From 71a77e07d0e33b57d4a50c173e5ce4fabceddbec Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 2 Feb 2011 12:13:49 +0000 Subject: drm/i915: Invalidate TLB caches on SNB BLT/BSD rings Signed-off-by: Chris Wilson Cc: stable@kernel.org --- drivers/gpu/drm/i915/i915_reg.h | 4 +++- drivers/gpu/drm/i915/intel_ringbuffer.c | 26 ++++++++++++++++---------- 2 files changed, 19 insertions(+), 11 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 5cfc68940f17..15d94c63918c 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -174,7 +174,9 @@ * address/value pairs. Don't overdue it, though, x <= 2^4 must hold! */ #define MI_LOAD_REGISTER_IMM(x) MI_INSTR(0x22, 2*x-1) -#define MI_FLUSH_DW MI_INSTR(0x26, 2) /* for GEN6 */ +#define MI_FLUSH_DW MI_INSTR(0x26, 1) /* for GEN6 */ +#define MI_INVALIDATE_TLB (1<<18) +#define MI_INVALIDATE_BSD (1<<7) #define MI_BATCH_BUFFER MI_INSTR(0x30, 1) #define MI_BATCH_NON_SECURE (1) #define MI_BATCH_NON_SECURE_I965 (1<<8) diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 6218fa97aa1e..445f27efe677 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1059,22 +1059,25 @@ static void gen6_bsd_ring_write_tail(struct intel_ring_buffer *ring, } static int gen6_ring_flush(struct intel_ring_buffer *ring, - u32 invalidate_domains, - u32 flush_domains) + u32 invalidate, u32 flush) { + uint32_t cmd; int ret; - if ((flush_domains & I915_GEM_DOMAIN_RENDER) == 0) + if (((invalidate | flush) & I915_GEM_GPU_DOMAINS) == 0) return 0; ret = intel_ring_begin(ring, 4); if (ret) return ret; - intel_ring_emit(ring, MI_FLUSH_DW); - intel_ring_emit(ring, 0); + cmd = MI_FLUSH_DW; + if (invalidate & I915_GEM_GPU_DOMAINS) + cmd |= MI_INVALIDATE_TLB | MI_INVALIDATE_BSD; + intel_ring_emit(ring, cmd); intel_ring_emit(ring, 0); intel_ring_emit(ring, 0); + intel_ring_emit(ring, MI_NOOP); intel_ring_advance(ring); return 0; } @@ -1230,22 +1233,25 @@ static int blt_ring_begin(struct intel_ring_buffer *ring, } static int blt_ring_flush(struct intel_ring_buffer *ring, - u32 invalidate_domains, - u32 flush_domains) + u32 invalidate, u32 flush) { + uint32_t cmd; int ret; - if ((flush_domains & I915_GEM_DOMAIN_RENDER) == 0) + if (((invalidate | flush) & I915_GEM_DOMAIN_RENDER) == 0) return 0; ret = blt_ring_begin(ring, 4); if (ret) return ret; - intel_ring_emit(ring, MI_FLUSH_DW); - intel_ring_emit(ring, 0); + cmd = MI_FLUSH_DW; + if (invalidate & I915_GEM_DOMAIN_RENDER) + cmd |= MI_INVALIDATE_TLB; + intel_ring_emit(ring, cmd); intel_ring_emit(ring, 0); intel_ring_emit(ring, 0); + intel_ring_emit(ring, MI_NOOP); intel_ring_advance(ring); return 0; } -- cgit v1.2.3 From 72389a33b8878e6091f7ab8080f5ed07054c7c39 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Sun, 6 Feb 2011 15:50:52 +0000 Subject: drm/i915/lvds: Restore dithering on native modes for gen2/3 A regression introduced in bee17e5 cleared the dithering bit for native modes on gen2/3. Bugzilla: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/711568 Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/intel_lvds.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c index ace8d5d30dd2..bcdba7bd5cfa 100644 --- a/drivers/gpu/drm/i915/intel_lvds.c +++ b/drivers/gpu/drm/i915/intel_lvds.c @@ -261,12 +261,6 @@ static bool intel_lvds_mode_fixup(struct drm_encoder *encoder, return true; } - /* Make sure pre-965s set dither correctly */ - if (INTEL_INFO(dev)->gen < 4) { - if (dev_priv->lvds_dither) - pfit_control |= PANEL_8TO6_DITHER_ENABLE; - } - /* Native modes don't need fitting */ if (adjusted_mode->hdisplay == mode->hdisplay && adjusted_mode->vdisplay == mode->vdisplay) @@ -374,10 +368,16 @@ static bool intel_lvds_mode_fixup(struct drm_encoder *encoder, } 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; + if (pfit_control != intel_lvds->pfit_control || pfit_pgm_ratios != intel_lvds->pfit_pgm_ratios) { intel_lvds->pfit_control = pfit_control; -- cgit v1.2.3 From ac66808814036b4c33dd98091b2176ae6157f1a8 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 9 Feb 2011 16:15:32 +0000 Subject: drm/i915: Disable RC6 on Ironlake The automatic powersaving feature is once again causing havoc, with 100% reliable hangs on boot and resume on affected machines. Reported-by: Francesco Allertsen Reported-by: Gui Rui Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=28582 Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/i915_drv.c | 5 +- drivers/gpu/drm/i915/i915_drv.h | 1 + drivers/gpu/drm/i915/intel_display.c | 92 ++++++++++++++++++++---------------- drivers/gpu/drm/i915/intel_drv.h | 1 - 4 files changed, 55 insertions(+), 44 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index cfb56d0ff367..0ad533f06af9 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -46,6 +46,9 @@ module_param_named(fbpercrtc, i915_fbpercrtc, int, 0400); unsigned int i915_powersave = 1; module_param_named(powersave, i915_powersave, int, 0600); +unsigned int i915_enable_rc6 = 0; +module_param_named(i915_enable_rc6, i915_enable_rc6, int, 0600); + unsigned int i915_lvds_downclock = 0; module_param_named(lvds_downclock, i915_lvds_downclock, int, 0400); @@ -360,7 +363,7 @@ static int i915_drm_thaw(struct drm_device *dev) /* Resume the modeset for every activated CRTC */ drm_helper_resume_force_mode(dev); - if (dev_priv->renderctx && dev_priv->pwrctx) + if (IS_IRONLAKE_M(dev)) ironlake_enable_rc6(dev); } diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index a0149c619cdd..65dfe81d0035 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -958,6 +958,7 @@ extern unsigned int i915_fbpercrtc; extern unsigned int i915_powersave; extern unsigned int i915_lvds_downclock; extern unsigned int i915_panel_use_ssc; +extern unsigned int i915_enable_rc6; extern int i915_suspend(struct drm_device *dev, pm_message_t state); extern int i915_resume(struct drm_device *dev); diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 7e42aa586504..94622e3a202e 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -6463,52 +6463,60 @@ void intel_enable_clock_gating(struct drm_device *dev) } } -void intel_disable_clock_gating(struct drm_device *dev) +static void ironlake_teardown_rc6(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; if (dev_priv->renderctx) { - struct drm_i915_gem_object *obj = dev_priv->renderctx; - - I915_WRITE(CCID, 0); - POSTING_READ(CCID); - - i915_gem_object_unpin(obj); - drm_gem_object_unreference(&obj->base); + i915_gem_object_unpin(dev_priv->renderctx); + drm_gem_object_unreference(&dev_priv->renderctx->base); dev_priv->renderctx = NULL; } if (dev_priv->pwrctx) { - struct drm_i915_gem_object *obj = dev_priv->pwrctx; + i915_gem_object_unpin(dev_priv->pwrctx); + drm_gem_object_unreference(&dev_priv->pwrctx->base); + dev_priv->pwrctx = NULL; + } +} + +static void ironlake_disable_rc6(struct drm_device *dev) +{ + struct drm_i915_private *dev_priv = dev->dev_private; + + if (I915_READ(PWRCTXA)) { + /* Wake the GPU, prevent RC6, then restore RSTDBYCTL */ + I915_WRITE(RSTDBYCTL, I915_READ(RSTDBYCTL) | RCX_SW_EXIT); + wait_for(((I915_READ(RSTDBYCTL) & RSX_STATUS_MASK) == RSX_STATUS_ON), + 50); I915_WRITE(PWRCTXA, 0); POSTING_READ(PWRCTXA); - i915_gem_object_unpin(obj); - drm_gem_object_unreference(&obj->base); - dev_priv->pwrctx = NULL; + I915_WRITE(RSTDBYCTL, I915_READ(RSTDBYCTL) & ~RCX_SW_EXIT); + POSTING_READ(RSTDBYCTL); } + + ironlake_disable_rc6(dev); } -static void ironlake_disable_rc6(struct drm_device *dev) +static int ironlake_setup_rc6(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - /* Wake the GPU, prevent RC6, then restore RSTDBYCTL */ - I915_WRITE(RSTDBYCTL, I915_READ(RSTDBYCTL) | RCX_SW_EXIT); - wait_for(((I915_READ(RSTDBYCTL) & RSX_STATUS_MASK) == RSX_STATUS_ON), - 10); - POSTING_READ(CCID); - I915_WRITE(PWRCTXA, 0); - POSTING_READ(PWRCTXA); - I915_WRITE(RSTDBYCTL, I915_READ(RSTDBYCTL) & ~RCX_SW_EXIT); - POSTING_READ(RSTDBYCTL); - i915_gem_object_unpin(dev_priv->renderctx); - drm_gem_object_unreference(&dev_priv->renderctx->base); - dev_priv->renderctx = NULL; - i915_gem_object_unpin(dev_priv->pwrctx); - drm_gem_object_unreference(&dev_priv->pwrctx->base); - dev_priv->pwrctx = NULL; + if (dev_priv->renderctx == NULL) + dev_priv->renderctx = intel_alloc_context_page(dev); + if (!dev_priv->renderctx) + return -ENOMEM; + + if (dev_priv->pwrctx == NULL) + dev_priv->pwrctx = intel_alloc_context_page(dev); + if (!dev_priv->pwrctx) { + ironlake_teardown_rc6(dev); + return -ENOMEM; + } + + return 0; } void ironlake_enable_rc6(struct drm_device *dev) @@ -6516,15 +6524,26 @@ void ironlake_enable_rc6(struct drm_device *dev) struct drm_i915_private *dev_priv = dev->dev_private; int ret; + /* rc6 disabled by default due to repeated reports of hanging during + * boot and resume. + */ + if (!i915_enable_rc6) + return; + + ret = ironlake_setup_rc6(dev); + if (ret) + return; + /* * GPU can automatically power down the render unit if given a page * to save state. */ ret = BEGIN_LP_RING(6); if (ret) { - ironlake_disable_rc6(dev); + ironlake_teardown_rc6(dev); return; } + OUT_RING(MI_SUSPEND_FLUSH | MI_SUSPEND_FLUSH_EN); OUT_RING(MI_SET_CONTEXT); OUT_RING(dev_priv->renderctx->gtt_offset | @@ -6541,6 +6560,7 @@ void ironlake_enable_rc6(struct drm_device *dev) I915_WRITE(RSTDBYCTL, I915_READ(RSTDBYCTL) & ~RCX_SW_EXIT); } + /* Set up chip specific display functions */ static void intel_init_display(struct drm_device *dev) { @@ -6783,21 +6803,9 @@ void intel_modeset_init(struct drm_device *dev) if (IS_GEN6(dev)) gen6_enable_rps(dev_priv); - if (IS_IRONLAKE_M(dev)) { - dev_priv->renderctx = intel_alloc_context_page(dev); - if (!dev_priv->renderctx) - goto skip_rc6; - dev_priv->pwrctx = intel_alloc_context_page(dev); - if (!dev_priv->pwrctx) { - i915_gem_object_unpin(dev_priv->renderctx); - drm_gem_object_unreference(&dev_priv->renderctx->base); - dev_priv->renderctx = NULL; - goto skip_rc6; - } + if (IS_IRONLAKE_M(dev)) ironlake_enable_rc6(dev); - } -skip_rc6: INIT_WORK(&dev_priv->idle_work, intel_idle_update); setup_timer(&dev_priv->idle_timer, intel_gpu_idle_timer, (unsigned long)dev); diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index 74db2557d644..2c431049963c 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -298,7 +298,6 @@ extern void intel_crtc_fb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green, extern void intel_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green, u16 *blue, int regno); extern void intel_enable_clock_gating(struct drm_device *dev); -extern void intel_disable_clock_gating(struct drm_device *dev); extern void ironlake_enable_drps(struct drm_device *dev); extern void ironlake_disable_drps(struct drm_device *dev); extern void gen6_enable_rps(struct drm_i915_private *dev_priv); -- cgit v1.2.3 From 139467433e50926d22338e9dc754feaaf94b9db0 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 9 Feb 2011 20:01:16 +0000 Subject: drm/i915/sdvo: If we have an EDID confirm it matches the mode of the connection If we have an EDID for a digital panel, but we are probing a non-TMDS connector then we know that this is a false detection, and vice versa. This should reduce the number of bogus outputs on multi-function adapters that report the same output on multiple connectors. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=34101 Reported-by: Sebastien Caty Tested-by: Sebastien Caty Signed-off-by: Chris Wilson Cc: stable@kernel.org --- drivers/gpu/drm/i915/intel_sdvo.c | 32 +++++++++++++++++++++++++++----- 1 file changed, 27 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c index 6a09c1413d60..d2dd90a9a101 100644 --- a/drivers/gpu/drm/i915/intel_sdvo.c +++ b/drivers/gpu/drm/i915/intel_sdvo.c @@ -46,6 +46,7 @@ SDVO_TV_MASK) #define IS_TV(c) (c->output_flag & SDVO_TV_MASK) +#define IS_TMDS(c) (c->output_flag & SDVO_TMDS_MASK) #define IS_LVDS(c) (c->output_flag & SDVO_LVDS_MASK) #define IS_TV_OR_LVDS(c) (c->output_flag & (SDVO_TV_MASK | SDVO_LVDS_MASK)) @@ -1359,7 +1360,8 @@ intel_sdvo_hdmi_sink_detect(struct drm_connector *connector) intel_sdvo->has_hdmi_monitor = drm_detect_hdmi_monitor(edid); intel_sdvo->has_hdmi_audio = drm_detect_monitor_audio(edid); } - } + } else + status = connector_status_disconnected; connector->display_info.raw_edid = NULL; kfree(edid); } @@ -1407,10 +1409,25 @@ intel_sdvo_detect(struct drm_connector *connector, bool force) if ((intel_sdvo_connector->output_flag & response) == 0) ret = connector_status_disconnected; - else if (response & SDVO_TMDS_MASK) + else if (IS_TMDS(intel_sdvo_connector)) ret = intel_sdvo_hdmi_sink_detect(connector); - else - ret = connector_status_connected; + else { + struct edid *edid; + + /* if we have an edid check it matches the connection */ + edid = intel_sdvo_get_edid(connector); + if (edid == NULL) + edid = intel_sdvo_get_analog_edid(connector); + if (edid != NULL) { + if (edid->input & DRM_EDID_INPUT_DIGITAL) + ret = connector_status_disconnected; + else + ret = connector_status_connected; + connector->display_info.raw_edid = NULL; + kfree(edid); + } else + ret = connector_status_connected; + } /* May update encoder flag for like clock for SDVO TV, etc.*/ if (ret == connector_status_connected) { @@ -1446,10 +1463,15 @@ static void intel_sdvo_get_ddc_modes(struct drm_connector *connector) edid = intel_sdvo_get_analog_edid(connector); if (edid != NULL) { - if (edid->input & DRM_EDID_INPUT_DIGITAL) { + struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector); + bool monitor_is_digital = !!(edid->input & DRM_EDID_INPUT_DIGITAL); + bool connector_is_digital = !!IS_TMDS(intel_sdvo_connector); + + if (connector_is_digital == monitor_is_digital) { drm_mode_connector_update_edid_property(connector, edid); drm_add_edid_modes(connector, edid); } + connector->display_info.raw_edid = NULL; kfree(edid); } -- cgit v1.2.3 From 1aad7ac0458f40e2d0365d488620084f3965f6e7 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Wed, 9 Feb 2011 18:46:58 +0000 Subject: drm/i915: Trigger modesetting if force-audio changes If the user changes the force-audio property and it no longer reflects the current configuration, then we need to trigger a mode set in order to update the registers. Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/intel_dp.c | 36 ++++++++++++++++++++++++++++++------ drivers/gpu/drm/i915/intel_hdmi.c | 39 +++++++++++++++++++++++++++++++++------ drivers/gpu/drm/i915/intel_sdvo.c | 34 ++++++++++++++++++++++++++++------ 3 files changed, 91 insertions(+), 18 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 1f4242b682c8..51cb4e36997f 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -1639,6 +1639,24 @@ static int intel_dp_get_modes(struct drm_connector *connector) return 0; } +static bool +intel_dp_detect_audio(struct drm_connector *connector) +{ + struct intel_dp *intel_dp = intel_attached_dp(connector); + struct edid *edid; + bool has_audio = false; + + edid = drm_get_edid(connector, &intel_dp->adapter); + if (edid) { + has_audio = drm_detect_monitor_audio(edid); + + connector->display_info.raw_edid = NULL; + kfree(edid); + } + + return has_audio; +} + static int intel_dp_set_property(struct drm_connector *connector, struct drm_property *property, @@ -1652,17 +1670,23 @@ intel_dp_set_property(struct drm_connector *connector, return ret; if (property == intel_dp->force_audio_property) { - if (val == intel_dp->force_audio) + int i = val; + bool has_audio; + + if (i == intel_dp->force_audio) return 0; - intel_dp->force_audio = val; + intel_dp->force_audio = i; - if (val > 0 && intel_dp->has_audio) - return 0; - if (val < 0 && !intel_dp->has_audio) + if (i == 0) + has_audio = intel_dp_detect_audio(connector); + else + has_audio = i > 0; + + if (has_audio == intel_dp->has_audio) return 0; - intel_dp->has_audio = val > 0; + intel_dp->has_audio = has_audio; goto done; } diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c index 0d0273e7b029..c635c9e357b9 100644 --- a/drivers/gpu/drm/i915/intel_hdmi.c +++ b/drivers/gpu/drm/i915/intel_hdmi.c @@ -251,6 +251,27 @@ static int intel_hdmi_get_modes(struct drm_connector *connector) &dev_priv->gmbus[intel_hdmi->ddc_bus].adapter); } +static bool +intel_hdmi_detect_audio(struct drm_connector *connector) +{ + struct intel_hdmi *intel_hdmi = intel_attached_hdmi(connector); + struct drm_i915_private *dev_priv = connector->dev->dev_private; + struct edid *edid; + bool has_audio = false; + + edid = drm_get_edid(connector, + &dev_priv->gmbus[intel_hdmi->ddc_bus].adapter); + if (edid) { + if (edid->input & DRM_EDID_INPUT_DIGITAL) + has_audio = drm_detect_monitor_audio(edid); + + connector->display_info.raw_edid = NULL; + kfree(edid); + } + + return has_audio; +} + static int intel_hdmi_set_property(struct drm_connector *connector, struct drm_property *property, @@ -264,17 +285,23 @@ intel_hdmi_set_property(struct drm_connector *connector, return ret; if (property == intel_hdmi->force_audio_property) { - if (val == intel_hdmi->force_audio) + int i = val; + bool has_audio; + + if (i == intel_hdmi->force_audio) return 0; - intel_hdmi->force_audio = val; + intel_hdmi->force_audio = i; - if (val > 0 && intel_hdmi->has_audio) - return 0; - if (val < 0 && !intel_hdmi->has_audio) + if (i == 0) + has_audio = intel_hdmi_detect_audio(connector); + else + has_audio = i > 0; + + if (has_audio == intel_hdmi->has_audio) return 0; - intel_hdmi->has_audio = val > 0; + intel_hdmi->has_audio = has_audio; goto done; } diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c index d2dd90a9a101..7c50cdce84f0 100644 --- a/drivers/gpu/drm/i915/intel_sdvo.c +++ b/drivers/gpu/drm/i915/intel_sdvo.c @@ -1690,6 +1690,22 @@ static void intel_sdvo_destroy(struct drm_connector *connector) kfree(connector); } +static bool intel_sdvo_detect_hdmi_audio(struct drm_connector *connector) +{ + struct intel_sdvo *intel_sdvo = intel_attached_sdvo(connector); + struct edid *edid; + bool has_audio = false; + + if (!intel_sdvo->is_hdmi) + return false; + + edid = intel_sdvo_get_edid(connector); + if (edid != NULL && edid->input & DRM_EDID_INPUT_DIGITAL) + has_audio = drm_detect_monitor_audio(edid); + + return has_audio; +} + static int intel_sdvo_set_property(struct drm_connector *connector, struct drm_property *property, @@ -1706,17 +1722,23 @@ intel_sdvo_set_property(struct drm_connector *connector, return ret; if (property == intel_sdvo_connector->force_audio_property) { - if (val == intel_sdvo_connector->force_audio) + int i = val; + bool has_audio; + + if (i == intel_sdvo_connector->force_audio) return 0; - intel_sdvo_connector->force_audio = val; + intel_sdvo_connector->force_audio = i; - if (val > 0 && intel_sdvo->has_hdmi_audio) - return 0; - if (val < 0 && !intel_sdvo->has_hdmi_audio) + if (i == 0) + has_audio = intel_sdvo_detect_hdmi_audio(connector); + else + has_audio = i > 0; + + if (has_audio == intel_sdvo->has_hdmi_audio) return 0; - intel_sdvo->has_hdmi_audio = val > 0; + intel_sdvo->has_hdmi_audio = has_audio; goto done; } -- cgit v1.2.3 From 8102e126c0827b5336065fd86d3d313b60fde23a Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 10 Feb 2011 10:05:35 +0000 Subject: drm/i915/tv: Use polling rather than interrupt-based hotplug The documentation recommends that we should use a polling method for TV detection as this is more power efficient than the interrupt based mechanism (as the encoder can be completely switched off). A secondary effect is that leaving the hotplug enabled seems to be causing pipe underruns as reported by Hugh Dickins on his Crestline. Tested-by: Hugh Dickins Signed-off-by: Chris Wilson [This is a candidate for stable, but needs minor porting to 2.6.37] --- drivers/gpu/drm/i915/intel_tv.c | 43 ++++++++++++++++++++++++++++------------- 1 file changed, 30 insertions(+), 13 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c index 93206e4eaa6f..fe4a53a50b83 100644 --- a/drivers/gpu/drm/i915/intel_tv.c +++ b/drivers/gpu/drm/i915/intel_tv.c @@ -1234,7 +1234,8 @@ static const struct drm_display_mode reported_modes[] = { * \return false if TV is disconnected. */ static int -intel_tv_detect_type (struct intel_tv *intel_tv) +intel_tv_detect_type (struct intel_tv *intel_tv, + struct drm_connector *connector) { struct drm_encoder *encoder = &intel_tv->base.base; struct drm_device *dev = encoder->dev; @@ -1245,11 +1246,13 @@ intel_tv_detect_type (struct intel_tv *intel_tv) int type; /* Disable TV interrupts around load detect or we'll recurse */ - spin_lock_irqsave(&dev_priv->irq_lock, irqflags); - i915_disable_pipestat(dev_priv, 0, - PIPE_HOTPLUG_INTERRUPT_ENABLE | - PIPE_HOTPLUG_TV_INTERRUPT_ENABLE); - spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + if (connector->polled & DRM_CONNECTOR_POLL_HPD) { + spin_lock_irqsave(&dev_priv->irq_lock, irqflags); + i915_disable_pipestat(dev_priv, 0, + PIPE_HOTPLUG_INTERRUPT_ENABLE | + PIPE_HOTPLUG_TV_INTERRUPT_ENABLE); + spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + } save_tv_dac = tv_dac = I915_READ(TV_DAC); save_tv_ctl = tv_ctl = I915_READ(TV_CTL); @@ -1302,11 +1305,13 @@ intel_tv_detect_type (struct intel_tv *intel_tv) I915_WRITE(TV_CTL, save_tv_ctl); /* Restore interrupt config */ - spin_lock_irqsave(&dev_priv->irq_lock, irqflags); - i915_enable_pipestat(dev_priv, 0, - PIPE_HOTPLUG_INTERRUPT_ENABLE | - PIPE_HOTPLUG_TV_INTERRUPT_ENABLE); - spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + if (connector->polled & DRM_CONNECTOR_POLL_HPD) { + spin_lock_irqsave(&dev_priv->irq_lock, irqflags); + i915_enable_pipestat(dev_priv, 0, + PIPE_HOTPLUG_INTERRUPT_ENABLE | + PIPE_HOTPLUG_TV_INTERRUPT_ENABLE); + spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags); + } return type; } @@ -1356,7 +1361,7 @@ intel_tv_detect(struct drm_connector *connector, bool force) drm_mode_set_crtcinfo(&mode, CRTC_INTERLACE_HALVE_V); if (intel_tv->base.base.crtc && intel_tv->base.base.crtc->enabled) { - type = intel_tv_detect_type(intel_tv); + type = intel_tv_detect_type(intel_tv, connector); } else if (force) { struct drm_crtc *crtc; int dpms_mode; @@ -1364,7 +1369,7 @@ intel_tv_detect(struct drm_connector *connector, bool force) crtc = intel_get_load_detect_pipe(&intel_tv->base, connector, &mode, &dpms_mode); if (crtc) { - type = intel_tv_detect_type(intel_tv); + type = intel_tv_detect_type(intel_tv, connector); intel_release_load_detect_pipe(&intel_tv->base, connector, dpms_mode); } else @@ -1658,6 +1663,18 @@ intel_tv_init(struct drm_device *dev) intel_encoder = &intel_tv->base; connector = &intel_connector->base; + /* The documentation, for the older chipsets at least, recommend + * using a polling method rather than hotplug detection for TVs. + * This is because in order to perform the hotplug detection, the PLLs + * for the TV must be kept alive increasing power drain and starving + * bandwidth from other encoders. Notably for instance, it causes + * pipe underruns on Crestline when this encoder is supposedly idle. + * + * More recent chipsets favour HDMI rather than integrated S-Video. + */ + connector->polled = + DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT; + drm_connector_init(dev, connector, &intel_tv_connector_funcs, DRM_MODE_CONNECTOR_SVIDEO); -- cgit v1.2.3 From 04dbff52600719017598f7439bf42e5a72e7de3b Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 10 Feb 2011 17:38:35 +0000 Subject: drm/i915: Fix resume regression from 5d1d0cc MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The irony of the patch to fix the resume regression on PineView causing a further regression on Ironlake is not lost on me. Reported-by: Jeff Chua Reported-by: Björn Schließmann Tested-by: Björn Schließmann Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=28802 Signed-off-by: Chris Wilson --- drivers/gpu/drm/i915/intel_display.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 94622e3a202e..3b006536b3d2 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -5558,9 +5558,7 @@ static void intel_crtc_reset(struct drm_crtc *crtc) /* Reset flags back to the 'unknown' status so that they * will be correctly set on the initial modeset. */ - intel_crtc->cursor_addr = 0; intel_crtc->dpms_mode = -1; - intel_crtc->active = true; /* force the pipe off on setup_init_config */ } static struct drm_crtc_helper_funcs intel_helper_funcs = { @@ -5666,6 +5664,7 @@ static void intel_crtc_init(struct drm_device *dev, int pipe) dev_priv->pipe_to_crtc_mapping[intel_crtc->pipe] = &intel_crtc->base; intel_crtc_reset(&intel_crtc->base); + intel_crtc->active = true; /* force the pipe off on setup_init_config */ if (HAS_PCH_SPLIT(dev)) { intel_helper_funcs.prepare = ironlake_crtc_prepare; -- cgit v1.2.3 From 16f9fdcbcce74102bed9a4b7ccc1fb05b5dd6ca3 Mon Sep 17 00:00:00 2001 From: Dave Airlie Date: Mon, 7 Feb 2011 12:00:51 +1000 Subject: drm/radeon: fix memory debugging since d961db75ce86a84f1f04e91ad1014653ed7d9f46 The old code dereferenced a value, the new code just needs to pass the ptr. fixes an oops looking at files in debugfs. cc: stable@kernel.org Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/radeon_ttm.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c index 1272e4b6a1d4..e5b2cf10cbf4 100644 --- a/drivers/gpu/drm/radeon/radeon_ttm.c +++ b/drivers/gpu/drm/radeon/radeon_ttm.c @@ -787,9 +787,9 @@ static int radeon_ttm_debugfs_init(struct radeon_device *rdev) radeon_mem_types_list[i].show = &radeon_mm_dump_table; radeon_mem_types_list[i].driver_features = 0; if (i == 0) - radeon_mem_types_list[i].data = &rdev->mman.bdev.man[TTM_PL_VRAM].priv; + radeon_mem_types_list[i].data = rdev->mman.bdev.man[TTM_PL_VRAM].priv; else - radeon_mem_types_list[i].data = &rdev->mman.bdev.man[TTM_PL_TT].priv; + radeon_mem_types_list[i].data = rdev->mman.bdev.man[TTM_PL_TT].priv; } /* Add ttm page pool to debugfs */ -- cgit v1.2.3 From c9417bdd4c6b1b92a21608c07e83afa419c7bb62 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Sun, 6 Feb 2011 14:23:26 -0500 Subject: drm/radeon/kms: fix interlaced modes on dce4+ - set scaler table clears the interleave bit, need to reset it in encoder quirks, this was already done for pre-dce4. - remove the interleave settings from set_base() functions this is now handled in the encoder quirks functions, and isn't technically part of the display base setup. - rename evergreen_do_set_base() to dce4_do_set_base() since it's used on both evergreen and NI asics. Fixes: https://bugzilla.kernel.org/show_bug.cgi?id=28182 Signed-off-by: Alex Deucher Cc: stable@kernel.org Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/atombios_crtc.c | 22 +++++----------------- drivers/gpu/drm/radeon/radeon_encoders.c | 20 +++++++++++++++----- 2 files changed, 20 insertions(+), 22 deletions(-) diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c index b1537000a104..dd4e3acf79c0 100644 --- a/drivers/gpu/drm/radeon/atombios_crtc.c +++ b/drivers/gpu/drm/radeon/atombios_crtc.c @@ -995,9 +995,9 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode } } -static int evergreen_crtc_do_set_base(struct drm_crtc *crtc, - struct drm_framebuffer *fb, - int x, int y, int atomic) +static int dce4_crtc_do_set_base(struct drm_crtc *crtc, + struct drm_framebuffer *fb, + int x, int y, int atomic) { struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc); struct drm_device *dev = crtc->dev; @@ -1137,12 +1137,6 @@ static int evergreen_crtc_do_set_base(struct drm_crtc *crtc, WREG32(EVERGREEN_VIEWPORT_SIZE + radeon_crtc->crtc_offset, (crtc->mode.hdisplay << 16) | crtc->mode.vdisplay); - if (crtc->mode.flags & DRM_MODE_FLAG_INTERLACE) - WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset, - EVERGREEN_INTERLEAVE_EN); - else - WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset, 0); - if (!atomic && fb && fb != crtc->fb) { radeon_fb = to_radeon_framebuffer(fb); rbo = radeon_fb->obj->driver_private; @@ -1300,12 +1294,6 @@ static int avivo_crtc_do_set_base(struct drm_crtc *crtc, WREG32(AVIVO_D1MODE_VIEWPORT_SIZE + radeon_crtc->crtc_offset, (crtc->mode.hdisplay << 16) | crtc->mode.vdisplay); - if (crtc->mode.flags & DRM_MODE_FLAG_INTERLACE) - WREG32(AVIVO_D1MODE_DATA_FORMAT + radeon_crtc->crtc_offset, - AVIVO_D1MODE_INTERLEAVE_EN); - else - WREG32(AVIVO_D1MODE_DATA_FORMAT + radeon_crtc->crtc_offset, 0); - if (!atomic && fb && fb != crtc->fb) { radeon_fb = to_radeon_framebuffer(fb); rbo = radeon_fb->obj->driver_private; @@ -1329,7 +1317,7 @@ int atombios_crtc_set_base(struct drm_crtc *crtc, int x, int y, struct radeon_device *rdev = dev->dev_private; if (ASIC_IS_DCE4(rdev)) - return evergreen_crtc_do_set_base(crtc, old_fb, x, y, 0); + return dce4_crtc_do_set_base(crtc, old_fb, x, y, 0); else if (ASIC_IS_AVIVO(rdev)) return avivo_crtc_do_set_base(crtc, old_fb, x, y, 0); else @@ -1344,7 +1332,7 @@ int atombios_crtc_set_base_atomic(struct drm_crtc *crtc, struct radeon_device *rdev = dev->dev_private; if (ASIC_IS_DCE4(rdev)) - return evergreen_crtc_do_set_base(crtc, fb, x, y, 1); + return dce4_crtc_do_set_base(crtc, fb, x, y, 1); else if (ASIC_IS_AVIVO(rdev)) return avivo_crtc_do_set_base(crtc, fb, x, y, 1); else diff --git a/drivers/gpu/drm/radeon/radeon_encoders.c b/drivers/gpu/drm/radeon/radeon_encoders.c index d4a542247618..5b38b73ccd12 100644 --- a/drivers/gpu/drm/radeon/radeon_encoders.c +++ b/drivers/gpu/drm/radeon/radeon_encoders.c @@ -1570,11 +1570,21 @@ atombios_apply_encoder_quirks(struct drm_encoder *encoder, } /* set scaler clears this on some chips */ - /* XXX check DCE4 */ - if (!(radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))) { - if (ASIC_IS_AVIVO(rdev) && (mode->flags & DRM_MODE_FLAG_INTERLACE)) - WREG32(AVIVO_D1MODE_DATA_FORMAT + radeon_crtc->crtc_offset, - AVIVO_D1MODE_INTERLEAVE_EN); + if (ASIC_IS_AVIVO(rdev) && + (!(radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)))) { + if (ASIC_IS_DCE4(rdev)) { + if (mode->flags & DRM_MODE_FLAG_INTERLACE) + WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset, + EVERGREEN_INTERLEAVE_EN); + else + WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset, 0); + } else { + if (mode->flags & DRM_MODE_FLAG_INTERLACE) + WREG32(AVIVO_D1MODE_DATA_FORMAT + radeon_crtc->crtc_offset, + AVIVO_D1MODE_INTERLEAVE_EN); + else + WREG32(AVIVO_D1MODE_DATA_FORMAT + radeon_crtc->crtc_offset, 0); + } } } -- cgit v1.2.3 From e917fd39eb35e5b2c464e67a80e759f3eb468e48 Mon Sep 17 00:00:00 2001 From: Jesper Juhl Date: Sat, 5 Feb 2011 20:51:53 +0100 Subject: radeon mkregtable: Add missing fclose() calls drivers/gpu/drm/radeon/mkregtable.c:parser_auth() almost always remembers to fclose(file) before returning, but it misses two spots. This is not really important since the process will exit shortly after and thus close the file for us, but being explicit prevents static analysis tools from complaining about leaked memory and missing fclose() calls and it also seems to be the prefered style of the existing code to explicitly close the file. So, here's a patch to add the two missing fclose() calls. Signed-off-by: Jesper Juhl Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/mkregtable.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/radeon/mkregtable.c b/drivers/gpu/drm/radeon/mkregtable.c index 607241c6a8a9..5a82b6b75849 100644 --- a/drivers/gpu/drm/radeon/mkregtable.c +++ b/drivers/gpu/drm/radeon/mkregtable.c @@ -673,8 +673,10 @@ static int parser_auth(struct table *t, const char *filename) last_reg = strtol(last_reg_s, NULL, 16); do { - if (fgets(buf, 1024, file) == NULL) + if (fgets(buf, 1024, file) == NULL) { + fclose(file); return -1; + } len = strlen(buf); if (ftell(file) == end) done = 1; @@ -685,6 +687,7 @@ static int parser_auth(struct table *t, const char *filename) fprintf(stderr, "Error matching regular expression %d in %s\n", r, filename); + fclose(file); return -1; } else { buf[match[0].rm_eo] = 0; -- cgit v1.2.3 From 9fad321ac6bedd96f449754a1a25289ea1789a49 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Mon, 7 Feb 2011 13:15:28 -0500 Subject: drm/radeon/kms: add connector table for mac g5 9600 PPC Mac cards do not provide connector tables in their vbios. Their connector/encoder configurations must be hardcoded in the driver. verified by nyef on #radeon Signed-off-by: Alex Deucher Cc: stable@kernel.org Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/radeon_combios.c | 47 +++++++++++++++++++++++++++++++++ drivers/gpu/drm/radeon/radeon_mode.h | 1 + 2 files changed, 48 insertions(+) diff --git a/drivers/gpu/drm/radeon/radeon_combios.c b/drivers/gpu/drm/radeon/radeon_combios.c index d27ef74590cd..cf7c8d5b4ec2 100644 --- a/drivers/gpu/drm/radeon/radeon_combios.c +++ b/drivers/gpu/drm/radeon/radeon_combios.c @@ -1504,6 +1504,11 @@ bool radeon_get_legacy_connector_info_from_table(struct drm_device *dev) (rdev->pdev->subsystem_device == 0x4a48)) { /* Mac X800 */ rdev->mode_info.connector_table = CT_MAC_X800; + } else if ((rdev->pdev->device == 0x4150) && + (rdev->pdev->subsystem_vendor == 0x1002) && + (rdev->pdev->subsystem_device == 0x4150)) { + /* Mac G5 9600 */ + rdev->mode_info.connector_table = CT_MAC_G5_9600; } else #endif /* CONFIG_PPC_PMAC */ #ifdef CONFIG_PPC64 @@ -2022,6 +2027,48 @@ bool radeon_get_legacy_connector_info_from_table(struct drm_device *dev) CONNECTOR_OBJECT_ID_DUAL_LINK_DVI_I, &hpd); break; + case CT_MAC_G5_9600: + DRM_INFO("Connector Table: %d (mac g5 9600)\n", + rdev->mode_info.connector_table); + /* DVI - tv dac, dvo */ + ddc_i2c = combios_setup_i2c_bus(rdev, DDC_DVI, 0, 0); + hpd.hpd = RADEON_HPD_1; /* ??? */ + radeon_add_legacy_encoder(dev, + radeon_get_encoder_enum(dev, + ATOM_DEVICE_DFP2_SUPPORT, + 0), + ATOM_DEVICE_DFP2_SUPPORT); + radeon_add_legacy_encoder(dev, + radeon_get_encoder_enum(dev, + ATOM_DEVICE_CRT2_SUPPORT, + 2), + ATOM_DEVICE_CRT2_SUPPORT); + radeon_add_legacy_connector(dev, 0, + ATOM_DEVICE_DFP2_SUPPORT | + ATOM_DEVICE_CRT2_SUPPORT, + DRM_MODE_CONNECTOR_DVII, &ddc_i2c, + CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_I, + &hpd); + /* ADC - primary dac, internal tmds */ + ddc_i2c = combios_setup_i2c_bus(rdev, DDC_VGA, 0, 0); + hpd.hpd = RADEON_HPD_2; /* ??? */ + radeon_add_legacy_encoder(dev, + radeon_get_encoder_enum(dev, + ATOM_DEVICE_DFP1_SUPPORT, + 0), + ATOM_DEVICE_DFP1_SUPPORT); + radeon_add_legacy_encoder(dev, + radeon_get_encoder_enum(dev, + ATOM_DEVICE_CRT1_SUPPORT, + 1), + ATOM_DEVICE_CRT1_SUPPORT); + radeon_add_legacy_connector(dev, 1, + ATOM_DEVICE_DFP1_SUPPORT | + ATOM_DEVICE_CRT1_SUPPORT, + DRM_MODE_CONNECTOR_DVII, &ddc_i2c, + CONNECTOR_OBJECT_ID_SINGLE_LINK_DVI_I, + &hpd); + break; default: DRM_INFO("Connector table: %d (invalid)\n", rdev->mode_info.connector_table); diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h index 6794cdf91f28..a670caaee29e 100644 --- a/drivers/gpu/drm/radeon/radeon_mode.h +++ b/drivers/gpu/drm/radeon/radeon_mode.h @@ -209,6 +209,7 @@ enum radeon_connector_table { CT_EMAC, CT_RN50_POWER, CT_MAC_X800, + CT_MAC_G5_9600, }; enum radeon_dvo_chip { -- cgit v1.2.3 From 01e2f533a234dc62d16c0d3d4fb9d71cf1ce50c3 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Fri, 11 Feb 2011 19:29:44 -0800 Subject: drm: do not leak kernel addresses via /proc/dri/*/vma In the continuing effort to avoid kernel addresses leaking to unprivileged users, this patch switches to %pK for /proc/dri/*/vma. Signed-off-by: Kees Cook Signed-off-by: Dave Airlie --- drivers/gpu/drm/drm_info.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/drm_info.c b/drivers/gpu/drm/drm_info.c index 3cdbaf379bb5..be9a9c07d152 100644 --- a/drivers/gpu/drm/drm_info.c +++ b/drivers/gpu/drm/drm_info.c @@ -283,17 +283,18 @@ int drm_vma_info(struct seq_file *m, void *data) #endif mutex_lock(&dev->struct_mutex); - seq_printf(m, "vma use count: %d, high_memory = %p, 0x%08llx\n", + seq_printf(m, "vma use count: %d, high_memory = %pK, 0x%pK\n", atomic_read(&dev->vma_count), - high_memory, (u64)virt_to_phys(high_memory)); + high_memory, (void *)virt_to_phys(high_memory)); list_for_each_entry(pt, &dev->vmalist, head) { vma = pt->vma; if (!vma) continue; seq_printf(m, - "\n%5d 0x%08lx-0x%08lx %c%c%c%c%c%c 0x%08lx000", - pt->pid, vma->vm_start, vma->vm_end, + "\n%5d 0x%pK-0x%pK %c%c%c%c%c%c 0x%08lx000", + pt->pid, + (void *)vma->vm_start, (void *)vma->vm_end, vma->vm_flags & VM_READ ? 'r' : '-', vma->vm_flags & VM_WRITE ? 'w' : '-', vma->vm_flags & VM_EXEC ? 'x' : '-', -- cgit v1.2.3 From 40b4a7599d5555b408e594f4c8dae8015ccaae8f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marek=20Ol=C5=A1=C3=A1k?= Date: Sat, 12 Feb 2011 19:21:35 +0100 Subject: drm/radeon/kms: optimize CS state checking for r100->r500 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The colorbuffer, zbuffer, and texture states are checked only once when they get changed. This improves performance in the apps which emit lots of draw packets and few state changes. This drops performance in glxgears by a 1% or so, but glxgears is not a benchmark we care about. The time spent in the kernel when running Torcs dropped from 33% to 23% and the frame rate is higher, which is a good thing. r600 might need something like this as well. Signed-off-by: Marek Olšák Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/r100.c | 40 +++++++++++++++++++++++++++++++++---- drivers/gpu/drm/radeon/r100_track.h | 11 ++++------ drivers/gpu/drm/radeon/r200.c | 18 +++++++++++++++++ drivers/gpu/drm/radeon/r300.c | 20 ++++++++++++++++++- 4 files changed, 77 insertions(+), 12 deletions(-) diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index 5f15820efe12..fdf4bc67ae58 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -1427,6 +1427,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, } track->zb.robj = reloc->robj; track->zb.offset = idx_value; + track->zb_dirty = true; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); break; case RADEON_RB3D_COLOROFFSET: @@ -1439,6 +1440,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, } track->cb[0].robj = reloc->robj; track->cb[0].offset = idx_value; + track->cb_dirty = true; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); break; case RADEON_PP_TXOFFSET_0: @@ -1454,6 +1456,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, } ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[i].robj = reloc->robj; + track->tex_dirty = true; break; case RADEON_PP_CUBIC_OFFSET_T0_0: case RADEON_PP_CUBIC_OFFSET_T0_1: @@ -1471,6 +1474,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->textures[0].cube_info[i].offset = idx_value; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[0].cube_info[i].robj = reloc->robj; + track->tex_dirty = true; break; case RADEON_PP_CUBIC_OFFSET_T1_0: case RADEON_PP_CUBIC_OFFSET_T1_1: @@ -1488,6 +1492,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->textures[1].cube_info[i].offset = idx_value; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[1].cube_info[i].robj = reloc->robj; + track->tex_dirty = true; break; case RADEON_PP_CUBIC_OFFSET_T2_0: case RADEON_PP_CUBIC_OFFSET_T2_1: @@ -1505,9 +1510,12 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->textures[2].cube_info[i].offset = idx_value; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[2].cube_info[i].robj = reloc->robj; + track->tex_dirty = true; break; case RADEON_RE_WIDTH_HEIGHT: track->maxy = ((idx_value >> 16) & 0x7FF); + track->cb_dirty = true; + track->zb_dirty = true; break; case RADEON_RB3D_COLORPITCH: r = r100_cs_packet_next_reloc(p, &reloc); @@ -1528,9 +1536,11 @@ static int r100_packet0_check(struct radeon_cs_parser *p, ib[idx] = tmp; track->cb[0].pitch = idx_value & RADEON_COLORPITCH_MASK; + track->cb_dirty = true; break; case RADEON_RB3D_DEPTHPITCH: track->zb.pitch = idx_value & RADEON_DEPTHPITCH_MASK; + track->zb_dirty = true; break; case RADEON_RB3D_CNTL: switch ((idx_value >> RADEON_RB3D_COLOR_FORMAT_SHIFT) & 0x1f) { @@ -1555,6 +1565,8 @@ static int r100_packet0_check(struct radeon_cs_parser *p, return -EINVAL; } track->z_enabled = !!(idx_value & RADEON_Z_ENABLE); + track->cb_dirty = true; + track->zb_dirty = true; break; case RADEON_RB3D_ZSTENCILCNTL: switch (idx_value & 0xf) { @@ -1572,6 +1584,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, default: break; } + track->zb_dirty = true; break; case RADEON_RB3D_ZPASS_ADDR: r = r100_cs_packet_next_reloc(p, &reloc); @@ -1588,6 +1601,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, uint32_t temp = idx_value >> 4; for (i = 0; i < track->num_texture; i++) track->textures[i].enabled = !!(temp & (1 << i)); + track->tex_dirty = true; } break; case RADEON_SE_VF_CNTL: @@ -1602,12 +1616,14 @@ static int r100_packet0_check(struct radeon_cs_parser *p, i = (reg - RADEON_PP_TEX_SIZE_0) / 8; track->textures[i].width = (idx_value & RADEON_TEX_USIZE_MASK) + 1; track->textures[i].height = ((idx_value & RADEON_TEX_VSIZE_MASK) >> RADEON_TEX_VSIZE_SHIFT) + 1; + track->tex_dirty = true; break; case RADEON_PP_TEX_PITCH_0: case RADEON_PP_TEX_PITCH_1: case RADEON_PP_TEX_PITCH_2: i = (reg - RADEON_PP_TEX_PITCH_0) / 8; track->textures[i].pitch = idx_value + 32; + track->tex_dirty = true; break; case RADEON_PP_TXFILTER_0: case RADEON_PP_TXFILTER_1: @@ -1621,6 +1637,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, tmp = (idx_value >> 27) & 0x7; if (tmp == 2 || tmp == 6) track->textures[i].roundup_h = false; + track->tex_dirty = true; break; case RADEON_PP_TXFORMAT_0: case RADEON_PP_TXFORMAT_1: @@ -1673,6 +1690,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, } track->textures[i].cube_info[4].width = 1 << ((idx_value >> 16) & 0xf); track->textures[i].cube_info[4].height = 1 << ((idx_value >> 20) & 0xf); + track->tex_dirty = true; break; case RADEON_PP_CUBIC_FACES_0: case RADEON_PP_CUBIC_FACES_1: @@ -1683,6 +1701,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p, track->textures[i].cube_info[face].width = 1 << ((tmp >> (face * 8)) & 0xf); track->textures[i].cube_info[face].height = 1 << ((tmp >> ((face * 8) + 4)) & 0xf); } + track->tex_dirty = true; break; default: printk(KERN_ERR "Forbidden register 0x%04X in cs at %d\n", @@ -3318,9 +3337,9 @@ int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track) unsigned long size; unsigned prim_walk; unsigned nverts; - unsigned num_cb = track->num_cb; + unsigned num_cb = track->cb_dirty ? track->num_cb : 0; - if (!track->zb_cb_clear && !track->color_channel_mask && + if (num_cb && !track->zb_cb_clear && !track->color_channel_mask && !track->blend_read_enable) num_cb = 0; @@ -3341,7 +3360,9 @@ int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track) return -EINVAL; } } - if (track->z_enabled) { + track->cb_dirty = false; + + if (track->zb_dirty && track->z_enabled) { if (track->zb.robj == NULL) { DRM_ERROR("[drm] No buffer for z buffer !