Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "Three nouveau fixes, all user visible issues, and one radeon
  regression fix"

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  drm/radeon: enforce use of radeon_get_ib_value when reading user cmd
  drm/nouveau: add lockdep annotations
  drm/nv50/fb: Fix nullptr-deref on IGPs
  drm/nouveau: use different register to wait for secret scrubber

8 files changed, 81 insertions, 71 deletions

diff --git a/drivers/gpu/drm/nouveau/core/core/falcon.c b/drivers/gpu/drm/nouveau/core/core/falcon.c
index 6b0843c3387..e05c1577758 100644
--- a/drivers/gpu/drm/nouveau/core/core/falcon.c
+++ b/drivers/gpu/drm/nouveau/core/core/falcon.c
@@ -73,8 +73,11 @@ _nouveau_falcon_init(struct nouveau_object *object)
 	nv_debug(falcon, "data limit: %d\n", falcon->data.limit);
 
 	/* wait for 'uc halted' to be signalled before continuing */
-	if (falcon->secret) {
-		nv_wait(falcon, 0x008, 0x00000010, 0x00000010);
+	if (falcon->secret && falcon->version < 4) {
+		if (!falcon->version)
+			nv_wait(falcon, 0x008, 0x00000010, 0x00000010);
+		else
+			nv_wait(falcon, 0x180, 0x80000000, 0);
 		nv_wo32(falcon, 0x004, 0x00000010);
 	}
 
diff --git a/drivers/gpu/drm/nouveau/core/core/subdev.c b/drivers/gpu/drm/nouveau/core/core/subdev.c
index f74c30aa33a..48f06378d3f 100644
--- a/drivers/gpu/drm/nouveau/core/core/subdev.c
+++ b/drivers/gpu/drm/nouveau/core/core/subdev.c
@@ -99,7 +99,7 @@ nouveau_subdev_create_(struct nouveau_object *parent,
 	if (ret)
 		return ret;
 
-	mutex_init(&subdev->mutex);
+	__mutex_init(&subdev->mutex, subname, &oclass->lock_class_key);
 	subdev->name = subname;
 
 	if (parent) {
diff --git a/drivers/gpu/drm/nouveau/core/include/core/object.h b/drivers/gpu/drm/nouveau/core/include/core/object.h
index 5982935ee23..106bb19fdd9 100644
--- a/drivers/gpu/drm/nouveau/core/include/core/object.h
+++ b/drivers/gpu/drm/nouveau/core/include/core/object.h
@@ -50,10 +50,13 @@ int  nouveau_object_fini(struct nouveau_object *, bool suspend);
 
 extern struct nouveau_ofuncs nouveau_object_ofuncs;
 
+/* Don't allocate dynamically, because lockdep needs lock_class_keys to be in
+ * ".data". */
 struct nouveau_oclass {
 	u32 handle;
-	struct nouveau_ofuncs *ofuncs;
-	struct nouveau_omthds *omthds;
+	struct nouveau_ofuncs * const ofuncs;
+	struct nouveau_omthds * const omthds;
+	struct lock_class_key lock_class_key;
 };
 
 #define nv_oclass(o)    nv_object(o)->oclass
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/base.c b/drivers/gpu/drm/nouveau/core/subdev/fb/base.c
index d6d16007ec1..d62045f454b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/base.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/base.c
@@ -86,8 +86,8 @@ nouveau_fb_preinit(struct nouveau_fb *pfb)
 			return ret;
 	}
 
-	if (!nouveau_mm_initialised(&pfb->tags) && tags) {
-		ret = nouveau_mm_init(&pfb->tags, 0, ++tags, 1);
+	if (!nouveau_mm_initialised(&pfb->tags)) {
+		ret = nouveau_mm_init(&pfb->tags, 0, tags ? ++tags : 0, 1);
 		if (ret)
 			return ret;
 	}
diff --git a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
index 487cb8c6c20..eac236ed19b 100644
--- a/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
+++ b/drivers/gpu/drm/nouveau/core/subdev/fb/nv50.c
@@ -99,7 +99,7 @@ nv50_fb_vram_init(struct nouveau_fb *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;
+	u32 size, tags = 0;
 	int ret;
 
 	pfb->ram.size = nv_rd32(pfb, 0x10020c);
@@ -140,10 +140,11 @@ nv50_fb_vram_init(struct nouveau_fb *pfb)
 			return ret;
 
 		pfb->ram.ranks = (nv_rd32(pfb, 0x100200) & 0x4) ? 2 : 1;
+		tags = nv_rd32(pfb, 0x100320);
 		break;
 	}
 
