = -1;
 int radeon_msi = -1;
 int radeon_lockup_timeout = 10000;
 int radeon_fastfb = 0;
+int radeon_dpm = -1;
+int radeon_aspm = -1;
 
 MODULE_PARM_DESC(no_wb, "Disable AGP writeback for scratch registers");
 module_param_named(no_wb, radeon_no_wb, int, 0444);
@@ -219,6 +223,12 @@ module_param_named(lockup_timeout, radeon_lockup_timeout, int, 0444);
 MODULE_PARM_DESC(fastfb, "Direct FB access for IGP chips (0 = disable, 1 = enable)");
 module_param_named(fastfb, radeon_fastfb, int, 0444);
 
+MODULE_PARM_DESC(dpm, "DPM support (1 = enable, 0 = disable, -1 = auto)");
+module_param_named(dpm, radeon_dpm, int, 0444);
+
+MODULE_PARM_DESC(aspm, "ASPM support (1 = enable, 0 = disable, -1 = auto)");
+module_param_named(aspm, radeon_aspm, int, 0444);
+
 static struct pci_device_id pciidlist[] = {
 	radeon_PCI_IDS
 };
@@ -422,6 +432,7 @@ static struct drm_driver kms_driver = {
 	.gem_prime_export = drm_gem_prime_export,
 	.gem_prime_import = drm_gem_prime_import,
 	.gem_prime_pin = radeon_gem_prime_pin,
+	.gem_prime_unpin = radeon_gem_prime_unpin,
 	.gem_prime_get_sg_table = radeon_gem_prime_get_sg_table,
 	.gem_prime_import_sg_table = radeon_gem_prime_import_sg_table,
 	.gem_prime_vmap = radeon_gem_prime_vmap,
diff --git a/drivers/gpu/drm/radeon/radeon_family.h b/drivers/gpu/drm/radeon/radeon_family.h
index 36e9803b077d..3c8289083f9d 100644
--- a/drivers/gpu/drm/radeon/radeon_family.h
+++ b/drivers/gpu/drm/radeon/radeon_family.h
@@ -93,6 +93,9 @@ enum radeon_family {
 	CHIP_VERDE,
 	CHIP_OLAND,
 	CHIP_HAINAN,
+	CHIP_BONAIRE,
+	CHIP_KAVERI,
+	CHIP_KABINI,
 	CHIP_LAST,
 };
 
diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c
index ddb8f8e04eb5..7ddb0efe2408 100644
--- a/drivers/gpu/drm/radeon/radeon_fence.c
+++ b/drivers/gpu/drm/radeon/radeon_fence.c
@@ -782,7 +782,7 @@ int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring)
 
 		} else {
 			/* put fence directly behind firmware */
-			index = ALIGN(rdev->uvd_fw->size, 8);
+			index = ALIGN(rdev->uvd.fw_size, 8);
 			rdev->fence_drv[ring].cpu_addr = rdev->uvd.cpu_addr + index;
 			rdev->fence_drv[ring].gpu_addr = rdev->uvd.gpu_addr + index;
 		}
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index 43ec4a401f07..d9d31a383276 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -466,7 +466,8 @@ int radeon_vm_manager_init(struct radeon_device *rdev)
 		size += rdev->vm_manager.max_pfn * 8;
 		size *= 2;
 		r = radeon_sa_bo_manager_init(rdev, &rdev->vm_manager.sa_manager,
-					      RADEON_GPU_PAGE_ALIGN(size),
+					      RADEON_VM_PTB_ALIGN(size),
+					      RADEON_VM_PTB_ALIGN_SIZE,
 					      RADEON_GEM_DOMAIN_VRAM);
 		if (r) {
 			dev_err(rdev->dev, "failed to allocate vm bo (%dKB)\n",
@@ -620,10 +621,10 @@ int radeon_vm_alloc_pt(struct radeon_device *rdev, struct radeon_vm *vm)
 	}
 
 retry:
-	pd_size = RADEON_GPU_PAGE_ALIGN(radeon_vm_directory_size(rdev));
+	pd_size = RADEON_VM_PTB_ALIGN(radeon_vm_directory_size(rdev));
 	r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager,
 			     &vm->page_directory, pd_size,
-			     RADEON_GPU_PAGE_SIZE, false);
+			     RADEON_VM_PTB_ALIGN_SIZE, false);
 	if (r == -ENOMEM) {
 		r = radeon_vm_evict(rdev, vm);
 		if (r)
@@ -952,8 +953,8 @@ static int radeon_vm_update_pdes(struct radeon_device *rdev,
 retry:
 		r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager,
 				     &vm->page_tables[pt_idx],
-				     RADEON_VM_PTE_COUNT * 8,
-				     RADEON_GPU_PAGE_SIZE, false);
+				     RADEON_VM_PTB_ALIGN(RADEON_VM_PTE_COUNT * 8),
+				     RADEON_VM_PTB_ALIGN_SIZE, false);
 
 		if (r == -ENOMEM) {
 			r = radeon_vm_evict(rdev, vm);
diff --git a/drivers/gpu/drm/radeon/radeon_irq_kms.c b/drivers/gpu/drm/radeon/radeon_irq_kms.c
index 5a99d433fc35..081886b0642d 100644
--- a/drivers/gpu/drm/radeon/radeon_irq_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_irq_kms.c
@@ -82,6 +82,23 @@ static void radeon_hotplug_work_func(struct work_struct *work)
 }
 
 /**
+ * radeon_irq_reset_work_func - execute gpu reset
+ *
+ * @work: work struct
+ *
+ * Execute scheduled gpu reset (cayman+).
+ * This function is called when the irq handler
+ * thinks we need a gpu reset.
+ */
+static void radeon_irq_reset_work_func(struct work_struct *work)
+{
+	struct radeon_device *rdev = container_of(work, struct radeon_device,
+						  reset_work);
+
+	radeon_gpu_reset(rdev);
+}
+
+/**
  * radeon_driver_irq_preinstall_kms - drm irq preinstall callback
  *
  * @dev: drm dev pointer
@@ -99,6 +116,7 @@ void radeon_driver_irq_preinstall_kms(struct drm_device *dev)
 	/* Disable *all* interrupts */
 	for (i = 0; i < RADEON_NUM_RINGS; i++)
 		atomic_set(&rdev->irq.ring_int[i], 0);
+	rdev->irq.dpm_thermal = false;
 	for (i = 0; i < RADEON_MAX_HPD_PINS; i++)
 		rdev->irq.hpd[i] = false;
 	for (i = 0; i < RADEON_MAX_CRTCS; i++) {
@@ -146,6 +164,7 @@ void radeon_driver_irq_uninstall_kms(struct drm_device *dev)
 	/* Disable *all* interrupts */
 	for (i = 0; i < RADEON_NUM_RINGS; i++)
 		atomic_set(&rdev->irq.ring_int[i], 0);
+	rdev->irq.dpm_thermal = false;
 	for (i = 0; i < RADEON_MAX_HPD_PINS; i++)
 		rdev->irq.hpd[i] = false;
 	for (i = 0; i < RADEON_MAX_CRTCS; i++) {
@@ -241,9 +260,6 @@ int radeon_irq_kms_init(struct radeon_device *rdev)
 {
 	int r = 0;
 
-	INIT_WORK(&rdev->hotplug_work, radeon_hotplug_work_func);
-	INIT_WORK(&rdev->audio_work, r600_audio_update_hdmi);
-
 	spin_lock_init(&rdev->irq.lock);
 	r = drm_vblank_init(rdev->ddev, rdev->num_crtc);
 	if (r) {
@@ -265,6 +281,11 @@ int radeon_irq_kms_init(struct radeon_device *rdev)
 		rdev->irq.installed = false;
 		return r;
 	}
+
+	INIT_WORK(&rdev->hotplug_work, radeon_hotplug_work_func);
+	INIT_WORK(&rdev->audio_work, r600_audio_update_hdmi);
+	INIT_WORK(&rdev->reset_work, radeon_irq_reset_work_func);
+
 	DRM_INFO("radeon: irq initialized.\n");
 	return 0;
 }
@@ -284,8 +305,8 @@ void radeon_irq_kms_fini(struct radeon_device *rdev)
 		rdev->irq.installed = false;
 		if (rdev->msi_enabled)
 			pci_disable_msi(rdev->pdev);
+		flush_work(&rdev->hotplug_work);
 	}
-	flush_work(&rdev->hotplug_work);
 }
 
 /**
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 4f2d4f4c1dab..49ff3d1a6102 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -229,7 +229,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		*value = rdev->accel_working;
 		break;
 	case RADEON_INFO_TILING_CONFIG:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			*value = rdev->config.cik.tile_config;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.tile_config;
 		else if (rdev->family >= CHIP_CAYMAN)
 			*value = rdev->config.cayman.tile_config;
@@ -281,7 +283,10 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 			*value = rdev->clock.spll.reference_freq * 10;
 		break;
 	case RADEON_INFO_NUM_BACKENDS:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			*value = rdev->config.cik.max_backends_per_se *
+				rdev->config.cik.max_shader_engines;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.max_backends_per_se *
 				rdev->config.si.max_shader_engines;
 		else if (rdev->family >= CHIP_CAYMAN)
@@ -298,7 +303,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		}
 		break;
 	case RADEON_INFO_NUM_TILE_PIPES:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			*value = rdev->config.cik.max_tile_pipes;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.max_tile_pipes;
 		else if (rdev->family >= CHIP_CAYMAN)
 			*value = rdev->config.cayman.max_tile_pipes;
@@ -316,7 +323,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		*value = 1;
 		break;
 	case RADEON_INFO_BACKEND_MAP:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			return -EINVAL;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.backend_map;
 		else if (rdev->family >= CHIP_CAYMAN)
 			*value = rdev->config.cayman.backend_map;
@@ -343,7 +352,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		*value = RADEON_IB_VM_MAX_SIZE;
 		break;
 	case RADEON_INFO_MAX_PIPES:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			*value = rdev->config.cik.max_cu_per_sh;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.max_cu_per_sh;
 		else if (rdev->family >= CHIP_CAYMAN)
 			*value = rdev->config.cayman.max_pipes_per_simd;
@@ -367,7 +378,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		value64 = radeon_get_gpu_clock_counter(rdev);
 		break;
 	case RADEON_INFO_MAX_SE:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			*value = rdev->config.cik.max_shader_engines;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.max_shader_engines;
 		else if (rdev->family >= CHIP_CAYMAN)
 			*value = rdev->config.cayman.max_shader_engines;
@@ -377,7 +390,9 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 			*value = 1;
 		break;
 	case RADEON_INFO_MAX_SH_PER_SE:
-		if (rdev->family >= CHIP_TAHITI)
+		if (rdev->family >= CHIP_BONAIRE)
+			*value = rdev->config.cik.max_sh_per_se;
+		else if (rdev->family >= CHIP_TAHITI)
 			*value = rdev->config.si.max_sh_per_se;
 		else
 			return -EINVAL;
@@ -407,12 +422,16 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		}
 		break;
 	case RADEON_INFO_SI_TILE_MODE_ARRAY:
-		if (rdev->family < CHIP_TAHITI) {
-			DRM_DEBUG_KMS("tile mode array is si only!\n");
+		if (rdev->family >= CHIP_BONAIRE) {
+			value = rdev->config.cik.tile_mode_array;
+			value_size = sizeof(uint32_t)*32;
+		} else if (rdev->family >= CHIP_TAHITI) {
+			value = rdev->config.si.tile_mode_array;
+			value_size = sizeof(uint32_t)*32;
+		} else {
+			DRM_DEBUG_KMS("tile mode array is si+ only!\n");
 			return -EINVAL;
 		}
-		value = rdev->config.si.tile_mode_array;
-		value_size = sizeof(uint32_t)*32;
 		break;
 	default:
 		DRM_DEBUG_KMS("Invalid request %d\n", info->request);
diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h
index 69ad4fe224c1..8296632a4235 100644
--- a/drivers/gpu/drm/radeon/radeon_mode.h
+++ b/drivers/gpu/drm/radeon/radeon_mode.h
@@ -307,6 +307,8 @@ struct radeon_crtc {
 	uint64_t cursor_addr;
 	int cursor_width;
 	int cursor_height;
+	int max_cursor_width;
+	int max_cursor_height;
 	uint32_t legacy_display_base_addr;
 	uint32_t legacy_cursor_offset;
 	enum radeon_rmx_type rmx_type;
@@ -329,6 +331,11 @@ struct radeon_crtc {
 	u32 pll_flags;
 	struct drm_encoder *encoder;
 	struct drm_connector *connector;
+	/* for dpm */
+	u32 line_time;
+	u32 wm_low;
+	u32 wm_high;
+	struct drm_display_mode hw_mode;
 };
 
 struct radeon_encoder_primary_dac {
@@ -512,12 +519,99 @@ struct atom_clock_dividers {
 	bool enable_dithen;
 	u32 vco_mode;
 	u32 real_clock;
+	/* added for CI */
+	u32 post_divider;
+	u32 flags;
+};
+
+struct atom_mpll_param {
+	union {
+		struct {
+#ifdef __BIG_ENDIAN
+			u32 reserved : 8;
+			u32 clkfrac : 12;
+			u32 clkf : 12;
+#else
+			u32 clkf : 12;
+			u32 clkfrac : 12;
+			u32 reserved : 8;
+#endif
+		};
+		u32 fb_div;
+	};
+	u32 post_div;
+	u32 bwcntl;
+	u32 dll_speed;
+	u32 vco_mode;
+	u32 yclk_sel;
+	u32 qdr;
+	u32 half_rate;
+};
+
+#define MEM_TYPE_GDDR5  0x50
+#define MEM_TYPE_GDDR4  0x40
+#define MEM_TYPE_GDDR3  0x30
+#define MEM_TYPE_DDR2   0x20
+#define MEM_TYPE_GDDR1  0x10
+#define MEM_TYPE_DDR3   0xb0
+#define MEM_TYPE_MASK   0xf0
+
+struct atom_memory_info {
+	u8 mem_vendor;
+	u8 mem_type;
+};
+
+#define MAX_AC_TIMING_ENTRIES 16
+
+struct atom_memory_clock_range_table
+{
+	u8 num_entries;
+	u8 rsv[3];
+	u32 mclk[MAX_AC_TIMING_ENTRIES];
+};
+
+#define VBIOS_MC_REGISTER_ARRAY_SIZE 32
+#define VBIOS_MAX_AC_TIMING_ENTRIES 20
+
+struct atom_mc_reg_entry {
+	u32 mclk_max;
+	u32 mc_data[VBIOS_MC_REGISTER_ARRAY_SIZE];
+};
+
+struct atom_mc_register_address {
+	u16 s1;
+	u8 pre_reg_data;
+};
+
+struct atom_mc_reg_table {
+	u8 last;
+	u8 num_entries;
+	struct atom_mc_reg_entry mc_reg_table_entry[VBIOS_MAX_AC_TIMING_ENTRIES];
+	struct atom_mc_register_address mc_reg_address[VBIOS_MC_REGISTER_ARRAY_SIZE];
+};
+
+#define MAX_VOLTAGE_ENTRIES 32
+
+struct atom_voltage_table_entry
+{
+	u16 value;
+	u32 smio_low;
+};
+
+struct atom_voltage_table
+{
+	u32 count;
+	u32 mask_low;
+	u32 phase_delay;
+	struct atom_voltage_table_entry entries[MAX_VOLTAGE_ENTRIES];
 };
 
 extern enum radeon_tv_std
 radeon_combios_get_tv_info(struct radeon_device *rdev);
 extern enum radeon_tv_std
 radeon_atombios_get_tv_info(struct radeon_device *rdev);
+extern void radeon_atombios_get_default_voltages(struct radeon_device *rdev,
+						 u16 *vddc, u16 *vddci, u16 *mvdd);
 
 extern struct drm_connector *
 radeon_get_connector_for_encoder(struct drm_encoder *encoder);
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index 1424ccde2377..2020bf4a3830 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -322,8 +322,8 @@ int radeon_bo_init(struct radeon_device *rdev)
 {
 	/* Add an MTRR for the VRAM */
 	if (!rdev->fastfb_working) {
-		rdev->mc.vram_mtrr = mtrr_add(rdev->mc.aper_base, rdev->mc.aper_size,
-			MTRR_TYPE_WRCOMB, 1);
+		rdev->mc.vram_mtrr = arch_phys_wc_add(rdev->mc.aper_base,
+						      rdev->mc.aper_size);
 	}
 	DRM_INFO("Detected VRAM RAM=%lluM, BAR=%lluM\n",
 		rdev->mc.mc_vram_size >> 20,
@@ -336,6 +336,7 @@ int radeon_bo_init(struct radeon_device *rdev)
 void radeon_bo_fini(struct radeon_device *rdev)
 {
 	radeon_ttm_fini(rdev);
+	arch_phys_wc_del(rdev->mc.vram_mtrr);
 }
 
 void radeon_bo_list_add_object(struct radeon_bo_list *lobj,
@@ -348,14 +349,15 @@ void radeon_bo_list_add_object(struct radeon_bo_list *lobj,
 	}
 }
 
-int radeon_bo_list_validate(struct list_head *head, int ring)
+int radeon_bo_list_validate(struct ww_acquire_ctx *ticket,
+			    struct list_head *head, int ring)
 {
 	struct radeon_bo_list *lobj;
 	struct radeon_bo *bo;
 	u32 domain;
 	int r;
 
-	r = ttm_eu_reserve_buffers(head);
+	r = ttm_eu_reserve_buffers(ticket, head);
 	if (unlikely(r != 0)) {
 		return r;
 	}
@@ -375,6 +377,7 @@ int radeon_bo_list_validate(struct list_head *head, int ring)
 					domain = lobj->alt_domain;
 					goto retry;
 				}
+				ttm_eu_backoff_reservation(ticket, head);
 				return r;
 			}
 		}
@@ -398,7 +401,7 @@ int radeon_bo_get_surface_reg(struct radeon_bo *bo)
 	int steal;
 	int i;
 
-	BUG_ON(!radeon_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->tbo.resv->lock.base);
 
 	if (!bo->tiling_flags)
 		return 0;
@@ -524,7 +527,8 @@ void radeon_bo_get_tiling_flags(struct radeon_bo *bo,
 				uint32_t *tiling_flags,
 				uint32_t *pitch)
 {
-	BUG_ON(!radeon_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->tbo.resv->lock.base);
+
 	if (tiling_flags)
 		*tiling_flags = bo->tiling_flags;
 	if (pitch)
@@ -534,7 +538,8 @@ void radeon_bo_get_tiling_flags(struct radeon_bo *bo,
 int radeon_bo_check_tiling(struct radeon_bo *bo, bool has_moved,
 				bool force_drop)
 {
-	BUG_ON(!radeon_bo_is_reserved(bo) && !force_drop);
+	if (!force_drop)
+		lockdep_assert_held(&bo->tbo.resv->lock.base);
 
 	if (!(bo->tiling_flags & RADEON_TILING_SURFACE))
 		return 0;
@@ -617,26 +622,3 @@ int radeon_bo_wait(struct radeon_bo *bo, u32 *mem_type, bool no_wait)
 	ttm_bo_unreserve(&bo->tbo);
 	return r;
 }
-
-
-/**
- * radeon_bo_reserve - reserve bo
- * @bo:		bo structure
- * @no_intr:	don't return -ERESTARTSYS on pending signal
- *
- * Returns:
- * -ERESTARTSYS: A wait for the buffer to become unreserved was interrupted by
- * a signal. Release all buffer reservations and return to user-space.
- */
-int radeon_bo_reserve(struct radeon_bo *bo, bool no_intr)
-{
-	int r;
-
-	r = ttm_bo_reserve(&bo->tbo, !no_intr, false, false, 0);
-	if (unlikely(r != 0)) {
-		if (r != -ERESTARTSYS)
-			dev_err(bo->rdev->dev, "%p reserve failed\n", bo);
-		return r;
-	}
-	return 0;
-}
diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h
index e2cb80a96b51..49c82c480013 100644
--- a/drivers/gpu/drm/radeon/radeon_object.h
+++ b/drivers/gpu/drm/radeon/radeon_object.h
@@ -52,7 +52,27 @@ static inline unsigned radeon_mem_type_to_domain(u32 mem_type)
 	return 0;
 }
 
-int radeon_bo_reserve(struct radeon_bo *bo, bool no_intr);
+/**
+ * radeon_bo_reserve - reserve bo
+ * @bo:		bo structure
+ * @no_intr:	don't return -ERESTARTSYS on pending signal
+ *
+ * Returns:
+ * -ERESTARTSYS: A wait for the buffer to become unreserved was interrupted by
+ * a signal. Release all buffer reservations and return to user-space.
+ */
+static inline int radeon_bo_reserve(struct radeon_bo *bo, bool no_intr)
+{
+	int r;
+
+	r = ttm_bo_reserve(&bo->tbo, !no_intr, false, false, 0);
+	if (unlikely(r != 0)) {
+		if (r != -ERESTARTSYS)
+			dev_err(bo->rdev->dev, "%p reserve failed\n", bo);
+		return r;
+	}
+	return 0;
+}
 
 static inline void radeon_bo_unreserve(struct radeon_bo *bo)
 {
@@ -78,11 +98,6 @@ static inline unsigned long radeon_bo_size(struct radeon_bo *bo)
 	return bo->tbo.num_pages << PAGE_SHIFT;
 }
 
-static inline bool radeon_bo_is_reserved(struct radeon_bo *bo)
-{
-	return ttm_bo_is_reserved(&bo->tbo);
-}
-
 static inline unsigned radeon_bo_ngpu_pages(struct radeon_bo *bo)
 {
 	return (bo->tbo.num_pages << PAGE_SHIFT) / RADEON_GPU_PAGE_SIZE;
@@ -128,7 +143,8 @@ extern int radeon_bo_init(struct radeon_device *rdev);
 extern void radeon_bo_fini(struct radeon_device *rdev);
 extern void radeon_bo_list_add_object(struct radeon_bo_list *lobj,
 				struct list_head *head);
-extern int radeon_bo_list_validate(struct list_head *head, int ring);
+extern int radeon_bo_list_validate(struct ww_acquire_ctx *ticket,
+				   struct list_head *head, int ring);
 extern int radeon_bo_fbdev_mmap(struct radeon_bo *bo,
 				struct vm_area_struct *vma);
 extern int radeon_bo_set_tiling_flags(struct radeon_bo *bo,
@@ -158,7 +174,7 @@ static inline void * radeon_sa_bo_cpu_addr(struct radeon_sa_bo *sa_bo)
 
 extern int radeon_sa_bo_manager_init(struct radeon_device *rdev,
 				     struct radeon_sa_manager *sa_manager,
-				     unsigned size, u32 domain);
+				     unsigned size, u32 align, u32 domain);
 extern void radeon_sa_bo_manager_fini(struct radeon_device *rdev,
 				      struct radeon_sa_manager *sa_manager);
 extern int radeon_sa_bo_manager_start(struct radeon_device *rdev,
diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c
index 788c64cb4b47..f374c467aaca 100644
--- a/drivers/gpu/drm/radeon/radeon_pm.c
+++ b/drivers/gpu/drm/radeon/radeon_pm.c
@@ -388,7 +388,8 @@ static ssize_t radeon_get_pm_method(struct device *dev,
 	int pm = rdev->pm.pm_method;
 
 	return snprintf(buf, PAGE_SIZE, "%s\n",
-			(pm == PM_METHOD_DYNPM) ? "dynpm" : "profile");
+			(pm == PM_METHOD_DYNPM) ? "dynpm" :
+			(pm == PM_METHOD_PROFILE) ? "profile" : "dpm");
 }
 
 static ssize_t radeon_set_pm_method(struct device *dev,
@@ -399,6 +400,11 @@ static ssize_t radeon_set_pm_method(struct device *dev,
 	struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev));
 	struct radeon_device *rdev = ddev->dev_private;
 
+	/* we don't support the legacy modes with dpm */
+	if (rdev->pm.pm_method == PM_METHOD_DPM) {
+		count = -EINVAL;
+		goto fail;
+	}
 
 	if (strncmp("dynpm", buf, strlen("dynpm")) == 0) {
 		mutex_lock(&rdev->pm.mutex);
@@ -423,8 +429,96 @@ fail:
 	return count;
 }
 
+static ssize_t radeon_get_dpm_state(struct device *dev,
+				    struct device_attribute *attr,
+				    char *buf)
+{
+	struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev));
+	struct radeon_device *rdev = ddev->dev_private;
+	enum radeon_pm_state_type pm = rdev->pm.dpm.user_state;
+
+	return snprintf(buf, PAGE_SIZE, "%s\n",
+			(pm == POWER_STATE_TYPE_BATTERY) ? "battery" :
+			(pm == POWER_STATE_TYPE_BALANCED) ? "balanced" : "performance");
+}
+
+static ssize_t radeon_set_dpm_state(struct device *dev,
+				    struct device_attribute *attr,
+				    const char *buf,
+				    size_t count)
+{
+	struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev));
+	struct radeon_device *rdev = ddev->dev_private;
+
+	mutex_lock(&rdev->pm.mutex);
+	if (strncmp("battery", buf, strlen("battery")) == 0)
+		rdev->pm.dpm.user_state = POWER_STATE_TYPE_BATTERY;
+	else if (strncmp("balanced", buf, strlen("balanced")) == 0)
+		rdev->pm.dpm.user_state = POWER_STATE_TYPE_BALANCED;
+	else if (strncmp("performance", buf, strlen("performance")) == 0)
+		rdev->pm.dpm.user_state = POWER_STATE_TYPE_PERFORMANCE;
+	else {
+		mutex_unlock(&rdev->pm.mutex);
+		count = -EINVAL;
+		goto fail;
+	}
+	mutex_unlock(&rdev->pm.mutex);
+	radeon_pm_compute_clocks(rdev);
+fail:
+	return count;
+}
+
+static ssize_t radeon_get_dpm_forced_performance_level(struct device *dev,
+						       struct device_attribute *attr,
+						       char *buf)
+{
+	struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev));
+	struct radeon_device *rdev = ddev->dev_private;
+	enum radeon_dpm_forced_level level = rdev->pm.dpm.forced_level;
+
+	return snprintf(buf, PAGE_SIZE, "%s\n",
+			(level == RADEON_DPM_FORCED_LEVEL_AUTO) ? "auto" :
+			(level == RADEON_DPM_FORCED_LEVEL_LOW) ? "low" : "high");
+}
+
+static ssize_t radeon_set_dpm_forced_performance_level(struct device *dev,
+						       struct device_attribute *attr,
+						       const char *buf,
+						       size_t count)
+{
+	struct drm_device *ddev = pci_get_drvdata(to_pci_dev(dev));
+	struct radeon_device *rdev = ddev->dev_private;
+	enum radeon_dpm_forced_level level;
+	int ret = 0;
+
+	mutex_lock(&rdev->pm.mutex);
+	if (strncmp("low", buf, strlen("low")) == 0) {
+		level = RADEON_DPM_FORCED_LEVEL_LOW;
+	} else if (strncmp("high", buf, strlen("high")) == 0) {
+		level = RADEON_DPM_FORCED_LEVEL_HIGH;
+	} else if (strncmp("auto", buf, strlen("auto")) == 0) {
+		level = RADEON_DPM_FORCED_LEVEL_AUTO;
+	} else {
+		mutex_unlock(&rdev->pm.mutex);
+		count = -EINVAL;
+		goto fail;
+	}
+	if (rdev->asic->dpm.force_performance_level) {
+		ret = radeon_dpm_force_performance_level(rdev, level);
+		if (ret)
+			count = -EINVAL;
+	}
+	mutex_unlock(&rdev->pm.mutex);
+fail:
+	return count;
+}
+
 static DEVICE_ATTR(power_profile, S_IRUGO | S_IWUSR, radeon_get_pm_profile, radeon_set_pm_profile);
 static DEVICE_ATTR(power_method, S_IRUGO | S_IWUSR, radeon_get_pm_method, radeon_set_pm_method);
+static DEVICE_ATTR(power_dpm_state, S_IRUGO | S_IWUSR, radeon_get_dpm_state, radeon_set_dpm_state);
+static DEVICE_ATTR(power_dpm_force_performance_level, S_IRUGO | S_IWUSR,
+		   radeon_get_dpm_forced_performance_level,
+		   radeon_set_dpm_forced_performance_level);
 
 static ssize_t radeon_hwmon_show_temp(struct device *dev,
 				      struct device_attribute *attr,
@@ -434,27 +528,10 @@ static ssize_t radeon_hwmon_show_temp(struct device *dev,
 	struct radeon_device *rdev = ddev->dev_private;
 	int temp;
 
-	switch (rdev->pm.int_thermal_type) {
-	case THERMAL_TYPE_RV6XX:
-		temp = rv6xx_get_temp(rdev);
-		break;
-	case THERMAL_TYPE_RV770:
-		temp = rv770_get_temp(rdev);
-		break;
-	case THERMAL_TYPE_EVERGREEN:
-	case THERMAL_TYPE_NI:
-		temp = evergreen_get_temp(rdev);
-		break;
-	case THERMAL_TYPE_SUMO:
-		temp = sumo_get_temp(rdev);
-		break;
-	case THERMAL_TYPE_SI:
-		temp = si_get_temp(rdev);
-		break;
-	default:
+	if (rdev->asic->pm.get_temperature)
+		temp = radeon_get_temperature(rdev);
+	else
 		temp = 0;
-		break;
-	}
 
 	return snprintf(buf, PAGE_SIZE, "%d\n", temp);
 }
@@ -492,8 +569,7 @@ static int radeon_hwmon_init(struct radeon_device *rdev)
 	case THERMAL_TYPE_NI:
 	case THERMAL_TYPE_SUMO:
 	case THERMAL_TYPE_SI:
-		/* No support for TN yet */
-		if (rdev->family == CHIP_ARUBA)
+		if (rdev->asic->pm.get_temperature == NULL)
 			return err;
 		rdev->pm.int_hwmon_dev = hwmon_device_register(rdev->dev);
 		if (IS_ERR(rdev->pm.int_hwmon_dev)) {
@@ -526,7 +602,289 @@ static void radeon_hwmon_fini(struct radeon_device *rdev)
 	}
 }
 
-void radeon_pm_suspend(struct radeon_device *rdev)
+static void radeon_dpm_thermal_work_handler(struct work_struct *work)
+{
+	struct radeon_device *rdev =
+		container_of(work, struct radeon_device,
+			     pm.dpm.thermal.work);
+	/* switch to the thermal state */
+	enum radeon_pm_state_type dpm_state = POWER_STATE_TYPE_INTERNAL_THERMAL;
+
+	if (!rdev->pm.dpm_enabled)
+		return;
+
+	if (rdev->asic->pm.get_temperature) {
+		int temp = radeon_get_temperature(rdev);
+
+		if (temp < rdev->pm.dpm.thermal.min_temp)
+			/* switch back the user state */
+			dpm_state = rdev->pm.dpm.user_state;
+	} else {
+		if (rdev->pm.dpm.thermal.high_to_low)
+			/* switch back the user state */
+			dpm_state = rdev->pm.dpm.user_state;
+	}
+	radeon_dpm_enable_power_state(rdev, dpm_state);
+}
+
+static struct radeon_ps *radeon_dpm_pick_power_state(struct radeon_device *rdev,
+						     enum radeon_pm_state_type dpm_state)
+{
+	int i;
+	struct radeon_ps *ps;
+	u32 ui_class;
+	bool single_display = (rdev->pm.dpm.new_active_crtc_count < 2) ?
+		true : false;
+
+	/* check if the vblank period is too short to adjust the mclk */
+	if (single_display && rdev->asic->dpm.vblank_too_short) {
+		if (radeon_dpm_vblank_too_short(rdev))
+			single_display = false;
+	}
+
+	/* certain older asics have a separare 3D performance state,
+	 * so try that first if the user selected performance
+	 */
+	if (dpm_state == POWER_STATE_TYPE_PERFORMANCE)
+		dpm_state = POWER_STATE_TYPE_INTERNAL_3DPERF;
+	/* balanced states don't exist at the moment */
+	if (dpm_state == POWER_STATE_TYPE_BALANCED)
+		dpm_state = POWER_STATE_TYPE_PERFORMANCE;
+
+restart_search:
+	/* Pick the best power state based on current conditions */
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		ps = &rdev->pm.dpm.ps[i];
+		ui_class = ps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK;
+		switch (dpm_state) {
+		/* user states */
+		case POWER_STATE_TYPE_BATTERY:
+			if (ui_class == ATOM_PPLIB_CLASSIFICATION_UI_BATTERY) {
+				if (ps->caps & ATOM_PPLIB_SINGLE_DISPLAY_ONLY) {
+					if (single_display)
+						return ps;
+				} else
+					return ps;
+			}
+			break;
+		case POWER_STATE_TYPE_BALANCED:
+			if (ui_class == ATOM_PPLIB_CLASSIFICATION_UI_BALANCED) {
+				if (ps->caps & ATOM_PPLIB_SINGLE_DISPLAY_ONLY) {
+					if (single_display)
+						return ps;
+				} else
+					return ps;
+			}
+			break;
+		case POWER_STATE_TYPE_PERFORMANCE:
+			if (ui_class == ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE) {
+				if (ps->caps & ATOM_PPLIB_SINGLE_DISPLAY_ONLY) {
+					if (single_display)
+						return ps;
+				} else
+					return ps;
+			}
+			break;
+		/* internal states */
+		case POWER_STATE_TYPE_INTERNAL_UVD:
+			return rdev->pm.dpm.uvd_ps;
+		case POWER_STATE_TYPE_INTERNAL_UVD_SD:
+			if (ps->class & ATOM_PPLIB_CLASSIFICATION_SDSTATE)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_UVD_HD:
+			if (ps->class & ATOM_PPLIB_CLASSIFICATION_HDSTATE)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_UVD_HD2:
+			if (ps->class & ATOM_PPLIB_CLASSIFICATION_HD2STATE)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_UVD_MVC:
+			if (ps->class2 & ATOM_PPLIB_CLASSIFICATION2_MVC)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_BOOT:
+			return rdev->pm.dpm.boot_ps;
+		case POWER_STATE_TYPE_INTERNAL_THERMAL:
+			if (ps->class & ATOM_PPLIB_CLASSIFICATION_THERMAL)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_ACPI:
+			if (ps->class & ATOM_PPLIB_CLASSIFICATION_ACPI)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_ULV:
+			if (ps->class2 & ATOM_PPLIB_CLASSIFICATION2_ULV)
+				return ps;
+			break;
+		case POWER_STATE_TYPE_INTERNAL_3DPERF:
+			if (ps->class & ATOM_PPLIB_CLASSIFICATION_3DPERFORMANCE)
+				return ps;
+			break;
+		default:
+			break;
+		}
+	}
+	/* use a fallback state if we didn't match */
+	switch (dpm_state) {
+	case POWER_STATE_TYPE_INTERNAL_UVD_SD:
+	case POWER_STATE_TYPE_INTERNAL_UVD_HD:
+	case POWER_STATE_TYPE_INTERNAL_UVD_HD2:
+	case POWER_STATE_TYPE_INTERNAL_UVD_MVC:
+		return rdev->pm.dpm.uvd_ps;
+	case POWER_STATE_TYPE_INTERNAL_THERMAL:
+		dpm_state = POWER_STATE_TYPE_INTERNAL_ACPI;
+		goto restart_search;
+	case POWER_STATE_TYPE_INTERNAL_ACPI:
+		dpm_state = POWER_STATE_TYPE_BATTERY;
+		goto restart_search;
+	case POWER_STATE_TYPE_BATTERY:
+	case POWER_STATE_TYPE_BALANCED:
+	case POWER_STATE_TYPE_INTERNAL_3DPERF:
+		dpm_state = POWER_STATE_TYPE_PERFORMANCE;
+		goto restart_search;
+	default:
+		break;
+	}
+
+	return NULL;
+}
+
+static void radeon_dpm_change_power_state_locked(struct radeon_device *rdev)
+{
+	int i;
+	struct radeon_ps *ps;
+	enum radeon_pm_state_type dpm_state;
+	int ret;
+
+	/* if dpm init failed */
+	if (!rdev->pm.dpm_enabled)
+		return;
+
+	if (rdev->pm.dpm.user_state != rdev->pm.dpm.state) {
+		/* add other state override checks here */
+		if ((!rdev->pm.dpm.thermal_active) &&
+		    (!rdev->pm.dpm.uvd_active))
+			rdev->pm.dpm.state = rdev->pm.dpm.user_state;
+	}
+	dpm_state = rdev->pm.dpm.state;
+
+	ps = radeon_dpm_pick_power_state(rdev, dpm_state);
+	if (ps)
+		rdev->pm.dpm.requested_ps = ps;
+	else
+		return;
+
+	/* no need to reprogram if nothing changed unless we are on BTC+ */
+	if (rdev->pm.dpm.current_ps == rdev->pm.dpm.requested_ps) {
+		if ((rdev->family < CHIP_BARTS) || (rdev->flags & RADEON_IS_IGP)) {
+			/* for pre-BTC and APUs if the num crtcs changed but state is the same,
+			 * all we need to do is update the display configuration.
+			 */
+			if (rdev->pm.dpm.new_active_crtcs != rdev->pm.dpm.current_active_crtcs) {
+				/* update display watermarks based on new power state */
+				radeon_bandwidth_update(rdev);
+				/* update displays */
+				radeon_dpm_display_configuration_changed(rdev);
+				rdev->pm.dpm.current_active_crtcs = rdev->pm.dpm.new_active_crtcs;
+				rdev->pm.dpm.current_active_crtc_count = rdev->pm.dpm.new_active_crtc_count;
+			}
+			return;
+		} else {
+			/* for BTC+ if the num crtcs hasn't changed and state is the same,
+			 * nothing to do, if the num crtcs is > 1 and state is the same,
+			 * update display configuration.
+			 */
+			if (rdev->pm.dpm.new_active_crtcs ==
+			    rdev->pm.dpm.current_active_crtcs) {
+				return;
+			} else {
+				if ((rdev->pm.dpm.current_active_crtc_count > 1) &&
+				    (rdev->pm.dpm.new_active_crtc_count > 1)) {
+					/* update display watermarks based on new power state */
+					radeon_bandwidth_update(rdev);
+					/* update displays */
+					radeon_dpm_display_configuration_changed(rdev);
+					rdev->pm.dpm.current_active_crtcs = rdev->pm.dpm.new_active_crtcs;
+					rdev->pm.dpm.current_active_crtc_count = rdev->pm.dpm.new_active_crtc_count;
+					return;
+				}
+			}
+		}
+	}
+
+	printk("switching from power state:\n");
+	radeon_dpm_print_power_state(rdev, rdev->pm.dpm.current_ps);
+	printk("switching to power state:\n");
+	radeon_dpm_print_power_state(rdev, rdev->pm.dpm.requested_ps);
+
+	mutex_lock(&rdev->ddev->struct_mutex);
+	down_write(&rdev->pm.mclk_lock);
+	mutex_lock(&rdev->ring_lock);
+
+	ret = radeon_dpm_pre_set_power_state(rdev);
+	if (ret)
+		goto done;
+
+	/* update display watermarks based on new power state */
+	radeon_bandwidth_update(rdev);
+	/* update displays */
+	radeon_dpm_display_configuration_changed(rdev);
+
+	rdev->pm.dpm.current_active_crtcs = rdev->pm.dpm.new_active_crtcs;
+	rdev->pm.dpm.current_active_crtc_count = rdev->pm.dpm.new_active_crtc_count;
+
+	/* wait for the rings to drain */
+	for (i = 0; i < RADEON_NUM_RINGS; i++) {
+		struct radeon_ring *ring = &rdev->ring[i];
+		if (ring->ready)
+			radeon_fence_wait_empty_locked(rdev, i);
+	}
+
+	/* program the new power state */
+	radeon_dpm_set_power_state(rdev);
+
+	/* update current power state */
+	rdev->pm.dpm.current_ps = rdev->pm.dpm.requested_ps;
+
+	radeon_dpm_post_set_power_state(rdev);
+
+done:
+	mutex_unlock(&rdev->ring_lock);
+	up_write(&rdev->pm.mclk_lock);
+	mutex_unlock(&rdev->ddev->struct_mutex);
+}
+
+void radeon_dpm_enable_power_state(struct radeon_device *rdev,
+				   enum radeon_pm_state_type dpm_state)
+{
+	if (!rdev->pm.dpm_enabled)
+		return;
+
+	mutex_lock(&rdev->pm.mutex);
+	switch (dpm_state) {
+	case POWER_STATE_TYPE_INTERNAL_THERMAL:
+		rdev->pm.dpm.thermal_active = true;
+		break;
+	case POWER_STATE_TYPE_INTERNAL_UVD:
+	case POWER_STATE_TYPE_INTERNAL_UVD_SD:
+	case POWER_STATE_TYPE_INTERNAL_UVD_HD:
+	case POWER_STATE_TYPE_INTERNAL_UVD_HD2:
+	case POWER_STATE_TYPE_INTERNAL_UVD_MVC:
+		rdev->pm.dpm.uvd_active = true;
+		break;
+	default:
+		rdev->pm.dpm.thermal_active = false;
+		rdev->pm.dpm.uvd_active = false;
+		break;
+	}
+	rdev->pm.dpm.state = dpm_state;
+	mutex_unlock(&rdev->pm.mutex);
+	radeon_pm_compute_clocks(rdev);
+}
+
+static void radeon_pm_suspend_old(struct radeon_device *rdev)
 {
 	mutex_lock(&rdev->pm.mutex);
 	if (rdev->pm.pm_method == PM_METHOD_DYNPM) {
@@ -538,11 +896,30 @@ void radeon_pm_suspend(struct radeon_device *rdev)
 	cancel_delayed_work_sync(&rdev->pm.dynpm_idle_work);
 }
 
-void radeon_pm_resume(struct radeon_device *rdev)
+static void radeon_pm_suspend_dpm(struct radeon_device *rdev)
+{
+	mutex_lock(&rdev->pm.mutex);
+	/* disable dpm */
+	radeon_dpm_disable(rdev);
+	/* reset the power state */
+	rdev->pm.dpm.current_ps = rdev->pm.dpm.requested_ps = rdev->pm.dpm.boot_ps;
+	rdev->pm.dpm_enabled = false;
+	mutex_unlock(&rdev->pm.mutex);
+}
+
+void radeon_pm_suspend(struct radeon_device *rdev)
+{
+	if (rdev->pm.pm_method == PM_METHOD_DPM)
+		radeon_pm_suspend_dpm(rdev);
+	else
+		radeon_pm_suspend_old(rdev);
+}
+
+static void radeon_pm_resume_old(struct radeon_device *rdev)
 {
 	/* set up the default clocks if the MC ucode is loaded */
 	if ((rdev->family >= CHIP_BARTS) &&
-	    (rdev->family <= CHIP_CAYMAN) &&
+	    (rdev->family <= CHIP_HAINAN) &&
 	    rdev->mc_fw) {
 		if (rdev->pm.default_vddc)
 			radeon_atom_set_voltage(rdev, rdev->pm.default_vddc,
@@ -573,12 +950,50 @@ void radeon_pm_resume(struct radeon_device *rdev)
 	radeon_pm_compute_clocks(rdev);
 }
 
-int radeon_pm_init(struct radeon_device *rdev)
+static void radeon_pm_resume_dpm(struct radeon_device *rdev)
+{
+	int ret;
+
+	/* asic init will reset to the boot state */
+	mutex_lock(&rdev->pm.mutex);
+	rdev->pm.dpm.current_ps = rdev->pm.dpm.requested_ps = rdev->pm.dpm.boot_ps;
+	radeon_dpm_setup_asic(rdev);
+	ret = radeon_dpm_enable(rdev);
+	mutex_unlock(&rdev->pm.mutex);
+	if (ret) {
+		DRM_ERROR("radeon: dpm resume failed\n");
+		if ((rdev->family >= CHIP_BARTS) &&
+		    (rdev->family <= CHIP_HAINAN) &&
+		    rdev->mc_fw) {
+			if (rdev->pm.default_vddc)
+				radeon_atom_set_voltage(rdev, rdev->pm.default_vddc,
+							SET_VOLTAGE_TYPE_ASIC_VDDC);
+			if (rdev->pm.default_vddci)
+				radeon_atom_set_voltage(rdev, rdev->pm.default_vddci,
+							SET_VOLTAGE_TYPE_ASIC_VDDCI);
+			if (rdev->pm.default_sclk)
+				radeon_set_engine_clock(rdev, rdev->pm.default_sclk);
+			if (rdev->pm.default_mclk)
+				radeon_set_memory_clock(rdev, rdev->pm.default_mclk);
+		}
+	} else {
+		rdev->pm.dpm_enabled = true;
+		radeon_pm_compute_clocks(rdev);
+	}
+}
+
+void radeon_pm_resume(struct radeon_device *rdev)
+{
+	if (rdev->pm.pm_method == PM_METHOD_DPM)
+		radeon_pm_resume_dpm(rdev);
+	else
+		radeon_pm_resume_old(rdev);
+}
+
+static int radeon_pm_init_old(struct radeon_device *rdev)
 {
 	int ret;
 
-	/* default to profile method */
-	rdev->pm.pm_method = PM_METHOD_PROFILE;
 	rdev->pm.profile = PM_PROFILE_DEFAULT;
 	rdev->pm.dynpm_state = DYNPM_STATE_DISABLED;
 	rdev->pm.dynpm_planned_action = DYNPM_ACTION_NONE;
@@ -599,7 +1014,7 @@ int radeon_pm_init(struct radeon_device *rdev)
 		radeon_pm_init_profile(rdev);
 		/* set up the default clocks if the MC ucode is loaded */
 		if ((rdev->family >= CHIP_BARTS) &&
-		    (rdev->family <= CHIP_CAYMAN) &&
+		    (rdev->family <= CHIP_HAINAN) &&
 		    rdev->mc_fw) {
 			if (rdev->pm.default_vddc)
 				radeon_atom_set_voltage(rdev, rdev->pm.default_vddc,
@@ -640,7 +1055,145 @@ int radeon_pm_init(struct radeon_device *rdev)
 	return 0;
 }
 
-void radeon_pm_fini(struct radeon_device *rdev)
+static void radeon_dpm_print_power_states(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		printk("== power state %d ==\n", i);
+		radeon_dpm_print_power_state(rdev, &rdev->pm.dpm.ps[i]);
+	}
+}
+
+static int radeon_pm_init_dpm(struct radeon_device *rdev)
+{
+	int ret;
+
+	/* default to performance state */
+	rdev->pm.dpm.state = POWER_STATE_TYPE_BALANCED;
+	rdev->pm.dpm.user_state = POWER_STATE_TYPE_BALANCED;
+	rdev->pm.default_sclk = rdev->clock.default_sclk;
+	rdev->pm.default_mclk = rdev->clock.default_mclk;
+	rdev->pm.current_sclk = rdev->clock.default_sclk;
+	rdev->pm.current_mclk = rdev->clock.default_mclk;
+	rdev->pm.int_thermal_type = THERMAL_TYPE_NONE;
+
+	if (rdev->bios && rdev->is_atom_bios)
+		radeon_atombios_get_power_modes(rdev);
+	else
+		return -EINVAL;
+
+	/* set up the internal thermal sensor if applicable */
+	ret = radeon_hwmon_init(rdev);
+	if (ret)
+		return ret;
+
+	INIT_WORK(&rdev->pm.dpm.thermal.work, radeon_dpm_thermal_work_handler);
+	mutex_lock(&rdev->pm.mutex);
+	radeon_dpm_init(rdev);
+	rdev->pm.dpm.current_ps = rdev->pm.dpm.requested_ps = rdev->pm.dpm.boot_ps;
+	radeon_dpm_print_power_states(rdev);
+	radeon_dpm_setup_asic(rdev);
+	ret = radeon_dpm_enable(rdev);
+	mutex_unlock(&rdev->pm.mutex);
+	if (ret) {
+		rdev->pm.dpm_enabled = false;
+		if ((rdev->family >= CHIP_BARTS) &&
+		    (rdev->family <= CHIP_HAINAN) &&
+		    rdev->mc_fw) {
+			if (rdev->pm.default_vddc)
+				radeon_atom_set_voltage(rdev, rdev->pm.default_vddc,
+							SET_VOLTAGE_TYPE_ASIC_VDDC);
+			if (rdev->pm.default_vddci)
+				radeon_atom_set_voltage(rdev, rdev->pm.default_vddci,
+							SET_VOLTAGE_TYPE_ASIC_VDDCI);
+			if (rdev->pm.default_sclk)
+				radeon_set_engine_clock(rdev, rdev->pm.default_sclk);
+			if (rdev->pm.default_mclk)
+				radeon_set_memory_clock(rdev, rdev->pm.default_mclk);
+		}
+		DRM_ERROR("radeon: dpm initialization failed\n");
+		return ret;
+	}
+	rdev->pm.dpm_enabled = true;
+	radeon_pm_compute_clocks(rdev);
+
+	if (rdev->pm.num_power_states > 1) {
+		ret = device_create_file(rdev->dev, &dev_attr_power_dpm_state);
+		if (ret)
+			DRM_ERROR("failed to create device file for dpm state\n");
+		ret = device_create_file(rdev->dev, &dev_attr_power_dpm_force_performance_level);
+		if (ret)
+			DRM_ERROR("failed to create device file for dpm state\n");
+		/* XXX: these are noops for dpm but are here for backwards compat */
+		ret = device_create_file(rdev->dev, &dev_attr_power_profile);
+		if (ret)
+			DRM_ERROR("failed to create device file for power profile\n");
+		ret = device_create_file(rdev->dev, &dev_attr_power_method);
+		if (ret)
+			DRM_ERROR("failed to create device file for power method\n");
+
+		if (radeon_debugfs_pm_init(rdev)) {
+			DRM_ERROR("Failed to register debugfs file for dpm!\n");
+		}
+
+		DRM_INFO("radeon: dpm initialized\n");
+	}
+
+	return 0;
+}
+
+int radeon_pm_init(struct radeon_device *rdev)
+{
+	/* enable dpm on rv6xx+ */
+	switch (rdev->family) {
+	case CHIP_RV610:
+	case CHIP_RV630:
+	case CHIP_RV620:
+	case CHIP_RV635:
+	case CHIP_RV670:
+	case CHIP_RS780:
+	case CHIP_RS880:
+	case CHIP_RV770:
+	case CHIP_RV730:
+	case CHIP_RV710:
+	case CHIP_RV740:
+	case CHIP_CEDAR:
+	case CHIP_REDWOOD:
+	case CHIP_JUNIPER:
+	case CHIP_CYPRESS:
+	case CHIP_HEMLOCK:
+	case CHIP_PALM:
+	case CHIP_SUMO:
+	case CHIP_SUMO2:
+	case CHIP_BARTS:
+	case CHIP_TURKS:
+	case CHIP_CAICOS:
+	case CHIP_CAYMAN:
+	case CHIP_ARUBA:
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+	case CHIP_VERDE:
+	case CHIP_OLAND:
+	case CHIP_HAINAN:
+		if (radeon_dpm == 1)
+			rdev->pm.pm_method = PM_METHOD_DPM;
+		else
+			rdev->pm.pm_method = PM_METHOD_PROFILE;
+		break;
+	default:
+		/* default to profile method */
+		rdev->pm.pm_method = PM_METHOD_PROFILE;
+		break;
+	}
+
+	if (rdev->pm.pm_method == PM_METHOD_DPM)
+		return radeon_pm_init_dpm(rdev);
+	else
+		return radeon_pm_init_old(rdev);
+}
+
+static void radeon_pm_fini_old(struct radeon_device *rdev)
 {
 	if (rdev->pm.num_power_states > 1) {
 		mutex_lock(&rdev->pm.mutex);
@@ -668,7 +1221,36 @@ void radeon_pm_fini(struct radeon_device *rdev)
 	radeon_hwmon_fini(rdev);
 }
 
-void radeon_pm_compute_clocks(struct radeon_device *rdev)
+static void radeon_pm_fini_dpm(struct radeon_device *rdev)
+{
+	if (rdev->pm.num_power_states > 1) {
+		mutex_lock(&rdev->pm.mutex);
+		radeon_dpm_disable(rdev);
+		mutex_unlock(&rdev->pm.mutex);
+
+		device_remove_file(rdev->dev, &dev_attr_power_dpm_state);
+		device_remove_file(rdev->dev, &dev_attr_power_dpm_force_performance_level);
+		/* XXX backwards compat */
+		device_remove_file(rdev->dev, &dev_attr_power_profile);
+		device_remove_file(rdev->dev, &dev_attr_power_method);
+	}
+	radeon_dpm_fini(rdev);
+
+	if (rdev->pm.power_state)
+		kfree(rdev->pm.power_state);
+
+	radeon_hwmon_fini(rdev);
+}
+
+void radeon_pm_fini(struct radeon_device *rdev)
+{
+	if (rdev->pm.pm_method == PM_METHOD_DPM)
+		radeon_pm_fini_dpm(rdev);
+	else
+		radeon_pm_fini_old(rdev);
+}
+
+static void radeon_pm_compute_clocks_old(struct radeon_device *rdev)
 {
 	struct drm_device *ddev = rdev->ddev;
 	struct drm_crtc *crtc;
@@ -739,6 +1321,46 @@ void radeon_pm_compute_clocks(struct radeon_device *rdev)
 	mutex_unlock(&rdev->pm.mutex);
 }
 
+static void radeon_pm_compute_clocks_dpm(struct radeon_device *rdev)
+{
+	struct drm_device *ddev = rdev->ddev;
+	struct drm_crtc *crtc;
+	struct radeon_crtc *radeon_crtc;
+
+	mutex_lock(&rdev->pm.mutex);
+
+	/* update active crtc counts */
+	rdev->pm.dpm.new_active_crtcs = 0;
+	rdev->pm.dpm.new_active_crtc_count = 0;
+	list_for_each_entry(crtc,
+		&ddev->mode_config.crtc_list, head) {
+		radeon_crtc = to_radeon_crtc(crtc);
+		if (crtc->enabled) {
+			rdev->pm.dpm.new_active_crtcs |= (1 << radeon_crtc->crtc_id);
+			rdev->pm.dpm.new_active_crtc_count++;
+		}
+	}
+
+	/* update battery/ac status */
+	if (power_supply_is_system_supplied() > 0)
+		rdev->pm.dpm.ac_power = true;
+	else
+		rdev->pm.dpm.ac_power = false;
+
+	radeon_dpm_change_power_state_locked(rdev);
+
+	mutex_unlock(&rdev->pm.mutex);
+
+}
+
+void radeon_pm_compute_clocks(struct radeon_device *rdev)
+{
+	if (rdev->pm.pm_method == PM_METHOD_DPM)
+		radeon_pm_compute_clocks_dpm(rdev);
+	else
+		radeon_pm_compute_clocks_old(rdev);
+}
+
 static bool radeon_pm_in_vbl(struct radeon_device *rdev)
 {
 	int  crtc, vpos, hpos, vbl_status;
@@ -842,19 +1464,28 @@ static int radeon_debugfs_pm_info(struct seq_file *m, void *data)
 	struct drm_device *dev = node->minor->dev;
 	struct radeon_device *rdev = dev->dev_private;
 
-	seq_printf(m, "default engine clock: %u0 kHz\n", rdev->pm.default_sclk);
-	/* radeon_get_engine_clock is not reliable on APUs so just print the current clock */
-	if ((rdev->family >= CHIP_PALM) && (rdev->flags & RADEON_IS_IGP))
-		seq_printf(m, "current engine clock: %u0 kHz\n", rdev->pm.current_sclk);
-	else
-		seq_printf(m, "current engine clock: %u0 kHz\n", radeon_get_engine_clock(rdev));
-	seq_printf(m, "default memory clock: %u0 kHz\n", rdev->pm.default_mclk);
-	if (rdev->asic->pm.get_memory_clock)
-		seq_printf(m, "current memory clock: %u0 kHz\n", radeon_get_memory_clock(rdev));
-	if (rdev->pm.current_vddc)
-		seq_printf(m, "voltage: %u mV\n", rdev->pm.current_vddc);
-	if (rdev->asic->pm.get_pcie_lanes)
-		seq_printf(m, "PCIE lanes: %d\n", radeon_get_pcie_lanes(rdev));
+	if (rdev->pm.dpm_enabled) {
+		mutex_lock(&rdev->pm.mutex);
+		if (rdev->asic->dpm.debugfs_print_current_performance_level)
+			radeon_dpm_debugfs_print_current_performance_level(rdev, m);
+		else
+			seq_printf(m, "Debugfs support not implemented for this asic\n");
+		mutex_unlock(&rdev->pm.mutex);
+	} else {
+		seq_printf(m, "default engine clock: %u0 kHz\n", rdev->pm.default_sclk);
+		/* radeon_get_engine_clock is not reliable on APUs so just print the current clock */
+		if ((rdev->family >= CHIP_PALM) && (rdev->flags & RADEON_IS_IGP))
+			seq_printf(m, "current engine clock: %u0 kHz\n", rdev->pm.current_sclk);
+		else
+			seq_printf(m, "current engine clock: %u0 kHz\n", radeon_get_engine_clock(rdev));
+		seq_printf(m, "default memory clock: %u0 kHz\n", rdev->pm.default_mclk);
+		if (rdev->asic->pm.get_memory_clock)
+			seq_printf(m, "current memory clock: %u0 kHz\n", radeon_get_memory_clock(rdev));
+		if (rdev->pm.current_vddc)
+			seq_printf(m, "voltage: %u mV\n", rdev->pm.current_vddc);
+		if (rdev->asic->pm.get_pcie_lanes)
+			seq_printf(m, "PCIE lanes: %d\n", radeon_get_pcie_lanes(rdev));
+	}
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/radeon/radeon_prime.c b/drivers/gpu/drm/radeon/radeon_prime.c
index 4940af7e75e6..65b9eabd5a2f 100644
--- a/drivers/gpu/drm/radeon/radeon_prime.c
+++ b/drivers/gpu/drm/radeon/radeon_prime.c
@@ -88,11 +88,19 @@ int radeon_gem_prime_pin(struct drm_gem_object *obj)
 
 	/* pin buffer into GTT */
 	ret = radeon_bo_pin(bo, RADEON_GEM_DOMAIN_GTT, NULL);
-	if (ret) {
-		radeon_bo_unreserve(bo);
-		return ret;
-	}
 	radeon_bo_unreserve(bo);
+	return ret;
+}
+
+void radeon_gem_prime_unpin(struct drm_gem_object *obj)
+{
+	struct radeon_bo *bo = gem_to_radeon_bo(obj);
+	int ret = 0;
 
-	return 0;
+	ret = radeon_bo_reserve(bo, false);
+	if (unlikely(ret != 0))
+		return;
+
+	radeon_bo_unpin(bo);
+	radeon_bo_unreserve(bo);
 }
diff --git a/drivers/gpu/drm/radeon/radeon_reg.h b/drivers/gpu/drm/radeon/radeon_reg.h
index 7e2c2b7cf188..62d54976d24e 100644
--- a/drivers/gpu/drm/radeon/radeon_reg.h
+++ b/drivers/gpu/drm/radeon/radeon_reg.h
@@ -57,6 +57,7 @@
 #include "evergreen_reg.h"
 #include "ni_reg.h"
 #include "si_reg.h"
+#include "cik_reg.h"
 
 #define RADEON_MC_AGP_LOCATION		0x014c
 #define		RADEON_MC_AGP_START_MASK	0x0000FFFF
diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c
index 82434018cbe8..fb5ea6208970 100644
--- a/drivers/gpu/drm/radeon/radeon_ring.c
+++ b/drivers/gpu/drm/radeon/radeon_ring.c
@@ -224,6 +224,7 @@ int radeon_ib_pool_init(struct radeon_device *rdev)
 	}
 	r = radeon_sa_bo_manager_init(rdev, &rdev->ring_tmp_bo,
 				      RADEON_IB_POOL_SIZE*64*1024,
+				      RADEON_GPU_PAGE_SIZE,
 				      RADEON_GEM_DOMAIN_GTT);
 	if (r) {
 		return r;
@@ -357,6 +358,38 @@ bool radeon_ring_supports_scratch_reg(struct radeon_device *rdev,
 	}
 }
 
+u32 radeon_ring_generic_get_rptr(struct radeon_device *rdev,
+				 struct radeon_ring *ring)
+{
+	u32 rptr;
+
+	if (rdev->wb.enabled && ring != &rdev->ring[R600_RING_TYPE_UVD_INDEX])
+		rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]);
+	else
+		rptr = RREG32(ring->rptr_reg);
+	rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+
+	return rptr;
+}
+
+u32 radeon_ring_generic_get_wptr(struct radeon_device *rdev,
+				 struct radeon_ring *ring)
+{
+	u32 wptr;
+
+	wptr = RREG32(ring->wptr_reg);
+	wptr = (wptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+
+	return wptr;
+}
+
+void radeon_ring_generic_set_wptr(struct radeon_device *rdev,
+				  struct radeon_ring *ring)
+{
+	WREG32(ring->wptr_reg, (ring->wptr << ring->ptr_reg_shift) & ring->ptr_reg_mask);
+	(void)RREG32(ring->wptr_reg);
+}
+
 /**
  * radeon_ring_free_size - update the free size
  *
@@ -367,13 +400,7 @@ bool radeon_ring_supports_scratch_reg(struct radeon_device *rdev,
  */
 void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *ring)
 {
-	u32 rptr;
-
-	if (rdev->wb.enabled && ring != &rdev->ring[R600_RING_TYPE_UVD_INDEX])
-		rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]);
-	else
-		rptr = RREG32(ring->rptr_reg);
-	ring->rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+	ring->rptr = radeon_ring_get_rptr(rdev, ring);
 	/* This works because ring_size is a power of 2 */
 	ring->ring_free_dw = (ring->rptr + (ring->ring_size / 4));
 	ring->ring_free_dw -= ring->wptr;
@@ -465,8 +492,7 @@ void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring)
 		radeon_ring_write(ring, ring->nop);
 	}
 	DRM_MEMORYBARRIER();
-	WREG32(ring->wptr_reg, (ring->wptr << ring->ptr_reg_shift) & ring->ptr_reg_mask);
-	(void)RREG32(ring->wptr_reg);
+	radeon_ring_set_wptr(rdev, ring);
 }
 
 /**
@@ -568,7 +594,6 @@ void radeon_ring_lockup_update(struct radeon_ring *ring)
 bool radeon_ring_test_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 {
 	unsigned long cjiffies, elapsed;
-	uint32_t rptr;
 
 	cjiffies = jiffies;
 	if (!time_after(cjiffies, ring->last_activity)) {
@@ -576,8 +601,7 @@ bool radeon_ring_test_lockup(struct radeon_device *rdev, struct radeon_ring *rin
 		radeon_ring_lockup_update(ring);
 		return false;
 	}
-	rptr = RREG32(ring->rptr_reg);
-	ring->rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+	ring->rptr = radeon_ring_get_rptr(rdev, ring);
 	if (ring->rptr != ring->last_rptr) {
 		/* CP is still working no lockup */
 		radeon_ring_lockup_update(ring);
@@ -804,9 +828,9 @@ static int radeon_debugfs_ring_info(struct seq_file *m, void *data)
 
 	radeon_ring_free_size(rdev, ring);
 	count = (ring->ring_size / 4) - ring->ring_free_dw;
-	tmp = RREG32(ring->wptr_reg) >> ring->ptr_reg_shift;
+	tmp = radeon_ring_get_wptr(rdev, ring);
 	seq_printf(m, "wptr(0x%04x): 0x%08x [%5d]\n", ring->wptr_reg, tmp, tmp);
-	tmp = RREG32(ring->rptr_reg) >> ring->ptr_reg_shift;
+	tmp = radeon_ring_get_rptr(rdev, ring);
 	seq_printf(m, "rptr(0x%04x): 0x%08x [%5d]\n", ring->rptr_reg, tmp, tmp);
 	if (ring->rptr_save_reg) {
 		seq_printf(m, "rptr next(0x%04x): 0x%08x\n", ring->rptr_save_reg,
diff --git a/drivers/gpu/drm/radeon/radeon_sa.c b/drivers/gpu/drm/radeon/radeon_sa.c
index 0abe5a9431bb..f0bac68254b7 100644
--- a/drivers/gpu/drm/radeon/radeon_sa.c
+++ b/drivers/gpu/drm/radeon/radeon_sa.c
@@ -49,7 +49,7 @@ static void radeon_sa_bo_try_free(struct radeon_sa_manager *sa_manager);
 
 int radeon_sa_bo_manager_init(struct radeon_device *rdev,
 			      struct radeon_sa_manager *sa_manager,
-			      unsigned size, u32 domain)
+			      unsigned size, u32 align, u32 domain)
 {
 	int i, r;
 
@@ -57,13 +57,14 @@ int radeon_sa_bo_manager_init(struct radeon_device *rdev,
 	sa_manager->bo = NULL;
 	sa_manager->size = size;
 	sa_manager->domain = domain;
+	sa_manager->align = align;
 	sa_manager->hole = &sa_manager->olist;
 	INIT_LIST_HEAD(&sa_manager->olist);
 	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
 		INIT_LIST_HEAD(&sa_manager->flist[i]);
 	}
 
-	r = radeon_bo_create(rdev, size, RADEON_GPU_PAGE_SIZE, true,
+	r = radeon_bo_create(rdev, size, align, true,
 			     domain, NULL, &sa_manager->bo);
 	if (r) {
 		dev_err(rdev->dev, "(%d) failed to allocate bo for manager\n", r);
@@ -317,7 +318,7 @@ int radeon_sa_bo_new(struct radeon_device *rdev,
 	unsigned tries[RADEON_NUM_RINGS];
 	int i, r;
 
-	BUG_ON(align > RADEON_GPU_PAGE_SIZE);
+	BUG_ON(align > sa_manager->align);
 	BUG_ON(size > sa_manager->size);
 
 	*sa_bo = kmalloc(sizeof(struct radeon_sa_bo), GFP_KERNEL);
diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c
index bbed4af8d0bc..f4d6bcee9006 100644
--- a/drivers/gpu/drm/radeon/radeon_test.c
+++ b/drivers/gpu/drm/radeon/radeon_test.c
@@ -35,7 +35,6 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 {
 	struct radeon_bo *vram_obj = NULL;
 	struct radeon_bo **gtt_obj = NULL;
-	struct radeon_fence *fence = NULL;
 	uint64_t gtt_addr, vram_addr;
 	unsigned i, n, size;
 	int r, ring;
@@ -81,37 +80,38 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 	}
 	r = radeon_bo_reserve(vram_obj, false);
 	if (unlikely(r != 0))
-		goto out_cleanup;
+		goto out_unref;
 	r = radeon_bo_pin(vram_obj, RADEON_GEM_DOMAIN_VRAM, &vram_addr);
 	if (r) {
 		DRM_ERROR("Failed to pin VRAM object\n");
-		goto out_cleanup;
+		goto out_unres;
 	}
 	for (i = 0; i < n; i++) {
 		void *gtt_map, *vram_map;
 		void **gtt_start, **gtt_end;
 		void **vram_start, **vram_end;
+		struct radeon_fence *fence = NULL;
 
 		r = radeon_bo_create(rdev, size, PAGE_SIZE, true,
 				     RADEON_GEM_DOMAIN_GTT, NULL, gtt_obj + i);
 		if (r) {
 			DRM_ERROR("Failed to create GTT object %d\n", i);
-			goto out_cleanup;
+			goto out_lclean;
 		}
 
 		r = radeon_bo_reserve(gtt_obj[i], false);
 		if (unlikely(r != 0))
-			goto out_cleanup;
+			goto out_lclean_unref;
 		r = radeon_bo_pin(gtt_obj[i], RADEON_GEM_DOMAIN_GTT, &gtt_addr);
 		if (r) {
 			DRM_ERROR("Failed to pin GTT object %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unres;
 		}
 
 		r = radeon_bo_kmap(gtt_obj[i], &gtt_map);
 		if (r) {
 			DRM_ERROR("Failed to map GTT object %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		for (gtt_start = gtt_map, gtt_end = gtt_map + size;
@@ -127,13 +127,13 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 			r = radeon_copy_blit(rdev, gtt_addr, vram_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
 		if (r) {
 			DRM_ERROR("Failed GTT->VRAM copy %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		r = radeon_fence_wait(fence, false);
 		if (r) {
 			DRM_ERROR("Failed to wait for GTT->VRAM fence %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		radeon_fence_unref(&fence);
@@ -141,7 +141,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 		r = radeon_bo_kmap(vram_obj, &vram_map);
 		if (r) {
 			DRM_ERROR("Failed to map VRAM object after copy %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		for (gtt_start = gtt_map, gtt_end = gtt_map + size,
@@ -160,7 +160,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 					  (vram_addr - rdev->mc.vram_start +
 					   (void*)gtt_start - gtt_map));
 				radeon_bo_kunmap(vram_obj);
-				goto out_cleanup;
+				goto out_lclean_unpin;
 			}
 			*vram_start = vram_start;
 		}
@@ -173,13 +173,13 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 			r = radeon_copy_blit(rdev, vram_addr, gtt_addr, size / RADEON_GPU_PAGE_SIZE, &fence);
 		if (r) {
 			DRM_ERROR("Failed VRAM->GTT copy %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		r = radeon_fence_wait(fence, false);
 		if (r) {
 			DRM_ERROR("Failed to wait for VRAM->GTT fence %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		radeon_fence_unref(&fence);
@@ -187,7 +187,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 		r = radeon_bo_kmap(gtt_obj[i], &gtt_map);
 		if (r) {
 			DRM_ERROR("Failed to map GTT object after copy %d\n", i);
-			goto out_cleanup;
+			goto out_lclean_unpin;
 		}
 
 		for (gtt_start = gtt_map, gtt_end = gtt_map + size,
@@ -206,7 +206,7 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 					  (gtt_addr - rdev->mc.gtt_start +
 					   (void*)vram_start - vram_map));
 				radeon_bo_kunmap(gtt_obj[i]);
-				goto out_cleanup;
+				goto out_lclean_unpin;
 			}
 		}
 
@@ -214,31 +214,32 @@ static void radeon_do_test_moves(struct radeon_device *rdev, int flag)
 
 		DRM_INFO("Tested GTT->VRAM and VRAM->GTT copy for GTT offset 0x%llx\n",
 			 gtt_addr - rdev->mc.gtt_start);
+		continue;
+
+out_lclean_unpin:
+		radeon_bo_unpin(gtt_obj[i]);
+out_lclean_unres:
+		radeon_bo_unreserve(gtt_obj[i]);
+out_lclean_unref:
+		radeon_bo_unref(&gtt_obj[i]);
+out_lclean:
+		for (--i; i >= 0; --i) {
+			radeon_bo_unpin(gtt_obj[i]);
+			radeon_bo_unreserve(gtt_obj[i]);
+			radeon_bo_unref(&gtt_obj[i]);
+		}
+		if (fence)
+			radeon_fence_unref(&fence);
+		break;
 	}
 
+	radeon_bo_unpin(vram_obj);
+out_unres:
+	radeon_bo_unreserve(vram_obj);
+out_unref:
+	radeon_bo_unref(&vram_obj);
 out_cleanup:
-	if (vram_obj) {
-		if (radeon_bo_is_reserved(vram_obj)) {
-			radeon_bo_unpin(vram_obj);
-			radeon_bo_unreserve(vram_obj);
-		}
-		radeon_bo_unref(&vram_obj);
-	}
-	if (gtt_obj) {
-		for (i = 0; i < n; i++) {
-			if (gtt_obj[i]) {
-				if (radeon_bo_is_reserved(gtt_obj[i])) {
-					radeon_bo_unpin(gtt_obj[i]);
-					radeon_bo_unreserve(gtt_obj[i]);
-				}
-				radeon_bo_unref(&gtt_obj[i]);
-			}
-		}
-		kfree(gtt_obj);
-	}
-	if (fence) {
-		radeon_fence_unref(&fence);
-	}
+	kfree(gtt_obj);
 	if (r) {
 		printk(KERN_WARNING "Error while testing BO move.\n");
 	}
diff --git a/drivers/gpu/drm/radeon/radeon_ucode.h b/drivers/gpu/drm/radeon/radeon_ucode.h
new file mode 100644
index 000000000000..d8b05f7bcf1a
--- /dev/null
+++ b/drivers/gpu/drm/radeon/radeon_ucode.h
@@ -0,0 +1,129 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __RADEON_UCODE_H__
+#define __RADEON_UCODE_H__
+
+/* CP */
+#define R600_PFP_UCODE_SIZE          576
+#define R600_PM4_UCODE_SIZE          1792
+#define R700_PFP_UCODE_SIZE          848
+#define R700_PM4_UCODE_SIZE          1360
+#define EVERGREEN_PFP_UCODE_SIZE     1120
+#define EVERGREEN_PM4_UCODE_SIZE     1376
+#define CAYMAN_PFP_UCODE_SIZE        2176
+#define CAYMAN_PM4_UCODE_SIZE        2176
+#define SI_PFP_UCODE_SIZE            2144
+#define SI_PM4_UCODE_SIZE            2144
+#define SI_CE_UCODE_SIZE             2144
+
+/* RLC */
+#define R600_RLC_UCODE_SIZE          768
+#define R700_RLC_UCODE_SIZE          1024
+#define EVERGREEN_RLC_UCODE_SIZE     768
+#define CAYMAN_RLC_UCODE_SIZE        1024
+#define ARUBA_RLC_UCODE_SIZE         1536
+#define SI_RLC_UCODE_SIZE            2048
+
+/* MC */
+#define BTC_MC_UCODE_SIZE            6024
+#define CAYMAN_MC_UCODE_SIZE         6037
+#define SI_MC_UCODE_SIZE             7769
+#define OLAND_MC_UCODE_SIZE          7863
+
+/* SMC */
+#define RV770_SMC_UCODE_START        0x0100
+#define RV770_SMC_UCODE_SIZE         0x410d
+#define RV770_SMC_INT_VECTOR_START   0xffc0
+#define RV770_SMC_INT_VECTOR_SIZE    0x0040
+
+#define RV730_SMC_UCODE_START        0x0100
+#define RV730_SMC_UCODE_SIZE         0x412c
+#define RV730_SMC_INT_VECTOR_START   0xffc0
+#define RV730_SMC_INT_VECTOR_SIZE    0x0040
+
+#define RV710_SMC_UCODE_START        0x0100
+#define RV710_SMC_UCODE_SIZE         0x3f1f
+#define RV710_SMC_INT_VECTOR_START   0xffc0
+#define RV710_SMC_INT_VECTOR_SIZE    0x0040
+
+#define RV740_SMC_UCODE_START        0x0100
+#define RV740_SMC_UCODE_SIZE         0x41c5
+#define RV740_SMC_INT_VECTOR_START   0xffc0
+#define RV740_SMC_INT_VECTOR_SIZE    0x0040
+
+#define CEDAR_SMC_UCODE_START        0x0100
+#define CEDAR_SMC_UCODE_SIZE         0x5d50
+#define CEDAR_SMC_INT_VECTOR_START   0xffc0
+#define CEDAR_SMC_INT_VECTOR_SIZE    0x0040
+
+#define REDWOOD_SMC_UCODE_START      0x0100
+#define REDWOOD_SMC_UCODE_SIZE       0x5f0a
+#define REDWOOD_SMC_INT_VECTOR_START 0xffc0
+#define REDWOOD_SMC_INT_VECTOR_SIZE  0x0040
+
+#define JUNIPER_SMC_UCODE_START      0x0100
+#define JUNIPER_SMC_UCODE_SIZE       0x5f1f
+#define JUNIPER_SMC_INT_VECTOR_START 0xffc0
+#define JUNIPER_SMC_INT_VECTOR_SIZE  0x0040
+
+#define CYPRESS_SMC_UCODE_START      0x0100
+#define CYPRESS_SMC_UCODE_SIZE       0x61f7
+#define CYPRESS_SMC_INT_VECTOR_START 0xffc0
+#define CYPRESS_SMC_INT_VECTOR_SIZE  0x0040
+
+#define BARTS_SMC_UCODE_START        0x0100
+#define BARTS_SMC_UCODE_SIZE         0x6107
+#define BARTS_SMC_INT_VECTOR_START   0xffc0
+#define BARTS_SMC_INT_VECTOR_SIZE    0x0040
+
+#define TURKS_SMC_UCODE_START        0x0100
+#define TURKS_SMC_UCODE_SIZE         0x605b
+#define TURKS_SMC_INT_VECTOR_START   0xffc0
+#define TURKS_SMC_INT_VECTOR_SIZE    0x0040
+
+#define CAICOS_SMC_UCODE_START       0x0100
+#define CAICOS_SMC_UCODE_SIZE        0x5fbd
+#define CAICOS_SMC_INT_VECTOR_START  0xffc0
+#define CAICOS_SMC_INT_VECTOR_SIZE   0x0040
+
+#define CAYMAN_SMC_UCODE_START       0x0100
+#define CAYMAN_SMC_UCODE_SIZE        0x79ec
+#define CAYMAN_SMC_INT_VECTOR_START  0xffc0
+#define CAYMAN_SMC_INT_VECTOR_SIZE   0x0040
+
+#define TAHITI_SMC_UCODE_START       0x10000
+#define TAHITI_SMC_UCODE_SIZE        0xf458
+
+#define PITCAIRN_SMC_UCODE_START     0x10000
+#define PITCAIRN_SMC_UCODE_SIZE      0xe9f4
+
+#define VERDE_SMC_UCODE_START        0x10000
+#define VERDE_SMC_UCODE_SIZE         0xebe4
+
+#define OLAND_SMC_UCODE_START        0x10000
+#define OLAND_SMC_UCODE_SIZE         0xe7b4
+
+#define HAINAN_SMC_UCODE_START       0x10000
+#define HAINAN_SMC_UCODE_SIZE        0xe67C
+
+#endif
diff --git a/drivers/gpu/drm/radeon/radeon_uvd.c b/drivers/gpu/drm/radeon/radeon_uvd.c
index cad735dd02c6..414fd145d20e 100644
--- a/drivers/gpu/drm/radeon/radeon_uvd.c
+++ b/drivers/gpu/drm/radeon/radeon_uvd.c
@@ -44,30 +44,25 @@
 #define FIRMWARE_CYPRESS	"radeon/CYPRESS_uvd.bin"
 #define FIRMWARE_SUMO		"radeon/SUMO_uvd.bin"
 #define FIRMWARE_TAHITI		"radeon/TAHITI_uvd.bin"
+#define FIRMWARE_BONAIRE	"radeon/BONAIRE_uvd.bin"
 
 MODULE_FIRMWARE(FIRMWARE_RV710);
 MODULE_FIRMWARE(FIRMWARE_CYPRESS);
 MODULE_FIRMWARE(FIRMWARE_SUMO);
 MODULE_FIRMWARE(FIRMWARE_TAHITI);
+MODULE_FIRMWARE(FIRMWARE_BONAIRE);
 
 static void radeon_uvd_idle_work_handler(struct work_struct *work);
 
 int radeon_uvd_init(struct radeon_device *rdev)
 {
-	struct platform_device *pdev;
+	const struct firmware *fw;
 	unsigned long bo_size;
 	const char *fw_name;
 	int i, r;
 
 	INIT_DELAYED_WORK(&rdev->uvd.idle_work, radeon_uvd_idle_work_handler);
 
-	pdev = platform_device_register_simple("radeon_uvd", 0, NULL, 0);
-	r = IS_ERR(pdev);
-	if (r) {
-		dev_err(rdev->dev, "radeon_uvd: Failed to register firmware\n");
-		return -EINVAL;
-	}
-
 	switch (rdev->family) {
 	case CHIP_RV710:
 	case CHIP_RV730:
@@ -100,21 +95,24 @@ int radeon_uvd_init(struct radeon_device *rdev)
 		fw_name = FIRMWARE_TAHITI;
 		break;
 
+	case CHIP_BONAIRE:
+	case CHIP_KABINI:
+	case CHIP_KAVERI:
+		fw_name = FIRMWARE_BONAIRE;
+		break;
+
 	default:
 		return -EINVAL;
 	}
 
-	r = request_firmware(&rdev->uvd_fw, fw_name, &pdev->dev);
+	r = request_firmware(&fw, fw_name, rdev->dev);
 	if (r) {
 		dev_err(rdev->dev, "radeon_uvd: Can't load firmware \"%s\"\n",
 			fw_name);
-		platform_device_unregister(pdev);
 		return r;
 	}
 
-	platform_device_unregister(pdev);
-
-	bo_size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 8) +
+	bo_size = RADEON_GPU_PAGE_ALIGN(fw->size + 8) +
 		  RADEON_UVD_STACK_SIZE + RADEON_UVD_HEAP_SIZE;
 	r = radeon_bo_create(rdev, bo_size, PAGE_SIZE, true,
 			     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->uvd.vcpu_bo);
@@ -123,16 +121,35 @@ int radeon_uvd_init(struct radeon_device *rdev)
 		return r;
 	}
 
-	r = radeon_uvd_resume(rdev);
-	if (r)
+	r = radeon_bo_reserve(rdev->uvd.vcpu_bo, false);
+	if (r) {
+		radeon_bo_unref(&rdev->uvd.vcpu_bo);
+		dev_err(rdev->dev, "(%d) failed to reserve UVD bo\n", r);
 		return r;
+	}
 
-	memset(rdev->uvd.cpu_addr, 0, bo_size);
-	memcpy(rdev->uvd.cpu_addr, rdev->uvd_fw->data, rdev->uvd_fw->size);
+	r = radeon_bo_pin(rdev->uvd.vcpu_bo, RADEON_GEM_DOMAIN_VRAM,
+			  &rdev->uvd.gpu_addr);
+	if (r) {
+		radeon_bo_unreserve(rdev->uvd.vcpu_bo);
+		radeon_bo_unref(&rdev->uvd.vcpu_bo);
+		dev_err(rdev->dev, "(%d) UVD bo pin failed\n", r);
+		return r;
+	}
 
-	r = radeon_uvd_suspend(rdev);
-	if (r)
+	r = radeon_bo_kmap(rdev->uvd.vcpu_bo, &rdev->uvd.cpu_addr);
+	if (r) {
+		dev_err(rdev->dev, "(%d) UVD map failed\n", r);
 		return r;
+	}
+
+	radeon_bo_unreserve(rdev->uvd.vcpu_bo);
+
+	rdev->uvd.fw_size = fw->size;
+	memset(rdev->uvd.cpu_addr, 0, bo_size);
+	memcpy(rdev->uvd.cpu_addr, fw->data, fw->size);
+
+	release_firmware(fw);
 
 	for (i = 0; i < RADEON_MAX_UVD_HANDLES; ++i) {
 		atomic_set(&rdev->uvd.handles[i], 0);
@@ -144,71 +161,47 @@ int radeon_uvd_init(struct radeon_device *rdev)
 
 void radeon_uvd_fini(struct radeon_device *rdev)
 {
-	radeon_uvd_suspend(rdev);
-	radeon_bo_unref(&rdev->uvd.vcpu_bo);
-}
-
-int radeon_uvd_suspend(struct radeon_device *rdev)
-{
 	int r;
 
 	if (rdev->uvd.vcpu_bo == NULL)
-		return 0;
+		return;
 
 	r = radeon_bo_reserve(rdev->uvd.vcpu_bo, false);
 	if (!r) {
 		radeon_bo_kunmap(rdev->uvd.vcpu_bo);
 		radeon_bo_unpin(rdev->uvd.vcpu_bo);
-		rdev->uvd.cpu_addr = NULL;
-		if (!radeon_bo_pin(rdev->uvd.vcpu_bo, RADEON_GEM_DOMAIN_CPU, NULL)) {
-			radeon_bo_kmap(rdev->uvd.vcpu_bo, &rdev->uvd.cpu_addr);
-		}
 		radeon_bo_unreserve(rdev->uvd.vcpu_bo);
-
-		if (rdev->uvd.cpu_addr) {
-			radeon_fence_driver_start_ring(rdev, R600_RING_TYPE_UVD_INDEX);
-		} else {
-			rdev->fence_drv[R600_RING_TYPE_UVD_INDEX].cpu_addr = NULL;
-		}
 	}
-	return r;
+
+	radeon_bo_unref(&rdev->uvd.vcpu_bo);
 }
 
-int radeon_uvd_resume(struct radeon_device *rdev)
+int radeon_uvd_suspend(struct radeon_device *rdev)
 {
-	int r;
+	unsigned size;
 
 	if (rdev->uvd.vcpu_bo == NULL)
-		return -EINVAL;
+		return 0;
 
-	r = radeon_bo_reserve(rdev->uvd.vcpu_bo, false);
-	if (r) {
-		radeon_bo_unref(&rdev->uvd.vcpu_bo);
-		dev_err(rdev->dev, "(%d) failed to reserve UVD bo\n", r);
-		return r;
-	}
+	size = radeon_bo_size(rdev->uvd.vcpu_bo);
+	rdev->uvd.saved_bo = kmalloc(size, GFP_KERNEL);
+	memcpy(rdev->uvd.saved_bo, rdev->uvd.cpu_addr, size);
 
-	/* Have been pin in cpu unmap unpin */
-	radeon_bo_kunmap(rdev->uvd.vcpu_bo);
-	radeon_bo_unpin(rdev->uvd.vcpu_bo);
+	return 0;
+}
 
-	r = radeon_bo_pin(rdev->uvd.vcpu_bo, RADEON_GEM_DOMAIN_VRAM,
-			  &rdev->uvd.gpu_addr);
-	if (r) {
-		radeon_bo_unreserve(rdev->uvd.vcpu_bo);
-		radeon_bo_unref(&rdev->uvd.vcpu_bo);
-		dev_err(rdev->dev, "(%d) UVD bo pin failed\n", r);
-		return r;
-	}
+int radeon_uvd_resume(struct radeon_device *rdev)
+{
+	if (rdev->uvd.vcpu_bo == NULL)
+		return -EINVAL;
 
-	r = radeon_bo_kmap(rdev->uvd.vcpu_bo, &rdev->uvd.cpu_addr);
-	if (r) {
-		dev_err(rdev->dev, "(%d) UVD map failed\n", r);
-		return r;
+	if (rdev->uvd.saved_bo != NULL) {
+		unsigned size = radeon_bo_size(rdev->uvd.vcpu_bo);
+		memcpy(rdev->uvd.cpu_addr, rdev->uvd.saved_bo, size);
+		kfree(rdev->uvd.saved_bo);
+		rdev->uvd.saved_bo = NULL;
 	}
 
-	radeon_bo_unreserve(rdev->uvd.vcpu_bo);
-
 	return 0;
 }
 
@@ -542,6 +535,7 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 			       struct radeon_fence **fence)
 {
 	struct ttm_validate_buffer tv;
+	struct ww_acquire_ctx ticket;
 	struct list_head head;
 	struct radeon_ib ib;
 	uint64_t addr;
@@ -553,7 +547,7 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 	INIT_LIST_HEAD(&head);
 	list_add(&tv.head, &head);
 
-	r = ttm_eu_reserve_buffers(&head);
+	r = ttm_eu_reserve_buffers(&ticket, &head);
 	if (r)
 		return r;
 
@@ -561,16 +555,12 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 	radeon_uvd_force_into_uvd_segment(bo);
 
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
-	if (r) {
-		ttm_eu_backoff_reservation(&head);
-		return r;
-	}
+	if (r) 
+		goto err;
 
 	r = radeon_ib_get(rdev, ring, &ib, NULL, 16);
-	if (r) {
-		ttm_eu_backoff_reservation(&head);
-		return r;
-	}
+	if (r)
+		goto err;
 
 	addr = radeon_bo_gpu_offset(bo);
 	ib.ptr[0] = PACKET0(UVD_GPCOM_VCPU_DATA0, 0);
@@ -584,11 +574,9 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 	ib.length_dw = 16;
 
 	r = radeon_ib_schedule(rdev, &ib, NULL);
-	if (r) {
-		ttm_eu_backoff_reservation(&head);
-		return r;
-	}
-	ttm_eu_fence_buffer_objects(&head, ib.fence);
+	if (r)
+		goto err;
+	ttm_eu_fence_buffer_objects(&ticket, &head, ib.fence);
 
 	if (fence)
 		*fence = radeon_fence_ref(ib.fence);
@@ -596,6 +584,10 @@ static int radeon_uvd_send_msg(struct radeon_device *rdev,
 	radeon_ib_free(rdev, &ib);
 	radeon_bo_unref(&bo);
 	return 0;
+
+err:
+	ttm_eu_backoff_reservation(&ticket, &head);
+	return r;
 }
 
 /* multiple fence commands without any stream commands in between can
@@ -691,11 +683,19 @@ static void radeon_uvd_idle_work_handler(struct work_struct *work)
 	struct radeon_device *rdev =
 		container_of(work, struct radeon_device, uvd.idle_work.work);
 
-	if (radeon_fence_count_emitted(rdev, R600_RING_TYPE_UVD_INDEX) == 0)
-		radeon_set_uvd_clocks(rdev, 0, 0);
-	else
+	if (radeon_fence_count_emitted(rdev, R600_RING_TYPE_UVD_INDEX) == 0) {
+		if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+			mutex_lock(&rdev->pm.mutex);
+			rdev->pm.dpm.uvd_active = false;
+			mutex_unlock(&rdev->pm.mutex);
+			radeon_pm_compute_clocks(rdev);
+		} else {
+			radeon_set_uvd_clocks(rdev, 0, 0);
+		}
+	} else {
 		schedule_delayed_work(&rdev->uvd.idle_work,
 				      msecs_to_jiffies(UVD_IDLE_TIMEOUT_MS));
+	}
 }
 
 void radeon_uvd_note_usage(struct radeon_device *rdev)
@@ -703,8 +703,14 @@ void radeon_uvd_note_usage(struct radeon_device *rdev)
 	bool set_clocks = !cancel_delayed_work_sync(&rdev->uvd.idle_work);
 	set_clocks &= schedule_delayed_work(&rdev->uvd.idle_work,
 					    msecs_to_jiffies(UVD_IDLE_TIMEOUT_MS));
-	if (set_clocks)
-		radeon_set_uvd_clocks(rdev, 53300, 40000);
+	if (set_clocks) {
+		if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+			/* XXX pick SD/HD/MVC */
+			radeon_dpm_enable_power_state(rdev, POWER_STATE_TYPE_INTERNAL_UVD);
+		} else {
+			radeon_set_uvd_clocks(rdev, 53300, 40000);
+		}
+	}
 }
 
 static unsigned radeon_uvd_calc_upll_post_div(unsigned vco_freq,
diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c
index 55880d5962c3..d8ddfb34545d 100644
--- a/drivers/gpu/drm/radeon/rs690.c
+++ b/drivers/gpu/drm/radeon/rs690.c
@@ -248,13 +248,16 @@ struct rs690_watermark {
 };
 
 static void rs690_crtc_bandwidth_compute(struct radeon_device *rdev,
-				  struct radeon_crtc *crtc,
-				  struct rs690_watermark *wm)
+					 struct radeon_crtc *crtc,
+					 struct rs690_watermark *wm,
+					 bool low)
 {
 	struct drm_display_mode *mode = &crtc->base.mode;
 	fixed20_12 a, b, c;
 	fixed20_12 pclk, request_fifo_depth, tolerable_latency, estimated_width;
 	fixed20_12 consumption_time, line_time, chunk_time, read_delay_latency;
+	fixed20_12 sclk, core_bandwidth, max_bandwidth;
+	u32 selected_sclk;
 
 	if (!crtc->base.enabled) {
 		/* FIXME: wouldn't it better to set priority mark to maximum */
@@ -262,6 +265,21 @@ static void rs690_crtc_bandwidth_compute(struct radeon_device *rdev,
 		return;
 	}
 
+	if (((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880)) &&
+	    (rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled)
+		selected_sclk = radeon_dpm_get_sclk(rdev, low);
+	else
+		selected_sclk = rdev->pm.current_sclk;
+
+	/* sclk in Mhz */
+	a.full = dfixed_const(100);
+	sclk.full = dfixed_const(selected_sclk);
+	sclk.full = dfixed_div(sclk, a);
+
+	/* core_bandwidth = sclk(Mhz) * 16 */
+	a.full = dfixed_const(16);
+	core_bandwidth.full = dfixed_div(rdev->pm.sclk, a);
+
 	if (crtc->vsc.full > dfixed_const(2))
 		wm->num_line_pair.full = dfixed_const(2);
 	else
@@ -322,36 +340,36 @@ static void rs690_crtc_bandwidth_compute(struct radeon_device *rdev,
 	wm->active_time.full = dfixed_div(wm->active_time, a);
 
 	/* Maximun bandwidth is the minimun bandwidth of all component */
-	rdev->pm.max_bandwidth = rdev->pm.core_bandwidth;
+	max_bandwidth = core_bandwidth;
 	if (rdev->mc.igp_sideport_enabled) {
-		if (rdev->pm.max_bandwidth.full > rdev->pm.sideport_bandwidth.full &&
+		if (max_bandwidth.full > rdev->pm.sideport_bandwidth.full &&
 			rdev->pm.sideport_bandwidth.full)
-			rdev->pm.max_bandwidth = rdev->pm.sideport_bandwidth;
+			max_bandwidth = rdev->pm.sideport_bandwidth;
 		read_delay_latency.full = dfixed_const(370 * 800 * 1000);
 		read_delay_latency.full = dfixed_div(read_delay_latency,
 			rdev->pm.igp_sideport_mclk);
 	} else {
-		if (rdev->pm.max_bandwidth.full > rdev->pm.k8_bandwidth.full &&
+		if (max_bandwidth.full > rdev->pm.k8_bandwidth.full &&
 			rdev->pm.k8_bandwidth.full)
-			rdev->pm.max_bandwidth = rdev->pm.k8_bandwidth;
-		if (rdev->pm.max_bandwidth.full > rdev->pm.ht_bandwidth.full &&
+			max_bandwidth = rdev->pm.k8_bandwidth;
+		if (max_bandwidth.full > rdev->pm.ht_bandwidth.full &&
 			rdev->pm.ht_bandwidth.full)
-			rdev->pm.max_bandwidth = rdev->pm.ht_bandwidth;
+			max_bandwidth = rdev->pm.ht_bandwidth;
 		read_delay_latency.full = dfixed_const(5000);
 	}
 
 	/* sclk = system clocks(ns) = 1000 / max_bandwidth / 16 */
 	a.full = dfixed_const(16);
-	rdev->pm.sclk.full = dfixed_mul(rdev->pm.max_bandwidth, a);
+	sclk.full = dfixed_mul(max_bandwidth, a);
 	a.full = dfixed_const(1000);
-	rdev->pm.sclk.full = dfixed_div(a, rdev->pm.sclk);
+	sclk.full = dfixed_div(a, sclk);
 	/* Determine chunk time
 	 * ChunkTime = the time it takes the DCP to send one chunk of data
 	 * to the LB which consists of pipeline delay and inter chunk gap
 	 * sclk = system clock(ns)
 	 */
 	a.full = dfixed_const(256 * 13);
-	chunk_time.full = dfixed_mul(rdev->pm.sclk, a);
+	chunk_time.full = dfixed_mul(sclk, a);
 	a.full = dfixed_const(10);
 	chunk_time.full = dfixed_div(chunk_time, a);
 
@@ -415,175 +433,200 @@ static void rs690_crtc_bandwidth_compute(struct radeon_device *rdev,
 	}
 }
 
-void rs690_bandwidth_update(struct radeon_device *rdev)
+static void rs690_compute_mode_priority(struct radeon_device *rdev,
+					struct rs690_watermark *wm0,
+					struct rs690_watermark *wm1,
+					struct drm_display_mode *mode0,
+					struct drm_display_mode *mode1,
+					u32 *d1mode_priority_a_cnt,
+					u32 *d2mode_priority_a_cnt)
 {
-	struct drm_display_mode *mode0 = NULL;
-	struct drm_display_mode *mode1 = NULL;
-	struct rs690_watermark wm0;
-	struct rs690_watermark wm1;
-	u32 tmp;
-	u32 d1mode_priority_a_cnt = S_006548_D1MODE_PRIORITY_A_OFF(1);
-	u32 d2mode_priority_a_cnt = S_006548_D1MODE_PRIORITY_A_OFF(1);
 	fixed20_12 priority_mark02, priority_mark12, fill_rate;
 	fixed20_12 a, b;
 
-	radeon_update_display_priority(rdev);
-
-	if (rdev->mode_info.crtcs[0]->base.enabled)
-		mode0 = &rdev->mode_info.crtcs[0]->base.mode;
-	if (rdev->mode_info.crtcs[1]->base.enabled)
-		mode1 = &rdev->mode_info.crtcs[1]->base.mode;
-	/*
-	 * Set display0/1 priority up in the memory controller for
-	 * modes if the user specifies HIGH for displaypriority
-	 * option.
-	 */
-	if ((rdev->disp_priority == 2) &&
-	    ((rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740))) {
-		tmp = RREG32_MC(R_000104_MC_INIT_MISC_LAT_TIMER);
-		tmp &= C_000104_MC_DISP0R_INIT_LAT;
-		tmp &= C_000104_MC_DISP1R_INIT_LAT;
-		if (mode0)
-			tmp |= S_000104_MC_DISP0R_INIT_LAT(1);
-		if (mode1)
-			tmp |= S_000104_MC_DISP1R_INIT_LAT(1);
-		WREG32_MC(R_000104_MC_INIT_MISC_LAT_TIMER, tmp);
-	}
-	rs690_line_buffer_adjust(rdev, mode0, mode1);
-
-	if ((rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740))
-		WREG32(R_006C9C_DCP_CONTROL, 0);
-	if ((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880))
-		WREG32(R_006C9C_DCP_CONTROL, 2);
-
-	rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0);
-	rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1);
-
-	tmp = (wm0.lb_request_fifo_depth - 1);
-	tmp |= (wm1.lb_request_fifo_depth - 1) << 16;
-	WREG32(R_006D58_LB_MAX_REQ_OUTSTANDING, tmp);
+	*d1mode_priority_a_cnt = S_006548_D1MODE_PRIORITY_A_OFF(1);
+	*d2mode_priority_a_cnt = S_006548_D1MODE_PRIORITY_A_OFF(1);
 
 	if (mode0 && mode1) {
-		if (dfixed_trunc(wm0.dbpp) > 64)
-			a.full = dfixed_mul(wm0.dbpp, wm0.num_line_pair);
+		if (dfixed_trunc(wm0->dbpp) > 64)
+			a.full = dfixed_mul(wm0->dbpp, wm0->num_line_pair);
 		else
-			a.full = wm0.num_line_pair.full;
-		if (dfixed_trunc(wm1.dbpp) > 64)
-			b.full = dfixed_mul(wm1.dbpp, wm1.num_line_pair);
+			a.full = wm0->num_line_pair.full;
+		if (dfixed_trunc(wm1->dbpp) > 64)
+			b.full = dfixed_mul(wm1->dbpp, wm1->num_line_pair);
 		else
-			b.full = wm1.num_line_pair.full;
+			b.full = wm1->num_line_pair.full;
 		a.full += b.full;
-		fill_rate.full = dfixed_div(wm0.sclk, a);
-		if (wm0.consumption_rate.full > fill_rate.full) {
-			b.full = wm0.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm0.active_time);
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+		fill_rate.full = dfixed_div(wm0->sclk, a);
+		if (wm0->consumption_rate.full > fill_rate.full) {
+			b.full = wm0->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm0->active_time);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			a.full = a.full + b.full;
 			b.full = dfixed_const(16 * 1000);
 			priority_mark02.full = dfixed_div(a, b);
 		} else {
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark02.full = dfixed_div(a, b);
 		}
-		if (wm1.consumption_rate.full > fill_rate.full) {
-			b.full = wm1.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm1.active_time);
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+		if (wm1->consumption_rate.full > fill_rate.full) {
+			b.full = wm1->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm1->active_time);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			a.full = a.full + b.full;
 			b.full = dfixed_const(16 * 1000);
 			priority_mark12.full = dfixed_div(a, b);
 		} else {
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark12.full = dfixed_div(a, b);
 		}
-		if (wm0.priority_mark.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark.full;
+		if (wm0->priority_mark.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark.full;
 		if (dfixed_trunc(priority_mark02) < 0)
 			priority_mark02.full = 0;
-		if (wm0.priority_mark_max.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark_max.full;
-		if (wm1.priority_mark.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark.full;
+		if (wm0->priority_mark_max.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark_max.full;
+		if (wm1->priority_mark.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark.full;
 		if (dfixed_trunc(priority_mark12) < 0)
 			priority_mark12.full = 0;
-		if (wm1.priority_mark_max.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark_max.full;
-		d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
-		d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
+		if (wm1->priority_mark_max.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark_max.full;
+		*d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
+		*d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
 		if (rdev->disp_priority == 2) {
-			d1mode_priority_a_cnt |= S_006548_D1MODE_PRIORITY_A_ALWAYS_ON(1);
-			d2mode_priority_a_cnt |= S_006D48_D2MODE_PRIORITY_A_ALWAYS_ON(1);
+			*d1mode_priority_a_cnt |= S_006548_D1MODE_PRIORITY_A_ALWAYS_ON(1);
+			*d2mode_priority_a_cnt |= S_006D48_D2MODE_PRIORITY_A_ALWAYS_ON(1);
 		}
 	} else if (mode0) {
-		if (dfixed_trunc(wm0.dbpp) > 64)
-			a.full = dfixed_mul(wm0.dbpp, wm0.num_line_pair);
+		if (dfixed_trunc(wm0->dbpp) > 64)
+			a.full = dfixed_mul(wm0->dbpp, wm0->num_line_pair);
 		else
-			a.full = wm0.num_line_pair.full;
-		fill_rate.full = dfixed_div(wm0.sclk, a);
-		if (wm0.consumption_rate.full > fill_rate.full) {
-			b.full = wm0.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm0.active_time);
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = wm0->num_line_pair.full;
+		fill_rate.full = dfixed_div(wm0->sclk, a);
+		if (wm0->consumption_rate.full > fill_rate.full) {
+			b.full = wm0->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm0->active_time);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			a.full = a.full + b.full;
 			b.full = dfixed_const(16 * 1000);
 			priority_mark02.full = dfixed_div(a, b);
 		} else {
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark02.full = dfixed_div(a, b);
 		}
-		if (wm0.priority_mark.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark.full;
+		if (wm0->priority_mark.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark.full;
 		if (dfixed_trunc(priority_mark02) < 0)
 			priority_mark02.full = 0;
-		if (wm0.priority_mark_max.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark_max.full;
-		d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
+		if (wm0->priority_mark_max.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark_max.full;
+		*d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
 		if (rdev->disp_priority == 2)
-			d1mode_priority_a_cnt |= S_006548_D1MODE_PRIORITY_A_ALWAYS_ON(1);
+			*d1mode_priority_a_cnt |= S_006548_D1MODE_PRIORITY_A_ALWAYS_ON(1);
 	} else if (mode1) {
-		if (dfixed_trunc(wm1.dbpp) > 64)
-			a.full = dfixed_mul(wm1.dbpp, wm1.num_line_pair);
+		if (dfixed_trunc(wm1->dbpp) > 64)
+			a.full = dfixed_mul(wm1->dbpp, wm1->num_line_pair);
 		else
-			a.full = wm1.num_line_pair.full;
-		fill_rate.full = dfixed_div(wm1.sclk, a);
-		if (wm1.consumption_rate.full > fill_rate.full) {
-			b.full = wm1.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm1.active_time);
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = wm1->num_line_pair.full;
+		fill_rate.full = dfixed_div(wm1->sclk, a);
+		if (wm1->consumption_rate.full > fill_rate.full) {
+			b.full = wm1->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm1->active_time);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			a.full = a.full + b.full;
 			b.full = dfixed_const(16 * 1000);
 			priority_mark12.full = dfixed_div(a, b);
 		} else {
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark12.full = dfixed_div(a, b);
 		}
-		if (wm1.priority_mark.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark.full;
+		if (wm1->priority_mark.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark.full;
 		if (dfixed_trunc(priority_mark12) < 0)
 			priority_mark12.full = 0;
-		if (wm1.priority_mark_max.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark_max.full;
-		d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
+		if (wm1->priority_mark_max.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark_max.full;
+		*d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
 		if (rdev->disp_priority == 2)
-			d2mode_priority_a_cnt |= S_006D48_D2MODE_PRIORITY_A_ALWAYS_ON(1);
+			*d2mode_priority_a_cnt |= S_006D48_D2MODE_PRIORITY_A_ALWAYS_ON(1);
 	}
+}
+
+void rs690_bandwidth_update(struct radeon_device *rdev)
+{
+	struct drm_display_mode *mode0 = NULL;
+	struct drm_display_mode *mode1 = NULL;
+	struct rs690_watermark wm0_high, wm0_low;
+	struct rs690_watermark wm1_high, wm1_low;
+	u32 tmp;
+	u32 d1mode_priority_a_cnt, d1mode_priority_b_cnt;
+	u32 d2mode_priority_a_cnt, d2mode_priority_b_cnt;
+
+	radeon_update_display_priority(rdev);
+
+	if (rdev->mode_info.crtcs[0]->base.enabled)
+		mode0 = &rdev->mode_info.crtcs[0]->base.mode;
+	if (rdev->mode_info.crtcs[1]->base.enabled)
+		mode1 = &rdev->mode_info.crtcs[1]->base.mode;
+	/*
+	 * Set display0/1 priority up in the memory controller for
+	 * modes if the user specifies HIGH for displaypriority
+	 * option.
+	 */
+	if ((rdev->disp_priority == 2) &&
+	    ((rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740))) {
+		tmp = RREG32_MC(R_000104_MC_INIT_MISC_LAT_TIMER);
+		tmp &= C_000104_MC_DISP0R_INIT_LAT;
+		tmp &= C_000104_MC_DISP1R_INIT_LAT;
+		if (mode0)
+			tmp |= S_000104_MC_DISP0R_INIT_LAT(1);
+		if (mode1)
+			tmp |= S_000104_MC_DISP1R_INIT_LAT(1);
+		WREG32_MC(R_000104_MC_INIT_MISC_LAT_TIMER, tmp);
+	}
+	rs690_line_buffer_adjust(rdev, mode0, mode1);
+
+	if ((rdev->family == CHIP_RS690) || (rdev->family == CHIP_RS740))
+		WREG32(R_006C9C_DCP_CONTROL, 0);
+	if ((rdev->family == CHIP_RS780) || (rdev->family == CHIP_RS880))
+		WREG32(R_006C9C_DCP_CONTROL, 2);
+
+	rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0_high, false);
+	rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1_high, false);
+
+	rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0_low, true);
+	rs690_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1_low, true);
+
+	tmp = (wm0_high.lb_request_fifo_depth - 1);
+	tmp |= (wm1_high.lb_request_fifo_depth - 1) << 16;
+	WREG32(R_006D58_LB_MAX_REQ_OUTSTANDING, tmp);
+
+	rs690_compute_mode_priority(rdev,
+				    &wm0_high, &wm1_high,
+				    mode0, mode1,
+				    &d1mode_priority_a_cnt, &d2mode_priority_a_cnt);
+	rs690_compute_mode_priority(rdev,
+				    &wm0_low, &wm1_low,
+				    mode0, mode1,
+				    &d1mode_priority_b_cnt, &d2mode_priority_b_cnt);
 
 	WREG32(R_006548_D1MODE_PRIORITY_A_CNT, d1mode_priority_a_cnt);
-	WREG32(R_00654C_D1MODE_PRIORITY_B_CNT, d1mode_priority_a_cnt);
+	WREG32(R_00654C_D1MODE_PRIORITY_B_CNT, d1mode_priority_b_cnt);
 	WREG32(R_006D48_D2MODE_PRIORITY_A_CNT, d2mode_priority_a_cnt);
-	WREG32(R_006D4C_D2MODE_PRIORITY_B_CNT, d2mode_priority_a_cnt);
+	WREG32(R_006D4C_D2MODE_PRIORITY_B_CNT, d2mode_priority_b_cnt);
 }
 
 uint32_t rs690_mc_rreg(struct radeon_device *rdev, uint32_t reg)
diff --git a/drivers/gpu/drm/radeon/rs780_dpm.c b/drivers/gpu/drm/radeon/rs780_dpm.c
new file mode 100644
index 000000000000..d1a1ce73bd45
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rs780_dpm.c
@@ -0,0 +1,988 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "rs780d.h"
+#include "r600_dpm.h"
+#include "rs780_dpm.h"
+#include "atom.h"
+#include <linux/seq_file.h>
+
+static struct igp_ps *rs780_get_ps(struct radeon_ps *rps)
+{
+	struct igp_ps *ps = rps->ps_priv;
+
+	return ps;
+}
+
+static struct igp_power_info *rs780_get_pi(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rdev->pm.dpm.priv;
+
+	return pi;
+}
+
+static void rs780_get_pm_mode_parameters(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+	struct radeon_mode_info *minfo = &rdev->mode_info;
+	struct drm_crtc *crtc;
+	struct radeon_crtc *radeon_crtc;
+	int i;
+
+	/* defaults */
+	pi->crtc_id = 0;
+	pi->refresh_rate = 60;
+
+	for (i = 0; i < rdev->num_crtc; i++) {
+		crtc = (struct drm_crtc *)minfo->crtcs[i];
+		if (crtc && crtc->enabled) {
+			radeon_crtc = to_radeon_crtc(crtc);
+			pi->crtc_id = radeon_crtc->crtc_id;
+			if (crtc->mode.htotal && crtc->mode.vtotal)
+				pi->refresh_rate =
+					(crtc->mode.clock * 1000) /
+					(crtc->mode.htotal * crtc->mode.vtotal);
+			break;
+		}
+	}
+}
+
+static void rs780_voltage_scaling_enable(struct radeon_device *rdev, bool enable);
+
+static int rs780_initialize_dpm_power_state(struct radeon_device *rdev,
+					    struct radeon_ps *boot_ps)
+{
+	struct atom_clock_dividers dividers;
+	struct igp_ps *default_state = rs780_get_ps(boot_ps);
+	int i, ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     default_state->sclk_low, false, &dividers);
+	if (ret)
+		return ret;
+
+	r600_engine_clock_entry_set_reference_divider(rdev, 0, dividers.ref_div);
+	r600_engine_clock_entry_set_feedback_divider(rdev, 0, dividers.fb_div);
+	r600_engine_clock_entry_set_post_divider(rdev, 0, dividers.post_div);
+
+	if (dividers.enable_post_div)
+		r600_engine_clock_entry_enable_post_divider(rdev, 0, true);
+	else
+		r600_engine_clock_entry_enable_post_divider(rdev, 0, false);
+
+	r600_engine_clock_entry_set_step_time(rdev, 0, R600_SST_DFLT);
+	r600_engine_clock_entry_enable_pulse_skipping(rdev, 0, false);
+
+	r600_engine_clock_entry_enable(rdev, 0, true);
+	for (i = 1; i < R600_PM_NUMBER_OF_SCLKS; i++)
+		r600_engine_clock_entry_enable(rdev, i, false);
+
+	r600_enable_mclk_control(rdev, false);
+	r600_voltage_control_enable_pins(rdev, 0);
+
+	return 0;
+}
+
+static int rs780_initialize_dpm_parameters(struct radeon_device *rdev,
+					   struct radeon_ps *boot_ps)
+{
+	int ret = 0;
+	int i;
+
+	r600_set_bsp(rdev, R600_BSU_DFLT, R600_BSP_DFLT);
+
+	r600_set_at(rdev, 0, 0, 0, 0);
+
+	r600_set_git(rdev, R600_GICST_DFLT);
+
+	for (i = 0; i < R600_PM_NUMBER_OF_TC; i++)
+		r600_set_tc(rdev, i, 0, 0);
+
+	r600_select_td(rdev, R600_TD_DFLT);
+	r600_set_vrc(rdev, 0);
+
+	r600_set_tpu(rdev, R600_TPU_DFLT);
+	r600_set_tpc(rdev, R600_TPC_DFLT);
+
+	r600_set_sstu(rdev, R600_SSTU_DFLT);
+	r600_set_sst(rdev, R600_SST_DFLT);
+
+	r600_set_fctu(rdev, R600_FCTU_DFLT);
+	r600_set_fct(rdev, R600_FCT_DFLT);
+
+	r600_set_vddc3d_oorsu(rdev, R600_VDDC3DOORSU_DFLT);
+	r600_set_vddc3d_oorphc(rdev, R600_VDDC3DOORPHC_DFLT);
+	r600_set_vddc3d_oorsdc(rdev, R600_VDDC3DOORSDC_DFLT);
+	r600_set_ctxcgtt3d_rphc(rdev, R600_CTXCGTT3DRPHC_DFLT);
+	r600_set_ctxcgtt3d_rsdc(rdev, R600_CTXCGTT3DRSDC_DFLT);
+
+	r600_vid_rt_set_vru(rdev, R600_VRU_DFLT);
+	r600_vid_rt_set_vrt(rdev, R600_VOLTAGERESPONSETIME_DFLT);
+	r600_vid_rt_set_ssu(rdev, R600_SPLLSTEPUNIT_DFLT);
+
+	ret = rs780_initialize_dpm_power_state(rdev, boot_ps);
+
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_LOW,     0);
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_MEDIUM,  0);
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_HIGH,    0);
+
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_LOW,    0);
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_MEDIUM, 0);
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_HIGH,   0);
+
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_LOW,    0);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_MEDIUM, 0);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_HIGH,   0);
+
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_LOW,    R600_DISPLAY_WATERMARK_HIGH);
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_MEDIUM, R600_DISPLAY_WATERMARK_HIGH);
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_HIGH,   R600_DISPLAY_WATERMARK_HIGH);
+
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_CTXSW, false);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_HIGH, false);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, false);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, true);
+
+	r600_power_level_set_enter_index(rdev, R600_POWER_LEVEL_LOW);
+
+	r600_set_vrc(rdev, RS780_CGFTV_DFLT);
+
+	return ret;
+}
+
+static void rs780_start_dpm(struct radeon_device *rdev)
+{
+	r600_enable_sclk_control(rdev, false);
+	r600_enable_mclk_control(rdev, false);
+
+	r600_dynamicpm_enable(rdev, true);
+
+	radeon_wait_for_vblank(rdev, 0);
+	radeon_wait_for_vblank(rdev, 1);
+
+	r600_enable_spll_bypass(rdev, true);
+	r600_wait_for_spll_change(rdev);
+	r600_enable_spll_bypass(rdev, false);
+	r600_wait_for_spll_change(rdev);
+
+	r600_enable_spll_bypass(rdev, true);
+	r600_wait_for_spll_change(rdev);
+	r600_enable_spll_bypass(rdev, false);
+	r600_wait_for_spll_change(rdev);
+
+	r600_enable_sclk_control(rdev, true);
+}
+
+
+static void rs780_preset_ranges_slow_clk_fbdiv_en(struct radeon_device *rdev)
+{
+	WREG32_P(FVTHROT_SLOW_CLK_FEEDBACK_DIV_REG1, RANGE_SLOW_CLK_FEEDBACK_DIV_EN,
+		 ~RANGE_SLOW_CLK_FEEDBACK_DIV_EN);
+
+	WREG32_P(FVTHROT_SLOW_CLK_FEEDBACK_DIV_REG1,
+		 RANGE0_SLOW_CLK_FEEDBACK_DIV(RS780_SLOWCLKFEEDBACKDIV_DFLT),
+		 ~RANGE0_SLOW_CLK_FEEDBACK_DIV_MASK);
+}
+
+static void rs780_preset_starting_fbdiv(struct radeon_device *rdev)
+{
+	u32 fbdiv = (RREG32(CG_SPLL_FUNC_CNTL) & SPLL_FB_DIV_MASK) >> SPLL_FB_DIV_SHIFT;
+
+	WREG32_P(FVTHROT_FBDIV_REG1, STARTING_FEEDBACK_DIV(fbdiv),
+		 ~STARTING_FEEDBACK_DIV_MASK);
+
+	WREG32_P(FVTHROT_FBDIV_REG2, FORCED_FEEDBACK_DIV(fbdiv),
+		 ~FORCED_FEEDBACK_DIV_MASK);
+
+	WREG32_P(FVTHROT_FBDIV_REG1, FORCE_FEEDBACK_DIV, ~FORCE_FEEDBACK_DIV);
+}
+
+static void rs780_voltage_scaling_init(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+	struct drm_device *dev = rdev->ddev;
+	u32 fv_throt_pwm_fb_div_range[3];
+	u32 fv_throt_pwm_range[4];
+
+	if (dev->pdev->device == 0x9614) {
+		fv_throt_pwm_fb_div_range[0] = RS780D_FVTHROTPWMFBDIVRANGEREG0_DFLT;
+		fv_throt_pwm_fb_div_range[1] = RS780D_FVTHROTPWMFBDIVRANGEREG1_DFLT;
+		fv_throt_pwm_fb_div_range[2] = RS780D_FVTHROTPWMFBDIVRANGEREG2_DFLT;
+	} else if ((dev->pdev->device == 0x9714) ||
+		   (dev->pdev->device == 0x9715)) {
+		fv_throt_pwm_fb_div_range[0] = RS880D_FVTHROTPWMFBDIVRANGEREG0_DFLT;
+		fv_throt_pwm_fb_div_range[1] = RS880D_FVTHROTPWMFBDIVRANGEREG1_DFLT;
+		fv_throt_pwm_fb_div_range[2] = RS880D_FVTHROTPWMFBDIVRANGEREG2_DFLT;
+	} else {
+		fv_throt_pwm_fb_div_range[0] = RS780_FVTHROTPWMFBDIVRANGEREG0_DFLT;
+		fv_throt_pwm_fb_div_range[1] = RS780_FVTHROTPWMFBDIVRANGEREG1_DFLT;
+		fv_throt_pwm_fb_div_range[2] = RS780_FVTHROTPWMFBDIVRANGEREG2_DFLT;
+	}
+
+	if (pi->pwm_voltage_control) {
+		fv_throt_pwm_range[0] = pi->min_voltage;
+		fv_throt_pwm_range[1] = pi->min_voltage;
+		fv_throt_pwm_range[2] = pi->max_voltage;
+		fv_throt_pwm_range[3] = pi->max_voltage;
+	} else {
+		fv_throt_pwm_range[0] = pi->invert_pwm_required ?
+			RS780_FVTHROTPWMRANGE3_GPIO_DFLT : RS780_FVTHROTPWMRANGE0_GPIO_DFLT;
+		fv_throt_pwm_range[1] = pi->invert_pwm_required ?
+			RS780_FVTHROTPWMRANGE2_GPIO_DFLT : RS780_FVTHROTPWMRANGE1_GPIO_DFLT;
+		fv_throt_pwm_range[2] = pi->invert_pwm_required ?
+			RS780_FVTHROTPWMRANGE1_GPIO_DFLT : RS780_FVTHROTPWMRANGE2_GPIO_DFLT;
+		fv_throt_pwm_range[3] = pi->invert_pwm_required ?
+			RS780_FVTHROTPWMRANGE0_GPIO_DFLT : RS780_FVTHROTPWMRANGE3_GPIO_DFLT;
+	}
+
+	WREG32_P(FVTHROT_PWM_CTRL_REG0,
+		 STARTING_PWM_HIGHTIME(pi->max_voltage),
+		 ~STARTING_PWM_HIGHTIME_MASK);
+
+	WREG32_P(FVTHROT_PWM_CTRL_REG0,
+		 NUMBER_OF_CYCLES_IN_PERIOD(pi->num_of_cycles_in_period),
+		 ~NUMBER_OF_CYCLES_IN_PERIOD_MASK);
+
+	WREG32_P(FVTHROT_PWM_CTRL_REG0, FORCE_STARTING_PWM_HIGHTIME,
+		 ~FORCE_STARTING_PWM_HIGHTIME);
+
+	if (pi->invert_pwm_required)
+		WREG32_P(FVTHROT_PWM_CTRL_REG0, INVERT_PWM_WAVEFORM, ~INVERT_PWM_WAVEFORM);
+	else
+		WREG32_P(FVTHROT_PWM_CTRL_REG0, 0, ~INVERT_PWM_WAVEFORM);
+
+	rs780_voltage_scaling_enable(rdev, true);
+
+	WREG32(FVTHROT_PWM_CTRL_REG1,
+	       (MIN_PWM_HIGHTIME(pi->min_voltage) |
+		MAX_PWM_HIGHTIME(pi->max_voltage)));
+
+	WREG32(FVTHROT_PWM_US_REG0, RS780_FVTHROTPWMUSREG0_DFLT);
+	WREG32(FVTHROT_PWM_US_REG1, RS780_FVTHROTPWMUSREG1_DFLT);
+	WREG32(FVTHROT_PWM_DS_REG0, RS780_FVTHROTPWMDSREG0_DFLT);
+	WREG32(FVTHROT_PWM_DS_REG1, RS780_FVTHROTPWMDSREG1_DFLT);
+
+	WREG32_P(FVTHROT_PWM_FEEDBACK_DIV_REG1,
+		 RANGE0_PWM_FEEDBACK_DIV(fv_throt_pwm_fb_div_range[0]),
+		 ~RANGE0_PWM_FEEDBACK_DIV_MASK);
+
+	WREG32(FVTHROT_PWM_FEEDBACK_DIV_REG2,
+	       (RANGE1_PWM_FEEDBACK_DIV(fv_throt_pwm_fb_div_range[1]) |
+		RANGE2_PWM_FEEDBACK_DIV(fv_throt_pwm_fb_div_range[2])));
+
+	WREG32(FVTHROT_PWM_FEEDBACK_DIV_REG3,
+	       (RANGE0_PWM(fv_throt_pwm_range[1]) |
+		RANGE1_PWM(fv_throt_pwm_range[2])));
+	WREG32(FVTHROT_PWM_FEEDBACK_DIV_REG4,
+	       (RANGE2_PWM(fv_throt_pwm_range[1]) |
+		RANGE3_PWM(fv_throt_pwm_range[2])));
+}
+
+static void rs780_clk_scaling_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(FVTHROT_CNTRL_REG, ENABLE_FV_THROT | ENABLE_FV_UPDATE,
+			 ~(ENABLE_FV_THROT | ENABLE_FV_UPDATE));
+	else
+		WREG32_P(FVTHROT_CNTRL_REG, 0,
+			 ~(ENABLE_FV_THROT | ENABLE_FV_UPDATE));
+}
+
+static void rs780_voltage_scaling_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(FVTHROT_CNTRL_REG, ENABLE_FV_THROT_IO, ~ENABLE_FV_THROT_IO);
+	else
+		WREG32_P(FVTHROT_CNTRL_REG, 0, ~ENABLE_FV_THROT_IO);
+}
+
+static void rs780_set_engine_clock_wfc(struct radeon_device *rdev)
+{
+	WREG32(FVTHROT_UTC0, RS780_FVTHROTUTC0_DFLT);
+	WREG32(FVTHROT_UTC1, RS780_FVTHROTUTC1_DFLT);
+	WREG32(FVTHROT_UTC2, RS780_FVTHROTUTC2_DFLT);
+	WREG32(FVTHROT_UTC3, RS780_FVTHROTUTC3_DFLT);
+	WREG32(FVTHROT_UTC4, RS780_FVTHROTUTC4_DFLT);
+
+	WREG32(FVTHROT_DTC0, RS780_FVTHROTDTC0_DFLT);
+	WREG32(FVTHROT_DTC1, RS780_FVTHROTDTC1_DFLT);
+	WREG32(FVTHROT_DTC2, RS780_FVTHROTDTC2_DFLT);
+	WREG32(FVTHROT_DTC3, RS780_FVTHROTDTC3_DFLT);
+	WREG32(FVTHROT_DTC4, RS780_FVTHROTDTC4_DFLT);
+}
+
+static void rs780_set_engine_clock_sc(struct radeon_device *rdev)
+{
+	WREG32_P(FVTHROT_FBDIV_REG2,
+		 FB_DIV_TIMER_VAL(RS780_FBDIVTIMERVAL_DFLT),
+		 ~FB_DIV_TIMER_VAL_MASK);
+
+	WREG32_P(FVTHROT_CNTRL_REG,
+		 REFRESH_RATE_DIVISOR(0) | MINIMUM_CIP(0xf),
+		 ~(REFRESH_RATE_DIVISOR_MASK | MINIMUM_CIP_MASK));
+}
+
+static void rs780_set_engine_clock_tdc(struct radeon_device *rdev)
+{
+	WREG32_P(FVTHROT_CNTRL_REG, 0, ~(FORCE_TREND_SEL | TREND_SEL_MODE));
+}
+
+static void rs780_set_engine_clock_ssc(struct radeon_device *rdev)
+{
+	WREG32(FVTHROT_FB_US_REG0, RS780_FVTHROTFBUSREG0_DFLT);
+	WREG32(FVTHROT_FB_US_REG1, RS780_FVTHROTFBUSREG1_DFLT);
+	WREG32(FVTHROT_FB_DS_REG0, RS780_FVTHROTFBDSREG0_DFLT);
+	WREG32(FVTHROT_FB_DS_REG1, RS780_FVTHROTFBDSREG1_DFLT);
+
+	WREG32_P(FVTHROT_FBDIV_REG1, MAX_FEEDBACK_STEP(1), ~MAX_FEEDBACK_STEP_MASK);
+}
+
+static void rs780_program_at(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+
+	WREG32(FVTHROT_TARGET_REG, 30000000 / pi->refresh_rate);
+	WREG32(FVTHROT_CB1, 1000000 * 5 / pi->refresh_rate);
+	WREG32(FVTHROT_CB2, 1000000 * 10 / pi->refresh_rate);
+	WREG32(FVTHROT_CB3, 1000000 * 30 / pi->refresh_rate);
+	WREG32(FVTHROT_CB4, 1000000 * 50 / pi->refresh_rate);
+}
+
+static void rs780_disable_vbios_powersaving(struct radeon_device *rdev)
+{
+	WREG32_P(CG_INTGFX_MISC, 0, ~0xFFF00000);
+}
+
+static void rs780_force_voltage_to_high(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+	struct igp_ps *current_state = rs780_get_ps(rdev->pm.dpm.current_ps);
+
+	if ((current_state->max_voltage == RS780_VDDC_LEVEL_HIGH) &&
+	    (current_state->min_voltage == RS780_VDDC_LEVEL_HIGH))
+		return;
+
+	WREG32_P(GFX_MACRO_BYPASS_CNTL, SPLL_BYPASS_CNTL, ~SPLL_BYPASS_CNTL);
+
+	udelay(1);
+
+	WREG32_P(FVTHROT_PWM_CTRL_REG0,
+		 STARTING_PWM_HIGHTIME(pi->max_voltage),
+		 ~STARTING_PWM_HIGHTIME_MASK);
+
+	WREG32_P(FVTHROT_PWM_CTRL_REG0,
+		 FORCE_STARTING_PWM_HIGHTIME, ~FORCE_STARTING_PWM_HIGHTIME);
+
+	WREG32_P(FVTHROT_PWM_FEEDBACK_DIV_REG1, 0,
+		~RANGE_PWM_FEEDBACK_DIV_EN);
+
+	udelay(1);
+
+	WREG32_P(GFX_MACRO_BYPASS_CNTL, 0, ~SPLL_BYPASS_CNTL);
+}
+
+static int rs780_set_engine_clock_scaling(struct radeon_device *rdev,
+					  struct radeon_ps *new_ps,
+					  struct radeon_ps *old_ps)
+{
+	struct atom_clock_dividers min_dividers, max_dividers, current_max_dividers;
+	struct igp_ps *new_state = rs780_get_ps(new_ps);
+	struct igp_ps *old_state = rs780_get_ps(old_ps);
+	int ret;
+
+	if ((new_state->sclk_high == old_state->sclk_high) &&
+	    (new_state->sclk_low == old_state->sclk_low))
+		return 0;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     new_state->sclk_low, false, &min_dividers);
+	if (ret)
+		return ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     new_state->sclk_high, false, &max_dividers);
+	if (ret)
+		return ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     old_state->sclk_high, false, &current_max_dividers);
+	if (ret)
+		return ret;
+
+	WREG32_P(GFX_MACRO_BYPASS_CNTL, SPLL_BYPASS_CNTL, ~SPLL_BYPASS_CNTL);
+
+	WREG32_P(FVTHROT_FBDIV_REG2, FORCED_FEEDBACK_DIV(max_dividers.fb_div),
+		 ~FORCED_FEEDBACK_DIV_MASK);
+	WREG32_P(FVTHROT_FBDIV_REG1, STARTING_FEEDBACK_DIV(max_dividers.fb_div),
+		 ~STARTING_FEEDBACK_DIV_MASK);
+	WREG32_P(FVTHROT_FBDIV_REG1, FORCE_FEEDBACK_DIV, ~FORCE_FEEDBACK_DIV);
+
+	udelay(100);
+
+	WREG32_P(GFX_MACRO_BYPASS_CNTL, 0, ~SPLL_BYPASS_CNTL);
+
+	if (max_dividers.fb_div > min_dividers.fb_div) {
+		WREG32_P(FVTHROT_FBDIV_REG0,
+			 MIN_FEEDBACK_DIV(min_dividers.fb_div) |
+			 MAX_FEEDBACK_DIV(max_dividers.fb_div),
+			 ~(MIN_FEEDBACK_DIV_MASK | MAX_FEEDBACK_DIV_MASK));
+
+		WREG32_P(FVTHROT_FBDIV_REG1, 0, ~FORCE_FEEDBACK_DIV);
+	}
+
+	return 0;
+}
+
+static void rs780_set_engine_clock_spc(struct radeon_device *rdev,
+				       struct radeon_ps *new_ps,
+				       struct radeon_ps *old_ps)
+{
+	struct igp_ps *new_state = rs780_get_ps(new_ps);
+	struct igp_ps *old_state = rs780_get_ps(old_ps);
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+
+	if ((new_state->sclk_high == old_state->sclk_high) &&
+	    (new_state->sclk_low == old_state->sclk_low))
+		return;
+
+	if (pi->crtc_id == 0)
+		WREG32_P(CG_INTGFX_MISC, 0, ~FVTHROT_VBLANK_SEL);
+	else
+		WREG32_P(CG_INTGFX_MISC, FVTHROT_VBLANK_SEL, ~FVTHROT_VBLANK_SEL);
+
+}
+
+static void rs780_activate_engine_clk_scaling(struct radeon_device *rdev,
+					      struct radeon_ps *new_ps,
+					      struct radeon_ps *old_ps)
+{
+	struct igp_ps *new_state = rs780_get_ps(new_ps);
+	struct igp_ps *old_state = rs780_get_ps(old_ps);
+
+	if ((new_state->sclk_high == old_state->sclk_high) &&
+	    (new_state->sclk_low == old_state->sclk_low))
+		return;
+
+	rs780_clk_scaling_enable(rdev, true);
+}
+
+static u32 rs780_get_voltage_for_vddc_level(struct radeon_device *rdev,
+					    enum rs780_vddc_level vddc)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+
+	if (vddc == RS780_VDDC_LEVEL_HIGH)
+		return pi->max_voltage;
+	else if (vddc == RS780_VDDC_LEVEL_LOW)
+		return pi->min_voltage;
+	else
+		return pi->max_voltage;
+}
+
+static void rs780_enable_voltage_scaling(struct radeon_device *rdev,
+					 struct radeon_ps *new_ps)
+{
+	struct igp_ps *new_state = rs780_get_ps(new_ps);
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+	enum rs780_vddc_level vddc_high, vddc_low;
+
+	udelay(100);
+
+	if ((new_state->max_voltage == RS780_VDDC_LEVEL_HIGH) &&
+	    (new_state->min_voltage == RS780_VDDC_LEVEL_HIGH))
+		return;
+
+	vddc_high = rs780_get_voltage_for_vddc_level(rdev,
+						     new_state->max_voltage);
+	vddc_low = rs780_get_voltage_for_vddc_level(rdev,
+						    new_state->min_voltage);
+
+	WREG32_P(GFX_MACRO_BYPASS_CNTL, SPLL_BYPASS_CNTL, ~SPLL_BYPASS_CNTL);
+
+	udelay(1);
+	if (vddc_high > vddc_low) {
+		WREG32_P(FVTHROT_PWM_FEEDBACK_DIV_REG1,
+			 RANGE_PWM_FEEDBACK_DIV_EN, ~RANGE_PWM_FEEDBACK_DIV_EN);
+
+		WREG32_P(FVTHROT_PWM_CTRL_REG0, 0, ~FORCE_STARTING_PWM_HIGHTIME);
+	} else if (vddc_high == vddc_low) {
+		if (pi->max_voltage != vddc_high) {
+			WREG32_P(FVTHROT_PWM_CTRL_REG0,
+				 STARTING_PWM_HIGHTIME(vddc_high),
+				 ~STARTING_PWM_HIGHTIME_MASK);
+
+			WREG32_P(FVTHROT_PWM_CTRL_REG0,
+				 FORCE_STARTING_PWM_HIGHTIME,
+				 ~FORCE_STARTING_PWM_HIGHTIME);
+		}
+	}
+
+	WREG32_P(GFX_MACRO_BYPASS_CNTL, 0, ~SPLL_BYPASS_CNTL);
+}
+
+static void rs780_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+						     struct radeon_ps *new_ps,
+						     struct radeon_ps *old_ps)
+{
+	struct igp_ps *new_state = rs780_get_ps(new_ps);
+	struct igp_ps *current_state = rs780_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->sclk_high >= current_state->sclk_high)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+static void rs780_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+						    struct radeon_ps *new_ps,
+						    struct radeon_ps *old_ps)
+{
+	struct igp_ps *new_state = rs780_get_ps(new_ps);
+	struct igp_ps *current_state = rs780_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->sclk_high < current_state->sclk_high)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+int rs780_dpm_enable(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	rs780_get_pm_mode_parameters(rdev);
+	rs780_disable_vbios_powersaving(rdev);
+
+	if (r600_dynamicpm_enabled(rdev))
+		return -EINVAL;
+	ret = rs780_initialize_dpm_parameters(rdev, boot_ps);
+	if (ret)
+		return ret;
+	rs780_start_dpm(rdev);
+
+	rs780_preset_ranges_slow_clk_fbdiv_en(rdev);
+	rs780_preset_starting_fbdiv(rdev);
+	if (pi->voltage_control)
+		rs780_voltage_scaling_init(rdev);
+	rs780_clk_scaling_enable(rdev, true);
+	rs780_set_engine_clock_sc(rdev);
+	rs780_set_engine_clock_wfc(rdev);
+	rs780_program_at(rdev);
+	rs780_set_engine_clock_tdc(rdev);
+	rs780_set_engine_clock_ssc(rdev);
+
+	if (pi->gfx_clock_gating)
+		r600_gfx_clockgating_enable(rdev, true);
+
+	if (rdev->irq.installed && (rdev->pm.int_thermal_type == THERMAL_TYPE_RV6XX)) {
+		ret = r600_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+	}
+
+	return 0;
+}
+
+void rs780_dpm_disable(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+
+	r600_dynamicpm_enable(rdev, false);
+
+	rs780_clk_scaling_enable(rdev, false);
+	rs780_voltage_scaling_enable(rdev, false);
+
+	if (pi->gfx_clock_gating)
+		r600_gfx_clockgating_enable(rdev, false);
+
+	if (rdev->irq.installed &&
+	    (rdev->pm.int_thermal_type == THERMAL_TYPE_RV6XX)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+}
+
+int rs780_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+	struct radeon_ps *new_ps = rdev->pm.dpm.requested_ps;
+	struct radeon_ps *old_ps = rdev->pm.dpm.current_ps;
+	int ret;
+
+	rs780_get_pm_mode_parameters(rdev);
+
+	rs780_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+
+	if (pi->voltage_control) {
+		rs780_force_voltage_to_high(rdev);
+		mdelay(5);
+	}
+
+	ret = rs780_set_engine_clock_scaling(rdev, new_ps, old_ps);
+	if (ret)
+		return ret;
+	rs780_set_engine_clock_spc(rdev, new_ps, old_ps);
+
+	rs780_activate_engine_clk_scaling(rdev, new_ps, old_ps);
+
+	if (pi->voltage_control)
+		rs780_enable_voltage_scaling(rdev, new_ps);
+
+	rs780_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+
+	return 0;
+}
+
+void rs780_dpm_setup_asic(struct radeon_device *rdev)
+{
+
+}
+
+void rs780_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+	rs780_get_pm_mode_parameters(rdev);
+	rs780_program_at(rdev);
+}
+
+union igp_info {
+	struct _ATOM_INTEGRATED_SYSTEM_INFO info;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V2 info_2;
+};
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void rs780_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					     struct radeon_ps *rps,
+					     struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info,
+					     u8 table_rev)
+{
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) {
+		rps->vclk = le32_to_cpu(non_clock_info->ulVCLK);
+		rps->dclk = le32_to_cpu(non_clock_info->ulDCLK);
+	} else if (r600_is_uvd_state(rps->class, rps->class2)) {
+		rps->vclk = RS780_DEFAULT_VCLK_FREQ;
+		rps->dclk = RS780_DEFAULT_DCLK_FREQ;
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT)
+		rdev->pm.dpm.boot_ps = rps;
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void rs780_parse_pplib_clock_info(struct radeon_device *rdev,
+					 struct radeon_ps *rps,
+					 union pplib_clock_info *clock_info)
+{
+	struct igp_ps *ps = rs780_get_ps(rps);
+	u32 sclk;
+
+	sclk = le16_to_cpu(clock_info->rs780.usLowEngineClockLow);
+	sclk |= clock_info->rs780.ucLowEngineClockHigh << 16;
+	ps->sclk_low = sclk;
+	sclk = le16_to_cpu(clock_info->rs780.usHighEngineClockLow);
+	sclk |= clock_info->rs780.ucHighEngineClockHigh << 16;
+	ps->sclk_high = sclk;
+	switch (le16_to_cpu(clock_info->rs780.usVDDC)) {
+	case ATOM_PPLIB_RS780_VOLTAGE_NONE:
+	default:
+		ps->min_voltage = RS780_VDDC_LEVEL_UNKNOWN;
+		ps->max_voltage = RS780_VDDC_LEVEL_UNKNOWN;
+		break;
+	case ATOM_PPLIB_RS780_VOLTAGE_LOW:
+		ps->min_voltage = RS780_VDDC_LEVEL_LOW;
+		ps->max_voltage = RS780_VDDC_LEVEL_LOW;
+		break;
+	case ATOM_PPLIB_RS780_VOLTAGE_HIGH:
+		ps->min_voltage = RS780_VDDC_LEVEL_HIGH;
+		ps->max_voltage = RS780_VDDC_LEVEL_HIGH;
+		break;
+	case ATOM_PPLIB_RS780_VOLTAGE_VARIABLE:
+		ps->min_voltage = RS780_VDDC_LEVEL_LOW;
+		ps->max_voltage = RS780_VDDC_LEVEL_HIGH;
+		break;
+	}
+	ps->flags = le32_to_cpu(clock_info->rs780.ulFlags);
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		ps->sclk_low = rdev->clock.default_sclk;
+		ps->sclk_high = rdev->clock.default_sclk;
+		ps->min_voltage = RS780_VDDC_LEVEL_HIGH;
+		ps->max_voltage = RS780_VDDC_LEVEL_HIGH;
+	}
+}
+
+static int rs780_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i;
+	union pplib_clock_info *clock_info;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	struct igp_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  power_info->pplib.ucNumStates, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+
+	for (i = 0; i < power_info->pplib.ucNumStates; i++) {
+		power_state = (union pplib_power_state *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usStateArrayOffset) +
+			 i * power_info->pplib.ucStateEntrySize);
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset) +
+			 (power_state->v1.ucNonClockStateIndex *
+			  power_info->pplib.ucNonClockSize));
+		if (power_info->pplib.ucStateEntrySize - 1) {
+			clock_info = (union pplib_clock_info *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) +
+				 (power_state->v1.ucClockStateIndices[0] *
+				  power_info->pplib.ucClockInfoSize));
+			ps = kzalloc(sizeof(struct igp_ps), GFP_KERNEL);
+			if (ps == NULL) {
+				kfree(rdev->pm.dpm.ps);
+				return -ENOMEM;
+			}
+			rdev->pm.dpm.ps[i].ps_priv = ps;
+			rs780_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+							 non_clock_info,
+							 power_info->pplib.ucNonClockSize);
+			rs780_parse_pplib_clock_info(rdev,
+						     &rdev->pm.dpm.ps[i],
+						     clock_info);
+		}
+	}
+	rdev->pm.dpm.num_ps = power_info->pplib.ucNumStates;
+	return 0;
+}
+
+int rs780_dpm_init(struct radeon_device *rdev)
+{
+	struct igp_power_info *pi;
+	int index = GetIndexIntoMasterTable(DATA, IntegratedSystemInfo);
+	union igp_info *info;
+	u16 data_offset;
+	u8 frev, crev;
+	int ret;
+
+	pi = kzalloc(sizeof(struct igp_power_info), GFP_KERNEL);
+	if (pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = pi;
+
+	ret = rs780_parse_power_table(rdev);
+	if (ret)
+		return ret;
+
+	pi->voltage_control = false;
+	pi->gfx_clock_gating = true;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, NULL,
+				   &frev, &crev, &data_offset)) {
+		info = (union igp_info *)(rdev->mode_info.atom_context->bios + data_offset);
+
+		/* Get various system informations from bios */
+		switch (crev) {
+		case 1:
+			pi->num_of_cycles_in_period =
+				info->info.ucNumberOfCyclesInPeriod;
+			pi->num_of_cycles_in_period |=
+				info->info.ucNumberOfCyclesInPeriodHi << 8;
+			pi->invert_pwm_required =
+				(pi->num_of_cycles_in_period & 0x8000) ? true : false;
+			pi->boot_voltage = info->info.ucStartingPWM_HighTime;
+			pi->max_voltage = info->info.ucMaxNBVoltage;
+			pi->max_voltage |= info->info.ucMaxNBVoltageHigh << 8;
+			pi->min_voltage = info->info.ucMinNBVoltage;
+			pi->min_voltage |= info->info.ucMinNBVoltageHigh << 8;
+			pi->inter_voltage_low =
+				le16_to_cpu(info->info.usInterNBVoltageLow);
+			pi->inter_voltage_high =
+				le16_to_cpu(info->info.usInterNBVoltageHigh);
+			pi->voltage_control = true;
+			pi->bootup_uma_clk = info->info.usK8MemoryClock * 100;
+			break;
+		case 2:
+			pi->num_of_cycles_in_period =
+				le16_to_cpu(info->info_2.usNumberOfCyclesInPeriod);
+			pi->invert_pwm_required =
+				(pi->num_of_cycles_in_period & 0x8000) ? true : false;
+			pi->boot_voltage =
+				le16_to_cpu(info->info_2.usBootUpNBVoltage);
+			pi->max_voltage =
+				le16_to_cpu(info->info_2.usMaxNBVoltage);
+			pi->min_voltage =
+				le16_to_cpu(info->info_2.usMinNBVoltage);
+			pi->system_config =
+				le32_to_cpu(info->info_2.ulSystemConfig);
+			pi->pwm_voltage_control =
+				(pi->system_config & 0x4) ? true : false;
+			pi->voltage_control = true;
+			pi->bootup_uma_clk = le32_to_cpu(info->info_2.ulBootUpUMAClock);
+			break;
+		default:
+			DRM_ERROR("No integrated system info for your GPU\n");
+			return -EINVAL;
+		}
+		if (pi->min_voltage > pi->max_voltage)
+			pi->voltage_control = false;
+		if (pi->pwm_voltage_control) {
+			if ((pi->num_of_cycles_in_period == 0) ||
+			    (pi->max_voltage == 0) ||
+			    (pi->min_voltage == 0))
+				pi->voltage_control = false;
+		} else {
+			if ((pi->num_of_cycles_in_period == 0) ||
+			    (pi->max_voltage == 0))
+				pi->voltage_control = false;
+		}
+
+		return 0;
+	}
+	radeon_dpm_fini(rdev);
+	return -EINVAL;
+}
+
+void rs780_dpm_print_power_state(struct radeon_device *rdev,
+				 struct radeon_ps *rps)
+{
+	struct igp_ps *ps = rs780_get_ps(rps);
+
+	r600_dpm_print_class_info(rps->class, rps->class2);
+	r600_dpm_print_cap_info(rps->caps);
+	printk("\tuvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+	printk("\t\tpower level 0    sclk: %u vddc_index: %d\n",
+	       ps->sclk_low, ps->min_voltage);
+	printk("\t\tpower level 1    sclk: %u vddc_index: %d\n",
+	       ps->sclk_high, ps->max_voltage);
+	r600_dpm_print_ps_status(rdev, rps);
+}
+
+void rs780_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+}
+
+u32 rs780_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct igp_ps *requested_state = rs780_get_ps(rdev->pm.dpm.requested_ps);
+
+	if (low)
+		return requested_state->sclk_low;
+	else
+		return requested_state->sclk_high;
+}
+
+u32 rs780_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct igp_power_info *pi = rs780_get_pi(rdev);
+
+	return pi->bootup_uma_clk;
+}
+
+void rs780_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						       struct seq_file *m)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct igp_ps *ps = rs780_get_ps(rps);
+	u32 current_fb_div = RREG32(FVTHROT_STATUS_REG0) & CURRENT_FEEDBACK_DIV_MASK;
+	u32 func_cntl = RREG32(CG_SPLL_FUNC_CNTL);
+	u32 ref_div = ((func_cntl & SPLL_REF_DIV_MASK) >> SPLL_REF_DIV_SHIFT) + 1;
+	u32 post_div = ((func_cntl & SPLL_SW_HILEN_MASK) >> SPLL_SW_HILEN_SHIFT) + 1 +
+		((func_cntl & SPLL_SW_LOLEN_MASK) >> SPLL_SW_LOLEN_SHIFT) + 1;
+	u32 sclk = (rdev->clock.spll.reference_freq * current_fb_div) /
+		(post_div * ref_div);
+
+	seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+
+	/* guess based on the current sclk */
+	if (sclk < (ps->sclk_low + 500))
+		seq_printf(m, "power level 0    sclk: %u vddc_index: %d\n",
+			   ps->sclk_low, ps->min_voltage);
+	else
+		seq_printf(m, "power level 1    sclk: %u vddc_index: %d\n",
+			   ps->sclk_high, ps->max_voltage);
+}
diff --git a/drivers/gpu/drm/radeon/rs780_dpm.h b/drivers/gpu/drm/radeon/rs780_dpm.h
new file mode 100644
index 000000000000..47a40b14fa43
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rs780_dpm.h
@@ -0,0 +1,109 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __RS780_DPM_H__
+#define __RS780_DPM_H__
+
+enum rs780_vddc_level {
+	RS780_VDDC_LEVEL_UNKNOWN = 0,
+	RS780_VDDC_LEVEL_LOW = 1,
+	RS780_VDDC_LEVEL_HIGH = 2,
+};
+
+struct igp_power_info {
+	/* flags */
+	bool invert_pwm_required;
+	bool pwm_voltage_control;
+	bool voltage_control;
+	bool gfx_clock_gating;
+	/* stored values */
+	u32 system_config;
+	u32 bootup_uma_clk;
+	u16 max_voltage;
+	u16 min_voltage;
+	u16 boot_voltage;
+	u16 inter_voltage_low;
+	u16 inter_voltage_high;
+	u16 num_of_cycles_in_period;
+	/* variable */
+	int crtc_id;
+	int refresh_rate;
+};
+
+struct igp_ps {
+	enum rs780_vddc_level min_voltage;
+	enum rs780_vddc_level max_voltage;
+	u32 sclk_low;
+	u32 sclk_high;
+	u32 flags;
+};
+
+#define RS780_CGFTV_DFLT                 0x0303000f
+#define RS780_FBDIVTIMERVAL_DFLT         0x2710
+
+#define RS780_FVTHROTUTC0_DFLT   0x04010040
+#define RS780_FVTHROTUTC1_DFLT   0x04010040
+#define RS780_FVTHROTUTC2_DFLT   0x04010040
+#define RS780_FVTHROTUTC3_DFLT   0x04010040
+#define RS780_FVTHROTUTC4_DFLT   0x04010040
+
+#define RS780_FVTHROTDTC0_DFLT 0x04010040
+#define RS780_FVTHROTDTC1_DFLT 0x04010040
+#define RS780_FVTHROTDTC2_DFLT 0x04010040
+#define RS780_FVTHROTDTC3_DFLT 0x04010040
+#define RS780_FVTHROTDTC4_DFLT 0x04010040
+
+#define RS780_FVTHROTFBUSREG0_DFLT       0x00001001
+#define RS780_FVTHROTFBUSREG1_DFLT       0x00002002
+#define RS780_FVTHROTFBDSREG0_DFLT       0x00004001
+#define RS780_FVTHROTFBDSREG1_DFLT       0x00020010
+
+#define RS780_FVTHROTPWMUSREG0_DFLT      0x00002001
+#define RS780_FVTHROTPWMUSREG1_DFLT      0x00004003
+#define RS780_FVTHROTPWMDSREG0_DFLT      0x00002001
+#define RS780_FVTHROTPWMDSREG1_DFLT      0x00004003
+
+#define RS780_FVTHROTPWMFBDIVRANGEREG0_DFLT  0x37
+#define RS780_FVTHROTPWMFBDIVRANGEREG1_DFLT  0x4b
+#define RS780_FVTHROTPWMFBDIVRANGEREG2_DFLT  0x8b
+
+#define RS780D_FVTHROTPWMFBDIVRANGEREG0_DFLT  0x8b
+#define RS780D_FVTHROTPWMFBDIVRANGEREG1_DFLT  0x8c
+#define RS780D_FVTHROTPWMFBDIVRANGEREG2_DFLT  0xb5
+
+#define RS880D_FVTHROTPWMFBDIVRANGEREG0_DFLT  0x8d
+#define RS880D_FVTHROTPWMFBDIVRANGEREG1_DFLT  0x8e
+#define RS880D_FVTHROTPWMFBDIVRANGEREG2_DFLT  0xBa
+
+#define RS780_FVTHROTPWMRANGE0_GPIO_DFLT  0x1a
+#define RS780_FVTHROTPWMRANGE1_GPIO_DFLT  0x1a
+#define RS780_FVTHROTPWMRANGE2_GPIO_DFLT  0x0
+#define RS780_FVTHROTPWMRANGE3_GPIO_DFLT  0x0
+
+#define RS780_SLOWCLKFEEDBACKDIV_DFLT 110
+
+#define RS780_CGCLKGATING_DFLT           0x0000E204
+
+#define RS780_DEFAULT_VCLK_FREQ  53300 /* 10 khz */
+#define RS780_DEFAULT_DCLK_FREQ  40000 /* 10 khz */
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rs780d.h b/drivers/gpu/drm/radeon/rs780d.h
new file mode 100644
index 000000000000..cfbe9a43d97b
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rs780d.h
@@ -0,0 +1,171 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __RS780D_H__
+#define __RS780D_H__
+
+#define CG_SPLL_FUNC_CNTL                                 0x600
+#       define SPLL_RESET                                (1 << 0)
+#       define SPLL_SLEEP                                (1 << 1)
+#       define SPLL_REF_DIV(x)                           ((x) << 2)
+#       define SPLL_REF_DIV_MASK                         (7 << 2)
+#       define SPLL_REF_DIV_SHIFT                        2
+#       define SPLL_FB_DIV(x)                            ((x) << 5)
+#       define SPLL_FB_DIV_MASK                          (0xff << 2)
+#       define SPLL_FB_DIV_SHIFT                         2
+#       define SPLL_PULSEEN                              (1 << 13)
+#       define SPLL_PULSENUM(x)                          ((x) << 14)
+#       define SPLL_PULSENUM_MASK                        (3 << 14)
+#       define SPLL_SW_HILEN(x)                          ((x) << 16)
+#       define SPLL_SW_HILEN_MASK                        (0xf << 16)
+#       define SPLL_SW_HILEN_SHIFT                       16
+#       define SPLL_SW_LOLEN(x)                          ((x) << 20)
+#       define SPLL_SW_LOLEN_MASK                        (0xf << 20)
+#       define SPLL_SW_LOLEN_SHIFT                       20
+#       define SPLL_DIVEN                                (1 << 24)
+#       define SPLL_BYPASS_EN                            (1 << 25)
+#       define SPLL_CHG_STATUS                           (1 << 29)
+#       define SPLL_CTLREQ                               (1 << 30)
+#       define SPLL_CTLACK                               (1 << 31)
+
+/* RS780/RS880 PM */
+#define	FVTHROT_CNTRL_REG				0x3000
+#define		DONT_WAIT_FOR_FBDIV_WRAP		(1 << 0)
+#define		MINIMUM_CIP(x)				((x) << 1)
+#define		MINIMUM_CIP_SHIFT			1
+#define		MINIMUM_CIP_MASK			0x1fffffe
+#define		REFRESH_RATE_DIVISOR(x)			((x) << 25)
+#define		REFRESH_RATE_DIVISOR_SHIFT		25
+#define		REFRESH_RATE_DIVISOR_MASK		(0x3 << 25)
+#define		ENABLE_FV_THROT				(1 << 27)
+#define		ENABLE_FV_UPDATE			(1 << 28)
+#define		TREND_SEL_MODE				(1 << 29)
+#define		FORCE_TREND_SEL				(1 << 30)
+#define		ENABLE_FV_THROT_IO			(1 << 31)
+#define	FVTHROT_TARGET_REG				0x3004
+#define		TARGET_IDLE_COUNT(x)			((x) << 0)
+#define		TARGET_IDLE_COUNT_MASK			0xffffff
+#define		TARGET_IDLE_COUNT_SHIFT			0
+#define	FVTHROT_CB1					0x3008
+#define	FVTHROT_CB2					0x300c
+#define	FVTHROT_CB3					0x3010
+#define	FVTHROT_CB4					0x3014
+#define	FVTHROT_UTC0					0x3018
+#define	FVTHROT_UTC1					0x301c
+#define	FVTHROT_UTC2					0x3020
+#define	FVTHROT_UTC3					0x3024
+#define	FVTHROT_UTC4					0x3028
+#define	FVTHROT_DTC0					0x302c
+#define	FVTHROT_DTC1					0x3030
+#define	FVTHROT_DTC2					0x3034
+#define	FVTHROT_DTC3					0x3038
+#define	FVTHROT_DTC4					0x303c
+#define	FVTHROT_FBDIV_REG0				0x3040
+#define		MIN_FEEDBACK_DIV(x)			((x) << 0)
+#define		MIN_FEEDBACK_DIV_MASK			0xfff
+#define		MIN_FEEDBACK_DIV_SHIFT			0
+#define		MAX_FEEDBACK_DIV(x)			((x) << 12)
+#define		MAX_FEEDBACK_DIV_MASK			(0xfff << 12)
+#define		MAX_FEEDBACK_DIV_SHIFT			12
+#define	FVTHROT_FBDIV_REG1				0x3044
+#define		MAX_FEEDBACK_STEP(x)			((x) << 0)
+#define		MAX_FEEDBACK_STEP_MASK			0xfff
+#define		MAX_FEEDBACK_STEP_SHIFT			0
+#define		STARTING_FEEDBACK_DIV(x)		((x) << 12)
+#define		STARTING_FEEDBACK_DIV_MASK		(0xfff << 12)
+#define		STARTING_FEEDBACK_DIV_SHIFT		12
+#define		FORCE_FEEDBACK_DIV			(1 << 24)
+#define	FVTHROT_FBDIV_REG2				0x3048
+#define		FORCED_FEEDBACK_DIV(x)			((x) << 0)
+#define		FORCED_FEEDBACK_DIV_MASK		0xfff
+#define		FORCED_FEEDBACK_DIV_SHIFT		0
+#define		FB_DIV_TIMER_VAL(x)			((x) << 12)
+#define		FB_DIV_TIMER_VAL_MASK			(0xffff << 12)
+#define		FB_DIV_TIMER_VAL_SHIFT			12
+#define	FVTHROT_FB_US_REG0				0x304c
+#define	FVTHROT_FB_US_REG1				0x3050
+#define	FVTHROT_FB_DS_REG0				0x3054
+#define	FVTHROT_FB_DS_REG1				0x3058
+#define	FVTHROT_PWM_CTRL_REG0				0x305c
+#define		STARTING_PWM_HIGHTIME(x)		((x) << 0)
+#define		STARTING_PWM_HIGHTIME_MASK		0xfff
+#define		STARTING_PWM_HIGHTIME_SHIFT		0
+#define		NUMBER_OF_CYCLES_IN_PERIOD(x)		((x) << 12)
+#define		NUMBER_OF_CYCLES_IN_PERIOD_MASK		(0xfff << 12)
+#define		NUMBER_OF_CYCLES_IN_PERIOD_SHIFT	12
+#define		FORCE_STARTING_PWM_HIGHTIME		(1 << 24)
+#define		INVERT_PWM_WAVEFORM			(1 << 25)
+#define	FVTHROT_PWM_CTRL_REG1				0x3060
+#define		MIN_PWM_HIGHTIME(x)			((x) << 0)
+#define		MIN_PWM_HIGHTIME_MASK			0xfff
+#define		MIN_PWM_HIGHTIME_SHIFT			0
+#define		MAX_PWM_HIGHTIME(x)			((x) << 12)
+#define		MAX_PWM_HIGHTIME_MASK			(0xfff << 12)
+#define		MAX_PWM_HIGHTIME_SHIFT			12
+#define	FVTHROT_PWM_US_REG0				0x3064
+#define	FVTHROT_PWM_US_REG1				0x3068
+#define	FVTHROT_PWM_DS_REG0				0x306c
+#define	FVTHROT_PWM_DS_REG1				0x3070
+#define	FVTHROT_STATUS_REG0				0x3074
+#define		CURRENT_FEEDBACK_DIV_MASK		0xfff
+#define		CURRENT_FEEDBACK_DIV_SHIFT		0
+#define	FVTHROT_STATUS_REG1				0x3078
+#define	FVTHROT_STATUS_REG2				0x307c
+#define	CG_INTGFX_MISC					0x3080
+#define		FVTHROT_VBLANK_SEL			(1 << 9)
+#define	FVTHROT_PWM_FEEDBACK_DIV_REG1			0x308c
+#define		RANGE0_PWM_FEEDBACK_DIV(x)		((x) << 0)
+#define		RANGE0_PWM_FEEDBACK_DIV_MASK		0xfff
+#define		RANGE0_PWM_FEEDBACK_DIV_SHIFT		0
+#define		RANGE_PWM_FEEDBACK_DIV_EN		(1 << 12)
+#define	FVTHROT_PWM_FEEDBACK_DIV_REG2			0x3090
+#define		RANGE1_PWM_FEEDBACK_DIV(x)		((x) << 0)
+#define		RANGE1_PWM_FEEDBACK_DIV_MASK		0xfff
+#define		RANGE1_PWM_FEEDBACK_DIV_SHIFT		0
+#define		RANGE2_PWM_FEEDBACK_DIV(x)		((x) << 12)
+#define		RANGE2_PWM_FEEDBACK_DIV_MASK		(0xfff << 12)
+#define		RANGE2_PWM_FEEDBACK_DIV_SHIFT		12
+#define	FVTHROT_PWM_FEEDBACK_DIV_REG3			0x3094
+#define		RANGE0_PWM(x)				((x) << 0)
+#define		RANGE0_PWM_MASK				0xfff
+#define		RANGE0_PWM_SHIFT			0
+#define		RANGE1_PWM(x)				((x) << 12)
+#define		RANGE1_PWM_MASK				(0xfff << 12)
+#define		RANGE1_PWM_SHIFT			12
+#define	FVTHROT_PWM_FEEDBACK_DIV_REG4			0x3098
+#define		RANGE2_PWM(x)				((x) << 0)
+#define		RANGE2_PWM_MASK				0xfff
+#define		RANGE2_PWM_SHIFT			0
+#define		RANGE3_PWM(x)				((x) << 12)
+#define		RANGE3_PWM_MASK				(0xfff << 12)
+#define		RANGE3_PWM_SHIFT			12
+#define	FVTHROT_SLOW_CLK_FEEDBACK_DIV_REG1		0x30ac
+#define		RANGE0_SLOW_CLK_FEEDBACK_DIV(x)		((x) << 0)
+#define		RANGE0_SLOW_CLK_FEEDBACK_DIV_MASK	0xfff
+#define		RANGE0_SLOW_CLK_FEEDBACK_DIV_SHIFT	0
+#define		RANGE_SLOW_CLK_FEEDBACK_DIV_EN		(1 << 12)
+
+#define	GFX_MACRO_BYPASS_CNTL				0x30c0
+#define		SPLL_BYPASS_CNTL			(1 << 0)
+#define		UPLL_BYPASS_CNTL			(1 << 1)
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c
index 21c7d7b26e55..8ea1573ae820 100644
--- a/drivers/gpu/drm/radeon/rv515.c
+++ b/drivers/gpu/drm/radeon/rv515.c
@@ -937,13 +937,16 @@ struct rv515_watermark {
 };
 
 static void rv515_crtc_bandwidth_compute(struct radeon_device *rdev,
-				  struct radeon_crtc *crtc,
-				  struct rv515_watermark *wm)
+					 struct radeon_crtc *crtc,
+					 struct rv515_watermark *wm,
+					 bool low)
 {
 	struct drm_display_mode *mode = &crtc->base.mode;
 	fixed20_12 a, b, c;
 	fixed20_12 pclk, request_fifo_depth, tolerable_latency, estimated_width;
 	fixed20_12 consumption_time, line_time, chunk_time, read_delay_latency;
+	fixed20_12 sclk;
+	u32 selected_sclk;
 
 	if (!crtc->base.enabled) {
 		/* FIXME: wouldn't it better to set priority mark to maximum */
@@ -951,6 +954,18 @@ static void rv515_crtc_bandwidth_compute(struct radeon_device *rdev,
 		return;
 	}
 
