Merge tag 'drm-intel-next-2017-03-06' of git://anongit.freedesktop.org/git/drm-intel into drm-next

4 weeks worth of stuff since I was traveling&lazy: - lspcon improvements (Imre) - proper atomic state for cdclk handling (Ville) - gpu reset improvements (Chris) - lots and lots of polish around fences, requests, waiting and everything related all over (both gem and modeset code), from Chris - atomic by default on gen5+ minus byt/bsw (Maarten did the patch to flip the default, really this is a massive joint team effort) - moar power domains, now 64bit (Ander) - big pile of in-kernel unit tests for various gem subsystems (Chris), including simple mock objects for i915 device and and the ggtt manager. - i915_gpu_info in debugfs, for taking a snapshot of the current gpu state. Same thing as i915_error_state, but useful if the kernel didn't notice something is stick. From Chris. - bxt dsi fixes (Umar Shankar) - bxt w/a updates (Jani) - no more struct_mutex for gem object unreference (Chris) - some execlist refactoring (Tvrtko) - color manager support for glk (Ander) - improve the power-well sync code to better take over from the firmware (Imre) - gem tracepoint polish (Tvrtko) - lots of glk fixes all around (Ander) - ctx switch improvements (Chris) - glk dsi support&fixes (Deepak M) - dsi fixes for vlv and clanups, lots of them (Hans de Goede) - switch to i915.ko types in lots of our internal modeset code (Ander) - byt/bsw atomic wm update code, yay (Ville) * tag 'drm-intel-next-2017-03-06' of git://anongit.freedesktop.org/git/drm-intel: (432 commits) drm/i915: Update DRIVER_DATE to 20170306 drm/i915: Don't use enums for hardware engine id drm/i915: Split breadcrumbs spinlock into two drm/i915: Refactor wakeup of the next breadcrumb waiter drm/i915: Take reference for signaling the request from hardirq drm/i915: Add FIFO underrun tracepoints drm/i915: Add cxsr toggle tracepoint drm/i915: Add VLV/CHV watermark/FIFO programming tracepoints drm/i915: Add plane update/disable tracepoints drm/i915: Kill level 0 wm hack for VLV/CHV drm/i915: Workaround VLV/CHV sprite1->sprite0 enable underrun drm/i915: Sanitize VLV/CHV watermarks properly drm/i915: Only use update_wm_{pre,post} for pre-ilk platforms drm/i915: Nuke crtc->wm.cxsr_allowed drm/i915: Compute proper intermediate wms for vlv/cvh drm/i915: Skip useless watermark/FIFO related work on VLV/CHV when not needed drm/i915: Compute vlv/chv wms the atomic way drm/i915: Compute VLV/CHV FIFO sizes based on the PM2 watermarks drm/i915: Plop vlv/chv fifo sizes into crtc state drm/i915: Plop vlv wm state into crtc_state ...
author: Dave Airlie <airlied@redhat.com> 2017-03-08 12:41:47 +1000
committer: Dave Airlie <airlied@redhat.com> 2017-03-08 12:41:47 +1000
commit: 2e16101780e9cc8c4c68566db002e7513a1530eb (patch)
tree: e3fa976f5fe68aea853c3653be13168bb6b43438 /drivers/gpu/drm
parent: b558dfd56a5c1f915327967ecfe1181cf2a7a494 (diff)
parent: 505b681539a7e14aeb866515d3ef1a67375839bc (diff)
download: linux-2e16101780e9cc8c4c68566db002e7513a1530eb.tar.gz
126 files changed, 18670 insertions, 8192 deletions
diff --git a/drivers/gpu/drm/i915/Kconfig b/drivers/gpu/drm/i915/Kconfig
index 183f5dc1c3f2..1ae0bb91ee60 100644
--- a/drivers/gpu/drm/i915/Kconfig
+++ b/drivers/gpu/drm/i915/Kconfig
@@ -19,6 +19,7 @@ config DRM_I915
 	select INPUT if ACPI
 	select ACPI_VIDEO if ACPI
 	select ACPI_BUTTON if ACPI
+	select SYNC_FILE
 	help
 	  Choose this option if you have a system that has "Intel Graphics
 	  Media Accelerator" or "HD Graphics" integrated graphics,
diff --git a/drivers/gpu/drm/i915/Kconfig.debug b/drivers/gpu/drm/i915/Kconfig.debug
index 597648c7a645..e091809a9a9e 100644
--- a/drivers/gpu/drm/i915/Kconfig.debug
+++ b/drivers/gpu/drm/i915/Kconfig.debug
@@ -24,7 +24,9 @@ config DRM_I915_DEBUG
         select X86_MSR # used by igt/pm_rpm
         select DRM_VGEM # used by igt/prime_vgem (dmabuf interop checks)
         select DRM_DEBUG_MM if DRM=y
+	select DRM_DEBUG_MM_SELFTEST
 	select DRM_I915_SW_FENCE_DEBUG_OBJECTS
+	select DRM_I915_SELFTEST
         default n
         help
           Choose this option to turn on extra driver debugging that may affect
@@ -58,3 +60,30 @@ config DRM_I915_SW_FENCE_DEBUG_OBJECTS
           Recommended for driver developers only.
 
           If in doubt, say "N".
+
+config DRM_I915_SELFTEST
+	bool "Enable selftests upon driver load"
+	depends on DRM_I915
+	default n
+	select FAULT_INJECTION
+	select PRIME_NUMBERS
+	help
+	  Choose this option to allow the driver to perform selftests upon
+	  loading; also requires the i915.selftest=1 module parameter. To
+	  exit the module after running the selftests (i.e. to prevent normal
+	  module initialisation afterwards) use i915.selftest=-1.
+
+	  Recommended for driver developers only.
+
+	  If in doubt, say "N".
+
+config DRM_I915_LOW_LEVEL_TRACEPOINTS
+        bool "Enable low level request tracing events"
+        depends on DRM_I915
+        default n
+        help
+          Choose this option to turn on low level request tracing events.
+          This provides the ability to precisely monitor engine utilisation
+          and also analyze the request dependency resolving timeline.
+
+          If in doubt, say "N".
diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index c62ab45683c0..b1b580337c7a 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -29,6 +29,7 @@ i915-$(CONFIG_DEBUG_FS) += i915_debugfs.o intel_pipe_crc.o
 # GEM code
 i915-y += i915_cmd_parser.o \
 	  i915_gem_batch_pool.o \
+	  i915_gem_clflush.o \
 	  i915_gem_context.o \
 	  i915_gem_dmabuf.o \
 	  i915_gem_evict.o \
@@ -72,6 +73,7 @@ i915-y += intel_audio.o \
 	  intel_atomic.o \
 	  intel_atomic_plane.o \
 	  intel_bios.o \
+	  intel_cdclk.o \
 	  intel_color.o \
 	  intel_display.o \
 	  intel_dpio_phy.o \
@@ -116,6 +118,9 @@ i915-y += dvo_ch7017.o \
 
 # Post-mortem debug and GPU hang state capture
 i915-$(CONFIG_DRM_I915_CAPTURE_ERROR) += i915_gpu_error.o
+i915-$(CONFIG_DRM_I915_SELFTEST) += \
+	selftests/i915_random.o \
+	selftests/i915_selftest.o
 
 # virtual gpu code
 i915-y += i915_vgpu.o
diff --git a/drivers/gpu/drm/i915/gvt/cmd_parser.c b/drivers/gpu/drm/i915/gvt/cmd_parser.c
index b9c8e2407682..81076d8ec41a 100644
--- a/drivers/gpu/drm/i915/gvt/cmd_parser.c
+++ b/drivers/gpu/drm/i915/gvt/cmd_parser.c
@@ -1530,7 +1530,7 @@ static int copy_gma_to_hva(struct intel_vgpu *vgpu, struct intel_vgpu_mm *mm,
 		len += copy_len;
 		gma += copy_len;
 	}
-	return 0;
+	return len;
 }
 
 
@@ -1644,7 +1644,7 @@ static int perform_bb_shadow(struct parser_exec_state *s)
 	ret = copy_gma_to_hva(s->vgpu, s->vgpu->gtt.ggtt_mm,
 			      gma, gma + bb_size,
 			      dst);
-	if (ret) {
+	if (ret < 0) {
 		gvt_err("fail to copy guest ring buffer\n");
 		goto unmap_src;
 	}
@@ -2608,11 +2608,8 @@ out:
 static int shadow_workload_ring_buffer(struct intel_vgpu_workload *workload)
 {
 	struct intel_vgpu *vgpu = workload->vgpu;
-	int ring_id = workload->ring_id;
-	struct i915_gem_context *shadow_ctx = vgpu->shadow_ctx;
-	struct intel_ring *ring = shadow_ctx->engine[ring_id].ring;
 	unsigned long gma_head, gma_tail, gma_top, guest_rb_size;
-	unsigned int copy_len = 0;
+	u32 *cs;
 	int ret;
 
 	guest_rb_size = _RING_CTL_BUF_SIZE(workload->rb_ctl);
@@ -2626,36 +2623,33 @@ static int shadow_workload_ring_buffer(struct intel_vgpu_workload *workload)
 	gma_top = workload->rb_start + guest_rb_size;
 
 	/* allocate shadow ring buffer */
-	ret = intel_ring_begin(workload->req, workload->rb_len / 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(workload->req, workload->rb_len / sizeof(u32));
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	/* get shadow ring buffer va */
-	workload->shadow_ring_buffer_va = ring->vaddr + ring->tail;
+	workload->shadow_ring_buffer_va = cs;
 
 	/* head > tail --> copy head <-> top */
 	if (gma_head > gma_tail) {
 		ret = copy_gma_to_hva(vgpu, vgpu->gtt.ggtt_mm,
-				gma_head, gma_top,
-				workload->shadow_ring_buffer_va);
-		if (ret) {
+				      gma_head, gma_top, cs);
+		if (ret < 0) {
 			gvt_err("fail to copy guest ring buffer\n");
 			return ret;
 		}
-		copy_len = gma_top - gma_head;
+		cs += ret / sizeof(u32);
 		gma_head = workload->rb_start;
 	}
 
 	/* copy head or start <-> tail */
-	ret = copy_gma_to_hva(vgpu, vgpu->gtt.ggtt_mm,
-			gma_head, gma_tail,
-			workload->shadow_ring_buffer_va + copy_len);
-	if (ret) {
+	ret = copy_gma_to_hva(vgpu, vgpu->gtt.ggtt_mm, gma_head, gma_tail, cs);
+	if (ret < 0) {
 		gvt_err("fail to copy guest ring buffer\n");
 		return ret;
 	}
-	ring->tail += workload->rb_len;
-	intel_ring_advance(ring);
+	cs += ret / sizeof(u32);
+	intel_ring_advance(workload->req, cs);
 	return 0;
 }
 
@@ -2709,7 +2703,7 @@ static int shadow_indirect_ctx(struct intel_shadow_wa_ctx *wa_ctx)
 				wa_ctx->workload->vgpu->gtt.ggtt_mm,
 				guest_gma, guest_gma + ctx_size,
 				map);
-	if (ret) {
+	if (ret < 0) {
 		gvt_err("fail to copy guest indirect ctx\n");
 		goto unmap_src;
 	}
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 7d7244798507..aa2d726b4349 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -35,6 +35,23 @@ static inline struct drm_i915_private *node_to_i915(struct drm_info_node *node)
 	return to_i915(node->minor->dev);
 }
 
+static __always_inline void seq_print_param(struct seq_file *m,
+					    const char *name,
+					    const char *type,
+					    const void *x)
+{
+	if (!__builtin_strcmp(type, "bool"))
+		seq_printf(m, "i915.%s=%s\n", name, yesno(*(const bool *)x));
+	else if (!__builtin_strcmp(type, "int"))
+		seq_printf(m, "i915.%s=%d\n", name, *(const int *)x);
+	else if (!__builtin_strcmp(type, "unsigned int"))
+		seq_printf(m, "i915.%s=%u\n", name, *(const unsigned int *)x);
+	else if (!__builtin_strcmp(type, "char *"))
+		seq_printf(m, "i915.%s=%s\n", name, *(const char **)x);
+	else
+		BUILD_BUG();
+}
+
 static int i915_capabilities(struct seq_file *m, void *data)
 {
 	struct drm_i915_private *dev_priv = node_to_i915(m->private);
@@ -43,10 +60,17 @@ static int i915_capabilities(struct seq_file *m, void *data)
 	seq_printf(m, "gen: %d\n", INTEL_GEN(dev_priv));
 	seq_printf(m, "platform: %s\n", intel_platform_name(info->platform));
 	seq_printf(m, "pch: %d\n", INTEL_PCH_TYPE(dev_priv));
+
 #define PRINT_FLAG(x)  seq_printf(m, #x ": %s\n", yesno(info->x))
 	DEV_INFO_FOR_EACH_FLAG(PRINT_FLAG);
 #undef PRINT_FLAG
 
+	kernel_param_lock(THIS_MODULE);
+#define PRINT_PARAM(T, x) seq_print_param(m, #x, #T, &i915.x);
+	I915_PARAMS_FOR_EACH(PRINT_PARAM);
+#undef PRINT_PARAM
+	kernel_param_unlock(THIS_MODULE);
+
 	return 0;
 }
 
@@ -428,7 +452,7 @@ static int i915_gem_object_info(struct seq_file *m, void *data)
 		   dpy_count, dpy_size);
 
 	seq_printf(m, "%llu [%llu] gtt total\n",
-		   ggtt->base.total, ggtt->mappable_end - ggtt->base.start);
+		   ggtt->base.total, ggtt->mappable_end);
 
 	seq_putc(m, '\n');
 	print_batch_pool_stats(m, dev_priv);
@@ -456,7 +480,7 @@ static int i915_gem_object_info(struct seq_file *m, void *data)
 		mutex_lock(&dev->struct_mutex);
 		request = list_first_entry_or_null(&file_priv->mm.request_list,
 						   struct drm_i915_gem_request,
-						   client_list);
+						   client_link);
 		rcu_read_lock();
 		task = pid_task(request && request->ctx->pid ?
 				request->ctx->pid : file->pid,
@@ -676,14 +700,14 @@ static void i915_ring_seqno_info(struct seq_file *m,
 	seq_printf(m, "Current sequence (%s): %x\n",
 		   engine->name, intel_engine_get_seqno(engine));
 
-	spin_lock_irq(&b->lock);
+	spin_lock_irq(&b->rb_lock);
 	for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) {
 		struct intel_wait *w = rb_entry(rb, typeof(*w), node);
 
 		seq_printf(m, "Waiting (%s): %s [%d] on %x\n",
 			   engine->name, w->tsk->comm, w->tsk->pid, w->seqno);
 	}
-	spin_unlock_irq(&b->lock);
+	spin_unlock_irq(&b->rb_lock);
 }
 
 static int i915_gem_seqno_info(struct seq_file *m, void *data)
@@ -827,10 +851,22 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 			   I915_READ(VLV_IIR_RW));
 		seq_printf(m, "Display IMR:\t%08x\n",
 			   I915_READ(VLV_IMR));
-		for_each_pipe(dev_priv, pipe)
+		for_each_pipe(dev_priv, pipe) {
+			enum intel_display_power_domain power_domain;
+
+			power_domain = POWER_DOMAIN_PIPE(pipe);
+			if (!intel_display_power_get_if_enabled(dev_priv,
+								power_domain)) {
+				seq_printf(m, "Pipe %c power disabled\n",
+					   pipe_name(pipe));
+				continue;
+			}
+
 			seq_printf(m, "Pipe %c stat:\t%08x\n",
 				   pipe_name(pipe),
 				   I915_READ(PIPESTAT(pipe)));
+			intel_display_power_put(dev_priv, power_domain);
+		}
 
 		seq_printf(m, "Master IER:\t%08x\n",
 			   I915_READ(VLV_MASTER_IER));
@@ -928,101 +964,96 @@ static int i915_gem_fence_regs_info(struct seq_file *m, void *data)
 }
 
 #if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
-
-static ssize_t
-i915_error_state_write(struct file *filp,
-		       const char __user *ubuf,
-		       size_t cnt,
-		       loff_t *ppos)
+static ssize_t gpu_state_read(struct file *file, char __user *ubuf,
+			      size_t count, loff_t *pos)
 {
-	struct i915_error_state_file_priv *error_priv = filp->private_data;
-
-	DRM_DEBUG_DRIVER("Resetting error state\n");
-	i915_destroy_error_state(error_priv->i915);
+	struct i915_gpu_state *error = file->private_data;
+	struct drm_i915_error_state_buf str;
+	ssize_t ret;
+	loff_t tmp;
 
-	return cnt;
-}
-
-static int i915_error_state_open(struct inode *inode, struct file *file)
-{
-	struct drm_i915_private *dev_priv = inode->i_private;
-	struct i915_error_state_file_priv *error_priv;
+	if (!error)
+		return 0;
 
-	error_priv = kzalloc(sizeof(*error_priv), GFP_KERNEL);
-	if (!error_priv)
-		return -ENOMEM;
+	ret = i915_error_state_buf_init(&str, error->i915, count, *pos);
+	if (ret)
+		return ret;
 
-	error_priv->i915 = dev_priv;
+	ret = i915_error_state_to_str(&str, error);
+	if (ret)
+		goto out;
 
-	i915_error_state_get(&dev_priv->drm, error_priv);
+	tmp = 0;
+	ret = simple_read_from_buffer(ubuf, count, &tmp, str.buf, str.bytes);
+	if (ret < 0)
+		goto out;
 
-	file->private_data = error_priv;
+	*pos = str.start + ret;
+out:
+	i915_error_state_buf_release(&str);
+	return ret;
+}
 
+static int gpu_state_release(struct inode *inode, struct file *file)
+{
+	i915_gpu_state_put(file->private_data);
 	return 0;
 }
 
-static int i915_error_state_release(struct inode *inode, struct file *file)
+static int i915_gpu_info_open(struct inode *inode, struct file *file)
 {
-	struct i915_error_state_file_priv *error_priv = file->private_data;
+	struct i915_gpu_state *gpu;
 
-	i915_error_state_put(error_priv);
-	kfree(error_priv);
+	gpu = i915_capture_gpu_state(inode->i_private);
+	if (!gpu)
+		return -ENOMEM;
 
+	file->private_data = gpu;
 	return 0;
 }
 
-static ssize_t i915_error_state_read(struct file *file, char __user *userbuf,
-				     size_t count, loff_t *pos)
+static const struct file_operations i915_gpu_info_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_gpu_info_open,
+	.read = gpu_state_read,
+	.llseek = default_llseek,
+	.release = gpu_state_release,
+};
+
+static ssize_t
+i915_error_state_write(struct file *filp,
+		       const char __user *ubuf,
+		       size_t cnt,
+		       loff_t *ppos)
 {
-	struct i915_error_state_file_priv *error_priv = file->private_data;
-	struct drm_i915_error_state_buf error_str;
-	loff_t tmp_pos = 0;
-	ssize_t ret_count = 0;
-	int ret;
+	struct i915_gpu_state *error = filp->private_data;
 
-	ret = i915_error_state_buf_init(&error_str, error_priv->i915,
-					count, *pos);
-	if (ret)
-		return ret;
+	if (!error)
+		return 0;
 
-	ret = i915_error_state_to_str(&error_str, error_priv);
-	if (ret)
-		goto out;
+	DRM_DEBUG_DRIVER("Resetting error state\n");
+	i915_reset_error_state(error->i915);
 
-	ret_count = simple_read_from_buffer(userbuf, count, &tmp_pos,
-					    error_str.buf,
-					    error_str.bytes);
+	return cnt;
+}
 
-	if (ret_count < 0)
-		ret = ret_count;
-	else
-		*pos = error_str.start + ret_count;
-out:
-	i915_error_state_buf_release(&error_str);
-	return ret ?: ret_count;
+static int i915_error_state_open(struct inode *inode, struct file *file)
+{
+	file->private_data = i915_first_error_state(inode->i_private);
+	return 0;
 }
 
 static const struct file_operations i915_error_state_fops = {
 	.owner = THIS_MODULE,
 	.open = i915_error_state_open,
-	.read = i915_error_state_read,
+	.read = gpu_state_read,
 	.write = i915_error_state_write,
 	.llseek = default_llseek,
-	.release = i915_error_state_release,
+	.release = gpu_state_release,
 };
-
 #endif
 
 static int
-i915_next_seqno_get(void *data, u64 *val)
-{
-	struct drm_i915_private *dev_priv = data;
-
-	*val = 1 + atomic_read(&dev_priv->gt.global_timeline.seqno);
-	return 0;
-}
-
-static int
 i915_next_seqno_set(void *data, u64 val)
 {
 	struct drm_i915_private *dev_priv = data;
@@ -1040,13 +1071,12 @@ i915_next_seqno_set(void *data, u64 val)
 }
 
 DEFINE_SIMPLE_ATTRIBUTE(i915_next_seqno_fops,
-			i915_next_seqno_get, i915_next_seqno_set,
+			NULL, i915_next_seqno_set,
 			"0x%llx\n");
 
 static int i915_frequency_info(struct seq_file *m, void *unused)
 {
 	struct drm_i915_private *dev_priv = node_to_i915(m->private);
-	struct drm_device *dev = &dev_priv->drm;
 	int ret = 0;
 
 	intel_runtime_pm_get(dev_priv);
@@ -1109,10 +1139,6 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 		}
 
 		/* RPSTAT1 is in the GT power well */
-		ret = mutex_lock_interruptible(&dev->struct_mutex);
-		if (ret)
-			goto out;
-
 		intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL);
 
 		reqf = I915_READ(GEN6_RPNSWREQ);
@@ -1147,7 +1173,6 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 		cagf = intel_gpu_freq(dev_priv, cagf);
 
 		intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);
-		mutex_unlock(&dev->struct_mutex);
 
 		if (IS_GEN6(dev_priv) || IS_GEN7(dev_priv)) {
 			pm_ier = I915_READ(GEN6_PMIER);
@@ -1198,21 +1223,18 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 
 		max_freq = (IS_GEN9_LP(dev_priv) ? rp_state_cap >> 0 :
 			    rp_state_cap >> 16) & 0xff;
-		max_freq *= (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv) ?
-			     GEN9_FREQ_SCALER : 1);
+		max_freq *= (IS_GEN9_BC(dev_priv) ? GEN9_FREQ_SCALER : 1);
 		seq_printf(m, "Lowest (RPN) frequency: %dMHz\n",
 			   intel_gpu_freq(dev_priv, max_freq));
 
 		max_freq = (rp_state_cap & 0xff00) >> 8;
-		max_freq *= (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv) ?
-			     GEN9_FREQ_SCALER : 1);
+		max_freq *= (IS_GEN9_BC(dev_priv) ? GEN9_FREQ_SCALER : 1);
 		seq_printf(m, "Nominal (RP1) frequency: %dMHz\n",
 			   intel_gpu_freq(dev_priv, max_freq));
 
 		max_freq = (IS_GEN9_LP(dev_priv) ? rp_state_cap >> 16 :
 			    rp_state_cap >> 0) & 0xff;
-		max_freq *= (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv) ?
-			     GEN9_FREQ_SCALER : 1);
+		max_freq *= (IS_GEN9_BC(dev_priv) ? GEN9_FREQ_SCALER : 1);
 		seq_printf(m, "Max non-overclocked (RP0) frequency: %dMHz\n",
 			   intel_gpu_freq(dev_priv, max_freq));
 		seq_printf(m, "Max overclocked frequency: %dMHz\n",
@@ -1236,11 +1258,10 @@ static int i915_frequency_info(struct seq_file *m, void *unused)
 		seq_puts(m, "no P-state info available\n");
 	}
 
-	seq_printf(m, "Current CD clock frequency: %d kHz\n", dev_priv->cdclk_freq);
+	seq_printf(m, "Current CD clock frequency: %d kHz\n", dev_priv->cdclk.hw.cdclk);
 	seq_printf(m, "Max CD clock frequency: %d kHz\n", dev_priv->max_cdclk_freq);
 	seq_printf(m, "Max pixel clock frequency: %d kHz\n", dev_priv->max_dotclk_freq);
 
-out:
 	intel_runtime_pm_put(dev_priv);
 	return ret;
 }
@@ -1307,35 +1328,40 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused)
 
 	intel_runtime_pm_put(dev_priv);
 
-	if (delayed_work_pending(&dev_priv->gpu_error.hangcheck_work)) {
-		seq_printf(m, "Hangcheck active, fires in %dms\n",
+	if (timer_pending(&dev_priv->gpu_error.hangcheck_work.timer))
+		seq_printf(m, "Hangcheck active, timer fires in %dms\n",
 			   jiffies_to_msecs(dev_priv->gpu_error.hangcheck_work.timer.expires -
 					    jiffies));
-	} else
-		seq_printf(m, "Hangcheck inactive\n");
+	else if (delayed_work_pending(&dev_priv->gpu_error.hangcheck_work))
+		seq_puts(m, "Hangcheck active, work pending\n");
+	else
+		seq_puts(m, "Hangcheck inactive\n");
+
+	seq_printf(m, "GT active? %s\n", yesno(dev_priv->gt.awake));
 
 	for_each_engine(engine, dev_priv, id) {
 		struct intel_breadcrumbs *b = &engine->breadcrumbs;
 		struct rb_node *rb;
 
 		seq_printf(m, "%s:\n", engine->name);
-		seq_printf(m, "\tseqno = %x [current %x, last %x]\n",
+		seq_printf(m, "\tseqno = %x [current %x, last %x], inflight %d\n",
 			   engine->hangcheck.seqno, seqno[id],
-			   intel_engine_last_submit(engine));
+			   intel_engine_last_submit(engine),
+			   engine->timeline->inflight_seqnos);
 		seq_printf(m, "\twaiters? %s, fake irq active? %s, stalled? %s\n",
 			   yesno(intel_engine_has_waiter(engine)),
 			   yesno(test_bit(engine->id,
 					  &dev_priv->gpu_error.missed_irq_rings)),
 			   yesno(engine->hangcheck.stalled));
 
-		spin_lock_irq(&b->lock);
+		spin_lock_irq(&b->rb_lock);
 		for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) {
 			struct intel_wait *w = rb_entry(rb, typeof(*w), node);
 
 			seq_printf(m, "\t%s [%d] waiting for %x\n",
 				   w->tsk->comm, w->tsk->pid, w->seqno);
 		}
-		spin_unlock_irq(&b->lock);
+		spin_unlock_irq(&b->rb_lock);
 
 		seq_printf(m, "\tACTHD = 0x%08llx [current 0x%08llx]\n",
 			   (long long)engine->hangcheck.acthd,
@@ -1788,7 +1814,7 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused)
 	if (ret)
 		goto out;
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	if (IS_GEN9_BC(dev_priv)) {
 		/* Convert GT frequency to 50 HZ units */
 		min_gpu_freq =
 			dev_priv->rps.min_freq_softlimit / GEN9_FREQ_SCALER;
@@ -1808,8 +1834,8 @@ static int i915_ring_freq_table(struct seq_file *m, void *unused)
 				       &ia_freq);
 		seq_printf(m, "%d\t\t%d\t\t\t\t%d\n",
 			   intel_gpu_freq(dev_priv, (gpu_freq *
-				(IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv) ?
-				 GEN9_FREQ_SCALER : 1))),
+						     (IS_GEN9_BC(dev_priv) ?
+						      GEN9_FREQ_SCALER : 1))),
 			   ((ia_freq >> 0) & 0xff) * 100,
 			   ((ia_freq >> 8) & 0xff) * 100);
 	}
@@ -2302,10 +2328,10 @@ static int i915_rps_boost_info(struct seq_file *m, void *data)
 		seq_printf(m, "\nRPS Autotuning (current \"%s\" window):\n",
 			   rps_power_to_str(dev_priv->rps.power));
 		seq_printf(m, "  Avg. up: %d%% [above threshold? %d%%]\n",
-			   100 * rpup / rpupei,
+			   rpup && rpupei ? 100 * rpup / rpupei : 0,
 			   dev_priv->rps.up_threshold);
 		seq_printf(m, "  Avg. down: %d%% [below threshold? %d%%]\n",
-			   100 * rpdown / rpdownei,
+			   rpdown && rpdownei ? 100 * rpdown / rpdownei : 0,
 			   dev_priv->rps.down_threshold);
 	} else {
 		seq_puts(m, "\nRPS Autotuning inactive\n");
@@ -2351,7 +2377,9 @@ static int i915_huc_load_status_info(struct seq_file *m, void *data)
 	seq_printf(m, "\tRSA: offset is %d; size = %d\n",
 		huc_fw->rsa_offset, huc_fw->rsa_size);
 
+	intel_runtime_pm_get(dev_priv);
 	seq_printf(m, "\nHuC status 0x%08x:\n", I915_READ(HUC_STATUS2));
+	intel_runtime_pm_put(dev_priv);
 
 	return 0;
 }
@@ -2383,6 +2411,8 @@ static int i915_guc_load_status_info(struct seq_file *m, void *data)
 	seq_printf(m, "\tRSA: offset is %d; size = %d\n",
 		guc_fw->rsa_offset, guc_fw->rsa_size);
 
+	intel_runtime_pm_get(dev_priv);
+
 	tmp = I915_READ(GUC_STATUS);
 
 	seq_printf(m, "\nGuC status 0x%08x:\n", tmp);
@@ -2396,6 +2426,8 @@ static int i915_guc_load_status_info(struct seq_file *m, void *data)
 	for (i = 0; i < 16; i++)
 		seq_printf(m, "\t%2d: \t0x%x\n", i, I915_READ(SOFT_SCRATCH(i)));
 
+	intel_runtime_pm_put(dev_priv);
+
 	return 0;
 }
 
@@ -2777,15 +2809,10 @@ static int i915_power_domain_info(struct seq_file *m, void *unused)
 		seq_printf(m, "%-25s %d\n", power_well->name,
 			   power_well->count);
 
-		for (power_domain = 0; power_domain < POWER_DOMAIN_NUM;
-		     power_domain++) {
-			if (!(BIT(power_domain) & power_well->domains))
-				continue;
-
+		for_each_power_domain(power_domain, power_well->domains)
 			seq_printf(m, "  %-23s %d\n",
 				 intel_display_power_domain_str(power_domain),
 				 power_domains->domain_use_count[power_domain]);
-		}
 	}
 
 	mutex_unlock(&power_domains->lock);
@@ -3205,6 +3232,11 @@ static int i915_engine_info(struct seq_file *m, void *unused)
 
 	intel_runtime_pm_get(dev_priv);
 
+	seq_printf(m, "GT awake? %s\n",
+		   yesno(dev_priv->gt.awake));
+	seq_printf(m, "Global active requests: %d\n",
+		   dev_priv->gt.active_requests);
+
 	for_each_engine(engine, dev_priv, id) {
 		struct intel_breadcrumbs *b = &engine->breadcrumbs;
 		struct drm_i915_gem_request *rq;
@@ -3212,11 +3244,12 @@ static int i915_engine_info(struct seq_file *m, void *unused)
 		u64 addr;
 
 		seq_printf(m, "%s\n", engine->name);
-		seq_printf(m, "\tcurrent seqno %x, last %x, hangcheck %x [%d ms]\n",
+		seq_printf(m, "\tcurrent seqno %x, last %x, hangcheck %x [%d ms], inflight %d\n",
 			   intel_engine_get_seqno(engine),
 			   intel_engine_last_submit(engine),
 			   engine->hangcheck.seqno,
-			   jiffies_to_msecs(jiffies - engine->hangcheck.action_timestamp));
+			   jiffies_to_msecs(jiffies - engine->hangcheck.action_timestamp),
+			   engine->timeline->inflight_seqnos);
 
 		rcu_read_lock();
 
@@ -3294,15 +3327,21 @@ static int i915_engine_info(struct seq_file *m, void *unused)
 
 			rcu_read_lock();
 			rq = READ_ONCE(engine->execlist_port[0].request);
-			if (rq)
-				print_request(m, rq, "\t\tELSP[0] ");
-			else
+			if (rq) {
+				seq_printf(m, "\t\tELSP[0] count=%d, ",
+					   engine->execlist_port[0].count);
+				print_request(m, rq, "rq: ");
+			} else {
 				seq_printf(m, "\t\tELSP[0] idle\n");
+			}
 			rq = READ_ONCE(engine->execlist_port[1].request);
-			if (rq)
-				print_request(m, rq, "\t\tELSP[1] ");
-			else
+			if (rq) {
+				seq_printf(m, "\t\tELSP[1] count=%d, ",
+					   engine->execlist_port[1].count);
+				print_request(m, rq, "rq: ");
+			} else {
 				seq_printf(m, "\t\tELSP[1] idle\n");
+			}
 			rcu_read_unlock();
 
 			spin_lock_irq(&engine->timeline->lock);
@@ -3320,14 +3359,14 @@ static int i915_engine_info(struct seq_file *m, void *unused)
 				   I915_READ(RING_PP_DIR_DCLV(engine)));
 		}
 
-		spin_lock_irq(&b->lock);
+		spin_lock_irq(&b->rb_lock);
 		for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) {
 			struct intel_wait *w = rb_entry(rb, typeof(*w), node);
 
 			seq_printf(m, "\t%s [%d] waiting for %x\n",
 				   w->tsk->comm, w->tsk->pid, w->seqno);
 		}
-		spin_unlock_irq(&b->lock);
+		spin_unlock_irq(&b->rb_lock);
 
 		seq_puts(m, "\n");
 	}
@@ -3746,7 +3785,19 @@ static int i915_displayport_test_data_show(struct seq_file *m, void *data)
 		if (connector->status == connector_status_connected &&
 		    connector->encoder != NULL) {
 			intel_dp = enc_to_intel_dp(connector->encoder);
-			seq_printf(m, "%lx", intel_dp->compliance.test_data.edid);
+			if (intel_dp->compliance.test_type ==
+			    DP_TEST_LINK_EDID_READ)
+				seq_printf(m, "%lx",
+					   intel_dp->compliance.test_data.edid);
+			else if (intel_dp->compliance.test_type ==
+				 DP_TEST_LINK_VIDEO_PATTERN) {
+				seq_printf(m, "hdisplay: %d\n",
+					   intel_dp->compliance.test_data.hdisplay);
+				seq_printf(m, "vdisplay: %d\n",
+					   intel_dp->compliance.test_data.vdisplay);
+				seq_printf(m, "bpc: %u\n",
+					   intel_dp->compliance.test_data.bpc);
+			}
 		} else
 			seq_puts(m, "0");
 	}
@@ -4237,7 +4288,8 @@ i915_max_freq_set(void *data, u64 val)
 
 	dev_priv->rps.max_freq_softlimit = val;
 
-	intel_set_rps(dev_priv, val);
+	if (intel_set_rps(dev_priv, val))
+		DRM_DEBUG_DRIVER("failed to update RPS to new softlimit\n");
 
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
@@ -4292,7 +4344,8 @@ i915_min_freq_set(void *data, u64 val)
 
 	dev_priv->rps.min_freq_softlimit = val;
 
-	intel_set_rps(dev_priv, val);
+	if (intel_set_rps(dev_priv, val))
+		DRM_DEBUG_DRIVER("failed to update RPS to new softlimit\n");
 
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
@@ -4418,7 +4471,7 @@ static void gen9_sseu_device_status(struct drm_i915_private *dev_priv,
 
 		sseu->slice_mask |= BIT(s);
 
-		if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+		if (IS_GEN9_BC(dev_priv))
 			sseu->subslice_mask =
 				INTEL_INFO(dev_priv)->sseu.subslice_mask;
 
@@ -4567,6 +4620,81 @@ static const struct file_operations i915_forcewake_fops = {
 	.release = i915_forcewake_release,
 };
 
+static int i915_hpd_storm_ctl_show(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = m->private;
+	struct i915_hotplug *hotplug = &dev_priv->hotplug;
+
+	seq_printf(m, "Threshold: %d\n", hotplug->hpd_storm_threshold);
+	seq_printf(m, "Detected: %s\n",
+		   yesno(delayed_work_pending(&hotplug->reenable_work)));
+
+	return 0;
+}
+
+static ssize_t i915_hpd_storm_ctl_write(struct file *file,
+					const char __user *ubuf, size_t len,
+					loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	struct drm_i915_private *dev_priv = m->private;
+	struct i915_hotplug *hotplug = &dev_priv->hotplug;
+	unsigned int new_threshold;
+	int i;
+	char *newline;
+	char tmp[16];
+
+	if (len >= sizeof(tmp))
+		return -EINVAL;
+
+	if (copy_from_user(tmp, ubuf, len))
+		return -EFAULT;
+
+	tmp[len] = '\0';
+
+	/* Strip newline, if any */
+	newline = strchr(tmp, '\n');
+	if (newline)
+		*newline = '\0';
+
+	if (strcmp(tmp, "reset") == 0)
+		new_threshold = HPD_STORM_DEFAULT_THRESHOLD;
+	else if (kstrtouint(tmp, 10, &new_threshold) != 0)
+		return -EINVAL;
+
+	if (new_threshold > 0)
+		DRM_DEBUG_KMS("Setting HPD storm detection threshold to %d\n",
+			      new_threshold);
+	else
+		DRM_DEBUG_KMS("Disabling HPD storm detection\n");
+
+	spin_lock_irq(&dev_priv->irq_lock);
+	hotplug->hpd_storm_threshold = new_threshold;
+	/* Reset the HPD storm stats so we don't accidentally trigger a storm */
+	for_each_hpd_pin(i)
+		hotplug->stats[i].count = 0;
+	spin_unlock_irq(&dev_priv->irq_lock);
+
+	/* Re-enable hpd immediately if we were in an irq storm */
+	flush_delayed_work(&dev_priv->hotplug.reenable_work);
+
+	return len;
+}
+
+static int i915_hpd_storm_ctl_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, i915_hpd_storm_ctl_show, inode->i_private);
+}
+
+static const struct file_operations i915_hpd_storm_ctl_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_hpd_storm_ctl_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = i915_hpd_storm_ctl_write
+};
+
 static const struct drm_info_list i915_debugfs_list[] = {
 	{"i915_capabilities", i915_capabilities, 0},
 	{"i915_gem_objects", i915_gem_object_info, 0},
@@ -4633,6 +4761,7 @@ static const struct i915_debugfs_files {
 	{"i915_gem_drop_caches", &i915_drop_caches_fops},
 #if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
 	{"i915_error_state", &i915_error_state_fops},
+	{"i915_gpu_info", &i915_gpu_info_fops},
 #endif
 	{"i915_next_seqno", &i915_next_seqno_fops},
 	{"i915_display_crc_ctl", &i915_display_crc_ctl_fops},
@@ -4643,7 +4772,8 @@ static const struct i915_debugfs_files {
 	{"i915_dp_test_data", &i915_displayport_test_data_fops},
 	{"i915_dp_test_type", &i915_displayport_test_type_fops},
 	{"i915_dp_test_active", &i915_displayport_test_active_fops},
-	{"i915_guc_log_control", &i915_guc_log_control_fops}
+	{"i915_guc_log_control", &i915_guc_log_control_fops},
+	{"i915_hpd_storm_ctl", &i915_hpd_storm_ctl_fops}
 };
 
 int i915_debugfs_register(struct drm_i915_private *dev_priv)
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 655d146e1126..704dbcf63866 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -43,6 +43,7 @@
 
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
+#include <drm/drm_atomic_helper.h>
 #include <drm/i915_drm.h>
 
 #include "i915_drv.h"
@@ -248,6 +249,7 @@ static int i915_getparam(struct drm_device *dev, void *data,
 	case I915_PARAM_IRQ_ACTIVE:
 	case I915_PARAM_ALLOW_BATCHBUFFER:
 	case I915_PARAM_LAST_DISPATCH:
+	case I915_PARAM_HAS_EXEC_CONSTANTS:
 		/* Reject all old ums/dri params. */
 		return -ENODEV;
 	case I915_PARAM_CHIPSET_ID:
@@ -274,9 +276,6 @@ static int i915_getparam(struct drm_device *dev, void *data,
 	case I915_PARAM_HAS_BSD2:
 		value = !!dev_priv->engine[VCS2];
 		break;
-	case I915_PARAM_HAS_EXEC_CONSTANTS:
-		value = INTEL_GEN(dev_priv) >= 4;
-		break;
 	case I915_PARAM_HAS_LLC:
 		value = HAS_LLC(dev_priv);
 		break;
@@ -318,10 +317,9 @@ static int i915_getparam(struct drm_device *dev, void *data,
 		value = INTEL_INFO(dev_priv)->sseu.min_eu_in_pool;
 		break;
 	case I915_PARAM_HUC_STATUS:
-		/* The register is already force-woken. We dont need
-		 * any rpm here
-		 */
+		intel_runtime_pm_get(dev_priv);
 		value = I915_READ(HUC_STATUS2) & HUC_FW_VERIFIED;
+		intel_runtime_pm_put(dev_priv);
 		break;
 	case I915_PARAM_MMAP_GTT_VERSION:
 		/* Though we've started our numbering from 1, and so class all
@@ -350,6 +348,8 @@ static int i915_getparam(struct drm_device *dev, void *data,
 	case I915_PARAM_HAS_EXEC_HANDLE_LUT:
 	case I915_PARAM_HAS_COHERENT_PHYS_GTT:
 	case I915_PARAM_HAS_EXEC_SOFTPIN:
+	case I915_PARAM_HAS_EXEC_ASYNC:
+	case I915_PARAM_HAS_EXEC_FENCE:
 		/* For the time being all of these are always true;
 		 * if some supported hardware does not have one of these
 		 * features this value needs to be provided from
@@ -756,6 +756,15 @@ out_err:
 	return -ENOMEM;
 }
 
+static void i915_engines_cleanup(struct drm_i915_private *i915)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, i915, id)
+		kfree(engine);
+}
+
 static void i915_workqueues_cleanup(struct drm_i915_private *dev_priv)
 {
 	destroy_workqueue(dev_priv->hotplug.dp_wq);
@@ -769,10 +778,17 @@ static void i915_workqueues_cleanup(struct drm_i915_private *dev_priv)
  */
 static void intel_detect_preproduction_hw(struct drm_i915_private *dev_priv)
 {
-	if (IS_HSW_EARLY_SDV(dev_priv) ||
-	    IS_SKL_REVID(dev_priv, 0, SKL_REVID_F0))
+	bool pre = false;
+
+	pre |= IS_HSW_EARLY_SDV(dev_priv);
+	pre |= IS_SKL_REVID(dev_priv, 0, SKL_REVID_F0);
+	pre |= IS_BXT_REVID(dev_priv, 0, BXT_REVID_B_LAST);
+
+	if (pre) {
 		DRM_ERROR("This is a pre-production stepping. "
 			  "It may not be fully functional.\n");
+		add_taint(TAINT_MACHINE_CHECK, LOCKDEP_STILL_OK);
+	}
 }
 
 /**
@@ -808,6 +824,7 @@ static int i915_driver_init_early(struct drm_i915_private *dev_priv,
 	spin_lock_init(&dev_priv->gpu_error.lock);
 	mutex_init(&dev_priv->backlight_lock);
 	spin_lock_init(&dev_priv->uncore.lock);
+
 	spin_lock_init(&dev_priv->mm.object_stat_lock);
 	spin_lock_init(&dev_priv->mmio_flip_lock);
 	spin_lock_init(&dev_priv->wm.dsparb_lock);
@@ -818,12 +835,15 @@ static int i915_driver_init_early(struct drm_i915_private *dev_priv,
 	mutex_init(&dev_priv->pps_mutex);
 
 	intel_uc_init_early(dev_priv);
-
 	i915_memcpy_init_early(dev_priv);
 
+	ret = intel_engines_init_early(dev_priv);
+	if (ret)
+		return ret;
+
 	ret = i915_workqueues_init(dev_priv);
 	if (ret < 0)
-		return ret;
+		goto err_engines;
 
 	/* This must be called before any calls to HAS_PCH_* */
 	intel_detect_pch(dev_priv);
@@ -852,6 +872,8 @@ static int i915_driver_init_early(struct drm_i915_private *dev_priv,
 
 err_workqueues:
 	i915_workqueues_cleanup(dev_priv);
+err_engines:
+	i915_engines_cleanup(dev_priv);
 	return ret;
 }
 
@@ -864,6 +886,7 @@ static void i915_driver_cleanup_early(struct drm_i915_private *dev_priv)
 	i915_perf_fini(dev_priv);
 	i915_gem_load_cleanup(dev_priv);
 	i915_workqueues_cleanup(dev_priv);
+	i915_engines_cleanup(dev_priv);
 }
 
 static int i915_mmio_setup(struct drm_i915_private *dev_priv)
@@ -930,6 +953,7 @@ static int i915_driver_init_mmio(struct drm_i915_private *dev_priv)
 		goto put_bridge;
 
 	intel_uncore_init(dev_priv);
+	i915_gem_init_mmio(dev_priv);
 
 	return 0;
 
@@ -967,7 +991,7 @@ static void intel_sanitize_options(struct drm_i915_private *dev_priv)
 	DRM_DEBUG_DRIVER("ppgtt mode: %i\n", i915.enable_ppgtt);
 
 	i915.semaphores = intel_sanitize_semaphores(dev_priv, i915.semaphores);
-	DRM_DEBUG_DRIVER("use GPU sempahores? %s\n", yesno(i915.semaphores));
+	DRM_DEBUG_DRIVER("use GPU semaphores? %s\n", yesno(i915.semaphores));
 }
 
 /**
@@ -1185,11 +1209,15 @@ static void i915_driver_unregister(struct drm_i915_private *dev_priv)
  */
 int i915_driver_load(struct pci_dev *pdev, const struct pci_device_id *ent)
 {
+	const struct intel_device_info *match_info =
+		(struct intel_device_info *)ent->driver_data;
 	struct drm_i915_private *dev_priv;
 	int ret;
 
-	if (i915.nuclear_pageflip)
-		driver.driver_features |= DRIVER_ATOMIC;
+	/* Enable nuclear pageflip on ILK+, except vlv/chv */
+	if (!i915.nuclear_pageflip &&
+	    (match_info->gen < 5 || match_info->has_gmch_display))
+		driver.driver_features &= ~DRIVER_ATOMIC;
 
 	ret = -ENOMEM;
 	dev_priv = kzalloc(sizeof(*dev_priv), GFP_KERNEL);
@@ -1197,8 +1225,7 @@ int i915_driver_load(struct pci_dev *pdev, const struct pci_device_id *ent)
 		ret = drm_dev_init(&dev_priv->drm, &driver, &pdev->dev);
 	if (ret) {
 		DRM_DEV_ERROR(&pdev->dev, "allocation failed\n");
-		kfree(dev_priv);
-		return ret;
+		goto out_free;
 	}
 
 	dev_priv->drm.pdev = pdev;
@@ -1206,7 +1233,7 @@ int i915_driver_load(struct pci_dev *pdev, const struct pci_device_id *ent)
 
 	ret = pci_enable_device(pdev);
 	if (ret)
-		goto out_free_priv;
+		goto out_fini;
 
 	pci_set_drvdata(pdev, &dev_priv->drm);
 
@@ -1270,9 +1297,11 @@ out_runtime_pm_put:
 	i915_driver_cleanup_early(dev_priv);
 out_pci_disable:
 	pci_disable_device(pdev);
-out_free_priv:
+out_fini:
 	i915_load_error(dev_priv, "Device initialization failed (%d)\n", ret);
-	drm_dev_unref(&dev_priv->drm);
+	drm_dev_fini(&dev_priv->drm);
+out_free:
+	kfree(dev_priv);
 	return ret;
 }
 
@@ -1280,6 +1309,8 @@ void i915_driver_unload(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct pci_dev *pdev = dev_priv->drm.pdev;
+	struct drm_modeset_acquire_ctx ctx;
+	int ret;
 
 	intel_fbdev_fini(dev);
 
@@ -1288,6 +1319,24 @@ void i915_driver_unload(struct drm_device *dev)
 
 	intel_display_power_get(dev_priv, POWER_DOMAIN_INIT);
 
+	drm_modeset_acquire_init(&ctx, 0);
+	while (1) {
+		ret = drm_modeset_lock_all_ctx(dev, &ctx);
+		if (!ret)
+			ret = drm_atomic_helper_disable_all(dev, &ctx);
+
+		if (ret != -EDEADLK)
+			break;
+
+		drm_modeset_backoff(&ctx);
+	}
+
+	if (ret)
+		DRM_ERROR("Disabling all crtc's during unload failed with %i\n", ret);
+
+	drm_modeset_drop_locks(&ctx);
+	drm_modeset_acquire_fini(&ctx);
+
 	intel_gvt_cleanup(dev_priv);
 
 	i915_driver_unregister(dev_priv);
@@ -1317,7 +1366,7 @@ void i915_driver_unload(struct drm_device *dev)
 
 	/* Free error state after interrupts are fully disabled. */
 	cancel_delayed_work_sync(&dev_priv->gpu_error.hangcheck_work);
-	i915_destroy_error_state(dev_priv);
+	i915_reset_error_state(dev_priv);
 
 	/* Flush any outstanding unpin_work. */
 	drain_workqueue(dev_priv->wq);
@@ -1333,8 +1382,16 @@ void i915_driver_unload(struct drm_device *dev)
 	i915_driver_cleanup_mmio(dev_priv);
 
 	intel_display_power_put(dev_priv, POWER_DOMAIN_INIT);
+}
+
+static void i915_driver_release(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	i915_driver_cleanup_early(dev_priv);
+	drm_dev_fini(&dev_priv->drm);
+
+	kfree(dev_priv);
 }
 
 static int i915_driver_open(struct drm_device *dev, struct drm_file *file)
@@ -1716,6 +1773,8 @@ static int i915_drm_resume_early(struct drm_device *dev)
 	    !(dev_priv->suspended_to_idle && dev_priv->csr.dmc_payload))
 		intel_power_domains_init_hw(dev_priv, true);
 
+	i915_gem_sanitize(dev_priv);
+
 	enable_rpm_wakeref_asserts(dev_priv);
 
 out:
@@ -1787,7 +1846,7 @@ void i915_reset(struct drm_i915_private *dev_priv)
 		goto error;
 	}
 
-	i915_gem_reset_finish(dev_priv);
+	i915_gem_reset(dev_priv);
 	intel_overlay_reset(dev_priv);
 
 	/* Ok, now get things going again... */
@@ -1813,6 +1872,7 @@ void i915_reset(struct drm_i915_private *dev_priv)
 	i915_queue_hangcheck(dev_priv);
 
 wakeup:
+	i915_gem_reset_finish(dev_priv);
 	enable_irq(dev_priv->drm.irq);
 	wake_up_bit(&error->flags, I915_RESET_IN_PROGRESS);
 	return;
@@ -2532,7 +2592,7 @@ static const struct drm_ioctl_desc i915_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(I915_HWS_ADDR, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(I915_GEM_INIT, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(I915_GEM_EXECBUFFER, i915_gem_execbuffer, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_GEM_EXECBUFFER2, i915_gem_execbuffer2, DRM_AUTH|DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(I915_GEM_EXECBUFFER2_WR, i915_gem_execbuffer2, DRM_AUTH|DRM_RENDER_ALLOW),
 	DRM_IOCTL_DEF_DRV(I915_GEM_PIN, i915_gem_reject_pin_ioctl, DRM_AUTH|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(I915_GEM_UNPIN, i915_gem_reject_pin_ioctl, DRM_AUTH|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(I915_GEM_BUSY, i915_gem_busy_ioctl, DRM_AUTH|DRM_RENDER_ALLOW),
@@ -2574,7 +2634,8 @@ static struct drm_driver driver = {
 	 */
 	.driver_features =
 	    DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED | DRIVER_GEM | DRIVER_PRIME |
-	    DRIVER_RENDER | DRIVER_MODESET,
+	    DRIVER_RENDER | DRIVER_MODESET | DRIVER_ATOMIC,
+	.release = i915_driver_release,
 	.open = i915_driver_open,
 	.lastclose = i915_driver_lastclose,
 	.preclose = i915_driver_preclose,
@@ -2603,3 +2664,7 @@ static struct drm_driver driver = {
 	.minor = DRIVER_MINOR,
 	.patchlevel = DRIVER_PATCHLEVEL,
 };
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/mock_drm.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 8df73751c367..9ae1e520f48c 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -79,8 +79,8 @@
 
 #define DRIVER_NAME		"i915"
 #define DRIVER_DESC		"Intel Graphics"
-#define DRIVER_DATE		"20170123"
-#define DRIVER_TIMESTAMP	1485156432
+#define DRIVER_DATE		"20170306"
+#define DRIVER_TIMESTAMP	1488785683
 
 #undef WARN_ON
 /* Many gcc seem to no see through this and fall over :( */
@@ -293,6 +293,7 @@ enum plane_id {
 	PLANE_PRIMARY,
 	PLANE_SPRITE0,
 	PLANE_SPRITE1,
+	PLANE_SPRITE2,
 	PLANE_CURSOR,
 	I915_MAX_PLANES,
 };
@@ -343,6 +344,11 @@ enum intel_display_power_domain {
 	POWER_DOMAIN_PORT_DDI_C_LANES,
 	POWER_DOMAIN_PORT_DDI_D_LANES,
 	POWER_DOMAIN_PORT_DDI_E_LANES,
+	POWER_DOMAIN_PORT_DDI_A_IO,
+	POWER_DOMAIN_PORT_DDI_B_IO,
+	POWER_DOMAIN_PORT_DDI_C_IO,
+	POWER_DOMAIN_PORT_DDI_D_IO,
+	POWER_DOMAIN_PORT_DDI_E_IO,
 	POWER_DOMAIN_PORT_DSI,
 	POWER_DOMAIN_PORT_CRT,
 	POWER_DOMAIN_PORT_OTHER,
@@ -384,6 +390,8 @@ enum hpd_pin {
 #define for_each_hpd_pin(__pin) \
 	for ((__pin) = (HPD_NONE + 1); (__pin) < HPD_NUM_PINS; (__pin)++)
 
+#define HPD_STORM_DEFAULT_THRESHOLD 5
+
 struct i915_hotplug {
 	struct work_struct hotplug_work;
 
@@ -407,6 +415,8 @@ struct i915_hotplug {
 	struct work_struct poll_init_work;
 	bool poll_enabled;
 
+	unsigned int hpd_storm_threshold;
+
 	/*
 	 * if we get a HPD irq from DP and a HPD irq from non-DP
 	 * the non-DP HPD could block the workqueue on a mode config
@@ -494,7 +504,35 @@ struct i915_hotplug {
 
 #define for_each_power_domain(domain, mask)				\
 	for ((domain) = 0; (domain) < POWER_DOMAIN_NUM; (domain)++)	\
-		for_each_if ((1 << (domain)) & (mask))
+		for_each_if (BIT_ULL(domain) & (mask))
+
+#define for_each_power_well(__dev_priv, __power_well)				\
+	for ((__power_well) = (__dev_priv)->power_domains.power_wells;	\
+	     (__power_well) - (__dev_priv)->power_domains.power_wells <	\
+		(__dev_priv)->power_domains.power_well_count;		\
+	     (__power_well)++)
+
+#define for_each_power_well_rev(__dev_priv, __power_well)			\
+	for ((__power_well) = (__dev_priv)->power_domains.power_wells +		\
+			      (__dev_priv)->power_domains.power_well_count - 1;	\
+	     (__power_well) - (__dev_priv)->power_domains.power_wells >= 0;	\
+	     (__power_well)--)
+
+#define for_each_power_domain_well(__dev_priv, __power_well, __domain_mask)	\
+	for_each_power_well(__dev_priv, __power_well)				\
+		for_each_if ((__power_well)->domains & (__domain_mask))
+
+#define for_each_power_domain_well_rev(__dev_priv, __power_well, __domain_mask) \
+	for_each_power_well_rev(__dev_priv, __power_well)		        \
+		for_each_if ((__power_well)->domains & (__domain_mask))
+
+#define for_each_intel_plane_in_state(__state, plane, plane_state, __i) \
+	for ((__i) = 0; \
+	     (__i) < (__state)->base.dev->mode_config.num_total_plane && \
+		     ((plane) = to_intel_plane((__state)->base.planes[__i].ptr), \
+		      (plane_state) = to_intel_plane_state((__state)->base.planes[__i].state), 1); \
+	     (__i)++) \
+		for_each_if (plane_state)
 
 struct drm_i915_private;
 struct i915_mm_struct;
@@ -600,9 +638,13 @@ struct intel_initial_plane_config;
 struct intel_crtc;
 struct intel_limit;
 struct dpll;
+struct intel_cdclk_state;
 
 struct drm_i915_display_funcs {
-	int (*get_display_clock_speed)(struct drm_i915_private *dev_priv);
+	void (*get_cdclk)(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_state *cdclk_state);
+	void (*set_cdclk)(struct drm_i915_private *dev_priv,
+			  const struct intel_cdclk_state *cdclk_state);
 	int (*get_fifo_size)(struct drm_i915_private *dev_priv, int plane);
 	int (*compute_pipe_wm)(struct intel_crtc_state *cstate);
 	int (*compute_intermediate_wm)(struct drm_device *dev,
@@ -617,7 +659,6 @@ struct drm_i915_display_funcs {
 	int (*compute_global_watermarks)(struct drm_atomic_state *state);
 	void (*update_wm)(struct intel_crtc *crtc);
 	int (*modeset_calc_cdclk)(struct drm_atomic_state *state);
-	void (*modeset_commit_cdclk)(struct drm_atomic_state *state);
 	/* Returns the active state of the crtc, and if the crtc is active,
 	 * fills out the pipe-config with the hw state. */
 	bool (*get_pipe_config)(struct intel_crtc *,
@@ -636,7 +677,8 @@ struct drm_i915_display_funcs {
 				   struct intel_encoder *encoder,
 				   const struct drm_display_mode *adjusted_mode);
 	void (*audio_codec_disable)(struct intel_encoder *encoder);
-	void (*fdi_link_train)(struct drm_crtc *crtc);
+	void (*fdi_link_train)(struct intel_crtc *crtc,
+			       const struct intel_crtc_state *crtc_state);
 	void (*init_clock_gating)(struct drm_i915_private *dev_priv);
 	int (*queue_flip)(struct drm_device *dev, struct drm_crtc *crtc,
 			  struct drm_framebuffer *fb,
@@ -856,6 +898,7 @@ enum intel_platform {
 	INTEL_BROXTON,
 	INTEL_KABYLAKE,
 	INTEL_GEMINILAKE,
+	INTEL_MAX_PLATFORMS
 };
 
 struct intel_device_info {
@@ -890,7 +933,7 @@ struct intel_device_info {
 
 struct intel_display_error_state;
 
-struct drm_i915_error_state {
+struct i915_gpu_state {
 	struct kref ref;
 	struct timeval time;
 	struct timeval boottime;
@@ -900,16 +943,20 @@ struct drm_i915_error_state {
 
 	char error_msg[128];
 	bool simulated;
+	bool awake;
+	bool wakelock;
+	bool suspended;
 	int iommu;
 	u32 reset_count;
 	u32 suspend_count;
 	struct intel_device_info device_info;
+	struct i915_params params;
 
 	/* Generic register state */
 	u32 eir;
 	u32 pgtbl_er;
 	u32 ier;
-	u32 gtier[4];
+	u32 gtier[4], ngtier;
 	u32 ccid;
 	u32 derrmr;
 	u32 forcewake;
@@ -923,6 +970,7 @@ struct drm_i915_error_state {
 	u32 gab_ctl;
 	u32 gfx_mode;
 
+	u32 nfence;
 	u64 fence[I915_MAX_NUM_FENCES];
 	struct intel_overlay_error_state *overlay;
 	struct intel_display_error_state *display;
@@ -970,6 +1018,16 @@ struct drm_i915_error_state {
 		u32 semaphore_mboxes[I915_NUM_ENGINES - 1];
 		struct intel_instdone instdone;
 
+		struct drm_i915_error_context {
+			char comm[TASK_COMM_LEN];
+			pid_t pid;
+			u32 handle;
+			u32 hw_id;
+			int ban_score;
+			int active;
+			int guilty;
+		} context;
+
 		struct drm_i915_error_object {
 			u64 gtt_offset;
 			u64 gtt_size;
@@ -1003,10 +1061,6 @@ struct drm_i915_error_state {
 				u32 pp_dir_base;
 			};
 		} vm_info;
-
-		pid_t pid;
-		char comm[TASK_COMM_LEN];
-		int context_bans;
 	} engine[I915_NUM_ENGINES];
 
 	struct drm_i915_error_buffer {
@@ -1395,7 +1449,7 @@ struct i915_power_well {
 	int count;
 	/* cached hw enabled state */
 	bool hw_enabled;
-	unsigned long domains;
+	u64 domains;
 	/* unique identifier for this power well */
 	unsigned long id;
 	/*
@@ -1456,7 +1510,7 @@ struct i915_gem_mm {
 	struct work_struct free_work;
 
 	/** Usable portion of the GTT for GEM */
-	phys_addr_t stolen_base; /* limited to low memory (32-bit) */
+	dma_addr_t stolen_base; /* limited to low memory (32-bit) */
 
 	/** PPGTT used for aliasing the PPGTT with the GTT */
 	struct i915_hw_ppgtt *aliasing_ppgtt;
@@ -1498,11 +1552,6 @@ struct drm_i915_error_state_buf {
 	loff_t pos;
 };
 
-struct i915_error_state_file_priv {
-	struct drm_i915_private *i915;
-	struct drm_i915_error_state *error;
-};
-
 #define I915_RESET_TIMEOUT (10 * HZ) /* 10s */
 #define I915_FENCE_TIMEOUT (10 * HZ) /* 10s */
 
@@ -1519,7 +1568,7 @@ struct i915_gpu_error {
 	/* For reset and error_state handling. */
 	spinlock_t lock;
 	/* Protected by the above dev->gpu_error.lock. */
-	struct drm_i915_error_state *first_error;
+	struct i915_gpu_state *first_error;
 
 	unsigned long missed_irq_rings;
 
@@ -2053,6 +2102,10 @@ struct i915_oa_ops {
 	bool (*oa_buffer_is_empty)(struct drm_i915_private *dev_priv);
 };
 
+struct intel_cdclk_state {
+	unsigned int cdclk, vco, ref;
+};
+
 struct drm_i915_private {
 	struct drm_device drm;
 
@@ -2063,8 +2116,6 @@ struct drm_i915_private {
 
 	const struct intel_device_info info;
 
-	int relative_constants_mode;
-
 	void __iomem *regs;
 
 	struct intel_uncore uncore;
@@ -2157,13 +2208,7 @@ struct drm_i915_private {
 
 	unsigned int fsb_freq, mem_freq, is_ddr3;
 	unsigned int skl_preferred_vco_freq;
-	unsigned int cdclk_freq, max_cdclk_freq;
-
-	/*
-	 * For reading holding any crtc lock is sufficient,
-	 * for writing must hold all of them.
-	 */
-	unsigned int atomic_cdclk_freq;
+	unsigned int max_cdclk_freq;
 
 	unsigned int max_dotclk_freq;
 	unsigned int rawclk_freq;
@@ -2171,8 +2216,22 @@ struct drm_i915_private {
 	unsigned int czclk_freq;
 
 	struct {
-		unsigned int vco, ref;
-	} cdclk_pll;
+		/*
+		 * The current logical cdclk state.
+		 * See intel_atomic_state.cdclk.logical
+		 *
+		 * For reading holding any crtc lock is sufficient,
+		 * for writing must hold all of them.
+		 */
+		struct intel_cdclk_state logical;
+		/*
+		 * The current actual cdclk state.
+		 * See intel_atomic_state.cdclk.actual
+		 */
+		struct intel_cdclk_state actual;
+		/* The current hardware cdclk state */
+		struct intel_cdclk_state hw;
+	} cdclk;
 
 	/**
 	 * wq - Driver workqueue for GEM.
@@ -2752,6 +2811,12 @@ intel_info(const struct drm_i915_private *dev_priv)
 #define IS_KBL_REVID(dev_priv, since, until) \
 	(IS_KABYLAKE(dev_priv) && IS_REVID(dev_priv, since, until))
 
+#define GLK_REVID_A0		0x0
+#define GLK_REVID_A1		0x1
+
+#define IS_GLK_REVID(dev_priv, since, until) \
+	(IS_GEMINILAKE(dev_priv) && IS_REVID(dev_priv, since, until))
+
 /*
  * The genX designation typically refers to the render engine, so render
  * capability related checks should use IS_GEN, while display and other checks
@@ -2767,8 +2832,9 @@ intel_info(const struct drm_i915_private *dev_priv)
 #define IS_GEN8(dev_priv)	(!!((dev_priv)->info.gen_mask & BIT(7)))
 #define IS_GEN9(dev_priv)	(!!((dev_priv)->info.gen_mask & BIT(8)))
 
-#define IS_GEN9_LP(dev_priv)	(IS_GEN9(dev_priv) && INTEL_INFO(dev_priv)->is_lp)
 #define IS_LP(dev_priv)	(INTEL_INFO(dev_priv)->is_lp)
+#define IS_GEN9_LP(dev_priv)	(IS_GEN9(dev_priv) && IS_LP(dev_priv))
+#define IS_GEN9_BC(dev_priv)	(IS_GEN9(dev_priv) && !IS_LP(dev_priv))
 
 #define ENGINE_MASK(id)	BIT(id)
 #define RENDER_RING	ENGINE_MASK(RCS)
@@ -2810,9 +2876,7 @@ intel_info(const struct drm_i915_private *dev_priv)
 
 /* WaRsDisableCoarsePowerGating:skl,bxt */
 #define NEEDS_WaRsDisableCoarsePowerGating(dev_priv) \
-	(IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1) || \
-	 IS_SKL_GT3(dev_priv) || \
-	 IS_SKL_GT4(dev_priv))
+	(IS_SKL_GT3(dev_priv) || IS_SKL_GT4(dev_priv))
 
 /*
  * dp aux and gmbus irq on gen4 seems to be able to generate legacy interrupts
@@ -2952,6 +3016,9 @@ extern unsigned long i915_gfx_val(struct drm_i915_private *dev_priv);
 extern void i915_update_gfx_val(struct drm_i915_private *dev_priv);
 int vlv_force_gfx_clock(struct drm_i915_private *dev_priv, bool on);
 
+int intel_engines_init_early(struct drm_i915_private *dev_priv);
+int intel_engines_init(struct drm_i915_private *dev_priv);
+
 /* intel_hotplug.c */
 void intel_hpd_irq_handler(struct drm_i915_private *dev_priv,
 			   u32 pin_mask, u32 long_mask);
@@ -3129,6 +3196,7 @@ int i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *file_priv);
 int i915_gem_wait_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *file_priv);
+void i915_gem_sanitize(struct drm_i915_private *i915);
 int i915_gem_load_init(struct drm_i915_private *dev_priv);
 void i915_gem_load_cleanup(struct drm_i915_private *dev_priv);
 void i915_gem_load_init_fences(struct drm_i915_private *dev_priv);
@@ -3341,15 +3409,17 @@ static inline u32 i915_reset_count(struct i915_gpu_error *error)
 }
 
 int i915_gem_reset_prepare(struct drm_i915_private *dev_priv);
+void i915_gem_reset(struct drm_i915_private *dev_priv);
 void i915_gem_reset_finish(struct drm_i915_private *dev_priv);
 void i915_gem_set_wedged(struct drm_i915_private *dev_priv);
-void i915_gem_clflush_object(struct drm_i915_gem_object *obj, bool force);
+
+void i915_gem_init_mmio(struct drm_i915_private *i915);
 int __must_check i915_gem_init(struct drm_i915_private *dev_priv);
 int __must_check i915_gem_init_hw(struct drm_i915_private *dev_priv);
 void i915_gem_init_swizzling(struct drm_i915_private *dev_priv);
 void i915_gem_cleanup_engines(struct drm_i915_private *dev_priv);
-int __must_check i915_gem_wait_for_idle(struct drm_i915_private *dev_priv,
-					unsigned int flags);
+int i915_gem_wait_for_idle(struct drm_i915_private *dev_priv,
+			   unsigned int flags);
 int __must_check i915_gem_suspend(struct drm_i915_private *dev_priv);
 void i915_gem_resume(struct drm_i915_private *dev_priv);
 int i915_gem_fault(struct vm_fault *vmf);
@@ -3543,7 +3613,7 @@ static inline void intel_display_crc_init(struct drm_i915_private *dev_priv) {}
 __printf(2, 3)
 void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...);
 int i915_error_state_to_str(struct drm_i915_error_state_buf *estr,
-			    const struct i915_error_state_file_priv *error);
+			    const struct i915_gpu_state *gpu);
 int i915_error_state_buf_init(struct drm_i915_error_state_buf *eb,
 			      struct drm_i915_private *i915,
 			      size_t count, loff_t pos);
@@ -3552,13 +3622,28 @@ static inline void i915_error_state_buf_release(
 {
 	kfree(eb->buf);
 }
+
+struct i915_gpu_state *i915_capture_gpu_state(struct drm_i915_private *i915);
 void i915_capture_error_state(struct drm_i915_private *dev_priv,
 			      u32 engine_mask,
 			      const char *error_msg);
-void i915_error_state_get(struct drm_device *dev,
-			  struct i915_error_state_file_priv *error_priv);
-void i915_error_state_put(struct i915_error_state_file_priv *error_priv);
-void i915_destroy_error_state(struct drm_i915_private *dev_priv);
+
+static inline struct i915_gpu_state *
+i915_gpu_state_get(struct i915_gpu_state *gpu)
+{
+	kref_get(&gpu->ref);
+	return gpu;
+}
+
+void __i915_gpu_state_free(struct kref *kref);
+static inline void i915_gpu_state_put(struct i915_gpu_state *gpu)
+{
+	if (gpu)
+		kref_put(&gpu->ref, __i915_gpu_state_free);
+}
+
+struct i915_gpu_state *i915_first_error_state(struct drm_i915_private *i915);
+void i915_reset_error_state(struct drm_i915_private *i915);
 
 #else
 
@@ -3568,7 +3653,13 @@ static inline void i915_capture_error_state(struct drm_i915_private *dev_priv,
 {
 }
 
-static inline void i915_destroy_error_state(struct drm_i915_private *dev_priv)
+static inline struct i915_gpu_state *
+i915_first_error_state(struct drm_i915_private *i915)
+{
+	return NULL;
+}
+
+static inline void i915_reset_error_state(struct drm_i915_private *i915)
 {
 }
 
@@ -3708,7 +3799,7 @@ extern void i915_redisable_vga(struct drm_i915_private *dev_priv);
 extern void i915_redisable_vga_power_on(struct drm_i915_private *dev_priv);
 extern bool ironlake_set_drps(struct drm_i915_private *dev_priv, u8 val);
 extern void intel_init_pch_refclk(struct drm_i915_private *dev_priv);
-extern void intel_set_rps(struct drm_i915_private *dev_priv, u8 val);
+extern int intel_set_rps(struct drm_i915_private *dev_priv, u8 val);
 extern bool intel_set_memory_cxsr(struct drm_i915_private *dev_priv,
 				  bool enable);
 
@@ -3724,7 +3815,6 @@ extern void intel_overlay_print_error_state(struct drm_i915_error_state_buf *e,
 extern struct intel_display_error_state *
 intel_display_capture_error_state(struct drm_i915_private *dev_priv);
 extern void intel_display_print_error_state(struct drm_i915_error_state_buf *e,
-					    struct drm_i915_private *dev_priv,
 					    struct intel_display_error_state *error);
 
 int sandybridge_pcode_read(struct drm_i915_private *dev_priv, u32 mbox, u32 *val);
@@ -3734,7 +3824,7 @@ int skl_pcode_request(struct drm_i915_private *dev_priv, u32 mbox, u32 request,
 
 /* intel_sideband.c */
 u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr);
-void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val);
+int vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val);
 u32 vlv_nc_read(struct drm_i915_private *dev_priv, u8 addr);
 u32 vlv_iosf_sb_read(struct drm_i915_private *dev_priv, u8 port, u32 reg);
 void vlv_iosf_sb_write(struct drm_i915_private *dev_priv, u8 port, u32 reg, u32 val);
@@ -3953,14 +4043,34 @@ wait_remaining_ms_from_jiffies(unsigned long timestamp_jiffies, int to_wait_ms)
 }
 
 static inline bool
-__i915_request_irq_complete(struct drm_i915_gem_request *req)
+__i915_request_irq_complete(const struct drm_i915_gem_request *req)
 {
 	struct intel_engine_cs *engine = req->engine;
+	u32 seqno;
+
+	/* Note that the engine may have wrapped around the seqno, and
+	 * so our request->global_seqno will be ahead of the hardware,
+	 * even though it completed the request before wrapping. We catch
+	 * this by kicking all the waiters before resetting the seqno
+	 * in hardware, and also signal the fence.
+	 */
+	if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &req->fence.flags))
+		return true;
+
+	/* The request was dequeued before we were awoken. We check after
+	 * inspecting the hw to confirm that this was the same request
+	 * that generated the HWS update. The memory barriers within
+	 * the request execution are sufficient to ensure that a check
+	 * after reading the value from hw matches this request.
+	 */
+	seqno = i915_gem_request_global_seqno(req);
+	if (!seqno)
+		return false;
 
 	/* Before we do the heavier coherent read of the seqno,
 	 * check the value (hopefully) in the CPU cacheline.
 	 */
-	if (__i915_gem_request_completed(req))
+	if (__i915_gem_request_completed(req, seqno))
 		return true;
 
 	/* Ensure our read of the seqno is coherent so that we
@@ -3975,9 +4085,9 @@ __i915_request_irq_complete(struct drm_i915_gem_request *req)
 	 * is woken.
 	 */
 	if (engine->irq_seqno_barrier &&
-	    rcu_access_pointer(engine->breadcrumbs.irq_seqno_bh) == current &&
-	    cmpxchg_relaxed(&engine->breadcrumbs.irq_posted, 1, 0)) {
-		struct task_struct *tsk;
+	    test_and_clear_bit(ENGINE_IRQ_BREADCRUMB, &engine->irq_posted)) {
+		struct intel_breadcrumbs *b = &engine->breadcrumbs;
+		unsigned long flags;
 
 		/* The ordering of irq_posted versus applying the barrier
 		 * is crucial. The clearing of the current irq_posted must
@@ -3999,19 +4109,18 @@ __i915_request_irq_complete(struct drm_i915_gem_request *req)
 		 * the seqno before we believe it coherent since they see
 		 * irq_posted == false but we are still running).
 		 */
-		rcu_read_lock();
-		tsk = rcu_dereference(engine->breadcrumbs.irq_seqno_bh);
-		if (tsk && tsk != current)
+		spin_lock_irqsave(&b->irq_lock, flags);
+		if (b->irq_wait && b->irq_wait->tsk != current)
 			/* Note that if the bottom-half is changed as we
 			 * are sending the wake-up, the new bottom-half will
 			 * be woken by whomever made the change. We only have
 			 * to worry about when we steal the irq-posted for
 			 * ourself.
 			 */
-			wake_up_process(tsk);
-		rcu_read_unlock();
+			wake_up_process(b->irq_wait->tsk);
+		spin_unlock_irqrestore(&b->irq_lock, flags);
 
-		if (__i915_gem_request_completed(req))
+		if (__i915_gem_request_completed(req, seqno))
 			return true;
 	}
 
@@ -4042,4 +4151,10 @@ int remap_io_mapping(struct vm_area_struct *vma,
 		     unsigned long addr, unsigned long pfn, unsigned long size,
 		     struct io_mapping *iomap);
 
+static inline bool i915_gem_object_is_coherent(struct drm_i915_gem_object *obj)
+{
+	return (obj->cache_level != I915_CACHE_NONE ||
+		HAS_LLC(to_i915(obj->base.dev)));
+}
+
 #endif
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 6908123162d1..7c20601fe1de 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -29,12 +29,14 @@
 #include <drm/drm_vma_manager.h>
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
+#include "i915_gem_clflush.h"
 #include "i915_vgpu.h"
 #include "i915_trace.h"
 #include "intel_drv.h"
 #include "intel_frontbuffer.h"
 #include "intel_mocs.h"
 #include <linux/dma-fence-array.h>
+#include <linux/kthread.h>
 #include <linux/reservation.h>
 #include <linux/shmem_fs.h>
 #include <linux/slab.h>
@@ -47,18 +49,12 @@ static void i915_gem_flush_free_objects(struct drm_i915_private *i915);
 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 bool cpu_cache_is_coherent(struct drm_device *dev,
-				  enum i915_cache_level level)
-{
-	return HAS_LLC(to_i915(dev)) || level != I915_CACHE_NONE;
-}
-
 static bool cpu_write_needs_clflush(struct drm_i915_gem_object *obj)
 {
 	if (obj->base.write_domain == I915_GEM_DOMAIN_CPU)
 		return false;
 
-	if (!cpu_cache_is_coherent(obj->base.dev, obj->cache_level))
+	if (!i915_gem_object_is_coherent(obj))
 		return true;
 
 	return obj->pin_display;
@@ -254,7 +250,7 @@ __i915_gem_object_release_shmem(struct drm_i915_gem_object *obj,
 
 	if (needs_clflush &&
 	    (obj->base.read_domains & I915_GEM_DOMAIN_CPU) == 0 &&
-	    !cpu_cache_is_coherent(obj->base.dev, obj->cache_level))
+	    !i915_gem_object_is_coherent(obj))
 		drm_clflush_sg(pages);
 
 	obj->base.read_domains = I915_GEM_DOMAIN_CPU;
@@ -312,6 +308,8 @@ static const struct drm_i915_gem_object_ops i915_gem_phys_ops = {
 	.release = i915_gem_object_release_phys,
 };
 
+static const struct drm_i915_gem_object_ops i915_gem_object_ops;
+
 int i915_gem_object_unbind(struct drm_i915_gem_object *obj)
 {
 	struct i915_vma *vma;
@@ -399,7 +397,7 @@ out:
 	if (flags & I915_WAIT_LOCKED && i915_gem_request_completed(rq))
 		i915_gem_request_retire_upto(rq);
 
-	if (rps && rq->global_seqno == intel_engine_last_submit(rq->engine)) {
+	if (rps && i915_gem_request_global_seqno(rq) == intel_engine_last_submit(rq->engine)) {
 		/* The GPU is now idle and this client has stalled.
 		 * Since no other client has submitted a request in the
 		 * meantime, assume that this client is the only one
@@ -424,7 +422,9 @@ i915_gem_object_wait_reservation(struct reservation_object *resv,
 				 long timeout,
 				 struct intel_rps_client *rps)
 {
+	unsigned int seq = __read_seqcount_begin(&resv->seq);
 	struct dma_fence *excl;
+	bool prune_fences = false;
 
 	if (flags & I915_WAIT_ALL) {
 		struct dma_fence **shared;
@@ -449,15 +449,26 @@ i915_gem_object_wait_reservation(struct reservation_object *resv,
 		for (; i < count; i++)
 			dma_fence_put(shared[i]);
 		kfree(shared);
+
+		prune_fences = count && timeout >= 0;
 	} else {
 		excl = reservation_object_get_excl_rcu(resv);
 	}
 
-	if (excl && timeout >= 0)
+	if (excl && timeout >= 0) {
 		timeout = i915_gem_object_wait_fence(excl, flags, timeout, rps);
+		prune_fences = timeout >= 0;
+	}
 
 	dma_fence_put(excl);
 
+	if (prune_fences && !__read_seqcount_retry(&resv->seq, seq)) {
+		reservation_object_lock(resv, NULL);
+		if (!__read_seqcount_retry(&resv->seq, seq))
+			reservation_object_add_excl_fence(resv, NULL);
+		reservation_object_unlock(resv);
+	}
+
 	return timeout;
 }
 
@@ -585,9 +596,18 @@ i915_gem_object_attach_phys(struct drm_i915_gem_object *obj,
 	if (obj->mm.pages)
 		return -EBUSY;
 
+	GEM_BUG_ON(obj->ops != &i915_gem_object_ops);
 	obj->ops = &i915_gem_phys_ops;
 
-	return i915_gem_object_pin_pages(obj);
+	ret = i915_gem_object_pin_pages(obj);
+	if (ret)
+		goto err_xfer;
+
+	return 0;
+
+err_xfer:
+	obj->ops = &i915_gem_object_ops;
+	return ret;
 }
 
 static int
@@ -608,7 +628,7 @@ i915_gem_phys_pwrite(struct drm_i915_gem_object *obj,
 	drm_clflush_virt_range(vaddr, args->size);
 	i915_gem_chipset_flush(to_i915(obj->base.dev));
 
-	intel_fb_obj_flush(obj, false, ORIGIN_CPU);
+	intel_fb_obj_flush(obj, ORIGIN_CPU);
 	return 0;
 }
 
@@ -771,8 +791,7 @@ int i915_gem_obj_prepare_shmem_read(struct drm_i915_gem_object *obj,
 	 * anyway again before the next pread happens.
 	 */
 	if (!(obj->base.read_domains & I915_GEM_DOMAIN_CPU))
-		*needs_clflush = !cpu_cache_is_coherent(obj->base.dev,
-							obj->cache_level);
+		*needs_clflush = !i915_gem_object_is_coherent(obj);
 
 	if (*needs_clflush && !static_cpu_has(X86_FEATURE_CLFLUSH)) {
 		ret = i915_gem_object_set_to_cpu_domain(obj, false);
@@ -828,8 +847,7 @@ int i915_gem_obj_prepare_shmem_write(struct drm_i915_gem_object *obj,
 	 * before writing.
 	 */
 	if (!(obj->base.read_domains & I915_GEM_DOMAIN_CPU))
-		*needs_clflush |= !cpu_cache_is_coherent(obj->base.dev,
-							 obj->cache_level);
+		*needs_clflush |= !i915_gem_object_is_coherent(obj);
 
 	if (*needs_clflush && !static_cpu_has(X86_FEATURE_CLFLUSH)) {
 		ret = i915_gem_object_set_to_cpu_domain(obj, true);
@@ -1257,7 +1275,7 @@ i915_gem_gtt_pwrite_fast(struct drm_i915_gem_object *obj,
 		user_data += page_length;
 		offset += page_length;
 	}
-	intel_fb_obj_flush(obj, false, ORIGIN_CPU);
+	intel_fb_obj_flush(obj, ORIGIN_CPU);
 
 	mutex_lock(&i915->drm.struct_mutex);
 out_unpin:
@@ -1393,7 +1411,7 @@ i915_gem_shmem_pwrite(struct drm_i915_gem_object *obj,
 		offset = 0;
 	}
 
-	intel_fb_obj_flush(obj, false, ORIGIN_CPU);
+	intel_fb_obj_flush(obj, ORIGIN_CPU);
 	i915_gem_obj_finish_shmem_access(obj);
 	return ret;
 }
@@ -1596,23 +1614,16 @@ i915_gem_sw_finish_ioctl(struct drm_device *dev, void *data,
 {
 	struct drm_i915_gem_sw_finish *args = data;
 	struct drm_i915_gem_object *obj;
-	int err = 0;
 
 	obj = i915_gem_object_lookup(file, args->handle);
 	if (!obj)
 		return -ENOENT;
 
 	/* Pinned buffers may be scanout, so flush the cache */
-	if (READ_ONCE(obj->pin_display)) {
-		err = i915_mutex_lock_interruptible(dev);
-		if (!err) {
-			i915_gem_object_flush_cpu_write_domain(obj);
-			mutex_unlock(&dev->struct_mutex);
-		}
-	}
-
+	i915_gem_object_flush_if_display(obj);
 	i915_gem_object_put(obj);
-	return err;
+
+	return 0;
 }
 
 /**
@@ -2223,17 +2234,17 @@ unlock:
 	mutex_unlock(&obj->mm.lock);
 }
 
-static void i915_sg_trim(struct sg_table *orig_st)
+static bool i915_sg_trim(struct sg_table *orig_st)
 {
 	struct sg_table new_st;
 	struct scatterlist *sg, *new_sg;
 	unsigned int i;
 
 	if (orig_st->nents == orig_st->orig_nents)
-		return;
+		return false;
 
 	if (sg_alloc_table(&new_st, orig_st->nents, GFP_KERNEL | __GFP_NOWARN))
-		return;
+		return false;
 
 	new_sg = new_st.sgl;
 	for_each_sg(orig_st->sgl, sg, orig_st->nents, i) {
@@ -2246,6 +2257,7 @@ static void i915_sg_trim(struct sg_table *orig_st)
 	sg_free_table(orig_st);
 
 	*orig_st = new_st;
+	return true;
 }
 
 static struct sg_table *
@@ -2596,7 +2608,8 @@ static void i915_gem_context_mark_innocent(struct i915_gem_context *ctx)
 struct drm_i915_gem_request *
 i915_gem_find_active_request(struct intel_engine_cs *engine)
 {
-	struct drm_i915_gem_request *request;
+	struct drm_i915_gem_request *request, *active = NULL;
+	unsigned long flags;
 
 	/* We are called by the error capture and reset at a random
 	 * point in time. In particular, note that neither is crucially
@@ -2606,15 +2619,22 @@ i915_gem_find_active_request(struct intel_engine_cs *engine)
 	 * extra delay for a recent interrupt is pointless. Hence, we do
 	 * not need an engine->irq_seqno_barrier() before the seqno reads.
 	 */
+	spin_lock_irqsave(&engine->timeline->lock, flags);
 	list_for_each_entry(request, &engine->timeline->requests, link) {
-		if (__i915_gem_request_completed(request))
+		if (__i915_gem_request_completed(request,
+						 request->global_seqno))
 			continue;
 
 		GEM_BUG_ON(request->engine != engine);
-		return request;
+		GEM_BUG_ON(test_bit(DMA_FENCE_FLAG_SIGNALED_BIT,
+				    &request->fence.flags));
+
+		active = request;
+		break;
 	}
+	spin_unlock_irqrestore(&engine->timeline->lock, flags);
 
-	return NULL;
+	return active;
 }
 
 static bool engine_stalled(struct intel_engine_cs *engine)
@@ -2641,7 +2661,30 @@ int i915_gem_reset_prepare(struct drm_i915_private *dev_priv)
 	for_each_engine(engine, dev_priv, id) {
 		struct drm_i915_gem_request *request;
 
+		/* Prevent the signaler thread from updating the request
+		 * state (by calling dma_fence_signal) as we are processing
+		 * the reset. The write from the GPU of the seqno is
+		 * asynchronous and the signaler thread may see a different
+		 * value to us and declare the request complete, even though
+		 * the reset routine have picked that request as the active
+		 * (incomplete) request. This conflict is not handled
+		 * gracefully!
+		 */
+		kthread_park(engine->breadcrumbs.signaler);
+
+		/* Prevent request submission to the hardware until we have
+		 * completed the reset in i915_gem_reset_finish(). If a request
+		 * is completed by one engine, it may then queue a request
+		 * to a second via its engine->irq_tasklet *just* as we are
+		 * calling engine->init_hw() and also writing the ELSP.
+		 * Turning off the engine->irq_tasklet until the reset is over
+		 * prevents the race.
+		 */
 		tasklet_kill(&engine->irq_tasklet);
+		tasklet_disable(&engine->irq_tasklet);
+
+		if (engine->irq_seqno_barrier)
+			engine->irq_seqno_barrier(engine);
 
 		if (engine_stalled(engine)) {
 			request = i915_gem_find_active_request(engine);
@@ -2739,9 +2782,6 @@ static void i915_gem_reset_engine(struct intel_engine_cs *engine)
 {
 	struct drm_i915_gem_request *request;
 
-	if (engine->irq_seqno_barrier)
-		engine->irq_seqno_barrier(engine);
-
 	request = i915_gem_find_active_request(engine);
 	if (request && i915_gem_reset_request(request)) {
 		DRM_DEBUG_DRIVER("resetting %s to restart from tail of request 0x%x\n",
@@ -2756,7 +2796,7 @@ static void i915_gem_reset_engine(struct intel_engine_cs *engine)
 	engine->reset_hw(engine, request);
 }
 
-void i915_gem_reset_finish(struct drm_i915_private *dev_priv)
+void i915_gem_reset(struct drm_i915_private *dev_priv)
 {
 	struct intel_engine_cs *engine;
 	enum intel_engine_id id;
@@ -2765,8 +2805,14 @@ void i915_gem_reset_finish(struct drm_i915_private *dev_priv)
 
 	i915_gem_retire_requests(dev_priv);
 
-	for_each_engine(engine, dev_priv, id)
+	for_each_engine(engine, dev_priv, id) {
+		struct i915_gem_context *ctx;
+
 		i915_gem_reset_engine(engine);
+		ctx = fetch_and_zero(&engine->last_retired_context);
+		if (ctx)
+			engine->context_unpin(engine, ctx);
+	}
 
 	i915_gem_restore_fences(dev_priv);
 
@@ -2778,6 +2824,19 @@ void i915_gem_reset_finish(struct drm_i915_private *dev_priv)
 	}
 }
 
+void i915_gem_reset_finish(struct drm_i915_private *dev_priv)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	lockdep_assert_held(&dev_priv->drm.struct_mutex);
+
+	for_each_engine(engine, dev_priv, id) {
+		tasklet_enable(&engine->irq_tasklet);
+		kthread_unpark(engine->breadcrumbs.signaler);
+	}
+}
+
 static void nop_submit_request(struct drm_i915_gem_request *request)
 {
 	dma_fence_set_error(&request->fence, -EIO);
@@ -2900,8 +2959,8 @@ i915_gem_idle_work_handler(struct work_struct *work)
 	 * new request is submitted.
 	 */
 	wait_for(READ_ONCE(dev_priv->gt.active_requests) ||
-		 intel_execlists_idle(dev_priv), 10);
-
+		 intel_engines_are_idle(dev_priv),
+		 10);
 	if (READ_ONCE(dev_priv->gt.active_requests))
 		return;
 
@@ -2926,11 +2985,13 @@ i915_gem_idle_work_handler(struct work_struct *work)
 	if (dev_priv->gt.active_requests)
 		goto out_unlock;
 
-	if (wait_for(intel_execlists_idle(dev_priv), 10))
+	if (wait_for(intel_engines_are_idle(dev_priv), 10))
 		DRM_ERROR("Timeout waiting for engines to idle\n");
 
-	for_each_engine(engine, dev_priv, id)
+	for_each_engine(engine, dev_priv, id) {
+		intel_engine_disarm_breadcrumbs(engine);
 		i915_gem_batch_pool_fini(&engine->batch_pool);
+	}
 
 	GEM_BUG_ON(!dev_priv->gt.awake);
 	dev_priv->gt.awake = false;
@@ -3029,6 +3090,16 @@ i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 		args->timeout_ns -= ktime_to_ns(ktime_sub(ktime_get(), start));
 		if (args->timeout_ns < 0)
 			args->timeout_ns = 0;
+
+		/*
+		 * Apparently ktime isn't accurate enough and occasionally has a
+		 * bit of mismatch in the jiffies<->nsecs<->ktime loop. So patch
+		 * things up to make the test happy. We allow up to 1 jiffy.
+		 *
+		 * This is a regression from the timespec->ktime conversion.
+		 */
+		if (ret == -ETIME && !nsecs_to_jiffies(args->timeout_ns))
+			args->timeout_ns = 0;
 	}
 
 	i915_gem_object_put(obj);
@@ -3071,41 +3142,6 @@ int i915_gem_wait_for_idle(struct drm_i915_private *i915, unsigned int flags)
 	return 0;
 }
 
-void i915_gem_clflush_object(struct drm_i915_gem_object *obj,
-			     bool force)
-{
-	/* If we don't have a page list set up, then we're not pinned
-	 * to GPU, and we can ignore the cache flush because it'll happen
-	 * again at bind time.
-	 */
-	if (!obj->mm.pages)
-		return;
-
-	/*
-	 * Stolen memory is always coherent with the GPU as it is explicitly
-	 * marked as wc by the system, or the system is cache-coherent.
-	 */
-	if (obj->stolen || obj->phys_handle)
-		return;
-
-	/* If the GPU is snooping the contents of the CPU cache,
-	 * we do not need to manually clear the CPU cache lines.  However,
-	 * the caches are only snooped when the render cache is
-	 * flushed/invalidated.  As we always have to emit invalidations
-	 * and flushes when moving into and out of the RENDER domain, correct
-	 * snooping behaviour occurs naturally as the result of our domain
-	 * tracking.
-	 */
-	if (!force && cpu_cache_is_coherent(obj->base.dev, obj->cache_level)) {
-		obj->cache_dirty = true;
-		return;
-	}
-
-	trace_i915_gem_object_clflush(obj);
-	drm_clflush_sg(obj->mm.pages);
-	obj->cache_dirty = false;
-}
-
 /** Flushes the GTT write domain for the object if it's dirty. */
 static void
 i915_gem_object_flush_gtt_write_domain(struct drm_i915_gem_object *obj)
@@ -3134,12 +3170,9 @@ i915_gem_object_flush_gtt_write_domain(struct drm_i915_gem_object *obj)
 	if (INTEL_GEN(dev_priv) >= 6 && !HAS_LLC(dev_priv))
 		POSTING_READ(RING_ACTHD(dev_priv->engine[RCS]->mmio_base));
 
-	intel_fb_obj_flush(obj, false, write_origin(obj, I915_GEM_DOMAIN_GTT));
+	intel_fb_obj_flush(obj, write_origin(obj, I915_GEM_DOMAIN_GTT));
 
 	obj->base.write_domain = 0;
-	trace_i915_gem_object_change_domain(obj,
-					    obj->base.read_domains,
-					    I915_GEM_DOMAIN_GTT);
 }
 
 /** Flushes the CPU write domain for the object if it's dirty. */
@@ -3149,13 +3182,27 @@ i915_gem_object_flush_cpu_write_domain(struct drm_i915_gem_object *obj)
 	if (obj->base.write_domain != I915_GEM_DOMAIN_CPU)
 		return;
 
-	i915_gem_clflush_object(obj, obj->pin_display);
-	intel_fb_obj_flush(obj, false, ORIGIN_CPU);
+	i915_gem_clflush_object(obj, I915_CLFLUSH_SYNC);
+	obj->base.write_domain = 0;
+}
+
+static void __i915_gem_object_flush_for_display(struct drm_i915_gem_object *obj)
+{
+	if (obj->base.write_domain != I915_GEM_DOMAIN_CPU && !obj->cache_dirty)
+		return;
 
+	i915_gem_clflush_object(obj, I915_CLFLUSH_FORCE);
 	obj->base.write_domain = 0;
-	trace_i915_gem_object_change_domain(obj,
-					    obj->base.read_domains,
-					    I915_GEM_DOMAIN_CPU);
+}
+
+void i915_gem_object_flush_if_display(struct drm_i915_gem_object *obj)
+{
+	if (!READ_ONCE(obj->pin_display))
+		return;
+
+	mutex_lock(&obj->base.dev->struct_mutex);
+	__i915_gem_object_flush_for_display(obj);
+	mutex_unlock(&obj->base.dev->struct_mutex);
 }
 
 /**
@@ -3169,7 +3216,6 @@ i915_gem_object_flush_cpu_write_domain(struct drm_i915_gem_object *obj)
 int
 i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
 {
-	uint32_t old_write_domain, old_read_domains;
 	int ret;
 
 	lockdep_assert_held(&obj->base.dev->struct_mutex);
@@ -3207,9 +3253,6 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
 	if ((obj->base.read_domains & I915_GEM_DOMAIN_GTT) == 0)
 		mb();
 
-	old_write_domain = obj->base.write_domain;
-	old_read_domains = obj->base.read_domains;
-
 	/* It should now be out of any other write domains, and we can update
 	 * the domain values for our changes.
 	 */
@@ -3221,10 +3264,6 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
 		obj->mm.dirty = true;
 	}
 
-	trace_i915_gem_object_change_domain(obj,
-					    old_read_domains,
-					    old_write_domain);
-
 	i915_gem_object_unpin_pages(obj);
 	return 0;
 }
@@ -3349,7 +3388,7 @@ restart:
 	}
 
 	if (obj->base.write_domain == I915_GEM_DOMAIN_CPU &&
-	    cpu_cache_is_coherent(obj->base.dev, obj->cache_level))
+	    i915_gem_object_is_coherent(obj))
 		obj->cache_dirty = true;
 
 	list_for_each_entry(vma, &obj->vma_list, obj_link)
@@ -3461,7 +3500,6 @@ i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj,
 				     const struct i915_ggtt_view *view)
 {
 	struct i915_vma *vma;
-	u32 old_read_domains, old_write_domain;
 	int ret;
 
 	lockdep_assert_held(&obj->base.dev->struct_mutex);
@@ -3521,24 +3559,14 @@ i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj,
 	vma->display_alignment = max_t(u64, vma->display_alignment, alignment);
 
 	/* Treat this as an end-of-frame, like intel_user_framebuffer_dirty() */
-	if (obj->cache_dirty || obj->base.write_domain == I915_GEM_DOMAIN_CPU) {
-		i915_gem_clflush_object(obj, true);
-		intel_fb_obj_flush(obj, false, ORIGIN_DIRTYFB);
-	}
-
-	old_write_domain = obj->base.write_domain;
-	old_read_domains = obj->base.read_domains;
+	__i915_gem_object_flush_for_display(obj);
+	intel_fb_obj_flush(obj, ORIGIN_DIRTYFB);
 
 	/* It should now be out of any other write domains, and we can update
 	 * the domain values for our changes.
 	 */
-	obj->base.write_domain = 0;
 	obj->base.read_domains |= I915_GEM_DOMAIN_GTT;
 
-	trace_i915_gem_object_change_domain(obj,
-					    old_read_domains,
-					    old_write_domain);
-
 	return vma;
 
 err_unpin_display:
@@ -3574,7 +3602,6 @@ i915_gem_object_unpin_from_display_plane(struct i915_vma *vma)
 int
 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;
 
 	lockdep_assert_held(&obj->base.dev->struct_mutex);
@@ -3593,13 +3620,9 @@ i915_gem_object_set_to_cpu_domain(struct drm_i915_gem_object *obj, bool write)
 
 	i915_gem_object_flush_gtt_write_domain(obj);
 
-	old_write_domain = obj->base.write_domain;
-	old_read_domains = obj->base.read_domains;
-
 	/* Flush the CPU cache if it's still invalid. */
 	if ((obj->base.read_domains & I915_GEM_DOMAIN_CPU) == 0) {
-		i915_gem_clflush_object(obj, false);
-
+		i915_gem_clflush_object(obj, I915_CLFLUSH_SYNC);
 		obj->base.read_domains |= I915_GEM_DOMAIN_CPU;
 	}
 
@@ -3616,10 +3639,6 @@ i915_gem_object_set_to_cpu_domain(struct drm_i915_gem_object *obj, bool write)
 		obj->base.write_domain = I915_GEM_DOMAIN_CPU;
 	}
 
-	trace_i915_gem_object_change_domain(obj,
-					    old_read_domains,
-					    old_write_domain);
-
 	return 0;
 }
 
@@ -3647,16 +3666,14 @@ i915_gem_ring_throttle(struct drm_device *dev, struct drm_file *file)
 		return -EIO;
 
 	spin_lock(&file_priv->mm.lock);
-	list_for_each_entry(request, &file_priv->mm.request_list, client_list) {
+	list_for_each_entry(request, &file_priv->mm.request_list, client_link) {
 		if (time_after_eq(request->emitted_jiffies, recent_enough))
 			break;
 
-		/*
-		 * Note that the request might not have been submitted yet.
-		 * In which case emitted_jiffies will be zero.
-		 */
-		if (!request->emitted_jiffies)
-			continue;
+		if (target) {
+			list_del(&target->client_link);
+			target->file_priv = NULL;
+		}
 
 		target = request;
 	}
@@ -3942,7 +3959,7 @@ frontbuffer_retire(struct i915_gem_active *active,
 	struct drm_i915_gem_object *obj =
 		container_of(active, typeof(*obj), frontbuffer_write);
 
-	intel_fb_obj_flush(obj, true, ORIGIN_CS);
+	intel_fb_obj_flush(obj, ORIGIN_CS);
 }
 
 void i915_gem_object_init(struct drm_i915_gem_object *obj,
@@ -4203,11 +4220,29 @@ static void assert_kernel_context_is_current(struct drm_i915_private *dev_priv)
 			   !i915_gem_context_is_kernel(engine->last_retired_context));
 }
 
+void i915_gem_sanitize(struct drm_i915_private *i915)
+{
+	/*
+	 * If we inherit context state from the BIOS or earlier occupants
+	 * of the GPU, the GPU may be in an inconsistent state when we
+	 * try to take over. The only way to remove the earlier state
+	 * is by resetting. However, resetting on earlier gen is tricky as
+	 * it may impact the display and we are uncertain about the stability
+	 * of the reset, so we only reset recent machines with logical
+	 * context support (that must be reset to remove any stray contexts).
+	 */
+	if (HAS_HW_CONTEXTS(i915)) {
+		int reset = intel_gpu_reset(i915, ALL_ENGINES);
+		WARN_ON(reset && reset != -ENODEV);
+	}
+}
+
 int i915_gem_suspend(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = &dev_priv->drm;
 	int ret;
 
+	intel_runtime_pm_get(dev_priv);
 	intel_suspend_gt_powersave(dev_priv);
 
 	mutex_lock(&dev->struct_mutex);
@@ -4222,13 +4257,13 @@ int i915_gem_suspend(struct drm_i915_private *dev_priv)
 	 */
 	ret = i915_gem_switch_to_kernel_context(dev_priv);
 	if (ret)
-		goto err;
+		goto err_unlock;
 
 	ret = i915_gem_wait_for_idle(dev_priv,
 				     I915_WAIT_INTERRUPTIBLE |
 				     I915_WAIT_LOCKED);
 	if (ret)
-		goto err;
+		goto err_unlock;
 
 	i915_gem_retire_requests(dev_priv);
 	GEM_BUG_ON(dev_priv->gt.active_requests);
@@ -4252,7 +4287,7 @@ int i915_gem_suspend(struct drm_i915_private *dev_priv)
 	 * reset the GPU back to its idle, low power state.
 	 */
 	WARN_ON(dev_priv->gt.awake);
-	WARN_ON(!intel_execlists_idle(dev_priv));
+	WARN_ON(!intel_engines_are_idle(dev_priv));
 
 	/*
 	 * Neither the BIOS, ourselves or any other kernel
@@ -4273,15 +4308,13 @@ int i915_gem_suspend(struct drm_i915_private *dev_priv)
 	 * machines is a good idea, we don't - just in case it leaves the
 	 * machine in an unusable condition.
 	 */
-	if (HAS_HW_CONTEXTS(dev_priv)) {
-		int reset = intel_gpu_reset(dev_priv, ALL_ENGINES);
-		WARN_ON(reset && reset != -ENODEV);
-	}
-
-	return 0;
+	i915_gem_sanitize(dev_priv);
+	goto out_rpm_put;
 
-err:
+err_unlock:
 	mutex_unlock(&dev->struct_mutex);
+out_rpm_put:
+	intel_runtime_pm_put(dev_priv);
 	return ret;
 }
 
@@ -4351,11 +4384,24 @@ static void init_unused_rings(struct drm_i915_private *dev_priv)
 	}
 }
 
-int
-i915_gem_init_hw(struct drm_i915_private *dev_priv)
+static int __i915_gem_restart_engines(void *data)
 {
+	struct drm_i915_private *i915 = data;
 	struct intel_engine_cs *engine;
 	enum intel_engine_id id;
+	int err;
+
+	for_each_engine(engine, i915, id) {
+		err = engine->init_hw(engine);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+int i915_gem_init_hw(struct drm_i915_private *dev_priv)
+{
 	int ret;
 
 	dev_priv->gt.last_init_time = ktime_get();
@@ -4401,11 +4447,9 @@ i915_gem_init_hw(struct drm_i915_private *dev_priv)
 	}
 
 	/* Need to do basic initialisation of all rings first: */
-	for_each_engine(engine, dev_priv, id) {
-		ret = engine->init_hw(engine);
-		if (ret)
-			goto out;
-	}
+	ret = __i915_gem_restart_engines(dev_priv);
+	if (ret)
+		goto out;
 
 	intel_mocs_init_l3cc_table(dev_priv);
 
@@ -4446,6 +4490,8 @@ int i915_gem_init(struct drm_i915_private *dev_priv)
 
 	mutex_lock(&dev_priv->drm.struct_mutex);
 
+	i915_gem_clflush_init(dev_priv);
+
 	if (!i915.enable_execlists) {
 		dev_priv->gt.resume = intel_legacy_submission_resume;
 		dev_priv->gt.cleanup_engine = intel_engine_cleanup;
@@ -4494,6 +4540,11 @@ out_unlock:
 	return ret;
 }
 
+void i915_gem_init_mmio(struct drm_i915_private *i915)
+{
+	i915_gem_sanitize(i915);
+}
+
 void
 i915_gem_cleanup_engines(struct drm_i915_private *dev_priv)
 {
@@ -4583,8 +4634,6 @@ i915_gem_load_init(struct drm_i915_private *dev_priv)
 	init_waitqueue_head(&dev_priv->gpu_error.wait_queue);
 	init_waitqueue_head(&dev_priv->gpu_error.reset_queue);
 
-	dev_priv->relative_constants_mode = I915_EXEC_CONSTANTS_REL_GENERAL;
-
 	init_waitqueue_head(&dev_priv->pending_flip_queue);
 
 	dev_priv->mm.interruptible = true;
@@ -4609,7 +4658,9 @@ err_out:
 
 void i915_gem_load_cleanup(struct drm_i915_private *dev_priv)
 {
+	i915_gem_drain_freed_objects(dev_priv);
 	WARN_ON(!llist_empty(&dev_priv->mm.free_list));
+	WARN_ON(dev_priv->mm.object_count);
 
 	mutex_lock(&dev_priv->drm.struct_mutex);
 	i915_gem_timeline_fini(&dev_priv->gt.global_timeline);
@@ -4627,14 +4678,10 @@ void i915_gem_load_cleanup(struct drm_i915_private *dev_priv)
 
 int i915_gem_freeze(struct drm_i915_private *dev_priv)
 {
-	intel_runtime_pm_get(dev_priv);
-
 	mutex_lock(&dev_priv->drm.struct_mutex);
 	i915_gem_shrink_all(dev_priv);
 	mutex_unlock(&dev_priv->drm.struct_mutex);
 
-	intel_runtime_pm_put(dev_priv);
-
 	return 0;
 }
 
@@ -4685,7 +4732,7 @@ void i915_gem_release(struct drm_device *dev, struct drm_file *file)
 	 * file_priv.
 	 */
 	spin_lock(&file_priv->mm.lock);
-	list_for_each_entry(request, &file_priv->mm.request_list, client_list)
+	list_for_each_entry(request, &file_priv->mm.request_list, client_link)
 		request->file_priv = NULL;
 	spin_unlock(&file_priv->mm.lock);
 
@@ -4949,3 +4996,11 @@ i915_gem_object_get_dma_address(struct drm_i915_gem_object *obj,
 	sg = i915_gem_object_get_sg(obj, n, &offset);
 	return sg_dma_address(sg) + (offset << PAGE_SHIFT);
 }
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/scatterlist.c"
+#include "selftests/mock_gem_device.c"
+#include "selftests/huge_gem_object.c"
+#include "selftests/i915_gem_object.c"
+#include "selftests/i915_gem_coherency.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_gem.h b/drivers/gpu/drm/i915/i915_gem.h
index a585d47c420a..5a49487368ca 100644
--- a/drivers/gpu/drm/i915/i915_gem.h
+++ b/drivers/gpu/drm/i915/i915_gem.h
@@ -28,9 +28,18 @@
 #ifdef CONFIG_DRM_I915_DEBUG_GEM
 #define GEM_BUG_ON(expr) BUG_ON(expr)
 #define GEM_WARN_ON(expr) WARN_ON(expr)
+
+#define GEM_DEBUG_DECL(var) var
+#define GEM_DEBUG_EXEC(expr) expr
+#define GEM_DEBUG_BUG_ON(expr) GEM_BUG_ON(expr)
+
 #else
 #define GEM_BUG_ON(expr) BUILD_BUG_ON_INVALID(expr)
 #define GEM_WARN_ON(expr) (BUILD_BUG_ON_INVALID(expr), 0)
+
+#define GEM_DEBUG_DECL(var)
+#define GEM_DEBUG_EXEC(expr) do { } while (0)
+#define GEM_DEBUG_BUG_ON(expr)
 #endif
 
 #define I915_NUM_ENGINES 5
diff --git a/drivers/gpu/drm/i915/i915_gem_batch_pool.c b/drivers/gpu/drm/i915/i915_gem_batch_pool.c
index b3bc119ec1bb..99ceae7855f8 100644
--- a/drivers/gpu/drm/i915/i915_gem_batch_pool.c
+++ b/drivers/gpu/drm/i915/i915_gem_batch_pool.c
@@ -122,9 +122,9 @@ i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool,
 
 		if (tmp->base.size >= size) {
 			/* Clear the set of shared fences early */
-			ww_mutex_lock(&tmp->resv->lock, NULL);
+			reservation_object_lock(tmp->resv, NULL);
 			reservation_object_add_excl_fence(tmp->resv, NULL);
-			ww_mutex_unlock(&tmp->resv->lock);
+			reservation_object_unlock(tmp->resv);
 
 			obj = tmp;
 			break;
diff --git a/drivers/gpu/drm/i915/i915_gem_clflush.c b/drivers/gpu/drm/i915/i915_gem_clflush.c
new file mode 100644
index 000000000000..d925fb582ba7
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem_clflush.c
@@ -0,0 +1,189 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "i915_drv.h"
+#include "intel_frontbuffer.h"
+#include "i915_gem_clflush.h"
+
+static DEFINE_SPINLOCK(clflush_lock);
+static u64 clflush_context;
+
+struct clflush {
+	struct dma_fence dma; /* Must be first for dma_fence_free() */
+	struct i915_sw_fence wait;
+	struct work_struct work;
+	struct drm_i915_gem_object *obj;
+};
+
+static const char *i915_clflush_get_driver_name(struct dma_fence *fence)
+{
+	return DRIVER_NAME;
+}
+
+static const char *i915_clflush_get_timeline_name(struct dma_fence *fence)
+{
+	return "clflush";
+}
+
+static bool i915_clflush_enable_signaling(struct dma_fence *fence)
+{
+	return true;
+}
+
+static void i915_clflush_release(struct dma_fence *fence)
+{
+	struct clflush *clflush = container_of(fence, typeof(*clflush), dma);
+
+	i915_sw_fence_fini(&clflush->wait);
+
+	BUILD_BUG_ON(offsetof(typeof(*clflush), dma));
+	dma_fence_free(&clflush->dma);
+}
+
+static const struct dma_fence_ops i915_clflush_ops = {
+	.get_driver_name = i915_clflush_get_driver_name,
+	.get_timeline_name = i915_clflush_get_timeline_name,
+	.enable_signaling = i915_clflush_enable_signaling,
+	.wait = dma_fence_default_wait,
+	.release = i915_clflush_release,
+};
+
+static void __i915_do_clflush(struct drm_i915_gem_object *obj)
+{
+	drm_clflush_sg(obj->mm.pages);
+	obj->cache_dirty = false;
+
+	intel_fb_obj_flush(obj, ORIGIN_CPU);
+}
+
+static void i915_clflush_work(struct work_struct *work)
+{
+	struct clflush *clflush = container_of(work, typeof(*clflush), work);
+	struct drm_i915_gem_object *obj = clflush->obj;
+
+	if (!obj->cache_dirty)
+		goto out;
+
+	if (i915_gem_object_pin_pages(obj)) {
+		DRM_ERROR("Failed to acquire obj->pages for clflushing\n");
+		goto out;
+	}
+
+	__i915_do_clflush(obj);
+
+	i915_gem_object_unpin_pages(obj);
+
+out:
+	i915_gem_object_put(obj);
+
+	dma_fence_signal(&clflush->dma);
+	dma_fence_put(&clflush->dma);
+}
+
+static int __i915_sw_fence_call
+i915_clflush_notify(struct i915_sw_fence *fence,
+		    enum i915_sw_fence_notify state)
+{
+	struct clflush *clflush = container_of(fence, typeof(*clflush), wait);
+
+	switch (state) {
+	case FENCE_COMPLETE:
+		schedule_work(&clflush->work);
+		break;
+
+	case FENCE_FREE:
+		dma_fence_put(&clflush->dma);
+		break;
+	}
+
+	return NOTIFY_DONE;
+}
+
+void i915_gem_clflush_object(struct drm_i915_gem_object *obj,
+			     unsigned int flags)
+{
+	struct clflush *clflush;
+
+	/*
+	 * Stolen memory is always coherent with the GPU as it is explicitly
+	 * marked as wc by the system, or the system is cache-coherent.
+	 * Similarly, we only access struct pages through the CPU cache, so
+	 * anything not backed by physical memory we consider to be always
+	 * coherent and not need clflushing.
+	 */
+	if (!i915_gem_object_has_struct_page(obj))
+		return;
+
+	obj->cache_dirty = true;
+
+	/* If the GPU is snooping the contents of the CPU cache,
+	 * we do not need to manually clear the CPU cache lines.  However,
+	 * the caches are only snooped when the render cache is
+	 * flushed/invalidated.  As we always have to emit invalidations
+	 * and flushes when moving into and out of the RENDER domain, correct
+	 * snooping behaviour occurs naturally as the result of our domain
+	 * tracking.
+	 */
+	if (!(flags & I915_CLFLUSH_FORCE) && i915_gem_object_is_coherent(obj))
+		return;
+
+	trace_i915_gem_object_clflush(obj);
+
+	clflush = NULL;
+	if (!(flags & I915_CLFLUSH_SYNC))
+		clflush = kmalloc(sizeof(*clflush), GFP_KERNEL);
+	if (clflush) {
+		dma_fence_init(&clflush->dma,
+			       &i915_clflush_ops,
+			       &clflush_lock,
+			       clflush_context,
+			       0);
+		i915_sw_fence_init(&clflush->wait, i915_clflush_notify);
+
+		clflush->obj = i915_gem_object_get(obj);
+		INIT_WORK(&clflush->work, i915_clflush_work);
+
+		dma_fence_get(&clflush->dma);
+
+		i915_sw_fence_await_reservation(&clflush->wait,
+						obj->resv, NULL,
+						false, I915_FENCE_TIMEOUT,
+						GFP_KERNEL);
+
+		reservation_object_lock(obj->resv, NULL);
+		reservation_object_add_excl_fence(obj->resv, &clflush->dma);
+		reservation_object_unlock(obj->resv);
+
+		i915_sw_fence_commit(&clflush->wait);
+	} else if (obj->mm.pages) {
+		__i915_do_clflush(obj);
+	} else {
+		GEM_BUG_ON(obj->base.write_domain != I915_GEM_DOMAIN_CPU);
+	}
+}
+
+void i915_gem_clflush_init(struct drm_i915_private *i915)
+{
+	clflush_context = dma_fence_context_alloc(1);
+}
diff --git a/drivers/gpu/drm/i915/i915_gem_clflush.h b/drivers/gpu/drm/i915/i915_gem_clflush.h
new file mode 100644
index 000000000000..b62d61a2d15f
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_gem_clflush.h
@@ -0,0 +1,37 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __I915_GEM_CLFLUSH_H__
+#define __I915_GEM_CLFLUSH_H__
+
+struct drm_i915_private;
+struct drm_i915_gem_object;
+
+void i915_gem_clflush_init(struct drm_i915_private *i915);
+void i915_gem_clflush_object(struct drm_i915_gem_object *obj,
+			     unsigned int flags);
+#define I915_CLFLUSH_FORCE BIT(0)
+#define I915_CLFLUSH_SYNC BIT(1)
+
+#endif /* __I915_GEM_CLFLUSH_H__ */
diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c
index 17f90c618208..baceca14f5e0 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/i915_gem_context.c
@@ -92,21 +92,6 @@
 
 #define ALL_L3_SLICES(dev) (1 << NUM_L3_SLICES(dev)) - 1
 
-/* This is a HW constraint. The value below is the largest known requirement
- * I've seen in a spec to date, and that was a workaround for a non-shipping
- * part. It should be safe to decrease this, but it's more future proof as is.
- */
-#define GEN6_CONTEXT_ALIGN (64<<10)
-#define GEN7_CONTEXT_ALIGN I915_GTT_MIN_ALIGNMENT
-
-static size_t get_context_alignment(struct drm_i915_private *dev_priv)
-{
-	if (IS_GEN6(dev_priv))
-		return GEN6_CONTEXT_ALIGN;
-
-	return GEN7_CONTEXT_ALIGN;
-}
-
 static int get_context_size(struct drm_i915_private *dev_priv)
 {
 	int ret;
@@ -236,6 +221,30 @@ static int assign_hw_id(struct drm_i915_private *dev_priv, unsigned *out)
 	return 0;
 }
 
+static u32 default_desc_template(const struct drm_i915_private *i915,
+				 const struct i915_hw_ppgtt *ppgtt)
+{
+	u32 address_mode;
+	u32 desc;
+
+	desc = GEN8_CTX_VALID | GEN8_CTX_PRIVILEGE;
+
+	address_mode = INTEL_LEGACY_32B_CONTEXT;
+	if (ppgtt && i915_vm_is_48bit(&ppgtt->base))
+		address_mode = INTEL_LEGACY_64B_CONTEXT;
+	desc |= address_mode << GEN8_CTX_ADDRESSING_MODE_SHIFT;
+
+	if (IS_GEN8(i915))
+		desc |= GEN8_CTX_L3LLC_COHERENT;
+
+	/* TODO: WaDisableLiteRestore when we start using semaphore
+	 * signalling between Command Streamers
+	 * ring->ctx_desc_template |= GEN8_CTX_FORCE_RESTORE;
+	 */
+
+	return desc;
+}
+
 static struct i915_gem_context *
 __create_hw_context(struct drm_i915_private *dev_priv,
 		    struct drm_i915_file_private *file_priv)
@@ -257,8 +266,6 @@ __create_hw_context(struct drm_i915_private *dev_priv,
 	list_add_tail(&ctx->link, &dev_priv->context_list);
 	ctx->i915 = dev_priv;
 
-	ctx->ggtt_alignment = get_context_alignment(dev_priv);
-
 	if (dev_priv->hw_context_size) {
 		struct drm_i915_gem_object *obj;
 		struct i915_vma *vma;
@@ -309,8 +316,8 @@ __create_hw_context(struct drm_i915_private *dev_priv,
 
 	i915_gem_context_set_bannable(ctx);
 	ctx->ring_size = 4 * PAGE_SIZE;
-	ctx->desc_template = GEN8_CTX_ADDRESSING_MODE(dev_priv) <<
-			     GEN8_CTX_ADDRESSING_MODE_SHIFT;
+	ctx->desc_template =
+		default_desc_template(dev_priv, dev_priv->mm.aliasing_ppgtt);
 	ATOMIC_INIT_NOTIFIER_HEAD(&ctx->status_notifier);
 
 	/* GuC requires the ring to be placed above GUC_WOPCM_TOP. If GuC is not
@@ -332,6 +339,13 @@ err_out:
 	return ERR_PTR(ret);
 }
 
+static void __destroy_hw_context(struct i915_gem_context *ctx,
+				 struct drm_i915_file_private *file_priv)
+{
+	idr_remove(&file_priv->context_idr, ctx->user_handle);
+	context_close(ctx);
+}
+
 /**
  * The default context needs to exist per ring that uses contexts. It stores the
  * context state of the GPU for applications that don't utilize HW contexts, as
@@ -356,12 +370,12 @@ i915_gem_create_context(struct drm_i915_private *dev_priv,
 		if (IS_ERR(ppgtt)) {
 			DRM_DEBUG_DRIVER("PPGTT setup failed (%ld)\n",
 					 PTR_ERR(ppgtt));
-			idr_remove(&file_priv->context_idr, ctx->user_handle);
-			context_close(ctx);
+			__destroy_hw_context(ctx, file_priv);
 			return ERR_CAST(ppgtt);
 		}
 
 		ctx->ppgtt = ppgtt;
+		ctx->desc_template = default_desc_template(dev_priv, ppgtt);
 	}
 
 	trace_i915_context_create(ctx);
@@ -400,7 +414,8 @@ i915_gem_context_create_gvt(struct drm_device *dev)
 	i915_gem_context_set_closed(ctx); /* not user accessible */
 	i915_gem_context_clear_bannable(ctx);
 	i915_gem_context_set_force_single_submission(ctx);
-	ctx->ring_size = 512 * PAGE_SIZE; /* Max ring buffer size */
+	if (!i915.enable_guc_submission)
+		ctx->ring_size = 512 * PAGE_SIZE; /* Max ring buffer size */
 
 	GEM_BUG_ON(i915_gem_context_is_kernel(ctx));
 out:
@@ -451,6 +466,11 @@ int i915_gem_context_init(struct drm_i915_private *dev_priv)
 		return PTR_ERR(ctx);
 	}
 
+	/* For easy recognisablity, we want the kernel context to be 0 and then
+	 * all user contexts will have non-zero hw_id.
+	 */
+	GEM_BUG_ON(ctx->hw_id);
+
 	i915_gem_context_clear_bannable(ctx);
 	ctx->priority = I915_PRIORITY_MIN; /* lowest priority; idle task */
 	dev_priv->kernel_context = ctx;
@@ -560,27 +580,15 @@ static inline int
 mi_set_context(struct drm_i915_gem_request *req, u32 hw_flags)
 {
 	struct drm_i915_private *dev_priv = req->i915;
-	struct intel_ring *ring = req->ring;
 	struct intel_engine_cs *engine = req->engine;
 	enum intel_engine_id id;
-	u32 flags = hw_flags | MI_MM_SPACE_GTT;
+	u32 *cs, flags = hw_flags | MI_MM_SPACE_GTT;
 	const int num_rings =
 		/* Use an extended w/a on ivb+ if signalling from other rings */
 		i915.semaphores ?
 		INTEL_INFO(dev_priv)->num_rings - 1 :
 		0;
-	int len, ret;
-
-	/* w/a: If Flush TLB Invalidation Mode is enabled, driver must do a TLB
-	 * invalidation prior to MI_SET_CONTEXT. On GEN6 we don't set the value
-	 * explicitly, so we rely on the value at ring init, stored in
-	 * itlb_before_ctx_switch.
-	 */
-	if (IS_GEN6(dev_priv)) {
-		ret = engine->emit_flush(req, EMIT_INVALIDATE);
-		if (ret)
-			return ret;
-	}
+	int len;
 
 	/* These flags are for resource streamer on HSW+ */
 	if (IS_HASWELL(dev_priv) || INTEL_GEN(dev_priv) >= 8)
@@ -593,99 +601,92 @@ mi_set_context(struct drm_i915_gem_request *req, u32 hw_flags)
 	if (INTEL_GEN(dev_priv) >= 7)
 		len += 2 + (num_rings ? 4*num_rings + 6 : 0);
 
-	ret = intel_ring_begin(req, len);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, len);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	/* WaProgramMiArbOnOffAroundMiSetContext:ivb,vlv,hsw,bdw,chv */
 	if (INTEL_GEN(dev_priv) >= 7) {
-		intel_ring_emit(ring, MI_ARB_ON_OFF | MI_ARB_DISABLE);
+		*cs++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
 		if (num_rings) {
 			struct intel_engine_cs *signaller;
 
-			intel_ring_emit(ring,
-					MI_LOAD_REGISTER_IMM(num_rings));
+			*cs++ = MI_LOAD_REGISTER_IMM(num_rings);
 			for_each_engine(signaller, dev_priv, id) {
 				if (signaller == engine)
 					continue;
 
-				intel_ring_emit_reg(ring,
-						    RING_PSMI_CTL(signaller->mmio_base));
-				intel_ring_emit(ring,
-						_MASKED_BIT_ENABLE(GEN6_PSMI_SLEEP_MSG_DISABLE));
+				*cs++ = i915_mmio_reg_offset(
+					   RING_PSMI_CTL(signaller->mmio_base));
+				*cs++ = _MASKED_BIT_ENABLE(
+						GEN6_PSMI_SLEEP_MSG_DISABLE);
 			}
 		}
 	}
 
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_emit(ring, MI_SET_CONTEXT);
-	intel_ring_emit(ring,
-			i915_ggtt_offset(req->ctx->engine[RCS].state) | flags);
+	*cs++ = MI_NOOP;
+	*cs++ = MI_SET_CONTEXT;
+	*cs++ = i915_ggtt_offset(req->ctx->engine[RCS].state) | flags;
 	/*
 	 * w/a: MI_SET_CONTEXT must always be followed by MI_NOOP
 	 * WaMiSetContext_Hang:snb,ivb,vlv
 	 */
-	intel_ring_emit(ring, MI_NOOP);
+	*cs++ = MI_NOOP;
 
 	if (INTEL_GEN(dev_priv) >= 7) {
 		if (num_rings) {
 			struct intel_engine_cs *signaller;
 			i915_reg_t last_reg = {}; /* keep gcc quiet */
 
-			intel_ring_emit(ring,
-					MI_LOAD_REGISTER_IMM(num_rings));
+			*cs++ = MI_LOAD_REGISTER_IMM(num_rings);
 			for_each_engine(signaller, dev_priv, id) {
 				if (signaller == engine)
 					continue;
 
 				last_reg = RING_PSMI_CTL(signaller->mmio_base);
-				intel_ring_emit_reg(ring, last_reg);
-				intel_ring_emit(ring,
-						_MASKED_BIT_DISABLE(GEN6_PSMI_SLEEP_MSG_DISABLE));
+				*cs++ = i915_mmio_reg_offset(last_reg);
+				*cs++ = _MASKED_BIT_DISABLE(
+						GEN6_PSMI_SLEEP_MSG_DISABLE);
 			}
 
 			/* Insert a delay before the next switch! */
-			intel_ring_emit(ring,
-					MI_STORE_REGISTER_MEM |
-					MI_SRM_LRM_GLOBAL_GTT);
-			intel_ring_emit_reg(ring, last_reg);
-			intel_ring_emit(ring,
-					i915_ggtt_offset(engine->scratch));
-			intel_ring_emit(ring, MI_NOOP);
+			*cs++ = MI_STORE_REGISTER_MEM | MI_SRM_LRM_GLOBAL_GTT;
+			*cs++ = i915_mmio_reg_offset(last_reg);
+			*cs++ = i915_ggtt_offset(engine->scratch);
+			*cs++ = MI_NOOP;
 		}
-		intel_ring_emit(ring, MI_ARB_ON_OFF | MI_ARB_ENABLE);
+		*cs++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
 	}
 
-	intel_ring_advance(ring);
+	intel_ring_advance(req, cs);
 
-	return ret;
+	return 0;
 }
 
 static int remap_l3(struct drm_i915_gem_request *req, int slice)
 {
-	u32 *remap_info = req->i915->l3_parity.remap_info[slice];
-	struct intel_ring *ring = req->ring;
-	int i, ret;
+	u32 *cs, *remap_info = req->i915->l3_parity.remap_info[slice];
+	int i;
 
 	if (!remap_info)
 		return 0;
 
-	ret = intel_ring_begin(req, GEN7_L3LOG_SIZE/4 * 2 + 2);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, GEN7_L3LOG_SIZE/4 * 2 + 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	/*
 	 * Note: We do not worry about the concurrent register cacheline hang
 	 * here because no other code should access these registers other than
 	 * at initialization time.
 	 */
-	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(GEN7_L3LOG_SIZE/4));
+	*cs++ = MI_LOAD_REGISTER_IMM(GEN7_L3LOG_SIZE/4);
 	for (i = 0; i < GEN7_L3LOG_SIZE/4; i++) {
-		intel_ring_emit_reg(ring, GEN7_L3LOG(slice, i));
-		intel_ring_emit(ring, remap_info[i]);
+		*cs++ = i915_mmio_reg_offset(GEN7_L3LOG(slice, i));
+		*cs++ = remap_info[i];
 	}
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -1014,8 +1015,7 @@ int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
 		return PTR_ERR(ctx);
 	}
 
-	idr_remove(&file_priv->context_idr, ctx->user_handle);
-	context_close(ctx);
+	__destroy_hw_context(ctx, file_priv);
 	mutex_unlock(&dev->struct_mutex);
 
 	DRM_DEBUG("HW context %d destroyed\n", args->ctx_id);
@@ -1164,3 +1164,8 @@ int i915_gem_context_reset_stats_ioctl(struct drm_device *dev,
 
 	return 0;
 }
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/mock_context.c"
+#include "selftests/i915_gem_context.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_gem_context.h b/drivers/gpu/drm/i915/i915_gem_context.h
index 0ac750b90f3d..81268c9770a6 100644
--- a/drivers/gpu/drm/i915/i915_gem_context.h
+++ b/drivers/gpu/drm/i915/i915_gem_context.h
@@ -140,8 +140,6 @@ struct i915_gem_context {
 	 */
 	int priority;
 
-	/** ggtt_alignment: alignment restriction for context objects */
-	u32 ggtt_alignment;
 	/** ggtt_offset_bias: placement restriction for context objects */
 	u32 ggtt_offset_bias;
 
diff --git a/drivers/gpu/drm/i915/i915_gem_dmabuf.c b/drivers/gpu/drm/i915/i915_gem_dmabuf.c
index 29bb8011dbc4..11898cd97596 100644
--- a/drivers/gpu/drm/i915/i915_gem_dmabuf.c
+++ b/drivers/gpu/drm/i915/i915_gem_dmabuf.c
@@ -307,3 +307,8 @@ fail_detach:
 
 	return ERR_PTR(ret);
 }
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/mock_dmabuf.c"
+#include "selftests/i915_gem_dmabuf.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
index c181b1bb3d2c..a0de5734f7d0 100644
--- a/drivers/gpu/drm/i915/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -258,6 +258,9 @@ int i915_gem_evict_for_node(struct i915_address_space *vm,
 	int ret = 0;
 
 	lockdep_assert_held(&vm->i915->drm.struct_mutex);
+	GEM_BUG_ON(!IS_ALIGNED(start, I915_GTT_PAGE_SIZE));
+	GEM_BUG_ON(!IS_ALIGNED(end, I915_GTT_PAGE_SIZE));
+
 	trace_i915_gem_evict_node(vm, target, flags);
 
 	/* Retire before we search the active list. Although we have
@@ -271,11 +274,13 @@ int i915_gem_evict_for_node(struct i915_address_space *vm,
 	check_color = vm->mm.color_adjust;
 	if (check_color) {
 		/* Expand search to cover neighbouring guard pages (or lack!) */
-		if (start > vm->start)
+		if (start)
 			start -= I915_GTT_PAGE_SIZE;
-		if (end < vm->start + vm->total)
-			end += I915_GTT_PAGE_SIZE;
+
+		/* Always look at the page afterwards to avoid the end-of-GTT */
+		end += I915_GTT_PAGE_SIZE;
 	}
+	GEM_BUG_ON(start >= end);
 
 	drm_mm_for_each_node_in_range(node, &vm->mm, start, end) {
 		/* If we find any non-objects (!vma), we cannot evict them */
@@ -284,6 +289,7 @@ int i915_gem_evict_for_node(struct i915_address_space *vm,
 			break;
 		}
 
+		GEM_BUG_ON(!node->allocated);
 		vma = container_of(node, typeof(*vma), node);
 
 		/* If we are using coloring to insert guard pages between
@@ -387,3 +393,7 @@ int i915_gem_evict_vm(struct i915_address_space *vm, bool do_idle)
 
 	return 0;
 }
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/i915_gem_evict.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index d02cfaefe1c8..dd7181ed5eca 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -28,12 +28,14 @@
 
 #include <linux/dma_remapping.h>
 #include <linux/reservation.h>
+#include <linux/sync_file.h>
 #include <linux/uaccess.h>
 
 #include <drm/drmP.h>
 #include <drm/i915_drm.h>
 
 #include "i915_drv.h"
+#include "i915_gem_clflush.h"
 #include "i915_trace.h"
 #include "intel_drv.h"
 #include "intel_frontbuffer.h"
@@ -1110,13 +1112,18 @@ i915_gem_execbuffer_move_to_gpu(struct drm_i915_gem_request *req,
 	list_for_each_entry(vma, vmas, exec_list) {
 		struct drm_i915_gem_object *obj = vma->obj;
 
+		if (vma->exec_entry->flags & EXEC_OBJECT_ASYNC)
+			continue;
+
+		if (obj->base.write_domain & I915_GEM_DOMAIN_CPU) {
+			i915_gem_clflush_object(obj, 0);
+			obj->base.write_domain = 0;
+		}
+
 		ret = i915_gem_request_await_object
 			(req, obj, obj->base.pending_write_domain);
 		if (ret)
 			return ret;
-
-		if (obj->base.write_domain & I915_GEM_DOMAIN_CPU)
-			i915_gem_clflush_object(obj, false);
 	}
 
 	/* Unconditionally flush any chipset caches (for streaming writes). */
@@ -1297,12 +1304,12 @@ static void eb_export_fence(struct drm_i915_gem_object *obj,
 	 * handle an error right now. Worst case should be missed
 	 * synchronisation leading to rendering corruption.
 	 */
-	ww_mutex_lock(&resv->lock, NULL);
+	reservation_object_lock(resv, NULL);
 	if (flags & EXEC_OBJECT_WRITE)
 		reservation_object_add_excl_fence(resv, &req->fence);
 	else if (reservation_object_reserve_shared(resv) == 0)
 		reservation_object_add_shared_fence(resv, &req->fence);
-	ww_mutex_unlock(&resv->lock);
+	reservation_object_unlock(resv);
 }
 
 static void
@@ -1313,8 +1320,6 @@ i915_gem_execbuffer_move_to_active(struct list_head *vmas,
 
 	list_for_each_entry(vma, vmas, exec_list) {
 		struct drm_i915_gem_object *obj = vma->obj;
-		u32 old_read = obj->base.read_domains;
-		u32 old_write = obj->base.write_domain;
 
 		obj->base.write_domain = obj->base.pending_write_domain;
 		if (obj->base.write_domain)
@@ -1325,32 +1330,31 @@ i915_gem_execbuffer_move_to_active(struct list_head *vmas,
 
 		i915_vma_move_to_active(vma, req, vma->exec_entry->flags);
 		eb_export_fence(obj, req, vma->exec_entry->flags);
-		trace_i915_gem_object_change_domain(obj, old_read, old_write);
 	}
 }
 
 static int
 i915_reset_gen7_sol_offsets(struct drm_i915_gem_request *req)
 {
-	struct intel_ring *ring = req->ring;
-	int ret, i;
+	u32 *cs;
+	int i;
 
 	if (!IS_GEN7(req->i915) || req->engine->id != RCS) {
 		DRM_DEBUG("sol reset is gen7/rcs only\n");
 		return -EINVAL;
 	}
 
-	ret = intel_ring_begin(req, 4 * 3);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 4 * 3);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	for (i = 0; i < 4; i++) {
-		intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
-		intel_ring_emit_reg(ring, GEN7_SO_WRITE_OFFSET(i));
-		intel_ring_emit(ring, 0);
+		*cs++ = MI_LOAD_REGISTER_IMM(1);
+		*cs++ = i915_mmio_reg_offset(GEN7_SO_WRITE_OFFSET(i));
+		*cs++ = 0;
 	}
 
-	intel_ring_advance(ring);
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -1403,15 +1407,20 @@ out:
 	return vma;
 }
 
+static void
+add_to_client(struct drm_i915_gem_request *req,
+	      struct drm_file *file)
+{
+	req->file_priv = file->driver_priv;
+	list_add_tail(&req->client_link, &req->file_priv->mm.request_list);
+}
+
 static int
 execbuf_submit(struct i915_execbuffer_params *params,
 	       struct drm_i915_gem_execbuffer2 *args,
 	       struct list_head *vmas)
 {
-	struct drm_i915_private *dev_priv = params->request->i915;
 	u64 exec_start, exec_len;
-	int instp_mode;
-	u32 instp_mask;
 	int ret;
 
 	ret = i915_gem_execbuffer_move_to_gpu(params->request, vmas);
@@ -1422,56 +1431,11 @@ execbuf_submit(struct i915_execbuffer_params *params,
 	if (ret)
 		return ret;
 
-	instp_mode = args->flags & I915_EXEC_CONSTANTS_MASK;
-	instp_mask = I915_EXEC_CONSTANTS_MASK;
-	switch (instp_mode) {
-	case I915_EXEC_CONSTANTS_REL_GENERAL:
-	case I915_EXEC_CONSTANTS_ABSOLUTE:
-	case I915_EXEC_CONSTANTS_REL_SURFACE:
-		if (instp_mode != 0 && params->engine->id != RCS) {
-			DRM_DEBUG("non-0 rel constants mode on non-RCS\n");
-			return -EINVAL;
-		}
-
-		if (instp_mode != dev_priv->relative_constants_mode) {
-			if (INTEL_INFO(dev_priv)->gen < 4) {
-				DRM_DEBUG("no rel constants on pre-gen4\n");
-				return -EINVAL;
-			}
-
-			if (INTEL_INFO(dev_priv)->gen > 5 &&
-			    instp_mode == I915_EXEC_CONSTANTS_REL_SURFACE) {
-				DRM_DEBUG("rel surface constants mode invalid on gen5+\n");
-				return -EINVAL;
-			}
-
-			/* The HW changed the meaning on this bit on gen6 */
-			if (INTEL_INFO(dev_priv)->gen >= 6)
-				instp_mask &= ~I915_EXEC_CONSTANTS_REL_SURFACE;
-		}
-		break;
-	default:
-		DRM_DEBUG("execbuf with unknown constants: %d\n", instp_mode);
+	if (args->flags & I915_EXEC_CONSTANTS_MASK) {
+		DRM_DEBUG("I915_EXEC_CONSTANTS_* unsupported\n");
 		return -EINVAL;
 	}
 
-	if (params->engine->id == RCS &&
-	    instp_mode != dev_priv->relative_constants_mode) {
-		struct intel_ring *ring = params->request->ring;
-
-		ret = intel_ring_begin(params->request, 4);
-		if (ret)
-			return ret;
-
-		intel_ring_emit(ring, MI_NOOP);
-		intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
-		intel_ring_emit_reg(ring, INSTPM);
-		intel_ring_emit(ring, instp_mask << 16 | instp_mode);
-		intel_ring_advance(ring);
-
-		dev_priv->relative_constants_mode = instp_mode;
-	}
-
 	if (args->flags & I915_EXEC_GEN7_SOL_RESET) {
 		ret = i915_reset_gen7_sol_offsets(params->request);
 		if (ret)
@@ -1491,8 +1455,6 @@ execbuf_submit(struct i915_execbuffer_params *params,
 	if (ret)
 		return ret;
 
-	trace_i915_gem_ring_dispatch(params->request, params->dispatch_flags);
-
 	i915_gem_execbuffer_move_to_active(vmas, params->request);
 
 	return 0;
@@ -1591,6 +1553,9 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	struct i915_execbuffer_params *params = &params_master;
 	const u32 ctx_id = i915_execbuffer2_get_context_id(*args);
 	u32 dispatch_flags;
+	struct dma_fence *in_fence = NULL;
+	struct sync_file *out_fence = NULL;
+	int out_fence_fd = -1;
 	int ret;
 	bool need_relocs;
 
@@ -1634,6 +1599,20 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		dispatch_flags |= I915_DISPATCH_RS;
 	}
 
+	if (args->flags & I915_EXEC_FENCE_IN) {
+		in_fence = sync_file_get_fence(lower_32_bits(args->rsvd2));
+		if (!in_fence)
+			return -EINVAL;
+	}
+
+	if (args->flags & I915_EXEC_FENCE_OUT) {
+		out_fence_fd = get_unused_fd_flags(O_CLOEXEC);
+		if (out_fence_fd < 0) {
+			ret = out_fence_fd;
+			goto err_in_fence;
+		}
+	}
+
 	/* Take a local wakeref for preparing to dispatch the execbuf as
 	 * we expect to access the hardware fairly frequently in the
 	 * process. Upon first dispatch, we acquire another prolonged
@@ -1778,6 +1757,21 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		goto err_batch_unpin;
 	}
 
+	if (in_fence) {
+		ret = i915_gem_request_await_dma_fence(params->request,
+						       in_fence);
+		if (ret < 0)
+			goto err_request;
+	}
+
+	if (out_fence_fd != -1) {
+		out_fence = sync_file_create(&params->request->fence);
+		if (!out_fence) {
+			ret = -ENOMEM;
+			goto err_request;
+		}
+	}
+
 	/* Whilst this request exists, batch_obj will be on the
 	 * active_list, and so will hold the active reference. Only when this
 	 * request is retired will the the batch_obj be moved onto the
@@ -1786,10 +1780,6 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	 */
 	params->request->batch = params->batch;
 
-	ret = i915_gem_request_add_to_client(params->request, file);
-	if (ret)
-		goto err_request;
-
 	/*
 	 * Save assorted stuff away to pass through to *_submission().
 	 * NB: This data should be 'persistent' and not local as it will
@@ -1802,9 +1792,23 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	params->dispatch_flags          = dispatch_flags;
 	params->ctx                     = ctx;
 
+	trace_i915_gem_request_queue(params->request, dispatch_flags);
+
 	ret = execbuf_submit(params, args, &eb->vmas);
 err_request:
 	__i915_add_request(params->request, ret == 0);
+	add_to_client(params->request, file);
+
+	if (out_fence) {
+		if (ret == 0) {
+			fd_install(out_fence_fd, out_fence->file);
+			args->rsvd2 &= GENMASK_ULL(0, 31); /* keep in-fence */
+			args->rsvd2 |= (u64)out_fence_fd << 32;
+			out_fence_fd = -1;
+		} else {
+			fput(out_fence->file);
+		}
+	}
 
 err_batch_unpin:
 	/*
@@ -1826,6 +1830,10 @@ pre_mutex_err:
 	/* intel_gpu_busy should also get a ref, so it will free when the device
 	 * is really idle. */
 	intel_runtime_pm_put(dev_priv);
+	if (out_fence_fd != -1)
+		put_unused_fd(out_fence_fd);
+err_in_fence:
+	dma_fence_put(in_fence);
 	return ret;
 }
 
@@ -1933,11 +1941,6 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 		return -EINVAL;
 	}
 
-	if (args->rsvd2 != 0) {
-		DRM_DEBUG("dirty rvsd2 field\n");
-		return -EINVAL;
-	}
-
 	exec2_list = drm_malloc_gfp(args->buffer_count,
 				    sizeof(*exec2_list),
 				    GFP_TEMPORARY);
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 2801a4d56324..cee9c4fec52a 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -23,6 +23,9 @@
  *
  */
 
+#include <linux/slab.h> /* fault-inject.h is not standalone! */
+
+#include <linux/fault-inject.h>
 #include <linux/log2.h>
 #include <linux/random.h>
 #include <linux/seq_file.h>
@@ -187,11 +190,17 @@ static int ppgtt_bind_vma(struct i915_vma *vma,
 			  enum i915_cache_level cache_level,
 			  u32 unused)
 {
-	u32 pte_flags = 0;
+	u32 pte_flags;
+	int ret;
+
+	ret = vma->vm->allocate_va_range(vma->vm, vma->node.start, vma->size);
+	if (ret)
+		return ret;
 
 	vma->pages = vma->obj->mm.pages;
 
 	/* Currently applicable only to VLV */
+	pte_flags = 0;
 	if (vma->obj->gt_ro)
 		pte_flags |= PTE_READ_ONLY;
 
@@ -203,9 +212,7 @@ static int ppgtt_bind_vma(struct i915_vma *vma,
 
 static void ppgtt_unbind_vma(struct i915_vma *vma)
 {
-	vma->vm->clear_range(vma->vm,
-			     vma->node.start,
-			     vma->size);
+	vma->vm->clear_range(vma->vm, vma->node.start, vma->size);
 }
 
 static gen8_pte_t gen8_pte_encode(dma_addr_t addr,
@@ -340,268 +347,229 @@ static gen6_pte_t iris_pte_encode(dma_addr_t addr,
 	return pte;
 }
 
-static int __setup_page_dma(struct drm_i915_private *dev_priv,
-			    struct i915_page_dma *p, gfp_t flags)
+static struct page *vm_alloc_page(struct i915_address_space *vm, gfp_t gfp)
 {
-	struct device *kdev = &dev_priv->drm.pdev->dev;
+	struct page *page;
 
-	p->page = alloc_page(flags);
-	if (!p->page)
-		return -ENOMEM;
+	if (I915_SELFTEST_ONLY(should_fail(&vm->fault_attr, 1)))
+		i915_gem_shrink_all(vm->i915);
 
-	p->daddr = dma_map_page(kdev,
-				p->page, 0, PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
+	if (vm->free_pages.nr)
+		return vm->free_pages.pages[--vm->free_pages.nr];
 
-	if (dma_mapping_error(kdev, p->daddr)) {
-		__free_page(p->page);
-		return -EINVAL;
-	}
+	page = alloc_page(gfp);
+	if (!page)
+		return NULL;
 
-	return 0;
+	if (vm->pt_kmap_wc)
+		set_pages_array_wc(&page, 1);
+
+	return page;
 }
 
-static int setup_page_dma(struct drm_i915_private *dev_priv,
-			  struct i915_page_dma *p)
+static void vm_free_pages_release(struct i915_address_space *vm)
 {
-	return __setup_page_dma(dev_priv, p, I915_GFP_DMA);
+	GEM_BUG_ON(!pagevec_count(&vm->free_pages));
+
+	if (vm->pt_kmap_wc)
+		set_pages_array_wb(vm->free_pages.pages,
+				   pagevec_count(&vm->free_pages));
+
+	__pagevec_release(&vm->free_pages);
 }
 
-static void cleanup_page_dma(struct drm_i915_private *dev_priv,
-			     struct i915_page_dma *p)
+static void vm_free_page(struct i915_address_space *vm, struct page *page)
 {
-	struct pci_dev *pdev = dev_priv->drm.pdev;
+	if (!pagevec_add(&vm->free_pages, page))
+		vm_free_pages_release(vm);
+}
 
-	if (WARN_ON(!p->page))
-		return;
+static int __setup_page_dma(struct i915_address_space *vm,
+			    struct i915_page_dma *p,
+			    gfp_t gfp)
+{
+	p->page = vm_alloc_page(vm, gfp | __GFP_NOWARN | __GFP_NORETRY);
+	if (unlikely(!p->page))
+		return -ENOMEM;
+
+	p->daddr = dma_map_page(vm->dma, p->page, 0, PAGE_SIZE,
+				PCI_DMA_BIDIRECTIONAL);
+	if (unlikely(dma_mapping_error(vm->dma, p->daddr))) {
+		vm_free_page(vm, p->page);
+		return -ENOMEM;
+	}
 
-	dma_unmap_page(&pdev->dev, p->daddr, PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
-	__free_page(p->page);
-	memset(p, 0, sizeof(*p));
+	return 0;
 }
 
-static void *kmap_page_dma(struct i915_page_dma *p)
+static int setup_page_dma(struct i915_address_space *vm,
+			  struct i915_page_dma *p)
 {
-	return kmap_atomic(p->page);
+	return __setup_page_dma(vm, p, I915_GFP_DMA);
 }
 
-/* We use the flushing unmap only with ppgtt structures:
- * page directories, page tables and scratch pages.
- */
-static void kunmap_page_dma(struct drm_i915_private *dev_priv, void *vaddr)
+static void cleanup_page_dma(struct i915_address_space *vm,
+			     struct i915_page_dma *p)
 {
-	/* There are only few exceptions for gen >=6. chv and bxt.
-	 * And we are not sure about the latter so play safe for now.
-	 */
-	if (IS_CHERRYVIEW(dev_priv) || IS_GEN9_LP(dev_priv))
-		drm_clflush_virt_range(vaddr, PAGE_SIZE);
-
-	kunmap_atomic(vaddr);
+	dma_unmap_page(vm->dma, p->daddr, PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
+	vm_free_page(vm, p->page);
 }
 
-#define kmap_px(px) kmap_page_dma(px_base(px))
-#define kunmap_px(ppgtt, vaddr) \
-		kunmap_page_dma((ppgtt)->base.i915, (vaddr))
+#define kmap_atomic_px(px) kmap_atomic(px_base(px)->page)
 
-#define setup_px(dev_priv, px) setup_page_dma((dev_priv), px_base(px))
-#define cleanup_px(dev_priv, px) cleanup_page_dma((dev_priv), px_base(px))
-#define fill_px(dev_priv, px, v) fill_page_dma((dev_priv), px_base(px), (v))
-#define fill32_px(dev_priv, px, v) \
-		fill_page_dma_32((dev_priv), px_base(px), (v))
+#define setup_px(vm, px) setup_page_dma((vm), px_base(px))
+#define cleanup_px(vm, px) cleanup_page_dma((vm), px_base(px))
+#define fill_px(ppgtt, px, v) fill_page_dma((vm), px_base(px), (v))
+#define fill32_px(ppgtt, px, v) fill_page_dma_32((vm), px_base(px), (v))
 
-static void fill_page_dma(struct drm_i915_private *dev_priv,
-			  struct i915_page_dma *p, const uint64_t val)
+static void fill_page_dma(struct i915_address_space *vm,
+			  struct i915_page_dma *p,
+			  const u64 val)
 {
+	u64 * const vaddr = kmap_atomic(p->page);
 	int i;
-	uint64_t * const vaddr = kmap_page_dma(p);
 
 	for (i = 0; i < 512; i++)
 		vaddr[i] = val;
 
-	kunmap_page_dma(dev_priv, vaddr);
+	kunmap_atomic(vaddr);
 }
 
-static void fill_page_dma_32(struct drm_i915_private *dev_priv,
-			     struct i915_page_dma *p, const uint32_t val32)
+static void fill_page_dma_32(struct i915_address_space *vm,
+			     struct i915_page_dma *p,
+			     const u32 v)
 {
-	uint64_t v = val32;
-
-	v = v << 32 | val32;
-
-	fill_page_dma(dev_priv, p, v);
+	fill_page_dma(vm, p, (u64)v << 32 | v);
 }
 
 static int
-setup_scratch_page(struct drm_i915_private *dev_priv,
-		   struct i915_page_dma *scratch,
-		   gfp_t gfp)
+setup_scratch_page(struct i915_address_space *vm, gfp_t gfp)
 {
-	return __setup_page_dma(dev_priv, scratch, gfp | __GFP_ZERO);
+	return __setup_page_dma(vm, &vm->scratch_page, gfp | __GFP_ZERO);
 }
 
-static void cleanup_scratch_page(struct drm_i915_private *dev_priv,
-				 struct i915_page_dma *scratch)
+static void cleanup_scratch_page(struct i915_address_space *vm)
 {
-	cleanup_page_dma(dev_priv, scratch);
+	cleanup_page_dma(vm, &vm->scratch_page);
 }
 
-static struct i915_page_table *alloc_pt(struct drm_i915_private *dev_priv)
+static struct i915_page_table *alloc_pt(struct i915_address_space *vm)
 {
 	struct i915_page_table *pt;
-	const size_t count = INTEL_GEN(dev_priv) >= 8 ? GEN8_PTES : GEN6_PTES;
-	int ret = -ENOMEM;
 
-	pt = kzalloc(sizeof(*pt), GFP_KERNEL);
-	if (!pt)
+	pt = kmalloc(sizeof(*pt), GFP_KERNEL | __GFP_NOWARN);
+	if (unlikely(!pt))
 		return ERR_PTR(-ENOMEM);
 
-	pt->used_ptes = kcalloc(BITS_TO_LONGS(count), sizeof(*pt->used_ptes),
-				GFP_KERNEL);
-
-	if (!pt->used_ptes)
-		goto fail_bitmap;
-
-	ret = setup_px(dev_priv, pt);
-	if (ret)
-		goto fail_page_m;
+	if (unlikely(setup_px(vm, pt))) {
+		kfree(pt);
+		return ERR_PTR(-ENOMEM);
+	}
 
+	pt->used_ptes = 0;
 	return pt;
-
-fail_page_m:
-	kfree(pt->used_ptes);
-fail_bitmap:
-	kfree(pt);
-
-	return ERR_PTR(ret);
 }
 
-static void free_pt(struct drm_i915_private *dev_priv,
-		    struct i915_page_table *pt)
+static void free_pt(struct i915_address_space *vm, struct i915_page_table *pt)
 {
-	cleanup_px(dev_priv, pt);
-	kfree(pt->used_ptes);
+	cleanup_px(vm, pt);
 	kfree(pt);
 }
 
 static void gen8_initialize_pt(struct i915_address_space *vm,
 			       struct i915_page_table *pt)
 {
-	gen8_pte_t scratch_pte;
-
-	scratch_pte = gen8_pte_encode(vm->scratch_page.daddr,
-				      I915_CACHE_LLC);
-
-	fill_px(vm->i915, pt, scratch_pte);
+	fill_px(vm, pt,
+		gen8_pte_encode(vm->scratch_page.daddr, I915_CACHE_LLC));
 }
 
 static void gen6_initialize_pt(struct i915_address_space *vm,
 			       struct i915_page_table *pt)
 {
-	gen6_pte_t scratch_pte;
-
-	WARN_ON(vm->scratch_page.daddr == 0);
-
-	scratch_pte = vm->pte_encode(vm->scratch_page.daddr,
-				     I915_CACHE_LLC, 0);
-
-	fill32_px(vm->i915, pt, scratch_pte);
+	fill32_px(vm, pt,
+		  vm->pte_encode(vm->scratch_page.daddr, I915_CACHE_LLC, 0));
 }
 
-static struct i915_page_directory *alloc_pd(struct drm_i915_private *dev_priv)
+static struct i915_page_directory *alloc_pd(struct i915_address_space *vm)
 {
 	struct i915_page_directory *pd;
-	int ret = -ENOMEM;
 
-	pd = kzalloc(sizeof(*pd), GFP_KERNEL);
-	if (!pd)
+	pd = kzalloc(sizeof(*pd), GFP_KERNEL | __GFP_NOWARN);
+	if (unlikely(!pd))
 		return ERR_PTR(-ENOMEM);
 
-	pd->used_pdes = kcalloc(BITS_TO_LONGS(I915_PDES),
-				sizeof(*pd->used_pdes), GFP_KERNEL);
-	if (!pd->used_pdes)
-		goto fail_bitmap;
-
-	ret = setup_px(dev_priv, pd);
-	if (ret)
-		goto fail_page_m;
+	if (unlikely(setup_px(vm, pd))) {
+		kfree(pd);
+		return ERR_PTR(-ENOMEM);
+	}
 
+	pd->used_pdes = 0;
 	return pd;
-
-fail_page_m:
-	kfree(pd->used_pdes);
-fail_bitmap:
-	kfree(pd);
-
-	return ERR_PTR(ret);
 }
 
-static void free_pd(struct drm_i915_private *dev_priv,
+static void free_pd(struct i915_address_space *vm,
 		    struct i915_page_directory *pd)
 {
-	if (px_page(pd)) {
-		cleanup_px(dev_priv, pd);
-		kfree(pd->used_pdes);
-		kfree(pd);
-	}
+	cleanup_px(vm, pd);
+	kfree(pd);
 }
 
 static void gen8_initialize_pd(struct i915_address_space *vm,
 			       struct i915_page_directory *pd)
 {
-	gen8_pde_t scratch_pde;
-
-	scratch_pde = gen8_pde_encode(px_dma(vm->scratch_pt), I915_CACHE_LLC);
+	unsigned int i;
 
-	fill_px(vm->i915, pd, scratch_pde);
+	fill_px(vm, pd,
+		gen8_pde_encode(px_dma(vm->scratch_pt), I915_CACHE_LLC));
+	for (i = 0; i < I915_PDES; i++)
+		pd->page_table[i] = vm->scratch_pt;
 }
 
-static int __pdp_init(struct drm_i915_private *dev_priv,
+static int __pdp_init(struct i915_address_space *vm,
 		      struct i915_page_directory_pointer *pdp)
 {
-	size_t pdpes = I915_PDPES_PER_PDP(dev_priv);
+	const unsigned int pdpes = i915_pdpes_per_pdp(vm);
+	unsigned int i;
 
-	pdp->used_pdpes = kcalloc(BITS_TO_LONGS(pdpes),
-				  sizeof(unsigned long),
-				  GFP_KERNEL);
-	if (!pdp->used_pdpes)
+	pdp->page_directory = kmalloc_array(pdpes, sizeof(*pdp->page_directory),
+					    GFP_KERNEL | __GFP_NOWARN);
+	if (unlikely(!pdp->page_directory))
 		return -ENOMEM;
 
-	pdp->page_directory = kcalloc(pdpes, sizeof(*pdp->page_directory),
-				      GFP_KERNEL);
-	if (!pdp->page_directory) {
-		kfree(pdp->used_pdpes);
-		/* the PDP might be the statically allocated top level. Keep it
-		 * as clean as possible */
-		pdp->used_pdpes = NULL;
-		return -ENOMEM;
-	}
+	for (i = 0; i < pdpes; i++)
+		pdp->page_directory[i] = vm->scratch_pd;
 
 	return 0;
 }
 
 static void __pdp_fini(struct i915_page_directory_pointer *pdp)
 {
-	kfree(pdp->used_pdpes);
 	kfree(pdp->page_directory);
 	pdp->page_directory = NULL;
 }
 
-static struct
-i915_page_directory_pointer *alloc_pdp(struct drm_i915_private *dev_priv)
+static inline bool use_4lvl(const struct i915_address_space *vm)
+{
+	return i915_vm_is_48bit(vm);
+}
+
+static struct i915_page_directory_pointer *
+alloc_pdp(struct i915_address_space *vm)
 {
 	struct i915_page_directory_pointer *pdp;
 	int ret = -ENOMEM;
 
-	WARN_ON(!USES_FULL_48BIT_PPGTT(dev_priv));
+	WARN_ON(!use_4lvl(vm));
 
 	pdp = kzalloc(sizeof(*pdp), GFP_KERNEL);
 	if (!pdp)
 		return ERR_PTR(-ENOMEM);
 
-	ret = __pdp_init(dev_priv, pdp);
+	ret = __pdp_init(vm, pdp);
 	if (ret)
 		goto fail_bitmap;
 
-	ret = setup_px(dev_priv, pdp);
+	ret = setup_px(vm, pdp);
 	if (ret)
 		goto fail_page_m;
 
@@ -615,14 +583,16 @@ fail_bitmap:
 	return ERR_PTR(ret);
 }
 
-static void free_pdp(struct drm_i915_private *dev_priv,
+static void free_pdp(struct i915_address_space *vm,
 		     struct i915_page_directory_pointer *pdp)
 {
 	__pdp_fini(pdp);
-	if (USES_FULL_48BIT_PPGTT(dev_priv)) {
-		cleanup_px(dev_priv, pdp);
-		kfree(pdp);
-	}
+
+	if (!use_4lvl(vm))
+		return;
+
+	cleanup_px(vm, pdp);
+	kfree(pdp);
 }
 
 static void gen8_initialize_pdp(struct i915_address_space *vm,
@@ -632,47 +602,18 @@ static void gen8_initialize_pdp(struct i915_address_space *vm,
 
 	scratch_pdpe = gen8_pdpe_encode(px_dma(vm->scratch_pd), I915_CACHE_LLC);
 
-	fill_px(vm->i915, pdp, scratch_pdpe);
+	fill_px(vm, pdp, scratch_pdpe);
 }
 
 static void gen8_initialize_pml4(struct i915_address_space *vm,
 				 struct i915_pml4 *pml4)
 {
-	gen8_ppgtt_pml4e_t scratch_pml4e;
-
-	scratch_pml4e = gen8_pml4e_encode(px_dma(vm->scratch_pdp),
-					  I915_CACHE_LLC);
-
-	fill_px(vm->i915, pml4, scratch_pml4e);
-}
-
-static void
-gen8_setup_pdpe(struct i915_hw_ppgtt *ppgtt,
-		struct i915_page_directory_pointer *pdp,
-		struct i915_page_directory *pd,
-		int index)
-{
-	gen8_ppgtt_pdpe_t *page_directorypo;
+	unsigned int i;
 
-	if (!USES_FULL_48BIT_PPGTT(to_i915(ppgtt->base.dev)))
-		return;
-
-	page_directorypo = kmap_px(pdp);
-	page_directorypo[index] = gen8_pdpe_encode(px_dma(pd), I915_CACHE_LLC);
-	kunmap_px(ppgtt, page_directorypo);
-}
-
-static void
-gen8_setup_pml4e(struct i915_hw_ppgtt *ppgtt,
-		 struct i915_pml4 *pml4,
-		 struct i915_page_directory_pointer *pdp,
-		 int index)
-{
-	gen8_ppgtt_pml4e_t *pagemap = kmap_px(pml4);
-
-	WARN_ON(!USES_FULL_48BIT_PPGTT(to_i915(ppgtt->base.dev)));
-	pagemap[index] = gen8_pml4e_encode(px_dma(pdp), I915_CACHE_LLC);
-	kunmap_px(ppgtt, pagemap);
+	fill_px(vm, pml4,
+		gen8_pml4e_encode(px_dma(vm->scratch_pdp), I915_CACHE_LLC));
+	for (i = 0; i < GEN8_PML4ES_PER_PML4; i++)
+		pml4->pdps[i] = vm->scratch_pdp;
 }
 
 /* Broadwell Page Directory Pointer Descriptors */
@@ -680,33 +621,32 @@ static int gen8_write_pdp(struct drm_i915_gem_request *req,
 			  unsigned entry,
 			  dma_addr_t addr)
 {
-	struct intel_ring *ring = req->ring;
 	struct intel_engine_cs *engine = req->engine;
-	int ret;
+	u32 *cs;
 
 	BUG_ON(entry >= 4);
 
-	ret = intel_ring_begin(req, 6);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
-	intel_ring_emit_reg(ring, GEN8_RING_PDP_UDW(engine, entry));
-	intel_ring_emit(ring, upper_32_bits(addr));
-	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
-	intel_ring_emit_reg(ring, GEN8_RING_PDP_LDW(engine, entry));
-	intel_ring_emit(ring, lower_32_bits(addr));
-	intel_ring_advance(ring);
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_UDW(engine, entry));
+	*cs++ = upper_32_bits(addr);
+	*cs++ = MI_LOAD_REGISTER_IMM(1);
+	*cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_LDW(engine, entry));
+	*cs++ = lower_32_bits(addr);
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
 
-static int gen8_legacy_mm_switch(struct i915_hw_ppgtt *ppgtt,
-				 struct drm_i915_gem_request *req)
+static int gen8_mm_switch_3lvl(struct i915_hw_ppgtt *ppgtt,
+			       struct drm_i915_gem_request *req)
 {
 	int i, ret;
 
-	for (i = GEN8_LEGACY_PDPES - 1; i >= 0; i--) {
+	for (i = GEN8_3LVL_PDPES - 1; i >= 0; i--) {
 		const dma_addr_t pd_daddr = i915_page_dir_dma_addr(ppgtt, i);
 
 		ret = gen8_write_pdp(req, i, pd_daddr);
@@ -717,8 +657,8 @@ static int gen8_legacy_mm_switch(struct i915_hw_ppgtt *ppgtt,
 	return 0;
 }
 
-static int gen8_48b_mm_switch(struct i915_hw_ppgtt *ppgtt,
-			      struct drm_i915_gem_request *req)
+static int gen8_mm_switch_4lvl(struct i915_hw_ppgtt *ppgtt,
+			       struct drm_i915_gem_request *req)
 {
 	return gen8_write_pdp(req, 0, px_dma(&ppgtt->pml4));
 }
@@ -738,70 +678,80 @@ static void mark_tlbs_dirty(struct i915_hw_ppgtt *ppgtt)
  */
 static bool gen8_ppgtt_clear_pt(struct i915_address_space *vm,
 				struct i915_page_table *pt,
-				uint64_t start,
-				uint64_t length)
+				u64 start, u64 length)
 {
-	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
 	unsigned int num_entries = gen8_pte_count(start, length);
 	unsigned int pte = gen8_pte_index(start);
 	unsigned int pte_end = pte + num_entries;
-	gen8_pte_t *pt_vaddr;
-	gen8_pte_t scratch_pte = gen8_pte_encode(vm->scratch_page.daddr,
-						 I915_CACHE_LLC);
+	const gen8_pte_t scratch_pte =
+		gen8_pte_encode(vm->scratch_page.daddr, I915_CACHE_LLC);
+	gen8_pte_t *vaddr;
 
-	if (WARN_ON(!px_page(pt)))
-		return false;
+	GEM_BUG_ON(num_entries > pt->used_ptes);
 
-	GEM_BUG_ON(pte_end > GEN8_PTES);
+	pt->used_ptes -= num_entries;
+	if (!pt->used_ptes)
+		return true;
 
-	bitmap_clear(pt->used_ptes, pte, num_entries);
-	if (USES_FULL_PPGTT(vm->i915)) {
-		if (bitmap_empty(pt->used_ptes, GEN8_PTES))
-			return true;
-	}
+	vaddr = kmap_atomic_px(pt);
+	while (pte < pte_end)
+		vaddr[pte++] = scratch_pte;
+	kunmap_atomic(vaddr);
 
-	pt_vaddr = kmap_px(pt);
+	return false;
+}
 
-	while (pte < pte_end)
-		pt_vaddr[pte++] = scratch_pte;
+static void gen8_ppgtt_set_pde(struct i915_address_space *vm,
+			       struct i915_page_directory *pd,
+			       struct i915_page_table *pt,
+			       unsigned int pde)
+{
+	gen8_pde_t *vaddr;
 
-	kunmap_px(ppgtt, pt_vaddr);
+	pd->page_table[pde] = pt;
 
-	return false;
+	vaddr = kmap_atomic_px(pd);
+	vaddr[pde] = gen8_pde_encode(px_dma(pt), I915_CACHE_LLC);
+	kunmap_atomic(vaddr);
 }
 
-/* Removes entries from a single page dir, releasing it if it's empty.
- * Caller can use the return value to update higher-level entries
- */
 static bool gen8_ppgtt_clear_pd(struct i915_address_space *vm,
 				struct i915_page_directory *pd,
-				uint64_t start,
-				uint64_t length)
+				u64 start, u64 length)
 {
-	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
 	struct i915_page_table *pt;
-	uint64_t pde;
-	gen8_pde_t *pde_vaddr;
-	gen8_pde_t scratch_pde = gen8_pde_encode(px_dma(vm->scratch_pt),
-						 I915_CACHE_LLC);
+	u32 pde;
 
 	gen8_for_each_pde(pt, pd, start, length, pde) {
-		if (WARN_ON(!pd->page_table[pde]))
-			break;
-
-		if (gen8_ppgtt_clear_pt(vm, pt, start, length)) {
-			__clear_bit(pde, pd->used_pdes);
-			pde_vaddr = kmap_px(pd);
-			pde_vaddr[pde] = scratch_pde;
-			kunmap_px(ppgtt, pde_vaddr);
-			free_pt(vm->i915, pt);
-		}
+		GEM_BUG_ON(pt == vm->scratch_pt);
+
+		if (!gen8_ppgtt_clear_pt(vm, pt, start, length))
+			continue;
+
+		gen8_ppgtt_set_pde(vm, pd, vm->scratch_pt, pde);
+		GEM_BUG_ON(!pd->used_pdes);
+		pd->used_pdes--;
+
+		free_pt(vm, pt);
 	}
 
-	if (bitmap_empty(pd->used_pdes, I915_PDES))
-		return true;
+	return !pd->used_pdes;
+}
 
-	return false;
+static void gen8_ppgtt_set_pdpe(struct i915_address_space *vm,
+				struct i915_page_directory_pointer *pdp,
+				struct i915_page_directory *pd,
+				unsigned int pdpe)
+{
+	gen8_ppgtt_pdpe_t *vaddr;
+
+	pdp->page_directory[pdpe] = pd;
+	if (!use_4lvl(vm))
+		return;
+
+	vaddr = kmap_atomic_px(pdp);
+	vaddr[pdpe] = gen8_pdpe_encode(px_dma(pd), I915_CACHE_LLC);
+	kunmap_atomic(vaddr);
 }
 
 /* Removes entries from a single page dir pointer, releasing it if it's empty.
@@ -809,138 +759,189 @@ static bool gen8_ppgtt_clear_pd(struct i915_address_space *vm,
  */
 static bool gen8_ppgtt_clear_pdp(struct i915_address_space *vm,
 				 struct i915_page_directory_pointer *pdp,
-				 uint64_t start,
-				 uint64_t length)
+				 u64 start, u64 length)
 {
-	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
 	struct i915_page_directory *pd;
-	uint64_t pdpe;
+	unsigned int pdpe;
 
 	gen8_for_each_pdpe(pd, pdp, start, length, pdpe) {
-		if (WARN_ON(!pdp->page_directory[pdpe]))
-			break;
+		GEM_BUG_ON(pd == vm->scratch_pd);
 
-		if (gen8_ppgtt_clear_pd(vm, pd, start, length)) {
-			__clear_bit(pdpe, pdp->used_pdpes);
-			gen8_setup_pdpe(ppgtt, pdp, vm->scratch_pd, pdpe);
-			free_pd(vm->i915, pd);
-		}
+		if (!gen8_ppgtt_clear_pd(vm, pd, start, length))
+			continue;
+
+		gen8_ppgtt_set_pdpe(vm, pdp, vm->scratch_pd, pdpe);
+		GEM_BUG_ON(!pdp->used_pdpes);
+		pdp->used_pdpes--;
+
+		free_pd(vm, pd);
 	}
 
-	mark_tlbs_dirty(ppgtt);
+	return !pdp->used_pdpes;
+}
 
-	if (bitmap_empty(pdp->used_pdpes, I915_PDPES_PER_PDP(dev_priv)))
-		return true;
+static void gen8_ppgtt_clear_3lvl(struct i915_address_space *vm,
+				  u64 start, u64 length)
+{
+	gen8_ppgtt_clear_pdp(vm, &i915_vm_to_ppgtt(vm)->pdp, start, length);
+}
 