-	return nv_rd32(pfb, 0x100320);
+	return tags;
 }
 
 static int
diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c
index 8b090f1eb51..5e7aef23825 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drm.c
@@ -245,6 +245,8 @@ 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)
 {
@@ -256,6 +258,7 @@ 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;
diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c
index 7a445666e71..ee4cff534f1 100644
--- a/drivers/gpu/drm/radeon/evergreen_cs.c
+++ b/drivers/gpu/drm/radeon/evergreen_cs.c
@@ -2909,14 +2909,14 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 				return -EINVAL;
 			}
 			if (tiled) {
-				dst_offset = ib[idx+1];
+				dst_offset = radeon_get_ib_value(p, idx+1);
 				dst_offset <<= 8;
 
 				ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
 				p->idx += count + 7;
 			} else {
-				dst_offset = ib[idx+1];
-				dst_offset |= ((u64)(ib[idx+2] & 0xff)) << 32;
+				dst_offset = radeon_get_ib_value(p, idx+1);
+				dst_offset |= ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32;
 
 				ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
 				ib[idx+2] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
@@ -2954,12 +2954,12 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 							DRM_ERROR("bad L2T, frame to fields DMA_PACKET_COPY\n");
 							return -EINVAL;
 						}
-						dst_offset = ib[idx+1];
+						dst_offset = radeon_get_ib_value(p, idx+1);
 						dst_offset <<= 8;
-						dst2_offset = ib[idx+2];
+						dst2_offset = radeon_get_ib_value(p, idx+2);
 						dst2_offset <<= 8;
-						src_offset = ib[idx+8];
-						src_offset |= ((u64)(ib[idx+9] & 0xff)) << 32;
+						src_offset = radeon_get_ib_value(p, idx+8);
+						src_offset |= ((u64)(radeon_get_ib_value(p, idx+9) & 0xff)) << 32;
 						if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
 							dev_warn(p->dev, "DMA L2T, frame to fields src buffer too small (%llu %lu)\n",
 								 src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
@@ -3014,12 +3014,12 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 							DRM_ERROR("bad L2T, broadcast DMA_PACKET_COPY\n");
 							return -EINVAL;
 						}
-						dst_offset = ib[idx+1];
+						dst_offset = radeon_get_ib_value(p, idx+1);
 						dst_offset <<= 8;
-						dst2_offset = ib[idx+2];
+						dst2_offset = radeon_get_ib_value(p, idx+2);
 						dst2_offset <<= 8;
-						src_offset = ib[idx+8];
-						src_offset |= ((u64)(ib[idx+9] & 0xff)) << 32;
+						src_offset = radeon_get_ib_value(p, idx+8);
+						src_offset |= ((u64)(radeon_get_ib_value(p, idx+9) & 0xff)) << 32;
 						if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
 							dev_warn(p->dev, "DMA L2T, broadcast src buffer too small (%llu %lu)\n",
 								 src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
@@ -3046,22 +3046,22 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 						/* detile bit */
 						if (idx_value & (1 << 31)) {
 							/* tiled src, linear dst */
-							src_offset = ib[idx+1];
+							src_offset = radeon_get_ib_value(p, idx+1);
 							src_offset <<= 8;
 							ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8);
 
-							dst_offset = ib[idx+7];
-							dst_offset |= ((u64)(ib[idx+8] & 0xff)) << 32;
+							dst_offset = radeon_get_ib_value(p, idx+7);
+							dst_offset |= ((u64)(radeon_get_ib_value(p, idx+8) & 0xff)) << 32;
 							ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
 							ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
 						} else {
 							/* linear src, tiled dst */
-							src_offset = ib[idx+7];
-							src_offset |= ((u64)(ib[idx+8] & 0xff)) << 32;
+							src_offset = radeon_get_ib_value(p, idx+7);
+							src_offset |= ((u64)(radeon_get_ib_value(p, idx+8) & 0xff)) << 32;
 							ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
 							ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
 
-							dst_offset = ib[idx+1];
+							dst_offset = radeon_get_ib_value(p, idx+1);
 							dst_offset <<= 8;
 							ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
 						}
@@ -3098,12 +3098,12 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 							DRM_ERROR("bad L2T, broadcast DMA_PACKET_COPY\n");
 							return -EINVAL;
 						}
-						dst_offset = ib[idx+1];
+						dst_offset = radeon_get_ib_value(p, idx+1);
 						dst_offset <<= 8;
-						dst2_offset = ib[idx+2];
+						dst2_offset = radeon_get_ib_value(p, idx+2);
 						dst2_offset <<= 8;
-						src_offset = ib[idx+8];
-						src_offset |= ((u64)(ib[idx+9] & 0xff)) << 32;
+						src_offset = radeon_get_ib_value(p, idx+8);
+						src_offset |= ((u64)(radeon_get_ib_value(p, idx+9) & 0xff)) << 32;
 						if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
 							dev_warn(p->dev, "DMA L2T, broadcast src buffer too small (%llu %lu)\n",
 								 src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
@@ -3135,22 +3135,22 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 						/* detile bit */
 						if (idx_value & (1 << 31)) {
 							/* tiled src, linear dst */
-							src_offset = ib[idx+1];
+							src_offset = radeon_get_ib_value(p, idx+1);
 							src_offset <<= 8;
 							ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8);
 