+	/* rv6xx, rv7xx */
+	if ((rdev->family >= CHIP_RV610) &&
+	    (rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled)
+		selected_sclk = radeon_dpm_get_sclk(rdev, low);
+	else
+		selected_sclk = rdev->pm.current_sclk;
+
+	/* sclk in Mhz */
+	a.full = dfixed_const(100);
+	sclk.full = dfixed_const(selected_sclk);
+	sclk.full = dfixed_div(sclk, a);
+
 	if (crtc->vsc.full > dfixed_const(2))
 		wm->num_line_pair.full = dfixed_const(2);
 	else
@@ -1016,7 +1031,7 @@ static void rv515_crtc_bandwidth_compute(struct radeon_device *rdev,
 	 * sclk = system clock(Mhz)
 	 */
 	a.full = dfixed_const(600 * 1000);
-	chunk_time.full = dfixed_div(a, rdev->pm.sclk);
+	chunk_time.full = dfixed_div(a, sclk);
 	read_delay_latency.full = dfixed_const(1000);
 
 	/* Determine the worst case latency
@@ -1077,152 +1092,177 @@ static void rv515_crtc_bandwidth_compute(struct radeon_device *rdev,
 	}
 }
 
-void rv515_bandwidth_avivo_update(struct radeon_device *rdev)
+static void rv515_compute_mode_priority(struct radeon_device *rdev,
+					struct rv515_watermark *wm0,
+					struct rv515_watermark *wm1,
+					struct drm_display_mode *mode0,
+					struct drm_display_mode *mode1,
+					u32 *d1mode_priority_a_cnt,
+					u32 *d2mode_priority_a_cnt)
 {
-	struct drm_display_mode *mode0 = NULL;
-	struct drm_display_mode *mode1 = NULL;
-	struct rv515_watermark wm0;
-	struct rv515_watermark wm1;
-	u32 tmp;
-	u32 d1mode_priority_a_cnt = MODE_PRIORITY_OFF;
-	u32 d2mode_priority_a_cnt = MODE_PRIORITY_OFF;
 	fixed20_12 priority_mark02, priority_mark12, fill_rate;
 	fixed20_12 a, b;
 
-	if (rdev->mode_info.crtcs[0]->base.enabled)
-		mode0 = &rdev->mode_info.crtcs[0]->base.mode;
-	if (rdev->mode_info.crtcs[1]->base.enabled)
-		mode1 = &rdev->mode_info.crtcs[1]->base.mode;
-	rs690_line_buffer_adjust(rdev, mode0, mode1);
-
-	rv515_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0);
-	rv515_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1);
-
-	tmp = wm0.lb_request_fifo_depth;
-	tmp |= wm1.lb_request_fifo_depth << 16;
-	WREG32(LB_MAX_REQ_OUTSTANDING, tmp);
+	*d1mode_priority_a_cnt = MODE_PRIORITY_OFF;
+	*d2mode_priority_a_cnt = MODE_PRIORITY_OFF;
 
 	if (mode0 && mode1) {
-		if (dfixed_trunc(wm0.dbpp) > 64)
-			a.full = dfixed_div(wm0.dbpp, wm0.num_line_pair);
+		if (dfixed_trunc(wm0->dbpp) > 64)
+			a.full = dfixed_div(wm0->dbpp, wm0->num_line_pair);
 		else
-			a.full = wm0.num_line_pair.full;
-		if (dfixed_trunc(wm1.dbpp) > 64)
-			b.full = dfixed_div(wm1.dbpp, wm1.num_line_pair);
+			a.full = wm0->num_line_pair.full;
+		if (dfixed_trunc(wm1->dbpp) > 64)
+			b.full = dfixed_div(wm1->dbpp, wm1->num_line_pair);
 		else
-			b.full = wm1.num_line_pair.full;
+			b.full = wm1->num_line_pair.full;
 		a.full += b.full;
-		fill_rate.full = dfixed_div(wm0.sclk, a);
-		if (wm0.consumption_rate.full > fill_rate.full) {
-			b.full = wm0.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm0.active_time);
+		fill_rate.full = dfixed_div(wm0->sclk, a);
+		if (wm0->consumption_rate.full > fill_rate.full) {
+			b.full = wm0->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm0->active_time);
 			a.full = dfixed_const(16);
 			b.full = dfixed_div(b, a);
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			priority_mark02.full = a.full + b.full;
 		} else {
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark02.full = dfixed_div(a, b);
 		}
-		if (wm1.consumption_rate.full > fill_rate.full) {
-			b.full = wm1.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm1.active_time);
+		if (wm1->consumption_rate.full > fill_rate.full) {
+			b.full = wm1->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm1->active_time);
 			a.full = dfixed_const(16);
 			b.full = dfixed_div(b, a);
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			priority_mark12.full = a.full + b.full;
 		} else {
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark12.full = dfixed_div(a, b);
 		}
-		if (wm0.priority_mark.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark.full;
+		if (wm0->priority_mark.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark.full;
 		if (dfixed_trunc(priority_mark02) < 0)
 			priority_mark02.full = 0;
-		if (wm0.priority_mark_max.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark_max.full;
-		if (wm1.priority_mark.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark.full;
+		if (wm0->priority_mark_max.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark_max.full;
+		if (wm1->priority_mark.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark.full;
 		if (dfixed_trunc(priority_mark12) < 0)
 			priority_mark12.full = 0;
-		if (wm1.priority_mark_max.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark_max.full;
-		d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
-		d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
+		if (wm1->priority_mark_max.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark_max.full;
+		*d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
+		*d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
 		if (rdev->disp_priority == 2) {
-			d1mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
-			d2mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
+			*d1mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
+			*d2mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
 		}
 	} else if (mode0) {
-		if (dfixed_trunc(wm0.dbpp) > 64)
-			a.full = dfixed_div(wm0.dbpp, wm0.num_line_pair);
+		if (dfixed_trunc(wm0->dbpp) > 64)
+			a.full = dfixed_div(wm0->dbpp, wm0->num_line_pair);
 		else
-			a.full = wm0.num_line_pair.full;
-		fill_rate.full = dfixed_div(wm0.sclk, a);
-		if (wm0.consumption_rate.full > fill_rate.full) {
-			b.full = wm0.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm0.active_time);
+			a.full = wm0->num_line_pair.full;
+		fill_rate.full = dfixed_div(wm0->sclk, a);
+		if (wm0->consumption_rate.full > fill_rate.full) {
+			b.full = wm0->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm0->active_time);
 			a.full = dfixed_const(16);
 			b.full = dfixed_div(b, a);
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			priority_mark02.full = a.full + b.full;
 		} else {
-			a.full = dfixed_mul(wm0.worst_case_latency,
-						wm0.consumption_rate);
+			a.full = dfixed_mul(wm0->worst_case_latency,
+						wm0->consumption_rate);
 			b.full = dfixed_const(16);
 			priority_mark02.full = dfixed_div(a, b);
 		}
-		if (wm0.priority_mark.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark.full;
+		if (wm0->priority_mark.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark.full;
 		if (dfixed_trunc(priority_mark02) < 0)
 			priority_mark02.full = 0;
-		if (wm0.priority_mark_max.full > priority_mark02.full)
-			priority_mark02.full = wm0.priority_mark_max.full;
-		d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
+		if (wm0->priority_mark_max.full > priority_mark02.full)
+			priority_mark02.full = wm0->priority_mark_max.full;
+		*d1mode_priority_a_cnt = dfixed_trunc(priority_mark02);
 		if (rdev->disp_priority == 2)
-			d1mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
+			*d1mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
 	} else if (mode1) {
-		if (dfixed_trunc(wm1.dbpp) > 64)
-			a.full = dfixed_div(wm1.dbpp, wm1.num_line_pair);
+		if (dfixed_trunc(wm1->dbpp) > 64)
+			a.full = dfixed_div(wm1->dbpp, wm1->num_line_pair);
 		else
-			a.full = wm1.num_line_pair.full;
-		fill_rate.full = dfixed_div(wm1.sclk, a);
-		if (wm1.consumption_rate.full > fill_rate.full) {
-			b.full = wm1.consumption_rate.full - fill_rate.full;
-			b.full = dfixed_mul(b, wm1.active_time);
+			a.full = wm1->num_line_pair.full;
+		fill_rate.full = dfixed_div(wm1->sclk, a);
+		if (wm1->consumption_rate.full > fill_rate.full) {
+			b.full = wm1->consumption_rate.full - fill_rate.full;
+			b.full = dfixed_mul(b, wm1->active_time);
 			a.full = dfixed_const(16);
 			b.full = dfixed_div(b, a);
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			priority_mark12.full = a.full + b.full;
 		} else {
-			a.full = dfixed_mul(wm1.worst_case_latency,
-						wm1.consumption_rate);
+			a.full = dfixed_mul(wm1->worst_case_latency,
+						wm1->consumption_rate);
 			b.full = dfixed_const(16 * 1000);
 			priority_mark12.full = dfixed_div(a, b);
 		}
-		if (wm1.priority_mark.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark.full;
+		if (wm1->priority_mark.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark.full;
 		if (dfixed_trunc(priority_mark12) < 0)
 			priority_mark12.full = 0;
-		if (wm1.priority_mark_max.full > priority_mark12.full)
-			priority_mark12.full = wm1.priority_mark_max.full;
-		d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
+		if (wm1->priority_mark_max.full > priority_mark12.full)
+			priority_mark12.full = wm1->priority_mark_max.full;
+		*d2mode_priority_a_cnt = dfixed_trunc(priority_mark12);
 		if (rdev->disp_priority == 2)
-			d2mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
+			*d2mode_priority_a_cnt |= MODE_PRIORITY_ALWAYS_ON;
 	}
+}
+
+void rv515_bandwidth_avivo_update(struct radeon_device *rdev)
+{
+	struct drm_display_mode *mode0 = NULL;
+	struct drm_display_mode *mode1 = NULL;
+	struct rv515_watermark wm0_high, wm0_low;
+	struct rv515_watermark wm1_high, wm1_low;
+	u32 tmp;
+	u32 d1mode_priority_a_cnt, d1mode_priority_b_cnt;
+	u32 d2mode_priority_a_cnt, d2mode_priority_b_cnt;
+
+	if (rdev->mode_info.crtcs[0]->base.enabled)
+		mode0 = &rdev->mode_info.crtcs[0]->base.mode;
+	if (rdev->mode_info.crtcs[1]->base.enabled)
+		mode1 = &rdev->mode_info.crtcs[1]->base.mode;
+	rs690_line_buffer_adjust(rdev, mode0, mode1);
+
+	rv515_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0_high, false);
+	rv515_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1_high, false);
+
+	rv515_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[0], &wm0_low, false);
+	rv515_crtc_bandwidth_compute(rdev, rdev->mode_info.crtcs[1], &wm1_low, false);
+
+	tmp = wm0_high.lb_request_fifo_depth;
+	tmp |= wm1_high.lb_request_fifo_depth << 16;
+	WREG32(LB_MAX_REQ_OUTSTANDING, tmp);
+
+	rv515_compute_mode_priority(rdev,
+				    &wm0_high, &wm1_high,
+				    mode0, mode1,
+				    &d1mode_priority_a_cnt, &d2mode_priority_a_cnt);
+	rv515_compute_mode_priority(rdev,
+				    &wm0_low, &wm1_low,
+				    mode0, mode1,
+				    &d1mode_priority_b_cnt, &d2mode_priority_b_cnt);
 
 	WREG32(D1MODE_PRIORITY_A_CNT, d1mode_priority_a_cnt);
-	WREG32(D1MODE_PRIORITY_B_CNT, d1mode_priority_a_cnt);
+	WREG32(D1MODE_PRIORITY_B_CNT, d1mode_priority_b_cnt);
 	WREG32(D2MODE_PRIORITY_A_CNT, d2mode_priority_a_cnt);
-	WREG32(D2MODE_PRIORITY_B_CNT, d2mode_priority_a_cnt);
+	WREG32(D2MODE_PRIORITY_B_CNT, d2mode_priority_b_cnt);
 }
 
 void rv515_bandwidth_update(struct radeon_device *rdev)
diff --git a/drivers/gpu/drm/radeon/rv6xx_dpm.c b/drivers/gpu/drm/radeon/rv6xx_dpm.c
new file mode 100644
index 000000000000..65e33f387341
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv6xx_dpm.c
@@ -0,0 +1,2087 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "rv6xxd.h"
+#include "r600_dpm.h"
+#include "rv6xx_dpm.h"
+#include "atom.h"
+#include <linux/seq_file.h>
+
+static u32 rv6xx_scale_count_given_unit(struct radeon_device *rdev,
+					u32 unscaled_count, u32 unit);
+
+static struct rv6xx_ps *rv6xx_get_ps(struct radeon_ps *rps)
+{
+	struct rv6xx_ps *ps = rps->ps_priv;
+
+	return ps;
+}
+
+static struct rv6xx_power_info *rv6xx_get_pi(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rdev->pm.dpm.priv;
+
+	return pi;
+}
+
+static void rv6xx_force_pcie_gen1(struct radeon_device *rdev)
+{
+	u32 tmp;
+	int i;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	tmp &= LC_GEN2_EN;
+	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	tmp |= LC_INITIATE_LINK_SPEED_CHANGE;
+	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (!(RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL) & LC_CURRENT_DATA_RATE))
+			break;
+		udelay(1);
+	}
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	tmp &= ~LC_INITIATE_LINK_SPEED_CHANGE;
+	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+}
+
+static void rv6xx_enable_pcie_gen2_support(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+
+	if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+	    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2)) {
+		tmp |= LC_GEN2_EN;
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+	}
+}
+
+static void rv6xx_enable_bif_dynamic_pcie_gen2(struct radeon_device *rdev,
+					       bool enable)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL) & ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+	if (enable)
+		tmp |= LC_HW_VOLTAGE_IF_CONTROL(1);
+	else
+		tmp |= LC_HW_VOLTAGE_IF_CONTROL(0);
+	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+}
+
+static void rv6xx_enable_l0s(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL) & ~LC_L0S_INACTIVITY_MASK;
+	tmp |= LC_L0S_INACTIVITY(3);
+	WREG32_PCIE_PORT(PCIE_LC_CNTL, tmp);
+}
+
+static void rv6xx_enable_l1(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL);
+	tmp &= ~LC_L1_INACTIVITY_MASK;
+	tmp |= LC_L1_INACTIVITY(4);
+	tmp &= ~LC_PMI_TO_L1_DIS;
+	tmp &= ~LC_ASPM_TO_L1_DIS;
+	WREG32_PCIE_PORT(PCIE_LC_CNTL, tmp);
+}
+
+static void rv6xx_enable_pll_sleep_in_l1(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL) & ~LC_L1_INACTIVITY_MASK;
+	tmp |= LC_L1_INACTIVITY(8);
+	WREG32_PCIE_PORT(PCIE_LC_CNTL, tmp);
+
+	/* NOTE, this is a PCIE indirect reg, not PCIE PORT */
+	tmp = RREG32_PCIE(PCIE_P_CNTL);
+	tmp |= P_PLL_PWRDN_IN_L1L23;
+	tmp &= ~P_PLL_BUF_PDNB;
+	tmp &= ~P_PLL_PDNB;
+	tmp |= P_ALLOW_PRX_FRONTEND_SHUTOFF;
+	WREG32_PCIE(PCIE_P_CNTL, tmp);
+}
+
+static int rv6xx_convert_clock_to_stepping(struct radeon_device *rdev,
+					   u32 clock, struct rv6xx_sclk_stepping *step)
+{
+	int ret;
+	struct atom_clock_dividers dividers;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	if (dividers.enable_post_div)
+		step->post_divider = 2 + (dividers.post_div & 0xF) + (dividers.post_div >> 4);
+	else
+		step->post_divider = 1;
+
+	step->vco_frequency = clock * step->post_divider;
+
+	return 0;
+}
+
+static void rv6xx_output_stepping(struct radeon_device *rdev,
+				  u32 step_index, struct rv6xx_sclk_stepping *step)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	u32 ref_clk = rdev->clock.spll.reference_freq;
+	u32 fb_divider;
+	u32 spll_step_count = rv6xx_scale_count_given_unit(rdev,
+							   R600_SPLLSTEPTIME_DFLT *
+							   pi->spll_ref_div,
+							   R600_SPLLSTEPUNIT_DFLT);
+
+	r600_engine_clock_entry_enable(rdev, step_index, true);
+	r600_engine_clock_entry_enable_pulse_skipping(rdev, step_index, false);
+
+	if (step->post_divider == 1)
+		r600_engine_clock_entry_enable_post_divider(rdev, step_index, false);
+	else {
+		u32 lo_len = (step->post_divider - 2) / 2;
+		u32 hi_len = step->post_divider - 2 - lo_len;
+
+		r600_engine_clock_entry_enable_post_divider(rdev, step_index, true);
+		r600_engine_clock_entry_set_post_divider(rdev, step_index, (hi_len << 4) | lo_len);
+	}
+
+	fb_divider = ((step->vco_frequency * pi->spll_ref_div) / ref_clk) >>
+		pi->fb_div_scale;
+
+	r600_engine_clock_entry_set_reference_divider(rdev, step_index,
+						      pi->spll_ref_div - 1);
+	r600_engine_clock_entry_set_feedback_divider(rdev, step_index, fb_divider);
+	r600_engine_clock_entry_set_step_time(rdev, step_index, spll_step_count);
+
+}
+
+static struct rv6xx_sclk_stepping rv6xx_next_vco_step(struct radeon_device *rdev,
+						      struct rv6xx_sclk_stepping *cur,
+						      bool increasing_vco, u32 step_size)
+{
+	struct rv6xx_sclk_stepping next;
+
+	next.post_divider = cur->post_divider;
+
+	if (increasing_vco)
+		next.vco_frequency = (cur->vco_frequency * (100 + step_size)) / 100;
+	else
+		next.vco_frequency = (cur->vco_frequency * 100 + 99 + step_size) / (100 + step_size);
+
+	return next;
+}
+
+static bool rv6xx_can_step_post_div(struct radeon_device *rdev,
+				    struct rv6xx_sclk_stepping *cur,
+                                    struct rv6xx_sclk_stepping *target)
+{
+	return (cur->post_divider > target->post_divider) &&
+		((cur->vco_frequency * target->post_divider) <=
+		 (target->vco_frequency * (cur->post_divider - 1)));
+}
+
+static struct rv6xx_sclk_stepping rv6xx_next_post_div_step(struct radeon_device *rdev,
+							   struct rv6xx_sclk_stepping *cur,
+							   struct rv6xx_sclk_stepping *target)
+{
+	struct rv6xx_sclk_stepping next = *cur;
+
+	while (rv6xx_can_step_post_div(rdev, &next, target))
+		next.post_divider--;
+
+	return next;
+}
+
+static bool rv6xx_reached_stepping_target(struct radeon_device *rdev,
+					  struct rv6xx_sclk_stepping *cur,
+					  struct rv6xx_sclk_stepping *target,
+					  bool increasing_vco)
+{
+	return (increasing_vco && (cur->vco_frequency >= target->vco_frequency)) ||
+		(!increasing_vco && (cur->vco_frequency <= target->vco_frequency));
+}
+
+static void rv6xx_generate_steps(struct radeon_device *rdev,
+				 u32 low, u32 high,
+                                 u32 start_index, u8 *end_index)
+{
+	struct rv6xx_sclk_stepping cur;
+	struct rv6xx_sclk_stepping target;
+	bool increasing_vco;
+	u32 step_index = start_index;
+
+	rv6xx_convert_clock_to_stepping(rdev, low, &cur);
+	rv6xx_convert_clock_to_stepping(rdev, high, &target);
+
+	rv6xx_output_stepping(rdev, step_index++, &cur);
+
+	increasing_vco = (target.vco_frequency >= cur.vco_frequency);
+
+	if (target.post_divider > cur.post_divider)
+		cur.post_divider = target.post_divider;
+
+	while (1) {
+		struct rv6xx_sclk_stepping next;
+
+		if (rv6xx_can_step_post_div(rdev, &cur, &target))
+			next = rv6xx_next_post_div_step(rdev, &cur, &target);
+		else
+			next = rv6xx_next_vco_step(rdev, &cur, increasing_vco, R600_VCOSTEPPCT_DFLT);
+
+		if (rv6xx_reached_stepping_target(rdev, &next, &target, increasing_vco)) {
+			struct rv6xx_sclk_stepping tiny =
+				rv6xx_next_vco_step(rdev, &target, !increasing_vco, R600_ENDINGVCOSTEPPCT_DFLT);
+			tiny.post_divider = next.post_divider;
+
+			if (!rv6xx_reached_stepping_target(rdev, &tiny, &cur, !increasing_vco))
+				rv6xx_output_stepping(rdev, step_index++, &tiny);
+
+			if ((next.post_divider != target.post_divider) &&
+			    (next.vco_frequency != target.vco_frequency)) {
+				struct rv6xx_sclk_stepping final_vco;
+
+				final_vco.vco_frequency = target.vco_frequency;
+				final_vco.post_divider = next.post_divider;
+
+				rv6xx_output_stepping(rdev, step_index++, &final_vco);
+			}
+
+			rv6xx_output_stepping(rdev, step_index++, &target);
+			break;
+		} else
+			rv6xx_output_stepping(rdev, step_index++, &next);
+
+		cur = next;
+	}
+
+	*end_index = (u8)step_index - 1;
+
+}
+
+static void rv6xx_generate_single_step(struct radeon_device *rdev,
+				       u32 clock, u32 index)
+{
+	struct rv6xx_sclk_stepping step;
+
+	rv6xx_convert_clock_to_stepping(rdev, clock, &step);
+	rv6xx_output_stepping(rdev, index, &step);
+}
+
+static void rv6xx_invalidate_intermediate_steps_range(struct radeon_device *rdev,
+						      u32 start_index, u32 end_index)
+{
+	u32 step_index;
+
+	for (step_index = start_index + 1; step_index < end_index; step_index++)
+		r600_engine_clock_entry_enable(rdev, step_index, false);
+}
+
+static void rv6xx_set_engine_spread_spectrum_clk_s(struct radeon_device *rdev,
+						   u32 index, u32 clk_s)
+{
+	WREG32_P(CG_SPLL_SPREAD_SPECTRUM_LOW + (index * 4),
+		 CLKS(clk_s), ~CLKS_MASK);
+}
+
+static void rv6xx_set_engine_spread_spectrum_clk_v(struct radeon_device *rdev,
+						   u32 index, u32 clk_v)
+{
+	WREG32_P(CG_SPLL_SPREAD_SPECTRUM_LOW + (index * 4),
+		 CLKV(clk_v), ~CLKV_MASK);
+}
+
+static void rv6xx_enable_engine_spread_spectrum(struct radeon_device *rdev,
+						u32 index, bool enable)
+{
+	if (enable)
+		WREG32_P(CG_SPLL_SPREAD_SPECTRUM_LOW + (index * 4),
+			 SSEN, ~SSEN);
+	else
+		WREG32_P(CG_SPLL_SPREAD_SPECTRUM_LOW + (index * 4),
+			 0, ~SSEN);
+}
+
+static void rv6xx_set_memory_spread_spectrum_clk_s(struct radeon_device *rdev,
+						   u32 clk_s)
+{
+	WREG32_P(CG_MPLL_SPREAD_SPECTRUM, CLKS(clk_s), ~CLKS_MASK);
+}
+
+static void rv6xx_set_memory_spread_spectrum_clk_v(struct radeon_device *rdev,
+						   u32 clk_v)
+{
+	WREG32_P(CG_MPLL_SPREAD_SPECTRUM, CLKV(clk_v), ~CLKV_MASK);
+}
+
+static void rv6xx_enable_memory_spread_spectrum(struct radeon_device *rdev,
+						bool enable)
+{
+	if (enable)
+		WREG32_P(CG_MPLL_SPREAD_SPECTRUM, SSEN, ~SSEN);
+	else
+		WREG32_P(CG_MPLL_SPREAD_SPECTRUM, 0, ~SSEN);
+}
+
+static void rv6xx_enable_dynamic_spread_spectrum(struct radeon_device *rdev,
+						 bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, ~DYN_SPREAD_SPECTRUM_EN);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~DYN_SPREAD_SPECTRUM_EN);
+}
+
+static void rv6xx_memory_clock_entry_enable_post_divider(struct radeon_device *rdev,
+							 u32 index, bool enable)
+{
+	if (enable)
+		WREG32_P(MPLL_FREQ_LEVEL_0 + (index * 4),
+			 LEVEL0_MPLL_DIV_EN, ~LEVEL0_MPLL_DIV_EN);
+	else
+		WREG32_P(MPLL_FREQ_LEVEL_0 + (index * 4), 0, ~LEVEL0_MPLL_DIV_EN);
+}
+
+static void rv6xx_memory_clock_entry_set_post_divider(struct radeon_device *rdev,
+						      u32 index, u32 divider)
+{
+	WREG32_P(MPLL_FREQ_LEVEL_0 + (index * 4),
+		 LEVEL0_MPLL_POST_DIV(divider), ~LEVEL0_MPLL_POST_DIV_MASK);
+}
+
+static void rv6xx_memory_clock_entry_set_feedback_divider(struct radeon_device *rdev,
+							  u32 index, u32 divider)
+{
+	WREG32_P(MPLL_FREQ_LEVEL_0 + (index * 4), LEVEL0_MPLL_FB_DIV(divider),
+		 ~LEVEL0_MPLL_FB_DIV_MASK);
+}
+
+static void rv6xx_memory_clock_entry_set_reference_divider(struct radeon_device *rdev,
+							   u32 index, u32 divider)
+{
+	WREG32_P(MPLL_FREQ_LEVEL_0 + (index * 4),
+		 LEVEL0_MPLL_REF_DIV(divider), ~LEVEL0_MPLL_REF_DIV_MASK);
+}
+
+static void rv6xx_vid_response_set_brt(struct radeon_device *rdev, u32 rt)
+{
+	WREG32_P(VID_RT, BRT(rt), ~BRT_MASK);
+}
+
+static void rv6xx_enable_engine_feedback_and_reference_sync(struct radeon_device *rdev)
+{
+	WREG32_P(SPLL_CNTL_MODE, SPLL_DIV_SYNC, ~SPLL_DIV_SYNC);
+}
+
+static u64 rv6xx_clocks_per_unit(u32 unit)
+{
+	u64 tmp = 1 << (2 * unit);
+
+	return tmp;
+}
+
+static u32 rv6xx_scale_count_given_unit(struct radeon_device *rdev,
+					u32 unscaled_count, u32 unit)
+{
+	u32 count_per_unit = (u32)rv6xx_clocks_per_unit(unit);
+
+	return (unscaled_count + count_per_unit - 1) / count_per_unit;
+}
+
+static u32 rv6xx_compute_count_for_delay(struct radeon_device *rdev,
+					 u32 delay_us, u32 unit)
+{
+	u32 ref_clk = rdev->clock.spll.reference_freq;
+
+	return rv6xx_scale_count_given_unit(rdev, delay_us * (ref_clk / 100), unit);
+}
+
+static void rv6xx_calculate_engine_speed_stepping_parameters(struct radeon_device *rdev,
+							     struct rv6xx_ps *state)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	pi->hw.sclks[R600_POWER_LEVEL_LOW] =
+		state->low.sclk;
+	pi->hw.sclks[R600_POWER_LEVEL_MEDIUM] =
+		state->medium.sclk;
+	pi->hw.sclks[R600_POWER_LEVEL_HIGH] =
+		state->high.sclk;
+
+	pi->hw.low_sclk_index = R600_POWER_LEVEL_LOW;
+	pi->hw.medium_sclk_index = R600_POWER_LEVEL_MEDIUM;
+	pi->hw.high_sclk_index = R600_POWER_LEVEL_HIGH;
+}
+
+static void rv6xx_calculate_memory_clock_stepping_parameters(struct radeon_device *rdev,
+							     struct rv6xx_ps *state)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	pi->hw.mclks[R600_POWER_LEVEL_CTXSW] =
+		state->high.mclk;
+	pi->hw.mclks[R600_POWER_LEVEL_HIGH] =
+		state->high.mclk;
+	pi->hw.mclks[R600_POWER_LEVEL_MEDIUM] =
+		state->medium.mclk;
+	pi->hw.mclks[R600_POWER_LEVEL_LOW] =
+		state->low.mclk;
+
+	pi->hw.high_mclk_index = R600_POWER_LEVEL_HIGH;
+
+	if (state->high.mclk == state->medium.mclk)
+		pi->hw.medium_mclk_index =
+			pi->hw.high_mclk_index;
+	else
+		pi->hw.medium_mclk_index = R600_POWER_LEVEL_MEDIUM;
+
+
+	if (state->medium.mclk == state->low.mclk)
+		pi->hw.low_mclk_index =
+			pi->hw.medium_mclk_index;
+	else
+		pi->hw.low_mclk_index = R600_POWER_LEVEL_LOW;
+}
+
+static void rv6xx_calculate_voltage_stepping_parameters(struct radeon_device *rdev,
+							struct rv6xx_ps *state)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	pi->hw.vddc[R600_POWER_LEVEL_CTXSW] = state->high.vddc;
+	pi->hw.vddc[R600_POWER_LEVEL_HIGH] = state->high.vddc;
+	pi->hw.vddc[R600_POWER_LEVEL_MEDIUM] = state->medium.vddc;
+	pi->hw.vddc[R600_POWER_LEVEL_LOW] = state->low.vddc;
+
+	pi->hw.backbias[R600_POWER_LEVEL_CTXSW] =
+		(state->high.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ? true : false;
+	pi->hw.backbias[R600_POWER_LEVEL_HIGH] =
+		(state->high.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ? true : false;
+	pi->hw.backbias[R600_POWER_LEVEL_MEDIUM] =
+		(state->medium.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ? true : false;
+	pi->hw.backbias[R600_POWER_LEVEL_LOW] =
+		(state->low.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ? true : false;
+
+	pi->hw.pcie_gen2[R600_POWER_LEVEL_HIGH] =
+		(state->high.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? true : false;
+	pi->hw.pcie_gen2[R600_POWER_LEVEL_MEDIUM] =
+		(state->medium.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? true : false;
+	pi->hw.pcie_gen2[R600_POWER_LEVEL_LOW] =
+		(state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? true : false;
+
+	pi->hw.high_vddc_index = R600_POWER_LEVEL_HIGH;
+
+	if ((state->high.vddc == state->medium.vddc) &&
+	    ((state->high.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ==
+	     (state->medium.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE)))
+		pi->hw.medium_vddc_index =
+			pi->hw.high_vddc_index;
+	else
+		pi->hw.medium_vddc_index = R600_POWER_LEVEL_MEDIUM;
+
+	if ((state->medium.vddc == state->low.vddc) &&
+	    ((state->medium.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ==
+	     (state->low.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE)))
+		pi->hw.low_vddc_index =
+			pi->hw.medium_vddc_index;
+	else
+		pi->hw.medium_vddc_index = R600_POWER_LEVEL_LOW;
+}
+
+static inline u32 rv6xx_calculate_vco_frequency(u32 ref_clock,
+						struct atom_clock_dividers *dividers,
+						u32 fb_divider_scale)
+{
+	return ref_clock * ((dividers->fb_div & ~1) << fb_divider_scale) /
+		(dividers->ref_div + 1);
+}
+
+static inline u32 rv6xx_calculate_spread_spectrum_clk_v(u32 vco_freq, u32 ref_freq,
+							u32 ss_rate, u32 ss_percent,
+							u32 fb_divider_scale)
+{
+	u32 fb_divider = vco_freq / ref_freq;
+
+	return (ss_percent * ss_rate * 4 * (fb_divider * fb_divider) /
+		(5375 * ((vco_freq * 10) / (4096 >> fb_divider_scale))));
+}
+
+static inline u32 rv6xx_calculate_spread_spectrum_clk_s(u32 ss_rate, u32 ref_freq)
+{
+	return (((ref_freq * 10) / (ss_rate * 2)) - 1) / 4;
+}
+
+static void rv6xx_program_engine_spread_spectrum(struct radeon_device *rdev,
+						 u32 clock, enum r600_power_level level)
+{
+	u32 ref_clk = rdev->clock.spll.reference_freq;
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	struct radeon_atom_ss ss;
+	u32 vco_freq, clk_v, clk_s;
+
+	rv6xx_enable_engine_spread_spectrum(rdev, level, false);
+
+	if (clock && pi->sclk_ss) {
+		if (radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM, clock, false, &dividers) == 0) {
+			vco_freq = rv6xx_calculate_vco_frequency(ref_clk, &dividers,
+								 pi->fb_div_scale);
+
+			if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+							     ASIC_INTERNAL_ENGINE_SS, vco_freq)) {
+				clk_v = rv6xx_calculate_spread_spectrum_clk_v(vco_freq,
+									      (ref_clk / (dividers.ref_div + 1)),
+									      ss.rate,
+									      ss.percentage,
+									      pi->fb_div_scale);
+
+				clk_s = rv6xx_calculate_spread_spectrum_clk_s(ss.rate,
+									      (ref_clk / (dividers.ref_div + 1)));
+
+				rv6xx_set_engine_spread_spectrum_clk_v(rdev, level, clk_v);
+				rv6xx_set_engine_spread_spectrum_clk_s(rdev, level, clk_s);
+				rv6xx_enable_engine_spread_spectrum(rdev, level, true);
+			}
+		}
+	}
+}
+
+static void rv6xx_program_sclk_spread_spectrum_parameters_except_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_program_engine_spread_spectrum(rdev,
+					     pi->hw.sclks[R600_POWER_LEVEL_HIGH],
+					     R600_POWER_LEVEL_HIGH);
+
+	rv6xx_program_engine_spread_spectrum(rdev,
+					     pi->hw.sclks[R600_POWER_LEVEL_MEDIUM],
+					     R600_POWER_LEVEL_MEDIUM);
+
+}
+
+static int rv6xx_program_mclk_stepping_entry(struct radeon_device *rdev,
+					     u32 entry, u32 clock)
+{
+	struct atom_clock_dividers dividers;
+
+	if (radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM, clock, false, &dividers))
+	    return -EINVAL;
+
+
+	rv6xx_memory_clock_entry_set_reference_divider(rdev, entry, dividers.ref_div);
+	rv6xx_memory_clock_entry_set_feedback_divider(rdev, entry, dividers.fb_div);
+	rv6xx_memory_clock_entry_set_post_divider(rdev, entry, dividers.post_div);
+
+	if (dividers.enable_post_div)
+		rv6xx_memory_clock_entry_enable_post_divider(rdev, entry, true);
+	else
+		rv6xx_memory_clock_entry_enable_post_divider(rdev, entry, false);
+
+	return 0;
+}
+
+static void rv6xx_program_mclk_stepping_parameters_except_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	int i;
+
+	for (i = 1; i < R600_PM_NUMBER_OF_MCLKS; i++) {
+		if (pi->hw.mclks[i])
+			rv6xx_program_mclk_stepping_entry(rdev, i,
+							  pi->hw.mclks[i]);
+	}
+}
+
+static void rv6xx_find_memory_clock_with_highest_vco(struct radeon_device *rdev,
+						     u32 requested_memory_clock,
+						     u32 ref_clk,
+						     struct atom_clock_dividers *dividers,
+						     u32 *vco_freq)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	struct atom_clock_dividers req_dividers;
+	u32 vco_freq_temp;
+
+	if (radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					   requested_memory_clock, false, &req_dividers) == 0) {
+		vco_freq_temp = rv6xx_calculate_vco_frequency(ref_clk, &req_dividers,
+							      pi->fb_div_scale);
+
+		if (vco_freq_temp > *vco_freq) {
+			*dividers = req_dividers;
+			*vco_freq = vco_freq_temp;
+		}
+	}
+}
+
+static void rv6xx_program_mclk_spread_spectrum_parameters(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	u32 ref_clk = rdev->clock.mpll.reference_freq;
+	struct atom_clock_dividers dividers;
+	struct radeon_atom_ss ss;
+	u32 vco_freq = 0, clk_v, clk_s;
+
+	rv6xx_enable_memory_spread_spectrum(rdev, false);
+
+	if (pi->mclk_ss) {
+		rv6xx_find_memory_clock_with_highest_vco(rdev,
+							 pi->hw.mclks[pi->hw.high_mclk_index],
+							 ref_clk,
+							 &dividers,
+							 &vco_freq);
+
+		rv6xx_find_memory_clock_with_highest_vco(rdev,
+							 pi->hw.mclks[pi->hw.medium_mclk_index],
+							 ref_clk,
+							 &dividers,
+							 &vco_freq);
+
+		rv6xx_find_memory_clock_with_highest_vco(rdev,
+							 pi->hw.mclks[pi->hw.low_mclk_index],
+							 ref_clk,
+							 &dividers,
+							 &vco_freq);
+
+		if (vco_freq) {
+			if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+							     ASIC_INTERNAL_MEMORY_SS, vco_freq)) {
+				clk_v = rv6xx_calculate_spread_spectrum_clk_v(vco_freq,
+									     (ref_clk / (dividers.ref_div + 1)),
+									     ss.rate,
+									     ss.percentage,
+									     pi->fb_div_scale);
+
+				clk_s = rv6xx_calculate_spread_spectrum_clk_s(ss.rate,
+									     (ref_clk / (dividers.ref_div + 1)));
+
+				rv6xx_set_memory_spread_spectrum_clk_v(rdev, clk_v);
+				rv6xx_set_memory_spread_spectrum_clk_s(rdev, clk_s);
+				rv6xx_enable_memory_spread_spectrum(rdev, true);
+			}
+		}
+	}
+}
+
+static int rv6xx_program_voltage_stepping_entry(struct radeon_device *rdev,
+						u32 entry, u16 voltage)
+{
+	u32 mask, set_pins;
+	int ret;
+
+	ret = radeon_atom_get_voltage_gpio_settings(rdev, voltage,
+						    SET_VOLTAGE_TYPE_ASIC_VDDC,
+						    &set_pins, &mask);
+	if (ret)
+		return ret;
+
+	r600_voltage_control_program_voltages(rdev, entry, set_pins);
+
+	return 0;
+}
+
+static void rv6xx_program_voltage_stepping_parameters_except_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	int i;
+
+	for (i = 1; i < R600_PM_NUMBER_OF_VOLTAGE_LEVELS; i++)
+		rv6xx_program_voltage_stepping_entry(rdev, i,
+						     pi->hw.vddc[i]);
+
+}
+
+static void rv6xx_program_backbias_stepping_parameters_except_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if (pi->hw.backbias[1])
+		WREG32_P(VID_UPPER_GPIO_CNTL, MEDIUM_BACKBIAS_VALUE, ~MEDIUM_BACKBIAS_VALUE);
+	else
+		WREG32_P(VID_UPPER_GPIO_CNTL, 0, ~MEDIUM_BACKBIAS_VALUE);
+
+	if (pi->hw.backbias[2])
+		WREG32_P(VID_UPPER_GPIO_CNTL, HIGH_BACKBIAS_VALUE, ~HIGH_BACKBIAS_VALUE);
+	else
+		WREG32_P(VID_UPPER_GPIO_CNTL, 0, ~HIGH_BACKBIAS_VALUE);
+}
+
+static void rv6xx_program_sclk_spread_spectrum_parameters_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_program_engine_spread_spectrum(rdev,
+					     pi->hw.sclks[R600_POWER_LEVEL_LOW],
+					     R600_POWER_LEVEL_LOW);
+}
+
+static void rv6xx_program_mclk_stepping_parameters_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if (pi->hw.mclks[0])
+		rv6xx_program_mclk_stepping_entry(rdev, 0,
+						  pi->hw.mclks[0]);
+}
+
+static void rv6xx_program_voltage_stepping_parameters_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_program_voltage_stepping_entry(rdev, 0,
+					     pi->hw.vddc[0]);
+
+}
+
+static void rv6xx_program_backbias_stepping_parameters_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if (pi->hw.backbias[0])
+		WREG32_P(VID_UPPER_GPIO_CNTL, LOW_BACKBIAS_VALUE, ~LOW_BACKBIAS_VALUE);
+	else
+		WREG32_P(VID_UPPER_GPIO_CNTL, 0, ~LOW_BACKBIAS_VALUE);
+}
+
+static u32 calculate_memory_refresh_rate(struct radeon_device *rdev,
+					 u32 engine_clock)
+{
+	u32 dram_rows, dram_refresh_rate;
+	u32 tmp;
+
+	tmp = (RREG32(RAMCFG) & NOOFROWS_MASK) >> NOOFROWS_SHIFT;
+	dram_rows = 1 << (tmp + 10);
+	dram_refresh_rate = 1 << ((RREG32(MC_SEQ_RESERVE_M) & 0x3) + 3);
+
+	return ((engine_clock * 10) * dram_refresh_rate / dram_rows - 32) / 64;
+}
+
+static void rv6xx_program_memory_timing_parameters(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	u32 sqm_ratio;
+	u32 arb_refresh_rate;
+	u32 high_clock;
+
+	if (pi->hw.sclks[R600_POWER_LEVEL_HIGH] <
+	    (pi->hw.sclks[R600_POWER_LEVEL_LOW] * 0xFF / 0x40))
+		high_clock = pi->hw.sclks[R600_POWER_LEVEL_HIGH];
+	else
+		high_clock =
+			pi->hw.sclks[R600_POWER_LEVEL_LOW] * 0xFF / 0x40;
+
+	radeon_atom_set_engine_dram_timings(rdev, high_clock, 0);
+
+	sqm_ratio = (STATE0(64 * high_clock / pi->hw.sclks[R600_POWER_LEVEL_LOW]) |
+		     STATE1(64 * high_clock / pi->hw.sclks[R600_POWER_LEVEL_MEDIUM]) |
+		     STATE2(64 * high_clock / pi->hw.sclks[R600_POWER_LEVEL_HIGH]) |
+		     STATE3(64 * high_clock / pi->hw.sclks[R600_POWER_LEVEL_HIGH]));
+	WREG32(SQM_RATIO, sqm_ratio);
+
+	arb_refresh_rate =
+		(POWERMODE0(calculate_memory_refresh_rate(rdev,
+							  pi->hw.sclks[R600_POWER_LEVEL_LOW])) |
+		 POWERMODE1(calculate_memory_refresh_rate(rdev,
+							  pi->hw.sclks[R600_POWER_LEVEL_MEDIUM])) |
+		 POWERMODE2(calculate_memory_refresh_rate(rdev,
+							  pi->hw.sclks[R600_POWER_LEVEL_MEDIUM])) |
+		 POWERMODE3(calculate_memory_refresh_rate(rdev,
+							  pi->hw.sclks[R600_POWER_LEVEL_HIGH])));
+	WREG32(ARB_RFSH_RATE, arb_refresh_rate);
+}
+
+static void rv6xx_program_mpll_timing_parameters(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	r600_set_mpll_lock_time(rdev, R600_MPLLLOCKTIME_DFLT *
+				pi->mpll_ref_div);
+	r600_set_mpll_reset_time(rdev, R600_MPLLRESETTIME_DFLT);
+}
+
+static void rv6xx_program_bsp(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	u32 ref_clk = rdev->clock.spll.reference_freq;
+
+	r600_calculate_u_and_p(R600_ASI_DFLT,
+			       ref_clk, 16,
+			       &pi->bsp,
+			       &pi->bsu);
+
+	r600_set_bsp(rdev, pi->bsu, pi->bsp);
+}
+
+static void rv6xx_program_at(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	r600_set_at(rdev,
+		    (pi->hw.rp[0] * pi->bsp) / 200,
+		    (pi->hw.rp[1] * pi->bsp) / 200,
+		    (pi->hw.lp[2] * pi->bsp) / 200,
+		    (pi->hw.lp[1] * pi->bsp) / 200);
+}
+
+static void rv6xx_program_git(struct radeon_device *rdev)
+{
+	r600_set_git(rdev, R600_GICST_DFLT);
+}
+
+static void rv6xx_program_tp(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < R600_PM_NUMBER_OF_TC; i++)
+		r600_set_tc(rdev, i, r600_utc[i], r600_dtc[i]);
+
+	r600_select_td(rdev, R600_TD_DFLT);
+}
+
+static void rv6xx_program_vc(struct radeon_device *rdev)
+{
+	r600_set_vrc(rdev, R600_VRC_DFLT);
+}
+
+static void rv6xx_clear_vc(struct radeon_device *rdev)
+{
+	r600_set_vrc(rdev, 0);
+}
+
+static void rv6xx_program_tpp(struct radeon_device *rdev)
+{
+	r600_set_tpu(rdev, R600_TPU_DFLT);
+	r600_set_tpc(rdev, R600_TPC_DFLT);
+}
+
+static void rv6xx_program_sstp(struct radeon_device *rdev)
+{
+	r600_set_sstu(rdev, R600_SSTU_DFLT);
+	r600_set_sst(rdev, R600_SST_DFLT);
+}
+
+static void rv6xx_program_fcp(struct radeon_device *rdev)
+{
+	r600_set_fctu(rdev, R600_FCTU_DFLT);
+	r600_set_fct(rdev, R600_FCT_DFLT);
+}
+
+static void rv6xx_program_vddc3d_parameters(struct radeon_device *rdev)
+{
+	r600_set_vddc3d_oorsu(rdev, R600_VDDC3DOORSU_DFLT);
+	r600_set_vddc3d_oorphc(rdev, R600_VDDC3DOORPHC_DFLT);
+	r600_set_vddc3d_oorsdc(rdev, R600_VDDC3DOORSDC_DFLT);
+	r600_set_ctxcgtt3d_rphc(rdev, R600_CTXCGTT3DRPHC_DFLT);
+	r600_set_ctxcgtt3d_rsdc(rdev, R600_CTXCGTT3DRSDC_DFLT);
+}
+
+static void rv6xx_program_voltage_timing_parameters(struct radeon_device *rdev)
+{
+	u32 rt;
+
+	r600_vid_rt_set_vru(rdev, R600_VRU_DFLT);
+
+	r600_vid_rt_set_vrt(rdev,
+			    rv6xx_compute_count_for_delay(rdev,
+							  rdev->pm.dpm.voltage_response_time,
+							  R600_VRU_DFLT));
+
+	rt = rv6xx_compute_count_for_delay(rdev,
+					   rdev->pm.dpm.backbias_response_time,
+					   R600_VRU_DFLT);
+
+	rv6xx_vid_response_set_brt(rdev, (rt + 0x1F) >> 5);
+}
+
+static void rv6xx_program_engine_speed_parameters(struct radeon_device *rdev)
+{
+	r600_vid_rt_set_ssu(rdev, R600_SPLLSTEPUNIT_DFLT);
+	rv6xx_enable_engine_feedback_and_reference_sync(rdev);
+}
+
+static u64 rv6xx_get_master_voltage_mask(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	u64 master_mask = 0;
+	int i;
+
+	for (i = 0; i < R600_PM_NUMBER_OF_VOLTAGE_LEVELS; i++) {
+		u32 tmp_mask, tmp_set_pins;
+		int ret;
+
+		ret = radeon_atom_get_voltage_gpio_settings(rdev,
+							    pi->hw.vddc[i],
+							    SET_VOLTAGE_TYPE_ASIC_VDDC,
+							    &tmp_set_pins, &tmp_mask);
+
+		if (ret == 0)
+			master_mask |= tmp_mask;
+	}
+
+	return master_mask;
+}
+
+static void rv6xx_program_voltage_gpio_pins(struct radeon_device *rdev)
+{
+	r600_voltage_control_enable_pins(rdev,
+					 rv6xx_get_master_voltage_mask(rdev));
+}
+
+static void rv6xx_enable_static_voltage_control(struct radeon_device *rdev,
+						struct radeon_ps *new_ps,
+						bool enable)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+
+	if (enable)
+		radeon_atom_set_voltage(rdev,
+					new_state->low.vddc,
+					SET_VOLTAGE_TYPE_ASIC_VDDC);
+	else
+		r600_voltage_control_deactivate_static_control(rdev,
+							       rv6xx_get_master_voltage_mask(rdev));
+}
+
+static void rv6xx_enable_display_gap(struct radeon_device *rdev, bool enable)
+{
+	if (enable) {
+		u32 tmp = (DISP1_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM) |
+			   DISP2_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM) |
+			   DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE) |
+			   DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE) |
+			   VBI_TIMER_COUNT(0x3FFF) |
+			   VBI_TIMER_UNIT(7));
+		WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+
+		WREG32_P(MCLK_PWRMGT_CNTL, USE_DISPLAY_GAP, ~USE_DISPLAY_GAP);
+	} else
+		WREG32_P(MCLK_PWRMGT_CNTL, 0, ~USE_DISPLAY_GAP);
+}
+
+static void rv6xx_program_power_level_enter_state(struct radeon_device *rdev)
+{
+	r600_power_level_set_enter_index(rdev, R600_POWER_LEVEL_MEDIUM);
+}
+
+static void rv6xx_calculate_t(u32 l_f, u32 h_f, int h,
+			      int d_l, int d_r, u8 *l, u8 *r)
+{
+	int a_n, a_d, h_r, l_r;
+
+	h_r = d_l;
+	l_r = 100 - d_r;
+
+	a_n = (int)h_f * d_l + (int)l_f * (h - d_r);
+	a_d = (int)l_f * l_r + (int)h_f * h_r;
+
+	if (a_d != 0) {
+		*l = d_l - h_r * a_n / a_d;
+		*r = d_r + l_r * a_n / a_d;
+	}
+}
+
+static void rv6xx_calculate_ap(struct radeon_device *rdev,
+			       struct rv6xx_ps *state)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	pi->hw.lp[0] = 0;
+	pi->hw.rp[R600_PM_NUMBER_OF_ACTIVITY_LEVELS - 1]
+		= 100;
+
+	rv6xx_calculate_t(state->low.sclk,
+			  state->medium.sclk,
+			  R600_AH_DFLT,
+			  R600_LMP_DFLT,
+			  R600_RLP_DFLT,
+			  &pi->hw.lp[1],
+			  &pi->hw.rp[0]);
+
+	rv6xx_calculate_t(state->medium.sclk,
+			  state->high.sclk,
+			  R600_AH_DFLT,
+			  R600_LHP_DFLT,
+			  R600_RMP_DFLT,
+			  &pi->hw.lp[2],
+			  &pi->hw.rp[1]);
+
+}
+
+static void rv6xx_calculate_stepping_parameters(struct radeon_device *rdev,
+						struct radeon_ps *new_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+
+	rv6xx_calculate_engine_speed_stepping_parameters(rdev, new_state);
+	rv6xx_calculate_memory_clock_stepping_parameters(rdev, new_state);
+	rv6xx_calculate_voltage_stepping_parameters(rdev, new_state);
+	rv6xx_calculate_ap(rdev, new_state);
+}
+
+static void rv6xx_program_stepping_parameters_except_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_program_mclk_stepping_parameters_except_lowest_entry(rdev);
+	if (pi->voltage_control)
+		rv6xx_program_voltage_stepping_parameters_except_lowest_entry(rdev);
+	rv6xx_program_backbias_stepping_parameters_except_lowest_entry(rdev);
+	rv6xx_program_sclk_spread_spectrum_parameters_except_lowest_entry(rdev);
+	rv6xx_program_mclk_spread_spectrum_parameters(rdev);
+	rv6xx_program_memory_timing_parameters(rdev);
+}
+
+static void rv6xx_program_stepping_parameters_lowest_entry(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_program_mclk_stepping_parameters_lowest_entry(rdev);
+	if (pi->voltage_control)
+		rv6xx_program_voltage_stepping_parameters_lowest_entry(rdev);
+	rv6xx_program_backbias_stepping_parameters_lowest_entry(rdev);
+	rv6xx_program_sclk_spread_spectrum_parameters_lowest_entry(rdev);
+}
+
+static void rv6xx_program_power_level_low(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_LOW,
+					   pi->hw.low_vddc_index);
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_LOW,
+					     pi->hw.low_mclk_index);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_LOW,
+					     pi->hw.low_sclk_index);
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_LOW,
+					  R600_DISPLAY_WATERMARK_LOW);
+	r600_power_level_set_pcie_gen2(rdev, R600_POWER_LEVEL_LOW,
+				       pi->hw.pcie_gen2[R600_POWER_LEVEL_LOW]);
+}
+
+static void rv6xx_program_power_level_low_to_lowest_state(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_LOW, 0);
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_LOW, 0);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_LOW, 0);
+
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_LOW,
+					  R600_DISPLAY_WATERMARK_LOW);
+
+	r600_power_level_set_pcie_gen2(rdev, R600_POWER_LEVEL_LOW,
+				       pi->hw.pcie_gen2[R600_POWER_LEVEL_LOW]);
+
+}
+
+static void rv6xx_program_power_level_medium(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_MEDIUM,
+					  pi->hw.medium_vddc_index);
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_MEDIUM,
+					    pi->hw.medium_mclk_index);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_MEDIUM,
+					    pi->hw.medium_sclk_index);
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_MEDIUM,
+					 R600_DISPLAY_WATERMARK_LOW);
+	r600_power_level_set_pcie_gen2(rdev, R600_POWER_LEVEL_MEDIUM,
+				      pi->hw.pcie_gen2[R600_POWER_LEVEL_MEDIUM]);
+}
+
+static void rv6xx_program_power_level_medium_for_transition(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_program_mclk_stepping_entry(rdev,
+					  R600_POWER_LEVEL_CTXSW,
+					  pi->hw.mclks[pi->hw.low_mclk_index]);
+
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_MEDIUM, 1);
+
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_MEDIUM,
+					     R600_POWER_LEVEL_CTXSW);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_MEDIUM,
+					     pi->hw.medium_sclk_index);
+
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_MEDIUM,
+					  R600_DISPLAY_WATERMARK_LOW);
+
+	rv6xx_enable_engine_spread_spectrum(rdev, R600_POWER_LEVEL_MEDIUM, false);
+
+	r600_power_level_set_pcie_gen2(rdev, R600_POWER_LEVEL_MEDIUM,
+				       pi->hw.pcie_gen2[R600_POWER_LEVEL_LOW]);
+}
+
+static void rv6xx_program_power_level_high(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	r600_power_level_set_voltage_index(rdev, R600_POWER_LEVEL_HIGH,
+					   pi->hw.high_vddc_index);
+	r600_power_level_set_mem_clock_index(rdev, R600_POWER_LEVEL_HIGH,
+					     pi->hw.high_mclk_index);
+	r600_power_level_set_eng_clock_index(rdev, R600_POWER_LEVEL_HIGH,
+					     pi->hw.high_sclk_index);
+
+	r600_power_level_set_watermark_id(rdev, R600_POWER_LEVEL_HIGH,
+					  R600_DISPLAY_WATERMARK_HIGH);
+
+	r600_power_level_set_pcie_gen2(rdev, R600_POWER_LEVEL_HIGH,
+				       pi->hw.pcie_gen2[R600_POWER_LEVEL_HIGH]);
+}
+
+static void rv6xx_enable_backbias(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, BACKBIAS_PAD_EN | BACKBIAS_DPM_CNTL,
+			 ~(BACKBIAS_PAD_EN | BACKBIAS_DPM_CNTL));
+	else
+		WREG32_P(GENERAL_PWRMGT, 0,
+			 ~(BACKBIAS_VALUE | BACKBIAS_PAD_EN | BACKBIAS_DPM_CNTL));
+}
+
+static void rv6xx_program_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL);
+
+	tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK);
+	if (RREG32(AVIVO_D1CRTC_CONTROL) & AVIVO_CRTC_EN) {
+		tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK);
+		tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+	} else if (RREG32(AVIVO_D2CRTC_CONTROL) & AVIVO_CRTC_EN) {
+		tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+		tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK);
+	} else {
+		tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+		tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+	}
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+}
+
+static void rv6xx_set_sw_voltage_to_safe(struct radeon_device *rdev,
+					 struct radeon_ps *new_ps,
+					 struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+	u16 safe_voltage;
+
+	safe_voltage = (new_state->low.vddc >= old_state->low.vddc) ?
+		new_state->low.vddc : old_state->low.vddc;
+
+	rv6xx_program_voltage_stepping_entry(rdev, R600_POWER_LEVEL_CTXSW,
+					     safe_voltage);
+
+	WREG32_P(GENERAL_PWRMGT, SW_GPIO_INDEX(R600_POWER_LEVEL_CTXSW),
+		 ~SW_GPIO_INDEX_MASK);
+}
+
+static void rv6xx_set_sw_voltage_to_low(struct radeon_device *rdev,
+					struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+
+	rv6xx_program_voltage_stepping_entry(rdev, R600_POWER_LEVEL_CTXSW,
+					     old_state->low.vddc);
+
+	WREG32_P(GENERAL_PWRMGT, SW_GPIO_INDEX(R600_POWER_LEVEL_CTXSW),
+		~SW_GPIO_INDEX_MASK);
+}
+
+static void rv6xx_set_safe_backbias(struct radeon_device *rdev,
+				    struct radeon_ps *new_ps,
+				    struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+
+	if ((new_state->low.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) &&
+	    (old_state->low.flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE))
+		WREG32_P(GENERAL_PWRMGT, BACKBIAS_VALUE, ~BACKBIAS_VALUE);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~BACKBIAS_VALUE);
+}
+
+static void rv6xx_set_safe_pcie_gen2(struct radeon_device *rdev,
+				     struct radeon_ps *new_ps,
+				     struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+
+	if ((new_state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) !=
+	    (old_state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2))
+		rv6xx_force_pcie_gen1(rdev);
+}
+
+static void rv6xx_enable_dynamic_voltage_control(struct radeon_device *rdev,
+						 bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, VOLT_PWRMGT_EN, ~VOLT_PWRMGT_EN);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~VOLT_PWRMGT_EN);
+}
+
+static void rv6xx_enable_dynamic_backbias_control(struct radeon_device *rdev,
+						  bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, BACKBIAS_DPM_CNTL, ~BACKBIAS_DPM_CNTL);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~BACKBIAS_DPM_CNTL);
+}
+
+static int rv6xx_step_sw_voltage(struct radeon_device *rdev,
+				 u16 initial_voltage,
+				 u16 target_voltage)
+{
+	u16 current_voltage;
+	u16 true_target_voltage;
+	u16 voltage_step;
+	int signed_voltage_step;
+
+	if ((radeon_atom_get_voltage_step(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+					  &voltage_step)) ||
+	    (radeon_atom_round_to_true_voltage(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+					       initial_voltage, &current_voltage)) ||
+	    (radeon_atom_round_to_true_voltage(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC,
+					       target_voltage, &true_target_voltage)))
+		return -EINVAL;
+
+	if (true_target_voltage < current_voltage)
+		signed_voltage_step = -(int)voltage_step;
+	else
+		signed_voltage_step = voltage_step;
+
+	while (current_voltage != true_target_voltage) {
+		current_voltage += signed_voltage_step;
+		rv6xx_program_voltage_stepping_entry(rdev, R600_POWER_LEVEL_CTXSW,
+						     current_voltage);
+		msleep((rdev->pm.dpm.voltage_response_time + 999) / 1000);
+	}
+
+	return 0;
+}
+
+static int rv6xx_step_voltage_if_increasing(struct radeon_device *rdev,
+					    struct radeon_ps *new_ps,
+					    struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+
+	if (new_state->low.vddc > old_state->low.vddc)
+		return rv6xx_step_sw_voltage(rdev,
+					     old_state->low.vddc,
+					     new_state->low.vddc);
+
+	return 0;
+}
+
+static int rv6xx_step_voltage_if_decreasing(struct radeon_device *rdev,
+					    struct radeon_ps *new_ps,
+					    struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+
+	if (new_state->low.vddc < old_state->low.vddc)
+		return rv6xx_step_sw_voltage(rdev,
+					     old_state->low.vddc,
+					     new_state->low.vddc);
+	else
+		return 0;
+}
+
+static void rv6xx_enable_high(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if ((pi->restricted_levels < 1) ||
+	    (pi->restricted_levels == 3))
+		r600_power_level_enable(rdev, R600_POWER_LEVEL_HIGH, true);
+}
+
+static void rv6xx_enable_medium(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if (pi->restricted_levels < 2)
+		r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, true);
+}
+
+static void rv6xx_set_dpm_event_sources(struct radeon_device *rdev, u32 sources)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	bool want_thermal_protection;
+	enum radeon_dpm_event_src dpm_event_src;
+
+	switch (sources) {
+        case 0:
+        default:
+		want_thermal_protection = false;
+		break;
+        case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_DIGITAL;
+		break;
+
+        case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_EXTERNAL;
+		break;
+
+        case ((1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL) |
+	      (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL)):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL;
+		break;
+	}
+
+	if (want_thermal_protection) {
+		WREG32_P(CG_THERMAL_CTRL, DPM_EVENT_SRC(dpm_event_src), ~DPM_EVENT_SRC_MASK);
+		if (pi->thermal_protection)
+			WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS);
+	} else {
+		WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS);
+	}
+}
+
+static void rv6xx_enable_auto_throttle_source(struct radeon_device *rdev,
+					      enum radeon_dpm_auto_throttle_src source,
+					      bool enable)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if (enable) {
+		if (!(pi->active_auto_throttle_sources & (1 << source))) {
+			pi->active_auto_throttle_sources |= 1 << source;
+			rv6xx_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources);
+		}
+	} else {
+		if (pi->active_auto_throttle_sources & (1 << source)) {
+			pi->active_auto_throttle_sources &= ~(1 << source);
+			rv6xx_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources);
+		}
+	}
+}
+
+
+static void rv6xx_enable_thermal_protection(struct radeon_device *rdev,
+					    bool enable)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	if (pi->active_auto_throttle_sources)
+		r600_enable_thermal_protection(rdev, enable);
+}
+
+static void rv6xx_generate_transition_stepping(struct radeon_device *rdev,
+					       struct radeon_ps *new_ps,
+					       struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *old_state = rv6xx_get_ps(old_ps);
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_generate_steps(rdev,
+			     old_state->low.sclk,
+			     new_state->low.sclk,
+			     0, &pi->hw.medium_sclk_index);
+}
+
+static void rv6xx_generate_low_step(struct radeon_device *rdev,
+				    struct radeon_ps *new_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	pi->hw.low_sclk_index = 0;
+	rv6xx_generate_single_step(rdev,
+				   new_state->low.sclk,
+				   0);
+}
+
+static void rv6xx_invalidate_intermediate_steps(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	rv6xx_invalidate_intermediate_steps_range(rdev, 0,
+						  pi->hw.medium_sclk_index);
+}
+
+static void rv6xx_generate_stepping_table(struct radeon_device *rdev,
+					  struct radeon_ps *new_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+
+	pi->hw.low_sclk_index = 0;
+
+	rv6xx_generate_steps(rdev,
+			     new_state->low.sclk,
+			     new_state->medium.sclk,
+			     0,
+			     &pi->hw.medium_sclk_index);
+	rv6xx_generate_steps(rdev,
+			     new_state->medium.sclk,
+			     new_state->high.sclk,
+			     pi->hw.medium_sclk_index,
+			     &pi->hw.high_sclk_index);
+}
+
+static void rv6xx_enable_spread_spectrum(struct radeon_device *rdev,
+					 bool enable)
+{
+	if (enable)
+		rv6xx_enable_dynamic_spread_spectrum(rdev, true);
+	else {
+		rv6xx_enable_engine_spread_spectrum(rdev, R600_POWER_LEVEL_LOW, false);
+		rv6xx_enable_engine_spread_spectrum(rdev, R600_POWER_LEVEL_MEDIUM, false);
+		rv6xx_enable_engine_spread_spectrum(rdev, R600_POWER_LEVEL_HIGH, false);
+		rv6xx_enable_dynamic_spread_spectrum(rdev, false);
+		rv6xx_enable_memory_spread_spectrum(rdev, false);
+	}
+}
+
+static void rv6xx_reset_lvtm_data_sync(struct radeon_device *rdev)
+{
+	if (ASIC_IS_DCE3(rdev))
+		WREG32_P(DCE3_LVTMA_DATA_SYNCHRONIZATION, LVTMA_PFREQCHG, ~LVTMA_PFREQCHG);
+	else
+		WREG32_P(LVTMA_DATA_SYNCHRONIZATION, LVTMA_PFREQCHG, ~LVTMA_PFREQCHG);
+}
+
+static void rv6xx_enable_dynamic_pcie_gen2(struct radeon_device *rdev,
+					   struct radeon_ps *new_ps,
+					   bool enable)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+
+	if (enable) {
+		rv6xx_enable_bif_dynamic_pcie_gen2(rdev, true);
+		rv6xx_enable_pcie_gen2_support(rdev);
+		r600_enable_dynamic_pcie_gen2(rdev, true);
+	} else {
+		if (!(new_state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2))
+			rv6xx_force_pcie_gen1(rdev);
+		rv6xx_enable_bif_dynamic_pcie_gen2(rdev, false);
+		r600_enable_dynamic_pcie_gen2(rdev, false);
+	}
+}
+
+static void rv6xx_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+						     struct radeon_ps *new_ps,
+						     struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *current_state = rv6xx_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->high.sclk >= current_state->high.sclk)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+static void rv6xx_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+						    struct radeon_ps *new_ps,
+						    struct radeon_ps *old_ps)
+{
+	struct rv6xx_ps *new_state = rv6xx_get_ps(new_ps);
+	struct rv6xx_ps *current_state = rv6xx_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->high.sclk < current_state->high.sclk)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+int rv6xx_dpm_enable(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	if (r600_dynamicpm_enabled(rdev))
+		return -EINVAL;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv6xx_enable_backbias(rdev, true);
+
+	if (pi->dynamic_ss)
+		rv6xx_enable_spread_spectrum(rdev, true);
+
+	rv6xx_program_mpll_timing_parameters(rdev);
+	rv6xx_program_bsp(rdev);
+	rv6xx_program_git(rdev);
+	rv6xx_program_tp(rdev);
+	rv6xx_program_tpp(rdev);
+	rv6xx_program_sstp(rdev);
+	rv6xx_program_fcp(rdev);
+	rv6xx_program_vddc3d_parameters(rdev);
+	rv6xx_program_voltage_timing_parameters(rdev);
+	rv6xx_program_engine_speed_parameters(rdev);
+
+	rv6xx_enable_display_gap(rdev, true);
+	if (pi->display_gap == false)
+		rv6xx_enable_display_gap(rdev, false);
+
+	rv6xx_program_power_level_enter_state(rdev);
+
+	rv6xx_calculate_stepping_parameters(rdev, boot_ps);
+
+	if (pi->voltage_control)
+		rv6xx_program_voltage_gpio_pins(rdev);
+
+	rv6xx_generate_stepping_table(rdev, boot_ps);
+
+	rv6xx_program_stepping_parameters_except_lowest_entry(rdev);
+	rv6xx_program_stepping_parameters_lowest_entry(rdev);
+
+	rv6xx_program_power_level_low(rdev);
+	rv6xx_program_power_level_medium(rdev);
+	rv6xx_program_power_level_high(rdev);
+	rv6xx_program_vc(rdev);
+	rv6xx_program_at(rdev);
+
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, true);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, true);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_HIGH, true);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		ret = r600_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+	}
+
+	rv6xx_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
+
+	r600_start_dpm(rdev);
+
+	if (pi->voltage_control)
+		rv6xx_enable_static_voltage_control(rdev, boot_ps, false);
+
+	if (pi->dynamic_pcie_gen2)
+		rv6xx_enable_dynamic_pcie_gen2(rdev, boot_ps, true);
+
+	if (pi->gfx_clock_gating)
+		r600_gfx_clockgating_enable(rdev, true);
+
+	return 0;
+}
+
+void rv6xx_dpm_disable(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+
+	if (!r600_dynamicpm_enabled(rdev))
+		return;
+
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, true);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, true);
+	rv6xx_enable_display_gap(rdev, false);
+	rv6xx_clear_vc(rdev);
+	r600_set_at(rdev, 0xFFFF, 0xFFFF, 0xFFFF, 0xFFFF);
+
+	if (pi->thermal_protection)
+		r600_enable_thermal_protection(rdev, false);
+
+	r600_wait_for_power_level(rdev, R600_POWER_LEVEL_LOW);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_HIGH, false);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, false);
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv6xx_enable_backbias(rdev, false);
+
+	rv6xx_enable_spread_spectrum(rdev, false);
+
+	if (pi->voltage_control)
+		rv6xx_enable_static_voltage_control(rdev, boot_ps, true);
+
+	if (pi->dynamic_pcie_gen2)
+		rv6xx_enable_dynamic_pcie_gen2(rdev, boot_ps, false);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	if (pi->gfx_clock_gating)
+		r600_gfx_clockgating_enable(rdev, false);
+
+	r600_stop_dpm(rdev);
+}
+
+int rv6xx_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct rv6xx_power_info *pi = rv6xx_get_pi(rdev);
+	struct radeon_ps *new_ps = rdev->pm.dpm.requested_ps;
+	struct radeon_ps *old_ps = rdev->pm.dpm.current_ps;
+	int ret;
+
+	rv6xx_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+
+	rv6xx_clear_vc(rdev);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, true);
+	r600_set_at(rdev, 0xFFFF, 0xFFFF, 0xFFFF, 0xFFFF);
+
+	if (pi->thermal_protection)
+		r600_enable_thermal_protection(rdev, false);
+
+	r600_wait_for_power_level(rdev, R600_POWER_LEVEL_LOW);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_HIGH, false);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, false);
+
+	rv6xx_generate_transition_stepping(rdev, new_ps, old_ps);
+	rv6xx_program_power_level_medium_for_transition(rdev);
+
+	if (pi->voltage_control) {
+		rv6xx_set_sw_voltage_to_safe(rdev, new_ps, old_ps);
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+			rv6xx_set_sw_voltage_to_low(rdev, old_ps);
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv6xx_set_safe_backbias(rdev, new_ps, old_ps);
+
+	if (pi->dynamic_pcie_gen2)
+		rv6xx_set_safe_pcie_gen2(rdev, new_ps, old_ps);
+
+	if (pi->voltage_control)
+		rv6xx_enable_dynamic_voltage_control(rdev, false);
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv6xx_enable_dynamic_backbias_control(rdev, false);
+
+	if (pi->voltage_control) {
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+			rv6xx_step_voltage_if_increasing(rdev, new_ps, old_ps);
+		msleep((rdev->pm.dpm.voltage_response_time + 999) / 1000);
+	}
+
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, true);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, false);
+	r600_wait_for_power_level_unequal(rdev, R600_POWER_LEVEL_LOW);
+
+	rv6xx_generate_low_step(rdev, new_ps);
+	rv6xx_invalidate_intermediate_steps(rdev);
+	rv6xx_calculate_stepping_parameters(rdev, new_ps);
+	rv6xx_program_stepping_parameters_lowest_entry(rdev);
+	rv6xx_program_power_level_low_to_lowest_state(rdev);
+
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_LOW, true);
+	r600_wait_for_power_level(rdev, R600_POWER_LEVEL_LOW);
+	r600_power_level_enable(rdev, R600_POWER_LEVEL_MEDIUM, false);
+
+	if (pi->voltage_control) {
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC) {
+			ret = rv6xx_step_voltage_if_decreasing(rdev, new_ps, old_ps);
+			if (ret)
+				return ret;
+		}
+		rv6xx_enable_dynamic_voltage_control(rdev, true);
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv6xx_enable_dynamic_backbias_control(rdev, true);
+
+	if (pi->dynamic_pcie_gen2)
+		rv6xx_enable_dynamic_pcie_gen2(rdev, new_ps, true);
+
+	rv6xx_reset_lvtm_data_sync(rdev);
+
+	rv6xx_generate_stepping_table(rdev, new_ps);
+	rv6xx_program_stepping_parameters_except_lowest_entry(rdev);
+	rv6xx_program_power_level_low(rdev);
+	rv6xx_program_power_level_medium(rdev);
+	rv6xx_program_power_level_high(rdev);
+	rv6xx_enable_medium(rdev);
+	rv6xx_enable_high(rdev);
+
+	if (pi->thermal_protection)
+		rv6xx_enable_thermal_protection(rdev, true);
+	rv6xx_program_vc(rdev);
+	rv6xx_program_at(rdev);
+
+	rv6xx_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+
+	return 0;
+}
+
+void rv6xx_setup_asic(struct radeon_device *rdev)
+{
+	r600_enable_acpi_pm(rdev);
+
+	if (radeon_aspm != 0) {
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L0s)
+			rv6xx_enable_l0s(rdev);
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L1)
+			rv6xx_enable_l1(rdev);
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_TURNOFFPLL_ASPML1)
+			rv6xx_enable_pll_sleep_in_l1(rdev);
+	}
+}
+
+void rv6xx_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+	rv6xx_program_display_gap(rdev);
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void rv6xx_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					     struct radeon_ps *rps,
+					     struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info)
+{
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (r600_is_uvd_state(rps->class, rps->class2)) {
+		rps->vclk = RV6XX_DEFAULT_VCLK_FREQ;
+		rps->dclk = RV6XX_DEFAULT_DCLK_FREQ;
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT)
+		rdev->pm.dpm.boot_ps = rps;
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void rv6xx_parse_pplib_clock_info(struct radeon_device *rdev,
+					 struct radeon_ps *rps, int index,
+					 union pplib_clock_info *clock_info)
+{
+	struct rv6xx_ps *ps = rv6xx_get_ps(rps);
+	u32 sclk, mclk;
+	u16 vddc;
+	struct rv6xx_pl *pl;
+
+	switch (index) {
+	case 0:
+		pl = &ps->low;
+		break;
+	case 1:
+		pl = &ps->medium;
+		break;
+	case 2:
+	default:
+		pl = &ps->high;
+		break;
+	}
+
+	sclk = le16_to_cpu(clock_info->r600.usEngineClockLow);
+	sclk |= clock_info->r600.ucEngineClockHigh << 16;
+	mclk = le16_to_cpu(clock_info->r600.usMemoryClockLow);
+	mclk |= clock_info->r600.ucMemoryClockHigh << 16;
+
+	pl->mclk = mclk;
+	pl->sclk = sclk;
+	pl->vddc = le16_to_cpu(clock_info->r600.usVDDC);
+	pl->flags = le32_to_cpu(clock_info->r600.ulFlags);
+
+	/* patch up vddc if necessary */
+	if (pl->vddc == 0xff01) {
+		if (radeon_atom_get_max_vddc(rdev, 0, 0, &vddc) == 0)
+			pl->vddc = vddc;
+	}
+
+	/* fix up pcie gen2 */
+	if (pl->flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) {
+		if ((rdev->family == CHIP_RV610) || (rdev->family == CHIP_RV630)) {
+			if (pl->vddc < 1100)
+				pl->flags &= ~ATOM_PPLIB_R600_FLAGS_PCIEGEN2;
+		}
+	}
+
+	/* patch up boot state */
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		u16 vddc, vddci, mvdd;
+		radeon_atombios_get_default_voltages(rdev, &vddc, &vddci, &mvdd);
+		pl->mclk = rdev->clock.default_mclk;
+		pl->sclk = rdev->clock.default_sclk;
+		pl->vddc = vddc;
+	}
+}
+
+static int rv6xx_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i, j;
+	union pplib_clock_info *clock_info;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	struct rv6xx_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  power_info->pplib.ucNumStates, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+
+	for (i = 0; i < power_info->pplib.ucNumStates; i++) {
+		power_state = (union pplib_power_state *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usStateArrayOffset) +
+			 i * power_info->pplib.ucStateEntrySize);
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset) +
+			 (power_state->v1.ucNonClockStateIndex *
+			  power_info->pplib.ucNonClockSize));
+		if (power_info->pplib.ucStateEntrySize - 1) {
+			ps = kzalloc(sizeof(struct rv6xx_ps), GFP_KERNEL);
+			if (ps == NULL) {
+				kfree(rdev->pm.dpm.ps);
+				return -ENOMEM;
+			}
+			rdev->pm.dpm.ps[i].ps_priv = ps;
+			rv6xx_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+							 non_clock_info);
+			for (j = 0; j < (power_info->pplib.ucStateEntrySize - 1); j++) {
+				clock_info = (union pplib_clock_info *)
+					(mode_info->atom_context->bios + data_offset +
+					 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) +
+					 (power_state->v1.ucClockStateIndices[j] *
+					  power_info->pplib.ucClockInfoSize));
+				rv6xx_parse_pplib_clock_info(rdev,
+							     &rdev->pm.dpm.ps[i], j,
+							     clock_info);
+			}
+		}
+	}
+	rdev->pm.dpm.num_ps = power_info->pplib.ucNumStates;
+	return 0;
+}
+
+int rv6xx_dpm_init(struct radeon_device *rdev)
+{
+	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	uint16_t data_offset, size;
+	uint8_t frev, crev;
+	struct atom_clock_dividers dividers;
+	struct rv6xx_power_info *pi;
+	int ret;
+
+	pi = kzalloc(sizeof(struct rv6xx_power_info), GFP_KERNEL);
+	if (pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = pi;
+
+	ret = rv6xx_parse_power_table(rdev);
+	if (ret)
+		return ret;
+
+	if (rdev->pm.dpm.voltage_response_time == 0)
+		rdev->pm.dpm.voltage_response_time = R600_VOLTAGERESPONSETIME_DFLT;
+	if (rdev->pm.dpm.backbias_response_time == 0)
+		rdev->pm.dpm.backbias_response_time = R600_BACKBIASRESPONSETIME_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->spll_ref_div = dividers.ref_div + 1;
+	else
+		pi->spll_ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->mpll_ref_div = dividers.ref_div + 1;
+	else
+		pi->mpll_ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	if (rdev->family >= CHIP_RV670)
+		pi->fb_div_scale = 1;
+	else
+		pi->fb_div_scale = 0;
+
+	pi->voltage_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, 0);
+
+	pi->gfx_clock_gating = true;
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+                                   &frev, &crev, &data_offset)) {
+		pi->sclk_ss = true;
+		pi->mclk_ss = true;
+		pi->dynamic_ss = true;
+	} else {
+		pi->sclk_ss = false;
+		pi->mclk_ss = false;
+		pi->dynamic_ss = false;
+	}
+
+	pi->dynamic_pcie_gen2 = true;
+
+	if (pi->gfx_clock_gating &&
+	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+		pi->thermal_protection = true;
+	else
+		pi->thermal_protection = false;
+
+	pi->display_gap = true;
+
+	return 0;
+}
+
+void rv6xx_dpm_print_power_state(struct radeon_device *rdev,
+				 struct radeon_ps *rps)
+{
+	struct rv6xx_ps *ps = rv6xx_get_ps(rps);
+	struct rv6xx_pl *pl;
+
+	r600_dpm_print_class_info(rps->class, rps->class2);
+	r600_dpm_print_cap_info(rps->caps);
+	printk("\tuvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+	pl = &ps->low;
+	printk("\t\tpower level 0    sclk: %u mclk: %u vddc: %u\n",
+	       pl->sclk, pl->mclk, pl->vddc);
+	pl = &ps->medium;
+	printk("\t\tpower level 1    sclk: %u mclk: %u vddc: %u\n",
+	       pl->sclk, pl->mclk, pl->vddc);
+	pl = &ps->high;
+	printk("\t\tpower level 2    sclk: %u mclk: %u vddc: %u\n",
+	       pl->sclk, pl->mclk, pl->vddc);
+	r600_dpm_print_ps_status(rdev, rps);
+}
+
+void rv6xx_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						       struct seq_file *m)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct rv6xx_ps *ps = rv6xx_get_ps(rps);
+	struct rv6xx_pl *pl;
+	u32 current_index =
+		(RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_PROFILE_INDEX_MASK) >>
+		CURRENT_PROFILE_INDEX_SHIFT;
+
+	if (current_index > 2) {
+		seq_printf(m, "invalid dpm profile %d\n", current_index);
+	} else {
+		if (current_index == 0)
+			pl = &ps->low;
+		else if (current_index == 1)
+			pl = &ps->medium;
+		else /* current_index == 2 */
+			pl = &ps->high;
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		seq_printf(m, "power level %d    sclk: %u mclk: %u vddc: %u\n",
+			   current_index, pl->sclk, pl->mclk, pl->vddc);
+	}
+}
+
+void rv6xx_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+}
+
+u32 rv6xx_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct rv6xx_ps *requested_state = rv6xx_get_ps(rdev->pm.dpm.requested_ps);
+
+	if (low)
+		return requested_state->low.sclk;
+	else
+		return requested_state->high.sclk;
+}
+
+u32 rv6xx_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct rv6xx_ps *requested_state = rv6xx_get_ps(rdev->pm.dpm.requested_ps);
+
+	if (low)
+		return requested_state->low.mclk;
+	else
+		return requested_state->high.mclk;
+}
diff --git a/drivers/gpu/drm/radeon/rv6xx_dpm.h b/drivers/gpu/drm/radeon/rv6xx_dpm.h
new file mode 100644
index 000000000000..8035d53ebea6
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv6xx_dpm.h
@@ -0,0 +1,95 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#ifndef __RV6XX_DPM_H__
+#define __RV6XX_DPM_H__
+
+#include "r600_dpm.h"
+
+/* Represents a single SCLK step. */
+struct rv6xx_sclk_stepping
+{
+    u32 vco_frequency;
+    u32 post_divider;
+};
+
+struct rv6xx_pm_hw_state {
+	u32 sclks[R600_PM_NUMBER_OF_ACTIVITY_LEVELS];
+	u32 mclks[R600_PM_NUMBER_OF_MCLKS];
+	u16 vddc[R600_PM_NUMBER_OF_VOLTAGE_LEVELS];
+	bool backbias[R600_PM_NUMBER_OF_VOLTAGE_LEVELS];
+	bool pcie_gen2[R600_PM_NUMBER_OF_ACTIVITY_LEVELS];
+	u8 high_sclk_index;
+	u8 medium_sclk_index;
+	u8 low_sclk_index;
+	u8 high_mclk_index;
+	u8 medium_mclk_index;
+	u8 low_mclk_index;
+	u8 high_vddc_index;
+	u8 medium_vddc_index;
+	u8 low_vddc_index;
+	u8 rp[R600_PM_NUMBER_OF_ACTIVITY_LEVELS];
+	u8 lp[R600_PM_NUMBER_OF_ACTIVITY_LEVELS];
+};
+
+struct rv6xx_power_info {
+	/* flags */
+	bool voltage_control;
+	bool sclk_ss;
+	bool mclk_ss;
+	bool dynamic_ss;
+	bool dynamic_pcie_gen2;
+	bool thermal_protection;
+	bool display_gap;
+	bool gfx_clock_gating;
+	/* clk values */
+	u32 fb_div_scale;
+	u32 spll_ref_div;
+	u32 mpll_ref_div;
+	u32 bsu;
+	u32 bsp;
+	/* */
+	u32 active_auto_throttle_sources;
+	/* current power state */
+	u32 restricted_levels;
+	struct rv6xx_pm_hw_state hw;
+};
+
+struct rv6xx_pl {
+	u32 sclk;
+	u32 mclk;
+	u16 vddc;
+	u32 flags;
+};
+
+struct rv6xx_ps {
+	struct rv6xx_pl high;
+	struct rv6xx_pl medium;
+	struct rv6xx_pl low;
+};
+
+#define RV6XX_DEFAULT_VCLK_FREQ  40000 /* 10 khz */
+#define RV6XX_DEFAULT_DCLK_FREQ  30000 /* 10 khz */
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv6xxd.h b/drivers/gpu/drm/radeon/rv6xxd.h
new file mode 100644
index 000000000000..34e86f90b431
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv6xxd.h
@@ -0,0 +1,246 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef RV6XXD_H
+#define RV6XXD_H
+
+/* RV6xx power management */
+#define SPLL_CNTL_MODE                                    0x60c
+#       define SPLL_DIV_SYNC                              (1 << 5)
+
+#define GENERAL_PWRMGT                                    0x618
+#       define GLOBAL_PWRMGT_EN                           (1 << 0)
+#       define STATIC_PM_EN                               (1 << 1)
+#       define MOBILE_SU                                  (1 << 2)
+#       define THERMAL_PROTECTION_DIS                     (1 << 3)
+#       define THERMAL_PROTECTION_TYPE                    (1 << 4)
+#       define ENABLE_GEN2PCIE                            (1 << 5)
+#       define SW_GPIO_INDEX(x)                           ((x) << 6)
+#       define SW_GPIO_INDEX_MASK                         (3 << 6)
+#       define LOW_VOLT_D2_ACPI                           (1 << 8)
+#       define LOW_VOLT_D3_ACPI                           (1 << 9)
+#       define VOLT_PWRMGT_EN                             (1 << 10)
+#       define BACKBIAS_PAD_EN                            (1 << 16)
+#       define BACKBIAS_VALUE                             (1 << 17)
+#       define BACKBIAS_DPM_CNTL                          (1 << 18)
+#       define DYN_SPREAD_SPECTRUM_EN                     (1 << 21)
+
+#define MCLK_PWRMGT_CNTL                                  0x624
+#       define MPLL_PWRMGT_OFF                            (1 << 0)
+#       define YCLK_TURNOFF                               (1 << 1)
+#       define MPLL_TURNOFF                               (1 << 2)
+#       define SU_MCLK_USE_BCLK                           (1 << 3)
+#       define DLL_READY                                  (1 << 4)
+#       define MC_BUSY                                    (1 << 5)
+#       define MC_INT_CNTL                                (1 << 7)
+#       define MRDCKA_SLEEP                               (1 << 8)
+#       define MRDCKB_SLEEP                               (1 << 9)
+#       define MRDCKC_SLEEP                               (1 << 10)
+#       define MRDCKD_SLEEP                               (1 << 11)
+#       define MRDCKE_SLEEP                               (1 << 12)
+#       define MRDCKF_SLEEP                               (1 << 13)
+#       define MRDCKG_SLEEP                               (1 << 14)
+#       define MRDCKH_SLEEP                               (1 << 15)
+#       define MRDCKA_RESET                               (1 << 16)
+#       define MRDCKB_RESET                               (1 << 17)
+#       define MRDCKC_RESET                               (1 << 18)
+#       define MRDCKD_RESET                               (1 << 19)
+#       define MRDCKE_RESET                               (1 << 20)
+#       define MRDCKF_RESET                               (1 << 21)
+#       define MRDCKG_RESET                               (1 << 22)
+#       define MRDCKH_RESET                               (1 << 23)
+#       define DLL_READY_READ                             (1 << 24)
+#       define USE_DISPLAY_GAP                            (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                  (1 << 26)
+#       define USE_DISPLAY_GAP_CTXSW                      (1 << 27)
+#       define MPLL_TURNOFF_D2                            (1 << 28)
+#       define USE_DISPLAY_URGENT_CTXSW                   (1 << 29)
+
+#define MPLL_FREQ_LEVEL_0                                 0x6e8
+#       define LEVEL0_MPLL_POST_DIV(x)                    ((x) << 0)
+#       define LEVEL0_MPLL_POST_DIV_MASK                  (0xff << 0)
+#       define LEVEL0_MPLL_FB_DIV(x)                      ((x) << 8)
+#       define LEVEL0_MPLL_FB_DIV_MASK                    (0xfff << 8)
+#       define LEVEL0_MPLL_REF_DIV(x)                     ((x) << 20)
+#       define LEVEL0_MPLL_REF_DIV_MASK                   (0x3f << 20)
+#       define LEVEL0_MPLL_DIV_EN                         (1 << 28)
+#       define LEVEL0_DLL_BYPASS                          (1 << 29)
+#       define LEVEL0_DLL_RESET                           (1 << 30)
+
+#define VID_RT                                            0x6f8
+#       define VID_CRT(x)                                 ((x) << 0)
+#       define VID_CRT_MASK                               (0x1fff << 0)
+#       define VID_CRTU(x)                                ((x) << 13)
+#       define VID_CRTU_MASK                              (7 << 13)
+#       define SSTU(x)                                    ((x) << 16)
+#       define SSTU_MASK                                  (7 << 16)
+#       define VID_SWT(x)                                 ((x) << 19)
+#       define VID_SWT_MASK                               (0x1f << 19)
+#       define BRT(x)                                     ((x) << 24)
+#       define BRT_MASK                                   (0xff << 24)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                  0x70c
+#       define TARGET_PROFILE_INDEX_MASK                  (3 << 0)
+#       define TARGET_PROFILE_INDEX_SHIFT                 0
+#       define CURRENT_PROFILE_INDEX_MASK                 (3 << 2)
+#       define CURRENT_PROFILE_INDEX_SHIFT                2
+#       define DYN_PWR_ENTER_INDEX(x)                     ((x) << 4)
+#       define DYN_PWR_ENTER_INDEX_MASK                   (3 << 4)
+#       define DYN_PWR_ENTER_INDEX_SHIFT                  4
+#       define CURR_MCLK_INDEX_MASK                       (3 << 6)
+#       define CURR_MCLK_INDEX_SHIFT                      6
+#       define CURR_SCLK_INDEX_MASK                       (0x1f << 8)
+#       define CURR_SCLK_INDEX_SHIFT                      8
+#       define CURR_VID_INDEX_MASK                        (3 << 13)
+#       define CURR_VID_INDEX_SHIFT                       13
+
+#define VID_UPPER_GPIO_CNTL                               0x740
+#       define CTXSW_UPPER_GPIO_VALUES(x)                 ((x) << 0)
+#       define CTXSW_UPPER_GPIO_VALUES_MASK               (7 << 0)
+#       define HIGH_UPPER_GPIO_VALUES(x)                  ((x) << 3)
+#       define HIGH_UPPER_GPIO_VALUES_MASK                (7 << 3)
+#       define MEDIUM_UPPER_GPIO_VALUES(x)                ((x) << 6)
+#       define MEDIUM_UPPER_GPIO_VALUES_MASK              (7 << 6)
+#       define LOW_UPPER_GPIO_VALUES(x)                   ((x) << 9)
+#       define LOW_UPPER_GPIO_VALUES_MASK                 (7 << 9)
+#       define CTXSW_BACKBIAS_VALUE                       (1 << 12)
+#       define HIGH_BACKBIAS_VALUE                        (1 << 13)
+#       define MEDIUM_BACKBIAS_VALUE                      (1 << 14)
+#       define LOW_BACKBIAS_VALUE                         (1 << 15)
+
+#define CG_DISPLAY_GAP_CNTL                               0x7dc
+#       define DISP1_GAP(x)                               ((x) << 0)
+#       define DISP1_GAP_MASK                             (3 << 0)
+#       define DISP2_GAP(x)                               ((x) << 2)
+#       define DISP2_GAP_MASK                             (3 << 2)
+#       define VBI_TIMER_COUNT(x)                         ((x) << 4)
+#       define VBI_TIMER_COUNT_MASK                       (0x3fff << 4)
+#       define VBI_TIMER_UNIT(x)                          ((x) << 20)
+#       define VBI_TIMER_UNIT_MASK                        (7 << 20)
+#       define DISP1_GAP_MCHG(x)                          ((x) << 24)
+#       define DISP1_GAP_MCHG_MASK                        (3 << 24)
+#       define DISP2_GAP_MCHG(x)                          ((x) << 26)
+#       define DISP2_GAP_MCHG_MASK                        (3 << 26)
+
+#define CG_THERMAL_CTRL                                   0x7f0
+#       define DPM_EVENT_SRC(x)                           ((x) << 0)
+#       define DPM_EVENT_SRC_MASK                         (7 << 0)
+#       define THERM_INC_CLK                              (1 << 3)
+#       define TOFFSET(x)                                 ((x) << 4)
+#       define TOFFSET_MASK                               (0xff << 4)
+#       define DIG_THERM_DPM(x)                           ((x) << 12)
+#       define DIG_THERM_DPM_MASK                         (0xff << 12)
+#       define CTF_SEL(x)                                 ((x) << 20)
+#       define CTF_SEL_MASK                               (7 << 20)
+#       define CTF_PAD_POLARITY                           (1 << 23)
+#       define CTF_PAD_EN                                 (1 << 24)
+
+#define CG_SPLL_SPREAD_SPECTRUM_LOW                       0x820
+#       define SSEN                                       (1 << 0)
+#       define CLKS(x)                                    ((x) << 3)
+#       define CLKS_MASK                                  (0xff << 3)
+#       define CLKS_SHIFT                                 3
+#       define CLKV(x)                                    ((x) << 11)
+#       define CLKV_MASK                                  (0x7ff << 11)
+#       define CLKV_SHIFT                                 11
+#define CG_MPLL_SPREAD_SPECTRUM                           0x830
+
+#define CITF_CNTL					0x200c
+#       define BLACKOUT_RD                              (1 << 0)
+#       define BLACKOUT_WR                              (1 << 1)
+
+#define RAMCFG						0x2408
+#define		NOOFBANK_SHIFT					0
+#define		NOOFBANK_MASK					0x00000001
+#define		NOOFRANK_SHIFT					1
+#define		NOOFRANK_MASK					0x00000002
+#define		NOOFROWS_SHIFT					2
+#define		NOOFROWS_MASK					0x0000001C
+#define		NOOFCOLS_SHIFT					5
+#define		NOOFCOLS_MASK					0x00000060
+#define		CHANSIZE_SHIFT					7
+#define		CHANSIZE_MASK					0x00000080
+#define		BURSTLENGTH_SHIFT				8
+#define		BURSTLENGTH_MASK				0x00000100
+#define		CHANSIZE_OVERRIDE				(1 << 10)
+
+#define SQM_RATIO					0x2424
+#       define STATE0(x)                                ((x) << 0)
+#       define STATE0_MASK                              (0xff << 0)
+#       define STATE1(x)                                ((x) << 8)
+#       define STATE1_MASK                              (0xff << 8)
+#       define STATE2(x)                                ((x) << 16)
+#       define STATE2_MASK                              (0xff << 16)
+#       define STATE3(x)                                ((x) << 24)
+#       define STATE3_MASK                              (0xff << 24)
+
+#define ARB_RFSH_CNTL					0x2460
+#       define ENABLE                                   (1 << 0)
+#define ARB_RFSH_RATE					0x2464
+#       define POWERMODE0(x)                            ((x) << 0)
+#       define POWERMODE0_MASK                          (0xff << 0)
+#       define POWERMODE1(x)                            ((x) << 8)
+#       define POWERMODE1_MASK                          (0xff << 8)
+#       define POWERMODE2(x)                            ((x) << 16)
+#       define POWERMODE2_MASK                          (0xff << 16)
+#       define POWERMODE3(x)                            ((x) << 24)
+#       define POWERMODE3_MASK                          (0xff << 24)
+
+#define MC_SEQ_DRAM					0x2608
+#       define CKE_DYN                                  (1 << 12)
+
+#define MC_SEQ_CMD					0x26c4
+
+#define MC_SEQ_RESERVE_S				0x2890
+#define MC_SEQ_RESERVE_M				0x2894
+
+#define LVTMA_DATA_SYNCHRONIZATION                      0x7adc
+#       define LVTMA_PFREQCHG                           (1 << 8)
+#define DCE3_LVTMA_DATA_SYNCHRONIZATION                 0x7f98
+
+/* PCIE indirect regs */
+#define PCIE_P_CNTL                                       0x40
+#       define P_PLL_PWRDN_IN_L1L23                       (1 << 3)
+#       define P_PLL_BUF_PDNB                             (1 << 4)
+#       define P_PLL_PDNB                                 (1 << 9)
+#       define P_ALLOW_PRX_FRONTEND_SHUTOFF               (1 << 12)
+/* PCIE PORT indirect regs */
+#define PCIE_LC_CNTL                                      0xa0
+#       define LC_L0S_INACTIVITY(x)                       ((x) << 8)
+#       define LC_L0S_INACTIVITY_MASK                     (0xf << 8)
+#       define LC_L0S_INACTIVITY_SHIFT                    8
+#       define LC_L1_INACTIVITY(x)                        ((x) << 12)
+#       define LC_L1_INACTIVITY_MASK                      (0xf << 12)
+#       define LC_L1_INACTIVITY_SHIFT                     12
+#       define LC_PMI_TO_L1_DIS                           (1 << 16)
+#       define LC_ASPM_TO_L1_DIS                          (1 << 24)
+#define PCIE_LC_SPEED_CNTL                                0xa4
+#       define LC_GEN2_EN                                 (1 << 0)
+#       define LC_INITIATE_LINK_SPEED_CHANGE              (1 << 7)
+#       define LC_CURRENT_DATA_RATE                       (1 << 11)
+#       define LC_HW_VOLTAGE_IF_CONTROL(x)                ((x) << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_MASK              (3 << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_SHIFT             12
+#       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 23)
+#       define LC_OTHER_SIDE_SUPPORTS_GEN2                (1 << 24)
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv730_dpm.c b/drivers/gpu/drm/radeon/rv730_dpm.c
new file mode 100644
index 000000000000..3f5e1cf138ba
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv730_dpm.c
@@ -0,0 +1,508 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "rv730d.h"
+#include "r600_dpm.h"
+#include "rv770_dpm.h"
+#include "atom.h"
+
+#define MC_CG_ARB_FREQ_F0           0x0a
+#define MC_CG_ARB_FREQ_F1           0x0b
+#define MC_CG_ARB_FREQ_F2           0x0c
+#define MC_CG_ARB_FREQ_F3           0x0d
+
+struct rv7xx_ps *rv770_get_ps(struct radeon_ps *rps);
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev);
+
+int rv730_populate_sclk_value(struct radeon_device *rdev,
+			      u32 engine_clock,
+			      RV770_SMC_SCLK_VALUE *sclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	u32 spll_func_cntl = pi->clk_regs.rv730.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 = pi->clk_regs.rv730.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 = pi->clk_regs.rv730.cg_spll_func_cntl_3;
+	u32 cg_spll_spread_spectrum = pi->clk_regs.rv730.cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2 = pi->clk_regs.rv730.cg_spll_spread_spectrum_2;
+	u64 tmp;
+	u32 reference_clock = rdev->clock.spll.reference_freq;
+	u32 reference_divider, post_divider;
+	u32 fbdiv;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     engine_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	reference_divider = 1 + dividers.ref_div;
+
+	if (dividers.enable_post_div)
+		post_divider = ((dividers.post_div >> 4) & 0xf) +
+			(dividers.post_div & 0xf) + 2;
+	else
+		post_divider = 1;
+
+	tmp = (u64) engine_clock * reference_divider * post_divider * 16384;
+	do_div(tmp, reference_clock);
+	fbdiv = (u32) tmp;
+
+	/* set up registers */
+	if (dividers.enable_post_div)
+		spll_func_cntl |= SPLL_DIVEN;
+	else
+		spll_func_cntl &= ~SPLL_DIVEN;
+	spll_func_cntl &= ~(SPLL_HILEN_MASK | SPLL_LOLEN_MASK | SPLL_REF_DIV_MASK);
+	spll_func_cntl |= SPLL_REF_DIV(dividers.ref_div);
+	spll_func_cntl |= SPLL_HILEN((dividers.post_div >> 4) & 0xf);
+	spll_func_cntl |= SPLL_LOLEN(dividers.post_div & 0xf);
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(2);
+
+	spll_func_cntl_3 &= ~SPLL_FB_DIV_MASK;
+	spll_func_cntl_3 |= SPLL_FB_DIV(fbdiv);
+	spll_func_cntl_3 |= SPLL_DITHEN;
+
+	if (pi->sclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = engine_clock * post_divider;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_ENGINE_SS, vco_freq)) {
+			u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate);
+			u32 clk_v = ss.percentage * fbdiv / (clk_s * 10000);
+
+			cg_spll_spread_spectrum &= ~CLK_S_MASK;
+			cg_spll_spread_spectrum |= CLK_S(clk_s);
+			cg_spll_spread_spectrum |= SSEN;
+
+			cg_spll_spread_spectrum_2 &= ~CLK_V_MASK;
+			cg_spll_spread_spectrum_2 |= CLK_V(clk_v);
+		}
+	}
+
+	sclk->sclk_value = cpu_to_be32(engine_clock);
+	sclk->vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	sclk->vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	sclk->vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+	sclk->vCG_SPLL_SPREAD_SPECTRUM = cpu_to_be32(cg_spll_spread_spectrum);
+	sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cpu_to_be32(cg_spll_spread_spectrum_2);
+
+	return 0;
+}
+
+int rv730_populate_mclk_value(struct radeon_device *rdev,
+			      u32 engine_clock, u32 memory_clock,
+			      LPRV7XX_SMC_MCLK_VALUE mclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 mclk_pwrmgt_cntl = pi->clk_regs.rv730.mclk_pwrmgt_cntl;
+	u32 dll_cntl = pi->clk_regs.rv730.dll_cntl;
+	u32 mpll_func_cntl = pi->clk_regs.rv730.mpll_func_cntl;
+	u32 mpll_func_cntl_2 = pi->clk_regs.rv730.mpll_func_cntl2;
+	u32 mpll_func_cntl_3 = pi->clk_regs.rv730.mpll_func_cntl3;
+	u32 mpll_ss = pi->clk_regs.rv730.mpll_ss;
+	u32 mpll_ss2 = pi->clk_regs.rv730.mpll_ss2;
+	struct atom_clock_dividers dividers;
+	u32 post_divider, reference_divider;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					     memory_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	reference_divider = dividers.ref_div + 1;
+
+	if (dividers.enable_post_div)
+		post_divider = ((dividers.post_div >> 4) & 0xf) +
+			(dividers.post_div & 0xf) + 2;
+	else
+		post_divider = 1;
+
+	/* setup the registers */
+	if (dividers.enable_post_div)
+		mpll_func_cntl |= MPLL_DIVEN;
+	else
+		mpll_func_cntl &= ~MPLL_DIVEN;
+
+	mpll_func_cntl &= ~(MPLL_REF_DIV_MASK | MPLL_HILEN_MASK | MPLL_LOLEN_MASK);
+	mpll_func_cntl |= MPLL_REF_DIV(dividers.ref_div);
+	mpll_func_cntl |= MPLL_HILEN((dividers.post_div >> 4) & 0xf);
+	mpll_func_cntl |= MPLL_LOLEN(dividers.post_div & 0xf);
+
+	mpll_func_cntl_3 &= ~MPLL_FB_DIV_MASK;
+	mpll_func_cntl_3 |= MPLL_FB_DIV(dividers.fb_div);
+	if (dividers.enable_dithen)
+		mpll_func_cntl_3 |= MPLL_DITHEN;
+	else
+		mpll_func_cntl_3 &= ~MPLL_DITHEN;
+
+	if (pi->mclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = memory_clock * post_divider;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_MEMORY_SS, vco_freq)) {
+			u32 reference_clock = rdev->clock.mpll.reference_freq;
+			u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate);
+			u32 clk_v = ss.percentage * dividers.fb_div / (clk_s * 10000);
+
+			mpll_ss &= ~CLK_S_MASK;
+			mpll_ss |= CLK_S(clk_s);
+			mpll_ss |= SSEN;
+
+			mpll_ss2 &= ~CLK_V_MASK;
+			mpll_ss |= CLK_V(clk_v);
+		}
+	}
+
+
+	mclk->mclk730.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	mclk->mclk730.vDLL_CNTL = cpu_to_be32(dll_cntl);
+	mclk->mclk730.mclk_value = cpu_to_be32(memory_clock);
+	mclk->mclk730.vMPLL_FUNC_CNTL = cpu_to_be32(mpll_func_cntl);
+	mclk->mclk730.vMPLL_FUNC_CNTL2 = cpu_to_be32(mpll_func_cntl_2);
+	mclk->mclk730.vMPLL_FUNC_CNTL3 = cpu_to_be32(mpll_func_cntl_3);
+	mclk->mclk730.vMPLL_SS = cpu_to_be32(mpll_ss);
+	mclk->mclk730.vMPLL_SS2 = cpu_to_be32(mpll_ss2);
+
+	return 0;
+}
+
+void rv730_read_clock_registers(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	pi->clk_regs.rv730.cg_spll_func_cntl =
+		RREG32(CG_SPLL_FUNC_CNTL);
+	pi->clk_regs.rv730.cg_spll_func_cntl_2 =
+		RREG32(CG_SPLL_FUNC_CNTL_2);
+	pi->clk_regs.rv730.cg_spll_func_cntl_3 =
+		RREG32(CG_SPLL_FUNC_CNTL_3);
+	pi->clk_regs.rv730.cg_spll_spread_spectrum =
+		RREG32(CG_SPLL_SPREAD_SPECTRUM);
+	pi->clk_regs.rv730.cg_spll_spread_spectrum_2 =
+		RREG32(CG_SPLL_SPREAD_SPECTRUM_2);
+
+	pi->clk_regs.rv730.mclk_pwrmgt_cntl =
+		RREG32(TCI_MCLK_PWRMGT_CNTL);
+	pi->clk_regs.rv730.dll_cntl =
+		RREG32(TCI_DLL_CNTL);
+	pi->clk_regs.rv730.mpll_func_cntl =
+		RREG32(CG_MPLL_FUNC_CNTL);
+	pi->clk_regs.rv730.mpll_func_cntl2 =
+		RREG32(CG_MPLL_FUNC_CNTL_2);
+	pi->clk_regs.rv730.mpll_func_cntl3 =
+		RREG32(CG_MPLL_FUNC_CNTL_3);
+	pi->clk_regs.rv730.mpll_ss =
+		RREG32(CG_TCI_MPLL_SPREAD_SPECTRUM);
+	pi->clk_regs.rv730.mpll_ss2 =
+		RREG32(CG_TCI_MPLL_SPREAD_SPECTRUM_2);
+}
+
+int rv730_populate_smc_acpi_state(struct radeon_device *rdev,
+				  RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 mpll_func_cntl = 0;
+	u32 mpll_func_cntl_2 = 0 ;
+	u32 mpll_func_cntl_3 = 0;
+	u32 mclk_pwrmgt_cntl;
+	u32 dll_cntl;
+	u32 spll_func_cntl;
+	u32 spll_func_cntl_2;
+	u32 spll_func_cntl_3;
+
+	table->ACPIState = table->initialState;
+	table->ACPIState.flags &= ~PPSMC_SWSTATE_FLAG_DC;
+
+	if (pi->acpi_vddc) {
+		rv770_populate_vddc_value(rdev, pi->acpi_vddc,
+					  &table->ACPIState.levels[0].vddc);
+		table->ACPIState.levels[0].gen2PCIE = pi->pcie_gen2 ?
+			pi->acpi_pcie_gen2 : 0;
+		table->ACPIState.levels[0].gen2XSP =
+			pi->acpi_pcie_gen2;
+	} else {
+		rv770_populate_vddc_value(rdev, pi->min_vddc_in_table,
+					  &table->ACPIState.levels[0].vddc);
+		table->ACPIState.levels[0].gen2PCIE = 0;
+	}
+
+	mpll_func_cntl = pi->clk_regs.rv730.mpll_func_cntl;
+	mpll_func_cntl_2 = pi->clk_regs.rv730.mpll_func_cntl2;
+	mpll_func_cntl_3 = pi->clk_regs.rv730.mpll_func_cntl3;
+
+	mpll_func_cntl |= MPLL_RESET | MPLL_BYPASS_EN;
+	mpll_func_cntl &= ~MPLL_SLEEP;
+
+	mpll_func_cntl_2 &= ~MCLK_MUX_SEL_MASK;
+	mpll_func_cntl_2 |= MCLK_MUX_SEL(1);
+
+	mclk_pwrmgt_cntl = (MRDCKA_RESET |
+			    MRDCKB_RESET |
+			    MRDCKC_RESET |
+			    MRDCKD_RESET |
+			    MRDCKE_RESET |
+			    MRDCKF_RESET |
+			    MRDCKG_RESET |
+			    MRDCKH_RESET |
+			    MRDCKA_SLEEP |
+			    MRDCKB_SLEEP |
+			    MRDCKC_SLEEP |
+			    MRDCKD_SLEEP |
+			    MRDCKE_SLEEP |
+			    MRDCKF_SLEEP |
+			    MRDCKG_SLEEP |
+			    MRDCKH_SLEEP);
+
+	dll_cntl = 0xff000000;
+
+	spll_func_cntl = pi->clk_regs.rv730.cg_spll_func_cntl;
+	spll_func_cntl_2 = pi->clk_regs.rv730.cg_spll_func_cntl_2;
+	spll_func_cntl_3 = pi->clk_regs.rv730.cg_spll_func_cntl_3;
+
+	spll_func_cntl |= SPLL_RESET | SPLL_BYPASS_EN;
+	spll_func_cntl &= ~SPLL_SLEEP;
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(4);
+
+	table->ACPIState.levels[0].mclk.mclk730.vMPLL_FUNC_CNTL = cpu_to_be32(mpll_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk730.vMPLL_FUNC_CNTL2 = cpu_to_be32(mpll_func_cntl_2);
+	table->ACPIState.levels[0].mclk.mclk730.vMPLL_FUNC_CNTL3 = cpu_to_be32(mpll_func_cntl_3);
+	table->ACPIState.levels[0].mclk.mclk730.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	table->ACPIState.levels[0].mclk.mclk730.vDLL_CNTL = cpu_to_be32(dll_cntl);
+
+	table->ACPIState.levels[0].mclk.mclk730.mclk_value = 0;
+
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+
+	table->ACPIState.levels[0].sclk.sclk_value = 0;
+
+	rv770_populate_mvdd_value(rdev, 0, &table->ACPIState.levels[0].mvdd);
+
+	table->ACPIState.levels[1] = table->ACPIState.levels[0];
+	table->ACPIState.levels[2] = table->ACPIState.levels[0];
+
+	return 0;
+}
+
+int rv730_populate_smc_initial_state(struct radeon_device *rdev,
+				     struct radeon_ps *radeon_state,
+				     RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_ps *initial_state = rv770_get_ps(radeon_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 a_t;
+
+	table->initialState.levels[0].mclk.mclk730.vMPLL_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv730.mpll_func_cntl);
+	table->initialState.levels[0].mclk.mclk730.vMPLL_FUNC_CNTL2 =
+		cpu_to_be32(pi->clk_regs.rv730.mpll_func_cntl2);
+	table->initialState.levels[0].mclk.mclk730.vMPLL_FUNC_CNTL3 =
+		cpu_to_be32(pi->clk_regs.rv730.mpll_func_cntl3);
+	table->initialState.levels[0].mclk.mclk730.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(pi->clk_regs.rv730.mclk_pwrmgt_cntl);
+	table->initialState.levels[0].mclk.mclk730.vDLL_CNTL =
+		cpu_to_be32(pi->clk_regs.rv730.dll_cntl);
+	table->initialState.levels[0].mclk.mclk730.vMPLL_SS =
+		cpu_to_be32(pi->clk_regs.rv730.mpll_ss);
+	table->initialState.levels[0].mclk.mclk730.vMPLL_SS2 =
+		cpu_to_be32(pi->clk_regs.rv730.mpll_ss2);
+
+	table->initialState.levels[0].mclk.mclk730.mclk_value =
+		cpu_to_be32(initial_state->low.mclk);
+
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv730.cg_spll_func_cntl);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv730.cg_spll_func_cntl_2);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(pi->clk_regs.rv730.cg_spll_func_cntl_3);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM =
+		cpu_to_be32(pi->clk_regs.rv730.cg_spll_spread_spectrum);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM_2 =
+		cpu_to_be32(pi->clk_regs.rv730.cg_spll_spread_spectrum_2);
+
+	table->initialState.levels[0].sclk.sclk_value =
+		cpu_to_be32(initial_state->low.sclk);
+
+	table->initialState.levels[0].arbValue = MC_CG_ARB_FREQ_F0;
+
+	table->initialState.levels[0].seqValue =
+		rv770_get_seq_value(rdev, &initial_state->low);
+
+	rv770_populate_vddc_value(rdev,
+				  initial_state->low.vddc,
+				  &table->initialState.levels[0].vddc);
+	rv770_populate_initial_mvdd_value(rdev,
+					  &table->initialState.levels[0].mvdd);
+
+	a_t = CG_R(0xffff) | CG_L(0);
+
+	table->initialState.levels[0].aT = cpu_to_be32(a_t);
+
+	table->initialState.levels[0].bSP = cpu_to_be32(pi->dsp);
+
+	if (pi->boot_in_gen2)
+		table->initialState.levels[0].gen2PCIE = 1;
+	else
+		table->initialState.levels[0].gen2PCIE = 0;
+	if (initial_state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2)
+		table->initialState.levels[0].gen2XSP = 1;
+	else
+		table->initialState.levels[0].gen2XSP = 0;
+
+	table->initialState.levels[1] = table->initialState.levels[0];
+	table->initialState.levels[2] = table->initialState.levels[0];
+
+	table->initialState.flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	return 0;
+}
+
+void rv730_program_memory_timing_parameters(struct radeon_device *rdev,
+					    struct radeon_ps *radeon_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+	u32 arb_refresh_rate = 0;
+	u32 dram_timing = 0;
+	u32 dram_timing2 = 0;
+	u32 old_dram_timing = 0;
+	u32 old_dram_timing2 = 0;
+
+	arb_refresh_rate = RREG32(MC_ARB_RFSH_RATE) &
+		~(POWERMODE1_MASK | POWERMODE2_MASK | POWERMODE3_MASK);
+	arb_refresh_rate |=
+		(POWERMODE1(rv770_calculate_memory_refresh_rate(rdev, state->low.sclk)) |
+		 POWERMODE2(rv770_calculate_memory_refresh_rate(rdev, state->medium.sclk)) |
+		 POWERMODE3(rv770_calculate_memory_refresh_rate(rdev, state->high.sclk)));
+	WREG32(MC_ARB_RFSH_RATE, arb_refresh_rate);
+
+	/* save the boot dram timings */
+	old_dram_timing = RREG32(MC_ARB_DRAM_TIMING);
+	old_dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+
+	radeon_atom_set_engine_dram_timings(rdev,
+					    state->high.sclk,
+					    state->high.mclk);
+
+	dram_timing = RREG32(MC_ARB_DRAM_TIMING);
+	dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+
+	WREG32(MC_ARB_DRAM_TIMING_3, dram_timing);
+	WREG32(MC_ARB_DRAM_TIMING2_3, dram_timing2);
+
+	radeon_atom_set_engine_dram_timings(rdev,
+					    state->medium.sclk,
+					    state->medium.mclk);
+
+	dram_timing = RREG32(MC_ARB_DRAM_TIMING);
+	dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+
+	WREG32(MC_ARB_DRAM_TIMING_2, dram_timing);
+	WREG32(MC_ARB_DRAM_TIMING2_2, dram_timing2);
+
+	radeon_atom_set_engine_dram_timings(rdev,
+					    state->low.sclk,
+					    state->low.mclk);
+
+	dram_timing = RREG32(MC_ARB_DRAM_TIMING);
+	dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+
+	WREG32(MC_ARB_DRAM_TIMING_1, dram_timing);
+	WREG32(MC_ARB_DRAM_TIMING2_1, dram_timing2);
+
+	/* restore the boot dram timings */
+	WREG32(MC_ARB_DRAM_TIMING, old_dram_timing);
+	WREG32(MC_ARB_DRAM_TIMING2, old_dram_timing2);
+
+}
+
+void rv730_start_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(SCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_OFF);
+
+	WREG32_P(TCI_MCLK_PWRMGT_CNTL, 0, ~MPLL_PWRMGT_OFF);
+
+	WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN);
+}
+
+void rv730_stop_dpm(struct radeon_device *rdev)
+{
+	PPSMC_Result result;
+
+	result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_TwoLevelsDisabled);
+
+	if (result != PPSMC_Result_OK)
+		DRM_ERROR("Could not force DPM to low\n");
+
+	WREG32_P(GENERAL_PWRMGT, 0, ~GLOBAL_PWRMGT_EN);
+
+	WREG32_P(SCLK_PWRMGT_CNTL, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF);
+
+	WREG32_P(TCI_MCLK_PWRMGT_CNTL, MPLL_PWRMGT_OFF, ~MPLL_PWRMGT_OFF);
+}
+
+void rv730_program_dcodt(struct radeon_device *rdev, bool use_dcodt)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 i = use_dcodt ? 0 : 1;
+	u32 mc4_io_pad_cntl;
+
+	mc4_io_pad_cntl = RREG32(MC4_IO_DQ_PAD_CNTL_D0_I0);
+	mc4_io_pad_cntl &= 0xFFFFFF00;
+	mc4_io_pad_cntl |= pi->odt_value_0[i];
+	WREG32(MC4_IO_DQ_PAD_CNTL_D0_I0, mc4_io_pad_cntl);
+	WREG32(MC4_IO_DQ_PAD_CNTL_D0_I1, mc4_io_pad_cntl);
+
+	mc4_io_pad_cntl = RREG32(MC4_IO_QS_PAD_CNTL_D0_I0);
+	mc4_io_pad_cntl &= 0xFFFFFF00;
+	mc4_io_pad_cntl |= pi->odt_value_1[i];
+	WREG32(MC4_IO_QS_PAD_CNTL_D0_I0, mc4_io_pad_cntl);
+	WREG32(MC4_IO_QS_PAD_CNTL_D0_I1, mc4_io_pad_cntl);
+}
+
+void rv730_get_odt_values(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 mc4_io_pad_cntl;
+
+	pi->odt_value_0[0] = (u8)0;
+	pi->odt_value_1[0] = (u8)0x80;
+
+	mc4_io_pad_cntl = RREG32(MC4_IO_DQ_PAD_CNTL_D0_I0);
+	pi->odt_value_0[1] = (u8)(mc4_io_pad_cntl & 0xff);
+
+	mc4_io_pad_cntl = RREG32(MC4_IO_QS_PAD_CNTL_D0_I0);
+	pi->odt_value_1[1] = (u8)(mc4_io_pad_cntl & 0xff);
+}
diff --git a/drivers/gpu/drm/radeon/rv730d.h b/drivers/gpu/drm/radeon/rv730d.h
new file mode 100644
index 000000000000..f0a7954fb1cb
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv730d.h
@@ -0,0 +1,165 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef RV730_H
+#define RV730_H
+
+#define	CG_SPLL_FUNC_CNTL				0x600
+#define		SPLL_RESET				(1 << 0)
+#define		SPLL_SLEEP				(1 << 1)
+#define		SPLL_DIVEN				(1 << 2)
+#define		SPLL_BYPASS_EN				(1 << 3)
+#define		SPLL_REF_DIV(x)				((x) << 4)
+#define		SPLL_REF_DIV_MASK			(0x3f << 4)
+#define		SPLL_HILEN(x)				((x) << 12)
+#define		SPLL_HILEN_MASK				(0xf << 12)
+#define		SPLL_LOLEN(x)				((x) << 16)
+#define		SPLL_LOLEN_MASK				(0xf << 16)
+#define	CG_SPLL_FUNC_CNTL_2				0x604
+#define		SCLK_MUX_SEL(x)				((x) << 0)
+#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_SPLL_FUNC_CNTL_3				0x608
+#define		SPLL_FB_DIV(x)				((x) << 0)
+#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		SPLL_DITHEN				(1 << 28)
+
+#define	CG_MPLL_FUNC_CNTL				0x624
+#define		MPLL_RESET				(1 << 0)
+#define		MPLL_SLEEP				(1 << 1)
+#define		MPLL_DIVEN				(1 << 2)
+#define		MPLL_BYPASS_EN				(1 << 3)
+#define		MPLL_REF_DIV(x)				((x) << 4)
+#define		MPLL_REF_DIV_MASK			(0x3f << 4)
+#define		MPLL_HILEN(x)				((x) << 12)
+#define		MPLL_HILEN_MASK				(0xf << 12)
+#define		MPLL_LOLEN(x)				((x) << 16)
+#define		MPLL_LOLEN_MASK				(0xf << 16)
+#define	CG_MPLL_FUNC_CNTL_2				0x628
+#define		MCLK_MUX_SEL(x)				((x) << 0)
+#define		MCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_MPLL_FUNC_CNTL_3				0x62c
+#define		MPLL_FB_DIV(x)				((x) << 0)
+#define		MPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		MPLL_DITHEN				(1 << 28)
+
+#define	CG_TCI_MPLL_SPREAD_SPECTRUM			0x634
+#define	CG_TCI_MPLL_SPREAD_SPECTRUM_2			0x638
+#define GENERAL_PWRMGT                                  0x63c
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+#       define STATIC_PM_EN                             (1 << 1)
+#       define THERMAL_PROTECTION_DIS                   (1 << 2)
+#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
+#       define ENABLE_GEN2PCIE                          (1 << 4)
+#       define ENABLE_GEN2XSP                           (1 << 5)
+#       define SW_SMIO_INDEX(x)                         ((x) << 6)
+#       define SW_SMIO_INDEX_MASK                       (3 << 6)
+#       define LOW_VOLT_D2_ACPI                         (1 << 8)
+#       define LOW_VOLT_D3_ACPI                         (1 << 9)
+#       define VOLT_PWRMGT_EN                           (1 << 10)
+#       define BACKBIAS_PAD_EN                          (1 << 18)
+#       define BACKBIAS_VALUE                           (1 << 19)
+#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
+#       define AC_DC_SW                                 (1 << 24)
+
+#define SCLK_PWRMGT_CNTL                                  0x644
+#       define SCLK_PWRMGT_OFF                            (1 << 0)
+#       define SCLK_LOW_D1                                (1 << 1)
+#       define FIR_RESET                                  (1 << 4)
+#       define FIR_FORCE_TREND_SEL                        (1 << 5)
+#       define FIR_TREND_MODE                             (1 << 6)
+#       define DYN_GFX_CLK_OFF_EN                         (1 << 7)
+#       define GFX_CLK_FORCE_ON                           (1 << 8)
+#       define GFX_CLK_REQUEST_OFF                        (1 << 9)
+#       define GFX_CLK_FORCE_OFF                          (1 << 10)
+#       define GFX_CLK_OFF_ACPI_D1                        (1 << 11)
+#       define GFX_CLK_OFF_ACPI_D2                        (1 << 12)
+#       define GFX_CLK_OFF_ACPI_D3                        (1 << 13)
+
+#define	TCI_MCLK_PWRMGT_CNTL				0x648
+#       define MPLL_PWRMGT_OFF                          (1 << 5)
+#       define DLL_READY                                (1 << 6)
+#       define MC_INT_CNTL                              (1 << 7)
+#       define MRDCKA_SLEEP                             (1 << 8)
+#       define MRDCKB_SLEEP                             (1 << 9)
+#       define MRDCKC_SLEEP                             (1 << 10)
+#       define MRDCKD_SLEEP                             (1 << 11)
+#       define MRDCKE_SLEEP                             (1 << 12)
+#       define MRDCKF_SLEEP                             (1 << 13)
+#       define MRDCKG_SLEEP                             (1 << 14)
+#       define MRDCKH_SLEEP                             (1 << 15)
+#       define MRDCKA_RESET                             (1 << 16)
+#       define MRDCKB_RESET                             (1 << 17)
+#       define MRDCKC_RESET                             (1 << 18)
+#       define MRDCKD_RESET                             (1 << 19)
+#       define MRDCKE_RESET                             (1 << 20)
+#       define MRDCKF_RESET                             (1 << 21)
+#       define MRDCKG_RESET                             (1 << 22)
+#       define MRDCKH_RESET                             (1 << 23)
+#       define DLL_READY_READ                           (1 << 24)
+#       define USE_DISPLAY_GAP                          (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                (1 << 26)
+#       define MPLL_TURNOFF_D2                          (1 << 28)
+#define	TCI_DLL_CNTL					0x64c
+
+#define	CG_PG_CNTL					0x858
+#       define PWRGATE_ENABLE                           (1 << 0)
+
+#define	CG_AT				                0x6d4
+#define		CG_R(x)					((x) << 0)
+#define		CG_R_MASK				(0xffff << 0)
+#define		CG_L(x)					((x) << 16)
+#define		CG_L_MASK				(0xffff << 16)
+
+#define	CG_SPLL_SPREAD_SPECTRUM				0x790
+#define		SSEN					(1 << 0)
+#define		CLK_S(x)				((x) << 4)
+#define		CLK_S_MASK				(0xfff << 4)
+#define	CG_SPLL_SPREAD_SPECTRUM_2			0x794
+#define		CLK_V(x)				((x) << 0)
+#define		CLK_V_MASK				(0x3ffffff << 0)
+
+#define	MC_ARB_DRAM_TIMING				0x2774
+#define	MC_ARB_DRAM_TIMING2				0x2778
+
+#define	MC_ARB_RFSH_RATE				0x27b0
+#define		POWERMODE0(x)				((x) << 0)
+#define		POWERMODE0_MASK				(0xff << 0)
+#define		POWERMODE1(x)				((x) << 8)
+#define		POWERMODE1_MASK				(0xff << 8)
+#define		POWERMODE2(x)				((x) << 16)
+#define		POWERMODE2_MASK				(0xff << 16)
+#define		POWERMODE3(x)				((x) << 24)
+#define		POWERMODE3_MASK				(0xff << 24)
+
+#define	MC_ARB_DRAM_TIMING_1				0x27f0
+#define	MC_ARB_DRAM_TIMING_2				0x27f4
+#define	MC_ARB_DRAM_TIMING_3				0x27f8
+#define	MC_ARB_DRAM_TIMING2_1				0x27fc
+#define	MC_ARB_DRAM_TIMING2_2				0x2800
+#define	MC_ARB_DRAM_TIMING2_3				0x2804
+
+#define	MC4_IO_DQ_PAD_CNTL_D0_I0			0x2978
+#define	MC4_IO_DQ_PAD_CNTL_D0_I1			0x297c
+#define	MC4_IO_QS_PAD_CNTL_D0_I0			0x2980
+#define	MC4_IO_QS_PAD_CNTL_D0_I1			0x2984
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv740_dpm.c b/drivers/gpu/drm/radeon/rv740_dpm.c
new file mode 100644
index 000000000000..c4c8da501da8
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv740_dpm.c
@@ -0,0 +1,416 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "rv740d.h"
+#include "r600_dpm.h"
+#include "rv770_dpm.h"
+#include "atom.h"
+
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev);
+
+u32 rv740_get_decoded_reference_divider(u32 encoded_ref)
+{
+	u32 ref = 0;
+
+	switch (encoded_ref) {
+        case 0:
+		ref = 1;
+		break;
+        case 16:
+		ref = 2;
+		break;
+        case 17:
+		ref = 3;
+		break;
+        case 18:
+		ref = 2;
+		break;
+        case 19:
+		ref = 3;
+		break;
+        case 20:
+		ref = 4;
+		break;
+        case 21:
+		ref = 5;
+		break;
+        default:
+		DRM_ERROR("Invalid encoded Reference Divider\n");
+		ref = 0;
+		break;
+	}
+
+	return ref;
+}
+
+struct dll_speed_setting {
+	u16 min;
+	u16 max;
+	u32 dll_speed;
+};
+
+static struct dll_speed_setting dll_speed_table[16] =
+{
+	{ 270, 320, 0x0f },
+	{ 240, 270, 0x0e },
+	{ 200, 240, 0x0d },
+	{ 180, 200, 0x0c },
+	{ 160, 180, 0x0b },
+	{ 140, 160, 0x0a },
+	{ 120, 140, 0x09 },
+	{ 110, 120, 0x08 },
+	{  95, 110, 0x07 },
+	{  85,  95, 0x06 },
+	{  78,  85, 0x05 },
+	{  70,  78, 0x04 },
+	{  65,  70, 0x03 },
+	{  60,  65, 0x02 },
+	{  42,  60, 0x01 },
+	{  00,  42, 0x00 }
+};
+
+u32 rv740_get_dll_speed(bool is_gddr5, u32 memory_clock)
+{
+	int i;
+	u32 factor;
+	u16 data_rate;
+
+	if (is_gddr5)
+		factor = 4;
+	else
+		factor = 2;
+
+	data_rate = (u16)(memory_clock * factor / 1000);
+
+	if (data_rate < dll_speed_table[0].max) {
+		for (i = 0; i < 16; i++) {
+			if (data_rate > dll_speed_table[i].min &&
+			    data_rate <= dll_speed_table[i].max)
+				return dll_speed_table[i].dll_speed;
+		}
+	}
+
+	DRM_DEBUG_KMS("Target MCLK greater than largest MCLK in DLL speed table\n");
+
+	return 0x0f;
+}
+
+int rv740_populate_sclk_value(struct radeon_device *rdev, u32 engine_clock,
+			      RV770_SMC_SCLK_VALUE *sclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	u32 spll_func_cntl = pi->clk_regs.rv770.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 = pi->clk_regs.rv770.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 = pi->clk_regs.rv770.cg_spll_func_cntl_3;
+	u32 cg_spll_spread_spectrum = pi->clk_regs.rv770.cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2 = pi->clk_regs.rv770.cg_spll_spread_spectrum_2;
+	u64 tmp;
+	u32 reference_clock = rdev->clock.spll.reference_freq;
+	u32 reference_divider;
+	u32 fbdiv;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     engine_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	reference_divider = 1 + dividers.ref_div;
+
+	tmp = (u64) engine_clock * reference_divider * dividers.post_div * 16384;
+	do_div(tmp, reference_clock);
+	fbdiv = (u32) tmp;
+
+	spll_func_cntl &= ~(SPLL_PDIV_A_MASK | SPLL_REF_DIV_MASK);
+	spll_func_cntl |= SPLL_REF_DIV(dividers.ref_div);
+	spll_func_cntl |= SPLL_PDIV_A(dividers.post_div);
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(2);
+
+	spll_func_cntl_3 &= ~SPLL_FB_DIV_MASK;
+	spll_func_cntl_3 |= SPLL_FB_DIV(fbdiv);
+	spll_func_cntl_3 |= SPLL_DITHEN;
+
+	if (pi->sclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = engine_clock * dividers.post_div;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_ENGINE_SS, vco_freq)) {
+			u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate);
+			u32 clk_v = 4 * ss.percentage * fbdiv / (clk_s * 10000);
+
+			cg_spll_spread_spectrum &= ~CLK_S_MASK;
+			cg_spll_spread_spectrum |= CLK_S(clk_s);
+			cg_spll_spread_spectrum |= SSEN;
+
+			cg_spll_spread_spectrum_2 &= ~CLK_V_MASK;
+			cg_spll_spread_spectrum_2 |= CLK_V(clk_v);
+		}
+	}
+
+	sclk->sclk_value = cpu_to_be32(engine_clock);
+	sclk->vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	sclk->vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	sclk->vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+	sclk->vCG_SPLL_SPREAD_SPECTRUM = cpu_to_be32(cg_spll_spread_spectrum);
+	sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cpu_to_be32(cg_spll_spread_spectrum_2);
+
+	return 0;
+}
+
+int rv740_populate_mclk_value(struct radeon_device *rdev,
+			      u32 engine_clock, u32 memory_clock,
+			      RV7XX_SMC_MCLK_VALUE *mclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 mpll_ad_func_cntl = pi->clk_regs.rv770.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 = pi->clk_regs.rv770.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl = pi->clk_regs.rv770.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 = pi->clk_regs.rv770.mpll_dq_func_cntl_2;
+	u32 mclk_pwrmgt_cntl = pi->clk_regs.rv770.mclk_pwrmgt_cntl;
+	u32 dll_cntl = pi->clk_regs.rv770.dll_cntl;
+	u32 mpll_ss1 = pi->clk_regs.rv770.mpll_ss1;
+	u32 mpll_ss2 = pi->clk_regs.rv770.mpll_ss2;
+	struct atom_clock_dividers dividers;
+	u32 ibias;
+	u32 dll_speed;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					     memory_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	ibias = rv770_map_clkf_to_ibias(rdev, dividers.whole_fb_div);
+
+	mpll_ad_func_cntl &= ~(CLKR_MASK |
+			       YCLK_POST_DIV_MASK |
+			       CLKF_MASK |
+			       CLKFRAC_MASK |
+			       IBIAS_MASK);
+	mpll_ad_func_cntl |= CLKR(dividers.ref_div);
+	mpll_ad_func_cntl |= YCLK_POST_DIV(dividers.post_div);
+	mpll_ad_func_cntl |= CLKF(dividers.whole_fb_div);
+	mpll_ad_func_cntl |= CLKFRAC(dividers.frac_fb_div);
+	mpll_ad_func_cntl |= IBIAS(ibias);
+
+	if (dividers.vco_mode)
+		mpll_ad_func_cntl_2 |= VCO_MODE;
+	else
+		mpll_ad_func_cntl_2 &= ~VCO_MODE;
+
+	if (pi->mem_gddr5) {
+		mpll_dq_func_cntl &= ~(CLKR_MASK |
+				       YCLK_POST_DIV_MASK |
+				       CLKF_MASK |
+				       CLKFRAC_MASK |
+				       IBIAS_MASK);
+		mpll_dq_func_cntl |= CLKR(dividers.ref_div);
+		mpll_dq_func_cntl |= YCLK_POST_DIV(dividers.post_div);
+		mpll_dq_func_cntl |= CLKF(dividers.whole_fb_div);
+		mpll_dq_func_cntl |= CLKFRAC(dividers.frac_fb_div);
+		mpll_dq_func_cntl |= IBIAS(ibias);
+
+		if (dividers.vco_mode)
+			mpll_dq_func_cntl_2 |= VCO_MODE;
+		else
+			mpll_dq_func_cntl_2 &= ~VCO_MODE;
+	}
+
+	if (pi->mclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = memory_clock * dividers.post_div;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_MEMORY_SS, vco_freq)) {
+			u32 reference_clock = rdev->clock.mpll.reference_freq;
+			u32 decoded_ref = rv740_get_decoded_reference_divider(dividers.ref_div);
+			u32 clk_s = reference_clock * 5 / (decoded_ref * ss.rate);
+			u32 clk_v = 0x40000 * ss.percentage *
+				(dividers.whole_fb_div + (dividers.frac_fb_div / 8)) / (clk_s * 10000);
+
+			mpll_ss1 &= ~CLKV_MASK;
+			mpll_ss1 |= CLKV(clk_v);
+
+			mpll_ss2 &= ~CLKS_MASK;
+			mpll_ss2 |= CLKS(clk_s);
+		}
+	}
+
+	dll_speed = rv740_get_dll_speed(pi->mem_gddr5,
+					memory_clock);
+
+	mclk_pwrmgt_cntl &= ~DLL_SPEED_MASK;
+	mclk_pwrmgt_cntl |= DLL_SPEED(dll_speed);
+
+	mclk->mclk770.mclk_value = cpu_to_be32(memory_clock);
+	mclk->mclk770.vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	mclk->mclk770.vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	mclk->mclk770.vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	mclk->mclk770.vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+	mclk->mclk770.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	mclk->mclk770.vDLL_CNTL = cpu_to_be32(dll_cntl);
+	mclk->mclk770.vMPLL_SS = cpu_to_be32(mpll_ss1);
+	mclk->mclk770.vMPLL_SS2 = cpu_to_be32(mpll_ss2);
+
+	return 0;
+}
+
+void rv740_read_clock_registers(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	pi->clk_regs.rv770.cg_spll_func_cntl =
+		RREG32(CG_SPLL_FUNC_CNTL);
+	pi->clk_regs.rv770.cg_spll_func_cntl_2 =
+		RREG32(CG_SPLL_FUNC_CNTL_2);
+	pi->clk_regs.rv770.cg_spll_func_cntl_3 =
+		RREG32(CG_SPLL_FUNC_CNTL_3);
+	pi->clk_regs.rv770.cg_spll_spread_spectrum =
+		RREG32(CG_SPLL_SPREAD_SPECTRUM);
+	pi->clk_regs.rv770.cg_spll_spread_spectrum_2 =
+		RREG32(CG_SPLL_SPREAD_SPECTRUM_2);
+
+	pi->clk_regs.rv770.mpll_ad_func_cntl =
+		RREG32(MPLL_AD_FUNC_CNTL);
+	pi->clk_regs.rv770.mpll_ad_func_cntl_2 =
+		RREG32(MPLL_AD_FUNC_CNTL_2);
+	pi->clk_regs.rv770.mpll_dq_func_cntl =
+		RREG32(MPLL_DQ_FUNC_CNTL);
+	pi->clk_regs.rv770.mpll_dq_func_cntl_2 =
+		RREG32(MPLL_DQ_FUNC_CNTL_2);
+	pi->clk_regs.rv770.mclk_pwrmgt_cntl =
+		RREG32(MCLK_PWRMGT_CNTL);
+	pi->clk_regs.rv770.dll_cntl = RREG32(DLL_CNTL);
+	pi->clk_regs.rv770.mpll_ss1 = RREG32(MPLL_SS1);
+	pi->clk_regs.rv770.mpll_ss2 = RREG32(MPLL_SS2);
+}
+
+int rv740_populate_smc_acpi_state(struct radeon_device *rdev,
+				  RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 mpll_ad_func_cntl = pi->clk_regs.rv770.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 = pi->clk_regs.rv770.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl = pi->clk_regs.rv770.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 = pi->clk_regs.rv770.mpll_dq_func_cntl_2;
+	u32 spll_func_cntl = pi->clk_regs.rv770.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 = pi->clk_regs.rv770.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 = pi->clk_regs.rv770.cg_spll_func_cntl_3;
+	u32 mclk_pwrmgt_cntl = pi->clk_regs.rv770.mclk_pwrmgt_cntl;
+	u32 dll_cntl = pi->clk_regs.rv770.dll_cntl;
+
+	table->ACPIState = table->initialState;
+
+	table->ACPIState.flags &= ~PPSMC_SWSTATE_FLAG_DC;
+
+	if (pi->acpi_vddc) {
+		rv770_populate_vddc_value(rdev, pi->acpi_vddc,
+					  &table->ACPIState.levels[0].vddc);
+		table->ACPIState.levels[0].gen2PCIE =
+			pi->pcie_gen2 ?
+			pi->acpi_pcie_gen2 : 0;
+		table->ACPIState.levels[0].gen2XSP =
+			pi->acpi_pcie_gen2;
+	} else {
+		rv770_populate_vddc_value(rdev, pi->min_vddc_in_table,
+					  &table->ACPIState.levels[0].vddc);
+		table->ACPIState.levels[0].gen2PCIE = 0;
+	}
+
+	mpll_ad_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN;
+
+	mpll_dq_func_cntl_2 |= BYPASS | BIAS_GEN_PDNB | RESET_EN;
+
+	mclk_pwrmgt_cntl |= (MRDCKA0_RESET |
+			     MRDCKA1_RESET |
+			     MRDCKB0_RESET |
+			     MRDCKB1_RESET |
+			     MRDCKC0_RESET |
+			     MRDCKC1_RESET |
+			     MRDCKD0_RESET |
+			     MRDCKD1_RESET);
+
+	dll_cntl |= (MRDCKA0_BYPASS |
+		     MRDCKA1_BYPASS |
+		     MRDCKB0_BYPASS |
+		     MRDCKB1_BYPASS |
+		     MRDCKC0_BYPASS |
+		     MRDCKC1_BYPASS |
+		     MRDCKD0_BYPASS |
+		     MRDCKD1_BYPASS);
+
+	spll_func_cntl |= SPLL_RESET | SPLL_SLEEP | SPLL_BYPASS_EN;
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(4);
+
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+	table->ACPIState.levels[0].mclk.mclk770.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vDLL_CNTL = cpu_to_be32(dll_cntl);
+
+	table->ACPIState.levels[0].mclk.mclk770.mclk_value = 0;
+
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+
+	table->ACPIState.levels[0].sclk.sclk_value = 0;
+
+	table->ACPIState.levels[1] = table->ACPIState.levels[0];
+	table->ACPIState.levels[2] = table->ACPIState.levels[0];
+
+	rv770_populate_mvdd_value(rdev, 0, &table->ACPIState.levels[0].mvdd);
+
+	return 0;
+}
+
+void rv740_enable_mclk_spread_spectrum(struct radeon_device *rdev,
+				       bool enable)
+{
+	if (enable)
+		WREG32_P(MPLL_CNTL_MODE, SS_SSEN, ~SS_SSEN);
+	else
+		WREG32_P(MPLL_CNTL_MODE, 0, ~SS_SSEN);
+}
+
+u8 rv740_get_mclk_frequency_ratio(u32 memory_clock)
+{
+	u8 mc_para_index;
+
+	if ((memory_clock < 10000) || (memory_clock > 47500))
+		mc_para_index = 0x00;
+	else
+		mc_para_index = (u8)((memory_clock - 10000) / 2500);
+
+	return mc_para_index;
+}
diff --git a/drivers/gpu/drm/radeon/rv740d.h b/drivers/gpu/drm/radeon/rv740d.h
new file mode 100644
index 000000000000..fe5ab075dc17
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv740d.h
@@ -0,0 +1,117 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef RV740_H
+#define RV740_H
+
+#define	CG_SPLL_FUNC_CNTL				0x600
+#define		SPLL_RESET				(1 << 0)
+#define		SPLL_SLEEP				(1 << 1)
+#define		SPLL_BYPASS_EN				(1 << 3)
+#define		SPLL_REF_DIV(x)				((x) << 4)
+#define		SPLL_REF_DIV_MASK			(0x3f << 4)
+#define		SPLL_PDIV_A(x)				((x) << 20)
+#define		SPLL_PDIV_A_MASK			(0x7f << 20)
+#define	CG_SPLL_FUNC_CNTL_2				0x604
+#define		SCLK_MUX_SEL(x)				((x) << 0)
+#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_SPLL_FUNC_CNTL_3				0x608
+#define		SPLL_FB_DIV(x)				((x) << 0)
+#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		SPLL_DITHEN				(1 << 28)
+
+#define	MPLL_CNTL_MODE					0x61c
+#define		SS_SSEN					(1 << 24)
+
+#define	MPLL_AD_FUNC_CNTL				0x624
+#define		CLKF(x)					((x) << 0)
+#define		CLKF_MASK				(0x7f << 0)
+#define		CLKR(x)					((x) << 7)
+#define		CLKR_MASK				(0x1f << 7)
+#define		CLKFRAC(x)				((x) << 12)
+#define		CLKFRAC_MASK				(0x1f << 12)
+#define		YCLK_POST_DIV(x)			((x) << 17)
+#define		YCLK_POST_DIV_MASK			(3 << 17)
+#define		IBIAS(x)				((x) << 20)
+#define		IBIAS_MASK				(0x3ff << 20)
+#define		RESET					(1 << 30)
+#define		PDNB					(1 << 31)
+#define	MPLL_AD_FUNC_CNTL_2				0x628
+#define		BYPASS					(1 << 19)
+#define		BIAS_GEN_PDNB				(1 << 24)
+#define		RESET_EN				(1 << 25)
+#define		VCO_MODE				(1 << 29)
+#define	MPLL_DQ_FUNC_CNTL				0x62c
+#define	MPLL_DQ_FUNC_CNTL_2				0x630
+
+#define	MCLK_PWRMGT_CNTL				0x648
+#define		DLL_SPEED(x)				((x) << 0)
+#define		DLL_SPEED_MASK				(0x1f << 0)
+#       define MPLL_PWRMGT_OFF                          (1 << 5)
+#       define DLL_READY                                (1 << 6)
+#       define MC_INT_CNTL                              (1 << 7)
+#       define MRDCKA0_SLEEP                            (1 << 8)
+#       define MRDCKA1_SLEEP                            (1 << 9)
+#       define MRDCKB0_SLEEP                            (1 << 10)
+#       define MRDCKB1_SLEEP                            (1 << 11)
+#       define MRDCKC0_SLEEP                            (1 << 12)
+#       define MRDCKC1_SLEEP                            (1 << 13)
+#       define MRDCKD0_SLEEP                            (1 << 14)
+#       define MRDCKD1_SLEEP                            (1 << 15)
+#       define MRDCKA0_RESET                            (1 << 16)
+#       define MRDCKA1_RESET                            (1 << 17)
+#       define MRDCKB0_RESET                            (1 << 18)
+#       define MRDCKB1_RESET                            (1 << 19)
+#       define MRDCKC0_RESET                            (1 << 20)
+#       define MRDCKC1_RESET                            (1 << 21)
+#       define MRDCKD0_RESET                            (1 << 22)
+#       define MRDCKD1_RESET                            (1 << 23)
+#       define DLL_READY_READ                           (1 << 24)
+#       define USE_DISPLAY_GAP                          (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                (1 << 26)
+#       define MPLL_TURNOFF_D2                          (1 << 28)
+#define	DLL_CNTL					0x64c
+#       define MRDCKA0_BYPASS                           (1 << 24)
+#       define MRDCKA1_BYPASS                           (1 << 25)
+#       define MRDCKB0_BYPASS                           (1 << 26)
+#       define MRDCKB1_BYPASS                           (1 << 27)
+#       define MRDCKC0_BYPASS                           (1 << 28)
+#       define MRDCKC1_BYPASS                           (1 << 29)
+#       define MRDCKD0_BYPASS                           (1 << 30)
+#       define MRDCKD1_BYPASS                           (1 << 31)
+
+#define	CG_SPLL_SPREAD_SPECTRUM				0x790
+#define		SSEN					(1 << 0)
+#define		CLK_S(x)				((x) << 4)
+#define		CLK_S_MASK				(0xfff << 4)
+#define	CG_SPLL_SPREAD_SPECTRUM_2			0x794
+#define		CLK_V(x)				((x) << 0)
+#define		CLK_V_MASK				(0x3ffffff << 0)
+
+#define	MPLL_SS1					0x85c
+#define		CLKV(x)					((x) << 0)
+#define		CLKV_MASK				(0x3ffffff << 0)
+#define	MPLL_SS2					0x860
+#define		CLKS(x)					((x) << 0)
+#define		CLKS_MASK				(0xfff << 0)
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c
index 4a62ad2e5399..30ea14e8854c 100644
--- a/drivers/gpu/drm/radeon/rv770.c
+++ b/drivers/gpu/drm/radeon/rv770.c
@@ -813,7 +813,7 @@ int rv770_uvd_resume(struct radeon_device *rdev)
 