\n"); return -EINVAL; @@ -3358,6 +3379,8 @@ int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track) return -EINVAL; } } + track->zb_dirty = false; + prim_walk = (track->vap_vf_cntl >> 4) & 0x3; if (track->vap_vf_cntl & (1 << 14)) { nverts = track->vap_alt_nverts; @@ -3417,13 +3440,22 @@ int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track) prim_walk); return -EINVAL; } - return r100_cs_track_texture_check(rdev, track); + + if (track->tex_dirty) { + track->tex_dirty = false; + return r100_cs_track_texture_check(rdev, track); + } + return 0; } void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track) { unsigned i, face; + track->cb_dirty = true; + track->zb_dirty = true; + track->tex_dirty = true; + if (rdev->family < CHIP_R300) { track->num_cb = 1; if (rdev->family <= CHIP_RS200) diff --git a/drivers/gpu/drm/radeon/r100_track.h b/drivers/gpu/drm/radeon/r100_track.h index af65600e6564..ee85c4a1fc08 100644 --- a/drivers/gpu/drm/radeon/r100_track.h +++ b/drivers/gpu/drm/radeon/r100_track.h @@ -52,14 +52,7 @@ struct r100_cs_track_texture { unsigned compress_format; }; -struct r100_cs_track_limits { - unsigned num_cb; - unsigned num_texture; - unsigned max_levels; -}; - struct r100_cs_track { - struct radeon_device *rdev; unsigned num_cb; unsigned num_texture; unsigned maxy; @@ -78,6 +71,10 @@ struct r100_cs_track { bool separate_cube; bool zb_cb_clear; bool blend_read_enable; + + bool cb_dirty; + bool zb_dirty; + bool tex_dirty; }; int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track); diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c index d2408c395619..f24058300413 100644 --- a/drivers/gpu/drm/radeon/r200.c +++ b/drivers/gpu/drm/radeon/r200.c @@ -184,6 +184,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, } track->zb.robj = reloc->robj; track->zb.offset = idx_value; + track->zb_dirty = true; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); break; case RADEON_RB3D_COLOROFFSET: @@ -196,6 +197,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, } track->cb[0].robj = reloc->robj; track->cb[0].offset = idx_value; + track->cb_dirty = true; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); break; case R200_PP_TXOFFSET_0: @@ -214,6 +216,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, } ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[i].robj = reloc->robj; + track->tex_dirty = true; break; case R200_PP_CUBIC_OFFSET_F1_0: case R200_PP_CUBIC_OFFSET_F2_0: @@ -257,9 +260,12 @@ int r200_packet0_check(struct radeon_cs_parser *p, track->textures[i].cube_info[face - 1].offset = idx_value; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); track->textures[i].cube_info[face - 1].robj = reloc->robj; + track->tex_dirty = true; break; case RADEON_RE_WIDTH_HEIGHT: track->maxy = ((idx_value >> 16) & 0x7FF); + track->cb_dirty = true; + track->zb_dirty = true; break; case RADEON_RB3D_COLORPITCH: r = r100_cs_packet_next_reloc(p, &reloc); @@ -280,9 +286,11 @@ int r200_packet0_check(struct radeon_cs_parser *p, ib[idx] = tmp; track->cb[0].pitch = idx_value & RADEON_COLORPITCH_MASK; + track->cb_dirty = true; break; case RADEON_RB3D_DEPTHPITCH: track->zb.pitch = idx_value & RADEON_DEPTHPITCH_MASK; + track->zb_dirty = true; break; case RADEON_RB3D_CNTL: switch ((idx_value >> RADEON_RB3D_COLOR_FORMAT_SHIFT) & 0x1f) { @@ -312,6 +320,8 @@ int r200_packet0_check(struct radeon_cs_parser *p, } track->z_enabled = !!(idx_value & RADEON_Z_ENABLE); + track->cb_dirty = true; + track->zb_dirty = true; break; case RADEON_RB3D_ZSTENCILCNTL: switch (idx_value & 0xf) { @@ -329,6 +339,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, default: break; } + track->zb_dirty = true; break; case RADEON_RB3D_ZPASS_ADDR: r = r100_cs_packet_next_reloc(p, &reloc); @@ -345,6 +356,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, uint32_t temp = idx_value >> 4; for (i = 0; i < track->num_texture; i++) track->textures[i].enabled = !!(temp & (1 << i)); + track->tex_dirty = true; } break; case RADEON_SE_VF_CNTL: @@ -369,6 +381,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, i = (reg - R200_PP_TXSIZE_0) / 32; track->textures[i].width = (idx_value & RADEON_TEX_USIZE_MASK) + 1; track->textures[i].height = ((idx_value & RADEON_TEX_VSIZE_MASK) >> RADEON_TEX_VSIZE_SHIFT) + 1; + track->tex_dirty = true; break; case R200_PP_TXPITCH_0: case R200_PP_TXPITCH_1: @@ -378,6 +391,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, case R200_PP_TXPITCH_5: i = (reg - R200_PP_TXPITCH_0) / 32; track->textures[i].pitch = idx_value + 32; + track->tex_dirty = true; break; case R200_PP_TXFILTER_0: case R200_PP_TXFILTER_1: @@ -394,6 +408,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, tmp = (idx_value >> 27) & 0x7; if (tmp == 2 || tmp == 6) track->textures[i].roundup_h = false; + track->tex_dirty = true; break; case R200_PP_TXMULTI_CTL_0: case R200_PP_TXMULTI_CTL_1: @@ -432,6 +447,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, track->textures[i].tex_coord_type = 1; break; } + track->tex_dirty = true; break; case R200_PP_TXFORMAT_0: case R200_PP_TXFORMAT_1: @@ -488,6 +504,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, } track->textures[i].cube_info[4].width = 1 << ((idx_value >> 16) & 0xf); track->textures[i].cube_info[4].height = 1 << ((idx_value >> 20) & 0xf); + track->tex_dirty = true; break; case R200_PP_CUBIC_FACES_0: case R200_PP_CUBIC_FACES_1: @@ -501,6 +518,7 @@ int r200_packet0_check(struct radeon_cs_parser *p, track->textures[i].cube_info[face].width = 1 << ((tmp >> (face * 8)) & 0xf); track->textures[i].cube_info[face].height = 1 << ((tmp >> ((face * 8) + 4)) & 0xf); } + track->tex_dirty = true; break; default: printk(KERN_ERR "Forbidden register 0x%04X in cs at %d\n", diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c index 55fe5ba7def3..15f94648f274 100644 --- a/drivers/gpu/drm/radeon/r300.c +++ b/drivers/gpu/drm/radeon/r300.c @@ -667,6 +667,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, } track->cb[i].robj = reloc->robj; track->cb[i].offset = idx_value; + track->cb_dirty = true; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); break; case R300_ZB_DEPTHOFFSET: @@ -679,6 +680,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, } track->zb.robj = reloc->robj; track->zb.offset = idx_value; + track->zb_dirty = true; ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); break; case R300_TX_OFFSET_0: @@ -717,6 +719,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, tmp |= tile_flags; ib[idx] = tmp; track->textures[i].robj = reloc->robj; + track->tex_dirty = true; break; /* Tracked registers */ case 0x2084: @@ -743,6 +746,8 @@ static int r300_packet0_check(struct radeon_cs_parser *p, if (p->rdev->family < CHIP_RV515) { track->maxy -= 1440; } + track->cb_dirty = true; + track->zb_dirty = true; break; case 0x4E00: /* RB3D_CCTL */ @@ -752,6 +757,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, return -EINVAL; } track->num_cb = ((idx_value >> 5) & 0x3) + 1; + track->cb_dirty = true; break; case 0x4E38: case 0x4E3C: @@ -814,6 +820,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, ((idx_value >> 21) & 0xF)); return -EINVAL; } + track->cb_dirty = true; break; case 0x4F00: /* ZB_CNTL */ @@ -822,6 +829,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, } else { track->z_enabled = false; } + track->zb_dirty = true; break; case 0x4F10: /* ZB_FORMAT */ @@ -838,6 +846,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, (idx_value & 0xF)); return -EINVAL; } + track->zb_dirty = true; break; case 0x4F24: /* ZB_DEPTHPITCH */ @@ -861,6 +870,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, ib[idx] = tmp; track->zb.pitch = idx_value & 0x3FFC; + track->zb_dirty = true; break; case 0x4104: for (i = 0; i < 16; i++) { @@ -869,6 +879,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, enabled = !!(idx_value & (1 << i)); track->textures[i].enabled = enabled; } + track->tex_dirty = true; break; case 0x44C0: case 0x44C4: @@ -951,8 +962,8 @@ static int r300_packet0_check(struct radeon_cs_parser *p, DRM_ERROR("Invalid texture format %u\n", (idx_value & 0x1F)); return -EINVAL; - break; } + track->tex_dirty = true; break; case 0x4400: case 0x4404: @@ -980,6 +991,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, if (tmp == 2 || tmp == 4 || tmp == 6) { track->textures[i].roundup_h = false; } + track->tex_dirty = true; break; case 0x4500: case 0x4504: @@ -1017,6 +1029,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, DRM_ERROR("Forbidden bit TXFORMAT_MSB\n"); return -EINVAL; } + track->tex_dirty = true; break; case 0x4480: case 0x4484: @@ -1046,6 +1059,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, track->textures[i].use_pitch = !!tmp; tmp = (idx_value >> 22) & 0xF; track->textures[i].txdepth = tmp; + track->tex_dirty = true; break; case R300_ZB_ZPASS_ADDR: r = r100_cs_packet_next_reloc(p, &reloc); @@ -1060,6 +1074,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, case 0x4e0c: /* RB3D_COLOR_CHANNEL_MASK */ track->color_channel_mask = idx_value; + track->cb_dirty = true; break; case 0x43a4: /* SC_HYPERZ_EN */ @@ -1073,6 +1088,8 @@ static int r300_packet0_check(struct radeon_cs_parser *p, case 0x4f1c: /* ZB_BW_CNTL */ track->zb_cb_clear = !!(idx_value & (1 << 5)); + track->cb_dirty = true; + track->zb_dirty = true; if (p->rdev->hyperz_filp != p->filp) { if (idx_value & (R300_HIZ_ENABLE | R300_RD_COMP_ENABLE | @@ -1084,6 +1101,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, case 0x4e04: /* RB3D_BLENDCNTL */ track->blend_read_enable = !!(idx_value & (1 << 2)); + track->cb_dirty = true; break; case 0x4f28: /* ZB_DEPTHCLEARVALUE */ break; -- cgit v1.2.3 From dee54c40a1a9898bcd156436a1d3524f530b5a90 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?C=C3=A9dric=20Cano?= Date: Fri, 11 Feb 2011 19:45:36 -0500 Subject: drm/radeon: 6xx/7xx non-kms endian fixes MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit agd5f: minor cleanups Signed-off-by: Cédric Cano Signed-off-by: Alex Deucher Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/r600_blit.c | 11 +++++++++-- drivers/gpu/drm/radeon/r600_cp.c | 31 ++++++++++++++++++++++++------- drivers/gpu/drm/radeon/radeon_drv.h | 1 + 3 files changed, 34 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/radeon/r600_blit.c b/drivers/gpu/drm/radeon/r600_blit.c index ca5c29f70779..7f1043448d25 100644 --- a/drivers/gpu/drm/radeon/r600_blit.c +++ b/drivers/gpu/drm/radeon/r600_blit.c @@ -137,9 +137,9 @@ set_shaders(struct drm_device *dev) ps = (u32 *) ((char *)dev->agp_buffer_map->handle + dev_priv->blit_vb->offset + 256); for (i = 0; i < r6xx_vs_size; i++) - vs[i] = r6xx_vs[i]; + vs[i] = cpu_to_le32(r6xx_vs[i]); for (i = 0; i < r6xx_ps_size; i++) - ps[i] = r6xx_ps[i]; + ps[i] = cpu_to_le32(r6xx_ps[i]); dev_priv->blit_vb->used = 512; @@ -192,6 +192,9 @@ set_vtx_resource(drm_radeon_private_t *dev_priv, u64 gpu_addr) DRM_DEBUG("\n"); sq_vtx_constant_word2 = (((gpu_addr >> 32) & 0xff) | (16 << 8)); +#ifdef __BIG_ENDIAN + sq_vtx_constant_word2 |= (2 << 30); +#endif BEGIN_RING(9); OUT_RING(CP_PACKET3(R600_IT_SET_RESOURCE, 7)); @@ -291,7 +294,11 @@ draw_auto(drm_radeon_private_t *dev_priv) OUT_RING(DI_PT_RECTLIST); OUT_RING(CP_PACKET3(R600_IT_INDEX_TYPE, 0)); +#ifdef __BIG_ENDIAN + OUT_RING((2 << 2) | DI_INDEX_SIZE_16_BIT); +#else OUT_RING(DI_INDEX_SIZE_16_BIT); +#endif OUT_RING(CP_PACKET3(R600_IT_NUM_INSTANCES, 0)); OUT_RING(1); diff --git a/drivers/gpu/drm/radeon/r600_cp.c b/drivers/gpu/drm/radeon/r600_cp.c index 4f4cd8b286d5..c3ab959bdc7c 100644 --- a/drivers/gpu/drm/radeon/r600_cp.c +++ b/drivers/gpu/drm/radeon/r600_cp.c @@ -396,6 +396,9 @@ static void r600_cp_load_microcode(drm_radeon_private_t *dev_priv) r600_do_cp_stop(dev_priv); RADEON_WRITE(R600_CP_RB_CNTL, +#ifdef __BIG_ENDIAN + R600_BUF_SWAP_32BIT | +#endif R600_RB_NO_UPDATE | R600_RB_BLKSZ(15) | R600_RB_BUFSZ(3)); @@ -486,9 +489,12 @@ static void r700_cp_load_microcode(drm_radeon_private_t *dev_priv) r600_do_cp_stop(dev_priv); RADEON_WRITE(R600_CP_RB_CNTL, +#ifdef __BIG_ENDIAN + R600_BUF_SWAP_32BIT | +#endif R600_RB_NO_UPDATE | - (15 << 8) | - (3 << 0)); + R600_RB_BLKSZ(15) | + R600_RB_BUFSZ(3)); RADEON_WRITE(R600_GRBM_SOFT_RESET, R600_SOFT_RESET_CP); RADEON_READ(R600_GRBM_SOFT_RESET); @@ -550,8 +556,12 @@ static void r600_test_writeback(drm_radeon_private_t *dev_priv) if (!dev_priv->writeback_works) { /* Disable writeback to avoid unnecessary bus master transfer */ - RADEON_WRITE(R600_CP_RB_CNTL, RADEON_READ(R600_CP_RB_CNTL) | - RADEON_RB_NO_UPDATE); + RADEON_WRITE(R600_CP_RB_CNTL, +#ifdef __BIG_ENDIAN + R600_BUF_SWAP_32BIT | +#endif + RADEON_READ(R600_CP_RB_CNTL) | + R600_RB_NO_UPDATE); RADEON_WRITE(R600_SCRATCH_UMSK, 0); } } @@ -575,7 +585,11 @@ int r600_do_engine_reset(struct drm_device *dev) RADEON_WRITE(R600_CP_RB_WPTR_DELAY, 0); cp_rb_cntl = RADEON_READ(R600_CP_RB_CNTL); - RADEON_WRITE(R600_CP_RB_CNTL, R600_RB_RPTR_WR_ENA); + RADEON_WRITE(R600_CP_RB_CNTL, +#ifdef __BIG_ENDIAN + R600_BUF_SWAP_32BIT | +#endif + R600_RB_RPTR_WR_ENA); RADEON_WRITE(R600_CP_RB_RPTR_WR, cp_ptr); RADEON_WRITE(R600_CP_RB_WPTR, cp_ptr); @@ -1838,7 +1852,10 @@ static void r600_cp_init_ring_buffer(struct drm_device *dev, + dev_priv->gart_vm_start; } RADEON_WRITE(R600_CP_RB_RPTR_ADDR, - rptr_addr & 0xffffffff); +#ifdef __BIG_ENDIAN + (2 << 0) | +#endif + (rptr_addr & 0xfffffffc)); RADEON_WRITE(R600_CP_RB_RPTR_ADDR_HI, upper_32_bits(rptr_addr)); @@ -1889,7 +1906,7 @@ static void r600_cp_init_ring_buffer(struct drm_device *dev, { u64 scratch_addr; - scratch_addr = RADEON_READ(R600_CP_RB_RPTR_ADDR); + scratch_addr = RADEON_READ(R600_CP_RB_RPTR_ADDR) & 0xFFFFFFFC; scratch_addr |= ((u64)RADEON_READ(R600_CP_RB_RPTR_ADDR_HI)) << 32; scratch_addr += R600_SCRATCH_REG_OFFSET; scratch_addr >>= 8; diff --git a/drivers/gpu/drm/radeon/radeon_drv.h b/drivers/gpu/drm/radeon/radeon_drv.h index 448eba89d1e6..5cba46b9779a 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.h +++ b/drivers/gpu/drm/radeon/radeon_drv.h @@ -1524,6 +1524,7 @@ extern u32 radeon_get_scratch(drm_radeon_private_t *dev_priv, int index); #define R600_CP_RB_CNTL 0xc104 # define R600_RB_BUFSZ(x) ((x) << 0) # define R600_RB_BLKSZ(x) ((x) << 8) +# define R600_BUF_SWAP_32BIT (2 << 16) # define R600_RB_NO_UPDATE (1 << 27) # define R600_RB_RPTR_WR_ENA (1 << 31) #define R600_CP_RB_RPTR_WR 0xc108 -- cgit v1.2.3 From 4589433c57bd34b7e49068549e07a43c8d41e39d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?C=C3=A9dric=20Cano?