-							dst_offset = ib[idx+7];
-							dst_offset |= ((u64)(ib[idx+8] & 0xff)) << 32;
+							dst_offset = radeon_get_ib_value(p, idx+7);
+							dst_offset |= ((u64)(radeon_get_ib_value(p, idx+8) & 0xff)) << 32;
 							ib[idx+7] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
 							ib[idx+8] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
 						} else {
 							/* linear src, tiled dst */
-							src_offset = ib[idx+7];
-							src_offset |= ((u64)(ib[idx+8] & 0xff)) << 32;
+							src_offset = radeon_get_ib_value(p, idx+7);
+							src_offset |= ((u64)(radeon_get_ib_value(p, idx+8) & 0xff)) << 32;
 							ib[idx+7] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
 							ib[idx+8] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
 
-							dst_offset = ib[idx+1];
+							dst_offset = radeon_get_ib_value(p, idx+1);
 							dst_offset <<= 8;
 							ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
 						}
@@ -3176,10 +3176,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 					switch (misc) {
 					case 0:
 						/* L2L, byte */
-						src_offset = ib[idx+2];
-						src_offset |= ((u64)(ib[idx+4] & 0xff)) << 32;
-						dst_offset = ib[idx+1];
-						dst_offset |= ((u64)(ib[idx+3] & 0xff)) << 32;
+						src_offset = radeon_get_ib_value(p, idx+2);
+						src_offset |= ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32;
+						dst_offset = radeon_get_ib_value(p, idx+1);
+						dst_offset |= ((u64)(radeon_get_ib_value(p, idx+3) & 0xff)) << 32;
 						if ((src_offset + count) > radeon_bo_size(src_reloc->robj)) {
 							dev_warn(p->dev, "DMA L2L, byte src buffer too small (%llu %lu)\n",
 								 src_offset + count, radeon_bo_size(src_reloc->robj));
@@ -3216,12 +3216,12 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 							DRM_ERROR("bad L2L, dw, broadcast DMA_PACKET_COPY\n");
 							return -EINVAL;
 						}
-						dst_offset = ib[idx+1];
-						dst_offset |= ((u64)(ib[idx+4] & 0xff)) << 32;
-						dst2_offset = ib[idx+2];
-						dst2_offset |= ((u64)(ib[idx+5] & 0xff)) << 32;
-						src_offset = ib[idx+3];
-						src_offset |= ((u64)(ib[idx+6] & 0xff)) << 32;
+						dst_offset = radeon_get_ib_value(p, idx+1);
+						dst_offset |= ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32;
+						dst2_offset = radeon_get_ib_value(p, idx+2);
+						dst2_offset |= ((u64)(radeon_get_ib_value(p, idx+5) & 0xff)) << 32;
+						src_offset = radeon_get_ib_value(p, idx+3);
+						src_offset |= ((u64)(radeon_get_ib_value(p, idx+6) & 0xff)) << 32;
 						if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
 							dev_warn(p->dev, "DMA L2L, dw, broadcast src buffer too small (%llu %lu)\n",
 								 src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
@@ -3251,10 +3251,10 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 					}
 				} else {
 					/* L2L, dw */
-					src_offset = ib[idx+2];
-					src_offset |= ((u64)(ib[idx+4] & 0xff)) << 32;
-					dst_offset = ib[idx+1];
-					dst_offset |= ((u64)(ib[idx+3] & 0xff)) << 32;
+					src_offset = radeon_get_ib_value(p, idx+2);
+					src_offset |= ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32;
+					dst_offset = radeon_get_ib_value(p, idx+1);
+					dst_offset |= ((u64)(radeon_get_ib_value(p, idx+3) & 0xff)) << 32;
 					if ((src_offset + (count * 4)) > radeon_bo_size(src_reloc->robj)) {
 						dev_warn(p->dev, "DMA L2L, dw src buffer too small (%llu %lu)\n",
 							 src_offset + (count * 4), radeon_bo_size(src_reloc->robj));
@@ -3279,8 +3279,8 @@ int evergreen_dma_cs_parse(struct radeon_cs_parser *p)
 				DRM_ERROR("bad DMA_PACKET_CONSTANT_FILL\n");
 				return -EINVAL;
 			}
-			dst_offset = ib[idx+1];
-			dst_offset |= ((u64)(ib[idx+3] & 0x00ff0000)) << 16;
+			dst_offset = radeon_get_ib_value(p, idx+1);
+			dst_offset |= ((u64)(radeon_get_ib_value(p, idx+3) & 0x00ff0000)) << 16;
 			if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
 				dev_warn(p->dev, "DMA constant fill buffer too small (%llu %lu)\n",
 					 dst_offset, radeon_bo_size(dst_reloc->robj));
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index 69ec24ab8d6..9b2512bf1a4 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -2623,14 +2623,14 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p)
 				return -EINVAL;
 			}
 			if (tiled) {
-				dst_offset = ib[idx+1];
+				dst_offset = radeon_get_ib_value(p, idx+1);
 				dst_offset <<= 8;
 