-	return false;
+static void gen8_ppgtt_set_pml4e(struct i915_pml4 *pml4,
+				 struct i915_page_directory_pointer *pdp,
+				 unsigned int pml4e)
+{
+	gen8_ppgtt_pml4e_t *vaddr;
+
+	pml4->pdps[pml4e] = pdp;
+
+	vaddr = kmap_atomic_px(pml4);
+	vaddr[pml4e] = gen8_pml4e_encode(px_dma(pdp), I915_CACHE_LLC);
+	kunmap_atomic(vaddr);
 }
 
 /* Removes entries from a single pml4.
  * This is the top-level structure in 4-level page tables used on gen8+.
  * Empty entries are always scratch pml4e.
  */
-static void gen8_ppgtt_clear_pml4(struct i915_address_space *vm,
-				  struct i915_pml4 *pml4,
-				  uint64_t start,
-				  uint64_t length)
+static void gen8_ppgtt_clear_4lvl(struct i915_address_space *vm,
+				  u64 start, u64 length)
 {
 	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
+	struct i915_pml4 *pml4 = &ppgtt->pml4;
 	struct i915_page_directory_pointer *pdp;
-	uint64_t pml4e;
+	unsigned int pml4e;
 
-	GEM_BUG_ON(!USES_FULL_48BIT_PPGTT(vm->i915));
+	GEM_BUG_ON(!use_4lvl(vm));
 
 	gen8_for_each_pml4e(pdp, pml4, start, length, pml4e) {
-		if (WARN_ON(!pml4->pdps[pml4e]))
-			break;
+		GEM_BUG_ON(pdp == vm->scratch_pdp);
 
-		if (gen8_ppgtt_clear_pdp(vm, pdp, start, length)) {
-			__clear_bit(pml4e, pml4->used_pml4es);
-			gen8_setup_pml4e(ppgtt, pml4, vm->scratch_pdp, pml4e);
-			free_pdp(vm->i915, pdp);
-		}
+		if (!gen8_ppgtt_clear_pdp(vm, pdp, start, length))
+			continue;
+
+		gen8_ppgtt_set_pml4e(pml4, vm->scratch_pdp, pml4e);
+
+		free_pdp(vm, pdp);
 	}
 }
 
-static void gen8_ppgtt_clear_range(struct i915_address_space *vm,
-				   uint64_t start, uint64_t length)
-{
-	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
+struct sgt_dma {
+	struct scatterlist *sg;
+	dma_addr_t dma, max;
+};
 
-	if (USES_FULL_48BIT_PPGTT(vm->i915))
-		gen8_ppgtt_clear_pml4(vm, &ppgtt->pml4, start, length);
-	else
-		gen8_ppgtt_clear_pdp(vm, &ppgtt->pdp, start, length);
+struct gen8_insert_pte {
+	u16 pml4e;
+	u16 pdpe;
+	u16 pde;
+	u16 pte;
+};
+
+static __always_inline struct gen8_insert_pte gen8_insert_pte(u64 start)
+{
+	return (struct gen8_insert_pte) {
+		 gen8_pml4e_index(start),
+		 gen8_pdpe_index(start),
+		 gen8_pde_index(start),
+		 gen8_pte_index(start),
+	};
 }
 
-static void
-gen8_ppgtt_insert_pte_entries(struct i915_address_space *vm,
+static __always_inline bool
+gen8_ppgtt_insert_pte_entries(struct i915_hw_ppgtt *ppgtt,
 			      struct i915_page_directory_pointer *pdp,
-			      struct sg_page_iter *sg_iter,
-			      uint64_t start,
+			      struct sgt_dma *iter,
+			      struct gen8_insert_pte *idx,
 			      enum i915_cache_level cache_level)
 {
-	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
-	gen8_pte_t *pt_vaddr;
-	unsigned pdpe = gen8_pdpe_index(start);
-	unsigned pde = gen8_pde_index(start);
-	unsigned pte = gen8_pte_index(start);
-
-	pt_vaddr = NULL;
-
-	while (__sg_page_iter_next(sg_iter)) {
-		if (pt_vaddr == NULL) {
-			struct i915_page_directory *pd = pdp->page_directory[pdpe];
-			struct i915_page_table *pt = pd->page_table[pde];
-			pt_vaddr = kmap_px(pt);
+	struct i915_page_directory *pd;
+	const gen8_pte_t pte_encode = gen8_pte_encode(0, cache_level);
+	gen8_pte_t *vaddr;
+	bool ret;
+
+	GEM_BUG_ON(idx->pdpe >= i915_pdpes_per_pdp(&ppgtt->base));
+	pd = pdp->page_directory[idx->pdpe];
+	vaddr = kmap_atomic_px(pd->page_table[idx->pde]);
+	do {
+		vaddr[idx->pte] = pte_encode | iter->dma;
+
+		iter->dma += PAGE_SIZE;
+		if (iter->dma >= iter->max) {
+			iter->sg = __sg_next(iter->sg);
+			if (!iter->sg) {
+				ret = false;
+				break;
+			}
+
+			iter->dma = sg_dma_address(iter->sg);
+			iter->max = iter->dma + iter->sg->length;
 		}
 
-		pt_vaddr[pte] =
-			gen8_pte_encode(sg_page_iter_dma_address(sg_iter),
-					cache_level);
-		if (++pte == GEN8_PTES) {
-			kunmap_px(ppgtt, pt_vaddr);
-			pt_vaddr = NULL;
-			if (++pde == I915_PDES) {
-				if (++pdpe == I915_PDPES_PER_PDP(vm->i915))
+		if (++idx->pte == GEN8_PTES) {
+			idx->pte = 0;
+
+			if (++idx->pde == I915_PDES) {
+				idx->pde = 0;
+
+				/* Limited by sg length for 3lvl */
+				if (++idx->pdpe == GEN8_PML4ES_PER_PML4) {
+					idx->pdpe = 0;
+					ret = true;
 					break;
-				pde = 0;
+				}
+
+				GEM_BUG_ON(idx->pdpe >= i915_pdpes_per_pdp(&ppgtt->base));
+				pd = pdp->page_directory[idx->pdpe];
 			}
-			pte = 0;
+
+			kunmap_atomic(vaddr);
+			vaddr = kmap_atomic_px(pd->page_table[idx->pde]);
 		}
-	}
+	} while (1);
+	kunmap_atomic(vaddr);
 
-	if (pt_vaddr)
-		kunmap_px(ppgtt, pt_vaddr);
+	return ret;
 }
 
-static void gen8_ppgtt_insert_entries(struct i915_address_space *vm,
-				      struct sg_table *pages,
-				      uint64_t start,
-				      enum i915_cache_level cache_level,
-				      u32 unused)
+static void gen8_ppgtt_insert_3lvl(struct i915_address_space *vm,
+				   struct sg_table *pages,
+				   u64 start,
+				   enum i915_cache_level cache_level,
+				   u32 unused)
 {
 	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
-	struct sg_page_iter sg_iter;
+	struct sgt_dma iter = {
+		.sg = pages->sgl,
+		.dma = sg_dma_address(iter.sg),
+		.max = iter.dma + iter.sg->length,
+	};
+	struct gen8_insert_pte idx = gen8_insert_pte(start);
 
-	__sg_page_iter_start(&sg_iter, pages->sgl, sg_nents(pages->sgl), 0);
+	gen8_ppgtt_insert_pte_entries(ppgtt, &ppgtt->pdp, &iter, &idx,
+				      cache_level);
+}
 
-	if (!USES_FULL_48BIT_PPGTT(vm->i915)) {
-		gen8_ppgtt_insert_pte_entries(vm, &ppgtt->pdp, &sg_iter, start,
-					      cache_level);
-	} else {
-		struct i915_page_directory_pointer *pdp;
-		uint64_t pml4e;
-		uint64_t length = (uint64_t)pages->orig_nents << PAGE_SHIFT;
+static void gen8_ppgtt_insert_4lvl(struct i915_address_space *vm,
+				   struct sg_table *pages,
+				   u64 start,
+				   enum i915_cache_level cache_level,
+				   u32 unused)
+{
+	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
+	struct sgt_dma iter = {
+		.sg = pages->sgl,
+		.dma = sg_dma_address(iter.sg),
+		.max = iter.dma + iter.sg->length,
+	};
+	struct i915_page_directory_pointer **pdps = ppgtt->pml4.pdps;
+	struct gen8_insert_pte idx = gen8_insert_pte(start);
 
-		gen8_for_each_pml4e(pdp, &ppgtt->pml4, start, length, pml4e) {
-			gen8_ppgtt_insert_pte_entries(vm, pdp, &sg_iter,
-						      start, cache_level);
-		}
-	}
+	while (gen8_ppgtt_insert_pte_entries(ppgtt, pdps[idx.pml4e++], &iter,
+					     &idx, cache_level))
+		GEM_BUG_ON(idx.pml4e >= GEN8_PML4ES_PER_PML4);
 }
 
-static void gen8_free_page_tables(struct drm_i915_private *dev_priv,
+static void gen8_free_page_tables(struct i915_address_space *vm,
 				  struct i915_page_directory *pd)
 {
 	int i;
@@ -948,38 +949,34 @@ static void gen8_free_page_tables(struct drm_i915_private *dev_priv,
 	if (!px_page(pd))
 		return;
 
-	for_each_set_bit(i, pd->used_pdes, I915_PDES) {
-		if (WARN_ON(!pd->page_table[i]))
-			continue;
-
-		free_pt(dev_priv, pd->page_table[i]);
-		pd->page_table[i] = NULL;
+	for (i = 0; i < I915_PDES; i++) {
+		if (pd->page_table[i] != vm->scratch_pt)
+			free_pt(vm, pd->page_table[i]);
 	}
 }
 
 static int gen8_init_scratch(struct i915_address_space *vm)
 {
-	struct drm_i915_private *dev_priv = vm->i915;
 	int ret;
 
-	ret = setup_scratch_page(dev_priv, &vm->scratch_page, I915_GFP_DMA);
+	ret = setup_scratch_page(vm, I915_GFP_DMA);
 	if (ret)
 		return ret;
 
-	vm->scratch_pt = alloc_pt(dev_priv);
+	vm->scratch_pt = alloc_pt(vm);
 	if (IS_ERR(vm->scratch_pt)) {
 		ret = PTR_ERR(vm->scratch_pt);
 		goto free_scratch_page;
 	}
 
-	vm->scratch_pd = alloc_pd(dev_priv);
+	vm->scratch_pd = alloc_pd(vm);
 	if (IS_ERR(vm->scratch_pd)) {
 		ret = PTR_ERR(vm->scratch_pd);
 		goto free_pt;
 	}
 
-	if (USES_FULL_48BIT_PPGTT(dev_priv)) {
-		vm->scratch_pdp = alloc_pdp(dev_priv);
+	if (use_4lvl(vm)) {
+		vm->scratch_pdp = alloc_pdp(vm);
 		if (IS_ERR(vm->scratch_pdp)) {
 			ret = PTR_ERR(vm->scratch_pdp);
 			goto free_pd;
@@ -988,29 +985,30 @@ static int gen8_init_scratch(struct i915_address_space *vm)
 
 	gen8_initialize_pt(vm, vm->scratch_pt);
 	gen8_initialize_pd(vm, vm->scratch_pd);
-	if (USES_FULL_48BIT_PPGTT(dev_priv))
+	if (use_4lvl(vm))
 		gen8_initialize_pdp(vm, vm->scratch_pdp);
 
 	return 0;
 
 free_pd:
-	free_pd(dev_priv, vm->scratch_pd);
+	free_pd(vm, vm->scratch_pd);
 free_pt:
-	free_pt(dev_priv, vm->scratch_pt);
+	free_pt(vm, vm->scratch_pt);
 free_scratch_page:
-	cleanup_scratch_page(dev_priv, &vm->scratch_page);
+	cleanup_scratch_page(vm);
 
 	return ret;
 }
 
 static int gen8_ppgtt_notify_vgt(struct i915_hw_ppgtt *ppgtt, bool create)
 {
+	struct i915_address_space *vm = &ppgtt->base;
+	struct drm_i915_private *dev_priv = vm->i915;
 	enum vgt_g2v_type msg;
-	struct drm_i915_private *dev_priv = ppgtt->base.i915;
 	int i;
 
-	if (USES_FULL_48BIT_PPGTT(dev_priv)) {
-		u64 daddr = px_dma(&ppgtt->pml4);
+	if (use_4lvl(vm)) {
+		const u64 daddr = px_dma(&ppgtt->pml4);
 
 		I915_WRITE(vgtif_reg(pdp[0].lo), lower_32_bits(daddr));
 		I915_WRITE(vgtif_reg(pdp[0].hi), upper_32_bits(daddr));
@@ -1018,8 +1016,8 @@ static int gen8_ppgtt_notify_vgt(struct i915_hw_ppgtt *ppgtt, bool create)
 		msg = (create ? VGT_G2V_PPGTT_L4_PAGE_TABLE_CREATE :
 				VGT_G2V_PPGTT_L4_PAGE_TABLE_DESTROY);
 	} else {
-		for (i = 0; i < GEN8_LEGACY_PDPES; i++) {
-			u64 daddr = i915_page_dir_dma_addr(ppgtt, i);
+		for (i = 0; i < GEN8_3LVL_PDPES; i++) {
+			const u64 daddr = i915_page_dir_dma_addr(ppgtt, i);
 
 			I915_WRITE(vgtif_reg(pdp[i].lo), lower_32_bits(daddr));
 			I915_WRITE(vgtif_reg(pdp[i].hi), upper_32_bits(daddr));
@@ -1036,44 +1034,42 @@ static int gen8_ppgtt_notify_vgt(struct i915_hw_ppgtt *ppgtt, bool create)
 
 static void gen8_free_scratch(struct i915_address_space *vm)
 {
-	struct drm_i915_private *dev_priv = vm->i915;
-
-	if (USES_FULL_48BIT_PPGTT(dev_priv))
-		free_pdp(dev_priv, vm->scratch_pdp);
-	free_pd(dev_priv, vm->scratch_pd);
-	free_pt(dev_priv, vm->scratch_pt);
-	cleanup_scratch_page(dev_priv, &vm->scratch_page);
+	if (use_4lvl(vm))
+		free_pdp(vm, vm->scratch_pdp);
+	free_pd(vm, vm->scratch_pd);
+	free_pt(vm, vm->scratch_pt);
+	cleanup_scratch_page(vm);
 }
 
-static void gen8_ppgtt_cleanup_3lvl(struct drm_i915_private *dev_priv,
+static void gen8_ppgtt_cleanup_3lvl(struct i915_address_space *vm,
 				    struct i915_page_directory_pointer *pdp)
 {
+	const unsigned int pdpes = i915_pdpes_per_pdp(vm);
 	int i;
 
-	for_each_set_bit(i, pdp->used_pdpes, I915_PDPES_PER_PDP(dev_priv)) {
-		if (WARN_ON(!pdp->page_directory[i]))
+	for (i = 0; i < pdpes; i++) {
+		if (pdp->page_directory[i] == vm->scratch_pd)
 			continue;
 
-		gen8_free_page_tables(dev_priv, pdp->page_directory[i]);
-		free_pd(dev_priv, pdp->page_directory[i]);
+		gen8_free_page_tables(vm, pdp->page_directory[i]);
+		free_pd(vm, pdp->page_directory[i]);
 	}
 
-	free_pdp(dev_priv, pdp);
+	free_pdp(vm, pdp);
 }
 
 static void gen8_ppgtt_cleanup_4lvl(struct i915_hw_ppgtt *ppgtt)
 {
-	struct drm_i915_private *dev_priv = ppgtt->base.i915;
 	int i;
 
-	for_each_set_bit(i, ppgtt->pml4.used_pml4es, GEN8_PML4ES_PER_PML4) {
-		if (WARN_ON(!ppgtt->pml4.pdps[i]))
+	for (i = 0; i < GEN8_PML4ES_PER_PML4; i++) {
+		if (ppgtt->pml4.pdps[i] == ppgtt->base.scratch_pdp)
 			continue;
 
-		gen8_ppgtt_cleanup_3lvl(dev_priv, ppgtt->pml4.pdps[i]);
+		gen8_ppgtt_cleanup_3lvl(&ppgtt->base, ppgtt->pml4.pdps[i]);
 	}
 
-	cleanup_px(dev_priv, &ppgtt->pml4);
+	cleanup_px(&ppgtt->base, &ppgtt->pml4);
 }
 
 static void gen8_ppgtt_cleanup(struct i915_address_space *vm)
@@ -1084,414 +1080,162 @@ static void gen8_ppgtt_cleanup(struct i915_address_space *vm)
 	if (intel_vgpu_active(dev_priv))
 		gen8_ppgtt_notify_vgt(ppgtt, false);
 
-	if (!USES_FULL_48BIT_PPGTT(dev_priv))
-		gen8_ppgtt_cleanup_3lvl(dev_priv, &ppgtt->pdp);
-	else
+	if (use_4lvl(vm))
 		gen8_ppgtt_cleanup_4lvl(ppgtt);
+	else
+		gen8_ppgtt_cleanup_3lvl(&ppgtt->base, &ppgtt->pdp);
 
 	gen8_free_scratch(vm);
 }
 
-/**
- * gen8_ppgtt_alloc_pagetabs() - Allocate page tables for VA range.
- * @vm:	Master vm structure.
- * @pd:	Page directory for this address range.
- * @start:	Starting virtual address to begin allocations.
- * @length:	Size of the allocations.
- * @new_pts:	Bitmap set by function with new allocations. Likely used by the
- *		caller to free on error.
- *
- * Allocate the required number of page tables. Extremely similar to
- * gen8_ppgtt_alloc_page_directories(). The main difference is here we are limited by
- * the page directory boundary (instead of the page directory pointer). That
- * boundary is 1GB virtual. Therefore, unlike gen8_ppgtt_alloc_page_directories(), it is
- * possible, and likely that the caller will need to use multiple calls of this
- * function to achieve the appropriate allocation.
- *
- * Return: 0 if success; negative error code otherwise.
- */
-static int gen8_ppgtt_alloc_pagetabs(struct i915_address_space *vm,
-				     struct i915_page_directory *pd,
-				     uint64_t start,
-				     uint64_t length,
-				     unsigned long *new_pts)
+static int gen8_ppgtt_alloc_pd(struct i915_address_space *vm,
+			       struct i915_page_directory *pd,
+			       u64 start, u64 length)
 {
-	struct drm_i915_private *dev_priv = vm->i915;
 	struct i915_page_table *pt;
-	uint32_t pde;
+	u64 from = start;
+	unsigned int pde;
 
 	gen8_for_each_pde(pt, pd, start, length, pde) {
-		/* Don't reallocate page tables */
-		if (test_bit(pde, pd->used_pdes)) {
-			/* Scratch is never allocated this way */
-			WARN_ON(pt == vm->scratch_pt);
-			continue;
-		}
-
-		pt = alloc_pt(dev_priv);
-		if (IS_ERR(pt))
-			goto unwind_out;
-
-		gen8_initialize_pt(vm, pt);
-		pd->page_table[pde] = pt;
-		__set_bit(pde, new_pts);
-		trace_i915_page_table_entry_alloc(vm, pde, start, GEN8_PDE_SHIFT);
-	}
-
-	return 0;
-
-unwind_out:
-	for_each_set_bit(pde, new_pts, I915_PDES)
-		free_pt(dev_priv, pd->page_table[pde]);
-
-	return -ENOMEM;
-}
-
-/**
- * gen8_ppgtt_alloc_page_directories() - Allocate page directories for VA range.
- * @vm:	Master vm structure.
- * @pdp:	Page directory pointer for this address range.
- * @start:	Starting virtual address to begin allocations.
- * @length:	Size of the allocations.
- * @new_pds:	Bitmap set by function with new allocations. Likely used by the
- *		caller to free on error.
- *
- * Allocate the required number of page directories starting at the pde index of
- * @start, and ending at the pde index @start + @length. This function will skip
- * over already allocated page directories within the range, and only allocate
- * new ones, setting the appropriate pointer within the pdp as well as the
- * correct position in the bitmap @new_pds.
- *
- * The function will only allocate the pages within the range for a give page
- * directory pointer. In other words, if @start + @length straddles a virtually
- * addressed PDP boundary (512GB for 4k pages), there will be more allocations
- * required by the caller, This is not currently possible, and the BUG in the
- * code will prevent it.
- *
- * Return: 0 if success; negative error code otherwise.
- */
-static int
-gen8_ppgtt_alloc_page_directories(struct i915_address_space *vm,
-				  struct i915_page_directory_pointer *pdp,
-				  uint64_t start,
-				  uint64_t length,
-				  unsigned long *new_pds)
-{
-	struct drm_i915_private *dev_priv = vm->i915;
-	struct i915_page_directory *pd;
-	uint32_t pdpe;
-	uint32_t pdpes = I915_PDPES_PER_PDP(dev_priv);
-
-	WARN_ON(!bitmap_empty(new_pds, pdpes));
-
-	gen8_for_each_pdpe(pd, pdp, start, length, pdpe) {
-		if (test_bit(pdpe, pdp->used_pdpes))
-			continue;
-
-		pd = alloc_pd(dev_priv);
-		if (IS_ERR(pd))
-			goto unwind_out;
-
-		gen8_initialize_pd(vm, pd);
-		pdp->page_directory[pdpe] = pd;
-		__set_bit(pdpe, new_pds);
-		trace_i915_page_directory_entry_alloc(vm, pdpe, start, GEN8_PDPE_SHIFT);
-	}
+		if (pt == vm->scratch_pt) {
+			pt = alloc_pt(vm);
+			if (IS_ERR(pt))
+				goto unwind;
 
-	return 0;
-
-unwind_out:
-	for_each_set_bit(pdpe, new_pds, pdpes)
-		free_pd(dev_priv, pdp->page_directory[pdpe]);
-
-	return -ENOMEM;
-}
-
-/**
- * gen8_ppgtt_alloc_page_dirpointers() - Allocate pdps for VA range.
- * @vm:	Master vm structure.
- * @pml4:	Page map level 4 for this address range.
- * @start:	Starting virtual address to begin allocations.
- * @length:	Size of the allocations.
- * @new_pdps:	Bitmap set by function with new allocations. Likely used by the
- *		caller to free on error.
- *
- * Allocate the required number of page directory pointers. Extremely similar to
- * gen8_ppgtt_alloc_page_directories() and gen8_ppgtt_alloc_pagetabs().
- * The main difference is here we are limited by the pml4 boundary (instead of
- * the page directory pointer).
- *
- * Return: 0 if success; negative error code otherwise.
- */
-static int
-gen8_ppgtt_alloc_page_dirpointers(struct i915_address_space *vm,
-				  struct i915_pml4 *pml4,
-				  uint64_t start,
-				  uint64_t length,
-				  unsigned long *new_pdps)
-{
-	struct drm_i915_private *dev_priv = vm->i915;
-	struct i915_page_directory_pointer *pdp;
-	uint32_t pml4e;
-
-	WARN_ON(!bitmap_empty(new_pdps, GEN8_PML4ES_PER_PML4));
-
-	gen8_for_each_pml4e(pdp, pml4, start, length, pml4e) {
-		if (!test_bit(pml4e, pml4->used_pml4es)) {
-			pdp = alloc_pdp(dev_priv);
-			if (IS_ERR(pdp))
-				goto unwind_out;
+			gen8_initialize_pt(vm, pt);
 
-			gen8_initialize_pdp(vm, pdp);
-			pml4->pdps[pml4e] = pdp;
-			__set_bit(pml4e, new_pdps);
-			trace_i915_page_directory_pointer_entry_alloc(vm,
-								      pml4e,
-								      start,
-								      GEN8_PML4E_SHIFT);
+			gen8_ppgtt_set_pde(vm, pd, pt, pde);
+			pd->used_pdes++;
+			GEM_BUG_ON(pd->used_pdes > I915_PDES);
 		}
-	}
-
-	return 0;
-
-unwind_out:
-	for_each_set_bit(pml4e, new_pdps, GEN8_PML4ES_PER_PML4)
-		free_pdp(dev_priv, pml4->pdps[pml4e]);
-
-	return -ENOMEM;
-}
-
-static void
-free_gen8_temp_bitmaps(unsigned long *new_pds, unsigned long *new_pts)
-{
-	kfree(new_pts);
-	kfree(new_pds);
-}
-
-/* Fills in the page directory bitmap, and the array of page tables bitmap. Both
- * of these are based on the number of PDPEs in the system.
- */
-static
-int __must_check alloc_gen8_temp_bitmaps(unsigned long **new_pds,
-					 unsigned long **new_pts,
-					 uint32_t pdpes)
-{
-	unsigned long *pds;
-	unsigned long *pts;
-
-	pds = kcalloc(BITS_TO_LONGS(pdpes), sizeof(unsigned long), GFP_TEMPORARY);
-	if (!pds)
-		return -ENOMEM;
-
-	pts = kcalloc(pdpes, BITS_TO_LONGS(I915_PDES) * sizeof(unsigned long),
-		      GFP_TEMPORARY);
-	if (!pts)
-		goto err_out;
-
-	*new_pds = pds;
-	*new_pts = pts;
 
+		pt->used_ptes += gen8_pte_count(start, length);
+	}
 	return 0;
 
-err_out:
-	free_gen8_temp_bitmaps(pds, pts);
+unwind:
+	gen8_ppgtt_clear_pd(vm, pd, from, start - from);
 	return -ENOMEM;
 }
 
-static int gen8_alloc_va_range_3lvl(struct i915_address_space *vm,
-				    struct i915_page_directory_pointer *pdp,
-				    uint64_t start,
-				    uint64_t length)
+static int gen8_ppgtt_alloc_pdp(struct i915_address_space *vm,
+				struct i915_page_directory_pointer *pdp,
+				u64 start, u64 length)
 {
-	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
-	unsigned long *new_page_dirs, *new_page_tables;
-	struct drm_i915_private *dev_priv = vm->i915;
 	struct i915_page_directory *pd;
-	const uint64_t orig_start = start;
-	const uint64_t orig_length = length;
-	uint32_t pdpe;
-	uint32_t pdpes = I915_PDPES_PER_PDP(dev_priv);
+	u64 from = start;
+	unsigned int pdpe;
 	int ret;
 
-	ret = alloc_gen8_temp_bitmaps(&new_page_dirs, &new_page_tables, pdpes);
-	if (ret)
-		return ret;
-
-	/* Do the allocations first so we can easily bail out */
-	ret = gen8_ppgtt_alloc_page_directories(vm, pdp, start, length,
-						new_page_dirs);
-	if (ret) {
-		free_gen8_temp_bitmaps(new_page_dirs, new_page_tables);
-		return ret;
-	}
-
-	/* For every page directory referenced, allocate page tables */
 	gen8_for_each_pdpe(pd, pdp, start, length, pdpe) {
-		ret = gen8_ppgtt_alloc_pagetabs(vm, pd, start, length,
-						new_page_tables + pdpe * BITS_TO_LONGS(I915_PDES));
-		if (ret)
-			goto err_out;
-	}
+		if (pd == vm->scratch_pd) {
+			pd = alloc_pd(vm);
+			if (IS_ERR(pd))
+				goto unwind;
 
-	start = orig_start;
-	length = orig_length;
+			gen8_initialize_pd(vm, pd);
+			gen8_ppgtt_set_pdpe(vm, pdp, pd, pdpe);
+			pdp->used_pdpes++;
+			GEM_BUG_ON(pdp->used_pdpes > i915_pdpes_per_pdp(vm));
 
-	/* Allocations have completed successfully, so set the bitmaps, and do
-	 * the mappings. */
-	gen8_for_each_pdpe(pd, pdp, start, length, pdpe) {
-		gen8_pde_t *const page_directory = kmap_px(pd);
-		struct i915_page_table *pt;
-		uint64_t pd_len = length;
-		uint64_t pd_start = start;
-		uint32_t pde;
-
-		/* Every pd should be allocated, we just did that above. */
-		WARN_ON(!pd);
-
-		gen8_for_each_pde(pt, pd, pd_start, pd_len, pde) {
-			/* Same reasoning as pd */
-			WARN_ON(!pt);
-			WARN_ON(!pd_len);
-			WARN_ON(!gen8_pte_count(pd_start, pd_len));
-
-			/* Set our used ptes within the page table */
-			bitmap_set(pt->used_ptes,
-				   gen8_pte_index(pd_start),
-				   gen8_pte_count(pd_start, pd_len));
-
-			/* Our pde is now pointing to the pagetable, pt */
-			__set_bit(pde, pd->used_pdes);
-
-			/* Map the PDE to the page table */
-			page_directory[pde] = gen8_pde_encode(px_dma(pt),
-							      I915_CACHE_LLC);
-			trace_i915_page_table_entry_map(&ppgtt->base, pde, pt,
-							gen8_pte_index(start),
-							gen8_pte_count(start, length),
-							GEN8_PTES);
-
-			/* NB: We haven't yet mapped ptes to pages. At this
-			 * point we're still relying on insert_entries() */
+			mark_tlbs_dirty(i915_vm_to_ppgtt(vm));
 		}
 
-		kunmap_px(ppgtt, page_directory);
-		__set_bit(pdpe, pdp->used_pdpes);
-		gen8_setup_pdpe(ppgtt, pdp, pd, pdpe);
+		ret = gen8_ppgtt_alloc_pd(vm, pd, start, length);
+		if (unlikely(ret))
+			goto unwind_pd;
 	}
 
-	free_gen8_temp_bitmaps(new_page_dirs, new_page_tables);
-	mark_tlbs_dirty(ppgtt);
 	return 0;
 
-err_out:
-	while (pdpe--) {
-		unsigned long temp;
-
-		for_each_set_bit(temp, new_page_tables + pdpe *
-				BITS_TO_LONGS(I915_PDES), I915_PDES)
-			free_pt(dev_priv,
-				pdp->page_directory[pdpe]->page_table[temp]);
+unwind_pd:
+	if (!pd->used_pdes) {
+		gen8_ppgtt_set_pdpe(vm, pdp, vm->scratch_pd, pdpe);
+		GEM_BUG_ON(!pdp->used_pdpes);
+		pdp->used_pdpes--;
+		free_pd(vm, pd);
 	}
+unwind:
+	gen8_ppgtt_clear_pdp(vm, pdp, from, start - from);
+	return -ENOMEM;
+}
 
-	for_each_set_bit(pdpe, new_page_dirs, pdpes)
-		free_pd(dev_priv, pdp->page_directory[pdpe]);
-
-	free_gen8_temp_bitmaps(new_page_dirs, new_page_tables);
-	mark_tlbs_dirty(ppgtt);
-	return ret;
+static int gen8_ppgtt_alloc_3lvl(struct i915_address_space *vm,
+				 u64 start, u64 length)
+{
+	return gen8_ppgtt_alloc_pdp(vm,
+				    &i915_vm_to_ppgtt(vm)->pdp, start, length);
 }
 
-static int gen8_alloc_va_range_4lvl(struct i915_address_space *vm,
-				    struct i915_pml4 *pml4,
-				    uint64_t start,
-				    uint64_t length)
+static int gen8_ppgtt_alloc_4lvl(struct i915_address_space *vm,
+				 u64 start, u64 length)
 {
-	DECLARE_BITMAP(new_pdps, GEN8_PML4ES_PER_PML4);
 	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
+	struct i915_pml4 *pml4 = &ppgtt->pml4;
 	struct i915_page_directory_pointer *pdp;
-	uint64_t pml4e;
-	int ret = 0;
-
-	/* Do the pml4 allocations first, so we don't need to track the newly
-	 * allocated tables below the pdp */
-	bitmap_zero(new_pdps, GEN8_PML4ES_PER_PML4);
-
-	/* The pagedirectory and pagetable allocations are done in the shared 3
-	 * and 4 level code. Just allocate the pdps.
-	 */
-	ret = gen8_ppgtt_alloc_page_dirpointers(vm, pml4, start, length,
-						new_pdps);
-	if (ret)
-		return ret;
-
-	WARN(bitmap_weight(new_pdps, GEN8_PML4ES_PER_PML4) > 2,
-	     "The allocation has spanned more than 512GB. "
-	     "It is highly likely this is incorrect.");
+	u64 from = start;
+	u32 pml4e;
+	int ret;
 
 	gen8_for_each_pml4e(pdp, pml4, start, length, pml4e) {
-		WARN_ON(!pdp);
+		if (pml4->pdps[pml4e] == vm->scratch_pdp) {
+			pdp = alloc_pdp(vm);
+			if (IS_ERR(pdp))
+				goto unwind;
 
-		ret = gen8_alloc_va_range_3lvl(vm, pdp, start, length);
-		if (ret)
-			goto err_out;
+			gen8_initialize_pdp(vm, pdp);
+			gen8_ppgtt_set_pml4e(pml4, pdp, pml4e);
+		}
 
-		gen8_setup_pml4e(ppgtt, pml4, pdp, pml4e);
+		ret = gen8_ppgtt_alloc_pdp(vm, pdp, start, length);
+		if (unlikely(ret))
+			goto unwind_pdp;
 	}
 
-	bitmap_or(pml4->used_pml4es, new_pdps, pml4->used_pml4es,
-		  GEN8_PML4ES_PER_PML4);
-
 	return 0;
 
-err_out:
-	for_each_set_bit(pml4e, new_pdps, GEN8_PML4ES_PER_PML4)
-		gen8_ppgtt_cleanup_3lvl(vm->i915, pml4->pdps[pml4e]);
-
-	return ret;
-}
-
-static int gen8_alloc_va_range(struct i915_address_space *vm,
-			       uint64_t start, uint64_t length)
-{
-	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
-
-	if (USES_FULL_48BIT_PPGTT(vm->i915))
-		return gen8_alloc_va_range_4lvl(vm, &ppgtt->pml4, start, length);
-	else
-		return gen8_alloc_va_range_3lvl(vm, &ppgtt->pdp, start, length);
+unwind_pdp:
+	if (!pdp->used_pdpes) {
+		gen8_ppgtt_set_pml4e(pml4, vm->scratch_pdp, pml4e);
+		free_pdp(vm, pdp);
+	}
+unwind:
+	gen8_ppgtt_clear_4lvl(vm, from, start - from);
+	return -ENOMEM;
 }
 
-static void gen8_dump_pdp(struct i915_page_directory_pointer *pdp,
-			  uint64_t start, uint64_t length,
+static void gen8_dump_pdp(struct i915_hw_ppgtt *ppgtt,
+			  struct i915_page_directory_pointer *pdp,
+			  u64 start, u64 length,
 			  gen8_pte_t scratch_pte,
 			  struct seq_file *m)
 {
+	struct i915_address_space *vm = &ppgtt->base;
 	struct i915_page_directory *pd;
-	uint32_t pdpe;
+	u32 pdpe;
 
 	gen8_for_each_pdpe(pd, pdp, start, length, pdpe) {
 		struct i915_page_table *pt;
-		uint64_t pd_len = length;
-		uint64_t pd_start = start;
-		uint32_t pde;
+		u64 pd_len = length;
+		u64 pd_start = start;
+		u32 pde;
 
-		if (!test_bit(pdpe, pdp->used_pdpes))
+		if (pdp->page_directory[pdpe] == ppgtt->base.scratch_pd)
 			continue;
 
 		seq_printf(m, "\tPDPE #%d\n", pdpe);
 		gen8_for_each_pde(pt, pd, pd_start, pd_len, pde) {
-			uint32_t  pte;
+			u32 pte;
 			gen8_pte_t *pt_vaddr;
 
-			if (!test_bit(pde, pd->used_pdes))
+			if (pd->page_table[pde] == ppgtt->base.scratch_pt)
 				continue;
 
-			pt_vaddr = kmap_px(pt);
+			pt_vaddr = kmap_atomic_px(pt);
 			for (pte = 0; pte < GEN8_PTES; pte += 4) {
-				uint64_t va =
-					(pdpe << GEN8_PDPE_SHIFT) |
-					(pde << GEN8_PDE_SHIFT) |
-					(pte << GEN8_PTE_SHIFT);
+				u64 va = (pdpe << GEN8_PDPE_SHIFT |
+					  pde << GEN8_PDE_SHIFT |
+					  pte << GEN8_PTE_SHIFT);
 				int i;
 				bool found = false;
 
@@ -1510,9 +1254,6 @@ static void gen8_dump_pdp(struct i915_page_directory_pointer *pdp,
 				}
 				seq_puts(m, "\n");
 			}
-			/* don't use kunmap_px, it could trigger
-			 * an unnecessary flush.
-			 */
 			kunmap_atomic(pt_vaddr);
 		}
 	}
@@ -1521,53 +1262,57 @@ static void gen8_dump_pdp(struct i915_page_directory_pointer *pdp,
 static void gen8_dump_ppgtt(struct i915_hw_ppgtt *ppgtt, struct seq_file *m)
 {
 	struct i915_address_space *vm = &ppgtt->base;
-	uint64_t start = ppgtt->base.start;
-	uint64_t length = ppgtt->base.total;
-	gen8_pte_t scratch_pte = gen8_pte_encode(vm->scratch_page.daddr,
-						 I915_CACHE_LLC);
+	const gen8_pte_t scratch_pte =
+		gen8_pte_encode(vm->scratch_page.daddr, I915_CACHE_LLC);
+	u64 start = 0, length = ppgtt->base.total;
 
-	if (!USES_FULL_48BIT_PPGTT(vm->i915)) {
-		gen8_dump_pdp(&ppgtt->pdp, start, length, scratch_pte, m);
-	} else {
-		uint64_t pml4e;
+	if (use_4lvl(vm)) {
+		u64 pml4e;
 		struct i915_pml4 *pml4 = &ppgtt->pml4;
 		struct i915_page_directory_pointer *pdp;
 
 		gen8_for_each_pml4e(pdp, pml4, start, length, pml4e) {
-			if (!test_bit(pml4e, pml4->used_pml4es))
+			if (pml4->pdps[pml4e] == ppgtt->base.scratch_pdp)
 				continue;
 
 			seq_printf(m, "    PML4E #%llu\n", pml4e);
-			gen8_dump_pdp(pdp, start, length, scratch_pte, m);
+			gen8_dump_pdp(ppgtt, pdp, start, length, scratch_pte, m);
 		}
+	} else {
+		gen8_dump_pdp(ppgtt, &ppgtt->pdp, start, length, scratch_pte, m);
 	}
 }
 
-static int gen8_preallocate_top_level_pdps(struct i915_hw_ppgtt *ppgtt)
+static int gen8_preallocate_top_level_pdp(struct i915_hw_ppgtt *ppgtt)
 {
-	unsigned long *new_page_dirs, *new_page_tables;
-	uint32_t pdpes = I915_PDPES_PER_PDP(to_i915(ppgtt->base.dev));
-	int ret;
+	struct i915_address_space *vm = &ppgtt->base;
+	struct i915_page_directory_pointer *pdp = &ppgtt->pdp;
+	struct i915_page_directory *pd;
+	u64 start = 0, length = ppgtt->base.total;
+	u64 from = start;
+	unsigned int pdpe;
 
-	/* We allocate temp bitmap for page tables for no gain
-	 * but as this is for init only, lets keep the things simple
-	 */
-	ret = alloc_gen8_temp_bitmaps(&new_page_dirs, &new_page_tables, pdpes);
-	if (ret)
-		return ret;
+	gen8_for_each_pdpe(pd, pdp, start, length, pdpe) {
+		pd = alloc_pd(vm);
+		if (IS_ERR(pd))
+			goto unwind;
 
-	/* Allocate for all pdps regardless of how the ppgtt
-	 * was defined.
-	 */
-	ret = gen8_ppgtt_alloc_page_directories(&ppgtt->base, &ppgtt->pdp,
-						0, 1ULL << 32,
-						new_page_dirs);
-	if (!ret)
-		*ppgtt->pdp.used_pdpes = *new_page_dirs;
+		gen8_initialize_pd(vm, pd);
+		gen8_ppgtt_set_pdpe(vm, pdp, pd, pdpe);
+		pdp->used_pdpes++;
+	}
 
-	free_gen8_temp_bitmaps(new_page_dirs, new_page_tables);
+	pdp->used_pdpes++; /* never remove */
+	return 0;
 
-	return ret;
+unwind:
+	start -= from;
+	gen8_for_each_pdpe(pd, pdp, from, start, pdpe) {
+		gen8_ppgtt_set_pdpe(vm, pdp, vm->scratch_pd, pdpe);
+		free_pd(vm, pd);
+	}
+	pdp->used_pdpes = 0;
+	return -ENOMEM;
 }
 
 /*
@@ -1579,52 +1324,64 @@ static int gen8_preallocate_top_level_pdps(struct i915_hw_ppgtt *ppgtt)
  */
 static int gen8_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 {
-	struct drm_i915_private *dev_priv = ppgtt->base.i915;
+	struct i915_address_space *vm = &ppgtt->base;
+	struct drm_i915_private *dev_priv = vm->i915;
 	int ret;
 
+	ppgtt->base.total = USES_FULL_48BIT_PPGTT(dev_priv) ?
+		1ULL << 48 :
+		1ULL << 32;
+
 	ret = gen8_init_scratch(&ppgtt->base);
-	if (ret)
+	if (ret) {
+		ppgtt->base.total = 0;
 		return ret;
+	}
 
-	ppgtt->base.start = 0;
-	ppgtt->base.cleanup = gen8_ppgtt_cleanup;
-	ppgtt->base.allocate_va_range = gen8_alloc_va_range;
-	ppgtt->base.insert_entries = gen8_ppgtt_insert_entries;
-	ppgtt->base.clear_range = gen8_ppgtt_clear_range;
-	ppgtt->base.unbind_vma = ppgtt_unbind_vma;
-	ppgtt->base.bind_vma = ppgtt_bind_vma;
-	ppgtt->debug_dump = gen8_dump_ppgtt;
+	/* There are only few exceptions for gen >=6. chv and bxt.
+	 * And we are not sure about the latter so play safe for now.
+	 */
+	if (IS_CHERRYVIEW(dev_priv) || IS_BROXTON(dev_priv))
+		ppgtt->base.pt_kmap_wc = true;
 
-	if (USES_FULL_48BIT_PPGTT(dev_priv)) {
-		ret = setup_px(dev_priv, &ppgtt->pml4);
+	if (use_4lvl(vm)) {
+		ret = setup_px(&ppgtt->base, &ppgtt->pml4);
 		if (ret)
 			goto free_scratch;
 
 		gen8_initialize_pml4(&ppgtt->base, &ppgtt->pml4);
 
-		ppgtt->base.total = 1ULL << 48;
-		ppgtt->switch_mm = gen8_48b_mm_switch;
+		ppgtt->switch_mm = gen8_mm_switch_4lvl;
+		ppgtt->base.allocate_va_range = gen8_ppgtt_alloc_4lvl;
+		ppgtt->base.insert_entries = gen8_ppgtt_insert_4lvl;
+		ppgtt->base.clear_range = gen8_ppgtt_clear_4lvl;
 	} else {
-		ret = __pdp_init(dev_priv, &ppgtt->pdp);
+		ret = __pdp_init(&ppgtt->base, &ppgtt->pdp);
 		if (ret)
 			goto free_scratch;
 
-		ppgtt->base.total = 1ULL << 32;
-		ppgtt->switch_mm = gen8_legacy_mm_switch;
-		trace_i915_page_directory_pointer_entry_alloc(&ppgtt->base,
-							      0, 0,
-							      GEN8_PML4E_SHIFT);
-
 		if (intel_vgpu_active(dev_priv)) {
-			ret = gen8_preallocate_top_level_pdps(ppgtt);
-			if (ret)
+			ret = gen8_preallocate_top_level_pdp(ppgtt);
+			if (ret) {
+				__pdp_fini(&ppgtt->pdp);
 				goto free_scratch;
+			}
 		}
+
+		ppgtt->switch_mm = gen8_mm_switch_3lvl;
+		ppgtt->base.allocate_va_range = gen8_ppgtt_alloc_3lvl;
+		ppgtt->base.insert_entries = gen8_ppgtt_insert_3lvl;
+		ppgtt->base.clear_range = gen8_ppgtt_clear_3lvl;
 	}
 
 	if (intel_vgpu_active(dev_priv))
 		gen8_ppgtt_notify_vgt(ppgtt, true);
 
+	ppgtt->base.cleanup = gen8_ppgtt_cleanup;
+	ppgtt->base.unbind_vma = ppgtt_unbind_vma;
+	ppgtt->base.bind_vma = ppgtt_bind_vma;
+	ppgtt->debug_dump = gen8_dump_ppgtt;
+
 	return 0;
 
 free_scratch:
@@ -1637,9 +1394,8 @@ static void gen6_dump_ppgtt(struct i915_hw_ppgtt *ppgtt, struct seq_file *m)
 	struct i915_address_space *vm = &ppgtt->base;
 	struct i915_page_table *unused;
 	gen6_pte_t scratch_pte;
-	uint32_t pd_entry;
-	uint32_t  pte, pde;
-	uint32_t start = ppgtt->base.start, length = ppgtt->base.total;
+	u32 pd_entry, pte, pde;
+	u32 start = 0, length = ppgtt->base.total;
 
 	scratch_pte = vm->pte_encode(vm->scratch_page.daddr,
 				     I915_CACHE_LLC, 0);
@@ -1658,7 +1414,7 @@ static void gen6_dump_ppgtt(struct i915_hw_ppgtt *ppgtt, struct seq_file *m)
 				   expected);
 		seq_printf(m, "\tPDE: %x\n", pd_entry);
 
-		pt_vaddr = kmap_px(ppgtt->pd.page_table[pde]);
+		pt_vaddr = kmap_atomic_px(ppgtt->pd.page_table[pde]);
 
 		for (pte = 0; pte < GEN6_PTES; pte+=4) {
 			unsigned long va =
@@ -1681,73 +1437,59 @@ static void gen6_dump_ppgtt(struct i915_hw_ppgtt *ppgtt, struct seq_file *m)
 			}
 			seq_puts(m, "\n");
 		}
-		kunmap_px(ppgtt, pt_vaddr);
+		kunmap_atomic(pt_vaddr);
 	}
 }
 
 /* Write pde (index) from the page directory @pd to the page table @pt */
-static void gen6_write_pde(struct i915_page_directory *pd,
-			    const int pde, struct i915_page_table *pt)
+static inline void gen6_write_pde(const struct i915_hw_ppgtt *ppgtt,
+				  const unsigned int pde,
+				  const struct i915_page_table *pt)
 {
 	/* Caller needs to make sure the write completes if necessary */
-	struct i915_hw_ppgtt *ppgtt =
-		container_of(pd, struct i915_hw_ppgtt, pd);
-	u32 pd_entry;
-
-	pd_entry = GEN6_PDE_ADDR_ENCODE(px_dma(pt));
-	pd_entry |= GEN6_PDE_VALID;
-
-	writel(pd_entry, ppgtt->pd_addr + pde);
+	writel_relaxed(GEN6_PDE_ADDR_ENCODE(px_dma(pt)) | GEN6_PDE_VALID,
+		       ppgtt->pd_addr + pde);
 }
 
 /* Write all the page tables found in the ppgtt structure to incrementing page
  * directories. */
-static void gen6_write_page_range(struct drm_i915_private *dev_priv,
-				  struct i915_page_directory *pd,
-				  uint32_t start, uint32_t length)
+static void gen6_write_page_range(struct i915_hw_ppgtt *ppgtt,
+				  u32 start, u32 length)
 {
-	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	struct i915_page_table *pt;
-	uint32_t pde;
+	unsigned int pde;
 
-	gen6_for_each_pde(pt, pd, start, length, pde)
-		gen6_write_pde(pd, pde, pt);
+	gen6_for_each_pde(pt, &ppgtt->pd, start, length, pde)
+		gen6_write_pde(ppgtt, pde, pt);
 
-	/* Make sure write is complete before other code can use this page
-	 * table. Also require for WC mapped PTEs */
-	readl(ggtt->gsm);
+	mark_tlbs_dirty(ppgtt);
+	wmb();
 }
 
-static uint32_t get_pd_offset(struct i915_hw_ppgtt *ppgtt)
+static inline u32 get_pd_offset(struct i915_hw_ppgtt *ppgtt)
 {
-	BUG_ON(ppgtt->pd.base.ggtt_offset & 0x3f);
-
-	return (ppgtt->pd.base.ggtt_offset / 64) << 16;
+	GEM_BUG_ON(ppgtt->pd.base.ggtt_offset & 0x3f);
+	return ppgtt->pd.base.ggtt_offset << 10;
 }
 
 static int hsw_mm_switch(struct i915_hw_ppgtt *ppgtt,
 			 struct drm_i915_gem_request *req)
 {
-	struct intel_ring *ring = req->ring;
 	struct intel_engine_cs *engine = req->engine;
-	int ret;
+	u32 *cs;
 
 	/* NB: TLBs must be flushed and invalidated before a switch */
-	ret = engine->emit_flush(req, EMIT_INVALIDATE | EMIT_FLUSH);
-	if (ret)
-		return ret;
-
-	ret = intel_ring_begin(req, 6);
-	if (ret)
-		return ret;
-
-	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(2));
-	intel_ring_emit_reg(ring, RING_PP_DIR_DCLV(engine));
-	intel_ring_emit(ring, PP_DIR_DCLV_2G);
-	intel_ring_emit_reg(ring, RING_PP_DIR_BASE(engine));
-	intel_ring_emit(ring, get_pd_offset(ppgtt));
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	cs = intel_ring_begin(req, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_LOAD_REGISTER_IMM(2);
+	*cs++ = i915_mmio_reg_offset(RING_PP_DIR_DCLV(engine));
+	*cs++ = PP_DIR_DCLV_2G;
+	*cs++ = i915_mmio_reg_offset(RING_PP_DIR_BASE(engine));
+	*cs++ = get_pd_offset(ppgtt);
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -1755,33 +1497,21 @@ static int hsw_mm_switch(struct i915_hw_ppgtt *ppgtt,
 static int gen7_mm_switch(struct i915_hw_ppgtt *ppgtt,
 			  struct drm_i915_gem_request *req)
 {
-	struct intel_ring *ring = req->ring;
 	struct intel_engine_cs *engine = req->engine;
-	int ret;
+	u32 *cs;
 
 	/* NB: TLBs must be flushed and invalidated before a switch */
-	ret = engine->emit_flush(req, EMIT_INVALIDATE | EMIT_FLUSH);
-	if (ret)
-		return ret;
-
-	ret = intel_ring_begin(req, 6);
-	if (ret)
-		return ret;
-
-	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(2));
-	intel_ring_emit_reg(ring, RING_PP_DIR_DCLV(engine));
-	intel_ring_emit(ring, PP_DIR_DCLV_2G);
-	intel_ring_emit_reg(ring, RING_PP_DIR_BASE(engine));
-	intel_ring_emit(ring, get_pd_offset(ppgtt));
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
-
-	/* XXX: RCS is the only one to auto invalidate the TLBs? */
-	if (engine->id != RCS) {
-		ret = engine->emit_flush(req, EMIT_INVALIDATE | EMIT_FLUSH);
-		if (ret)
-			return ret;
-	}
+	cs = intel_ring_begin(req, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_LOAD_REGISTER_IMM(2);
+	*cs++ = i915_mmio_reg_offset(RING_PP_DIR_DCLV(engine));
+	*cs++ = PP_DIR_DCLV_2G;
+	*cs++ = i915_mmio_reg_offset(RING_PP_DIR_BASE(engine));
+	*cs++ = get_pd_offset(ppgtt);
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -1813,7 +1543,7 @@ static void gen8_ppgtt_enable(struct drm_i915_private *dev_priv)
 static void gen7_ppgtt_enable(struct drm_i915_private *dev_priv)
 {
 	struct intel_engine_cs *engine;
-	uint32_t ecochk, ecobits;
+	u32 ecochk, ecobits;
 	enum intel_engine_id id;
 
 	ecobits = I915_READ(GAC_ECO_BITS);
@@ -1837,7 +1567,7 @@ static void gen7_ppgtt_enable(struct drm_i915_private *dev_priv)
 
 static void gen6_ppgtt_enable(struct drm_i915_private *dev_priv)
 {
-	uint32_t ecochk, gab_ctl, ecobits;
+	u32 ecochk, gab_ctl, ecobits;
 
 	ecobits = I915_READ(GAC_ECO_BITS);
 	I915_WRITE(GAC_ECO_BITS, ecobits | ECOBITS_SNB_BIT |
@@ -1854,168 +1584,124 @@ static void gen6_ppgtt_enable(struct drm_i915_private *dev_priv)
 
 /* PPGTT support for Sandybdrige/Gen6 and later */
 static void gen6_ppgtt_clear_range(struct i915_address_space *vm,
-				   uint64_t start,
-				   uint64_t length)
+				   u64 start, u64 length)
 {
 	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
-	gen6_pte_t *pt_vaddr, scratch_pte;
-	unsigned first_entry = start >> PAGE_SHIFT;
-	unsigned num_entries = length >> PAGE_SHIFT;
-	unsigned act_pt = first_entry / GEN6_PTES;
-	unsigned first_pte = first_entry % GEN6_PTES;
-	unsigned last_pte, i;
-
-	scratch_pte = vm->pte_encode(vm->scratch_page.daddr,
-				     I915_CACHE_LLC, 0);
+	unsigned int first_entry = start >> PAGE_SHIFT;
+	unsigned int pde = first_entry / GEN6_PTES;
+	unsigned int pte = first_entry % GEN6_PTES;
+	unsigned int num_entries = length >> PAGE_SHIFT;
+	gen6_pte_t scratch_pte =
+		vm->pte_encode(vm->scratch_page.daddr, I915_CACHE_LLC, 0);
 
 	while (num_entries) {
-		last_pte = first_pte + num_entries;
-		if (last_pte > GEN6_PTES)
-			last_pte = GEN6_PTES;
+		struct i915_page_table *pt = ppgtt->pd.page_table[pde++];
+		unsigned int end = min(pte + num_entries, GEN6_PTES);
+		gen6_pte_t *vaddr;
 
-		pt_vaddr = kmap_px(ppgtt->pd.page_table[act_pt]);
+		num_entries -= end - pte;
 
-		for (i = first_pte; i < last_pte; i++)
-			pt_vaddr[i] = scratch_pte;
+		/* Note that the hw doesn't support removing PDE on the fly
+		 * (they are cached inside the context with no means to
+		 * invalidate the cache), so we can only reset the PTE
+		 * entries back to scratch.
+		 */
 
-		kunmap_px(ppgtt, pt_vaddr);
+		vaddr = kmap_atomic_px(pt);
+		do {
+			vaddr[pte++] = scratch_pte;
+		} while (pte < end);
+		kunmap_atomic(vaddr);
 
-		num_entries -= last_pte - first_pte;
-		first_pte = 0;
-		act_pt++;
+		pte = 0;
 	}
 }
 
 static void gen6_ppgtt_insert_entries(struct i915_address_space *vm,
 				      struct sg_table *pages,
-				      uint64_t start,
-				      enum i915_cache_level cache_level, u32 flags)
+				      u64 start,
+				      enum i915_cache_level cache_level,
+				      u32 flags)
 {
 	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
 	unsigned first_entry = start >> PAGE_SHIFT;
 	unsigned act_pt = first_entry / GEN6_PTES;
 	unsigned act_pte = first_entry % GEN6_PTES;
-	gen6_pte_t *pt_vaddr = NULL;
-	struct sgt_iter sgt_iter;
-	dma_addr_t addr;
+	const u32 pte_encode = vm->pte_encode(0, cache_level, flags);
+	struct sgt_dma iter;
+	gen6_pte_t *vaddr;
+
+	vaddr = kmap_atomic_px(ppgtt->pd.page_table[act_pt]);
+	iter.sg = pages->sgl;
+	iter.dma = sg_dma_address(iter.sg);
+	iter.max = iter.dma + iter.sg->length;
+	do {
+		vaddr[act_pte] = pte_encode | GEN6_PTE_ADDR_ENCODE(iter.dma);
 
-	for_each_sgt_dma(addr, sgt_iter, pages) {
-		if (pt_vaddr == NULL)
-			pt_vaddr = kmap_px(ppgtt->pd.page_table[act_pt]);
+		iter.dma += PAGE_SIZE;
+		if (iter.dma == iter.max) {
+			iter.sg = __sg_next(iter.sg);
+			if (!iter.sg)
+				break;
 
-		pt_vaddr[act_pte] =
-			vm->pte_encode(addr, cache_level, flags);
+			iter.dma = sg_dma_address(iter.sg);
+			iter.max = iter.dma + iter.sg->length;
+		}
 
 		if (++act_pte == GEN6_PTES) {
-			kunmap_px(ppgtt, pt_vaddr);
-			pt_vaddr = NULL;
-			act_pt++;
+			kunmap_atomic(vaddr);
+			vaddr = kmap_atomic_px(ppgtt->pd.page_table[++act_pt]);
 			act_pte = 0;
 		}
-	}
-
-	if (pt_vaddr)
-		kunmap_px(ppgtt, pt_vaddr);
+	} while (1);
+	kunmap_atomic(vaddr);
 }
 
 static int gen6_alloc_va_range(struct i915_address_space *vm,
-			       uint64_t start_in, uint64_t length_in)
+			       u64 start, u64 length)
 {
-	DECLARE_BITMAP(new_page_tables, I915_PDES);
-	struct drm_i915_private *dev_priv = vm->i915;
-	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
 	struct i915_page_table *pt;
-	uint32_t start, length, start_save, length_save;
-	uint32_t pde;
-	int ret;
+	u64 from = start;
+	unsigned int pde;
+	bool flush = false;
 
-	start = start_save = start_in;
-	length = length_save = length_in;
-
-	bitmap_zero(new_page_tables, I915_PDES);
-
-	/* The allocation is done in two stages so that we can bail out with
-	 * minimal amount of pain. The first stage finds new page tables that
-	 * need allocation. The second stage marks use ptes within the page
-	 * tables.
-	 */
 	gen6_for_each_pde(pt, &ppgtt->pd, start, length, pde) {
-		if (pt != vm->scratch_pt) {
-			WARN_ON(bitmap_empty(pt->used_ptes, GEN6_PTES));
-			continue;
-		}
-
-		/* We've already allocated a page table */
-		WARN_ON(!bitmap_empty(pt->used_ptes, GEN6_PTES));
+		if (pt == vm->scratch_pt) {
+			pt = alloc_pt(vm);
+			if (IS_ERR(pt))
+				goto unwind_out;
 
-		pt = alloc_pt(dev_priv);
-		if (IS_ERR(pt)) {
-			ret = PTR_ERR(pt);
-			goto unwind_out;
+			gen6_initialize_pt(vm, pt);
+			ppgtt->pd.page_table[pde] = pt;
+			gen6_write_pde(ppgtt, pde, pt);
+			flush = true;
 		}
-
-		gen6_initialize_pt(vm, pt);
-
-		ppgtt->pd.page_table[pde] = pt;
-		__set_bit(pde, new_page_tables);
-		trace_i915_page_table_entry_alloc(vm, pde, start, GEN6_PDE_SHIFT);
 	}
 
-	start = start_save;
-	length = length_save;
-
-	gen6_for_each_pde(pt, &ppgtt->pd, start, length, pde) {
-		DECLARE_BITMAP(tmp_bitmap, GEN6_PTES);
-
-		bitmap_zero(tmp_bitmap, GEN6_PTES);
-		bitmap_set(tmp_bitmap, gen6_pte_index(start),
-			   gen6_pte_count(start, length));
-
-		if (__test_and_clear_bit(pde, new_page_tables))
-			gen6_write_pde(&ppgtt->pd, pde, pt);
-
-		trace_i915_page_table_entry_map(vm, pde, pt,
-					 gen6_pte_index(start),
-					 gen6_pte_count(start, length),
-					 GEN6_PTES);
-		bitmap_or(pt->used_ptes, tmp_bitmap, pt->used_ptes,
-				GEN6_PTES);
+	if (flush) {
+		mark_tlbs_dirty(ppgtt);
+		wmb();
 	}
 
-	WARN_ON(!bitmap_empty(new_page_tables, I915_PDES));
-
-	/* Make sure write is complete before other code can use this page
-	 * table. Also require for WC mapped PTEs */
-	readl(ggtt->gsm);
-
-	mark_tlbs_dirty(ppgtt);
 	return 0;
 
 unwind_out:
-	for_each_set_bit(pde, new_page_tables, I915_PDES) {
-		struct i915_page_table *pt = ppgtt->pd.page_table[pde];
-
-		ppgtt->pd.page_table[pde] = vm->scratch_pt;
-		free_pt(dev_priv, pt);
-	}
-
-	mark_tlbs_dirty(ppgtt);
-	return ret;
+	gen6_ppgtt_clear_range(vm, from, start);
+	return -ENOMEM;
 }
 
 static int gen6_init_scratch(struct i915_address_space *vm)
 {
-	struct drm_i915_private *dev_priv = vm->i915;
 	int ret;
 
-	ret = setup_scratch_page(dev_priv, &vm->scratch_page, I915_GFP_DMA);
+	ret = setup_scratch_page(vm, I915_GFP_DMA);
 	if (ret)
 		return ret;
 
-	vm->scratch_pt = alloc_pt(dev_priv);
+	vm->scratch_pt = alloc_pt(vm);
 	if (IS_ERR(vm->scratch_pt)) {
-		cleanup_scratch_page(dev_priv, &vm->scratch_page);
+		cleanup_scratch_page(vm);
 		return PTR_ERR(vm->scratch_pt);
 	}
 
@@ -2026,25 +1712,22 @@ static int gen6_init_scratch(struct i915_address_space *vm)
 
 static void gen6_free_scratch(struct i915_address_space *vm)
 {
-	struct drm_i915_private *dev_priv = vm->i915;
-
-	free_pt(dev_priv, vm->scratch_pt);
-	cleanup_scratch_page(dev_priv, &vm->scratch_page);
+	free_pt(vm, vm->scratch_pt);
+	cleanup_scratch_page(vm);
 }
 
 static void gen6_ppgtt_cleanup(struct i915_address_space *vm)
 {
 	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
 	struct i915_page_directory *pd = &ppgtt->pd;
-	struct drm_i915_private *dev_priv = vm->i915;
 	struct i915_page_table *pt;
-	uint32_t pde;
+	u32 pde;
 
 	drm_mm_remove_node(&ppgtt->node);
 
 	gen6_for_all_pdes(pt, pd, pde)
 		if (pt != vm->scratch_pt)
-			free_pt(dev_priv, pt);
+			free_pt(vm, pt);
 
 	gen6_free_scratch(vm);
 }
@@ -2077,6 +1760,12 @@ static int gen6_ppgtt_allocate_page_directories(struct i915_hw_ppgtt *ppgtt)
 	if (ppgtt->node.start < ggtt->mappable_end)
 		DRM_DEBUG("Forced to use aperture for PDEs\n");
 
+	ppgtt->pd.base.ggtt_offset =
+		ppgtt->node.start / PAGE_SIZE * sizeof(gen6_pte_t);
+
+	ppgtt->pd_addr = (gen6_pte_t __iomem *)ggtt->gsm +
+		ppgtt->pd.base.ggtt_offset / sizeof(gen6_pte_t);
+
 	return 0;
 
 err_out:
@@ -2090,10 +1779,10 @@ static int gen6_ppgtt_alloc(struct i915_hw_ppgtt *ppgtt)
 }
 
 static void gen6_scratch_va_range(struct i915_hw_ppgtt *ppgtt,
-				  uint64_t start, uint64_t length)
+				  u64 start, u64 length)
 {
 	struct i915_page_table *unused;
-	uint32_t pde;
+	u32 pde;
 
 	gen6_for_each_pde(unused, &ppgtt->pd, start, length, pde)
 		ppgtt->pd.page_table[pde] = ppgtt->base.scratch_pt;
@@ -2119,32 +1808,30 @@ static int gen6_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 	if (ret)
 		return ret;
 
-	ppgtt->base.allocate_va_range = gen6_alloc_va_range;
+	ppgtt->base.total = I915_PDES * GEN6_PTES * PAGE_SIZE;
+
+	gen6_scratch_va_range(ppgtt, 0, ppgtt->base.total);
+	gen6_write_page_range(ppgtt, 0, ppgtt->base.total);
+
+	ret = gen6_alloc_va_range(&ppgtt->base, 0, ppgtt->base.total);
+	if (ret) {
+		gen6_ppgtt_cleanup(&ppgtt->base);
+		return ret;
+	}
+
 	ppgtt->base.clear_range = gen6_ppgtt_clear_range;
 	ppgtt->base.insert_entries = gen6_ppgtt_insert_entries;
 	ppgtt->base.unbind_vma = ppgtt_unbind_vma;
 	ppgtt->base.bind_vma = ppgtt_bind_vma;
 	ppgtt->base.cleanup = gen6_ppgtt_cleanup;
-	ppgtt->base.start = 0;
-	ppgtt->base.total = I915_PDES * GEN6_PTES * PAGE_SIZE;
 	ppgtt->debug_dump = gen6_dump_ppgtt;
 
-	ppgtt->pd.base.ggtt_offset =
-		ppgtt->node.start / PAGE_SIZE * sizeof(gen6_pte_t);
-
-	ppgtt->pd_addr = (gen6_pte_t __iomem *)ggtt->gsm +
-		ppgtt->pd.base.ggtt_offset / sizeof(gen6_pte_t);
-
-	gen6_scratch_va_range(ppgtt, 0, ppgtt->base.total);
-
-	gen6_write_page_range(dev_priv, &ppgtt->pd, 0, ppgtt->base.total);
-
 	DRM_DEBUG_DRIVER("Allocated pde space (%lldM) at GTT entry: %llx\n",
 			 ppgtt->node.size >> 20,
 			 ppgtt->node.start / PAGE_SIZE);
 
-	DRM_DEBUG("Adding PPGTT at offset %x\n",
-		  ppgtt->pd.base.ggtt_offset << 10);
+	DRM_DEBUG_DRIVER("Adding PPGTT at offset %x\n",
+			 ppgtt->pd.base.ggtt_offset << 10);
 
 	return 0;
 }
@@ -2153,6 +1840,7 @@ static int __hw_ppgtt_init(struct i915_hw_ppgtt *ppgtt,
 			   struct drm_i915_private *dev_priv)
 {
 	ppgtt->base.i915 = dev_priv;
+	ppgtt->base.dma = &dev_priv->drm.pdev->dev;
 
 	if (INTEL_INFO(dev_priv)->gen < 8)
 		return gen6_ppgtt_init(ppgtt);
@@ -2165,15 +1853,23 @@ static void i915_address_space_init(struct i915_address_space *vm,
 				    const char *name)
 {
 	i915_gem_timeline_init(dev_priv, &vm->timeline, name);
-	drm_mm_init(&vm->mm, vm->start, vm->total);
+
+	drm_mm_init(&vm->mm, 0, vm->total);
+	vm->mm.head_node.color = I915_COLOR_UNEVICTABLE;
+
 	INIT_LIST_HEAD(&vm->active_list);
 	INIT_LIST_HEAD(&vm->inactive_list);
 	INIT_LIST_HEAD(&vm->unbound_list);
+
 	list_add_tail(&vm->global_link, &dev_priv->vm_list);
+	pagevec_init(&vm->free_pages, false);
 }
 
 static void i915_address_space_fini(struct i915_address_space *vm)
 {
+	if (pagevec_count(&vm->free_pages))
+		vm_free_pages_release(vm);
+
 	i915_gem_timeline_fini(&vm->timeline);
 	drm_mm_takedown(&vm->mm);
 	list_del(&vm->global_link);
@@ -2185,34 +1881,17 @@ static void gtt_write_workarounds(struct drm_i915_private *dev_priv)
 	 * called on driver load and after a GPU reset, so you can place
 	 * workarounds here even if they get overwritten by GPU reset.
 	 */
-	/* WaIncreaseDefaultTLBEntries:chv,bdw,skl,bxt */
+	/* WaIncreaseDefaultTLBEntries:chv,bdw,skl,bxt,kbl,glk */
 	if (IS_BROADWELL(dev_priv))
 		I915_WRITE(GEN8_L3_LRA_1_GPGPU, GEN8_L3_LRA_1_GPGPU_DEFAULT_VALUE_BDW);
 	else if (IS_CHERRYVIEW(dev_priv))
 		I915_WRITE(GEN8_L3_LRA_1_GPGPU, GEN8_L3_LRA_1_GPGPU_DEFAULT_VALUE_CHV);
-	else if (IS_SKYLAKE(dev_priv))
+	else if (IS_GEN9_BC(dev_priv))
 		I915_WRITE(GEN8_L3_LRA_1_GPGPU, GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_SKL);
-	else if (IS_BROXTON(dev_priv))
+	else if (IS_GEN9_LP(dev_priv))
 		I915_WRITE(GEN8_L3_LRA_1_GPGPU, GEN9_L3_LRA_1_GPGPU_DEFAULT_VALUE_BXT);
 }
 
-static int i915_ppgtt_init(struct i915_hw_ppgtt *ppgtt,
-			   struct drm_i915_private *dev_priv,
-			   struct drm_i915_file_private *file_priv,
-			   const char *name)
-{
-	int ret;
-
-	ret = __hw_ppgtt_init(ppgtt, dev_priv);
-	if (ret == 0) {
-		kref_init(&ppgtt->ref);
-		i915_address_space_init(&ppgtt->base, dev_priv, name);
-		ppgtt->base.file = file_priv;
-	}
-
-	return ret;
-}
-
 int i915_ppgtt_init_hw(struct drm_i915_private *dev_priv)
 {
 	gtt_write_workarounds(dev_priv);
@@ -2250,12 +1929,16 @@ i915_ppgtt_create(struct drm_i915_private *dev_priv,
 	if (!ppgtt)
 		return ERR_PTR(-ENOMEM);
 
-	ret = i915_ppgtt_init(ppgtt, dev_priv, fpriv, name);
+	ret = __hw_ppgtt_init(ppgtt, dev_priv);
 	if (ret) {
 		kfree(ppgtt);
 		return ERR_PTR(ret);
 	}
 
+	kref_init(&ppgtt->ref);
+	i915_address_space_init(&ppgtt->base, dev_priv, name);
+	ppgtt->base.file = fpriv;
+
 	trace_i915_ppgtt_create(&ppgtt->base);
 
 	return ppgtt;
@@ -2294,9 +1977,8 @@ void i915_ppgtt_release(struct kref *kref)
 	WARN_ON(!list_empty(&ppgtt->base.inactive_list));
 	WARN_ON(!list_empty(&ppgtt->base.unbound_list));
 
-	i915_address_space_fini(&ppgtt->base);
-
 	ppgtt->base.cleanup(&ppgtt->base);
+	i915_address_space_fini(&ppgtt->base);
 	kfree(ppgtt);
 }
 
@@ -2358,7 +2040,7 @@ void i915_gem_suspend_gtt_mappings(struct drm_i915_private *dev_priv)
 
 	i915_check_and_clear_faults(dev_priv);
 
-	ggtt->base.clear_range(&ggtt->base, ggtt->base.start, ggtt->base.total);
+	ggtt->base.clear_range(&ggtt->base, 0, ggtt->base.total);
 
 	i915_ggtt_invalidate(dev_priv);
 }
@@ -2395,7 +2077,7 @@ static void gen8_set_pte(void __iomem *addr, gen8_pte_t pte)
 
 static void gen8_ggtt_insert_page(struct i915_address_space *vm,
 				  dma_addr_t addr,
-				  uint64_t offset,
+				  u64 offset,
 				  enum i915_cache_level level,
 				  u32 unused)
 {
@@ -2410,32 +2092,22 @@ static void gen8_ggtt_insert_page(struct i915_address_space *vm,
 
 static void gen8_ggtt_insert_entries(struct i915_address_space *vm,
 				     struct sg_table *st,
-				     uint64_t start,
-				     enum i915_cache_level level, u32 unused)
+				     u64 start,
+				     enum i915_cache_level level,
+				     u32 unused)
 {
 	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
 	struct sgt_iter sgt_iter;
 	gen8_pte_t __iomem *gtt_entries;
-	gen8_pte_t gtt_entry;
+	const gen8_pte_t pte_encode = gen8_pte_encode(0, level);
 	dma_addr_t addr;
-	int i = 0;
 
-	gtt_entries = (gen8_pte_t __iomem *)ggtt->gsm + (start >> PAGE_SHIFT);
+	gtt_entries = (gen8_pte_t __iomem *)ggtt->gsm;
+	gtt_entries += start >> PAGE_SHIFT;
+	for_each_sgt_dma(addr, sgt_iter, st)
+		gen8_set_pte(gtt_entries++, pte_encode | addr);
 
-	for_each_sgt_dma(addr, sgt_iter, st) {
-		gtt_entry = gen8_pte_encode(addr, level);
-		gen8_set_pte(&gtt_entries[i++], gtt_entry);
-	}
-
-	/*
-	 * XXX: This serves as a posting read to make sure that the PTE has
-	 * actually been updated. There is some concern that even though
-	 * registers and PTEs are within the same BAR that they are potentially
-	 * of NUMA access patterns. Therefore, even with the way we assume
-	 * hardware should work, we must keep this posting read for paranoia.
-	 */
-	if (i != 0)
-		WARN_ON(readq(&gtt_entries[i-1]) != gtt_entry);
+	wmb();
 
 	/* This next bit makes the above posting read even more important. We
 	 * want to flush the TLBs only after we're certain all the PTE updates
@@ -2444,35 +2116,9 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm,
 	ggtt->invalidate(vm->i915);
 }
 
-struct insert_entries {
-	struct i915_address_space *vm;
-	struct sg_table *st;
-	uint64_t start;
-	enum i915_cache_level level;
-	u32 flags;
-};
-
-static int gen8_ggtt_insert_entries__cb(void *_arg)
-{
-	struct insert_entries *arg = _arg;
-	gen8_ggtt_insert_entries(arg->vm, arg->st,
-				 arg->start, arg->level, arg->flags);
-	return 0;
-}
-
-static void gen8_ggtt_insert_entries__BKL(struct i915_address_space *vm,
-					  struct sg_table *st,
-					  uint64_t start,
-					  enum i915_cache_level level,
-					  u32 flags)
-{
-	struct insert_entries arg = { vm, st, start, level, flags };
-	stop_machine(gen8_ggtt_insert_entries__cb, &arg, NULL);
-}
-
 static void gen6_ggtt_insert_page(struct i915_address_space *vm,
 				  dma_addr_t addr,
-				  uint64_t offset,
+				  u64 offset,
 				  enum i915_cache_level level,
 				  u32 flags)
 {
@@ -2493,31 +2139,18 @@ static void gen6_ggtt_insert_page(struct i915_address_space *vm,
  */
 static void gen6_ggtt_insert_entries(struct i915_address_space *vm,
 				     struct sg_table *st,
-				     uint64_t start,
-				     enum i915_cache_level level, u32 flags)
+				     u64 start,
+				     enum i915_cache_level level,
+				     u32 flags)
 {
 	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
-	struct sgt_iter sgt_iter;
-	gen6_pte_t __iomem *gtt_entries;
-	gen6_pte_t gtt_entry;
+	gen6_pte_t __iomem *entries = (gen6_pte_t __iomem *)ggtt->gsm;
+	unsigned int i = start >> PAGE_SHIFT;
+	struct sgt_iter iter;
 	dma_addr_t addr;
-	int i = 0;
-
-	gtt_entries = (gen6_pte_t __iomem *)ggtt->gsm + (start >> PAGE_SHIFT);
-
-	for_each_sgt_dma(addr, sgt_iter, st) {
-		gtt_entry = vm->pte_encode(addr, level, flags);
-		iowrite32(gtt_entry, &gtt_entries[i++]);
-	}
-
-	/* XXX: This serves as a posting read to make sure that the PTE has
-	 * actually been updated. There is some concern that even though
-	 * registers and PTEs are within the same BAR that they are potentially
-	 * of NUMA access patterns. Therefore, even with the way we assume
-	 * hardware should work, we must keep this posting read for paranoia.
-	 */
-	if (i != 0)
-		WARN_ON(readl(&gtt_entries[i-1]) != gtt_entry);
+	for_each_sgt_dma(addr, iter, st)
+		iowrite32(vm->pte_encode(addr, level, flags), &entries[i++]);
+	wmb();
 
 	/* This next bit makes the above posting read even more important. We
 	 * want to flush the TLBs only after we're certain all the PTE updates
@@ -2527,17 +2160,19 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm,
 }
 
 static void nop_clear_range(struct i915_address_space *vm,
-			    uint64_t start, uint64_t length)
+			    u64 start, u64 length)
 {
 }
 
 static void gen8_ggtt_clear_range(struct i915_address_space *vm,
-				  uint64_t start, uint64_t length)
+				  u64 start, u64 length)
 {
 	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
 	unsigned first_entry = start >> PAGE_SHIFT;
 	unsigned num_entries = length >> PAGE_SHIFT;
-	gen8_pte_t scratch_pte, __iomem *gtt_base =
+	const gen8_pte_t scratch_pte =
+		gen8_pte_encode(vm->scratch_page.daddr, I915_CACHE_LLC);
+	gen8_pte_t __iomem *gtt_base =
 		(gen8_pte_t __iomem *)ggtt->gsm + first_entry;
 	const int max_entries = ggtt_total_entries(ggtt) - first_entry;
 	int i;
@@ -2547,16 +2182,12 @@ static void gen8_ggtt_clear_range(struct i915_address_space *vm,
 		 first_entry, num_entries, max_entries))
 		num_entries = max_entries;
 
-	scratch_pte = gen8_pte_encode(vm->scratch_page.daddr,
-				      I915_CACHE_LLC);
 	for (i = 0; i < num_entries; i++)
 		gen8_set_pte(&gtt_base[i], scratch_pte);
-	readl(gtt_base);
 }
 
 static void gen6_ggtt_clear_range(struct i915_address_space *vm,
-				  uint64_t start,
-				  uint64_t length)
+				  u64 start, u64 length)
 {
 	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
 	unsigned first_entry = start >> PAGE_SHIFT;
@@ -2576,12 +2207,11 @@ static void gen6_ggtt_clear_range(struct i915_address_space *vm,
 
 	for (i = 0; i < num_entries; i++)
 		iowrite32(scratch_pte, &gtt_base[i]);
-	readl(gtt_base);
 }
 
 static void i915_ggtt_insert_page(struct i915_address_space *vm,
 				  dma_addr_t addr,
-				  uint64_t offset,
+				  u64 offset,
 				  enum i915_cache_level cache_level,
 				  u32 unused)
 {
@@ -2593,19 +2223,18 @@ static void i915_ggtt_insert_page(struct i915_address_space *vm,
 
 static void i915_ggtt_insert_entries(struct i915_address_space *vm,
 				     struct sg_table *pages,
-				     uint64_t start,
-				     enum i915_cache_level cache_level, u32 unused)
+				     u64 start,
+				     enum i915_cache_level cache_level,
+				     u32 unused)
 {
 	unsigned int flags = (cache_level == I915_CACHE_NONE) ?
 		AGP_USER_MEMORY : AGP_USER_CACHED_MEMORY;
 
 	intel_gtt_insert_sg_entries(pages, start >> PAGE_SHIFT, flags);
-
 }
 
 static void i915_ggtt_clear_range(struct i915_address_space *vm,
-				  uint64_t start,
-				  uint64_t length)
+				  u64 start, u64 length)
 {
 	intel_gtt_clear_range(start >> PAGE_SHIFT, length >> PAGE_SHIFT);
 }
@@ -2616,14 +2245,16 @@ static int ggtt_bind_vma(struct i915_vma *vma,
 {
 	struct drm_i915_private *i915 = vma->vm->i915;
 	struct drm_i915_gem_object *obj = vma->obj;
-	u32 pte_flags = 0;
-	int ret;
+	u32 pte_flags;
 
-	ret = i915_get_ggtt_vma_pages(vma);
-	if (ret)
-		return ret;
+	if (unlikely(!vma->pages)) {
+		int ret = i915_get_ggtt_vma_pages(vma);
+		if (ret)
+			return ret;
+	}
 
 	/* Currently applicable only to VLV */
+	pte_flags = 0;
 	if (obj->gt_ro)
 		pte_flags |= PTE_READ_ONLY;
 
@@ -2642,6 +2273,15 @@ static int ggtt_bind_vma(struct i915_vma *vma,
 	return 0;
 }
 
+static void ggtt_unbind_vma(struct i915_vma *vma)
+{
+	struct drm_i915_private *i915 = vma->vm->i915;
+
+	intel_runtime_pm_get(i915);
+	vma->vm->clear_range(vma->vm, vma->node.start, vma->size);
+	intel_runtime_pm_put(i915);
+}
+
 static int aliasing_gtt_bind_vma(struct i915_vma *vma,
 				 enum i915_cache_level cache_level,
 				 u32 flags)
@@ -2650,15 +2290,32 @@ static int aliasing_gtt_bind_vma(struct i915_vma *vma,
 	u32 pte_flags;
 	int ret;
 
-	ret = i915_get_ggtt_vma_pages(vma);
-	if (ret)
-		return ret;
+	if (unlikely(!vma->pages)) {
+		ret = i915_get_ggtt_vma_pages(vma);
+		if (ret)
+			return ret;
+	}
 
 	/* Currently applicable only to VLV */
 	pte_flags = 0;
 	if (vma->obj->gt_ro)
 		pte_flags |= PTE_READ_ONLY;
 
+	if (flags & I915_VMA_LOCAL_BIND) {
+		struct i915_hw_ppgtt *appgtt = i915->mm.aliasing_ppgtt;
+
+		if (appgtt->base.allocate_va_range) {
+			ret = appgtt->base.allocate_va_range(&appgtt->base,
+							     vma->node.start,
+							     vma->node.size);
+			if (ret)
+				goto err_pages;
+		}
+
+		appgtt->base.insert_entries(&appgtt->base,
+					    vma->pages, vma->node.start,
+					    cache_level, pte_flags);
+	}
 
 	if (flags & I915_VMA_GLOBAL_BIND) {
 		intel_runtime_pm_get(i915);
@@ -2668,32 +2325,35 @@ static int aliasing_gtt_bind_vma(struct i915_vma *vma,
 		intel_runtime_pm_put(i915);
 	}
 
-	if (flags & I915_VMA_LOCAL_BIND) {
-		struct i915_hw_ppgtt *appgtt = i915->mm.aliasing_ppgtt;
-		appgtt->base.insert_entries(&appgtt->base,
-					    vma->pages, vma->node.start,
-					    cache_level, pte_flags);
-	}
-
 	return 0;
+
+err_pages:
+	if (!(vma->flags & (I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND))) {
+		if (vma->pages != vma->obj->mm.pages) {
+			GEM_BUG_ON(!vma->pages);
+			sg_free_table(vma->pages);
+			kfree(vma->pages);
+		}
+		vma->pages = NULL;
+	}
+	return ret;
 }
 
-static void ggtt_unbind_vma(struct i915_vma *vma)
+static void aliasing_gtt_unbind_vma(struct i915_vma *vma)
 {
 	struct drm_i915_private *i915 = vma->vm->i915;
-	struct i915_hw_ppgtt *appgtt = i915->mm.aliasing_ppgtt;
-	const u64 size = min(vma->size, vma->node.size);
 
 	if (vma->flags & I915_VMA_GLOBAL_BIND) {
 		intel_runtime_pm_get(i915);
-		vma->vm->clear_range(vma->vm,
-				     vma->node.start, size);
+		vma->vm->clear_range(vma->vm, vma->node.start, vma->size);
 		intel_runtime_pm_put(i915);
 	}
 
-	if (vma->flags & I915_VMA_LOCAL_BIND && appgtt)
-		appgtt->base.clear_range(&appgtt->base,
-					 vma->node.start, size);
+	if (vma->flags & I915_VMA_LOCAL_BIND) {
+		struct i915_address_space *vm = &i915->mm.aliasing_ppgtt->base;
+
+		vm->clear_range(vm, vma->node.start, vma->size);
+	}
 }
 
 void i915_gem_gtt_finish_pages(struct drm_i915_gem_object *obj,
@@ -2719,14 +2379,76 @@ static void i915_gtt_color_adjust(const struct drm_mm_node *node,
 				  u64 *start,
 				  u64 *end)
 {
-	if (node->color != color)
+	if (node->allocated && node->color != color)
 		*start += I915_GTT_PAGE_SIZE;
 
+	/* Also leave a space between the unallocated reserved node after the
+	 * GTT and any objects within the GTT, i.e. we use the color adjustment
+	 * to insert a guard page to prevent prefetches crossing over the
+	 * GTT boundary.
+	 */
 	node = list_next_entry(node, node_list);
-	if (node->allocated && node->color != color)
+	if (node->color != color)
 		*end -= I915_GTT_PAGE_SIZE;
 }
 
+int i915_gem_init_aliasing_ppgtt(struct drm_i915_private *i915)
+{
+	struct i915_ggtt *ggtt = &i915->ggtt;
+	struct i915_hw_ppgtt *ppgtt;
+	int err;
+
+	ppgtt = i915_ppgtt_create(i915, ERR_PTR(-EPERM), "[alias]");
+	if (IS_ERR(ppgtt))
+		return PTR_ERR(ppgtt);
+
+	if (WARN_ON(ppgtt->base.total < ggtt->base.total)) {
+		err = -ENODEV;
+		goto err_ppgtt;
+	}
+
+	if (ppgtt->base.allocate_va_range) {
+		/* Note we only pre-allocate as far as the end of the global
+		 * GTT. On 48b / 4-level page-tables, the difference is very,
+		 * very significant! We have to preallocate as GVT/vgpu does
+		 * not like the page directory disappearing.
+		 */
+		err = ppgtt->base.allocate_va_range(&ppgtt->base,
+						    0, ggtt->base.total);
+		if (err)
+			goto err_ppgtt;
+	}
+
+	i915->mm.aliasing_ppgtt = ppgtt;
+
+	WARN_ON(ggtt->base.bind_vma != ggtt_bind_vma);
+	ggtt->base.bind_vma = aliasing_gtt_bind_vma;
+
+	WARN_ON(ggtt->base.unbind_vma != ggtt_unbind_vma);
+	ggtt->base.unbind_vma = aliasing_gtt_unbind_vma;
+
+	return 0;
+
+err_ppgtt:
+	i915_ppgtt_put(ppgtt);
+	return err;
+}
+
+void i915_gem_fini_aliasing_ppgtt(struct drm_i915_private *i915)
+{
+	struct i915_ggtt *ggtt = &i915->ggtt;
+	struct i915_hw_ppgtt *ppgtt;
+
+	ppgtt = fetch_and_zero(&i915->mm.aliasing_ppgtt);
+	if (!ppgtt)
+		return;
+
+	i915_ppgtt_put(ppgtt);
+
+	ggtt->base.bind_vma = ggtt_bind_vma;
+	ggtt->base.unbind_vma = ggtt_unbind_vma;
+}
+
 int i915_gem_init_ggtt(struct drm_i915_private *dev_priv)
 {
 	/* Let GEM Manage all of the aperture.
@@ -2740,7 +2462,6 @@ int i915_gem_init_ggtt(struct drm_i915_private *dev_priv)
 	 */
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	unsigned long hole_start, hole_end;
-	struct i915_hw_ppgtt *ppgtt;
 	struct drm_mm_node *entry;
 	int ret;
 
@@ -2769,38 +2490,13 @@ int i915_gem_init_ggtt(struct drm_i915_private *dev_priv)
 			       ggtt->base.total - PAGE_SIZE, PAGE_SIZE);
 
 	if (USES_PPGTT(dev_priv) && !USES_FULL_PPGTT(dev_priv)) {
-		ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
-		if (!ppgtt) {
-			ret = -ENOMEM;
-			goto err;
-		}
-
-		ret = __hw_ppgtt_init(ppgtt, dev_priv);
+		ret = i915_gem_init_aliasing_ppgtt(dev_priv);
 		if (ret)
-			goto err_ppgtt;
-
-		if (ppgtt->base.allocate_va_range) {
-			ret = ppgtt->base.allocate_va_range(&ppgtt->base, 0,
-							    ppgtt->base.total);
-			if (ret)
-				goto err_ppgtt_cleanup;
-		}
-
-		ppgtt->base.clear_range(&ppgtt->base,
-					ppgtt->base.start,
-					ppgtt->base.total);
-
-		dev_priv->mm.aliasing_ppgtt = ppgtt;
-		WARN_ON(ggtt->base.bind_vma != ggtt_bind_vma);
-		ggtt->base.bind_vma = aliasing_gtt_bind_vma;
+			goto err;
 	}
 
 	return 0;
 
-err_ppgtt_cleanup:
-	ppgtt->base.cleanup(&ppgtt->base);
-err_ppgtt:
-	kfree(ppgtt);
 err:
 	drm_mm_remove_node(&ggtt->error_capture);
 	return ret;
@@ -2813,27 +2509,31 @@ err:
 void i915_ggtt_cleanup_hw(struct drm_i915_private *dev_priv)
 {
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
+	struct i915_vma *vma, *vn;
 
-	if (dev_priv->mm.aliasing_ppgtt) {
-		struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
-		ppgtt->base.cleanup(&ppgtt->base);
-		kfree(ppgtt);
-	}
+	ggtt->base.closed = true;
+
+	mutex_lock(&dev_priv->drm.struct_mutex);
+	WARN_ON(!list_empty(&ggtt->base.active_list));
+	list_for_each_entry_safe(vma, vn, &ggtt->base.inactive_list, vm_link)
+		WARN_ON(i915_vma_unbind(vma));
+	mutex_unlock(&dev_priv->drm.struct_mutex);
 
 	i915_gem_cleanup_stolen(&dev_priv->drm);
 
+	mutex_lock(&dev_priv->drm.struct_mutex);
+	i915_gem_fini_aliasing_ppgtt(dev_priv);
+
 	if (drm_mm_node_allocated(&ggtt->error_capture))
 		drm_mm_remove_node(&ggtt->error_capture);
 
 	if (drm_mm_initialized(&ggtt->base.mm)) {
 		intel_vgt_deballoon(dev_priv);
-
-		mutex_lock(&dev_priv->drm.struct_mutex);
 		i915_address_space_fini(&ggtt->base);
-		mutex_unlock(&dev_priv->drm.struct_mutex);
 	}
 
 	ggtt->base.cleanup(&ggtt->base);
+	mutex_unlock(&dev_priv->drm.struct_mutex);
 
 	arch_phys_wc_del(ggtt->mtrr);
 	io_mapping_fini(&ggtt->mappable);
@@ -2943,7 +2643,7 @@ static int ggtt_probe_common(struct i915_ggtt *ggtt, u64 size)
 		return -ENOMEM;
 	}
 
-	ret = setup_scratch_page(dev_priv, &ggtt->base.scratch_page, GFP_DMA32);
+	ret = setup_scratch_page(&ggtt->base, GFP_DMA32);
 	if (ret) {
 		DRM_ERROR("Scratch setup failed\n");
 		/* iounmap will also get called at remove, but meh */
@@ -2959,7 +2659,7 @@ static int ggtt_probe_common(struct i915_ggtt *ggtt, u64 size)
  * writing this data shouldn't be harmful even in those cases. */
 static void bdw_setup_private_ppat(struct drm_i915_private *dev_priv)
 {
-	uint64_t pat;
+	u64 pat;
 
 	pat = GEN8_PPAT(0, GEN8_PPAT_WB | GEN8_PPAT_LLC)     | /* for normal objects, no eLLC */
 	      GEN8_PPAT(1, GEN8_PPAT_WC | GEN8_PPAT_LLCELLC) | /* for something pointing to ptes? */
@@ -2994,7 +2694,7 @@ static void bdw_setup_private_ppat(struct drm_i915_private *dev_priv)
 
 static void chv_setup_private_ppat(struct drm_i915_private *dev_priv)
 {
-	uint64_t pat;
+	u64 pat;
 
 	/*
 	 * Map WB on BDW to snooped on CHV.
@@ -3032,7 +2732,7 @@ static void gen6_gmch_remove(struct i915_address_space *vm)
 	struct i915_ggtt *ggtt = i915_vm_to_ggtt(vm);
 
 	iounmap(ggtt->gsm);
-	cleanup_scratch_page(vm->i915, &vm->scratch_page);
+	cleanup_scratch_page(vm);
 }
 
 static int gen8_gmch_probe(struct i915_ggtt *ggtt)
@@ -3078,8 +2778,6 @@ static int gen8_gmch_probe(struct i915_ggtt *ggtt)
 		ggtt->base.clear_range = gen8_ggtt_clear_range;
 
 	ggtt->base.insert_entries = gen8_ggtt_insert_entries;
-	if (IS_CHERRYVIEW(dev_priv))
-		ggtt->base.insert_entries = gen8_ggtt_insert_entries__BKL;
 
 	ggtt->invalidate = gen6_ggtt_invalidate;
 
@@ -3183,6 +2881,7 @@ int i915_ggtt_probe_hw(struct drm_i915_private *dev_priv)
 	int ret;
 
 	ggtt->base.i915 = dev_priv;
+	ggtt->base.dma = &dev_priv->drm.pdev->dev;
 
 	if (INTEL_GEN(dev_priv) <= 5)
 		ret = i915_gmch_probe(ggtt);
@@ -3242,14 +2941,14 @@ int i915_ggtt_init_hw(struct drm_i915_private *dev_priv)
 
 	INIT_LIST_HEAD(&dev_priv->vm_list);
 
-	/* Subtract the guard page before address space initialization to
-	 * shrink the range used by drm_mm.
+	/* Note that we use page colouring to enforce a guard page at the
+	 * end of the address space. This is required as the CS may prefetch
+	 * beyond the end of the batch buffer, across the page boundary,
+	 * and beyond the end of the GTT if we do not provide a guard.
 	 */
 	mutex_lock(&dev_priv->drm.struct_mutex);
-	ggtt->base.total -= PAGE_SIZE;
 	i915_address_space_init(&ggtt->base, dev_priv, "[global]");
-	ggtt->base.total += PAGE_SIZE;
-	if (!HAS_LLC(dev_priv))
+	if (!HAS_LLC(dev_priv) && !USES_PPGTT(dev_priv))
 		ggtt->base.mm.color_adjust = i915_gtt_color_adjust;
 	mutex_unlock(&dev_priv->drm.struct_mutex);
 
@@ -3303,7 +3002,7 @@ void i915_gem_restore_gtt_mappings(struct drm_i915_private *dev_priv)
 	i915_check_and_clear_faults(dev_priv);
 
 	/* First fill our portion of the GTT with scratch pages */
-	ggtt->base.clear_range(&ggtt->base, ggtt->base.start, ggtt->base.total);
+	ggtt->base.clear_range(&ggtt->base, 0, ggtt->base.total);
 
 	ggtt->base.closed = true; /* skip rewriting PTE on VMA unbind */
 
@@ -3344,8 +3043,6 @@ void i915_gem_restore_gtt_mappings(struct drm_i915_private *dev_priv)
 		struct i915_address_space *vm;
 
 		list_for_each_entry(vm, &dev_priv->vm_list, global_link) {
-			/* TODO: Perhaps it shouldn't be gen6 specific */
-
 			struct i915_hw_ppgtt *ppgtt;
 
 			if (i915_is_ggtt(vm))
@@ -3353,8 +3050,7 @@ void i915_gem_restore_gtt_mappings(struct drm_i915_private *dev_priv)
 			else
 				ppgtt = i915_vm_to_ppgtt(vm);
 
-			gen6_write_page_range(dev_priv, &ppgtt->pd,
-					      0, ppgtt->base.total);
+			gen6_write_page_range(ppgtt, 0, ppgtt->base.total);
 		}
 	}
 
@@ -3389,11 +3085,11 @@ rotate_pages(const dma_addr_t *in, unsigned int offset,
 	return sg;
 }
 
-static struct sg_table *
-intel_rotate_fb_obj_pages(const struct intel_rotation_info *rot_info,
-			  struct drm_i915_gem_object *obj)
+static noinline struct sg_table *
+intel_rotate_pages(struct intel_rotation_info *rot_info,
+		   struct drm_i915_gem_object *obj)
 {
-	const size_t n_pages = obj->base.size / PAGE_SIZE;
+	const unsigned long n_pages = obj->base.size / PAGE_SIZE;
 	unsigned int size = intel_rotation_info_size(rot_info);
 	struct sgt_iter sgt_iter;
 	dma_addr_t dma_addr;
@@ -3452,7 +3148,7 @@ err_st_alloc:
 	return ERR_PTR(ret);
 }
 
-static struct sg_table *
+static noinline struct sg_table *
 intel_partial_pages(const struct i915_ggtt_view *view,
 		    struct drm_i915_gem_object *obj)
 {
@@ -3506,7 +3202,7 @@ err_st_alloc:
 static int
 i915_get_ggtt_vma_pages(struct i915_vma *vma)
 {
-	int ret = 0;
+	int ret;
 
 	/* The vma->pages are only valid within the lifespan of the borrowed
 	 * obj->mm.pages. When the obj->mm.pages sg_table is regenerated, so
@@ -3515,32 +3211,33 @@ i915_get_ggtt_vma_pages(struct i915_vma *vma)
 	 */
 	GEM_BUG_ON(!i915_gem_object_has_pinned_pages(vma->obj));
 
-	if (vma->pages)
+	switch (vma->ggtt_view.type) {
+	case I915_GGTT_VIEW_NORMAL:
+		vma->pages = vma->obj->mm.pages;
 		return 0;
 
-	if (vma->ggtt_view.type == I915_GGTT_VIEW_NORMAL)
-		vma->pages = vma->obj->mm.pages;
-	else if (vma->ggtt_view.type == I915_GGTT_VIEW_ROTATED)
+	case I915_GGTT_VIEW_ROTATED:
 		vma->pages =
-			intel_rotate_fb_obj_pages(&vma->ggtt_view.rotated,
-						  vma->obj);
-	else if (vma->ggtt_view.type == I915_GGTT_VIEW_PARTIAL)
+			intel_rotate_pages(&vma->ggtt_view.rotated, vma->obj);
+		break;
+
+	case I915_GGTT_VIEW_PARTIAL:
 		vma->pages = intel_partial_pages(&vma->ggtt_view, vma->obj);
-	else
+		break;
+
+	default:
 		WARN_ONCE(1, "GGTT view %u not implemented!\n",
 			  vma->ggtt_view.type);
+		return -EINVAL;
+	}
 
-	if (!vma->pages) {
-		DRM_ERROR("Failed to get pages for GGTT view type %u!\n",
-			  vma->ggtt_view.type);
-		ret = -EINVAL;
-	} else if (IS_ERR(vma->pages)) {
+	ret = 0;
+	if (unlikely(IS_ERR(vma->pages))) {
 		ret = PTR_ERR(vma->pages);
 		vma->pages = NULL;
 		DRM_ERROR("Failed to get pages for VMA view type %u (%d)!\n",
 			  vma->ggtt_view.type, ret);
 	}
-
 	return ret;
 }
 
@@ -3743,3 +3440,8 @@ int i915_gem_gtt_insert(struct i915_address_space *vm,
 					   size, alignment, color,
 					   start, end, DRM_MM_INSERT_EVICT);
 }
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/mock_gtt.c"
+#include "selftests/i915_gem_gtt.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h
index 3c5ef5358cef..fb15684c1d83 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.h
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
@@ -36,9 +36,11 @@
 
 #include <linux/io-mapping.h>
 #include <linux/mm.h>
+#include <linux/pagevec.h>
 
 #include "i915_gem_timeline.h"
 #include "i915_gem_request.h"
+#include "i915_selftest.h"
 
 #define I915_GTT_PAGE_SIZE 4096UL
 #define I915_GTT_MIN_ALIGNMENT I915_GTT_PAGE_SIZE
@@ -51,11 +53,11 @@
 struct drm_i915_file_private;
 struct drm_i915_fence_reg;
 
-typedef uint32_t gen6_pte_t;
-typedef uint64_t gen8_pte_t;
-typedef uint64_t gen8_pde_t;
-typedef uint64_t gen8_ppgtt_pdpe_t;
-typedef uint64_t gen8_ppgtt_pml4e_t;
+typedef u32 gen6_pte_t;
+typedef u64 gen8_pte_t;
+typedef u64 gen8_pde_t;
+typedef u64 gen8_ppgtt_pdpe_t;
+typedef u64 gen8_ppgtt_pml4e_t;
 
 #define ggtt_total_entries(ggtt) ((ggtt)->base.total >> PAGE_SHIFT)
 
@@ -67,7 +69,7 @@ typedef uint64_t gen8_ppgtt_pml4e_t;
 #define GEN6_PTE_UNCACHED		(1 << 1)
 #define GEN6_PTE_VALID			(1 << 0)
 
-#define I915_PTES(pte_len)		(PAGE_SIZE / (pte_len))
+#define I915_PTES(pte_len)		((unsigned int)(PAGE_SIZE / (pte_len)))
 #define I915_PTE_MASK(pte_len)		(I915_PTES(pte_len) - 1)
 #define I915_PDES			512
 #define I915_PDE_MASK			(I915_PDES - 1)
@@ -99,13 +101,20 @@ typedef uint64_t gen8_ppgtt_pml4e_t;
 #define HSW_GTT_ADDR_ENCODE(addr)	((addr) | (((addr) >> 28) & 0x7f0))
 #define HSW_PTE_ADDR_ENCODE(addr)	HSW_GTT_ADDR_ENCODE(addr)
 
-/* GEN8 legacy style address is defined as a 3 level page table:
+/* GEN8 32b style address is defined as a 3 level page table:
  * 31:30 | 29:21 | 20:12 |  11:0
  * PDPE  |  PDE  |  PTE  | offset
  * The difference as compared to normal x86 3 level page table is the PDPEs are
  * programmed via register.
- *
- * GEN8 48b legacy style address is defined as a 4 level page table:
+ */
+#define GEN8_3LVL_PDPES			4
+#define GEN8_PDE_SHIFT			21
+#define GEN8_PDE_MASK			0x1ff
+#define GEN8_PTE_SHIFT			12
+#define GEN8_PTE_MASK			0x1ff
+#define GEN8_PTES			I915_PTES(sizeof(gen8_pte_t))
+
+/* GEN8 48b style address is defined as a 4 level page table:
  * 47:39 | 38:30 | 29:21 | 20:12 |  11:0
  * PML4E | PDPE  |  PDE  |  PTE  | offset
  */
@@ -116,15 +125,6 @@ typedef uint64_t gen8_ppgtt_pml4e_t;
 /* NB: GEN8_PDPE_MASK is untrue for 32b platforms, but it has no impact on 32b page
  * tables */
 #define GEN8_PDPE_MASK			0x1ff
-#define GEN8_PDE_SHIFT			21
-#define GEN8_PDE_MASK			0x1ff
-#define GEN8_PTE_SHIFT			12
-#define GEN8_PTE_MASK			0x1ff
-#define GEN8_LEGACY_PDPES		4
-#define GEN8_PTES			I915_PTES(sizeof(gen8_pte_t))
-
-#define I915_PDPES_PER_PDP(dev_priv)	(USES_FULL_48BIT_PPGTT(dev_priv) ?\
-					GEN8_PML4ES_PER_PML4 : GEN8_LEGACY_PDPES)
 
 #define PPAT_UNCACHED_INDEX		(_PAGE_PWT | _PAGE_PCD)
 #define PPAT_CACHED_PDE_INDEX		0 /* WB LLC */
@@ -141,7 +141,7 @@ typedef uint64_t gen8_ppgtt_pml4e_t;
 #define GEN8_PPAT_WC			(1<<0)
 #define GEN8_PPAT_UC			(0<<0)
 #define GEN8_PPAT_ELLC_OVERRIDE		(0<<2)
-#define GEN8_PPAT(i, x)			((uint64_t) (x) << ((i) * 8))
+#define GEN8_PPAT(i, x)			((u64)(x) << ((i) * 8))
 
 struct sg_table;
 
@@ -208,7 +208,7 @@ struct i915_page_dma {
 		/* For gen6/gen7 only. This is the offset in the GGTT
 		 * where the page directory entries for PPGTT begin
 		 */
-		uint32_t ggtt_offset;
+		u32 ggtt_offset;
 	};
 };
 
@@ -218,28 +218,24 @@ struct i915_page_dma {
 
 struct i915_page_table {
 	struct i915_page_dma base;
-
-	unsigned long *used_ptes;
+	unsigned int used_ptes;
 };
 
 struct i915_page_directory {
 	struct i915_page_dma base;
 
-	unsigned long *used_pdes;
 	struct i915_page_table *page_table[I915_PDES]; /* PDEs */
+	unsigned int used_pdes;
 };
 
 struct i915_page_directory_pointer {
 	struct i915_page_dma base;
-
-	unsigned long *used_pdpes;
 	struct i915_page_directory **page_directory;
+	unsigned int used_pdpes;
 };
 
 struct i915_pml4 {
 	struct i915_page_dma base;
-
-	DECLARE_BITMAP(used_pml4es, GEN8_PML4ES_PER_PML4);
 	struct i915_page_directory_pointer *pdps[GEN8_PML4ES_PER_PML4];
 };
 
@@ -247,6 +243,7 @@ struct i915_address_space {
 	struct drm_mm mm;
 	struct i915_gem_timeline timeline;
 	struct drm_i915_private *i915;
+	struct device *dma;
 	/* Every address space belongs to a struct file - except for the global
 	 * GTT that is owned by the driver (and so @file is set to NULL). In
 	 * principle, no information should leak from one context to another
@@ -257,7 +254,6 @@ struct i915_address_space {
 	 */
 	struct drm_i915_file_private *file;
 	struct list_head global_link;
-	u64 start;		/* Start offset always 0 for dri2 */
 	u64 total;		/* size addr space maps (ex. 2GB for ggtt) */
 
 	bool closed;
@@ -297,6 +293,9 @@ struct i915_address_space {
 	 */
 	struct list_head unbound_list;
 
+	struct pagevec free_pages;
+	bool pt_kmap_wc;
+
 	/* FIXME: Need a more generic return type */
 	gen6_pte_t (*pte_encode)(dma_addr_t addr,
 				 enum i915_cache_level level,
@@ -304,20 +303,19 @@ struct i915_address_space {
 	/* flags for pte_encode */
 #define PTE_READ_ONLY	(1<<0)
 	int (*allocate_va_range)(struct i915_address_space *vm,
-				 uint64_t start,
-				 uint64_t length);
+				 u64 start, u64 length);
 	void (*clear_range)(struct i915_address_space *vm,
-			    uint64_t start,
-			    uint64_t length);
+			    u64 start, u64 length);
 	void (*insert_page)(struct i915_address_space *vm,
 			    dma_addr_t addr,
-			    uint64_t offset,
+			    u64 offset,
 			    enum i915_cache_level cache_level,
 			    u32 flags);
 	void (*insert_entries)(struct i915_address_space *vm,
 			       struct sg_table *st,
-			       uint64_t start,
-			       enum i915_cache_level cache_level, u32 flags);
+			       u64 start,
+			       enum i915_cache_level cache_level,
+			       u32 flags);
 	void (*cleanup)(struct i915_address_space *vm);
 	/** Unmap an object from an address space. This usually consists of
 	 * setting the valid PTE entries to a reserved scratch page. */
@@ -326,10 +324,18 @@ struct i915_address_space {
 	int (*bind_vma)(struct i915_vma *vma,
 			enum i915_cache_level cache_level,
 			u32 flags);
+
+	I915_SELFTEST_DECLARE(struct fault_attr fault_attr);
 };
 
 #define i915_is_ggtt(V) (!(V)->file)
 
+static inline bool
+i915_vm_is_48bit(const struct i915_address_space *vm)
+{
+	return (vm->total - 1) >> 32;
+}
+
 /* The Graphics Translation Table is the way in which GEN hardware translates a
  * Graphics Virtual Address into a Physical Address. In addition to the normal
  * collateral associated with any va->pa translations GEN hardware also has a
@@ -381,7 +387,6 @@ struct i915_hw_ppgtt {
 
 	gen6_pte_t __iomem *pd_addr;
 
-	int (*enable)(struct i915_hw_ppgtt *ppgtt);
 	int (*switch_mm)(struct i915_hw_ppgtt *ppgtt,
 			 struct drm_i915_gem_request *req);
 	void (*debug_dump)(struct i915_hw_ppgtt *ppgtt, struct seq_file *m);
@@ -409,9 +414,9 @@ struct i915_hw_ppgtt {
 		(pt = (pd)->page_table[iter], true);			\
 	     ++iter)
 
-static inline uint32_t i915_pte_index(uint64_t address, uint32_t pde_shift)
+static inline u32 i915_pte_index(u64 address, unsigned int pde_shift)
 {
-	const uint32_t mask = NUM_PTE(pde_shift) - 1;
+	const u32 mask = NUM_PTE(pde_shift) - 1;
 
 	return (address >> PAGE_SHIFT) & mask;
 }
@@ -420,11 +425,10 @@ static inline uint32_t i915_pte_index(uint64_t address, uint32_t pde_shift)
  * does not cross a page table boundary, so the max value would be
  * GEN6_PTES for GEN6, and GEN8_PTES for GEN8.
 */
-static inline uint32_t i915_pte_count(uint64_t addr, size_t length,
-				      uint32_t pde_shift)
+static inline u32 i915_pte_count(u64 addr, u64 length, unsigned int pde_shift)
 {
-	const uint64_t mask = ~((1ULL << pde_shift) - 1);
-	uint64_t end;
+	const u64 mask = ~((1ULL << pde_shift) - 1);
+	u64 end;
 
 	WARN_ON(length == 0);
 	WARN_ON(offset_in_page(addr|length));
@@ -437,26 +441,35 @@ static inline uint32_t i915_pte_count(uint64_t addr, size_t length,
 	return i915_pte_index(end, pde_shift) - i915_pte_index(addr, pde_shift);
 }
 
-static inline uint32_t i915_pde_index(uint64_t addr, uint32_t shift)
+static inline u32 i915_pde_index(u64 addr, u32 shift)
 {
 	return (addr >> shift) & I915_PDE_MASK;
 }
 
-static inline uint32_t gen6_pte_index(uint32_t addr)
+static inline u32 gen6_pte_index(u32 addr)
 {
 	return i915_pte_index(addr, GEN6_PDE_SHIFT);
 }
 
-static inline size_t gen6_pte_count(uint32_t addr, uint32_t length)
+static inline u32 gen6_pte_count(u32 addr, u32 length)
 {
 	return i915_pte_count(addr, length, GEN6_PDE_SHIFT);
 }
 
-static inline uint32_t gen6_pde_index(uint32_t addr)
+static inline u32 gen6_pde_index(u32 addr)
 {
 	return i915_pde_index(addr, GEN6_PDE_SHIFT);
 }
 
+static inline unsigned int
+i915_pdpes_per_pdp(const struct i915_address_space *vm)
+{
+	if (i915_vm_is_48bit(vm))
+		return GEN8_PML4ES_PER_PML4;
+
+	return GEN8_3LVL_PDPES;
+}
+
 /* Equivalent to the gen6 version, For each pde iterates over every pde
  * between from start until start + length. On gen8+ it simply iterates
  * over every page directory entry in a page directory.
@@ -471,7 +484,7 @@ static inline uint32_t gen6_pde_index(uint32_t addr)
 
 #define gen8_for_each_pdpe(pd, pdp, start, length, iter)		\
 	for (iter = gen8_pdpe_index(start);				\
-	     length > 0 && iter < I915_PDPES_PER_PDP(dev) &&		\
+	     length > 0 && iter < i915_pdpes_per_pdp(vm) &&		\
 		(pd = (pdp)->page_directory[iter], true);		\
 	     ({ u64 temp = ALIGN(start+1, 1 << GEN8_PDPE_SHIFT);	\
 		    temp = min(temp - start, length);			\
@@ -485,27 +498,27 @@ static inline uint32_t gen6_pde_index(uint32_t addr)
 		    temp = min(temp - start, length);			\
 		    start += temp, length -= temp; }), ++iter)
 
-static inline uint32_t gen8_pte_index(uint64_t address)
+static inline u32 gen8_pte_index(u64 address)
 {
 	return i915_pte_index(address, GEN8_PDE_SHIFT);
 }
 
-static inline uint32_t gen8_pde_index(uint64_t address)
+static inline u32 gen8_pde_index(u64 address)
 {
 	return i915_pde_index(address, GEN8_PDE_SHIFT);
 }
 
-static inline uint32_t gen8_pdpe_index(uint64_t address)
+static inline u32 gen8_pdpe_index(u64 address)
 {
 	return (address >> GEN8_PDPE_SHIFT) & GEN8_PDPE_MASK;
 }
 
-static inline uint32_t gen8_pml4e_index(uint64_t address)
+static inline u32 gen8_pml4e_index(u64 address)
 {
 	return (address >> GEN8_PML4E_SHIFT) & GEN8_PML4E_MASK;
 }
 
-static inline size_t gen8_pte_count(uint64_t address, uint64_t length)
+static inline u64 gen8_pte_count(u64 address, u64 length)
 {
 	return i915_pte_count(address, length, GEN8_PDE_SHIFT);
 }
@@ -513,9 +526,7 @@ static inline size_t gen8_pte_count(uint64_t address, uint64_t length)
 static inline dma_addr_t
 i915_page_dir_dma_addr(const struct i915_hw_ppgtt *ppgtt, const unsigned n)
 {
-	return test_bit(n, ppgtt->pdp.used_pdpes) ?
-		px_dma(ppgtt->pdp.page_directory[n]) :
-		px_dma(ppgtt->base.scratch_pd);
+	return px_dma(ppgtt->pdp.page_directory[n]);
 }
 
 static inline struct i915_ggtt *
@@ -525,6 +536,9 @@ i915_vm_to_ggtt(struct i915_address_space *vm)
 	return container_of(vm, struct i915_ggtt, base);
 }
 
+int i915_gem_init_aliasing_ppgtt(struct drm_i915_private *i915);
+void i915_gem_fini_aliasing_ppgtt(struct drm_i915_private *i915);
+
 int i915_ggtt_probe_hw(struct drm_i915_private *dev_priv);
 int i915_ggtt_init_hw(struct drm_i915_private *dev_priv);
 int i915_ggtt_enable_hw(struct drm_i915_private *dev_priv);
diff --git a/drivers/gpu/drm/i915/i915_gem_internal.c b/drivers/gpu/drm/i915/i915_gem_internal.c
index 933019e1b206..fc950abbe400 100644
--- a/drivers/gpu/drm/i915/i915_gem_internal.c
+++ b/drivers/gpu/drm/i915/i915_gem_internal.c
@@ -35,8 +35,10 @@ static void internal_free_pages(struct sg_table *st)
 {
 	struct scatterlist *sg;
 
-	for (sg = st->sgl; sg; sg = __sg_next(sg))
-		__free_pages(sg_page(sg), get_order(sg->length));
+	for (sg = st->sgl; sg; sg = __sg_next(sg)) {
+		if (sg_page(sg))
+			__free_pages(sg_page(sg), get_order(sg->length));
+	}
 
 	sg_free_table(st);
 	kfree(st);
@@ -133,6 +135,7 @@ create_st:
 	return st;
 
 err:
+	sg_set_page(sg, NULL, 0, 0);
 	sg_mark_end(sg);
 	internal_free_pages(st);
 	return ERR_PTR(-ENOMEM);
diff --git a/drivers/gpu/drm/i915/i915_gem_object.h b/drivers/gpu/drm/i915/i915_gem_object.h
index bf90b07163d1..33b0dc4782a9 100644
--- a/drivers/gpu/drm/i915/i915_gem_object.h
+++ b/drivers/gpu/drm/i915/i915_gem_object.h
@@ -33,6 +33,8 @@
 
 #include <drm/i915_drm.h>
 
+#include "i915_selftest.h"
+
 struct drm_i915_gem_object_ops {
 	unsigned int flags;
 #define I915_GEM_OBJECT_HAS_STRUCT_PAGE 0x1
@@ -84,6 +86,7 @@ struct drm_i915_gem_object {
 	struct list_head obj_exec_link;
 
 	struct list_head batch_pool_link;
+	I915_SELFTEST_DECLARE(struct list_head st_link);
 
 	unsigned long flags;
 
@@ -162,19 +165,23 @@ struct drm_i915_gem_object {
 	struct reservation_object *resv;
 
 	/** References from framebuffers, locks out tiling changes. */
-	unsigned long framebuffer_references;
+	unsigned int framebuffer_references;
 
 	/** Record of address bit 17 of each page at last unbind. */
 	unsigned long *bit_17;
 
-	struct i915_gem_userptr {
-		uintptr_t ptr;
-		unsigned read_only :1;
+	union {
+		struct i915_gem_userptr {
+			uintptr_t ptr;
+			unsigned read_only :1;
+
+			struct i915_mm_struct *mm;
+			struct i915_mmu_object *mmu_object;
+			struct work_struct *work;
+		} userptr;
 
-		struct i915_mm_struct *mm;
-		struct i915_mmu_object *mmu_object;
-		struct work_struct *work;
-	} userptr;
+		unsigned long scratch;
+	};
 
 	/** for phys allocated objects */
 	struct drm_dma_handle *phys_handle;
@@ -253,6 +260,16 @@ extern void drm_gem_object_unreference(struct drm_gem_object *);
 __deprecated
 extern void drm_gem_object_unreference_unlocked(struct drm_gem_object *);
 
+static inline void i915_gem_object_lock(struct drm_i915_gem_object *obj)
+{
+	reservation_object_lock(obj->resv, NULL);
+}
+
+static inline void i915_gem_object_unlock(struct drm_i915_gem_object *obj)
+{
+	reservation_object_unlock(obj->resv);
+}
+
 static inline bool
 i915_gem_object_is_dead(const struct drm_i915_gem_object *obj)
 {
@@ -299,6 +316,12 @@ i915_gem_object_clear_active_reference(struct drm_i915_gem_object *obj)
 
 void __i915_gem_object_release_unless_active(struct drm_i915_gem_object *obj);
 
+static inline bool
+i915_gem_object_is_framebuffer(const struct drm_i915_gem_object *obj)
+{
+	return READ_ONCE(obj->framebuffer_references);
+}
+
 static inline unsigned int
 i915_gem_object_get_tiling(struct drm_i915_gem_object *obj)
 {
@@ -357,5 +380,7 @@ i915_gem_object_last_write_engine(struct drm_i915_gem_object *obj)
 	return engine;
 }
 
+void i915_gem_object_flush_if_display(struct drm_i915_gem_object *obj);
+
 #endif
 
diff --git a/drivers/gpu/drm/i915/i915_gem_request.c b/drivers/gpu/drm/i915/i915_gem_request.c
index e7c3c0318ff6..1e1d9f2072cd 100644
--- a/drivers/gpu/drm/i915/i915_gem_request.c
+++ b/drivers/gpu/drm/i915/i915_gem_request.c
@@ -72,7 +72,6 @@ static void i915_fence_release(struct dma_fence *fence)
 	 * caught trying to reuse dead objects.
 	 */
 	i915_sw_fence_fini(&req->submit);
-	i915_sw_fence_fini(&req->execute);
 
 	kmem_cache_free(req->i915->requests, req);
 }
@@ -86,42 +85,20 @@ const struct dma_fence_ops i915_fence_ops = {
 	.release = i915_fence_release,
 };
 
-int i915_gem_request_add_to_client(struct drm_i915_gem_request *req,
-				   struct drm_file *file)
-{
-	struct drm_i915_private *dev_private;
-	struct drm_i915_file_private *file_priv;
-
-	WARN_ON(!req || !file || req->file_priv);
-
-	if (!req || !file)
-		return -EINVAL;
-
-	if (req->file_priv)
-		return -EINVAL;
-
-	dev_private = req->i915;
-	file_priv = file->driver_priv;
-
-	spin_lock(&file_priv->mm.lock);
-	req->file_priv = file_priv;
-	list_add_tail(&req->client_list, &file_priv->mm.request_list);
-	spin_unlock(&file_priv->mm.lock);
-
-	return 0;
-}
-
 static inline void
 i915_gem_request_remove_from_client(struct drm_i915_gem_request *request)
 {
-	struct drm_i915_file_private *file_priv = request->file_priv;
+	struct drm_i915_file_private *file_priv;
 
+	file_priv = request->file_priv;
 	if (!file_priv)
 		return;
 
 	spin_lock(&file_priv->mm.lock);
-	list_del(&request->client_list);
-	request->file_priv = NULL;
+	if (request->file_priv) {
+		list_del(&request->client_link);
+		request->file_priv = NULL;
+	}
 	spin_unlock(&file_priv->mm.lock);
 }
 
@@ -201,6 +178,92 @@ i915_priotree_init(struct i915_priotree *pt)
 	pt->priority = INT_MIN;
 }
 
+static int reset_all_global_seqno(struct drm_i915_private *i915, u32 seqno)
+{
+	struct i915_gem_timeline *timeline = &i915->gt.global_timeline;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	int ret;
+
+	/* Carefully retire all requests without writing to the rings */
+	ret = i915_gem_wait_for_idle(i915,
+				     I915_WAIT_INTERRUPTIBLE |
+				     I915_WAIT_LOCKED);
+	if (ret)
+		return ret;
+
+	i915_gem_retire_requests(i915);
+	GEM_BUG_ON(i915->gt.active_requests > 1);
+
+	/* If the seqno wraps around, we need to clear the breadcrumb rbtree */
+	for_each_engine(engine, i915, id) {
+		struct intel_timeline *tl = &timeline->engine[id];
+
+		if (wait_for(intel_engine_is_idle(engine), 50))
+			return -EBUSY;
+
+		if (!i915_seqno_passed(seqno, tl->seqno)) {
+			/* spin until threads are complete */
+			while (intel_breadcrumbs_busy(engine))
+				cond_resched();
+		}
+
+		/* Finally reset hw state */
+		tl->seqno = seqno;
+		intel_engine_init_global_seqno(engine, seqno);
+	}
+
+	list_for_each_entry(timeline, &i915->gt.timelines, link) {
+		for_each_engine(engine, i915, id) {
+			struct intel_timeline *tl = &timeline->engine[id];
+
+			memset(tl->sync_seqno, 0, sizeof(tl->sync_seqno));
+		}
+	}
+
+	return 0;
+}
+
+int i915_gem_set_global_seqno(struct drm_device *dev, u32 seqno)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+
+	lockdep_assert_held(&dev_priv->drm.struct_mutex);
+
+	if (seqno == 0)
+		return -EINVAL;
+
+	/* HWS page needs to be set less than what we
+	 * will inject to ring
+	 */
+	return reset_all_global_seqno(dev_priv, seqno - 1);
+}
+
+static int reserve_seqno(struct intel_engine_cs *engine)
+{
+	u32 active = ++engine->timeline->inflight_seqnos;
+	u32 seqno = engine->timeline->seqno;
+	int ret;
+
+	/* Reservation is fine until we need to wrap around */
+	if (likely(!add_overflows(seqno, active)))
+		return 0;
+
+	ret = reset_all_global_seqno(engine->i915, 0);
+	if (ret) {
+		engine->timeline->inflight_seqnos--;
+		return ret;
+	}
+
+	return 0;
+}
+
+static void unreserve_seqno(struct intel_engine_cs *engine)
+{
+	GEM_BUG_ON(!engine->timeline->inflight_seqnos);
+	engine->timeline->inflight_seqnos--;
+}
+
 void i915_gem_retire_noop(struct i915_gem_active *active,
 			  struct drm_i915_gem_request *request)
 {
@@ -214,7 +277,6 @@ static void i915_gem_request_retire(struct drm_i915_gem_request *request)
 
 	lockdep_assert_held(&request->i915->drm.struct_mutex);
 	GEM_BUG_ON(!i915_sw_fence_signaled(&request->submit));
-	GEM_BUG_ON(!i915_sw_fence_signaled(&request->execute));
 	GEM_BUG_ON(!i915_gem_request_completed(request));
 	GEM_BUG_ON(!request->i915->gt.active_requests);
 
@@ -240,6 +302,7 @@ static void i915_gem_request_retire(struct drm_i915_gem_request *request)
 				 &request->i915->gt.idle_work,
 				 msecs_to_jiffies(100));
 	}
+	unreserve_seqno(request->engine);
 
 	/* Walk through the active list, calling retire on each. This allows
 	 * objects to track their GPU activity and mark themselves as idle
@@ -310,88 +373,9 @@ void i915_gem_request_retire_upto(struct drm_i915_gem_request *req)
 	} while (tmp != req);
 }
 
-static int i915_gem_init_global_seqno(struct drm_i915_private *i915, u32 seqno)
-{
-	struct i915_gem_timeline *timeline = &i915->gt.global_timeline;
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	int ret;
-
-	/* Carefully retire all requests without writing to the rings */
-	ret = i915_gem_wait_for_idle(i915,
-				     I915_WAIT_INTERRUPTIBLE |
-				     I915_WAIT_LOCKED);
-	if (ret)
-		return ret;
-
-	i915_gem_retire_requests(i915);
-	GEM_BUG_ON(i915->gt.active_requests > 1);
-
-	/* If the seqno wraps around, we need to clear the breadcrumb rbtree */
-	if (!i915_seqno_passed(seqno, atomic_read(&timeline->seqno))) {
-		while (intel_breadcrumbs_busy(i915))
-			cond_resched(); /* spin until threads are complete */
-	}
-	atomic_set(&timeline->seqno, seqno);
-
-	/* Finally reset hw state */
-	for_each_engine(engine, i915, id)
-		intel_engine_init_global_seqno(engine, seqno);
-
-	list_for_each_entry(timeline, &i915->gt.timelines, link) {
-		for_each_engine(engine, i915, id) {
-			struct intel_timeline *tl = &timeline->engine[id];
-
-			memset(tl->sync_seqno, 0, sizeof(tl->sync_seqno));
-		}
-	}
-
-	return 0;
-}
-
-int i915_gem_set_global_seqno(struct drm_device *dev, u32 seqno)
+static u32 timeline_get_seqno(struct intel_timeline *tl)
 {
-	struct drm_i915_private *dev_priv = to_i915(dev);
-
-	lockdep_assert_held(&dev_priv->drm.struct_mutex);
-
-	if (seqno == 0)
-		return -EINVAL;
-
-	/* HWS page needs to be set less than what we
-	 * will inject to ring
-	 */
-	return i915_gem_init_global_seqno(dev_priv, seqno - 1);
-}
-
-static int reserve_global_seqno(struct drm_i915_private *i915)
-{
-	u32 active_requests = ++i915->gt.active_requests;
-	u32 seqno = atomic_read(&i915->gt.global_timeline.seqno);
-	int ret;
-
-	/* Reservation is fine until we need to wrap around */
-	if (likely(seqno + active_requests > seqno))
-		return 0;
-
-	ret = i915_gem_init_global_seqno(i915, 0);
-	if (ret) {
-		i915->gt.active_requests--;
-		return ret;
-	}
-
-	return 0;
-}
-
-static u32 __timeline_get_seqno(struct i915_gem_timeline *tl)
-{
-	/* seqno only incremented under a mutex */
-	return ++tl->seqno.counter;
-}
-
-static u32 timeline_get_seqno(struct i915_gem_timeline *tl)
-{
-	return atomic_inc_return(&tl->seqno);
+	return ++tl->seqno;
 }
 
 void __i915_gem_request_submit(struct drm_i915_gem_request *request)
@@ -400,19 +384,19 @@ void __i915_gem_request_submit(struct drm_i915_gem_request *request)
 	struct intel_timeline *timeline;
 	u32 seqno;
 
+	GEM_BUG_ON(!irqs_disabled());
+	lockdep_assert_held(&engine->timeline->lock);
+
+	trace_i915_gem_request_execute(request);
+
 	/* Transfer from per-context onto the global per-engine timeline */
 	timeline = engine->timeline;
 	GEM_BUG_ON(timeline == request->timeline);
-	assert_spin_locked(&timeline->lock);
 
-	seqno = timeline_get_seqno(timeline->common);
+	seqno = timeline_get_seqno(timeline);
 	GEM_BUG_ON(!seqno);
 	GEM_BUG_ON(i915_seqno_passed(intel_engine_get_seqno(engine), seqno));
 
-	GEM_BUG_ON(i915_seqno_passed(timeline->last_submitted_seqno, seqno));
-	request->previous_seqno = timeline->last_submitted_seqno;
-	timeline->last_submitted_seqno = seqno;
-
 	/* We may be recursing from the signal callback of another i915 fence */
 	spin_lock_nested(&request->lock, SINGLE_DEPTH_NESTING);
 	request->global_seqno = seqno;
@@ -420,7 +404,6 @@ void __i915_gem_request_submit(struct drm_i915_gem_request *request)
 		intel_engine_enable_signaling(request);
 	spin_unlock(&request->lock);
 
-	GEM_BUG_ON(!request->global_seqno);
 	engine->emit_breadcrumb(request,
 				request->ring->vaddr + request->postfix);
 
@@ -428,7 +411,7 @@ void __i915_gem_request_submit(struct drm_i915_gem_request *request)
 	list_move_tail(&request->link, &timeline->requests);
 	spin_unlock(&request->timeline->lock);
 
-	i915_sw_fence_commit(&request->execute);
+	wake_up_all(&request->execute);
 }
 
 void i915_gem_request_submit(struct drm_i915_gem_request *request)
@@ -444,33 +427,66 @@ void i915_gem_request_submit(struct drm_i915_gem_request *request)
 	spin_unlock_irqrestore(&engine->timeline->lock, flags);
 }
 
-static int __i915_sw_fence_call
-submit_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
+void __i915_gem_request_unsubmit(struct drm_i915_gem_request *request)
 {
-	struct drm_i915_gem_request *request =
-		container_of(fence, typeof(*request), submit);
+	struct intel_engine_cs *engine = request->engine;
+	struct intel_timeline *timeline;
 
-	switch (state) {
-	case FENCE_COMPLETE:
-		request->engine->submit_request(request);
-		break;
+	GEM_BUG_ON(!irqs_disabled());
+	lockdep_assert_held(&engine->timeline->lock);
 
-	case FENCE_FREE:
-		i915_gem_request_put(request);
-		break;
-	}
+	/* Only unwind in reverse order, required so that the per-context list
+	 * is kept in seqno/ring order.
+	 */
+	GEM_BUG_ON(request->global_seqno != engine->timeline->seqno);
+	engine->timeline->seqno--;
 
-	return NOTIFY_DONE;
+	/* We may be recursing from the signal callback of another i915 fence */
+	spin_lock_nested(&request->lock, SINGLE_DEPTH_NESTING);
+	request->global_seqno = 0;
+	if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &request->fence.flags))
+		intel_engine_cancel_signaling(request);
+	spin_unlock(&request->lock);
+
+	/* Transfer back from the global per-engine timeline to per-context */
+	timeline = request->timeline;
+	GEM_BUG_ON(timeline == engine->timeline);
+
+	spin_lock(&timeline->lock);
+	list_move(&request->link, &timeline->requests);
+	spin_unlock(&timeline->lock);
+
+	/* We don't need to wake_up any waiters on request->execute, they
+	 * will get woken by any other event or us re-adding this request
+	 * to the engine timeline (__i915_gem_request_submit()). The waiters
+	 * should be quite adapt at finding that the request now has a new
+	 * global_seqno to the one they went to sleep on.
+	 */
+}
+
+void i915_gem_request_unsubmit(struct drm_i915_gem_request *request)
+{
+	struct intel_engine_cs *engine = request->engine;
+	unsigned long flags;
+
+	/* Will be called from irq-context when using foreign fences. */
+	spin_lock_irqsave(&engine->timeline->lock, flags);
+
+	__i915_gem_request_unsubmit(request);
+
+	spin_unlock_irqrestore(&engine->timeline->lock, flags);
 }
 
 static int __i915_sw_fence_call
-execute_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
+submit_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
 {
 	struct drm_i915_gem_request *request =
-		container_of(fence, typeof(*request), execute);
+		container_of(fence, typeof(*request), submit);
 
 	switch (state) {
 	case FENCE_COMPLETE:
+		trace_i915_gem_request_submit(request);
+		request->engine->submit_request(request);
 		break;
 
 	case FENCE_FREE:
@@ -517,14 +533,14 @@ i915_gem_request_alloc(struct intel_engine_cs *engine,
 	if (ret)
 		return ERR_PTR(ret);
 
-	ret = reserve_global_seqno(dev_priv);
+	ret = reserve_seqno(engine);
 	if (ret)
 		goto err_unpin;
 
 	/* Move the oldest request to the slab-cache (if not in use!) */
 	req = list_first_entry_or_null(&engine->timeline->requests,
 				       typeof(*req), link);
-	if (req && __i915_gem_request_completed(req))
+	if (req && i915_gem_request_completed(req))
 		i915_gem_request_retire(req);
 
 	/* Beware: Dragons be flying overhead.
@@ -569,17 +585,11 @@ i915_gem_request_alloc(struct intel_engine_cs *engine,
 		       &i915_fence_ops,
 		       &req->lock,
 		       req->timeline->fence_context,
-		       __timeline_get_seqno(req->timeline->common));
+		       timeline_get_seqno(req->timeline));
 
 	/* We bump the ref for the fence chain */
 	i915_sw_fence_init(&i915_gem_request_get(req)->submit, submit_notify);
-	i915_sw_fence_init(&i915_gem_request_get(req)->execute, execute_notify);
-
-	/* Ensure that the execute fence completes after the submit fence -
-	 * as we complete the execute fence from within the submit fence
-	 * callback, its completion would otherwise be visible first.
-	 */
-	i915_sw_fence_await_sw_fence(&req->execute, &req->submit, &req->execq);
+	init_waitqueue_head(&req->execute);
 
 	i915_priotree_init(&req->priotree);
 
@@ -614,6 +624,8 @@ i915_gem_request_alloc(struct intel_engine_cs *engine,
 	 */
 	req->head = req->ring->tail;
 
+	/* Check that we didn't interrupt ourselves with a new request */
+	GEM_BUG_ON(req->timeline->seqno != req->fence.seqno);
 	return req;
 
 err_ctx:
@@ -624,7 +636,7 @@ err_ctx:
 
 	kmem_cache_free(dev_priv->requests, req);
 err_unreserve:
-	dev_priv->gt.active_requests--;
+	unreserve_seqno(engine);
 err_unpin:
 	engine->context_unpin(engine, ctx);
 	return ERR_PTR(ret);
@@ -634,6 +646,7 @@ static int
 i915_gem_request_await_request(struct drm_i915_gem_request *to,
 			       struct drm_i915_gem_request *from)
 {
+	u32 seqno;
 	int ret;
 
 	GEM_BUG_ON(to == from);
@@ -656,14 +669,15 @@ i915_gem_request_await_request(struct drm_i915_gem_request *to,
 		return ret < 0 ? ret : 0;
 	}
 
-	if (!from->global_seqno) {
+	seqno = i915_gem_request_global_seqno(from);
+	if (!seqno) {
 		ret = i915_sw_fence_await_dma_fence(&to->submit,
 						    &from->fence, 0,
 						    GFP_KERNEL);
 		return ret < 0 ? ret : 0;
 	}
 
-	if (from->global_seqno <= to->timeline->sync_seqno[from->engine->id])
+	if (seqno <= to->timeline->sync_seqno[from->engine->id])
 		return 0;
 
 	trace_i915_gem_ring_sync_to(to, from);
@@ -681,7 +695,7 @@ i915_gem_request_await_request(struct drm_i915_gem_request *to,
 			return ret;
 	}
 
-	to->timeline->sync_seqno[from->engine->id] = from->global_seqno;
+	to->timeline->sync_seqno[from->engine->id] = seqno;
 	return 0;
 }
 
@@ -827,6 +841,7 @@ void __i915_add_request(struct drm_i915_gem_request *request, bool flush_caches)
 	struct intel_ring *ring = request->ring;
 	struct intel_timeline *timeline = request->timeline;
 	struct drm_i915_gem_request *prev;
+	u32 *cs;
 	int err;
 
 	lockdep_assert_held(&request->i915->drm.struct_mutex);
@@ -836,8 +851,7 @@ void __i915_add_request(struct drm_i915_gem_request *request, bool flush_caches)
 	 * our i915_gem_request_alloc() and called __i915_add_request() before
 	 * us, the timeline will hold its seqno which is later than ours.
 	 */
-	GEM_BUG_ON(i915_seqno_passed(timeline->last_submitted_seqno,
-				     request->fence.seqno));
+	GEM_BUG_ON(timeline->seqno != request->fence.seqno);
 
 	/*
 	 * To ensure that this call will not fail, space for its emissions
@@ -865,10 +879,9 @@ void __i915_add_request(struct drm_i915_gem_request *request, bool flush_caches)
 	 * GPU processing the request, we never over-estimate the
 	 * position of the ring's HEAD.
 	 */
-	err = intel_ring_begin(request, engine->emit_breadcrumb_sz);
-	GEM_BUG_ON(err);
-	request->postfix = ring->tail;
-	ring->tail += engine->emit_breadcrumb_sz * sizeof(u32);
+	cs = intel_ring_begin(request, engine->emit_breadcrumb_sz);
+	GEM_BUG_ON(IS_ERR(cs));
+	request->postfix = intel_ring_offset(request, cs);
 
 	/* Seal the request and mark it as pending execution. Note that
 	 * we may inspect this state, without holding any locks, during
@@ -892,16 +905,14 @@ void __i915_add_request(struct drm_i915_gem_request *request, bool flush_caches)
 	list_add_tail(&request->link, &timeline->requests);
 	spin_unlock_irq(&timeline->lock);
 
-	GEM_BUG_ON(i915_seqno_passed(timeline->last_submitted_seqno,
-				     request->fence.seqno));
-
-	timeline->last_submitted_seqno = request->fence.seqno;
+	GEM_BUG_ON(timeline->seqno != request->fence.seqno);
 	i915_gem_active_set(&timeline->last_request, request);
 
 	list_add_tail(&request->ring_link, &ring->request_list);
 	request->emitted_jiffies = jiffies;
 
-	i915_gem_mark_busy(engine);
+	if (!request->i915->gt.active_requests++)
+		i915_gem_mark_busy(engine);
 
 	/* Let the backend know a new request has arrived that may need
 	 * to adjust the existing execution schedule due to a high priority
@@ -921,16 +932,6 @@ void __i915_add_request(struct drm_i915_gem_request *request, bool flush_caches)
 	local_bh_enable(); /* Kick the execlists tasklet if just scheduled */
 }
 
-static void reset_wait_queue(wait_queue_head_t *q, wait_queue_t *wait)
-{
-	unsigned long flags;
-
-	spin_lock_irqsave(&q->lock, flags);
-	if (list_empty(&wait->task_list))
-		__add_wait_queue(q, wait);
-	spin_unlock_irqrestore(&q->lock, flags);
-}
-
 static unsigned long local_clock_us(unsigned int *cpu)
 {
 	unsigned long t;
@@ -964,9 +965,10 @@ static bool busywait_stop(unsigned long timeout, unsigned int cpu)
 }
 
 bool __i915_spin_request(const struct drm_i915_gem_request *req,
-			 int state, unsigned long timeout_us)
+			 u32 seqno, int state, unsigned long timeout_us)
 {
-	unsigned int cpu;
+	struct intel_engine_cs *engine = req->engine;
+	unsigned int irq, cpu;
 
 	/* When waiting for high frequency requests, e.g. during synchronous
 	 * rendering split between the CPU and GPU, the finite amount of time
@@ -978,11 +980,24 @@ bool __i915_spin_request(const struct drm_i915_gem_request *req,
 	 * takes to sleep on a request, on the order of a microsecond.
 	 */
 
+	irq = atomic_read(&engine->irq_count);
 	timeout_us += local_clock_us(&cpu);
 	do {
-		if (__i915_gem_request_completed(req))
+		if (seqno != i915_gem_request_global_seqno(req))
+			break;
+
+		if (i915_seqno_passed(intel_engine_get_seqno(req->engine),
+				      seqno))
 			return true;
 
+		/* Seqno are meant to be ordered *before* the interrupt. If
+		 * we see an interrupt without a corresponding seqno advance,
+		 * assume we won't see one in the near future but require
+		 * the engine->seqno_barrier() to fixup coherency.
+		 */
+		if (atomic_read(&engine->irq_count) != irq)
+			break;
+
 		if (signal_pending_state(state, current))
 			break;
 
@@ -995,52 +1010,14 @@ bool __i915_spin_request(const struct drm_i915_gem_request *req,
 	return false;
 }
 
-static long
-__i915_request_wait_for_execute(struct drm_i915_gem_request *request,
-				unsigned int flags,
-				long timeout)
+static bool __i915_wait_request_check_and_reset(struct drm_i915_gem_request *request)
 {
-	const int state = flags & I915_WAIT_INTERRUPTIBLE ?
-		TASK_INTERRUPTIBLE : TASK_UNINTERRUPTIBLE;
-	wait_queue_head_t *q = &request->i915->gpu_error.wait_queue;
-	DEFINE_WAIT(reset);
-	DEFINE_WAIT(wait);
-
-	if (flags & I915_WAIT_LOCKED)
-		add_wait_queue(q, &reset);
-
-	do {
-		prepare_to_wait(&request->execute.wait, &wait, state);
-
-		if (i915_sw_fence_done(&request->execute))
-			break;
-
-		if (flags & I915_WAIT_LOCKED &&
-		    i915_reset_in_progress(&request->i915->gpu_error)) {
-			__set_current_state(TASK_RUNNING);
-			i915_reset(request->i915);
-			reset_wait_queue(q, &reset);
-			continue;
-		}
-
-		if (signal_pending_state(state, current)) {
-			timeout = -ERESTARTSYS;
-			break;
-		}
-
-		if (!timeout) {
-			timeout = -ETIME;
-			break;
-		}
-
-		timeout = io_schedule_timeout(timeout);
-	} while (1);
-	finish_wait(&request->execute.wait, &wait);
-
-	if (flags & I915_WAIT_LOCKED)
-		remove_wait_queue(q, &reset);
+	if (likely(!i915_reset_in_progress(&request->i915->gpu_error)))
+		return false;
 
-	return timeout;
+	__set_current_state(TASK_RUNNING);
+	i915_reset(request->i915);
+	return true;
 }
 
 /**
@@ -1068,7 +1045,9 @@ long i915_wait_request(struct drm_i915_gem_request *req,
 {
 	const int state = flags & I915_WAIT_INTERRUPTIBLE ?
 		TASK_INTERRUPTIBLE : TASK_UNINTERRUPTIBLE;
-	DEFINE_WAIT(reset);
+	wait_queue_head_t *errq = &req->i915->gpu_error.wait_queue;
+	DEFINE_WAIT_FUNC(reset, default_wake_function);
+	DEFINE_WAIT_FUNC(exec, default_wake_function);
 	struct intel_wait wait;
 
 	might_sleep();
@@ -1085,27 +1064,45 @@ long i915_wait_request(struct drm_i915_gem_request *req,
 	if (!timeout)
 		return -ETIME;
 
-	trace_i915_gem_request_wait_begin(req);
+	trace_i915_gem_request_wait_begin(req, flags);
+
+	add_wait_queue(&req->execute, &exec);
+	if (flags & I915_WAIT_LOCKED)
+		add_wait_queue(errq, &reset);
+
+	intel_wait_init(&wait, req);
+
+restart:
+	do {
+		set_current_state(state);
+		if (intel_wait_update_request(&wait, req))
+			break;
+
+		if (flags & I915_WAIT_LOCKED &&
+		    __i915_wait_request_check_and_reset(req))
+			continue;
 
-	if (!i915_sw_fence_done(&req->execute)) {
-		timeout = __i915_request_wait_for_execute(req, flags, timeout);
-		if (timeout < 0)
+		if (signal_pending_state(state, current)) {
+			timeout = -ERESTARTSYS;
 			goto complete;
+		}
 
-		GEM_BUG_ON(!i915_sw_fence_done(&req->execute));
-	}
-	GEM_BUG_ON(!i915_sw_fence_done(&req->submit));
-	GEM_BUG_ON(!req->global_seqno);
+		if (!timeout) {
+			timeout = -ETIME;
+			goto complete;
+		}
+
+		timeout = io_schedule_timeout(timeout);
+	} while (1);
+
+	GEM_BUG_ON(!intel_wait_has_seqno(&wait));
+	GEM_BUG_ON(!i915_sw_fence_signaled(&req->submit));
 
 	/* Optimistic short spin before touching IRQs */
 	if (i915_spin_request(req, state, 5))
 		goto complete;
 
 	set_current_state(state);
-	if (flags & I915_WAIT_LOCKED)
-		add_wait_queue(&req->i915->gpu_error.wait_queue, &reset);
-
-	intel_wait_init(&wait, req->global_seqno);
 	if (intel_engine_add_wait(req->engine, &wait))
 		/* In order to check that we haven't missed the interrupt
 		 * as we enabled it, we need to kick ourselves to do a
@@ -1113,6 +1110,9 @@ long i915_wait_request(struct drm_i915_gem_request *req,
 		 */
 		goto wakeup;
 
+	if (flags & I915_WAIT_LOCKED)
+		__i915_wait_request_check_and_reset(req);
+
 	for (;;) {
 		if (signal_pending_state(state, current)) {
 			timeout = -ERESTARTSYS;
@@ -1126,7 +1126,8 @@ long i915_wait_request(struct drm_i915_gem_request *req,
 
 		timeout = io_schedule_timeout(timeout);
 
-		if (intel_wait_complete(&wait))
+		if (intel_wait_complete(&wait) &&
+		    intel_wait_check_request(&wait, req))
 			break;
 
 		set_current_state(state);
@@ -1151,25 +1152,25 @@ wakeup:
 		 * itself, or indirectly by recovering the GPU).
 		 */
 		if (flags & I915_WAIT_LOCKED &&
-		    i915_reset_in_progress(&req->i915->gpu_error)) {
-			__set_current_state(TASK_RUNNING);
-			i915_reset(req->i915);
-			reset_wait_queue(&req->i915->gpu_error.wait_queue,
-					 &reset);
+		    __i915_wait_request_check_and_reset(req))
 			continue;
-		}
 
 		/* Only spin if we know the GPU is processing this request */
 		if (i915_spin_request(req, state, 2))
 			break;
+
+		if (!intel_wait_check_request(&wait, req)) {
+			intel_engine_remove_wait(req->engine, &wait);
+			goto restart;
+		}
 	}
 
 	intel_engine_remove_wait(req->engine, &wait);
-	if (flags & I915_WAIT_LOCKED)
-		remove_wait_queue(&req->i915->gpu_error.wait_queue, &reset);
-	__set_current_state(TASK_RUNNING);
-
 complete:
+	__set_current_state(TASK_RUNNING);
+	if (flags & I915_WAIT_LOCKED)
+		remove_wait_queue(errq, &reset);
+	remove_wait_queue(&req->execute, &exec);
 	trace_i915_gem_request_wait_end(req);
 
 	return timeout;
@@ -1178,14 +1179,21 @@ complete:
 static void engine_retire_requests(struct intel_engine_cs *engine)
 {
 	struct drm_i915_gem_request *request, *next;
+	u32 seqno = intel_engine_get_seqno(engine);
+	LIST_HEAD(retire);
 
+	spin_lock_irq(&engine->timeline->lock);
 	list_for_each_entry_safe(request, next,
 				 &engine->timeline->requests, link) {
-		if (!__i915_gem_request_completed(request))
-			return;
+		if (!i915_seqno_passed(seqno, request->global_seqno))
+			break;
 
-		i915_gem_request_retire(request);
+		list_move_tail(&request->link, &retire);
 	}
+	spin_unlock_irq(&engine->timeline->lock);
+
+	list_for_each_entry_safe(request, next, &retire, link)
+		i915_gem_request_retire(request);
 }
 
 void i915_gem_retire_requests(struct drm_i915_private *dev_priv)
@@ -1201,3 +1209,8 @@ void i915_gem_retire_requests(struct drm_i915_private *dev_priv)
 	for_each_engine(engine, dev_priv, id)
 		engine_retire_requests(engine);
 }
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/mock_request.c"
+#include "selftests/i915_gem_request.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_gem_request.h b/drivers/gpu/drm/i915/i915_gem_request.h
index ea511f06efaf..5018e55922f0 100644
--- a/drivers/gpu/drm/i915/i915_gem_request.h
+++ b/drivers/gpu/drm/i915/i915_gem_request.h
@@ -32,10 +32,12 @@
 
 struct drm_file;
 struct drm_i915_gem_object;
+struct drm_i915_gem_request;
 
 struct intel_wait {
 	struct rb_node node;
 	struct task_struct *tsk;
+	struct drm_i915_gem_request *request;
 	u32 seqno;
 };
 
@@ -119,18 +121,10 @@ struct drm_i915_gem_request {
 	 * The submit fence is used to await upon all of the request's
 	 * dependencies. When it is signaled, the request is ready to run.
 	 * It is used by the driver to then queue the request for execution.
-	 *
-	 * The execute fence is used to signal when the request has been
-	 * sent to hardware.
-	 *
-	 * It is illegal for the submit fence of one request to wait upon the
-	 * execute fence of an earlier request. It should be sufficient to
-	 * wait upon the submit fence of the earlier request.
 	 */
 	struct i915_sw_fence submit;
-	struct i915_sw_fence execute;
 	wait_queue_t submitq;
-	wait_queue_t execq;
+	wait_queue_head_t execute;
 
 	/* A list of everyone we wait upon, and everyone who waits upon us.
 	 * Even though we will not be submitted to the hardware before the
@@ -143,13 +137,12 @@ struct drm_i915_gem_request {
 	struct i915_priotree priotree;
 	struct i915_dependency dep;
 
-	u32 global_seqno;
-
-	/** GEM sequence number associated with the previous request,
-	 * when the HWS breadcrumb is equal to this the GPU is processing
-	 * this request.
+	/** GEM sequence number associated with this request on the
+	 * global execution timeline. It is zero when the request is not
+	 * on the HW queue (i.e. not on the engine timeline list).
+	 * Its value is guarded by the timeline spinlock.
 	 */
-	u32 previous_seqno;
+	u32 global_seqno;
 
 	/** Position in the ring of the start of the request */
 	u32 head;
@@ -187,7 +180,7 @@ struct drm_i915_gem_request {
 
 	struct drm_i915_file_private *file_priv;
 	/** file_priv list entry for this request */
-	struct list_head client_list;
+	struct list_head client_link;
 };
 
 extern const struct dma_fence_ops i915_fence_ops;
@@ -200,8 +193,6 @@ static inline bool dma_fence_is_i915(const struct dma_fence *fence)
 struct drm_i915_gem_request * __must_check
 i915_gem_request_alloc(struct intel_engine_cs *engine,
 		       struct i915_gem_context *ctx);
-int i915_gem_request_add_to_client(struct drm_i915_gem_request *req,
-				   struct drm_file *file);
 void i915_gem_request_retire_upto(struct drm_i915_gem_request *req);
 
 static inline struct drm_i915_gem_request *
@@ -243,6 +234,30 @@ static inline void i915_gem_request_assign(struct drm_i915_gem_request **pdst,
 	*pdst = src;
 }
 
+/**
+ * i915_gem_request_global_seqno - report the current global seqno
+ * @request - the request
+ *
+ * A request is assigned a global seqno only when it is on the hardware
+ * execution queue. The global seqno can be used to maintain a list of
+ * requests on the same engine in retirement order, for example for
+ * constructing a priority queue for waiting. Prior to its execution, or
+ * if it is subsequently removed in the event of preemption, its global
+ * seqno is zero. As both insertion and removal from the execution queue
+ * may operate in IRQ context, it is not guarded by the usual struct_mutex
+ * BKL. Instead those relying on the global seqno must be prepared for its
+ * value to change between reads. Only when the request is complete can
+ * the global seqno be stable (due to the memory barriers on submitting
+ * the commands to the hardware to write the breadcrumb, if the HWS shows
+ * that it has passed the global seqno and the global seqno is unchanged
+ * after the read, it is indeed complete).
+ */
+static u32
+i915_gem_request_global_seqno(const struct drm_i915_gem_request *request)
+{
+	return READ_ONCE(request->global_seqno);
+}
+
 int
 i915_gem_request_await_object(struct drm_i915_gem_request *to,
 			      struct drm_i915_gem_object *obj,
@@ -259,6 +274,9 @@ void __i915_add_request(struct drm_i915_gem_request *req, bool flush_caches);
 void __i915_gem_request_submit(struct drm_i915_gem_request *request);
 void i915_gem_request_submit(struct drm_i915_gem_request *request);
 
+void __i915_gem_request_unsubmit(struct drm_i915_gem_request *request);
+void i915_gem_request_unsubmit(struct drm_i915_gem_request *request);
+
 struct intel_rps_client;
 #define NO_WAITBOOST ERR_PTR(-1)
 #define IS_RPS_CLIENT(p) (!IS_ERR(p))
@@ -283,46 +301,55 @@ static inline bool i915_seqno_passed(u32 seq1, u32 seq2)
 }
 
 static inline bool
-__i915_gem_request_started(const struct drm_i915_gem_request *req)
+__i915_gem_request_started(const struct drm_i915_gem_request *req, u32 seqno)
 {
-	GEM_BUG_ON(!req->global_seqno);
+	GEM_BUG_ON(!seqno);
 	return i915_seqno_passed(intel_engine_get_seqno(req->engine),
-				 req->previous_seqno);
+				 seqno - 1);
 }
 
 static inline bool
 i915_gem_request_started(const struct drm_i915_gem_request *req)
 {
-	if (!req->global_seqno)
+	u32 seqno;
+
+	seqno = i915_gem_request_global_seqno(req);
+	if (!seqno)
 		return false;
 
-	return __i915_gem_request_started(req);
+	return __i915_gem_request_started(req, seqno);
 }
 
 static inline bool
-__i915_gem_request_completed(const struct drm_i915_gem_request *req)
+__i915_gem_request_completed(const struct drm_i915_gem_request *req, u32 seqno)
 {
-	GEM_BUG_ON(!req->global_seqno);
-	return i915_seqno_passed(intel_engine_get_seqno(req->engine),
-				 req->global_seqno);
+	GEM_BUG_ON(!seqno);
+	return i915_seqno_passed(intel_engine_get_seqno(req->engine), seqno) &&
+		seqno == i915_gem_request_global_seqno(req);
 }
 
 static inline bool
 i915_gem_request_completed(const struct drm_i915_gem_request *req)
 {
-	if (!req->global_seqno)
+	u32 seqno;
+
+	seqno = i915_gem_request_global_seqno(req);
+	if (!seqno)
 		return false;
 
-	return __i915_gem_request_completed(req);
+	return __i915_gem_request_completed(req, seqno);
 }
 
 bool __i915_spin_request(const struct drm_i915_gem_request *request,
-			 int state, unsigned long timeout_us);
+			 u32 seqno, int state, unsigned long timeout_us);
 static inline bool i915_spin_request(const struct drm_i915_gem_request *request,
 				     int state, unsigned long timeout_us)
 {
-	return (__i915_gem_request_started(request) &&
-		__i915_spin_request(request, state, timeout_us));
+	u32 seqno;
+
+	seqno = i915_gem_request_global_seqno(request);
+	return (__i915_gem_request_started(request, seqno) &&
+		__i915_spin_request(request, seqno, state, timeout_us));
 }
 
 /* We treat requests as fences. This is not be to confused with our
diff --git a/drivers/gpu/drm/i915/i915_gem_shrinker.c b/drivers/gpu/drm/i915/i915_gem_shrinker.c
index 401006b4c6a3..006a8b908f77 100644
--- a/drivers/gpu/drm/i915/i915_gem_shrinker.c
+++ b/drivers/gpu/drm/i915/i915_gem_shrinker.c
@@ -207,7 +207,7 @@ i915_gem_shrink(struct drm_i915_private *dev_priv,
 
 			if (!(flags & I915_SHRINK_ACTIVE) &&
 			    (i915_gem_object_is_active(obj) ||
-			     obj->framebuffer_references))
+			     i915_gem_object_is_framebuffer(obj)))
 				continue;
 
 			if (!can_release_pages(obj))
@@ -259,10 +259,13 @@ unsigned long i915_gem_shrink_all(struct drm_i915_private *dev_priv)
 {
 	unsigned long freed;
 
+	intel_runtime_pm_get(dev_priv);
 	freed = i915_gem_shrink(dev_priv, -1UL,
 				I915_SHRINK_BOUND |
 				I915_SHRINK_UNBOUND |
 				I915_SHRINK_ACTIVE);
+	intel_runtime_pm_put(dev_priv);
+
 	rcu_barrier(); /* wait until our RCU delayed slab frees are completed */
 
 	return freed;
@@ -380,9 +383,7 @@ i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)
 	if (!i915_gem_shrinker_lock_uninterruptible(dev_priv, &slu, 5000))
 		return NOTIFY_DONE;
 
-	intel_runtime_pm_get(dev_priv);
 	freed_pages = i915_gem_shrink_all(dev_priv);
-	intel_runtime_pm_put(dev_priv);
 
 	/* Because we may be allocating inside our own driver, we cannot
 	 * assert that there are no objects with pinned pages that are not
diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c
index 9673bcc3b6ad..f3abdc27c5dd 100644
--- a/drivers/gpu/drm/i915/i915_gem_stolen.c
+++ b/drivers/gpu/drm/i915/i915_gem_stolen.c
@@ -79,12 +79,12 @@ void i915_gem_stolen_remove_node(struct drm_i915_private *dev_priv,
 	mutex_unlock(&dev_priv->mm.stolen_lock);
 }
 
-static unsigned long i915_stolen_to_physical(struct drm_i915_private *dev_priv)
+static dma_addr_t i915_stolen_to_dma(struct drm_i915_private *dev_priv)
 {
 	struct pci_dev *pdev = dev_priv->drm.pdev;
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	struct resource *r;
-	u32 base;
+	dma_addr_t base;
 
 	/* Almost universally we can find the Graphics Base of Stolen Memory
 	 * at register BSM (0x5c) in the igfx configuration space. On a few
@@ -189,14 +189,14 @@ static unsigned long i915_stolen_to_physical(struct drm_i915_private *dev_priv)
 		base = tom - tseg_size - ggtt->stolen_size;
 	}
 
-	if (base == 0)
+	if (base == 0 || add_overflows(base, ggtt->stolen_size))
 		return 0;
 
 	/* make sure we don't clobber the GTT if it's within stolen memory */
 	if (INTEL_GEN(dev_priv) <= 4 &&
 	    !IS_G33(dev_priv) && !IS_PINEVIEW(dev_priv) && !IS_G4X(dev_priv)) {
 		struct {
-			u32 start, end;
+			dma_addr_t start, end;
 		} stolen[2] = {
 			{ .start = base, .end = base + ggtt->stolen_size, },
 			{ .start = base, .end = base + ggtt->stolen_size, },
@@ -228,11 +228,13 @@ static unsigned long i915_stolen_to_physical(struct drm_i915_private *dev_priv)
 
 		if (stolen[0].start != stolen[1].start ||
 		    stolen[0].end != stolen[1].end) {
+			dma_addr_t end = base + ggtt->stolen_size - 1;
+
 			DRM_DEBUG_KMS("GTT within stolen memory at 0x%llx-0x%llx\n",
 				      (unsigned long long)ggtt_start,
 				      (unsigned long long)ggtt_end - 1);
-			DRM_DEBUG_KMS("Stolen memory adjusted to 0x%x-0x%x\n",
-				      base, base + (u32)ggtt->stolen_size - 1);
+			DRM_DEBUG_KMS("Stolen memory adjusted to %pad - %pad\n",
+				      &base, &end);
 		}
 	}
 
@@ -261,8 +263,10 @@ static unsigned long i915_stolen_to_physical(struct drm_i915_private *dev_priv)
 		 * range. Apparently this works.
 		 */
 		if (r == NULL && !IS_GEN3(dev_priv)) {
-			DRM_ERROR("conflict detected with stolen region: [0x%08x - 0x%08x]\n",
-				  base, base + (uint32_t)ggtt->stolen_size);
+			dma_addr_t end = base + ggtt->stolen_size;
+
+			DRM_ERROR("conflict detected with stolen region: [%pad - %pad]\n",
+				  &base, &end);
 			base = 0;
 		}
 	}
@@ -281,13 +285,13 @@ void i915_gem_cleanup_stolen(struct drm_device *dev)
 }
 
 static void g4x_get_stolen_reserved(struct drm_i915_private *dev_priv,
-				    phys_addr_t *base, u32 *size)
+				    dma_addr_t *base, u32 *size)
 {
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	uint32_t reg_val = I915_READ(IS_GM45(dev_priv) ?
 				     CTG_STOLEN_RESERVED :
 				     ELK_STOLEN_RESERVED);
-	phys_addr_t stolen_top = dev_priv->mm.stolen_base + ggtt->stolen_size;
+	dma_addr_t stolen_top = dev_priv->mm.stolen_base + ggtt->stolen_size;
 
 	*base = (reg_val & G4X_STOLEN_RESERVED_ADDR2_MASK) << 16;
 
@@ -304,7 +308,7 @@ static void g4x_get_stolen_reserved(struct drm_i915_private *dev_priv,
 }
 
 static void gen6_get_stolen_reserved(struct drm_i915_private *dev_priv,
-				     phys_addr_t *base, u32 *size)
+				     dma_addr_t *base, u32 *size)
 {
 	uint32_t reg_val = I915_READ(GEN6_STOLEN_RESERVED);
 
@@ -330,7 +334,7 @@ static void gen6_get_stolen_reserved(struct drm_i915_private *dev_priv,
 }
 
 static void gen7_get_stolen_reserved(struct drm_i915_private *dev_priv,
-				     phys_addr_t *base, u32 *size)
+				     dma_addr_t *base, u32 *size)
 {
 	uint32_t reg_val = I915_READ(GEN6_STOLEN_RESERVED);
 
@@ -350,7 +354,7 @@ static void gen7_get_stolen_reserved(struct drm_i915_private *dev_priv,
 }
 
 static void chv_get_stolen_reserved(struct drm_i915_private *dev_priv,
-				    phys_addr_t *base, u32 *size)
+				    dma_addr_t *base, u32 *size)
 {
 	uint32_t reg_val = I915_READ(GEN6_STOLEN_RESERVED);
 
@@ -376,11 +380,11 @@ static void chv_get_stolen_reserved(struct drm_i915_private *dev_priv,
 }
 
 static void bdw_get_stolen_reserved(struct drm_i915_private *dev_priv,
-				    phys_addr_t *base, u32 *size)
+				    dma_addr_t *base, u32 *size)
 {
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	uint32_t reg_val = I915_READ(GEN6_STOLEN_RESERVED);
-	phys_addr_t stolen_top;
+	dma_addr_t stolen_top;
 
 	stolen_top = dev_priv->mm.stolen_base + ggtt->stolen_size;
 
@@ -399,7 +403,7 @@ static void bdw_get_stolen_reserved(struct drm_i915_private *dev_priv,
 int i915_gem_init_stolen(struct drm_i915_private *dev_priv)
 {
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
-	phys_addr_t reserved_base, stolen_top;
+	dma_addr_t reserved_base, stolen_top;
 	u32 reserved_total, reserved_size;
 	u32 stolen_usable_start;
 
@@ -420,7 +424,7 @@ int i915_gem_init_stolen(struct drm_i915_private *dev_priv)
 	if (ggtt->stolen_size == 0)
 		return 0;
 
-	dev_priv->mm.stolen_base = i915_stolen_to_physical(dev_priv);
+	dev_priv->mm.stolen_base = i915_stolen_to_dma(dev_priv);
 	if (dev_priv->mm.stolen_base == 0)
 		return 0;
 
@@ -469,8 +473,8 @@ int i915_gem_init_stolen(struct drm_i915_private *dev_priv)
 
 	if (reserved_base < dev_priv->mm.stolen_base ||
 	    reserved_base + reserved_size > stolen_top) {
-		phys_addr_t reserved_top = reserved_base + reserved_size;
-		DRM_DEBUG_KMS("Stolen reserved area [%pa - %pa] outside stolen memory [%pa - %pa]\n",
+		dma_addr_t reserved_top = reserved_base + reserved_size;
+		DRM_DEBUG_KMS("Stolen reserved area [%pad - %pad] outside stolen memory [%pad - %pad]\n",
 			      &reserved_base, &reserved_top,
 			      &dev_priv->mm.stolen_base, &stolen_top);
 		return 0;
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index 974ac08df473..a0d6d4317a49 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -158,13 +158,8 @@ i915_tiling_ok(struct drm_i915_gem_object *obj,
 		if (stride > 8192)
 			return false;
 
-		if (IS_GEN3(i915)) {
-			if (obj->base.size > I830_FENCE_MAX_SIZE_VAL << 20)
-				return false;
-		} else {
-			if (obj->base.size > I830_FENCE_MAX_SIZE_VAL << 19)
-				return false;
-		}
+		if (!is_power_of_2(stride))
+			return false;
 	}
 
 	if (IS_GEN2(i915) ||
@@ -176,12 +171,7 @@ i915_tiling_ok(struct drm_i915_gem_object *obj,
 	if (!stride || !IS_ALIGNED(stride, tile_width))
 		return false;
 
-	/* 965+ just needs multiples of tile width */
-	if (INTEL_GEN(i915) >= 4)
-		return true;
-
-	/* Pre-965 needs power of two tile widths */
-	return is_power_of_2(stride);
+	return true;
 }
 
 static bool i915_vma_fence_prepare(struct i915_vma *vma,
@@ -248,7 +238,7 @@ i915_gem_object_set_tiling(struct drm_i915_gem_object *obj,
 	if ((tiling | stride) == obj->tiling_and_stride)
 		return 0;
 
-	if (obj->framebuffer_references)
+	if (i915_gem_object_is_framebuffer(obj))
 		return -EBUSY;
 
 	/* We need to rebind the object if its current allocation
@@ -268,6 +258,12 @@ i915_gem_object_set_tiling(struct drm_i915_gem_object *obj,
 	if (err)
 		return err;
 
+	i915_gem_object_lock(obj);
+	if (i915_gem_object_is_framebuffer(obj)) {
+		i915_gem_object_unlock(obj);
+		return -EBUSY;
+	}
+
 	/* If the memory has unknown (i.e. varying) swizzling, we pin the
 	 * pages to prevent them being swapped out and causing corruption
 	 * due to the change in swizzling.
@@ -304,6 +300,7 @@ i915_gem_object_set_tiling(struct drm_i915_gem_object *obj,
 	}
 
 	obj->tiling_and_stride = tiling | stride;
+	i915_gem_object_unlock(obj);
 
 	/* Force the fence to be reacquired for GTT access */
 	i915_gem_release_mmap(obj);
diff --git a/drivers/gpu/drm/i915/i915_gem_timeline.h b/drivers/gpu/drm/i915/i915_gem_timeline.h
index f2e51f42cc2f..6c53e14cab2a 100644
--- a/drivers/gpu/drm/i915/i915_gem_timeline.h
+++ b/drivers/gpu/drm/i915/i915_gem_timeline.h
@@ -33,7 +33,13 @@ struct i915_gem_timeline;
 
 struct intel_timeline {
 	u64 fence_context;
-	u32 last_submitted_seqno;
+	u32 seqno;
+
+	/**
+	 * Count of outstanding requests, from the time they are constructed
+	 * to the moment they are retired. Loosely coupled to hardware.
+	 */
+	u32 inflight_seqnos;
 
 	spinlock_t lock;
 
@@ -56,7 +62,6 @@ struct intel_timeline {
 
 struct i915_gem_timeline {
 	struct list_head link;
-	atomic_t seqno;
 
 	struct drm_i915_private *i915;
 	const char *name;
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
index 9cd22cda17af..8effc59f5cb5 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.c
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -342,7 +342,7 @@ static void print_error_buffers(struct drm_i915_error_state_buf *m,
 }
 
 static void error_print_instdone(struct drm_i915_error_state_buf *m,
-				 struct drm_i915_error_engine *ee)
+				 const struct drm_i915_error_engine *ee)
 {
 	int slice;
 	int subslice;
@@ -372,7 +372,7 @@ static void error_print_instdone(struct drm_i915_error_state_buf *m,
 
 static void error_print_request(struct drm_i915_error_state_buf *m,
 				const char *prefix,
-				struct drm_i915_error_request *erq)
+				const struct drm_i915_error_request *erq)
 {
 	if (!erq->seqno)
 		return;
@@ -384,8 +384,17 @@ static void error_print_request(struct drm_i915_error_state_buf *m,
 		   erq->head, erq->tail);
 }
 
+static void error_print_context(struct drm_i915_error_state_buf *m,
+				const char *header,
+				const struct drm_i915_error_context *ctx)
+{
+	err_printf(m, "%s%s[%d] user_handle %d hw_id %d, ban score %d guilty %d active %d\n",
+		   header, ctx->comm, ctx->pid, ctx->handle, ctx->hw_id,
+		   ctx->ban_score, ctx->guilty, ctx->active);
+}
+
 static void error_print_engine(struct drm_i915_error_state_buf *m,
-			       struct drm_i915_error_engine *ee)
+			       const struct drm_i915_error_engine *ee)
 {
 	err_printf(m, "%s command stream:\n", engine_str(ee->engine_id));
 	err_printf(m, "  START: 0x%08x\n", ee->start);
@@ -457,6 +466,7 @@ static void error_print_engine(struct drm_i915_error_state_buf *m,
 
 	error_print_request(m, "  ELSP[0]: ", &ee->execlist[0]);
 	error_print_request(m, "  ELSP[1]: ", &ee->execlist[1]);
+	error_print_context(m, "  Active context: ", &ee->context);
 }
 
 void i915_error_printf(struct drm_i915_error_state_buf *e, const char *f, ...)
@@ -536,21 +546,57 @@ static void err_print_capabilities(struct drm_i915_error_state_buf *m,
 #undef PRINT_FLAG
 }
 