 	/* programm the VCPU memory controller bits 0-27 */
 	addr = rdev->uvd.gpu_addr >> 3;
-	size = RADEON_GPU_PAGE_ALIGN(rdev->uvd_fw->size + 4) >> 3;
+	size = RADEON_GPU_PAGE_ALIGN(rdev->uvd.fw_size + 4) >> 3;
 	WREG32(UVD_VCPU_CACHE_OFFSET0, addr);
 	WREG32(UVD_VCPU_CACHE_SIZE0, size);
 
diff --git a/drivers/gpu/drm/radeon/rv770_dpm.c b/drivers/gpu/drm/radeon/rv770_dpm.c
new file mode 100644
index 000000000000..2d347925f77d
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv770_dpm.c
@@ -0,0 +1,2523 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "rv770d.h"
+#include "r600_dpm.h"
+#include "rv770_dpm.h"
+#include "cypress_dpm.h"
+#include "atom.h"
+#include <linux/seq_file.h>
+
+#define MC_CG_ARB_FREQ_F0           0x0a
+#define MC_CG_ARB_FREQ_F1           0x0b
+#define MC_CG_ARB_FREQ_F2           0x0c
+#define MC_CG_ARB_FREQ_F3           0x0d
+
+#define MC_CG_SEQ_DRAMCONF_S0       0x05
+#define MC_CG_SEQ_DRAMCONF_S1       0x06
+
+#define PCIE_BUS_CLK                10000
+#define TCLK                        (PCIE_BUS_CLK / 10)
+
+#define SMC_RAM_END 0xC000
+
+struct rv7xx_ps *rv770_get_ps(struct radeon_ps *rps)
+{
+	struct rv7xx_ps *ps = rps->ps_priv;
+
+	return ps;
+}
+
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rdev->pm.dpm.priv;
+
+	return pi;
+}
+
+struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *pi = rdev->pm.dpm.priv;
+
+	return pi;
+}
+
+static void rv770_enable_bif_dynamic_pcie_gen2(struct radeon_device *rdev,
+					       bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	if (enable) {
+		tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+		tmp |= LC_HW_VOLTAGE_IF_CONTROL(1);
+		tmp |= LC_GEN2_EN_STRAP;
+	} else {
+		if (!pi->boot_in_gen2) {
+			tmp &= ~LC_HW_VOLTAGE_IF_CONTROL_MASK;
+			tmp &= ~LC_GEN2_EN_STRAP;
+		}
+	}
+	if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) ||
+	    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2))
+		WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, tmp);
+
+}
+
+static void rv770_enable_l0s(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL) & ~LC_L0S_INACTIVITY_MASK;
+	tmp |= LC_L0S_INACTIVITY(3);
+	WREG32_PCIE_PORT(PCIE_LC_CNTL, tmp);
+}
+
+static void rv770_enable_l1(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL);
+	tmp &= ~LC_L1_INACTIVITY_MASK;
+	tmp |= LC_L1_INACTIVITY(4);
+	tmp &= ~LC_PMI_TO_L1_DIS;
+	tmp &= ~LC_ASPM_TO_L1_DIS;
+	WREG32_PCIE_PORT(PCIE_LC_CNTL, tmp);
+}
+
+static void rv770_enable_pll_sleep_in_l1(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL) & ~LC_L1_INACTIVITY_MASK;
+	tmp |= LC_L1_INACTIVITY(8);
+	WREG32_PCIE_PORT(PCIE_LC_CNTL, tmp);
+
+	/* NOTE, this is a PCIE indirect reg, not PCIE PORT */
+	tmp = RREG32_PCIE(PCIE_P_CNTL);
+	tmp |= P_PLL_PWRDN_IN_L1L23;
+	tmp &= ~P_PLL_BUF_PDNB;
+	tmp &= ~P_PLL_PDNB;
+	tmp |= P_ALLOW_PRX_FRONTEND_SHUTOFF;
+	WREG32_PCIE(PCIE_P_CNTL, tmp);
+}
+
+static void rv770_gfx_clock_gating_enable(struct radeon_device *rdev,
+					  bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+	else {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+		RREG32(GB_TILING_CONFIG);
+	}
+}
+
+static void rv770_mg_clock_gating_enable(struct radeon_device *rdev,
+					 bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (enable) {
+		u32 mgcg_cgtt_local0;
+
+		if (rdev->family == CHIP_RV770)
+			mgcg_cgtt_local0 = RV770_MGCGTTLOCAL0_DFLT;
+		else
+			mgcg_cgtt_local0 = RV7XX_MGCGTTLOCAL0_DFLT;
+
+		WREG32(CG_CGTT_LOCAL_0, mgcg_cgtt_local0);
+		WREG32(CG_CGTT_LOCAL_1, (RV770_MGCGTTLOCAL1_DFLT & 0xFFFFCFFF));
+
+		if (pi->mgcgtssm)
+			WREG32(CGTS_SM_CTRL_REG, RV770_MGCGCGTSSMCTRL_DFLT);
+	} else {
+		WREG32(CG_CGTT_LOCAL_0, 0xFFFFFFFF);
+		WREG32(CG_CGTT_LOCAL_1, 0xFFFFCFFF);
+	}
+}
+
+void rv770_restore_cgcg(struct radeon_device *rdev)
+{
+	bool dpm_en = false, cg_en = false;
+
+	if (RREG32(GENERAL_PWRMGT) & GLOBAL_PWRMGT_EN)
+		dpm_en = true;
+	if (RREG32(SCLK_PWRMGT_CNTL) & DYN_GFX_CLK_OFF_EN)
+		cg_en = true;
+
+	if (dpm_en && !cg_en)
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+}
+
+static void rv770_start_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(SCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_OFF);
+
+	WREG32_P(MCLK_PWRMGT_CNTL, 0, ~MPLL_PWRMGT_OFF);
+
+	WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN);
+}
+
+void rv770_stop_dpm(struct radeon_device *rdev)
+{
+	PPSMC_Result result;
+
+	result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_TwoLevelsDisabled);
+
+	if (result != PPSMC_Result_OK)
+		DRM_ERROR("Could not force DPM to low.\n");
+
+	WREG32_P(GENERAL_PWRMGT, 0, ~GLOBAL_PWRMGT_EN);
+
+	WREG32_P(SCLK_PWRMGT_CNTL, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF);
+
+	WREG32_P(MCLK_PWRMGT_CNTL, MPLL_PWRMGT_OFF, ~MPLL_PWRMGT_OFF);
+}
+
+bool rv770_dpm_enabled(struct radeon_device *rdev)
+{
+	if (RREG32(GENERAL_PWRMGT) & GLOBAL_PWRMGT_EN)
+		return true;
+	else
+		return false;
+}
+
+void rv770_enable_thermal_protection(struct radeon_device *rdev,
+				     bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS);
+	else
+		WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS);
+}
+
+void rv770_enable_acpi_pm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, STATIC_PM_EN, ~STATIC_PM_EN);
+}
+
+u8 rv770_get_seq_value(struct radeon_device *rdev,
+		       struct rv7xx_pl *pl)
+{
+	return (pl->flags & ATOM_PPLIB_R600_FLAGS_LOWPOWER) ?
+		MC_CG_SEQ_DRAMCONF_S0 : MC_CG_SEQ_DRAMCONF_S1;
+}
+
+int rv770_read_smc_soft_register(struct radeon_device *rdev,
+				 u16 reg_offset, u32 *value)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	return rv770_read_smc_sram_dword(rdev,
+					 pi->soft_regs_start + reg_offset,
+					 value, pi->sram_end);
+}
+
+int rv770_write_smc_soft_register(struct radeon_device *rdev,
+				  u16 reg_offset, u32 value)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	return rv770_write_smc_sram_dword(rdev,
+					  pi->soft_regs_start + reg_offset,
+					  value, pi->sram_end);
+}
+
+int rv770_populate_smc_t(struct radeon_device *rdev,
+			 struct radeon_ps *radeon_state,
+			 RV770_SMC_SWSTATE *smc_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+	int a_n;
+	int a_d;
+	u8 l[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE];
+	u8 r[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE];
+	u32 a_t;
+
+	l[0] = 0;
+	r[2] = 100;
+
+	a_n = (int)state->medium.sclk * pi->lmp +
+		(int)state->low.sclk * (R600_AH_DFLT - pi->rlp);
+	a_d = (int)state->low.sclk * (100 - (int)pi->rlp) +
+		(int)state->medium.sclk * pi->lmp;
+
+	l[1] = (u8)(pi->lmp - (int)pi->lmp * a_n / a_d);
+	r[0] = (u8)(pi->rlp + (100 - (int)pi->rlp) * a_n / a_d);
+
+	a_n = (int)state->high.sclk * pi->lhp + (int)state->medium.sclk *
+		(R600_AH_DFLT - pi->rmp);
+	a_d = (int)state->medium.sclk * (100 - (int)pi->rmp) +
+		(int)state->high.sclk * pi->lhp;
+
+	l[2] = (u8)(pi->lhp - (int)pi->lhp * a_n / a_d);
+	r[1] = (u8)(pi->rmp + (100 - (int)pi->rmp) * a_n / a_d);
+
+	for (i = 0; i < (RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1); i++) {
+		a_t = CG_R(r[i] * pi->bsp / 200) | CG_L(l[i] * pi->bsp / 200);
+		smc_state->levels[i].aT = cpu_to_be32(a_t);
+	}
+
+	a_t = CG_R(r[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1] * pi->pbsp / 200) |
+		CG_L(l[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1] * pi->pbsp / 200);
+
+	smc_state->levels[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1].aT =
+		cpu_to_be32(a_t);
+
+	return 0;
+}
+
+int rv770_populate_smc_sp(struct radeon_device *rdev,
+			  struct radeon_ps *radeon_state,
+			  RV770_SMC_SWSTATE *smc_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+
+	for (i = 0; i < (RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1); i++)
+		smc_state->levels[i].bSP = cpu_to_be32(pi->dsp);
+
+	smc_state->levels[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1].bSP =
+		cpu_to_be32(pi->psp);
+
+	return 0;
+}
+
+static void rv770_calculate_fractional_mpll_feedback_divider(u32 memory_clock,
+							     u32 reference_clock,
+							     bool gddr5,
+							     struct atom_clock_dividers *dividers,
+							     u32 *clkf,
+							     u32 *clkfrac)
+{
+	u32 post_divider, reference_divider, feedback_divider8;
+	u32 fyclk;
+
+	if (gddr5)
+		fyclk = (memory_clock * 8) / 2;
+	else
+		fyclk = (memory_clock * 4) / 2;
+
+	post_divider = dividers->post_div;
+	reference_divider = dividers->ref_div;
+
+	feedback_divider8 =
+		(8 * fyclk * reference_divider * post_divider) / reference_clock;
+
+	*clkf = feedback_divider8 / 8;
+	*clkfrac = feedback_divider8 % 8;
+}
+
+static int rv770_encode_yclk_post_div(u32 postdiv, u32 *encoded_postdiv)
+{
+	int ret = 0;
+
+	switch (postdiv) {
+        case 1:
+		*encoded_postdiv = 0;
+		break;
+        case 2:
+		*encoded_postdiv = 1;
+		break;
+        case 4:
+		*encoded_postdiv = 2;
+		break;
+        case 8:
+		*encoded_postdiv = 3;
+		break;
+        case 16:
+		*encoded_postdiv = 4;
+		break;
+        default:
+		ret = -EINVAL;
+		break;
+	}
+
+    return ret;
+}
+
+u32 rv770_map_clkf_to_ibias(struct radeon_device *rdev, u32 clkf)
+{
+	if (clkf <= 0x10)
+		return 0x4B;
+	if (clkf <= 0x19)
+		return 0x5B;
+	if (clkf <= 0x21)
+		return 0x2B;
+	if (clkf <= 0x27)
+		return 0x6C;
+	if (clkf <= 0x31)
+		return 0x9D;
+	return 0xC6;
+}
+
+static int rv770_populate_mclk_value(struct radeon_device *rdev,
+				     u32 engine_clock, u32 memory_clock,
+				     RV7XX_SMC_MCLK_VALUE *mclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 encoded_reference_dividers[] = { 0, 16, 17, 20, 21 };
+	u32 mpll_ad_func_cntl =
+		pi->clk_regs.rv770.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl =
+		pi->clk_regs.rv770.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_dq_func_cntl_2;
+	u32 mclk_pwrmgt_cntl =
+		pi->clk_regs.rv770.mclk_pwrmgt_cntl;
+	u32 dll_cntl = pi->clk_regs.rv770.dll_cntl;
+	struct atom_clock_dividers dividers;
+	u32 reference_clock = rdev->clock.mpll.reference_freq;
+	u32 clkf, clkfrac;
+	u32 postdiv_yclk;
+	u32 ibias;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_MEMORY_PLL_PARAM,
+					     memory_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	if ((dividers.ref_div < 1) || (dividers.ref_div > 5))
+		return -EINVAL;
+
+	rv770_calculate_fractional_mpll_feedback_divider(memory_clock, reference_clock,
+							 pi->mem_gddr5,
+							 &dividers, &clkf, &clkfrac);
+
+	ret = rv770_encode_yclk_post_div(dividers.post_div, &postdiv_yclk);
+	if (ret)
+		return ret;
+
+	ibias = rv770_map_clkf_to_ibias(rdev, clkf);
+
+	mpll_ad_func_cntl &= ~(CLKR_MASK |
+			       YCLK_POST_DIV_MASK |
+			       CLKF_MASK |
+			       CLKFRAC_MASK |
+			       IBIAS_MASK);
+	mpll_ad_func_cntl |= CLKR(encoded_reference_dividers[dividers.ref_div - 1]);
+	mpll_ad_func_cntl |= YCLK_POST_DIV(postdiv_yclk);
+	mpll_ad_func_cntl |= CLKF(clkf);
+	mpll_ad_func_cntl |= CLKFRAC(clkfrac);
+	mpll_ad_func_cntl |= IBIAS(ibias);
+
+	if (dividers.vco_mode)
+		mpll_ad_func_cntl_2 |= VCO_MODE;
+	else
+		mpll_ad_func_cntl_2 &= ~VCO_MODE;
+
+	if (pi->mem_gddr5) {
+		rv770_calculate_fractional_mpll_feedback_divider(memory_clock,
+								 reference_clock,
+								 pi->mem_gddr5,
+								 &dividers, &clkf, &clkfrac);
+
+		ibias = rv770_map_clkf_to_ibias(rdev, clkf);
+
+		ret = rv770_encode_yclk_post_div(dividers.post_div, &postdiv_yclk);
+		if (ret)
+			return ret;
+
+		mpll_dq_func_cntl &= ~(CLKR_MASK |
+				       YCLK_POST_DIV_MASK |
+				       CLKF_MASK |
+				       CLKFRAC_MASK |
+				       IBIAS_MASK);
+		mpll_dq_func_cntl |= CLKR(encoded_reference_dividers[dividers.ref_div - 1]);
+		mpll_dq_func_cntl |= YCLK_POST_DIV(postdiv_yclk);
+		mpll_dq_func_cntl |= CLKF(clkf);
+		mpll_dq_func_cntl |= CLKFRAC(clkfrac);
+		mpll_dq_func_cntl |= IBIAS(ibias);
+
+		if (dividers.vco_mode)
+			mpll_dq_func_cntl_2 |= VCO_MODE;
+		else
+			mpll_dq_func_cntl_2 &= ~VCO_MODE;
+	}
+
+	mclk->mclk770.mclk_value = cpu_to_be32(memory_clock);
+	mclk->mclk770.vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	mclk->mclk770.vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	mclk->mclk770.vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	mclk->mclk770.vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+	mclk->mclk770.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	mclk->mclk770.vDLL_CNTL = cpu_to_be32(dll_cntl);
+
+	return 0;
+}
+
+static int rv770_populate_sclk_value(struct radeon_device *rdev,
+				     u32 engine_clock,
+				     RV770_SMC_SCLK_VALUE *sclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	u32 spll_func_cntl =
+		pi->clk_regs.rv770.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 =
+		pi->clk_regs.rv770.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 =
+		pi->clk_regs.rv770.cg_spll_func_cntl_3;
+	u32 cg_spll_spread_spectrum =
+		pi->clk_regs.rv770.cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2 =
+		pi->clk_regs.rv770.cg_spll_spread_spectrum_2;
+	u64 tmp;
+	u32 reference_clock = rdev->clock.spll.reference_freq;
+	u32 reference_divider, post_divider;
+	u32 fbdiv;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     engine_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	reference_divider = 1 + dividers.ref_div;
+
+	if (dividers.enable_post_div)
+		post_divider = (0x0f & (dividers.post_div >> 4)) + (0x0f & dividers.post_div) + 2;
+	else
+		post_divider = 1;
+
+	tmp = (u64) engine_clock * reference_divider * post_divider * 16384;
+	do_div(tmp, reference_clock);
+	fbdiv = (u32) tmp;
+
+	if (dividers.enable_post_div)
+		spll_func_cntl |= SPLL_DIVEN;
+	else
+		spll_func_cntl &= ~SPLL_DIVEN;
+	spll_func_cntl &= ~(SPLL_HILEN_MASK | SPLL_LOLEN_MASK | SPLL_REF_DIV_MASK);
+	spll_func_cntl |= SPLL_REF_DIV(dividers.ref_div);
+	spll_func_cntl |= SPLL_HILEN((dividers.post_div >> 4) & 0xf);
+	spll_func_cntl |= SPLL_LOLEN(dividers.post_div & 0xf);
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(2);
+
+	spll_func_cntl_3 &= ~SPLL_FB_DIV_MASK;
+	spll_func_cntl_3 |= SPLL_FB_DIV(fbdiv);
+	spll_func_cntl_3 |= SPLL_DITHEN;
+
+	if (pi->sclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = engine_clock * post_divider;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_ENGINE_SS, vco_freq)) {
+			u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate);
+			u32 clk_v = ss.percentage * fbdiv / (clk_s * 10000);
+
+			cg_spll_spread_spectrum &= ~CLKS_MASK;
+			cg_spll_spread_spectrum |= CLKS(clk_s);
+			cg_spll_spread_spectrum |= SSEN;
+
+			cg_spll_spread_spectrum_2 &= ~CLKV_MASK;
+			cg_spll_spread_spectrum_2 |= CLKV(clk_v);
+		}
+	}
+
+	sclk->sclk_value = cpu_to_be32(engine_clock);
+	sclk->vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	sclk->vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	sclk->vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+	sclk->vCG_SPLL_SPREAD_SPECTRUM = cpu_to_be32(cg_spll_spread_spectrum);
+	sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cpu_to_be32(cg_spll_spread_spectrum_2);
+
+	return 0;
+}
+
+int rv770_populate_vddc_value(struct radeon_device *rdev, u16 vddc,
+			      RV770_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+
+	if (!pi->voltage_control) {
+		voltage->index = 0;
+		voltage->value = 0;
+		return 0;
+	}
+
+	for (i = 0; i < pi->valid_vddc_entries; i++) {
+		if (vddc <= pi->vddc_table[i].vddc) {
+			voltage->index = pi->vddc_table[i].vddc_index;
+			voltage->value = cpu_to_be16(vddc);
+			break;
+		}
+	}
+
+	if (i == pi->valid_vddc_entries)
+		return -EINVAL;
+
+	return 0;
+}
+
+int rv770_populate_mvdd_value(struct radeon_device *rdev, u32 mclk,
+			      RV770_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (!pi->mvdd_control) {
+		voltage->index = MVDD_HIGH_INDEX;
+		voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+		return 0;
+	}
+
+	if (mclk <= pi->mvdd_split_frequency) {
+		voltage->index = MVDD_LOW_INDEX;
+		voltage->value = cpu_to_be16(MVDD_LOW_VALUE);
+	} else {
+		voltage->index = MVDD_HIGH_INDEX;
+		voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+	}
+
+	return 0;
+}
+
+static int rv770_convert_power_level_to_smc(struct radeon_device *rdev,
+					    struct rv7xx_pl *pl,
+					    RV770_SMC_HW_PERFORMANCE_LEVEL *level,
+					    u8 watermark_level)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int ret;
+
+	level->gen2PCIE = pi->pcie_gen2 ?
+		((pl->flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? 1 : 0) : 0;
+	level->gen2XSP  = (pl->flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2) ? 1 : 0;
+	level->backbias = (pl->flags & ATOM_PPLIB_R600_FLAGS_BACKBIASENABLE) ? 1 : 0;
+	level->displayWatermark = watermark_level;
+
+	if (rdev->family == CHIP_RV740)
+		ret = rv740_populate_sclk_value(rdev, pl->sclk,
+						&level->sclk);
+	else if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		ret = rv730_populate_sclk_value(rdev, pl->sclk,
+						&level->sclk);
+	else
+		ret = rv770_populate_sclk_value(rdev, pl->sclk,
+						&level->sclk);
+	if (ret)
+		return ret;
+
+	if (rdev->family == CHIP_RV740) {
+		if (pi->mem_gddr5) {
+			if (pl->mclk <= pi->mclk_strobe_mode_threshold)
+				level->strobeMode =
+					rv740_get_mclk_frequency_ratio(pl->mclk) | 0x10;
+			else
+				level->strobeMode = 0;
+
+			if (pl->mclk > pi->mclk_edc_enable_threshold)
+				level->mcFlags = SMC_MC_EDC_RD_FLAG | SMC_MC_EDC_WR_FLAG;
+			else
+				level->mcFlags =  0;
+		}
+		ret = rv740_populate_mclk_value(rdev, pl->sclk,
+						pl->mclk, &level->mclk);
+	} else if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		ret = rv730_populate_mclk_value(rdev, pl->sclk,
+						pl->mclk, &level->mclk);
+	else
+		ret = rv770_populate_mclk_value(rdev, pl->sclk,
+						pl->mclk, &level->mclk);
+	if (ret)
+		return ret;
+
+	ret = rv770_populate_vddc_value(rdev, pl->vddc,
+					&level->vddc);
+	if (ret)
+		return ret;
+
+	ret = rv770_populate_mvdd_value(rdev, pl->mclk, &level->mvdd);
+
+	return ret;
+}
+
+static int rv770_convert_power_state_to_smc(struct radeon_device *rdev,
+					    struct radeon_ps *radeon_state,
+					    RV770_SMC_SWSTATE *smc_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+	int ret;
+
+	if (!(radeon_state->caps & ATOM_PPLIB_DISALLOW_ON_DC))
+		smc_state->flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	ret = rv770_convert_power_level_to_smc(rdev,
+					       &state->low,
+					       &smc_state->levels[0],
+					       PPSMC_DISPLAY_WATERMARK_LOW);
+	if (ret)
+		return ret;
+
+	ret = rv770_convert_power_level_to_smc(rdev,
+					       &state->medium,
+					       &smc_state->levels[1],
+					       PPSMC_DISPLAY_WATERMARK_LOW);
+	if (ret)
+		return ret;
+
+	ret = rv770_convert_power_level_to_smc(rdev,
+					       &state->high,
+					       &smc_state->levels[2],
+					       PPSMC_DISPLAY_WATERMARK_HIGH);
+	if (ret)
+		return ret;
+
+	smc_state->levels[0].arbValue = MC_CG_ARB_FREQ_F1;
+	smc_state->levels[1].arbValue = MC_CG_ARB_FREQ_F2;
+	smc_state->levels[2].arbValue = MC_CG_ARB_FREQ_F3;
+
+	smc_state->levels[0].seqValue = rv770_get_seq_value(rdev,
+							    &state->low);
+	smc_state->levels[1].seqValue = rv770_get_seq_value(rdev,
+							    &state->medium);
+	smc_state->levels[2].seqValue = rv770_get_seq_value(rdev,
+							    &state->high);
+
+	rv770_populate_smc_sp(rdev, radeon_state, smc_state);
+
+	return rv770_populate_smc_t(rdev, radeon_state, smc_state);
+
+}
+
+u32 rv770_calculate_memory_refresh_rate(struct radeon_device *rdev,
+					u32 engine_clock)
+{
+	u32 dram_rows;
+	u32 dram_refresh_rate;
+	u32 mc_arb_rfsh_rate;
+	u32 tmp;
+
+	tmp = (RREG32(MC_ARB_RAMCFG) & NOOFROWS_MASK) >> NOOFROWS_SHIFT;
+	dram_rows = 1 << (tmp + 10);
+	tmp = RREG32(MC_SEQ_MISC0) & 3;
+	dram_refresh_rate = 1 << (tmp + 3);
+	mc_arb_rfsh_rate = ((engine_clock * 10) * dram_refresh_rate / dram_rows - 32) / 64;
+
+	return mc_arb_rfsh_rate;
+}
+
+static void rv770_program_memory_timing_parameters(struct radeon_device *rdev,
+						   struct radeon_ps *radeon_state)
+{
+	struct rv7xx_ps *state = rv770_get_ps(radeon_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 sqm_ratio;
+	u32 arb_refresh_rate;
+	u32 high_clock;
+
+	if (state->high.sclk < (state->low.sclk * 0xFF / 0x40))
+		high_clock = state->high.sclk;
+	else
+		high_clock = (state->low.sclk * 0xFF / 0x40);
+
+	radeon_atom_set_engine_dram_timings(rdev, high_clock,
+					    state->high.mclk);
+
+	sqm_ratio =
+		STATE0(64 * high_clock / pi->boot_sclk) |
+		STATE1(64 * high_clock / state->low.sclk) |
+		STATE2(64 * high_clock / state->medium.sclk) |
+		STATE3(64 * high_clock / state->high.sclk);
+	WREG32(MC_ARB_SQM_RATIO, sqm_ratio);
+
+	arb_refresh_rate =
+		POWERMODE0(rv770_calculate_memory_refresh_rate(rdev, pi->boot_sclk)) |
+		POWERMODE1(rv770_calculate_memory_refresh_rate(rdev, state->low.sclk)) |
+		POWERMODE2(rv770_calculate_memory_refresh_rate(rdev, state->medium.sclk)) |
+		POWERMODE3(rv770_calculate_memory_refresh_rate(rdev, state->high.sclk));
+	WREG32(MC_ARB_RFSH_RATE, arb_refresh_rate);
+}
+
+void rv770_enable_backbias(struct radeon_device *rdev,
+			   bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, BACKBIAS_PAD_EN, ~BACKBIAS_PAD_EN);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~(BACKBIAS_VALUE | BACKBIAS_PAD_EN));
+}
+
+static void rv770_enable_spread_spectrum(struct radeon_device *rdev,
+					 bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (enable) {
+		if (pi->sclk_ss)
+			WREG32_P(GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, ~DYN_SPREAD_SPECTRUM_EN);
+
+		if (pi->mclk_ss) {
+			if (rdev->family == CHIP_RV740)
+				rv740_enable_mclk_spread_spectrum(rdev, true);
+		}
+	} else {
+		WREG32_P(CG_SPLL_SPREAD_SPECTRUM, 0, ~SSEN);
+
+		WREG32_P(GENERAL_PWRMGT, 0, ~DYN_SPREAD_SPECTRUM_EN);
+
+		WREG32_P(CG_MPLL_SPREAD_SPECTRUM, 0, ~SSEN);
+
+		if (rdev->family == CHIP_RV740)
+			rv740_enable_mclk_spread_spectrum(rdev, false);
+	}
+}
+
+static void rv770_program_mpll_timing_parameters(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if ((rdev->family == CHIP_RV770) && !pi->mem_gddr5) {
+		WREG32(MPLL_TIME,
+		       (MPLL_LOCK_TIME(R600_MPLLLOCKTIME_DFLT * pi->ref_div) |
+			MPLL_RESET_TIME(R600_MPLLRESETTIME_DFLT)));
+	}
+}
+
+void rv770_setup_bsp(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 xclk = radeon_get_xclk(rdev);
+
+	r600_calculate_u_and_p(pi->asi,
+			       xclk,
+			       16,
+			       &pi->bsp,
+			       &pi->bsu);
+
+	r600_calculate_u_and_p(pi->pasi,
+			       xclk,
+			       16,
+			       &pi->pbsp,
+			       &pi->pbsu);
+
+	pi->dsp = BSP(pi->bsp) | BSU(pi->bsu);
+	pi->psp = BSP(pi->pbsp) | BSU(pi->pbsu);
+
+	WREG32(CG_BSP, pi->dsp);
+
+}
+
+void rv770_program_git(struct radeon_device *rdev)
+{
+	WREG32_P(CG_GIT, CG_GICST(R600_GICST_DFLT), ~CG_GICST_MASK);
+}
+
+void rv770_program_tp(struct radeon_device *rdev)
+{
+	int i;
+	enum r600_td td = R600_TD_DFLT;
+
+	for (i = 0; i < R600_PM_NUMBER_OF_TC; i++)
+		WREG32(CG_FFCT_0 + (i * 4), (UTC_0(r600_utc[i]) | DTC_0(r600_dtc[i])));
+
+	if (td == R600_TD_AUTO)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_FORCE_TREND_SEL);
+	else
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_FORCE_TREND_SEL, ~FIR_FORCE_TREND_SEL);
+	if (td == R600_TD_UP)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_TREND_MODE);
+	if (td == R600_TD_DOWN)
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_TREND_MODE, ~FIR_TREND_MODE);
+}
+
+void rv770_program_tpp(struct radeon_device *rdev)
+{
+	WREG32(CG_TPC, R600_TPC_DFLT);
+}
+
+void rv770_program_sstp(struct radeon_device *rdev)
+{
+	WREG32(CG_SSP, (SSTU(R600_SSTU_DFLT) | SST(R600_SST_DFLT)));
+}
+
+void rv770_program_engine_speed_parameters(struct radeon_device *rdev)
+{
+	WREG32_P(SPLL_CNTL_MODE, SPLL_DIV_SYNC, ~SPLL_DIV_SYNC);
+}
+
+static void rv770_enable_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL);
+
+	tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK);
+	tmp |= (DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE) |
+		DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE));
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+}
+
+void rv770_program_vc(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	WREG32(CG_FTV, pi->vrc);
+}
+
+void rv770_clear_vc(struct radeon_device *rdev)
+{
+	WREG32(CG_FTV, 0);
+}
+
+int rv770_upload_firmware(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int ret;
+
+	rv770_reset_smc(rdev);
+	rv770_stop_smc_clock(rdev);
+
+	ret = rv770_load_smc_ucode(rdev, pi->sram_end);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int rv770_populate_smc_acpi_state(struct radeon_device *rdev,
+					 RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	u32 mpll_ad_func_cntl =
+		pi->clk_regs.rv770.mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl =
+		pi->clk_regs.rv770.mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2 =
+		pi->clk_regs.rv770.mpll_dq_func_cntl_2;
+	u32 spll_func_cntl =
+		pi->clk_regs.rv770.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 =
+		pi->clk_regs.rv770.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 =
+		pi->clk_regs.rv770.cg_spll_func_cntl_3;
+	u32 mclk_pwrmgt_cntl;
+	u32 dll_cntl;
+
+	table->ACPIState = table->initialState;
+
+	table->ACPIState.flags &= ~PPSMC_SWSTATE_FLAG_DC;
+
+	if (pi->acpi_vddc) {
+		rv770_populate_vddc_value(rdev, pi->acpi_vddc,
+					  &table->ACPIState.levels[0].vddc);
+		if (pi->pcie_gen2) {
+			if (pi->acpi_pcie_gen2)
+				table->ACPIState.levels[0].gen2PCIE = 1;
+			else
+				table->ACPIState.levels[0].gen2PCIE = 0;
+		} else
+			table->ACPIState.levels[0].gen2PCIE = 0;
+		if (pi->acpi_pcie_gen2)
+			table->ACPIState.levels[0].gen2XSP = 1;
+		else
+			table->ACPIState.levels[0].gen2XSP = 0;
+	} else {
+		rv770_populate_vddc_value(rdev, pi->min_vddc_in_table,
+					  &table->ACPIState.levels[0].vddc);
+		table->ACPIState.levels[0].gen2PCIE = 0;
+	}
+
+
+	mpll_ad_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN;
+
+	mpll_dq_func_cntl_2 |= BIAS_GEN_PDNB | RESET_EN;
+
+	mclk_pwrmgt_cntl = (MRDCKA0_RESET |
+			    MRDCKA1_RESET |
+			    MRDCKB0_RESET |
+			    MRDCKB1_RESET |
+			    MRDCKC0_RESET |
+			    MRDCKC1_RESET |
+			    MRDCKD0_RESET |
+			    MRDCKD1_RESET);
+
+	dll_cntl = 0xff000000;
+
+	spll_func_cntl |= SPLL_RESET | SPLL_SLEEP | SPLL_BYPASS_EN;
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(4);
+
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL_2 = cpu_to_be32(mpll_ad_func_cntl_2);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL_2 = cpu_to_be32(mpll_dq_func_cntl_2);
+
+	table->ACPIState.levels[0].mclk.mclk770.vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	table->ACPIState.levels[0].mclk.mclk770.vDLL_CNTL = cpu_to_be32(dll_cntl);
+
+	table->ACPIState.levels[0].mclk.mclk770.mclk_value = 0;
+
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL = cpu_to_be32(spll_func_cntl);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(spll_func_cntl_2);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(spll_func_cntl_3);
+
+	table->ACPIState.levels[0].sclk.sclk_value = 0;
+
+	rv770_populate_mvdd_value(rdev, 0, &table->ACPIState.levels[0].mvdd);
+
+	table->ACPIState.levels[1] = table->ACPIState.levels[0];
+	table->ACPIState.levels[2] = table->ACPIState.levels[0];
+
+	return 0;
+}
+
+int rv770_populate_initial_mvdd_value(struct radeon_device *rdev,
+				      RV770_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if ((pi->s0_vid_lower_smio_cntl & pi->mvdd_mask_low) ==
+	     (pi->mvdd_low_smio[MVDD_LOW_INDEX] & pi->mvdd_mask_low) ) {
+		voltage->index = MVDD_LOW_INDEX;
+		voltage->value = cpu_to_be16(MVDD_LOW_VALUE);
+	} else {
+		voltage->index = MVDD_HIGH_INDEX;
+		voltage->value = cpu_to_be16(MVDD_HIGH_VALUE);
+	}
+
+	return 0;
+}
+
+static int rv770_populate_smc_initial_state(struct radeon_device *rdev,
+					    struct radeon_ps *radeon_state,
+					    RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_ps *initial_state = rv770_get_ps(radeon_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 a_t;
+
+	table->initialState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ad_func_cntl);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_AD_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ad_func_cntl_2);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_dq_func_cntl);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_DQ_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_dq_func_cntl_2);
+	table->initialState.levels[0].mclk.mclk770.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.mclk_pwrmgt_cntl);
+	table->initialState.levels[0].mclk.mclk770.vDLL_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.dll_cntl);
+
+	table->initialState.levels[0].mclk.mclk770.vMPLL_SS =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ss1);
+	table->initialState.levels[0].mclk.mclk770.vMPLL_SS2 =
+		cpu_to_be32(pi->clk_regs.rv770.mpll_ss2);
+
+	table->initialState.levels[0].mclk.mclk770.mclk_value =
+		cpu_to_be32(initial_state->low.mclk);
+
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_func_cntl);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_func_cntl_2);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_func_cntl_3);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_spread_spectrum);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM_2 =
+		cpu_to_be32(pi->clk_regs.rv770.cg_spll_spread_spectrum_2);
+
+	table->initialState.levels[0].sclk.sclk_value =
+		cpu_to_be32(initial_state->low.sclk);
+
+	table->initialState.levels[0].arbValue = MC_CG_ARB_FREQ_F0;
+
+	table->initialState.levels[0].seqValue =
+		rv770_get_seq_value(rdev, &initial_state->low);
+
+	rv770_populate_vddc_value(rdev,
+				  initial_state->low.vddc,
+				  &table->initialState.levels[0].vddc);
+	rv770_populate_initial_mvdd_value(rdev,
+					  &table->initialState.levels[0].mvdd);
+
+	a_t = CG_R(0xffff) | CG_L(0);
+	table->initialState.levels[0].aT = cpu_to_be32(a_t);
+
+	table->initialState.levels[0].bSP = cpu_to_be32(pi->dsp);
+
+	if (pi->boot_in_gen2)
+		table->initialState.levels[0].gen2PCIE = 1;
+	else
+		table->initialState.levels[0].gen2PCIE = 0;
+	if (initial_state->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2)
+		table->initialState.levels[0].gen2XSP = 1;
+	else
+		table->initialState.levels[0].gen2XSP = 0;
+
+	if (rdev->family == CHIP_RV740) {
+		if (pi->mem_gddr5) {
+			if (initial_state->low.mclk <= pi->mclk_strobe_mode_threshold)
+				table->initialState.levels[0].strobeMode =
+					rv740_get_mclk_frequency_ratio(initial_state->low.mclk) | 0x10;
+			else
+				table->initialState.levels[0].strobeMode = 0;
+
+			if (initial_state->low.mclk >= pi->mclk_edc_enable_threshold)
+				table->initialState.levels[0].mcFlags = SMC_MC_EDC_RD_FLAG | SMC_MC_EDC_WR_FLAG;
+			else
+				table->initialState.levels[0].mcFlags =  0;
+		}
+	}
+
+	table->initialState.levels[1] = table->initialState.levels[0];
+	table->initialState.levels[2] = table->initialState.levels[0];
+
+	table->initialState.flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	return 0;
+}
+
+static int rv770_populate_smc_vddc_table(struct radeon_device *rdev,
+					 RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+
+	for (i = 0; i < pi->valid_vddc_entries; i++) {
+		table->highSMIO[pi->vddc_table[i].vddc_index] =
+			pi->vddc_table[i].high_smio;
+		table->lowSMIO[pi->vddc_table[i].vddc_index] =
+			cpu_to_be32(pi->vddc_table[i].low_smio);
+	}
+
+	table->voltageMaskTable.highMask[RV770_SMC_VOLTAGEMASK_VDDC] = 0;
+	table->voltageMaskTable.lowMask[RV770_SMC_VOLTAGEMASK_VDDC] =
+		cpu_to_be32(pi->vddc_mask_low);
+
+	for (i = 0;
+	     ((i < pi->valid_vddc_entries) &&
+	      (pi->max_vddc_in_table >
+	       pi->vddc_table[i].vddc));
+	     i++);
+
+	table->maxVDDCIndexInPPTable =
+		pi->vddc_table[i].vddc_index;
+
+	return 0;
+}
+
+static int rv770_populate_smc_mvdd_table(struct radeon_device *rdev,
+					 RV770_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (pi->mvdd_control) {
+		table->lowSMIO[MVDD_HIGH_INDEX] |=
+			cpu_to_be32(pi->mvdd_low_smio[MVDD_HIGH_INDEX]);
+		table->lowSMIO[MVDD_LOW_INDEX] |=
+			cpu_to_be32(pi->mvdd_low_smio[MVDD_LOW_INDEX]);
+
+		table->voltageMaskTable.highMask[RV770_SMC_VOLTAGEMASK_MVDD] = 0;
+		table->voltageMaskTable.lowMask[RV770_SMC_VOLTAGEMASK_MVDD] =
+			cpu_to_be32(pi->mvdd_mask_low);
+	}
+
+	return 0;
+}
+
+static int rv770_init_smc_table(struct radeon_device *rdev,
+				struct radeon_ps *radeon_boot_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct rv7xx_ps *boot_state = rv770_get_ps(radeon_boot_state);
+	RV770_SMC_STATETABLE *table = &pi->smc_statetable;
+	int ret;
+
+	memset(table, 0, sizeof(RV770_SMC_STATETABLE));
+
+	pi->boot_sclk = boot_state->low.sclk;
+
+	rv770_populate_smc_vddc_table(rdev, table);
+	rv770_populate_smc_mvdd_table(rdev, table);
+
+	switch (rdev->pm.int_thermal_type) {
+        case THERMAL_TYPE_RV770:
+        case THERMAL_TYPE_ADT7473_WITH_INTERNAL:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_INTERNAL;
+		break;
+        case THERMAL_TYPE_NONE:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_NONE;
+		break;
+        case THERMAL_TYPE_EXTERNAL_GPIO:
+        default:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_EXTERNAL;
+		break;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_HARDWAREDC) {
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GPIO_DC;
+
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_DONT_WAIT_FOR_VBLANK_ON_ALERT)
+			table->extraFlags |= PPSMC_EXTRAFLAGS_AC2DC_DONT_WAIT_FOR_VBLANK;
+
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_GOTO_BOOT_ON_ALERT)
+			table->extraFlags |= PPSMC_EXTRAFLAGS_AC2DC_ACTION_GOTOINITIALSTATE;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_STEPVDDC;
+
+	if (pi->mem_gddr5)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GDDR5;
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		ret = rv730_populate_smc_initial_state(rdev, radeon_boot_state, table);
+	else
+		ret = rv770_populate_smc_initial_state(rdev, radeon_boot_state, table);
+	if (ret)
+		return ret;
+
+	if (rdev->family == CHIP_RV740)
+		ret = rv740_populate_smc_acpi_state(rdev, table);
+	else if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		ret = rv730_populate_smc_acpi_state(rdev, table);
+	else
+		ret = rv770_populate_smc_acpi_state(rdev, table);
+	if (ret)
+		return ret;
+
+	table->driverState = table->initialState;
+
+	return rv770_copy_bytes_to_smc(rdev,
+				       pi->state_table_start,
+				       (const u8 *)table,
+				       sizeof(RV770_SMC_STATETABLE),
+				       pi->sram_end);
+}
+
+static int rv770_construct_vddc_table(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u16 min, max, step;
+	u32 steps = 0;
+	u8 vddc_index = 0;
+	u32 i;
+
+	radeon_atom_get_min_voltage(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, &min);
+	radeon_atom_get_max_voltage(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, &max);
+	radeon_atom_get_voltage_step(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, &step);
+
+	steps = (max - min) / step + 1;
+
+	if (steps > MAX_NO_VREG_STEPS)
+		return -EINVAL;
+
+	for (i = 0; i < steps; i++) {
+		u32 gpio_pins, gpio_mask;
+
+		pi->vddc_table[i].vddc = (u16)(min + i * step);
+		radeon_atom_get_voltage_gpio_settings(rdev,
+						      pi->vddc_table[i].vddc,
+						      SET_VOLTAGE_TYPE_ASIC_VDDC,
+						      &gpio_pins, &gpio_mask);
+		pi->vddc_table[i].low_smio = gpio_pins & gpio_mask;
+		pi->vddc_table[i].high_smio = 0;
+		pi->vddc_mask_low = gpio_mask;
+		if (i > 0) {
+			if ((pi->vddc_table[i].low_smio !=
+			     pi->vddc_table[i - 1].low_smio ) ||
+			     (pi->vddc_table[i].high_smio !=
+			      pi->vddc_table[i - 1].high_smio))
+				vddc_index++;
+		}
+		pi->vddc_table[i].vddc_index = vddc_index;
+	}
+
+	pi->valid_vddc_entries = (u8)steps;
+
+	return 0;
+}
+
+static u32 rv770_get_mclk_split_point(struct atom_memory_info *memory_info)
+{
+	if (memory_info->mem_type == MEM_TYPE_GDDR3)
+		return 30000;
+
+	return 0;
+}
+
+static int rv770_get_mvdd_pin_configuration(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 gpio_pins, gpio_mask;
+
+	radeon_atom_get_voltage_gpio_settings(rdev,
+					      MVDD_HIGH_VALUE, SET_VOLTAGE_TYPE_ASIC_MVDDC,
+					      &gpio_pins, &gpio_mask);
+	pi->mvdd_mask_low = gpio_mask;
+	pi->mvdd_low_smio[MVDD_HIGH_INDEX] =
+		gpio_pins & gpio_mask;
+
+	radeon_atom_get_voltage_gpio_settings(rdev,
+					      MVDD_LOW_VALUE, SET_VOLTAGE_TYPE_ASIC_MVDDC,
+					      &gpio_pins, &gpio_mask);
+	pi->mvdd_low_smio[MVDD_LOW_INDEX] =
+		gpio_pins & gpio_mask;
+
+	return 0;
+}
+
+u8 rv770_get_memory_module_index(struct radeon_device *rdev)
+{
+	return (u8) ((RREG32(BIOS_SCRATCH_4) >> 16) & 0xff);
+}
+
+static int rv770_get_mvdd_configuration(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 memory_module_index;
+	struct atom_memory_info memory_info;
+
+	memory_module_index = rv770_get_memory_module_index(rdev);
+
+	if (radeon_atom_get_memory_info(rdev, memory_module_index, &memory_info)) {
+		pi->mvdd_control = false;
+		return 0;
+	}
+
+	pi->mvdd_split_frequency =
+		rv770_get_mclk_split_point(&memory_info);
+
+	if (pi->mvdd_split_frequency == 0) {
+		pi->mvdd_control = false;
+		return 0;
+	}
+
+	return rv770_get_mvdd_pin_configuration(rdev);
+}
+
+void rv770_enable_voltage_control(struct radeon_device *rdev,
+				  bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, VOLT_PWRMGT_EN, ~VOLT_PWRMGT_EN);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~VOLT_PWRMGT_EN);
+}
+
+static void rv770_program_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL);
+
+	tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK);
+	if (rdev->pm.dpm.new_active_crtcs & 1) {
+		tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK);
+		tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+	} else if (rdev->pm.dpm.new_active_crtcs & 2) {
+		tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+		tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_VBLANK);
+	} else {
+		tmp |= DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+		tmp |= DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE);
+	}
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+}
+
+static void rv770_enable_dynamic_pcie_gen2(struct radeon_device *rdev,
+					   bool enable)
+{
+	rv770_enable_bif_dynamic_pcie_gen2(rdev, enable);
+
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, ENABLE_GEN2PCIE, ~ENABLE_GEN2PCIE);
+	else
+		WREG32_P(GENERAL_PWRMGT, 0, ~ENABLE_GEN2PCIE);
+}
+
+static void r7xx_program_memory_timing_parameters(struct radeon_device *rdev,
+						  struct radeon_ps *radeon_new_state)
+{
+	if ((rdev->family == CHIP_RV730) ||
+	    (rdev->family == CHIP_RV710) ||
+	    (rdev->family == CHIP_RV740))
+		rv730_program_memory_timing_parameters(rdev, radeon_new_state);
+	else
+		rv770_program_memory_timing_parameters(rdev, radeon_new_state);
+}
+
+static int rv770_upload_sw_state(struct radeon_device *rdev,
+				 struct radeon_ps *radeon_new_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u16 address = pi->state_table_start +
+		offsetof(RV770_SMC_STATETABLE, driverState);
+	RV770_SMC_SWSTATE state = { 0 };
+	int ret;
+
+	ret = rv770_convert_power_state_to_smc(rdev, radeon_new_state, &state);
+	if (ret)
+		return ret;
+
+	return rv770_copy_bytes_to_smc(rdev, address, (const u8 *)&state,
+				       sizeof(RV770_SMC_SWSTATE),
+				       pi->sram_end);
+}
+
+int rv770_halt_smc(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_Halt) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	if (rv770_wait_for_smc_inactive(rdev) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return 0;
+}
+
+int rv770_resume_smc(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_Resume) != PPSMC_Result_OK)
+		return -EINVAL;
+	return 0;
+}
+
+int rv770_set_sw_state(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_SwitchToSwState) != PPSMC_Result_OK)
+		return -EINVAL;
+	return 0;
+}
+
+int rv770_set_boot_state(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_SwitchToInitialState) != PPSMC_Result_OK)
+		return -EINVAL;
+	return 0;
+}
+
+void rv770_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+					      struct radeon_ps *new_ps,
+					      struct radeon_ps *old_ps)
+{
+	struct rv7xx_ps *new_state = rv770_get_ps(new_ps);
+	struct rv7xx_ps *current_state = rv770_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->high.sclk >= current_state->high.sclk)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+void rv770_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+					     struct radeon_ps *new_ps,
+					     struct radeon_ps *old_ps)
+{
+	struct rv7xx_ps *new_state = rv770_get_ps(new_ps);
+	struct rv7xx_ps *current_state = rv770_get_ps(old_ps);
+
+	if ((new_ps->vclk == old_ps->vclk) &&
+	    (new_ps->dclk == old_ps->dclk))
+		return;
+
+	if (new_state->high.sclk < current_state->high.sclk)
+		return;
+
+	radeon_set_uvd_clocks(rdev, new_ps->vclk, new_ps->dclk);
+}
+
+int rv770_restrict_performance_levels_before_switch(struct radeon_device *rdev)
+{
+	if (rv770_send_msg_to_smc(rdev, (PPSMC_Msg)(PPSMC_MSG_NoForcedLevel)) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	if (rv770_send_msg_to_smc(rdev, (PPSMC_Msg)(PPSMC_MSG_TwoLevelsDisabled)) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return 0;
+}
+
+int rv770_dpm_force_performance_level(struct radeon_device *rdev,
+				      enum radeon_dpm_forced_level level)
+{
+	PPSMC_Msg msg;
+
+	if (level == RADEON_DPM_FORCED_LEVEL_HIGH) {
+		if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_ZeroLevelsDisabled) != PPSMC_Result_OK)
+			return -EINVAL;
+		msg = PPSMC_MSG_ForceHigh;
+	} else if (level == RADEON_DPM_FORCED_LEVEL_LOW) {
+		if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_NoForcedLevel) != PPSMC_Result_OK)
+			return -EINVAL;
+		msg = (PPSMC_Msg)(PPSMC_MSG_TwoLevelsDisabled);
+	} else {
+		if (rv770_send_msg_to_smc(rdev, PPSMC_MSG_NoForcedLevel) != PPSMC_Result_OK)
+			return -EINVAL;
+		msg = (PPSMC_Msg)(PPSMC_MSG_ZeroLevelsDisabled);
+	}
+
+	if (rv770_send_msg_to_smc(rdev, msg) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	rdev->pm.dpm.forced_level = level;
+
+	return 0;
+}
+
+void r7xx_start_smc(struct radeon_device *rdev)
+{
+	rv770_start_smc(rdev);
+	rv770_start_smc_clock(rdev);
+}
+
+
+void r7xx_stop_smc(struct radeon_device *rdev)
+{
+	rv770_reset_smc(rdev);
+	rv770_stop_smc_clock(rdev);
+}
+
+static void rv770_read_clock_registers(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	pi->clk_regs.rv770.cg_spll_func_cntl =
+		RREG32(CG_SPLL_FUNC_CNTL);
+	pi->clk_regs.rv770.cg_spll_func_cntl_2 =
+		RREG32(CG_SPLL_FUNC_CNTL_2);
+	pi->clk_regs.rv770.cg_spll_func_cntl_3 =
+		RREG32(CG_SPLL_FUNC_CNTL_3);
+	pi->clk_regs.rv770.cg_spll_spread_spectrum =
+		RREG32(CG_SPLL_SPREAD_SPECTRUM);
+	pi->clk_regs.rv770.cg_spll_spread_spectrum_2 =
+		RREG32(CG_SPLL_SPREAD_SPECTRUM_2);
+	pi->clk_regs.rv770.mpll_ad_func_cntl =
+		RREG32(MPLL_AD_FUNC_CNTL);
+	pi->clk_regs.rv770.mpll_ad_func_cntl_2 =
+		RREG32(MPLL_AD_FUNC_CNTL_2);
+	pi->clk_regs.rv770.mpll_dq_func_cntl =
+		RREG32(MPLL_DQ_FUNC_CNTL);
+	pi->clk_regs.rv770.mpll_dq_func_cntl_2 =
+		RREG32(MPLL_DQ_FUNC_CNTL_2);
+	pi->clk_regs.rv770.mclk_pwrmgt_cntl =
+		RREG32(MCLK_PWRMGT_CNTL);
+	pi->clk_regs.rv770.dll_cntl = RREG32(DLL_CNTL);
+}
+
+static void r7xx_read_clock_registers(struct radeon_device *rdev)
+{
+	if (rdev->family == CHIP_RV740)
+		rv740_read_clock_registers(rdev);
+	else if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		rv730_read_clock_registers(rdev);
+	else
+		rv770_read_clock_registers(rdev);
+}
+
+void rv770_read_voltage_smio_registers(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	pi->s0_vid_lower_smio_cntl =
+		RREG32(S0_VID_LOWER_SMIO_CNTL);
+}
+
+void rv770_reset_smio_status(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 sw_smio_index, vid_smio_cntl;
+
+	sw_smio_index =
+		(RREG32(GENERAL_PWRMGT) & SW_SMIO_INDEX_MASK) >> SW_SMIO_INDEX_SHIFT;
+	switch (sw_smio_index) {
+        case 3:
+		vid_smio_cntl = RREG32(S3_VID_LOWER_SMIO_CNTL);
+		break;
+        case 2:
+		vid_smio_cntl = RREG32(S2_VID_LOWER_SMIO_CNTL);
+		break;
+        case 1:
+		vid_smio_cntl = RREG32(S1_VID_LOWER_SMIO_CNTL);
+		break;
+        case 0:
+		return;
+        default:
+		vid_smio_cntl = pi->s0_vid_lower_smio_cntl;
+		break;
+	}
+
+	WREG32(S0_VID_LOWER_SMIO_CNTL, vid_smio_cntl);
+	WREG32_P(GENERAL_PWRMGT, SW_SMIO_INDEX(0), ~SW_SMIO_INDEX_MASK);
+}
+
+void rv770_get_memory_type(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp;
+
+	tmp = RREG32(MC_SEQ_MISC0);
+
+	if (((tmp & MC_SEQ_MISC0_GDDR5_MASK) >> MC_SEQ_MISC0_GDDR5_SHIFT) ==
+	    MC_SEQ_MISC0_GDDR5_VALUE)
+		pi->mem_gddr5 = true;
+	else
+		pi->mem_gddr5 = false;
+
+}
+
+void rv770_get_pcie_gen2_status(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 tmp;
+
+	tmp = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+
+	if ((tmp & LC_OTHER_SIDE_EVER_SENT_GEN2) &&
+	    (tmp & LC_OTHER_SIDE_SUPPORTS_GEN2))
+		pi->pcie_gen2 = true;
+	else
+		pi->pcie_gen2 = false;
+
+	if (pi->pcie_gen2) {
+		if (tmp & LC_CURRENT_DATA_RATE)
+			pi->boot_in_gen2 = true;
+		else
+			pi->boot_in_gen2 = false;
+	} else
+		pi->boot_in_gen2 = false;
+}
+
+#if 0
+static int rv770_enter_ulp_state(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (pi->gfx_clock_gating) {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+		RREG32(GB_TILING_CONFIG);
+	}
+
+	WREG32_P(SMC_MSG, HOST_SMC_MSG(PPSMC_MSG_SwitchToMinimumPower),
+		 ~HOST_SMC_MSG_MASK);
+
+	udelay(7000);
+
+	return 0;
+}
+
+static int rv770_exit_ulp_state(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+
+	WREG32_P(SMC_MSG, HOST_SMC_MSG(PPSMC_MSG_ResumeFromMinimumPower),
+		 ~HOST_SMC_MSG_MASK);
+
+	udelay(7000);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (((RREG32(SMC_MSG) & HOST_SMC_RESP_MASK) >> HOST_SMC_RESP_SHIFT) == 1)
+			break;
+		udelay(1000);
+	}
+
+	if (pi->gfx_clock_gating)
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+
+	return 0;
+}
+#endif
+
+static void rv770_get_mclk_odt_threshold(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 memory_module_index;
+	struct atom_memory_info memory_info;
+
+	pi->mclk_odt_threshold = 0;
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710)) {
+		memory_module_index = rv770_get_memory_module_index(rdev);
+
+		if (radeon_atom_get_memory_info(rdev, memory_module_index, &memory_info))
+			return;
+
+		if (memory_info.mem_type == MEM_TYPE_DDR2 ||
+		    memory_info.mem_type == MEM_TYPE_DDR3)
+			pi->mclk_odt_threshold = 30000;
+	}
+}
+
+void rv770_get_max_vddc(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u16 vddc;
+
+	if (radeon_atom_get_max_vddc(rdev, 0, 0, &vddc))
+		pi->max_vddc = 0;
+	else
+		pi->max_vddc = vddc;
+}
+
+void rv770_program_response_times(struct radeon_device *rdev)
+{
+	u32 voltage_response_time, backbias_response_time;
+	u32 acpi_delay_time, vbi_time_out;
+	u32 vddc_dly, bb_dly, acpi_dly, vbi_dly;
+	u32 reference_clock;
+
+	voltage_response_time = (u32)rdev->pm.dpm.voltage_response_time;
+	backbias_response_time = (u32)rdev->pm.dpm.backbias_response_time;
+
+	if (voltage_response_time == 0)
+		voltage_response_time = 1000;
+
+	if (backbias_response_time == 0)
+		backbias_response_time = 1000;
+
+	acpi_delay_time = 15000;
+	vbi_time_out = 100000;
+
+	reference_clock = radeon_get_xclk(rdev);
+
+	vddc_dly = (voltage_response_time  * reference_clock) / 1600;
+	bb_dly = (backbias_response_time * reference_clock) / 1600;
+	acpi_dly = (acpi_delay_time * reference_clock) / 1600;
+	vbi_dly = (vbi_time_out * reference_clock) / 1600;
+
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_delay_vreg, vddc_dly);
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_delay_bbias, bb_dly);
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_delay_acpi, acpi_dly);
+	rv770_write_smc_soft_register(rdev,
+				      RV770_SMC_SOFT_REGISTER_mclk_chg_timeout, vbi_dly);
+#if 0
+	/* XXX look up hw revision */
+	if (WEKIVA_A21)
+		rv770_write_smc_soft_register(rdev,
+					      RV770_SMC_SOFT_REGISTER_baby_step_timer,
+					      0x10);
+#endif
+}
+
+static void rv770_program_dcodt_before_state_switch(struct radeon_device *rdev,
+						    struct radeon_ps *radeon_new_state,
+						    struct radeon_ps *radeon_current_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct rv7xx_ps *new_state = rv770_get_ps(radeon_new_state);
+	struct rv7xx_ps *current_state = rv770_get_ps(radeon_current_state);
+	bool current_use_dc = false;
+	bool new_use_dc = false;
+
+	if (pi->mclk_odt_threshold == 0)
+		return;
+
+	if (current_state->high.mclk <= pi->mclk_odt_threshold)
+		current_use_dc = true;
+
+	if (new_state->high.mclk <= pi->mclk_odt_threshold)
+		new_use_dc = true;
+
+	if (current_use_dc == new_use_dc)
+		return;
+
+	if (!current_use_dc && new_use_dc)
+		return;
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		rv730_program_dcodt(rdev, new_use_dc);
+}
+
+static void rv770_program_dcodt_after_state_switch(struct radeon_device *rdev,
+						   struct radeon_ps *radeon_new_state,
+						   struct radeon_ps *radeon_current_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct rv7xx_ps *new_state = rv770_get_ps(radeon_new_state);
+	struct rv7xx_ps *current_state = rv770_get_ps(radeon_current_state);
+	bool current_use_dc = false;
+	bool new_use_dc = false;
+
+	if (pi->mclk_odt_threshold == 0)
+		return;
+
+	if (current_state->high.mclk <= pi->mclk_odt_threshold)
+		current_use_dc = true;
+
+	if (new_state->high.mclk <= pi->mclk_odt_threshold)
+		new_use_dc = true;
+
+	if (current_use_dc == new_use_dc)
+		return;
+
+	if (current_use_dc && !new_use_dc)
+		return;
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		rv730_program_dcodt(rdev, new_use_dc);
+}
+
+static void rv770_retrieve_odt_values(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (pi->mclk_odt_threshold == 0)
+		return;
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		rv730_get_odt_values(rdev);
+}
+
+static void rv770_set_dpm_event_sources(struct radeon_device *rdev, u32 sources)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	bool want_thermal_protection;
+	enum radeon_dpm_event_src dpm_event_src;
+
+	switch (sources) {
+        case 0:
+        default:
+		want_thermal_protection = false;
+		break;
+        case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_DIGITAL;
+		break;
+
+        case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_EXTERNAL;
+		break;
+
+        case ((1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL) |
+	      (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL)):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL;
+		break;
+	}
+
+	if (want_thermal_protection) {
+		WREG32_P(CG_THERMAL_CTRL, DPM_EVENT_SRC(dpm_event_src), ~DPM_EVENT_SRC_MASK);
+		if (pi->thermal_protection)
+			WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS);
+	} else {
+		WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS);
+	}
+}
+
+void rv770_enable_auto_throttle_source(struct radeon_device *rdev,
+				       enum radeon_dpm_auto_throttle_src source,
+				       bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (enable) {
+		if (!(pi->active_auto_throttle_sources & (1 << source))) {
+			pi->active_auto_throttle_sources |= 1 << source;
+			rv770_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources);
+		}
+	} else {
+		if (pi->active_auto_throttle_sources & (1 << source)) {
+			pi->active_auto_throttle_sources &= ~(1 << source);
+			rv770_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources);
+		}
+	}
+}
+
+int rv770_set_thermal_temperature_range(struct radeon_device *rdev,
+					int min_temp, int max_temp)
+{
+	int low_temp = 0 * 1000;
+	int high_temp = 255 * 1000;
+
+	if (low_temp < min_temp)
+		low_temp = min_temp;
+	if (high_temp > max_temp)
+		high_temp = max_temp;
+	if (high_temp < low_temp) {
+		DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp);
+		return -EINVAL;
+	}
+
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTH(high_temp / 1000), ~DIG_THERM_INTH_MASK);
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTL(low_temp / 1000), ~DIG_THERM_INTL_MASK);
+	WREG32_P(CG_THERMAL_CTRL, DIG_THERM_DPM(high_temp / 1000), ~DIG_THERM_DPM_MASK);
+
+	rdev->pm.dpm.thermal.min_temp = low_temp;
+	rdev->pm.dpm.thermal.max_temp = high_temp;
+
+	return 0;
+}
+
+int rv770_dpm_enable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	if (pi->gfx_clock_gating)
+		rv770_restore_cgcg(rdev);
+
+	if (rv770_dpm_enabled(rdev))
+		return -EINVAL;
+
+	if (pi->voltage_control) {
+		rv770_enable_voltage_control(rdev, true);
+		ret = rv770_construct_vddc_table(rdev);
+		if (ret) {
+			DRM_ERROR("rv770_construct_vddc_table failed\n");
+			return ret;
+		}
+	}
+
+	if (pi->dcodt)
+		rv770_retrieve_odt_values(rdev);
+
+	if (pi->mvdd_control) {
+		ret = rv770_get_mvdd_configuration(rdev);
+		if (ret) {
+			DRM_ERROR("rv770_get_mvdd_configuration failed\n");
+			return ret;
+		}
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_BACKBIAS)
+		rv770_enable_backbias(rdev, true);
+
+	rv770_enable_spread_spectrum(rdev, true);
+
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, true);
+
+	rv770_program_mpll_timing_parameters(rdev);
+	rv770_setup_bsp(rdev);
+	rv770_program_git(rdev);
+	rv770_program_tp(rdev);
+	rv770_program_tpp(rdev);
+	rv770_program_sstp(rdev);
+	rv770_program_engine_speed_parameters(rdev);
+	rv770_enable_display_gap(rdev);
+	rv770_program_vc(rdev);
+
+	if (pi->dynamic_pcie_gen2)
+		rv770_enable_dynamic_pcie_gen2(rdev, true);
+
+	ret = rv770_upload_firmware(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_upload_firmware failed\n");
+		return ret;
+	}
+	ret = rv770_init_smc_table(rdev, boot_ps);
+	if (ret) {
+		DRM_ERROR("rv770_init_smc_table failed\n");
+		return ret;
+	}
+
+	rv770_program_response_times(rdev);
+	r7xx_start_smc(rdev);
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		rv730_start_dpm(rdev);
+	else
+		rv770_start_dpm(rdev);
+
+	if (pi->gfx_clock_gating)
+		rv770_gfx_clock_gating_enable(rdev, true);
+
+	if (pi->mg_clock_gating)
+		rv770_mg_clock_gating_enable(rdev, true);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		PPSMC_Result result;
+
+		ret = rv770_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+		result = rv770_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+
+		if (result != PPSMC_Result_OK)
+			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
+	}
+
+	rv770_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
+
+	return 0;
+}
+
+void rv770_dpm_disable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (!rv770_dpm_enabled(rdev))
+		return;
+
+	rv770_clear_vc(rdev);
+
+	if (pi->thermal_protection)
+		rv770_enable_thermal_protection(rdev, false);
+
+	rv770_enable_spread_spectrum(rdev, false);
+
+	if (pi->dynamic_pcie_gen2)
+		rv770_enable_dynamic_pcie_gen2(rdev, false);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	if (pi->gfx_clock_gating)
+		rv770_gfx_clock_gating_enable(rdev, false);
+
+	if (pi->mg_clock_gating)
+		rv770_mg_clock_gating_enable(rdev, false);
+
+	if ((rdev->family == CHIP_RV730) || (rdev->family == CHIP_RV710))
+		rv730_stop_dpm(rdev);
+	else
+		rv770_stop_dpm(rdev);
+
+	r7xx_stop_smc(rdev);
+	rv770_reset_smio_status(rdev);
+}
+
+int rv770_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct radeon_ps *new_ps = rdev->pm.dpm.requested_ps;
+	struct radeon_ps *old_ps = rdev->pm.dpm.current_ps;
+	int ret;
+
+	ret = rv770_restrict_performance_levels_before_switch(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_restrict_performance_levels_before_switch failed\n");
+		return ret;
+	}
+	rv770_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+	ret = rv770_halt_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_halt_smc failed\n");
+		return ret;
+	}
+	ret = rv770_upload_sw_state(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("rv770_upload_sw_state failed\n");
+		return ret;
+	}
+	r7xx_program_memory_timing_parameters(rdev, new_ps);
+	if (pi->dcodt)
+		rv770_program_dcodt_before_state_switch(rdev, new_ps, old_ps);
+	ret = rv770_resume_smc(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_resume_smc failed\n");
+		return ret;
+	}
+	ret = rv770_set_sw_state(rdev);
+	if (ret) {
+		DRM_ERROR("rv770_set_sw_state failed\n");
+		return ret;
+	}
+	if (pi->dcodt)
+		rv770_program_dcodt_after_state_switch(rdev, new_ps, old_ps);
+	rv770_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+
+	ret = rv770_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO);
+	if (ret) {
+		DRM_ERROR("rv770_dpm_force_performance_level failed\n");
+		return ret;
+	}
+
+	return 0;
+}
+
+void rv770_dpm_reset_asic(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+
+	rv770_restrict_performance_levels_before_switch(rdev);
+	if (pi->dcodt)
+		rv770_program_dcodt_before_state_switch(rdev, boot_ps, boot_ps);
+	rv770_set_boot_state(rdev);
+	if (pi->dcodt)
+		rv770_program_dcodt_after_state_switch(rdev, boot_ps, boot_ps);
+}
+
+void rv770_dpm_setup_asic(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	r7xx_read_clock_registers(rdev);
+	rv770_read_voltage_smio_registers(rdev);
+	rv770_get_memory_type(rdev);
+	if (pi->dcodt)
+		rv770_get_mclk_odt_threshold(rdev);
+	rv770_get_pcie_gen2_status(rdev);
+
+	rv770_enable_acpi_pm(rdev);
+
+	if (radeon_aspm != 0) {
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L0s)
+			rv770_enable_l0s(rdev);
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_ASPM_L1)
+			rv770_enable_l1(rdev);
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_TURNOFFPLL_ASPML1)
+			rv770_enable_pll_sleep_in_l1(rdev);
+	}
+}
+
+void rv770_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+	rv770_program_display_gap(rdev);
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void rv7xx_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					     struct radeon_ps *rps,
+					     struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info,
+					     u8 table_rev)
+{
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) {
+		rps->vclk = le32_to_cpu(non_clock_info->ulVCLK);
+		rps->dclk = le32_to_cpu(non_clock_info->ulDCLK);
+	} else if (r600_is_uvd_state(rps->class, rps->class2)) {
+		rps->vclk = RV770_DEFAULT_VCLK_FREQ;
+		rps->dclk = RV770_DEFAULT_DCLK_FREQ;
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT)
+		rdev->pm.dpm.boot_ps = rps;
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void rv7xx_parse_pplib_clock_info(struct radeon_device *rdev,
+					 struct radeon_ps *rps, int index,
+					 union pplib_clock_info *clock_info)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct rv7xx_ps *ps = rv770_get_ps(rps);
+	u32 sclk, mclk;
+	u16 vddc;
+	struct rv7xx_pl *pl;
+
+	switch (index) {
+	case 0:
+		pl = &ps->low;
+		break;
+	case 1:
+		pl = &ps->medium;
+		break;
+	case 2:
+	default:
+		pl = &ps->high;
+		break;
+	}
+
+	if (rdev->family >= CHIP_CEDAR) {
+		sclk = le16_to_cpu(clock_info->evergreen.usEngineClockLow);
+		sclk |= clock_info->evergreen.ucEngineClockHigh << 16;
+		mclk = le16_to_cpu(clock_info->evergreen.usMemoryClockLow);
+		mclk |= clock_info->evergreen.ucMemoryClockHigh << 16;
+
+		pl->vddc = le16_to_cpu(clock_info->evergreen.usVDDC);
+		pl->vddci = le16_to_cpu(clock_info->evergreen.usVDDCI);
+		pl->flags = le32_to_cpu(clock_info->evergreen.ulFlags);
+	} else {
+		sclk = le16_to_cpu(clock_info->r600.usEngineClockLow);
+		sclk |= clock_info->r600.ucEngineClockHigh << 16;
+		mclk = le16_to_cpu(clock_info->r600.usMemoryClockLow);
+		mclk |= clock_info->r600.ucMemoryClockHigh << 16;
+
+		pl->vddc = le16_to_cpu(clock_info->r600.usVDDC);
+		pl->flags = le32_to_cpu(clock_info->r600.ulFlags);
+	}
+
+	pl->mclk = mclk;
+	pl->sclk = sclk;
+
+	/* patch up vddc if necessary */
+	if (pl->vddc == 0xff01) {
+		if (radeon_atom_get_max_vddc(rdev, 0, 0, &vddc) == 0)
+			pl->vddc = vddc;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_ACPI) {
+		pi->acpi_vddc = pl->vddc;
+		if (rdev->family >= CHIP_CEDAR)
+			eg_pi->acpi_vddci = pl->vddci;
+		if (ps->low.flags & ATOM_PPLIB_R600_FLAGS_PCIEGEN2)
+			pi->acpi_pcie_gen2 = true;
+		else
+			pi->acpi_pcie_gen2 = false;
+	}
+
+	if (rps->class2 & ATOM_PPLIB_CLASSIFICATION2_ULV) {
+		if (rdev->family >= CHIP_BARTS) {
+			eg_pi->ulv.supported = true;
+			eg_pi->ulv.pl = pl;
+		}
+	}
+
+	if (pi->min_vddc_in_table > pl->vddc)
+		pi->min_vddc_in_table = pl->vddc;
+
+	if (pi->max_vddc_in_table < pl->vddc)
+		pi->max_vddc_in_table = pl->vddc;
+
+	/* patch up boot state */
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		u16 vddc, vddci, mvdd;
+		radeon_atombios_get_default_voltages(rdev, &vddc, &vddci, &mvdd);
+		pl->mclk = rdev->clock.default_mclk;
+		pl->sclk = rdev->clock.default_sclk;
+		pl->vddc = vddc;
+		pl->vddci = vddci;
+	}
+
+	if (rdev->family >= CHIP_BARTS) {
+		if ((rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) ==
+		    ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE) {
+			rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk = pl->sclk;
+			rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.mclk = pl->mclk;
+			rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddc = pl->vddc;
+			rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddci = pl->vddci;
+		}
+	}
+}
+
+int rv7xx_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i, j;
+	union pplib_clock_info *clock_info;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	struct rv7xx_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  power_info->pplib.ucNumStates, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+
+	for (i = 0; i < power_info->pplib.ucNumStates; i++) {
+		power_state = (union pplib_power_state *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usStateArrayOffset) +
+			 i * power_info->pplib.ucStateEntrySize);
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			(mode_info->atom_context->bios + data_offset +
+			 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset) +
+			 (power_state->v1.ucNonClockStateIndex *
+			  power_info->pplib.ucNonClockSize));
+		if (power_info->pplib.ucStateEntrySize - 1) {
+			ps = kzalloc(sizeof(struct rv7xx_ps), GFP_KERNEL);
+			if (ps == NULL) {
+				kfree(rdev->pm.dpm.ps);
+				return -ENOMEM;
+			}
+			rdev->pm.dpm.ps[i].ps_priv = ps;
+			rv7xx_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+							 non_clock_info,
+							 power_info->pplib.ucNonClockSize);
+			for (j = 0; j < (power_info->pplib.ucStateEntrySize - 1); j++) {
+				clock_info = (union pplib_clock_info *)
+					(mode_info->atom_context->bios + data_offset +
+					 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) +
+					 (power_state->v1.ucClockStateIndices[j] *
+					  power_info->pplib.ucClockInfoSize));
+				rv7xx_parse_pplib_clock_info(rdev,
+							     &rdev->pm.dpm.ps[i], j,
+							     clock_info);
+			}
+		}
+	}
+	rdev->pm.dpm.num_ps = power_info->pplib.ucNumStates;
+	return 0;
+}
+
+int rv770_dpm_init(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi;
+	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	uint16_t data_offset, size;
+	uint8_t frev, crev;
+	struct atom_clock_dividers dividers;
+	int ret;
+
+	pi = kzalloc(sizeof(struct rv7xx_power_info), GFP_KERNEL);
+	if (pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = pi;
+
+	rv770_get_max_vddc(rdev);
+
+	pi->acpi_vddc = 0;
+	pi->min_vddc_in_table = 0;
+	pi->max_vddc_in_table = 0;
+
+	ret = rv7xx_parse_power_table(rdev);
+	if (ret)
+		return ret;
+
+	if (rdev->pm.dpm.voltage_response_time == 0)
+		rdev->pm.dpm.voltage_response_time = R600_VOLTAGERESPONSETIME_DFLT;
+	if (rdev->pm.dpm.backbias_response_time == 0)
+		rdev->pm.dpm.backbias_response_time = R600_BACKBIASRESPONSETIME_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->ref_div = dividers.ref_div + 1;
+	else
+		pi->ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	pi->mclk_strobe_mode_threshold = 30000;
+	pi->mclk_edc_enable_threshold = 30000;
+
+	pi->rlp = RV770_RLP_DFLT;
+	pi->rmp = RV770_RMP_DFLT;
+	pi->lhp = RV770_LHP_DFLT;
+	pi->lmp = RV770_LMP_DFLT;
+
+	pi->voltage_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, 0);
+
+	pi->mvdd_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, 0);
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+                                   &frev, &crev, &data_offset)) {
+		pi->sclk_ss = true;
+		pi->mclk_ss = true;
+		pi->dynamic_ss = true;
+	} else {
+		pi->sclk_ss = false;
+		pi->mclk_ss = false;
+		pi->dynamic_ss = false;
+	}
+
+	pi->asi = RV770_ASI_DFLT;
+	pi->pasi = RV770_HASI_DFLT;
+	pi->vrc = RV770_VRC_DFLT;
+
+	pi->power_gating = false;
+
+	pi->gfx_clock_gating = true;
+
+	pi->mg_clock_gating = true;
+	pi->mgcgtssm = true;
+
+	pi->dynamic_pcie_gen2 = true;
+
+	if (pi->gfx_clock_gating &&
+	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+		pi->thermal_protection = true;
+	else
+		pi->thermal_protection = false;
+
+	pi->display_gap = true;
+
+	if (rdev->flags & RADEON_IS_MOBILITY)
+		pi->dcodt = true;
+	else
+		pi->dcodt = false;
+
+	pi->ulps = true;
+
+	pi->mclk_stutter_mode_threshold = 0;
+
+	pi->sram_end = SMC_RAM_END;
+	pi->state_table_start = RV770_SMC_TABLE_ADDRESS;
+	pi->soft_regs_start = RV770_SMC_SOFT_REGISTERS_START;
+
+	return 0;
+}
+
+void rv770_dpm_print_power_state(struct radeon_device *rdev,
+				 struct radeon_ps *rps)
+{
+	struct rv7xx_ps *ps = rv770_get_ps(rps);
+	struct rv7xx_pl *pl;
+
+	r600_dpm_print_class_info(rps->class, rps->class2);
+	r600_dpm_print_cap_info(rps->caps);
+	printk("\tuvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+	if (rdev->family >= CHIP_CEDAR) {
+		pl = &ps->low;
+		printk("\t\tpower level 0    sclk: %u mclk: %u vddc: %u vddci: %u\n",
+		       pl->sclk, pl->mclk, pl->vddc, pl->vddci);
+		pl = &ps->medium;
+		printk("\t\tpower level 1    sclk: %u mclk: %u vddc: %u vddci: %u\n",
+		       pl->sclk, pl->mclk, pl->vddc, pl->vddci);
+		pl = &ps->high;
+		printk("\t\tpower level 2    sclk: %u mclk: %u vddc: %u vddci: %u\n",
+		       pl->sclk, pl->mclk, pl->vddc, pl->vddci);
+	} else {
+		pl = &ps->low;
+		printk("\t\tpower level 0    sclk: %u mclk: %u vddc: %u\n",
+		       pl->sclk, pl->mclk, pl->vddc);
+		pl = &ps->medium;
+		printk("\t\tpower level 1    sclk: %u mclk: %u vddc: %u\n",
+		       pl->sclk, pl->mclk, pl->vddc);
+		pl = &ps->high;
+		printk("\t\tpower level 2    sclk: %u mclk: %u vddc: %u\n",
+		       pl->sclk, pl->mclk, pl->vddc);
+	}
+	r600_dpm_print_ps_status(rdev, rps);
+}
+
+void rv770_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						       struct seq_file *m)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct rv7xx_ps *ps = rv770_get_ps(rps);
+	struct rv7xx_pl *pl;
+	u32 current_index =
+		(RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_PROFILE_INDEX_MASK) >>
+		CURRENT_PROFILE_INDEX_SHIFT;
+
+	if (current_index > 2) {
+		seq_printf(m, "invalid dpm profile %d\n", current_index);
+	} else {
+		if (current_index == 0)
+			pl = &ps->low;
+		else if (current_index == 1)
+			pl = &ps->medium;
+		else /* current_index == 2 */
+			pl = &ps->high;
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		if (rdev->family >= CHIP_CEDAR) {
+			seq_printf(m, "power level %d    sclk: %u mclk: %u vddc: %u vddci: %u\n",
+				   current_index, pl->sclk, pl->mclk, pl->vddc, pl->vddci);
+		} else {
+			seq_printf(m, "power level %d    sclk: %u mclk: %u vddc: %u\n",
+				   current_index, pl->sclk, pl->mclk, pl->vddc);
+		}
+	}
+}
+
+void rv770_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+}
+
+u32 rv770_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct rv7xx_ps *requested_state = rv770_get_ps(rdev->pm.dpm.requested_ps);
+
+	if (low)
+		return requested_state->low.sclk;
+	else
+		return requested_state->high.sclk;
+}
+
+u32 rv770_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct rv7xx_ps *requested_state = rv770_get_ps(rdev->pm.dpm.requested_ps);
+
+	if (low)
+		return requested_state->low.mclk;
+	else
+		return requested_state->high.mclk;
+}
+
+bool rv770_dpm_vblank_too_short(struct radeon_device *rdev)
+{
+	u32 vblank_time = r600_dpm_get_vblank_time(rdev);
+
+	if (vblank_time < 300)
+		return true;
+	else
+		return false;
+
+}
diff --git a/drivers/gpu/drm/radeon/rv770_dpm.h b/drivers/gpu/drm/radeon/rv770_dpm.h
new file mode 100644
index 000000000000..96b1b2a62a8a
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv770_dpm.h
@@ -0,0 +1,289 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __RV770_DPM_H__
+#define __RV770_DPM_H__
+
+#include "rv770_smc.h"
+
+struct rv770_clock_registers {
+	u32 cg_spll_func_cntl;
+	u32 cg_spll_func_cntl_2;
+	u32 cg_spll_func_cntl_3;
+	u32 cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2;
+	u32 mpll_ad_func_cntl;
+	u32 mpll_ad_func_cntl_2;
+	u32 mpll_dq_func_cntl;
+	u32 mpll_dq_func_cntl_2;
+	u32 mclk_pwrmgt_cntl;
+	u32 dll_cntl;
+	u32 mpll_ss1;
+	u32 mpll_ss2;
+};
+
+struct rv730_clock_registers {
+	u32 cg_spll_func_cntl;
+	u32 cg_spll_func_cntl_2;
+	u32 cg_spll_func_cntl_3;
+	u32 cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2;
+	u32 mclk_pwrmgt_cntl;
+	u32 dll_cntl;
+	u32 mpll_func_cntl;
+	u32 mpll_func_cntl2;
+	u32 mpll_func_cntl3;
+	u32 mpll_ss;
+	u32 mpll_ss2;
+};
+
+union r7xx_clock_registers {
+	struct rv770_clock_registers rv770;
+	struct rv730_clock_registers rv730;
+};
+
+struct vddc_table_entry {
+	u16 vddc;
+	u8 vddc_index;
+	u8 high_smio;
+	u32 low_smio;
+};
+
+#define MAX_NO_OF_MVDD_VALUES 2
+#define MAX_NO_VREG_STEPS 32
+
+struct rv7xx_power_info {
+	/* flags */
+	bool mem_gddr5;
+	bool pcie_gen2;
+	bool dynamic_pcie_gen2;
+	bool acpi_pcie_gen2;
+	bool boot_in_gen2;
+	bool voltage_control; /* vddc */
+	bool mvdd_control;
+	bool sclk_ss;
+	bool mclk_ss;
+	bool dynamic_ss;
+	bool gfx_clock_gating;
+	bool mg_clock_gating;
+	bool mgcgtssm;
+	bool power_gating;
+	bool thermal_protection;
+	bool display_gap;
+	bool dcodt;
+	bool ulps;
+	/* registers */
+	union r7xx_clock_registers clk_regs;
+	u32 s0_vid_lower_smio_cntl;
+	/* voltage */
+	u32 vddc_mask_low;
+	u32 mvdd_mask_low;
+	u32 mvdd_split_frequency;
+	u32 mvdd_low_smio[MAX_NO_OF_MVDD_VALUES];
+	u16 max_vddc;
+	u16 max_vddc_in_table;
+	u16 min_vddc_in_table;
+	struct vddc_table_entry vddc_table[MAX_NO_VREG_STEPS];
+	u8 valid_vddc_entries;
+	/* dc odt */
+	u32 mclk_odt_threshold;
+	u8 odt_value_0[2];
+	u8 odt_value_1[2];
+	/* stored values */
+	u32 boot_sclk;
+	u16 acpi_vddc;
+	u32 ref_div;
+	u32 active_auto_throttle_sources;
+	u32 mclk_stutter_mode_threshold;
+	u32 mclk_strobe_mode_threshold;
+	u32 mclk_edc_enable_threshold;
+	u32 bsp;
+	u32 bsu;
+	u32 pbsp;
+	u32 pbsu;
+	u32 dsp;
+	u32 psp;
+	u32 asi;
+	u32 pasi;
+	u32 vrc;
+	u32 restricted_levels;
+	u32 rlp;
+	u32 rmp;
+	u32 lhp;
+	u32 lmp;
+	/* smc offsets */
+	u16 state_table_start;
+	u16 soft_regs_start;
+	u16 sram_end;
+	/* scratch structs */
+	RV770_SMC_STATETABLE smc_statetable;
+};
+
+struct rv7xx_pl {
+	u32 sclk;
+	u32 mclk;
+	u16 vddc;
+	u16 vddci; /* eg+ only */
+	u32 flags;
+	enum radeon_pcie_gen pcie_gen; /* si+ only */
+};
+
+struct rv7xx_ps {
+	struct rv7xx_pl high;
+	struct rv7xx_pl medium;
+	struct rv7xx_pl low;
+	bool dc_compatible;
+};
+
+#define RV770_RLP_DFLT                                10
+#define RV770_RMP_DFLT                                25
+#define RV770_LHP_DFLT                                25
+#define RV770_LMP_DFLT                                10
+#define RV770_VRC_DFLT                                0x003f
+#define RV770_ASI_DFLT                                1000
+#define RV770_HASI_DFLT                               200000
+#define RV770_MGCGTTLOCAL0_DFLT                       0x00100000
+#define RV7XX_MGCGTTLOCAL0_DFLT                       0
+#define RV770_MGCGTTLOCAL1_DFLT                       0xFFFF0000
+#define RV770_MGCGCGTSSMCTRL_DFLT                     0x55940000
+
+#define MVDD_LOW_INDEX  0
+#define MVDD_HIGH_INDEX 1
+
+#define MVDD_LOW_VALUE  0
+#define MVDD_HIGH_VALUE 0xffff
+
+#define RV770_DEFAULT_VCLK_FREQ  53300 /* 10 khz */
+#define RV770_DEFAULT_DCLK_FREQ  40000 /* 10 khz */
+
+/* rv730/rv710 */
+int rv730_populate_sclk_value(struct radeon_device *rdev,
+			      u32 engine_clock,
+			      RV770_SMC_SCLK_VALUE *sclk);
+int rv730_populate_mclk_value(struct radeon_device *rdev,
+			      u32 engine_clock, u32 memory_clock,
+			      LPRV7XX_SMC_MCLK_VALUE mclk);
+void rv730_read_clock_registers(struct radeon_device *rdev);
+int rv730_populate_smc_acpi_state(struct radeon_device *rdev,
+				  RV770_SMC_STATETABLE *table);
+int rv730_populate_smc_initial_state(struct radeon_device *rdev,
+				     struct radeon_ps *radeon_initial_state,
+				     RV770_SMC_STATETABLE *table);
+void rv730_program_memory_timing_parameters(struct radeon_device *rdev,
+					    struct radeon_ps *radeon_state);
+void rv730_power_gating_enable(struct radeon_device *rdev,
+			       bool enable);
+void rv730_start_dpm(struct radeon_device *rdev);
+void rv730_stop_dpm(struct radeon_device *rdev);
+void rv730_program_dcodt(struct radeon_device *rdev, bool use_dcodt);
+void rv730_get_odt_values(struct radeon_device *rdev);
+
+/* rv740 */
+int rv740_populate_sclk_value(struct radeon_device *rdev, u32 engine_clock,
+			      RV770_SMC_SCLK_VALUE *sclk);
+int rv740_populate_mclk_value(struct radeon_device *rdev,
+			      u32 engine_clock, u32 memory_clock,
+			      RV7XX_SMC_MCLK_VALUE *mclk);
+void rv740_read_clock_registers(struct radeon_device *rdev);
+int rv740_populate_smc_acpi_state(struct radeon_device *rdev,
+				  RV770_SMC_STATETABLE *table);
+void rv740_enable_mclk_spread_spectrum(struct radeon_device *rdev,
+				       bool enable);
+u8 rv740_get_mclk_frequency_ratio(u32 memory_clock);
+u32 rv740_get_dll_speed(bool is_gddr5, u32 memory_clock);
+u32 rv740_get_decoded_reference_divider(u32 encoded_ref);
+
+/* rv770 */
+u32 rv770_map_clkf_to_ibias(struct radeon_device *rdev, u32 clkf);
+int rv770_populate_vddc_value(struct radeon_device *rdev, u16 vddc,
+			      RV770_SMC_VOLTAGE_VALUE *voltage);
+int rv770_populate_mvdd_value(struct radeon_device *rdev, u32 mclk,
+			      RV770_SMC_VOLTAGE_VALUE *voltage);
+u8 rv770_get_seq_value(struct radeon_device *rdev,
+		       struct rv7xx_pl *pl);
+int rv770_populate_initial_mvdd_value(struct radeon_device *rdev,
+				      RV770_SMC_VOLTAGE_VALUE *voltage);
+u32 rv770_calculate_memory_refresh_rate(struct radeon_device *rdev,
+					u32 engine_clock);
+void rv770_program_response_times(struct radeon_device *rdev);
+int rv770_populate_smc_sp(struct radeon_device *rdev,
+			  struct radeon_ps *radeon_state,
+			  RV770_SMC_SWSTATE *smc_state);
+int rv770_populate_smc_t(struct radeon_device *rdev,
+			 struct radeon_ps *radeon_state,
+			 RV770_SMC_SWSTATE *smc_state);
+void rv770_read_voltage_smio_registers(struct radeon_device *rdev);
+void rv770_get_memory_type(struct radeon_device *rdev);
+void r7xx_start_smc(struct radeon_device *rdev);
+u8 rv770_get_memory_module_index(struct radeon_device *rdev);
+void rv770_get_max_vddc(struct radeon_device *rdev);
+void rv770_get_pcie_gen2_status(struct radeon_device *rdev);
+void rv770_enable_acpi_pm(struct radeon_device *rdev);
+void rv770_restore_cgcg(struct radeon_device *rdev);
+bool rv770_dpm_enabled(struct radeon_device *rdev);
+void rv770_enable_voltage_control(struct radeon_device *rdev,
+				  bool enable);
+void rv770_enable_backbias(struct radeon_device *rdev,
+			   bool enable);
+void rv770_enable_thermal_protection(struct radeon_device *rdev,
+				     bool enable);
+void rv770_enable_auto_throttle_source(struct radeon_device *rdev,
+				       enum radeon_dpm_auto_throttle_src source,
+				       bool enable);
+void rv770_setup_bsp(struct radeon_device *rdev);
+void rv770_program_git(struct radeon_device *rdev);
+void rv770_program_tp(struct radeon_device *rdev);
+void rv770_program_tpp(struct radeon_device *rdev);
+void rv770_program_sstp(struct radeon_device *rdev);
+void rv770_program_engine_speed_parameters(struct radeon_device *rdev);
+void rv770_program_vc(struct radeon_device *rdev);
+void rv770_clear_vc(struct radeon_device *rdev);
+int rv770_upload_firmware(struct radeon_device *rdev);
+void rv770_stop_dpm(struct radeon_device *rdev);
+void r7xx_stop_smc(struct radeon_device *rdev);
+void rv770_reset_smio_status(struct radeon_device *rdev);
+int rv770_restrict_performance_levels_before_switch(struct radeon_device *rdev);
+int rv770_dpm_force_performance_level(struct radeon_device *rdev,
+				      enum radeon_dpm_forced_level level);
+int rv770_halt_smc(struct radeon_device *rdev);
+int rv770_resume_smc(struct radeon_device *rdev);
+int rv770_set_sw_state(struct radeon_device *rdev);
+int rv770_set_boot_state(struct radeon_device *rdev);
+int rv7xx_parse_power_table(struct radeon_device *rdev);
+void rv770_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+					      struct radeon_ps *new_ps,
+					      struct radeon_ps *old_ps);
+void rv770_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+					     struct radeon_ps *new_ps,
+					     struct radeon_ps *old_ps);
+
+/* smc */
+int rv770_read_smc_soft_register(struct radeon_device *rdev,
+				 u16 reg_offset, u32 *value);
+int rv770_write_smc_soft_register(struct radeon_device *rdev,
+				  u16 reg_offset, u32 value);
+
+/* thermal */
+int rv770_set_thermal_temperature_range(struct radeon_device *rdev,
+					int min_temp, int max_temp);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv770_smc.c b/drivers/gpu/drm/radeon/rv770_smc.c
new file mode 100644
index 000000000000..ab95da570215
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv770_smc.c
@@ -0,0 +1,621 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include <linux/firmware.h>
+#include "drmP.h"
+#include "radeon.h"
+#include "rv770d.h"
+#include "rv770_dpm.h"
+#include "rv770_smc.h"
+#include "atom.h"
+#include "radeon_ucode.h"
+
+#define FIRST_SMC_INT_VECT_REG 0xFFD8
+#define FIRST_INT_VECT_S19     0xFFC0
+
+static const u8 rv770_smc_int_vectors[] =
+{
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x0C, 0xD7,
+	0x08, 0x2B, 0x08, 0x10,
+	0x03, 0x51, 0x03, 0x51,
+	0x03, 0x51, 0x03, 0x51
+};
+
+static const u8 rv730_smc_int_vectors[] =
+{
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x08, 0x15,
+	0x08, 0x15, 0x0C, 0xBB,
+	0x08, 0x30, 0x08, 0x15,
+	0x03, 0x56, 0x03, 0x56,
+	0x03, 0x56, 0x03, 0x56
+};
+
+static const u8 rv710_smc_int_vectors[] =
+{
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x08, 0x04,
+	0x08, 0x04, 0x0C, 0xCB,
+	0x08, 0x1F, 0x08, 0x04,
+	0x03, 0x51, 0x03, 0x51,
+	0x03, 0x51, 0x03, 0x51
+};
+
+static const u8 rv740_smc_int_vectors[] =
+{
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x08, 0x10,
+	0x08, 0x10, 0x0C, 0xD7,
+	0x08, 0x2B, 0x08, 0x10,
+	0x03, 0x51, 0x03, 0x51,
+	0x03, 0x51, 0x03, 0x51
+};
+
+static const u8 cedar_smc_int_vectors[] =
+{
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x11, 0x8B,
+	0x0B, 0x20, 0x0B, 0x05,
+	0x04, 0xF6, 0x04, 0xF6,
+	0x04, 0xF6, 0x04, 0xF6
+};
+
+static const u8 redwood_smc_int_vectors[] =
+{
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x11, 0x8B,
+	0x0B, 0x20, 0x0B, 0x05,
+	0x04, 0xF6, 0x04, 0xF6,
+	0x04, 0xF6, 0x04, 0xF6
+};
+
+static const u8 juniper_smc_int_vectors[] =
+{
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x11, 0x8B,
+	0x0B, 0x20, 0x0B, 0x05,
+	0x04, 0xF6, 0x04, 0xF6,
+	0x04, 0xF6, 0x04, 0xF6
+};
+
+static const u8 cypress_smc_int_vectors[] =
+{
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x0B, 0x05,
+	0x0B, 0x05, 0x11, 0x8B,
+	0x0B, 0x20, 0x0B, 0x05,
+	0x04, 0xF6, 0x04, 0xF6,
+	0x04, 0xF6, 0x04, 0xF6
+};
+
+static const u8 barts_smc_int_vectors[] =
+{
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x12, 0xAA,
+	0x0C, 0x2F, 0x15, 0xF6,
+	0x15, 0xF6, 0x05, 0x0A,
+	0x05, 0x0A, 0x05, 0x0A
+};
+
+static const u8 turks_smc_int_vectors[] =
+{
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x12, 0xAA,
+	0x0C, 0x2F, 0x15, 0xF6,
+	0x15, 0xF6, 0x05, 0x0A,
+	0x05, 0x0A, 0x05, 0x0A
+};
+
+static const u8 caicos_smc_int_vectors[] =
+{
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x0C, 0x14,
+	0x0C, 0x14, 0x12, 0xAA,
+	0x0C, 0x2F, 0x15, 0xF6,
+	0x15, 0xF6, 0x05, 0x0A,
+	0x05, 0x0A, 0x05, 0x0A
+};
+
+static const u8 cayman_smc_int_vectors[] =
+{
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x12, 0x05,
+	0x12, 0x05, 0x18, 0xEA,
+	0x12, 0x20, 0x1C, 0x34,
+	0x1C, 0x34, 0x08, 0x72,
+	0x08, 0x72, 0x08, 0x72
+};
+
+int rv770_set_smc_sram_address(struct radeon_device *rdev,
+			       u16 smc_address, u16 limit)
+{
+	u32 addr;
+
+	if (smc_address & 3)
+		return -EINVAL;
+	if ((smc_address + 3) > limit)
+		return -EINVAL;
+
+	addr = smc_address;
+	addr |= SMC_SRAM_AUTO_INC_DIS;
+
+	WREG32(SMC_SRAM_ADDR, addr);
+
+	return 0;
+}
+
+int rv770_copy_bytes_to_smc(struct radeon_device *rdev,
+			    u16 smc_start_address, const u8 *src,
+			    u16 byte_count, u16 limit)
+{
+	u32 data, original_data, extra_shift;
+	u16 addr;
+	int ret;
+
+	if (smc_start_address & 3)
+		return -EINVAL;
+	if ((smc_start_address + byte_count) > limit)
+		return -EINVAL;
+
+	addr = smc_start_address;
+
+	while (byte_count >= 4) {
+		/* SMC address space is BE */
+		data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3];
+
+		ret = rv770_set_smc_sram_address(rdev, addr, limit);
+		if (ret)
+			return ret;
+
+		WREG32(SMC_SRAM_DATA, data);
+
+		src += 4;
+		byte_count -= 4;
+		addr += 4;
+	}
+
+	/* RMW for final bytes */
+	if (byte_count > 0) {
+		data = 0;
+
+		ret = rv770_set_smc_sram_address(rdev, addr, limit);
+		if (ret)
+			return ret;
+
+		original_data = RREG32(SMC_SRAM_DATA);
+
+		extra_shift = 8 * (4 - byte_count);
+
+		while (byte_count > 0) {
+			/* SMC address space is BE */
+			data = (data << 8) + *src++;
+			byte_count--;
+		}
+
+		data <<= extra_shift;
+
+		data |= (original_data & ~((~0UL) << extra_shift));
+
+		ret = rv770_set_smc_sram_address(rdev, addr, limit);
+		if (ret)
+			return ret;
+
+		WREG32(SMC_SRAM_DATA, data);
+	}
+
+	return 0;
+}
+
+static int rv770_program_interrupt_vectors(struct radeon_device *rdev,
+					   u32 smc_first_vector, const u8 *src,
+					   u32 byte_count)
+{
+	u32 tmp, i;
+
+	if (byte_count % 4)
+		return -EINVAL;
+
+	if (smc_first_vector < FIRST_SMC_INT_VECT_REG) {
+		tmp = FIRST_SMC_INT_VECT_REG - smc_first_vector;
+
+		if (tmp > byte_count)
+			return 0;
+
+		byte_count -= tmp;
+		src += tmp;
+		smc_first_vector = FIRST_SMC_INT_VECT_REG;
+	}
+
+	for (i = 0; i < byte_count; i += 4) {
+		/* SMC address space is BE */
+		tmp = (src[i] << 24) | (src[i + 1] << 16) | (src[i + 2] << 8) | src[i + 3];
+
+		WREG32(SMC_ISR_FFD8_FFDB + i, tmp);
+	}
+
+	return 0;
+}
+
+void rv770_start_smc(struct radeon_device *rdev)
+{
+	WREG32_P(SMC_IO, SMC_RST_N, ~SMC_RST_N);
+}
+
+void rv770_reset_smc(struct radeon_device *rdev)
+{
+	WREG32_P(SMC_IO, 0, ~SMC_RST_N);
+}
+
+void rv770_stop_smc_clock(struct radeon_device *rdev)
+{
+	WREG32_P(SMC_IO, 0, ~SMC_CLK_EN);
+}
+
+void rv770_start_smc_clock(struct radeon_device *rdev)
+{
+	WREG32_P(SMC_IO, SMC_CLK_EN, ~SMC_CLK_EN);
+}
+
+bool rv770_is_smc_running(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	tmp = RREG32(SMC_IO);
+
+	if ((tmp & SMC_RST_N) && (tmp & SMC_CLK_EN))
+		return true;
+	else
+		return false;
+}
+
+PPSMC_Result rv770_send_msg_to_smc(struct radeon_device *rdev, PPSMC_Msg msg)
+{
+	u32 tmp;
+	int i;
+	PPSMC_Result result;
+
+	if (!rv770_is_smc_running(rdev))
+		return PPSMC_Result_Failed;
+
+	WREG32_P(SMC_MSG, HOST_SMC_MSG(msg), ~HOST_SMC_MSG_MASK);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32(SMC_MSG) & HOST_SMC_RESP_MASK;
+		tmp >>= HOST_SMC_RESP_SHIFT;
+		if (tmp != 0)
+			break;
+		udelay(1);
+	}
+
+	tmp = RREG32(SMC_MSG) & HOST_SMC_RESP_MASK;
+	tmp >>= HOST_SMC_RESP_SHIFT;
+
+	result = (PPSMC_Result)tmp;
+	return result;
+}
+
+PPSMC_Result rv770_wait_for_smc_inactive(struct radeon_device *rdev)
+{
+	int i;
+	PPSMC_Result result = PPSMC_Result_OK;
+
+	if (!rv770_is_smc_running(rdev))
+		return result;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(SMC_IO) & SMC_STOP_MODE)
+			break;
+		udelay(1);
+	}
+
+	return result;
+}
+
+static void rv770_clear_smc_sram(struct radeon_device *rdev, u16 limit)
+{
+	u16 i;
+
+	for (i = 0;  i < limit; i += 4) {
+		rv770_set_smc_sram_address(rdev, i, limit);
+		WREG32(SMC_SRAM_DATA, 0);
+	}
+}
+
+int rv770_load_smc_ucode(struct radeon_device *rdev,
+			 u16 limit)
+{
+	int ret;
+	const u8 *int_vect;
+	u16 int_vect_start_address;
+	u16 int_vect_size;
+	const u8 *ucode_data;
+	u16 ucode_start_address;
+	u16 ucode_size;
+
+	if (!rdev->smc_fw)
+		return -EINVAL;
+
+	rv770_clear_smc_sram(rdev, limit);
+
+	switch (rdev->family) {
+	case CHIP_RV770:
+		ucode_start_address = RV770_SMC_UCODE_START;
+		ucode_size = RV770_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&rv770_smc_int_vectors;
+		int_vect_start_address = RV770_SMC_INT_VECTOR_START;
+		int_vect_size = RV770_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_RV730:
+		ucode_start_address = RV730_SMC_UCODE_START;
+		ucode_size = RV730_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&rv730_smc_int_vectors;
+		int_vect_start_address = RV730_SMC_INT_VECTOR_START;
+		int_vect_size = RV730_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_RV710:
+		ucode_start_address = RV710_SMC_UCODE_START;
+		ucode_size = RV710_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&rv710_smc_int_vectors;
+		int_vect_start_address = RV710_SMC_INT_VECTOR_START;
+		int_vect_size = RV710_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_RV740:
+		ucode_start_address = RV740_SMC_UCODE_START;
+		ucode_size = RV740_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&rv740_smc_int_vectors;
+		int_vect_start_address = RV740_SMC_INT_VECTOR_START;
+		int_vect_size = RV740_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_CEDAR:
+		ucode_start_address = CEDAR_SMC_UCODE_START;
+		ucode_size = CEDAR_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&cedar_smc_int_vectors;
+		int_vect_start_address = CEDAR_SMC_INT_VECTOR_START;
+		int_vect_size = CEDAR_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_REDWOOD:
+		ucode_start_address = REDWOOD_SMC_UCODE_START;
+		ucode_size = REDWOOD_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&redwood_smc_int_vectors;
+		int_vect_start_address = REDWOOD_SMC_INT_VECTOR_START;
+		int_vect_size = REDWOOD_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_JUNIPER:
+		ucode_start_address = JUNIPER_SMC_UCODE_START;
+		ucode_size = JUNIPER_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&juniper_smc_int_vectors;
+		int_vect_start_address = JUNIPER_SMC_INT_VECTOR_START;
+		int_vect_size = JUNIPER_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_CYPRESS:
+	case CHIP_HEMLOCK:
+		ucode_start_address = CYPRESS_SMC_UCODE_START;
+		ucode_size = CYPRESS_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&cypress_smc_int_vectors;
+		int_vect_start_address = CYPRESS_SMC_INT_VECTOR_START;
+		int_vect_size = CYPRESS_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_BARTS:
+		ucode_start_address = BARTS_SMC_UCODE_START;
+		ucode_size = BARTS_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&barts_smc_int_vectors;
+		int_vect_start_address = BARTS_SMC_INT_VECTOR_START;
+		int_vect_size = BARTS_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_TURKS:
+		ucode_start_address = TURKS_SMC_UCODE_START;
+		ucode_size = TURKS_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&turks_smc_int_vectors;
+		int_vect_start_address = TURKS_SMC_INT_VECTOR_START;
+		int_vect_size = TURKS_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_CAICOS:
+		ucode_start_address = CAICOS_SMC_UCODE_START;
+		ucode_size = CAICOS_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&caicos_smc_int_vectors;
+		int_vect_start_address = CAICOS_SMC_INT_VECTOR_START;
+		int_vect_size = CAICOS_SMC_INT_VECTOR_SIZE;
+		break;
+	case CHIP_CAYMAN:
+		ucode_start_address = CAYMAN_SMC_UCODE_START;
+		ucode_size = CAYMAN_SMC_UCODE_SIZE;
+		int_vect = (const u8 *)&cayman_smc_int_vectors;
+		int_vect_start_address = CAYMAN_SMC_INT_VECTOR_START;
+		int_vect_size = CAYMAN_SMC_INT_VECTOR_SIZE;
+		break;
+	default:
+		DRM_ERROR("unknown asic in smc ucode loader\n");
+		BUG();
+	}
+
+	/* load the ucode */
+	ucode_data = (const u8 *)rdev->smc_fw->data;
+	ret = rv770_copy_bytes_to_smc(rdev, ucode_start_address,
+				      ucode_data, ucode_size, limit);
+	if (ret)
+		return ret;
+
+	/* set up the int vectors */
+	ret = rv770_program_interrupt_vectors(rdev, int_vect_start_address,
+					      int_vect, int_vect_size);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+int rv770_read_smc_sram_dword(struct radeon_device *rdev,
+			      u16 smc_address, u32 *value, u16 limit)
+{
+	int ret;
+
+	ret = rv770_set_smc_sram_address(rdev, smc_address, limit);
+	if (ret)
+		return ret;
+
+	*value = RREG32(SMC_SRAM_DATA);
+
+	return 0;
+}
+
+int rv770_write_smc_sram_dword(struct radeon_device *rdev,
+			       u16 smc_address, u32 value, u16 limit)
+{
+	int ret;
+
+	ret = rv770_set_smc_sram_address(rdev, smc_address, limit);
+	if (ret)
+		return ret;
+
+	WREG32(SMC_SRAM_DATA, value);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/radeon/rv770_smc.h b/drivers/gpu/drm/radeon/rv770_smc.h
new file mode 100644
index 000000000000..f78d92a4b325
--- /dev/null
+++ b/drivers/gpu/drm/radeon/rv770_smc.h
@@ -0,0 +1,209 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __RV770_SMC_H__
+#define __RV770_SMC_H__
+
+#include "ppsmc.h"
+
+#pragma pack(push, 1)
+
+#define RV770_SMC_TABLE_ADDRESS 0xB000
+
+#define RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE    3
+
+struct RV770_SMC_SCLK_VALUE
+{
+    uint32_t        vCG_SPLL_FUNC_CNTL;
+    uint32_t        vCG_SPLL_FUNC_CNTL_2;
+    uint32_t        vCG_SPLL_FUNC_CNTL_3;
+    uint32_t        vCG_SPLL_SPREAD_SPECTRUM;
+    uint32_t        vCG_SPLL_SPREAD_SPECTRUM_2;
+    uint32_t        sclk_value;
+};
+
+typedef struct RV770_SMC_SCLK_VALUE RV770_SMC_SCLK_VALUE;
+
+struct RV770_SMC_MCLK_VALUE
+{
+    uint32_t        vMPLL_AD_FUNC_CNTL;
+    uint32_t        vMPLL_AD_FUNC_CNTL_2;
+    uint32_t        vMPLL_DQ_FUNC_CNTL;
+    uint32_t        vMPLL_DQ_FUNC_CNTL_2;
+    uint32_t        vMCLK_PWRMGT_CNTL;
+    uint32_t        vDLL_CNTL;
+    uint32_t        vMPLL_SS;
+    uint32_t        vMPLL_SS2;
+    uint32_t        mclk_value;
+};
+
+typedef struct RV770_SMC_MCLK_VALUE RV770_SMC_MCLK_VALUE;
+
+
+struct RV730_SMC_MCLK_VALUE
+{
+    uint32_t        vMCLK_PWRMGT_CNTL;
+    uint32_t        vDLL_CNTL;
+    uint32_t        vMPLL_FUNC_CNTL;
+    uint32_t        vMPLL_FUNC_CNTL2;
+    uint32_t        vMPLL_FUNC_CNTL3;
+    uint32_t        vMPLL_SS;
+    uint32_t        vMPLL_SS2;
+    uint32_t        mclk_value;
+};
+
+typedef struct RV730_SMC_MCLK_VALUE RV730_SMC_MCLK_VALUE;
+
+struct RV770_SMC_VOLTAGE_VALUE
+{
+    uint16_t             value;
+    uint8_t              index;
+    uint8_t              padding;
+};
+
+typedef struct RV770_SMC_VOLTAGE_VALUE RV770_SMC_VOLTAGE_VALUE;
+
+union RV7XX_SMC_MCLK_VALUE
+{
+    RV770_SMC_MCLK_VALUE    mclk770;
+    RV730_SMC_MCLK_VALUE    mclk730;
+};
+
+typedef union RV7XX_SMC_MCLK_VALUE RV7XX_SMC_MCLK_VALUE, *LPRV7XX_SMC_MCLK_VALUE;
+
+struct RV770_SMC_HW_PERFORMANCE_LEVEL
+{
+    uint8_t                 arbValue;
+    union{
+        uint8_t             seqValue;
+        uint8_t             ACIndex;
+    };
+    uint8_t                 displayWatermark;
+    uint8_t                 gen2PCIE;
+    uint8_t                 gen2XSP;
+    uint8_t                 backbias;
+    uint8_t                 strobeMode;
+    uint8_t                 mcFlags;
+    uint32_t                aT;
+    uint32_t                bSP;
+    RV770_SMC_SCLK_VALUE    sclk;
+    RV7XX_SMC_MCLK_VALUE    mclk;
+    RV770_SMC_VOLTAGE_VALUE vddc;
+    RV770_SMC_VOLTAGE_VALUE mvdd;
+    RV770_SMC_VOLTAGE_VALUE vddci;
+    uint8_t                 reserved1;
+    uint8_t                 reserved2;
+    uint8_t                 stateFlags;
+    uint8_t                 padding;
+};
+
+#define SMC_STROBE_RATIO    0x0F
+#define SMC_STROBE_ENABLE   0x10
+
+#define SMC_MC_EDC_RD_FLAG  0x01
+#define SMC_MC_EDC_WR_FLAG  0x02
+#define SMC_MC_RTT_ENABLE   0x04
+#define SMC_MC_STUTTER_EN   0x08
+
+typedef struct RV770_SMC_HW_PERFORMANCE_LEVEL RV770_SMC_HW_PERFORMANCE_LEVEL;
+
+struct RV770_SMC_SWSTATE
+{
+    uint8_t           flags;
+    uint8_t           padding1;
+    uint8_t           padding2;
+    uint8_t           padding3;
+    RV770_SMC_HW_PERFORMANCE_LEVEL levels[RV770_SMC_PERFORMANCE_LEVELS_PER_SWSTATE];
+};
+
+typedef struct RV770_SMC_SWSTATE RV770_SMC_SWSTATE;
+
+#define RV770_SMC_VOLTAGEMASK_VDDC 0
+#define RV770_SMC_VOLTAGEMASK_MVDD 1
+#define RV770_SMC_VOLTAGEMASK_VDDCI 2
+#define RV770_SMC_VOLTAGEMASK_MAX  4
+
+struct RV770_SMC_VOLTAGEMASKTABLE
+{
+    uint8_t  highMask[RV770_SMC_VOLTAGEMASK_MAX];
+    uint32_t lowMask[RV770_SMC_VOLTAGEMASK_MAX];
+};
+
+typedef struct RV770_SMC_VOLTAGEMASKTABLE RV770_SMC_VOLTAGEMASKTABLE;
+
+#define MAX_NO_VREG_STEPS 32
+
+struct RV770_SMC_STATETABLE
+{
+    uint8_t             thermalProtectType;
+    uint8_t             systemFlags;
+    uint8_t             maxVDDCIndexInPPTable;
+    uint8_t             extraFlags;
+    uint8_t             highSMIO[MAX_NO_VREG_STEPS];
+    uint32_t            lowSMIO[MAX_NO_VREG_STEPS];
+    RV770_SMC_VOLTAGEMASKTABLE voltageMaskTable;
+    RV770_SMC_SWSTATE   initialState;
+    RV770_SMC_SWSTATE   ACPIState;
+    RV770_SMC_SWSTATE   driverState;
+    RV770_SMC_SWSTATE   ULVState;
+};
+
+typedef struct RV770_SMC_STATETABLE RV770_SMC_STATETABLE;
+
+#define PPSMC_STATEFLAG_AUTO_PULSE_SKIP 0x01
+
+#pragma pack(pop)
+
+#define RV770_SMC_SOFT_REGISTERS_START        0x104
+
+#define RV770_SMC_SOFT_REGISTER_mclk_chg_timeout        0x0
+#define RV770_SMC_SOFT_REGISTER_baby_step_timer         0x8
+#define RV770_SMC_SOFT_REGISTER_delay_bbias             0xC
+#define RV770_SMC_SOFT_REGISTER_delay_vreg              0x10
+#define RV770_SMC_SOFT_REGISTER_delay_acpi              0x2C
+#define RV770_SMC_SOFT_REGISTER_seq_index               0x64
+#define RV770_SMC_SOFT_REGISTER_mvdd_chg_time           0x68
+#define RV770_SMC_SOFT_REGISTER_mclk_switch_lim         0x78
+#define RV770_SMC_SOFT_REGISTER_mc_block_delay          0x90
+#define RV770_SMC_SOFT_REGISTER_uvd_enabled             0x9C
+#define RV770_SMC_SOFT_REGISTER_is_asic_lombok          0xA0
+
+int rv770_set_smc_sram_address(struct radeon_device *rdev,
+			       u16 smc_address, u16 limit);
+int rv770_copy_bytes_to_smc(struct radeon_device *rdev,
+			    u16 smc_start_address, const u8 *src,
+			    u16 byte_count, u16 limit);
+void rv770_start_smc(struct radeon_device *rdev);
+void rv770_reset_smc(struct radeon_device *rdev);
+void rv770_stop_smc_clock(struct radeon_device *rdev);
+void rv770_start_smc_clock(struct radeon_device *rdev);
+bool rv770_is_smc_running(struct radeon_device *rdev);
+PPSMC_Result rv770_send_msg_to_smc(struct radeon_device *rdev, PPSMC_Msg msg);
+PPSMC_Result rv770_wait_for_smc_inactive(struct radeon_device *rdev);
+int rv770_read_smc_sram_dword(struct radeon_device *rdev,
+			      u16 smc_address, u32 *value, u16 limit);
+int rv770_write_smc_sram_dword(struct radeon_device *rdev,
+			       u16 smc_address, u32 value, u16 limit);
+int rv770_load_smc_ucode(struct radeon_device *rdev,
+			 u16 limit);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/rv770d.h b/drivers/gpu/drm/radeon/rv770d.h
index 85b16266f748..6bef2b7d601b 100644
--- a/drivers/gpu/drm/radeon/rv770d.h
+++ b/drivers/gpu/drm/radeon/rv770d.h
@@ -62,6 +62,246 @@
 #	define UPLL_FB_DIV(x)				((x) << 0)
 #	define UPLL_FB_DIV_MASK				0x01FFFFFF
 
+/* pm registers */
+#define	SMC_SRAM_ADDR					0x200
+#define		SMC_SRAM_AUTO_INC_DIS				(1 << 16)
+#define	SMC_SRAM_DATA					0x204
+#define	SMC_IO						0x208
+#define		SMC_RST_N					(1 << 0)
+#define		SMC_STOP_MODE					(1 << 2)
+#define		SMC_CLK_EN					(1 << 11)
+#define	SMC_MSG						0x20c
+#define		HOST_SMC_MSG(x)					((x) << 0)
+#define		HOST_SMC_MSG_MASK				(0xff << 0)
+#define		HOST_SMC_MSG_SHIFT				0
+#define		HOST_SMC_RESP(x)				((x) << 8)
+#define		HOST_SMC_RESP_MASK				(0xff << 8)
+#define		HOST_SMC_RESP_SHIFT				8
+#define		SMC_HOST_MSG(x)					((x) << 16)
+#define		SMC_HOST_MSG_MASK				(0xff << 16)
+#define		SMC_HOST_MSG_SHIFT				16
+#define		SMC_HOST_RESP(x)				((x) << 24)
+#define		SMC_HOST_RESP_MASK				(0xff << 24)
+#define		SMC_HOST_RESP_SHIFT				24
+
+#define	SMC_ISR_FFD8_FFDB				0x218
+
+#define	CG_SPLL_FUNC_CNTL				0x600
+#define		SPLL_RESET				(1 << 0)
+#define		SPLL_SLEEP				(1 << 1)
+#define		SPLL_DIVEN				(1 << 2)
+#define		SPLL_BYPASS_EN				(1 << 3)
+#define		SPLL_REF_DIV(x)				((x) << 4)
+#define		SPLL_REF_DIV_MASK			(0x3f << 4)
+#define		SPLL_HILEN(x)				((x) << 12)
+#define		SPLL_HILEN_MASK				(0xf << 12)
+#define		SPLL_LOLEN(x)				((x) << 16)
+#define		SPLL_LOLEN_MASK				(0xf << 16)
+#define	CG_SPLL_FUNC_CNTL_2				0x604
+#define		SCLK_MUX_SEL(x)				((x) << 0)
+#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_SPLL_FUNC_CNTL_3				0x608
+#define		SPLL_FB_DIV(x)				((x) << 0)
+#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		SPLL_DITHEN				(1 << 28)
+
+#define	SPLL_CNTL_MODE					0x610
+#define		SPLL_DIV_SYNC				(1 << 5)
+
+#define	MPLL_AD_FUNC_CNTL				0x624
+#define		CLKF(x)					((x) << 0)
+#define		CLKF_MASK				(0x7f << 0)
+#define		CLKR(x)					((x) << 7)
+#define		CLKR_MASK				(0x1f << 7)
+#define		CLKFRAC(x)				((x) << 12)
+#define		CLKFRAC_MASK				(0x1f << 12)
+#define		YCLK_POST_DIV(x)			((x) << 17)
+#define		YCLK_POST_DIV_MASK			(3 << 17)
+#define		IBIAS(x)				((x) << 20)
+#define		IBIAS_MASK				(0x3ff << 20)
+#define		RESET					(1 << 30)
+#define		PDNB					(1 << 31)
+#define	MPLL_AD_FUNC_CNTL_2				0x628
+#define		BYPASS					(1 << 19)
+#define		BIAS_GEN_PDNB				(1 << 24)
+#define		RESET_EN				(1 << 25)
+#define		VCO_MODE				(1 << 29)
+#define	MPLL_DQ_FUNC_CNTL				0x62c
+#define	MPLL_DQ_FUNC_CNTL_2				0x630
+
+#define GENERAL_PWRMGT                                  0x63c
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+#       define STATIC_PM_EN                             (1 << 1)
+#       define THERMAL_PROTECTION_DIS                   (1 << 2)
+#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
+#       define ENABLE_GEN2PCIE                          (1 << 4)
+#       define ENABLE_GEN2XSP                           (1 << 5)
+#       define SW_SMIO_INDEX(x)                         ((x) << 6)
+#       define SW_SMIO_INDEX_MASK                       (3 << 6)
+#       define SW_SMIO_INDEX_SHIFT                      6
+#       define LOW_VOLT_D2_ACPI                         (1 << 8)
+#       define LOW_VOLT_D3_ACPI                         (1 << 9)
+#       define VOLT_PWRMGT_EN                           (1 << 10)
+#       define BACKBIAS_PAD_EN                          (1 << 18)
+#       define BACKBIAS_VALUE                           (1 << 19)
+#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
+#       define AC_DC_SW                                 (1 << 24)
+
+#define CG_TPC                                            0x640
+#define SCLK_PWRMGT_CNTL                                  0x644
+#       define SCLK_PWRMGT_OFF                            (1 << 0)
+#       define SCLK_LOW_D1                                (1 << 1)
+#       define FIR_RESET                                  (1 << 4)
+#       define FIR_FORCE_TREND_SEL                        (1 << 5)
+#       define FIR_TREND_MODE                             (1 << 6)
+#       define DYN_GFX_CLK_OFF_EN                         (1 << 7)
+#       define GFX_CLK_FORCE_ON                           (1 << 8)
+#       define GFX_CLK_REQUEST_OFF                        (1 << 9)
+#       define GFX_CLK_FORCE_OFF                          (1 << 10)
+#       define GFX_CLK_OFF_ACPI_D1                        (1 << 11)
+#       define GFX_CLK_OFF_ACPI_D2                        (1 << 12)
+#       define GFX_CLK_OFF_ACPI_D3                        (1 << 13)
+#define	MCLK_PWRMGT_CNTL				0x648
+#       define DLL_SPEED(x)				((x) << 0)
+#       define DLL_SPEED_MASK				(0x1f << 0)
+#       define MPLL_PWRMGT_OFF                          (1 << 5)
+#       define DLL_READY                                (1 << 6)
+#       define MC_INT_CNTL                              (1 << 7)
+#       define MRDCKA0_SLEEP                            (1 << 8)
+#       define MRDCKA1_SLEEP                            (1 << 9)
+#       define MRDCKB0_SLEEP                            (1 << 10)
+#       define MRDCKB1_SLEEP                            (1 << 11)
+#       define MRDCKC0_SLEEP                            (1 << 12)
+#       define MRDCKC1_SLEEP                            (1 << 13)
+#       define MRDCKD0_SLEEP                            (1 << 14)
+#       define MRDCKD1_SLEEP                            (1 << 15)
+#       define MRDCKA0_RESET                            (1 << 16)
+#       define MRDCKA1_RESET                            (1 << 17)
+#       define MRDCKB0_RESET                            (1 << 18)
+#       define MRDCKB1_RESET                            (1 << 19)
+#       define MRDCKC0_RESET                            (1 << 20)
+#       define MRDCKC1_RESET                            (1 << 21)
+#       define MRDCKD0_RESET                            (1 << 22)
+#       define MRDCKD1_RESET                            (1 << 23)
+#       define DLL_READY_READ                           (1 << 24)
+#       define USE_DISPLAY_GAP                          (1 << 25)
+#       define USE_DISPLAY_URGENT_NORMAL                (1 << 26)
+#       define MPLL_TURNOFF_D2                          (1 << 28)
+#define	DLL_CNTL					0x64c
+#       define MRDCKA0_BYPASS                           (1 << 24)
+#       define MRDCKA1_BYPASS                           (1 << 25)
+#       define MRDCKB0_BYPASS                           (1 << 26)
+#       define MRDCKB1_BYPASS                           (1 << 27)
+#       define MRDCKC0_BYPASS                           (1 << 28)
+#       define MRDCKC1_BYPASS                           (1 << 29)
+#       define MRDCKD0_BYPASS                           (1 << 30)
+#       define MRDCKD1_BYPASS                           (1 << 31)
+
+#define MPLL_TIME                                         0x654
+#       define MPLL_LOCK_TIME(x)			((x) << 0)
+#       define MPLL_LOCK_TIME_MASK			(0xffff << 0)
+#       define MPLL_RESET_TIME(x)			((x) << 16)
+#       define MPLL_RESET_TIME_MASK			(0xffff << 16)
+
+#define CG_CLKPIN_CNTL                                    0x660
+#       define MUX_TCLK_TO_XCLK                           (1 << 8)
+#       define XTALIN_DIVIDE                              (1 << 9)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                  0x66c
+#       define CURRENT_PROFILE_INDEX_MASK                 (0xf << 4)
+#       define CURRENT_PROFILE_INDEX_SHIFT                4
+
+#define S0_VID_LOWER_SMIO_CNTL                            0x678
+#define S1_VID_LOWER_SMIO_CNTL                            0x67c
+#define S2_VID_LOWER_SMIO_CNTL                            0x680
+#define S3_VID_LOWER_SMIO_CNTL                            0x684
+
+#define CG_FTV                                            0x690
+#define CG_FFCT_0                                         0x694
+#       define UTC_0(x)                                   ((x) << 0)
+#       define UTC_0_MASK                                 (0x3ff << 0)
+#       define DTC_0(x)                                   ((x) << 10)
+#       define DTC_0_MASK                                 (0x3ff << 10)
+
+#define CG_BSP                                          0x6d0
+#       define BSP(x)					((x) << 0)
+#       define BSP_MASK					(0xffff << 0)
+#       define BSU(x)					((x) << 16)
+#       define BSU_MASK					(0xf << 16)
+#define CG_AT                                           0x6d4
+#       define CG_R(x)					((x) << 0)
+#       define CG_R_MASK				(0xffff << 0)
+#       define CG_L(x)					((x) << 16)
+#       define CG_L_MASK				(0xffff << 16)
+#define CG_GIT                                          0x6d8
+#       define CG_GICST(x)                              ((x) << 0)
+#       define CG_GICST_MASK                            (0xffff << 0)
+#       define CG_GIPOT(x)                              ((x) << 16)
+#       define CG_GIPOT_MASK                            (0xffff << 16)
+
+#define CG_SSP                                            0x6e8
+#       define SST(x)                                     ((x) << 0)
+#       define SST_MASK                                   (0xffff << 0)
+#       define SSTU(x)                                    ((x) << 16)
+#       define SSTU_MASK                                  (0xf << 16)
+
+#define CG_DISPLAY_GAP_CNTL                               0x714
+#       define DISP1_GAP(x)                               ((x) << 0)
+#       define DISP1_GAP_MASK                             (3 << 0)
+#       define DISP2_GAP(x)                               ((x) << 2)
+#       define DISP2_GAP_MASK                             (3 << 2)
+#       define VBI_TIMER_COUNT(x)                         ((x) << 4)
+#       define VBI_TIMER_COUNT_MASK                       (0x3fff << 4)
+#       define VBI_TIMER_UNIT(x)                          ((x) << 20)
+#       define VBI_TIMER_UNIT_MASK                        (7 << 20)
+#       define DISP1_GAP_MCHG(x)                          ((x) << 24)
+#       define DISP1_GAP_MCHG_MASK                        (3 << 24)
+#       define DISP2_GAP_MCHG(x)                          ((x) << 26)
+#       define DISP2_GAP_MCHG_MASK                        (3 << 26)
+
+#define	CG_SPLL_SPREAD_SPECTRUM				0x790
+#define		SSEN					(1 << 0)
+#define		CLKS(x)					((x) << 4)
+#define		CLKS_MASK				(0xfff << 4)
+#define	CG_SPLL_SPREAD_SPECTRUM_2			0x794
+#define		CLKV(x)					((x) << 0)
+#define		CLKV_MASK				(0x3ffffff << 0)
+#define	CG_MPLL_SPREAD_SPECTRUM				0x798
+#define CG_UPLL_SPREAD_SPECTRUM				0x79c
+#	define SSEN_MASK				0x00000001
+
+#define CG_CGTT_LOCAL_0                                   0x7d0
+#define CG_CGTT_LOCAL_1                                   0x7d4
+
+#define BIOS_SCRATCH_4                                    0x1734
+
+#define MC_SEQ_MISC0                                      0x2a00
+#define         MC_SEQ_MISC0_GDDR5_SHIFT                  28
+#define         MC_SEQ_MISC0_GDDR5_MASK                   0xf0000000
+#define         MC_SEQ_MISC0_GDDR5_VALUE                  5
+
+#define MC_ARB_SQM_RATIO                                  0x2770
+#define		STATE0(x)				((x) << 0)
+#define		STATE0_MASK				(0xff << 0)
+#define		STATE1(x)				((x) << 8)
+#define		STATE1_MASK				(0xff << 8)
+#define		STATE2(x)				((x) << 16)
+#define		STATE2_MASK				(0xff << 16)
+#define		STATE3(x)				((x) << 24)
+#define		STATE3_MASK				(0xff << 24)
+
+#define	MC_ARB_RFSH_RATE				0x27b0
+#define		POWERMODE0(x)				((x) << 0)
+#define		POWERMODE0_MASK				(0xff << 0)
+#define		POWERMODE1(x)				((x) << 8)
+#define		POWERMODE1_MASK				(0xff << 8)
+#define		POWERMODE2(x)				((x) << 16)
+#define		POWERMODE2_MASK				(0xff << 16)
+#define		POWERMODE3(x)				((x) << 24)
+#define		POWERMODE3_MASK				(0xff << 24)
+
+#define CGTS_SM_CTRL_REG                                  0x9150
+
 /* Registers */
 #define	CB_COLOR0_BASE					0x28040
 #define	CB_COLOR1_BASE					0x28044
@@ -86,8 +326,8 @@
 #define	CONFIG_MEMSIZE					0x5428
 
 #define	CP_ME_CNTL					0x86D8
-#define		CP_ME_HALT					(1<<28)
-#define		CP_PFP_HALT					(1<<26)
+#define		CP_ME_HALT					(1 << 28)
+#define		CP_PFP_HALT					(1 << 26)
 #define	CP_ME_RAM_DATA					0xC160
 #define	CP_ME_RAM_RADDR					0xC158
 #define	CP_ME_RAM_WADDR					0xC15C
@@ -157,9 +397,22 @@
 #define		GUI_ACTIVE					(1<<31)
 #define	GRBM_STATUS2					0x8014
 
