Merge branch 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6

* 'drm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6: (39 commits)
  i915/gtt: fix ordering causing DMAR errors on object teardown.
  i915/gtt: fix ordering issues with status setup and DMAR
  drm/i915/execbuffer: Reorder binding of objects to favour restrictions
  drm/i915: If we hit OOM when allocating GTT pages, clear the aperture
  drm/i915/evict: Ensure we completely cleanup on failure
  drm/i915/execbuffer: Correctly clear the current object list upon EFAULT
  drm/i915/debugfs: Show all objects in the gtt
  drm/i915: Record AGP memory type upon error
  drm/i915: Periodically flush the active lists and requests
  drm/i915/gtt: Unmap the PCI pages after unbinding them from the GTT
  drm/i915: Record the error batchbuffer on each ring
  drm/i915: Include TLB miss overhead for computing WM
  drm/i915: Propagate error from flushing the ring
  drm/i915: detect & report PCH display error interrupts
  drm/i915: cleanup rc6 code
  drm/i915: fix rc6 enabling around suspend/resume
  drm/i915: re-enable rc6 support for Ironlake+
  drm/i915: Make the ring IMR handling private
  drm/i915/ringbuffer: Simplify the ring irq refcounting
  drm/i915/debugfs: Show the per-ring IMR
  ...

23 files changed, 1082 insertions, 637 deletions

diff --git a/drivers/char/agp/intel-agp.h b/drivers/char/agp/intel-agp.h
index 010e3defd6c3..c195bfeade11 100644
--- a/drivers/char/agp/intel-agp.h
+++ b/drivers/char/agp/intel-agp.h
@@ -94,6 +94,8 @@
 #define G4x_GMCH_SIZE_VT_1_5M	(0xa << 8)
 #define G4x_GMCH_SIZE_VT_2M	(0xc << 8)
 
+#define GFX_FLSH_CNTL		0x2170 /* 915+ */
+
 #define I810_DRAM_CTL		0x3000
 #define I810_DRAM_ROW_0		0x00000001
 #define I810_DRAM_ROW_0_SDRAM	0x00000001
diff --git a/drivers/char/agp/intel-gtt.c b/drivers/char/agp/intel-gtt.c
index 356f73e0d17e..e921b693412b 100644
--- a/drivers/char/agp/intel-gtt.c
+++ b/drivers/char/agp/intel-gtt.c
@@ -688,14 +688,14 @@ static int intel_gtt_init(void)
 
 	intel_private.base.stolen_size = intel_gtt_stolen_size();
 
+	intel_private.base.needs_dmar = USE_PCI_DMA_API && INTEL_GTT_GEN > 2;
+
 	ret = intel_gtt_setup_scratch_page();
 	if (ret != 0) {
 		intel_gtt_cleanup();
 		return ret;
 	}
 
-	intel_private.base.needs_dmar = USE_PCI_DMA_API && INTEL_GTT_GEN > 2;
-
 	return 0;
 }
 
@@ -814,6 +814,12 @@ static bool intel_enable_gtt(void)
 		}
 	}
 
+	/* On the resume path we may be adjusting the PGTBL value, so
+	 * be paranoid and flush all chipset write buffers...
+	 */
+	if (INTEL_GTT_GEN >= 3)
+		writel(0, intel_private.registers+GFX_FLSH_CNTL);
+
 	reg = intel_private.registers+I810_PGETBL_CTL;
 	writel(intel_private.PGETBL_save, reg);
 	if (HAS_PGTBL_EN && (readl(reg) & I810_PGETBL_ENABLED) == 0) {
@@ -823,6 +829,9 @@ static bool intel_enable_gtt(void)
 		return false;
 	}
 
+	if (INTEL_GTT_GEN >= 3)
+		writel(0, intel_private.registers+GFX_FLSH_CNTL);
+
 	return true;
 }
 
@@ -991,14 +1000,14 @@ static int intel_fake_agp_remove_entries(struct agp_memory *mem,
 	if (mem->page_count == 0)
 		return 0;
 
+	intel_gtt_clear_range(pg_start, mem->page_count);
+
 	if (intel_private.base.needs_dmar) {
 		intel_gtt_unmap_memory(mem->sg_list, mem->num_sg);
 		mem->sg_list = NULL;
 		mem->num_sg = 0;
 	}
 
-	intel_gtt_clear_range(pg_start, mem->page_count);
-
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 92f75782c332..19a3d58044dd 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -106,10 +106,19 @@ static const char *get_tiling_flag(struct drm_i915_gem_object *obj)
     }
 }
 
+static const char *agp_type_str(int type)
+{
+	switch (type) {
+	case 0: return " uncached";
+	case 1: return " snooped";
+	default: return "";
+	}
+}
+
 static void
 describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
 {
-	seq_printf(m, "%p: %s%s %8zd %04x %04x %d %d%s%s",
+	seq_printf(m, "%p: %s%s %8zd %04x %04x %d %d%s%s%s",
 		   &obj->base,
 		   get_pin_flag(obj),
 		   get_tiling_flag(obj),
@@ -118,6 +127,7 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
 		   obj->base.write_domain,
 		   obj->last_rendering_seqno,
 		   obj->last_fenced_seqno,
+		   agp_type_str(obj->agp_type == AGP_USER_CACHED_MEMORY),
 		   obj->dirty ? " dirty" : "",
 		   obj->madv == I915_MADV_DONTNEED ? " purgeable" : "");
 	if (obj->base.name)
@@ -276,6 +286,37 @@ static int i915_gem_object_info(struct seq_file *m, void* data)
 	return 0;
 }
 
+static int i915_gem_gtt_info(struct seq_file *m, void* data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_gem_object *obj;
+	size_t total_obj_size, total_gtt_size;
+	int count, ret;
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+
+	total_obj_size = total_gtt_size = count = 0;
+	list_for_each_entry(obj, &dev_priv->mm.gtt_list, gtt_list) {
+		seq_printf(m, "   ");
+		describe_obj(m, obj);
+		seq_printf(m, "\n");
+		total_obj_size += obj->base.size;
+		total_gtt_size += obj->gtt_space->size;
+		count++;
+	}
+
+	mutex_unlock(&dev->struct_mutex);
+
+	seq_printf(m, "Total %d objects, %zu bytes, %zu GTT size\n",
+		   count, total_obj_size, total_gtt_size);
+
+	return 0;
+}
+
 
 static int i915_gem_pageflip_info(struct seq_file *m, void *data)
 {
@@ -456,8 +497,14 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 	}
 	seq_printf(m, "Interrupts received: %d\n",
 		   atomic_read(&dev_priv->irq_received));
-	for (i = 0; i < I915_NUM_RINGS; i++)
+	for (i = 0; i < I915_NUM_RINGS; i++) {
+		if (IS_GEN6(dev)) {
+			seq_printf(m, "Graphics Interrupt mask (%s):	%08x\n",
+				   dev_priv->ring[i].name,
+				   I915_READ_IMR(&dev_priv->ring[i]));
+		}
 		i915_ring_seqno_info(m, &dev_priv->ring[i]);
+	}
 	mutex_unlock(&dev->struct_mutex);
 
 	return 0;
@@ -656,7 +703,7 @@ static void print_error_buffers(struct seq_file *m,
 	seq_printf(m, "%s [%d]:\n", name, count);
 
 	while (count--) {
-		seq_printf(m, "  %08x %8zd %04x %04x %08x%s%s%s%s%s",
+		seq_printf(m, "  %08x %8zd %04x %04x %08x%s%s%s%s%s%s",
 			   err->gtt_offset,
 			   err->size,
 			   err->read_domains,
@@ -666,7 +713,8 @@ static void print_error_buffers(struct seq_file *m,
 			   tiling_flag(err->tiling),
 			   dirty_flag(err->dirty),
 			   purgeable_flag(err->purgeable),
-			   ring_str(err->ring));
+			   ring_str(err->ring),
+			   agp_type_str(err->agp_type));
 
 		if (err->name)
 			seq_printf(m, " (name: %d)", err->name);
@@ -744,7 +792,9 @@ static int i915_error_state(struct seq_file *m, void *unused)
 		if (error->batchbuffer[i]) {
 			struct drm_i915_error_object *obj = error->batchbuffer[i];
 
-			seq_printf(m, "--- gtt_offset = 0x%08x\n", obj->gtt_offset);
+			seq_printf(m, "%s --- gtt_offset = 0x%08x\n",
+				   dev_priv->ring[i].name,
+				   obj->gtt_offset);
 			offset = 0;
 			for (page = 0; page < obj->page_count; page++) {
 				for (elt = 0; elt < PAGE_SIZE/4; elt++) {
@@ -890,7 +940,7 @@ static int i915_drpc_info(struct seq_file *m, void *unused)
 	struct drm_device *dev = node->minor->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	u32 rgvmodectl = I915_READ(MEMMODECTL);
-	u32 rstdbyctl = I915_READ(MCHBAR_RENDER_STANDBY);
+	u32 rstdbyctl = I915_READ(RSTDBYCTL);
 	u16 crstandvid = I915_READ16(CRSTANDVID);
 
 	seq_printf(m, "HD boost: %s\n", (rgvmodectl & MEMMODE_BOOST_EN) ?
@@ -913,6 +963,30 @@ static int i915_drpc_info(struct seq_file *m, void *unused)
 	seq_printf(m, "RS2 VID: %d\n", ((crstandvid >> 8) & 0x3f));
 	seq_printf(m, "Render standby enabled: %s\n",
 		   (rstdbyctl & RCX_SW_EXIT) ? "no" : "yes");
+	seq_printf(m, "Current RS state: ");
+	switch (rstdbyctl & RSX_STATUS_MASK) {
+	case RSX_STATUS_ON:
+		seq_printf(m, "on\n");
+		break;
+	case RSX_STATUS_RC1:
+		seq_printf(m, "RC1\n");
+		break;
+	case RSX_STATUS_RC1E:
+		seq_printf(m, "RC1E\n");
+		break;
+	case RSX_STATUS_RS1:
+		seq_printf(m, "RS1\n");
+		break;
+	case RSX_STATUS_RS2:
+		seq_printf(m, "RS2 (RC6)\n");
+		break;
+	case RSX_STATUS_RS3:
+		seq_printf(m, "RC3 (RC6+)\n");
+		break;
+	default:
+		seq_printf(m, "unknown\n");
+		break;
+	}
 
 	return 0;
 }
@@ -1187,6 +1261,7 @@ static int i915_wedged_create(struct dentry *root, struct drm_minor *minor)
 static struct drm_info_list i915_debugfs_list[] = {
 	{"i915_capabilities", i915_capabilities, 0, 0},
 	{"i915_gem_objects", i915_gem_object_info, 0},
+	{"i915_gem_gtt", i915_gem_gtt_info, 0},
 	{"i915_gem_active", i915_gem_object_list_info, 0, (void *) ACTIVE_LIST},
 	{"i915_gem_flushing", i915_gem_object_list_info, 0, (void *) FLUSHING_LIST},
 	{"i915_gem_inactive", i915_gem_object_list_info, 0, (void *) INACTIVE_LIST},
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 0568dbdc10ef..844f3c972b04 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -1962,13 +1962,6 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	/* enable GEM by default */
 	dev_priv->has_gem = 1;
 
-	if (dev_priv->has_gem == 0 &&
-	    drm_core_check_feature(dev, DRIVER_MODESET)) {
-		DRM_ERROR("kernel modesetting requires GEM, disabling driver.\n");
-		ret = -ENODEV;
-		goto out_workqueue_free;
-	}
-
 	dev->driver->get_vblank_counter = i915_get_vblank_counter;
 	dev->max_vblank_count = 0xffffff; /* only 24 bits of frame count */
 	if (IS_G4X(dev) || IS_GEN5(dev) || IS_GEN6(dev)) {
@@ -2055,7 +2048,6 @@ out_gem_unload:
 
 	intel_teardown_gmbus(dev);
 	intel_teardown_mchbar(dev);
-out_workqueue_free:
 	destroy_workqueue(dev_priv->wq);
 out_iomapfree:
 	io_mapping_free(dev_priv->mm.gtt_mapping);
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 872493331988..0de75a23f8e7 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -49,6 +49,9 @@ module_param_named(powersave, i915_powersave, int, 0600);
 unsigned int i915_lvds_downclock = 0;
 module_param_named(lvds_downclock, i915_lvds_downclock, int, 0400);
 
+bool i915_try_reset = true;
+module_param_named(reset, i915_try_reset, bool, 0600);
+
 static struct drm_driver driver;
 extern int intel_agp_enabled;
 
@@ -352,6 +355,9 @@ 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)
+			ironlake_enable_rc6(dev);
 	}
 
 	intel_opregion_init(dev);
@@ -475,6 +481,9 @@ int i915_reset(struct drm_device *dev, u8 flags)
 	bool need_display = true;
 	int ret;
 
+	if (!i915_try_reset)
+		return 0;
+
 	if (!mutex_trylock(&dev->struct_mutex))
 		return -EBUSY;
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index aac1bf332f75..385fc7ec39d3 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -172,20 +172,21 @@ struct drm_i915_error_state {
 		int page_count;
 		u32 gtt_offset;
 		u32 *pages[0];
-	} *ringbuffer, *batchbuffer[2];
+	} *ringbuffer, *batchbuffer[I915_NUM_RINGS];
 	struct drm_i915_error_buffer {
-		size_t size;
+		u32 size;
 		u32 name;
 		u32 seqno;
 		u32 gtt_offset;
 		u32 read_domains;
 		u32 write_domain;
-		u32 fence_reg;
+		s32 fence_reg:5;
 		s32 pinned:2;
 		u32 tiling:2;
 		u32 dirty:1;
 		u32 purgeable:1;
 		u32 ring:4;
+		u32 agp_type:1;
 	} *active_bo, *pinned_bo;
 	u32 active_bo_count, pinned_bo_count;
 	struct intel_overlay_error_state *overlay;
@@ -332,6 +333,7 @@ typedef struct drm_i915_private {
 
 	/* LVDS info */
 	int backlight_level;  /* restore backlight to this value */
+	bool backlight_enabled;
 	struct drm_display_mode *panel_fixed_mode;
 	struct drm_display_mode *lfp_lvds_vbt_mode; /* if any */
 	struct drm_display_mode *sdvo_lvds_vbt_mode; /* if any */
@@ -794,6 +796,7 @@ struct drm_i915_gem_object {
 	 */
 	struct hlist_node exec_node;
 	unsigned long exec_handle;
+	struct drm_i915_gem_exec_object2 *exec_entry;
 
 	/**
 	 * Current offset of the object in GTT space.
@@ -1006,12 +1009,6 @@ extern u32 i915_get_vblank_counter(struct drm_device *dev, int crtc);
 extern u32 gm45_get_vblank_counter(struct drm_device *dev, int crtc);
 extern int i915_vblank_swap(struct drm_device *dev, void *data,
 			    struct drm_file *file_priv);
-extern void i915_enable_irq(drm_i915_private_t *dev_priv, u32 mask);
-extern void i915_disable_irq(drm_i915_private_t *dev_priv, u32 mask);
-extern void ironlake_enable_graphics_irq(drm_i915_private_t *dev_priv,
-		u32 mask);
-extern void ironlake_disable_graphics_irq(drm_i915_private_t *dev_priv,
-		u32 mask);
 
 void
 i915_enable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask);
@@ -1091,10 +1088,10 @@ int i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *file_priv);
 void i915_gem_load(struct drm_device *dev);
 int i915_gem_init_object(struct drm_gem_object *obj);
-void i915_gem_flush_ring(struct drm_device *dev,
-			 struct intel_ring_buffer *ring,
-			 uint32_t invalidate_domains,
-			 uint32_t flush_domains);
+int __must_check i915_gem_flush_ring(struct drm_device *dev,
+				     struct intel_ring_buffer *ring,
+				     uint32_t invalidate_domains,
+				     uint32_t flush_domains);
 struct drm_i915_gem_object *i915_gem_alloc_object(struct drm_device *dev,
 						  size_t size);
 void i915_gem_free_object(struct drm_gem_object *obj);
@@ -1265,6 +1262,7 @@ extern void intel_disable_fbc(struct drm_device *dev);
 extern void intel_enable_fbc(struct drm_crtc *crtc, unsigned long interval);
 extern bool intel_fbc_enabled(struct drm_device *dev);
 extern bool ironlake_set_drps(struct drm_device *dev, u8 val);
+extern void ironlake_enable_rc6(struct drm_device *dev);
 extern void gen6_set_rps(struct drm_device *dev, u8 val);
 extern void intel_detect_pch (struct drm_device *dev);
 extern int intel_trans_dp_port_sel (struct drm_crtc *crtc);
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index c79c0b62ef60..3dfc848ff755 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -35,18 +35,18 @@
 #include <linux/swap.h>
 #include <linux/pci.h>
 