= Date: Fri, 11 Feb 2011 19:45:37 -0500 Subject: drm/radeon/kms: atombios big endian fixes MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit agd5f: additional cleanups/fixes Signed-off-by: Cédric Cano Signed-off-by: Alex Deucher Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/atombios_crtc.c | 52 ++++++++++++++++---------------- drivers/gpu/drm/radeon/radeon_atombios.c | 48 ++++++++++++++--------------- drivers/gpu/drm/radeon/radeon_encoders.c | 4 +-- 3 files changed, 52 insertions(+), 52 deletions(-) diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c index dd4e3acf79c0..1bf61220a450 100644 --- a/drivers/gpu/drm/radeon/atombios_crtc.c +++ b/drivers/gpu/drm/radeon/atombios_crtc.c @@ -48,29 +48,29 @@ static void atombios_overscan_setup(struct drm_crtc *crtc, switch (radeon_crtc->rmx_type) { case RMX_CENTER: - args.usOverscanTop = (adjusted_mode->crtc_vdisplay - mode->crtc_vdisplay) / 2; - args.usOverscanBottom = (adjusted_mode->crtc_vdisplay - mode->crtc_vdisplay) / 2; - args.usOverscanLeft = (adjusted_mode->crtc_hdisplay - mode->crtc_hdisplay) / 2; - args.usOverscanRight = (adjusted_mode->crtc_hdisplay - mode->crtc_hdisplay) / 2; + args.usOverscanTop = cpu_to_le16((adjusted_mode->crtc_vdisplay - mode->crtc_vdisplay) / 2); + args.usOverscanBottom = cpu_to_le16((adjusted_mode->crtc_vdisplay - mode->crtc_vdisplay) / 2); + args.usOverscanLeft = cpu_to_le16((adjusted_mode->crtc_hdisplay - mode->crtc_hdisplay) / 2); + args.usOverscanRight = cpu_to_le16((adjusted_mode->crtc_hdisplay - mode->crtc_hdisplay) / 2); break; case RMX_ASPECT: a1 = mode->crtc_vdisplay * adjusted_mode->crtc_hdisplay; a2 = adjusted_mode->crtc_vdisplay * mode->crtc_hdisplay; if (a1 > a2) { - args.usOverscanLeft = (adjusted_mode->crtc_hdisplay - (a2 / mode->crtc_vdisplay)) / 2; - args.usOverscanRight = (adjusted_mode->crtc_hdisplay - (a2 / mode->crtc_vdisplay)) / 2; + args.usOverscanLeft = cpu_to_le16((adjusted_mode->crtc_hdisplay - (a2 / mode->crtc_vdisplay)) / 2); + args.usOverscanRight = cpu_to_le16((adjusted_mode->crtc_hdisplay - (a2 / mode->crtc_vdisplay)) / 2); } else if (a2 > a1) { - args.usOverscanLeft = (adjusted_mode->crtc_vdisplay - (a1 / mode->crtc_hdisplay)) / 2; - args.usOverscanRight = (adjusted_mode->crtc_vdisplay - (a1 / mode->crtc_hdisplay)) / 2; + args.usOverscanLeft = cpu_to_le16((adjusted_mode->crtc_vdisplay - (a1 / mode->crtc_hdisplay)) / 2); + args.usOverscanRight = cpu_to_le16((adjusted_mode->crtc_vdisplay - (a1 / mode->crtc_hdisplay)) / 2); } break; case RMX_FULL: default: - args.usOverscanRight = radeon_crtc->h_border; - args.usOverscanLeft = radeon_crtc->h_border; - args.usOverscanBottom = radeon_crtc->v_border; - args.usOverscanTop = radeon_crtc->v_border; + args.usOverscanRight = cpu_to_le16(radeon_crtc->h_border); + args.usOverscanLeft = cpu_to_le16(radeon_crtc->h_border); + args.usOverscanBottom = cpu_to_le16(radeon_crtc->v_border); + args.usOverscanTop = cpu_to_le16(radeon_crtc->v_border); break; } atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); @@ -419,23 +419,23 @@ static void atombios_crtc_program_ss(struct drm_crtc *crtc, memset(&args, 0, sizeof(args)); if (ASIC_IS_DCE5(rdev)) { - args.v3.usSpreadSpectrumAmountFrac = 0; + args.v3.usSpreadSpectrumAmountFrac = cpu_to_le16(0); args.v3.ucSpreadSpectrumType = ss->type; switch (pll_id) { case ATOM_PPLL1: args.v3.ucSpreadSpectrumType |= ATOM_PPLL_SS_TYPE_V3_P1PLL; - args.v3.usSpreadSpectrumAmount = ss->amount; - args.v3.usSpreadSpectrumStep = ss->step; + args.v3.usSpreadSpectrumAmount = cpu_to_le16(ss->amount); + args.v3.usSpreadSpectrumStep = cpu_to_le16(ss->step); break; case ATOM_PPLL2: args.v3.ucSpreadSpectrumType |= ATOM_PPLL_SS_TYPE_V3_P2PLL; - args.v3.usSpreadSpectrumAmount = ss->amount; - args.v3.usSpreadSpectrumStep = ss->step; + args.v3.usSpreadSpectrumAmount = cpu_to_le16(ss->amount); + args.v3.usSpreadSpectrumStep = cpu_to_le16(ss->step); break; case ATOM_DCPLL: args.v3.ucSpreadSpectrumType |= ATOM_PPLL_SS_TYPE_V3_DCPLL; - args.v3.usSpreadSpectrumAmount = 0; - args.v3.usSpreadSpectrumStep = 0; + args.v3.usSpreadSpectrumAmount = cpu_to_le16(0); + args.v3.usSpreadSpectrumStep = cpu_to_le16(0); break; case ATOM_PPLL_INVALID: return; @@ -447,18 +447,18 @@ static void atombios_crtc_program_ss(struct drm_crtc *crtc, switch (pll_id) { case ATOM_PPLL1: args.v2.ucSpreadSpectrumType |= ATOM_PPLL_SS_TYPE_V2_P1PLL; - args.v2.usSpreadSpectrumAmount = ss->amount; - args.v2.usSpreadSpectrumStep = ss->step; + args.v2.usSpreadSpectrumAmount = cpu_to_le16(ss->amount); + args.v2.usSpreadSpectrumStep = cpu_to_le16(ss->step); break; case ATOM_PPLL2: args.v2.ucSpreadSpectrumType |= ATOM_PPLL_SS_TYPE_V2_P2PLL; - args.v2.usSpreadSpectrumAmount = ss->amount; - args.v2.usSpreadSpectrumStep = ss->step; + args.v2.usSpreadSpectrumAmount = cpu_to_le16(ss->amount); + args.v2.usSpreadSpectrumStep = cpu_to_le16(ss->step); break; case ATOM_DCPLL: args.v2.ucSpreadSpectrumType |= ATOM_PPLL_SS_TYPE_V2_DCPLL; - args.v2.usSpreadSpectrumAmount = 0; - args.v2.usSpreadSpectrumStep = 0; + args.v2.usSpreadSpectrumAmount = cpu_to_le16(0); + args.v2.usSpreadSpectrumStep = cpu_to_le16(0); break; case ATOM_PPLL_INVALID: return; @@ -721,7 +721,7 @@ static void atombios_crtc_set_dcpll(struct drm_crtc *crtc, * SetPixelClock provides the dividers */ args.v5.ucCRTC = ATOM_CRTC_INVALID; - args.v5.usPixelClock = dispclk; + args.v5.usPixelClock = cpu_to_le16(dispclk); args.v5.ucPpll = ATOM_DCPLL; break; case 6: diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c index 5c1cc7ad9a15..02d5c415f499 100644 --- a/drivers/gpu/drm/radeon/radeon_atombios.c +++ b/drivers/gpu/drm/radeon/radeon_atombios.c @@ -88,7 +88,7 @@ static inline struct radeon_i2c_bus_rec radeon_lookup_i2c_gpio(struct radeon_dev /* some evergreen boards have bad data for this entry */ if (ASIC_IS_DCE4(rdev)) { if ((i == 7) && - (gpio->usClkMaskRegisterIndex == 0x1936) && + (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x1936) && (gpio->sucI2cId.ucAccess == 0)) { gpio->sucI2cId.ucAccess = 0x97; gpio->ucDataMaskShift = 8; @@ -101,7 +101,7 @@ static inline struct radeon_i2c_bus_rec radeon_lookup_i2c_gpio(struct radeon_dev /* some DCE3 boards have bad data for this entry */ if (ASIC_IS_DCE3(rdev)) { if ((i == 4) && - (gpio->usClkMaskRegisterIndex == 0x1fda) && + (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x1fda) && (gpio->sucI2cId.ucAccess == 0x94)) gpio->sucI2cId.ucAccess = 0x14; } @@ -172,7 +172,7 @@ void radeon_atombios_i2c_init(struct radeon_device *rdev) /* some evergreen boards have bad data for this entry */ if (ASIC_IS_DCE4(rdev)) { if ((i == 7) && - (gpio->usClkMaskRegisterIndex == 0x1936) && + (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x1936) && (gpio->sucI2cId.ucAccess == 0)) { gpio->sucI2cId.ucAccess = 0x97; gpio->ucDataMaskShift = 8; @@ -185,7 +185,7 @@ void radeon_atombios_i2c_init(struct radeon_device *rdev) /* some DCE3 boards have bad data for this entry */ if (ASIC_IS_DCE3(rdev)) { if ((i == 4) && - (gpio->usClkMaskRegisterIndex == 0x1fda) && + (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x1fda) && (gpio->sucI2cId.ucAccess == 0x94)) gpio->sucI2cId.ucAccess = 0x14; } @@ -252,7 +252,7 @@ static inline struct radeon_gpio_rec radeon_lookup_gpio(struct radeon_device *rd pin = &gpio_info->asGPIO_Pin[i]; if (id == pin->ucGPIO_ID) { gpio.id = pin->ucGPIO_ID; - gpio.reg = pin->usGpioPin_AIndex * 4; + gpio.reg = le16_to_cpu(pin->usGpioPin_AIndex) * 4; gpio.mask = (1 << pin->ucGpioPinBitShift); gpio.valid = true; break; @@ -1274,11 +1274,11 @@ bool radeon_atombios_sideport_present(struct radeon_device *rdev) data_offset); switch (crev) { case 1: - if (igp_info->info.ulBootUpMemoryClock) + if (le32_to_cpu(igp_info->info.ulBootUpMemoryClock)) return true; break; case 2: - if (igp_info->info_2.ulBootUpSidePortClock) + if (le32_to_cpu(igp_info->info_2.ulBootUpSidePortClock)) return true; break; default: @@ -1442,7 +1442,7 @@ bool radeon_atombios_get_asic_ss_info(struct radeon_device *rdev, for (i = 0; i < num_indices; i++) { if ((ss_info->info.asSpreadSpectrum[i].ucClockIndication == id) && - (clock <= ss_info->info.asSpreadSpectrum[i].ulTargetClockRange)) { + (clock <= le32_to_cpu(ss_info->info.asSpreadSpectrum[i].ulTargetClockRange))) { ss->percentage = le16_to_cpu(ss_info->info.asSpreadSpectrum[i].usSpreadSpectrumPercentage); ss->type = ss_info->info.asSpreadSpectrum[i].ucSpreadSpectrumMode; @@ -1456,7 +1456,7 @@ bool radeon_atombios_get_asic_ss_info(struct radeon_device *rdev, sizeof(ATOM_ASIC_SS_ASSIGNMENT_V2); for (i = 0; i < num_indices; i++) { if ((ss_info->info_2.asSpreadSpectrum[i].ucClockIndication == id) && - (clock <= ss_info->info_2.asSpreadSpectrum[i].ulTargetClockRange)) { + (clock <= le32_to_cpu(ss_info->info_2.asSpreadSpectrum[i].ulTargetClockRange))) { ss->percentage = le16_to_cpu(ss_info->info_2.asSpreadSpectrum[i].usSpreadSpectrumPercentage); ss->type = ss_info->info_2.asSpreadSpectrum[i].ucSpreadSpectrumMode; @@ -1470,7 +1470,7 @@ bool radeon_atombios_get_asic_ss_info(struct radeon_device *rdev, sizeof(ATOM_ASIC_SS_ASSIGNMENT_V3); for (i = 0; i < num_indices; i++) { if ((ss_info->info_3.asSpreadSpectrum[i].ucClockIndication == id) && - (clock <= ss_info->info_3.asSpreadSpectrum[i].ulTargetClockRange)) { + (clock <= le32_to_cpu(ss_info->info_3.asSpreadSpectrum[i].ulTargetClockRange))) { ss->percentage = le16_to_cpu(ss_info->info_3.asSpreadSpectrum[i].usSpreadSpectrumPercentage); ss->type = ss_info->info_3.asSpreadSpectrum[i].ucSpreadSpectrumMode; @@ -1553,8 +1553,8 @@ struct radeon_encoder_atom_dig *radeon_atombios_get_lvds_info(struct if (misc & ATOM_DOUBLE_CLOCK_MODE) lvds->native_mode.flags |= DRM_MODE_FLAG_DBLSCAN; - lvds->native_mode.width_mm = lvds_info->info.sLCDTiming.usImageHSize; - lvds->native_mode.height_mm = lvds_info->info.sLCDTiming.usImageVSize; + lvds->native_mode.width_mm = le16_to_cpu(lvds_info->info.sLCDTiming.usImageHSize); + lvds->native_mode.height_mm = le16_to_cpu(lvds_info->info.sLCDTiming.usImageVSize); /* set crtc values */ drm_mode_set_crtcinfo(&lvds->native_mode, CRTC_INTERLACE_HALVE_V); @@ -1569,13 +1569,13 @@ struct radeon_encoder_atom_dig *radeon_atombios_get_lvds_info(struct lvds->linkb = false; /* parse the lcd record table */ - if (lvds_info->info.usModePatchTableOffset) { + if (le16_to_cpu(lvds_info->info.usModePatchTableOffset)) { ATOM_FAKE_EDID_PATCH_RECORD *fake_edid_record; ATOM_PANEL_RESOLUTION_PATCH_RECORD *panel_res_record; bool bad_record = false; u8 *record = (u8 *)(mode_info->atom_context->bios + data_offset + - lvds_info->info.usModePatchTableOffset); + le16_to_cpu(lvds_info->info.usModePatchTableOffset)); while (*record != ATOM_RECORD_END_TYPE) { switch (*record) { case LCD_MODE_PATCH_RECORD_MODE_TYPE: @@ -2189,7 +2189,7 @@ static u16 radeon_atombios_get_default_vddc(struct radeon_device *rdev) firmware_info = (union firmware_info *)(mode_info->atom_context->bios + data_offset); - vddc = firmware_info->info_14.usBootUpVDDCVoltage; + vddc = le16_to_cpu(firmware_info->info_14.usBootUpVDDCVoltage); } return vddc; @@ -2284,7 +2284,7 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev, rdev->pm.power_state[state_index].clock_info[mode_index].voltage.type = VOLTAGE_SW; rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage = - clock_info->evergreen.usVDDC; + le16_to_cpu(clock_info->evergreen.usVDDC); } else { sclk = le16_to_cpu(clock_info->r600.usEngineClockLow); sclk |= clock_info->r600.ucEngineClockHigh << 16; @@ -2295,7 +2295,7 @@ static bool radeon_atombios_parse_pplib_clock_info(struct radeon_device *rdev, rdev->pm.power_state[state_index].clock_info[mode_index].voltage.type = VOLTAGE_SW; rdev->pm.power_state[state_index].clock_info[mode_index].voltage.voltage = - clock_info->r600.usVDDC; + le16_to_cpu(clock_info->r600.usVDDC); } if (rdev->flags & RADEON_IS_IGP) { @@ -2408,13 +2408,13 @@ static int radeon_atombios_parse_power_table_6(struct radeon_device *rdev) radeon_atombios_add_pplib_thermal_controller(rdev, &power_info->pplib.sThermalController); state_array = (struct StateArray *) (mode_info->atom_context->bios + data_offset + - power_info->pplib.usStateArrayOffset); + le16_to_cpu(power_info->pplib.usStateArrayOffset)); clock_info_array = (struct ClockInfoArray *) (mode_info->atom_context->bios + data_offset + - power_info->pplib.usClockInfoArrayOffset); + le16_to_cpu(power_info->pplib.usClockInfoArrayOffset)); non_clock_info_array = (struct NonClockInfoArray *) (mode_info->atom_context->bios + data_offset + - power_info->pplib.usNonClockInfoArrayOffset); + le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset)); rdev->pm.power_state = kzalloc(sizeof(struct radeon_power_state) * state_array->ucNumEntries, GFP_KERNEL); if (!rdev->pm.power_state) @@ -2533,7 +2533,7 @@ uint32_t radeon_atom_get_engine_clock(struct radeon_device *rdev) int index = GetIndexIntoMasterTable(COMMAND, GetEngineClock); atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); - return args.ulReturnEngineClock; + return le32_to_cpu(args.ulReturnEngineClock); } uint32_t radeon_atom_get_memory_clock(struct radeon_device *rdev) @@ -2542,7 +2542,7 @@ uint32_t radeon_atom_get_memory_clock(struct radeon_device *rdev) int index = GetIndexIntoMasterTable(COMMAND, GetMemoryClock); atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); - return args.ulReturnMemoryClock; + return le32_to_cpu(args.ulReturnMemoryClock); } void radeon_atom_set_engine_clock(struct radeon_device *rdev, @@ -2551,7 +2551,7 @@ void radeon_atom_set_engine_clock(struct radeon_device *rdev, SET_ENGINE_CLOCK_PS_ALLOCATION args; int index = GetIndexIntoMasterTable(COMMAND, SetEngineClock); - args.ulTargetEngineClock = eng_clock; /* 10 khz */ + args.ulTargetEngineClock = cpu_to_le32(eng_clock); /* 10 khz */ atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } @@ -2565,7 +2565,7 @@ void radeon_atom_set_memory_clock(struct radeon_device *rdev, if (rdev->flags & RADEON_IS_IGP) return; - args.ulTargetMemoryClock = mem_clock; /* 10 khz */ + args.ulTargetMemoryClock = cpu_to_le32(mem_clock); /* 10 khz */ atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args); } diff --git a/drivers/gpu/drm/radeon/radeon_encoders.c b/drivers/gpu/drm/radeon/radeon_encoders.c index 5b38b73ccd12..b4274883227f 100644 --- a/drivers/gpu/drm/radeon/radeon_encoders.c +++ b/drivers/gpu/drm/radeon/radeon_encoders.c @@ -910,7 +910,7 @@ atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t args.v1.ucAction = action; if (action == ATOM_TRANSMITTER_ACTION_INIT) { - args.v1.usInitInfo = connector_object_id; + args.v1.usInitInfo = cpu_to_le16(connector_object_id); } else if (action == ATOM_TRANSMITTER_ACTION_SETUP_VSEMPH) { args.v1.asMode.ucLaneSel = lane_num; args.v1.asMode.ucLaneSet = lane_set; @@ -1140,7 +1140,7 @@ atombios_external_encoder_setup(struct drm_encoder *encoder, case 3: args.v3.sExtEncoder.ucAction = action; if (action == EXTERNAL_ENCODER_ACTION_V3_ENCODER_INIT) - args.v3.sExtEncoder.usConnectorId = connector_object_id; + args.v3.sExtEncoder.usConnectorId = cpu_to_le16(connector_object_id); else args.v3.sExtEncoder.