+static __always_inline void err_print_param(struct drm_i915_error_state_buf *m,
+					    const char *name,
+					    const char *type,
+					    const void *x)
+{
+	if (!__builtin_strcmp(type, "bool"))
+		err_printf(m, "i915.%s=%s\n", name, yesno(*(const bool *)x));
+	else if (!__builtin_strcmp(type, "int"))
+		err_printf(m, "i915.%s=%d\n", name, *(const int *)x);
+	else if (!__builtin_strcmp(type, "unsigned int"))
+		err_printf(m, "i915.%s=%u\n", name, *(const unsigned int *)x);
+	else if (!__builtin_strcmp(type, "char *"))
+		err_printf(m, "i915.%s=%s\n", name, *(const char **)x);
+	else
+		BUILD_BUG();
+}
+
+static void err_print_params(struct drm_i915_error_state_buf *m,
+			     const struct i915_params *p)
+{
+#define PRINT(T, x) err_print_param(m, #x, #T, &p->x);
+	I915_PARAMS_FOR_EACH(PRINT);
+#undef PRINT
+}
+
+static void err_print_pciid(struct drm_i915_error_state_buf *m,
+			    struct drm_i915_private *i915)
+{
+	struct pci_dev *pdev = i915->drm.pdev;
+
+	err_printf(m, "PCI ID: 0x%04x\n", pdev->device);
+	err_printf(m, "PCI Revision: 0x%02x\n", pdev->revision);
+	err_printf(m, "PCI Subsystem: %04x:%04x\n",
+		   pdev->subsystem_vendor,
+		   pdev->subsystem_device);
+}
+
 int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
-			    const struct i915_error_state_file_priv *error_priv)
+			    const struct i915_gpu_state *error)
 {
-	struct drm_i915_private *dev_priv = error_priv->i915;
-	struct pci_dev *pdev = dev_priv->drm.pdev;
-	struct drm_i915_error_state *error = error_priv->error;
+	struct drm_i915_private *dev_priv = m->i915;
 	struct drm_i915_error_object *obj;
 	int i, j;
 
 	if (!error) {
-		err_printf(m, "no error state collected\n");
-		goto out;
+		err_printf(m, "No error state collected\n");
+		return 0;
 	}
 
-	err_printf(m, "%s\n", error->error_msg);
+	if (*error->error_msg)
+		err_printf(m, "%s\n", error->error_msg);
 	err_printf(m, "Kernel: " UTS_RELEASE "\n");
 	err_printf(m, "Time: %ld s %ld us\n",
 		   error->time.tv_sec, error->time.tv_usec);
@@ -558,26 +604,22 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 		   error->boottime.tv_sec, error->boottime.tv_usec);
 	err_printf(m, "Uptime: %ld s %ld us\n",
 		   error->uptime.tv_sec, error->uptime.tv_usec);
-	err_print_capabilities(m, &error->device_info);
 
 	for (i = 0; i < ARRAY_SIZE(error->engine); i++) {
 		if (error->engine[i].hangcheck_stalled &&
-		    error->engine[i].pid != -1) {
-			err_printf(m, "Active process (on ring %s): %s [%d], context bans %d\n",
+		    error->engine[i].context.pid) {
+			err_printf(m, "Active process (on ring %s): %s [%d], score %d\n",
 				   engine_str(i),
-				   error->engine[i].comm,
-				   error->engine[i].pid,
-				   error->engine[i].context_bans);
+				   error->engine[i].context.comm,
+				   error->engine[i].context.pid,
+				   error->engine[i].context.ban_score);
 		}
 	}
 	err_printf(m, "Reset count: %u\n", error->reset_count);
 	err_printf(m, "Suspend count: %u\n", error->suspend_count);
 	err_printf(m, "Platform: %s\n", intel_platform_name(error->device_info.platform));
-	err_printf(m, "PCI ID: 0x%04x\n", pdev->device);
-	err_printf(m, "PCI Revision: 0x%02x\n", pdev->revision);
-	err_printf(m, "PCI Subsystem: %04x:%04x\n",
-		   pdev->subsystem_vendor,
-		   pdev->subsystem_device);
+	err_print_pciid(m, error->i915);
+
 	err_printf(m, "IOMMU enabled?: %d\n", error->iommu);
 
 	if (HAS_CSR(dev_priv)) {
@@ -590,21 +632,20 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 			   CSR_VERSION_MINOR(csr->version));
 	}
 
+	err_printf(m, "GT awake: %s\n", yesno(error->awake));
+	err_printf(m, "RPM wakelock: %s\n", yesno(error->wakelock));
+	err_printf(m, "PM suspended: %s\n", yesno(error->suspended));
 	err_printf(m, "EIR: 0x%08x\n", error->eir);
 	err_printf(m, "IER: 0x%08x\n", error->ier);
-	if (INTEL_GEN(dev_priv) >= 8) {
-		for (i = 0; i < 4; i++)
-			err_printf(m, "GTIER gt %d: 0x%08x\n", i,
-				   error->gtier[i]);
-	} else if (HAS_PCH_SPLIT(dev_priv) || IS_VALLEYVIEW(dev_priv))
-		err_printf(m, "GTIER: 0x%08x\n", error->gtier[0]);
+	for (i = 0; i < error->ngtier; i++)
+		err_printf(m, "GTIER[%d]: 0x%08x\n", i, error->gtier[i]);
 	err_printf(m, "PGTBL_ER: 0x%08x\n", error->pgtbl_er);
 	err_printf(m, "FORCEWAKE: 0x%08x\n", error->forcewake);
 	err_printf(m, "DERRMR: 0x%08x\n", error->derrmr);
 	err_printf(m, "CCID: 0x%08x\n", error->ccid);
 	err_printf(m, "Missed interrupts: 0x%08lx\n", dev_priv->gpu_error.missed_irq_rings);
 
-	for (i = 0; i < dev_priv->num_fence_regs; i++)
+	for (i = 0; i < error->nfence; i++)
 		err_printf(m, "  fence[%d] = %08llx\n", i, error->fence[i]);
 
 	if (INTEL_GEN(dev_priv) >= 6) {
@@ -653,16 +694,18 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 			    error->pinned_bo_count);
 
 	for (i = 0; i < ARRAY_SIZE(error->engine); i++) {
-		struct drm_i915_error_engine *ee = &error->engine[i];
+		const struct drm_i915_error_engine *ee = &error->engine[i];
 
 		obj = ee->batchbuffer;
 		if (obj) {
 			err_puts(m, dev_priv->engine[i]->name);
-			if (ee->pid != -1)
-				err_printf(m, " (submitted by %s [%d], bans %d)",
-					   ee->comm,
-					   ee->pid,
-					   ee->context_bans);
+			if (ee->context.pid)
+				err_printf(m, " (submitted by %s [%d], ctx %d [%d], score %d)",
+					   ee->context.comm,
+					   ee->context.pid,
+					   ee->context.handle,
+					   ee->context.hw_id,
+					   ee->context.ban_score);
 			err_printf(m, " --- gtt_offset = 0x%08x %08x\n",
 				   upper_32_bits(obj->gtt_offset),
 				   lower_32_bits(obj->gtt_offset));
@@ -716,9 +759,11 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 		intel_overlay_print_error_state(m, error->overlay);
 
 	if (error->display)
-		intel_display_print_error_state(m, dev_priv, error->display);
+		intel_display_print_error_state(m, error->display);
+
+	err_print_capabilities(m, &error->device_info);
+	err_print_params(m, &error->params);
 
-out:
 	if (m->bytes == 0 && m->err)
 		return m->err;
 
@@ -770,10 +815,16 @@ static void i915_error_object_free(struct drm_i915_error_object *obj)
 	kfree(obj);
 }
 
-static void i915_error_state_free(struct kref *error_ref)
+static __always_inline void free_param(const char *type, void *x)
+{
+	if (!__builtin_strcmp(type, "char *"))
+		kfree(*(void **)x);
+}
+
+void __i915_gpu_state_free(struct kref *error_ref)
 {
-	struct drm_i915_error_state *error = container_of(error_ref,
-							  typeof(*error), ref);
+	struct i915_gpu_state *error =
+		container_of(error_ref, typeof(*error), ref);
 	int i;
 
 	for (i = 0; i < ARRAY_SIZE(error->engine); i++) {
@@ -800,6 +851,11 @@ static void i915_error_state_free(struct kref *error_ref)
 
 	kfree(error->overlay);
 	kfree(error->display);
+
+#define FREE(T, x) free_param(#T, &error->params.x);
+	I915_PARAMS_FOR_EACH(FREE);
+#undef FREE
+
 	kfree(error);
 }
 
@@ -938,7 +994,7 @@ static u32 capture_error_bo(struct drm_i915_error_buffer *err,
  * It's only a small step better than a random number in its current form.
  */
 static uint32_t i915_error_generate_code(struct drm_i915_private *dev_priv,
-					 struct drm_i915_error_state *error,
+					 struct i915_gpu_state *error,
 					 int *engine_id)
 {
 	uint32_t error_code = 0;
@@ -963,20 +1019,21 @@ static uint32_t i915_error_generate_code(struct drm_i915_private *dev_priv,
 }
 
 static void i915_gem_record_fences(struct drm_i915_private *dev_priv,
-				   struct drm_i915_error_state *error)
+				   struct i915_gpu_state *error)
 {
 	int i;
 
-	if (IS_GEN3(dev_priv) || IS_GEN2(dev_priv)) {
+	if (INTEL_GEN(dev_priv) >= 6) {
 		for (i = 0; i < dev_priv->num_fence_regs; i++)
-			error->fence[i] = I915_READ(FENCE_REG(i));
-	} else if (IS_GEN5(dev_priv) || IS_GEN4(dev_priv)) {
+			error->fence[i] = I915_READ64(FENCE_REG_GEN6_LO(i));
+	} else if (INTEL_GEN(dev_priv) >= 4) {
 		for (i = 0; i < dev_priv->num_fence_regs; i++)
 			error->fence[i] = I915_READ64(FENCE_REG_965_LO(i));
-	} else if (INTEL_GEN(dev_priv) >= 6) {
+	} else {
 		for (i = 0; i < dev_priv->num_fence_regs; i++)
-			error->fence[i] = I915_READ64(FENCE_REG_GEN6_LO(i));
+			error->fence[i] = I915_READ(FENCE_REG(i));
 	}
+	error->nfence = i;
 }
 
 static inline u32
@@ -1000,7 +1057,7 @@ gen8_engine_sync_index(struct intel_engine_cs *engine,
 	return idx;
 }
 
-static void gen8_record_semaphore_state(struct drm_i915_error_state *error,
+static void gen8_record_semaphore_state(struct i915_gpu_state *error,
 					struct intel_engine_cs *engine,
 					struct drm_i915_error_engine *ee)
 {
@@ -1054,7 +1111,7 @@ static void error_record_engine_waiters(struct intel_engine_cs *engine,
 	if (RB_EMPTY_ROOT(&b->waiters))
 		return;
 
-	if (!spin_trylock_irq(&b->lock)) {
+	if (!spin_trylock_irq(&b->rb_lock)) {
 		ee->waiters = ERR_PTR(-EDEADLK);
 		return;
 	}
@@ -1062,7 +1119,7 @@ static void error_record_engine_waiters(struct intel_engine_cs *engine,
 	count = 0;
 	for (rb = rb_first(&b->waiters); rb != NULL; rb = rb_next(rb))
 		count++;
-	spin_unlock_irq(&b->lock);
+	spin_unlock_irq(&b->rb_lock);
 
 	waiter = NULL;
 	if (count)
@@ -1072,7 +1129,7 @@ static void error_record_engine_waiters(struct intel_engine_cs *engine,
 	if (!waiter)
 		return;
 
-	if (!spin_trylock_irq(&b->lock)) {
+	if (!spin_trylock_irq(&b->rb_lock)) {
 		kfree(waiter);
 		ee->waiters = ERR_PTR(-EDEADLK);
 		return;
@@ -1080,7 +1137,7 @@ static void error_record_engine_waiters(struct intel_engine_cs *engine,
 
 	ee->waiters = waiter;
 	for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) {
-		struct intel_wait *w = container_of(rb, typeof(*w), node);
+		struct intel_wait *w = rb_entry(rb, typeof(*w), node);
 
 		strcpy(waiter->comm, w->tsk->comm);
 		waiter->pid = w->tsk->pid;
@@ -1090,10 +1147,10 @@ static void error_record_engine_waiters(struct intel_engine_cs *engine,
 		if (++ee->num_waiters == count)
 			break;
 	}
-	spin_unlock_irq(&b->lock);
+	spin_unlock_irq(&b->rb_lock);
 }
 
-static void error_record_engine_registers(struct drm_i915_error_state *error,
+static void error_record_engine_registers(struct i915_gpu_state *error,
 					  struct intel_engine_cs *engine,
 					  struct drm_i915_error_engine *ee)
 {
@@ -1267,8 +1324,30 @@ static void error_record_engine_execlists(struct intel_engine_cs *engine,
 				       &ee->execlist[n]);
 }
 
+static void record_context(struct drm_i915_error_context *e,
+			   struct i915_gem_context *ctx)
+{
+	if (ctx->pid) {
+		struct task_struct *task;
+
+		rcu_read_lock();
+		task = pid_task(ctx->pid, PIDTYPE_PID);
+		if (task) {
+			strcpy(e->comm, task->comm);
+			e->pid = task->pid;
+		}
+		rcu_read_unlock();
+	}
+
+	e->handle = ctx->user_handle;
+	e->hw_id = ctx->hw_id;
+	e->ban_score = ctx->ban_score;
+	e->guilty = ctx->guilty_count;
+	e->active = ctx->active_count;
+}
+
 static void i915_gem_record_rings(struct drm_i915_private *dev_priv,
-				  struct drm_i915_error_state *error)
+				  struct i915_gpu_state *error)
 {
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	int i;
@@ -1281,7 +1360,6 @@ static void i915_gem_record_rings(struct drm_i915_private *dev_priv,
 		struct drm_i915_error_engine *ee = &error->engine[i];
 		struct drm_i915_gem_request *request;
 
-		ee->pid = -1;
 		ee->engine_id = -1;
 
 		if (!engine)
@@ -1296,11 +1374,12 @@ static void i915_gem_record_rings(struct drm_i915_private *dev_priv,
 		request = i915_gem_find_active_request(engine);
 		if (request) {
 			struct intel_ring *ring;
-			struct pid *pid;
 
 			ee->vm = request->ctx->ppgtt ?
 				&request->ctx->ppgtt->base : &ggtt->base;
 
+			record_context(&ee->context, request->ctx);
+
 			/* We need to copy these to an anonymous buffer
 			 * as the simplest method to avoid being overwritten
 			 * by userspace.
@@ -1318,19 +1397,6 @@ static void i915_gem_record_rings(struct drm_i915_private *dev_priv,
 				i915_error_object_create(dev_priv,
 							 request->ctx->engine[i].state);
 
-			pid = request->ctx->pid;
-			if (pid) {
-				struct task_struct *task;
-
-				rcu_read_lock();
-				task = pid_task(pid, PIDTYPE_PID);
-				if (task) {
-					strcpy(ee->comm, task->comm);
-					ee->pid = task->pid;
-				}
-				rcu_read_unlock();
-			}
-
 			error->simulated |=
 				i915_gem_context_no_error_capture(request->ctx);
 
@@ -1357,7 +1423,7 @@ static void i915_gem_record_rings(struct drm_i915_private *dev_priv,
 }
 
 static void i915_gem_capture_vm(struct drm_i915_private *dev_priv,
-				struct drm_i915_error_state *error,
+				struct i915_gpu_state *error,
 				struct i915_address_space *vm,
 				int idx)
 {
@@ -1383,7 +1449,7 @@ static void i915_gem_capture_vm(struct drm_i915_private *dev_priv,
 }
 
 static void i915_capture_active_buffers(struct drm_i915_private *dev_priv,
-					struct drm_i915_error_state *error)
+					struct i915_gpu_state *error)
 {
 	int cnt = 0, i, j;
 
@@ -1408,7 +1474,7 @@ static void i915_capture_active_buffers(struct drm_i915_private *dev_priv,
 }
 
 static void i915_capture_pinned_buffers(struct drm_i915_private *dev_priv,
-					struct drm_i915_error_state *error)
+					struct i915_gpu_state *error)
 {
 	struct i915_address_space *vm = &dev_priv->ggtt.base;
 	struct drm_i915_error_buffer *bo;
@@ -1439,7 +1505,7 @@ static void i915_capture_pinned_buffers(struct drm_i915_private *dev_priv,
 }
 
 static void i915_gem_capture_guc_log_buffer(struct drm_i915_private *dev_priv,
-					    struct drm_i915_error_state *error)
+					    struct i915_gpu_state *error)
 {
 	/* Capturing log buf contents won't be useful if logging was disabled */
 	if (!dev_priv->guc.log.vma || (i915.guc_log_level < 0))
@@ -1451,7 +1517,7 @@ static void i915_gem_capture_guc_log_buffer(struct drm_i915_private *dev_priv,
 
 /* Capture all registers which don't fit into another category. */
 static void i915_capture_reg_state(struct drm_i915_private *dev_priv,
-				   struct drm_i915_error_state *error)
+				   struct i915_gpu_state *error)
 {
 	int i;
 
@@ -1508,9 +1574,11 @@ static void i915_capture_reg_state(struct drm_i915_private *dev_priv,
 		error->ier = I915_READ(GEN8_DE_MISC_IER);
 		for (i = 0; i < 4; i++)
 			error->gtier[i] = I915_READ(GEN8_GT_IER(i));
+		error->ngtier = 4;
 	} else if (HAS_PCH_SPLIT(dev_priv)) {
 		error->ier = I915_READ(DEIER);
 		error->gtier[0] = I915_READ(GTIER);
+		error->ngtier = 1;
 	} else if (IS_GEN2(dev_priv)) {
 		error->ier = I915_READ16(IER);
 	} else if (!IS_VALLEYVIEW(dev_priv)) {
@@ -1521,7 +1589,7 @@ static void i915_capture_reg_state(struct drm_i915_private *dev_priv,
 }
 
 static void i915_error_capture_msg(struct drm_i915_private *dev_priv,
-				   struct drm_i915_error_state *error,
+				   struct i915_gpu_state *error,
 				   u32 engine_mask,
 				   const char *error_msg)
 {
@@ -1534,12 +1602,12 @@ static void i915_error_capture_msg(struct drm_i915_private *dev_priv,
 			"GPU HANG: ecode %d:%d:0x%08x",
 			INTEL_GEN(dev_priv), engine_id, ecode);
 
-	if (engine_id != -1 && error->engine[engine_id].pid != -1)
+	if (engine_id != -1 && error->engine[engine_id].context.pid)
 		len += scnprintf(error->error_msg + len,
 				 sizeof(error->error_msg) - len,
 				 ", in %s [%d]",
-				 error->engine[engine_id].comm,
-				 error->engine[engine_id].pid);
+				 error->engine[engine_id].context.comm,
+				 error->engine[engine_id].context.pid);
 
 	scnprintf(error->error_msg + len, sizeof(error->error_msg) - len,
 		  ", reason: %s, action: %s",
@@ -1548,8 +1616,12 @@ static void i915_error_capture_msg(struct drm_i915_private *dev_priv,
 }
 
 static void i915_capture_gen_state(struct drm_i915_private *dev_priv,
-				   struct drm_i915_error_state *error)
+				   struct i915_gpu_state *error)
 {
+	error->awake = dev_priv->gt.awake;
+	error->wakelock = atomic_read(&dev_priv->pm.wakeref_count);
+	error->suspended = dev_priv->pm.suspended;
+
 	error->iommu = -1;
 #ifdef CONFIG_INTEL_IOMMU
 	error->iommu = intel_iommu_gfx_mapped;
@@ -1562,9 +1634,26 @@ static void i915_capture_gen_state(struct drm_i915_private *dev_priv,
 	       sizeof(error->device_info));
 }
 
+static __always_inline void dup_param(const char *type, void *x)
+{
+	if (!__builtin_strcmp(type, "char *"))
+		*(void **)x = kstrdup(*(void **)x, GFP_ATOMIC);
+}
+
 static int capture(void *data)
 {
-	struct drm_i915_error_state *error = data;
+	struct i915_gpu_state *error = data;
+
+	do_gettimeofday(&error->time);
+	error->boottime = ktime_to_timeval(ktime_get_boottime());
+	error->uptime =
+		ktime_to_timeval(ktime_sub(ktime_get(),
+					   error->i915->gt.last_init_time));
+
+	error->params = i915;
+#define DUP(T, x) dup_param(#T, &error->params.x);
+	I915_PARAMS_FOR_EACH(DUP);
+#undef DUP
 
 	i915_capture_gen_state(error->i915, error);
 	i915_capture_reg_state(error->i915, error);
@@ -1574,12 +1663,6 @@ static int capture(void *data)
 	i915_capture_pinned_buffers(error->i915, error);
 	i915_gem_capture_guc_log_buffer(error->i915, error);
 
-	do_gettimeofday(&error->time);
-	error->boottime = ktime_to_timeval(ktime_get_boottime());
-	error->uptime =
-		ktime_to_timeval(ktime_sub(ktime_get(),
-					   error->i915->gt.last_init_time));
-
 	error->overlay = intel_overlay_capture_error_state(error->i915);
 	error->display = intel_display_capture_error_state(error->i915);
 
@@ -1588,6 +1671,23 @@ static int capture(void *data)
 
 #define DAY_AS_SECONDS(x) (24 * 60 * 60 * (x))
 
+struct i915_gpu_state *
+i915_capture_gpu_state(struct drm_i915_private *i915)
+{
+	struct i915_gpu_state *error;
+
+	error = kzalloc(sizeof(*error), GFP_ATOMIC);
+	if (!error)
+		return NULL;
+
+	kref_init(&error->ref);
+	error->i915 = i915;
+
+	stop_machine(capture, error, NULL);
+
+	return error;
+}
+
 /**
  * i915_capture_error_state - capture an error record for later analysis
  * @dev: drm device
@@ -1602,7 +1702,7 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 			      const char *error_msg)
 {
 	static bool warned;
-	struct drm_i915_error_state *error;
+	struct i915_gpu_state *error;
 	unsigned long flags;
 
 	if (!i915.error_capture)
@@ -1611,18 +1711,12 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 	if (READ_ONCE(dev_priv->gpu_error.first_error))
 		return;
 
-	/* Account for pipe specific data like PIPE*STAT */
-	error = kzalloc(sizeof(*error), GFP_ATOMIC);
+	error = i915_capture_gpu_state(dev_priv);
 	if (!error) {
 		DRM_DEBUG_DRIVER("out of memory, not capturing error state\n");
 		return;
 	}
 
-	kref_init(&error->ref);
-	error->i915 = dev_priv;
-
-	stop_machine(capture, error, NULL);
-
 	i915_error_capture_msg(dev_priv, error, engine_mask, error_msg);
 	DRM_INFO("%s\n", error->error_msg);
 
@@ -1636,7 +1730,7 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 	}
 
 	if (error) {
-		i915_error_state_free(&error->ref);
+		__i915_gpu_state_free(&error->ref);
 		return;
 	}
 
@@ -1652,33 +1746,28 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 	}
 }
 
-void i915_error_state_get(struct drm_device *dev,
-			  struct i915_error_state_file_priv *error_priv)
+struct i915_gpu_state *
+i915_first_error_state(struct drm_i915_private *i915)
 {
-	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct i915_gpu_state *error;
 
-	spin_lock_irq(&dev_priv->gpu_error.lock);
-	error_priv->error = dev_priv->gpu_error.first_error;
-	if (error_priv->error)
-		kref_get(&error_priv->error->ref);
-	spin_unlock_irq(&dev_priv->gpu_error.lock);
-}
+	spin_lock_irq(&i915->gpu_error.lock);
+	error = i915->gpu_error.first_error;
+	if (error)
+		i915_gpu_state_get(error);
+	spin_unlock_irq(&i915->gpu_error.lock);
 
-void i915_error_state_put(struct i915_error_state_file_priv *error_priv)
-{
-	if (error_priv->error)
-		kref_put(&error_priv->error->ref, i915_error_state_free);
+	return error;
 }
 
-void i915_destroy_error_state(struct drm_i915_private *dev_priv)
+void i915_reset_error_state(struct drm_i915_private *i915)
 {
-	struct drm_i915_error_state *error;
+	struct i915_gpu_state *error;
 
-	spin_lock_irq(&dev_priv->gpu_error.lock);
-	error = dev_priv->gpu_error.first_error;
-	dev_priv->gpu_error.first_error = NULL;
-	spin_unlock_irq(&dev_priv->gpu_error.lock);
+	spin_lock_irq(&i915->gpu_error.lock);
+	error = i915->gpu_error.first_error;
+	i915->gpu_error.first_error = NULL;
+	spin_unlock_irq(&i915->gpu_error.lock);
 
-	if (error)
-		kref_put(&error->ref, i915_error_state_free);
+	i915_gpu_state_put(error);
 }
diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c
index 8ced9e26f075..beb38e30d0e9 100644
--- a/drivers/gpu/drm/i915/i915_guc_submission.c
+++ b/drivers/gpu/drm/i915/i915_guc_submission.c
@@ -348,7 +348,7 @@ int i915_guc_wq_reserve(struct drm_i915_gem_request *request)
 	u32 freespace;
 	int ret;
 
-	spin_lock(&client->wq_lock);
+	spin_lock_irq(&client->wq_lock);
 	freespace = CIRC_SPACE(client->wq_tail, desc->head, client->wq_size);
 	freespace -= client->wq_rsvd;
 	if (likely(freespace >= wqi_size)) {
@@ -358,21 +358,27 @@ int i915_guc_wq_reserve(struct drm_i915_gem_request *request)
 		client->no_wq_space++;
 		ret = -EAGAIN;
 	}
-	spin_unlock(&client->wq_lock);
+	spin_unlock_irq(&client->wq_lock);
 
 	return ret;
 }
 
+static void guc_client_update_wq_rsvd(struct i915_guc_client *client, int size)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&client->wq_lock, flags);
+	client->wq_rsvd += size;
+	spin_unlock_irqrestore(&client->wq_lock, flags);
+}
+
 void i915_guc_wq_unreserve(struct drm_i915_gem_request *request)
 {
-	const size_t wqi_size = sizeof(struct guc_wq_item);
+	const int wqi_size = sizeof(struct guc_wq_item);
 	struct i915_guc_client *client = request->i915->guc.execbuf_client;
 
 	GEM_BUG_ON(READ_ONCE(client->wq_rsvd) < wqi_size);
-
-	spin_lock(&client->wq_lock);
-	client->wq_rsvd -= wqi_size;
-	spin_unlock(&client->wq_lock);
+	guc_client_update_wq_rsvd(client, -wqi_size);
 }
 
 /* Construct a Work Item and append it to the GuC's Work Queue */
@@ -509,15 +515,18 @@ static void __i915_guc_submit(struct drm_i915_gem_request *rq)
 	unsigned int engine_id = engine->id;
 	struct intel_guc *guc = &rq->i915->guc;
 	struct i915_guc_client *client = guc->execbuf_client;
+	unsigned long flags;
 	int b_ret;
 
-	spin_lock(&client->wq_lock);
-	guc_wq_item_append(client, rq);
-
 	/* WA to flush out the pending GMADR writes to ring buffer. */
 	if (i915_vma_is_map_and_fenceable(rq->ring->vma))
 		POSTING_READ_FW(GUC_STATUS);
 
+	trace_i915_gem_request_in(rq, 0);
+
+	spin_lock_irqsave(&client->wq_lock, flags);
+
+	guc_wq_item_append(client, rq);
 	b_ret = guc_ring_doorbell(client);
 
 	client->submissions[engine_id] += 1;
@@ -527,7 +536,8 @@ static void __i915_guc_submit(struct drm_i915_gem_request *rq)
 
 	guc->submissions[engine_id] += 1;
 	guc->last_seqno[engine_id] = rq->global_seqno;
-	spin_unlock(&client->wq_lock);
+
+	spin_unlock_irqrestore(&client->wq_lock, flags);
 }
 
 static void i915_guc_submit(struct drm_i915_gem_request *rq)
@@ -943,16 +953,19 @@ int i915_guc_submission_enable(struct drm_i915_private *dev_priv)
 
 	/* Take over from manual control of ELSP (execlists) */
 	for_each_engine(engine, dev_priv, id) {
+		const int wqi_size = sizeof(struct guc_wq_item);
 		struct drm_i915_gem_request *rq;
 
 		engine->submit_request = i915_guc_submit;
 		engine->schedule = NULL;
 
 		/* Replay the current set of previously submitted requests */
+		spin_lock_irq(&engine->timeline->lock);
 		list_for_each_entry(rq, &engine->timeline->requests, link) {
-			client->wq_rsvd += sizeof(struct guc_wq_item);
+			guc_client_update_wq_rsvd(client, wqi_size);
 			__i915_guc_submit(rq);
 		}
+		spin_unlock_irq(&engine->timeline->lock);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index a62feb686895..df95733cf112 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -180,7 +180,7 @@ i915_hotplug_interrupt_update_locked(struct drm_i915_private *dev_priv,
 {
 	uint32_t val;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 	WARN_ON(bits & ~mask);
 
 	val = I915_READ(PORT_HOTPLUG_EN);
@@ -222,7 +222,7 @@ void ilk_update_display_irq(struct drm_i915_private *dev_priv,
 {
 	uint32_t new_val;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	WARN_ON(enabled_irq_mask & ~interrupt_mask);
 
@@ -250,7 +250,7 @@ static void ilk_update_gt_irq(struct drm_i915_private *dev_priv,
 			      uint32_t interrupt_mask,
 			      uint32_t enabled_irq_mask)
 {
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	WARN_ON(enabled_irq_mask & ~interrupt_mask);
 
@@ -302,7 +302,7 @@ static void snb_update_pm_irq(struct drm_i915_private *dev_priv,
 
 	WARN_ON(enabled_irq_mask & ~interrupt_mask);
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	new_val = dev_priv->pm_imr;
 	new_val &= ~interrupt_mask;
@@ -340,7 +340,7 @@ void gen6_reset_pm_iir(struct drm_i915_private *dev_priv, u32 reset_mask)
 {
 	i915_reg_t reg = gen6_pm_iir(dev_priv);
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	I915_WRITE(reg, reset_mask);
 	I915_WRITE(reg, reset_mask);
@@ -349,7 +349,7 @@ void gen6_reset_pm_iir(struct drm_i915_private *dev_priv, u32 reset_mask)
 
 void gen6_enable_pm_irq(struct drm_i915_private *dev_priv, u32 enable_mask)
 {
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	dev_priv->pm_ier |= enable_mask;
 	I915_WRITE(gen6_pm_ier(dev_priv), dev_priv->pm_ier);
@@ -359,7 +359,7 @@ void gen6_enable_pm_irq(struct drm_i915_private *dev_priv, u32 enable_mask)
 
 void gen6_disable_pm_irq(struct drm_i915_private *dev_priv, u32 disable_mask)
 {
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	dev_priv->pm_ier &= ~disable_mask;
 	__gen6_mask_pm_irq(dev_priv, disable_mask);
@@ -463,7 +463,7 @@ static void bdw_update_port_irq(struct drm_i915_private *dev_priv,
 	uint32_t new_val;
 	uint32_t old_val;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	WARN_ON(enabled_irq_mask & ~interrupt_mask);
 
@@ -496,7 +496,7 @@ void bdw_update_pipe_irq(struct drm_i915_private *dev_priv,
 {
 	uint32_t new_val;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	WARN_ON(enabled_irq_mask & ~interrupt_mask);
 
@@ -530,7 +530,7 @@ void ibx_display_interrupt_update(struct drm_i915_private *dev_priv,
 
 	WARN_ON(enabled_irq_mask & ~interrupt_mask);
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	if (WARN_ON(!intel_irqs_enabled(dev_priv)))
 		return;
@@ -546,7 +546,7 @@ __i915_enable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
 	i915_reg_t reg = PIPESTAT(pipe);
 	u32 pipestat = I915_READ(reg) & PIPESTAT_INT_ENABLE_MASK;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 	WARN_ON(!intel_irqs_enabled(dev_priv));
 
 	if (WARN_ONCE(enable_mask & ~PIPESTAT_INT_ENABLE_MASK ||
@@ -573,7 +573,7 @@ __i915_disable_pipestat(struct drm_i915_private *dev_priv, enum pipe pipe,
 	i915_reg_t reg = PIPESTAT(pipe);
 	u32 pipestat = I915_READ(reg) & PIPESTAT_INT_ENABLE_MASK;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 	WARN_ON(!intel_irqs_enabled(dev_priv));
 
 	if (WARN_ONCE(enable_mask & ~PIPESTAT_INT_ENABLE_MASK ||
@@ -783,6 +783,9 @@ static int __intel_get_crtc_scanline(struct intel_crtc *crtc)
 	enum pipe pipe = crtc->pipe;
 	int position, vtotal;
 
+	if (!crtc->active)
+		return -1;
+
 	vtotal = mode->crtc_vtotal;
 	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
 		vtotal /= 2;
@@ -1033,9 +1036,42 @@ static void ironlake_rps_change_irq_handler(struct drm_i915_private *dev_priv)
 
 static void notify_ring(struct intel_engine_cs *engine)
 {
-	smp_store_mb(engine->breadcrumbs.irq_posted, true);
-	if (intel_engine_wakeup(engine))
-		trace_i915_gem_request_notify(engine);
+	struct drm_i915_gem_request *rq = NULL;
+	struct intel_wait *wait;
+
+	atomic_inc(&engine->irq_count);
+	set_bit(ENGINE_IRQ_BREADCRUMB, &engine->irq_posted);
+
+	spin_lock(&engine->breadcrumbs.irq_lock);
+	wait = engine->breadcrumbs.irq_wait;
+	if (wait) {
+		/* We use a callback from the dma-fence to submit
+		 * requests after waiting on our own requests. To
+		 * ensure minimum delay in queuing the next request to
+		 * hardware, signal the fence now rather than wait for
+		 * the signaler to be woken up. We still wake up the
+		 * waiter in order to handle the irq-seqno coherency
+		 * issues (we may receive the interrupt before the
+		 * seqno is written, see __i915_request_irq_complete())
+		 * and to handle coalescing of multiple seqno updates
+		 * and many waiters.
+		 */
+		if (i915_seqno_passed(intel_engine_get_seqno(engine),
+				      wait->seqno))
+			rq = i915_gem_request_get(wait->request);
+
+		wake_up_process(wait->tsk);
+	} else {
+		__intel_engine_disarm_breadcrumbs(engine);
+	}
+	spin_unlock(&engine->breadcrumbs.irq_lock);
+
+	if (rq) {
+		dma_fence_signal(&rq->fence);
+		i915_gem_request_put(rq);
+	}
+
+	trace_intel_engine_notify(engine, wait);
 }
 
 static void vlv_c0_read(struct drm_i915_private *dev_priv,
@@ -1173,20 +1209,12 @@ static void gen6_pm_rps_work(struct work_struct *work)
 
 		if (new_delay >= dev_priv->rps.max_freq_softlimit)
 			adj = 0;
-		/*
-		 * For better performance, jump directly
-		 * to RPe if we're below it.
-		 */
-		if (new_delay < dev_priv->rps.efficient_freq - adj) {
-			new_delay = dev_priv->rps.efficient_freq;
-			adj = 0;
-		}
 	} else if (client_boost || any_waiters(dev_priv)) {
 		adj = 0;
 	} else if (pm_iir & GEN6_PM_RP_DOWN_TIMEOUT) {
 		if (dev_priv->rps.cur_freq > dev_priv->rps.efficient_freq)
 			new_delay = dev_priv->rps.efficient_freq;
-		else
+		else if (dev_priv->rps.cur_freq > dev_priv->rps.min_freq_softlimit)
 			new_delay = dev_priv->rps.min_freq_softlimit;
 		adj = 0;
 	} else if (pm_iir & GEN6_PM_RP_DOWN_THRESHOLD) {
@@ -1209,7 +1237,10 @@ static void gen6_pm_rps_work(struct work_struct *work)
 	new_delay += adj;
 	new_delay = clamp_t(int, new_delay, min, max);
 
-	intel_set_rps(dev_priv, new_delay);
+	if (intel_set_rps(dev_priv, new_delay)) {
+		DRM_DEBUG_DRIVER("Failed to set new GPU frequency\n");
+		dev_priv->rps.last_adj = 0;
+	}
 
 	mutex_unlock(&dev_priv->rps.hw_lock);
 }
@@ -1349,8 +1380,11 @@ gen8_cs_irq_handler(struct intel_engine_cs *engine, u32 iir, int test_shift)
 {
 	if (iir & (GT_RENDER_USER_INTERRUPT << test_shift))
 		notify_ring(engine);
-	if (iir & (GT_CONTEXT_SWITCH_INTERRUPT << test_shift))
-		tasklet_schedule(&engine->irq_tasklet);
+
+	if (iir & (GT_CONTEXT_SWITCH_INTERRUPT << test_shift)) {
+		set_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted);
+		tasklet_hi_schedule(&engine->irq_tasklet);
+	}
 }
 
 static irqreturn_t gen8_gt_irq_ack(struct drm_i915_private *dev_priv,
@@ -3106,19 +3140,9 @@ static u32 intel_hpd_enabled_irqs(struct drm_i915_private *dev_priv,
 	return enabled_irqs;
 }
 
-static void ibx_hpd_irq_setup(struct drm_i915_private *dev_priv)
+static void ibx_hpd_detection_setup(struct drm_i915_private *dev_priv)
 {
-	u32 hotplug_irqs, hotplug, enabled_irqs;
-
-	if (HAS_PCH_IBX(dev_priv)) {
-		hotplug_irqs = SDE_HOTPLUG_MASK;
-		enabled_irqs = intel_hpd_enabled_irqs(dev_priv, hpd_ibx);
-	} else {
-		hotplug_irqs = SDE_HOTPLUG_MASK_CPT;
-		enabled_irqs = intel_hpd_enabled_irqs(dev_priv, hpd_cpt);
-	}
-
-	ibx_display_interrupt_update(dev_priv, hotplug_irqs, enabled_irqs);
+	u32 hotplug;
 
 	/*
 	 * Enable digital hotplug on the PCH, and configure the DP short pulse
@@ -3126,10 +3150,12 @@ static void ibx_hpd_irq_setup(struct drm_i915_private *dev_priv)
 	 * The pulse duration bits are reserved on LPT+.
 	 */
 	hotplug = I915_READ(PCH_PORT_HOTPLUG);
-	hotplug &= ~(PORTD_PULSE_DURATION_MASK|PORTC_PULSE_DURATION_MASK|PORTB_PULSE_DURATION_MASK);
-	hotplug |= PORTD_HOTPLUG_ENABLE | PORTD_PULSE_DURATION_2ms;
-	hotplug |= PORTC_HOTPLUG_ENABLE | PORTC_PULSE_DURATION_2ms;
+	hotplug &= ~(PORTB_PULSE_DURATION_MASK |
+		     PORTC_PULSE_DURATION_MASK |
+		     PORTD_PULSE_DURATION_MASK);
 	hotplug |= PORTB_HOTPLUG_ENABLE | PORTB_PULSE_DURATION_2ms;
+	hotplug |= PORTC_HOTPLUG_ENABLE | PORTC_PULSE_DURATION_2ms;
+	hotplug |= PORTD_HOTPLUG_ENABLE | PORTD_PULSE_DURATION_2ms;
 	/*
 	 * When CPU and PCH are on the same package, port A
 	 * HPD must be enabled in both north and south.
@@ -3139,6 +3165,23 @@ static void ibx_hpd_irq_setup(struct drm_i915_private *dev_priv)
 	I915_WRITE(PCH_PORT_HOTPLUG, hotplug);
 }
 
+static void ibx_hpd_irq_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug_irqs, enabled_irqs;
+
+	if (HAS_PCH_IBX(dev_priv)) {
+		hotplug_irqs = SDE_HOTPLUG_MASK;
+		enabled_irqs = intel_hpd_enabled_irqs(dev_priv, hpd_ibx);
+	} else {
+		hotplug_irqs = SDE_HOTPLUG_MASK_CPT;
+		enabled_irqs = intel_hpd_enabled_irqs(dev_priv, hpd_cpt);
+	}
+
+	ibx_display_interrupt_update(dev_priv, hotplug_irqs, enabled_irqs);
+
+	ibx_hpd_detection_setup(dev_priv);
+}
+
 static void spt_hpd_detection_setup(struct drm_i915_private *dev_priv)
 {
 	u32 hotplug;
@@ -3168,9 +3211,25 @@ static void spt_hpd_irq_setup(struct drm_i915_private *dev_priv)
 	spt_hpd_detection_setup(dev_priv);
 }
 
+static void ilk_hpd_detection_setup(struct drm_i915_private *dev_priv)
+{
+	u32 hotplug;
+
+	/*
+	 * Enable digital hotplug on the CPU, and configure the DP short pulse
+	 * duration to 2ms (which is the minimum in the Display Port spec)
+	 * The pulse duration bits are reserved on HSW+.
+	 */
+	hotplug = I915_READ(DIGITAL_PORT_HOTPLUG_CNTRL);
+	hotplug &= ~DIGITAL_PORTA_PULSE_DURATION_MASK;
+	hotplug |= DIGITAL_PORTA_HOTPLUG_ENABLE |
+		   DIGITAL_PORTA_PULSE_DURATION_2ms;
+	I915_WRITE(DIGITAL_PORT_HOTPLUG_CNTRL, hotplug);
+}
+
 static void ilk_hpd_irq_setup(struct drm_i915_private *dev_priv)
 {
-	u32 hotplug_irqs, hotplug, enabled_irqs;
+	u32 hotplug_irqs, enabled_irqs;
 
 	if (INTEL_GEN(dev_priv) >= 8) {
 		hotplug_irqs = GEN8_PORT_DP_A_HOTPLUG;
@@ -3189,15 +3248,7 @@ static void ilk_hpd_irq_setup(struct drm_i915_private *dev_priv)
 		ilk_update_display_irq(dev_priv, hotplug_irqs, enabled_irqs);
 	}
 
-	/*
-	 * Enable digital hotplug on the CPU, and configure the DP short pulse
-	 * duration to 2ms (which is the minimum in the Display Port spec)
-	 * The pulse duration bits are reserved on HSW+.
-	 */
-	hotplug = I915_READ(DIGITAL_PORT_HOTPLUG_CNTRL);
-	hotplug &= ~DIGITAL_PORTA_PULSE_DURATION_MASK;
-	hotplug |= DIGITAL_PORTA_HOTPLUG_ENABLE | DIGITAL_PORTA_PULSE_DURATION_2ms;
-	I915_WRITE(DIGITAL_PORT_HOTPLUG_CNTRL, hotplug);
+	ilk_hpd_detection_setup(dev_priv);
 
 	ibx_hpd_irq_setup(dev_priv);
 }
@@ -3268,7 +3319,7 @@ static void ibx_irq_postinstall(struct drm_device *dev)
 
 	if (HAS_PCH_IBX(dev_priv) || HAS_PCH_CPT(dev_priv) ||
 	    HAS_PCH_LPT(dev_priv))
-		; /* TODO: Enable HPD detection on older PCH platforms too */
+		ibx_hpd_detection_setup(dev_priv);
 	else
 		spt_hpd_detection_setup(dev_priv);
 }
@@ -3345,6 +3396,8 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 
 	gen5_gt_irq_postinstall(dev);
 
+	ilk_hpd_detection_setup(dev_priv);
+
 	ibx_irq_postinstall(dev);
 
 	if (IS_IRONLAKE_M(dev_priv)) {
@@ -3363,7 +3416,7 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 
 void valleyview_enable_display_irqs(struct drm_i915_private *dev_priv)
 {
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	if (dev_priv->display_irqs_enabled)
 		return;
@@ -3378,7 +3431,7 @@ void valleyview_enable_display_irqs(struct drm_i915_private *dev_priv)
 
 void valleyview_disable_display_irqs(struct drm_i915_private *dev_priv)
 {
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	if (!dev_priv->display_irqs_enabled)
 		return;
@@ -3485,6 +3538,8 @@ static void gen8_de_irq_postinstall(struct drm_i915_private *dev_priv)
 
 	if (IS_GEN9_LP(dev_priv))
 		bxt_hpd_detection_setup(dev_priv);
+	else if (IS_BROADWELL(dev_priv))
+		ilk_hpd_detection_setup(dev_priv);
 }
 
 static int gen8_irq_postinstall(struct drm_device *dev)
@@ -4052,7 +4107,7 @@ static void i915_hpd_irq_setup(struct drm_i915_private *dev_priv)
 {
 	u32 hotplug_en;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	/* Note HDMI and DP share hotplug bits */
 	/* enable bits are the same for all generations */
@@ -4265,6 +4320,18 @@ void intel_irq_init(struct drm_i915_private *dev_priv)
 	if (!IS_GEN2(dev_priv))
 		dev->vblank_disable_immediate = true;
 
+	/* Most platforms treat the display irq block as an always-on
+	 * power domain. vlv/chv can disable it at runtime and need
+	 * special care to avoid writing any of the display block registers
+	 * outside of the power domain. We defer setting up the display irqs
+	 * in this case to the runtime pm.
+	 */
+	dev_priv->display_irqs_enabled = true;
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		dev_priv->display_irqs_enabled = false;
+
+	dev_priv->hotplug.hpd_storm_threshold = HPD_STORM_DEFAULT_THRESHOLD;
+
 	dev->driver->get_vblank_timestamp = i915_get_vblank_timestamp;
 	dev->driver->get_scanout_position = i915_get_crtc_scanoutpos;
 
diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c
index 0e280fbd52f1..2e9645e6555a 100644
--- a/drivers/gpu/drm/i915/i915_params.c
+++ b/drivers/gpu/drm/i915/i915_params.c
@@ -145,7 +145,7 @@ MODULE_PARM_DESC(enable_psr, "Enable PSR "
 		 "(0=disabled, 1=enabled - link mode chosen per-platform, 2=force link-standby mode, 3=force link-off mode) "
 		 "Default: -1 (use per-chip default)");
 
-module_param_named_unsafe(alpha_support, i915.alpha_support, int, 0400);
+module_param_named_unsafe(alpha_support, i915.alpha_support, bool, 0400);
 MODULE_PARM_DESC(alpha_support,
 	"Enable alpha quality driver support for latest hardware. "
 	"See also CONFIG_DRM_I915_ALPHA_SUPPORT.");
@@ -205,9 +205,9 @@ module_param_named(verbose_state_checks, i915.verbose_state_checks, bool, 0600);
 MODULE_PARM_DESC(verbose_state_checks,
 	"Enable verbose logs (ie. WARN_ON()) in case of unexpected hw state conditions.");
 
-module_param_named_unsafe(nuclear_pageflip, i915.nuclear_pageflip, bool, 0600);
+module_param_named_unsafe(nuclear_pageflip, i915.nuclear_pageflip, bool, 0400);
 MODULE_PARM_DESC(nuclear_pageflip,
-		 "Force atomic modeset functionality; asynchronous mode is not yet supported. (default: false).");
+		 "Force enable atomic functionality on platforms that don't have full support yet.");
 
 /* WA to get away with the default setting in VBT for early platforms.Will be removed */
 module_param_named_unsafe(edp_vswing, i915.edp_vswing, int, 0400);
diff --git a/drivers/gpu/drm/i915/i915_params.h b/drivers/gpu/drm/i915/i915_params.h
index 8e433de04679..55d47eea172e 100644
--- a/drivers/gpu/drm/i915/i915_params.h
+++ b/drivers/gpu/drm/i915/i915_params.h
@@ -27,46 +27,51 @@
 
 #include <linux/cache.h> /* for __read_mostly */
 
+#define I915_PARAMS_FOR_EACH(func) \
+	func(int, modeset); \
+	func(int, panel_ignore_lid); \
+	func(int, semaphores); \
+	func(int, lvds_channel_mode); \
+	func(int, panel_use_ssc); \
+	func(int, vbt_sdvo_panel_type); \
+	func(int, enable_rc6); \
+	func(int, enable_dc); \
+	func(int, enable_fbc); \
+	func(int, enable_ppgtt); \
+	func(int, enable_execlists); \
+	func(int, enable_psr); \
+	func(int, disable_power_well); \
+	func(int, enable_ips); \
+	func(int, invert_brightness); \
+	func(int, enable_guc_loading); \
+	func(int, enable_guc_submission); \
+	func(int, guc_log_level); \
+	func(int, use_mmio_flip); \
+	func(int, mmio_debug); \
+	func(int, edp_vswing); \
+	func(unsigned int, inject_load_failure); \
+	/* leave bools at the end to not create holes */ \
+	func(bool, alpha_support); \
+	func(bool, enable_cmd_parser); \
+	func(bool, enable_hangcheck); \
+	func(bool, fastboot); \
+	func(bool, prefault_disable); \
+	func(bool, load_detect_test); \
+	func(bool, force_reset_modeset_test); \
+	func(bool, reset); \
+	func(bool, error_capture); \
+	func(bool, disable_display); \
+	func(bool, verbose_state_checks); \
+	func(bool, nuclear_pageflip); \
+	func(bool, enable_dp_mst); \
+	func(bool, enable_dpcd_backlight); \
+	func(bool, enable_gvt)
+
+#define MEMBER(T, member) T member
 struct i915_params {
-	int modeset;
-	int panel_ignore_lid;
-	int semaphores;
-	int lvds_channel_mode;
-	int panel_use_ssc;
-	int vbt_sdvo_panel_type;
-	int enable_rc6;
-	int enable_dc;
-	int enable_fbc;
-	int enable_ppgtt;
-	int enable_execlists;
-	int enable_psr;
-	unsigned int alpha_support;
-	int disable_power_well;
-	int enable_ips;
-	int invert_brightness;
-	int enable_guc_loading;
-	int enable_guc_submission;
-	int guc_log_level;
-	int use_mmio_flip;
-	int mmio_debug;
-	int edp_vswing;
-	unsigned int inject_load_failure;
-	/* leave bools at the end to not create holes */
-	bool enable_cmd_parser;
-	bool enable_hangcheck;
-	bool fastboot;
-	bool prefault_disable;
-	bool load_detect_test;
-	bool force_reset_modeset_test;
-	bool reset;
-	bool error_capture;
-	bool disable_display;
-	bool verbose_state_checks;
-	bool nuclear_pageflip;
-	bool enable_dp_mst;
-	bool enable_dpcd_backlight;
-	bool enable_gvt;
+	I915_PARAMS_FOR_EACH(MEMBER);
 };
+#undef MEMBER
 
 extern struct i915_params i915 __read_mostly;
 
diff --git a/drivers/gpu/drm/i915/i915_pci.c b/drivers/gpu/drm/i915/i915_pci.c
index ecb487b5356f..732101ed57fb 100644
--- a/drivers/gpu/drm/i915/i915_pci.c
+++ b/drivers/gpu/drm/i915/i915_pci.c
@@ -27,6 +27,7 @@
 #include <linux/vga_switcheroo.h>
 
 #include "i915_drv.h"
+#include "i915_selftest.h"
 
 #define GEN_DEFAULT_PIPEOFFSETS \
 	.pipe_offsets = { PIPE_A_OFFSET, PIPE_B_OFFSET, \
@@ -403,6 +404,7 @@ static const struct intel_device_info intel_geminilake_info = {
 	.platform = INTEL_GEMINILAKE,
 	.is_alpha_support = 1,
 	.ddb_size = 1024,
+	.color = { .degamma_lut_size = 0, .gamma_lut_size = 1024 }
 };
 
 static const struct intel_device_info intel_kabylake_info = {
@@ -472,10 +474,19 @@ static const struct pci_device_id pciidlist[] = {
 };
 MODULE_DEVICE_TABLE(pci, pciidlist);
 
+static void i915_pci_remove(struct pci_dev *pdev)
+{
+	struct drm_device *dev = pci_get_drvdata(pdev);
+
+	i915_driver_unload(dev);
+	drm_dev_unref(dev);
+}
+
 static int i915_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 {
 	struct intel_device_info *intel_info =
 		(struct intel_device_info *) ent->driver_data;
+	int err;
 
 	if (IS_ALPHA_SUPPORT(intel_info) && !i915.alpha_support) {
 		DRM_INFO("The driver support for your hardware in this kernel version is alpha quality\n"
@@ -499,15 +510,17 @@ static int i915_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	if (vga_switcheroo_client_probe_defer(pdev))
 		return -EPROBE_DEFER;
 
-	return i915_driver_load(pdev, ent);
-}
+	err = i915_driver_load(pdev, ent);
+	if (err)
+		return err;
 
-static void i915_pci_remove(struct pci_dev *pdev)
-{
-	struct drm_device *dev = pci_get_drvdata(pdev);
+	err = i915_live_selftests(pdev);
+	if (err) {
+		i915_pci_remove(pdev);
+		return err > 0 ? -ENOTTY : err;
+	}
 
-	i915_driver_unload(dev);
-	drm_dev_unref(dev);
+	return 0;
 }
 
 static struct pci_driver i915_pci_driver = {
@@ -521,6 +534,11 @@ static struct pci_driver i915_pci_driver = {
 static int __init i915_init(void)
 {
 	bool use_kms = true;
+	int err;
+
+	err = i915_mock_selftests();
+	if (err)
+		return err > 0 ? 0 : err;
 
 	/*
 	 * Enable KMS by default, unless explicitly overriden by
diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c
index a1b7eec58be2..8c121187ff39 100644
--- a/drivers/gpu/drm/i915/i915_perf.c
+++ b/drivers/gpu/drm/i915/i915_perf.c
@@ -1008,7 +1008,7 @@ static void hsw_disable_metric_set(struct drm_i915_private *dev_priv)
 
 static void gen7_update_oacontrol_locked(struct drm_i915_private *dev_priv)
 {
-	assert_spin_locked(&dev_priv->perf.hook_lock);
+	lockdep_assert_held(&dev_priv->perf.hook_lock);
 
 	if (dev_priv->perf.oa.exclusive_stream->enabled) {
 		struct i915_gem_context *ctx =
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 1c8f5b9a7fcd..cc843f96576f 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -48,6 +48,8 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg)
 	return !i915_mmio_reg_equal(reg, INVALID_MMIO_REG);
 }
 
+#define _PICK(__index, ...) (((const u32 []){ __VA_ARGS__ })[__index])
+
 #define _PIPE(pipe, a, b) ((a) + (pipe)*((b)-(a)))
 #define _MMIO_PIPE(pipe, a, b) _MMIO(_PIPE(pipe, a, b))
 #define _PLANE(plane, a, b) _PIPE(plane, a, b)
@@ -56,14 +58,11 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg)
 #define _MMIO_TRANS(tran, a, b) _MMIO(_TRANS(tran, a, b))
 #define _PORT(port, a, b) ((a) + (port)*((b)-(a)))
 #define _MMIO_PORT(port, a, b) _MMIO(_PORT(port, a, b))
-#define _PIPE3(pipe, a, b, c) ((pipe) == PIPE_A ? (a) : \
-			       (pipe) == PIPE_B ? (b) : (c))
+#define _PIPE3(pipe, ...) _PICK(pipe, __VA_ARGS__)
 #define _MMIO_PIPE3(pipe, a, b, c) _MMIO(_PIPE3(pipe, a, b, c))
-#define _PORT3(port, a, b, c) ((port) == PORT_A ? (a) : \
-			       (port) == PORT_B ? (b) : (c))
+#define _PORT3(port, ...) _PICK(port, __VA_ARGS__)
 #define _MMIO_PORT3(pipe, a, b, c) _MMIO(_PORT3(pipe, a, b, c))
-#define _PHY3(phy, a, b, c) ((phy) == DPIO_PHY0 ? (a) : \
-			     (phy) == DPIO_PHY1 ? (b) : (c))
+#define _PHY3(phy, ...) _PICK(phy, __VA_ARGS__)
 #define _MMIO_PHY3(phy, a, b, c) _MMIO(_PHY3(phy, a, b, c))
 
 #define _MASKED_FIELD(mask, value) ({					   \
@@ -78,7 +77,13 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg)
 #define _MASKED_BIT_ENABLE(a)	({ typeof(a) _a = (a); _MASKED_FIELD(_a, _a); })
 #define _MASKED_BIT_DISABLE(a)	(_MASKED_FIELD((a), 0))
 
+/* Engine ID */
 
+#define RCS_HW		0
+#define VCS_HW		1
+#define BCS_HW		2
+#define VECS_HW		3
+#define VCS2_HW		4
 
 /* PCI config space */
 
@@ -120,7 +125,7 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg)
 #define GCFGC	0xf0 /* 915+ only */
 #define   GC_LOW_FREQUENCY_ENABLE	(1 << 7)
 #define   GC_DISPLAY_CLOCK_190_200_MHZ	(0 << 4)
-#define   GC_DISPLAY_CLOCK_333_MHZ	(4 << 4)
+#define   GC_DISPLAY_CLOCK_333_320_MHZ	(4 << 4)
 #define   GC_DISPLAY_CLOCK_267_MHZ_PNV	(0 << 4)
 #define   GC_DISPLAY_CLOCK_333_MHZ_PNV	(1 << 4)
 #define   GC_DISPLAY_CLOCK_444_MHZ_PNV	(2 << 4)
@@ -1553,6 +1558,7 @@ enum skl_disp_power_wells {
 	_MMIO(_BXT_PHY_CH(phy, ch, reg_ch0, reg_ch1))
 
 #define BXT_P_CR_GT_DISP_PWRON		_MMIO(0x138090)
+#define  MIPIO_RST_CTRL				(1 << 2)
 
 #define _BXT_PHY_CTL_DDI_A		0x64C00
 #define _BXT_PHY_CTL_DDI_B		0x64C10
@@ -3376,10 +3382,22 @@ enum {
 	INTEL_LEGACY_64B_CONTEXT
 };
 
+enum {
+	FAULT_AND_HANG = 0,
+	FAULT_AND_HALT, /* Debug only */
+	FAULT_AND_STREAM,
+	FAULT_AND_CONTINUE /* Unsupported */
+};
+
+#define GEN8_CTX_VALID (1<<0)
+#define GEN8_CTX_FORCE_PD_RESTORE (1<<1)
+#define GEN8_CTX_FORCE_RESTORE (1<<2)
+#define GEN8_CTX_L3LLC_COHERENT (1<<5)
+#define GEN8_CTX_PRIVILEGE (1<<8)
 #define GEN8_CTX_ADDRESSING_MODE_SHIFT 3
-#define GEN8_CTX_ADDRESSING_MODE(dev_priv) (USES_FULL_48BIT_PPGTT(dev_priv) ?\
-				INTEL_LEGACY_64B_CONTEXT : \
-				INTEL_LEGACY_32B_CONTEXT)
+
+#define GEN8_CTX_ID_SHIFT 32
+#define GEN8_CTX_ID_WIDTH 21
 
 #define CHV_CLK_CTL1			_MMIO(0x101100)
 #define VLV_CLK_CTL2			_MMIO(0x101104)
@@ -5887,11 +5905,18 @@ enum {
 #define _PLANE_KEYMSK_2_A			0x70298
 #define _PLANE_KEYMAX_1_A			0x701a0
 #define _PLANE_KEYMAX_2_A			0x702a0
+#define _PLANE_COLOR_CTL_1_A			0x701CC /* GLK+ */
+#define _PLANE_COLOR_CTL_2_A			0x702CC /* GLK+ */
+#define _PLANE_COLOR_CTL_3_A			0x703CC /* GLK+ */
+#define   PLANE_COLOR_PIPE_GAMMA_ENABLE		(1 << 30)
+#define   PLANE_COLOR_PIPE_CSC_ENABLE		(1 << 23)
+#define   PLANE_COLOR_PLANE_GAMMA_DISABLE	(1 << 13)
 #define _PLANE_BUF_CFG_1_A			0x7027c
 #define _PLANE_BUF_CFG_2_A			0x7037c
 #define _PLANE_NV12_BUF_CFG_1_A		0x70278
 #define _PLANE_NV12_BUF_CFG_2_A		0x70378
 
+
 #define _PLANE_CTL_1_B				0x71180
 #define _PLANE_CTL_2_B				0x71280
 #define _PLANE_CTL_3_B				0x71380
@@ -5986,7 +6011,17 @@ enum {
 #define PLANE_NV12_BUF_CFG(pipe, plane)	\
 	_MMIO_PLANE(plane, _PLANE_NV12_BUF_CFG_1(pipe), _PLANE_NV12_BUF_CFG_2(pipe))
 
-/* SKL new cursor registers */
+#define _PLANE_COLOR_CTL_1_B			0x711CC
+#define _PLANE_COLOR_CTL_2_B			0x712CC
+#define _PLANE_COLOR_CTL_3_B			0x713CC
+#define _PLANE_COLOR_CTL_1(pipe)	\
+	_PIPE(pipe, _PLANE_COLOR_CTL_1_A, _PLANE_COLOR_CTL_1_B)
+#define _PLANE_COLOR_CTL_2(pipe)	\
+	_PIPE(pipe, _PLANE_COLOR_CTL_2_A, _PLANE_COLOR_CTL_2_B)
+#define PLANE_COLOR_CTL(pipe, plane)	\
+	_MMIO_PLANE(plane, _PLANE_COLOR_CTL_1(pipe), _PLANE_COLOR_CTL_2(pipe))
+
+#/* SKL new cursor registers */
 #define _CUR_BUF_CFG_A				0x7017c
 #define _CUR_BUF_CFG_B				0x7117c
 #define CUR_BUF_CFG(pipe)	_MMIO_PIPE(pipe, _CUR_BUF_CFG_A, _CUR_BUF_CFG_B)
@@ -6466,6 +6501,11 @@ enum {
 #define CHICKEN_PAR2_1		_MMIO(0x42090)
 #define  KVM_CONFIG_CHANGE_NOTIFICATION_SELECT	(1 << 14)
 
+#define CHICKEN_MISC_2		_MMIO(0x42084)
+#define  GLK_CL0_PWR_DOWN	(1 << 10)
+#define  GLK_CL1_PWR_DOWN	(1 << 11)
+#define  GLK_CL2_PWR_DOWN	(1 << 12)
+
 #define _CHICKEN_PIPESL_1_A	0x420b0
 #define _CHICKEN_PIPESL_1_B	0x420b4
 #define  HSW_FBCQ_DIS			(1 << 22)
@@ -8167,6 +8207,7 @@ enum {
 #define   PAL_PREC_10_12_BIT		(0 << 31)
 #define   PAL_PREC_SPLIT_MODE		(1 << 31)
 #define   PAL_PREC_AUTO_INCREMENT	(1 << 15)
+#define   PAL_PREC_INDEX_VALUE_MASK	(0x3ff << 0)
 #define _PAL_PREC_DATA_A	0x4A404
 #define _PAL_PREC_DATA_B	0x4AC04
 #define _PAL_PREC_DATA_C	0x4B404
@@ -8176,12 +8217,26 @@ enum {
 #define _PAL_PREC_EXT_GC_MAX_A	0x4A420
 #define _PAL_PREC_EXT_GC_MAX_B	0x4AC20
 #define _PAL_PREC_EXT_GC_MAX_C	0x4B420
+#define _PAL_PREC_EXT2_GC_MAX_A	0x4A430
+#define _PAL_PREC_EXT2_GC_MAX_B	0x4AC30
+#define _PAL_PREC_EXT2_GC_MAX_C	0x4B430
 
 #define PREC_PAL_INDEX(pipe)		_MMIO_PIPE(pipe, _PAL_PREC_INDEX_A, _PAL_PREC_INDEX_B)
 #define PREC_PAL_DATA(pipe)		_MMIO_PIPE(pipe, _PAL_PREC_DATA_A, _PAL_PREC_DATA_B)
 #define PREC_PAL_GC_MAX(pipe, i)	_MMIO(_PIPE(pipe, _PAL_PREC_GC_MAX_A, _PAL_PREC_GC_MAX_B) + (i) * 4)
 #define PREC_PAL_EXT_GC_MAX(pipe, i)	_MMIO(_PIPE(pipe, _PAL_PREC_EXT_GC_MAX_A, _PAL_PREC_EXT_GC_MAX_B) + (i) * 4)
 
+#define _PRE_CSC_GAMC_INDEX_A	0x4A484
+#define _PRE_CSC_GAMC_INDEX_B	0x4AC84
+#define _PRE_CSC_GAMC_INDEX_C	0x4B484
+#define   PRE_CSC_GAMC_AUTO_INCREMENT	(1 << 10)
+#define _PRE_CSC_GAMC_DATA_A	0x4A488
+#define _PRE_CSC_GAMC_DATA_B	0x4AC88
+#define _PRE_CSC_GAMC_DATA_C	0x4B488
+
+#define PRE_CSC_GAMC_INDEX(pipe)	_MMIO_PIPE(pipe, _PRE_CSC_GAMC_INDEX_A, _PRE_CSC_GAMC_INDEX_B)
+#define PRE_CSC_GAMC_DATA(pipe)		_MMIO_PIPE(pipe, _PRE_CSC_GAMC_DATA_A, _PRE_CSC_GAMC_DATA_B)
+
 /* pipe CSC & degamma/gamma LUTs on CHV */
 #define _CGM_PIPE_A_CSC_COEFF01	(VLV_DISPLAY_BASE + 0x67900)
 #define _CGM_PIPE_A_CSC_COEFF23	(VLV_DISPLAY_BASE + 0x67904)
@@ -8215,9 +8270,14 @@ enum {
 
 /* MIPI DSI registers */
 
-#define _MIPI_PORT(port, a, c)	_PORT3(port, a, 0, c)	/* ports A and C only */
+#define _MIPI_PORT(port, a, c)	((port) ? c : a)	/* ports A and C only */
 #define _MMIO_MIPI(port, a, c)	_MMIO(_MIPI_PORT(port, a, c))
 
+#define MIPIO_TXESC_CLK_DIV1			_MMIO(0x160004)
+#define  GLK_TX_ESC_CLK_DIV1_MASK			0x3FF
+#define MIPIO_TXESC_CLK_DIV2			_MMIO(0x160008)
+#define  GLK_TX_ESC_CLK_DIV2_MASK			0x3FF
+
 /* BXT MIPI clock controls */
 #define BXT_MAX_VAR_OUTPUT_KHZ			39500
 
@@ -8304,10 +8364,12 @@ enum {
 #define  BXT_DSI_PLL_PVD_RATIO_SHIFT	16
 #define  BXT_DSI_PLL_PVD_RATIO_MASK	(3 << BXT_DSI_PLL_PVD_RATIO_SHIFT)
 #define  BXT_DSI_PLL_PVD_RATIO_1	(1 << BXT_DSI_PLL_PVD_RATIO_SHIFT)
+#define  BXT_DSIC_16X_BY1		(0 << 10)
 #define  BXT_DSIC_16X_BY2		(1 << 10)
 #define  BXT_DSIC_16X_BY3		(2 << 10)
 #define  BXT_DSIC_16X_BY4		(3 << 10)
 #define  BXT_DSIC_16X_MASK		(3 << 10)
+#define  BXT_DSIA_16X_BY1		(0 << 8)
 #define  BXT_DSIA_16X_BY2		(1 << 8)
 #define  BXT_DSIA_16X_BY3		(2 << 8)
 #define  BXT_DSIA_16X_BY4		(3 << 8)
@@ -8317,6 +8379,8 @@ enum {
 
 #define BXT_DSI_PLL_RATIO_MAX		0x7D
 #define BXT_DSI_PLL_RATIO_MIN		0x22
+#define GLK_DSI_PLL_RATIO_MAX		0x6F
+#define GLK_DSI_PLL_RATIO_MIN		0x22
 #define BXT_DSI_PLL_RATIO_MASK		0xFF
 #define BXT_REF_CLOCK_KHZ		19200
 
@@ -8333,6 +8397,12 @@ enum {
 #define _BXT_MIPIC_PORT_CTRL				0x6B8C0
 #define BXT_MIPI_PORT_CTRL(tc)	_MMIO_MIPI(tc, _BXT_MIPIA_PORT_CTRL, _BXT_MIPIC_PORT_CTRL)
 
+#define BXT_P_DSI_REGULATOR_CFG			_MMIO(0x160020)
+#define  STAP_SELECT					(1 << 0)
+
+#define BXT_P_DSI_REGULATOR_TX_CTRL		_MMIO(0x160054)
+#define  HS_IO_CTRL_SELECT				(1 << 0)
+
 #define  DPI_ENABLE					(1 << 31) /* A + C */
 #define  MIPIA_MIPI4DPHY_DELAY_COUNT_SHIFT		27
 #define  MIPIA_MIPI4DPHY_DELAY_COUNT_MASK		(0xf << 27)
@@ -8586,6 +8656,14 @@ enum {
 #define  LP_BYTECLK_SHIFT				0
 #define  LP_BYTECLK_MASK				(0xffff << 0)
 
+#define _MIPIA_TLPX_TIME_COUNT		(dev_priv->mipi_mmio_base + 0xb0a4)
+#define _MIPIC_TLPX_TIME_COUNT		(dev_priv->mipi_mmio_base + 0xb8a4)
+#define MIPI_TLPX_TIME_COUNT(port)	 _MMIO_MIPI(port, _MIPIA_TLPX_TIME_COUNT, _MIPIC_TLPX_TIME_COUNT)
+
+#define _MIPIA_CLK_LANE_TIMING		(dev_priv->mipi_mmio_base + 0xb098)
+#define _MIPIC_CLK_LANE_TIMING		(dev_priv->mipi_mmio_base + 0xb898)
+#define MIPI_CLK_LANE_TIMING(port)	 _MMIO_MIPI(port, _MIPIA_CLK_LANE_TIMING, _MIPIC_CLK_LANE_TIMING)
+
 /* bits 31:0 */
 #define _MIPIA_LP_GEN_DATA		(dev_priv->mipi_mmio_base + 0xb064)
 #define _MIPIC_LP_GEN_DATA		(dev_priv->mipi_mmio_base + 0xb864)
diff --git a/drivers/gpu/drm/i915/i915_selftest.h b/drivers/gpu/drm/i915/i915_selftest.h
new file mode 100644
index 000000000000..9d7d86f1733d
--- /dev/null
+++ b/drivers/gpu/drm/i915/i915_selftest.h
@@ -0,0 +1,106 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#ifndef __I915_SELFTEST_H__
+#define __I915_SELFTEST_H__
+
+struct pci_dev;
+struct drm_i915_private;
+
+struct i915_selftest {
+	unsigned long timeout_jiffies;
+	unsigned int timeout_ms;
+	unsigned int random_seed;
+	int mock;
+	int live;
+};
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include <linux/fault-inject.h>
+
+extern struct i915_selftest i915_selftest;
+
+int i915_mock_selftests(void);
+int i915_live_selftests(struct pci_dev *pdev);
+
+/* We extract the function declarations from i915_mock_selftests.h and
+ * i915_live_selftests.h Add your unit test declarations there!
+ *
+ * Mock unit tests are run very early upon module load, before the driver
+ * is probed. All hardware interactions, as well as other subsystems, must
+ * be "mocked".
+ *
+ * Live unit tests are run after the driver is loaded - all hardware
+ * interactions are real.
+ */
+#define selftest(name, func) int func(void);
+#include "selftests/i915_mock_selftests.h"
+#undef selftest
+#define selftest(name, func) int func(struct drm_i915_private *i915);
+#include "selftests/i915_live_selftests.h"
+#undef selftest
+
+struct i915_subtest {
+	int (*func)(void *data);
+	const char *name;
+};
+
+int __i915_subtests(const char *caller,
+		    const struct i915_subtest *st,
+		    unsigned int count,
+		    void *data);
+#define i915_subtests(T, data) \
+	__i915_subtests(__func__, T, ARRAY_SIZE(T), data)
+
+#define SUBTEST(x) { x, #x }
+
+#define I915_SELFTEST_DECLARE(x) x
+#define I915_SELFTEST_ONLY(x) unlikely(x)
+
+#else /* !IS_ENABLED(CONFIG_DRM_I915_SELFTEST) */
+
+static inline int i915_mock_selftests(void) { return 0; }
+static inline int i915_live_selftests(struct pci_dev *pdev) { return 0; }
+
+#define I915_SELFTEST_DECLARE(x)
+#define I915_SELFTEST_ONLY(x) 0
+
+#endif
+
+/* Using the i915_selftest_ prefix becomes a little unwieldy with the helpers.
+ * Instead we use the igt_ shorthand, in reference to the intel-gpu-tools
+ * suite of uabi test cases (which includes a test runner for our selftests).
+ */
+
+#define IGT_TIMEOUT(name__) \
+	unsigned long name__ = jiffies + i915_selftest.timeout_jiffies
+
+__printf(2, 3)
+bool __igt_timeout(unsigned long timeout, const char *fmt, ...);
+
+#define igt_timeout(t, fmt, ...) \
+	__igt_timeout((t), KERN_WARNING pr_fmt(fmt), ##__VA_ARGS__)
+
+#define igt_can_mi_store_dword_imm(D) (INTEL_GEN(D) > 2)
+
+#endif /* !__I915_SELFTEST_H__ */
diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c
index 376ac957cd1c..af0ac9f261fd 100644
--- a/drivers/gpu/drm/i915/i915_sysfs.c
+++ b/drivers/gpu/drm/i915/i915_sysfs.c
@@ -395,13 +395,13 @@ static ssize_t gt_max_freq_mhz_store(struct device *kdev,
 	/* We still need *_set_rps to process the new max_delay and
 	 * update the interrupt limits and PMINTRMSK even though
 	 * frequency request may be unchanged. */
-	intel_set_rps(dev_priv, val);
+	ret = intel_set_rps(dev_priv, val);
 
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	intel_runtime_pm_put(dev_priv);
 
-	return count;
+	return ret ?: count;
 }
 
 static ssize_t gt_min_freq_mhz_show(struct device *kdev, struct device_attribute *attr, char *buf)
@@ -448,14 +448,13 @@ static ssize_t gt_min_freq_mhz_store(struct device *kdev,
 	/* We still need *_set_rps to process the new min_delay and
 	 * update the interrupt limits and PMINTRMSK even though
 	 * frequency request may be unchanged. */
-	intel_set_rps(dev_priv, val);
+	ret = intel_set_rps(dev_priv, val);
 
 	mutex_unlock(&dev_priv->rps.hw_lock);
 
 	intel_runtime_pm_put(dev_priv);
 
-	return count;
-
+	return ret ?: count;
 }
 
 static DEVICE_ATTR(gt_act_freq_mhz, S_IRUGO, gt_act_freq_mhz_show, NULL);
@@ -523,33 +522,27 @@ static ssize_t error_state_read(struct file *filp, struct kobject *kobj,
 
 	struct device *kdev = kobj_to_dev(kobj);
 	struct drm_i915_private *dev_priv = kdev_minor_to_i915(kdev);
-	struct drm_device *dev = &dev_priv->drm;
-	struct i915_error_state_file_priv error_priv;
 	struct drm_i915_error_state_buf error_str;
-	ssize_t ret_count = 0;
-	int ret;
-
-	memset(&error_priv, 0, sizeof(error_priv));
+	struct i915_gpu_state *gpu;
+	ssize_t ret;
 
-	ret = i915_error_state_buf_init(&error_str, to_i915(dev), count, off);
+	ret = i915_error_state_buf_init(&error_str, dev_priv, count, off);
 	if (ret)
 		return ret;
 
-	error_priv.i915 = dev_priv;
-	i915_error_state_get(dev, &error_priv);
-
-	ret = i915_error_state_to_str(&error_str, &error_priv);
+	gpu = i915_first_error_state(dev_priv);
+	ret = i915_error_state_to_str(&error_str, gpu);
 	if (ret)
 		goto out;
 
-	ret_count = count < error_str.bytes ? count : error_str.bytes;
+	ret = count < error_str.bytes ? count : error_str.bytes;
+	memcpy(buf, error_str.buf, ret);
 
-	memcpy(buf, error_str.buf, ret_count);
 out:
-	i915_error_state_put(&error_priv);
+	i915_gpu_state_put(gpu);
 	i915_error_state_buf_release(&error_str);
 
-	return ret ?: ret_count;
+	return ret;
 }
 
 static ssize_t error_state_write(struct file *file, struct kobject *kobj,
@@ -560,7 +553,7 @@ static ssize_t error_state_write(struct file *file, struct kobject *kobj,
 	struct drm_i915_private *dev_priv = kdev_minor_to_i915(kdev);
 
 	DRM_DEBUG_DRIVER("Resetting error state\n");
-	i915_destroy_error_state(dev_priv);
+	i915_reset_error_state(dev_priv);
 
 	return count;
 }
diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h
index 4461df5a94fe..5503f5ab1e98 100644
--- a/drivers/gpu/drm/i915/i915_trace.h
+++ b/drivers/gpu/drm/i915/i915_trace.h
@@ -14,6 +14,206 @@
 #define TRACE_SYSTEM i915
 #define TRACE_INCLUDE_FILE i915_trace
 
+/* watermark/fifo updates */
+
+TRACE_EVENT(intel_cpu_fifo_underrun,
+	    TP_PROTO(struct drm_i915_private *dev_priv, enum pipe pipe),
+	    TP_ARGS(dev_priv, pipe),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = pipe;
+			   __entry->frame = dev_priv->drm.driver->get_vblank_counter(&dev_priv->drm, pipe);
+			   __entry->scanline = intel_get_crtc_scanline(intel_get_crtc_for_pipe(dev_priv, pipe));
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe),
+		      __entry->frame, __entry->scanline)
+);
+
+TRACE_EVENT(intel_pch_fifo_underrun,
+	    TP_PROTO(struct drm_i915_private *dev_priv, enum transcoder pch_transcoder),
+	    TP_ARGS(dev_priv, pch_transcoder),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     ),
+
+	    TP_fast_assign(
+			   enum pipe pipe = (enum pipe)pch_transcoder;
+			   __entry->pipe = pipe;
+			   __entry->frame = dev_priv->drm.driver->get_vblank_counter(&dev_priv->drm, pipe);
+			   __entry->scanline = intel_get_crtc_scanline(intel_get_crtc_for_pipe(dev_priv, pipe));
+			   ),
+
+	    TP_printk("pch transcoder %c, frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe),
+		      __entry->frame, __entry->scanline)
+);
+
+TRACE_EVENT(intel_memory_cxsr,
+	    TP_PROTO(struct drm_i915_private *dev_priv, bool old, bool new),
+	    TP_ARGS(dev_priv, old, new),
+
+	    TP_STRUCT__entry(
+			     __array(u32, frame, 3)
+			     __array(u32, scanline, 3)
+			     __field(bool, old)
+			     __field(bool, new)
+			     ),
+
+	    TP_fast_assign(
+			   enum pipe pipe;
+			   for_each_pipe(dev_priv, pipe) {
+				   __entry->frame[pipe] =
+					   dev_priv->drm.driver->get_vblank_counter(&dev_priv->drm, pipe);
+				   __entry->scanline[pipe] =
+					   intel_get_crtc_scanline(intel_get_crtc_for_pipe(dev_priv, pipe));
+			   }
+			   __entry->old = old;
+			   __entry->new = new;
+			   ),
+
+	    TP_printk("%s->%s, pipe A: frame=%u, scanline=%u, pipe B: frame=%u, scanline=%u, pipe C: frame=%u, scanline=%u",
+		      onoff(__entry->old), onoff(__entry->new),
+		      __entry->frame[PIPE_A], __entry->scanline[PIPE_A],
+		      __entry->frame[PIPE_B], __entry->scanline[PIPE_B],
+		      __entry->frame[PIPE_C], __entry->scanline[PIPE_C])
+);
+
+TRACE_EVENT(vlv_wm,
+	    TP_PROTO(struct intel_crtc *crtc, const struct vlv_wm_values *wm),
+	    TP_ARGS(crtc, wm),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     __field(u32, level)
+			     __field(u32, cxsr)
+			     __field(u32, primary)
+			     __field(u32, sprite0)
+			     __field(u32, sprite1)
+			     __field(u32, cursor)
+			     __field(u32, sr_plane)
+			     __field(u32, sr_cursor)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = crtc->base.dev->driver->get_vblank_counter(crtc->base.dev,
+										       crtc->pipe);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   __entry->level = wm->level;
+			   __entry->cxsr = wm->cxsr;
+			   __entry->primary = wm->pipe[crtc->pipe].plane[PLANE_PRIMARY];
+			   __entry->sprite0 = wm->pipe[crtc->pipe].plane[PLANE_SPRITE0];
+			   __entry->sprite1 = wm->pipe[crtc->pipe].plane[PLANE_SPRITE1];
+			   __entry->cursor = wm->pipe[crtc->pipe].plane[PLANE_CURSOR];
+			   __entry->sr_plane = wm->sr.plane;
+			   __entry->sr_cursor = wm->sr.cursor;
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u, level=%d, cxsr=%d, wm %d/%d/%d/%d, sr %d/%d",
+		      pipe_name(__entry->pipe), __entry->frame,
+		      __entry->scanline, __entry->level, __entry->cxsr,
+		      __entry->primary, __entry->sprite0, __entry->sprite1, __entry->cursor,
+		      __entry->sr_plane, __entry->sr_cursor)
+);
+
+TRACE_EVENT(vlv_fifo_size,
+	    TP_PROTO(struct intel_crtc *crtc, u32 sprite0_start, u32 sprite1_start, u32 fifo_size),
+	    TP_ARGS(crtc, sprite0_start, sprite1_start, fifo_size),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     __field(u32, sprite0_start)
+			     __field(u32, sprite1_start)
+			     __field(u32, fifo_size)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->frame = crtc->base.dev->driver->get_vblank_counter(crtc->base.dev,
+										       crtc->pipe);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   __entry->sprite0_start = sprite0_start;
+			   __entry->sprite1_start = sprite1_start;
+			   __entry->fifo_size = fifo_size;
+			   ),
+
+	    TP_printk("pipe %c, frame=%u, scanline=%u, %d/%d/%d",
+		      pipe_name(__entry->pipe), __entry->frame,
+		      __entry->scanline, __entry->sprite0_start,
+		      __entry->sprite1_start, __entry->fifo_size)
+);
+
+/* plane updates */
+
+TRACE_EVENT(intel_update_plane,
+	    TP_PROTO(struct drm_plane *plane, struct intel_crtc *crtc),
+	    TP_ARGS(plane, crtc),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(const char *, name)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     __array(int, src, 4)
+			     __array(int, dst, 4)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->name = plane->name;
+			   __entry->frame = crtc->base.dev->driver->get_vblank_counter(crtc->base.dev,
+										       crtc->pipe);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   memcpy(__entry->src, &plane->state->src, sizeof(__entry->src));
+			   memcpy(__entry->dst, &plane->state->dst, sizeof(__entry->dst));
+			   ),
+
+	    TP_printk("pipe %c, plane %s, frame=%u, scanline=%u, " DRM_RECT_FP_FMT " -> " DRM_RECT_FMT,
+		      pipe_name(__entry->pipe), __entry->name,
+		      __entry->frame, __entry->scanline,
+		      DRM_RECT_FP_ARG((const struct drm_rect *)__entry->src),
+		      DRM_RECT_ARG((const struct drm_rect *)__entry->dst))
+);
+
+TRACE_EVENT(intel_disable_plane,
+	    TP_PROTO(struct drm_plane *plane, struct intel_crtc *crtc),
+	    TP_ARGS(plane, crtc),
+
+	    TP_STRUCT__entry(
+			     __field(enum pipe, pipe)
+			     __field(const char *, name)
+			     __field(u32, frame)
+			     __field(u32, scanline)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pipe = crtc->pipe;
+			   __entry->name = plane->name;
+			   __entry->frame = crtc->base.dev->driver->get_vblank_counter(crtc->base.dev,
+										       crtc->pipe);
+			   __entry->scanline = intel_get_crtc_scanline(crtc);
+			   ),
+
+	    TP_printk("pipe %c, plane %s, frame=%u, scanline=%u",
+		      pipe_name(__entry->pipe), __entry->name,
+		      __entry->frame, __entry->scanline)
+);
+
 /* pipe updates */
 
 TRACE_EVENT(i915_pipe_update_start,
@@ -175,134 +375,6 @@ TRACE_EVENT(i915_vma_unbind,
 		      __entry->obj, __entry->offset, __entry->size, __entry->vm)
 );
 
-TRACE_EVENT(i915_va_alloc,
-	TP_PROTO(struct i915_vma *vma),
-	TP_ARGS(vma),
-
-	TP_STRUCT__entry(
-		__field(struct i915_address_space *, vm)
-		__field(u64, start)
-		__field(u64, end)
-	),
-
-	TP_fast_assign(
-		__entry->vm = vma->vm;
-		__entry->start = vma->node.start;
-		__entry->end = vma->node.start + vma->node.size - 1;
-	),
-
-	TP_printk("vm=%p (%c), 0x%llx-0x%llx",
-		  __entry->vm, i915_is_ggtt(__entry->vm) ? 'G' : 'P',  __entry->start, __entry->end)
-);
-
-DECLARE_EVENT_CLASS(i915_px_entry,
-	TP_PROTO(struct i915_address_space *vm, u32 px, u64 start, u64 px_shift),
-	TP_ARGS(vm, px, start, px_shift),
-
-	TP_STRUCT__entry(
-		__field(struct i915_address_space *, vm)
-		__field(u32, px)
-		__field(u64, start)
-		__field(u64, end)
-	),
-
-	TP_fast_assign(
-		__entry->vm = vm;
-		__entry->px = px;
-		__entry->start = start;
-		__entry->end = ((start + (1ULL << px_shift)) & ~((1ULL << px_shift)-1)) - 1;
-	),
-
-	TP_printk("vm=%p, pde=%d (0x%llx-0x%llx)",
-		  __entry->vm, __entry->px, __entry->start, __entry->end)
-);
-
-DEFINE_EVENT(i915_px_entry, i915_page_table_entry_alloc,
-	     TP_PROTO(struct i915_address_space *vm, u32 pde, u64 start, u64 pde_shift),
-	     TP_ARGS(vm, pde, start, pde_shift)
-);
-
-DEFINE_EVENT_PRINT(i915_px_entry, i915_page_directory_entry_alloc,
-		   TP_PROTO(struct i915_address_space *vm, u32 pdpe, u64 start, u64 pdpe_shift),
-		   TP_ARGS(vm, pdpe, start, pdpe_shift),
-
-		   TP_printk("vm=%p, pdpe=%d (0x%llx-0x%llx)",
-			     __entry->vm, __entry->px, __entry->start, __entry->end)
-);
-
-DEFINE_EVENT_PRINT(i915_px_entry, i915_page_directory_pointer_entry_alloc,
-		   TP_PROTO(struct i915_address_space *vm, u32 pml4e, u64 start, u64 pml4e_shift),
-		   TP_ARGS(vm, pml4e, start, pml4e_shift),
-
-		   TP_printk("vm=%p, pml4e=%d (0x%llx-0x%llx)",
-			     __entry->vm, __entry->px, __entry->start, __entry->end)
-);
-
-/* Avoid extra math because we only support two sizes. The format is defined by
- * bitmap_scnprintf. Each 32 bits is 8 HEX digits followed by comma */
-#define TRACE_PT_SIZE(bits) \
-	((((bits) == 1024) ? 288 : 144) + 1)
-
-DECLARE_EVENT_CLASS(i915_page_table_entry_update,
-	TP_PROTO(struct i915_address_space *vm, u32 pde,
-		 struct i915_page_table *pt, u32 first, u32 count, u32 bits),
-	TP_ARGS(vm, pde, pt, first, count, bits),
-
-	TP_STRUCT__entry(
-		__field(struct i915_address_space *, vm)
-		__field(u32, pde)
-		__field(u32, first)
-		__field(u32, last)
-		__dynamic_array(char, cur_ptes, TRACE_PT_SIZE(bits))
-	),
-
-	TP_fast_assign(
-		__entry->vm = vm;
-		__entry->pde = pde;
-		__entry->first = first;
-		__entry->last = first + count - 1;
-		scnprintf(__get_str(cur_ptes),
-			  TRACE_PT_SIZE(bits),
-			  "%*pb",
-			  bits,
-			  pt->used_ptes);
-	),
-
-	TP_printk("vm=%p, pde=%d, updating %u:%u\t%s",
-		  __entry->vm, __entry->pde, __entry->last, __entry->first,
-		  __get_str(cur_ptes))
-);
-
-DEFINE_EVENT(i915_page_table_entry_update, i915_page_table_entry_map,
-	TP_PROTO(struct i915_address_space *vm, u32 pde,
-		 struct i915_page_table *pt, u32 first, u32 count, u32 bits),
-	TP_ARGS(vm, pde, pt, first, count, bits)
-);
-
-TRACE_EVENT(i915_gem_object_change_domain,
-	    TP_PROTO(struct drm_i915_gem_object *obj, u32 old_read, u32 old_write),
-	    TP_ARGS(obj, old_read, old_write),
-
-	    TP_STRUCT__entry(
-			     __field(struct drm_i915_gem_object *, obj)
-			     __field(u32, read_domains)
-			     __field(u32, write_domain)
-			     ),
-
-	    TP_fast_assign(
-			   __entry->obj = obj;
-			   __entry->read_domains = obj->base.read_domains | (old_read << 16);
-			   __entry->write_domain = obj->base.write_domain | (old_write << 16);
-			   ),
-
-	    TP_printk("obj=%p, read=%02x=>%02x, write=%02x=>%02x",
-		      __entry->obj,
-		      __entry->read_domains >> 16,
-		      __entry->read_domains & 0xffff,
-		      __entry->write_domain >> 16,
-		      __entry->write_domain & 0xffff)
-);
-
 TRACE_EVENT(i915_gem_object_pwrite,
 	    TP_PROTO(struct drm_i915_gem_object *obj, u32 offset, u32 len),
 	    TP_ARGS(obj, offset, len),
@@ -503,13 +575,14 @@ TRACE_EVENT(i915_gem_ring_sync_to,
 		      __entry->seqno)
 );
 
-TRACE_EVENT(i915_gem_ring_dispatch,
+TRACE_EVENT(i915_gem_request_queue,
 	    TP_PROTO(struct drm_i915_gem_request *req, u32 flags),
 	    TP_ARGS(req, flags),
 
 	    TP_STRUCT__entry(
 			     __field(u32, dev)
 			     __field(u32, ring)
+			     __field(u32, ctx)
 			     __field(u32, seqno)
 			     __field(u32, flags)
 			     ),
@@ -517,13 +590,14 @@ TRACE_EVENT(i915_gem_ring_dispatch,
 	    TP_fast_assign(
 			   __entry->dev = req->i915->drm.primary->index;
 			   __entry->ring = req->engine->id;
-			   __entry->seqno = req->global_seqno;
+			   __entry->ctx = req->ctx->hw_id;
+			   __entry->seqno = req->fence.seqno;
 			   __entry->flags = flags;
-			   dma_fence_enable_sw_signaling(&req->fence);
 			   ),
 
-	    TP_printk("dev=%u, ring=%u, seqno=%u, flags=%x",
-		      __entry->dev, __entry->ring, __entry->seqno, __entry->flags)
+	    TP_printk("dev=%u, ring=%u, ctx=%u, seqno=%u, flags=0x%x",
+		      __entry->dev, __entry->ring, __entry->ctx, __entry->seqno,
+		      __entry->flags)
 );
 
 TRACE_EVENT(i915_gem_ring_flush,
@@ -555,18 +629,23 @@ DECLARE_EVENT_CLASS(i915_gem_request,
 
 	    TP_STRUCT__entry(
 			     __field(u32, dev)
+			     __field(u32, ctx)
 			     __field(u32, ring)
 			     __field(u32, seqno)
+			     __field(u32, global)
 			     ),
 
 	    TP_fast_assign(
 			   __entry->dev = req->i915->drm.primary->index;
+			   __entry->ctx = req->ctx->hw_id;
 			   __entry->ring = req->engine->id;
-			   __entry->seqno = req->global_seqno;
+			   __entry->seqno = req->fence.seqno;
+			   __entry->global = req->global_seqno;
 			   ),
 
-	    TP_printk("dev=%u, ring=%u, seqno=%u",
-		      __entry->dev, __entry->ring, __entry->seqno)
+	    TP_printk("dev=%u, ring=%u, ctx=%u, seqno=%u, global=%u",
+		      __entry->dev, __entry->ring, __entry->ctx, __entry->seqno,
+		      __entry->global)
 );
 
 DEFINE_EVENT(i915_gem_request, i915_gem_request_add,
@@ -574,24 +653,100 @@ DEFINE_EVENT(i915_gem_request, i915_gem_request_add,
 	    TP_ARGS(req)
 );
 
-TRACE_EVENT(i915_gem_request_notify,
-	    TP_PROTO(struct intel_engine_cs *engine),
-	    TP_ARGS(engine),
+#if defined(CONFIG_DRM_I915_LOW_LEVEL_TRACEPOINTS)
+DEFINE_EVENT(i915_gem_request, i915_gem_request_submit,
+	     TP_PROTO(struct drm_i915_gem_request *req),
+	     TP_ARGS(req)
+);
+
+DEFINE_EVENT(i915_gem_request, i915_gem_request_execute,
+	     TP_PROTO(struct drm_i915_gem_request *req),
+	     TP_ARGS(req)
+);
+
+DECLARE_EVENT_CLASS(i915_gem_request_hw,
+		    TP_PROTO(struct drm_i915_gem_request *req,
+			     unsigned int port),
+		    TP_ARGS(req, port),
+
+		    TP_STRUCT__entry(
+				     __field(u32, dev)
+				     __field(u32, ring)
+				     __field(u32, seqno)
+				     __field(u32, global_seqno)
+				     __field(u32, ctx)
+				     __field(u32, port)
+				    ),
+
+		    TP_fast_assign(
+			           __entry->dev = req->i915->drm.primary->index;
+			           __entry->ring = req->engine->id;
+			           __entry->ctx = req->ctx->hw_id;
+			           __entry->seqno = req->fence.seqno;
+			           __entry->global_seqno = req->global_seqno;
+			           __entry->port = port;
+			          ),
+
+		    TP_printk("dev=%u, ring=%u, ctx=%u, seqno=%u, global=%u, port=%u",
+			      __entry->dev, __entry->ring, __entry->ctx,
+			      __entry->seqno, __entry->global_seqno,
+			      __entry->port)
+);
+
+DEFINE_EVENT(i915_gem_request_hw, i915_gem_request_in,
+	     TP_PROTO(struct drm_i915_gem_request *req, unsigned int port),
+	     TP_ARGS(req, port)
+);
+
+DEFINE_EVENT(i915_gem_request, i915_gem_request_out,
+	     TP_PROTO(struct drm_i915_gem_request *req),
+	     TP_ARGS(req)
+);
+#else
+#if !defined(TRACE_HEADER_MULTI_READ)
+static inline void
+trace_i915_gem_request_submit(struct drm_i915_gem_request *req)
+{
+}
+
+static inline void
+trace_i915_gem_request_execute(struct drm_i915_gem_request *req)
+{
+}
+
+static inline void
+trace_i915_gem_request_in(struct drm_i915_gem_request *req, unsigned int port)
+{
+}
+
+static inline void
+trace_i915_gem_request_out(struct drm_i915_gem_request *req)
+{
+}
+#endif
+#endif
+
+TRACE_EVENT(intel_engine_notify,
+	    TP_PROTO(struct intel_engine_cs *engine, bool waiters),
+	    TP_ARGS(engine, waiters),
 
 	    TP_STRUCT__entry(
 			     __field(u32, dev)
 			     __field(u32, ring)
 			     __field(u32, seqno)
+			     __field(bool, waiters)
 			     ),
 
 	    TP_fast_assign(
 			   __entry->dev = engine->i915->drm.primary->index;
 			   __entry->ring = engine->id;
 			   __entry->seqno = intel_engine_get_seqno(engine);
+			   __entry->waiters = waiters;
 			   ),
 
-	    TP_printk("dev=%u, ring=%u, seqno=%u",
-		      __entry->dev, __entry->ring, __entry->seqno)
+	    TP_printk("dev=%u, ring=%u, seqno=%u, waiters=%u",
+		      __entry->dev, __entry->ring, __entry->seqno,
+		      __entry->waiters)
 );
 
 DEFINE_EVENT(i915_gem_request, i915_gem_request_retire,
@@ -599,20 +754,17 @@ DEFINE_EVENT(i915_gem_request, i915_gem_request_retire,
 	    TP_ARGS(req)
 );
 
-DEFINE_EVENT(i915_gem_request, i915_gem_request_complete,
-	    TP_PROTO(struct drm_i915_gem_request *req),
-	    TP_ARGS(req)
-);
-
 TRACE_EVENT(i915_gem_request_wait_begin,
-	    TP_PROTO(struct drm_i915_gem_request *req),
-	    TP_ARGS(req),
+	    TP_PROTO(struct drm_i915_gem_request *req, unsigned int flags),
+	    TP_ARGS(req, flags),
 
 	    TP_STRUCT__entry(
 			     __field(u32, dev)
 			     __field(u32, ring)
+			     __field(u32, ctx)
 			     __field(u32, seqno)
-			     __field(bool, blocking)
+			     __field(u32, global)
+			     __field(unsigned int, flags)
 			     ),
 
 	    /* NB: the blocking information is racy since mutex_is_locked
@@ -624,14 +776,16 @@ TRACE_EVENT(i915_gem_request_wait_begin,
 	    TP_fast_assign(
 			   __entry->dev = req->i915->drm.primary->index;
 			   __entry->ring = req->engine->id;
-			   __entry->seqno = req->global_seqno;
-			   __entry->blocking =
-				     mutex_is_locked(&req->i915->drm.struct_mutex);
+			   __entry->ctx = req->ctx->hw_id;
+			   __entry->seqno = req->fence.seqno;
+			   __entry->global = req->global_seqno;
+			   __entry->flags = flags;
 			   ),
 
-	    TP_printk("dev=%u, ring=%u, seqno=%u, blocking=%s",
-		      __entry->dev, __entry->ring,
-		      __entry->seqno, __entry->blocking ?  "yes (NB)" : "no")
+	    TP_printk("dev=%u, ring=%u, ctx=%u, seqno=%u, global=%u, blocking=%u, flags=0x%x",
+		      __entry->dev, __entry->ring, __entry->ctx, __entry->seqno,
+		      __entry->global, !!(__entry->flags & I915_WAIT_LOCKED),
+		      __entry->flags)
 );
 
 DEFINE_EVENT(i915_gem_request, i915_gem_request_wait_end,
@@ -769,17 +923,19 @@ DECLARE_EVENT_CLASS(i915_context,
 	TP_STRUCT__entry(
 			__field(u32, dev)
 			__field(struct i915_gem_context *, ctx)
+			__field(u32, hw_id)
 			__field(struct i915_address_space *, vm)
 	),
 
 	TP_fast_assign(
+			__entry->dev = ctx->i915->drm.primary->index;
 			__entry->ctx = ctx;
+			__entry->hw_id = ctx->hw_id;
 			__entry->vm = ctx->ppgtt ? &ctx->ppgtt->base : NULL;
-			__entry->dev = ctx->i915->drm.primary->index;
 	),
 
-	TP_printk("dev=%u, ctx=%p, ctx_vm=%p",
-		  __entry->dev, __entry->ctx, __entry->vm)
+	TP_printk("dev=%u, ctx=%p, ctx_vm=%p, hw_id=%u",
+		  __entry->dev, __entry->ctx, __entry->vm, __entry->hw_id)
 )
 
 DEFINE_EVENT(i915_context, i915_context_create,
diff --git a/drivers/gpu/drm/i915/i915_utils.h b/drivers/gpu/drm/i915/i915_utils.h
index 34020873e1f6..b8ba0f2f92af 100644
--- a/drivers/gpu/drm/i915/i915_utils.h
+++ b/drivers/gpu/drm/i915/i915_utils.h
@@ -25,6 +25,17 @@
 #ifndef __I915_UTILS_H
 #define __I915_UTILS_H
 
+#if GCC_VERSION >= 70000
+#define add_overflows(A, B) \
+	__builtin_add_overflow_p((A), (B), (typeof((A) + (B)))0)
+#else
+#define add_overflows(A, B) ({ \
+	typeof(A) a = (A); \
+	typeof(B) b = (B); \
+	a + b < a; \
+})
+#endif
+
 #define range_overflows(start, size, max) ({ \
 	typeof(start) start__ = (start); \
 	typeof(size) size__ = (size); \
diff --git a/drivers/gpu/drm/i915/i915_vgpu.c b/drivers/gpu/drm/i915/i915_vgpu.c
index d0abfd08a01c..14014068dfcf 100644
--- a/drivers/gpu/drm/i915/i915_vgpu.c
+++ b/drivers/gpu/drm/i915/i915_vgpu.c
@@ -179,7 +179,7 @@ static int vgt_balloon_space(struct i915_ggtt *ggtt,
 int intel_vgt_balloon(struct drm_i915_private *dev_priv)
 {
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
-	unsigned long ggtt_end = ggtt->base.start + ggtt->base.total;
+	unsigned long ggtt_end = ggtt->base.total;
 
 	unsigned long mappable_base, mappable_size, mappable_end;
 	unsigned long unmappable_base, unmappable_size, unmappable_end;
@@ -202,8 +202,7 @@ int intel_vgt_balloon(struct drm_i915_private *dev_priv)
 	DRM_INFO("Unmappable graphic memory: base 0x%lx size %ldKiB\n",
 		 unmappable_base, unmappable_size / 1024);
 
-	if (mappable_base < ggtt->base.start ||
-	    mappable_end > ggtt->mappable_end ||
+	if (mappable_end > ggtt->mappable_end ||
 	    unmappable_base < ggtt->mappable_end ||
 	    unmappable_end > ggtt_end) {
 		DRM_ERROR("Invalid ballooning configuration!\n");
@@ -231,9 +230,9 @@ int intel_vgt_balloon(struct drm_i915_private *dev_priv)
 	}
 
 	/* Mappable graphic memory ballooning */
-	if (mappable_base > ggtt->base.start) {
+	if (mappable_base) {
 		ret = vgt_balloon_space(ggtt, &bl_info.space[0],
-					ggtt->base.start, mappable_base);
+					0, mappable_base);
 
 		if (ret)
 			goto err;
diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c
index 155906e84812..1aba47024656 100644
--- a/drivers/gpu/drm/i915/i915_vma.c
+++ b/drivers/gpu/drm/i915/i915_vma.c
@@ -78,6 +78,9 @@ vma_create(struct drm_i915_gem_object *obj,
 	struct rb_node *rb, **p;
 	int i;
 
+	/* The aliasing_ppgtt should never be used directly! */
+	GEM_BUG_ON(vm == &vm->i915->mm.aliasing_ppgtt->base);
+
 	vma = kmem_cache_zalloc(vm->i915->vmas, GFP_KERNEL);
 	if (vma == NULL)
 		return ERR_PTR(-ENOMEM);
@@ -238,7 +241,15 @@ int i915_vma_bind(struct i915_vma *vma, enum i915_cache_level cache_level,
 	u32 vma_flags;
 	int ret;
 
-	if (WARN_ON(flags == 0))
+	GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+	GEM_BUG_ON(vma->size > vma->node.size);
+
+	if (GEM_WARN_ON(range_overflows(vma->node.start,
+					vma->node.size,
+					vma->vm->total)))
+		return -ENODEV;
+
+	if (GEM_WARN_ON(!flags))
 		return -EINVAL;
 
 	bind_flags = 0;
@@ -255,20 +266,6 @@ int i915_vma_bind(struct i915_vma *vma, enum i915_cache_level cache_level,
 	if (bind_flags == 0)
 		return 0;
 
-	if (GEM_WARN_ON(range_overflows(vma->node.start,
-					vma->node.size,
-					vma->vm->total)))
-		return -ENODEV;
-
-	if (vma_flags == 0 && vma->vm->allocate_va_range) {
-		trace_i915_va_alloc(vma);
-		ret = vma->vm->allocate_va_range(vma->vm,
-						 vma->node.start,
-						 vma->node.size);
-		if (ret)
-			return ret;
-	}
-
 	trace_i915_vma_bind(vma, bind_flags);
 	ret = vma->vm->bind_vma(vma, cache_level, bind_flags);
 	if (ret)
@@ -324,8 +321,8 @@ void i915_vma_unpin_and_release(struct i915_vma **p_vma)
 	__i915_gem_object_release_unless_active(obj);
 }
 
-bool
-i915_vma_misplaced(struct i915_vma *vma, u64 size, u64 alignment, u64 flags)
+bool i915_vma_misplaced(const struct i915_vma *vma,
+			u64 size, u64 alignment, u64 flags)
 {
 	if (!drm_mm_node_allocated(&vma->node))
 		return false;
@@ -512,10 +509,36 @@ err_unpin:
 	return ret;
 }
 
+static void
+i915_vma_remove(struct i915_vma *vma)
+{
+	struct drm_i915_gem_object *obj = vma->obj;
+
+	GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+	GEM_BUG_ON(vma->flags & (I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND));
+
+	drm_mm_remove_node(&vma->node);
+	list_move_tail(&vma->vm_link, &vma->vm->unbound_list);
+
+	/* Since the unbound list is global, only move to that list if
+	 * no more VMAs exist.
+	 */
+	if (--obj->bind_count == 0)
+		list_move_tail(&obj->global_link,
+			       &to_i915(obj->base.dev)->mm.unbound_list);
+
+	/* And finally now the object is completely decoupled from this vma,
+	 * we can drop its hold on the backing storage and allow it to be
+	 * reaped by the shrinker.
+	 */
+	i915_gem_object_unpin_pages(obj);
+	GEM_BUG_ON(atomic_read(&obj->mm.pages_pin_count) < obj->bind_count);
+}
+
 int __i915_vma_do_pin(struct i915_vma *vma,
 		      u64 size, u64 alignment, u64 flags)
 {
-	unsigned int bound = vma->flags;
+	const unsigned int bound = vma->flags;
 	int ret;
 
 	lockdep_assert_held(&vma->vm->i915->drm.struct_mutex);
@@ -524,18 +547,18 @@ int __i915_vma_do_pin(struct i915_vma *vma,
 
 	if (WARN_ON(bound & I915_VMA_PIN_OVERFLOW)) {
 		ret = -EBUSY;
-		goto err;
+		goto err_unpin;
 	}
 
 	if ((bound & I915_VMA_BIND_MASK) == 0) {
 		ret = i915_vma_insert(vma, size, alignment, flags);
 		if (ret)
-			goto err;
+			goto err_unpin;
 	}
 
 	ret = i915_vma_bind(vma, vma->obj->cache_level, flags);
 	if (ret)
-		goto err;
+		goto err_remove;
 
 	if ((bound ^ vma->flags) & I915_VMA_GLOBAL_BIND)
 		__i915_vma_set_map_and_fenceable(vma);
@@ -544,7 +567,12 @@ int __i915_vma_do_pin(struct i915_vma *vma,
 	GEM_BUG_ON(i915_vma_misplaced(vma, size, alignment, flags));
 	return 0;
 
-err:
+err_remove:
+	if ((bound & I915_VMA_BIND_MASK) == 0) {
+		GEM_BUG_ON(vma->pages);
+		i915_vma_remove(vma);
+	}
+err_unpin:
 	__i915_vma_unpin(vma);
 	return ret;
 }
@@ -657,9 +685,6 @@ int i915_vma_unbind(struct i915_vma *vma)
 	}
 	vma->flags &= ~(I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND);
 
-	drm_mm_remove_node(&vma->node);
-	list_move_tail(&vma->vm_link, &vma->vm->unbound_list);
-
 	if (vma->pages != obj->mm.pages) {
 		GEM_BUG_ON(!vma->pages);
 		sg_free_table(vma->pages);
@@ -667,18 +692,7 @@ int i915_vma_unbind(struct i915_vma *vma)
 	}
 	vma->pages = NULL;
 
-	/* Since the unbound list is global, only move to that list if
-	 * no more VMAs exist. */
-	if (--obj->bind_count == 0)
-		list_move_tail(&obj->global_link,
-			       &to_i915(obj->base.dev)->mm.unbound_list);
-
-	/* And finally now the object is completely decoupled from this vma,
-	 * we can drop its hold on the backing storage and allow it to be
-	 * reaped by the shrinker.
-	 */
-	i915_gem_object_unpin_pages(obj);
-	GEM_BUG_ON(atomic_read(&obj->mm.pages_pin_count) < obj->bind_count);
+	i915_vma_remove(vma);
 
 destroy:
 	if (unlikely(i915_vma_is_closed(vma)))
@@ -687,3 +701,6 @@ destroy:
 	return 0;
 }
 
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/i915_vma.c"
+#endif
diff --git a/drivers/gpu/drm/i915/i915_vma.h b/drivers/gpu/drm/i915/i915_vma.h
index e39d922cfb6f..2e03f81dddbe 100644
--- a/drivers/gpu/drm/i915/i915_vma.h
+++ b/drivers/gpu/drm/i915/i915_vma.h
@@ -228,8 +228,8 @@ i915_vma_compare(struct i915_vma *vma,
 int i915_vma_bind(struct i915_vma *vma, enum i915_cache_level cache_level,
 		  u32 flags);
 bool i915_gem_valid_gtt_space(struct i915_vma *vma, unsigned long cache_level);
-bool
-i915_vma_misplaced(struct i915_vma *vma, u64 size, u64 alignment, u64 flags);
+bool i915_vma_misplaced(const struct i915_vma *vma,
+			u64 size, u64 alignment, u64 flags);
 void __i915_vma_set_map_and_fenceable(struct i915_vma *vma);
 int __must_check i915_vma_unbind(struct i915_vma *vma);
 void i915_vma_close(struct i915_vma *vma);
diff --git a/drivers/gpu/drm/i915/intel_atomic.c b/drivers/gpu/drm/i915/intel_atomic.c
index aa9160e7f1d8..50fb1f76cc5f 100644
--- a/drivers/gpu/drm/i915/intel_atomic.c
+++ b/drivers/gpu/drm/i915/intel_atomic.c
@@ -99,6 +99,7 @@ intel_crtc_duplicate_state(struct drm_crtc *crtc)
 	crtc_state->update_wm_pre = false;
 	crtc_state->update_wm_post = false;
 	crtc_state->fb_changed = false;
+	crtc_state->fifo_changed = false;
 	crtc_state->wm.need_postvbl_update = false;
 	crtc_state->fb_bits = 0;
 
@@ -121,7 +122,7 @@ intel_crtc_destroy_state(struct drm_crtc *crtc,
 
 /**
  * intel_atomic_setup_scalers() - setup scalers for crtc per staged requests
- * @dev: DRM device
+ * @dev_priv: i915 device
  * @crtc: intel crtc
  * @crtc_state: incoming crtc_state to validate and setup scalers
  *
@@ -136,9 +137,9 @@ intel_crtc_destroy_state(struct drm_crtc *crtc,
  *         0 - scalers were setup succesfully
  *         error code - otherwise
  */
-int intel_atomic_setup_scalers(struct drm_device *dev,
-	struct intel_crtc *intel_crtc,
-	struct intel_crtc_state *crtc_state)
+int intel_atomic_setup_scalers(struct drm_i915_private *dev_priv,
+			       struct intel_crtc *intel_crtc,
+			       struct intel_crtc_state *crtc_state)
 {
 	struct drm_plane *plane = NULL;
 	struct intel_plane *intel_plane;
@@ -199,7 +200,7 @@ int intel_atomic_setup_scalers(struct drm_device *dev,
 			 */
 			if (!plane) {
 				struct drm_plane_state *state;
-				plane = drm_plane_from_index(dev, i);
+				plane = drm_plane_from_index(&dev_priv->drm, i);
 				state = drm_atomic_get_plane_state(drm_state, plane);
 				if (IS_ERR(state)) {
 					DRM_DEBUG_KMS("Failed to add [PLANE:%d] to drm_state\n",
@@ -247,7 +248,9 @@ int intel_atomic_setup_scalers(struct drm_device *dev,
 		}
 
 		/* set scaler mode */
-		if (num_scalers_need == 1 && intel_crtc->pipe != PIPE_C) {
+		if (IS_GEMINILAKE(dev_priv)) {
+			scaler_state->scalers[*scaler_id].mode = 0;
+		} else if (num_scalers_need == 1 && intel_crtc->pipe != PIPE_C) {
 			/*
 			 * when only 1 scaler is in use on either pipe A or B,
 			 * scaler 0 operates in high quality (HQ) mode.
diff --git a/drivers/gpu/drm/i915/intel_atomic_plane.c b/drivers/gpu/drm/i915/intel_atomic_plane.c
index 41fd94e62d3c..cfb47293fd53 100644
--- a/drivers/gpu/drm/i915/intel_atomic_plane.c
+++ b/drivers/gpu/drm/i915/intel_atomic_plane.c
@@ -189,6 +189,12 @@ int intel_plane_atomic_check_with_state(struct intel_crtc_state *crtc_state,
 	if (ret)
 		return ret;
 
+	/* FIXME pre-g4x don't work like this */
+	if (intel_state->base.visible)
+		crtc_state->active_planes |= BIT(intel_plane->id);
+	else
+		crtc_state->active_planes &= ~BIT(intel_plane->id);
+
 	return intel_plane_atomic_calc_changes(&crtc_state->base, state);
 }
 
@@ -225,12 +231,19 @@ static void intel_plane_atomic_update(struct drm_plane *plane,
 		to_intel_plane_state(plane->state);
 	struct drm_crtc *crtc = plane->state->crtc ?: old_state->crtc;
 
-	if (intel_state->base.visible)
+	if (intel_state->base.visible) {
+		trace_intel_update_plane(plane,
+					 to_intel_crtc(crtc));
+
 		intel_plane->update_plane(plane,
 					  to_intel_crtc_state(crtc->state),
 					  intel_state);
-	else
+	} else {
+		trace_intel_disable_plane(plane,
+					  to_intel_crtc(crtc));
+
 		intel_plane->disable_plane(plane, crtc);
+	}
 }
 
 const struct drm_plane_helper_funcs intel_plane_helper_funcs = {
diff --git a/drivers/gpu/drm/i915/intel_audio.c b/drivers/gpu/drm/i915/intel_audio.c
index d76f3033e890..52c207e81f41 100644
--- a/drivers/gpu/drm/i915/intel_audio.c
+++ b/drivers/gpu/drm/i915/intel_audio.c
@@ -720,7 +720,7 @@ static void i915_audio_component_codec_wake_override(struct device *kdev,
 	struct drm_i915_private *dev_priv = kdev_to_i915(kdev);
 	u32 tmp;
 
-	if (!IS_SKYLAKE(dev_priv) && !IS_KABYLAKE(dev_priv))
+	if (!IS_GEN9_BC(dev_priv))
 		return;
 
 	i915_audio_component_get_power(kdev);
@@ -752,7 +752,7 @@ static int i915_audio_component_get_cdclk_freq(struct device *kdev)
 	if (WARN_ON_ONCE(!HAS_DDI(dev_priv)))
 		return -ENODEV;
 
-	return dev_priv->cdclk_freq;
+	return dev_priv->cdclk.hw.cdclk;
 }
 
 /*
diff --git a/drivers/gpu/drm/i915/intel_breadcrumbs.c b/drivers/gpu/drm/i915/intel_breadcrumbs.c
index 7044e9a6abf7..2393bb9fe665 100644
--- a/drivers/gpu/drm/i915/intel_breadcrumbs.c
+++ b/drivers/gpu/drm/i915/intel_breadcrumbs.c
@@ -27,43 +27,90 @@
 
 #include "i915_drv.h"
 
+static unsigned int __intel_breadcrumbs_wakeup(struct intel_breadcrumbs *b)
+{
+	struct intel_wait *wait;
+	unsigned int result = 0;
+
+	lockdep_assert_held(&b->irq_lock);
+
+	wait = b->irq_wait;
+	if (wait) {
+		result = ENGINE_WAKEUP_WAITER;
+		if (wake_up_process(wait->tsk))
+			result |= ENGINE_WAKEUP_ASLEEP;
+	}
+
+	return result;
+}
+
+unsigned int intel_engine_wakeup(struct intel_engine_cs *engine)
+{
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+	unsigned long flags;
+	unsigned int result;
+
+	spin_lock_irqsave(&b->irq_lock, flags);
+	result = __intel_breadcrumbs_wakeup(b);
+	spin_unlock_irqrestore(&b->irq_lock, flags);
+
+	return result;
+}
+
+static unsigned long wait_timeout(void)
+{
+	return round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES);
+}
+
+static noinline void missed_breadcrumb(struct intel_engine_cs *engine)
+{
+	DRM_DEBUG_DRIVER("%s missed breadcrumb at %pF, irq posted? %s\n",
+			 engine->name, __builtin_return_address(0),
+			 yesno(test_bit(ENGINE_IRQ_BREADCRUMB,
+					&engine->irq_posted)));
+
+	set_bit(engine->id, &engine->i915->gpu_error.missed_irq_rings);
+}
+
 static void intel_breadcrumbs_hangcheck(unsigned long data)
 {
 	struct intel_engine_cs *engine = (struct intel_engine_cs *)data;
 	struct intel_breadcrumbs *b = &engine->breadcrumbs;
 
-	if (!b->irq_enabled)
+	if (!b->irq_armed)
 		return;
 
-	if (time_before(jiffies, b->timeout)) {
-		mod_timer(&b->hangcheck, b->timeout);
+	if (b->hangcheck_interrupts != atomic_read(&engine->irq_count)) {
+		b->hangcheck_interrupts = atomic_read(&engine->irq_count);
+		mod_timer(&b->hangcheck, wait_timeout());
 		return;
 	}
 
-	DRM_DEBUG("Hangcheck timer elapsed... %s idle\n", engine->name);
-	set_bit(engine->id, &engine->i915->gpu_error.missed_irq_rings);
-	mod_timer(&engine->breadcrumbs.fake_irq, jiffies + 1);
-
-	/* Ensure that even if the GPU hangs, we get woken up.
+	/* We keep the hangcheck time alive until we disarm the irq, even
+	 * if there are no waiters at present.
 	 *
-	 * However, note that if no one is waiting, we never notice
-	 * a gpu hang. Eventually, we will have to wait for a resource
-	 * held by the GPU and so trigger a hangcheck. In the most
-	 * pathological case, this will be upon memory starvation! To
-	 * prevent this, we also queue the hangcheck from the retire
-	 * worker.
+	 * If the waiter was currently running, assume it hasn't had a chance
+	 * to process the pending interrupt (e.g, low priority task on a loaded
+	 * system) and wait until it sleeps before declaring a missed interrupt.
+	 *
+	 * If the waiter was asleep (and not even pending a wakeup), then we
+	 * must have missed an interrupt as the GPU has stopped advancing
+	 * but we still have a waiter. Assuming all batches complete within
+	 * DRM_I915_HANGCHECK_JIFFIES [1.5s]!
 	 */
-	i915_queue_hangcheck(engine->i915);
-}
-
-static unsigned long wait_timeout(void)
-{
-	return round_jiffies_up(jiffies + DRM_I915_HANGCHECK_JIFFIES);
+	if (intel_engine_wakeup(engine) & ENGINE_WAKEUP_ASLEEP) {
+		missed_breadcrumb(engine);
+		mod_timer(&engine->breadcrumbs.fake_irq, jiffies + 1);
+	} else {
+		mod_timer(&b->hangcheck, wait_timeout());
+	}
 }
 
 static void intel_breadcrumbs_fake_irq(unsigned long data)
 {
 	struct intel_engine_cs *engine = (struct intel_engine_cs *)data;
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+	unsigned long flags;
 
 	/*
 	 * The timer persists in case we cannot enable interrupts,
@@ -72,8 +119,26 @@ static void intel_breadcrumbs_fake_irq(unsigned long data)
 	 * every jiffie in order to kick the oldest waiter to do the
 	 * coherent seqno check.
 	 */
-	if (intel_engine_wakeup(engine))
-		mod_timer(&engine->breadcrumbs.fake_irq, jiffies + 1);
+
+	spin_lock_irqsave(&b->irq_lock, flags);
+	if (!__intel_breadcrumbs_wakeup(b))
+		__intel_engine_disarm_breadcrumbs(engine);
+	spin_unlock_irqrestore(&b->irq_lock, flags);
+	if (!b->irq_armed)
+		return;
+
+	mod_timer(&b->fake_irq, jiffies + 1);
+
+	/* Ensure that even if the GPU hangs, we get woken up.
+	 *
+	 * However, note that if no one is waiting, we never notice
+	 * a gpu hang. Eventually, we will have to wait for a resource
+	 * held by the GPU and so trigger a hangcheck. In the most
+	 * pathological case, this will be upon memory starvation! To
+	 * prevent this, we also queue the hangcheck from the retire
+	 * worker.
+	 */
+	i915_queue_hangcheck(engine->i915);
 }
 
 static void irq_enable(struct intel_engine_cs *engine)
@@ -82,7 +147,7 @@ static void irq_enable(struct intel_engine_cs *engine)
 	 * we still need to force the barrier before reading the seqno,
 	 * just in case.
 	 */
-	engine->breadcrumbs.irq_posted = true;
+	set_bit(ENGINE_IRQ_BREADCRUMB, &engine->irq_posted);
 
 	/* Caller disables interrupts */
 	spin_lock(&engine->i915->irq_lock);
@@ -96,61 +161,114 @@ static void irq_disable(struct intel_engine_cs *engine)
 	spin_lock(&engine->i915->irq_lock);
 	engine->irq_disable(engine);
 	spin_unlock(&engine->i915->irq_lock);
+}
+
+void __intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine)
+{
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+
+	lockdep_assert_held(&b->irq_lock);
 
-	engine->breadcrumbs.irq_posted = false;
+	if (b->irq_enabled) {
+		irq_disable(engine);
+		b->irq_enabled = false;
+	}
+
+	b->irq_armed = false;
 }
 
-static void __intel_breadcrumbs_enable_irq(struct intel_breadcrumbs *b)
+void intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine)
 {
-	struct intel_engine_cs *engine =
-		container_of(b, struct intel_engine_cs, breadcrumbs);
-	struct drm_i915_private *i915 = engine->i915;
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+	unsigned long flags;
 
-	assert_spin_locked(&b->lock);
-	if (b->rpm_wakelock)
+	if (!b->irq_armed)
 		return;
 
-	/* Since we are waiting on a request, the GPU should be busy
-	 * and should have its own rpm reference. For completeness,
-	 * record an rpm reference for ourselves to cover the
-	 * interrupt we unmask.
+	spin_lock_irqsave(&b->irq_lock, flags);
+
+	/* We only disarm the irq when we are idle (all requests completed),
+	 * so if there remains a sleeping waiter, it missed the request
+	 * completion.
 	 */
-	intel_runtime_pm_get_noresume(i915);
-	b->rpm_wakelock = true;
+	if (__intel_breadcrumbs_wakeup(b) & ENGINE_WAKEUP_ASLEEP)
+		missed_breadcrumb(engine);
 
-	/* No interrupts? Kick the waiter every jiffie! */
-	if (intel_irqs_enabled(i915)) {
-		if (!test_bit(engine->id, &i915->gpu_error.test_irq_rings))
-			irq_enable(engine);
-		b->irq_enabled = true;
-	}
+	__intel_engine_disarm_breadcrumbs(engine);
+
+	spin_unlock_irqrestore(&b->irq_lock, flags);
+}
+
+static bool use_fake_irq(const struct intel_breadcrumbs *b)
+{
+	const struct intel_engine_cs *engine =
+		container_of(b, struct intel_engine_cs, breadcrumbs);
+
+	if (!test_bit(engine->id, &engine->i915->gpu_error.missed_irq_rings))
+		return false;
 
-	if (!b->irq_enabled ||
-	    test_bit(engine->id, &i915->gpu_error.missed_irq_rings)) {
+	/* Only start with the heavy weight fake irq timer if we have not
+	 * seen any interrupts since enabling it the first time. If the
+	 * interrupts are still arriving, it means we made a mistake in our
+	 * engine->seqno_barrier(), a timing error that should be transient
+	 * and unlikely to reoccur.
+	 */
+	return atomic_read(&engine->irq_count) == b->hangcheck_interrupts;
+}
+
+static void enable_fake_irq(struct intel_breadcrumbs *b)
+{
+	/* Ensure we never sleep indefinitely */
+	if (!b->irq_enabled || use_fake_irq(b))
 		mod_timer(&b->fake_irq, jiffies + 1);
-	} else {
-		/* Ensure we never sleep indefinitely */
-		GEM_BUG_ON(!time_after(b->timeout, jiffies));
-		mod_timer(&b->hangcheck, b->timeout);
-	}
+	else
+		mod_timer(&b->hangcheck, wait_timeout());
 }
 
-static void __intel_breadcrumbs_disable_irq(struct intel_breadcrumbs *b)
+static void __intel_breadcrumbs_enable_irq(struct intel_breadcrumbs *b)
 {
 	struct intel_engine_cs *engine =
 		container_of(b, struct intel_engine_cs, breadcrumbs);
+	struct drm_i915_private *i915 = engine->i915;
 
-	assert_spin_locked(&b->lock);
-	if (!b->rpm_wakelock)
+	lockdep_assert_held(&b->irq_lock);
+	if (b->irq_armed)
 		return;
 
-	if (b->irq_enabled) {
-		irq_disable(engine);
-		b->irq_enabled = false;
+	/* The breadcrumb irq will be disarmed on the interrupt after the
+	 * waiters are signaled. This gives us a single interrupt window in
+	 * which we can add a new waiter and avoid the cost of re-enabling
+	 * the irq.
+	 */
+	b->irq_armed = true;
+	GEM_BUG_ON(b->irq_enabled);
+
+	if (I915_SELFTEST_ONLY(b->mock)) {
+		/* For our mock objects we want to avoid interaction
+		 * with the real hardware (which is not set up). So
+		 * we simply pretend we have enabled the powerwell
+		 * and the irq, and leave it up to the mock
+		 * implementation to call intel_engine_wakeup()
+		 * itself when it wants to simulate a user interrupt,
+		 */
+		return;
+	}
+
+	/* Since we are waiting on a request, the GPU should be busy
+	 * and should have its own rpm reference. This is tracked
+	 * by i915->gt.awake, we can forgo holding our own wakref
+	 * for the interrupt as before i915->gt.awake is released (when
+	 * the driver is idle) we disarm the breadcrumbs.
+	 */
+
+	/* No interrupts? Kick the waiter every jiffie! */
+	if (intel_irqs_enabled(i915)) {
+		if (!test_bit(engine->id, &i915->gpu_error.test_irq_rings))
+			irq_enable(engine);
+		b->irq_enabled = true;
 	}
 
-	intel_runtime_pm_put(engine->i915);
-	b->rpm_wakelock = false;
+	enable_fake_irq(b);
 }
 
 static inline struct intel_wait *to_wait(struct rb_node *node)
@@ -161,7 +279,7 @@ static inline struct intel_wait *to_wait(struct rb_node *node)
 static inline void __intel_breadcrumbs_finish(struct intel_breadcrumbs *b,
 					      struct intel_wait *wait)
 {
-	assert_spin_locked(&b->lock);
+	lockdep_assert_held(&b->rb_lock);
 
 	/* This request is completed, so remove it from the tree, mark it as
 	 * complete, and *then* wake up the associated task.
@@ -172,6 +290,24 @@ static inline void __intel_breadcrumbs_finish(struct intel_breadcrumbs *b,
 	wake_up_process(wait->tsk); /* implicit smp_wmb() */
 }
 
+static inline void __intel_breadcrumbs_next(struct intel_engine_cs *engine,
+					    struct rb_node *next)
+{
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+
+	spin_lock(&b->irq_lock);
+	GEM_BUG_ON(!b->irq_armed);
+	b->irq_wait = to_wait(next);
+	spin_unlock(&b->irq_lock);
+
+	/* We always wake up the next waiter that takes over as the bottom-half
+	 * as we may delegate not only the irq-seqno barrier to the next waiter
+	 * but also the task of waking up concurrent waiters.
+	 */
+	if (next)
+		wake_up_process(to_wait(next)->tsk);
+}
+
 static bool __intel_engine_add_wait(struct intel_engine_cs *engine,
 				    struct intel_wait *wait)
 {
@@ -235,7 +371,6 @@ static bool __intel_engine_add_wait(struct intel_engine_cs *engine,
 	}
 	rb_link_node(&wait->node, parent, p);
 	rb_insert_color(&wait->node, &b->waiters);
-	GEM_BUG_ON(!first && !rcu_access_pointer(b->irq_seqno_bh));
 
 	if (completed) {
 		struct rb_node *next = rb_next(completed);
@@ -243,22 +378,7 @@ static bool __intel_engine_add_wait(struct intel_engine_cs *engine,
 		GEM_BUG_ON(!next && !first);
 		if (next && next != &wait->node) {
 			GEM_BUG_ON(first);
-			b->timeout = wait_timeout();
-			b->first_wait = to_wait(next);
-			rcu_assign_pointer(b->irq_seqno_bh, b->first_wait->tsk);
-			/* As there is a delay between reading the current
-			 * seqno, processing the completed tasks and selecting
-			 * the next waiter, we may have missed the interrupt
-			 * and so need for the next bottom-half to wakeup.
-			 *
-			 * Also as we enable the IRQ, we may miss the
-			 * interrupt for that seqno, so we have to wake up
-			 * the next bottom-half in order to do a coherent check
-			 * in case the seqno passed.
-			 */
-			__intel_breadcrumbs_enable_irq(b);
-			if (READ_ONCE(b->irq_posted))
-				wake_up_process(to_wait(next)->tsk);
+			__intel_breadcrumbs_next(engine, next);
 		}
 
 		do {
@@ -269,10 +389,9 @@ static bool __intel_engine_add_wait(struct intel_engine_cs *engine,
 	}
 
 	if (first) {
+		spin_lock(&b->irq_lock);
 		GEM_BUG_ON(rb_first(&b->waiters) != &wait->node);
-		b->timeout = wait_timeout();
-		b->first_wait = wait;
-		rcu_assign_pointer(b->irq_seqno_bh, wait->tsk);
+		b->irq_wait = wait;
 		/* After assigning ourselves as the new bottom-half, we must
 		 * perform a cursory check to prevent a missed interrupt.
 		 * Either we miss the interrupt whilst programming the hardware,
@@ -282,10 +401,10 @@ static bool __intel_engine_add_wait(struct intel_engine_cs *engine,
 		 * and so we miss the wake up.
 		 */
 		__intel_breadcrumbs_enable_irq(b);
+		spin_unlock(&b->irq_lock);
 	}
-	GEM_BUG_ON(!rcu_access_pointer(b->irq_seqno_bh));
-	GEM_BUG_ON(!b->first_wait);
-	GEM_BUG_ON(rb_first(&b->waiters) != &b->first_wait->node);
+	GEM_BUG_ON(!b->irq_wait);
+	GEM_BUG_ON(rb_first(&b->waiters) != &b->irq_wait->node);
 
 	return first;
 }
@@ -296,9 +415,9 @@ bool intel_engine_add_wait(struct intel_engine_cs *engine,
 	struct intel_breadcrumbs *b = &engine->breadcrumbs;
 	bool first;
 
-	spin_lock_irq(&b->lock);
+	spin_lock_irq(&b->rb_lock);
 	first = __intel_engine_add_wait(engine, wait);
-	spin_unlock_irq(&b->lock);
+	spin_unlock_irq(&b->rb_lock);
 
 	return first;
 }
@@ -317,29 +436,20 @@ static inline int wakeup_priority(struct intel_breadcrumbs *b,
 		return tsk->prio;
 }
 
-void intel_engine_remove_wait(struct intel_engine_cs *engine,
-			      struct intel_wait *wait)
+static void __intel_engine_remove_wait(struct intel_engine_cs *engine,
+				       struct intel_wait *wait)
 {
 	struct intel_breadcrumbs *b = &engine->breadcrumbs;
 
-	/* Quick check to see if this waiter was already decoupled from
-	 * the tree by the bottom-half to avoid contention on the spinlock
-	 * by the herd.
-	 */
-	if (RB_EMPTY_NODE(&wait->node))
-		return;
-
-	spin_lock_irq(&b->lock);
+	lockdep_assert_held(&b->rb_lock);
 
 	if (RB_EMPTY_NODE(&wait->node))
-		goto out_unlock;
+		goto out;
 
-	if (b->first_wait == wait) {
+	if (b->irq_wait == wait) {
 		const int priority = wakeup_priority(b, wait->tsk);
 		struct rb_node *next;
 
-		GEM_BUG_ON(rcu_access_pointer(b->irq_seqno_bh) != wait->tsk);
-
 		/* We are the current bottom-half. Find the next candidate,
 		 * the first waiter in the queue on the remaining oldest
 		 * request. As multiple seqnos may complete in the time it
@@ -372,25 +482,7 @@ void intel_engine_remove_wait(struct intel_engine_cs *engine,
 			}
 		}
 
-		if (next) {
-			/* In our haste, we may have completed the first waiter
-			 * before we enabled the interrupt. Do so now as we
-			 * have a second waiter for a future seqno. Afterwards,
-			 * we have to wake up that waiter in case we missed
-			 * the interrupt, or if we have to handle an
-			 * exception rather than a seqno completion.
-			 */
-			b->timeout = wait_timeout();
-			b->first_wait = to_wait(next);
-			rcu_assign_pointer(b->irq_seqno_bh, b->first_wait->tsk);
-			if (b->first_wait->seqno != wait->seqno)
-				__intel_breadcrumbs_enable_irq(b);
-			wake_up_process(b->first_wait->tsk);
-		} else {
-			b->first_wait = NULL;
-			rcu_assign_pointer(b->irq_seqno_bh, NULL);
-			__intel_breadcrumbs_disable_irq(b);
-		}
+		__intel_breadcrumbs_next(engine, next);
 	} else {
 		GEM_BUG_ON(rb_first(&b->waiters) == &wait->node);
 	}
@@ -398,15 +490,35 @@ void intel_engine_remove_wait(struct intel_engine_cs *engine,
 	GEM_BUG_ON(RB_EMPTY_NODE(&wait->node));
 	rb_erase(&wait->node, &b->waiters);
 
-out_unlock:
-	GEM_BUG_ON(b->first_wait == wait);
+out:
+	GEM_BUG_ON(b->irq_wait == wait);
 	GEM_BUG_ON(rb_first(&b->waiters) !=
-		   (b->first_wait ? &b->first_wait->node : NULL));
-	GEM_BUG_ON(!rcu_access_pointer(b->irq_seqno_bh) ^ RB_EMPTY_ROOT(&b->waiters));
-	spin_unlock_irq(&b->lock);
+		   (b->irq_wait ? &b->irq_wait->node : NULL));
+}
+
+void intel_engine_remove_wait(struct intel_engine_cs *engine,
+			      struct intel_wait *wait)
+{
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+
+	/* Quick check to see if this waiter was already decoupled from
+	 * the tree by the bottom-half to avoid contention on the spinlock
+	 * by the herd.
+	 */
+	if (RB_EMPTY_NODE(&wait->node))
+		return;
+
+	spin_lock_irq(&b->rb_lock);
+	__intel_engine_remove_wait(engine, wait);
+	spin_unlock_irq(&b->rb_lock);
+}
+
+static bool signal_valid(const struct drm_i915_gem_request *request)
+{
+	return intel_wait_check_request(&request->signaling.wait, request);
 }
 
-static bool signal_complete(struct drm_i915_gem_request *request)
+static bool signal_complete(const struct drm_i915_gem_request *request)
 {
 	if (!request)
 		return false;
@@ -415,7 +527,7 @@ static bool signal_complete(struct drm_i915_gem_request *request)
 	 * signalled that this wait is already completed.
 	 */
 	if (intel_wait_complete(&request->signaling.wait))
-		return true;
+		return signal_valid(request);
 
 	/* Carefully check if the request is complete, giving time for the
 	 * seqno to be visible or if the GPU hung.
@@ -458,40 +570,62 @@ static int intel_breadcrumbs_signaler(void *arg)
 		 * need to wait for a new interrupt from the GPU or for
 		 * a new client.
 		 */
-		request = READ_ONCE(b->first_signal);
+		rcu_read_lock();
+		request = rcu_dereference(b->first_signal);
+		if (request)
+			request = i915_gem_request_get_rcu(request);
+		rcu_read_unlock();
 		if (signal_complete(request)) {
-			/* Wake up all other completed waiters and select the
-			 * next bottom-half for the next user interrupt.
-			 */
-			intel_engine_remove_wait(engine,
-						 &request->signaling.wait);
-
 			local_bh_disable();
 			dma_fence_signal(&request->fence);
 			local_bh_enable(); /* kick start the tasklets */
 
+			spin_lock_irq(&b->rb_lock);
+
+			/* Wake up all other completed waiters and select the
+			 * next bottom-half for the next user interrupt.
+			 */
+			__intel_engine_remove_wait(engine,
+						   &request->signaling.wait);
+
 			/* Find the next oldest signal. Note that as we have
 			 * not been holding the lock, another client may
 			 * have installed an even older signal than the one
 			 * we just completed - so double check we are still
 			 * the oldest before picking the next one.
 			 */
-			spin_lock_irq(&b->lock);
-			if (request == b->first_signal) {
+			if (request == rcu_access_pointer(b->first_signal)) {
 				struct rb_node *rb =
 					rb_next(&request->signaling.node);
-				b->first_signal = rb ? to_signaler(rb) : NULL;
+				rcu_assign_pointer(b->first_signal,
+						   rb ? to_signaler(rb) : NULL);
 			}
 			rb_erase(&request->signaling.node, &b->signals);
-			spin_unlock_irq(&b->lock);
+			RB_CLEAR_NODE(&request->signaling.node);
+
+			spin_unlock_irq(&b->rb_lock);
 
 			i915_gem_request_put(request);
 		} else {
-			if (kthread_should_stop())
+			DEFINE_WAIT(exec);
+
+			if (kthread_should_stop()) {
+				GEM_BUG_ON(request);
 				break;
+			}
+
+			if (request)
+				add_wait_queue(&request->execute, &exec);
 
 			schedule();
+
+			if (request)
+				remove_wait_queue(&request->execute, &exec);
+
+			if (kthread_should_park())
+				kthread_parkme();
 		}
+		i915_gem_request_put(request);
 	} while (1);
 	__set_current_state(TASK_RUNNING);
 
@@ -504,6 +638,7 @@ void intel_engine_enable_signaling(struct drm_i915_gem_request *request)
 	struct intel_breadcrumbs *b = &engine->breadcrumbs;
 	struct rb_node *parent, **p;
 	bool first, wakeup;
+	u32 seqno;
 
 	/* Note that we may be called from an interrupt handler on another
 	 * device (e.g. nouveau signaling a fence completion causing us
@@ -513,15 +648,19 @@ void intel_engine_enable_signaling(struct drm_i915_gem_request *request)
 	 */
 
 	/* locked by dma_fence_enable_sw_signaling() (irqsafe fence->lock) */
-	assert_spin_locked(&request->lock);
-	if (!request->global_seqno)
+	GEM_BUG_ON(!irqs_disabled());
+	lockdep_assert_held(&request->lock);
+
+	seqno = i915_gem_request_global_seqno(request);
+	if (!seqno)
 		return;
 
 	request->signaling.wait.tsk = b->signaler;
-	request->signaling.wait.seqno = request->global_seqno;
+	request->signaling.wait.request = request;
+	request->signaling.wait.seqno = seqno;
 	i915_gem_request_get(request);
 
-	spin_lock(&b->lock);
+	spin_lock(&b->rb_lock);
 
 	/* First add ourselves into the list of waiters, but register our
 	 * bottom-half as the signaller thread. As per usual, only the oldest
@@ -542,8 +681,8 @@ void intel_engine_enable_signaling(struct drm_i915_gem_request *request)
 	p = &b->signals.rb_node;
 	while (*p) {
 		parent = *p;
-		if (i915_seqno_passed(request->global_seqno,
-				      to_signaler(parent)->global_seqno)) {
+		if (i915_seqno_passed(seqno,
+				      to_signaler(parent)->signaling.wait.seqno)) {
 			p = &parent->rb_right;
 			first = false;
 		} else {
@@ -553,20 +692,52 @@ void intel_engine_enable_signaling(struct drm_i915_gem_request *request)
 	rb_link_node(&request->signaling.node, parent, p);
 	rb_insert_color(&request->signaling.node, &b->signals);
 	if (first)
-		smp_store_mb(b->first_signal, request);
+		rcu_assign_pointer(b->first_signal, request);
 
-	spin_unlock(&b->lock);
+	spin_unlock(&b->rb_lock);
 
 	if (wakeup)
 		wake_up_process(b->signaler);
 }
 
+void intel_engine_cancel_signaling(struct drm_i915_gem_request *request)
+{
+	struct intel_engine_cs *engine = request->engine;
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+
+	GEM_BUG_ON(!irqs_disabled());
+	lockdep_assert_held(&request->lock);
+	GEM_BUG_ON(!request->signaling.wait.seqno);
+
+	spin_lock(&b->rb_lock);
+
+	if (!RB_EMPTY_NODE(&request->signaling.node)) {
+		if (request == rcu_access_pointer(b->first_signal)) {
+			struct rb_node *rb =
+				rb_next(&request->signaling.node);
+			rcu_assign_pointer(b->first_signal,
+					   rb ? to_signaler(rb) : NULL);
+		}
+		rb_erase(&request->signaling.node, &b->signals);
+		RB_CLEAR_NODE(&request->signaling.node);
+		i915_gem_request_put(request);
+	}
+
+	__intel_engine_remove_wait(engine, &request->signaling.wait);
+
+	spin_unlock(&b->rb_lock);
+
+	request->signaling.wait.seqno = 0;
+}
+
 int intel_engine_init_breadcrumbs(struct intel_engine_cs *engine)
 {
 	struct intel_breadcrumbs *b = &engine->breadcrumbs;
 	struct task_struct *tsk;
 
-	spin_lock_init(&b->lock);
+	spin_lock_init(&b->rb_lock);
+	spin_lock_init(&b->irq_lock);
+
 	setup_timer(&b->fake_irq,
 		    intel_breadcrumbs_fake_irq,
 		    (unsigned long)engine);
@@ -604,20 +775,26 @@ void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine)
 	struct intel_breadcrumbs *b = &engine->breadcrumbs;
 
 	cancel_fake_irq(engine);
-	spin_lock_irq(&b->lock);
+	spin_lock_irq(&b->irq_lock);
 
-	__intel_breadcrumbs_disable_irq(b);
-	if (intel_engine_has_waiter(engine)) {
-		b->timeout = wait_timeout();
-		__intel_breadcrumbs_enable_irq(b);
-		if (READ_ONCE(b->irq_posted))
-			wake_up_process(b->first_wait->tsk);
-	} else {
-		/* sanitize the IMR and unmask any auxiliary interrupts */
+	if (b->irq_enabled)
+		irq_enable(engine);
+	else
 		irq_disable(engine);
-	}
 
-	spin_unlock_irq(&b->lock);
+	/* We set the IRQ_BREADCRUMB bit when we enable the irq presuming the
+	 * GPU is active and may have already executed the MI_USER_INTERRUPT
+	 * before the CPU is ready to receive. However, the engine is currently
+	 * idle (we haven't started it yet), there is no possibility for a
+	 * missed interrupt as we enabled the irq and so we can clear the
+	 * immediate wakeup (until a real interrupt arrives for the waiter).
+	 */
+	clear_bit(ENGINE_IRQ_BREADCRUMB, &engine->irq_posted);
+
+	if (b->irq_armed)
+		enable_fake_irq(b);
+
+	spin_unlock_irq(&b->irq_lock);
 }
 
 void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine)
@@ -625,9 +802,9 @@ void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine)
 	struct intel_breadcrumbs *b = &engine->breadcrumbs;
 
 	/* The engines should be idle and all requests accounted for! */
-	WARN_ON(READ_ONCE(b->first_wait));
+	WARN_ON(READ_ONCE(b->irq_wait));
 	WARN_ON(!RB_EMPTY_ROOT(&b->waiters));
-	WARN_ON(READ_ONCE(b->first_signal));
+	WARN_ON(rcu_access_pointer(b->first_signal));
 	WARN_ON(!RB_EMPTY_ROOT(&b->signals));
 
 	if (!IS_ERR_OR_NULL(b->signaler))
@@ -636,29 +813,28 @@ void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine)
 	cancel_fake_irq(engine);
 }
 
-unsigned int intel_breadcrumbs_busy(struct drm_i915_private *i915)
+bool intel_breadcrumbs_busy(struct intel_engine_cs *engine)
 {
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-	unsigned int mask = 0;
-
-	for_each_engine(engine, i915, id) {
-		struct intel_breadcrumbs *b = &engine->breadcrumbs;
-
-		spin_lock_irq(&b->lock);
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+	bool busy = false;
 
-		if (b->first_wait) {
-			wake_up_process(b->first_wait->tsk);
-			mask |= intel_engine_flag(engine);
-		}
+	spin_lock_irq(&b->rb_lock);
 
-		if (b->first_signal) {
-			wake_up_process(b->signaler);
-			mask |= intel_engine_flag(engine);
-		}
+	if (b->irq_wait) {
+		wake_up_process(b->irq_wait->tsk);
+		busy |= intel_engine_flag(engine);
+	}
 
-		spin_unlock_irq(&b->lock);
+	if (rcu_access_pointer(b->first_signal)) {
+		wake_up_process(b->signaler);
+		busy |= intel_engine_flag(engine);
 	}
 