-static void i915_gem_object_flush_gpu_write_domain(struct drm_i915_gem_object *obj);
+static __must_check int i915_gem_object_flush_gpu_write_domain(struct drm_i915_gem_object *obj);
 static void i915_gem_object_flush_gtt_write_domain(struct drm_i915_gem_object *obj);
 static void i915_gem_object_flush_cpu_write_domain(struct drm_i915_gem_object *obj);
-static int i915_gem_object_set_to_cpu_domain(struct drm_i915_gem_object *obj,
-					     bool write);
-static int i915_gem_object_set_cpu_read_domain_range(struct drm_i915_gem_object *obj,
-						     uint64_t offset,
-						     uint64_t size);
+static __must_check int i915_gem_object_set_to_cpu_domain(struct drm_i915_gem_object *obj,
+							  bool write);
+static __must_check int i915_gem_object_set_cpu_read_domain_range(struct drm_i915_gem_object *obj,
+								  uint64_t offset,
+								  uint64_t size);
 static void i915_gem_object_set_to_full_cpu_read_domain(struct drm_i915_gem_object *obj);
-static int i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
-				       unsigned alignment,
-				       bool map_and_fenceable);
+static __must_check int i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
+						    unsigned alignment,
+						    bool map_and_fenceable);
 static void i915_gem_clear_fence_reg(struct drm_device *dev,
 				     struct drm_i915_fence_reg *reg);
 static int i915_gem_phys_pwrite(struct drm_device *dev,
@@ -1935,6 +1935,8 @@ i915_gem_retire_work_handler(struct work_struct *work)
 {
 	drm_i915_private_t *dev_priv;
 	struct drm_device *dev;
+	bool idle;
+	int i;
 
 	dev_priv = container_of(work, drm_i915_private_t,
 				mm.retire_work.work);
@@ -1948,11 +1950,31 @@ i915_gem_retire_work_handler(struct work_struct *work)
 
 	i915_gem_retire_requests(dev);
 
-	if (!dev_priv->mm.suspended &&
-		(!list_empty(&dev_priv->ring[RCS].request_list) ||
-		 !list_empty(&dev_priv->ring[VCS].request_list) ||
-		 !list_empty(&dev_priv->ring[BCS].request_list)))
+	/* Send a periodic flush down the ring so we don't hold onto GEM
+	 * objects indefinitely.
+	 */
+	idle = true;
+	for (i = 0; i < I915_NUM_RINGS; i++) {
+		struct intel_ring_buffer *ring = &dev_priv->ring[i];
+
+		if (!list_empty(&ring->gpu_write_list)) {
+			struct drm_i915_gem_request *request;
+			int ret;
+
+			ret = i915_gem_flush_ring(dev, ring, 0,
+						  I915_GEM_GPU_DOMAINS);
+			request = kzalloc(sizeof(*request), GFP_KERNEL);
+			if (ret || request == NULL ||
+			    i915_add_request(dev, NULL, request, ring))
+			    kfree(request);
+		}
+
+		idle &= list_empty(&ring->request_list);
+	}
+
+	if (!dev_priv->mm.suspended && !idle)
 		queue_delayed_work(dev_priv->wq, &dev_priv->mm.retire_work, HZ);
+
 	mutex_unlock(&dev->struct_mutex);
 }
 
@@ -2142,25 +2164,37 @@ i915_gem_object_unbind(struct drm_i915_gem_object *obj)
 	return ret;
 }
 
-void
+int
 i915_gem_flush_ring(struct drm_device *dev,
 		    struct intel_ring_buffer *ring,
 		    uint32_t invalidate_domains,
 		    uint32_t flush_domains)
 {
-	ring->flush(ring, invalidate_domains, flush_domains);
+	int ret;
+
+	ret = ring->flush(ring, invalidate_domains, flush_domains);
+	if (ret)
+		return ret;
+
 	i915_gem_process_flushing_list(dev, flush_domains, ring);
+	return 0;
 }
 
 static int i915_ring_idle(struct drm_device *dev,
 			  struct intel_ring_buffer *ring)
 {
+	int ret;
+
 	if (list_empty(&ring->gpu_write_list) && list_empty(&ring->active_list))
 		return 0;
 
-	if (!list_empty(&ring->gpu_write_list))
-		i915_gem_flush_ring(dev, ring,
+	if (!list_empty(&ring->gpu_write_list)) {
+		ret = i915_gem_flush_ring(dev, ring,
 				    I915_GEM_GPU_DOMAINS, I915_GEM_GPU_DOMAINS);
+		if (ret)
+			return ret;
+	}
+
 	return i915_wait_request(dev,
 				 i915_gem_next_request_seqno(dev, ring),
 				 ring);
@@ -2370,10 +2404,13 @@ i915_gem_object_flush_fence(struct drm_i915_gem_object *obj,
 	int ret;
 
 	if (obj->fenced_gpu_access) {
-		if (obj->base.write_domain & I915_GEM_GPU_DOMAINS)
-			i915_gem_flush_ring(obj->base.dev,
-					    obj->last_fenced_ring,
-					    0, obj->base.write_domain);
+		if (obj->base.write_domain & I915_GEM_GPU_DOMAINS) {
+			ret = i915_gem_flush_ring(obj->base.dev,
+						  obj->last_fenced_ring,
+						  0, obj->base.write_domain);
+			if (ret)
+				return ret;
+		}
 
 		obj->fenced_gpu_access = false;
 	}
@@ -2393,6 +2430,12 @@ i915_gem_object_flush_fence(struct drm_i915_gem_object *obj,
 		obj->last_fenced_ring = NULL;
 	}
 
+	/* Ensure that all CPU reads are completed before installing a fence
+	 * and all writes before removing the fence.
+	 */
+	if (obj->base.read_domains & I915_GEM_DOMAIN_GTT)
+		mb();
+
 	return 0;
 }
 
@@ -2523,9 +2566,12 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj,
 				return ret;
 		} else if (obj->tiling_changed) {
 			if (obj->fenced_gpu_access) {
-				if (obj->base.write_domain & I915_GEM_GPU_DOMAINS)
-					i915_gem_flush_ring(obj->base.dev, obj->ring,
-							    0, obj->base.write_domain);
+				if (obj->base.write_domain & I915_GEM_GPU_DOMAINS) {
+					ret = i915_gem_flush_ring(obj->base.dev, obj->ring,
+								  0, obj->base.write_domain);
+					if (ret)
+						return ret;
+				}
 
 				obj->fenced_gpu_access = false;
 			}
@@ -2736,10 +2782,8 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 		obj->gtt_space = NULL;
 
 		if (ret == -ENOMEM) {
-			/* first try to clear up some space from the GTT */
-			ret = i915_gem_evict_something(dev, size,
-						       alignment,
-						       map_and_fenceable);
+			/* first try to reclaim some memory by clearing the GTT */
+			ret = i915_gem_evict_everything(dev, false);
 			if (ret) {
 				/* now try to shrink everyone else */
 				if (gfpmask) {
@@ -2747,7 +2791,7 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 					goto search_free;
 				}
 
-				return ret;
+				return -ENOMEM;
 			}
 
 			goto search_free;
@@ -2762,9 +2806,7 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 		drm_mm_put_block(obj->gtt_space);
 		obj->gtt_space = NULL;
 
-		ret = i915_gem_evict_something(dev, size,
-					       alignment, map_and_fenceable);
-		if (ret)
+		if (i915_gem_evict_everything(dev, false))
 			return ret;
 
 		goto search_free;
@@ -2811,17 +2853,16 @@ i915_gem_clflush_object(struct drm_i915_gem_object *obj)
 }
 
 /** Flushes any GPU write domain for the object if it's dirty. */
-static void
+static int
 i915_gem_object_flush_gpu_write_domain(struct drm_i915_gem_object *obj)
 {
 	struct drm_device *dev = obj->base.dev;
 
 	if ((obj->base.write_domain & I915_GEM_GPU_DOMAINS) == 0)
-		return;
+		return 0;
 
 	/* Queue the GPU write cache flushing we need. */
-	i915_gem_flush_ring(dev, obj->ring, 0, obj->base.write_domain);
-	BUG_ON(obj->base.write_domain);
+	return i915_gem_flush_ring(dev, obj->ring, 0, obj->base.write_domain);
 }
 
 /** Flushes the GTT write domain for the object if it's dirty. */
@@ -2833,10 +2874,16 @@ i915_gem_object_flush_gtt_write_domain(struct drm_i915_gem_object *obj)
 	if (obj->base.write_domain != I915_GEM_DOMAIN_GTT)
 		return;
 
-	/* No actual flushing is required for the GTT write domain.   Writes
+	/* No actual flushing is required for the GTT write domain.  Writes
 	 * to it immediately go to main memory as far as we know, so there's
 	 * no chipset flush.  It also doesn't land in render cache.
+	 *
+	 * However, we do have to enforce the order so that all writes through
+	 * the GTT land before any writes to the device, such as updates to
+	 * the GATT itself.
 	 */
+	wmb();
+
 	i915_gem_release_mmap(obj);
 
 	old_write_domain = obj->base.write_domain;
@@ -2882,7 +2929,10 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
 	if (obj->gtt_space == NULL)
 		return -EINVAL;
 
-	i915_gem_object_flush_gpu_write_domain(obj);
+	ret = i915_gem_object_flush_gpu_write_domain(obj);
+	if (ret)
+		return ret;
+
 	if (obj->pending_gpu_write || write) {
 		ret = i915_gem_object_wait_rendering(obj, true);
 		if (ret)
@@ -2927,7 +2977,10 @@ i915_gem_object_set_to_display_plane(struct drm_i915_gem_object *obj,
 	if (obj->gtt_space == NULL)
 		return -EINVAL;
 
-	i915_gem_object_flush_gpu_write_domain(obj);
+	ret = i915_gem_object_flush_gpu_write_domain(obj);
+	if (ret)
+		return ret;
+
 
 	/* Currently, we are always called from an non-interruptible context. */
 	if (pipelined != obj->ring) {
@@ -2952,12 +3005,17 @@ int
 i915_gem_object_flush_gpu(struct drm_i915_gem_object *obj,
 			  bool interruptible)
 {
+	int ret;
+
 	if (!obj->active)
 		return 0;
 
-	if (obj->base.write_domain & I915_GEM_GPU_DOMAINS)
-		i915_gem_flush_ring(obj->base.dev, obj->ring,
-				    0, obj->base.write_domain);
+	if (obj->base.write_domain & I915_GEM_GPU_DOMAINS) {
+		ret = i915_gem_flush_ring(obj->base.dev, obj->ring,
+					  0, obj->base.write_domain);
+		if (ret)
+			return ret;
+	}
 
 	return i915_gem_object_wait_rendering(obj, interruptible);
 }
@@ -2974,7 +3032,10 @@ i915_gem_object_set_to_cpu_domain(struct drm_i915_gem_object *obj, bool write)
 	uint32_t old_write_domain, old_read_domains;
 	int ret;
 
-	i915_gem_object_flush_gpu_write_domain(obj);
+	ret = i915_gem_object_flush_gpu_write_domain(obj);
+	if (ret)
+		return ret;
+
 	ret = i915_gem_object_wait_rendering(obj, true);
 	if (ret)
 		return ret;
@@ -3069,7 +3130,10 @@ i915_gem_object_set_cpu_read_domain_range(struct drm_i915_gem_object *obj,
 	if (offset == 0 && size == obj->base.size)
 		return i915_gem_object_set_to_cpu_domain(obj, 0);
 
-	i915_gem_object_flush_gpu_write_domain(obj);
+	ret = i915_gem_object_flush_gpu_write_domain(obj);
+	if (ret)
+		return ret;
+
 	ret = i915_gem_object_wait_rendering(obj, true);
 	if (ret)
 		return ret;
@@ -3362,8 +3426,8 @@ i915_gem_busy_ioctl(struct drm_device *dev, void *data,
 		 * flush earlier is beneficial.
 		 */
 		if (obj->base.write_domain & I915_GEM_GPU_DOMAINS) {
-			i915_gem_flush_ring(dev, obj->ring,
-					    0, obj->base.write_domain);
+			ret = i915_gem_flush_ring(dev, obj->ring,
+						  0, obj->base.write_domain);
 		} else if (obj->ring->outstanding_lazy_request ==
 			   obj->last_rendering_seqno) {
 			struct drm_i915_gem_request *request;
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
index 78b8cf90c922..3d39005540aa 100644
--- a/drivers/gpu/drm/i915/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -127,9 +127,15 @@ i915_gem_evict_something(struct drm_device *dev, int min_size,
 	}
 
 	/* Nothing found, clean up and bail out! */
-	list_for_each_entry(obj, &unwind_list, exec_list) {
+	while (!list_empty(&unwind_list)) {
+		obj = list_first_entry(&unwind_list,
+				       struct drm_i915_gem_object,
+				       exec_list);
+
 		ret = drm_mm_scan_remove_block(obj->gtt_space);
 		BUG_ON(ret);
+
+		list_del_init(&obj->exec_list);
 		drm_gem_object_unreference(&obj->base);
 	}
 
@@ -162,6 +168,7 @@ found:
 				       exec_list);
 		if (ret == 0)
 			ret = i915_gem_object_unbind(obj);
+
 		list_del_init(&obj->exec_list);
 		drm_gem_object_unreference(&obj->base);
 	}
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 61129e6759eb..e69834341ef0 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -268,7 +268,6 @@ eb_destroy(struct eb_objects *eb)
 static int
 i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 				   struct eb_objects *eb,
-				   struct drm_i915_gem_exec_object2 *entry,
 				   struct drm_i915_gem_relocation_entry *reloc)
 {
 	struct drm_device *dev = obj->base.dev;
@@ -411,10 +410,10 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 
 static int
 i915_gem_execbuffer_relocate_object(struct drm_i915_gem_object *obj,
-				    struct eb_objects *eb,
-				    struct drm_i915_gem_exec_object2 *entry)
+				    struct eb_objects *eb)
 {
 	struct drm_i915_gem_relocation_entry __user *user_relocs;
+	struct drm_i915_gem_exec_object2 *entry = obj->exec_entry;
 	int i, ret;
 
 	user_relocs = (void __user *)(uintptr_t)entry->relocs_ptr;
@@ -426,7 +425,7 @@ i915_gem_execbuffer_relocate_object(struct drm_i915_gem_object *obj,
 					      sizeof(reloc)))
 			return -EFAULT;
 
-		ret = i915_gem_execbuffer_relocate_entry(obj, eb, entry, &reloc);
+		ret = i915_gem_execbuffer_relocate_entry(obj, eb, &reloc);
 		if (ret)
 			return ret;
 
@@ -442,13 +441,13 @@ i915_gem_execbuffer_relocate_object(struct drm_i915_gem_object *obj,
 static int
 i915_gem_execbuffer_relocate_object_slow(struct drm_i915_gem_object *obj,
 					 struct eb_objects *eb,
-					 struct drm_i915_gem_exec_object2 *entry,
 					 struct drm_i915_gem_relocation_entry *relocs)
 {
+	const struct drm_i915_gem_exec_object2 *entry = obj->exec_entry;
 	int i, ret;
 
 	for (i = 0; i < entry->relocation_count; i++) {
-		ret = i915_gem_execbuffer_relocate_entry(obj, eb, entry, &relocs[i]);
+		ret = i915_gem_execbuffer_relocate_entry(obj, eb, &relocs[i]);
 		if (ret)
 			return ret;
 	}
@@ -459,8 +458,7 @@ i915_gem_execbuffer_relocate_object_slow(struct drm_i915_gem_object *obj,
 static int
 i915_gem_execbuffer_relocate(struct drm_device *dev,
 			     struct eb_objects *eb,
-			     struct list_head *objects,
-			     struct drm_i915_gem_exec_object2 *exec)
+			     struct list_head *objects)
 {
 	struct drm_i915_gem_object *obj;
 	int ret;
@@ -468,7 +466,7 @@ i915_gem_execbuffer_relocate(struct drm_device *dev,
 	list_for_each_entry(obj, objects, exec_list) {
 		obj->base.pending_read_domains = 0;
 		obj->base.pending_write_domain = 0;
-		ret = i915_gem_execbuffer_relocate_object(obj, eb, exec++);
+		ret = i915_gem_execbuffer_relocate_object(obj, eb);
 		if (ret)
 			return ret;
 	}
@@ -479,13 +477,36 @@ i915_gem_execbuffer_relocate(struct drm_device *dev,
 static int
 i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 			    struct drm_file *file,
-			    struct list_head *objects,
-			    struct drm_i915_gem_exec_object2 *exec)
+			    struct list_head *objects)
 {
 	struct drm_i915_gem_object *obj;
-	struct drm_i915_gem_exec_object2 *entry;
 	int ret, retry;
 	bool has_fenced_gpu_access = INTEL_INFO(ring->dev)->gen < 4;
+	struct list_head ordered_objects;
+
+	INIT_LIST_HEAD(&ordered_objects);
+	while (!list_empty(objects)) {
+		struct drm_i915_gem_exec_object2 *entry;
+		bool need_fence, need_mappable;
+
+		obj = list_first_entry(objects,
+				       struct drm_i915_gem_object,
+				       exec_list);
+		entry = obj->exec_entry;
+
+		need_fence =
+			has_fenced_gpu_access &&
+			entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
+			obj->tiling_mode != I915_TILING_NONE;
+		need_mappable =
+			entry->relocation_count ? true : need_fence;
+
+		if (need_mappable)
+			list_move(&obj->exec_list, &ordered_objects);
+		else
+			list_move_tail(&obj->exec_list, &ordered_objects);
+	}
+	list_splice(&ordered_objects, objects);
 