 				ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
 				p->idx += count + 5;
 			} else {
-				dst_offset = ib[idx+1];
-				dst_offset |= ((u64)(ib[idx+2] & 0xff)) << 32;
+				dst_offset = radeon_get_ib_value(p, idx+1);
+				dst_offset |= ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32;
 
 				ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
 				ib[idx+2] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
@@ -2658,32 +2658,32 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p)
 				/* detile bit */
 				if (idx_value & (1 << 31)) {
 					/* tiled src, linear dst */
-					src_offset = ib[idx+1];
+					src_offset = radeon_get_ib_value(p, idx+1);
 					src_offset <<= 8;
 					ib[idx+1] += (u32)(src_reloc->lobj.gpu_offset >> 8);
 
-					dst_offset = ib[idx+5];
-					dst_offset |= ((u64)(ib[idx+6] & 0xff)) << 32;
+					dst_offset = radeon_get_ib_value(p, idx+5);
+					dst_offset |= ((u64)(radeon_get_ib_value(p, idx+6) & 0xff)) << 32;
 					ib[idx+5] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
 					ib[idx+6] += upper_32_bits(dst_reloc->lobj.gpu_offset) & 0xff;
 				} else {
 					/* linear src, tiled dst */
-					src_offset = ib[idx+5];
-					src_offset |= ((u64)(ib[idx+6] & 0xff)) << 32;
+					src_offset = radeon_get_ib_value(p, idx+5);
+					src_offset |= ((u64)(radeon_get_ib_value(p, idx+6) & 0xff)) << 32;
 					ib[idx+5] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
 					ib[idx+6] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
 
-					dst_offset = ib[idx+1];
+					dst_offset = radeon_get_ib_value(p, idx+1);
 					dst_offset <<= 8;
 					ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset >> 8);
 				}
 				p->idx += 7;
 			} else {
 				if (p->family >= CHIP_RV770) {
-					src_offset = ib[idx+2];
-					src_offset |= ((u64)(ib[idx+4] & 0xff)) << 32;
-					dst_offset = ib[idx+1];
-					dst_offset |= ((u64)(ib[idx+3] & 0xff)) << 32;
+					src_offset = radeon_get_ib_value(p, idx+2);
+					src_offset |= ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32;
+					dst_offset = radeon_get_ib_value(p, idx+1);
+					dst_offset |= ((u64)(radeon_get_ib_value(p, idx+3) & 0xff)) << 32;
 
 					ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
 					ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
@@ -2691,10 +2691,10 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p)
 					ib[idx+4] += upper_32_bits(src_reloc->lobj.gpu_offset) & 0xff;
 					p->idx += 5;
 				} else {
-					src_offset = ib[idx+2];
-					src_offset |= ((u64)(ib[idx+3] & 0xff)) << 32;
-					dst_offset = ib[idx+1];
-					dst_offset |= ((u64)(ib[idx+3] & 0xff0000)) << 16;
+					src_offset = radeon_get_ib_value(p, idx+2);
+					src_offset |= ((u64)(radeon_get_ib_value(p, idx+3) & 0xff)) << 32;
+					dst_offset = radeon_get_ib_value(p, idx+1);
+					dst_offset |= ((u64)(radeon_get_ib_value(p, idx+3) & 0xff0000)) << 16;
 
 					ib[idx+1] += (u32)(dst_reloc->lobj.gpu_offset & 0xfffffffc);
 					ib[idx+2] += (u32)(src_reloc->lobj.gpu_offset & 0xfffffffc);
@@ -2724,8 +2724,8 @@ int r600_dma_cs_parse(struct radeon_cs_parser *p)
 				DRM_ERROR("bad DMA_PACKET_WRITE\n");
 				return -EINVAL;
 			}
-			dst_offset = ib[idx+1];
-			dst_offset |= ((u64)(ib[idx+3] & 0x00ff0000)) << 16;
+			dst_offset = radeon_get_ib_value(p, idx+1);
+			dst_offset |= ((u64)(radeon_get_ib_value(p, idx+3) & 0x00ff0000)) << 16;
 			if ((dst_offset + (count * 4)) > radeon_bo_size(dst_reloc->robj)) {
 				dev_warn(p->dev, "DMA constant fill buffer too small (%llu %lu)\n",
 					 dst_offset + (count * 4), radeon_bo_size(dst_reloc->robj));