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10); args.v3.sExtEncoder.ucEncoderMode = atombios_get_encoder_mode(encoder); -- cgit v1.2.3 From 4eace7fdfa1f8ac2f0a833e12bd07eeb453ec9ef Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?C=C3=A9dric=20Cano?= Date: Fri, 11 Feb 2011 19:45:38 -0500 Subject: drm/radeon/kms: 6xx/7xx big endian fixes MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit agd5f: minor cleanups Signed-off-by: Cédric Cano Signed-off-by: Alex Deucher Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/r600.c | 22 +++++++++++++++++----- drivers/gpu/drm/radeon/r600_blit_kms.c | 25 +++++++++++++++++++------ drivers/gpu/drm/radeon/r600_blit_shaders.c | 4 ++++ drivers/gpu/drm/radeon/r600d.h | 9 +++++---- drivers/gpu/drm/radeon/rv770.c | 6 +++++- drivers/gpu/drm/radeon/rv770d.h | 8 ++++---- 6 files changed, 54 insertions(+), 20 deletions(-) diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c index 650672a0f5ad..de88624d5f87 100644 --- a/drivers/gpu/drm/radeon/r600.c +++ b/drivers/gpu/drm/radeon/r600.c @@ -2105,7 +2105,11 @@ static int r600_cp_load_microcode(struct radeon_device *rdev) r600_cp_stop(rdev); - WREG32(CP_RB_CNTL, RB_NO_UPDATE | RB_BLKSZ(15) | RB_BUFSZ(3)); + WREG32(CP_RB_CNTL, +#ifdef __BIG_ENDIAN + BUF_SWAP_32BIT | +#endif + RB_NO_UPDATE | RB_BLKSZ(15) | RB_BUFSZ(3)); /* Reset cp */ WREG32(GRBM_SOFT_RESET, SOFT_RESET_CP); @@ -2192,7 +2196,11 @@ int r600_cp_resume(struct radeon_device *rdev) WREG32(CP_RB_WPTR, 0); /* set the wb address whether it's enabled or not */ - WREG32(CP_RB_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC); + WREG32(CP_RB_RPTR_ADDR, +#ifdef __BIG_ENDIAN + RB_RPTR_SWAP(2) | +#endif + ((rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC)); WREG32(CP_RB_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFF); WREG32(SCRATCH_ADDR, ((rdev->wb.gpu_addr + RADEON_WB_SCRATCH_OFFSET) >> 8) & 0xFFFFFFFF); @@ -2628,7 +2636,11 @@ void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) { /* FIXME: implement */ radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); - radeon_ring_write(rdev, ib->gpu_addr & 0xFFFFFFFC); + radeon_ring_write(rdev, +#ifdef __BIG_ENDIAN + (2 << 0) | +#endif + (ib->gpu_addr & 0xFFFFFFFC)); radeon_ring_write(rdev, upper_32_bits(ib->gpu_addr) & 0xFF); radeon_ring_write(rdev, ib->length_dw); } @@ -3297,8 +3309,8 @@ restart_ih: while (rptr != wptr) { /* wptr/rptr are in bytes! */ ring_index = rptr / 4; - src_id = rdev->ih.ring[ring_index] & 0xff; - src_data = rdev->ih.ring[ring_index + 1] & 0xfffffff; + src_id = le32_to_cpu(rdev->ih.ring[ring_index]) & 0xff; + src_data = le32_to_cpu(rdev->ih.ring[ring_index + 1]) & 0xfffffff; switch (src_id) { case 1: /* D1 vblank/vline */ diff --git a/drivers/gpu/drm/radeon/r600_blit_kms.c b/drivers/gpu/drm/radeon/r600_blit_kms.c index 86e5aa07f0db..69d94dd4db21 100644 --- a/drivers/gpu/drm/radeon/r600_blit_kms.c +++ b/drivers/gpu/drm/radeon/r600_blit_kms.c @@ -165,6 +165,9 @@ set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr) u32 sq_vtx_constant_word2; sq_vtx_constant_word2 = ((upper_32_bits(gpu_addr) & 0xff) | (16 << 8)); +#ifdef __BIG_ENDIAN + sq_vtx_constant_word2 |= (2 << 30); +#endif radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 7)); radeon_ring_write(rdev, 0x460); @@ -253,7 +256,11 @@ draw_auto(struct radeon_device *rdev) radeon_ring_write(rdev, DI_PT_RECTLIST); radeon_ring_write(rdev, PACKET3(PACKET3_INDEX_TYPE, 0)); - radeon_ring_write(rdev, DI_INDEX_SIZE_16_BIT); + radeon_ring_write(rdev, +#ifdef __BIG_ENDIAN + (2 << 2) | +#endif + DI_INDEX_SIZE_16_BIT); radeon_ring_write(rdev, PACKET3(PACKET3_NUM_INSTANCES, 0)); radeon_ring_write(rdev, 1); @@ -424,7 +431,11 @@ set_default_state(struct radeon_device *rdev) dwords = ALIGN(rdev->r600_blit.state_len, 0x10); gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.state_offset; radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); - radeon_ring_write(rdev, gpu_addr & 0xFFFFFFFC); + radeon_ring_write(rdev, +#ifdef __BIG_ENDIAN + (2 << 0) | +#endif + (gpu_addr & 0xFFFFFFFC)); radeon_ring_write(rdev, upper_32_bits(gpu_addr) & 0xFF); radeon_ring_write(rdev, dwords); @@ -467,7 +478,7 @@ static inline uint32_t i2f(uint32_t input) int r600_blit_init(struct radeon_device *rdev) { u32 obj_size; - int r, dwords; + int i, r, dwords; void *ptr; u32 packet2s[16]; int num_packet2s = 0; @@ -486,7 +497,7 @@ int r600_blit_init(struct radeon_device *rdev) dwords = rdev->r600_blit.state_len; while (dwords & 0xf) { - packet2s[num_packet2s++] = PACKET2(0); + packet2s[num_packet2s++] = cpu_to_le32(PACKET2(0)); dwords++; } @@ -529,8 +540,10 @@ int r600_blit_init(struct radeon_device *rdev) if (num_packet2s) memcpy_toio(ptr + rdev->r600_blit.state_offset + (rdev->r600_blit.state_len * 4), packet2s, num_packet2s * 4); - memcpy(ptr + rdev->r600_blit.vs_offset, r6xx_vs, r6xx_vs_size * 4); - memcpy(ptr + rdev->r600_blit.ps_offset, r6xx_ps, r6xx_ps_size * 4); + for (i = 0; i < r6xx_vs_size; i++) + *(u32 *)((unsigned long)ptr + rdev->r600_blit.vs_offset + i * 4) = cpu_to_le32(r6xx_vs[i]); + for (i = 0; i < r6xx_ps_size; i++) + *(u32 *)((unsigned long)ptr + rdev->r600_blit.ps_offset + i * 4) = cpu_to_le32(r6xx_ps[i]); radeon_bo_kunmap(rdev->r600_blit.shader_obj); radeon_bo_unreserve(rdev->r600_blit.shader_obj); diff --git a/drivers/gpu/drm/radeon/r600_blit_shaders.c b/drivers/gpu/drm/radeon/r600_blit_shaders.c index e8151c1d55b2..2d1f6c5ee2a7 100644 --- a/drivers/gpu/drm/radeon/r600_blit_shaders.c +++ b/drivers/gpu/drm/radeon/r600_blit_shaders.c @@ -684,7 +684,11 @@ const u32 r6xx_vs[] = 0x00000000, 0x3c000000, 0x68cd1000, +#ifdef __BIG_ENDIAN + 0x000a0000, +#else 0x00080000, +#endif 0x00000000, }; diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h index a5d898b4bad2..04bac0bbd3ec 100644 --- a/drivers/gpu/drm/radeon/r600d.h +++ b/drivers/gpu/drm/radeon/r600d.h @@ -154,13 +154,14 @@ #define ROQ_IB2_START(x) ((x) << 8) #define CP_RB_BASE 0xC100 #define CP_RB_CNTL 0xC104 -#define RB_BUFSZ(x) ((x)<<0) -#define RB_BLKSZ(x) ((x)<<8) -#define RB_NO_UPDATE (1<<27) -#define RB_RPTR_WR_ENA (1<<31) +#define RB_BUFSZ(x) ((x) << 0) +#define RB_BLKSZ(x) ((x) << 8) +#define RB_NO_UPDATE (1 << 27) +#define RB_RPTR_WR_ENA (1 << 31) #define BUF_SWAP_32BIT (2 << 16) #define CP_RB_RPTR 0x8700 #define CP_RB_RPTR_ADDR 0xC10C +#define RB_RPTR_SWAP(x) ((x) << 0) #define CP_RB_RPTR_ADDR_HI 0xC110 #define CP_RB_RPTR_WR 0xC108 #define CP_RB_WPTR 0xC114 diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c index 2211a323db41..d8ba67690656 100644 --- a/drivers/gpu/drm/radeon/rv770.c +++ b/drivers/gpu/drm/radeon/rv770.c @@ -321,7 +321,11 @@ static int rv770_cp_load_microcode(struct radeon_device *rdev) return -EINVAL; r700_cp_stop(rdev); - WREG32(CP_RB_CNTL, RB_NO_UPDATE | (15 << 8) | (3 << 0)); + WREG32(CP_RB_CNTL, +#ifdef __BIG_ENDIAN + BUF_SWAP_32BIT | +#endif + RB_NO_UPDATE | RB_BLKSZ(15) | RB_BUFSZ(3)); /* Reset cp */ WREG32(GRBM_SOFT_RESET, SOFT_RESET_CP); diff --git a/drivers/gpu/drm/radeon/rv770d.h b/drivers/gpu/drm/radeon/rv770d.h index abc8cf5a3672..79fa588e9ed5 100644 --- a/drivers/gpu/drm/radeon/rv770d.h +++ b/drivers/gpu/drm/radeon/rv770d.h @@ -76,10 +76,10 @@ #define ROQ_IB1_START(x) ((x) << 0) #define ROQ_IB2_START(x) ((x) << 8) #define CP_RB_CNTL 0xC104 -#define RB_BUFSZ(x) ((x)<<0) -#define RB_BLKSZ(x) ((x)<<8) -#define RB_NO_UPDATE (1<<27) -#define RB_RPTR_WR_ENA (1<<31) +#define RB_BUFSZ(x) ((x) << 0) +#define RB_BLKSZ(x) ((x) << 8) +#define RB_NO_UPDATE (1 << 27) +#define RB_RPTR_WR_ENA (1 << 31) #define BUF_SWAP_32BIT (2 << 16) #define CP_RB_RPTR 0x8700 #define CP_RB_RPTR_ADDR 0xC10C -- cgit v1.2.3 From 0f234f5fdca1e31c7a6333c3633edc653cf3e598 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Sun, 13 Feb 2011 19:06:33 -0500 Subject: drm/radeon/kms: evergreen/ni big endian fixes (v2) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Based on 6xx/7xx endian fixes from Cédric Cano. v2: fix typo in shader Signed-off-by: Alex Deucher Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/evergreen.c | 22 +++++++++++++++++----- drivers/gpu/drm/radeon/evergreen_blit_kms.c | 19 ++++++++++++++----- drivers/gpu/drm/radeon/evergreen_blit_shaders.c | 8 ++++++++ drivers/gpu/drm/radeon/evergreend.h | 1 + 4 files changed, 40 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c index ffdc8332b76e..d270b3ff896b 100644 --- a/drivers/gpu/drm/radeon/evergreen.c +++ b/drivers/gpu/drm/radeon/evergreen.c @@ -1192,7 +1192,11 @@ void evergreen_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib) radeon_ring_write(rdev, 1); /* FIXME: implement */ radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2)); - radeon_ring_write(rdev, ib->gpu_addr & 0xFFFFFFFC); + radeon_ring_write(rdev, +#ifdef __BIG_ENDIAN + (2 << 0) | +#endif + (ib->gpu_addr & 0xFFFFFFFC)); radeon_ring_write(rdev, upper_32_bits(ib->gpu_addr) & 0xFF); radeon_ring_write(rdev, ib->length_dw); } @@ -1207,7 +1211,11 @@ static int evergreen_cp_load_microcode(struct radeon_device *rdev) return -EINVAL; r700_cp_stop(rdev); - WREG32(CP_RB_CNTL, RB_NO_UPDATE | (15 << 8) | (3 << 0)); + WREG32(CP_RB_CNTL, +#ifdef __BIG_ENDIAN + BUF_SWAP_32BIT | +#endif + RB_NO_UPDATE | RB_BLKSZ(15) | RB_BUFSZ(3)); fw_data = (const __be32 *)rdev->pfp_fw->data; WREG32(CP_PFP_UCODE_ADDR, 0); @@ -1326,7 +1334,11 @@ int evergreen_cp_resume(struct radeon_device *rdev) WREG32(CP_RB_WPTR, 0); /* set the wb address wether it's enabled or not */ - WREG32(CP_RB_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC); + WREG32(CP_RB_RPTR_ADDR, +#ifdef __BIG_ENDIAN + RB_RPTR_SWAP(2) | +#endif + ((rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC)); WREG32(CP_RB_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFF); WREG32(SCRATCH_ADDR, ((rdev->wb.gpu_addr + RADEON_WB_SCRATCH_OFFSET) >> 8) & 0xFFFFFFFF); @@ -2627,8 +2639,8 @@ restart_ih: while (rptr != wptr) { /* wptr/rptr are in bytes! */ ring_index = rptr / 4; - src_id = rdev->ih.ring[ring_index] & 0xff; - src_data = rdev->ih.ring[ring_index + 1] & 0xfffffff; + src_id = le32_to_cpu(rdev->ih.ring[ring_index]) & 0xff; + src_data = le32_to_cpu(rdev->ih.ring[ring_index + 1]) & 0xfffffff; switch (src_id) { case 1: /* D1 vblank/vline */ diff --git a/drivers/gpu/drm/radeon/evergreen_blit_kms.c b/drivers/gpu/drm/radeon/evergreen_blit_kms.c index a1ba4b3053d0..a7b7a33eaf3a 100644 --- a/drivers/gpu/drm/radeon/evergreen_blit_kms.c +++ b/drivers/gpu/drm/radeon/evergreen_blit_kms.c @@ -133,6 +133,9 @@ set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr) /* high addr, stride */ sq_vtx_constant_word2 = ((upper_32_bits(gpu_addr) & 0xff) | (16 << 8)); +#ifdef __BIG_ENDIAN + sq_vtx_constant_word2 |= (2 << 30); +#endif /* xyzw swizzles */ sq_vtx_constant_word3 = (0 << 3) | (1 << 6) | (2 << 9) | (3 << 12); @@ -221,7 +224,11 @@ draw_auto(struct radeon_device *rdev) radeon_ring_write(rdev, DI_PT_RECTLIST); radeon_ring_write(rdev, PACKET3(PACKET3_INDEX_TYPE, 0)); - radeon_ring_write(rdev, DI_INDEX_SIZE_16_BIT); + radeon_ring_write(rdev, +#ifdef __BIG_ENDIAN + (2 << 2) | +#endif + DI_INDEX_SIZE_16_BIT); radeon_ring_write(rdev, PACKET3(PACKET3_NUM_INSTANCES, 0)); radeon_ring_write(rdev, 1); @@ -541,7 +548,7 @@ static inline uint32_t i2f(uint32_t input) int evergreen_blit_init(struct radeon_device *rdev) { u32 obj_size; - int r, dwords; + int i, r, dwords; void *ptr; u32 packet2s[16]; int num_packet2s = 0; @@ -557,7 +564,7 @@ int evergreen_blit_init(struct radeon_device *rdev) dwords = rdev->r600_blit.state_len; while (dwords & 0xf) { - packet2s[num_packet2s++] = PACKET2(0); + packet2s[num_packet2s++] = cpu_to_le32(PACKET2(0)); dwords++; } @@ -598,8 +605,10 @@ int evergreen_blit_init(struct radeon_device *rdev) if (num_packet2s) memcpy_toio(ptr + rdev->r600_blit.state_offset + (rdev->r600_blit.state_len * 4), packet2s, num_packet2s * 4); - memcpy(ptr + rdev->r600_blit.vs_offset, evergreen_vs, evergreen_vs_size * 4); - memcpy(ptr + rdev->r600_blit.ps_offset, evergreen_ps, evergreen_ps_size * 4); + for (i = 0; i < evergreen_vs_size; i++) + *(u32 *)((unsigned long)ptr + rdev->r600_blit.vs_offset + i * 4) = cpu_to_le32(evergreen_vs[i]); + for (i = 0; i < evergreen_ps_size; i++) + *(u32 *)((unsigned long)ptr + rdev->r600_blit.ps_offset + i * 4) = cpu_to_le32(evergreen_ps[i]); radeon_bo_kunmap(rdev->r600_blit.shader_obj); radeon_bo_unreserve(rdev->r600_blit.shader_obj); diff --git a/drivers/gpu/drm/radeon/evergreen_blit_shaders.c b/drivers/gpu/drm/radeon/evergreen_blit_shaders.c index ef1d28c07fbf..3a10399e0066 100644 --- a/drivers/gpu/drm/radeon/evergreen_blit_shaders.c +++ b/drivers/gpu/drm/radeon/evergreen_blit_shaders.c @@ -311,11 +311,19 @@ const u32 evergreen_vs[] = 0x00000000, 0x3c000000, 0x67961001, +#ifdef __BIG_ENDIAN + 0x000a0000, +#else 0x00080000, +#endif 0x00000000, 0x1c000000, 0x67961000, +#ifdef __BIG_ENDIAN + 0x00020008, +#else 0x00000008, +#endif 0x00000000, }; diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h index afec1aca2a73..eb4acf4528ff 100644 --- a/drivers/gpu/drm/radeon/evergreend.h +++ b/drivers/gpu/drm/radeon/evergreend.h @@ -98,6 +98,7 @@ #define BUF_SWAP_32BIT (2 << 16) #define CP_RB_RPTR 0x8700 #define CP_RB_RPTR_ADDR 0xC10C +#define RB_RPTR_SWAP(x) ((x) << 0) #define CP_RB_RPTR_ADDR_HI 0xC110 #define CP_RB_RPTR_WR 0xC108 #define CP_RB_WPTR 0xC114 -- cgit v1.2.3 From 8fd1b84cc9d32e7e5c44e990a9c9e27504b232ed Mon Sep 17 00:00:00 2001 From: Dave Airlie Date: Thu, 10 Feb 2011 14:46:06 +1000 Subject: drm/radeon: fix race between GPU reset and TTM delayed delete thread. My evergreen has been in a remote PC for week and reset has never once saved me from certain doom, I finally relocated to the box with a serial cable and noticed an oops when the GPU resets, and the TTM delayed delete thread tries to remove something from the GTT. This stops the delayed delete thread from executing across the GPU reset handler, and woot I can GPU reset now. Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/radeon_device.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c index 0d478932b1a9..4954e2d6ffa2 100644 --- a/drivers/gpu/drm/radeon/radeon_device.c +++ b/drivers/gpu/drm/radeon/radeon_device.c @@ -936,8 +936,11 @@ int radeon_resume_kms(struct drm_device *dev) int radeon_gpu_reset(struct radeon_device *rdev) { int r; + int resched; radeon_save_bios_scratch_regs(rdev); + /* block TTM */ + resched = ttm_bo_lock_delayed_workqueue(&rdev->mman.bdev); radeon_suspend(rdev); r = radeon_asic_reset(rdev); @@ -946,6 +949,7 @@ int radeon_gpu_reset(struct radeon_device *rdev) radeon_resume(rdev); radeon_restore_bios_scratch_regs(rdev); drm_helper_resume_force_mode(rdev->ddev); + ttm_bo_unlock_delayed_workqueue(&rdev->mman.bdev, resched); return 0; } /* bad news, how to tell it to userspace ? */ -- cgit v1.2.3 From 1ea9dbf250ae6706400dc0e3d6e1cc7540830731 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Thu, 10 Feb 2011 14:51:33 -0500 Subject: drm/radeon/kms: use linear aligned for 6xx/7xx bo blits Not only is linear aligned supposedly more performant, linear general is only supported by the CB in single slice mode. The texture hardware doesn't support linear general, but I think the hw automatically upgrades it to linear aligned. Signed-off-by: Alex Deucher Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/r600_blit_kms.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/radeon/r600_blit_kms.c b/drivers/gpu/drm/radeon/r600_blit_kms.c index 69d94dd4db21..41f7aafc97c4 100644 --- a/drivers/gpu/drm/radeon/r600_blit_kms.c +++ b/drivers/gpu/drm/radeon/r600_blit_kms.c @@ -54,7 +54,7 @@ set_render_target(struct radeon_device *rdev, int format, if (h < 8) h = 8; - cb_color_info = ((format << 2) | (1 << 27)); + cb_color_info = ((format << 2) | (1 << 27) | (1 << 8)); pitch = (w / 8) - 1; slice = ((w * h) / 64) - 1; @@ -202,7 +202,7 @@ set_tex_resource(struct radeon_device *rdev, if (h < 1) h = 1; - sq_tex_resource_word0 = (1 << 0); + sq_tex_resource_word0 = (1 << 0) | (1 << 3); sq_tex_resource_word0 |= ((((pitch >> 3) - 1) << 8) | ((w - 1) << 19)); -- cgit v1.2.3 From 27dcfc102279867ef0080d3b27e0f8306cac53d1 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Thu, 10 Feb 2011 14:51:34 -0500 Subject: drm/radeon/kms: use linear aligned for evergreen/ni bo blits Not only is linear aligned supposedly more performant, linear general is only supported by the CB in single slice mode. The texture hardware doesn't support linear general, but I think the hw automatically upgrades it to linear aligned. Signed-off-by: Alex Deucher Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/evergreen_blit_kms.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/radeon/evergreen_blit_kms.c b/drivers/gpu/drm/radeon/evergreen_blit_kms.c index a7b7a33eaf3a..2adfb03f479b 100644 --- a/drivers/gpu/drm/radeon/evergreen_blit_kms.c +++ b/drivers/gpu/drm/radeon/evergreen_blit_kms.c @@ -55,7 +55,7 @@ set_render_target(struct radeon_device *rdev, int format, if (h < 8) h = 8; - cb_color_info = ((format << 2) | (1 << 24)); + cb_color_info = ((format << 2) | (1 << 24) | (1 << 8)); pitch = (w / 8) - 1; slice = ((w * h) / 64) - 1; @@ -176,7 +176,7 @@ set_tex_resource(struct radeon_device *rdev, sq_tex_resource_word0 = (1 << 0); /* 2D */ sq_tex_resource_word0 |= ((((pitch >> 3) - 1) << 6) | ((w - 1) << 18)); - sq_tex_resource_word1 = ((h - 1) << 0); + sq_tex_resource_word1 = ((h - 1) << 0) | (1 << 28); /* xyzw swizzles */ sq_tex_resource_word4 = (0 << 16) | (1 << 19) | (2 << 22) | (3 << 25); -- cgit v1.2.3 From 501834349e872ed4115eea3beef65ca9eeb5528e Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marek=20Ol=C5=A1=C3=A1k?= Date: Mon, 14 Feb 2011 01:01:09 +0100 Subject: drm/radeon/kms: fix tracking of BLENDCNTL, COLOR_CHANNEL_MASK, and GB_Z on r300 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Also move ZB_DEPTHCLEARVALUE to the list of safe regs. Signed-off-by: Marek Olšák Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/r300.c | 3 +-- drivers/gpu/drm/radeon/reg_srcs/r300 | 3 +-- drivers/gpu/drm/radeon/reg_srcs/r420 | 4 +--- drivers/gpu/drm/radeon/reg_srcs/rs600 | 3 +-- drivers/gpu/drm/radeon/reg_srcs/rv515 | 4 +--- 5 files changed, 5 insertions(+), 12 deletions(-) diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c index 15f94648f274..862b61742b82 100644 --- a/drivers/gpu/drm/radeon/r300.c +++ b/drivers/gpu/drm/radeon/r300.c @@ -873,6 +873,7 @@ static int r300_packet0_check(struct radeon_cs_parser *p, track->zb_dirty = true; break; case 0x4104: + /* TX_ENABLE */ for (i = 0; i < 16; i++) { bool enabled; @@ -1103,8 +1104,6 @@ static int r300_packet0_check(struct radeon_cs_parser *p, track->blend_read_enable = !!(idx_value & (1 << 2)); track->cb_dirty = true; break; - case 0x4f28: /* ZB_DEPTHCLEARVALUE */ - break; case 0x4f30: /* ZB_MASK_OFFSET */ case 0x4f34: /* ZB_ZMASK_PITCH */ case 0x4f44: /* ZB_HIZ_OFFSET */ diff --git a/drivers/gpu/drm/radeon/reg_srcs/r300 b/drivers/gpu/drm/radeon/reg_srcs/r300 index b506ec1cab4b..13a94e2ee03b 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/r300 +++ b/drivers/gpu/drm/radeon/reg_srcs/r300 @@ -683,9 +683,7 @@ r300 0x4f60 0x4DF4 US_ALU_CONST_G_31 0x4DF8 US_ALU_CONST_B_31 0x4DFC US_ALU_CONST_A_31 -0x4E04 RB3D_BLENDCNTL_R3 0x4E08 RB3D_ABLENDCNTL_R3 -0x4E0C RB3D_COLOR_CHANNEL_MASK 0x4E10 RB3D_CONSTANT_COLOR 0x4E14 RB3D_COLOR_CLEAR_VALUE 0x4E18 RB3D_ROPCNTL_R3 @@ -715,4 +713,5 @@ r300 0x4f60 0x4F08 ZB_STENCILREFMASK 0x4F14 ZB_ZTOP 0x4F18 ZB_ZCACHE_CTLSTAT +0x4F28 ZB_DEPTHCLEARVALUE 0x4F58 ZB_ZPASS_DATA diff --git a/drivers/gpu/drm/radeon/reg_srcs/r420 b/drivers/gpu/drm/radeon/reg_srcs/r420 index 8c1214c2390f..5c95cf87f7f2 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/r420 +++ b/drivers/gpu/drm/radeon/reg_srcs/r420 @@ -130,7 +130,6 @@ r420 0x4f60 0x401C GB_SELECT 0x4020 GB_AA_CONFIG 0x4024 GB_FIFO_SIZE -0x4028 GB_Z_PEQ_CONFIG 0x4100 TX_INVALTAGS 0x4200 GA_POINT_S0 0x4204 GA_POINT_T0 @@ -750,9 +749,7 @@ r420 0x4f60 0x4DF4 US_ALU_CONST_G_31 0x4DF8 US_ALU_CONST_B_31 0x4DFC US_ALU_CONST_A_31 -0x4E04 RB3D_BLENDCNTL_R3 0x4E08 RB3D_ABLENDCNTL_R3 -0x4E0C RB3D_COLOR_CHANNEL_MASK 0x4E10 RB3D_CONSTANT_COLOR 0x4E14 RB3D_COLOR_CLEAR_VALUE 0x4E18 RB3D_ROPCNTL_R3 @@ -782,4 +779,5 @@ r420 0x4f60 0x4F08 ZB_STENCILREFMASK 0x4F14 ZB_ZTOP 0x4F18 ZB_ZCACHE_CTLSTAT +0x4F28 ZB_DEPTHCLEARVALUE 0x4F58 ZB_ZPASS_DATA diff --git a/drivers/gpu/drm/radeon/reg_srcs/rs600 b/drivers/gpu/drm/radeon/reg_srcs/rs600 index 0828d80396f2..263109c1d0c8 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/rs600 +++ b/drivers/gpu/drm/radeon/reg_srcs/rs600 @@ -749,9 +749,7 @@ rs600 0x6d40 0x4DF4 US_ALU_CONST_G_31 0x4DF8 US_ALU_CONST_B_31 0x4DFC US_ALU_CONST_A_31 -0x4E04 RB3D_BLENDCNTL_R3 0x4E08 RB3D_ABLENDCNTL_R3 -0x4E0C RB3D_COLOR_CHANNEL_MASK 0x4E10 RB3D_CONSTANT_COLOR 0x4E14 RB3D_COLOR_CLEAR_VALUE 0x4E18 RB3D_ROPCNTL_R3 @@ -781,4 +779,5 @@ rs600 0x6d40 0x4F08 ZB_STENCILREFMASK 0x4F14 ZB_ZTOP 0x4F18 ZB_ZCACHE_CTLSTAT +0x4F28 ZB_DEPTHCLEARVALUE 0x4F58 ZB_ZPASS_DATA diff --git a/drivers/gpu/drm/radeon/reg_srcs/rv515 b/drivers/gpu/drm/radeon/reg_srcs/rv515 index ef422bbacfc1..eeed003f14c7 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/rv515 +++ b/drivers/gpu/drm/radeon/reg_srcs/rv515 @@ -164,7 +164,6 @@ rv515 0x6d40 0x401C GB_SELECT 0x4020 GB_AA_CONFIG 0x4024 GB_FIFO_SIZE -0x4028 GB_Z_PEQ_CONFIG 0x4100 TX_INVALTAGS 0x4114 SU_TEX_WRAP_PS3 0x4118 PS3_ENABLE @@ -461,9 +460,7 @@ rv515 0x6d40 0x4DF4 US_ALU_CONST_G_31 0x4DF8 US_ALU_CONST_B_31 0x4DFC US_ALU_CONST_A_31 -0x4E04 RB3D_BLENDCNTL_R3 0x4E08 RB3D_ABLENDCNTL_R3 -0x4E0C RB3D_COLOR_CHANNEL_MASK 0x4E10 RB3D_CONSTANT_COLOR 0x4E14 RB3D_COLOR_CLEAR_VALUE 0x4E18 RB3D_ROPCNTL_R3 @@ -496,4 +493,5 @@ rv515 0x6d40 0x4F14 ZB_ZTOP 0x4F18 ZB_ZCACHE_CTLSTAT 0x4F58 ZB_ZPASS_DATA +0x4F28 ZB_DEPTHCLEARVALUE 0x4FD4 ZB_STENCILREFMASK_BF -- cgit v1.2.3 From fff1ce4dc6113b6fdc4e3a815ca5fd229408f8ef Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marek=20Ol=C5=A1=C3=A1k?= Date: Mon, 14 Feb 2011 01:01:10 +0100 Subject: drm/radeon/kms: check AA resolve registers on r300 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This is an important security fix because we allowed arbitrary values to be passed to AARESOLVE_OFFSET. This also puts the right buffer address in the register. Signed-off-by: Marek Olšák Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/r100.c | 23 +++++++++++++++++++++++ drivers/gpu/drm/radeon/r100_track.h | 4 +++- drivers/gpu/drm/radeon/r300.c | 21 +++++++++++++++++++++ drivers/gpu/drm/radeon/r300_reg.h | 2 ++ drivers/gpu/drm/radeon/reg_srcs/r300 | 3 --- drivers/gpu/drm/radeon/reg_srcs/r420 | 3 --- drivers/gpu/drm/radeon/reg_srcs/rs600 | 3 --- drivers/gpu/drm/radeon/reg_srcs/rv515 | 3 --- 8 files changed, 49 insertions(+), 13 deletions(-) diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index fdf4bc67ae58..56deae5bf02e 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -3381,6 +3381,26 @@ int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track) } track->zb_dirty = false; + if (track->aa_dirty && track->aaresolve) { + if (track->aa.robj == NULL) { + DRM_ERROR("[drm] No buffer for AA resolve buffer %d !\n", i); + return -EINVAL; + } + /* I believe the format comes from colorbuffer0. */ + size = track->aa.pitch * track->cb[0].cpp * track->maxy; + size += track->aa.offset; + if (size > radeon_bo_size(track->aa.robj)) { + DRM_ERROR("[drm] Buffer too small for AA resolve buffer %d " + "(need %lu have %lu) !\n", i, size, + radeon_bo_size(track->aa.robj)); + DRM_ERROR("[drm] AA resolve buffer %d (%u %u %u %u)\n", + i, track->aa.pitch, track->cb[0].cpp, + track->aa.offset, track->maxy); + return -EINVAL; + } + } + track->aa_dirty = false; + prim_walk = (track->vap_vf_cntl >> 4) & 0x3; if (track->vap_vf_cntl & (1 << 14)) { nverts = track->vap_alt_nverts; @@ -3455,6 +3475,7 @@ void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track track->cb_dirty = true; track->zb_dirty = true; track->tex_dirty = true; + track->aa_dirty = true; if (rdev->family < CHIP_R300) { track->num_cb = 1; @@ -3469,6 +3490,8 @@ void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track track->num_texture = 16; track->maxy = 4096; track->separate_cube = 0; + track->aaresolve = true; + track->aa.robj = NULL; } for (i = 0; i < track->num_cb; i++) { diff --git a/drivers/gpu/drm/radeon/r100_track.h b/drivers/gpu/drm/radeon/r100_track.h index ee85c4a1fc08..2fef9de7f363 100644 --- a/drivers/gpu/drm/radeon/r100_track.h +++ b/drivers/gpu/drm/radeon/r100_track.h @@ -66,15 +66,17 @@ struct r100_cs_track { struct r100_cs_track_array arrays[11]; struct r100_cs_track_cb cb[R300_MAX_CB]; struct r100_cs_track_cb zb; + struct r100_cs_track_cb aa; struct r100_cs_track_texture textures[R300_TRACK_MAX_TEXTURE]; bool z_enabled; bool separate_cube; bool zb_cb_clear; bool blend_read_enable; - bool cb_dirty; bool zb_dirty; bool tex_dirty; + bool aa_dirty; + bool aaresolve; }; int r100_cs_track_check(struct radeon_device *rdev, struct r100_cs_track *track); diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c index 862b61742b82..768c60ee4ab6 100644 --- a/drivers/gpu/drm/radeon/r300.c +++ b/drivers/gpu/drm/radeon/r300.c @@ -1104,6 +1104,27 @@ static int r300_packet0_check(struct radeon_cs_parser *p, track->blend_read_enable = !!