 	/* Attempt to pin all of the buffers into the GTT.
 	 * This is done in 3 phases:
@@ -504,14 +525,11 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 		ret = 0;
 
 		/* Unbind any ill-fitting objects or pin. */
-		entry = exec;
 		list_for_each_entry(obj, objects, exec_list) {
+			struct drm_i915_gem_exec_object2 *entry = obj->exec_entry;
 			bool need_fence, need_mappable;
-
-			if (!obj->gtt_space) {
-				entry++;
+			if (!obj->gtt_space)
 				continue;
-			}
 
 			need_fence =
 				has_fenced_gpu_access &&
@@ -534,8 +552,8 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 		}
 
 		/* Bind fresh objects */
-		entry = exec;
 		list_for_each_entry(obj, objects, exec_list) {
+			struct drm_i915_gem_exec_object2 *entry = obj->exec_entry;
 			bool need_fence;
 
 			need_fence =
@@ -570,7 +588,6 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 			}
 
 			entry->offset = obj->gtt_offset;
-			entry++;
 		}
 
 		/* Decrement pin count for bound objects */
@@ -622,7 +639,7 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 	int i, total, ret;
 
 	/* We may process another execbuffer during the unlock... */
-	while (list_empty(objects)) {
+	while (!list_empty(objects)) {
 		obj = list_first_entry(objects,
 				       struct drm_i915_gem_object,
 				       exec_list);
@@ -665,7 +682,6 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 	}
 
 	/* reacquire the objects */
-	INIT_LIST_HEAD(objects);
 	eb_reset(eb);
 	for (i = 0; i < count; i++) {
 		struct drm_i915_gem_object *obj;
@@ -681,10 +697,11 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 
 		list_add_tail(&obj->exec_list, objects);
 		obj->exec_handle = exec[i].handle;
+		obj->exec_entry = &exec[i];
 		eb_add_object(eb, obj);
 	}
 
-	ret = i915_gem_execbuffer_reserve(ring, file, objects, exec);
+	ret = i915_gem_execbuffer_reserve(ring, file, objects);
 	if (ret)
 		goto err;
 
@@ -693,7 +710,6 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 		obj->base.pending_read_domains = 0;
 		obj->base.pending_write_domain = 0;
 		ret = i915_gem_execbuffer_relocate_object_slow(obj, eb,
-							       exec,
 							       reloc + total);
 		if (ret)
 			goto err;
@@ -713,25 +729,34 @@ err:
 	return ret;
 }
 
-static void
+static int
 i915_gem_execbuffer_flush(struct drm_device *dev,
 			  uint32_t invalidate_domains,
 			  uint32_t flush_domains,
 			  uint32_t flush_rings)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	int i;
+	int i, ret;
 
 	if (flush_domains & I915_GEM_DOMAIN_CPU)
 		intel_gtt_chipset_flush();
 
+	if (flush_domains & I915_GEM_DOMAIN_GTT)
+		wmb();
+
 	if ((flush_domains | invalidate_domains) & I915_GEM_GPU_DOMAINS) {
 		for (i = 0; i < I915_NUM_RINGS; i++)
-			if (flush_rings & (1 << i))
-				i915_gem_flush_ring(dev, &dev_priv->ring[i],
-						    invalidate_domains,
-						    flush_domains);
+			if (flush_rings & (1 << i)) {
+				ret = i915_gem_flush_ring(dev,
+							  &dev_priv->ring[i],
+							  invalidate_domains,
+							  flush_domains);
+				if (ret)
+					return ret;
+			}
 	}
+
+	return 0;
 }
 
 static int
@@ -795,10 +820,12 @@ i915_gem_execbuffer_move_to_gpu(struct intel_ring_buffer *ring,
 			 cd.invalidate_domains,
 			 cd.flush_domains);
 #endif
-		i915_gem_execbuffer_flush(ring->dev,
-					  cd.invalidate_domains,
-					  cd.flush_domains,
-					  cd.flush_rings);
+		ret = i915_gem_execbuffer_flush(ring->dev,
+						cd.invalidate_domains,
+						cd.flush_domains,
+						cd.flush_rings);
+		if (ret)
+			return ret;
 	}
 
 	list_for_each_entry(obj, objects, exec_list) {
@@ -921,7 +948,7 @@ i915_gem_execbuffer_retire_commands(struct drm_device *dev,
 				    struct intel_ring_buffer *ring)
 {
 	struct drm_i915_gem_request *request;
-	u32 flush_domains;
+	u32 invalidate;
 
 	/*
 	 * Ensure that the commands in the batch buffer are
@@ -929,11 +956,13 @@ i915_gem_execbuffer_retire_commands(struct drm_device *dev,
 	 *
 	 * The sampler always gets flushed on i965 (sigh).
 	 */
-	flush_domains = 0;
+	invalidate = I915_GEM_DOMAIN_COMMAND;
 	if (INTEL_INFO(dev)->gen >= 4)
-		flush_domains |= I915_GEM_DOMAIN_SAMPLER;
-
-	ring->flush(ring, I915_GEM_DOMAIN_COMMAND, flush_domains);
+		invalidate |= I915_GEM_DOMAIN_SAMPLER;
+	if (ring->flush(ring, invalidate, 0)) {
+		i915_gem_next_request_seqno(dev, ring);
+		return;
+	}
 
 	/* Add a breadcrumb for the completion of the batch buffer */
 	request = kzalloc(sizeof(*request), GFP_KERNEL);
@@ -1098,16 +1127,22 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 
 		list_add_tail(&obj->exec_list, &objects);
 		obj->exec_handle = exec[i].handle;
+		obj->exec_entry = &exec[i];
 		eb_add_object(eb, obj);
 	}
 
+	/* take note of the batch buffer before we might reorder the lists */
+	batch_obj = list_entry(objects.prev,
+			       struct drm_i915_gem_object,
+			       exec_list);
+
 	/* Move the objects en-masse into the GTT, evicting if necessary. */
-	ret = i915_gem_execbuffer_reserve(ring, file, &objects, exec);
+	ret = i915_gem_execbuffer_reserve(ring, file, &objects);
 	if (ret)
 		goto err;
 
 	/* The objects are in their final locations, apply the relocations. */
-	ret = i915_gem_execbuffer_relocate(dev, eb, &objects, exec);
+	ret = i915_gem_execbuffer_relocate(dev, eb, &objects);
 	if (ret) {
 		if (ret == -EFAULT) {
 			ret = i915_gem_execbuffer_relocate_slow(dev, file, ring,
@@ -1121,9 +1156,6 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	}
 
 	/* Set the pending read domains for the batch buffer to COMMAND */
-	batch_obj = list_entry(objects.prev,
-			       struct drm_i915_gem_object,
-			       exec_list);
 	if (batch_obj->base.pending_write_domain) {
 		DRM_ERROR("Attempting to use self-modifying batch buffer\n");
 		ret = -EINVAL;
@@ -1340,4 +1372,3 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 	drm_free_large(exec2_list);
 	return ret;
 }
-
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 86673e77d7cb..70433ae50ac8 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -85,15 +85,11 @@ int i915_gem_gtt_bind_object(struct drm_i915_gem_object *obj)
 
 void i915_gem_gtt_unbind_object(struct drm_i915_gem_object *obj)
 {
-	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	intel_gtt_clear_range(obj->gtt_space->start >> PAGE_SHIFT,
+			      obj->base.size >> PAGE_SHIFT);
 
-	if (dev_priv->mm.gtt->needs_dmar) {
+	if (obj->sg_list) {
 		intel_gtt_unmap_memory(obj->sg_list, obj->num_sg);
 		obj->sg_list = NULL;
-		obj->num_sg = 0;
 	}
-
-	intel_gtt_clear_range(obj->gtt_space->start >> PAGE_SHIFT,
-			      obj->base.size >> PAGE_SHIFT);
 }
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 0dadc025b77b..e418e8bb61e6 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -64,26 +64,6 @@
 #define DRM_I915_VBLANK_PIPE_ALL	(DRM_I915_VBLANK_PIPE_A | \
 					 DRM_I915_VBLANK_PIPE_B)
 
-void
-ironlake_enable_graphics_irq(drm_i915_private_t *dev_priv, u32 mask)
-{
-	if ((dev_priv->gt_irq_mask & mask) != 0) {
-		dev_priv->gt_irq_mask &= ~mask;
-		I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
-		POSTING_READ(GTIMR);
-	}
-}
-
-void
-ironlake_disable_graphics_irq(drm_i915_private_t *dev_priv, u32 mask)
-{
-	if ((dev_priv->gt_irq_mask & mask) != mask) {
-		dev_priv->gt_irq_mask |= mask;
-		I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
-		POSTING_READ(GTIMR);
-	}
-}
-
 /* For display hotplug interrupt */
 static void
 ironlake_enable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
@@ -105,26 +85,6 @@ ironlake_disable_display_irq(drm_i915_private_t *dev_priv, u32 mask)
 	}
 }
 
-void
-i915_enable_irq(drm_i915_private_t *dev_priv, u32 mask)
-{
-	if ((dev_priv->irq_mask & mask) != 0) {
-		dev_priv->irq_mask &= ~mask;
-		I915_WRITE(IMR, dev_priv->irq_mask);
-		POSTING_READ(IMR);
-	}
-}
-
-void
-i915_disable_irq(drm_i915_private_t *dev_priv, u32 mask)
-{
-	if ((dev_priv->irq_mask & mask) != mask) {
-		dev_priv->irq_mask |= mask;
-		I915_WRITE(IMR, dev_priv->irq_mask);
-		POSTING_READ(IMR);
-	}
-}
-
 static inline u32
 i915_pipestat(int pipe)
 {
@@ -389,9 +349,12 @@ static void notify_ring(struct drm_device *dev,
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 seqno = ring->get_seqno(ring);
-	ring->irq_seqno = seqno;
+
 	trace_i915_gem_request_complete(dev, seqno);
+
+	ring->irq_seqno = seqno;
 	wake_up_all(&ring->irq_queue);
+
 	dev_priv->hangcheck_count = 0;
 	mod_timer(&dev_priv->hangcheck_timer,
 		  jiffies + msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD));
@@ -435,6 +398,50 @@ static void gen6_pm_irq_handler(struct drm_device *dev)
 	I915_WRITE(GEN6_PMIIR, pm_iir);
 }
 
+static void pch_irq_handler(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
+	u32 pch_iir;
+
+	pch_iir = I915_READ(SDEIIR);
+
+	if (pch_iir & SDE_AUDIO_POWER_MASK)
+		DRM_DEBUG_DRIVER("PCH audio power change on port %d\n",
+				 (pch_iir & SDE_AUDIO_POWER_MASK) >>
+				 SDE_AUDIO_POWER_SHIFT);
+
+	if (pch_iir & SDE_GMBUS)
+		DRM_DEBUG_DRIVER("PCH GMBUS interrupt\n");
+
+	if (pch_iir & SDE_AUDIO_HDCP_MASK)
+		DRM_DEBUG_DRIVER("PCH HDCP audio interrupt\n");
+
+	if (pch_iir & SDE_AUDIO_TRANS_MASK)
+		DRM_DEBUG_DRIVER("PCH transcoder audio interrupt\n");
+
+	if (pch_iir & SDE_POISON)
+		DRM_ERROR("PCH poison interrupt\n");
+
+	if (pch_iir & SDE_FDI_MASK) {
+		u32 fdia, fdib;
+
+		fdia = I915_READ(FDI_RXA_IIR);
+		fdib = I915_READ(FDI_RXB_IIR);
+		DRM_DEBUG_DRIVER("PCH FDI RX interrupt; FDI RXA IIR: 0x%08x, FDI RXB IIR: 0x%08x\n", fdia, fdib);
+	}
+
+	if (pch_iir & (SDE_TRANSB_CRC_DONE | SDE_TRANSA_CRC_DONE))
+		DRM_DEBUG_DRIVER("PCH transcoder CRC done interrupt\n");
+
+	if (pch_iir & (SDE_TRANSB_CRC_ERR | SDE_TRANSA_CRC_ERR))
+		DRM_DEBUG_DRIVER("PCH transcoder CRC error interrupt\n");
+
+	if (pch_iir & SDE_TRANSB_FIFO_UNDER)
+		DRM_DEBUG_DRIVER("PCH transcoder B underrun interrupt\n");
+	if (pch_iir & SDE_TRANSA_FIFO_UNDER)
+		DRM_DEBUG_DRIVER("PCH transcoder A underrun interrupt\n");
+}
+
 static irqreturn_t ironlake_irq_handler(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
@@ -502,8 +509,11 @@ static irqreturn_t ironlake_irq_handler(struct drm_device *dev)
 		drm_handle_vblank(dev, 1);
 
 	/* check event from PCH */
-	if ((de_iir & DE_PCH_EVENT) && (pch_iir & hotplug_mask))
-		queue_work(dev_priv->wq, &dev_priv->hotplug_work);
+	if (de_iir & DE_PCH_EVENT) {
+		if (pch_iir & hotplug_mask)
+			queue_work(dev_priv->wq, &dev_priv->hotplug_work);
+		pch_irq_handler(dev);
+	}
 
 	if (de_iir & DE_PCU_EVENT) {
 		I915_WRITE16(MEMINTRSTS, I915_READ(MEMINTRSTS));
@@ -556,10 +566,9 @@ static void i915_error_work_func(struct work_struct *work)
 
 #ifdef CONFIG_DEBUG_FS
 static struct drm_i915_error_object *
-i915_error_object_create(struct drm_device *dev,
+i915_error_object_create(struct drm_i915_private *dev_priv,
 			 struct drm_i915_gem_object *src)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct drm_i915_error_object *dst;
 	int page, page_count;
 	u32 reloc_offset;
@@ -632,52 +641,6 @@ i915_error_state_free(struct drm_device *dev,
 	kfree(error);
 }
 
-static u32
-i915_get_bbaddr(struct drm_device *dev, u32 *ring)
-{
-	u32 cmd;
-
-	if (IS_I830(dev) || IS_845G(dev))
-		cmd = MI_BATCH_BUFFER;
-	else if (INTEL_INFO(dev)->gen >= 4)
-		cmd = (MI_BATCH_BUFFER_START | (2 << 6) |
-		       MI_BATCH_NON_SECURE_I965);
-	else
-		cmd = (MI_BATCH_BUFFER_START | (2 << 6));
-
-	return ring[0] == cmd ? ring[1] : 0;
-}
-
-static u32
-i915_ringbuffer_last_batch(struct drm_device *dev,
-			   struct intel_ring_buffer *ring)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 head, bbaddr;
-	u32 *val;
-
-	/* Locate the current position in the ringbuffer and walk back
-	 * to find the most recently dispatched batch buffer.
-	 */
-	head = I915_READ_HEAD(ring) & HEAD_ADDR;
-
-	val = (u32 *)(ring->virtual_start + head);
-	while (--val >= (u32 *)ring->virtual_start) {
-		bbaddr = i915_get_bbaddr(dev, val);
-		if (bbaddr)
-			return bbaddr;
-	}
-
-	val = (u32 *)(ring->virtual_start + ring->size);
-	while (--val >= (u32 *)ring->virtual_start) {
-		bbaddr = i915_get_bbaddr(dev, val);
-		if (bbaddr)
-			return bbaddr;
-	}
-
-	return 0;
-}
-
 static u32 capture_bo_list(struct drm_i915_error_buffer *err,
 			   int count,
 			   struct list_head *head)
@@ -702,6 +665,7 @@ static u32 capture_bo_list(struct drm_i915_error_buffer *err,
 		err->dirty = obj->dirty;
 		err->purgeable = obj->madv != I915_MADV_WILLNEED;
 		err->ring = obj->ring ? obj->ring->id : 0;
+		err->agp_type = obj->agp_type == AGP_USER_CACHED_MEMORY;
 
 		if (++i == count)
 			break;
@@ -741,6 +705,36 @@ static void i915_gem_record_fences(struct drm_device *dev,
 	}
 }
 