-	return mask;
+	spin_unlock_irq(&b->rb_lock);
+
+	return busy;
 }
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/intel_breadcrumbs.c"
+#endif
diff --git a/drivers/gpu/drm/i915/intel_cdclk.c b/drivers/gpu/drm/i915/intel_cdclk.c
new file mode 100644
index 000000000000..d643c0c5321b
--- /dev/null
+++ b/drivers/gpu/drm/i915/intel_cdclk.c
@@ -0,0 +1,1891 @@
+/*
+ * Copyright © 2006-2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#include "intel_drv.h"
+
+/**
+ * DOC: CDCLK / RAWCLK
+ *
+ * The display engine uses several different clocks to do its work. There
+ * are two main clocks involved that aren't directly related to the actual
+ * pixel clock or any symbol/bit clock of the actual output port. These
+ * are the core display clock (CDCLK) and RAWCLK.
+ *
+ * CDCLK clocks most of the display pipe logic, and thus its frequency
+ * must be high enough to support the rate at which pixels are flowing
+ * through the pipes. Downscaling must also be accounted as that increases
+ * the effective pixel rate.
+ *
+ * On several platforms the CDCLK frequency can be changed dynamically
+ * to minimize power consumption for a given display configuration.
+ * Typically changes to the CDCLK frequency require all the display pipes
+ * to be shut down while the frequency is being changed.
+ *
+ * On SKL+ the DMC will toggle the CDCLK off/on during DC5/6 entry/exit.
+ * DMC will not change the active CDCLK frequency however, so that part
+ * will still be performed by the driver directly.
+ *
+ * RAWCLK is a fixed frequency clock, often used by various auxiliary
+ * blocks such as AUX CH or backlight PWM. Hence the only thing we
+ * really need to know about RAWCLK is its frequency so that various
+ * dividers can be programmed correctly.
+ */
+
+static void fixed_133mhz_get_cdclk(struct drm_i915_private *dev_priv,
+				   struct intel_cdclk_state *cdclk_state)
+{
+	cdclk_state->cdclk = 133333;
+}
+
+static void fixed_200mhz_get_cdclk(struct drm_i915_private *dev_priv,
+				   struct intel_cdclk_state *cdclk_state)
+{
+	cdclk_state->cdclk = 200000;
+}
+
+static void fixed_266mhz_get_cdclk(struct drm_i915_private *dev_priv,
+				   struct intel_cdclk_state *cdclk_state)
+{
+	cdclk_state->cdclk = 266667;
+}
+
+static void fixed_333mhz_get_cdclk(struct drm_i915_private *dev_priv,
+				   struct intel_cdclk_state *cdclk_state)
+{
+	cdclk_state->cdclk = 333333;
+}
+
+static void fixed_400mhz_get_cdclk(struct drm_i915_private *dev_priv,
+				   struct intel_cdclk_state *cdclk_state)
+{
+	cdclk_state->cdclk = 400000;
+}
+
+static void fixed_450mhz_get_cdclk(struct drm_i915_private *dev_priv,
+				   struct intel_cdclk_state *cdclk_state)
+{
+	cdclk_state->cdclk = 450000;
+}
+
+static void i85x_get_cdclk(struct drm_i915_private *dev_priv,
+			   struct intel_cdclk_state *cdclk_state)
+{
+	struct pci_dev *pdev = dev_priv->drm.pdev;
+	u16 hpllcc = 0;
+
+	/*
+	 * 852GM/852GMV only supports 133 MHz and the HPLLCC
+	 * encoding is different :(
+	 * FIXME is this the right way to detect 852GM/852GMV?
+	 */
+	if (pdev->revision == 0x1) {
+		cdclk_state->cdclk = 133333;
+		return;
+	}
+
+	pci_bus_read_config_word(pdev->bus,
+				 PCI_DEVFN(0, 3), HPLLCC, &hpllcc);
+
+	/* Assume that the hardware is in the high speed state.  This
+	 * should be the default.
+	 */
+	switch (hpllcc & GC_CLOCK_CONTROL_MASK) {
+	case GC_CLOCK_133_200:
+	case GC_CLOCK_133_200_2:
+	case GC_CLOCK_100_200:
+		cdclk_state->cdclk = 200000;
+		break;
+	case GC_CLOCK_166_250:
+		cdclk_state->cdclk = 250000;
+		break;
+	case GC_CLOCK_100_133:
+		cdclk_state->cdclk = 133333;
+		break;
+	case GC_CLOCK_133_266:
+	case GC_CLOCK_133_266_2:
+	case GC_CLOCK_166_266:
+		cdclk_state->cdclk = 266667;
+		break;
+	}
+}
+
+static void i915gm_get_cdclk(struct drm_i915_private *dev_priv,
+			     struct intel_cdclk_state *cdclk_state)
+{
+	struct pci_dev *pdev = dev_priv->drm.pdev;
+	u16 gcfgc = 0;
+
+	pci_read_config_word(pdev, GCFGC, &gcfgc);
+
+	if (gcfgc & GC_LOW_FREQUENCY_ENABLE) {
+		cdclk_state->cdclk = 133333;
+		return;
+	}
+
+	switch (gcfgc & GC_DISPLAY_CLOCK_MASK) {
+	case GC_DISPLAY_CLOCK_333_320_MHZ:
+		cdclk_state->cdclk = 333333;
+		break;
+	default:
+	case GC_DISPLAY_CLOCK_190_200_MHZ:
+		cdclk_state->cdclk = 190000;
+		break;
+	}
+}
+
+static void i945gm_get_cdclk(struct drm_i915_private *dev_priv,
+			     struct intel_cdclk_state *cdclk_state)
+{
+	struct pci_dev *pdev = dev_priv->drm.pdev;
+	u16 gcfgc = 0;
+
+	pci_read_config_word(pdev, GCFGC, &gcfgc);
+
+	if (gcfgc & GC_LOW_FREQUENCY_ENABLE) {
+		cdclk_state->cdclk = 133333;
+		return;
+	}
+
+	switch (gcfgc & GC_DISPLAY_CLOCK_MASK) {
+	case GC_DISPLAY_CLOCK_333_320_MHZ:
+		cdclk_state->cdclk = 320000;
+		break;
+	default:
+	case GC_DISPLAY_CLOCK_190_200_MHZ:
+		cdclk_state->cdclk = 200000;
+		break;
+	}
+}
+
+static unsigned int intel_hpll_vco(struct drm_i915_private *dev_priv)
+{
+	static const unsigned int blb_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 4800000,
+		[4] = 6400000,
+	};
+	static const unsigned int pnv_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 4800000,
+		[4] = 2666667,
+	};
+	static const unsigned int cl_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 6400000,
+		[4] = 3333333,
+		[5] = 3566667,
+		[6] = 4266667,
+	};
+	static const unsigned int elk_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 4800000,
+	};
+	static const unsigned int ctg_vco[8] = {
+		[0] = 3200000,
+		[1] = 4000000,
+		[2] = 5333333,
+		[3] = 6400000,
+		[4] = 2666667,
+		[5] = 4266667,
+	};
+	const unsigned int *vco_table;
+	unsigned int vco;
+	uint8_t tmp = 0;
+
+	/* FIXME other chipsets? */
+	if (IS_GM45(dev_priv))
+		vco_table = ctg_vco;
+	else if (IS_G4X(dev_priv))
+		vco_table = elk_vco;
+	else if (IS_I965GM(dev_priv))
+		vco_table = cl_vco;
+	else if (IS_PINEVIEW(dev_priv))
+		vco_table = pnv_vco;
+	else if (IS_G33(dev_priv))
+		vco_table = blb_vco;
+	else
+		return 0;
+
+	tmp = I915_READ(IS_MOBILE(dev_priv) ? HPLLVCO_MOBILE : HPLLVCO);
+
+	vco = vco_table[tmp & 0x7];
+	if (vco == 0)
+		DRM_ERROR("Bad HPLL VCO (HPLLVCO=0x%02x)\n", tmp);
+	else
+		DRM_DEBUG_KMS("HPLL VCO %u kHz\n", vco);
+
+	return vco;
+}
+
+static void g33_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_state *cdclk_state)
+{
+	struct pci_dev *pdev = dev_priv->drm.pdev;
+	static const uint8_t div_3200[] = { 12, 10,  8,  7, 5, 16 };
+	static const uint8_t div_4000[] = { 14, 12, 10,  8, 6, 20 };
+	static const uint8_t div_4800[] = { 20, 14, 12, 10, 8, 24 };
+	static const uint8_t div_5333[] = { 20, 16, 12, 12, 8, 28 };
+	const uint8_t *div_table;
+	unsigned int cdclk_sel;
+	uint16_t tmp = 0;
+
+	cdclk_state->vco = intel_hpll_vco(dev_priv);
+
+	pci_read_config_word(pdev, GCFGC, &tmp);
+
+	cdclk_sel = (tmp >> 4) & 0x7;
+
+	if (cdclk_sel >= ARRAY_SIZE(div_3200))
+		goto fail;
+
+	switch (cdclk_state->vco) {
+	case 3200000:
+		div_table = div_3200;
+		break;
+	case 4000000:
+		div_table = div_4000;
+		break;
+	case 4800000:
+		div_table = div_4800;
+		break;
+	case 5333333:
+		div_table = div_5333;
+		break;
+	default:
+		goto fail;
+	}
+
+	cdclk_state->cdclk = DIV_ROUND_CLOSEST(cdclk_state->vco,
+					       div_table[cdclk_sel]);
+	return;
+
+fail:
+	DRM_ERROR("Unable to determine CDCLK. HPLL VCO=%u kHz, CFGC=0x%08x\n",
+		  cdclk_state->vco, tmp);
+	cdclk_state->cdclk = 190476;
+}
+
+static void pnv_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_state *cdclk_state)
+{
+	struct pci_dev *pdev = dev_priv->drm.pdev;
+	u16 gcfgc = 0;
+
+	pci_read_config_word(pdev, GCFGC, &gcfgc);
+
+	switch (gcfgc & GC_DISPLAY_CLOCK_MASK) {
+	case GC_DISPLAY_CLOCK_267_MHZ_PNV:
+		cdclk_state->cdclk = 266667;
+		break;
+	case GC_DISPLAY_CLOCK_333_MHZ_PNV:
+		cdclk_state->cdclk = 333333;
+		break;
+	case GC_DISPLAY_CLOCK_444_MHZ_PNV:
+		cdclk_state->cdclk = 444444;
+		break;
+	case GC_DISPLAY_CLOCK_200_MHZ_PNV:
+		cdclk_state->cdclk = 200000;
+		break;
+	default:
+		DRM_ERROR("Unknown pnv display core clock 0x%04x\n", gcfgc);
+	case GC_DISPLAY_CLOCK_133_MHZ_PNV:
+		cdclk_state->cdclk = 133333;
+		break;
+	case GC_DISPLAY_CLOCK_167_MHZ_PNV:
+		cdclk_state->cdclk = 166667;
+		break;
+	}
+}
+
+static void i965gm_get_cdclk(struct drm_i915_private *dev_priv,
+			     struct intel_cdclk_state *cdclk_state)
+{
+	struct pci_dev *pdev = dev_priv->drm.pdev;
+	static const uint8_t div_3200[] = { 16, 10,  8 };
+	static const uint8_t div_4000[] = { 20, 12, 10 };
+	static const uint8_t div_5333[] = { 24, 16, 14 };
+	const uint8_t *div_table;
+	unsigned int cdclk_sel;
+	uint16_t tmp = 0;
+
+	cdclk_state->vco = intel_hpll_vco(dev_priv);
+
+	pci_read_config_word(pdev, GCFGC, &tmp);
+
+	cdclk_sel = ((tmp >> 8) & 0x1f) - 1;
+
+	if (cdclk_sel >= ARRAY_SIZE(div_3200))
+		goto fail;
+
+	switch (cdclk_state->vco) {
+	case 3200000:
+		div_table = div_3200;
+		break;
+	case 4000000:
+		div_table = div_4000;
+		break;
+	case 5333333:
+		div_table = div_5333;
+		break;
+	default:
+		goto fail;
+	}
+
+	cdclk_state->cdclk = DIV_ROUND_CLOSEST(cdclk_state->vco,
+					       div_table[cdclk_sel]);
+	return;
+
+fail:
+	DRM_ERROR("Unable to determine CDCLK. HPLL VCO=%u kHz, CFGC=0x%04x\n",
+		  cdclk_state->vco, tmp);
+	cdclk_state->cdclk = 200000;
+}
+
+static void gm45_get_cdclk(struct drm_i915_private *dev_priv,
+			   struct intel_cdclk_state *cdclk_state)
+{
+	struct pci_dev *pdev = dev_priv->drm.pdev;
+	unsigned int cdclk_sel;
+	uint16_t tmp = 0;
+
+	cdclk_state->vco = intel_hpll_vco(dev_priv);
+
+	pci_read_config_word(pdev, GCFGC, &tmp);
+
+	cdclk_sel = (tmp >> 12) & 0x1;
+
+	switch (cdclk_state->vco) {
+	case 2666667:
+	case 4000000:
+	case 5333333:
+		cdclk_state->cdclk = cdclk_sel ? 333333 : 222222;
+		break;
+	case 3200000:
+		cdclk_state->cdclk = cdclk_sel ? 320000 : 228571;
+		break;
+	default:
+		DRM_ERROR("Unable to determine CDCLK. HPLL VCO=%u, CFGC=0x%04x\n",
+			  cdclk_state->vco, tmp);
+		cdclk_state->cdclk = 222222;
+		break;
+	}
+}
+
+static void hsw_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_state *cdclk_state)
+{
+	uint32_t lcpll = I915_READ(LCPLL_CTL);
+	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
+
+	if (lcpll & LCPLL_CD_SOURCE_FCLK)
+		cdclk_state->cdclk = 800000;
+	else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
+		cdclk_state->cdclk = 450000;
+	else if (freq == LCPLL_CLK_FREQ_450)
+		cdclk_state->cdclk = 450000;
+	else if (IS_HSW_ULT(dev_priv))
+		cdclk_state->cdclk = 337500;
+	else
+		cdclk_state->cdclk = 540000;
+}
+
+static int vlv_calc_cdclk(struct drm_i915_private *dev_priv,
+			  int max_pixclk)
+{
+	int freq_320 = (dev_priv->hpll_freq <<  1) % 320000 != 0 ?
+		333333 : 320000;
+	int limit = IS_CHERRYVIEW(dev_priv) ? 95 : 90;
+
+	/*
+	 * We seem to get an unstable or solid color picture at 200MHz.
+	 * Not sure what's wrong. For now use 200MHz only when all pipes
+	 * are off.
+	 */
+	if (!IS_CHERRYVIEW(dev_priv) &&
+	    max_pixclk > freq_320*limit/100)
+		return 400000;
+	else if (max_pixclk > 266667*limit/100)
+		return freq_320;
+	else if (max_pixclk > 0)
+		return 266667;
+	else
+		return 200000;
+}
+
+static void vlv_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_state *cdclk_state)
+{
+	cdclk_state->vco = vlv_get_hpll_vco(dev_priv);
+	cdclk_state->cdclk = vlv_get_cck_clock(dev_priv, "cdclk",
+					       CCK_DISPLAY_CLOCK_CONTROL,
+					       cdclk_state->vco);
+}
+
+static void vlv_program_pfi_credits(struct drm_i915_private *dev_priv)
+{
+	unsigned int credits, default_credits;
+
+	if (IS_CHERRYVIEW(dev_priv))
+		default_credits = PFI_CREDIT(12);
+	else
+		default_credits = PFI_CREDIT(8);
+
+	if (dev_priv->cdclk.hw.cdclk >= dev_priv->czclk_freq) {
+		/* CHV suggested value is 31 or 63 */
+		if (IS_CHERRYVIEW(dev_priv))
+			credits = PFI_CREDIT_63;
+		else
+			credits = PFI_CREDIT(15);
+	} else {
+		credits = default_credits;
+	}
+
+	/*
+	 * WA - write default credits before re-programming
+	 * FIXME: should we also set the resend bit here?
+	 */
+	I915_WRITE(GCI_CONTROL, VGA_FAST_MODE_DISABLE |
+		   default_credits);
+
+	I915_WRITE(GCI_CONTROL, VGA_FAST_MODE_DISABLE |
+		   credits | PFI_CREDIT_RESEND);
+
+	/*
+	 * FIXME is this guaranteed to clear
+	 * immediately or should we poll for it?
+	 */
+	WARN_ON(I915_READ(GCI_CONTROL) & PFI_CREDIT_RESEND);
+}
+
+static void vlv_set_cdclk(struct drm_i915_private *dev_priv,
+			  const struct intel_cdclk_state *cdclk_state)
+{
+	int cdclk = cdclk_state->cdclk;
+	u32 val, cmd;
+
+	if (cdclk >= 320000) /* jump to highest voltage for 400MHz too */
+		cmd = 2;
+	else if (cdclk == 266667)
+		cmd = 1;
+	else
+		cmd = 0;
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+	val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
+	val &= ~DSPFREQGUAR_MASK;
+	val |= (cmd << DSPFREQGUAR_SHIFT);
+	vlv_punit_write(dev_priv, PUNIT_REG_DSPFREQ, val);
+	if (wait_for((vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) &
+		      DSPFREQSTAT_MASK) == (cmd << DSPFREQSTAT_SHIFT),
+		     50)) {
+		DRM_ERROR("timed out waiting for CDclk change\n");
+	}
+	mutex_unlock(&dev_priv->rps.hw_lock);
+
+	mutex_lock(&dev_priv->sb_lock);
+
+	if (cdclk == 400000) {
+		u32 divider;
+
+		divider = DIV_ROUND_CLOSEST(dev_priv->hpll_freq << 1,
+					    cdclk) - 1;
+
+		/* adjust cdclk divider */
+		val = vlv_cck_read(dev_priv, CCK_DISPLAY_CLOCK_CONTROL);
+		val &= ~CCK_FREQUENCY_VALUES;
+		val |= divider;
+		vlv_cck_write(dev_priv, CCK_DISPLAY_CLOCK_CONTROL, val);
+
+		if (wait_for((vlv_cck_read(dev_priv, CCK_DISPLAY_CLOCK_CONTROL) &
+			      CCK_FREQUENCY_STATUS) == (divider << CCK_FREQUENCY_STATUS_SHIFT),
+			     50))
+			DRM_ERROR("timed out waiting for CDclk change\n");
+	}
+
+	/* adjust self-refresh exit latency value */
+	val = vlv_bunit_read(dev_priv, BUNIT_REG_BISOC);
+	val &= ~0x7f;
+
+	/*
+	 * For high bandwidth configs, we set a higher latency in the bunit
+	 * so that the core display fetch happens in time to avoid underruns.
+	 */
+	if (cdclk == 400000)
+		val |= 4500 / 250; /* 4.5 usec */
+	else
+		val |= 3000 / 250; /* 3.0 usec */
+	vlv_bunit_write(dev_priv, BUNIT_REG_BISOC, val);
+
+	mutex_unlock(&dev_priv->sb_lock);
+
+	intel_update_cdclk(dev_priv);
+
+	vlv_program_pfi_credits(dev_priv);
+}
+
+static void chv_set_cdclk(struct drm_i915_private *dev_priv,
+			  const struct intel_cdclk_state *cdclk_state)
+{
+	int cdclk = cdclk_state->cdclk;
+	u32 val, cmd;
+
+	switch (cdclk) {
+	case 333333:
+	case 320000:
+	case 266667:
+	case 200000:
+		break;
+	default:
+		MISSING_CASE(cdclk);
+		return;
+	}
+
+	/*
+	 * Specs are full of misinformation, but testing on actual
+	 * hardware has shown that we just need to write the desired
+	 * CCK divider into the Punit register.
+	 */
+	cmd = DIV_ROUND_CLOSEST(dev_priv->hpll_freq << 1, cdclk) - 1;
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+	val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
+	val &= ~DSPFREQGUAR_MASK_CHV;
+	val |= (cmd << DSPFREQGUAR_SHIFT_CHV);
+	vlv_punit_write(dev_priv, PUNIT_REG_DSPFREQ, val);
+	if (wait_for((vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) &
+		      DSPFREQSTAT_MASK_CHV) == (cmd << DSPFREQSTAT_SHIFT_CHV),
+		     50)) {
+		DRM_ERROR("timed out waiting for CDclk change\n");
+	}
+	mutex_unlock(&dev_priv->rps.hw_lock);
+
+	intel_update_cdclk(dev_priv);
+
+	vlv_program_pfi_credits(dev_priv);
+}
+
+static int bdw_calc_cdclk(int max_pixclk)
+{
+	if (max_pixclk > 540000)
+		return 675000;
+	else if (max_pixclk > 450000)
+		return 540000;
+	else if (max_pixclk > 337500)
+		return 450000;
+	else
+		return 337500;
+}
+
+static void bdw_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_state *cdclk_state)
+{
+	uint32_t lcpll = I915_READ(LCPLL_CTL);
+	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
+
+	if (lcpll & LCPLL_CD_SOURCE_FCLK)
+		cdclk_state->cdclk = 800000;
+	else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
+		cdclk_state->cdclk = 450000;
+	else if (freq == LCPLL_CLK_FREQ_450)
+		cdclk_state->cdclk = 450000;
+	else if (freq == LCPLL_CLK_FREQ_54O_BDW)
+		cdclk_state->cdclk = 540000;
+	else if (freq == LCPLL_CLK_FREQ_337_5_BDW)
+		cdclk_state->cdclk = 337500;
+	else
+		cdclk_state->cdclk = 675000;
+}
+
+static void bdw_set_cdclk(struct drm_i915_private *dev_priv,
+			  const struct intel_cdclk_state *cdclk_state)
+{
+	int cdclk = cdclk_state->cdclk;
+	uint32_t val, data;
+	int ret;
+
+	if (WARN((I915_READ(LCPLL_CTL) &
+		  (LCPLL_PLL_DISABLE | LCPLL_PLL_LOCK |
+		   LCPLL_CD_CLOCK_DISABLE | LCPLL_ROOT_CD_CLOCK_DISABLE |
+		   LCPLL_CD2X_CLOCK_DISABLE | LCPLL_POWER_DOWN_ALLOW |
+		   LCPLL_CD_SOURCE_FCLK)) != LCPLL_PLL_LOCK,
+		 "trying to change cdclk frequency with cdclk not enabled\n"))
+		return;
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+	ret = sandybridge_pcode_write(dev_priv,
+				      BDW_PCODE_DISPLAY_FREQ_CHANGE_REQ, 0x0);
+	mutex_unlock(&dev_priv->rps.hw_lock);
+	if (ret) {
+		DRM_ERROR("failed to inform pcode about cdclk change\n");
+		return;
+	}
+
+	val = I915_READ(LCPLL_CTL);
+	val |= LCPLL_CD_SOURCE_FCLK;
+	I915_WRITE(LCPLL_CTL, val);
+
+	if (wait_for_us(I915_READ(LCPLL_CTL) &
+			LCPLL_CD_SOURCE_FCLK_DONE, 1))
+		DRM_ERROR("Switching to FCLK failed\n");
+
+	val = I915_READ(LCPLL_CTL);
+	val &= ~LCPLL_CLK_FREQ_MASK;
+
+	switch (cdclk) {
+	case 450000:
+		val |= LCPLL_CLK_FREQ_450;
+		data = 0;
+		break;
+	case 540000:
+		val |= LCPLL_CLK_FREQ_54O_BDW;
+		data = 1;
+		break;
+	case 337500:
+		val |= LCPLL_CLK_FREQ_337_5_BDW;
+		data = 2;
+		break;
+	case 675000:
+		val |= LCPLL_CLK_FREQ_675_BDW;
+		data = 3;
+		break;
+	default:
+		WARN(1, "invalid cdclk frequency\n");
+		return;
+	}
+
+	I915_WRITE(LCPLL_CTL, val);
+
+	val = I915_READ(LCPLL_CTL);
+	val &= ~LCPLL_CD_SOURCE_FCLK;
+	I915_WRITE(LCPLL_CTL, val);
+
+	if (wait_for_us((I915_READ(LCPLL_CTL) &
+			LCPLL_CD_SOURCE_FCLK_DONE) == 0, 1))
+		DRM_ERROR("Switching back to LCPLL failed\n");
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+	sandybridge_pcode_write(dev_priv, HSW_PCODE_DE_WRITE_FREQ_REQ, data);
+	mutex_unlock(&dev_priv->rps.hw_lock);
+
+	I915_WRITE(CDCLK_FREQ, DIV_ROUND_CLOSEST(cdclk, 1000) - 1);
+
+	intel_update_cdclk(dev_priv);
+
+	WARN(cdclk != dev_priv->cdclk.hw.cdclk,
+	     "cdclk requested %d kHz but got %d kHz\n",
+	     cdclk, dev_priv->cdclk.hw.cdclk);
+}
+
+static int skl_calc_cdclk(int max_pixclk, int vco)
+{
+	if (vco == 8640000) {
+		if (max_pixclk > 540000)
+			return 617143;
+		else if (max_pixclk > 432000)
+			return 540000;
+		else if (max_pixclk > 308571)
+			return 432000;
+		else
+			return 308571;
+	} else {
+		if (max_pixclk > 540000)
+			return 675000;
+		else if (max_pixclk > 450000)
+			return 540000;
+		else if (max_pixclk > 337500)
+			return 450000;
+		else
+			return 337500;
+	}
+}
+
+static void skl_dpll0_update(struct drm_i915_private *dev_priv,
+			     struct intel_cdclk_state *cdclk_state)
+{
+	u32 val;
+
+	cdclk_state->ref = 24000;
+	cdclk_state->vco = 0;
+
+	val = I915_READ(LCPLL1_CTL);
+	if ((val & LCPLL_PLL_ENABLE) == 0)
+		return;
+
+	if (WARN_ON((val & LCPLL_PLL_LOCK) == 0))
+		return;
+
+	val = I915_READ(DPLL_CTRL1);
+
+	if (WARN_ON((val & (DPLL_CTRL1_HDMI_MODE(SKL_DPLL0) |
+			    DPLL_CTRL1_SSC(SKL_DPLL0) |
+			    DPLL_CTRL1_OVERRIDE(SKL_DPLL0))) !=
+		    DPLL_CTRL1_OVERRIDE(SKL_DPLL0)))
+		return;
+
+	switch (val & DPLL_CTRL1_LINK_RATE_MASK(SKL_DPLL0)) {
+	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_810, SKL_DPLL0):
+	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1350, SKL_DPLL0):
+	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1620, SKL_DPLL0):
+	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_2700, SKL_DPLL0):
+		cdclk_state->vco = 8100000;
+		break;
+	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1080, SKL_DPLL0):
+	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_2160, SKL_DPLL0):
+		cdclk_state->vco = 8640000;
+		break;
+	default:
+		MISSING_CASE(val & DPLL_CTRL1_LINK_RATE_MASK(SKL_DPLL0));
+		break;
+	}
+}
+
+static void skl_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_state *cdclk_state)
+{
+	u32 cdctl;
+
+	skl_dpll0_update(dev_priv, cdclk_state);
+
+	cdclk_state->cdclk = cdclk_state->ref;
+
+	if (cdclk_state->vco == 0)
+		return;
+
+	cdctl = I915_READ(CDCLK_CTL);
+
+	if (cdclk_state->vco == 8640000) {
+		switch (cdctl & CDCLK_FREQ_SEL_MASK) {
+		case CDCLK_FREQ_450_432:
+			cdclk_state->cdclk = 432000;
+			break;
+		case CDCLK_FREQ_337_308:
+			cdclk_state->cdclk = 308571;
+			break;
+		case CDCLK_FREQ_540:
+			cdclk_state->cdclk = 540000;
+			break;
+		case CDCLK_FREQ_675_617:
+			cdclk_state->cdclk = 617143;
+			break;
+		default:
+			MISSING_CASE(cdctl & CDCLK_FREQ_SEL_MASK);
+			break;
+		}
+	} else {
+		switch (cdctl & CDCLK_FREQ_SEL_MASK) {
+		case CDCLK_FREQ_450_432:
+			cdclk_state->cdclk = 450000;
+			break;
+		case CDCLK_FREQ_337_308:
+			cdclk_state->cdclk = 337500;
+			break;
+		case CDCLK_FREQ_540:
+			cdclk_state->cdclk = 540000;
+			break;
+		case CDCLK_FREQ_675_617:
+			cdclk_state->cdclk = 675000;
+			break;
+		default:
+			MISSING_CASE(cdctl & CDCLK_FREQ_SEL_MASK);
+			break;
+		}
+	}
+}
+
+/* convert from kHz to .1 fixpoint MHz with -1MHz offset */
+static int skl_cdclk_decimal(int cdclk)
+{
+	return DIV_ROUND_CLOSEST(cdclk - 1000, 500);
+}
+
+static void skl_set_preferred_cdclk_vco(struct drm_i915_private *dev_priv,
+					int vco)
+{
+	bool changed = dev_priv->skl_preferred_vco_freq != vco;
+
+	dev_priv->skl_preferred_vco_freq = vco;
+
+	if (changed)
+		intel_update_max_cdclk(dev_priv);
+}
+
+static void skl_dpll0_enable(struct drm_i915_private *dev_priv, int vco)
+{
+	int min_cdclk = skl_calc_cdclk(0, vco);
+	u32 val;
+
+	WARN_ON(vco != 8100000 && vco != 8640000);
+
+	/* select the minimum CDCLK before enabling DPLL 0 */
+	val = CDCLK_FREQ_337_308 | skl_cdclk_decimal(min_cdclk);
+	I915_WRITE(CDCLK_CTL, val);
+	POSTING_READ(CDCLK_CTL);
+
+	/*
+	 * We always enable DPLL0 with the lowest link rate possible, but still
+	 * taking into account the VCO required to operate the eDP panel at the
+	 * desired frequency. The usual DP link rates operate with a VCO of
+	 * 8100 while the eDP 1.4 alternate link rates need a VCO of 8640.
+	 * The modeset code is responsible for the selection of the exact link
+	 * rate later on, with the constraint of choosing a frequency that
+	 * works with vco.
+	 */
+	val = I915_READ(DPLL_CTRL1);
+
+	val &= ~(DPLL_CTRL1_HDMI_MODE(SKL_DPLL0) | DPLL_CTRL1_SSC(SKL_DPLL0) |
+		 DPLL_CTRL1_LINK_RATE_MASK(SKL_DPLL0));
+	val |= DPLL_CTRL1_OVERRIDE(SKL_DPLL0);
+	if (vco == 8640000)
+		val |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1080,
+					    SKL_DPLL0);
+	else
+		val |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_810,
+					    SKL_DPLL0);
+
+	I915_WRITE(DPLL_CTRL1, val);
+	POSTING_READ(DPLL_CTRL1);
+
+	I915_WRITE(LCPLL1_CTL, I915_READ(LCPLL1_CTL) | LCPLL_PLL_ENABLE);
+
+	if (intel_wait_for_register(dev_priv,
+				    LCPLL1_CTL, LCPLL_PLL_LOCK, LCPLL_PLL_LOCK,
+				    5))
+		DRM_ERROR("DPLL0 not locked\n");
+
+	dev_priv->cdclk.hw.vco = vco;
+
+	/* We'll want to keep using the current vco from now on. */
+	skl_set_preferred_cdclk_vco(dev_priv, vco);
+}
+
+static void skl_dpll0_disable(struct drm_i915_private *dev_priv)
+{
+	I915_WRITE(LCPLL1_CTL, I915_READ(LCPLL1_CTL) & ~LCPLL_PLL_ENABLE);
+	if (intel_wait_for_register(dev_priv,
+				   LCPLL1_CTL, LCPLL_PLL_LOCK, 0,
+				   1))
+		DRM_ERROR("Couldn't disable DPLL0\n");
+
+	dev_priv->cdclk.hw.vco = 0;
+}
+
+static void skl_set_cdclk(struct drm_i915_private *dev_priv,
+			  const struct intel_cdclk_state *cdclk_state)
+{
+	int cdclk = cdclk_state->cdclk;
+	int vco = cdclk_state->vco;
+	u32 freq_select, pcu_ack;
+	int ret;
+
+	WARN_ON((cdclk == 24000) != (vco == 0));
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+	ret = skl_pcode_request(dev_priv, SKL_PCODE_CDCLK_CONTROL,
+				SKL_CDCLK_PREPARE_FOR_CHANGE,
+				SKL_CDCLK_READY_FOR_CHANGE,
+				SKL_CDCLK_READY_FOR_CHANGE, 3);
+	mutex_unlock(&dev_priv->rps.hw_lock);
+	if (ret) {
+		DRM_ERROR("Failed to inform PCU about cdclk change (%d)\n",
+			  ret);
+		return;
+	}
+
+	/* set CDCLK_CTL */
+	switch (cdclk) {
+	case 450000:
+	case 432000:
+		freq_select = CDCLK_FREQ_450_432;
+		pcu_ack = 1;
+		break;
+	case 540000:
+		freq_select = CDCLK_FREQ_540;
+		pcu_ack = 2;
+		break;
+	case 308571:
+	case 337500:
+	default:
+		freq_select = CDCLK_FREQ_337_308;
+		pcu_ack = 0;
+		break;
+	case 617143:
+	case 675000:
+		freq_select = CDCLK_FREQ_675_617;
+		pcu_ack = 3;
+		break;
+	}
+
+	if (dev_priv->cdclk.hw.vco != 0 &&
+	    dev_priv->cdclk.hw.vco != vco)
+		skl_dpll0_disable(dev_priv);
+
+	if (dev_priv->cdclk.hw.vco != vco)
+		skl_dpll0_enable(dev_priv, vco);
+
+	I915_WRITE(CDCLK_CTL, freq_select | skl_cdclk_decimal(cdclk));
+	POSTING_READ(CDCLK_CTL);
+
+	/* inform PCU of the change */
+	mutex_lock(&dev_priv->rps.hw_lock);
+	sandybridge_pcode_write(dev_priv, SKL_PCODE_CDCLK_CONTROL, pcu_ack);
+	mutex_unlock(&dev_priv->rps.hw_lock);
+
+	intel_update_cdclk(dev_priv);
+}
+
+static void skl_sanitize_cdclk(struct drm_i915_private *dev_priv)
+{
+	uint32_t cdctl, expected;
+
+	/*
+	 * check if the pre-os initialized the display
+	 * There is SWF18 scratchpad register defined which is set by the
+	 * pre-os which can be used by the OS drivers to check the status
+	 */
+	if ((I915_READ(SWF_ILK(0x18)) & 0x00FFFFFF) == 0)
+		goto sanitize;
+
+	intel_update_cdclk(dev_priv);
+	/* Is PLL enabled and locked ? */
+	if (dev_priv->cdclk.hw.vco == 0 ||
+	    dev_priv->cdclk.hw.cdclk == dev_priv->cdclk.hw.ref)
+		goto sanitize;
+
+	/* DPLL okay; verify the cdclock
+	 *
+	 * Noticed in some instances that the freq selection is correct but
+	 * decimal part is programmed wrong from BIOS where pre-os does not
+	 * enable display. Verify the same as well.
+	 */
+	cdctl = I915_READ(CDCLK_CTL);
+	expected = (cdctl & CDCLK_FREQ_SEL_MASK) |
+		skl_cdclk_decimal(dev_priv->cdclk.hw.cdclk);
+	if (cdctl == expected)
+		/* All well; nothing to sanitize */
+		return;
+
+sanitize:
+	DRM_DEBUG_KMS("Sanitizing cdclk programmed by pre-os\n");
+
+	/* force cdclk programming */
+	dev_priv->cdclk.hw.cdclk = 0;
+	/* force full PLL disable + enable */
+	dev_priv->cdclk.hw.vco = -1;
+}
+
+/**
+ * skl_init_cdclk - Initialize CDCLK on SKL
+ * @dev_priv: i915 device
+ *
+ * Initialize CDCLK for SKL and derivatives. This is generally
+ * done only during the display core initialization sequence,
+ * after which the DMC will take care of turning CDCLK off/on
+ * as needed.
+ */
+void skl_init_cdclk(struct drm_i915_private *dev_priv)
+{
+	struct intel_cdclk_state cdclk_state;
+
+	skl_sanitize_cdclk(dev_priv);
+
+	if (dev_priv->cdclk.hw.cdclk != 0 &&
+	    dev_priv->cdclk.hw.vco != 0) {
+		/*
+		 * Use the current vco as our initial
+		 * guess as to what the preferred vco is.
+		 */
+		if (dev_priv->skl_preferred_vco_freq == 0)
+			skl_set_preferred_cdclk_vco(dev_priv,
+						    dev_priv->cdclk.hw.vco);
+		return;
+	}
+
+	cdclk_state = dev_priv->cdclk.hw;
+
+	cdclk_state.vco = dev_priv->skl_preferred_vco_freq;
+	if (cdclk_state.vco == 0)
+		cdclk_state.vco = 8100000;
+	cdclk_state.cdclk = skl_calc_cdclk(0, cdclk_state.vco);
+
+	skl_set_cdclk(dev_priv, &cdclk_state);
+}
+
+/**
+ * skl_uninit_cdclk - Uninitialize CDCLK on SKL
+ * @dev_priv: i915 device
+ *
+ * Uninitialize CDCLK for SKL and derivatives. This is done only
+ * during the display core uninitialization sequence.
+ */
+void skl_uninit_cdclk(struct drm_i915_private *dev_priv)
+{
+	struct intel_cdclk_state cdclk_state = dev_priv->cdclk.hw;
+
+	cdclk_state.cdclk = cdclk_state.ref;
+	cdclk_state.vco = 0;
+
+	skl_set_cdclk(dev_priv, &cdclk_state);
+}
+
+static int bxt_calc_cdclk(int max_pixclk)
+{
+	if (max_pixclk > 576000)
+		return 624000;
+	else if (max_pixclk > 384000)
+		return 576000;
+	else if (max_pixclk > 288000)
+		return 384000;
+	else if (max_pixclk > 144000)
+		return 288000;
+	else
+		return 144000;
+}
+
+static int glk_calc_cdclk(int max_pixclk)
+{
+	if (max_pixclk > 2 * 158400)
+		return 316800;
+	else if (max_pixclk > 2 * 79200)
+		return 158400;
+	else
+		return 79200;
+}
+
+static int bxt_de_pll_vco(struct drm_i915_private *dev_priv, int cdclk)
+{
+	int ratio;
+
+	if (cdclk == dev_priv->cdclk.hw.ref)
+		return 0;
+
+	switch (cdclk) {
+	default:
+		MISSING_CASE(cdclk);
+	case 144000:
+	case 288000:
+	case 384000:
+	case 576000:
+		ratio = 60;
+		break;
+	case 624000:
+		ratio = 65;
+		break;
+	}
+
+	return dev_priv->cdclk.hw.ref * ratio;
+}
+
+static int glk_de_pll_vco(struct drm_i915_private *dev_priv, int cdclk)
+{
+	int ratio;
+
+	if (cdclk == dev_priv->cdclk.hw.ref)
+		return 0;
+
+	switch (cdclk) {
+	default:
+		MISSING_CASE(cdclk);
+	case  79200:
+	case 158400:
+	case 316800:
+		ratio = 33;
+		break;
+	}
+
+	return dev_priv->cdclk.hw.ref * ratio;
+}
+
+static void bxt_de_pll_update(struct drm_i915_private *dev_priv,
+			      struct intel_cdclk_state *cdclk_state)
+{
+	u32 val;
+
+	cdclk_state->ref = 19200;
+	cdclk_state->vco = 0;
+
+	val = I915_READ(BXT_DE_PLL_ENABLE);
+	if ((val & BXT_DE_PLL_PLL_ENABLE) == 0)
+		return;
+
+	if (WARN_ON((val & BXT_DE_PLL_LOCK) == 0))
+		return;
+
+	val = I915_READ(BXT_DE_PLL_CTL);
+	cdclk_state->vco = (val & BXT_DE_PLL_RATIO_MASK) * cdclk_state->ref;
+}
+
+static void bxt_get_cdclk(struct drm_i915_private *dev_priv,
+			  struct intel_cdclk_state *cdclk_state)
+{
+	u32 divider;
+	int div;
+
+	bxt_de_pll_update(dev_priv, cdclk_state);
+
+	cdclk_state->cdclk = cdclk_state->ref;
+
+	if (cdclk_state->vco == 0)
+		return;
+
+	divider = I915_READ(CDCLK_CTL) & BXT_CDCLK_CD2X_DIV_SEL_MASK;
+
+	switch (divider) {
+	case BXT_CDCLK_CD2X_DIV_SEL_1:
+		div = 2;
+		break;
+	case BXT_CDCLK_CD2X_DIV_SEL_1_5:
+		WARN(IS_GEMINILAKE(dev_priv), "Unsupported divider\n");
+		div = 3;
+		break;
+	case BXT_CDCLK_CD2X_DIV_SEL_2:
+		div = 4;
+		break;
+	case BXT_CDCLK_CD2X_DIV_SEL_4:
+		div = 8;
+		break;
+	default:
+		MISSING_CASE(divider);
+		return;
+	}
+
+	cdclk_state->cdclk = DIV_ROUND_CLOSEST(cdclk_state->vco, div);
+}
+
+static void bxt_de_pll_disable(struct drm_i915_private *dev_priv)
+{
+	I915_WRITE(BXT_DE_PLL_ENABLE, 0);
+
+	/* Timeout 200us */
+	if (intel_wait_for_register(dev_priv,
+				    BXT_DE_PLL_ENABLE, BXT_DE_PLL_LOCK, 0,
+				    1))
+		DRM_ERROR("timeout waiting for DE PLL unlock\n");
+
+	dev_priv->cdclk.hw.vco = 0;
+}
+
+static void bxt_de_pll_enable(struct drm_i915_private *dev_priv, int vco)
+{
+	int ratio = DIV_ROUND_CLOSEST(vco, dev_priv->cdclk.hw.ref);
+	u32 val;
+
+	val = I915_READ(BXT_DE_PLL_CTL);
+	val &= ~BXT_DE_PLL_RATIO_MASK;
+	val |= BXT_DE_PLL_RATIO(ratio);
+	I915_WRITE(BXT_DE_PLL_CTL, val);
+
+	I915_WRITE(BXT_DE_PLL_ENABLE, BXT_DE_PLL_PLL_ENABLE);
+
+	/* Timeout 200us */
+	if (intel_wait_for_register(dev_priv,
+				    BXT_DE_PLL_ENABLE,
+				    BXT_DE_PLL_LOCK,
+				    BXT_DE_PLL_LOCK,
+				    1))
+		DRM_ERROR("timeout waiting for DE PLL lock\n");
+
+	dev_priv->cdclk.hw.vco = vco;
+}
+
+static void bxt_set_cdclk(struct drm_i915_private *dev_priv,
+			  const struct intel_cdclk_state *cdclk_state)
+{
+	int cdclk = cdclk_state->cdclk;
+	int vco = cdclk_state->vco;
+	u32 val, divider;
+	int ret;
+
+	/* cdclk = vco / 2 / div{1,1.5,2,4} */
+	switch (DIV_ROUND_CLOSEST(vco, cdclk)) {
+	case 8:
+		divider = BXT_CDCLK_CD2X_DIV_SEL_4;
+		break;
+	case 4:
+		divider = BXT_CDCLK_CD2X_DIV_SEL_2;
+		break;
+	case 3:
+		WARN(IS_GEMINILAKE(dev_priv), "Unsupported divider\n");
+		divider = BXT_CDCLK_CD2X_DIV_SEL_1_5;
+		break;
+	case 2:
+		divider = BXT_CDCLK_CD2X_DIV_SEL_1;
+		break;
+	default:
+		WARN_ON(cdclk != dev_priv->cdclk.hw.ref);
+		WARN_ON(vco != 0);
+
+		divider = BXT_CDCLK_CD2X_DIV_SEL_1;
+		break;
+	}
+
+	/* Inform power controller of upcoming frequency change */
+	mutex_lock(&dev_priv->rps.hw_lock);
+	ret = sandybridge_pcode_write(dev_priv, HSW_PCODE_DE_WRITE_FREQ_REQ,
+				      0x80000000);
+	mutex_unlock(&dev_priv->rps.hw_lock);
+
+	if (ret) {
+		DRM_ERROR("PCode CDCLK freq change notify failed (err %d, freq %d)\n",
+			  ret, cdclk);
+		return;
+	}
+
+	if (dev_priv->cdclk.hw.vco != 0 &&
+	    dev_priv->cdclk.hw.vco != vco)
+		bxt_de_pll_disable(dev_priv);
+
+	if (dev_priv->cdclk.hw.vco != vco)
+		bxt_de_pll_enable(dev_priv, vco);
+
+	val = divider | skl_cdclk_decimal(cdclk);
+	/*
+	 * FIXME if only the cd2x divider needs changing, it could be done
+	 * without shutting off the pipe (if only one pipe is active).
+	 */
+	val |= BXT_CDCLK_CD2X_PIPE_NONE;
+	/*
+	 * Disable SSA Precharge when CD clock frequency < 500 MHz,
+	 * enable otherwise.
+	 */
+	if (cdclk >= 500000)
+		val |= BXT_CDCLK_SSA_PRECHARGE_ENABLE;
+	I915_WRITE(CDCLK_CTL, val);
+
+	mutex_lock(&dev_priv->rps.hw_lock);
+	ret = sandybridge_pcode_write(dev_priv, HSW_PCODE_DE_WRITE_FREQ_REQ,
+				      DIV_ROUND_UP(cdclk, 25000));
+	mutex_unlock(&dev_priv->rps.hw_lock);
+
+	if (ret) {
+		DRM_ERROR("PCode CDCLK freq set failed, (err %d, freq %d)\n",
+			  ret, cdclk);
+		return;
+	}
+
+	intel_update_cdclk(dev_priv);
+}
+
+static void bxt_sanitize_cdclk(struct drm_i915_private *dev_priv)
+{
+	u32 cdctl, expected;
+
+	intel_update_cdclk(dev_priv);
+
+	if (dev_priv->cdclk.hw.vco == 0 ||
+	    dev_priv->cdclk.hw.cdclk == dev_priv->cdclk.hw.ref)
+		goto sanitize;
+
+	/* DPLL okay; verify the cdclock
+	 *
+	 * Some BIOS versions leave an incorrect decimal frequency value and
+	 * set reserved MBZ bits in CDCLK_CTL at least during exiting from S4,
+	 * so sanitize this register.
+	 */
+	cdctl = I915_READ(CDCLK_CTL);
+	/*
+	 * Let's ignore the pipe field, since BIOS could have configured the
+	 * dividers both synching to an active pipe, or asynchronously
+	 * (PIPE_NONE).
+	 */
+	cdctl &= ~BXT_CDCLK_CD2X_PIPE_NONE;
+
+	expected = (cdctl & BXT_CDCLK_CD2X_DIV_SEL_MASK) |
+		skl_cdclk_decimal(dev_priv->cdclk.hw.cdclk);
+	/*
+	 * Disable SSA Precharge when CD clock frequency < 500 MHz,
+	 * enable otherwise.
+	 */
+	if (dev_priv->cdclk.hw.cdclk >= 500000)
+		expected |= BXT_CDCLK_SSA_PRECHARGE_ENABLE;
+
+	if (cdctl == expected)
+		/* All well; nothing to sanitize */
+		return;
+
+sanitize:
+	DRM_DEBUG_KMS("Sanitizing cdclk programmed by pre-os\n");
+
+	/* force cdclk programming */
+	dev_priv->cdclk.hw.cdclk = 0;
+
+	/* force full PLL disable + enable */
+	dev_priv->cdclk.hw.vco = -1;
+}
+
+/**
+ * bxt_init_cdclk - Initialize CDCLK on BXT
+ * @dev_priv: i915 device
+ *
+ * Initialize CDCLK for BXT and derivatives. This is generally
+ * done only during the display core initialization sequence,
+ * after which the DMC will take care of turning CDCLK off/on
+ * as needed.
+ */
+void bxt_init_cdclk(struct drm_i915_private *dev_priv)
+{
+	struct intel_cdclk_state cdclk_state;
+
+	bxt_sanitize_cdclk(dev_priv);
+
+	if (dev_priv->cdclk.hw.cdclk != 0 &&
+	    dev_priv->cdclk.hw.vco != 0)
+		return;
+
+	cdclk_state = dev_priv->cdclk.hw;
+
+	/*
+	 * FIXME:
+	 * - The initial CDCLK needs to be read from VBT.
+	 *   Need to make this change after VBT has changes for BXT.
+	 */
+	if (IS_GEMINILAKE(dev_priv)) {
+		cdclk_state.cdclk = glk_calc_cdclk(0);
+		cdclk_state.vco = glk_de_pll_vco(dev_priv, cdclk_state.cdclk);
+	} else {
+		cdclk_state.cdclk = bxt_calc_cdclk(0);
+		cdclk_state.vco = bxt_de_pll_vco(dev_priv, cdclk_state.cdclk);
+	}
+
+	bxt_set_cdclk(dev_priv, &cdclk_state);
+}
+
+/**
+ * bxt_uninit_cdclk - Uninitialize CDCLK on BXT
+ * @dev_priv: i915 device
+ *
+ * Uninitialize CDCLK for BXT and derivatives. This is done only
+ * during the display core uninitialization sequence.
+ */
+void bxt_uninit_cdclk(struct drm_i915_private *dev_priv)
+{
+	struct intel_cdclk_state cdclk_state = dev_priv->cdclk.hw;
+
+	cdclk_state.cdclk = cdclk_state.ref;
+	cdclk_state.vco = 0;
+
+	bxt_set_cdclk(dev_priv, &cdclk_state);
+}
+
+/**
+ * intel_cdclk_state_compare - Determine if two CDCLK states differ
+ * @a: first CDCLK state
+ * @b: second CDCLK state
+ *
+ * Returns:
+ * True if the CDCLK states are identical, false if they differ.
+ */
+bool intel_cdclk_state_compare(const struct intel_cdclk_state *a,
+			       const struct intel_cdclk_state *b)
+{
+	return memcmp(a, b, sizeof(*a)) == 0;
+}
+
+/**
+ * intel_set_cdclk - Push the CDCLK state to the hardware
+ * @dev_priv: i915 device
+ * @cdclk_state: new CDCLK state
+ *
+ * Program the hardware based on the passed in CDCLK state,
+ * if necessary.
+ */
+void intel_set_cdclk(struct drm_i915_private *dev_priv,
+		     const struct intel_cdclk_state *cdclk_state)
+{
+	if (intel_cdclk_state_compare(&dev_priv->cdclk.hw, cdclk_state))
+		return;
+
+	if (WARN_ON_ONCE(!dev_priv->display.set_cdclk))
+		return;
+
+	DRM_DEBUG_DRIVER("Changing CDCLK to %d kHz, VCO %d kHz, ref %d kHz\n",
+			 cdclk_state->cdclk, cdclk_state->vco,
+			 cdclk_state->ref);
+
+	dev_priv->display.set_cdclk(dev_priv, cdclk_state);
+}
+
+static int bdw_adjust_min_pipe_pixel_rate(struct intel_crtc_state *crtc_state,
+					  int pixel_rate)
+{
+	struct drm_i915_private *dev_priv =
+		to_i915(crtc_state->base.crtc->dev);
+
+	/* pixel rate mustn't exceed 95% of cdclk with IPS on BDW */
+	if (IS_BROADWELL(dev_priv) && crtc_state->ips_enabled)
+		pixel_rate = DIV_ROUND_UP(pixel_rate * 100, 95);
+
+	/* BSpec says "Do not use DisplayPort with CDCLK less than
+	 * 432 MHz, audio enabled, port width x4, and link rate
+	 * HBR2 (5.4 GHz), or else there may be audio corruption or
+	 * screen corruption."
+	 */
+	if (intel_crtc_has_dp_encoder(crtc_state) &&
+	    crtc_state->has_audio &&
+	    crtc_state->port_clock >= 540000 &&
+	    crtc_state->lane_count == 4)
+		pixel_rate = max(432000, pixel_rate);
+
+	return pixel_rate;
+}
+
+/* compute the max rate for new configuration */
+static int intel_max_pixel_rate(struct drm_atomic_state *state)
+{
+	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
+	struct drm_i915_private *dev_priv = to_i915(state->dev);
+	struct drm_crtc *crtc;
+	struct drm_crtc_state *cstate;
+	struct intel_crtc_state *crtc_state;
+	unsigned int max_pixel_rate = 0, i;
+	enum pipe pipe;
+
+	memcpy(intel_state->min_pixclk, dev_priv->min_pixclk,
+	       sizeof(intel_state->min_pixclk));
+
+	for_each_crtc_in_state(state, crtc, cstate, i) {
+		int pixel_rate;
+
+		crtc_state = to_intel_crtc_state(cstate);
+		if (!crtc_state->base.enable) {
+			intel_state->min_pixclk[i] = 0;
+			continue;
+		}
+
+		pixel_rate = crtc_state->pixel_rate;
+
+		if (IS_BROADWELL(dev_priv) || IS_GEN9(dev_priv))
+			pixel_rate =
+				bdw_adjust_min_pipe_pixel_rate(crtc_state,
+							       pixel_rate);
+
+		intel_state->min_pixclk[i] = pixel_rate;
+	}
+
+	for_each_pipe(dev_priv, pipe)
+		max_pixel_rate = max(intel_state->min_pixclk[pipe],
+				     max_pixel_rate);
+
+	return max_pixel_rate;
+}
+
+static int vlv_modeset_calc_cdclk(struct drm_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->dev);
+	int max_pixclk = intel_max_pixel_rate(state);
+	struct intel_atomic_state *intel_state =
+		to_intel_atomic_state(state);
+	int cdclk;
+
+	cdclk = vlv_calc_cdclk(dev_priv, max_pixclk);
+
+	if (cdclk > dev_priv->max_cdclk_freq) {
+		DRM_DEBUG_KMS("requested cdclk (%d kHz) exceeds max (%d kHz)\n",
+			      cdclk, dev_priv->max_cdclk_freq);
+		return -EINVAL;
+	}
+
+	intel_state->cdclk.logical.cdclk = cdclk;
+
+	if (!intel_state->active_crtcs) {
+		cdclk = vlv_calc_cdclk(dev_priv, 0);
+
+		intel_state->cdclk.actual.cdclk = cdclk;
+	} else {
+		intel_state->cdclk.actual =
+			intel_state->cdclk.logical;
+	}
+
+	return 0;
+}
+
+static int bdw_modeset_calc_cdclk(struct drm_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->dev);
+	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
+	int max_pixclk = intel_max_pixel_rate(state);
+	int cdclk;
+
+	/*
+	 * FIXME should also account for plane ratio
+	 * once 64bpp pixel formats are supported.
+	 */
+	cdclk = bdw_calc_cdclk(max_pixclk);
+
+	if (cdclk > dev_priv->max_cdclk_freq) {
+		DRM_DEBUG_KMS("requested cdclk (%d kHz) exceeds max (%d kHz)\n",
+			      cdclk, dev_priv->max_cdclk_freq);
+		return -EINVAL;
+	}
+
+	intel_state->cdclk.logical.cdclk = cdclk;
+
+	if (!intel_state->active_crtcs) {
+		cdclk = bdw_calc_cdclk(0);
+
+		intel_state->cdclk.actual.cdclk = cdclk;
+	} else {
+		intel_state->cdclk.actual =
+			intel_state->cdclk.logical;
+	}
+
+	return 0;
+}
+
+static int skl_modeset_calc_cdclk(struct drm_atomic_state *state)
+{
+	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
+	struct drm_i915_private *dev_priv = to_i915(state->dev);
+	const int max_pixclk = intel_max_pixel_rate(state);
+	int cdclk, vco;
+
+	vco = intel_state->cdclk.logical.vco;
+	if (!vco)
+		vco = dev_priv->skl_preferred_vco_freq;
+
+	/*
+	 * FIXME should also account for plane ratio
+	 * once 64bpp pixel formats are supported.
+	 */
+	cdclk = skl_calc_cdclk(max_pixclk, vco);
+
+	if (cdclk > dev_priv->max_cdclk_freq) {
+		DRM_DEBUG_KMS("requested cdclk (%d kHz) exceeds max (%d kHz)\n",
+			      cdclk, dev_priv->max_cdclk_freq);
+		return -EINVAL;
+	}
+
+	intel_state->cdclk.logical.vco = vco;
+	intel_state->cdclk.logical.cdclk = cdclk;
+
+	if (!intel_state->active_crtcs) {
+		cdclk = skl_calc_cdclk(0, vco);
+
+		intel_state->cdclk.actual.vco = vco;
+		intel_state->cdclk.actual.cdclk = cdclk;
+	} else {
+		intel_state->cdclk.actual =
+			intel_state->cdclk.logical;
+	}
+
+	return 0;
+}
+
+static int bxt_modeset_calc_cdclk(struct drm_atomic_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->dev);
+	int max_pixclk = intel_max_pixel_rate(state);
+	struct intel_atomic_state *intel_state =
+		to_intel_atomic_state(state);
+	int cdclk, vco;
+
+	if (IS_GEMINILAKE(dev_priv)) {
+		cdclk = glk_calc_cdclk(max_pixclk);
+		vco = glk_de_pll_vco(dev_priv, cdclk);
+	} else {
+		cdclk = bxt_calc_cdclk(max_pixclk);
+		vco = bxt_de_pll_vco(dev_priv, cdclk);
+	}
+
+	if (cdclk > dev_priv->max_cdclk_freq) {
+		DRM_DEBUG_KMS("requested cdclk (%d kHz) exceeds max (%d kHz)\n",
+			      cdclk, dev_priv->max_cdclk_freq);
+		return -EINVAL;
+	}
+
+	intel_state->cdclk.logical.vco = vco;
+	intel_state->cdclk.logical.cdclk = cdclk;
+
+	if (!intel_state->active_crtcs) {
+		if (IS_GEMINILAKE(dev_priv)) {
+			cdclk = glk_calc_cdclk(0);
+			vco = glk_de_pll_vco(dev_priv, cdclk);
+		} else {
+			cdclk = bxt_calc_cdclk(0);
+			vco = bxt_de_pll_vco(dev_priv, cdclk);
+		}
+
+		intel_state->cdclk.actual.vco = vco;
+		intel_state->cdclk.actual.cdclk = cdclk;
+	} else {
+		intel_state->cdclk.actual =
+			intel_state->cdclk.logical;
+	}
+
+	return 0;
+}
+
+static int intel_compute_max_dotclk(struct drm_i915_private *dev_priv)
+{
+	int max_cdclk_freq = dev_priv->max_cdclk_freq;
+
+	if (IS_GEMINILAKE(dev_priv))
+		return 2 * max_cdclk_freq;
+	else if (INTEL_INFO(dev_priv)->gen >= 9 ||
+		 IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
+		return max_cdclk_freq;
+	else if (IS_CHERRYVIEW(dev_priv))
+		return max_cdclk_freq*95/100;
+	else if (INTEL_INFO(dev_priv)->gen < 4)
+		return 2*max_cdclk_freq*90/100;
+	else
+		return max_cdclk_freq*90/100;
+}
+
+/**
+ * intel_update_max_cdclk - Determine the maximum support CDCLK frequency
+ * @dev_priv: i915 device
+ *
+ * Determine the maximum CDCLK frequency the platform supports, and also
+ * derive the maximum dot clock frequency the maximum CDCLK frequency
+ * allows.
+ */
+void intel_update_max_cdclk(struct drm_i915_private *dev_priv)
+{
+	if (IS_GEN9_BC(dev_priv)) {
+		u32 limit = I915_READ(SKL_DFSM) & SKL_DFSM_CDCLK_LIMIT_MASK;
+		int max_cdclk, vco;
+
+		vco = dev_priv->skl_preferred_vco_freq;
+		WARN_ON(vco != 8100000 && vco != 8640000);
+
+		/*
+		 * Use the lower (vco 8640) cdclk values as a
+		 * first guess. skl_calc_cdclk() will correct it
+		 * if the preferred vco is 8100 instead.
+		 */
+		if (limit == SKL_DFSM_CDCLK_LIMIT_675)
+			max_cdclk = 617143;
+		else if (limit == SKL_DFSM_CDCLK_LIMIT_540)
+			max_cdclk = 540000;
+		else if (limit == SKL_DFSM_CDCLK_LIMIT_450)
+			max_cdclk = 432000;
+		else
+			max_cdclk = 308571;
+
+		dev_priv->max_cdclk_freq = skl_calc_cdclk(max_cdclk, vco);
+	} else if (IS_GEMINILAKE(dev_priv)) {
+		dev_priv->max_cdclk_freq = 316800;
+	} else if (IS_BROXTON(dev_priv)) {
+		dev_priv->max_cdclk_freq = 624000;
+	} else if (IS_BROADWELL(dev_priv))  {
+		/*
+		 * FIXME with extra cooling we can allow
+		 * 540 MHz for ULX and 675 Mhz for ULT.
+		 * How can we know if extra cooling is
+		 * available? PCI ID, VTB, something else?
+		 */
+		if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
+			dev_priv->max_cdclk_freq = 450000;
+		else if (IS_BDW_ULX(dev_priv))
+			dev_priv->max_cdclk_freq = 450000;
+		else if (IS_BDW_ULT(dev_priv))
+			dev_priv->max_cdclk_freq = 540000;
+		else
+			dev_priv->max_cdclk_freq = 675000;
+	} else if (IS_CHERRYVIEW(dev_priv)) {
+		dev_priv->max_cdclk_freq = 320000;
+	} else if (IS_VALLEYVIEW(dev_priv)) {
+		dev_priv->max_cdclk_freq = 400000;
+	} else {
+		/* otherwise assume cdclk is fixed */
+		dev_priv->max_cdclk_freq = dev_priv->cdclk.hw.cdclk;
+	}
+
+	dev_priv->max_dotclk_freq = intel_compute_max_dotclk(dev_priv);
+
+	DRM_DEBUG_DRIVER("Max CD clock rate: %d kHz\n",
+			 dev_priv->max_cdclk_freq);
+
+	DRM_DEBUG_DRIVER("Max dotclock rate: %d kHz\n",
+			 dev_priv->max_dotclk_freq);
+}
+
+/**
+ * intel_update_cdclk - Determine the current CDCLK frequency
+ * @dev_priv: i915 device
+ *
+ * Determine the current CDCLK frequency.
+ */
+void intel_update_cdclk(struct drm_i915_private *dev_priv)
+{
+	dev_priv->display.get_cdclk(dev_priv, &dev_priv->cdclk.hw);
+
+	DRM_DEBUG_DRIVER("Current CD clock rate: %d kHz, VCO: %d kHz, ref: %d kHz\n",
+			 dev_priv->cdclk.hw.cdclk, dev_priv->cdclk.hw.vco,
+			 dev_priv->cdclk.hw.ref);
+
+	/*
+	 * 9:0 CMBUS [sic] CDCLK frequency (cdfreq):
+	 * Programmng [sic] note: bit[9:2] should be programmed to the number
+	 * of cdclk that generates 4MHz reference clock freq which is used to
+	 * generate GMBus clock. This will vary with the cdclk freq.
+	 */
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		I915_WRITE(GMBUSFREQ_VLV,
+			   DIV_ROUND_UP(dev_priv->cdclk.hw.cdclk, 1000));
+}
+
+static int pch_rawclk(struct drm_i915_private *dev_priv)
+{
+	return (I915_READ(PCH_RAWCLK_FREQ) & RAWCLK_FREQ_MASK) * 1000;
+}
+
+static int vlv_hrawclk(struct drm_i915_private *dev_priv)
+{
+	/* RAWCLK_FREQ_VLV register updated from power well code */
+	return vlv_get_cck_clock_hpll(dev_priv, "hrawclk",
+				      CCK_DISPLAY_REF_CLOCK_CONTROL);
+}
+
+static int g4x_hrawclk(struct drm_i915_private *dev_priv)
+{
+	uint32_t clkcfg;
+
+	/* hrawclock is 1/4 the FSB frequency */
+	clkcfg = I915_READ(CLKCFG);
+	switch (clkcfg & CLKCFG_FSB_MASK) {
+	case CLKCFG_FSB_400:
+		return 100000;
+	case CLKCFG_FSB_533:
+		return 133333;
+	case CLKCFG_FSB_667:
+		return 166667;
+	case CLKCFG_FSB_800:
+		return 200000;
+	case CLKCFG_FSB_1067:
+		return 266667;
+	case CLKCFG_FSB_1333:
+		return 333333;
+	/* these two are just a guess; one of them might be right */
+	case CLKCFG_FSB_1600:
+	case CLKCFG_FSB_1600_ALT:
+		return 400000;
+	default:
+		return 133333;
+	}
+}
+
+/**
+ * intel_update_rawclk - Determine the current RAWCLK frequency
+ * @dev_priv: i915 device
+ *
+ * Determine the current RAWCLK frequency. RAWCLK is a fixed
+ * frequency clock so this needs to done only once.
+ */
+void intel_update_rawclk(struct drm_i915_private *dev_priv)
+{
+	if (HAS_PCH_SPLIT(dev_priv))
+		dev_priv->rawclk_freq = pch_rawclk(dev_priv);
+	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		dev_priv->rawclk_freq = vlv_hrawclk(dev_priv);
+	else if (IS_G4X(dev_priv) || IS_PINEVIEW(dev_priv))
+		dev_priv->rawclk_freq = g4x_hrawclk(dev_priv);
+	else
+		/* no rawclk on other platforms, or no need to know it */
+		return;
+
+	DRM_DEBUG_DRIVER("rawclk rate: %d kHz\n", dev_priv->rawclk_freq);
+}
+
+/**
+ * intel_init_cdclk_hooks - Initialize CDCLK related modesetting hooks
+ * @dev_priv: i915 device
+ */
+void intel_init_cdclk_hooks(struct drm_i915_private *dev_priv)
+{
+	if (IS_CHERRYVIEW(dev_priv)) {
+		dev_priv->display.set_cdclk = chv_set_cdclk;
+		dev_priv->display.modeset_calc_cdclk =
+			vlv_modeset_calc_cdclk;
+	} else if (IS_VALLEYVIEW(dev_priv)) {
+		dev_priv->display.set_cdclk = vlv_set_cdclk;
+		dev_priv->display.modeset_calc_cdclk =
+			vlv_modeset_calc_cdclk;
+	} else if (IS_BROADWELL(dev_priv)) {
+		dev_priv->display.set_cdclk = bdw_set_cdclk;
+		dev_priv->display.modeset_calc_cdclk =
+			bdw_modeset_calc_cdclk;
+	} else if (IS_GEN9_LP(dev_priv)) {
+		dev_priv->display.set_cdclk = bxt_set_cdclk;
+		dev_priv->display.modeset_calc_cdclk =
+			bxt_modeset_calc_cdclk;
+	} else if (IS_GEN9_BC(dev_priv)) {
+		dev_priv->display.set_cdclk = skl_set_cdclk;
+		dev_priv->display.modeset_calc_cdclk =
+			skl_modeset_calc_cdclk;
+	}
+
+	if (IS_GEN9_BC(dev_priv))
+		dev_priv->display.get_cdclk = skl_get_cdclk;
+	else if (IS_GEN9_LP(dev_priv))
+		dev_priv->display.get_cdclk = bxt_get_cdclk;
+	else if (IS_BROADWELL(dev_priv))
+		dev_priv->display.get_cdclk = bdw_get_cdclk;
+	else if (IS_HASWELL(dev_priv))
+		dev_priv->display.get_cdclk = hsw_get_cdclk;
+	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		dev_priv->display.get_cdclk = vlv_get_cdclk;
+	else if (IS_GEN6(dev_priv) || IS_IVYBRIDGE(dev_priv))
+		dev_priv->display.get_cdclk = fixed_400mhz_get_cdclk;
+	else if (IS_GEN5(dev_priv))
+		dev_priv->display.get_cdclk = fixed_450mhz_get_cdclk;
+	else if (IS_GM45(dev_priv))
+		dev_priv->display.get_cdclk = gm45_get_cdclk;
+	else if (IS_G4X(dev_priv))
+		dev_priv->display.get_cdclk = g33_get_cdclk;
+	else if (IS_I965GM(dev_priv))
+		dev_priv->display.get_cdclk = i965gm_get_cdclk;
+	else if (IS_I965G(dev_priv))
+		dev_priv->display.get_cdclk = fixed_400mhz_get_cdclk;
+	else if (IS_PINEVIEW(dev_priv))
+		dev_priv->display.get_cdclk = pnv_get_cdclk;
+	else if (IS_G33(dev_priv))
+		dev_priv->display.get_cdclk = g33_get_cdclk;
+	else if (IS_I945GM(dev_priv))
+		dev_priv->display.get_cdclk = i945gm_get_cdclk;
+	else if (IS_I945G(dev_priv))
+		dev_priv->display.get_cdclk = fixed_400mhz_get_cdclk;
+	else if (IS_I915GM(dev_priv))
+		dev_priv->display.get_cdclk = i915gm_get_cdclk;
+	else if (IS_I915G(dev_priv))
+		dev_priv->display.get_cdclk = fixed_333mhz_get_cdclk;
+	else if (IS_I865G(dev_priv))
+		dev_priv->display.get_cdclk = fixed_266mhz_get_cdclk;
+	else if (IS_I85X(dev_priv))
+		dev_priv->display.get_cdclk = i85x_get_cdclk;
+	else if (IS_I845G(dev_priv))
+		dev_priv->display.get_cdclk = fixed_200mhz_get_cdclk;
+	else { /* 830 */
+		WARN(!IS_I830(dev_priv),
+		     "Unknown platform. Assuming 133 MHz CDCLK\n");
+		dev_priv->display.get_cdclk = fixed_133mhz_get_cdclk;
+	}
+}
diff --git a/drivers/gpu/drm/i915/intel_color.c b/drivers/gpu/drm/i915/intel_color.c
index d81232b79f00..b9e5266d933b 100644
--- a/drivers/gpu/drm/i915/intel_color.c
+++ b/drivers/gpu/drm/i915/intel_color.c
@@ -340,20 +340,12 @@ static void haswell_load_luts(struct drm_crtc_state *crtc_state)
 		hsw_enable_ips(intel_crtc);
 }
 
-/* Loads the palette/gamma unit for the CRTC on Broadwell+. */
-static void broadwell_load_luts(struct drm_crtc_state *state)
+static void bdw_load_degamma_lut(struct drm_crtc_state *state)
 {
-	struct drm_crtc *crtc = state->crtc;
-	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
-	struct intel_crtc_state *intel_state = to_intel_crtc_state(state);
-	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+	struct drm_i915_private *dev_priv = to_i915(state->crtc->dev);
+	enum pipe pipe = to_intel_crtc(state->crtc)->pipe;
 	uint32_t i, lut_size = INTEL_INFO(dev_priv)->color.degamma_lut_size;
 
-	if (crtc_state_is_legacy(state)) {
-		haswell_load_luts(state);
-		return;
-	}
-
 	I915_WRITE(PREC_PAL_INDEX(pipe),
 		   PAL_PREC_SPLIT_MODE | PAL_PREC_AUTO_INCREMENT);
 
@@ -377,6 +369,20 @@ static void broadwell_load_luts(struct drm_crtc_state *state)
 				   (v << 20) | (v << 10) | v);
 		}
 	}
+}
+
+static void bdw_load_gamma_lut(struct drm_crtc_state *state, u32 offset)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->crtc->dev);
+	enum pipe pipe = to_intel_crtc(state->crtc)->pipe;
+	uint32_t i, lut_size = INTEL_INFO(dev_priv)->color.gamma_lut_size;
+
+	WARN_ON(offset & ~PAL_PREC_INDEX_VALUE_MASK);
+
+	I915_WRITE(PREC_PAL_INDEX(pipe),
+		   (offset ? PAL_PREC_SPLIT_MODE : 0) |
+		   PAL_PREC_AUTO_INCREMENT |
+		   offset);
 
 	if (state->gamma_lut) {
 		struct drm_color_lut *lut =
@@ -410,6 +416,23 @@ static void broadwell_load_luts(struct drm_crtc_state *state)
 		I915_WRITE(PREC_PAL_GC_MAX(pipe, 1), (1 << 16) - 1);
 		I915_WRITE(PREC_PAL_GC_MAX(pipe, 2), (1 << 16) - 1);
 	}
+}
+
+/* Loads the palette/gamma unit for the CRTC on Broadwell+. */
+static void broadwell_load_luts(struct drm_crtc_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->crtc->dev);
+	struct intel_crtc_state *intel_state = to_intel_crtc_state(state);
+	enum pipe pipe = to_intel_crtc(state->crtc)->pipe;
+
+	if (crtc_state_is_legacy(state)) {
+		haswell_load_luts(state);
+		return;
+	}
+
+	bdw_load_degamma_lut(state);
+	bdw_load_gamma_lut(state,
+			   INTEL_INFO(dev_priv)->color.degamma_lut_size);
 
 	intel_state->gamma_mode = GAMMA_MODE_MODE_SPLIT;
 	I915_WRITE(GAMMA_MODE(pipe), GAMMA_MODE_MODE_SPLIT);
@@ -422,6 +445,58 @@ static void broadwell_load_luts(struct drm_crtc_state *state)
 	I915_WRITE(PREC_PAL_INDEX(pipe), 0);
 }
 
+static void glk_load_degamma_lut(struct drm_crtc_state *state)
+{
+	struct drm_i915_private *dev_priv = to_i915(state->crtc->dev);
+	enum pipe pipe = to_intel_crtc(state->crtc)->pipe;
+	const uint32_t lut_size = 33;
+	uint32_t i;
+
+	/*
+	 * When setting the auto-increment bit, the hardware seems to
+	 * ignore the index bits, so we need to reset it to index 0
+	 * separately.
+	 */
+	I915_WRITE(PRE_CSC_GAMC_INDEX(pipe), 0);
+	I915_WRITE(PRE_CSC_GAMC_INDEX(pipe), PRE_CSC_GAMC_AUTO_INCREMENT);
+
+	/*
+	 *  FIXME: The pipe degamma table in geminilake doesn't support
+	 *  different values per channel, so this just loads a linear table.
+	 */
+	for (i = 0; i < lut_size; i++) {
+		uint32_t v = (i * ((1 << 16) - 1)) / (lut_size - 1);
+
+		I915_WRITE(PRE_CSC_GAMC_DATA(pipe), v);
+	}
+
+	/* Clamp values > 1.0. */
+	while (i++ < 35)
+		I915_WRITE(PRE_CSC_GAMC_DATA(pipe), (1 << 16) - 1);
+}
+
+static void glk_load_luts(struct drm_crtc_state *state)
+{
+	struct drm_crtc *crtc = state->crtc;
+	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_crtc_state *intel_state = to_intel_crtc_state(state);
+	enum pipe pipe = to_intel_crtc(crtc)->pipe;
+
+	glk_load_degamma_lut(state);
+
+	if (crtc_state_is_legacy(state)) {
+		haswell_load_luts(state);
+		return;
+	}
+
+	bdw_load_gamma_lut(state, 0);
+
+	intel_state->gamma_mode = GAMMA_MODE_MODE_10BIT;
+	I915_WRITE(GAMMA_MODE(pipe), GAMMA_MODE_MODE_10BIT);
+	POSTING_READ(GAMMA_MODE(pipe));
+}
+
 /* Loads the palette/gamma unit for the CRTC on CherryView. */
 static void cherryview_load_luts(struct drm_crtc_state *state)
 {
@@ -536,10 +611,13 @@ void intel_color_init(struct drm_crtc *crtc)
 	} else if (IS_HASWELL(dev_priv)) {
 		dev_priv->display.load_csc_matrix = i9xx_load_csc_matrix;
 		dev_priv->display.load_luts = haswell_load_luts;
-	} else if (IS_BROADWELL(dev_priv) || IS_SKYLAKE(dev_priv) ||
-		   IS_BROXTON(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	} else if (IS_BROADWELL(dev_priv) || IS_GEN9_BC(dev_priv) ||
+		   IS_BROXTON(dev_priv)) {
 		dev_priv->display.load_csc_matrix = i9xx_load_csc_matrix;
 		dev_priv->display.load_luts = broadwell_load_luts;
+	} else if (IS_GEMINILAKE(dev_priv)) {
+		dev_priv->display.load_csc_matrix = i9xx_load_csc_matrix;
+		dev_priv->display.load_luts = glk_load_luts;
 	} else {
 		dev_priv->display.load_luts = i9xx_load_luts;
 	}
diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index 2bf5aca6e37c..8c82607294c6 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -69,12 +69,11 @@ static bool intel_crt_get_hw_state(struct intel_encoder *encoder,
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crt *crt = intel_encoder_to_crt(encoder);
-	enum intel_display_power_domain power_domain;
 	u32 tmp;
 	bool ret;
 
-	power_domain = intel_display_port_power_domain(encoder);
-	if (!intel_display_power_get_if_enabled(dev_priv, power_domain))
+	if (!intel_display_power_get_if_enabled(dev_priv,
+						encoder->power_domain))
 		return false;
 
 	ret = false;
@@ -91,7 +90,7 @@ static bool intel_crt_get_hw_state(struct intel_encoder *encoder,
 
 	ret = true;
 out:
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, encoder->power_domain);
 
 	return ret;
 }
@@ -676,7 +675,6 @@ intel_crt_detect(struct drm_connector *connector, bool force)
 	struct drm_i915_private *dev_priv = to_i915(connector->dev);
 	struct intel_crt *crt = intel_attached_crt(connector);
 	struct intel_encoder *intel_encoder = &crt->base;
-	enum intel_display_power_domain power_domain;
 	enum drm_connector_status status;
 	struct intel_load_detect_pipe tmp;
 	struct drm_modeset_acquire_ctx ctx;
@@ -689,8 +687,7 @@ intel_crt_detect(struct drm_connector *connector, bool force)
 	if (dmi_check_system(intel_spurious_crt_detect))
 		return connector_status_disconnected;
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
+	intel_display_power_get(dev_priv, intel_encoder->power_domain);
 
 	if (I915_HAS_HOTPLUG(dev_priv)) {
 		/* We can not rely on the HPD pin always being correctly wired
@@ -745,7 +742,7 @@ intel_crt_detect(struct drm_connector *connector, bool force)
 	drm_modeset_acquire_fini(&ctx);
 
 out:
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, intel_encoder->power_domain);
 	return status;
 }
 
@@ -761,12 +758,10 @@ static int intel_crt_get_modes(struct drm_connector *connector)
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crt *crt = intel_attached_crt(connector);
 	struct intel_encoder *intel_encoder = &crt->base;
-	enum intel_display_power_domain power_domain;
 	int ret;
 	struct i2c_adapter *i2c;
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
+	intel_display_power_get(dev_priv, intel_encoder->power_domain);
 
 	i2c = intel_gmbus_get_adapter(dev_priv, dev_priv->vbt.crt_ddc_pin);
 	ret = intel_crt_ddc_get_modes(connector, i2c);
@@ -778,7 +773,7 @@ static int intel_crt_get_modes(struct drm_connector *connector)
 	ret = intel_crt_ddc_get_modes(connector, i2c);
 
 out:
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, intel_encoder->power_domain);
 
 	return ret;
 }
@@ -904,6 +899,8 @@ void intel_crt_init(struct drm_i915_private *dev_priv)
 
 	crt->adpa_reg = adpa_reg;
 
+	crt->base.power_domain = POWER_DOMAIN_PORT_CRT;
+
 	crt->base.compute_config = intel_crt_compute_config;
 	if (HAS_PCH_SPLIT(dev_priv)) {
 		crt->base.disable = pch_disable_crt;
diff --git a/drivers/gpu/drm/i915/intel_csr.c b/drivers/gpu/drm/i915/intel_csr.c
index 0085bc745f6a..14659c7e2858 100644
--- a/drivers/gpu/drm/i915/intel_csr.c
+++ b/drivers/gpu/drm/i915/intel_csr.c
@@ -34,9 +34,9 @@
  * low-power state and comes back to normal.
  */
 
-#define I915_CSR_GLK "i915/glk_dmc_ver1_01.bin"
+#define I915_CSR_GLK "i915/glk_dmc_ver1_03.bin"
 MODULE_FIRMWARE(I915_CSR_GLK);
-#define GLK_CSR_VERSION_REQUIRED	CSR_VERSION(1, 1)
+#define GLK_CSR_VERSION_REQUIRED	CSR_VERSION(1, 3)
 
 #define I915_CSR_KBL "i915/kbl_dmc_ver1_01.bin"
 MODULE_FIRMWARE(I915_CSR_KBL);
@@ -396,13 +396,11 @@ static void csr_load_work_fn(struct work_struct *work)
 	struct drm_i915_private *dev_priv;
 	struct intel_csr *csr;
 	const struct firmware *fw = NULL;
-	int ret;
 
 	dev_priv = container_of(work, typeof(*dev_priv), csr.work);
 	csr = &dev_priv->csr;
 
-	ret = request_firmware(&fw, dev_priv->csr.fw_path,
-			       &dev_priv->drm.pdev->dev);
+	request_firmware(&fw, dev_priv->csr.fw_path, &dev_priv->drm.pdev->dev);
 	if (fw)
 		dev_priv->csr.dmc_payload = parse_csr_fw(dev_priv, fw);
 
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c
index 66b367d0771a..04676760e6fd 100644
--- a/drivers/gpu/drm/i915/intel_ddi.c
+++ b/drivers/gpu/drm/i915/intel_ddi.c
@@ -34,6 +34,19 @@ struct ddi_buf_trans {
 	u8 i_boost;	/* SKL: I_boost; valid: 0x0, 0x1, 0x3, 0x7 */
 };
 
+static const u8 index_to_dp_signal_levels[] = {
+	[0] = DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_0,
+	[1] = DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_1,
+	[2] = DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_2,
+	[3] = DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_3,
+	[4] = DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_0,
+	[5] = DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_1,
+	[6] = DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_2,
+	[7] = DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_0,
+	[8] = DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_1,
+	[9] = DP_TRAIN_VOLTAGE_SWING_LEVEL_3 | DP_TRAIN_PRE_EMPH_LEVEL_0,
+};
+
 /* HDMI/DVI modes ignore everything but the last 2 items. So we share
  * them for both DP and FDI transports, allowing those ports to
  * automatically adapt to HDMI connections as well
@@ -445,7 +458,7 @@ static int intel_ddi_hdmi_level(struct drm_i915_private *dev_priv, enum port por
 	if (IS_GEN9_LP(dev_priv))
 		return hdmi_level;
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	if (IS_GEN9_BC(dev_priv)) {
 		skl_get_buf_trans_hdmi(dev_priv, &n_hdmi_entries);
 		hdmi_default_entry = 8;
 	} else if (IS_BROADWELL(dev_priv)) {
@@ -468,6 +481,59 @@ static int intel_ddi_hdmi_level(struct drm_i915_private *dev_priv, enum port por
 	return hdmi_level;
 }
 
+static const struct ddi_buf_trans *
+intel_ddi_get_buf_trans_dp(struct drm_i915_private *dev_priv,
+			   int *n_entries)
+{
+	if (IS_KABYLAKE(dev_priv)) {
+		return kbl_get_buf_trans_dp(dev_priv, n_entries);
+	} else if (IS_SKYLAKE(dev_priv)) {
+		return skl_get_buf_trans_dp(dev_priv, n_entries);
+	} else if (IS_BROADWELL(dev_priv)) {
+		*n_entries = ARRAY_SIZE(bdw_ddi_translations_dp);
+		return  bdw_ddi_translations_dp;
+	} else if (IS_HASWELL(dev_priv)) {
+		*n_entries = ARRAY_SIZE(hsw_ddi_translations_dp);
+		return hsw_ddi_translations_dp;
+	}
+
+	*n_entries = 0;
+	return NULL;
+}
+
+static const struct ddi_buf_trans *
+intel_ddi_get_buf_trans_edp(struct drm_i915_private *dev_priv,
+			    int *n_entries)
+{
+	if (IS_KABYLAKE(dev_priv) || IS_SKYLAKE(dev_priv)) {
+		return skl_get_buf_trans_edp(dev_priv, n_entries);
+	} else if (IS_BROADWELL(dev_priv)) {
+		return bdw_get_buf_trans_edp(dev_priv, n_entries);
+	} else if (IS_HASWELL(dev_priv)) {
+		*n_entries = ARRAY_SIZE(hsw_ddi_translations_dp);
+		return hsw_ddi_translations_dp;
+	}
+
+	*n_entries = 0;
+	return NULL;
+}
+
+static const struct ddi_buf_trans *
+intel_ddi_get_buf_trans_fdi(struct drm_i915_private *dev_priv,
+			    int *n_entries)
+{
+	if (IS_BROADWELL(dev_priv)) {
+		*n_entries = ARRAY_SIZE(hsw_ddi_translations_fdi);
+		return hsw_ddi_translations_fdi;
+	} else if (IS_HASWELL(dev_priv)) {
+		*n_entries = ARRAY_SIZE(hsw_ddi_translations_fdi);
+		return hsw_ddi_translations_fdi;
+	}
+
+	*n_entries = 0;
+	return NULL;
+}
+
 /*
  * Starting with Haswell, DDI port buffers must be programmed with correct
  * values in advance. This function programs the correct values for
@@ -477,76 +543,43 @@ void intel_prepare_dp_ddi_buffers(struct intel_encoder *encoder)
 {
 	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	u32 iboost_bit = 0;
-	int i, n_dp_entries, n_edp_entries, size;
+	int i, n_entries;
 	enum port port = intel_ddi_get_encoder_port(encoder);
-	const struct ddi_buf_trans *ddi_translations_fdi;
-	const struct ddi_buf_trans *ddi_translations_dp;
-	const struct ddi_buf_trans *ddi_translations_edp;
 	const struct ddi_buf_trans *ddi_translations;
 
 	if (IS_GEN9_LP(dev_priv))
 		return;
 
-	if (IS_KABYLAKE(dev_priv)) {
-		ddi_translations_fdi = NULL;
-		ddi_translations_dp =
-				kbl_get_buf_trans_dp(dev_priv, &n_dp_entries);
-		ddi_translations_edp =
-				skl_get_buf_trans_edp(dev_priv, &n_edp_entries);
-	} else if (IS_SKYLAKE(dev_priv)) {
-		ddi_translations_fdi = NULL;
-		ddi_translations_dp =
-				skl_get_buf_trans_dp(dev_priv, &n_dp_entries);
-		ddi_translations_edp =
-				skl_get_buf_trans_edp(dev_priv, &n_edp_entries);
-	} else if (IS_BROADWELL(dev_priv)) {
-		ddi_translations_fdi = bdw_ddi_translations_fdi;
-		ddi_translations_dp = bdw_ddi_translations_dp;
-		ddi_translations_edp = bdw_get_buf_trans_edp(dev_priv, &n_edp_entries);
-		n_dp_entries = ARRAY_SIZE(bdw_ddi_translations_dp);
-	} else if (IS_HASWELL(dev_priv)) {
-		ddi_translations_fdi = hsw_ddi_translations_fdi;
-		ddi_translations_dp = hsw_ddi_translations_dp;
-		ddi_translations_edp = hsw_ddi_translations_dp;
-		n_dp_entries = n_edp_entries = ARRAY_SIZE(hsw_ddi_translations_dp);
-	} else {
-		WARN(1, "ddi translation table missing\n");
-		ddi_translations_edp = bdw_ddi_translations_dp;
-		ddi_translations_fdi = bdw_ddi_translations_fdi;
-		ddi_translations_dp = bdw_ddi_translations_dp;
-		n_edp_entries = ARRAY_SIZE(bdw_ddi_translations_edp);
-		n_dp_entries = ARRAY_SIZE(bdw_ddi_translations_dp);
-	}
-
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
-		/* If we're boosting the current, set bit 31 of trans1 */
-		if (dev_priv->vbt.ddi_port_info[port].dp_boost_level)
-			iboost_bit = DDI_BUF_BALANCE_LEG_ENABLE;
-
-		if (WARN_ON(encoder->type == INTEL_OUTPUT_EDP &&
-			    port != PORT_A && port != PORT_E &&
-			    n_edp_entries > 9))
-			n_edp_entries = 9;
-	}
-
 	switch (encoder->type) {
 	case INTEL_OUTPUT_EDP:
-		ddi_translations = ddi_translations_edp;
-		size = n_edp_entries;
+		ddi_translations = intel_ddi_get_buf_trans_edp(dev_priv,
+							       &n_entries);
 		break;
 	case INTEL_OUTPUT_DP:
-		ddi_translations = ddi_translations_dp;
-		size = n_dp_entries;
+		ddi_translations = intel_ddi_get_buf_trans_dp(dev_priv,
+							      &n_entries);
 		break;
 	case INTEL_OUTPUT_ANALOG:
-		ddi_translations = ddi_translations_fdi;
-		size = n_dp_entries;
+		ddi_translations = intel_ddi_get_buf_trans_fdi(dev_priv,
+							       &n_entries);
 		break;
 	default:
-		BUG();
+		MISSING_CASE(encoder->type);
+		return;
+	}
+
+	if (IS_GEN9_BC(dev_priv)) {
+		/* If we're boosting the current, set bit 31 of trans1 */
+		if (dev_priv->vbt.ddi_port_info[port].dp_boost_level)
+			iboost_bit = DDI_BUF_BALANCE_LEG_ENABLE;
+
+		if (WARN_ON(encoder->type == INTEL_OUTPUT_EDP &&
+			    port != PORT_A && port != PORT_E &&
+			    n_entries > 9))
+			n_entries = 9;
 	}
 
-	for (i = 0; i < size; i++) {
+	for (i = 0; i < n_entries; i++) {
 		I915_WRITE(DDI_BUF_TRANS_LO(port, i),
 			   ddi_translations[i].trans1 | iboost_bit);
 		I915_WRITE(DDI_BUF_TRANS_HI(port, i),
@@ -572,7 +605,7 @@ static void intel_prepare_hdmi_ddi_buffers(struct intel_encoder *encoder)
 
 	hdmi_level = intel_ddi_hdmi_level(dev_priv, port);
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	if (IS_GEN9_BC(dev_priv)) {
 		ddi_translations_hdmi = skl_get_buf_trans_hdmi(dev_priv, &n_hdmi_entries);
 
 		/* If we're boosting the current, set bit 31 of trans1 */
@@ -641,15 +674,15 @@ static uint32_t hsw_pll_to_ddi_pll_sel(struct intel_shared_dpll *pll)
  * DDI A (which is used for eDP)
  */
 
-void hsw_fdi_link_train(struct drm_crtc *crtc)
+void hsw_fdi_link_train(struct intel_crtc *crtc,
+			const struct intel_crtc_state *crtc_state)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_encoder *encoder;
 	u32 temp, i, rx_ctl_val, ddi_pll_sel;
 
-	for_each_encoder_on_crtc(dev, crtc, encoder) {
+	for_each_encoder_on_crtc(dev, &crtc->base, encoder) {
 		WARN_ON(encoder->type != INTEL_OUTPUT_ANALOG);
 		intel_prepare_dp_ddi_buffers(encoder);
 	}
@@ -668,7 +701,7 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 	/* Enable the PCH Receiver FDI PLL */
 	rx_ctl_val = dev_priv->fdi_rx_config | FDI_RX_ENHANCE_FRAME_ENABLE |
 		     FDI_RX_PLL_ENABLE |
-		     FDI_DP_PORT_WIDTH(intel_crtc->config->fdi_lanes);
+		     FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
 	I915_WRITE(FDI_RX_CTL(PIPE_A), rx_ctl_val);
 	POSTING_READ(FDI_RX_CTL(PIPE_A));
 	udelay(220);
@@ -678,7 +711,7 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 	I915_WRITE(FDI_RX_CTL(PIPE_A), rx_ctl_val);
 
 	/* Configure Port Clock Select */
-	ddi_pll_sel = hsw_pll_to_ddi_pll_sel(intel_crtc->config->shared_dpll);
+	ddi_pll_sel = hsw_pll_to_ddi_pll_sel(crtc_state->shared_dpll);
 	I915_WRITE(PORT_CLK_SEL(PORT_E), ddi_pll_sel);
 	WARN_ON(ddi_pll_sel != PORT_CLK_SEL_SPLL);
 
@@ -698,7 +731,7 @@ void hsw_fdi_link_train(struct drm_crtc *crtc)
 		 * port reversal bit */
 		I915_WRITE(DDI_BUF_CTL(PORT_E),
 			   DDI_BUF_CTL_ENABLE |
-			   ((intel_crtc->config->fdi_lanes - 1) << 1) |
+			   ((crtc_state->fdi_lanes - 1) << 1) |
 			   DDI_BUF_TRANS_SELECT(i / 2));
 		POSTING_READ(DDI_BUF_CTL(PORT_E));
 
@@ -785,21 +818,20 @@ void intel_ddi_init_dp_buf_reg(struct intel_encoder *encoder)
 }
 
 static struct intel_encoder *
-intel_ddi_get_crtc_encoder(struct drm_crtc *crtc)
+intel_ddi_get_crtc_encoder(struct intel_crtc *crtc)
 {
-	struct drm_device *dev = crtc->dev;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct drm_device *dev = crtc->base.dev;
 	struct intel_encoder *intel_encoder, *ret = NULL;
 	int num_encoders = 0;
 
-	for_each_encoder_on_crtc(dev, crtc, intel_encoder) {
+	for_each_encoder_on_crtc(dev, &crtc->base, intel_encoder) {
 		ret = intel_encoder;
 		num_encoders++;
 	}
 
 	if (num_encoders != 1)
 		WARN(1, "%d encoders on crtc for pipe %c\n", num_encoders,
-		     pipe_name(intel_crtc->pipe));
+		     pipe_name(crtc->pipe));
 
 	BUG_ON(ret == NULL);
 	return ret;
@@ -1089,7 +1121,7 @@ void intel_ddi_clock_get(struct intel_encoder *encoder,
 
 	if (INTEL_GEN(dev_priv) <= 8)
 		hsw_ddi_clock_get(encoder, pipe_config);
-	else if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+	else if (IS_GEN9_BC(dev_priv))
 		skl_ddi_clock_get(encoder, pipe_config);
 	else if (IS_GEN9_LP(dev_priv))
 		bxt_ddi_clock_get(encoder, pipe_config);
@@ -1150,7 +1182,7 @@ bool intel_ddi_pll_select(struct intel_crtc *intel_crtc,
 	struct intel_encoder *intel_encoder =
 		intel_ddi_get_crtc_new_encoder(crtc_state);
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+	if (IS_GEN9_BC(dev_priv))
 		return skl_ddi_pll_select(intel_crtc, crtc_state,
 					  intel_encoder);
 	else if (IS_GEN9_LP(dev_priv))
@@ -1161,12 +1193,12 @@ bool intel_ddi_pll_select(struct intel_crtc *intel_crtc,
 					  intel_encoder);
 }
 
-void intel_ddi_set_pipe_settings(struct drm_crtc *crtc)
+void intel_ddi_set_pipe_settings(const struct intel_crtc_state *crtc_state)
 {
-	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
-	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
 	int type = intel_encoder->type;
 	uint32_t temp;
 
@@ -1174,7 +1206,7 @@ void intel_ddi_set_pipe_settings(struct drm_crtc *crtc)
 		WARN_ON(transcoder_is_dsi(cpu_transcoder));
 
 		temp = TRANS_MSA_SYNC_CLK;
-		switch (intel_crtc->config->pipe_bpp) {
+		switch (crtc_state->pipe_bpp) {
 		case 18:
 			temp |= TRANS_MSA_6_BPC;
 			break;
@@ -1194,12 +1226,12 @@ void intel_ddi_set_pipe_settings(struct drm_crtc *crtc)
 	}
 }
 
-void intel_ddi_set_vc_payload_alloc(struct drm_crtc *crtc, bool state)
+void intel_ddi_set_vc_payload_alloc(const struct intel_crtc_state *crtc_state,
+				    bool state)
 {
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
 	uint32_t temp;
 	temp = I915_READ(TRANS_DDI_FUNC_CTL(cpu_transcoder));
 	if (state == true)
@@ -1209,14 +1241,13 @@ void intel_ddi_set_vc_payload_alloc(struct drm_crtc *crtc, bool state)
 	I915_WRITE(TRANS_DDI_FUNC_CTL(cpu_transcoder), temp);
 }
 
-void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
+void intel_ddi_enable_transcoder_func(const struct intel_crtc_state *crtc_state)
 {
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc);
 	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum pipe pipe = intel_crtc->pipe;
-	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum pipe pipe = crtc->pipe;
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
 	int type = intel_encoder->type;
 	uint32_t temp;
@@ -1225,7 +1256,7 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 	temp = TRANS_DDI_FUNC_ENABLE;
 	temp |= TRANS_DDI_SELECT_PORT(port);
 
-	switch (intel_crtc->config->pipe_bpp) {
+	switch (crtc_state->pipe_bpp) {
 	case 18:
 		temp |= TRANS_DDI_BPC_6;
 		break;
@@ -1242,9 +1273,9 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 		BUG();
 	}
 
-	if (intel_crtc->config->base.adjusted_mode.flags & DRM_MODE_FLAG_PVSYNC)
+	if (crtc_state->base.adjusted_mode.flags & DRM_MODE_FLAG_PVSYNC)
 		temp |= TRANS_DDI_PVSYNC;
-	if (intel_crtc->config->base.adjusted_mode.flags & DRM_MODE_FLAG_PHSYNC)
+	if (crtc_state->base.adjusted_mode.flags & DRM_MODE_FLAG_PHSYNC)
 		temp |= TRANS_DDI_PHSYNC;
 
 	if (cpu_transcoder == TRANSCODER_EDP) {
@@ -1255,8 +1286,8 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 			 * using motion blur mitigation (which we don't
 			 * support). */
 			if (IS_HASWELL(dev_priv) &&
-			    (intel_crtc->config->pch_pfit.enabled ||
-			     intel_crtc->config->pch_pfit.force_thru))
+			    (crtc_state->pch_pfit.enabled ||
+			     crtc_state->pch_pfit.force_thru))
 				temp |= TRANS_DDI_EDP_INPUT_A_ONOFF;
 			else
 				temp |= TRANS_DDI_EDP_INPUT_A_ON;
@@ -1274,20 +1305,20 @@ void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc)
 	}
 
 	if (type == INTEL_OUTPUT_HDMI) {
-		if (intel_crtc->config->has_hdmi_sink)
+		if (crtc_state->has_hdmi_sink)
 			temp |= TRANS_DDI_MODE_SELECT_HDMI;
 		else
 			temp |= TRANS_DDI_MODE_SELECT_DVI;
 	} else if (type == INTEL_OUTPUT_ANALOG) {
 		temp |= TRANS_DDI_MODE_SELECT_FDI;
-		temp |= (intel_crtc->config->fdi_lanes - 1) << 1;
+		temp |= (crtc_state->fdi_lanes - 1) << 1;
 	} else if (type == INTEL_OUTPUT_DP ||
 		   type == INTEL_OUTPUT_EDP) {
 		temp |= TRANS_DDI_MODE_SELECT_DP_SST;
-		temp |= DDI_PORT_WIDTH(intel_crtc->config->lane_count);
+		temp |= DDI_PORT_WIDTH(crtc_state->lane_count);
 	} else if (type == INTEL_OUTPUT_DP_MST) {
 		temp |= TRANS_DDI_MODE_SELECT_DP_MST;
-		temp |= DDI_PORT_WIDTH(intel_crtc->config->lane_count);
+		temp |= DDI_PORT_WIDTH(crtc_state->lane_count);
 	} else {
 		WARN(1, "Invalid encoder type %d for pipe %c\n",
 		     intel_encoder->type, pipe_name(pipe));
@@ -1316,12 +1347,11 @@ bool intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector)
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
 	enum pipe pipe = 0;
 	enum transcoder cpu_transcoder;
-	enum intel_display_power_domain power_domain;
 	uint32_t tmp;
 	bool ret;
 
-	power_domain = intel_display_port_power_domain(intel_encoder);
-	if (!intel_display_power_get_if_enabled(dev_priv, power_domain))
+	if (!intel_display_power_get_if_enabled(dev_priv,
+						intel_encoder->power_domain))
 		return false;
 
 	if (!intel_encoder->get_hw_state(intel_encoder, &pipe)) {
@@ -1363,7 +1393,7 @@ bool intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector)
 	}
 
 out:
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, intel_encoder->power_domain);
 
 	return ret;
 }
@@ -1374,13 +1404,12 @@ bool intel_ddi_get_hw_state(struct intel_encoder *encoder,
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	enum port port = intel_ddi_get_encoder_port(encoder);
-	enum intel_display_power_domain power_domain;
 	u32 tmp;
 	int i;
 	bool ret;
 
-	power_domain = intel_display_port_power_domain(encoder);
-	if (!intel_display_power_get_if_enabled(dev_priv, power_domain))
+	if (!intel_display_power_get_if_enabled(dev_priv,
+						encoder->power_domain))
 		return false;
 
 	ret = false;
@@ -1437,29 +1466,39 @@ out:
 				  "(PHY_CTL %08x)\n", port_name(port), tmp);
 	}
 
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, encoder->power_domain);
 
 	return ret;
 }
 
-void intel_ddi_enable_pipe_clock(struct intel_crtc *intel_crtc)
+static u64 intel_ddi_get_power_domains(struct intel_encoder *encoder)
 {
-	struct drm_crtc *crtc = &intel_crtc->base;
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct intel_digital_port *dig_port = enc_to_dig_port(&encoder->base);
+	enum pipe pipe;
+
+	if (intel_ddi_get_hw_state(encoder, &pipe))
+		return BIT_ULL(dig_port->ddi_io_power_domain);
+
+	return 0;
+}
+
+void intel_ddi_enable_pipe_clock(const struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	struct intel_encoder *intel_encoder = intel_ddi_get_crtc_encoder(crtc);
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
-	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
 
 	if (cpu_transcoder != TRANSCODER_EDP)
 		I915_WRITE(TRANS_CLK_SEL(cpu_transcoder),
 			   TRANS_CLK_SEL_PORT(port));
 }
 
-void intel_ddi_disable_pipe_clock(struct intel_crtc *intel_crtc)
+void intel_ddi_disable_pipe_clock(const struct intel_crtc_state *crtc_state)
 {
-	struct drm_i915_private *dev_priv = to_i915(intel_crtc->base.dev);
-	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
 
 	if (cpu_transcoder != TRANSCODER_EDP)
 		I915_WRITE(TRANS_CLK_SEL(cpu_transcoder),
@@ -1582,50 +1621,38 @@ static void bxt_ddi_vswing_sequence(struct drm_i915_private *dev_priv,
 				     ddi_translations[level].deemphasis);
 }
 
-static uint32_t translate_signal_level(int signal_levels)
+u8 intel_ddi_dp_voltage_max(struct intel_encoder *encoder)
 {
-	uint32_t level;
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	int n_entries;
 
-	switch (signal_levels) {
-	default:
-		DRM_DEBUG_KMS("Unsupported voltage swing/pre-emphasis level: 0x%x\n",
-			      signal_levels);
-	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_0:
-		level = 0;
-		break;
-	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_1:
-		level = 1;
-		break;
-	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_2:
-		level = 2;
-		break;
-	case DP_TRAIN_VOLTAGE_SWING_LEVEL_0 | DP_TRAIN_PRE_EMPH_LEVEL_3:
-		level = 3;
-		break;
+	if (encoder->type == INTEL_OUTPUT_EDP)
+		intel_ddi_get_buf_trans_edp(dev_priv, &n_entries);
+	else
+		intel_ddi_get_buf_trans_dp(dev_priv, &n_entries);
 
-	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_0:
-		level = 4;
-		break;
-	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_1:
-		level = 5;
-		break;
-	case DP_TRAIN_VOLTAGE_SWING_LEVEL_1 | DP_TRAIN_PRE_EMPH_LEVEL_2:
-		level = 6;
-		break;
+	if (WARN_ON(n_entries < 1))
+		n_entries = 1;
+	if (WARN_ON(n_entries > ARRAY_SIZE(index_to_dp_signal_levels)))
+		n_entries = ARRAY_SIZE(index_to_dp_signal_levels);
 
-	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_0:
-		level = 7;
-		break;
-	case DP_TRAIN_VOLTAGE_SWING_LEVEL_2 | DP_TRAIN_PRE_EMPH_LEVEL_1:
-		level = 8;
-		break;
+	return index_to_dp_signal_levels[n_entries - 1] &
+		DP_TRAIN_VOLTAGE_SWING_MASK;
+}
 
-	case DP_TRAIN_VOLTAGE_SWING_LEVEL_3 | DP_TRAIN_PRE_EMPH_LEVEL_0:
-		level = 9;
-		break;
+static uint32_t translate_signal_level(int signal_levels)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(index_to_dp_signal_levels); i++) {
+		if (index_to_dp_signal_levels[i] == signal_levels)
+			return i;
 	}
 
-	return level;
+	WARN(1, "Unsupported voltage swing/pre-emphasis level: 0x%x\n",
+	     signal_levels);
+
+	return 0;
 }
 
 uint32_t ddi_signal_levels(struct intel_dp *intel_dp)
@@ -1641,7 +1668,7 @@ uint32_t ddi_signal_levels(struct intel_dp *intel_dp)
 
 	level = translate_signal_level(signal_levels);
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+	if (IS_GEN9_BC(dev_priv))
 		skl_ddi_set_iboost(encoder, level);
 	else if (IS_GEN9_LP(dev_priv))
 		bxt_ddi_vswing_sequence(dev_priv, level, port, encoder->type);
@@ -1658,7 +1685,7 @@ void intel_ddi_clk_select(struct intel_encoder *encoder,
 	if (WARN_ON(!pll))
 		return;
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	if (IS_GEN9_BC(dev_priv)) {
 		uint32_t val;
 
 		/* DDI -> PLL mapping  */
@@ -1684,6 +1711,9 @@ static void intel_ddi_pre_enable_dp(struct intel_encoder *encoder,
 	struct intel_dp *intel_dp = enc_to_intel_dp(&encoder->base);
 	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	enum port port = intel_ddi_get_encoder_port(encoder);
+	struct intel_digital_port *dig_port = enc_to_dig_port(&encoder->base);
+
+	WARN_ON(link_mst && (port == PORT_A || port == PORT_E));
 
 	intel_dp_set_link_params(intel_dp, link_rate, lane_count,
 				 link_mst);
@@ -1691,6 +1721,9 @@ static void intel_ddi_pre_enable_dp(struct intel_encoder *encoder,
 		intel_edp_panel_on(intel_dp);
 
 	intel_ddi_clk_select(encoder, pll);
+
+	intel_display_power_get(dev_priv, dig_port->ddi_io_power_domain);
+
 	intel_prepare_dp_ddi_buffers(encoder);
 	intel_ddi_init_dp_buf_reg(encoder);
 	intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON);
@@ -1710,11 +1743,15 @@ static void intel_ddi_pre_enable_hdmi(struct intel_encoder *encoder,
 	struct drm_encoder *drm_encoder = &encoder->base;
 	enum port port = intel_ddi_get_encoder_port(encoder);
 	int level = intel_ddi_hdmi_level(dev_priv, port);
+	struct intel_digital_port *dig_port = enc_to_dig_port(&encoder->base);
 
 	intel_dp_dual_mode_set_tmds_output(intel_hdmi, true);
 	intel_ddi_clk_select(encoder, pll);
+
+	intel_display_power_get(dev_priv, dig_port->ddi_io_power_domain);
+
 	intel_prepare_hdmi_ddi_buffers(encoder);
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+	if (IS_GEN9_BC(dev_priv))
 		skl_ddi_set_iboost(encoder, level);
 	else if (IS_GEN9_LP(dev_priv))
 		bxt_ddi_vswing_sequence(dev_priv, level, port,
@@ -1729,23 +1766,21 @@ static void intel_ddi_pre_enable(struct intel_encoder *intel_encoder,
 				 struct intel_crtc_state *pipe_config,
 				 struct drm_connector_state *conn_state)
 {
-	struct drm_encoder *encoder = &intel_encoder->base;
-	struct intel_crtc *crtc = to_intel_crtc(encoder->crtc);
 	int type = intel_encoder->type;
 
 	if (type == INTEL_OUTPUT_DP || type == INTEL_OUTPUT_EDP) {
 		intel_ddi_pre_enable_dp(intel_encoder,
-					crtc->config->port_clock,
-					crtc->config->lane_count,
-					crtc->config->shared_dpll,
-					intel_crtc_has_type(crtc->config,
+					pipe_config->port_clock,
+					pipe_config->lane_count,
+					pipe_config->shared_dpll,
+					intel_crtc_has_type(pipe_config,
 							    INTEL_OUTPUT_DP_MST));
 	}
 	if (type == INTEL_OUTPUT_HDMI) {
 		intel_ddi_pre_enable_hdmi(intel_encoder,
 					  pipe_config->has_hdmi_sink,
 					  pipe_config, conn_state,
-					  crtc->config->shared_dpll);
+					  pipe_config->shared_dpll);
 	}
 }
 
@@ -1756,6 +1791,7 @@ static void intel_ddi_post_disable(struct intel_encoder *intel_encoder,
 	struct drm_encoder *encoder = &intel_encoder->base;
 	struct drm_i915_private *dev_priv = to_i915(encoder->dev);
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
+	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
 	int type = intel_encoder->type;
 	uint32_t val;
 	bool wait = false;
@@ -1784,7 +1820,10 @@ static void intel_ddi_post_disable(struct intel_encoder *intel_encoder,
 		intel_edp_panel_off(intel_dp);
 	}
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+	if (dig_port)
+		intel_display_power_put(dev_priv, dig_port->ddi_io_power_domain);
+
+	if (IS_GEN9_BC(dev_priv))
 		I915_WRITE(DPLL_CTRL2, (I915_READ(DPLL_CTRL2) |
 					DPLL_CTRL2_DDI_CLK_OFF(port)));
 	else if (INTEL_GEN(dev_priv) < 9)
@@ -1835,8 +1874,6 @@ static void intel_enable_ddi(struct intel_encoder *intel_encoder,
 			     struct drm_connector_state *conn_state)
 {
 	struct drm_encoder *encoder = &intel_encoder->base;
-	struct drm_crtc *crtc = encoder->crtc;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct drm_i915_private *dev_priv = to_i915(encoder->dev);
 	enum port port = intel_ddi_get_encoder_port(intel_encoder);
 	int type = intel_encoder->type;
@@ -1863,10 +1900,8 @@ static void intel_enable_ddi(struct intel_encoder *intel_encoder,
 		intel_edp_drrs_enable(intel_dp, pipe_config);
 	}
 
-	if (intel_crtc->config->has_audio) {
-		intel_display_power_get(dev_priv, POWER_DOMAIN_AUDIO);
+	if (pipe_config->has_audio)
 		intel_audio_codec_enable(intel_encoder, pipe_config, conn_state);
-	}
 }
 
 static void intel_disable_ddi(struct intel_encoder *intel_encoder,
@@ -1874,16 +1909,10 @@ static void intel_disable_ddi(struct intel_encoder *intel_encoder,
 			      struct drm_connector_state *old_conn_state)
 {
 	struct drm_encoder *encoder = &intel_encoder->base;
-	struct drm_crtc *crtc = encoder->crtc;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int type = intel_encoder->type;
-	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
 
-	if (intel_crtc->config->has_audio) {
+	if (old_crtc_state->has_audio)
 		intel_audio_codec_disable(intel_encoder);
-		intel_display_power_put(dev_priv, POWER_DOMAIN_AUDIO);
-	}
 
 	if (type == INTEL_OUTPUT_EDP) {
 		struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
@@ -1898,8 +1927,7 @@ static void bxt_ddi_pre_pll_enable(struct intel_encoder *encoder,
 				   struct intel_crtc_state *pipe_config,
 				   struct drm_connector_state *conn_state)
 {
-	struct intel_crtc *intel_crtc = to_intel_crtc(encoder->base.crtc);
-	uint8_t mask = intel_crtc->config->lane_lat_optim_mask;
+	uint8_t mask = pipe_config->lane_lat_optim_mask;
 
 	bxt_ddi_phy_set_lane_optim_mask(encoder, mask);
 }
@@ -2126,45 +2154,6 @@ intel_ddi_init_hdmi_connector(struct intel_digital_port *intel_dig_port)
 	return connector;
 }
 
-struct intel_shared_dpll *
-intel_ddi_get_link_dpll(struct intel_dp *intel_dp, int clock)
-{
-	struct intel_connector *connector = intel_dp->attached_connector;
-	struct intel_encoder *encoder = connector->encoder;
-	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
-	struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp);
-	struct intel_shared_dpll *pll = NULL;
-	struct intel_shared_dpll_state tmp_pll_state;
-	enum intel_dpll_id dpll_id;
-
-	if (IS_GEN9_LP(dev_priv)) {
-		dpll_id =  (enum intel_dpll_id)dig_port->port;
-		/*
-		 * Select the required PLL. This works for platforms where
-		 * there is no shared DPLL.
-		 */
-		pll = &dev_priv->shared_dplls[dpll_id];
-		if (WARN_ON(pll->active_mask)) {
-
-			DRM_ERROR("Shared DPLL in use. active_mask:%x\n",
-				  pll->active_mask);
-			return NULL;
-		}
-		tmp_pll_state = pll->state;
-		if (!bxt_ddi_dp_set_dpll_hw_state(clock,
-						  &pll->state.hw_state)) {
-			DRM_ERROR("Could not setup DPLL\n");
-			pll->state = tmp_pll_state;
-			return NULL;
-		}
-	} else if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
-		pll = skl_find_link_pll(dev_priv, clock);
-	} else if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) {
-		pll = hsw_ddi_dp_get_dpll(encoder, clock);
-	}
-	return pll;
-}
-
 void intel_ddi_init(struct drm_i915_private *dev_priv, enum port port)
 {
 	struct intel_digital_port *intel_dig_port;
@@ -2241,12 +2230,38 @@ void intel_ddi_init(struct drm_i915_private *dev_priv, enum port port)
 	intel_encoder->get_hw_state = intel_ddi_get_hw_state;
 	intel_encoder->get_config = intel_ddi_get_config;
 	intel_encoder->suspend = intel_dp_encoder_suspend;
+	intel_encoder->get_power_domains = intel_ddi_get_power_domains;
 
 	intel_dig_port->port = port;
 	intel_dig_port->saved_port_bits = I915_READ(DDI_BUF_CTL(port)) &
 					  (DDI_BUF_PORT_REVERSAL |
 					   DDI_A_4_LANES);
 
+	switch (port) {
+	case PORT_A:
+		intel_dig_port->ddi_io_power_domain =
+			POWER_DOMAIN_PORT_DDI_A_IO;
+		break;
+	case PORT_B:
+		intel_dig_port->ddi_io_power_domain =
+			POWER_DOMAIN_PORT_DDI_B_IO;
+		break;
+	case PORT_C:
+		intel_dig_port->ddi_io_power_domain =
+			POWER_DOMAIN_PORT_DDI_C_IO;
+		break;
+	case PORT_D:
+		intel_dig_port->ddi_io_power_domain =
+			POWER_DOMAIN_PORT_DDI_D_IO;
+		break;
+	case PORT_E:
+		intel_dig_port->ddi_io_power_domain =
+			POWER_DOMAIN_PORT_DDI_E_IO;
+		break;
+	default:
+		MISSING_CASE(port);
+	}
+
 	/*
 	 * Bspec says that DDI_A_4_LANES is the only supported configuration
 	 * for Broxton.  Yet some BIOS fail to set this bit on port A if eDP
@@ -2265,6 +2280,7 @@ void intel_ddi_init(struct drm_i915_private *dev_priv, enum port port)
 	intel_dig_port->max_lanes = max_lanes;
 
 	intel_encoder->type = INTEL_OUTPUT_UNKNOWN;
+	intel_encoder->power_domain = intel_port_to_power_domain(port);
 	intel_encoder->port = port;
 	intel_encoder->crtc_mask = (1 << 0) | (1 << 1) | (1 << 2);
 	intel_encoder->cloneable = 0;
@@ -2274,14 +2290,7 @@ void intel_ddi_init(struct drm_i915_private *dev_priv, enum port port)
 			goto err;
 
 		intel_dig_port->hpd_pulse = intel_dp_hpd_pulse;
-		/*
-		 * On BXT A0/A1, sw needs to activate DDIA HPD logic and
-		 * interrupts to check the external panel connection.
-		 */
-		if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1) && port == PORT_B)
-			dev_priv->hotplug.irq_port[PORT_A] = intel_dig_port;
-		else
-			dev_priv->hotplug.irq_port[port] = intel_dig_port;
+		dev_priv->hotplug.irq_port[port] = intel_dig_port;
 	}
 
 	/* In theory we don't need the encoder->type check, but leave it just in
diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c
index fcf81815daff..9fc6ab783008 100644
--- a/drivers/gpu/drm/i915/intel_device_info.c
+++ b/drivers/gpu/drm/i915/intel_device_info.c
@@ -56,6 +56,8 @@ static const char * const platform_names[] = {
 
 const char *intel_platform_name(enum intel_platform platform)
 {
+	BUILD_BUG_ON(ARRAY_SIZE(platform_names) != INTEL_MAX_PLATFORMS);
+
 	if (WARN_ON_ONCE(platform >= ARRAY_SIZE(platform_names) ||
 			 platform_names[platform] == NULL))
 		return "<unknown>";
@@ -234,7 +236,7 @@ static void broadwell_sseu_info_init(struct drm_i915_private *dev_priv)
 	 * The subslice disable field is global, i.e. it applies
 	 * to each of the enabled slices.
 	 */
-	sseu->subslice_mask = BIT(ss_max) - 1;
+	sseu->subslice_mask = GENMASK(ss_max - 1, 0);
 	sseu->subslice_mask &= ~((fuse2 & GEN8_F2_SS_DIS_MASK) >>
 				 GEN8_F2_SS_DIS_SHIFT);
 
@@ -410,10 +412,6 @@ void intel_device_info_runtime_init(struct drm_i915_private *dev_priv)
 
 	info->has_snoop = !info->has_llc;
 
-	/* Snooping is broken on BXT A stepping. */
-	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1))
-		info->has_snoop = false;
-
 	DRM_DEBUG_DRIVER("slice mask: %04x\n", info->sseu.slice_mask);
 	DRM_DEBUG_DRIVER("slice total: %u\n", hweight8(info->sseu.slice_mask));
 	DRM_DEBUG_DRIVER("subslice total: %u\n",
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index a2fece5e9fb3..7369ee31ad91 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -37,6 +37,7 @@
 #include "intel_frontbuffer.h"
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
+#include "i915_gem_clflush.h"
 #include "intel_dsi.h"
 #include "i915_trace.h"
 #include <drm/drm_atomic.h>
@@ -96,10 +97,9 @@ static void i9xx_crtc_clock_get(struct intel_crtc *crtc,
 static void ironlake_pch_clock_get(struct intel_crtc *crtc,
 				   struct intel_crtc_state *pipe_config);
 
-static int intel_framebuffer_init(struct drm_device *dev,
-				  struct intel_framebuffer *ifb,
-				  struct drm_mode_fb_cmd2 *mode_cmd,
-				  struct drm_i915_gem_object *obj);
+static int intel_framebuffer_init(struct intel_framebuffer *ifb,
+				  struct drm_i915_gem_object *obj,
+				  struct drm_mode_fb_cmd2 *mode_cmd);
 static void i9xx_set_pipeconf(struct intel_crtc *intel_crtc);
 static void intel_set_pipe_timings(struct intel_crtc *intel_crtc);
 static void intel_set_pipe_src_size(struct intel_crtc *intel_crtc);
@@ -122,9 +122,6 @@ static void ironlake_pfit_disable(struct intel_crtc *crtc, bool force);
 static void ironlake_pfit_enable(struct intel_crtc *crtc);
 static void intel_modeset_setup_hw_state(struct drm_device *dev);
 static void intel_pre_disable_primary_noatomic(struct drm_crtc *crtc);
-static int ilk_max_pixel_rate(struct drm_atomic_state *state);
-static int glk_calc_cdclk(int max_pixclk);
-static int bxt_calc_cdclk(int max_pixclk);
 
 struct intel_limit {
 	struct {
@@ -138,7 +135,7 @@ struct intel_limit {
 };
 
 /* returns HPLL frequency in kHz */
-static int valleyview_get_vco(struct drm_i915_private *dev_priv)
+int vlv_get_hpll_vco(struct drm_i915_private *dev_priv)
 {
 	int hpll_freq, vco_freq[] = { 800, 1600, 2000, 2400 };
 
@@ -170,73 +167,16 @@ int vlv_get_cck_clock(struct drm_i915_private *dev_priv,
 	return DIV_ROUND_CLOSEST(ref_freq << 1, divider + 1);
 }
 
-static int vlv_get_cck_clock_hpll(struct drm_i915_private *dev_priv,
-				  const char *name, u32 reg)
+int vlv_get_cck_clock_hpll(struct drm_i915_private *dev_priv,
+			   const char *name, u32 reg)
 {
 	if (dev_priv->hpll_freq == 0)
-		dev_priv->hpll_freq = valleyview_get_vco(dev_priv);
+		dev_priv->hpll_freq = vlv_get_hpll_vco(dev_priv);
 
 	return vlv_get_cck_clock(dev_priv, name, reg,
 				 dev_priv->hpll_freq);
 }
 
-static int
-intel_pch_rawclk(struct drm_i915_private *dev_priv)
-{
-	return (I915_READ(PCH_RAWCLK_FREQ) & RAWCLK_FREQ_MASK) * 1000;
-}
-
-static int
-intel_vlv_hrawclk(struct drm_i915_private *dev_priv)
-{
-	/* RAWCLK_FREQ_VLV register updated from power well code */
-	return vlv_get_cck_clock_hpll(dev_priv, "hrawclk",
-				      CCK_DISPLAY_REF_CLOCK_CONTROL);
-}
-
-static int
-intel_g4x_hrawclk(struct drm_i915_private *dev_priv)
-{
-	uint32_t clkcfg;
-
-	/* hrawclock is 1/4 the FSB frequency */
-	clkcfg = I915_READ(CLKCFG);
-	switch (clkcfg & CLKCFG_FSB_MASK) {
-	case CLKCFG_FSB_400:
-		return 100000;
-	case CLKCFG_FSB_533:
-		return 133333;
-	case CLKCFG_FSB_667:
-		return 166667;
-	case CLKCFG_FSB_800:
-		return 200000;
-	case CLKCFG_FSB_1067:
-		return 266667;
-	case CLKCFG_FSB_1333:
-		return 333333;
-	/* these two are just a guess; one of them might be right */
-	case CLKCFG_FSB_1600:
-	case CLKCFG_FSB_1600_ALT:
-		return 400000;
-	default:
-		return 133333;
-	}
-}
-
-void intel_update_rawclk(struct drm_i915_private *dev_priv)
-{
-	if (HAS_PCH_SPLIT(dev_priv))
-		dev_priv->rawclk_freq = intel_pch_rawclk(dev_priv);
-	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-		dev_priv->rawclk_freq = intel_vlv_hrawclk(dev_priv);
-	else if (IS_G4X(dev_priv) || IS_PINEVIEW(dev_priv))
-		dev_priv->rawclk_freq = intel_g4x_hrawclk(dev_priv);
-	else
-		return; /* no rawclk on other platforms, or no need to know it */
-
-	DRM_DEBUG_DRIVER("rawclk rate: %d kHz\n", dev_priv->rawclk_freq);
-}
-
 static void intel_update_czclk(struct drm_i915_private *dev_priv)
 {
 	if (!(IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)))
@@ -2112,11 +2052,13 @@ static void intel_tile_dims(const struct drm_i915_private *dev_priv,
 }
 
 unsigned int
-intel_fb_align_height(struct drm_device *dev, unsigned int height,
-		      uint32_t pixel_format, uint64_t fb_modifier)
+intel_fb_align_height(struct drm_i915_private *dev_priv,
+		      unsigned int height,
+		      uint32_t pixel_format,
+		      uint64_t fb_modifier)
 {
 	unsigned int cpp = drm_format_plane_cpp(pixel_format, 0);
-	unsigned int tile_height = intel_tile_height(to_i915(dev), fb_modifier, cpp);
+	unsigned int tile_height = intel_tile_height(dev_priv, fb_modifier, cpp);
 
 	return ALIGN(height, tile_height);
 }
@@ -2682,15 +2624,13 @@ intel_alloc_initial_plane_obj(struct intel_crtc *crtc,
 		return false;
 
 	mutex_lock(&dev->struct_mutex);
-
 	obj = i915_gem_object_create_stolen_for_preallocated(dev_priv,
 							     base_aligned,
 							     base_aligned,
 							     size_aligned);
-	if (!obj) {
-		mutex_unlock(&dev->struct_mutex);
+	mutex_unlock(&dev->struct_mutex);
+	if (!obj)
 		return false;
-	}
 
 	if (plane_config->tiling == I915_TILING_X)
 		obj->tiling_and_stride = fb->pitches[0] | I915_TILING_X;
@@ -2702,20 +2642,17 @@ intel_alloc_initial_plane_obj(struct intel_crtc *crtc,
 	mode_cmd.modifier[0] = fb->modifier;
 	mode_cmd.flags = DRM_MODE_FB_MODIFIERS;
 
-	if (intel_framebuffer_init(dev, to_intel_framebuffer(fb),
-				   &mode_cmd, obj)) {
+	if (intel_framebuffer_init(to_intel_framebuffer(fb), obj, &mode_cmd)) {
 		DRM_DEBUG_KMS("intel fb init failed\n");
 		goto out_unref_obj;
 	}
 
-	mutex_unlock(&dev->struct_mutex);
 
 	DRM_DEBUG_KMS("initial plane fb obj %p\n", obj);
 	return true;
 
 out_unref_obj:
 	i915_gem_object_put(obj);
-	mutex_unlock(&dev->struct_mutex);
 	return false;
 }
 
@@ -2734,6 +2671,29 @@ update_state_fb(struct drm_plane *plane)
 }
 
 static void
+intel_set_plane_visible(struct intel_crtc_state *crtc_state,
+			struct intel_plane_state *plane_state,
+			bool visible)
+{
+	struct intel_plane *plane = to_intel_plane(plane_state->base.plane);
+
+	plane_state->base.visible = visible;
+
+	/* FIXME pre-g4x don't work like this */
+	if (visible) {
+		crtc_state->base.plane_mask |= BIT(drm_plane_index(&plane->base));
+		crtc_state->active_planes |= BIT(plane->id);
+	} else {
+		crtc_state->base.plane_mask &= ~BIT(drm_plane_index(&plane->base));
+		crtc_state->active_planes &= ~BIT(plane->id);
+	}
+
+	DRM_DEBUG_KMS("%s active planes 0x%x\n",
+		      crtc_state->base.crtc->name,
+		      crtc_state->active_planes);
+}
+
+static void
 intel_find_initial_plane_obj(struct intel_crtc *intel_crtc,
 			     struct intel_initial_plane_config *plane_config)
 {
@@ -2790,9 +2750,11 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc,
 	 * simplest solution is to just disable the primary plane now and
 	 * pretend the BIOS never had it enabled.
 	 */
-	plane_state->visible = false;
-	crtc_state->plane_mask &= ~(1 << drm_plane_index(primary));
+	intel_set_plane_visible(to_intel_crtc_state(crtc_state),
+				to_intel_plane_state(plane_state),
+				false);
 	intel_pre_disable_primary_noatomic(&intel_crtc->base);
+	trace_intel_disable_plane(primary, intel_crtc);
 	intel_plane->disable_plane(primary, &intel_crtc->base);
 
 	return;
@@ -2831,7 +2793,11 @@ valid_fb:
 	drm_framebuffer_reference(fb);
 	primary->fb = primary->state->fb = fb;
 	primary->crtc = primary->state->crtc = &intel_crtc->base;
-	intel_crtc->base.state->plane_mask |= (1 << drm_plane_index(primary));
+
+	intel_set_plane_visible(to_intel_crtc_state(crtc_state),
+				to_intel_plane_state(plane_state),
+				true);
+
 	atomic_or(to_intel_plane(primary)->frontbuffer_bit,
 		  &obj->frontbuffer_bits);
 }
@@ -3386,13 +3352,22 @@ static void skylake_update_primary_plane(struct drm_plane *plane,
 	int dst_w = drm_rect_width(&plane_state->base.dst);
 	int dst_h = drm_rect_height(&plane_state->base.dst);
 
-	plane_ctl = PLANE_CTL_ENABLE |
-		    PLANE_CTL_PIPE_GAMMA_ENABLE |
-		    PLANE_CTL_PIPE_CSC_ENABLE;
+	plane_ctl = PLANE_CTL_ENABLE;
+
+	if (IS_GEMINILAKE(dev_priv)) {
+		I915_WRITE(PLANE_COLOR_CTL(pipe, plane_id),
+			   PLANE_COLOR_PIPE_GAMMA_ENABLE |
+			   PLANE_COLOR_PIPE_CSC_ENABLE |
+			   PLANE_COLOR_PLANE_GAMMA_DISABLE);
+	} else {
+		plane_ctl |=
+			PLANE_CTL_PIPE_GAMMA_ENABLE |
+			PLANE_CTL_PIPE_CSC_ENABLE |
+			PLANE_CTL_PLANE_GAMMA_DISABLE;
+	}
 
 	plane_ctl |= skl_plane_ctl_format(fb->format->format);
 	plane_ctl |= skl_plane_ctl_tiling(fb->modifier);
-	plane_ctl |= PLANE_CTL_PLANE_GAMMA_DISABLE;
 	plane_ctl |= skl_plane_ctl_rotation(rotation);
 
 	/* Sizes are 0 based */
@@ -3473,10 +3448,14 @@ static void intel_update_primary_planes(struct drm_device *dev)
 		struct intel_plane_state *plane_state =
 			to_intel_plane_state(plane->base.state);
 
-		if (plane_state->base.visible)
+		if (plane_state->base.visible) {
+			trace_intel_update_plane(&plane->base,
+						 to_intel_crtc(crtc));
+
 			plane->update_plane(&plane->base,
 					    to_intel_crtc_state(crtc->state),
 					    plane_state);
+		}
 	}
 }
 
@@ -3505,7 +3484,8 @@ __intel_display_resume(struct drm_device *dev,
 	}
 
 	/* ignore any reset values/BIOS leftovers in the WM registers */
-	to_intel_atomic_state(state)->skip_intermediate_wm = true;
+	if (!HAS_GMCH_DISPLAY(to_i915(dev)))
+		to_intel_atomic_state(state)->skip_intermediate_wm = true;
 
 	ret = drm_atomic_helper_commit_duplicated_state(state, ctx);
 
@@ -3701,12 +3681,11 @@ static void intel_update_pipe_config(struct intel_crtc *crtc,
 	}
 }
 
-static void intel_fdi_normal_train(struct drm_crtc *crtc)
+static void intel_fdi_normal_train(struct intel_crtc *crtc)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
+	int pipe = crtc->pipe;
 	i915_reg_t reg;
 	u32 temp;
 
@@ -3744,12 +3723,12 @@ static void intel_fdi_normal_train(struct drm_crtc *crtc)
 }
 
 /* The FDI link training functions for ILK/Ibexpeak. */
-static void ironlake_fdi_link_train(struct drm_crtc *crtc)
+static void ironlake_fdi_link_train(struct intel_crtc *crtc,
+				    const struct intel_crtc_state *crtc_state)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
+	int pipe = crtc->pipe;
 	i915_reg_t reg;
 	u32 temp, tries;
 
@@ -3770,7 +3749,7 @@ static void ironlake_fdi_link_train(struct drm_crtc *crtc)
 	reg = FDI_TX_CTL(pipe);
 	temp = I915_READ(reg);
 	temp &= ~FDI_DP_PORT_WIDTH_MASK;
-	temp |= FDI_DP_PORT_WIDTH(intel_crtc->config->fdi_lanes);
+	temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
 	temp &= ~FDI_LINK_TRAIN_NONE;
 	temp |= FDI_LINK_TRAIN_PATTERN_1;
 	I915_WRITE(reg, temp | FDI_TX_ENABLE);
@@ -3845,12 +3824,12 @@ static const int snb_b_fdi_train_param[] = {
 };
 
 /* The FDI link training functions for SNB/Cougarpoint. */
-static void gen6_fdi_link_train(struct drm_crtc *crtc)
+static void gen6_fdi_link_train(struct intel_crtc *crtc,
+				const struct intel_crtc_state *crtc_state)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
+	int pipe = crtc->pipe;
 	i915_reg_t reg;
 	u32 temp, i, retry;
 
@@ -3869,7 +3848,7 @@ static void gen6_fdi_link_train(struct drm_crtc *crtc)
 	reg = FDI_TX_CTL(pipe);
 	temp = I915_READ(reg);
 	temp &= ~FDI_DP_PORT_WIDTH_MASK;
-	temp |= FDI_DP_PORT_WIDTH(intel_crtc->config->fdi_lanes);
+	temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
 	temp &= ~FDI_LINK_TRAIN_NONE;
 	temp |= FDI_LINK_TRAIN_PATTERN_1;
 	temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
@@ -3978,12 +3957,12 @@ static void gen6_fdi_link_train(struct drm_crtc *crtc)
 }
 
 /* Manual link training for Ivy Bridge A0 parts */
-static void ivb_manual_fdi_link_train(struct drm_crtc *crtc)
+static void ivb_manual_fdi_link_train(struct intel_crtc *crtc,
+				      const struct intel_crtc_state *crtc_state)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
+	int pipe = crtc->pipe;
 	i915_reg_t reg;
 	u32 temp, i, j;
 
@@ -4021,7 +4000,7 @@ static void ivb_manual_fdi_link_train(struct drm_crtc *crtc)
 		reg = FDI_TX_CTL(pipe);
 		temp = I915_READ(reg);
 		temp &= ~FDI_DP_PORT_WIDTH_MASK;
-		temp |= FDI_DP_PORT_WIDTH(intel_crtc->config->fdi_lanes);
+		temp |= FDI_DP_PORT_WIDTH(crtc_state->fdi_lanes);
 		temp |= FDI_LINK_TRAIN_PATTERN_1_IVB;
 		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
 		temp |= snb_b_fdi_train_param[j/2];
@@ -4308,10 +4287,10 @@ void lpt_disable_iclkip(struct drm_i915_private *dev_priv)
 }
 
 /* Program iCLKIP clock to the desired frequency */
-static void lpt_program_iclkip(struct drm_crtc *crtc)
+static void lpt_program_iclkip(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
-	int clock = to_intel_crtc(crtc)->config->base.adjusted_mode.crtc_clock;
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	int clock = crtc->config->base.adjusted_mode.crtc_clock;
 	u32 divsel, phaseinc, auxdiv, phasedir = 0;
 	u32 temp;
 
@@ -4492,12 +4471,12 @@ static void ivybridge_update_fdi_bc_bifurcation(struct intel_crtc *intel_crtc)
 
 /* Return which DP Port should be selected for Transcoder DP control */
 static enum port
-intel_trans_dp_port_sel(struct drm_crtc *crtc)
+intel_trans_dp_port_sel(struct intel_crtc *crtc)
 {
-	struct drm_device *dev = crtc->dev;
+	struct drm_device *dev = crtc->base.dev;
 	struct intel_encoder *encoder;
 
-	for_each_encoder_on_crtc(dev, crtc, encoder) {
+	for_each_encoder_on_crtc(dev, &crtc->base, encoder) {
 		if (encoder->type == INTEL_OUTPUT_DP ||
 		    encoder->type == INTEL_OUTPUT_EDP)
 			return enc_to_dig_port(&encoder->base)->port;
@@ -4514,18 +4493,18 @@ intel_trans_dp_port_sel(struct drm_crtc *crtc)
  *   - DP transcoding bits
  *   - transcoder
  */
-static void ironlake_pch_enable(struct drm_crtc *crtc)
+static void ironlake_pch_enable(const struct intel_crtc_state *crtc_state)
 {
-	struct drm_device *dev = crtc->dev;
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc);
+	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
+	int pipe = crtc->pipe;
 	u32 temp;
 
 	assert_pch_transcoder_disabled(dev_priv, pipe);
 
 	if (IS_IVYBRIDGE(dev_priv))
-		ivybridge_update_fdi_bc_bifurcation(intel_crtc);
+		ivybridge_update_fdi_bc_bifurcation(crtc);
 
 	/* Write the TU size bits before fdi link training, so that error
 	 * detection works. */
@@ -4533,7 +4512,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 		   I915_READ(PIPE_DATA_M1(pipe)) & TU_SIZE_MASK);
 
 	/* For PCH output, training FDI link */
-	dev_priv->display.fdi_link_train(crtc);
+	dev_priv->display.fdi_link_train(crtc, crtc_state);
 
 	/* We need to program the right clock selection before writing the pixel
 	 * mutliplier into the DPLL. */
@@ -4543,7 +4522,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 		temp = I915_READ(PCH_DPLL_SEL);
 		temp |= TRANS_DPLL_ENABLE(pipe);
 		sel = TRANS_DPLLB_SEL(pipe);
-		if (intel_crtc->config->shared_dpll ==
+		if (crtc_state->shared_dpll ==
 		    intel_get_shared_dpll_by_id(dev_priv, DPLL_ID_PCH_PLL_B))
 			temp |= sel;
 		else
@@ -4558,19 +4537,19 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	 * Note that enable_shared_dpll tries to do the right thing, but
 	 * get_shared_dpll unconditionally resets the pll - we need that to have
 	 * the right LVDS enable sequence. */
-	intel_enable_shared_dpll(intel_crtc);
+	intel_enable_shared_dpll(crtc);
 
 	/* set transcoder timing, panel must allow it */
 	assert_panel_unlocked(dev_priv, pipe);
-	ironlake_pch_transcoder_set_timings(intel_crtc, pipe);
+	ironlake_pch_transcoder_set_timings(crtc, pipe);
 
 	intel_fdi_normal_train(crtc);
 
 	/* For PCH DP, enable TRANS_DP_CTL */
 	if (HAS_PCH_CPT(dev_priv) &&
-	    intel_crtc_has_dp_encoder(intel_crtc->config)) {
+	    intel_crtc_has_dp_encoder(crtc_state)) {
 		const struct drm_display_mode *adjusted_mode =
-			&intel_crtc->config->base.adjusted_mode;
+			&crtc_state->base.adjusted_mode;
 		u32 bpc = (I915_READ(PIPECONF(pipe)) & PIPECONF_BPC_MASK) >> 5;
 		i915_reg_t reg = TRANS_DP_CTL(pipe);
 		temp = I915_READ(reg);
@@ -4605,19 +4584,18 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	ironlake_enable_pch_transcoder(dev_priv, pipe);
 }
 
-static void lpt_pch_enable(struct drm_crtc *crtc)
+static void lpt_pch_enable(const struct intel_crtc_state *crtc_state)
 {
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	enum transcoder cpu_transcoder = crtc_state->cpu_transcoder;
 
 	assert_pch_transcoder_disabled(dev_priv, TRANSCODER_A);
 
 	lpt_program_iclkip(crtc);
 
 	/* Set transcoder timing. */
-	ironlake_pch_transcoder_set_timings(intel_crtc, PIPE_A);
+	ironlake_pch_transcoder_set_timings(crtc, PIPE_A);
 
 	lpt_enable_pch_transcoder(dev_priv, cpu_transcoder);
 }
@@ -5026,8 +5004,6 @@ static void intel_post_plane_update(struct intel_crtc_state *old_crtc_state)
 
 	intel_frontbuffer_flip(to_i915(crtc->base.dev), pipe_config->fb_bits);
 
-	crtc->wm.cxsr_allowed = true;
-
 	if (pipe_config->update_wm_post && pipe_config->base.active)
 		intel_update_watermarks(crtc);
 
@@ -5074,22 +5050,18 @@ static void intel_pre_plane_update(struct intel_crtc_state *old_crtc_state)
 			intel_pre_disable_primary(&crtc->base);
 	}
 
-	if (pipe_config->disable_cxsr && HAS_GMCH_DISPLAY(dev_priv)) {
-		crtc->wm.cxsr_allowed = false;
-
-		/*
-		 * Vblank time updates from the shadow to live plane control register
-		 * are blocked if the memory self-refresh mode is active at that
-		 * moment. So to make sure the plane gets truly disabled, disable
-		 * first the self-refresh mode. The self-refresh enable bit in turn
-		 * will be checked/applied by the HW only at the next frame start
-		 * event which is after the vblank start event, so we need to have a
-		 * wait-for-vblank between disabling the plane and the pipe.
-		 */
-		if (old_crtc_state->base.active &&
-		    intel_set_memory_cxsr(dev_priv, false))
-			intel_wait_for_vblank(dev_priv, crtc->pipe);
-	}
+	/*
+	 * Vblank time updates from the shadow to live plane control register
+	 * are blocked if the memory self-refresh mode is active at that
+	 * moment. So to make sure the plane gets truly disabled, disable
+	 * first the self-refresh mode. The self-refresh enable bit in turn
+	 * will be checked/applied by the HW only at the next frame start
+	 * event which is after the vblank start event, so we need to have a
+	 * wait-for-vblank between disabling the plane and the pipe.
+	 */
+	if (HAS_GMCH_DISPLAY(dev_priv) && old_crtc_state->base.active &&
+	    pipe_config->disable_cxsr && intel_set_memory_cxsr(dev_priv, false))
+		intel_wait_for_vblank(dev_priv, crtc->pipe);
 
 	/*
 	 * IVB workaround: must disable low power watermarks for at least
@@ -5344,7 +5316,7 @@ static void ironlake_crtc_enable(struct intel_crtc_state *pipe_config,
 	intel_enable_pipe(intel_crtc);
 
 	if (intel_crtc->config->has_pch_encoder)
-		ironlake_pch_enable(crtc);
+		ironlake_pch_enable(pipe_config);
 
 	assert_vblank_disabled(crtc);
 	drm_crtc_vblank_on(crtc);
@@ -5426,10 +5398,10 @@ static void haswell_crtc_enable(struct intel_crtc_state *pipe_config,
 	intel_encoders_pre_enable(crtc, pipe_config, old_state);
 
 	if (intel_crtc->config->has_pch_encoder)
-		dev_priv->display.fdi_link_train(crtc);
+		dev_priv->display.fdi_link_train(intel_crtc, pipe_config);
 
 	if (!transcoder_is_dsi(cpu_transcoder))
-		intel_ddi_enable_pipe_clock(intel_crtc);
+		intel_ddi_enable_pipe_clock(pipe_config);
 
 	if (INTEL_GEN(dev_priv) >= 9)
 		skylake_pfit_enable(intel_crtc);
@@ -5442,9 +5414,9 @@ static void haswell_crtc_enable(struct intel_crtc_state *pipe_config,
 	 */
 	intel_color_load_luts(&pipe_config->base);
 
-	intel_ddi_set_pipe_settings(crtc);
+	intel_ddi_set_pipe_settings(pipe_config);
 	if (!transcoder_is_dsi(cpu_transcoder))
-		intel_ddi_enable_transcoder_func(crtc);
+		intel_ddi_enable_transcoder_func(pipe_config);
 
 	if (dev_priv->display.initial_watermarks != NULL)
 		dev_priv->display.initial_watermarks(old_intel_state, pipe_config);
@@ -5454,10 +5426,10 @@ static void haswell_crtc_enable(struct intel_crtc_state *pipe_config,
 		intel_enable_pipe(intel_crtc);
 
 	if (intel_crtc->config->has_pch_encoder)
-		lpt_pch_enable(crtc);
+		lpt_pch_enable(pipe_config);
 
 	if (intel_crtc_has_type(intel_crtc->config, INTEL_OUTPUT_DP_MST))
-		intel_ddi_set_vc_payload_alloc(crtc, true);
+		intel_ddi_set_vc_payload_alloc(pipe_config, true);
 
 	assert_vblank_disabled(crtc);
 	drm_crtc_vblank_on(crtc);
@@ -5579,7 +5551,7 @@ static void haswell_crtc_disable(struct intel_crtc_state *old_crtc_state,
 		intel_disable_pipe(intel_crtc);
 
 	if (intel_crtc_has_type(intel_crtc->config, INTEL_OUTPUT_DP_MST))
-		intel_ddi_set_vc_payload_alloc(crtc, false);
+		intel_ddi_set_vc_payload_alloc(intel_crtc->config, false);
 
 	if (!transcoder_is_dsi(cpu_transcoder))
 		intel_ddi_disable_transcoder_func(dev_priv, cpu_transcoder);
@@ -5590,7 +5562,7 @@ static void haswell_crtc_disable(struct intel_crtc_state *old_crtc_state,
 		ironlake_pfit_disable(intel_crtc, false);
 
 	if (!transcoder_is_dsi(cpu_transcoder))
-		intel_ddi_disable_pipe_clock(intel_crtc);
+		intel_ddi_disable_pipe_clock(intel_crtc->config);
 
 	intel_encoders_post_disable(crtc, old_crtc_state, old_state);
 
@@ -5623,7 +5595,7 @@ static void i9xx_pfit_enable(struct intel_crtc *crtc)
 	I915_WRITE(BCLRPAT(crtc->pipe), 0);
 }
 
-static enum intel_display_power_domain port_to_power_domain(enum port port)
+enum intel_display_power_domain intel_port_to_power_domain(enum port port)
 {
 	switch (port) {
 	case PORT_A:
@@ -5642,91 +5614,15 @@ static enum intel_display_power_domain port_to_power_domain(enum port port)
 	}
 }
 
-static enum intel_display_power_domain port_to_aux_power_domain(enum port port)
-{
-	switch (port) {
-	case PORT_A:
-		return POWER_DOMAIN_AUX_A;
-	case PORT_B:
-		return POWER_DOMAIN_AUX_B;
-	case PORT_C:
-		return POWER_DOMAIN_AUX_C;
-	case PORT_D:
-		return POWER_DOMAIN_AUX_D;
-	case PORT_E:
-		/* FIXME: Check VBT for actual wiring of PORT E */
-		return POWER_DOMAIN_AUX_D;
-	default:
-		MISSING_CASE(port);
-		return POWER_DOMAIN_AUX_A;
-	}
-}
-
-enum intel_display_power_domain
-intel_display_port_power_domain(struct intel_encoder *intel_encoder)
-{
-	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
-	struct intel_digital_port *intel_dig_port;
-
-	switch (intel_encoder->type) {
-	case INTEL_OUTPUT_UNKNOWN:
-		/* Only DDI platforms should ever use this output type */
-		WARN_ON_ONCE(!HAS_DDI(dev_priv));
-	case INTEL_OUTPUT_DP:
-	case INTEL_OUTPUT_HDMI:
-	case INTEL_OUTPUT_EDP:
-		intel_dig_port = enc_to_dig_port(&intel_encoder->base);
-		return port_to_power_domain(intel_dig_port->port);
-	case INTEL_OUTPUT_DP_MST:
-		intel_dig_port = enc_to_mst(&intel_encoder->base)->primary;
-		return port_to_power_domain(intel_dig_port->port);
-	case INTEL_OUTPUT_ANALOG:
-		return POWER_DOMAIN_PORT_CRT;
-	case INTEL_OUTPUT_DSI:
-		return POWER_DOMAIN_PORT_DSI;
-	default:
-		return POWER_DOMAIN_PORT_OTHER;
-	}
-}
-
-enum intel_display_power_domain
-intel_display_port_aux_power_domain(struct intel_encoder *intel_encoder)
-{
-	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
-	struct intel_digital_port *intel_dig_port;
-
-	switch (intel_encoder->type) {
-	case INTEL_OUTPUT_UNKNOWN:
-	case INTEL_OUTPUT_HDMI:
-		/*
-		 * Only DDI platforms should ever use these output types.
-		 * We can get here after the HDMI detect code has already set
-		 * the type of the shared encoder. Since we can't be sure
-		 * what's the status of the given connectors, play safe and
-		 * run the DP detection too.
-		 */
-		WARN_ON_ONCE(!HAS_DDI(dev_priv));
-	case INTEL_OUTPUT_DP:
-	case INTEL_OUTPUT_EDP:
-		intel_dig_port = enc_to_dig_port(&intel_encoder->base);
-		return port_to_aux_power_domain(intel_dig_port->port);
-	case INTEL_OUTPUT_DP_MST:
-		intel_dig_port = enc_to_mst(&intel_encoder->base)->primary;
-		return port_to_aux_power_domain(intel_dig_port->port);
-	default:
-		MISSING_CASE(intel_encoder->type);
-		return POWER_DOMAIN_AUX_A;
-	}
-}
-
-static unsigned long get_crtc_power_domains(struct drm_crtc *crtc,
-					    struct intel_crtc_state *crtc_state)
+static u64 get_crtc_power_domains(struct drm_crtc *crtc,
+				  struct intel_crtc_state *crtc_state)
 {
 	struct drm_device *dev = crtc->dev;
+	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct drm_encoder *encoder;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum pipe pipe = intel_crtc->pipe;
-	unsigned long mask;
+	u64 mask;
 	enum transcoder transcoder = crtc_state->cpu_transcoder;
 
 	if (!crtc_state->base.active)
@@ -5736,28 +5632,31 @@ static unsigned long get_crtc_power_domains(struct drm_crtc *crtc,
 	mask |= BIT(POWER_DOMAIN_TRANSCODER(transcoder));
 	if (crtc_state->pch_pfit.enabled ||
 	    crtc_state->pch_pfit.force_thru)
-		mask |= BIT(POWER_DOMAIN_PIPE_PANEL_FITTER(pipe));
+		mask |= BIT_ULL(POWER_DOMAIN_PIPE_PANEL_FITTER(pipe));
 
 	drm_for_each_encoder_mask(encoder, dev, crtc_state->base.encoder_mask) {
 		struct intel_encoder *intel_encoder = to_intel_encoder(encoder);
 
-		mask |= BIT(intel_display_port_power_domain(intel_encoder));
+		mask |= BIT_ULL(intel_encoder->power_domain);
 	}
 
+	if (HAS_DDI(dev_priv) && crtc_state->has_audio)
+		mask |= BIT(POWER_DOMAIN_AUDIO);
+
 	if (crtc_state->shared_dpll)
-		mask |= BIT(POWER_DOMAIN_PLLS);
+		mask |= BIT_ULL(POWER_DOMAIN_PLLS);
 
 	return mask;
 }
 
-static unsigned long
+static u64
 modeset_get_crtc_power_domains(struct drm_crtc *crtc,
 			       struct intel_crtc_state *crtc_state)
 {
 	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum intel_display_power_domain domain;
-	unsigned long domains, new_domains, old_domains;
+	u64 domains, new_domains, old_domains;
 
 	old_domains = intel_crtc->enabled_power_domains;
 	intel_crtc->enabled_power_domains = new_domains =
@@ -5772,7 +5671,7 @@ modeset_get_crtc_power_domains(struct drm_crtc *crtc,
 }
 
 static void modeset_put_power_domains(struct drm_i915_private *dev_priv,
-				      unsigned long domains)
+				      u64 domains)
 {
 	enum intel_display_power_domain domain;
 
@@ -5780,922 +5679,11 @@ static void modeset_put_power_domains(struct drm_i915_private *dev_priv,
 		intel_display_power_put(dev_priv, domain);
 }
 
-static int intel_compute_max_dotclk(struct drm_i915_private *dev_priv)
-{
-	int max_cdclk_freq = dev_priv->max_cdclk_freq;
-
-	if (IS_GEMINILAKE(dev_priv))
-		return 2 * max_cdclk_freq;
-	else if (INTEL_INFO(dev_priv)->gen >= 9 ||
-		 IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
-		return max_cdclk_freq;
-	else if (IS_CHERRYVIEW(dev_priv))
-		return max_cdclk_freq*95/100;
-	else if (INTEL_INFO(dev_priv)->gen < 4)
-		return 2*max_cdclk_freq*90/100;
-	else
-		return max_cdclk_freq*90/100;
-}
-
-static int skl_calc_cdclk(int max_pixclk, int vco);
-
-static void intel_update_max_cdclk(struct drm_i915_private *dev_priv)
-{
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
-		u32 limit = I915_READ(SKL_DFSM) & SKL_DFSM_CDCLK_LIMIT_MASK;
-		int max_cdclk, vco;
-
-		vco = dev_priv->skl_preferred_vco_freq;
-		WARN_ON(vco != 8100000 && vco != 8640000);
-
-		/*
-		 * Use the lower (vco 8640) cdclk values as a
-		 * first guess. skl_calc_cdclk() will correct it
-		 * if the preferred vco is 8100 instead.
-		 */
-		if (limit == SKL_DFSM_CDCLK_LIMIT_675)
-			max_cdclk = 617143;
-		else if (limit == SKL_DFSM_CDCLK_LIMIT_540)
-			max_cdclk = 540000;
-		else if (limit == SKL_DFSM_CDCLK_LIMIT_450)
-			max_cdclk = 432000;
-		else
-			max_cdclk = 308571;
-
-		dev_priv->max_cdclk_freq = skl_calc_cdclk(max_cdclk, vco);
-	} else if (IS_GEMINILAKE(dev_priv)) {
-		dev_priv->max_cdclk_freq = 316800;
-	} else if (IS_BROXTON(dev_priv)) {
-		dev_priv->max_cdclk_freq = 624000;
-	} else if (IS_BROADWELL(dev_priv))  {
-		/*
-		 * FIXME with extra cooling we can allow
-		 * 540 MHz for ULX and 675 Mhz for ULT.
-		 * How can we know if extra cooling is
-		 * available? PCI ID, VTB, something else?
-		 */
-		if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
-			dev_priv->max_cdclk_freq = 450000;
-		else if (IS_BDW_ULX(dev_priv))
-			dev_priv->max_cdclk_freq = 450000;
-		else if (IS_BDW_ULT(dev_priv))
-			dev_priv->max_cdclk_freq = 540000;
-		else
-			dev_priv->max_cdclk_freq = 675000;
-	} else if (IS_CHERRYVIEW(dev_priv)) {
-		dev_priv->max_cdclk_freq = 320000;
-	} else if (IS_VALLEYVIEW(dev_priv)) {
-		dev_priv->max_cdclk_freq = 400000;
-	} else {
-		/* otherwise assume cdclk is fixed */
-		dev_priv->max_cdclk_freq = dev_priv->cdclk_freq;
-	}
-
-	dev_priv->max_dotclk_freq = intel_compute_max_dotclk(dev_priv);
-
-	DRM_DEBUG_DRIVER("Max CD clock rate: %d kHz\n",
-			 dev_priv->max_cdclk_freq);
-
-	DRM_DEBUG_DRIVER("Max dotclock rate: %d kHz\n",
-			 dev_priv->max_dotclk_freq);
-}
-
-static void intel_update_cdclk(struct drm_i915_private *dev_priv)
-{
-	dev_priv->cdclk_freq = dev_priv->display.get_display_clock_speed(dev_priv);
-
-	if (INTEL_GEN(dev_priv) >= 9)
-		DRM_DEBUG_DRIVER("Current CD clock rate: %d kHz, VCO: %d kHz, ref: %d kHz\n",
-				 dev_priv->cdclk_freq, dev_priv->cdclk_pll.vco,
-				 dev_priv->cdclk_pll.ref);
-	else
-		DRM_DEBUG_DRIVER("Current CD clock rate: %d kHz\n",
-				 dev_priv->cdclk_freq);
-
-	/*
-	 * 9:0 CMBUS [sic] CDCLK frequency (cdfreq):
-	 * Programmng [sic] note: bit[9:2] should be programmed to the number
-	 * of cdclk that generates 4MHz reference clock freq which is used to
-	 * generate GMBus clock. This will vary with the cdclk freq.
-	 */
-	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-		I915_WRITE(GMBUSFREQ_VLV, DIV_ROUND_UP(dev_priv->cdclk_freq, 1000));
-}
-
-/* convert from kHz to .1 fixpoint MHz with -1MHz offset */
-static int skl_cdclk_decimal(int cdclk)
-{
-	return DIV_ROUND_CLOSEST(cdclk - 1000, 500);
-}
-
-static int bxt_de_pll_vco(struct drm_i915_private *dev_priv, int cdclk)
-{
-	int ratio;
-
-	if (cdclk == dev_priv->cdclk_pll.ref)
-		return 0;
-
-	switch (cdclk) {
-	default:
-		MISSING_CASE(cdclk);
-	case 144000:
-	case 288000:
-	case 384000:
-	case 576000:
-		ratio = 60;
-		break;
-	case 624000:
-		ratio = 65;
-		break;
-	}
-
-	return dev_priv->cdclk_pll.ref * ratio;
-}
-
-static int glk_de_pll_vco(struct drm_i915_private *dev_priv, int cdclk)
-{
-	int ratio;
-
-	if (cdclk == dev_priv->cdclk_pll.ref)
-		return 0;
-
-	switch (cdclk) {
-	default:
-		MISSING_CASE(cdclk);
-	case  79200:
-	case 158400:
-	case 316800:
-		ratio = 33;
-		break;
-	}
-
-	return dev_priv->cdclk_pll.ref * ratio;
-}
-
-static void bxt_de_pll_disable(struct drm_i915_private *dev_priv)
-{
-	I915_WRITE(BXT_DE_PLL_ENABLE, 0);
-
-	/* Timeout 200us */
-	if (intel_wait_for_register(dev_priv,
-				    BXT_DE_PLL_ENABLE, BXT_DE_PLL_LOCK, 0,
-				    1))
-		DRM_ERROR("timeout waiting for DE PLL unlock\n");
-
-	dev_priv->cdclk_pll.vco = 0;
-}
-
-static void bxt_de_pll_enable(struct drm_i915_private *dev_priv, int vco)
-{
-	int ratio = DIV_ROUND_CLOSEST(vco, dev_priv->cdclk_pll.ref);
-	u32 val;
-
-	val = I915_READ(BXT_DE_PLL_CTL);
-	val &= ~BXT_DE_PLL_RATIO_MASK;
-	val |= BXT_DE_PLL_RATIO(ratio);
-	I915_WRITE(BXT_DE_PLL_CTL, val);
-
-	I915_WRITE(BXT_DE_PLL_ENABLE, BXT_DE_PLL_PLL_ENABLE);
-
-	/* Timeout 200us */
-	if (intel_wait_for_register(dev_priv,
-				    BXT_DE_PLL_ENABLE,
-				    BXT_DE_PLL_LOCK,
-				    BXT_DE_PLL_LOCK,
-				    1))
-		DRM_ERROR("timeout waiting for DE PLL lock\n");
-
-	dev_priv->cdclk_pll.vco = vco;
-}
-
-static void bxt_set_cdclk(struct drm_i915_private *dev_priv, int cdclk)
-{
-	u32 val, divider;
-	int vco, ret;
-
-	if (IS_GEMINILAKE(dev_priv))
-		vco = glk_de_pll_vco(dev_priv, cdclk);
-	else
-		vco = bxt_de_pll_vco(dev_priv, cdclk);
-
-	DRM_DEBUG_DRIVER("Changing CDCLK to %d kHz (VCO %d kHz)\n", cdclk, vco);
-
-	/* cdclk = vco / 2 / div{1,1.5,2,4} */
-	switch (DIV_ROUND_CLOSEST(vco, cdclk)) {
-	case 8:
-		divider = BXT_CDCLK_CD2X_DIV_SEL_4;
-		break;
-	case 4:
-		divider = BXT_CDCLK_CD2X_DIV_SEL_2;
-		break;
-	case 3:
-		WARN(IS_GEMINILAKE(dev_priv), "Unsupported divider\n");
-		divider = BXT_CDCLK_CD2X_DIV_SEL_1_5;
-		break;
-	case 2:
-		divider = BXT_CDCLK_CD2X_DIV_SEL_1;
-		break;
-	default:
-		WARN_ON(cdclk != dev_priv->cdclk_pll.ref);
-		WARN_ON(vco != 0);
-
-		divider = BXT_CDCLK_CD2X_DIV_SEL_1;
-		break;
-	}
-
-	/* Inform power controller of upcoming frequency change */
-	mutex_lock(&dev_priv->rps.hw_lock);
-	ret = sandybridge_pcode_write(dev_priv, HSW_PCODE_DE_WRITE_FREQ_REQ,
-				      0x80000000);
-	mutex_unlock(&dev_priv->rps.hw_lock);
-
-	if (ret) {
-		DRM_ERROR("PCode CDCLK freq change notify failed (err %d, freq %d)\n",
-			  ret, cdclk);
-		return;
-	}
-
-	if (dev_priv->cdclk_pll.vco != 0 &&
-	    dev_priv->cdclk_pll.vco != vco)
-		bxt_de_pll_disable(dev_priv);
-
-	if (dev_priv->cdclk_pll.vco != vco)
-		bxt_de_pll_enable(dev_priv, vco);
-
-	val = divider | skl_cdclk_decimal(cdclk);
-	/*
-	 * FIXME if only the cd2x divider needs changing, it could be done
-	 * without shutting off the pipe (if only one pipe is active).
-	 */
-	val |= BXT_CDCLK_CD2X_PIPE_NONE;
-	/*
-	 * Disable SSA Precharge when CD clock frequency < 500 MHz,
-	 * enable otherwise.
-	 */
-	if (cdclk >= 500000)
-		val |= BXT_CDCLK_SSA_PRECHARGE_ENABLE;
-	I915_WRITE(CDCLK_CTL, val);
-
-	mutex_lock(&dev_priv->rps.hw_lock);
-	ret = sandybridge_pcode_write(dev_priv, HSW_PCODE_DE_WRITE_FREQ_REQ,
-				      DIV_ROUND_UP(cdclk, 25000));
-	mutex_unlock(&dev_priv->rps.hw_lock);
-
-	if (ret) {
-		DRM_ERROR("PCode CDCLK freq set failed, (err %d, freq %d)\n",
-			  ret, cdclk);
-		return;
-	}
-
-	intel_update_cdclk(dev_priv);
-}
-
-static void bxt_sanitize_cdclk(struct drm_i915_private *dev_priv)
-{
-	u32 cdctl, expected;
-
-	intel_update_cdclk(dev_priv);
-
-	if (dev_priv->cdclk_pll.vco == 0 ||
-	    dev_priv->cdclk_freq == dev_priv->cdclk_pll.ref)
-		goto sanitize;
-
-	/* DPLL okay; verify the cdclock
-	 *
-	 * Some BIOS versions leave an incorrect decimal frequency value and
-	 * set reserved MBZ bits in CDCLK_CTL at least during exiting from S4,
-	 * so sanitize this register.
-	 */
-	cdctl = I915_READ(CDCLK_CTL);
-	/*
-	 * Let's ignore the pipe field, since BIOS could have configured the
-	 * dividers both synching to an active pipe, or asynchronously
-	 * (PIPE_NONE).
-	 */
-	cdctl &= ~BXT_CDCLK_CD2X_PIPE_NONE;
-
-	expected = (cdctl & BXT_CDCLK_CD2X_DIV_SEL_MASK) |
-		   skl_cdclk_decimal(dev_priv->cdclk_freq);
-	/*
-	 * Disable SSA Precharge when CD clock frequency < 500 MHz,
-	 * enable otherwise.
-	 */
-	if (dev_priv->cdclk_freq >= 500000)
-		expected |= BXT_CDCLK_SSA_PRECHARGE_ENABLE;
-
-	if (cdctl == expected)
-		/* All well; nothing to sanitize */
-		return;
-
-sanitize:
-	DRM_DEBUG_KMS("Sanitizing cdclk programmed by pre-os\n");
-
-	/* force cdclk programming */
-	dev_priv->cdclk_freq = 0;
-
-	/* force full PLL disable + enable */
-	dev_priv->cdclk_pll.vco = -1;
-}
-
-void bxt_init_cdclk(struct drm_i915_private *dev_priv)
-{
-	int cdclk;
-
-	bxt_sanitize_cdclk(dev_priv);
-
-	if (dev_priv->cdclk_freq != 0 && dev_priv->cdclk_pll.vco != 0)
-		return;
-
-	/*
-	 * FIXME:
-	 * - The initial CDCLK needs to be read from VBT.
-	 *   Need to make this change after VBT has changes for BXT.
-	 */
-	if (IS_GEMINILAKE(dev_priv))
-		cdclk = glk_calc_cdclk(0);
-	else
-		cdclk = bxt_calc_cdclk(0);
-
-	bxt_set_cdclk(dev_priv, cdclk);
-}
-
-void bxt_uninit_cdclk(struct drm_i915_private *dev_priv)
-{
-	bxt_set_cdclk(dev_priv, dev_priv->cdclk_pll.ref);
-}
-
-static int skl_calc_cdclk(int max_pixclk, int vco)
-{
-	if (vco == 8640000) {
-		if (max_pixclk > 540000)
-			return 617143;
-		else if (max_pixclk > 432000)
-			return 540000;
-		else if (max_pixclk > 308571)
-			return 432000;
-		else
-			return 308571;
-	} else {
-		if (max_pixclk > 540000)
-			return 675000;
-		else if (max_pixclk > 450000)
-			return 540000;
-		else if (max_pixclk > 337500)
-			return 450000;
-		else
-			return 337500;
-	}
-}
-
-static void
-skl_dpll0_update(struct drm_i915_private *dev_priv)
-{
-	u32 val;
-
-	dev_priv->cdclk_pll.ref = 24000;
-	dev_priv->cdclk_pll.vco = 0;
-
-	val = I915_READ(LCPLL1_CTL);
-	if ((val & LCPLL_PLL_ENABLE) == 0)
-		return;
-
-	if (WARN_ON((val & LCPLL_PLL_LOCK) == 0))
-		return;
-
-	val = I915_READ(DPLL_CTRL1);
-
-	if (WARN_ON((val & (DPLL_CTRL1_HDMI_MODE(SKL_DPLL0) |
-			    DPLL_CTRL1_SSC(SKL_DPLL0) |
-			    DPLL_CTRL1_OVERRIDE(SKL_DPLL0))) !=
-		    DPLL_CTRL1_OVERRIDE(SKL_DPLL0)))
-		return;
-
-	switch (val & DPLL_CTRL1_LINK_RATE_MASK(SKL_DPLL0)) {
-	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_810, SKL_DPLL0):
-	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1350, SKL_DPLL0):
-	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1620, SKL_DPLL0):
-	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_2700, SKL_DPLL0):
-		dev_priv->cdclk_pll.vco = 8100000;
-		break;
-	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1080, SKL_DPLL0):
-	case DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_2160, SKL_DPLL0):
-		dev_priv->cdclk_pll.vco = 8640000;
-		break;
-	default:
-		MISSING_CASE(val & DPLL_CTRL1_LINK_RATE_MASK(SKL_DPLL0));
-		break;
-	}
-}
-
-void skl_set_preferred_cdclk_vco(struct drm_i915_private *dev_priv, int vco)
-{
-	bool changed = dev_priv->skl_preferred_vco_freq != vco;
-
-	dev_priv->skl_preferred_vco_freq = vco;
-
-	if (changed)
-		intel_update_max_cdclk(dev_priv);
-}
-
-static void
-skl_dpll0_enable(struct drm_i915_private *dev_priv, int vco)
-{
-	int min_cdclk = skl_calc_cdclk(0, vco);
-	u32 val;
-
-	WARN_ON(vco != 8100000 && vco != 8640000);
-
-	/* select the minimum CDCLK before enabling DPLL 0 */
-	val = CDCLK_FREQ_337_308 | skl_cdclk_decimal(min_cdclk);
-	I915_WRITE(CDCLK_CTL, val);
-	POSTING_READ(CDCLK_CTL);
-
-	/*
-	 * We always enable DPLL0 with the lowest link rate possible, but still
-	 * taking into account the VCO required to operate the eDP panel at the
-	 * desired frequency. The usual DP link rates operate with a VCO of
-	 * 8100 while the eDP 1.4 alternate link rates need a VCO of 8640.
-	 * The modeset code is responsible for the selection of the exact link
-	 * rate later on, with the constraint of choosing a frequency that
-	 * works with vco.
-	 */
-	val = I915_READ(DPLL_CTRL1);
-
-	val &= ~(DPLL_CTRL1_HDMI_MODE(SKL_DPLL0) | DPLL_CTRL1_SSC(SKL_DPLL0) |
-		 DPLL_CTRL1_LINK_RATE_MASK(SKL_DPLL0));
-	val |= DPLL_CTRL1_OVERRIDE(SKL_DPLL0);
-	if (vco == 8640000)
-		val |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_1080,
-					    SKL_DPLL0);
-	else
-		val |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_810,
-					    SKL_DPLL0);
-
-	I915_WRITE(DPLL_CTRL1, val);
-	POSTING_READ(DPLL_CTRL1);
-
-	I915_WRITE(LCPLL1_CTL, I915_READ(LCPLL1_CTL) | LCPLL_PLL_ENABLE);
-
-	if (intel_wait_for_register(dev_priv,
-				    LCPLL1_CTL, LCPLL_PLL_LOCK, LCPLL_PLL_LOCK,
-				    5))
-		DRM_ERROR("DPLL0 not locked\n");
-
-	dev_priv->cdclk_pll.vco = vco;
-
-	/* We'll want to keep using the current vco from now on. */
-	skl_set_preferred_cdclk_vco(dev_priv, vco);
-}
-
-static void
-skl_dpll0_disable(struct drm_i915_private *dev_priv)
-{
-	I915_WRITE(LCPLL1_CTL, I915_READ(LCPLL1_CTL) & ~LCPLL_PLL_ENABLE);
-	if (intel_wait_for_register(dev_priv,
-				   LCPLL1_CTL, LCPLL_PLL_LOCK, 0,
-				   1))
-		DRM_ERROR("Couldn't disable DPLL0\n");
-
-	dev_priv->cdclk_pll.vco = 0;
-}
-
-static void skl_set_cdclk(struct drm_i915_private *dev_priv, int cdclk, int vco)
-{
-	u32 freq_select, pcu_ack;
-	int ret;
-
-	WARN_ON((cdclk == 24000) != (vco == 0));
-
-	DRM_DEBUG_DRIVER("Changing CDCLK to %d kHz (VCO %d kHz)\n", cdclk, vco);
-
-	mutex_lock(&dev_priv->rps.hw_lock);
-	ret = skl_pcode_request(dev_priv, SKL_PCODE_CDCLK_CONTROL,
-				SKL_CDCLK_PREPARE_FOR_CHANGE,
-				SKL_CDCLK_READY_FOR_CHANGE,
-				SKL_CDCLK_READY_FOR_CHANGE, 3);
-	mutex_unlock(&dev_priv->rps.hw_lock);
-	if (ret) {
-		DRM_ERROR("Failed to inform PCU about cdclk change (%d)\n",
-			  ret);
-		return;
-	}
-
-	/* set CDCLK_CTL */
-	switch (cdclk) {
-	case 450000:
-	case 432000:
-		freq_select = CDCLK_FREQ_450_432;
-		pcu_ack = 1;
-		break;
-	case 540000:
-		freq_select = CDCLK_FREQ_540;
-		pcu_ack = 2;
-		break;
-	case 308571:
-	case 337500:
-	default:
-		freq_select = CDCLK_FREQ_337_308;
-		pcu_ack = 0;
-		break;
-	case 617143:
-	case 675000:
-		freq_select = CDCLK_FREQ_675_617;
-		pcu_ack = 3;
-		break;
-	}
-
-	if (dev_priv->cdclk_pll.vco != 0 &&
-	    dev_priv->cdclk_pll.vco != vco)
-		skl_dpll0_disable(dev_priv);
-
-	if (dev_priv->cdclk_pll.vco != vco)
-		skl_dpll0_enable(dev_priv, vco);
-
-	I915_WRITE(CDCLK_CTL, freq_select | skl_cdclk_decimal(cdclk));
-	POSTING_READ(CDCLK_CTL);
-
-	/* inform PCU of the change */
-	mutex_lock(&dev_priv->rps.hw_lock);
-	sandybridge_pcode_write(dev_priv, SKL_PCODE_CDCLK_CONTROL, pcu_ack);
-	mutex_unlock(&dev_priv->rps.hw_lock);
-
-	intel_update_cdclk(dev_priv);
-}
-
-static void skl_sanitize_cdclk(struct drm_i915_private *dev_priv);
-
-void skl_uninit_cdclk(struct drm_i915_private *dev_priv)
-{
-	skl_set_cdclk(dev_priv, dev_priv->cdclk_pll.ref, 0);
-}
-
-void skl_init_cdclk(struct drm_i915_private *dev_priv)
-{
-	int cdclk, vco;
-
-	skl_sanitize_cdclk(dev_priv);
-
-	if (dev_priv->cdclk_freq != 0 && dev_priv->cdclk_pll.vco != 0) {
-		/*
-		 * Use the current vco as our initial
-		 * guess as to what the preferred vco is.
-		 */
-		if (dev_priv->skl_preferred_vco_freq == 0)
-			skl_set_preferred_cdclk_vco(dev_priv,
-						    dev_priv->cdclk_pll.vco);
-		return;
-	}
-
-	vco = dev_priv->skl_preferred_vco_freq;
-	if (vco == 0)
-		vco = 8100000;
-	cdclk = skl_calc_cdclk(0, vco);
-
-	skl_set_cdclk(dev_priv, cdclk, vco);
-}
-
-static void skl_sanitize_cdclk(struct drm_i915_private *dev_priv)
-{
-	uint32_t cdctl, expected;
-
-	/*
-	 * check if the pre-os intialized the display
-	 * There is SWF18 scratchpad register defined which is set by the
-	 * pre-os which can be used by the OS drivers to check the status
-	 */
-	if ((I915_READ(SWF_ILK(0x18)) & 0x00FFFFFF) == 0)
-		goto sanitize;
-
-	intel_update_cdclk(dev_priv);
-	/* Is PLL enabled and locked ? */
-	if (dev_priv->cdclk_pll.vco == 0 ||
-	    dev_priv->cdclk_freq == dev_priv->cdclk_pll.ref)
-		goto sanitize;
-
-	/* DPLL okay; verify the cdclock
-	 *
-	 * Noticed in some instances that the freq selection is correct but
-	 * decimal part is programmed wrong from BIOS where pre-os does not
-	 * enable display. Verify the same as well.
-	 */
-	cdctl = I915_READ(CDCLK_CTL);
-	expected = (cdctl & CDCLK_FREQ_SEL_MASK) |
-		skl_cdclk_decimal(dev_priv->cdclk_freq);
-	if (cdctl == expected)
-		/* All well; nothing to sanitize */
-		return;
-
-sanitize:
-	DRM_DEBUG_KMS("Sanitizing cdclk programmed by pre-os\n");
-
-	/* force cdclk programming */
-	dev_priv->cdclk_freq = 0;
-	/* force full PLL disable + enable */
-	dev_priv->cdclk_pll.vco = -1;
-}
-
-/* Adjust CDclk dividers to allow high res or save power if possible */
-static void valleyview_set_cdclk(struct drm_device *dev, int cdclk)
-{
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	u32 val, cmd;
-
-	WARN_ON(dev_priv->display.get_display_clock_speed(dev_priv)
-					!= dev_priv->cdclk_freq);
-
-	if (cdclk >= 320000) /* jump to highest voltage for 400MHz too */
-		cmd = 2;
-	else if (cdclk == 266667)
-		cmd = 1;
-	else
-		cmd = 0;
-
-	mutex_lock(&dev_priv->rps.hw_lock);
-	val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
-	val &= ~DSPFREQGUAR_MASK;
-	val |= (cmd << DSPFREQGUAR_SHIFT);
-	vlv_punit_write(dev_priv, PUNIT_REG_DSPFREQ, val);
-	if (wait_for((vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) &
-		      DSPFREQSTAT_MASK) == (cmd << DSPFREQSTAT_SHIFT),
-		     50)) {
-		DRM_ERROR("timed out waiting for CDclk change\n");
-	}
-	mutex_unlock(&dev_priv->rps.hw_lock);
-
-	mutex_lock(&dev_priv->sb_lock);
-
-	if (cdclk == 400000) {
-		u32 divider;
-
-		divider = DIV_ROUND_CLOSEST(dev_priv->hpll_freq << 1, cdclk) - 1;
-
-		/* adjust cdclk divider */
-		val = vlv_cck_read(dev_priv, CCK_DISPLAY_CLOCK_CONTROL);
-		val &= ~CCK_FREQUENCY_VALUES;
-		val |= divider;
-		vlv_cck_write(dev_priv, CCK_DISPLAY_CLOCK_CONTROL, val);
-
-		if (wait_for((vlv_cck_read(dev_priv, CCK_DISPLAY_CLOCK_CONTROL) &
-			      CCK_FREQUENCY_STATUS) == (divider << CCK_FREQUENCY_STATUS_SHIFT),
-			     50))
-			DRM_ERROR("timed out waiting for CDclk change\n");
-	}
-
-	/* adjust self-refresh exit latency value */
-	val = vlv_bunit_read(dev_priv, BUNIT_REG_BISOC);
-	val &= ~0x7f;
-
-	/*
-	 * For high bandwidth configs, we set a higher latency in the bunit
-	 * so that the core display fetch happens in time to avoid underruns.
-	 */
-	if (cdclk == 400000)
-		val |= 4500 / 250; /* 4.5 usec */
-	else
-		val |= 3000 / 250; /* 3.0 usec */
-	vlv_bunit_write(dev_priv, BUNIT_REG_BISOC, val);
-
-	mutex_unlock(&dev_priv->sb_lock);
-
-	intel_update_cdclk(dev_priv);
-}
-
-static void cherryview_set_cdclk(struct drm_device *dev, int cdclk)
-{
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	u32 val, cmd;
-
-	WARN_ON(dev_priv->display.get_display_clock_speed(dev_priv)
-						!= dev_priv->cdclk_freq);
-
-	switch (cdclk) {
-	case 333333:
-	case 320000:
-	case 266667:
-	case 200000:
-		break;
-	default:
-		MISSING_CASE(cdclk);
-		return;
-	}
-
-	/*
-	 * Specs are full of misinformation, but testing on actual
-	 * hardware has shown that we just need to write the desired
-	 * CCK divider into the Punit register.
-	 */
-	cmd = DIV_ROUND_CLOSEST(dev_priv->hpll_freq << 1, cdclk) - 1;
-
-	mutex_lock(&dev_priv->rps.hw_lock);
-	val = vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ);
-	val &= ~DSPFREQGUAR_MASK_CHV;
-	val |= (cmd << DSPFREQGUAR_SHIFT_CHV);
-	vlv_punit_write(dev_priv, PUNIT_REG_DSPFREQ, val);
-	if (wait_for((vlv_punit_read(dev_priv, PUNIT_REG_DSPFREQ) &
-		      DSPFREQSTAT_MASK_CHV) == (cmd << DSPFREQSTAT_SHIFT_CHV),
-		     50)) {
-		DRM_ERROR("timed out waiting for CDclk change\n");
-	}
-	mutex_unlock(&dev_priv->rps.hw_lock);
-
-	intel_update_cdclk(dev_priv);
-}
-
-static int valleyview_calc_cdclk(struct drm_i915_private *dev_priv,
-				 int max_pixclk)
-{
-	int freq_320 = (dev_priv->hpll_freq <<  1) % 320000 != 0 ? 333333 : 320000;
-	int limit = IS_CHERRYVIEW(dev_priv) ? 95 : 90;
-
-	/*
-	 * Really only a few cases to deal with, as only 4 CDclks are supported:
-	 *   200MHz
-	 *   267MHz
-	 *   320/333MHz (depends on HPLL freq)
-	 *   400MHz (VLV only)
-	 * So we check to see whether we're above 90% (VLV) or 95% (CHV)
-	 * of the lower bin and adjust if needed.
-	 *
-	 * We seem to get an unstable or solid color picture at 200MHz.
-	 * Not sure what's wrong. For now use 200MHz only when all pipes
-	 * are off.
-	 */
-	if (!IS_CHERRYVIEW(dev_priv) &&
-	    max_pixclk > freq_320*limit/100)
-		return 400000;
-	else if (max_pixclk > 266667*limit/100)
-		return freq_320;
-	else if (max_pixclk > 0)
-		return 266667;
-	else
-		return 200000;
-}
-
-static int glk_calc_cdclk(int max_pixclk)
-{
-	if (max_pixclk > 2 * 158400)
-		return 316800;
-	else if (max_pixclk > 2 * 79200)
-		return 158400;
-	else
-		return 79200;
-}
-
-static int bxt_calc_cdclk(int max_pixclk)
-{
-	if (max_pixclk > 576000)
-		return 624000;
-	else if (max_pixclk > 384000)
-		return 576000;
-	else if (max_pixclk > 288000)
-		return 384000;
-	else if (max_pixclk > 144000)
-		return 288000;
-	else
-		return 144000;
-}
-
-/* Compute the max pixel clock for new configuration. */
-static int intel_mode_max_pixclk(struct drm_device *dev,
-				 struct drm_atomic_state *state)
-{
-	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct drm_crtc *crtc;
-	struct drm_crtc_state *crtc_state;
-	unsigned max_pixclk = 0, i;
-	enum pipe pipe;
-
-	memcpy(intel_state->min_pixclk, dev_priv->min_pixclk,
-	       sizeof(intel_state->min_pixclk));
-
-	for_each_crtc_in_state(state, crtc, crtc_state, i) {
-		int pixclk = 0;
-
-		if (crtc_state->enable)
-			pixclk = crtc_state->adjusted_mode.crtc_clock;
-
-		intel_state->min_pixclk[i] = pixclk;
-	}
-
-	for_each_pipe(dev_priv, pipe)
-		max_pixclk = max(intel_state->min_pixclk[pipe], max_pixclk);
-
-	return max_pixclk;
-}
-
-static int valleyview_modeset_calc_cdclk(struct drm_atomic_state *state)
-{
-	struct drm_device *dev = state->dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	int max_pixclk = intel_mode_max_pixclk(dev, state);
-	struct intel_atomic_state *intel_state =
-		to_intel_atomic_state(state);
-
-	intel_state->cdclk = intel_state->dev_cdclk =
-		valleyview_calc_cdclk(dev_priv, max_pixclk);
-
-	if (!intel_state->active_crtcs)
-		intel_state->dev_cdclk = valleyview_calc_cdclk(dev_priv, 0);
-
-	return 0;
-}
-
-static int bxt_modeset_calc_cdclk(struct drm_atomic_state *state)
-{
-	struct drm_i915_private *dev_priv = to_i915(state->dev);
-	int max_pixclk = ilk_max_pixel_rate(state);
-	struct intel_atomic_state *intel_state =
-		to_intel_atomic_state(state);
-	int cdclk;
-
-	if (IS_GEMINILAKE(dev_priv))
-		cdclk = glk_calc_cdclk(max_pixclk);
-	else
-		cdclk = bxt_calc_cdclk(max_pixclk);
-
-	intel_state->cdclk = intel_state->dev_cdclk = cdclk;
-
-	if (!intel_state->active_crtcs) {
-		if (IS_GEMINILAKE(dev_priv))
-			cdclk = glk_calc_cdclk(0);
-		else
-			cdclk = bxt_calc_cdclk(0);
-
-		intel_state->dev_cdclk = cdclk;
-	}
-
-	return 0;
-}
-
-static void vlv_program_pfi_credits(struct drm_i915_private *dev_priv)
-{
-	unsigned int credits, default_credits;
-
-	if (IS_CHERRYVIEW(dev_priv))
-		default_credits = PFI_CREDIT(12);
-	else
-		default_credits = PFI_CREDIT(8);
-
-	if (dev_priv->cdclk_freq >= dev_priv->czclk_freq) {
-		/* CHV suggested value is 31 or 63 */
-		if (IS_CHERRYVIEW(dev_priv))
-			credits = PFI_CREDIT_63;
-		else
-			credits = PFI_CREDIT(15);
-	} else {
-		credits = default_credits;
-	}
-
-	/*
-	 * WA - write default credits before re-programming
-	 * FIXME: should we also set the resend bit here?
-	 */
-	I915_WRITE(GCI_CONTROL, VGA_FAST_MODE_DISABLE |
-		   default_credits);
-
-	I915_WRITE(GCI_CONTROL, VGA_FAST_MODE_DISABLE |
-		   credits | PFI_CREDIT_RESEND);
-
-	/*
-	 * FIXME is this guaranteed to clear
-	 * immediately or should we poll for it?
-	 */
-	WARN_ON(I915_READ(GCI_CONTROL) & PFI_CREDIT_RESEND);
-}
-
-static void valleyview_modeset_commit_cdclk(struct drm_atomic_state *old_state)
-{
-	struct drm_device *dev = old_state->dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct intel_atomic_state *old_intel_state =
-		to_intel_atomic_state(old_state);
-	unsigned req_cdclk = old_intel_state->dev_cdclk;
-
-	/*
-	 * FIXME: We can end up here with all power domains off, yet
-	 * with a CDCLK frequency other than the minimum. To account
-	 * for this take the PIPE-A power domain, which covers the HW
-	 * blocks needed for the following programming. This can be
-	 * removed once it's guaranteed that we get here either with
-	 * the minimum CDCLK set, or the required power domains
-	 * enabled.
-	 */
-	intel_display_power_get(dev_priv, POWER_DOMAIN_PIPE_A);
-
-	if (IS_CHERRYVIEW(dev_priv))
-		cherryview_set_cdclk(dev, req_cdclk);
-	else
-		valleyview_set_cdclk(dev, req_cdclk);
-
-	vlv_program_pfi_credits(dev_priv);
-
-	intel_display_power_put(dev_priv, POWER_DOMAIN_PIPE_A);
-}
-
 static void valleyview_crtc_enable(struct intel_crtc_state *pipe_config,
 				   struct drm_atomic_state *old_state)
 {
+	struct intel_atomic_state *old_intel_state =
+		to_intel_atomic_state(old_state);
 	struct drm_crtc *crtc = pipe_config->base.crtc;
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
@@ -6740,7 +5728,8 @@ static void valleyview_crtc_enable(struct intel_crtc_state *pipe_config,
 
 	intel_color_load_luts(&pipe_config->base);
 
-	intel_update_watermarks(intel_crtc);
+	dev_priv->display.initial_watermarks(old_intel_state,
+					     pipe_config);
 	intel_enable_pipe(intel_crtc);
 
 	assert_vblank_disabled(crtc);
@@ -6857,6 +5846,9 @@ static void i9xx_crtc_disable(struct intel_crtc_state *old_crtc_state,
 
 	if (!IS_GEN2(dev_priv))
 		intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false);
+
+	if (!dev_priv->display.initial_watermarks)
+		intel_update_watermarks(intel_crtc);
 }
 
 static void intel_crtc_disable_noatomic(struct drm_crtc *crtc)
@@ -6865,7 +5857,7 @@ static void intel_crtc_disable_noatomic(struct drm_crtc *crtc)
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct drm_i915_private *dev_priv = to_i915(crtc->dev);
 	enum intel_display_power_domain domain;
-	unsigned long domains;
+	u64 domains;
 	struct drm_atomic_state *state;
 	struct intel_crtc_state *crtc_state;
 	int ret;
@@ -7173,7 +6165,7 @@ static bool pipe_config_supports_ips(struct drm_i915_private *dev_priv,
 	 *
 	 * Should measure whether using a lower cdclk w/o IPS
 	 */
-	return ilk_pipe_pixel_rate(pipe_config) <=
+	return pipe_config->pixel_rate <=
 		dev_priv->max_cdclk_freq * 95 / 100;
 }
 
@@ -7197,6 +6189,54 @@ static bool intel_crtc_supports_double_wide(const struct intel_crtc *crtc)
 		(crtc->pipe == PIPE_A || IS_I915G(dev_priv));
 }
 
+static uint32_t ilk_pipe_pixel_rate(const struct intel_crtc_state *pipe_config)
+{
+	uint32_t pixel_rate;
+
+	pixel_rate = pipe_config->base.adjusted_mode.crtc_clock;
+
+	/*
+	 * We only use IF-ID interlacing. If we ever use
+	 * PF-ID we'll need to adjust the pixel_rate here.
+	 */
+
+	if (pipe_config->pch_pfit.enabled) {
+		uint64_t pipe_w, pipe_h, pfit_w, pfit_h;
+		uint32_t pfit_size = pipe_config->pch_pfit.size;
+
+		pipe_w = pipe_config->pipe_src_w;
+		pipe_h = pipe_config->pipe_src_h;
+
+		pfit_w = (pfit_size >> 16) & 0xFFFF;
+		pfit_h = pfit_size & 0xFFFF;
+		if (pipe_w < pfit_w)
+			pipe_w = pfit_w;
+		if (pipe_h < pfit_h)
+			pipe_h = pfit_h;
+
+		if (WARN_ON(!pfit_w || !pfit_h))
+			return pixel_rate;
+
+		pixel_rate = div_u64((uint64_t) pixel_rate * pipe_w * pipe_h,
+				     pfit_w * pfit_h);
+	}
+
+	return pixel_rate;
+}
+
+static void intel_crtc_compute_pixel_rate(struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev);
+
+	if (HAS_GMCH_DISPLAY(dev_priv))
+		/* FIXME calculate proper pipe pixel rate for GMCH pfit */
+		crtc_state->pixel_rate =
+			crtc_state->base.adjusted_mode.crtc_clock;
+	else
+		crtc_state->pixel_rate =
+			ilk_pipe_pixel_rate(crtc_state);
+}
+
 static int intel_crtc_compute_config(struct intel_crtc *crtc,
 				     struct intel_crtc_state *pipe_config)
 {
@@ -7243,6 +6283,8 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 		adjusted_mode->crtc_hsync_start == adjusted_mode->crtc_hdisplay)
 		return -EINVAL;
 
+	intel_crtc_compute_pixel_rate(pipe_config);
+
 	if (HAS_IPS(dev_priv))
 		hsw_compute_ips_config(crtc, pipe_config);
 
@@ -7252,428 +6294,6 @@ static int intel_crtc_compute_config(struct intel_crtc *crtc,
 	return 0;
 }
 
