Merge branch 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/ickle/drm-intel

* 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/ickle/drm-intel:
  drm/i915: Fix LVDS fixed-mode regression from 219adae1
  drm/i915/ringbuffer: Use the HEAD auto-reporting mechanism
  drm/i915: Avoid might_fault during pwrite whilst holding our mutex
  agp/intel: fix cache control for sandybridge
  agp/intel: restore cache behavior on sandybridge
  drm/i915; Don't apply Ironlake FDI clock workaround to Sandybridge
  drm/i915: Fix KMS regression on Sandybridge/CPT
  i915: reprogram power monitoring registers on resume
  drm/i915: SNB BLT workaround
  drm/i915: Fix the graphics frequency clamping at init and when IPS is active.
  drm/i915: Allow powersave modparam to be adjusted at runtime.
  drm/i915: Apply big hammer to serialise buffer access between rings
  drm/i915: opregion_setup: iounmap correct address
  drm/i915: Flush read-only buffers from the active list upon idle as well
  i915: signedness bug in check_overlay_src()
  drm/i915: Fix typo from "Enable DisplayPort Audio"

14 files changed, 267 insertions, 96 deletions

diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index 6b6760ea243..9272c38dd3c 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -1210,14 +1210,14 @@ static void gen6_write_entry(dma_addr_t addr, unsigned int entry,
 	unsigned int gfdt = flags & AGP_USER_CACHED_MEMORY_GFDT;
 	u32 pte_flags;
 
-	if (type_mask == AGP_USER_UNCACHED_MEMORY)
+	if (type_mask == AGP_USER_MEMORY)
 		pte_flags = GEN6_PTE_UNCACHED | I810_PTE_VALID;
 	else if (type_mask == AGP_USER_CACHED_MEMORY_LLC_MLC) {
-		pte_flags = GEN6_PTE_LLC | I810_PTE_VALID;
+		pte_flags = GEN6_PTE_LLC_MLC | I810_PTE_VALID;
 		if (gfdt)
 			pte_flags |= GEN6_PTE_GFDT;
 	} else { /* set 'normal'/'cached' to LLC by default */
-		pte_flags = GEN6_PTE_LLC_MLC | I810_PTE_VALID;
+		pte_flags = GEN6_PTE_LLC | I810_PTE_VALID;
 		if (gfdt)
 			pte_flags |= GEN6_PTE_GFDT;
 	}
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 3467dd42076..80745f85902 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -44,7 +44,7 @@ unsigned int i915_fbpercrtc = 0;
 module_param_named(fbpercrtc, i915_fbpercrtc, int, 0400);
 
 unsigned int i915_powersave = 1;
-module_param_named(powersave, i915_powersave, int, 0400);
+module_param_named(powersave, i915_powersave, int, 0600);
 
 unsigned int i915_lvds_downclock = 0;
 module_param_named(lvds_downclock, i915_lvds_downclock, int, 0400);
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 2c2c19b6285..90414ae86af 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -1321,6 +1321,7 @@ static inline void i915_write(struct drm_i915_private *dev_priv, u32 reg,
 
 #define INTEL_PCH_TYPE(dev) (((struct drm_i915_private *)(dev)->dev_private)->pch_type)
 #define HAS_PCH_CPT(dev) (INTEL_PCH_TYPE(dev) == PCH_CPT)
+#define HAS_PCH_IBX(dev) (INTEL_PCH_TYPE(dev) == PCH_IBX)
 
 #define PRIMARY_RINGBUFFER_SIZE         (128*1024)
 