+static struct drm_i915_error_object *
+i915_error_first_batchbuffer(struct drm_i915_private *dev_priv,
+			     struct intel_ring_buffer *ring)
+{
+	struct drm_i915_gem_object *obj;
+	u32 seqno;
+
+	if (!ring->get_seqno)
+		return NULL;
+
+	seqno = ring->get_seqno(ring);
+	list_for_each_entry(obj, &dev_priv->mm.active_list, mm_list) {
+		if (obj->ring != ring)
+			continue;
+
+		if (!i915_seqno_passed(obj->last_rendering_seqno, seqno))
+			continue;
+
+		if ((obj->base.read_domains & I915_GEM_DOMAIN_COMMAND) == 0)
+			continue;
+
+		/* We need to copy these to an anonymous buffer as the simplest
+		 * method to avoid being overwritten by userspace.
+		 */
+		return i915_error_object_create(dev_priv, obj);
+	}
+
+	return NULL;
+}
+
 /**
  * i915_capture_error_state - capture an error record for later analysis
  * @dev: drm device
@@ -755,10 +749,8 @@ static void i915_capture_error_state(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_gem_object *obj;
 	struct drm_i915_error_state *error;
-	struct drm_i915_gem_object *batchbuffer[2];
 	unsigned long flags;
-	u32 bbaddr;
-	int count;
+	int i;
 
 	spin_lock_irqsave(&dev_priv->error_lock, flags);
 	error = dev_priv->first_error;
@@ -817,83 +809,30 @@ static void i915_capture_error_state(struct drm_device *dev)
 	}
 	i915_gem_record_fences(dev, error);
 
-	bbaddr = i915_ringbuffer_last_batch(dev, &dev_priv->ring[RCS]);
-
-	/* Grab the current batchbuffer, most likely to have crashed. */
-	batchbuffer[0] = NULL;
-	batchbuffer[1] = NULL;
-	count = 0;
-	list_for_each_entry(obj, &dev_priv->mm.active_list, mm_list) {
-		if (batchbuffer[0] == NULL &&
-		    bbaddr >= obj->gtt_offset &&
-		    bbaddr < obj->gtt_offset + obj->base.size)
-			batchbuffer[0] = obj;
-
-		if (batchbuffer[1] == NULL &&
-		    error->acthd >= obj->gtt_offset &&
-		    error->acthd < obj->gtt_offset + obj->base.size)
-			batchbuffer[1] = obj;
-
-		count++;
-	}
-	/* Scan the other lists for completeness for those bizarre errors. */
-	if (batchbuffer[0] == NULL || batchbuffer[1] == NULL) {
-		list_for_each_entry(obj, &dev_priv->mm.flushing_list, mm_list) {
-			if (batchbuffer[0] == NULL &&
-			    bbaddr >= obj->gtt_offset &&
-			    bbaddr < obj->gtt_offset + obj->base.size)
-				batchbuffer[0] = obj;
-
-			if (batchbuffer[1] == NULL &&
-			    error->acthd >= obj->gtt_offset &&
-			    error->acthd < obj->gtt_offset + obj->base.size)
-				batchbuffer[1] = obj;
-
-			if (batchbuffer[0] && batchbuffer[1])
-				break;
-		}
-	}
-	if (batchbuffer[0] == NULL || batchbuffer[1] == NULL) {
-		list_for_each_entry(obj, &dev_priv->mm.inactive_list, mm_list) {
-			if (batchbuffer[0] == NULL &&
-			    bbaddr >= obj->gtt_offset &&
-			    bbaddr < obj->gtt_offset + obj->base.size)
-				batchbuffer[0] = obj;
-
-			if (batchbuffer[1] == NULL &&
-			    error->acthd >= obj->gtt_offset &&
-			    error->acthd < obj->gtt_offset + obj->base.size)
-				batchbuffer[1] = obj;
-
-			if (batchbuffer[0] && batchbuffer[1])
-				break;
-		}
-	}
-
-	/* We need to copy these to an anonymous buffer as the simplest
-	 * method to avoid being overwritten by userspace.
-	 */
-	error->batchbuffer[0] = i915_error_object_create(dev, batchbuffer[0]);
-	if (batchbuffer[1] != batchbuffer[0])
-		error->batchbuffer[1] = i915_error_object_create(dev, batchbuffer[1]);
-	else
-		error->batchbuffer[1] = NULL;
+	/* Record the active batchbuffers */
+	for (i = 0; i < I915_NUM_RINGS; i++)
+		error->batchbuffer[i] =
+			i915_error_first_batchbuffer(dev_priv,
+						     &dev_priv->ring[i]);
 
 	/* Record the ringbuffer */
-	error->ringbuffer = i915_error_object_create(dev,
+	error->ringbuffer = i915_error_object_create(dev_priv,
 						     dev_priv->ring[RCS].obj);
 
 	/* Record buffers on the active and pinned lists. */
 	error->active_bo = NULL;
 	error->pinned_bo = NULL;
 
-	error->active_bo_count = count;
+	i = 0;
+	list_for_each_entry(obj, &dev_priv->mm.active_list, mm_list)
+		i++;
+	error->active_bo_count = i;
 	list_for_each_entry(obj, &dev_priv->mm.pinned_list, mm_list)
-		count++;
-	error->pinned_bo_count = count - error->active_bo_count;
+		i++;
+	error->pinned_bo_count = i - error->active_bo_count;
 
-	if (count) {
-		error->active_bo = kmalloc(sizeof(*error->active_bo)*count,
+	if (i) {
+		error->active_bo = kmalloc(sizeof(*error->active_bo)*i,
 					   GFP_ATOMIC);
 		if (error->active_bo)
 			error->pinned_bo =
@@ -1673,11 +1612,6 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 
 	I915_WRITE(GTIIR, I915_READ(GTIIR));
 	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
-	if (IS_GEN6(dev)) {
-		I915_WRITE(GEN6_RENDER_IMR, ~GEN6_RENDER_USER_INTERRUPT);
-		I915_WRITE(GEN6_BSD_IMR, ~GEN6_BSD_USER_INTERRUPT);
-		I915_WRITE(GEN6_BLITTER_IMR, ~GEN6_BLITTER_USER_INTERRUPT);
-	}
 
 	if (IS_GEN6(dev))
 		render_irqs =
@@ -1698,6 +1632,9 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 	} else {
 		hotplug_mask = SDE_CRT_HOTPLUG | SDE_PORTB_HOTPLUG |
 			       SDE_PORTC_HOTPLUG | SDE_PORTD_HOTPLUG;
+		hotplug_mask |= SDE_AUX_MASK | SDE_FDI_MASK | SDE_TRANS_MASK;
+		I915_WRITE(FDI_RXA_IMR, 0);
+		I915_WRITE(FDI_RXB_IMR, 0);
 	}
 
 	dev_priv->pch_irq_mask = ~hotplug_mask;
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 8f948a6fbc1c..40a407f41f61 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -145,6 +145,8 @@
 #define   MI_END_SCENE		(1 << 4) /* flush binner and incr scene count */
 #define   MI_INVALIDATE_ISP	(1 << 5) /* invalidate indirect state pointers */
 #define MI_BATCH_BUFFER_END	MI_INSTR(0x0a, 0)
+#define MI_SUSPEND_FLUSH	MI_INSTR(0x0b, 0)
+#define   MI_SUSPEND_FLUSH_EN	(1<<0)
 #define MI_REPORT_HEAD		MI_INSTR(0x07, 0)
 #define MI_OVERLAY_FLIP		MI_INSTR(0x11,0)
 #define   MI_OVERLAY_CONTINUE	(0x0<<21)
@@ -159,6 +161,7 @@
 #define   MI_MM_SPACE_PHYSICAL		(0<<8)
 #define   MI_SAVE_EXT_STATE_EN		(1<<3)
 #define   MI_RESTORE_EXT_STATE_EN	(1<<2)
+#define   MI_FORCE_RESTORE		(1<<1)
 #define   MI_RESTORE_INHIBIT		(1<<0)
 #define MI_STORE_DWORD_IMM	MI_INSTR(0x20, 1)
 #define   MI_MEM_VIRTUAL	(1 << 22) /* 965+ only */
@@ -288,6 +291,7 @@
 #define RING_HWS_PGA_GEN6(base)	((base)+0x2080)
 #define RING_ACTHD(base)	((base)+0x74)
 #define RING_NOPID(base)	((base)+0x94)
+#define RING_IMR(base)		((base)+0xa8)
 #define   TAIL_ADDR		0x001FFFF8
 #define   HEAD_WRAP_COUNT	0xFFE00000
 #define   HEAD_WRAP_ONE		0x00200000
@@ -1130,9 +1134,50 @@
 #define RCBMINAVG		0x111a0
 #define RCUPEI			0x111b0
 #define RCDNEI			0x111b4
-#define MCHBAR_RENDER_STANDBY		0x111b8
-#define   RCX_SW_EXIT		(1<<23)
-#define   RSX_STATUS_MASK	0x00700000
+#define RSTDBYCTL		0x111b8
+#define   RS1EN			(1<<31)
+#define   RS2EN			(1<<30)
+#define   RS3EN			(1<<29)
+#define   D3RS3EN		(1<<28) /* Display D3 imlies RS3 */
+#define   SWPROMORSX		(1<<27) /* RSx promotion timers ignored */
+#define   RCWAKERW		(1<<26) /* Resetwarn from PCH causes wakeup */
+#define   DPRSLPVREN		(1<<25) /* Fast voltage ramp enable */
+#define   GFXTGHYST		(1<<24) /* Hysteresis to allow trunk gating */
+#define   RCX_SW_EXIT		(1<<23) /* Leave RSx and prevent re-entry */
+#define   RSX_STATUS_MASK	(7<<20)
+#define   RSX_STATUS_ON		(0<<20)
+#define   RSX_STATUS_RC1	(1<<20)
+#define   RSX_STATUS_RC1E	(2<<20)
+#define   RSX_STATUS_RS1	(3<<20)
+#define   RSX_STATUS_RS2	(4<<20) /* aka rc6 */
+#define   RSX_STATUS_RSVD	(5<<20) /* deep rc6 unsupported on ilk */
+#define   RSX_STATUS_RS3	(6<<20) /* rs3 unsupported on ilk */
+#define   RSX_STATUS_RSVD2	(7<<20)
+#define   UWRCRSXE		(1<<19) /* wake counter limit prevents rsx */
+#define   RSCRP			(1<<18) /* rs requests control on rs1/2 reqs */
+#define   JRSC			(1<<17) /* rsx coupled to cpu c-state */
+#define   RS2INC0		(1<<16) /* allow rs2 in cpu c0 */
+#define   RS1CONTSAV_MASK	(3<<14)
+#define   RS1CONTSAV_NO_RS1	(0<<14) /* rs1 doesn't save/restore context */
+#define   RS1CONTSAV_RSVD	(1<<14)
+#define   RS1CONTSAV_SAVE_RS1	(2<<14) /* rs1 saves context */
+#define   RS1CONTSAV_FULL_RS1	(3<<14) /* rs1 saves and restores context */
+#define   NORMSLEXLAT_MASK	(3<<12)
+#define   SLOW_RS123		(0<<12)
+#define   SLOW_RS23		(1<<12)
+#define   SLOW_RS3		(2<<12)
+#define   NORMAL_RS123		(3<<12)
+#define   RCMODE_TIMEOUT	(1<<11) /* 0 is eval interval method */
+#define   IMPROMOEN		(1<<10) /* promo is immediate or delayed until next idle interval (only for timeout method above) */
+#define   RCENTSYNC		(1<<9) /* rs coupled to cpu c-state (3/6/7) */
+#define   STATELOCK		(1<<7) /* locked to rs_cstate if 0 */
+#define   RS_CSTATE_MASK	(3<<4)
+#define   RS_CSTATE_C367_RS1	(0<<4)
+#define   RS_CSTATE_C36_RS1_C7_RS2 (1<<4)
+#define   RS_CSTATE_RSVD	(2<<4)
+#define   RS_CSTATE_C367_RS2	(3<<4)
+#define   REDSAVES		(1<<3) /* no context save if was idle during rs0 */
+#define   REDRESTORES		(1<<2) /* no restore if was idle during rs0 */
 #define VIDCTL			0x111c0
 #define VIDSTS			0x111c8
 #define VIDSTART		0x111cc /* 8 bits */
@@ -2345,8 +2390,13 @@
 
 /* Memory latency timer register */
 #define MLTR_ILK		0x11222
+#define  MLTR_WM1_SHIFT		0
+#define  MLTR_WM2_SHIFT		8
 /* the unit of memory self-refresh latency time is 0.5us */
 #define  ILK_SRLT_MASK		0x3f
+#define ILK_LATENCY(shift)	(I915_READ(MLTR_ILK) >> (shift) & ILK_SRLT_MASK)
+#define ILK_READ_WM1_LATENCY()	ILK_LATENCY(MLTR_WM1_SHIFT)
+#define ILK_READ_WM2_LATENCY()	ILK_LATENCY(MLTR_WM2_SHIFT)
 
 /* define the fifo size on Ironlake */
 #define ILK_DISPLAY_FIFO	128
@@ -2728,12 +2778,41 @@
 /* PCH */
 
 /* south display engine interrupt */
+#define SDE_AUDIO_POWER_D	(1 << 27)
+#define SDE_AUDIO_POWER_C	(1 << 26)
+#define SDE_AUDIO_POWER_B	(1 << 25)
+#define SDE_AUDIO_POWER_SHIFT	(25)
+#define SDE_AUDIO_POWER_MASK	(7 << SDE_AUDIO_POWER_SHIFT)
+#define SDE_GMBUS		(1 << 24)
+#define SDE_AUDIO_HDCP_TRANSB	(1 << 23)
+#define SDE_AUDIO_HDCP_TRANSA	(1 << 22)
+#define SDE_AUDIO_HDCP_MASK	(3 << 22)
+#define SDE_AUDIO_TRANSB	(1 << 21)
+#define SDE_AUDIO_TRANSA	(1 << 20)
+#define SDE_AUDIO_TRANS_MASK	(3 << 20)
+#define SDE_POISON		(1 << 19)
+/* 18 reserved */
+#define SDE_FDI_RXB		(1 << 17)
+#define SDE_FDI_RXA		(1 << 16)
+#define SDE_FDI_MASK		(3 << 16)
+#define SDE_AUXD		(1 << 15)
+#define SDE_AUXC		(1 << 14)
+#define SDE_AUXB		(1 << 13)
+#define SDE_AUX_MASK		(7 << 13)
+/* 12 reserved */
 #define SDE_CRT_HOTPLUG         (1 << 11)
 #define SDE_PORTD_HOTPLUG       (1 << 10)
 #define SDE_PORTC_HOTPLUG       (1 << 9)
 #define SDE_PORTB_HOTPLUG       (1 << 8)
 #define SDE_SDVOB_HOTPLUG       (1 << 6)
 #define SDE_HOTPLUG_MASK	(0xf << 8)
+#define SDE_TRANSB_CRC_DONE	(1 << 5)
+#define SDE_TRANSB_CRC_ERR	(1 << 4)
+#define SDE_TRANSB_FIFO_UNDER	(1 << 3)
+#define SDE_TRANSA_CRC_DONE	(1 << 2)
+#define SDE_TRANSA_CRC_ERR	(1 << 1)
+#define SDE_TRANSA_FIFO_UNDER	(1 << 0)
+#define SDE_TRANS_MASK		(0x3f)
 /* CPT */
 #define SDE_CRT_HOTPLUG_CPT	(1 << 19)
 #define SDE_PORTD_HOTPLUG_CPT	(1 << 23)
@@ -3174,10 +3253,11 @@
 #define  EDP_LINK_TRAIN_600MV_3_5DB_SNB_A	(0x01<<22)
 #define  EDP_LINK_TRAIN_800MV_0DB_SNB_A		(0x0<<22)
 /* SNB B-stepping */
-#define  EDP_LINK_TRAIN_400MV_0DB_SNB_B		(0x0<<22)
-#define  EDP_LINK_TRAIN_400MV_6DB_SNB_B		(0x3a<<22)
-#define  EDP_LINK_TRAIN_600MV_3_5DB_SNB_B	(0x39<<22)
-#define  EDP_LINK_TRAIN_800MV_0DB_SNB_B		(0x38<<22)
+#define  EDP_LINK_TRAIN_400_600MV_0DB_SNB_B	(0x0<<22)
+#define  EDP_LINK_TRAIN_400MV_3_5DB_SNB_B	(0x1<<22)
+#define  EDP_LINK_TRAIN_400_600MV_6DB_SNB_B	(0x3a<<22)
+#define  EDP_LINK_TRAIN_600_800MV_3_5DB_SNB_B	(0x39<<22)
+#define  EDP_LINK_TRAIN_800_1200MV_0DB_SNB_B	(0x38<<22)
 #define  EDP_LINK_TRAIN_VOL_EMP_MASK_SNB	(0x3f<<22)
 