-static int skylake_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	u32 cdctl;
-
-	skl_dpll0_update(dev_priv);
-
-	if (dev_priv->cdclk_pll.vco == 0)
-		return dev_priv->cdclk_pll.ref;
-
-	cdctl = I915_READ(CDCLK_CTL);
-
-	if (dev_priv->cdclk_pll.vco == 8640000) {
-		switch (cdctl & CDCLK_FREQ_SEL_MASK) {
-		case CDCLK_FREQ_450_432:
-			return 432000;
-		case CDCLK_FREQ_337_308:
-			return 308571;
-		case CDCLK_FREQ_540:
-			return 540000;
-		case CDCLK_FREQ_675_617:
-			return 617143;
-		default:
-			MISSING_CASE(cdctl & CDCLK_FREQ_SEL_MASK);
-		}
-	} else {
-		switch (cdctl & CDCLK_FREQ_SEL_MASK) {
-		case CDCLK_FREQ_450_432:
-			return 450000;
-		case CDCLK_FREQ_337_308:
-			return 337500;
-		case CDCLK_FREQ_540:
-			return 540000;
-		case CDCLK_FREQ_675_617:
-			return 675000;
-		default:
-			MISSING_CASE(cdctl & CDCLK_FREQ_SEL_MASK);
-		}
-	}
-
-	return dev_priv->cdclk_pll.ref;
-}
-
-static void bxt_de_pll_update(struct drm_i915_private *dev_priv)
-{
-	u32 val;
-
-	dev_priv->cdclk_pll.ref = 19200;
-	dev_priv->cdclk_pll.vco = 0;
-
-	val = I915_READ(BXT_DE_PLL_ENABLE);
-	if ((val & BXT_DE_PLL_PLL_ENABLE) == 0)
-		return;
-
-	if (WARN_ON((val & BXT_DE_PLL_LOCK) == 0))
-		return;
-
-	val = I915_READ(BXT_DE_PLL_CTL);
-	dev_priv->cdclk_pll.vco = (val & BXT_DE_PLL_RATIO_MASK) *
-		dev_priv->cdclk_pll.ref;
-}
-
-static int broxton_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	u32 divider;
-	int div, vco;
-
-	bxt_de_pll_update(dev_priv);
-
-	vco = dev_priv->cdclk_pll.vco;
-	if (vco == 0)
-		return dev_priv->cdclk_pll.ref;
-
-	divider = I915_READ(CDCLK_CTL) & BXT_CDCLK_CD2X_DIV_SEL_MASK;
-
-	switch (divider) {
-	case BXT_CDCLK_CD2X_DIV_SEL_1:
-		div = 2;
-		break;
-	case BXT_CDCLK_CD2X_DIV_SEL_1_5:
-		WARN(IS_GEMINILAKE(dev_priv), "Unsupported divider\n");
-		div = 3;
-		break;
-	case BXT_CDCLK_CD2X_DIV_SEL_2:
-		div = 4;
-		break;
-	case BXT_CDCLK_CD2X_DIV_SEL_4:
-		div = 8;
-		break;
-	default:
-		MISSING_CASE(divider);
-		return dev_priv->cdclk_pll.ref;
-	}
-
-	return DIV_ROUND_CLOSEST(vco, div);
-}
-
-static int broadwell_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	uint32_t lcpll = I915_READ(LCPLL_CTL);
-	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
-
-	if (lcpll & LCPLL_CD_SOURCE_FCLK)
-		return 800000;
-	else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
-		return 450000;
-	else if (freq == LCPLL_CLK_FREQ_450)
-		return 450000;
-	else if (freq == LCPLL_CLK_FREQ_54O_BDW)
-		return 540000;
-	else if (freq == LCPLL_CLK_FREQ_337_5_BDW)
-		return 337500;
-	else
-		return 675000;
-}
-
-static int haswell_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	uint32_t lcpll = I915_READ(LCPLL_CTL);
-	uint32_t freq = lcpll & LCPLL_CLK_FREQ_MASK;
-
-	if (lcpll & LCPLL_CD_SOURCE_FCLK)
-		return 800000;
-	else if (I915_READ(FUSE_STRAP) & HSW_CDCLK_LIMIT)
-		return 450000;
-	else if (freq == LCPLL_CLK_FREQ_450)
-		return 450000;
-	else if (IS_HSW_ULT(dev_priv))
-		return 337500;
-	else
-		return 540000;
-}
-
-static int valleyview_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	return vlv_get_cck_clock_hpll(dev_priv, "cdclk",
-				      CCK_DISPLAY_CLOCK_CONTROL);
-}
-
-static int ilk_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	return 450000;
-}
-
-static int i945_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	return 400000;
-}
-
-static int i915_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	return 333333;
-}
-
-static int i9xx_misc_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	return 200000;
-}
-
-static int pnv_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	struct pci_dev *pdev = dev_priv->drm.pdev;
-	u16 gcfgc = 0;
-
-	pci_read_config_word(pdev, GCFGC, &gcfgc);
-
-	switch (gcfgc & GC_DISPLAY_CLOCK_MASK) {
-	case GC_DISPLAY_CLOCK_267_MHZ_PNV:
-		return 266667;
-	case GC_DISPLAY_CLOCK_333_MHZ_PNV:
-		return 333333;
-	case GC_DISPLAY_CLOCK_444_MHZ_PNV:
-		return 444444;
-	case GC_DISPLAY_CLOCK_200_MHZ_PNV:
-		return 200000;
-	default:
-		DRM_ERROR("Unknown pnv display core clock 0x%04x\n", gcfgc);
-	case GC_DISPLAY_CLOCK_133_MHZ_PNV:
-		return 133333;
-	case GC_DISPLAY_CLOCK_167_MHZ_PNV:
-		return 166667;
-	}
-}
-
-static int i915gm_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	struct pci_dev *pdev = dev_priv->drm.pdev;
-	u16 gcfgc = 0;
-
-	pci_read_config_word(pdev, GCFGC, &gcfgc);
-
-	if (gcfgc & GC_LOW_FREQUENCY_ENABLE)
-		return 133333;
-	else {
-		switch (gcfgc & GC_DISPLAY_CLOCK_MASK) {
-		case GC_DISPLAY_CLOCK_333_MHZ:
-			return 333333;
-		default:
-		case GC_DISPLAY_CLOCK_190_200_MHZ:
-			return 190000;
-		}
-	}
-}
-
-static int i865_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	return 266667;
-}
-
-static int i85x_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	struct pci_dev *pdev = dev_priv->drm.pdev;
-	u16 hpllcc = 0;
-
-	/*
-	 * 852GM/852GMV only supports 133 MHz and the HPLLCC
-	 * encoding is different :(
-	 * FIXME is this the right way to detect 852GM/852GMV?
-	 */
-	if (pdev->revision == 0x1)
-		return 133333;
-
-	pci_bus_read_config_word(pdev->bus,
-				 PCI_DEVFN(0, 3), HPLLCC, &hpllcc);
-
-	/* Assume that the hardware is in the high speed state.  This
-	 * should be the default.
-	 */
-	switch (hpllcc & GC_CLOCK_CONTROL_MASK) {
-	case GC_CLOCK_133_200:
-	case GC_CLOCK_133_200_2:
-	case GC_CLOCK_100_200:
-		return 200000;
-	case GC_CLOCK_166_250:
-		return 250000;
-	case GC_CLOCK_100_133:
-		return 133333;
-	case GC_CLOCK_133_266:
-	case GC_CLOCK_133_266_2:
-	case GC_CLOCK_166_266:
-		return 266667;
-	}
-
-	/* Shouldn't happen */
-	return 0;
-}
-
-static int i830_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	return 133333;
-}
-
-static unsigned int intel_hpll_vco(struct drm_i915_private *dev_priv)
-{
-	static const unsigned int blb_vco[8] = {
-		[0] = 3200000,
-		[1] = 4000000,
-		[2] = 5333333,
-		[3] = 4800000,
-		[4] = 6400000,
-	};
-	static const unsigned int pnv_vco[8] = {
-		[0] = 3200000,
-		[1] = 4000000,
-		[2] = 5333333,
-		[3] = 4800000,
-		[4] = 2666667,
-	};
-	static const unsigned int cl_vco[8] = {
-		[0] = 3200000,
-		[1] = 4000000,
-		[2] = 5333333,
-		[3] = 6400000,
-		[4] = 3333333,
-		[5] = 3566667,
-		[6] = 4266667,
-	};
-	static const unsigned int elk_vco[8] = {
-		[0] = 3200000,
-		[1] = 4000000,
-		[2] = 5333333,
-		[3] = 4800000,
-	};
-	static const unsigned int ctg_vco[8] = {
-		[0] = 3200000,
-		[1] = 4000000,
-		[2] = 5333333,
-		[3] = 6400000,
-		[4] = 2666667,
-		[5] = 4266667,
-	};
-	const unsigned int *vco_table;
-	unsigned int vco;
-	uint8_t tmp = 0;
-
-	/* FIXME other chipsets? */
-	if (IS_GM45(dev_priv))
-		vco_table = ctg_vco;
-	else if (IS_G4X(dev_priv))
-		vco_table = elk_vco;
-	else if (IS_I965GM(dev_priv))
-		vco_table = cl_vco;
-	else if (IS_PINEVIEW(dev_priv))
-		vco_table = pnv_vco;
-	else if (IS_G33(dev_priv))
-		vco_table = blb_vco;
-	else
-		return 0;
-
-	tmp = I915_READ(IS_MOBILE(dev_priv) ? HPLLVCO_MOBILE : HPLLVCO);
-
-	vco = vco_table[tmp & 0x7];
-	if (vco == 0)
-		DRM_ERROR("Bad HPLL VCO (HPLLVCO=0x%02x)\n", tmp);
-	else
-		DRM_DEBUG_KMS("HPLL VCO %u kHz\n", vco);
-
-	return vco;
-}
-
-static int gm45_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	struct pci_dev *pdev = dev_priv->drm.pdev;
-	unsigned int cdclk_sel, vco = intel_hpll_vco(dev_priv);
-	uint16_t tmp = 0;
-
-	pci_read_config_word(pdev, GCFGC, &tmp);
-
-	cdclk_sel = (tmp >> 12) & 0x1;
-
-	switch (vco) {
-	case 2666667:
-	case 4000000:
-	case 5333333:
-		return cdclk_sel ? 333333 : 222222;
-	case 3200000:
-		return cdclk_sel ? 320000 : 228571;
-	default:
-		DRM_ERROR("Unable to determine CDCLK. HPLL VCO=%u, CFGC=0x%04x\n", vco, tmp);
-		return 222222;
-	}
-}
-
-static int i965gm_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	struct pci_dev *pdev = dev_priv->drm.pdev;
-	static const uint8_t div_3200[] = { 16, 10,  8 };
-	static const uint8_t div_4000[] = { 20, 12, 10 };
-	static const uint8_t div_5333[] = { 24, 16, 14 };
-	const uint8_t *div_table;
-	unsigned int cdclk_sel, vco = intel_hpll_vco(dev_priv);
-	uint16_t tmp = 0;
-
-	pci_read_config_word(pdev, GCFGC, &tmp);
-
-	cdclk_sel = ((tmp >> 8) & 0x1f) - 1;
-
-	if (cdclk_sel >= ARRAY_SIZE(div_3200))
-		goto fail;
-
-	switch (vco) {
-	case 3200000:
-		div_table = div_3200;
-		break;
-	case 4000000:
-		div_table = div_4000;
-		break;
-	case 5333333:
-		div_table = div_5333;
-		break;
-	default:
-		goto fail;
-	}
-
-	return DIV_ROUND_CLOSEST(vco, div_table[cdclk_sel]);
-
-fail:
-	DRM_ERROR("Unable to determine CDCLK. HPLL VCO=%u kHz, CFGC=0x%04x\n", vco, tmp);
-	return 200000;
-}
-
-static int g33_get_display_clock_speed(struct drm_i915_private *dev_priv)
-{
-	struct pci_dev *pdev = dev_priv->drm.pdev;
-	static const uint8_t div_3200[] = { 12, 10,  8,  7, 5, 16 };
-	static const uint8_t div_4000[] = { 14, 12, 10,  8, 6, 20 };
-	static const uint8_t div_4800[] = { 20, 14, 12, 10, 8, 24 };
-	static const uint8_t div_5333[] = { 20, 16, 12, 12, 8, 28 };
-	const uint8_t *div_table;
-	unsigned int cdclk_sel, vco = intel_hpll_vco(dev_priv);
-	uint16_t tmp = 0;
-
-	pci_read_config_word(pdev, GCFGC, &tmp);
-
-	cdclk_sel = (tmp >> 4) & 0x7;
-
-	if (cdclk_sel >= ARRAY_SIZE(div_3200))
-		goto fail;
-
-	switch (vco) {
-	case 3200000:
-		div_table = div_3200;
-		break;
-	case 4000000:
-		div_table = div_4000;
-		break;
-	case 4800000:
-		div_table = div_4800;
-		break;
-	case 5333333:
-		div_table = div_5333;
-		break;
-	default:
-		goto fail;
-	}
-
-	return DIV_ROUND_CLOSEST(vco, div_table[cdclk_sel]);
-
-fail:
-	DRM_ERROR("Unable to determine CDCLK. HPLL VCO=%u kHz, CFGC=0x%08x\n", vco, tmp);
-	return 190476;
-}
-
 static void
 intel_reduce_m_n_ratio(uint32_t *num, uint32_t *den)
 {
@@ -8768,7 +7388,8 @@ i9xx_get_initial_plane_config(struct intel_crtc *crtc,
 	val = I915_READ(DSPSTRIDE(pipe));
 	fb->pitches[0] = val & 0xffffffc0;
 
-	aligned_height = intel_fb_align_height(dev, fb->height,
+	aligned_height = intel_fb_align_height(dev_priv,
+					       fb->height,
 					       fb->format->format,
 					       fb->modifier);
 
@@ -9809,7 +8430,8 @@ skylake_get_initial_plane_config(struct intel_crtc *crtc,
 						fb->format->format);
 	fb->pitches[0] = (val & 0x3ff) * stride_mult;
 
-	aligned_height = intel_fb_align_height(dev, fb->height,
+	aligned_height = intel_fb_align_height(dev_priv,
+					       fb->height,
 					       fb->format->format,
 					       fb->modifier);
 
@@ -9907,7 +8529,8 @@ ironlake_get_initial_plane_config(struct intel_crtc *crtc,
 	val = I915_READ(DSPSTRIDE(pipe));
 	fb->pitches[0] = val & 0xffffffc0;
 
-	aligned_height = intel_fb_align_height(dev, fb->height,
+	aligned_height = intel_fb_align_height(dev_priv,
+					       fb->height,
 					       fb->format->format,
 					       fb->modifier);
 
@@ -10235,245 +8858,6 @@ void hsw_disable_pc8(struct drm_i915_private *dev_priv)
 	}
 }
 
-static void bxt_modeset_commit_cdclk(struct drm_atomic_state *old_state)
-{
-	struct drm_device *dev = old_state->dev;
-	struct intel_atomic_state *old_intel_state =
-		to_intel_atomic_state(old_state);
-	unsigned int req_cdclk = old_intel_state->dev_cdclk;
-
-	bxt_set_cdclk(to_i915(dev), req_cdclk);
-}
-
-static int bdw_adjust_min_pipe_pixel_rate(struct intel_crtc_state *crtc_state,
-					  int pixel_rate)
-{
-	struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev);
-
-	/* pixel rate mustn't exceed 95% of cdclk with IPS on BDW */
-	if (IS_BROADWELL(dev_priv) && crtc_state->ips_enabled)
-		pixel_rate = DIV_ROUND_UP(pixel_rate * 100, 95);
-
-	/* BSpec says "Do not use DisplayPort with CDCLK less than
-	 * 432 MHz, audio enabled, port width x4, and link rate
-	 * HBR2 (5.4 GHz), or else there may be audio corruption or
-	 * screen corruption."
-	 */
-	if (intel_crtc_has_dp_encoder(crtc_state) &&
-	    crtc_state->has_audio &&
-	    crtc_state->port_clock >= 540000 &&
-	    crtc_state->lane_count == 4)
-		pixel_rate = max(432000, pixel_rate);
-
-	return pixel_rate;
-}
-
-/* compute the max rate for new configuration */
-static int ilk_max_pixel_rate(struct drm_atomic_state *state)
-{
-	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
-	struct drm_i915_private *dev_priv = to_i915(state->dev);
-	struct drm_crtc *crtc;
-	struct drm_crtc_state *cstate;
-	struct intel_crtc_state *crtc_state;
-	unsigned max_pixel_rate = 0, i;
-	enum pipe pipe;
-
-	memcpy(intel_state->min_pixclk, dev_priv->min_pixclk,
-	       sizeof(intel_state->min_pixclk));
-
-	for_each_crtc_in_state(state, crtc, cstate, i) {
-		int pixel_rate;
-
-		crtc_state = to_intel_crtc_state(cstate);
-		if (!crtc_state->base.enable) {
-			intel_state->min_pixclk[i] = 0;
-			continue;
-		}
-
-		pixel_rate = ilk_pipe_pixel_rate(crtc_state);
-
-		if (IS_BROADWELL(dev_priv) || IS_GEN9(dev_priv))
-			pixel_rate = bdw_adjust_min_pipe_pixel_rate(crtc_state,
-								    pixel_rate);
-
-		intel_state->min_pixclk[i] = pixel_rate;
-	}
-
-	for_each_pipe(dev_priv, pipe)
-		max_pixel_rate = max(intel_state->min_pixclk[pipe], max_pixel_rate);
-
-	return max_pixel_rate;
-}
-
-static void broadwell_set_cdclk(struct drm_device *dev, int cdclk)
-{
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	uint32_t val, data;
-	int ret;
-
-	if (WARN((I915_READ(LCPLL_CTL) &
-		  (LCPLL_PLL_DISABLE | LCPLL_PLL_LOCK |
-		   LCPLL_CD_CLOCK_DISABLE | LCPLL_ROOT_CD_CLOCK_DISABLE |
-		   LCPLL_CD2X_CLOCK_DISABLE | LCPLL_POWER_DOWN_ALLOW |
-		   LCPLL_CD_SOURCE_FCLK)) != LCPLL_PLL_LOCK,
-		 "trying to change cdclk frequency with cdclk not enabled\n"))
-		return;
-
-	mutex_lock(&dev_priv->rps.hw_lock);
-	ret = sandybridge_pcode_write(dev_priv,
-				      BDW_PCODE_DISPLAY_FREQ_CHANGE_REQ, 0x0);
-	mutex_unlock(&dev_priv->rps.hw_lock);
-	if (ret) {
-		DRM_ERROR("failed to inform pcode about cdclk change\n");
-		return;
-	}
-
-	val = I915_READ(LCPLL_CTL);
-	val |= LCPLL_CD_SOURCE_FCLK;
-	I915_WRITE(LCPLL_CTL, val);
-
-	if (wait_for_us(I915_READ(LCPLL_CTL) &
-			LCPLL_CD_SOURCE_FCLK_DONE, 1))
-		DRM_ERROR("Switching to FCLK failed\n");
-
-	val = I915_READ(LCPLL_CTL);
-	val &= ~LCPLL_CLK_FREQ_MASK;
-
-	switch (cdclk) {
-	case 450000:
-		val |= LCPLL_CLK_FREQ_450;
-		data = 0;
-		break;
-	case 540000:
-		val |= LCPLL_CLK_FREQ_54O_BDW;
-		data = 1;
-		break;
-	case 337500:
-		val |= LCPLL_CLK_FREQ_337_5_BDW;
-		data = 2;
-		break;
-	case 675000:
-		val |= LCPLL_CLK_FREQ_675_BDW;
-		data = 3;
-		break;
-	default:
-		WARN(1, "invalid cdclk frequency\n");
-		return;
-	}
-
-	I915_WRITE(LCPLL_CTL, val);
-
-	val = I915_READ(LCPLL_CTL);
-	val &= ~LCPLL_CD_SOURCE_FCLK;
-	I915_WRITE(LCPLL_CTL, val);
-
-	if (wait_for_us((I915_READ(LCPLL_CTL) &
-			LCPLL_CD_SOURCE_FCLK_DONE) == 0, 1))
-		DRM_ERROR("Switching back to LCPLL failed\n");
-
-	mutex_lock(&dev_priv->rps.hw_lock);
-	sandybridge_pcode_write(dev_priv, HSW_PCODE_DE_WRITE_FREQ_REQ, data);
-	mutex_unlock(&dev_priv->rps.hw_lock);
-
-	I915_WRITE(CDCLK_FREQ, DIV_ROUND_CLOSEST(cdclk, 1000) - 1);
-
-	intel_update_cdclk(dev_priv);
-
-	WARN(cdclk != dev_priv->cdclk_freq,
-	     "cdclk requested %d kHz but got %d kHz\n",
-	     cdclk, dev_priv->cdclk_freq);
-}
-
-static int broadwell_calc_cdclk(int max_pixclk)
-{
-	if (max_pixclk > 540000)
-		return 675000;
-	else if (max_pixclk > 450000)
-		return 540000;
-	else if (max_pixclk > 337500)
-		return 450000;
-	else
-		return 337500;
-}
-
-static int broadwell_modeset_calc_cdclk(struct drm_atomic_state *state)
-{
-	struct drm_i915_private *dev_priv = to_i915(state->dev);
-	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
-	int max_pixclk = ilk_max_pixel_rate(state);
-	int cdclk;
-
-	/*
-	 * FIXME should also account for plane ratio
-	 * once 64bpp pixel formats are supported.
-	 */
-	cdclk = broadwell_calc_cdclk(max_pixclk);
-
-	if (cdclk > dev_priv->max_cdclk_freq) {
-		DRM_DEBUG_KMS("requested cdclk (%d kHz) exceeds max (%d kHz)\n",
-			      cdclk, dev_priv->max_cdclk_freq);
-		return -EINVAL;
-	}
-
-	intel_state->cdclk = intel_state->dev_cdclk = cdclk;
-	if (!intel_state->active_crtcs)
-		intel_state->dev_cdclk = broadwell_calc_cdclk(0);
-
-	return 0;
-}
-
-static void broadwell_modeset_commit_cdclk(struct drm_atomic_state *old_state)
-{
-	struct drm_device *dev = old_state->dev;
-	struct intel_atomic_state *old_intel_state =
-		to_intel_atomic_state(old_state);
-	unsigned req_cdclk = old_intel_state->dev_cdclk;
-
-	broadwell_set_cdclk(dev, req_cdclk);
-}
-
-static int skl_modeset_calc_cdclk(struct drm_atomic_state *state)
-{
-	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
-	struct drm_i915_private *dev_priv = to_i915(state->dev);
-	const int max_pixclk = ilk_max_pixel_rate(state);
-	int vco = intel_state->cdclk_pll_vco;
-	int cdclk;
-
-	/*
-	 * FIXME should also account for plane ratio
-	 * once 64bpp pixel formats are supported.
-	 */
-	cdclk = skl_calc_cdclk(max_pixclk, vco);
-
-	/*
-	 * FIXME move the cdclk caclulation to
-	 * compute_config() so we can fail gracegully.
-	 */
-	if (cdclk > dev_priv->max_cdclk_freq) {
-		DRM_ERROR("requested cdclk (%d kHz) exceeds max (%d kHz)\n",
-			  cdclk, dev_priv->max_cdclk_freq);
-		cdclk = dev_priv->max_cdclk_freq;
-	}
-
-	intel_state->cdclk = intel_state->dev_cdclk = cdclk;
-	if (!intel_state->active_crtcs)
-		intel_state->dev_cdclk = skl_calc_cdclk(0, vco);
-
-	return 0;
-}
-
-static void skl_modeset_commit_cdclk(struct drm_atomic_state *old_state)
-{
-	struct drm_i915_private *dev_priv = to_i915(old_state->dev);
-	struct intel_atomic_state *intel_state = to_intel_atomic_state(old_state);
-	unsigned int req_cdclk = intel_state->dev_cdclk;
-	unsigned int req_vco = intel_state->cdclk_pll_vco;
-
-	skl_set_cdclk(dev_priv, req_cdclk, req_vco);
-}
-
 static int haswell_crtc_compute_clock(struct intel_crtc *crtc,
 				      struct intel_crtc_state *crtc_state)
 {
@@ -10565,7 +8949,7 @@ static void haswell_get_ddi_pll(struct drm_i915_private *dev_priv,
 
 static bool hsw_get_transcoder_state(struct intel_crtc *crtc,
 				     struct intel_crtc_state *pipe_config,
-				     unsigned long *power_domain_mask)
+				     u64 *power_domain_mask)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
@@ -10607,7 +8991,7 @@ static bool hsw_get_transcoder_state(struct intel_crtc *crtc,
 	power_domain = POWER_DOMAIN_TRANSCODER(pipe_config->cpu_transcoder);
 	if (!intel_display_power_get_if_enabled(dev_priv, power_domain))
 		return false;
-	*power_domain_mask |= BIT(power_domain);
+	*power_domain_mask |= BIT_ULL(power_domain);
 
 	tmp = I915_READ(PIPECONF(pipe_config->cpu_transcoder));
 
@@ -10616,7 +9000,7 @@ static bool hsw_get_transcoder_state(struct intel_crtc *crtc,
 
 static bool bxt_get_dsi_transcoder_state(struct intel_crtc *crtc,
 					 struct intel_crtc_state *pipe_config,
-					 unsigned long *power_domain_mask)
+					 u64 *power_domain_mask)
 {
 	struct drm_device *dev = crtc->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
@@ -10634,7 +9018,7 @@ static bool bxt_get_dsi_transcoder_state(struct intel_crtc *crtc,
 		power_domain = POWER_DOMAIN_TRANSCODER(cpu_transcoder);
 		if (!intel_display_power_get_if_enabled(dev_priv, power_domain))
 			continue;
-		*power_domain_mask |= BIT(power_domain);
+		*power_domain_mask |= BIT_ULL(power_domain);
 
 		/*
 		 * The PLL needs to be enabled with a valid divider
@@ -10674,7 +9058,7 @@ static void haswell_get_ddi_port_state(struct intel_crtc *crtc,
 
 	port = (tmp & TRANS_DDI_PORT_MASK) >> TRANS_DDI_PORT_SHIFT;
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+	if (IS_GEN9_BC(dev_priv))
 		skylake_get_ddi_pll(dev_priv, port, pipe_config);
 	else if (IS_GEN9_LP(dev_priv))
 		bxt_get_ddi_pll(dev_priv, port, pipe_config);
@@ -10709,13 +9093,13 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc,
 {
 	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
 	enum intel_display_power_domain power_domain;
-	unsigned long power_domain_mask;
+	u64 power_domain_mask;
 	bool active;
 
 	power_domain = POWER_DOMAIN_PIPE(crtc->pipe);
 	if (!intel_display_power_get_if_enabled(dev_priv, power_domain))
 		return false;
-	power_domain_mask = BIT(power_domain);
+	power_domain_mask = BIT_ULL(power_domain);
 
 	pipe_config->shared_dpll = NULL;
 
@@ -10749,7 +9133,7 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc,
 
 	power_domain = POWER_DOMAIN_PIPE_PANEL_FITTER(crtc->pipe);
 	if (intel_display_power_get_if_enabled(dev_priv, power_domain)) {
-		power_domain_mask |= BIT(power_domain);
+		power_domain_mask |= BIT_ULL(power_domain);
 		if (INTEL_GEN(dev_priv) >= 9)
 			skylake_get_pfit_config(crtc, pipe_config);
 		else
@@ -10972,9 +9356,8 @@ static struct drm_display_mode load_detect_mode = {
 };
 
 struct drm_framebuffer *
-__intel_framebuffer_create(struct drm_device *dev,
-			   struct drm_mode_fb_cmd2 *mode_cmd,
-			   struct drm_i915_gem_object *obj)
+intel_framebuffer_create(struct drm_i915_gem_object *obj,
+			 struct drm_mode_fb_cmd2 *mode_cmd)
 {
 	struct intel_framebuffer *intel_fb;
 	int ret;
@@ -10983,7 +9366,7 @@ __intel_framebuffer_create(struct drm_device *dev,
 	if (!intel_fb)
 		return ERR_PTR(-ENOMEM);
 
-	ret = intel_framebuffer_init(dev, intel_fb, mode_cmd, obj);
+	ret = intel_framebuffer_init(intel_fb, obj, mode_cmd);
 	if (ret)
 		goto err;
 
@@ -10994,23 +9377,6 @@ err:
 	return ERR_PTR(ret);
 }
 
-static struct drm_framebuffer *
-intel_framebuffer_create(struct drm_device *dev,
-			 struct drm_mode_fb_cmd2 *mode_cmd,
-			 struct drm_i915_gem_object *obj)
-{
-	struct drm_framebuffer *fb;
-	int ret;
-
-	ret = i915_mutex_lock_interruptible(dev);
-	if (ret)
-		return ERR_PTR(ret);
-	fb = __intel_framebuffer_create(dev, mode_cmd, obj);
-	mutex_unlock(&dev->struct_mutex);
-
-	return fb;
-}
-
 static u32
 intel_framebuffer_pitch_for_width(int width, int bpp)
 {
@@ -11045,7 +9411,7 @@ intel_framebuffer_create_for_mode(struct drm_device *dev,
 								bpp);
 	mode_cmd.pixel_format = drm_mode_legacy_fb_format(bpp, depth);
 
-	fb = intel_framebuffer_create(dev, &mode_cmd, obj);
+	fb = intel_framebuffer_create(obj, &mode_cmd);
 	if (IS_ERR(fb))
 		i915_gem_object_put(obj);
 
@@ -11731,14 +10097,12 @@ static int intel_gen2_queue_flip(struct drm_device *dev,
 				 struct drm_i915_gem_request *req,
 				 uint32_t flags)
 {
-	struct intel_ring *ring = req->ring;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	u32 flip_mask;
-	int ret;
+	u32 flip_mask, *cs;
 
-	ret = intel_ring_begin(req, 6);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	/* Can't queue multiple flips, so wait for the previous
 	 * one to finish before executing the next.
@@ -11747,13 +10111,12 @@ static int intel_gen2_queue_flip(struct drm_device *dev,
 		flip_mask = MI_WAIT_FOR_PLANE_B_FLIP;
 	else
 		flip_mask = MI_WAIT_FOR_PLANE_A_FLIP;
-	intel_ring_emit(ring, MI_WAIT_FOR_EVENT | flip_mask);
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_emit(ring, MI_DISPLAY_FLIP |
-			MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
-	intel_ring_emit(ring, fb->pitches[0]);
-	intel_ring_emit(ring, intel_crtc->flip_work->gtt_offset);
-	intel_ring_emit(ring, 0); /* aux display base address, unused */
+	*cs++ = MI_WAIT_FOR_EVENT | flip_mask;
+	*cs++ = MI_NOOP;
+	*cs++ = MI_DISPLAY_FLIP | MI_DISPLAY_FLIP_PLANE(intel_crtc->plane);
+	*cs++ = fb->pitches[0];
+	*cs++ = intel_crtc->flip_work->gtt_offset;
+	*cs++ = 0; /* aux display base address, unused */
 
 	return 0;
 }
@@ -11765,26 +10128,23 @@ static int intel_gen3_queue_flip(struct drm_device *dev,
 				 struct drm_i915_gem_request *req,
 				 uint32_t flags)
 {
-	struct intel_ring *ring = req->ring;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	u32 flip_mask;
-	int ret;
+	u32 flip_mask, *cs;
 
-	ret = intel_ring_begin(req, 6);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	if (intel_crtc->plane)
 		flip_mask = MI_WAIT_FOR_PLANE_B_FLIP;
 	else
 		flip_mask = MI_WAIT_FOR_PLANE_A_FLIP;
-	intel_ring_emit(ring, MI_WAIT_FOR_EVENT | flip_mask);
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_emit(ring, MI_DISPLAY_FLIP_I915 |
-			MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
-	intel_ring_emit(ring, fb->pitches[0]);
-	intel_ring_emit(ring, intel_crtc->flip_work->gtt_offset);
-	intel_ring_emit(ring, MI_NOOP);
+	*cs++ = MI_WAIT_FOR_EVENT | flip_mask;
+	*cs++ = MI_NOOP;
+	*cs++ = MI_DISPLAY_FLIP_I915 | MI_DISPLAY_FLIP_PLANE(intel_crtc->plane);
+	*cs++ = fb->pitches[0];
+	*cs++ = intel_crtc->flip_work->gtt_offset;
+	*cs++ = MI_NOOP;
 
 	return 0;
 }
@@ -11796,25 +10156,22 @@ static int intel_gen4_queue_flip(struct drm_device *dev,
 				 struct drm_i915_gem_request *req,
 				 uint32_t flags)
 {
-	struct intel_ring *ring = req->ring;
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	uint32_t pf, pipesrc;
-	int ret;
+	u32 pf, pipesrc, *cs;
 
-	ret = intel_ring_begin(req, 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	/* i965+ uses the linear or tiled offsets from the
 	 * Display Registers (which do not change across a page-flip)
 	 * so we need only reprogram the base address.
 	 */
-	intel_ring_emit(ring, MI_DISPLAY_FLIP |
-			MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
-	intel_ring_emit(ring, fb->pitches[0]);
-	intel_ring_emit(ring, intel_crtc->flip_work->gtt_offset |
-			intel_fb_modifier_to_tiling(fb->modifier));
+	*cs++ = MI_DISPLAY_FLIP | MI_DISPLAY_FLIP_PLANE(intel_crtc->plane);
+	*cs++ = fb->pitches[0];
+	*cs++ = intel_crtc->flip_work->gtt_offset |
+		intel_fb_modifier_to_tiling(fb->modifier);
 
 	/* XXX Enabling the panel-fitter across page-flip is so far
 	 * untested on non-native modes, so ignore it for now.
@@ -11822,7 +10179,7 @@ static int intel_gen4_queue_flip(struct drm_device *dev,
 	 */
 	pf = 0;
 	pipesrc = I915_READ(PIPESRC(intel_crtc->pipe)) & 0x0fff0fff;
-	intel_ring_emit(ring, pf | pipesrc);
+	*cs++ = pf | pipesrc;
 
 	return 0;
 }
@@ -11834,21 +10191,17 @@ static int intel_gen6_queue_flip(struct drm_device *dev,
 				 struct drm_i915_gem_request *req,
 				 uint32_t flags)
 {
-	struct intel_ring *ring = req->ring;
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	uint32_t pf, pipesrc;
-	int ret;
+	u32 pf, pipesrc, *cs;
 
-	ret = intel_ring_begin(req, 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, MI_DISPLAY_FLIP |
-			MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
-	intel_ring_emit(ring, fb->pitches[0] |
-			intel_fb_modifier_to_tiling(fb->modifier));
-	intel_ring_emit(ring, intel_crtc->flip_work->gtt_offset);
+	*cs++ = MI_DISPLAY_FLIP | MI_DISPLAY_FLIP_PLANE(intel_crtc->plane);
+	*cs++ = fb->pitches[0] | intel_fb_modifier_to_tiling(fb->modifier);
+	*cs++ = intel_crtc->flip_work->gtt_offset;
 
 	/* Contrary to the suggestions in the documentation,
 	 * "Enable Panel Fitter" does not seem to be required when page
@@ -11858,7 +10211,7 @@ static int intel_gen6_queue_flip(struct drm_device *dev,
 	 */
 	pf = 0;
 	pipesrc = I915_READ(PIPESRC(intel_crtc->pipe)) & 0x0fff0fff;
-	intel_ring_emit(ring, pf | pipesrc);
+	*cs++ = pf | pipesrc;
 
 	return 0;
 }
@@ -11871,9 +10224,8 @@ static int intel_gen7_queue_flip(struct drm_device *dev,
 				 uint32_t flags)
 {
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct intel_ring *ring = req->ring;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	uint32_t plane_bit = 0;
+	u32 *cs, plane_bit = 0;
 	int len, ret;
 
 	switch (intel_crtc->plane) {
@@ -11917,9 +10269,9 @@ static int intel_gen7_queue_flip(struct drm_device *dev,
 	if (ret)
 		return ret;
 
-	ret = intel_ring_begin(req, len);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, len);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	/* Unmask the flip-done completion message. Note that the bspec says that
 	 * we should do this for both the BCS and RCS, and that we must not unmask
@@ -11931,31 +10283,28 @@ static int intel_gen7_queue_flip(struct drm_device *dev,
 	 * to zero does lead to lockups within MI_DISPLAY_FLIP.
 	 */
 	if (req->engine->id == RCS) {
-		intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
-		intel_ring_emit_reg(ring, DERRMR);
-		intel_ring_emit(ring, ~(DERRMR_PIPEA_PRI_FLIP_DONE |
-					  DERRMR_PIPEB_PRI_FLIP_DONE |
-					  DERRMR_PIPEC_PRI_FLIP_DONE));
+		*cs++ = MI_LOAD_REGISTER_IMM(1);
+		*cs++ = i915_mmio_reg_offset(DERRMR);
+		*cs++ = ~(DERRMR_PIPEA_PRI_FLIP_DONE |
+			  DERRMR_PIPEB_PRI_FLIP_DONE |
+			  DERRMR_PIPEC_PRI_FLIP_DONE);
 		if (IS_GEN8(dev_priv))
-			intel_ring_emit(ring, MI_STORE_REGISTER_MEM_GEN8 |
-					      MI_SRM_LRM_GLOBAL_GTT);
+			*cs++ = MI_STORE_REGISTER_MEM_GEN8 |
+				MI_SRM_LRM_GLOBAL_GTT;
 		else
-			intel_ring_emit(ring, MI_STORE_REGISTER_MEM |
-					      MI_SRM_LRM_GLOBAL_GTT);
-		intel_ring_emit_reg(ring, DERRMR);
-		intel_ring_emit(ring,
-				i915_ggtt_offset(req->engine->scratch) + 256);
+			*cs++ = MI_STORE_REGISTER_MEM | MI_SRM_LRM_GLOBAL_GTT;
+		*cs++ = i915_mmio_reg_offset(DERRMR);
+		*cs++ = i915_ggtt_offset(req->engine->scratch) + 256;
 		if (IS_GEN8(dev_priv)) {
-			intel_ring_emit(ring, 0);
-			intel_ring_emit(ring, MI_NOOP);
+			*cs++ = 0;
+			*cs++ = MI_NOOP;
 		}
 	}
 
-	intel_ring_emit(ring, MI_DISPLAY_FLIP_I915 | plane_bit);
-	intel_ring_emit(ring, fb->pitches[0] |
-			intel_fb_modifier_to_tiling(fb->modifier));
-	intel_ring_emit(ring, intel_crtc->flip_work->gtt_offset);
-	intel_ring_emit(ring, (MI_NOOP));
+	*cs++ = MI_DISPLAY_FLIP_I915 | plane_bit;
+	*cs++ = fb->pitches[0] | intel_fb_modifier_to_tiling(fb->modifier);
+	*cs++ = intel_crtc->flip_work->gtt_offset;
+	*cs++ = MI_NOOP;
 
 	return 0;
 }
@@ -12147,6 +10496,7 @@ void intel_check_page_flip(struct drm_i915_private *dev_priv, int pipe)
 	spin_unlock(&dev->event_lock);
 }
 
+__maybe_unused
 static int intel_crtc_page_flip(struct drm_crtc *crtc,
 				struct drm_framebuffer *fb,
 				struct drm_pending_vblank_event *event,
@@ -12441,11 +10791,11 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state,
 	struct intel_crtc_state *pipe_config = to_intel_crtc_state(crtc_state);
 	struct drm_crtc *crtc = crtc_state->crtc;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_plane *plane = plane_state->plane;
+	struct intel_plane *plane = to_intel_plane(plane_state->plane);
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_plane_state *old_plane_state =
-		to_intel_plane_state(plane->state);
+		to_intel_plane_state(plane->base.state);
 	bool mode_changed = needs_modeset(crtc_state);
 	bool was_crtc_enabled = crtc->state->active;
 	bool is_crtc_enabled = crtc_state->active;
@@ -12453,7 +10803,7 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state,
 	struct drm_framebuffer *fb = plane_state->fb;
 	int ret;
 
-	if (INTEL_GEN(dev_priv) >= 9 && plane->type != DRM_PLANE_TYPE_CURSOR) {
+	if (INTEL_GEN(dev_priv) >= 9 && plane->id != PLANE_CURSOR) {
 		ret = skl_update_scaler_plane(
 			to_intel_crtc_state(crtc_state),
 			to_intel_plane_state(plane_state));
@@ -12477,8 +10827,10 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state,
 	 * per-plane wm computation to the .check_plane() hook, and
 	 * only combine the results from all planes in the current place?
 	 */
-	if (!is_crtc_enabled)
+	if (!is_crtc_enabled) {
 		plane_state->visible = visible = false;
+		to_intel_crtc_state(crtc_state)->active_planes &= ~BIT(plane->id);
+	}
 
 	if (!was_visible && !visible)
 		return 0;
@@ -12490,41 +10842,39 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state,
 	turn_on = visible && (!was_visible || mode_changed);
 
 	DRM_DEBUG_ATOMIC("[CRTC:%d:%s] has [PLANE:%d:%s] with fb %i\n",
-			 intel_crtc->base.base.id,
-			 intel_crtc->base.name,
-			 plane->base.id, plane->name,
+			 intel_crtc->base.base.id, intel_crtc->base.name,
+			 plane->base.base.id, plane->base.name,
 			 fb ? fb->base.id : -1);
 
 	DRM_DEBUG_ATOMIC("[PLANE:%d:%s] visible %i -> %i, off %i, on %i, ms %i\n",
-			 plane->base.id, plane->name,
+			 plane->base.base.id, plane->base.name,
 			 was_visible, visible,
 			 turn_off, turn_on, mode_changed);
 
 	if (turn_on) {
-		pipe_config->update_wm_pre = true;
+		if (INTEL_GEN(dev_priv) < 5)
+			pipe_config->update_wm_pre = true;
 
 		/* must disable cxsr around plane enable/disable */
-		if (plane->type != DRM_PLANE_TYPE_CURSOR)
+		if (plane->id != PLANE_CURSOR)
 			pipe_config->disable_cxsr = true;
 	} else if (turn_off) {
-		pipe_config->update_wm_post = true;
+		if (INTEL_GEN(dev_priv) < 5)
+			pipe_config->update_wm_post = true;
 
 		/* must disable cxsr around plane enable/disable */
-		if (plane->type != DRM_PLANE_TYPE_CURSOR)
+		if (plane->id != PLANE_CURSOR)
 			pipe_config->disable_cxsr = true;
-	} else if (intel_wm_need_update(plane, plane_state)) {
-		/* FIXME bollocks */
-		pipe_config->update_wm_pre = true;
-		pipe_config->update_wm_post = true;
+	} else if (intel_wm_need_update(&plane->base, plane_state)) {
+		if (INTEL_GEN(dev_priv) < 5) {
+			/* FIXME bollocks */
+			pipe_config->update_wm_pre = true;
+			pipe_config->update_wm_post = true;
+		}
 	}
 
-	/* Pre-gen9 platforms need two-step watermark updates */
-	if ((pipe_config->update_wm_pre || pipe_config->update_wm_post) &&
-	    INTEL_GEN(dev_priv) < 9 && dev_priv->display.optimize_watermarks)
-		to_intel_crtc_state(crtc_state)->wm.need_postvbl_update = true;
-
 	if (visible || was_visible)
-		pipe_config->fb_bits |= to_intel_plane(plane)->frontbuffer_bit;
+		pipe_config->fb_bits |= plane->frontbuffer_bit;
 
 	/*
 	 * WaCxSRDisabledForSpriteScaling:ivb
@@ -12532,7 +10882,7 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state,
 	 * cstate->update_wm was already set above, so this flag will
 	 * take effect when we commit and program watermarks.
 	 */
-	if (plane->type == DRM_PLANE_TYPE_OVERLAY && IS_IVYBRIDGE(dev_priv) &&
+	if (plane->id == PLANE_SPRITE0 && IS_IVYBRIDGE(dev_priv) &&
 	    needs_scaling(to_intel_plane_state(plane_state)) &&
 	    !needs_scaling(old_plane_state))
 		pipe_config->disable_lp_wm = true;
@@ -12642,7 +10992,7 @@ static int intel_crtc_atomic_check(struct drm_crtc *crtc,
 			ret = skl_update_scaler_crtc(pipe_config);
 
 		if (!ret)
-			ret = intel_atomic_setup_scalers(dev, intel_crtc,
+			ret = intel_atomic_setup_scalers(dev_priv, intel_crtc,
 							 pipe_config);
 	}
 
@@ -12800,9 +11150,10 @@ static void intel_dump_pipe_config(struct intel_crtc *crtc,
 	DRM_DEBUG_KMS("adjusted mode:\n");
 	drm_mode_debug_printmodeline(&pipe_config->base.adjusted_mode);
 	intel_dump_crtc_timings(&pipe_config->base.adjusted_mode);
-	DRM_DEBUG_KMS("port clock: %d, pipe src size: %dx%d\n",
+	DRM_DEBUG_KMS("port clock: %d, pipe src size: %dx%d, pixel rate %d\n",
 		      pipe_config->port_clock,
-		      pipe_config->pipe_src_w, pipe_config->pipe_src_h);
+		      pipe_config->pipe_src_w, pipe_config->pipe_src_h,
+		      pipe_config->pixel_rate);
 
 	if (INTEL_GEN(dev_priv) >= 9)
 		DRM_DEBUG_KMS("num_scalers: %d, scaler_users: 0x%x, scaler_id: %d\n",
@@ -12920,10 +11271,13 @@ static bool check_digital_port_conflicts(struct drm_atomic_state *state)
 static void
 clear_intel_crtc_state(struct intel_crtc_state *crtc_state)
 {
+	struct drm_i915_private *dev_priv =
+		to_i915(crtc_state->base.crtc->dev);
 	struct drm_crtc_state tmp_state;
 	struct intel_crtc_scaler_state scaler_state;
 	struct intel_dpll_hw_state dpll_hw_state;
 	struct intel_shared_dpll *shared_dpll;
+	struct intel_crtc_wm_state wm_state;
 	bool force_thru;
 
 	/* FIXME: before the switch to atomic started, a new pipe_config was
@@ -12936,6 +11290,8 @@ clear_intel_crtc_state(struct intel_crtc_state *crtc_state)
 	shared_dpll = crtc_state->shared_dpll;
 	dpll_hw_state = crtc_state->dpll_hw_state;
 	force_thru = crtc_state->pch_pfit.force_thru;
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		wm_state = crtc_state->wm;
 
 	memset(crtc_state, 0, sizeof *crtc_state);
 
@@ -12944,6 +11300,8 @@ clear_intel_crtc_state(struct intel_crtc_state *crtc_state)
 	crtc_state->shared_dpll = shared_dpll;
 	crtc_state->dpll_hw_state = dpll_hw_state;
 	crtc_state->pch_pfit.force_thru = force_thru;
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+		crtc_state->wm = wm_state;
 }
 
 static int
@@ -13060,8 +11418,11 @@ encoder_retry:
 	}
 
 	/* Dithering seems to not pass-through bits correctly when it should, so
-	 * only enable it on 6bpc panels. */
-	pipe_config->dither = pipe_config->pipe_bpp == 6*3;
+	 * only enable it on 6bpc panels and when its not a compliance
+	 * test requesting 6bpc video pattern.
+	 */
+	pipe_config->dither = (pipe_config->pipe_bpp == 6*3) &&
+		!pipe_config->dither_force_disable;
 	DRM_DEBUG_KMS("hw max bpp: %i, pipe bpp: %i, dithering: %i\n",
 		      base_bpp, pipe_config->pipe_bpp, pipe_config->dither);
 
@@ -13375,6 +11736,7 @@ intel_pipe_config_compare(struct drm_i915_private *dev_priv,
 		}
 
 		PIPE_CONF_CHECK_I(scaler_state.scaler_id);
+		PIPE_CONF_CHECK_CLOCK_FUZZY(pixel_rate);
 	}
 
 	/* BDW+ don't expose a synchronous way to read the state */
@@ -13666,6 +12028,8 @@ verify_crtc_state(struct drm_crtc *crtc,
 		}
 	}
 
+	intel_crtc_compute_pixel_rate(pipe_config);
+
 	if (!new_crtc_state->active)
 		return;
 
@@ -13993,6 +12357,8 @@ static int intel_modeset_checks(struct drm_atomic_state *state)
 
 	intel_state->modeset = true;
 	intel_state->active_crtcs = dev_priv->active_crtcs;
+	intel_state->cdclk.logical = dev_priv->cdclk.logical;
+	intel_state->cdclk.actual = dev_priv->cdclk.actual;
 
 	for_each_crtc_in_state(state, crtc, crtc_state, i) {
 		if (crtc_state->active)
@@ -14012,38 +12378,35 @@ static int intel_modeset_checks(struct drm_atomic_state *state)
 	 * adjusted_mode bits in the crtc directly.
 	 */
 	if (dev_priv->display.modeset_calc_cdclk) {
-		if (!intel_state->cdclk_pll_vco)
-			intel_state->cdclk_pll_vco = dev_priv->cdclk_pll.vco;
-		if (!intel_state->cdclk_pll_vco)
-			intel_state->cdclk_pll_vco = dev_priv->skl_preferred_vco_freq;
-
 		ret = dev_priv->display.modeset_calc_cdclk(state);
 		if (ret < 0)
 			return ret;
 
 		/*
-		 * Writes to dev_priv->atomic_cdclk_freq must protected by
+		 * Writes to dev_priv->cdclk.logical must protected by
 		 * holding all the crtc locks, even if we don't end up
 		 * touching the hardware
 		 */
-		if (intel_state->cdclk != dev_priv->atomic_cdclk_freq) {
+		if (!intel_cdclk_state_compare(&dev_priv->cdclk.logical,
+					       &intel_state->cdclk.logical)) {
 			ret = intel_lock_all_pipes(state);
 			if (ret < 0)
 				return ret;
 		}
 
 		/* All pipes must be switched off while we change the cdclk. */
-		if (intel_state->dev_cdclk != dev_priv->cdclk_freq ||
-		    intel_state->cdclk_pll_vco != dev_priv->cdclk_pll.vco) {
+		if (!intel_cdclk_state_compare(&dev_priv->cdclk.actual,
+					       &intel_state->cdclk.actual)) {
 			ret = intel_modeset_all_pipes(state);
 			if (ret < 0)
 				return ret;
 		}
 
-		DRM_DEBUG_KMS("New cdclk calculated to be atomic %u, actual %u\n",
-			      intel_state->cdclk, intel_state->dev_cdclk);
+		DRM_DEBUG_KMS("New cdclk calculated to be logical %u kHz, actual %u kHz\n",
+			      intel_state->cdclk.logical.cdclk,
+			      intel_state->cdclk.actual.cdclk);
 	} else {
-		to_intel_atomic_state(state)->cdclk = dev_priv->atomic_cdclk_freq;
+		to_intel_atomic_state(state)->cdclk.logical = dev_priv->cdclk.logical;
 	}
 
 	intel_modeset_clear_plls(state);
@@ -14146,7 +12509,7 @@ static int intel_atomic_check(struct drm_device *dev,
 		if (ret)
 			return ret;
 	} else {
-		intel_state->cdclk = dev_priv->atomic_cdclk_freq;
+		intel_state->cdclk.logical = dev_priv->cdclk.logical;
 	}
 
 	ret = drm_atomic_helper_check_planes(dev, state);
@@ -14253,12 +12616,7 @@ static bool needs_vblank_wait(struct intel_crtc_state *crtc_state)
 	if (crtc_state->update_wm_post)
 		return true;
 
-	/*
-	 * cxsr is re-enabled after vblank.
-	 * This is already handled by crtc_state->update_wm_post,
-	 * but added for clarity.
-	 */
-	if (crtc_state->disable_cxsr)
+	if (crtc_state->wm.need_postvbl_update)
 		return true;
 
 	return false;
@@ -14380,6 +12738,24 @@ static void skl_update_crtcs(struct drm_atomic_state *state,
 	} while (progress);
 }
 
+static void intel_atomic_helper_free_state(struct drm_i915_private *dev_priv)
+{
+	struct intel_atomic_state *state, *next;
+	struct llist_node *freed;
+
+	freed = llist_del_all(&dev_priv->atomic_helper.free_list);
+	llist_for_each_entry_safe(state, next, freed, freed)
+		drm_atomic_state_put(&state->base);
+}
+
+static void intel_atomic_helper_free_state_worker(struct work_struct *work)
+{
+	struct drm_i915_private *dev_priv =
+		container_of(work, typeof(*dev_priv), atomic_helper.free_work);
+
+	intel_atomic_helper_free_state(dev_priv);
+}
+
 static void intel_atomic_commit_tail(struct drm_atomic_state *state)
 {
 	struct drm_device *dev = state->dev;
@@ -14389,7 +12765,7 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state)
 	struct drm_crtc *crtc;
 	struct intel_crtc_state *intel_cstate;
 	bool hw_check = intel_state->modeset;
-	unsigned long put_domains[I915_MAX_PIPES] = {};
+	u64 put_domains[I915_MAX_PIPES] = {};
 	unsigned crtc_vblank_mask = 0;
 	int i;
 
@@ -14433,12 +12809,12 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state)
 				/*
 				 * Make sure we don't call initial_watermarks
 				 * for ILK-style watermark updates.
+				 *
+				 * No clue what this is supposed to achieve.
 				 */
-				if (dev_priv->display.atomic_update_watermarks)
+				if (INTEL_GEN(dev_priv) >= 9)
 					dev_priv->display.initial_watermarks(intel_state,
 									     to_intel_crtc_state(crtc->state));
-				else
-					intel_update_watermarks(intel_crtc);
 			}
 		}
 	}
@@ -14450,10 +12826,7 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state)
 	if (intel_state->modeset) {
 		drm_atomic_helper_update_legacy_modeset_state(state->dev, state);
 
-		if (dev_priv->display.modeset_commit_cdclk &&
-		    (intel_state->dev_cdclk != dev_priv->cdclk_freq ||
-		     intel_state->cdclk_pll_vco != dev_priv->cdclk_pll.vco))
-			dev_priv->display.modeset_commit_cdclk(state);
+		intel_set_cdclk(dev_priv, &dev_priv->cdclk.actual);
 
 		/*
 		 * SKL workaround: bspec recommends we disable the SAGV when we
@@ -14546,6 +12919,8 @@ static void intel_atomic_commit_tail(struct drm_atomic_state *state)
 	 * can happen also when the device is completely off.
 	 */
 	intel_uncore_arm_unclaimed_mmio_detection(dev_priv);
+
+	intel_atomic_helper_free_state(dev_priv);
 }
 
 static void intel_atomic_commit_work(struct work_struct *work)
@@ -14615,6 +12990,17 @@ static int intel_atomic_commit(struct drm_device *dev,
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	int ret = 0;
 
+	/*
+	 * The intel_legacy_cursor_update() fast path takes care
+	 * of avoiding the vblank waits for simple cursor
+	 * movement and flips. For cursor on/off and size changes,
+	 * we want to perform the vblank waits so that watermark
+	 * updates happen during the correct frames. Gen9+ have
+	 * double buffered watermarks and so shouldn't need this.
+	 */
+	if (INTEL_GEN(dev_priv) < 9)
+		state->legacy_cursor_update = false;
+
 	ret = drm_atomic_helper_setup_commit(state, nonblock);
 	if (ret)
 		return ret;
@@ -14639,7 +13025,8 @@ static int intel_atomic_commit(struct drm_device *dev,
 		memcpy(dev_priv->min_pixclk, intel_state->min_pixclk,
 		       sizeof(intel_state->min_pixclk));
 		dev_priv->active_crtcs = intel_state->active_crtcs;
-		dev_priv->atomic_cdclk_freq = intel_state->cdclk;
+		dev_priv->cdclk.logical = intel_state->cdclk.logical;
+		dev_priv->cdclk.actual = intel_state->cdclk.actual;
 	}
 
 	drm_atomic_state_get(state);
@@ -14739,7 +13126,7 @@ static const struct drm_crtc_funcs intel_crtc_funcs = {
 	.set_config = drm_atomic_helper_set_config,
 	.set_property = drm_atomic_helper_crtc_set_property,
 	.destroy = intel_crtc_destroy,
-	.page_flip = intel_crtc_page_flip,
+	.page_flip = drm_atomic_helper_page_flip,
 	.atomic_duplicate_state = intel_crtc_duplicate_state,
 	.atomic_destroy_state = intel_crtc_destroy_state,
 	.set_crc_source = intel_crtc_set_crc_source,
@@ -14771,6 +13158,29 @@ intel_prepare_plane_fb(struct drm_plane *plane,
 	struct drm_i915_gem_object *old_obj = intel_fb_obj(plane->state->fb);
 	int ret;
 
+	if (obj) {
+		if (plane->type == DRM_PLANE_TYPE_CURSOR &&
+		    INTEL_INFO(dev_priv)->cursor_needs_physical) {
+			const int align = IS_I830(dev_priv) ? 16 * 1024 : 256;
+
+			ret = i915_gem_object_attach_phys(obj, align);
+			if (ret) {
+				DRM_DEBUG_KMS("failed to attach phys object\n");
+				return ret;
+			}
+		} else {
+			struct i915_vma *vma;
+
+			vma = intel_pin_and_fence_fb_obj(fb, new_state->rotation);
+			if (IS_ERR(vma)) {
+				DRM_DEBUG_KMS("failed to pin object\n");
+				return PTR_ERR(vma);
+			}
+
+			to_intel_plane_state(new_state)->vma = vma;
+		}
+	}
+
 	if (!obj && !old_obj)
 		return 0;
 
@@ -14823,26 +13233,6 @@ intel_prepare_plane_fb(struct drm_plane *plane,
 		i915_gem_object_wait_priority(obj, 0, I915_PRIORITY_DISPLAY);
 	}
 
-	if (plane->type == DRM_PLANE_TYPE_CURSOR &&
-	    INTEL_INFO(dev_priv)->cursor_needs_physical) {
-		int align = IS_I830(dev_priv) ? 16 * 1024 : 256;
-		ret = i915_gem_object_attach_phys(obj, align);
-		if (ret) {
-			DRM_DEBUG_KMS("failed to attach phys object\n");
-			return ret;
-		}
-	} else {
-		struct i915_vma *vma;
-
-		vma = intel_pin_and_fence_fb_obj(fb, new_state->rotation);
-		if (IS_ERR(vma)) {
-			DRM_DEBUG_KMS("failed to pin object\n");
-			return PTR_ERR(vma);
-		}
-
-		to_intel_plane_state(new_state)->vma = vma;
-	}
-
 	return 0;
 }
 
@@ -14870,16 +13260,22 @@ intel_cleanup_plane_fb(struct drm_plane *plane,
 int
 skl_max_scale(struct intel_crtc *intel_crtc, struct intel_crtc_state *crtc_state)
 {
+	struct drm_i915_private *dev_priv;
 	int max_scale;
-	int crtc_clock, cdclk;
+	int crtc_clock, max_dotclk;
 
 	if (!intel_crtc || !crtc_state->base.enable)
 		return DRM_PLANE_HELPER_NO_SCALING;
 
+	dev_priv = to_i915(intel_crtc->base.dev);
+
 	crtc_clock = crtc_state->base.adjusted_mode.crtc_clock;
-	cdclk = to_intel_atomic_state(crtc_state->base.state)->cdclk;
+	max_dotclk = to_intel_atomic_state(crtc_state->base.state)->cdclk.logical.cdclk;
 
-	if (WARN_ON_ONCE(!crtc_clock || cdclk < crtc_clock))
+	if (IS_GEMINILAKE(dev_priv))
+		max_dotclk *= 2;
+
+	if (WARN_ON_ONCE(!crtc_clock || max_dotclk < crtc_clock))
 		return DRM_PLANE_HELPER_NO_SCALING;
 
 	/*
@@ -14888,7 +13284,8 @@ skl_max_scale(struct intel_crtc *intel_crtc, struct intel_crtc_state *crtc_state
 	 *            or
 	 *    cdclk/crtc_clock
 	 */
-	max_scale = min((1 << 16) * 3 - 1, (1 << 8) * ((cdclk << 8) / crtc_clock));
+	max_scale = min((1 << 16) * 3 - 1,
+			(1 << 8) * ((max_dotclk << 8) / crtc_clock));
 
 	return max_scale;
 }
@@ -15038,8 +13435,7 @@ intel_legacy_cursor_update(struct drm_plane *plane,
 	    old_plane_state->src_h != src_h ||
 	    old_plane_state->crtc_w != crtc_w ||
 	    old_plane_state->crtc_h != crtc_h ||
-	    !old_plane_state->visible ||
-	    old_plane_state->fb->modifier != fb->modifier)
+	    !old_plane_state->fb != !fb)
 		goto slow;
 
 	new_plane_state = intel_plane_duplicate_state(plane);
@@ -15062,10 +13458,6 @@ intel_legacy_cursor_update(struct drm_plane *plane,
 	if (ret)
 		goto out_free;
 
-	/* Visibility changed, must take slowpath. */
-	if (!new_plane_state->visible)
-		goto slow_free;
-
 	ret = mutex_lock_interruptible(&dev_priv->drm.struct_mutex);
 	if (ret)
 		goto out_free;
@@ -15105,9 +13497,15 @@ intel_legacy_cursor_update(struct drm_plane *plane,
 	new_plane_state->fb = old_fb;
 	to_intel_plane_state(new_plane_state)->vma = old_vma;
 
-	intel_plane->update_plane(plane,
-				  to_intel_crtc_state(crtc->state),
-				  to_intel_plane_state(plane->state));
+	if (plane->state->visible) {
+		trace_intel_update_plane(plane, to_intel_crtc(crtc));
+		intel_plane->update_plane(plane,
+					  to_intel_crtc_state(crtc->state),
+					  to_intel_plane_state(plane->state));
+	} else {
+		trace_intel_disable_plane(plane, to_intel_crtc(crtc));
+		intel_plane->disable_plane(plane, crtc);
+	}
 
 	intel_cleanup_plane_fb(plane, new_plane_state);
 
@@ -15117,8 +13515,6 @@ out_free:
 	intel_plane_destroy_state(plane, new_plane_state);
 	return ret;
 
-slow_free:
-	intel_plane_destroy_state(plane, new_plane_state);
 slow:
 	return drm_atomic_helper_update_plane(plane, crtc, fb,
 					      crtc_x, crtc_y, crtc_w, crtc_h,
@@ -15492,8 +13888,6 @@ static int intel_crtc_init(struct drm_i915_private *dev_priv, enum pipe pipe)
 	intel_crtc->cursor_cntl = ~0;
 	intel_crtc->cursor_size = ~0;
 
-	intel_crtc->wm.cxsr_allowed = true;
-
 	/* initialize shared scalers */
 	intel_crtc_init_scalers(intel_crtc, crtc_state);
 
@@ -15681,7 +14075,7 @@ static void intel_setup_outputs(struct drm_i915_private *dev_priv)
 		 */
 		found = I915_READ(DDI_BUF_CTL(PORT_A)) & DDI_INIT_DISPLAY_DETECTED;
 		/* WaIgnoreDDIAStrap: skl */
-		if (found || IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+		if (found || IS_GEN9_BC(dev_priv))
 			intel_ddi_init(dev_priv, PORT_A);
 
 		/* DDI B, C and D detection is indicated by the SFUSE_STRAP
@@ -15697,7 +14091,7 @@ static void intel_setup_outputs(struct drm_i915_private *dev_priv)
 		/*
 		 * On SKL we don't have a way to detect DDI-E so we rely on VBT.
 		 */
-		if ((IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) &&
+		if (IS_GEN9_BC(dev_priv) &&
 		    (dev_priv->vbt.ddi_port_info[PORT_E].supports_dp ||
 		     dev_priv->vbt.ddi_port_info[PORT_E].supports_dvi ||
 		     dev_priv->vbt.ddi_port_info[PORT_E].supports_hdmi))
@@ -15830,14 +14224,16 @@ static void intel_setup_outputs(struct drm_i915_private *dev_priv)
 
 static void intel_user_framebuffer_destroy(struct drm_framebuffer *fb)
 {
-	struct drm_device *dev = fb->dev;
 	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
 
 	drm_framebuffer_cleanup(fb);
-	mutex_lock(&dev->struct_mutex);
+
+	i915_gem_object_lock(intel_fb->obj);
 	WARN_ON(!intel_fb->obj->framebuffer_references--);
+	i915_gem_object_unlock(intel_fb->obj);
+
 	i915_gem_object_put(intel_fb->obj);
-	mutex_unlock(&dev->struct_mutex);
+
 	kfree(intel_fb);
 }
 
@@ -15862,15 +14258,10 @@ static int intel_user_framebuffer_dirty(struct drm_framebuffer *fb,
 					struct drm_clip_rect *clips,
 					unsigned num_clips)
 {
-	struct drm_device *dev = fb->dev;
-	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
-	struct drm_i915_gem_object *obj = intel_fb->obj;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 
-	mutex_lock(&dev->struct_mutex);
-	if (obj->pin_display && obj->cache_dirty)
-		i915_gem_clflush_object(obj, true);
-	intel_fb_obj_flush(obj, false, ORIGIN_DIRTYFB);
-	mutex_unlock(&dev->struct_mutex);
+	i915_gem_object_flush_if_display(obj);
+	intel_fb_obj_flush(obj, ORIGIN_DIRTYFB);
 
 	return 0;
 }
@@ -15885,7 +14276,7 @@ static
 u32 intel_fb_pitch_limit(struct drm_i915_private *dev_priv,
 			 uint64_t fb_modifier, uint32_t pixel_format)
 {
-	u32 gen = INTEL_INFO(dev_priv)->gen;
+	u32 gen = INTEL_GEN(dev_priv);
 
 	if (gen >= 9) {
 		int cpp = drm_format_plane_cpp(pixel_format, 0);
@@ -15894,8 +14285,7 @@ u32 intel_fb_pitch_limit(struct drm_i915_private *dev_priv,
 		 *  pixels and 32K bytes."
 		 */
 		return min(8192 * cpp, 32768);
-	} else if (gen >= 5 && !IS_VALLEYVIEW(dev_priv) &&
-		   !IS_CHERRYVIEW(dev_priv)) {
+	} else if (gen >= 5 && !HAS_GMCH_DISPLAY(dev_priv)) {
 		return 32*1024;
 	} else if (gen >= 4) {
 		if (fb_modifier == I915_FORMAT_MOD_X_TILED)
@@ -15913,18 +14303,21 @@ u32 intel_fb_pitch_limit(struct drm_i915_private *dev_priv,
 	}
 }
 
-static int intel_framebuffer_init(struct drm_device *dev,
-				  struct intel_framebuffer *intel_fb,
-				  struct drm_mode_fb_cmd2 *mode_cmd,
-				  struct drm_i915_gem_object *obj)
+static int intel_framebuffer_init(struct intel_framebuffer *intel_fb,
+				  struct drm_i915_gem_object *obj,
+				  struct drm_mode_fb_cmd2 *mode_cmd)
 {
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	unsigned int tiling = i915_gem_object_get_tiling(obj);
-	int ret;
-	u32 pitch_limit, stride_alignment;
+	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 	struct drm_format_name_buf format_name;
+	u32 pitch_limit, stride_alignment;
+	unsigned int tiling, stride;
+	int ret = -EINVAL;
 
-	WARN_ON(!mutex_is_locked(&dev->struct_mutex));
+	i915_gem_object_lock(obj);
+	obj->framebuffer_references++;
+	tiling = i915_gem_object_get_tiling(obj);
+	stride = i915_gem_object_get_stride(obj);
+	i915_gem_object_unlock(obj);
 
 	if (mode_cmd->flags & DRM_MODE_FB_MODIFIERS) {
 		/*
@@ -15934,14 +14327,14 @@ static int intel_framebuffer_init(struct drm_device *dev,
 		if (tiling != I915_TILING_NONE &&
 		    tiling != intel_fb_modifier_to_tiling(mode_cmd->modifier[0])) {
 			DRM_DEBUG("tiling_mode doesn't match fb modifier\n");
-			return -EINVAL;
+			goto err;
 		}
 	} else {
 		if (tiling == I915_TILING_X) {
 			mode_cmd->modifier[0] = I915_FORMAT_MOD_X_TILED;
 		} else if (tiling == I915_TILING_Y) {
 			DRM_DEBUG("No Y tiling for legacy addfb\n");
-			return -EINVAL;
+			goto err;
 		}
 	}
 
@@ -15952,7 +14345,7 @@ static int intel_framebuffer_init(struct drm_device *dev,
 		if (INTEL_GEN(dev_priv) < 9) {
 			DRM_DEBUG("Unsupported tiling 0x%llx!\n",
 				  mode_cmd->modifier[0]);
-			return -EINVAL;
+			goto err;
 		}
 	case DRM_FORMAT_MOD_NONE:
 	case I915_FORMAT_MOD_X_TILED:
@@ -15960,7 +14353,7 @@ static int intel_framebuffer_init(struct drm_device *dev,
 	default:
 		DRM_DEBUG("Unsupported fb modifier 0x%llx!\n",
 			  mode_cmd->modifier[0]);
-		return -EINVAL;
+		goto err;
 	}
 
 	/*
@@ -15970,7 +14363,7 @@ static int intel_framebuffer_init(struct drm_device *dev,
 	if (INTEL_INFO(dev_priv)->gen < 4 &&
 	    tiling != intel_fb_modifier_to_tiling(mode_cmd->modifier[0])) {
 		DRM_DEBUG("tiling_mode must match fb modifier exactly on gen2/3\n");
-		return -EINVAL;
+		goto err;
 	}
 
 	stride_alignment = intel_fb_stride_alignment(dev_priv,
@@ -15979,7 +14372,7 @@ static int intel_framebuffer_init(struct drm_device *dev,
 	if (mode_cmd->pitches[0] & (stride_alignment - 1)) {
 		DRM_DEBUG("pitch (%d) must be at least %u byte aligned\n",
 			  mode_cmd->pitches[0], stride_alignment);
-		return -EINVAL;
+		goto err;
 	}
 
 	pitch_limit = intel_fb_pitch_limit(dev_priv, mode_cmd->modifier[0],
@@ -15989,19 +14382,17 @@ static int intel_framebuffer_init(struct drm_device *dev,
 			  mode_cmd->modifier[0] != DRM_FORMAT_MOD_NONE ?
 			  "tiled" : "linear",
 			  mode_cmd->pitches[0], pitch_limit);
-		return -EINVAL;
+		goto err;
 	}
 
 	/*
 	 * If there's a fence, enforce that
 	 * the fb pitch and fence stride match.
 	 */
-	if (tiling != I915_TILING_NONE &&
-	    mode_cmd->pitches[0] != i915_gem_object_get_stride(obj)) {
+	if (tiling != I915_TILING_NONE && mode_cmd->pitches[0] !=  stride) {
 		DRM_DEBUG("pitch (%d) must match tiling stride (%d)\n",
-			  mode_cmd->pitches[0],
-			  i915_gem_object_get_stride(obj));
-		return -EINVAL;
+			  mode_cmd->pitches[0], stride);
+		goto err;
 	}
 
 	/* Reject formats not supported by any plane early. */
@@ -16015,7 +14406,7 @@ static int intel_framebuffer_init(struct drm_device *dev,
 		if (INTEL_GEN(dev_priv) > 3) {
 			DRM_DEBUG("unsupported pixel format: %s\n",
 			          drm_get_format_name(mode_cmd->pixel_format, &format_name));
-			return -EINVAL;
+			goto err;
 		}
 		break;
 	case DRM_FORMAT_ABGR8888:
@@ -16023,7 +14414,7 @@ static int intel_framebuffer_init(struct drm_device *dev,
 		    INTEL_GEN(dev_priv) < 9) {
 			DRM_DEBUG("unsupported pixel format: %s\n",
 			          drm_get_format_name(mode_cmd->pixel_format, &format_name));
-			return -EINVAL;
+			goto err;
 		}
 		break;
 	case DRM_FORMAT_XBGR8888:
@@ -16032,14 +14423,14 @@ static int intel_framebuffer_init(struct drm_device *dev,
 		if (INTEL_GEN(dev_priv) < 4) {
 			DRM_DEBUG("unsupported pixel format: %s\n",
 			          drm_get_format_name(mode_cmd->pixel_format, &format_name));
-			return -EINVAL;
+			goto err;
 		}
 		break;
 	case DRM_FORMAT_ABGR2101010:
 		if (!IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv)) {
 			DRM_DEBUG("unsupported pixel format: %s\n",
 			          drm_get_format_name(mode_cmd->pixel_format, &format_name));
-			return -EINVAL;
+			goto err;
 		}
 		break;
 	case DRM_FORMAT_YUYV:
@@ -16049,35 +14440,42 @@ static int intel_framebuffer_init(struct drm_device *dev,
 		if (INTEL_GEN(dev_priv) < 5) {
 			DRM_DEBUG("unsupported pixel format: %s\n",
 			          drm_get_format_name(mode_cmd->pixel_format, &format_name));
-			return -EINVAL;
+			goto err;
 		}
 		break;
 	default:
 		DRM_DEBUG("unsupported pixel format: %s\n",
 		          drm_get_format_name(mode_cmd->pixel_format, &format_name));
-		return -EINVAL;
+		goto err;
 	}
 
 	/* FIXME need to adjust LINOFF/TILEOFF accordingly. */
 	if (mode_cmd->offsets[0] != 0)
-		return -EINVAL;
+		goto err;
 
-	drm_helper_mode_fill_fb_struct(dev, &intel_fb->base, mode_cmd);
+	drm_helper_mode_fill_fb_struct(&dev_priv->drm,
+				       &intel_fb->base, mode_cmd);
 	intel_fb->obj = obj;
 
 	ret = intel_fill_fb_info(dev_priv, &intel_fb->base);
 	if (ret)
-		return ret;
+		goto err;
 
-	ret = drm_framebuffer_init(dev, &intel_fb->base, &intel_fb_funcs);
+	ret = drm_framebuffer_init(obj->base.dev,
+				   &intel_fb->base,
+				   &intel_fb_funcs);
 	if (ret) {
 		DRM_ERROR("framebuffer init failed %d\n", ret);
-		return ret;
+		goto err;
 	}
 
-	intel_fb->obj->framebuffer_references++;
-
 	return 0;
+
+err:
+	i915_gem_object_lock(obj);
+	obj->framebuffer_references--;
+	i915_gem_object_unlock(obj);
+	return ret;
 }
 
 static struct drm_framebuffer *
@@ -16093,7 +14491,7 @@ intel_user_framebuffer_create(struct drm_device *dev,
 	if (!obj)
 		return ERR_PTR(-ENOENT);
 
-	fb = intel_framebuffer_create(dev, &mode_cmd, obj);
+	fb = intel_framebuffer_create(obj, &mode_cmd);
 	if (IS_ERR(fb))
 		i915_gem_object_put(obj);
 
@@ -16127,6 +14525,8 @@ static const struct drm_mode_config_funcs intel_mode_funcs = {
  */
 void intel_init_display_hooks(struct drm_i915_private *dev_priv)
 {
+	intel_init_cdclk_hooks(dev_priv);
+
 	if (INTEL_INFO(dev_priv)->gen >= 9) {
 		dev_priv->display.get_pipe_config = haswell_get_pipe_config;
 		dev_priv->display.get_initial_plane_config =
@@ -16195,62 +14595,6 @@ void intel_init_display_hooks(struct drm_i915_private *dev_priv)
 		dev_priv->display.crtc_disable = i9xx_crtc_disable;
 	}
 
-	/* Returns the core display clock speed */
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			skylake_get_display_clock_speed;
-	else if (IS_GEN9_LP(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			broxton_get_display_clock_speed;
-	else if (IS_BROADWELL(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			broadwell_get_display_clock_speed;
-	else if (IS_HASWELL(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			haswell_get_display_clock_speed;
-	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			valleyview_get_display_clock_speed;
-	else if (IS_GEN5(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			ilk_get_display_clock_speed;
-	else if (IS_I945G(dev_priv) || IS_I965G(dev_priv) ||
-		 IS_GEN6(dev_priv) || IS_IVYBRIDGE(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			i945_get_display_clock_speed;
-	else if (IS_GM45(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			gm45_get_display_clock_speed;
-	else if (IS_I965GM(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			i965gm_get_display_clock_speed;
-	else if (IS_PINEVIEW(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			pnv_get_display_clock_speed;
-	else if (IS_G33(dev_priv) || IS_G4X(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			g33_get_display_clock_speed;
-	else if (IS_I915G(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			i915_get_display_clock_speed;
-	else if (IS_I945GM(dev_priv) || IS_I845G(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			i9xx_misc_get_display_clock_speed;
-	else if (IS_I915GM(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			i915gm_get_display_clock_speed;
-	else if (IS_I865G(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			i865_get_display_clock_speed;
-	else if (IS_I85X(dev_priv))
-		dev_priv->display.get_display_clock_speed =
-			i85x_get_display_clock_speed;
-	else { /* 830 */
-		WARN(!IS_I830(dev_priv), "Unknown platform. Assuming 133 MHz CDCLK\n");
-		dev_priv->display.get_display_clock_speed =
-			i830_get_display_clock_speed;
-	}
-
 	if (IS_GEN5(dev_priv)) {
 		dev_priv->display.fdi_link_train = ironlake_fdi_link_train;
 	} else if (IS_GEN6(dev_priv)) {
@@ -16262,28 +14606,6 @@ void intel_init_display_hooks(struct drm_i915_private *dev_priv)
 		dev_priv->display.fdi_link_train = hsw_fdi_link_train;
 	}
 
-	if (IS_BROADWELL(dev_priv)) {
-		dev_priv->display.modeset_commit_cdclk =
-			broadwell_modeset_commit_cdclk;
-		dev_priv->display.modeset_calc_cdclk =
-			broadwell_modeset_calc_cdclk;
-	} else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
-		dev_priv->display.modeset_commit_cdclk =
-			valleyview_modeset_commit_cdclk;
-		dev_priv->display.modeset_calc_cdclk =
-			valleyview_modeset_calc_cdclk;
-	} else if (IS_GEN9_LP(dev_priv)) {
-		dev_priv->display.modeset_commit_cdclk =
-			bxt_modeset_commit_cdclk;
-		dev_priv->display.modeset_calc_cdclk =
-			bxt_modeset_calc_cdclk;
-	} else if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
-		dev_priv->display.modeset_commit_cdclk =
-			skl_modeset_commit_cdclk;
-		dev_priv->display.modeset_calc_cdclk =
-			skl_modeset_calc_cdclk;
-	}
-
 	if (dev_priv->info.gen >= 9)
 		dev_priv->display.update_crtcs = skl_update_crtcs;
 	else
@@ -16510,8 +14832,7 @@ void intel_modeset_init_hw(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	intel_update_cdclk(dev_priv);
-
-	dev_priv->atomic_cdclk_freq = dev_priv->cdclk_freq;
+	dev_priv->cdclk.logical = dev_priv->cdclk.actual = dev_priv->cdclk.hw;
 
 	intel_init_clock_gating(dev_priv);
 }
@@ -16566,7 +14887,8 @@ retry:
 	 * intermediate watermarks (since we don't trust the current
 	 * watermarks).
 	 */
-	intel_state->skip_intermediate_wm = true;
+	if (!HAS_GMCH_DISPLAY(dev_priv))
+		intel_state->skip_intermediate_wm = true;
 
 	ret = intel_atomic_check(dev, state);
 	if (ret) {
@@ -16600,18 +14922,6 @@ fail:
 	drm_modeset_acquire_fini(&ctx);
 }
 
-static void intel_atomic_helper_free_state(struct work_struct *work)
-{
-	struct drm_i915_private *dev_priv =
-		container_of(work, typeof(*dev_priv), atomic_helper.free_work);
-	struct intel_atomic_state *state, *next;
-	struct llist_node *freed;
-
-	freed = llist_del_all(&dev_priv->atomic_helper.free_list);
-	llist_for_each_entry_safe(state, next, freed, freed)
-		drm_atomic_state_put(&state->base);
-}
-
 int intel_modeset_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = to_i915(dev);
@@ -16632,7 +14942,7 @@ int intel_modeset_init(struct drm_device *dev)
 	dev->mode_config.funcs = &intel_mode_funcs;
 
 	INIT_WORK(&dev_priv->atomic_helper.free_work,
-		  intel_atomic_helper_free_state);
+		  intel_atomic_helper_free_state_worker);
 
 	intel_init_quirks(dev);
 
@@ -16697,12 +15007,11 @@ int intel_modeset_init(struct drm_device *dev)
 		}
 	}
 
-	intel_update_czclk(dev_priv);
-	intel_update_cdclk(dev_priv);
-	dev_priv->atomic_cdclk_freq = dev_priv->cdclk_freq;
-
 	intel_shared_dpll_init(dev);
 
+	intel_update_czclk(dev_priv);
+	intel_modeset_init_hw(dev);
+
 	if (dev_priv->max_cdclk_freq == 0)
 		intel_update_max_cdclk(dev_priv);
 
@@ -16742,7 +15051,8 @@ int intel_modeset_init(struct drm_device *dev)
 	 * Note that we need to do this after reconstructing the BIOS fb's
 	 * since the watermark calculation done here will use pstate->fb.
 	 */
-	sanitize_watermarks(dev);
+	if (!HAS_GMCH_DISPLAY(dev_priv))
+		sanitize_watermarks(dev);
 
 	return 0;
 }
@@ -16844,6 +15154,7 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc)
 			if (plane->base.type == DRM_PLANE_TYPE_PRIMARY)
 				continue;
 
+			trace_intel_disable_plane(&plane->base, crtc);
 			plane->disable_plane(&plane->base, &crtc->base);
 		}
 	}
@@ -16990,15 +15301,14 @@ static bool primary_get_hw_state(struct intel_plane *plane)
 /* FIXME read out full plane state for all planes */
 static void readout_plane_state(struct intel_crtc *crtc)
 {
-	struct drm_plane *primary = crtc->base.primary;
-	struct intel_plane_state *plane_state =
-		to_intel_plane_state(primary->state);
+	struct intel_plane *primary = to_intel_plane(crtc->base.primary);
+	bool visible;
 
-	plane_state->base.visible = crtc->active &&
-		primary_get_hw_state(to_intel_plane(primary));
+	visible = crtc->active && primary_get_hw_state(primary);
 
-	if (plane_state->base.visible)
-		crtc->base.state->plane_mask |= 1 << drm_plane_index(primary);
+	intel_set_plane_visible(to_intel_crtc_state(crtc->base.state),
+				to_intel_plane_state(primary->base.state),
+				visible);
 }
 
 static void intel_modeset_readout_hw_state(struct drm_device *dev)
@@ -17131,10 +15441,11 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev)
 			 */
 			crtc_state->base.mode.private_flags = I915_MODE_FLAG_INHERITED;
 
-			if (INTEL_GEN(dev_priv) >= 9 || IS_BROADWELL(dev_priv))
-				pixclk = ilk_pipe_pixel_rate(crtc_state);
-			else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-				pixclk = crtc_state->base.adjusted_mode.crtc_clock;
+			intel_crtc_compute_pixel_rate(crtc_state);
+
+			if (INTEL_GEN(dev_priv) >= 9 || IS_BROADWELL(dev_priv) ||
+			    IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+				pixclk = crtc_state->pixel_rate;
 			else
 				WARN_ON(dev_priv->display.modeset_calc_cdclk);
 
@@ -17152,6 +15463,24 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev)
 	}
 }
 
+static void
+get_encoder_power_domains(struct drm_i915_private *dev_priv)
+{
+	struct intel_encoder *encoder;
+
+	for_each_intel_encoder(&dev_priv->drm, encoder) {
+		u64 get_domains;
+		enum intel_display_power_domain domain;
+
+		if (!encoder->get_power_domains)
+			continue;
+
+		get_domains = encoder->get_power_domains(encoder);
+		for_each_power_domain(domain, get_domains)
+			intel_display_power_get(dev_priv, domain);
+	}
+}
+
 /* Scan out the current hw modeset state,
  * and sanitizes it to the current state
  */
@@ -17167,6 +15496,8 @@ intel_modeset_setup_hw_state(struct drm_device *dev)
 	intel_modeset_readout_hw_state(dev);
 
 	/* HW state is read out, now we need to sanitize this mess. */
+	get_encoder_power_domains(dev_priv);
+
 	for_each_intel_encoder(dev, encoder) {
 		intel_sanitize_encoder(encoder);
 	}
@@ -17193,15 +15524,17 @@ intel_modeset_setup_hw_state(struct drm_device *dev)
 		pll->on = false;
 	}
 
-	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
 		vlv_wm_get_hw_state(dev);
-	else if (IS_GEN9(dev_priv))
+		vlv_wm_sanitize(dev_priv);
+	} else if (IS_GEN9(dev_priv)) {
 		skl_wm_get_hw_state(dev);
-	else if (HAS_PCH_SPLIT(dev_priv))
+	} else if (HAS_PCH_SPLIT(dev_priv)) {
 		ilk_wm_get_hw_state(dev);
+	}
 
 	for_each_intel_crtc(dev, crtc) {
-		unsigned long put_domains;
+		u64 put_domains;
 
 		put_domains = modeset_get_crtc_power_domains(&crtc->base, crtc->config);
 		if (WARN_ON(put_domains))
@@ -17209,6 +15542,8 @@ intel_modeset_setup_hw_state(struct drm_device *dev)
 	}
 	intel_display_set_init_power(dev_priv, false);
 
+	intel_power_domains_verify_state(dev_priv);
+
 	intel_fbc_init_pipe_state(dev_priv);
 }
 
@@ -17259,8 +15594,6 @@ void intel_modeset_gem_init(struct drm_device *dev)
 
 	intel_init_gt_powersave(dev_priv);
 
-	intel_modeset_init_hw(dev);
-
 	intel_setup_overlay(dev_priv);
 }
 
@@ -17492,9 +15825,9 @@ intel_display_capture_error_state(struct drm_i915_private *dev_priv)
 
 void
 intel_display_print_error_state(struct drm_i915_error_state_buf *m,
-				struct drm_i915_private *dev_priv,
 				struct intel_display_error_state *error)
 {
+	struct drm_i915_private *dev_priv = m->i915;
 	int i;
 
 	if (!error)
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index d1670b8afbf5..fd96a6cf7326 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -28,8 +28,10 @@
 #include <linux/i2c.h>
 #include <linux/slab.h>
 #include <linux/export.h>
+#include <linux/types.h>
 #include <linux/notifier.h>
 #include <linux/reboot.h>
+#include <asm/byteorder.h>
 #include <drm/drmP.h>
 #include <drm/drm_atomic_helper.h>
 #include <drm/drm_crtc.h>
@@ -226,7 +228,7 @@ intel_dp_source_rates(struct intel_dp *intel_dp, const int **source_rates)
 	if (IS_GEN9_LP(dev_priv)) {
 		*source_rates = bxt_rates;
 		size = ARRAY_SIZE(bxt_rates);
-	} else if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	} else if (IS_GEN9_BC(dev_priv)) {
 		*source_rates = skl_rates;
 		size = ARRAY_SIZE(skl_rates);
 	} else {
@@ -394,14 +396,12 @@ static void pps_lock(struct intel_dp *intel_dp)
 	struct intel_encoder *encoder = &intel_dig_port->base;
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum intel_display_power_domain power_domain;
 
 	/*
 	 * See vlv_power_sequencer_reset() why we need
 	 * a power domain reference here.
 	 */
-	power_domain = intel_display_port_aux_power_domain(encoder);
-	intel_display_power_get(dev_priv, power_domain);
+	intel_display_power_get(dev_priv, intel_dp->aux_power_domain);
 
 	mutex_lock(&dev_priv->pps_mutex);
 }
@@ -412,12 +412,10 @@ static void pps_unlock(struct intel_dp *intel_dp)
 	struct intel_encoder *encoder = &intel_dig_port->base;
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum intel_display_power_domain power_domain;
 
 	mutex_unlock(&dev_priv->pps_mutex);
 
-	power_domain = intel_display_port_aux_power_domain(encoder);
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, intel_dp->aux_power_domain);
 }
 
 static void
@@ -916,7 +914,7 @@ static uint32_t ilk_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
 	 * divide by 2000 and use that
 	 */
 	if (intel_dig_port->port == PORT_A)
-		return DIV_ROUND_CLOSEST(dev_priv->cdclk_freq, 2000);
+		return DIV_ROUND_CLOSEST(dev_priv->cdclk.hw.cdclk, 2000);
 	else
 		return DIV_ROUND_CLOSEST(dev_priv->rawclk_freq, 2000);
 }
@@ -1593,6 +1591,13 @@ static int intel_dp_compute_bpp(struct intel_dp *intel_dp,
 	if (bpc > 0)
 		bpp = min(bpp, 3*bpc);
 
+	/* For DP Compliance we override the computed bpp for the pipe */
+	if (intel_dp->compliance.test_data.bpc != 0) {
+		pipe_config->pipe_bpp =	3*intel_dp->compliance.test_data.bpc;
+		pipe_config->dither_force_disable = pipe_config->pipe_bpp == 6*3;
+		DRM_DEBUG_KMS("Setting pipe_bpp to %d\n",
+			      pipe_config->pipe_bpp);
+	}
 	return bpp;
 }
 
@@ -1613,6 +1618,7 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 	/* Conveniently, the link BW constants become indices with a shift...*/
 	int min_clock = 0;
 	int max_clock;
+	int link_rate_index;
 	int bpp, mode_rate;
 	int link_avail, link_clock;
 	int common_rates[DP_MAX_SUPPORTED_RATES] = {};
@@ -1654,6 +1660,15 @@ intel_dp_compute_config(struct intel_encoder *encoder,
 	if (adjusted_mode->flags & DRM_MODE_FLAG_DBLCLK)
 		return false;
 
+	/* Use values requested by Compliance Test Request */
+	if (intel_dp->compliance.test_type == DP_TEST_LINK_TRAINING) {
+		link_rate_index = intel_dp_link_rate_index(intel_dp,
+							   common_rates,
+							   intel_dp->compliance.test_link_rate);
+		if (link_rate_index >= 0)
+			min_clock = max_clock = link_rate_index;
+		min_lane_count = max_lane_count = intel_dp->compliance.test_lane_count;
+	}
 	DRM_DEBUG_KMS("DP link computation with max lane count %i "
 		      "max bw %d pixel clock %iKHz\n",
 		      max_lane_count, common_rates[max_clock],
@@ -1753,8 +1768,7 @@ found:
 	 * DPLL0 VCO may need to be adjusted to get the correct
 	 * clock for eDP. This will affect cdclk as well.
 	 */
-	if (is_edp(intel_dp) &&
-	    (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))) {
+	if (is_edp(intel_dp) && IS_GEN9_BC(dev_priv)) {
 		int vco;
 
 		switch (pipe_config->port_clock / 2) {
@@ -1767,7 +1781,7 @@ found:
 			break;
 		}
 
-		to_intel_atomic_state(pipe_config->base.state)->cdclk_pll_vco = vco;
+		to_intel_atomic_state(pipe_config->base.state)->cdclk.logical.vco = vco;
 	}
 
 	if (!HAS_DDI(dev_priv))
@@ -1987,9 +2001,7 @@ static bool edp_panel_vdd_on(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum intel_display_power_domain power_domain;
 	u32 pp;
 	i915_reg_t pp_stat_reg, pp_ctrl_reg;
 	bool need_to_disable = !intel_dp->want_panel_vdd;
@@ -2005,8 +2017,7 @@ static bool edp_panel_vdd_on(struct intel_dp *intel_dp)
 	if (edp_have_panel_vdd(intel_dp))
 		return need_to_disable;
 
-	power_domain = intel_display_port_aux_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
+	intel_display_power_get(dev_priv, intel_dp->aux_power_domain);
 
 	DRM_DEBUG_KMS("Turning eDP port %c VDD on\n",
 		      port_name(intel_dig_port->port));
@@ -2064,8 +2075,6 @@ static void edp_panel_vdd_off_sync(struct intel_dp *intel_dp)
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_digital_port *intel_dig_port =
 		dp_to_dig_port(intel_dp);
-	struct intel_encoder *intel_encoder = &intel_dig_port->base;
-	enum intel_display_power_domain power_domain;
 	u32 pp;
 	i915_reg_t pp_stat_reg, pp_ctrl_reg;
 
@@ -2095,8 +2104,7 @@ static void edp_panel_vdd_off_sync(struct intel_dp *intel_dp)
 	if ((pp & PANEL_POWER_ON) == 0)
 		intel_dp->panel_power_off_time = ktime_get_boottime();
 
-	power_domain = intel_display_port_aux_power_domain(intel_encoder);
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, intel_dp->aux_power_domain);
 }
 
 static void edp_panel_vdd_work(struct work_struct *__work)
@@ -2209,11 +2217,8 @@ void intel_edp_panel_on(struct intel_dp *intel_dp)
 
 static void edp_panel_off(struct intel_dp *intel_dp)
 {
-	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
-	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum intel_display_power_domain power_domain;
 	u32 pp;
 	i915_reg_t pp_ctrl_reg;
 
@@ -2245,8 +2250,7 @@ static void edp_panel_off(struct intel_dp *intel_dp)
 	wait_panel_off(intel_dp);
 
 	/* We got a reference when we enabled the VDD. */
-	power_domain = intel_display_port_aux_power_domain(intel_encoder);
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, intel_dp->aux_power_domain);
 }
 
 void intel_edp_panel_off(struct intel_dp *intel_dp)
@@ -2492,12 +2496,11 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 	enum port port = dp_to_dig_port(intel_dp)->port;
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum intel_display_power_domain power_domain;
 	u32 tmp;
 	bool ret;
 
-	power_domain = intel_display_port_power_domain(encoder);
-	if (!intel_display_power_get_if_enabled(dev_priv, power_domain))
+	if (!intel_display_power_get_if_enabled(dev_priv,
+						encoder->power_domain))
 		return false;
 
 	ret = false;
@@ -2533,7 +2536,7 @@ static bool intel_dp_get_hw_state(struct intel_encoder *encoder,
 	ret = true;
 
 out:
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, encoder->power_domain);
 
 	return ret;
 }
@@ -3080,9 +3083,8 @@ intel_dp_voltage_max(struct intel_dp *intel_dp)
 	if (IS_GEN9_LP(dev_priv))
 		return DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
 	else if (INTEL_GEN(dev_priv) >= 9) {
-		if (dev_priv->vbt.edp.low_vswing && port == PORT_A)
-			return DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
-		return DP_TRAIN_VOLTAGE_SWING_LEVEL_2;
+		struct intel_encoder *encoder = &dp_to_dig_port(intel_dp)->base;
+		return intel_ddi_dp_voltage_max(encoder);
 	} else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
 		return DP_TRAIN_VOLTAGE_SWING_LEVEL_3;
 	else if (IS_GEN7(dev_priv) && port == PORT_A)
@@ -3922,19 +3924,112 @@ intel_dp_get_sink_irq_esi(struct intel_dp *intel_dp, u8 *sink_irq_vector)
 
 static uint8_t intel_dp_autotest_link_training(struct intel_dp *intel_dp)
 {
-	uint8_t test_result = DP_TEST_ACK;
-	return test_result;
+	int status = 0;
+	int min_lane_count = 1;
+	int common_rates[DP_MAX_SUPPORTED_RATES] = {};
+	int link_rate_index, test_link_rate;
+	uint8_t test_lane_count, test_link_bw;
+	/* (DP CTS 1.2)
+	 * 4.3.1.11
+	 */
+	/* Read the TEST_LANE_COUNT and TEST_LINK_RTAE fields (DP CTS 3.1.4) */
+	status = drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_LANE_COUNT,
+				   &test_lane_count);
+
+	if (status <= 0) {
+		DRM_DEBUG_KMS("Lane count read failed\n");
+		return DP_TEST_NAK;
+	}
+	test_lane_count &= DP_MAX_LANE_COUNT_MASK;
+	/* Validate the requested lane count */
+	if (test_lane_count < min_lane_count ||
+	    test_lane_count > intel_dp->max_sink_lane_count)
+		return DP_TEST_NAK;
+
+	status = drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_LINK_RATE,
+				   &test_link_bw);
+	if (status <= 0) {
+		DRM_DEBUG_KMS("Link Rate read failed\n");
+		return DP_TEST_NAK;
+	}
+	/* Validate the requested link rate */
+	test_link_rate = drm_dp_bw_code_to_link_rate(test_link_bw);
+	link_rate_index = intel_dp_link_rate_index(intel_dp,
+						   common_rates,
+						   test_link_rate);
+	if (link_rate_index < 0)
+		return DP_TEST_NAK;
+
+	intel_dp->compliance.test_lane_count = test_lane_count;
+	intel_dp->compliance.test_link_rate = test_link_rate;
+
+	return DP_TEST_ACK;
 }
 
 static uint8_t intel_dp_autotest_video_pattern(struct intel_dp *intel_dp)
 {
-	uint8_t test_result = DP_TEST_NAK;
-	return test_result;
+	uint8_t test_pattern;
+	uint16_t test_misc;
+	__be16 h_width, v_height;
+	int status = 0;
+
+	/* Read the TEST_PATTERN (DP CTS 3.1.5) */
+	status = drm_dp_dpcd_read(&intel_dp->aux, DP_TEST_PATTERN,
+				  &test_pattern, 1);
+	if (status <= 0) {
+		DRM_DEBUG_KMS("Test pattern read failed\n");
+		return DP_TEST_NAK;
+	}
+	if (test_pattern != DP_COLOR_RAMP)
+		return DP_TEST_NAK;
+
+	status = drm_dp_dpcd_read(&intel_dp->aux, DP_TEST_H_WIDTH_HI,
+				  &h_width, 2);
+	if (status <= 0) {
+		DRM_DEBUG_KMS("H Width read failed\n");
+		return DP_TEST_NAK;
+	}
+
+	status = drm_dp_dpcd_read(&intel_dp->aux, DP_TEST_V_HEIGHT_HI,
+				  &v_height, 2);
+	if (status <= 0) {
+		DRM_DEBUG_KMS("V Height read failed\n");
+		return DP_TEST_NAK;
+	}
+
+	status = drm_dp_dpcd_read(&intel_dp->aux, DP_TEST_MISC0,
+				  &test_misc, 1);
+	if (status <= 0) {
+		DRM_DEBUG_KMS("TEST MISC read failed\n");
+		return DP_TEST_NAK;
+	}
+	if ((test_misc & DP_TEST_COLOR_FORMAT_MASK) != DP_COLOR_FORMAT_RGB)
+		return DP_TEST_NAK;
+	if (test_misc & DP_TEST_DYNAMIC_RANGE_CEA)
+		return DP_TEST_NAK;
+	switch (test_misc & DP_TEST_BIT_DEPTH_MASK) {
+	case DP_TEST_BIT_DEPTH_6:
+		intel_dp->compliance.test_data.bpc = 6;
+		break;
+	case DP_TEST_BIT_DEPTH_8:
+		intel_dp->compliance.test_data.bpc = 8;
+		break;
+	default:
+		return DP_TEST_NAK;
+	}
+
+	intel_dp->compliance.test_data.video_pattern = test_pattern;
+	intel_dp->compliance.test_data.hdisplay = be16_to_cpu(h_width);
+	intel_dp->compliance.test_data.vdisplay = be16_to_cpu(v_height);
+	/* Set test active flag here so userspace doesn't interrupt things */
+	intel_dp->compliance.test_active = 1;
+
+	return DP_TEST_ACK;
 }
 
 static uint8_t intel_dp_autotest_edid(struct intel_dp *intel_dp)
 {
-	uint8_t test_result = DP_TEST_NAK;
+	uint8_t test_result = DP_TEST_ACK;
 	struct intel_connector *intel_connector = intel_dp->attached_connector;
 	struct drm_connector *connector = &intel_connector->base;
 
@@ -3969,7 +4064,7 @@ static uint8_t intel_dp_autotest_edid(struct intel_dp *intel_dp)
 			DRM_DEBUG_KMS("Failed to write EDID checksum\n");
 
 		test_result = DP_TEST_ACK | DP_TEST_EDID_CHECKSUM_WRITE;
-		intel_dp->compliance.test_data.edid = INTEL_DP_RESOLUTION_STANDARD;
+		intel_dp->compliance.test_data.edid = INTEL_DP_RESOLUTION_PREFERRED;
 	}
 
 	/* Set test active flag here so userspace doesn't interrupt things */
@@ -3987,45 +4082,42 @@ static uint8_t intel_dp_autotest_phy_pattern(struct intel_dp *intel_dp)
 static void intel_dp_handle_test_request(struct intel_dp *intel_dp)
 {
 	uint8_t response = DP_TEST_NAK;
-	uint8_t rxdata = 0;
-	int status = 0;
+	uint8_t request = 0;
+	int status;
 
-	status = drm_dp_dpcd_read(&intel_dp->aux, DP_TEST_REQUEST, &rxdata, 1);
+	status = drm_dp_dpcd_readb(&intel_dp->aux, DP_TEST_REQUEST, &request);
 	if (status <= 0) {
 		DRM_DEBUG_KMS("Could not read test request from sink\n");
 		goto update_status;
 	}
 
-	switch (rxdata) {
+	switch (request) {
 	case DP_TEST_LINK_TRAINING:
 		DRM_DEBUG_KMS("LINK_TRAINING test requested\n");
-		intel_dp->compliance.test_type = DP_TEST_LINK_TRAINING;
 		response = intel_dp_autotest_link_training(intel_dp);
 		break;
 	case DP_TEST_LINK_VIDEO_PATTERN:
 		DRM_DEBUG_KMS("TEST_PATTERN test requested\n");
-		intel_dp->compliance.test_type = DP_TEST_LINK_VIDEO_PATTERN;
 		response = intel_dp_autotest_video_pattern(intel_dp);
 		break;
 	case DP_TEST_LINK_EDID_READ:
 		DRM_DEBUG_KMS("EDID test requested\n");
-		intel_dp->compliance.test_type = DP_TEST_LINK_EDID_READ;
 		response = intel_dp_autotest_edid(intel_dp);
 		break;
 	case DP_TEST_LINK_PHY_TEST_PATTERN:
 		DRM_DEBUG_KMS("PHY_PATTERN test requested\n");
-		intel_dp->compliance.test_type = DP_TEST_LINK_PHY_TEST_PATTERN;
 		response = intel_dp_autotest_phy_pattern(intel_dp);
 		break;
 	default:
-		DRM_DEBUG_KMS("Invalid test request '%02x'\n", rxdata);
+		DRM_DEBUG_KMS("Invalid test request '%02x'\n", request);
 		break;
 	}
 
+	if (response & DP_TEST_ACK)
+		intel_dp->compliance.test_type = request;
+
 update_status:
-	status = drm_dp_dpcd_write(&intel_dp->aux,
-				   DP_TEST_RESPONSE,
-				   &response, 1);
+	status = drm_dp_dpcd_writeb(&intel_dp->aux, DP_TEST_RESPONSE, response);
 	if (status <= 0)
 		DRM_DEBUG_KMS("Could not write test response to sink\n");
 }
@@ -4137,9 +4229,8 @@ intel_dp_check_link_status(struct intel_dp *intel_dp)
 	if (!intel_dp->lane_count)
 		return;
 
-	/* if link training is requested we should perform it always */
-	if ((intel_dp->compliance.test_type == DP_TEST_LINK_TRAINING) ||
-	    (!drm_dp_channel_eq_ok(link_status, intel_dp->lane_count))) {
+	/* Retrain if Channel EQ or CR not ok */
+	if (!drm_dp_channel_eq_ok(link_status, intel_dp->lane_count)) {
 		DRM_DEBUG_KMS("%s: channel EQ not ok, retraining\n",
 			      intel_encoder->base.name);
 
@@ -4164,6 +4255,7 @@ static bool
 intel_dp_short_pulse(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp_to_dev(intel_dp);
+	struct intel_encoder *intel_encoder = &dp_to_dig_port(intel_dp)->base;
 	u8 sink_irq_vector = 0;
 	u8 old_sink_count = intel_dp->sink_count;
 	bool ret;
@@ -4197,7 +4289,7 @@ intel_dp_short_pulse(struct intel_dp *intel_dp)
 				   sink_irq_vector);
 
 		if (sink_irq_vector & DP_AUTOMATED_TEST_REQUEST)
-			DRM_DEBUG_DRIVER("Test request in short pulse not handled\n");
+			intel_dp_handle_test_request(intel_dp);
 		if (sink_irq_vector & (DP_CP_IRQ | DP_SINK_SPECIFIC_IRQ))
 			DRM_DEBUG_DRIVER("CP or sink specific irq unhandled\n");
 	}
@@ -4205,6 +4297,11 @@ intel_dp_short_pulse(struct intel_dp *intel_dp)
 	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
 	intel_dp_check_link_status(intel_dp);
 	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+	if (intel_dp->compliance.test_type == DP_TEST_LINK_TRAINING) {
+		DRM_DEBUG_KMS("Link Training Compliance Test requested\n");
+		/* Send a Hotplug Uevent to userspace to start modeset */
+		drm_kms_helper_hotplug_event(intel_encoder->base.dev);
+	}
 
 	return true;
 }
@@ -4213,9 +4310,13 @@ intel_dp_short_pulse(struct intel_dp *intel_dp)
 static enum drm_connector_status
 intel_dp_detect_dpcd(struct intel_dp *intel_dp)
 {
+	struct intel_lspcon *lspcon = dp_to_lspcon(intel_dp);
 	uint8_t *dpcd = intel_dp->dpcd;
 	uint8_t type;
 
+	if (lspcon->active)
+		lspcon_resume(lspcon);
+
 	if (!intel_dp_get_dpcd(intel_dp))
 		return connector_status_disconnected;
 
@@ -4474,11 +4575,9 @@ intel_dp_long_pulse(struct intel_connector *intel_connector)
 	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_device *dev = connector->dev;
 	enum drm_connector_status status;
-	enum intel_display_power_domain power_domain;
 	u8 sink_irq_vector = 0;
 
-	power_domain = intel_display_port_aux_power_domain(intel_encoder);
-	intel_display_power_get(to_i915(dev), power_domain);
+	intel_display_power_get(to_i915(dev), intel_dp->aux_power_domain);
 
 	/* Can't disconnect eDP, but you can close the lid... */
 	if (is_edp(intel_dp))
@@ -4511,11 +4610,15 @@ intel_dp_long_pulse(struct intel_connector *intel_connector)
 		      yesno(intel_dp_source_supports_hbr2(intel_dp)),
 		      yesno(drm_dp_tps3_supported(intel_dp->dpcd)));
 
-	/* Set the max lane count for sink */
-	intel_dp->max_sink_lane_count = drm_dp_max_lane_count(intel_dp->dpcd);
+	if (intel_dp->reset_link_params) {
+		/* Set the max lane count for sink */
+		intel_dp->max_sink_lane_count = drm_dp_max_lane_count(intel_dp->dpcd);
+
+		/* Set the max link BW for sink */
+		intel_dp->max_sink_link_bw = intel_dp_max_link_bw(intel_dp);
 
-	/* Set the max link BW for sink */
-	intel_dp->max_sink_link_bw = intel_dp_max_link_bw(intel_dp);
+		intel_dp->reset_link_params = false;
+	}
 
 	intel_dp_print_rates(intel_dp);
 
@@ -4575,7 +4678,7 @@ out:
 	if (status != connector_status_connected && !intel_dp->is_mst)
 		intel_dp_unset_edid(intel_dp);
 
-	intel_display_power_put(to_i915(dev), power_domain);
+	intel_display_power_put(to_i915(dev), intel_dp->aux_power_domain);
 	return status;
 }
 
@@ -4603,7 +4706,6 @@ intel_dp_force(struct drm_connector *connector)
 	struct intel_dp *intel_dp = intel_attached_dp(connector);
 	struct intel_encoder *intel_encoder = &dp_to_dig_port(intel_dp)->base;
 	struct drm_i915_private *dev_priv = to_i915(intel_encoder->base.dev);
-	enum intel_display_power_domain power_domain;
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
 		      connector->base.id, connector->name);
@@ -4612,12 +4714,11 @@ intel_dp_force(struct drm_connector *connector)
 	if (connector->status != connector_status_connected)
 		return;
 
-	power_domain = intel_display_port_aux_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
+	intel_display_power_get(dev_priv, intel_dp->aux_power_domain);
 
 	intel_dp_set_edid(intel_dp);
 
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, intel_dp->aux_power_domain);
 
 	if (intel_encoder->type != INTEL_OUTPUT_EDP)
 		intel_encoder->type = INTEL_OUTPUT_DP;
@@ -4852,7 +4953,6 @@ static void intel_edp_panel_vdd_sanitize(struct intel_dp *intel_dp)
 	struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum intel_display_power_domain power_domain;
 
 	lockdep_assert_held(&dev_priv->pps_mutex);
 
@@ -4866,8 +4966,7 @@ static void intel_edp_panel_vdd_sanitize(struct intel_dp *intel_dp)
 	 * indefinitely.
 	 */
 	DRM_DEBUG_KMS("VDD left on by BIOS, adjusting state tracking\n");
-	power_domain = intel_display_port_aux_power_domain(&intel_dig_port->base);
-	intel_display_power_get(dev_priv, power_domain);
+	intel_display_power_get(dev_priv, intel_dp->aux_power_domain);
 
 	edp_panel_vdd_schedule_off(intel_dp);
 }
@@ -4897,6 +4996,8 @@ void intel_dp_encoder_reset(struct drm_encoder *encoder)
 	if (lspcon->active)
 		lspcon_resume(lspcon);
 
+	intel_dp->reset_link_params = true;
+
 	pps_lock(intel_dp);
 
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
@@ -4939,10 +5040,8 @@ enum irqreturn
 intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd)
 {
 	struct intel_dp *intel_dp = &intel_dig_port->dp;
-	struct intel_encoder *intel_encoder = &intel_dig_port->base;
 	struct drm_device *dev = intel_dig_port->base.base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	enum intel_display_power_domain power_domain;
 	enum irqreturn ret = IRQ_NONE;
 
 	if (intel_dig_port->base.type != INTEL_OUTPUT_EDP &&
@@ -4966,12 +5065,12 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd)
 		      long_hpd ? "long" : "short");
 
 	if (long_hpd) {
+		intel_dp->reset_link_params = true;
 		intel_dp->detect_done = false;
 		return IRQ_NONE;
 	}
 
-	power_domain = intel_display_port_aux_power_domain(intel_encoder);
-	intel_display_power_get(dev_priv, power_domain);
+	intel_display_power_get(dev_priv, intel_dp->aux_power_domain);
 
 	if (intel_dp->is_mst) {
 		if (intel_dp_check_mst_status(intel_dp) == -EINVAL) {
@@ -4999,7 +5098,7 @@ intel_dp_hpd_pulse(struct intel_digital_port *intel_dig_port, bool long_hpd)
 	ret = IRQ_HANDLED;
 
 put_power:
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, intel_dp->aux_power_domain);
 
 	return ret;
 }
@@ -5790,6 +5889,41 @@ out_vdd_off:
 	return false;
 }
 
+/* Set up the hotplug pin and aux power domain. */
+static void
+intel_dp_init_connector_port_info(struct intel_digital_port *intel_dig_port)
+{
+	struct intel_encoder *encoder = &intel_dig_port->base;
+	struct intel_dp *intel_dp = &intel_dig_port->dp;
+
+	switch (intel_dig_port->port) {
+	case PORT_A:
+		encoder->hpd_pin = HPD_PORT_A;
+		intel_dp->aux_power_domain = POWER_DOMAIN_AUX_A;
+		break;
+	case PORT_B:
+		encoder->hpd_pin = HPD_PORT_B;
+		intel_dp->aux_power_domain = POWER_DOMAIN_AUX_B;
+		break;
+	case PORT_C:
+		encoder->hpd_pin = HPD_PORT_C;
+		intel_dp->aux_power_domain = POWER_DOMAIN_AUX_C;
+		break;
+	case PORT_D:
+		encoder->hpd_pin = HPD_PORT_D;
+		intel_dp->aux_power_domain = POWER_DOMAIN_AUX_D;
+		break;
+	case PORT_E:
+		encoder->hpd_pin = HPD_PORT_E;
+
+		/* FIXME: Check VBT for actual wiring of PORT E */
+		intel_dp->aux_power_domain = POWER_DOMAIN_AUX_D;
+		break;
+	default:
+		MISSING_CASE(intel_dig_port->port);
+	}
+}
+
 bool
 intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 			struct intel_connector *intel_connector)
@@ -5807,6 +5941,7 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 		 intel_dig_port->max_lanes, port_name(port)))
 		return false;
 
+	intel_dp->reset_link_params = true;
 	intel_dp->pps_pipe = INVALID_PIPE;
 	intel_dp->active_pipe = INVALID_PIPE;
 
@@ -5863,6 +5998,8 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	connector->interlace_allowed = true;
 	connector->doublescan_allowed = 0;
 
+	intel_dp_init_connector_port_info(intel_dig_port);
+
 	intel_dp_aux_init(intel_dp);
 
 	INIT_DELAYED_WORK(&intel_dp->panel_vdd_work,
@@ -5875,29 +6012,6 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port,
 	else
 		intel_connector->get_hw_state = intel_connector_get_hw_state;
 
-	/* Set up the hotplug pin. */
-	switch (port) {
-	case PORT_A:
-		intel_encoder->hpd_pin = HPD_PORT_A;
-		break;
-	case PORT_B:
-		intel_encoder->hpd_pin = HPD_PORT_B;
-		if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1))
-			intel_encoder->hpd_pin = HPD_PORT_A;
-		break;
-	case PORT_C:
-		intel_encoder->hpd_pin = HPD_PORT_C;
-		break;
-	case PORT_D:
-		intel_encoder->hpd_pin = HPD_PORT_D;
-		break;
-	case PORT_E:
-		intel_encoder->hpd_pin = HPD_PORT_E;
-		break;
-	default:
-		BUG();
-	}
-
 	/* init MST on ports that can support it */
 	if (HAS_DP_MST(dev_priv) && !is_edp(intel_dp) &&
 	    (port == PORT_B || port == PORT_C || port == PORT_D))
@@ -5982,6 +6096,7 @@ bool intel_dp_init(struct drm_i915_private *dev_priv,
 	intel_dig_port->max_lanes = 4;
 
 	intel_encoder->type = INTEL_OUTPUT_DP;
+	intel_encoder->power_domain = intel_port_to_power_domain(port);
 	if (IS_CHERRYVIEW(dev_priv)) {
 		if (port == PORT_D)
 			intel_encoder->crtc_mask = 1 << 2;
diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c
index 38e3ca2f6f18..094cbdcbcd6d 100644
--- a/drivers/gpu/drm/i915/intel_dp_mst.c
+++ b/drivers/gpu/drm/i915/intel_dp_mst.c
@@ -47,6 +47,11 @@ static bool intel_dp_mst_compute_config(struct intel_encoder *encoder,
 
 	pipe_config->has_pch_encoder = false;
 	bpp = 24;
+	if (intel_dp->compliance.test_data.bpc) {
+		bpp = intel_dp->compliance.test_data.bpc * 3;
+		DRM_DEBUG_KMS("Setting pipe bpp to %d\n",
+			      bpp);
+	}
 	/*
 	 * for MST we always configure max link bw - the spec doesn't
 	 * seem to suggest we should do otherwise.
@@ -55,7 +60,7 @@ static bool intel_dp_mst_compute_config(struct intel_encoder *encoder,
 
 	pipe_config->lane_count = lane_count;
 
-	pipe_config->pipe_bpp = 24;
+	pipe_config->pipe_bpp = bpp;
 	pipe_config->port_clock = intel_dp_max_link_rate(intel_dp);
 
 	state = pipe_config->base.state;
@@ -87,7 +92,6 @@ static void intel_mst_disable_dp(struct intel_encoder *encoder,
 	struct intel_dp *intel_dp = &intel_dig_port->dp;
 	struct intel_connector *connector =
 		to_intel_connector(old_conn_state->connector);
-	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	int ret;
 
 	DRM_DEBUG_KMS("%d\n", intel_dp->active_mst_links);
@@ -98,10 +102,8 @@ static void intel_mst_disable_dp(struct intel_encoder *encoder,
 	if (ret) {
 		DRM_ERROR("failed to update payload %d\n", ret);
 	}
-	if (old_crtc_state->has_audio) {
+	if (old_crtc_state->has_audio)
 		intel_audio_codec_disable(encoder);
-		intel_display_power_put(dev_priv, POWER_DOMAIN_AUDIO);
-	}
 }
 
 static void intel_mst_post_disable_dp(struct intel_encoder *encoder,
@@ -157,23 +159,9 @@ static void intel_mst_pre_enable_dp(struct intel_encoder *encoder,
 
 	DRM_DEBUG_KMS("%d\n", intel_dp->active_mst_links);
 
-	if (intel_dp->active_mst_links == 0) {
-		intel_ddi_clk_select(&intel_dig_port->base,
-				     pipe_config->shared_dpll);
-
-		intel_prepare_dp_ddi_buffers(&intel_dig_port->base);
-		intel_dp_set_link_params(intel_dp,
-					 pipe_config->port_clock,
-					 pipe_config->lane_count,
-					 true);
-
-		intel_ddi_init_dp_buf_reg(&intel_dig_port->base);
-
-		intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON);
-
-		intel_dp_start_link_train(intel_dp);
-		intel_dp_stop_link_train(intel_dp);
-	}
+	if (intel_dp->active_mst_links == 0)
+		intel_dig_port->base.pre_enable(&intel_dig_port->base,
+						pipe_config, NULL);
 
 	ret = drm_dp_mst_allocate_vcpi(&intel_dp->mst_mgr,
 				       connector->port,
@@ -214,10 +202,8 @@ static void intel_mst_enable_dp(struct intel_encoder *encoder,
 	ret = drm_dp_check_act_status(&intel_dp->mst_mgr);
 
 	ret = drm_dp_update_payload_part2(&intel_dp->mst_mgr);
-	if (pipe_config->has_audio) {
-		intel_display_power_get(dev_priv, POWER_DOMAIN_AUDIO);
+	if (pipe_config->has_audio)
 		intel_audio_codec_enable(encoder, pipe_config, conn_state);
-	}
 }
 
 static bool intel_dp_mst_enc_get_hw_state(struct intel_encoder *encoder,
@@ -548,6 +534,7 @@ intel_dp_create_fake_mst_encoder(struct intel_digital_port *intel_dig_port, enum
 			 DRM_MODE_ENCODER_DPMST, "DP-MST %c", pipe_name(pipe));
 
 	intel_encoder->type = INTEL_OUTPUT_DP_MST;
+	intel_encoder->power_domain = intel_dig_port->base.power_domain;
 	intel_encoder->port = intel_dig_port->port;
 	intel_encoder->crtc_mask = 0x7;
 	intel_encoder->cloneable = 0;
diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.c b/drivers/gpu/drm/i915/intel_dpll_mgr.c
index e59e43a9f3a6..b4de632f1158 100644
--- a/drivers/gpu/drm/i915/intel_dpll_mgr.c
+++ b/drivers/gpu/drm/i915/intel_dpll_mgr.c
@@ -42,44 +42,6 @@
  * commit phase.
  */
 
-struct intel_shared_dpll *
-skl_find_link_pll(struct drm_i915_private *dev_priv, int clock)
-{
-	struct intel_shared_dpll *pll = NULL;
-	struct intel_dpll_hw_state dpll_hw_state;
-	enum intel_dpll_id i;
-	bool found = false;
-
-	if (!skl_ddi_dp_set_dpll_hw_state(clock, &dpll_hw_state))
-		return pll;
-
-	for (i = DPLL_ID_SKL_DPLL1; i <= DPLL_ID_SKL_DPLL3; i++) {
-		pll = &dev_priv->shared_dplls[i];
-
-		/* Only want to check enabled timings first */
-		if (pll->state.crtc_mask == 0)
-			continue;
-
-		if (memcmp(&dpll_hw_state, &pll->state.hw_state,
-			   sizeof(pll->state.hw_state)) == 0) {
-			found = true;
-			break;
-		}
-	}
-
-	/* Ok no matching timings, maybe there's a free one? */
-	for (i = DPLL_ID_SKL_DPLL1;
-	     ((found == false) && (i <= DPLL_ID_SKL_DPLL3)); i++) {
-		pll = &dev_priv->shared_dplls[i];
-		if (pll->state.crtc_mask == 0) {
-			pll->state.hw_state = dpll_hw_state;
-			break;
-		}
-	}
-
-	return pll;
-}
-
 static void
 intel_atomic_duplicate_dpll_state(struct drm_i915_private *dev_priv,
 				  struct intel_shared_dpll_state *shared_dpll)
@@ -811,8 +773,8 @@ static struct intel_shared_dpll *hsw_ddi_hdmi_get_dpll(int clock,
 	return pll;
 }
 
-struct intel_shared_dpll *hsw_ddi_dp_get_dpll(struct intel_encoder *encoder,
-					      int clock)
+static struct intel_shared_dpll *
+hsw_ddi_dp_get_dpll(struct intel_encoder *encoder, int clock)
 {
 	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_shared_dpll *pll;
@@ -1360,8 +1322,9 @@ static bool skl_ddi_hdmi_pll_dividers(struct intel_crtc *crtc,
 }
 
 
-bool skl_ddi_dp_set_dpll_hw_state(int clock,
-				  struct intel_dpll_hw_state *dpll_hw_state)
+static bool
+skl_ddi_dp_set_dpll_hw_state(int clock,
+			     struct intel_dpll_hw_state *dpll_hw_state)
 {
 	uint32_t ctrl1;
 
@@ -1816,8 +1779,9 @@ static bool bxt_ddi_set_dpll_hw_state(int clock,
 	return true;
 }
 
-bool bxt_ddi_dp_set_dpll_hw_state(int clock,
-			  struct intel_dpll_hw_state *dpll_hw_state)
+static bool
+bxt_ddi_dp_set_dpll_hw_state(int clock,
+			     struct intel_dpll_hw_state *dpll_hw_state)
 {
 	struct bxt_clk_div clk_div = {0};
 
@@ -2016,7 +1980,7 @@ void intel_shared_dpll_init(struct drm_device *dev)
 	const struct dpll_info *dpll_info;
 	int i;
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+	if (IS_GEN9_BC(dev_priv))
 		dpll_mgr = &skl_pll_mgr;
 	else if (IS_GEN9_LP(dev_priv))
 		dpll_mgr = &bxt_pll_mgr;
diff --git a/drivers/gpu/drm/i915/intel_dpll_mgr.h b/drivers/gpu/drm/i915/intel_dpll_mgr.h
index af1497eb4f9c..f8d13a947c13 100644
--- a/drivers/gpu/drm/i915/intel_dpll_mgr.h
+++ b/drivers/gpu/drm/i915/intel_dpll_mgr.h
@@ -282,20 +282,4 @@ void intel_shared_dpll_init(struct drm_device *dev);
 void intel_dpll_dump_hw_state(struct drm_i915_private *dev_priv,
 			      struct intel_dpll_hw_state *hw_state);
 
-/* BXT dpll related functions */
-bool bxt_ddi_dp_set_dpll_hw_state(int clock,
-			  struct intel_dpll_hw_state *dpll_hw_state);
-
-
-/* SKL dpll related functions */
-bool skl_ddi_dp_set_dpll_hw_state(int clock,
-				  struct intel_dpll_hw_state *dpll_hw_state);
-struct intel_shared_dpll *skl_find_link_pll(struct drm_i915_private *dev_priv,
-					    int clock);
-
-
-/* HSW dpll related functions */
-struct intel_shared_dpll *hsw_ddi_dp_get_dpll(struct intel_encoder *encoder,
-					      int clock);
-
 #endif /* _INTEL_DPLL_MGR_H_ */
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 064582963ff6..0f766f83a31b 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -242,6 +242,9 @@ struct intel_encoder {
 	 * be set correctly before calling this function. */
 	void (*get_config)(struct intel_encoder *,
 			   struct intel_crtc_state *pipe_config);
+	/* Returns a mask of power domains that need to be referenced as part
+	 * of the hardware state readout code. */
+	u64 (*get_power_domains)(struct intel_encoder *encoder);
 	/*
 	 * Called during system suspend after all pending requests for the
 	 * encoder are flushed (for example for DP AUX transactions) and
@@ -250,6 +253,7 @@ struct intel_encoder {
 	void (*suspend)(struct intel_encoder *);
 	int crtc_mask;
 	enum hpd_pin hpd_pin;
+	enum intel_display_power_domain power_domain;
 	/* for communication with audio component; protected by av_mutex */
 	const struct drm_connector *audio_connector;
 };
@@ -334,13 +338,20 @@ struct dpll {
 struct intel_atomic_state {
 	struct drm_atomic_state base;
 
-	unsigned int cdclk;
-
-	/*
-	 * Calculated device cdclk, can be different from cdclk
-	 * only when all crtc's are DPMS off.
-	 */
-	unsigned int dev_cdclk;
+	struct {
+		/*
+		 * Logical state of cdclk (used for all scaling, watermark,
+		 * etc. calculations and checks). This is computed as if all
+		 * enabled crtcs were active.
+		 */
+		struct intel_cdclk_state logical;
+
+		/*
+		 * Actual state of cdclk, can be different from the logical
+		 * state only when all crtc's are DPMS off.
+		 */
+		struct intel_cdclk_state actual;
+	} cdclk;
 
 	bool dpll_set, modeset;
 
@@ -357,9 +368,6 @@ struct intel_atomic_state {
 	unsigned int active_crtcs;
 	unsigned int min_pixclk[I915_MAX_PIPES];
 
-	/* SKL/KBL Only */
-	unsigned int cdclk_pll_vco;
-
 	struct intel_shared_dpll_state shared_dpll[I915_NUM_PLLS];
 
 	/*
@@ -485,6 +493,24 @@ struct skl_pipe_wm {
 	uint32_t linetime;
 };
 
+enum vlv_wm_level {
+	VLV_WM_LEVEL_PM2,
+	VLV_WM_LEVEL_PM5,
+	VLV_WM_LEVEL_DDR_DVFS,
+	NUM_VLV_WM_LEVELS,
+};
+
+struct vlv_wm_state {
+	struct vlv_pipe_wm wm[NUM_VLV_WM_LEVELS];
+	struct vlv_sr_wm sr[NUM_VLV_WM_LEVELS];
+	uint8_t num_levels;
+	bool cxsr;
+};
+
+struct vlv_fifo_state {
+	u16 plane[I915_MAX_PLANES];
+};
+
 struct intel_crtc_wm_state {
 	union {
 		struct {
@@ -509,6 +535,17 @@ struct intel_crtc_wm_state {
 			struct skl_pipe_wm optimal;
 			struct skl_ddb_entry ddb;
 		} skl;
+
+		struct {
+			/* "raw" watermarks (not inverted) */
+			struct vlv_pipe_wm raw[NUM_VLV_WM_LEVELS];
+			/* intermediate watermarks (inverted) */
+			struct vlv_wm_state intermediate;
+			/* optimal watermarks (inverted) */
+			struct vlv_wm_state optimal;
+			/* display FIFO split */
+			struct vlv_fifo_state fifo_state;
+		} vlv;
 	};
 
 	/*
@@ -539,12 +576,19 @@ struct intel_crtc_state {
 	bool disable_cxsr;
 	bool update_wm_pre, update_wm_post; /* watermarks are updated */
 	bool fb_changed; /* fb on any of the planes is changed */
+	bool fifo_changed; /* FIFO split is changed */
 
 	/* Pipe source size (ie. panel fitter input size)
 	 * All planes will be positioned inside this space,
 	 * and get clipped at the edges. */
 	int pipe_src_w, pipe_src_h;
 
+	/*
+	 * Pipe pixel rate, adjusted for
+	 * panel fitter/pipe scaler downscaling.
+	 */
+	unsigned int pixel_rate;
+
 	/* Whether to set up the PCH/FDI. Note that we never allow sharing
 	 * between pch encoders and cpu encoders. */
 	bool has_pch_encoder;
@@ -581,6 +625,14 @@ struct intel_crtc_state {
 	 */
 	bool dither;
 
+	/*
+	 * Dither gets enabled for 18bpp which causes CRC mismatch errors for
+	 * compliance video pattern tests.
+	 * Disable dither only if it is a compliance test request for
+	 * 18bpp.
+	 */
+	bool dither_force_disable;
+
 	/* Controls for the clock computation, to override various stages. */
 	bool clock_set;
 
@@ -674,15 +726,9 @@ struct intel_crtc_state {
 
 	/* Gamma mode programmed on the pipe */
 	uint32_t gamma_mode;
-};
 
-struct vlv_wm_state {
-	struct vlv_pipe_wm wm[3];
-	struct vlv_sr_wm sr[3];
-	uint8_t num_active_planes;
-	uint8_t num_levels;
-	uint8_t level;
-	bool cxsr;
+	/* bitmask of visible planes (enum plane_id) */
+	u8 active_planes;
 };
 
 struct intel_crtc {
@@ -698,7 +744,7 @@ struct intel_crtc {
 	bool active;
 	bool lowfreq_avail;
 	u8 plane_ids_mask;
-	unsigned long enabled_power_domains;
+	unsigned long long enabled_power_domains;
 	struct intel_overlay *overlay;
 	struct intel_flip_work *flip_work;
 
@@ -730,10 +776,8 @@ struct intel_crtc {
 		/* watermarks currently being used  */
 		union {
 			struct intel_pipe_wm ilk;
+			struct vlv_wm_state vlv;
 		} active;
-
-		/* allow CxSR on this pipe */
-		bool cxsr_allowed;
 	} wm;
 
 	int scanline_offset;
@@ -747,27 +791,6 @@ struct intel_crtc {
 
 	/* scalers available on this crtc */
 	int num_scalers;
-
-	struct vlv_wm_state wm_state;
-};
-
-struct intel_plane_wm_parameters {
-	uint32_t horiz_pixels;
-	uint32_t vert_pixels;
-	/*
-	 *   For packed pixel formats:
-	 *     bytes_per_pixel - holds bytes per pixel
-	 *   For planar pixel formats:
-	 *     bytes_per_pixel - holds bytes per pixel for uv-plane
-	 *     y_bytes_per_pixel - holds bytes per pixel for y-plane
-	 */
-	uint8_t bytes_per_pixel;
-	uint8_t y_bytes_per_pixel;
-	bool enabled;
-	bool scaled;
-	u64 tiling;
-	unsigned int rotation;
-	uint16_t fifo_size;
 };
 
 struct intel_plane {
@@ -779,13 +802,6 @@ struct intel_plane {
 	int max_downscale;
 	uint32_t frontbuffer_bit;
 
-	/* Since we need to change the watermarks before/after
-	 * enabling/disabling the planes, we need to store the parameters here
-	 * as the other pieces of the struct may not reflect the values we want
-	 * for the watermark calculations. Currently only Haswell uses this.
-	 */
-	struct intel_plane_wm_parameters wm;
-
 	/*
 	 * NOTE: Do not place new plane state fields here (e.g., when adding
 	 * new plane properties).  New runtime state should now be placed in
@@ -891,12 +907,17 @@ struct intel_dp_desc {
 
 struct intel_dp_compliance_data {
 	unsigned long edid;
+	uint8_t video_pattern;
+	uint16_t hdisplay, vdisplay;
+	uint8_t bpc;
 };
 
 struct intel_dp_compliance {
 	unsigned long test_type;
 	struct intel_dp_compliance_data test_data;
 	bool test_active;
+	int test_link_rate;
+	u8 test_lane_count;
 };
 
 struct intel_dp {
@@ -911,6 +932,7 @@ struct intel_dp {
 	bool has_audio;
 	bool detect_done;
 	bool channel_eq_status;
+	bool reset_link_params;
 	enum hdmi_force_audio force_audio;
 	bool limited_color_range;
 	bool color_range_auto;
@@ -928,6 +950,7 @@ struct intel_dp {
 	/* sink or branch descriptor */
 	struct intel_dp_desc desc;
 	struct drm_dp_aux aux;
+	enum intel_display_power_domain aux_power_domain;
 	uint8_t train_set[4];
 	int panel_power_up_delay;
 	int panel_power_down_delay;
@@ -990,7 +1013,6 @@ struct intel_dp {
 struct intel_lspcon {
 	bool active;
 	enum drm_lspcon_mode mode;
-	bool desc_valid;
 };
 
 struct intel_digital_port {
@@ -1003,6 +1025,7 @@ struct intel_digital_port {
 	enum irqreturn (*hpd_pulse)(struct intel_digital_port *, bool);
 	bool release_cl2_override;
 	uint8_t max_lanes;
+	enum intel_display_power_domain ddi_io_power_domain;
 };
 
 struct intel_dp_mst_encoder {
@@ -1097,7 +1120,19 @@ intel_attached_encoder(struct drm_connector *connector)
 static inline struct intel_digital_port *
 enc_to_dig_port(struct drm_encoder *encoder)
 {
-	return container_of(encoder, struct intel_digital_port, base.base);
+	struct intel_encoder *intel_encoder = to_intel_encoder(encoder);
+
+	switch (intel_encoder->type) {
+	case INTEL_OUTPUT_UNKNOWN:
+		WARN_ON(!HAS_DDI(to_i915(encoder->dev)));
+	case INTEL_OUTPUT_DP:
+	case INTEL_OUTPUT_EDP:
+	case INTEL_OUTPUT_HDMI:
+		return container_of(encoder, struct intel_digital_port,
+				    base.base);
+	default:
+		return NULL;
+	}
 }
 
 static inline struct intel_dp_mst_encoder *
@@ -1185,18 +1220,19 @@ void intel_ddi_fdi_post_disable(struct intel_encoder *intel_encoder,
 				struct intel_crtc_state *old_crtc_state,
 				struct drm_connector_state *old_conn_state);
 void intel_prepare_dp_ddi_buffers(struct intel_encoder *encoder);
-void hsw_fdi_link_train(struct drm_crtc *crtc);
+void hsw_fdi_link_train(struct intel_crtc *crtc,
+			const struct intel_crtc_state *crtc_state);
 void intel_ddi_init(struct drm_i915_private *dev_priv, enum port port);
 enum port intel_ddi_get_encoder_port(struct intel_encoder *intel_encoder);
 bool intel_ddi_get_hw_state(struct intel_encoder *encoder, enum pipe *pipe);
-void intel_ddi_enable_transcoder_func(struct drm_crtc *crtc);
+void intel_ddi_enable_transcoder_func(const struct intel_crtc_state *crtc_state);
 void intel_ddi_disable_transcoder_func(struct drm_i915_private *dev_priv,
 				       enum transcoder cpu_transcoder);
-void intel_ddi_enable_pipe_clock(struct intel_crtc *intel_crtc);
-void intel_ddi_disable_pipe_clock(struct intel_crtc *intel_crtc);
+void intel_ddi_enable_pipe_clock(const struct intel_crtc_state *crtc_state);
+void intel_ddi_disable_pipe_clock(const  struct intel_crtc_state *crtc_state);
 bool intel_ddi_pll_select(struct intel_crtc *crtc,
 			  struct intel_crtc_state *crtc_state);
-void intel_ddi_set_pipe_settings(struct drm_crtc *crtc);
+void intel_ddi_set_pipe_settings(const struct intel_crtc_state *crtc_state);
 void intel_ddi_prepare_link_retrain(struct intel_dp *intel_dp);
 bool intel_ddi_connector_get_hw_state(struct intel_connector *intel_connector);
 bool intel_ddi_is_audio_enabled(struct drm_i915_private *dev_priv,
@@ -1209,11 +1245,12 @@ intel_ddi_get_crtc_new_encoder(struct intel_crtc_state *crtc_state);
 void intel_ddi_init_dp_buf_reg(struct intel_encoder *encoder);
 void intel_ddi_clock_get(struct intel_encoder *encoder,
 			 struct intel_crtc_state *pipe_config);
-void intel_ddi_set_vc_payload_alloc(struct drm_crtc *crtc, bool state);
+void intel_ddi_set_vc_payload_alloc(const struct intel_crtc_state *crtc_state,
+				    bool state);
 uint32_t ddi_signal_levels(struct intel_dp *intel_dp);
-struct intel_shared_dpll *intel_ddi_get_link_dpll(struct intel_dp *intel_dp,
-						  int clock);
-unsigned int intel_fb_align_height(struct drm_device *dev,
+u8 intel_ddi_dp_voltage_max(struct intel_encoder *encoder);
+
+unsigned int intel_fb_align_height(struct drm_i915_private *dev_priv,
 				   unsigned int height,
 				   uint32_t pixel_format,
 				   uint64_t fb_format_modifier);
@@ -1231,12 +1268,24 @@ void i915_audio_component_cleanup(struct drm_i915_private *dev_priv);
 void intel_audio_init(struct drm_i915_private *dev_priv);
 void intel_audio_deinit(struct drm_i915_private *dev_priv);
 
+/* intel_cdclk.c */
+void intel_init_cdclk_hooks(struct drm_i915_private *dev_priv);
+void intel_update_max_cdclk(struct drm_i915_private *dev_priv);
+void intel_update_cdclk(struct drm_i915_private *dev_priv);
+void intel_update_rawclk(struct drm_i915_private *dev_priv);
+bool intel_cdclk_state_compare(const struct intel_cdclk_state *a,
+			       const struct intel_cdclk_state *b);
+void intel_set_cdclk(struct drm_i915_private *dev_priv,
+		     const struct intel_cdclk_state *cdclk_state);
+
 /* intel_display.c */
 enum transcoder intel_crtc_pch_transcoder(struct intel_crtc *crtc);
-void skl_set_preferred_cdclk_vco(struct drm_i915_private *dev_priv, int vco);
 void intel_update_rawclk(struct drm_i915_private *dev_priv);
+int vlv_get_hpll_vco(struct drm_i915_private *dev_priv);
 int vlv_get_cck_clock(struct drm_i915_private *dev_priv,
 		      const char *name, u32 reg, int ref_freq);
+int vlv_get_cck_clock_hpll(struct drm_i915_private *dev_priv,
+			   const char *name, u32 reg);
 void lpt_disable_pch_transcoder(struct drm_i915_private *dev_priv);
 void lpt_disable_iclkip(struct drm_i915_private *dev_priv);
 extern const struct drm_plane_funcs intel_plane_funcs;
@@ -1311,9 +1360,8 @@ struct i915_vma *
 intel_pin_and_fence_fb_obj(struct drm_framebuffer *fb, unsigned int rotation);
 void intel_unpin_fb_vma(struct i915_vma *vma);
 struct drm_framebuffer *
-__intel_framebuffer_create(struct drm_device *dev,
-			   struct drm_mode_fb_cmd2 *mode_cmd,
-			   struct drm_i915_gem_object *obj);
+intel_framebuffer_create(struct drm_i915_gem_object *obj,
+			 struct drm_mode_fb_cmd2 *mode_cmd);
 void intel_finish_page_flip_cs(struct drm_i915_private *dev_priv, int pipe);
 void intel_finish_page_flip_mmio(struct drm_i915_private *dev_priv, int pipe);
 void intel_check_page_flip(struct drm_i915_private *dev_priv, int pipe);
@@ -1388,10 +1436,7 @@ int chv_calc_dpll_params(int refclk, struct dpll *pll_clock);
 bool intel_crtc_active(struct intel_crtc *crtc);
 void hsw_enable_ips(struct intel_crtc *crtc);
 void hsw_disable_ips(struct intel_crtc *crtc);
-enum intel_display_power_domain
-intel_display_port_power_domain(struct intel_encoder *intel_encoder);
-enum intel_display_power_domain
-intel_display_port_aux_power_domain(struct intel_encoder *intel_encoder);
+enum intel_display_power_domain intel_port_to_power_domain(enum port port);
 void intel_mode_from_pipe_config(struct drm_display_mode *mode,
 				 struct intel_crtc_state *pipe_config);
 
@@ -1664,6 +1709,7 @@ int intel_power_domains_init(struct drm_i915_private *);
 void intel_power_domains_fini(struct drm_i915_private *);
 void intel_power_domains_init_hw(struct drm_i915_private *dev_priv, bool resume);
 void intel_power_domains_suspend(struct drm_i915_private *dev_priv);
+void intel_power_domains_verify_state(struct drm_i915_private *dev_priv);
 void bxt_display_core_init(struct drm_i915_private *dev_priv, bool resume);
 void bxt_display_core_uninit(struct drm_i915_private *dev_priv);
 void intel_runtime_pm_enable(struct drm_i915_private *dev_priv);
@@ -1692,10 +1738,8 @@ static inline void
 assert_rpm_wakelock_held(struct drm_i915_private *dev_priv)
 {
 	assert_rpm_device_not_suspended(dev_priv);
-	/* FIXME: Needs to be converted back to WARN_ONCE, but currently causes
-	 * too much noise. */
-	if (!atomic_read(&dev_priv->pm.wakeref_count))
-		DRM_DEBUG_DRIVER("RPM wakelock ref not held during HW access");
+	WARN_ONCE(!atomic_read(&dev_priv->pm.wakeref_count),
+		  "RPM wakelock ref not held during HW access");
 }
 
 /**
@@ -1783,6 +1827,7 @@ void skl_ddb_get_hw_state(struct drm_i915_private *dev_priv,
 			  struct skl_ddb_allocation *ddb /* out */);
 void skl_pipe_wm_get_hw_state(struct drm_crtc *crtc,
 			      struct skl_pipe_wm *out);
+void vlv_wm_sanitize(struct drm_i915_private *dev_priv);
 bool intel_can_enable_sagv(struct drm_atomic_state *state);
 int intel_enable_sagv(struct drm_i915_private *dev_priv);
 int intel_disable_sagv(struct drm_i915_private *dev_priv);
@@ -1791,7 +1836,6 @@ bool skl_wm_level_equals(const struct skl_wm_level *l1,
 bool skl_ddb_allocation_overlaps(const struct skl_ddb_entry **entries,
 				 const struct skl_ddb_entry *ddb,
 				 int ignore);
-uint32_t ilk_pipe_pixel_rate(const struct intel_crtc_state *pipe_config);
 bool ilk_disable_lp_wm(struct drm_device *dev);
 int sanitize_rc6_option(struct drm_i915_private *dev_priv, int enable_rc6);
 static inline int intel_enable_rc6(void)
@@ -1865,9 +1909,9 @@ intel_atomic_get_existing_plane_state(struct drm_atomic_state *state,
 	return to_intel_plane_state(plane_state);
 }
 
-int intel_atomic_setup_scalers(struct drm_device *dev,
-	struct intel_crtc *intel_crtc,
-	struct intel_crtc_state *crtc_state);
+int intel_atomic_setup_scalers(struct drm_i915_private *dev_priv,
+			       struct intel_crtc *intel_crtc,
+			       struct intel_crtc_state *crtc_state);
 
 /* intel_atomic_plane.c */
 struct intel_plane_state *intel_create_plane_state(struct drm_plane *plane);
diff --git a/drivers/gpu/drm/i915/intel_dsi.c b/drivers/gpu/drm/i915/intel_dsi.c
index 16732e7bc08e..323fc097c3ee 100644
--- a/drivers/gpu/drm/i915/intel_dsi.c
+++ b/drivers/gpu/drm/i915/intel_dsi.c
@@ -80,7 +80,7 @@ enum mipi_dsi_pixel_format pixel_format_from_register_bits(u32 fmt)
 	}
 }
 
-static void wait_for_dsi_fifo_empty(struct intel_dsi *intel_dsi, enum port port)
+void wait_for_dsi_fifo_empty(struct intel_dsi *intel_dsi, enum port port)
 {
 	struct drm_encoder *encoder = &intel_dsi->base.base;
 	struct drm_device *dev = encoder->dev;
@@ -357,41 +357,132 @@ static bool intel_dsi_compute_config(struct intel_encoder *encoder,
 	return true;
 }
 
-static void bxt_dsi_device_ready(struct intel_encoder *encoder)
+static void glk_dsi_device_ready(struct intel_encoder *encoder)
 {
 	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	enum port port;
-	u32 val;
+	u32 tmp, val;
 
-	DRM_DEBUG_KMS("\n");
+	/* Set the MIPI mode
+	 * If MIPI_Mode is off, then writing to LP_Wake bit is not reflecting.
+	 * Power ON MIPI IO first and then write into IO reset and LP wake bits
+	 */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		tmp = I915_READ(MIPI_CTRL(port));
+		I915_WRITE(MIPI_CTRL(port), tmp | GLK_MIPIIO_ENABLE);
+	}
+
+	/* Put the IO into reset */
+	tmp = I915_READ(MIPI_CTRL(PORT_A));
+	tmp &= ~GLK_MIPIIO_RESET_RELEASED;
+	I915_WRITE(MIPI_CTRL(PORT_A), tmp);
 
-	/* Exit Low power state in 4 steps*/
+	/* Program LP Wake */
 	for_each_dsi_port(port, intel_dsi->ports) {
+		tmp = I915_READ(MIPI_CTRL(port));
+		tmp |= GLK_LP_WAKE;
+		I915_WRITE(MIPI_CTRL(port), tmp);
+	}
 
-		/* 1. Enable MIPI PHY transparent latch */
-		val = I915_READ(BXT_MIPI_PORT_CTRL(port));
-		I915_WRITE(BXT_MIPI_PORT_CTRL(port), val | LP_OUTPUT_HOLD);
-		usleep_range(2000, 2500);
+	/* Wait for Pwr ACK */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_wait_for_register(dev_priv,
+				MIPI_CTRL(port), GLK_MIPIIO_PORT_POWERED,
+				GLK_MIPIIO_PORT_POWERED, 20))
+			DRM_ERROR("MIPIO port is powergated\n");
+	}
+
+	/* Wait for MIPI PHY status bit to set */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_wait_for_register(dev_priv,
+				MIPI_CTRL(port), GLK_PHY_STATUS_PORT_READY,
+				GLK_PHY_STATUS_PORT_READY, 20))
+			DRM_ERROR("PHY is not ON\n");
+	}
+
+	/* Get IO out of reset */
+	tmp = I915_READ(MIPI_CTRL(PORT_A));
+	I915_WRITE(MIPI_CTRL(PORT_A), tmp | GLK_MIPIIO_RESET_RELEASED);
 
-		/* 2. Enter ULPS */
+	/* Get IO out of Low power state*/
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (!(I915_READ(MIPI_DEVICE_READY(port)) & DEVICE_READY)) {
+			val = I915_READ(MIPI_DEVICE_READY(port));
+			val &= ~ULPS_STATE_MASK;
+			val |= DEVICE_READY;
+			I915_WRITE(MIPI_DEVICE_READY(port), val);
+			usleep_range(10, 15);
+		}
+
+		/* Enter ULPS */
 		val = I915_READ(MIPI_DEVICE_READY(port));
 		val &= ~ULPS_STATE_MASK;
 		val |= (ULPS_STATE_ENTER | DEVICE_READY);
 		I915_WRITE(MIPI_DEVICE_READY(port), val);
-		/* at least 2us - relaxed for hrtimer subsystem optimization */
-		usleep_range(10, 50);
 
-		/* 3. Exit ULPS */
+		/* Wait for ULPS Not active */
+		if (intel_wait_for_register(dev_priv,
+				MIPI_CTRL(port), GLK_ULPS_NOT_ACTIVE,
+				GLK_ULPS_NOT_ACTIVE, 20))
+			DRM_ERROR("ULPS is still active\n");
+
+		/* Exit ULPS */
 		val = I915_READ(MIPI_DEVICE_READY(port));
 		val &= ~ULPS_STATE_MASK;
 		val |= (ULPS_STATE_EXIT | DEVICE_READY);
 		I915_WRITE(MIPI_DEVICE_READY(port), val);
-		usleep_range(1000, 1500);
 
-		/* Clear ULPS and set device ready */
+		/* Enter Normal Mode */
+		val = I915_READ(MIPI_DEVICE_READY(port));
+		val &= ~ULPS_STATE_MASK;
+		val |= (ULPS_STATE_NORMAL_OPERATION | DEVICE_READY);
+		I915_WRITE(MIPI_DEVICE_READY(port), val);
+
+		tmp = I915_READ(MIPI_CTRL(port));
+		tmp &= ~GLK_LP_WAKE;
+		I915_WRITE(MIPI_CTRL(port), tmp);
+	}
+
+	/* Wait for Stop state */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_wait_for_register(dev_priv,
+				MIPI_CTRL(port), GLK_DATA_LANE_STOP_STATE,
+				GLK_DATA_LANE_STOP_STATE, 20))
+			DRM_ERROR("Date lane not in STOP state\n");
+	}
+
+	/* Wait for AFE LATCH */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_wait_for_register(dev_priv,
+				BXT_MIPI_PORT_CTRL(port), AFE_LATCHOUT,
+				AFE_LATCHOUT, 20))
+			DRM_ERROR("D-PHY not entering LP-11 state\n");
+	}
+}
+
+static void bxt_dsi_device_ready(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
+	enum port port;
+	u32 val;
+
+	DRM_DEBUG_KMS("\n");
+
+	/* Enable MIPI PHY transparent latch */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		val = I915_READ(BXT_MIPI_PORT_CTRL(port));
+		I915_WRITE(BXT_MIPI_PORT_CTRL(port), val | LP_OUTPUT_HOLD);
+		usleep_range(2000, 2500);
+	}
+
+	/* Clear ULPS and set device ready */
+	for_each_dsi_port(port, intel_dsi->ports) {
 		val = I915_READ(MIPI_DEVICE_READY(port));
 		val &= ~ULPS_STATE_MASK;
+		I915_WRITE(MIPI_DEVICE_READY(port), val);
+		usleep_range(2000, 2500);
 		val |= DEVICE_READY;
 		I915_WRITE(MIPI_DEVICE_READY(port), val);
 	}
@@ -442,8 +533,121 @@ static void intel_dsi_device_ready(struct intel_encoder *encoder)
 
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
 		vlv_dsi_device_ready(encoder);
-	else if (IS_GEN9_LP(dev_priv))
+	else if (IS_BROXTON(dev_priv))
 		bxt_dsi_device_ready(encoder);
+	else if (IS_GEMINILAKE(dev_priv))
+		glk_dsi_device_ready(encoder);
+}
+
+static void glk_dsi_enter_low_power_mode(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
+	enum port port;
+	u32 val;
+
+	/* Enter ULPS */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		val = I915_READ(MIPI_DEVICE_READY(port));
+		val &= ~ULPS_STATE_MASK;
+		val |= (ULPS_STATE_ENTER | DEVICE_READY);
+		I915_WRITE(MIPI_DEVICE_READY(port), val);
+	}
+
+	/* Wait for MIPI PHY status bit to unset */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_wait_for_register(dev_priv,
+					    MIPI_CTRL(port),
+					    GLK_PHY_STATUS_PORT_READY, 0, 20))
+			DRM_ERROR("PHY is not turning OFF\n");
+	}
+
+	/* Wait for Pwr ACK bit to unset */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_wait_for_register(dev_priv,
+					    MIPI_CTRL(port),
+					    GLK_MIPIIO_PORT_POWERED, 0, 20))
+			DRM_ERROR("MIPI IO Port is not powergated\n");
+	}
+}
+
+static void glk_dsi_disable_mipi_io(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
+	enum port port;
+	u32 tmp;
+
+	/* Put the IO into reset */
+	tmp = I915_READ(MIPI_CTRL(PORT_A));
+	tmp &= ~GLK_MIPIIO_RESET_RELEASED;
+	I915_WRITE(MIPI_CTRL(PORT_A), tmp);
+
+	/* Wait for MIPI PHY status bit to unset */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		if (intel_wait_for_register(dev_priv,
+					    MIPI_CTRL(port),
+					    GLK_PHY_STATUS_PORT_READY, 0, 20))
+			DRM_ERROR("PHY is not turning OFF\n");
+	}
+
+	/* Clear MIPI mode */
+	for_each_dsi_port(port, intel_dsi->ports) {
+		tmp = I915_READ(MIPI_CTRL(port));
+		tmp &= ~GLK_MIPIIO_ENABLE;
+		I915_WRITE(MIPI_CTRL(port), tmp);
+	}
+}
+
+static void glk_dsi_clear_device_ready(struct intel_encoder *encoder)
+{
+	glk_dsi_enter_low_power_mode(encoder);
+	glk_dsi_disable_mipi_io(encoder);
+}
+
+static void vlv_dsi_clear_device_ready(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
+	enum port port;
+
+	DRM_DEBUG_KMS("\n");
+	for_each_dsi_port(port, intel_dsi->ports) {
+		/* Common bit for both MIPI Port A & MIPI Port C on VLV/CHV */
+		i915_reg_t port_ctrl = IS_GEN9_LP(dev_priv) ?
+			BXT_MIPI_PORT_CTRL(port) : MIPI_PORT_CTRL(PORT_A);
+		u32 val;
+
+		I915_WRITE(MIPI_DEVICE_READY(port), DEVICE_READY |
+							ULPS_STATE_ENTER);
+		usleep_range(2000, 2500);
+
+		I915_WRITE(MIPI_DEVICE_READY(port), DEVICE_READY |
+							ULPS_STATE_EXIT);
+		usleep_range(2000, 2500);
+
+		I915_WRITE(MIPI_DEVICE_READY(port), DEVICE_READY |
+							ULPS_STATE_ENTER);
+		usleep_range(2000, 2500);
+
+		/*
+		 * On VLV/CHV, wait till Clock lanes are in LP-00 state for MIPI
+		 * Port A only. MIPI Port C has no similar bit for checking.
+		 */
+		if ((IS_GEN9_LP(dev_priv) || port == PORT_A) &&
+		    intel_wait_for_register(dev_priv,
+					    port_ctrl, AFE_LATCHOUT, 0,
+					    30))
+			DRM_ERROR("DSI LP not going Low\n");
+
+		/* Disable MIPI PHY transparent latch */
+		val = I915_READ(port_ctrl);
+		I915_WRITE(port_ctrl, val & ~LP_OUTPUT_HOLD);
+		usleep_range(1000, 1500);
+
+		I915_WRITE(MIPI_DEVICE_READY(port), 0x00);
+		usleep_range(2000, 2500);
+	}
 }
 
 static void intel_dsi_port_enable(struct intel_encoder *encoder)
@@ -456,12 +660,21 @@ static void intel_dsi_port_enable(struct intel_encoder *encoder)
 
 	if (intel_dsi->dual_link == DSI_DUAL_LINK_FRONT_BACK) {
 		u32 temp;
-
-		temp = I915_READ(VLV_CHICKEN_3);
-		temp &= ~PIXEL_OVERLAP_CNT_MASK |
+		if (IS_GEN9_LP(dev_priv)) {
+			for_each_dsi_port(port, intel_dsi->ports) {
+				temp = I915_READ(MIPI_CTRL(port));
+				temp &= ~BXT_PIXEL_OVERLAP_CNT_MASK |
+					intel_dsi->pixel_overlap <<
+					BXT_PIXEL_OVERLAP_CNT_SHIFT;
+				I915_WRITE(MIPI_CTRL(port), temp);
+			}
+		} else {
+			temp = I915_READ(VLV_CHICKEN_3);
+			temp &= ~PIXEL_OVERLAP_CNT_MASK |
 					intel_dsi->pixel_overlap <<
 					PIXEL_OVERLAP_CNT_SHIFT;
-		I915_WRITE(VLV_CHICKEN_3, temp);
+			I915_WRITE(VLV_CHICKEN_3, temp);
+		}
 	}
 
 	for_each_dsi_port(port, intel_dsi->ports) {
@@ -509,37 +722,57 @@ static void intel_dsi_port_disable(struct intel_encoder *encoder)
 	}
 }
 
-static void intel_dsi_enable(struct intel_encoder *encoder)
-{
-	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
-	enum port port;
-
-	DRM_DEBUG_KMS("\n");
-
-	if (is_cmd_mode(intel_dsi)) {
-		for_each_dsi_port(port, intel_dsi->ports)
-			I915_WRITE(MIPI_MAX_RETURN_PKT_SIZE(port), 8 * 4);
-	} else {
-		msleep(20); /* XXX */
-		for_each_dsi_port(port, intel_dsi->ports)
-			dpi_send_cmd(intel_dsi, TURN_ON, false, port);
-		msleep(100);
-
-		drm_panel_enable(intel_dsi->panel);
+static void intel_dsi_prepare(struct intel_encoder *intel_encoder,
+			      struct intel_crtc_state *pipe_config);
+static void intel_dsi_unprepare(struct intel_encoder *encoder);
 
-		for_each_dsi_port(port, intel_dsi->ports)
-			wait_for_dsi_fifo_empty(intel_dsi, port);
+static void intel_dsi_msleep(struct intel_dsi *intel_dsi, int msec)
+{
+	struct drm_i915_private *dev_priv = to_i915(intel_dsi->base.base.dev);
 
-		intel_dsi_port_enable(encoder);
-	}
+	/* For v3 VBTs in vid-mode the delays are part of the VBT sequences */
+	if (is_vid_mode(intel_dsi) && dev_priv->vbt.dsi.seq_version >= 3)
+		return;
 
-	intel_panel_enable_backlight(intel_dsi->attached_connector);
+	msleep(msec);
 }
 
-static void intel_dsi_prepare(struct intel_encoder *intel_encoder,
-			      struct intel_crtc_state *pipe_config);
+/*
+ * Panel enable/disable sequences from the VBT spec.
+ *
+ * Note the spec has AssertReset / DeassertReset swapped from their
+ * usual naming. We use the normal names to avoid confusion (so below
+ * they are swapped compared to the spec).
+ *
+ * Steps starting with MIPI refer to VBT sequences, note that for v2
+ * VBTs several steps which have a VBT in v2 are expected to be handled
+ * directly by the driver, by directly driving gpios for example.
+ *
+ * v2 video mode seq         v3 video mode seq         command mode seq
+ * - power on                - MIPIPanelPowerOn        - power on
+ * - wait t1+t2                                        - wait t1+t2
+ * - MIPIDeassertResetPin    - MIPIDeassertResetPin    - MIPIDeassertResetPin
+ * - io lines to lp-11       - io lines to lp-11       - io lines to lp-11
+ * - MIPISendInitialDcsCmds  - MIPISendInitialDcsCmds  - MIPISendInitialDcsCmds
+ *                                                     - MIPITearOn
+ *                                                     - MIPIDisplayOn
+ * - turn on DPI             - turn on DPI             - set pipe to dsr mode
+ * - MIPIDisplayOn           - MIPIDisplayOn
+ * - wait t5                                           - wait t5
+ * - backlight on            - MIPIBacklightOn         - backlight on
+ * ...                       ...                       ... issue mem cmds ...
+ * - backlight off           - MIPIBacklightOff        - backlight off
+ * - wait t6                                           - wait t6
+ * - MIPIDisplayOff
+ * - turn off DPI            - turn off DPI            - disable pipe dsr mode
+ *                                                     - MIPITearOff
+ *                           - MIPIDisplayOff          - MIPIDisplayOff
+ * - io lines to lp-00       - io lines to lp-00       - io lines to lp-00
+ * - MIPIAssertResetPin      - MIPIAssertResetPin      - MIPIAssertResetPin
+ * - wait t3                                           - wait t3
+ * - power off               - MIPIPanelPowerOff       - power off
+ * - wait t4                                           - wait t4
+ */
 
 static void intel_dsi_pre_enable(struct intel_encoder *encoder,
 				 struct intel_crtc_state *pipe_config,
@@ -548,6 +781,7 @@ static void intel_dsi_pre_enable(struct intel_encoder *encoder,
 	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	enum port port;
+	u32 val;
 
 	DRM_DEBUG_KMS("\n");
 
@@ -558,13 +792,16 @@ static void intel_dsi_pre_enable(struct intel_encoder *encoder,
 	intel_disable_dsi_pll(encoder);
 	intel_enable_dsi_pll(encoder, pipe_config);
 
-	intel_dsi_prepare(encoder, pipe_config);
-
-	/* Panel Enable over CRC PMIC */
-	if (intel_dsi->gpio_panel)
-		gpiod_set_value_cansleep(intel_dsi->gpio_panel, 1);
+	if (IS_BROXTON(dev_priv)) {
+		/* Add MIPI IO reset programming for modeset */
+		val = I915_READ(BXT_P_CR_GT_DISP_PWRON);
+		I915_WRITE(BXT_P_CR_GT_DISP_PWRON,
+					val | MIPIO_RST_CTRL);
 
-	msleep(intel_dsi->panel_on_delay);
+		/* Power up DSI regulator */
+		I915_WRITE(BXT_P_DSI_REGULATOR_CFG, STAP_SELECT);
+		I915_WRITE(BXT_P_DSI_REGULATOR_TX_CTRL, 0);
+	}
 
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
 		u32 val;
@@ -575,17 +812,43 @@ static void intel_dsi_pre_enable(struct intel_encoder *encoder,
 		I915_WRITE(DSPCLK_GATE_D, val);
 	}
 
-	/* put device in ready state */
-	intel_dsi_device_ready(encoder);
+	intel_dsi_prepare(encoder, pipe_config);
+
+	/* Power on, try both CRC pmic gpio and VBT */
+	if (intel_dsi->gpio_panel)
+		gpiod_set_value_cansleep(intel_dsi->gpio_panel, 1);
+	intel_dsi_exec_vbt_sequence(intel_dsi, MIPI_SEQ_POWER_ON);
+	intel_dsi_msleep(intel_dsi, intel_dsi->panel_on_delay);
 
-	drm_panel_prepare(intel_dsi->panel);
+	/* Deassert reset */
+	intel_dsi_exec_vbt_sequence(intel_dsi, MIPI_SEQ_DEASSERT_RESET);
 
-	for_each_dsi_port(port, intel_dsi->ports)
-		wait_for_dsi_fifo_empty(intel_dsi, port);
+	/* Put device in ready state (LP-11) */
+	intel_dsi_device_ready(encoder);
+
+	/* Send initialization commands in LP mode */
+	intel_dsi_exec_vbt_sequence(intel_dsi, MIPI_SEQ_INIT_OTP);
 
 	/* Enable port in pre-enable phase itself because as per hw team
 	 * recommendation, port should be enabled befor plane & pipe */
-	intel_dsi_enable(encoder);
+	if (is_cmd_mode(intel_dsi)) {
+		for_each_dsi_port(port, intel_dsi->ports)
+			I915_WRITE(MIPI_MAX_RETURN_PKT_SIZE(port), 8 * 4);
+		intel_dsi_exec_vbt_sequence(intel_dsi, MIPI_SEQ_TEAR_ON);
+		intel_dsi_exec_vbt_sequence(intel_dsi, MIPI_SEQ_DISPLAY_ON);
+	} else {
+		msleep(20); /* XXX */
+		for_each_dsi_port(port, intel_dsi->ports)
+			dpi_send_cmd(intel_dsi, TURN_ON, false, port);
+		intel_dsi_msleep(intel_dsi, 100);
+
+		intel_dsi_exec_vbt_sequence(intel_dsi, MIPI_SEQ_DISPLAY_ON);
+
+		intel_dsi_port_enable(encoder);
+	}
+
+	intel_panel_enable_backlight(intel_dsi->attached_connector);
+	intel_dsi_exec_vbt_sequence(intel_dsi, MIPI_SEQ_BACKLIGHT_ON);
 }
 
 static void intel_dsi_enable_nop(struct intel_encoder *encoder,
@@ -604,13 +867,30 @@ static void intel_dsi_pre_disable(struct intel_encoder *encoder,
 				  struct intel_crtc_state *old_crtc_state,
 				  struct drm_connector_state *old_conn_state)
 {
+	struct drm_device *dev = encoder->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	enum port port;
 
 	DRM_DEBUG_KMS("\n");
 
+	intel_dsi_exec_vbt_sequence(intel_dsi, MIPI_SEQ_BACKLIGHT_OFF);
 	intel_panel_disable_backlight(intel_dsi->attached_connector);
 
+	/*
+	 * Disable Device ready before the port shutdown in order
+	 * to avoid split screen
+	 */
+	if (IS_BROXTON(dev_priv)) {
+		for_each_dsi_port(port, intel_dsi->ports)
+			I915_WRITE(MIPI_DEVICE_READY(port), 0);
+	}
+
+	/*
+	 * According to the spec we should send SHUTDOWN before
+	 * MIPI_SEQ_DISPLAY_OFF only for v3+ VBTs, but field testing
+	 * has shown that the v3 sequence works for v2 VBTs too
+	 */
 	if (is_vid_mode(intel_dsi)) {
 		/* Send Shutdown command to the panel in LP mode */
 		for_each_dsi_port(port, intel_dsi->ports)
@@ -619,13 +899,25 @@ static void intel_dsi_pre_disable(struct intel_encoder *encoder,
 	}
 }
 
-static void intel_dsi_disable(struct intel_encoder *encoder)
+static void intel_dsi_clear_device_ready(struct intel_encoder *encoder)
 {
-	struct drm_device *dev = encoder->base.dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+
+	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv) ||
+	    IS_BROXTON(dev_priv))
+		vlv_dsi_clear_device_ready(encoder);
+	else if (IS_GEMINILAKE(dev_priv))
+		glk_dsi_clear_device_ready(encoder);
+}
+
+static void intel_dsi_post_disable(struct intel_encoder *encoder,
+				   struct intel_crtc_state *pipe_config,
+				   struct drm_connector_state *conn_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
 	enum port port;
-	u32 temp;
+	u32 val;
 
 	DRM_DEBUG_KMS("\n");
 
@@ -634,85 +926,32 @@ static void intel_dsi_disable(struct intel_encoder *encoder)
 			wait_for_dsi_fifo_empty(intel_dsi, port);
 
 		intel_dsi_port_disable(encoder);
-		msleep(2);
-	}
-
-	for_each_dsi_port(port, intel_dsi->ports) {
-		/* Panel commands can be sent when clock is in LP11 */
-		I915_WRITE(MIPI_DEVICE_READY(port), 0x0);
-
-		intel_dsi_reset_clocks(encoder, port);
-		I915_WRITE(MIPI_EOT_DISABLE(port), CLOCKSTOP);
-
-		temp = I915_READ(MIPI_DSI_FUNC_PRG(port));
-		temp &= ~VID_MODE_FORMAT_MASK;
-		I915_WRITE(MIPI_DSI_FUNC_PRG(port), temp);
-
-		I915_WRITE(MIPI_DEVICE_READY(port), 0x1);
+		usleep_range(2000, 5000);
 	}
-	/* if disable packets are sent before sending shutdown packet then in
-	 * some next enable sequence send turn on packet error is observed */
-	drm_panel_disable(intel_dsi->panel);
 
-	for_each_dsi_port(port, intel_dsi->ports)
-		wait_for_dsi_fifo_empty(intel_dsi, port);
-}
-
-static void intel_dsi_clear_device_ready(struct intel_encoder *encoder)
-{
-	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
-	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
-	enum port port;
-
-	DRM_DEBUG_KMS("\n");
-	for_each_dsi_port(port, intel_dsi->ports) {
-		/* Common bit for both MIPI Port A & MIPI Port C on VLV/CHV */
-		i915_reg_t port_ctrl = IS_GEN9_LP(dev_priv) ?
-			BXT_MIPI_PORT_CTRL(port) : MIPI_PORT_CTRL(PORT_A);
-		u32 val;
+	intel_dsi_unprepare(encoder);
 
-		I915_WRITE(MIPI_DEVICE_READY(port), DEVICE_READY |
-							ULPS_STATE_ENTER);
-		usleep_range(2000, 2500);
-
-		I915_WRITE(MIPI_DEVICE_READY(port), DEVICE_READY |
-							ULPS_STATE_EXIT);
-		usleep_range(2000, 2500);
-
-		I915_WRITE(MIPI_DEVICE_READY(port), DEVICE_READY |
-							ULPS_STATE_ENTER);
-		usleep_range(2000, 2500);
+	/*
+	 * if disable packets are sent before sending shutdown packet then in
+	 * some next enable sequence send turn on packet error is observed
+	 */
+	if (is_cmd_mode(intel_dsi))
+		intel_dsi_exec_vbt_sequence(intel_dsi, MIPI_SEQ_TEAR_OFF);
+	intel_dsi_exec_vbt_sequence(intel_dsi, MIPI_SEQ_DISPLAY_OFF);
 
-		/* Wait till Clock lanes are in LP-00 state for MIPI Port A
-		 * only. MIPI Port C has no similar bit for checking
-		 */
-		if (intel_wait_for_register(dev_priv,
-					    port_ctrl, AFE_LATCHOUT, 0,
-					    30))
-			DRM_ERROR("DSI LP not going Low\n");
+	/* Transition to LP-00 */
+	intel_dsi_clear_device_ready(encoder);
 
-		/* Disable MIPI PHY transparent latch */
-		val = I915_READ(port_ctrl);
-		I915_WRITE(port_ctrl, val & ~LP_OUTPUT_HOLD);
-		usleep_range(1000, 1500);
+	if (IS_BROXTON(dev_priv)) {
+		/* Power down DSI regulator to save power */
+		I915_WRITE(BXT_P_DSI_REGULATOR_CFG, STAP_SELECT);
+		I915_WRITE(BXT_P_DSI_REGULATOR_TX_CTRL, HS_IO_CTRL_SELECT);
 
-		I915_WRITE(MIPI_DEVICE_READY(port), 0x00);
-		usleep_range(2000, 2500);
+		/* Add MIPI IO reset programming for modeset */
+		val = I915_READ(BXT_P_CR_GT_DISP_PWRON);
+		I915_WRITE(BXT_P_CR_GT_DISP_PWRON,
+				val & ~MIPIO_RST_CTRL);
 	}
-}
-
-static void intel_dsi_post_disable(struct intel_encoder *encoder,
-				   struct intel_crtc_state *pipe_config,
-				   struct drm_connector_state *conn_state)
-{
-	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
-	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
-
-	DRM_DEBUG_KMS("\n");
-
-	intel_dsi_disable(encoder);
-
-	intel_dsi_clear_device_ready(encoder);
 
 	intel_disable_dsi_pll(encoder);
 
@@ -724,11 +963,12 @@ static void intel_dsi_post_disable(struct intel_encoder *encoder,
 		I915_WRITE(DSPCLK_GATE_D, val);
 	}
 
-	drm_panel_unprepare(intel_dsi->panel);
-
-	msleep(intel_dsi->panel_off_delay);
+	/* Assert reset */
+	intel_dsi_exec_vbt_sequence(intel_dsi, MIPI_SEQ_ASSERT_RESET);
 
-	/* Panel Disable over CRC PMIC */
+	/* Power off, try both CRC pmic gpio and VBT */
+	intel_dsi_msleep(intel_dsi, intel_dsi->panel_off_delay);
+	intel_dsi_exec_vbt_sequence(intel_dsi, MIPI_SEQ_POWER_OFF);
 	if (intel_dsi->gpio_panel)
 		gpiod_set_value_cansleep(intel_dsi->gpio_panel, 0);
 
@@ -736,7 +976,7 @@ static void intel_dsi_post_disable(struct intel_encoder *encoder,
 	 * FIXME As we do with eDP, just make a note of the time here
 	 * and perform the wait before the next panel power on.
 	 */
-	msleep(intel_dsi->panel_pwr_cycle_delay);
+	intel_dsi_msleep(intel_dsi, intel_dsi->panel_pwr_cycle_delay);
 }
 
 static bool intel_dsi_get_hw_state(struct intel_encoder *encoder,
@@ -744,14 +984,13 @@ static bool intel_dsi_get_hw_state(struct intel_encoder *encoder,
 {
 	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
-	enum intel_display_power_domain power_domain;
 	enum port port;
 	bool active = false;
 
 	DRM_DEBUG_KMS("\n");
 
-	power_domain = intel_display_port_power_domain(encoder);
-	if (!intel_display_power_get_if_enabled(dev_priv, power_domain))
+	if (!intel_display_power_get_if_enabled(dev_priv,
+						encoder->power_domain))
 		return false;
 
 	/*
@@ -807,7 +1046,7 @@ static bool intel_dsi_get_hw_state(struct intel_encoder *encoder,
 	}
 
 out_put_power:
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, encoder->power_domain);
 
 	return active;
 }
@@ -1279,6 +1518,14 @@ static void intel_dsi_prepare(struct intel_encoder *intel_encoder,
 		 */
 		I915_WRITE(MIPI_LP_BYTECLK(port), intel_dsi->lp_byte_clk);
 
+		if (IS_GEMINILAKE(dev_priv)) {
+			I915_WRITE(MIPI_TLPX_TIME_COUNT(port),
+					intel_dsi->lp_byte_clk);
+			/* Shadow of DPHY reg */
+			I915_WRITE(MIPI_CLK_LANE_TIMING(port),
+					intel_dsi->dphy_reg);
+		}
+
 		/* the bw essential for transmitting 16 long packets containing
 		 * 252 bytes meant for dcs write memory command is programmed in
 		 * this register in terms of byte clocks. based on dsi transfer
@@ -1302,6 +1549,30 @@ static void intel_dsi_prepare(struct intel_encoder *intel_encoder,
 	}
 }
 
+static void intel_dsi_unprepare(struct intel_encoder *encoder)
+{
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
+	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
+	enum port port;
+	u32 val;
+
+	if (!IS_GEMINILAKE(dev_priv)) {
+		for_each_dsi_port(port, intel_dsi->ports) {
+			/* Panel commands can be sent when clock is in LP11 */
+			I915_WRITE(MIPI_DEVICE_READY(port), 0x0);
+
+			intel_dsi_reset_clocks(encoder, port);
+			I915_WRITE(MIPI_EOT_DISABLE(port), CLOCKSTOP);
+
+			val = I915_READ(MIPI_DSI_FUNC_PRG(port));
+			val &= ~VID_MODE_FORMAT_MASK;
+			I915_WRITE(MIPI_DSI_FUNC_PRG(port), val);
+
+			I915_WRITE(MIPI_DEVICE_READY(port), 0x1);
+		}
+	}
+}
+
 static int intel_dsi_get_modes(struct drm_connector *connector)
 {
 	struct intel_connector *intel_connector = to_intel_connector(connector);
@@ -1485,6 +1756,7 @@ void intel_dsi_init(struct drm_i915_private *dev_priv)
 	intel_connector->get_hw_state = intel_connector_get_hw_state;
 
 	intel_encoder->port = port;
+
 	/*
 	 * On BYT/CHV, pipe A maps to MIPI DSI port A, pipe B maps to MIPI DSI
 	 * port C. BXT isn't limited like this.
@@ -1560,7 +1832,8 @@ void intel_dsi_init(struct drm_i915_private *dev_priv)
 	 * In case of BYT with CRC PMIC, we need to use GPIO for
 	 * Panel control.
 	 */
-	if (dev_priv->vbt.dsi.config->pwm_blc == PPS_BLC_PMIC) {
+	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
+	    (dev_priv->vbt.dsi.config->pwm_blc == PPS_BLC_PMIC)) {
 		intel_dsi->gpio_panel =
 			gpiod_get(dev->dev, "panel", GPIOD_OUT_HIGH);
 
@@ -1571,6 +1844,7 @@ void intel_dsi_init(struct drm_i915_private *dev_priv)
 	}
 
 	intel_encoder->type = INTEL_OUTPUT_DSI;
+	intel_encoder->power_domain = POWER_DOMAIN_PORT_DSI;
 	intel_encoder->cloneable = 0;
 	drm_connector_init(dev, connector, &intel_dsi_connector_funcs,
 			   DRM_MODE_CONNECTOR_DSI);
diff --git a/drivers/gpu/drm/i915/intel_dsi.h b/drivers/gpu/drm/i915/intel_dsi.h
index 5967ea6d6045..548649158abd 100644
--- a/drivers/gpu/drm/i915/intel_dsi.h
+++ b/drivers/gpu/drm/i915/intel_dsi.h
@@ -130,6 +130,11 @@ static inline struct intel_dsi *enc_to_intel_dsi(struct drm_encoder *encoder)
 	return container_of(encoder, struct intel_dsi, base.base);
 }
 