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 8eb8453208b..eba9b161522 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -2172,7 +2172,7 @@ i915_gem_object_unbind(struct drm_gem_object *obj)
 static int i915_ring_idle(struct drm_device *dev,
 			  struct intel_ring_buffer *ring)
 {
-	if (list_empty(&ring->gpu_write_list))
+	if (list_empty(&ring->gpu_write_list) && list_empty(&ring->active_list))
 		return 0;
 
 	i915_gem_flush_ring(dev, NULL, ring,
@@ -2190,9 +2190,7 @@ i915_gpu_idle(struct drm_device *dev)
 	int ret;
 
 	lists_empty = (list_empty(&dev_priv->mm.flushing_list) &&
-		       list_empty(&dev_priv->render_ring.active_list) &&
-		       list_empty(&dev_priv->bsd_ring.active_list) &&
-		       list_empty(&dev_priv->blt_ring.active_list));
+		       list_empty(&dev_priv->mm.active_list));
 	if (lists_empty)
 		return 0;
 
@@ -3108,7 +3106,8 @@ i915_gem_object_set_to_gpu_domain(struct drm_gem_object *obj,
 	 * write domain
 	 */
 	if (obj->write_domain &&
-	    obj->write_domain != obj->pending_read_domains) {
+	    (obj->write_domain != obj->pending_read_domains ||
+	     obj_priv->ring != ring)) {
 		flush_domains |= obj->write_domain;
 		invalidate_domains |=
 			obj->pending_read_domains & ~obj->write_domain;
@@ -3497,6 +3496,52 @@ i915_gem_execbuffer_pin(struct drm_device *dev,
 	return 0;
 }
 
+static int
+i915_gem_execbuffer_move_to_gpu(struct drm_device *dev,
+				struct drm_file *file,
+				struct intel_ring_buffer *ring,
+				struct drm_gem_object **objects,
+				int count)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret, i;
+
+	/* Zero the global flush/invalidate flags. These
+	 * will be modified as new domains are computed
+	 * for each object
+	 */
+	dev->invalidate_domains = 0;
+	dev->flush_domains = 0;
+	dev_priv->mm.flush_rings = 0;
+	for (i = 0; i < count; i++)
+		i915_gem_object_set_to_gpu_domain(objects[i], ring);
+
+	if (dev->invalidate_domains | dev->flush_domains) {
+#if WATCH_EXEC
+		DRM_INFO("%s: invalidate_domains %08x flush_domains %08x\n",
+			  __func__,
+			 dev->invalidate_domains,
+			 dev->flush_domains);
+#endif
+		i915_gem_flush(dev, file,
+			       dev->invalidate_domains,
+			       dev->flush_domains,
+			       dev_priv->mm.flush_rings);
+	}
+
+	for (i = 0; i < count; i++) {
+		struct drm_i915_gem_object *obj = to_intel_bo(objects[i]);
+		/* XXX replace with semaphores */
+		if (obj->ring && ring != obj->ring) {
+			ret = i915_gem_object_wait_rendering(&obj->base, true);
+			if (ret)
+				return ret;
+		}
+	}
+
+	return 0;
+}
+
 /* Throttle our rendering by waiting until the ring has completed our requests
  * emitted over 20 msec ago.
  *
@@ -3757,33 +3802,10 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		goto err;
 	}
 
-	/* Zero the global flush/invalidate flags. These
-	 * will be modified as new domains are computed
-	 * for each object
-	 */
-	dev->invalidate_domains = 0;
-	dev->flush_domains = 0;
-	dev_priv->mm.flush_rings = 0;
-
-	for (i = 0; i < args->buffer_count; i++) {
-		struct drm_gem_object *obj = object_list[i];
-
-		/* Compute new gpu domains and update invalidate/flush */
-		i915_gem_object_set_to_gpu_domain(obj, ring);
-	}
-
-	if (dev->invalidate_domains | dev->flush_domains) {
-#if WATCH_EXEC
-		DRM_INFO("%s: invalidate_domains %08x flush_domains %08x\n",
-			  __func__,
-			 dev->invalidate_domains,
-			 dev->flush_domains);
-#endif
-		i915_gem_flush(dev, file,
-			       dev->invalidate_domains,
-			       dev->flush_domains,
-			       dev_priv->mm.flush_rings);
-	}
+	ret = i915_gem_execbuffer_move_to_gpu(dev, file, ring,
+					      object_list, args->buffer_count);
+	if (ret)
+		goto err;
 