 #define  FORCEWAKE				0xA18C
@@ -3239,6 +3319,7 @@
 
 #define GEN6_PCODE_MAILBOX			0x138124
 #define   GEN6_PCODE_READY			(1<<31)
+#define   GEN6_READ_OC_PARAMS			0xc
 #define   GEN6_PCODE_WRITE_MIN_FREQ_TABLE	0x9
 #define GEN6_PCODE_DATA				0x138128
 
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 410772466fa7..0521ecf26017 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -740,7 +740,7 @@ void i915_restore_display(struct drm_device *dev)
 		I915_WRITE(PCH_PP_OFF_DELAYS, dev_priv->savePP_OFF_DELAYS);
 		I915_WRITE(PCH_PP_DIVISOR, dev_priv->savePP_DIVISOR);
 		I915_WRITE(PCH_PP_CONTROL, dev_priv->savePP_CONTROL);
-		I915_WRITE(MCHBAR_RENDER_STANDBY,
+		I915_WRITE(RSTDBYCTL,
 			   dev_priv->saveMCHBAR_RENDER_STANDBY);
 	} else {
 		I915_WRITE(PFIT_PGM_RATIOS, dev_priv->savePFIT_PGM_RATIOS);
@@ -811,7 +811,7 @@ int i915_save_state(struct drm_device *dev)
 		dev_priv->saveFDI_RXA_IMR = I915_READ(FDI_RXA_IMR);
 		dev_priv->saveFDI_RXB_IMR = I915_READ(FDI_RXB_IMR);
 		dev_priv->saveMCHBAR_RENDER_STANDBY =
-			I915_READ(MCHBAR_RENDER_STANDBY);
+			I915_READ(RSTDBYCTL);
 	} else {
 		dev_priv->saveIER = I915_READ(IER);
 		dev_priv->saveIMR = I915_READ(IMR);
@@ -822,10 +822,6 @@ int i915_save_state(struct drm_device *dev)
 	if (IS_GEN6(dev))
 		gen6_disable_rps(dev);
 
-	/* XXX disabling the clock gating breaks suspend on gm45
-	intel_disable_clock_gating(dev);
-	 */
-
 	/* Cache mode state */
 	dev_priv->saveCACHE_MODE_0 = I915_READ(CACHE_MODE_0);
 
diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index 8df574316063..17035b87ee46 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -30,6 +30,7 @@
 #include "drm.h"
 #include "drm_crtc.h"
 #include "drm_crtc_helper.h"
+#include "drm_edid.h"
 #include "intel_drv.h"
 #include "i915_drm.h"
 #include "i915_drv.h"
@@ -287,8 +288,9 @@ static bool intel_crt_ddc_probe(struct drm_i915_private *dev_priv, int ddc_bus)
 	return i2c_transfer(&dev_priv->gmbus[ddc_bus].adapter, msgs, 1) == 1;
 }
 
-static bool intel_crt_detect_ddc(struct intel_crt *crt)
+static bool intel_crt_detect_ddc(struct drm_connector *connector)
 {
+	struct intel_crt *crt = intel_attached_crt(connector);
 	struct drm_i915_private *dev_priv = crt->base.base.dev->dev_private;
 
 	/* CRT should always be at 0, but check anyway */
@@ -301,8 +303,26 @@ static bool intel_crt_detect_ddc(struct intel_crt *crt)
 	}
 
 	if (intel_ddc_probe(&crt->base, dev_priv->crt_ddc_pin)) {
-		DRM_DEBUG_KMS("CRT detected via DDC:0x50 [EDID]\n");
-		return true;
+		struct edid *edid;
+		bool is_digital = false;
+
+		edid = drm_get_edid(connector,
+			&dev_priv->gmbus[dev_priv->crt_ddc_pin].adapter);
+		/*
+		 * This may be a DVI-I connector with a shared DDC
+		 * link between analog and digital outputs, so we
+		 * have to check the EDID input spec of the attached device.
+		 */
+		if (edid != NULL) {
+			is_digital = edid->input & DRM_EDID_INPUT_DIGITAL;
+			connector->display_info.raw_edid = NULL;
+			kfree(edid);
+		}
+
+		if (!is_digital) {
+			DRM_DEBUG_KMS("CRT detected via DDC:0x50 [EDID]\n");
+			return true;
+		}
 	}
 
 	return false;
@@ -458,7 +478,7 @@ intel_crt_detect(struct drm_connector *connector, bool force)
 		}
 	}
 
-	if (intel_crt_detect_ddc(crt))
+	if (intel_crt_detect_ddc(connector))
 		return connector_status_connected;
 
 	if (!force)
@@ -472,7 +492,7 @@ intel_crt_detect(struct drm_connector *connector, bool force)
 		crtc = intel_get_load_detect_pipe(&crt->base, connector,
 						  NULL, &dpms_mode);
 		if (crtc) {
-			if (intel_crt_detect_ddc(crt))
+			if (intel_crt_detect_ddc(connector))
 				status = connector_status_connected;
 			else
 				status = intel_crt_load_detect(crtc, crt);
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 0abe79fb6385..25d96889d7d2 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -3418,15 +3418,16 @@ static void i830_update_wm(struct drm_device *dev, int planea_clock, int unused,
 static bool ironlake_compute_wm0(struct drm_device *dev,
 				 int pipe,
 				 const struct intel_watermark_params *display,
-				 int display_latency,
+				 int display_latency_ns,
 				 const struct intel_watermark_params *cursor,
-				 int cursor_latency,
+				 int cursor_latency_ns,
 				 int *plane_wm,
 				 int *cursor_wm)
 {
 	struct drm_crtc *crtc;
-	int htotal, hdisplay, clock, pixel_size = 0;
-	int line_time_us, line_count, entries;
+	int htotal, hdisplay, clock, pixel_size;
+	int line_time_us, line_count;
+	int entries, tlb_miss;
 
 	crtc = intel_get_crtc_for_pipe(dev, pipe);
 	if (crtc->fb == NULL || !crtc->enabled)
@@ -3438,7 +3439,10 @@ static bool ironlake_compute_wm0(struct drm_device *dev,
 	pixel_size = crtc->fb->bits_per_pixel / 8;
 
 	/* Use the small buffer method to calculate plane watermark */
-	entries = ((clock * pixel_size / 1000) * display_latency * 100) / 1000;
+	entries = ((clock * pixel_size / 1000) * display_latency_ns) / 1000;
+	tlb_miss = display->fifo_size*display->cacheline_size - hdisplay * 8;
+	if (tlb_miss > 0)
+		entries += tlb_miss;
 	entries = DIV_ROUND_UP(entries, display->cacheline_size);
 	*plane_wm = entries + display->guard_size;
 	if (*plane_wm > (int)display->max_wm)
@@ -3446,8 +3450,11 @@ static bool ironlake_compute_wm0(struct drm_device *dev,
 
 	/* Use the large buffer method to calculate cursor watermark */
 	line_time_us = ((htotal * 1000) / clock);
-	line_count = (cursor_latency * 100 / line_time_us + 1000) / 1000;
+	line_count = (cursor_latency_ns / line_time_us + 1000) / 1000;
 	entries = line_count * 64 * pixel_size;
+	tlb_miss = cursor->fifo_size*cursor->cacheline_size - hdisplay * 8;
+	if (tlb_miss > 0)
+		entries += tlb_miss;
 	entries = DIV_ROUND_UP(entries, cursor->cacheline_size);
 	*cursor_wm = entries + cursor->guard_size;
 	if (*cursor_wm > (int)cursor->max_wm)
@@ -3456,113 +3463,17 @@ static bool ironlake_compute_wm0(struct drm_device *dev,
 	return true;
 }
 
-static void ironlake_update_wm(struct drm_device *dev,
-			       int planea_clock, int planeb_clock,
-			       int sr_hdisplay, int sr_htotal,
-			       int pixel_size)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int plane_wm, cursor_wm, enabled;
-	int tmp;
-
-	enabled = 0;
-	if (ironlake_compute_wm0(dev, 0,
-				 &ironlake_display_wm_info,
-				 ILK_LP0_PLANE_LATENCY,
-				 &ironlake_cursor_wm_info,
-				 ILK_LP0_CURSOR_LATENCY,
-				 &plane_wm, &cursor_wm)) {
-		I915_WRITE(WM0_PIPEA_ILK,
-			   (plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm);
-		DRM_DEBUG_KMS("FIFO watermarks For pipe A -"
-			      " plane %d, " "cursor: %d\n",
-			      plane_wm, cursor_wm);
-		enabled++;
-	}
-
-	if (ironlake_compute_wm0(dev, 1,
-				 &ironlake_display_wm_info,
-				 ILK_LP0_PLANE_LATENCY,
-				 &ironlake_cursor_wm_info,
-				 ILK_LP0_CURSOR_LATENCY,
-				 &plane_wm, &cursor_wm)) {
-		I915_WRITE(WM0_PIPEB_ILK,
-			   (plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm);
-		DRM_DEBUG_KMS("FIFO watermarks For pipe B -"
-			      " plane %d, cursor: %d\n",
-			      plane_wm, cursor_wm);
-		enabled++;
-	}
-
-	/*
-	 * Calculate and update the self-refresh watermark only when one
-	 * display plane is used.
-	 */
-	tmp = 0;
-	if (enabled == 1) {
-		unsigned long line_time_us;
-		int small, large, plane_fbc;
-		int sr_clock, entries;
-		int line_count, line_size;
-		/* Read the self-refresh latency. The unit is 0.5us */
-		int ilk_sr_latency = I915_READ(MLTR_ILK) & ILK_SRLT_MASK;
-
-		sr_clock = planea_clock ? planea_clock : planeb_clock;
-		line_time_us = (sr_htotal * 1000) / sr_clock;
-
-		/* Use ns/us then divide to preserve precision */
-		line_count = ((ilk_sr_latency * 500) / line_time_us + 1000)
-			/ 1000;
-		line_size = sr_hdisplay * pixel_size;
-
-		/* Use the minimum of the small and large buffer method for primary */
-		small = ((sr_clock * pixel_size / 1000) * (ilk_sr_latency * 500)) / 1000;
-		large = line_count * line_size;
-
-		entries = DIV_ROUND_UP(min(small, large),
-				       ironlake_display_srwm_info.cacheline_size);
-
-		plane_fbc = entries * 64;
-		plane_fbc = DIV_ROUND_UP(plane_fbc, line_size);
-
-		plane_wm = entries + ironlake_display_srwm_info.guard_size;
-		if (plane_wm > (int)ironlake_display_srwm_info.max_wm)
-			plane_wm = ironlake_display_srwm_info.max_wm;
-
-		/* calculate the self-refresh watermark for display cursor */
-		entries = line_count * pixel_size * 64;
-		entries = DIV_ROUND_UP(entries,
-				       ironlake_cursor_srwm_info.cacheline_size);
-
-		cursor_wm = entries + ironlake_cursor_srwm_info.guard_size;
-		if (cursor_wm > (int)ironlake_cursor_srwm_info.max_wm)
-			cursor_wm = ironlake_cursor_srwm_info.max_wm;
-
-		/* configure watermark and enable self-refresh */
-		tmp = (WM1_LP_SR_EN |
-		       (ilk_sr_latency << WM1_LP_LATENCY_SHIFT) |
-		       (plane_fbc << WM1_LP_FBC_SHIFT) |
-		       (plane_wm << WM1_LP_SR_SHIFT) |
-		       cursor_wm);
-		DRM_DEBUG_KMS("self-refresh watermark: display plane %d, fbc lines %d,"
-			      " cursor %d\n", plane_wm, plane_fbc, cursor_wm);
-	}
-	I915_WRITE(WM1_LP_ILK, tmp);
-	/* XXX setup WM2 and WM3 */
-}
-
 /*
  * Check the wm result.
  *
  * If any calculated watermark values is larger than the maximum value that
  * can be programmed into the associated watermark register, that watermark
  * must be disabled.
- *
- * Also return true if all of those watermark values is 0, which is set by
- * sandybridge_compute_srwm, to indicate the latency is ZERO.
  */
-static bool sandybridge_check_srwm(struct drm_device *dev, int level,
-				   int fbc_wm, int display_wm, int cursor_wm)
+static bool ironlake_check_srwm(struct drm_device *dev, int level,
+				int fbc_wm, int display_wm, int cursor_wm,
+				const struct intel_watermark_params *display,
+				const struct intel_watermark_params *cursor)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
@@ -3571,7 +3482,7 @@ static bool sandybridge_check_srwm(struct drm_device *dev, int level,
 
 	if (fbc_wm > SNB_FBC_MAX_SRWM) {
 		DRM_DEBUG_KMS("fbc watermark(%d) is too large(%d), disabling wm%d+\n",
-				fbc_wm, SNB_FBC_MAX_SRWM, level);
+			      fbc_wm, SNB_FBC_MAX_SRWM, level);
 
 		/* fbc has it's own way to disable FBC WM */
 		I915_WRITE(DISP_ARB_CTL,
@@ -3579,15 +3490,15 @@ static bool sandybridge_check_srwm(struct drm_device *dev, int level,
 		return false;
 	}
 
-	if (display_wm > SNB_DISPLAY_MAX_SRWM) {
+	if (display_wm > display->max_wm) {
 		DRM_DEBUG_KMS("display watermark(%d) is too large(%d), disabling wm%d+\n",
-				display_wm, SNB_DISPLAY_MAX_SRWM, level);
+			      display_wm, SNB_DISPLAY_MAX_SRWM, level);
 		return false;
 	}
 
-	if (cursor_wm > SNB_CURSOR_MAX_SRWM) {
+	if (cursor_wm > cursor->max_wm) {
 		DRM_DEBUG_KMS("cursor watermark(%d) is too large(%d), disabling wm%d+\n",
-				cursor_wm, SNB_CURSOR_MAX_SRWM, level);
+			      cursor_wm, SNB_CURSOR_MAX_SRWM, level);
 		return false;
 	}
 
@@ -3602,16 +3513,18 @@ static bool sandybridge_check_srwm(struct drm_device *dev, int level,
 /*
  * Compute watermark values of WM[1-3],
  */
-static bool sandybridge_compute_srwm(struct drm_device *dev, int level,
-				     int hdisplay, int htotal, int pixel_size,
-				     int clock, int latency_ns, int *fbc_wm,
-				     int *display_wm, int *cursor_wm)
+static bool ironlake_compute_srwm(struct drm_device *dev, int level,
+				  int hdisplay, int htotal,
+				  int pixel_size, int clock, int latency_ns,
+				  const struct intel_watermark_params *display,
+				  const struct intel_watermark_params *cursor,
+				  int *fbc_wm, int *display_wm, int *cursor_wm)
 {
 
 	unsigned long line_time_us;
+	int line_count, line_size;
 	int small, large;
 	int entries;
-	int line_count, line_size;
 
 	if (!latency_ns) {
 		*fbc_wm = *display_wm = *cursor_wm = 0;
@@ -3626,24 +3539,110 @@ static bool sandybridge_compute_srwm(struct drm_device *dev, int level,
 	small = ((clock * pixel_size / 1000) * latency_ns) / 1000;
 	large = line_count * line_size;
 