-#define CG_CLKPIN_CNTL                                    0x660
-#       define MUX_TCLK_TO_XCLK                           (1 << 8)
-#       define XTALIN_DIVIDE                              (1 << 9)
+#define	CG_THERMAL_CTRL					0x72C
+#define 	DPM_EVENT_SRC(x)			((x) << 0)
+#define 	DPM_EVENT_SRC_MASK			(7 << 0)
+#define		DIG_THERM_DPM(x)			((x) << 14)
+#define		DIG_THERM_DPM_MASK			0x003FC000
+#define		DIG_THERM_DPM_SHIFT			14
+
+#define	CG_THERMAL_INT					0x734
+#define		DIG_THERM_INTH(x)			((x) << 8)
+#define		DIG_THERM_INTH_MASK			0x0000FF00
+#define		DIG_THERM_INTH_SHIFT			8
+#define		DIG_THERM_INTL(x)			((x) << 16)
+#define		DIG_THERM_INTL_MASK			0x00FF0000
+#define		DIG_THERM_INTL_SHIFT			16
+#define 	THERM_INT_MASK_HIGH			(1 << 24)
+#define 	THERM_INT_MASK_LOW			(1 << 25)
 
 #define	CG_MULT_THERMAL_STATUS				0x740
 #define		ASIC_T(x)			        ((x) << 16)
@@ -662,7 +915,22 @@
 #define D1GRPH_SECONDARY_SURFACE_ADDRESS_HIGH             0x691c
 #define D2GRPH_SECONDARY_SURFACE_ADDRESS_HIGH             0x611c
 
-/* PCIE link stuff */
+/* PCIE indirect regs */
+#define PCIE_P_CNTL                                       0x40
+#       define P_PLL_PWRDN_IN_L1L23                       (1 << 3)
+#       define P_PLL_BUF_PDNB                             (1 << 4)
+#       define P_PLL_PDNB                                 (1 << 9)
+#       define P_ALLOW_PRX_FRONTEND_SHUTOFF               (1 << 12)
+/* PCIE PORT regs */
+#define PCIE_LC_CNTL                                      0xa0
+#       define LC_L0S_INACTIVITY(x)                       ((x) << 8)
+#       define LC_L0S_INACTIVITY_MASK                     (0xf << 8)
+#       define LC_L0S_INACTIVITY_SHIFT                    8
+#       define LC_L1_INACTIVITY(x)                        ((x) << 12)
+#       define LC_L1_INACTIVITY_MASK                      (0xf << 12)
+#       define LC_L1_INACTIVITY_SHIFT                     12
+#       define LC_PMI_TO_L1_DIS                           (1 << 16)
+#       define LC_ASPM_TO_L1_DIS                          (1 << 24)
 #define PCIE_LC_TRAINING_CNTL                             0xa1 /* PCIE_P */
 #define PCIE_LC_LINK_WIDTH_CNTL                           0xa2 /* PCIE_P */
 #       define LC_LINK_WIDTH_SHIFT                        0
@@ -690,6 +958,9 @@
 #       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK      (0x3 << 8)
 #       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT     3
 #       define LC_CURRENT_DATA_RATE                       (1 << 11)
+#       define LC_HW_VOLTAGE_IF_CONTROL(x)                ((x) << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_MASK              (3 << 12)
+#       define LC_HW_VOLTAGE_IF_CONTROL_SHIFT             12
 #       define LC_VOLTAGE_TIMER_SEL_MASK                  (0xf << 14)
 #       define LC_CLR_FAILED_SPD_CHANGE_CNT               (1 << 21)
 #       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 23)
diff --git a/drivers/gpu/drm/radeon/si.c b/drivers/gpu/drm/radeon/si.c
index a1b0da6b5808..d325280e2f9f 100644
--- a/drivers/gpu/drm/radeon/si.c
+++ b/drivers/gpu/drm/radeon/si.c
@@ -22,7 +22,6 @@
  * Authors: Alex Deucher
  */
 #include <linux/firmware.h>
-#include <linux/platform_device.h>
 #include <linux/slab.h>
 #include <linux/module.h>
 #include <drm/drmP.h>
@@ -32,40 +31,43 @@
 #include "sid.h"
 #include "atom.h"
 #include "si_blit_shaders.h"
+#include "clearstate_si.h"
+#include "radeon_ucode.h"
 
-#define SI_PFP_UCODE_SIZE 2144
-#define SI_PM4_UCODE_SIZE 2144
-#define SI_CE_UCODE_SIZE 2144
-#define SI_RLC_UCODE_SIZE 2048
-#define SI_MC_UCODE_SIZE 7769
-#define OLAND_MC_UCODE_SIZE 7863
 
 MODULE_FIRMWARE("radeon/TAHITI_pfp.bin");
 MODULE_FIRMWARE("radeon/TAHITI_me.bin");
 MODULE_FIRMWARE("radeon/TAHITI_ce.bin");
 MODULE_FIRMWARE("radeon/TAHITI_mc.bin");
 MODULE_FIRMWARE("radeon/TAHITI_rlc.bin");
+MODULE_FIRMWARE("radeon/TAHITI_smc.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_pfp.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_me.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_ce.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_mc.bin");
 MODULE_FIRMWARE("radeon/PITCAIRN_rlc.bin");
+MODULE_FIRMWARE("radeon/PITCAIRN_smc.bin");
 MODULE_FIRMWARE("radeon/VERDE_pfp.bin");
 MODULE_FIRMWARE("radeon/VERDE_me.bin");
 MODULE_FIRMWARE("radeon/VERDE_ce.bin");
 MODULE_FIRMWARE("radeon/VERDE_mc.bin");
 MODULE_FIRMWARE("radeon/VERDE_rlc.bin");
+MODULE_FIRMWARE("radeon/VERDE_smc.bin");
 MODULE_FIRMWARE("radeon/OLAND_pfp.bin");
 MODULE_FIRMWARE("radeon/OLAND_me.bin");
 MODULE_FIRMWARE("radeon/OLAND_ce.bin");
 MODULE_FIRMWARE("radeon/OLAND_mc.bin");
 MODULE_FIRMWARE("radeon/OLAND_rlc.bin");
+MODULE_FIRMWARE("radeon/OLAND_smc.bin");
 MODULE_FIRMWARE("radeon/HAINAN_pfp.bin");
 MODULE_FIRMWARE("radeon/HAINAN_me.bin");
 MODULE_FIRMWARE("radeon/HAINAN_ce.bin");
 MODULE_FIRMWARE("radeon/HAINAN_mc.bin");
 MODULE_FIRMWARE("radeon/HAINAN_rlc.bin");
+MODULE_FIRMWARE("radeon/HAINAN_smc.bin");
 
+static void si_pcie_gen3_enable(struct radeon_device *rdev);
+static void si_program_aspm(struct radeon_device *rdev);
 extern int r600_ih_ring_alloc(struct radeon_device *rdev);
 extern void r600_ih_ring_fini(struct radeon_device *rdev);
 extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev);
@@ -75,6 +77,228 @@ extern u32 evergreen_get_number_of_dram_channels(struct radeon_device *rdev);
 extern void evergreen_print_gpu_status_regs(struct radeon_device *rdev);
 extern bool evergreen_is_display_hung(struct radeon_device *rdev);
 
+static const u32 verde_rlc_save_restore_register_list[] =
+{
+	(0x8000 << 16) | (0x98f4 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x98f4 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0xe80 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0xe80 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x89bc >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x89bc >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x8c1c >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x8c1c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x98f0 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xe7c >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9148 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9148 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9150 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x897c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8d8c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac54 >> 2),
+	0X00000000,
+	0x3,
+	(0x9c00 << 16) | (0x98f8 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9910 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9914 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9918 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x991c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9920 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9924 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9928 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x992c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9930 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9934 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9938 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x993c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9940 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9944 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9948 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x994c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9950 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9954 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9958 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x995c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9960 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9964 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9968 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x996c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9970 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9974 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9978 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x997c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9980 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9984 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9988 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x998c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c00 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c14 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c04 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c08 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9b7c >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9b7c >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0xe84 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0xe84 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x89c0 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x89c0 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x914c >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x914c >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x8c20 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x8c20 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9354 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9354 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9060 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9364 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9100 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x913c >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x90e0 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x90e4 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x90e8 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x90e0 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x90e4 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x90e8 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8bcc >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8b24 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x88c4 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8e50 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c0c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8e58 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8e5c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9508 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x950c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9494 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac0c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac10 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac14 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xae00 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac08 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x88d4 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x88c8 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x88cc >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x89b0 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8b10 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8a14 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9830 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9834 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9838 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9a10 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9870 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9874 >> 2),
+	0x00000000,
+	(0x8001 << 16) | (0x9870 >> 2),
+	0x00000000,
+	(0x8001 << 16) | (0x9874 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9870 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9874 >> 2),
+	0x00000000,
+	(0x8041 << 16) | (0x9870 >> 2),
+	0x00000000,
+	(0x8041 << 16) | (0x9874 >> 2),
+	0x00000000,
+	0x00000000
+};
+
 static const u32 tahiti_golden_rlc_registers[] =
 {
 	0xc424, 0xffffffff, 0x00601005,
@@ -1316,22 +1540,15 @@ static int si_mc_load_microcode(struct radeon_device *rdev)
 
 static int si_init_microcode(struct radeon_device *rdev)
 {
-	struct platform_device *pdev;
 	const char *chip_name;
 	const char *rlc_chip_name;
 	size_t pfp_req_size, me_req_size, ce_req_size, rlc_req_size, mc_req_size;
+	size_t smc_req_size;
 	char fw_name[30];
 	int err;
 
 	DRM_DEBUG("\n");
 
-	pdev = platform_device_register_simple("radeon_cp", 0, NULL, 0);
-	err = IS_ERR(pdev);
-	if (err) {
-		printk(KERN_ERR "radeon_cp: Failed to register firmware\n");
-		return -EINVAL;
-	}
-
 	switch (rdev->family) {
 	case CHIP_TAHITI:
 		chip_name = "TAHITI";
@@ -1341,6 +1558,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
 		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
 		mc_req_size = SI_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(TAHITI_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_PITCAIRN:
 		chip_name = "PITCAIRN";
@@ -1350,6 +1568,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
 		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
 		mc_req_size = SI_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(PITCAIRN_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_VERDE:
 		chip_name = "VERDE";
@@ -1359,6 +1578,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
 		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
 		mc_req_size = SI_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(VERDE_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_OLAND:
 		chip_name = "OLAND";
@@ -1368,6 +1588,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
 		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
 		mc_req_size = OLAND_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(OLAND_SMC_UCODE_SIZE, 4);
 		break;
 	case CHIP_HAINAN:
 		chip_name = "HAINAN";
@@ -1377,6 +1598,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 		ce_req_size = SI_CE_UCODE_SIZE * 4;
 		rlc_req_size = SI_RLC_UCODE_SIZE * 4;
 		mc_req_size = OLAND_MC_UCODE_SIZE * 4;
+		smc_req_size = ALIGN(HAINAN_SMC_UCODE_SIZE, 4);
 		break;
 	default: BUG();
 	}
@@ -1384,7 +1606,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 	DRM_INFO("Loading %s Microcode\n", chip_name);
 
 	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
-	err = request_firmware(&rdev->pfp_fw, fw_name, &pdev->dev);
+	err = request_firmware(&rdev->pfp_fw, fw_name, rdev->dev);
 	if (err)
 		goto out;
 	if (rdev->pfp_fw->size != pfp_req_size) {
@@ -1396,7 +1618,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 	}
 
 	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
-	err = request_firmware(&rdev->me_fw, fw_name, &pdev->dev);
+	err = request_firmware(&rdev->me_fw, fw_name, rdev->dev);
 	if (err)
 		goto out;
 	if (rdev->me_fw->size != me_req_size) {
@@ -1407,7 +1629,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 	}
 
 	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
-	err = request_firmware(&rdev->ce_fw, fw_name, &pdev->dev);
+	err = request_firmware(&rdev->ce_fw, fw_name, rdev->dev);
 	if (err)
 		goto out;
 	if (rdev->ce_fw->size != ce_req_size) {
@@ -1418,7 +1640,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 	}
 
 	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", rlc_chip_name);
-	err = request_firmware(&rdev->rlc_fw, fw_name, &pdev->dev);
+	err = request_firmware(&rdev->rlc_fw, fw_name, rdev->dev);
 	if (err)
 		goto out;
 	if (rdev->rlc_fw->size != rlc_req_size) {
@@ -1429,7 +1651,7 @@ static int si_init_microcode(struct radeon_device *rdev)
 	}
 
 	snprintf(fw_name, sizeof(fw_name), "radeon/%s_mc.bin", chip_name);
-	err = request_firmware(&rdev->mc_fw, fw_name, &pdev->dev);
+	err = request_firmware(&rdev->mc_fw, fw_name, rdev->dev);
 	if (err)
 		goto out;
 	if (rdev->mc_fw->size != mc_req_size) {
@@ -1439,9 +1661,18 @@ static int si_init_microcode(struct radeon_device *rdev)
 		err = -EINVAL;
 	}
 
-out:
-	platform_device_unregister(pdev);
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_smc.bin", chip_name);
+	err = request_firmware(&rdev->smc_fw, fw_name, rdev->dev);
+	if (err)
+		goto out;
+	if (rdev->smc_fw->size != smc_req_size) {
+		printk(KERN_ERR
+		       "si_smc: Bogus length %zu in firmware \"%s\"\n",
+		       rdev->smc_fw->size, fw_name);
+		err = -EINVAL;
+	}
 
+out:
 	if (err) {
 		if (err != -EINVAL)
 			printk(KERN_ERR
@@ -1457,6 +1688,8 @@ out:
 		rdev->rlc_fw = NULL;
 		release_firmware(rdev->mc_fw);
 		rdev->mc_fw = NULL;
+		release_firmware(rdev->smc_fw);
+		rdev->smc_fw = NULL;
 	}
 	return err;
 }
@@ -1792,7 +2025,8 @@ static void dce6_program_watermarks(struct radeon_device *rdev,
 					 u32 lb_size, u32 num_heads)
 {
 	struct drm_display_mode *mode = &radeon_crtc->base.mode;
-	struct dce6_wm_params wm;
+	struct dce6_wm_params wm_low, wm_high;
+	u32 dram_channels;
 	u32 pixel_period;
 	u32 line_time = 0;
 	u32 latency_watermark_a = 0, latency_watermark_b = 0;
@@ -1808,38 +2042,83 @@ static void dce6_program_watermarks(struct radeon_device *rdev,
 		priority_a_cnt = 0;
 		priority_b_cnt = 0;
 
-		wm.yclk = rdev->pm.current_mclk * 10;
-		wm.sclk = rdev->pm.current_sclk * 10;
-		wm.disp_clk = mode->clock;
-		wm.src_width = mode->crtc_hdisplay;
-		wm.active_time = mode->crtc_hdisplay * pixel_period;
-		wm.blank_time = line_time - wm.active_time;
-		wm.interlaced = false;
-		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-			wm.interlaced = true;
-		wm.vsc = radeon_crtc->vsc;
-		wm.vtaps = 1;
-		if (radeon_crtc->rmx_type != RMX_OFF)
-			wm.vtaps = 2;
-		wm.bytes_per_pixel = 4; /* XXX: get this from fb config */
-		wm.lb_size = lb_size;
 		if (rdev->family == CHIP_ARUBA)
-			wm.dram_channels = evergreen_get_number_of_dram_channels(rdev);
+			dram_channels = evergreen_get_number_of_dram_channels(rdev);
 		else
-			wm.dram_channels = si_get_number_of_dram_channels(rdev);
-		wm.num_heads = num_heads;
+			dram_channels = si_get_number_of_dram_channels(rdev);
+
+		/* watermark for high clocks */
+		if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+			wm_high.yclk =
+				radeon_dpm_get_mclk(rdev, false) * 10;
+			wm_high.sclk =
+				radeon_dpm_get_sclk(rdev, false) * 10;
+		} else {
+			wm_high.yclk = rdev->pm.current_mclk * 10;
+			wm_high.sclk = rdev->pm.current_sclk * 10;
+		}
+
+		wm_high.disp_clk = mode->clock;
+		wm_high.src_width = mode->crtc_hdisplay;
+		wm_high.active_time = mode->crtc_hdisplay * pixel_period;
+		wm_high.blank_time = line_time - wm_high.active_time;
+		wm_high.interlaced = false;
+		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+			wm_high.interlaced = true;
+		wm_high.vsc = radeon_crtc->vsc;
+		wm_high.vtaps = 1;
+		if (radeon_crtc->rmx_type != RMX_OFF)
+			wm_high.vtaps = 2;
+		wm_high.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm_high.lb_size = lb_size;
+		wm_high.dram_channels = dram_channels;
+		wm_high.num_heads = num_heads;
+
+		/* watermark for low clocks */
+		if ((rdev->pm.pm_method == PM_METHOD_DPM) && rdev->pm.dpm_enabled) {
+			wm_low.yclk =
+				radeon_dpm_get_mclk(rdev, true) * 10;
+			wm_low.sclk =
+				radeon_dpm_get_sclk(rdev, true) * 10;
+		} else {
+			wm_low.yclk = rdev->pm.current_mclk * 10;
+			wm_low.sclk = rdev->pm.current_sclk * 10;
+		}
+
+		wm_low.disp_clk = mode->clock;
+		wm_low.src_width = mode->crtc_hdisplay;
+		wm_low.active_time = mode->crtc_hdisplay * pixel_period;
+		wm_low.blank_time = line_time - wm_low.active_time;
+		wm_low.interlaced = false;
+		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+			wm_low.interlaced = true;
+		wm_low.vsc = radeon_crtc->vsc;
+		wm_low.vtaps = 1;
+		if (radeon_crtc->rmx_type != RMX_OFF)
+			wm_low.vtaps = 2;
+		wm_low.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm_low.lb_size = lb_size;
+		wm_low.dram_channels = dram_channels;
+		wm_low.num_heads = num_heads;
 
 		/* set for high clocks */
-		latency_watermark_a = min(dce6_latency_watermark(&wm), (u32)65535);
+		latency_watermark_a = min(dce6_latency_watermark(&wm_high), (u32)65535);
 		/* set for low clocks */
-		/* wm.yclk = low clk; wm.sclk = low clk */
-		latency_watermark_b = min(dce6_latency_watermark(&wm), (u32)65535);
+		latency_watermark_b = min(dce6_latency_watermark(&wm_low), (u32)65535);
 
 		/* possibly force display priority to high */
 		/* should really do this at mode validation time... */
-		if (!dce6_average_bandwidth_vs_dram_bandwidth_for_display(&wm) ||
-		    !dce6_average_bandwidth_vs_available_bandwidth(&wm) ||
-		    !dce6_check_latency_hiding(&wm) ||
+		if (!dce6_average_bandwidth_vs_dram_bandwidth_for_display(&wm_high) ||
+		    !dce6_average_bandwidth_vs_available_bandwidth(&wm_high) ||
+		    !dce6_check_latency_hiding(&wm_high) ||
+		    (rdev->disp_priority == 2)) {
+			DRM_DEBUG_KMS("force priority to high\n");
+			priority_a_cnt |= PRIORITY_ALWAYS_ON;
+			priority_b_cnt |= PRIORITY_ALWAYS_ON;
+		}
+		if (!dce6_average_bandwidth_vs_dram_bandwidth_for_display(&wm_low) ||
+		    !dce6_average_bandwidth_vs_available_bandwidth(&wm_low) ||
+		    !dce6_check_latency_hiding(&wm_low) ||
 		    (rdev->disp_priority == 2)) {
 			DRM_DEBUG_KMS("force priority to high\n");
 			priority_a_cnt |= PRIORITY_ALWAYS_ON;
@@ -1895,6 +2174,10 @@ static void dce6_program_watermarks(struct radeon_device *rdev,
 	WREG32(PRIORITY_A_CNT + radeon_crtc->crtc_offset, priority_a_cnt);
 	WREG32(PRIORITY_B_CNT + radeon_crtc->crtc_offset, priority_b_cnt);
 
+	/* save values for DPM */
+	radeon_crtc->line_time = line_time;
+	radeon_crtc->wm_high = latency_watermark_a;
+	radeon_crtc->wm_low = latency_watermark_b;
 }
 
 void dce6_bandwidth_update(struct radeon_device *rdev)
@@ -3535,8 +3818,8 @@ static void si_mc_program(struct radeon_device *rdev)
 	}
 }
 
-static void si_vram_gtt_location(struct radeon_device *rdev,
-				 struct radeon_mc *mc)
+void si_vram_gtt_location(struct radeon_device *rdev,
+			  struct radeon_mc *mc)
 {
 	if (mc->mc_vram_size > 0xFFC0000000ULL) {
 		/* leave room for at least 1024M GTT */
@@ -4107,6 +4390,270 @@ void si_vm_fini(struct radeon_device *rdev)
 }
 
 /**
+ * si_vm_decode_fault - print human readable fault info
+ *
+ * @rdev: radeon_device pointer
+ * @status: VM_CONTEXT1_PROTECTION_FAULT_STATUS register value
+ * @addr: VM_CONTEXT1_PROTECTION_FAULT_ADDR register value
+ *
+ * Print human readable fault information (SI).
+ */
+static void si_vm_decode_fault(struct radeon_device *rdev,
+			       u32 status, u32 addr)
+{
+	u32 mc_id = (status & MEMORY_CLIENT_ID_MASK) >> MEMORY_CLIENT_ID_SHIFT;
+	u32 vmid = (status & FAULT_VMID_MASK) >> FAULT_VMID_SHIFT;
+	u32 protections = (status & PROTECTIONS_MASK) >> PROTECTIONS_SHIFT;
+	char *block;
+
+	if (rdev->family == CHIP_TAHITI) {
+		switch (mc_id) {
+		case 160:
+		case 144:
+		case 96:
+		case 80:
+		case 224:
+		case 208:
+		case 32:
+		case 16:
+			block = "CB";
+			break;
+		case 161:
+		case 145:
+		case 97:
+		case 81:
+		case 225:
+		case 209:
+		case 33:
+		case 17:
+			block = "CB_FMASK";
+			break;
+		case 162:
+		case 146:
+		case 98:
+		case 82:
+		case 226:
+		case 210:
+		case 34:
+		case 18:
+			block = "CB_CMASK";
+			break;
+		case 163:
+		case 147:
+		case 99:
+		case 83:
+		case 227:
+		case 211:
+		case 35:
+		case 19:
+			block = "CB_IMMED";
+			break;
+		case 164:
+		case 148:
+		case 100:
+		case 84:
+		case 228:
+		case 212:
+		case 36:
+		case 20:
+			block = "DB";
+			break;
+		case 165:
+		case 149:
+		case 101:
+		case 85:
+		case 229:
+		case 213:
+		case 37:
+		case 21:
+			block = "DB_HTILE";
+			break;
+		case 167:
+		case 151:
+		case 103:
+		case 87:
+		case 231:
+		case 215:
+		case 39:
+		case 23:
+			block = "DB_STEN";
+			break;
+		case 72:
+		case 68:
+		case 64:
+		case 8:
+		case 4:
+		case 0:
+		case 136:
+		case 132:
+		case 128:
+		case 200:
+		case 196:
+		case 192:
+			block = "TC";
+			break;
+		case 112:
+		case 48:
+			block = "CP";
+			break;
+		case 49:
+		case 177:
+		case 50:
+		case 178:
+			block = "SH";
+			break;
+		case 53:
+		case 190:
+			block = "VGT";
+			break;
+		case 117:
+			block = "IH";
+			break;
+		case 51:
+		case 115:
+			block = "RLC";
+			break;
+		case 119:
+		case 183:
+			block = "DMA0";
+			break;
+		case 61:
+			block = "DMA1";
+			break;
+		case 248:
+		case 120:
+			block = "HDP";
+			break;
+		default:
+			block = "unknown";
+			break;
+		}
+	} else {
+		switch (mc_id) {
+		case 32:
+		case 16:
+		case 96:
+		case 80:
+		case 160:
+		case 144:
+		case 224:
+		case 208:
+			block = "CB";
+			break;
+		case 33:
+		case 17:
+		case 97:
+		case 81:
+		case 161:
+		case 145:
+		case 225:
+		case 209:
+			block = "CB_FMASK";
+			break;
+		case 34:
+		case 18:
+		case 98:
+		case 82:
+		case 162:
+		case 146:
+		case 226:
+		case 210:
+			block = "CB_CMASK";
+			break;
+		case 35:
+		case 19:
+		case 99:
+		case 83:
+		case 163:
+		case 147:
+		case 227:
+		case 211:
+			block = "CB_IMMED";
+			break;
+		case 36:
+		case 20:
+		case 100:
+		case 84:
+		case 164:
+		case 148:
+		case 228:
+		case 212:
+			block = "DB";
+			break;
+		case 37:
+		case 21:
+		case 101:
+		case 85:
+		case 165:
+		case 149:
+		case 229:
+		case 213:
+			block = "DB_HTILE";
+			break;
+		case 39:
+		case 23:
+		case 103:
+		case 87:
+		case 167:
+		case 151:
+		case 231:
+		case 215:
+			block = "DB_STEN";
+			break;
+		case 72:
+		case 68:
+		case 8:
+		case 4:
+		case 136:
+		case 132:
+		case 200:
+		case 196:
+			block = "TC";
+			break;
+		case 112:
+		case 48:
+			block = "CP";
+			break;
+		case 49:
+		case 177:
+		case 50:
+		case 178:
+			block = "SH";
+			break;
+		case 53:
+			block = "VGT";
+			break;
+		case 117:
+			block = "IH";
+			break;
+		case 51:
+		case 115:
+			block = "RLC";
+			break;
+		case 119:
+		case 183:
+			block = "DMA0";
+			break;
+		case 61:
+			block = "DMA1";
+			break;
+		case 248:
+		case 120:
+			block = "HDP";
+			break;
+		default:
+			block = "unknown";
+			break;
+		}
+	}
+
+	printk("VM fault (0x%02x, vmid %d) at page %u, %s from %s (%d)\n",
+	       protections, vmid, addr,
+	       (status & MEMORY_CLIENT_RW_MASK) ? "write" : "read",
+	       block, mc_id);
+}
+
+/**
  * si_vm_set_page - update the page tables using the CP
  *
  * @rdev: radeon_device pointer
@@ -4282,6 +4829,450 @@ void si_dma_vm_flush(struct radeon_device *rdev, int ridx, struct radeon_vm *vm)
 }
 
 /*
+ *  Power and clock gating
+ */
+static void si_wait_for_rlc_serdes(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(RLC_SERDES_MASTER_BUSY_0) == 0)
+			break;
+		udelay(1);
+	}
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(RLC_SERDES_MASTER_BUSY_1) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+static void si_enable_gui_idle_interrupt(struct radeon_device *rdev,
+					 bool enable)
+{
+	u32 tmp = RREG32(CP_INT_CNTL_RING0);
+	u32 mask;
+	int i;
+
+	if (enable)
+		tmp |= (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	else
+		tmp &= ~(CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	WREG32(CP_INT_CNTL_RING0, tmp);
+
+	if (!enable) {
+		/* read a gfx register */
+		tmp = RREG32(DB_DEPTH_INFO);
+
+		mask = RLC_BUSY_STATUS | GFX_POWER_STATUS | GFX_CLOCK_STATUS | GFX_LS_STATUS;
+		for (i = 0; i < rdev->usec_timeout; i++) {
+			if ((RREG32(RLC_STAT) & mask) == (GFX_CLOCK_STATUS | GFX_POWER_STATUS))
+				break;
+			udelay(1);
+		}
+	}
+}
+
+static void si_set_uvd_dcm(struct radeon_device *rdev,
+			   bool sw_mode)
+{
+	u32 tmp, tmp2;
+
+	tmp = RREG32(UVD_CGC_CTRL);
+	tmp &= ~(CLK_OD_MASK | CG_DT_MASK);
+	tmp |= DCM | CG_DT(1) | CLK_OD(4);
+
+	if (sw_mode) {
+		tmp &= ~0x7ffff800;
+		tmp2 = DYN_OR_EN | DYN_RR_EN | G_DIV_ID(7);
+	} else {
+		tmp |= 0x7ffff800;
+		tmp2 = 0;
+	}
+
+	WREG32(UVD_CGC_CTRL, tmp);
+	WREG32_UVD_CTX(UVD_CGC_CTRL2, tmp2);
+}
+
+static void si_init_uvd_internal_cg(struct radeon_device *rdev)
+{
+	bool hw_mode = true;
+
+	if (hw_mode) {
+		si_set_uvd_dcm(rdev, false);
+	} else {
+		u32 tmp = RREG32(UVD_CGC_CTRL);
+		tmp &= ~DCM;
+		WREG32(UVD_CGC_CTRL, tmp);
+	}
+}
+
+static u32 si_halt_rlc(struct radeon_device *rdev)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(RLC_CNTL);
+
+	if (data & RLC_ENABLE) {
+		data &= ~RLC_ENABLE;
+		WREG32(RLC_CNTL, data);
+
+		si_wait_for_rlc_serdes(rdev);
+	}
+
+	return orig;
+}
+
+static void si_update_rlc(struct radeon_device *rdev, u32 rlc)
+{
+	u32 tmp;
+
+	tmp = RREG32(RLC_CNTL);
+	if (tmp != rlc)
+		WREG32(RLC_CNTL, rlc);
+}
+
+static void si_enable_dma_pg(struct radeon_device *rdev, bool enable)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(DMA_PG);
+	if (enable)
+		data |= PG_CNTL_ENABLE;
+	else
+		data &= ~PG_CNTL_ENABLE;
+	if (orig != data)
+		WREG32(DMA_PG, data);
+}
+
+static void si_init_dma_pg(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	WREG32(DMA_PGFSM_WRITE,  0x00002000);
+	WREG32(DMA_PGFSM_CONFIG, 0x100010ff);
+
+	for (tmp = 0; tmp < 5; tmp++)
+		WREG32(DMA_PGFSM_WRITE, 0);
+}
+
+static void si_enable_gfx_cgpg(struct radeon_device *rdev,
+			       bool enable)
+{
+	u32 tmp;
+
+	if (enable) {
+		tmp = RLC_PUD(0x10) | RLC_PDD(0x10) | RLC_TTPD(0x10) | RLC_MSD(0x10);
+		WREG32(RLC_TTOP_D, tmp);
+
+		tmp = RREG32(RLC_PG_CNTL);
+		tmp |= GFX_PG_ENABLE;
+		WREG32(RLC_PG_CNTL, tmp);
+
+		tmp = RREG32(RLC_AUTO_PG_CTRL);
+		tmp |= AUTO_PG_EN;
+		WREG32(RLC_AUTO_PG_CTRL, tmp);
+	} else {
+		tmp = RREG32(RLC_AUTO_PG_CTRL);
+		tmp &= ~AUTO_PG_EN;
+		WREG32(RLC_AUTO_PG_CTRL, tmp);
+
+		tmp = RREG32(DB_RENDER_CONTROL);
+	}
+}
+
+static void si_init_gfx_cgpg(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
+
+	tmp = RREG32(RLC_PG_CNTL);
+	tmp |= GFX_PG_SRC;
+	WREG32(RLC_PG_CNTL, tmp);
+
+	WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+
+	tmp = RREG32(RLC_AUTO_PG_CTRL);
+
+	tmp &= ~GRBM_REG_SGIT_MASK;
+	tmp |= GRBM_REG_SGIT(0x700);
+	tmp &= ~PG_AFTER_GRBM_REG_ST_MASK;
+	WREG32(RLC_AUTO_PG_CTRL, tmp);
+}
+
+static u32 si_get_cu_active_bitmap(struct radeon_device *rdev, u32 se, u32 sh)
+{
+	u32 mask = 0, tmp, tmp1;
+	int i;
+
+	si_select_se_sh(rdev, se, sh);
+	tmp = RREG32(CC_GC_SHADER_ARRAY_CONFIG);
+	tmp1 = RREG32(GC_USER_SHADER_ARRAY_CONFIG);
+	si_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+
+	tmp &= 0xffff0000;
+
+	tmp |= tmp1;
+	tmp >>= 16;
+
+	for (i = 0; i < rdev->config.si.max_cu_per_sh; i ++) {
+		mask <<= 1;
+		mask |= 1;
+	}
+
+	return (~tmp) & mask;
+}
+
+static void si_init_ao_cu_mask(struct radeon_device *rdev)
+{
+	u32 i, j, k, active_cu_number = 0;
+	u32 mask, counter, cu_bitmap;
+	u32 tmp = 0;
+
+	for (i = 0; i < rdev->config.si.max_shader_engines; i++) {
+		for (j = 0; j < rdev->config.si.max_sh_per_se; j++) {
+			mask = 1;
+			cu_bitmap = 0;
+			counter  = 0;
+			for (k = 0; k < rdev->config.si.max_cu_per_sh; k++) {
+				if (si_get_cu_active_bitmap(rdev, i, j) & mask) {
+					if (counter < 2)
+						cu_bitmap |= mask;
+					counter++;
+				}
+				mask <<= 1;
+			}
+
+			active_cu_number += counter;
+			tmp |= (cu_bitmap << (i * 16 + j * 8));
+		}
+	}
+
+	WREG32(RLC_PG_AO_CU_MASK, tmp);
+
+	tmp = RREG32(RLC_MAX_PG_CU);
+	tmp &= ~MAX_PU_CU_MASK;
+	tmp |= MAX_PU_CU(active_cu_number);
+	WREG32(RLC_MAX_PG_CU, tmp);
+}
+
+static void si_enable_cgcg(struct radeon_device *rdev,
+			   bool enable)
+{
+	u32 data, orig, tmp;
+
+	orig = data = RREG32(RLC_CGCG_CGLS_CTRL);
+
+	si_enable_gui_idle_interrupt(rdev, enable);
+
+	if (enable) {
+		WREG32(RLC_GCPM_GENERAL_3, 0x00000080);
+
+		tmp = si_halt_rlc(rdev);
+
+		WREG32(RLC_SERDES_WR_MASTER_MASK_0, 0xffffffff);
+		WREG32(RLC_SERDES_WR_MASTER_MASK_1, 0xffffffff);
+		WREG32(RLC_SERDES_WR_CTRL, 0x00b000ff);
+
+		si_wait_for_rlc_serdes(rdev);
+
+		si_update_rlc(rdev, tmp);
+
+		WREG32(RLC_SERDES_WR_CTRL, 0x007000ff);
+
+		data |= CGCG_EN | CGLS_EN;
+	} else {
+		RREG32(CB_CGTT_SCLK_CTRL);
+		RREG32(CB_CGTT_SCLK_CTRL);
+		RREG32(CB_CGTT_SCLK_CTRL);
+		RREG32(CB_CGTT_SCLK_CTRL);
+
+		data &= ~(CGCG_EN | CGLS_EN);
+	}
+
+	if (orig != data)
+		WREG32(RLC_CGCG_CGLS_CTRL, data);
+}
+
+static void si_enable_mgcg(struct radeon_device *rdev,
+			   bool enable)
+{
+	u32 data, orig, tmp = 0;
+
+	if (enable) {
+		orig = data = RREG32(CGTS_SM_CTRL_REG);
+		data = 0x96940200;
+		if (orig != data)
+			WREG32(CGTS_SM_CTRL_REG, data);
+
+		orig = data = RREG32(CP_MEM_SLP_CNTL);
+		data |= CP_MEM_LS_EN;
+		if (orig != data)
+			WREG32(CP_MEM_SLP_CNTL, data);
+
+		orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE);
+		data &= 0xffffffc0;
+		if (orig != data)
+			WREG32(RLC_CGTT_MGCG_OVERRIDE, data);
+
+		tmp = si_halt_rlc(rdev);
+
+		WREG32(RLC_SERDES_WR_MASTER_MASK_0, 0xffffffff);
+		WREG32(RLC_SERDES_WR_MASTER_MASK_1, 0xffffffff);
+		WREG32(RLC_SERDES_WR_CTRL, 0x00d000ff);
+
+		si_update_rlc(rdev, tmp);
+	} else {
+		orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE);
+		data |= 0x00000003;
+		if (orig != data)
+			WREG32(RLC_CGTT_MGCG_OVERRIDE, data);
+
+		data = RREG32(CP_MEM_SLP_CNTL);
+		if (data & CP_MEM_LS_EN) {
+			data &= ~CP_MEM_LS_EN;
+			WREG32(CP_MEM_SLP_CNTL, data);
+		}
+		orig = data = RREG32(CGTS_SM_CTRL_REG);
+		data |= LS_OVERRIDE | OVERRIDE;
+		if (orig != data)
+			WREG32(CGTS_SM_CTRL_REG, data);
+
+		tmp = si_halt_rlc(rdev);
+
+		WREG32(RLC_SERDES_WR_MASTER_MASK_0, 0xffffffff);
+		WREG32(RLC_SERDES_WR_MASTER_MASK_1, 0xffffffff);
+		WREG32(RLC_SERDES_WR_CTRL, 0x00e000ff);
+
+		si_update_rlc(rdev, tmp);
+	}
+}
+
+static void si_enable_uvd_mgcg(struct radeon_device *rdev,
+			       bool enable)
+{
+	u32 orig, data, tmp;
+
+	if (enable) {
+		tmp = RREG32_UVD_CTX(UVD_CGC_MEM_CTRL);
+		tmp |= 0x3fff;
+		WREG32_UVD_CTX(UVD_CGC_MEM_CTRL, tmp);
+
+		orig = data = RREG32(UVD_CGC_CTRL);
+		data |= DCM;
+		if (orig != data)
+			WREG32(UVD_CGC_CTRL, data);
+
+		WREG32_SMC(SMC_CG_IND_START + CG_CGTT_LOCAL_0, 0);
+		WREG32_SMC(SMC_CG_IND_START + CG_CGTT_LOCAL_1, 0);
+	} else {
+		tmp = RREG32_UVD_CTX(UVD_CGC_MEM_CTRL);
+		tmp &= ~0x3fff;
+		WREG32_UVD_CTX(UVD_CGC_MEM_CTRL, tmp);
+
+		orig = data = RREG32(UVD_CGC_CTRL);
+		data &= ~DCM;
+		if (orig != data)
+			WREG32(UVD_CGC_CTRL, data);
+
+		WREG32_SMC(SMC_CG_IND_START + CG_CGTT_LOCAL_0, 0xffffffff);
+		WREG32_SMC(SMC_CG_IND_START + CG_CGTT_LOCAL_1, 0xffffffff);
+	}
+}
+
+static const u32 mc_cg_registers[] =
+{
+	MC_HUB_MISC_HUB_CG,
+	MC_HUB_MISC_SIP_CG,
+	MC_HUB_MISC_VM_CG,
+	MC_XPB_CLK_GAT,
+	ATC_MISC_CG,
+	MC_CITF_MISC_WR_CG,
+	MC_CITF_MISC_RD_CG,
+	MC_CITF_MISC_VM_CG,
+	VM_L2_CG,
+};
+
+static void si_enable_mc_ls(struct radeon_device *rdev,
+			    bool enable)
+{
+	int i;
+	u32 orig, data;
+
+	for (i = 0; i < ARRAY_SIZE(mc_cg_registers); i++) {
+		orig = data = RREG32(mc_cg_registers[i]);
+		if (enable)
+			data |= MC_LS_ENABLE;
+		else
+			data &= ~MC_LS_ENABLE;
+		if (data != orig)
+			WREG32(mc_cg_registers[i], data);
+	}
+}
+
+
+static void si_init_cg(struct radeon_device *rdev)
+{
+	bool has_uvd = true;
+
+	si_enable_mgcg(rdev, true);
+	si_enable_cgcg(rdev, true);
+	/* disable MC LS on Tahiti */
+	if (rdev->family == CHIP_TAHITI)
+		si_enable_mc_ls(rdev, false);
+	if (has_uvd) {
+		si_enable_uvd_mgcg(rdev, true);
+		si_init_uvd_internal_cg(rdev);
+	}
+}
+
+static void si_fini_cg(struct radeon_device *rdev)
+{
+	bool has_uvd = true;
+
+	if (has_uvd)
+		si_enable_uvd_mgcg(rdev, false);
+	si_enable_cgcg(rdev, false);
+	si_enable_mgcg(rdev, false);
+}
+
+static void si_init_pg(struct radeon_device *rdev)
+{
+	bool has_pg = false;
+
+	/* only cape verde supports PG */
+	if (rdev->family == CHIP_VERDE)
+		has_pg = true;
+
+	if (has_pg) {
+		si_init_ao_cu_mask(rdev);
+		si_init_dma_pg(rdev);
+		si_enable_dma_pg(rdev, true);
+		si_init_gfx_cgpg(rdev);
+		si_enable_gfx_cgpg(rdev, true);
+	} else {
+		WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
+		WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+	}
+}
+
+static void si_fini_pg(struct radeon_device *rdev)
+{
+	bool has_pg = false;
+
+	/* only cape verde supports PG */
+	if (rdev->family == CHIP_VERDE)
+		has_pg = true;
+
+	if (has_pg) {
+		si_enable_dma_pg(rdev, false);
+		si_enable_gfx_cgpg(rdev, false);
+	}
+}
+
+/*
  * RLC
  */
 void si_rlc_fini(struct radeon_device *rdev)
@@ -4313,8 +5304,15 @@ void si_rlc_fini(struct radeon_device *rdev)
 	}
 }
 
+#define RLC_CLEAR_STATE_END_MARKER          0x00000001
+
 int si_rlc_init(struct radeon_device *rdev)
 {
+	volatile u32 *dst_ptr;
+	u32 dws, data, i, j, k, reg_num;
+	u32 reg_list_num, reg_list_hdr_blk_index, reg_list_blk_index;
+	u64 reg_list_mc_addr;
+	const struct cs_section_def *cs_data = si_cs_data;
 	int r;
 
 	/* save restore block */
@@ -4335,18 +5333,44 @@ int si_rlc_init(struct radeon_device *rdev)
 	}
 	r = radeon_bo_pin(rdev->rlc.save_restore_obj, RADEON_GEM_DOMAIN_VRAM,
 			  &rdev->rlc.save_restore_gpu_addr);
-	radeon_bo_unreserve(rdev->rlc.save_restore_obj);
 	if (r) {
+		radeon_bo_unreserve(rdev->rlc.save_restore_obj);
 		dev_warn(rdev->dev, "(%d) pin RLC sr bo failed\n", r);
 		si_rlc_fini(rdev);
 		return r;
 	}
 
+	if (rdev->family == CHIP_VERDE) {
+		r = radeon_bo_kmap(rdev->rlc.save_restore_obj, (void **)&rdev->rlc.sr_ptr);
+		if (r) {
+			dev_warn(rdev->dev, "(%d) map RLC sr bo failed\n", r);
+			si_rlc_fini(rdev);
+		return r;
+		}
+		/* write the sr buffer */
+		dst_ptr = rdev->rlc.sr_ptr;
+		for (i = 0; i < ARRAY_SIZE(verde_rlc_save_restore_register_list); i++) {
+			dst_ptr[i] = verde_rlc_save_restore_register_list[i];
+		}
+		radeon_bo_kunmap(rdev->rlc.save_restore_obj);
+	}
+	radeon_bo_unreserve(rdev->rlc.save_restore_obj);
+
 	/* clear state block */
+	reg_list_num = 0;
+	dws = 0;
+	for (i = 0; cs_data[i].section != NULL; i++) {
+		for (j = 0; cs_data[i].section[j].extent != NULL; j++) {
+			reg_list_num++;
+			dws += cs_data[i].section[j].reg_count;
+		}
+	}
+	reg_list_blk_index = (3 * reg_list_num + 2);
+	dws += reg_list_blk_index;
+
 	if (rdev->rlc.clear_state_obj == NULL) {
-		r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE, PAGE_SIZE, true,
-				     RADEON_GEM_DOMAIN_VRAM, NULL,
-				     &rdev->rlc.clear_state_obj);
+		r = radeon_bo_create(rdev, dws * 4, PAGE_SIZE, true,
+				     RADEON_GEM_DOMAIN_VRAM, NULL, &rdev->rlc.clear_state_obj);
 		if (r) {
 			dev_warn(rdev->dev, "(%d) create RLC c bo failed\n", r);
 			si_rlc_fini(rdev);
@@ -4360,24 +5384,113 @@ int si_rlc_init(struct radeon_device *rdev)
 	}
 	r = radeon_bo_pin(rdev->rlc.clear_state_obj, RADEON_GEM_DOMAIN_VRAM,
 			  &rdev->rlc.clear_state_gpu_addr);
-	radeon_bo_unreserve(rdev->rlc.clear_state_obj);
 	if (r) {
+
+		radeon_bo_unreserve(rdev->rlc.clear_state_obj);
 		dev_warn(rdev->dev, "(%d) pin RLC c bo failed\n", r);
 		si_rlc_fini(rdev);
 		return r;
 	}
+	r = radeon_bo_kmap(rdev->rlc.clear_state_obj, (void **)&rdev->rlc.cs_ptr);
+	if (r) {
+		dev_warn(rdev->dev, "(%d) map RLC c bo failed\n", r);
+		si_rlc_fini(rdev);
+		return r;
+	}
+	/* set up the cs buffer */
+	dst_ptr = rdev->rlc.cs_ptr;
+	reg_list_hdr_blk_index = 0;
+	reg_list_mc_addr = rdev->rlc.clear_state_gpu_addr + (reg_list_blk_index * 4);
+	data = upper_32_bits(reg_list_mc_addr);
+	dst_ptr[reg_list_hdr_blk_index] = data;
+	reg_list_hdr_blk_index++;
+	for (i = 0; cs_data[i].section != NULL; i++) {
+		for (j = 0; cs_data[i].section[j].extent != NULL; j++) {
+			reg_num = cs_data[i].section[j].reg_count;
+			data = reg_list_mc_addr & 0xffffffff;
+			dst_ptr[reg_list_hdr_blk_index] = data;
+			reg_list_hdr_blk_index++;
+
+			data = (cs_data[i].section[j].reg_index * 4) & 0xffffffff;
+			dst_ptr[reg_list_hdr_blk_index] = data;
+			reg_list_hdr_blk_index++;
+
+			data = 0x08000000 | (reg_num * 4);
+			dst_ptr[reg_list_hdr_blk_index] = data;
+			reg_list_hdr_blk_index++;
+
+			for (k = 0; k < reg_num; k++) {
+				data = cs_data[i].section[j].extent[k];
+				dst_ptr[reg_list_blk_index + k] = data;
+			}
+			reg_list_mc_addr += reg_num * 4;
+			reg_list_blk_index += reg_num;
+		}
+	}
+	dst_ptr[reg_list_hdr_blk_index] = RLC_CLEAR_STATE_END_MARKER;
+
+	radeon_bo_kunmap(rdev->rlc.clear_state_obj);
+	radeon_bo_unreserve(rdev->rlc.clear_state_obj);
 
 	return 0;
 }
 
+static void si_rlc_reset(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(GRBM_SOFT_RESET);
+
+	tmp |= SOFT_RESET_RLC;
+	WREG32(GRBM_SOFT_RESET, tmp);
+	udelay(50);
+	tmp &= ~SOFT_RESET_RLC;
+	WREG32(GRBM_SOFT_RESET, tmp);
+	udelay(50);
+}
+
 static void si_rlc_stop(struct radeon_device *rdev)
 {
 	WREG32(RLC_CNTL, 0);
+
+	si_enable_gui_idle_interrupt(rdev, false);
+
+	si_wait_for_rlc_serdes(rdev);
 }
 
 static void si_rlc_start(struct radeon_device *rdev)
 {
 	WREG32(RLC_CNTL, RLC_ENABLE);
+
+	si_enable_gui_idle_interrupt(rdev, true);
+
+	udelay(50);
+}
+
+static bool si_lbpw_supported(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	/* Enable LBPW only for DDR3 */
+	tmp = RREG32(MC_SEQ_MISC0);
+	if ((tmp & 0xF0000000) == 0xB0000000)
+		return true;
+	return false;
+}
+
+static void si_enable_lbpw(struct radeon_device *rdev, bool enable)
+{
+	u32 tmp;
+
+	tmp = RREG32(RLC_LB_CNTL);
+	if (enable)
+		tmp |= LOAD_BALANCE_ENABLE;
+	else
+		tmp &= ~LOAD_BALANCE_ENABLE;
+	WREG32(RLC_LB_CNTL, tmp);
+
+	if (!enable) {
+		si_select_se_sh(rdev, 0xffffffff, 0xffffffff);
+		WREG32(SPI_LB_CU_MASK, 0x00ff);
+	}
 }
 
 static int si_rlc_resume(struct radeon_device *rdev)
@@ -4390,14 +5503,18 @@ static int si_rlc_resume(struct radeon_device *rdev)
 
 	si_rlc_stop(rdev);
 
+	si_rlc_reset(rdev);
+
+	si_init_pg(rdev);
+
+	si_init_cg(rdev);
+
 	WREG32(RLC_RL_BASE, 0);
 	WREG32(RLC_RL_SIZE, 0);
 	WREG32(RLC_LB_CNTL, 0);
 	WREG32(RLC_LB_CNTR_MAX, 0xffffffff);
 	WREG32(RLC_LB_CNTR_INIT, 0);
-
-	WREG32(RLC_SAVE_AND_RESTORE_BASE, rdev->rlc.save_restore_gpu_addr >> 8);
-	WREG32(RLC_CLEAR_STATE_RESTORE_BASE, rdev->rlc.clear_state_gpu_addr >> 8);
+	WREG32(RLC_LB_INIT_CU_MASK, 0xffffffff);
 
 	WREG32(RLC_MC_CNTL, 0);
 	WREG32(RLC_UCODE_CNTL, 0);
@@ -4409,6 +5526,8 @@ static int si_rlc_resume(struct radeon_device *rdev)
 	}
 	WREG32(RLC_UCODE_ADDR, 0);
 
+	si_enable_lbpw(rdev, si_lbpw_supported(rdev));
+
 	si_rlc_start(rdev);
 
 	return 0;
@@ -4578,6 +5697,7 @@ int si_irq_set(struct radeon_device *rdev)
 	u32 grbm_int_cntl = 0;
 	u32 grph1 = 0, grph2 = 0, grph3 = 0, grph4 = 0, grph5 = 0, grph6 = 0;
 	u32 dma_cntl, dma_cntl1;
+	u32 thermal_int = 0;
 
 	if (!rdev->irq.installed) {
 		WARN(1, "Can't enable IRQ/MSI because no handler is installed\n");
@@ -4603,6 +5723,9 @@ int si_irq_set(struct radeon_device *rdev)
 	dma_cntl = RREG32(DMA_CNTL + DMA0_REGISTER_OFFSET) & ~TRAP_ENABLE;
 	dma_cntl1 = RREG32(DMA_CNTL + DMA1_REGISTER_OFFSET) & ~TRAP_ENABLE;
 
+	thermal_int = RREG32(CG_THERMAL_INT) &
+		~(THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW);
+
 	/* enable CP interrupts on all rings */
 	if (atomic_read(&rdev->irq.ring_int[RADEON_RING_TYPE_GFX_INDEX])) {
 		DRM_DEBUG("si_irq_set: sw int gfx\n");
@@ -4689,6 +5812,11 @@ int si_irq_set(struct radeon_device *rdev)
 
 	WREG32(GRBM_INT_CNTL, grbm_int_cntl);
 
+	if (rdev->irq.dpm_thermal) {
+		DRM_DEBUG("dpm thermal\n");
+		thermal_int |= THERM_INT_MASK_HIGH | THERM_INT_MASK_LOW;
+	}
+
 	if (rdev->num_crtc >= 2) {
 		WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, crtc1);
 		WREG32(INT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, crtc2);
@@ -4724,6 +5852,8 @@ int si_irq_set(struct radeon_device *rdev)
 		WREG32(DC_HPD6_INT_CONTROL, hpd6);
 	}
 
+	WREG32(CG_THERMAL_INT, thermal_int);
+
 	return 0;
 }
 
@@ -4888,6 +6018,8 @@ int si_irq_process(struct radeon_device *rdev)
 	u32 src_id, src_data, ring_id;
 	u32 ring_index;
 	bool queue_hotplug = false;
+	bool queue_thermal = false;
+	u32 status, addr;
 
 	if (!rdev->ih.enabled || rdev->shutdown)
 		return IRQ_NONE;
@@ -5123,11 +6255,14 @@ restart_ih:
 			break;
 		case 146:
 		case 147:
+			addr = RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR);
+			status = RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS);
 			dev_err(rdev->dev, "GPU fault detected: %d 0x%08x\n", src_id, src_data);
 			dev_err(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x%08X\n",
-				RREG32(VM_CONTEXT1_PROTECTION_FAULT_ADDR));
+				addr);
 			dev_err(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
-				RREG32(VM_CONTEXT1_PROTECTION_FAULT_STATUS));
+				status);
+			si_vm_decode_fault(rdev, status, addr);
 			/* reset addr and status */
 			WREG32_P(VM_CONTEXT1_CNTL2, 1, ~1);
 			break;
@@ -5158,6 +6293,16 @@ restart_ih:
 			DRM_DEBUG("IH: DMA trap\n");
 			radeon_fence_process(rdev, R600_RING_TYPE_DMA_INDEX);
 			break;
+		case 230: /* thermal low to high */
+			DRM_DEBUG("IH: thermal low to high\n");
+			rdev->pm.dpm.thermal.high_to_low = false;
+			queue_thermal = true;
+			break;
+		case 231: /* thermal high to low */
+			DRM_DEBUG("IH: thermal high to low\n");
+			rdev->pm.dpm.thermal.high_to_low = true;
+			queue_thermal = true;
+			break;
 		case 233: /* GUI IDLE */
 			DRM_DEBUG("IH: GUI idle\n");
 			break;
@@ -5176,6 +6321,8 @@ restart_ih:
 	}
 	if (queue_hotplug)
 		schedule_work(&rdev->hotplug_work);
+	if (queue_thermal && rdev->pm.dpm_enabled)
+		schedule_work(&rdev->pm.dpm.thermal.work);
 	rdev->ih.rptr = rptr;
 	WREG32(IH_RB_RPTR, rdev->ih.rptr);
 	atomic_set(&rdev->ih.lock, 0);
@@ -5270,6 +6417,11 @@ static int si_startup(struct radeon_device *rdev)
 	struct radeon_ring *ring;
 	int r;
 
+	/* enable pcie gen2/3 link */
+	si_pcie_gen3_enable(rdev);
+	/* enable aspm */
+	si_program_aspm(rdev);
+
 	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->ce_fw ||
 	    !rdev->rlc_fw || !rdev->mc_fw) {
 		r = si_init_microcode(rdev);
@@ -5609,6 +6761,8 @@ void si_fini(struct radeon_device *rdev)
 	cayman_dma_fini(rdev);
 	si_irq_fini(rdev);
 	si_rlc_fini(rdev);
+	si_fini_cg(rdev);
+	si_fini_pg(rdev);
 	radeon_wb_fini(rdev);
 	radeon_vm_manager_fini(rdev);
 	radeon_ib_pool_fini(rdev);
@@ -5735,3 +6889,364 @@ int si_set_uvd_clocks(struct radeon_device *rdev, u32 vclk, u32 dclk)
 