 	for (i = 0; i < args->buffer_count; i++) {
 		struct drm_gem_object *obj = object_list[i];
@@ -4856,17 +4878,24 @@ i915_gem_phys_pwrite(struct drm_device *dev, struct drm_gem_object *obj,
 		     struct drm_file *file_priv)
 {
 	struct drm_i915_gem_object *obj_priv = to_intel_bo(obj);
-	void *obj_addr;
-	int ret;
-	char __user *user_data;
+	void *vaddr = obj_priv->phys_obj->handle->vaddr + args->offset;
+	char __user *user_data = (char __user *) (uintptr_t) args->data_ptr;
 
-	user_data = (char __user *) (uintptr_t) args->data_ptr;
-	obj_addr = obj_priv->phys_obj->handle->vaddr + args->offset;
+	DRM_DEBUG_DRIVER("vaddr %p, %lld\n", vaddr, args->size);
 
-	DRM_DEBUG_DRIVER("obj_addr %p, %lld\n", obj_addr, args->size);
-	ret = copy_from_user(obj_addr, user_data, args->size);
-	if (ret)
-		return -EFAULT;
+	if (__copy_from_user_inatomic_nocache(vaddr, user_data, args->size)) {
+		unsigned long unwritten;
+
+		/* The physical object once assigned is fixed for the lifetime
+		 * of the obj, so we can safely drop the lock and continue
+		 * to access vaddr.
+		 */
+		mutex_unlock(&dev->struct_mutex);
+		unwritten = copy_from_user(vaddr, user_data, args->size);
+		mutex_lock(&dev->struct_mutex);
+		if (unwritten)
+			return -EFAULT;
+	}
 
 	drm_agp_chipset_flush(dev);
 	return 0;
@@ -4900,9 +4929,7 @@ i915_gpu_is_active(struct drm_device *dev)
 	int lists_empty;
 
 	lists_empty = list_empty(&dev_priv->mm.flushing_list) &&
-		      list_empty(&dev_priv->render_ring.active_list) &&
-		      list_empty(&dev_priv->bsd_ring.active_list) &&
-		      list_empty(&dev_priv->blt_ring.active_list);
+		      list_empty(&dev_priv->mm.active_list);
 
 	return !lists_empty;
 }
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
index 43a4013f53f..d8ae7d1d0cc 100644
--- a/drivers/gpu/drm/i915/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -165,9 +165,7 @@ i915_gem_evict_everything(struct drm_device *dev)
 
 	lists_empty = (list_empty(&dev_priv->mm.inactive_list) &&
 		       list_empty(&dev_priv->mm.flushing_list) &&
-		       list_empty(&dev_priv->render_ring.active_list) &&
-		       list_empty(&dev_priv->bsd_ring.active_list) &&
-		       list_empty(&dev_priv->blt_ring.active_list));
+		       list_empty(&dev_priv->mm.active_list));
 	if (lists_empty)
 		return -ENOSPC;
 
@@ -184,9 +182,7 @@ i915_gem_evict_everything(struct drm_device *dev)
 
 	lists_empty = (list_empty(&dev_priv->mm.inactive_list) &&
 		       list_empty(&dev_priv->mm.flushing_list) &&
-		       list_empty(&dev_priv->render_ring.active_list) &&
-		       list_empty(&dev_priv->bsd_ring.active_list) &&
-		       list_empty(&dev_priv->blt_ring.active_list));
+		       list_empty(&dev_priv->mm.active_list));
 	BUG_ON(!lists_empty);
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 989c19d2d95..454c064f8ef 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -862,8 +862,10 @@ int i915_restore_state(struct drm_device *dev)
 	/* Clock gating state */
 	intel_init_clock_gating(dev);
 