-	entries = DIV_ROUND_UP(min(small, large),
-				sandybridge_display_srwm_info.cacheline_size);
-	*display_wm = entries + sandybridge_display_srwm_info.guard_size;
+	entries = DIV_ROUND_UP(min(small, large), display->cacheline_size);
+	*display_wm = entries + display->guard_size;
 
 	/*
-	 * Spec said:
+	 * Spec says:
 	 * FBC WM = ((Final Primary WM * 64) / number of bytes per line) + 2
 	 */
 	*fbc_wm = DIV_ROUND_UP(*display_wm * 64, line_size) + 2;
 
 	/* calculate the self-refresh watermark for display cursor */
 	entries = line_count * pixel_size * 64;
-	entries = DIV_ROUND_UP(entries,
-			       sandybridge_cursor_srwm_info.cacheline_size);
-	*cursor_wm = entries + sandybridge_cursor_srwm_info.guard_size;
+	entries = DIV_ROUND_UP(entries, cursor->cacheline_size);
+	*cursor_wm = entries + cursor->guard_size;
 
-	return sandybridge_check_srwm(dev, level,
-				      *fbc_wm, *display_wm, *cursor_wm);
+	return ironlake_check_srwm(dev, level,
+				   *fbc_wm, *display_wm, *cursor_wm,
+				   display, cursor);
+}
+
+static void ironlake_update_wm(struct drm_device *dev,
+			       int planea_clock, int planeb_clock,
+			       int hdisplay, int htotal,
+			       int pixel_size)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int fbc_wm, plane_wm, cursor_wm, enabled;
+	int clock;
+
+	enabled = 0;
+	if (ironlake_compute_wm0(dev, 0,
+				 &ironlake_display_wm_info,
+				 ILK_LP0_PLANE_LATENCY,
+				 &ironlake_cursor_wm_info,
+				 ILK_LP0_CURSOR_LATENCY,
+				 &plane_wm, &cursor_wm)) {
+		I915_WRITE(WM0_PIPEA_ILK,
+			   (plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm);
+		DRM_DEBUG_KMS("FIFO watermarks For pipe A -"
+			      " plane %d, " "cursor: %d\n",
+			      plane_wm, cursor_wm);
+		enabled++;
+	}
+
+	if (ironlake_compute_wm0(dev, 1,
+				 &ironlake_display_wm_info,
+				 ILK_LP0_PLANE_LATENCY,
+				 &ironlake_cursor_wm_info,
+				 ILK_LP0_CURSOR_LATENCY,
+				 &plane_wm, &cursor_wm)) {
+		I915_WRITE(WM0_PIPEB_ILK,
+			   (plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm);
+		DRM_DEBUG_KMS("FIFO watermarks For pipe B -"
+			      " plane %d, cursor: %d\n",
+			      plane_wm, cursor_wm);
+		enabled++;
+	}
+
+	/*
+	 * Calculate and update the self-refresh watermark only when one
+	 * display plane is used.
+	 */
+	I915_WRITE(WM3_LP_ILK, 0);
+	I915_WRITE(WM2_LP_ILK, 0);
+	I915_WRITE(WM1_LP_ILK, 0);
+
+	if (enabled != 1)
+		return;
+
+	clock = planea_clock ? planea_clock : planeb_clock;
+
+	/* WM1 */
+	if (!ironlake_compute_srwm(dev, 1, hdisplay, htotal, pixel_size,
+				   clock, ILK_READ_WM1_LATENCY() * 500,
+				   &ironlake_display_srwm_info,
+				   &ironlake_cursor_srwm_info,
+				   &fbc_wm, &plane_wm, &cursor_wm))
+		return;
+
+	I915_WRITE(WM1_LP_ILK,
+		   WM1_LP_SR_EN |
+		   (ILK_READ_WM1_LATENCY() << WM1_LP_LATENCY_SHIFT) |
+		   (fbc_wm << WM1_LP_FBC_SHIFT) |
+		   (plane_wm << WM1_LP_SR_SHIFT) |
+		   cursor_wm);
+
+	/* WM2 */
+	if (!ironlake_compute_srwm(dev, 2, hdisplay, htotal, pixel_size,
+				   clock, ILK_READ_WM2_LATENCY() * 500,
+				   &ironlake_display_srwm_info,
+				   &ironlake_cursor_srwm_info,
+				   &fbc_wm, &plane_wm, &cursor_wm))
+		return;
+
+	I915_WRITE(WM2_LP_ILK,
+		   WM2_LP_EN |
+		   (ILK_READ_WM2_LATENCY() << WM1_LP_LATENCY_SHIFT) |
+		   (fbc_wm << WM1_LP_FBC_SHIFT) |
+		   (plane_wm << WM1_LP_SR_SHIFT) |
+		   cursor_wm);
+
+	/*
+	 * WM3 is unsupported on ILK, probably because we don't have latency
+	 * data for that power state
+	 */
 }
 
 static void sandybridge_update_wm(struct drm_device *dev,
@@ -3652,7 +3651,7 @@ static void sandybridge_update_wm(struct drm_device *dev,
 			       int pixel_size)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int latency = SNB_READ_WM0_LATENCY();
+	int latency = SNB_READ_WM0_LATENCY() * 100;	/* In unit 0.1us */
 	int fbc_wm, plane_wm, cursor_wm, enabled;
 	int clock;
 
@@ -3701,9 +3700,11 @@ static void sandybridge_update_wm(struct drm_device *dev,
 	clock = planea_clock ? planea_clock : planeb_clock;
 
 	/* WM1 */
-	if (!sandybridge_compute_srwm(dev, 1, hdisplay, htotal, pixel_size,
-				      clock, SNB_READ_WM1_LATENCY() * 500,
-				      &fbc_wm, &plane_wm, &cursor_wm))
+	if (!ironlake_compute_srwm(dev, 1, hdisplay, htotal, pixel_size,
+				   clock, SNB_READ_WM1_LATENCY() * 500,
+				   &sandybridge_display_srwm_info,
+				   &sandybridge_cursor_srwm_info,
+				   &fbc_wm, &plane_wm, &cursor_wm))
 		return;
 
 	I915_WRITE(WM1_LP_ILK,
@@ -3714,10 +3715,12 @@ static void sandybridge_update_wm(struct drm_device *dev,
 		   cursor_wm);
 
 	/* WM2 */
-	if (!sandybridge_compute_srwm(dev, 2,
-				      hdisplay, htotal, pixel_size,
-				      clock, SNB_READ_WM2_LATENCY() * 500,
-				      &fbc_wm, &plane_wm, &cursor_wm))
+	if (!ironlake_compute_srwm(dev, 2,
+				   hdisplay, htotal, pixel_size,
+				   clock, SNB_READ_WM2_LATENCY() * 500,
+				   &sandybridge_display_srwm_info,
+				   &sandybridge_cursor_srwm_info,
+				   &fbc_wm, &plane_wm, &cursor_wm))
 		return;
 
 	I915_WRITE(WM2_LP_ILK,
@@ -3728,10 +3731,12 @@ static void sandybridge_update_wm(struct drm_device *dev,
 		   cursor_wm);
 
 	/* WM3 */
-	if (!sandybridge_compute_srwm(dev, 3,
-				      hdisplay, htotal, pixel_size,
-				      clock, SNB_READ_WM3_LATENCY() * 500,
-				      &fbc_wm, &plane_wm, &cursor_wm))
+	if (!ironlake_compute_srwm(dev, 3,
+				   hdisplay, htotal, pixel_size,
+				   clock, SNB_READ_WM3_LATENCY() * 500,
+				   &sandybridge_display_srwm_info,
+				   &sandybridge_cursor_srwm_info,
+				   &fbc_wm, &plane_wm, &cursor_wm))
 		return;
 
 	I915_WRITE(WM3_LP_ILK,
@@ -3951,7 +3956,7 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc,
 		int lane = 0, link_bw, bpp;
 		/* CPU eDP doesn't require FDI link, so just set DP M/N
 		   according to current link config */
-		if (has_edp_encoder && !intel_encoder_is_pch_edp(&encoder->base)) {
+		if (has_edp_encoder && !intel_encoder_is_pch_edp(&has_edp_encoder->base)) {
 			target_clock = mode->clock;
 			intel_edp_link_config(has_edp_encoder,
 					      &lane, &link_bw);
@@ -5038,8 +5043,8 @@ static void intel_increase_pllclock(struct drm_crtc *crtc)
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
-	int dpll_reg = (pipe == 0) ? DPLL_A : DPLL_B;
-	int dpll = I915_READ(dpll_reg);
+	int dpll_reg = DPLL(pipe);
+	int dpll;
 
 	if (HAS_PCH_SPLIT(dev))
 		return;
@@ -5047,17 +5052,19 @@ static void intel_increase_pllclock(struct drm_crtc *crtc)
 	if (!dev_priv->lvds_downclock_avail)
 		return;
 
+	dpll = I915_READ(dpll_reg);
 	if (!HAS_PIPE_CXSR(dev) && (dpll & DISPLAY_RATE_SELECT_FPA1)) {
 		DRM_DEBUG_DRIVER("upclocking LVDS\n");
 
 		/* Unlock panel regs */
-		I915_WRITE(PP_CONTROL, I915_READ(PP_CONTROL) |
-			   PANEL_UNLOCK_REGS);
+		I915_WRITE(PP_CONTROL,
+			   I915_READ(PP_CONTROL) | PANEL_UNLOCK_REGS);
 
 		dpll &= ~DISPLAY_RATE_SELECT_FPA1;
 		I915_WRITE(dpll_reg, dpll);
-		dpll = I915_READ(dpll_reg);
+		POSTING_READ(dpll_reg);
 		intel_wait_for_vblank(dev, pipe);
+
 		dpll = I915_READ(dpll_reg);
 		if (dpll & DISPLAY_RATE_SELECT_FPA1)
 			DRM_DEBUG_DRIVER("failed to upclock LVDS!\n");
@@ -5802,6 +5809,8 @@ static void intel_setup_outputs(struct drm_device *dev)
 		encoder->base.possible_clones =
 			intel_encoder_clones(dev, encoder->clone_mask);
 	}
+
+	intel_panel_setup_backlight(dev);
 }
 
 static void intel_user_framebuffer_destroy(struct drm_framebuffer *fb)
@@ -6145,6 +6154,10 @@ void intel_init_emon(struct drm_device *dev)
 
 void gen6_enable_rps(struct drm_i915_private *dev_priv)
 {
+	u32 rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
+	u32 gt_perf_status = I915_READ(GEN6_GT_PERF_STATUS);
+	u32 pcu_mbox;
+	int cur_freq, min_freq, max_freq;
 	int i;
 
 	/* Here begins a magic sequence of register writes to enable
@@ -6216,6 +6229,29 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv)
 		     500))
 		DRM_ERROR("timeout waiting for pcode mailbox to finish\n");
 
+	min_freq = (rp_state_cap & 0xff0000) >> 16;
+	max_freq = rp_state_cap & 0xff;
+	cur_freq = (gt_perf_status & 0xff00) >> 8;
+
+	/* Check for overclock support */
+	if (wait_for((I915_READ(GEN6_PCODE_MAILBOX) & GEN6_PCODE_READY) == 0,
+		     500))
+		DRM_ERROR("timeout waiting for pcode mailbox to become idle\n");
+	I915_WRITE(GEN6_PCODE_MAILBOX, GEN6_READ_OC_PARAMS);
+	pcu_mbox = I915_READ(GEN6_PCODE_DATA);
+	if (wait_for((I915_READ(GEN6_PCODE_MAILBOX) & GEN6_PCODE_READY) == 0,
+		     500))
+		DRM_ERROR("timeout waiting for pcode mailbox to finish\n");
+	if (pcu_mbox & (1<<31)) { /* OC supported */
+		max_freq = pcu_mbox & 0xff;
+		DRM_DEBUG_DRIVER("overclocking supported, adjusting frequency max to %dMHz\n", pcu_mbox * 100);
+	}
+
+	/* In units of 100MHz */
+	dev_priv->max_delay = max_freq;
+	dev_priv->min_delay = min_freq;
+	dev_priv->cur_delay = cur_freq;
+
 	/* requires MSI enabled */
 	I915_WRITE(GEN6_PMIER,
 		   GEN6_PM_MBOX_EVENT |
@@ -6386,42 +6422,6 @@ void intel_enable_clock_gating(struct drm_device *dev)
 	} else if (IS_I830(dev)) {
 		I915_WRITE(DSPCLK_GATE_D, OVRUNIT_CLOCK_GATE_DISABLE);
 	}
-
-	/*
-	 * GPU can automatically power down the render unit if given a page
-	 * to save state.
-	 */
-	if (IS_IRONLAKE_M(dev) && 0) { /* XXX causes a failure during suspend */
-		if (dev_priv->renderctx == NULL)
-			dev_priv->renderctx = intel_alloc_context_page(dev);
-		if (dev_priv->renderctx) {
-			struct drm_i915_gem_object *obj = dev_priv->renderctx;
-			if (BEGIN_LP_RING(4) == 0) {
-				OUT_RING(MI_SET_CONTEXT);
-				OUT_RING(obj->gtt_offset |
-					 MI_MM_SPACE_GTT |
-					 MI_SAVE_EXT_STATE_EN |
-					 MI_RESTORE_EXT_STATE_EN |
-					 MI_RESTORE_INHIBIT);
-				OUT_RING(MI_NOOP);
-				OUT_RING(MI_FLUSH);
-				ADVANCE_LP_RING();
-			}
-		} else
-			DRM_DEBUG_KMS("Failed to allocate render context."
-				       "Disable RC6\n");
-	}
-
-	if (IS_GEN4(dev) && IS_MOBILE(dev)) {
-		if (dev_priv->pwrctx == NULL)
-			dev_priv->pwrctx = intel_alloc_context_page(dev);
-		if (dev_priv->pwrctx) {
-			struct drm_i915_gem_object *obj = dev_priv->pwrctx;
-			I915_WRITE(PWRCTXA, obj->gtt_offset | PWRCTX_EN);
-			I915_WRITE(MCHBAR_RENDER_STANDBY,
-				   I915_READ(MCHBAR_RENDER_STANDBY) & ~RCX_SW_EXIT);
-		}
-	}
 }
 
 void intel_disable_clock_gating(struct drm_device *dev)
@@ -6451,6 +6451,57 @@ void intel_disable_clock_gating(struct drm_device *dev)
 	}
 }
 
+static void ironlake_disable_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;
+}
+
+void ironlake_enable_rc6(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int ret;
+
+	/*
+	 * 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);
+		return;
+	}
+	OUT_RING(MI_SUSPEND_FLUSH | MI_SUSPEND_FLUSH_EN);
+	OUT_RING(MI_SET_CONTEXT);
+	OUT_RING(dev_priv->renderctx->gtt_offset |
+		 MI_MM_SPACE_GTT |
+		 MI_SAVE_EXT_STATE_EN |
+		 MI_RESTORE_EXT_STATE_EN |
+		 MI_RESTORE_INHIBIT);
+	OUT_RING(MI_SUSPEND_FLUSH);
+	OUT_RING(MI_NOOP);
+	OUT_RING(MI_FLUSH);
+	ADVANCE_LP_RING();
+
+	I915_WRITE(PWRCTXA, dev_priv->pwrctx->gtt_offset | PWRCTX_EN);
+	I915_WRITE(RSTDBYCTL, I915_READ(RSTDBYCTL) & ~RCX_SW_EXIT);
+}
+
 /* Set up chip specific display functions */
 static void intel_init_display(struct drm_device *dev)
 {
@@ -6665,12 +6716,7 @@ void intel_modeset_init(struct drm_device *dev)
 		dev->mode_config.max_width = 8192;
 		dev->mode_config.max_height = 8192;
 	}
-
-	/* set memory base */
-	if (IS_GEN2(dev))
-		dev->mode_config.fb_base = pci_resource_start(dev->pdev, 0);
-	else
-		dev->mode_config.fb_base = pci_resource_start(dev->pdev, 2);
+	dev->mode_config.fb_base = dev->agp->base;
 
 	if (IS_MOBILE(dev) || !IS_GEN2(dev))
 		dev_priv->num_pipe = 2;
@@ -6698,6 +6744,21 @@ 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;
+		}
+		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);
@@ -6734,7 +6795,8 @@ void intel_modeset_cleanup(struct drm_device *dev)
 	if (IS_GEN6(dev))
 		gen6_disable_rps(dev);
 