+void wait_for_dsi_fifo_empty(struct intel_dsi *intel_dsi, enum port port);
+
+void intel_dsi_exec_vbt_sequence(struct intel_dsi *intel_dsi,
+				 enum mipi_seq seq_id);
+
 bool intel_dsi_pll_is_enabled(struct drm_i915_private *dev_priv);
 int intel_compute_dsi_pll(struct intel_encoder *encoder,
 			  struct intel_crtc_state *config);
diff --git a/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c b/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c
index 8f683b8b1816..ab922b6eb36a 100644
--- a/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c
+++ b/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c
@@ -192,6 +192,8 @@ static const u8 *mipi_exec_send_packet(struct intel_dsi *intel_dsi,
 		break;
 	}
 
+	wait_for_dsi_fifo_empty(intel_dsi, port);
+
 out:
 	data += len;
 
@@ -424,10 +426,9 @@ static const char *sequence_name(enum mipi_seq seq_id)
 		return "(unknown)";
 }
 
-static void generic_exec_sequence(struct drm_panel *panel, enum mipi_seq seq_id)
+void intel_dsi_exec_vbt_sequence(struct intel_dsi *intel_dsi,
+				 enum mipi_seq seq_id)
 {
-	struct vbt_panel *vbt_panel = to_vbt_panel(panel);
-	struct intel_dsi *intel_dsi = vbt_panel->intel_dsi;
 	struct drm_i915_private *dev_priv = to_i915(intel_dsi->base.base.dev);
 	const u8 *data;
 	fn_mipi_elem_exec mipi_elem_exec;
@@ -491,40 +492,6 @@ static void generic_exec_sequence(struct drm_panel *panel, enum mipi_seq seq_id)
 	}
 }
 
-static int vbt_panel_prepare(struct drm_panel *panel)
-{
-	generic_exec_sequence(panel, MIPI_SEQ_ASSERT_RESET);
-	generic_exec_sequence(panel, MIPI_SEQ_POWER_ON);
-	generic_exec_sequence(panel, MIPI_SEQ_DEASSERT_RESET);
-	generic_exec_sequence(panel, MIPI_SEQ_INIT_OTP);
-
-	return 0;
-}
-
-static int vbt_panel_unprepare(struct drm_panel *panel)
-{
-	generic_exec_sequence(panel, MIPI_SEQ_ASSERT_RESET);
-	generic_exec_sequence(panel, MIPI_SEQ_POWER_OFF);
-
-	return 0;
-}
-
-static int vbt_panel_enable(struct drm_panel *panel)
-{
-	generic_exec_sequence(panel, MIPI_SEQ_DISPLAY_ON);
-	generic_exec_sequence(panel, MIPI_SEQ_BACKLIGHT_ON);
-
-	return 0;
-}
-
-static int vbt_panel_disable(struct drm_panel *panel)
-{
-	generic_exec_sequence(panel, MIPI_SEQ_BACKLIGHT_OFF);
-	generic_exec_sequence(panel, MIPI_SEQ_DISPLAY_OFF);
-
-	return 0;
-}
-
 static int vbt_panel_get_modes(struct drm_panel *panel)
 {
 	struct vbt_panel *vbt_panel = to_vbt_panel(panel);
@@ -548,10 +515,6 @@ static int vbt_panel_get_modes(struct drm_panel *panel)
 }
 
 static const struct drm_panel_funcs vbt_panel_funcs = {
-	.disable = vbt_panel_disable,
-	.unprepare = vbt_panel_unprepare,
-	.prepare = vbt_panel_prepare,
-	.enable = vbt_panel_enable,
 	.get_modes = vbt_panel_get_modes,
 };
 
@@ -571,6 +534,7 @@ struct drm_panel *vbt_panel_init(struct intel_dsi *intel_dsi, u16 panel_id)
 	u32 tclk_prepare_clkzero, ths_prepare_hszero;
 	u32 lp_to_hs_switch, hs_to_lp_switch;
 	u32 pclk, computed_ddr;
+	u32 mul;
 	u16 burst_mode_ratio;
 	enum port port;
 
@@ -674,11 +638,6 @@ struct drm_panel *vbt_panel_init(struct intel_dsi *intel_dsi, u16 panel_id)
 		break;
 	}
 
-	/*
-	 * ui(s) = 1/f [f in hz]
-	 * ui(ns) = 10^9 / (f*10^6) [f in Mhz] -> 10^3/f(Mhz)
-	 */
-
 	/* in Kbps */
 	ui_num = NS_KHZ_RATIO;
 	ui_den = bitrate;
@@ -692,21 +651,26 @@ struct drm_panel *vbt_panel_init(struct intel_dsi *intel_dsi, u16 panel_id)
 	 */
 	intel_dsi->lp_byte_clk = DIV_ROUND_UP(tlpx_ns * ui_den, 8 * ui_num);
 
-	/* count values in UI = (ns value) * (bitrate / (2 * 10^6))
-	 *
-	 * Since txddrclkhs_i is 2xUI, all the count values programmed in
-	 * DPHY param register are divided by 2
+	/* DDR clock period = 2 * UI
+	 * UI(sec) = 1/(bitrate * 10^3) (bitrate is in KHZ)
+	 * UI(nsec) = 10^6 / bitrate
+	 * DDR clock period (nsec) = 2 * UI = (2 * 10^6)/ bitrate
+	 * DDR clock count  = ns_value / DDR clock period
 	 *
-	 * prepare count
+	 * For GEMINILAKE dphy_param_reg will be programmed in terms of
+	 * HS byte clock count for other platform in HS ddr clock count
 	 */
+	mul = IS_GEMINILAKE(dev_priv) ? 8 : 2;
 	ths_prepare_ns = max(mipi_config->ths_prepare,
 			     mipi_config->tclk_prepare);
-	prepare_cnt = DIV_ROUND_UP(ths_prepare_ns * ui_den, ui_num * 2);
+
+	/* prepare count */
+	prepare_cnt = DIV_ROUND_UP(ths_prepare_ns * ui_den, ui_num * mul);
 
 	/* exit zero count */
 	exit_zero_cnt = DIV_ROUND_UP(
 				(ths_prepare_hszero - ths_prepare_ns) * ui_den,
-				ui_num * 2
+				ui_num * mul
 				);
 
 	/*
@@ -720,12 +684,12 @@ struct drm_panel *vbt_panel_init(struct intel_dsi *intel_dsi, u16 panel_id)
 
 	/* clk zero count */
 	clk_zero_cnt = DIV_ROUND_UP(
-			(tclk_prepare_clkzero -	ths_prepare_ns)
-			* ui_den, 2 * ui_num);
+				(tclk_prepare_clkzero -	ths_prepare_ns)
+				* ui_den, ui_num * mul);
 
 	/* trail count */
 	tclk_trail_ns = max(mipi_config->tclk_trail, mipi_config->ths_trail);
-	trail_cnt = DIV_ROUND_UP(tclk_trail_ns * ui_den, 2 * ui_num);
+	trail_cnt = DIV_ROUND_UP(tclk_trail_ns * ui_den, ui_num * mul);
 
 	if (prepare_cnt > PREPARE_CNT_MAX ||
 		exit_zero_cnt > EXIT_ZERO_CNT_MAX ||
@@ -801,6 +765,19 @@ struct drm_panel *vbt_panel_init(struct intel_dsi *intel_dsi, u16 panel_id)
 			8);
 	intel_dsi->clk_hs_to_lp_count += extra_byte_count;
 
+	DRM_DEBUG_KMS("Pclk %d\n", intel_dsi->pclk);
+	DRM_DEBUG_KMS("Pixel overlap %d\n", intel_dsi->pixel_overlap);
+	DRM_DEBUG_KMS("Lane count %d\n", intel_dsi->lane_count);
+	DRM_DEBUG_KMS("DPHY param reg 0x%x\n", intel_dsi->dphy_reg);
+	DRM_DEBUG_KMS("Video mode format %s\n",
+		      intel_dsi->video_mode_format == VIDEO_MODE_NON_BURST_WITH_SYNC_PULSE ?
+		      "non-burst with sync pulse" :
+		      intel_dsi->video_mode_format == VIDEO_MODE_NON_BURST_WITH_SYNC_EVENTS ?
+		      "non-burst with sync events" :
+		      intel_dsi->video_mode_format == VIDEO_MODE_BURST ?
+		      "burst" : "<unknown>");
+	DRM_DEBUG_KMS("Burst mode ratio %d\n", intel_dsi->burst_mode_ratio);
+	DRM_DEBUG_KMS("Reset timer %d\n", intel_dsi->rst_timer_val);
 	DRM_DEBUG_KMS("Eot %s\n", enableddisabled(intel_dsi->eotp_pkt));
 	DRM_DEBUG_KMS("Clockstop %s\n", enableddisabled(!intel_dsi->clock_stop));
 	DRM_DEBUG_KMS("Mode %s\n", intel_dsi->operation_mode ? "command" : "video");
diff --git a/drivers/gpu/drm/i915/intel_dsi_pll.c b/drivers/gpu/drm/i915/intel_dsi_pll.c
index 61440e5c2563..2ff2ee7f3b78 100644
--- a/drivers/gpu/drm/i915/intel_dsi_pll.c
+++ b/drivers/gpu/drm/i915/intel_dsi_pll.c
@@ -206,17 +206,24 @@ static bool bxt_dsi_pll_is_enabled(struct drm_i915_private *dev_priv)
 		return false;
 
 	/*
-	 * Both dividers must be programmed with valid values even if only one
-	 * of the PLL is used, see BSpec/Broxton Clocks. Check this here for
+	 * Dividers must be programmed with valid values. As per BSEPC, for
+	 * GEMINLAKE only PORT A divider values are checked while for BXT
+	 * both divider values are validated. Check this here for
 	 * paranoia, since BIOS is known to misconfigure PLLs in this way at
 	 * times, and since accessing DSI registers with invalid dividers
 	 * causes a system hang.
 	 */
 	val = I915_READ(BXT_DSI_PLL_CTL);
-	if (!(val & BXT_DSIA_16X_MASK) || !(val & BXT_DSIC_16X_MASK)) {
-		DRM_DEBUG_DRIVER("PLL is enabled with invalid divider settings (%08x)\n",
-				 val);
-		enabled = false;
+	if (IS_GEMINILAKE(dev_priv)) {
+		if (!(val & BXT_DSIA_16X_MASK)) {
+			DRM_DEBUG_DRIVER("Invalid PLL divider (%08x)\n", val);
+			enabled = false;
+		}
+	} else {
+		if (!(val & BXT_DSIA_16X_MASK) || !(val & BXT_DSIC_16X_MASK)) {
+			DRM_DEBUG_DRIVER("Invalid PLL divider (%08x)\n", val);
+			enabled = false;
+		}
 	}
 
 	return enabled;
@@ -372,6 +379,53 @@ static void vlv_dsi_reset_clocks(struct intel_encoder *encoder, enum port port)
 			ESCAPE_CLOCK_DIVIDER_SHIFT);
 }
 
+static void glk_dsi_program_esc_clock(struct drm_device *dev,
+				   const struct intel_crtc_state *config)
+{
+	struct drm_i915_private *dev_priv = to_i915(dev);
+	u32 dsi_rate = 0;
+	u32 pll_ratio = 0;
+	u32 ddr_clk = 0;
+	u32 div1_value = 0;
+	u32 div2_value = 0;
+	u32 txesc1_div = 0;
+	u32 txesc2_div = 0;
+
+	pll_ratio = config->dsi_pll.ctrl & BXT_DSI_PLL_RATIO_MASK;
+
+	dsi_rate = (BXT_REF_CLOCK_KHZ * pll_ratio) / 2;
+
+	ddr_clk = dsi_rate / 2;
+
+	/* Variable divider value */
+	div1_value = DIV_ROUND_CLOSEST(ddr_clk, 20000);
+
+	/* Calculate TXESC1 divider */
+	if (div1_value <= 10)
+		txesc1_div = div1_value;
+	else if ((div1_value > 10) && (div1_value <= 20))
+		txesc1_div = DIV_ROUND_UP(div1_value, 2);
+	else if ((div1_value > 20) && (div1_value <= 30))
+		txesc1_div = DIV_ROUND_UP(div1_value, 4);
+	else if ((div1_value > 30) && (div1_value <= 40))
+		txesc1_div = DIV_ROUND_UP(div1_value, 6);
+	else if ((div1_value > 40) && (div1_value <= 50))
+		txesc1_div = DIV_ROUND_UP(div1_value, 8);
+	else
+		txesc1_div = 10;
+
+	/* Calculate TXESC2 divider */
+	div2_value = DIV_ROUND_UP(div1_value, txesc1_div);
+
+	if (div2_value < 10)
+		txesc2_div = div2_value;
+	else
+		txesc2_div = 10;
+
+	I915_WRITE(MIPIO_TXESC_CLK_DIV1, txesc1_div & GLK_TX_ESC_CLK_DIV1_MASK);
+	I915_WRITE(MIPIO_TXESC_CLK_DIV2, txesc2_div & GLK_TX_ESC_CLK_DIV2_MASK);
+}
+
 /* Program BXT Mipi clocks and dividers */
 static void bxt_dsi_program_clocks(struct drm_device *dev, enum port port,
 				   const struct intel_crtc_state *config)
@@ -416,11 +470,7 @@ static void bxt_dsi_program_clocks(struct drm_device *dev, enum port port,
 	rx_div_lower = rx_div & RX_DIVIDER_BIT_1_2;
 	rx_div_upper = (rx_div & RX_DIVIDER_BIT_3_4) >> 2;
 
-	/* As per bpsec program the 8/3X clock divider to the below value */
-	if (dev_priv->vbt.dsi.config->is_cmd_mode)
-		mipi_8by3_divider = 0x2;
-	else
-		mipi_8by3_divider = 0x3;
+	mipi_8by3_divider = 0x2;
 
 	tmp |= BXT_MIPI_8X_BY3_DIVIDER(port, mipi_8by3_divider);
 	tmp |= BXT_MIPI_TX_ESCLK_DIVIDER(port, tx_div);
@@ -430,11 +480,12 @@ static void bxt_dsi_program_clocks(struct drm_device *dev, enum port port,
 	I915_WRITE(BXT_MIPI_CLOCK_CTL, tmp);
 }
 
-static int bxt_compute_dsi_pll(struct intel_encoder *encoder,
+static int gen9lp_compute_dsi_pll(struct intel_encoder *encoder,
 			       struct intel_crtc_state *config)
 {
+	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 	struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base);
-	u8 dsi_ratio;
+	u8 dsi_ratio, dsi_ratio_min, dsi_ratio_max;
 	u32 dsi_clk;
 
 	dsi_clk = dsi_clk_from_pclk(intel_dsi->pclk, intel_dsi->pixel_format,
@@ -446,11 +497,20 @@ static int bxt_compute_dsi_pll(struct intel_encoder *encoder,
 	 * round 'up' the result
 	 */
 	dsi_ratio = DIV_ROUND_UP(dsi_clk * 2, BXT_REF_CLOCK_KHZ);
-	if (dsi_ratio < BXT_DSI_PLL_RATIO_MIN ||
-	    dsi_ratio > BXT_DSI_PLL_RATIO_MAX) {
+
+	if (IS_BROXTON(dev_priv)) {
+		dsi_ratio_min = BXT_DSI_PLL_RATIO_MIN;
+		dsi_ratio_max = BXT_DSI_PLL_RATIO_MAX;
+	} else {
+		dsi_ratio_min = GLK_DSI_PLL_RATIO_MIN;
+		dsi_ratio_max = GLK_DSI_PLL_RATIO_MAX;
+	}
+
+	if (dsi_ratio < dsi_ratio_min || dsi_ratio > dsi_ratio_max) {
 		DRM_ERROR("Cant get a suitable ratio from DSI PLL ratios\n");
 		return -ECHRNG;
-	}
+	} else
+		DRM_DEBUG_KMS("DSI PLL calculation is Done!!\n");
 
 	/*
 	 * Program DSI ratio and Select MIPIC and MIPIA PLL output as 8x
@@ -462,13 +522,13 @@ static int bxt_compute_dsi_pll(struct intel_encoder *encoder,
 	/* As per recommendation from hardware team,
 	 * Prog PVD ratio =1 if dsi ratio <= 50
 	 */
-	if (dsi_ratio <= 50)
+	if (IS_BROXTON(dev_priv) && dsi_ratio <= 50)
 		config->dsi_pll.ctrl |= BXT_DSI_PLL_PVD_RATIO_1;
 
 	return 0;
 }
 
-static void bxt_enable_dsi_pll(struct intel_encoder *encoder,
+static void gen9lp_enable_dsi_pll(struct intel_encoder *encoder,
 			       const struct intel_crtc_state *config)
 {
 	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
@@ -483,8 +543,12 @@ static void bxt_enable_dsi_pll(struct intel_encoder *encoder,
 	POSTING_READ(BXT_DSI_PLL_CTL);
 
 	/* Program TX, RX, Dphy clocks */
-	for_each_dsi_port(port, intel_dsi->ports)
-		bxt_dsi_program_clocks(encoder->base.dev, port, config);
+	if (IS_BROXTON(dev_priv)) {
+		for_each_dsi_port(port, intel_dsi->ports)
+			bxt_dsi_program_clocks(encoder->base.dev, port, config);
+	} else {
+		glk_dsi_program_esc_clock(encoder->base.dev, config);
+	}
 
 	/* Enable DSI PLL */
 	val = I915_READ(BXT_DSI_PLL_ENABLE);
@@ -522,7 +586,7 @@ int intel_compute_dsi_pll(struct intel_encoder *encoder,
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
 		return vlv_compute_dsi_pll(encoder, config);
 	else if (IS_GEN9_LP(dev_priv))
-		return bxt_compute_dsi_pll(encoder, config);
+		return gen9lp_compute_dsi_pll(encoder, config);
 
 	return -ENODEV;
 }
@@ -535,7 +599,7 @@ void intel_enable_dsi_pll(struct intel_encoder *encoder,
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
 		vlv_enable_dsi_pll(encoder, config);
 	else if (IS_GEN9_LP(dev_priv))
-		bxt_enable_dsi_pll(encoder, config);
+		gen9lp_enable_dsi_pll(encoder, config);
 }
 
 void intel_disable_dsi_pll(struct intel_encoder *encoder)
@@ -548,19 +612,30 @@ void intel_disable_dsi_pll(struct intel_encoder *encoder)
 		bxt_disable_dsi_pll(encoder);
 }
 
-static void bxt_dsi_reset_clocks(struct intel_encoder *encoder, enum port port)
+static void gen9lp_dsi_reset_clocks(struct intel_encoder *encoder,
+				    enum port port)
 {
 	u32 tmp;
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
 
 	/* Clear old configurations */
-	tmp = I915_READ(BXT_MIPI_CLOCK_CTL);
-	tmp &= ~(BXT_MIPI_TX_ESCLK_FIXDIV_MASK(port));
-	tmp &= ~(BXT_MIPI_RX_ESCLK_UPPER_FIXDIV_MASK(port));
-	tmp &= ~(BXT_MIPI_8X_BY3_DIVIDER_MASK(port));
-	tmp &= ~(BXT_MIPI_RX_ESCLK_LOWER_FIXDIV_MASK(port));
-	I915_WRITE(BXT_MIPI_CLOCK_CTL, tmp);
+	if (IS_BROXTON(dev_priv)) {
+		tmp = I915_READ(BXT_MIPI_CLOCK_CTL);
+		tmp &= ~(BXT_MIPI_TX_ESCLK_FIXDIV_MASK(port));
+		tmp &= ~(BXT_MIPI_RX_ESCLK_UPPER_FIXDIV_MASK(port));
+		tmp &= ~(BXT_MIPI_8X_BY3_DIVIDER_MASK(port));
+		tmp &= ~(BXT_MIPI_RX_ESCLK_LOWER_FIXDIV_MASK(port));
+		I915_WRITE(BXT_MIPI_CLOCK_CTL, tmp);
+	} else {
+		tmp = I915_READ(MIPIO_TXESC_CLK_DIV1);
+		tmp &= ~GLK_TX_ESC_CLK_DIV1_MASK;
+		I915_WRITE(MIPIO_TXESC_CLK_DIV1, tmp);
+
+		tmp = I915_READ(MIPIO_TXESC_CLK_DIV2);
+		tmp &= ~GLK_TX_ESC_CLK_DIV2_MASK;
+		I915_WRITE(MIPIO_TXESC_CLK_DIV2, tmp);
+	}
 	I915_WRITE(MIPI_EOT_DISABLE(port), CLOCKSTOP);
 }
 
@@ -569,7 +644,7 @@ void intel_dsi_reset_clocks(struct intel_encoder *encoder, enum port port)
 	struct drm_i915_private *dev_priv = to_i915(encoder->base.dev);
 
 	if (IS_GEN9_LP(dev_priv))
-		bxt_dsi_reset_clocks(encoder, port);
+		gen9lp_dsi_reset_clocks(encoder, port);
 	else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
 		vlv_dsi_reset_clocks(encoder, port);
 }
diff --git a/drivers/gpu/drm/i915/intel_dvo.c b/drivers/gpu/drm/i915/intel_dvo.c
index 50da89dcb92b..6025839ed3b7 100644
--- a/drivers/gpu/drm/i915/intel_dvo.c
+++ b/drivers/gpu/drm/i915/intel_dvo.c
@@ -515,6 +515,7 @@ void intel_dvo_init(struct drm_i915_private *dev_priv)
 				 "DVO %c", port_name(port));
 
 		intel_encoder->type = INTEL_OUTPUT_DVO;
+		intel_encoder->power_domain = POWER_DOMAIN_PORT_OTHER;
 		intel_encoder->port = port;
 		intel_encoder->crtc_mask = (1 << 0) | (1 << 1);
 
diff --git a/drivers/gpu/drm/i915/intel_engine_cs.c b/drivers/gpu/drm/i915/intel_engine_cs.c
index 371acf109e34..73fe718516a5 100644
--- a/drivers/gpu/drm/i915/intel_engine_cs.c
+++ b/drivers/gpu/drm/i915/intel_engine_cs.c
@@ -28,8 +28,8 @@
 
 static const struct engine_info {
 	const char *name;
-	unsigned exec_id;
-	enum intel_engine_hw_id hw_id;
+	unsigned int exec_id;
+	unsigned int hw_id;
 	u32 mmio_base;
 	unsigned irq_shift;
 	int (*init_legacy)(struct intel_engine_cs *engine);
@@ -110,21 +110,20 @@ intel_engine_setup(struct drm_i915_private *dev_priv,
 }
 
 /**
- * intel_engines_init() - allocate, populate and init the Engine Command Streamers
+ * intel_engines_init_early() - allocate the Engine Command Streamers
  * @dev_priv: i915 device private
  *
  * Return: non-zero if the initialization failed.
  */
-int intel_engines_init(struct drm_i915_private *dev_priv)
+int intel_engines_init_early(struct drm_i915_private *dev_priv)
 {
 	struct intel_device_info *device_info = mkwrite_device_info(dev_priv);
 	unsigned int ring_mask = INTEL_INFO(dev_priv)->ring_mask;
 	unsigned int mask = 0;
-	int (*init)(struct intel_engine_cs *engine);
 	struct intel_engine_cs *engine;
 	enum intel_engine_id id;
 	unsigned int i;
-	int ret;
+	int err;
 
 	WARN_ON(ring_mask == 0);
 	WARN_ON(ring_mask &
@@ -134,20 +133,8 @@ int intel_engines_init(struct drm_i915_private *dev_priv)
 		if (!HAS_ENGINE(dev_priv, i))
 			continue;
 
-		if (i915.enable_execlists)
-			init = intel_engines[i].init_execlists;
-		else
-			init = intel_engines[i].init_legacy;
-
-		if (!init)
-			continue;
-
-		ret = intel_engine_setup(dev_priv, i);
-		if (ret)
-			goto cleanup;
-
-		ret = init(dev_priv->engine[i]);
-		if (ret)
+		err = intel_engine_setup(dev_priv, i);
+		if (err)
 			goto cleanup;
 
 		mask |= ENGINE_MASK(i);
@@ -166,14 +153,67 @@ int intel_engines_init(struct drm_i915_private *dev_priv)
 	return 0;
 
 cleanup:
+	for_each_engine(engine, dev_priv, id)
+		kfree(engine);
+	return err;
+}
+
+/**
+ * intel_engines_init() - allocate, populate and init the Engine Command Streamers
+ * @dev_priv: i915 device private
+ *
+ * Return: non-zero if the initialization failed.
+ */
+int intel_engines_init(struct drm_i915_private *dev_priv)
+{
+	struct intel_device_info *device_info = mkwrite_device_info(dev_priv);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id, err_id;
+	unsigned int mask = 0;
+	int err = 0;
+
 	for_each_engine(engine, dev_priv, id) {
+		int (*init)(struct intel_engine_cs *engine);
+
 		if (i915.enable_execlists)
-			intel_logical_ring_cleanup(engine);
+			init = intel_engines[id].init_execlists;
 		else
-			intel_engine_cleanup(engine);
+			init = intel_engines[id].init_legacy;
+		if (!init) {
+			kfree(engine);
+			dev_priv->engine[id] = NULL;
+			continue;
+		}
+
+		err = init(engine);
+		if (err) {
+			err_id = id;
+			goto cleanup;
+		}
+
+		mask |= ENGINE_MASK(id);
 	}
 
-	return ret;
+	/*
+	 * Catch failures to update intel_engines table when the new engines
+	 * are added to the driver by a warning and disabling the forgotten
+	 * engines.
+	 */
+	if (WARN_ON(mask != INTEL_INFO(dev_priv)->ring_mask))
+		device_info->ring_mask = mask;
+
+	device_info->num_rings = hweight32(mask);
+
+	return 0;
+
+cleanup:
+	for_each_engine(engine, dev_priv, id) {
+		if (id >= err_id)
+			kfree(engine);
+		else
+			dev_priv->gt.cleanup_engine(engine);
+	}
+	return err;
 }
 
 void intel_engine_init_global_seqno(struct intel_engine_cs *engine, u32 seqno)
@@ -212,8 +252,6 @@ void intel_engine_init_global_seqno(struct intel_engine_cs *engine, u32 seqno)
 		engine->irq_seqno_barrier(engine);
 
 	GEM_BUG_ON(i915_gem_active_isset(&engine->timeline->last_request));
-	engine->timeline->last_submitted_seqno = seqno;
-
 	engine->hangcheck.seqno = seqno;
 
 	/* After manually advancing the seqno, fake the interrupt in case
@@ -482,3 +520,601 @@ void intel_engine_get_instdone(struct intel_engine_cs *engine,
 		break;
 	}
 }
+
+static int wa_add(struct drm_i915_private *dev_priv,
+		  i915_reg_t addr,
+		  const u32 mask, const u32 val)
+{
+	const u32 idx = dev_priv->workarounds.count;
+
+	if (WARN_ON(idx >= I915_MAX_WA_REGS))
+		return -ENOSPC;
+
+	dev_priv->workarounds.reg[idx].addr = addr;
+	dev_priv->workarounds.reg[idx].value = val;
+	dev_priv->workarounds.reg[idx].mask = mask;
+
+	dev_priv->workarounds.count++;
+
+	return 0;
+}
+
+#define WA_REG(addr, mask, val) do { \
+		const int r = wa_add(dev_priv, (addr), (mask), (val)); \
+		if (r) \
+			return r; \
+	} while (0)
+
+#define WA_SET_BIT_MASKED(addr, mask) \
+	WA_REG(addr, (mask), _MASKED_BIT_ENABLE(mask))
+
+#define WA_CLR_BIT_MASKED(addr, mask) \
+	WA_REG(addr, (mask), _MASKED_BIT_DISABLE(mask))
+
+#define WA_SET_FIELD_MASKED(addr, mask, value) \
+	WA_REG(addr, mask, _MASKED_FIELD(mask, value))
+
+#define WA_SET_BIT(addr, mask) WA_REG(addr, mask, I915_READ(addr) | (mask))
+#define WA_CLR_BIT(addr, mask) WA_REG(addr, mask, I915_READ(addr) & ~(mask))
+
+#define WA_WRITE(addr, val) WA_REG(addr, 0xffffffff, val)
+
+static int wa_ring_whitelist_reg(struct intel_engine_cs *engine,
+				 i915_reg_t reg)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+	struct i915_workarounds *wa = &dev_priv->workarounds;
+	const uint32_t index = wa->hw_whitelist_count[engine->id];
+
+	if (WARN_ON(index >= RING_MAX_NONPRIV_SLOTS))
+		return -EINVAL;
+
+	WA_WRITE(RING_FORCE_TO_NONPRIV(engine->mmio_base, index),
+		 i915_mmio_reg_offset(reg));
+	wa->hw_whitelist_count[engine->id]++;
+
+	return 0;
+}
+
+static int gen8_init_workarounds(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+
+	WA_SET_BIT_MASKED(INSTPM, INSTPM_FORCE_ORDERING);
+
+	/* WaDisableAsyncFlipPerfMode:bdw,chv */
+	WA_SET_BIT_MASKED(MI_MODE, ASYNC_FLIP_PERF_DISABLE);
+
+	/* WaDisablePartialInstShootdown:bdw,chv */
+	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN,
+			  PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE);
+
+	/* Use Force Non-Coherent whenever executing a 3D context. This is a
+	 * workaround for for a possible hang in the unlikely event a TLB
+	 * invalidation occurs during a PSD flush.
+	 */
+	/* WaForceEnableNonCoherent:bdw,chv */
+	/* WaHdcDisableFetchWhenMasked:bdw,chv */
+	WA_SET_BIT_MASKED(HDC_CHICKEN0,
+			  HDC_DONOT_FETCH_MEM_WHEN_MASKED |
+			  HDC_FORCE_NON_COHERENT);
+
+	/* From the Haswell PRM, Command Reference: Registers, CACHE_MODE_0:
+	 * "The Hierarchical Z RAW Stall Optimization allows non-overlapping
+	 *  polygons in the same 8x4 pixel/sample area to be processed without
+	 *  stalling waiting for the earlier ones to write to Hierarchical Z
+	 *  buffer."
+	 *
+	 * This optimization is off by default for BDW and CHV; turn it on.
+	 */
+	WA_CLR_BIT_MASKED(CACHE_MODE_0_GEN7, HIZ_RAW_STALL_OPT_DISABLE);
+
+	/* Wa4x4STCOptimizationDisable:bdw,chv */
+	WA_SET_BIT_MASKED(CACHE_MODE_1, GEN8_4x4_STC_OPTIMIZATION_DISABLE);
+
+	/*
+	 * BSpec recommends 8x4 when MSAA is used,
+	 * however in practice 16x4 seems fastest.
+	 *
+	 * Note that PS/WM thread counts depend on the WIZ hashing
+	 * disable bit, which we don't touch here, but it's good
+	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
+	 */
+	WA_SET_FIELD_MASKED(GEN7_GT_MODE,
+			    GEN6_WIZ_HASHING_MASK,
+			    GEN6_WIZ_HASHING_16x4);
+
+	return 0;
+}
+
+static int bdw_init_workarounds(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+	int ret;
+
+	ret = gen8_init_workarounds(engine);
+	if (ret)
+		return ret;
+
+	/* WaDisableThreadStallDopClockGating:bdw (pre-production) */
+	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE);
+
+	/* WaDisableDopClockGating:bdw
+	 *
+	 * Also see the related UCGTCL1 write in broadwell_init_clock_gating()
+	 * to disable EUTC clock gating.
+	 */
+	WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2,
+			  DOP_CLOCK_GATING_DISABLE);
+
+	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3,
+			  GEN8_SAMPLER_POWER_BYPASS_DIS);
+
+	WA_SET_BIT_MASKED(HDC_CHICKEN0,
+			  /* WaForceContextSaveRestoreNonCoherent:bdw */
+			  HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT |
+			  /* WaDisableFenceDestinationToSLM:bdw (pre-prod) */
+			  (IS_BDW_GT3(dev_priv) ? HDC_FENCE_DEST_SLM_DISABLE : 0));
+
+	return 0;
+}
+
+static int chv_init_workarounds(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+	int ret;
+
+	ret = gen8_init_workarounds(engine);
+	if (ret)
+		return ret;
+
+	/* WaDisableThreadStallDopClockGating:chv */
+	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE);
+
+	/* Improve HiZ throughput on CHV. */
+	WA_SET_BIT_MASKED(HIZ_CHICKEN, CHV_HZ_8X8_MODE_IN_1X);
+
+	return 0;
+}
+
+static int gen9_init_workarounds(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+	int ret;
+
+	/* WaConextSwitchWithConcurrentTLBInvalidate:skl,bxt,kbl,glk */
+	I915_WRITE(GEN9_CSFE_CHICKEN1_RCS, _MASKED_BIT_ENABLE(GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE));
+
+	/* WaEnableLbsSlaRetryTimerDecrement:skl,bxt,kbl,glk */
+	I915_WRITE(BDW_SCRATCH1, I915_READ(BDW_SCRATCH1) |
+		   GEN9_LBS_SLA_RETRY_TIMER_DECREMENT_ENABLE);
+
+	/* WaDisableKillLogic:bxt,skl,kbl */
+	I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) |
+		   ECOCHK_DIS_TLB);
+
+	/* WaClearFlowControlGpgpuContextSave:skl,bxt,kbl,glk */
+	/* WaDisablePartialInstShootdown:skl,bxt,kbl,glk */
+	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN,
+			  FLOW_CONTROL_ENABLE |
+			  PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE);
+
+	/* Syncing dependencies between camera and graphics:skl,bxt,kbl */
+	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3,
+			  GEN9_DISABLE_OCL_OOB_SUPPRESS_LOGIC);
+
+	/* WaDisableDgMirrorFixInHalfSliceChicken5:bxt */
+	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1))
+		WA_CLR_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN5,
+				  GEN9_DG_MIRROR_FIX_ENABLE);
+
+	/* WaSetDisablePixMaskCammingAndRhwoInCommonSliceChicken:bxt */
+	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1)) {
+		WA_SET_BIT_MASKED(GEN7_COMMON_SLICE_CHICKEN1,
+				  GEN9_RHWO_OPTIMIZATION_DISABLE);
+		/*
+		 * WA also requires GEN9_SLICE_COMMON_ECO_CHICKEN0[14:14] to be set
+		 * but we do that in per ctx batchbuffer as there is an issue
+		 * with this register not getting restored on ctx restore
+		 */
+	}
+
+	/* WaEnableSamplerGPGPUPreemptionSupport:skl,bxt,kbl */
+	WA_SET_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN7,
+			  GEN9_ENABLE_GPGPU_PREEMPTION);
+
+	/* Wa4x4STCOptimizationDisable:skl,bxt,kbl,glk */
+	/* WaDisablePartialResolveInVc:skl,bxt,kbl */
+	WA_SET_BIT_MASKED(CACHE_MODE_1, (GEN8_4x4_STC_OPTIMIZATION_DISABLE |
+					 GEN9_PARTIAL_RESOLVE_IN_VC_DISABLE));
+
+	/* WaCcsTlbPrefetchDisable:skl,bxt,kbl,glk */
+	WA_CLR_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN5,
+			  GEN9_CCS_TLB_PREFETCH_ENABLE);
+
+	/* WaDisableMaskBasedCammingInRCC:bxt */
+	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1))
+		WA_SET_BIT_MASKED(SLICE_ECO_CHICKEN0,
+				  PIXEL_MASK_CAMMING_DISABLE);
+
+	/* WaForceContextSaveRestoreNonCoherent:skl,bxt,kbl */
+	WA_SET_BIT_MASKED(HDC_CHICKEN0,
+			  HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT |
+			  HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE);
+
+	/* WaForceEnableNonCoherent and WaDisableHDCInvalidation are
+	 * both tied to WaForceContextSaveRestoreNonCoherent
+	 * in some hsds for skl. We keep the tie for all gen9. The
+	 * documentation is a bit hazy and so we want to get common behaviour,
+	 * even though there is no clear evidence we would need both on kbl/bxt.
+	 * This area has been source of system hangs so we play it safe
+	 * and mimic the skl regardless of what bspec says.
+	 *
+	 * Use Force Non-Coherent whenever executing a 3D context. This
+	 * is a workaround for a possible hang in the unlikely event
+	 * a TLB invalidation occurs during a PSD flush.
+	 */
+
+	/* WaForceEnableNonCoherent:skl,bxt,kbl */
+	WA_SET_BIT_MASKED(HDC_CHICKEN0,
+			  HDC_FORCE_NON_COHERENT);
+
+	/* WaDisableHDCInvalidation:skl,bxt,kbl */
+	I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) |
+		   BDW_DISABLE_HDC_INVALIDATION);
+
+	/* WaDisableSamplerPowerBypassForSOPingPong:skl,bxt,kbl */
+	if (IS_SKYLAKE(dev_priv) ||
+	    IS_KABYLAKE(dev_priv) ||
+	    IS_BXT_REVID(dev_priv, 0, BXT_REVID_B0))
+		WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3,
+				  GEN8_SAMPLER_POWER_BYPASS_DIS);
+
+	/* WaDisableSTUnitPowerOptimization:skl,bxt,kbl,glk */
+	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN2, GEN8_ST_PO_DISABLE);
+
+	/* WaOCLCoherentLineFlush:skl,bxt,kbl */
+	I915_WRITE(GEN8_L3SQCREG4, (I915_READ(GEN8_L3SQCREG4) |
+				    GEN8_LQSC_FLUSH_COHERENT_LINES));
+
+	/* WaVFEStateAfterPipeControlwithMediaStateClear:skl,bxt,glk */
+	ret = wa_ring_whitelist_reg(engine, GEN9_CTX_PREEMPT_REG);
+	if (ret)
+		return ret;
+
+	/* WaEnablePreemptionGranularityControlByUMD:skl,bxt,kbl */
+	ret= wa_ring_whitelist_reg(engine, GEN8_CS_CHICKEN1);
+	if (ret)
+		return ret;
+
+	/* WaAllowUMDToModifyHDCChicken1:skl,bxt,kbl,glk */
+	ret = wa_ring_whitelist_reg(engine, GEN8_HDC_CHICKEN1);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int skl_tune_iz_hashing(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+	u8 vals[3] = { 0, 0, 0 };
+	unsigned int i;
+
+	for (i = 0; i < 3; i++) {
+		u8 ss;
+
+		/*
+		 * Only consider slices where one, and only one, subslice has 7
+		 * EUs
+		 */
+		if (!is_power_of_2(INTEL_INFO(dev_priv)->sseu.subslice_7eu[i]))
+			continue;
+
+		/*
+		 * subslice_7eu[i] != 0 (because of the check above) and
+		 * ss_max == 4 (maximum number of subslices possible per slice)
+		 *
+		 * ->    0 <= ss <= 3;
+		 */
+		ss = ffs(INTEL_INFO(dev_priv)->sseu.subslice_7eu[i]) - 1;
+		vals[i] = 3 - ss;
+	}
+
+	if (vals[0] == 0 && vals[1] == 0 && vals[2] == 0)
+		return 0;
+
+	/* Tune IZ hashing. See intel_device_info_runtime_init() */
+	WA_SET_FIELD_MASKED(GEN7_GT_MODE,
+			    GEN9_IZ_HASHING_MASK(2) |
+			    GEN9_IZ_HASHING_MASK(1) |
+			    GEN9_IZ_HASHING_MASK(0),
+			    GEN9_IZ_HASHING(2, vals[2]) |
+			    GEN9_IZ_HASHING(1, vals[1]) |
+			    GEN9_IZ_HASHING(0, vals[0]));
+
+	return 0;
+}
+
+static int skl_init_workarounds(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+	int ret;
+
+	ret = gen9_init_workarounds(engine);
+	if (ret)
+		return ret;
+
+	/*
+	 * Actual WA is to disable percontext preemption granularity control
+	 * until D0 which is the default case so this is equivalent to
+	 * !WaDisablePerCtxtPreemptionGranularityControl:skl
+	 */
+	I915_WRITE(GEN7_FF_SLICE_CS_CHICKEN1,
+		   _MASKED_BIT_ENABLE(GEN9_FFSC_PERCTX_PREEMPT_CTRL));
+
+	/* WaEnableGapsTsvCreditFix:skl */
+	I915_WRITE(GEN8_GARBCNTL, (I915_READ(GEN8_GARBCNTL) |
+				   GEN9_GAPS_TSV_CREDIT_DISABLE));
+
+	/* WaDisableGafsUnitClkGating:skl */
+	WA_SET_BIT(GEN7_UCGCTL4, GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE);
+
+	/* WaInPlaceDecompressionHang:skl */
+	if (IS_SKL_REVID(dev_priv, SKL_REVID_H0, REVID_FOREVER))
+		WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA,
+			   GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS);
+
+	/* WaDisableLSQCROPERFforOCL:skl */
+	ret = wa_ring_whitelist_reg(engine, GEN8_L3SQCREG4);
+	if (ret)
+		return ret;
+
+	return skl_tune_iz_hashing(engine);
+}
+
+static int bxt_init_workarounds(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+	int ret;
+
+	ret = gen9_init_workarounds(engine);
+	if (ret)
+		return ret;
+
+	/* WaStoreMultiplePTEenable:bxt */
+	/* This is a requirement according to Hardware specification */
+	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1))
+		I915_WRITE(TILECTL, I915_READ(TILECTL) | TILECTL_TLBPF);
+
+	/* WaSetClckGatingDisableMedia:bxt */
+	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1)) {
+		I915_WRITE(GEN7_MISCCPCTL, (I915_READ(GEN7_MISCCPCTL) &
+					    ~GEN8_DOP_CLOCK_GATE_MEDIA_ENABLE));
+	}
+
+	/* WaDisableThreadStallDopClockGating:bxt */
+	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN,
+			  STALL_DOP_GATING_DISABLE);
+
+	/* WaDisablePooledEuLoadBalancingFix:bxt */
+	if (IS_BXT_REVID(dev_priv, BXT_REVID_B0, REVID_FOREVER)) {
+		WA_SET_BIT_MASKED(FF_SLICE_CS_CHICKEN2,
+				  GEN9_POOLED_EU_LOAD_BALANCING_FIX_DISABLE);
+	}
+
+	/* WaDisableSbeCacheDispatchPortSharing:bxt */
+	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_B0)) {
+		WA_SET_BIT_MASKED(
+			GEN7_HALF_SLICE_CHICKEN1,
+			GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
+	}
+
+	/* WaDisableObjectLevelPreemptionForTrifanOrPolygon:bxt */
+	/* WaDisableObjectLevelPreemptionForInstancedDraw:bxt */
+	/* WaDisableObjectLevelPreemtionForInstanceId:bxt */
+	/* WaDisableLSQCROPERFforOCL:bxt */
+	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1)) {
+		ret = wa_ring_whitelist_reg(engine, GEN9_CS_DEBUG_MODE1);
+		if (ret)
+			return ret;
+
+		ret = wa_ring_whitelist_reg(engine, GEN8_L3SQCREG4);
+		if (ret)
+			return ret;
+	}
+
+	/* WaProgramL3SqcReg1DefaultForPerf:bxt */
+	if (IS_BXT_REVID(dev_priv, BXT_REVID_B0, REVID_FOREVER))
+		I915_WRITE(GEN8_L3SQCREG1, L3_GENERAL_PRIO_CREDITS(62) |
+					   L3_HIGH_PRIO_CREDITS(2));
+
+	/* WaToEnableHwFixForPushConstHWBug:bxt */
+	if (IS_BXT_REVID(dev_priv, BXT_REVID_C0, REVID_FOREVER))
+		WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
+				  GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
+
+	/* WaInPlaceDecompressionHang:bxt */
+	if (IS_BXT_REVID(dev_priv, BXT_REVID_C0, REVID_FOREVER))
+		WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA,
+			   GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS);
+
+	return 0;
+}
+
+static int kbl_init_workarounds(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+	int ret;
+
+	ret = gen9_init_workarounds(engine);
+	if (ret)
+		return ret;
+
+	/* WaEnableGapsTsvCreditFix:kbl */
+	I915_WRITE(GEN8_GARBCNTL, (I915_READ(GEN8_GARBCNTL) |
+				   GEN9_GAPS_TSV_CREDIT_DISABLE));
+
+	/* WaDisableDynamicCreditSharing:kbl */
+	if (IS_KBL_REVID(dev_priv, 0, KBL_REVID_B0))
+		WA_SET_BIT(GAMT_CHKN_BIT_REG,
+			   GAMT_CHKN_DISABLE_DYNAMIC_CREDIT_SHARING);
+
+	/* WaDisableFenceDestinationToSLM:kbl (pre-prod) */
+	if (IS_KBL_REVID(dev_priv, KBL_REVID_A0, KBL_REVID_A0))
+		WA_SET_BIT_MASKED(HDC_CHICKEN0,
+				  HDC_FENCE_DEST_SLM_DISABLE);
+
+	/* WaToEnableHwFixForPushConstHWBug:kbl */
+	if (IS_KBL_REVID(dev_priv, KBL_REVID_C0, REVID_FOREVER))
+		WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
+				  GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
+
+	/* WaDisableGafsUnitClkGating:kbl */
+	WA_SET_BIT(GEN7_UCGCTL4, GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE);
+
+	/* WaDisableSbeCacheDispatchPortSharing:kbl */
+	WA_SET_BIT_MASKED(
+		GEN7_HALF_SLICE_CHICKEN1,
+		GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
+
+	/* WaInPlaceDecompressionHang:kbl */
+	WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA,
+		   GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS);
+
+	/* WaDisableLSQCROPERFforOCL:kbl */
+	ret = wa_ring_whitelist_reg(engine, GEN8_L3SQCREG4);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int glk_init_workarounds(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+	int ret;
+
+	ret = gen9_init_workarounds(engine);
+	if (ret)
+		return ret;
+
+	/* WaToEnableHwFixForPushConstHWBug:glk */
+	WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
+			  GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
+
+	return 0;
+}
+
+int init_workarounds_ring(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+	int err;
+
+	WARN_ON(engine->id != RCS);
+
+	dev_priv->workarounds.count = 0;
+	dev_priv->workarounds.hw_whitelist_count[engine->id] = 0;
+
+	if (IS_BROADWELL(dev_priv))
+		err = bdw_init_workarounds(engine);
+	else if (IS_CHERRYVIEW(dev_priv))
+		err = chv_init_workarounds(engine);
+	else if (IS_SKYLAKE(dev_priv))
+		err =  skl_init_workarounds(engine);
+	else if (IS_BROXTON(dev_priv))
+		err = bxt_init_workarounds(engine);
+	else if (IS_KABYLAKE(dev_priv))
+		err = kbl_init_workarounds(engine);
+	else if (IS_GEMINILAKE(dev_priv))
+		err =  glk_init_workarounds(engine);
+	else
+		err = 0;
+	if (err)
+		return err;
+
+	DRM_DEBUG_DRIVER("%s: Number of context specific w/a: %d\n",
+			 engine->name, dev_priv->workarounds.count);
+	return 0;
+}
+
+int intel_ring_workarounds_emit(struct drm_i915_gem_request *req)
+{
+	struct i915_workarounds *w = &req->i915->workarounds;
+	u32 *cs;
+	int ret, i;
+
+	if (w->count == 0)
+		return 0;
+
+	ret = req->engine->emit_flush(req, EMIT_BARRIER);
+	if (ret)
+		return ret;
+
+	cs = intel_ring_begin(req, (w->count * 2 + 2));
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = MI_LOAD_REGISTER_IMM(w->count);
+	for (i = 0; i < w->count; i++) {
+		*cs++ = i915_mmio_reg_offset(w->reg[i].addr);
+		*cs++ = w->reg[i].value;
+	}
+	*cs++ = MI_NOOP;
+
+	intel_ring_advance(req, cs);
+
+	ret = req->engine->emit_flush(req, EMIT_BARRIER);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+/**
+ * intel_engine_is_idle() - Report if the engine has finished process all work
+ * @engine: the intel_engine_cs
+ *
+ * Return true if there are no requests pending, nothing left to be submitted
+ * to hardware, and that the engine is idle.
+ */
+bool intel_engine_is_idle(struct intel_engine_cs *engine)
+{
+	struct drm_i915_private *dev_priv = engine->i915;
+
+	/* Any inflight/incomplete requests? */
+	if (!i915_seqno_passed(intel_engine_get_seqno(engine),
+			       intel_engine_last_submit(engine)))
+		return false;
+
+	/* Interrupt/tasklet pending? */
+	if (test_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted))
+		return false;
+
+	/* Both ports drained, no more ELSP submission? */
+	if (engine->execlist_port[0].request)
+		return false;
+
+	/* Ring stopped? */
+	if (INTEL_GEN(dev_priv) > 2 && !(I915_READ_MODE(engine) & MODE_IDLE))
+		return false;
+
+	return true;
+}
+
+bool intel_engines_are_idle(struct drm_i915_private *dev_priv)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	for_each_engine(engine, dev_priv, id) {
+		if (!intel_engine_is_idle(engine))
+			return false;
+	}
+
+	return true;
+}
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/mock_engine.c"
+#endif
diff --git a/drivers/gpu/drm/i915/intel_fbc.c b/drivers/gpu/drm/i915/intel_fbc.c
index 89fe5c8464df..17d418b23d77 100644
--- a/drivers/gpu/drm/i915/intel_fbc.c
+++ b/drivers/gpu/drm/i915/intel_fbc.c
@@ -537,8 +537,7 @@ static int find_compression_threshold(struct drm_i915_private *dev_priv,
 	 * reserved range size, so it always assumes the maximum (8mb) is used.
 	 * If we enable FBC using a CFB on that memory range we'll get FIFO
 	 * underruns, even if that range is not reserved by the BIOS. */
-	if (IS_BROADWELL(dev_priv) ||
-	    IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+	if (IS_BROADWELL(dev_priv) || IS_GEN9_BC(dev_priv))
 		end = ggtt->stolen_size - 8 * 1024 * 1024;
 	else
 		end = U64_MAX;
@@ -628,7 +627,8 @@ err_fb:
 	kfree(compressed_llb);
 	i915_gem_stolen_remove_node(dev_priv, &fbc->compressed_fb);
 err_llb:
-	pr_info_once("drm: not enough stolen space for compressed buffer (need %d more bytes), disabling. Hint: you may be able to increase stolen memory size in the BIOS to avoid this.\n", size);
+	if (drm_mm_initialized(&dev_priv->mm.stolen))
+		pr_info_once("drm: not enough stolen space for compressed buffer (need %d more bytes), disabling. Hint: you may be able to increase stolen memory size in the BIOS to avoid this.\n", size);
 	return -ENOSPC;
 }
 
@@ -743,8 +743,7 @@ static void intel_fbc_update_state_cache(struct intel_crtc *crtc,
 
 	cache->crtc.mode_flags = crtc_state->base.adjusted_mode.flags;
 	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
-		cache->crtc.hsw_bdw_pixel_rate =
-			ilk_pipe_pixel_rate(crtc_state);
+		cache->crtc.hsw_bdw_pixel_rate = crtc_state->pixel_rate;
 
 	cache->plane.rotation = plane_state->base.rotation;
 	cache->plane.src_w = drm_rect_width(&plane_state->base.src) >> 16;
@@ -819,7 +818,7 @@ static bool intel_fbc_can_activate(struct intel_crtc *crtc)
 
 	/* WaFbcExceedCdClockThreshold:hsw,bdw */
 	if ((IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv)) &&
-	    cache->crtc.hsw_bdw_pixel_rate >= dev_priv->cdclk_freq * 95 / 100) {
+	    cache->crtc.hsw_bdw_pixel_rate >= dev_priv->cdclk.hw.cdclk * 95 / 100) {
 		fbc->no_fbc_reason = "pixel rate is too big";
 		return false;
 	}
diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c
index 281c5c48a84d..f7e9a4e69595 100644
--- a/drivers/gpu/drm/i915/intel_fbdev.c
+++ b/drivers/gpu/drm/i915/intel_fbdev.c
@@ -45,6 +45,14 @@
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
 
+static void intel_fbdev_invalidate(struct intel_fbdev *ifbdev)
+{
+	struct drm_i915_gem_object *obj = ifbdev->fb->obj;
+	unsigned int origin = ifbdev->vma->fence ? ORIGIN_GTT : ORIGIN_CPU;
+
+	intel_fb_obj_invalidate(obj, origin);
+}
+
 static int intel_fbdev_set_par(struct fb_info *info)
 {
 	struct drm_fb_helper *fb_helper = info->par;
@@ -53,12 +61,8 @@ static int intel_fbdev_set_par(struct fb_info *info)
 	int ret;
 
 	ret = drm_fb_helper_set_par(info);
-
-	if (ret == 0) {
-		mutex_lock(&fb_helper->dev->struct_mutex);
-		intel_fb_obj_invalidate(ifbdev->fb->obj, ORIGIN_GTT);
-		mutex_unlock(&fb_helper->dev->struct_mutex);
-	}
+	if (ret == 0)
+		intel_fbdev_invalidate(ifbdev);
 
 	return ret;
 }
@@ -71,12 +75,8 @@ static int intel_fbdev_blank(int blank, struct fb_info *info)
 	int ret;
 
 	ret = drm_fb_helper_blank(blank, info);
-
-	if (ret == 0) {
-		mutex_lock(&fb_helper->dev->struct_mutex);
-		intel_fb_obj_invalidate(ifbdev->fb->obj, ORIGIN_GTT);
-		mutex_unlock(&fb_helper->dev->struct_mutex);
-	}
+	if (ret == 0)
+		intel_fbdev_invalidate(ifbdev);
 
 	return ret;
 }
@@ -87,15 +87,11 @@ static int intel_fbdev_pan_display(struct fb_var_screeninfo *var,
 	struct drm_fb_helper *fb_helper = info->par;
 	struct intel_fbdev *ifbdev =
 		container_of(fb_helper, struct intel_fbdev, helper);
-
 	int ret;
-	ret = drm_fb_helper_pan_display(var, info);
 
-	if (ret == 0) {
-		mutex_lock(&fb_helper->dev->struct_mutex);
-		intel_fb_obj_invalidate(ifbdev->fb->obj, ORIGIN_GTT);
-		mutex_unlock(&fb_helper->dev->struct_mutex);
-	}
+	ret = drm_fb_helper_pan_display(var, info);
+	if (ret == 0)
+		intel_fbdev_invalidate(ifbdev);
 
 	return ret;
 }
@@ -121,7 +117,7 @@ static int intelfb_alloc(struct drm_fb_helper *helper,
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	struct drm_mode_fb_cmd2 mode_cmd = {};
-	struct drm_i915_gem_object *obj = NULL;
+	struct drm_i915_gem_object *obj;
 	int size, ret;
 
 	/* we don't do packed 24bpp */
@@ -136,14 +132,13 @@ static int intelfb_alloc(struct drm_fb_helper *helper,
 	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
 							  sizes->surface_depth);
 
-	mutex_lock(&dev->struct_mutex);
-
 	size = mode_cmd.pitches[0] * mode_cmd.height;
 	size = PAGE_ALIGN(size);
 
 	/* If the FB is too big, just don't use it since fbdev is not very
 	 * important and we should probably use that space with FBC or other
 	 * features. */
+	obj = NULL;
 	if (size * 2 < ggtt->stolen_usable_size)
 		obj = i915_gem_object_create_stolen(dev_priv, size);
 	if (obj == NULL)
@@ -151,24 +146,22 @@ static int intelfb_alloc(struct drm_fb_helper *helper,
 	if (IS_ERR(obj)) {
 		DRM_ERROR("failed to allocate framebuffer\n");
 		ret = PTR_ERR(obj);
-		goto out;
+		goto err;
 	}
 
-	fb = __intel_framebuffer_create(dev, &mode_cmd, obj);
+	fb = intel_framebuffer_create(obj, &mode_cmd);
 	if (IS_ERR(fb)) {
-		i915_gem_object_put(obj);
 		ret = PTR_ERR(fb);
-		goto out;
+		goto err_obj;
 	}
 
-	mutex_unlock(&dev->struct_mutex);
-
 	ifbdev->fb = to_intel_framebuffer(fb);
 
 	return 0;
 
-out:
-	mutex_unlock(&dev->struct_mutex);
+err_obj:
+	i915_gem_object_put(obj);
+err:
 	return ret;
 }
 
@@ -355,23 +348,23 @@ static bool intel_fb_initial_config(struct drm_fb_helper *fb_helper,
 				    bool *enabled, int width, int height)
 {
 	struct drm_i915_private *dev_priv = to_i915(fb_helper->dev);
-	unsigned long conn_configured, mask;
+	unsigned long conn_configured, conn_seq, mask;
 	unsigned int count = min(fb_helper->connector_count, BITS_PER_LONG);
 	int i, j;
 	bool *save_enabled;
 	bool fallback = true;
 	int num_connectors_enabled = 0;
 	int num_connectors_detected = 0;
-	int pass = 0;
 
 	save_enabled = kcalloc(count, sizeof(bool), GFP_KERNEL);
 	if (!save_enabled)
 		return false;
 
 	memcpy(save_enabled, enabled, count);
-	mask = BIT(count) - 1;
+	mask = GENMASK(count - 1, 0);
 	conn_configured = 0;
 retry:
+	conn_seq = conn_configured;
 	for (i = 0; i < count; i++) {
 		struct drm_fb_helper_connector *fb_conn;
 		struct drm_connector *connector;
@@ -385,7 +378,7 @@ retry:
 		if (conn_configured & BIT(i))
 			continue;
 
-		if (pass == 0 && !connector->has_tile)
+		if (conn_seq == 0 && !connector->has_tile)
 			continue;
 
 		if (connector->status == connector_status_connected)
@@ -496,10 +489,8 @@ retry:
 		conn_configured |= BIT(i);
 	}
 
-	if ((conn_configured & mask) != mask) {
-		pass++;
+	if ((conn_configured & mask) != mask && conn_configured != conn_seq)
 		goto retry;
-	}
 
 	/*
 	 * If the BIOS didn't enable everything it could, fall back to have the
@@ -628,7 +619,7 @@ static bool intel_fbdev_init_bios(struct drm_device *dev,
 		}
 
 		cur_size = intel_crtc->config->base.adjusted_mode.crtc_vdisplay;
-		cur_size = intel_fb_align_height(dev, cur_size,
+		cur_size = intel_fb_align_height(to_i915(dev), cur_size,
 						 fb->base.format->format,
 						 fb->base.modifier);
 		cur_size *= fb->base.pitches[0];
@@ -838,11 +829,6 @@ void intel_fbdev_restore_mode(struct drm_device *dev)
 	if (!ifbdev->fb)
 		return;
 
-	if (drm_fb_helper_restore_fbdev_mode_unlocked(&ifbdev->helper)) {
-		DRM_DEBUG("failed to restore crtc mode\n");
-	} else {
-		mutex_lock(&dev->struct_mutex);
-		intel_fb_obj_invalidate(ifbdev->fb->obj, ORIGIN_GTT);
-		mutex_unlock(&dev->struct_mutex);
-	}
+	if (drm_fb_helper_restore_fbdev_mode_unlocked(&ifbdev->helper) == 0)
+		intel_fbdev_invalidate(ifbdev);
 }
diff --git a/drivers/gpu/drm/i915/intel_fifo_underrun.c b/drivers/gpu/drm/i915/intel_fifo_underrun.c
index e660d8b4bbc3..966e255ca053 100644
--- a/drivers/gpu/drm/i915/intel_fifo_underrun.c
+++ b/drivers/gpu/drm/i915/intel_fifo_underrun.c
@@ -54,7 +54,7 @@ static bool ivb_can_enable_err_int(struct drm_device *dev)
 	struct intel_crtc *crtc;
 	enum pipe pipe;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	for_each_pipe(dev_priv, pipe) {
 		crtc = intel_get_crtc_for_pipe(dev_priv, pipe);
@@ -72,7 +72,7 @@ static bool cpt_can_enable_serr_int(struct drm_device *dev)
 	enum pipe pipe;
 	struct intel_crtc *crtc;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	for_each_pipe(dev_priv, pipe) {
 		crtc = intel_get_crtc_for_pipe(dev_priv, pipe);
@@ -90,7 +90,7 @@ static void i9xx_check_fifo_underruns(struct intel_crtc *crtc)
 	i915_reg_t reg = PIPESTAT(crtc->pipe);
 	u32 pipestat = I915_READ(reg) & 0xffff0000;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	if ((pipestat & PIPE_FIFO_UNDERRUN_STATUS) == 0)
 		return;
@@ -98,6 +98,7 @@ static void i9xx_check_fifo_underruns(struct intel_crtc *crtc)
 	I915_WRITE(reg, pipestat | PIPE_FIFO_UNDERRUN_STATUS);
 	POSTING_READ(reg);
 
+	trace_intel_cpu_fifo_underrun(dev_priv, crtc->pipe);
 	DRM_ERROR("pipe %c underrun\n", pipe_name(crtc->pipe));
 }
 
@@ -109,7 +110,7 @@ static void i9xx_set_fifo_underrun_reporting(struct drm_device *dev,
 	i915_reg_t reg = PIPESTAT(pipe);
 	u32 pipestat = I915_READ(reg) & 0xffff0000;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	if (enable) {
 		I915_WRITE(reg, pipestat | PIPE_FIFO_UNDERRUN_STATUS);
@@ -139,7 +140,7 @@ static void ivybridge_check_fifo_underruns(struct intel_crtc *crtc)
 	enum pipe pipe = crtc->pipe;
 	uint32_t err_int = I915_READ(GEN7_ERR_INT);
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	if ((err_int & ERR_INT_FIFO_UNDERRUN(pipe)) == 0)
 		return;
@@ -147,6 +148,7 @@ static void ivybridge_check_fifo_underruns(struct intel_crtc *crtc)
 	I915_WRITE(GEN7_ERR_INT, ERR_INT_FIFO_UNDERRUN(pipe));
 	POSTING_READ(GEN7_ERR_INT);
 
+	trace_intel_cpu_fifo_underrun(dev_priv, pipe);
 	DRM_ERROR("fifo underrun on pipe %c\n", pipe_name(pipe));
 }
 
@@ -204,7 +206,7 @@ static void cpt_check_pch_fifo_underruns(struct intel_crtc *crtc)
 	enum transcoder pch_transcoder = (enum transcoder) crtc->pipe;
 	uint32_t serr_int = I915_READ(SERR_INT);
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	if ((serr_int & SERR_INT_TRANS_FIFO_UNDERRUN(pch_transcoder)) == 0)
 		return;
@@ -212,6 +214,7 @@ static void cpt_check_pch_fifo_underruns(struct intel_crtc *crtc)
 	I915_WRITE(SERR_INT, SERR_INT_TRANS_FIFO_UNDERRUN(pch_transcoder));
 	POSTING_READ(SERR_INT);
 
+	trace_intel_pch_fifo_underrun(dev_priv, pch_transcoder);
 	DRM_ERROR("pch fifo underrun on pch transcoder %s\n",
 		  transcoder_name(pch_transcoder));
 }
@@ -248,7 +251,7 @@ static bool __intel_set_cpu_fifo_underrun_reporting(struct drm_device *dev,
 	struct intel_crtc *crtc = intel_get_crtc_for_pipe(dev_priv, pipe);
 	bool old;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	old = !crtc->cpu_fifo_underrun_disabled;
 	crtc->cpu_fifo_underrun_disabled = !enable;
@@ -368,9 +371,11 @@ void intel_cpu_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv,
 	    crtc->cpu_fifo_underrun_disabled)
 		return;
 
-	if (intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false))
+	if (intel_set_cpu_fifo_underrun_reporting(dev_priv, pipe, false)) {
+		trace_intel_cpu_fifo_underrun(dev_priv, pipe);
 		DRM_ERROR("CPU pipe %c FIFO underrun\n",
 			  pipe_name(pipe));
+	}
 
 	intel_fbc_handle_fifo_underrun_irq(dev_priv);
 }
@@ -388,9 +393,11 @@ void intel_pch_fifo_underrun_irq_handler(struct drm_i915_private *dev_priv,
 					 enum transcoder pch_transcoder)
 {
 	if (intel_set_pch_fifo_underrun_reporting(dev_priv, pch_transcoder,
-						  false))
+						  false)) {
+		trace_intel_pch_fifo_underrun(dev_priv, pch_transcoder);
 		DRM_ERROR("PCH transcoder %s FIFO underrun\n",
 			  transcoder_name(pch_transcoder));
+	}
 }
 
 /**
diff --git a/drivers/gpu/drm/i915/intel_frontbuffer.c b/drivers/gpu/drm/i915/intel_frontbuffer.c
index 966de4c7c7a2..fcfc217e754e 100644
--- a/drivers/gpu/drm/i915/intel_frontbuffer.c
+++ b/drivers/gpu/drm/i915/intel_frontbuffer.c
@@ -114,13 +114,12 @@ static void intel_frontbuffer_flush(struct drm_i915_private *dev_priv,
 }
 
 void __intel_fb_obj_flush(struct drm_i915_gem_object *obj,
-			  bool retire,
 			  enum fb_op_origin origin,
 			  unsigned int frontbuffer_bits)
 {
 	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 
-	if (retire) {
+	if (origin == ORIGIN_CS) {
 		spin_lock(&dev_priv->fb_tracking.lock);
 		/* Filter out new bits since rendering started. */
 		frontbuffer_bits &= dev_priv->fb_tracking.busy_bits;
diff --git a/drivers/gpu/drm/i915/intel_frontbuffer.h b/drivers/gpu/drm/i915/intel_frontbuffer.h
index 7bab41218cf7..63cd9a753a72 100644
--- a/drivers/gpu/drm/i915/intel_frontbuffer.h
+++ b/drivers/gpu/drm/i915/intel_frontbuffer.h
@@ -38,7 +38,6 @@ void __intel_fb_obj_invalidate(struct drm_i915_gem_object *obj,
 			       enum fb_op_origin origin,
 			       unsigned int frontbuffer_bits);
 void __intel_fb_obj_flush(struct drm_i915_gem_object *obj,
-			  bool retire,
 			  enum fb_op_origin origin,
 			  unsigned int frontbuffer_bits);
 
@@ -69,15 +68,12 @@ static inline bool intel_fb_obj_invalidate(struct drm_i915_gem_object *obj,
 /**
  * intel_fb_obj_flush - flush frontbuffer object
  * @obj: GEM object to flush
- * @retire: set when retiring asynchronous rendering
  * @origin: which operation caused the flush
  *
  * This function gets called every time rendering on the given object has
- * completed and frontbuffer caching can be started again. If @retire is true
- * then any delayed flushes will be unblocked.
+ * completed and frontbuffer caching can be started again.
  */
 static inline void intel_fb_obj_flush(struct drm_i915_gem_object *obj,
-				      bool retire,
 				      enum fb_op_origin origin)
 {
 	unsigned int frontbuffer_bits;
@@ -86,7 +82,7 @@ static inline void intel_fb_obj_flush(struct drm_i915_gem_object *obj,
 	if (!frontbuffer_bits)
 		return;
 
-	__intel_fb_obj_flush(obj, retire, origin, frontbuffer_bits);
+	__intel_fb_obj_flush(obj, origin, frontbuffer_bits);
 }
 
 #endif /* __INTEL_FRONTBUFFER_H__ */
diff --git a/drivers/gpu/drm/i915/intel_guc_loader.c b/drivers/gpu/drm/i915/intel_guc_loader.c
index 2f1cf9aea04e..9885f760f2ef 100644
--- a/drivers/gpu/drm/i915/intel_guc_loader.c
+++ b/drivers/gpu/drm/i915/intel_guc_loader.c
@@ -520,10 +520,6 @@ int intel_guc_setup(struct drm_i915_private *dev_priv)
 
 	guc_fw->load_status = INTEL_UC_FIRMWARE_SUCCESS;
 
-	DRM_DEBUG_DRIVER("GuC fw status: fetch %s, load %s\n",
-		intel_uc_fw_status_repr(guc_fw->fetch_status),
-		intel_uc_fw_status_repr(guc_fw->load_status));
-
 	intel_guc_auth_huc(dev_priv);
 
 	if (i915.enable_guc_submission) {
@@ -536,6 +532,11 @@ int intel_guc_setup(struct drm_i915_private *dev_priv)
 		guc_interrupts_capture(dev_priv);
 	}
 
+	DRM_INFO("GuC %s (firmware %s [version %u.%u])\n",
+		 i915.enable_guc_submission ? "submission enabled" : "loaded",
+		 guc_fw->path,
+		 guc_fw->major_ver_found, guc_fw->minor_ver_found);
+
 	return 0;
 
 fail:
@@ -713,12 +714,9 @@ fail:
 	DRM_DEBUG_DRIVER("uC fw fetch status FAIL; err %d, fw %p, obj %p\n",
 		err, fw, uc_fw->obj);
 
-	mutex_lock(&dev_priv->drm.struct_mutex);
-	obj = uc_fw->obj;
+	obj = fetch_and_zero(&uc_fw->obj);
 	if (obj)
 		i915_gem_object_put(obj);
-	uc_fw->obj = NULL;
-	mutex_unlock(&dev_priv->drm.struct_mutex);
 
 	release_firmware(fw);		/* OK even if fw is NULL */
 	uc_fw->fetch_status = INTEL_UC_FIRMWARE_FAIL;
@@ -792,16 +790,17 @@ void intel_guc_init(struct drm_i915_private *dev_priv)
 void intel_guc_fini(struct drm_i915_private *dev_priv)
 {
 	struct intel_uc_fw *guc_fw = &dev_priv->guc.fw;
+	struct drm_i915_gem_object *obj;
 
 	mutex_lock(&dev_priv->drm.struct_mutex);
 	guc_interrupts_release(dev_priv);
 	i915_guc_submission_disable(dev_priv);
 	i915_guc_submission_fini(dev_priv);
-
-	if (guc_fw->obj)
-		i915_gem_object_put(guc_fw->obj);
-	guc_fw->obj = NULL;
 	mutex_unlock(&dev_priv->drm.struct_mutex);
 
+	obj = fetch_and_zero(&guc_fw->obj);
+	if (obj)
+		i915_gem_object_put(obj);
+
 	guc_fw->fetch_status = INTEL_UC_FIRMWARE_NONE;
 }
diff --git a/drivers/gpu/drm/i915/intel_hangcheck.c b/drivers/gpu/drm/i915/intel_hangcheck.c
index f05971f5586f..dce742243ba6 100644
--- a/drivers/gpu/drm/i915/intel_hangcheck.c
+++ b/drivers/gpu/drm/i915/intel_hangcheck.c
@@ -480,3 +480,7 @@ void intel_hangcheck_init(struct drm_i915_private *i915)
 	INIT_DELAYED_WORK(&i915->gpu_error.hangcheck_work,
 			  i915_hangcheck_elapsed);
 }
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/intel_hangcheck.c"
+#endif
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index ebae2bd83918..c2184f755ec6 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -902,12 +902,11 @@ static bool intel_hdmi_get_hw_state(struct intel_encoder *encoder,
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(&encoder->base);
-	enum intel_display_power_domain power_domain;
 	u32 tmp;
 	bool ret;
 
-	power_domain = intel_display_port_power_domain(encoder);
-	if (!intel_display_power_get_if_enabled(dev_priv, power_domain))
+	if (!intel_display_power_get_if_enabled(dev_priv,
+						encoder->power_domain))
 		return false;
 
 	ret = false;
@@ -927,7 +926,7 @@ static bool intel_hdmi_get_hw_state(struct intel_encoder *encoder,
 	ret = true;
 
 out:
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, encoder->power_domain);
 
 	return ret;
 }
@@ -1869,14 +1868,7 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port,
 
 	switch (port) {
 	case PORT_B:
-		/*
-		 * On BXT A0/A1, sw needs to activate DDIA HPD logic and
-		 * interrupts to check the external panel connection.
-		 */
-		if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1))
-			intel_encoder->hpd_pin = HPD_PORT_A;
-		else
-			intel_encoder->hpd_pin = HPD_PORT_B;
+		intel_encoder->hpd_pin = HPD_PORT_B;
 		break;
 	case PORT_C:
 		intel_encoder->hpd_pin = HPD_PORT_C;
@@ -1988,6 +1980,7 @@ void intel_hdmi_init(struct drm_i915_private *dev_priv,
 	}
 
 	intel_encoder->type = INTEL_OUTPUT_HDMI;
+	intel_encoder->power_domain = intel_port_to_power_domain(port);
 	intel_encoder->port = port;
 	if (IS_CHERRYVIEW(dev_priv)) {
 		if (port == PORT_D)
diff --git a/drivers/gpu/drm/i915/intel_hotplug.c b/drivers/gpu/drm/i915/intel_hotplug.c
index b62e3f8ad415..ba763e7d7dcf 100644
--- a/drivers/gpu/drm/i915/intel_hotplug.c
+++ b/drivers/gpu/drm/i915/intel_hotplug.c
@@ -100,7 +100,6 @@ bool intel_hpd_pin_to_port(enum hpd_pin pin, enum port *port)
 }
 
 #define HPD_STORM_DETECT_PERIOD		1000
-#define HPD_STORM_THRESHOLD		5
 #define HPD_STORM_REENABLE_DELAY	(2 * 60 * 1000)
 
 /**
@@ -112,9 +111,13 @@ bool intel_hpd_pin_to_port(enum hpd_pin pin, enum port *port)
  * storms. Only the pin specific stats and state are changed, the caller is
  * responsible for further action.
  *
- * @HPD_STORM_THRESHOLD irqs are allowed within @HPD_STORM_DETECT_PERIOD ms,
- * otherwise it's considered an irq storm, and the irq state is set to
- * @HPD_MARK_DISABLED.
+ * The number of irqs that are allowed within @HPD_STORM_DETECT_PERIOD is
+ * stored in @dev_priv->hotplug.hpd_storm_threshold which defaults to
+ * @HPD_STORM_DEFAULT_THRESHOLD. If this threshold is exceeded, it's
+ * considered an irq storm and the irq state is set to @HPD_MARK_DISABLED.
+ *
+ * The HPD threshold can be controlled through i915_hpd_storm_ctl in debugfs,
+ * and should only be adjusted for automated hotplug testing.
  *
  * Return true if an irq storm was detected on @pin.
  */
@@ -123,13 +126,15 @@ static bool intel_hpd_irq_storm_detect(struct drm_i915_private *dev_priv,
 {
 	unsigned long start = dev_priv->hotplug.stats[pin].last_jiffies;
 	unsigned long end = start + msecs_to_jiffies(HPD_STORM_DETECT_PERIOD);
+	const int threshold = dev_priv->hotplug.hpd_storm_threshold;
 	bool storm = false;
 
 	if (!time_in_range(jiffies, start, end)) {
 		dev_priv->hotplug.stats[pin].last_jiffies = jiffies;
 		dev_priv->hotplug.stats[pin].count = 0;
 		DRM_DEBUG_KMS("Received HPD interrupt on PIN %d - cnt: 0\n", pin);
-	} else if (dev_priv->hotplug.stats[pin].count > HPD_STORM_THRESHOLD) {
+	} else if (dev_priv->hotplug.stats[pin].count > threshold &&
+		   threshold) {
 		dev_priv->hotplug.stats[pin].state = HPD_MARK_DISABLED;
 		DRM_DEBUG_KMS("HPD interrupt storm detected on PIN %d\n", pin);
 		storm = true;
@@ -152,7 +157,7 @@ static void intel_hpd_irq_storm_disable(struct drm_i915_private *dev_priv)
 	enum hpd_pin pin;
 	bool hpd_disabled = false;
 
-	assert_spin_locked(&dev_priv->irq_lock);
+	lockdep_assert_held(&dev_priv->irq_lock);
 
 	list_for_each_entry(connector, &mode_config->connector_list, head) {
 		if (connector->polled != DRM_CONNECTOR_POLL_HPD)
@@ -219,7 +224,7 @@ static void intel_hpd_irq_storm_reenable_work(struct work_struct *work)
 			}
 		}
 	}
-	if (dev_priv->display.hpd_irq_setup)
+	if (dev_priv->display_irqs_enabled && dev_priv->display.hpd_irq_setup)
 		dev_priv->display.hpd_irq_setup(dev_priv);
 	spin_unlock_irq(&dev_priv->irq_lock);
 
@@ -425,7 +430,7 @@ void intel_hpd_irq_handler(struct drm_i915_private *dev_priv,
 		}
 	}
 
-	if (storm_detected)
+	if (storm_detected && dev_priv->display_irqs_enabled)
 		dev_priv->display.hpd_irq_setup(dev_priv);
 	spin_unlock(&dev_priv->irq_lock);
 
@@ -471,10 +476,12 @@ void intel_hpd_init(struct drm_i915_private *dev_priv)
 	 * Interrupt setup is already guaranteed to be single-threaded, this is
 	 * just to make the assert_spin_locked checks happy.
 	 */
-	spin_lock_irq(&dev_priv->irq_lock);
-	if (dev_priv->display.hpd_irq_setup)
-		dev_priv->display.hpd_irq_setup(dev_priv);
-	spin_unlock_irq(&dev_priv->irq_lock);
+	if (dev_priv->display_irqs_enabled && dev_priv->display.hpd_irq_setup) {
+		spin_lock_irq(&dev_priv->irq_lock);
+		if (dev_priv->display_irqs_enabled)
+			dev_priv->display.hpd_irq_setup(dev_priv);
+		spin_unlock_irq(&dev_priv->irq_lock);
+	}
 }
 
 static void i915_hpd_poll_init_work(struct work_struct *work)
diff --git a/drivers/gpu/drm/i915/intel_huc.c b/drivers/gpu/drm/i915/intel_huc.c
index c144609425f6..e660109fc51e 100644
--- a/drivers/gpu/drm/i915/intel_huc.c
+++ b/drivers/gpu/drm/i915/intel_huc.c
@@ -181,12 +181,14 @@ void intel_huc_init(struct drm_i915_private *dev_priv)
 	}
 
 	huc_fw->path = fw_path;
+
+	if (huc_fw->path == NULL)
+		return;
+
 	huc_fw->fetch_status = INTEL_UC_FIRMWARE_PENDING;
 
 	DRM_DEBUG_DRIVER("HuC firmware pending, path %s\n", fw_path);
 
-	WARN(huc_fw->path == NULL, "HuC present but no fw path\n");
-
 	intel_uc_fw_fetch(dev_priv, huc_fw);
 }
 
@@ -274,12 +276,11 @@ fail:
 void intel_huc_fini(struct drm_i915_private *dev_priv)
 {
 	struct intel_uc_fw *huc_fw = &dev_priv->huc.fw;
+	struct drm_i915_gem_object *obj;
 
-	mutex_lock(&dev_priv->drm.struct_mutex);
-	if (huc_fw->obj)
-		i915_gem_object_put(huc_fw->obj);
-	huc_fw->obj = NULL;
-	mutex_unlock(&dev_priv->drm.struct_mutex);
+	obj = fetch_and_zero(&huc_fw->obj);
+	if (obj)
+		i915_gem_object_put(obj);
 
 	huc_fw->fetch_status = INTEL_UC_FIRMWARE_NONE;
 }
diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c
index bce1ba80f277..b6401e8f1bd6 100644
--- a/drivers/gpu/drm/i915/intel_i2c.c
+++ b/drivers/gpu/drm/i915/intel_i2c.c
@@ -74,7 +74,7 @@ static const struct gmbus_pin *get_gmbus_pin(struct drm_i915_private *dev_priv,
 {
 	if (IS_GEN9_LP(dev_priv))
 		return &gmbus_pins_bxt[pin];
-	else if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+	else if (IS_GEN9_BC(dev_priv))
 		return &gmbus_pins_skl[pin];
 	else if (IS_BROADWELL(dev_priv))
 		return &gmbus_pins_bdw[pin];
@@ -89,7 +89,7 @@ bool intel_gmbus_is_valid_pin(struct drm_i915_private *dev_priv,
 
 	if (IS_GEN9_LP(dev_priv))
 		size = ARRAY_SIZE(gmbus_pins_bxt);
-	else if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+	else if (IS_GEN9_BC(dev_priv))
 		size = ARRAY_SIZE(gmbus_pins_skl);
 	else if (IS_BROADWELL(dev_priv))
 		size = ARRAY_SIZE(gmbus_pins_bdw);
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index ebf8023d21e6..89f38e7def9f 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -190,13 +190,7 @@
 #define CTX_R_PWR_CLK_STATE		0x42
 #define CTX_GPGPU_CSR_BASE_ADDRESS	0x44
 
-#define GEN8_CTX_VALID (1<<0)
-#define GEN8_CTX_FORCE_PD_RESTORE (1<<1)
-#define GEN8_CTX_FORCE_RESTORE (1<<2)
-#define GEN8_CTX_L3LLC_COHERENT (1<<5)
-#define GEN8_CTX_PRIVILEGE (1<<8)
-
-#define ASSIGN_CTX_REG(reg_state, pos, reg, val) do { \
+#define CTX_REG(reg_state, pos, reg, val) do { \
 	(reg_state)[(pos)+0] = i915_mmio_reg_offset(reg); \
 	(reg_state)[(pos)+1] = (val); \
 } while (0)
@@ -212,14 +206,6 @@
 	reg_state[CTX_PDP0_LDW + 1] = lower_32_bits(px_dma(&ppgtt->pml4)); \
 } while (0)
 
-enum {
-	FAULT_AND_HANG = 0,
-	FAULT_AND_HALT, /* Debug only */
-	FAULT_AND_STREAM,
-	FAULT_AND_CONTINUE /* Unsupported */
-};
-#define GEN8_CTX_ID_SHIFT 32
-#define GEN8_CTX_ID_WIDTH 21
 #define GEN8_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT	0x17
 #define GEN9_CTX_RCS_INDIRECT_CTX_OFFSET_DEFAULT	0x26
 
@@ -267,30 +253,6 @@ int intel_sanitize_enable_execlists(struct drm_i915_private *dev_priv, int enabl
 	return 0;
 }
 
-static void
-logical_ring_init_platform_invariants(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-
-	engine->disable_lite_restore_wa =
-		IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1) &&
-		(engine->id == VCS || engine->id == VCS2);
-
-	engine->ctx_desc_template = GEN8_CTX_VALID;
-	if (IS_GEN8(dev_priv))
-		engine->ctx_desc_template |= GEN8_CTX_L3LLC_COHERENT;
-	engine->ctx_desc_template |= GEN8_CTX_PRIVILEGE;
-
-	/* TODO: WaDisableLiteRestore when we start using semaphore
-	 * signalling between Command Streamers */
-	/* ring->ctx_desc_template |= GEN8_CTX_FORCE_RESTORE; */
-
-	/* WaEnableForceRestoreInCtxtDescForVCS:skl */
-	/* WaEnableForceRestoreInCtxtDescForVCS:bxt */
-	if (engine->disable_lite_restore_wa)
-		engine->ctx_desc_template |= GEN8_CTX_FORCE_RESTORE;
-}
-
 /**
  * intel_lr_context_descriptor_update() - calculate & cache the descriptor
  * 					  descriptor for a pinned context
@@ -304,7 +266,7 @@ logical_ring_init_platform_invariants(struct intel_engine_cs *engine)
  *
  * This is what a descriptor looks like, from LSB to MSB::
  *
- *      bits  0-11:    flags, GEN8_CTX_* (cached in ctx_desc_template)
+ *      bits  0-11:    flags, GEN8_CTX_* (cached in ctx->desc_template)
  *      bits 12-31:    LRCA, GTT address of (the HWSP of) this context
  *      bits 32-52:    ctx ID, a globally unique tag
  *      bits 53-54:    mbz, reserved for use by hardware
@@ -319,8 +281,7 @@ intel_lr_context_descriptor_update(struct i915_gem_context *ctx,
 
 	BUILD_BUG_ON(MAX_CONTEXT_HW_ID > (1<<GEN8_CTX_ID_WIDTH));
 
-	desc = ctx->desc_template;				/* bits  3-4  */
-	desc |= engine->ctx_desc_template;			/* bits  0-11 */
+	desc = ctx->desc_template;				/* bits  0-11 */
 	desc |= i915_ggtt_offset(ce->state) + LRC_PPHWSP_PN * PAGE_SIZE;
 								/* bits 12-31 */
 	desc |= (u64)ctx->hw_id << GEN8_CTX_ID_SHIFT;		/* bits 32-52 */
@@ -364,6 +325,7 @@ static u64 execlists_update_context(struct drm_i915_gem_request *rq)
 		rq->ctx->ppgtt ?: rq->i915->mm.aliasing_ppgtt;
 	u32 *reg_state = ce->lrc_reg_state;
 
+	GEM_BUG_ON(!IS_ALIGNED(rq->tail, 8));
 	reg_state[CTX_RING_TAIL+1] = rq->tail;
 
 	/* True 32b PPGTT with dynamic page allocation: update PDP
@@ -371,7 +333,7 @@ static u64 execlists_update_context(struct drm_i915_gem_request *rq)
 	 * PML4 is allocated during ppgtt init, so this is not needed
 	 * in 48-bit mode.
 	 */
-	if (ppgtt && !USES_FULL_48BIT_PPGTT(ppgtt->base.dev))
+	if (ppgtt && !i915_vm_is_48bit(&ppgtt->base))
 		execlists_update_context_pdps(ppgtt, reg_state);
 
 	return ce->lrc_desc;
@@ -385,17 +347,20 @@ static void execlists_submit_ports(struct intel_engine_cs *engine)
 		dev_priv->regs + i915_mmio_reg_offset(RING_ELSP(engine));
 	u64 desc[2];
 
+	GEM_BUG_ON(port[0].count > 1);
 	if (!port[0].count)
 		execlists_context_status_change(port[0].request,
 						INTEL_CONTEXT_SCHEDULE_IN);
 	desc[0] = execlists_update_context(port[0].request);
-	engine->preempt_wa = port[0].count++; /* bdw only? fixed on skl? */
+	GEM_DEBUG_EXEC(port[0].context_id = upper_32_bits(desc[0]));
+	port[0].count++;
 
 	if (port[1].request) {
 		GEM_BUG_ON(port[1].count);
 		execlists_context_status_change(port[1].request,
 						INTEL_CONTEXT_SCHEDULE_IN);
 		desc[1] = execlists_update_context(port[1].request);
+		GEM_DEBUG_EXEC(port[1].context_id = upper_32_bits(desc[1]));
 		port[1].count = 1;
 	} else {
 		desc[1] = 0;
@@ -514,6 +479,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 		cursor->priotree.priority = INT_MAX;
 
 		__i915_gem_request_submit(cursor);
+		trace_i915_gem_request_in(cursor, port - engine->execlist_port);
 		last = cursor;
 		submit = true;
 	}
@@ -532,37 +498,11 @@ static bool execlists_elsp_idle(struct intel_engine_cs *engine)
 	return !engine->execlist_port[0].request;
 }
 
-/**
- * intel_execlists_idle() - Determine if all engine submission ports are idle
- * @dev_priv: i915 device private
- *
- * Return true if there are no requests pending on any of the submission ports
- * of any engines.
- */
-bool intel_execlists_idle(struct drm_i915_private *dev_priv)
-{
-	struct intel_engine_cs *engine;
-	enum intel_engine_id id;
-
-	if (!i915.enable_execlists)
-		return true;
-
-	for_each_engine(engine, dev_priv, id)
-		if (!execlists_elsp_idle(engine))
-			return false;
-
-	return true;
-}
-
-static bool execlists_elsp_ready(struct intel_engine_cs *engine)
+static bool execlists_elsp_ready(const struct intel_engine_cs *engine)
 {
-	int port;
-
-	port = 1; /* wait for a free slot */
-	if (engine->disable_lite_restore_wa || engine->preempt_wa)
-		port = 0; /* wait for GPU to be idle before continuing */
+	const struct execlist_port *port = engine->execlist_port;
 
-	return !engine->execlist_port[port].request;
+	return port[0].count + port[1].count < 2;
 }
 
 /*
@@ -577,7 +517,7 @@ static void intel_lrc_irq_handler(unsigned long data)
 
 	intel_uncore_forcewake_get(dev_priv, engine->fw_domains);
 
-	if (!execlists_elsp_idle(engine)) {
+	while (test_and_clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted)) {
 		u32 __iomem *csb_mmio =
 			dev_priv->regs + i915_mmio_reg_offset(RING_CONTEXT_STATUS_PTR(engine));
 		u32 __iomem *buf =
@@ -587,31 +527,55 @@ static void intel_lrc_irq_handler(unsigned long data)
 		csb = readl(csb_mmio);
 		head = GEN8_CSB_READ_PTR(csb);
 		tail = GEN8_CSB_WRITE_PTR(csb);
+		if (head == tail)
+			break;
+
 		if (tail < head)
 			tail += GEN8_CSB_ENTRIES;
-		while (head < tail) {
+		do {
 			unsigned int idx = ++head % GEN8_CSB_ENTRIES;
 			unsigned int status = readl(buf + 2 * idx);
 
+			/* We are flying near dragons again.
+			 *
+			 * We hold a reference to the request in execlist_port[]
+			 * but no more than that. We are operating in softirq
+			 * context and so cannot hold any mutex or sleep. That
+			 * prevents us stopping the requests we are processing
+			 * in port[] from being retired simultaneously (the
+			 * breadcrumb will be complete before we see the
+			 * context-switch). As we only hold the reference to the
+			 * request, any pointer chasing underneath the request
+			 * is subject to a potential use-after-free. Thus we
+			 * store all of the bookkeeping within port[] as
+			 * required, and avoid using unguarded pointers beneath
+			 * request itself. The same applies to the atomic
+			 * status notifier.
+			 */
+
 			if (!(status & GEN8_CTX_STATUS_COMPLETED_MASK))
 				continue;
 
+			/* Check the context/desc id for this event matches */
+			GEM_DEBUG_BUG_ON(readl(buf + 2 * idx + 1) !=
+					 port[0].context_id);
+
 			GEM_BUG_ON(port[0].count == 0);
 			if (--port[0].count == 0) {
 				GEM_BUG_ON(status & GEN8_CTX_STATUS_PREEMPTED);
+				GEM_BUG_ON(!i915_gem_request_completed(port[0].request));
 				execlists_context_status_change(port[0].request,
 								INTEL_CONTEXT_SCHEDULE_OUT);
 
+				trace_i915_gem_request_out(port[0].request);
 				i915_gem_request_put(port[0].request);
 				port[0] = port[1];
 				memset(&port[1], 0, sizeof(port[1]));
-
-				engine->preempt_wa = false;
 			}
 
 			GEM_BUG_ON(port[0].count == 0 &&
 				   !(status & GEN8_CTX_STATUS_ACTIVE_IDLE));
-		}
+		} while (head < tail);
 
 		writel(_MASKED_FIELD(GEN8_CSB_READ_PTR_MASK,
 				     GEN8_CSB_WRITE_PTR(csb) << 8),
@@ -658,10 +622,11 @@ static void execlists_submit_request(struct drm_i915_gem_request *request)
 	/* Will be called from irq-context when using foreign fences. */
 	spin_lock_irqsave(&engine->timeline->lock, flags);
 
-	if (insert_request(&request->priotree, &engine->execlist_queue))
+	if (insert_request(&request->priotree, &engine->execlist_queue)) {
 		engine->execlist_first = &request->priotree.node;
-	if (execlists_elsp_idle(engine))
-		tasklet_hi_schedule(&engine->irq_tasklet);
+		if (execlists_elsp_ready(engine))
+			tasklet_hi_schedule(&engine->irq_tasklet);
+	}
 
 	spin_unlock_irqrestore(&engine->timeline->lock, flags);
 }
@@ -784,11 +749,9 @@ static int execlists_context_pin(struct intel_engine_cs *engine,
 	}
 	GEM_BUG_ON(!ce->state);
 
-	flags = PIN_GLOBAL;
+	flags = PIN_GLOBAL | PIN_HIGH;
 	if (ctx->ggtt_offset_bias)
 		flags |= PIN_OFFSET_BIAS | ctx->ggtt_offset_bias;
-	if (i915_gem_context_is_kernel(ctx))
-		flags |= PIN_HIGH;
 
 	ret = i915_vma_pin(ce->state, 0, GEN8_LR_CONTEXT_ALIGN, flags);
 	if (ret)
@@ -847,6 +810,7 @@ static int execlists_request_alloc(struct drm_i915_gem_request *request)
 {
 	struct intel_engine_cs *engine = request->engine;
 	struct intel_context *ce = &request->ctx->engine[engine->id];
+	u32 *cs;
 	int ret;
 
 	GEM_BUG_ON(!ce->pin_count);
@@ -871,9 +835,11 @@ static int execlists_request_alloc(struct drm_i915_gem_request *request)
 			goto err;
 	}
 
-	ret = intel_ring_begin(request, 0);
-	if (ret)
+	cs = intel_ring_begin(request, 0);
+	if (IS_ERR(cs)) {
+		ret = PTR_ERR(cs);
 		goto err_unreserve;
+	}
 
 	if (!ce->initialised) {
 		ret = engine->init_context(request);
@@ -900,51 +866,6 @@ err:
 	return ret;
 }
 
-static int intel_logical_ring_workarounds_emit(struct drm_i915_gem_request *req)
-{
-	int ret, i;
-	struct intel_ring *ring = req->ring;
-	struct i915_workarounds *w = &req->i915->workarounds;
-
-	if (w->count == 0)
-		return 0;
-
-	ret = req->engine->emit_flush(req, EMIT_BARRIER);
-	if (ret)
-		return ret;
-
-	ret = intel_ring_begin(req, w->count * 2 + 2);
-	if (ret)
-		return ret;
-
-	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(w->count));
-	for (i = 0; i < w->count; i++) {
-		intel_ring_emit_reg(ring, w->reg[i].addr);
-		intel_ring_emit(ring, w->reg[i].value);
-	}
-	intel_ring_emit(ring, MI_NOOP);
-
-	intel_ring_advance(ring);
-
-	ret = req->engine->emit_flush(req, EMIT_BARRIER);
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
-#define wa_ctx_emit(batch, index, cmd)					\
-	do {								\
-		int __index = (index)++;				\
-		if (WARN_ON(__index >= (PAGE_SIZE / sizeof(uint32_t)))) { \
-			return -ENOSPC;					\
-		}							\
-		batch[__index] = (cmd);					\
-	} while (0)
-
-#define wa_ctx_emit_reg(batch, index, reg) \
-	wa_ctx_emit((batch), (index), i915_mmio_reg_offset(reg))
-
 /*
  * In this WA we need to set GEN8_L3SQCREG4[21:21] and reset it after
  * PIPE_CONTROL instruction. This is required for the flush to happen correctly
@@ -961,56 +882,29 @@ static int intel_logical_ring_workarounds_emit(struct drm_i915_gem_request *req)
  * This WA is also required for Gen9 so extracting as a function avoids
  * code duplication.
  */
-static inline int gen8_emit_flush_coherentl3_wa(struct intel_engine_cs *engine,
-						uint32_t *batch,
-						uint32_t index)
+static u32 *
+gen8_emit_flush_coherentl3_wa(struct intel_engine_cs *engine, u32 *batch)
 {
-	uint32_t l3sqc4_flush = (0x40400000 | GEN8_LQSC_FLUSH_COHERENT_LINES);
-
-	wa_ctx_emit(batch, index, (MI_STORE_REGISTER_MEM_GEN8 |
-				   MI_SRM_LRM_GLOBAL_GTT));
-	wa_ctx_emit_reg(batch, index, GEN8_L3SQCREG4);
-	wa_ctx_emit(batch, index, i915_ggtt_offset(engine->scratch) + 256);
-	wa_ctx_emit(batch, index, 0);
-
-	wa_ctx_emit(batch, index, MI_LOAD_REGISTER_IMM(1));
-	wa_ctx_emit_reg(batch, index, GEN8_L3SQCREG4);
-	wa_ctx_emit(batch, index, l3sqc4_flush);
-
-	wa_ctx_emit(batch, index, GFX_OP_PIPE_CONTROL(6));
-	wa_ctx_emit(batch, index, (PIPE_CONTROL_CS_STALL |
-				   PIPE_CONTROL_DC_FLUSH_ENABLE));
-	wa_ctx_emit(batch, index, 0);
-	wa_ctx_emit(batch, index, 0);
-	wa_ctx_emit(batch, index, 0);
-	wa_ctx_emit(batch, index, 0);
-
-	wa_ctx_emit(batch, index, (MI_LOAD_REGISTER_MEM_GEN8 |
-				   MI_SRM_LRM_GLOBAL_GTT));
-	wa_ctx_emit_reg(batch, index, GEN8_L3SQCREG4);
-	wa_ctx_emit(batch, index, i915_ggtt_offset(engine->scratch) + 256);
-	wa_ctx_emit(batch, index, 0);
-
-	return index;
-}
+	*batch++ = MI_STORE_REGISTER_MEM_GEN8 | MI_SRM_LRM_GLOBAL_GTT;
+	*batch++ = i915_mmio_reg_offset(GEN8_L3SQCREG4);
+	*batch++ = i915_ggtt_offset(engine->scratch) + 256;
+	*batch++ = 0;
 
-static inline uint32_t wa_ctx_start(struct i915_wa_ctx_bb *wa_ctx,
-				    uint32_t offset,
-				    uint32_t start_alignment)
-{
-	return wa_ctx->offset = ALIGN(offset, start_alignment);
-}
+	*batch++ = MI_LOAD_REGISTER_IMM(1);
+	*batch++ = i915_mmio_reg_offset(GEN8_L3SQCREG4);
+	*batch++ = 0x40400000 | GEN8_LQSC_FLUSH_COHERENT_LINES;
 
-static inline int wa_ctx_end(struct i915_wa_ctx_bb *wa_ctx,
-			     uint32_t offset,
-			     uint32_t size_alignment)
-{
-	wa_ctx->size = offset - wa_ctx->offset;
+	batch = gen8_emit_pipe_control(batch,
+				       PIPE_CONTROL_CS_STALL |
+				       PIPE_CONTROL_DC_FLUSH_ENABLE,
+				       0);
 
-	WARN(wa_ctx->size % size_alignment,
-	     "wa_ctx_bb failed sanity checks: size %d is not aligned to %d\n",
-	     wa_ctx->size, size_alignment);
-	return 0;
+	*batch++ = MI_LOAD_REGISTER_MEM_GEN8 | MI_SRM_LRM_GLOBAL_GTT;
+	*batch++ = i915_mmio_reg_offset(GEN8_L3SQCREG4);
+	*batch++ = i915_ggtt_offset(engine->scratch) + 256;
+	*batch++ = 0;
+
+	return batch;
 }
 
 /*
@@ -1028,42 +922,28 @@ static inline int wa_ctx_end(struct i915_wa_ctx_bb *wa_ctx,
  * MI_BATCH_BUFFER_END will be added to perctx batch and both of them together
  * makes a complete batch buffer.
  */
-static int gen8_init_indirectctx_bb(struct intel_engine_cs *engine,
-				    struct i915_wa_ctx_bb *wa_ctx,
-				    uint32_t *batch,
-				    uint32_t *offset)
+static u32 *gen8_init_indirectctx_bb(struct intel_engine_cs *engine, u32 *batch)
 {
-	uint32_t scratch_addr;
-	uint32_t index = wa_ctx_start(wa_ctx, *offset, CACHELINE_DWORDS);
-
 	/* WaDisableCtxRestoreArbitration:bdw,chv */
-	wa_ctx_emit(batch, index, MI_ARB_ON_OFF | MI_ARB_DISABLE);
+	*batch++ = MI_ARB_ON_OFF | MI_ARB_DISABLE;
 
 	/* WaFlushCoherentL3CacheLinesAtContextSwitch:bdw */
-	if (IS_BROADWELL(engine->i915)) {
-		int rc = gen8_emit_flush_coherentl3_wa(engine, batch, index);
-		if (rc < 0)
-			return rc;
-		index = rc;
-	}
+	if (IS_BROADWELL(engine->i915))
+		batch = gen8_emit_flush_coherentl3_wa(engine, batch);
 
 	/* WaClearSlmSpaceAtContextSwitch:bdw,chv */
 	/* Actual scratch location is at 128 bytes offset */
-	scratch_addr = i915_ggtt_offset(engine->scratch) + 2 * CACHELINE_BYTES;
-
-	wa_ctx_emit(batch, index, GFX_OP_PIPE_CONTROL(6));
-	wa_ctx_emit(batch, index, (PIPE_CONTROL_FLUSH_L3 |
-				   PIPE_CONTROL_GLOBAL_GTT_IVB |
-				   PIPE_CONTROL_CS_STALL |
-				   PIPE_CONTROL_QW_WRITE));
-	wa_ctx_emit(batch, index, scratch_addr);
-	wa_ctx_emit(batch, index, 0);
-	wa_ctx_emit(batch, index, 0);
-	wa_ctx_emit(batch, index, 0);
+	batch = gen8_emit_pipe_control(batch,
+				       PIPE_CONTROL_FLUSH_L3 |
+				       PIPE_CONTROL_GLOBAL_GTT_IVB |
+				       PIPE_CONTROL_CS_STALL |
+				       PIPE_CONTROL_QW_WRITE,
+				       i915_ggtt_offset(engine->scratch) +
+				       2 * CACHELINE_BYTES);
 
 	/* Pad to end of cacheline */
-	while (index % CACHELINE_DWORDS)
-		wa_ctx_emit(batch, index, MI_NOOP);
+	while ((unsigned long)batch % CACHELINE_BYTES)
+		*batch++ = MI_NOOP;
 
 	/*
 	 * MI_BATCH_BUFFER_END is not required in Indirect ctx BB because
@@ -1071,7 +951,7 @@ static int gen8_init_indirectctx_bb(struct intel_engine_cs *engine,
 	 * in the register CTX_RCS_INDIRECT_CTX
 	 */
 
-	return wa_ctx_end(wa_ctx, *offset = index, CACHELINE_DWORDS);
+	return batch;
 }
 
 /*
@@ -1083,65 +963,40 @@ static int gen8_init_indirectctx_bb(struct intel_engine_cs *engine,
  *  This batch is terminated with MI_BATCH_BUFFER_END and so we need not add padding
  *  to align it with cacheline as padding after MI_BATCH_BUFFER_END is redundant.
  */
-static int gen8_init_perctx_bb(struct intel_engine_cs *engine,
-			       struct i915_wa_ctx_bb *wa_ctx,
-			       uint32_t *batch,
-			       uint32_t *offset)
+static u32 *gen8_init_perctx_bb(struct intel_engine_cs *engine, u32 *batch)
 {
-	uint32_t index = wa_ctx_start(wa_ctx, *offset, CACHELINE_DWORDS);
-
 	/* WaDisableCtxRestoreArbitration:bdw,chv */
-	wa_ctx_emit(batch, index, MI_ARB_ON_OFF | MI_ARB_ENABLE);
-
-	wa_ctx_emit(batch, index, MI_BATCH_BUFFER_END);
+	*batch++ = MI_ARB_ON_OFF | MI_ARB_ENABLE;
+	*batch++ = MI_BATCH_BUFFER_END;
 
-	return wa_ctx_end(wa_ctx, *offset = index, 1);
+	return batch;
 }
 
-static int gen9_init_indirectctx_bb(struct intel_engine_cs *engine,
-				    struct i915_wa_ctx_bb *wa_ctx,
-				    uint32_t *batch,
-				    uint32_t *offset)
+static u32 *gen9_init_indirectctx_bb(struct intel_engine_cs *engine, u32 *batch)
 {
-	int ret;
-	struct drm_i915_private *dev_priv = engine->i915;
-	uint32_t index = wa_ctx_start(wa_ctx, *offset, CACHELINE_DWORDS);
+	/* WaFlushCoherentL3CacheLinesAtContextSwitch:skl,bxt,glk */
+	batch = gen8_emit_flush_coherentl3_wa(engine, batch);
 
-	/* WaDisableCtxRestoreArbitration:bxt */
-	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1))
-		wa_ctx_emit(batch, index, MI_ARB_ON_OFF | MI_ARB_DISABLE);
-
-	/* WaFlushCoherentL3CacheLinesAtContextSwitch:skl,bxt */
-	ret = gen8_emit_flush_coherentl3_wa(engine, batch, index);
-	if (ret < 0)
-		return ret;
-	index = ret;
-
-	/* WaDisableGatherAtSetShaderCommonSlice:skl,bxt,kbl */
-	wa_ctx_emit(batch, index, MI_LOAD_REGISTER_IMM(1));
-	wa_ctx_emit_reg(batch, index, COMMON_SLICE_CHICKEN2);
-	wa_ctx_emit(batch, index, _MASKED_BIT_DISABLE(
-			    GEN9_DISABLE_GATHER_AT_SET_SHADER_COMMON_SLICE));
-	wa_ctx_emit(batch, index, MI_NOOP);
+	/* WaDisableGatherAtSetShaderCommonSlice:skl,bxt,kbl,glk */
+	*batch++ = MI_LOAD_REGISTER_IMM(1);
+	*batch++ = i915_mmio_reg_offset(COMMON_SLICE_CHICKEN2);
+	*batch++ = _MASKED_BIT_DISABLE(
+			GEN9_DISABLE_GATHER_AT_SET_SHADER_COMMON_SLICE);
+	*batch++ = MI_NOOP;
 
 	/* WaClearSlmSpaceAtContextSwitch:kbl */
 	/* Actual scratch location is at 128 bytes offset */
-	if (IS_KBL_REVID(dev_priv, 0, KBL_REVID_A0)) {
-		u32 scratch_addr =
-			i915_ggtt_offset(engine->scratch) + 2 * CACHELINE_BYTES;
-
-		wa_ctx_emit(batch, index, GFX_OP_PIPE_CONTROL(6));
-		wa_ctx_emit(batch, index, (PIPE_CONTROL_FLUSH_L3 |
-					   PIPE_CONTROL_GLOBAL_GTT_IVB |
-					   PIPE_CONTROL_CS_STALL |
-					   PIPE_CONTROL_QW_WRITE));
-		wa_ctx_emit(batch, index, scratch_addr);
-		wa_ctx_emit(batch, index, 0);
-		wa_ctx_emit(batch, index, 0);
-		wa_ctx_emit(batch, index, 0);
+	if (IS_KBL_REVID(engine->i915, 0, KBL_REVID_A0)) {
+		batch = gen8_emit_pipe_control(batch,
+					       PIPE_CONTROL_FLUSH_L3 |
+					       PIPE_CONTROL_GLOBAL_GTT_IVB |
+					       PIPE_CONTROL_CS_STALL |
+					       PIPE_CONTROL_QW_WRITE,
+					       i915_ggtt_offset(engine->scratch)
+					       + 2 * CACHELINE_BYTES);
 	}
 
-	/* WaMediaPoolStateCmdInWABB:bxt */
+	/* WaMediaPoolStateCmdInWABB:bxt,glk */
 	if (HAS_POOLED_EU(engine->i915)) {
 		/*
 		 * EU pool configuration is setup along with golden context
@@ -1156,73 +1011,37 @@ static int gen9_init_indirectctx_bb(struct intel_engine_cs *engine,
 		 * possible configurations, to avoid duplication they are
 		 * not shown here again.
 		 */
-		u32 eu_pool_config = 0x00777000;
-		wa_ctx_emit(batch, index, GEN9_MEDIA_POOL_STATE);
-		wa_ctx_emit(batch, index, GEN9_MEDIA_POOL_ENABLE);
-		wa_ctx_emit(batch, index, eu_pool_config);
-		wa_ctx_emit(batch, index, 0);
-		wa_ctx_emit(batch, index, 0);
-		wa_ctx_emit(batch, index, 0);
+		*batch++ = GEN9_MEDIA_POOL_STATE;
+		*batch++ = GEN9_MEDIA_POOL_ENABLE;
+		*batch++ = 0x00777000;
+		*batch++ = 0;
+		*batch++ = 0;
+		*batch++ = 0;
 	}
 
 	/* Pad to end of cacheline */
-	while (index % CACHELINE_DWORDS)
-		wa_ctx_emit(batch, index, MI_NOOP);
+	while ((unsigned long)batch % CACHELINE_BYTES)
+		*batch++ = MI_NOOP;
 
-	return wa_ctx_end(wa_ctx, *offset = index, CACHELINE_DWORDS);
+	return batch;
 }
 
-static int gen9_init_perctx_bb(struct intel_engine_cs *engine,
-			       struct i915_wa_ctx_bb *wa_ctx,
-			       uint32_t *batch,
-			       uint32_t *offset)
+static u32 *gen9_init_perctx_bb(struct intel_engine_cs *engine, u32 *batch)
 {
-	uint32_t index = wa_ctx_start(wa_ctx, *offset, CACHELINE_DWORDS);
-
-	/* WaSetDisablePixMaskCammingAndRhwoInCommonSliceChicken:bxt */
-	if (IS_BXT_REVID(engine->i915, 0, BXT_REVID_A1)) {
-		wa_ctx_emit(batch, index, MI_LOAD_REGISTER_IMM(1));
-		wa_ctx_emit_reg(batch, index, GEN9_SLICE_COMMON_ECO_CHICKEN0);
-		wa_ctx_emit(batch, index,
-			    _MASKED_BIT_ENABLE(DISABLE_PIXEL_MASK_CAMMING));
-		wa_ctx_emit(batch, index, MI_NOOP);
-	}
-
-	/* WaClearTdlStateAckDirtyBits:bxt */
-	if (IS_BXT_REVID(engine->i915, 0, BXT_REVID_B0)) {
-		wa_ctx_emit(batch, index, MI_LOAD_REGISTER_IMM(4));
+	*batch++ = MI_BATCH_BUFFER_END;
 
-		wa_ctx_emit_reg(batch, index, GEN8_STATE_ACK);
-		wa_ctx_emit(batch, index, _MASKED_BIT_DISABLE(GEN9_SUBSLICE_TDL_ACK_BITS));
-
-		wa_ctx_emit_reg(batch, index, GEN9_STATE_ACK_SLICE1);
-		wa_ctx_emit(batch, index, _MASKED_BIT_DISABLE(GEN9_SUBSLICE_TDL_ACK_BITS));
-
-		wa_ctx_emit_reg(batch, index, GEN9_STATE_ACK_SLICE2);
-		wa_ctx_emit(batch, index, _MASKED_BIT_DISABLE(GEN9_SUBSLICE_TDL_ACK_BITS));
-
-		wa_ctx_emit_reg(batch, index, GEN7_ROW_CHICKEN2);
-		/* dummy write to CS, mask bits are 0 to ensure the register is not modified */
-		wa_ctx_emit(batch, index, 0x0);
-		wa_ctx_emit(batch, index, MI_NOOP);
-	}
-
-	/* WaDisableCtxRestoreArbitration:bxt */
-	if (IS_BXT_REVID(engine->i915, 0, BXT_REVID_A1))
-		wa_ctx_emit(batch, index, MI_ARB_ON_OFF | MI_ARB_ENABLE);
-
-	wa_ctx_emit(batch, index, MI_BATCH_BUFFER_END);
-
-	return wa_ctx_end(wa_ctx, *offset = index, 1);
+	return batch;
 }
 
-static int lrc_setup_wa_ctx_obj(struct intel_engine_cs *engine, u32 size)
+#define CTX_WA_BB_OBJ_SIZE (PAGE_SIZE)
+
+static int lrc_setup_wa_ctx(struct intel_engine_cs *engine)
 {
 	struct drm_i915_gem_object *obj;
 	struct i915_vma *vma;
 	int err;
 
-	obj = i915_gem_object_create(engine->i915, PAGE_ALIGN(size));
+	obj = i915_gem_object_create(engine->i915, CTX_WA_BB_OBJ_SIZE);
 	if (IS_ERR(obj))
 		return PTR_ERR(obj);
 
@@ -1244,82 +1063,79 @@ err:
 	return err;
 }
 
-static void lrc_destroy_wa_ctx_obj(struct intel_engine_cs *engine)
+static void lrc_destroy_wa_ctx(struct intel_engine_cs *engine)
 {
 	i915_vma_unpin_and_release(&engine->wa_ctx.vma);
 }
 
+typedef u32 *(*wa_bb_func_t)(struct intel_engine_cs *engine, u32 *batch);
+
 static int intel_init_workaround_bb(struct intel_engine_cs *engine)
 {
 	struct i915_ctx_workarounds *wa_ctx = &engine->wa_ctx;
-	uint32_t *batch;
-	uint32_t offset;
+	struct i915_wa_ctx_bb *wa_bb[2] = { &wa_ctx->indirect_ctx,
+					    &wa_ctx->per_ctx };
+	wa_bb_func_t wa_bb_fn[2];
 	struct page *page;
+	void *batch, *batch_ptr;
+	unsigned int i;
 	int ret;
 
-	WARN_ON(engine->id != RCS);
+	if (WARN_ON(engine->id != RCS || !engine->scratch))
+		return -EINVAL;
 
-	/* update this when WA for higher Gen are added */
-	if (INTEL_GEN(engine->i915) > 9) {
-		DRM_ERROR("WA batch buffer is not initialized for Gen%d\n",
-			  INTEL_GEN(engine->i915));
+	switch (INTEL_GEN(engine->i915)) {
+	case 9:
+		wa_bb_fn[0] = gen9_init_indirectctx_bb;
+		wa_bb_fn[1] = gen9_init_perctx_bb;
+		break;
+	case 8:
+		wa_bb_fn[0] = gen8_init_indirectctx_bb;
+		wa_bb_fn[1] = gen8_init_perctx_bb;
+		break;
+	default:
+		MISSING_CASE(INTEL_GEN(engine->i915));
 		return 0;
 	}
 
-	/* some WA perform writes to scratch page, ensure it is valid */
-	if (!engine->scratch) {
-		DRM_ERROR("scratch page not allocated for %s\n", engine->name);
-		return -EINVAL;
-	}
-
-	ret = lrc_setup_wa_ctx_obj(engine, PAGE_SIZE);
+	ret = lrc_setup_wa_ctx(engine);
 	if (ret) {
 		DRM_DEBUG_DRIVER("Failed to setup context WA page: %d\n", ret);
 		return ret;
 	}
 
 	page = i915_gem_object_get_dirty_page(wa_ctx->vma->obj, 0);
-	batch = kmap_atomic(page);
-	offset = 0;
-
-	if (IS_GEN8(engine->i915)) {
-		ret = gen8_init_indirectctx_bb(engine,
-					       &wa_ctx->indirect_ctx,
-					       batch,
-					       &offset);
-		if (ret)
-			goto out;
-
-		ret = gen8_init_perctx_bb(engine,
-					  &wa_ctx->per_ctx,
-					  batch,
-					  &offset);
-		if (ret)
-			goto out;
-	} else if (IS_GEN9(engine->i915)) {
-		ret = gen9_init_indirectctx_bb(engine,
-					       &wa_ctx->indirect_ctx,
-					       batch,
-					       &offset);
-		if (ret)
-			goto out;
+	batch = batch_ptr = kmap_atomic(page);
 
-		ret = gen9_init_perctx_bb(engine,
-					  &wa_ctx->per_ctx,
-					  batch,
-					  &offset);
-		if (ret)
-			goto out;
+	/*
+	 * Emit the two workaround batch buffers, recording the offset from the
+	 * start of the workaround batch buffer object for each and their
+	 * respective sizes.
+	 */
+	for (i = 0; i < ARRAY_SIZE(wa_bb_fn); i++) {
+		wa_bb[i]->offset = batch_ptr - batch;
+		if (WARN_ON(!IS_ALIGNED(wa_bb[i]->offset, CACHELINE_BYTES))) {
+			ret = -EINVAL;
+			break;
+		}
+		batch_ptr = wa_bb_fn[i](engine, batch_ptr);
+		wa_bb[i]->size = batch_ptr - (batch + wa_bb[i]->offset);
 	}
 
-out:
+	BUG_ON(batch_ptr - batch > CTX_WA_BB_OBJ_SIZE);
+
 	kunmap_atomic(batch);
 	if (ret)
-		lrc_destroy_wa_ctx_obj(engine);
+		lrc_destroy_wa_ctx(engine);
 
 	return ret;
 }
 
+static u32 port_seqno(struct execlist_port *port)
+{
+	return port->request ? port->request->global_seqno : 0;
+}
+
 static int gen8_init_common_ring(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
@@ -1334,7 +1150,6 @@ static int gen8_init_common_ring(struct intel_engine_cs *engine)
 
 	I915_WRITE(RING_HWSTAM(engine->mmio_base), 0xffffffff);
 	I915_WRITE(RING_MODE_GEN7(engine),
-		   _MASKED_BIT_DISABLE(GFX_REPLAY_MODE) |
 		   _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE));
 	I915_WRITE(RING_HWS_PGA(engine->mmio_base),
 		   engine->status_page.ggtt_offset);
@@ -1343,7 +1158,12 @@ static int gen8_init_common_ring(struct intel_engine_cs *engine)
 	DRM_DEBUG_DRIVER("Execlists enabled for %s\n", engine->name);
 
 	/* After a GPU reset, we may have requests to replay */
+	clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted);
 	if (!execlists_elsp_idle(engine)) {
+		DRM_DEBUG_DRIVER("Restarting %s from requests [0x%x, 0x%x]\n",
+				 engine->name,
+				 port_seqno(&engine->execlist_port[0]),
+				 port_seqno(&engine->execlist_port[1]));
 		engine->execlist_port[0].count = 0;
 		engine->execlist_port[1].count = 0;
 		execlists_submit_ports(engine);
@@ -1388,7 +1208,6 @@ static int gen9_init_render_ring(struct intel_engine_cs *engine)
 static void reset_common_ring(struct intel_engine_cs *engine,
 			      struct drm_i915_gem_request *request)
 {
-	struct drm_i915_private *dev_priv = engine->i915;
 	struct execlist_port *port = engine->execlist_port;
 	struct intel_context *ce;
 
@@ -1429,7 +1248,6 @@ static void reset_common_ring(struct intel_engine_cs *engine,
 		return;
 
 	/* Catch up with any missed context-switch interrupts */
-	I915_WRITE(RING_CONTEXT_STATUS_PTR(engine), _MASKED_FIELD(0xffff, 0));
 	if (request->ctx != port[0].request->ctx) {
 		i915_gem_request_put(port[0].request);
 		port[0] = port[1];
@@ -1440,42 +1258,42 @@ static void reset_common_ring(struct intel_engine_cs *engine,
 
 	/* Reset WaIdleLiteRestore:bdw,skl as well */
 	request->tail = request->wa_tail - WA_TAIL_DWORDS * sizeof(u32);
+	GEM_BUG_ON(!IS_ALIGNED(request->tail, 8));
 }
 
 static int intel_logical_ring_emit_pdps(struct drm_i915_gem_request *req)
 {
 	struct i915_hw_ppgtt *ppgtt = req->ctx->ppgtt;
-	struct intel_ring *ring = req->ring;
 	struct intel_engine_cs *engine = req->engine;
-	const int num_lri_cmds = GEN8_LEGACY_PDPES * 2;
-	int i, ret;
+	const int num_lri_cmds = GEN8_3LVL_PDPES * 2;
+	u32 *cs;
+	int i;
 
-	ret = intel_ring_begin(req, num_lri_cmds * 2 + 2);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, num_lri_cmds * 2 + 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(num_lri_cmds));
-	for (i = GEN8_LEGACY_PDPES - 1; i >= 0; i--) {
+	*cs++ = MI_LOAD_REGISTER_IMM(num_lri_cmds);
+	for (i = GEN8_3LVL_PDPES - 1; i >= 0; i--) {
 		const dma_addr_t pd_daddr = i915_page_dir_dma_addr(ppgtt, i);
 
-		intel_ring_emit_reg(ring, GEN8_RING_PDP_UDW(engine, i));
-		intel_ring_emit(ring, upper_32_bits(pd_daddr));
-		intel_ring_emit_reg(ring, GEN8_RING_PDP_LDW(engine, i));
-		intel_ring_emit(ring, lower_32_bits(pd_daddr));
+		*cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_UDW(engine, i));
+		*cs++ = upper_32_bits(pd_daddr);
+		*cs++ = i915_mmio_reg_offset(GEN8_RING_PDP_LDW(engine, i));
+		*cs++ = lower_32_bits(pd_daddr);
 	}
 
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
 
 static int gen8_emit_bb_start(struct drm_i915_gem_request *req,
 			      u64 offset, u32 len,
-			      unsigned int dispatch_flags)
+			      const unsigned int flags)
 {
-	struct intel_ring *ring = req->ring;
-	bool ppgtt = !(dispatch_flags & I915_DISPATCH_SECURE);
+	u32 *cs;
 	int ret;
 
 	/* Don't rely in hw updating PDPs, specially in lite-restore.
@@ -1485,30 +1303,28 @@ static int gen8_emit_bb_start(struct drm_i915_gem_request *req,
 	 * not idle). PML4 is allocated during ppgtt init so this is
 	 * not needed in 48-bit.*/
 	if (req->ctx->ppgtt &&
-	    (intel_engine_flag(req->engine) & req->ctx->ppgtt->pd_dirty_rings)) {
-		if (!USES_FULL_48BIT_PPGTT(req->i915) &&
-		    !intel_vgpu_active(req->i915)) {
-			ret = intel_logical_ring_emit_pdps(req);
-			if (ret)
-				return ret;
-		}
+	    (intel_engine_flag(req->engine) & req->ctx->ppgtt->pd_dirty_rings) &&
+	    !i915_vm_is_48bit(&req->ctx->ppgtt->base) &&
+	    !intel_vgpu_active(req->i915)) {
+		ret = intel_logical_ring_emit_pdps(req);
+		if (ret)
+			return ret;
 
 		req->ctx->ppgtt->pd_dirty_rings &= ~intel_engine_flag(req->engine);
 	}
 
-	ret = intel_ring_begin(req, 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	/* FIXME(BDW): Address space and security selectors. */
-	intel_ring_emit(ring, MI_BATCH_BUFFER_START_GEN8 |
-			(ppgtt<<8) |
-			(dispatch_flags & I915_DISPATCH_RS ?
-			 MI_BATCH_RESOURCE_STREAMER : 0));
-	intel_ring_emit(ring, lower_32_bits(offset));
-	intel_ring_emit(ring, upper_32_bits(offset));
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	*cs++ = MI_BATCH_BUFFER_START_GEN8 |
+		(flags & I915_DISPATCH_SECURE ? 0 : BIT(8)) |
+		(flags & I915_DISPATCH_RS ? MI_BATCH_RESOURCE_STREAMER : 0);
+	*cs++ = lower_32_bits(offset);
+	*cs++ = upper_32_bits(offset);
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -1529,13 +1345,11 @@ static void gen8_logical_ring_disable_irq(struct intel_engine_cs *engine)
 
 static int gen8_emit_flush(struct drm_i915_gem_request *request, u32 mode)
 {
-	struct intel_ring *ring = request->ring;
-	u32 cmd;
-	int ret;
+	u32 cmd, *cs;
 
-	ret = intel_ring_begin(request, 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(request, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	cmd = MI_FLUSH_DW + 1;
 
@@ -1552,13 +1366,11 @@ static int gen8_emit_flush(struct drm_i915_gem_request *request, u32 mode)
 			cmd |= MI_INVALIDATE_BSD;
 	}
 
-	intel_ring_emit(ring, cmd);
-	intel_ring_emit(ring,
-			I915_GEM_HWS_SCRATCH_ADDR |
-			MI_FLUSH_DW_USE_GTT);
-	intel_ring_emit(ring, 0); /* upper addr */
-	intel_ring_emit(ring, 0); /* value */
-	intel_ring_advance(ring);
+	*cs++ = cmd;
+	*cs++ = I915_GEM_HWS_SCRATCH_ADDR | MI_FLUSH_DW_USE_GTT;
+	*cs++ = 0; /* upper addr */
+	*cs++ = 0; /* value */
+	intel_ring_advance(request, cs);
 
 	return 0;
 }
@@ -1566,13 +1378,11 @@ static int gen8_emit_flush(struct drm_i915_gem_request *request, u32 mode)
 static int gen8_emit_flush_render(struct drm_i915_gem_request *request,
 				  u32 mode)
 {
-	struct intel_ring *ring = request->ring;
 	struct intel_engine_cs *engine = request->engine;
 	u32 scratch_addr =
 		i915_ggtt_offset(engine->scratch) + 2 * CACHELINE_BYTES;
 	bool vf_flush_wa = false, dc_flush_wa = false;
-	u32 flags = 0;
-	int ret;
+	u32 *cs, flags = 0;
 	int len;
 
 	flags |= PIPE_CONTROL_CS_STALL;
@@ -1614,62 +1424,25 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request,
 	if (dc_flush_wa)
 		len += 12;
 
-	ret = intel_ring_begin(request, len);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(request, len);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	if (vf_flush_wa) {
-		intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(6));
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-	}
+	if (vf_flush_wa)
+		cs = gen8_emit_pipe_control(cs, 0, 0);
 
-	if (dc_flush_wa) {
-		intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(6));
-		intel_ring_emit(ring, PIPE_CONTROL_DC_FLUSH_ENABLE);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-	}
+	if (dc_flush_wa)
+		cs = gen8_emit_pipe_control(cs, PIPE_CONTROL_DC_FLUSH_ENABLE,
+					    0);
 
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(6));
-	intel_ring_emit(ring, flags);
-	intel_ring_emit(ring, scratch_addr);
-	intel_ring_emit(ring, 0);
-	intel_ring_emit(ring, 0);
-	intel_ring_emit(ring, 0);
-
-	if (dc_flush_wa) {
-		intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(6));
-		intel_ring_emit(ring, PIPE_CONTROL_CS_STALL);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, 0);
-	}
+	cs = gen8_emit_pipe_control(cs, flags, scratch_addr);
 
-	intel_ring_advance(ring);
+	if (dc_flush_wa)
+		cs = gen8_emit_pipe_control(cs, PIPE_CONTROL_CS_STALL, 0);
 
-	return 0;
-}
+	intel_ring_advance(request, cs);
 
-static void bxt_a_seqno_barrier(struct intel_engine_cs *engine)
-{
-	/*
-	 * On BXT A steppings there is a HW coherency issue whereby the
-	 * MI_STORE_DATA_IMM storing the completed request's seqno
-	 * occasionally doesn't invalidate the CPU cache. Work around this by
-	 * clflushing the corresponding cacheline whenever the caller wants
-	 * the coherency to be guaranteed. Note that this cacheline is known
-	 * to be clean at this point, since we only write it in
-	 * bxt_a_set_seqno(), where we also do a clflush after the write. So
-	 * this clflush in practice becomes an invalidate operation.
-	 */
-	intel_flush_status_page(engine, I915_GEM_HWS_INDEX);
+	return 0;
 }
 
 /*
@@ -1677,34 +1450,34 @@ static void bxt_a_seqno_barrier(struct intel_engine_cs *engine)
  * used as a workaround for not being allowed to do lite
  * restore with HEAD==TAIL (WaIdleLiteRestore).
  */
-static void gen8_emit_wa_tail(struct drm_i915_gem_request *request, u32 *out)
+static void gen8_emit_wa_tail(struct drm_i915_gem_request *request, u32 *cs)
 {
-	*out++ = MI_NOOP;
-	*out++ = MI_NOOP;
-	request->wa_tail = intel_ring_offset(request->ring, out);
+	*cs++ = MI_NOOP;
+	*cs++ = MI_NOOP;
+	request->wa_tail = intel_ring_offset(request, cs);
 }
 
-static void gen8_emit_breadcrumb(struct drm_i915_gem_request *request,
-				 u32 *out)
+static void gen8_emit_breadcrumb(struct drm_i915_gem_request *request, u32 *cs)
 {
 	/* w/a: bit 5 needs to be zero for MI_FLUSH_DW address. */
 	BUILD_BUG_ON(I915_GEM_HWS_INDEX_ADDR & (1 << 5));
 
-	*out++ = (MI_FLUSH_DW + 1) | MI_FLUSH_DW_OP_STOREDW;
-	*out++ = intel_hws_seqno_address(request->engine) | MI_FLUSH_DW_USE_GTT;
-	*out++ = 0;
-	*out++ = request->global_seqno;
-	*out++ = MI_USER_INTERRUPT;
-	*out++ = MI_NOOP;
-	request->tail = intel_ring_offset(request->ring, out);
+	*cs++ = (MI_FLUSH_DW + 1) | MI_FLUSH_DW_OP_STOREDW;
+	*cs++ = intel_hws_seqno_address(request->engine) | MI_FLUSH_DW_USE_GTT;
+	*cs++ = 0;
+	*cs++ = request->global_seqno;
+	*cs++ = MI_USER_INTERRUPT;
+	*cs++ = MI_NOOP;
+	request->tail = intel_ring_offset(request, cs);
+	GEM_BUG_ON(!IS_ALIGNED(request->tail, 8));
 
-	gen8_emit_wa_tail(request, out);
+	gen8_emit_wa_tail(request, cs);
 }
 
 static const int gen8_emit_breadcrumb_sz = 6 + WA_TAIL_DWORDS;
 
 static void gen8_emit_breadcrumb_render(struct drm_i915_gem_request *request,
-					u32 *out)
+					u32 *cs)
 {
 	/* We're using qword write, seqno should be aligned to 8 bytes. */
 	BUILD_BUG_ON(I915_GEM_HWS_INDEX & 1);
@@ -1713,20 +1486,20 @@ static void gen8_emit_breadcrumb_render(struct drm_i915_gem_request *request,
 	 * need a prior CS_STALL, which is emitted by the flush
 	 * following the batch.
 	 */
-	*out++ = GFX_OP_PIPE_CONTROL(6);
-	*out++ = (PIPE_CONTROL_GLOBAL_GTT_IVB |
-		  PIPE_CONTROL_CS_STALL |
-		  PIPE_CONTROL_QW_WRITE);
-	*out++ = intel_hws_seqno_address(request->engine);
-	*out++ = 0;
-	*out++ = request->global_seqno;
+	*cs++ = GFX_OP_PIPE_CONTROL(6);
+	*cs++ = PIPE_CONTROL_GLOBAL_GTT_IVB | PIPE_CONTROL_CS_STALL |
+		PIPE_CONTROL_QW_WRITE;
+	*cs++ = intel_hws_seqno_address(request->engine);
+	*cs++ = 0;
+	*cs++ = request->global_seqno;
 	/* We're thrashing one dword of HWS. */
-	*out++ = 0;
-	*out++ = MI_USER_INTERRUPT;
-	*out++ = MI_NOOP;
-	request->tail = intel_ring_offset(request->ring, out);
+	*cs++ = 0;
+	*cs++ = MI_USER_INTERRUPT;
+	*cs++ = MI_NOOP;
+	request->tail = intel_ring_offset(request, cs);
+	GEM_BUG_ON(!IS_ALIGNED(request->tail, 8));
 
-	gen8_emit_wa_tail(request, out);
+	gen8_emit_wa_tail(request, cs);
 }
 
 static const int gen8_emit_breadcrumb_render_sz = 8 + WA_TAIL_DWORDS;
@@ -1735,7 +1508,7 @@ static int gen8_init_rcs_context(struct drm_i915_gem_request *req)
 {
 	int ret;
 
-	ret = intel_logical_ring_workarounds_emit(req);
+	ret = intel_ring_workarounds_emit(req);
 	if (ret)
 		return ret;
 
@@ -1781,7 +1554,7 @@ void intel_logical_ring_cleanup(struct intel_engine_cs *engine)
 
 	intel_engine_cleanup_common(engine);
 
-	lrc_destroy_wa_ctx_obj(engine);
+	lrc_destroy_wa_ctx(engine);
 	engine->i915 = NULL;
 	dev_priv->engine[engine->id] = NULL;
 	kfree(engine);
@@ -1819,8 +1592,6 @@ logical_ring_default_vfuncs(struct intel_engine_cs *engine)
 	engine->irq_enable = gen8_logical_ring_enable_irq;
 	engine->irq_disable = gen8_logical_ring_disable_irq;
 	engine->emit_bb_start = gen8_emit_bb_start;
-	if (IS_BXT_REVID(engine->i915, 0, BXT_REVID_A1))
-		engine->irq_seqno_barrier = bxt_a_seqno_barrier;
 }
 
 static inline void
@@ -1877,7 +1648,6 @@ logical_ring_setup(struct intel_engine_cs *engine)
 	tasklet_init(&engine->irq_tasklet,
 		     intel_lrc_irq_handler, (unsigned long)engine);
 
-	logical_ring_init_platform_invariants(engine);
 	logical_ring_default_vfuncs(engine);
 	logical_ring_default_irqs(engine);
 }
@@ -2015,105 +1785,89 @@ static u32 intel_lr_indirect_ctx_offset(struct intel_engine_cs *engine)
 	return indirect_ctx_offset;
 }
 
-static void execlists_init_reg_state(u32 *reg_state,
+static void execlists_init_reg_state(u32 *regs,
 				     struct i915_gem_context *ctx,
 				     struct intel_engine_cs *engine,
 				     struct intel_ring *ring)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
 	struct i915_hw_ppgtt *ppgtt = ctx->ppgtt ?: dev_priv->mm.aliasing_ppgtt;
+	u32 base = engine->mmio_base;
+	bool rcs = engine->id == RCS;
+
+	/* A context is actually a big batch buffer with several
+	 * MI_LOAD_REGISTER_IMM commands followed by (reg, value) pairs. The
+	 * values we are setting here are only for the first context restore:
+	 * on a subsequent save, the GPU will recreate this batchbuffer with new
+	 * values (including all the missing MI_LOAD_REGISTER_IMM commands that
+	 * we are not initializing here).
+	 */
+	regs[CTX_LRI_HEADER_0] = MI_LOAD_REGISTER_IMM(rcs ? 14 : 11) |
+				 MI_LRI_FORCE_POSTED;
+
+	CTX_REG(regs, CTX_CONTEXT_CONTROL, RING_CONTEXT_CONTROL(engine),
+		_MASKED_BIT_ENABLE(CTX_CTRL_INHIBIT_SYN_CTX_SWITCH |
+				   CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT |
+				   (HAS_RESOURCE_STREAMER(dev_priv) ?
+				   CTX_CTRL_RS_CTX_ENABLE : 0)));
+	CTX_REG(regs, CTX_RING_HEAD, RING_HEAD(base), 0);
+	CTX_REG(regs, CTX_RING_TAIL, RING_TAIL(base), 0);
+	CTX_REG(regs, CTX_RING_BUFFER_START, RING_START(base), 0);
+	CTX_REG(regs, CTX_RING_BUFFER_CONTROL, RING_CTL(base),
+		RING_CTL_SIZE(ring->size) | RING_VALID);
+	CTX_REG(regs, CTX_BB_HEAD_U, RING_BBADDR_UDW(base), 0);
+	CTX_REG(regs, CTX_BB_HEAD_L, RING_BBADDR(base), 0);
+	CTX_REG(regs, CTX_BB_STATE, RING_BBSTATE(base), RING_BB_PPGTT);
+	CTX_REG(regs, CTX_SECOND_BB_HEAD_U, RING_SBBADDR_UDW(base), 0);
+	CTX_REG(regs, CTX_SECOND_BB_HEAD_L, RING_SBBADDR(base), 0);
+	CTX_REG(regs, CTX_SECOND_BB_STATE, RING_SBBSTATE(base), 0);
+	if (rcs) {
+		CTX_REG(regs, CTX_BB_PER_CTX_PTR, RING_BB_PER_CTX_PTR(base), 0);
+		CTX_REG(regs, CTX_RCS_INDIRECT_CTX, RING_INDIRECT_CTX(base), 0);
+		CTX_REG(regs, CTX_RCS_INDIRECT_CTX_OFFSET,
+			RING_INDIRECT_CTX_OFFSET(base), 0);
 
-	/* A context is actually a big batch buffer with several MI_LOAD_REGISTER_IMM
-	 * commands followed by (reg, value) pairs. The values we are setting here are
-	 * only for the first context restore: on a subsequent save, the GPU will
-	 * recreate this batchbuffer with new values (including all the missing
-	 * MI_LOAD_REGISTER_IMM commands that we are not initializing here). */
-	reg_state[CTX_LRI_HEADER_0] =
-		MI_LOAD_REGISTER_IMM(engine->id == RCS ? 14 : 11) | MI_LRI_FORCE_POSTED;
-	ASSIGN_CTX_REG(reg_state, CTX_CONTEXT_CONTROL,
-		       RING_CONTEXT_CONTROL(engine),
-		       _MASKED_BIT_ENABLE(CTX_CTRL_INHIBIT_SYN_CTX_SWITCH |
-					  CTX_CTRL_ENGINE_CTX_RESTORE_INHIBIT |
-					  (HAS_RESOURCE_STREAMER(dev_priv) ?
-					   CTX_CTRL_RS_CTX_ENABLE : 0)));
-	ASSIGN_CTX_REG(reg_state, CTX_RING_HEAD, RING_HEAD(engine->mmio_base),
-		       0);
-	ASSIGN_CTX_REG(reg_state, CTX_RING_TAIL, RING_TAIL(engine->mmio_base),
-		       0);
-	ASSIGN_CTX_REG(reg_state, CTX_RING_BUFFER_START,
-		       RING_START(engine->mmio_base), 0);
-	ASSIGN_CTX_REG(reg_state, CTX_RING_BUFFER_CONTROL,
-		       RING_CTL(engine->mmio_base),
-		       RING_CTL_SIZE(ring->size) | RING_VALID);
-	ASSIGN_CTX_REG(reg_state, CTX_BB_HEAD_U,
-		       RING_BBADDR_UDW(engine->mmio_base), 0);
-	ASSIGN_CTX_REG(reg_state, CTX_BB_HEAD_L,
-		       RING_BBADDR(engine->mmio_base), 0);
-	ASSIGN_CTX_REG(reg_state, CTX_BB_STATE,
-		       RING_BBSTATE(engine->mmio_base),
-		       RING_BB_PPGTT);
-	ASSIGN_CTX_REG(reg_state, CTX_SECOND_BB_HEAD_U,
-		       RING_SBBADDR_UDW(engine->mmio_base), 0);
-	ASSIGN_CTX_REG(reg_state, CTX_SECOND_BB_HEAD_L,
-		       RING_SBBADDR(engine->mmio_base), 0);
-	ASSIGN_CTX_REG(reg_state, CTX_SECOND_BB_STATE,
-		       RING_SBBSTATE(engine->mmio_base), 0);
-	if (engine->id == RCS) {
-		ASSIGN_CTX_REG(reg_state, CTX_BB_PER_CTX_PTR,
-			       RING_BB_PER_CTX_PTR(engine->mmio_base), 0);
-		ASSIGN_CTX_REG(reg_state, CTX_RCS_INDIRECT_CTX,
-			       RING_INDIRECT_CTX(engine->mmio_base), 0);
-		ASSIGN_CTX_REG(reg_state, CTX_RCS_INDIRECT_CTX_OFFSET,
-			       RING_INDIRECT_CTX_OFFSET(engine->mmio_base), 0);
 		if (engine->wa_ctx.vma) {
 			struct i915_ctx_workarounds *wa_ctx = &engine->wa_ctx;
 			u32 ggtt_offset = i915_ggtt_offset(wa_ctx->vma);
 
-			reg_state[CTX_RCS_INDIRECT_CTX+1] =
-				(ggtt_offset + wa_ctx->indirect_ctx.offset * sizeof(uint32_t)) |
-				(wa_ctx->indirect_ctx.size / CACHELINE_DWORDS);
+			regs[CTX_RCS_INDIRECT_CTX + 1] =
+				(ggtt_offset + wa_ctx->indirect_ctx.offset) |
+				(wa_ctx->indirect_ctx.size / CACHELINE_BYTES);
 
-			reg_state[CTX_RCS_INDIRECT_CTX_OFFSET+1] =
+			regs[CTX_RCS_INDIRECT_CTX_OFFSET + 1] =
 				intel_lr_indirect_ctx_offset(engine) << 6;
 
-			reg_state[CTX_BB_PER_CTX_PTR+1] =
-				(ggtt_offset + wa_ctx->per_ctx.offset * sizeof(uint32_t)) |
-				0x01;
+			regs[CTX_BB_PER_CTX_PTR + 1] =
+				(ggtt_offset + wa_ctx->per_ctx.offset) | 0x01;
 		}
 	}
-	reg_state[CTX_LRI_HEADER_1] = MI_LOAD_REGISTER_IMM(9) | MI_LRI_FORCE_POSTED;
-	ASSIGN_CTX_REG(reg_state, CTX_CTX_TIMESTAMP,
-		       RING_CTX_TIMESTAMP(engine->mmio_base), 0);
+
+	regs[CTX_LRI_HEADER_1] = MI_LOAD_REGISTER_IMM(9) | MI_LRI_FORCE_POSTED;
+
+	CTX_REG(regs, CTX_CTX_TIMESTAMP, RING_CTX_TIMESTAMP(base), 0);
 	/* PDP values well be assigned later if needed */
-	ASSIGN_CTX_REG(reg_state, CTX_PDP3_UDW, GEN8_RING_PDP_UDW(engine, 3),
-		       0);
-	ASSIGN_CTX_REG(reg_state, CTX_PDP3_LDW, GEN8_RING_PDP_LDW(engine, 3),
-		       0);
-	ASSIGN_CTX_REG(reg_state, CTX_PDP2_UDW, GEN8_RING_PDP_UDW(engine, 2),
-		       0);
-	ASSIGN_CTX_REG(reg_state, CTX_PDP2_LDW, GEN8_RING_PDP_LDW(engine, 2),
-		       0);
-	ASSIGN_CTX_REG(reg_state, CTX_PDP1_UDW, GEN8_RING_PDP_UDW(engine, 1),
-		       0);
-	ASSIGN_CTX_REG(reg_state, CTX_PDP1_LDW, GEN8_RING_PDP_LDW(engine, 1),
-		       0);
-	ASSIGN_CTX_REG(reg_state, CTX_PDP0_UDW, GEN8_RING_PDP_UDW(engine, 0),
-		       0);
-	ASSIGN_CTX_REG(reg_state, CTX_PDP0_LDW, GEN8_RING_PDP_LDW(engine, 0),
-		       0);
-
-	if (ppgtt && USES_FULL_48BIT_PPGTT(ppgtt->base.dev)) {
+	CTX_REG(regs, CTX_PDP3_UDW, GEN8_RING_PDP_UDW(engine, 3), 0);
+	CTX_REG(regs, CTX_PDP3_LDW, GEN8_RING_PDP_LDW(engine, 3), 0);
+	CTX_REG(regs, CTX_PDP2_UDW, GEN8_RING_PDP_UDW(engine, 2), 0);
+	CTX_REG(regs, CTX_PDP2_LDW, GEN8_RING_PDP_LDW(engine, 2), 0);
+	CTX_REG(regs, CTX_PDP1_UDW, GEN8_RING_PDP_UDW(engine, 1), 0);
+	CTX_REG(regs, CTX_PDP1_LDW, GEN8_RING_PDP_LDW(engine, 1), 0);
+	CTX_REG(regs, CTX_PDP0_UDW, GEN8_RING_PDP_UDW(engine, 0), 0);
+	CTX_REG(regs, CTX_PDP0_LDW, GEN8_RING_PDP_LDW(engine, 0), 0);
+
+	if (ppgtt && i915_vm_is_48bit(&ppgtt->base)) {
 		/* 64b PPGTT (48bit canonical)
 		 * PDP0_DESCRIPTOR contains the base address to PML4 and
 		 * other PDP Descriptors are ignored.
 		 */
-		ASSIGN_CTX_PML4(ppgtt, reg_state);
+		ASSIGN_CTX_PML4(ppgtt, regs);
 	}
 
-	if (engine->id == RCS) {
-		reg_state[CTX_LRI_HEADER_2] = MI_LOAD_REGISTER_IMM(1);
-		ASSIGN_CTX_REG(reg_state, CTX_R_PWR_CLK_STATE, GEN8_R_PWR_CLK_STATE,
-			       make_rpcs(dev_priv));
+	if (rcs) {
+		regs[CTX_LRI_HEADER_2] = MI_LOAD_REGISTER_IMM(1);
+		CTX_REG(regs, CTX_R_PWR_CLK_STATE, GEN8_R_PWR_CLK_STATE,
+			make_rpcs(dev_priv));
 	}
 }
 
diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h
index 0c852c024227..5fc07761caff 100644
--- a/drivers/gpu/drm/i915/intel_lrc.h
+++ b/drivers/gpu/drm/i915/intel_lrc.h
@@ -68,8 +68,6 @@ void intel_logical_ring_cleanup(struct intel_engine_cs *engine);
 int logical_render_ring_init(struct intel_engine_cs *engine);
 int logical_xcs_ring_init(struct intel_engine_cs *engine);
 
-int intel_engines_init(struct drm_i915_private *dev_priv);
-
 /* Logical Ring Contexts */
 
 /* One extra page is added before LRC for GuC as shared data */
@@ -90,6 +88,5 @@ uint64_t intel_lr_context_descriptor(struct i915_gem_context *ctx,
 int intel_sanitize_enable_execlists(struct drm_i915_private *dev_priv,
 				    int enable_execlists);
 void intel_execlists_enable_submission(struct drm_i915_private *dev_priv);
-bool intel_execlists_idle(struct drm_i915_private *dev_priv);
 
 #endif /* _INTEL_LRC_H_ */
diff --git a/drivers/gpu/drm/i915/intel_lspcon.c b/drivers/gpu/drm/i915/intel_lspcon.c
index c300647ef604..71cbe9c08932 100644
--- a/drivers/gpu/drm/i915/intel_lspcon.c
+++ b/drivers/gpu/drm/i915/intel_lspcon.c
@@ -162,21 +162,8 @@ static void lspcon_resume_in_pcon_wa(struct intel_lspcon *lspcon)
 	struct drm_i915_private *dev_priv = to_i915(dig_port->base.base.dev);
 	unsigned long start = jiffies;
 
-	if (!lspcon->desc_valid)
-		return;
-
 	while (1) {
-		struct intel_dp_desc desc;
-
-		/*
-		 * The w/a only applies in PCON mode and we don't expect any
-		 * AUX errors.
-		 */
-		if (!__intel_dp_read_desc(intel_dp, &desc))
-			return;
-
-		if (intel_digital_port_connected(dev_priv, dig_port) &&
-		    !memcmp(&intel_dp->desc, &desc, sizeof(desc))) {
+		if (intel_digital_port_connected(dev_priv, dig_port)) {
 			DRM_DEBUG_KMS("LSPCON recovering in PCON mode after %u ms\n",
 				      jiffies_to_msecs(jiffies - start));
 			return;
@@ -253,7 +240,7 @@ bool lspcon_init(struct intel_digital_port *intel_dig_port)
 		return false;
 	}
 
-	lspcon->desc_valid = intel_dp_read_desc(dp);
+	intel_dp_read_desc(dp);
 
 	DRM_DEBUG_KMS("Success: LSPCON init\n");
 	return true;
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index 9ca4dc4d2378..8b942ef2b3ec 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -91,12 +91,11 @@ static bool intel_lvds_get_hw_state(struct intel_encoder *encoder,
 	struct drm_device *dev = encoder->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_lvds_encoder *lvds_encoder = to_lvds_encoder(&encoder->base);
-	enum intel_display_power_domain power_domain;
 	u32 tmp;
 	bool ret;
 
-	power_domain = intel_display_port_power_domain(encoder);
-	if (!intel_display_power_get_if_enabled(dev_priv, power_domain))
+	if (!intel_display_power_get_if_enabled(dev_priv,
+						encoder->power_domain))
 		return false;
 
 	ret = false;
@@ -114,7 +113,7 @@ static bool intel_lvds_get_hw_state(struct intel_encoder *encoder,
 	ret = true;
 
 out:
-	intel_display_power_put(dev_priv, power_domain);
+	intel_display_power_put(dev_priv, encoder->power_domain);
 
 	return ret;
 }
@@ -1066,6 +1065,7 @@ void intel_lvds_init(struct drm_i915_private *dev_priv)
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 
 	intel_encoder->type = INTEL_OUTPUT_LVDS;
+	intel_encoder->power_domain = POWER_DOMAIN_PORT_OTHER;
 	intel_encoder->port = PORT_NONE;
 	intel_encoder->cloneable = 0;
 	if (HAS_PCH_SPLIT(dev_priv))
diff --git a/drivers/gpu/drm/i915/intel_mocs.c b/drivers/gpu/drm/i915/intel_mocs.c
index c787fc4e6eb9..92e461c68385 100644
--- a/drivers/gpu/drm/i915/intel_mocs.c
+++ b/drivers/gpu/drm/i915/intel_mocs.c
@@ -178,7 +178,7 @@ static bool get_mocs_settings(struct drm_i915_private *dev_priv,
 {
 	bool result = false;
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	if (IS_GEN9_BC(dev_priv)) {
 		table->size  = ARRAY_SIZE(skylake_mocs_table);
 		table->table = skylake_mocs_table;
 		result = true;
@@ -191,7 +191,7 @@ static bool get_mocs_settings(struct drm_i915_private *dev_priv,
 			  "Platform that should have a MOCS table does not.\n");
 	}
 
-	/* WaDisableSkipCaching:skl,bxt,kbl */
+	/* WaDisableSkipCaching:skl,bxt,kbl,glk */
 	if (IS_GEN9(dev_priv)) {
 		int i;
 
@@ -276,23 +276,22 @@ int intel_mocs_init_engine(struct intel_engine_cs *engine)
 static int emit_mocs_control_table(struct drm_i915_gem_request *req,
 				   const struct drm_i915_mocs_table *table)
 {
-	struct intel_ring *ring = req->ring;
 	enum intel_engine_id engine = req->engine->id;
 	unsigned int index;
-	int ret;
+	u32 *cs;
 
 	if (WARN_ON(table->size > GEN9_NUM_MOCS_ENTRIES))
 		return -ENODEV;
 
-	ret = intel_ring_begin(req, 2 + 2 * GEN9_NUM_MOCS_ENTRIES);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 2 + 2 * GEN9_NUM_MOCS_ENTRIES);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(GEN9_NUM_MOCS_ENTRIES));
+	*cs++ = MI_LOAD_REGISTER_IMM(GEN9_NUM_MOCS_ENTRIES);
 
 	for (index = 0; index < table->size; index++) {
-		intel_ring_emit_reg(ring, mocs_register(engine, index));
-		intel_ring_emit(ring, table->table[index].control_value);
+		*cs++ = i915_mmio_reg_offset(mocs_register(engine, index));
+		*cs++ = table->table[index].control_value;
 	}
 
 	/*
@@ -304,12 +303,12 @@ static int emit_mocs_control_table(struct drm_i915_gem_request *req,
 	 * that value to all the used entries.
 	 */
 	for (; index < GEN9_NUM_MOCS_ENTRIES; index++) {
-		intel_ring_emit_reg(ring, mocs_register(engine, index));
-		intel_ring_emit(ring, table->table[0].control_value);
+		*cs++ = i915_mmio_reg_offset(mocs_register(engine, index));
+		*cs++ = table->table[0].control_value;
 	}
 
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -336,29 +335,27 @@ static inline u32 l3cc_combine(const struct drm_i915_mocs_table *table,
 static int emit_mocs_l3cc_table(struct drm_i915_gem_request *req,
 				const struct drm_i915_mocs_table *table)
 {
-	struct intel_ring *ring = req->ring;
 	unsigned int i;
-	int ret;
+	u32 *cs;
 
 	if (WARN_ON(table->size > GEN9_NUM_MOCS_ENTRIES))
 		return -ENODEV;
 
-	ret = intel_ring_begin(req, 2 + GEN9_NUM_MOCS_ENTRIES);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 2 + GEN9_NUM_MOCS_ENTRIES);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring,
-			MI_LOAD_REGISTER_IMM(GEN9_NUM_MOCS_ENTRIES / 2));
+	*cs++ = MI_LOAD_REGISTER_IMM(GEN9_NUM_MOCS_ENTRIES / 2);
 
 	for (i = 0; i < table->size/2; i++) {
-		intel_ring_emit_reg(ring, GEN9_LNCFCMOCS(i));
-		intel_ring_emit(ring, l3cc_combine(table, 2*i, 2*i+1));
+		*cs++ = i915_mmio_reg_offset(GEN9_LNCFCMOCS(i));
+		*cs++ = l3cc_combine(table, 2 * i, 2 * i + 1);
 	}
 
 	if (table->size & 0x01) {
 		/* Odd table size - 1 left over */
-		intel_ring_emit_reg(ring, GEN9_LNCFCMOCS(i));
-		intel_ring_emit(ring, l3cc_combine(table, 2*i, 0));
+		*cs++ = i915_mmio_reg_offset(GEN9_LNCFCMOCS(i));
+		*cs++ = l3cc_combine(table, 2 * i, 0);
 		i++;
 	}
 
@@ -368,12 +365,12 @@ static int emit_mocs_l3cc_table(struct drm_i915_gem_request *req,
 	 * they are reserved by the hardware.
 	 */
 	for (; i < GEN9_NUM_MOCS_ENTRIES / 2; i++) {
-		intel_ring_emit_reg(ring, GEN9_LNCFCMOCS(i));
-		intel_ring_emit(ring, l3cc_combine(table, 0, 0));
+		*cs++ = i915_mmio_reg_offset(GEN9_LNCFCMOCS(i));
+		*cs++ = l3cc_combine(table, 0, 0);
 	}
 
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index 0608fad7f593..5ef9f5bfb92c 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -267,8 +267,7 @@ static int intel_overlay_on(struct intel_overlay *overlay)
 {
 	struct drm_i915_private *dev_priv = overlay->i915;
 	struct drm_i915_gem_request *req;
-	struct intel_ring *ring;
-	int ret;
+	u32 *cs;
 
 	WARN_ON(overlay->active);
 	WARN_ON(IS_I830(dev_priv) && !(dev_priv->quirks & QUIRK_PIPEA_FORCE));
@@ -277,10 +276,10 @@ static int intel_overlay_on(struct intel_overlay *overlay)
 	if (IS_ERR(req))
 		return PTR_ERR(req);
 
-	ret = intel_ring_begin(req, 4);
-	if (ret) {
+	cs = intel_ring_begin(req, 4);
+	if (IS_ERR(cs)) {
 		i915_add_request_no_flush(req);
-		return ret;
+		return PTR_ERR(cs);
 	}
 
 	overlay->active = true;
@@ -288,12 +287,11 @@ static int intel_overlay_on(struct intel_overlay *overlay)
 	if (IS_I830(dev_priv))
 		i830_overlay_clock_gating(dev_priv, false);
 
-	ring = req->ring;
-	intel_ring_emit(ring, MI_OVERLAY_FLIP | MI_OVERLAY_ON);
-	intel_ring_emit(ring, overlay->flip_addr | OFC_UPDATE);
-	intel_ring_emit(ring, MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP);
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	*cs++ = MI_OVERLAY_FLIP | MI_OVERLAY_ON;
+	*cs++ = overlay->flip_addr | OFC_UPDATE;
+	*cs++ = MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return intel_overlay_do_wait_request(overlay, req, NULL);
 }
@@ -326,10 +324,8 @@ static int intel_overlay_continue(struct intel_overlay *overlay,
 {
 	struct drm_i915_private *dev_priv = overlay->i915;
 	struct drm_i915_gem_request *req;
-	struct intel_ring *ring;
 	u32 flip_addr = overlay->flip_addr;
-	u32 tmp;
-	int ret;
+	u32 tmp, *cs;
 
 	WARN_ON(!overlay->active);
 
@@ -345,16 +341,15 @@ static int intel_overlay_continue(struct intel_overlay *overlay,
 	if (IS_ERR(req))
 		return PTR_ERR(req);
 
-	ret = intel_ring_begin(req, 2);
-	if (ret) {
+	cs = intel_ring_begin(req, 2);
+	if (IS_ERR(cs)) {
 		i915_add_request_no_flush(req);
-		return ret;
+		return PTR_ERR(cs);
 	}
 
-	ring = req->ring;
-	intel_ring_emit(ring, MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE);
-	intel_ring_emit(ring, flip_addr);
-	intel_ring_advance(ring);
+	*cs++ = MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE;
+	*cs++ = flip_addr;
+	intel_ring_advance(req, cs);
 
 	intel_overlay_flip_prepare(overlay, vma);
 
@@ -408,9 +403,7 @@ static void intel_overlay_off_tail(struct i915_gem_active *active,
 static int intel_overlay_off(struct intel_overlay *overlay)
 {
 	struct drm_i915_gem_request *req;
-	struct intel_ring *ring;
-	u32 flip_addr = overlay->flip_addr;
-	int ret;
+	u32 *cs, flip_addr = overlay->flip_addr;
 
 	WARN_ON(!overlay->active);
 
@@ -424,25 +417,23 @@ static int intel_overlay_off(struct intel_overlay *overlay)
 	if (IS_ERR(req))
 		return PTR_ERR(req);
 
-	ret = intel_ring_begin(req, 6);
-	if (ret) {
+	cs = intel_ring_begin(req, 6);
+	if (IS_ERR(cs)) {
 		i915_add_request_no_flush(req);
-		return ret;
+		return PTR_ERR(cs);
 	}
 
-	ring = req->ring;
-
 	/* wait for overlay to go idle */
-	intel_ring_emit(ring, MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE);
-	intel_ring_emit(ring, flip_addr);
-	intel_ring_emit(ring, MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP);
+	*cs++ = MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE;
+	*cs++ = flip_addr;
+	*cs++ = MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP;
 
 	/* turn overlay off */
-	intel_ring_emit(ring, MI_OVERLAY_FLIP | MI_OVERLAY_OFF);
-	intel_ring_emit(ring, flip_addr);
-	intel_ring_emit(ring, MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP);
+	*cs++ = MI_OVERLAY_FLIP | MI_OVERLAY_OFF;
+	*cs++ = flip_addr;
+	*cs++ = MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP;
 
-	intel_ring_advance(ring);
+	intel_ring_advance(req, cs);
 
 	intel_overlay_flip_prepare(overlay, NULL);
 
@@ -465,6 +456,7 @@ static int intel_overlay_recover_from_interrupt(struct intel_overlay *overlay)
 static int intel_overlay_release_old_vid(struct intel_overlay *overlay)
 {
 	struct drm_i915_private *dev_priv = overlay->i915;
+	u32 *cs;
 	int ret;
 
 	lockdep_assert_held(&dev_priv->drm.struct_mutex);
@@ -478,23 +470,20 @@ static int intel_overlay_release_old_vid(struct intel_overlay *overlay)
 	if (I915_READ(ISR) & I915_OVERLAY_PLANE_FLIP_PENDING_INTERRUPT) {
 		/* synchronous slowpath */
 		struct drm_i915_gem_request *req;
-		struct intel_ring *ring;
 
 		req = alloc_request(overlay);
 		if (IS_ERR(req))
 			return PTR_ERR(req);
 
-		ret = intel_ring_begin(req, 2);
-		if (ret) {
+		cs = intel_ring_begin(req, 2);
+		if (IS_ERR(cs)) {
 			i915_add_request_no_flush(req);
-			return ret;
+			return PTR_ERR(cs);
 		}
 
-		ring = req->ring;
-		intel_ring_emit(ring,
-				MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP);
-		intel_ring_emit(ring, MI_NOOP);
-		intel_ring_advance(ring);
+		*cs++ = MI_WAIT_FOR_EVENT | MI_WAIT_FOR_OVERLAY_FLIP;
+		*cs++ = MI_NOOP;
+		intel_ring_advance(req, cs);
 
 		ret = intel_overlay_do_wait_request(overlay, req,
 						    intel_overlay_release_old_vid_tail);
diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index 1a6ff26dea20..cb50c527401f 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -1315,7 +1315,7 @@ static u32 i9xx_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz)
 	if (IS_PINEVIEW(dev_priv))
 		clock = KHz(dev_priv->rawclk_freq);
 	else
-		clock = KHz(dev_priv->cdclk_freq);
+		clock = KHz(dev_priv->cdclk.hw.cdclk);
 
 	return DIV_ROUND_CLOSEST(clock, pwm_freq_hz * 32);
 }
@@ -1333,7 +1333,7 @@ static u32 i965_hz_to_pwm(struct intel_connector *connector, u32 pwm_freq_hz)
 	if (IS_G4X(dev_priv))
 		clock = KHz(dev_priv->rawclk_freq);
 	else
-		clock = KHz(dev_priv->cdclk_freq);
+		clock = KHz(dev_priv->cdclk.hw.cdclk);
 
 	return DIV_ROUND_CLOSEST(clock, pwm_freq_hz * 128);
 }
diff --git a/drivers/gpu/drm/i915/intel_pipe_crc.c b/drivers/gpu/drm/i915/intel_pipe_crc.c
index 5aa524e32df7..9fd9c70baeed 100644
--- a/drivers/gpu/drm/i915/intel_pipe_crc.c
+++ b/drivers/gpu/drm/i915/intel_pipe_crc.c
@@ -80,7 +80,7 @@ static int i915_pipe_crc_release(struct inode *inode, struct file *filep)
 
 static int pipe_crc_data_count(struct intel_pipe_crc *pipe_crc)
 {
-	assert_spin_locked(&pipe_crc->lock);
+	lockdep_assert_held(&pipe_crc->lock);
 	return CIRC_CNT(pipe_crc->head, pipe_crc->tail,
 			INTEL_PIPE_CRC_ENTRIES_NR);
 }
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 249623d45be0..99e09f63d4b3 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -65,12 +65,12 @@ static void gen9_init_clock_gating(struct drm_i915_private *dev_priv)
 	I915_WRITE(GEN8_CONFIG0,
 		   I915_READ(GEN8_CONFIG0) | GEN9_DEFAULT_FIXES);
 
-	/* WaEnableChickenDCPR:skl,bxt,kbl */
+	/* WaEnableChickenDCPR:skl,bxt,kbl,glk */
 	I915_WRITE(GEN8_CHICKEN_DCPR_1,
 		   I915_READ(GEN8_CHICKEN_DCPR_1) | MASK_WAKEMEM);
 
 	/* WaFbcTurnOffFbcWatermark:skl,bxt,kbl */
-	/* WaFbcWakeMemOn:skl,bxt,kbl */
+	/* WaFbcWakeMemOn:skl,bxt,kbl,glk */
 	I915_WRITE(DISP_ARB_CTL, I915_READ(DISP_ARB_CTL) |
 		   DISP_FBC_WM_DIS |
 		   DISP_FBC_MEMORY_WAKE);
@@ -99,9 +99,31 @@ static void bxt_init_clock_gating(struct drm_i915_private *dev_priv)
 	 * Wa: Backlight PWM may stop in the asserted state, causing backlight
 	 * to stay fully on.
 	 */
-	if (IS_BXT_REVID(dev_priv, BXT_REVID_B0, REVID_FOREVER))
-		I915_WRITE(GEN9_CLKGATE_DIS_0, I915_READ(GEN9_CLKGATE_DIS_0) |
-			   PWM1_GATING_DIS | PWM2_GATING_DIS);
+	I915_WRITE(GEN9_CLKGATE_DIS_0, I915_READ(GEN9_CLKGATE_DIS_0) |
+		   PWM1_GATING_DIS | PWM2_GATING_DIS);
+}
+
+static void glk_init_clock_gating(struct drm_i915_private *dev_priv)
+{
+	gen9_init_clock_gating(dev_priv);
+
+	/*
+	 * WaDisablePWMClockGating:glk
+	 * Backlight PWM may stop in the asserted state, causing backlight
+	 * to stay fully on.
+	 */
+	I915_WRITE(GEN9_CLKGATE_DIS_0, I915_READ(GEN9_CLKGATE_DIS_0) |
+		   PWM1_GATING_DIS | PWM2_GATING_DIS);
+
+	/* WaDDIIOTimeout:glk */
+	if (IS_GLK_REVID(dev_priv, 0, GLK_REVID_A1)) {
+		u32 val = I915_READ(CHICKEN_MISC_2);
+		val &= ~(GLK_CL0_PWR_DOWN |
+			 GLK_CL1_PWR_DOWN |
+			 GLK_CL2_PWR_DOWN);
+		I915_WRITE(CHICKEN_MISC_2, val);
+	}
+
 }
 
 static void i915_pineview_get_mem_freq(struct drm_i915_private *dev_priv)
@@ -355,6 +377,8 @@ static bool _intel_set_memory_cxsr(struct drm_i915_private *dev_priv, bool enabl
 		return false;
 	}
 
+	trace_intel_memory_cxsr(dev_priv, was_enabled, enable);
+
 	DRM_DEBUG_KMS("memory self-refresh is %s (was %s)\n",
 		      enableddisabled(enable),
 		      enableddisabled(was_enabled));
@@ -393,15 +417,15 @@ static const int pessimal_latency_ns = 5000;
 #define VLV_FIFO_START(dsparb, dsparb2, lo_shift, hi_shift) \
 	((((dsparb) >> (lo_shift)) & 0xff) | ((((dsparb2) >> (hi_shift)) & 0x1) << 8))
 
-static int vlv_get_fifo_size(struct intel_plane *plane)
+static void vlv_get_fifo_size(struct intel_crtc_state *crtc_state)
 {
-	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
-	int sprite0_start, sprite1_start, size;
-
-	if (plane->id == PLANE_CURSOR)
-		return 63;
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct vlv_fifo_state *fifo_state = &crtc_state->wm.vlv.fifo_state;
+	enum pipe pipe = crtc->pipe;
+	int sprite0_start, sprite1_start;
 
-	switch (plane->pipe) {
+	switch (pipe) {
 		uint32_t dsparb, dsparb2, dsparb3;
 	case PIPE_A:
 		dsparb = I915_READ(DSPARB);
@@ -422,26 +446,21 @@ static int vlv_get_fifo_size(struct intel_plane *plane)
 		sprite1_start = VLV_FIFO_START(dsparb3, dsparb2, 8, 20);
 		break;
 	default:
-		return 0;
-	}
-
-	switch (plane->id) {
-	case PLANE_PRIMARY:
-		size = sprite0_start;
-		break;
-	case PLANE_SPRITE0:
-		size = sprite1_start - sprite0_start;
-		break;
-	case PLANE_SPRITE1:
-		size = 512 - 1 - sprite1_start;
-		break;
-	default:
-		return 0;
+		MISSING_CASE(pipe);
+		return;
 	}
 
-	DRM_DEBUG_KMS("%s FIFO size: %d\n", plane->base.name, size);
+	fifo_state->plane[PLANE_PRIMARY] = sprite0_start;
+	fifo_state->plane[PLANE_SPRITE0] = sprite1_start - sprite0_start;
+	fifo_state->plane[PLANE_SPRITE1] = 511 - sprite1_start;
+	fifo_state->plane[PLANE_CURSOR] = 63;
 
-	return size;
+	DRM_DEBUG_KMS("Pipe %c FIFO size: %d/%d/%d/%d\n",
+		      pipe_name(pipe),
+		      fifo_state->plane[PLANE_PRIMARY],
+		      fifo_state->plane[PLANE_SPRITE0],
+		      fifo_state->plane[PLANE_SPRITE1],
+		      fifo_state->plane[PLANE_CURSOR]);
 }
 
 static int i9xx_get_fifo_size(struct drm_i915_private *dev_priv, int plane)
@@ -871,6 +890,8 @@ static void vlv_write_wm_values(struct drm_i915_private *dev_priv,
 	enum pipe pipe;
 
 	for_each_pipe(dev_priv, pipe) {
+		trace_vlv_wm(intel_get_crtc_for_pipe(dev_priv, pipe), wm);
+
 		I915_WRITE(VLV_DDL(pipe),
 			   (wm->ddl[pipe].plane[PLANE_CURSOR] << DDL_CURSOR_SHIFT) |
 			   (wm->ddl[pipe].plane[PLANE_SPRITE1] << DDL_SPRITE_SHIFT(1)) |
@@ -941,12 +962,6 @@ static void vlv_write_wm_values(struct drm_i915_private *dev_priv,
 
 #undef FW_WM_VLV
 
-enum vlv_wm_level {
-	VLV_WM_LEVEL_PM2,
-	VLV_WM_LEVEL_PM5,
-	VLV_WM_LEVEL_DDR_DVFS,
-};
-
 /* latency must be in 0.1us units. */
 static unsigned int vlv_wm_method2(unsigned int pixel_rate,
 				   unsigned int pipe_htotal,
@@ -1017,71 +1032,114 @@ static uint16_t vlv_compute_wm_level(const struct intel_crtc_state *crtc_state,
 	return min_t(int, wm, USHRT_MAX);
 }
 
-static void vlv_compute_fifo(struct intel_crtc *crtc)
+static bool vlv_need_sprite0_fifo_workaround(unsigned int active_planes)
 {
-	struct drm_device *dev = crtc->base.dev;
-	struct vlv_wm_state *wm_state = &crtc->wm_state;
-	struct intel_plane *plane;
-	unsigned int total_rate = 0;
-	const int fifo_size = 512 - 1;
+	return (active_planes & (BIT(PLANE_SPRITE0) |
+				 BIT(PLANE_SPRITE1))) == BIT(PLANE_SPRITE1);
+}
+
+static int vlv_compute_fifo(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc);
+	const struct vlv_pipe_wm *raw =
+		&crtc_state->wm.vlv.raw[VLV_WM_LEVEL_PM2];
+	struct vlv_fifo_state *fifo_state = &crtc_state->wm.vlv.fifo_state;
+	unsigned int active_planes = crtc_state->active_planes & ~BIT(PLANE_CURSOR);
+	int num_active_planes = hweight32(active_planes);
+	const int fifo_size = 511;
 	int fifo_extra, fifo_left = fifo_size;
+	int sprite0_fifo_extra = 0;
+	unsigned int total_rate;
+	enum plane_id plane_id;
 