-	if (HAS_PCH_SPLIT(dev))
+	if (HAS_PCH_SPLIT(dev)) {
 		ironlake_enable_drps(dev);
+		intel_init_emon(dev);
+	}
 
 	/* Cache mode state */
 	I915_WRITE (CACHE_MODE_0, dev_priv->saveCACHE_MODE_0 | 0xffff0000);
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 990f065374b..48d8fd686ea 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -1681,6 +1681,37 @@ static void ironlake_set_pll_edp(struct drm_crtc *crtc, int clock)
 	udelay(500);
 }
 
+static void intel_fdi_normal_train(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	int pipe = intel_crtc->pipe;
+	u32 reg, temp;
+
+	/* enable normal train */
+	reg = FDI_TX_CTL(pipe);
+	temp = I915_READ(reg);
+	temp &= ~FDI_LINK_TRAIN_NONE;
+	temp |= FDI_LINK_TRAIN_NONE | FDI_TX_ENHANCE_FRAME_ENABLE;
+	I915_WRITE(reg, temp);
+
+	reg = FDI_RX_CTL(pipe);
+	temp = I915_READ(reg);
+	if (HAS_PCH_CPT(dev)) {
+		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
+		temp |= FDI_LINK_TRAIN_NORMAL_CPT;
+	} else {
+		temp &= ~FDI_LINK_TRAIN_NONE;
+		temp |= FDI_LINK_TRAIN_NONE;
+	}
+	I915_WRITE(reg, temp | FDI_RX_ENHANCE_FRAME_ENABLE);
+
+	/* wait one idle pattern time */
+	POSTING_READ(reg);
+	udelay(1000);
+}
+
 /* The FDI link training functions for ILK/Ibexpeak. */
 static void ironlake_fdi_link_train(struct drm_crtc *crtc)
 {
@@ -1767,27 +1798,6 @@ static void ironlake_fdi_link_train(struct drm_crtc *crtc)
 
 	DRM_DEBUG_KMS("FDI train done\n");
 
-	/* enable normal train */
-	reg = FDI_TX_CTL(pipe);
-	temp = I915_READ(reg);
-	temp &= ~FDI_LINK_TRAIN_NONE;
-	temp |= FDI_LINK_TRAIN_NONE | FDI_TX_ENHANCE_FRAME_ENABLE;
-	I915_WRITE(reg, temp);
-
-	reg = FDI_RX_CTL(pipe);
-	temp = I915_READ(reg);
-	if (HAS_PCH_CPT(dev)) {
-		temp &= ~FDI_LINK_TRAIN_PATTERN_MASK_CPT;
-		temp |= FDI_LINK_TRAIN_NORMAL_CPT;
-	} else {
-		temp &= ~FDI_LINK_TRAIN_NONE;
-		temp |= FDI_LINK_TRAIN_NONE;
-	}
-	I915_WRITE(reg, temp | FDI_RX_ENHANCE_FRAME_ENABLE);
-
-	/* wait one idle pattern time */
-	POSTING_READ(reg);
-	udelay(1000);
 }
 
 static const int const snb_b_fdi_train_param [] = {
@@ -2090,6 +2100,8 @@ static void ironlake_crtc_enable(struct drm_crtc *crtc)
 	I915_WRITE(TRANS_VBLANK(pipe), I915_READ(VBLANK(pipe)));
 	I915_WRITE(TRANS_VSYNC(pipe),  I915_READ(VSYNC(pipe)));
 
+	intel_fdi_normal_train(crtc);
+
 	/* For PCH DP, enable TRANS_DP_CTL */
 	if (HAS_PCH_CPT(dev) &&
 	    intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)) {
@@ -2200,9 +2212,10 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
 	udelay(100);
 