 	return 0;
 }
+
+static void si_pcie_gen3_enable(struct radeon_device *rdev)
+{
+	struct pci_dev *root = rdev->pdev->bus->self;
+	int bridge_pos, gpu_pos;
+	u32 speed_cntl, mask, current_data_rate;
+	int ret, i;
+	u16 tmp16;
+
+	if (radeon_pcie_gen2 == 0)
+		return;
+
+	if (rdev->flags & RADEON_IS_IGP)
+		return;
+
+	if (!(rdev->flags & RADEON_IS_PCIE))
+		return;
+
+	ret = drm_pcie_get_speed_cap_mask(rdev->ddev, &mask);
+	if (ret != 0)
+		return;
+
+	if (!(mask & (DRM_PCIE_SPEED_50 | DRM_PCIE_SPEED_80)))
+		return;
+
+	speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	current_data_rate = (speed_cntl & LC_CURRENT_DATA_RATE_MASK) >>
+		LC_CURRENT_DATA_RATE_SHIFT;
+	if (mask & DRM_PCIE_SPEED_80) {
+		if (current_data_rate == 2) {
+			DRM_INFO("PCIE gen 3 link speeds already enabled\n");
+			return;
+		}
+		DRM_INFO("enabling PCIE gen 3 link speeds, disable with radeon.pcie_gen2=0\n");
+	} else if (mask & DRM_PCIE_SPEED_50) {
+		if (current_data_rate == 1) {
+			DRM_INFO("PCIE gen 2 link speeds already enabled\n");
+			return;
+		}
+		DRM_INFO("enabling PCIE gen 2 link speeds, disable with radeon.pcie_gen2=0\n");
+	}
+
+	bridge_pos = pci_pcie_cap(root);
+	if (!bridge_pos)
+		return;
+
+	gpu_pos = pci_pcie_cap(rdev->pdev);
+	if (!gpu_pos)
+		return;
+
+	if (mask & DRM_PCIE_SPEED_80) {
+		/* re-try equalization if gen3 is not already enabled */
+		if (current_data_rate != 2) {
+			u16 bridge_cfg, gpu_cfg;
+			u16 bridge_cfg2, gpu_cfg2;
+			u32 max_lw, current_lw, tmp;
+
+			pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL, &bridge_cfg);
+			pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, &gpu_cfg);
+
+			tmp16 = bridge_cfg | PCI_EXP_LNKCTL_HAWD;
+			pci_write_config_word(root, bridge_pos + PCI_EXP_LNKCTL, tmp16);
+
+			tmp16 = gpu_cfg | PCI_EXP_LNKCTL_HAWD;
+			pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, tmp16);
+
+			tmp = RREG32_PCIE(PCIE_LC_STATUS1);
+			max_lw = (tmp & LC_DETECTED_LINK_WIDTH_MASK) >> LC_DETECTED_LINK_WIDTH_SHIFT;
+			current_lw = (tmp & LC_OPERATING_LINK_WIDTH_MASK) >> LC_OPERATING_LINK_WIDTH_SHIFT;
+
+			if (current_lw < max_lw) {
+				tmp = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
+				if (tmp & LC_RENEGOTIATION_SUPPORT) {
+					tmp &= ~(LC_LINK_WIDTH_MASK | LC_UPCONFIGURE_DIS);
+					tmp |= (max_lw << LC_LINK_WIDTH_SHIFT);
+					tmp |= LC_UPCONFIGURE_SUPPORT | LC_RENEGOTIATE_EN | LC_RECONFIG_NOW;
+					WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, tmp);
+				}
+			}
+
+			for (i = 0; i < 10; i++) {
+				/* check status */
+				pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_DEVSTA, &tmp16);
+				if (tmp16 & PCI_EXP_DEVSTA_TRPND)
+					break;
+
+				pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL, &bridge_cfg);
+				pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, &gpu_cfg);
+
+				pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL2, &bridge_cfg2);
+				pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, &gpu_cfg2);
+
+				tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4);
+				tmp |= LC_SET_QUIESCE;
+				WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp);
+
+				tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4);
+				tmp |= LC_REDO_EQ;
+				WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp);
+
+				mdelay(100);
+
+				/* linkctl */
+				pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL, &tmp16);
+				tmp16 &= ~PCI_EXP_LNKCTL_HAWD;
+				tmp16 |= (bridge_cfg & PCI_EXP_LNKCTL_HAWD);
+				pci_write_config_word(root, bridge_pos + PCI_EXP_LNKCTL, tmp16);
+
+				pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, &tmp16);
+				tmp16 &= ~PCI_EXP_LNKCTL_HAWD;
+				tmp16 |= (gpu_cfg & PCI_EXP_LNKCTL_HAWD);
+				pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL, tmp16);
+
+				/* linkctl2 */
+				pci_read_config_word(root, bridge_pos + PCI_EXP_LNKCTL2, &tmp16);
+				tmp16 &= ~((1 << 4) | (7 << 9));
+				tmp16 |= (bridge_cfg2 & ((1 << 4) | (7 << 9)));
+				pci_write_config_word(root, bridge_pos + PCI_EXP_LNKCTL2, tmp16);
+
+				pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, &tmp16);
+				tmp16 &= ~((1 << 4) | (7 << 9));
+				tmp16 |= (gpu_cfg2 & ((1 << 4) | (7 << 9)));
+				pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, tmp16);
+
+				tmp = RREG32_PCIE_PORT(PCIE_LC_CNTL4);
+				tmp &= ~LC_SET_QUIESCE;
+				WREG32_PCIE_PORT(PCIE_LC_CNTL4, tmp);
+			}
+		}
+	}
+
+	/* set the link speed */
+	speed_cntl |= LC_FORCE_EN_SW_SPEED_CHANGE | LC_FORCE_DIS_HW_SPEED_CHANGE;
+	speed_cntl &= ~LC_FORCE_DIS_SW_SPEED_CHANGE;
+	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
+
+	pci_read_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, &tmp16);
+	tmp16 &= ~0xf;
+	if (mask & DRM_PCIE_SPEED_80)
+		tmp16 |= 3; /* gen3 */
+	else if (mask & DRM_PCIE_SPEED_50)
+		tmp16 |= 2; /* gen2 */
+	else
+		tmp16 |= 1; /* gen1 */
+	pci_write_config_word(rdev->pdev, gpu_pos + PCI_EXP_LNKCTL2, tmp16);
+
+	speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+	speed_cntl |= LC_INITIATE_LINK_SPEED_CHANGE;
+	WREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL, speed_cntl);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL);
+		if ((speed_cntl & LC_INITIATE_LINK_SPEED_CHANGE) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+static void si_program_aspm(struct radeon_device *rdev)
+{
+	u32 data, orig;
+	bool disable_l0s = false, disable_l1 = false, disable_plloff_in_l1 = false;
+	bool disable_clkreq = false;
+
+	if (radeon_aspm == 0)
+		return;
+
+	if (!(rdev->flags & RADEON_IS_PCIE))
+		return;
+
+	orig = data = RREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL);
+	data &= ~LC_XMIT_N_FTS_MASK;
+	data |= LC_XMIT_N_FTS(0x24) | LC_XMIT_N_FTS_OVERRIDE_EN;
+	if (orig != data)
+		WREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL, data);
+
+	orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL3);
+	data |= LC_GO_TO_RECOVERY;
+	if (orig != data)
+		WREG32_PCIE_PORT(PCIE_LC_CNTL3, data);
+
+	orig = data = RREG32_PCIE(PCIE_P_CNTL);
+	data |= P_IGNORE_EDB_ERR;
+	if (orig != data)
+		WREG32_PCIE(PCIE_P_CNTL, data);
+
+	orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL);
+	data &= ~(LC_L0S_INACTIVITY_MASK | LC_L1_INACTIVITY_MASK);
+	data |= LC_PMI_TO_L1_DIS;
+	if (!disable_l0s)
+		data |= LC_L0S_INACTIVITY(7);
+
+	if (!disable_l1) {
+		data |= LC_L1_INACTIVITY(7);
+		data &= ~LC_PMI_TO_L1_DIS;
+		if (orig != data)
+			WREG32_PCIE_PORT(PCIE_LC_CNTL, data);
+
+		if (!disable_plloff_in_l1) {
+			bool clk_req_support;
+
+			orig = data = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0);
+			data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7);
+			if (orig != data)
+				WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0, data);
+
+			orig = data = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1);
+			data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7);
+			if (orig != data)
+				WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1, data);
+
+			orig = data = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0);
+			data &= ~(PLL_POWER_STATE_IN_OFF_0_MASK | PLL_POWER_STATE_IN_TXS2_0_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_0(7) | PLL_POWER_STATE_IN_TXS2_0(7);
+			if (orig != data)
+				WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0, data);
+
+			orig = data = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1);
+			data &= ~(PLL_POWER_STATE_IN_OFF_1_MASK | PLL_POWER_STATE_IN_TXS2_1_MASK);
+			data |= PLL_POWER_STATE_IN_OFF_1(7) | PLL_POWER_STATE_IN_TXS2_1(7);
+			if (orig != data)
+				WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1, data);
+
+			if ((rdev->family != CHIP_OLAND) && (rdev->family != CHIP_HAINAN)) {
+				orig = data = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0);
+				data &= ~PLL_RAMP_UP_TIME_0_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_0, data);
+
+				orig = data = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1);
+				data &= ~PLL_RAMP_UP_TIME_1_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_1, data);
+
+				orig = data = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_2);
+				data &= ~PLL_RAMP_UP_TIME_2_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_2, data);
+
+				orig = data = RREG32_PIF_PHY0(PB0_PIF_PWRDOWN_3);
+				data &= ~PLL_RAMP_UP_TIME_3_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY0(PB0_PIF_PWRDOWN_3, data);
+
+				orig = data = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0);
+				data &= ~PLL_RAMP_UP_TIME_0_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_0, data);
+
+				orig = data = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1);
+				data &= ~PLL_RAMP_UP_TIME_1_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_1, data);
+
+				orig = data = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_2);
+				data &= ~PLL_RAMP_UP_TIME_2_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_2, data);
+
+				orig = data = RREG32_PIF_PHY1(PB1_PIF_PWRDOWN_3);
+				data &= ~PLL_RAMP_UP_TIME_3_MASK;
+				if (orig != data)
+					WREG32_PIF_PHY1(PB1_PIF_PWRDOWN_3, data);
+			}
+			orig = data = RREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL);
+			data &= ~LC_DYN_LANES_PWR_STATE_MASK;
+			data |= LC_DYN_LANES_PWR_STATE(3);
+			if (orig != data)
+				WREG32_PCIE_PORT(PCIE_LC_LINK_WIDTH_CNTL, data);
+
+			orig = data = RREG32_PIF_PHY0(PB0_PIF_CNTL);
+			data &= ~LS2_EXIT_TIME_MASK;
+			if ((rdev->family == CHIP_OLAND) || (rdev->family == CHIP_HAINAN))
+				data |= LS2_EXIT_TIME(5);
+			if (orig != data)
+				WREG32_PIF_PHY0(PB0_PIF_CNTL, data);
+
+			orig = data = RREG32_PIF_PHY1(PB1_PIF_CNTL);
+			data &= ~LS2_EXIT_TIME_MASK;
+			if ((rdev->family == CHIP_OLAND) || (rdev->family == CHIP_HAINAN))
+				data |= LS2_EXIT_TIME(5);
+			if (orig != data)
+				WREG32_PIF_PHY1(PB1_PIF_CNTL, data);
+
+			if (!disable_clkreq) {
+				struct pci_dev *root = rdev->pdev->bus->self;
+				u32 lnkcap;
+
+				clk_req_support = false;
+				pcie_capability_read_dword(root, PCI_EXP_LNKCAP, &lnkcap);
+				if (lnkcap & PCI_EXP_LNKCAP_CLKPM)
+					clk_req_support = true;
+			} else {
+				clk_req_support = false;
+			}
+
+			if (clk_req_support) {
+				orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL2);
+				data |= LC_ALLOW_PDWN_IN_L1 | LC_ALLOW_PDWN_IN_L23;
+				if (orig != data)
+					WREG32_PCIE_PORT(PCIE_LC_CNTL2, data);
+
+				orig = data = RREG32(THM_CLK_CNTL);
+				data &= ~(CMON_CLK_SEL_MASK | TMON_CLK_SEL_MASK);
+				data |= CMON_CLK_SEL(1) | TMON_CLK_SEL(1);
+				if (orig != data)
+					WREG32(THM_CLK_CNTL, data);
+
+				orig = data = RREG32(MISC_CLK_CNTL);
+				data &= ~(DEEP_SLEEP_CLK_SEL_MASK | ZCLK_SEL_MASK);
+				data |= DEEP_SLEEP_CLK_SEL(1) | ZCLK_SEL(1);
+				if (orig != data)
+					WREG32(MISC_CLK_CNTL, data);
+
+				orig = data = RREG32(CG_CLKPIN_CNTL);
+				data &= ~BCLK_AS_XCLK;
+				if (orig != data)
+					WREG32(CG_CLKPIN_CNTL, data);
+
+				orig = data = RREG32(CG_CLKPIN_CNTL_2);
+				data &= ~FORCE_BIF_REFCLK_EN;
+				if (orig != data)
+					WREG32(CG_CLKPIN_CNTL_2, data);
+
+				orig = data = RREG32(MPLL_BYPASSCLK_SEL);
+				data &= ~MPLL_CLKOUT_SEL_MASK;
+				data |= MPLL_CLKOUT_SEL(4);
+				if (orig != data)
+					WREG32(MPLL_BYPASSCLK_SEL, data);
+
+				orig = data = RREG32(SPLL_CNTL_MODE);
+				data &= ~SPLL_REFCLK_SEL_MASK;
+				if (orig != data)
+					WREG32(SPLL_CNTL_MODE, data);
+			}
+		}
+	} else {
+		if (orig != data)
+			WREG32_PCIE_PORT(PCIE_LC_CNTL, data);
+	}
+
+	orig = data = RREG32_PCIE(PCIE_CNTL2);
+	data |= SLV_MEM_LS_EN | MST_MEM_LS_EN | REPLAY_MEM_LS_EN;
+	if (orig != data)
+		WREG32_PCIE(PCIE_CNTL2, data);
+
+	if (!disable_l0s) {
+		data = RREG32_PCIE_PORT(PCIE_LC_N_FTS_CNTL);
+		if((data & LC_N_FTS_MASK) == LC_N_FTS_MASK) {
+			data = RREG32_PCIE(PCIE_LC_STATUS1);
+			if ((data & LC_REVERSE_XMIT) && (data & LC_REVERSE_RCVR)) {
+				orig = data = RREG32_PCIE_PORT(PCIE_LC_CNTL);
+				data &= ~LC_L0S_INACTIVITY_MASK;
+				if (orig != data)
+					WREG32_PCIE_PORT(PCIE_LC_CNTL, data);
+			}
+		}
+	}
+}
diff --git a/drivers/gpu/drm/radeon/si_dpm.c b/drivers/gpu/drm/radeon/si_dpm.c
new file mode 100644
index 000000000000..73aaa2e4c312
--- /dev/null
+++ b/drivers/gpu/drm/radeon/si_dpm.c
@@ -0,0 +1,6432 @@
+/*
+ * Copyright 2013 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "sid.h"
+#include "r600_dpm.h"
+#include "si_dpm.h"
+#include "atom.h"
+#include <linux/math64.h>
+#include <linux/seq_file.h>
+
+#define MC_CG_ARB_FREQ_F0           0x0a
+#define MC_CG_ARB_FREQ_F1           0x0b
+#define MC_CG_ARB_FREQ_F2           0x0c
+#define MC_CG_ARB_FREQ_F3           0x0d
+
+#define SMC_RAM_END                 0x20000
+
+#define DDR3_DRAM_ROWS              0x2000
+
+#define SCLK_MIN_DEEPSLEEP_FREQ     1350
+
+static const struct si_cac_config_reg cac_weights_tahiti[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0xc, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x101, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0xc, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x8fc, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x95, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x34e, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x1a1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0xda, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x46, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x208, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0xe7, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x948, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x167, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x31, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x18e, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg lcac_tahiti[] =
+{
+	{ 0x143, 0x0001fffe, 1, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x146, 0x0001fffe, 1, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x146, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x149, 0x0001fffe, 1, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x149, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14c, 0x0001fffe, 1, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x98, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x98, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9b, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9e, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x101, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x101, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x107, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x107, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10a, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10d, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8f, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x92, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x92, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x95, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x95, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x152, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x152, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x155, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x155, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x158, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x158, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x113, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x113, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x116, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x116, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x119, 0x0001fffe, 1, 0x8, SISLANDS_CACCONFIG_CGIND },
+	{ 0x119, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x122, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x122, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x125, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x125, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x128, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x128, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12b, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+
+};
+
+static const struct si_cac_config_reg cac_override_tahiti[] =
+{
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_powertune_data powertune_data_tahiti =
+{
+	((1 << 16) | 27027),
+	6,
+	0,
+	4,
+	95,
+	{
+		0UL,
+		0UL,
+		4521550UL,
+		309631529UL,
+		-1270850L,
+		4513710L,
+		40
+	},
+	595000000UL,
+	12,
+	{
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0
+	},
+	true
+};
+
+static const struct si_dte_data dte_data_tahiti =
+{
+	{ 1159409, 0, 0, 0, 0 },
+	{ 777, 0, 0, 0, 0 },
+	2,
+	54000,
+	127000,
+	25,
+	2,
+	10,
+	13,
+	{ 27, 31, 35, 39, 43, 47, 54, 61, 67, 74, 81, 88, 95, 0, 0, 0 },
+	{ 240888759, 221057860, 235370597, 162287531, 158510299, 131423027, 116673180, 103067515, 87941937, 76209048, 68209175, 64090048, 58301890, 0, 0, 0 },
+	{ 12024, 11189, 11451, 8411, 7939, 6666, 5681, 4905, 4241, 3720, 3354, 3122, 2890, 0, 0, 0 },
+	85,
+	false
+};
+
+static const struct si_dte_data dte_data_tahiti_le =
+{
+	{ 0x1E8480, 0x7A1200, 0x2160EC0, 0x3938700, 0 },
+	{ 0x7D, 0x7D, 0x4E4, 0xB00, 0 },
+	0x5,
+	0xAFC8,
+	0x64,
+	0x32,
+	1,
+	0,
+	0x10,
+	{ 0x78, 0x7C, 0x82, 0x88, 0x8E, 0x94, 0x9A, 0xA0, 0xA6, 0xAC, 0xB0, 0xB4, 0xB8, 0xBC, 0xC0, 0xC4 },
+	{ 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700, 0x3938700 },
+	{ 0x2AF8, 0x2AF8, 0x29BB, 0x27F9, 0x2637, 0x2475, 0x22B3, 0x20F1, 0x1F2F, 0x1D6D, 0x1734, 0x1414, 0x10F4, 0xDD4, 0xAB4, 0x794 },
+	85,
+	true
+};
+
+static const struct si_dte_data dte_data_tahiti_pro =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	45000,
+	100,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0x7D0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_new_zealand =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0 },
+	{ 0x29B, 0x3E9, 0x537, 0x7D2, 0 },
+	0x5,
+	0xAFC8,
+	0x69,
+	0x32,
+	1,
+	0,
+	0x10,
+	{ 0x82, 0xA0, 0xB4, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE, 0xFE },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0xDAC, 0x1388, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685, 0x685 },
+	85,
+	true
+};
+
+static const struct si_dte_data dte_data_aruba_pro =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	45000,
+	100,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0x1000, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_malta =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	45000,
+	100,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+struct si_cac_config_reg cac_weights_pitcairn[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x8a, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x24d, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x19, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x76, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0xc11, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x7f3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x403, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x367, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x4c9, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x45d, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x36d, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x534, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x5da, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x880, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0x201, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x9f, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x1f, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5de, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x7b, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x13, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0xf9, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x66, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x13, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x186, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg lcac_pitcairn[] =
+{
+	{ 0x98, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x98, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9b, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x107, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x107, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x113, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x113, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x152, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x152, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8f, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x146, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x146, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9e, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10a, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x116, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x116, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x155, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x155, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x92, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x92, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x149, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x149, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x101, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x101, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10d, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x119, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x119, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x158, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x158, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x95, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x95, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x122, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x122, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x125, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x125, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x128, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x128, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12b, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_override_pitcairn[] =
+{
+    { 0xFFFFFFFF }
+};
+
+static const struct si_powertune_data powertune_data_pitcairn =
+{
+	((1 << 16) | 27027),
+	5,
+	0,
+	6,
+	100,
+	{
+		51600000UL,
+		1800000UL,
+		7194395UL,
+		309631529UL,
+		-1270850L,
+		4513710L,
+		100
+	},
+	117830498UL,
+	12,
+	{
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0
+	},
+	true
+};
+
+static const struct si_dte_data dte_data_pitcairn =
+{
+	{ 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0 },
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	0,
+	false
+};
+
+static const struct si_dte_data dte_data_curacao_xt =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	45000,
+	100,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0x1D17, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_curacao_pro =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	45000,
+	100,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0x1D17, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_neptune_xt =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	45000,
+	100,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0x3A2F, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_cac_config_reg cac_weights_chelsea_pro[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x82, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x153, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x52, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0xAC, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0xBE, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x110, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x4CD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x37, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x27, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0xC3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x28, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x26C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3B2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x99D, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA3F, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x34, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x2BD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x100, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_chelsea_xt[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x82, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x153, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x52, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0xAC, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0xBE, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x110, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x4CD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x37, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x27, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0xC3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x28, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x26C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3B2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x99D, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA3F, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x34, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x30A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x100, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_heathrow[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x82, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x153, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x52, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0xAC, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0xBE, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x110, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x4CD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x37, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x27, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0xC3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x28, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x26C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3B2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x99D, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA3F, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x34, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x362, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x100, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_cape_verde_pro[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x82, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x153, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x52, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0xAC, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0xBE, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x110, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x4CD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x37, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x27, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0xC3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x28, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x26C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3B2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x99D, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA3F, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x34, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x315, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x100, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_cape_verde[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x82, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x153, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x52, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0xAC, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0xBE, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x110, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x4CD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x37, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x27, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0xC3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x28, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x26C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3B2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x99D, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA3F, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x34, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x3BA, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x100, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg lcac_cape_verde[] =
+{
+	{ 0x98, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x98, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9b, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x107, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x107, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x113, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x113, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x152, 0x0001fffe, 1, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x152, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8f, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x146, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x146, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_override_cape_verde[] =
+{
+    { 0xFFFFFFFF }
+};
+
+static const struct si_powertune_data powertune_data_cape_verde =
+{
+	((1 << 16) | 0x6993),
+	5,
+	0,
+	7,
+	105,
+	{
+		0UL,
+		0UL,
+		7194395UL,
+		309631529UL,
+		-1270850L,
+		4513710L,
+		100
+	},
+	117830498UL,
+	12,
+	{
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0
+	},
+	true
+};
+
+static const struct si_dte_data dte_data_cape_verde =
+{
+	{ 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0 },
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	0,
+	false
+};
+
+static const struct si_dte_data dte_data_venus_xtx =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x71C, 0xAAB, 0xE39, 0x11C7, 0x0 },
+	5,
+	55000,
+	0x69,
+	0xA,
+	1,
+	0,
+	0x3,
+	{ 0x96, 0xB4, 0xFF, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	{ 0x895440, 0x3D0900, 0x989680, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	{ 0xD6D8, 0x88B8, 0x1555, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_venus_xt =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0xBDA, 0x11C7, 0x17B4, 0x1DA1, 0x0 },
+	5,
+	55000,
+	0x69,
+	0xA,
+	1,
+	0,
+	0x3,
+	{ 0x96, 0xB4, 0xFF, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	{ 0x895440, 0x3D0900, 0x989680, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	{ 0xAFC8, 0x88B8, 0x238E, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_venus_pro =
+{
+	{  0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x11C7, 0x1AAB, 0x238E, 0x2C72, 0x0 },
+	5,
+	55000,
+	0x69,
+	0xA,
+	1,
+	0,
+	0x3,
+	{ 0x96, 0xB4, 0xFF, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	{ 0x895440, 0x3D0900, 0x989680, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	{ 0x88B8, 0x88B8, 0x3555, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+struct si_cac_config_reg cac_weights_oland[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x82, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x153, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x52, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x4F, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x135, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0xAC, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x118, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0xBE, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x110, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x4CD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x37, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x27, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0xC3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x35, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0x28, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x26C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3B2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x99D, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA3F, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0xA, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x34, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x3BA, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x30, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x100, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_mars_pro[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x43, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0xAF, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x2A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x59, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x1A5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x1D6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x2A3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x8FD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x76, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x8A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0xA3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x71, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0xA6, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x81, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3D2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x27C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA96, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0xB, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x32, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7E, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0x280, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0x3C, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0x203, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0xB4, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_mars_xt[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x43, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0xAF, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x2A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x59, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x1A5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x1D6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x2A3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x8FD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x76, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x8A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0xA3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x71, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0xA6, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x81, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3D2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x27C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA96, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0xB, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x60, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x32, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7E, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0x280, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0x3C, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0x203, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0xB4, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_oland_pro[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x43, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0xAF, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x2A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x59, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x1A5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x1D6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x2A3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x8FD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x76, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x8A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0xA3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x71, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0xA6, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x81, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3D2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x27C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA96, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0xB, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x90, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x32, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7E, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0x280, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0x3C, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0x203, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0xB4, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_weights_oland_xt[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x43, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0xAF, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x2A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x29, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0xA0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x59, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x1A5, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x1D6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0x2A3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x8FD, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x76, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x8A, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0xA3, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x71, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0xA6, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x81, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0x3D2, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0x27C, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xA96, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x5, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0xB, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x15, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x36, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x10, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x120, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x32, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x7E, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0x280, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0x7, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0x3C, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0x203, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0xB4, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg lcac_oland[] =
+{
+	{ 0x98, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x98, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x0001fffe, 1, 0x4, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg lcac_mars_pro[] =
+{
+	{ 0x98, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x98, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x104, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x110, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x0001fffe, 1, 0x6, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x143, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x164, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x167, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16a, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15e, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x161, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15b, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x0001fffe, 1, 0x2, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16d, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x170, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x173, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x176, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x179, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17c, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x0001fffe, 1, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17f, 0x00000001, 0, 0x1, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_cac_config_reg cac_override_oland[] =
+{
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_powertune_data powertune_data_oland =
+{
+	((1 << 16) | 0x6993),
+	5,
+	0,
+	7,
+	105,
+	{
+		0UL,
+		0UL,
+		7194395UL,
+		309631529UL,
+		-1270850L,
+		4513710L,
+		100
+	},
+	117830498UL,
+	12,
+	{
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0
+	},
+	true
+};
+
+static const struct si_powertune_data powertune_data_mars_pro =
+{
+	((1 << 16) | 0x6993),
+	5,
+	0,
+	7,
+	105,
+	{
+		0UL,
+		0UL,
+		7194395UL,
+		309631529UL,
+		-1270850L,
+		4513710L,
+		100
+	},
+	117830498UL,
+	12,
+	{
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0
+	},
+	true
+};
+
+static const struct si_dte_data dte_data_oland =
+{
+	{ 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0 },
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	0,
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 },
+	0,
+	false
+};
+
+static const struct si_dte_data dte_data_mars_pro =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	55000,
+	105,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0xF627, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+static const struct si_dte_data dte_data_sun_xt =
+{
+	{ 0x1E8480, 0x3D0900, 0x989680, 0x2625A00, 0x0 },
+	{ 0x0, 0x0, 0x0, 0x0, 0x0 },
+	5,
+	55000,
+	105,
+	0xA,
+	1,
+	0,
+	0x10,
+	{ 0x96, 0xB4, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF },
+	{ 0x895440, 0x3D0900, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680, 0x989680 },
+	{ 0xD555, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0 },
+	90,
+	true
+};
+
+
+static const struct si_cac_config_reg cac_weights_hainan[] =
+{
+	{ 0x0, 0x0000ffff, 0, 0x2d9, SISLANDS_CACCONFIG_CGIND },
+	{ 0x0, 0xffff0000, 16, 0x22b, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0x0000ffff, 0, 0x21c, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1, 0xffff0000, 16, 0x1dc, SISLANDS_CACCONFIG_CGIND },
+	{ 0x2, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0x0000ffff, 0, 0x24e, SISLANDS_CACCONFIG_CGIND },
+	{ 0x3, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0x0000ffff, 0, 0x35e, SISLANDS_CACCONFIG_CGIND },
+	{ 0x5, 0xffff0000, 16, 0x1143, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0x0000ffff, 0, 0xe17, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6, 0xffff0000, 16, 0x441, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18f, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0x0000ffff, 0, 0x28b, SISLANDS_CACCONFIG_CGIND },
+	{ 0x7, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x8, 0xffff0000, 16, 0xabe, SISLANDS_CACCONFIG_CGIND },
+	{ 0x9, 0x0000ffff, 0, 0xf11, SISLANDS_CACCONFIG_CGIND },
+	{ 0xa, 0x0000ffff, 0, 0x907, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0x0000ffff, 0, 0xb45, SISLANDS_CACCONFIG_CGIND },
+	{ 0xb, 0xffff0000, 16, 0xd1e, SISLANDS_CACCONFIG_CGIND },
+	{ 0xc, 0x0000ffff, 0, 0xa2c, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0x0000ffff, 0, 0x62, SISLANDS_CACCONFIG_CGIND },
+	{ 0xd, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0xe, 0x0000ffff, 0, 0x1f3, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0x0000ffff, 0, 0x42, SISLANDS_CACCONFIG_CGIND },
+	{ 0xf, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x10, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0x0000ffff, 0, 0x709, SISLANDS_CACCONFIG_CGIND },
+	{ 0x11, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x12, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x13, 0xffff0000, 16, 0x3a, SISLANDS_CACCONFIG_CGIND },
+	{ 0x14, 0x0000ffff, 0, 0x357, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0x0000ffff, 0, 0x9f, SISLANDS_CACCONFIG_CGIND },
+	{ 0x15, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x4e, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0x0000ffff, 0, 0x314, SISLANDS_CACCONFIG_CGIND },
+	{ 0x16, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x17, 0x0000ffff, 0, 0x6d, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x18, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0x0000ffff, 0, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x19, 0xffff0000, 16, 0x0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1a, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1b, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1c, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1d, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1e, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x1f, 0xffff0000, 16, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x20, 0x0000ffff, 0, 0, SISLANDS_CACCONFIG_CGIND },
+	{ 0x6d, 0x0000ffff, 0, 0x1b9, SISLANDS_CACCONFIG_CGIND },
+	{ 0xFFFFFFFF }
+};
+
+static const struct si_powertune_data powertune_data_hainan =
+{
+	((1 << 16) | 0x6993),
+	5,
+	0,
+	9,
+	105,
+	{
+		0UL,
+		0UL,
+		7194395UL,
+		309631529UL,
+		-1270850L,
+		4513710L,
+		100
+	},
+	117830498UL,
+	12,
+	{
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0,
+		0
+	},
+	true
+};
+
+struct rv7xx_power_info *rv770_get_pi(struct radeon_device *rdev);
+struct evergreen_power_info *evergreen_get_pi(struct radeon_device *rdev);
+struct ni_power_info *ni_get_pi(struct radeon_device *rdev);
+struct ni_ps *ni_get_ps(struct radeon_ps *rps);
+
+static int si_populate_voltage_value(struct radeon_device *rdev,
+				     const struct atom_voltage_table *table,
+				     u16 value, SISLANDS_SMC_VOLTAGE_VALUE *voltage);
+static int si_get_std_voltage_value(struct radeon_device *rdev,
+				    SISLANDS_SMC_VOLTAGE_VALUE *voltage,
+				    u16 *std_voltage);
+static int si_write_smc_soft_register(struct radeon_device *rdev,
+				      u16 reg_offset, u32 value);
+static int si_convert_power_level_to_smc(struct radeon_device *rdev,
+					 struct rv7xx_pl *pl,
+					 SISLANDS_SMC_HW_PERFORMANCE_LEVEL *level);
+static int si_calculate_sclk_params(struct radeon_device *rdev,
+				    u32 engine_clock,
+				    SISLANDS_SMC_SCLK_VALUE *sclk);
+
+static struct si_power_info *si_get_pi(struct radeon_device *rdev)
+{
+        struct si_power_info *pi = rdev->pm.dpm.priv;
+
+        return pi;
+}
+
+static void si_calculate_leakage_for_v_and_t_formula(const struct ni_leakage_coeffients *coeff,
+						     u16 v, s32 t, u32 ileakage, u32 *leakage)
+{
+	s64 kt, kv, leakage_w, i_leakage, vddc;
+	s64 temperature, t_slope, t_intercept, av, bv, t_ref;
+
+	i_leakage = drm_int2fixp(ileakage / 100);
+	vddc = div64_s64(drm_int2fixp(v), 1000);
+	temperature = div64_s64(drm_int2fixp(t), 1000);
+
+	t_slope = div64_s64(drm_int2fixp(coeff->t_slope), 100000000);
+	t_intercept = div64_s64(drm_int2fixp(coeff->t_intercept), 100000000);
+	av = div64_s64(drm_int2fixp(coeff->av), 100000000);
+	bv = div64_s64(drm_int2fixp(coeff->bv), 100000000);
+	t_ref = drm_int2fixp(coeff->t_ref);
+
+	kt = drm_fixp_div(drm_fixp_exp(drm_fixp_mul(drm_fixp_mul(t_slope, vddc) + t_intercept, temperature)),
+			  drm_fixp_exp(drm_fixp_mul(drm_fixp_mul(t_slope, vddc) + t_intercept, t_ref)));
+	kv = drm_fixp_mul(av, drm_fixp_exp(drm_fixp_mul(bv, vddc)));
+
+	leakage_w = drm_fixp_mul(drm_fixp_mul(drm_fixp_mul(i_leakage, kt), kv), vddc);
+
+	*leakage = drm_fixp2int(leakage_w * 1000);
+}
+
+static void si_calculate_leakage_for_v_and_t(struct radeon_device *rdev,
+					     const struct ni_leakage_coeffients *coeff,
+					     u16 v,
+					     s32 t,
+					     u32 i_leakage,
+					     u32 *leakage)
+{
+	si_calculate_leakage_for_v_and_t_formula(coeff, v, t, i_leakage, leakage);
+}
+
+static void si_calculate_leakage_for_v_formula(const struct ni_leakage_coeffients *coeff,
+					       const u32 fixed_kt, u16 v,
+					       u32 ileakage, u32 *leakage)
+{
+	s64 kt, kv, leakage_w, i_leakage, vddc;
+
+	i_leakage = div64_s64(drm_int2fixp(ileakage), 100);
+	vddc = div64_s64(drm_int2fixp(v), 1000);
+
+	kt = div64_s64(drm_int2fixp(fixed_kt), 100000000);
+	kv = drm_fixp_mul(div64_s64(drm_int2fixp(coeff->av), 100000000),
+			  drm_fixp_exp(drm_fixp_mul(div64_s64(drm_int2fixp(coeff->bv), 100000000), vddc)));
+
+	leakage_w = drm_fixp_mul(drm_fixp_mul(drm_fixp_mul(i_leakage, kt), kv), vddc);
+
+	*leakage = drm_fixp2int(leakage_w * 1000);
+}
+
+static void si_calculate_leakage_for_v(struct radeon_device *rdev,
+				       const struct ni_leakage_coeffients *coeff,
+				       const u32 fixed_kt,
+				       u16 v,
+				       u32 i_leakage,
+				       u32 *leakage)
+{
+	si_calculate_leakage_for_v_formula(coeff, fixed_kt, v, i_leakage, leakage);
+}
+
+
+static void si_update_dte_from_pl2(struct radeon_device *rdev,
+				   struct si_dte_data *dte_data)
+{
+	u32 p_limit1 = rdev->pm.dpm.tdp_limit;
+	u32 p_limit2 = rdev->pm.dpm.near_tdp_limit;
+	u32 k = dte_data->k;
+	u32 t_max = dte_data->max_t;
+	u32 t_split[5] = { 10, 15, 20, 25, 30 };
+	u32 t_0 = dte_data->t0;
+	u32 i;
+
+	if (p_limit2 != 0 && p_limit2 <= p_limit1) {
+		dte_data->tdep_count = 3;
+
+		for (i = 0; i < k; i++) {
+			dte_data->r[i] =
+				(t_split[i] * (t_max - t_0/(u32)1000) * (1 << 14)) /
+				(p_limit2  * (u32)100);
+		}
+
+		dte_data->tdep_r[1] = dte_data->r[4] * 2;
+
+		for (i = 2; i < SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE; i++) {
+			dte_data->tdep_r[i] = dte_data->r[4];
+		}
+	} else {
+		DRM_ERROR("Invalid PL2! DTE will not be updated.\n");
+	}
+}
+
+static void si_initialize_powertune_defaults(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	bool update_dte_from_pl2 = false;
+
+	if (rdev->family == CHIP_TAHITI) {
+		si_pi->cac_weights = cac_weights_tahiti;
+		si_pi->lcac_config = lcac_tahiti;
+		si_pi->cac_override = cac_override_tahiti;
+		si_pi->powertune_data = &powertune_data_tahiti;
+		si_pi->dte_data = dte_data_tahiti;
+
+		switch (rdev->pdev->device) {
+		case 0x6798:
+			si_pi->dte_data.enable_dte_by_default = true;
+			break;
+		case 0x6799:
+			si_pi->dte_data = dte_data_new_zealand;
+			break;
+		case 0x6790:
+		case 0x6791:
+		case 0x6792:
+		case 0x679E:
+			si_pi->dte_data = dte_data_aruba_pro;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x679B:
+			si_pi->dte_data = dte_data_malta;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x679A:
+			si_pi->dte_data = dte_data_tahiti_pro;
+			update_dte_from_pl2 = true;
+			break;
+		default:
+			if (si_pi->dte_data.enable_dte_by_default == true)
+				DRM_ERROR("DTE is not enabled!\n");
+			break;
+		}
+	} else if (rdev->family == CHIP_PITCAIRN) {
+		switch (rdev->pdev->device) {
+		case 0x6810:
+		case 0x6818:
+			si_pi->cac_weights = cac_weights_pitcairn;
+			si_pi->lcac_config = lcac_pitcairn;
+			si_pi->cac_override = cac_override_pitcairn;
+			si_pi->powertune_data = &powertune_data_pitcairn;
+			si_pi->dte_data = dte_data_curacao_xt;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x6819:
+		case 0x6811:
+			si_pi->cac_weights = cac_weights_pitcairn;
+			si_pi->lcac_config = lcac_pitcairn;
+			si_pi->cac_override = cac_override_pitcairn;
+			si_pi->powertune_data = &powertune_data_pitcairn;
+			si_pi->dte_data = dte_data_curacao_pro;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x6800:
+		case 0x6806:
+			si_pi->cac_weights = cac_weights_pitcairn;
+			si_pi->lcac_config = lcac_pitcairn;
+			si_pi->cac_override = cac_override_pitcairn;
+			si_pi->powertune_data = &powertune_data_pitcairn;
+			si_pi->dte_data = dte_data_neptune_xt;
+			update_dte_from_pl2 = true;
+			break;
+		default:
+			si_pi->cac_weights = cac_weights_pitcairn;
+			si_pi->lcac_config = lcac_pitcairn;
+			si_pi->cac_override = cac_override_pitcairn;
+			si_pi->powertune_data = &powertune_data_pitcairn;
+			si_pi->dte_data = dte_data_pitcairn;
+		}
+	} else if (rdev->family == CHIP_VERDE) {
+		si_pi->lcac_config = lcac_cape_verde;
+		si_pi->cac_override = cac_override_cape_verde;
+		si_pi->powertune_data = &powertune_data_cape_verde;
+
+		switch (rdev->pdev->device) {
+		case 0x683B:
+		case 0x683F:
+		case 0x6829:
+			si_pi->cac_weights = cac_weights_cape_verde_pro;
+			si_pi->dte_data = dte_data_cape_verde;
+			break;
+		case 0x6825:
+		case 0x6827:
+			si_pi->cac_weights = cac_weights_heathrow;
+			si_pi->dte_data = dte_data_cape_verde;
+			break;
+		case 0x6824:
+		case 0x682D:
+			si_pi->cac_weights = cac_weights_chelsea_xt;
+			si_pi->dte_data = dte_data_cape_verde;
+			break;
+		case 0x682F:
+			si_pi->cac_weights = cac_weights_chelsea_pro;
+			si_pi->dte_data = dte_data_cape_verde;
+			break;
+		case 0x6820:
+			si_pi->cac_weights = cac_weights_heathrow;
+			si_pi->dte_data = dte_data_venus_xtx;
+			break;
+		case 0x6821:
+			si_pi->cac_weights = cac_weights_heathrow;
+			si_pi->dte_data = dte_data_venus_xt;
+			break;
+		case 0x6823:
+			si_pi->cac_weights = cac_weights_chelsea_pro;
+			si_pi->dte_data = dte_data_venus_pro;
+			break;
+		case 0x682B:
+			si_pi->cac_weights = cac_weights_chelsea_pro;
+			si_pi->dte_data = dte_data_venus_pro;
+			break;
+		default:
+			si_pi->cac_weights = cac_weights_cape_verde;
+			si_pi->dte_data = dte_data_cape_verde;
+			break;
+		}
+	} else if (rdev->family == CHIP_OLAND) {
+		switch (rdev->pdev->device) {
+		case 0x6601:
+		case 0x6621:
+		case 0x6603:
+			si_pi->cac_weights = cac_weights_mars_pro;
+			si_pi->lcac_config = lcac_mars_pro;
+			si_pi->cac_override = cac_override_oland;
+			si_pi->powertune_data = &powertune_data_mars_pro;
+			si_pi->dte_data = dte_data_mars_pro;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x6600:
+		case 0x6606:
+		case 0x6620:
+			si_pi->cac_weights = cac_weights_mars_xt;
+			si_pi->lcac_config = lcac_mars_pro;
+			si_pi->cac_override = cac_override_oland;
+			si_pi->powertune_data = &powertune_data_mars_pro;
+			si_pi->dte_data = dte_data_mars_pro;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x6611:
+			si_pi->cac_weights = cac_weights_oland_pro;
+			si_pi->lcac_config = lcac_mars_pro;
+			si_pi->cac_override = cac_override_oland;
+			si_pi->powertune_data = &powertune_data_mars_pro;
+			si_pi->dte_data = dte_data_mars_pro;
+			update_dte_from_pl2 = true;
+			break;
+		case 0x6610:
+			si_pi->cac_weights = cac_weights_oland_xt;
+			si_pi->lcac_config = lcac_mars_pro;
+			si_pi->cac_override = cac_override_oland;
+			si_pi->powertune_data = &powertune_data_mars_pro;
+			si_pi->dte_data = dte_data_mars_pro;
+			update_dte_from_pl2 = true;
+			break;
+		default:
+			si_pi->cac_weights = cac_weights_oland;
+			si_pi->lcac_config = lcac_oland;
+			si_pi->cac_override = cac_override_oland;
+			si_pi->powertune_data = &powertune_data_oland;
+			si_pi->dte_data = dte_data_oland;
+			break;
+		}
+	} else if (rdev->family == CHIP_HAINAN) {
+		si_pi->cac_weights = cac_weights_hainan;
+		si_pi->lcac_config = lcac_oland;
+		si_pi->cac_override = cac_override_oland;
+		si_pi->powertune_data = &powertune_data_hainan;
+		si_pi->dte_data = dte_data_sun_xt;
+		update_dte_from_pl2 = true;
+	} else {
+		DRM_ERROR("Unknown SI asic revision, failed to initialize PowerTune!\n");
+		return;
+	}
+
+	ni_pi->enable_power_containment = false;
+	ni_pi->enable_cac = false;
+	ni_pi->enable_sq_ramping = false;
+	si_pi->enable_dte = false;
+
+	if (si_pi->powertune_data->enable_powertune_by_default) {
+		ni_pi->enable_power_containment= true;
+		ni_pi->enable_cac = true;
+		if (si_pi->dte_data.enable_dte_by_default) {
+			si_pi->enable_dte = true;
+			if (update_dte_from_pl2)
+				si_update_dte_from_pl2(rdev, &si_pi->dte_data);
+
+		}
+		ni_pi->enable_sq_ramping = true;
+	}
+
+	ni_pi->driver_calculate_cac_leakage = true;
+	ni_pi->cac_configuration_required = true;
+
+	if (ni_pi->cac_configuration_required) {
+		ni_pi->support_cac_long_term_average = true;
+		si_pi->dyn_powertune_data.l2_lta_window_size =
+			si_pi->powertune_data->l2_lta_window_size_default;
+		si_pi->dyn_powertune_data.lts_truncate =
+			si_pi->powertune_data->lts_truncate_default;
+	} else {
+		ni_pi->support_cac_long_term_average = false;
+		si_pi->dyn_powertune_data.l2_lta_window_size = 0;
+		si_pi->dyn_powertune_data.lts_truncate = 0;
+	}
+
+	si_pi->dyn_powertune_data.disable_uvd_powertune = false;
+}
+
+static u32 si_get_smc_power_scaling_factor(struct radeon_device *rdev)
+{
+	return 1;
+}
+
+static u32 si_calculate_cac_wintime(struct radeon_device *rdev)
+{
+	u32 xclk;
+	u32 wintime;
+	u32 cac_window;
+	u32 cac_window_size;
+
+	xclk = radeon_get_xclk(rdev);
+
+	if (xclk == 0)
+		return 0;
+
+	cac_window = RREG32(CG_CAC_CTRL) & CAC_WINDOW_MASK;
+	cac_window_size = ((cac_window & 0xFFFF0000) >> 16) * (cac_window & 0x0000FFFF);
+
+	wintime = (cac_window_size * 100) / xclk;
+
+	return wintime;
+}
+
+static u32 si_scale_power_for_smc(u32 power_in_watts, u32 scaling_factor)
+{
+	return power_in_watts;
+}
+
+static int si_calculate_adjusted_tdp_limits(struct radeon_device *rdev,
+					    bool adjust_polarity,
+					    u32 tdp_adjustment,
+					    u32 *tdp_limit,
+					    u32 *near_tdp_limit)
+{
+	u32 adjustment_delta, max_tdp_limit;
+
+	if (tdp_adjustment > (u32)rdev->pm.dpm.tdp_od_limit)
+		return -EINVAL;
+
+	max_tdp_limit = ((100 + 100) * rdev->pm.dpm.tdp_limit) / 100;
+
+	if (adjust_polarity) {
+		*tdp_limit = ((100 + tdp_adjustment) * rdev->pm.dpm.tdp_limit) / 100;
+		*near_tdp_limit = rdev->pm.dpm.near_tdp_limit_adjusted + (*tdp_limit - rdev->pm.dpm.tdp_limit);
+	} else {
+		*tdp_limit = ((100 - tdp_adjustment) * rdev->pm.dpm.tdp_limit) / 100;
+		adjustment_delta  = rdev->pm.dpm.tdp_limit - *tdp_limit;
+		if (adjustment_delta < rdev->pm.dpm.near_tdp_limit_adjusted)
+			*near_tdp_limit = rdev->pm.dpm.near_tdp_limit_adjusted - adjustment_delta;
+		else
+			*near_tdp_limit = 0;
+	}
+
+	if ((*tdp_limit <= 0) || (*tdp_limit > max_tdp_limit))
+		return -EINVAL;
+	if ((*near_tdp_limit <= 0) || (*near_tdp_limit > *tdp_limit))
+		return -EINVAL;
+
+	return 0;
+}
+
+static int si_populate_smc_tdp_limits(struct radeon_device *rdev,
+				      struct radeon_ps *radeon_state)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	if (ni_pi->enable_power_containment) {
+		SISLANDS_SMC_STATETABLE *smc_table = &si_pi->smc_statetable;
+		PP_SIslands_PAPMParameters *papm_parm;
+		struct radeon_ppm_table *ppm = rdev->pm.dpm.dyn_state.ppm_table;
+		u32 scaling_factor = si_get_smc_power_scaling_factor(rdev);
+		u32 tdp_limit;
+		u32 near_tdp_limit;
+		int ret;
+
+		if (scaling_factor == 0)
+			return -EINVAL;
+
+		memset(smc_table, 0, sizeof(SISLANDS_SMC_STATETABLE));
+
+		ret = si_calculate_adjusted_tdp_limits(rdev,
+						       false, /* ??? */
+						       rdev->pm.dpm.tdp_adjustment,
+						       &tdp_limit,
+						       &near_tdp_limit);
+		if (ret)
+			return ret;
+
+		smc_table->dpm2Params.TDPLimit =
+			cpu_to_be32(si_scale_power_for_smc(tdp_limit, scaling_factor) * 1000);
+		smc_table->dpm2Params.NearTDPLimit =
+			cpu_to_be32(si_scale_power_for_smc(near_tdp_limit, scaling_factor) * 1000);
+		smc_table->dpm2Params.SafePowerLimit =
+			cpu_to_be32(si_scale_power_for_smc((near_tdp_limit * SISLANDS_DPM2_TDP_SAFE_LIMIT_PERCENT) / 100, scaling_factor) * 1000);
+
+		ret = si_copy_bytes_to_smc(rdev,
+					   (si_pi->state_table_start + offsetof(SISLANDS_SMC_STATETABLE, dpm2Params) +
+						 offsetof(PP_SIslands_DPM2Parameters, TDPLimit)),
+					   (u8 *)(&(smc_table->dpm2Params.TDPLimit)),
+					   sizeof(u32) * 3,
+					   si_pi->sram_end);
+		if (ret)
+			return ret;
+
+		if (si_pi->enable_ppm) {
+			papm_parm = &si_pi->papm_parm;
+			memset(papm_parm, 0, sizeof(PP_SIslands_PAPMParameters));
+			papm_parm->NearTDPLimitTherm = cpu_to_be32(ppm->dgpu_tdp);
+			papm_parm->dGPU_T_Limit = cpu_to_be32(ppm->tj_max);
+			papm_parm->dGPU_T_Warning = cpu_to_be32(95);
+			papm_parm->dGPU_T_Hysteresis = cpu_to_be32(5);
+			papm_parm->PlatformPowerLimit = 0xffffffff;
+			papm_parm->NearTDPLimitPAPM = 0xffffffff;
+
+			ret = si_copy_bytes_to_smc(rdev, si_pi->papm_cfg_table_start,
+						   (u8 *)papm_parm,
+						   sizeof(PP_SIslands_PAPMParameters),
+						   si_pi->sram_end);
+			if (ret)
+				return ret;
+		}
+	}
+	return 0;
+}
+
+static int si_populate_smc_tdp_limits_2(struct radeon_device *rdev,
+					struct radeon_ps *radeon_state)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	if (ni_pi->enable_power_containment) {
+		SISLANDS_SMC_STATETABLE *smc_table = &si_pi->smc_statetable;
+		u32 scaling_factor = si_get_smc_power_scaling_factor(rdev);
+		int ret;
+
+		memset(smc_table, 0, sizeof(SISLANDS_SMC_STATETABLE));
+
+		smc_table->dpm2Params.NearTDPLimit =
+			cpu_to_be32(si_scale_power_for_smc(rdev->pm.dpm.near_tdp_limit_adjusted, scaling_factor) * 1000);
+		smc_table->dpm2Params.SafePowerLimit =
+			cpu_to_be32(si_scale_power_for_smc((rdev->pm.dpm.near_tdp_limit_adjusted * SISLANDS_DPM2_TDP_SAFE_LIMIT_PERCENT) / 100, scaling_factor) * 1000);
+
+		ret = si_copy_bytes_to_smc(rdev,
+					   (si_pi->state_table_start +
+					    offsetof(SISLANDS_SMC_STATETABLE, dpm2Params) +
+					    offsetof(PP_SIslands_DPM2Parameters, NearTDPLimit)),
+					   (u8 *)(&(smc_table->dpm2Params.NearTDPLimit)),
+					   sizeof(u32) * 2,
+					   si_pi->sram_end);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static u16 si_calculate_power_efficiency_ratio(struct radeon_device *rdev,
+					       const u16 prev_std_vddc,
+					       const u16 curr_std_vddc)
+{
+	u64 margin = (u64)SISLANDS_DPM2_PWREFFICIENCYRATIO_MARGIN;
+	u64 prev_vddc = (u64)prev_std_vddc;
+	u64 curr_vddc = (u64)curr_std_vddc;
+	u64 pwr_efficiency_ratio, n, d;
+
+	if ((prev_vddc == 0) || (curr_vddc == 0))
+		return 0;
+
+	n = div64_u64((u64)1024 * curr_vddc * curr_vddc * ((u64)1000 + margin), (u64)1000);
+	d = prev_vddc * prev_vddc;
+	pwr_efficiency_ratio = div64_u64(n, d);
+
+	if (pwr_efficiency_ratio > (u64)0xFFFF)
+		return 0;
+
+	return (u16)pwr_efficiency_ratio;
+}
+
+static bool si_should_disable_uvd_powertune(struct radeon_device *rdev,
+					    struct radeon_ps *radeon_state)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	if (si_pi->dyn_powertune_data.disable_uvd_powertune &&
+	    radeon_state->vclk && radeon_state->dclk)
+		return true;
+
+	return false;
+}
+
+static int si_populate_power_containment_values(struct radeon_device *rdev,
+						struct radeon_ps *radeon_state,
+						SISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	SISLANDS_SMC_VOLTAGE_VALUE vddc;
+	u32 prev_sclk;
+	u32 max_sclk;
+	u32 min_sclk;
+	u16 prev_std_vddc;
+	u16 curr_std_vddc;
+	int i;
+	u16 pwr_efficiency_ratio;
+	u8 max_ps_percent;
+	bool disable_uvd_power_tune;
+	int ret;
+
+	if (ni_pi->enable_power_containment == false)
+		return 0;
+
+	if (state->performance_level_count == 0)
+		return -EINVAL;
+
+	if (smc_state->levelCount != state->performance_level_count)
+		return -EINVAL;
+
+	disable_uvd_power_tune = si_should_disable_uvd_powertune(rdev, radeon_state);
+
+	smc_state->levels[0].dpm2.MaxPS = 0;
+	smc_state->levels[0].dpm2.NearTDPDec = 0;
+	smc_state->levels[0].dpm2.AboveSafeInc = 0;
+	smc_state->levels[0].dpm2.BelowSafeInc = 0;
+	smc_state->levels[0].dpm2.PwrEfficiencyRatio = 0;
+
+	for (i = 1; i < state->performance_level_count; i++) {
+		prev_sclk = state->performance_levels[i-1].sclk;
+		max_sclk  = state->performance_levels[i].sclk;
+		if (i == 1)
+			max_ps_percent = SISLANDS_DPM2_MAXPS_PERCENT_M;
+		else
+			max_ps_percent = SISLANDS_DPM2_MAXPS_PERCENT_H;
+
+		if (prev_sclk > max_sclk)
+			return -EINVAL;
+
+		if ((max_ps_percent == 0) ||
+		    (prev_sclk == max_sclk) ||
+		    disable_uvd_power_tune) {
+			min_sclk = max_sclk;
+		} else if (i == 1) {
+			min_sclk = prev_sclk;
+		} else {
+			min_sclk = (prev_sclk * (u32)max_ps_percent) / 100;
+		}
+
+		if (min_sclk < state->performance_levels[0].sclk)
+			min_sclk = state->performance_levels[0].sclk;
+
+		if (min_sclk == 0)
+			return -EINVAL;
+
+		ret = si_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+						state->performance_levels[i-1].vddc, &vddc);
+		if (ret)
+			return ret;
+
+		ret = si_get_std_voltage_value(rdev, &vddc, &prev_std_vddc);
+		if (ret)
+			return ret;
+
+		ret = si_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+						state->performance_levels[i].vddc, &vddc);
+		if (ret)
+			return ret;
+
+		ret = si_get_std_voltage_value(rdev, &vddc, &curr_std_vddc);
+		if (ret)
+			return ret;
+
+		pwr_efficiency_ratio = si_calculate_power_efficiency_ratio(rdev,
+									   prev_std_vddc, curr_std_vddc);
+
+		smc_state->levels[i].dpm2.MaxPS = (u8)((SISLANDS_DPM2_MAX_PULSE_SKIP * (max_sclk - min_sclk)) / max_sclk);
+		smc_state->levels[i].dpm2.NearTDPDec = SISLANDS_DPM2_NEAR_TDP_DEC;
+		smc_state->levels[i].dpm2.AboveSafeInc = SISLANDS_DPM2_ABOVE_SAFE_INC;
+		smc_state->levels[i].dpm2.BelowSafeInc = SISLANDS_DPM2_BELOW_SAFE_INC;
+		smc_state->levels[i].dpm2.PwrEfficiencyRatio = cpu_to_be16(pwr_efficiency_ratio);
+	}
+
+	return 0;
+}
+
+static int si_populate_sq_ramping_values(struct radeon_device *rdev,
+					 struct radeon_ps *radeon_state,
+					 SISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	u32 sq_power_throttle, sq_power_throttle2;
+	bool enable_sq_ramping = ni_pi->enable_sq_ramping;
+	int i;
+
+	if (state->performance_level_count == 0)
+		return -EINVAL;
+
+	if (smc_state->levelCount != state->performance_level_count)
+		return -EINVAL;
+
+	if (rdev->pm.dpm.sq_ramping_threshold == 0)
+		return -EINVAL;
+
+	if (SISLANDS_DPM2_SQ_RAMP_MAX_POWER > (MAX_POWER_MASK >> MAX_POWER_SHIFT))
+		enable_sq_ramping = false;
+
+	if (SISLANDS_DPM2_SQ_RAMP_MIN_POWER > (MIN_POWER_MASK >> MIN_POWER_SHIFT))
+		enable_sq_ramping = false;
+
+	if (SISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA > (MAX_POWER_DELTA_MASK >> MAX_POWER_DELTA_SHIFT))
+		enable_sq_ramping = false;
+
+	if (SISLANDS_DPM2_SQ_RAMP_STI_SIZE > (STI_SIZE_MASK >> STI_SIZE_SHIFT))
+		enable_sq_ramping = false;
+
+	if (NISLANDS_DPM2_SQ_RAMP_LTI_RATIO <= (LTI_RATIO_MASK >> LTI_RATIO_SHIFT))
+		enable_sq_ramping = false;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		sq_power_throttle = 0;
+		sq_power_throttle2 = 0;
+
+		if ((state->performance_levels[i].sclk >= rdev->pm.dpm.sq_ramping_threshold) &&
+		    enable_sq_ramping) {
+			sq_power_throttle |= MAX_POWER(SISLANDS_DPM2_SQ_RAMP_MAX_POWER);
+			sq_power_throttle |= MIN_POWER(SISLANDS_DPM2_SQ_RAMP_MIN_POWER);
+			sq_power_throttle2 |= MAX_POWER_DELTA(SISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA);
+			sq_power_throttle2 |= STI_SIZE(SISLANDS_DPM2_SQ_RAMP_STI_SIZE);
+			sq_power_throttle2 |= LTI_RATIO(SISLANDS_DPM2_SQ_RAMP_LTI_RATIO);
+		} else {
+			sq_power_throttle |= MAX_POWER_MASK | MIN_POWER_MASK;
+			sq_power_throttle2 |= MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK;
+		}
+
+		smc_state->levels[i].SQPowerThrottle = cpu_to_be32(sq_power_throttle);
+		smc_state->levels[i].SQPowerThrottle_2 = cpu_to_be32(sq_power_throttle2);
+	}
+
+	return 0;
+}
+
+static int si_enable_power_containment(struct radeon_device *rdev,
+				       struct radeon_ps *radeon_new_state,
+				       bool enable)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	PPSMC_Result smc_result;
+	int ret = 0;
+
+	if (ni_pi->enable_power_containment) {
+		if (enable) {
+			if (!si_should_disable_uvd_powertune(rdev, radeon_new_state)) {
+				smc_result = si_send_msg_to_smc(rdev, PPSMC_TDPClampingActive);
+				if (smc_result != PPSMC_Result_OK) {
+					ret = -EINVAL;
+					ni_pi->pc_enabled = false;
+				} else {
+					ni_pi->pc_enabled = true;
+				}
+			}
+		} else {
+			smc_result = si_send_msg_to_smc(rdev, PPSMC_TDPClampingInactive);
+			if (smc_result != PPSMC_Result_OK)
+				ret = -EINVAL;
+			ni_pi->pc_enabled = false;
+		}
+	}
+
+	return ret;
+}
+
+static int si_initialize_smc_dte_tables(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	int ret = 0;
+	struct si_dte_data *dte_data = &si_pi->dte_data;
+	Smc_SIslands_DTE_Configuration *dte_tables = NULL;
+	u32 table_size;
+	u8 tdep_count;
+	u32 i;
+
+	if (dte_data == NULL)
+		si_pi->enable_dte = false;
+
+	if (si_pi->enable_dte == false)
+		return 0;
+
+	if (dte_data->k <= 0)
+		return -EINVAL;
+
+	dte_tables = kzalloc(sizeof(Smc_SIslands_DTE_Configuration), GFP_KERNEL);
+	if (dte_tables == NULL) {
+		si_pi->enable_dte = false;
+		return -ENOMEM;
+	}
+
+	table_size = dte_data->k;
+
+	if (table_size > SMC_SISLANDS_DTE_MAX_FILTER_STAGES)
+		table_size = SMC_SISLANDS_DTE_MAX_FILTER_STAGES;
+
+	tdep_count = dte_data->tdep_count;
+	if (tdep_count > SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE)
+		tdep_count = SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE;
+
+	dte_tables->K = cpu_to_be32(table_size);
+	dte_tables->T0 = cpu_to_be32(dte_data->t0);
+	dte_tables->MaxT = cpu_to_be32(dte_data->max_t);
+	dte_tables->WindowSize = dte_data->window_size;
+	dte_tables->temp_select = dte_data->temp_select;
+	dte_tables->DTE_mode = dte_data->dte_mode;
+	dte_tables->Tthreshold = cpu_to_be32(dte_data->t_threshold);
+
+	if (tdep_count > 0)
+		table_size--;
+
+	for (i = 0; i < table_size; i++) {
+		dte_tables->tau[i] = cpu_to_be32(dte_data->tau[i]);
+		dte_tables->R[i]   = cpu_to_be32(dte_data->r[i]);
+	}
+
+	dte_tables->Tdep_count = tdep_count;
+
+	for (i = 0; i < (u32)tdep_count; i++) {
+		dte_tables->T_limits[i] = dte_data->t_limits[i];
+		dte_tables->Tdep_tau[i] = cpu_to_be32(dte_data->tdep_tau[i]);
+		dte_tables->Tdep_R[i] = cpu_to_be32(dte_data->tdep_r[i]);
+	}
+
+	ret = si_copy_bytes_to_smc(rdev, si_pi->dte_table_start, (u8 *)dte_tables,
+				   sizeof(Smc_SIslands_DTE_Configuration), si_pi->sram_end);
+	kfree(dte_tables);
+
+	return ret;
+}
+
+static int si_get_cac_std_voltage_max_min(struct radeon_device *rdev,
+					  u16 *max, u16 *min)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct radeon_cac_leakage_table *table =
+		&rdev->pm.dpm.dyn_state.cac_leakage_table;
+	u32 i;
+	u32 v0_loadline;
+
+
+	if (table == NULL)
+		return -EINVAL;
+
+	*max = 0;
+	*min = 0xFFFF;
+
+	for (i = 0; i < table->count; i++) {
+		if (table->entries[i].vddc > *max)
+			*max = table->entries[i].vddc;
+		if (table->entries[i].vddc < *min)
+			*min = table->entries[i].vddc;
+	}
+
+	if (si_pi->powertune_data->lkge_lut_v0_percent > 100)
+		return -EINVAL;
+
+	v0_loadline = (*min) * (100 - si_pi->powertune_data->lkge_lut_v0_percent) / 100;
+
+	if (v0_loadline > 0xFFFFUL)
+		return -EINVAL;
+
+	*min = (u16)v0_loadline;
+
+	if ((*min > *max) || (*max == 0) || (*min == 0))
+		return -EINVAL;
+
+	return 0;
+}
+
+static u16 si_get_cac_std_voltage_step(u16 max, u16 min)
+{
+	return ((max - min) + (SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES - 1)) /
+		SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES;
+}
+
+static int si_init_dte_leakage_table(struct radeon_device *rdev,
+				     PP_SIslands_CacConfig *cac_tables,
+				     u16 vddc_max, u16 vddc_min, u16 vddc_step,
+				     u16 t0, u16 t_step)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 leakage;
+	unsigned int i, j;
+	s32 t;
+	u32 smc_leakage;
+	u32 scaling_factor;
+	u16 voltage;
+
+	scaling_factor = si_get_smc_power_scaling_factor(rdev);
+
+	for (i = 0; i < SMC_SISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES ; i++) {
+		t = (1000 * (i * t_step + t0));
+
+		for (j = 0; j < SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES; j++) {
+			voltage = vddc_max - (vddc_step * j);
+
+			si_calculate_leakage_for_v_and_t(rdev,
+							 &si_pi->powertune_data->leakage_coefficients,
+							 voltage,
+							 t,
+							 si_pi->dyn_powertune_data.cac_leakage,
+							 &leakage);
+
+			smc_leakage = si_scale_power_for_smc(leakage, scaling_factor) / 4;
+
+			if (smc_leakage > 0xFFFF)
+				smc_leakage = 0xFFFF;
+
+			cac_tables->cac_lkge_lut[i][SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES-1-j] =
+				cpu_to_be16((u16)smc_leakage);
+		}
+	}
+	return 0;
+}
+
+static int si_init_simplified_leakage_table(struct radeon_device *rdev,
+					    PP_SIslands_CacConfig *cac_tables,
+					    u16 vddc_max, u16 vddc_min, u16 vddc_step)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 leakage;
+	unsigned int i, j;
+	u32 smc_leakage;
+	u32 scaling_factor;
+	u16 voltage;
+
+	scaling_factor = si_get_smc_power_scaling_factor(rdev);
+
+	for (j = 0; j < SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES; j++) {
+		voltage = vddc_max - (vddc_step * j);
+
+		si_calculate_leakage_for_v(rdev,
+					   &si_pi->powertune_data->leakage_coefficients,
+					   si_pi->powertune_data->fixed_kt,
+					   voltage,
+					   si_pi->dyn_powertune_data.cac_leakage,
+					   &leakage);
+
+		smc_leakage = si_scale_power_for_smc(leakage, scaling_factor) / 4;
+
+		if (smc_leakage > 0xFFFF)
+			smc_leakage = 0xFFFF;
+
+		for (i = 0; i < SMC_SISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES ; i++)
+			cac_tables->cac_lkge_lut[i][SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES-1-j] =
+				cpu_to_be16((u16)smc_leakage);
+	}
+	return 0;
+}
+
+static int si_initialize_smc_cac_tables(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	PP_SIslands_CacConfig *cac_tables = NULL;
+	u16 vddc_max, vddc_min, vddc_step;
+	u16 t0, t_step;
+	u32 load_line_slope, reg;
+	int ret = 0;
+	u32 ticks_per_us = radeon_get_xclk(rdev) / 100;
+
+	if (ni_pi->enable_cac == false)
+		return 0;
+
+	cac_tables = kzalloc(sizeof(PP_SIslands_CacConfig), GFP_KERNEL);
+	if (!cac_tables)
+		return -ENOMEM;
+
+	reg = RREG32(CG_CAC_CTRL) & ~CAC_WINDOW_MASK;
+	reg |= CAC_WINDOW(si_pi->powertune_data->cac_window);
+	WREG32(CG_CAC_CTRL, reg);
+
+	si_pi->dyn_powertune_data.cac_leakage = rdev->pm.dpm.cac_leakage;
+	si_pi->dyn_powertune_data.dc_pwr_value =
+		si_pi->powertune_data->dc_cac[NISLANDS_DCCAC_LEVEL_0];
+	si_pi->dyn_powertune_data.wintime = si_calculate_cac_wintime(rdev);
+	si_pi->dyn_powertune_data.shift_n = si_pi->powertune_data->shift_n_default;
+
+	si_pi->dyn_powertune_data.leakage_minimum_temperature = 80 * 1000;
+
+	ret = si_get_cac_std_voltage_max_min(rdev, &vddc_max, &vddc_min);
+	if (ret)
+		goto done_free;
+
+	vddc_step = si_get_cac_std_voltage_step(vddc_max, vddc_min);
+	vddc_min = vddc_max - (vddc_step * (SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES - 1));
+	t_step = 4;
+	t0 = 60;
+
+	if (si_pi->enable_dte || ni_pi->driver_calculate_cac_leakage)
+		ret = si_init_dte_leakage_table(rdev, cac_tables,
+						vddc_max, vddc_min, vddc_step,
+						t0, t_step);
+	else
+		ret = si_init_simplified_leakage_table(rdev, cac_tables,
+						       vddc_max, vddc_min, vddc_step);
+	if (ret)
+		goto done_free;
+
+	load_line_slope = ((u32)rdev->pm.dpm.load_line_slope << SMC_SISLANDS_SCALE_R) / 100;
+
+	cac_tables->l2numWin_TDP = cpu_to_be32(si_pi->dyn_powertune_data.l2_lta_window_size);
+	cac_tables->lts_truncate_n = si_pi->dyn_powertune_data.lts_truncate;
+	cac_tables->SHIFT_N = si_pi->dyn_powertune_data.shift_n;
+	cac_tables->lkge_lut_V0 = cpu_to_be32((u32)vddc_min);
+	cac_tables->lkge_lut_Vstep = cpu_to_be32((u32)vddc_step);
+	cac_tables->R_LL = cpu_to_be32(load_line_slope);
+	cac_tables->WinTime = cpu_to_be32(si_pi->dyn_powertune_data.wintime);
+	cac_tables->calculation_repeats = cpu_to_be32(2);
+	cac_tables->dc_cac = cpu_to_be32(0);
+	cac_tables->log2_PG_LKG_SCALE = 12;
+	cac_tables->cac_temp = si_pi->powertune_data->operating_temp;
+	cac_tables->lkge_lut_T0 = cpu_to_be32((u32)t0);
+	cac_tables->lkge_lut_Tstep = cpu_to_be32((u32)t_step);
+
+	ret = si_copy_bytes_to_smc(rdev, si_pi->cac_table_start, (u8 *)cac_tables,
+				   sizeof(PP_SIslands_CacConfig), si_pi->sram_end);
+
+	if (ret)
+		goto done_free;
+
+	ret = si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_ticks_per_us, ticks_per_us);
+
+done_free:
+	if (ret) {
+		ni_pi->enable_cac = false;
+		ni_pi->enable_power_containment = false;
+	}
+
+	kfree(cac_tables);
+
+	return 0;
+}
+
+static int si_program_cac_config_registers(struct radeon_device *rdev,
+					   const struct si_cac_config_reg *cac_config_regs)
+{
+	const struct si_cac_config_reg *config_regs = cac_config_regs;
+	u32 data = 0, offset;
+
+	if (!config_regs)
+		return -EINVAL;
+
+	while (config_regs->offset != 0xFFFFFFFF) {
+		switch (config_regs->type) {
+		case SISLANDS_CACCONFIG_CGIND:
+			offset = SMC_CG_IND_START + config_regs->offset;
+			if (offset < SMC_CG_IND_END)
+				data = RREG32_SMC(offset);
+			break;
+		default:
+			data = RREG32(config_regs->offset << 2);
+			break;
+		}
+
+		data &= ~config_regs->mask;
+		data |= ((config_regs->value << config_regs->shift) & config_regs->mask);
+
+		switch (config_regs->type) {
+		case SISLANDS_CACCONFIG_CGIND:
+			offset = SMC_CG_IND_START + config_regs->offset;
+			if (offset < SMC_CG_IND_END)
+				WREG32_SMC(offset, data);
+			break;
+		default:
+			WREG32(config_regs->offset << 2, data);
+			break;
+		}
+		config_regs++;
+	}
+	return 0;
+}
+
+static int si_initialize_hardware_cac_manager(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	int ret;
+
+	if ((ni_pi->enable_cac == false) ||
+	    (ni_pi->cac_configuration_required == false))
+		return 0;
+
+	ret = si_program_cac_config_registers(rdev, si_pi->lcac_config);
+	if (ret)
+		return ret;
+	ret = si_program_cac_config_registers(rdev, si_pi->cac_override);
+	if (ret)
+		return ret;
+	ret = si_program_cac_config_registers(rdev, si_pi->cac_weights);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int si_enable_smc_cac(struct radeon_device *rdev,
+			     struct radeon_ps *radeon_new_state,
+			     bool enable)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	PPSMC_Result smc_result;
+	int ret = 0;
+
+	if (ni_pi->enable_cac) {
+		if (enable) {
+			if (!si_should_disable_uvd_powertune(rdev, radeon_new_state)) {
+				if (ni_pi->support_cac_long_term_average) {
+					smc_result = si_send_msg_to_smc(rdev, PPSMC_CACLongTermAvgEnable);
+					if (smc_result != PPSMC_Result_OK)
+						ni_pi->support_cac_long_term_average = false;
+				}
+
+				smc_result = si_send_msg_to_smc(rdev, PPSMC_MSG_EnableCac);
+				if (smc_result != PPSMC_Result_OK) {
+					ret = -EINVAL;
+					ni_pi->cac_enabled = false;
+				} else {
+					ni_pi->cac_enabled = true;
+				}
+
+				if (si_pi->enable_dte) {
+					smc_result = si_send_msg_to_smc(rdev, PPSMC_MSG_EnableDTE);
+					if (smc_result != PPSMC_Result_OK)
+						ret = -EINVAL;
+				}
+			}
+		} else if (ni_pi->cac_enabled) {
+			if (si_pi->enable_dte)
+				smc_result = si_send_msg_to_smc(rdev, PPSMC_MSG_DisableDTE);
+
+			smc_result = si_send_msg_to_smc(rdev, PPSMC_MSG_DisableCac);
+
+			ni_pi->cac_enabled = false;
+
+			if (ni_pi->support_cac_long_term_average)
+				smc_result = si_send_msg_to_smc(rdev, PPSMC_CACLongTermAvgDisable);
+		}
+	}
+	return ret;
+}
+
+static int si_init_smc_spll_table(struct radeon_device *rdev)
+{
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	SMC_SISLANDS_SPLL_DIV_TABLE *spll_table;
+	SISLANDS_SMC_SCLK_VALUE sclk_params;
+	u32 fb_div, p_div;
+	u32 clk_s, clk_v;
+	u32 sclk = 0;
+	int ret = 0;
+	u32 tmp;
+	int i;
+
+	if (si_pi->spll_table_start == 0)
+		return -EINVAL;
+
+	spll_table = kzalloc(sizeof(SMC_SISLANDS_SPLL_DIV_TABLE), GFP_KERNEL);
+	if (spll_table == NULL)
+		return -ENOMEM;
+
+	for (i = 0; i < 256; i++) {
+		ret = si_calculate_sclk_params(rdev, sclk, &sclk_params);
+		if (ret)
+			break;
+
+		p_div = (sclk_params.vCG_SPLL_FUNC_CNTL & SPLL_PDIV_A_MASK) >> SPLL_PDIV_A_SHIFT;
+		fb_div = (sclk_params.vCG_SPLL_FUNC_CNTL_3 & SPLL_FB_DIV_MASK) >> SPLL_FB_DIV_SHIFT;
+		clk_s = (sclk_params.vCG_SPLL_SPREAD_SPECTRUM & CLK_S_MASK) >> CLK_S_SHIFT;
+		clk_v = (sclk_params.vCG_SPLL_SPREAD_SPECTRUM_2 & CLK_V_MASK) >> CLK_V_SHIFT;
+
+		fb_div &= ~0x00001FFF;
+		fb_div >>= 1;
+		clk_v >>= 6;
+
+		if (p_div & ~(SMC_SISLANDS_SPLL_DIV_TABLE_PDIV_MASK >> SMC_SISLANDS_SPLL_DIV_TABLE_PDIV_SHIFT))
+			ret = -EINVAL;
+		if (fb_div & ~(SMC_SISLANDS_SPLL_DIV_TABLE_FBDIV_MASK >> SMC_SISLANDS_SPLL_DIV_TABLE_FBDIV_SHIFT))
+			ret = -EINVAL;
+		if (clk_s & ~(SMC_SISLANDS_SPLL_DIV_TABLE_CLKS_MASK >> SMC_SISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT))
+			ret = -EINVAL;
+		if (clk_v & ~(SMC_SISLANDS_SPLL_DIV_TABLE_CLKV_MASK >> SMC_SISLANDS_SPLL_DIV_TABLE_CLKV_SHIFT))
+			ret = -EINVAL;
+
+		if (ret)
+			break;
+
+		tmp = ((fb_div << SMC_SISLANDS_SPLL_DIV_TABLE_FBDIV_SHIFT) & SMC_SISLANDS_SPLL_DIV_TABLE_FBDIV_MASK) |
+			((p_div << SMC_SISLANDS_SPLL_DIV_TABLE_PDIV_SHIFT) & SMC_SISLANDS_SPLL_DIV_TABLE_PDIV_MASK);
+		spll_table->freq[i] = cpu_to_be32(tmp);
+
+		tmp = ((clk_v << SMC_SISLANDS_SPLL_DIV_TABLE_CLKV_SHIFT) & SMC_SISLANDS_SPLL_DIV_TABLE_CLKV_MASK) |
+			((clk_s << SMC_SISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT) & SMC_SISLANDS_SPLL_DIV_TABLE_CLKS_MASK);
+		spll_table->ss[i] = cpu_to_be32(tmp);
+
+		sclk += 512;
+	}
+
+
+	if (!ret)
+		ret = si_copy_bytes_to_smc(rdev, si_pi->spll_table_start,
+					   (u8 *)spll_table, sizeof(SMC_SISLANDS_SPLL_DIV_TABLE),
+					   si_pi->sram_end);
+
+	if (ret)
+		ni_pi->enable_power_containment = false;
+
+	kfree(spll_table);
+
+	return ret;
+}
+
+static void si_apply_state_adjust_rules(struct radeon_device *rdev,
+					struct radeon_ps *rps)
+{
+	struct ni_ps *ps = ni_get_ps(rps);
+	struct radeon_clock_and_voltage_limits *max_limits;
+	bool disable_mclk_switching;
+	u32 mclk, sclk;
+	u16 vddc, vddci;
+	int i;
+
+	if ((rdev->pm.dpm.new_active_crtc_count > 1) ||
+	    ni_dpm_vblank_too_short(rdev))
+		disable_mclk_switching = true;
+	else
+		disable_mclk_switching = false;
+
+	if (rdev->pm.dpm.ac_power)
+		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac;
+	else
+		max_limits = &rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc;
+
+	for (i = ps->performance_level_count - 2; i >= 0; i--) {
+		if (ps->performance_levels[i].vddc > ps->performance_levels[i+1].vddc)
+			ps->performance_levels[i].vddc = ps->performance_levels[i+1].vddc;
+	}
+	if (rdev->pm.dpm.ac_power == false) {
+		for (i = 0; i < ps->performance_level_count; i++) {
+			if (ps->performance_levels[i].mclk > max_limits->mclk)
+				ps->performance_levels[i].mclk = max_limits->mclk;
+			if (ps->performance_levels[i].sclk > max_limits->sclk)
+				ps->performance_levels[i].sclk = max_limits->sclk;
+			if (ps->performance_levels[i].vddc > max_limits->vddc)
+				ps->performance_levels[i].vddc = max_limits->vddc;
+			if (ps->performance_levels[i].vddci > max_limits->vddci)
+				ps->performance_levels[i].vddci = max_limits->vddci;
+		}
+	}
+
+	/* XXX validate the min clocks required for display */
+
+	if (disable_mclk_switching) {
+		mclk  = ps->performance_levels[ps->performance_level_count - 1].mclk;
+		sclk = ps->performance_levels[0].sclk;
+		vddc = ps->performance_levels[0].vddc;
+		vddci = ps->performance_levels[ps->performance_level_count - 1].vddci;
+	} else {
+		sclk = ps->performance_levels[0].sclk;
+		mclk = ps->performance_levels[0].mclk;
+		vddc = ps->performance_levels[0].vddc;
+		vddci = ps->performance_levels[0].vddci;
+	}
+
+	/* adjusted low state */
+	ps->performance_levels[0].sclk = sclk;
+	ps->performance_levels[0].mclk = mclk;
+	ps->performance_levels[0].vddc = vddc;
+	ps->performance_levels[0].vddci = vddci;
+
+	for (i = 1; i < ps->performance_level_count; i++) {
+		if (ps->performance_levels[i].sclk < ps->performance_levels[i - 1].sclk)
+			ps->performance_levels[i].sclk = ps->performance_levels[i - 1].sclk;
+		if (ps->performance_levels[i].vddc < ps->performance_levels[i - 1].vddc)
+			ps->performance_levels[i].vddc = ps->performance_levels[i - 1].vddc;
+	}
+
+	if (disable_mclk_switching) {
+		mclk = ps->performance_levels[0].mclk;
+		for (i = 1; i < ps->performance_level_count; i++) {
+			if (mclk < ps->performance_levels[i].mclk)
+				mclk = ps->performance_levels[i].mclk;
+		}
+		for (i = 0; i < ps->performance_level_count; i++) {
+			ps->performance_levels[i].mclk = mclk;
+			ps->performance_levels[i].vddci = vddci;
+		}
+	} else {
+		for (i = 1; i < ps->performance_level_count; i++) {
+			if (ps->performance_levels[i].mclk < ps->performance_levels[i - 1].mclk)
+				ps->performance_levels[i].mclk = ps->performance_levels[i - 1].mclk;
+			if (ps->performance_levels[i].vddci < ps->performance_levels[i - 1].vddci)
+				ps->performance_levels[i].vddci = ps->performance_levels[i - 1].vddci;
+		}
+	}
+
+        for (i = 0; i < ps->performance_level_count; i++)
+                btc_adjust_clock_combinations(rdev, max_limits,
+                                              &ps->performance_levels[i]);
+
+	for (i = 0; i < ps->performance_level_count; i++) {
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk,
+						   ps->performance_levels[i].sclk,
+						   max_limits->vddc,  &ps->performance_levels[i].vddc);
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk,
+						   ps->performance_levels[i].mclk,
+						   max_limits->vddci, &ps->performance_levels[i].vddci);
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk,
+						   ps->performance_levels[i].mclk,
+						   max_limits->vddc,  &ps->performance_levels[i].vddc);
+		btc_apply_voltage_dependency_rules(&rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk,
+						   rdev->clock.current_dispclk,
+						   max_limits->vddc,  &ps->performance_levels[i].vddc);
+	}
+
+	for (i = 0; i < ps->performance_level_count; i++) {
+		btc_apply_voltage_delta_rules(rdev,
+					      max_limits->vddc, max_limits->vddci,
+					      &ps->performance_levels[i].vddc,
+					      &ps->performance_levels[i].vddci);
+	}
+
+	ps->dc_compatible = true;
+	for (i = 0; i < ps->performance_level_count; i++) {
+		if (ps->performance_levels[i].vddc > rdev->pm.dpm.dyn_state.max_clock_voltage_on_dc.vddc)
+			ps->dc_compatible = false;
+	}
+
+}
+
+#if 0
+static int si_read_smc_soft_register(struct radeon_device *rdev,
+				     u16 reg_offset, u32 *value)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	return si_read_smc_sram_dword(rdev,
+				      si_pi->soft_regs_start + reg_offset, value,
+				      si_pi->sram_end);
+}
+#endif
+
+static int si_write_smc_soft_register(struct radeon_device *rdev,
+				      u16 reg_offset, u32 value)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	return si_write_smc_sram_dword(rdev,
+				       si_pi->soft_regs_start + reg_offset,
+				       value, si_pi->sram_end);
+}
+
+static bool si_is_special_1gb_platform(struct radeon_device *rdev)
+{
+	bool ret = false;
+	u32 tmp, width, row, column, bank, density;
+	bool is_memory_gddr5, is_special;
+
+	tmp = RREG32(MC_SEQ_MISC0);
+	is_memory_gddr5 = (MC_SEQ_MISC0_GDDR5_VALUE == ((tmp & MC_SEQ_MISC0_GDDR5_MASK) >> MC_SEQ_MISC0_GDDR5_SHIFT));
+	is_special = (MC_SEQ_MISC0_REV_ID_VALUE == ((tmp & MC_SEQ_MISC0_REV_ID_MASK) >> MC_SEQ_MISC0_REV_ID_SHIFT))
+		& (MC_SEQ_MISC0_VEN_ID_VALUE == ((tmp & MC_SEQ_MISC0_VEN_ID_MASK) >> MC_SEQ_MISC0_VEN_ID_SHIFT));
+
+	WREG32(MC_SEQ_IO_DEBUG_INDEX, 0xb);
+	width = ((RREG32(MC_SEQ_IO_DEBUG_DATA) >> 1) & 1) ? 16 : 32;
+
+	tmp = RREG32(MC_ARB_RAMCFG);
+	row = ((tmp & NOOFROWS_MASK) >> NOOFROWS_SHIFT) + 10;
+	column = ((tmp & NOOFCOLS_MASK) >> NOOFCOLS_SHIFT) + 8;
+	bank = ((tmp & NOOFBANK_MASK) >> NOOFBANK_SHIFT) + 2;
+
+	density = (1 << (row + column - 20 + bank)) * width;
+
+	if ((rdev->pdev->device == 0x6819) &&
+	    is_memory_gddr5 && is_special && (density == 0x400))
+		ret = true;
+
+	return ret;
+}
+
+static void si_get_leakage_vddc(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u16 vddc, count = 0;
+	int i, ret;
+
+	for (i = 0; i < SISLANDS_MAX_LEAKAGE_COUNT; i++) {
+		ret = radeon_atom_get_leakage_vddc_based_on_leakage_idx(rdev, &vddc, SISLANDS_LEAKAGE_INDEX0 + i);
+
+		if (!ret && (vddc > 0) && (vddc != (SISLANDS_LEAKAGE_INDEX0 + i))) {
+			si_pi->leakage_voltage.entries[count].voltage = vddc;
+			si_pi->leakage_voltage.entries[count].leakage_index =
+				SISLANDS_LEAKAGE_INDEX0 + i;
+			count++;
+		}
+	}
+	si_pi->leakage_voltage.count = count;
+}
+
+static int si_get_leakage_voltage_from_leakage_index(struct radeon_device *rdev,
+						     u32 index, u16 *leakage_voltage)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	int i;
+
+	if (leakage_voltage == NULL)
+		return -EINVAL;
+
+	if ((index & 0xff00) != 0xff00)
+		return -EINVAL;
+
+	if ((index & 0xff) > SISLANDS_MAX_LEAKAGE_COUNT + 1)
+		return -EINVAL;
+
+	if (index < SISLANDS_LEAKAGE_INDEX0)
+		return -EINVAL;
+
+	for (i = 0; i < si_pi->leakage_voltage.count; i++) {
+		if (si_pi->leakage_voltage.entries[i].leakage_index == index) {
+			*leakage_voltage = si_pi->leakage_voltage.entries[i].voltage;
+			return 0;
+		}
+	}
+	return -EAGAIN;
+}
+
+static void si_set_dpm_event_sources(struct radeon_device *rdev, u32 sources)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	bool want_thermal_protection;
+	enum radeon_dpm_event_src dpm_event_src;
+
+	switch (sources) {
+	case 0:
+	default:
+		want_thermal_protection = false;
+                break;
+	case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_DIGITAL;
+		break;
+	case (1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_EXTERNAL;
+		break;
+	case ((1 << RADEON_DPM_AUTO_THROTTLE_SRC_EXTERNAL) |
+	      (1 << RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL)):
+		want_thermal_protection = true;
+		dpm_event_src = RADEON_DPM_EVENT_SRC_DIGIAL_OR_EXTERNAL;
+		break;
+	}
+
+	if (want_thermal_protection) {
+		WREG32_P(CG_THERMAL_CTRL, DPM_EVENT_SRC(dpm_event_src), ~DPM_EVENT_SRC_MASK);
+		if (pi->thermal_protection)
+			WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS);
+	} else {
+		WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS);
+	}
+}
+
+static void si_enable_auto_throttle_source(struct radeon_device *rdev,
+					   enum radeon_dpm_auto_throttle_src source,
+					   bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (enable) {
+		if (!(pi->active_auto_throttle_sources & (1 << source))) {
+			pi->active_auto_throttle_sources |= 1 << source;
+			si_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources);
+		}
+	} else {
+		if (pi->active_auto_throttle_sources & (1 << source)) {
+			pi->active_auto_throttle_sources &= ~(1 << source);
+			si_set_dpm_event_sources(rdev, pi->active_auto_throttle_sources);
+		}
+	}
+}
+
+static void si_start_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN);
+}
+
+static void si_stop_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, 0, ~GLOBAL_PWRMGT_EN);
+}
+
+static void si_enable_sclk_control(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~SCLK_PWRMGT_OFF);
+	else
+		WREG32_P(SCLK_PWRMGT_CNTL, SCLK_PWRMGT_OFF, ~SCLK_PWRMGT_OFF);
+
+}
+
+#if 0
+static int si_notify_hardware_of_thermal_state(struct radeon_device *rdev,
+					       u32 thermal_level)
+{
+	PPSMC_Result ret;
+
+	if (thermal_level == 0) {
+		ret = si_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+		if (ret == PPSMC_Result_OK)
+			return 0;
+		else
+			return -EINVAL;
+	}
+	return 0;
+}
+
+static void si_notify_hardware_vpu_recovery_event(struct radeon_device *rdev)
+{
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_tdr_is_about_to_happen, true);
+}
+#endif
+
+#if 0
+static int si_notify_hw_of_powersource(struct radeon_device *rdev, bool ac_power)
+{
+	if (ac_power)
+		return (si_send_msg_to_smc(rdev, PPSMC_MSG_RunningOnAC) == PPSMC_Result_OK) ?
+			0 : -EINVAL;
+
+	return 0;
+}
+#endif
+
+static PPSMC_Result si_send_msg_to_smc_with_parameter(struct radeon_device *rdev,
+						      PPSMC_Msg msg, u32 parameter)
+{
+	WREG32(SMC_SCRATCH0, parameter);
+	return si_send_msg_to_smc(rdev, msg);
+}
+
+static int si_restrict_performance_levels_before_switch(struct radeon_device *rdev)
+{
+	if (si_send_msg_to_smc(rdev, PPSMC_MSG_NoForcedLevel) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 1) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+int si_dpm_force_performance_level(struct radeon_device *rdev,
+				   enum radeon_dpm_forced_level level)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct ni_ps *ps = ni_get_ps(rps);
+	u32 levels;
+
+	if (level == RADEON_DPM_FORCED_LEVEL_HIGH) {
+		if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+
+		if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 1) != PPSMC_Result_OK)
+			return -EINVAL;
+	} else if (level == RADEON_DPM_FORCED_LEVEL_LOW) {
+		if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+
+		levels = ps->performance_level_count - 1;
+		if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, levels) != PPSMC_Result_OK)
+			return -EINVAL;
+	} else if (level == RADEON_DPM_FORCED_LEVEL_AUTO) {
+		if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetForcedLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+
+		if (si_send_msg_to_smc_with_parameter(rdev, PPSMC_MSG_SetEnabledLevels, 0) != PPSMC_Result_OK)
+			return -EINVAL;
+	}
+
+	rdev->pm.dpm.forced_level = level;
+
+	return 0;
+}
+
+static int si_set_boot_state(struct radeon_device *rdev)
+{
+	return (si_send_msg_to_smc(rdev, PPSMC_MSG_SwitchToInitialState) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+static int si_set_sw_state(struct radeon_device *rdev)
+{
+	return (si_send_msg_to_smc(rdev, PPSMC_MSG_SwitchToSwState) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+static int si_halt_smc(struct radeon_device *rdev)
+{
+	if (si_send_msg_to_smc(rdev, PPSMC_MSG_Halt) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return (si_wait_for_smc_inactive(rdev) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+static int si_resume_smc(struct radeon_device *rdev)
+{
+	if (si_send_msg_to_smc(rdev, PPSMC_FlushDataCache) != PPSMC_Result_OK)
+		return -EINVAL;
+
+	return (si_send_msg_to_smc(rdev, PPSMC_MSG_Resume) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+static void si_dpm_start_smc(struct radeon_device *rdev)
+{
+	si_program_jump_on_start(rdev);
+	si_start_smc(rdev);
+	si_start_smc_clock(rdev);
+}
+
+static void si_dpm_stop_smc(struct radeon_device *rdev)
+{
+	si_reset_smc(rdev);
+	si_stop_smc_clock(rdev);
+}
+
+static int si_process_firmware_header(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_stateTable,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+        si_pi->state_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_softRegisters,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->soft_regs_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_mcRegisterTable,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->mc_reg_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_mcArbDramAutoRefreshTable,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->arb_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_CacConfigTable,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->cac_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_DteConfiguration,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->dte_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_spllTable,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->spll_table_start = tmp;
+
+	ret = si_read_smc_sram_dword(rdev,
+				     SISLANDS_SMC_FIRMWARE_HEADER_LOCATION +
+				     SISLANDS_SMC_FIRMWARE_HEADER_PAPMParameters,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	si_pi->papm_cfg_table_start = tmp;
+
+	return ret;
+}
+
+static void si_read_clock_registers(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	si_pi->clock_registers.cg_spll_func_cntl = RREG32(CG_SPLL_FUNC_CNTL);
+	si_pi->clock_registers.cg_spll_func_cntl_2 = RREG32(CG_SPLL_FUNC_CNTL_2);
+	si_pi->clock_registers.cg_spll_func_cntl_3 = RREG32(CG_SPLL_FUNC_CNTL_3);
+	si_pi->clock_registers.cg_spll_func_cntl_4 = RREG32(CG_SPLL_FUNC_CNTL_4);
+	si_pi->clock_registers.cg_spll_spread_spectrum = RREG32(CG_SPLL_SPREAD_SPECTRUM);
+	si_pi->clock_registers.cg_spll_spread_spectrum_2 = RREG32(CG_SPLL_SPREAD_SPECTRUM_2);
+	si_pi->clock_registers.dll_cntl = RREG32(DLL_CNTL);
+	si_pi->clock_registers.mclk_pwrmgt_cntl = RREG32(MCLK_PWRMGT_CNTL);
+	si_pi->clock_registers.mpll_ad_func_cntl = RREG32(MPLL_AD_FUNC_CNTL);
+	si_pi->clock_registers.mpll_dq_func_cntl = RREG32(MPLL_DQ_FUNC_CNTL);
+	si_pi->clock_registers.mpll_func_cntl = RREG32(MPLL_FUNC_CNTL);
+	si_pi->clock_registers.mpll_func_cntl_1 = RREG32(MPLL_FUNC_CNTL_1);
+	si_pi->clock_registers.mpll_func_cntl_2 = RREG32(MPLL_FUNC_CNTL_2);
+	si_pi->clock_registers.mpll_ss1 = RREG32(MPLL_SS1);
+	si_pi->clock_registers.mpll_ss2 = RREG32(MPLL_SS2);
+}
+
+static void si_enable_thermal_protection(struct radeon_device *rdev,
+					  bool enable)
+{
+	if (enable)
+		WREG32_P(GENERAL_PWRMGT, 0, ~THERMAL_PROTECTION_DIS);
+	else
+		WREG32_P(GENERAL_PWRMGT, THERMAL_PROTECTION_DIS, ~THERMAL_PROTECTION_DIS);
+}
+
+static void si_enable_acpi_power_management(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, STATIC_PM_EN, ~STATIC_PM_EN);
+}
+
+#if 0
+static int si_enter_ulp_state(struct radeon_device *rdev)
+{
+	WREG32(SMC_MESSAGE_0, PPSMC_MSG_SwitchToMinimumPower);
+
+	udelay(25000);
+
+	return 0;
+}
+
+static int si_exit_ulp_state(struct radeon_device *rdev)
+{
+	int i;
+
+	WREG32(SMC_MESSAGE_0, PPSMC_MSG_ResumeFromMinimumPower);
+
+	udelay(7000);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(SMC_RESP_0) == 1)
+			break;
+		udelay(1000);
+	}
+
+	return 0;
+}
+#endif
+
+static int si_notify_smc_display_change(struct radeon_device *rdev,
+				     bool has_display)
+{
+	PPSMC_Msg msg = has_display ?
+		PPSMC_MSG_HasDisplay : PPSMC_MSG_NoDisplay;
+
+	return (si_send_msg_to_smc(rdev, msg) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+static void si_program_response_times(struct radeon_device *rdev)
+{
+	u32 voltage_response_time, backbias_response_time, acpi_delay_time, vbi_time_out;
+	u32 vddc_dly, acpi_dly, vbi_dly;
+	u32 reference_clock;
+
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_mvdd_chg_time, 1);
+
+	voltage_response_time = (u32)rdev->pm.dpm.voltage_response_time;
+        backbias_response_time = (u32)rdev->pm.dpm.backbias_response_time;
+
+	if (voltage_response_time == 0)
+		voltage_response_time = 1000;
+
+	acpi_delay_time = 15000;
+	vbi_time_out = 100000;
+
+	reference_clock = radeon_get_xclk(rdev);
+
+	vddc_dly = (voltage_response_time  * reference_clock) / 100;
+	acpi_dly = (acpi_delay_time * reference_clock) / 100;
+	vbi_dly  = (vbi_time_out * reference_clock) / 100;
+
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_delay_vreg,  vddc_dly);
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_delay_acpi,  acpi_dly);
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_mclk_chg_timeout, vbi_dly);
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_mc_block_delay, 0xAA);
+}
+
+static void si_program_ds_registers(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	u32 tmp = 1; /* XXX: 0x10 on tahiti A0 */
+
+	if (eg_pi->sclk_deep_sleep) {
+		WREG32_P(MISC_CLK_CNTL, DEEP_SLEEP_CLK_SEL(tmp), ~DEEP_SLEEP_CLK_SEL_MASK);
+		WREG32_P(CG_SPLL_AUTOSCALE_CNTL, AUTOSCALE_ON_SS_CLEAR,
+			 ~AUTOSCALE_ON_SS_CLEAR);
+	}
+}
+
+static void si_program_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp, pipe;
+	int i;
+
+	tmp = RREG32(CG_DISPLAY_GAP_CNTL) & ~(DISP1_GAP_MASK | DISP2_GAP_MASK);
+	if (rdev->pm.dpm.new_active_crtc_count > 0)
+		tmp |= DISP1_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM);
+	else
+		tmp |= DISP1_GAP(R600_PM_DISPLAY_GAP_IGNORE);
+
+	if (rdev->pm.dpm.new_active_crtc_count > 1)
+		tmp |= DISP2_GAP(R600_PM_DISPLAY_GAP_VBLANK_OR_WM);
+	else
+		tmp |= DISP2_GAP(R600_PM_DISPLAY_GAP_IGNORE);
+
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+
+	tmp = RREG32(DCCG_DISP_SLOW_SELECT_REG);
+	pipe = (tmp & DCCG_DISP1_SLOW_SELECT_MASK) >> DCCG_DISP1_SLOW_SELECT_SHIFT;
+
+	if ((rdev->pm.dpm.new_active_crtc_count > 0) &&
+	    (!(rdev->pm.dpm.new_active_crtcs & (1 << pipe)))) {
+		/* find the first active crtc */
+		for (i = 0; i < rdev->num_crtc; i++) {
+			if (rdev->pm.dpm.new_active_crtcs & (1 << i))
+				break;
+		}
+		if (i == rdev->num_crtc)
+			pipe = 0;
+		else
+			pipe = i;
+
+		tmp &= ~DCCG_DISP1_SLOW_SELECT_MASK;
+		tmp |= DCCG_DISP1_SLOW_SELECT(pipe);
+		WREG32(DCCG_DISP_SLOW_SELECT_REG, tmp);
+	}
+
+	si_notify_smc_display_change(rdev, rdev->pm.dpm.new_active_crtc_count > 0);
+}
+
+static void si_enable_spread_spectrum(struct radeon_device *rdev, bool enable)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	if (enable) {
+		if (pi->sclk_ss)
+			WREG32_P(GENERAL_PWRMGT, DYN_SPREAD_SPECTRUM_EN, ~DYN_SPREAD_SPECTRUM_EN);
+	} else {
+		WREG32_P(CG_SPLL_SPREAD_SPECTRUM, 0, ~SSEN);
+		WREG32_P(GENERAL_PWRMGT, 0, ~DYN_SPREAD_SPECTRUM_EN);
+	}
+}
+
+static void si_setup_bsp(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 xclk = radeon_get_xclk(rdev);
+
+	r600_calculate_u_and_p(pi->asi,
+			       xclk,
+			       16,
+			       &pi->bsp,
+			       &pi->bsu);
+
+	r600_calculate_u_and_p(pi->pasi,
+			       xclk,
+			       16,
+			       &pi->pbsp,
+			       &pi->pbsu);
+
+
+        pi->dsp = BSP(pi->bsp) | BSU(pi->bsu);
+	pi->psp = BSP(pi->pbsp) | BSU(pi->pbsu);
+
+	WREG32(CG_BSP, pi->dsp);
+}
+
+static void si_program_git(struct radeon_device *rdev)
+{
+	WREG32_P(CG_GIT, CG_GICST(R600_GICST_DFLT), ~CG_GICST_MASK);
+}
+
+static void si_program_tp(struct radeon_device *rdev)
+{
+	int i;
+	enum r600_td td = R600_TD_DFLT;
+
+	for (i = 0; i < R600_PM_NUMBER_OF_TC; i++)
+		WREG32(CG_FFCT_0 + (i * 4), (UTC_0(r600_utc[i]) | DTC_0(r600_dtc[i])));
+
+	if (td == R600_TD_AUTO)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_FORCE_TREND_SEL);
+	else
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_FORCE_TREND_SEL, ~FIR_FORCE_TREND_SEL);
+
+	if (td == R600_TD_UP)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_TREND_MODE);
+
+	if (td == R600_TD_DOWN)
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_TREND_MODE, ~FIR_TREND_MODE);
+}
+
+static void si_program_tpp(struct radeon_device *rdev)
+{
+	WREG32(CG_TPC, R600_TPC_DFLT);
+}
+
+static void si_program_sstp(struct radeon_device *rdev)
+{
+	WREG32(CG_SSP, (SSTU(R600_SSTU_DFLT) | SST(R600_SST_DFLT)));
+}
+
+static void si_enable_display_gap(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32(CG_DISPLAY_GAP_CNTL);
+
+	tmp &= ~(DISP1_GAP_MCHG_MASK | DISP2_GAP_MCHG_MASK);
+	tmp |= (DISP1_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE) |
+		DISP2_GAP_MCHG(R600_PM_DISPLAY_GAP_IGNORE));
+	WREG32(CG_DISPLAY_GAP_CNTL, tmp);
+}
+
+static void si_program_vc(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	WREG32(CG_FTV, pi->vrc);
+}
+
+static void si_clear_vc(struct radeon_device *rdev)
+{
+	WREG32(CG_FTV, 0);
+}
+
+static u8 si_get_ddr3_mclk_frequency_ratio(u32 memory_clock)
+{
+	u8 mc_para_index;
+
+	if (memory_clock < 10000)
+		mc_para_index = 0;
+	else if (memory_clock >= 80000)
+		mc_para_index = 0x0f;
+	else
+		mc_para_index = (u8)((memory_clock - 10000) / 5000 + 1);
+	return mc_para_index;
+}
+
+static u8 si_get_mclk_frequency_ratio(u32 memory_clock, bool strobe_mode)
+{
+	u8 mc_para_index;
+
+	if (strobe_mode) {
+		if (memory_clock < 12500)
+			mc_para_index = 0x00;
+		else if (memory_clock > 47500)
+			mc_para_index = 0x0f;
+		else
+			mc_para_index = (u8)((memory_clock - 10000) / 2500);
+	} else {
+		if (memory_clock < 65000)
+			mc_para_index = 0x00;
+		else if (memory_clock > 135000)
+			mc_para_index = 0x0f;
+		else
+			mc_para_index = (u8)((memory_clock - 60000) / 5000);
+	}
+	return mc_para_index;
+}
+
+static u8 si_get_strobe_mode_settings(struct radeon_device *rdev, u32 mclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	bool strobe_mode = false;
+	u8 result = 0;
+
+	if (mclk <= pi->mclk_strobe_mode_threshold)
+		strobe_mode = true;
+
+	if (pi->mem_gddr5)
+		result = si_get_mclk_frequency_ratio(mclk, strobe_mode);
+	else
+		result = si_get_ddr3_mclk_frequency_ratio(mclk);
+
+	if (strobe_mode)
+		result |= SISLANDS_SMC_STROBE_ENABLE;
+
+	return result;
+}
+
+static int si_upload_firmware(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	int ret;
+
+	si_reset_smc(rdev);
+	si_stop_smc_clock(rdev);
+
+	ret = si_load_smc_ucode(rdev, si_pi->sram_end);
+
+	return ret;
+}
+
+static bool si_validate_phase_shedding_tables(struct radeon_device *rdev,
+					      const struct atom_voltage_table *table,
+					      const struct radeon_phase_shedding_limits_table *limits)
+{
+	u32 data, num_bits, num_levels;
+
+	if ((table == NULL) || (limits == NULL))
+		return false;
+
+	data = table->mask_low;
+
+	num_bits = hweight32(data);
+
+	if (num_bits == 0)
+		return false;
+
+	num_levels = (1 << num_bits);
+
+	if (table->count != num_levels)
+		return false;
+
+	if (limits->count != (num_levels - 1))
+		return false;
+
+	return true;
+}
+
+static void si_trim_voltage_table_to_fit_state_table(struct radeon_device *rdev,
+						     struct atom_voltage_table *voltage_table)
+{
+	unsigned int i, diff;
+
+	if (voltage_table->count <= SISLANDS_MAX_NO_VREG_STEPS)
+		return;
+
+	diff = voltage_table->count - SISLANDS_MAX_NO_VREG_STEPS;
+
+	for (i= 0; i < SISLANDS_MAX_NO_VREG_STEPS; i++)
+		voltage_table->entries[i] = voltage_table->entries[i + diff];
+
+	voltage_table->count = SISLANDS_MAX_NO_VREG_STEPS;
+}
+
+static int si_construct_voltage_tables(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	int ret;
+
+	ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDC,
+					    VOLTAGE_OBJ_GPIO_LUT, &eg_pi->vddc_voltage_table);
+	if (ret)
+		return ret;
+
+	if (eg_pi->vddc_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS)
+		si_trim_voltage_table_to_fit_state_table(rdev, &eg_pi->vddc_voltage_table);
+
+	if (eg_pi->vddci_control) {
+		ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDCI,
+						    VOLTAGE_OBJ_GPIO_LUT, &eg_pi->vddci_voltage_table);
+		if (ret)
+			return ret;
+
+		if (eg_pi->vddci_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS)
+			si_trim_voltage_table_to_fit_state_table(rdev, &eg_pi->vddci_voltage_table);
+	}
+
+	if (pi->mvdd_control) {
+		ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_MVDDC,
+						    VOLTAGE_OBJ_GPIO_LUT, &si_pi->mvdd_voltage_table);
+
+		if (ret) {
+			pi->mvdd_control = false;
+			return ret;
+		}
+
+		if (si_pi->mvdd_voltage_table.count == 0) {
+			pi->mvdd_control = false;
+			return -EINVAL;
+		}
+
+		if (si_pi->mvdd_voltage_table.count > SISLANDS_MAX_NO_VREG_STEPS)
+			si_trim_voltage_table_to_fit_state_table(rdev, &si_pi->mvdd_voltage_table);
+	}
+
+	if (si_pi->vddc_phase_shed_control) {
+		ret = radeon_atom_get_voltage_table(rdev, VOLTAGE_TYPE_VDDC,
+						    VOLTAGE_OBJ_PHASE_LUT, &si_pi->vddc_phase_shed_table);
+		if (ret)
+			si_pi->vddc_phase_shed_control = false;
+
+		if ((si_pi->vddc_phase_shed_table.count == 0) ||
+		    (si_pi->vddc_phase_shed_table.count > SISLANDS_MAX_NO_VREG_STEPS))
+			si_pi->vddc_phase_shed_control = false;
+	}
+
+	return 0;
+}
+
+static void si_populate_smc_voltage_table(struct radeon_device *rdev,
+					  const struct atom_voltage_table *voltage_table,
+					  SISLANDS_SMC_STATETABLE *table)
+{
+	unsigned int i;
+
+	for (i = 0; i < voltage_table->count; i++)
+		table->lowSMIO[i] |= cpu_to_be32(voltage_table->entries[i].smio_low);
+}
+
+static int si_populate_smc_voltage_tables(struct radeon_device *rdev,
+					  SISLANDS_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u8 i;
+
+	if (eg_pi->vddc_voltage_table.count) {
+		si_populate_smc_voltage_table(rdev, &eg_pi->vddc_voltage_table, table);
+		table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDC] =
+			cpu_to_be32(eg_pi->vddc_voltage_table.mask_low);
+
+		for (i = 0; i < eg_pi->vddc_voltage_table.count; i++) {
+			if (pi->max_vddc_in_table <= eg_pi->vddc_voltage_table.entries[i].value) {
+				table->maxVDDCIndexInPPTable = i;
+				break;
+			}
+		}
+	}
+
+	if (eg_pi->vddci_voltage_table.count) {
+		si_populate_smc_voltage_table(rdev, &eg_pi->vddci_voltage_table, table);
+
+		table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDCI] =
+			cpu_to_be32(eg_pi->vddci_voltage_table.mask_low);
+	}
+
+
+	if (si_pi->mvdd_voltage_table.count) {
+		si_populate_smc_voltage_table(rdev, &si_pi->mvdd_voltage_table, table);
+
+		table->voltageMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_MVDD] =
+			cpu_to_be32(si_pi->mvdd_voltage_table.mask_low);
+	}
+
+	if (si_pi->vddc_phase_shed_control) {
+		if (si_validate_phase_shedding_tables(rdev, &si_pi->vddc_phase_shed_table,
+						      &rdev->pm.dpm.dyn_state.phase_shedding_limits_table)) {
+			si_populate_smc_voltage_table(rdev, &si_pi->vddc_phase_shed_table, table);
+
+			table->phaseMaskTable.lowMask[SISLANDS_SMC_VOLTAGEMASK_VDDC] =
+				cpu_to_be32(si_pi->vddc_phase_shed_table.mask_low);
+
+			si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_phase_shedding_delay,
+						   (u32)si_pi->vddc_phase_shed_table.phase_delay);
+		} else {
+			si_pi->vddc_phase_shed_control = false;
+		}
+	}
+
+	return 0;
+}
+
+static int si_populate_voltage_value(struct radeon_device *rdev,
+				     const struct atom_voltage_table *table,
+				     u16 value, SISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	unsigned int i;
+
+	for (i = 0; i < table->count; i++) {
+		if (value <= table->entries[i].value) {
+			voltage->index = (u8)i;
+			voltage->value = cpu_to_be16(table->entries[i].value);
+			break;
+		}
+	}
+
+	if (i >= table->count)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int si_populate_mvdd_value(struct radeon_device *rdev, u32 mclk,
+				  SISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	if (pi->mvdd_control) {
+		if (mclk <= pi->mvdd_split_frequency)
+			voltage->index = 0;
+		else
+			voltage->index = (u8)(si_pi->mvdd_voltage_table.count) - 1;
+
+		voltage->value = cpu_to_be16(si_pi->mvdd_voltage_table.entries[voltage->index].value);
+	}
+	return 0;
+}
+
+static int si_get_std_voltage_value(struct radeon_device *rdev,
+				    SISLANDS_SMC_VOLTAGE_VALUE *voltage,
+				    u16 *std_voltage)
+{
+	u16 v_index;
+	bool voltage_found = false;
+	*std_voltage = be16_to_cpu(voltage->value);
+
+	if (rdev->pm.dpm.dyn_state.cac_leakage_table.entries) {
+		if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_NEW_CAC_VOLTAGE) {
+			if (rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries == NULL)
+				return -EINVAL;
+
+			for (v_index = 0; (u32)v_index < rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count; v_index++) {
+				if (be16_to_cpu(voltage->value) ==
+				    (u16)rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[v_index].v) {
+					voltage_found = true;
+					if ((u32)v_index < rdev->pm.dpm.dyn_state.cac_leakage_table.count)
+						*std_voltage =
+							rdev->pm.dpm.dyn_state.cac_leakage_table.entries[v_index].vddc;
+					else
+						*std_voltage =
+							rdev->pm.dpm.dyn_state.cac_leakage_table.entries[rdev->pm.dpm.dyn_state.cac_leakage_table.count-1].vddc;
+					break;
+				}
+			}
+
+			if (!voltage_found) {
+				for (v_index = 0; (u32)v_index < rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count; v_index++) {
+					if (be16_to_cpu(voltage->value) <=
+					    (u16)rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[v_index].v) {
+						voltage_found = true;
+						if ((u32)v_index < rdev->pm.dpm.dyn_state.cac_leakage_table.count)
+							*std_voltage =
+								rdev->pm.dpm.dyn_state.cac_leakage_table.entries[v_index].vddc;
+						else
+							*std_voltage =
+								rdev->pm.dpm.dyn_state.cac_leakage_table.entries[rdev->pm.dpm.dyn_state.cac_leakage_table.count-1].vddc;
+						break;
+					}
+				}
+			}
+		} else {
+			if ((u32)voltage->index < rdev->pm.dpm.dyn_state.cac_leakage_table.count)
+				*std_voltage = rdev->pm.dpm.dyn_state.cac_leakage_table.entries[voltage->index].vddc;
+		}
+	}
+
+	return 0;
+}
+
+static int si_populate_std_voltage_value(struct radeon_device *rdev,
+					 u16 value, u8 index,
+					 SISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	voltage->index = index;
+	voltage->value = cpu_to_be16(value);
+
+	return 0;
+}
+
+static int si_populate_phase_shedding_value(struct radeon_device *rdev,
+					    const struct radeon_phase_shedding_limits_table *limits,
+					    u16 voltage, u32 sclk, u32 mclk,
+					    SISLANDS_SMC_VOLTAGE_VALUE *smc_voltage)
+{
+	unsigned int i;
+
+	for (i = 0; i < limits->count; i++) {
+		if ((voltage <= limits->entries[i].voltage) &&
+		    (sclk <= limits->entries[i].sclk) &&
+		    (mclk <= limits->entries[i].mclk))
+			break;
+	}
+
+	smc_voltage->phase_settings = (u8)i;
+
+	return 0;
+}
+
+static int si_init_arb_table_index(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = si_read_smc_sram_dword(rdev, si_pi->arb_table_start, &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	tmp &= 0x00FFFFFF;
+	tmp |= MC_CG_ARB_FREQ_F1 << 24;
+
+	return si_write_smc_sram_dword(rdev, si_pi->arb_table_start,  tmp, si_pi->sram_end);
+}
+
+static int si_initial_switch_from_arb_f0_to_f1(struct radeon_device *rdev)
+{
+	return ni_copy_and_switch_arb_sets(rdev, MC_CG_ARB_FREQ_F0, MC_CG_ARB_FREQ_F1);
+}
+
+static int si_reset_to_default(struct radeon_device *rdev)
+{
+	return (si_send_msg_to_smc(rdev, PPSMC_MSG_ResetToDefaults) == PPSMC_Result_OK) ?
+		0 : -EINVAL;
+}
+
+static int si_force_switch_to_arb_f0(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 tmp;
+	int ret;
+
+	ret = si_read_smc_sram_dword(rdev, si_pi->arb_table_start,
+				     &tmp, si_pi->sram_end);
+	if (ret)
+		return ret;
+
+	tmp = (tmp >> 24) & 0xff;
+
+	if (tmp == MC_CG_ARB_FREQ_F0)
+		return 0;
+
+	return ni_copy_and_switch_arb_sets(rdev, tmp, MC_CG_ARB_FREQ_F0);
+}
+
+static u32 si_calculate_memory_refresh_rate(struct radeon_device *rdev,
+					    u32 engine_clock)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u32 dram_rows;
+	u32 dram_refresh_rate;
+	u32 mc_arb_rfsh_rate;
+	u32 tmp = (RREG32(MC_ARB_RAMCFG) & NOOFROWS_MASK) >> NOOFROWS_SHIFT;
+
+	if (pi->mem_gddr5)
+		dram_rows = 1 << (tmp + 10);
+	else
+		dram_rows = DDR3_DRAM_ROWS;
+
+	dram_refresh_rate = 1 << ((RREG32(MC_SEQ_MISC0) & 0x3) + 3);
+	mc_arb_rfsh_rate = ((engine_clock * 10) * dram_refresh_rate / dram_rows - 32) / 64;
+
+	return mc_arb_rfsh_rate;
+}
+
+static int si_populate_memory_timing_parameters(struct radeon_device *rdev,
+						struct rv7xx_pl *pl,
+						SMC_SIslands_MCArbDramTimingRegisterSet *arb_regs)
+{
+	u32 dram_timing;
+	u32 dram_timing2;
+	u32 burst_time;
+
+	arb_regs->mc_arb_rfsh_rate =
+		(u8)si_calculate_memory_refresh_rate(rdev, pl->sclk);
+
+	radeon_atom_set_engine_dram_timings(rdev,
+					    pl->sclk,
+                                            pl->mclk);
+
+	dram_timing  = RREG32(MC_ARB_DRAM_TIMING);
+	dram_timing2 = RREG32(MC_ARB_DRAM_TIMING2);
+	burst_time = RREG32(MC_ARB_BURST_TIME) & STATE0_MASK;
+
+	arb_regs->mc_arb_dram_timing  = cpu_to_be32(dram_timing);
+	arb_regs->mc_arb_dram_timing2 = cpu_to_be32(dram_timing2);
+	arb_regs->mc_arb_burst_time = (u8)burst_time;
+
+	return 0;
+}
+
+static int si_do_program_memory_timing_parameters(struct radeon_device *rdev,
+						  struct radeon_ps *radeon_state,
+						  unsigned int first_arb_set)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	SMC_SIslands_MCArbDramTimingRegisterSet arb_regs = { 0 };
+	int i, ret = 0;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		ret = si_populate_memory_timing_parameters(rdev, &state->performance_levels[i], &arb_regs);
+		if (ret)
+			break;
+		ret = si_copy_bytes_to_smc(rdev,
+					   si_pi->arb_table_start +
+					   offsetof(SMC_SIslands_MCArbDramTimingRegisters, data) +
+					   sizeof(SMC_SIslands_MCArbDramTimingRegisterSet) * (first_arb_set + i),
+					   (u8 *)&arb_regs,
+					   sizeof(SMC_SIslands_MCArbDramTimingRegisterSet),
+					   si_pi->sram_end);
+		if (ret)
+			break;
+        }
+
+	return ret;
+}
+
+static int si_program_memory_timing_parameters(struct radeon_device *rdev,
+					       struct radeon_ps *radeon_new_state)
+{
+	return si_do_program_memory_timing_parameters(rdev, radeon_new_state,
+						      SISLANDS_DRIVER_STATE_ARB_INDEX);
+}
+
+static int si_populate_initial_mvdd_value(struct radeon_device *rdev,
+					  struct SISLANDS_SMC_VOLTAGE_VALUE *voltage)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	if (pi->mvdd_control)
+		return si_populate_voltage_value(rdev, &si_pi->mvdd_voltage_table,
+						 si_pi->mvdd_bootup_value, voltage);
+
+	return 0;
+}
+
+static int si_populate_smc_initial_state(struct radeon_device *rdev,
+					 struct radeon_ps *radeon_initial_state,
+					 SISLANDS_SMC_STATETABLE *table)
+{
+	struct ni_ps *initial_state = ni_get_ps(radeon_initial_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 reg;
+	int ret;
+
+	table->initialState.levels[0].mclk.vDLL_CNTL =
+		cpu_to_be32(si_pi->clock_registers.dll_cntl);
+	table->initialState.levels[0].mclk.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(si_pi->clock_registers.mclk_pwrmgt_cntl);
+	table->initialState.levels[0].mclk.vMPLL_AD_FUNC_CNTL =
+		cpu_to_be32(si_pi->clock_registers.mpll_ad_func_cntl);
+	table->initialState.levels[0].mclk.vMPLL_DQ_FUNC_CNTL =
+		cpu_to_be32(si_pi->clock_registers.mpll_dq_func_cntl);
+	table->initialState.levels[0].mclk.vMPLL_FUNC_CNTL =
+		cpu_to_be32(si_pi->clock_registers.mpll_func_cntl);
+	table->initialState.levels[0].mclk.vMPLL_FUNC_CNTL_1 =
+		cpu_to_be32(si_pi->clock_registers.mpll_func_cntl_1);
+	table->initialState.levels[0].mclk.vMPLL_FUNC_CNTL_2 =
+		cpu_to_be32(si_pi->clock_registers.mpll_func_cntl_2);
+	table->initialState.levels[0].mclk.vMPLL_SS =
+		cpu_to_be32(si_pi->clock_registers.mpll_ss1);
+	table->initialState.levels[0].mclk.vMPLL_SS2 =
+		cpu_to_be32(si_pi->clock_registers.mpll_ss2);
+
+	table->initialState.levels[0].mclk.mclk_value =
+		cpu_to_be32(initial_state->performance_levels[0].mclk);
+
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(si_pi->clock_registers.cg_spll_func_cntl);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(si_pi->clock_registers.cg_spll_func_cntl_2);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(si_pi->clock_registers.cg_spll_func_cntl_3);
+	table->initialState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_4 =
+		cpu_to_be32(si_pi->clock_registers.cg_spll_func_cntl_4);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM =
+		cpu_to_be32(si_pi->clock_registers.cg_spll_spread_spectrum);
+	table->initialState.levels[0].sclk.vCG_SPLL_SPREAD_SPECTRUM_2  =
+		cpu_to_be32(si_pi->clock_registers.cg_spll_spread_spectrum_2);
+
+	table->initialState.levels[0].sclk.sclk_value =
+		cpu_to_be32(initial_state->performance_levels[0].sclk);
+
+	table->initialState.levels[0].arbRefreshState =
+		SISLANDS_INITIAL_STATE_ARB_INDEX;
+
+	table->initialState.levels[0].ACIndex = 0;
+
+	ret = si_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+					initial_state->performance_levels[0].vddc,
+					&table->initialState.levels[0].vddc);
+
+	if (!ret) {
+		u16 std_vddc;
+
+		ret = si_get_std_voltage_value(rdev,
+					       &table->initialState.levels[0].vddc,
+					       &std_vddc);
+		if (!ret)
+			si_populate_std_voltage_value(rdev, std_vddc,
+						      table->initialState.levels[0].vddc.index,
+						      &table->initialState.levels[0].std_vddc);
+	}
+
+	if (eg_pi->vddci_control)
+		si_populate_voltage_value(rdev,
+					  &eg_pi->vddci_voltage_table,
+					  initial_state->performance_levels[0].vddci,
+					  &table->initialState.levels[0].vddci);
+
+	if (si_pi->vddc_phase_shed_control)
+		si_populate_phase_shedding_value(rdev,
+						 &rdev->pm.dpm.dyn_state.phase_shedding_limits_table,
+						 initial_state->performance_levels[0].vddc,
+						 initial_state->performance_levels[0].sclk,
+						 initial_state->performance_levels[0].mclk,
+						 &table->initialState.levels[0].vddc);
+
+	si_populate_initial_mvdd_value(rdev, &table->initialState.levels[0].mvdd);
+
+	reg = CG_R(0xffff) | CG_L(0);
+	table->initialState.levels[0].aT = cpu_to_be32(reg);
+
+	table->initialState.levels[0].bSP = cpu_to_be32(pi->dsp);
+
+	table->initialState.levels[0].gen2PCIE = (u8)si_pi->boot_pcie_gen;
+
+	if (pi->mem_gddr5) {
+		table->initialState.levels[0].strobeMode =
+			si_get_strobe_mode_settings(rdev,
+						    initial_state->performance_levels[0].mclk);
+
+		if (initial_state->performance_levels[0].mclk > pi->mclk_edc_enable_threshold)
+			table->initialState.levels[0].mcFlags = SISLANDS_SMC_MC_EDC_RD_FLAG | SISLANDS_SMC_MC_EDC_WR_FLAG;
+		else
+			table->initialState.levels[0].mcFlags =  0;
+	}
+
+	table->initialState.levelCount = 1;
+
+	table->initialState.flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	table->initialState.levels[0].dpm2.MaxPS = 0;
+	table->initialState.levels[0].dpm2.NearTDPDec = 0;
+	table->initialState.levels[0].dpm2.AboveSafeInc = 0;
+	table->initialState.levels[0].dpm2.BelowSafeInc = 0;
+	table->initialState.levels[0].dpm2.PwrEfficiencyRatio = 0;
+
+	reg = MIN_POWER_MASK | MAX_POWER_MASK;
+	table->initialState.levels[0].SQPowerThrottle = cpu_to_be32(reg);
+
+	reg = MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK;
+	table->initialState.levels[0].SQPowerThrottle_2 = cpu_to_be32(reg);
+
+	return 0;
+}
+
+static int si_populate_smc_acpi_state(struct radeon_device *rdev,
+				      SISLANDS_SMC_STATETABLE *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 spll_func_cntl = si_pi->clock_registers.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 = si_pi->clock_registers.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 = si_pi->clock_registers.cg_spll_func_cntl_3;
+	u32 spll_func_cntl_4 = si_pi->clock_registers.cg_spll_func_cntl_4;
+	u32 dll_cntl = si_pi->clock_registers.dll_cntl;
+	u32 mclk_pwrmgt_cntl = si_pi->clock_registers.mclk_pwrmgt_cntl;
+	u32 mpll_ad_func_cntl = si_pi->clock_registers.mpll_ad_func_cntl;
+	u32 mpll_dq_func_cntl = si_pi->clock_registers.mpll_dq_func_cntl;
+	u32 mpll_func_cntl = si_pi->clock_registers.mpll_func_cntl;
+	u32 mpll_func_cntl_1 = si_pi->clock_registers.mpll_func_cntl_1;
+	u32 mpll_func_cntl_2 = si_pi->clock_registers.mpll_func_cntl_2;
+	u32 reg;
+	int ret;
+
+	table->ACPIState = table->initialState;
+
+	table->ACPIState.flags &= ~PPSMC_SWSTATE_FLAG_DC;
+
+	if (pi->acpi_vddc) {
+		ret = si_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+						pi->acpi_vddc, &table->ACPIState.levels[0].vddc);
+		if (!ret) {
+			u16 std_vddc;
+
+			ret = si_get_std_voltage_value(rdev,
+						       &table->ACPIState.levels[0].vddc, &std_vddc);
+			if (!ret)
+				si_populate_std_voltage_value(rdev, std_vddc,
+							      table->ACPIState.levels[0].vddc.index,
+							      &table->ACPIState.levels[0].std_vddc);
+		}
+		table->ACPIState.levels[0].gen2PCIE = si_pi->acpi_pcie_gen;
+
+		if (si_pi->vddc_phase_shed_control) {
+			si_populate_phase_shedding_value(rdev,
+							 &rdev->pm.dpm.dyn_state.phase_shedding_limits_table,
+							 pi->acpi_vddc,
+							 0,
+							 0,
+							 &table->ACPIState.levels[0].vddc);
+		}
+	} else {
+		ret = si_populate_voltage_value(rdev, &eg_pi->vddc_voltage_table,
+						pi->min_vddc_in_table, &table->ACPIState.levels[0].vddc);
+		if (!ret) {
+			u16 std_vddc;
+
+			ret = si_get_std_voltage_value(rdev,
+						       &table->ACPIState.levels[0].vddc, &std_vddc);
+
+			if (!ret)
+				si_populate_std_voltage_value(rdev, std_vddc,
+							      table->ACPIState.levels[0].vddc.index,
+							      &table->ACPIState.levels[0].std_vddc);
+		}
+		table->ACPIState.levels[0].gen2PCIE = (u8)r600_get_pcie_gen_support(rdev,
+										    si_pi->sys_pcie_mask,
+										    si_pi->boot_pcie_gen,
+										    RADEON_PCIE_GEN1);
+
+		if (si_pi->vddc_phase_shed_control)
+			si_populate_phase_shedding_value(rdev,
+							 &rdev->pm.dpm.dyn_state.phase_shedding_limits_table,
+							 pi->min_vddc_in_table,
+							 0,
+							 0,
+							 &table->ACPIState.levels[0].vddc);
+	}
+
+	if (pi->acpi_vddc) {
+		if (eg_pi->acpi_vddci)
+			si_populate_voltage_value(rdev, &eg_pi->vddci_voltage_table,
+						  eg_pi->acpi_vddci,
+						  &table->ACPIState.levels[0].vddci);
+	}
+
+	mclk_pwrmgt_cntl |= MRDCK0_RESET | MRDCK1_RESET;
+	mclk_pwrmgt_cntl &= ~(MRDCK0_PDNB | MRDCK1_PDNB);
+
+	dll_cntl &= ~(MRDCK0_BYPASS | MRDCK1_BYPASS);
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(4);
+
+	table->ACPIState.levels[0].mclk.vDLL_CNTL =
+		cpu_to_be32(dll_cntl);
+	table->ACPIState.levels[0].mclk.vMCLK_PWRMGT_CNTL =
+		cpu_to_be32(mclk_pwrmgt_cntl);
+	table->ACPIState.levels[0].mclk.vMPLL_AD_FUNC_CNTL =
+		cpu_to_be32(mpll_ad_func_cntl);
+	table->ACPIState.levels[0].mclk.vMPLL_DQ_FUNC_CNTL =
+		cpu_to_be32(mpll_dq_func_cntl);
+	table->ACPIState.levels[0].mclk.vMPLL_FUNC_CNTL =
+		cpu_to_be32(mpll_func_cntl);
+	table->ACPIState.levels[0].mclk.vMPLL_FUNC_CNTL_1 =
+		cpu_to_be32(mpll_func_cntl_1);
+	table->ACPIState.levels[0].mclk.vMPLL_FUNC_CNTL_2 =
+		cpu_to_be32(mpll_func_cntl_2);
+	table->ACPIState.levels[0].mclk.vMPLL_SS =
+		cpu_to_be32(si_pi->clock_registers.mpll_ss1);
+	table->ACPIState.levels[0].mclk.vMPLL_SS2 =
+		cpu_to_be32(si_pi->clock_registers.mpll_ss2);
+
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL =
+		cpu_to_be32(spll_func_cntl);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_2 =
+		cpu_to_be32(spll_func_cntl_2);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_3 =
+		cpu_to_be32(spll_func_cntl_3);
+	table->ACPIState.levels[0].sclk.vCG_SPLL_FUNC_CNTL_4 =
+		cpu_to_be32(spll_func_cntl_4);
+
+	table->ACPIState.levels[0].mclk.mclk_value = 0;
+	table->ACPIState.levels[0].sclk.sclk_value = 0;
+
+	si_populate_mvdd_value(rdev, 0, &table->ACPIState.levels[0].mvdd);
+
+	if (eg_pi->dynamic_ac_timing)
+		table->ACPIState.levels[0].ACIndex = 0;
+
+	table->ACPIState.levels[0].dpm2.MaxPS = 0;
+	table->ACPIState.levels[0].dpm2.NearTDPDec = 0;
+	table->ACPIState.levels[0].dpm2.AboveSafeInc = 0;
+	table->ACPIState.levels[0].dpm2.BelowSafeInc = 0;
+	table->ACPIState.levels[0].dpm2.PwrEfficiencyRatio = 0;
+
+	reg = MIN_POWER_MASK | MAX_POWER_MASK;
+	table->ACPIState.levels[0].SQPowerThrottle = cpu_to_be32(reg);
+
+	reg = MAX_POWER_DELTA_MASK | STI_SIZE_MASK | LTI_RATIO_MASK;
+	table->ACPIState.levels[0].SQPowerThrottle_2 = cpu_to_be32(reg);
+
+	return 0;
+}
+
+static int si_populate_ulv_state(struct radeon_device *rdev,
+				 SISLANDS_SMC_SWSTATE *state)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct si_ulv_param *ulv = &si_pi->ulv;
+	u32 sclk_in_sr = 1350; /* ??? */
+	int ret;
+
+	ret = si_convert_power_level_to_smc(rdev, &ulv->pl,
+					    &state->levels[0]);
+	if (!ret) {
+		if (eg_pi->sclk_deep_sleep) {
+			if (sclk_in_sr <= SCLK_MIN_DEEPSLEEP_FREQ)
+				state->levels[0].stateFlags |= PPSMC_STATEFLAG_DEEPSLEEP_BYPASS;
+			else
+				state->levels[0].stateFlags |= PPSMC_STATEFLAG_DEEPSLEEP_THROTTLE;
+		}
+		if (ulv->one_pcie_lane_in_ulv)
+			state->flags |= PPSMC_SWSTATE_FLAG_PCIE_X1;
+		state->levels[0].arbRefreshState = (u8)(SISLANDS_ULV_STATE_ARB_INDEX);
+		state->levels[0].ACIndex = 1;
+		state->levels[0].std_vddc = state->levels[0].vddc;
+		state->levelCount = 1;
+
+		state->flags |= PPSMC_SWSTATE_FLAG_DC;
+	}
+
+	return ret;
+}
+
+static int si_program_ulv_memory_timing_parameters(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct si_ulv_param *ulv = &si_pi->ulv;
+	SMC_SIslands_MCArbDramTimingRegisterSet arb_regs = { 0 };
+	int ret;
+
+	ret = si_populate_memory_timing_parameters(rdev, &ulv->pl,
+						   &arb_regs);
+	if (ret)
+		return ret;
+
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_ulv_volt_change_delay,
+				   ulv->volt_change_delay);
+
+	ret = si_copy_bytes_to_smc(rdev,
+				   si_pi->arb_table_start +
+				   offsetof(SMC_SIslands_MCArbDramTimingRegisters, data) +
+				   sizeof(SMC_SIslands_MCArbDramTimingRegisterSet) * SISLANDS_ULV_STATE_ARB_INDEX,
+				   (u8 *)&arb_regs,
+				   sizeof(SMC_SIslands_MCArbDramTimingRegisterSet),
+				   si_pi->sram_end);
+
+	return ret;
+}
+
+static void si_get_mvdd_configuration(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+
+	pi->mvdd_split_frequency = 30000;
+}
+
+static int si_init_smc_table(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct radeon_ps *radeon_boot_state = rdev->pm.dpm.boot_ps;
+	const struct si_ulv_param *ulv = &si_pi->ulv;
+	SISLANDS_SMC_STATETABLE  *table = &si_pi->smc_statetable;
+	int ret;
+	u32 lane_width;
+	u32 vr_hot_gpio;
+
+	si_populate_smc_voltage_tables(rdev, table);
+
+	switch (rdev->pm.int_thermal_type) {
+	case THERMAL_TYPE_SI:
+	case THERMAL_TYPE_EMC2103_WITH_INTERNAL:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_INTERNAL;
+		break;
+	case THERMAL_TYPE_NONE:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_NONE;
+		break;
+	default:
+		table->thermalProtectType = PPSMC_THERMAL_PROTECT_TYPE_EXTERNAL;
+		break;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_HARDWAREDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GPIO_DC;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REGULATOR_HOT) {
+		if ((rdev->pdev->device != 0x6818) && (rdev->pdev->device != 0x6819))
+			table->systemFlags |= PPSMC_SYSTEMFLAG_REGULATOR_HOT;
+	}
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_STEPVDDC)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_STEPVDDC;
+
+	if (pi->mem_gddr5)
+		table->systemFlags |= PPSMC_SYSTEMFLAG_GDDR5;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_REVERT_GPIO5_POLARITY)
+		table->systemFlags |= PPSMC_EXTRAFLAGS_AC2DC_GPIO5_POLARITY_HIGH;
+
+	if (rdev->pm.dpm.platform_caps & ATOM_PP_PLATFORM_CAP_VRHOT_GPIO_CONFIGURABLE) {
+		table->systemFlags |= PPSMC_SYSTEMFLAG_REGULATOR_HOT_PROG_GPIO;
+		vr_hot_gpio = rdev->pm.dpm.backbias_response_time;
+		si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_vr_hot_gpio,
+					   vr_hot_gpio);
+	}
+
+	ret = si_populate_smc_initial_state(rdev, radeon_boot_state, table);
+	if (ret)
+		return ret;
+
+	ret = si_populate_smc_acpi_state(rdev, table);
+	if (ret)
+		return ret;
+
+	table->driverState = table->initialState;
+
+	ret = si_do_program_memory_timing_parameters(rdev, radeon_boot_state,
+						     SISLANDS_INITIAL_STATE_ARB_INDEX);
+	if (ret)
+		return ret;
+
+	if (ulv->supported && ulv->pl.vddc) {
+		ret = si_populate_ulv_state(rdev, &table->ULVState);
+		if (ret)
+			return ret;
+
+		ret = si_program_ulv_memory_timing_parameters(rdev);
+		if (ret)
+			return ret;
+
+		WREG32(CG_ULV_CONTROL, ulv->cg_ulv_control);
+		WREG32(CG_ULV_PARAMETER, ulv->cg_ulv_parameter);
+
+		lane_width = radeon_get_pcie_lanes(rdev);
+		si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_non_ulv_pcie_link_width, lane_width);
+	} else {
+		table->ULVState = table->initialState;
+	}
+
+	return si_copy_bytes_to_smc(rdev, si_pi->state_table_start,
+				    (u8 *)table, sizeof(SISLANDS_SMC_STATETABLE),
+				    si_pi->sram_end);
+}
+
+static int si_calculate_sclk_params(struct radeon_device *rdev,
+				    u32 engine_clock,
+				    SISLANDS_SMC_SCLK_VALUE *sclk)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	u32 spll_func_cntl = si_pi->clock_registers.cg_spll_func_cntl;
+	u32 spll_func_cntl_2 = si_pi->clock_registers.cg_spll_func_cntl_2;
+	u32 spll_func_cntl_3 = si_pi->clock_registers.cg_spll_func_cntl_3;
+	u32 spll_func_cntl_4 = si_pi->clock_registers.cg_spll_func_cntl_4;
+	u32 cg_spll_spread_spectrum = si_pi->clock_registers.cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2 = si_pi->clock_registers.cg_spll_spread_spectrum_2;
+	u64 tmp;
+	u32 reference_clock = rdev->clock.spll.reference_freq;
+	u32 reference_divider;
+	u32 fbdiv;
+	int ret;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     engine_clock, false, &dividers);
+	if (ret)
+		return ret;
+
+	reference_divider = 1 + dividers.ref_div;
+
+	tmp = (u64) engine_clock * reference_divider * dividers.post_div * 16384;
+	do_div(tmp, reference_clock);
+	fbdiv = (u32) tmp;
+
+	spll_func_cntl &= ~(SPLL_PDIV_A_MASK | SPLL_REF_DIV_MASK);
+	spll_func_cntl |= SPLL_REF_DIV(dividers.ref_div);
+	spll_func_cntl |= SPLL_PDIV_A(dividers.post_div);
+
+	spll_func_cntl_2 &= ~SCLK_MUX_SEL_MASK;
+	spll_func_cntl_2 |= SCLK_MUX_SEL(2);
+
+        spll_func_cntl_3 &= ~SPLL_FB_DIV_MASK;
+        spll_func_cntl_3 |= SPLL_FB_DIV(fbdiv);
+        spll_func_cntl_3 |= SPLL_DITHEN;
+
+	if (pi->sclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 vco_freq = engine_clock * dividers.post_div;
+
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+						     ASIC_INTERNAL_ENGINE_SS, vco_freq)) {
+			u32 clk_s = reference_clock * 5 / (reference_divider * ss.rate);
+			u32 clk_v = 4 * ss.percentage * fbdiv / (clk_s * 10000);
+
+			cg_spll_spread_spectrum &= ~CLK_S_MASK;
+			cg_spll_spread_spectrum |= CLK_S(clk_s);
+			cg_spll_spread_spectrum |= SSEN;
+
+			cg_spll_spread_spectrum_2 &= ~CLK_V_MASK;
+			cg_spll_spread_spectrum_2 |= CLK_V(clk_v);
+		}
+	}
+
+	sclk->sclk_value = engine_clock;
+	sclk->vCG_SPLL_FUNC_CNTL = spll_func_cntl;
+	sclk->vCG_SPLL_FUNC_CNTL_2 = spll_func_cntl_2;
+	sclk->vCG_SPLL_FUNC_CNTL_3 = spll_func_cntl_3;
+	sclk->vCG_SPLL_FUNC_CNTL_4 = spll_func_cntl_4;
+	sclk->vCG_SPLL_SPREAD_SPECTRUM = cg_spll_spread_spectrum;
+	sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cg_spll_spread_spectrum_2;
+
+	return 0;
+}
+
+static int si_populate_sclk_value(struct radeon_device *rdev,
+				  u32 engine_clock,
+				  SISLANDS_SMC_SCLK_VALUE *sclk)
+{
+	SISLANDS_SMC_SCLK_VALUE sclk_tmp;
+	int ret;
+
+	ret = si_calculate_sclk_params(rdev, engine_clock, &sclk_tmp);
+	if (!ret) {
+		sclk->sclk_value = cpu_to_be32(sclk_tmp.sclk_value);
+		sclk->vCG_SPLL_FUNC_CNTL = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL);
+		sclk->vCG_SPLL_FUNC_CNTL_2 = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL_2);
+		sclk->vCG_SPLL_FUNC_CNTL_3 = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL_3);
+		sclk->vCG_SPLL_FUNC_CNTL_4 = cpu_to_be32(sclk_tmp.vCG_SPLL_FUNC_CNTL_4);
+		sclk->vCG_SPLL_SPREAD_SPECTRUM = cpu_to_be32(sclk_tmp.vCG_SPLL_SPREAD_SPECTRUM);
+		sclk->vCG_SPLL_SPREAD_SPECTRUM_2 = cpu_to_be32(sclk_tmp.vCG_SPLL_SPREAD_SPECTRUM_2);
+	}
+
+	return ret;
+}
+
+static int si_populate_mclk_value(struct radeon_device *rdev,
+				  u32 engine_clock,
+				  u32 memory_clock,
+				  SISLANDS_SMC_MCLK_VALUE *mclk,
+				  bool strobe_mode,
+				  bool dll_state_on)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32  dll_cntl = si_pi->clock_registers.dll_cntl;
+	u32  mclk_pwrmgt_cntl = si_pi->clock_registers.mclk_pwrmgt_cntl;
+	u32  mpll_ad_func_cntl = si_pi->clock_registers.mpll_ad_func_cntl;
+	u32  mpll_dq_func_cntl = si_pi->clock_registers.mpll_dq_func_cntl;
+	u32  mpll_func_cntl = si_pi->clock_registers.mpll_func_cntl;
+	u32  mpll_func_cntl_1 = si_pi->clock_registers.mpll_func_cntl_1;
+	u32  mpll_func_cntl_2 = si_pi->clock_registers.mpll_func_cntl_2;
+	u32  mpll_ss1 = si_pi->clock_registers.mpll_ss1;
+	u32  mpll_ss2 = si_pi->clock_registers.mpll_ss2;
+	struct atom_mpll_param mpll_param;
+	int ret;
+
+	ret = radeon_atom_get_memory_pll_dividers(rdev, memory_clock, strobe_mode, &mpll_param);
+	if (ret)
+		return ret;
+
+	mpll_func_cntl &= ~BWCTRL_MASK;
+	mpll_func_cntl |= BWCTRL(mpll_param.bwcntl);
+
+	mpll_func_cntl_1 &= ~(CLKF_MASK | CLKFRAC_MASK | VCO_MODE_MASK);
+	mpll_func_cntl_1 |= CLKF(mpll_param.clkf) |
+		CLKFRAC(mpll_param.clkfrac) | VCO_MODE(mpll_param.vco_mode);
+
+	mpll_ad_func_cntl &= ~YCLK_POST_DIV_MASK;
+	mpll_ad_func_cntl |= YCLK_POST_DIV(mpll_param.post_div);
+
+	if (pi->mem_gddr5) {
+		mpll_dq_func_cntl &= ~(YCLK_SEL_MASK | YCLK_POST_DIV_MASK);
+		mpll_dq_func_cntl |= YCLK_SEL(mpll_param.yclk_sel) |
+			YCLK_POST_DIV(mpll_param.post_div);
+	}
+
+	if (pi->mclk_ss) {
+		struct radeon_atom_ss ss;
+		u32 freq_nom;
+		u32 tmp;
+		u32 reference_clock = rdev->clock.mpll.reference_freq;
+
+		if (pi->mem_gddr5)
+			freq_nom = memory_clock * 4;
+		else
+			freq_nom = memory_clock * 2;
+
+		tmp = freq_nom / reference_clock;
+		tmp = tmp * tmp;
+		if (radeon_atombios_get_asic_ss_info(rdev, &ss,
+                                                     ASIC_INTERNAL_MEMORY_SS, freq_nom)) {
+			u32 clks = reference_clock * 5 / ss.rate;
+			u32 clkv = (u32)((((131 * ss.percentage * ss.rate) / 100) * tmp) / freq_nom);
+
+                        mpll_ss1 &= ~CLKV_MASK;
+                        mpll_ss1 |= CLKV(clkv);
+
+                        mpll_ss2 &= ~CLKS_MASK;
+                        mpll_ss2 |= CLKS(clks);
+		}
+	}
+
+	mclk_pwrmgt_cntl &= ~DLL_SPEED_MASK;
+	mclk_pwrmgt_cntl |= DLL_SPEED(mpll_param.dll_speed);
+
+	if (dll_state_on)
+		mclk_pwrmgt_cntl |= MRDCK0_PDNB | MRDCK1_PDNB;
+	else
+		mclk_pwrmgt_cntl &= ~(MRDCK0_PDNB | MRDCK1_PDNB);
+
+	mclk->mclk_value = cpu_to_be32(memory_clock);
+	mclk->vMPLL_FUNC_CNTL = cpu_to_be32(mpll_func_cntl);
+	mclk->vMPLL_FUNC_CNTL_1 = cpu_to_be32(mpll_func_cntl_1);
+	mclk->vMPLL_FUNC_CNTL_2 = cpu_to_be32(mpll_func_cntl_2);
+	mclk->vMPLL_AD_FUNC_CNTL = cpu_to_be32(mpll_ad_func_cntl);
+	mclk->vMPLL_DQ_FUNC_CNTL = cpu_to_be32(mpll_dq_func_cntl);
+	mclk->vMCLK_PWRMGT_CNTL = cpu_to_be32(mclk_pwrmgt_cntl);
+	mclk->vDLL_CNTL = cpu_to_be32(dll_cntl);
+	mclk->vMPLL_SS = cpu_to_be32(mpll_ss1);
+	mclk->vMPLL_SS2 = cpu_to_be32(mpll_ss2);
+
+	return 0;
+}
+
+static void si_populate_smc_sp(struct radeon_device *rdev,
+			       struct radeon_ps *radeon_state,
+			       SISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct ni_ps *ps = ni_get_ps(radeon_state);
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	int i;
+
+	for (i = 0; i < ps->performance_level_count - 1; i++)
+		smc_state->levels[i].bSP = cpu_to_be32(pi->dsp);
+
+	smc_state->levels[ps->performance_level_count - 1].bSP =
+		cpu_to_be32(pi->psp);
+}
+
+static int si_convert_power_level_to_smc(struct radeon_device *rdev,
+					 struct rv7xx_pl *pl,
+					 SISLANDS_SMC_HW_PERFORMANCE_LEVEL *level)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	int ret;
+	bool dll_state_on;
+	u16 std_vddc;
+	bool gmc_pg = false;
+
+	if (eg_pi->pcie_performance_request &&
+	    (si_pi->force_pcie_gen != RADEON_PCIE_GEN_INVALID))
+		level->gen2PCIE = (u8)si_pi->force_pcie_gen;
+	else
+		level->gen2PCIE = (u8)pl->pcie_gen;
+
+	ret = si_populate_sclk_value(rdev, pl->sclk, &level->sclk);
+	if (ret)
+		return ret;
+
+	level->mcFlags =  0;
+
+	if (pi->mclk_stutter_mode_threshold &&
+	    (pl->mclk <= pi->mclk_stutter_mode_threshold) &&
+	    !eg_pi->uvd_enabled &&
+	    (RREG32(DPG_PIPE_STUTTER_CONTROL) & STUTTER_ENABLE) &&
+	    (rdev->pm.dpm.new_active_crtc_count <= 2)) {
+		level->mcFlags |= SISLANDS_SMC_MC_STUTTER_EN;
+
+		if (gmc_pg)
+			level->mcFlags |= SISLANDS_SMC_MC_PG_EN;
+	}
+
+	if (pi->mem_gddr5) {
+		if (pl->mclk > pi->mclk_edc_enable_threshold)
+			level->mcFlags |= SISLANDS_SMC_MC_EDC_RD_FLAG;
+
+		if (pl->mclk > eg_pi->mclk_edc_wr_enable_threshold)
+			level->mcFlags |= SISLANDS_SMC_MC_EDC_WR_FLAG;
+
+		level->strobeMode = si_get_strobe_mode_settings(rdev, pl->mclk);
+
+		if (level->strobeMode & SISLANDS_SMC_STROBE_ENABLE) {
+			if (si_get_mclk_frequency_ratio(pl->mclk, true) >=
+			    ((RREG32(MC_SEQ_MISC7) >> 16) & 0xf))
+				dll_state_on = ((RREG32(MC_SEQ_MISC5) >> 1) & 0x1) ? true : false;
+			else
+				dll_state_on = ((RREG32(MC_SEQ_MISC6) >> 1) & 0x1) ? true : false;
+		} else {
+			dll_state_on = false;
+		}
+	} else {
+		level->strobeMode = si_get_strobe_mode_settings(rdev,
+								pl->mclk);
+
+		dll_state_on = ((RREG32(MC_SEQ_MISC5) >> 1) & 0x1) ? true : false;
+	}
+
+	ret = si_populate_mclk_value(rdev,
+				     pl->sclk,
+				     pl->mclk,
+				     &level->mclk,
+				     (level->strobeMode & SISLANDS_SMC_STROBE_ENABLE) != 0, dll_state_on);
+	if (ret)
+		return ret;
+
+	ret = si_populate_voltage_value(rdev,
+					&eg_pi->vddc_voltage_table,
+					pl->vddc, &level->vddc);
+	if (ret)
+		return ret;
+
+
+	ret = si_get_std_voltage_value(rdev, &level->vddc, &std_vddc);
+	if (ret)
+		return ret;
+
+	ret = si_populate_std_voltage_value(rdev, std_vddc,
+					    level->vddc.index, &level->std_vddc);
+	if (ret)
+		return ret;
+
+	if (eg_pi->vddci_control) {
+		ret = si_populate_voltage_value(rdev, &eg_pi->vddci_voltage_table,
+						pl->vddci, &level->vddci);
+		if (ret)
+			return ret;
+	}
+
+	if (si_pi->vddc_phase_shed_control) {
+		ret = si_populate_phase_shedding_value(rdev,
+						       &rdev->pm.dpm.dyn_state.phase_shedding_limits_table,
+						       pl->vddc,
+						       pl->sclk,
+						       pl->mclk,
+						       &level->vddc);
+		if (ret)
+			return ret;
+	}
+
+	level->MaxPoweredUpCU = si_pi->max_cu;
+
+	ret = si_populate_mvdd_value(rdev, pl->mclk, &level->mvdd);
+
+	return ret;
+}
+
+static int si_populate_smc_t(struct radeon_device *rdev,
+			     struct radeon_ps *radeon_state,
+			     SISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	u32 a_t;
+	u32 t_l, t_h;
+	u32 high_bsp;
+	int i, ret;
+
+	if (state->performance_level_count >= 9)
+		return -EINVAL;
+
+	if (state->performance_level_count < 2) {
+		a_t = CG_R(0xffff) | CG_L(0);
+		smc_state->levels[0].aT = cpu_to_be32(a_t);
+		return 0;
+	}
+
+	smc_state->levels[0].aT = cpu_to_be32(0);
+
+	for (i = 0; i <= state->performance_level_count - 2; i++) {
+		ret = r600_calculate_at(
+			(50 / SISLANDS_MAX_HARDWARE_POWERLEVELS) * 100 * (i + 1),
+			100 * R600_AH_DFLT,
+			state->performance_levels[i + 1].sclk,
+			state->performance_levels[i].sclk,
+			&t_l,
+			&t_h);
+
+		if (ret) {
+			t_h = (i + 1) * 1000 - 50 * R600_AH_DFLT;
+			t_l = (i + 1) * 1000 + 50 * R600_AH_DFLT;
+		}
+
+		a_t = be32_to_cpu(smc_state->levels[i].aT) & ~CG_R_MASK;
+		a_t |= CG_R(t_l * pi->bsp / 20000);
+		smc_state->levels[i].aT = cpu_to_be32(a_t);
+
+		high_bsp = (i == state->performance_level_count - 2) ?
+			pi->pbsp : pi->bsp;
+		a_t = CG_R(0xffff) | CG_L(t_h * high_bsp / 20000);
+		smc_state->levels[i + 1].aT = cpu_to_be32(a_t);
+	}
+
+	return 0;
+}
+
+static int si_disable_ulv(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct si_ulv_param *ulv = &si_pi->ulv;
+
+	if (ulv->supported)
+		return (si_send_msg_to_smc(rdev, PPSMC_MSG_DisableULV) == PPSMC_Result_OK) ?
+			0 : -EINVAL;
+
+	return 0;
+}
+
+static bool si_is_state_ulv_compatible(struct radeon_device *rdev,
+				       struct radeon_ps *radeon_state)
+{
+	const struct si_power_info *si_pi = si_get_pi(rdev);
+	const struct si_ulv_param *ulv = &si_pi->ulv;
+	const struct ni_ps *state = ni_get_ps(radeon_state);
+	int i;
+
+	if (state->performance_levels[0].mclk != ulv->pl.mclk)
+		return false;
+
+	/* XXX validate against display requirements! */
+
+	for (i = 0; i < rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.count; i++) {
+		if (rdev->clock.current_dispclk <=
+		    rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[i].clk) {
+			if (ulv->pl.vddc <
+			    rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[i].v)
+				return false;
+		}
+	}
+
+	if ((radeon_state->vclk != 0) || (radeon_state->dclk != 0))
+		return false;
+
+	return true;
+}
+
+static int si_set_power_state_conditionally_enable_ulv(struct radeon_device *rdev,
+						       struct radeon_ps *radeon_new_state)
+{
+	const struct si_power_info *si_pi = si_get_pi(rdev);
+	const struct si_ulv_param *ulv = &si_pi->ulv;
+
+	if (ulv->supported) {
+		if (si_is_state_ulv_compatible(rdev, radeon_new_state))
+			return (si_send_msg_to_smc(rdev, PPSMC_MSG_EnableULV) == PPSMC_Result_OK) ?
+				0 : -EINVAL;
+	}
+	return 0;
+}
+
+static int si_convert_power_state_to_smc(struct radeon_device *rdev,
+					 struct radeon_ps *radeon_state,
+					 SISLANDS_SMC_SWSTATE *smc_state)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct ni_power_info *ni_pi = ni_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	int i, ret;
+	u32 threshold;
+	u32 sclk_in_sr = 1350; /* ??? */
+
+	if (state->performance_level_count > SISLANDS_MAX_HARDWARE_POWERLEVELS)
+		return -EINVAL;
+
+	threshold = state->performance_levels[state->performance_level_count-1].sclk * 100 / 100;
+
+	if (radeon_state->vclk && radeon_state->dclk) {
+		eg_pi->uvd_enabled = true;
+		if (eg_pi->smu_uvd_hs)
+			smc_state->flags |= PPSMC_SWSTATE_FLAG_UVD;
+	} else {
+		eg_pi->uvd_enabled = false;
+	}
+
+	if (state->dc_compatible)
+		smc_state->flags |= PPSMC_SWSTATE_FLAG_DC;
+
+	smc_state->levelCount = 0;
+	for (i = 0; i < state->performance_level_count; i++) {
+		if (eg_pi->sclk_deep_sleep) {
+			if ((i == 0) || si_pi->sclk_deep_sleep_above_low) {
+				if (sclk_in_sr <= SCLK_MIN_DEEPSLEEP_FREQ)
+					smc_state->levels[i].stateFlags |= PPSMC_STATEFLAG_DEEPSLEEP_BYPASS;
+				else
+					smc_state->levels[i].stateFlags |= PPSMC_STATEFLAG_DEEPSLEEP_THROTTLE;
+			}
+		}
+
+		ret = si_convert_power_level_to_smc(rdev, &state->performance_levels[i],
+						    &smc_state->levels[i]);
+		smc_state->levels[i].arbRefreshState =
+			(u8)(SISLANDS_DRIVER_STATE_ARB_INDEX + i);
+
+		if (ret)
+			return ret;
+
+		if (ni_pi->enable_power_containment)
+			smc_state->levels[i].displayWatermark =
+				(state->performance_levels[i].sclk < threshold) ?
+				PPSMC_DISPLAY_WATERMARK_LOW : PPSMC_DISPLAY_WATERMARK_HIGH;
+		else
+			smc_state->levels[i].displayWatermark = (i < 2) ?
+				PPSMC_DISPLAY_WATERMARK_LOW : PPSMC_DISPLAY_WATERMARK_HIGH;
+
+		if (eg_pi->dynamic_ac_timing)
+			smc_state->levels[i].ACIndex = SISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT + i;
+		else
+			smc_state->levels[i].ACIndex = 0;
+
+		smc_state->levelCount++;
+	}
+
+	si_write_smc_soft_register(rdev,
+				   SI_SMC_SOFT_REGISTER_watermark_threshold,
+				   threshold / 512);
+
+	si_populate_smc_sp(rdev, radeon_state, smc_state);
+
+	ret = si_populate_power_containment_values(rdev, radeon_state, smc_state);
+	if (ret)
+		ni_pi->enable_power_containment = false;
+
+	ret = si_populate_sq_ramping_values(rdev, radeon_state, smc_state);
+        if (ret)
+		ni_pi->enable_sq_ramping = false;
+
+	return si_populate_smc_t(rdev, radeon_state, smc_state);
+}
+
+static int si_upload_sw_state(struct radeon_device *rdev,
+			      struct radeon_ps *radeon_new_state)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct ni_ps *new_state = ni_get_ps(radeon_new_state);
+	int ret;
+	u32 address = si_pi->state_table_start +
+		offsetof(SISLANDS_SMC_STATETABLE, driverState);
+	u32 state_size = sizeof(SISLANDS_SMC_SWSTATE) +
+		((new_state->performance_level_count - 1) *
+		 sizeof(SISLANDS_SMC_HW_PERFORMANCE_LEVEL));
+	SISLANDS_SMC_SWSTATE *smc_state = &si_pi->smc_statetable.driverState;
+
+	memset(smc_state, 0, state_size);
+
+	ret = si_convert_power_state_to_smc(rdev, radeon_new_state, smc_state);
+	if (ret)
+		return ret;
+
+	ret = si_copy_bytes_to_smc(rdev, address, (u8 *)smc_state,
+				   state_size, si_pi->sram_end);
+
+	return ret;
+}
+
+static int si_upload_ulv_state(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct si_ulv_param *ulv = &si_pi->ulv;
+	int ret = 0;
+
+	if (ulv->supported && ulv->pl.vddc) {
+		u32 address = si_pi->state_table_start +
+			offsetof(SISLANDS_SMC_STATETABLE, ULVState);
+		SISLANDS_SMC_SWSTATE *smc_state = &si_pi->smc_statetable.ULVState;
+		u32 state_size = sizeof(SISLANDS_SMC_SWSTATE);
+
+		memset(smc_state, 0, state_size);
+
+		ret = si_populate_ulv_state(rdev, smc_state);
+		if (!ret)
+			ret = si_copy_bytes_to_smc(rdev, address, (u8 *)smc_state,
+						   state_size, si_pi->sram_end);
+	}
+
+	return ret;
+}
+
+static int si_upload_smc_data(struct radeon_device *rdev)
+{
+	struct radeon_crtc *radeon_crtc = NULL;
+	int i;
+
+	if (rdev->pm.dpm.new_active_crtc_count == 0)
+		return 0;
+
+	for (i = 0; i < rdev->num_crtc; i++) {
+		if (rdev->pm.dpm.new_active_crtcs & (1 << i)) {
+			radeon_crtc = rdev->mode_info.crtcs[i];
+			break;
+		}
+	}
+
+	if (radeon_crtc == NULL)
+		return 0;
+
+	if (radeon_crtc->line_time <= 0)
+		return 0;
+
+	if (si_write_smc_soft_register(rdev,
+				       SI_SMC_SOFT_REGISTER_crtc_index,
+				       radeon_crtc->crtc_id) != PPSMC_Result_OK)
+		return 0;
+
+	if (si_write_smc_soft_register(rdev,
+				       SI_SMC_SOFT_REGISTER_mclk_change_block_cp_min,
+				       radeon_crtc->wm_high / radeon_crtc->line_time) != PPSMC_Result_OK)
+		return 0;
+
+	if (si_write_smc_soft_register(rdev,
+				       SI_SMC_SOFT_REGISTER_mclk_change_block_cp_max,
+				       radeon_crtc->wm_low / radeon_crtc->line_time) != PPSMC_Result_OK)
+		return 0;
+
+	return 0;
+}
+
+static int si_set_mc_special_registers(struct radeon_device *rdev,
+				       struct si_mc_reg_table *table)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	u8 i, j, k;
+	u32 temp_reg;
+
+	for (i = 0, j = table->last; i < table->last; i++) {
+		if (j >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE)
+			return -EINVAL;
+		switch (table->mc_reg_address[i].s1 << 2) {
+		case MC_SEQ_MISC1:
+			temp_reg = RREG32(MC_PMG_CMD_EMRS);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_EMRS >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+			for (k = 0; k < table->num_entries; k++)
+				table->mc_reg_table_entry[k].mc_data[j] =
+					((temp_reg & 0xffff0000)) |
+					((table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16);
+			j++;
+			if (j >= SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+
+			temp_reg = RREG32(MC_PMG_CMD_MRS);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+			for (k = 0; k < table->num_entries; k++) {
+				table->mc_reg_table_entry[k].mc_data[j] =
+					(temp_reg & 0xffff0000) |
+					(table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff);
+				if (!pi->mem_gddr5)
+					table->mc_reg_table_entry[k].mc_data[j] |= 0x100;
+			}
+			j++;
+			if (j > SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+
+			if (!pi->mem_gddr5) {
+				table->mc_reg_address[j].s1 = MC_PMG_AUTO_CMD >> 2;
+				table->mc_reg_address[j].s0 = MC_PMG_AUTO_CMD >> 2;
+				for (k = 0; k < table->num_entries; k++)
+					table->mc_reg_table_entry[k].mc_data[j] =
+						(table->mc_reg_table_entry[k].mc_data[i] & 0xffff0000) >> 16;
+				j++;
+				if (j > SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE)
+					return -EINVAL;
+			}
+			break;
+		case MC_SEQ_RESERVE_M:
+			temp_reg = RREG32(MC_PMG_CMD_MRS1);
+			table->mc_reg_address[j].s1 = MC_PMG_CMD_MRS1 >> 2;
+			table->mc_reg_address[j].s0 = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+			for(k = 0; k < table->num_entries; k++)
+				table->mc_reg_table_entry[k].mc_data[j] =
+					(temp_reg & 0xffff0000) |
+					(table->mc_reg_table_entry[k].mc_data[i] & 0x0000ffff);
+			j++;
+			if (j > SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE)
+				return -EINVAL;
+			break;
+		default:
+			break;
+		}
+	}
+
+	table->last = j;
+
+	return 0;
+}
+
+static bool si_check_s0_mc_reg_index(u16 in_reg, u16 *out_reg)
+{
+	bool result = true;
+
+	switch (in_reg) {
+	case  MC_SEQ_RAS_TIMING >> 2:
+		*out_reg = MC_SEQ_RAS_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_CAS_TIMING >> 2:
+		*out_reg = MC_SEQ_CAS_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_MISC_TIMING >> 2:
+		*out_reg = MC_SEQ_MISC_TIMING_LP >> 2;
+		break;
+        case MC_SEQ_MISC_TIMING2 >> 2:
+		*out_reg = MC_SEQ_MISC_TIMING2_LP >> 2;
+		break;
+        case MC_SEQ_RD_CTL_D0 >> 2:
+		*out_reg = MC_SEQ_RD_CTL_D0_LP >> 2;
+		break;
+        case MC_SEQ_RD_CTL_D1 >> 2:
+		*out_reg = MC_SEQ_RD_CTL_D1_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_D0 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_D0_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_D1 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_D1_LP >> 2;
+		break;
+        case MC_PMG_CMD_EMRS >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_EMRS_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS1 >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS1_LP >> 2;
+		break;
+        case MC_SEQ_PMG_TIMING >> 2:
+		*out_reg = MC_SEQ_PMG_TIMING_LP >> 2;
+		break;
+        case MC_PMG_CMD_MRS2 >> 2:
+		*out_reg = MC_SEQ_PMG_CMD_MRS2_LP >> 2;
+		break;
+        case MC_SEQ_WR_CTL_2 >> 2:
+		*out_reg = MC_SEQ_WR_CTL_2_LP >> 2;
+		break;
+        default:
+		result = false;
+		break;
+	}
+
+	return result;
+}
+
+static void si_set_valid_flag(struct si_mc_reg_table *table)
+{
+	u8 i, j;
+
+	for (i = 0; i < table->last; i++) {
+		for (j = 1; j < table->num_entries; j++) {
+			if (table->mc_reg_table_entry[j-1].mc_data[i] != table->mc_reg_table_entry[j].mc_data[i]) {
+				table->valid_flag |= 1 << i;
+				break;
+			}
+		}
+	}
+}
+
+static void si_set_s0_mc_reg_index(struct si_mc_reg_table *table)
+{
+	u32 i;
+	u16 address;
+
+	for (i = 0; i < table->last; i++)
+		table->mc_reg_address[i].s0 = si_check_s0_mc_reg_index(table->mc_reg_address[i].s1, &address) ?
+			address : table->mc_reg_address[i].s1;
+
+}
+
+static int si_copy_vbios_mc_reg_table(struct atom_mc_reg_table *table,
+				      struct si_mc_reg_table *si_table)
+{
+	u8 i, j;
+
+	if (table->last > SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE)
+		return -EINVAL;
+	if (table->num_entries > MAX_AC_TIMING_ENTRIES)
+		return -EINVAL;
+
+	for (i = 0; i < table->last; i++)
+		si_table->mc_reg_address[i].s1 = table->mc_reg_address[i].s1;
+	si_table->last = table->last;
+
+	for (i = 0; i < table->num_entries; i++) {
+		si_table->mc_reg_table_entry[i].mclk_max =
+			table->mc_reg_table_entry[i].mclk_max;
+		for (j = 0; j < table->last; j++) {
+			si_table->mc_reg_table_entry[i].mc_data[j] =
+				table->mc_reg_table_entry[i].mc_data[j];
+		}
+	}
+	si_table->num_entries = table->num_entries;
+
+	return 0;
+}
+
+static int si_initialize_mc_reg_table(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct atom_mc_reg_table *table;
+	struct si_mc_reg_table *si_table = &si_pi->mc_reg_table;
+	u8 module_index = rv770_get_memory_module_index(rdev);
+	int ret;
+
+	table = kzalloc(sizeof(struct atom_mc_reg_table), GFP_KERNEL);
+	if (!table)
+		return -ENOMEM;
+
+	WREG32(MC_SEQ_RAS_TIMING_LP, RREG32(MC_SEQ_RAS_TIMING));
+	WREG32(MC_SEQ_CAS_TIMING_LP, RREG32(MC_SEQ_CAS_TIMING));
+	WREG32(MC_SEQ_MISC_TIMING_LP, RREG32(MC_SEQ_MISC_TIMING));
+	WREG32(MC_SEQ_MISC_TIMING2_LP, RREG32(MC_SEQ_MISC_TIMING2));
+	WREG32(MC_SEQ_PMG_CMD_EMRS_LP, RREG32(MC_PMG_CMD_EMRS));
+	WREG32(MC_SEQ_PMG_CMD_MRS_LP, RREG32(MC_PMG_CMD_MRS));
+	WREG32(MC_SEQ_PMG_CMD_MRS1_LP, RREG32(MC_PMG_CMD_MRS1));
+	WREG32(MC_SEQ_WR_CTL_D0_LP, RREG32(MC_SEQ_WR_CTL_D0));
+	WREG32(MC_SEQ_WR_CTL_D1_LP, RREG32(MC_SEQ_WR_CTL_D1));
+	WREG32(MC_SEQ_RD_CTL_D0_LP, RREG32(MC_SEQ_RD_CTL_D0));
+	WREG32(MC_SEQ_RD_CTL_D1_LP, RREG32(MC_SEQ_RD_CTL_D1));
+	WREG32(MC_SEQ_PMG_TIMING_LP, RREG32(MC_SEQ_PMG_TIMING));
+	WREG32(MC_SEQ_PMG_CMD_MRS2_LP, RREG32(MC_PMG_CMD_MRS2));
+	WREG32(MC_SEQ_WR_CTL_2_LP, RREG32(MC_SEQ_WR_CTL_2));
+
+        ret = radeon_atom_init_mc_reg_table(rdev, module_index, table);
+        if (ret)
+                goto init_mc_done;
+
+        ret = si_copy_vbios_mc_reg_table(table, si_table);
+        if (ret)
+                goto init_mc_done;
+
+	si_set_s0_mc_reg_index(si_table);
+
+	ret = si_set_mc_special_registers(rdev, si_table);
+        if (ret)
+                goto init_mc_done;
+
+	si_set_valid_flag(si_table);
+
+init_mc_done:
+	kfree(table);
+
+	return ret;
+
+}
+
+static void si_populate_mc_reg_addresses(struct radeon_device *rdev,
+					 SMC_SIslands_MCRegisters *mc_reg_table)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 i, j;
+
+	for (i = 0, j = 0; j < si_pi->mc_reg_table.last; j++) {
+		if (si_pi->mc_reg_table.valid_flag & (1 << j)) {
+			if (i >= SMC_NISLANDS_MC_REGISTER_ARRAY_SIZE)
+				break;
+			mc_reg_table->address[i].s0 =
+				cpu_to_be16(si_pi->mc_reg_table.mc_reg_address[j].s0);
+			mc_reg_table->address[i].s1 =
+				cpu_to_be16(si_pi->mc_reg_table.mc_reg_address[j].s1);
+			i++;
+		}
+	}
+	mc_reg_table->last = (u8)i;
+}
+
+static void si_convert_mc_registers(const struct si_mc_reg_entry *entry,
+				    SMC_SIslands_MCRegisterSet *data,
+				    u32 num_entries, u32 valid_flag)
+{
+	u32 i, j;
+
+	for(i = 0, j = 0; j < num_entries; j++) {
+		if (valid_flag & (1 << j)) {
+			data->value[i] = cpu_to_be32(entry->mc_data[j]);
+			i++;
+		}
+	}
+}
+
+static void si_convert_mc_reg_table_entry_to_smc(struct radeon_device *rdev,
+						 struct rv7xx_pl *pl,
+						 SMC_SIslands_MCRegisterSet *mc_reg_table_data)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 i = 0;
+
+	for (i = 0; i < si_pi->mc_reg_table.num_entries; i++) {
+		if (pl->mclk <= si_pi->mc_reg_table.mc_reg_table_entry[i].mclk_max)
+			break;
+	}
+
+	if ((i == si_pi->mc_reg_table.num_entries) && (i > 0))
+		--i;
+
+	si_convert_mc_registers(&si_pi->mc_reg_table.mc_reg_table_entry[i],
+				mc_reg_table_data, si_pi->mc_reg_table.last,
+				si_pi->mc_reg_table.valid_flag);
+}
+
+static void si_convert_mc_reg_table_to_smc(struct radeon_device *rdev,
+					   struct radeon_ps *radeon_state,
+					   SMC_SIslands_MCRegisters *mc_reg_table)
+{
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	int i;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		si_convert_mc_reg_table_entry_to_smc(rdev,
+						     &state->performance_levels[i],
+						     &mc_reg_table->data[SISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT + i]);
+	}
+}
+
+static int si_populate_mc_reg_table(struct radeon_device *rdev,
+				    struct radeon_ps *radeon_boot_state)
+{
+	struct ni_ps *boot_state = ni_get_ps(radeon_boot_state);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct si_ulv_param *ulv = &si_pi->ulv;
+	SMC_SIslands_MCRegisters *smc_mc_reg_table = &si_pi->smc_mc_reg_table;
+
+	memset(smc_mc_reg_table, 0, sizeof(SMC_SIslands_MCRegisters));
+
+	si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_seq_index, 1);
+
+	si_populate_mc_reg_addresses(rdev, smc_mc_reg_table);
+
+	si_convert_mc_reg_table_entry_to_smc(rdev, &boot_state->performance_levels[0],
+					     &smc_mc_reg_table->data[SISLANDS_MCREGISTERTABLE_INITIAL_SLOT]);
+
+	si_convert_mc_registers(&si_pi->mc_reg_table.mc_reg_table_entry[0],
+				&smc_mc_reg_table->data[SISLANDS_MCREGISTERTABLE_ACPI_SLOT],
+				si_pi->mc_reg_table.last,
+				si_pi->mc_reg_table.valid_flag);
+
+	if (ulv->supported && ulv->pl.vddc != 0)
+		si_convert_mc_reg_table_entry_to_smc(rdev, &ulv->pl,
+						     &smc_mc_reg_table->data[SISLANDS_MCREGISTERTABLE_ULV_SLOT]);
+	else
+		si_convert_mc_registers(&si_pi->mc_reg_table.mc_reg_table_entry[0],
+					&smc_mc_reg_table->data[SISLANDS_MCREGISTERTABLE_ULV_SLOT],
+					si_pi->mc_reg_table.last,
+					si_pi->mc_reg_table.valid_flag);
+
+	si_convert_mc_reg_table_to_smc(rdev, radeon_boot_state, smc_mc_reg_table);
+
+	return si_copy_bytes_to_smc(rdev, si_pi->mc_reg_table_start,
+				    (u8 *)smc_mc_reg_table,
+				    sizeof(SMC_SIslands_MCRegisters), si_pi->sram_end);
+}
+
+static int si_upload_mc_reg_table(struct radeon_device *rdev,
+				  struct radeon_ps *radeon_new_state)
+{
+	struct ni_ps *new_state = ni_get_ps(radeon_new_state);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	u32 address = si_pi->mc_reg_table_start +
+		offsetof(SMC_SIslands_MCRegisters,
+			 data[SISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT]);
+	SMC_SIslands_MCRegisters *smc_mc_reg_table = &si_pi->smc_mc_reg_table;
+
+	memset(smc_mc_reg_table, 0, sizeof(SMC_SIslands_MCRegisters));
+
+	si_convert_mc_reg_table_to_smc(rdev, radeon_new_state, smc_mc_reg_table);
+
+
+	return si_copy_bytes_to_smc(rdev, address,
+				    (u8 *)&smc_mc_reg_table->data[SISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT],
+				    sizeof(SMC_SIslands_MCRegisterSet) * new_state->performance_level_count,
+				    si_pi->sram_end);
+
+}
+
+static void si_enable_voltage_control(struct radeon_device *rdev, bool enable)
+{
+        if (enable)
+                WREG32_P(GENERAL_PWRMGT, VOLT_PWRMGT_EN, ~VOLT_PWRMGT_EN);
+        else
+                WREG32_P(GENERAL_PWRMGT, 0, ~VOLT_PWRMGT_EN);
+}
+
+static enum radeon_pcie_gen si_get_maximum_link_speed(struct radeon_device *rdev,
+						      struct radeon_ps *radeon_state)
+{
+	struct ni_ps *state = ni_get_ps(radeon_state);
+	int i;
+	u16 pcie_speed, max_speed = 0;
+
+	for (i = 0; i < state->performance_level_count; i++) {
+		pcie_speed = state->performance_levels[i].pcie_gen;
+		if (max_speed < pcie_speed)
+			max_speed = pcie_speed;
+	}
+	return max_speed;
+}
+
+static u16 si_get_current_pcie_speed(struct radeon_device *rdev)
+{
+	u32 speed_cntl;
+
+	speed_cntl = RREG32_PCIE_PORT(PCIE_LC_SPEED_CNTL) & LC_CURRENT_DATA_RATE_MASK;
+	speed_cntl >>= LC_CURRENT_DATA_RATE_SHIFT;
+
+	return (u16)speed_cntl;
+}
+
+static void si_request_link_speed_change_before_state_change(struct radeon_device *rdev,
+							     struct radeon_ps *radeon_new_state,
+							     struct radeon_ps *radeon_current_state)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	enum radeon_pcie_gen target_link_speed = si_get_maximum_link_speed(rdev, radeon_new_state);
+	enum radeon_pcie_gen current_link_speed;
+
+	if (si_pi->force_pcie_gen == RADEON_PCIE_GEN_INVALID)
+		current_link_speed = si_get_maximum_link_speed(rdev, radeon_current_state);
+	else
+		current_link_speed = si_pi->force_pcie_gen;
+
+	si_pi->force_pcie_gen = RADEON_PCIE_GEN_INVALID;
+	si_pi->pspp_notify_required = false;
+	if (target_link_speed > current_link_speed) {
+		switch (target_link_speed) {
+#if defined(CONFIG_ACPI)
+		case RADEON_PCIE_GEN3:
+			if (radeon_acpi_pcie_performance_request(rdev, PCIE_PERF_REQ_PECI_GEN3, false) == 0)
+				break;
+			si_pi->force_pcie_gen = RADEON_PCIE_GEN2;
+			if (current_link_speed == RADEON_PCIE_GEN2)
+				break;
+		case RADEON_PCIE_GEN2:
+			if (radeon_acpi_pcie_performance_request(rdev, PCIE_PERF_REQ_PECI_GEN2, false) == 0)
+				break;
+#endif
+		default:
+			si_pi->force_pcie_gen = si_get_current_pcie_speed(rdev);
+			break;
+		}
+	} else {
+		if (target_link_speed < current_link_speed)
+			si_pi->pspp_notify_required = true;
+	}
+}
+
+static void si_notify_link_speed_change_after_state_change(struct radeon_device *rdev,
+							   struct radeon_ps *radeon_new_state,
+							   struct radeon_ps *radeon_current_state)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	enum radeon_pcie_gen target_link_speed = si_get_maximum_link_speed(rdev, radeon_new_state);
+	u8 request;
+
+	if (si_pi->pspp_notify_required) {
+		if (target_link_speed == RADEON_PCIE_GEN3)
+			request = PCIE_PERF_REQ_PECI_GEN3;
+		else if (target_link_speed == RADEON_PCIE_GEN2)
+			request = PCIE_PERF_REQ_PECI_GEN2;
+		else
+			request = PCIE_PERF_REQ_PECI_GEN1;
+
+		if ((request == PCIE_PERF_REQ_PECI_GEN1) &&
+		    (si_get_current_pcie_speed(rdev) > 0))
+			return;
+
+#if defined(CONFIG_ACPI)
+		radeon_acpi_pcie_performance_request(rdev, request, false);
+#endif
+	}
+}
+
+#if 0
+static int si_ds_request(struct radeon_device *rdev,
+			 bool ds_status_on, u32 count_write)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+
+	if (eg_pi->sclk_deep_sleep) {
+		if (ds_status_on)
+			return (si_send_msg_to_smc(rdev, PPSMC_MSG_CancelThrottleOVRDSCLKDS) ==
+				PPSMC_Result_OK) ?
+				0 : -EINVAL;
+		else
+			return (si_send_msg_to_smc(rdev, PPSMC_MSG_ThrottleOVRDSCLKDS) ==
+				PPSMC_Result_OK) ? 0 : -EINVAL;
+	}
+	return 0;
+}
+#endif
+
+static void si_set_max_cu_value(struct radeon_device *rdev)
+{
+	struct si_power_info *si_pi = si_get_pi(rdev);
+
+	if (rdev->family == CHIP_VERDE) {
+		switch (rdev->pdev->device) {
+		case 0x6820:
+		case 0x6825:
+		case 0x6821:
+		case 0x6823:
+		case 0x6827:
+			si_pi->max_cu = 10;
+			break;
+		case 0x682D:
+		case 0x6824:
+		case 0x682F:
+		case 0x6826:
+			si_pi->max_cu = 8;
+			break;
+		case 0x6828:
+		case 0x6830:
+		case 0x6831:
+		case 0x6838:
+		case 0x6839:
+		case 0x683D:
+			si_pi->max_cu = 10;
+			break;
+		case 0x683B:
+		case 0x683F:
+		case 0x6829:
+			si_pi->max_cu = 8;
+			break;
+		default:
+			si_pi->max_cu = 0;
+			break;
+		}
+	} else {
+		si_pi->max_cu = 0;
+	}
+}
+
+static int si_patch_single_dependency_table_based_on_leakage(struct radeon_device *rdev,
+							     struct radeon_clock_voltage_dependency_table *table)
+{
+	u32 i;
+	int j;
+	u16 leakage_voltage;
+
+	if (table) {
+		for (i = 0; i < table->count; i++) {
+			switch (si_get_leakage_voltage_from_leakage_index(rdev,
+									  table->entries[i].v,
+									  &leakage_voltage)) {
+			case 0:
+				table->entries[i].v = leakage_voltage;
+				break;
+			case -EAGAIN:
+				return -EINVAL;
+			case -EINVAL:
+			default:
+				break;
+			}
+		}
+
+		for (j = (table->count - 2); j >= 0; j--) {
+			table->entries[j].v = (table->entries[j].v <= table->entries[j + 1].v) ?
+				table->entries[j].v : table->entries[j + 1].v;
+		}
+	}
+	return 0;
+}
+
+static int si_patch_dependency_tables_based_on_leakage(struct radeon_device *rdev)
+{
+	int ret = 0;
+
+	ret = si_patch_single_dependency_table_based_on_leakage(rdev,
+								&rdev->pm.dpm.dyn_state.vddc_dependency_on_sclk);
+	ret = si_patch_single_dependency_table_based_on_leakage(rdev,
+								&rdev->pm.dpm.dyn_state.vddc_dependency_on_mclk);
+	ret = si_patch_single_dependency_table_based_on_leakage(rdev,
+								&rdev->pm.dpm.dyn_state.vddci_dependency_on_mclk);
+	return ret;
+}
+
+static void si_set_pcie_lane_width_in_smc(struct radeon_device *rdev,
+					  struct radeon_ps *radeon_new_state,
+					  struct radeon_ps *radeon_current_state)
+{
+	u32 lane_width;
+	u32 new_lane_width =
+		(radeon_new_state->caps & ATOM_PPLIB_PCIE_LINK_WIDTH_MASK) >> ATOM_PPLIB_PCIE_LINK_WIDTH_SHIFT;
+	u32 current_lane_width =
+		(radeon_current_state->caps & ATOM_PPLIB_PCIE_LINK_WIDTH_MASK) >> ATOM_PPLIB_PCIE_LINK_WIDTH_SHIFT;
+
+	if (new_lane_width != current_lane_width) {
+		radeon_set_pcie_lanes(rdev, new_lane_width);
+		lane_width = radeon_get_pcie_lanes(rdev);
+		si_write_smc_soft_register(rdev, SI_SMC_SOFT_REGISTER_non_ulv_pcie_link_width, lane_width);
+	}
+}
+
+void si_dpm_setup_asic(struct radeon_device *rdev)
+{
+	rv770_get_memory_type(rdev);
+	si_read_clock_registers(rdev);
+	si_enable_acpi_power_management(rdev);
+}
+
+static int si_set_thermal_temperature_range(struct radeon_device *rdev,
+					int min_temp, int max_temp)
+{
+	int low_temp = 0 * 1000;
+	int high_temp = 255 * 1000;
+
+	if (low_temp < min_temp)
+		low_temp = min_temp;
+	if (high_temp > max_temp)
+		high_temp = max_temp;
+	if (high_temp < low_temp) {
+		DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp);
+		return -EINVAL;
+	}
+
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTH(high_temp / 1000), ~DIG_THERM_INTH_MASK);
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTL(low_temp / 1000), ~DIG_THERM_INTL_MASK);
+	WREG32_P(CG_THERMAL_CTRL, DIG_THERM_DPM(high_temp / 1000), ~DIG_THERM_DPM_MASK);
+
+	rdev->pm.dpm.thermal.min_temp = low_temp;
+	rdev->pm.dpm.thermal.max_temp = high_temp;
+
+	return 0;
+}
+
+int si_dpm_enable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+	int ret;
+
+	if (si_is_smc_running(rdev))
+		return -EINVAL;
+	if (pi->voltage_control)
+		si_enable_voltage_control(rdev, true);
+	if (pi->mvdd_control)
+		si_get_mvdd_configuration(rdev);
+	if (pi->voltage_control) {
+		ret = si_construct_voltage_tables(rdev);
+		if (ret) {
+			DRM_ERROR("si_construct_voltage_tables failed\n");
+			return ret;
+		}
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = si_initialize_mc_reg_table(rdev);
+		if (ret)
+			eg_pi->dynamic_ac_timing = false;
+	}
+	if (pi->dynamic_ss)
+		si_enable_spread_spectrum(rdev, true);
+	if (pi->thermal_protection)
+		si_enable_thermal_protection(rdev, true);
+	si_setup_bsp(rdev);
+	si_program_git(rdev);
+	si_program_tp(rdev);
+	si_program_tpp(rdev);
+	si_program_sstp(rdev);
+	si_enable_display_gap(rdev);
+	si_program_vc(rdev);
+	ret = si_upload_firmware(rdev);
+	if (ret) {
+		DRM_ERROR("si_upload_firmware failed\n");
+		return ret;
+	}
+	ret = si_process_firmware_header(rdev);
+	if (ret) {
+		DRM_ERROR("si_process_firmware_header failed\n");
+		return ret;
+	}
+	ret = si_initial_switch_from_arb_f0_to_f1(rdev);
+	if (ret) {
+		DRM_ERROR("si_initial_switch_from_arb_f0_to_f1 failed\n");
+		return ret;
+	}
+	ret = si_init_smc_table(rdev);
+	if (ret) {
+		DRM_ERROR("si_init_smc_table failed\n");
+		return ret;
+	}
+	ret = si_init_smc_spll_table(rdev);
+	if (ret) {
+		DRM_ERROR("si_init_smc_spll_table failed\n");
+		return ret;
+	}
+	ret = si_init_arb_table_index(rdev);
+	if (ret) {
+		DRM_ERROR("si_init_arb_table_index failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = si_populate_mc_reg_table(rdev, boot_ps);
+		if (ret) {
+			DRM_ERROR("si_populate_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+	ret = si_initialize_smc_cac_tables(rdev);
+	if (ret) {
+		DRM_ERROR("si_initialize_smc_cac_tables failed\n");
+		return ret;
+	}
+	ret = si_initialize_hardware_cac_manager(rdev);
+	if (ret) {
+		DRM_ERROR("si_initialize_hardware_cac_manager failed\n");
+		return ret;
+	}
+	ret = si_initialize_smc_dte_tables(rdev);
+	if (ret) {
+		DRM_ERROR("si_initialize_smc_dte_tables failed\n");
+		return ret;
+	}
+	ret = si_populate_smc_tdp_limits(rdev, boot_ps);
+	if (ret) {
+		DRM_ERROR("si_populate_smc_tdp_limits failed\n");
+		return ret;
+	}
+	ret = si_populate_smc_tdp_limits_2(rdev, boot_ps);
+	if (ret) {
+		DRM_ERROR("si_populate_smc_tdp_limits_2 failed\n");
+		return ret;
+	}
+	si_program_response_times(rdev);
+	si_program_ds_registers(rdev);
+	si_dpm_start_smc(rdev);
+	ret = si_notify_smc_display_change(rdev, false);
+	if (ret) {
+		DRM_ERROR("si_notify_smc_display_change failed\n");
+		return ret;
+	}
+	si_enable_sclk_control(rdev, true);
+	si_start_dpm(rdev);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		PPSMC_Result result;
+
+		ret = si_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+		result = si_send_msg_to_smc(rdev, PPSMC_MSG_EnableThermalInterrupt);
+
+		if (result != PPSMC_Result_OK)
+			DRM_DEBUG_KMS("Could not enable thermal interrupts.\n");
+	}
+
+	si_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, true);
+
+	ni_update_current_ps(rdev, boot_ps);
+
+	return 0;
+}
+
+void si_dpm_disable(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct radeon_ps *boot_ps = rdev->pm.dpm.boot_ps;
+
+	if (!si_is_smc_running(rdev))
+		return;
+	si_disable_ulv(rdev);
+	si_clear_vc(rdev);
+	if (pi->thermal_protection)
+		si_enable_thermal_protection(rdev, false);
+	si_enable_power_containment(rdev, boot_ps, false);
+	si_enable_smc_cac(rdev, boot_ps, false);
+	si_enable_spread_spectrum(rdev, false);
+	si_enable_auto_throttle_source(rdev, RADEON_DPM_AUTO_THROTTLE_SRC_THERMAL, false);
+	si_stop_dpm(rdev);
+	si_reset_to_default(rdev);
+	si_dpm_stop_smc(rdev);
+	si_force_switch_to_arb_f0(rdev);
+
+	ni_update_current_ps(rdev, boot_ps);
+}
+
+int si_dpm_pre_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps requested_ps = *rdev->pm.dpm.requested_ps;
+	struct radeon_ps *new_ps = &requested_ps;
+
+	ni_update_requested_ps(rdev, new_ps);
+
+	si_apply_state_adjust_rules(rdev, &eg_pi->requested_rps);
+
+	return 0;
+}
+
+static int si_power_control_set_level(struct radeon_device *rdev)
+{
+	struct radeon_ps *new_ps = rdev->pm.dpm.requested_ps;
+	int ret;
+
+	ret = si_restrict_performance_levels_before_switch(rdev);
+	if (ret)
+		return ret;
+	ret = si_halt_smc(rdev);
+	if (ret)
+		return ret;
+	ret = si_populate_smc_tdp_limits(rdev, new_ps);
+	if (ret)
+		return ret;
+	ret = si_populate_smc_tdp_limits_2(rdev, new_ps);
+	if (ret)
+		return ret;
+	ret = si_resume_smc(rdev);
+	if (ret)
+		return ret;
+	ret = si_set_sw_state(rdev);
+	if (ret)
+		return ret;
+	return 0;
+}
+
+int si_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = &eg_pi->requested_rps;
+	struct radeon_ps *old_ps = &eg_pi->current_rps;
+	int ret;
+
+	ret = si_disable_ulv(rdev);
+	if (ret) {
+		DRM_ERROR("si_disable_ulv failed\n");
+		return ret;
+	}
+	ret = si_restrict_performance_levels_before_switch(rdev);
+	if (ret) {
+		DRM_ERROR("si_restrict_performance_levels_before_switch failed\n");
+		return ret;
+	}
+	if (eg_pi->pcie_performance_request)
+		si_request_link_speed_change_before_state_change(rdev, new_ps, old_ps);
+	ni_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+	ret = si_enable_power_containment(rdev, new_ps, false);
+	if (ret) {
+		DRM_ERROR("si_enable_power_containment failed\n");
+		return ret;
+	}
+	ret = si_enable_smc_cac(rdev, new_ps, false);
+	if (ret) {
+		DRM_ERROR("si_enable_smc_cac failed\n");
+		return ret;
+	}
+	ret = si_halt_smc(rdev);
+	if (ret) {
+		DRM_ERROR("si_halt_smc failed\n");
+		return ret;
+	}
+	ret = si_upload_sw_state(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("si_upload_sw_state failed\n");
+		return ret;
+	}
+	ret = si_upload_smc_data(rdev);
+	if (ret) {
+		DRM_ERROR("si_upload_smc_data failed\n");
+		return ret;
+	}
+	ret = si_upload_ulv_state(rdev);
+	if (ret) {
+		DRM_ERROR("si_upload_ulv_state failed\n");
+		return ret;
+	}
+	if (eg_pi->dynamic_ac_timing) {
+		ret = si_upload_mc_reg_table(rdev, new_ps);
+		if (ret) {
+			DRM_ERROR("si_upload_mc_reg_table failed\n");
+			return ret;
+		}
+	}
+	ret = si_program_memory_timing_parameters(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("si_program_memory_timing_parameters failed\n");
+		return ret;
+	}
+	si_set_pcie_lane_width_in_smc(rdev, new_ps, old_ps);
+
+	ret = si_resume_smc(rdev);
+	if (ret) {
+		DRM_ERROR("si_resume_smc failed\n");
+		return ret;
+	}
+	ret = si_set_sw_state(rdev);
+	if (ret) {
+		DRM_ERROR("si_set_sw_state failed\n");
+		return ret;
+	}
+	ni_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+	if (eg_pi->pcie_performance_request)
+		si_notify_link_speed_change_after_state_change(rdev, new_ps, old_ps);
+	ret = si_set_power_state_conditionally_enable_ulv(rdev, new_ps);
+	if (ret) {
+		DRM_ERROR("si_set_power_state_conditionally_enable_ulv failed\n");
+		return ret;
+	}
+	ret = si_enable_smc_cac(rdev, new_ps, true);
+	if (ret) {
+		DRM_ERROR("si_enable_smc_cac failed\n");
+		return ret;
+	}
+	ret = si_enable_power_containment(rdev, new_ps, true);
+	if (ret) {
+		DRM_ERROR("si_enable_power_containment failed\n");
+		return ret;
+	}
+
+	ret = si_power_control_set_level(rdev);
+	if (ret) {
+		DRM_ERROR("si_power_control_set_level failed\n");
+		return ret;
+	}
+
+#if 0
+	/* XXX */
+	ret = si_dpm_force_performance_level(rdev, RADEON_DPM_FORCED_LEVEL_AUTO);
+	if (ret) {
+		DRM_ERROR("si_dpm_force_performance_level failed\n");
+		return ret;
+	}
+#else
+	rdev->pm.dpm.forced_level = RADEON_DPM_FORCED_LEVEL_AUTO;
+#endif
+
+	return 0;
+}
+
+void si_dpm_post_set_power_state(struct radeon_device *rdev)
+{
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct radeon_ps *new_ps = &eg_pi->requested_rps;
+
+	ni_update_current_ps(rdev, new_ps);
+}
+
+
+void si_dpm_reset_asic(struct radeon_device *rdev)
+{
+	si_restrict_performance_levels_before_switch(rdev);
+	si_disable_ulv(rdev);
+	si_set_boot_state(rdev);
+}
+
+void si_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+	si_program_display_gap(rdev);
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+	struct _ATOM_PPLIB_SI_CLOCK_INFO si;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void si_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					  struct radeon_ps *rps,
+					  struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info,
+					  u8 table_rev)
+{
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) {
+		rps->vclk = le32_to_cpu(non_clock_info->ulVCLK);
+		rps->dclk = le32_to_cpu(non_clock_info->ulDCLK);
+	} else if (r600_is_uvd_state(rps->class, rps->class2)) {
+		rps->vclk = RV770_DEFAULT_VCLK_FREQ;
+		rps->dclk = RV770_DEFAULT_DCLK_FREQ;
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT)
+		rdev->pm.dpm.boot_ps = rps;
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void si_parse_pplib_clock_info(struct radeon_device *rdev,
+				      struct radeon_ps *rps, int index,
+				      union pplib_clock_info *clock_info)
+{
+	struct rv7xx_power_info *pi = rv770_get_pi(rdev);
+	struct evergreen_power_info *eg_pi = evergreen_get_pi(rdev);
+	struct si_power_info *si_pi = si_get_pi(rdev);
+	struct ni_ps *ps = ni_get_ps(rps);
+	u16 leakage_voltage;
+	struct rv7xx_pl *pl = &ps->performance_levels[index];
+	int ret;
+
+	ps->performance_level_count = index + 1;
+
+	pl->sclk = le16_to_cpu(clock_info->si.usEngineClockLow);
+	pl->sclk |= clock_info->si.ucEngineClockHigh << 16;
+	pl->mclk = le16_to_cpu(clock_info->si.usMemoryClockLow);
+	pl->mclk |= clock_info->si.ucMemoryClockHigh << 16;
+
+	pl->vddc = le16_to_cpu(clock_info->si.usVDDC);
+	pl->vddci = le16_to_cpu(clock_info->si.usVDDCI);
+	pl->flags = le32_to_cpu(clock_info->si.ulFlags);
+	pl->pcie_gen = r600_get_pcie_gen_support(rdev,
+						 si_pi->sys_pcie_mask,
+						 si_pi->boot_pcie_gen,
+						 clock_info->si.ucPCIEGen);
+
+	/* patch up vddc if necessary */
+	ret = si_get_leakage_voltage_from_leakage_index(rdev, pl->vddc,
+							&leakage_voltage);
+	if (ret == 0)
+		pl->vddc = leakage_voltage;
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_ACPI) {
+		pi->acpi_vddc = pl->vddc;
+		eg_pi->acpi_vddci = pl->vddci;
+		si_pi->acpi_pcie_gen = pl->pcie_gen;
+	}
+
+	if ((rps->class2 & ATOM_PPLIB_CLASSIFICATION2_ULV) &&
+	    index == 0) {
+		/* XXX disable for A0 tahiti */
+		si_pi->ulv.supported = true;
+		si_pi->ulv.pl = *pl;
+		si_pi->ulv.one_pcie_lane_in_ulv = false;
+		si_pi->ulv.volt_change_delay = SISLANDS_ULVVOLTAGECHANGEDELAY_DFLT;
+		si_pi->ulv.cg_ulv_parameter = SISLANDS_CGULVPARAMETER_DFLT;
+		si_pi->ulv.cg_ulv_control = SISLANDS_CGULVCONTROL_DFLT;
+	}
+
+	if (pi->min_vddc_in_table > pl->vddc)
+		pi->min_vddc_in_table = pl->vddc;
+
+	if (pi->max_vddc_in_table < pl->vddc)
+		pi->max_vddc_in_table = pl->vddc;
+
+	/* patch up boot state */
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		u16 vddc, vddci, mvdd;
+		radeon_atombios_get_default_voltages(rdev, &vddc, &vddci, &mvdd);
+		pl->mclk = rdev->clock.default_mclk;
+		pl->sclk = rdev->clock.default_sclk;
+		pl->vddc = vddc;
+		pl->vddci = vddci;
+		si_pi->mvdd_bootup_value = mvdd;
+	}
+
+	if ((rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) ==
+	    ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE) {
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.sclk = pl->sclk;
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.mclk = pl->mclk;
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddc = pl->vddc;
+		rdev->pm.dpm.dyn_state.max_clock_voltage_on_ac.vddci = pl->vddci;
+	}
+}
+
+static int si_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i, j, k, non_clock_array_index, clock_array_index;
+	union pplib_clock_info *clock_info;
+	struct _StateArray *state_array;
+	struct _ClockInfoArray *clock_info_array;
+	struct _NonClockInfoArray *non_clock_info_array;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	u8 *power_state_offset;
+	struct ni_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	state_array = (struct _StateArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usStateArrayOffset));
+	clock_info_array = (struct _ClockInfoArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset));
+	non_clock_info_array = (struct _NonClockInfoArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset));
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  state_array->ucNumEntries, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	power_state_offset = (u8 *)state_array->states;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+	for (i = 0; i < state_array->ucNumEntries; i++) {
+		power_state = (union pplib_power_state *)power_state_offset;
+		non_clock_array_index = power_state->v2.nonClockInfoIndex;
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			&non_clock_info_array->nonClockInfo[non_clock_array_index];
+		if (!rdev->pm.power_state[i].clock_info)
+			return -EINVAL;
+		ps = kzalloc(sizeof(struct ni_ps), GFP_KERNEL);
+		if (ps == NULL) {
+			kfree(rdev->pm.dpm.ps);
+			return -ENOMEM;
+		}
+		rdev->pm.dpm.ps[i].ps_priv = ps;
+		si_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+					      non_clock_info,
+					      non_clock_info_array->ucEntrySize);
+		k = 0;
+		for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) {
+			clock_array_index = power_state->v2.clockInfoIndex[j];
+			if (clock_array_index >= clock_info_array->ucNumEntries)
+				continue;
+			if (k >= SISLANDS_MAX_HARDWARE_POWERLEVELS)
+				break;
+			clock_info = (union pplib_clock_info *)
+				&clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize];
+			si_parse_pplib_clock_info(rdev,
+						  &rdev->pm.dpm.ps[i], k,
+						  clock_info);
+			k++;
+		}
+		power_state_offset += 2 + power_state->v2.ucNumDPMLevels;
+	}
+	rdev->pm.dpm.num_ps = state_array->ucNumEntries;
+	return 0;
+}
+
+int si_dpm_init(struct radeon_device *rdev)
+{
+	struct rv7xx_power_info *pi;
+	struct evergreen_power_info *eg_pi;
+	struct ni_power_info *ni_pi;
+	struct si_power_info *si_pi;
+	int index = GetIndexIntoMasterTable(DATA, ASIC_InternalSS_Info);
+	u16 data_offset, size;
+	u8 frev, crev;
+	struct atom_clock_dividers dividers;
+	int ret;
+	u32 mask;
+
+	si_pi = kzalloc(sizeof(struct si_power_info), GFP_KERNEL);
+	if (si_pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = si_pi;
+	ni_pi = &si_pi->ni;
+	eg_pi = &ni_pi->eg;
+	pi = &eg_pi->rv7xx;
+
+	ret = drm_pcie_get_speed_cap_mask(rdev->ddev, &mask);
+	if (ret)
+		si_pi->sys_pcie_mask = 0;
+	else
+		si_pi->sys_pcie_mask = mask;
+	si_pi->force_pcie_gen = RADEON_PCIE_GEN_INVALID;
+	si_pi->boot_pcie_gen = si_get_current_pcie_speed(rdev);
+
+	si_set_max_cu_value(rdev);
+
+	rv770_get_max_vddc(rdev);
+	si_get_leakage_vddc(rdev);
+	si_patch_dependency_tables_based_on_leakage(rdev);
+
+	pi->acpi_vddc = 0;
+	eg_pi->acpi_vddci = 0;
+	pi->min_vddc_in_table = 0;
+	pi->max_vddc_in_table = 0;
+
+	ret = si_parse_power_table(rdev);
+	if (ret)
+		return ret;
+	ret = r600_parse_extended_power_table(rdev);
+	if (ret)
+		return ret;
+
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries =
+		kzalloc(4 * sizeof(struct radeon_clock_voltage_dependency_entry), GFP_KERNEL);
+	if (!rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries) {
+		r600_free_extended_power_table(rdev);
+		return -ENOMEM;
+	}
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.count = 4;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].clk = 0;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[0].v = 0;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].clk = 36000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[1].v = 720;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].clk = 54000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[2].v = 810;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].clk = 72000;
+	rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries[3].v = 900;
+
+	if (rdev->pm.dpm.voltage_response_time == 0)
+		rdev->pm.dpm.voltage_response_time = R600_VOLTAGERESPONSETIME_DFLT;
+	if (rdev->pm.dpm.backbias_response_time == 0)
+		rdev->pm.dpm.backbias_response_time = R600_BACKBIASRESPONSETIME_DFLT;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     0, false, &dividers);
+	if (ret)
+		pi->ref_div = dividers.ref_div + 1;
+	else
+		pi->ref_div = R600_REFERENCEDIVIDER_DFLT;
+
+	eg_pi->smu_uvd_hs = false;
+
+	pi->mclk_strobe_mode_threshold = 40000;
+	if (si_is_special_1gb_platform(rdev))
+		pi->mclk_stutter_mode_threshold = 0;
+	else
+		pi->mclk_stutter_mode_threshold = pi->mclk_strobe_mode_threshold;
+	pi->mclk_edc_enable_threshold = 40000;
+	eg_pi->mclk_edc_wr_enable_threshold = 40000;
+
+	ni_pi->mclk_rtt_mode_threshold = eg_pi->mclk_edc_wr_enable_threshold;
+
+	pi->voltage_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, VOLTAGE_OBJ_GPIO_LUT);
+
+	pi->mvdd_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_MVDDC, VOLTAGE_OBJ_GPIO_LUT);
+
+	eg_pi->vddci_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDCI, VOLTAGE_OBJ_GPIO_LUT);
+
+	si_pi->vddc_phase_shed_control =
+		radeon_atom_is_voltage_gpio(rdev, SET_VOLTAGE_TYPE_ASIC_VDDC, VOLTAGE_OBJ_PHASE_LUT);
+
+	if (atom_parse_data_header(rdev->mode_info.atom_context, index, &size,
+                                   &frev, &crev, &data_offset)) {
+		pi->sclk_ss = true;
+		pi->mclk_ss = true;
+		pi->dynamic_ss = true;
+	} else {
+		pi->sclk_ss = false;
+		pi->mclk_ss = false;
+		pi->dynamic_ss = true;
+	}
+
+	pi->asi = RV770_ASI_DFLT;
+	pi->pasi = CYPRESS_HASI_DFLT;
+	pi->vrc = SISLANDS_VRC_DFLT;
+
+	pi->gfx_clock_gating = true;
+
+	eg_pi->sclk_deep_sleep = true;
+	si_pi->sclk_deep_sleep_above_low = false;
+
+	if (pi->gfx_clock_gating &&
+	    (rdev->pm.int_thermal_type != THERMAL_TYPE_NONE))
+		pi->thermal_protection = true;
+	else
+		pi->thermal_protection = false;
+
+	eg_pi->dynamic_ac_timing = true;
+
+	eg_pi->light_sleep = true;
+#if defined(CONFIG_ACPI)
+	eg_pi->pcie_performance_request =
+		radeon_acpi_is_pcie_performance_request_supported(rdev);
+#else
+	eg_pi->pcie_performance_request = false;
+#endif
+
+	si_pi->sram_end = SMC_RAM_END;
+
+	rdev->pm.dpm.dyn_state.mclk_sclk_ratio = 4;
+	rdev->pm.dpm.dyn_state.sclk_mclk_delta = 15000;
+	rdev->pm.dpm.dyn_state.vddc_vddci_delta = 200;
+	rdev->pm.dpm.dyn_state.valid_sclk_values.count = 0;
+	rdev->pm.dpm.dyn_state.valid_sclk_values.values = NULL;
+	rdev->pm.dpm.dyn_state.valid_mclk_values.count = 0;
+	rdev->pm.dpm.dyn_state.valid_mclk_values.values = NULL;
+
+	si_initialize_powertune_defaults(rdev);
+
+	return 0;
+}
+
+void si_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+	kfree(rdev->pm.dpm.dyn_state.vddc_dependency_on_dispclk.entries);
+	r600_free_extended_power_table(rdev);
+}
+
+void si_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						    struct seq_file *m)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct ni_ps *ps = ni_get_ps(rps);
+	struct rv7xx_pl *pl;
+	u32 current_index =
+		(RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_STATE_INDEX_MASK) >>
+		CURRENT_STATE_INDEX_SHIFT;
+
+	if (current_index >= ps->performance_level_count) {
+		seq_printf(m, "invalid dpm profile %d\n", current_index);
+	} else {
+		pl = &ps->performance_levels[current_index];
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		seq_printf(m, "power level %d    sclk: %u mclk: %u vddc: %u vddci: %u pcie gen: %u\n",
+			   current_index, pl->sclk, pl->mclk, pl->vddc, pl->vddci, pl->pcie_gen + 1);
+	}
+}
diff --git a/drivers/gpu/drm/radeon/si_dpm.h b/drivers/gpu/drm/radeon/si_dpm.h
new file mode 100644
index 000000000000..4ce5032cdf49
--- /dev/null
+++ b/drivers/gpu/drm/radeon/si_dpm.h
@@ -0,0 +1,227 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __SI_DPM_H__
+#define __SI_DPM_H__
+
+#include "ni_dpm.h"
+#include "sislands_smc.h"
+
+enum si_cac_config_reg_type
+{
+	SISLANDS_CACCONFIG_MMR = 0,
+	SISLANDS_CACCONFIG_CGIND,
+	SISLANDS_CACCONFIG_MAX
+};
+
+struct si_cac_config_reg
+{
+	u32 offset;
+	u32 mask;
+	u32 shift;
+	u32 value;
+	enum si_cac_config_reg_type type;
+};
+
+struct si_powertune_data
+{
+	u32 cac_window;
+	u32 l2_lta_window_size_default;
+	u8 lts_truncate_default;
+	u8 shift_n_default;
+	u8 operating_temp;
+	struct ni_leakage_coeffients leakage_coefficients;
+	u32 fixed_kt;
+	u32 lkge_lut_v0_percent;
+	u8 dc_cac[NISLANDS_DCCAC_MAX_LEVELS];
+	bool enable_powertune_by_default;
+};
+
+struct si_dyn_powertune_data
+{
+	u32 cac_leakage;
+	s32 leakage_minimum_temperature;
+	u32 wintime;
+	u32 l2_lta_window_size;
+	u8 lts_truncate;
+	u8 shift_n;
+	u8 dc_pwr_value;
+	bool disable_uvd_powertune;
+};
+
+struct si_dte_data
+{
+	u32 tau[SMC_SISLANDS_DTE_MAX_FILTER_STAGES];
+	u32 r[SMC_SISLANDS_DTE_MAX_FILTER_STAGES];
+	u32 k;
+	u32 t0;
+	u32 max_t;
+	u8 window_size;
+	u8 temp_select;
+	u8 dte_mode;
+	u8 tdep_count;
+	u8 t_limits[SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE];
+	u32 tdep_tau[SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE];
+	u32 tdep_r[SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE];
+	u32 t_threshold;
+	bool enable_dte_by_default;
+};
+
+struct si_clock_registers {
+	u32 cg_spll_func_cntl;
+	u32 cg_spll_func_cntl_2;
+	u32 cg_spll_func_cntl_3;
+	u32 cg_spll_func_cntl_4;
+	u32 cg_spll_spread_spectrum;
+	u32 cg_spll_spread_spectrum_2;
+	u32 dll_cntl;
+	u32 mclk_pwrmgt_cntl;
+	u32 mpll_ad_func_cntl;
+	u32 mpll_dq_func_cntl;
+	u32 mpll_func_cntl;
+	u32 mpll_func_cntl_1;
+	u32 mpll_func_cntl_2;
+	u32 mpll_ss1;
+	u32 mpll_ss2;
+};
+
+struct si_mc_reg_entry {
+	u32 mclk_max;
+	u32 mc_data[SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE];
+};
+
+struct si_mc_reg_table {
+	u8 last;
+	u8 num_entries;
+	u16 valid_flag;
+	struct si_mc_reg_entry mc_reg_table_entry[MAX_AC_TIMING_ENTRIES];
+	SMC_NIslands_MCRegisterAddress mc_reg_address[SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE];
+};
+
+#define SISLANDS_MCREGISTERTABLE_INITIAL_SLOT               0
+#define SISLANDS_MCREGISTERTABLE_ACPI_SLOT                  1
+#define SISLANDS_MCREGISTERTABLE_ULV_SLOT                   2
+#define SISLANDS_MCREGISTERTABLE_FIRST_DRIVERSTATE_SLOT     3
+
+struct si_leakage_voltage_entry
+{
+	u16 voltage;
+	u16 leakage_index;
+};
+
+#define SISLANDS_LEAKAGE_INDEX0     0xff01
+#define SISLANDS_MAX_LEAKAGE_COUNT  4
+
+struct si_leakage_voltage
+{
+	u16 count;
+	struct si_leakage_voltage_entry entries[SISLANDS_MAX_LEAKAGE_COUNT];
+};
+
+#define SISLANDS_MAX_HARDWARE_POWERLEVELS 5
+
+struct si_ulv_param {
+	bool supported;
+	u32 cg_ulv_control;
+	u32 cg_ulv_parameter;
+	u32 volt_change_delay;
+	struct rv7xx_pl pl;
+	bool one_pcie_lane_in_ulv;
+};
+
+struct si_power_info {
+	/* must be first! */
+	struct ni_power_info ni;
+	struct si_clock_registers clock_registers;
+	struct si_mc_reg_table mc_reg_table;
+	struct atom_voltage_table mvdd_voltage_table;
+	struct atom_voltage_table vddc_phase_shed_table;
+	struct si_leakage_voltage leakage_voltage;
+	u16 mvdd_bootup_value;
+	struct si_ulv_param ulv;
+	u32 max_cu;
+	/* pcie gen */
+	enum radeon_pcie_gen force_pcie_gen;
+	enum radeon_pcie_gen boot_pcie_gen;
+	enum radeon_pcie_gen acpi_pcie_gen;
+	u32 sys_pcie_mask;
+	/* flags */
+	bool enable_dte;
+	bool enable_ppm;
+	bool vddc_phase_shed_control;
+	bool pspp_notify_required;
+	bool sclk_deep_sleep_above_low;
+	/* smc offsets */
+	u32 sram_end;
+	u32 state_table_start;
+	u32 soft_regs_start;
+	u32 mc_reg_table_start;
+	u32 arb_table_start;
+	u32 cac_table_start;
+	u32 dte_table_start;
+	u32 spll_table_start;
+	u32 papm_cfg_table_start;
+	/* CAC stuff */
+	const struct si_cac_config_reg *cac_weights;
+	const struct si_cac_config_reg *lcac_config;
+	const struct si_cac_config_reg *cac_override;
+	const struct si_powertune_data *powertune_data;
+	struct si_dyn_powertune_data dyn_powertune_data;
+	/* DTE stuff */
+	struct si_dte_data dte_data;
+	/* scratch structs */
+	SMC_SIslands_MCRegisters smc_mc_reg_table;
+	SISLANDS_SMC_STATETABLE smc_statetable;
+	PP_SIslands_PAPMParameters papm_parm;
+};
+
+#define SISLANDS_INITIAL_STATE_ARB_INDEX    0
+#define SISLANDS_ACPI_STATE_ARB_INDEX       1
+#define SISLANDS_ULV_STATE_ARB_INDEX        2
+#define SISLANDS_DRIVER_STATE_ARB_INDEX     3
+
+#define SISLANDS_DPM2_MAX_PULSE_SKIP        256
+
+#define SISLANDS_DPM2_NEAR_TDP_DEC          10
+#define SISLANDS_DPM2_ABOVE_SAFE_INC        5
+#define SISLANDS_DPM2_BELOW_SAFE_INC        20
+
+#define SISLANDS_DPM2_TDP_SAFE_LIMIT_PERCENT            80
+
+#define SISLANDS_DPM2_MAXPS_PERCENT_H                   99
+#define SISLANDS_DPM2_MAXPS_PERCENT_M                   99
+
+#define SISLANDS_DPM2_SQ_RAMP_MAX_POWER                 0x3FFF
+#define SISLANDS_DPM2_SQ_RAMP_MIN_POWER                 0x12
+#define SISLANDS_DPM2_SQ_RAMP_MAX_POWER_DELTA           0x15
+#define SISLANDS_DPM2_SQ_RAMP_STI_SIZE                  0x1E
+#define SISLANDS_DPM2_SQ_RAMP_LTI_RATIO                 0xF
+
+#define SISLANDS_DPM2_PWREFFICIENCYRATIO_MARGIN         10
+
+#define SISLANDS_VRC_DFLT                               0xC000B3
+#define SISLANDS_ULVVOLTAGECHANGEDELAY_DFLT             1687
+#define SISLANDS_CGULVPARAMETER_DFLT                    0x00040035
+#define SISLANDS_CGULVCONTROL_DFLT                      0x1f007550
+
+
+#endif
diff --git a/drivers/gpu/drm/radeon/si_smc.c b/drivers/gpu/drm/radeon/si_smc.c
new file mode 100644
index 000000000000..5f524c0a541e
--- /dev/null
+++ b/drivers/gpu/drm/radeon/si_smc.c
@@ -0,0 +1,284 @@
+/*
+ * Copyright 2011 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+
+#include <linux/firmware.h>
+#include "drmP.h"
+#include "radeon.h"
+#include "sid.h"
+#include "ppsmc.h"
+#include "radeon_ucode.h"
+
+int si_set_smc_sram_address(struct radeon_device *rdev,
+			    u32 smc_address, u32 limit)
+{
+	if (smc_address & 3)
+		return -EINVAL;
+	if ((smc_address + 3) > limit)
+		return -EINVAL;
+
+	WREG32(SMC_IND_INDEX_0, smc_address);
+	WREG32_P(SMC_IND_ACCESS_CNTL, 0, ~AUTO_INCREMENT_IND_0);
+
+	return 0;
+}
+
+int si_copy_bytes_to_smc(struct radeon_device *rdev,
+			 u32 smc_start_address,
+			 const u8 *src, u32 byte_count, u32 limit)
+{
+	int ret;
+	u32 data, original_data, addr, extra_shift;
+
+	if (smc_start_address & 3)
+		return -EINVAL;
+	if ((smc_start_address + byte_count) > limit)
+		return -EINVAL;
+
+	addr = smc_start_address;
+
+	while (byte_count >= 4) {
+		/* SMC address space is BE */
+		data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3];
+
+		ret = si_set_smc_sram_address(rdev, addr, limit);
+		if (ret)
+			return ret;
+
+		WREG32(SMC_IND_DATA_0, data);
+
+		src += 4;
+		byte_count -= 4;
+		addr += 4;
+	}
+
+	/* RMW for the final bytes */
+	if (byte_count > 0) {
+		data = 0;
+
+		ret = si_set_smc_sram_address(rdev, addr, limit);
+		if (ret)
+			return ret;
+
+		original_data = RREG32(SMC_IND_DATA_0);
+
+		extra_shift = 8 * (4 - byte_count);
+
+		while (byte_count > 0) {
+			/* SMC address space is BE */
+			data = (data << 8) + *src++;
+			byte_count--;
+		}
+
+		data <<= extra_shift;
+
+		data |= (original_data & ~((~0UL) << extra_shift));
+
+		ret = si_set_smc_sram_address(rdev, addr, limit);
+		if (ret)
+			return ret;
+
+		WREG32(SMC_IND_DATA_0, data);
+	}
+	return 0;
+}
+
+void si_start_smc(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32_SMC(SMC_SYSCON_RESET_CNTL);
+
+	tmp &= ~RST_REG;
+
+	WREG32_SMC(SMC_SYSCON_RESET_CNTL, tmp);
+}
+
+void si_reset_smc(struct radeon_device *rdev)
+{
+	u32 tmp;
+
+	RREG32(CB_CGTT_SCLK_CTRL);
+	RREG32(CB_CGTT_SCLK_CTRL);
+	RREG32(CB_CGTT_SCLK_CTRL);
+	RREG32(CB_CGTT_SCLK_CTRL);
+
+	tmp = RREG32_SMC(SMC_SYSCON_RESET_CNTL);
+	tmp |= RST_REG;
+	WREG32_SMC(SMC_SYSCON_RESET_CNTL, tmp);
+}
+
+int si_program_jump_on_start(struct radeon_device *rdev)
+{
+	static u8 data[] = { 0x0E, 0x00, 0x40, 0x40 };
+
+	return si_copy_bytes_to_smc(rdev, 0x0, data, 4, sizeof(data)+1);
+}
+
+void si_stop_smc_clock(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0);
+
+	tmp |= CK_DISABLE;
+
+	WREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0, tmp);
+}
+
+void si_start_smc_clock(struct radeon_device *rdev)
+{
+	u32 tmp = RREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0);
+
+	tmp &= ~CK_DISABLE;
+
+	WREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0, tmp);
+}
+
+bool si_is_smc_running(struct radeon_device *rdev)
+{
+	u32 rst = RREG32_SMC(SMC_SYSCON_RESET_CNTL);
+	u32 clk = RREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0);
+
+	if (!(rst & RST_REG) && !(clk & CK_DISABLE))
+		return true;
+
+	return false;
+}
+
+PPSMC_Result si_send_msg_to_smc(struct radeon_device *rdev, PPSMC_Msg msg)
+{
+	u32 tmp;
+	int i;
+
+	if (!si_is_smc_running(rdev))
+		return PPSMC_Result_Failed;
+
+	WREG32(SMC_MESSAGE_0, msg);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32(SMC_RESP_0);
+		if (tmp != 0)
+			break;
+		udelay(1);
+	}
+	tmp = RREG32(SMC_RESP_0);
+
+	return (PPSMC_Result)tmp;
+}
+
+PPSMC_Result si_wait_for_smc_inactive(struct radeon_device *rdev)
+{
+	u32 tmp;
+	int i;
+
+	if (!si_is_smc_running(rdev))
+		return PPSMC_Result_OK;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		tmp = RREG32_SMC(SMC_SYSCON_CLOCK_CNTL_0);
+		if ((tmp & CKEN) == 0)
+			break;
+		udelay(1);
+	}
+
+	return PPSMC_Result_OK;
+}
+
+int si_load_smc_ucode(struct radeon_device *rdev, u32 limit)
+{
+	u32 ucode_start_address;
+	u32 ucode_size;
+	const u8 *src;
+	u32 data;
+
+	if (!rdev->smc_fw)
+		return -EINVAL;
+
+	switch (rdev->family) {
+	case CHIP_TAHITI:
+		ucode_start_address = TAHITI_SMC_UCODE_START;
+		ucode_size = TAHITI_SMC_UCODE_SIZE;
+		break;
+	case CHIP_PITCAIRN:
+		ucode_start_address = PITCAIRN_SMC_UCODE_START;
+		ucode_size = PITCAIRN_SMC_UCODE_SIZE;
+		break;
+	case CHIP_VERDE:
+		ucode_start_address = VERDE_SMC_UCODE_START;
+		ucode_size = VERDE_SMC_UCODE_SIZE;
+		break;
+	case CHIP_OLAND:
+		ucode_start_address = OLAND_SMC_UCODE_START;
+		ucode_size = OLAND_SMC_UCODE_SIZE;
+		break;
+	case CHIP_HAINAN:
+		ucode_start_address = HAINAN_SMC_UCODE_START;
+		ucode_size = HAINAN_SMC_UCODE_SIZE;
+		break;
+	default:
+		DRM_ERROR("unknown asic in smc ucode loader\n");
+		BUG();
+	}
+
+	if (ucode_size & 3)
+		return -EINVAL;
+
+	src = (const u8 *)rdev->smc_fw->data;
+	WREG32(SMC_IND_INDEX_0, ucode_start_address);
+	WREG32_P(SMC_IND_ACCESS_CNTL, AUTO_INCREMENT_IND_0, ~AUTO_INCREMENT_IND_0);
+	while (ucode_size >= 4) {
+		/* SMC address space is BE */
+		data = (src[0] << 24) | (src[1] << 16) | (src[2] << 8) | src[3];
+
+		WREG32(SMC_IND_DATA_0, data);
+
+		src += 4;
+		ucode_size -= 4;
+	}
+	WREG32_P(SMC_IND_ACCESS_CNTL, 0, ~AUTO_INCREMENT_IND_0);
+
+	return 0;
+}
+
+int si_read_smc_sram_dword(struct radeon_device *rdev, u32 smc_address,
+			   u32 *value, u32 limit)
+{
+	int ret;
+
+	ret = si_set_smc_sram_address(rdev, smc_address, limit);
+	if (ret)
+		return ret;
+
+	*value = RREG32(SMC_IND_DATA_0);
+	return 0;
+}
+
+int si_write_smc_sram_dword(struct radeon_device *rdev, u32 smc_address,
+			    u32 value, u32 limit)
+{
+	int ret;
+
+	ret = si_set_smc_sram_address(rdev, smc_address, limit);
+	if (ret)
+		return ret;
+
+	WREG32(SMC_IND_DATA_0, value);
+	return 0;
+}
diff --git a/drivers/gpu/drm/radeon/sid.h b/drivers/gpu/drm/radeon/sid.h
index 8f2d7d4f9b28..2c8da27a929f 100644
--- a/drivers/gpu/drm/radeon/sid.h
+++ b/drivers/gpu/drm/radeon/sid.h
@@ -30,6 +30,94 @@
 #define VERDE_GB_ADDR_CONFIG_GOLDEN         0x12010002
 #define HAINAN_GB_ADDR_CONFIG_GOLDEN        0x02010001
 
+#define SI_MAX_SH_GPRS           256
+#define SI_MAX_TEMP_GPRS         16
+#define SI_MAX_SH_THREADS        256
+#define SI_MAX_SH_STACK_ENTRIES  4096
+#define SI_MAX_FRC_EOV_CNT       16384
+#define SI_MAX_BACKENDS          8
+#define SI_MAX_BACKENDS_MASK     0xFF
+#define SI_MAX_BACKENDS_PER_SE_MASK     0x0F
+#define SI_MAX_SIMDS             12
+#define SI_MAX_SIMDS_MASK        0x0FFF
+#define SI_MAX_SIMDS_PER_SE_MASK        0x00FF
+#define SI_MAX_PIPES             8
+#define SI_MAX_PIPES_MASK        0xFF
+#define SI_MAX_PIPES_PER_SIMD_MASK      0x3F
+#define SI_MAX_LDS_NUM           0xFFFF
+#define SI_MAX_TCC               16
+#define SI_MAX_TCC_MASK          0xFFFF
+
+/* SMC IND accessor regs */
+#define SMC_IND_INDEX_0                              0x200
+#define SMC_IND_DATA_0                               0x204
+
+#define SMC_IND_ACCESS_CNTL                          0x228
+#       define AUTO_INCREMENT_IND_0                  (1 << 0)
+#define SMC_MESSAGE_0                                0x22c
+#define SMC_RESP_0                                   0x230
+
+/* CG IND registers are accessed via SMC indirect space + SMC_CG_IND_START */
+#define SMC_CG_IND_START                    0xc0030000
+#define SMC_CG_IND_END                      0xc0040000
+
+#define	CG_CGTT_LOCAL_0				0x400
+#define	CG_CGTT_LOCAL_1				0x401
+
+/* SMC IND registers */
+#define	SMC_SYSCON_RESET_CNTL				0x80000000
+#       define RST_REG                                  (1 << 0)
+#define	SMC_SYSCON_CLOCK_CNTL_0				0x80000004
+#       define CK_DISABLE                               (1 << 0)
+#       define CKEN                                     (1 << 24)
+
+#define VGA_HDP_CONTROL  				0x328
+#define		VGA_MEMORY_DISABLE				(1 << 4)
+
+#define DCCG_DISP_SLOW_SELECT_REG                       0x4fc
+#define		DCCG_DISP1_SLOW_SELECT(x)		((x) << 0)
+#define		DCCG_DISP1_SLOW_SELECT_MASK		(7 << 0)
+#define		DCCG_DISP1_SLOW_SELECT_SHIFT		0
+#define		DCCG_DISP2_SLOW_SELECT(x)		((x) << 4)
+#define		DCCG_DISP2_SLOW_SELECT_MASK		(7 << 4)
+#define		DCCG_DISP2_SLOW_SELECT_SHIFT		4
+
+#define	CG_SPLL_FUNC_CNTL				0x600
+#define		SPLL_RESET				(1 << 0)
+#define		SPLL_SLEEP				(1 << 1)
+#define		SPLL_BYPASS_EN				(1 << 3)
+#define		SPLL_REF_DIV(x)				((x) << 4)
+#define		SPLL_REF_DIV_MASK			(0x3f << 4)
+#define		SPLL_PDIV_A(x)				((x) << 20)
+#define		SPLL_PDIV_A_MASK			(0x7f << 20)
+#define		SPLL_PDIV_A_SHIFT			20
+#define	CG_SPLL_FUNC_CNTL_2				0x604
+#define		SCLK_MUX_SEL(x)				((x) << 0)
+#define		SCLK_MUX_SEL_MASK			(0x1ff << 0)
+#define	CG_SPLL_FUNC_CNTL_3				0x608
+#define		SPLL_FB_DIV(x)				((x) << 0)
+#define		SPLL_FB_DIV_MASK			(0x3ffffff << 0)
+#define		SPLL_FB_DIV_SHIFT			0
+#define		SPLL_DITHEN				(1 << 28)
+#define	CG_SPLL_FUNC_CNTL_4				0x60c
+
+#define	SPLL_CNTL_MODE					0x618
+#	define SPLL_REFCLK_SEL(x)			((x) << 8)
+#	define SPLL_REFCLK_SEL_MASK			0xFF00
+
+#define	CG_SPLL_SPREAD_SPECTRUM				0x620
+#define		SSEN					(1 << 0)
+#define		CLK_S(x)				((x) << 4)
+#define		CLK_S_MASK				(0xfff << 4)
+#define		CLK_S_SHIFT				4
+#define	CG_SPLL_SPREAD_SPECTRUM_2			0x624
+#define		CLK_V(x)				((x) << 0)
+#define		CLK_V_MASK				(0x3ffffff << 0)
+#define		CLK_V_SHIFT				0
+
+#define	CG_SPLL_AUTOSCALE_CNTL				0x62c
+#       define AUTOSCALE_ON_SS_CLEAR                    (1 << 9)
+
 /* discrete uvd clocks */
 #define	CG_UPLL_FUNC_CNTL				0x634
 #	define UPLL_RESET_MASK				0x00000001
@@ -59,6 +147,45 @@
 #define	CG_UPLL_SPREAD_SPECTRUM				0x650
 #	define SSEN_MASK				0x00000001
 
+#define	MPLL_BYPASSCLK_SEL				0x65c
+#	define MPLL_CLKOUT_SEL(x)			((x) << 8)
+#	define MPLL_CLKOUT_SEL_MASK			0xFF00
+
+#define CG_CLKPIN_CNTL                                    0x660
+#       define XTALIN_DIVIDE                              (1 << 1)
+#       define BCLK_AS_XCLK                               (1 << 2)
+#define CG_CLKPIN_CNTL_2                                  0x664
+#       define FORCE_BIF_REFCLK_EN                        (1 << 3)
+#       define MUX_TCLK_TO_XCLK                           (1 << 8)
+
+#define	THM_CLK_CNTL					0x66c
+#	define CMON_CLK_SEL(x)				((x) << 0)
+#	define CMON_CLK_SEL_MASK			0xFF
+#	define TMON_CLK_SEL(x)				((x) << 8)
+#	define TMON_CLK_SEL_MASK			0xFF00
+#define	MISC_CLK_CNTL					0x670
+#	define DEEP_SLEEP_CLK_SEL(x)			((x) << 0)
+#	define DEEP_SLEEP_CLK_SEL_MASK			0xFF
+#	define ZCLK_SEL(x)				((x) << 8)
+#	define ZCLK_SEL_MASK				0xFF00
+
+#define	CG_THERMAL_CTRL					0x700
+#define 	DPM_EVENT_SRC(x)			((x) << 0)
+#define 	DPM_EVENT_SRC_MASK			(7 << 0)
+#define		DIG_THERM_DPM(x)			((x) << 14)
+#define		DIG_THERM_DPM_MASK			0x003FC000
+#define		DIG_THERM_DPM_SHIFT			14
+
+#define	CG_THERMAL_INT					0x708
+#define		DIG_THERM_INTH(x)			((x) << 8)
+#define		DIG_THERM_INTH_MASK			0x0000FF00
+#define		DIG_THERM_INTH_SHIFT			8
+#define		DIG_THERM_INTL(x)			((x) << 16)
+#define		DIG_THERM_INTL_MASK			0x00FF0000
+#define		DIG_THERM_INTL_SHIFT			16
+#define 	THERM_INT_MASK_HIGH			(1 << 24)
+#define 	THERM_INT_MASK_LOW			(1 << 25)
+
 #define	CG_MULT_THERMAL_STATUS					0x714
 #define		ASIC_MAX_TEMP(x)				((x) << 0)
 #define		ASIC_MAX_TEMP_MASK				0x000001ff
@@ -67,31 +194,89 @@
 #define		CTF_TEMP_MASK					0x0003fe00
 #define		CTF_TEMP_SHIFT					9
 
-#define SI_MAX_SH_GPRS           256
-#define SI_MAX_TEMP_GPRS         16
-#define SI_MAX_SH_THREADS        256
-#define SI_MAX_SH_STACK_ENTRIES  4096
-#define SI_MAX_FRC_EOV_CNT       16384
-#define SI_MAX_BACKENDS          8
-#define SI_MAX_BACKENDS_MASK     0xFF
-#define SI_MAX_BACKENDS_PER_SE_MASK     0x0F
-#define SI_MAX_SIMDS             12
-#define SI_MAX_SIMDS_MASK        0x0FFF
-#define SI_MAX_SIMDS_PER_SE_MASK        0x00FF
-#define SI_MAX_PIPES             8
-#define SI_MAX_PIPES_MASK        0xFF
-#define SI_MAX_PIPES_PER_SIMD_MASK      0x3F
-#define SI_MAX_LDS_NUM           0xFFFF
-#define SI_MAX_TCC               16
-#define SI_MAX_TCC_MASK          0xFFFF
-
-#define VGA_HDP_CONTROL  				0x328
-#define		VGA_MEMORY_DISABLE				(1 << 4)
-
-#define CG_CLKPIN_CNTL                                    0x660
-#       define XTALIN_DIVIDE                              (1 << 1)
-#define CG_CLKPIN_CNTL_2                                  0x664
-#       define MUX_TCLK_TO_XCLK                           (1 << 8)
+#define GENERAL_PWRMGT                                  0x780
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+#       define STATIC_PM_EN                             (1 << 1)
+#       define THERMAL_PROTECTION_DIS                   (1 << 2)
+#       define THERMAL_PROTECTION_TYPE                  (1 << 3)
+#       define SW_SMIO_INDEX(x)                         ((x) << 6)
+#       define SW_SMIO_INDEX_MASK                       (1 << 6)
+#       define SW_SMIO_INDEX_SHIFT                      6
+#       define VOLT_PWRMGT_EN                           (1 << 10)
+#       define DYN_SPREAD_SPECTRUM_EN                   (1 << 23)
+#define CG_TPC                                            0x784
+#define SCLK_PWRMGT_CNTL                                  0x788
+#       define SCLK_PWRMGT_OFF                            (1 << 0)
+#       define SCLK_LOW_D1                                (1 << 1)
+#       define FIR_RESET                                  (1 << 4)
+#       define FIR_FORCE_TREND_SEL                        (1 << 5)
+#       define FIR_TREND_MODE                             (1 << 6)
+#       define DYN_GFX_CLK_OFF_EN                         (1 << 7)
+#       define GFX_CLK_FORCE_ON                           (1 << 8)
+#       define GFX_CLK_REQUEST_OFF                        (1 << 9)
+#       define GFX_CLK_FORCE_OFF                          (1 << 10)
+#       define GFX_CLK_OFF_ACPI_D1                        (1 << 11)
+#       define GFX_CLK_OFF_ACPI_D2                        (1 << 12)
+#       define GFX_CLK_OFF_ACPI_D3                        (1 << 13)
+#       define DYN_LIGHT_SLEEP_EN                         (1 << 14)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                  0x798
+#       define CURRENT_STATE_INDEX_MASK                   (0xf << 4)
+#       define CURRENT_STATE_INDEX_SHIFT                  4
+
+#define CG_FTV                                            0x7bc
+
+#define CG_FFCT_0                                         0x7c0
+#       define UTC_0(x)                                   ((x) << 0)
+#       define UTC_0_MASK                                 (0x3ff << 0)
+#       define DTC_0(x)                                   ((x) << 10)
+#       define DTC_0_MASK                                 (0x3ff << 10)
+
+#define CG_BSP                                          0x7fc
+#       define BSP(x)					((x) << 0)
+#       define BSP_MASK					(0xffff << 0)
+#       define BSU(x)					((x) << 16)
+#       define BSU_MASK					(0xf << 16)
+#define CG_AT                                           0x800
+#       define CG_R(x)					((x) << 0)
+#       define CG_R_MASK				(0xffff << 0)
+#       define CG_L(x)					((x) << 16)
+#       define CG_L_MASK				(0xffff << 16)
+
+#define CG_GIT                                          0x804
+#       define CG_GICST(x)                              ((x) << 0)
+#       define CG_GICST_MASK                            (0xffff << 0)
+#       define CG_GIPOT(x)                              ((x) << 16)
+#       define CG_GIPOT_MASK                            (0xffff << 16)
+
+#define CG_SSP                                            0x80c
+#       define SST(x)                                     ((x) << 0)
+#       define SST_MASK                                   (0xffff << 0)
+#       define SSTU(x)                                    ((x) << 16)
+#       define SSTU_MASK                                  (0xf << 16)
+
+#define CG_DISPLAY_GAP_CNTL                               0x828
+#       define DISP1_GAP(x)                               ((x) << 0)
+#       define DISP1_GAP_MASK                             (3 << 0)
+#       define DISP2_GAP(x)                               ((x) << 2)
+#       define DISP2_GAP_MASK                             (3 << 2)
+#       define VBI_TIMER_COUNT(x)                         ((x) << 4)
+#       define VBI_TIMER_COUNT_MASK                       (0x3fff << 4)
+#       define VBI_TIMER_UNIT(x)                          ((x) << 20)
+#       define VBI_TIMER_UNIT_MASK                        (7 << 20)
+#       define DISP1_GAP_MCHG(x)                          ((x) << 24)
+#       define DISP1_GAP_MCHG_MASK                        (3 << 24)
+#       define DISP2_GAP_MCHG(x)                          ((x) << 26)
+#       define DISP2_GAP_MCHG_MASK                        (3 << 26)
+
+#define	CG_ULV_CONTROL					0x878
+#define	CG_ULV_PARAMETER				0x87c
+
+#define	SMC_SCRATCH0					0x884
+
+#define	CG_CAC_CTRL					0x8b8
+#	define CAC_WINDOW(x)				((x) << 0)
+#	define CAC_WINDOW_MASK				0x00ffffff
 