-	for_each_intel_plane_on_crtc(dev, crtc, plane) {
-		struct intel_plane_state *state =
-			to_intel_plane_state(plane->base.state);
+	/*
+	 * When enabling sprite0 after sprite1 has already been enabled
+	 * we tend to get an underrun unless sprite0 already has some
+	 * FIFO space allcoated. Hence we always allocate at least one
+	 * cacheline for sprite0 whenever sprite1 is enabled.
+	 *
+	 * All other plane enable sequences appear immune to this problem.
+	 */
+	if (vlv_need_sprite0_fifo_workaround(active_planes))
+		sprite0_fifo_extra = 1;
 
-		if (plane->base.type == DRM_PLANE_TYPE_CURSOR)
-			continue;
+	total_rate = raw->plane[PLANE_PRIMARY] +
+		raw->plane[PLANE_SPRITE0] +
+		raw->plane[PLANE_SPRITE1] +
+		sprite0_fifo_extra;
 
-		if (state->base.visible) {
-			wm_state->num_active_planes++;
-			total_rate += state->base.fb->format->cpp[0];
-		}
-	}
+	if (total_rate > fifo_size)
+		return -EINVAL;
 
-	for_each_intel_plane_on_crtc(dev, crtc, plane) {
-		struct intel_plane_state *state =
-			to_intel_plane_state(plane->base.state);
-		unsigned int rate;
+	if (total_rate == 0)
+		total_rate = 1;
 
-		if (plane->base.type == DRM_PLANE_TYPE_CURSOR) {
-			plane->wm.fifo_size = 63;
-			continue;
-		}
+	for_each_plane_id_on_crtc(crtc, plane_id) {
+		unsigned int rate;
 
-		if (!state->base.visible) {
-			plane->wm.fifo_size = 0;
+		if ((active_planes & BIT(plane_id)) == 0) {
+			fifo_state->plane[plane_id] = 0;
 			continue;
 		}
 
-		rate = state->base.fb->format->cpp[0];
-		plane->wm.fifo_size = fifo_size * rate / total_rate;
-		fifo_left -= plane->wm.fifo_size;
+		rate = raw->plane[plane_id];
+		fifo_state->plane[plane_id] = fifo_size * rate / total_rate;
+		fifo_left -= fifo_state->plane[plane_id];
 	}
 
-	fifo_extra = DIV_ROUND_UP(fifo_left, wm_state->num_active_planes ?: 1);
+	fifo_state->plane[PLANE_SPRITE0] += sprite0_fifo_extra;
+	fifo_left -= sprite0_fifo_extra;
+
+	fifo_state->plane[PLANE_CURSOR] = 63;
+
+	fifo_extra = DIV_ROUND_UP(fifo_left, num_active_planes ?: 1);
 
 	/* spread the remainder evenly */
-	for_each_intel_plane_on_crtc(dev, crtc, plane) {
+	for_each_plane_id_on_crtc(crtc, plane_id) {
 		int plane_extra;
 
 		if (fifo_left == 0)
 			break;
 
-		if (plane->base.type == DRM_PLANE_TYPE_CURSOR)
-			continue;
-
-		/* give it all to the first plane if none are active */
-		if (plane->wm.fifo_size == 0 &&
-		    wm_state->num_active_planes)
+		if ((active_planes & BIT(plane_id)) == 0)
 			continue;
 
 		plane_extra = min(fifo_extra, fifo_left);
-		plane->wm.fifo_size += plane_extra;
+		fifo_state->plane[plane_id] += plane_extra;
 		fifo_left -= plane_extra;
 	}
 
-	WARN_ON(fifo_left != 0);
+	WARN_ON(active_planes != 0 && fifo_left != 0);
+
+	/* give it all to the first plane if none are active */
+	if (active_planes == 0) {
+		WARN_ON(fifo_left != fifo_size);
+		fifo_state->plane[PLANE_PRIMARY] = fifo_left;
+	}
+
+	return 0;
+}
+
+static int vlv_num_wm_levels(struct drm_i915_private *dev_priv)
+{
+	return dev_priv->wm.max_level + 1;
+}
+
+/* mark all levels starting from 'level' as invalid */
+static void vlv_invalidate_wms(struct intel_crtc *crtc,
+			       struct vlv_wm_state *wm_state, int level)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+
+	for (; level < vlv_num_wm_levels(dev_priv); level++) {
+		enum plane_id plane_id;
+
+		for_each_plane_id_on_crtc(crtc, plane_id)
+			wm_state->wm[level].plane[plane_id] = USHRT_MAX;
+
+		wm_state->sr[level].cursor = USHRT_MAX;
+		wm_state->sr[level].plane = USHRT_MAX;
+	}
 }
 
 static u16 vlv_invert_wm_value(u16 wm, u16 fifo_size)
@@ -1092,136 +1150,213 @@ static u16 vlv_invert_wm_value(u16 wm, u16 fifo_size)
 		return fifo_size - wm;
 }
 
-static void vlv_invert_wms(struct intel_crtc *crtc)
+/*
+ * Starting from 'level' set all higher
+ * levels to 'value' in the "raw" watermarks.
+ */
+static bool vlv_raw_plane_wm_set(struct intel_crtc_state *crtc_state,
+				 int level, enum plane_id plane_id, u16 value)
 {
-	struct vlv_wm_state *wm_state = &crtc->wm_state;
-	int level;
-
-	for (level = 0; level < wm_state->num_levels; level++) {
-		struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-		const int sr_fifo_size =
-			INTEL_INFO(dev_priv)->num_pipes * 512 - 1;
-		struct intel_plane *plane;
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev);
+	int num_levels = vlv_num_wm_levels(dev_priv);
+	bool dirty = false;
 
-		wm_state->sr[level].plane =
-			vlv_invert_wm_value(wm_state->sr[level].plane,
-					    sr_fifo_size);
-		wm_state->sr[level].cursor =
-			vlv_invert_wm_value(wm_state->sr[level].cursor,
-					    63);
+	for (; level < num_levels; level++) {
+		struct vlv_pipe_wm *raw = &crtc_state->wm.vlv.raw[level];
 
-		for_each_intel_plane_on_crtc(&dev_priv->drm, crtc, plane) {
-			wm_state->wm[level].plane[plane->id] =
-				vlv_invert_wm_value(wm_state->wm[level].plane[plane->id],
-						    plane->wm.fifo_size);
-		}
+		dirty |= raw->plane[plane_id] != value;
+		raw->plane[plane_id] = value;
 	}
+
+	return dirty;
 }
 
-static void vlv_compute_wm(struct intel_crtc *crtc)
+static bool vlv_plane_wm_compute(struct intel_crtc_state *crtc_state,
+				 const struct intel_plane_state *plane_state)
 {
-	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-	struct vlv_wm_state *wm_state = &crtc->wm_state;
-	struct intel_plane *plane;
+	struct intel_plane *plane = to_intel_plane(plane_state->base.plane);
+	enum plane_id plane_id = plane->id;
+	int num_levels = vlv_num_wm_levels(to_i915(plane->base.dev));
 	int level;
+	bool dirty = false;
+
+	if (!plane_state->base.visible) {
+		dirty |= vlv_raw_plane_wm_set(crtc_state, 0, plane_id, 0);
+		goto out;
+	}
+
+	for (level = 0; level < num_levels; level++) {
+		struct vlv_pipe_wm *raw = &crtc_state->wm.vlv.raw[level];
+		int wm = vlv_compute_wm_level(crtc_state, plane_state, level);
+		int max_wm = plane_id == PLANE_CURSOR ? 63 : 511;
+
+		if (wm > max_wm)
+			break;
 
-	memset(wm_state, 0, sizeof(*wm_state));
+		dirty |= raw->plane[plane_id] != wm;
+		raw->plane[plane_id] = wm;
+	}
 
-	wm_state->cxsr = crtc->pipe != PIPE_C && crtc->wm.cxsr_allowed;
-	wm_state->num_levels = dev_priv->wm.max_level + 1;
+	/* mark all higher levels as invalid */
+	dirty |= vlv_raw_plane_wm_set(crtc_state, level, plane_id, USHRT_MAX);
 
-	wm_state->num_active_planes = 0;
+out:
+	if (dirty)
+		DRM_DEBUG_KMS("%s wms: [0]=%d,[1]=%d,[2]=%d\n",
+			      plane->base.name,
+			      crtc_state->wm.vlv.raw[VLV_WM_LEVEL_PM2].plane[plane_id],
+			      crtc_state->wm.vlv.raw[VLV_WM_LEVEL_PM5].plane[plane_id],
+			      crtc_state->wm.vlv.raw[VLV_WM_LEVEL_DDR_DVFS].plane[plane_id]);
 
-	vlv_compute_fifo(crtc);
+	return dirty;
+}
 
-	if (wm_state->num_active_planes != 1)
-		wm_state->cxsr = false;
+static bool vlv_plane_wm_is_valid(const struct intel_crtc_state *crtc_state,
+				  enum plane_id plane_id, int level)
+{
+	const struct vlv_pipe_wm *raw =
+		&crtc_state->wm.vlv.raw[level];
+	const struct vlv_fifo_state *fifo_state =
+		&crtc_state->wm.vlv.fifo_state;
 
-	for_each_intel_plane_on_crtc(&dev_priv->drm, crtc, plane) {
-		struct intel_plane_state *state =
+	return raw->plane[plane_id] <= fifo_state->plane[plane_id];
+}
+
+static bool vlv_crtc_wm_is_valid(const struct intel_crtc_state *crtc_state, int level)
+{
+	return vlv_plane_wm_is_valid(crtc_state, PLANE_PRIMARY, level) &&
+		vlv_plane_wm_is_valid(crtc_state, PLANE_SPRITE0, level) &&
+		vlv_plane_wm_is_valid(crtc_state, PLANE_SPRITE1, level) &&
+		vlv_plane_wm_is_valid(crtc_state, PLANE_CURSOR, level);
+}
+
+static int vlv_compute_pipe_wm(struct intel_crtc_state *crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_atomic_state *state =
+		to_intel_atomic_state(crtc_state->base.state);
+	struct vlv_wm_state *wm_state = &crtc_state->wm.vlv.optimal;
+	const struct vlv_fifo_state *fifo_state =
+		&crtc_state->wm.vlv.fifo_state;
+	int num_active_planes = hweight32(crtc_state->active_planes &
+					  ~BIT(PLANE_CURSOR));
+	bool needs_modeset = drm_atomic_crtc_needs_modeset(&crtc_state->base);
+	struct intel_plane_state *plane_state;
+	struct intel_plane *plane;
+	enum plane_id plane_id;
+	int level, ret, i;
+	unsigned int dirty = 0;
+
+	for_each_intel_plane_in_state(state, plane, plane_state, i) {
+		const struct intel_plane_state *old_plane_state =
 			to_intel_plane_state(plane->base.state);
-		int level;
 
-		if (!state->base.visible)
+		if (plane_state->base.crtc != &crtc->base &&
+		    old_plane_state->base.crtc != &crtc->base)
 			continue;
 
-		/* normal watermarks */
-		for (level = 0; level < wm_state->num_levels; level++) {
-			int wm = vlv_compute_wm_level(crtc->config, state, level);
-			int max_wm = plane->wm.fifo_size;
+		if (vlv_plane_wm_compute(crtc_state, plane_state))
+			dirty |= BIT(plane->id);
+	}
 
-			/* hack */
-			if (WARN_ON(level == 0 && wm > max_wm))
-				wm = max_wm;
+	/*
+	 * DSPARB registers may have been reset due to the
+	 * power well being turned off. Make sure we restore
+	 * them to a consistent state even if no primary/sprite
+	 * planes are initially active.
+	 */
+	if (needs_modeset)
+		crtc_state->fifo_changed = true;
 
-			if (wm > max_wm)
-				break;
+	if (!dirty)
+		return 0;
 
-			wm_state->wm[level].plane[plane->id] = wm;
-		}
+	/* cursor changes don't warrant a FIFO recompute */
+	if (dirty & ~BIT(PLANE_CURSOR)) {
+		const struct intel_crtc_state *old_crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		const struct vlv_fifo_state *old_fifo_state =
+			&old_crtc_state->wm.vlv.fifo_state;
 
-		wm_state->num_levels = level;
+		ret = vlv_compute_fifo(crtc_state);
+		if (ret)
+			return ret;
 
-		if (!wm_state->cxsr)
-			continue;
+		if (needs_modeset ||
+		    memcmp(old_fifo_state, fifo_state,
+			   sizeof(*fifo_state)) != 0)
+			crtc_state->fifo_changed = true;
+	}
 
-		/* maxfifo watermarks */
-		if (plane->id == PLANE_CURSOR) {
-			for (level = 0; level < wm_state->num_levels; level++)
-				wm_state->sr[level].cursor =
-					wm_state->wm[level].plane[PLANE_CURSOR];
-		} else {
-			for (level = 0; level < wm_state->num_levels; level++)
-				wm_state->sr[level].plane =
-					max(wm_state->sr[level].plane,
-					    wm_state->wm[level].plane[plane->id]);
+	/* initially allow all levels */
+	wm_state->num_levels = vlv_num_wm_levels(dev_priv);
+	/*
+	 * Note that enabling cxsr with no primary/sprite planes
+	 * enabled can wedge the pipe. Hence we only allow cxsr
+	 * with exactly one enabled primary/sprite plane.
+	 */
+	wm_state->cxsr = crtc->pipe != PIPE_C && num_active_planes == 1;
+
+	for (level = 0; level < wm_state->num_levels; level++) {
+		const struct vlv_pipe_wm *raw = &crtc_state->wm.vlv.raw[level];
+		const int sr_fifo_size = INTEL_INFO(dev_priv)->num_pipes * 512 - 1;
+
+		if (!vlv_crtc_wm_is_valid(crtc_state, level))
+			break;
+
+		for_each_plane_id_on_crtc(crtc, plane_id) {
+			wm_state->wm[level].plane[plane_id] =
+				vlv_invert_wm_value(raw->plane[plane_id],
+						    fifo_state->plane[plane_id]);
 		}
-	}
 
-	/* clear any (partially) filled invalid levels */
-	for (level = wm_state->num_levels; level < dev_priv->wm.max_level + 1; level++) {
-		memset(&wm_state->wm[level], 0, sizeof(wm_state->wm[level]));
-		memset(&wm_state->sr[level], 0, sizeof(wm_state->sr[level]));
+		wm_state->sr[level].plane =
+			vlv_invert_wm_value(max3(raw->plane[PLANE_PRIMARY],
+						 raw->plane[PLANE_SPRITE0],
+						 raw->plane[PLANE_SPRITE1]),
+					    sr_fifo_size);
+
+		wm_state->sr[level].cursor =
+			vlv_invert_wm_value(raw->plane[PLANE_CURSOR],
+					    63);
 	}
 
-	vlv_invert_wms(crtc);
+	if (level == 0)
+		return -EINVAL;
+
+	/* limit to only levels we can actually handle */
+	wm_state->num_levels = level;
+
+	/* invalidate the higher levels */
+	vlv_invalidate_wms(crtc, wm_state, level);
+
+	return 0;
 }
 
 #define VLV_FIFO(plane, value) \
 	(((value) << DSPARB_ ## plane ## _SHIFT_VLV) & DSPARB_ ## plane ## _MASK_VLV)
 
-static void vlv_pipe_set_fifo_size(struct intel_crtc *crtc)
+static void vlv_atomic_update_fifo(struct intel_atomic_state *state,
+				   struct intel_crtc_state *crtc_state)
 {
-	struct drm_device *dev = crtc->base.dev;
-	struct drm_i915_private *dev_priv = to_i915(dev);
-	struct intel_plane *plane;
-	int sprite0_start = 0, sprite1_start = 0, fifo_size = 0;
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	const struct vlv_fifo_state *fifo_state =
+		&crtc_state->wm.vlv.fifo_state;
+	int sprite0_start, sprite1_start, fifo_size;
 
-	for_each_intel_plane_on_crtc(dev, crtc, plane) {
-		switch (plane->id) {
-		case PLANE_PRIMARY:
-			sprite0_start = plane->wm.fifo_size;
-			break;
-		case PLANE_SPRITE0:
-			sprite1_start = sprite0_start + plane->wm.fifo_size;
-			break;
-		case PLANE_SPRITE1:
-			fifo_size = sprite1_start + plane->wm.fifo_size;
-			break;
-		case PLANE_CURSOR:
-			WARN_ON(plane->wm.fifo_size != 63);
-			break;
-		default:
-			MISSING_CASE(plane->id);
-			break;
-		}
-	}
+	if (!crtc_state->fifo_changed)
+		return;
+
+	sprite0_start = fifo_state->plane[PLANE_PRIMARY];
+	sprite1_start = fifo_state->plane[PLANE_SPRITE0] + sprite0_start;
+	fifo_size = fifo_state->plane[PLANE_SPRITE1] + sprite1_start;
 
-	WARN_ON(fifo_size != 512 - 1);
+	WARN_ON(fifo_state->plane[PLANE_CURSOR] != 63);
+	WARN_ON(fifo_size != 511);
 
-	DRM_DEBUG_KMS("Pipe %c FIFO split %d / %d / %d\n",
-		      pipe_name(crtc->pipe), sprite0_start,
-		      sprite1_start, fifo_size);
+	trace_vlv_fifo_size(crtc, sprite0_start, sprite1_start, fifo_size);
 
 	spin_lock(&dev_priv->wm.dsparb_lock);
 
@@ -1289,6 +1424,46 @@ static void vlv_pipe_set_fifo_size(struct intel_crtc *crtc)
 
 #undef VLV_FIFO
 
+static int vlv_compute_intermediate_wm(struct drm_device *dev,
+				       struct intel_crtc *crtc,
+				       struct intel_crtc_state *crtc_state)
+{
+	struct vlv_wm_state *intermediate = &crtc_state->wm.vlv.intermediate;
+	const struct vlv_wm_state *optimal = &crtc_state->wm.vlv.optimal;
+	const struct vlv_wm_state *active = &crtc->wm.active.vlv;
+	int level;
+
+	intermediate->num_levels = min(optimal->num_levels, active->num_levels);
+	intermediate->cxsr = optimal->cxsr && active->cxsr &&
+		!crtc_state->disable_cxsr;
+
+	for (level = 0; level < intermediate->num_levels; level++) {
+		enum plane_id plane_id;
+
+		for_each_plane_id_on_crtc(crtc, plane_id) {
+			intermediate->wm[level].plane[plane_id] =
+				min(optimal->wm[level].plane[plane_id],
+				    active->wm[level].plane[plane_id]);
+		}
+
+		intermediate->sr[level].plane = min(optimal->sr[level].plane,
+						    active->sr[level].plane);
+		intermediate->sr[level].cursor = min(optimal->sr[level].cursor,
+						     active->sr[level].cursor);
+	}
+
+	vlv_invalidate_wms(crtc, intermediate, level);
+
+	/*
+	 * If our intermediate WM are identical to the final WM, then we can
+	 * omit the post-vblank programming; only update if it's different.
+	 */
+	if (memcmp(intermediate, optimal, sizeof(*intermediate)) != 0)
+		crtc_state->wm.need_postvbl_update = true;
+
+	return 0;
+}
+
 static void vlv_merge_wm(struct drm_i915_private *dev_priv,
 			 struct vlv_wm_values *wm)
 {
@@ -1299,7 +1474,7 @@ static void vlv_merge_wm(struct drm_i915_private *dev_priv,
 	wm->cxsr = true;
 
 	for_each_intel_crtc(&dev_priv->drm, crtc) {
-		const struct vlv_wm_state *wm_state = &crtc->wm_state;
+		const struct vlv_wm_state *wm_state = &crtc->wm.active.vlv;
 
 		if (!crtc->active)
 			continue;
@@ -1318,14 +1493,11 @@ static void vlv_merge_wm(struct drm_i915_private *dev_priv,
 		wm->level = VLV_WM_LEVEL_PM2;
 
 	for_each_intel_crtc(&dev_priv->drm, crtc) {
-		struct vlv_wm_state *wm_state = &crtc->wm_state;
+		const struct vlv_wm_state *wm_state = &crtc->wm.active.vlv;
 		enum pipe pipe = crtc->pipe;
 
-		if (!crtc->active)
-			continue;
-
 		wm->pipe[pipe] = wm_state->wm[wm->level];
-		if (wm->cxsr)
+		if (crtc->active && wm->cxsr)
 			wm->sr = wm_state->sr[wm->level];
 
 		wm->ddl[pipe].plane[PLANE_PRIMARY] = DDL_PRECISION_HIGH | 2;
@@ -1345,22 +1517,15 @@ static bool is_enabling(int old, int new, int threshold)
 	return old < threshold && new >= threshold;
 }
 
-static void vlv_update_wm(struct intel_crtc *crtc)
+static void vlv_program_watermarks(struct drm_i915_private *dev_priv)
 {
-	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
-	enum pipe pipe = crtc->pipe;
 	struct vlv_wm_values *old_wm = &dev_priv->wm.vlv;
 	struct vlv_wm_values new_wm = {};
 
-	vlv_compute_wm(crtc);
 	vlv_merge_wm(dev_priv, &new_wm);
 
-	if (memcmp(old_wm, &new_wm, sizeof(new_wm)) == 0) {
-		/* FIXME should be part of crtc atomic commit */
-		vlv_pipe_set_fifo_size(crtc);
-
+	if (memcmp(old_wm, &new_wm, sizeof(new_wm)) == 0)
 		return;
-	}
 
 	if (is_disabling(old_wm->level, new_wm.level, VLV_WM_LEVEL_DDR_DVFS))
 		chv_set_memory_dvfs(dev_priv, false);
@@ -1371,17 +1536,8 @@ static void vlv_update_wm(struct intel_crtc *crtc)
 	if (is_disabling(old_wm->cxsr, new_wm.cxsr, true))
 		_intel_set_memory_cxsr(dev_priv, false);
 
-	/* FIXME should be part of crtc atomic commit */
-	vlv_pipe_set_fifo_size(crtc);
-
 	vlv_write_wm_values(dev_priv, &new_wm);
 
-	DRM_DEBUG_KMS("Setting FIFO watermarks - %c: plane=%d, cursor=%d, "
-		      "sprite0=%d, sprite1=%d, SR: plane=%d, cursor=%d level=%d cxsr=%d\n",
-		      pipe_name(pipe), new_wm.pipe[pipe].plane[PLANE_PRIMARY], new_wm.pipe[pipe].plane[PLANE_CURSOR],
-		      new_wm.pipe[pipe].plane[PLANE_SPRITE0], new_wm.pipe[pipe].plane[PLANE_SPRITE1],
-		      new_wm.sr.plane, new_wm.sr.cursor, new_wm.level, new_wm.cxsr);
-
 	if (is_enabling(old_wm->cxsr, new_wm.cxsr, true))
 		_intel_set_memory_cxsr(dev_priv, true);
 
@@ -1394,6 +1550,33 @@ static void vlv_update_wm(struct intel_crtc *crtc)
 	*old_wm = new_wm;
 }
 
+static void vlv_initial_watermarks(struct intel_atomic_state *state,
+				   struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev);
+	struct intel_crtc *crtc = to_intel_crtc(crtc_state->base.crtc);
+
+	mutex_lock(&dev_priv->wm.wm_mutex);
+	crtc->wm.active.vlv = crtc_state->wm.vlv.intermediate;
+	vlv_program_watermarks(dev_priv);
+	mutex_unlock(&dev_priv->wm.wm_mutex);
+}
+
+static void vlv_optimize_watermarks(struct intel_atomic_state *state,
+				    struct intel_crtc_state *crtc_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(crtc_state->base.crtc->dev);
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc_state->base.crtc);
+
+	if (!crtc_state->wm.need_postvbl_update)
+		return;
+
+	mutex_lock(&dev_priv->wm.wm_mutex);
+	intel_crtc->wm.active.vlv = crtc_state->wm.vlv.optimal;
+	vlv_program_watermarks(dev_priv);
+	mutex_unlock(&dev_priv->wm.wm_mutex);
+}
+
 #define single_plane_enabled(mask) is_power_of_2(mask)
 
 static void g4x_update_wm(struct intel_crtc *crtc)
@@ -1701,39 +1884,6 @@ static void i845_update_wm(struct intel_crtc *unused_crtc)
 	I915_WRITE(FW_BLC, fwater_lo);
 }
 
-uint32_t ilk_pipe_pixel_rate(const struct intel_crtc_state *pipe_config)
-{
-	uint32_t pixel_rate;
-
-	pixel_rate = pipe_config->base.adjusted_mode.crtc_clock;
-
-	/* We only use IF-ID interlacing. If we ever use PF-ID we'll need to
-	 * adjust the pixel_rate here. */
-
-	if (pipe_config->pch_pfit.enabled) {
-		uint64_t pipe_w, pipe_h, pfit_w, pfit_h;
-		uint32_t pfit_size = pipe_config->pch_pfit.size;
-
-		pipe_w = pipe_config->pipe_src_w;
-		pipe_h = pipe_config->pipe_src_h;
-
-		pfit_w = (pfit_size >> 16) & 0xFFFF;
-		pfit_h = pfit_size & 0xFFFF;
-		if (pipe_w < pfit_w)
-			pipe_w = pfit_w;
-		if (pipe_h < pfit_h)
-			pipe_h = pfit_h;
-
-		if (WARN_ON(!pfit_w || !pfit_h))
-			return pixel_rate;
-
-		pixel_rate = div_u64((uint64_t) pixel_rate * pipe_w * pipe_h,
-				     pfit_w * pfit_h);
-	}
-
-	return pixel_rate;
-}
-
 /* latency must be in 0.1us units. */
 static uint32_t ilk_wm_method1(uint32_t pixel_rate, uint8_t cpp, uint32_t latency)
 {
@@ -1807,12 +1957,12 @@ static uint32_t ilk_compute_pri_wm(const struct intel_crtc_state *cstate,
 
 	cpp = pstate->base.fb->format->cpp[0];
 
-	method1 = ilk_wm_method1(ilk_pipe_pixel_rate(cstate), cpp, mem_value);
+	method1 = ilk_wm_method1(cstate->pixel_rate, cpp, mem_value);
 
 	if (!is_lp)
 		return method1;
 
-	method2 = ilk_wm_method2(ilk_pipe_pixel_rate(cstate),
+	method2 = ilk_wm_method2(cstate->pixel_rate,
 				 cstate->base.adjusted_mode.crtc_htotal,
 				 drm_rect_width(&pstate->base.dst),
 				 cpp, mem_value);
@@ -1836,8 +1986,8 @@ static uint32_t ilk_compute_spr_wm(const struct intel_crtc_state *cstate,
 
 	cpp = pstate->base.fb->format->cpp[0];
 
-	method1 = ilk_wm_method1(ilk_pipe_pixel_rate(cstate), cpp, mem_value);
-	method2 = ilk_wm_method2(ilk_pipe_pixel_rate(cstate),
+	method1 = ilk_wm_method1(cstate->pixel_rate, cpp, mem_value);
+	method2 = ilk_wm_method2(cstate->pixel_rate,
 				 cstate->base.adjusted_mode.crtc_htotal,
 				 drm_rect_width(&pstate->base.dst),
 				 cpp, mem_value);
@@ -1852,20 +2002,24 @@ static uint32_t ilk_compute_cur_wm(const struct intel_crtc_state *cstate,
 				   const struct intel_plane_state *pstate,
 				   uint32_t mem_value)
 {
+	int cpp;
+
 	/*
-	 * We treat the cursor plane as always-on for the purposes of watermark
-	 * calculation.  Until we have two-stage watermark programming merged,
-	 * this is necessary to avoid flickering.
+	 * Treat cursor with fb as always visible since cursor updates
+	 * can happen faster than the vrefresh rate, and the current
+	 * watermark code doesn't handle that correctly. Cursor updates
+	 * which set/clear the fb or change the cursor size are going
+	 * to get throttled by intel_legacy_cursor_update() to work
+	 * around this problem with the watermark code.
 	 */
-	int cpp = 4;
-	int width = pstate->base.visible ? pstate->base.crtc_w : 64;
-
-	if (!cstate->base.active)
+	if (!cstate->base.active || !pstate->base.fb)
 		return 0;
 
-	return ilk_wm_method2(ilk_pipe_pixel_rate(cstate),
+	cpp = pstate->base.fb->format->cpp[0];
+
+	return ilk_wm_method2(cstate->pixel_rate,
 			      cstate->base.adjusted_mode.crtc_htotal,
-			      width, cpp, mem_value);
+			      pstate->base.crtc_w, cpp, mem_value);
 }
 
 /* Only for WM_LP. */
@@ -2095,7 +2249,7 @@ hsw_compute_linetime_wm(const struct intel_crtc_state *cstate)
 		return 0;
 	if (WARN_ON(adjusted_mode->crtc_clock == 0))
 		return 0;
-	if (WARN_ON(intel_state->cdclk == 0))
+	if (WARN_ON(intel_state->cdclk.logical.cdclk == 0))
 		return 0;
 
 	/* The WM are computed with base on how long it takes to fill a single
@@ -2104,7 +2258,7 @@ hsw_compute_linetime_wm(const struct intel_crtc_state *cstate)
 	linetime = DIV_ROUND_CLOSEST(adjusted_mode->crtc_htotal * 1000 * 8,
 				     adjusted_mode->crtc_clock);
 	ips_linetime = DIV_ROUND_CLOSEST(adjusted_mode->crtc_htotal * 1000 * 8,
-					 intel_state->cdclk);
+					 intel_state->cdclk.logical.cdclk);
 
 	return PIPE_WM_LINETIME_IPS_LINETIME(ips_linetime) |
 	       PIPE_WM_LINETIME_TIME(linetime);
@@ -2173,7 +2327,7 @@ static void intel_read_wm_latency(struct drm_i915_private *dev_priv,
 		}
 
 		/*
-		 * WaWmMemoryReadLatency:skl
+		 * WaWmMemoryReadLatency:skl,glk
 		 *
 		 * punit doesn't take into account the read latency so we need
 		 * to add 2us to the various latency levels we retrieve from the
@@ -2498,8 +2652,8 @@ static int ilk_compute_intermediate_wm(struct drm_device *dev,
 	 * If our intermediate WM are identical to the final WM, then we can
 	 * omit the post-vblank programming; only update if it's different.
 	 */
-	if (memcmp(a, &newstate->wm.ilk.optimal, sizeof(*a)) == 0)
-		newstate->wm.need_postvbl_update = false;
+	if (memcmp(a, &newstate->wm.ilk.optimal, sizeof(*a)) != 0)
+		newstate->wm.need_postvbl_update = true;
 
 	return 0;
 }
@@ -2895,8 +3049,7 @@ static bool skl_needs_memory_bw_wa(struct intel_atomic_state *state)
 {
 	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 
-	if (IS_SKYLAKE(dev_priv) || IS_BROXTON(dev_priv) ||
-	    IS_KABYLAKE(dev_priv))
+	if (IS_GEN9_BC(dev_priv) || IS_BROXTON(dev_priv))
 		return true;
 
 	return false;
@@ -3547,7 +3700,7 @@ static uint32_t skl_adjusted_plane_pixel_rate(const struct intel_crtc_state *cst
 	 * Adjusted plane pixel rate is just the pipe's adjusted pixel rate
 	 * with additional adjustments for plane-specific scaling.
 	 */
-	adjusted_pixel_rate = ilk_pipe_pixel_rate(cstate);
+	adjusted_pixel_rate = cstate->pixel_rate;
 	downscale_amount = skl_plane_downscale_amount(pstate);
 
 	pixel_rate = adjusted_pixel_rate * downscale_amount >> 16;
@@ -3775,7 +3928,7 @@ skl_compute_linetime_wm(struct intel_crtc_state *cstate)
 	if (!cstate->base.active)
 		return 0;
 
-	pixel_rate = ilk_pipe_pixel_rate(cstate);
+	pixel_rate = cstate->pixel_rate;
 
 	if (WARN_ON(pixel_rate == 0))
 		return 0;
@@ -4539,15 +4692,11 @@ void vlv_wm_get_hw_state(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct vlv_wm_values *wm = &dev_priv->wm.vlv;
-	struct intel_plane *plane;
-	enum pipe pipe;
+	struct intel_crtc *crtc;
 	u32 val;
 
 	vlv_read_wm_values(dev_priv, wm);
 
-	for_each_intel_plane(dev, plane)
-		plane->wm.fifo_size = vlv_get_fifo_size(plane);
-
 	wm->cxsr = I915_READ(FW_BLC_SELF_VLV) & FW_CSPWRDWNEN;
 	wm->level = VLV_WM_LEVEL_PM2;
 
@@ -4585,18 +4734,107 @@ void vlv_wm_get_hw_state(struct drm_device *dev)
 		mutex_unlock(&dev_priv->rps.hw_lock);
 	}
 
-	for_each_pipe(dev_priv, pipe)
+	for_each_intel_crtc(dev, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		struct vlv_wm_state *active = &crtc->wm.active.vlv;
+		const struct vlv_fifo_state *fifo_state =
+			&crtc_state->wm.vlv.fifo_state;
+		enum pipe pipe = crtc->pipe;
+		enum plane_id plane_id;
+		int level;
+
+		vlv_get_fifo_size(crtc_state);
+
+		active->num_levels = wm->level + 1;
+		active->cxsr = wm->cxsr;
+
+		for (level = 0; level < active->num_levels; level++) {
+			struct vlv_pipe_wm *raw =
+				&crtc_state->wm.vlv.raw[level];
+
+			active->sr[level].plane = wm->sr.plane;
+			active->sr[level].cursor = wm->sr.cursor;
+
+			for_each_plane_id_on_crtc(crtc, plane_id) {
+				active->wm[level].plane[plane_id] =
+					wm->pipe[pipe].plane[plane_id];
+
+				raw->plane[plane_id] =
+					vlv_invert_wm_value(active->wm[level].plane[plane_id],
+							    fifo_state->plane[plane_id]);
+			}
+		}
+
+		for_each_plane_id_on_crtc(crtc, plane_id)
+			vlv_raw_plane_wm_set(crtc_state, level,
+					     plane_id, USHRT_MAX);
+		vlv_invalidate_wms(crtc, active, level);
+
+		crtc_state->wm.vlv.optimal = *active;
+		crtc_state->wm.vlv.intermediate = *active;
+
 		DRM_DEBUG_KMS("Initial watermarks: pipe %c, plane=%d, cursor=%d, sprite0=%d, sprite1=%d\n",
 			      pipe_name(pipe),
 			      wm->pipe[pipe].plane[PLANE_PRIMARY],
 			      wm->pipe[pipe].plane[PLANE_CURSOR],
 			      wm->pipe[pipe].plane[PLANE_SPRITE0],
 			      wm->pipe[pipe].plane[PLANE_SPRITE1]);
+	}
 
 	DRM_DEBUG_KMS("Initial watermarks: SR plane=%d, SR cursor=%d level=%d cxsr=%d\n",
 		      wm->sr.plane, wm->sr.cursor, wm->level, wm->cxsr);
 }
 
+void vlv_wm_sanitize(struct drm_i915_private *dev_priv)
+{
+	struct intel_plane *plane;
+	struct intel_crtc *crtc;
+
+	mutex_lock(&dev_priv->wm.wm_mutex);
+
+	for_each_intel_plane(&dev_priv->drm, plane) {
+		struct intel_crtc *crtc =
+			intel_get_crtc_for_pipe(dev_priv, plane->pipe);
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+		struct intel_plane_state *plane_state =
+			to_intel_plane_state(plane->base.state);
+		struct vlv_wm_state *wm_state = &crtc_state->wm.vlv.optimal;
+		const struct vlv_fifo_state *fifo_state =
+			&crtc_state->wm.vlv.fifo_state;
+		enum plane_id plane_id = plane->id;
+		int level;
+
+		if (plane_state->base.visible)
+			continue;
+
+		for (level = 0; level < wm_state->num_levels; level++) {
+			struct vlv_pipe_wm *raw =
+				&crtc_state->wm.vlv.raw[level];
+
+			raw->plane[plane_id] = 0;
+
+			wm_state->wm[level].plane[plane_id] =
+				vlv_invert_wm_value(raw->plane[plane_id],
+						    fifo_state->plane[plane_id]);
+		}
+	}
+
+	for_each_intel_crtc(&dev_priv->drm, crtc) {
+		struct intel_crtc_state *crtc_state =
+			to_intel_crtc_state(crtc->base.state);
+
+		crtc_state->wm.vlv.intermediate =
+			crtc_state->wm.vlv.optimal;
+		crtc->wm.active.vlv = crtc_state->wm.vlv.optimal;
+	}
+
+	vlv_program_watermarks(dev_priv);
+
+	mutex_unlock(&dev_priv->wm.wm_mutex);
+}
+
 void ilk_wm_get_hw_state(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = to_i915(dev);
@@ -4680,7 +4918,7 @@ bool ironlake_set_drps(struct drm_i915_private *dev_priv, u8 val)
 {
 	u16 rgvswctl;
 
-	assert_spin_locked(&mchdev_lock);
+	lockdep_assert_held(&mchdev_lock);
 
 	rgvswctl = I915_READ16(MEMSWCTL);
 	if (rgvswctl & MEMCTL_CMD_STS) {
@@ -4891,6 +5129,12 @@ static void gen6_set_rps_thresholds(struct drm_i915_private *dev_priv, u8 val)
 		break;
 	}
 
+	/* When byt can survive without system hang with dynamic
+	 * sw freq adjustments, this restriction can be lifted.
+	 */
+	if (IS_VALLEYVIEW(dev_priv))
+		goto skip_hw_write;
+
 	I915_WRITE(GEN6_RP_UP_EI,
 		   GT_INTERVAL_FROM_US(dev_priv, ei_up));
 	I915_WRITE(GEN6_RP_UP_THRESHOLD,
@@ -4911,6 +5155,7 @@ static void gen6_set_rps_thresholds(struct drm_i915_private *dev_priv, u8 val)
 		   GEN6_RP_UP_BUSY_AVG |
 		   GEN6_RP_DOWN_IDLE_AVG);
 
+skip_hw_write:
 	dev_priv->rps.power = new_power;
 	dev_priv->rps.up_threshold = threshold_up;
 	dev_priv->rps.down_threshold = threshold_down;
@@ -4934,16 +5179,8 @@ static u32 gen6_rps_pm_mask(struct drm_i915_private *dev_priv, u8 val)
 /* gen6_set_rps is called to update the frequency request, but should also be
  * called when the range (min_delay and max_delay) is modified so that we can
  * update the GEN6_RP_INTERRUPT_LIMITS register accordingly. */
-static void gen6_set_rps(struct drm_i915_private *dev_priv, u8 val)
+static int gen6_set_rps(struct drm_i915_private *dev_priv, u8 val)
 {
-	/* WaGsvDisableTurbo: Workaround to disable turbo on BXT A* */
-	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1))
-		return;
-
-	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
-	WARN_ON(val > dev_priv->rps.max_freq);
-	WARN_ON(val < dev_priv->rps.min_freq);
-
 	/* min/max delay may still have been modified so be sure to
 	 * write the limits value.
 	 */
@@ -4969,17 +5206,15 @@ static void gen6_set_rps(struct drm_i915_private *dev_priv, u8 val)
 	I915_WRITE(GEN6_RP_INTERRUPT_LIMITS, intel_rps_limits(dev_priv, val));
 	I915_WRITE(GEN6_PMINTRMSK, gen6_rps_pm_mask(dev_priv, val));
 
-	POSTING_READ(GEN6_RPNSWREQ);
-
 	dev_priv->rps.cur_freq = val;
 	trace_intel_gpu_freq_change(intel_gpu_freq(dev_priv, val));
+
+	return 0;
 }
 
-static void valleyview_set_rps(struct drm_i915_private *dev_priv, u8 val)
+static int valleyview_set_rps(struct drm_i915_private *dev_priv, u8 val)
 {
-	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
-	WARN_ON(val > dev_priv->rps.max_freq);
-	WARN_ON(val < dev_priv->rps.min_freq);
+	int err;
 
 	if (WARN_ONCE(IS_CHERRYVIEW(dev_priv) && (val & 1),
 		      "Odd GPU freq value\n"))
@@ -4988,13 +5223,17 @@ static void valleyview_set_rps(struct drm_i915_private *dev_priv, u8 val)
 	I915_WRITE(GEN6_PMINTRMSK, gen6_rps_pm_mask(dev_priv, val));
 
 	if (val != dev_priv->rps.cur_freq) {
-		vlv_punit_write(dev_priv, PUNIT_REG_GPU_FREQ_REQ, val);
-		if (!IS_CHERRYVIEW(dev_priv))
-			gen6_set_rps_thresholds(dev_priv, val);
+		err = vlv_punit_write(dev_priv, PUNIT_REG_GPU_FREQ_REQ, val);
+		if (err)
+			return err;
+
+		gen6_set_rps_thresholds(dev_priv, val);
 	}
 
 	dev_priv->rps.cur_freq = val;
 	trace_intel_gpu_freq_change(intel_gpu_freq(dev_priv, val));
+
+	return 0;
 }
 
 /* vlv_set_rps_idle: Set the frequency to idle, if Gfx clocks are down
@@ -5007,6 +5246,7 @@ static void valleyview_set_rps(struct drm_i915_private *dev_priv, u8 val)
 static void vlv_set_rps_idle(struct drm_i915_private *dev_priv)
 {
 	u32 val = dev_priv->rps.idle_freq;
+	int err;
 
 	if (dev_priv->rps.cur_freq <= val)
 		return;
@@ -5024,14 +5264,19 @@ static void vlv_set_rps_idle(struct drm_i915_private *dev_priv)
 	 * power than the render powerwell.
 	 */
 	intel_uncore_forcewake_get(dev_priv, FORCEWAKE_MEDIA);
-	valleyview_set_rps(dev_priv, val);
+	err = valleyview_set_rps(dev_priv, val);
 	intel_uncore_forcewake_put(dev_priv, FORCEWAKE_MEDIA);
+
+	if (err)
+		DRM_ERROR("Failed to set RPS for idle\n");
 }
 
 void gen6_rps_busy(struct drm_i915_private *dev_priv)
 {
 	mutex_lock(&dev_priv->rps.hw_lock);
 	if (dev_priv->rps.enabled) {
+		u8 freq;
+
 		if (dev_priv->pm_rps_events & (GEN6_PM_RP_DOWN_EI_EXPIRED | GEN6_PM_RP_UP_EI_EXPIRED))
 			gen6_rps_reset_ei(dev_priv);
 		I915_WRITE(GEN6_PMINTRMSK,
@@ -5039,11 +5284,17 @@ void gen6_rps_busy(struct drm_i915_private *dev_priv)
 
 		gen6_enable_rps_interrupts(dev_priv);
 
-		/* Ensure we start at the user's desired frequency */
-		intel_set_rps(dev_priv,
-			      clamp(dev_priv->rps.cur_freq,
-				    dev_priv->rps.min_freq_softlimit,
-				    dev_priv->rps.max_freq_softlimit));
+		/* Use the user's desired frequency as a guide, but for better
+		 * performance, jump directly to RPe as our starting frequency.
+		 */
+		freq = max(dev_priv->rps.cur_freq,
+			   dev_priv->rps.efficient_freq);
+
+		if (intel_set_rps(dev_priv,
+				  clamp(freq,
+					dev_priv->rps.min_freq_softlimit,
+					dev_priv->rps.max_freq_softlimit)))
+			DRM_DEBUG_DRIVER("Failed to set idle frequency\n");
 	}
 	mutex_unlock(&dev_priv->rps.hw_lock);
 }
@@ -5111,12 +5362,25 @@ void gen6_rps_boost(struct drm_i915_private *dev_priv,
 	spin_unlock(&dev_priv->rps.client_lock);
 }
 
-void intel_set_rps(struct drm_i915_private *dev_priv, u8 val)
+int intel_set_rps(struct drm_i915_private *dev_priv, u8 val)
 {
+	int err;
+
+	lockdep_assert_held(&dev_priv->rps.hw_lock);
+	GEM_BUG_ON(val > dev_priv->rps.max_freq);
+	GEM_BUG_ON(val < dev_priv->rps.min_freq);
+
+	if (!dev_priv->rps.enabled) {
+		dev_priv->rps.cur_freq = val;
+		return 0;
+	}
+
 	if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
-		valleyview_set_rps(dev_priv, val);
+		err = valleyview_set_rps(dev_priv, val);
 	else
-		gen6_set_rps(dev_priv, val);
+		err = gen6_set_rps(dev_priv, val);
+
+	return err;
 }
 
 static void gen9_disable_rc6(struct drm_i915_private *dev_priv)
@@ -5294,7 +5558,7 @@ static void gen6_init_rps_frequencies(struct drm_i915_private *dev_priv)
 
 	dev_priv->rps.efficient_freq = dev_priv->rps.rp1_freq;
 	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv) ||
-	    IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	    IS_GEN9_BC(dev_priv)) {
 		u32 ddcc_status = 0;
 
 		if (sandybridge_pcode_read(dev_priv,
@@ -5307,7 +5571,7 @@ static void gen6_init_rps_frequencies(struct drm_i915_private *dev_priv)
 					dev_priv->rps.max_freq);
 	}
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	if (IS_GEN9_BC(dev_priv)) {
 		/* Store the frequency values in 16.66 MHZ units, which is
 		 * the natural hardware unit for SKL
 		 */
@@ -5320,7 +5584,7 @@ static void gen6_init_rps_frequencies(struct drm_i915_private *dev_priv)
 }
 
 static void reset_rps(struct drm_i915_private *dev_priv,
-		      void (*set)(struct drm_i915_private *, u8))
+		      int (*set)(struct drm_i915_private *, u8))
 {
 	u8 freq = dev_priv->rps.cur_freq;
 
@@ -5328,7 +5592,8 @@ static void reset_rps(struct drm_i915_private *dev_priv,
 	dev_priv->rps.power = -1;
 	dev_priv->rps.cur_freq = -1;
 
-	set(dev_priv, freq);
+	if (set(dev_priv, freq))
+		DRM_ERROR("Failed to reset RPS to initial values\n");
 }
 
 /* See the Gen9_GT_PM_Programming_Guide doc for the below */
@@ -5336,22 +5601,6 @@ static void gen9_enable_rps(struct drm_i915_private *dev_priv)
 {
 	intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL);
 
-	/* WaGsvDisableTurbo: Workaround to disable turbo on BXT A* */
-	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1)) {
-		/*
-		 * BIOS could leave the Hw Turbo enabled, so need to explicitly
-		 * clear out the Control register just to avoid inconsitency
-		 * with debugfs interface, which will show  Turbo as enabled
-		 * only and that is not expected by the User after adding the
-		 * WaGsvDisableTurbo. Apart from this there is no problem even
-		 * if the Turbo is left enabled in the Control register, as the
-		 * Up/Down interrupts would remain masked.
-		 */
-		gen9_disable_rps(dev_priv);
-		intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);
-		return;
-	}
-
 	/* Program defaults and thresholds for RPS*/
 	I915_WRITE(GEN6_RC_VIDEO_FREQ,
 		GEN9_FREQUENCY(dev_priv->rps.rp1_freq));
@@ -5411,18 +5660,9 @@ static void gen9_enable_rc6(struct drm_i915_private *dev_priv)
 	if (intel_enable_rc6() & INTEL_RC6_ENABLE)
 		rc6_mask = GEN6_RC_CTL_RC6_ENABLE;
 	DRM_INFO("RC6 %s\n", onoff(rc6_mask & GEN6_RC_CTL_RC6_ENABLE));
-	/* WaRsUseTimeoutMode:bxt */
-	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1)) {
-		I915_WRITE(GEN6_RC6_THRESHOLD, 625); /* 800us */
-		I915_WRITE(GEN6_RC_CONTROL, GEN6_RC_CTL_HW_ENABLE |
-			   GEN7_RC_CTL_TO_MODE |
-			   rc6_mask);
-	} else {
-		I915_WRITE(GEN6_RC6_THRESHOLD, 37500); /* 37.5/125ms per EI */
-		I915_WRITE(GEN6_RC_CONTROL, GEN6_RC_CTL_HW_ENABLE |
-			   GEN6_RC_CTL_EI_MODE(1) |
-			   rc6_mask);
-	}
+	I915_WRITE(GEN6_RC6_THRESHOLD, 37500); /* 37.5/125ms per EI */
+	I915_WRITE(GEN6_RC_CONTROL,
+		   GEN6_RC_CTL_HW_ENABLE | GEN6_RC_CTL_EI_MODE(1) | rc6_mask);
 
 	/*
 	 * 3b: Enable Coarse Power Gating only when RC6 is enabled.
@@ -5637,7 +5877,7 @@ static void gen6_update_ring_freq(struct drm_i915_private *dev_priv)
 	/* convert DDR frequency from units of 266.6MHz to bandwidth */
 	min_ring_freq = mult_frac(min_ring_freq, 8, 3);
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	if (IS_GEN9_BC(dev_priv)) {
 		/* Convert GT frequency to 50 HZ units */
 		min_gpu_freq = dev_priv->rps.min_freq / GEN9_FREQ_SCALER;
 		max_gpu_freq = dev_priv->rps.max_freq / GEN9_FREQ_SCALER;
@@ -5655,7 +5895,7 @@ static void gen6_update_ring_freq(struct drm_i915_private *dev_priv)
 		int diff = max_gpu_freq - gpu_freq;
 		unsigned int ia_freq = 0, ring_freq = 0;
 
-		if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+		if (IS_GEN9_BC(dev_priv)) {
 			/*
 			 * ring_freq = 2 * GT. ring_freq is in 100MHz units
 			 * No floor required for ring frequency on SKL.
@@ -5739,6 +5979,17 @@ static int cherryview_rps_guar_freq(struct drm_i915_private *dev_priv)
 	return rp1;
 }
 
+static u32 cherryview_rps_min_freq(struct drm_i915_private *dev_priv)
+{
+	u32 val, rpn;
+
+	val = vlv_punit_read(dev_priv, FB_GFX_FMIN_AT_VMIN_FUSE);
+	rpn = ((val >> FB_GFX_FMIN_AT_VMIN_FUSE_SHIFT) &
+		       FB_GFX_FREQ_FUSE_MASK);
+
+	return rpn;
+}
+
 static int valleyview_rps_guar_freq(struct drm_i915_private *dev_priv)
 {
 	u32 val, rp1;
@@ -5975,8 +6226,7 @@ static void cherryview_init_gt_powersave(struct drm_i915_private *dev_priv)
 			 intel_gpu_freq(dev_priv, dev_priv->rps.rp1_freq),
 			 dev_priv->rps.rp1_freq);
 
-	/* PUnit validated range is only [RPe, RP0] */
-	dev_priv->rps.min_freq = dev_priv->rps.efficient_freq;
+	dev_priv->rps.min_freq = cherryview_rps_min_freq(dev_priv);
 	DRM_DEBUG_DRIVER("min GPU freq: %d MHz (%u)\n",
 			 intel_gpu_freq(dev_priv, dev_priv->rps.min_freq),
 			 dev_priv->rps.min_freq);
@@ -6199,7 +6449,7 @@ static unsigned long __i915_chipset_val(struct drm_i915_private *dev_priv)
 	unsigned long now = jiffies_to_msecs(jiffies), diff1;
 	int i;
 
-	assert_spin_locked(&mchdev_lock);
+	lockdep_assert_held(&mchdev_lock);
 
 	diff1 = now - dev_priv->ips.last_time1;
 
@@ -6304,7 +6554,7 @@ static void __i915_update_gfx_val(struct drm_i915_private *dev_priv)
 	u64 now, diff, diffms;
 	u32 count;
 
-	assert_spin_locked(&mchdev_lock);
+	lockdep_assert_held(&mchdev_lock);
 
 	now = ktime_get_raw_ns();
 	diffms = now - dev_priv->ips.last_time2;
@@ -6349,7 +6599,7 @@ static unsigned long __i915_gfx_val(struct drm_i915_private *dev_priv)
 	unsigned long t, corr, state1, corr2, state2;
 	u32 pxvid, ext_v;
 
-	assert_spin_locked(&mchdev_lock);
+	lockdep_assert_held(&mchdev_lock);
 
 	pxvid = I915_READ(PXVFREQ(dev_priv->rps.cur_freq));
 	pxvid = (pxvid >> 24) & 0x7f;
@@ -6775,7 +7025,7 @@ void intel_enable_gt_powersave(struct drm_i915_private *dev_priv)
 	} else if (INTEL_GEN(dev_priv) >= 9) {
 		gen9_enable_rc6(dev_priv);
 		gen9_enable_rps(dev_priv);
-		if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+		if (IS_GEN9_BC(dev_priv))
 			gen6_update_ring_freq(dev_priv);
 	} else if (IS_BROADWELL(dev_priv)) {
 		gen8_enable_rps(dev_priv);
@@ -7260,6 +7510,14 @@ static void broadwell_init_clock_gating(struct drm_i915_private *dev_priv)
 		   | KVM_CONFIG_CHANGE_NOTIFICATION_SELECT);
 
 	lpt_init_clock_gating(dev_priv);
+
+	/* WaDisableDopClockGating:bdw
+	 *
+	 * Also see the CHICKEN2 write in bdw_init_workarounds() to disable DOP
+	 * clock gating.
+	 */
+	I915_WRITE(GEN6_UCGCTL1,
+		   I915_READ(GEN6_UCGCTL1) | GEN6_EU_TCUNIT_CLOCK_GATE_DISABLE);
 }
 
 static void haswell_init_clock_gating(struct drm_i915_private *dev_priv)
@@ -7656,8 +7914,10 @@ void intel_init_clock_gating_hooks(struct drm_i915_private *dev_priv)
 		dev_priv->display.init_clock_gating = skylake_init_clock_gating;
 	else if (IS_KABYLAKE(dev_priv))
 		dev_priv->display.init_clock_gating = kabylake_init_clock_gating;
-	else if (IS_GEN9_LP(dev_priv))
+	else if (IS_BROXTON(dev_priv))
 		dev_priv->display.init_clock_gating = bxt_init_clock_gating;
+	else if (IS_GEMINILAKE(dev_priv))
+		dev_priv->display.init_clock_gating = glk_init_clock_gating;
 	else if (IS_BROADWELL(dev_priv))
 		dev_priv->display.init_clock_gating = broadwell_init_clock_gating;
 	else if (IS_CHERRYVIEW(dev_priv))
@@ -7727,7 +7987,11 @@ void intel_init_pm(struct drm_i915_private *dev_priv)
 		}
 	} else if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) {
 		vlv_setup_wm_latency(dev_priv);
-		dev_priv->display.update_wm = vlv_update_wm;
+		dev_priv->display.compute_pipe_wm = vlv_compute_pipe_wm;
+		dev_priv->display.compute_intermediate_wm = vlv_compute_intermediate_wm;
+		dev_priv->display.initial_watermarks = vlv_initial_watermarks;
+		dev_priv->display.optimize_watermarks = vlv_optimize_watermarks;
+		dev_priv->display.atomic_update_watermarks = vlv_atomic_update_fifo;
 	} else if (IS_PINEVIEW(dev_priv)) {
 		if (!intel_get_cxsr_latency(IS_PINEVIEW_G(dev_priv),
 					    dev_priv->is_ddr3,
@@ -7916,10 +8180,10 @@ static bool skl_pcode_try_request(struct drm_i915_private *dev_priv, u32 mbox,
  * @timeout_base_ms: timeout for polling with preemption enabled
  *
  * Keep resending the @request to @mbox until PCODE acknowledges it, PCODE
- * reports an error or an overall timeout of @timeout_base_ms+10 ms expires.
+ * reports an error or an overall timeout of @timeout_base_ms+50 ms expires.
  * The request is acknowledged once the PCODE reply dword equals @reply after
  * applying @reply_mask. Polling is first attempted with preemption enabled
- * for @timeout_base_ms and if this times out for another 10 ms with
+ * for @timeout_base_ms and if this times out for another 50 ms with
  * preemption disabled.
  *
  * Returns 0 on success, %-ETIMEDOUT in case of a timeout, <0 in case of some
@@ -7955,14 +8219,15 @@ int skl_pcode_request(struct drm_i915_private *dev_priv, u32 mbox, u32 request,
 	 * worst case) _and_ PCODE was busy for some reason even after a
 	 * (queued) request and @timeout_base_ms delay. As a workaround retry
 	 * the poll with preemption disabled to maximize the number of
-	 * requests. Increase the timeout from @timeout_base_ms to 10ms to
+	 * requests. Increase the timeout from @timeout_base_ms to 50ms to
 	 * account for interrupts that could reduce the number of these
-	 * requests.
+	 * requests, and for any quirks of the PCODE firmware that delays
+	 * the request completion.
 	 */
 	DRM_DEBUG_KMS("PCODE timeout, retrying with preemption disabled\n");
 	WARN_ON_ONCE(timeout_base_ms > 3);
 	preempt_disable();
-	ret = wait_for_atomic(COND, 10);
+	ret = wait_for_atomic(COND, 50);
 	preempt_enable();
 
 out:
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index 91bc4abf5d3e..4a864f8c9387 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -39,7 +39,7 @@
  */
 #define LEGACY_REQUEST_SIZE 200
 
-int __intel_ring_space(int head, int tail, int size)
+static int __intel_ring_space(int head, int tail, int size)
 {
 	int space = head - tail;
 	if (space <= 0)
@@ -61,22 +61,20 @@ void intel_ring_update_space(struct intel_ring *ring)
 static int
 gen2_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 {
-	struct intel_ring *ring = req->ring;
-	u32 cmd;
-	int ret;
+	u32 cmd, *cs;
 
 	cmd = MI_FLUSH;
 
 	if (mode & EMIT_INVALIDATE)
 		cmd |= MI_READ_FLUSH;
 
-	ret = intel_ring_begin(req, 2);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, cmd);
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	*cs++ = cmd;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -84,9 +82,7 @@ gen2_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 static int
 gen4_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 {
-	struct intel_ring *ring = req->ring;
-	u32 cmd;
-	int ret;
+	u32 cmd, *cs;
 
 	/*
 	 * read/write caches:
@@ -123,13 +119,13 @@ gen4_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 			cmd |= MI_INVALIDATE_ISP;
 	}
 
-	ret = intel_ring_begin(req, 2);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, cmd);
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	*cs++ = cmd;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -174,35 +170,33 @@ gen4_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 static int
 intel_emit_post_sync_nonzero_flush(struct drm_i915_gem_request *req)
 {
-	struct intel_ring *ring = req->ring;
 	u32 scratch_addr =
 		i915_ggtt_offset(req->engine->scratch) + 2 * CACHELINE_BYTES;
-	int ret;
-
-	ret = intel_ring_begin(req, 6);
-	if (ret)
-		return ret;
-
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(5));
-	intel_ring_emit(ring, PIPE_CONTROL_CS_STALL |
-			PIPE_CONTROL_STALL_AT_SCOREBOARD);
-	intel_ring_emit(ring, scratch_addr | PIPE_CONTROL_GLOBAL_GTT);
-	intel_ring_emit(ring, 0); /* low dword */
-	intel_ring_emit(ring, 0); /* high dword */
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
-
-	ret = intel_ring_begin(req, 6);
-	if (ret)
-		return ret;
-
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(5));
-	intel_ring_emit(ring, PIPE_CONTROL_QW_WRITE);
-	intel_ring_emit(ring, scratch_addr | PIPE_CONTROL_GLOBAL_GTT);
-	intel_ring_emit(ring, 0);
-	intel_ring_emit(ring, 0);
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	u32 *cs;
+
+	cs = intel_ring_begin(req, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = GFX_OP_PIPE_CONTROL(5);
+	*cs++ = PIPE_CONTROL_CS_STALL | PIPE_CONTROL_STALL_AT_SCOREBOARD;
+	*cs++ = scratch_addr | PIPE_CONTROL_GLOBAL_GTT;
+	*cs++ = 0; /* low dword */
+	*cs++ = 0; /* high dword */
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
+
+	cs = intel_ring_begin(req, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	*cs++ = GFX_OP_PIPE_CONTROL(5);
+	*cs++ = PIPE_CONTROL_QW_WRITE;
+	*cs++ = scratch_addr | PIPE_CONTROL_GLOBAL_GTT;
+	*cs++ = 0;
+	*cs++ = 0;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -210,10 +204,9 @@ intel_emit_post_sync_nonzero_flush(struct drm_i915_gem_request *req)
 static int
 gen6_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 {
-	struct intel_ring *ring = req->ring;
 	u32 scratch_addr =
 		i915_ggtt_offset(req->engine->scratch) + 2 * CACHELINE_BYTES;
-	u32 flags = 0;
+	u32 *cs, flags = 0;
 	int ret;
 
 	/* Force SNB workarounds for PIPE_CONTROL flushes */
@@ -247,15 +240,15 @@ gen6_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 		flags |= PIPE_CONTROL_QW_WRITE | PIPE_CONTROL_CS_STALL;
 	}
 
-	ret = intel_ring_begin(req, 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(4));
-	intel_ring_emit(ring, flags);
-	intel_ring_emit(ring, scratch_addr | PIPE_CONTROL_GLOBAL_GTT);
-	intel_ring_emit(ring, 0);
-	intel_ring_advance(ring);
+	*cs++ = GFX_OP_PIPE_CONTROL(4);
+	*cs++ = flags;
+	*cs++ = scratch_addr | PIPE_CONTROL_GLOBAL_GTT;
+	*cs++ = 0;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -263,20 +256,17 @@ gen6_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 static int
 gen7_render_ring_cs_stall_wa(struct drm_i915_gem_request *req)
 {
-	struct intel_ring *ring = req->ring;
-	int ret;
+	u32 *cs;
 
-	ret = intel_ring_begin(req, 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(4));
-	intel_ring_emit(ring,
-			PIPE_CONTROL_CS_STALL |
-			PIPE_CONTROL_STALL_AT_SCOREBOARD);
-	intel_ring_emit(ring, 0);
-	intel_ring_emit(ring, 0);
-	intel_ring_advance(ring);
+	*cs++ = GFX_OP_PIPE_CONTROL(4);
+	*cs++ = PIPE_CONTROL_CS_STALL | PIPE_CONTROL_STALL_AT_SCOREBOARD;
+	*cs++ = 0;
+	*cs++ = 0;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -284,11 +274,9 @@ gen7_render_ring_cs_stall_wa(struct drm_i915_gem_request *req)
 static int
 gen7_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 {
-	struct intel_ring *ring = req->ring;
 	u32 scratch_addr =
 		i915_ggtt_offset(req->engine->scratch) + 2 * CACHELINE_BYTES;
-	u32 flags = 0;
-	int ret;
+	u32 *cs, flags = 0;
 
 	/*
 	 * Ensure that any following seqno writes only happen when the render
@@ -332,37 +320,15 @@ gen7_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 		gen7_render_ring_cs_stall_wa(req);
 	}
 
-	ret = intel_ring_begin(req, 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(4));
-	intel_ring_emit(ring, flags);
-	intel_ring_emit(ring, scratch_addr);
-	intel_ring_emit(ring, 0);
-	intel_ring_advance(ring);
-
-	return 0;
-}
-
-static int
-gen8_emit_pipe_control(struct drm_i915_gem_request *req,
-		       u32 flags, u32 scratch_addr)
-{
-	struct intel_ring *ring = req->ring;
-	int ret;
-
-	ret = intel_ring_begin(req, 6);
-	if (ret)
-		return ret;
-
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(6));
-	intel_ring_emit(ring, flags);
-	intel_ring_emit(ring, scratch_addr);
-	intel_ring_emit(ring, 0);
-	intel_ring_emit(ring, 0);
-	intel_ring_emit(ring, 0);
-	intel_ring_advance(ring);
+	*cs++ = GFX_OP_PIPE_CONTROL(4);
+	*cs++ = flags;
+	*cs++ = scratch_addr;
+	*cs++ = 0;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -370,12 +336,14 @@ gen8_emit_pipe_control(struct drm_i915_gem_request *req,
 static int
 gen8_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 {
-	u32 scratch_addr =
-		i915_ggtt_offset(req->engine->scratch) + 2 * CACHELINE_BYTES;
-	u32 flags = 0;
-	int ret;
+	u32 flags;
+	u32 *cs;
 
-	flags |= PIPE_CONTROL_CS_STALL;
+	cs = intel_ring_begin(req, mode & EMIT_INVALIDATE ? 12 : 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
+
+	flags = PIPE_CONTROL_CS_STALL;
 
 	if (mode & EMIT_FLUSH) {
 		flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
@@ -394,15 +362,19 @@ gen8_render_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 		flags |= PIPE_CONTROL_GLOBAL_GTT_IVB;
 
 		/* WaCsStallBeforeStateCacheInvalidate:bdw,chv */
-		ret = gen8_emit_pipe_control(req,
-					     PIPE_CONTROL_CS_STALL |
-					     PIPE_CONTROL_STALL_AT_SCOREBOARD,
-					     0);
-		if (ret)
-			return ret;
+		cs = gen8_emit_pipe_control(cs,
+					    PIPE_CONTROL_CS_STALL |
+					    PIPE_CONTROL_STALL_AT_SCOREBOARD,
+					    0);
 	}
 
-	return gen8_emit_pipe_control(req, flags, scratch_addr);
+	cs = gen8_emit_pipe_control(cs, flags,
+				    i915_ggtt_offset(req->engine->scratch) +
+				    2 * CACHELINE_BYTES);
+
+	intel_ring_advance(req, cs);
+
+	return 0;
 }
 
 static void ring_setup_phys_status_page(struct intel_engine_cs *engine)
@@ -657,41 +629,6 @@ static void reset_ring_common(struct intel_engine_cs *engine,
 	}
 }
 
-static int intel_ring_workarounds_emit(struct drm_i915_gem_request *req)
-{
-	struct intel_ring *ring = req->ring;
-	struct i915_workarounds *w = &req->i915->workarounds;
-	int ret, i;
-
-	if (w->count == 0)
-		return 0;
-
-	ret = req->engine->emit_flush(req, EMIT_BARRIER);
-	if (ret)
-		return ret;
-
-	ret = intel_ring_begin(req, (w->count * 2 + 2));
-	if (ret)
-		return ret;
-
-	intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(w->count));
-	for (i = 0; i < w->count; i++) {
-		intel_ring_emit_reg(ring, w->reg[i].addr);
-		intel_ring_emit(ring, w->reg[i].value);
-	}
-	intel_ring_emit(ring, MI_NOOP);
-
-	intel_ring_advance(ring);
-
-	ret = req->engine->emit_flush(req, EMIT_BARRIER);
-	if (ret)
-		return ret;
-
-	DRM_DEBUG_DRIVER("Number of Workarounds emitted: %d\n", w->count);
-
-	return 0;
-}
-
 static int intel_rcs_ctx_init(struct drm_i915_gem_request *req)
 {
 	int ret;
@@ -707,498 +644,6 @@ static int intel_rcs_ctx_init(struct drm_i915_gem_request *req)
 	return 0;
 }
 
-static int wa_add(struct drm_i915_private *dev_priv,
-		  i915_reg_t addr,
-		  const u32 mask, const u32 val)
-{
-	const u32 idx = dev_priv->workarounds.count;
-
-	if (WARN_ON(idx >= I915_MAX_WA_REGS))
-		return -ENOSPC;
-
-	dev_priv->workarounds.reg[idx].addr = addr;
-	dev_priv->workarounds.reg[idx].value = val;
-	dev_priv->workarounds.reg[idx].mask = mask;
-
-	dev_priv->workarounds.count++;
-
-	return 0;
-}
-
-#define WA_REG(addr, mask, val) do { \
-		const int r = wa_add(dev_priv, (addr), (mask), (val)); \
-		if (r) \
-			return r; \
-	} while (0)
-
-#define WA_SET_BIT_MASKED(addr, mask) \
-	WA_REG(addr, (mask), _MASKED_BIT_ENABLE(mask))
-
-#define WA_CLR_BIT_MASKED(addr, mask) \
-	WA_REG(addr, (mask), _MASKED_BIT_DISABLE(mask))
-
-#define WA_SET_FIELD_MASKED(addr, mask, value) \
-	WA_REG(addr, mask, _MASKED_FIELD(mask, value))
-
-#define WA_SET_BIT(addr, mask) WA_REG(addr, mask, I915_READ(addr) | (mask))
-#define WA_CLR_BIT(addr, mask) WA_REG(addr, mask, I915_READ(addr) & ~(mask))
-
-#define WA_WRITE(addr, val) WA_REG(addr, 0xffffffff, val)
-
-static int wa_ring_whitelist_reg(struct intel_engine_cs *engine,
-				 i915_reg_t reg)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-	struct i915_workarounds *wa = &dev_priv->workarounds;
-	const uint32_t index = wa->hw_whitelist_count[engine->id];
-
-	if (WARN_ON(index >= RING_MAX_NONPRIV_SLOTS))
-		return -EINVAL;
-
-	WA_WRITE(RING_FORCE_TO_NONPRIV(engine->mmio_base, index),
-		 i915_mmio_reg_offset(reg));
-	wa->hw_whitelist_count[engine->id]++;
-
-	return 0;
-}
-
-static int gen8_init_workarounds(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-
-	WA_SET_BIT_MASKED(INSTPM, INSTPM_FORCE_ORDERING);
-
-	/* WaDisableAsyncFlipPerfMode:bdw,chv */
-	WA_SET_BIT_MASKED(MI_MODE, ASYNC_FLIP_PERF_DISABLE);
-
-	/* WaDisablePartialInstShootdown:bdw,chv */
-	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN,
-			  PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE);
-
-	/* Use Force Non-Coherent whenever executing a 3D context. This is a
-	 * workaround for for a possible hang in the unlikely event a TLB
-	 * invalidation occurs during a PSD flush.
-	 */
-	/* WaForceEnableNonCoherent:bdw,chv */
-	/* WaHdcDisableFetchWhenMasked:bdw,chv */
-	WA_SET_BIT_MASKED(HDC_CHICKEN0,
-			  HDC_DONOT_FETCH_MEM_WHEN_MASKED |
-			  HDC_FORCE_NON_COHERENT);
-
-	/* From the Haswell PRM, Command Reference: Registers, CACHE_MODE_0:
-	 * "The Hierarchical Z RAW Stall Optimization allows non-overlapping
-	 *  polygons in the same 8x4 pixel/sample area to be processed without
-	 *  stalling waiting for the earlier ones to write to Hierarchical Z
-	 *  buffer."
-	 *
-	 * This optimization is off by default for BDW and CHV; turn it on.
-	 */
-	WA_CLR_BIT_MASKED(CACHE_MODE_0_GEN7, HIZ_RAW_STALL_OPT_DISABLE);
-
-	/* Wa4x4STCOptimizationDisable:bdw,chv */
-	WA_SET_BIT_MASKED(CACHE_MODE_1, GEN8_4x4_STC_OPTIMIZATION_DISABLE);
-
-	/*
-	 * BSpec recommends 8x4 when MSAA is used,
-	 * however in practice 16x4 seems fastest.
-	 *
-	 * Note that PS/WM thread counts depend on the WIZ hashing
-	 * disable bit, which we don't touch here, but it's good
-	 * to keep in mind (see 3DSTATE_PS and 3DSTATE_WM).
-	 */
-	WA_SET_FIELD_MASKED(GEN7_GT_MODE,
-			    GEN6_WIZ_HASHING_MASK,
-			    GEN6_WIZ_HASHING_16x4);
-
-	return 0;
-}
-
-static int bdw_init_workarounds(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-	int ret;
-
-	ret = gen8_init_workarounds(engine);
-	if (ret)
-		return ret;
-
-	/* WaDisableThreadStallDopClockGating:bdw (pre-production) */
-	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE);
-
-	/* WaDisableDopClockGating:bdw */
-	WA_SET_BIT_MASKED(GEN7_ROW_CHICKEN2,
-			  DOP_CLOCK_GATING_DISABLE);
-
-	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3,
-			  GEN8_SAMPLER_POWER_BYPASS_DIS);
-
-	WA_SET_BIT_MASKED(HDC_CHICKEN0,
-			  /* WaForceContextSaveRestoreNonCoherent:bdw */
-			  HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT |
-			  /* WaDisableFenceDestinationToSLM:bdw (pre-prod) */
-			  (IS_BDW_GT3(dev_priv) ? HDC_FENCE_DEST_SLM_DISABLE : 0));
-
-	return 0;
-}
-
-static int chv_init_workarounds(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-	int ret;
-
-	ret = gen8_init_workarounds(engine);
-	if (ret)
-		return ret;
-
-	/* WaDisableThreadStallDopClockGating:chv */
-	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN, STALL_DOP_GATING_DISABLE);
-
-	/* Improve HiZ throughput on CHV. */
-	WA_SET_BIT_MASKED(HIZ_CHICKEN, CHV_HZ_8X8_MODE_IN_1X);
-
-	return 0;
-}
-
-static int gen9_init_workarounds(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-	int ret;
-
-	/* WaConextSwitchWithConcurrentTLBInvalidate:skl,bxt,kbl */
-	I915_WRITE(GEN9_CSFE_CHICKEN1_RCS, _MASKED_BIT_ENABLE(GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE));
-
-	/* WaEnableLbsSlaRetryTimerDecrement:skl,bxt,kbl */
-	I915_WRITE(BDW_SCRATCH1, I915_READ(BDW_SCRATCH1) |
-		   GEN9_LBS_SLA_RETRY_TIMER_DECREMENT_ENABLE);
-
-	/* WaDisableKillLogic:bxt,skl,kbl */
-	I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) |
-		   ECOCHK_DIS_TLB);
-
-	/* WaClearFlowControlGpgpuContextSave:skl,bxt,kbl */
-	/* WaDisablePartialInstShootdown:skl,bxt,kbl */
-	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN,
-			  FLOW_CONTROL_ENABLE |
-			  PARTIAL_INSTRUCTION_SHOOTDOWN_DISABLE);
-
-	/* Syncing dependencies between camera and graphics:skl,bxt,kbl */
-	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3,
-			  GEN9_DISABLE_OCL_OOB_SUPPRESS_LOGIC);
-
-	/* WaDisableDgMirrorFixInHalfSliceChicken5:bxt */
-	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1))
-		WA_CLR_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN5,
-				  GEN9_DG_MIRROR_FIX_ENABLE);
-
-	/* WaSetDisablePixMaskCammingAndRhwoInCommonSliceChicken:bxt */
-	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1)) {
-		WA_SET_BIT_MASKED(GEN7_COMMON_SLICE_CHICKEN1,
-				  GEN9_RHWO_OPTIMIZATION_DISABLE);
-		/*
-		 * WA also requires GEN9_SLICE_COMMON_ECO_CHICKEN0[14:14] to be set
-		 * but we do that in per ctx batchbuffer as there is an issue
-		 * with this register not getting restored on ctx restore
-		 */
-	}
-
-	/* WaEnableSamplerGPGPUPreemptionSupport:skl,bxt,kbl */
-	WA_SET_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN7,
-			  GEN9_ENABLE_GPGPU_PREEMPTION);
-
-	/* Wa4x4STCOptimizationDisable:skl,bxt,kbl */
-	/* WaDisablePartialResolveInVc:skl,bxt,kbl */
-	WA_SET_BIT_MASKED(CACHE_MODE_1, (GEN8_4x4_STC_OPTIMIZATION_DISABLE |
-					 GEN9_PARTIAL_RESOLVE_IN_VC_DISABLE));
-
-	/* WaCcsTlbPrefetchDisable:skl,bxt,kbl */
-	WA_CLR_BIT_MASKED(GEN9_HALF_SLICE_CHICKEN5,
-			  GEN9_CCS_TLB_PREFETCH_ENABLE);
-
-	/* WaDisableMaskBasedCammingInRCC:bxt */
-	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1))
-		WA_SET_BIT_MASKED(SLICE_ECO_CHICKEN0,
-				  PIXEL_MASK_CAMMING_DISABLE);
-
-	/* WaForceContextSaveRestoreNonCoherent:skl,bxt,kbl */
-	WA_SET_BIT_MASKED(HDC_CHICKEN0,
-			  HDC_FORCE_CONTEXT_SAVE_RESTORE_NON_COHERENT |
-			  HDC_FORCE_CSR_NON_COHERENT_OVR_DISABLE);
-
-	/* WaForceEnableNonCoherent and WaDisableHDCInvalidation are
-	 * both tied to WaForceContextSaveRestoreNonCoherent
-	 * in some hsds for skl. We keep the tie for all gen9. The
-	 * documentation is a bit hazy and so we want to get common behaviour,
-	 * even though there is no clear evidence we would need both on kbl/bxt.
-	 * This area has been source of system hangs so we play it safe
-	 * and mimic the skl regardless of what bspec says.
-	 *
-	 * Use Force Non-Coherent whenever executing a 3D context. This
-	 * is a workaround for a possible hang in the unlikely event
-	 * a TLB invalidation occurs during a PSD flush.
-	 */
-
-	/* WaForceEnableNonCoherent:skl,bxt,kbl */
-	WA_SET_BIT_MASKED(HDC_CHICKEN0,
-			  HDC_FORCE_NON_COHERENT);
-
-	/* WaDisableHDCInvalidation:skl,bxt,kbl */
-	I915_WRITE(GAM_ECOCHK, I915_READ(GAM_ECOCHK) |
-		   BDW_DISABLE_HDC_INVALIDATION);
-
-	/* WaDisableSamplerPowerBypassForSOPingPong:skl,bxt,kbl */
-	if (IS_SKYLAKE(dev_priv) ||
-	    IS_KABYLAKE(dev_priv) ||
-	    IS_BXT_REVID(dev_priv, 0, BXT_REVID_B0))
-		WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN3,
-				  GEN8_SAMPLER_POWER_BYPASS_DIS);
-
-	/* WaDisableSTUnitPowerOptimization:skl,bxt,kbl */
-	WA_SET_BIT_MASKED(HALF_SLICE_CHICKEN2, GEN8_ST_PO_DISABLE);
-
-	/* WaOCLCoherentLineFlush:skl,bxt,kbl */
-	I915_WRITE(GEN8_L3SQCREG4, (I915_READ(GEN8_L3SQCREG4) |
-				    GEN8_LQSC_FLUSH_COHERENT_LINES));
-
-	/* WaVFEStateAfterPipeControlwithMediaStateClear:skl,bxt */
-	ret = wa_ring_whitelist_reg(engine, GEN9_CTX_PREEMPT_REG);
-	if (ret)
-		return ret;
-
-	/* WaEnablePreemptionGranularityControlByUMD:skl,bxt,kbl */
-	ret= wa_ring_whitelist_reg(engine, GEN8_CS_CHICKEN1);
-	if (ret)
-		return ret;
-
-	/* WaAllowUMDToModifyHDCChicken1:skl,bxt,kbl */
-	ret = wa_ring_whitelist_reg(engine, GEN8_HDC_CHICKEN1);
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
-static int skl_tune_iz_hashing(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-	u8 vals[3] = { 0, 0, 0 };
-	unsigned int i;
-
-	for (i = 0; i < 3; i++) {
-		u8 ss;
-
-		/*
-		 * Only consider slices where one, and only one, subslice has 7
-		 * EUs
-		 */
-		if (!is_power_of_2(INTEL_INFO(dev_priv)->sseu.subslice_7eu[i]))
-			continue;
-
-		/*
-		 * subslice_7eu[i] != 0 (because of the check above) and
-		 * ss_max == 4 (maximum number of subslices possible per slice)
-		 *
-		 * ->    0 <= ss <= 3;
-		 */
-		ss = ffs(INTEL_INFO(dev_priv)->sseu.subslice_7eu[i]) - 1;
-		vals[i] = 3 - ss;
-	}
-
-	if (vals[0] == 0 && vals[1] == 0 && vals[2] == 0)
-		return 0;
-
-	/* Tune IZ hashing. See intel_device_info_runtime_init() */
-	WA_SET_FIELD_MASKED(GEN7_GT_MODE,
-			    GEN9_IZ_HASHING_MASK(2) |
-			    GEN9_IZ_HASHING_MASK(1) |
-			    GEN9_IZ_HASHING_MASK(0),
-			    GEN9_IZ_HASHING(2, vals[2]) |
-			    GEN9_IZ_HASHING(1, vals[1]) |
-			    GEN9_IZ_HASHING(0, vals[0]));
-
-	return 0;
-}
-
-static int skl_init_workarounds(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-	int ret;
-
-	ret = gen9_init_workarounds(engine);
-	if (ret)
-		return ret;
-
-	/*
-	 * Actual WA is to disable percontext preemption granularity control
-	 * until D0 which is the default case so this is equivalent to
-	 * !WaDisablePerCtxtPreemptionGranularityControl:skl
-	 */
-	I915_WRITE(GEN7_FF_SLICE_CS_CHICKEN1,
-		   _MASKED_BIT_ENABLE(GEN9_FFSC_PERCTX_PREEMPT_CTRL));
-
-	/* WaEnableGapsTsvCreditFix:skl */
-	I915_WRITE(GEN8_GARBCNTL, (I915_READ(GEN8_GARBCNTL) |
-				   GEN9_GAPS_TSV_CREDIT_DISABLE));
-
-	/* WaDisableGafsUnitClkGating:skl */
-	WA_SET_BIT(GEN7_UCGCTL4, GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE);
-
-	/* WaInPlaceDecompressionHang:skl */
-	if (IS_SKL_REVID(dev_priv, SKL_REVID_H0, REVID_FOREVER))
-		WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA,
-			   GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS);
-
-	/* WaDisableLSQCROPERFforOCL:skl */
-	ret = wa_ring_whitelist_reg(engine, GEN8_L3SQCREG4);
-	if (ret)
-		return ret;
-
-	return skl_tune_iz_hashing(engine);
-}
-
-static int bxt_init_workarounds(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-	int ret;
-
-	ret = gen9_init_workarounds(engine);
-	if (ret)
-		return ret;
-
-	/* WaStoreMultiplePTEenable:bxt */
-	/* This is a requirement according to Hardware specification */
-	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1))
-		I915_WRITE(TILECTL, I915_READ(TILECTL) | TILECTL_TLBPF);
-
-	/* WaSetClckGatingDisableMedia:bxt */
-	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1)) {
-		I915_WRITE(GEN7_MISCCPCTL, (I915_READ(GEN7_MISCCPCTL) &
-					    ~GEN8_DOP_CLOCK_GATE_MEDIA_ENABLE));
-	}
-
-	/* WaDisableThreadStallDopClockGating:bxt */
-	WA_SET_BIT_MASKED(GEN8_ROW_CHICKEN,
-			  STALL_DOP_GATING_DISABLE);
-
-	/* WaDisablePooledEuLoadBalancingFix:bxt */
-	if (IS_BXT_REVID(dev_priv, BXT_REVID_B0, REVID_FOREVER)) {
-		WA_SET_BIT_MASKED(FF_SLICE_CS_CHICKEN2,
-				  GEN9_POOLED_EU_LOAD_BALANCING_FIX_DISABLE);
-	}
-
-	/* WaDisableSbeCacheDispatchPortSharing:bxt */
-	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_B0)) {
-		WA_SET_BIT_MASKED(
-			GEN7_HALF_SLICE_CHICKEN1,
-			GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
-	}
-
-	/* WaDisableObjectLevelPreemptionForTrifanOrPolygon:bxt */
-	/* WaDisableObjectLevelPreemptionForInstancedDraw:bxt */
-	/* WaDisableObjectLevelPreemtionForInstanceId:bxt */
-	/* WaDisableLSQCROPERFforOCL:bxt */
-	if (IS_BXT_REVID(dev_priv, 0, BXT_REVID_A1)) {
-		ret = wa_ring_whitelist_reg(engine, GEN9_CS_DEBUG_MODE1);
-		if (ret)
-			return ret;
-
-		ret = wa_ring_whitelist_reg(engine, GEN8_L3SQCREG4);
-		if (ret)
-			return ret;
-	}
-
-	/* WaProgramL3SqcReg1DefaultForPerf:bxt */
-	if (IS_BXT_REVID(dev_priv, BXT_REVID_B0, REVID_FOREVER))
-		I915_WRITE(GEN8_L3SQCREG1, L3_GENERAL_PRIO_CREDITS(62) |
-					   L3_HIGH_PRIO_CREDITS(2));
-
-	/* WaToEnableHwFixForPushConstHWBug:bxt */
-	if (IS_BXT_REVID(dev_priv, BXT_REVID_C0, REVID_FOREVER))
-		WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
-				  GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
-
-	/* WaInPlaceDecompressionHang:bxt */
-	if (IS_BXT_REVID(dev_priv, BXT_REVID_C0, REVID_FOREVER))
-		WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA,
-			   GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS);
-
-	return 0;
-}
-
-static int kbl_init_workarounds(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-	int ret;
-
-	ret = gen9_init_workarounds(engine);
-	if (ret)
-		return ret;
-
-	/* WaEnableGapsTsvCreditFix:kbl */
-	I915_WRITE(GEN8_GARBCNTL, (I915_READ(GEN8_GARBCNTL) |
-				   GEN9_GAPS_TSV_CREDIT_DISABLE));
-
-	/* WaDisableDynamicCreditSharing:kbl */
-	if (IS_KBL_REVID(dev_priv, 0, KBL_REVID_B0))
-		WA_SET_BIT(GAMT_CHKN_BIT_REG,
-			   GAMT_CHKN_DISABLE_DYNAMIC_CREDIT_SHARING);
-
-	/* WaDisableFenceDestinationToSLM:kbl (pre-prod) */
-	if (IS_KBL_REVID(dev_priv, KBL_REVID_A0, KBL_REVID_A0))
-		WA_SET_BIT_MASKED(HDC_CHICKEN0,
-				  HDC_FENCE_DEST_SLM_DISABLE);
-
-	/* WaToEnableHwFixForPushConstHWBug:kbl */
-	if (IS_KBL_REVID(dev_priv, KBL_REVID_C0, REVID_FOREVER))
-		WA_SET_BIT_MASKED(COMMON_SLICE_CHICKEN2,
-				  GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION);
-
-	/* WaDisableGafsUnitClkGating:kbl */
-	WA_SET_BIT(GEN7_UCGCTL4, GEN8_EU_GAUNIT_CLOCK_GATE_DISABLE);
-
-	/* WaDisableSbeCacheDispatchPortSharing:kbl */
-	WA_SET_BIT_MASKED(
-		GEN7_HALF_SLICE_CHICKEN1,
-		GEN7_SBE_SS_CACHE_DISPATCH_PORT_SHARING_DISABLE);
-
-	/* WaInPlaceDecompressionHang:kbl */
-	WA_SET_BIT(GEN9_GAMT_ECO_REG_RW_IA,
-		   GAMT_ECO_ENABLE_IN_PLACE_DECOMPRESS);
-
-	/* WaDisableLSQCROPERFforOCL:kbl */
-	ret = wa_ring_whitelist_reg(engine, GEN8_L3SQCREG4);
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
-int init_workarounds_ring(struct intel_engine_cs *engine)
-{
-	struct drm_i915_private *dev_priv = engine->i915;
-
-	WARN_ON(engine->id != RCS);
-
-	dev_priv->workarounds.count = 0;
-	dev_priv->workarounds.hw_whitelist_count[RCS] = 0;
-
-	if (IS_BROADWELL(dev_priv))
-		return bdw_init_workarounds(engine);
-
-	if (IS_CHERRYVIEW(dev_priv))
-		return chv_init_workarounds(engine);
-
-	if (IS_SKYLAKE(dev_priv))
-		return skl_init_workarounds(engine);
-
-	if (IS_BROXTON(dev_priv))
-		return bxt_init_workarounds(engine);
-
-	if (IS_KABYLAKE(dev_priv))
-		return kbl_init_workarounds(engine);
-
-	return 0;
-}
-
 static int init_render_ring(struct intel_engine_cs *engine)
 {
 	struct drm_i915_private *dev_priv = engine->i915;
@@ -1257,7 +702,7 @@ static void render_ring_cleanup(struct intel_engine_cs *engine)
 	i915_vma_unpin_and_release(&dev_priv->semaphore);
 }
 
-static u32 *gen8_rcs_signal(struct drm_i915_gem_request *req, u32 *out)
+static u32 *gen8_rcs_signal(struct drm_i915_gem_request *req, u32 *cs)
 {
 	struct drm_i915_private *dev_priv = req->i915;
 	struct intel_engine_cs *waiter;
@@ -1268,23 +713,22 @@ static u32 *gen8_rcs_signal(struct drm_i915_gem_request *req, u32 *out)
 		if (gtt_offset == MI_SEMAPHORE_SYNC_INVALID)
 			continue;
 
-		*out++ = GFX_OP_PIPE_CONTROL(6);
-		*out++ = (PIPE_CONTROL_GLOBAL_GTT_IVB |
-			  PIPE_CONTROL_QW_WRITE |
-			  PIPE_CONTROL_CS_STALL);
-		*out++ = lower_32_bits(gtt_offset);
-		*out++ = upper_32_bits(gtt_offset);
-		*out++ = req->global_seqno;
-		*out++ = 0;
-		*out++ = (MI_SEMAPHORE_SIGNAL |
-			  MI_SEMAPHORE_TARGET(waiter->hw_id));
-		*out++ = 0;
+		*cs++ = GFX_OP_PIPE_CONTROL(6);
+		*cs++ = PIPE_CONTROL_GLOBAL_GTT_IVB | PIPE_CONTROL_QW_WRITE |
+			PIPE_CONTROL_CS_STALL;
+		*cs++ = lower_32_bits(gtt_offset);
+		*cs++ = upper_32_bits(gtt_offset);
+		*cs++ = req->global_seqno;
+		*cs++ = 0;
+		*cs++ = MI_SEMAPHORE_SIGNAL |
+			MI_SEMAPHORE_TARGET(waiter->hw_id);
+		*cs++ = 0;
 	}
 
-	return out;
+	return cs;
 }
 
-static u32 *gen8_xcs_signal(struct drm_i915_gem_request *req, u32 *out)
+static u32 *gen8_xcs_signal(struct drm_i915_gem_request *req, u32 *cs)
 {
 	struct drm_i915_private *dev_priv = req->i915;
 	struct intel_engine_cs *waiter;
@@ -1295,19 +739,19 @@ static u32 *gen8_xcs_signal(struct drm_i915_gem_request *req, u32 *out)
 		if (gtt_offset == MI_SEMAPHORE_SYNC_INVALID)
 			continue;
 
-		*out++ = (MI_FLUSH_DW + 1) | MI_FLUSH_DW_OP_STOREDW;
-		*out++ = lower_32_bits(gtt_offset) | MI_FLUSH_DW_USE_GTT;
-		*out++ = upper_32_bits(gtt_offset);
-		*out++ = req->global_seqno;
-		*out++ = (MI_SEMAPHORE_SIGNAL |
-			  MI_SEMAPHORE_TARGET(waiter->hw_id));
-		*out++ = 0;
+		*cs++ = (MI_FLUSH_DW + 1) | MI_FLUSH_DW_OP_STOREDW;
+		*cs++ = lower_32_bits(gtt_offset) | MI_FLUSH_DW_USE_GTT;
+		*cs++ = upper_32_bits(gtt_offset);
+		*cs++ = req->global_seqno;
+		*cs++ = MI_SEMAPHORE_SIGNAL |
+			MI_SEMAPHORE_TARGET(waiter->hw_id);
+		*cs++ = 0;
 	}
 
-	return out;
+	return cs;
 }
 
-static u32 *gen6_signal(struct drm_i915_gem_request *req, u32 *out)
+static u32 *gen6_signal(struct drm_i915_gem_request *req, u32 *cs)
 {
 	struct drm_i915_private *dev_priv = req->i915;
 	struct intel_engine_cs *engine;
@@ -1322,16 +766,16 @@ static u32 *gen6_signal(struct drm_i915_gem_request *req, u32 *out)
 
 		mbox_reg = req->engine->semaphore.mbox.signal[engine->hw_id];
 		if (i915_mmio_reg_valid(mbox_reg)) {
-			*out++ = MI_LOAD_REGISTER_IMM(1);
-			*out++ = i915_mmio_reg_offset(mbox_reg);
-			*out++ = req->global_seqno;
+			*cs++ = MI_LOAD_REGISTER_IMM(1);
+			*cs++ = i915_mmio_reg_offset(mbox_reg);
+			*cs++ = req->global_seqno;
 			num_rings++;
 		}
 	}
 	if (num_rings & 1)
-		*out++ = MI_NOOP;
+		*cs++ = MI_NOOP;
 
-	return out;
+	return cs;
 }
 
 static void i9xx_submit_request(struct drm_i915_gem_request *request)
@@ -1340,18 +784,19 @@ static void i9xx_submit_request(struct drm_i915_gem_request *request)
 
 	i915_gem_request_submit(request);
 
+	GEM_BUG_ON(!IS_ALIGNED(request->tail, 8));
 	I915_WRITE_TAIL(request->engine, request->tail);
 }
 
-static void i9xx_emit_breadcrumb(struct drm_i915_gem_request *req,
-				 u32 *out)
+static void i9xx_emit_breadcrumb(struct drm_i915_gem_request *req, u32 *cs)
 {
-	*out++ = MI_STORE_DWORD_INDEX;
-	*out++ = I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT;
-	*out++ = req->global_seqno;
-	*out++ = MI_USER_INTERRUPT;
+	*cs++ = MI_STORE_DWORD_INDEX;
+	*cs++ = I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT;
+	*cs++ = req->global_seqno;
+	*cs++ = MI_USER_INTERRUPT;
 
-	req->tail = intel_ring_offset(req->ring, out);
+	req->tail = intel_ring_offset(req, cs);
+	GEM_BUG_ON(!IS_ALIGNED(req->tail, 8));
 }
 
 static const int i9xx_emit_breadcrumb_sz = 4;
@@ -1364,34 +809,33 @@ static const int i9xx_emit_breadcrumb_sz = 4;
  * Update the mailbox registers in the *other* rings with the current seqno.
  * This acts like a signal in the canonical semaphore.
  */
-static void gen6_sema_emit_breadcrumb(struct drm_i915_gem_request *req,
-				      u32 *out)
+static void gen6_sema_emit_breadcrumb(struct drm_i915_gem_request *req, u32 *cs)
 {
 	return i9xx_emit_breadcrumb(req,
-				    req->engine->semaphore.signal(req, out));
+				    req->engine->semaphore.signal(req, cs));
 }
 
 static void gen8_render_emit_breadcrumb(struct drm_i915_gem_request *req,
-					u32 *out)
+					u32 *cs)
 {
 	struct intel_engine_cs *engine = req->engine;
 
 	if (engine->semaphore.signal)
-		out = engine->semaphore.signal(req, out);
-
-	*out++ = GFX_OP_PIPE_CONTROL(6);
-	*out++ = (PIPE_CONTROL_GLOBAL_GTT_IVB |
-			       PIPE_CONTROL_CS_STALL |
-			       PIPE_CONTROL_QW_WRITE);
-	*out++ = intel_hws_seqno_address(engine);
-	*out++ = 0;
-	*out++ = req->global_seqno;
+		cs = engine->semaphore.signal(req, cs);
+
+	*cs++ = GFX_OP_PIPE_CONTROL(6);
+	*cs++ = PIPE_CONTROL_GLOBAL_GTT_IVB | PIPE_CONTROL_CS_STALL |
+		PIPE_CONTROL_QW_WRITE;
+	*cs++ = intel_hws_seqno_address(engine);
+	*cs++ = 0;
+	*cs++ = req->global_seqno;
 	/* We're thrashing one dword of HWS. */
-	*out++ = 0;
-	*out++ = MI_USER_INTERRUPT;
-	*out++ = MI_NOOP;
+	*cs++ = 0;
+	*cs++ = MI_USER_INTERRUPT;
+	*cs++ = MI_NOOP;
 
-	req->tail = intel_ring_offset(req->ring, out);
+	req->tail = intel_ring_offset(req, cs);
+	GEM_BUG_ON(!IS_ALIGNED(req->tail, 8));
 }
 
 static const int gen8_render_emit_breadcrumb_sz = 8;
@@ -1408,24 +852,21 @@ static int
 gen8_ring_sync_to(struct drm_i915_gem_request *req,
 		  struct drm_i915_gem_request *signal)
 {
-	struct intel_ring *ring = req->ring;
 	struct drm_i915_private *dev_priv = req->i915;
 	u64 offset = GEN8_WAIT_OFFSET(req->engine, signal->engine->id);
 	struct i915_hw_ppgtt *ppgtt;
-	int ret;
+	u32 *cs;
 
-	ret = intel_ring_begin(req, 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring,
-			MI_SEMAPHORE_WAIT |
-			MI_SEMAPHORE_GLOBAL_GTT |
-			MI_SEMAPHORE_SAD_GTE_SDD);
-	intel_ring_emit(ring, signal->global_seqno);
-	intel_ring_emit(ring, lower_32_bits(offset));
-	intel_ring_emit(ring, upper_32_bits(offset));
-	intel_ring_advance(ring);
+	*cs++ = MI_SEMAPHORE_WAIT | MI_SEMAPHORE_GLOBAL_GTT |
+		MI_SEMAPHORE_SAD_GTE_SDD;
+	*cs++ = signal->global_seqno;
+	*cs++ = lower_32_bits(offset);
+	*cs++ = upper_32_bits(offset);
+	intel_ring_advance(req, cs);
 
 	/* When the !RCS engines idle waiting upon a semaphore, they lose their
 	 * pagetables and we must reload them before executing the batch.
@@ -1442,28 +883,27 @@ static int
 gen6_ring_sync_to(struct drm_i915_gem_request *req,
 		  struct drm_i915_gem_request *signal)
 {
-	struct intel_ring *ring = req->ring;
 	u32 dw1 = MI_SEMAPHORE_MBOX |
 		  MI_SEMAPHORE_COMPARE |
 		  MI_SEMAPHORE_REGISTER;
 	u32 wait_mbox = signal->engine->semaphore.mbox.wait[req->engine->hw_id];
-	int ret;
+	u32 *cs;
 
 	WARN_ON(wait_mbox == MI_SEMAPHORE_SYNC_INVALID);
 
-	ret = intel_ring_begin(req, 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, dw1 | wait_mbox);
+	*cs++ = dw1 | wait_mbox;
 	/* Throughout all of the GEM code, seqno passed implies our current
 	 * seqno is >= the last seqno executed. However for hardware the
 	 * comparison is strictly greater than.
 	 */
-	intel_ring_emit(ring, signal->global_seqno - 1);
-	intel_ring_emit(ring, 0);
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	*cs++ = signal->global_seqno - 1;
+	*cs++ = 0;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -1564,16 +1004,15 @@ i8xx_irq_disable(struct intel_engine_cs *engine)
 static int
 bsd_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 {
-	struct intel_ring *ring = req->ring;
-	int ret;
+	u32 *cs;
 
-	ret = intel_ring_begin(req, 2);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, MI_FLUSH);
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	*cs++ = MI_FLUSH;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 	return 0;
 }
 
@@ -1639,20 +1078,16 @@ i965_emit_bb_start(struct drm_i915_gem_request *req,
 		   u64 offset, u32 length,
 		   unsigned int dispatch_flags)
 {
-	struct intel_ring *ring = req->ring;
-	int ret;
+	u32 *cs;
 
-	ret = intel_ring_begin(req, 2);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring,
-			MI_BATCH_BUFFER_START |
-			MI_BATCH_GTT |
-			(dispatch_flags & I915_DISPATCH_SECURE ?
-			 0 : MI_BATCH_NON_SECURE_I965));
-	intel_ring_emit(ring, offset);
-	intel_ring_advance(ring);
+	*cs++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT | (dispatch_flags &
+		I915_DISPATCH_SECURE ? 0 : MI_BATCH_NON_SECURE_I965);
+	*cs++ = offset;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -1666,59 +1101,56 @@ i830_emit_bb_start(struct drm_i915_gem_request *req,
 		   u64 offset, u32 len,
 		   unsigned int dispatch_flags)
 {
-	struct intel_ring *ring = req->ring;
-	u32 cs_offset = i915_ggtt_offset(req->engine->scratch);
-	int ret;
+	u32 *cs, cs_offset = i915_ggtt_offset(req->engine->scratch);
 
-	ret = intel_ring_begin(req, 6);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 6);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	/* Evict the invalid PTE TLBs */
-	intel_ring_emit(ring, COLOR_BLT_CMD | BLT_WRITE_RGBA);
-	intel_ring_emit(ring, BLT_DEPTH_32 | BLT_ROP_COLOR_COPY | 4096);
-	intel_ring_emit(ring, I830_TLB_ENTRIES << 16 | 4); /* load each page */
-	intel_ring_emit(ring, cs_offset);
-	intel_ring_emit(ring, 0xdeadbeef);
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	*cs++ = COLOR_BLT_CMD | BLT_WRITE_RGBA;
+	*cs++ = BLT_DEPTH_32 | BLT_ROP_COLOR_COPY | 4096;
+	*cs++ = I830_TLB_ENTRIES << 16 | 4; /* load each page */
+	*cs++ = cs_offset;
+	*cs++ = 0xdeadbeef;
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	if ((dispatch_flags & I915_DISPATCH_PINNED) == 0) {
 		if (len > I830_BATCH_LIMIT)
 			return -ENOSPC;
 
-		ret = intel_ring_begin(req, 6 + 2);
-		if (ret)
-			return ret;
+		cs = intel_ring_begin(req, 6 + 2);
+		if (IS_ERR(cs))
+			return PTR_ERR(cs);
 
 		/* Blit the batch (which has now all relocs applied) to the
 		 * stable batch scratch bo area (so that the CS never
 		 * stumbles over its tlb invalidation bug) ...
 		 */
-		intel_ring_emit(ring, SRC_COPY_BLT_CMD | BLT_WRITE_RGBA);
-		intel_ring_emit(ring,
-				BLT_DEPTH_32 | BLT_ROP_SRC_COPY | 4096);
-		intel_ring_emit(ring, DIV_ROUND_UP(len, 4096) << 16 | 4096);
-		intel_ring_emit(ring, cs_offset);
-		intel_ring_emit(ring, 4096);
-		intel_ring_emit(ring, offset);
-
-		intel_ring_emit(ring, MI_FLUSH);
-		intel_ring_emit(ring, MI_NOOP);
-		intel_ring_advance(ring);
+		*cs++ = SRC_COPY_BLT_CMD | BLT_WRITE_RGBA;
+		*cs++ = BLT_DEPTH_32 | BLT_ROP_SRC_COPY | 4096;
+		*cs++ = DIV_ROUND_UP(len, 4096) << 16 | 4096;
+		*cs++ = cs_offset;
+		*cs++ = 4096;
+		*cs++ = offset;
+
+		*cs++ = MI_FLUSH;
+		*cs++ = MI_NOOP;
+		intel_ring_advance(req, cs);
 
 		/* ... and execute it. */
 		offset = cs_offset;
 	}
 
-	ret = intel_ring_begin(req, 2);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, MI_BATCH_BUFFER_START | MI_BATCH_GTT);
-	intel_ring_emit(ring, offset | (dispatch_flags & I915_DISPATCH_SECURE ?
-					0 : MI_BATCH_NON_SECURE));
-	intel_ring_advance(ring);
+	*cs++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT;
+	*cs++ = offset | (dispatch_flags & I915_DISPATCH_SECURE ? 0 :
+		MI_BATCH_NON_SECURE);
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -1728,17 +1160,16 @@ i915_emit_bb_start(struct drm_i915_gem_request *req,
 		   u64 offset, u32 len,
 		   unsigned int dispatch_flags)
 {
-	struct intel_ring *ring = req->ring;
-	int ret;
+	u32 *cs;
 
-	ret = intel_ring_begin(req, 2);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring, MI_BATCH_BUFFER_START | MI_BATCH_GTT);
-	intel_ring_emit(ring, offset | (dispatch_flags & I915_DISPATCH_SECURE ?
-					0 : MI_BATCH_NON_SECURE));
-	intel_ring_advance(ring);
+	*cs++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT;
+	*cs++ = offset | (dispatch_flags & I915_DISPATCH_SECURE ? 0 :
+		MI_BATCH_NON_SECURE);
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -1985,7 +1416,7 @@ intel_ring_free(struct intel_ring *ring)
 	kfree(ring);
 }
 
-static int context_pin(struct i915_gem_context *ctx, unsigned int flags)
+static int context_pin(struct i915_gem_context *ctx)
 {
 	struct i915_vma *vma = ctx->engine[RCS].state;
 	int ret;
@@ -2000,7 +1431,8 @@ static int context_pin(struct i915_gem_context *ctx, unsigned int flags)
 			return ret;
 	}
 
-	return i915_vma_pin(vma, 0, ctx->ggtt_alignment, PIN_GLOBAL | flags);
+	return i915_vma_pin(vma, 0, I915_GTT_MIN_ALIGNMENT,
+			    PIN_GLOBAL | PIN_HIGH);
 }
 
 static int intel_ring_context_pin(struct intel_engine_cs *engine,
@@ -2015,13 +1447,7 @@ static int intel_ring_context_pin(struct intel_engine_cs *engine,
 		return 0;
 
 	if (ce->state) {
-		unsigned int flags;
-
-		flags = 0;
-		if (i915_gem_context_is_kernel(ctx))
-			flags = PIN_HIGH;
-
-		ret = context_pin(ctx, flags);
+		ret = context_pin(ctx);
 		if (ret)
 			goto error;
 	}
@@ -2152,7 +1578,7 @@ void intel_legacy_submission_resume(struct drm_i915_private *dev_priv)
 
 static int ring_request_alloc(struct drm_i915_gem_request *request)
 {
-	int ret;
+	u32 *cs;
 
 	GEM_BUG_ON(!request->ctx->engine[request->engine->id].pin_count);
 
@@ -2165,9 +1591,9 @@ static int ring_request_alloc(struct drm_i915_gem_request *request)
 	GEM_BUG_ON(!request->engine->buffer);
 	request->ring = request->engine->buffer;
 
-	ret = intel_ring_begin(request, 0);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(request, 0);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	request->reserved_space -= LEGACY_REQUEST_SIZE;
 	return 0;
@@ -2222,7 +1648,7 @@ static int wait_for_space(struct drm_i915_gem_request *req, int bytes)
 	return 0;
 }
 
-int intel_ring_begin(struct drm_i915_gem_request *req, int num_dwords)
+u32 *intel_ring_begin(struct drm_i915_gem_request *req, int num_dwords)
 {
 	struct intel_ring *ring = req->ring;
 	int remain_actual = ring->size - ring->tail;
@@ -2230,6 +1656,7 @@ int intel_ring_begin(struct drm_i915_gem_request *req, int num_dwords)
 	int bytes = num_dwords * sizeof(u32);
 	int total_bytes, wait_bytes;
 	bool need_wrap = false;
+	u32 *cs;
 
 	total_bytes = bytes + req->reserved_space;
 
@@ -2256,7 +1683,7 @@ int intel_ring_begin(struct drm_i915_gem_request *req, int num_dwords)
 	if (wait_bytes > ring->space) {
 		int ret = wait_for_space(req, wait_bytes);
 		if (unlikely(ret))
-			return ret;
+			return ERR_PTR(ret);
 	}
 
 	if (unlikely(need_wrap)) {
@@ -2269,31 +1696,34 @@ int intel_ring_begin(struct drm_i915_gem_request *req, int num_dwords)
 		ring->space -= remain_actual;
 	}
 
+	GEM_BUG_ON(ring->tail > ring->size - bytes);
+	cs = ring->vaddr + ring->tail;
+	ring->tail += bytes;
 	ring->space -= bytes;
 	GEM_BUG_ON(ring->space < 0);
-	return 0;
+
+	return cs;
 }
 
 /* Align the ring tail to a cacheline boundary */
 int intel_ring_cacheline_align(struct drm_i915_gem_request *req)
 {
-	struct intel_ring *ring = req->ring;
 	int num_dwords =
-		(ring->tail & (CACHELINE_BYTES - 1)) / sizeof(uint32_t);
-	int ret;
+		(req->ring->tail & (CACHELINE_BYTES - 1)) / sizeof(uint32_t);
+	u32 *cs;
 
 	if (num_dwords == 0)
 		return 0;
 
 	num_dwords = CACHELINE_BYTES / sizeof(uint32_t) - num_dwords;
-	ret = intel_ring_begin(req, num_dwords);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, num_dwords);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	while (num_dwords--)
-		intel_ring_emit(ring, MI_NOOP);
+		*cs++ = MI_NOOP;
 
-	intel_ring_advance(ring);
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -2337,13 +1767,11 @@ static void gen6_bsd_submit_request(struct drm_i915_gem_request *request)
 
 static int gen6_bsd_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 {
-	struct intel_ring *ring = req->ring;
-	uint32_t cmd;
-	int ret;
+	u32 cmd, *cs;
 
-	ret = intel_ring_begin(req, 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	cmd = MI_FLUSH_DW;
 	if (INTEL_GEN(req->i915) >= 8)
@@ -2365,16 +1793,16 @@ static int gen6_bsd_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 	if (mode & EMIT_INVALIDATE)
 		cmd |= MI_INVALIDATE_TLB | MI_INVALIDATE_BSD;
 
-	intel_ring_emit(ring, cmd);
-	intel_ring_emit(ring, I915_GEM_HWS_SCRATCH_ADDR | MI_FLUSH_DW_USE_GTT);
+	*cs++ = cmd;
+	*cs++ = I915_GEM_HWS_SCRATCH_ADDR | MI_FLUSH_DW_USE_GTT;
 	if (INTEL_GEN(req->i915) >= 8) {
-		intel_ring_emit(ring, 0); /* upper addr */
-		intel_ring_emit(ring, 0); /* value */
+		*cs++ = 0; /* upper addr */
+		*cs++ = 0; /* value */
 	} else  {
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, MI_NOOP);
+		*cs++ = 0;
+		*cs++ = MI_NOOP;
 	}
-	intel_ring_advance(ring);
+	intel_ring_advance(req, cs);
 	return 0;
 }
 
@@ -2383,23 +1811,21 @@ gen8_emit_bb_start(struct drm_i915_gem_request *req,
 		   u64 offset, u32 len,
 		   unsigned int dispatch_flags)
 {
-	struct intel_ring *ring = req->ring;
 	bool ppgtt = USES_PPGTT(req->i915) &&
 			!(dispatch_flags & I915_DISPATCH_SECURE);
-	int ret;
+	u32 *cs;
 
-	ret = intel_ring_begin(req, 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	/* FIXME(BDW): Address space and security selectors. */
-	intel_ring_emit(ring, MI_BATCH_BUFFER_START_GEN8 | (ppgtt<<8) |
-			(dispatch_flags & I915_DISPATCH_RS ?
-			 MI_BATCH_RESOURCE_STREAMER : 0));
-	intel_ring_emit(ring, lower_32_bits(offset));
-	intel_ring_emit(ring, upper_32_bits(offset));
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
+	*cs++ = MI_BATCH_BUFFER_START_GEN8 | (ppgtt << 8) | (dispatch_flags &
+		I915_DISPATCH_RS ? MI_BATCH_RESOURCE_STREAMER : 0);
+	*cs++ = lower_32_bits(offset);
+	*cs++ = upper_32_bits(offset);
+	*cs++ = MI_NOOP;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -2409,22 +1835,19 @@ hsw_emit_bb_start(struct drm_i915_gem_request *req,
 		  u64 offset, u32 len,
 		  unsigned int dispatch_flags)
 {
-	struct intel_ring *ring = req->ring;
-	int ret;
+	u32 *cs;
 
-	ret = intel_ring_begin(req, 2);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring,
-			MI_BATCH_BUFFER_START |
-			(dispatch_flags & I915_DISPATCH_SECURE ?
-			 0 : MI_BATCH_PPGTT_HSW | MI_BATCH_NON_SECURE_HSW) |
-			(dispatch_flags & I915_DISPATCH_RS ?
-			 MI_BATCH_RESOURCE_STREAMER : 0));
+	*cs++ = MI_BATCH_BUFFER_START | (dispatch_flags & I915_DISPATCH_SECURE ?
+		0 : MI_BATCH_PPGTT_HSW | MI_BATCH_NON_SECURE_HSW) |
+		(dispatch_flags & I915_DISPATCH_RS ?
+		MI_BATCH_RESOURCE_STREAMER : 0);
 	/* bit0-7 is the length on GEN6+ */
-	intel_ring_emit(ring, offset);
-	intel_ring_advance(ring);
+	*cs++ = offset;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -2434,20 +1857,17 @@ gen6_emit_bb_start(struct drm_i915_gem_request *req,
 		   u64 offset, u32 len,
 		   unsigned int dispatch_flags)
 {
-	struct intel_ring *ring = req->ring;
-	int ret;
+	u32 *cs;
 
-	ret = intel_ring_begin(req, 2);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 2);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
-	intel_ring_emit(ring,
-			MI_BATCH_BUFFER_START |
-			(dispatch_flags & I915_DISPATCH_SECURE ?
-			 0 : MI_BATCH_NON_SECURE_I965));
+	*cs++ = MI_BATCH_BUFFER_START | (dispatch_flags & I915_DISPATCH_SECURE ?
+		0 : MI_BATCH_NON_SECURE_I965);
 	/* bit0-7 is the length on GEN6+ */
-	intel_ring_emit(ring, offset);
-	intel_ring_advance(ring);
+	*cs++ = offset;
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
@@ -2456,13 +1876,11 @@ gen6_emit_bb_start(struct drm_i915_gem_request *req,
 
 static int gen6_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 {
-	struct intel_ring *ring = req->ring;
-	uint32_t cmd;
-	int ret;
+	u32 cmd, *cs;
 
-	ret = intel_ring_begin(req, 4);
-	if (ret)
-		return ret;
+	cs = intel_ring_begin(req, 4);
+	if (IS_ERR(cs))
+		return PTR_ERR(cs);
 
 	cmd = MI_FLUSH_DW;
 	if (INTEL_GEN(req->i915) >= 8)
@@ -2483,17 +1901,16 @@ static int gen6_ring_flush(struct drm_i915_gem_request *req, u32 mode)
 	 */
 	if (mode & EMIT_INVALIDATE)
 		cmd |= MI_INVALIDATE_TLB;
-	intel_ring_emit(ring, cmd);
-	intel_ring_emit(ring,
-			I915_GEM_HWS_SCRATCH_ADDR | MI_FLUSH_DW_USE_GTT);
+	*cs++ = cmd;
+	*cs++ = I915_GEM_HWS_SCRATCH_ADDR | MI_FLUSH_DW_USE_GTT;
 	if (INTEL_GEN(req->i915) >= 8) {
-		intel_ring_emit(ring, 0); /* upper addr */
-		intel_ring_emit(ring, 0); /* value */
+		*cs++ = 0; /* upper addr */
+		*cs++ = 0; /* value */
 	} else  {
-		intel_ring_emit(ring, 0);
-		intel_ring_emit(ring, MI_NOOP);
+		*cs++ = 0;
+		*cs++ = MI_NOOP;
 	}
-	intel_ring_advance(ring);
+	intel_ring_advance(req, cs);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index 79c2b8d72322..0ef491df5b4e 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -5,6 +5,7 @@
 #include "i915_gem_batch_pool.h"
 #include "i915_gem_request.h"
 #include "i915_gem_timeline.h"
+#include "i915_selftest.h"
 
 #define I915_CMD_HASH_ORDER 9
 
@@ -144,6 +145,7 @@ struct intel_ring {
 
 	u32 head;
 	u32 tail;
+
 	int space;
 	int size;
 	int effective_size;
@@ -184,26 +186,26 @@ struct i915_ctx_workarounds {
 struct drm_i915_gem_request;
 struct intel_render_state;
 
+/*
+ * Engine IDs definitions.
+ * Keep instances of the same type engine together.
+ */
+enum intel_engine_id {
+	RCS = 0,
+	BCS,
+	VCS,
+	VCS2,
+#define _VCS(n) (VCS + (n))
+	VECS
+};
+
 struct intel_engine_cs {
 	struct drm_i915_private *i915;
 	const char	*name;
-	enum intel_engine_id {
-		RCS = 0,
-		BCS,
-		VCS,
-		VCS2,	/* Keep instances of the same type engine together. */
-		VECS
-	} id;
-#define _VCS(n) (VCS + (n))
+	enum intel_engine_id id;
 	unsigned int exec_id;
-	enum intel_engine_hw_id {
-		RCS_HW = 0,
-		VCS_HW,
-		BCS_HW,
-		VECS_HW,
-		VCS2_HW
-	} hw_id;
-	enum intel_engine_hw_id guc_id; /* XXX same as hw_id? */
+	unsigned int hw_id;
+	unsigned int guc_id;
 	u32		mmio_base;
 	unsigned int irq_shift;
 	struct intel_ring *buffer;
@@ -211,6 +213,11 @@ struct intel_engine_cs {
 
 	struct intel_render_state *render_state;
 
+	atomic_t irq_count;
+	unsigned long irq_posted;
+#define ENGINE_IRQ_BREADCRUMB 0
+#define ENGINE_IRQ_EXECLIST 1
+
 	/* Rather than have every client wait upon all user interrupts,
 	 * with the herd waking after every interrupt and each doing the
 	 * heavyweight seqno dance, we delegate the task (of being the
@@ -228,22 +235,22 @@ struct intel_engine_cs {
 	 * the overhead of waking that client is much preferred.
 	 */
 	struct intel_breadcrumbs {
-		struct task_struct __rcu *irq_seqno_bh; /* bh for interrupts */
-		bool irq_posted;
+		spinlock_t irq_lock; /* protects irq_*; irqsafe */
+		struct intel_wait *irq_wait; /* oldest waiter by retirement */
 
-		spinlock_t lock; /* protects the lists of requests; irqsafe */
+		spinlock_t rb_lock; /* protects the rb and wraps irq_lock */
 		struct rb_root waiters; /* sorted by retirement, priority */
 		struct rb_root signals; /* sorted by retirement */
-		struct intel_wait *first_wait; /* oldest waiter by retirement */
 		struct task_struct *signaler; /* used for fence signalling */
-		struct drm_i915_gem_request *first_signal;
+		struct drm_i915_gem_request __rcu *first_signal;
 		struct timer_list fake_irq; /* used after a missed interrupt */
 		struct timer_list hangcheck; /* detect missed interrupts */
 
-		unsigned long timeout;
+		unsigned int hangcheck_interrupts;
 
+		bool irq_armed : 1;
 		bool irq_enabled : 1;
-		bool rpm_wakelock : 1;
+		I915_SELFTEST_DECLARE(bool mock : 1);
 	} breadcrumbs;
 
 	/*
@@ -285,7 +292,7 @@ struct intel_engine_cs {
 #define I915_DISPATCH_PINNED BIT(1)
 #define I915_DISPATCH_RS     BIT(2)
 	void		(*emit_breadcrumb)(struct drm_i915_gem_request *req,
-					   u32 *out);
+					   u32 *cs);
 	int		emit_breadcrumb_sz;
 
 	/* Pass the request to the hardware queue (e.g. directly into
@@ -368,7 +375,7 @@ struct intel_engine_cs {
 		/* AKA wait() */
 		int	(*sync_to)(struct drm_i915_gem_request *req,
 				   struct drm_i915_gem_request *signal);
-		u32	*(*signal)(struct drm_i915_gem_request *req, u32 *out);
+		u32	*(*signal)(struct drm_i915_gem_request *req, u32 *cs);
 	} semaphore;
 
 	/* Execlists */
@@ -376,13 +383,11 @@ struct intel_engine_cs {
 	struct execlist_port {
 		struct drm_i915_gem_request *request;
 		unsigned int count;
+		GEM_DEBUG_DECL(u32 context_id);
 	} execlist_port[2];
 	struct rb_root execlist_queue;
 	struct rb_node *execlist_first;
 	unsigned int fw_domains;
-	bool disable_lite_restore_wa;
-	bool preempt_wa;
-	u32 ctx_desc_template;
 
 	/* Contexts are pinned whilst they are active on the GPU. The last
 	 * context executed remains active whilst the GPU is idle - the
@@ -492,21 +497,12 @@ void intel_engine_cleanup(struct intel_engine_cs *engine);
 
 void intel_legacy_submission_resume(struct drm_i915_private *dev_priv);
 
-int __must_check intel_ring_begin(struct drm_i915_gem_request *req, int n);
 int __must_check intel_ring_cacheline_align(struct drm_i915_gem_request *req);
 
-static inline void intel_ring_emit(struct intel_ring *ring, u32 data)
-{
-	*(uint32_t *)(ring->vaddr + ring->tail) = data;
-	ring->tail += 4;
-}
+u32 __must_check *intel_ring_begin(struct drm_i915_gem_request *req, int n);
 
-static inline void intel_ring_emit_reg(struct intel_ring *ring, i915_reg_t reg)
-{
-	intel_ring_emit(ring, i915_mmio_reg_offset(reg));
-}
-
-static inline void intel_ring_advance(struct intel_ring *ring)
+static inline void
+intel_ring_advance(struct drm_i915_gem_request *req, u32 *cs)
 {
 	/* Dummy function.
 	 *
@@ -516,16 +512,18 @@ static inline void intel_ring_advance(struct intel_ring *ring)
 	 * reserved for the command packet (i.e. the value passed to
 	 * intel_ring_begin()).
 	 */
+	GEM_BUG_ON((req->ring->vaddr + req->ring->tail) != cs);
 }
 
-static inline u32 intel_ring_offset(struct intel_ring *ring, void *addr)
+static inline u32
+intel_ring_offset(struct drm_i915_gem_request *req, void *addr)
 {
 	/* Don't write ring->size (equivalent to 0) as that hangs some GPUs. */
-	u32 offset = addr - ring->vaddr;
-	return offset & (ring->size - 1);
+	u32 offset = addr - req->ring->vaddr;
+	GEM_BUG_ON(offset > req->ring->size);
+	return offset & (req->ring->size - 1);
 }
 
-int __intel_ring_space(int head, int tail, int size);
 void intel_ring_update_space(struct intel_ring *ring);
 
 void intel_engine_init_global_seqno(struct intel_engine_cs *engine, u32 seqno);
@@ -558,10 +556,11 @@ static inline u32 intel_engine_last_submit(struct intel_engine_cs *engine)
 	 * wtih serialising this hint with anything, so document it as
 	 * a hint and nothing more.
 	 */
-	return READ_ONCE(engine->timeline->last_submitted_seqno);
+	return READ_ONCE(engine->timeline->seqno);
 }
 
 int init_workarounds_ring(struct intel_engine_cs *engine);
+int intel_ring_workarounds_emit(struct drm_i915_gem_request *req);
 
 void intel_engine_get_instdone(struct intel_engine_cs *engine,
 			       struct intel_instdone *instdone);
@@ -583,12 +582,51 @@ static inline u32 intel_hws_seqno_address(struct intel_engine_cs *engine)
 /* intel_breadcrumbs.c -- user interrupt bottom-half for waiters */
 int intel_engine_init_breadcrumbs(struct intel_engine_cs *engine);
 
-static inline void intel_wait_init(struct intel_wait *wait, u32 seqno)
+static inline void intel_wait_init(struct intel_wait *wait,
+				   struct drm_i915_gem_request *rq)
+{
+	wait->tsk = current;
+	wait->request = rq;
+}
+
+static inline void intel_wait_init_for_seqno(struct intel_wait *wait, u32 seqno)
 {
 	wait->tsk = current;
 	wait->seqno = seqno;
 }
 
+static inline bool intel_wait_has_seqno(const struct intel_wait *wait)
+{
+	return wait->seqno;
+}
+
+static inline bool
+intel_wait_update_seqno(struct intel_wait *wait, u32 seqno)
+{
+	wait->seqno = seqno;
+	return intel_wait_has_seqno(wait);
+}
+
+static inline bool
+intel_wait_update_request(struct intel_wait *wait,
+			  const struct drm_i915_gem_request *rq)
+{
+	return intel_wait_update_seqno(wait, i915_gem_request_global_seqno(rq));
+}
+
+static inline bool
+intel_wait_check_seqno(const struct intel_wait *wait, u32 seqno)
+{
+	return wait->seqno == seqno;
+}
+
+static inline bool
+intel_wait_check_request(const struct intel_wait *wait,
+			 const struct drm_i915_gem_request *rq)
+{
+	return intel_wait_check_seqno(wait, i915_gem_request_global_seqno(rq));
+}
+
 static inline bool intel_wait_complete(const struct intel_wait *wait)
 {
 	return RB_EMPTY_NODE(&wait->node);
@@ -599,38 +637,36 @@ bool intel_engine_add_wait(struct intel_engine_cs *engine,
 void intel_engine_remove_wait(struct intel_engine_cs *engine,
 			      struct intel_wait *wait);
 void intel_engine_enable_signaling(struct drm_i915_gem_request *request);
+void intel_engine_cancel_signaling(struct drm_i915_gem_request *request);
 
 static inline bool intel_engine_has_waiter(const struct intel_engine_cs *engine)
 {
-	return rcu_access_pointer(engine->breadcrumbs.irq_seqno_bh);
+	return READ_ONCE(engine->breadcrumbs.irq_wait);
 }
 
-static inline bool intel_engine_wakeup(const struct intel_engine_cs *engine)
+unsigned int intel_engine_wakeup(struct intel_engine_cs *engine);
+#define ENGINE_WAKEUP_WAITER BIT(0)
+#define ENGINE_WAKEUP_ASLEEP BIT(1)
+
+void __intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine);
+void intel_engine_disarm_breadcrumbs(struct intel_engine_cs *engine);
+
+void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine);
+void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine);
+bool intel_breadcrumbs_busy(struct intel_engine_cs *engine);
+
+static inline u32 *gen8_emit_pipe_control(u32 *batch, u32 flags, u32 offset)
 {
-	bool wakeup = false;
+	memset(batch, 0, 6 * sizeof(u32));
 
-	/* Note that for this not to dangerously chase a dangling pointer,
-	 * we must hold the rcu_read_lock here.
-	 *
-	 * Also note that tsk is likely to be in !TASK_RUNNING state so an
-	 * early test for tsk->state != TASK_RUNNING before wake_up_process()
-	 * is unlikely to be beneficial.
-	 */
-	if (intel_engine_has_waiter(engine)) {
-		struct task_struct *tsk;
-
-		rcu_read_lock();
-		tsk = rcu_dereference(engine->breadcrumbs.irq_seqno_bh);
-		if (tsk)
-			wakeup = wake_up_process(tsk);
-		rcu_read_unlock();
-	}
+	batch[0] = GFX_OP_PIPE_CONTROL(6);
+	batch[1] = flags;
+	batch[2] = offset;
 
-	return wakeup;
+	return batch + 6;
 }
 
-void intel_engine_reset_breadcrumbs(struct intel_engine_cs *engine);
-void intel_engine_fini_breadcrumbs(struct intel_engine_cs *engine);
-unsigned int intel_breadcrumbs_busy(struct drm_i915_private *i915);
+bool intel_engine_is_idle(struct intel_engine_cs *engine);
+bool intel_engines_are_idle(struct drm_i915_private *dev_priv);
 
 #endif /* _INTEL_RINGBUFFER_H_ */
diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
index c0b7e95b5b8e..012bc358a33a 100644
--- a/drivers/gpu/drm/i915/intel_runtime_pm.c
+++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
@@ -49,19 +49,6 @@
  * present for a given platform.
  */
 
-#define for_each_power_well(i, power_well, domain_mask, power_domains)	\
-	for (i = 0;							\
-	     i < (power_domains)->power_well_count &&			\
-		 ((power_well) = &(power_domains)->power_wells[i]);	\
-	     i++)							\
-		for_each_if ((power_well)->domains & (domain_mask))
-
-#define for_each_power_well_rev(i, power_well, domain_mask, power_domains) \
-	for (i = (power_domains)->power_well_count - 1;			 \
-	     i >= 0 && ((power_well) = &(power_domains)->power_wells[i]);\
-	     i--)							 \
-		for_each_if ((power_well)->domains & (domain_mask))
-
 bool intel_display_power_well_is_enabled(struct drm_i915_private *dev_priv,
 				    int power_well_id);
 
@@ -106,6 +93,16 @@ intel_display_power_domain_str(enum intel_display_power_domain domain)
 		return "PORT_DDI_D_LANES";
 	case POWER_DOMAIN_PORT_DDI_E_LANES:
 		return "PORT_DDI_E_LANES";
+	case POWER_DOMAIN_PORT_DDI_A_IO:
+		return "PORT_DDI_A_IO";
+	case POWER_DOMAIN_PORT_DDI_B_IO:
+		return "PORT_DDI_B_IO";
+	case POWER_DOMAIN_PORT_DDI_C_IO:
+		return "PORT_DDI_C_IO";
+	case POWER_DOMAIN_PORT_DDI_D_IO:
+		return "PORT_DDI_D_IO";
+	case POWER_DOMAIN_PORT_DDI_E_IO:
+		return "PORT_DDI_E_IO";
 	case POWER_DOMAIN_PORT_DSI:
 		return "PORT_DSI";
 	case POWER_DOMAIN_PORT_CRT:
@@ -198,19 +195,15 @@ static bool hsw_power_well_enabled(struct drm_i915_private *dev_priv,
 bool __intel_display_power_is_enabled(struct drm_i915_private *dev_priv,
 				      enum intel_display_power_domain domain)
 {
-	struct i915_power_domains *power_domains;
 	struct i915_power_well *power_well;
 	bool is_enabled;
-	int i;
 
 	if (dev_priv->pm.suspended)
 		return false;
 
-	power_domains = &dev_priv->power_domains;
-
 	is_enabled = true;
 
-	for_each_power_well_rev(i, power_well, BIT(domain), power_domains) {
+	for_each_power_domain_well_rev(dev_priv, power_well, BIT_ULL(domain)) {
 		if (power_well->always_on)
 			continue;
 
@@ -385,124 +378,121 @@ static void hsw_set_power_well(struct drm_i915_private *dev_priv,
 }
 
 #define SKL_DISPLAY_POWERWELL_2_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_TRANSCODER_A) |		\
-	BIT(POWER_DOMAIN_PIPE_B) |			\
-	BIT(POWER_DOMAIN_TRANSCODER_B) |		\
-	BIT(POWER_DOMAIN_PIPE_C) |			\
-	BIT(POWER_DOMAIN_TRANSCODER_C) |		\
-	BIT(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |		\
-	BIT(POWER_DOMAIN_PIPE_C_PANEL_FITTER) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_D_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_E_LANES) |		\
-	BIT(POWER_DOMAIN_AUX_B) |                       \
-	BIT(POWER_DOMAIN_AUX_C) |			\
-	BIT(POWER_DOMAIN_AUX_D) |			\
-	BIT(POWER_DOMAIN_AUDIO) |			\
-	BIT(POWER_DOMAIN_VGA) |				\
-	BIT(POWER_DOMAIN_INIT))
-#define SKL_DISPLAY_DDI_A_E_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PORT_DDI_A_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_E_LANES) |		\
-	BIT(POWER_DOMAIN_INIT))
-#define SKL_DISPLAY_DDI_B_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
-	BIT(POWER_DOMAIN_INIT))
-#define SKL_DISPLAY_DDI_C_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
-	BIT(POWER_DOMAIN_INIT))
-#define SKL_DISPLAY_DDI_D_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PORT_DDI_D_LANES) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_A) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_B) |			\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_B) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_C) |			\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_C) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_D_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_E_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_B) |                       \
+	BIT_ULL(POWER_DOMAIN_AUX_C) |			\
+	BIT_ULL(POWER_DOMAIN_AUX_D) |			\
+	BIT_ULL(POWER_DOMAIN_AUDIO) |			\
+	BIT_ULL(POWER_DOMAIN_VGA) |				\
+	BIT_ULL(POWER_DOMAIN_INIT))
+#define SKL_DISPLAY_DDI_IO_A_E_POWER_DOMAINS (		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_A_IO) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_E_IO) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
+#define SKL_DISPLAY_DDI_IO_B_POWER_DOMAINS (		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_IO) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
+#define SKL_DISPLAY_DDI_IO_C_POWER_DOMAINS (		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_IO) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
+#define SKL_DISPLAY_DDI_IO_D_POWER_DOMAINS (		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_D_IO) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 #define SKL_DISPLAY_DC_OFF_POWER_DOMAINS (		\
 	SKL_DISPLAY_POWERWELL_2_POWER_DOMAINS |		\
-	BIT(POWER_DOMAIN_MODESET) |			\
-	BIT(POWER_DOMAIN_AUX_A) |			\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_MODESET) |			\
+	BIT_ULL(POWER_DOMAIN_AUX_A) |			\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 #define BXT_DISPLAY_POWERWELL_2_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_TRANSCODER_A) |		\
-	BIT(POWER_DOMAIN_PIPE_B) |			\
-	BIT(POWER_DOMAIN_TRANSCODER_B) |		\
-	BIT(POWER_DOMAIN_PIPE_C) |			\
-	BIT(POWER_DOMAIN_TRANSCODER_C) |		\
-	BIT(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |		\
-	BIT(POWER_DOMAIN_PIPE_C_PANEL_FITTER) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
-	BIT(POWER_DOMAIN_AUX_B) |			\
-	BIT(POWER_DOMAIN_AUX_C) |			\
-	BIT(POWER_DOMAIN_AUDIO) |			\
-	BIT(POWER_DOMAIN_VGA) |				\
-	BIT(POWER_DOMAIN_GMBUS) |			\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_A) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_B) |			\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_B) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_C) |			\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_C) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_B) |			\
+	BIT_ULL(POWER_DOMAIN_AUX_C) |			\
+	BIT_ULL(POWER_DOMAIN_AUDIO) |			\
+	BIT_ULL(POWER_DOMAIN_VGA) |				\
+	BIT_ULL(POWER_DOMAIN_GMBUS) |			\
+	BIT_ULL(POWER_DOMAIN_INIT))
 #define BXT_DISPLAY_DC_OFF_POWER_DOMAINS (		\
 	BXT_DISPLAY_POWERWELL_2_POWER_DOMAINS |		\
-	BIT(POWER_DOMAIN_MODESET) |			\
-	BIT(POWER_DOMAIN_AUX_A) |			\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_MODESET) |			\
+	BIT_ULL(POWER_DOMAIN_AUX_A) |			\
+	BIT_ULL(POWER_DOMAIN_INIT))
 #define BXT_DPIO_CMN_A_POWER_DOMAINS (			\
-	BIT(POWER_DOMAIN_PORT_DDI_A_LANES) |		\
-	BIT(POWER_DOMAIN_AUX_A) |			\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_A_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_A) |			\
+	BIT_ULL(POWER_DOMAIN_INIT))
 #define BXT_DPIO_CMN_BC_POWER_DOMAINS (			\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
-	BIT(POWER_DOMAIN_AUX_B) |			\
-	BIT(POWER_DOMAIN_AUX_C) |			\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_B) |			\
+	BIT_ULL(POWER_DOMAIN_AUX_C) |			\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 #define GLK_DISPLAY_POWERWELL_2_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_TRANSCODER_A) |		\
-	BIT(POWER_DOMAIN_PIPE_B) |			\
-	BIT(POWER_DOMAIN_TRANSCODER_B) |		\
-	BIT(POWER_DOMAIN_PIPE_C) |			\
-	BIT(POWER_DOMAIN_TRANSCODER_C) |		\
-	BIT(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |		\
-	BIT(POWER_DOMAIN_PIPE_C_PANEL_FITTER) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
-	BIT(POWER_DOMAIN_AUX_B) |                       \
-	BIT(POWER_DOMAIN_AUX_C) |			\
-	BIT(POWER_DOMAIN_AUDIO) |			\
-	BIT(POWER_DOMAIN_VGA) |				\
-	BIT(POWER_DOMAIN_INIT))
-#define GLK_DISPLAY_DDI_A_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PORT_DDI_A_LANES) |		\
-	BIT(POWER_DOMAIN_INIT))
-#define GLK_DISPLAY_DDI_B_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
-	BIT(POWER_DOMAIN_INIT))
-#define GLK_DISPLAY_DDI_C_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_A) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_B) |			\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_B) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_C) |			\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_C) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_B) |                       \
+	BIT_ULL(POWER_DOMAIN_AUX_C) |			\
+	BIT_ULL(POWER_DOMAIN_AUDIO) |			\
+	BIT_ULL(POWER_DOMAIN_VGA) |				\
+	BIT_ULL(POWER_DOMAIN_INIT))
+#define GLK_DISPLAY_DDI_IO_A_POWER_DOMAINS (		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_A_IO))
+#define GLK_DISPLAY_DDI_IO_B_POWER_DOMAINS (		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_IO))
+#define GLK_DISPLAY_DDI_IO_C_POWER_DOMAINS (		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_IO))
 #define GLK_DPIO_CMN_A_POWER_DOMAINS (			\
-	BIT(POWER_DOMAIN_PORT_DDI_A_LANES) |		\
-	BIT(POWER_DOMAIN_AUX_A) |			\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_A_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_A) |			\
+	BIT_ULL(POWER_DOMAIN_INIT))
 #define GLK_DPIO_CMN_B_POWER_DOMAINS (			\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
-	BIT(POWER_DOMAIN_AUX_B) |			\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_B) |			\
+	BIT_ULL(POWER_DOMAIN_INIT))
 #define GLK_DPIO_CMN_C_POWER_DOMAINS (			\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
-	BIT(POWER_DOMAIN_AUX_C) |			\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_C) |			\
+	BIT_ULL(POWER_DOMAIN_INIT))
 #define GLK_DISPLAY_AUX_A_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_AUX_A) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_AUX_A) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 #define GLK_DISPLAY_AUX_B_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_AUX_B) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_AUX_B) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 #define GLK_DISPLAY_AUX_C_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_AUX_C) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_AUX_C) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 #define GLK_DISPLAY_DC_OFF_POWER_DOMAINS (		\
 	GLK_DISPLAY_POWERWELL_2_POWER_DOMAINS |		\
-	BIT(POWER_DOMAIN_MODESET) |			\
-	BIT(POWER_DOMAIN_AUX_A) |			\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_MODESET) |			\
+	BIT_ULL(POWER_DOMAIN_AUX_A) |			\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 static void assert_can_enable_dc9(struct drm_i915_private *dev_priv)
 {
@@ -732,7 +722,7 @@ gen9_sanitize_power_well_requests(struct drm_i915_private *dev_priv,
 	 * other request bits to be set, so WARN for those.
 	 */
 	if (power_well_id == SKL_DISP_PW_1 ||
-	    ((IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) &&
+	    (IS_GEN9_BC(dev_priv) &&
 	     power_well_id == SKL_DISP_PW_MISC_IO))
 		DRM_DEBUG_DRIVER("Clearing auxiliary requests for %s forced on "
 				 "by DMC\n", power_well->name);
@@ -847,14 +837,14 @@ static void skl_set_power_well(struct drm_i915_private *dev_priv,
 static void hsw_power_well_sync_hw(struct drm_i915_private *dev_priv,
 				   struct i915_power_well *power_well)
 {
-	hsw_set_power_well(dev_priv, power_well, power_well->count > 0);
-
-	/*
-	 * We're taking over the BIOS, so clear any requests made by it since
-	 * the driver is in charge now.
-	 */
-	if (I915_READ(HSW_PWR_WELL_BIOS) & HSW_PWR_WELL_ENABLE_REQUEST)
+	/* Take over the request bit if set by BIOS. */
+	if (I915_READ(HSW_PWR_WELL_BIOS) & HSW_PWR_WELL_ENABLE_REQUEST) {
+		if (!(I915_READ(HSW_PWR_WELL_DRIVER) &
+		      HSW_PWR_WELL_ENABLE_REQUEST))
+			I915_WRITE(HSW_PWR_WELL_DRIVER,
+				   HSW_PWR_WELL_ENABLE_REQUEST);
 		I915_WRITE(HSW_PWR_WELL_BIOS, 0);
+	}
 }
 
 static void hsw_power_well_enable(struct drm_i915_private *dev_priv,
@@ -881,10 +871,17 @@ static bool skl_power_well_enabled(struct drm_i915_private *dev_priv,
 static void skl_power_well_sync_hw(struct drm_i915_private *dev_priv,
 				struct i915_power_well *power_well)
 {
-	skl_set_power_well(dev_priv, power_well, power_well->count > 0);
+	uint32_t mask = SKL_POWER_WELL_REQ(power_well->id);
+	uint32_t bios_req = I915_READ(HSW_PWR_WELL_BIOS);
+
+	/* Take over the request bit if set by BIOS. */
+	if (bios_req & mask) {
+		uint32_t drv_req = I915_READ(HSW_PWR_WELL_DRIVER);
 
-	/* Clear any request made by BIOS as driver is taking over */
-	I915_WRITE(HSW_PWR_WELL_BIOS, 0);
+		if (!(drv_req & mask))
+			I915_WRITE(HSW_PWR_WELL_DRIVER, drv_req | mask);
+		I915_WRITE(HSW_PWR_WELL_BIOS, bios_req & ~mask);
+	}
 }
 
 static void skl_power_well_enable(struct drm_i915_private *dev_priv,
@@ -917,16 +914,6 @@ static bool bxt_dpio_cmn_power_well_enabled(struct drm_i915_private *dev_priv,
 	return bxt_ddi_phy_is_enabled(dev_priv, power_well->data);
 }
 
-static void bxt_dpio_cmn_power_well_sync_hw(struct drm_i915_private *dev_priv,
-					    struct i915_power_well *power_well)
-{
-	if (power_well->count > 0)
-		bxt_dpio_cmn_power_well_enable(dev_priv, power_well);
-	else
-		bxt_dpio_cmn_power_well_disable(dev_priv, power_well);
-}
-
-
 static void bxt_verify_ddi_phy_power_wells(struct drm_i915_private *dev_priv)
 {
 	struct i915_power_well *power_well;
@@ -964,10 +951,12 @@ static void gen9_assert_dbuf_enabled(struct drm_i915_private *dev_priv)
 static void gen9_dc_off_power_well_enable(struct drm_i915_private *dev_priv,
 					  struct i915_power_well *power_well)
 {
+	struct intel_cdclk_state cdclk_state = {};
+
 	gen9_set_dc_state(dev_priv, DC_STATE_DISABLE);
 
-	WARN_ON(dev_priv->cdclk_freq !=
-		dev_priv->display.get_display_clock_speed(dev_priv));
+	dev_priv->display.get_cdclk(dev_priv, &cdclk_state);
+	WARN_ON(!intel_cdclk_state_compare(&dev_priv->cdclk.hw, &cdclk_state));
 
 	gen9_assert_dbuf_enabled(dev_priv);
 
@@ -987,13 +976,9 @@ static void gen9_dc_off_power_well_disable(struct drm_i915_private *dev_priv,
 		gen9_enable_dc5(dev_priv);
 }
 
-static void gen9_dc_off_power_well_sync_hw(struct drm_i915_private *dev_priv,
-					   struct i915_power_well *power_well)
+static void i9xx_power_well_sync_hw_noop(struct drm_i915_private *dev_priv,
+					 struct i915_power_well *power_well)
 {
-	if (power_well->count > 0)
-		gen9_dc_off_power_well_enable(dev_priv, power_well);
-	else
-		gen9_dc_off_power_well_disable(dev_priv, power_well);
 }
 
 static void i9xx_always_on_power_well_noop(struct drm_i915_private *dev_priv,
@@ -1043,12 +1028,6 @@ out:
 	mutex_unlock(&dev_priv->rps.hw_lock);
 }
 
-static void vlv_power_well_sync_hw(struct drm_i915_private *dev_priv,
-				   struct i915_power_well *power_well)
-{
-	vlv_set_power_well(dev_priv, power_well, power_well->count > 0);
-}
-
 static void vlv_power_well_enable(struct drm_i915_private *dev_priv,
 				  struct i915_power_well *power_well)
 {
@@ -1249,7 +1228,7 @@ static void vlv_dpio_cmn_power_well_disable(struct drm_i915_private *dev_priv,
 	vlv_set_power_well(dev_priv, power_well, false);
 }
 
-#define POWER_DOMAIN_MASK (BIT(POWER_DOMAIN_NUM) - 1)
+#define POWER_DOMAIN_MASK (GENMASK_ULL(POWER_DOMAIN_NUM - 1, 0))
 
 static struct i915_power_well *lookup_power_well(struct drm_i915_private *dev_priv,
 						 int power_well_id)
@@ -1659,14 +1638,6 @@ out:
 	mutex_unlock(&dev_priv->rps.hw_lock);
 }
 
-static void chv_pipe_power_well_sync_hw(struct drm_i915_private *dev_priv,
-					struct i915_power_well *power_well)
-{
-	WARN_ON_ONCE(power_well->id != PIPE_A);
-
-	chv_set_pipe_power_well(dev_priv, power_well, power_well->count > 0);
-}
-
 static void chv_pipe_power_well_enable(struct drm_i915_private *dev_priv,
 				       struct i915_power_well *power_well)
 {
@@ -1693,9 +1664,8 @@ __intel_display_power_get_domain(struct drm_i915_private *dev_priv,
 {
 	struct i915_power_domains *power_domains = &dev_priv->power_domains;
 	struct i915_power_well *power_well;
-	int i;
 
-	for_each_power_well(i, power_well, BIT(domain), power_domains)
+	for_each_power_domain_well(dev_priv, power_well, BIT_ULL(domain))
 		intel_power_well_get(dev_priv, power_well);
 
 	power_domains->domain_use_count[domain]++;
@@ -1779,7 +1749,6 @@ void intel_display_power_put(struct drm_i915_private *dev_priv,
 {
 	struct i915_power_domains *power_domains;
 	struct i915_power_well *power_well;
-	int i;
 
 	power_domains = &dev_priv->power_domains;
 
@@ -1790,7 +1759,7 @@ void intel_display_power_put(struct drm_i915_private *dev_priv,
 	     intel_display_power_domain_str(domain));
 	power_domains->domain_use_count[domain]--;
 
-	for_each_power_well_rev(i, power_well, BIT(domain), power_domains)
+	for_each_power_domain_well_rev(dev_priv, power_well, BIT_ULL(domain))
 		intel_power_well_put(dev_priv, power_well);
 
 	mutex_unlock(&power_domains->lock);
@@ -1799,134 +1768,134 @@ void intel_display_power_put(struct drm_i915_private *dev_priv,
 }
 
 #define HSW_DISPLAY_POWER_DOMAINS (			\
-	BIT(POWER_DOMAIN_PIPE_B) |			\
-	BIT(POWER_DOMAIN_PIPE_C) |			\
-	BIT(POWER_DOMAIN_PIPE_A_PANEL_FITTER) |		\
-	BIT(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |		\
-	BIT(POWER_DOMAIN_PIPE_C_PANEL_FITTER) |		\
-	BIT(POWER_DOMAIN_TRANSCODER_A) |		\
-	BIT(POWER_DOMAIN_TRANSCODER_B) |		\
-	BIT(POWER_DOMAIN_TRANSCODER_C) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_D_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_CRT) | /* DDI E */	\
-	BIT(POWER_DOMAIN_VGA) |				\
-	BIT(POWER_DOMAIN_AUDIO) |			\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PIPE_B) |			\
+	BIT_ULL(POWER_DOMAIN_PIPE_C) |			\
+	BIT_ULL(POWER_DOMAIN_PIPE_A_PANEL_FITTER) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) |		\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_A) |		\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_B) |		\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_C) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_D_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_CRT) | /* DDI E */	\
+	BIT_ULL(POWER_DOMAIN_VGA) |				\
+	BIT_ULL(POWER_DOMAIN_AUDIO) |			\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 #define BDW_DISPLAY_POWER_DOMAINS (			\
-	BIT(POWER_DOMAIN_PIPE_B) |			\
-	BIT(POWER_DOMAIN_PIPE_C) |			\
-	BIT(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |		\
-	BIT(POWER_DOMAIN_PIPE_C_PANEL_FITTER) |		\
-	BIT(POWER_DOMAIN_TRANSCODER_A) |		\
-	BIT(POWER_DOMAIN_TRANSCODER_B) |		\
-	BIT(POWER_DOMAIN_TRANSCODER_C) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_DDI_D_LANES) |		\
-	BIT(POWER_DOMAIN_PORT_CRT) | /* DDI E */	\
-	BIT(POWER_DOMAIN_VGA) |				\
-	BIT(POWER_DOMAIN_AUDIO) |			\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PIPE_B) |			\
+	BIT_ULL(POWER_DOMAIN_PIPE_C) |			\
+	BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) |		\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_A) |		\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_B) |		\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_C) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_D_LANES) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_CRT) | /* DDI E */	\
+	BIT_ULL(POWER_DOMAIN_VGA) |				\
+	BIT_ULL(POWER_DOMAIN_AUDIO) |			\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 #define VLV_DISPLAY_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PIPE_A) |		\
-	BIT(POWER_DOMAIN_PIPE_B) |		\
-	BIT(POWER_DOMAIN_PIPE_A_PANEL_FITTER) |	\
-	BIT(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |	\
-	BIT(POWER_DOMAIN_TRANSCODER_A) |	\
-	BIT(POWER_DOMAIN_TRANSCODER_B) |	\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_DSI) |		\
-	BIT(POWER_DOMAIN_PORT_CRT) |		\
-	BIT(POWER_DOMAIN_VGA) |			\
-	BIT(POWER_DOMAIN_AUDIO) |		\
-	BIT(POWER_DOMAIN_AUX_B) |		\
-	BIT(POWER_DOMAIN_AUX_C) |		\
-	BIT(POWER_DOMAIN_GMBUS) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PIPE_A) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_B) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_A_PANEL_FITTER) |	\
+	BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |	\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_A) |	\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_B) |	\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_PORT_DSI) |		\
+	BIT_ULL(POWER_DOMAIN_PORT_CRT) |		\
+	BIT_ULL(POWER_DOMAIN_VGA) |			\
+	BIT_ULL(POWER_DOMAIN_AUDIO) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_B) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_C) |		\
+	BIT_ULL(POWER_DOMAIN_GMBUS) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 #define VLV_DPIO_CMN_BC_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_CRT) |		\
-	BIT(POWER_DOMAIN_AUX_B) |		\
-	BIT(POWER_DOMAIN_AUX_C) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_PORT_CRT) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_B) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_C) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 #define VLV_DPIO_TX_B_LANES_01_POWER_DOMAINS (	\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |	\
-	BIT(POWER_DOMAIN_AUX_B) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_AUX_B) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 #define VLV_DPIO_TX_B_LANES_23_POWER_DOMAINS (	\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |	\
-	BIT(POWER_DOMAIN_AUX_B) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_AUX_B) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 #define VLV_DPIO_TX_C_LANES_01_POWER_DOMAINS (	\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |	\
-	BIT(POWER_DOMAIN_AUX_C) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_AUX_C) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 #define VLV_DPIO_TX_C_LANES_23_POWER_DOMAINS (	\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |	\
-	BIT(POWER_DOMAIN_AUX_C) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_AUX_C) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 #define CHV_DISPLAY_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PIPE_A) |		\
-	BIT(POWER_DOMAIN_PIPE_B) |		\
-	BIT(POWER_DOMAIN_PIPE_C) |		\
-	BIT(POWER_DOMAIN_PIPE_A_PANEL_FITTER) |	\
-	BIT(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |	\
-	BIT(POWER_DOMAIN_PIPE_C_PANEL_FITTER) |	\
-	BIT(POWER_DOMAIN_TRANSCODER_A) |	\
-	BIT(POWER_DOMAIN_TRANSCODER_B) |	\
-	BIT(POWER_DOMAIN_TRANSCODER_C) |	\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_DDI_D_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_DSI) |		\
-	BIT(POWER_DOMAIN_VGA) |			\
-	BIT(POWER_DOMAIN_AUDIO) |		\
-	BIT(POWER_DOMAIN_AUX_B) |		\
-	BIT(POWER_DOMAIN_AUX_C) |		\
-	BIT(POWER_DOMAIN_AUX_D) |		\
-	BIT(POWER_DOMAIN_GMBUS) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PIPE_A) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_B) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_C) |		\
+	BIT_ULL(POWER_DOMAIN_PIPE_A_PANEL_FITTER) |	\
+	BIT_ULL(POWER_DOMAIN_PIPE_B_PANEL_FITTER) |	\
+	BIT_ULL(POWER_DOMAIN_PIPE_C_PANEL_FITTER) |	\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_A) |	\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_B) |	\
+	BIT_ULL(POWER_DOMAIN_TRANSCODER_C) |	\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_D_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_PORT_DSI) |		\
+	BIT_ULL(POWER_DOMAIN_VGA) |			\
+	BIT_ULL(POWER_DOMAIN_AUDIO) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_B) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_C) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_D) |		\
+	BIT_ULL(POWER_DOMAIN_GMBUS) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 #define CHV_DPIO_CMN_BC_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PORT_DDI_B_LANES) |	\
-	BIT(POWER_DOMAIN_PORT_DDI_C_LANES) |	\
-	BIT(POWER_DOMAIN_AUX_B) |		\
-	BIT(POWER_DOMAIN_AUX_C) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_B_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_C_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_AUX_B) |		\
+	BIT_ULL(POWER_DOMAIN_AUX_C) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 #define CHV_DPIO_CMN_D_POWER_DOMAINS (		\
-	BIT(POWER_DOMAIN_PORT_DDI_D_LANES) |	\
-	BIT(POWER_DOMAIN_AUX_D) |		\
-	BIT(POWER_DOMAIN_INIT))
+	BIT_ULL(POWER_DOMAIN_PORT_DDI_D_LANES) |	\
+	BIT_ULL(POWER_DOMAIN_AUX_D) |		\
+	BIT_ULL(POWER_DOMAIN_INIT))
 
 static const struct i915_power_well_ops i9xx_always_on_power_well_ops = {
-	.sync_hw = i9xx_always_on_power_well_noop,
+	.sync_hw = i9xx_power_well_sync_hw_noop,
 	.enable = i9xx_always_on_power_well_noop,
 	.disable = i9xx_always_on_power_well_noop,
 	.is_enabled = i9xx_always_on_power_well_enabled,
 };
 
 static const struct i915_power_well_ops chv_pipe_power_well_ops = {
-	.sync_hw = chv_pipe_power_well_sync_hw,
+	.sync_hw = i9xx_power_well_sync_hw_noop,
 	.enable = chv_pipe_power_well_enable,
 	.disable = chv_pipe_power_well_disable,
 	.is_enabled = chv_pipe_power_well_enabled,
 };
 
 static const struct i915_power_well_ops chv_dpio_cmn_power_well_ops = {
-	.sync_hw = vlv_power_well_sync_hw,
+	.sync_hw = i9xx_power_well_sync_hw_noop,
 	.enable = chv_dpio_cmn_power_well_enable,
 	.disable = chv_dpio_cmn_power_well_disable,
 	.is_enabled = vlv_power_well_enabled,
@@ -1956,14 +1925,14 @@ static const struct i915_power_well_ops skl_power_well_ops = {
 };
 
 static const struct i915_power_well_ops gen9_dc_off_power_well_ops = {
-	.sync_hw = gen9_dc_off_power_well_sync_hw,
+	.sync_hw = i9xx_power_well_sync_hw_noop,
 	.enable = gen9_dc_off_power_well_enable,
 	.disable = gen9_dc_off_power_well_disable,
 	.is_enabled = gen9_dc_off_power_well_enabled,
 };
 
 static const struct i915_power_well_ops bxt_dpio_cmn_power_well_ops = {
-	.sync_hw = bxt_dpio_cmn_power_well_sync_hw,
+	.sync_hw = i9xx_power_well_sync_hw_noop,
 	.enable = bxt_dpio_cmn_power_well_enable,
 	.disable = bxt_dpio_cmn_power_well_disable,
 	.is_enabled = bxt_dpio_cmn_power_well_enabled,
@@ -1998,21 +1967,21 @@ static struct i915_power_well bdw_power_wells[] = {
 };
 
 static const struct i915_power_well_ops vlv_display_power_well_ops = {
-	.sync_hw = vlv_power_well_sync_hw,
+	.sync_hw = i9xx_power_well_sync_hw_noop,
 	.enable = vlv_display_power_well_enable,
 	.disable = vlv_display_power_well_disable,
 	.is_enabled = vlv_power_well_enabled,
 };
 
 static const struct i915_power_well_ops vlv_dpio_cmn_power_well_ops = {
-	.sync_hw = vlv_power_well_sync_hw,
+	.sync_hw = i9xx_power_well_sync_hw_noop,
 	.enable = vlv_dpio_cmn_power_well_enable,
 	.disable = vlv_dpio_cmn_power_well_disable,
 	.is_enabled = vlv_power_well_enabled,
 };
 
 static const struct i915_power_well_ops vlv_dpio_power_well_ops = {
-	.sync_hw = vlv_power_well_sync_hw,
+	.sync_hw = i9xx_power_well_sync_hw_noop,
 	.enable = vlv_power_well_enable,
 	.disable = vlv_power_well_disable,
 	.is_enabled = vlv_power_well_enabled,
@@ -2155,26 +2124,26 @@ static struct i915_power_well skl_power_wells[] = {
 		.id = SKL_DISP_PW_2,
 	},
 	{
-		.name = "DDI A/E power well",
-		.domains = SKL_DISPLAY_DDI_A_E_POWER_DOMAINS,
+		.name = "DDI A/E IO power well",
+		.domains = SKL_DISPLAY_DDI_IO_A_E_POWER_DOMAINS,
 		.ops = &skl_power_well_ops,
 		.id = SKL_DISP_PW_DDI_A_E,
 	},
 	{
-		.name = "DDI B power well",
-		.domains = SKL_DISPLAY_DDI_B_POWER_DOMAINS,
+		.name = "DDI B IO power well",
+		.domains = SKL_DISPLAY_DDI_IO_B_POWER_DOMAINS,
 		.ops = &skl_power_well_ops,
 		.id = SKL_DISP_PW_DDI_B,
 	},
 	{
-		.name = "DDI C power well",
-		.domains = SKL_DISPLAY_DDI_C_POWER_DOMAINS,
+		.name = "DDI C IO power well",
+		.domains = SKL_DISPLAY_DDI_IO_C_POWER_DOMAINS,
 		.ops = &skl_power_well_ops,
 		.id = SKL_DISP_PW_DDI_C,
 	},
 	{
-		.name = "DDI D power well",
-		.domains = SKL_DISPLAY_DDI_D_POWER_DOMAINS,
+		.name = "DDI D IO power well",
+		.domains = SKL_DISPLAY_DDI_IO_D_POWER_DOMAINS,
 		.ops = &skl_power_well_ops,
 		.id = SKL_DISP_PW_DDI_D,
 	},
@@ -2287,20 +2256,20 @@ static struct i915_power_well glk_power_wells[] = {
 		.id = GLK_DISP_PW_AUX_C,
 	},
 	{
-		.name = "DDI A power well",
-		.domains = GLK_DISPLAY_DDI_A_POWER_DOMAINS,
+		.name = "DDI A IO power well",
+		.domains = GLK_DISPLAY_DDI_IO_A_POWER_DOMAINS,
 		.ops = &skl_power_well_ops,
 		.id = GLK_DISP_PW_DDI_A,
 	},
 	{
-		.name = "DDI B power well",
-		.domains = GLK_DISPLAY_DDI_B_POWER_DOMAINS,
+		.name = "DDI B IO power well",
+		.domains = GLK_DISPLAY_DDI_IO_B_POWER_DOMAINS,
 		.ops = &skl_power_well_ops,
 		.id = SKL_DISP_PW_DDI_B,
 	},
 	{
-		.name = "DDI C power well",
-		.domains = GLK_DISPLAY_DDI_C_POWER_DOMAINS,
+		.name = "DDI C IO power well",
+		.domains = GLK_DISPLAY_DDI_IO_C_POWER_DOMAINS,
 		.ops = &skl_power_well_ops,
 		.id = SKL_DISP_PW_DDI_C,
 	},
@@ -2323,7 +2292,7 @@ static uint32_t get_allowed_dc_mask(const struct drm_i915_private *dev_priv,
 	int requested_dc;
 	int max_dc;
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	if (IS_GEN9_BC(dev_priv)) {
 		max_dc = 2;
 		mask = 0;
 	} else if (IS_GEN9_LP(dev_priv)) {
@@ -2386,7 +2355,7 @@ int intel_power_domains_init(struct drm_i915_private *dev_priv)
 	dev_priv->csr.allowed_dc_mask = get_allowed_dc_mask(dev_priv,
 							    i915.enable_dc);
 
-	BUILD_BUG_ON(POWER_DOMAIN_NUM > 31);
+	BUILD_BUG_ON(POWER_DOMAIN_NUM > 64);
 
 	mutex_init(&power_domains->lock);
 
@@ -2398,7 +2367,7 @@ int intel_power_domains_init(struct drm_i915_private *dev_priv)
 		set_power_wells(power_domains, hsw_power_wells);
 	} else if (IS_BROADWELL(dev_priv)) {
 		set_power_wells(power_domains, bdw_power_wells);
-	} else if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	} else if (IS_GEN9_BC(dev_priv)) {
 		set_power_wells(power_domains, skl_power_wells);
 	} else if (IS_BROXTON(dev_priv)) {
 		set_power_wells(power_domains, bxt_power_wells);
@@ -2454,10 +2423,9 @@ static void intel_power_domains_sync_hw(struct drm_i915_private *dev_priv)
 {
 	struct i915_power_domains *power_domains = &dev_priv->power_domains;
 	struct i915_power_well *power_well;
-	int i;
 
 	mutex_lock(&power_domains->lock);
-	for_each_power_well(i, power_well, POWER_DOMAIN_MASK, power_domains) {
+	for_each_power_well(dev_priv, power_well) {
 		power_well->ops->sync_hw(dev_priv, power_well);
 		power_well->hw_enabled = power_well->ops->is_enabled(dev_priv,
 								     power_well);
@@ -2722,7 +2690,10 @@ static void vlv_cmnlane_wa(struct drm_i915_private *dev_priv)
  * @resume: Called from resume code paths or not
  *
  * This function initializes the hardware power domain state and enables all
- * power domains using intel_display_set_init_power().
+ * power wells belonging to the INIT power domain. Power wells in other
+ * domains (and not in the INIT domain) are referenced or disabled during the
+ * modeset state HW readout. After that the reference count of each power well
+ * must match its HW enabled state, see intel_power_domains_verify_state().
  */
 void intel_power_domains_init_hw(struct drm_i915_private *dev_priv, bool resume)
 {
@@ -2730,7 +2701,7 @@ void intel_power_domains_init_hw(struct drm_i915_private *dev_priv, bool resume)
 
 	power_domains->initializing = true;
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	if (IS_GEN9_BC(dev_priv)) {
 		skl_display_core_init(dev_priv, resume);
 	} else if (IS_GEN9_LP(dev_priv)) {
 		bxt_display_core_init(dev_priv, resume);
@@ -2769,12 +2740,92 @@ void intel_power_domains_suspend(struct drm_i915_private *dev_priv)
 	if (!i915.disable_power_well)
 		intel_display_power_put(dev_priv, POWER_DOMAIN_INIT);
 
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
+	if (IS_GEN9_BC(dev_priv))
 		skl_display_core_uninit(dev_priv);
 	else if (IS_GEN9_LP(dev_priv))
 		bxt_display_core_uninit(dev_priv);
 }
 
+static void intel_power_domains_dump_info(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_domains *power_domains = &dev_priv->power_domains;
+	struct i915_power_well *power_well;
+
+	for_each_power_well(dev_priv, power_well) {
+		enum intel_display_power_domain domain;
+
+		DRM_DEBUG_DRIVER("%-25s %d\n",
+				 power_well->name, power_well->count);
+
+		for_each_power_domain(domain, power_well->domains)
+			DRM_DEBUG_DRIVER("  %-23s %d\n",
+					 intel_display_power_domain_str(domain),
+					 power_domains->domain_use_count[domain]);
+	}
+}
+
+/**
+ * intel_power_domains_verify_state - verify the HW/SW state for all power wells
+ * @dev_priv: i915 device instance
+ *
+ * Verify if the reference count of each power well matches its HW enabled
+ * state and the total refcount of the domains it belongs to. This must be
+ * called after modeset HW state sanitization, which is responsible for
+ * acquiring reference counts for any power wells in use and disabling the
+ * ones left on by BIOS but not required by any active output.
+ */
+void intel_power_domains_verify_state(struct drm_i915_private *dev_priv)
+{
+	struct i915_power_domains *power_domains = &dev_priv->power_domains;
+	struct i915_power_well *power_well;
+	bool dump_domain_info;
+
+	mutex_lock(&power_domains->lock);
+
+	dump_domain_info = false;
+	for_each_power_well(dev_priv, power_well) {
+		enum intel_display_power_domain domain;
+		int domains_count;
+		bool enabled;
+
+		/*
+		 * Power wells not belonging to any domain (like the MISC_IO
+		 * and PW1 power wells) are under FW control, so ignore them,
+		 * since their state can change asynchronously.
+		 */
+		if (!power_well->domains)
+			continue;
+
+		enabled = power_well->ops->is_enabled(dev_priv, power_well);
+		if ((power_well->count || power_well->always_on) != enabled)
+			DRM_ERROR("power well %s state mismatch (refcount %d/enabled %d)",
+				  power_well->name, power_well->count, enabled);
+
+		domains_count = 0;
+		for_each_power_domain(domain, power_well->domains)
+			domains_count += power_domains->domain_use_count[domain];
+
+		if (power_well->count != domains_count) {
+			DRM_ERROR("power well %s refcount/domain refcount mismatch "
+				  "(refcount %d/domains refcount %d)\n",
+				  power_well->name, power_well->count,
+				  domains_count);
+			dump_domain_info = true;
+		}
+	}
+
+	if (dump_domain_info) {
+		static bool dumped;
+
+		if (!dumped) {
+			intel_power_domains_dump_info(dev_priv);
+			dumped = true;
+		}
+	}
+
+	mutex_unlock(&power_domains->lock);
+}
+
 /**
  * intel_runtime_pm_get - grab a runtime pm reference
  * @dev_priv: i915 device instance
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 2ad13903a054..816a6f5a3fd9 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -2981,6 +2981,7 @@ bool intel_sdvo_init(struct drm_i915_private *dev_priv,
 	/* encoder type will be decided later */
 	intel_encoder = &intel_sdvo->base;
 	intel_encoder->type = INTEL_OUTPUT_SDVO;
+	intel_encoder->power_domain = POWER_DOMAIN_PORT_OTHER;
 	intel_encoder->port = port;
 	drm_encoder_init(&dev_priv->drm, &intel_encoder->base,
 			 &intel_sdvo_enc_funcs, 0,
diff --git a/drivers/gpu/drm/i915/intel_sideband.c b/drivers/gpu/drm/i915/intel_sideband.c
index 1a840bf92eea..7d971cb56116 100644
--- a/drivers/gpu/drm/i915/intel_sideband.c
+++ b/drivers/gpu/drm/i915/intel_sideband.c
@@ -60,8 +60,7 @@ static int vlv_sideband_rw(struct drm_i915_private *dev_priv, u32 devfn,
 	}
 
 	I915_WRITE(VLV_IOSF_ADDR, addr);
-	if (!is_read)
-		I915_WRITE(VLV_IOSF_DATA, *val);
+	I915_WRITE(VLV_IOSF_DATA, is_read ? 0 : *val);
 	I915_WRITE(VLV_IOSF_DOORBELL_REQ, cmd);
 
 	if (intel_wait_for_register(dev_priv,
@@ -74,7 +73,6 @@ static int vlv_sideband_rw(struct drm_i915_private *dev_priv, u32 devfn,
 
 	if (is_read)
 		*val = I915_READ(VLV_IOSF_DATA);
-	I915_WRITE(VLV_IOSF_DATA, 0);
 
 	return 0;
 }
@@ -93,14 +91,18 @@ u32 vlv_punit_read(struct drm_i915_private *dev_priv, u32 addr)
 	return val;
 }
 
-void vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val)
+int vlv_punit_write(struct drm_i915_private *dev_priv, u32 addr, u32 val)
 {
+	int err;
+
 	WARN_ON(!mutex_is_locked(&dev_priv->rps.hw_lock));
 
 	mutex_lock(&dev_priv->sb_lock);
-	vlv_sideband_rw(dev_priv, PCI_DEVFN(0, 0), IOSF_PORT_PUNIT,
-			SB_CRWRDA_NP, addr, &val);
+	err = vlv_sideband_rw(dev_priv, PCI_DEVFN(0, 0), IOSF_PORT_PUNIT,
+			      SB_CRWRDA_NP, addr, &val);
 	mutex_unlock(&dev_priv->sb_lock);
+
+	return err;
 }
 
 u32 vlv_bunit_read(struct drm_i915_private *dev_priv, u32 reg)
@@ -214,6 +216,7 @@ u32 intel_sbi_read(struct drm_i915_private *dev_priv, u16 reg,
 	}
 
 	I915_WRITE(SBI_ADDR, (reg << 16));
+	I915_WRITE(SBI_DATA, 0);
 
 	if (destination == SBI_ICLK)
 		value = SBI_CTL_DEST_ICLK | SBI_CTL_OP_CRRD;
@@ -223,10 +226,15 @@ u32 intel_sbi_read(struct drm_i915_private *dev_priv, u16 reg,
 
 	if (intel_wait_for_register(dev_priv,
 				    SBI_CTL_STAT,
-				    SBI_BUSY | SBI_RESPONSE_FAIL,
+				    SBI_BUSY,
 				    0,
 				    100)) {
-		DRM_ERROR("timeout waiting for SBI to complete read transaction\n");
+		DRM_ERROR("timeout waiting for SBI to complete read\n");
+		return 0;
+	}
+
+	if (I915_READ(SBI_CTL_STAT) & SBI_RESPONSE_FAIL) {
+		DRM_ERROR("error during SBI read of reg %x\n", reg);
 		return 0;
 	}
 
@@ -258,10 +266,16 @@ void intel_sbi_write(struct drm_i915_private *dev_priv, u16 reg, u32 value,
 
 	if (intel_wait_for_register(dev_priv,
 				    SBI_CTL_STAT,
-				    SBI_BUSY | SBI_RESPONSE_FAIL,
+				    SBI_BUSY,
 				    0,
 				    100)) {
-		DRM_ERROR("timeout waiting for SBI to complete write transaction\n");
+		DRM_ERROR("timeout waiting for SBI to complete write\n");
+		return;
+	}
+
+	if (I915_READ(SBI_CTL_STAT) & SBI_RESPONSE_FAIL) {
+		DRM_ERROR("error during SBI write of %x to reg %x\n",
+			  value, reg);
 		return;
 	}
 }
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index 9ef54688872a..27e0752d1578 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -219,13 +219,22 @@ skl_update_plane(struct drm_plane *drm_plane,
 	uint32_t src_w = drm_rect_width(&plane_state->base.src) >> 16;
 	uint32_t src_h = drm_rect_height(&plane_state->base.src) >> 16;
 