 	/* Ironlake workaround, disable clock pointer after downing FDI */
-	I915_WRITE(FDI_RX_CHICKEN(pipe),
-		   I915_READ(FDI_RX_CHICKEN(pipe) &
-			     ~FDI_RX_PHASE_SYNC_POINTER_ENABLE));
+	if (HAS_PCH_IBX(dev))
+		I915_WRITE(FDI_RX_CHICKEN(pipe),
+			   I915_READ(FDI_RX_CHICKEN(pipe) &
+				     ~FDI_RX_PHASE_SYNC_POINTER_ENABLE));
 
 	/* still set train pattern 1 */
 	reg = FDI_TX_CTL(pipe);
@@ -5581,20 +5594,19 @@ void ironlake_enable_drps(struct drm_device *dev)
 	fmin = (rgvmodectl & MEMMODE_FMIN_MASK);
 	fstart = (rgvmodectl & MEMMODE_FSTART_MASK) >>
 		MEMMODE_FSTART_SHIFT;
-	fstart = fmax;
 
 	vstart = (I915_READ(PXVFREQ_BASE + (fstart * 4)) & PXVFREQ_PX_MASK) >>
 		PXVFREQ_PX_SHIFT;
 
-	dev_priv->fmax = fstart; /* IPS callback will increase this */
+	dev_priv->fmax = fmax; /* IPS callback will increase this */
 	dev_priv->fstart = fstart;
 
-	dev_priv->max_delay = fmax;
+	dev_priv->max_delay = fstart;
 	dev_priv->min_delay = fmin;
 	dev_priv->cur_delay = fstart;
 
-	DRM_DEBUG_DRIVER("fmax: %d, fmin: %d, fstart: %d\n", fmax, fmin,
-			 fstart);
+	DRM_DEBUG_DRIVER("fmax: %d, fmin: %d, fstart: %d\n",
+			 fmax, fmin, fstart);
 
 	I915_WRITE(MEMINTREN, MEMINT_CX_SUPR_EN | MEMINT_EVAL_CHG_EN);
 
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 891f4f1d63b..c8e00555331 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -1517,7 +1517,7 @@ g4x_dp_detect(struct intel_dp *intel_dp)
 			status = connector_status_connected;
 	}
 
-	return bit;
+	return status;
 }
 
 /**
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 9af9f86a876..21551fe7454 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -296,6 +296,7 @@ extern void intel_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green,
 extern void intel_init_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 intel_init_emon(struct drm_device *dev);
 
 extern int intel_pin_and_fence_fb_obj(struct drm_device *dev,
 				      struct drm_gem_object *obj,
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index f1a649990ea..4324a326f98 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -481,11 +481,8 @@ static int intel_lvds_get_modes(struct drm_connector *connector)
 	struct drm_device *dev = connector->dev;
 	struct drm_display_mode *mode;
 
-	if (intel_lvds->edid) {
-		drm_mode_connector_update_edid_property(connector,
-							intel_lvds->edid);
+	if (intel_lvds->edid)
 		return drm_add_edid_modes(connector, intel_lvds->edid);
-	}
 
 	mode = drm_mode_duplicate(dev, intel_lvds->fixed_mode);
 	if (mode == 0)
@@ -939,7 +936,16 @@ void intel_lvds_init(struct drm_device *dev)
 	 */
 	intel_lvds->edid = drm_get_edid(connector,
 					&dev_priv->gmbus[pin].adapter);
-
+	if (intel_lvds->edid) {
+		if (drm_add_edid_modes(connector,
+				       intel_lvds->edid)) {
+			drm_mode_connector_update_edid_property(connector,
+								intel_lvds->edid);
+		} else {
+			kfree(intel_lvds->edid);
+			intel_lvds->edid = NULL;
+		}
+	}
 	if (!intel_lvds->edid) {
 		/* Didn't get an EDID, so
 		 * Set wide sync ranges so we get all modes
diff --git a/drivers/gpu/drm/i915/intel_opregion.c b/drivers/gpu/drm/i915/intel_opregion.c
index 917c7dc3cd6..9b0d9a867ae 100644
--- a/drivers/gpu/drm/i915/intel_opregion.c
+++ b/drivers/gpu/drm/i915/intel_opregion.c
@@ -512,6 +512,6 @@ int intel_opregion_setup(struct drm_device *dev)
 	return 0;
 