 #define DMIF_ADDR_CONFIG  				0xBD4
 
@@ -182,6 +367,20 @@
 
 #define	VM_CONTEXT1_PROTECTION_FAULT_ADDR		0x14FC
 #define	VM_CONTEXT1_PROTECTION_FAULT_STATUS		0x14DC
+#define		PROTECTIONS_MASK			(0xf << 0)
+#define		PROTECTIONS_SHIFT			0
+		/* bit 0: range
+		 * bit 1: pde0
+		 * bit 2: valid
+		 * bit 3: read
+		 * bit 4: write
+		 */
+#define		MEMORY_CLIENT_ID_MASK			(0xff << 12)
+#define		MEMORY_CLIENT_ID_SHIFT			12
+#define		MEMORY_CLIENT_RW_MASK			(1 << 24)
+#define		MEMORY_CLIENT_RW_SHIFT			24
+#define		FAULT_VMID_MASK				(0xf << 25)
+#define		FAULT_VMID_SHIFT			25
 
 #define VM_INVALIDATE_REQUEST				0x1478
 #define VM_INVALIDATE_RESPONSE				0x147c
@@ -203,6 +402,10 @@
 #define	VM_CONTEXT0_PAGE_TABLE_END_ADDR			0x157C
 #define	VM_CONTEXT1_PAGE_TABLE_END_ADDR			0x1580
 
+#define VM_L2_CG           				0x15c0
+#define		MC_CG_ENABLE				(1 << 18)
+#define		MC_LS_ENABLE				(1 << 19)
+
 #define MC_SHARED_CHMAP						0x2004
 #define		NOOFCHAN_SHIFT					12
 #define		NOOFCHAN_MASK					0x0000f000
@@ -228,6 +431,17 @@
 
 #define MC_SHARED_BLACKOUT_CNTL           		0x20ac
 
+#define MC_HUB_MISC_HUB_CG           			0x20b8
+#define MC_HUB_MISC_VM_CG           			0x20bc
+
+#define MC_HUB_MISC_SIP_CG           			0x20c0
+
+#define MC_XPB_CLK_GAT           			0x2478
+
+#define MC_CITF_MISC_RD_CG           			0x2648
+#define MC_CITF_MISC_WR_CG           			0x264c
+#define MC_CITF_MISC_VM_CG           			0x2650
+
 #define	MC_ARB_RAMCFG					0x2760
 #define		NOOFBANK_SHIFT					0
 #define		NOOFBANK_MASK					0x00000003
@@ -243,6 +457,23 @@
 #define		NOOFGROUPS_SHIFT				12
 #define		NOOFGROUPS_MASK					0x00001000
 
+#define	MC_ARB_DRAM_TIMING				0x2774
+#define	MC_ARB_DRAM_TIMING2				0x2778
+
+#define MC_ARB_BURST_TIME                               0x2808
+#define		STATE0(x)				((x) << 0)
+#define		STATE0_MASK				(0x1f << 0)
+#define		STATE0_SHIFT				0
+#define		STATE1(x)				((x) << 5)
+#define		STATE1_MASK				(0x1f << 5)
+#define		STATE1_SHIFT				5
+#define		STATE2(x)				((x) << 10)
+#define		STATE2_MASK				(0x1f << 10)
+#define		STATE2_SHIFT				10
+#define		STATE3(x)				((x) << 15)
+#define		STATE3_MASK				(0x1f << 15)
+#define		STATE3_SHIFT				15
+
 #define	MC_SEQ_TRAIN_WAKEUP_CNTL			0x2808
 #define		TRAIN_DONE_D0      			(1 << 30)
 #define		TRAIN_DONE_D1      			(1 << 31)
@@ -250,13 +481,105 @@
 #define MC_SEQ_SUP_CNTL           			0x28c8
 #define		RUN_MASK      				(1 << 0)
 #define MC_SEQ_SUP_PGM           			0x28cc
+#define MC_PMG_AUTO_CMD           			0x28d0
 
 #define MC_IO_PAD_CNTL_D0           			0x29d0
 #define		MEM_FALL_OUT_CMD      			(1 << 8)
 
+#define MC_SEQ_RAS_TIMING                               0x28a0
+#define MC_SEQ_CAS_TIMING                               0x28a4
+#define MC_SEQ_MISC_TIMING                              0x28a8
+#define MC_SEQ_MISC_TIMING2                             0x28ac
+#define MC_SEQ_PMG_TIMING                               0x28b0
+#define MC_SEQ_RD_CTL_D0                                0x28b4
+#define MC_SEQ_RD_CTL_D1                                0x28b8
+#define MC_SEQ_WR_CTL_D0                                0x28bc
+#define MC_SEQ_WR_CTL_D1                                0x28c0
+
+#define MC_SEQ_MISC0           				0x2a00
+#define 	MC_SEQ_MISC0_VEN_ID_SHIFT               8
+#define 	MC_SEQ_MISC0_VEN_ID_MASK                0x00000f00
+#define 	MC_SEQ_MISC0_VEN_ID_VALUE               3
+#define 	MC_SEQ_MISC0_REV_ID_SHIFT               12
+#define 	MC_SEQ_MISC0_REV_ID_MASK                0x0000f000
+#define 	MC_SEQ_MISC0_REV_ID_VALUE               1
+#define 	MC_SEQ_MISC0_GDDR5_SHIFT                28
+#define 	MC_SEQ_MISC0_GDDR5_MASK                 0xf0000000
+#define 	MC_SEQ_MISC0_GDDR5_VALUE                5
+#define MC_SEQ_MISC1                                    0x2a04
+#define MC_SEQ_RESERVE_M                                0x2a08
+#define MC_PMG_CMD_EMRS                                 0x2a0c
+
 #define MC_SEQ_IO_DEBUG_INDEX           		0x2a44
 #define MC_SEQ_IO_DEBUG_DATA           			0x2a48
 
+#define MC_SEQ_MISC5                                    0x2a54
+#define MC_SEQ_MISC6                                    0x2a58
+
+#define MC_SEQ_MISC7                                    0x2a64
+
+#define MC_SEQ_RAS_TIMING_LP                            0x2a6c
+#define MC_SEQ_CAS_TIMING_LP                            0x2a70
+#define MC_SEQ_MISC_TIMING_LP                           0x2a74
+#define MC_SEQ_MISC_TIMING2_LP                          0x2a78
+#define MC_SEQ_WR_CTL_D0_LP                             0x2a7c
+#define MC_SEQ_WR_CTL_D1_LP                             0x2a80
+#define MC_SEQ_PMG_CMD_EMRS_LP                          0x2a84
+#define MC_SEQ_PMG_CMD_MRS_LP                           0x2a88
+
+#define MC_PMG_CMD_MRS                                  0x2aac
+
+#define MC_SEQ_RD_CTL_D0_LP                             0x2b1c
+#define MC_SEQ_RD_CTL_D1_LP                             0x2b20
+
+#define MC_PMG_CMD_MRS1                                 0x2b44
+#define MC_SEQ_PMG_CMD_MRS1_LP                          0x2b48
+#define MC_SEQ_PMG_TIMING_LP                            0x2b4c
+
+#define MC_SEQ_WR_CTL_2                                 0x2b54
+#define MC_SEQ_WR_CTL_2_LP                              0x2b58
+#define MC_PMG_CMD_MRS2                                 0x2b5c
+#define MC_SEQ_PMG_CMD_MRS2_LP                          0x2b60
+
+#define	MCLK_PWRMGT_CNTL				0x2ba0
+#       define DLL_SPEED(x)				((x) << 0)
+#       define DLL_SPEED_MASK				(0x1f << 0)
+#       define DLL_READY                                (1 << 6)
+#       define MC_INT_CNTL                              (1 << 7)
+#       define MRDCK0_PDNB                              (1 << 8)
+#       define MRDCK1_PDNB                              (1 << 9)
+#       define MRDCK0_RESET                             (1 << 16)
+#       define MRDCK1_RESET                             (1 << 17)
+#       define DLL_READY_READ                           (1 << 24)
+#define	DLL_CNTL					0x2ba4
+#       define MRDCK0_BYPASS                            (1 << 24)
+#       define MRDCK1_BYPASS                            (1 << 25)
+
+#define	MPLL_FUNC_CNTL					0x2bb4
+#define		BWCTRL(x)				((x) << 20)
+#define		BWCTRL_MASK				(0xff << 20)
+#define	MPLL_FUNC_CNTL_1				0x2bb8
+#define		VCO_MODE(x)				((x) << 0)
+#define		VCO_MODE_MASK				(3 << 0)
+#define		CLKFRAC(x)				((x) << 4)
+#define		CLKFRAC_MASK				(0xfff << 4)
+#define		CLKF(x)					((x) << 16)
+#define		CLKF_MASK				(0xfff << 16)
+#define	MPLL_FUNC_CNTL_2				0x2bbc
+#define	MPLL_AD_FUNC_CNTL				0x2bc0
+#define		YCLK_POST_DIV(x)			((x) << 0)
+#define		YCLK_POST_DIV_MASK			(7 << 0)
+#define	MPLL_DQ_FUNC_CNTL				0x2bc4
+#define		YCLK_SEL(x)				((x) << 4)
+#define		YCLK_SEL_MASK				(1 << 4)
+
+#define	MPLL_SS1					0x2bcc
+#define		CLKV(x)					((x) << 0)
+#define		CLKV_MASK				(0x3ffffff << 0)
+#define	MPLL_SS2					0x2bd0
+#define		CLKS(x)					((x) << 0)
+#define		CLKS_MASK				(0xfff << 0)
+
 #define	HDP_HOST_PATH_CNTL				0x2C00
 #define	HDP_NONSURFACE_BASE				0x2C04
 #define	HDP_NONSURFACE_INFO				0x2C08
@@ -266,6 +589,8 @@
 #define HDP_MISC_CNTL					0x2F4C
 #define 	HDP_FLUSH_INVALIDATE_CACHE			(1 << 0)
 
+#define ATC_MISC_CG           				0x3350
+
 #define IH_RB_CNTL                                        0x3e00
 #       define IH_RB_ENABLE                               (1 << 0)
 #       define IH_IB_SIZE(x)                              ((x) << 1) /* log2 */
@@ -424,6 +749,9 @@
 #       define DC_HPDx_RX_INT_TIMER(x)                    ((x) << 16)
 #       define DC_HPDx_EN                                 (1 << 28)
 
+#define DPG_PIPE_STUTTER_CONTROL                          0x6cd4
+#       define STUTTER_ENABLE                             (1 << 0)
+
 /* 0x6e98, 0x7a98, 0x10698, 0x11298, 0x11e98, 0x12a98 */
 #define CRTC_STATUS_FRAME_COUNT                         0x6e98
 
@@ -599,6 +927,24 @@
 
 #define	SQC_CACHES					0x8C08
 
+#define SQ_POWER_THROTTLE                               0x8e58
+#define		MIN_POWER(x)				((x) << 0)
+#define		MIN_POWER_MASK				(0x3fff << 0)
+#define		MIN_POWER_SHIFT				0
+#define		MAX_POWER(x)				((x) << 16)
+#define		MAX_POWER_MASK				(0x3fff << 16)
+#define		MAX_POWER_SHIFT				0
+#define SQ_POWER_THROTTLE2                              0x8e5c
+#define		MAX_POWER_DELTA(x)			((x) << 0)
+#define		MAX_POWER_DELTA_MASK			(0x3fff << 0)
+#define		MAX_POWER_DELTA_SHIFT			0
+#define		STI_SIZE(x)				((x) << 16)
+#define		STI_SIZE_MASK				(0x3ff << 16)
+#define		STI_SIZE_SHIFT				16
+#define		LTI_RATIO(x)				((x) << 27)
+#define		LTI_RATIO_MASK				(0xf << 27)
+#define		LTI_RATIO_SHIFT				27
+
 #define	SX_DEBUG_1					0x9060
 
 #define	SPI_STATIC_THREAD_MGMT_1			0x90E0
@@ -616,6 +962,11 @@
 #define	CGTS_USER_TCC_DISABLE				0x914C
 #define		TCC_DISABLE_MASK				0xFFFF0000
 #define		TCC_DISABLE_SHIFT				16
+#define	CGTS_SM_CTRL_REG				0x9150
+#define		OVERRIDE				(1 << 21)
+#define		LS_OVERRIDE				(1 << 22)
+
+#define	SPI_LB_CU_MASK					0x9354
 
 #define	TA_CNTL_AUX					0x9508
 
@@ -705,6 +1056,8 @@
 #define	CB_PERFCOUNTER3_SELECT0				0x9a38
 #define	CB_PERFCOUNTER3_SELECT1				0x9a3c
 
+#define	CB_CGTT_SCLK_CTRL				0x9a60
+
 #define	GC_USER_RB_BACKEND_DISABLE			0x9B7C
 #define		BACKEND_DISABLE_MASK			0x00FF0000
 #define		BACKEND_DISABLE_SHIFT			16
@@ -762,6 +1115,9 @@
 #       define CP_RINGID1_INT_STAT                      (1 << 30)
 #       define CP_RINGID0_INT_STAT                      (1 << 31)
 
+#define	CP_MEM_SLP_CNTL					0xC1E4
+#       define CP_MEM_LS_EN                             (1 << 0)
+
 #define	CP_DEBUG					0xC1FC
 
 #define RLC_CNTL                                          0xC300
@@ -769,6 +1125,7 @@
 #define RLC_RL_BASE                                       0xC304
 #define RLC_RL_SIZE                                       0xC308
 #define RLC_LB_CNTL                                       0xC30C
+#       define LOAD_BALANCE_ENABLE                        (1 << 0)
 #define RLC_SAVE_AND_RESTORE_BASE                         0xC310
 #define RLC_LB_CNTR_MAX                                   0xC314
 #define RLC_LB_CNTR_INIT                                  0xC318
@@ -783,6 +1140,56 @@
 #define RLC_CAPTURE_GPU_CLOCK_COUNT                       0xC340
 #define RLC_MC_CNTL                                       0xC344
 #define RLC_UCODE_CNTL                                    0xC348
+#define RLC_STAT                                          0xC34C
+#       define RLC_BUSY_STATUS                            (1 << 0)
+#       define GFX_POWER_STATUS                           (1 << 1)
+#       define GFX_CLOCK_STATUS                           (1 << 2)
+#       define GFX_LS_STATUS                              (1 << 3)
+
+#define	RLC_PG_CNTL					0xC35C
+#	define GFX_PG_ENABLE				(1 << 0)
+#	define GFX_PG_SRC				(1 << 1)
+
+#define	RLC_CGTT_MGCG_OVERRIDE				0xC400
+#define	RLC_CGCG_CGLS_CTRL				0xC404
+#	define CGCG_EN					(1 << 0)
+#	define CGLS_EN					(1 << 1)
+
+#define	RLC_TTOP_D					0xC414
+#	define RLC_PUD(x)				((x) << 0)
+#	define RLC_PUD_MASK				(0xff << 0)
+#	define RLC_PDD(x)				((x) << 8)
+#	define RLC_PDD_MASK				(0xff << 8)
+#	define RLC_TTPD(x)				((x) << 16)
+#	define RLC_TTPD_MASK				(0xff << 16)
+#	define RLC_MSD(x)				((x) << 24)
+#	define RLC_MSD_MASK				(0xff << 24)
+
+#define RLC_LB_INIT_CU_MASK                               0xC41C
+
+#define	RLC_PG_AO_CU_MASK				0xC42C
+#define	RLC_MAX_PG_CU					0xC430
+#	define MAX_PU_CU(x)				((x) << 0)
+#	define MAX_PU_CU_MASK				(0xff << 0)
+#define	RLC_AUTO_PG_CTRL				0xC434
+#	define AUTO_PG_EN				(1 << 0)
+#	define GRBM_REG_SGIT(x)				((x) << 3)
+#	define GRBM_REG_SGIT_MASK			(0xffff << 3)
+#	define PG_AFTER_GRBM_REG_ST(x)			((x) << 19)
+#	define PG_AFTER_GRBM_REG_ST_MASK		(0x1fff << 19)
+
+#define RLC_SERDES_WR_MASTER_MASK_0                       0xC454
+#define RLC_SERDES_WR_MASTER_MASK_1                       0xC458
+#define RLC_SERDES_WR_CTRL                                0xC45C
+
+#define RLC_SERDES_MASTER_BUSY_0                          0xC464
+#define RLC_SERDES_MASTER_BUSY_1                          0xC468
+
+#define RLC_GCPM_GENERAL_3                                0xC478
+
+#define	DB_RENDER_CONTROL				0x28000
+
+#define DB_DEPTH_INFO                                   0x2803c
 
 #define PA_SC_RASTER_CONFIG                             0x28350
 #       define RASTER_CONFIG_RB_MAP_0                   0
@@ -829,6 +1236,146 @@
 #       define THREAD_TRACE_FLUSH                       (54 << 0)
 #       define THREAD_TRACE_FINISH                      (55 << 0)
 
+/* PIF PHY0 registers idx/data 0x8/0xc */
+#define PB0_PIF_CNTL                                      0x10
+#       define LS2_EXIT_TIME(x)                           ((x) << 17)
+#       define LS2_EXIT_TIME_MASK                         (0x7 << 17)
+#       define LS2_EXIT_TIME_SHIFT                        17
+#define PB0_PIF_PAIRING                                   0x11
+#       define MULTI_PIF                                  (1 << 25)
+#define PB0_PIF_PWRDOWN_0                                 0x12
+#       define PLL_POWER_STATE_IN_TXS2_0(x)               ((x) << 7)
+#       define PLL_POWER_STATE_IN_TXS2_0_MASK             (0x7 << 7)
+#       define PLL_POWER_STATE_IN_TXS2_0_SHIFT            7
+#       define PLL_POWER_STATE_IN_OFF_0(x)                ((x) << 10)
+#       define PLL_POWER_STATE_IN_OFF_0_MASK              (0x7 << 10)
+#       define PLL_POWER_STATE_IN_OFF_0_SHIFT             10
+#       define PLL_RAMP_UP_TIME_0(x)                      ((x) << 24)
+#       define PLL_RAMP_UP_TIME_0_MASK                    (0x7 << 24)
+#       define PLL_RAMP_UP_TIME_0_SHIFT                   24
+#define PB0_PIF_PWRDOWN_1                                 0x13
+#       define PLL_POWER_STATE_IN_TXS2_1(x)               ((x) << 7)
+#       define PLL_POWER_STATE_IN_TXS2_1_MASK             (0x7 << 7)
+#       define PLL_POWER_STATE_IN_TXS2_1_SHIFT            7
+#       define PLL_POWER_STATE_IN_OFF_1(x)                ((x) << 10)
+#       define PLL_POWER_STATE_IN_OFF_1_MASK              (0x7 << 10)
+#       define PLL_POWER_STATE_IN_OFF_1_SHIFT             10
+#       define PLL_RAMP_UP_TIME_1(x)                      ((x) << 24)
+#       define PLL_RAMP_UP_TIME_1_MASK                    (0x7 << 24)
+#       define PLL_RAMP_UP_TIME_1_SHIFT                   24
+
+#define PB0_PIF_PWRDOWN_2                                 0x17
+#       define PLL_POWER_STATE_IN_TXS2_2(x)               ((x) << 7)
+#       define PLL_POWER_STATE_IN_TXS2_2_MASK             (0x7 << 7)
+#       define PLL_POWER_STATE_IN_TXS2_2_SHIFT            7
+#       define PLL_POWER_STATE_IN_OFF_2(x)                ((x) << 10)
+#       define PLL_POWER_STATE_IN_OFF_2_MASK              (0x7 << 10)
+#       define PLL_POWER_STATE_IN_OFF_2_SHIFT             10
+#       define PLL_RAMP_UP_TIME_2(x)                      ((x) << 24)
+#       define PLL_RAMP_UP_TIME_2_MASK                    (0x7 << 24)
+#       define PLL_RAMP_UP_TIME_2_SHIFT                   24
+#define PB0_PIF_PWRDOWN_3                                 0x18
+#       define PLL_POWER_STATE_IN_TXS2_3(x)               ((x) << 7)
+#       define PLL_POWER_STATE_IN_TXS2_3_MASK             (0x7 << 7)
+#       define PLL_POWER_STATE_IN_TXS2_3_SHIFT            7
+#       define PLL_POWER_STATE_IN_OFF_3(x)                ((x) << 10)
+#       define PLL_POWER_STATE_IN_OFF_3_MASK              (0x7 << 10)
+#       define PLL_POWER_STATE_IN_OFF_3_SHIFT             10
+#       define PLL_RAMP_UP_TIME_3(x)                      ((x) << 24)
+#       define PLL_RAMP_UP_TIME_3_MASK                    (0x7 << 24)
+#       define PLL_RAMP_UP_TIME_3_SHIFT                   24
+/* PIF PHY1 registers idx/data 0x10/0x14 */
+#define PB1_PIF_CNTL                                      0x10
+#define PB1_PIF_PAIRING                                   0x11
+#define PB1_PIF_PWRDOWN_0                                 0x12
+#define PB1_PIF_PWRDOWN_1                                 0x13
+
+#define PB1_PIF_PWRDOWN_2                                 0x17
+#define PB1_PIF_PWRDOWN_3                                 0x18
+/* PCIE registers idx/data 0x30/0x34 */
+#define PCIE_CNTL2                                        0x1c /* PCIE */
+#       define SLV_MEM_LS_EN                              (1 << 16)
+#       define MST_MEM_LS_EN                              (1 << 18)
+#       define REPLAY_MEM_LS_EN                           (1 << 19)
+#define PCIE_LC_STATUS1                                   0x28 /* PCIE */
+#       define LC_REVERSE_RCVR                            (1 << 0)
+#       define LC_REVERSE_XMIT                            (1 << 1)
+#       define LC_OPERATING_LINK_WIDTH_MASK               (0x7 << 2)
+#       define LC_OPERATING_LINK_WIDTH_SHIFT              2
+#       define LC_DETECTED_LINK_WIDTH_MASK                (0x7 << 5)
+#       define LC_DETECTED_LINK_WIDTH_SHIFT               5
+
+#define PCIE_P_CNTL                                       0x40 /* PCIE */
+#       define P_IGNORE_EDB_ERR                           (1 << 6)
+
+/* PCIE PORT registers idx/data 0x38/0x3c */
+#define PCIE_LC_CNTL                                      0xa0
+#       define LC_L0S_INACTIVITY(x)                       ((x) << 8)
+#       define LC_L0S_INACTIVITY_MASK                     (0xf << 8)
+#       define LC_L0S_INACTIVITY_SHIFT                    8
+#       define LC_L1_INACTIVITY(x)                        ((x) << 12)
+#       define LC_L1_INACTIVITY_MASK                      (0xf << 12)
+#       define LC_L1_INACTIVITY_SHIFT                     12
+#       define LC_PMI_TO_L1_DIS                           (1 << 16)
+#       define LC_ASPM_TO_L1_DIS                          (1 << 24)
+#define PCIE_LC_LINK_WIDTH_CNTL                           0xa2 /* PCIE_P */
+#       define LC_LINK_WIDTH_SHIFT                        0
+#       define LC_LINK_WIDTH_MASK                         0x7
+#       define LC_LINK_WIDTH_X0                           0
+#       define LC_LINK_WIDTH_X1                           1
+#       define LC_LINK_WIDTH_X2                           2
+#       define LC_LINK_WIDTH_X4                           3
+#       define LC_LINK_WIDTH_X8                           4
+#       define LC_LINK_WIDTH_X16                          6
+#       define LC_LINK_WIDTH_RD_SHIFT                     4
+#       define LC_LINK_WIDTH_RD_MASK                      0x70
+#       define LC_RECONFIG_ARC_MISSING_ESCAPE             (1 << 7)
+#       define LC_RECONFIG_NOW                            (1 << 8)
+#       define LC_RENEGOTIATION_SUPPORT                   (1 << 9)
+#       define LC_RENEGOTIATE_EN                          (1 << 10)
+#       define LC_SHORT_RECONFIG_EN                       (1 << 11)
+#       define LC_UPCONFIGURE_SUPPORT                     (1 << 12)
+#       define LC_UPCONFIGURE_DIS                         (1 << 13)
+#       define LC_DYN_LANES_PWR_STATE(x)                  ((x) << 21)
+#       define LC_DYN_LANES_PWR_STATE_MASK                (0x3 << 21)
+#       define LC_DYN_LANES_PWR_STATE_SHIFT               21
+#define PCIE_LC_N_FTS_CNTL                                0xa3 /* PCIE_P */
+#       define LC_XMIT_N_FTS(x)                           ((x) << 0)
+#       define LC_XMIT_N_FTS_MASK                         (0xff << 0)
+#       define LC_XMIT_N_FTS_SHIFT                        0
+#       define LC_XMIT_N_FTS_OVERRIDE_EN                  (1 << 8)
+#       define LC_N_FTS_MASK                              (0xff << 24)
+#define PCIE_LC_SPEED_CNTL                                0xa4 /* PCIE_P */
+#       define LC_GEN2_EN_STRAP                           (1 << 0)
+#       define LC_GEN3_EN_STRAP                           (1 << 1)
+#       define LC_TARGET_LINK_SPEED_OVERRIDE_EN           (1 << 2)
+#       define LC_TARGET_LINK_SPEED_OVERRIDE_MASK         (0x3 << 3)
+#       define LC_TARGET_LINK_SPEED_OVERRIDE_SHIFT        3
+#       define LC_FORCE_EN_SW_SPEED_CHANGE                (1 << 5)
+#       define LC_FORCE_DIS_SW_SPEED_CHANGE               (1 << 6)
+#       define LC_FORCE_EN_HW_SPEED_CHANGE                (1 << 7)
+#       define LC_FORCE_DIS_HW_SPEED_CHANGE               (1 << 8)
+#       define LC_INITIATE_LINK_SPEED_CHANGE              (1 << 9)
+#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_MASK      (0x3 << 10)
+#       define LC_SPEED_CHANGE_ATTEMPTS_ALLOWED_SHIFT     10
+#       define LC_CURRENT_DATA_RATE_MASK                  (0x3 << 13) /* 0/1/2 = gen1/2/3 */
+#       define LC_CURRENT_DATA_RATE_SHIFT                 13
+#       define LC_CLR_FAILED_SPD_CHANGE_CNT               (1 << 16)
+#       define LC_OTHER_SIDE_EVER_SENT_GEN2               (1 << 18)
+#       define LC_OTHER_SIDE_SUPPORTS_GEN2                (1 << 19)
+#       define LC_OTHER_SIDE_EVER_SENT_GEN3               (1 << 20)
+#       define LC_OTHER_SIDE_SUPPORTS_GEN3                (1 << 21)
+
+#define PCIE_LC_CNTL2                                     0xb1
+#       define LC_ALLOW_PDWN_IN_L1                        (1 << 17)
+#       define LC_ALLOW_PDWN_IN_L23                       (1 << 18)
+
+#define PCIE_LC_CNTL3                                     0xb5 /* PCIE_P */
+#       define LC_GO_TO_RECOVERY                          (1 << 30)
+#define PCIE_LC_CNTL4                                     0xb6 /* PCIE_P */
+#       define LC_REDO_EQ                                 (1 << 5)
+#       define LC_SET_QUIESCE                             (1 << 13)
+
 /*
  * UVD
  */
@@ -838,6 +1385,21 @@
 #define UVD_RBC_RB_RPTR					0xF690
 #define UVD_RBC_RB_WPTR					0xF694
 
+#define	UVD_CGC_CTRL					0xF4B0
+#	define DCM					(1 << 0)
+#	define CG_DT(x)					((x) << 2)
+#	define CG_DT_MASK				(0xf << 2)
+#	define CLK_OD(x)				((x) << 6)
+#	define CLK_OD_MASK				(0x1f << 6)
+
+ /* UVD CTX indirect */
+#define	UVD_CGC_MEM_CTRL				0xC0
+#define	UVD_CGC_CTRL2					0xC1
+#	define DYN_OR_EN				(1 << 0)
+#	define DYN_RR_EN				(1 << 1)
+#	define G_DIV_ID(x)				((x) << 2)
+#	define G_DIV_ID_MASK				(0x7 << 2)
+
 /*
  * PM4
  */
@@ -1082,6 +1644,11 @@
 #       define DMA_IDLE                                   (1 << 0)
 #define DMA_TILING_CONFIG  				  0xd0b8
 