-	intel_disable_clock_gating(dev);
+	if (IS_IRONLAKE_M(dev))
+		ironlake_disable_rc6(dev);
 
 	mutex_unlock(&dev->struct_mutex);
 
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 1dc60408d5b8..1f4242b682c8 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -1153,18 +1153,27 @@ intel_dp_signal_levels(uint8_t train_set, int lane_count)
 static uint32_t
 intel_gen6_edp_signal_levels(uint8_t train_set)
 {
-	switch (train_set & (DP_TRAIN_VOLTAGE_SWING_MASK|DP_TRAIN_PRE_EMPHASIS_MASK)) {
+	int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
+					 DP_TRAIN_PRE_EMPHASIS_MASK);
+	switch (signal_levels) {
 	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_0:
-		return EDP_LINK_TRAIN_400MV_0DB_SNB_B;
+	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_0:
+		return EDP_LINK_TRAIN_400_600MV_0DB_SNB_B;
+	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_3_5:
+		return EDP_LINK_TRAIN_400MV_3_5DB_SNB_B;
 	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_6:
-		return EDP_LINK_TRAIN_400MV_6DB_SNB_B;
+	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_6:
+		return EDP_LINK_TRAIN_400_600MV_6DB_SNB_B;
 	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return EDP_LINK_TRAIN_600MV_3_5DB_SNB_B;
+	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_3_5:
+		return EDP_LINK_TRAIN_600_800MV_3_5DB_SNB_B;
 	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_0:
-		return EDP_LINK_TRAIN_800MV_0DB_SNB_B;
+	case DP_TRAIN_VOLTAGE_SWING_1200 | DP_TRAIN_PRE_EMPHASIS_0:
+		return EDP_LINK_TRAIN_800_1200MV_0DB_SNB_B;
 	default:
-		DRM_DEBUG_KMS("Unsupported voltage swing/pre-emphasis level\n");
-		return EDP_LINK_TRAIN_400MV_0DB_SNB_B;
+		DRM_DEBUG_KMS("Unsupported voltage swing/pre-emphasis level:"
+			      "0x%x\n", signal_levels);
+		return EDP_LINK_TRAIN_400_600MV_0DB_SNB_B;
 	}
 }
 
@@ -1334,17 +1343,24 @@ intel_dp_complete_link_train(struct intel_dp *intel_dp)
 	struct drm_device *dev = intel_dp->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	bool channel_eq = false;
-	int tries;
+	int tries, cr_tries;
 	u32 reg;
 	uint32_t DP = intel_dp->DP;
 
 	/* channel equalization */
 	tries = 0;
+	cr_tries = 0;
 	channel_eq = false;
 	for (;;) {
 		/* Use intel_dp->train_set[0] to set the voltage and pre emphasis values */
 		uint32_t    signal_levels;
 
+		if (cr_tries > 5) {
+			DRM_ERROR("failed to train DP, aborting\n");
+			intel_dp_link_down(intel_dp);
+			break;
+		}
+
 		if (IS_GEN6(dev) && is_edp(intel_dp)) {
 			signal_levels = intel_gen6_edp_signal_levels(intel_dp->train_set[0]);
 			DP = (DP & ~EDP_LINK_TRAIN_VOL_EMP_MASK_SNB) | signal_levels;
@@ -1367,14 +1383,26 @@ intel_dp_complete_link_train(struct intel_dp *intel_dp)
 		if (!intel_dp_get_link_status(intel_dp))
 			break;
 
+		/* Make sure clock is still ok */
+		if (!intel_clock_recovery_ok(intel_dp->link_status, intel_dp->lane_count)) {
+			intel_dp_start_link_train(intel_dp);
+			cr_tries++;
+			continue;
+		}
+
 		if (intel_channel_eq_ok(intel_dp)) {
 			channel_eq = true;
 			break;
 		}
 
-		/* Try 5 times */
-		if (tries > 5)
-			break;
+		/* Try 5 times, then try clock recovery if that fails */
+		if (tries > 5) {
+			intel_dp_link_down(intel_dp);
+			intel_dp_start_link_train(intel_dp);
+			tries = 0;
+			cr_tries++;
+			continue;
+		}
 
 		/* Compute new intel_dp->train_set as requested by target */
 		intel_get_adjust_train(intel_dp);
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index d782ad9fd6db..74db2557d644 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -257,6 +257,9 @@ extern void intel_pch_panel_fitting(struct drm_device *dev,
 extern u32 intel_panel_get_max_backlight(struct drm_device *dev);
 extern u32 intel_panel_get_backlight(struct drm_device *dev);
 extern void intel_panel_set_backlight(struct drm_device *dev, u32 level);
+extern void intel_panel_setup_backlight(struct drm_device *dev);
+extern void intel_panel_enable_backlight(struct drm_device *dev);
+extern void intel_panel_disable_backlight(struct drm_device *dev);
 
 extern void intel_crtc_load_lut(struct drm_crtc *crtc);
 extern void intel_encoder_prepare (struct drm_encoder *encoder);
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index 701e830d0012..ee145a257287 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -62,6 +62,7 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 			  struct drm_fb_helper_surface_size *sizes)
 {
 	struct drm_device *dev = ifbdev->helper.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct fb_info *info;
 	struct drm_framebuffer *fb;
 	struct drm_mode_fb_cmd mode_cmd;
@@ -77,7 +78,7 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 	mode_cmd.height = sizes->surface_height;
 
 	mode_cmd.bpp = sizes->surface_bpp;
-	mode_cmd.pitch = ALIGN(mode_cmd.width * ((mode_cmd.bpp + 1) / 8), 64);
+	mode_cmd.pitch = ALIGN(mode_cmd.width * ((mode_cmd.bpp + 7) / 8), 64);
 	mode_cmd.depth = sizes->surface_depth;
 
 	size = mode_cmd.pitch * mode_cmd.height;
@@ -120,6 +121,11 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 	info->flags = FBINFO_DEFAULT | FBINFO_CAN_FORCE_OUTPUT;
 	info->fbops = &intelfb_ops;
 
+	ret = fb_alloc_cmap(&info->cmap, 256, 0);
+	if (ret) {
+		ret = -ENOMEM;
+		goto out_unpin;
+	}
 	/* setup aperture base/size for vesafb takeover */
 	info->apertures = alloc_apertures(1);
 	if (!info->apertures) {
@@ -127,10 +133,8 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 		goto out_unpin;
 	}
 	info->apertures->ranges[0].base = dev->mode_config.fb_base;
-	if (!IS_GEN2(dev))
-		info->apertures->ranges[0].size = pci_resource_len(dev->pdev, 2);
-	else
-		info->apertures->ranges[0].size = pci_resource_len(dev->pdev, 0);
+	info->apertures->ranges[0].size =
+		dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT;
 
 	info->fix.smem_start = dev->mode_config.fb_base + obj->gtt_offset;
 	info->fix.smem_len = size;
@@ -140,12 +144,6 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 		ret = -ENOSPC;
 		goto out_unpin;
 	}
-
-	ret = fb_alloc_cmap(&info->cmap, 256, 0);
-	if (ret) {
-		ret = -ENOMEM;
-		goto out_unpin;
-	}
 	info->screen_size = size;
 
 //	memset(info->screen_base, 0, size);
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index aa2307080be2..8f4f6bd33ee9 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -106,7 +106,7 @@ static void intel_lvds_enable(struct intel_lvds *intel_lvds)
 	I915_WRITE(ctl_reg, I915_READ(ctl_reg) | POWER_TARGET_ON);
 	POSTING_READ(lvds_reg);
 
-	intel_panel_set_backlight(dev, dev_priv->backlight_level);
+	intel_panel_enable_backlight(dev);
 }
 
 static void intel_lvds_disable(struct intel_lvds *intel_lvds)
@@ -123,8 +123,7 @@ static void intel_lvds_disable(struct intel_lvds *intel_lvds)
 		lvds_reg = LVDS;
 	}
 
-	dev_priv->backlight_level = intel_panel_get_backlight(dev);
-	intel_panel_set_backlight(dev, 0);
+	intel_panel_disable_backlight(dev);
 
 	I915_WRITE(ctl_reg, I915_READ(ctl_reg) & ~POWER_TARGET_ON);
 
@@ -375,6 +374,10 @@ static bool intel_lvds_mode_fixup(struct drm_encoder *encoder,
 	}
 
 out:
+	if ((pfit_control & PFIT_ENABLE) == 0) {
+		pfit_control = 0;
+		pfit_pgm_ratios = 0;
+	}
 	if (pfit_control != intel_lvds->pfit_control ||
 	    pfit_pgm_ratios != intel_lvds->pfit_pgm_ratios) {
 		intel_lvds->pfit_control = pfit_control;
@@ -398,8 +401,6 @@ static void intel_lvds_prepare(struct drm_encoder *encoder)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_lvds *intel_lvds = to_intel_lvds(encoder);
 
-	dev_priv->backlight_level = intel_panel_get_backlight(dev);
-
 	/* We try to do the minimum that is necessary in order to unlock
 	 * the registers for mode setting.
 	 *
@@ -430,9 +431,6 @@ static void intel_lvds_commit(struct drm_encoder *encoder)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_lvds *intel_lvds = to_intel_lvds(encoder);
 
-	if (dev_priv->backlight_level == 0)
-		dev_priv->backlight_level = intel_panel_get_max_backlight(dev);
-
 	/* Undo any unlocking done in prepare to prevent accidental
 	 * adjustment of the registers.
 	 */
diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index 7350ec2515c6..e00d200df3db 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -250,3 +250,34 @@ void intel_panel_set_backlight(struct drm_device *dev, u32 level)
 		tmp &= ~BACKLIGHT_DUTY_CYCLE_MASK;
 	I915_WRITE(BLC_PWM_CTL, tmp | level);
 }
+
+void intel_panel_disable_backlight(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->backlight_enabled) {
+		dev_priv->backlight_level = intel_panel_get_backlight(dev);
+		dev_priv->backlight_enabled = false;
+	}
+
+	intel_panel_set_backlight(dev, 0);
+}
+
+void intel_panel_enable_backlight(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->backlight_level == 0)
+		dev_priv->backlight_level = intel_panel_get_max_backlight(dev);
+
+	intel_panel_set_backlight(dev, dev_priv->backlight_level);
+	dev_priv->backlight_enabled = true;
+}
+
+void intel_panel_setup_backlight(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	dev_priv->backlight_level = intel_panel_get_max_backlight(dev);
+	dev_priv->backlight_enabled = dev_priv->backlight_level != 0;
+}
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 56bc95c056dd..03e337072517 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -48,7 +48,7 @@ static u32 i915_gem_get_seqno(struct drm_device *dev)
 	return seqno;
 }
 
-static void
+static int
 render_ring_flush(struct intel_ring_buffer *ring,
 		  u32	invalidate_domains,
 		  u32	flush_domains)
@@ -56,6 +56,7 @@ render_ring_flush(struct intel_ring_buffer *ring,
 	struct drm_device *dev = ring->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	u32 cmd;
+	int ret;
 
 #if WATCH_EXEC
 	DRM_INFO("%s: invalidate %08x flush %08x\n", __func__,
@@ -116,12 +117,16 @@ render_ring_flush(struct intel_ring_buffer *ring,
 #if WATCH_EXEC
 		DRM_INFO("%s: queue flush %08x to ring\n", __func__, cmd);
 #endif
-		if (intel_ring_begin(ring, 2) == 0) {
-			intel_ring_emit(ring, cmd);
-			intel_ring_emit(ring, MI_NOOP);
-			intel_ring_advance(ring);
-		}
+		ret = intel_ring_begin(ring, 2);
+		if (ret)
+			return ret;
+
+		intel_ring_emit(ring, cmd);
+		intel_ring_emit(ring, MI_NOOP);
+		intel_ring_advance(ring);
 	}
+
+	return 0;
 }
 
 static void ring_write_tail(struct intel_ring_buffer *ring,
@@ -480,26 +485,56 @@ pc_render_get_seqno(struct intel_ring_buffer *ring)
 	return pc->cpu_page[0];
 }
 
+static void
+ironlake_enable_irq(drm_i915_private_t *dev_priv, u32 mask)
+{
+	dev_priv->gt_irq_mask &= ~mask;
+	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
+	POSTING_READ(GTIMR);
+}
+
+static void
+ironlake_disable_irq(drm_i915_private_t *dev_priv, u32 mask)
+{
+	dev_priv->gt_irq_mask |= mask;
+	I915_WRITE(GTIMR, dev_priv->gt_irq_mask);
+	POSTING_READ(GTIMR);
+}
+
+static void
+i915_enable_irq(drm_i915_private_t *dev_priv, u32 mask)
+{
+	dev_priv->irq_mask &= ~mask;
+	I915_WRITE(IMR, dev_priv->irq_mask);
+	POSTING_READ(IMR);
+}
+
+static void
+i915_disable_irq(drm_i915_private_t *dev_priv, u32 mask)
+{
+	dev_priv->irq_mask |= mask;
+	I915_WRITE(IMR, dev_priv->irq_mask);
+	POSTING_READ(IMR);
+}
+
 static bool
 render_ring_get_irq(struct intel_ring_buffer *ring)
 {
 	struct drm_device *dev = ring->dev;
+	drm_i915_private_t *dev_priv = dev->dev_private;
 
 	if (!dev->irq_enabled)
 		return false;
 
-	if (atomic_inc_return(&ring->irq_refcount) == 1) {
-		drm_i915_private_t *dev_priv = dev->dev_private;
-		unsigned long irqflags;
-
-		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	spin_lock(&ring->irq_lock);
+	if (ring->irq_refcount++ == 0) {
 		if (HAS_PCH_SPLIT(dev))
-			ironlake_enable_graphics_irq(dev_priv,
-						     GT_PIPE_NOTIFY | GT_USER_INTERRUPT);
+			ironlake_enable_irq(dev_priv,
+					    GT_PIPE_NOTIFY | GT_USER_INTERRUPT);
 		else
 			i915_enable_irq(dev_priv, I915_USER_INTERRUPT);
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 	}
+	spin_unlock(&ring->irq_lock);
 
 	return true;
 }
@@ -508,20 +543,18 @@ static void
 render_ring_put_irq(struct intel_ring_buffer *ring)
 {
 	struct drm_device *dev = ring->dev;
+	drm_i915_private_t *dev_priv = dev->dev_private;
 
-	if (atomic_dec_and_test(&ring->irq_refcount)) {
-		drm_i915_private_t *dev_priv = dev->dev_private;
-		unsigned long irqflags;
-
-		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
+	spin_lock(&ring->irq_lock);
+	if (--ring->irq_refcount == 0) {
 		if (HAS_PCH_SPLIT(dev))
-			ironlake_disable_graphics_irq(dev_priv,
-						      GT_USER_INTERRUPT |
-						      GT_PIPE_NOTIFY);
+			ironlake_disable_irq(dev_priv,
+					     GT_USER_INTERRUPT |
+					     GT_PIPE_NOTIFY);
 		else
 			i915_disable_irq(dev_priv, I915_USER_INTERRUPT);
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 	}
+	spin_unlock(&ring->irq_lock);
 }
 
 void intel_ring_setup_status_page(struct intel_ring_buffer *ring)
@@ -534,19 +567,24 @@ void intel_ring_setup_status_page(struct intel_ring_buffer *ring)
 	POSTING_READ(mmio);
 }
 
-static void
+static int
 bsd_ring_flush(struct intel_ring_buffer *ring,
 	       u32     invalidate_domains,
 	       u32     flush_domains)
 {
+	int ret;
+
 	if ((flush_domains & I915_GEM_DOMAIN_RENDER) == 0)
-		return;
+		return 0;
 
-	if (intel_ring_begin(ring, 2) == 0) {
-		intel_ring_emit(ring, MI_FLUSH);
-		intel_ring_emit(ring, MI_NOOP);
-		intel_ring_advance(ring);
-	}
+	ret = intel_ring_begin(ring, 2);
+	if (ret)
+		return ret;
+
+	intel_ring_emit(ring, MI_FLUSH);
+	intel_ring_emit(ring, MI_NOOP);
+	intel_ring_advance(ring);
+	return 0;
 }
 
 static int
@@ -577,18 +615,15 @@ static bool
 ring_get_irq(struct intel_ring_buffer *ring, u32 flag)
 {
 	struct drm_device *dev = ring->dev;
+	drm_i915_private_t *dev_priv = dev->dev_private;
 
 	if (!dev->irq_enabled)
 	       return false;
 