 err_out:
-	iounmap(opregion->header);
+	iounmap(base);
 	return err;
 }
diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index afb96d25219..02ff0a481f4 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -946,7 +946,9 @@ static int check_overlay_src(struct drm_device *dev,
 {
 	int uv_hscale = uv_hsubsampling(rec->flags);
 	int uv_vscale = uv_vsubsampling(rec->flags);
-	u32 stride_mask, depth, tmp;
+	u32 stride_mask;
+	int depth;
+	u32 tmp;
 
 	/* check src dimensions */
 	if (IS_845G(dev) || IS_I830(dev)) {
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 09f2dc353ae..b83306f9244 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -177,7 +177,7 @@ static int init_ring_common(struct drm_device *dev,
 
 	I915_WRITE_CTL(ring,
 			((ring->gem_object->size - PAGE_SIZE) & RING_NR_PAGES)
-			| RING_NO_REPORT | RING_VALID);
+			| RING_REPORT_64K | RING_VALID);
 
 	head = I915_READ_HEAD(ring) & HEAD_ADDR;
 	/* If the head is still not zero, the ring is dead */
@@ -654,6 +654,10 @@ void intel_cleanup_ring_buffer(struct drm_device *dev,
 	i915_gem_object_unpin(ring->gem_object);
 	drm_gem_object_unreference(ring->gem_object);
 	ring->gem_object = NULL;
+
+	if (ring->cleanup)
+		ring->cleanup(ring);
+
 	cleanup_status_page(dev, ring);
 }
 
@@ -688,6 +692,17 @@ int intel_wait_ring_buffer(struct drm_device *dev,
 {
 	unsigned long end;
 	drm_i915_private_t *dev_priv = dev->dev_private;
+	u32 head;
+
+	head = intel_read_status_page(ring, 4);
+	if (head) {
+		ring->head = head & HEAD_ADDR;
+		ring->space = ring->head - (ring->tail + 8);
+		if (ring->space < 0)
+			ring->space += ring->size;
+		if (ring->space >= n)
+			return 0;
+	}
 
 	trace_i915_ring_wait_begin (dev);
 	end = jiffies + 3 * HZ;
@@ -854,19 +869,125 @@ blt_ring_put_user_irq(struct drm_device *dev,
 	/* do nothing */
 }
 