+#define	DMA_PG						0xd0d4
+#	define PG_CNTL_ENABLE				(1 << 0)
+#define	DMA_PGFSM_CONFIG				0xd0d8
+#define	DMA_PGFSM_WRITE					0xd0dc
+
 #define DMA_PACKET(cmd, b, t, s, n)	((((cmd) & 0xF) << 28) |	\
 					 (((b) & 0x1) << 26) |		\
 					 (((t) & 0x1) << 23) |		\
diff --git a/drivers/gpu/drm/radeon/sislands_smc.h b/drivers/gpu/drm/radeon/sislands_smc.h
new file mode 100644
index 000000000000..5578e9837026
--- /dev/null
+++ b/drivers/gpu/drm/radeon/sislands_smc.h
@@ -0,0 +1,397 @@
+/*
+ * Copyright 2013 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef PP_SISLANDS_SMC_H
+#define PP_SISLANDS_SMC_H
+
+#include "ppsmc.h"
+
+#pragma pack(push, 1)
+
+#define SISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE 16
+
+struct PP_SIslands_Dpm2PerfLevel
+{
+    uint8_t MaxPS;
+    uint8_t TgtAct;
+    uint8_t MaxPS_StepInc;
+    uint8_t MaxPS_StepDec;
+    uint8_t PSSamplingTime;
+    uint8_t NearTDPDec;
+    uint8_t AboveSafeInc;
+    uint8_t BelowSafeInc;
+    uint8_t PSDeltaLimit;
+    uint8_t PSDeltaWin;
+    uint16_t PwrEfficiencyRatio;
+    uint8_t Reserved[4];
+};
+
+typedef struct PP_SIslands_Dpm2PerfLevel PP_SIslands_Dpm2PerfLevel;
+
+struct PP_SIslands_DPM2Status
+{
+    uint32_t    dpm2Flags;
+    uint8_t     CurrPSkip;
+    uint8_t     CurrPSkipPowerShift;
+    uint8_t     CurrPSkipTDP;
+    uint8_t     CurrPSkipOCP;
+    uint8_t     MaxSPLLIndex;
+    uint8_t     MinSPLLIndex;
+    uint8_t     CurrSPLLIndex;
+    uint8_t     InfSweepMode;
+    uint8_t     InfSweepDir;
+    uint8_t     TDPexceeded;
+    uint8_t     reserved;
+    uint8_t     SwitchDownThreshold;
+    uint32_t    SwitchDownCounter;
+    uint32_t    SysScalingFactor;
+};
+
+typedef struct PP_SIslands_DPM2Status PP_SIslands_DPM2Status;
+
+struct PP_SIslands_DPM2Parameters
+{
+    uint32_t    TDPLimit;
+    uint32_t    NearTDPLimit;
+    uint32_t    SafePowerLimit;
+    uint32_t    PowerBoostLimit;
+    uint32_t    MinLimitDelta;
+};
+typedef struct PP_SIslands_DPM2Parameters PP_SIslands_DPM2Parameters;
+
+struct PP_SIslands_PAPMStatus
+{
+    uint32_t    EstimatedDGPU_T;
+    uint32_t    EstimatedDGPU_P;
+    uint32_t    EstimatedAPU_T;
+    uint32_t    EstimatedAPU_P;
+    uint8_t     dGPU_T_Limit_Exceeded;
+    uint8_t     reserved[3];
+};
+typedef struct PP_SIslands_PAPMStatus PP_SIslands_PAPMStatus;
+
+struct PP_SIslands_PAPMParameters
+{
+    uint32_t    NearTDPLimitTherm;
+    uint32_t    NearTDPLimitPAPM;
+    uint32_t    PlatformPowerLimit;
+    uint32_t    dGPU_T_Limit;
+    uint32_t    dGPU_T_Warning;
+    uint32_t    dGPU_T_Hysteresis;
+};
+typedef struct PP_SIslands_PAPMParameters PP_SIslands_PAPMParameters;
+
+struct SISLANDS_SMC_SCLK_VALUE
+{
+    uint32_t    vCG_SPLL_FUNC_CNTL;
+    uint32_t    vCG_SPLL_FUNC_CNTL_2;
+    uint32_t    vCG_SPLL_FUNC_CNTL_3;
+    uint32_t    vCG_SPLL_FUNC_CNTL_4;
+    uint32_t    vCG_SPLL_SPREAD_SPECTRUM;
+    uint32_t    vCG_SPLL_SPREAD_SPECTRUM_2;
+    uint32_t    sclk_value;
+};
+
+typedef struct SISLANDS_SMC_SCLK_VALUE SISLANDS_SMC_SCLK_VALUE;
+
+struct SISLANDS_SMC_MCLK_VALUE
+{
+    uint32_t    vMPLL_FUNC_CNTL;
+    uint32_t    vMPLL_FUNC_CNTL_1;
+    uint32_t    vMPLL_FUNC_CNTL_2;
+    uint32_t    vMPLL_AD_FUNC_CNTL;
+    uint32_t    vMPLL_DQ_FUNC_CNTL;
+    uint32_t    vMCLK_PWRMGT_CNTL;
+    uint32_t    vDLL_CNTL;
+    uint32_t    vMPLL_SS;
+    uint32_t    vMPLL_SS2;
+    uint32_t    mclk_value;
+};
+
+typedef struct SISLANDS_SMC_MCLK_VALUE SISLANDS_SMC_MCLK_VALUE;
+
+struct SISLANDS_SMC_VOLTAGE_VALUE
+{
+    uint16_t    value;
+    uint8_t     index;
+    uint8_t     phase_settings;
+};
+
+typedef struct SISLANDS_SMC_VOLTAGE_VALUE SISLANDS_SMC_VOLTAGE_VALUE;
+
+struct SISLANDS_SMC_HW_PERFORMANCE_LEVEL
+{
+    uint8_t                     ACIndex;
+    uint8_t                     displayWatermark;
+    uint8_t                     gen2PCIE;
+    uint8_t                     UVDWatermark;
+    uint8_t                     VCEWatermark;
+    uint8_t                     strobeMode;
+    uint8_t                     mcFlags;
+    uint8_t                     padding;
+    uint32_t                    aT;
+    uint32_t                    bSP;
+    SISLANDS_SMC_SCLK_VALUE     sclk;
+    SISLANDS_SMC_MCLK_VALUE     mclk;
+    SISLANDS_SMC_VOLTAGE_VALUE  vddc;
+    SISLANDS_SMC_VOLTAGE_VALUE  mvdd;
+    SISLANDS_SMC_VOLTAGE_VALUE  vddci;
+    SISLANDS_SMC_VOLTAGE_VALUE  std_vddc;
+    uint8_t                     hysteresisUp;
+    uint8_t                     hysteresisDown;
+    uint8_t                     stateFlags;
+    uint8_t                     arbRefreshState;
+    uint32_t                    SQPowerThrottle;
+    uint32_t                    SQPowerThrottle_2;
+    uint32_t                    MaxPoweredUpCU;
+    SISLANDS_SMC_VOLTAGE_VALUE  high_temp_vddc;
+    SISLANDS_SMC_VOLTAGE_VALUE  low_temp_vddc;
+    uint32_t                    reserved[2];
+    PP_SIslands_Dpm2PerfLevel   dpm2;
+};
+
+#define SISLANDS_SMC_STROBE_RATIO    0x0F
+#define SISLANDS_SMC_STROBE_ENABLE   0x10
+
+#define SISLANDS_SMC_MC_EDC_RD_FLAG  0x01
+#define SISLANDS_SMC_MC_EDC_WR_FLAG  0x02
+#define SISLANDS_SMC_MC_RTT_ENABLE   0x04
+#define SISLANDS_SMC_MC_STUTTER_EN   0x08
+#define SISLANDS_SMC_MC_PG_EN        0x10
+
+typedef struct SISLANDS_SMC_HW_PERFORMANCE_LEVEL SISLANDS_SMC_HW_PERFORMANCE_LEVEL;
+
+struct SISLANDS_SMC_SWSTATE
+{
+    uint8_t                             flags;
+    uint8_t                             levelCount;
+    uint8_t                             padding2;
+    uint8_t                             padding3;
+    SISLANDS_SMC_HW_PERFORMANCE_LEVEL   levels[1];
+};
+
+typedef struct SISLANDS_SMC_SWSTATE SISLANDS_SMC_SWSTATE;
+
+#define SISLANDS_SMC_VOLTAGEMASK_VDDC  0
+#define SISLANDS_SMC_VOLTAGEMASK_MVDD  1
+#define SISLANDS_SMC_VOLTAGEMASK_VDDCI 2
+#define SISLANDS_SMC_VOLTAGEMASK_MAX   4
+
+struct SISLANDS_SMC_VOLTAGEMASKTABLE
+{
+    uint32_t lowMask[SISLANDS_SMC_VOLTAGEMASK_MAX];
+};
+
+typedef struct SISLANDS_SMC_VOLTAGEMASKTABLE SISLANDS_SMC_VOLTAGEMASKTABLE;
+
+#define SISLANDS_MAX_NO_VREG_STEPS 32
+
+struct SISLANDS_SMC_STATETABLE
+{
+    uint8_t                             thermalProtectType;
+    uint8_t                             systemFlags;
+    uint8_t                             maxVDDCIndexInPPTable;
+    uint8_t                             extraFlags;
+    uint32_t                            lowSMIO[SISLANDS_MAX_NO_VREG_STEPS];
+    SISLANDS_SMC_VOLTAGEMASKTABLE       voltageMaskTable;
+    SISLANDS_SMC_VOLTAGEMASKTABLE       phaseMaskTable;
+    PP_SIslands_DPM2Parameters          dpm2Params;
+    SISLANDS_SMC_SWSTATE                initialState;
+    SISLANDS_SMC_SWSTATE                ACPIState;
+    SISLANDS_SMC_SWSTATE                ULVState;
+    SISLANDS_SMC_SWSTATE                driverState;
+    SISLANDS_SMC_HW_PERFORMANCE_LEVEL   dpmLevels[SISLANDS_MAX_SMC_PERFORMANCE_LEVELS_PER_SWSTATE - 1];
+};
+
+typedef struct SISLANDS_SMC_STATETABLE SISLANDS_SMC_STATETABLE;
+
+#define SI_SMC_SOFT_REGISTER_mclk_chg_timeout         0x0
+#define SI_SMC_SOFT_REGISTER_delay_vreg               0xC
+#define SI_SMC_SOFT_REGISTER_delay_acpi               0x28
+#define SI_SMC_SOFT_REGISTER_seq_index                0x5C
+#define SI_SMC_SOFT_REGISTER_mvdd_chg_time            0x60
+#define SI_SMC_SOFT_REGISTER_mclk_switch_lim          0x70
+#define SI_SMC_SOFT_REGISTER_watermark_threshold      0x78
+#define SI_SMC_SOFT_REGISTER_phase_shedding_delay     0x88
+#define SI_SMC_SOFT_REGISTER_ulv_volt_change_delay    0x8C
+#define SI_SMC_SOFT_REGISTER_mc_block_delay           0x98
+#define SI_SMC_SOFT_REGISTER_ticks_per_us             0xA8
+#define SI_SMC_SOFT_REGISTER_crtc_index               0xC4
+#define SI_SMC_SOFT_REGISTER_mclk_change_block_cp_min 0xC8
+#define SI_SMC_SOFT_REGISTER_mclk_change_block_cp_max 0xCC
+#define SI_SMC_SOFT_REGISTER_non_ulv_pcie_link_width  0xF4
+#define SI_SMC_SOFT_REGISTER_tdr_is_about_to_happen   0xFC
+#define SI_SMC_SOFT_REGISTER_vr_hot_gpio              0x100
+
+#define SMC_SISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES 16
+#define SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES 32
+
+#define SMC_SISLANDS_SCALE_I  7
+#define SMC_SISLANDS_SCALE_R 12
+
+struct PP_SIslands_CacConfig
+{
+    uint16_t   cac_lkge_lut[SMC_SISLANDS_LKGE_LUT_NUM_OF_TEMP_ENTRIES][SMC_SISLANDS_LKGE_LUT_NUM_OF_VOLT_ENTRIES];
+    uint32_t   lkge_lut_V0;
+    uint32_t   lkge_lut_Vstep;
+    uint32_t   WinTime;
+    uint32_t   R_LL;
+    uint32_t   calculation_repeats;
+    uint32_t   l2numWin_TDP;
+    uint32_t   dc_cac;
+    uint8_t    lts_truncate_n;
+    uint8_t    SHIFT_N;
+    uint8_t    log2_PG_LKG_SCALE;
+    uint8_t    cac_temp;
+    uint32_t   lkge_lut_T0;
+    uint32_t   lkge_lut_Tstep;
+};
+
+typedef struct PP_SIslands_CacConfig PP_SIslands_CacConfig;
+
+#define SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE 16
+#define SMC_SISLANDS_MC_REGISTER_ARRAY_SET_COUNT 20
+
+struct SMC_SIslands_MCRegisterAddress
+{
+    uint16_t s0;
+    uint16_t s1;
+};
+
+typedef struct SMC_SIslands_MCRegisterAddress SMC_SIslands_MCRegisterAddress;
+
+struct SMC_SIslands_MCRegisterSet
+{
+    uint32_t value[SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE];
+};
+
+typedef struct SMC_SIslands_MCRegisterSet SMC_SIslands_MCRegisterSet;
+
+struct SMC_SIslands_MCRegisters
+{
+    uint8_t                             last;
+    uint8_t                             reserved[3];
+    SMC_SIslands_MCRegisterAddress      address[SMC_SISLANDS_MC_REGISTER_ARRAY_SIZE];
+    SMC_SIslands_MCRegisterSet          data[SMC_SISLANDS_MC_REGISTER_ARRAY_SET_COUNT];
+};
+
+typedef struct SMC_SIslands_MCRegisters SMC_SIslands_MCRegisters;
+
+struct SMC_SIslands_MCArbDramTimingRegisterSet
+{
+    uint32_t mc_arb_dram_timing;
+    uint32_t mc_arb_dram_timing2;
+    uint8_t  mc_arb_rfsh_rate;
+    uint8_t  mc_arb_burst_time;
+    uint8_t  padding[2];
+};
+
+typedef struct SMC_SIslands_MCArbDramTimingRegisterSet SMC_SIslands_MCArbDramTimingRegisterSet;
+
+struct SMC_SIslands_MCArbDramTimingRegisters
+{
+    uint8_t                                     arb_current;
+    uint8_t                                     reserved[3];
+    SMC_SIslands_MCArbDramTimingRegisterSet     data[16];
+};
+
+typedef struct SMC_SIslands_MCArbDramTimingRegisters SMC_SIslands_MCArbDramTimingRegisters;
+
+struct SMC_SISLANDS_SPLL_DIV_TABLE
+{
+    uint32_t    freq[256];
+    uint32_t    ss[256];
+};
+
+#define SMC_SISLANDS_SPLL_DIV_TABLE_FBDIV_MASK  0x01ffffff
+#define SMC_SISLANDS_SPLL_DIV_TABLE_FBDIV_SHIFT 0
+#define SMC_SISLANDS_SPLL_DIV_TABLE_PDIV_MASK   0xfe000000
+#define SMC_SISLANDS_SPLL_DIV_TABLE_PDIV_SHIFT  25
+#define SMC_SISLANDS_SPLL_DIV_TABLE_CLKV_MASK   0x000fffff
+#define SMC_SISLANDS_SPLL_DIV_TABLE_CLKV_SHIFT  0
+#define SMC_SISLANDS_SPLL_DIV_TABLE_CLKS_MASK   0xfff00000
+#define SMC_SISLANDS_SPLL_DIV_TABLE_CLKS_SHIFT  20
+
+typedef struct SMC_SISLANDS_SPLL_DIV_TABLE SMC_SISLANDS_SPLL_DIV_TABLE;
+
+#define SMC_SISLANDS_DTE_MAX_FILTER_STAGES 5
+
+#define SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE 16
+
+struct Smc_SIslands_DTE_Configuration
+{
+    uint32_t tau[SMC_SISLANDS_DTE_MAX_FILTER_STAGES];
+    uint32_t R[SMC_SISLANDS_DTE_MAX_FILTER_STAGES];
+    uint32_t K;
+    uint32_t T0;
+    uint32_t MaxT;
+    uint8_t  WindowSize;
+    uint8_t  Tdep_count;
+    uint8_t  temp_select;
+    uint8_t  DTE_mode;
+    uint8_t  T_limits[SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE];
+    uint32_t Tdep_tau[SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE];
+    uint32_t Tdep_R[SMC_SISLANDS_DTE_MAX_TEMPERATURE_DEPENDENT_ARRAY_SIZE];
+    uint32_t Tthreshold;
+};
+
+typedef struct Smc_SIslands_DTE_Configuration Smc_SIslands_DTE_Configuration;
+
+#define SMC_SISLANDS_DTE_STATUS_FLAG_DTE_ON 1
+
+#define SISLANDS_SMC_FIRMWARE_HEADER_LOCATION 0x10000
+
+#define SISLANDS_SMC_FIRMWARE_HEADER_version                   0x0
+#define SISLANDS_SMC_FIRMWARE_HEADER_flags                     0x4
+#define SISLANDS_SMC_FIRMWARE_HEADER_softRegisters             0xC
+#define SISLANDS_SMC_FIRMWARE_HEADER_stateTable                0x10
+#define SISLANDS_SMC_FIRMWARE_HEADER_fanTable                  0x14
+#define SISLANDS_SMC_FIRMWARE_HEADER_CacConfigTable            0x18
+#define SISLANDS_SMC_FIRMWARE_HEADER_mcRegisterTable           0x24
+#define SISLANDS_SMC_FIRMWARE_HEADER_mcArbDramAutoRefreshTable 0x30
+#define SISLANDS_SMC_FIRMWARE_HEADER_spllTable                 0x38
+#define SISLANDS_SMC_FIRMWARE_HEADER_DteConfiguration          0x40
+#define SISLANDS_SMC_FIRMWARE_HEADER_PAPMParameters            0x48
+
+#pragma pack(pop)
+
+int si_set_smc_sram_address(struct radeon_device *rdev,
+			    u32 smc_address, u32 limit);
+int si_copy_bytes_to_smc(struct radeon_device *rdev,
+			 u32 smc_start_address,
+			 const u8 *src, u32 byte_count, u32 limit);
+void si_start_smc(struct radeon_device *rdev);
+void si_reset_smc(struct radeon_device *rdev);
+int si_program_jump_on_start(struct radeon_device *rdev);
+void si_stop_smc_clock(struct radeon_device *rdev);
+void si_start_smc_clock(struct radeon_device *rdev);
+bool si_is_smc_running(struct radeon_device *rdev);
+PPSMC_Result si_send_msg_to_smc(struct radeon_device *rdev, PPSMC_Msg msg);
+PPSMC_Result si_wait_for_smc_inactive(struct radeon_device *rdev);
+int si_load_smc_ucode(struct radeon_device *rdev, u32 limit);
+int si_read_smc_sram_dword(struct radeon_device *rdev, u32 smc_address,
+			   u32 *value, u32 limit);
+int si_write_smc_sram_dword(struct radeon_device *rdev, u32 smc_address,
+			    u32 value, u32 limit);
+
+#endif
+
diff --git a/drivers/gpu/drm/radeon/sumo_dpm.c b/drivers/gpu/drm/radeon/sumo_dpm.c
new file mode 100644
index 000000000000..c0a850319908
--- /dev/null
+++ b/drivers/gpu/drm/radeon/sumo_dpm.c
@@ -0,0 +1,1888 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "sumod.h"
+#include "r600_dpm.h"
+#include "cypress_dpm.h"
+#include "sumo_dpm.h"
+#include <linux/seq_file.h>
+
+#define SUMO_MAX_DEEPSLEEP_DIVIDER_ID 5
+#define SUMO_MINIMUM_ENGINE_CLOCK 800
+#define BOOST_DPM_LEVEL 7
+
+static const u32 sumo_utc[SUMO_PM_NUMBER_OF_TC] =
+{
+	SUMO_UTC_DFLT_00,
+	SUMO_UTC_DFLT_01,
+	SUMO_UTC_DFLT_02,
+	SUMO_UTC_DFLT_03,
+	SUMO_UTC_DFLT_04,
+	SUMO_UTC_DFLT_05,
+	SUMO_UTC_DFLT_06,
+	SUMO_UTC_DFLT_07,
+	SUMO_UTC_DFLT_08,
+	SUMO_UTC_DFLT_09,
+	SUMO_UTC_DFLT_10,
+	SUMO_UTC_DFLT_11,
+	SUMO_UTC_DFLT_12,
+	SUMO_UTC_DFLT_13,
+	SUMO_UTC_DFLT_14,
+};
+
+static const u32 sumo_dtc[SUMO_PM_NUMBER_OF_TC] =
+{
+	SUMO_DTC_DFLT_00,
+	SUMO_DTC_DFLT_01,
+	SUMO_DTC_DFLT_02,
+	SUMO_DTC_DFLT_03,
+	SUMO_DTC_DFLT_04,
+	SUMO_DTC_DFLT_05,
+	SUMO_DTC_DFLT_06,
+	SUMO_DTC_DFLT_07,
+	SUMO_DTC_DFLT_08,
+	SUMO_DTC_DFLT_09,
+	SUMO_DTC_DFLT_10,
+	SUMO_DTC_DFLT_11,
+	SUMO_DTC_DFLT_12,
+	SUMO_DTC_DFLT_13,
+	SUMO_DTC_DFLT_14,
+};
+
+struct sumo_ps *sumo_get_ps(struct radeon_ps *rps)
+{
+	struct sumo_ps *ps = rps->ps_priv;
+
+	return ps;
+}
+
+struct sumo_power_info *sumo_get_pi(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = rdev->pm.dpm.priv;
+
+	return pi;
+}
+
+static void sumo_gfx_clockgating_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+	else {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+		RREG32(GB_ADDR_CONFIG);
+	}
+}
+
+#define CGCG_CGTT_LOCAL0_MASK 0xE5BFFFFF
+#define CGCG_CGTT_LOCAL1_MASK 0xEFFF07FF
+
+static void sumo_mg_clockgating_enable(struct radeon_device *rdev, bool enable)
+{
+	u32 local0;
+	u32 local1;
+
+	local0 = RREG32(CG_CGTT_LOCAL_0);
+	local1 = RREG32(CG_CGTT_LOCAL_1);
+
+	if (enable) {
+		WREG32(CG_CGTT_LOCAL_0, (0 & CGCG_CGTT_LOCAL0_MASK) | (local0 & ~CGCG_CGTT_LOCAL0_MASK) );
+		WREG32(CG_CGTT_LOCAL_1, (0 & CGCG_CGTT_LOCAL1_MASK) | (local1 & ~CGCG_CGTT_LOCAL1_MASK) );
+	} else {
+		WREG32(CG_CGTT_LOCAL_0, (0xFFFFFFFF & CGCG_CGTT_LOCAL0_MASK) | (local0 & ~CGCG_CGTT_LOCAL0_MASK) );
+		WREG32(CG_CGTT_LOCAL_1, (0xFFFFCFFF & CGCG_CGTT_LOCAL1_MASK) | (local1 & ~CGCG_CGTT_LOCAL1_MASK) );
+	}
+}
+
+static void sumo_program_git(struct radeon_device *rdev)
+{
+	u32 p, u;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	r600_calculate_u_and_p(SUMO_GICST_DFLT,
+			       xclk, 16, &p, &u);
+
+	WREG32_P(CG_GIT, CG_GICST(p), ~CG_GICST_MASK);
+}
+
+static void sumo_program_grsd(struct radeon_device *rdev)
+{
+	u32 p, u;
+	u32 xclk = radeon_get_xclk(rdev);
+	u32 grs = 256 * 25 / 100;
+
+	r600_calculate_u_and_p(1, xclk, 14, &p, &u);
+
+	WREG32(CG_GCOOR, PHC(grs) | SDC(p) | SU(u));
+}
+
+void sumo_gfx_clockgating_initialize(struct radeon_device *rdev)
+{
+	sumo_program_git(rdev);
+	sumo_program_grsd(rdev);
+}
+
+static void sumo_gfx_powergating_initialize(struct radeon_device *rdev)
+{
+	u32 rcu_pwr_gating_cntl;
+	u32 p, u;
+	u32 p_c, p_p, d_p;
+	u32 r_t, i_t;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	if (rdev->family == CHIP_PALM) {
+		p_c = 4;
+		d_p = 10;
+		r_t = 10;
+		i_t = 4;
+		p_p = 50 + 1000/200 + 6 * 32;
+	} else {
+		p_c = 16;
+		d_p = 50;
+		r_t = 50;
+		i_t  = 50;
+		p_p = 113;
+	}
+
+	WREG32(CG_SCRATCH2, 0x01B60A17);
+
+	r600_calculate_u_and_p(SUMO_GFXPOWERGATINGT_DFLT,
+			       xclk, 16, &p, &u);
+
+	WREG32_P(CG_PWR_GATING_CNTL, PGP(p) | PGU(u),
+		 ~(PGP_MASK | PGU_MASK));
+
+	r600_calculate_u_and_p(SUMO_VOLTAGEDROPT_DFLT,
+			       xclk, 16, &p, &u);
+
+	WREG32_P(CG_CG_VOLTAGE_CNTL, PGP(p) | PGU(u),
+		 ~(PGP_MASK | PGU_MASK));
+
+	if (rdev->family == CHIP_PALM) {
+		WREG32_RCU(RCU_PWR_GATING_SEQ0, 0x10103210);
+		WREG32_RCU(RCU_PWR_GATING_SEQ1, 0x10101010);
+	} else {
+		WREG32_RCU(RCU_PWR_GATING_SEQ0, 0x76543210);
+		WREG32_RCU(RCU_PWR_GATING_SEQ1, 0xFEDCBA98);
+	}
+
+	rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL);
+	rcu_pwr_gating_cntl &=
+		~(RSVD_MASK | PCV_MASK | PGS_MASK);
+	rcu_pwr_gating_cntl |= PCV(p_c) | PGS(1) | PWR_GATING_EN;
+	if (rdev->family == CHIP_PALM) {
+		rcu_pwr_gating_cntl &= ~PCP_MASK;
+		rcu_pwr_gating_cntl |= PCP(0x77);
+	}
+	WREG32_RCU(RCU_PWR_GATING_CNTL, rcu_pwr_gating_cntl);
+
+	rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_2);
+	rcu_pwr_gating_cntl &= ~(MPPU_MASK | MPPD_MASK);
+	rcu_pwr_gating_cntl |= MPPU(p_p) | MPPD(50);
+	WREG32_RCU(RCU_PWR_GATING_CNTL_2, rcu_pwr_gating_cntl);
+
+	rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_3);
+	rcu_pwr_gating_cntl &= ~(DPPU_MASK | DPPD_MASK);
+	rcu_pwr_gating_cntl |= DPPU(d_p) | DPPD(50);
+	WREG32_RCU(RCU_PWR_GATING_CNTL_3, rcu_pwr_gating_cntl);
+
+	rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_4);
+	rcu_pwr_gating_cntl &= ~(RT_MASK | IT_MASK);
+	rcu_pwr_gating_cntl |= RT(r_t) | IT(i_t);
+	WREG32_RCU(RCU_PWR_GATING_CNTL_4, rcu_pwr_gating_cntl);
+
+	if (rdev->family == CHIP_PALM)
+		WREG32_RCU(RCU_PWR_GATING_CNTL_5, 0xA02);
+
+	sumo_smu_pg_init(rdev);
+
+	rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL);
+	rcu_pwr_gating_cntl &=
+		~(RSVD_MASK | PCV_MASK | PGS_MASK);
+	rcu_pwr_gating_cntl |= PCV(p_c) | PGS(4) | PWR_GATING_EN;
+	if (rdev->family == CHIP_PALM) {
+		rcu_pwr_gating_cntl &= ~PCP_MASK;
+		rcu_pwr_gating_cntl |= PCP(0x77);
+	}
+	WREG32_RCU(RCU_PWR_GATING_CNTL, rcu_pwr_gating_cntl);
+
+	if (rdev->family == CHIP_PALM) {
+		rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_2);
+		rcu_pwr_gating_cntl &= ~(MPPU_MASK | MPPD_MASK);
+		rcu_pwr_gating_cntl |= MPPU(113) | MPPD(50);
+		WREG32_RCU(RCU_PWR_GATING_CNTL_2, rcu_pwr_gating_cntl);
+
+		rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_3);
+		rcu_pwr_gating_cntl &= ~(DPPU_MASK | DPPD_MASK);
+		rcu_pwr_gating_cntl |= DPPU(16) | DPPD(50);
+		WREG32_RCU(RCU_PWR_GATING_CNTL_3, rcu_pwr_gating_cntl);
+	}
+
+	sumo_smu_pg_init(rdev);
+
+	rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL);
+	rcu_pwr_gating_cntl &=
+		~(RSVD_MASK | PCV_MASK | PGS_MASK);
+	rcu_pwr_gating_cntl |= PGS(5) | PWR_GATING_EN;
+
+	if (rdev->family == CHIP_PALM) {
+		rcu_pwr_gating_cntl |= PCV(4);
+		rcu_pwr_gating_cntl &= ~PCP_MASK;
+		rcu_pwr_gating_cntl |= PCP(0x77);
+	} else
+		rcu_pwr_gating_cntl |= PCV(11);
+	WREG32_RCU(RCU_PWR_GATING_CNTL, rcu_pwr_gating_cntl);
+
+	if (rdev->family == CHIP_PALM) {
+		rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_2);
+		rcu_pwr_gating_cntl &= ~(MPPU_MASK | MPPD_MASK);
+		rcu_pwr_gating_cntl |= MPPU(113) | MPPD(50);
+		WREG32_RCU(RCU_PWR_GATING_CNTL_2, rcu_pwr_gating_cntl);
+
+		rcu_pwr_gating_cntl = RREG32_RCU(RCU_PWR_GATING_CNTL_3);
+		rcu_pwr_gating_cntl &= ~(DPPU_MASK | DPPD_MASK);
+		rcu_pwr_gating_cntl |= DPPU(22) | DPPD(50);
+		WREG32_RCU(RCU_PWR_GATING_CNTL_3, rcu_pwr_gating_cntl);
+	}
+
+	sumo_smu_pg_init(rdev);
+}
+
+static void sumo_gfx_powergating_enable(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(CG_PWR_GATING_CNTL, DYN_PWR_DOWN_EN, ~DYN_PWR_DOWN_EN);
+	else {
+		WREG32_P(CG_PWR_GATING_CNTL, 0, ~DYN_PWR_DOWN_EN);
+		RREG32(GB_ADDR_CONFIG);
+	}
+}
+
+static int sumo_enable_clock_power_gating(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	if (pi->enable_gfx_clock_gating)
+		sumo_gfx_clockgating_initialize(rdev);
+	if (pi->enable_gfx_power_gating)
+		sumo_gfx_powergating_initialize(rdev);
+	if (pi->enable_mg_clock_gating)
+		sumo_mg_clockgating_enable(rdev, true);
+	if (pi->enable_gfx_clock_gating)
+		sumo_gfx_clockgating_enable(rdev, true);
+	if (pi->enable_gfx_power_gating)
+		sumo_gfx_powergating_enable(rdev, true);
+
+	return 0;
+}
+
+static void sumo_disable_clock_power_gating(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	if (pi->enable_gfx_clock_gating)
+		sumo_gfx_clockgating_enable(rdev, false);
+	if (pi->enable_gfx_power_gating)
+		sumo_gfx_powergating_enable(rdev, false);
+	if (pi->enable_mg_clock_gating)
+		sumo_mg_clockgating_enable(rdev, false);
+}
+
+static void sumo_calculate_bsp(struct radeon_device *rdev,
+			       u32 high_clk)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 xclk = radeon_get_xclk(rdev);
+
+	pi->pasi = 65535 * 100 / high_clk;
+	pi->asi = 65535 * 100 / high_clk;
+
+	r600_calculate_u_and_p(pi->asi,
+			       xclk, 16, &pi->bsp, &pi->bsu);
+
+	r600_calculate_u_and_p(pi->pasi,
+			       xclk, 16, &pi->pbsp, &pi->pbsu);
+
+	pi->dsp = BSP(pi->bsp) | BSU(pi->bsu);
+	pi->psp = BSP(pi->pbsp) | BSU(pi->pbsu);
+}
+
+static void sumo_init_bsp(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	WREG32(CG_BSP_0, pi->psp);
+}
+
+
+static void sumo_program_bsp(struct radeon_device *rdev,
+			     struct radeon_ps *rps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *ps = sumo_get_ps(rps);
+	u32 i;
+	u32 highest_engine_clock = ps->levels[ps->num_levels - 1].sclk;
+
+	if (ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE)
+		highest_engine_clock = pi->boost_pl.sclk;
+
+	sumo_calculate_bsp(rdev, highest_engine_clock);
+
+	for (i = 0; i < ps->num_levels - 1; i++)
+		WREG32(CG_BSP_0 + (i * 4), pi->dsp);
+
+	WREG32(CG_BSP_0 + (i * 4), pi->psp);
+
+	if (ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE)
+		WREG32(CG_BSP_0 + (BOOST_DPM_LEVEL * 4), pi->psp);
+}
+
+static void sumo_write_at(struct radeon_device *rdev,
+			  u32 index, u32 value)
+{
+	if (index == 0)
+		WREG32(CG_AT_0, value);
+	else if (index == 1)
+		WREG32(CG_AT_1, value);
+	else if (index == 2)
+		WREG32(CG_AT_2, value);
+	else if (index == 3)
+		WREG32(CG_AT_3, value);
+	else if (index == 4)
+		WREG32(CG_AT_4, value);
+	else if (index == 5)
+		WREG32(CG_AT_5, value);
+	else if (index == 6)
+		WREG32(CG_AT_6, value);
+	else if (index == 7)
+		WREG32(CG_AT_7, value);
+}
+
+static void sumo_program_at(struct radeon_device *rdev,
+			    struct radeon_ps *rps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *ps = sumo_get_ps(rps);
+	u32 asi;
+	u32 i;
+	u32 m_a;
+	u32 a_t;
+	u32 r[SUMO_MAX_HARDWARE_POWERLEVELS];
+	u32 l[SUMO_MAX_HARDWARE_POWERLEVELS];
+
+	r[0] = SUMO_R_DFLT0;
+	r[1] = SUMO_R_DFLT1;
+	r[2] = SUMO_R_DFLT2;
+	r[3] = SUMO_R_DFLT3;
+	r[4] = SUMO_R_DFLT4;
+
+	l[0] = SUMO_L_DFLT0;
+	l[1] = SUMO_L_DFLT1;
+	l[2] = SUMO_L_DFLT2;
+	l[3] = SUMO_L_DFLT3;
+	l[4] = SUMO_L_DFLT4;
+
+	for (i = 0; i < ps->num_levels; i++) {
+		asi = (i == ps->num_levels - 1) ? pi->pasi : pi->asi;
+
+		m_a = asi * ps->levels[i].sclk / 100;
+
+		a_t = CG_R(m_a * r[i] / 100) | CG_L(m_a * l[i] / 100);
+
+		sumo_write_at(rdev, i, a_t);
+	}
+
+	if (ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE) {
+		asi = pi->pasi;
+
+		m_a = asi * pi->boost_pl.sclk / 100;
+
+		a_t = CG_R(m_a * r[ps->num_levels - 1] / 100) |
+			CG_L(m_a * l[ps->num_levels - 1] / 100);
+
+		sumo_write_at(rdev, BOOST_DPM_LEVEL, a_t);
+	}
+}
+
+static void sumo_program_tp(struct radeon_device *rdev)
+{
+	int i;
+	enum r600_td td = R600_TD_DFLT;
+
+	for (i = 0; i < SUMO_PM_NUMBER_OF_TC; i++) {
+		WREG32_P(CG_FFCT_0 + (i * 4), UTC_0(sumo_utc[i]), ~UTC_0_MASK);
+		WREG32_P(CG_FFCT_0 + (i * 4), DTC_0(sumo_dtc[i]), ~DTC_0_MASK);
+	}
+
+	if (td == R600_TD_AUTO)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_FORCE_TREND_SEL);
+	else
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_FORCE_TREND_SEL, ~FIR_FORCE_TREND_SEL);
+
+	if (td == R600_TD_UP)
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_TREND_MODE);
+
+	if (td == R600_TD_DOWN)
+		WREG32_P(SCLK_PWRMGT_CNTL, FIR_TREND_MODE, ~FIR_TREND_MODE);
+}
+
+void sumo_program_vc(struct radeon_device *rdev, u32 vrc)
+{
+	WREG32(CG_FTV, vrc);
+}
+
+void sumo_clear_vc(struct radeon_device *rdev)
+{
+	WREG32(CG_FTV, 0);
+}
+
+void sumo_program_sstp(struct radeon_device *rdev)
+{
+	u32 p, u;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	r600_calculate_u_and_p(SUMO_SST_DFLT,
+			       xclk, 16, &p, &u);
+
+	WREG32(CG_SSP, SSTU(u) | SST(p));
+}
+
+static void sumo_set_divider_value(struct radeon_device *rdev,
+				   u32 index, u32 divider)
+{
+	u32 reg_index = index / 4;
+	u32 field_index = index % 4;
+
+	if (field_index == 0)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 SCLK_FSTATE_0_DIV(divider), ~SCLK_FSTATE_0_DIV_MASK);
+	else if (field_index == 1)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 SCLK_FSTATE_1_DIV(divider), ~SCLK_FSTATE_1_DIV_MASK);
+	else if (field_index == 2)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 SCLK_FSTATE_2_DIV(divider), ~SCLK_FSTATE_2_DIV_MASK);
+	else if (field_index == 3)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 SCLK_FSTATE_3_DIV(divider), ~SCLK_FSTATE_3_DIV_MASK);
+}
+
+static void sumo_set_ds_dividers(struct radeon_device *rdev,
+				 u32 index, u32 divider)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	if (pi->enable_sclk_ds) {
+		u32 dpm_ctrl = RREG32(CG_SCLK_DPM_CTRL_6);
+
+		dpm_ctrl &= ~(0x7 << (index * 3));
+		dpm_ctrl |= (divider << (index * 3));
+		WREG32(CG_SCLK_DPM_CTRL_6, dpm_ctrl);
+	}
+}
+
+static void sumo_set_ss_dividers(struct radeon_device *rdev,
+				 u32 index, u32 divider)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	if (pi->enable_sclk_ds) {
+		u32 dpm_ctrl = RREG32(CG_SCLK_DPM_CTRL_11);
+
+		dpm_ctrl &= ~(0x7 << (index * 3));
+		dpm_ctrl |= (divider << (index * 3));
+		WREG32(CG_SCLK_DPM_CTRL_11, dpm_ctrl);
+	}
+}
+
+static void sumo_set_vid(struct radeon_device *rdev, u32 index, u32 vid)
+{
+	u32 voltage_cntl = RREG32(CG_DPM_VOLTAGE_CNTL);
+
+	voltage_cntl &= ~(DPM_STATE0_LEVEL_MASK << (index * 2));
+	voltage_cntl |= (vid << (DPM_STATE0_LEVEL_SHIFT + index * 2));
+	WREG32(CG_DPM_VOLTAGE_CNTL, voltage_cntl);
+}
+
+static void sumo_set_allos_gnb_slow(struct radeon_device *rdev, u32 index, u32 gnb_slow)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 temp = gnb_slow;
+	u32 cg_sclk_dpm_ctrl_3;
+
+	if (pi->driver_nbps_policy_disable)
+		temp = 1;
+
+	cg_sclk_dpm_ctrl_3 = RREG32(CG_SCLK_DPM_CTRL_3);
+	cg_sclk_dpm_ctrl_3 &= ~(GNB_SLOW_FSTATE_0_MASK << index);
+	cg_sclk_dpm_ctrl_3 |= (temp << (GNB_SLOW_FSTATE_0_SHIFT + index));
+
+	WREG32(CG_SCLK_DPM_CTRL_3, cg_sclk_dpm_ctrl_3);
+}
+
+static void sumo_program_power_level(struct radeon_device *rdev,
+				     struct sumo_pl *pl, u32 index)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	int ret;
+	struct atom_clock_dividers dividers;
+	u32 ds_en = RREG32(DEEP_SLEEP_CNTL) & ENABLE_DS;
+
+	ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+					     pl->sclk, false, &dividers);
+	if (ret)
+		return;
+
+	sumo_set_divider_value(rdev, index, dividers.post_div);
+
+	sumo_set_vid(rdev, index, pl->vddc_index);
+
+	if (pl->ss_divider_index == 0 || pl->ds_divider_index == 0) {
+		if (ds_en)
+			WREG32_P(DEEP_SLEEP_CNTL, 0, ~ENABLE_DS);
+	} else {
+		sumo_set_ss_dividers(rdev, index, pl->ss_divider_index);
+		sumo_set_ds_dividers(rdev, index, pl->ds_divider_index);
+
+		if (!ds_en)
+			WREG32_P(DEEP_SLEEP_CNTL, ENABLE_DS, ~ENABLE_DS);
+	}
+
+	sumo_set_allos_gnb_slow(rdev, index, pl->allow_gnb_slow);
+
+	if (pi->enable_boost)
+		sumo_set_tdp_limit(rdev, index, pl->sclk_dpm_tdp_limit);
+}
+
+static void sumo_power_level_enable(struct radeon_device *rdev, u32 index, bool enable)
+{
+	u32 reg_index = index / 4;
+	u32 field_index = index % 4;
+
+	if (field_index == 0)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 enable ? SCLK_FSTATE_0_VLD : 0, ~SCLK_FSTATE_0_VLD);
+	else if (field_index == 1)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 enable ? SCLK_FSTATE_1_VLD : 0, ~SCLK_FSTATE_1_VLD);
+	else if (field_index == 2)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 enable ? SCLK_FSTATE_2_VLD : 0, ~SCLK_FSTATE_2_VLD);
+	else if (field_index == 3)
+		WREG32_P(CG_SCLK_DPM_CTRL + (reg_index * 4),
+			 enable ? SCLK_FSTATE_3_VLD : 0, ~SCLK_FSTATE_3_VLD);
+}
+
+static bool sumo_dpm_enabled(struct radeon_device *rdev)
+{
+	if (RREG32(CG_SCLK_DPM_CTRL_3) & DPM_SCLK_ENABLE)
+		return true;
+	else
+		return false;
+}
+
+static void sumo_start_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(CG_SCLK_DPM_CTRL_3, DPM_SCLK_ENABLE, ~DPM_SCLK_ENABLE);
+}
+
+static void sumo_stop_dpm(struct radeon_device *rdev)
+{
+	WREG32_P(CG_SCLK_DPM_CTRL_3, 0, ~DPM_SCLK_ENABLE);
+}
+
+static void sumo_set_forced_mode(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_P(CG_SCLK_DPM_CTRL_3, FORCE_SCLK_STATE_EN, ~FORCE_SCLK_STATE_EN);
+	else
+		WREG32_P(CG_SCLK_DPM_CTRL_3, 0, ~FORCE_SCLK_STATE_EN);
+}
+
+static void sumo_set_forced_mode_enabled(struct radeon_device *rdev)
+{
+	int i;
+
+	sumo_set_forced_mode(rdev, true);
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(CG_SCLK_STATUS) & SCLK_OVERCLK_DETECT)
+			break;
+		udelay(1);
+	}
+}
+
+static void sumo_wait_for_level_0(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if ((RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURR_SCLK_INDEX_MASK) == 0)
+			break;
+		udelay(1);
+	}
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if ((RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURR_INDEX_MASK) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+static void sumo_set_forced_mode_disabled(struct radeon_device *rdev)
+{
+	sumo_set_forced_mode(rdev, false);
+}
+
+static void sumo_enable_power_level_0(struct radeon_device *rdev)
+{
+	sumo_power_level_enable(rdev, 0, true);
+}
+
+static void sumo_patch_boost_state(struct radeon_device *rdev,
+				   struct radeon_ps *rps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *new_ps = sumo_get_ps(rps);
+
+	if (new_ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE) {
+		pi->boost_pl = new_ps->levels[new_ps->num_levels - 1];
+		pi->boost_pl.sclk = pi->sys_info.boost_sclk;
+		pi->boost_pl.vddc_index = pi->sys_info.boost_vid_2bit;
+		pi->boost_pl.sclk_dpm_tdp_limit = pi->sys_info.sclk_dpm_tdp_limit_boost;
+	}
+}
+
+static void sumo_pre_notify_alt_vddnb_change(struct radeon_device *rdev,
+					     struct radeon_ps *new_rps,
+					     struct radeon_ps *old_rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(new_rps);
+	struct sumo_ps *old_ps = sumo_get_ps(old_rps);
+	u32 nbps1_old = 0;
+	u32 nbps1_new = 0;
+
+	if (old_ps != NULL)
+		nbps1_old = (old_ps->flags & SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE) ? 1 : 0;
+
+	nbps1_new = (new_ps->flags & SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE) ? 1 : 0;
+
+	if (nbps1_old == 1 && nbps1_new == 0)
+		sumo_smu_notify_alt_vddnb_change(rdev, 0, 0);
+}
+
+static void sumo_post_notify_alt_vddnb_change(struct radeon_device *rdev,
+					      struct radeon_ps *new_rps,
+					      struct radeon_ps *old_rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(new_rps);
+	struct sumo_ps *old_ps = sumo_get_ps(old_rps);
+	u32 nbps1_old = 0;
+	u32 nbps1_new = 0;
+
+	if (old_ps != NULL)
+		nbps1_old = (old_ps->flags & SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE)? 1 : 0;
+
+	nbps1_new = (new_ps->flags & SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE)? 1 : 0;
+
+	if (nbps1_old == 0 && nbps1_new == 1)
+		sumo_smu_notify_alt_vddnb_change(rdev, 1, 1);
+}
+
+static void sumo_enable_boost(struct radeon_device *rdev,
+			      struct radeon_ps *rps,
+			      bool enable)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(rps);
+
+	if (enable) {
+		if (new_ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE)
+			sumo_boost_state_enable(rdev, true);
+	} else
+		sumo_boost_state_enable(rdev, false);
+}
+
+static void sumo_set_forced_level(struct radeon_device *rdev, u32 index)
+{
+	WREG32_P(CG_SCLK_DPM_CTRL_3, FORCE_SCLK_STATE(index), ~FORCE_SCLK_STATE_MASK);
+}
+
+static void sumo_set_forced_level_0(struct radeon_device *rdev)
+{
+	sumo_set_forced_level(rdev, 0);
+}
+
+static void sumo_program_wl(struct radeon_device *rdev,
+			    struct radeon_ps *rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(rps);
+	u32 dpm_ctrl4 = RREG32(CG_SCLK_DPM_CTRL_4);
+
+	dpm_ctrl4 &= 0xFFFFFF00;
+	dpm_ctrl4 |= (1 << (new_ps->num_levels - 1));
+
+	if (new_ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE)
+		dpm_ctrl4 |= (1 << BOOST_DPM_LEVEL);
+
+	WREG32(CG_SCLK_DPM_CTRL_4, dpm_ctrl4);
+}
+
+static void sumo_program_power_levels_0_to_n(struct radeon_device *rdev,
+					     struct radeon_ps *new_rps,
+					     struct radeon_ps *old_rps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *new_ps = sumo_get_ps(new_rps);
+	struct sumo_ps *old_ps = sumo_get_ps(old_rps);
+	u32 i;
+	u32 n_current_state_levels = (old_ps == NULL) ? 1 : old_ps->num_levels;
+
+	for (i = 0; i < new_ps->num_levels; i++) {
+		sumo_program_power_level(rdev, &new_ps->levels[i], i);
+		sumo_power_level_enable(rdev, i, true);
+	}
+
+	for (i = new_ps->num_levels; i < n_current_state_levels; i++)
+		sumo_power_level_enable(rdev, i, false);
+
+	if (new_ps->flags & SUMO_POWERSTATE_FLAGS_BOOST_STATE)
+		sumo_program_power_level(rdev, &pi->boost_pl, BOOST_DPM_LEVEL);
+}
+
+static void sumo_enable_acpi_pm(struct radeon_device *rdev)
+{
+	WREG32_P(GENERAL_PWRMGT, STATIC_PM_EN, ~STATIC_PM_EN);
+}
+
+static void sumo_program_power_level_enter_state(struct radeon_device *rdev)
+{
+	WREG32_P(CG_SCLK_DPM_CTRL_5, SCLK_FSTATE_BOOTUP(0), ~SCLK_FSTATE_BOOTUP_MASK);
+}
+
+static void sumo_program_acpi_power_level(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct atom_clock_dividers dividers;
+	int ret;
+
+        ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+                                             pi->acpi_pl.sclk,
+					     false, &dividers);
+	if (ret)
+		return;
+
+	WREG32_P(CG_ACPI_CNTL, SCLK_ACPI_DIV(dividers.post_div), ~SCLK_ACPI_DIV_MASK);
+	WREG32_P(CG_ACPI_VOLTAGE_CNTL, 0, ~ACPI_VOLTAGE_EN);
+}
+
+static void sumo_program_bootup_state(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 dpm_ctrl4 = RREG32(CG_SCLK_DPM_CTRL_4);
+	u32 i;
+
+	sumo_program_power_level(rdev, &pi->boot_pl, 0);
+
+	dpm_ctrl4 &= 0xFFFFFF00;
+	WREG32(CG_SCLK_DPM_CTRL_4, dpm_ctrl4);
+
+	for (i = 1; i < 8; i++)
+		sumo_power_level_enable(rdev, i, false);
+}
+
+static void sumo_setup_uvd_clocks(struct radeon_device *rdev,
+				  struct radeon_ps *new_rps,
+				  struct radeon_ps *old_rps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	if (pi->enable_gfx_power_gating) {
+		sumo_gfx_powergating_enable(rdev, false);
+	}
+
+	radeon_set_uvd_clocks(rdev, new_rps->vclk, new_rps->dclk);
+
+	if (pi->enable_gfx_power_gating) {
+		if (!pi->disable_gfx_power_gating_in_uvd ||
+		    !r600_is_uvd_state(new_rps->class, new_rps->class2))
+			sumo_gfx_powergating_enable(rdev, true);
+	}
+}
+
+static void sumo_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+						    struct radeon_ps *new_rps,
+						    struct radeon_ps *old_rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(new_rps);
+	struct sumo_ps *current_ps = sumo_get_ps(old_rps);
+
+	if ((new_rps->vclk == old_rps->vclk) &&
+	    (new_rps->dclk == old_rps->dclk))
+		return;
+
+	if (new_ps->levels[new_ps->num_levels - 1].sclk >=
+	    current_ps->levels[current_ps->num_levels - 1].sclk)
+		return;
+
+	sumo_setup_uvd_clocks(rdev, new_rps, old_rps);
+}
+
+static void sumo_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+						   struct radeon_ps *new_rps,
+						   struct radeon_ps *old_rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(new_rps);
+	struct sumo_ps *current_ps = sumo_get_ps(old_rps);
+
+	if ((new_rps->vclk == old_rps->vclk) &&
+	    (new_rps->dclk == old_rps->dclk))
+		return;
+
+	if (new_ps->levels[new_ps->num_levels - 1].sclk <
+	    current_ps->levels[current_ps->num_levels - 1].sclk)
+		return;
+
+	sumo_setup_uvd_clocks(rdev, new_rps, old_rps);
+}
+
+void sumo_take_smu_control(struct radeon_device *rdev, bool enable)
+{
+/* This bit selects who handles display phy powergating.
+ * Clear the bit to let atom handle it.
+ * Set it to let the driver handle it.
+ * For now we just let atom handle it.
+ */
+#if 0
+	u32 v = RREG32(DOUT_SCRATCH3);
+
+	if (enable)
+		v |= 0x4;
+	else
+		v &= 0xFFFFFFFB;
+
+	WREG32(DOUT_SCRATCH3, v);
+#endif
+}
+
+static void sumo_enable_sclk_ds(struct radeon_device *rdev, bool enable)
+{
+	if (enable) {
+		u32 deep_sleep_cntl = RREG32(DEEP_SLEEP_CNTL);
+		u32 deep_sleep_cntl2 = RREG32(DEEP_SLEEP_CNTL2);
+		u32 t = 1;
+
+		deep_sleep_cntl &= ~R_DIS;
+		deep_sleep_cntl &= ~HS_MASK;
+		deep_sleep_cntl |= HS(t > 4095 ? 4095 : t);
+
+		deep_sleep_cntl2 |= LB_UFP_EN;
+		deep_sleep_cntl2 &= INOUT_C_MASK;
+		deep_sleep_cntl2 |= INOUT_C(0xf);
+
+		WREG32(DEEP_SLEEP_CNTL2, deep_sleep_cntl2);
+		WREG32(DEEP_SLEEP_CNTL, deep_sleep_cntl);
+	} else
+		WREG32_P(DEEP_SLEEP_CNTL, 0, ~ENABLE_DS);
+}
+
+static void sumo_program_bootup_at(struct radeon_device *rdev)
+{
+	WREG32_P(CG_AT_0, CG_R(0xffff), ~CG_R_MASK);
+	WREG32_P(CG_AT_0, CG_L(0), ~CG_L_MASK);
+}
+
+static void sumo_reset_am(struct radeon_device *rdev)
+{
+	WREG32_P(SCLK_PWRMGT_CNTL, FIR_RESET, ~FIR_RESET);
+}
+
+static void sumo_start_am(struct radeon_device *rdev)
+{
+	WREG32_P(SCLK_PWRMGT_CNTL, 0, ~FIR_RESET);
+}
+
+static void sumo_program_ttp(struct radeon_device *rdev)
+{
+	u32 xclk = radeon_get_xclk(rdev);
+	u32 p, u;
+	u32 cg_sclk_dpm_ctrl_5 = RREG32(CG_SCLK_DPM_CTRL_5);
+
+	r600_calculate_u_and_p(1000,
+			       xclk, 16, &p, &u);
+
+	cg_sclk_dpm_ctrl_5 &= ~(TT_TP_MASK | TT_TU_MASK);
+	cg_sclk_dpm_ctrl_5 |= TT_TP(p) | TT_TU(u);
+
+	WREG32(CG_SCLK_DPM_CTRL_5, cg_sclk_dpm_ctrl_5);
+}
+
+static void sumo_program_ttt(struct radeon_device *rdev)
+{
+	u32 cg_sclk_dpm_ctrl_3 = RREG32(CG_SCLK_DPM_CTRL_3);
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	cg_sclk_dpm_ctrl_3 &= ~(GNB_TT_MASK | GNB_THERMTHRO_MASK);
+	cg_sclk_dpm_ctrl_3 |= GNB_TT(pi->thermal_auto_throttling + 49);
+
+	WREG32(CG_SCLK_DPM_CTRL_3, cg_sclk_dpm_ctrl_3);
+}
+
+
+static void sumo_enable_voltage_scaling(struct radeon_device *rdev, bool enable)
+{
+	if (enable) {
+		WREG32_P(CG_DPM_VOLTAGE_CNTL, DPM_VOLTAGE_EN, ~DPM_VOLTAGE_EN);
+		WREG32_P(CG_CG_VOLTAGE_CNTL, 0, ~CG_VOLTAGE_EN);
+	} else {
+		WREG32_P(CG_CG_VOLTAGE_CNTL, CG_VOLTAGE_EN, ~CG_VOLTAGE_EN);
+		WREG32_P(CG_DPM_VOLTAGE_CNTL, 0, ~DPM_VOLTAGE_EN);
+	}
+}
+
+static void sumo_override_cnb_thermal_events(struct radeon_device *rdev)
+{
+	WREG32_P(CG_SCLK_DPM_CTRL_3, CNB_THERMTHRO_MASK_SCLK,
+		 ~CNB_THERMTHRO_MASK_SCLK);
+}
+
+static void sumo_program_dc_hto(struct radeon_device *rdev)
+{
+	u32 cg_sclk_dpm_ctrl_4 = RREG32(CG_SCLK_DPM_CTRL_4);
+	u32 p, u;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	r600_calculate_u_and_p(100000,
+			       xclk, 14, &p, &u);
+
+	cg_sclk_dpm_ctrl_4 &= ~(DC_HDC_MASK | DC_HU_MASK);
+	cg_sclk_dpm_ctrl_4 |= DC_HDC(p) | DC_HU(u);
+
+	WREG32(CG_SCLK_DPM_CTRL_4, cg_sclk_dpm_ctrl_4);
+}
+
+static void sumo_force_nbp_state(struct radeon_device *rdev,
+				 struct radeon_ps *rps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *new_ps = sumo_get_ps(rps);
+
+	if (!pi->driver_nbps_policy_disable) {
+		if (new_ps->flags & SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE)
+			WREG32_P(CG_SCLK_DPM_CTRL_3, FORCE_NB_PSTATE_1, ~FORCE_NB_PSTATE_1);
+		else
+			WREG32_P(CG_SCLK_DPM_CTRL_3, 0, ~FORCE_NB_PSTATE_1);
+	}
+}
+
+u32 sumo_get_sleep_divider_from_id(u32 id)
+{
+	return 1 << id;
+}
+
+u32 sumo_get_sleep_divider_id_from_clock(struct radeon_device *rdev,
+					 u32 sclk,
+					 u32 min_sclk_in_sr)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 i;
+	u32 temp;
+	u32 min = (min_sclk_in_sr > SUMO_MINIMUM_ENGINE_CLOCK) ?
+		min_sclk_in_sr : SUMO_MINIMUM_ENGINE_CLOCK;
+
+	if (sclk < min)
+		return 0;
+
+	if (!pi->enable_sclk_ds)
+		return 0;
+
+	for (i = SUMO_MAX_DEEPSLEEP_DIVIDER_ID;  ; i--) {
+		temp = sclk / sumo_get_sleep_divider_from_id(i);
+
+		if (temp >= min || i == 0)
+			break;
+	}
+	return i;
+}
+
+static u32 sumo_get_valid_engine_clock(struct radeon_device *rdev,
+				       u32 lower_limit)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 i;
+
+	for (i = 0; i < pi->sys_info.sclk_voltage_mapping_table.num_max_dpm_entries; i++) {
+		if (pi->sys_info.sclk_voltage_mapping_table.entries[i].sclk_frequency >= lower_limit)
+			return pi->sys_info.sclk_voltage_mapping_table.entries[i].sclk_frequency;
+	}
+
+	return pi->sys_info.sclk_voltage_mapping_table.entries[pi->sys_info.sclk_voltage_mapping_table.num_max_dpm_entries - 1].sclk_frequency;
+}
+
+static void sumo_patch_thermal_state(struct radeon_device *rdev,
+				     struct sumo_ps *ps,
+				     struct sumo_ps *current_ps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 sclk_in_sr = pi->sys_info.min_sclk; /* ??? */
+	u32 current_vddc;
+	u32 current_sclk;
+	u32 current_index = 0;
+
+	if (current_ps) {
+		current_vddc = current_ps->levels[current_index].vddc_index;
+		current_sclk = current_ps->levels[current_index].sclk;
+	} else {
+		current_vddc = pi->boot_pl.vddc_index;
+		current_sclk = pi->boot_pl.sclk;
+	}
+
+	ps->levels[0].vddc_index = current_vddc;
+
+	if (ps->levels[0].sclk > current_sclk)
+		ps->levels[0].sclk = current_sclk;
+
+	ps->levels[0].ss_divider_index =
+		sumo_get_sleep_divider_id_from_clock(rdev, ps->levels[0].sclk, sclk_in_sr);
+
+	ps->levels[0].ds_divider_index =
+		sumo_get_sleep_divider_id_from_clock(rdev, ps->levels[0].sclk, SUMO_MINIMUM_ENGINE_CLOCK);
+
+	if (ps->levels[0].ds_divider_index > ps->levels[0].ss_divider_index + 1)
+		ps->levels[0].ds_divider_index = ps->levels[0].ss_divider_index + 1;
+
+	if (ps->levels[0].ss_divider_index == ps->levels[0].ds_divider_index) {
+		if (ps->levels[0].ss_divider_index > 1)
+			ps->levels[0].ss_divider_index = ps->levels[0].ss_divider_index - 1;
+	}
+
+	if (ps->levels[0].ss_divider_index == 0)
+		ps->levels[0].ds_divider_index = 0;
+
+	if (ps->levels[0].ds_divider_index == 0)
+		ps->levels[0].ss_divider_index = 0;
+}
+
+static void sumo_apply_state_adjust_rules(struct radeon_device *rdev,
+					  struct radeon_ps *new_rps,
+					  struct radeon_ps *old_rps)
+{
+	struct sumo_ps *ps = sumo_get_ps(new_rps);
+	struct sumo_ps *current_ps = sumo_get_ps(old_rps);
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 min_voltage = 0; /* ??? */
+	u32 min_sclk = pi->sys_info.min_sclk; /* XXX check against disp reqs */
+	u32 sclk_in_sr = pi->sys_info.min_sclk; /* ??? */
+	u32 i;
+
+	if (new_rps->class & ATOM_PPLIB_CLASSIFICATION_THERMAL)
+		return sumo_patch_thermal_state(rdev, ps, current_ps);
+
+	if (pi->enable_boost) {
+		if (new_rps->class & ATOM_PPLIB_CLASSIFICATION_UI_PERFORMANCE)
+			ps->flags |= SUMO_POWERSTATE_FLAGS_BOOST_STATE;
+	}
+
+	if ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_UI_BATTERY) ||
+	    (new_rps->class & ATOM_PPLIB_CLASSIFICATION_SDSTATE) ||
+	    (new_rps->class & ATOM_PPLIB_CLASSIFICATION_HDSTATE))
+		ps->flags |= SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE;
+
+	for (i = 0; i < ps->num_levels; i++) {
+		if (ps->levels[i].vddc_index < min_voltage)
+			ps->levels[i].vddc_index = min_voltage;
+
+		if (ps->levels[i].sclk < min_sclk)
+			ps->levels[i].sclk =
+				sumo_get_valid_engine_clock(rdev, min_sclk);
+
+		ps->levels[i].ss_divider_index =
+			sumo_get_sleep_divider_id_from_clock(rdev, ps->levels[i].sclk, sclk_in_sr);
+
+		ps->levels[i].ds_divider_index =
+			sumo_get_sleep_divider_id_from_clock(rdev, ps->levels[i].sclk, SUMO_MINIMUM_ENGINE_CLOCK);
+
+		if (ps->levels[i].ds_divider_index > ps->levels[i].ss_divider_index + 1)
+			ps->levels[i].ds_divider_index = ps->levels[i].ss_divider_index + 1;
+
+		if (ps->levels[i].ss_divider_index == ps->levels[i].ds_divider_index) {
+			if (ps->levels[i].ss_divider_index > 1)
+				ps->levels[i].ss_divider_index = ps->levels[i].ss_divider_index - 1;
+		}
+
+		if (ps->levels[i].ss_divider_index == 0)
+			ps->levels[i].ds_divider_index = 0;
+
+		if (ps->levels[i].ds_divider_index == 0)
+			ps->levels[i].ss_divider_index = 0;
+
+		if (ps->flags & SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE)
+			ps->levels[i].allow_gnb_slow = 1;
+		else if ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE) ||
+			 (new_rps->class2 & ATOM_PPLIB_CLASSIFICATION2_MVC))
+			ps->levels[i].allow_gnb_slow = 0;
+		else if (i == ps->num_levels - 1)
+			ps->levels[i].allow_gnb_slow = 0;
+		else
+			ps->levels[i].allow_gnb_slow = 1;
+	}
+}
+
+static void sumo_cleanup_asic(struct radeon_device *rdev)
+{
+	sumo_take_smu_control(rdev, false);
+}
+
+static int sumo_set_thermal_temperature_range(struct radeon_device *rdev,
+					      int min_temp, int max_temp)
+{
+	int low_temp = 0 * 1000;
+	int high_temp = 255 * 1000;
+
+	if (low_temp < min_temp)
+		low_temp = min_temp;
+	if (high_temp > max_temp)
+		high_temp = max_temp;
+	if (high_temp < low_temp) {
+		DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp);
+		return -EINVAL;
+	}
+
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTH(49 + (high_temp / 1000)), ~DIG_THERM_INTH_MASK);
+	WREG32_P(CG_THERMAL_INT, DIG_THERM_INTL(49 + (low_temp / 1000)), ~DIG_THERM_INTL_MASK);
+
+	rdev->pm.dpm.thermal.min_temp = low_temp;
+	rdev->pm.dpm.thermal.max_temp = high_temp;
+
+	return 0;
+}
+
+static void sumo_update_current_ps(struct radeon_device *rdev,
+				   struct radeon_ps *rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(rps);
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	pi->current_rps = *rps;
+	pi->current_ps = *new_ps;
+	pi->current_rps.ps_priv = &pi->current_ps;
+}
+
+static void sumo_update_requested_ps(struct radeon_device *rdev,
+				     struct radeon_ps *rps)
+{
+	struct sumo_ps *new_ps = sumo_get_ps(rps);
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	pi->requested_rps = *rps;
+	pi->requested_ps = *new_ps;
+	pi->requested_rps.ps_priv = &pi->requested_ps;
+}
+
+int sumo_dpm_enable(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	int ret;
+
+	if (sumo_dpm_enabled(rdev))
+		return -EINVAL;
+
+	ret = sumo_enable_clock_power_gating(rdev);
+	if (ret)
+		return ret;
+	sumo_program_bootup_state(rdev);
+	sumo_init_bsp(rdev);
+	sumo_reset_am(rdev);
+	sumo_program_tp(rdev);
+	sumo_program_bootup_at(rdev);
+	sumo_start_am(rdev);
+	if (pi->enable_auto_thermal_throttling) {
+		sumo_program_ttp(rdev);
+		sumo_program_ttt(rdev);
+	}
+	sumo_program_dc_hto(rdev);
+	sumo_program_power_level_enter_state(rdev);
+	sumo_enable_voltage_scaling(rdev, true);
+	sumo_program_sstp(rdev);
+	sumo_program_vc(rdev, SUMO_VRC_DFLT);
+	sumo_override_cnb_thermal_events(rdev);
+	sumo_start_dpm(rdev);
+	sumo_wait_for_level_0(rdev);
+	if (pi->enable_sclk_ds)
+		sumo_enable_sclk_ds(rdev, true);
+	if (pi->enable_boost)
+		sumo_enable_boost_timer(rdev);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		ret = sumo_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret)
+			return ret;
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+	}
+
+	sumo_update_current_ps(rdev, rdev->pm.dpm.boot_ps);
+
+	return 0;
+}
+
+void sumo_dpm_disable(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	if (!sumo_dpm_enabled(rdev))
+		return;
+	sumo_disable_clock_power_gating(rdev);
+	if (pi->enable_sclk_ds)
+		sumo_enable_sclk_ds(rdev, false);
+	sumo_clear_vc(rdev);
+	sumo_wait_for_level_0(rdev);
+	sumo_stop_dpm(rdev);
+	sumo_enable_voltage_scaling(rdev, false);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	sumo_update_current_ps(rdev, rdev->pm.dpm.boot_ps);
+}
+
+int sumo_dpm_pre_set_power_state(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct radeon_ps requested_ps = *rdev->pm.dpm.requested_ps;
+	struct radeon_ps *new_ps = &requested_ps;
+
+	sumo_update_requested_ps(rdev, new_ps);
+
+	if (pi->enable_dynamic_patch_ps)
+		sumo_apply_state_adjust_rules(rdev,
+					      &pi->requested_rps,
+					      &pi->current_rps);
+
+	return 0;
+}
+
+int sumo_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct radeon_ps *new_ps = &pi->requested_rps;
+	struct radeon_ps *old_ps = &pi->current_rps;
+
+	if (pi->enable_dpm)
+		sumo_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+	if (pi->enable_boost) {
+		sumo_enable_boost(rdev, new_ps, false);
+		sumo_patch_boost_state(rdev, new_ps);
+	}
+	if (pi->enable_dpm) {
+		sumo_pre_notify_alt_vddnb_change(rdev, new_ps, old_ps);
+		sumo_enable_power_level_0(rdev);
+		sumo_set_forced_level_0(rdev);
+		sumo_set_forced_mode_enabled(rdev);
+		sumo_wait_for_level_0(rdev);
+		sumo_program_power_levels_0_to_n(rdev, new_ps, old_ps);
+		sumo_program_wl(rdev, new_ps);
+		sumo_program_bsp(rdev, new_ps);
+		sumo_program_at(rdev, new_ps);
+		sumo_force_nbp_state(rdev, new_ps);
+		sumo_set_forced_mode_disabled(rdev);
+		sumo_set_forced_mode_enabled(rdev);
+		sumo_set_forced_mode_disabled(rdev);
+		sumo_post_notify_alt_vddnb_change(rdev, new_ps, old_ps);
+	}
+	if (pi->enable_boost)
+		sumo_enable_boost(rdev, new_ps, true);
+	if (pi->enable_dpm)
+		sumo_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+
+	rdev->pm.dpm.forced_level = RADEON_DPM_FORCED_LEVEL_AUTO;
+
+	return 0;
+}
+
+void sumo_dpm_post_set_power_state(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct radeon_ps *new_ps = &pi->requested_rps;
+
+	sumo_update_current_ps(rdev, new_ps);
+}
+
+void sumo_dpm_reset_asic(struct radeon_device *rdev)
+{
+	sumo_program_bootup_state(rdev);
+	sumo_enable_power_level_0(rdev);
+	sumo_set_forced_level_0(rdev);
+	sumo_set_forced_mode_enabled(rdev);
+	sumo_wait_for_level_0(rdev);
+	sumo_set_forced_mode_disabled(rdev);
+	sumo_set_forced_mode_enabled(rdev);
+	sumo_set_forced_mode_disabled(rdev);
+}
+
+void sumo_dpm_setup_asic(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	sumo_initialize_m3_arb(rdev);
+	pi->fw_version = sumo_get_running_fw_version(rdev);
+	DRM_INFO("Found smc ucode version: 0x%08x\n", pi->fw_version);
+	sumo_program_acpi_power_level(rdev);
+	sumo_enable_acpi_pm(rdev);
+	sumo_take_smu_control(rdev, true);
+}
+
+void sumo_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void sumo_patch_boot_state(struct radeon_device *rdev,
+				  struct sumo_ps *ps)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	ps->num_levels = 1;
+	ps->flags = 0;
+	ps->levels[0] = pi->boot_pl;
+}
+
+static void sumo_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					    struct radeon_ps *rps,
+					    struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info,
+					    u8 table_rev)
+{
+	struct sumo_ps *ps = sumo_get_ps(rps);
+
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) {
+		rps->vclk = le32_to_cpu(non_clock_info->ulVCLK);
+		rps->dclk = le32_to_cpu(non_clock_info->ulDCLK);
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		rdev->pm.dpm.boot_ps = rps;
+		sumo_patch_boot_state(rdev, ps);
+	}
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void sumo_parse_pplib_clock_info(struct radeon_device *rdev,
+					struct radeon_ps *rps, int index,
+					union pplib_clock_info *clock_info)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *ps = sumo_get_ps(rps);
+	struct sumo_pl *pl = &ps->levels[index];
+	u32 sclk;
+
+	sclk = le16_to_cpu(clock_info->sumo.usEngineClockLow);
+	sclk |= clock_info->sumo.ucEngineClockHigh << 16;
+	pl->sclk = sclk;
+	pl->vddc_index = clock_info->sumo.vddcIndex;
+	pl->sclk_dpm_tdp_limit = clock_info->sumo.tdpLimit;
+
+	ps->num_levels = index + 1;
+
+	if (pi->enable_sclk_ds) {
+		pl->ds_divider_index = 5;
+		pl->ss_divider_index = 4;
+	}
+}
+
+static int sumo_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i, j, k, non_clock_array_index, clock_array_index;
+	union pplib_clock_info *clock_info;
+	struct _StateArray *state_array;
+	struct _ClockInfoArray *clock_info_array;
+	struct _NonClockInfoArray *non_clock_info_array;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	u8 *power_state_offset;
+	struct sumo_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	state_array = (struct _StateArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usStateArrayOffset));
+	clock_info_array = (struct _ClockInfoArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset));
+	non_clock_info_array = (struct _NonClockInfoArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset));
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  state_array->ucNumEntries, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	power_state_offset = (u8 *)state_array->states;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+	for (i = 0; i < state_array->ucNumEntries; i++) {
+		power_state = (union pplib_power_state *)power_state_offset;
+		non_clock_array_index = power_state->v2.nonClockInfoIndex;
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			&non_clock_info_array->nonClockInfo[non_clock_array_index];
+		if (!rdev->pm.power_state[i].clock_info)
+			return -EINVAL;
+		ps = kzalloc(sizeof(struct sumo_ps), GFP_KERNEL);
+		if (ps == NULL) {
+			kfree(rdev->pm.dpm.ps);
+			return -ENOMEM;
+		}
+		rdev->pm.dpm.ps[i].ps_priv = ps;
+		k = 0;
+		for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) {
+			clock_array_index = power_state->v2.clockInfoIndex[j];
+			if (k >= SUMO_MAX_HARDWARE_POWERLEVELS)
+				break;
+			clock_info = (union pplib_clock_info *)
+				&clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize];
+			sumo_parse_pplib_clock_info(rdev,
+						    &rdev->pm.dpm.ps[i], k,
+						    clock_info);
+			k++;
+		}
+		sumo_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+						non_clock_info,
+						non_clock_info_array->ucEntrySize);
+		power_state_offset += 2 + power_state->v2.ucNumDPMLevels;
+	}
+	rdev->pm.dpm.num_ps = state_array->ucNumEntries;
+	return 0;
+}
+
+u32 sumo_convert_vid2_to_vid7(struct radeon_device *rdev,
+			      struct sumo_vid_mapping_table *vid_mapping_table,
+			      u32 vid_2bit)
+{
+	u32 i;
+
+	for (i = 0; i < vid_mapping_table->num_entries; i++) {
+		if (vid_mapping_table->entries[i].vid_2bit == vid_2bit)
+			return vid_mapping_table->entries[i].vid_7bit;
+	}
+
+	return vid_mapping_table->entries[vid_mapping_table->num_entries - 1].vid_7bit;
+}
+
+static u16 sumo_convert_voltage_index_to_value(struct radeon_device *rdev,
+					       u32 vid_2bit)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 vid_7bit = sumo_convert_vid2_to_vid7(rdev, &pi->sys_info.vid_mapping_table, vid_2bit);
+
+	if (vid_7bit > 0x7C)
+		return 0;
+
+	return (15500 - vid_7bit * 125 + 5) / 10;
+}
+
+static void sumo_construct_display_voltage_mapping_table(struct radeon_device *rdev,
+							 struct sumo_disp_clock_voltage_mapping_table *disp_clk_voltage_mapping_table,
+							 ATOM_CLK_VOLT_CAPABILITY *table)
+{
+	u32 i;
+
+	for (i = 0; i < SUMO_MAX_NUMBER_VOLTAGES; i++) {
+		if (table[i].ulMaximumSupportedCLK == 0)
+			break;
+
+		disp_clk_voltage_mapping_table->display_clock_frequency[i] =
+			table[i].ulMaximumSupportedCLK;
+	}
+
+	disp_clk_voltage_mapping_table->num_max_voltage_levels = i;
+
+	if (disp_clk_voltage_mapping_table->num_max_voltage_levels == 0) {
+		disp_clk_voltage_mapping_table->display_clock_frequency[0] = 80000;
+		disp_clk_voltage_mapping_table->num_max_voltage_levels = 1;
+	}
+}
+
+void sumo_construct_sclk_voltage_mapping_table(struct radeon_device *rdev,
+					       struct sumo_sclk_voltage_mapping_table *sclk_voltage_mapping_table,
+					       ATOM_AVAILABLE_SCLK_LIST *table)
+{
+	u32 i;
+	u32 n = 0;
+	u32 prev_sclk = 0;
+
+	for (i = 0; i < SUMO_MAX_HARDWARE_POWERLEVELS; i++) {
+		if (table[i].ulSupportedSCLK > prev_sclk) {
+			sclk_voltage_mapping_table->entries[n].sclk_frequency =
+				table[i].ulSupportedSCLK;
+			sclk_voltage_mapping_table->entries[n].vid_2bit =
+				table[i].usVoltageIndex;
+			prev_sclk = table[i].ulSupportedSCLK;
+			n++;
+		}
+	}
+
+	sclk_voltage_mapping_table->num_max_dpm_entries = n;
+}
+
+void sumo_construct_vid_mapping_table(struct radeon_device *rdev,
+				      struct sumo_vid_mapping_table *vid_mapping_table,
+				      ATOM_AVAILABLE_SCLK_LIST *table)
+{
+	u32 i, j;
+
+	for (i = 0; i < SUMO_MAX_HARDWARE_POWERLEVELS; i++) {
+		if (table[i].ulSupportedSCLK != 0) {
+			vid_mapping_table->entries[table[i].usVoltageIndex].vid_7bit =
+				table[i].usVoltageID;
+			vid_mapping_table->entries[table[i].usVoltageIndex].vid_2bit =
+				table[i].usVoltageIndex;
+		}
+	}
+
+	for (i = 0; i < SUMO_MAX_NUMBER_VOLTAGES; i++) {
+		if (vid_mapping_table->entries[i].vid_7bit == 0) {
+			for (j = i + 1; j < SUMO_MAX_NUMBER_VOLTAGES; j++) {
+				if (vid_mapping_table->entries[j].vid_7bit != 0) {
+					vid_mapping_table->entries[i] =
+						vid_mapping_table->entries[j];
+					vid_mapping_table->entries[j].vid_7bit = 0;
+					break;
+				}
+			}
+
+			if (j == SUMO_MAX_NUMBER_VOLTAGES)
+				break;
+		}
+	}
+
+	vid_mapping_table->num_entries = i;
+}
+
+union igp_info {
+	struct _ATOM_INTEGRATED_SYSTEM_INFO info;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V2 info_2;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V5 info_5;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V6 info_6;
+};
+
+static int sumo_parse_sys_info_table(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	int index = GetIndexIntoMasterTable(DATA, IntegratedSystemInfo);
+	union igp_info *igp_info;
+	u8 frev, crev;
+	u16 data_offset;
+	int i;
+
+	if (atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset)) {
+		igp_info = (union igp_info *)(mode_info->atom_context->bios +
+					      data_offset);
+
+		if (crev != 6) {
+			DRM_ERROR("Unsupported IGP table: %d %d\n", frev, crev);
+			return -EINVAL;
+		}
+		pi->sys_info.bootup_sclk = le32_to_cpu(igp_info->info_6.ulBootUpEngineClock);
+		pi->sys_info.min_sclk = le32_to_cpu(igp_info->info_6.ulMinEngineClock);
+		pi->sys_info.bootup_uma_clk = le32_to_cpu(igp_info->info_6.ulBootUpUMAClock);
+		pi->sys_info.bootup_nb_voltage_index =
+			le16_to_cpu(igp_info->info_6.usBootUpNBVoltage);
+		if (igp_info->info_6.ucHtcTmpLmt == 0)
+			pi->sys_info.htc_tmp_lmt = 203;
+		else
+			pi->sys_info.htc_tmp_lmt = igp_info->info_6.ucHtcTmpLmt;
+		if (igp_info->info_6.ucHtcHystLmt == 0)
+			pi->sys_info.htc_hyst_lmt = 5;
+		else
+			pi->sys_info.htc_hyst_lmt = igp_info->info_6.ucHtcHystLmt;
+		if (pi->sys_info.htc_tmp_lmt <= pi->sys_info.htc_hyst_lmt) {
+			DRM_ERROR("The htcTmpLmt should be larger than htcHystLmt.\n");
+		}
+		for (i = 0; i < NUMBER_OF_M3ARB_PARAM_SETS; i++) {
+			pi->sys_info.csr_m3_arb_cntl_default[i] =
+				le32_to_cpu(igp_info->info_6.ulCSR_M3_ARB_CNTL_DEFAULT[i]);
+			pi->sys_info.csr_m3_arb_cntl_uvd[i] =
+				le32_to_cpu(igp_info->info_6.ulCSR_M3_ARB_CNTL_UVD[i]);
+			pi->sys_info.csr_m3_arb_cntl_fs3d[i] =
+				le32_to_cpu(igp_info->info_6.ulCSR_M3_ARB_CNTL_FS3D[i]);
+		}
+		pi->sys_info.sclk_dpm_boost_margin =
+			le32_to_cpu(igp_info->info_6.SclkDpmBoostMargin);
+		pi->sys_info.sclk_dpm_throttle_margin =
+			le32_to_cpu(igp_info->info_6.SclkDpmThrottleMargin);
+		pi->sys_info.sclk_dpm_tdp_limit_pg =
+			le16_to_cpu(igp_info->info_6.SclkDpmTdpLimitPG);
+		pi->sys_info.gnb_tdp_limit = le16_to_cpu(igp_info->info_6.GnbTdpLimit);
+		pi->sys_info.sclk_dpm_tdp_limit_boost =
+			le16_to_cpu(igp_info->info_6.SclkDpmTdpLimitBoost);
+		pi->sys_info.boost_sclk = le32_to_cpu(igp_info->info_6.ulBoostEngineCLock);
+		pi->sys_info.boost_vid_2bit = igp_info->info_6.ulBoostVid_2bit;
+		if (igp_info->info_6.EnableBoost)
+			pi->sys_info.enable_boost = true;
+		else
+			pi->sys_info.enable_boost = false;
+		sumo_construct_display_voltage_mapping_table(rdev,
+							     &pi->sys_info.disp_clk_voltage_mapping_table,
+							     igp_info->info_6.sDISPCLK_Voltage);
+		sumo_construct_sclk_voltage_mapping_table(rdev,
+							  &pi->sys_info.sclk_voltage_mapping_table,
+							  igp_info->info_6.sAvail_SCLK);
+		sumo_construct_vid_mapping_table(rdev, &pi->sys_info.vid_mapping_table,
+						 igp_info->info_6.sAvail_SCLK);
+
+	}
+	return 0;
+}
+
+static void sumo_construct_boot_and_acpi_state(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	pi->boot_pl.sclk = pi->sys_info.bootup_sclk;
+	pi->boot_pl.vddc_index = pi->sys_info.bootup_nb_voltage_index;
+	pi->boot_pl.ds_divider_index = 0;
+	pi->boot_pl.ss_divider_index = 0;
+	pi->boot_pl.allow_gnb_slow = 1;
+	pi->acpi_pl = pi->boot_pl;
+	pi->current_ps.num_levels = 1;
+	pi->current_ps.levels[0] = pi->boot_pl;
+}
+
+int sumo_dpm_init(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi;
+	u32 hw_rev = (RREG32(HW_REV) & ATI_REV_ID_MASK) >> ATI_REV_ID_SHIFT;
+	int ret;
+
+	pi = kzalloc(sizeof(struct sumo_power_info), GFP_KERNEL);
+	if (pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = pi;
+
+	pi->driver_nbps_policy_disable = false;
+	if ((rdev->family == CHIP_PALM) && (hw_rev < 3))
+		pi->disable_gfx_power_gating_in_uvd = true;
+	else
+		pi->disable_gfx_power_gating_in_uvd = false;
+	pi->enable_alt_vddnb = true;
+	pi->enable_sclk_ds = true;
+	pi->enable_dynamic_m3_arbiter = false;
+	pi->enable_dynamic_patch_ps = true;
+	/* Some PALM chips don't seem to properly ungate gfx when UVD is in use;
+	 * for now just disable gfx PG.
+	 */
+	if (rdev->family == CHIP_PALM)
+		pi->enable_gfx_power_gating = false;
+	else
+		pi->enable_gfx_power_gating = true;
+	pi->enable_gfx_clock_gating = true;
+	pi->enable_mg_clock_gating = true;
+	pi->enable_auto_thermal_throttling = true;
+
+	ret = sumo_parse_sys_info_table(rdev);
+	if (ret)
+		return ret;
+
+	sumo_construct_boot_and_acpi_state(rdev);
+
+	ret = sumo_parse_power_table(rdev);
+	if (ret)
+		return ret;
+
+	pi->pasi = CYPRESS_HASI_DFLT;
+	pi->asi = RV770_ASI_DFLT;
+	pi->thermal_auto_throttling = pi->sys_info.htc_tmp_lmt;
+	pi->enable_boost = pi->sys_info.enable_boost;
+	pi->enable_dpm = true;
+
+	return 0;
+}
+
+void sumo_dpm_print_power_state(struct radeon_device *rdev,
+				struct radeon_ps *rps)
+{
+	int i;
+	struct sumo_ps *ps = sumo_get_ps(rps);
+
+	r600_dpm_print_class_info(rps->class, rps->class2);
+	r600_dpm_print_cap_info(rps->caps);
+	printk("\tuvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+	for (i = 0; i < ps->num_levels; i++) {
+		struct sumo_pl *pl = &ps->levels[i];
+		printk("\t\tpower level %d    sclk: %u vddc: %u\n",
+		       i, pl->sclk,
+		       sumo_convert_voltage_index_to_value(rdev, pl->vddc_index));
+	}
+	r600_dpm_print_ps_status(rdev, rps);
+}
+
+void sumo_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+						      struct seq_file *m)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct sumo_ps *ps = sumo_get_ps(rps);
+	struct sumo_pl *pl;
+	u32 current_index =
+		(RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURR_INDEX_MASK) >>
+		CURR_INDEX_SHIFT;
+
+	if (current_index == BOOST_DPM_LEVEL) {
+		pl = &pi->boost_pl;
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		seq_printf(m, "power level %d    sclk: %u vddc: %u\n",
+			   current_index, pl->sclk,
+			   sumo_convert_voltage_index_to_value(rdev, pl->vddc_index));
+	} else if (current_index >= ps->num_levels) {
+		seq_printf(m, "invalid dpm profile %d\n", current_index);
+	} else {
+		pl = &ps->levels[current_index];
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		seq_printf(m, "power level %d    sclk: %u vddc: %u\n",
+			   current_index, pl->sclk,
+			   sumo_convert_voltage_index_to_value(rdev, pl->vddc_index));
+	}
+}
+
+void sumo_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	sumo_cleanup_asic(rdev); /* ??? */
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+}
+
+u32 sumo_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct sumo_ps *requested_state = sumo_get_ps(&pi->requested_rps);
+
+	if (low)
+		return requested_state->levels[0].sclk;
+	else
+		return requested_state->levels[requested_state->num_levels - 1].sclk;
+}
+
+u32 sumo_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+
+	return pi->sys_info.bootup_uma_clk;
+}
+
+int sumo_dpm_force_performance_level(struct radeon_device *rdev,
+				     enum radeon_dpm_forced_level level)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	struct radeon_ps *rps = &pi->current_rps;
+	struct sumo_ps *ps = sumo_get_ps(rps);
+	int i;
+
+	if (ps->num_levels <= 1)
+		return 0;
+
+	if (level == RADEON_DPM_FORCED_LEVEL_HIGH) {
+		if (pi->enable_boost)
+			sumo_enable_boost(rdev, rps, false);
+		sumo_power_level_enable(rdev, ps->num_levels - 1, true);
+		sumo_set_forced_level(rdev, ps->num_levels - 1);
+		sumo_set_forced_mode_enabled(rdev);
+		for (i = 0; i < ps->num_levels - 1; i++) {
+			sumo_power_level_enable(rdev, i, false);
+		}
+		sumo_set_forced_mode(rdev, false);
+		sumo_set_forced_mode_enabled(rdev);
+		sumo_set_forced_mode(rdev, false);
+	} else if (level == RADEON_DPM_FORCED_LEVEL_LOW) {
+		if (pi->enable_boost)
+			sumo_enable_boost(rdev, rps, false);
+		sumo_power_level_enable(rdev, 0, true);
+		sumo_set_forced_level(rdev, 0);
+		sumo_set_forced_mode_enabled(rdev);
+		for (i = 1; i < ps->num_levels; i++) {
+			sumo_power_level_enable(rdev, i, false);
+		}
+		sumo_set_forced_mode(rdev, false);
+		sumo_set_forced_mode_enabled(rdev);
+		sumo_set_forced_mode(rdev, false);
+	} else {
+		for (i = 0; i < ps->num_levels; i++) {
+			sumo_power_level_enable(rdev, i, true);
+		}
+		if (pi->enable_boost)
+			sumo_enable_boost(rdev, rps, true);
+	}
+
+	rdev->pm.dpm.forced_level = level;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/radeon/sumo_dpm.h b/drivers/gpu/drm/radeon/sumo_dpm.h
new file mode 100644
index 000000000000..07dda299c784
--- /dev/null
+++ b/drivers/gpu/drm/radeon/sumo_dpm.h
@@ -0,0 +1,220 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __SUMO_DPM_H__
+#define __SUMO_DPM_H__
+
+#include "atom.h"
+
+#define SUMO_MAX_HARDWARE_POWERLEVELS 5
+#define SUMO_PM_NUMBER_OF_TC 15
+
+struct sumo_pl {
+	u32 sclk;
+	u32 vddc_index;
+	u32 ds_divider_index;
+	u32 ss_divider_index;
+	u32 allow_gnb_slow;
+	u32 sclk_dpm_tdp_limit;
+};
+
+/* used for the flags field */
+#define SUMO_POWERSTATE_FLAGS_FORCE_NBPS1_STATE (1 << 0)
+#define SUMO_POWERSTATE_FLAGS_BOOST_STATE       (1 << 1)
+
+struct sumo_ps {
+	struct sumo_pl levels[SUMO_MAX_HARDWARE_POWERLEVELS];
+	u32 num_levels;
+	/* flags */
+	u32 flags;
+};
+
+#define NUMBER_OF_M3ARB_PARAM_SETS 10
+#define SUMO_MAX_NUMBER_VOLTAGES    4
+
+struct sumo_disp_clock_voltage_mapping_table {
+	u32 num_max_voltage_levels;
+	u32 display_clock_frequency[SUMO_MAX_NUMBER_VOLTAGES];
+};
+
+struct sumo_vid_mapping_entry {
+	u16 vid_2bit;
+	u16 vid_7bit;
+};
+
+struct sumo_vid_mapping_table {
+	u32 num_entries;
+	struct sumo_vid_mapping_entry entries[SUMO_MAX_NUMBER_VOLTAGES];
+};
+
+struct sumo_sclk_voltage_mapping_entry {
+	u32 sclk_frequency;
+	u16 vid_2bit;
+	u16 rsv;
+};
+
+struct sumo_sclk_voltage_mapping_table {
+	u32 num_max_dpm_entries;
+	struct sumo_sclk_voltage_mapping_entry entries[SUMO_MAX_HARDWARE_POWERLEVELS];
+};
+
+struct sumo_sys_info {
+	u32 bootup_sclk;
+	u32 min_sclk;
+	u32 bootup_uma_clk;
+	u16 bootup_nb_voltage_index;
+	u8 htc_tmp_lmt;
+	u8 htc_hyst_lmt;
+	struct sumo_sclk_voltage_mapping_table sclk_voltage_mapping_table;
+	struct sumo_disp_clock_voltage_mapping_table disp_clk_voltage_mapping_table;
+	struct sumo_vid_mapping_table vid_mapping_table;
+	u32 csr_m3_arb_cntl_default[NUMBER_OF_M3ARB_PARAM_SETS];
+	u32 csr_m3_arb_cntl_uvd[NUMBER_OF_M3ARB_PARAM_SETS];
+	u32 csr_m3_arb_cntl_fs3d[NUMBER_OF_M3ARB_PARAM_SETS];
+	u32 sclk_dpm_boost_margin;
+	u32 sclk_dpm_throttle_margin;
+	u32 sclk_dpm_tdp_limit_pg;
+	u32 gnb_tdp_limit;
+	u32 sclk_dpm_tdp_limit_boost;
+	u32 boost_sclk;
+	u32 boost_vid_2bit;
+	bool enable_boost;
+};
+
+struct sumo_power_info {
+	u32 asi;
+	u32 pasi;
+	u32 bsp;
+	u32 bsu;
+	u32 pbsp;
+	u32 pbsu;
+	u32 dsp;
+	u32 psp;
+	u32 thermal_auto_throttling;
+	u32 uvd_m3_arbiter;
+	u32 fw_version;
+	struct sumo_sys_info sys_info;
+	struct sumo_pl acpi_pl;
+	struct sumo_pl boot_pl;
+	struct sumo_pl boost_pl;
+	bool disable_gfx_power_gating_in_uvd;
+	bool driver_nbps_policy_disable;
+	bool enable_alt_vddnb;
+	bool enable_dynamic_m3_arbiter;
+	bool enable_gfx_clock_gating;
+	bool enable_gfx_power_gating;
+	bool enable_mg_clock_gating;
+	bool enable_sclk_ds;
+	bool enable_auto_thermal_throttling;
+	bool enable_dynamic_patch_ps;
+	bool enable_dpm;
+	bool enable_boost;
+	struct radeon_ps current_rps;
+	struct sumo_ps current_ps;
+	struct radeon_ps requested_rps;
+	struct sumo_ps requested_ps;
+};
+
+#define SUMO_UTC_DFLT_00                     0x48
+#define SUMO_UTC_DFLT_01                     0x44
+#define SUMO_UTC_DFLT_02                     0x44
+#define SUMO_UTC_DFLT_03                     0x44
+#define SUMO_UTC_DFLT_04                     0x44
+#define SUMO_UTC_DFLT_05                     0x44
+#define SUMO_UTC_DFLT_06                     0x44
+#define SUMO_UTC_DFLT_07                     0x44
+#define SUMO_UTC_DFLT_08                     0x44
+#define SUMO_UTC_DFLT_09                     0x44
+#define SUMO_UTC_DFLT_10                     0x44
+#define SUMO_UTC_DFLT_11                     0x44
+#define SUMO_UTC_DFLT_12                     0x44
+#define SUMO_UTC_DFLT_13                     0x44
+#define SUMO_UTC_DFLT_14                     0x44
+
+#define SUMO_DTC_DFLT_00                     0x48
+#define SUMO_DTC_DFLT_01                     0x44
+#define SUMO_DTC_DFLT_02                     0x44
+#define SUMO_DTC_DFLT_03                     0x44
+#define SUMO_DTC_DFLT_04                     0x44
+#define SUMO_DTC_DFLT_05                     0x44
+#define SUMO_DTC_DFLT_06                     0x44
+#define SUMO_DTC_DFLT_07                     0x44
+#define SUMO_DTC_DFLT_08                     0x44
+#define SUMO_DTC_DFLT_09                     0x44
+#define SUMO_DTC_DFLT_10                     0x44
+#define SUMO_DTC_DFLT_11                     0x44
+#define SUMO_DTC_DFLT_12                     0x44
+#define SUMO_DTC_DFLT_13                     0x44
+#define SUMO_DTC_DFLT_14                     0x44
+
+#define SUMO_AH_DFLT               5
+
+#define SUMO_R_DFLT0               70
+#define SUMO_R_DFLT1               70
+#define SUMO_R_DFLT2               70
+#define SUMO_R_DFLT3               70
+#define SUMO_R_DFLT4               100
+
+#define SUMO_L_DFLT0               0
+#define SUMO_L_DFLT1               20
+#define SUMO_L_DFLT2               20
+#define SUMO_L_DFLT3               20
+#define SUMO_L_DFLT4               20
+#define SUMO_VRC_DFLT              0x30033
+#define SUMO_MGCGTTLOCAL0_DFLT     0
+#define SUMO_MGCGTTLOCAL1_DFLT     0
+#define SUMO_GICST_DFLT            19
+#define SUMO_SST_DFLT              8
+#define SUMO_VOLTAGEDROPT_DFLT     1
+#define SUMO_GFXPOWERGATINGT_DFLT  100
+
+/* sumo_dpm.c */
+void sumo_gfx_clockgating_initialize(struct radeon_device *rdev);
+void sumo_program_vc(struct radeon_device *rdev, u32 vrc);
+void sumo_clear_vc(struct radeon_device *rdev);
+void sumo_program_sstp(struct radeon_device *rdev);
+void sumo_take_smu_control(struct radeon_device *rdev, bool enable);
+void sumo_construct_sclk_voltage_mapping_table(struct radeon_device *rdev,
+					       struct sumo_sclk_voltage_mapping_table *sclk_voltage_mapping_table,
+					       ATOM_AVAILABLE_SCLK_LIST *table);
+void sumo_construct_vid_mapping_table(struct radeon_device *rdev,
+				      struct sumo_vid_mapping_table *vid_mapping_table,
+				      ATOM_AVAILABLE_SCLK_LIST *table);
+u32 sumo_convert_vid2_to_vid7(struct radeon_device *rdev,
+			      struct sumo_vid_mapping_table *vid_mapping_table,
+			      u32 vid_2bit);
+u32 sumo_get_sleep_divider_from_id(u32 id);
+u32 sumo_get_sleep_divider_id_from_clock(struct radeon_device *rdev,
+					 u32 sclk,
+					 u32 min_sclk_in_sr);
+
+/* sumo_smc.c */
+void sumo_initialize_m3_arb(struct radeon_device *rdev);
+void sumo_smu_pg_init(struct radeon_device *rdev);
+void sumo_set_tdp_limit(struct radeon_device *rdev, u32 index, u32 tdp_limit);
+void sumo_smu_notify_alt_vddnb_change(struct radeon_device *rdev,
+				      bool powersaving, bool force_nbps1);
+void sumo_boost_state_enable(struct radeon_device *rdev, bool enable);
+void sumo_enable_boost_timer(struct radeon_device *rdev);
+u32 sumo_get_running_fw_version(struct radeon_device *rdev);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/sumo_smc.c b/drivers/gpu/drm/radeon/sumo_smc.c
new file mode 100644
index 000000000000..18abba5b5810
--- /dev/null
+++ b/drivers/gpu/drm/radeon/sumo_smc.c
@@ -0,0 +1,222 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "sumod.h"
+#include "sumo_dpm.h"
+#include "ppsmc.h"
+
+#define SUMO_SMU_SERVICE_ROUTINE_PG_INIT        1
+#define SUMO_SMU_SERVICE_ROUTINE_ALTVDDNB_NOTIFY  27
+#define SUMO_SMU_SERVICE_ROUTINE_GFX_SRV_ID_20  20
+
+struct sumo_ps *sumo_get_ps(struct radeon_ps *rps);
+struct sumo_power_info *sumo_get_pi(struct radeon_device *rdev);
+
+static void sumo_send_msg_to_smu(struct radeon_device *rdev, u32 id)
+{
+	u32 gfx_int_req;
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(GFX_INT_STATUS) & INT_DONE)
+			break;
+		udelay(1);
+	}
+
+	gfx_int_req = SERV_INDEX(id) | INT_REQ;
+	WREG32(GFX_INT_REQ, gfx_int_req);
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(GFX_INT_REQ) & INT_REQ)
+			break;
+		udelay(1);
+	}
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(GFX_INT_STATUS) & INT_ACK)
+			break;
+		udelay(1);
+	}
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(GFX_INT_STATUS) & INT_DONE)
+			break;
+		udelay(1);
+	}
+
+	gfx_int_req &= ~INT_REQ;
+	WREG32(GFX_INT_REQ, gfx_int_req);
+}
+
+void sumo_initialize_m3_arb(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 i;
+
+	if (!pi->enable_dynamic_m3_arbiter)
+		return;
+
+	for (i = 0; i < NUMBER_OF_M3ARB_PARAM_SETS; i++)
+		WREG32_RCU(MCU_M3ARB_PARAMS + (i * 4),
+			   pi->sys_info.csr_m3_arb_cntl_default[i]);
+
+	for (; i < NUMBER_OF_M3ARB_PARAM_SETS * 2; i++)
+		WREG32_RCU(MCU_M3ARB_PARAMS + (i * 4),
+			   pi->sys_info.csr_m3_arb_cntl_uvd[i % NUMBER_OF_M3ARB_PARAM_SETS]);
+
+	for (; i < NUMBER_OF_M3ARB_PARAM_SETS * 3; i++)
+		WREG32_RCU(MCU_M3ARB_PARAMS + (i * 4),
+			   pi->sys_info.csr_m3_arb_cntl_fs3d[i % NUMBER_OF_M3ARB_PARAM_SETS]);
+}
+
+static bool sumo_is_alt_vddnb_supported(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	bool return_code = false;
+
+	if (!pi->enable_alt_vddnb)
+		return return_code;
+
+	if ((rdev->family == CHIP_SUMO) || (rdev->family == CHIP_SUMO2)) {
+		if (pi->fw_version >= 0x00010C00)
+			return_code = true;
+	}
+
+	return return_code;
+}
+
+void sumo_smu_notify_alt_vddnb_change(struct radeon_device *rdev,
+				      bool powersaving, bool force_nbps1)
+{
+	u32 param = 0;
+
+	if (!sumo_is_alt_vddnb_supported(rdev))
+		return;
+
+	if (powersaving)
+		param |= 1;
+
+	if (force_nbps1)
+		param |= 2;
+
+	WREG32_RCU(RCU_ALTVDDNB_NOTIFY, param);
+
+	sumo_send_msg_to_smu(rdev, SUMO_SMU_SERVICE_ROUTINE_ALTVDDNB_NOTIFY);
+}
+
+void sumo_smu_pg_init(struct radeon_device *rdev)
+{
+	sumo_send_msg_to_smu(rdev, SUMO_SMU_SERVICE_ROUTINE_PG_INIT);
+}
+
+static u32 sumo_power_of_4(u32 unit)
+{
+	u32 ret = 1;
+	u32 i;
+
+	for (i = 0; i < unit; i++)
+		ret *= 4;
+
+	return ret;
+}
+
+void sumo_enable_boost_timer(struct radeon_device *rdev)
+{
+	struct sumo_power_info *pi = sumo_get_pi(rdev);
+	u32 period, unit, timer_value;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	unit = (RREG32_RCU(RCU_LCLK_SCALING_CNTL) & LCLK_SCALING_TIMER_PRESCALER_MASK)
+		>> LCLK_SCALING_TIMER_PRESCALER_SHIFT;
+
+	period = 100 * (xclk / 100 / sumo_power_of_4(unit));
+
+	timer_value = (period << 16) | (unit << 4);
+
+	WREG32_RCU(RCU_GNB_PWR_REP_TIMER_CNTL, timer_value);
+	WREG32_RCU(RCU_BOOST_MARGIN, pi->sys_info.sclk_dpm_boost_margin);
+	WREG32_RCU(RCU_THROTTLE_MARGIN, pi->sys_info.sclk_dpm_throttle_margin);
+	WREG32_RCU(GNB_TDP_LIMIT, pi->sys_info.gnb_tdp_limit);
+	WREG32_RCU(RCU_SclkDpmTdpLimitPG, pi->sys_info.sclk_dpm_tdp_limit_pg);
+
+	sumo_send_msg_to_smu(rdev, SUMO_SMU_SERVICE_ROUTINE_GFX_SRV_ID_20);
+}
+
+void sumo_set_tdp_limit(struct radeon_device *rdev, u32 index, u32 tdp_limit)
+{
+	u32 regoffset = 0;
+	u32 shift = 0;
+	u32 mask = 0xFFF;
+	u32 sclk_dpm_tdp_limit;
+
+	switch (index) {
+	case 0:
+		regoffset = RCU_SclkDpmTdpLimit01;
+		shift = 16;
+		break;
+	case 1:
+		regoffset = RCU_SclkDpmTdpLimit01;
+		shift = 0;
+		break;
+	case 2:
+		regoffset = RCU_SclkDpmTdpLimit23;
+		shift = 16;
+		break;
+	case 3:
+		regoffset = RCU_SclkDpmTdpLimit23;
+		shift = 0;
+		break;
+	case 4:
+		regoffset = RCU_SclkDpmTdpLimit47;
+		shift = 16;
+		break;
+	case 7:
+		regoffset = RCU_SclkDpmTdpLimit47;
+		shift = 0;
+		break;
+	default:
+		break;
+	}
+
+	sclk_dpm_tdp_limit = RREG32_RCU(regoffset);
+	sclk_dpm_tdp_limit &= ~(mask << shift);
+	sclk_dpm_tdp_limit |= (tdp_limit << shift);
+	WREG32_RCU(regoffset, sclk_dpm_tdp_limit);
+}
+
+void sumo_boost_state_enable(struct radeon_device *rdev, bool enable)
+{
+	u32 boost_disable = RREG32_RCU(RCU_GPU_BOOST_DISABLE);
+
+	boost_disable &= 0xFFFFFFFE;
+	boost_disable |= (enable ? 0 : 1);
+	WREG32_RCU(RCU_GPU_BOOST_DISABLE, boost_disable);
+}
+
+u32 sumo_get_running_fw_version(struct radeon_device *rdev)
+{
+	return RREG32_RCU(RCU_FW_VERSION);
+}
+
diff --git a/drivers/gpu/drm/radeon/sumod.h b/drivers/gpu/drm/radeon/sumod.h
new file mode 100644
index 000000000000..7c9c2d4b86c0
--- /dev/null
+++ b/drivers/gpu/drm/radeon/sumod.h
@@ -0,0 +1,372 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+#ifndef _SUMOD_H_
+#define _SUMOD_H_
+
+/* pm registers */
+
+/* rcu */
+#define RCU_FW_VERSION                                  0x30c
+
+#define RCU_PWR_GATING_SEQ0                             0x408
+#define RCU_PWR_GATING_SEQ1                             0x40c
+#define RCU_PWR_GATING_CNTL                             0x410
+#       define PWR_GATING_EN                            (1 << 0)
+#       define RSVD_MASK                                (0x3 << 1)
+#       define PCV(x)                                   ((x) << 3)
+#       define PCV_MASK                                 (0x1f << 3)
+#       define PCV_SHIFT                                3
+#       define PCP(x)                                   ((x) << 8)
+#       define PCP_MASK                                 (0xf << 8)
+#       define PCP_SHIFT                                8
+#       define RPW(x)                                   ((x) << 16)
+#       define RPW_MASK                                 (0xf << 16)
+#       define RPW_SHIFT                                16
+#       define ID(x)                                    ((x) << 24)
+#       define ID_MASK                                  (0xf << 24)
+#       define ID_SHIFT                                 24
+#       define PGS(x)                                   ((x) << 28)
+#       define PGS_MASK                                 (0xf << 28)
+#       define PGS_SHIFT                                28
+
+#define RCU_ALTVDDNB_NOTIFY                             0x430
+#define RCU_LCLK_SCALING_CNTL                           0x434
+#       define LCLK_SCALING_EN                          (1 << 0)
+#       define LCLK_SCALING_TYPE                        (1 << 1)
+#       define LCLK_SCALING_TIMER_PRESCALER(x)          ((x) << 4)
+#       define LCLK_SCALING_TIMER_PRESCALER_MASK        (0xf << 4)
+#       define LCLK_SCALING_TIMER_PRESCALER_SHIFT       4
+#       define LCLK_SCALING_TIMER_PERIOD(x)             ((x) << 16)
+#       define LCLK_SCALING_TIMER_PERIOD_MASK           (0xf << 16)
+#       define LCLK_SCALING_TIMER_PERIOD_SHIFT          16
+
+#define RCU_PWR_GATING_CNTL_2                           0x4a0
+#       define MPPU(x)                                  ((x) << 0)
+#       define MPPU_MASK                                (0xffff << 0)
+#       define MPPU_SHIFT                               0
+#       define MPPD(x)                                  ((x) << 16)
+#       define MPPD_MASK                                (0xffff << 16)
+#       define MPPD_SHIFT                               16
+#define RCU_PWR_GATING_CNTL_3                           0x4a4
+#       define DPPU(x)                                  ((x) << 0)
+#       define DPPU_MASK                                (0xffff << 0)
+#       define DPPU_SHIFT                               0
+#       define DPPD(x)                                  ((x) << 16)
+#       define DPPD_MASK                                (0xffff << 16)
+#       define DPPD_SHIFT                               16
+#define RCU_PWR_GATING_CNTL_4                           0x4a8
+#       define RT(x)                                    ((x) << 0)
+#       define RT_MASK                                  (0xffff << 0)
+#       define RT_SHIFT                                 0
+#       define IT(x)                                    ((x) << 16)
+#       define IT_MASK                                  (0xffff << 16)
+#       define IT_SHIFT                                 16
+
+/* yes these two have the same address */
+#define RCU_PWR_GATING_CNTL_5                           0x504
+#define RCU_GPU_BOOST_DISABLE                           0x508
+
+#define MCU_M3ARB_INDEX                                 0x504
+#define MCU_M3ARB_PARAMS                                0x508
+
+#define RCU_GNB_PWR_REP_TIMER_CNTL                      0x50C
+
+#define RCU_SclkDpmTdpLimit01                           0x514
+#define RCU_SclkDpmTdpLimit23                           0x518
+#define RCU_SclkDpmTdpLimit47                           0x51C
+#define RCU_SclkDpmTdpLimitPG                           0x520
+
+#define GNB_TDP_LIMIT                                   0x540
+#define RCU_BOOST_MARGIN                                0x544
+#define RCU_THROTTLE_MARGIN                             0x548
+
+#define SMU_PCIE_PG_ARGS                                0x58C
+#define SMU_PCIE_PG_ARGS_2                              0x598
+#define SMU_PCIE_PG_ARGS_3                              0x59C
+
+/* mmio */
+#define RCU_STATUS                                      0x11c
+#       define GMC_PWR_GATER_BUSY                       (1 << 8)
+#       define GFX_PWR_GATER_BUSY                       (1 << 9)
+#       define UVD_PWR_GATER_BUSY                       (1 << 10)
+#       define PCIE_PWR_GATER_BUSY                      (1 << 11)
+#       define GMC_PWR_GATER_STATE                      (1 << 12)
+#       define GFX_PWR_GATER_STATE                      (1 << 13)
+#       define UVD_PWR_GATER_STATE                      (1 << 14)
+#       define PCIE_PWR_GATER_STATE                     (1 << 15)
+#       define GFX1_PWR_GATER_BUSY                      (1 << 16)
+#       define GFX2_PWR_GATER_BUSY                      (1 << 17)
+#       define GFX1_PWR_GATER_STATE                     (1 << 18)
+#       define GFX2_PWR_GATER_STATE                     (1 << 19)
+
+#define GFX_INT_REQ                                     0x120
+#       define INT_REQ                                  (1 << 0)
+#       define SERV_INDEX(x)                            ((x) << 1)
+#       define SERV_INDEX_MASK                          (0xff << 1)
+#       define SERV_INDEX_SHIFT                         1
+#define GFX_INT_STATUS                                  0x124
+#       define INT_ACK                                  (1 << 0)
+#       define INT_DONE                                 (1 << 1)
+
+#define CG_SCLK_CNTL                                    0x600
+#       define SCLK_DIVIDER(x)                          ((x) << 0)
+#       define SCLK_DIVIDER_MASK                        (0x7f << 0)
+#       define SCLK_DIVIDER_SHIFT                       0
+#define CG_SCLK_STATUS                                  0x604
+#       define SCLK_OVERCLK_DETECT                      (1 << 2)
+
+#define CG_DCLK_CNTL                                    0x610
+#       define DCLK_DIVIDER_MASK                        0x7f
+#       define DCLK_DIR_CNTL_EN                         (1 << 8)
+#define CG_DCLK_STATUS                                  0x614
+#       define DCLK_STATUS                              (1 << 0)
+#define CG_VCLK_CNTL                                    0x618
+#       define VCLK_DIVIDER_MASK                        0x7f
+#       define VCLK_DIR_CNTL_EN                         (1 << 8)
+#define CG_VCLK_STATUS                                  0x61c
+
+#define GENERAL_PWRMGT                                  0x63c
+#       define STATIC_PM_EN                             (1 << 1)
+
+#define SCLK_PWRMGT_CNTL                                0x644
+#       define SCLK_PWRMGT_OFF                          (1 << 0)
+#       define SCLK_LOW_D1                              (1 << 1)
+#       define FIR_RESET                                (1 << 4)
+#       define FIR_FORCE_TREND_SEL                      (1 << 5)
+#       define FIR_TREND_MODE                           (1 << 6)
+#       define DYN_GFX_CLK_OFF_EN                       (1 << 7)
+#       define GFX_CLK_FORCE_ON                         (1 << 8)
+#       define GFX_CLK_REQUEST_OFF                      (1 << 9)
+#       define GFX_CLK_FORCE_OFF                        (1 << 10)
+#       define GFX_CLK_OFF_ACPI_D1                      (1 << 11)
+#       define GFX_CLK_OFF_ACPI_D2                      (1 << 12)
+#       define GFX_CLK_OFF_ACPI_D3                      (1 << 13)
+#       define GFX_VOLTAGE_CHANGE_EN                    (1 << 16)
+#       define GFX_VOLTAGE_CHANGE_MODE                  (1 << 17)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                0x66c
+#       define TARG_SCLK_INDEX(x)                       ((x) << 6)
+#       define TARG_SCLK_INDEX_MASK                     (0x7 << 6)
+#       define TARG_SCLK_INDEX_SHIFT                    6
+#       define CURR_SCLK_INDEX(x)                       ((x) << 9)
+#       define CURR_SCLK_INDEX_MASK                     (0x7 << 9)
+#       define CURR_SCLK_INDEX_SHIFT                    9
+#       define TARG_INDEX(x)                            ((x) << 12)
+#       define TARG_INDEX_MASK                          (0x7 << 12)
+#       define TARG_INDEX_SHIFT                         12
+#       define CURR_INDEX(x)                            ((x) << 15)
+#       define CURR_INDEX_MASK                          (0x7 << 15)
+#       define CURR_INDEX_SHIFT                         15
+
+#define CG_SCLK_DPM_CTRL                                0x684
+#       define SCLK_FSTATE_0_DIV(x)                     ((x) << 0)
+#       define SCLK_FSTATE_0_DIV_MASK                   (0x7f << 0)
+#       define SCLK_FSTATE_0_DIV_SHIFT                  0
+#       define SCLK_FSTATE_0_VLD                        (1 << 7)
+#       define SCLK_FSTATE_1_DIV(x)                     ((x) << 8)
+#       define SCLK_FSTATE_1_DIV_MASK                   (0x7f << 8)
+#       define SCLK_FSTATE_1_DIV_SHIFT                  8
+#       define SCLK_FSTATE_1_VLD                        (1 << 15)
+#       define SCLK_FSTATE_2_DIV(x)                     ((x) << 16)
+#       define SCLK_FSTATE_2_DIV_MASK                   (0x7f << 16)
+#       define SCLK_FSTATE_2_DIV_SHIFT                  16
+#       define SCLK_FSTATE_2_VLD                        (1 << 23)
+#       define SCLK_FSTATE_3_DIV(x)                     ((x) << 24)
+#       define SCLK_FSTATE_3_DIV_MASK                   (0x7f << 24)
+#       define SCLK_FSTATE_3_DIV_SHIFT                  24
+#       define SCLK_FSTATE_3_VLD                        (1 << 31)
+#define CG_SCLK_DPM_CTRL_2                              0x688
+#define CG_GCOOR                                        0x68c
+#       define PHC(x)                                   ((x) << 0)
+#       define PHC_MASK                                 (0x1f << 0)
+#       define PHC_SHIFT                                0
+#       define SDC(x)                                   ((x) << 9)
+#       define SDC_MASK                                 (0x3ff << 9)
+#       define SDC_SHIFT                                9
+#       define SU(x)                                    ((x) << 23)
+#       define SU_MASK                                  (0xf << 23)
+#       define SU_SHIFT                                 23
+#       define DIV_ID(x)                                ((x) << 28)
+#       define DIV_ID_MASK                              (0x7 << 28)
+#       define DIV_ID_SHIFT                             28
+
+#define CG_FTV                                          0x690
+#define CG_FFCT_0                                       0x694
+#       define UTC_0(x)                                 ((x) << 0)
+#       define UTC_0_MASK                               (0x3ff << 0)
+#       define UTC_0_SHIFT                              0
+#       define DTC_0(x)                                 ((x) << 10)
+#       define DTC_0_MASK                               (0x3ff << 10)
+#       define DTC_0_SHIFT                              10
+
+#define CG_GIT                                          0x6d8
+#       define CG_GICST(x)                              ((x) << 0)
+#       define CG_GICST_MASK                            (0xffff << 0)
+#       define CG_GICST_SHIFT                           0
+#       define CG_GIPOT(x)                              ((x) << 16)
+#       define CG_GIPOT_MASK                            (0xffff << 16)
+#       define CG_GIPOT_SHIFT                           16
+
+#define CG_SCLK_DPM_CTRL_3                              0x6e0
+#       define FORCE_SCLK_STATE(x)                      ((x) << 0)
+#       define FORCE_SCLK_STATE_MASK                    (0x7 << 0)
+#       define FORCE_SCLK_STATE_SHIFT                   0
+#       define FORCE_SCLK_STATE_EN                      (1 << 3)
+#       define GNB_TT(x)                                ((x) << 8)
+#       define GNB_TT_MASK                              (0xff << 8)
+#       define GNB_TT_SHIFT                             8
+#       define GNB_THERMTHRO_MASK                       (1 << 16)
+#       define CNB_THERMTHRO_MASK_SCLK                  (1 << 17)
+#       define DPM_SCLK_ENABLE                          (1 << 18)
+#       define GNB_SLOW_FSTATE_0_MASK                   (1 << 23)
+#       define GNB_SLOW_FSTATE_0_SHIFT                  23
+#       define FORCE_NB_PSTATE_1                        (1 << 31)
+
+#define CG_SSP                                          0x6e8
+#       define SST(x)                                   ((x) << 0)
+#       define SST_MASK                                 (0xffff << 0)
+#       define SST_SHIFT                                0
+#       define SSTU(x)                                  ((x) << 16)
+#       define SSTU_MASK                                (0xffff << 16)
+#       define SSTU_SHIFT                               16
+
+#define CG_ACPI_CNTL                                    0x70c
+#       define SCLK_ACPI_DIV(x)                         ((x) << 0)
+#       define SCLK_ACPI_DIV_MASK                       (0x7f << 0)
+#       define SCLK_ACPI_DIV_SHIFT                      0
+
+#define CG_SCLK_DPM_CTRL_4                              0x71c
+#       define DC_HDC(x)                                ((x) << 14)
+#       define DC_HDC_MASK                              (0x3fff << 14)
+#       define DC_HDC_SHIFT                             14
+#       define DC_HU(x)                                 ((x) << 28)
+#       define DC_HU_MASK                               (0xf << 28)
+#       define DC_HU_SHIFT                              28
+#define CG_SCLK_DPM_CTRL_5                              0x720
+#       define SCLK_FSTATE_BOOTUP(x)                    ((x) << 0)
+#       define SCLK_FSTATE_BOOTUP_MASK                  (0x7 << 0)
+#       define SCLK_FSTATE_BOOTUP_SHIFT                 0
+#       define TT_TP(x)                                 ((x) << 3)
+#       define TT_TP_MASK                               (0xffff << 3)
+#       define TT_TP_SHIFT                              3
+#       define TT_TU(x)                                 ((x) << 19)
+#       define TT_TU_MASK                               (0xff << 19)
+#       define TT_TU_SHIFT                              19
+#define CG_SCLK_DPM_CTRL_6                              0x724
+#define CG_AT_0                                         0x728
+#       define CG_R(x)                                  ((x) << 0)
+#       define CG_R_MASK                                (0xffff << 0)
+#       define CG_R_SHIFT                               0
+#       define CG_L(x)                                  ((x) << 16)
+#       define CG_L_MASK                                (0xffff << 16)
+#       define CG_L_SHIFT                               16
+#define CG_AT_1                                         0x72c
+#define CG_AT_2                                         0x730
+#define	CG_THERMAL_INT					0x734
+#define		DIG_THERM_INTH(x)			((x) << 8)
+#define		DIG_THERM_INTH_MASK			0x0000FF00
+#define		DIG_THERM_INTH_SHIFT			8
+#define		DIG_THERM_INTL(x)			((x) << 16)
+#define		DIG_THERM_INTL_MASK			0x00FF0000
+#define		DIG_THERM_INTL_SHIFT			16
+#define 	THERM_INT_MASK_HIGH			(1 << 24)
+#define 	THERM_INT_MASK_LOW			(1 << 25)
+#define CG_AT_3                                         0x738
+#define CG_AT_4                                         0x73c
+#define CG_AT_5                                         0x740
+#define CG_AT_6                                         0x744
+#define CG_AT_7                                         0x748
+
+#define CG_BSP_0                                        0x750
+#       define BSP(x)                                   ((x) << 0)
+#       define BSP_MASK                                 (0xffff << 0)
+#       define BSP_SHIFT                                0
+#       define BSU(x)                                   ((x) << 16)
+#       define BSU_MASK                                 (0xf << 16)
+#       define BSU_SHIFT                                16
+
+#define CG_CG_VOLTAGE_CNTL                              0x770
+#       define REQ                                      (1 << 0)
+#       define LEVEL(x)                                 ((x) << 1)
+#       define LEVEL_MASK                               (0x3 << 1)
+#       define LEVEL_SHIFT                              1
+#       define CG_VOLTAGE_EN                            (1 << 3)
+#       define FORCE                                    (1 << 4)
+#       define PERIOD(x)                                ((x) << 8)
+#       define PERIOD_MASK                              (0xffff << 8)
+#       define PERIOD_SHIFT                             8
+#       define UNIT(x)                                  ((x) << 24)
+#       define UNIT_MASK                                (0xf << 24)
+#       define UNIT_SHIFT                               24
+
+#define CG_ACPI_VOLTAGE_CNTL                            0x780
+#       define ACPI_VOLTAGE_EN                          (1 << 8)
+
+#define CG_DPM_VOLTAGE_CNTL                             0x788
+#       define DPM_STATE0_LEVEL_MASK                    (0x3 << 0)
+#       define DPM_STATE0_LEVEL_SHIFT                   0
+#       define DPM_VOLTAGE_EN                           (1 << 16)
+
+#define CG_PWR_GATING_CNTL                              0x7ac
+#       define DYN_PWR_DOWN_EN                          (1 << 0)
+#       define ACPI_PWR_DOWN_EN                         (1 << 1)
+#       define GFX_CLK_OFF_PWR_DOWN_EN                  (1 << 2)
+#       define IOC_DISGPU_PWR_DOWN_EN                   (1 << 3)
+#       define FORCE_POWR_ON                            (1 << 4)
+#       define PGP(x)                                   ((x) << 8)
+#       define PGP_MASK                                 (0xffff << 8)
+#       define PGP_SHIFT                                8
+#       define PGU(x)                                   ((x) << 24)
+#       define PGU_MASK                                 (0xf << 24)
+#       define PGU_SHIFT                                24
+
+#define CG_CGTT_LOCAL_0                                 0x7d0
+#define CG_CGTT_LOCAL_1                                 0x7d4
+
+#define DEEP_SLEEP_CNTL                                 0x818
+#       define R_DIS                                    (1 << 3)
+#       define HS(x)                                    ((x) << 4)
+#       define HS_MASK                                  (0xfff << 4)
+#       define HS_SHIFT                                 4
+#       define ENABLE_DS                                (1 << 31)
+#define DEEP_SLEEP_CNTL2                                0x81c
+#       define LB_UFP_EN                                (1 << 0)
+#       define INOUT_C(x)                               ((x) << 4)
+#       define INOUT_C_MASK                             (0xff << 4)
+#       define INOUT_C_SHIFT                            4
+
+#define CG_SCRATCH2                                     0x824
+
+#define CG_SCLK_DPM_CTRL_11                             0x830
+
+#define HW_REV   					0x5564
+#       define ATI_REV_ID_MASK                          (0xf << 28)
+#       define ATI_REV_ID_SHIFT                         28
+/* 0 = A0, 1 = A1, 2 = B0, 3 = C0, etc. */
+
+#define DOUT_SCRATCH3   				0x611c
+
+#define GB_ADDR_CONFIG  				0x98f8
+
+#endif
diff --git a/drivers/gpu/drm/radeon/trinity_dpm.c b/drivers/gpu/drm/radeon/trinity_dpm.c
new file mode 100644
index 000000000000..a1eb5f59939f
--- /dev/null
+++ b/drivers/gpu/drm/radeon/trinity_dpm.c
@@ -0,0 +1,1949 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "trinityd.h"
+#include "r600_dpm.h"
+#include "trinity_dpm.h"
+#include <linux/seq_file.h>
+
+#define TRINITY_MAX_DEEPSLEEP_DIVIDER_ID 5
+#define TRINITY_MINIMUM_ENGINE_CLOCK 800
+#define SCLK_MIN_DIV_INTV_SHIFT     12
+#define TRINITY_DISPCLK_BYPASS_THRESHOLD 10000
+
+#ifndef TRINITY_MGCG_SEQUENCE
+#define TRINITY_MGCG_SEQUENCE  100
+
+static const u32 trinity_mgcg_shls_default[] =
+{
+	/* Register, Value, Mask */
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00003fc4, 0xc0000000, 0xffffffff,
+	0x00005448, 0x00000100, 0xffffffff,
+	0x000055e4, 0x00000100, 0xffffffff,
+	0x0000160c, 0x00000100, 0xffffffff,
+	0x00008984, 0x06000100, 0xffffffff,
+	0x0000c164, 0x00000100, 0xffffffff,
+	0x00008a18, 0x00000100, 0xffffffff,
+	0x0000897c, 0x06000100, 0xffffffff,
+	0x00008b28, 0x00000100, 0xffffffff,
+	0x00009144, 0x00800200, 0xffffffff,
+	0x00009a60, 0x00000100, 0xffffffff,
+	0x00009868, 0x00000100, 0xffffffff,
+	0x00008d58, 0x00000100, 0xffffffff,
+	0x00009510, 0x00000100, 0xffffffff,
+	0x0000949c, 0x00000100, 0xffffffff,
+	0x00009654, 0x00000100, 0xffffffff,
+	0x00009030, 0x00000100, 0xffffffff,
+	0x00009034, 0x00000100, 0xffffffff,
+	0x00009038, 0x00000100, 0xffffffff,
+	0x0000903c, 0x00000100, 0xffffffff,
+	0x00009040, 0x00000100, 0xffffffff,
+	0x0000a200, 0x00000100, 0xffffffff,
+	0x0000a204, 0x00000100, 0xffffffff,
+	0x0000a208, 0x00000100, 0xffffffff,
+	0x0000a20c, 0x00000100, 0xffffffff,
+	0x00009744, 0x00000100, 0xffffffff,
+	0x00003f80, 0x00000100, 0xffffffff,
+	0x0000a210, 0x00000100, 0xffffffff,
+	0x0000a214, 0x00000100, 0xffffffff,
+	0x000004d8, 0x00000100, 0xffffffff,
+	0x00009664, 0x00000100, 0xffffffff,
+	0x00009698, 0x00000100, 0xffffffff,
+	0x000004d4, 0x00000200, 0xffffffff,
+	0x000004d0, 0x00000000, 0xffffffff,
+	0x000030cc, 0x00000104, 0xffffffff,
+	0x0000d0c0, 0x00000100, 0xffffffff,
+	0x0000d8c0, 0x00000100, 0xffffffff,
+	0x0000951c, 0x00010000, 0xffffffff,
+	0x00009160, 0x00030002, 0xffffffff,
+	0x00009164, 0x00050004, 0xffffffff,
+	0x00009168, 0x00070006, 0xffffffff,
+	0x00009178, 0x00070000, 0xffffffff,
+	0x0000917c, 0x00030002, 0xffffffff,
+	0x00009180, 0x00050004, 0xffffffff,
+	0x0000918c, 0x00010006, 0xffffffff,
+	0x00009190, 0x00090008, 0xffffffff,
+	0x00009194, 0x00070000, 0xffffffff,
+	0x00009198, 0x00030002, 0xffffffff,
+	0x0000919c, 0x00050004, 0xffffffff,
+	0x000091a8, 0x00010006, 0xffffffff,
+	0x000091ac, 0x00090008, 0xffffffff,
+	0x000091b0, 0x00070000, 0xffffffff,
+	0x000091b4, 0x00030002, 0xffffffff,
+	0x000091b8, 0x00050004, 0xffffffff,
+	0x000091c4, 0x00010006, 0xffffffff,
+	0x000091c8, 0x00090008, 0xffffffff,
+	0x000091cc, 0x00070000, 0xffffffff,
+	0x000091d0, 0x00030002, 0xffffffff,
+	0x000091d4, 0x00050004, 0xffffffff,
+	0x000091e0, 0x00010006, 0xffffffff,
+	0x000091e4, 0x00090008, 0xffffffff,
+	0x000091e8, 0x00000000, 0xffffffff,
+	0x000091ec, 0x00070000, 0xffffffff,
+	0x000091f0, 0x00030002, 0xffffffff,
+	0x000091f4, 0x00050004, 0xffffffff,
+	0x00009200, 0x00010006, 0xffffffff,
+	0x00009204, 0x00090008, 0xffffffff,
+	0x00009208, 0x00070000, 0xffffffff,
+	0x0000920c, 0x00030002, 0xffffffff,
+	0x00009210, 0x00050004, 0xffffffff,
+	0x0000921c, 0x00010006, 0xffffffff,
+	0x00009220, 0x00090008, 0xffffffff,
+	0x00009294, 0x00000000, 0xffffffff
+};
+
+static const u32 trinity_mgcg_shls_enable[] =
+{
+	/* Register, Value, Mask */
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0x00000000, 0x000133FF,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0x00000000, 0xE00B03FC,
+	0x00009150, 0x96944200, 0xffffffff
+};
+
+static const u32 trinity_mgcg_shls_disable[] =
+{
+	/* Register, Value, Mask */
+	0x0000802c, 0xc0000000, 0xffffffff,
+	0x00009150, 0x00600000, 0xffffffff,
+	0x000008f8, 0x00000000, 0xffffffff,
+	0x000008fc, 0xffffffff, 0x000133FF,
+	0x000008f8, 0x00000001, 0xffffffff,
+	0x000008fc, 0xffffffff, 0xE00B03FC
+};
+#endif
+
+#ifndef TRINITY_SYSLS_SEQUENCE
+#define TRINITY_SYSLS_SEQUENCE  100
+
+static const u32 trinity_sysls_default[] =
+{
+	/* Register, Value, Mask */
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x0000d8bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x00002f50, 0x00000404, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x0000641c, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+
+static const u32 trinity_sysls_disable[] =
+{
+	/* Register, Value, Mask */
+	0x0000d0c0, 0x00000000, 0xffffffff,
+	0x0000d8c0, 0x00000000, 0xffffffff,
+	0x000055e8, 0x00000000, 0xffffffff,
+	0x0000d0bc, 0x00000000, 0xffffffff,
+	0x0000d8bc, 0x00000000, 0xffffffff,
+	0x000015c0, 0x00041401, 0xffffffff,
+	0x0000264c, 0x00040400, 0xffffffff,
+	0x00002648, 0x00040400, 0xffffffff,
+	0x00002650, 0x00040400, 0xffffffff,
+	0x000020b8, 0x00040400, 0xffffffff,
+	0x000020bc, 0x00040400, 0xffffffff,
+	0x000020c0, 0x00040c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680000, 0xffffffff,
+	0x00002f50, 0x00000404, 0xffffffff,
+	0x000004c8, 0x00000001, 0xffffffff,
+	0x0000641c, 0x00007ffd, 0xffffffff,
+	0x00000c7c, 0x0000ff00, 0xffffffff,
+	0x00006dfc, 0x0000007f, 0xffffffff
+};
+
+static const u32 trinity_sysls_enable[] =
+{
+	/* Register, Value, Mask */
+	0x000055e8, 0x00000001, 0xffffffff,
+	0x0000d0bc, 0x00000100, 0xffffffff,
+	0x0000d8bc, 0x00000100, 0xffffffff,
+	0x000015c0, 0x000c1401, 0xffffffff,
+	0x0000264c, 0x000c0400, 0xffffffff,
+	0x00002648, 0x000c0400, 0xffffffff,
+	0x00002650, 0x000c0400, 0xffffffff,
+	0x000020b8, 0x000c0400, 0xffffffff,
+	0x000020bc, 0x000c0400, 0xffffffff,
+	0x000020c0, 0x000c0c80, 0xffffffff,
+	0x0000f4a0, 0x000000c0, 0xffffffff,
+	0x0000f4a4, 0x00680fff, 0xffffffff,
+	0x00002f50, 0x00000903, 0xffffffff,
+	0x000004c8, 0x00000000, 0xffffffff,
+	0x0000641c, 0x00000000, 0xffffffff,
+	0x00000c7c, 0x00000000, 0xffffffff,
+	0x00006dfc, 0x00000000, 0xffffffff
+};
+#endif
+
+static const u32 trinity_override_mgpg_sequences[] =
+{
+	/* Register, Value */
+	0x00000200, 0xE030032C,
+	0x00000204, 0x00000FFF,
+	0x00000200, 0xE0300058,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE0300054,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE0300074,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE0300070,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE0300090,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE030008C,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE03000AC,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE03000A8,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE03000C8,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE03000C4,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE03000E4,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE03000E0,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE0300100,
+	0x00000204, 0x00030301,
+	0x00000200, 0xE03000FC,
+	0x00000204, 0x500010FF,
+	0x00000200, 0xE0300058,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE0300054,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE0300074,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE0300070,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE0300090,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE030008C,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE03000AC,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000A8,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE03000C8,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000C4,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE03000E4,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000E0,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE0300100,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000FC,
+	0x00000204, 0x600010FF,
+	0x00000200, 0xE0300058,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE0300054,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE0300074,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE0300070,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE0300090,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE030008C,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE03000AC,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000A8,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE03000C8,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000C4,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE03000E4,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000E0,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE0300100,
+	0x00000204, 0x00030303,
+	0x00000200, 0xE03000FC,
+	0x00000204, 0x700010FF,
+	0x00000200, 0xE0300058,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE0300054,
+	0x00000204, 0x800010FF,
+	0x00000200, 0xE0300074,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE0300070,
+	0x00000204, 0x800010FF,
+	0x00000200, 0xE0300090,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE030008C,
+	0x00000204, 0x800010FF,
+	0x00000200, 0xE03000AC,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE03000A8,
+	0x00000204, 0x800010FF,
+	0x00000200, 0xE03000C4,
+	0x00000204, 0x800010FF,
+	0x00000200, 0xE03000C8,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE03000E4,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE03000E0,
+	0x00000204, 0x800010FF,
+	0x00000200, 0xE0300100,
+	0x00000204, 0x00010303,
+	0x00000200, 0xE03000FC,
+	0x00000204, 0x800010FF,
+	0x00000200, 0x0001f198,
+	0x00000204, 0x0003ffff,
+	0x00000200, 0x0001f19C,
+	0x00000204, 0x3fffffff,
+	0x00000200, 0xE030032C,
+	0x00000204, 0x00000000,
+};
+
+static void trinity_program_clk_gating_hw_sequence(struct radeon_device *rdev,
+						   const u32 *seq, u32 count);
+static void trinity_override_dynamic_mg_powergating(struct radeon_device *rdev);
+static void trinity_apply_state_adjust_rules(struct radeon_device *rdev,
+					     struct radeon_ps *new_rps,
+					     struct radeon_ps *old_rps);
+
+struct trinity_ps *trinity_get_ps(struct radeon_ps *rps)
+{
+	struct trinity_ps *ps = rps->ps_priv;
+
+	return ps;
+}
+
+struct trinity_power_info *trinity_get_pi(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = rdev->pm.dpm.priv;
+
+	return pi;
+}
+
+static void trinity_gfx_powergating_initialize(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 p, u;
+	u32 value;
+	struct atom_clock_dividers dividers;
+	u32 xclk = radeon_get_xclk(rdev);
+	u32 sssd = 1;
+	int ret;
+	u32 hw_rev = (RREG32(HW_REV) & ATI_REV_ID_MASK) >> ATI_REV_ID_SHIFT;
+
+        ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+                                             25000, false, &dividers);
+	if (ret)
+		return;
+
+	value = RREG32_SMC(GFX_POWER_GATING_CNTL);
+	value &= ~(SSSD_MASK | PDS_DIV_MASK);
+	if (sssd)
+		value |= SSSD(1);
+	value |= PDS_DIV(dividers.post_div);
+	WREG32_SMC(GFX_POWER_GATING_CNTL, value);
+
+	r600_calculate_u_and_p(500, xclk, 16, &p, &u);
+
+	WREG32(CG_PG_CTRL, SP(p) | SU(u));
+
+	WREG32_P(CG_GIPOTS, CG_GIPOT(p), ~CG_GIPOT_MASK);
+
+	/* XXX double check hw_rev */
+	if (pi->override_dynamic_mgpg && (hw_rev == 0))
+		trinity_override_dynamic_mg_powergating(rdev);
+
+}
+
+#define CGCG_CGTT_LOCAL0_MASK       0xFFFF33FF
+#define CGCG_CGTT_LOCAL1_MASK       0xFFFB0FFE
+#define CGTS_SM_CTRL_REG_DISABLE    0x00600000
+#define CGTS_SM_CTRL_REG_ENABLE     0x96944200
+
+static void trinity_mg_clockgating_enable(struct radeon_device *rdev,
+					  bool enable)
+{
+	u32 local0;
+	u32 local1;
+
+	if (enable) {
+		local0 = RREG32_CG(CG_CGTT_LOCAL_0);
+		local1 = RREG32_CG(CG_CGTT_LOCAL_1);
+
+		WREG32_CG(CG_CGTT_LOCAL_0,
+			  (0x00380000 & CGCG_CGTT_LOCAL0_MASK) | (local0 & ~CGCG_CGTT_LOCAL0_MASK) );
+		WREG32_CG(CG_CGTT_LOCAL_1,
+			  (0x0E000000 & CGCG_CGTT_LOCAL1_MASK) | (local1 & ~CGCG_CGTT_LOCAL1_MASK) );
+
+		WREG32(CGTS_SM_CTRL_REG, CGTS_SM_CTRL_REG_ENABLE);
+	} else {
+		WREG32(CGTS_SM_CTRL_REG, CGTS_SM_CTRL_REG_DISABLE);
+
+		local0 = RREG32_CG(CG_CGTT_LOCAL_0);
+		local1 = RREG32_CG(CG_CGTT_LOCAL_1);
+
+		WREG32_CG(CG_CGTT_LOCAL_0,
+			  CGCG_CGTT_LOCAL0_MASK | (local0 & ~CGCG_CGTT_LOCAL0_MASK) );
+		WREG32_CG(CG_CGTT_LOCAL_1,
+			  CGCG_CGTT_LOCAL1_MASK | (local1 & ~CGCG_CGTT_LOCAL1_MASK) );
+	}
+}
+
+static void trinity_mg_clockgating_initialize(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *seq = NULL;
+
+	seq = &trinity_mgcg_shls_default[0];
+	count = sizeof(trinity_mgcg_shls_default) / (3 * sizeof(u32));
+
+	trinity_program_clk_gating_hw_sequence(rdev, seq, count);
+}
+
+static void trinity_gfx_clockgating_enable(struct radeon_device *rdev,
+					   bool enable)
+{
+	if (enable) {
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_GFX_CLK_OFF_EN, ~DYN_GFX_CLK_OFF_EN);
+	} else {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_GFX_CLK_OFF_EN);
+		WREG32_P(SCLK_PWRMGT_CNTL, GFX_CLK_FORCE_ON, ~GFX_CLK_FORCE_ON);
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~GFX_CLK_FORCE_ON);
+		RREG32(GB_ADDR_CONFIG);
+	}
+}
+
+static void trinity_program_clk_gating_hw_sequence(struct radeon_device *rdev,
+						   const u32 *seq, u32 count)
+{
+	u32 i, length = count * 3;
+
+	for (i = 0; i < length; i += 3)
+		WREG32_P(seq[i], seq[i+1], ~seq[i+2]);
+}
+
+static void trinity_program_override_mgpg_sequences(struct radeon_device *rdev,
+						    const u32 *seq, u32 count)
+{
+	u32  i, length = count * 2;
+
+	for (i = 0; i < length; i += 2)
+		WREG32(seq[i], seq[i+1]);
+
+}
+
+static void trinity_override_dynamic_mg_powergating(struct radeon_device *rdev)
+{
+	u32 count;
+	const u32 *seq = NULL;
+
+	seq = &trinity_override_mgpg_sequences[0];
+	count = sizeof(trinity_override_mgpg_sequences) / (2 * sizeof(u32));
+
+	trinity_program_override_mgpg_sequences(rdev, seq, count);
+}
+
+static void trinity_ls_clockgating_enable(struct radeon_device *rdev,
+					  bool enable)
+{
+	u32 count;
+	const u32 *seq = NULL;
+
+	if (enable) {
+		seq = &trinity_sysls_enable[0];
+		count = sizeof(trinity_sysls_enable) / (3 * sizeof(u32));
+	} else {
+		seq = &trinity_sysls_disable[0];
+		count = sizeof(trinity_sysls_disable) / (3 * sizeof(u32));
+	}
+
+	trinity_program_clk_gating_hw_sequence(rdev, seq, count);
+}
+
+static void trinity_gfx_powergating_enable(struct radeon_device *rdev,
+					   bool enable)
+{
+	if (enable) {
+		if (RREG32_SMC(CC_SMU_TST_EFUSE1_MISC) & RB_BACKEND_DISABLE_MASK)
+			WREG32_SMC(SMU_SCRATCH_A, (RREG32_SMC(SMU_SCRATCH_A) | 0x01));
+
+		WREG32_P(SCLK_PWRMGT_CNTL, DYN_PWR_DOWN_EN, ~DYN_PWR_DOWN_EN);
+	} else {
+		WREG32_P(SCLK_PWRMGT_CNTL, 0, ~DYN_PWR_DOWN_EN);
+		RREG32(GB_ADDR_CONFIG);
+	}
+}
+
+static void trinity_gfx_dynamic_mgpg_enable(struct radeon_device *rdev,
+					    bool enable)
+{
+	u32 value;
+
+	if (enable) {
+		value = RREG32_SMC(PM_I_CNTL_1);
+		value &= ~DS_PG_CNTL_MASK;
+		value |= DS_PG_CNTL(1);
+		WREG32_SMC(PM_I_CNTL_1, value);
+
+		value = RREG32_SMC(SMU_S_PG_CNTL);
+		value &= ~DS_PG_EN_MASK;
+		value |= DS_PG_EN(1);
+		WREG32_SMC(SMU_S_PG_CNTL, value);
+	} else {
+		value = RREG32_SMC(SMU_S_PG_CNTL);
+		value &= ~DS_PG_EN_MASK;
+		WREG32_SMC(SMU_S_PG_CNTL, value);
+
+		value = RREG32_SMC(PM_I_CNTL_1);
+		value &= ~DS_PG_CNTL_MASK;
+		WREG32_SMC(PM_I_CNTL_1, value);
+	}
+
+	trinity_gfx_dynamic_mgpg_config(rdev);
+
+}
+
+static void trinity_enable_clock_power_gating(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	if (pi->enable_gfx_clock_gating)
+		sumo_gfx_clockgating_initialize(rdev);
+	if (pi->enable_mg_clock_gating)
+		trinity_mg_clockgating_initialize(rdev);
+	if (pi->enable_gfx_power_gating)
+		trinity_gfx_powergating_initialize(rdev);
+	if (pi->enable_mg_clock_gating) {
+		trinity_ls_clockgating_enable(rdev, true);
+		trinity_mg_clockgating_enable(rdev, true);
+	}
+	if (pi->enable_gfx_clock_gating)
+		trinity_gfx_clockgating_enable(rdev, true);
+	if (pi->enable_gfx_dynamic_mgpg)
+		trinity_gfx_dynamic_mgpg_enable(rdev, true);
+	if (pi->enable_gfx_power_gating)
+		trinity_gfx_powergating_enable(rdev, true);
+}
+
+static void trinity_disable_clock_power_gating(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	if (pi->enable_gfx_power_gating)
+		trinity_gfx_powergating_enable(rdev, false);
+	if (pi->enable_gfx_dynamic_mgpg)
+		trinity_gfx_dynamic_mgpg_enable(rdev, false);
+	if (pi->enable_gfx_clock_gating)
+		trinity_gfx_clockgating_enable(rdev, false);
+	if (pi->enable_mg_clock_gating) {
+		trinity_mg_clockgating_enable(rdev, false);
+		trinity_ls_clockgating_enable(rdev, false);
+	}
+}
+
+static void trinity_set_divider_value(struct radeon_device *rdev,
+				      u32 index, u32 sclk)
+{
+	struct atom_clock_dividers  dividers;
+	int ret;
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+        ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+                                             sclk, false, &dividers);
+	if (ret)
+		return;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix);
+	value &= ~CLK_DIVIDER_MASK;
+	value |= CLK_DIVIDER(dividers.post_div);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix, value);
+
+        ret = radeon_atom_get_clock_dividers(rdev, COMPUTE_ENGINE_PLL_PARAM,
+                                             sclk/2, false, &dividers);
+	if (ret)
+		return;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_PG_CNTL + ix);
+	value &= ~PD_SCLK_DIVIDER_MASK;
+	value |= PD_SCLK_DIVIDER(dividers.post_div);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_PG_CNTL + ix, value);
+}
+
+static void trinity_set_ds_dividers(struct radeon_device *rdev,
+				    u32 index, u32 divider)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix);
+	value &= ~DS_DIV_MASK;
+	value |= DS_DIV(divider);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix, value);
+}
+
+static void trinity_set_ss_dividers(struct radeon_device *rdev,
+				    u32 index, u32 divider)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix);
+	value &= ~DS_SH_DIV_MASK;
+	value |= DS_SH_DIV(divider);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix, value);
+}
+
+static void trinity_set_vid(struct radeon_device *rdev, u32 index, u32 vid)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 vid_7bit = sumo_convert_vid2_to_vid7(rdev, &pi->sys_info.vid_mapping_table, vid);
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix);
+	value &= ~VID_MASK;
+	value |= VID(vid_7bit);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix, value);
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix);
+	value &= ~LVRT_MASK;
+	value |= LVRT(0);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix, value);
+}
+
+static void trinity_set_allos_gnb_slow(struct radeon_device *rdev,
+				       u32 index, u32 gnb_slow)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_3 + ix);
+	value &= ~GNB_SLOW_MASK;
+	value |= GNB_SLOW(gnb_slow);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_3 + ix, value);
+}
+
+static void trinity_set_force_nbp_state(struct radeon_device *rdev,
+					u32 index, u32 force_nbp_state)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_3 + ix);
+	value &= ~FORCE_NBPS1_MASK;
+	value |= FORCE_NBPS1(force_nbp_state);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_3 + ix, value);
+}
+
+static void trinity_set_display_wm(struct radeon_device *rdev,
+				   u32 index, u32 wm)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix);
+	value &= ~DISPLAY_WM_MASK;
+	value |= DISPLAY_WM(wm);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix, value);
+}
+
+static void trinity_set_vce_wm(struct radeon_device *rdev,
+			       u32 index, u32 wm)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix);
+	value &= ~VCE_WM_MASK;
+	value |= VCE_WM(wm);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_1 + ix, value);
+}
+
+static void trinity_set_at(struct radeon_device *rdev,
+			   u32 index, u32 at)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_AT + ix);
+	value &= ~AT_MASK;
+	value |= AT(at);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_AT + ix, value);
+}
+
+static void trinity_program_power_level(struct radeon_device *rdev,
+					struct trinity_pl *pl, u32 index)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	if (index >= SUMO_MAX_HARDWARE_POWERLEVELS)
+		return;
+
+	trinity_set_divider_value(rdev, index, pl->sclk);
+	trinity_set_vid(rdev, index, pl->vddc_index);
+	trinity_set_ss_dividers(rdev, index, pl->ss_divider_index);
+	trinity_set_ds_dividers(rdev, index, pl->ds_divider_index);
+	trinity_set_allos_gnb_slow(rdev, index, pl->allow_gnb_slow);
+	trinity_set_force_nbp_state(rdev, index, pl->force_nbp_state);
+	trinity_set_display_wm(rdev, index, pl->display_wm);
+	trinity_set_vce_wm(rdev, index, pl->vce_wm);
+	trinity_set_at(rdev, index, pi->at[index]);
+}
+
+static void trinity_power_level_enable_disable(struct radeon_device *rdev,
+					       u32 index, bool enable)
+{
+	u32 value;
+	u32 ix = index * TRINITY_SIZEOF_DPM_STATE_TABLE;
+
+	value = RREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix);
+	value &= ~STATE_VALID_MASK;
+	if (enable)
+		value |= STATE_VALID(1);
+	WREG32_SMC(SMU_SCLK_DPM_STATE_0_CNTL_0 + ix, value);
+}
+
+static bool trinity_dpm_enabled(struct radeon_device *rdev)
+{
+	if (RREG32_SMC(SMU_SCLK_DPM_CNTL) & SCLK_DPM_EN(1))
+		return true;
+	else
+		return false;
+}
+
+static void trinity_start_dpm(struct radeon_device *rdev)
+{
+	u32 value = RREG32_SMC(SMU_SCLK_DPM_CNTL);
+
+	value &= ~(SCLK_DPM_EN_MASK | SCLK_DPM_BOOT_STATE_MASK | VOLTAGE_CHG_EN_MASK);
+	value |= SCLK_DPM_EN(1) | SCLK_DPM_BOOT_STATE(0) | VOLTAGE_CHG_EN(1);
+	WREG32_SMC(SMU_SCLK_DPM_CNTL, value);
+
+	WREG32_P(GENERAL_PWRMGT, GLOBAL_PWRMGT_EN, ~GLOBAL_PWRMGT_EN);
+	WREG32_P(CG_CG_VOLTAGE_CNTL, 0, ~EN);
+
+	trinity_dpm_config(rdev, true);
+}
+
+static void trinity_wait_for_dpm_enabled(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(SCLK_PWRMGT_CNTL) & DYNAMIC_PM_EN)
+			break;
+		udelay(1);
+	}
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if ((RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & TARGET_STATE_MASK) == 0)
+			break;
+		udelay(1);
+	}
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if ((RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_STATE_MASK) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+static void trinity_stop_dpm(struct radeon_device *rdev)
+{
+	u32 sclk_dpm_cntl;
+
+	WREG32_P(CG_CG_VOLTAGE_CNTL, EN, ~EN);
+
+	sclk_dpm_cntl = RREG32_SMC(SMU_SCLK_DPM_CNTL);
+	sclk_dpm_cntl &= ~(SCLK_DPM_EN_MASK | VOLTAGE_CHG_EN_MASK);
+	WREG32_SMC(SMU_SCLK_DPM_CNTL, sclk_dpm_cntl);
+
+	trinity_dpm_config(rdev, false);
+}
+
+static void trinity_start_am(struct radeon_device *rdev)
+{
+	WREG32_P(SCLK_PWRMGT_CNTL, 0, ~(RESET_SCLK_CNT | RESET_BUSY_CNT));
+}
+
+static void trinity_reset_am(struct radeon_device *rdev)
+{
+	WREG32_P(SCLK_PWRMGT_CNTL, RESET_SCLK_CNT | RESET_BUSY_CNT,
+		 ~(RESET_SCLK_CNT | RESET_BUSY_CNT));
+}
+
+static void trinity_wait_for_level_0(struct radeon_device *rdev)
+{
+	int i;
+
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if ((RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_STATE_MASK) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+static void trinity_enable_power_level_0(struct radeon_device *rdev)
+{
+	trinity_power_level_enable_disable(rdev, 0, true);
+}
+
+static void trinity_force_level_0(struct radeon_device *rdev)
+{
+	trinity_dpm_force_state(rdev, 0);
+}
+
+static void trinity_unforce_levels(struct radeon_device *rdev)
+{
+	trinity_dpm_no_forced_level(rdev);
+}
+
+static void trinity_program_power_levels_0_to_n(struct radeon_device *rdev,
+						struct radeon_ps *new_rps,
+						struct radeon_ps *old_rps)
+{
+	struct trinity_ps *new_ps = trinity_get_ps(new_rps);
+	struct trinity_ps *old_ps = trinity_get_ps(old_rps);
+	u32 i;
+	u32 n_current_state_levels = (old_ps == NULL) ? 1 : old_ps->num_levels;
+
+	for (i = 0; i < new_ps->num_levels; i++) {
+		trinity_program_power_level(rdev, &new_ps->levels[i], i);
+		trinity_power_level_enable_disable(rdev, i, true);
+	}
+
+	for (i = new_ps->num_levels; i < n_current_state_levels; i++)
+		trinity_power_level_enable_disable(rdev, i, false);
+}
+
+static void trinity_program_bootup_state(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 i;
+
+	trinity_program_power_level(rdev, &pi->boot_pl, 0);
+	trinity_power_level_enable_disable(rdev, 0, true);
+
+	for (i = 1; i < 8; i++)
+		trinity_power_level_enable_disable(rdev, i, false);
+}
+
+static void trinity_setup_uvd_clock_table(struct radeon_device *rdev,
+					  struct radeon_ps *rps)
+{
+	struct trinity_ps *ps = trinity_get_ps(rps);
+	u32 uvdstates = (ps->vclk_low_divider |
+			 ps->vclk_high_divider << 8 |
+			 ps->dclk_low_divider << 16 |
+			 ps->dclk_high_divider << 24);
+
+	WREG32_SMC(SMU_UVD_DPM_STATES, uvdstates);
+}
+
+static void trinity_setup_uvd_dpm_interval(struct radeon_device *rdev,
+					   u32 interval)
+{
+	u32 p, u;
+	u32 tp = RREG32_SMC(PM_TP);
+	u32 val;
+	u32 xclk = radeon_get_xclk(rdev);
+
+	r600_calculate_u_and_p(interval, xclk, 16, &p, &u);
+
+	val = (p + tp - 1) / tp;
+
+	WREG32_SMC(SMU_UVD_DPM_CNTL, val);
+}
+
+static bool trinity_uvd_clocks_zero(struct radeon_ps *rps)
+{
+	if ((rps->vclk == 0) && (rps->dclk == 0))
+		return true;
+	else
+		return false;
+}
+
+static bool trinity_uvd_clocks_equal(struct radeon_ps *rps1,
+				     struct radeon_ps *rps2)
+{
+	struct trinity_ps *ps1 = trinity_get_ps(rps1);
+	struct trinity_ps *ps2 = trinity_get_ps(rps2);
+
+	if ((rps1->vclk == rps2->vclk) &&
+	    (rps1->dclk == rps2->dclk) &&
+	    (ps1->vclk_low_divider == ps2->vclk_low_divider) &&
+	    (ps1->vclk_high_divider == ps2->vclk_high_divider) &&
+	    (ps1->dclk_low_divider == ps2->dclk_low_divider) &&
+	    (ps1->dclk_high_divider == ps2->dclk_high_divider))
+		return true;
+	else
+		return false;
+}
+
+static void trinity_setup_uvd_clocks(struct radeon_device *rdev,
+				     struct radeon_ps *new_rps,
+				     struct radeon_ps *old_rps)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	if (pi->enable_gfx_power_gating) {
+		trinity_gfx_powergating_enable(rdev, false);
+	}
+
+	if (pi->uvd_dpm) {
+		if (trinity_uvd_clocks_zero(new_rps) &&
+		    !trinity_uvd_clocks_zero(old_rps)) {
+			trinity_setup_uvd_dpm_interval(rdev, 0);
+		} else if (!trinity_uvd_clocks_zero(new_rps)) {
+			trinity_setup_uvd_clock_table(rdev, new_rps);
+
+			if (trinity_uvd_clocks_zero(old_rps)) {
+				u32 tmp = RREG32(CG_MISC_REG);
+				tmp &= 0xfffffffd;
+				WREG32(CG_MISC_REG, tmp);
+
+				radeon_set_uvd_clocks(rdev, new_rps->vclk, new_rps->dclk);
+
+				trinity_setup_uvd_dpm_interval(rdev, 3000);
+			}
+		}
+		trinity_uvd_dpm_config(rdev);
+	} else {
+		if (trinity_uvd_clocks_zero(new_rps) ||
+		    trinity_uvd_clocks_equal(new_rps, old_rps))
+			return;
+
+		radeon_set_uvd_clocks(rdev, new_rps->vclk, new_rps->dclk);
+	}
+
+	if (pi->enable_gfx_power_gating) {
+		trinity_gfx_powergating_enable(rdev, true);
+	}
+}
+
+static void trinity_set_uvd_clock_before_set_eng_clock(struct radeon_device *rdev,
+						       struct radeon_ps *new_rps,
+						       struct radeon_ps *old_rps)
+{
+	struct trinity_ps *new_ps = trinity_get_ps(new_rps);
+	struct trinity_ps *current_ps = trinity_get_ps(new_rps);
+
+	if (new_ps->levels[new_ps->num_levels - 1].sclk >=
+	    current_ps->levels[current_ps->num_levels - 1].sclk)
+		return;
+
+	trinity_setup_uvd_clocks(rdev, new_rps, old_rps);
+}
+
+static void trinity_set_uvd_clock_after_set_eng_clock(struct radeon_device *rdev,
+						      struct radeon_ps *new_rps,
+						      struct radeon_ps *old_rps)
+{
+	struct trinity_ps *new_ps = trinity_get_ps(new_rps);
+	struct trinity_ps *current_ps = trinity_get_ps(old_rps);
+
+	if (new_ps->levels[new_ps->num_levels - 1].sclk <
+	    current_ps->levels[current_ps->num_levels - 1].sclk)
+		return;
+
+	trinity_setup_uvd_clocks(rdev, new_rps, old_rps);
+}
+
+static void trinity_program_ttt(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 value = RREG32_SMC(SMU_SCLK_DPM_TTT);
+
+	value &= ~(HT_MASK | LT_MASK);
+	value |= HT((pi->thermal_auto_throttling + 49) * 8);
+	value |= LT((pi->thermal_auto_throttling + 49 - pi->sys_info.htc_hyst_lmt) * 8);
+	WREG32_SMC(SMU_SCLK_DPM_TTT, value);
+}
+
+static void trinity_enable_att(struct radeon_device *rdev)
+{
+	u32 value = RREG32_SMC(SMU_SCLK_DPM_TT_CNTL);
+
+	value &= ~SCLK_TT_EN_MASK;
+	value |= SCLK_TT_EN(1);
+	WREG32_SMC(SMU_SCLK_DPM_TT_CNTL, value);
+}
+
+static void trinity_program_sclk_dpm(struct radeon_device *rdev)
+{
+	u32 p, u;
+	u32 tp = RREG32_SMC(PM_TP);
+	u32 ni;
+	u32 xclk = radeon_get_xclk(rdev);
+	u32 value;
+
+	r600_calculate_u_and_p(400, xclk, 16, &p, &u);
+
+	ni = (p + tp - 1) / tp;
+
+	value = RREG32_SMC(PM_I_CNTL_1);
+	value &= ~SCLK_DPM_MASK;
+	value |= SCLK_DPM(ni);
+	WREG32_SMC(PM_I_CNTL_1, value);
+}
+
+static int trinity_set_thermal_temperature_range(struct radeon_device *rdev,
+						 int min_temp, int max_temp)
+{
+	int low_temp = 0 * 1000;
+	int high_temp = 255 * 1000;
+
+        if (low_temp < min_temp)
+		low_temp = min_temp;
+        if (high_temp > max_temp)
+		high_temp = max_temp;
+        if (high_temp < low_temp) {
+		DRM_ERROR("invalid thermal range: %d - %d\n", low_temp, high_temp);
+                return -EINVAL;
+        }
+
+	WREG32_P(CG_THERMAL_INT_CTRL, DIG_THERM_INTH(49 + (high_temp / 1000)), ~DIG_THERM_INTH_MASK);
+	WREG32_P(CG_THERMAL_INT_CTRL, DIG_THERM_INTL(49 + (low_temp / 1000)), ~DIG_THERM_INTL_MASK);
+
+	rdev->pm.dpm.thermal.min_temp = low_temp;
+	rdev->pm.dpm.thermal.max_temp = high_temp;
+
+	return 0;
+}
+
+static void trinity_update_current_ps(struct radeon_device *rdev,
+				      struct radeon_ps *rps)
+{
+	struct trinity_ps *new_ps = trinity_get_ps(rps);
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	pi->current_rps = *rps;
+	pi->current_ps = *new_ps;
+	pi->current_rps.ps_priv = &pi->current_ps;
+}
+
+static void trinity_update_requested_ps(struct radeon_device *rdev,
+					struct radeon_ps *rps)
+{
+	struct trinity_ps *new_ps = trinity_get_ps(rps);
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	pi->requested_rps = *rps;
+	pi->requested_ps = *new_ps;
+	pi->requested_rps.ps_priv = &pi->requested_ps;
+}
+
+int trinity_dpm_enable(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	int ret;
+
+	trinity_acquire_mutex(rdev);
+
+	if (trinity_dpm_enabled(rdev)) {
+		trinity_release_mutex(rdev);
+		return -EINVAL;
+	}
+
+	trinity_enable_clock_power_gating(rdev);
+	trinity_program_bootup_state(rdev);
+	sumo_program_vc(rdev, 0x00C00033);
+	trinity_start_am(rdev);
+	if (pi->enable_auto_thermal_throttling) {
+		trinity_program_ttt(rdev);
+		trinity_enable_att(rdev);
+	}
+	trinity_program_sclk_dpm(rdev);
+	trinity_start_dpm(rdev);
+	trinity_wait_for_dpm_enabled(rdev);
+	trinity_release_mutex(rdev);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		ret = trinity_set_thermal_temperature_range(rdev, R600_TEMP_RANGE_MIN, R600_TEMP_RANGE_MAX);
+		if (ret) {
+			trinity_release_mutex(rdev);
+			return ret;
+		}
+		rdev->irq.dpm_thermal = true;
+		radeon_irq_set(rdev);
+	}
+
+	trinity_update_current_ps(rdev, rdev->pm.dpm.boot_ps);
+
+	return 0;
+}
+
+void trinity_dpm_disable(struct radeon_device *rdev)
+{
+	trinity_acquire_mutex(rdev);
+	if (!trinity_dpm_enabled(rdev)) {
+		trinity_release_mutex(rdev);
+		return;
+	}
+	trinity_disable_clock_power_gating(rdev);
+	sumo_clear_vc(rdev);
+	trinity_wait_for_level_0(rdev);
+	trinity_stop_dpm(rdev);
+	trinity_reset_am(rdev);
+	trinity_release_mutex(rdev);
+
+	if (rdev->irq.installed &&
+	    r600_is_internal_thermal_sensor(rdev->pm.int_thermal_type)) {
+		rdev->irq.dpm_thermal = false;
+		radeon_irq_set(rdev);
+	}
+
+	trinity_update_current_ps(rdev, rdev->pm.dpm.boot_ps);
+}
+
+static void trinity_get_min_sclk_divider(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	pi->min_sclk_did =
+		(RREG32_SMC(CC_SMU_MISC_FUSES) & MinSClkDid_MASK) >> MinSClkDid_SHIFT;
+}
+
+static void trinity_setup_nbp_sim(struct radeon_device *rdev,
+				  struct radeon_ps *rps)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct trinity_ps *new_ps = trinity_get_ps(rps);
+	u32 nbpsconfig;
+
+	if (pi->sys_info.nb_dpm_enable) {
+		nbpsconfig = RREG32_SMC(NB_PSTATE_CONFIG);
+		nbpsconfig &= ~(Dpm0PgNbPsLo_MASK | Dpm0PgNbPsHi_MASK | DpmXNbPsLo_MASK | DpmXNbPsHi_MASK);
+		nbpsconfig |= (Dpm0PgNbPsLo(new_ps->Dpm0PgNbPsLo) |
+			       Dpm0PgNbPsHi(new_ps->Dpm0PgNbPsHi) |
+			       DpmXNbPsLo(new_ps->DpmXNbPsLo) |
+			       DpmXNbPsHi(new_ps->DpmXNbPsHi));
+		WREG32_SMC(NB_PSTATE_CONFIG, nbpsconfig);
+	}
+}
+
+int trinity_dpm_force_performance_level(struct radeon_device *rdev,
+					enum radeon_dpm_forced_level level)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct radeon_ps *rps = &pi->current_rps;
+	struct trinity_ps *ps = trinity_get_ps(rps);
+	int i, ret;
+
+	if (ps->num_levels <= 1)
+		return 0;
+
+	if (level == RADEON_DPM_FORCED_LEVEL_HIGH) {
+		/* not supported by the hw */
+		return -EINVAL;
+	} else if (level == RADEON_DPM_FORCED_LEVEL_LOW) {
+		ret = trinity_dpm_n_levels_disabled(rdev, ps->num_levels - 1);
+		if (ret)
+			return ret;
+	} else {
+		for (i = 0; i < ps->num_levels; i++) {
+			ret = trinity_dpm_n_levels_disabled(rdev, 0);
+			if (ret)
+				return ret;
+		}
+	}
+
+	rdev->pm.dpm.forced_level = level;
+
+	return 0;
+}
+
+int trinity_dpm_pre_set_power_state(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct radeon_ps requested_ps = *rdev->pm.dpm.requested_ps;
+	struct radeon_ps *new_ps = &requested_ps;
+
+	trinity_update_requested_ps(rdev, new_ps);
+
+	trinity_apply_state_adjust_rules(rdev,
+					 &pi->requested_rps,
+					 &pi->current_rps);
+
+	return 0;
+}
+
+int trinity_dpm_set_power_state(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct radeon_ps *new_ps = &pi->requested_rps;
+	struct radeon_ps *old_ps = &pi->current_rps;
+
+	trinity_acquire_mutex(rdev);
+	if (pi->enable_dpm) {
+		trinity_set_uvd_clock_before_set_eng_clock(rdev, new_ps, old_ps);
+		trinity_enable_power_level_0(rdev);
+		trinity_force_level_0(rdev);
+		trinity_wait_for_level_0(rdev);
+		trinity_setup_nbp_sim(rdev, new_ps);
+		trinity_program_power_levels_0_to_n(rdev, new_ps, old_ps);
+		trinity_force_level_0(rdev);
+		trinity_unforce_levels(rdev);
+		trinity_set_uvd_clock_after_set_eng_clock(rdev, new_ps, old_ps);
+		rdev->pm.dpm.forced_level = RADEON_DPM_FORCED_LEVEL_AUTO;
+	}
+	trinity_release_mutex(rdev);
+
+	return 0;
+}
+
+void trinity_dpm_post_set_power_state(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct radeon_ps *new_ps = &pi->requested_rps;
+
+	trinity_update_current_ps(rdev, new_ps);
+}
+
+void trinity_dpm_setup_asic(struct radeon_device *rdev)
+{
+	trinity_acquire_mutex(rdev);
+	sumo_program_sstp(rdev);
+	sumo_take_smu_control(rdev, true);
+	trinity_get_min_sclk_divider(rdev);
+	trinity_release_mutex(rdev);
+}
+
+void trinity_dpm_reset_asic(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	trinity_acquire_mutex(rdev);
+	if (pi->enable_dpm) {
+		trinity_enable_power_level_0(rdev);
+		trinity_force_level_0(rdev);
+		trinity_wait_for_level_0(rdev);
+		trinity_program_bootup_state(rdev);
+		trinity_force_level_0(rdev);
+		trinity_unforce_levels(rdev);
+	}
+	trinity_release_mutex(rdev);
+}
+
+static u16 trinity_convert_voltage_index_to_value(struct radeon_device *rdev,
+						  u32 vid_2bit)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 vid_7bit = sumo_convert_vid2_to_vid7(rdev, &pi->sys_info.vid_mapping_table, vid_2bit);
+	u32 svi_mode = (RREG32_SMC(PM_CONFIG) & SVI_Mode) ? 1 : 0;
+	u32 step = (svi_mode == 0) ? 1250 : 625;
+	u32 delta = vid_7bit * step + 50;
+
+	if (delta > 155000)
+		return 0;
+
+	return (155000 - delta) / 100;
+}
+
+static void trinity_patch_boot_state(struct radeon_device *rdev,
+				     struct trinity_ps *ps)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	ps->num_levels = 1;
+	ps->nbps_flags = 0;
+	ps->bapm_flags = 0;
+	ps->levels[0] = pi->boot_pl;
+}
+
+static u8 trinity_calculate_vce_wm(struct radeon_device *rdev, u32 sclk)
+{
+	if (sclk < 20000)
+		return 1;
+	return 0;
+}
+
+static void trinity_construct_boot_state(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	pi->boot_pl.sclk = pi->sys_info.bootup_sclk;
+	pi->boot_pl.vddc_index = pi->sys_info.bootup_nb_voltage_index;
+	pi->boot_pl.ds_divider_index = 0;
+	pi->boot_pl.ss_divider_index = 0;
+	pi->boot_pl.allow_gnb_slow = 1;
+	pi->boot_pl.force_nbp_state = 0;
+	pi->boot_pl.display_wm = 0;
+	pi->boot_pl.vce_wm = 0;
+	pi->current_ps.num_levels = 1;
+	pi->current_ps.levels[0] = pi->boot_pl;
+}
+
+static u8 trinity_get_sleep_divider_id_from_clock(struct radeon_device *rdev,
+						  u32 sclk, u32 min_sclk_in_sr)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 i;
+	u32 temp;
+	u32 min = (min_sclk_in_sr > TRINITY_MINIMUM_ENGINE_CLOCK) ?
+		min_sclk_in_sr : TRINITY_MINIMUM_ENGINE_CLOCK;
+
+	if (sclk < min)
+		return 0;
+
+	if (!pi->enable_sclk_ds)
+		return 0;
+
+	for (i = TRINITY_MAX_DEEPSLEEP_DIVIDER_ID;  ; i--) {
+		temp = sclk / sumo_get_sleep_divider_from_id(i);
+		if (temp >= min || i == 0)
+			break;
+	}
+
+	return (u8)i;
+}
+
+static u32 trinity_get_valid_engine_clock(struct radeon_device *rdev,
+					  u32 lower_limit)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 i;
+
+	for (i = 0; i < pi->sys_info.sclk_voltage_mapping_table.num_max_dpm_entries; i++) {
+		if (pi->sys_info.sclk_voltage_mapping_table.entries[i].sclk_frequency >= lower_limit)
+			return pi->sys_info.sclk_voltage_mapping_table.entries[i].sclk_frequency;
+	}
+
+	if (i == pi->sys_info.sclk_voltage_mapping_table.num_max_dpm_entries)
+		DRM_ERROR("engine clock out of range!");
+
+	return 0;
+}
+
+static void trinity_patch_thermal_state(struct radeon_device *rdev,
+					struct trinity_ps *ps,
+					struct trinity_ps *current_ps)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 sclk_in_sr = pi->sys_info.min_sclk; /* ??? */
+	u32 current_vddc;
+	u32 current_sclk;
+	u32 current_index = 0;
+
+	if (current_ps) {
+		current_vddc = current_ps->levels[current_index].vddc_index;
+		current_sclk = current_ps->levels[current_index].sclk;
+	} else {
+		current_vddc = pi->boot_pl.vddc_index;
+		current_sclk = pi->boot_pl.sclk;
+	}
+
+	ps->levels[0].vddc_index = current_vddc;
+
+	if (ps->levels[0].sclk > current_sclk)
+		ps->levels[0].sclk = current_sclk;
+
+	ps->levels[0].ds_divider_index =
+		trinity_get_sleep_divider_id_from_clock(rdev, ps->levels[0].sclk, sclk_in_sr);
+	ps->levels[0].ss_divider_index = ps->levels[0].ds_divider_index;
+	ps->levels[0].allow_gnb_slow = 1;
+	ps->levels[0].force_nbp_state = 0;
+	ps->levels[0].display_wm = 0;
+	ps->levels[0].vce_wm =
+		trinity_calculate_vce_wm(rdev, ps->levels[0].sclk);
+}
+
+static u8 trinity_calculate_display_wm(struct radeon_device *rdev,
+				       struct trinity_ps *ps, u32 index)
+{
+	if (ps == NULL || ps->num_levels <= 1)
+		return 0;
+	else if (ps->num_levels == 2) {
+		if (index == 0)
+			return 0;
+		else
+			return 1;
+	} else {
+		if (index == 0)
+			return 0;
+		else if (ps->levels[index].sclk < 30000)
+			return 0;
+		else
+			return 1;
+	}
+}
+
+static u32 trinity_get_uvd_clock_index(struct radeon_device *rdev,
+				       struct radeon_ps *rps)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 i = 0;
+
+	for (i = 0; i < 4; i++) {
+		if ((rps->vclk == pi->sys_info.uvd_clock_table_entries[i].vclk) &&
+		    (rps->dclk == pi->sys_info.uvd_clock_table_entries[i].dclk))
+		    break;
+	}
+
+	if (i >= 4) {
+		DRM_ERROR("UVD clock index not found!\n");
+		i = 3;
+	}
+	return i;
+}
+
+static void trinity_adjust_uvd_state(struct radeon_device *rdev,
+				     struct radeon_ps *rps)
+{
+	struct trinity_ps *ps = trinity_get_ps(rps);
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 high_index = 0;
+	u32 low_index = 0;
+
+	if (pi->uvd_dpm && r600_is_uvd_state(rps->class, rps->class2)) {
+		high_index = trinity_get_uvd_clock_index(rdev, rps);
+
+		switch(high_index) {
+		case 3:
+		case 2:
+			low_index = 1;
+			break;
+		case 1:
+		case 0:
+		default:
+			low_index = 0;
+			break;
+		}
+
+		ps->vclk_low_divider =
+			pi->sys_info.uvd_clock_table_entries[high_index].vclk_did;
+		ps->dclk_low_divider =
+			pi->sys_info.uvd_clock_table_entries[high_index].dclk_did;
+		ps->vclk_high_divider =
+			pi->sys_info.uvd_clock_table_entries[low_index].vclk_did;
+		ps->dclk_high_divider =
+			pi->sys_info.uvd_clock_table_entries[low_index].dclk_did;
+	}
+}
+
+
+
+static void trinity_apply_state_adjust_rules(struct radeon_device *rdev,
+					     struct radeon_ps *new_rps,
+					     struct radeon_ps *old_rps)
+{
+	struct trinity_ps *ps = trinity_get_ps(new_rps);
+	struct trinity_ps *current_ps = trinity_get_ps(old_rps);
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 min_voltage = 0; /* ??? */
+	u32 min_sclk = pi->sys_info.min_sclk; /* XXX check against disp reqs */
+	u32 sclk_in_sr = pi->sys_info.min_sclk; /* ??? */
+	u32 i;
+	bool force_high;
+	u32 num_active_displays = rdev->pm.dpm.new_active_crtc_count;
+
+	if (new_rps->class & ATOM_PPLIB_CLASSIFICATION_THERMAL)
+		return trinity_patch_thermal_state(rdev, ps, current_ps);
+
+	trinity_adjust_uvd_state(rdev, new_rps);
+
+	for (i = 0; i < ps->num_levels; i++) {
+		if (ps->levels[i].vddc_index < min_voltage)
+			ps->levels[i].vddc_index = min_voltage;
+
+		if (ps->levels[i].sclk < min_sclk)
+			ps->levels[i].sclk =
+				trinity_get_valid_engine_clock(rdev, min_sclk);
+
+		ps->levels[i].ds_divider_index =
+			sumo_get_sleep_divider_id_from_clock(rdev, ps->levels[i].sclk, sclk_in_sr);
+
+		ps->levels[i].ss_divider_index = ps->levels[i].ds_divider_index;
+
+		ps->levels[i].allow_gnb_slow = 1;
+		ps->levels[i].force_nbp_state = 0;
+		ps->levels[i].display_wm =
+			trinity_calculate_display_wm(rdev, ps, i);
+		ps->levels[i].vce_wm =
+			trinity_calculate_vce_wm(rdev, ps->levels[0].sclk);
+	}
+
+	if ((new_rps->class & (ATOM_PPLIB_CLASSIFICATION_HDSTATE | ATOM_PPLIB_CLASSIFICATION_SDSTATE)) ||
+	    ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) == ATOM_PPLIB_CLASSIFICATION_UI_BATTERY))
+		ps->bapm_flags |= TRINITY_POWERSTATE_FLAGS_BAPM_DISABLE;
+
+	if (pi->sys_info.nb_dpm_enable) {
+		ps->Dpm0PgNbPsLo = 0x1;
+		ps->Dpm0PgNbPsHi = 0x0;
+		ps->DpmXNbPsLo = 0x2;
+		ps->DpmXNbPsHi = 0x1;
+
+		if ((new_rps->class & (ATOM_PPLIB_CLASSIFICATION_HDSTATE | ATOM_PPLIB_CLASSIFICATION_SDSTATE)) ||
+		    ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) == ATOM_PPLIB_CLASSIFICATION_UI_BATTERY)) {
+			force_high = ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_HDSTATE) ||
+				      ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_SDSTATE) &&
+				       (pi->sys_info.uma_channel_number == 1)));
+			force_high = (num_active_displays >= 3) || force_high;
+			ps->Dpm0PgNbPsLo = force_high ? 0x2 : 0x3;
+			ps->Dpm0PgNbPsHi = 0x1;
+			ps->DpmXNbPsLo = force_high ? 0x2 : 0x3;
+			ps->DpmXNbPsHi = 0x2;
+			ps->levels[ps->num_levels - 1].allow_gnb_slow = 0;
+		}
+	}
+}
+
+static void trinity_cleanup_asic(struct radeon_device *rdev)
+{
+	sumo_take_smu_control(rdev, false);
+}
+
+#if 0
+static void trinity_pre_display_configuration_change(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	if (pi->voltage_drop_in_dce)
+		trinity_dce_enable_voltage_adjustment(rdev, false);
+}
+#endif
+
+static void trinity_add_dccac_value(struct radeon_device *rdev)
+{
+	u32 gpu_cac_avrg_cntl_window_size;
+	u32 num_active_displays = rdev->pm.dpm.new_active_crtc_count;
+	u64 disp_clk = rdev->clock.default_dispclk / 100;
+	u32 dc_cac_value;
+
+	gpu_cac_avrg_cntl_window_size =
+		(RREG32_SMC(GPU_CAC_AVRG_CNTL) & WINDOW_SIZE_MASK) >> WINDOW_SIZE_SHIFT;
+
+	dc_cac_value = (u32)((14213 * disp_clk * disp_clk * (u64)num_active_displays) >>
+			     (32 - gpu_cac_avrg_cntl_window_size));
+
+	WREG32_SMC(DC_CAC_VALUE, dc_cac_value);
+}
+
+void trinity_dpm_display_configuration_changed(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	if (pi->voltage_drop_in_dce)
+		trinity_dce_enable_voltage_adjustment(rdev, true);
+	trinity_add_dccac_value(rdev);
+}
+
+union power_info {
+	struct _ATOM_POWERPLAY_INFO info;
+	struct _ATOM_POWERPLAY_INFO_V2 info_2;
+	struct _ATOM_POWERPLAY_INFO_V3 info_3;
+	struct _ATOM_PPLIB_POWERPLAYTABLE pplib;
+	struct _ATOM_PPLIB_POWERPLAYTABLE2 pplib2;
+	struct _ATOM_PPLIB_POWERPLAYTABLE3 pplib3;
+};
+
+union pplib_clock_info {
+	struct _ATOM_PPLIB_R600_CLOCK_INFO r600;
+	struct _ATOM_PPLIB_RS780_CLOCK_INFO rs780;
+	struct _ATOM_PPLIB_EVERGREEN_CLOCK_INFO evergreen;
+	struct _ATOM_PPLIB_SUMO_CLOCK_INFO sumo;
+};
+
+union pplib_power_state {
+	struct _ATOM_PPLIB_STATE v1;
+	struct _ATOM_PPLIB_STATE_V2 v2;
+};
+
+static void trinity_parse_pplib_non_clock_info(struct radeon_device *rdev,
+					       struct radeon_ps *rps,
+					       struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info,
+					       u8 table_rev)
+{
+	struct trinity_ps *ps = trinity_get_ps(rps);
+
+	rps->caps = le32_to_cpu(non_clock_info->ulCapsAndSettings);
+	rps->class = le16_to_cpu(non_clock_info->usClassification);
+	rps->class2 = le16_to_cpu(non_clock_info->usClassification2);
+
+	if (ATOM_PPLIB_NONCLOCKINFO_VER1 < table_rev) {
+		rps->vclk = le32_to_cpu(non_clock_info->ulVCLK);
+		rps->dclk = le32_to_cpu(non_clock_info->ulDCLK);
+	} else {
+		rps->vclk = 0;
+		rps->dclk = 0;
+	}
+
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_BOOT) {
+		rdev->pm.dpm.boot_ps = rps;
+		trinity_patch_boot_state(rdev, ps);
+	}
+	if (rps->class & ATOM_PPLIB_CLASSIFICATION_UVDSTATE)
+		rdev->pm.dpm.uvd_ps = rps;
+}
+
+static void trinity_parse_pplib_clock_info(struct radeon_device *rdev,
+					   struct radeon_ps *rps, int index,
+					   union pplib_clock_info *clock_info)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct trinity_ps *ps = trinity_get_ps(rps);
+	struct trinity_pl *pl = &ps->levels[index];
+	u32 sclk;
+
+	sclk = le16_to_cpu(clock_info->sumo.usEngineClockLow);
+	sclk |= clock_info->sumo.ucEngineClockHigh << 16;
+	pl->sclk = sclk;
+	pl->vddc_index = clock_info->sumo.vddcIndex;
+
+	ps->num_levels = index + 1;
+
+	if (pi->enable_sclk_ds) {
+		pl->ds_divider_index = 5;
+		pl->ss_divider_index = 5;
+	}
+}
+
+static int trinity_parse_power_table(struct radeon_device *rdev)
+{
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	struct _ATOM_PPLIB_NONCLOCK_INFO *non_clock_info;
+	union pplib_power_state *power_state;
+	int i, j, k, non_clock_array_index, clock_array_index;
+	union pplib_clock_info *clock_info;
+	struct _StateArray *state_array;
+	struct _ClockInfoArray *clock_info_array;
+	struct _NonClockInfoArray *non_clock_info_array;
+	union power_info *power_info;
+	int index = GetIndexIntoMasterTable(DATA, PowerPlayInfo);
+        u16 data_offset;
+	u8 frev, crev;
+	u8 *power_state_offset;
+	struct sumo_ps *ps;
+
+	if (!atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset))
+		return -EINVAL;
+	power_info = (union power_info *)(mode_info->atom_context->bios + data_offset);
+
+	state_array = (struct _StateArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usStateArrayOffset));
+	clock_info_array = (struct _ClockInfoArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset));
+	non_clock_info_array = (struct _NonClockInfoArray *)
+		(mode_info->atom_context->bios + data_offset +
+		 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset));
+
+	rdev->pm.dpm.ps = kzalloc(sizeof(struct radeon_ps) *
+				  state_array->ucNumEntries, GFP_KERNEL);
+	if (!rdev->pm.dpm.ps)
+		return -ENOMEM;
+	power_state_offset = (u8 *)state_array->states;
+	rdev->pm.dpm.platform_caps = le32_to_cpu(power_info->pplib.ulPlatformCaps);
+	rdev->pm.dpm.backbias_response_time = le16_to_cpu(power_info->pplib.usBackbiasTime);
+	rdev->pm.dpm.voltage_response_time = le16_to_cpu(power_info->pplib.usVoltageTime);
+	for (i = 0; i < state_array->ucNumEntries; i++) {
+		power_state = (union pplib_power_state *)power_state_offset;
+		non_clock_array_index = power_state->v2.nonClockInfoIndex;
+		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
+			&non_clock_info_array->nonClockInfo[non_clock_array_index];
+		if (!rdev->pm.power_state[i].clock_info)
+			return -EINVAL;
+		ps = kzalloc(sizeof(struct sumo_ps), GFP_KERNEL);
+		if (ps == NULL) {
+			kfree(rdev->pm.dpm.ps);
+			return -ENOMEM;
+		}
+		rdev->pm.dpm.ps[i].ps_priv = ps;
+		k = 0;
+		for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) {
+			clock_array_index = power_state->v2.clockInfoIndex[j];
+			if (clock_array_index >= clock_info_array->ucNumEntries)
+				continue;
+			if (k >= SUMO_MAX_HARDWARE_POWERLEVELS)
+				break;
+			clock_info = (union pplib_clock_info *)
+				&clock_info_array->clockInfo[clock_array_index * clock_info_array->ucEntrySize];
+			trinity_parse_pplib_clock_info(rdev,
+						       &rdev->pm.dpm.ps[i], k,
+						       clock_info);
+			k++;
+		}
+		trinity_parse_pplib_non_clock_info(rdev, &rdev->pm.dpm.ps[i],
+						   non_clock_info,
+						   non_clock_info_array->ucEntrySize);
+		power_state_offset += 2 + power_state->v2.ucNumDPMLevels;
+	}
+	rdev->pm.dpm.num_ps = state_array->ucNumEntries;
+	return 0;
+}
+
+union igp_info {
+	struct _ATOM_INTEGRATED_SYSTEM_INFO info;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V2 info_2;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V5 info_5;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V6 info_6;
+	struct _ATOM_INTEGRATED_SYSTEM_INFO_V1_7 info_7;
+};
+
+static u32 trinity_convert_did_to_freq(struct radeon_device *rdev, u8 did)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	u32 divider;
+
+	if (did >= 8 && did <= 0x3f)
+		divider = did * 25;
+	else if (did > 0x3f && did <= 0x5f)
+		divider = (did - 64) * 50 + 1600;
+	else if (did > 0x5f && did <= 0x7e)
+		divider = (did - 96) * 100 + 3200;
+	else if (did == 0x7f)
+		divider = 128 * 100;
+	else
+		return 10000;
+
+	return ((pi->sys_info.dentist_vco_freq * 100) + (divider - 1)) / divider;
+}
+
+static int trinity_parse_sys_info_table(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct radeon_mode_info *mode_info = &rdev->mode_info;
+	int index = GetIndexIntoMasterTable(DATA, IntegratedSystemInfo);
+	union igp_info *igp_info;
+	u8 frev, crev;
+	u16 data_offset;
+	int i;
+
+	if (atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset)) {
+		igp_info = (union igp_info *)(mode_info->atom_context->bios +
+					      data_offset);
+
+		if (crev != 7) {
+			DRM_ERROR("Unsupported IGP table: %d %d\n", frev, crev);
+			return -EINVAL;
+		}
+		pi->sys_info.bootup_sclk = le32_to_cpu(igp_info->info_7.ulBootUpEngineClock);
+		pi->sys_info.min_sclk = le32_to_cpu(igp_info->info_7.ulMinEngineClock);
+		pi->sys_info.bootup_uma_clk = le32_to_cpu(igp_info->info_7.ulBootUpUMAClock);
+		pi->sys_info.dentist_vco_freq = le32_to_cpu(igp_info->info_7.ulDentistVCOFreq);
+		pi->sys_info.bootup_nb_voltage_index =
+			le16_to_cpu(igp_info->info_7.usBootUpNBVoltage);
+		if (igp_info->info_7.ucHtcTmpLmt == 0)
+			pi->sys_info.htc_tmp_lmt = 203;
+		else
+			pi->sys_info.htc_tmp_lmt = igp_info->info_7.ucHtcTmpLmt;
+		if (igp_info->info_7.ucHtcHystLmt == 0)
+			pi->sys_info.htc_hyst_lmt = 5;
+		else
+			pi->sys_info.htc_hyst_lmt = igp_info->info_7.ucHtcHystLmt;
+		if (pi->sys_info.htc_tmp_lmt <= pi->sys_info.htc_hyst_lmt) {
+			DRM_ERROR("The htcTmpLmt should be larger than htcHystLmt.\n");
+		}
+
+		if (pi->enable_nbps_policy)
+			pi->sys_info.nb_dpm_enable = igp_info->info_7.ucNBDPMEnable;
+		else
+			pi->sys_info.nb_dpm_enable = 0;
+
+		for (i = 0; i < TRINITY_NUM_NBPSTATES; i++) {
+			pi->sys_info.nbp_mclk[i] = le32_to_cpu(igp_info->info_7.ulNbpStateMemclkFreq[i]);
+			pi->sys_info.nbp_nclk[i] = le32_to_cpu(igp_info->info_7.ulNbpStateNClkFreq[i]);
+		}
+
+		pi->sys_info.nbp_voltage_index[0] = le16_to_cpu(igp_info->info_7.usNBP0Voltage);
+		pi->sys_info.nbp_voltage_index[1] = le16_to_cpu(igp_info->info_7.usNBP1Voltage);
+		pi->sys_info.nbp_voltage_index[2] = le16_to_cpu(igp_info->info_7.usNBP2Voltage);
+		pi->sys_info.nbp_voltage_index[3] = le16_to_cpu(igp_info->info_7.usNBP3Voltage);
+
+		if (!pi->sys_info.nb_dpm_enable) {
+			for (i = 1; i < TRINITY_NUM_NBPSTATES; i++) {
+				pi->sys_info.nbp_mclk[i] = pi->sys_info.nbp_mclk[0];
+				pi->sys_info.nbp_nclk[i] = pi->sys_info.nbp_nclk[0];
+				pi->sys_info.nbp_voltage_index[i] = pi->sys_info.nbp_voltage_index[0];
+			}
+		}
+
+		pi->sys_info.uma_channel_number = igp_info->info_7.ucUMAChannelNumber;
+
+		sumo_construct_sclk_voltage_mapping_table(rdev,
+							  &pi->sys_info.sclk_voltage_mapping_table,
+							  igp_info->info_7.sAvail_SCLK);
+		sumo_construct_vid_mapping_table(rdev, &pi->sys_info.vid_mapping_table,
+						 igp_info->info_7.sAvail_SCLK);
+
+		pi->sys_info.uvd_clock_table_entries[0].vclk_did =
+			igp_info->info_7.ucDPMState0VclkFid;
+		pi->sys_info.uvd_clock_table_entries[1].vclk_did =
+			igp_info->info_7.ucDPMState1VclkFid;
+		pi->sys_info.uvd_clock_table_entries[2].vclk_did =
+			igp_info->info_7.ucDPMState2VclkFid;
+		pi->sys_info.uvd_clock_table_entries[3].vclk_did =
+			igp_info->info_7.ucDPMState3VclkFid;
+
+		pi->sys_info.uvd_clock_table_entries[0].dclk_did =
+			igp_info->info_7.ucDPMState0DclkFid;
+		pi->sys_info.uvd_clock_table_entries[1].dclk_did =
+			igp_info->info_7.ucDPMState1DclkFid;
+		pi->sys_info.uvd_clock_table_entries[2].dclk_did =
+			igp_info->info_7.ucDPMState2DclkFid;
+		pi->sys_info.uvd_clock_table_entries[3].dclk_did =
+			igp_info->info_7.ucDPMState3DclkFid;
+
+		for (i = 0; i < 4; i++) {
+			pi->sys_info.uvd_clock_table_entries[i].vclk =
+				trinity_convert_did_to_freq(rdev,
+							    pi->sys_info.uvd_clock_table_entries[i].vclk_did);
+			pi->sys_info.uvd_clock_table_entries[i].dclk =
+				trinity_convert_did_to_freq(rdev,
+							    pi->sys_info.uvd_clock_table_entries[i].dclk_did);
+		}
+
+
+
+	}
+	return 0;
+}
+
+int trinity_dpm_init(struct radeon_device *rdev)
+{
+	struct trinity_power_info *pi;
+	int ret, i;
+
+	pi = kzalloc(sizeof(struct trinity_power_info), GFP_KERNEL);
+	if (pi == NULL)
+		return -ENOMEM;
+	rdev->pm.dpm.priv = pi;
+
+	for (i = 0; i < SUMO_MAX_HARDWARE_POWERLEVELS; i++)
+		pi->at[i] = TRINITY_AT_DFLT;
+
+	pi->enable_nbps_policy = true;
+	pi->enable_sclk_ds = true;
+	pi->enable_gfx_power_gating = true;
+	pi->enable_gfx_clock_gating = true;
+	pi->enable_mg_clock_gating = true;
+	pi->enable_gfx_dynamic_mgpg = true; /* ??? */
+	pi->override_dynamic_mgpg = true;
+	pi->enable_auto_thermal_throttling = true;
+	pi->voltage_drop_in_dce = false; /* need to restructure dpm/modeset interaction */
+	pi->uvd_dpm = true; /* ??? */
+
+	ret = trinity_parse_sys_info_table(rdev);
+	if (ret)
+		return ret;
+
+	trinity_construct_boot_state(rdev);
+
+	ret = trinity_parse_power_table(rdev);
+	if (ret)
+		return ret;
+
+	pi->thermal_auto_throttling = pi->sys_info.htc_tmp_lmt;
+	pi->enable_dpm = true;
+
+	return 0;
+}
+
+void trinity_dpm_print_power_state(struct radeon_device *rdev,
+				   struct radeon_ps *rps)
+{
+	int i;
+	struct trinity_ps *ps = trinity_get_ps(rps);
+
+	r600_dpm_print_class_info(rps->class, rps->class2);
+	r600_dpm_print_cap_info(rps->caps);
+	printk("\tuvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+	for (i = 0; i < ps->num_levels; i++) {
+		struct trinity_pl *pl = &ps->levels[i];
+		printk("\t\tpower level %d    sclk: %u vddc: %u\n",
+		       i, pl->sclk,
+		       trinity_convert_voltage_index_to_value(rdev, pl->vddc_index));
+	}
+	r600_dpm_print_ps_status(rdev, rps);
+}
+
+void trinity_dpm_debugfs_print_current_performance_level(struct radeon_device *rdev,
+							 struct seq_file *m)
+{
+	struct radeon_ps *rps = rdev->pm.dpm.current_ps;
+	struct trinity_ps *ps = trinity_get_ps(rps);
+	struct trinity_pl *pl;
+	u32 current_index =
+		(RREG32(TARGET_AND_CURRENT_PROFILE_INDEX) & CURRENT_STATE_MASK) >>
+		CURRENT_STATE_SHIFT;
+
+	if (current_index >= ps->num_levels) {
+		seq_printf(m, "invalid dpm profile %d\n", current_index);
+	} else {
+		pl = &ps->levels[current_index];
+		seq_printf(m, "uvd    vclk: %d dclk: %d\n", rps->vclk, rps->dclk);
+		seq_printf(m, "power level %d    sclk: %u vddc: %u\n",
+			   current_index, pl->sclk,
+			   trinity_convert_voltage_index_to_value(rdev, pl->vddc_index));
+	}
+}
+
+void trinity_dpm_fini(struct radeon_device *rdev)
+{
+	int i;
+
+	trinity_cleanup_asic(rdev); /* ??? */
+
+	for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
+		kfree(rdev->pm.dpm.ps[i].ps_priv);
+	}
+	kfree(rdev->pm.dpm.ps);
+	kfree(rdev->pm.dpm.priv);
+}
+
+u32 trinity_dpm_get_sclk(struct radeon_device *rdev, bool low)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+	struct trinity_ps *requested_state = trinity_get_ps(&pi->requested_rps);
+
+	if (low)
+		return requested_state->levels[0].sclk;
+	else
+		return requested_state->levels[requested_state->num_levels - 1].sclk;
+}
+
+u32 trinity_dpm_get_mclk(struct radeon_device *rdev, bool low)
+{
+	struct trinity_power_info *pi = trinity_get_pi(rdev);
+
+	return pi->sys_info.bootup_uma_clk;
+}
diff --git a/drivers/gpu/drm/radeon/trinity_dpm.h b/drivers/gpu/drm/radeon/trinity_dpm.h
new file mode 100644
index 000000000000..e82df071f8b3
--- /dev/null
+++ b/drivers/gpu/drm/radeon/trinity_dpm.h
@@ -0,0 +1,132 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#ifndef __TRINITY_DPM_H__
+#define __TRINITY_DPM_H__
+
+#include "sumo_dpm.h"
+
+#define TRINITY_SIZEOF_DPM_STATE_TABLE (SMU_SCLK_DPM_STATE_1_CNTL_0 - SMU_SCLK_DPM_STATE_0_CNTL_0)
+
+struct trinity_pl {
+	u32 sclk;
+	u8 vddc_index;
+	u8 ds_divider_index;
+	u8 ss_divider_index;
+	u8 allow_gnb_slow;
+	u8 force_nbp_state;
+	u8 display_wm;
+	u8 vce_wm;
+};
+
+#define TRINITY_POWERSTATE_FLAGS_NBPS_FORCEHIGH  (1 << 0)
+#define TRINITY_POWERSTATE_FLAGS_NBPS_LOCKTOHIGH (1 << 1)
+#define TRINITY_POWERSTATE_FLAGS_NBPS_LOCKTOLOW  (1 << 2)
+
+#define TRINITY_POWERSTATE_FLAGS_BAPM_DISABLE    (1 << 0)
+
+struct trinity_ps {
+	u32 num_levels;
+	struct trinity_pl levels[SUMO_MAX_HARDWARE_POWERLEVELS];
+
+	u32 nbps_flags;
+	u32 bapm_flags;
+
+	u8 Dpm0PgNbPsLo;
+	u8 Dpm0PgNbPsHi;
+	u8 DpmXNbPsLo;
+	u8 DpmXNbPsHi;
+
+	u32 vclk_low_divider;
+	u32 vclk_high_divider;
+	u32 dclk_low_divider;
+	u32 dclk_high_divider;
+};
+
+#define TRINITY_NUM_NBPSTATES   4
+
+struct trinity_uvd_clock_table_entry
+{
+	u32 vclk;
+	u32 dclk;
+	u8 vclk_did;
+	u8 dclk_did;
+	u8 rsv[2];
+};
+
+struct trinity_sys_info {
+	u32 bootup_uma_clk;
+	u32 bootup_sclk;
+	u32 min_sclk;
+	u32 dentist_vco_freq;
+	u32 nb_dpm_enable;
+	u32 nbp_mclk[TRINITY_NUM_NBPSTATES];
+	u32 nbp_nclk[TRINITY_NUM_NBPSTATES];
+	u16 nbp_voltage_index[TRINITY_NUM_NBPSTATES];
+	u16 bootup_nb_voltage_index;
+	u8 htc_tmp_lmt;
+	u8 htc_hyst_lmt;
+	struct sumo_sclk_voltage_mapping_table sclk_voltage_mapping_table;
+	struct sumo_vid_mapping_table vid_mapping_table;
+	u32 uma_channel_number;
+	struct trinity_uvd_clock_table_entry uvd_clock_table_entries[4];
+};
+
+struct trinity_power_info {
+	u32 at[SUMO_MAX_HARDWARE_POWERLEVELS];
+	u32 dpm_interval;
+	u32 thermal_auto_throttling;
+	struct trinity_sys_info sys_info;
+	struct trinity_pl boot_pl;
+	u32 min_sclk_did;
+	bool enable_nbps_policy;
+	bool voltage_drop_in_dce;
+	bool override_dynamic_mgpg;
+	bool enable_gfx_clock_gating;
+	bool enable_gfx_power_gating;
+	bool enable_mg_clock_gating;
+	bool enable_gfx_dynamic_mgpg;
+	bool enable_auto_thermal_throttling;
+	bool enable_dpm;
+	bool enable_sclk_ds;
+	bool uvd_dpm;
+	struct radeon_ps current_rps;
+	struct trinity_ps current_ps;
+	struct radeon_ps requested_rps;
+	struct trinity_ps requested_ps;
+};
+
+#define TRINITY_AT_DFLT            30
+
+/* trinity_smc.c */
+int trinity_dpm_config(struct radeon_device *rdev, bool enable);
+int trinity_uvd_dpm_config(struct radeon_device *rdev);
+int trinity_dpm_force_state(struct radeon_device *rdev, u32 n);
+int trinity_dpm_n_levels_disabled(struct radeon_device *rdev, u32 n);
+int trinity_dpm_no_forced_level(struct radeon_device *rdev);
+int trinity_dce_enable_voltage_adjustment(struct radeon_device *rdev,
+					  bool enable);
+int trinity_gfx_dynamic_mgpg_config(struct radeon_device *rdev);
+void trinity_acquire_mutex(struct radeon_device *rdev);
+void trinity_release_mutex(struct radeon_device *rdev);
+
+#endif
diff --git a/drivers/gpu/drm/radeon/trinity_smc.c b/drivers/gpu/drm/radeon/trinity_smc.c
new file mode 100644
index 000000000000..a42d89f1830c
--- /dev/null
+++ b/drivers/gpu/drm/radeon/trinity_smc.c
@@ -0,0 +1,122 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "drmP.h"
+#include "radeon.h"
+#include "trinityd.h"
+#include "trinity_dpm.h"
+#include "ppsmc.h"
+
+struct trinity_ps *trinity_get_ps(struct radeon_ps *rps);
+struct trinity_power_info *trinity_get_pi(struct radeon_device *rdev);
+
+static int trinity_notify_message_to_smu(struct radeon_device *rdev, u32 id)
+{
+	int i;
+	u32 v = 0;
+
+	WREG32(SMC_MESSAGE_0, id);
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if (RREG32(SMC_RESP_0) != 0)
+			break;
+		udelay(1);
+	}
+	v = RREG32(SMC_RESP_0);
+
+	if (v != 1) {
+		if (v == 0xFF) {
+			DRM_ERROR("SMC failed to handle the message!\n");
+			return -EINVAL;
+		} else if (v == 0xFE) {
+			DRM_ERROR("Unknown SMC message!\n");
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+int trinity_dpm_config(struct radeon_device *rdev, bool enable)
+{
+	if (enable)
+		WREG32_SMC(SMU_SCRATCH0, 1);
+	else
+		WREG32_SMC(SMU_SCRATCH0, 0);
+
+	return trinity_notify_message_to_smu(rdev, PPSMC_MSG_DPM_Config);
+}
+
+int trinity_dpm_force_state(struct radeon_device *rdev, u32 n)
+{
+	WREG32_SMC(SMU_SCRATCH0, n);
+
+	return trinity_notify_message_to_smu(rdev, PPSMC_MSG_DPM_ForceState);
+}
+
+int trinity_dpm_n_levels_disabled(struct radeon_device *rdev, u32 n)
+{
+	WREG32_SMC(SMU_SCRATCH0, n);
+
+	return trinity_notify_message_to_smu(rdev, PPSMC_MSG_DPM_N_LevelsDisabled);
+}
+
+int trinity_uvd_dpm_config(struct radeon_device *rdev)
+{
+	return trinity_notify_message_to_smu(rdev, PPSMC_MSG_UVD_DPM_Config);
+}
+
+int trinity_dpm_no_forced_level(struct radeon_device *rdev)
+{
+	return trinity_notify_message_to_smu(rdev, PPSMC_MSG_NoForcedLevel);
+}
+
+int trinity_dce_enable_voltage_adjustment(struct radeon_device *rdev,
+					  bool enable)
+{
+	if (enable)
+		return trinity_notify_message_to_smu(rdev, PPSMC_MSG_DCE_AllowVoltageAdjustment);
+	else
+		return trinity_notify_message_to_smu(rdev, PPSMC_MSG_DCE_RemoveVoltageAdjustment);
+}
+
+int trinity_gfx_dynamic_mgpg_config(struct radeon_device *rdev)
+{
+	return trinity_notify_message_to_smu(rdev, PPSMC_MSG_PG_SIMD_Config);
+}
+
+void trinity_acquire_mutex(struct radeon_device *rdev)
+{
+	int i;
+
+	WREG32(SMC_INT_REQ, 1);
+	for (i = 0; i < rdev->usec_timeout; i++) {
+		if ((RREG32(SMC_INT_REQ) & 0xffff) == 1)
+			break;
+		udelay(1);
+	}
+}
+
+void trinity_release_mutex(struct radeon_device *rdev)
+{
+	WREG32(SMC_INT_REQ, 0);
+}
diff --git a/drivers/gpu/drm/radeon/trinityd.h b/drivers/gpu/drm/radeon/trinityd.h
new file mode 100644
index 000000000000..fd32e2771755
--- /dev/null
+++ b/drivers/gpu/drm/radeon/trinityd.h
@@ -0,0 +1,228 @@
+/*
+ * Copyright 2012 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Alex Deucher
+ */
+#ifndef _TRINITYD_H_
+#define _TRINITYD_H_
+
+/* pm registers */
+
+/* cg */
+#define CG_CGTT_LOCAL_0                                 0x0
+#define CG_CGTT_LOCAL_1                                 0x1
+
+/* smc */
+#define SMU_SCLK_DPM_STATE_0_CNTL_0                     0x1f000
+#       define STATE_VALID(x)                           ((x) << 0)
+#       define STATE_VALID_MASK                         (0xff << 0)
+#       define STATE_VALID_SHIFT                        0
+#       define CLK_DIVIDER(x)                           ((x) << 8)
+#       define CLK_DIVIDER_MASK                         (0xff << 8)
+#       define CLK_DIVIDER_SHIFT                        8
+#       define VID(x)                                   ((x) << 16)
+#       define VID_MASK                                 (0xff << 16)
+#       define VID_SHIFT                                16
+#       define LVRT(x)                                  ((x) << 24)
+#       define LVRT_MASK                                (0xff << 24)
+#       define LVRT_SHIFT                               24
+#define SMU_SCLK_DPM_STATE_0_CNTL_1                     0x1f004
+#       define DS_DIV(x)                                ((x) << 0)
+#       define DS_DIV_MASK                              (0xff << 0)
+#       define DS_DIV_SHIFT                             0
+#       define DS_SH_DIV(x)                             ((x) << 8)
+#       define DS_SH_DIV_MASK                           (0xff << 8)
+#       define DS_SH_DIV_SHIFT                          8
+#       define DISPLAY_WM(x)                            ((x) << 16)
+#       define DISPLAY_WM_MASK                          (0xff << 16)
+#       define DISPLAY_WM_SHIFT                         16
+#       define VCE_WM(x)                                ((x) << 24)
+#       define VCE_WM_MASK                              (0xff << 24)
+#       define VCE_WM_SHIFT                             24
+
+#define SMU_SCLK_DPM_STATE_0_CNTL_3                     0x1f00c
+#       define GNB_SLOW(x)                              ((x) << 0)
+#       define GNB_SLOW_MASK                            (0xff << 0)
+#       define GNB_SLOW_SHIFT                           0
+#       define FORCE_NBPS1(x)                           ((x) << 8)
+#       define FORCE_NBPS1_MASK                         (0xff << 8)
+#       define FORCE_NBPS1_SHIFT                        8
+#define SMU_SCLK_DPM_STATE_0_AT                         0x1f010
+#       define AT(x)                                    ((x) << 0)
+#       define AT_MASK                                  (0xff << 0)
+#       define AT_SHIFT                                 0
+
+#define SMU_SCLK_DPM_STATE_0_PG_CNTL                    0x1f014
+#       define PD_SCLK_DIVIDER(x)                       ((x) << 16)
+#       define PD_SCLK_DIVIDER_MASK                     (0xff << 16)
+#       define PD_SCLK_DIVIDER_SHIFT                    16
+
+#define SMU_SCLK_DPM_STATE_1_CNTL_0                     0x1f020
+
+#define SMU_SCLK_DPM_CNTL                               0x1f100
+#       define SCLK_DPM_EN(x)                           ((x) << 0)
+#       define SCLK_DPM_EN_MASK                         (0xff << 0)
+#       define SCLK_DPM_EN_SHIFT                        0
+#       define SCLK_DPM_BOOT_STATE(x)                   ((x) << 16)
+#       define SCLK_DPM_BOOT_STATE_MASK                 (0xff << 16)
+#       define SCLK_DPM_BOOT_STATE_SHIFT                16
+#       define VOLTAGE_CHG_EN(x)                        ((x) << 24)
+#       define VOLTAGE_CHG_EN_MASK                      (0xff << 24)
+#       define VOLTAGE_CHG_EN_SHIFT                     24
+
+#define SMU_SCLK_DPM_TT_CNTL                            0x1f108
+#       define SCLK_TT_EN(x)                            ((x) << 0)
+#       define SCLK_TT_EN_MASK                          (0xff << 0)
+#       define SCLK_TT_EN_SHIFT                         0
+#define SMU_SCLK_DPM_TTT                                0x1f10c
+#       define LT(x)                                    ((x) << 0)
+#       define LT_MASK                                  (0xffff << 0)
+#       define LT_SHIFT                                 0
+#       define HT(x)                                    ((x) << 16)
+#       define HT_MASK                                  (0xffff << 16)
+#       define HT_SHIFT                                 16
+
+#define SMU_UVD_DPM_STATES                              0x1f1a0
+#define SMU_UVD_DPM_CNTL                                0x1f1a4
+
+#define SMU_S_PG_CNTL                                   0x1f118
+#       define DS_PG_EN(x)                              ((x) << 16)
+#       define DS_PG_EN_MASK                            (0xff << 16)
+#       define DS_PG_EN_SHIFT                           16
+
+#define GFX_POWER_GATING_CNTL                           0x1f38c
+#       define PDS_DIV(x)                               ((x) << 0)
+#       define PDS_DIV_MASK                             (0xff << 0)
+#       define PDS_DIV_SHIFT                            0
+#       define SSSD(x)                                  ((x) << 8)
+#       define SSSD_MASK                                (0xff << 8)
+#       define SSSD_SHIFT                               8
+
+#define PM_CONFIG                                       0x1f428
+#       define SVI_Mode                                 (1 << 29)
+
+#define PM_I_CNTL_1                                     0x1f464
+#       define SCLK_DPM(x)                              ((x) << 0)
+#       define SCLK_DPM_MASK                            (0xff << 0)
+#       define SCLK_DPM_SHIFT                           0
+#       define DS_PG_CNTL(x)                            ((x) << 16)
+#       define DS_PG_CNTL_MASK                          (0xff << 16)
+#       define DS_PG_CNTL_SHIFT                         16
+#define PM_TP                                           0x1f468
+
+#define NB_PSTATE_CONFIG                                0x1f5f8
+#       define Dpm0PgNbPsLo(x)                          ((x) << 0)
+#       define Dpm0PgNbPsLo_MASK                        (3 << 0)
+#       define Dpm0PgNbPsLo_SHIFT                       0
+#       define Dpm0PgNbPsHi(x)                          ((x) << 2)
+#       define Dpm0PgNbPsHi_MASK                        (3 << 2)
+#       define Dpm0PgNbPsHi_SHIFT                       2
+#       define DpmXNbPsLo(x)                            ((x) << 4)
+#       define DpmXNbPsLo_MASK                          (3 << 4)
+#       define DpmXNbPsLo_SHIFT                         4
+#       define DpmXNbPsHi(x)                            ((x) << 6)
+#       define DpmXNbPsHi_MASK                          (3 << 6)
+#       define DpmXNbPsHi_SHIFT                         6
+
+#define DC_CAC_VALUE                                    0x1f908
+
+#define GPU_CAC_AVRG_CNTL                               0x1f920
+#       define WINDOW_SIZE(x)                           ((x) << 0)
+#       define WINDOW_SIZE_MASK                         (0xff << 0)
+#       define WINDOW_SIZE_SHIFT                        0
+
+#define CC_SMU_MISC_FUSES                               0xe0001004
+#       define MinSClkDid(x)                   ((x) << 2)
+#       define MinSClkDid_MASK                 (0x7f << 2)
+#       define MinSClkDid_SHIFT                2
+
+#define CC_SMU_TST_EFUSE1_MISC                          0xe000101c
+#       define RB_BACKEND_DISABLE(x)                    ((x) << 16)
+#       define RB_BACKEND_DISABLE_MASK                  (3 << 16)
+#       define RB_BACKEND_DISABLE_SHIFT                 16
+
+#define SMU_SCRATCH_A                                   0xe0003024
+
+#define SMU_SCRATCH0                                    0xe0003040
+
+/* mmio */
+#define SMC_INT_REQ                                     0x220
+
+#define SMC_MESSAGE_0                                   0x22c
+#define SMC_RESP_0                                      0x230
+
+#define GENERAL_PWRMGT                                  0x670
+#       define GLOBAL_PWRMGT_EN                         (1 << 0)
+
+#define SCLK_PWRMGT_CNTL                                0x678
+#       define DYN_PWR_DOWN_EN                          (1 << 2)
+#       define RESET_BUSY_CNT                           (1 << 4)
+#       define RESET_SCLK_CNT                           (1 << 5)
+#       define DYN_GFX_CLK_OFF_EN                       (1 << 7)
+#       define GFX_CLK_FORCE_ON                         (1 << 8)
+#       define DYNAMIC_PM_EN                            (1 << 21)
+
+#define TARGET_AND_CURRENT_PROFILE_INDEX                0x684
+#       define TARGET_STATE(x)                          ((x) << 0)
+#       define TARGET_STATE_MASK                        (0xf << 0)
+#       define TARGET_STATE_SHIFT                       0
+#       define CURRENT_STATE(x)                         ((x) << 4)
+#       define CURRENT_STATE_MASK                       (0xf << 4)
+#       define CURRENT_STATE_SHIFT                      4
+
+#define CG_GIPOTS                                       0x6d8
+#       define CG_GIPOT(x)                              ((x) << 16)
+#       define CG_GIPOT_MASK                            (0xffff << 16)
+#       define CG_GIPOT_SHIFT                           16
+
+#define CG_PG_CTRL                                      0x6e0
+#       define SP(x)                                    ((x) << 0)
+#       define SP_MASK                                  (0xffff << 0)
+#       define SP_SHIFT                                 0
+#       define SU(x)                                    ((x) << 16)
+#       define SU_MASK                                  (0xffff << 16)
+#       define SU_SHIFT                                 16
+
+#define CG_MISC_REG                                     0x708
+
+#define CG_THERMAL_INT_CTRL                             0x738
+#       define DIG_THERM_INTH(x)                        ((x) << 0)
+#       define DIG_THERM_INTH_MASK                      (0xff << 0)
+#       define DIG_THERM_INTH_SHIFT                     0
+#       define DIG_THERM_INTL(x)                        ((x) << 8)
+#       define DIG_THERM_INTL_MASK                      (0xff << 8)
+#       define DIG_THERM_INTL_SHIFT                     8
+#       define THERM_INTH_MASK                          (1 << 24)
+#       define THERM_INTL_MASK                          (1 << 25)
+
+#define CG_CG_VOLTAGE_CNTL                              0x770
+#       define EN                                       (1 << 9)
+
+#define HW_REV   					0x5564
+#       define ATI_REV_ID_MASK                          (0xf << 28)
+#       define ATI_REV_ID_SHIFT                         28
+/* 0 = A0, 1 = A1, 2 = B0, 3 = C0, etc. */
+
+#define CGTS_SM_CTRL_REG                                0x9150
+
+#define GB_ADDR_CONFIG                                  0x98f8
+
+#endif
diff --git a/drivers/gpu/drm/rcar-du/Kconfig b/drivers/gpu/drm/rcar-du/Kconfig
new file mode 100644
index 000000000000..72887df8dd76
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/Kconfig
@@ -0,0 +1,9 @@
+config DRM_RCAR_DU
+	tristate "DRM Support for R-Car Display Unit"
+	depends on DRM && ARM
+	select DRM_KMS_HELPER
+	select DRM_KMS_CMA_HELPER
+	select DRM_GEM_CMA_HELPER
+	help
+	  Choose this option if you have an R-Car chipset.
+	  If M is selected the module will be called rcar-du-drm.
diff --git a/drivers/gpu/drm/rcar-du/Makefile b/drivers/gpu/drm/rcar-du/Makefile
new file mode 100644
index 000000000000..7333c0094015
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/Makefile
@@ -0,0 +1,8 @@
+rcar-du-drm-y := rcar_du_crtc.o \
+		 rcar_du_drv.o \
+		 rcar_du_kms.o \
+		 rcar_du_lvds.o \
+		 rcar_du_plane.o \
+		 rcar_du_vga.o
+
+obj-$(CONFIG_DRM_RCAR_DU)	+= rcar-du-drm.o
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_crtc.c b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c
new file mode 100644
index 000000000000..24183fb93592
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c
@@ -0,0 +1,595 @@
+/*
+ * rcar_du_crtc.c  --  R-Car Display Unit CRTCs
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/clk.h>
+#include <linux/mutex.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+
+#include "rcar_du_crtc.h"
+#include "rcar_du_drv.h"
+#include "rcar_du_kms.h"
+#include "rcar_du_lvds.h"
+#include "rcar_du_plane.h"
+#include "rcar_du_regs.h"
+#include "rcar_du_vga.h"
+
+#define to_rcar_crtc(c)	container_of(c, struct rcar_du_crtc, crtc)
+
+static u32 rcar_du_crtc_read(struct rcar_du_crtc *rcrtc, u32 reg)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+
+	return rcar_du_read(rcdu, rcrtc->mmio_offset + reg);
+}
+
+static void rcar_du_crtc_write(struct rcar_du_crtc *rcrtc, u32 reg, u32 data)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+
+	rcar_du_write(rcdu, rcrtc->mmio_offset + reg, data);
+}
+
+static void rcar_du_crtc_clr(struct rcar_du_crtc *rcrtc, u32 reg, u32 clr)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+
+	rcar_du_write(rcdu, rcrtc->mmio_offset + reg,
+		      rcar_du_read(rcdu, rcrtc->mmio_offset + reg) & ~clr);
+}
+
+static void rcar_du_crtc_set(struct rcar_du_crtc *rcrtc, u32 reg, u32 set)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+
+	rcar_du_write(rcdu, rcrtc->mmio_offset + reg,
+		      rcar_du_read(rcdu, rcrtc->mmio_offset + reg) | set);
+}
+
+static void rcar_du_crtc_clr_set(struct rcar_du_crtc *rcrtc, u32 reg,
+				 u32 clr, u32 set)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+	u32 value = rcar_du_read(rcdu, rcrtc->mmio_offset + reg);
+
+	rcar_du_write(rcdu, rcrtc->mmio_offset + reg, (value & ~clr) | set);
+}
+
+static void rcar_du_crtc_set_display_timing(struct rcar_du_crtc *rcrtc)
+{
+	struct drm_crtc *crtc = &rcrtc->crtc;
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+	const struct drm_display_mode *mode = &crtc->mode;
+	unsigned long clk;
+	u32 value;
+	u32 div;
+
+	/* Dot clock */
+	clk = clk_get_rate(rcdu->clock);
+	div = DIV_ROUND_CLOSEST(clk, mode->clock * 1000);
+	div = clamp(div, 1U, 64U) - 1;
+
+	rcar_du_write(rcdu, rcrtc->index ? ESCR2 : ESCR,
+		      ESCR_DCLKSEL_CLKS | div);
+	rcar_du_write(rcdu, rcrtc->index ? OTAR2 : OTAR, 0);
+
+	/* Signal polarities */
+	value = ((mode->flags & DRM_MODE_FLAG_PVSYNC) ? 0 : DSMR_VSL)
+	      | ((mode->flags & DRM_MODE_FLAG_PHSYNC) ? 0 : DSMR_HSL)
+	      | DSMR_DIPM_DE;
+	rcar_du_crtc_write(rcrtc, DSMR, value);
+
+	/* Display timings */
+	rcar_du_crtc_write(rcrtc, HDSR, mode->htotal - mode->hsync_start - 19);
+	rcar_du_crtc_write(rcrtc, HDER, mode->htotal - mode->hsync_start +
+					mode->hdisplay - 19);
+	rcar_du_crtc_write(rcrtc, HSWR, mode->hsync_end -
+					mode->hsync_start - 1);
+	rcar_du_crtc_write(rcrtc, HCR,  mode->htotal - 1);
+
+	rcar_du_crtc_write(rcrtc, VDSR, mode->vtotal - mode->vsync_end - 2);
+	rcar_du_crtc_write(rcrtc, VDER, mode->vtotal - mode->vsync_end +
+					mode->vdisplay - 2);
+	rcar_du_crtc_write(rcrtc, VSPR, mode->vtotal - mode->vsync_end +
+					mode->vsync_start - 1);
+	rcar_du_crtc_write(rcrtc, VCR,  mode->vtotal - 1);
+
+	rcar_du_crtc_write(rcrtc, DESR,  mode->htotal - mode->hsync_start);
+	rcar_du_crtc_write(rcrtc, DEWR,  mode->hdisplay);
+}
+
+static void rcar_du_crtc_set_routing(struct rcar_du_crtc *rcrtc)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+	u32 dorcr = rcar_du_read(rcdu, DORCR);
+
+	dorcr &= ~(DORCR_PG2T | DORCR_DK2S | DORCR_PG2D_MASK);
+
+	/* Set the DU1 pins sources. Select CRTC 0 if explicitly requested and
+	 * CRTC 1 in all other cases to avoid cloning CRTC 0 to DU0 and DU1 by
+	 * default.
+	 */
+	if (rcrtc->outputs & (1 << 1) && rcrtc->index == 0)
+		dorcr |= DORCR_PG2D_DS1;
+	else
+		dorcr |= DORCR_PG2T | DORCR_DK2S | DORCR_PG2D_DS2;
+
+	rcar_du_write(rcdu, DORCR, dorcr);
+}
+
+static void __rcar_du_start_stop(struct rcar_du_device *rcdu, bool start)
+{
+	rcar_du_write(rcdu, DSYSR,
+		      (rcar_du_read(rcdu, DSYSR) & ~(DSYSR_DRES | DSYSR_DEN)) |
+		      (start ? DSYSR_DEN : DSYSR_DRES));
+}
+
+static void rcar_du_start_stop(struct rcar_du_device *rcdu, bool start)
+{
+	/* Many of the configuration bits are only updated when the display
+	 * reset (DRES) bit in DSYSR is set to 1, disabling *both* CRTCs. Some
+	 * of those bits could be pre-configured, but others (especially the
+	 * bits related to plane assignment to display timing controllers) need
+	 * to be modified at runtime.
+	 *
+	 * Restart the display controller if a start is requested. Sorry for the
+	 * flicker. It should be possible to move most of the "DRES-update" bits
+	 * setup to driver initialization time and minimize the number of cases
+	 * when the display controller will have to be restarted.
+	 */
+	if (start) {
+		if (rcdu->used_crtcs++ != 0)
+			__rcar_du_start_stop(rcdu, false);
+		__rcar_du_start_stop(rcdu, true);
+	} else {
+		if (--rcdu->used_crtcs == 0)
+			__rcar_du_start_stop(rcdu, false);
+	}
+}
+
+void rcar_du_crtc_route_output(struct drm_crtc *crtc, unsigned int output)
+{
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+
+	/* Store the route from the CRTC output to the DU output. The DU will be
+	 * configured when starting the CRTC.
+	 */
+	rcrtc->outputs |= 1 << output;
+}
+
+void rcar_du_crtc_update_planes(struct drm_crtc *crtc)
+{
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+	struct rcar_du_plane *planes[RCAR_DU_NUM_HW_PLANES];
+	unsigned int num_planes = 0;
+	unsigned int prio = 0;
+	unsigned int i;
+	u32 dptsr = 0;
+	u32 dspr = 0;
+
+	for (i = 0; i < ARRAY_SIZE(rcdu->planes.planes); ++i) {
+		struct rcar_du_plane *plane = &rcdu->planes.planes[i];
+		unsigned int j;
+
+		if (plane->crtc != &rcrtc->crtc || !plane->enabled)
+			continue;
+
+		/* Insert the plane in the sorted planes array. */
+		for (j = num_planes++; j > 0; --j) {
+			if (planes[j-1]->zpos <= plane->zpos)
+				break;
+			planes[j] = planes[j-1];
+		}
+
+		planes[j] = plane;
+		prio += plane->format->planes * 4;
+	}
+
+	for (i = 0; i < num_planes; ++i) {
+		struct rcar_du_plane *plane = planes[i];
+		unsigned int index = plane->hwindex;
+
+		prio -= 4;
+		dspr |= (index + 1) << prio;
+		dptsr |= DPTSR_PnDK(index) |  DPTSR_PnTS(index);
+
+		if (plane->format->planes == 2) {
+			index = (index + 1) % 8;
+
+			prio -= 4;
+			dspr |= (index + 1) << prio;
+			dptsr |= DPTSR_PnDK(index) |  DPTSR_PnTS(index);
+		}
+	}
+
+	/* Select display timing and dot clock generator 2 for planes associated
+	 * with superposition controller 2.
+	 */
+	if (rcrtc->index) {
+		u32 value = rcar_du_read(rcdu, DPTSR);
+
+		/* The DPTSR register is updated when the display controller is
+		 * stopped. We thus need to restart the DU. Once again, sorry
+		 * for the flicker. One way to mitigate the issue would be to
+		 * pre-associate planes with CRTCs (either with a fixed 4/4
+		 * split, or through a module parameter). Flicker would then
+		 * occur only if we need to break the pre-association.
+		 */
+		if (value != dptsr) {
+			rcar_du_write(rcdu, DPTSR, dptsr);
+			if (rcdu->used_crtcs) {
+				__rcar_du_start_stop(rcdu, false);
+				__rcar_du_start_stop(rcdu, true);
+			}
+		}
+	}
+
+	rcar_du_write(rcdu, rcrtc->index ? DS2PR : DS1PR, dspr);
+}
+
+static void rcar_du_crtc_start(struct rcar_du_crtc *rcrtc)
+{
+	struct drm_crtc *crtc = &rcrtc->crtc;
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+	unsigned int i;
+
+	if (rcrtc->started)
+		return;
+
+	if (WARN_ON(rcrtc->plane->format == NULL))
+		return;
+
+	/* Set display off and background to black */
+	rcar_du_crtc_write(rcrtc, DOOR, DOOR_RGB(0, 0, 0));
+	rcar_du_crtc_write(rcrtc, BPOR, BPOR_RGB(0, 0, 0));
+
+	/* Configure display timings and output routing */
+	rcar_du_crtc_set_display_timing(rcrtc);
+	rcar_du_crtc_set_routing(rcrtc);
+
+	mutex_lock(&rcdu->planes.lock);
+	rcrtc->plane->enabled = true;
+	rcar_du_crtc_update_planes(crtc);
+	mutex_unlock(&rcdu->planes.lock);
+
+	/* Setup planes. */
+	for (i = 0; i < ARRAY_SIZE(rcdu->planes.planes); ++i) {
+		struct rcar_du_plane *plane = &rcdu->planes.planes[i];
+
+		if (plane->crtc != crtc || !plane->enabled)
+			continue;
+
+		rcar_du_plane_setup(plane);
+	}
+
+	/* Select master sync mode. This enables display operation in master
+	 * sync mode (with the HSYNC and VSYNC signals configured as outputs and
+	 * actively driven).
+	 */
+	rcar_du_crtc_clr_set(rcrtc, DSYSR, DSYSR_TVM_MASK, DSYSR_TVM_MASTER);
+
+	rcar_du_start_stop(rcdu, true);
+
+	rcrtc->started = true;
+}
+
+static void rcar_du_crtc_stop(struct rcar_du_crtc *rcrtc)
+{
+	struct drm_crtc *crtc = &rcrtc->crtc;
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+
+	if (!rcrtc->started)
+		return;
+
+	mutex_lock(&rcdu->planes.lock);
+	rcrtc->plane->enabled = false;
+	rcar_du_crtc_update_planes(crtc);
+	mutex_unlock(&rcdu->planes.lock);
+
+	/* Select switch sync mode. This stops display operation and configures
+	 * the HSYNC and VSYNC signals as inputs.
+	 */
+	rcar_du_crtc_clr_set(rcrtc, DSYSR, DSYSR_TVM_MASK, DSYSR_TVM_SWITCH);
+
+	rcar_du_start_stop(rcdu, false);
+
+	rcrtc->started = false;
+}
+
+void rcar_du_crtc_suspend(struct rcar_du_crtc *rcrtc)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+
+	rcar_du_crtc_stop(rcrtc);
+	rcar_du_put(rcdu);
+}
+
+void rcar_du_crtc_resume(struct rcar_du_crtc *rcrtc)
+{
+	struct rcar_du_device *rcdu = rcrtc->crtc.dev->dev_private;
+
+	if (rcrtc->dpms != DRM_MODE_DPMS_ON)
+		return;
+
+	rcar_du_get(rcdu);
+	rcar_du_crtc_start(rcrtc);
+}
+
+static void rcar_du_crtc_update_base(struct rcar_du_crtc *rcrtc)
+{
+	struct drm_crtc *crtc = &rcrtc->crtc;
+
+	rcar_du_plane_compute_base(rcrtc->plane, crtc->fb);
+	rcar_du_plane_update_base(rcrtc->plane);
+}
+
+static void rcar_du_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+
+	if (rcrtc->dpms == mode)
+		return;
+
+	if (mode == DRM_MODE_DPMS_ON) {
+		rcar_du_get(rcdu);
+		rcar_du_crtc_start(rcrtc);
+	} else {
+		rcar_du_crtc_stop(rcrtc);
+		rcar_du_put(rcdu);
+	}
+
+	rcrtc->dpms = mode;
+}
+
+static bool rcar_du_crtc_mode_fixup(struct drm_crtc *crtc,
+				    const struct drm_display_mode *mode,
+				    struct drm_display_mode *adjusted_mode)
+{
+	/* TODO Fixup modes */
+	return true;
+}
+
+static void rcar_du_crtc_mode_prepare(struct drm_crtc *crtc)
+{
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+
+	/* We need to access the hardware during mode set, acquire a reference
+	 * to the DU.
+	 */
+	rcar_du_get(rcdu);
+
+	/* Stop the CRTC and release the plane. Force the DPMS mode to off as a
+	 * result.
+	 */
+	rcar_du_crtc_stop(rcrtc);
+	rcar_du_plane_release(rcrtc->plane);
+
+	rcrtc->dpms = DRM_MODE_DPMS_OFF;
+}
+
+static int rcar_du_crtc_mode_set(struct drm_crtc *crtc,
+				 struct drm_display_mode *mode,
+				 struct drm_display_mode *adjusted_mode,
+				 int x, int y,
+				 struct drm_framebuffer *old_fb)
+{
+	struct rcar_du_device *rcdu = crtc->dev->dev_private;
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+	const struct rcar_du_format_info *format;
+	int ret;
+
+	format = rcar_du_format_info(crtc->fb->pixel_format);
+	if (format == NULL) {
+		dev_dbg(rcdu->dev, "mode_set: unsupported format %08x\n",
+			crtc->fb->pixel_format);
+		ret = -EINVAL;
+		goto error;
+	}
+
+	ret = rcar_du_plane_reserve(rcrtc->plane, format);
+	if (ret < 0)
+		goto error;
+
+	rcrtc->plane->format = format;
+	rcrtc->plane->pitch = crtc->fb->pitches[0];
+
+	rcrtc->plane->src_x = x;
+	rcrtc->plane->src_y = y;
+	rcrtc->plane->width = mode->hdisplay;
+	rcrtc->plane->height = mode->vdisplay;
+
+	rcar_du_plane_compute_base(rcrtc->plane, crtc->fb);
+
+	rcrtc->outputs = 0;
+
+	return 0;
+
+error:
+	/* There's no rollback/abort operation to clean up in case of error. We
+	 * thus need to release the reference to the DU acquired in prepare()
+	 * here.
+	 */
+	rcar_du_put(rcdu);
+	return ret;
+}
+
+static void rcar_du_crtc_mode_commit(struct drm_crtc *crtc)
+{
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+
+	/* We're done, restart the CRTC and set the DPMS mode to on. The
+	 * reference to the DU acquired at prepare() time will thus be released
+	 * by the DPMS handler (possibly called by the disable() handler).
+	 */
+	rcar_du_crtc_start(rcrtc);
+	rcrtc->dpms = DRM_MODE_DPMS_ON;
+}
+
+static int rcar_du_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
+				      struct drm_framebuffer *old_fb)
+{
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+
+	rcrtc->plane->src_x = x;
+	rcrtc->plane->src_y = y;
+
+	rcar_du_crtc_update_base(to_rcar_crtc(crtc));
+
+	return 0;
+}
+
+static void rcar_du_crtc_disable(struct drm_crtc *crtc)
+{
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+
+	rcar_du_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
+	rcar_du_plane_release(rcrtc->plane);
+}
+
+static const struct drm_crtc_helper_funcs crtc_helper_funcs = {
+	.dpms = rcar_du_crtc_dpms,
+	.mode_fixup = rcar_du_crtc_mode_fixup,
+	.prepare = rcar_du_crtc_mode_prepare,
+	.commit = rcar_du_crtc_mode_commit,
+	.mode_set = rcar_du_crtc_mode_set,
+	.mode_set_base = rcar_du_crtc_mode_set_base,
+	.disable = rcar_du_crtc_disable,
+};
+
+void rcar_du_crtc_cancel_page_flip(struct rcar_du_crtc *rcrtc,
+				   struct drm_file *file)
+{
+	struct drm_pending_vblank_event *event;
+	struct drm_device *dev = rcrtc->crtc.dev;
+	unsigned long flags;
+
+	/* Destroy the pending vertical blanking event associated with the
+	 * pending page flip, if any, and disable vertical blanking interrupts.
+	 */
+	spin_lock_irqsave(&dev->event_lock, flags);
+	event = rcrtc->event;
+	if (event && event->base.file_priv == file) {
+		rcrtc->event = NULL;
+		event->base.destroy(&event->base);
+		drm_vblank_put(dev, rcrtc->index);
+	}
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+}
+
+static void rcar_du_crtc_finish_page_flip(struct rcar_du_crtc *rcrtc)
+{
+	struct drm_pending_vblank_event *event;
+	struct drm_device *dev = rcrtc->crtc.dev;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev->event_lock, flags);
+	event = rcrtc->event;
+	rcrtc->event = NULL;
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+
+	if (event == NULL)
+		return;
+
+	spin_lock_irqsave(&dev->event_lock, flags);
+	drm_send_vblank_event(dev, rcrtc->index, event);
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+
+	drm_vblank_put(dev, rcrtc->index);
+}
+
+static int rcar_du_crtc_page_flip(struct drm_crtc *crtc,
+				  struct drm_framebuffer *fb,
+				  struct drm_pending_vblank_event *event)
+{
+	struct rcar_du_crtc *rcrtc = to_rcar_crtc(crtc);
+	struct drm_device *dev = rcrtc->crtc.dev;
+	unsigned long flags;
+
+	spin_lock_irqsave(&dev->event_lock, flags);
+	if (rcrtc->event != NULL) {
+		spin_unlock_irqrestore(&dev->event_lock, flags);
+		return -EBUSY;
+	}
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+
+	crtc->fb = fb;
+	rcar_du_crtc_update_base(rcrtc);
+
+	if (event) {
+		event->pipe = rcrtc->index;
+		drm_vblank_get(dev, rcrtc->index);
+		spin_lock_irqsave(&dev->event_lock, flags);
+		rcrtc->event = event;
+		spin_unlock_irqrestore(&dev->event_lock, flags);
+	}
+
+	return 0;
+}
+
+static const struct drm_crtc_funcs crtc_funcs = {
+	.destroy = drm_crtc_cleanup,
+	.set_config = drm_crtc_helper_set_config,
+	.page_flip = rcar_du_crtc_page_flip,
+};
+
+int rcar_du_crtc_create(struct rcar_du_device *rcdu, unsigned int index)
+{
+	struct rcar_du_crtc *rcrtc = &rcdu->crtcs[index];
+	struct drm_crtc *crtc = &rcrtc->crtc;
+	int ret;
+
+	rcrtc->mmio_offset = index ? DISP2_REG_OFFSET : 0;
+	rcrtc->index = index;
+	rcrtc->dpms = DRM_MODE_DPMS_OFF;
+	rcrtc->plane = &rcdu->planes.planes[index];
+
+	rcrtc->plane->crtc = crtc;
+
+	ret = drm_crtc_init(rcdu->ddev, crtc, &crtc_funcs);
+	if (ret < 0)
+		return ret;
+
+	drm_crtc_helper_add(crtc, &crtc_helper_funcs);
+
+	return 0;
+}
+
+void rcar_du_crtc_enable_vblank(struct rcar_du_crtc *rcrtc, bool enable)
+{
+	if (enable) {
+		rcar_du_crtc_write(rcrtc, DSRCR, DSRCR_VBCL);
+		rcar_du_crtc_set(rcrtc, DIER, DIER_VBE);
+	} else {
+		rcar_du_crtc_clr(rcrtc, DIER, DIER_VBE);
+	}
+}
+
+void rcar_du_crtc_irq(struct rcar_du_crtc *rcrtc)
+{
+	u32 status;
+
+	status = rcar_du_crtc_read(rcrtc, DSSR);
+	rcar_du_crtc_write(rcrtc, DSRCR, status & DSRCR_MASK);
+
+	if (status & DSSR_VBK) {
+		drm_handle_vblank(rcrtc->crtc.dev, rcrtc->index);
+		rcar_du_crtc_finish_page_flip(rcrtc);
+	}
+}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_crtc.h b/drivers/gpu/drm/rcar-du/rcar_du_crtc.h
new file mode 100644
index 000000000000..2a0365bcbd14
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_crtc.h
@@ -0,0 +1,50 @@
+/*
+ * rcar_du_crtc.h  --  R-Car Display Unit CRTCs
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_CRTC_H__
+#define __RCAR_DU_CRTC_H__
+
+#include <linux/mutex.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+
+struct rcar_du_device;
+struct rcar_du_plane;
+
+struct rcar_du_crtc {
+	struct drm_crtc crtc;
+
+	unsigned int mmio_offset;
+	unsigned int index;
+	bool started;
+
+	struct drm_pending_vblank_event *event;
+	unsigned int outputs;
+	int dpms;
+
+	struct rcar_du_plane *plane;
+};
+
+int rcar_du_crtc_create(struct rcar_du_device *rcdu, unsigned int index);
+void rcar_du_crtc_enable_vblank(struct rcar_du_crtc *rcrtc, bool enable);
+void rcar_du_crtc_irq(struct rcar_du_crtc *rcrtc);
+void rcar_du_crtc_cancel_page_flip(struct rcar_du_crtc *rcrtc,
+				   struct drm_file *file);
+void rcar_du_crtc_suspend(struct rcar_du_crtc *rcrtc);
+void rcar_du_crtc_resume(struct rcar_du_crtc *rcrtc);
+
+void rcar_du_crtc_route_output(struct drm_crtc *crtc, unsigned int output);
+void rcar_du_crtc_update_planes(struct drm_crtc *crtc);
+
+#endif /* __RCAR_DU_CRTC_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_drv.c b/drivers/gpu/drm/rcar-du/rcar_du_drv.c
new file mode 100644
index 000000000000..dc0fe09b2ba1
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_drv.c
@@ -0,0 +1,330 @@
+/*
+ * rcar_du_drv.c  --  R-Car Display Unit DRM driver
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <linux/clk.h>
+#include <linux/io.h>
+#include <linux/mm.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/pm.h>
+#include <linux/slab.h>
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+
+#include "rcar_du_crtc.h"
+#include "rcar_du_drv.h"
+#include "rcar_du_kms.h"
+#include "rcar_du_regs.h"
+
+/* -----------------------------------------------------------------------------
+ * Core device operations
+ */
+
+/*
+ * rcar_du_get - Acquire a reference to the DU
+ *
+ * Acquiring a reference enables the device clock and setup core registers. A
+ * reference must be held before accessing any hardware registers.
+ *
+ * This function must be called with the DRM mode_config lock held.
+ *
+ * Return 0 in case of success or a negative error code otherwise.
+ */
+int rcar_du_get(struct rcar_du_device *rcdu)
+{
+	int ret;
+
+	if (rcdu->use_count)
+		goto done;
+
+	/* Enable clocks before accessing the hardware. */
+	ret = clk_prepare_enable(rcdu->clock);
+	if (ret < 0)
+		return ret;
+
+	/* Enable extended features */
+	rcar_du_write(rcdu, DEFR, DEFR_CODE | DEFR_DEFE);
+	rcar_du_write(rcdu, DEFR2, DEFR2_CODE | DEFR2_DEFE2G);
+	rcar_du_write(rcdu, DEFR3, DEFR3_CODE | DEFR3_DEFE3);
+	rcar_du_write(rcdu, DEFR4, DEFR4_CODE);
+	rcar_du_write(rcdu, DEFR5, DEFR5_CODE | DEFR5_DEFE5);
+
+	/* Use DS1PR and DS2PR to configure planes priorities and connects the
+	 * superposition 0 to DU0 pins. DU1 pins will be configured dynamically.
+	 */
+	rcar_du_write(rcdu, DORCR, DORCR_PG1D_DS1 | DORCR_DPRS);
+
+done:
+	rcdu->use_count++;
+	return 0;
+}
+
+/*
+ * rcar_du_put - Release a reference to the DU
+ *
+ * Releasing the last reference disables the device clock.
+ *
+ * This function must be called with the DRM mode_config lock held.
+ */
+void rcar_du_put(struct rcar_du_device *rcdu)
+{
+	if (--rcdu->use_count)
+		return;
+
+	clk_disable_unprepare(rcdu->clock);
+}
+
+/* -----------------------------------------------------------------------------
+ * DRM operations
+ */
+
+static int rcar_du_unload(struct drm_device *dev)
+{
+	drm_kms_helper_poll_fini(dev);
+	drm_mode_config_cleanup(dev);
+	drm_vblank_cleanup(dev);
+	drm_irq_uninstall(dev);
+
+	dev->dev_private = NULL;
+
+	return 0;
+}
+
+static int rcar_du_load(struct drm_device *dev, unsigned long flags)
+{
+	struct platform_device *pdev = dev->platformdev;
+	struct rcar_du_platform_data *pdata = pdev->dev.platform_data;
+	struct rcar_du_device *rcdu;
+	struct resource *ioarea;
+	struct resource *mem;
+	int ret;
+
+	if (pdata == NULL) {
+		dev_err(dev->dev, "no platform data\n");
+		return -ENODEV;
+	}
+
+	rcdu = devm_kzalloc(&pdev->dev, sizeof(*rcdu), GFP_KERNEL);
+	if (rcdu == NULL) {
+		dev_err(dev->dev, "failed to allocate private data\n");
+		return -ENOMEM;
+	}
+
+	rcdu->dev = &pdev->dev;
+	rcdu->pdata = pdata;
+	rcdu->ddev = dev;
+	dev->dev_private = rcdu;
+
+	/* I/O resources and clocks */
+	mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (mem == NULL) {
+		dev_err(&pdev->dev, "failed to get memory resource\n");
+		return -EINVAL;
+	}
+
+	ioarea = devm_request_mem_region(&pdev->dev, mem->start,
+					 resource_size(mem), pdev->name);
+	if (ioarea == NULL) {
+		dev_err(&pdev->dev, "failed to request memory region\n");
+		return -EBUSY;
+	}
+
+	rcdu->mmio = devm_ioremap_nocache(&pdev->dev, ioarea->start,
+					  resource_size(ioarea));
+	if (rcdu->mmio == NULL) {
+		dev_err(&pdev->dev, "failed to remap memory resource\n");
+		return -ENOMEM;
+	}
+
+	rcdu->clock = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(rcdu->clock)) {
+		dev_err(&pdev->dev, "failed to get clock\n");
+		return -ENOENT;
+	}
+
+	/* DRM/KMS objects */
+	ret = rcar_du_modeset_init(rcdu);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "failed to initialize DRM/KMS\n");
+		goto done;
+	}
+
+	/* IRQ and vblank handling */
+	ret = drm_vblank_init(dev, (1 << rcdu->num_crtcs) - 1);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "failed to initialize vblank\n");
+		goto done;
+	}
+
+	ret = drm_irq_install(dev);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "failed to install IRQ handler\n");
+		goto done;
+	}
+
+	platform_set_drvdata(pdev, rcdu);
+
+done:
+	if (ret)
+		rcar_du_unload(dev);
+
+	return ret;
+}
+
+static void rcar_du_preclose(struct drm_device *dev, struct drm_file *file)
+{
+	struct rcar_du_device *rcdu = dev->dev_private;
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(rcdu->crtcs); ++i)
+		rcar_du_crtc_cancel_page_flip(&rcdu->crtcs[i], file);
+}
+
+static irqreturn_t rcar_du_irq(int irq, void *arg)
+{
+	struct drm_device *dev = arg;
+	struct rcar_du_device *rcdu = dev->dev_private;
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(rcdu->crtcs); ++i)
+		rcar_du_crtc_irq(&rcdu->crtcs[i]);
+
+	return IRQ_HANDLED;
+}
+
+static int rcar_du_enable_vblank(struct drm_device *dev, int crtc)
+{
+	struct rcar_du_device *rcdu = dev->dev_private;
+
+	rcar_du_crtc_enable_vblank(&rcdu->crtcs[crtc], true);
+
+	return 0;
+}
+
+static void rcar_du_disable_vblank(struct drm_device *dev, int crtc)
+{
+	struct rcar_du_device *rcdu = dev->dev_private;
+
+	rcar_du_crtc_enable_vblank(&rcdu->crtcs[crtc], false);
+}
+
+static const struct file_operations rcar_du_fops = {
+	.owner		= THIS_MODULE,
+	.open		= drm_open,
+	.release	= drm_release,
+	.unlocked_ioctl	= drm_ioctl,
+#ifdef CONFIG_COMPAT
+	.compat_ioctl	= drm_compat_ioctl,
+#endif
+	.poll		= drm_poll,
+	.read		= drm_read,
+	.fasync		= drm_fasync,
+	.llseek		= no_llseek,
+	.mmap		= drm_gem_cma_mmap,
+};
+
+static struct drm_driver rcar_du_driver = {
+	.driver_features	= DRIVER_HAVE_IRQ | DRIVER_GEM | DRIVER_MODESET
+				| DRIVER_PRIME,
+	.load			= rcar_du_load,
+	.unload			= rcar_du_unload,
+	.preclose		= rcar_du_preclose,
+	.irq_handler		= rcar_du_irq,
+	.get_vblank_counter	= drm_vblank_count,
+	.enable_vblank		= rcar_du_enable_vblank,
+	.disable_vblank		= rcar_du_disable_vblank,
+	.gem_free_object	= drm_gem_cma_free_object,
+	.gem_vm_ops		= &drm_gem_cma_vm_ops,
+	.prime_handle_to_fd	= drm_gem_prime_handle_to_fd,
+	.prime_fd_to_handle	= drm_gem_prime_fd_to_handle,
+	.gem_prime_import	= drm_gem_prime_import,
+	.gem_prime_export	= drm_gem_prime_export,
+	.gem_prime_get_sg_table	= drm_gem_cma_prime_get_sg_table,
+	.gem_prime_import_sg_table = drm_gem_cma_prime_import_sg_table,
+	.gem_prime_vmap		= drm_gem_cma_prime_vmap,
+	.gem_prime_vunmap	= drm_gem_cma_prime_vunmap,
+	.gem_prime_mmap		= drm_gem_cma_prime_mmap,
+	.dumb_create		= rcar_du_dumb_create,
+	.dumb_map_offset	= drm_gem_cma_dumb_map_offset,
+	.dumb_destroy		= drm_gem_cma_dumb_destroy,
+	.fops			= &rcar_du_fops,
+	.name			= "rcar-du",
+	.desc			= "Renesas R-Car Display Unit",
+	.date			= "20130110",
+	.major			= 1,
+	.minor			= 0,
+};
+
+/* -----------------------------------------------------------------------------
+ * Power management
+ */
+
+#if CONFIG_PM_SLEEP
+static int rcar_du_pm_suspend(struct device *dev)
+{
+	struct rcar_du_device *rcdu = dev_get_drvdata(dev);
+
+	drm_kms_helper_poll_disable(rcdu->ddev);
+	/* TODO Suspend the CRTC */
+
+	return 0;
+}
+
+static int rcar_du_pm_resume(struct device *dev)
+{
+	struct rcar_du_device *rcdu = dev_get_drvdata(dev);
+
+	/* TODO Resume the CRTC */
+
+	drm_kms_helper_poll_enable(rcdu->ddev);
+	return 0;
+}
+#endif
+
+static const struct dev_pm_ops rcar_du_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(rcar_du_pm_suspend, rcar_du_pm_resume)
+};
+
+/* -----------------------------------------------------------------------------
+ * Platform driver
+ */
+
+static int rcar_du_probe(struct platform_device *pdev)
+{
+	return drm_platform_init(&rcar_du_driver, pdev);
+}
+
+static int rcar_du_remove(struct platform_device *pdev)
+{
+	drm_platform_exit(&rcar_du_driver, pdev);
+
+	return 0;
+}
+
+static struct platform_driver rcar_du_platform_driver = {
+	.probe		= rcar_du_probe,
+	.remove		= rcar_du_remove,
+	.driver		= {
+		.owner	= THIS_MODULE,
+		.name	= "rcar-du",
+		.pm	= &rcar_du_pm_ops,
+	},
+};
+
+module_platform_driver(rcar_du_platform_driver);
+
+MODULE_AUTHOR("Laurent Pinchart <laurent.pinchart@ideasonboard.com>");
+MODULE_DESCRIPTION("Renesas R-Car Display Unit DRM Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_drv.h b/drivers/gpu/drm/rcar-du/rcar_du_drv.h
new file mode 100644
index 000000000000..193cc59d495c
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_drv.h
@@ -0,0 +1,66 @@
+/*
+ * rcar_du_drv.h  --  R-Car Display Unit DRM driver
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_DRV_H__
+#define __RCAR_DU_DRV_H__
+
+#include <linux/kernel.h>
+#include <linux/mutex.h>
+#include <linux/platform_data/rcar-du.h>
+
+#include "rcar_du_crtc.h"
+#include "rcar_du_plane.h"
+
+struct clk;
+struct device;
+struct drm_device;
+
+struct rcar_du_device {
+	struct device *dev;
+	const struct rcar_du_platform_data *pdata;
+
+	void __iomem *mmio;
+	struct clk *clock;
+	unsigned int use_count;
+
+	struct drm_device *ddev;
+
+	struct rcar_du_crtc crtcs[2];
+	unsigned int used_crtcs;
+	unsigned int num_crtcs;
+
+	struct {
+		struct rcar_du_plane planes[RCAR_DU_NUM_SW_PLANES];
+		unsigned int free;
+		struct mutex lock;
+
+		struct drm_property *alpha;
+		struct drm_property *colorkey;
+		struct drm_property *zpos;
+	} planes;
+};
+
+int rcar_du_get(struct rcar_du_device *rcdu);
+void rcar_du_put(struct rcar_du_device *rcdu);
+
+static inline u32 rcar_du_read(struct rcar_du_device *rcdu, u32 reg)
+{
+	return ioread32(rcdu->mmio + reg);
+}
+
+static inline void rcar_du_write(struct rcar_du_device *rcdu, u32 reg, u32 data)
+{
+	iowrite32(data, rcdu->mmio + reg);
+}
+
+#endif /* __RCAR_DU_DRV_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
new file mode 100644
index 000000000000..d30c2e29bee2
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
@@ -0,0 +1,265 @@
+/*
+ * rcar_du_kms.c  --  R-Car Display Unit Mode Setting
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+
+#include "rcar_du_crtc.h"
+#include "rcar_du_drv.h"
+#include "rcar_du_kms.h"
+#include "rcar_du_lvds.h"
+#include "rcar_du_regs.h"
+#include "rcar_du_vga.h"
+
+/* -----------------------------------------------------------------------------
+ * Format helpers
+ */
+
+static const struct rcar_du_format_info rcar_du_format_infos[] = {
+	{
+		.fourcc = DRM_FORMAT_RGB565,
+		.bpp = 16,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_TP | PnMR_DDDF_16BPP,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		.fourcc = DRM_FORMAT_ARGB1555,
+		.bpp = 16,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_ALP | PnMR_DDDF_ARGB,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		.fourcc = DRM_FORMAT_XRGB1555,
+		.bpp = 16,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_ALP | PnMR_DDDF_ARGB,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		.fourcc = DRM_FORMAT_XRGB8888,
+		.bpp = 32,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_TP | PnMR_DDDF_16BPP,
+		.edf = PnDDCR4_EDF_RGB888,
+	}, {
+		.fourcc = DRM_FORMAT_ARGB8888,
+		.bpp = 32,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_ALP | PnMR_DDDF_16BPP,
+		.edf = PnDDCR4_EDF_ARGB8888,
+	}, {
+		.fourcc = DRM_FORMAT_UYVY,
+		.bpp = 16,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_TP_OFF | PnMR_DDDF_YC,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		.fourcc = DRM_FORMAT_YUYV,
+		.bpp = 16,
+		.planes = 1,
+		.pnmr = PnMR_SPIM_TP_OFF | PnMR_DDDF_YC,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		.fourcc = DRM_FORMAT_NV12,
+		.bpp = 12,
+		.planes = 2,
+		.pnmr = PnMR_SPIM_TP_OFF | PnMR_DDDF_YC,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		.fourcc = DRM_FORMAT_NV21,
+		.bpp = 12,
+		.planes = 2,
+		.pnmr = PnMR_SPIM_TP_OFF | PnMR_DDDF_YC,
+		.edf = PnDDCR4_EDF_NONE,
+	}, {
+		/* In YUV 4:2:2, only NV16 is supported (NV61 isn't) */
+		.fourcc = DRM_FORMAT_NV16,
+		.bpp = 16,
+		.planes = 2,
+		.pnmr = PnMR_SPIM_TP_OFF | PnMR_DDDF_YC,
+		.edf = PnDDCR4_EDF_NONE,
+	},
+};
+
+const struct rcar_du_format_info *rcar_du_format_info(u32 fourcc)
+{
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(rcar_du_format_infos); ++i) {
+		if (rcar_du_format_infos[i].fourcc == fourcc)
+			return &rcar_du_format_infos[i];
+	}
+
+	return NULL;
+}
+
+/* -----------------------------------------------------------------------------
+ * Common connector and encoder functions
+ */
+
+struct drm_encoder *
+rcar_du_connector_best_encoder(struct drm_connector *connector)
+{
+	struct rcar_du_connector *rcon = to_rcar_connector(connector);
+
+	return &rcon->encoder->encoder;
+}
+
+void rcar_du_encoder_mode_prepare(struct drm_encoder *encoder)
+{
+}
+
+void rcar_du_encoder_mode_set(struct drm_encoder *encoder,
+			      struct drm_display_mode *mode,
+			      struct drm_display_mode *adjusted_mode)
+{
+	struct rcar_du_encoder *renc = to_rcar_encoder(encoder);
+
+	rcar_du_crtc_route_output(encoder->crtc, renc->output);
+}
+
+void rcar_du_encoder_mode_commit(struct drm_encoder *encoder)
+{
+}
+
+/* -----------------------------------------------------------------------------
+ * Frame buffer
+ */
+
+int rcar_du_dumb_create(struct drm_file *file, struct drm_device *dev,
+			struct drm_mode_create_dumb *args)
+{
+	unsigned int min_pitch = DIV_ROUND_UP(args->width * args->bpp, 8);
+	unsigned int align;
+
+	/* The pitch must be aligned to a 16 pixels boundary. */
+	align = 16 * args->bpp / 8;
+	args->pitch = roundup(max(args->pitch, min_pitch), align);
+
+	return drm_gem_cma_dumb_create(file, dev, args);
+}
+
+static struct drm_framebuffer *
+rcar_du_fb_create(struct drm_device *dev, struct drm_file *file_priv,
+		  struct drm_mode_fb_cmd2 *mode_cmd)
+{
+	const struct rcar_du_format_info *format;
+	unsigned int align;
+
+	format = rcar_du_format_info(mode_cmd->pixel_format);
+	if (format == NULL) {
+		dev_dbg(dev->dev, "unsupported pixel format %08x\n",
+			mode_cmd->pixel_format);
+		return ERR_PTR(-EINVAL);
+	}
+
+	align = 16 * format->bpp / 8;
+
+	if (mode_cmd->pitches[0] & (align - 1) ||
+	    mode_cmd->pitches[0] >= 8192) {
+		dev_dbg(dev->dev, "invalid pitch value %u\n",
+			mode_cmd->pitches[0]);
+		return ERR_PTR(-EINVAL);
+	}
+
+	if (format->planes == 2) {
+		if (mode_cmd->pitches[1] != mode_cmd->pitches[0]) {
+			dev_dbg(dev->dev,
+				"luma and chroma pitches do not match\n");
+			return ERR_PTR(-EINVAL);
+		}
+	}
+
+	return drm_fb_cma_create(dev, file_priv, mode_cmd);
+}
+
+static const struct drm_mode_config_funcs rcar_du_mode_config_funcs = {
+	.fb_create = rcar_du_fb_create,
+};
+
+int rcar_du_modeset_init(struct rcar_du_device *rcdu)
+{
+	struct drm_device *dev = rcdu->ddev;
+	struct drm_encoder *encoder;
+	unsigned int i;
+	int ret;
+
+	drm_mode_config_init(rcdu->ddev);
+
+	rcdu->ddev->mode_config.min_width = 0;
+	rcdu->ddev->mode_config.min_height = 0;
+	rcdu->ddev->mode_config.max_width = 4095;
+	rcdu->ddev->mode_config.max_height = 2047;
+	rcdu->ddev->mode_config.funcs = &rcar_du_mode_config_funcs;
+
+	ret = rcar_du_plane_init(rcdu);
+	if (ret < 0)
+		return ret;
+
+	for (i = 0; i < ARRAY_SIZE(rcdu->crtcs); ++i) {
+		ret = rcar_du_crtc_create(rcdu, i);
+		if (ret < 0)
+			return ret;
+	}
+
+	rcdu->used_crtcs = 0;
+	rcdu->num_crtcs = i;
+
+	for (i = 0; i < rcdu->pdata->num_encoders; ++i) {
+		const struct rcar_du_encoder_data *pdata =
+			&rcdu->pdata->encoders[i];
+
+		if (pdata->output >= ARRAY_SIZE(rcdu->crtcs)) {
+			dev_warn(rcdu->dev,
+				 "encoder %u references unexisting output %u, skipping\n",
+				 i, pdata->output);
+			continue;
+		}
+
+		switch (pdata->encoder) {
+		case RCAR_DU_ENCODER_VGA:
+			rcar_du_vga_init(rcdu, &pdata->u.vga, pdata->output);
+			break;
+
+		case RCAR_DU_ENCODER_LVDS:
+			rcar_du_lvds_init(rcdu, &pdata->u.lvds, pdata->output);
+			break;
+
+		default:
+			break;
+		}
+	}
+
+	/* Set the possible CRTCs and possible clones. All encoders can be
+	 * driven by the CRTC associated with the output they're connected to,
+	 * as well as by CRTC 0.
+	 */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		struct rcar_du_encoder *renc = to_rcar_encoder(encoder);
+
+		encoder->possible_crtcs = (1 << 0) | (1 << renc->output);
+		encoder->possible_clones = 1 << 0;
+	}
+
+	ret = rcar_du_plane_register(rcdu);
+	if (ret < 0)
+		return ret;
+
+	drm_kms_helper_poll_init(rcdu->ddev);
+
+	drm_helper_disable_unused_functions(rcdu->ddev);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.h b/drivers/gpu/drm/rcar-du/rcar_du_kms.h
new file mode 100644
index 000000000000..dba472263486
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.h
@@ -0,0 +1,62 @@
+/*
+ * rcar_du_kms.h  --  R-Car Display Unit Mode Setting
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_KMS_H__
+#define __RCAR_DU_KMS_H__
+
+#include <linux/types.h>
+
+#include <drm/drm_crtc.h>
+
+struct rcar_du_device;
+
+struct rcar_du_format_info {
+	u32 fourcc;
+	unsigned int bpp;
+	unsigned int planes;
+	unsigned int pnmr;
+	unsigned int edf;
+};
+
+struct rcar_du_encoder {
+	struct drm_encoder encoder;
+	unsigned int output;
+};
+
+#define to_rcar_encoder(e) \
+	container_of(e, struct rcar_du_encoder, encoder)
+
+struct rcar_du_connector {
+	struct drm_connector connector;
+	struct rcar_du_encoder *encoder;
+};
+
+#define to_rcar_connector(c) \
+	container_of(c, struct rcar_du_connector, connector)
+
+const struct rcar_du_format_info *rcar_du_format_info(u32 fourcc);
+
+struct drm_encoder *
+rcar_du_connector_best_encoder(struct drm_connector *connector);
+void rcar_du_encoder_mode_prepare(struct drm_encoder *encoder);
+void rcar_du_encoder_mode_set(struct drm_encoder *encoder,
+			      struct drm_display_mode *mode,
+			      struct drm_display_mode *adjusted_mode);
+void rcar_du_encoder_mode_commit(struct drm_encoder *encoder);
+
+int rcar_du_modeset_init(struct rcar_du_device *rcdu);
+
+int rcar_du_dumb_create(struct drm_file *file, struct drm_device *dev,
+			struct drm_mode_create_dumb *args);
+
+#endif /* __RCAR_DU_KMS_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_lvds.c b/drivers/gpu/drm/rcar-du/rcar_du_lvds.c
new file mode 100644
index 000000000000..7aefe7267e1d
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_lvds.c
@@ -0,0 +1,216 @@
+/*
+ * rcar_du_lvds.c  --  R-Car Display Unit LVDS Encoder and Connector
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+
+#include "rcar_du_drv.h"
+#include "rcar_du_kms.h"
+#include "rcar_du_lvds.h"
+
+struct rcar_du_lvds_connector {
+	struct rcar_du_connector connector;
+
+	const struct rcar_du_panel_data *panel;
+};
+
+#define to_rcar_lvds_connector(c) \
+	container_of(c, struct rcar_du_lvds_connector, connector.connector)
+
+/* -----------------------------------------------------------------------------
+ * Connector
+ */
+
+static int rcar_du_lvds_connector_get_modes(struct drm_connector *connector)
+{
+	struct rcar_du_lvds_connector *lvdscon = to_rcar_lvds_connector(connector);
+	struct drm_display_mode *mode;
+
+	mode = drm_mode_create(connector->dev);
+	if (mode == NULL)
+		return 0;
+
+	mode->type = DRM_MODE_TYPE_PREFERRED | DRM_MODE_TYPE_DRIVER;
+	mode->clock = lvdscon->panel->mode.clock;
+	mode->hdisplay = lvdscon->panel->mode.hdisplay;
+	mode->hsync_start = lvdscon->panel->mode.hsync_start;
+	mode->hsync_end = lvdscon->panel->mode.hsync_end;
+	mode->htotal = lvdscon->panel->mode.htotal;
+	mode->vdisplay = lvdscon->panel->mode.vdisplay;
+	mode->vsync_start = lvdscon->panel->mode.vsync_start;
+	mode->vsync_end = lvdscon->panel->mode.vsync_end;
+	mode->vtotal = lvdscon->panel->mode.vtotal;
+	mode->flags = lvdscon->panel->mode.flags;
+
+	drm_mode_set_name(mode);
+	drm_mode_probed_add(connector, mode);
+
+	return 1;
+}
+
+static int rcar_du_lvds_connector_mode_valid(struct drm_connector *connector,
+					    struct drm_display_mode *mode)
+{
+	return MODE_OK;
+}
+
+static const struct drm_connector_helper_funcs connector_helper_funcs = {
+	.get_modes = rcar_du_lvds_connector_get_modes,
+	.mode_valid = rcar_du_lvds_connector_mode_valid,
+	.best_encoder = rcar_du_connector_best_encoder,
+};
+
+static void rcar_du_lvds_connector_destroy(struct drm_connector *connector)
+{
+	drm_sysfs_connector_remove(connector);
+	drm_connector_cleanup(connector);
+}
+
+static enum drm_connector_status
+rcar_du_lvds_connector_detect(struct drm_connector *connector, bool force)
+{
+	return connector_status_connected;
+}
+
+static const struct drm_connector_funcs connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.detect = rcar_du_lvds_connector_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = rcar_du_lvds_connector_destroy,
+};
+
+static int rcar_du_lvds_connector_init(struct rcar_du_device *rcdu,
+				       struct rcar_du_encoder *renc,
+				       const struct rcar_du_panel_data *panel)
+{
+	struct rcar_du_lvds_connector *lvdscon;
+	struct drm_connector *connector;
+	int ret;
+
+	lvdscon = devm_kzalloc(rcdu->dev, sizeof(*lvdscon), GFP_KERNEL);
+	if (lvdscon == NULL)
+		return -ENOMEM;
+
+	lvdscon->panel = panel;
+
+	connector = &lvdscon->connector.connector;
+	connector->display_info.width_mm = panel->width_mm;
+	connector->display_info.height_mm = panel->height_mm;
+
+	ret = drm_connector_init(rcdu->ddev, connector, &connector_funcs,
+				 DRM_MODE_CONNECTOR_LVDS);
+	if (ret < 0)
+		return ret;
+
+	drm_connector_helper_add(connector, &connector_helper_funcs);
+	ret = drm_sysfs_connector_add(connector);
+	if (ret < 0)
+		return ret;
+
+	drm_helper_connector_dpms(connector, DRM_MODE_DPMS_OFF);
+	drm_object_property_set_value(&connector->base,
+		rcdu->ddev->mode_config.dpms_property, DRM_MODE_DPMS_OFF);
+
+	ret = drm_mode_connector_attach_encoder(connector, &renc->encoder);
+	if (ret < 0)
+		return ret;
+
+	connector->encoder = &renc->encoder;
+	lvdscon->connector.encoder = renc;
+
+	return 0;
+}
+
+/* -----------------------------------------------------------------------------
+ * Encoder
+ */
+
+static void rcar_du_lvds_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+}
+
+static bool rcar_du_lvds_encoder_mode_fixup(struct drm_encoder *encoder,
+					   const struct drm_display_mode *mode,
+					   struct drm_display_mode *adjusted_mode)
+{
+	const struct drm_display_mode *panel_mode;
+	struct drm_device *dev = encoder->dev;
+	struct drm_connector *connector;
+	bool found = false;
+
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
+		if (connector->encoder == encoder) {
+			found = true;
+			break;
+		}
+	}
+
+	if (!found) {
+		dev_dbg(dev->dev, "mode_fixup: no connector found\n");
+		return false;
+	}
+
+	if (list_empty(&connector->modes)) {
+		dev_dbg(dev->dev, "mode_fixup: empty modes list\n");
+		return false;
+	}
+
+	panel_mode = list_first_entry(&connector->modes,
+				      struct drm_display_mode, head);
+
+	/* We're not allowed to modify the resolution. */
+	if (mode->hdisplay != panel_mode->hdisplay ||
+	    mode->vdisplay != panel_mode->vdisplay)
+		return false;
+
+	/* The flat panel mode is fixed, just copy it to the adjusted mode. */
+	drm_mode_copy(adjusted_mode, panel_mode);
+
+	return true;
+}
+
+static const struct drm_encoder_helper_funcs encoder_helper_funcs = {
+	.dpms = rcar_du_lvds_encoder_dpms,
+	.mode_fixup = rcar_du_lvds_encoder_mode_fixup,
+	.prepare = rcar_du_encoder_mode_prepare,
+	.commit = rcar_du_encoder_mode_commit,
+	.mode_set = rcar_du_encoder_mode_set,
+};
+
+static const struct drm_encoder_funcs encoder_funcs = {
+	.destroy = drm_encoder_cleanup,
+};
+
+int rcar_du_lvds_init(struct rcar_du_device *rcdu,
+		      const struct rcar_du_encoder_lvds_data *data,
+		      unsigned int output)
+{
+	struct rcar_du_encoder *renc;
+	int ret;
+
+	renc = devm_kzalloc(rcdu->dev, sizeof(*renc), GFP_KERNEL);
+	if (renc == NULL)
+		return -ENOMEM;
+
+	renc->output = output;
+
+	ret = drm_encoder_init(rcdu->ddev, &renc->encoder, &encoder_funcs,
+			       DRM_MODE_ENCODER_LVDS);
+	if (ret < 0)
+		return ret;
+
+	drm_encoder_helper_add(&renc->encoder, &encoder_helper_funcs);
+
+	return rcar_du_lvds_connector_init(rcdu, renc, &data->panel);
+}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_lvds.h b/drivers/gpu/drm/rcar-du/rcar_du_lvds.h
new file mode 100644
index 000000000000..b47f8328e103
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_lvds.h
@@ -0,0 +1,24 @@
+/*
+ * rcar_du_lvds.h  --  R-Car Display Unit LVDS Encoder and Connector
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_LVDS_H__
+#define __RCAR_DU_LVDS_H__
+
+struct rcar_du_device;
+struct rcar_du_encoder_lvds_data;
+
+int rcar_du_lvds_init(struct rcar_du_device *rcdu,
+		      const struct rcar_du_encoder_lvds_data *data,
+		      unsigned int output);
+
+#endif /* __RCAR_DU_LVDS_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_plane.c b/drivers/gpu/drm/rcar-du/rcar_du_plane.c
new file mode 100644
index 000000000000..a65f81ddf51d
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_plane.c
@@ -0,0 +1,507 @@
+/*
+ * rcar_du_plane.c  --  R-Car Display Unit Planes
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_fb_cma_helper.h>
+#include <drm/drm_gem_cma_helper.h>
+
+#include "rcar_du_drv.h"
+#include "rcar_du_kms.h"
+#include "rcar_du_plane.h"
+#include "rcar_du_regs.h"
+
+#define RCAR_DU_COLORKEY_NONE		(0 << 24)
+#define RCAR_DU_COLORKEY_SOURCE		(1 << 24)
+#define RCAR_DU_COLORKEY_MASK		(1 << 24)
+
+struct rcar_du_kms_plane {
+	struct drm_plane plane;
+	struct rcar_du_plane *hwplane;
+};
+
+static inline struct rcar_du_plane *to_rcar_plane(struct drm_plane *plane)
+{
+	return container_of(plane, struct rcar_du_kms_plane, plane)->hwplane;
+}
+
+static u32 rcar_du_plane_read(struct rcar_du_device *rcdu,
+			      unsigned int index, u32 reg)
+{
+	return rcar_du_read(rcdu, index * PLANE_OFF + reg);
+}
+
+static void rcar_du_plane_write(struct rcar_du_device *rcdu,
+				unsigned int index, u32 reg, u32 data)
+{
+	rcar_du_write(rcdu, index * PLANE_OFF + reg, data);
+}
+
+int rcar_du_plane_reserve(struct rcar_du_plane *plane,
+			  const struct rcar_du_format_info *format)
+{
+	struct rcar_du_device *rcdu = plane->dev;
+	unsigned int i;
+	int ret = -EBUSY;
+
+	mutex_lock(&rcdu->planes.lock);
+
+	for (i = 0; i < ARRAY_SIZE(rcdu->planes.planes); ++i) {
+		if (!(rcdu->planes.free & (1 << i)))
+			continue;
+
+		if (format->planes == 1 ||
+		    rcdu->planes.free & (1 << ((i + 1) % 8)))
+			break;
+	}
+
+	if (i == ARRAY_SIZE(rcdu->planes.planes))
+		goto done;
+
+	rcdu->planes.free &= ~(1 << i);
+	if (format->planes == 2)
+		rcdu->planes.free &= ~(1 << ((i + 1) % 8));
+
+	plane->hwindex = i;
+
+	ret = 0;
+
+done:
+	mutex_unlock(&rcdu->planes.lock);
+	return ret;
+}
+
+void rcar_du_plane_release(struct rcar_du_plane *plane)
+{
+	struct rcar_du_device *rcdu = plane->dev;
+
+	if (plane->hwindex == -1)
+		return;
+
+	mutex_lock(&rcdu->planes.lock);
+	rcdu->planes.free |= 1 << plane->hwindex;
+	if (plane->format->planes == 2)
+		rcdu->planes.free |= 1 << ((plane->hwindex + 1) % 8);
+	mutex_unlock(&rcdu->planes.lock);
+
+	plane->hwindex = -1;
+}
+
+void rcar_du_plane_update_base(struct rcar_du_plane *plane)
+{
+	struct rcar_du_device *rcdu = plane->dev;
+	unsigned int index = plane->hwindex;
+
+	/* According to the datasheet the Y position is expressed in raster line
+	 * units. However, 32bpp formats seem to require a doubled Y position
+	 * value. Similarly, for the second plane, NV12 and NV21 formats seem to
+	 * require a halved Y position value.
+	 */
+	rcar_du_plane_write(rcdu, index, PnSPXR, plane->src_x);
+	rcar_du_plane_write(rcdu, index, PnSPYR, plane->src_y *
+			    (plane->format->bpp == 32 ? 2 : 1));
+	rcar_du_plane_write(rcdu, index, PnDSA0R, plane->dma[0]);
+
+	if (plane->format->planes == 2) {
+		index = (index + 1) % 8;
+
+		rcar_du_plane_write(rcdu, index, PnSPXR, plane->src_x);
+		rcar_du_plane_write(rcdu, index, PnSPYR, plane->src_y *
+				    (plane->format->bpp == 16 ? 2 : 1) / 2);
+		rcar_du_plane_write(rcdu, index, PnDSA0R, plane->dma[1]);
+	}
+}
+
+void rcar_du_plane_compute_base(struct rcar_du_plane *plane,
+				struct drm_framebuffer *fb)
+{
+	struct drm_gem_cma_object *gem;
+
+	gem = drm_fb_cma_get_gem_obj(fb, 0);
+	plane->dma[0] = gem->paddr + fb->offsets[0];
+
+	if (plane->format->planes == 2) {
+		gem = drm_fb_cma_get_gem_obj(fb, 1);
+		plane->dma[1] = gem->paddr + fb->offsets[1];
+	}
+}
+
+static void rcar_du_plane_setup_mode(struct rcar_du_plane *plane,
+				     unsigned int index)
+{
+	struct rcar_du_device *rcdu = plane->dev;
+	u32 colorkey;
+	u32 pnmr;
+
+	/* The PnALPHAR register controls alpha-blending in 16bpp formats
+	 * (ARGB1555 and XRGB1555).
+	 *
+	 * For ARGB, set the alpha value to 0, and enable alpha-blending when
+	 * the A bit is 0. This maps A=0 to alpha=0 and A=1 to alpha=255.
+	 *
+	 * For XRGB, set the alpha value to the plane-wide alpha value and
+	 * enable alpha-blending regardless of the X bit value.
+	 */
+	if (plane->format->fourcc != DRM_FORMAT_XRGB1555)
+		rcar_du_plane_write(rcdu, index, PnALPHAR, PnALPHAR_ABIT_0);
+	else
+		rcar_du_plane_write(rcdu, index, PnALPHAR,
+				    PnALPHAR_ABIT_X | plane->alpha);
+
+	pnmr = PnMR_BM_MD | plane->format->pnmr;
+
+	/* Disable color keying when requested. YUV formats have the
+	 * PnMR_SPIM_TP_OFF bit set in their pnmr field, disabling color keying
+	 * automatically.
+	 */
+	if ((plane->colorkey & RCAR_DU_COLORKEY_MASK) == RCAR_DU_COLORKEY_NONE)
+		pnmr |= PnMR_SPIM_TP_OFF;
+
+	/* For packed YUV formats we need to select the U/V order. */
+	if (plane->format->fourcc == DRM_FORMAT_YUYV)
+		pnmr |= PnMR_YCDF_YUYV;
+
+	rcar_du_plane_write(rcdu, index, PnMR, pnmr);
+
+	switch (plane->format->fourcc) {
+	case DRM_FORMAT_RGB565:
+		colorkey = ((plane->colorkey & 0xf80000) >> 8)
+			 | ((plane->colorkey & 0x00fc00) >> 5)
+			 | ((plane->colorkey & 0x0000f8) >> 3);
+		rcar_du_plane_write(rcdu, index, PnTC2R, colorkey);
+		break;
+
+	case DRM_FORMAT_ARGB1555:
+	case DRM_FORMAT_XRGB1555:
+		colorkey = ((plane->colorkey & 0xf80000) >> 9)
+			 | ((plane->colorkey & 0x00f800) >> 6)
+			 | ((plane->colorkey & 0x0000f8) >> 3);
+		rcar_du_plane_write(rcdu, index, PnTC2R, colorkey);
+		break;
+
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_ARGB8888:
+		rcar_du_plane_write(rcdu, index, PnTC3R,
+				    PnTC3R_CODE | (plane->colorkey & 0xffffff));
+		break;
+	}
+}
+
+static void __rcar_du_plane_setup(struct rcar_du_plane *plane,
+				  unsigned int index)
+{
+	struct rcar_du_device *rcdu = plane->dev;
+	u32 ddcr2 = PnDDCR2_CODE;
+	u32 ddcr4;
+	u32 mwr;
+
+	/* Data format
+	 *
+	 * The data format is selected by the DDDF field in PnMR and the EDF
+	 * field in DDCR4.
+	 */
+	ddcr4 = rcar_du_plane_read(rcdu, index, PnDDCR4);
+	ddcr4 &= ~PnDDCR4_EDF_MASK;
+	ddcr4 |= plane->format->edf | PnDDCR4_CODE;
+
+	rcar_du_plane_setup_mode(plane, index);
+
+	if (plane->format->planes == 2) {
+		if (plane->hwindex != index) {
+			if (plane->format->fourcc == DRM_FORMAT_NV12 ||
+			    plane->format->fourcc == DRM_FORMAT_NV21)
+				ddcr2 |= PnDDCR2_Y420;
+
+			if (plane->format->fourcc == DRM_FORMAT_NV21)
+				ddcr2 |= PnDDCR2_NV21;
+
+			ddcr2 |= PnDDCR2_DIVU;
+		} else {
+			ddcr2 |= PnDDCR2_DIVY;
+		}
+	}
+
+	rcar_du_plane_write(rcdu, index, PnDDCR2, ddcr2);
+	rcar_du_plane_write(rcdu, index, PnDDCR4, ddcr4);
+
+	/* Memory pitch (expressed in pixels) */
+	if (plane->format->planes == 2)
+		mwr = plane->pitch;
+	else
+		mwr = plane->pitch * 8 / plane->format->bpp;
+
+	rcar_du_plane_write(rcdu, index, PnMWR, mwr);
+
+	/* Destination position and size */
+	rcar_du_plane_write(rcdu, index, PnDSXR, plane->width);
+	rcar_du_plane_write(rcdu, index, PnDSYR, plane->height);
+	rcar_du_plane_write(rcdu, index, PnDPXR, plane->dst_x);
+	rcar_du_plane_write(rcdu, index, PnDPYR, plane->dst_y);
+
+	/* Wrap-around and blinking, disabled */
+	rcar_du_plane_write(rcdu, index, PnWASPR, 0);
+	rcar_du_plane_write(rcdu, index, PnWAMWR, 4095);
+	rcar_du_plane_write(rcdu, index, PnBTR, 0);
+	rcar_du_plane_write(rcdu, index, PnMLR, 0);
+}
+
+void rcar_du_plane_setup(struct rcar_du_plane *plane)
+{
+	__rcar_du_plane_setup(plane, plane->hwindex);
+	if (plane->format->planes == 2)
+		__rcar_du_plane_setup(plane, (plane->hwindex + 1) % 8);
+
+	rcar_du_plane_update_base(plane);
+}
+
+static int
+rcar_du_plane_update(struct drm_plane *plane, struct drm_crtc *crtc,
+		       struct drm_framebuffer *fb, int crtc_x, int crtc_y,
+		       unsigned int crtc_w, unsigned int crtc_h,
+		       uint32_t src_x, uint32_t src_y,
+		       uint32_t src_w, uint32_t src_h)
+{
+	struct rcar_du_plane *rplane = to_rcar_plane(plane);
+	struct rcar_du_device *rcdu = plane->dev->dev_private;
+	const struct rcar_du_format_info *format;
+	unsigned int nplanes;
+	int ret;
+
+	format = rcar_du_format_info(fb->pixel_format);
+	if (format == NULL) {
+		dev_dbg(rcdu->dev, "%s: unsupported format %08x\n", __func__,
+			fb->pixel_format);
+		return -EINVAL;
+	}
+
+	if (src_w >> 16 != crtc_w || src_h >> 16 != crtc_h) {
+		dev_dbg(rcdu->dev, "%s: scaling not supported\n", __func__);
+		return -EINVAL;
+	}
+
+	nplanes = rplane->format ? rplane->format->planes : 0;
+
+	/* Reallocate hardware planes if the number of required planes has
+	 * changed.
+	 */
+	if (format->planes != nplanes) {
+		rcar_du_plane_release(rplane);
+		ret = rcar_du_plane_reserve(rplane, format);
+		if (ret < 0)
+			return ret;
+	}
+
+	rplane->crtc = crtc;
+	rplane->format = format;
+	rplane->pitch = fb->pitches[0];
+
+	rplane->src_x = src_x >> 16;
+	rplane->src_y = src_y >> 16;
+	rplane->dst_x = crtc_x;
+	rplane->dst_y = crtc_y;
+	rplane->width = crtc_w;
+	rplane->height = crtc_h;
+
+	rcar_du_plane_compute_base(rplane, fb);
+	rcar_du_plane_setup(rplane);
+
+	mutex_lock(&rcdu->planes.lock);
+	rplane->enabled = true;
+	rcar_du_crtc_update_planes(rplane->crtc);
+	mutex_unlock(&rcdu->planes.lock);
+
+	return 0;
+}
+
+static int rcar_du_plane_disable(struct drm_plane *plane)
+{
+	struct rcar_du_device *rcdu = plane->dev->dev_private;
+	struct rcar_du_plane *rplane = to_rcar_plane(plane);
+
+	if (!rplane->enabled)
+		return 0;
+
+	mutex_lock(&rcdu->planes.lock);
+	rplane->enabled = false;
+	rcar_du_crtc_update_planes(rplane->crtc);
+	mutex_unlock(&rcdu->planes.lock);
+
+	rcar_du_plane_release(rplane);
+
+	rplane->crtc = NULL;
+	rplane->format = NULL;
+
+	return 0;
+}
+
+/* Both the .set_property and the .update_plane operations are called with the
+ * mode_config lock held. There is this no need to explicitly protect access to
+ * the alpha and colorkey fields and the mode register.
+ */
+static void rcar_du_plane_set_alpha(struct rcar_du_plane *plane, u32 alpha)
+{
+	if (plane->alpha == alpha)
+		return;
+
+	plane->alpha = alpha;
+	if (!plane->enabled || plane->format->fourcc != DRM_FORMAT_XRGB1555)
+		return;
+
+	rcar_du_plane_setup_mode(plane, plane->hwindex);
+}
+
+static void rcar_du_plane_set_colorkey(struct rcar_du_plane *plane,
+				       u32 colorkey)
+{
+	if (plane->colorkey == colorkey)
+		return;
+
+	plane->colorkey = colorkey;
+	if (!plane->enabled)
+		return;
+
+	rcar_du_plane_setup_mode(plane, plane->hwindex);
+}
+
+static void rcar_du_plane_set_zpos(struct rcar_du_plane *plane,
+				   unsigned int zpos)
+{
+	struct rcar_du_device *rcdu = plane->dev;
+
+	mutex_lock(&rcdu->planes.lock);
+	if (plane->zpos == zpos)
+		goto done;
+
+	plane->zpos = zpos;
+	if (!plane->enabled)
+		goto done;
+
+	rcar_du_crtc_update_planes(plane->crtc);
+
+done:
+	mutex_unlock(&rcdu->planes.lock);
+}
+
+static int rcar_du_plane_set_property(struct drm_plane *plane,
+				      struct drm_property *property,
+				      uint64_t value)
+{
+	struct rcar_du_device *rcdu = plane->dev->dev_private;
+	struct rcar_du_plane *rplane = to_rcar_plane(plane);
+
+	if (property == rcdu->planes.alpha)
+		rcar_du_plane_set_alpha(rplane, value);
+	else if (property == rcdu->planes.colorkey)
+		rcar_du_plane_set_colorkey(rplane, value);
+	else if (property == rcdu->planes.zpos)
+		rcar_du_plane_set_zpos(rplane, value);
+	else
+		return -EINVAL;
+
+	return 0;
+}
+
+static const struct drm_plane_funcs rcar_du_plane_funcs = {
+	.update_plane = rcar_du_plane_update,
+	.disable_plane = rcar_du_plane_disable,
+	.set_property = rcar_du_plane_set_property,
+	.destroy = drm_plane_cleanup,
+};
+
+static const uint32_t formats[] = {
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_ARGB1555,
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_NV12,
+	DRM_FORMAT_NV21,
+	DRM_FORMAT_NV16,
+};
+
+int rcar_du_plane_init(struct rcar_du_device *rcdu)
+{
+	unsigned int i;
+
+	mutex_init(&rcdu->planes.lock);
+	rcdu->planes.free = 0xff;
+
+	rcdu->planes.alpha =
+		drm_property_create_range(rcdu->ddev, 0, "alpha", 0, 255);
+	if (rcdu->planes.alpha == NULL)
+		return -ENOMEM;
+
+	/* The color key is expressed as an RGB888 triplet stored in a 32-bit
+	 * integer in XRGB8888 format. Bit 24 is used as a flag to disable (0)
+	 * or enable source color keying (1).
+	 */
+	rcdu->planes.colorkey =
+		drm_property_create_range(rcdu->ddev, 0, "colorkey",
+					  0, 0x01ffffff);
+	if (rcdu->planes.colorkey == NULL)
+		return -ENOMEM;
+
+	rcdu->planes.zpos =
+		drm_property_create_range(rcdu->ddev, 0, "zpos", 1, 7);
+	if (rcdu->planes.zpos == NULL)
+		return -ENOMEM;
+
+	for (i = 0; i < ARRAY_SIZE(rcdu->planes.planes); ++i) {
+		struct rcar_du_plane *plane = &rcdu->planes.planes[i];
+
+		plane->dev = rcdu;
+		plane->hwindex = -1;
+		plane->alpha = 255;
+		plane->colorkey = RCAR_DU_COLORKEY_NONE;
+		plane->zpos = 0;
+	}
+
+	return 0;
+}
+
+int rcar_du_plane_register(struct rcar_du_device *rcdu)
+{
+	unsigned int i;
+	int ret;
+
+	for (i = 0; i < RCAR_DU_NUM_KMS_PLANES; ++i) {
+		struct rcar_du_kms_plane *plane;
+
+		plane = devm_kzalloc(rcdu->dev, sizeof(*plane), GFP_KERNEL);
+		if (plane == NULL)
+			return -ENOMEM;
+
+		plane->hwplane = &rcdu->planes.planes[i + 2];
+		plane->hwplane->zpos = 1;
+
+		ret = drm_plane_init(rcdu->ddev, &plane->plane,
+				     (1 << rcdu->num_crtcs) - 1,
+				     &rcar_du_plane_funcs, formats,
+				     ARRAY_SIZE(formats), false);
+		if (ret < 0)
+			return ret;
+
+		drm_object_attach_property(&plane->plane.base,
+					   rcdu->planes.alpha, 255);
+		drm_object_attach_property(&plane->plane.base,
+					   rcdu->planes.colorkey,
+					   RCAR_DU_COLORKEY_NONE);
+		drm_object_attach_property(&plane->plane.base,
+					   rcdu->planes.zpos, 1);
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_plane.h b/drivers/gpu/drm/rcar-du/rcar_du_plane.h
new file mode 100644
index 000000000000..5397dba2fe57
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_plane.h
@@ -0,0 +1,67 @@
+/*
+ * rcar_du_plane.h  --  R-Car Display Unit Planes
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_PLANE_H__
+#define __RCAR_DU_PLANE_H__
+
+struct drm_crtc;
+struct drm_framebuffer;
+struct rcar_du_device;
+struct rcar_du_format_info;
+
+/* The RCAR DU has 8 hardware planes, shared between KMS planes and CRTCs. As
+ * using KMS planes requires at least one of the CRTCs being enabled, no more
+ * than 7 KMS planes can be available. We thus create 7 KMS planes and
+ * 9 software planes (one for each KMS planes and one for each CRTC).
+ */
+
+#define RCAR_DU_NUM_KMS_PLANES		7
+#define RCAR_DU_NUM_HW_PLANES		8
+#define RCAR_DU_NUM_SW_PLANES		9
+
+struct rcar_du_plane {
+	struct rcar_du_device *dev;
+	struct drm_crtc *crtc;
+
+	bool enabled;
+
+	int hwindex;		/* 0-based, -1 means unused */
+	unsigned int alpha;
+	unsigned int colorkey;
+	unsigned int zpos;
+
+	const struct rcar_du_format_info *format;
+
+	unsigned long dma[2];
+	unsigned int pitch;
+
+	unsigned int width;
+	unsigned int height;
+
+	unsigned int src_x;
+	unsigned int src_y;
+	unsigned int dst_x;
+	unsigned int dst_y;
+};
+
+int rcar_du_plane_init(struct rcar_du_device *rcdu);
+int rcar_du_plane_register(struct rcar_du_device *rcdu);
+void rcar_du_plane_setup(struct rcar_du_plane *plane);
+void rcar_du_plane_update_base(struct rcar_du_plane *plane);
+void rcar_du_plane_compute_base(struct rcar_du_plane *plane,
+				struct drm_framebuffer *fb);
+int rcar_du_plane_reserve(struct rcar_du_plane *plane,
+			  const struct rcar_du_format_info *format);
+void rcar_du_plane_release(struct rcar_du_plane *plane);
+
+#endif /* __RCAR_DU_PLANE_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_regs.h b/drivers/gpu/drm/rcar-du/rcar_du_regs.h
new file mode 100644
index 000000000000..69f21f19b51c
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_regs.h
@@ -0,0 +1,445 @@
+/*
+ * rcar_du_regs.h  --  R-Car Display Unit Registers Definitions
+ *
+ * Copyright (C) 2013 Renesas Electronics Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation.
+ */
+
+#ifndef __RCAR_DU_REGS_H__
+#define __RCAR_DU_REGS_H__
+
+#define DISP2_REG_OFFSET	 0x30000
+
+/* -----------------------------------------------------------------------------
+ * Display Control Registers
+ */
+
+#define DSYSR			0x00000	/* display 1 */
+#define D2SYSR			0x30000	/* display 2 */
+#define DSYSR_ILTS		(1 << 29)
+#define DSYSR_DSEC		(1 << 20)
+#define DSYSR_IUPD		(1 << 16)
+#define DSYSR_DRES		(1 << 9)
+#define DSYSR_DEN		(1 << 8)
+#define DSYSR_TVM_MASTER	(0 << 6)
+#define DSYSR_TVM_SWITCH	(1 << 6)
+#define DSYSR_TVM_TVSYNC	(2 << 6)
+#define DSYSR_TVM_MASK		(3 << 6)
+#define DSYSR_SCM_INT_NONE	(0 << 4)
+#define DSYSR_SCM_INT_SYNC	(2 << 4)
+#define DSYSR_SCM_INT_VIDEO	(3 << 4)
+
+#define DSMR			0x00004
+#define D2SMR			0x30004
+#define DSMR_VSPM		(1 << 28)
+#define DSMR_ODPM		(1 << 27)
+#define DSMR_DIPM_DISP		(0 << 25)
+#define DSMR_DIPM_CSYNC		(1 << 25)
+#define DSMR_DIPM_DE		(3 << 25)
+#define DSMR_DIPM_MASK		(3 << 25)
+#define DSMR_CSPM		(1 << 24)
+#define DSMR_DIL		(1 << 19)
+#define DSMR_VSL		(1 << 18)
+#define DSMR_HSL		(1 << 17)
+#define DSMR_DDIS		(1 << 16)
+#define DSMR_CDEL		(1 << 15)
+#define DSMR_CDEM_CDE		(0 << 13)
+#define DSMR_CDEM_LOW		(2 << 13)
+#define DSMR_CDEM_HIGH		(3 << 13)
+#define DSMR_CDEM_MASK		(3 << 13)
+#define DSMR_CDED		(1 << 12)
+#define DSMR_ODEV		(1 << 8)
+#define DSMR_CSY_VH_OR		(0 << 6)
+#define DSMR_CSY_333		(2 << 6)
+#define DSMR_CSY_222		(3 << 6)
+#define DSMR_CSY_MASK		(3 << 6)
+
+#define DSSR			0x00008
+#define D2SSR			0x30008
+#define DSSR_VC1FB_DSA0		(0 << 30)
+#define DSSR_VC1FB_DSA1		(1 << 30)
+#define DSSR_VC1FB_DSA2		(2 << 30)
+#define DSSR_VC1FB_INIT		(3 << 30)
+#define DSSR_VC1FB_MASK		(3 << 30)
+#define DSSR_VC0FB_DSA0		(0 << 28)
+#define DSSR_VC0FB_DSA1		(1 << 28)
+#define DSSR_VC0FB_DSA2		(2 << 28)
+#define DSSR_VC0FB_INIT		(3 << 28)
+#define DSSR_VC0FB_MASK		(3 << 28)
+#define DSSR_DFB(n)		(1 << ((n)+15))
+#define DSSR_TVR		(1 << 15)
+#define DSSR_FRM		(1 << 14)
+#define DSSR_VBK		(1 << 11)
+#define DSSR_RINT		(1 << 9)
+#define DSSR_HBK		(1 << 8)
+#define DSSR_ADC(n)		(1 << ((n)-1))
+
+#define DSRCR			0x0000c
+#define D2SRCR			0x3000c
+#define DSRCR_TVCL		(1 << 15)
+#define DSRCR_FRCL		(1 << 14)
+#define DSRCR_VBCL		(1 << 11)
+#define DSRCR_RICL		(1 << 9)
+#define DSRCR_HBCL		(1 << 8)
+#define DSRCR_ADCL(n)		(1 << ((n)-1))
+#define DSRCR_MASK		0x0000cbff
+
+#define DIER			0x00010
+#define D2IER			0x30010
+#define DIER_TVE		(1 << 15)
+#define DIER_FRE		(1 << 14)
+#define DIER_VBE		(1 << 11)
+#define DIER_RIE		(1 << 9)
+#define DIER_HBE		(1 << 8)
+#define DIER_ADCE(n)		(1 << ((n)-1))
+
+#define CPCR			0x00014
+#define CPCR_CP4CE		(1 << 19)
+#define CPCR_CP3CE		(1 << 18)
+#define CPCR_CP2CE		(1 << 17)
+#define CPCR_CP1CE		(1 << 16)
+
+#define DPPR			0x00018
+#define DPPR_DPE(n)		(1 << ((n)*4-1))
+#define DPPR_DPS(n, p)		(((p)-1) << DPPR_DPS_SHIFT(n))
+#define DPPR_DPS_SHIFT(n)	(((n)-1)*4)
+#define DPPR_BPP16		(DPPR_DPE(8) | DPPR_DPS(8, 1))	/* plane1 */
+#define DPPR_BPP32_P1		(DPPR_DPE(7) | DPPR_DPS(7, 1))
+#define DPPR_BPP32_P2		(DPPR_DPE(8) | DPPR_DPS(8, 2))
+#define DPPR_BPP32		(DPPR_BPP32_P1 | DPPR_BPP32_P2)	/* plane1 & 2 */
+
+#define DEFR			0x00020
+#define D2EFR			0x30020
+#define DEFR_CODE		(0x7773 << 16)
+#define DEFR_EXSL		(1 << 12)
+#define DEFR_EXVL		(1 << 11)
+#define DEFR_EXUP		(1 << 5)
+#define DEFR_VCUP		(1 << 4)
+#define DEFR_DEFE		(1 << 0)
+
+#define DAPCR			0x00024
+#define DAPCR_CODE		(0x7773 << 16)
+#define DAPCR_AP2E		(1 << 4)
+#define DAPCR_AP1E		(1 << 0)
+
+#define DCPCR			0x00028
+#define DCPCR_CODE		(0x7773 << 16)
+#define DCPCR_CA2B		(1 << 13)
+#define DCPCR_CD2F		(1 << 12)
+#define DCPCR_DC2E		(1 << 8)
+#define DCPCR_CAB		(1 << 5)
+#define DCPCR_CDF		(1 << 4)
+#define DCPCR_DCE		(1 << 0)
+
+#define DEFR2			0x00034
+#define D2EFR2			0x30034
+#define DEFR2_CODE		(0x7775 << 16)
+#define DEFR2_DEFE2G		(1 << 0)
+
+#define DEFR3			0x00038
+#define D2EFR3			0x30038
+#define DEFR3_CODE		(0x7776 << 16)
+#define DEFR3_EVDA		(1 << 14)
+#define DEFR3_EVDM_1		(1 << 12)
+#define DEFR3_EVDM_2		(2 << 12)
+#define DEFR3_EVDM_3		(3 << 12)
+#define DEFR3_VMSM2_EMA		(1 << 6)
+#define DEFR3_VMSM1_ENA		(1 << 4)
+#define DEFR3_DEFE3		(1 << 0)
+
+#define DEFR4			0x0003c
+#define D2EFR4			0x3003c
+#define DEFR4_CODE		(0x7777 << 16)
+#define DEFR4_LRUO		(1 << 5)
+#define DEFR4_SPCE		(1 << 4)
+
+#define DVCSR			0x000d0
+#define DVCSR_VCnFB2_DSA0(n)	(0 << ((n)*2+16))
+#define DVCSR_VCnFB2_DSA1(n)	(1 << ((n)*2+16))
+#define DVCSR_VCnFB2_DSA2(n)	(2 << ((n)*2+16))
+#define DVCSR_VCnFB2_INIT(n)	(3 << ((n)*2+16))
+#define DVCSR_VCnFB2_MASK(n)	(3 << ((n)*2+16))
+#define DVCSR_VCnFB_DSA0(n)	(0 << ((n)*2))
+#define DVCSR_VCnFB_DSA1(n)	(1 << ((n)*2))
+#define DVCSR_VCnFB_DSA2(n)	(2 << ((n)*2))
+#define DVCSR_VCnFB_INIT(n)	(3 << ((n)*2))
+#define DVCSR_VCnFB_MASK(n)	(3 << ((n)*2))
+
+#define DEFR5			0x000e0
+#define DEFR5_CODE		(0x66 << 24)
+#define DEFR5_YCRGB2_DIS	(0 << 14)
+#define DEFR5_YCRGB2_PRI1	(1 << 14)
+#define DEFR5_YCRGB2_PRI2	(2 << 14)
+#define DEFR5_YCRGB2_PRI3	(3 << 14)
+#define DEFR5_YCRGB2_MASK	(3 << 14)
+#define DEFR5_YCRGB1_DIS	(0 << 12)
+#define DEFR5_YCRGB1_PRI1	(1 << 12)
+#define DEFR5_YCRGB1_PRI2	(2 << 12)
+#define DEFR5_YCRGB1_PRI3	(3 << 12)
+#define DEFR5_YCRGB1_MASK	(3 << 12)
+#define DEFR5_DEFE5		(1 << 0)
+
+#define DDLTR			0x000e4
+#define DDLTR_CODE		(0x7766 << 16)
+#define DDLTR_DLAR2		(1 << 6)
+#define DDLTR_DLAY2		(1 << 5)
+#define DDLTR_DLAY1		(1 << 1)
+
+#define DEFR6			0x000e8
+#define DEFR6_CODE		(0x7778 << 16)
+#define DEFR6_ODPM22_D2SMR	(0 << 10)
+#define DEFR6_ODPM22_DISP	(2 << 10)
+#define DEFR6_ODPM22_CDE	(3 << 10)
+#define DEFR6_ODPM22_MASK	(3 << 10)
+#define DEFR6_ODPM12_DSMR	(0 << 8)
+#define DEFR6_ODPM12_DISP	(2 << 8)
+#define DEFR6_ODPM12_CDE	(3 << 8)
+#define DEFR6_ODPM12_MASK	(3 << 8)
+#define DEFR6_TCNE2		(1 << 6)
+#define DEFR6_MLOS1		(1 << 2)
+#define DEFR6_DEFAULT		(DEFR6_CODE | DEFR6_TCNE2)
+
+/* -----------------------------------------------------------------------------
+ * Display Timing Generation Registers
+ */
+
+#define HDSR			0x00040
+#define HDER			0x00044
+#define VDSR			0x00048
+#define VDER			0x0004c
+#define HCR			0x00050
+#define HSWR			0x00054
+#define VCR			0x00058
+#define VSPR			0x0005c
+#define EQWR			0x00060
+#define SPWR			0x00064
+#define CLAMPSR			0x00070
+#define CLAMPWR			0x00074
+#define DESR			0x00078
+#define DEWR			0x0007c
+
+/* -----------------------------------------------------------------------------
+ * Display Attribute Registers
+ */
+
+#define CP1TR			0x00080
+#define CP2TR			0x00084
+#define CP3TR			0x00088
+#define CP4TR			0x0008c
+
+#define DOOR			0x00090
+#define DOOR_RGB(r, g, b)	(((r) << 18) | ((g) << 10) | ((b) << 2))
+#define CDER			0x00094
+#define CDER_RGB(r, g, b)	(((r) << 18) | ((g) << 10) | ((b) << 2))
+#define BPOR			0x00098
+#define BPOR_RGB(r, g, b)	(((r) << 18) | ((g) << 10) | ((b) << 2))
+
+#define RINTOFSR		0x0009c
+
+#define DSHPR			0x000c8
+#define DSHPR_CODE		(0x7776 << 16)
+#define DSHPR_PRIH		(0xa << 4)
+#define DSHPR_PRIL_BPP16	(0x8 << 0)
+#define DSHPR_PRIL_BPP32	(0x9 << 0)
+
+/* -----------------------------------------------------------------------------
+ * Display Plane Registers
+ */
+
+#define PLANE_OFF		0x00100
+
+#define PnMR			0x00100 /* plane 1 */
+#define PnMR_VISL_VIN0		(0 << 26)	/* use Video Input 0 */
+#define PnMR_VISL_VIN1		(1 << 26)	/* use Video Input 1 */
+#define PnMR_VISL_VIN2		(2 << 26)	/* use Video Input 2 */
+#define PnMR_VISL_VIN3		(3 << 26)	/* use Video Input 3 */
+#define PnMR_YCDF_YUYV		(1 << 20)	/* YUYV format */
+#define PnMR_TC_R		(0 << 17)	/* Tranparent color is PnTC1R */
+#define PnMR_TC_CP		(1 << 17)	/* Tranparent color is color palette */
+#define PnMR_WAE		(1 << 16)	/* Wrap around Enable */
+#define PnMR_SPIM_TP		(0 << 12)	/* Transparent Color */
+#define PnMR_SPIM_ALP		(1 << 12)	/* Alpha Blending */
+#define PnMR_SPIM_EOR		(2 << 12)	/* EOR */
+#define PnMR_SPIM_TP_OFF	(1 << 14)	/* No Transparent Color */
+#define PnMR_CPSL_CP1		(0 << 8)	/* Color Palette selected 1 */
+#define PnMR_CPSL_CP2		(1 << 8)	/* Color Palette selected 2 */
+#define PnMR_CPSL_CP3		(2 << 8)	/* Color Palette selected 3 */
+#define PnMR_CPSL_CP4		(3 << 8)	/* Color Palette selected 4 */
+#define PnMR_DC			(1 << 7)	/* Display Area Change */
+#define PnMR_BM_MD		(0 << 4)	/* Manual Display Change Mode */
+#define PnMR_BM_AR		(1 << 4)	/* Auto Rendering Mode */
+#define PnMR_BM_AD		(2 << 4)	/* Auto Display Change Mode */
+#define PnMR_BM_VC		(3 << 4)	/* Video Capture Mode */
+#define PnMR_DDDF_8BPP		(0 << 0)	/* 8bit */
+#define PnMR_DDDF_16BPP		(1 << 0)	/* 16bit or 32bit */
+#define PnMR_DDDF_ARGB		(2 << 0)	/* ARGB */
+#define PnMR_DDDF_YC		(3 << 0)	/* YC */
+#define PnMR_DDDF_MASK		(3 << 0)
+
+#define PnMWR			0x00104
+
+#define PnALPHAR		0x00108
+#define PnALPHAR_ABIT_1		(0 << 12)
+#define PnALPHAR_ABIT_0		(1 << 12)
+#define PnALPHAR_ABIT_X		(2 << 12)
+
+#define PnDSXR			0x00110
+#define PnDSYR			0x00114
+#define PnDPXR			0x00118
+#define PnDPYR			0x0011c
+
+#define PnDSA0R			0x00120
+#define PnDSA1R			0x00124
+#define PnDSA2R			0x00128
+#define PnDSA_MASK		0xfffffff0
+
+#define PnSPXR			0x00130
+#define PnSPYR			0x00134
+#define PnWASPR			0x00138
+#define PnWAMWR			0x0013c
+
+#define PnBTR			0x00140
+
+#define PnTC1R			0x00144
+#define PnTC2R			0x00148
+#define PnTC3R			0x0014c
+#define PnTC3R_CODE		(0x66 << 24)
+
+#define PnMLR			0x00150
+
+#define PnSWAPR			0x00180
+#define PnSWAPR_DIGN		(1 << 4)
+#define PnSWAPR_SPQW		(1 << 3)
+#define PnSWAPR_SPLW		(1 << 2)
+#define PnSWAPR_SPWD		(1 << 1)
+#define PnSWAPR_SPBY		(1 << 0)
+
+#define PnDDCR			0x00184
+#define PnDDCR_CODE		(0x7775 << 16)
+#define PnDDCR_LRGB1		(1 << 11)
+#define PnDDCR_LRGB0		(1 << 10)
+
+#define PnDDCR2			0x00188
+#define PnDDCR2_CODE		(0x7776 << 16)
+#define PnDDCR2_NV21		(1 << 5)
+#define PnDDCR2_Y420		(1 << 4)
+#define PnDDCR2_DIVU		(1 << 1)
+#define PnDDCR2_DIVY		(1 << 0)
+
+#define PnDDCR4			0x00190
+#define PnDDCR4_CODE		(0x7766 << 16)
+#define PnDDCR4_SDFS_RGB	(0 << 4)
+#define PnDDCR4_SDFS_YC		(5 << 4)
+#define PnDDCR4_SDFS_MASK	(7 << 4)
+#define PnDDCR4_EDF_NONE	(0 << 0)
+#define PnDDCR4_EDF_ARGB8888	(1 << 0)
+#define PnDDCR4_EDF_RGB888	(2 << 0)
+#define PnDDCR4_EDF_RGB666	(3 << 0)
+#define PnDDCR4_EDF_MASK	(7 << 0)
+
+#define APnMR			0x0a100
+#define APnMR_WAE		(1 << 16)	/* Wrap around Enable */
+#define APnMR_DC		(1 << 7)	/* Display Area Change */
+#define APnMR_BM_MD		(0 << 4)	/* Manual Display Change Mode */
+#define APnMR_BM_AD		(2 << 4)	/* Auto Display Change Mode */
+
+#define APnMWR			0x0a104
+#define APnDSA0R		0x0a120
+#define APnDSA1R		0x0a124
+#define APnDSA2R		0x0a128
+#define APnMLR			0x0a150
+
+/* -----------------------------------------------------------------------------
+ * Display Capture Registers
+ */
+
+#define DCMWR			0x0c104
+#define DC2MWR			0x0c204
+#define DCSAR			0x0c120
+#define DC2SAR			0x0c220
+#define DCMLR			0x0c150
+#define DC2MLR			0x0c250
+
+/* -----------------------------------------------------------------------------
+ * Color Palette Registers
+ */
+
+#define CP1_000R		0x01000
+#define CP1_255R		0x013fc
+#define CP2_000R		0x02000
+#define CP2_255R		0x023fc
+#define CP3_000R		0x03000
+#define CP3_255R		0x033fc
+#define CP4_000R		0x04000
+#define CP4_255R		0x043fc
+
+/* -----------------------------------------------------------------------------
+ * External Synchronization Control Registers
+ */
+
+#define ESCR			0x10000
+#define ESCR2			0x31000
+#define ESCR_DCLKOINV		(1 << 25)
+#define ESCR_DCLKSEL_DCLKIN	(0 << 20)
+#define ESCR_DCLKSEL_CLKS	(1 << 20)
+#define ESCR_DCLKSEL_MASK	(1 << 20)
+#define ESCR_DCLKDIS		(1 << 16)
+#define ESCR_SYNCSEL_OFF	(0 << 8)
+#define ESCR_SYNCSEL_EXVSYNC	(2 << 8)
+#define ESCR_SYNCSEL_EXHSYNC	(3 << 8)
+#define ESCR_FRQSEL_MASK	(0x3f << 0)
+
+#define OTAR			0x10004
+#define OTAR2			0x31004
+
+/* -----------------------------------------------------------------------------
+ * Dual Display Output Control Registers
+ */
+
+#define DORCR			0x11000
+#define DORCR_PG2T		(1 << 30)
+#define DORCR_DK2S		(1 << 28)
+#define DORCR_PG2D_DS1		(0 << 24)
+#define DORCR_PG2D_DS2		(1 << 24)
+#define DORCR_PG2D_FIX0		(2 << 24)
+#define DORCR_PG2D_DOOR		(3 << 24)
+#define DORCR_PG2D_MASK		(3 << 24)
+#define DORCR_DR1D		(1 << 21)
+#define DORCR_PG1D_DS1		(0 << 16)
+#define DORCR_PG1D_DS2		(1 << 16)
+#define DORCR_PG1D_FIX0		(2 << 16)
+#define DORCR_PG1D_DOOR		(3 << 16)
+#define DORCR_PG1D_MASK		(3 << 16)
+#define DORCR_RGPV		(1 << 4)
+#define DORCR_DPRS		(1 << 0)
+
+#define DPTSR			0x11004
+#define DPTSR_PnDK(n)		(1 << ((n) + 16))
+#define DPTSR_PnTS(n)		(1 << (n))
+
+#define DAPTSR			0x11008
+#define DAPTSR_APnDK(n)		(1 << ((n) + 16))
+#define DAPTSR_APnTS(n)		(1 << (n))
+
+#define DS1PR			0x11020
+#define DS2PR			0x11024
+
+/* -----------------------------------------------------------------------------
+ * YC-RGB Conversion Coefficient Registers
+ */
+
+#define YNCR			0x11080
+#define YNOR			0x11084
+#define CRNOR			0x11088
+#define CBNOR			0x1108c
+#define RCRCR			0x11090
+#define GCRCR			0x11094
+#define GCBCR			0x11098
+#define BCBCR			0x1109c
+
+#endif /* __RCAR_DU_REGS_H__ */
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vga.c b/drivers/gpu/drm/rcar-du/rcar_du_vga.c
new file mode 100644
index 000000000000..327289ec380d
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_vga.c
@@ -0,0 +1,149 @@
+/*
+ * rcar_du_vga.c  --  R-Car Display Unit VGA DAC and Connector
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <drm/drmP.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_crtc_helper.h>
+
+#include "rcar_du_drv.h"
+#include "rcar_du_kms.h"
+#include "rcar_du_vga.h"
+
+/* -----------------------------------------------------------------------------
+ * Connector
+ */
+
+static int rcar_du_vga_connector_get_modes(struct drm_connector *connector)
+{
+	return 0;
+}
+
+static int rcar_du_vga_connector_mode_valid(struct drm_connector *connector,
+					    struct drm_display_mode *mode)
+{
+	return MODE_OK;
+}
+
+static const struct drm_connector_helper_funcs connector_helper_funcs = {
+	.get_modes = rcar_du_vga_connector_get_modes,
+	.mode_valid = rcar_du_vga_connector_mode_valid,
+	.best_encoder = rcar_du_connector_best_encoder,
+};
+
+static void rcar_du_vga_connector_destroy(struct drm_connector *connector)
+{
+	drm_sysfs_connector_remove(connector);
+	drm_connector_cleanup(connector);
+}
+
+static enum drm_connector_status
+rcar_du_vga_connector_detect(struct drm_connector *connector, bool force)
+{
+	return connector_status_unknown;
+}
+
+static const struct drm_connector_funcs connector_funcs = {
+	.dpms = drm_helper_connector_dpms,
+	.detect = rcar_du_vga_connector_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = rcar_du_vga_connector_destroy,
+};
+
+static int rcar_du_vga_connector_init(struct rcar_du_device *rcdu,
+				      struct rcar_du_encoder *renc)
+{
+	struct rcar_du_connector *rcon;
+	struct drm_connector *connector;
+	int ret;
+
+	rcon = devm_kzalloc(rcdu->dev, sizeof(*rcon), GFP_KERNEL);
+	if (rcon == NULL)
+		return -ENOMEM;
+
+	connector = &rcon->connector;
+	connector->display_info.width_mm = 0;
+	connector->display_info.height_mm = 0;
+
+	ret = drm_connector_init(rcdu->ddev, connector, &connector_funcs,
+				 DRM_MODE_CONNECTOR_VGA);
+	if (ret < 0)
+		return ret;
+
+	drm_connector_helper_add(connector, &connector_helper_funcs);
+	ret = drm_sysfs_connector_add(connector);
+	if (ret < 0)
+		return ret;
+
+	drm_helper_connector_dpms(connector, DRM_MODE_DPMS_OFF);
+	drm_object_property_set_value(&connector->base,
+		rcdu->ddev->mode_config.dpms_property, DRM_MODE_DPMS_OFF);
+
+	ret = drm_mode_connector_attach_encoder(connector, &renc->encoder);
+	if (ret < 0)
+		return ret;
+
+	connector->encoder = &renc->encoder;
+	rcon->encoder = renc;
+
+	return 0;
+}
+
+/* -----------------------------------------------------------------------------
+ * Encoder
+ */
+
+static void rcar_du_vga_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+}
+
+static bool rcar_du_vga_encoder_mode_fixup(struct drm_encoder *encoder,
+					   const struct drm_display_mode *mode,
+					   struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static const struct drm_encoder_helper_funcs encoder_helper_funcs = {
+	.dpms = rcar_du_vga_encoder_dpms,
+	.mode_fixup = rcar_du_vga_encoder_mode_fixup,
+	.prepare = rcar_du_encoder_mode_prepare,
+	.commit = rcar_du_encoder_mode_commit,
+	.mode_set = rcar_du_encoder_mode_set,
+};
+
+static const struct drm_encoder_funcs encoder_funcs = {
+	.destroy = drm_encoder_cleanup,
+};
+
+int rcar_du_vga_init(struct rcar_du_device *rcdu,
+		     const struct rcar_du_encoder_vga_data *data,
+		     unsigned int output)
+{
+	struct rcar_du_encoder *renc;
+	int ret;
+
+	renc = devm_kzalloc(rcdu->dev, sizeof(*renc), GFP_KERNEL);
+	if (renc == NULL)
+		return -ENOMEM;
+
+	renc->output = output;
+
+	ret = drm_encoder_init(rcdu->ddev, &renc->encoder, &encoder_funcs,
+			       DRM_MODE_ENCODER_DAC);
+	if (ret < 0)
+		return ret;
+
+	drm_encoder_helper_add(&renc->encoder, &encoder_helper_funcs);
+
+	return rcar_du_vga_connector_init(rcdu, renc);
+}
diff --git a/drivers/gpu/drm/rcar-du/rcar_du_vga.h b/drivers/gpu/drm/rcar-du/rcar_du_vga.h
new file mode 100644
index 000000000000..66b4d2d7190d
--- /dev/null
+++ b/drivers/gpu/drm/rcar-du/rcar_du_vga.h
@@ -0,0 +1,24 @@
+/*
+ * rcar_du_vga.h  --  R-Car Display Unit VGA DAC and Connector
+ *
+ * Copyright (C) 2013 Renesas Corporation
+ *
+ * Contact: Laurent Pinchart (laurent.pinchart@ideasonboard.com)
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ */
+
+#ifndef __RCAR_DU_VGA_H__
+#define __RCAR_DU_VGA_H__
+
+struct rcar_du_device;
+struct rcar_du_encoder_vga_data;
+
+int rcar_du_vga_init(struct rcar_du_device *rcdu,
+		     const struct rcar_du_encoder_vga_data *data,
+		     unsigned int output);
+
+#endif /* __RCAR_DU_VGA_H__ */
diff --git a/drivers/gpu/drm/savage/savage_bci.c b/drivers/gpu/drm/savage/savage_bci.c
index b55c1d661147..bd6b2cf508d5 100644
--- a/drivers/gpu/drm/savage/savage_bci.c
+++ b/drivers/gpu/drm/savage/savage_bci.c
@@ -570,9 +570,6 @@ int savage_driver_firstopen(struct drm_device *dev)
 	unsigned int fb_rsrc, aper_rsrc;
 	int ret = 0;
 