(idx_value & (1 << 2)); track->cb_dirty = true; break; + case R300_RB3D_AARESOLVE_OFFSET: + r = r100_cs_packet_next_reloc(p, &reloc); + if (r) { + DRM_ERROR("No reloc for ib[%d]=0x%04X\n", + idx, reg); + r100_cs_dump_packet(p, pkt); + return r; + } + track->aa.robj = reloc->robj; + track->aa.offset = idx_value; + track->aa_dirty = true; + ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset); + break; + case R300_RB3D_AARESOLVE_PITCH: + track->aa.pitch = idx_value & 0x3FFE; + track->aa_dirty = true; + break; + case R300_RB3D_AARESOLVE_CTL: + track->aaresolve = idx_value & 0x1; + track->aa_dirty = true; + break; case 0x4f30: /* ZB_MASK_OFFSET */ case 0x4f34: /* ZB_ZMASK_PITCH */ case 0x4f44: /* ZB_HIZ_OFFSET */ diff --git a/drivers/gpu/drm/radeon/r300_reg.h b/drivers/gpu/drm/radeon/r300_reg.h index 1a0d5362cd79..f0bce399c9f3 100644 --- a/drivers/gpu/drm/radeon/r300_reg.h +++ b/drivers/gpu/drm/radeon/r300_reg.h @@ -1371,6 +1371,8 @@ #define R300_RB3D_COLORPITCH2 0x4E40 /* GUESS */ #define R300_RB3D_COLORPITCH3 0x4E44 /* GUESS */ +#define R300_RB3D_AARESOLVE_OFFSET 0x4E80 +#define R300_RB3D_AARESOLVE_PITCH 0x4E84 #define R300_RB3D_AARESOLVE_CTL 0x4E88 /* gap */ diff --git a/drivers/gpu/drm/radeon/reg_srcs/r300 b/drivers/gpu/drm/radeon/reg_srcs/r300 index 13a94e2ee03b..e8a1786b6426 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/r300 +++ b/drivers/gpu/drm/radeon/reg_srcs/r300 @@ -704,9 +704,6 @@ r300 0x4f60 0x4E74 RB3D_CMASK_WRINDEX 0x4E78 RB3D_CMASK_DWORD 0x4E7C RB3D_CMASK_RDINDEX -0x4E80 RB3D_AARESOLVE_OFFSET -0x4E84 RB3D_AARESOLVE_PITCH -0x4E88 RB3D_AARESOLVE_CTL 0x4EA0 RB3D_DISCARD_SRC_PIXEL_LTE_THRESHOLD 0x4EA4 RB3D_DISCARD_SRC_PIXEL_GTE_THRESHOLD 0x4F04 ZB_ZSTENCILCNTL diff --git a/drivers/gpu/drm/radeon/reg_srcs/r420 b/drivers/gpu/drm/radeon/reg_srcs/r420 index 5c95cf87f7f2..722074e21e2f 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/r420 +++ b/drivers/gpu/drm/radeon/reg_srcs/r420 @@ -770,9 +770,6 @@ r420 0x4f60 0x4E74 RB3D_CMASK_WRINDEX 0x4E78 RB3D_CMASK_DWORD 0x4E7C RB3D_CMASK_RDINDEX -0x4E80 RB3D_AARESOLVE_OFFSET -0x4E84 RB3D_AARESOLVE_PITCH -0x4E88 RB3D_AARESOLVE_CTL 0x4EA0 RB3D_DISCARD_SRC_PIXEL_LTE_THRESHOLD 0x4EA4 RB3D_DISCARD_SRC_PIXEL_GTE_THRESHOLD 0x4F04 ZB_ZSTENCILCNTL diff --git a/drivers/gpu/drm/radeon/reg_srcs/rs600 b/drivers/gpu/drm/radeon/reg_srcs/rs600 index 263109c1d0c8..d9f62866bbc1 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/rs600 +++ b/drivers/gpu/drm/radeon/reg_srcs/rs600 @@ -770,9 +770,6 @@ rs600 0x6d40 0x4E74 RB3D_CMASK_WRINDEX 0x4E78 RB3D_CMASK_DWORD 0x4E7C RB3D_CMASK_RDINDEX -0x4E80 RB3D_AARESOLVE_OFFSET -0x4E84 RB3D_AARESOLVE_PITCH -0x4E88 RB3D_AARESOLVE_CTL 0x4EA0 RB3D_DISCARD_SRC_PIXEL_LTE_THRESHOLD 0x4EA4 RB3D_DISCARD_SRC_PIXEL_GTE_THRESHOLD 0x4F04 ZB_ZSTENCILCNTL diff --git a/drivers/gpu/drm/radeon/reg_srcs/rv515 b/drivers/gpu/drm/radeon/reg_srcs/rv515 index eeed003f14c7..911a8fbd32bb 100644 --- a/drivers/gpu/drm/radeon/reg_srcs/rv515 +++ b/drivers/gpu/drm/radeon/reg_srcs/rv515 @@ -481,9 +481,6 @@ rv515 0x6d40 0x4E74 RB3D_CMASK_WRINDEX 0x4E78 RB3D_CMASK_DWORD 0x4E7C RB3D_CMASK_RDINDEX -0x4E80 RB3D_AARESOLVE_OFFSET -0x4E84 RB3D_AARESOLVE_PITCH -0x4E88 RB3D_AARESOLVE_CTL 0x4EA0 RB3D_DISCARD_SRC_PIXEL_LTE_THRESHOLD 0x4EA4 RB3D_DISCARD_SRC_PIXEL_GTE_THRESHOLD 0x4EF8 RB3D_CONSTANT_COLOR_AR -- cgit v1.2.3 From c2049b3d29f47ed3750226dc51251a3404c85876 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Sun, 13 Feb 2011 18:42:41 -0500 Subject: drm/radeon/kms: improve 6xx/7xx CS error output Makes debugging CS rejections much easier. Signed-off-by: Alex Deucher Cc: stable@kernel.org Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/r600_cs.c | 46 +++++++++++++++++++++++----------------- 1 file changed, 26 insertions(+), 20 deletions(-) diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c index 7831e0890210..153095fba62f 100644 --- a/drivers/gpu/drm/radeon/r600_cs.c +++ b/drivers/gpu/drm/radeon/r600_cs.c @@ -295,17 +295,18 @@ static inline int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i) } if (!IS_ALIGNED(pitch, pitch_align)) { - dev_warn(p->dev, "%s:%d cb pitch (%d) invalid\n", - __func__, __LINE__, pitch); + dev_warn(p->dev, "%s:%d cb pitch (%d, 0x%x, %d) invalid\n", + __func__, __LINE__, pitch, pitch_align, array_mode); return -EINVAL; } if (!IS_ALIGNED(height, height_align)) { - dev_warn(p->dev, "%s:%d cb height (%d) invalid\n", - __func__, __LINE__, height); + dev_warn(p->dev, "%s:%d cb height (%d, 0x%x, %d) invalid\n", + __func__, __LINE__, height, height_align, array_mode); return -EINVAL; } if (!IS_ALIGNED(base_offset, base_align)) { - dev_warn(p->dev, "%s offset[%d] 0x%llx not aligned\n", __func__, i, base_offset); + dev_warn(p->dev, "%s offset[%d] 0x%llx 0x%llx, %d not aligned\n", __func__, i, + base_offset, base_align, array_mode); return -EINVAL; } @@ -320,7 +321,10 @@ static inline int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i) * broken userspace. */ } else { - dev_warn(p->dev, "%s offset[%d] %d %d %lu too big\n", __func__, i, track->cb_color_bo_offset[i], tmp, radeon_bo_size(track->cb_color_bo[i])); + dev_warn(p->dev, "%s offset[%d] %d %d %d %lu too big\n", __func__, i, + array_mode, + track->cb_color_bo_offset[i], tmp, + radeon_bo_size(track->cb_color_bo[i])); return -EINVAL; } } @@ -455,17 +459,18 @@ static int r600_cs_track_check(struct radeon_cs_parser *p) } if (!IS_ALIGNED(pitch, pitch_align)) { - dev_warn(p->dev, "%s:%d db pitch (%d) invalid\n", - __func__, __LINE__, pitch); + dev_warn(p->dev, "%s:%d db pitch (%d, 0x%x, %d) invalid\n", + __func__, __LINE__, pitch, pitch_align, array_mode); return -EINVAL; } if (!IS_ALIGNED(height, height_align)) { - dev_warn(p->dev, "%s:%d db height (%d) invalid\n", - __func__, __LINE__, height); + dev_warn(p->dev, "%s:%d db height (%d, 0x%x, %d) invalid\n", + __func__, __LINE__, height, height_align, array_mode); return -EINVAL; } if (!IS_ALIGNED(base_offset, base_align)) { - dev_warn(p->dev, "%s offset[%d] 0x%llx not aligned\n", __func__, i, base_offset); + dev_warn(p->dev, "%s offset[%d] 0x%llx, 0x%llx, %d not aligned\n", __func__, i, + base_offset, base_align, array_mode); return -EINVAL; } @@ -473,9 +478,10 @@ static int r600_cs_track_check(struct radeon_cs_parser *p) nviews = G_028004_SLICE_MAX(track->db_depth_view) + 1; tmp = ntiles * bpe * 64 * nviews; if ((tmp + track->db_offset) > radeon_bo_size(track->db_bo)) { - dev_warn(p->dev, "z/stencil buffer too small (0x%08X %d %d %d -> %u have %lu)\n", - track->db_depth_size, ntiles, nviews, bpe, tmp + track->db_offset, - radeon_bo_size(track->db_bo)); + dev_warn(p->dev, "z/stencil buffer (%d) too small (0x%08X %d %d %d -> %u have %lu)\n", + array_mode, + track->db_depth_size, ntiles, nviews, bpe, tmp + track->db_offset, + radeon_bo_size(track->db_bo)); return -EINVAL; } } @@ -1227,18 +1233,18 @@ static inline int r600_check_texture_resource(struct radeon_cs_parser *p, u32 i /* XXX check height as well... */ if (!IS_ALIGNED(pitch, pitch_align)) { - dev_warn(p->dev, "%s:%d tex pitch (%d) invalid\n", - __func__, __LINE__, pitch); + dev_warn(p->dev, "%s:%d tex pitch (%d, 0x%x, %d) invalid\n", + __func__, __LINE__, pitch, pitch_align, G_038000_TILE_MODE(word0)); return -EINVAL; } if (!IS_ALIGNED(base_offset, base_align)) { - dev_warn(p->dev, "%s:%d tex base offset (0x%llx) invalid\n", - __func__, __LINE__, base_offset); + dev_warn(p->dev, "%s:%d tex base offset (0x%llx, 0x%llx, %d) invalid\n", + __func__, __LINE__, base_offset, base_align, G_038000_TILE_MODE(word0)); return -EINVAL; } if (!IS_ALIGNED(mip_offset, base_align)) { - dev_warn(p->dev, "%s:%d tex mip offset (0x%llx) invalid\n", - __func__, __LINE__, mip_offset); + dev_warn(p->dev, "%s:%d tex mip offset (0x%llx, 0x%llx, %d) invalid\n", + __func__, __LINE__, mip_offset, base_align, G_038000_TILE_MODE(word0)); return -EINVAL; } -- cgit v1.2.3 From 265aa6c8d8822c9074a2174e8c9f31a37fa02e50 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Mon, 14 Feb 2011 16:16:22 -0500 Subject: drm/radeon/kms: fix a few more atombios endian issues Signed-off-by: Alex Deucher Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/atombios_crtc.c | 2 +- drivers/gpu/drm/radeon/rs690.c | 12 ++++++------ 2 files changed, 7 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c index 1bf61220a450..0b10b0e14945 100644 --- a/drivers/gpu/drm/radeon/atombios_crtc.c +++ b/drivers/gpu/drm/radeon/atombios_crtc.c @@ -728,7 +728,7 @@ static void atombios_crtc_set_dcpll(struct drm_crtc *crtc, /* if the default dcpll clock is specified, * SetPixelClock provides the dividers */ - args.v6.ulDispEngClkFreq = dispclk; + args.v6.ulDispEngClkFreq = cpu_to_le32(dispclk); args.v6.ucPpll = ATOM_DCPLL; break; default: diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c index 0137d3e3728d..6638c8e4c81b 100644 --- a/drivers/gpu/drm/radeon/rs690.c +++ b/drivers/gpu/drm/radeon/rs690.c @@ -77,9 +77,9 @@ void rs690_pm_info(struct radeon_device *rdev) switch (crev) { case 1: tmp.full = dfixed_const(100); - rdev->pm.igp_sideport_mclk.full = dfixed_const(info->info.ulBootUpMemoryClock); + rdev->pm.igp_sideport_mclk.full = dfixed_const(le32_to_cpu(info->info.ulBootUpMemoryClock)); rdev->pm.igp_sideport_mclk.full = dfixed_div(rdev->pm.igp_sideport_mclk, tmp); - if (info->info.usK8MemoryClock) + if (le16_to_cpu(info->info.usK8MemoryClock)) rdev->pm.igp_system_mclk.full = dfixed_const(le16_to_cpu(info->info.usK8MemoryClock)); else if (rdev->clock.default_mclk) { rdev->pm.igp_system_mclk.full = dfixed_const(rdev->clock.default_mclk); @@ -91,16 +91,16 @@ void rs690_pm_info(struct radeon_device *rdev) break; case 2: tmp.full = dfixed_const(100); - rdev->pm.igp_sideport_mclk.full = dfixed_const(info->info_v2.ulBootUpSidePortClock); + rdev->pm.igp_sideport_mclk.full = dfixed_const(le32_to_cpu(info->info_v2.ulBootUpSidePortClock)); rdev->pm.igp_sideport_mclk.full = dfixed_div(rdev->pm.igp_sideport_mclk, tmp); - if (info->info_v2.ulBootUpUMAClock) - rdev->pm.igp_system_mclk.full = dfixed_const(info->info_v2.ulBootUpUMAClock); + if (le32_to_cpu(info->info_v2.ulBootUpUMAClock)) + rdev->pm.igp_system_mclk.full = dfixed_const(le32_to_cpu(info->info_v2.ulBootUpUMAClock)); else if (rdev->clock.default_mclk) rdev->pm.igp_system_mclk.full = dfixed_const(rdev->clock.default_mclk); else rdev->pm.igp_system_mclk.full = dfixed_const(66700); rdev->pm.igp_system_mclk.full = dfixed_div(rdev->pm.igp_system_mclk, tmp); - rdev->pm.igp_ht_link_clk.full = dfixed_const(info->info_v2.ulHTLinkFreq); + rdev->pm.igp_ht_link_clk.full = dfixed_const(le32_to_cpu(info->info_v2.ulHTLinkFreq)); rdev->pm.igp_ht_link_clk.full = dfixed_div(rdev->pm.igp_ht_link_clk, tmp); rdev->pm.igp_ht_link_width.full = dfixed_const(le16_to_cpu(info->info_v2.usMinHTLinkWidth)); break; -- cgit v1.2.3 From bd91572e77013aad242e2d2d565bd730c043f8b8 Mon Sep 17 00:00:00 2001 From: Xiao Jiang Date: Mon, 14 Feb 2011 17:58:23 +0800 Subject: drm: fix wrong usages of drm_device in DRM Developer's Guide A few wrong usages of drm_device, which should be drm_driver. Signed-off-by: Xiao Jiang Signed-off-by: Dave Airlie --- Documentation/DocBook/drm.tmpl | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/Documentation/DocBook/drm.tmpl b/Documentation/DocBook/drm.tmpl index 2861055afd7a..c27915893974 100644 --- a/Documentation/DocBook/drm.tmpl +++ b/Documentation/DocBook/drm.tmpl @@ -73,8 +73,8 @@ services. - The core of every DRM driver is struct drm_device. Drivers - will typically statically initialize a drm_device structure, + The core of every DRM driver is struct drm_driver. Drivers + will typically statically initialize a drm_driver structure, then pass it to drm_init() at load time. @@ -84,7 +84,7 @@ Driver initialization Before calling the DRM initialization routines, the driver must - first create and fill out a struct drm_device structure. + first create and fill out a struct drm_driver structure. static struct drm_driver driver = { -- cgit v1.2.3 From a4b40d5d97f5c9ad0b7f4bf2818291ca184bb433 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Mon, 14 Feb 2011 11:43:10 -0500 Subject: drm/radeon/kms: add bounds checking to avivo pll algo Prevent divider overflow. Fixes: https://bugzilla.kernel.org/show_bug.cgi?id=28932 Signed-off-by: Alex Deucher Cc: stable@kernel.org Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/radeon_display.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c index 2eff98cfd728..0e657095de7c 100644 --- a/drivers/gpu/drm/radeon/radeon_display.c +++ b/drivers/gpu/drm/radeon/radeon_display.c @@ -793,6 +793,11 @@ static void avivo_get_fb_div(struct radeon_pll *pll, tmp *= target_clock; *fb_div = tmp / pll->reference_freq; *frac_fb_div = tmp % pll->reference_freq; + + if (*fb_div > pll->max_feedback_div) + *fb_div = pll->max_feedback_div; + else if (*fb_div < pll->min_feedback_div) + *fb_div = pll->min_feedback_div; } static u32 avivo_get_post_div(struct radeon_pll *pll, @@ -826,6 +831,11 @@ static u32 avivo_get_post_div(struct radeon_pll *pll, post_div--; } + if (post_div > pll->max_post_div) + post_div = pll->max_post_div; + else if (post_div < pll->min_post_div) + post_div = pll->min_post_div; + return post_div; } -- cgit v1.2.3 From 5b40ddf888398ce4cccbf3b9d0a18d90149ed7ff Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Mon, 14 Feb 2011 11:43:11 -0500 Subject: drm/radeon/kms: hopefully fix pll issues for real (v3) The problematic boards have a recommended reference divider to be used when spread spectrum is enabled on the laptop panel. Enable the use of the recommended reference divider along with the new pll algo. v2: testing options v3: When using the fixed reference divider with LVDS, prefer min m to max p and use fractional feedback dividers. Fixes: https://bugzilla.kernel.org/show_bug.cgi?id=28852 https://bugzilla.kernel.org/show_bug.cgi?id=24462 https://bugzilla.kernel.org/show_bug.cgi?id=26552 MacbookPro issues reported by Justin Mattock Signed-off-by: Alex Deucher Cc: stable@kernel.org Signed-off-by: Dave Airlie --- drivers/gpu/drm/radeon/atombios_crtc.c | 22 ++++++++-------------- 1 file changed, 8 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c index 0b10b0e14945..095bc507fb16 100644 --- a/drivers/gpu/drm/radeon/atombios_crtc.c +++ b/drivers/gpu/drm/radeon/atombios_crtc.c @@ -538,7 +538,6 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, pll->flags |= RADEON_PLL_PREFER_HIGH_FB_DIV; else pll->flags |= RADEON_PLL_PREFER_LOW_REF_DIV; - } list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) { @@ -555,29 +554,28 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc, dp_clock = dig_connector->dp_clock; } } -/* this might work properly with the new pll algo */ -#if 0 /* doesn't work properly on some laptops */ + /* use recommended ref_div for ss */ if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) { + pll->flags |= RADEON_PLL_PREFER_MINM_OVER_MAXP; if (ss_enabled) { if (ss->refdiv) { pll->flags |= RADEON_PLL_USE_REF_DIV; pll->reference_div = ss->refdiv; + if (ASIC_IS_AVIVO(rdev)) + pll->flags |= RADEON_PLL_USE_FRAC_FB_DIV; } } } -#endif + if (ASIC_IS_AVIVO(rdev)) { /* DVO wants 2x pixel clock if the DVO chip is in 12 bit mode */ if (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1) adjusted_clock = mode->clock * 2; if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)) pll->flags |= RADEON_PLL_PREFER_CLOSEST_LOWER; - /* rv515 needs more testing with this option */ - if (rdev->family != CHIP_RV515) { - if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) - pll->flags |= RADEON_PLL_IS_LCD; - } + if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) + pll->flags |= RADEON_PLL_IS_LCD; } else { if (encoder->encoder_type != DRM_MODE_ENCODER_DAC) pll->flags |= RADEON_PLL_NO_ODD_POST_DIV; @@ -957,11 +955,7 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode /* adjust pixel clock as needed */ adjusted_clock = atombios_adjust_pll(crtc, mode, pll, ss_enabled, &ss); - /* rv515 seems happier with the old algo */ - if (rdev->family == CHIP_RV515) - radeon_compute_pll_legacy(pll, adjusted_clock, &pll_clock, &fb_div, &frac_fb_div, - &ref_div, &post_div); - else if (ASIC_IS_AVIVO(rdev)) + if (ASIC_IS_AVIVO(rdev)) radeon_compute_pll_avivo(pll, adjusted_clock, &pll_clock, &fb_div, &frac_fb_div, &ref_div, &post_div); else -- cgit v1.2.3