+
+/* Workaround for some stepping of SNB,
+ * each time when BLT engine ring tail moved,
+ * the first command in the ring to be parsed
+ * should be MI_BATCH_BUFFER_START
+ */
+#define NEED_BLT_WORKAROUND(dev) \
+	(IS_GEN6(dev) && (dev->pdev->revision < 8))
+
+static inline struct drm_i915_gem_object *
+to_blt_workaround(struct intel_ring_buffer *ring)
+{
+	return ring->private;
+}
+
+static int blt_ring_init(struct drm_device *dev,
+			 struct intel_ring_buffer *ring)
+{
+	if (NEED_BLT_WORKAROUND(dev)) {
+		struct drm_i915_gem_object *obj;
+		u32 __iomem *ptr;
+		int ret;
+
+		obj = to_intel_bo(i915_gem_alloc_object(dev, 4096));
+		if (obj == NULL)
+			return -ENOMEM;
+
+		ret = i915_gem_object_pin(&obj->base, 4096);
+		if (ret) {
+			drm_gem_object_unreference(&obj->base);
+			return ret;
+		}
+
+		ptr = kmap(obj->pages[0]);
+		iowrite32(MI_BATCH_BUFFER_END, ptr);
+		iowrite32(MI_NOOP, ptr+1);
+		kunmap(obj->pages[0]);
+
+		ret = i915_gem_object_set_to_gtt_domain(&obj->base, false);
+		if (ret) {
+			i915_gem_object_unpin(&obj->base);
+			drm_gem_object_unreference(&obj->base);
+			return ret;
+		}
+
+		ring->private = obj;
+	}
+
+	return init_ring_common(dev, ring);
+}
+
+static void blt_ring_begin(struct drm_device *dev,
+			   struct intel_ring_buffer *ring,
+			  int num_dwords)
+{
+	if (ring->private) {
+		intel_ring_begin(dev, ring, num_dwords+2);
+		intel_ring_emit(dev, ring, MI_BATCH_BUFFER_START);
+		intel_ring_emit(dev, ring, to_blt_workaround(ring)->gtt_offset);
+	} else
+		intel_ring_begin(dev, ring, 4);
+}
+
+static void blt_ring_flush(struct drm_device *dev,
+			   struct intel_ring_buffer *ring,
+			   u32 invalidate_domains,
+			   u32 flush_domains)
+{
+	blt_ring_begin(dev, ring, 4);
+	intel_ring_emit(dev, ring, MI_FLUSH_DW);
+	intel_ring_emit(dev, ring, 0);
+	intel_ring_emit(dev, ring, 0);
+	intel_ring_emit(dev, ring, 0);
+	intel_ring_advance(dev, ring);
+}
+
+static u32
+blt_ring_add_request(struct drm_device *dev,
+		     struct intel_ring_buffer *ring,
+		     u32 flush_domains)
+{
+	u32 seqno = i915_gem_get_seqno(dev);
+
+	blt_ring_begin(dev, ring, 4);
+	intel_ring_emit(dev, ring, MI_STORE_DWORD_INDEX);
+	intel_ring_emit(dev, ring,
+			I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT);
+	intel_ring_emit(dev, ring, seqno);
+	intel_ring_emit(dev, ring, MI_USER_INTERRUPT);
+	intel_ring_advance(dev, ring);
+
+	DRM_DEBUG_DRIVER("%s %d\n", ring->name, seqno);
+	return seqno;
+}
+
+static void blt_ring_cleanup(struct intel_ring_buffer *ring)
+{
+	if (!ring->private)
+		return;
+
+	i915_gem_object_unpin(ring->private);
+	drm_gem_object_unreference(ring->private);
+	ring->private = NULL;
+}
+
 static const struct intel_ring_buffer gen6_blt_ring = {
        .name			= "blt ring",
        .id			= RING_BLT,
        .mmio_base		= BLT_RING_BASE,
        .size			= 32 * PAGE_SIZE,
-       .init			= init_ring_common,
+       .init			= blt_ring_init,
        .write_tail		= ring_write_tail,
-       .flush			= gen6_ring_flush,
-       .add_request		= ring_add_request,
+       .flush			= blt_ring_flush,
+       .add_request		= blt_ring_add_request,
        .get_seqno		= ring_status_page_get_seqno,
        .user_irq_get		= blt_ring_get_user_irq,
        .user_irq_put		= blt_ring_put_user_irq,
        .dispatch_gem_execbuffer	= gen6_ring_dispatch_gem_execbuffer,
+       .cleanup			= blt_ring_cleanup,
 };
 
 int intel_init_render_ring_buffer(struct drm_device *dev)
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index a05aff0e576..3126c268198 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -63,6 +63,7 @@ struct  intel_ring_buffer {
 			struct drm_i915_gem_execbuffer2 *exec,
 			struct drm_clip_rect *cliprects,
 			uint64_t exec_offset);
+	void		(*cleanup)(struct intel_ring_buffer *ring);
 
 	/**
 	 * List of objects currently involved in rendering from the
@@ -98,6 +99,8 @@ struct  intel_ring_buffer {
 
 	wait_queue_head_t irq_queue;
 	drm_local_map_t map;
+
+	void *private;
 };
 
 static inline u32