-	if (atomic_inc_return(&ring->irq_refcount) == 1) {
-		drm_i915_private_t *dev_priv = dev->dev_private;
-		unsigned long irqflags;
-
-		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
-		ironlake_enable_graphics_irq(dev_priv, flag);
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
-	}
+	spin_lock(&ring->irq_lock);
+	if (ring->irq_refcount++ == 0)
+		ironlake_enable_irq(dev_priv, flag);
+	spin_unlock(&ring->irq_lock);
 
 	return true;
 }
@@ -597,15 +632,47 @@ static void
 ring_put_irq(struct intel_ring_buffer *ring, u32 flag)
 {
 	struct drm_device *dev = ring->dev;
+	drm_i915_private_t *dev_priv = dev->dev_private;
 
-	if (atomic_dec_and_test(&ring->irq_refcount)) {
-		drm_i915_private_t *dev_priv = dev->dev_private;
-		unsigned long irqflags;
+	spin_lock(&ring->irq_lock);
+	if (--ring->irq_refcount == 0)
+		ironlake_disable_irq(dev_priv, flag);
+	spin_unlock(&ring->irq_lock);
+}
 
-		spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
-		ironlake_disable_graphics_irq(dev_priv, flag);
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
+static bool
+gen6_ring_get_irq(struct intel_ring_buffer *ring, u32 gflag, u32 rflag)
+{
+	struct drm_device *dev = ring->dev;
+	drm_i915_private_t *dev_priv = dev->dev_private;
+
+	if (!dev->irq_enabled)
+	       return false;
+
+	spin_lock(&ring->irq_lock);
+	if (ring->irq_refcount++ == 0) {
+		ring->irq_mask &= ~rflag;
+		I915_WRITE_IMR(ring, ring->irq_mask);
+		ironlake_enable_irq(dev_priv, gflag);
 	}
+	spin_unlock(&ring->irq_lock);
+
+	return true;
+}
+
+static void
+gen6_ring_put_irq(struct intel_ring_buffer *ring, u32 gflag, u32 rflag)
+{
+	struct drm_device *dev = ring->dev;
+	drm_i915_private_t *dev_priv = dev->dev_private;
+
+	spin_lock(&ring->irq_lock);
+	if (--ring->irq_refcount == 0) {
+		ring->irq_mask |= rflag;
+		I915_WRITE_IMR(ring, ring->irq_mask);
+		ironlake_disable_irq(dev_priv, gflag);
+	}
+	spin_unlock(&ring->irq_lock);
 }
 
 static bool
@@ -748,6 +815,9 @@ int intel_init_ring_buffer(struct drm_device *dev,
 	INIT_LIST_HEAD(&ring->request_list);
 	INIT_LIST_HEAD(&ring->gpu_write_list);
 
+	spin_lock_init(&ring->irq_lock);
+	ring->irq_mask = ~0;
+
 	if (I915_NEED_GFX_HWS(dev)) {
 		ret = init_status_page(ring);
 		if (ret)
@@ -785,6 +855,14 @@ int intel_init_ring_buffer(struct drm_device *dev,
 	if (ret)
 		goto err_unmap;
 
+	/* Workaround an erratum on the i830 which causes a hang if
+	 * the TAIL pointer points to within the last 2 cachelines
+	 * of the buffer.
+	 */
+	ring->effective_size = ring->size;
+	if (IS_I830(ring->dev))
+		ring->effective_size -= 128;
+
 	return 0;
 
 err_unmap:
@@ -827,8 +905,7 @@ void intel_cleanup_ring_buffer(struct intel_ring_buffer *ring)
 static int intel_wrap_ring_buffer(struct intel_ring_buffer *ring)
 {
 	unsigned int *virt;
-	int rem;
-	rem = ring->size - ring->tail;
+	int rem = ring->size - ring->tail;
 
 	if (ring->space < rem) {
 		int ret = intel_wait_ring_buffer(ring, rem);
@@ -895,7 +972,7 @@ int intel_ring_begin(struct intel_ring_buffer *ring,
 	int n = 4*num_dwords;
 	int ret;
 
-	if (unlikely(ring->tail + n > ring->size)) {
+	if (unlikely(ring->tail + n > ring->effective_size)) {
 		ret = intel_wrap_ring_buffer(ring);
 		if (unlikely(ret))
 			return ret;
@@ -973,20 +1050,25 @@ static void gen6_bsd_ring_write_tail(struct intel_ring_buffer *ring,
 	       GEN6_BSD_SLEEP_PSMI_CONTROL_RC_ILDL_MESSAGE_ENABLE);
 }
 
-static void gen6_ring_flush(struct intel_ring_buffer *ring,
-			    u32 invalidate_domains,
-			    u32 flush_domains)
+static int gen6_ring_flush(struct intel_ring_buffer *ring,
+			   u32 invalidate_domains,
+			   u32 flush_domains)
 {
+	int ret;
+
 	if ((flush_domains & I915_GEM_DOMAIN_RENDER) == 0)
-		return;
+		return 0;
 
-	if (intel_ring_begin(ring, 4) == 0) {
-		intel_ring_emit(ring, MI_FLUSH_DW);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-		intel_ring_advance(ring);
-	}
+	ret = intel_ring_begin(ring, 4);
+	if (ret)
+		return ret;
+
+	intel_ring_emit(ring, MI_FLUSH_DW);
+	intel_ring_emit(ring, 0);
+	intel_ring_emit(ring, 0);
+	intel_ring_emit(ring, 0);
+	intel_ring_advance(ring);
+	return 0;
 }
 
 static int
@@ -1008,15 +1090,35 @@ gen6_ring_dispatch_execbuffer(struct intel_ring_buffer *ring,
 }
 
 static bool
+gen6_render_ring_get_irq(struct intel_ring_buffer *ring)
+{
+	return gen6_ring_get_irq(ring,
+				 GT_USER_INTERRUPT,
+				 GEN6_RENDER_USER_INTERRUPT);
+}
+
+static void
+gen6_render_ring_put_irq(struct intel_ring_buffer *ring)
+{
+	return gen6_ring_put_irq(ring,
+				 GT_USER_INTERRUPT,
+				 GEN6_RENDER_USER_INTERRUPT);
+}
+
+static bool
 gen6_bsd_ring_get_irq(struct intel_ring_buffer *ring)
 {
-	return ring_get_irq(ring, GT_GEN6_BSD_USER_INTERRUPT);
+	return gen6_ring_get_irq(ring,
+				 GT_GEN6_BSD_USER_INTERRUPT,
+				 GEN6_BSD_USER_INTERRUPT);
 }
 
 static void
 gen6_bsd_ring_put_irq(struct intel_ring_buffer *ring)
 {
-	ring_put_irq(ring, GT_GEN6_BSD_USER_INTERRUPT);
+	return gen6_ring_put_irq(ring,
+				 GT_GEN6_BSD_USER_INTERRUPT,
+				 GEN6_BSD_USER_INTERRUPT);
 }
 
 /* ring buffer for Video Codec for Gen6+ */
@@ -1040,13 +1142,17 @@ static const struct intel_ring_buffer gen6_bsd_ring = {
 static bool
 blt_ring_get_irq(struct intel_ring_buffer *ring)
 {
-	return ring_get_irq(ring, GT_BLT_USER_INTERRUPT);
+	return gen6_ring_get_irq(ring,
+				 GT_BLT_USER_INTERRUPT,
+				 GEN6_BLITTER_USER_INTERRUPT);
 }
 
 static void
 blt_ring_put_irq(struct intel_ring_buffer *ring)
 {
-	ring_put_irq(ring, GT_BLT_USER_INTERRUPT);
+	gen6_ring_put_irq(ring,
+			  GT_BLT_USER_INTERRUPT,
+			  GEN6_BLITTER_USER_INTERRUPT);
 }
 
 
@@ -1115,20 +1221,25 @@ static int blt_ring_begin(struct intel_ring_buffer *ring,
 		return intel_ring_begin(ring, 4);
 }
 
-static void blt_ring_flush(struct intel_ring_buffer *ring,
+static int blt_ring_flush(struct intel_ring_buffer *ring,
 			   u32 invalidate_domains,
 			   u32 flush_domains)
 {
+	int ret;
+
 	if ((flush_domains & I915_GEM_DOMAIN_RENDER) == 0)
-		return;
+		return 0;
 
-	if (blt_ring_begin(ring, 4) == 0) {
-		intel_ring_emit(ring, MI_FLUSH_DW);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-		intel_ring_advance(ring);
-	}
+	ret = blt_ring_begin(ring, 4);
+	if (ret)
+		return ret;
+
+	intel_ring_emit(ring, MI_FLUSH_DW);
+	intel_ring_emit(ring, 0);
+	intel_ring_emit(ring, 0);
+	intel_ring_emit(ring, 0);
+	intel_ring_advance(ring);
+	return 0;
 }
 
 static void blt_ring_cleanup(struct intel_ring_buffer *ring)
@@ -1165,6 +1276,8 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 	*ring = render_ring;
 	if (INTEL_INFO(dev)->gen >= 6) {
 		ring->add_request = gen6_add_request;
+		ring->irq_get = gen6_render_ring_get_irq;
+		ring->irq_put = gen6_render_ring_put_irq;
 	} else if (IS_GEN5(dev)) {
 		ring->add_request = pc_render_add_request;
 		ring->get_seqno = pc_render_get_seqno;
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index 8e2e357ad6ee..be9087e4c9be 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -16,21 +16,24 @@ struct  intel_hw_status_page {
 
 #define I915_RING_READ(reg) i915_safe_read(dev_priv, reg)
 
-#define I915_READ_TAIL(ring) I915_RING_READ(RING_TAIL(ring->mmio_base))
-#define I915_WRITE_TAIL(ring, val) I915_WRITE(RING_TAIL(ring->mmio_base), val)
+#define I915_READ_TAIL(ring) I915_RING_READ(RING_TAIL((ring)->mmio_base))
+#define I915_WRITE_TAIL(ring, val) I915_WRITE(RING_TAIL((ring)->mmio_base), val)
 
-#define I915_READ_START(ring) I915_RING_READ(RING_START(ring->mmio_base))
-#define I915_WRITE_START(ring, val) I915_WRITE(RING_START(ring->mmio_base), val)
+#define I915_READ_START(ring) I915_RING_READ(RING_START((ring)->mmio_base))
+#define I915_WRITE_START(ring, val) I915_WRITE(RING_START((ring)->mmio_base), val)
 
-#define I915_READ_HEAD(ring)  I915_RING_READ(RING_HEAD(ring->mmio_base))
-#define I915_WRITE_HEAD(ring, val) I915_WRITE(RING_HEAD(ring->mmio_base), val)
+#define I915_READ_HEAD(ring)  I915_RING_READ(RING_HEAD((ring)->mmio_base))
+#define I915_WRITE_HEAD(ring, val) I915_WRITE(RING_HEAD((ring)->mmio_base), val)
 
-#define I915_READ_CTL(ring) I915_RING_READ(RING_CTL(ring->mmio_base))
-#define I915_WRITE_CTL(ring, val) I915_WRITE(RING_CTL(ring->mmio_base), val)
+#define I915_READ_CTL(ring) I915_RING_READ(RING_CTL((ring)->mmio_base))
+#define I915_WRITE_CTL(ring, val) I915_WRITE(RING_CTL((ring)->mmio_base), val)
 
-#define I915_READ_NOPID(ring) I915_RING_READ(RING_NOPID(ring->mmio_base))
-#define I915_READ_SYNC_0(ring) I915_RING_READ(RING_SYNC_0(ring->mmio_base))
-#define I915_READ_SYNC_1(ring) I915_RING_READ(RING_SYNC_1(ring->mmio_base))
+#define I915_WRITE_IMR(ring, val) I915_WRITE(RING_IMR((ring)->mmio_base), val)
+#define I915_READ_IMR(ring) I915_RING_READ(RING_IMR((ring)->mmio_base))
+
+#define I915_READ_NOPID(ring) I915_RING_READ(RING_NOPID((ring)->mmio_base))
+#define I915_READ_SYNC_0(ring) I915_RING_READ(RING_SYNC_0((ring)->mmio_base))
+#define I915_READ_SYNC_1(ring) I915_RING_READ(RING_SYNC_1((ring)->mmio_base))
 
 struct  intel_ring_buffer {
 	const char	*name;
@@ -49,12 +52,15 @@ struct  intel_ring_buffer {
 	u32		tail;
 	int		space;
 	int		size;
+	int		effective_size;
 	struct intel_hw_status_page status_page;
 
+	spinlock_t	irq_lock;
+	u32		irq_refcount;
+	u32		irq_mask;
 	u32		irq_seqno;		/* last seq seem at irq time */
 	u32		waiting_seqno;
 	u32		sync_seqno[I915_NUM_RINGS-1];
-	atomic_t	irq_refcount;
 	bool __must_check (*irq_get)(struct intel_ring_buffer *ring);
 	void		(*irq_put)(struct intel_ring_buffer *ring);
 
@@ -62,9 +68,9 @@ struct  intel_ring_buffer {
 
 	void		(*write_tail)(struct intel_ring_buffer *ring,
 				      u32 value);
-	void		(*flush)(struct intel_ring_buffer *ring,
-				 u32	invalidate_domains,
-				 u32	flush_domains);
+	int __must_check (*flush)(struct intel_ring_buffer *ring,
+				  u32	invalidate_domains,
+				  u32	flush_domains);
 	int		(*add_request)(struct intel_ring_buffer *ring,
 				       u32 *seqno);
 	u32		(*get_seqno)(struct intel_ring_buffer *ring);
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 9d0af36a13ec..45cd37652a37 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -1024,9 +1024,13 @@ static void intel_sdvo_mode_set(struct drm_encoder *encoder,
 	if (!intel_sdvo_set_target_input(intel_sdvo))
 		return;
 
-	if (intel_sdvo->has_hdmi_monitor &&
-	    !intel_sdvo_set_avi_infoframe(intel_sdvo))
-		return;
+	if (intel_sdvo->has_hdmi_monitor) {
+		intel_sdvo_set_encode(intel_sdvo, SDVO_ENCODE_HDMI);
+		intel_sdvo_set_colorimetry(intel_sdvo,
+					   SDVO_COLORIMETRY_RGB256);
+		intel_sdvo_set_avi_infoframe(intel_sdvo);
+	} else
+		intel_sdvo_set_encode(intel_sdvo, SDVO_ENCODE_DVI);
 
 	if (intel_sdvo->is_tv &&
 	    !intel_sdvo_set_tv_format(intel_sdvo))
@@ -1398,6 +1402,9 @@ intel_sdvo_detect(struct drm_connector *connector, bool force)
 
 	intel_sdvo->attached_output = response;
 
+	intel_sdvo->has_hdmi_monitor = false;
+	intel_sdvo->has_hdmi_audio = false;
+
 	if ((intel_sdvo_connector->output_flag & response) == 0)
 		ret = connector_status_disconnected;
 	else if (response & SDVO_TMDS_MASK)
@@ -1922,20 +1929,7 @@ intel_sdvo_select_i2c_bus(struct drm_i915_private *dev_priv,
 static bool
 intel_sdvo_is_hdmi_connector(struct intel_sdvo *intel_sdvo, int device)
 {
-	int is_hdmi;
-
-	if (!intel_sdvo_check_supp_encode(intel_sdvo))
-		return false;
-
-	if (!intel_sdvo_set_target_output(intel_sdvo,
-					  device == 0 ? SDVO_OUTPUT_TMDS0 : SDVO_OUTPUT_TMDS1))
-		return false;
-
-	is_hdmi = 0;
-	if (!intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_ENCODE, &is_hdmi, 1))
-		return false;
-
-	return !!is_hdmi;
+	return intel_sdvo_check_supp_encode(intel_sdvo);
 }
 
 static u8
@@ -2037,12 +2031,7 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 	connector->connector_type = DRM_MODE_CONNECTOR_DVID;
 
 	if (intel_sdvo_is_hdmi_connector(intel_sdvo, device)) {
-		/* enable hdmi encoding mode if supported */
-		intel_sdvo_set_encode(intel_sdvo, SDVO_ENCODE_HDMI);
-		intel_sdvo_set_colorimetry(intel_sdvo,
-					   SDVO_COLORIMETRY_RGB256);
 		connector->connector_type = DRM_MODE_CONNECTOR_HDMIA;
-
 		intel_sdvo->is_hdmi = true;
 	}
 	intel_sdvo->base.clone_mask = ((1 << INTEL_SDVO_NON_TV_CLONE_BIT) |