-	dev_priv->mtrr[0].handle = -1;
-	dev_priv->mtrr[1].handle = -1;
-	dev_priv->mtrr[2].handle = -1;
 	if (S3_SAVAGE3D_SERIES(dev_priv->chipset)) {
 		fb_rsrc = 0;
 		fb_base = pci_resource_start(dev->pdev, 0);
@@ -584,21 +581,14 @@ int savage_driver_firstopen(struct drm_device *dev)
 		if (pci_resource_len(dev->pdev, 0) == 0x08000000) {
 			/* Don't make MMIO write-cobining! We need 3
 			 * MTRRs. */
-			dev_priv->mtrr[0].base = fb_base;
-			dev_priv->mtrr[0].size = 0x01000000;
-			dev_priv->mtrr[0].handle =
-			    drm_mtrr_add(dev_priv->mtrr[0].base,
-				         dev_priv->mtrr[0].size, DRM_MTRR_WC);
-			dev_priv->mtrr[1].base = fb_base + 0x02000000;
-			dev_priv->mtrr[1].size = 0x02000000;
-			dev_priv->mtrr[1].handle =
-			    drm_mtrr_add(dev_priv->mtrr[1].base,
-					 dev_priv->mtrr[1].size, DRM_MTRR_WC);
-			dev_priv->mtrr[2].base = fb_base + 0x04000000;
-			dev_priv->mtrr[2].size = 0x04000000;
-			dev_priv->mtrr[2].handle =
-			    drm_mtrr_add(dev_priv->mtrr[2].base,
-					 dev_priv->mtrr[2].size, DRM_MTRR_WC);
+			dev_priv->mtrr_handles[0] =
+				arch_phys_wc_add(fb_base, 0x01000000);
+			dev_priv->mtrr_handles[1] =
+				arch_phys_wc_add(fb_base + 0x02000000,
+						 0x02000000);
+			dev_priv->mtrr_handles[2] =
+				arch_phys_wc_add(fb_base + 0x04000000,
+						0x04000000);
 		} else {
 			DRM_ERROR("strange pci_resource_len %08llx\n",
 				  (unsigned long long)
@@ -616,11 +606,9 @@ int savage_driver_firstopen(struct drm_device *dev)
 		if (pci_resource_len(dev->pdev, 1) == 0x08000000) {
 			/* Can use one MTRR to cover both fb and
 			 * aperture. */
-			dev_priv->mtrr[0].base = fb_base;
-			dev_priv->mtrr[0].size = 0x08000000;
-			dev_priv->mtrr[0].handle =
-			    drm_mtrr_add(dev_priv->mtrr[0].base,
-					 dev_priv->mtrr[0].size, DRM_MTRR_WC);
+			dev_priv->mtrr_handles[0] =
+				arch_phys_wc_add(fb_base,
+						 0x08000000);
 		} else {
 			DRM_ERROR("strange pci_resource_len %08llx\n",
 				  (unsigned long long)
@@ -660,11 +648,10 @@ void savage_driver_lastclose(struct drm_device *dev)
 	drm_savage_private_t *dev_priv = dev->dev_private;
 	int i;
 
-	for (i = 0; i < 3; ++i)
-		if (dev_priv->mtrr[i].handle >= 0)
-			drm_mtrr_del(dev_priv->mtrr[i].handle,
-				 dev_priv->mtrr[i].base,
-				 dev_priv->mtrr[i].size, DRM_MTRR_WC);
+	for (i = 0; i < 3; ++i) {
+		arch_phys_wc_del(dev_priv->mtrr_handles[i]);
+		dev_priv->mtrr_handles[i] = 0;
+	}
 }
 
 int savage_driver_unload(struct drm_device *dev)
diff --git a/drivers/gpu/drm/savage/savage_drv.h b/drivers/gpu/drm/savage/savage_drv.h
index df2aac6636f7..c05082a59f6f 100644
--- a/drivers/gpu/drm/savage/savage_drv.h
+++ b/drivers/gpu/drm/savage/savage_drv.h
@@ -160,10 +160,7 @@ typedef struct drm_savage_private {
 	drm_local_map_t *cmd_dma;
 	drm_local_map_t fake_dma;
 
-	struct {
-		int handle;
-		unsigned long base, size;
-	} mtrr[3];
+	int mtrr_handles[3];
 
 	/* BCI and status-related stuff */
 	volatile uint32_t *status_ptr, *bci_ptr;
diff --git a/drivers/gpu/drm/shmobile/Kconfig b/drivers/gpu/drm/shmobile/Kconfig
index 7e7d52b2a2fc..ca498d151a76 100644
--- a/drivers/gpu/drm/shmobile/Kconfig
+++ b/drivers/gpu/drm/shmobile/Kconfig
@@ -1,6 +1,6 @@
 config DRM_SHMOBILE
 	tristate "DRM Support for SH Mobile"
-	depends on DRM && (SUPERH || ARCH_SHMOBILE)
+	depends on DRM && (ARM || SUPERH)
 	select DRM_KMS_HELPER
 	select DRM_KMS_CMA_HELPER
 	select DRM_GEM_CMA_HELPER
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_drv.c b/drivers/gpu/drm/shmobile/shmob_drm_drv.c
index f6e0b5395051..5f83f9a3ef59 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_drv.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_drv.c
@@ -90,7 +90,7 @@ static int shmob_drm_setup_clocks(struct shmob_drm_device *sdev,
 		return -EINVAL;
 	}
 
-	clk = clk_get(sdev->dev, clkname);
+	clk = devm_clk_get(sdev->dev, clkname);
 	if (IS_ERR(clk)) {
 		dev_err(sdev->dev, "cannot get dot clock %s\n", clkname);
 		return PTR_ERR(clk);
@@ -106,21 +106,12 @@ static int shmob_drm_setup_clocks(struct shmob_drm_device *sdev,
 
 static int shmob_drm_unload(struct drm_device *dev)
 {
-	struct shmob_drm_device *sdev = dev->dev_private;
-
 	drm_kms_helper_poll_fini(dev);
 	drm_mode_config_cleanup(dev);
 	drm_vblank_cleanup(dev);
 	drm_irq_uninstall(dev);
 
-	if (sdev->clock)
-		clk_put(sdev->clock);
-
-	if (sdev->mmio)
-		iounmap(sdev->mmio);
-
 	dev->dev_private = NULL;
-	kfree(sdev);
 
 	return 0;
 }
@@ -139,7 +130,7 @@ static int shmob_drm_load(struct drm_device *dev, unsigned long flags)
 		return -EINVAL;
 	}
 
-	sdev = kzalloc(sizeof(*sdev), GFP_KERNEL);
+	sdev = devm_kzalloc(&pdev->dev, sizeof(*sdev), GFP_KERNEL);
 	if (sdev == NULL) {
 		dev_err(dev->dev, "failed to allocate private data\n");
 		return -ENOMEM;
@@ -156,29 +147,28 @@ static int shmob_drm_load(struct drm_device *dev, unsigned long flags)
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	if (res == NULL) {
 		dev_err(&pdev->dev, "failed to get memory resource\n");
-		ret = -EINVAL;
-		goto done;
+		return -EINVAL;
 	}
 
-	sdev->mmio = ioremap_nocache(res->start, resource_size(res));
+	sdev->mmio = devm_ioremap_nocache(&pdev->dev, res->start,
+					  resource_size(res));
 	if (sdev->mmio == NULL) {
 		dev_err(&pdev->dev, "failed to remap memory resource\n");
-		ret = -ENOMEM;
-		goto done;
+		return -ENOMEM;
 	}
 
 	ret = shmob_drm_setup_clocks(sdev, pdata->clk_source);
 	if (ret < 0)
-		goto done;
+		return ret;
 
 	ret = shmob_drm_init_interface(sdev);
 	if (ret < 0)
-		goto done;
+		return ret;
 
 	ret = shmob_drm_modeset_init(sdev);
 	if (ret < 0) {
 		dev_err(&pdev->dev, "failed to initialize mode setting\n");
-		goto done;
+		return ret;
 	}
 
 	for (i = 0; i < 4; ++i) {
@@ -273,7 +263,8 @@ static const struct file_operations shmob_drm_fops = {
 };
 
 static struct drm_driver shmob_drm_driver = {
-	.driver_features	= DRIVER_HAVE_IRQ | DRIVER_GEM | DRIVER_MODESET,
+	.driver_features	= DRIVER_HAVE_IRQ | DRIVER_GEM | DRIVER_MODESET
+				| DRIVER_PRIME,
 	.load			= shmob_drm_load,
 	.unload			= shmob_drm_unload,
 	.preclose		= shmob_drm_preclose,
@@ -283,6 +274,15 @@ static struct drm_driver shmob_drm_driver = {
 	.disable_vblank		= shmob_drm_disable_vblank,
 	.gem_free_object	= drm_gem_cma_free_object,
 	.gem_vm_ops		= &drm_gem_cma_vm_ops,
+	.prime_handle_to_fd	= drm_gem_prime_handle_to_fd,
+	.prime_fd_to_handle	= drm_gem_prime_fd_to_handle,
+	.gem_prime_import	= drm_gem_prime_import,
+	.gem_prime_export	= drm_gem_prime_export,
+	.gem_prime_get_sg_table	= drm_gem_cma_prime_get_sg_table,
+	.gem_prime_import_sg_table = drm_gem_cma_prime_import_sg_table,
+	.gem_prime_vmap		= drm_gem_cma_prime_vmap,
+	.gem_prime_vunmap	= drm_gem_cma_prime_vunmap,
+	.gem_prime_mmap		= drm_gem_cma_prime_mmap,
 	.dumb_create		= drm_gem_cma_dumb_create,
 	.dumb_map_offset	= drm_gem_cma_dumb_map_offset,
 	.dumb_destroy		= drm_gem_cma_dumb_destroy,
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_kms.c b/drivers/gpu/drm/shmobile/shmob_drm_kms.c
index c291ee385b4f..fc0ef0ca7d04 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_kms.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_kms.c
@@ -116,7 +116,7 @@ shmob_drm_fb_create(struct drm_device *dev, struct drm_file *file_priv,
 	}
 
 	if (mode_cmd->pitches[0] & 7 || mode_cmd->pitches[0] >= 65536) {
-		dev_dbg(dev->dev, "valid pitch value %u\n",
+		dev_dbg(dev->dev, "invalid pitch value %u\n",
 			mode_cmd->pitches[0]);
 		return ERR_PTR(-EINVAL);
 	}
diff --git a/drivers/gpu/drm/shmobile/shmob_drm_plane.c b/drivers/gpu/drm/shmobile/shmob_drm_plane.c
index e1eb899b0288..060ae03e5f9b 100644
--- a/drivers/gpu/drm/shmobile/shmob_drm_plane.c
+++ b/drivers/gpu/drm/shmobile/shmob_drm_plane.c
@@ -166,7 +166,7 @@ void shmob_drm_plane_setup(struct drm_plane *plane)
 {
 	struct shmob_drm_plane *splane = to_shmob_plane(plane);
 
-	if (plane->fb == NULL || !plane->enabled)
+	if (plane->fb == NULL)
 		return;
 
 	__shmob_drm_plane_setup(splane, plane->fb);
@@ -221,11 +221,8 @@ static int shmob_drm_plane_disable(struct drm_plane *plane)
 
 static void shmob_drm_plane_destroy(struct drm_plane *plane)
 {
-	struct shmob_drm_plane *splane = to_shmob_plane(plane);
-
 	shmob_drm_plane_disable(plane);
 	drm_plane_cleanup(plane);
-	kfree(splane);
 }
 
 static const struct drm_plane_funcs shmob_drm_plane_funcs = {
@@ -251,7 +248,7 @@ int shmob_drm_plane_create(struct shmob_drm_device *sdev, unsigned int index)
 	struct shmob_drm_plane *splane;
 	int ret;
 
-	splane = kzalloc(sizeof(*splane), GFP_KERNEL);
+	splane = devm_kzalloc(sdev->dev, sizeof(*splane), GFP_KERNEL);
 	if (splane == NULL)
 		return -ENOMEM;
 
@@ -261,8 +258,6 @@ int shmob_drm_plane_create(struct shmob_drm_device *sdev, unsigned int index)
 	ret = drm_plane_init(sdev->ddev, &splane->plane, 1,
 			     &shmob_drm_plane_funcs, formats,
 			     ARRAY_SIZE(formats), false);
-	if (ret < 0)
-		kfree(splane);
 
 	return ret;
 }
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c
index 5dd3c7d031d5..7418dcd986d3 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c
@@ -42,7 +42,8 @@ struct tilcdc_crtc {
 
 static void unref_worker(struct work_struct *work)
 {
-	struct tilcdc_crtc *tilcdc_crtc = container_of(work, struct tilcdc_crtc, work);
+	struct tilcdc_crtc *tilcdc_crtc =
+		container_of(work, struct tilcdc_crtc, work);
 	struct drm_device *dev = tilcdc_crtc->base.dev;
 	struct drm_framebuffer *fb;
 
@@ -55,10 +56,12 @@ static void unref_worker(struct work_struct *work)
 static void set_scanout(struct drm_crtc *crtc, int n)
 {
 	static const uint32_t base_reg[] = {
-			LCDC_DMA_FB_BASE_ADDR_0_REG, LCDC_DMA_FB_BASE_ADDR_1_REG,
+			LCDC_DMA_FB_BASE_ADDR_0_REG,
+			LCDC_DMA_FB_BASE_ADDR_1_REG,
 	};
 	static const uint32_t ceil_reg[] = {
-			LCDC_DMA_FB_CEILING_ADDR_0_REG, LCDC_DMA_FB_CEILING_ADDR_1_REG,
+			LCDC_DMA_FB_CEILING_ADDR_0_REG,
+			LCDC_DMA_FB_CEILING_ADDR_1_REG,
 	};
 	static const uint32_t stat[] = {
 			LCDC_END_OF_FRAME0, LCDC_END_OF_FRAME1,
@@ -194,7 +197,8 @@ static void tilcdc_crtc_dpms(struct drm_crtc *crtc, int mode)
 		tilcdc_crtc->frame_done = false;
 		stop(crtc);
 
-		/* if necessary wait for framedone irq which will still come
+		/*
+		 * if necessary wait for framedone irq which will still come
 		 * before putting things to sleep..
 		 */
 		if (priv->rev == 2) {
@@ -289,17 +293,24 @@ static int tilcdc_crtc_mode_set(struct drm_crtc *crtc,
 	reg = tilcdc_read(dev, LCDC_RASTER_TIMING_2_REG) & ~0x000fff00;
 	reg |= LCDC_AC_BIAS_FREQUENCY(info->ac_bias) |
 		LCDC_AC_BIAS_TRANSITIONS_PER_INT(info->ac_bias_intrpt);
+
+	/*
+	 * subtract one from hfp, hbp, hsw because the hardware uses
+	 * a value of 0 as 1
+	 */
 	if (priv->rev == 2) {
-		reg |= (hfp & 0x300) >> 8;
-		reg |= (hbp & 0x300) >> 4;
-		reg |= (hsw & 0x3c0) << 21;
+		/* clear bits we're going to set */
+		reg &= ~0x78000033;
+		reg |= ((hfp-1) & 0x300) >> 8;
+		reg |= ((hbp-1) & 0x300) >> 4;
+		reg |= ((hsw-1) & 0x3c0) << 21;
 	}
 	tilcdc_write(dev, LCDC_RASTER_TIMING_2_REG, reg);
 
 	reg = (((mode->hdisplay >> 4) - 1) << 4) |
-		((hbp & 0xff) << 24) |
-		((hfp & 0xff) << 16) |
-		((hsw & 0x3f) << 10);
+		(((hbp-1) & 0xff) << 24) |
+		(((hfp-1) & 0xff) << 16) |
+		(((hsw-1) & 0x3f) << 10);
 	if (priv->rev == 2)
 		reg |= (((mode->hdisplay >> 4) - 1) & 0x40) >> 3;
 	tilcdc_write(dev, LCDC_RASTER_TIMING_0_REG, reg);
@@ -307,9 +318,24 @@ static int tilcdc_crtc_mode_set(struct drm_crtc *crtc,
 	reg = ((mode->vdisplay - 1) & 0x3ff) |
 		((vbp & 0xff) << 24) |
 		((vfp & 0xff) << 16) |
-		((vsw & 0x3f) << 10);
+		(((vsw-1) & 0x3f) << 10);
 	tilcdc_write(dev, LCDC_RASTER_TIMING_1_REG, reg);
 
+	/*
+	 * be sure to set Bit 10 for the V2 LCDC controller,
+	 * otherwise limited to 1024 pixels width, stopping
+	 * 1920x1080 being suppoted.
+	 */
+	if (priv->rev == 2) {
+		if ((mode->vdisplay - 1) & 0x400) {
+			tilcdc_set(dev, LCDC_RASTER_TIMING_2_REG,
+				LCDC_LPP_B10);
+		} else {
+			tilcdc_clear(dev, LCDC_RASTER_TIMING_2_REG,
+				LCDC_LPP_B10);
+		}
+	}
+
 	/* Configure display type: */
 	reg = tilcdc_read(dev, LCDC_RASTER_CTRL_REG) &
 		~(LCDC_TFT_MODE | LCDC_MONO_8BIT_MODE | LCDC_MONOCHROME_MODE |
@@ -384,10 +410,6 @@ static int tilcdc_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 	return 0;
 }
 
-static void tilcdc_crtc_load_lut(struct drm_crtc *crtc)
-{
-}
-
 static const struct drm_crtc_funcs tilcdc_crtc_funcs = {
 		.destroy        = tilcdc_crtc_destroy,
 		.set_config     = drm_crtc_helper_set_config,
@@ -401,7 +423,6 @@ static const struct drm_crtc_helper_funcs tilcdc_crtc_helper_funcs = {
 		.commit         = tilcdc_crtc_commit,
 		.mode_set       = tilcdc_crtc_mode_set,
 		.mode_set_base  = tilcdc_crtc_mode_set_base,
-		.load_lut       = tilcdc_crtc_load_lut,
 };
 
 int tilcdc_crtc_max_width(struct drm_crtc *crtc)
@@ -422,7 +443,12 @@ int tilcdc_crtc_mode_valid(struct drm_crtc *crtc, struct drm_display_mode *mode)
 {
 	struct tilcdc_drm_private *priv = crtc->dev->dev_private;
 	unsigned int bandwidth;
+	uint32_t hbp, hfp, hsw, vbp, vfp, vsw;
 
+	/*
+	 * check to see if the width is within the range that
+	 * the LCD Controller physically supports
+	 */
 	if (mode->hdisplay > tilcdc_crtc_max_width(crtc))
 		return MODE_VIRTUAL_X;
 
@@ -433,10 +459,70 @@ int tilcdc_crtc_mode_valid(struct drm_crtc *crtc, struct drm_display_mode *mode)
 	if (mode->vdisplay > 2048)
 		return MODE_VIRTUAL_Y;
 
+	DBG("Processing mode %dx%d@%d with pixel clock %d",
+		mode->hdisplay, mode->vdisplay,
+		drm_mode_vrefresh(mode), mode->clock);
+
+	hbp = mode->htotal - mode->hsync_end;
+	hfp = mode->hsync_start - mode->hdisplay;
+	hsw = mode->hsync_end - mode->hsync_start;
+	vbp = mode->vtotal - mode->vsync_end;
+	vfp = mode->vsync_start - mode->vdisplay;
+	vsw = mode->vsync_end - mode->vsync_start;
+
+	if ((hbp-1) & ~0x3ff) {
+		DBG("Pruning mode: Horizontal Back Porch out of range");
+		return MODE_HBLANK_WIDE;
+	}
+
+	if ((hfp-1) & ~0x3ff) {
+		DBG("Pruning mode: Horizontal Front Porch out of range");
+		return MODE_HBLANK_WIDE;
+	}
+
+	if ((hsw-1) & ~0x3ff) {
+		DBG("Pruning mode: Horizontal Sync Width out of range");
+		return MODE_HSYNC_WIDE;
+	}
+
+	if (vbp & ~0xff) {
+		DBG("Pruning mode: Vertical Back Porch out of range");
+		return MODE_VBLANK_WIDE;
+	}
+
+	if (vfp & ~0xff) {
+		DBG("Pruning mode: Vertical Front Porch out of range");
+		return MODE_VBLANK_WIDE;
+	}
+
+	if ((vsw-1) & ~0x3f) {
+		DBG("Pruning mode: Vertical Sync Width out of range");
+		return MODE_VSYNC_WIDE;
+	}
+
+	/*
+	 * some devices have a maximum allowed pixel clock
+	 * configured from the DT
+	 */
+	if (mode->clock > priv->max_pixelclock) {
+		DBG("Pruning mode: pixel clock too high");
+		return MODE_CLOCK_HIGH;
+	}
+
+	/*
+	 * some devices further limit the max horizontal resolution
+	 * configured from the DT
+	 */
+	if (mode->hdisplay > priv->max_width)
+		return MODE_BAD_WIDTH;
+
 	/* filter out modes that would require too much memory bandwidth: */
-	bandwidth = mode->hdisplay * mode->vdisplay * drm_mode_vrefresh(mode);
-	if (bandwidth > priv->max_bandwidth)
+	bandwidth = mode->hdisplay * mode->vdisplay *
+		drm_mode_vrefresh(mode);
+	if (bandwidth > priv->max_bandwidth) {
+		DBG("Pruning mode: exceeds defined bandwidth limit");
 		return MODE_BAD;
+	}
 
 	return MODE_OK;
 }
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.c b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
index 2b5461bcd9fb..40b71da5a214 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_drv.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.c
@@ -26,6 +26,7 @@
 #include "drm_fb_helper.h"
 
 static LIST_HEAD(module_list);
+static bool slave_probing;
 
 void tilcdc_module_init(struct tilcdc_module *mod, const char *name,
 		const struct tilcdc_module_ops *funcs)
@@ -41,6 +42,11 @@ void tilcdc_module_cleanup(struct tilcdc_module *mod)
 	list_del(&mod->list);
 }
 
+void tilcdc_slave_probedefer(bool defered)
+{
+	slave_probing = defered;
+}
+
 static struct of_device_id tilcdc_of_match[];
 
 static struct drm_framebuffer *tilcdc_fb_create(struct drm_device *dev,
@@ -157,7 +163,9 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 	struct platform_device *pdev = dev->platformdev;
 	struct device_node *node = pdev->dev.of_node;
 	struct tilcdc_drm_private *priv;
+	struct tilcdc_module *mod;
 	struct resource *res;
+	u32 bpp = 0;
 	int ret;
 
 	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
@@ -210,7 +218,20 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 #endif
 
 	if (of_property_read_u32(node, "max-bandwidth", &priv->max_bandwidth))
-		priv->max_bandwidth = 1280 * 1024 * 60;
+		priv->max_bandwidth = TILCDC_DEFAULT_MAX_BANDWIDTH;
+
+	DBG("Maximum Bandwidth Value %d", priv->max_bandwidth);
+
+	if (of_property_read_u32(node, "ti,max-width", &priv->max_width))
+		priv->max_width = TILCDC_DEFAULT_MAX_WIDTH;
+
+	DBG("Maximum Horizontal Pixel Width Value %dpixels", priv->max_width);
+
+	if (of_property_read_u32(node, "ti,max-pixelclock",
+					&priv->max_pixelclock))
+		priv->max_pixelclock = TILCDC_DEFAULT_MAX_PIXELCLOCK;
+
+	DBG("Maximum Pixel Clock Value %dKHz", priv->max_pixelclock);
 
 	pm_runtime_enable(dev->dev);
 
@@ -256,7 +277,15 @@ static int tilcdc_load(struct drm_device *dev, unsigned long flags)
 
 	platform_set_drvdata(pdev, dev);
 
-	priv->fbdev = drm_fbdev_cma_init(dev, 16,
+
+	list_for_each_entry(mod, &module_list, list) {
+		DBG("%s: preferred_bpp: %d", mod->name, mod->preferred_bpp);
+		bpp = mod->preferred_bpp;
+		if (bpp > 0)
+			break;
+	}
+
+	priv->fbdev = drm_fbdev_cma_init(dev, bpp,
 			dev->mode_config.num_crtc,
 			dev->mode_config.num_connector);
 
@@ -557,6 +586,10 @@ static int tilcdc_pdev_probe(struct platform_device *pdev)
 		return -ENXIO;
 	}
 
+	/* defer probing if slave is in deferred probing */
+	if (slave_probing == true)
+		return -EPROBE_DEFER;
+
 	return drm_platform_init(&tilcdc_driver, pdev);
 }
 
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.h b/drivers/gpu/drm/tilcdc/tilcdc_drv.h
index 8242b5a4307b..093803683b25 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_drv.h
+++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.h
@@ -34,6 +34,18 @@
 #include <drm/drm_gem_cma_helper.h>
 #include <drm/drm_fb_cma_helper.h>
 
+/* Defaulting to pixel clock defined on AM335x */
+#define TILCDC_DEFAULT_MAX_PIXELCLOCK  126000
+/* Defaulting to max width as defined on AM335x */
+#define TILCDC_DEFAULT_MAX_WIDTH  2048
+/*
+ * This may need some tweaking, but want to allow at least 1280x1024@60
+ * with optimized DDR & EMIF settings tweaked 1920x1080@24 appears to
+ * be supportable
+ */
+#define TILCDC_DEFAULT_MAX_BANDWIDTH  (1280*1024*60)
+
+
 struct tilcdc_drm_private {
 	void __iomem *mmio;
 
@@ -43,6 +55,16 @@ struct tilcdc_drm_private {
 
 	/* don't attempt resolutions w/ higher W * H * Hz: */
 	uint32_t max_bandwidth;
+	/*
+	 * Pixel Clock will be restricted to some value as
+	 * defined in the device datasheet measured in KHz
+	 */
+	uint32_t max_pixelclock;
+	/*
+	 * Max allowable width is limited on a per device basis
+	 * measured in pixels
+	 */
+	uint32_t max_width;
 
 	/* register contents saved across suspend/resume: */
 	u32 saved_register[12];
@@ -89,12 +111,13 @@ struct tilcdc_module {
 	const char *name;
 	struct list_head list;
 	const struct tilcdc_module_ops *funcs;
+	unsigned int preferred_bpp;
 };
 
 void tilcdc_module_init(struct tilcdc_module *mod, const char *name,
 		const struct tilcdc_module_ops *funcs);
 void tilcdc_module_cleanup(struct tilcdc_module *mod);
-
+void tilcdc_slave_probedefer(bool defered);
 
 /* Panel config that needs to be set in the crtc, but is not coming from
  * the mode timings.  The display module is expected to call
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_panel.c b/drivers/gpu/drm/tilcdc/tilcdc_panel.c
index 09176654fddb..86c67329b605 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_panel.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_panel.c
@@ -393,6 +393,8 @@ static int panel_probe(struct platform_device *pdev)
 		goto fail;
 	}
 
+	mod->preferred_bpp = panel_mod->info->bpp;
+
 	panel_mod->backlight = of_find_backlight_by_node(node);
 	if (panel_mod->backlight)
 		dev_info(&pdev->dev, "found backlight\n");
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_regs.h b/drivers/gpu/drm/tilcdc/tilcdc_regs.h
index 17fd1b45428a..1bf5e2553acc 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_regs.h
+++ b/drivers/gpu/drm/tilcdc/tilcdc_regs.h
@@ -80,6 +80,7 @@
 #define LCDC_INVERT_PIXEL_CLOCK                  BIT(22)
 #define LCDC_INVERT_HSYNC                        BIT(21)
 #define LCDC_INVERT_VSYNC                        BIT(20)
+#define LCDC_LPP_B10                             BIT(26)
 
 /* LCDC Block */
 #define LCDC_PID_REG                             0x0
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_slave.c b/drivers/gpu/drm/tilcdc/tilcdc_slave.c
index db1d2fc9dfb5..dfffaf014022 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_slave.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_slave.c
@@ -298,6 +298,7 @@ static int slave_probe(struct platform_device *pdev)
 	struct tilcdc_module *mod;
 	struct pinctrl *pinctrl;
 	uint32_t i2c_phandle;
+	struct i2c_adapter *slavei2c;
 	int ret = -EINVAL;
 
 	/* bail out early if no DT data: */
@@ -306,42 +307,48 @@ static int slave_probe(struct platform_device *pdev)
 		return -ENXIO;
 	}
 
-	slave_mod = kzalloc(sizeof(*slave_mod), GFP_KERNEL);
-	if (!slave_mod)
-		return -ENOMEM;
-
-	mod = &slave_mod->base;
-
-	tilcdc_module_init(mod, "slave", &slave_module_ops);
-
-	pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
-	if (IS_ERR(pinctrl))
-		dev_warn(&pdev->dev, "pins are not configured\n");
-
+	/* Bail out early if i2c not specified */
 	if (of_property_read_u32(node, "i2c", &i2c_phandle)) {
 		dev_err(&pdev->dev, "could not get i2c bus phandle\n");
-		goto fail;
+		return ret;
 	}
 
 	i2c_node = of_find_node_by_phandle(i2c_phandle);
 	if (!i2c_node) {
 		dev_err(&pdev->dev, "could not get i2c bus node\n");
-		goto fail;
+		return ret;
 	}
 
-	slave_mod->i2c = of_find_i2c_adapter_by_node(i2c_node);
-	if (!slave_mod->i2c) {
+	/* but defer the probe if it can't be initialized it might come later */
+	slavei2c = of_find_i2c_adapter_by_node(i2c_node);
+	of_node_put(i2c_node);
+
+	if (!slavei2c) {
+		ret = -EPROBE_DEFER;
+		tilcdc_slave_probedefer(true);
 		dev_err(&pdev->dev, "could not get i2c\n");
-		goto fail;
+		return ret;
 	}
 
-	of_node_put(i2c_node);
+	slave_mod = kzalloc(sizeof(*slave_mod), GFP_KERNEL);
+	if (!slave_mod)
+		return -ENOMEM;
 
-	return 0;
+	mod = &slave_mod->base;
 
-fail:
-	slave_destroy(mod);
-	return ret;
+	mod->preferred_bpp = slave_info.bpp;
+
+	slave_mod->i2c = slavei2c;
+
+	tilcdc_module_init(mod, "slave", &slave_module_ops);
+
+	pinctrl = devm_pinctrl_get_select_default(&pdev->dev);
+	if (IS_ERR(pinctrl))
+		dev_warn(&pdev->dev, "pins are not configured\n");
+
+	tilcdc_slave_probedefer(false);
+
+	return 0;
 }
 
 static int slave_remove(struct platform_device *pdev)
diff --git a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
index a36788fbcd98..925c7cddeff9 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
+++ b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c
@@ -354,6 +354,8 @@ static int tfp410_probe(struct platform_device *pdev)
 		goto fail;
 	}
 
+	mod->preferred_bpp = dvi_info.bpp;
+
 	i2c_node = of_find_node_by_phandle(i2c_phandle);
 	if (!i2c_node) {
 		dev_err(&pdev->dev, "could not get i2c bus node\n");
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 9b07b7d44a58..cb9dd674670c 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -150,6 +150,9 @@ static void ttm_bo_release_list(struct kref *list_kref)
 	if (bo->ttm)
 		ttm_tt_destroy(bo->ttm);
 	atomic_dec(&bo->glob->bo_count);
+	if (bo->resv == &bo->ttm_resv)
+		reservation_object_fini(&bo->ttm_resv);
+
 	if (bo->destroy)
 		bo->destroy(bo);
 	else {
@@ -158,24 +161,12 @@ static void ttm_bo_release_list(struct kref *list_kref)
 	ttm_mem_global_free(bdev->glob->mem_glob, acc_size);
 }
 
-static int ttm_bo_wait_unreserved(struct ttm_buffer_object *bo,
-				  bool interruptible)
-{
-	if (interruptible) {
-		return wait_event_interruptible(bo->event_queue,
-					       !ttm_bo_is_reserved(bo));
-	} else {
-		wait_event(bo->event_queue, !ttm_bo_is_reserved(bo));
-		return 0;
-	}
-}
-
 void ttm_bo_add_to_lru(struct ttm_buffer_object *bo)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
 	struct ttm_mem_type_manager *man;
 
-	BUG_ON(!ttm_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->resv->lock.base);
 
 	if (!(bo->mem.placement & TTM_PL_FLAG_NO_EVICT)) {
 
@@ -191,6 +182,7 @@ void ttm_bo_add_to_lru(struct ttm_buffer_object *bo)
 		}
 	}
 }
+EXPORT_SYMBOL(ttm_bo_add_to_lru);
 
 int ttm_bo_del_from_lru(struct ttm_buffer_object *bo)
 {
@@ -213,71 +205,6 @@ int ttm_bo_del_from_lru(struct ttm_buffer_object *bo)
 	return put_count;
 }
 
-int ttm_bo_reserve_nolru(struct ttm_buffer_object *bo,
-			  bool interruptible,
-			  bool no_wait, bool use_sequence, uint32_t sequence)
-{
-	int ret;
-
-	while (unlikely(atomic_xchg(&bo->reserved, 1) != 0)) {
-		/**
-		 * Deadlock avoidance for multi-bo reserving.
-		 */
-		if (use_sequence && bo->seq_valid) {
-			/**
-			 * We've already reserved this one.
-			 */
-			if (unlikely(sequence == bo->val_seq))
-				return -EDEADLK;
-			/**
-			 * Already reserved by a thread that will not back
-			 * off for us. We need to back off.
-			 */
-			if (unlikely(sequence - bo->val_seq < (1 << 31)))
-				return -EAGAIN;
-		}
-
-		if (no_wait)
-			return -EBUSY;
-
-		ret = ttm_bo_wait_unreserved(bo, interruptible);
-
-		if (unlikely(ret))
-			return ret;
-	}
-
-	if (use_sequence) {
-		bool wake_up = false;
-		/**
-		 * Wake up waiters that may need to recheck for deadlock,
-		 * if we decreased the sequence number.
-		 */
-		if (unlikely((bo->val_seq - sequence < (1 << 31))
-			     || !bo->seq_valid))
-			wake_up = true;
-
-		/*
-		 * In the worst case with memory ordering these values can be
-		 * seen in the wrong order. However since we call wake_up_all
-		 * in that case, this will hopefully not pose a problem,
-		 * and the worst case would only cause someone to accidentally
-		 * hit -EAGAIN in ttm_bo_reserve when they see old value of
-		 * val_seq. However this would only happen if seq_valid was
-		 * written before val_seq was, and just means some slightly
-		 * increased cpu usage
-		 */
-		bo->val_seq = sequence;
-		bo->seq_valid = true;
-		if (wake_up)
-			wake_up_all(&bo->event_queue);
-	} else {
-		bo->seq_valid = false;
-	}
-
-	return 0;
-}
-EXPORT_SYMBOL(ttm_bo_reserve);
-
 static void ttm_bo_ref_bug(struct kref *list_kref)
 {
 	BUG();
@@ -290,89 +217,16 @@ void ttm_bo_list_ref_sub(struct ttm_buffer_object *bo, int count,
 		 (never_free) ? ttm_bo_ref_bug : ttm_bo_release_list);
 }
 
-int ttm_bo_reserve(struct ttm_buffer_object *bo,
-		   bool interruptible,
-		   bool no_wait, bool use_sequence, uint32_t sequence)
-{
-	struct ttm_bo_global *glob = bo->glob;
-	int put_count = 0;
-	int ret;
-
-	ret = ttm_bo_reserve_nolru(bo, interruptible, no_wait, use_sequence,
-				   sequence);
-	if (likely(ret == 0)) {
-		spin_lock(&glob->lru_lock);
-		put_count = ttm_bo_del_from_lru(bo);
-		spin_unlock(&glob->lru_lock);
-		ttm_bo_list_ref_sub(bo, put_count, true);
-	}
-
-	return ret;
-}
-
-int ttm_bo_reserve_slowpath_nolru(struct ttm_buffer_object *bo,
-				  bool interruptible, uint32_t sequence)
-{
-	bool wake_up = false;
-	int ret;
-
-	while (unlikely(atomic_xchg(&bo->reserved, 1) != 0)) {
-		WARN_ON(bo->seq_valid && sequence == bo->val_seq);
-
-		ret = ttm_bo_wait_unreserved(bo, interruptible);
-
-		if (unlikely(ret))
-			return ret;
-	}
-
-	if ((bo->val_seq - sequence < (1 << 31)) || !bo->seq_valid)
-		wake_up = true;
-
-	/**
-	 * Wake up waiters that may need to recheck for deadlock,
-	 * if we decreased the sequence number.
-	 */
-	bo->val_seq = sequence;
-	bo->seq_valid = true;
-	if (wake_up)
-		wake_up_all(&bo->event_queue);
-
-	return 0;
-}
-
-int ttm_bo_reserve_slowpath(struct ttm_buffer_object *bo,
-			    bool interruptible, uint32_t sequence)
-{
-	struct ttm_bo_global *glob = bo->glob;
-	int put_count, ret;
-
-	ret = ttm_bo_reserve_slowpath_nolru(bo, interruptible, sequence);
-	if (likely(!ret)) {
-		spin_lock(&glob->lru_lock);
-		put_count = ttm_bo_del_from_lru(bo);
-		spin_unlock(&glob->lru_lock);
-		ttm_bo_list_ref_sub(bo, put_count, true);
-	}
-	return ret;
-}
-EXPORT_SYMBOL(ttm_bo_reserve_slowpath);
-
-void ttm_bo_unreserve_locked(struct ttm_buffer_object *bo)
-{
-	ttm_bo_add_to_lru(bo);
-	atomic_set(&bo->reserved, 0);
-	wake_up_all(&bo->event_queue);
-}
-
-void ttm_bo_unreserve(struct ttm_buffer_object *bo)
+void ttm_bo_del_sub_from_lru(struct ttm_buffer_object *bo)
 {
-	struct ttm_bo_global *glob = bo->glob;
+	int put_count;
 
-	spin_lock(&glob->lru_lock);
-	ttm_bo_unreserve_locked(bo);
-	spin_unlock(&glob->lru_lock);
+	spin_lock(&bo->glob->lru_lock);
+	put_count = ttm_bo_del_from_lru(bo);
+	spin_unlock(&bo->glob->lru_lock);
+	ttm_bo_list_ref_sub(bo, put_count, true);
 }
-EXPORT_SYMBOL(ttm_bo_unreserve);
+EXPORT_SYMBOL(ttm_bo_del_sub_from_lru);
 
 /*
  * Call bo->mutex locked.
@@ -544,17 +398,7 @@ static void ttm_bo_cleanup_memtype_use(struct ttm_buffer_object *bo)
 	}
 	ttm_bo_mem_put(bo, &bo->mem);
 
-	atomic_set(&bo->reserved, 0);
-	wake_up_all(&bo->event_queue);
-
-	/*
-	 * Since the final reference to this bo may not be dropped by
-	 * the current task we have to put a memory barrier here to make
-	 * sure the changes done in this function are always visible.
-	 *
-	 * This function only needs protection against the final kref_put.
-	 */
-	smp_mb__before_atomic_dec();
+	ww_mutex_unlock (&bo->resv->lock);
 }
 
 static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
@@ -586,10 +430,8 @@ static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
 		sync_obj = driver->sync_obj_ref(bo->sync_obj);
 	spin_unlock(&bdev->fence_lock);
 
-	if (!ret) {
-		atomic_set(&bo->reserved, 0);
-		wake_up_all(&bo->event_queue);
-	}
+	if (!ret)
+		ww_mutex_unlock(&bo->resv->lock);
 
 	kref_get(&bo->list_kref);
 	list_add_tail(&bo->ddestroy, &bdev->ddestroy);
@@ -639,8 +481,7 @@ static int ttm_bo_cleanup_refs_and_unlock(struct ttm_buffer_object *bo,
 		sync_obj = driver->sync_obj_ref(bo->sync_obj);
 		spin_unlock(&bdev->fence_lock);
 
-		atomic_set(&bo->reserved, 0);
-		wake_up_all(&bo->event_queue);
+		ww_mutex_unlock(&bo->resv->lock);
 		spin_unlock(&glob->lru_lock);
 
 		ret = driver->sync_obj_wait(sync_obj, false, interruptible);
@@ -678,8 +519,7 @@ static int ttm_bo_cleanup_refs_and_unlock(struct ttm_buffer_object *bo,
 		spin_unlock(&bdev->fence_lock);
 
 	if (ret || unlikely(list_empty(&bo->ddestroy))) {
-		atomic_set(&bo->reserved, 0);
-		wake_up_all(&bo->event_queue);
+		ww_mutex_unlock(&bo->resv->lock);
 		spin_unlock(&glob->lru_lock);
 		return ret;
 	}
@@ -831,7 +671,7 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 		goto out;
 	}
 
-	BUG_ON(!ttm_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->resv->lock.base);
 
 	evict_mem = bo->mem;
 	evict_mem.mm_node = NULL;
@@ -1121,7 +961,7 @@ int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 	struct ttm_mem_reg mem;
 	struct ttm_bo_device *bdev = bo->bdev;
 
-	BUG_ON(!ttm_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->resv->lock.base);
 
 	/*
 	 * FIXME: It's possible to pipeline buffer moves.
@@ -1180,7 +1020,7 @@ int ttm_bo_validate(struct ttm_buffer_object *bo,
 {
 	int ret;
 
-	BUG_ON(!ttm_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->resv->lock.base);
 	/* Check that range is valid */
 	if (placement->lpfn || placement->fpfn)
 		if (placement->fpfn > placement->lpfn ||
@@ -1239,6 +1079,7 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	int ret = 0;
 	unsigned long num_pages;
 	struct ttm_mem_global *mem_glob = bdev->glob->mem_glob;
+	bool locked;
 
 	ret = ttm_mem_global_alloc(mem_glob, acc_size, false, false);
 	if (ret) {
@@ -1265,8 +1106,6 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	kref_init(&bo->kref);
 	kref_init(&bo->list_kref);
 	atomic_set(&bo->cpu_writers, 0);
-	atomic_set(&bo->reserved, 1);
-	init_waitqueue_head(&bo->event_queue);
 	INIT_LIST_HEAD(&bo->lru);
 	INIT_LIST_HEAD(&bo->ddestroy);
 	INIT_LIST_HEAD(&bo->swap);
@@ -1284,37 +1123,34 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	bo->mem.bus.io_reserved_count = 0;
 	bo->priv_flags = 0;
 	bo->mem.placement = (TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED);
-	bo->seq_valid = false;
 	bo->persistent_swap_storage = persistent_swap_storage;
 	bo->acc_size = acc_size;
 	bo->sg = sg;
+	bo->resv = &bo->ttm_resv;
+	reservation_object_init(bo->resv);
 	atomic_inc(&bo->glob->bo_count);
 
 	ret = ttm_bo_check_placement(bo, placement);
-	if (unlikely(ret != 0))
-		goto out_err;
 
 	/*
 	 * For ttm_bo_type_device buffers, allocate
 	 * address space from the device.
 	 */
-	if (bo->type == ttm_bo_type_device ||
-	    bo->type == ttm_bo_type_sg) {
+	if (likely(!ret) &&
+	    (bo->type == ttm_bo_type_device ||
+	     bo->type == ttm_bo_type_sg))
 		ret = ttm_bo_setup_vm(bo);
-		if (ret)
-			goto out_err;
-	}
 
-	ret = ttm_bo_validate(bo, placement, interruptible, false);
-	if (ret)
-		goto out_err;
+	locked = ww_mutex_trylock(&bo->resv->lock);
+	WARN_ON(!locked);
 
-	ttm_bo_unreserve(bo);
-	return 0;
+	if (likely(!ret))
+		ret = ttm_bo_validate(bo, placement, interruptible, false);
 
-out_err:
 	ttm_bo_unreserve(bo);
-	ttm_bo_unref(&bo);
+
+	if (unlikely(ret))
+		ttm_bo_unref(&bo);
 
 	return ret;
 }
@@ -1619,9 +1455,7 @@ int ttm_bo_device_init(struct ttm_bo_device *bdev,
 		goto out_no_sys;
 
 	bdev->addr_space_rb = RB_ROOT;
-	ret = drm_mm_init(&bdev->addr_space_mm, file_page_offset, 0x10000000);
-	if (unlikely(ret != 0))
-		goto out_no_addr_mm;
+	drm_mm_init(&bdev->addr_space_mm, file_page_offset, 0x10000000);
 
 	INIT_DELAYED_WORK(&bdev->wq, ttm_bo_delayed_workqueue);
 	INIT_LIST_HEAD(&bdev->ddestroy);
@@ -1635,8 +1469,6 @@ int ttm_bo_device_init(struct ttm_bo_device *bdev,
 	mutex_unlock(&glob->device_list_mutex);
 
 	return 0;
-out_no_addr_mm:
-	ttm_bo_clean_mm(bdev, 0);
 out_no_sys:
 	return ret;
 }
@@ -1927,8 +1759,7 @@ out:
 	 * already swapped buffer.
 	 */
 
-	atomic_set(&bo->reserved, 0);
-	wake_up_all(&bo->event_queue);
+	ww_mutex_unlock(&bo->resv->lock);
 	kref_put(&bo->list_kref, ttm_bo_release_list);
 	return ret;
 }
diff --git a/drivers/gpu/drm/ttm/ttm_bo_manager.c b/drivers/gpu/drm/ttm/ttm_bo_manager.c
index 9212494e9072..e4367f91472a 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_manager.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_manager.c
@@ -103,18 +103,12 @@ static int ttm_bo_man_init(struct ttm_mem_type_manager *man,
 			   unsigned long p_size)
 {
 	struct ttm_range_manager *rman;
-	int ret;
 
 	rman = kzalloc(sizeof(*rman), GFP_KERNEL);
 	if (!rman)
 		return -ENOMEM;
 
-	ret = drm_mm_init(&rman->mm, 0, p_size);
-	if (ret) {
-		kfree(rman);
-		return ret;
-	}
-
+	drm_mm_init(&rman->mm, 0, p_size);
 	spin_lock_init(&rman->lock);
 	man->priv = rman;
 	return 0;
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index af894584dd90..319cf4127c5b 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -433,6 +433,7 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	struct ttm_buffer_object *fbo;
 	struct ttm_bo_device *bdev = bo->bdev;
 	struct ttm_bo_driver *driver = bdev->driver;
+	int ret;
 
 	fbo = kmalloc(sizeof(*fbo), GFP_KERNEL);
 	if (!fbo)
@@ -445,7 +446,6 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	 * TODO: Explicit member copy would probably be better here.
 	 */
 
-	init_waitqueue_head(&fbo->event_queue);
 	INIT_LIST_HEAD(&fbo->ddestroy);
 	INIT_LIST_HEAD(&fbo->lru);
 	INIT_LIST_HEAD(&fbo->swap);
@@ -463,6 +463,10 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	kref_init(&fbo->kref);
 	fbo->destroy = &ttm_transfered_destroy;
 	fbo->acc_size = 0;
+	fbo->resv = &fbo->ttm_resv;
+	reservation_object_init(fbo->resv);
+	ret = ww_mutex_trylock(&fbo->resv->lock);
+	WARN_ON(!ret);
 
 	*new_obj = fbo;
 	return 0;
diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
index 7b90def15674..6c911789ae5c 100644
--- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c
+++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
@@ -32,7 +32,8 @@
 #include <linux/sched.h>
 #include <linux/module.h>
 
-static void ttm_eu_backoff_reservation_locked(struct list_head *list)
+static void ttm_eu_backoff_reservation_locked(struct list_head *list,
+					      struct ww_acquire_ctx *ticket)
 {
 	struct ttm_validate_buffer *entry;
 
@@ -41,14 +42,12 @@ static void ttm_eu_backoff_reservation_locked(struct list_head *list)
 		if (!entry->reserved)
 			continue;
 
+		entry->reserved = false;
 		if (entry->removed) {
 			ttm_bo_add_to_lru(bo);
 			entry->removed = false;
-
 		}
-		entry->reserved = false;
-		atomic_set(&bo->reserved, 0);
-		wake_up_all(&bo->event_queue);
+		ww_mutex_unlock(&bo->resv->lock);
 	}
 }
 
@@ -82,7 +81,8 @@ static void ttm_eu_list_ref_sub(struct list_head *list)
 	}
 }
 
-void ttm_eu_backoff_reservation(struct list_head *list)
+void ttm_eu_backoff_reservation(struct ww_acquire_ctx *ticket,
+				struct list_head *list)
 {
 	struct ttm_validate_buffer *entry;
 	struct ttm_bo_global *glob;
@@ -93,7 +93,8 @@ void ttm_eu_backoff_reservation(struct list_head *list)
 	entry = list_first_entry(list, struct ttm_validate_buffer, head);
 	glob = entry->bo->glob;
 	spin_lock(&glob->lru_lock);
-	ttm_eu_backoff_reservation_locked(list);
+	ttm_eu_backoff_reservation_locked(list, ticket);
+	ww_acquire_fini(ticket);
 	spin_unlock(&glob->lru_lock);
 }
 EXPORT_SYMBOL(ttm_eu_backoff_reservation);
@@ -110,12 +111,12 @@ EXPORT_SYMBOL(ttm_eu_backoff_reservation);
  * buffers in different orders.
  */
 
-int ttm_eu_reserve_buffers(struct list_head *list)
+int ttm_eu_reserve_buffers(struct ww_acquire_ctx *ticket,
+			   struct list_head *list)
 {
 	struct ttm_bo_global *glob;
 	struct ttm_validate_buffer *entry;
 	int ret;
-	uint32_t val_seq;
 
 	if (list_empty(list))
 		return 0;
@@ -129,9 +130,7 @@ int ttm_eu_reserve_buffers(struct list_head *list)
 	entry = list_first_entry(list, struct ttm_validate_buffer, head);
 	glob = entry->bo->glob;
 
-	spin_lock(&glob->lru_lock);
-	val_seq = entry->bo->bdev->val_seq++;
-
+	ww_acquire_init(ticket, &reservation_ww_class);
 retry:
 	list_for_each_entry(entry, list, head) {
 		struct ttm_buffer_object *bo = entry->bo;
@@ -140,49 +139,34 @@ retry:
 		if (entry->reserved)
 			continue;
 
-		ret = ttm_bo_reserve_nolru(bo, true, true, true, val_seq);
-		switch (ret) {
-		case 0:
-			break;
-		case -EBUSY:
-			ttm_eu_del_from_lru_locked(list);
-			spin_unlock(&glob->lru_lock);
-			ret = ttm_bo_reserve_nolru(bo, true, false,
-						   true, val_seq);
-			spin_lock(&glob->lru_lock);
-			if (!ret)
-				break;
-
-			if (unlikely(ret != -EAGAIN))
-				goto err;
 
-			/* fallthrough */
-		case -EAGAIN:
-			ttm_eu_backoff_reservation_locked(list);
+		ret = ttm_bo_reserve_nolru(bo, true, false, true, ticket);
 
-			/*
-			 * temporarily increase sequence number every retry,
-			 * to prevent us from seeing our old reservation
-			 * sequence when someone else reserved the buffer,
-			 * but hasn't updated the seq_valid/seqno members yet.
+		if (ret == -EDEADLK) {
+			/* uh oh, we lost out, drop every reservation and try
+			 * to only reserve this buffer, then start over if
+			 * this succeeds.
 			 */
-			val_seq = entry->bo->bdev->val_seq++;
-
+			spin_lock(&glob->lru_lock);
+			ttm_eu_backoff_reservation_locked(list, ticket);
 			spin_unlock(&glob->lru_lock);
 			ttm_eu_list_ref_sub(list);
-			ret = ttm_bo_reserve_slowpath_nolru(bo, true, val_seq);
-			if (unlikely(ret != 0))
-				return ret;
-			spin_lock(&glob->lru_lock);
+			ret = ww_mutex_lock_slow_interruptible(&bo->resv->lock,
+							       ticket);
+			if (unlikely(ret != 0)) {
+				if (ret == -EINTR)
+					ret = -ERESTARTSYS;
+				goto err_fini;
+			}
+
 			entry->reserved = true;
 			if (unlikely(atomic_read(&bo->cpu_writers) > 0)) {
 				ret = -EBUSY;
 				goto err;
 			}
 			goto retry;
-		default:
+		} else if (ret)
 			goto err;
-		}
 
 		entry->reserved = true;
 		if (unlikely(atomic_read(&bo->cpu_writers) > 0)) {
@@ -191,21 +175,27 @@ retry:
 		}
 	}
 
+	ww_acquire_done(ticket);
+	spin_lock(&glob->lru_lock);
 	ttm_eu_del_from_lru_locked(list);
 	spin_unlock(&glob->lru_lock);
 	ttm_eu_list_ref_sub(list);
-
 	return 0;
 
 err:
-	ttm_eu_backoff_reservation_locked(list);
+	spin_lock(&glob->lru_lock);
+	ttm_eu_backoff_reservation_locked(list, ticket);
 	spin_unlock(&glob->lru_lock);
 	ttm_eu_list_ref_sub(list);
+err_fini:
+	ww_acquire_done(ticket);
+	ww_acquire_fini(ticket);
 	return ret;
 }
 EXPORT_SYMBOL(ttm_eu_reserve_buffers);
 
-void ttm_eu_fence_buffer_objects(struct list_head *list, void *sync_obj)
+void ttm_eu_fence_buffer_objects(struct ww_acquire_ctx *ticket,
+				 struct list_head *list, void *sync_obj)
 {
 	struct ttm_validate_buffer *entry;
 	struct ttm_buffer_object *bo;
@@ -228,11 +218,13 @@ void ttm_eu_fence_buffer_objects(struct list_head *list, void *sync_obj)
 		bo = entry->bo;
 		entry->old_sync_obj = bo->sync_obj;
 		bo->sync_obj = driver->sync_obj_ref(sync_obj);
-		ttm_bo_unreserve_locked(bo);
+		ttm_bo_add_to_lru(bo);
+		ww_mutex_unlock(&bo->resv->lock);
 		entry->reserved = false;
 	}
 	spin_unlock(&bdev->fence_lock);
 	spin_unlock(&glob->lru_lock);
+	ww_acquire_fini(ticket);
 
 	list_for_each_entry(entry, list, head) {
 		if (entry->old_sync_obj)
diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c
index dc0c065f8d39..97e9d614700f 100644
--- a/drivers/gpu/drm/udl/udl_fb.c
+++ b/drivers/gpu/drm/udl/udl_fb.c
@@ -393,19 +393,6 @@ static struct fb_ops udlfb_ops = {
 	.fb_release = udl_fb_release,
 };
 
-static void udl_crtc_fb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green,
-			   u16 blue, int regno)
-{
-}
-
-static void udl_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green,
-			     u16 *blue, int regno)
-{
-	*red = 0;
-	*green = 0;
-	*blue = 0;
-}
-
 static int udl_user_framebuffer_dirty(struct drm_framebuffer *fb,
 				      struct drm_file *file,
 				      unsigned flags, unsigned color,
@@ -558,8 +545,6 @@ out:
 }
 
 static struct drm_fb_helper_funcs udl_fb_helper_funcs = {
-	.gamma_set = udl_crtc_fb_gamma_set,
-	.gamma_get = udl_crtc_fb_gamma_get,
 	.fb_probe = udlfb_create,
 };
 
diff --git a/drivers/gpu/drm/udl/udl_modeset.c b/drivers/gpu/drm/udl/udl_modeset.c
index e96d2349bd54..2ae1eb7d1635 100644
--- a/drivers/gpu/drm/udl/udl_modeset.c
+++ b/drivers/gpu/drm/udl/udl_modeset.c
@@ -363,10 +363,6 @@ static void udl_crtc_destroy(struct drm_crtc *crtc)
 	kfree(crtc);
 }
 
-static void udl_load_lut(struct drm_crtc *crtc)
-{
-}
-
 static void udl_crtc_prepare(struct drm_crtc *crtc)
 {
 }
@@ -383,7 +379,6 @@ static struct drm_crtc_helper_funcs udl_helper_funcs = {
 	.prepare = udl_crtc_prepare,
 	.commit = udl_crtc_commit,
 	.disable = udl_crtc_disable,
-	.load_lut = udl_load_lut,
 };
 
 static const struct drm_crtc_funcs udl_crtc_funcs = {
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
index 5fae06ad7e25..d4e54fcc0acd 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
@@ -302,7 +302,7 @@ void vmw_bo_pin(struct ttm_buffer_object *bo, bool pin)
 	uint32_t old_mem_type = bo->mem.mem_type;
 	int ret;
 
-	BUG_ON(!ttm_bo_is_reserved(bo));
+	lockdep_assert_held(&bo->resv->lock.base);
 	BUG_ON(old_mem_type != TTM_PL_VRAM &&
 	       old_mem_type != VMW_PL_GMR);
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index 07dfd823cc30..78e21649d48a 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -565,8 +565,8 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 		dev_priv->has_gmr = false;
 	}
 
-	dev_priv->mmio_mtrr = drm_mtrr_add(dev_priv->mmio_start,
-					   dev_priv->mmio_size, DRM_MTRR_WC);
+	dev_priv->mmio_mtrr = arch_phys_wc_add(dev_priv->mmio_start,
+					       dev_priv->mmio_size);
 
 	dev_priv->mmio_virt = ioremap_wc(dev_priv->mmio_start,
 					 dev_priv->mmio_size);
@@ -664,8 +664,7 @@ out_no_device:
 out_err4:
 	iounmap(dev_priv->mmio_virt);
 out_err3:
-	drm_mtrr_del(dev_priv->mmio_mtrr, dev_priv->mmio_start,
-		     dev_priv->mmio_size, DRM_MTRR_WC);
+	arch_phys_wc_del(dev_priv->mmio_mtrr);
 	if (dev_priv->has_gmr)
 		(void) ttm_bo_clean_mm(&dev_priv->bdev, VMW_PL_GMR);
 	(void)ttm_bo_clean_mm(&dev_priv->bdev, TTM_PL_VRAM);
@@ -709,8 +708,7 @@ static int vmw_driver_unload(struct drm_device *dev)
 
 	ttm_object_device_release(&dev_priv->tdev);
 	iounmap(dev_priv->mmio_virt);
-	drm_mtrr_del(dev_priv->mmio_mtrr, dev_priv->mmio_start,
-		     dev_priv->mmio_size, DRM_MTRR_WC);
+	arch_phys_wc_del(dev_priv->mmio_mtrr);
 	if (dev_priv->has_gmr)
 		(void)ttm_bo_clean_mm(&dev_priv->bdev, VMW_PL_GMR);
 	(void)ttm_bo_clean_mm(&dev_priv->bdev, TTM_PL_VRAM);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
index 394e6476105b..599f6469a1eb 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
@@ -1432,6 +1432,7 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 	struct vmw_fence_obj *fence = NULL;
 	struct vmw_resource *error_resource;
 	struct list_head resource_list;
+	struct ww_acquire_ctx ticket;
 	uint32_t handle;
 	void *cmd;
 	int ret;
@@ -1488,7 +1489,7 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 	if (unlikely(ret != 0))
 		goto out_err;
 
-	ret = ttm_eu_reserve_buffers(&sw_context->validate_nodes);
+	ret = ttm_eu_reserve_buffers(&ticket, &sw_context->validate_nodes);
 	if (unlikely(ret != 0))
 		goto out_err;
 
@@ -1537,7 +1538,7 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 		DRM_ERROR("Fence submission error. Syncing.\n");
 
 	vmw_resource_list_unreserve(&sw_context->resource_list, false);
-	ttm_eu_fence_buffer_objects(&sw_context->validate_nodes,
+	ttm_eu_fence_buffer_objects(&ticket, &sw_context->validate_nodes,
 				    (void *) fence);
 
 	if (unlikely(dev_priv->pinned_bo != NULL &&
@@ -1570,7 +1571,7 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 out_err:
 	vmw_resource_relocations_free(&sw_context->res_relocations);
 	vmw_free_relocations(sw_context);
-	ttm_eu_backoff_reservation(&sw_context->validate_nodes);
+	ttm_eu_backoff_reservation(&ticket, &sw_context->validate_nodes);
 	vmw_resource_list_unreserve(&sw_context->resource_list, true);
 	vmw_clear_validations(sw_context);
 	if (unlikely(dev_priv->pinned_bo != NULL &&
@@ -1644,6 +1645,7 @@ void __vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
 	struct list_head validate_list;
 	struct ttm_validate_buffer pinned_val, query_val;
 	struct vmw_fence_obj *lfence = NULL;
+	struct ww_acquire_ctx ticket;
 
 	if (dev_priv->pinned_bo == NULL)
 		goto out_unlock;
@@ -1657,7 +1659,7 @@ void __vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
 	list_add_tail(&query_val.head, &validate_list);
 
 	do {
-		ret = ttm_eu_reserve_buffers(&validate_list);
+		ret = ttm_eu_reserve_buffers(&ticket, &validate_list);
 	} while (ret == -ERESTARTSYS);
 
 	if (unlikely(ret != 0)) {
@@ -1684,7 +1686,7 @@ void __vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
 						  NULL);
 		fence = lfence;
 	}
-	ttm_eu_fence_buffer_objects(&validate_list, (void *) fence);
+	ttm_eu_fence_buffer_objects(&ticket, &validate_list, (void *) fence);
 	if (lfence != NULL)
 		vmw_fence_obj_unreference(&lfence);
 
@@ -1696,7 +1698,7 @@ out_unlock:
 	return;
 
 out_no_emit:
-	ttm_eu_backoff_reservation(&validate_list);
+	ttm_eu_backoff_reservation(&ticket, &validate_list);
 out_no_reserve:
 	ttm_bo_unref(&query_val.bo);
 	ttm_bo_unref(&pinned_val.bo);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index 3e3c7ab33ca2..d4607b2530d6 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -174,7 +174,6 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 			   uint32_t handle, uint32_t width, uint32_t height)
 {
 	struct vmw_private *dev_priv = vmw_priv(crtc->dev);
-	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
 	struct vmw_display_unit *du = vmw_crtc_to_du(crtc);
 	struct vmw_surface *surface = NULL;
 	struct vmw_dma_buffer *dmabuf = NULL;
@@ -197,6 +196,8 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 	}
 
 	if (handle) {
+		struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
+
 		ret = vmw_user_lookup_handle(dev_priv, tfile,
 					     handle, &surface, &dmabuf);
 		if (ret) {
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
index bc784254e78e..7953d1f90b63 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
@@ -958,13 +958,13 @@ void vmw_resource_unreserve(struct vmw_resource *res,
 	if (new_backup && new_backup != res->backup) {
 
 		if (res->backup) {
-			BUG_ON(!ttm_bo_is_reserved(&res->backup->base));
+			lockdep_assert_held(&res->backup->base.resv->lock.base);
 			list_del_init(&res->mob_head);
 			vmw_dmabuf_unreference(&res->backup);
 		}
 
 		res->backup = vmw_dmabuf_reference(new_backup);
-		BUG_ON(!ttm_bo_is_reserved(&new_backup->base));
+		lockdep_assert_held(&new_backup->base.resv->lock.base);
 		list_add_tail(&res->mob_head, &new_backup->res_list);
 	}
 	if (new_backup)
@@ -990,9 +990,11 @@ void vmw_resource_unreserve(struct vmw_resource *res,
  * @val_buf:        On successful return contains data about the
  *                  reserved and validated backup buffer.
  */
-int vmw_resource_check_buffer(struct vmw_resource *res,
-			      bool interruptible,
-			      struct ttm_validate_buffer *val_buf)
+static int
+vmw_resource_check_buffer(struct vmw_resource *res,
+			  struct ww_acquire_ctx *ticket,
+			  bool interruptible,
+			  struct ttm_validate_buffer *val_buf)
 {
 	struct list_head val_list;
 	bool backup_dirty = false;
@@ -1007,7 +1009,7 @@ int vmw_resource_check_buffer(struct vmw_resource *res,
 	INIT_LIST_HEAD(&val_list);
 	val_buf->bo = ttm_bo_reference(&res->backup->base);
 	list_add_tail(&val_buf->head, &val_list);
-	ret = ttm_eu_reserve_buffers(&val_list);
+	ret = ttm_eu_reserve_buffers(ticket, &val_list);
 	if (unlikely(ret != 0))
 		goto out_no_reserve;
 
@@ -1025,7 +1027,7 @@ int vmw_resource_check_buffer(struct vmw_resource *res,
 	return 0;
 
 out_no_validate:
-	ttm_eu_backoff_reservation(&val_list);
+	ttm_eu_backoff_reservation(ticket, &val_list);
 out_no_reserve:
 	ttm_bo_unref(&val_buf->bo);
 	if (backup_dirty)
@@ -1069,7 +1071,9 @@ int vmw_resource_reserve(struct vmw_resource *res, bool no_backup)
  *.
  * @val_buf:        Backup buffer information.
  */
-void vmw_resource_backoff_reservation(struct ttm_validate_buffer *val_buf)
+static void
+vmw_resource_backoff_reservation(struct ww_acquire_ctx *ticket,
+				 struct ttm_validate_buffer *val_buf)
 {
 	struct list_head val_list;
 
@@ -1078,7 +1082,7 @@ void vmw_resource_backoff_reservation(struct ttm_validate_buffer *val_buf)
 
 	INIT_LIST_HEAD(&val_list);
 	list_add_tail(&val_buf->head, &val_list);
-	ttm_eu_backoff_reservation(&val_list);
+	ttm_eu_backoff_reservation(ticket, &val_list);
 	ttm_bo_unref(&val_buf->bo);
 }
 
@@ -1092,12 +1096,13 @@ int vmw_resource_do_evict(struct vmw_resource *res)
 {
 	struct ttm_validate_buffer val_buf;
 	const struct vmw_res_func *func = res->func;
+	struct ww_acquire_ctx ticket;
 	int ret;
 
 	BUG_ON(!func->may_evict);
 
 	val_buf.bo = NULL;
-	ret = vmw_resource_check_buffer(res, true, &val_buf);
+	ret = vmw_resource_check_buffer(res, &ticket, true, &val_buf);
 	if (unlikely(ret != 0))
 		return ret;
 
@@ -1112,7 +1117,7 @@ int vmw_resource_do_evict(struct vmw_resource *res)
 	res->backup_dirty = true;
 	res->res_dirty = false;
 out_no_unbind:
-	vmw_resource_backoff_reservation(&val_buf);
+	vmw_resource_backoff_reservation(&ticket, &val_buf);
 
 	return ret;
 }
diff --git a/drivers/gpu/host1x/dev.h b/drivers/gpu/host1x/dev.h
index a1607d6e135b..790ddf114e58 100644
--- a/drivers/gpu/host1x/dev.h
+++ b/drivers/gpu/host1x/dev.h
@@ -73,7 +73,7 @@ struct host1x_syncpt_ops {
 	void (*restore_wait_base)(struct host1x_syncpt *syncpt);
 	void (*load_wait_base)(struct host1x_syncpt *syncpt);
 	u32 (*load)(struct host1x_syncpt *syncpt);
-	void (*cpu_incr)(struct host1x_syncpt *syncpt);
+	int (*cpu_incr)(struct host1x_syncpt *syncpt);
 	int (*patch_wait)(struct host1x_syncpt *syncpt, void *patch_addr);
 };
 
@@ -157,10 +157,10 @@ static inline u32 host1x_hw_syncpt_load(struct host1x *host,
 	return host->syncpt_op->load(sp);
 }
 
-static inline void host1x_hw_syncpt_cpu_incr(struct host1x *host,
-					     struct host1x_syncpt *sp)
+static inline int host1x_hw_syncpt_cpu_incr(struct host1x *host,
+					    struct host1x_syncpt *sp)
 {
-	host->syncpt_op->cpu_incr(sp);
+	return host->syncpt_op->cpu_incr(sp);
 }
 
 static inline int host1x_hw_syncpt_patch_wait(struct host1x *host,
diff --git a/drivers/gpu/host1x/drm/dc.c b/drivers/gpu/host1x/drm/dc.c
index 8c04943f82e3..5360e5a57ecc 100644
--- a/drivers/gpu/host1x/drm/dc.c
+++ b/drivers/gpu/host1x/drm/dc.c
@@ -79,6 +79,9 @@ static int tegra_plane_disable(struct drm_plane *plane)
 	struct tegra_plane *p = to_tegra_plane(plane);
 	unsigned long value;
 
+	if (!plane->crtc)
+		return 0;
+
 	value = WINDOW_A_SELECT << p->index;
 	tegra_dc_writel(dc, value, DC_CMD_DISPLAY_WINDOW_HEADER);
 
@@ -140,6 +143,7 @@ static int tegra_dc_add_planes(struct drm_device *drm, struct tegra_dc *dc)
 static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 			     struct drm_framebuffer *fb)
 {
+	unsigned int format = tegra_dc_format(fb->pixel_format);
 	struct tegra_bo *bo = tegra_fb_get_plane(fb, 0);
 	unsigned long value;
 
@@ -150,6 +154,7 @@ static int tegra_dc_set_base(struct tegra_dc *dc, int x, int y,
 
 	tegra_dc_writel(dc, bo->paddr + value, DC_WINBUF_START_ADDR);
 	tegra_dc_writel(dc, fb->pitches[0], DC_WIN_LINE_STRIDE);
+	tegra_dc_writel(dc, format, DC_WIN_COLOR_DEPTH);
 
 	value = GENERAL_UPDATE | WIN_A_UPDATE;
 	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
diff --git a/drivers/gpu/host1x/drm/drm.c b/drivers/gpu/host1x/drm/drm.c
index 2b561c9118c6..e184b00faacd 100644
--- a/drivers/gpu/host1x/drm/drm.c
+++ b/drivers/gpu/host1x/drm/drm.c
@@ -148,6 +148,7 @@ int host1x_drm_init(struct host1x_drm *host1x, struct drm_device *drm)
 				dev_err(host1x->dev,
 					"DRM setup failed for %s: %d\n",
 					dev_name(client->dev), err);
+				mutex_unlock(&host1x->clients_lock);
 				return err;
 			}
 		}
@@ -175,6 +176,7 @@ int host1x_drm_exit(struct host1x_drm *host1x)
 				dev_err(host1x->dev,
 					"DRM cleanup failed for %s: %d\n",
 					dev_name(client->dev), err);
+				mutex_unlock(&host1x->clients_lock);
 				return err;
 			}
 		}
@@ -257,6 +259,13 @@ static int tegra_drm_load(struct drm_device *drm, unsigned long flags)
 	if (err < 0)
 		return err;
 
+	/*
+	 * We don't use the drm_irq_install() helpers provided by the DRM
+	 * core, so we need to set this manually in order to allow the
+	 * DRM_IOCTL_WAIT_VBLANK to operate correctly.
+	 */
+	drm->irq_enabled = 1;
+
 	err = drm_vblank_init(drm, drm->mode_config.num_crtc);
 	if (err < 0)
 		return err;
@@ -378,8 +387,7 @@ static int tegra_syncpt_incr(struct drm_device *drm, void *data,
 	if (!sp)
 		return -EINVAL;
 
-	host1x_syncpt_incr(sp);
-	return 0;
+	return host1x_syncpt_incr(sp);
 }
 
 static int tegra_syncpt_wait(struct drm_device *drm, void *data,
@@ -605,7 +613,7 @@ static void tegra_debugfs_cleanup(struct drm_minor *minor)
 #endif
 
 struct drm_driver tegra_drm_driver = {
-	.driver_features = DRIVER_BUS_PLATFORM | DRIVER_MODESET | DRIVER_GEM,
+	.driver_features = DRIVER_MODESET | DRIVER_GEM,
 	.load = tegra_drm_load,
 	.unload = tegra_drm_unload,
 	.open = tegra_drm_open,
diff --git a/drivers/gpu/host1x/drm/gr2d.c b/drivers/gpu/host1x/drm/gr2d.c
index 6a45ae090ee7..27ffcf15a4b4 100644
--- a/drivers/gpu/host1x/drm/gr2d.c
+++ b/drivers/gpu/host1x/drm/gr2d.c
@@ -84,7 +84,7 @@ static struct host1x_bo *host1x_bo_lookup(struct drm_device *drm,
 
 	gem = drm_gem_object_lookup(drm, file, handle);
 	if (!gem)
-		return 0;
+		return NULL;
 
 	mutex_lock(&drm->struct_mutex);
 	drm_gem_object_unreference(gem);
@@ -135,8 +135,10 @@ static int gr2d_submit(struct host1x_drm_context *context,
 			goto fail;
 
 		bo = host1x_bo_lookup(drm, file, cmdbuf.handle);
-		if (!bo)
+		if (!bo) {
+			err = -ENOENT;
 			goto fail;
+		}
 
 		host1x_job_add_gather(job, bo, cmdbuf.words, cmdbuf.offset);
 		num_cmdbufs--;
@@ -158,8 +160,10 @@ static int gr2d_submit(struct host1x_drm_context *context,
 		reloc->cmdbuf = cmdbuf;
 		reloc->target = target;
 
-		if (!reloc->target || !reloc->cmdbuf)
+		if (!reloc->target || !reloc->cmdbuf) {
+			err = -ENOENT;
 			goto fail;
+		}
 	}
 
 	err = copy_from_user(job->waitchk, waitchks,
@@ -281,7 +285,7 @@ static int gr2d_probe(struct platform_device *pdev)
 	if (!gr2d->channel)
 		return -ENOMEM;
 
-	*syncpts = host1x_syncpt_request(dev, 0);
+	*syncpts = host1x_syncpt_request(dev, false);
 	if (!(*syncpts)) {
 		host1x_channel_free(gr2d->channel);
 		return -ENOMEM;
diff --git a/drivers/gpu/host1x/hw/cdma_hw.c b/drivers/gpu/host1x/hw/cdma_hw.c
index 590b69d91dab..2ee4ad55c4db 100644
--- a/drivers/gpu/host1x/hw/cdma_hw.c
+++ b/drivers/gpu/host1x/hw/cdma_hw.c
@@ -44,7 +44,7 @@ static void cdma_timeout_cpu_incr(struct host1x_cdma *cdma, u32 getptr,
 	u32 i;
 
 	for (i = 0; i < syncpt_incrs; i++)
-		host1x_syncpt_cpu_incr(cdma->timeout.syncpt);
+		host1x_syncpt_incr(cdma->timeout.syncpt);
 
 	/* after CPU incr, ensure shadow is up to date */
 	host1x_syncpt_load(cdma->timeout.syncpt);
diff --git a/drivers/gpu/host1x/hw/syncpt_hw.c b/drivers/gpu/host1x/hw/syncpt_hw.c
index 61174990102a..0cf6095d3367 100644
--- a/drivers/gpu/host1x/hw/syncpt_hw.c
+++ b/drivers/gpu/host1x/hw/syncpt_hw.c
@@ -77,21 +77,19 @@ static u32 syncpt_load(struct host1x_syncpt *sp)
  * Write a cpu syncpoint increment to the hardware, without touching
  * the cache.
  */
-static void syncpt_cpu_incr(struct host1x_syncpt *sp)
+static int syncpt_cpu_incr(struct host1x_syncpt *sp)
 {
 	struct host1x *host = sp->host;
 	u32 reg_offset = sp->id / 32;
 
 	if (!host1x_syncpt_client_managed(sp) &&
-	    host1x_syncpt_idle(sp)) {
-		dev_err(host->dev, "Trying to increment syncpoint id %d beyond max\n",
-			sp->id);
-		host1x_debug_dump(sp->host);
-		return;
-	}
+	    host1x_syncpt_idle(sp))
+		return -EINVAL;
 	host1x_sync_writel(host, BIT_MASK(sp->id),
 			   HOST1X_SYNC_SYNCPT_CPU_INCR(reg_offset));
 	wmb();
+
+	return 0;
 }
 
 /* remove a wait pointed to by patch_addr */
diff --git a/drivers/gpu/host1x/job.c b/drivers/gpu/host1x/job.c
index f665d679031c..cc807667d8f1 100644
--- a/drivers/gpu/host1x/job.c
+++ b/drivers/gpu/host1x/job.c
@@ -228,17 +228,15 @@ static unsigned int do_relocs(struct host1x_job *job, struct host1x_bo *cmdbuf)
 	void *cmdbuf_page_addr = NULL;
 
 	/* pin & patch the relocs for one gather */
-	while (i < job->num_relocs) {
+	for (i = 0; i < job->num_relocs; i++) {
 		struct host1x_reloc *reloc = &job->relocarray[i];
 		u32 reloc_addr = (job->reloc_addr_phys[i] +
 			reloc->target_offset) >> reloc->shift;
 		u32 *target;
 
 		/* skip all other gathers */
-		if (!(reloc->cmdbuf && cmdbuf == reloc->cmdbuf)) {
-			i++;
+		if (cmdbuf != reloc->cmdbuf)
 			continue;
-		}
 
 		if (last_page != reloc->cmdbuf_offset >> PAGE_SHIFT) {
 			if (cmdbuf_page_addr)
@@ -257,9 +255,6 @@ static unsigned int do_relocs(struct host1x_job *job, struct host1x_bo *cmdbuf)
 
 		target = cmdbuf_page_addr + (reloc->cmdbuf_offset & ~PAGE_MASK);
 		*target = reloc_addr;
-
-		/* mark this gather as handled */
-		reloc->cmdbuf = 0;
 	}
 
 	if (cmdbuf_page_addr)
@@ -268,15 +263,15 @@ static unsigned int do_relocs(struct host1x_job *job, struct host1x_bo *cmdbuf)
 	return 0;
 }
 
-static int check_reloc(struct host1x_reloc *reloc, struct host1x_bo *cmdbuf,
+static bool check_reloc(struct host1x_reloc *reloc, struct host1x_bo *cmdbuf,
 		       unsigned int offset)
 {
 	offset *= sizeof(u32);
 
 	if (reloc->cmdbuf != cmdbuf || reloc->cmdbuf_offset != offset)
-		return -EINVAL;
+		return false;
 
-	return 0;
+	return true;
 }
 
 struct host1x_firewall {
@@ -307,10 +302,10 @@ static int check_mask(struct host1x_firewall *fw)
 
 		if (mask & 1) {
 			if (fw->job->is_addr_reg(fw->dev, fw->class, reg)) {
-				bool bad_reloc = check_reloc(fw->reloc,
-							     fw->cmdbuf_id,
-							     fw->offset);
-				if (!fw->num_relocs || bad_reloc)
+				if (!fw->num_relocs)
+					return -EINVAL;
+				if (!check_reloc(fw->reloc, fw->cmdbuf_id,
+						 fw->offset))
 					return -EINVAL;
 				fw->reloc++;
 				fw->num_relocs--;
@@ -330,14 +325,14 @@ static int check_incr(struct host1x_firewall *fw)
 	u32 count = fw->count;
 	u32 reg = fw->reg;
 
-	while (fw) {
+	while (count) {
 		if (fw->words == 0)
 			return -EINVAL;
 
 		if (fw->job->is_addr_reg(fw->dev, fw->class, reg)) {
-			bool bad_reloc = check_reloc(fw->reloc, fw->cmdbuf_id,
-						     fw->offset);
-			if (!fw->num_relocs || bad_reloc)
+			if (!fw->num_relocs)
+				return -EINVAL;
+			if (!check_reloc(fw->reloc, fw->cmdbuf_id, fw->offset))
 				return -EINVAL;
 			fw->reloc++;
 			fw->num_relocs--;
@@ -361,9 +356,9 @@ static int check_nonincr(struct host1x_firewall *fw)
 			return -EINVAL;
 
 		if (is_addr_reg) {
-			bool bad_reloc = check_reloc(fw->reloc, fw->cmdbuf_id,
-						     fw->offset);
-			if (!fw->num_relocs || bad_reloc)
+			if (!fw->num_relocs)
+				return -EINVAL;
+			if (!check_reloc(fw->reloc, fw->cmdbuf_id, fw->offset))
 				return -EINVAL;
 			fw->reloc++;
 			fw->num_relocs--;
@@ -376,69 +371,58 @@ static int check_nonincr(struct host1x_firewall *fw)
 	return 0;
 }
 
-static int validate(struct host1x_job *job, struct device *dev,
-		    struct host1x_job_gather *g)
+static int validate(struct host1x_firewall *fw, struct host1x_job_gather *g)
 {
-	u32 *cmdbuf_base;
+	u32 *cmdbuf_base = (u32 *)fw->job->gather_copy_mapped +
+		(g->offset / sizeof(u32));
 	int err = 0;
-	struct host1x_firewall fw;
 
-	fw.job = job;
-	fw.dev = dev;
-	fw.reloc = job->relocarray;
-	fw.num_relocs = job->num_relocs;
-	fw.cmdbuf_id = g->bo;
-
-	fw.offset = 0;
-	fw.class = 0;
-
-	if (!job->is_addr_reg)
+	if (!fw->job->is_addr_reg)
 		return 0;
 
-	cmdbuf_base = host1x_bo_mmap(g->bo);
-	if (!cmdbuf_base)
-		return -ENOMEM;
+	fw->words = g->words;
+	fw->cmdbuf_id = g->bo;
+	fw->offset = 0;
 
-	fw.words = g->words;
-	while (fw.words && !err) {
-		u32 word = cmdbuf_base[fw.offset];
+	while (fw->words && !err) {
+		u32 word = cmdbuf_base[fw->offset];
 		u32 opcode = (word & 0xf0000000) >> 28;
 
-		fw.mask = 0;
-		fw.reg = 0;
-		fw.count = 0;
-		fw.words--;
-		fw.offset++;
+		fw->mask = 0;
+		fw->reg = 0;
+		fw->count = 0;
+		fw->words--;
+		fw->offset++;
 
 		switch (opcode) {
 		case 0:
-			fw.class = word >> 6 & 0x3ff;
-			fw.mask = word & 0x3f;
-			fw.reg = word >> 16 & 0xfff;
-			err = check_mask(&fw);
+			fw->class = word >> 6 & 0x3ff;
+			fw->mask = word & 0x3f;
+			fw->reg = word >> 16 & 0xfff;
+			err = check_mask(fw);
 			if (err)
 				goto out;
 			break;
 		case 1:
-			fw.reg = word >> 16 & 0xfff;
-			fw.count = word & 0xffff;
-			err = check_incr(&fw);
+			fw->reg = word >> 16 & 0xfff;
+			fw->count = word & 0xffff;
+			err = check_incr(fw);
 			if (err)
 				goto out;
 			break;
 
 		case 2:
-			fw.reg = word >> 16 & 0xfff;
-			fw.count = word & 0xffff;
-			err = check_nonincr(&fw);
+			fw->reg = word >> 16 & 0xfff;
+			fw->count = word & 0xffff;
+			err = check_nonincr(fw);
 			if (err)
 				goto out;
 			break;
 
 		case 3:
-			fw.mask = word & 0xffff;
-			fw.reg = word >> 16 & 0xfff;
-			err = check_mask(&fw);
+			fw->mask = word & 0xffff;
+			fw->reg = word >> 16 & 0xfff;
+			err = check_mask(fw);
 			if (err)
 				goto out;
 			break;
@@ -453,21 +437,26 @@ static int validate(struct host1x_job *job, struct device *dev,
 	}
 
 	/* No relocs should remain at this point */
-	if (fw.num_relocs)
+	if (fw->num_relocs)
 		err = -EINVAL;
 
 out:
-	host1x_bo_munmap(g->bo, cmdbuf_base);
-
 	return err;
 }
 
 static inline int copy_gathers(struct host1x_job *job, struct device *dev)
 {
+	struct host1x_firewall fw;
 	size_t size = 0;
 	size_t offset = 0;
 	int i;
 
+	fw.job = job;
+	fw.dev = dev;
+	fw.reloc = job->relocarray;
+	fw.num_relocs = job->num_relocs;
+	fw.class = 0;
+
 	for (i = 0; i < job->num_gathers; i++) {
 		struct host1x_job_gather *g = &job->gathers[i];
 		size += g->words * sizeof(u32);
@@ -488,14 +477,19 @@ static inline int copy_gathers(struct host1x_job *job, struct device *dev)
 		struct host1x_job_gather *g = &job->gathers[i];
 		void *gather;
 
+		/* Copy the gather */
 		gather = host1x_bo_mmap(g->bo);
 		memcpy(job->gather_copy_mapped + offset, gather + g->offset,
 		       g->words * sizeof(u32));
 		host1x_bo_munmap(g->bo, gather);
 
+		/* Store the location in the buffer */
 		g->base = job->gather_copy;
 		g->offset = offset;
-		g->bo = NULL;
+
+		/* Validate the job */
+		if (validate(&fw, g))
+			return -EINVAL;
 
 		offset += g->words * sizeof(u32);
 	}
@@ -540,20 +534,11 @@ int host1x_job_pin(struct host1x_job *job, struct device *dev)
 			if (job->gathers[j].bo == g->bo)
 				job->gathers[j].handled = true;
 
-		err = 0;
-
-		if (IS_ENABLED(CONFIG_TEGRA_HOST1X_FIREWALL))
-			err = validate(job, dev, g);
-
+		err = do_relocs(job, g->bo);
 		if (err)
-			dev_err(dev, "Job invalid (err=%d)\n", err);
-
-		if (!err)
-			err = do_relocs(job, g->bo);
-
-		if (!err)
-			err = do_waitchks(job, host, g->bo);
+			break;
 
+		err = do_waitchks(job, host, g->bo);
 		if (err)
 			break;
 	}
diff --git a/drivers/gpu/host1x/syncpt.c b/drivers/gpu/host1x/syncpt.c
index 4b493453e805..409745b949db 100644
--- a/drivers/gpu/host1x/syncpt.c
+++ b/drivers/gpu/host1x/syncpt.c
@@ -32,7 +32,7 @@
 
 static struct host1x_syncpt *_host1x_syncpt_alloc(struct host1x *host,
 						  struct device *dev,
-						  int client_managed)
+						  bool client_managed)
 {
 	int i;
 	struct host1x_syncpt *sp = host->syncpt;
@@ -40,7 +40,8 @@ static struct host1x_syncpt *_host1x_syncpt_alloc(struct host1x *host,
 
 	for (i = 0; i < host->info->nb_pts && sp->name; i++, sp++)
 		;
-	if (sp->dev)
+
+	if (i >= host->info->nb_pts)
 		return NULL;
 
 	name = kasprintf(GFP_KERNEL, "%02d-%s", sp->id,
@@ -128,22 +129,11 @@ u32 host1x_syncpt_load_wait_base(struct host1x_syncpt *sp)
 }
 
 /*
- * Write a cpu syncpoint increment to the hardware, without touching
- * the cache. Caller is responsible for host being powered.
- */
-void host1x_syncpt_cpu_incr(struct host1x_syncpt *sp)
-{
-	host1x_hw_syncpt_cpu_incr(sp->host, sp);
-}
-
-/*
  * Increment syncpoint value from cpu, updating cache
  */
-void host1x_syncpt_incr(struct host1x_syncpt *sp)
+int host1x_syncpt_incr(struct host1x_syncpt *sp)
 {
-	if (host1x_syncpt_client_managed(sp))
-		host1x_syncpt_incr_max(sp, 1);
-	host1x_syncpt_cpu_incr(sp);
+	return host1x_hw_syncpt_cpu_incr(sp->host, sp);
 }
 
 /*
@@ -331,7 +321,7 @@ int host1x_syncpt_init(struct host1x *host)
 	host1x_syncpt_restore(host);
 
 	/* Allocate sync point to use for clearing waits for expired fences */
-	host->nop_sp = _host1x_syncpt_alloc(host, NULL, 0);
+	host->nop_sp = _host1x_syncpt_alloc(host, NULL, false);
 	if (!host->nop_sp)
 		return -ENOMEM;
 
@@ -339,7 +329,7 @@ int host1x_syncpt_init(struct host1x *host)
 }
 
 struct host1x_syncpt *host1x_syncpt_request(struct device *dev,
-					    int client_managed)
+					    bool client_managed)
 {
 	struct host1x *host = dev_get_drvdata(dev->parent);
 	return _host1x_syncpt_alloc(host, dev, client_managed);
@@ -353,7 +343,7 @@ void host1x_syncpt_free(struct host1x_syncpt *sp)
 	kfree(sp->name);
 	sp->dev = NULL;
 	sp->name = NULL;
-	sp->client_managed = 0;
+	sp->client_managed = false;
 }
 
 void host1x_syncpt_deinit(struct host1x *host)
diff --git a/drivers/gpu/host1x/syncpt.h b/drivers/gpu/host1x/syncpt.h
index c99806130f2e..267c0b9d3647 100644
--- a/drivers/gpu/host1x/syncpt.h
+++ b/drivers/gpu/host1x/syncpt.h
@@ -36,7 +36,7 @@ struct host1x_syncpt {
 	atomic_t max_val;
 	u32 base_val;
 	const char *name;
-	int client_managed;
+	bool client_managed;
 	struct host1x *host;
 	struct device *dev;
 
@@ -94,7 +94,7 @@ static inline bool host1x_syncpt_check_max(struct host1x_syncpt *sp, u32 real)
 }
 
 /* Return true if sync point is client managed. */
-static inline int host1x_syncpt_client_managed(struct host1x_syncpt *sp)
+static inline bool host1x_syncpt_client_managed(struct host1x_syncpt *sp)
 {
 	return sp->client_managed;
 }
@@ -115,9 +115,6 @@ static inline bool host1x_syncpt_idle(struct host1x_syncpt *sp)
 /* Return pointer to struct denoting sync point id. */
 struct host1x_syncpt *host1x_syncpt_get(struct host1x *host, u32 id);
 
-/* Request incrementing a sync point. */
-void host1x_syncpt_cpu_incr(struct host1x_syncpt *sp);
-
 /* Load current value from hardware to the shadow register. */
 u32 host1x_syncpt_load(struct host1x_syncpt *sp);
 
@@ -133,8 +130,8 @@ void host1x_syncpt_restore(struct host1x *host);
 /* Read current wait base value into shadow register and return it. */
 u32 host1x_syncpt_load_wait_base(struct host1x_syncpt *sp);
 
-/* Increment sync point and its max. */
-void host1x_syncpt_incr(struct host1x_syncpt *sp);
+/* Request incrementing a sync point. */
+int host1x_syncpt_incr(struct host1x_syncpt *sp);
 
 /* Indicate future operations by incrementing the sync point max. */
 u32 host1x_syncpt_incr_max(struct host1x_syncpt *sp, u32 incrs);
@@ -157,7 +154,7 @@ u32 host1x_syncpt_id(struct host1x_syncpt *sp);
 
 /* Allocate a sync point for a device. */
 struct host1x_syncpt *host1x_syncpt_request(struct device *dev,
-		int client_managed);
+					    bool client_managed);
 
 /* Free a sync point. */
 void host1x_syncpt_free(struct host1x_syncpt *sp);