-	plane_ctl = PLANE_CTL_ENABLE |
-		PLANE_CTL_PIPE_GAMMA_ENABLE |
-		PLANE_CTL_PIPE_CSC_ENABLE;
+	plane_ctl = PLANE_CTL_ENABLE;
+
+	if (IS_GEMINILAKE(dev_priv)) {
+		I915_WRITE(PLANE_COLOR_CTL(pipe, plane_id),
+			   PLANE_COLOR_PIPE_GAMMA_ENABLE |
+			   PLANE_COLOR_PIPE_CSC_ENABLE |
+			   PLANE_COLOR_PLANE_GAMMA_DISABLE);
+	} else {
+		plane_ctl |=
+			PLANE_CTL_PIPE_GAMMA_ENABLE |
+			PLANE_CTL_PIPE_CSC_ENABLE |
+			PLANE_CTL_PLANE_GAMMA_DISABLE;
+	}
 
 	plane_ctl |= skl_plane_ctl_format(fb->format->format);
 	plane_ctl |= skl_plane_ctl_tiling(fb->modifier);
-
 	plane_ctl |= skl_plane_ctl_rotation(rotation);
 
 	if (key->flags) {
diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c
index eb692e4ffe01..6ed1a3ce47b7 100644
--- a/drivers/gpu/drm/i915/intel_tv.c
+++ b/drivers/gpu/drm/i915/intel_tv.c
@@ -1621,6 +1621,7 @@ intel_tv_init(struct drm_i915_private *dev_priv)
 	intel_connector_attach_encoder(intel_connector, intel_encoder);
 
 	intel_encoder->type = INTEL_OUTPUT_TVOUT;
+	intel_encoder->power_domain = POWER_DOMAIN_PORT_OTHER;
 	intel_encoder->port = PORT_NONE;
 	intel_encoder->crtc_mask = (1 << 0) | (1 << 1);
 	intel_encoder->cloneable = 0;
diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
index abe08885a5ba..b35b7a03eeaf 100644
--- a/drivers/gpu/drm/i915/intel_uncore.c
+++ b/drivers/gpu/drm/i915/intel_uncore.c
@@ -133,6 +133,13 @@ fw_domains_put(struct drm_i915_private *dev_priv, enum forcewake_domains fw_doma
 }
 
 static void
+vgpu_fw_domains_nop(struct drm_i915_private *dev_priv,
+		    enum forcewake_domains fw_domains)
+{
+	/* Guest driver doesn't need to takes care forcewake. */
+}
+
+static void
 fw_domains_posting_read(struct drm_i915_private *dev_priv)
 {
 	struct intel_uncore_forcewake_domain *d;
@@ -499,7 +506,7 @@ void intel_uncore_forcewake_get(struct drm_i915_private *dev_priv,
 void intel_uncore_forcewake_get__locked(struct drm_i915_private *dev_priv,
 					enum forcewake_domains fw_domains)
 {
-	assert_spin_locked(&dev_priv->uncore.lock);
+	lockdep_assert_held(&dev_priv->uncore.lock);
 
 	if (!dev_priv->uncore.funcs.force_wake_get)
 		return;
@@ -557,7 +564,7 @@ void intel_uncore_forcewake_put(struct drm_i915_private *dev_priv,
 void intel_uncore_forcewake_put__locked(struct drm_i915_private *dev_priv,
 					enum forcewake_domains fw_domains)
 {
-	assert_spin_locked(&dev_priv->uncore.lock);
+	lockdep_assert_held(&dev_priv->uncore.lock);
 
 	if (!dev_priv->uncore.funcs.force_wake_put)
 		return;
@@ -635,33 +642,6 @@ find_fw_domain(struct drm_i915_private *dev_priv, u32 offset)
 	return entry->domains;
 }
 
-static void
-intel_fw_table_check(struct drm_i915_private *dev_priv)
-{
-	const struct intel_forcewake_range *ranges;
-	unsigned int num_ranges;
-	s32 prev;
-	unsigned int i;
-
-	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG))
-		return;
-
-	ranges = dev_priv->uncore.fw_domains_table;
-	if (!ranges)
-		return;
-
-	num_ranges = dev_priv->uncore.fw_domains_table_entries;
-
-	for (i = 0, prev = -1; i < num_ranges; i++, ranges++) {
-		WARN_ON_ONCE(IS_GEN9(dev_priv) &&
-			     (prev + 1) != (s32)ranges->start);
-		WARN_ON_ONCE(prev >= (s32)ranges->start);
-		prev = ranges->start;
-		WARN_ON_ONCE(prev >= (s32)ranges->end);
-		prev = ranges->end;
-	}
-}
-
 #define GEN_FW_RANGE(s, e, d) \
 	{ .start = (s), .end = (e), .domains = (d) }
 
@@ -700,23 +680,6 @@ static const i915_reg_t gen8_shadowed_regs[] = {
 	/* TODO: Other registers are not yet used */
 };
 
-static void intel_shadow_table_check(void)
-{
-	const i915_reg_t *reg = gen8_shadowed_regs;
-	s32 prev;
-	u32 offset;
-	unsigned int i;
-
-	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG))
-		return;
-
-	for (i = 0, prev = -1; i < ARRAY_SIZE(gen8_shadowed_regs); i++, reg++) {
-		offset = i915_mmio_reg_offset(*reg);
-		WARN_ON_ONCE(prev >= (s32)offset);
-		prev = offset;
-	}
-}
-
 static int mmio_reg_cmp(u32 key, const i915_reg_t *reg)
 {
 	u32 offset = i915_mmio_reg_offset(*reg);
@@ -985,29 +948,19 @@ static inline void __force_wake_auto(struct drm_i915_private *dev_priv,
 		___force_wake_auto(dev_priv, fw_domains);
 }
 
-#define __gen6_read(x) \
-static u##x \
-gen6_read##x(struct drm_i915_private *dev_priv, i915_reg_t reg, bool trace) { \
-	enum forcewake_domains fw_engine; \
-	GEN6_READ_HEADER(x); \
-	fw_engine = __gen6_reg_read_fw_domains(offset); \
-	if (fw_engine) \
-		__force_wake_auto(dev_priv, fw_engine); \
-	val = __raw_i915_read##x(dev_priv, reg); \
-	GEN6_READ_FOOTER; \
-}
-
-#define __fwtable_read(x) \
+#define __gen_read(func, x) \
 static u##x \
-fwtable_read##x(struct drm_i915_private *dev_priv, i915_reg_t reg, bool trace) { \
+func##_read##x(struct drm_i915_private *dev_priv, i915_reg_t reg, bool trace) { \
 	enum forcewake_domains fw_engine; \
 	GEN6_READ_HEADER(x); \
-	fw_engine = __fwtable_reg_read_fw_domains(offset); \
+	fw_engine = __##func##_reg_read_fw_domains(offset); \
 	if (fw_engine) \
 		__force_wake_auto(dev_priv, fw_engine); \
 	val = __raw_i915_read##x(dev_priv, reg); \
 	GEN6_READ_FOOTER; \
 }
+#define __gen6_read(x) __gen_read(gen6, x)
+#define __fwtable_read(x) __gen_read(fwtable, x)
 
 #define __gen9_decoupled_read(x) \
 static u##x \
@@ -1045,34 +998,6 @@ __gen6_read(64)
 #undef GEN6_READ_FOOTER
 #undef GEN6_READ_HEADER
 
-#define VGPU_READ_HEADER(x) \
-	unsigned long irqflags; \
-	u##x val = 0; \
-	assert_rpm_device_not_suspended(dev_priv); \
-	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags)
-
-#define VGPU_READ_FOOTER \
-	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags); \
-	trace_i915_reg_rw(false, reg, val, sizeof(val), trace); \
-	return val
-
-#define __vgpu_read(x) \
-static u##x \
-vgpu_read##x(struct drm_i915_private *dev_priv, i915_reg_t reg, bool trace) { \
-	VGPU_READ_HEADER(x); \
-	val = __raw_i915_read##x(dev_priv, reg); \
-	VGPU_READ_FOOTER; \
-}
-
-__vgpu_read(8)
-__vgpu_read(16)
-__vgpu_read(32)
-__vgpu_read(64)
-
-#undef __vgpu_read
-#undef VGPU_READ_FOOTER
-#undef VGPU_READ_HEADER
-
 #define GEN2_WRITE_HEADER \
 	trace_i915_reg_rw(true, reg, val, sizeof(val), trace); \
 	assert_rpm_wakelock_held(dev_priv); \
@@ -1136,29 +1061,19 @@ gen6_write##x(struct drm_i915_private *dev_priv, i915_reg_t reg, u##x val, bool
 	GEN6_WRITE_FOOTER; \
 }
 
-#define __gen8_write(x) \
-static void \
-gen8_write##x(struct drm_i915_private *dev_priv, i915_reg_t reg, u##x val, bool trace) { \
-	enum forcewake_domains fw_engine; \
-	GEN6_WRITE_HEADER; \
-	fw_engine = __gen8_reg_write_fw_domains(offset); \
-	if (fw_engine) \
-		__force_wake_auto(dev_priv, fw_engine); \
-	__raw_i915_write##x(dev_priv, reg, val); \
-	GEN6_WRITE_FOOTER; \
-}
-
-#define __fwtable_write(x) \
+#define __gen_write(func, x) \
 static void \
-fwtable_write##x(struct drm_i915_private *dev_priv, i915_reg_t reg, u##x val, bool trace) { \
+func##_write##x(struct drm_i915_private *dev_priv, i915_reg_t reg, u##x val, bool trace) { \
 	enum forcewake_domains fw_engine; \
 	GEN6_WRITE_HEADER; \
-	fw_engine = __fwtable_reg_write_fw_domains(offset); \
+	fw_engine = __##func##_reg_write_fw_domains(offset); \
 	if (fw_engine) \
 		__force_wake_auto(dev_priv, fw_engine); \
 	__raw_i915_write##x(dev_priv, reg, val); \
 	GEN6_WRITE_FOOTER; \
 }
+#define __gen8_write(x) __gen_write(gen8, x)
+#define __fwtable_write(x) __gen_write(fwtable, x)
 
 #define __gen9_decoupled_write(x) \
 static void \
@@ -1195,31 +1110,6 @@ __gen6_write(32)
 #undef GEN6_WRITE_FOOTER
 #undef GEN6_WRITE_HEADER
 
-#define VGPU_WRITE_HEADER \
-	unsigned long irqflags; \
-	trace_i915_reg_rw(true, reg, val, sizeof(val), trace); \
-	assert_rpm_device_not_suspended(dev_priv); \
-	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags)
-
-#define VGPU_WRITE_FOOTER \
-	spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags)
-
-#define __vgpu_write(x) \
-static void vgpu_write##x(struct drm_i915_private *dev_priv, \
-			  i915_reg_t reg, u##x val, bool trace) { \
-	VGPU_WRITE_HEADER; \
-	__raw_i915_write##x(dev_priv, reg, val); \
-	VGPU_WRITE_FOOTER; \
-}
-
-__vgpu_write(8)
-__vgpu_write(16)
-__vgpu_write(32)
-
-#undef __vgpu_write
-#undef VGPU_WRITE_FOOTER
-#undef VGPU_WRITE_HEADER
-
 #define ASSIGN_WRITE_MMIO_VFUNCS(x) \
 do { \
 	dev_priv->uncore.funcs.mmio_writeb = x##_write8; \
@@ -1375,6 +1265,11 @@ static void intel_uncore_fw_domains_init(struct drm_i915_private *dev_priv)
 			       FORCEWAKE, FORCEWAKE_ACK);
 	}
 
+	if (intel_vgpu_active(dev_priv)) {
+		dev_priv->uncore.funcs.force_wake_get = vgpu_fw_domains_nop;
+		dev_priv->uncore.funcs.force_wake_put = vgpu_fw_domains_nop;
+	}
+
 	/* All future platforms are expected to require complex power gating */
 	WARN_ON(dev_priv->uncore.fw_domains == 0);
 }
@@ -1445,15 +1340,6 @@ void intel_uncore_init(struct drm_i915_private *dev_priv)
 		break;
 	}
 
-	intel_fw_table_check(dev_priv);
-	if (INTEL_GEN(dev_priv) >= 8)
-		intel_shadow_table_check();
-
-	if (intel_vgpu_active(dev_priv)) {
-		ASSIGN_WRITE_MMIO_VFUNCS(vgpu);
-		ASSIGN_READ_MMIO_VFUNCS(vgpu);
-	}
-
 	i915_check_and_clear_faults(dev_priv);
 }
 #undef ASSIGN_WRITE_MMIO_VFUNCS
@@ -1971,3 +1857,7 @@ intel_uncore_forcewake_for_reg(struct drm_i915_private *dev_priv,
 
 	return fw_domains;
 }
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftests/intel_uncore.c"
+#endif
diff --git a/drivers/gpu/drm/i915/selftests/huge_gem_object.c b/drivers/gpu/drm/i915/selftests/huge_gem_object.c
new file mode 100644
index 000000000000..4e681fc13be4
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/huge_gem_object.c
@@ -0,0 +1,135 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "huge_gem_object.h"
+
+static void huge_free_pages(struct drm_i915_gem_object *obj,
+			    struct sg_table *pages)
+{
+	unsigned long nreal = obj->scratch / PAGE_SIZE;
+	struct scatterlist *sg;
+
+	for (sg = pages->sgl; sg && nreal--; sg = __sg_next(sg))
+		__free_page(sg_page(sg));
+
+	sg_free_table(pages);
+	kfree(pages);
+}
+
+static struct sg_table *
+huge_get_pages(struct drm_i915_gem_object *obj)
+{
+#define GFP (GFP_KERNEL | __GFP_NOWARN | __GFP_NORETRY)
+	const unsigned long nreal = obj->scratch / PAGE_SIZE;
+	const unsigned long npages = obj->base.size / PAGE_SIZE;
+	struct scatterlist *sg, *src, *end;
+	struct sg_table *pages;
+	unsigned long n;
+
+	pages = kmalloc(sizeof(*pages), GFP);
+	if (!pages)
+		return ERR_PTR(-ENOMEM);
+
+	if (sg_alloc_table(pages, npages, GFP)) {
+		kfree(pages);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	sg = pages->sgl;
+	for (n = 0; n < nreal; n++) {
+		struct page *page;
+
+		page = alloc_page(GFP | __GFP_HIGHMEM);
+		if (!page) {
+			sg_mark_end(sg);
+			goto err;
+		}
+
+		sg_set_page(sg, page, PAGE_SIZE, 0);
+		sg = __sg_next(sg);
+	}
+	if (nreal < npages) {
+		for (end = sg, src = pages->sgl; sg; sg = __sg_next(sg)) {
+			sg_set_page(sg, sg_page(src), PAGE_SIZE, 0);
+			src = __sg_next(src);
+			if (src == end)
+				src = pages->sgl;
+		}
+	}
+
+	if (i915_gem_gtt_prepare_pages(obj, pages))
+		goto err;
+
+	return pages;
+
+err:
+	huge_free_pages(obj, pages);
+	return ERR_PTR(-ENOMEM);
+#undef GFP
+}
+
+static void huge_put_pages(struct drm_i915_gem_object *obj,
+			   struct sg_table *pages)
+{
+	i915_gem_gtt_finish_pages(obj, pages);
+	huge_free_pages(obj, pages);
+
+	obj->mm.dirty = false;
+}
+
+static const struct drm_i915_gem_object_ops huge_ops = {
+	.flags = I915_GEM_OBJECT_HAS_STRUCT_PAGE |
+		 I915_GEM_OBJECT_IS_SHRINKABLE,
+	.get_pages = huge_get_pages,
+	.put_pages = huge_put_pages,
+};
+
+struct drm_i915_gem_object *
+huge_gem_object(struct drm_i915_private *i915,
+		phys_addr_t phys_size,
+		dma_addr_t dma_size)
+{
+	struct drm_i915_gem_object *obj;
+
+	GEM_BUG_ON(!phys_size || phys_size > dma_size);
+	GEM_BUG_ON(!IS_ALIGNED(phys_size, PAGE_SIZE));
+	GEM_BUG_ON(!IS_ALIGNED(dma_size, I915_GTT_PAGE_SIZE));
+
+	if (overflows_type(dma_size, obj->base.size))
+		return ERR_PTR(-E2BIG);
+
+	obj = i915_gem_object_alloc(i915);
+	if (!obj)
+		return ERR_PTR(-ENOMEM);
+
+	drm_gem_private_object_init(&i915->drm, &obj->base, dma_size);
+	i915_gem_object_init(obj, &huge_ops);
+
+	obj->base.write_domain = I915_GEM_DOMAIN_CPU;
+	obj->base.read_domains = I915_GEM_DOMAIN_CPU;
+	obj->cache_level = HAS_LLC(i915) ? I915_CACHE_LLC : I915_CACHE_NONE;
+	obj->scratch = phys_size;
+
+	return obj;
+}
diff --git a/drivers/gpu/drm/i915/selftests/huge_gem_object.h b/drivers/gpu/drm/i915/selftests/huge_gem_object.h
new file mode 100644
index 000000000000..a6133a9e8029
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/huge_gem_object.h
@@ -0,0 +1,45 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __HUGE_GEM_OBJECT_H
+#define __HUGE_GEM_OBJECT_H
+
+struct drm_i915_gem_object *
+huge_gem_object(struct drm_i915_private *i915,
+		phys_addr_t phys_size,
+		dma_addr_t dma_size);
+
+static inline phys_addr_t
+huge_gem_object_phys_size(struct drm_i915_gem_object *obj)
+{
+	return obj->scratch;
+}
+
+static inline dma_addr_t
+huge_gem_object_dma_size(struct drm_i915_gem_object *obj)
+{
+	return obj->base.size;
+}
+
+#endif /* !__HUGE_GEM_OBJECT_H */
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_coherency.c b/drivers/gpu/drm/i915/selftests/i915_gem_coherency.c
new file mode 100644
index 000000000000..f08d0179b3df
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_coherency.c
@@ -0,0 +1,385 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/prime_numbers.h>
+
+#include "../i915_selftest.h"
+#include "i915_random.h"
+
+static int cpu_set(struct drm_i915_gem_object *obj,
+		   unsigned long offset,
+		   u32 v)
+{
+	unsigned int needs_clflush;
+	struct page *page;
+	typeof(v) *map;
+	int err;
+
+	err = i915_gem_obj_prepare_shmem_write(obj, &needs_clflush);
+	if (err)
+		return err;
+
+	page = i915_gem_object_get_page(obj, offset >> PAGE_SHIFT);
+	map = kmap_atomic(page);
+	if (needs_clflush & CLFLUSH_BEFORE)
+		clflush(map+offset_in_page(offset) / sizeof(*map));
+	map[offset_in_page(offset) / sizeof(*map)] = v;
+	if (needs_clflush & CLFLUSH_AFTER)
+		clflush(map+offset_in_page(offset) / sizeof(*map));
+	kunmap_atomic(map);
+
+	i915_gem_obj_finish_shmem_access(obj);
+	return 0;
+}
+
+static int cpu_get(struct drm_i915_gem_object *obj,
+		   unsigned long offset,
+		   u32 *v)
+{
+	unsigned int needs_clflush;
+	struct page *page;
+	typeof(v) map;
+	int err;
+
+	err = i915_gem_obj_prepare_shmem_read(obj, &needs_clflush);
+	if (err)
+		return err;
+
+	page = i915_gem_object_get_page(obj, offset >> PAGE_SHIFT);
+	map = kmap_atomic(page);
+	if (needs_clflush & CLFLUSH_BEFORE)
+		clflush(map+offset_in_page(offset) / sizeof(*map));
+	*v = map[offset_in_page(offset) / sizeof(*map)];
+	kunmap_atomic(map);
+
+	i915_gem_obj_finish_shmem_access(obj);
+	return 0;
+}
+
+static int gtt_set(struct drm_i915_gem_object *obj,
+		   unsigned long offset,
+		   u32 v)
+{
+	struct i915_vma *vma;
+	typeof(v) *map;
+	int err;
+
+	err = i915_gem_object_set_to_gtt_domain(obj, true);
+	if (err)
+		return err;
+
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, PIN_MAPPABLE);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	map = i915_vma_pin_iomap(vma);
+	i915_vma_unpin(vma);
+	if (IS_ERR(map))
+		return PTR_ERR(map);
+
+	map[offset / sizeof(*map)] = v;
+	i915_vma_unpin_iomap(vma);
+
+	return 0;
+}
+
+static int gtt_get(struct drm_i915_gem_object *obj,
+		   unsigned long offset,
+		   u32 *v)
+{
+	struct i915_vma *vma;
+	typeof(v) map;
+	int err;
+
+	err = i915_gem_object_set_to_gtt_domain(obj, false);
+	if (err)
+		return err;
+
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, PIN_MAPPABLE);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	map = i915_vma_pin_iomap(vma);
+	i915_vma_unpin(vma);
+	if (IS_ERR(map))
+		return PTR_ERR(map);
+
+	*v = map[offset / sizeof(*map)];
+	i915_vma_unpin_iomap(vma);
+
+	return 0;
+}
+
+static int wc_set(struct drm_i915_gem_object *obj,
+		  unsigned long offset,
+		  u32 v)
+{
+	typeof(v) *map;
+	int err;
+
+	/* XXX GTT write followed by WC write go missing */
+	i915_gem_object_flush_gtt_write_domain(obj);
+
+	err = i915_gem_object_set_to_gtt_domain(obj, true);
+	if (err)
+		return err;
+
+	map = i915_gem_object_pin_map(obj, I915_MAP_WC);
+	if (IS_ERR(map))
+		return PTR_ERR(map);
+
+	map[offset / sizeof(*map)] = v;
+	i915_gem_object_unpin_map(obj);
+
+	return 0;
+}
+
+static int wc_get(struct drm_i915_gem_object *obj,
+		  unsigned long offset,
+		  u32 *v)
+{
+	typeof(v) map;
+	int err;
+
+	/* XXX WC write followed by GTT write go missing */
+	i915_gem_object_flush_gtt_write_domain(obj);
+
+	err = i915_gem_object_set_to_gtt_domain(obj, false);
+	if (err)
+		return err;
+
+	map = i915_gem_object_pin_map(obj, I915_MAP_WC);
+	if (IS_ERR(map))
+		return PTR_ERR(map);
+
+	*v = map[offset / sizeof(*map)];
+	i915_gem_object_unpin_map(obj);
+
+	return 0;
+}
+
+static int gpu_set(struct drm_i915_gem_object *obj,
+		   unsigned long offset,
+		   u32 v)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct drm_i915_gem_request *rq;
+	struct i915_vma *vma;
+	u32 *cs;
+	int err;
+
+	err = i915_gem_object_set_to_gtt_domain(obj, true);
+	if (err)
+		return err;
+
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, 0);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	rq = i915_gem_request_alloc(i915->engine[RCS], i915->kernel_context);
+	if (IS_ERR(rq)) {
+		i915_vma_unpin(vma);
+		return PTR_ERR(rq);
+	}
+
+	cs = intel_ring_begin(rq, 4);
+	if (IS_ERR(cs)) {
+		__i915_add_request(rq, false);
+		i915_vma_unpin(vma);
+		return PTR_ERR(cs);
+	}
+
+	if (INTEL_GEN(i915) >= 8) {
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | 1 << 22;
+		*cs++ = lower_32_bits(i915_ggtt_offset(vma) + offset);
+		*cs++ = upper_32_bits(i915_ggtt_offset(vma) + offset);
+		*cs++ = v;
+	} else if (INTEL_GEN(i915) >= 4) {
+		*cs++ = MI_STORE_DWORD_IMM_GEN4 | 1 << 22;
+		*cs++ = 0;
+		*cs++ = i915_ggtt_offset(vma) + offset;
+		*cs++ = v;
+	} else {
+		*cs++ = MI_STORE_DWORD_IMM | 1 << 22;
+		*cs++ = i915_ggtt_offset(vma) + offset;
+		*cs++ = v;
+		*cs++ = MI_NOOP;
+	}
+	intel_ring_advance(rq, cs);
+
+	i915_vma_move_to_active(vma, rq, EXEC_OBJECT_WRITE);
+	i915_vma_unpin(vma);
+
+	reservation_object_lock(obj->resv, NULL);
+	reservation_object_add_excl_fence(obj->resv, &rq->fence);
+	reservation_object_unlock(obj->resv);
+
+	__i915_add_request(rq, true);
+
+	return 0;
+}
+
+static bool always_valid(struct drm_i915_private *i915)
+{
+	return true;
+}
+
+static bool needs_mi_store_dword(struct drm_i915_private *i915)
+{
+	return igt_can_mi_store_dword_imm(i915);
+}
+
+static const struct igt_coherency_mode {
+	const char *name;
+	int (*set)(struct drm_i915_gem_object *, unsigned long offset, u32 v);
+	int (*get)(struct drm_i915_gem_object *, unsigned long offset, u32 *v);
+	bool (*valid)(struct drm_i915_private *i915);
+} igt_coherency_mode[] = {
+	{ "cpu", cpu_set, cpu_get, always_valid },
+	{ "gtt", gtt_set, gtt_get, always_valid },
+	{ "wc", wc_set, wc_get, always_valid },
+	{ "gpu", gpu_set, NULL, needs_mi_store_dword },
+	{ },
+};
+
+static int igt_gem_coherency(void *arg)
+{
+	const unsigned int ncachelines = PAGE_SIZE/64;
+	I915_RND_STATE(prng);
+	struct drm_i915_private *i915 = arg;
+	const struct igt_coherency_mode *read, *write, *over;
+	struct drm_i915_gem_object *obj;
+	unsigned long count, n;
+	u32 *offsets, *values;
+	int err = 0;
+
+	/* We repeatedly write, overwrite and read from a sequence of
+	 * cachelines in order to try and detect incoherency (unflushed writes
+	 * from either the CPU or GPU). Each setter/getter uses our cache
+	 * domain API which should prevent incoherency.
+	 */
+
+	offsets = kmalloc_array(ncachelines, 2*sizeof(u32), GFP_KERNEL);
+	if (!offsets)
+		return -ENOMEM;
+	for (count = 0; count < ncachelines; count++)
+		offsets[count] = count * 64 + 4 * (count % 16);
+
+	values = offsets + ncachelines;
+
+	mutex_lock(&i915->drm.struct_mutex);
+	for (over = igt_coherency_mode; over->name; over++) {
+		if (!over->set)
+			continue;
+
+		if (!over->valid(i915))
+			continue;
+
+		for (write = igt_coherency_mode; write->name; write++) {
+			if (!write->set)
+				continue;
+
+			if (!write->valid(i915))
+				continue;
+
+			for (read = igt_coherency_mode; read->name; read++) {
+				if (!read->get)
+					continue;
+
+				if (!read->valid(i915))
+					continue;
+
+				for_each_prime_number_from(count, 1, ncachelines) {
+					obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+					if (IS_ERR(obj)) {
+						err = PTR_ERR(obj);
+						goto unlock;
+					}
+
+					i915_random_reorder(offsets, ncachelines, &prng);
+					for (n = 0; n < count; n++)
+						values[n] = prandom_u32_state(&prng);
+
+					for (n = 0; n < count; n++) {
+						err = over->set(obj, offsets[n], ~values[n]);
+						if (err) {
+							pr_err("Failed to set stale value[%ld/%ld] in object using %s, err=%d\n",
+							       n, count, over->name, err);
+							goto put_object;
+						}
+					}
+
+					for (n = 0; n < count; n++) {
+						err = write->set(obj, offsets[n], values[n]);
+						if (err) {
+							pr_err("Failed to set value[%ld/%ld] in object using %s, err=%d\n",
+							       n, count, write->name, err);
+							goto put_object;
+						}
+					}
+
+					for (n = 0; n < count; n++) {
+						u32 found;
+
+						err = read->get(obj, offsets[n], &found);
+						if (err) {
+							pr_err("Failed to get value[%ld/%ld] in object using %s, err=%d\n",
+							       n, count, read->name, err);
+							goto put_object;
+						}
+
+						if (found != values[n]) {
+							pr_err("Value[%ld/%ld] mismatch, (overwrite with %s) wrote [%s] %x read [%s] %x (inverse %x), at offset %x\n",
+							       n, count, over->name,
+							       write->name, values[n],
+							       read->name, found,
+							       ~values[n], offsets[n]);
+							err = -EINVAL;
+							goto put_object;
+						}
+					}
+
+					__i915_gem_object_release_unless_active(obj);
+				}
+			}
+		}
+	}
+unlock:
+	mutex_unlock(&i915->drm.struct_mutex);
+	kfree(offsets);
+	return err;
+
+put_object:
+	__i915_gem_object_release_unless_active(obj);
+	goto unlock;
+}
+
+int i915_gem_coherency_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_gem_coherency),
+	};
+
+	return i915_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_context.c b/drivers/gpu/drm/i915/selftests/i915_gem_context.c
new file mode 100644
index 000000000000..3813a19a6179
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_context.c
@@ -0,0 +1,459 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "../i915_selftest.h"
+
+#include "mock_drm.h"
+#include "huge_gem_object.h"
+
+#define DW_PER_PAGE (PAGE_SIZE / sizeof(u32))
+
+static struct i915_vma *
+gpu_fill_dw(struct i915_vma *vma, u64 offset, unsigned long count, u32 value)
+{
+	struct drm_i915_gem_object *obj;
+	const int gen = INTEL_GEN(vma->vm->i915);
+	unsigned long n, size;
+	u32 *cmd;
+	int err;
+
+	GEM_BUG_ON(!igt_can_mi_store_dword_imm(vma->vm->i915));
+
+	size = (4 * count + 1) * sizeof(u32);
+	size = round_up(size, PAGE_SIZE);
+	obj = i915_gem_object_create_internal(vma->vm->i915, size);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	cmd = i915_gem_object_pin_map(obj, I915_MAP_WB);
+	if (IS_ERR(cmd)) {
+		err = PTR_ERR(cmd);
+		goto err;
+	}
+
+	GEM_BUG_ON(offset + (count - 1) * PAGE_SIZE > vma->node.size);
+	offset += vma->node.start;
+
+	for (n = 0; n < count; n++) {
+		if (gen >= 8) {
+			*cmd++ = MI_STORE_DWORD_IMM_GEN4;
+			*cmd++ = lower_32_bits(offset);
+			*cmd++ = upper_32_bits(offset);
+			*cmd++ = value;
+		} else if (gen >= 4) {
+			*cmd++ = MI_STORE_DWORD_IMM_GEN4 |
+				(gen < 6 ? 1 << 22 : 0);
+			*cmd++ = 0;
+			*cmd++ = offset;
+			*cmd++ = value;
+		} else {
+			*cmd++ = MI_STORE_DWORD_IMM | 1 << 22;
+			*cmd++ = offset;
+			*cmd++ = value;
+		}
+		offset += PAGE_SIZE;
+	}
+	*cmd = MI_BATCH_BUFFER_END;
+	i915_gem_object_unpin_map(obj);
+
+	err = i915_gem_object_set_to_gtt_domain(obj, false);
+	if (err)
+		goto err;
+
+	vma = i915_vma_instance(obj, vma->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto err;
+
+	return vma;
+
+err:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+static unsigned long real_page_count(struct drm_i915_gem_object *obj)
+{
+	return huge_gem_object_phys_size(obj) >> PAGE_SHIFT;
+}
+
+static unsigned long fake_page_count(struct drm_i915_gem_object *obj)
+{
+	return huge_gem_object_dma_size(obj) >> PAGE_SHIFT;
+}
+
+static int gpu_fill(struct drm_i915_gem_object *obj,
+		    struct i915_gem_context *ctx,
+		    struct intel_engine_cs *engine,
+		    unsigned int dw)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct i915_address_space *vm =
+		ctx->ppgtt ? &ctx->ppgtt->base : &i915->ggtt.base;
+	struct drm_i915_gem_request *rq;
+	struct i915_vma *vma;
+	struct i915_vma *batch;
+	unsigned int flags;
+	int err;
+
+	GEM_BUG_ON(obj->base.size > vm->total);
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	err = i915_gem_object_set_to_gtt_domain(obj, false);
+	if (err)
+		return err;
+
+	err = i915_vma_pin(vma, 0, 0, PIN_HIGH | PIN_USER);
+	if (err)
+		return err;
+
+	/* Within the GTT the huge objects maps every page onto
+	 * its 1024 real pages (using phys_pfn = dma_pfn % 1024).
+	 * We set the nth dword within the page using the nth
+	 * mapping via the GTT - this should exercise the GTT mapping
+	 * whilst checking that each context provides a unique view
+	 * into the object.
+	 */
+	batch = gpu_fill_dw(vma,
+			    (dw * real_page_count(obj)) << PAGE_SHIFT |
+			    (dw * sizeof(u32)),
+			    real_page_count(obj),
+			    dw);
+	if (IS_ERR(batch)) {
+		err = PTR_ERR(batch);
+		goto err_vma;
+	}
+
+	rq = i915_gem_request_alloc(engine, ctx);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto err_batch;
+	}
+
+	err = engine->emit_flush(rq, EMIT_INVALIDATE);
+	if (err)
+		goto err_request;
+
+	err = i915_switch_context(rq);
+	if (err)
+		goto err_request;
+
+	flags = 0;
+	if (INTEL_GEN(vm->i915) <= 5)
+		flags |= I915_DISPATCH_SECURE;
+
+	err = engine->emit_bb_start(rq,
+				    batch->node.start, batch->node.size,
+				    flags);
+	if (err)
+		goto err_request;
+
+	i915_vma_move_to_active(batch, rq, 0);
+	i915_gem_object_set_active_reference(batch->obj);
+	i915_vma_unpin(batch);
+	i915_vma_close(batch);
+
+	i915_vma_move_to_active(vma, rq, 0);
+	i915_vma_unpin(vma);
+
+	reservation_object_lock(obj->resv, NULL);
+	reservation_object_add_excl_fence(obj->resv, &rq->fence);
+	reservation_object_unlock(obj->resv);
+
+	__i915_add_request(rq, true);
+
+	return 0;
+
+err_request:
+	__i915_add_request(rq, false);
+err_batch:
+	i915_vma_unpin(batch);
+err_vma:
+	i915_vma_unpin(vma);
+	return err;
+}
+
+static int cpu_fill(struct drm_i915_gem_object *obj, u32 value)
+{
+	const bool has_llc = HAS_LLC(to_i915(obj->base.dev));
+	unsigned int n, m, need_flush;
+	int err;
+
+	err = i915_gem_obj_prepare_shmem_write(obj, &need_flush);
+	if (err)
+		return err;
+
+	for (n = 0; n < real_page_count(obj); n++) {
+		u32 *map;
+
+		map = kmap_atomic(i915_gem_object_get_page(obj, n));
+		for (m = 0; m < DW_PER_PAGE; m++)
+			map[m] = value;
+		if (!has_llc)
+			drm_clflush_virt_range(map, PAGE_SIZE);
+		kunmap_atomic(map);
+	}
+
+	i915_gem_obj_finish_shmem_access(obj);
+	obj->base.read_domains = I915_GEM_DOMAIN_GTT | I915_GEM_DOMAIN_CPU;
+	obj->base.write_domain = 0;
+	return 0;
+}
+
+static int cpu_check(struct drm_i915_gem_object *obj, unsigned int max)
+{
+	unsigned int n, m, needs_flush;
+	int err;
+
+	err = i915_gem_obj_prepare_shmem_read(obj, &needs_flush);
+	if (err)
+		return err;
+
+	for (n = 0; n < real_page_count(obj); n++) {
+		u32 *map;
+
+		map = kmap_atomic(i915_gem_object_get_page(obj, n));
+		if (needs_flush & CLFLUSH_BEFORE)
+			drm_clflush_virt_range(map, PAGE_SIZE);
+
+		for (m = 0; m < max; m++) {
+			if (map[m] != m) {
+				pr_err("Invalid value at page %d, offset %d: found %x expected %x\n",
+				       n, m, map[m], m);
+				err = -EINVAL;
+				goto out_unmap;
+			}
+		}
+
+		for (; m < DW_PER_PAGE; m++) {
+			if (map[m] != 0xdeadbeef) {
+				pr_err("Invalid value at page %d, offset %d: found %x expected %x\n",
+				       n, m, map[m], 0xdeadbeef);
+				err = -EINVAL;
+				goto out_unmap;
+			}
+		}
+
+out_unmap:
+		kunmap_atomic(map);
+		if (err)
+			break;
+	}
+
+	i915_gem_obj_finish_shmem_access(obj);
+	return err;
+}
+
+static struct drm_i915_gem_object *
+create_test_object(struct i915_gem_context *ctx,
+		   struct drm_file *file,
+		   struct list_head *objects)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_address_space *vm =
+		ctx->ppgtt ? &ctx->ppgtt->base : &ctx->i915->ggtt.base;
+	u64 size;
+	u32 handle;
+	int err;
+
+	size = min(vm->total / 2, 1024ull * DW_PER_PAGE * PAGE_SIZE);
+	size = round_down(size, DW_PER_PAGE * PAGE_SIZE);
+
+	obj = huge_gem_object(ctx->i915, DW_PER_PAGE * PAGE_SIZE, size);
+	if (IS_ERR(obj))
+		return obj;
+
+	/* tie the handle to the drm_file for easy reaping */
+	err = drm_gem_handle_create(file, &obj->base, &handle);
+	i915_gem_object_put(obj);
+	if (err)
+		return ERR_PTR(err);
+
+	err = cpu_fill(obj, 0xdeadbeef);
+	if (err) {
+		pr_err("Failed to fill object with cpu, err=%d\n",
+		       err);
+		return ERR_PTR(err);
+	}
+
+	list_add_tail(&obj->st_link, objects);
+	return obj;
+}
+
+static unsigned long max_dwords(struct drm_i915_gem_object *obj)
+{
+	unsigned long npages = fake_page_count(obj);
+
+	GEM_BUG_ON(!IS_ALIGNED(npages, DW_PER_PAGE));
+	return npages / DW_PER_PAGE;
+}
+
+static int igt_ctx_exec(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_file *file = mock_file(i915);
+	struct drm_i915_gem_object *obj;
+	IGT_TIMEOUT(end_time);
+	LIST_HEAD(objects);
+	unsigned long ncontexts, ndwords, dw;
+	bool first_shared_gtt = true;
+	int err;
+
+	/* Create a few different contexts (with different mm) and write
+	 * through each ctx/mm using the GPU making sure those writes end
+	 * up in the expected pages of our obj.
+	 */
+
+	mutex_lock(&i915->drm.struct_mutex);
+
+	ncontexts = 0;
+	ndwords = 0;
+	dw = 0;
+	while (!time_after(jiffies, end_time)) {
+		struct intel_engine_cs *engine;
+		struct i915_gem_context *ctx;
+		unsigned int id;
+
+		if (first_shared_gtt) {
+			ctx = __create_hw_context(i915, file->driver_priv);
+			first_shared_gtt = false;
+		} else {
+			ctx = i915_gem_create_context(i915, file->driver_priv);
+		}
+		if (IS_ERR(ctx)) {
+			err = PTR_ERR(ctx);
+			goto out_unlock;
+		}
+
+		for_each_engine(engine, i915, id) {
+			if (dw == 0) {
+				obj = create_test_object(ctx, file, &objects);
+				if (IS_ERR(obj)) {
+					err = PTR_ERR(obj);
+					goto out_unlock;
+				}
+			}
+
+			err = gpu_fill(obj, ctx, engine, dw);
+			if (err) {
+				pr_err("Failed to fill dword %lu [%lu/%lu] with gpu (%s) in ctx %u [full-ppgtt? %s], err=%d\n",
+				       ndwords, dw, max_dwords(obj),
+				       engine->name, ctx->hw_id,
+				       yesno(!!ctx->ppgtt), err);
+				goto out_unlock;
+			}
+
+			if (++dw == max_dwords(obj))
+				dw = 0;
+			ndwords++;
+		}
+		ncontexts++;
+	}
+	pr_info("Submitted %lu contexts (across %u engines), filling %lu dwords\n",
+		ncontexts, INTEL_INFO(i915)->num_rings, ndwords);
+
+	dw = 0;
+	list_for_each_entry(obj, &objects, st_link) {
+		unsigned int rem =
+			min_t(unsigned int, ndwords - dw, max_dwords(obj));
+
+		err = cpu_check(obj, rem);
+		if (err)
+			break;
+
+		dw += rem;
+	}
+
+out_unlock:
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	mock_file_free(i915, file);
+	return err;
+}
+
+static int fake_aliasing_ppgtt_enable(struct drm_i915_private *i915)
+{
+	struct drm_i915_gem_object *obj;
+	int err;
+
+	err = i915_gem_init_aliasing_ppgtt(i915);
+	if (err)
+		return err;
+
+	list_for_each_entry(obj, &i915->mm.bound_list, global_link) {
+		struct i915_vma *vma;
+
+		vma = i915_vma_instance(obj, &i915->ggtt.base, NULL);
+		if (IS_ERR(vma))
+			continue;
+
+		vma->flags &= ~I915_VMA_LOCAL_BIND;
+	}
+
+	return 0;
+}
+
+static void fake_aliasing_ppgtt_disable(struct drm_i915_private *i915)
+{
+	i915_gem_fini_aliasing_ppgtt(i915);
+}
+
+int i915_gem_context_live_selftests(struct drm_i915_private *dev_priv)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_ctx_exec),
+	};
+	bool fake_alias = false;
+	int err;
+
+	/* Install a fake aliasing gtt for exercise */
+	if (USES_PPGTT(dev_priv) && !dev_priv->mm.aliasing_ppgtt) {
+		mutex_lock(&dev_priv->drm.struct_mutex);
+		err = fake_aliasing_ppgtt_enable(dev_priv);
+		mutex_unlock(&dev_priv->drm.struct_mutex);
+		if (err)
+			return err;
+
+		GEM_BUG_ON(!dev_priv->mm.aliasing_ppgtt);
+		fake_alias = true;
+	}
+
+	err = i915_subtests(tests, dev_priv);
+
+	if (fake_alias) {
+		mutex_lock(&dev_priv->drm.struct_mutex);
+		fake_aliasing_ppgtt_disable(dev_priv);
+		mutex_unlock(&dev_priv->drm.struct_mutex);
+	}
+
+	return err;
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_dmabuf.c b/drivers/gpu/drm/i915/selftests/i915_gem_dmabuf.c
new file mode 100644
index 000000000000..817bef74bbcb
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_dmabuf.c
@@ -0,0 +1,303 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "../i915_selftest.h"
+
+#include "mock_gem_device.h"
+#include "mock_dmabuf.h"
+
+static int igt_dmabuf_export(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	struct dma_buf *dmabuf;
+
+	obj = i915_gem_object_create(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	dmabuf = i915_gem_prime_export(&i915->drm, &obj->base, 0);
+	i915_gem_object_put(obj);
+	if (IS_ERR(dmabuf)) {
+		pr_err("i915_gem_prime_export failed with err=%d\n",
+		       (int)PTR_ERR(dmabuf));
+		return PTR_ERR(dmabuf);
+	}
+
+	dma_buf_put(dmabuf);
+	return 0;
+}
+
+static int igt_dmabuf_import_self(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	struct drm_gem_object *import;
+	struct dma_buf *dmabuf;
+	int err;
+
+	obj = i915_gem_object_create(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	dmabuf = i915_gem_prime_export(&i915->drm, &obj->base, 0);
+	if (IS_ERR(dmabuf)) {
+		pr_err("i915_gem_prime_export failed with err=%d\n",
+		       (int)PTR_ERR(dmabuf));
+		err = PTR_ERR(dmabuf);
+		goto out;
+	}
+
+	import = i915_gem_prime_import(&i915->drm, dmabuf);
+	if (IS_ERR(import)) {
+		pr_err("i915_gem_prime_import failed with err=%d\n",
+		       (int)PTR_ERR(import));
+		err = PTR_ERR(import);
+		goto out_dmabuf;
+	}
+
+	if (import != &obj->base) {
+		pr_err("i915_gem_prime_import created a new object!\n");
+		err = -EINVAL;
+		goto out_import;
+	}
+
+	err = 0;
+out_import:
+	i915_gem_object_put(to_intel_bo(import));
+out_dmabuf:
+	dma_buf_put(dmabuf);
+out:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static int igt_dmabuf_import(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	struct dma_buf *dmabuf;
+	void *obj_map, *dma_map;
+	u32 pattern[] = { 0, 0xaa, 0xcc, 0x55, 0xff };
+	int err, i;
+
+	dmabuf = mock_dmabuf(1);
+	if (IS_ERR(dmabuf))
+		return PTR_ERR(dmabuf);
+
+	obj = to_intel_bo(i915_gem_prime_import(&i915->drm, dmabuf));
+	if (IS_ERR(obj)) {
+		pr_err("i915_gem_prime_import failed with err=%d\n",
+		       (int)PTR_ERR(obj));
+		err = PTR_ERR(obj);
+		goto out_dmabuf;
+	}
+
+	if (obj->base.dev != &i915->drm) {
+		pr_err("i915_gem_prime_import created a non-i915 object!\n");
+		err = -EINVAL;
+		goto out_obj;
+	}
+
+	if (obj->base.size != PAGE_SIZE) {
+		pr_err("i915_gem_prime_import is wrong size found %lld, expected %ld\n",
+		       (long long)obj->base.size, PAGE_SIZE);
+		err = -EINVAL;
+		goto out_obj;
+	}
+
+	dma_map = dma_buf_vmap(dmabuf);
+	if (!dma_map) {
+		pr_err("dma_buf_vmap failed\n");
+		err = -ENOMEM;
+		goto out_obj;
+	}
+
+	if (0) { /* Can not yet map dmabuf */
+		obj_map = i915_gem_object_pin_map(obj, I915_MAP_WB);
+		if (IS_ERR(obj_map)) {
+			err = PTR_ERR(obj_map);
+			pr_err("i915_gem_object_pin_map failed with err=%d\n", err);
+			goto out_dma_map;
+		}
+
+		for (i = 0; i < ARRAY_SIZE(pattern); i++) {
+			memset(dma_map, pattern[i], PAGE_SIZE);
+			if (memchr_inv(obj_map, pattern[i], PAGE_SIZE)) {
+				err = -EINVAL;
+				pr_err("imported vmap not all set to %x!\n", pattern[i]);
+				i915_gem_object_unpin_map(obj);
+				goto out_dma_map;
+			}
+		}
+
+		for (i = 0; i < ARRAY_SIZE(pattern); i++) {
+			memset(obj_map, pattern[i], PAGE_SIZE);
+			if (memchr_inv(dma_map, pattern[i], PAGE_SIZE)) {
+				err = -EINVAL;
+				pr_err("exported vmap not all set to %x!\n", pattern[i]);
+				i915_gem_object_unpin_map(obj);
+				goto out_dma_map;
+			}
+		}
+
+		i915_gem_object_unpin_map(obj);
+	}
+
+	err = 0;
+out_dma_map:
+	dma_buf_vunmap(dmabuf, dma_map);
+out_obj:
+	i915_gem_object_put(obj);
+out_dmabuf:
+	dma_buf_put(dmabuf);
+	return err;
+}
+
+static int igt_dmabuf_import_ownership(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	struct dma_buf *dmabuf;
+	void *ptr;
+	int err;
+
+	dmabuf = mock_dmabuf(1);
+	if (IS_ERR(dmabuf))
+		return PTR_ERR(dmabuf);
+
+	ptr = dma_buf_vmap(dmabuf);
+	if (!ptr) {
+		pr_err("dma_buf_vmap failed\n");
+		err = -ENOMEM;
+		goto err_dmabuf;
+	}
+
+	memset(ptr, 0xc5, PAGE_SIZE);
+	dma_buf_vunmap(dmabuf, ptr);
+
+	obj = to_intel_bo(i915_gem_prime_import(&i915->drm, dmabuf));
+	if (IS_ERR(obj)) {
+		pr_err("i915_gem_prime_import failed with err=%d\n",
+		       (int)PTR_ERR(obj));
+		err = PTR_ERR(obj);
+		goto err_dmabuf;
+	}
+
+	dma_buf_put(dmabuf);
+
+	err = i915_gem_object_pin_pages(obj);
+	if (err) {
+		pr_err("i915_gem_object_pin_pages failed with err=%d\n", err);
+		goto out_obj;
+	}
+
+	err = 0;
+	i915_gem_object_unpin_pages(obj);
+out_obj:
+	i915_gem_object_put(obj);
+	return err;
+
+err_dmabuf:
+	dma_buf_put(dmabuf);
+	return err;
+}
+
+static int igt_dmabuf_export_vmap(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	struct dma_buf *dmabuf;
+	void *ptr;
+	int err;
+
+	obj = i915_gem_object_create(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	dmabuf = i915_gem_prime_export(&i915->drm, &obj->base, 0);
+	if (IS_ERR(dmabuf)) {
+		pr_err("i915_gem_prime_export failed with err=%d\n",
+		       (int)PTR_ERR(dmabuf));
+		err = PTR_ERR(dmabuf);
+		goto err_obj;
+	}
+	i915_gem_object_put(obj);
+
+	ptr = dma_buf_vmap(dmabuf);
+	if (IS_ERR(ptr)) {
+		err = PTR_ERR(ptr);
+		pr_err("dma_buf_vmap failed with err=%d\n", err);
+		goto out;
+	}
+
+	if (memchr_inv(ptr, 0, dmabuf->size)) {
+		pr_err("Exported object not initialiased to zero!\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	memset(ptr, 0xc5, dmabuf->size);
+
+	err = 0;
+	dma_buf_vunmap(dmabuf, ptr);
+out:
+	dma_buf_put(dmabuf);
+	return err;
+
+err_obj:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+int i915_gem_dmabuf_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_dmabuf_export),
+		SUBTEST(igt_dmabuf_import_self),
+		SUBTEST(igt_dmabuf_import),
+		SUBTEST(igt_dmabuf_import_ownership),
+		SUBTEST(igt_dmabuf_export_vmap),
+	};
+	struct drm_i915_private *i915;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	err = i915_subtests(tests, i915);
+
+	drm_dev_unref(&i915->drm);
+	return err;
+}
+
+int i915_gem_dmabuf_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_dmabuf_export),
+	};
+
+	return i915_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_evict.c b/drivers/gpu/drm/i915/selftests/i915_gem_evict.c
new file mode 100644
index 000000000000..97af353db218
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_evict.c
@@ -0,0 +1,260 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "../i915_selftest.h"
+
+#include "mock_gem_device.h"
+
+static int populate_ggtt(struct drm_i915_private *i915)
+{
+	struct drm_i915_gem_object *obj;
+	u64 size;
+
+	for (size = 0;
+	     size + I915_GTT_PAGE_SIZE <= i915->ggtt.base.total;
+	     size += I915_GTT_PAGE_SIZE) {
+		struct i915_vma *vma;
+
+		obj = i915_gem_object_create_internal(i915, I915_GTT_PAGE_SIZE);
+		if (IS_ERR(obj))
+			return PTR_ERR(obj);
+
+		vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, 0);
+		if (IS_ERR(vma))
+			return PTR_ERR(vma);
+	}
+
+	if (!list_empty(&i915->mm.unbound_list)) {
+		size = 0;
+		list_for_each_entry(obj, &i915->mm.unbound_list, global_link)
+			size++;
+
+		pr_err("Found %lld objects unbound!\n", size);
+		return -EINVAL;
+	}
+
+	if (list_empty(&i915->ggtt.base.inactive_list)) {
+		pr_err("No objects on the GGTT inactive list!\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void unpin_ggtt(struct drm_i915_private *i915)
+{
+	struct i915_vma *vma;
+
+	list_for_each_entry(vma, &i915->ggtt.base.inactive_list, vm_link)
+		i915_vma_unpin(vma);
+}
+
+static void cleanup_objects(struct drm_i915_private *i915)
+{
+	struct drm_i915_gem_object *obj, *on;
+
+	list_for_each_entry_safe(obj, on, &i915->mm.unbound_list, global_link)
+		i915_gem_object_put(obj);
+
+	list_for_each_entry_safe(obj, on, &i915->mm.bound_list, global_link)
+		i915_gem_object_put(obj);
+
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	i915_gem_drain_freed_objects(i915);
+
+	mutex_lock(&i915->drm.struct_mutex);
+}
+
+static int igt_evict_something(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_ggtt *ggtt = &i915->ggtt;
+	int err;
+
+	/* Fill the GGTT with pinned objects and try to evict one. */
+
+	err = populate_ggtt(i915);
+	if (err)
+		goto cleanup;
+
+	/* Everything is pinned, nothing should happen */
+	err = i915_gem_evict_something(&ggtt->base,
+				       I915_GTT_PAGE_SIZE, 0, 0,
+				       0, U64_MAX,
+				       0);
+	if (err != -ENOSPC) {
+		pr_err("i915_gem_evict_something failed on a full GGTT with err=%d\n",
+		       err);
+		goto cleanup;
+	}
+
+	unpin_ggtt(i915);
+
+	/* Everything is unpinned, we should be able to evict something */
+	err = i915_gem_evict_something(&ggtt->base,
+				       I915_GTT_PAGE_SIZE, 0, 0,
+				       0, U64_MAX,
+				       0);
+	if (err) {
+		pr_err("i915_gem_evict_something failed on a full GGTT with err=%d\n",
+		       err);
+		goto cleanup;
+	}
+
+cleanup:
+	cleanup_objects(i915);
+	return err;
+}
+
+static int igt_overcommit(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err;
+
+	/* Fill the GGTT with pinned objects and then try to pin one more.
+	 * We expect it to fail.
+	 */
+
+	err = populate_ggtt(i915);
+	if (err)
+		goto cleanup;
+
+	obj = i915_gem_object_create_internal(i915, I915_GTT_PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto cleanup;
+	}
+
+	list_move(&obj->global_link, &i915->mm.unbound_list);
+
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, 0);
+	if (!IS_ERR(vma) || PTR_ERR(vma) != -ENOSPC) {
+		pr_err("Failed to evict+insert, i915_gem_object_ggtt_pin returned err=%d\n", (int)PTR_ERR(vma));
+		err = -EINVAL;
+		goto cleanup;
+	}
+
+cleanup:
+	cleanup_objects(i915);
+	return err;
+}
+
+static int igt_evict_for_vma(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_ggtt *ggtt = &i915->ggtt;
+	struct drm_mm_node target = {
+		.start = 0,
+		.size = 4096,
+	};
+	int err;
+
+	/* Fill the GGTT with pinned objects and try to evict a range. */
+
+	err = populate_ggtt(i915);
+	if (err)
+		goto cleanup;
+
+	/* Everything is pinned, nothing should happen */
+	err = i915_gem_evict_for_node(&ggtt->base, &target, 0);
+	if (err != -ENOSPC) {
+		pr_err("i915_gem_evict_for_node on a full GGTT returned err=%d\n",
+		       err);
+		goto cleanup;
+	}
+
+	unpin_ggtt(i915);
+
+	/* Everything is unpinned, we should be able to evict the node */
+	err = i915_gem_evict_for_node(&ggtt->base, &target, 0);
+	if (err) {
+		pr_err("i915_gem_evict_for_node returned err=%d\n",
+		       err);
+		goto cleanup;
+	}
+
+cleanup:
+	cleanup_objects(i915);
+	return err;
+}
+
+static int igt_evict_vm(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_ggtt *ggtt = &i915->ggtt;
+	int err;
+
+	/* Fill the GGTT with pinned objects and try to evict everything. */
+
+	err = populate_ggtt(i915);
+	if (err)
+		goto cleanup;
+
+	/* Everything is pinned, nothing should happen */
+	err = i915_gem_evict_vm(&ggtt->base, false);
+	if (err) {
+		pr_err("i915_gem_evict_vm on a full GGTT returned err=%d]\n",
+		       err);
+		goto cleanup;
+	}
+
+	unpin_ggtt(i915);
+
+	err = i915_gem_evict_vm(&ggtt->base, false);
+	if (err) {
+		pr_err("i915_gem_evict_vm on a full GGTT returned err=%d]\n",
+		       err);
+		goto cleanup;
+	}
+
+cleanup:
+	cleanup_objects(i915);
+	return err;
+}
+
+int i915_gem_evict_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_evict_something),
+		SUBTEST(igt_evict_for_vma),
+		SUBTEST(igt_evict_vm),
+		SUBTEST(igt_overcommit),
+	};
+	struct drm_i915_private *i915;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	mutex_lock(&i915->drm.struct_mutex);
+	err = i915_subtests(tests, i915);
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	drm_dev_unref(&i915->drm);
+	return err;
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
new file mode 100644
index 000000000000..0f3fa34377c6
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_gtt.c
@@ -0,0 +1,1556 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/list_sort.h>
+#include <linux/prime_numbers.h>
+
+#include "../i915_selftest.h"
+#include "i915_random.h"
+
+#include "mock_context.h"
+#include "mock_drm.h"
+#include "mock_gem_device.h"
+
+static void fake_free_pages(struct drm_i915_gem_object *obj,
+			    struct sg_table *pages)
+{
+	sg_free_table(pages);
+	kfree(pages);
+}
+
+static struct sg_table *
+fake_get_pages(struct drm_i915_gem_object *obj)
+{
+#define GFP (GFP_KERNEL | __GFP_NOWARN | __GFP_NORETRY)
+#define PFN_BIAS 0x1000
+	struct sg_table *pages;
+	struct scatterlist *sg;
+	typeof(obj->base.size) rem;
+
+	pages = kmalloc(sizeof(*pages), GFP);
+	if (!pages)
+		return ERR_PTR(-ENOMEM);
+
+	rem = round_up(obj->base.size, BIT(31)) >> 31;
+	if (sg_alloc_table(pages, rem, GFP)) {
+		kfree(pages);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	rem = obj->base.size;
+	for (sg = pages->sgl; sg; sg = sg_next(sg)) {
+		unsigned long len = min_t(typeof(rem), rem, BIT(31));
+
+		GEM_BUG_ON(!len);
+		sg_set_page(sg, pfn_to_page(PFN_BIAS), len, 0);
+		sg_dma_address(sg) = page_to_phys(sg_page(sg));
+		sg_dma_len(sg) = len;
+
+		rem -= len;
+	}
+	GEM_BUG_ON(rem);
+
+	obj->mm.madv = I915_MADV_DONTNEED;
+	return pages;
+#undef GFP
+}
+
+static void fake_put_pages(struct drm_i915_gem_object *obj,
+			   struct sg_table *pages)
+{
+	fake_free_pages(obj, pages);
+	obj->mm.dirty = false;
+	obj->mm.madv = I915_MADV_WILLNEED;
+}
+
+static const struct drm_i915_gem_object_ops fake_ops = {
+	.flags = I915_GEM_OBJECT_IS_SHRINKABLE,
+	.get_pages = fake_get_pages,
+	.put_pages = fake_put_pages,
+};
+
+static struct drm_i915_gem_object *
+fake_dma_object(struct drm_i915_private *i915, u64 size)
+{
+	struct drm_i915_gem_object *obj;
+
+	GEM_BUG_ON(!size);
+	GEM_BUG_ON(!IS_ALIGNED(size, I915_GTT_PAGE_SIZE));
+
+	if (overflows_type(size, obj->base.size))
+		return ERR_PTR(-E2BIG);
+
+	obj = i915_gem_object_alloc(i915);
+	if (!obj)
+		return ERR_PTR(-ENOMEM);
+
+	drm_gem_private_object_init(&i915->drm, &obj->base, size);
+	i915_gem_object_init(obj, &fake_ops);
+
+	obj->base.write_domain = I915_GEM_DOMAIN_CPU;
+	obj->base.read_domains = I915_GEM_DOMAIN_CPU;
+	obj->cache_level = I915_CACHE_NONE;
+
+	/* Preallocate the "backing storage" */
+	if (i915_gem_object_pin_pages(obj))
+		return ERR_PTR(-ENOMEM);
+
+	i915_gem_object_unpin_pages(obj);
+	return obj;
+}
+
+static int igt_ppgtt_alloc(void *arg)
+{
+	struct drm_i915_private *dev_priv = arg;
+	struct i915_hw_ppgtt *ppgtt;
+	u64 size, last;
+	int err;
+
+	/* Allocate a ppggt and try to fill the entire range */
+
+	if (!USES_PPGTT(dev_priv))
+		return 0;
+
+	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
+	if (!ppgtt)
+		return -ENOMEM;
+
+	mutex_lock(&dev_priv->drm.struct_mutex);
+	err = __hw_ppgtt_init(ppgtt, dev_priv);
+	if (err)
+		goto err_ppgtt;
+
+	if (!ppgtt->base.allocate_va_range)
+		goto err_ppgtt_cleanup;
+
+	/* Check we can allocate the entire range */
+	for (size = 4096;
+	     size <= ppgtt->base.total;
+	     size <<= 2) {
+		err = ppgtt->base.allocate_va_range(&ppgtt->base, 0, size);
+		if (err) {
+			if (err == -ENOMEM) {
+				pr_info("[1] Ran out of memory for va_range [0 + %llx] [bit %d]\n",
+					size, ilog2(size));
+				err = 0; /* virtual space too large! */
+			}
+			goto err_ppgtt_cleanup;
+		}
+
+		ppgtt->base.clear_range(&ppgtt->base, 0, size);
+	}
+
+	/* Check we can incrementally allocate the entire range */
+	for (last = 0, size = 4096;
+	     size <= ppgtt->base.total;
+	     last = size, size <<= 2) {
+		err = ppgtt->base.allocate_va_range(&ppgtt->base,
+						    last, size - last);
+		if (err) {
+			if (err == -ENOMEM) {
+				pr_info("[2] Ran out of memory for va_range [%llx + %llx] [bit %d]\n",
+					last, size - last, ilog2(size));
+				err = 0; /* virtual space too large! */
+			}
+			goto err_ppgtt_cleanup;
+		}
+	}
+
+err_ppgtt_cleanup:
+	ppgtt->base.cleanup(&ppgtt->base);
+err_ppgtt:
+	mutex_unlock(&dev_priv->drm.struct_mutex);
+	kfree(ppgtt);
+	return err;
+}
+
+static int lowlevel_hole(struct drm_i915_private *i915,
+			 struct i915_address_space *vm,
+			 u64 hole_start, u64 hole_end,
+			 unsigned long end_time)
+{
+	I915_RND_STATE(seed_prng);
+	unsigned int size;
+
+	/* Keep creating larger objects until one cannot fit into the hole */
+	for (size = 12; (hole_end - hole_start) >> size; size++) {
+		I915_RND_SUBSTATE(prng, seed_prng);
+		struct drm_i915_gem_object *obj;
+		unsigned int *order, count, n;
+		u64 hole_size;
+
+		hole_size = (hole_end - hole_start) >> size;
+		if (hole_size > KMALLOC_MAX_SIZE / sizeof(u32))
+			hole_size = KMALLOC_MAX_SIZE / sizeof(u32);
+		count = hole_size;
+		do {
+			count >>= 1;
+			order = i915_random_order(count, &prng);
+		} while (!order && count);
+		if (!order)
+			break;
+
+		GEM_BUG_ON(count * BIT_ULL(size) > vm->total);
+		GEM_BUG_ON(hole_start + count * BIT_ULL(size) > hole_end);
+
+		/* Ignore allocation failures (i.e. don't report them as
+		 * a test failure) as we are purposefully allocating very
+		 * large objects without checking that we have sufficient
+		 * memory. We expect to hit -ENOMEM.
+		 */
+
+		obj = fake_dma_object(i915, BIT_ULL(size));
+		if (IS_ERR(obj)) {
+			kfree(order);
+			break;
+		}
+
+		GEM_BUG_ON(obj->base.size != BIT_ULL(size));
+
+		if (i915_gem_object_pin_pages(obj)) {
+			i915_gem_object_put(obj);
+			kfree(order);
+			break;
+		}
+
+		for (n = 0; n < count; n++) {
+			u64 addr = hole_start + order[n] * BIT_ULL(size);
+
+			GEM_BUG_ON(addr + BIT_ULL(size) > vm->total);
+
+			if (igt_timeout(end_time,
+					"%s timed out before %d/%d\n",
+					__func__, n, count)) {
+				hole_end = hole_start; /* quit */
+				break;
+			}
+
+			if (vm->allocate_va_range &&
+			    vm->allocate_va_range(vm, addr, BIT_ULL(size)))
+				break;
+
+			vm->insert_entries(vm, obj->mm.pages, addr,
+					   I915_CACHE_NONE, 0);
+		}
+		count = n;
+
+		i915_random_reorder(order, count, &prng);
+		for (n = 0; n < count; n++) {
+			u64 addr = hole_start + order[n] * BIT_ULL(size);
+
+			GEM_BUG_ON(addr + BIT_ULL(size) > vm->total);
+			vm->clear_range(vm, addr, BIT_ULL(size));
+		}
+
+		i915_gem_object_unpin_pages(obj);
+		i915_gem_object_put(obj);
+
+		kfree(order);
+	}
+
+	return 0;
+}
+
+static void close_object_list(struct list_head *objects,
+			      struct i915_address_space *vm)
+{
+	struct drm_i915_gem_object *obj, *on;
+	int ignored;
+
+	list_for_each_entry_safe(obj, on, objects, st_link) {
+		struct i915_vma *vma;
+
+		vma = i915_vma_instance(obj, vm, NULL);
+		if (!IS_ERR(vma))
+			ignored = i915_vma_unbind(vma);
+		/* Only ppgtt vma may be closed before the object is freed */
+		if (!IS_ERR(vma) && !i915_vma_is_ggtt(vma))
+			i915_vma_close(vma);
+
+		list_del(&obj->st_link);
+		i915_gem_object_put(obj);
+	}
+}
+
+static int fill_hole(struct drm_i915_private *i915,
+		     struct i915_address_space *vm,
+		     u64 hole_start, u64 hole_end,
+		     unsigned long end_time)
+{
+	const u64 hole_size = hole_end - hole_start;
+	struct drm_i915_gem_object *obj;
+	const unsigned long max_pages =
+		min_t(u64, ULONG_MAX - 1, hole_size/2 >> PAGE_SHIFT);
+	const unsigned long max_step = max(int_sqrt(max_pages), 2UL);
+	unsigned long npages, prime, flags;
+	struct i915_vma *vma;
+	LIST_HEAD(objects);
+	int err;
+
+	/* Try binding many VMA working inwards from either edge */
+
+	flags = PIN_OFFSET_FIXED | PIN_USER;
+	if (i915_is_ggtt(vm))
+		flags |= PIN_GLOBAL;
+
+	for_each_prime_number_from(prime, 2, max_step) {
+		for (npages = 1; npages <= max_pages; npages *= prime) {
+			const u64 full_size = npages << PAGE_SHIFT;
+			const struct {
+				const char *name;
+				u64 offset;
+				int step;
+			} phases[] = {
+				{ "top-down", hole_end, -1, },
+				{ "bottom-up", hole_start, 1, },
+				{ }
+			}, *p;
+
+			obj = fake_dma_object(i915, full_size);
+			if (IS_ERR(obj))
+				break;
+
+			list_add(&obj->st_link, &objects);
+
+			/* Align differing sized objects against the edges, and
+			 * check we don't walk off into the void when binding
+			 * them into the GTT.
+			 */
+			for (p = phases; p->name; p++) {
+				u64 offset;
+
+				offset = p->offset;
+				list_for_each_entry(obj, &objects, st_link) {
+					vma = i915_vma_instance(obj, vm, NULL);
+					if (IS_ERR(vma))
+						continue;
+
+					if (p->step < 0) {
+						if (offset < hole_start + obj->base.size)
+							break;
+						offset -= obj->base.size;
+					}
+
+					err = i915_vma_pin(vma, 0, 0, offset | flags);
+					if (err) {
+						pr_err("%s(%s) pin (forward) failed with err=%d on size=%lu pages (prime=%lu), offset=%llx\n",
+						       __func__, p->name, err, npages, prime, offset);
+						goto err;
+					}
+
+					if (!drm_mm_node_allocated(&vma->node) ||
+					    i915_vma_misplaced(vma, 0, 0, offset | flags)) {
+						pr_err("%s(%s) (forward) insert failed: vma.node=%llx + %llx [allocated? %d], expected offset %llx\n",
+						       __func__, p->name, vma->node.start, vma->node.size, drm_mm_node_allocated(&vma->node),
+						       offset);
+						err = -EINVAL;
+						goto err;
+					}
+
+					i915_vma_unpin(vma);
+
+					if (p->step > 0) {
+						if (offset + obj->base.size > hole_end)
+							break;
+						offset += obj->base.size;
+					}
+				}
+
+				offset = p->offset;
+				list_for_each_entry(obj, &objects, st_link) {
+					vma = i915_vma_instance(obj, vm, NULL);
+					if (IS_ERR(vma))
+						continue;
+
+					if (p->step < 0) {
+						if (offset < hole_start + obj->base.size)
+							break;
+						offset -= obj->base.size;
+					}
+
+					if (!drm_mm_node_allocated(&vma->node) ||
+					    i915_vma_misplaced(vma, 0, 0, offset | flags)) {
+						pr_err("%s(%s) (forward) moved vma.node=%llx + %llx, expected offset %llx\n",
+						       __func__, p->name, vma->node.start, vma->node.size,
+						       offset);
+						err = -EINVAL;
+						goto err;
+					}
+
+					err = i915_vma_unbind(vma);
+					if (err) {
+						pr_err("%s(%s) (forward) unbind of vma.node=%llx + %llx failed with err=%d\n",
+						       __func__, p->name, vma->node.start, vma->node.size,
+						       err);
+						goto err;
+					}
+
+					if (p->step > 0) {
+						if (offset + obj->base.size > hole_end)
+							break;
+						offset += obj->base.size;
+					}
+				}
+
+				offset = p->offset;
+				list_for_each_entry_reverse(obj, &objects, st_link) {
+					vma = i915_vma_instance(obj, vm, NULL);
+					if (IS_ERR(vma))
+						continue;
+
+					if (p->step < 0) {
+						if (offset < hole_start + obj->base.size)
+							break;
+						offset -= obj->base.size;
+					}
+
+					err = i915_vma_pin(vma, 0, 0, offset | flags);
+					if (err) {
+						pr_err("%s(%s) pin (backward) failed with err=%d on size=%lu pages (prime=%lu), offset=%llx\n",
+						       __func__, p->name, err, npages, prime, offset);
+						goto err;
+					}
+
+					if (!drm_mm_node_allocated(&vma->node) ||
+					    i915_vma_misplaced(vma, 0, 0, offset | flags)) {
+						pr_err("%s(%s) (backward) insert failed: vma.node=%llx + %llx [allocated? %d], expected offset %llx\n",
+						       __func__, p->name, vma->node.start, vma->node.size, drm_mm_node_allocated(&vma->node),
+						       offset);
+						err = -EINVAL;
+						goto err;
+					}
+
+					i915_vma_unpin(vma);
+
+					if (p->step > 0) {
+						if (offset + obj->base.size > hole_end)
+							break;
+						offset += obj->base.size;
+					}
+				}
+
+				offset = p->offset;
+				list_for_each_entry_reverse(obj, &objects, st_link) {
+					vma = i915_vma_instance(obj, vm, NULL);
+					if (IS_ERR(vma))
+						continue;
+
+					if (p->step < 0) {
+						if (offset < hole_start + obj->base.size)
+							break;
+						offset -= obj->base.size;
+					}
+
+					if (!drm_mm_node_allocated(&vma->node) ||
+					    i915_vma_misplaced(vma, 0, 0, offset | flags)) {
+						pr_err("%s(%s) (backward) moved vma.node=%llx + %llx [allocated? %d], expected offset %llx\n",
+						       __func__, p->name, vma->node.start, vma->node.size, drm_mm_node_allocated(&vma->node),
+						       offset);
+						err = -EINVAL;
+						goto err;
+					}
+
+					err = i915_vma_unbind(vma);
+					if (err) {
+						pr_err("%s(%s) (backward) unbind of vma.node=%llx + %llx failed with err=%d\n",
+						       __func__, p->name, vma->node.start, vma->node.size,
+						       err);
+						goto err;
+					}
+
+					if (p->step > 0) {
+						if (offset + obj->base.size > hole_end)
+							break;
+						offset += obj->base.size;
+					}
+				}
+			}
+
+			if (igt_timeout(end_time, "%s timed out (npages=%lu, prime=%lu)\n",
+					__func__, npages, prime)) {
+				err = -EINTR;
+				goto err;
+			}
+		}
+
+		close_object_list(&objects, vm);
+	}
+
+	return 0;
+
+err:
+	close_object_list(&objects, vm);
+	return err;
+}
+
+static int walk_hole(struct drm_i915_private *i915,
+		     struct i915_address_space *vm,
+		     u64 hole_start, u64 hole_end,
+		     unsigned long end_time)
+{
+	const u64 hole_size = hole_end - hole_start;
+	const unsigned long max_pages =
+		min_t(u64, ULONG_MAX - 1, hole_size >> PAGE_SHIFT);
+	unsigned long flags;
+	u64 size;
+
+	/* Try binding a single VMA in different positions within the hole */
+
+	flags = PIN_OFFSET_FIXED | PIN_USER;
+	if (i915_is_ggtt(vm))
+		flags |= PIN_GLOBAL;
+
+	for_each_prime_number_from(size, 1, max_pages) {
+		struct drm_i915_gem_object *obj;
+		struct i915_vma *vma;
+		u64 addr;
+		int err = 0;
+
+		obj = fake_dma_object(i915, size << PAGE_SHIFT);
+		if (IS_ERR(obj))
+			break;
+
+		vma = i915_vma_instance(obj, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto err;
+		}
+
+		for (addr = hole_start;
+		     addr + obj->base.size < hole_end;
+		     addr += obj->base.size) {
+			err = i915_vma_pin(vma, 0, 0, addr | flags);
+			if (err) {
+				pr_err("%s bind failed at %llx + %llx [hole %llx- %llx] with err=%d\n",
+				       __func__, addr, vma->size,
+				       hole_start, hole_end, err);
+				goto err;
+			}
+			i915_vma_unpin(vma);
+
+			if (!drm_mm_node_allocated(&vma->node) ||
+			    i915_vma_misplaced(vma, 0, 0, addr | flags)) {
+				pr_err("%s incorrect at %llx + %llx\n",
+				       __func__, addr, vma->size);
+				err = -EINVAL;
+				goto err;
+			}
+
+			err = i915_vma_unbind(vma);
+			if (err) {
+				pr_err("%s unbind failed at %llx + %llx  with err=%d\n",
+				       __func__, addr, vma->size, err);
+				goto err;
+			}
+
+			GEM_BUG_ON(drm_mm_node_allocated(&vma->node));
+
+			if (igt_timeout(end_time,
+					"%s timed out at %llx\n",
+					__func__, addr)) {
+				err = -EINTR;
+				goto err;
+			}
+		}
+
+err:
+		if (!i915_vma_is_ggtt(vma))
+			i915_vma_close(vma);
+		i915_gem_object_put(obj);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int pot_hole(struct drm_i915_private *i915,
+		    struct i915_address_space *vm,
+		    u64 hole_start, u64 hole_end,
+		    unsigned long end_time)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	unsigned long flags;
+	unsigned int pot;
+	int err = 0;
+
+	flags = PIN_OFFSET_FIXED | PIN_USER;
+	if (i915_is_ggtt(vm))
+		flags |= PIN_GLOBAL;
+
+	obj = i915_gem_object_create_internal(i915, 2 * I915_GTT_PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_obj;
+	}
+
+	/* Insert a pair of pages across every pot boundary within the hole */
+	for (pot = fls64(hole_end - 1) - 1;
+	     pot > ilog2(2 * I915_GTT_PAGE_SIZE);
+	     pot--) {
+		u64 step = BIT_ULL(pot);
+		u64 addr;
+
+		for (addr = round_up(hole_start + I915_GTT_PAGE_SIZE, step) - I915_GTT_PAGE_SIZE;
+		     addr <= round_down(hole_end - 2*I915_GTT_PAGE_SIZE, step) - I915_GTT_PAGE_SIZE;
+		     addr += step) {
+			err = i915_vma_pin(vma, 0, 0, addr | flags);
+			if (err) {
+				pr_err("%s failed to pin object at %llx in hole [%llx - %llx], with err=%d\n",
+				       __func__,
+				       addr,
+				       hole_start, hole_end,
+				       err);
+				goto err;
+			}
+
+			if (!drm_mm_node_allocated(&vma->node) ||
+			    i915_vma_misplaced(vma, 0, 0, addr | flags)) {
+				pr_err("%s incorrect at %llx + %llx\n",
+				       __func__, addr, vma->size);
+				i915_vma_unpin(vma);
+				err = i915_vma_unbind(vma);
+				err = -EINVAL;
+				goto err;
+			}
+
+			i915_vma_unpin(vma);
+			err = i915_vma_unbind(vma);
+			GEM_BUG_ON(err);
+		}
+
+		if (igt_timeout(end_time,
+				"%s timed out after %d/%d\n",
+				__func__, pot, fls64(hole_end - 1) - 1)) {
+			err = -EINTR;
+			goto err;
+		}
+	}
+
+err:
+	if (!i915_vma_is_ggtt(vma))
+		i915_vma_close(vma);
+err_obj:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static int drunk_hole(struct drm_i915_private *i915,
+		      struct i915_address_space *vm,
+		      u64 hole_start, u64 hole_end,
+		      unsigned long end_time)
+{
+	I915_RND_STATE(prng);
+	unsigned int size;
+	unsigned long flags;
+
+	flags = PIN_OFFSET_FIXED | PIN_USER;
+	if (i915_is_ggtt(vm))
+		flags |= PIN_GLOBAL;
+
+	/* Keep creating larger objects until one cannot fit into the hole */
+	for (size = 12; (hole_end - hole_start) >> size; size++) {
+		struct drm_i915_gem_object *obj;
+		unsigned int *order, count, n;
+		struct i915_vma *vma;
+		u64 hole_size;
+		int err;
+
+		hole_size = (hole_end - hole_start) >> size;
+		if (hole_size > KMALLOC_MAX_SIZE / sizeof(u32))
+			hole_size = KMALLOC_MAX_SIZE / sizeof(u32);
+		count = hole_size;
+		do {
+			count >>= 1;
+			order = i915_random_order(count, &prng);
+		} while (!order && count);
+		if (!order)
+			break;
+
+		/* Ignore allocation failures (i.e. don't report them as
+		 * a test failure) as we are purposefully allocating very
+		 * large objects without checking that we have sufficient
+		 * memory. We expect to hit -ENOMEM.
+		 */
+
+		obj = fake_dma_object(i915, BIT_ULL(size));
+		if (IS_ERR(obj)) {
+			kfree(order);
+			break;
+		}
+
+		vma = i915_vma_instance(obj, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto err_obj;
+		}
+
+		GEM_BUG_ON(vma->size != BIT_ULL(size));
+
+		for (n = 0; n < count; n++) {
+			u64 addr = hole_start + order[n] * BIT_ULL(size);
+
+			err = i915_vma_pin(vma, 0, 0, addr | flags);
+			if (err) {
+				pr_err("%s failed to pin object at %llx + %llx in hole [%llx - %llx], with err=%d\n",
+				       __func__,
+				       addr, BIT_ULL(size),
+				       hole_start, hole_end,
+				       err);
+				goto err;
+			}
+
+			if (!drm_mm_node_allocated(&vma->node) ||
+			    i915_vma_misplaced(vma, 0, 0, addr | flags)) {
+				pr_err("%s incorrect at %llx + %llx\n",
+				       __func__, addr, BIT_ULL(size));
+				i915_vma_unpin(vma);
+				err = i915_vma_unbind(vma);
+				err = -EINVAL;
+				goto err;
+			}
+
+			i915_vma_unpin(vma);
+			err = i915_vma_unbind(vma);
+			GEM_BUG_ON(err);
+
+			if (igt_timeout(end_time,
+					"%s timed out after %d/%d\n",
+					__func__, n, count)) {
+				err = -EINTR;
+				goto err;
+			}
+		}
+
+err:
+		if (!i915_vma_is_ggtt(vma))
+			i915_vma_close(vma);
+err_obj:
+		i915_gem_object_put(obj);
+		kfree(order);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int __shrink_hole(struct drm_i915_private *i915,
+			 struct i915_address_space *vm,
+			 u64 hole_start, u64 hole_end,
+			 unsigned long end_time)
+{
+	struct drm_i915_gem_object *obj;
+	unsigned long flags = PIN_OFFSET_FIXED | PIN_USER;
+	unsigned int order = 12;
+	LIST_HEAD(objects);
+	int err = 0;
+	u64 addr;
+
+	/* Keep creating larger objects until one cannot fit into the hole */
+	for (addr = hole_start; addr < hole_end; ) {
+		struct i915_vma *vma;
+		u64 size = BIT_ULL(order++);
+
+		size = min(size, hole_end - addr);
+		obj = fake_dma_object(i915, size);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			break;
+		}
+
+		list_add(&obj->st_link, &objects);
+
+		vma = i915_vma_instance(obj, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			break;
+		}
+
+		GEM_BUG_ON(vma->size != size);
+
+		err = i915_vma_pin(vma, 0, 0, addr | flags);
+		if (err) {
+			pr_err("%s failed to pin object at %llx + %llx in hole [%llx - %llx], with err=%d\n",
+			       __func__, addr, size, hole_start, hole_end, err);
+			break;
+		}
+
+		if (!drm_mm_node_allocated(&vma->node) ||
+		    i915_vma_misplaced(vma, 0, 0, addr | flags)) {
+			pr_err("%s incorrect at %llx + %llx\n",
+			       __func__, addr, size);
+			i915_vma_unpin(vma);
+			err = i915_vma_unbind(vma);
+			err = -EINVAL;
+			break;
+		}
+
+		i915_vma_unpin(vma);
+		addr += size;
+
+		if (igt_timeout(end_time,
+				"%s timed out at ofset %llx [%llx - %llx]\n",
+				__func__, addr, hole_start, hole_end)) {
+			err = -EINTR;
+			break;
+		}
+	}
+
+	close_object_list(&objects, vm);
+	return err;
+}
+
+static int shrink_hole(struct drm_i915_private *i915,
+		       struct i915_address_space *vm,
+		       u64 hole_start, u64 hole_end,
+		       unsigned long end_time)
+{
+	unsigned long prime;
+	int err;
+
+	vm->fault_attr.probability = 999;
+	atomic_set(&vm->fault_attr.times, -1);
+
+	for_each_prime_number_from(prime, 0, ULONG_MAX - 1) {
+		vm->fault_attr.interval = prime;
+		err = __shrink_hole(i915, vm, hole_start, hole_end, end_time);
+		if (err)
+			break;
+	}
+
+	memset(&vm->fault_attr, 0, sizeof(vm->fault_attr));
+
+	return err;
+}
+
+static int exercise_ppgtt(struct drm_i915_private *dev_priv,
+			  int (*func)(struct drm_i915_private *i915,
+				      struct i915_address_space *vm,
+				      u64 hole_start, u64 hole_end,
+				      unsigned long end_time))
+{
+	struct drm_file *file;
+	struct i915_hw_ppgtt *ppgtt;
+	IGT_TIMEOUT(end_time);
+	int err;
+
+	if (!USES_FULL_PPGTT(dev_priv))
+		return 0;
+
+	file = mock_file(dev_priv);
+	if (IS_ERR(file))
+		return PTR_ERR(file);
+
+	mutex_lock(&dev_priv->drm.struct_mutex);
+	ppgtt = i915_ppgtt_create(dev_priv, file->driver_priv, "mock");
+	if (IS_ERR(ppgtt)) {
+		err = PTR_ERR(ppgtt);
+		goto out_unlock;
+	}
+	GEM_BUG_ON(offset_in_page(ppgtt->base.total));
+	GEM_BUG_ON(ppgtt->base.closed);
+
+	err = func(dev_priv, &ppgtt->base, 0, ppgtt->base.total, end_time);
+
+	i915_ppgtt_close(&ppgtt->base);
+	i915_ppgtt_put(ppgtt);
+out_unlock:
+	mutex_unlock(&dev_priv->drm.struct_mutex);
+
+	mock_file_free(dev_priv, file);
+	return err;
+}
+
+static int igt_ppgtt_fill(void *arg)
+{
+	return exercise_ppgtt(arg, fill_hole);
+}
+
+static int igt_ppgtt_walk(void *arg)
+{
+	return exercise_ppgtt(arg, walk_hole);
+}
+
+static int igt_ppgtt_pot(void *arg)
+{
+	return exercise_ppgtt(arg, pot_hole);
+}
+
+static int igt_ppgtt_drunk(void *arg)
+{
+	return exercise_ppgtt(arg, drunk_hole);
+}
+
+static int igt_ppgtt_lowlevel(void *arg)
+{
+	return exercise_ppgtt(arg, lowlevel_hole);
+}
+
+static int igt_ppgtt_shrink(void *arg)
+{
+	return exercise_ppgtt(arg, shrink_hole);
+}
+
+static int sort_holes(void *priv, struct list_head *A, struct list_head *B)
+{
+	struct drm_mm_node *a = list_entry(A, typeof(*a), hole_stack);
+	struct drm_mm_node *b = list_entry(B, typeof(*b), hole_stack);
+
+	if (a->start < b->start)
+		return -1;
+	else
+		return 1;
+}
+
+static int exercise_ggtt(struct drm_i915_private *i915,
+			 int (*func)(struct drm_i915_private *i915,
+				     struct i915_address_space *vm,
+				     u64 hole_start, u64 hole_end,
+				     unsigned long end_time))
+{
+	struct i915_ggtt *ggtt = &i915->ggtt;
+	u64 hole_start, hole_end, last = 0;
+	struct drm_mm_node *node;
+	IGT_TIMEOUT(end_time);
+	int err;
+
+	mutex_lock(&i915->drm.struct_mutex);
+restart:
+	list_sort(NULL, &ggtt->base.mm.hole_stack, sort_holes);
+	drm_mm_for_each_hole(node, &ggtt->base.mm, hole_start, hole_end) {
+		if (hole_start < last)
+			continue;
+
+		if (ggtt->base.mm.color_adjust)
+			ggtt->base.mm.color_adjust(node, 0,
+						   &hole_start, &hole_end);
+		if (hole_start >= hole_end)
+			continue;
+
+		err = func(i915, &ggtt->base, hole_start, hole_end, end_time);
+		if (err)
+			break;
+
+		/* As we have manipulated the drm_mm, the list may be corrupt */
+		last = hole_end;
+		goto restart;
+	}
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	return err;
+}
+
+static int igt_ggtt_fill(void *arg)
+{
+	return exercise_ggtt(arg, fill_hole);
+}
+
+static int igt_ggtt_walk(void *arg)
+{
+	return exercise_ggtt(arg, walk_hole);
+}
+
+static int igt_ggtt_pot(void *arg)
+{
+	return exercise_ggtt(arg, pot_hole);
+}
+
+static int igt_ggtt_drunk(void *arg)
+{
+	return exercise_ggtt(arg, drunk_hole);
+}
+
+static int igt_ggtt_lowlevel(void *arg)
+{
+	return exercise_ggtt(arg, lowlevel_hole);
+}
+
+static int igt_ggtt_page(void *arg)
+{
+	const unsigned int count = PAGE_SIZE/sizeof(u32);
+	I915_RND_STATE(prng);
+	struct drm_i915_private *i915 = arg;
+	struct i915_ggtt *ggtt = &i915->ggtt;
+	struct drm_i915_gem_object *obj;
+	struct drm_mm_node tmp;
+	unsigned int *order, n;
+	int err;
+
+	mutex_lock(&i915->drm.struct_mutex);
+
+	obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto out_unlock;
+	}
+
+	err = i915_gem_object_pin_pages(obj);
+	if (err)
+		goto out_free;
+
+	memset(&tmp, 0, sizeof(tmp));
+	err = drm_mm_insert_node_in_range(&ggtt->base.mm, &tmp,
+					  1024 * PAGE_SIZE, 0,
+					  I915_COLOR_UNEVICTABLE,
+					  0, ggtt->mappable_end,
+					  DRM_MM_INSERT_LOW);
+	if (err)
+		goto out_unpin;
+
+	order = i915_random_order(count, &prng);
+	if (!order) {
+		err = -ENOMEM;
+		goto out_remove;
+	}
+
+	for (n = 0; n < count; n++) {
+		u64 offset = tmp.start + order[n] * PAGE_SIZE;
+		u32 __iomem *vaddr;
+
+		ggtt->base.insert_page(&ggtt->base,
+				       i915_gem_object_get_dma_address(obj, 0),
+				       offset, I915_CACHE_NONE, 0);
+
+		vaddr = io_mapping_map_atomic_wc(&ggtt->mappable, offset);
+		iowrite32(n, vaddr + n);
+		io_mapping_unmap_atomic(vaddr);
+
+		wmb();
+		ggtt->base.clear_range(&ggtt->base, offset, PAGE_SIZE);
+	}
+
+	i915_random_reorder(order, count, &prng);
+	for (n = 0; n < count; n++) {
+		u64 offset = tmp.start + order[n] * PAGE_SIZE;
+		u32 __iomem *vaddr;
+		u32 val;
+
+		ggtt->base.insert_page(&ggtt->base,
+				       i915_gem_object_get_dma_address(obj, 0),
+				       offset, I915_CACHE_NONE, 0);
+
+		vaddr = io_mapping_map_atomic_wc(&ggtt->mappable, offset);
+		val = ioread32(vaddr + n);
+		io_mapping_unmap_atomic(vaddr);
+
+		ggtt->base.clear_range(&ggtt->base, offset, PAGE_SIZE);
+
+		if (val != n) {
+			pr_err("insert page failed: found %d, expected %d\n",
+			       val, n);
+			err = -EINVAL;
+			break;
+		}
+	}
+
+	kfree(order);
+out_remove:
+	drm_mm_remove_node(&tmp);
+out_unpin:
+	i915_gem_object_unpin_pages(obj);
+out_free:
+	i915_gem_object_put(obj);
+out_unlock:
+	mutex_unlock(&i915->drm.struct_mutex);
+	return err;
+}
+
+static void track_vma_bind(struct i915_vma *vma)
+{
+	struct drm_i915_gem_object *obj = vma->obj;
+
+	obj->bind_count++; /* track for eviction later */
+	__i915_gem_object_pin_pages(obj);
+
+	vma->pages = obj->mm.pages;
+	list_move_tail(&vma->vm_link, &vma->vm->inactive_list);
+}
+
+static int exercise_mock(struct drm_i915_private *i915,
+			 int (*func)(struct drm_i915_private *i915,
+				     struct i915_address_space *vm,
+				     u64 hole_start, u64 hole_end,
+				     unsigned long end_time))
+{
+	struct i915_gem_context *ctx;
+	struct i915_hw_ppgtt *ppgtt;
+	IGT_TIMEOUT(end_time);
+	int err;
+
+	ctx = mock_context(i915, "mock");
+	if (!ctx)
+		return -ENOMEM;
+
+	ppgtt = ctx->ppgtt;
+	GEM_BUG_ON(!ppgtt);
+
+	err = func(i915, &ppgtt->base, 0, ppgtt->base.total, end_time);
+
+	mock_context_close(ctx);
+	return err;
+}
+
+static int igt_mock_fill(void *arg)
+{
+	return exercise_mock(arg, fill_hole);
+}
+
+static int igt_mock_walk(void *arg)
+{
+	return exercise_mock(arg, walk_hole);
+}
+
+static int igt_mock_pot(void *arg)
+{
+	return exercise_mock(arg, pot_hole);
+}
+
+static int igt_mock_drunk(void *arg)
+{
+	return exercise_mock(arg, drunk_hole);
+}
+
+static int igt_gtt_reserve(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj, *on;
+	LIST_HEAD(objects);
+	u64 total;
+	int err;
+
+	/* i915_gem_gtt_reserve() tries to reserve the precise range
+	 * for the node, and evicts if it has to. So our test checks that
+	 * it can give us the requsted space and prevent overlaps.
+	 */
+
+	/* Start by filling the GGTT */
+	for (total = 0;
+	     total + 2*I915_GTT_PAGE_SIZE <= i915->ggtt.base.total;
+	     total += 2*I915_GTT_PAGE_SIZE) {
+		struct i915_vma *vma;
+
+		obj = i915_gem_object_create_internal(i915, 2*PAGE_SIZE);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			goto out;
+		}
+
+		err = i915_gem_object_pin_pages(obj);
+		if (err) {
+			i915_gem_object_put(obj);
+			goto out;
+		}
+
+		list_add(&obj->st_link, &objects);
+
+		vma = i915_vma_instance(obj, &i915->ggtt.base, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		err = i915_gem_gtt_reserve(&i915->ggtt.base, &vma->node,
+					   obj->base.size,
+					   total,
+					   obj->cache_level,
+					   0);
+		if (err) {
+			pr_err("i915_gem_gtt_reserve (pass 1) failed at %llu/%llu with err=%d\n",
+			       total, i915->ggtt.base.total, err);
+			goto out;
+		}
+		track_vma_bind(vma);
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+		if (vma->node.start != total ||
+		    vma->node.size != 2*I915_GTT_PAGE_SIZE) {
+			pr_err("i915_gem_gtt_reserve (pass 1) placement failed, found (%llx + %llx), expected (%llx + %lx)\n",
+			       vma->node.start, vma->node.size,
+			       total, 2*I915_GTT_PAGE_SIZE);
+			err = -EINVAL;
+			goto out;
+		}
+	}
+
+	/* Now we start forcing evictions */
+	for (total = I915_GTT_PAGE_SIZE;
+	     total + 2*I915_GTT_PAGE_SIZE <= i915->ggtt.base.total;
+	     total += 2*I915_GTT_PAGE_SIZE) {
+		struct i915_vma *vma;
+
+		obj = i915_gem_object_create_internal(i915, 2*PAGE_SIZE);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			goto out;
+		}
+
+		err = i915_gem_object_pin_pages(obj);
+		if (err) {
+			i915_gem_object_put(obj);
+			goto out;
+		}
+
+		list_add(&obj->st_link, &objects);
+
+		vma = i915_vma_instance(obj, &i915->ggtt.base, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		err = i915_gem_gtt_reserve(&i915->ggtt.base, &vma->node,
+					   obj->base.size,
+					   total,
+					   obj->cache_level,
+					   0);
+		if (err) {
+			pr_err("i915_gem_gtt_reserve (pass 2) failed at %llu/%llu with err=%d\n",
+			       total, i915->ggtt.base.total, err);
+			goto out;
+		}
+		track_vma_bind(vma);
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+		if (vma->node.start != total ||
+		    vma->node.size != 2*I915_GTT_PAGE_SIZE) {
+			pr_err("i915_gem_gtt_reserve (pass 2) placement failed, found (%llx + %llx), expected (%llx + %lx)\n",
+			       vma->node.start, vma->node.size,
+			       total, 2*I915_GTT_PAGE_SIZE);
+			err = -EINVAL;
+			goto out;
+		}
+	}
+
+	/* And then try at random */
+	list_for_each_entry_safe(obj, on, &objects, st_link) {
+		struct i915_vma *vma;
+		u64 offset;
+
+		vma = i915_vma_instance(obj, &i915->ggtt.base, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		err = i915_vma_unbind(vma);
+		if (err) {
+			pr_err("i915_vma_unbind failed with err=%d!\n", err);
+			goto out;
+		}
+
+		offset = random_offset(0, i915->ggtt.base.total,
+				       2*I915_GTT_PAGE_SIZE,
+				       I915_GTT_MIN_ALIGNMENT);
+
+		err = i915_gem_gtt_reserve(&i915->ggtt.base, &vma->node,
+					   obj->base.size,
+					   offset,
+					   obj->cache_level,
+					   0);
+		if (err) {
+			pr_err("i915_gem_gtt_reserve (pass 3) failed at %llu/%llu with err=%d\n",
+			       total, i915->ggtt.base.total, err);
+			goto out;
+		}
+		track_vma_bind(vma);
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+		if (vma->node.start != offset ||
+		    vma->node.size != 2*I915_GTT_PAGE_SIZE) {
+			pr_err("i915_gem_gtt_reserve (pass 3) placement failed, found (%llx + %llx), expected (%llx + %lx)\n",
+			       vma->node.start, vma->node.size,
+			       offset, 2*I915_GTT_PAGE_SIZE);
+			err = -EINVAL;
+			goto out;
+		}
+	}
+
+out:
+	list_for_each_entry_safe(obj, on, &objects, st_link) {
+		i915_gem_object_unpin_pages(obj);
+		i915_gem_object_put(obj);
+	}
+	return err;
+}
+
+static int igt_gtt_insert(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj, *on;
+	struct drm_mm_node tmp = {};
+	const struct invalid_insert {
+		u64 size;
+		u64 alignment;
+		u64 start, end;
+	} invalid_insert[] = {
+		{
+			i915->ggtt.base.total + I915_GTT_PAGE_SIZE, 0,
+			0, i915->ggtt.base.total,
+		},
+		{
+			2*I915_GTT_PAGE_SIZE, 0,
+			0, I915_GTT_PAGE_SIZE,
+		},
+		{
+			-(u64)I915_GTT_PAGE_SIZE, 0,
+			0, 4*I915_GTT_PAGE_SIZE,
+		},
+		{
+			-(u64)2*I915_GTT_PAGE_SIZE, 2*I915_GTT_PAGE_SIZE,
+			0, 4*I915_GTT_PAGE_SIZE,
+		},
+		{
+			I915_GTT_PAGE_SIZE, I915_GTT_MIN_ALIGNMENT << 1,
+			I915_GTT_MIN_ALIGNMENT, I915_GTT_MIN_ALIGNMENT << 1,
+		},
+		{}
+	}, *ii;
+	LIST_HEAD(objects);
+	u64 total;
+	int err;
+
+	/* i915_gem_gtt_insert() tries to allocate some free space in the GTT
+	 * to the node, evicting if required.
+	 */
+
+	/* Check a couple of obviously invalid requests */
+	for (ii = invalid_insert; ii->size; ii++) {
+		err = i915_gem_gtt_insert(&i915->ggtt.base, &tmp,
+					  ii->size, ii->alignment,
+					  I915_COLOR_UNEVICTABLE,
+					  ii->start, ii->end,
+					  0);
+		if (err != -ENOSPC) {
+			pr_err("Invalid i915_gem_gtt_insert(.size=%llx, .alignment=%llx, .start=%llx, .end=%llx) succeeded (err=%d)\n",
+			       ii->size, ii->alignment, ii->start, ii->end,
+			       err);
+			return -EINVAL;
+		}
+	}
+
+	/* Start by filling the GGTT */
+	for (total = 0;
+	     total + I915_GTT_PAGE_SIZE <= i915->ggtt.base.total;
+	     total += I915_GTT_PAGE_SIZE) {
+		struct i915_vma *vma;
+
+		obj = i915_gem_object_create_internal(i915, I915_GTT_PAGE_SIZE);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			goto out;
+		}
+
+		err = i915_gem_object_pin_pages(obj);
+		if (err) {
+			i915_gem_object_put(obj);
+			goto out;
+		}
+
+		list_add(&obj->st_link, &objects);
+
+		vma = i915_vma_instance(obj, &i915->ggtt.base, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		err = i915_gem_gtt_insert(&i915->ggtt.base, &vma->node,
+					  obj->base.size, 0, obj->cache_level,
+					  0, i915->ggtt.base.total,
+					  0);
+		if (err == -ENOSPC) {
+			/* maxed out the GGTT space */
+			i915_gem_object_put(obj);
+			break;
+		}
+		if (err) {
+			pr_err("i915_gem_gtt_insert (pass 1) failed at %llu/%llu with err=%d\n",
+			       total, i915->ggtt.base.total, err);
+			goto out;
+		}
+		track_vma_bind(vma);
+		__i915_vma_pin(vma);
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+	}
+
+	list_for_each_entry(obj, &objects, st_link) {
+		struct i915_vma *vma;
+
+		vma = i915_vma_instance(obj, &i915->ggtt.base, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		if (!drm_mm_node_allocated(&vma->node)) {
+			pr_err("VMA was unexpectedly evicted!\n");
+			err = -EINVAL;
+			goto out;
+		}
+
+		__i915_vma_unpin(vma);
+	}
+
+	/* If we then reinsert, we should find the same hole */
+	list_for_each_entry_safe(obj, on, &objects, st_link) {
+		struct i915_vma *vma;
+		u64 offset;
+
+		vma = i915_vma_instance(obj, &i915->ggtt.base, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+		offset = vma->node.start;
+
+		err = i915_vma_unbind(vma);
+		if (err) {
+			pr_err("i915_vma_unbind failed with err=%d!\n", err);
+			goto out;
+		}
+
+		err = i915_gem_gtt_insert(&i915->ggtt.base, &vma->node,
+					  obj->base.size, 0, obj->cache_level,
+					  0, i915->ggtt.base.total,
+					  0);
+		if (err) {
+			pr_err("i915_gem_gtt_insert (pass 2) failed at %llu/%llu with err=%d\n",
+			       total, i915->ggtt.base.total, err);
+			goto out;
+		}
+		track_vma_bind(vma);
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+		if (vma->node.start != offset) {
+			pr_err("i915_gem_gtt_insert did not return node to its previous location (the only hole), expected address %llx, found %llx\n",
+			       offset, vma->node.start);
+			err = -EINVAL;
+			goto out;
+		}
+	}
+
+	/* And then force evictions */
+	for (total = 0;
+	     total + 2*I915_GTT_PAGE_SIZE <= i915->ggtt.base.total;
+	     total += 2*I915_GTT_PAGE_SIZE) {
+		struct i915_vma *vma;
+
+		obj = i915_gem_object_create_internal(i915, 2*I915_GTT_PAGE_SIZE);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			goto out;
+		}
+
+		err = i915_gem_object_pin_pages(obj);
+		if (err) {
+			i915_gem_object_put(obj);
+			goto out;
+		}
+
+		list_add(&obj->st_link, &objects);
+
+		vma = i915_vma_instance(obj, &i915->ggtt.base, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out;
+		}
+
+		err = i915_gem_gtt_insert(&i915->ggtt.base, &vma->node,
+					  obj->base.size, 0, obj->cache_level,
+					  0, i915->ggtt.base.total,
+					  0);
+		if (err) {
+			pr_err("i915_gem_gtt_insert (pass 3) failed at %llu/%llu with err=%d\n",
+			       total, i915->ggtt.base.total, err);
+			goto out;
+		}
+		track_vma_bind(vma);
+
+		GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+	}
+
+out:
+	list_for_each_entry_safe(obj, on, &objects, st_link) {
+		i915_gem_object_unpin_pages(obj);
+		i915_gem_object_put(obj);
+	}
+	return err;
+}
+
+int i915_gem_gtt_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_mock_drunk),
+		SUBTEST(igt_mock_walk),
+		SUBTEST(igt_mock_pot),
+		SUBTEST(igt_mock_fill),
+		SUBTEST(igt_gtt_reserve),
+		SUBTEST(igt_gtt_insert),
+	};
+	struct drm_i915_private *i915;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	mutex_lock(&i915->drm.struct_mutex);
+	err = i915_subtests(tests, i915);
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	drm_dev_unref(&i915->drm);
+	return err;
+}
+
+int i915_gem_gtt_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_ppgtt_alloc),
+		SUBTEST(igt_ppgtt_lowlevel),
+		SUBTEST(igt_ppgtt_drunk),
+		SUBTEST(igt_ppgtt_walk),
+		SUBTEST(igt_ppgtt_pot),
+		SUBTEST(igt_ppgtt_fill),
+		SUBTEST(igt_ppgtt_shrink),
+		SUBTEST(igt_ggtt_lowlevel),
+		SUBTEST(igt_ggtt_drunk),
+		SUBTEST(igt_ggtt_walk),
+		SUBTEST(igt_ggtt_pot),
+		SUBTEST(igt_ggtt_fill),
+		SUBTEST(igt_ggtt_page),
+	};
+
+	GEM_BUG_ON(offset_in_page(i915->ggtt.base.total));
+
+	return i915_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_object.c b/drivers/gpu/drm/i915/selftests/i915_gem_object.c
new file mode 100644
index 000000000000..67d82bf1407f
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_object.c
@@ -0,0 +1,600 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "../i915_selftest.h"
+
+#include "mock_gem_device.h"
+#include "huge_gem_object.h"
+
+static int igt_gem_object(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	int err = -ENOMEM;
+
+	/* Basic test to ensure we can create an object */
+
+	obj = i915_gem_object_create(i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		pr_err("i915_gem_object_create failed, err=%d\n", err);
+		goto out;
+	}
+
+	err = 0;
+	i915_gem_object_put(obj);
+out:
+	return err;
+}
+
+static int igt_phys_object(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	int err;
+
+	/* Create an object and bind it to a contiguous set of physical pages,
+	 * i.e. exercise the i915_gem_object_phys API.
+	 */
+
+	obj = i915_gem_object_create(i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		pr_err("i915_gem_object_create failed, err=%d\n", err);
+		goto out;
+	}
+
+	mutex_lock(&i915->drm.struct_mutex);
+	err = i915_gem_object_attach_phys(obj, PAGE_SIZE);
+	mutex_unlock(&i915->drm.struct_mutex);
+	if (err) {
+		pr_err("i915_gem_object_attach_phys failed, err=%d\n", err);
+		goto out_obj;
+	}
+
+	if (obj->ops != &i915_gem_phys_ops) {
+		pr_err("i915_gem_object_attach_phys did not create a phys object\n");
+		err = -EINVAL;
+		goto out_obj;
+	}
+
+	if (!atomic_read(&obj->mm.pages_pin_count)) {
+		pr_err("i915_gem_object_attach_phys did not pin its phys pages\n");
+		err = -EINVAL;
+		goto out_obj;
+	}
+
+	/* Make the object dirty so that put_pages must do copy back the data */
+	mutex_lock(&i915->drm.struct_mutex);
+	err = i915_gem_object_set_to_gtt_domain(obj, true);
+	mutex_unlock(&i915->drm.struct_mutex);
+	if (err) {
+		pr_err("i915_gem_object_set_to_gtt_domain failed with err=%d\n",
+		       err);
+		goto out_obj;
+	}
+
+out_obj:
+	i915_gem_object_put(obj);
+out:
+	return err;
+}
+
+static int igt_gem_huge(void *arg)
+{
+	const unsigned int nreal = 509; /* just to be awkward */
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	unsigned int n;
+	int err;
+
+	/* Basic sanitycheck of our huge fake object allocation */
+
+	obj = huge_gem_object(i915,
+			      nreal * PAGE_SIZE,
+			      i915->ggtt.base.total + PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	err = i915_gem_object_pin_pages(obj);
+	if (err) {
+		pr_err("Failed to allocate %u pages (%lu total), err=%d\n",
+		       nreal, obj->base.size / PAGE_SIZE, err);
+		goto out;
+	}
+
+	for (n = 0; n < obj->base.size / PAGE_SIZE; n++) {
+		if (i915_gem_object_get_page(obj, n) !=
+		    i915_gem_object_get_page(obj, n % nreal)) {
+			pr_err("Page lookup mismatch at index %u [%u]\n",
+			       n, n % nreal);
+			err = -EINVAL;
+			goto out_unpin;
+		}
+	}
+
+out_unpin:
+	i915_gem_object_unpin_pages(obj);
+out:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+struct tile {
+	unsigned int width;
+	unsigned int height;
+	unsigned int stride;
+	unsigned int size;
+	unsigned int tiling;
+	unsigned int swizzle;
+};
+
+static u64 swizzle_bit(unsigned int bit, u64 offset)
+{
+	return (offset & BIT_ULL(bit)) >> (bit - 6);
+}
+
+static u64 tiled_offset(const struct tile *tile, u64 v)
+{
+	u64 x, y;
+
+	if (tile->tiling == I915_TILING_NONE)
+		return v;
+
+	y = div64_u64_rem(v, tile->stride, &x);
+	v = div64_u64_rem(y, tile->height, &y) * tile->stride * tile->height;
+
+	if (tile->tiling == I915_TILING_X) {
+		v += y * tile->width;
+		v += div64_u64_rem(x, tile->width, &x) << tile->size;
+		v += x;
+	} else {
+		const unsigned int ytile_span = 16;
+		const unsigned int ytile_height = 32 * ytile_span;
+
+		v += y * ytile_span;
+		v += div64_u64_rem(x, ytile_span, &x) * ytile_height;
+		v += x;
+	}
+
+	switch (tile->swizzle) {
+	case I915_BIT_6_SWIZZLE_9:
+		v ^= swizzle_bit(9, v);
+		break;
+	case I915_BIT_6_SWIZZLE_9_10:
+		v ^= swizzle_bit(9, v) ^ swizzle_bit(10, v);
+		break;
+	case I915_BIT_6_SWIZZLE_9_11:
+		v ^= swizzle_bit(9, v) ^ swizzle_bit(11, v);
+		break;
+	case I915_BIT_6_SWIZZLE_9_10_11:
+		v ^= swizzle_bit(9, v) ^ swizzle_bit(10, v) ^ swizzle_bit(11, v);
+		break;
+	}
+
+	return v;
+}
+
+static int check_partial_mapping(struct drm_i915_gem_object *obj,
+				 const struct tile *tile,
+				 unsigned long end_time)
+{
+	const unsigned int nreal = obj->scratch / PAGE_SIZE;
+	const unsigned long npages = obj->base.size / PAGE_SIZE;
+	struct i915_vma *vma;
+	unsigned long page;
+	int err;
+
+	if (igt_timeout(end_time,
+			"%s: timed out before tiling=%d stride=%d\n",
+			__func__, tile->tiling, tile->stride))
+		return -EINTR;
+
+	err = i915_gem_object_set_tiling(obj, tile->tiling, tile->stride);
+	if (err)
+		return err;
+
+	GEM_BUG_ON(i915_gem_object_get_tiling(obj) != tile->tiling);
+	GEM_BUG_ON(i915_gem_object_get_stride(obj) != tile->stride);
+
+	for_each_prime_number_from(page, 1, npages) {
+		struct i915_ggtt_view view =
+			compute_partial_view(obj, page, MIN_CHUNK_PAGES);
+		u32 __iomem *io;
+		struct page *p;
+		unsigned int n;
+		u64 offset;
+		u32 *cpu;
+
+		GEM_BUG_ON(view.partial.size > nreal);
+
+		err = i915_gem_object_set_to_gtt_domain(obj, true);
+		if (err)
+			return err;
+
+		vma = i915_gem_object_ggtt_pin(obj, &view, 0, 0, PIN_MAPPABLE);
+		if (IS_ERR(vma)) {
+			pr_err("Failed to pin partial view: offset=%lu\n",
+			       page);
+			return PTR_ERR(vma);
+		}
+
+		n = page - view.partial.offset;
+		GEM_BUG_ON(n >= view.partial.size);
+
+		io = i915_vma_pin_iomap(vma);
+		i915_vma_unpin(vma);
+		if (IS_ERR(io)) {
+			pr_err("Failed to iomap partial view: offset=%lu\n",
+			       page);
+			return PTR_ERR(io);
+		}
+
+		err = i915_vma_get_fence(vma);
+		if (err) {
+			pr_err("Failed to get fence for partial view: offset=%lu\n",
+			       page);
+			i915_vma_unpin_iomap(vma);
+			return err;
+		}
+
+		iowrite32(page, io + n * PAGE_SIZE/sizeof(*io));
+		i915_vma_unpin_iomap(vma);
+
+		offset = tiled_offset(tile, page << PAGE_SHIFT);
+		if (offset >= obj->base.size)
+			continue;
+
+		i915_gem_object_flush_gtt_write_domain(obj);
+
+		p = i915_gem_object_get_page(obj, offset >> PAGE_SHIFT);
+		cpu = kmap(p) + offset_in_page(offset);
+		drm_clflush_virt_range(cpu, sizeof(*cpu));
+		if (*cpu != (u32)page) {
+			pr_err("Partial view for %lu [%u] (offset=%llu, size=%u [%llu, row size %u], fence=%d, tiling=%d, stride=%d) misalignment, expected write to page (%llu + %u [0x%llx]) of 0x%x, found 0x%x\n",
+			       page, n,
+			       view.partial.offset,
+			       view.partial.size,
+			       vma->size >> PAGE_SHIFT,
+			       tile_row_pages(obj),
+			       vma->fence ? vma->fence->id : -1, tile->tiling, tile->stride,
+			       offset >> PAGE_SHIFT,
+			       (unsigned int)offset_in_page(offset),
+			       offset,
+			       (u32)page, *cpu);
+			err = -EINVAL;
+		}
+		*cpu = 0;
+		drm_clflush_virt_range(cpu, sizeof(*cpu));
+		kunmap(p);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+static int igt_partial_tiling(void *arg)
+{
+	const unsigned int nreal = 1 << 12; /* largest tile row x2 */
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj;
+	int tiling;
+	int err;
+
+	/* We want to check the page mapping and fencing of a large object
+	 * mmapped through the GTT. The object we create is larger than can
+	 * possibly be mmaped as a whole, and so we must use partial GGTT vma.
+	 * We then check that a write through each partial GGTT vma ends up
+	 * in the right set of pages within the object, and with the expected
+	 * tiling, which we verify by manual swizzling.
+	 */
+
+	obj = huge_gem_object(i915,
+			      nreal << PAGE_SHIFT,
+			      (1 + next_prime_number(i915->ggtt.base.total >> PAGE_SHIFT)) << PAGE_SHIFT);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	err = i915_gem_object_pin_pages(obj);
+	if (err) {
+		pr_err("Failed to allocate %u pages (%lu total), err=%d\n",
+		       nreal, obj->base.size / PAGE_SIZE, err);
+		goto out;
+	}
+
+	mutex_lock(&i915->drm.struct_mutex);
+
+	if (1) {
+		IGT_TIMEOUT(end);
+		struct tile tile;
+
+		tile.height = 1;
+		tile.width = 1;
+		tile.size = 0;
+		tile.stride = 0;
+		tile.swizzle = I915_BIT_6_SWIZZLE_NONE;
+		tile.tiling = I915_TILING_NONE;
+
+		err = check_partial_mapping(obj, &tile, end);
+		if (err && err != -EINTR)
+			goto out_unlock;
+	}
+
+	for (tiling = I915_TILING_X; tiling <= I915_TILING_Y; tiling++) {
+		IGT_TIMEOUT(end);
+		unsigned int max_pitch;
+		unsigned int pitch;
+		struct tile tile;
+
+		tile.tiling = tiling;
+		switch (tiling) {
+		case I915_TILING_X:
+			tile.swizzle = i915->mm.bit_6_swizzle_x;
+			break;
+		case I915_TILING_Y:
+			tile.swizzle = i915->mm.bit_6_swizzle_y;
+			break;
+		}
+
+		if (tile.swizzle == I915_BIT_6_SWIZZLE_UNKNOWN ||
+		    tile.swizzle == I915_BIT_6_SWIZZLE_9_10_17)
+			continue;
+
+		if (INTEL_GEN(i915) <= 2) {
+			tile.height = 16;
+			tile.width = 128;
+			tile.size = 11;
+		} else if (tile.tiling == I915_TILING_Y &&
+			   HAS_128_BYTE_Y_TILING(i915)) {
+			tile.height = 32;
+			tile.width = 128;
+			tile.size = 12;
+		} else {
+			tile.height = 8;
+			tile.width = 512;
+			tile.size = 12;
+		}
+
+		if (INTEL_GEN(i915) < 4)
+			max_pitch = 8192 / tile.width;
+		else if (INTEL_GEN(i915) < 7)
+			max_pitch = 128 * I965_FENCE_MAX_PITCH_VAL / tile.width;
+		else
+			max_pitch = 128 * GEN7_FENCE_MAX_PITCH_VAL / tile.width;
+
+		for (pitch = max_pitch; pitch; pitch >>= 1) {
+			tile.stride = tile.width * pitch;
+			err = check_partial_mapping(obj, &tile, end);
+			if (err == -EINTR)
+				goto next_tiling;
+			if (err)
+				goto out_unlock;
+
+			if (pitch > 2 && INTEL_GEN(i915) >= 4) {
+				tile.stride = tile.width * (pitch - 1);
+				err = check_partial_mapping(obj, &tile, end);
+				if (err == -EINTR)
+					goto next_tiling;
+				if (err)
+					goto out_unlock;
+			}
+
+			if (pitch < max_pitch && INTEL_GEN(i915) >= 4) {
+				tile.stride = tile.width * (pitch + 1);
+				err = check_partial_mapping(obj, &tile, end);
+				if (err == -EINTR)
+					goto next_tiling;
+				if (err)
+					goto out_unlock;
+			}
+		}
+
+		if (INTEL_GEN(i915) >= 4) {
+			for_each_prime_number(pitch, max_pitch) {
+				tile.stride = tile.width * pitch;
+				err = check_partial_mapping(obj, &tile, end);
+				if (err == -EINTR)
+					goto next_tiling;
+				if (err)
+					goto out_unlock;
+			}
+		}
+
+next_tiling: ;
+	}
+
+out_unlock:
+	mutex_unlock(&i915->drm.struct_mutex);
+	i915_gem_object_unpin_pages(obj);
+out:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static int make_obj_busy(struct drm_i915_gem_object *obj)
+{
+	struct drm_i915_private *i915 = to_i915(obj->base.dev);
+	struct drm_i915_gem_request *rq;
+	struct i915_vma *vma;
+	int err;
+
+	vma = i915_vma_instance(obj, &i915->ggtt.base, NULL);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		return err;
+
+	rq = i915_gem_request_alloc(i915->engine[RCS], i915->kernel_context);
+	if (IS_ERR(rq)) {
+		i915_vma_unpin(vma);
+		return PTR_ERR(rq);
+	}
+
+	i915_vma_move_to_active(vma, rq, 0);
+	i915_add_request(rq);
+
+	i915_gem_object_set_active_reference(obj);
+	i915_vma_unpin(vma);
+	return 0;
+}
+
+static bool assert_mmap_offset(struct drm_i915_private *i915,
+			       unsigned long size,
+			       int expected)
+{
+	struct drm_i915_gem_object *obj;
+	int err;
+
+	obj = i915_gem_object_create_internal(i915, size);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	err = i915_gem_object_create_mmap_offset(obj);
+	i915_gem_object_put(obj);
+
+	return err == expected;
+}
+
+static int igt_mmap_offset_exhaustion(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_mm *mm = &i915->drm.vma_offset_manager->vm_addr_space_mm;
+	struct drm_i915_gem_object *obj;
+	struct drm_mm_node resv, *hole;
+	u64 hole_start, hole_end;
+	int loop, err;
+
+	/* Trim the device mmap space to only a page */
+	memset(&resv, 0, sizeof(resv));
+	drm_mm_for_each_hole(hole, mm, hole_start, hole_end) {
+		resv.start = hole_start;
+		resv.size = hole_end - hole_start - 1; /* PAGE_SIZE units */
+		err = drm_mm_reserve_node(mm, &resv);
+		if (err) {
+			pr_err("Failed to trim VMA manager, err=%d\n", err);
+			return err;
+		}
+		break;
+	}
+
+	/* Just fits! */
+	if (!assert_mmap_offset(i915, PAGE_SIZE, 0)) {
+		pr_err("Unable to insert object into single page hole\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	/* Too large */
+	if (!assert_mmap_offset(i915, 2*PAGE_SIZE, -ENOSPC)) {
+		pr_err("Unexpectedly succeeded in inserting too large object into single page hole\n");
+		err = -EINVAL;
+		goto out;
+	}
+
+	/* Fill the hole, further allocation attempts should then fail */
+	obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj)) {
+		err = PTR_ERR(obj);
+		goto out;
+	}
+
+	err = i915_gem_object_create_mmap_offset(obj);
+	if (err) {
+		pr_err("Unable to insert object into reclaimed hole\n");
+		goto err_obj;
+	}
+
+	if (!assert_mmap_offset(i915, PAGE_SIZE, -ENOSPC)) {
+		pr_err("Unexpectedly succeeded in inserting object into no holes!\n");
+		err = -EINVAL;
+		goto err_obj;
+	}
+
+	i915_gem_object_put(obj);
+
+	/* Now fill with busy dead objects that we expect to reap */
+	for (loop = 0; loop < 3; loop++) {
+		obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+		if (IS_ERR(obj)) {
+			err = PTR_ERR(obj);
+			goto out;
+		}
+
+		mutex_lock(&i915->drm.struct_mutex);
+		err = make_obj_busy(obj);
+		mutex_unlock(&i915->drm.struct_mutex);
+		if (err) {
+			pr_err("[loop %d] Failed to busy the object\n", loop);
+			goto err_obj;
+		}
+
+		GEM_BUG_ON(!i915_gem_object_is_active(obj));
+		err = i915_gem_object_create_mmap_offset(obj);
+		if (err) {
+			pr_err("[loop %d] i915_gem_object_create_mmap_offset failed with err=%d\n",
+			       loop, err);
+			goto out;
+		}
+	}
+
+out:
+	drm_mm_remove_node(&resv);
+	return err;
+err_obj:
+	i915_gem_object_put(obj);
+	goto out;
+}
+
+int i915_gem_object_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_gem_object),
+		SUBTEST(igt_phys_object),
+	};
+	struct drm_i915_private *i915;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	err = i915_subtests(tests, i915);
+
+	drm_dev_unref(&i915->drm);
+	return err;
+}
+
+int i915_gem_object_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_gem_huge),
+		SUBTEST(igt_partial_tiling),
+		SUBTEST(igt_mmap_offset_exhaustion),
+	};
+
+	return i915_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_gem_request.c b/drivers/gpu/drm/i915/selftests/i915_gem_request.c
new file mode 100644
index 000000000000..926b24c117d6
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_gem_request.c
@@ -0,0 +1,882 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/prime_numbers.h>
+
+#include "../i915_selftest.h"
+
+#include "mock_context.h"
+#include "mock_gem_device.h"
+
+static int igt_add_request(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_request *request;
+	int err = -ENOMEM;
+
+	/* Basic preliminary test to create a request and let it loose! */
+
+	mutex_lock(&i915->drm.struct_mutex);
+	request = mock_request(i915->engine[RCS],
+			       i915->kernel_context,
+			       HZ / 10);
+	if (!request)
+		goto out_unlock;
+
+	i915_add_request(request);
+
+	err = 0;
+out_unlock:
+	mutex_unlock(&i915->drm.struct_mutex);
+	return err;
+}
+
+static int igt_wait_request(void *arg)
+{
+	const long T = HZ / 4;
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_request *request;
+	int err = -EINVAL;
+
+	/* Submit a request, then wait upon it */
+
+	mutex_lock(&i915->drm.struct_mutex);
+	request = mock_request(i915->engine[RCS], i915->kernel_context, T);
+	if (!request) {
+		err = -ENOMEM;
+		goto out_unlock;
+	}
+
+	if (i915_wait_request(request, I915_WAIT_LOCKED, 0) != -ETIME) {
+		pr_err("request wait (busy query) succeeded (expected timeout before submit!)\n");
+		goto out_unlock;
+	}
+
+	if (i915_wait_request(request, I915_WAIT_LOCKED, T) != -ETIME) {
+		pr_err("request wait succeeded (expected timeout before submit!)\n");
+		goto out_unlock;
+	}
+
+	if (i915_gem_request_completed(request)) {
+		pr_err("request completed before submit!!\n");
+		goto out_unlock;
+	}
+
+	i915_add_request(request);
+
+	if (i915_wait_request(request, I915_WAIT_LOCKED, 0) != -ETIME) {
+		pr_err("request wait (busy query) succeeded (expected timeout after submit!)\n");
+		goto out_unlock;
+	}
+
+	if (i915_gem_request_completed(request)) {
+		pr_err("request completed immediately!\n");
+		goto out_unlock;
+	}
+
+	if (i915_wait_request(request, I915_WAIT_LOCKED, T / 2) != -ETIME) {
+		pr_err("request wait succeeded (expected timeout!)\n");
+		goto out_unlock;
+	}
+
+	if (i915_wait_request(request, I915_WAIT_LOCKED, T) == -ETIME) {
+		pr_err("request wait timed out!\n");
+		goto out_unlock;
+	}
+
+	if (!i915_gem_request_completed(request)) {
+		pr_err("request not complete after waiting!\n");
+		goto out_unlock;
+	}
+
+	if (i915_wait_request(request, I915_WAIT_LOCKED, T) == -ETIME) {
+		pr_err("request wait timed out when already complete!\n");
+		goto out_unlock;
+	}
+
+	err = 0;
+out_unlock:
+	mock_device_flush(i915);
+	mutex_unlock(&i915->drm.struct_mutex);
+	return err;
+}
+
+static int igt_fence_wait(void *arg)
+{
+	const long T = HZ / 4;
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_request *request;
+	int err = -EINVAL;
+
+	/* Submit a request, treat it as a fence and wait upon it */
+
+	mutex_lock(&i915->drm.struct_mutex);
+	request = mock_request(i915->engine[RCS], i915->kernel_context, T);
+	if (!request) {
+		err = -ENOMEM;
+		goto out_locked;
+	}
+	mutex_unlock(&i915->drm.struct_mutex); /* safe as we are single user */
+
+	if (dma_fence_wait_timeout(&request->fence, false, T) != -ETIME) {
+		pr_err("fence wait success before submit (expected timeout)!\n");
+		goto out_device;
+	}
+
+	mutex_lock(&i915->drm.struct_mutex);
+	i915_add_request(request);
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	if (dma_fence_is_signaled(&request->fence)) {
+		pr_err("fence signaled immediately!\n");
+		goto out_device;
+	}
+
+	if (dma_fence_wait_timeout(&request->fence, false, T / 2) != -ETIME) {
+		pr_err("fence wait success after submit (expected timeout)!\n");
+		goto out_device;
+	}
+
+	if (dma_fence_wait_timeout(&request->fence, false, T) <= 0) {
+		pr_err("fence wait timed out (expected success)!\n");
+		goto out_device;
+	}
+
+	if (!dma_fence_is_signaled(&request->fence)) {
+		pr_err("fence unsignaled after waiting!\n");
+		goto out_device;
+	}
+
+	if (dma_fence_wait_timeout(&request->fence, false, T) <= 0) {
+		pr_err("fence wait timed out when complete (expected success)!\n");
+		goto out_device;
+	}
+
+	err = 0;
+out_device:
+	mutex_lock(&i915->drm.struct_mutex);
+out_locked:
+	mock_device_flush(i915);
+	mutex_unlock(&i915->drm.struct_mutex);
+	return err;
+}
+
+static int igt_request_rewind(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_request *request, *vip;
+	struct i915_gem_context *ctx[2];
+	int err = -EINVAL;
+
+	mutex_lock(&i915->drm.struct_mutex);
+	ctx[0] = mock_context(i915, "A");
+	request = mock_request(i915->engine[RCS], ctx[0], 2 * HZ);
+	if (!request) {
+		err = -ENOMEM;
+		goto err_context_0;
+	}
+
+	i915_gem_request_get(request);
+	i915_add_request(request);
+
+	ctx[1] = mock_context(i915, "B");
+	vip = mock_request(i915->engine[RCS], ctx[1], 0);
+	if (!vip) {
+		err = -ENOMEM;
+		goto err_context_1;
+	}
+
+	/* Simulate preemption by manual reordering */
+	if (!mock_cancel_request(request)) {
+		pr_err("failed to cancel request (already executed)!\n");
+		i915_add_request(vip);
+		goto err_context_1;
+	}
+	i915_gem_request_get(vip);
+	i915_add_request(vip);
+	request->engine->submit_request(request);
+
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	if (i915_wait_request(vip, 0, HZ) == -ETIME) {
+		pr_err("timed out waiting for high priority request, vip.seqno=%d, current seqno=%d\n",
+		       vip->global_seqno, intel_engine_get_seqno(i915->engine[RCS]));
+		goto err;
+	}
+
+	if (i915_gem_request_completed(request)) {
+		pr_err("low priority request already completed\n");
+		goto err;
+	}
+
+	err = 0;
+err:
+	i915_gem_request_put(vip);
+	mutex_lock(&i915->drm.struct_mutex);
+err_context_1:
+	mock_context_close(ctx[1]);
+	i915_gem_request_put(request);
+err_context_0:
+	mock_context_close(ctx[0]);
+	mock_device_flush(i915);
+	mutex_unlock(&i915->drm.struct_mutex);
+	return err;
+}
+
+int i915_gem_request_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_add_request),
+		SUBTEST(igt_wait_request),
+		SUBTEST(igt_fence_wait),
+		SUBTEST(igt_request_rewind),
+	};
+	struct drm_i915_private *i915;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	err = i915_subtests(tests, i915);
+	drm_dev_unref(&i915->drm);
+
+	return err;
+}
+
+struct live_test {
+	struct drm_i915_private *i915;
+	const char *func;
+	const char *name;
+
+	unsigned int reset_count;
+};
+
+static int begin_live_test(struct live_test *t,
+			   struct drm_i915_private *i915,
+			   const char *func,
+			   const char *name)
+{
+	int err;
+
+	t->i915 = i915;
+	t->func = func;
+	t->name = name;
+
+	err = i915_gem_wait_for_idle(i915, I915_WAIT_LOCKED);
+	if (err) {
+		pr_err("%s(%s): failed to idle before, with err=%d!",
+		       func, name, err);
+		return err;
+	}
+
+	i915_gem_retire_requests(i915);
+
+	i915->gpu_error.missed_irq_rings = 0;
+	t->reset_count = i915_reset_count(&i915->gpu_error);
+
+	return 0;
+}
+
+static int end_live_test(struct live_test *t)
+{
+	struct drm_i915_private *i915 = t->i915;
+
+	if (wait_for(intel_engines_are_idle(i915), 1)) {
+		pr_err("%s(%s): GPU not idle\n", t->func, t->name);
+		return -EIO;
+	}
+
+	if (t->reset_count != i915_reset_count(&i915->gpu_error)) {
+		pr_err("%s(%s): GPU was reset %d times!\n",
+		       t->func, t->name,
+		       i915_reset_count(&i915->gpu_error) - t->reset_count);
+		return -EIO;
+	}
+
+	if (i915->gpu_error.missed_irq_rings) {
+		pr_err("%s(%s): Missed interrupts on engines %lx\n",
+		       t->func, t->name, i915->gpu_error.missed_irq_rings);
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static int live_nop_request(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_engine_cs *engine;
+	struct live_test t;
+	unsigned int id;
+	int err;
+
+	/* Submit various sized batches of empty requests, to each engine
+	 * (individually), and wait for the batch to complete. We can check
+	 * the overhead of submitting requests to the hardware.
+	 */
+
+	mutex_lock(&i915->drm.struct_mutex);
+
+	for_each_engine(engine, i915, id) {
+		IGT_TIMEOUT(end_time);
+		struct drm_i915_gem_request *request;
+		unsigned long n, prime;
+		ktime_t times[2] = {};
+
+		err = begin_live_test(&t, i915, __func__, engine->name);
+		if (err)
+			goto out_unlock;
+
+		for_each_prime_number_from(prime, 1, 8192) {
+			times[1] = ktime_get_raw();
+
+			for (n = 0; n < prime; n++) {
+				request = i915_gem_request_alloc(engine,
+								 i915->kernel_context);
+				if (IS_ERR(request)) {
+					err = PTR_ERR(request);
+					goto out_unlock;
+				}
+
+				/* This space is left intentionally blank.
+				 *
+				 * We do not actually want to perform any
+				 * action with this request, we just want
+				 * to measure the latency in allocation
+				 * and submission of our breadcrumbs -
+				 * ensuring that the bare request is sufficient
+				 * for the system to work (i.e. proper HEAD
+				 * tracking of the rings, interrupt handling,
+				 * etc). It also gives us the lowest bounds
+				 * for latency.
+				 */
+
+				i915_add_request(request);
+			}
+			i915_wait_request(request,
+					  I915_WAIT_LOCKED,
+					  MAX_SCHEDULE_TIMEOUT);
+
+			times[1] = ktime_sub(ktime_get_raw(), times[1]);
+			if (prime == 1)
+				times[0] = times[1];
+
+			if (__igt_timeout(end_time, NULL))
+				break;
+		}
+
+		err = end_live_test(&t);
+		if (err)
+			goto out_unlock;
+
+		pr_info("Request latencies on %s: 1 = %lluns, %lu = %lluns\n",
+			engine->name,
+			ktime_to_ns(times[0]),
+			prime, div64_u64(ktime_to_ns(times[1]), prime));
+	}
+
+out_unlock:
+	mutex_unlock(&i915->drm.struct_mutex);
+	return err;
+}
+
+static struct i915_vma *empty_batch(struct drm_i915_private *i915)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	u32 *cmd;
+	int err;
+
+	obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	cmd = i915_gem_object_pin_map(obj, I915_MAP_WB);
+	if (IS_ERR(cmd)) {
+		err = PTR_ERR(cmd);
+		goto err;
+	}
+	*cmd = MI_BATCH_BUFFER_END;
+	i915_gem_object_unpin_map(obj);
+
+	err = i915_gem_object_set_to_gtt_domain(obj, false);
+	if (err)
+		goto err;
+
+	vma = i915_vma_instance(obj, &i915->ggtt.base, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER | PIN_GLOBAL);
+	if (err)
+		goto err;
+
+	return vma;
+
+err:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+static struct drm_i915_gem_request *
+empty_request(struct intel_engine_cs *engine,
+	      struct i915_vma *batch)
+{
+	struct drm_i915_gem_request *request;
+	int err;
+
+	request = i915_gem_request_alloc(engine,
+					 engine->i915->kernel_context);
+	if (IS_ERR(request))
+		return request;
+
+	err = engine->emit_flush(request, EMIT_INVALIDATE);
+	if (err)
+		goto out_request;
+
+	err = i915_switch_context(request);
+	if (err)
+		goto out_request;
+
+	err = engine->emit_bb_start(request,
+				    batch->node.start,
+				    batch->node.size,
+				    I915_DISPATCH_SECURE);
+	if (err)
+		goto out_request;
+
+out_request:
+	__i915_add_request(request, err == 0);
+	return err ? ERR_PTR(err) : request;
+}
+
+static int live_empty_request(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_engine_cs *engine;
+	struct live_test t;
+	struct i915_vma *batch;
+	unsigned int id;
+	int err = 0;
+
+	/* Submit various sized batches of empty requests, to each engine
+	 * (individually), and wait for the batch to complete. We can check
+	 * the overhead of submitting requests to the hardware.
+	 */
+
+	mutex_lock(&i915->drm.struct_mutex);
+
+	batch = empty_batch(i915);
+	if (IS_ERR(batch)) {
+		err = PTR_ERR(batch);
+		goto out_unlock;
+	}
+
+	for_each_engine(engine, i915, id) {
+		IGT_TIMEOUT(end_time);
+		struct drm_i915_gem_request *request;
+		unsigned long n, prime;
+		ktime_t times[2] = {};
+
+		err = begin_live_test(&t, i915, __func__, engine->name);
+		if (err)
+			goto out_batch;
+
+		/* Warmup / preload */
+		request = empty_request(engine, batch);
+		if (IS_ERR(request)) {
+			err = PTR_ERR(request);
+			goto out_batch;
+		}
+		i915_wait_request(request,
+				  I915_WAIT_LOCKED,
+				  MAX_SCHEDULE_TIMEOUT);
+
+		for_each_prime_number_from(prime, 1, 8192) {
+			times[1] = ktime_get_raw();
+
+			for (n = 0; n < prime; n++) {
+				request = empty_request(engine, batch);
+				if (IS_ERR(request)) {
+					err = PTR_ERR(request);
+					goto out_batch;
+				}
+			}
+			i915_wait_request(request,
+					  I915_WAIT_LOCKED,
+					  MAX_SCHEDULE_TIMEOUT);
+
+			times[1] = ktime_sub(ktime_get_raw(), times[1]);
+			if (prime == 1)
+				times[0] = times[1];
+
+			if (__igt_timeout(end_time, NULL))
+				break;
+		}
+
+		err = end_live_test(&t);
+		if (err)
+			goto out_batch;
+
+		pr_info("Batch latencies on %s: 1 = %lluns, %lu = %lluns\n",
+			engine->name,
+			ktime_to_ns(times[0]),
+			prime, div64_u64(ktime_to_ns(times[1]), prime));
+	}
+
+out_batch:
+	i915_vma_unpin(batch);
+	i915_vma_put(batch);
+out_unlock:
+	mutex_unlock(&i915->drm.struct_mutex);
+	return err;
+}
+
+static struct i915_vma *recursive_batch(struct drm_i915_private *i915)
+{
+	struct i915_gem_context *ctx = i915->kernel_context;
+	struct i915_address_space *vm = ctx->ppgtt ? &ctx->ppgtt->base : &i915->ggtt.base;
+	struct drm_i915_gem_object *obj;
+	const int gen = INTEL_GEN(i915);
+	struct i915_vma *vma;
+	u32 *cmd;
+	int err;
+
+	obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return ERR_CAST(obj);
+
+	vma = i915_vma_instance(obj, vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		goto err;
+
+	err = i915_gem_object_set_to_gtt_domain(obj, true);
+	if (err)
+		goto err;
+
+	cmd = i915_gem_object_pin_map(obj, I915_MAP_WC);
+	if (IS_ERR(cmd)) {
+		err = PTR_ERR(cmd);
+		goto err;
+	}
+
+	if (gen >= 8) {
+		*cmd++ = MI_BATCH_BUFFER_START | 1 << 8 | 1;
+		*cmd++ = lower_32_bits(vma->node.start);
+		*cmd++ = upper_32_bits(vma->node.start);
+	} else if (gen >= 6) {
+		*cmd++ = MI_BATCH_BUFFER_START | 1 << 8;
+		*cmd++ = lower_32_bits(vma->node.start);
+	} else if (gen >= 4) {
+		*cmd++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT;
+		*cmd++ = lower_32_bits(vma->node.start);
+	} else {
+		*cmd++ = MI_BATCH_BUFFER_START | MI_BATCH_GTT | 1;
+		*cmd++ = lower_32_bits(vma->node.start);
+	}
+	*cmd++ = MI_BATCH_BUFFER_END; /* terminate early in case of error */
+
+	wmb();
+	i915_gem_object_unpin_map(obj);
+
+	return vma;
+
+err:
+	i915_gem_object_put(obj);
+	return ERR_PTR(err);
+}
+
+static int recursive_batch_resolve(struct i915_vma *batch)
+{
+	u32 *cmd;
+
+	cmd = i915_gem_object_pin_map(batch->obj, I915_MAP_WC);
+	if (IS_ERR(cmd))
+		return PTR_ERR(cmd);
+
+	*cmd = MI_BATCH_BUFFER_END;
+	wmb();
+
+	i915_gem_object_unpin_map(batch->obj);
+
+	return 0;
+}
+
+static int live_all_engines(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_engine_cs *engine;
+	struct drm_i915_gem_request *request[I915_NUM_ENGINES];
+	struct i915_vma *batch;
+	struct live_test t;
+	unsigned int id;
+	int err;
+
+	/* Check we can submit requests to all engines simultaneously. We
+	 * send a recursive batch to each engine - checking that we don't
+	 * block doing so, and that they don't complete too soon.
+	 */
+
+	mutex_lock(&i915->drm.struct_mutex);
+
+	err = begin_live_test(&t, i915, __func__, "");
+	if (err)
+		goto out_unlock;
+
+	batch = recursive_batch(i915);
+	if (IS_ERR(batch)) {
+		err = PTR_ERR(batch);
+		pr_err("%s: Unable to create batch, err=%d\n", __func__, err);
+		goto out_unlock;
+	}
+
+	for_each_engine(engine, i915, id) {
+		request[id] = i915_gem_request_alloc(engine,
+						     i915->kernel_context);
+		if (IS_ERR(request[id])) {
+			err = PTR_ERR(request[id]);
+			pr_err("%s: Request allocation failed with err=%d\n",
+			       __func__, err);
+			goto out_request;
+		}
+
+		err = engine->emit_flush(request[id], EMIT_INVALIDATE);
+		GEM_BUG_ON(err);
+
+		err = i915_switch_context(request[id]);
+		GEM_BUG_ON(err);
+
+		err = engine->emit_bb_start(request[id],
+					    batch->node.start,
+					    batch->node.size,
+					    0);
+		GEM_BUG_ON(err);
+		request[id]->batch = batch;
+
+		if (!i915_gem_object_has_active_reference(batch->obj)) {
+			i915_gem_object_get(batch->obj);
+			i915_gem_object_set_active_reference(batch->obj);
+		}
+
+		i915_vma_move_to_active(batch, request[id], 0);
+		i915_gem_request_get(request[id]);
+		i915_add_request(request[id]);
+	}
+
+	for_each_engine(engine, i915, id) {
+		if (i915_gem_request_completed(request[id])) {
+			pr_err("%s(%s): request completed too early!\n",
+			       __func__, engine->name);
+			err = -EINVAL;
+			goto out_request;
+		}
+	}
+
+	err = recursive_batch_resolve(batch);
+	if (err) {
+		pr_err("%s: failed to resolve batch, err=%d\n", __func__, err);
+		goto out_request;
+	}
+
+	for_each_engine(engine, i915, id) {
+		long timeout;
+
+		timeout = i915_wait_request(request[id],
+					    I915_WAIT_LOCKED,
+					    MAX_SCHEDULE_TIMEOUT);
+		if (timeout < 0) {
+			err = timeout;
+			pr_err("%s: error waiting for request on %s, err=%d\n",
+			       __func__, engine->name, err);
+			goto out_request;
+		}
+
+		GEM_BUG_ON(!i915_gem_request_completed(request[id]));
+		i915_gem_request_put(request[id]);
+		request[id] = NULL;
+	}
+
+	err = end_live_test(&t);
+
+out_request:
+	for_each_engine(engine, i915, id)
+		if (request[id])
+			i915_gem_request_put(request[id]);
+	i915_vma_unpin(batch);
+	i915_vma_put(batch);
+out_unlock:
+	mutex_unlock(&i915->drm.struct_mutex);
+	return err;
+}
+
+static int live_sequential_engines(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_request *request[I915_NUM_ENGINES] = {};
+	struct drm_i915_gem_request *prev = NULL;
+	struct intel_engine_cs *engine;
+	struct live_test t;
+	unsigned int id;
+	int err;
+
+	/* Check we can submit requests to all engines sequentially, such
+	 * that each successive request waits for the earlier ones. This
+	 * tests that we don't execute requests out of order, even though
+	 * they are running on independent engines.
+	 */
+
+	mutex_lock(&i915->drm.struct_mutex);
+
+	err = begin_live_test(&t, i915, __func__, "");
+	if (err)
+		goto out_unlock;
+
+	for_each_engine(engine, i915, id) {
+		struct i915_vma *batch;
+
+		batch = recursive_batch(i915);
+		if (IS_ERR(batch)) {
+			err = PTR_ERR(batch);
+			pr_err("%s: Unable to create batch for %s, err=%d\n",
+			       __func__, engine->name, err);
+			goto out_unlock;
+		}
+
+		request[id] = i915_gem_request_alloc(engine,
+						     i915->kernel_context);
+		if (IS_ERR(request[id])) {
+			err = PTR_ERR(request[id]);
+			pr_err("%s: Request allocation failed for %s with err=%d\n",
+			       __func__, engine->name, err);
+			goto out_request;
+		}
+
+		if (prev) {
+			err = i915_gem_request_await_dma_fence(request[id],
+							       &prev->fence);
+			if (err) {
+				i915_add_request(request[id]);
+				pr_err("%s: Request await failed for %s with err=%d\n",
+				       __func__, engine->name, err);
+				goto out_request;
+			}
+		}
+
+		err = engine->emit_flush(request[id], EMIT_INVALIDATE);
+		GEM_BUG_ON(err);
+
+		err = i915_switch_context(request[id]);
+		GEM_BUG_ON(err);
+
+		err = engine->emit_bb_start(request[id],
+					    batch->node.start,
+					    batch->node.size,
+					    0);
+		GEM_BUG_ON(err);
+		request[id]->batch = batch;
+
+		i915_vma_move_to_active(batch, request[id], 0);
+		i915_gem_object_set_active_reference(batch->obj);
+		i915_vma_get(batch);
+
+		i915_gem_request_get(request[id]);
+		i915_add_request(request[id]);
+
+		prev = request[id];
+	}
+
+	for_each_engine(engine, i915, id) {
+		long timeout;
+
+		if (i915_gem_request_completed(request[id])) {
+			pr_err("%s(%s): request completed too early!\n",
+			       __func__, engine->name);
+			err = -EINVAL;
+			goto out_request;
+		}
+
+		err = recursive_batch_resolve(request[id]->batch);
+		if (err) {
+			pr_err("%s: failed to resolve batch, err=%d\n",
+			       __func__, err);
+			goto out_request;
+		}
+
+		timeout = i915_wait_request(request[id],
+					    I915_WAIT_LOCKED,
+					    MAX_SCHEDULE_TIMEOUT);
+		if (timeout < 0) {
+			err = timeout;
+			pr_err("%s: error waiting for request on %s, err=%d\n",
+			       __func__, engine->name, err);
+			goto out_request;
+		}
+
+		GEM_BUG_ON(!i915_gem_request_completed(request[id]));
+	}
+
+	err = end_live_test(&t);
+
+out_request:
+	for_each_engine(engine, i915, id) {
+		u32 *cmd;
+
+		if (!request[id])
+			break;
+
+		cmd = i915_gem_object_pin_map(request[id]->batch->obj,
+					      I915_MAP_WC);
+		if (!IS_ERR(cmd)) {
+			*cmd = MI_BATCH_BUFFER_END;
+			wmb();
+			i915_gem_object_unpin_map(request[id]->batch->obj);
+		}
+
+		i915_vma_put(request[id]->batch);
+		i915_gem_request_put(request[id]);
+	}
+out_unlock:
+	mutex_unlock(&i915->drm.struct_mutex);
+	return err;
+}
+
+int i915_gem_request_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_nop_request),
+		SUBTEST(live_all_engines),
+		SUBTEST(live_sequential_engines),
+		SUBTEST(live_empty_request),
+	};
+	return i915_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_live_selftests.h b/drivers/gpu/drm/i915/selftests/i915_live_selftests.h
new file mode 100644
index 000000000000..18b174d855ca
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_live_selftests.h
@@ -0,0 +1,19 @@
+/* List each unit test as selftest(name, function)
+ *
+ * The name is used as both an enum and expanded as subtest__name to create
+ * a module parameter. It must be unique and legal for a C identifier.
+ *
+ * The function should be of type int function(void). It may be conditionally
+ * compiled using #if IS_ENABLED(DRM_I915_SELFTEST).
+ *
+ * Tests are executed in order by igt/drv_selftest
+ */
+selftest(sanitycheck, i915_live_sanitycheck) /* keep first (igt selfcheck) */
+selftest(uncore, intel_uncore_live_selftests)
+selftest(requests, i915_gem_request_live_selftests)
+selftest(objects, i915_gem_object_live_selftests)
+selftest(dmabuf, i915_gem_dmabuf_live_selftests)
+selftest(coherency, i915_gem_coherency_live_selftests)
+selftest(gtt, i915_gem_gtt_live_selftests)
+selftest(contexts, i915_gem_context_live_selftests)
+selftest(hangcheck, intel_hangcheck_live_selftests)
diff --git a/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h b/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h
new file mode 100644
index 000000000000..be9a9ebf5692
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_mock_selftests.h
@@ -0,0 +1,20 @@
+/* List each unit test as selftest(name, function)
+ *
+ * The name is used as both an enum and expanded as subtest__name to create
+ * a module parameter. It must be unique and legal for a C identifier.
+ *
+ * The function should be of type int function(void). It may be conditionally
+ * compiled using #if IS_ENABLED(DRM_I915_SELFTEST).
+ *
+ * Tests are executed in order by igt/drv_selftest
+ */
+selftest(sanitycheck, i915_mock_sanitycheck) /* keep first (igt selfcheck) */
+selftest(scatterlist, scatterlist_mock_selftests)
+selftest(uncore, intel_uncore_mock_selftests)
+selftest(breadcrumbs, intel_breadcrumbs_mock_selftests)
+selftest(requests, i915_gem_request_mock_selftests)
+selftest(objects, i915_gem_object_mock_selftests)
+selftest(dmabuf, i915_gem_dmabuf_mock_selftests)
+selftest(vma, i915_vma_mock_selftests)
+selftest(evict, i915_gem_evict_mock_selftests)
+selftest(gtt, i915_gem_gtt_mock_selftests)
diff --git a/drivers/gpu/drm/i915/selftests/i915_random.c b/drivers/gpu/drm/i915/selftests/i915_random.c
new file mode 100644
index 000000000000..c17c83c30637
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_random.c
@@ -0,0 +1,63 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/bitops.h>
+#include <linux/kernel.h>
+#include <linux/random.h>
+#include <linux/slab.h>
+#include <linux/types.h>
+
+#include "i915_random.h"
+
+static inline u32 i915_prandom_u32_max_state(u32 ep_ro, struct rnd_state *state)
+{
+	return upper_32_bits((u64)prandom_u32_state(state) * ep_ro);
+}
+
+void i915_random_reorder(unsigned int *order, unsigned int count,
+			 struct rnd_state *state)
+{
+	unsigned int i, j;
+
+	for (i = 0; i < count; i++) {
+		BUILD_BUG_ON(sizeof(unsigned int) > sizeof(u32));
+		j = i915_prandom_u32_max_state(count, state);
+		swap(order[i], order[j]);
+	}
+}
+
+unsigned int *i915_random_order(unsigned int count, struct rnd_state *state)
+{
+	unsigned int *order, i;
+
+	order = kmalloc_array(count, sizeof(*order), GFP_TEMPORARY);
+	if (!order)
+		return order;
+
+	for (i = 0; i < count; i++)
+		order[i] = i;
+
+	i915_random_reorder(order, count, state);
+	return order;
+}
diff --git a/drivers/gpu/drm/i915/selftests/i915_random.h b/drivers/gpu/drm/i915/selftests/i915_random.h
new file mode 100644
index 000000000000..b9c334ce6cd9
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_random.h
@@ -0,0 +1,50 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __I915_SELFTESTS_RANDOM_H__
+#define __I915_SELFTESTS_RANDOM_H__
+
+#include <linux/random.h>
+
+#include "../i915_selftest.h"
+
+#define I915_RND_STATE_INITIALIZER(x) ({				\
+	struct rnd_state state__;					\
+	prandom_seed_state(&state__, (x));				\
+	state__;							\
+})
+
+#define I915_RND_STATE(name__) \
+	struct rnd_state name__ = I915_RND_STATE_INITIALIZER(i915_selftest.random_seed)
+
+#define I915_RND_SUBSTATE(name__, parent__) \
+	struct rnd_state name__ = I915_RND_STATE_INITIALIZER(prandom_u32_state(&(parent__)))
+
+unsigned int *i915_random_order(unsigned int count,
+				struct rnd_state *state);
+void i915_random_reorder(unsigned int *order,
+			 unsigned int count,
+			 struct rnd_state *state);
+
+#endif /* !__I915_SELFTESTS_RANDOM_H__ */
diff --git a/drivers/gpu/drm/i915/selftests/i915_selftest.c b/drivers/gpu/drm/i915/selftests/i915_selftest.c
new file mode 100644
index 000000000000..6ba3abb10c6f
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_selftest.c
@@ -0,0 +1,250 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/random.h>
+
+#include "../i915_drv.h"
+#include "../i915_selftest.h"
+
+struct i915_selftest i915_selftest __read_mostly = {
+	.timeout_ms = 1000,
+};
+
+int i915_mock_sanitycheck(void)
+{
+	pr_info(DRIVER_NAME ": %s() - ok!\n", __func__);
+	return 0;
+}
+
+int i915_live_sanitycheck(struct drm_i915_private *i915)
+{
+	pr_info("%s: %s() - ok!\n", i915->drm.driver->name, __func__);
+	return 0;
+}
+
+enum {
+#define selftest(name, func) mock_##name,
+#include "i915_mock_selftests.h"
+#undef selftest
+};
+
+enum {
+#define selftest(name, func) live_##name,
+#include "i915_live_selftests.h"
+#undef selftest
+};
+
+struct selftest {
+	bool enabled;
+	const char *name;
+	union {
+		int (*mock)(void);
+		int (*live)(struct drm_i915_private *);
+	};
+};
+
+#define selftest(n, f) [mock_##n] = { .name = #n, .mock = f },
+static struct selftest mock_selftests[] = {
+#include "i915_mock_selftests.h"
+};
+#undef selftest
+
+#define selftest(n, f) [live_##n] = { .name = #n, .live = f },
+static struct selftest live_selftests[] = {
+#include "i915_live_selftests.h"
+};
+#undef selftest
+
+/* Embed the line number into the parameter name so that we can order tests */
+#define selftest(n, func) selftest_0(n, func, param(n))
+#define param(n) __PASTE(igt__, __PASTE(__LINE__, __mock_##n))
+#define selftest_0(n, func, id) \
+module_param_named(id, mock_selftests[mock_##n].enabled, bool, 0400);
+#include "i915_mock_selftests.h"
+#undef selftest_0
+#undef param
+
+#define param(n) __PASTE(igt__, __PASTE(__LINE__, __live_##n))
+#define selftest_0(n, func, id) \
+module_param_named(id, live_selftests[live_##n].enabled, bool, 0400);
+#include "i915_live_selftests.h"
+#undef selftest_0
+#undef param
+#undef selftest
+
+static void set_default_test_all(struct selftest *st, unsigned int count)
+{
+	unsigned int i;
+
+	for (i = 0; i < count; i++)
+		if (st[i].enabled)
+			return;
+
+	for (i = 0; i < count; i++)
+		st[i].enabled = true;
+}
+
+static int __run_selftests(const char *name,
+			   struct selftest *st,
+			   unsigned int count,
+			   void *data)
+{
+	int err = 0;
+
+	while (!i915_selftest.random_seed)
+		i915_selftest.random_seed = get_random_int();
+
+	i915_selftest.timeout_jiffies =
+		i915_selftest.timeout_ms ?
+		msecs_to_jiffies_timeout(i915_selftest.timeout_ms) :
+		MAX_SCHEDULE_TIMEOUT;
+
+	set_default_test_all(st, count);
+
+	pr_info(DRIVER_NAME ": Performing %s selftests with st_random_seed=0x%x st_timeout=%u\n",
+		name, i915_selftest.random_seed, i915_selftest.timeout_ms);
+
+	/* Tests are listed in order in i915_*_selftests.h */
+	for (; count--; st++) {
+		if (!st->enabled)
+			continue;
+
+		cond_resched();
+		if (signal_pending(current))
+			return -EINTR;
+
+		pr_debug(DRIVER_NAME ": Running %s\n", st->name);
+		if (data)
+			err = st->live(data);
+		else
+			err = st->mock();
+		if (err == -EINTR && !signal_pending(current))
+			err = 0;
+		if (err)
+			break;
+	}
+
+	if (WARN(err > 0 || err == -ENOTTY,
+		 "%s returned %d, conflicting with selftest's magic values!\n",
+		 st->name, err))
+		err = -1;
+
+	return err;
+}
+
+#define run_selftests(x, data) \
+	__run_selftests(#x, x##_selftests, ARRAY_SIZE(x##_selftests), data)
+
+int i915_mock_selftests(void)
+{
+	int err;
+
+	if (!i915_selftest.mock)
+		return 0;
+
+	err = run_selftests(mock, NULL);
+	if (err) {
+		i915_selftest.mock = err;
+		return err;
+	}
+
+	if (i915_selftest.mock < 0) {
+		i915_selftest.mock = -ENOTTY;
+		return 1;
+	}
+
+	return 0;
+}
+
+int i915_live_selftests(struct pci_dev *pdev)
+{
+	int err;
+
+	if (!i915_selftest.live)
+		return 0;
+
+	err = run_selftests(live, to_i915(pci_get_drvdata(pdev)));
+	if (err) {
+		i915_selftest.live = err;
+		return err;
+	}
+
+	if (i915_selftest.live < 0) {
+		i915_selftest.live = -ENOTTY;
+		return 1;
+	}
+
+	return 0;
+}
+
+int __i915_subtests(const char *caller,
+		    const struct i915_subtest *st,
+		    unsigned int count,
+		    void *data)
+{
+	int err;
+
+	for (; count--; st++) {
+		cond_resched();
+		if (signal_pending(current))
+			return -EINTR;
+
+		pr_debug(DRIVER_NAME ": Running %s/%s\n", caller, st->name);
+		err = st->func(data);
+		if (err && err != -EINTR) {
+			pr_err(DRIVER_NAME "/%s: %s failed with error %d\n",
+			       caller, st->name, err);
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+bool __igt_timeout(unsigned long timeout, const char *fmt, ...)
+{
+	va_list va;
+
+	if (!signal_pending(current)) {
+		cond_resched();
+		if (time_before(jiffies, timeout))
+			return false;
+	}
+
+	if (fmt) {
+		va_start(va, fmt);
+		vprintk(fmt, va);
+		va_end(va);
+	}
+
+	return true;
+}
+
+module_param_named(st_random_seed, i915_selftest.random_seed, uint, 0400);
+module_param_named(st_timeout, i915_selftest.timeout_ms, uint, 0400);
+
+module_param_named_unsafe(mock_selftests, i915_selftest.mock, int, 0400);
+MODULE_PARM_DESC(mock_selftests, "Run selftests before loading, using mock hardware (0:disabled [default], 1:run tests then load driver, -1:run tests then exit module)");
+
+module_param_named_unsafe(live_selftests, i915_selftest.live, int, 0400);
+MODULE_PARM_DESC(live_selftests, "Run selftests after driver initialisation on the live system (0:disabled [default], 1:run tests then continue, -1:run tests then exit module)");
diff --git a/drivers/gpu/drm/i915/selftests/i915_vma.c b/drivers/gpu/drm/i915/selftests/i915_vma.c
new file mode 100644
index 000000000000..ad56566e24db
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/i915_vma.c
@@ -0,0 +1,746 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/prime_numbers.h>
+
+#include "../i915_selftest.h"
+
+#include "mock_gem_device.h"
+#include "mock_context.h"
+
+static bool assert_vma(struct i915_vma *vma,
+		       struct drm_i915_gem_object *obj,
+		       struct i915_gem_context *ctx)
+{
+	bool ok = true;
+
+	if (vma->vm != &ctx->ppgtt->base) {
+		pr_err("VMA created with wrong VM\n");
+		ok = false;
+	}
+
+	if (vma->size != obj->base.size) {
+		pr_err("VMA created with wrong size, found %llu, expected %zu\n",
+		       vma->size, obj->base.size);
+		ok = false;
+	}
+
+	if (vma->ggtt_view.type != I915_GGTT_VIEW_NORMAL) {
+		pr_err("VMA created with wrong type [%d]\n",
+		       vma->ggtt_view.type);
+		ok = false;
+	}
+
+	return ok;
+}
+
+static struct i915_vma *
+checked_vma_instance(struct drm_i915_gem_object *obj,
+		     struct i915_address_space *vm,
+		     struct i915_ggtt_view *view)
+{
+	struct i915_vma *vma;
+	bool ok = true;
+
+	vma = i915_vma_instance(obj, vm, view);
+	if (IS_ERR(vma))
+		return vma;
+
+	/* Manual checks, will be reinforced by i915_vma_compare! */
+	if (vma->vm != vm) {
+		pr_err("VMA's vm [%p] does not match request [%p]\n",
+		       vma->vm, vm);
+		ok = false;
+	}
+
+	if (i915_is_ggtt(vm) != i915_vma_is_ggtt(vma)) {
+		pr_err("VMA ggtt status [%d] does not match parent [%d]\n",
+		       i915_vma_is_ggtt(vma), i915_is_ggtt(vm));
+		ok = false;
+	}
+
+	if (i915_vma_compare(vma, vm, view)) {
+		pr_err("i915_vma_compare failed with create parmaters!\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	if (i915_vma_compare(vma, vma->vm,
+			     i915_vma_is_ggtt(vma) ? &vma->ggtt_view : NULL)) {
+		pr_err("i915_vma_compare failed with itself\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	if (!ok) {
+		pr_err("i915_vma_compare failed to detect the difference!\n");
+		return ERR_PTR(-EINVAL);
+	}
+
+	return vma;
+}
+
+static int create_vmas(struct drm_i915_private *i915,
+		       struct list_head *objects,
+		       struct list_head *contexts)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_gem_context *ctx;
+	int pinned;
+
+	list_for_each_entry(obj, objects, st_link) {
+		for (pinned = 0; pinned <= 1; pinned++) {
+			list_for_each_entry(ctx, contexts, link) {
+				struct i915_address_space *vm =
+					&ctx->ppgtt->base;
+				struct i915_vma *vma;
+				int err;
+
+				vma = checked_vma_instance(obj, vm, NULL);
+				if (IS_ERR(vma))
+					return PTR_ERR(vma);
+
+				if (!assert_vma(vma, obj, ctx)) {
+					pr_err("VMA lookup/create failed\n");
+					return -EINVAL;
+				}
+
+				if (!pinned) {
+					err = i915_vma_pin(vma, 0, 0, PIN_USER);
+					if (err) {
+						pr_err("Failed to pin VMA\n");
+						return err;
+					}
+				} else {
+					i915_vma_unpin(vma);
+				}
+			}
+		}
+	}
+
+	return 0;
+}
+
+static int igt_vma_create(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_object *obj, *on;
+	struct i915_gem_context *ctx, *cn;
+	unsigned long num_obj, num_ctx;
+	unsigned long no, nc;
+	IGT_TIMEOUT(end_time);
+	LIST_HEAD(contexts);
+	LIST_HEAD(objects);
+	int err;
+
+	/* Exercise creating many vma amonst many objections, checking the
+	 * vma creation and lookup routines.
+	 */
+
+	no = 0;
+	for_each_prime_number(num_obj, ULONG_MAX - 1) {
+		for (; no < num_obj; no++) {
+			obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+			if (IS_ERR(obj))
+				goto out;
+
+			list_add(&obj->st_link, &objects);
+		}
+
+		nc = 0;
+		for_each_prime_number(num_ctx, MAX_CONTEXT_HW_ID) {
+			for (; nc < num_ctx; nc++) {
+				ctx = mock_context(i915, "mock");
+				if (!ctx)
+					goto out;
+
+				list_move(&ctx->link, &contexts);
+			}
+
+			err = create_vmas(i915, &objects, &contexts);
+			if (err)
+				goto out;
+
+			if (igt_timeout(end_time,
+					"%s timed out: after %lu objects in %lu contexts\n",
+					__func__, no, nc))
+				goto end;
+		}
+
+		list_for_each_entry_safe(ctx, cn, &contexts, link)
+			mock_context_close(ctx);
+	}
+
+end:
+	/* Final pass to lookup all created contexts */
+	err = create_vmas(i915, &objects, &contexts);
+out:
+	list_for_each_entry_safe(ctx, cn, &contexts, link)
+		mock_context_close(ctx);
+
+	list_for_each_entry_safe(obj, on, &objects, st_link)
+		i915_gem_object_put(obj);
+	return err;
+}
+
+struct pin_mode {
+	u64 size;
+	u64 flags;
+	bool (*assert)(const struct i915_vma *,
+		       const struct pin_mode *mode,
+		       int result);
+	const char *string;
+};
+
+static bool assert_pin_valid(const struct i915_vma *vma,
+			     const struct pin_mode *mode,
+			     int result)
+{
+	if (result)
+		return false;
+
+	if (i915_vma_misplaced(vma, mode->size, 0, mode->flags))
+		return false;
+
+	return true;
+}
+
+__maybe_unused
+static bool assert_pin_e2big(const struct i915_vma *vma,
+			     const struct pin_mode *mode,
+			     int result)
+{
+	return result == -E2BIG;
+}
+
+__maybe_unused
+static bool assert_pin_enospc(const struct i915_vma *vma,
+			      const struct pin_mode *mode,
+			      int result)
+{
+	return result == -ENOSPC;
+}
+
+__maybe_unused
+static bool assert_pin_einval(const struct i915_vma *vma,
+			      const struct pin_mode *mode,
+			      int result)
+{
+	return result == -EINVAL;
+}
+
+static int igt_vma_pin1(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	const struct pin_mode modes[] = {
+#define VALID(sz, fl) { .size = (sz), .flags = (fl), .assert = assert_pin_valid, .string = #sz ", " #fl ", (valid) " }
+#define __INVALID(sz, fl, check, eval) { .size = (sz), .flags = (fl), .assert = (check), .string = #sz ", " #fl ", (invalid " #eval ")" }
+#define INVALID(sz, fl) __INVALID(sz, fl, assert_pin_einval, EINVAL)
+#define TOOBIG(sz, fl) __INVALID(sz, fl, assert_pin_e2big, E2BIG)
+#define NOSPACE(sz, fl) __INVALID(sz, fl, assert_pin_enospc, ENOSPC)
+		VALID(0, PIN_GLOBAL),
+		VALID(0, PIN_GLOBAL | PIN_MAPPABLE),
+
+		VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | 4096),
+		VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | 8192),
+		VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | (i915->ggtt.mappable_end - 4096)),
+		VALID(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_BIAS | (i915->ggtt.mappable_end - 4096)),
+		VALID(0, PIN_GLOBAL | PIN_OFFSET_BIAS | (i915->ggtt.base.total - 4096)),
+
+		VALID(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_FIXED | (i915->ggtt.mappable_end - 4096)),
+		INVALID(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_FIXED | i915->ggtt.mappable_end),
+		VALID(0, PIN_GLOBAL | PIN_OFFSET_FIXED | (i915->ggtt.base.total - 4096)),
+		INVALID(0, PIN_GLOBAL | PIN_OFFSET_FIXED | i915->ggtt.base.total),
+		INVALID(0, PIN_GLOBAL | PIN_OFFSET_FIXED | round_down(U64_MAX, PAGE_SIZE)),
+
+		VALID(4096, PIN_GLOBAL),
+		VALID(8192, PIN_GLOBAL),
+		VALID(i915->ggtt.mappable_end - 4096, PIN_GLOBAL | PIN_MAPPABLE),
+		VALID(i915->ggtt.mappable_end, PIN_GLOBAL | PIN_MAPPABLE),
+		TOOBIG(i915->ggtt.mappable_end + 4096, PIN_GLOBAL | PIN_MAPPABLE),
+		VALID(i915->ggtt.base.total - 4096, PIN_GLOBAL),
+		VALID(i915->ggtt.base.total, PIN_GLOBAL),
+		TOOBIG(i915->ggtt.base.total + 4096, PIN_GLOBAL),
+		TOOBIG(round_down(U64_MAX, PAGE_SIZE), PIN_GLOBAL),
+		INVALID(8192, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_FIXED | (i915->ggtt.mappable_end - 4096)),
+		INVALID(8192, PIN_GLOBAL | PIN_OFFSET_FIXED | (i915->ggtt.base.total - 4096)),
+		INVALID(8192, PIN_GLOBAL | PIN_OFFSET_FIXED | (round_down(U64_MAX, PAGE_SIZE) - 4096)),
+
+		VALID(8192, PIN_GLOBAL | PIN_OFFSET_BIAS | (i915->ggtt.mappable_end - 4096)),
+
+#if !IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM)
+		/* Misusing BIAS is a programming error (it is not controllable
+		 * from userspace) so when debugging is enabled, it explodes.
+		 * However, the tests are still quite interesting for checking
+		 * variable start, end and size.
+		 */
+		NOSPACE(0, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_BIAS | i915->ggtt.mappable_end),
+		NOSPACE(0, PIN_GLOBAL | PIN_OFFSET_BIAS | i915->ggtt.base.total),
+		NOSPACE(8192, PIN_GLOBAL | PIN_MAPPABLE | PIN_OFFSET_BIAS | (i915->ggtt.mappable_end - 4096)),
+		NOSPACE(8192, PIN_GLOBAL | PIN_OFFSET_BIAS | (i915->ggtt.base.total - 4096)),
+#endif
+		{ },
+#undef NOSPACE
+#undef TOOBIG
+#undef INVALID
+#undef __INVALID
+#undef VALID
+	}, *m;
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err = -EINVAL;
+
+	/* Exercise all the weird and wonderful i915_vma_pin requests,
+	 * focusing on error handling of boundary conditions.
+	 */
+
+	GEM_BUG_ON(!drm_mm_clean(&i915->ggtt.base.mm));
+
+	obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vma = checked_vma_instance(obj, &i915->ggtt.base, NULL);
+	if (IS_ERR(vma))
+		goto out;
+
+	for (m = modes; m->assert; m++) {
+		err = i915_vma_pin(vma, m->size, 0, m->flags);
+		if (!m->assert(vma, m, err)) {
+			pr_err("%s to pin single page into GGTT with mode[%d:%s]: size=%llx flags=%llx, err=%d\n",
+			       m->assert == assert_pin_valid ? "Failed" : "Unexpectedly succeeded",
+			       (int)(m - modes), m->string, m->size, m->flags,
+			       err);
+			if (!err)
+				i915_vma_unpin(vma);
+			err = -EINVAL;
+			goto out;
+		}
+
+		if (!err) {
+			i915_vma_unpin(vma);
+			err = i915_vma_unbind(vma);
+			if (err) {
+				pr_err("Failed to unbind single page from GGTT, err=%d\n", err);
+				goto out;
+			}
+		}
+	}
+
+	err = 0;
+out:
+	i915_gem_object_put(obj);
+	return err;
+}
+
+static unsigned long rotated_index(const struct intel_rotation_info *r,
+				   unsigned int n,
+				   unsigned int x,
+				   unsigned int y)
+{
+	return (r->plane[n].stride * (r->plane[n].height - y - 1) +
+		r->plane[n].offset + x);
+}
+
+static struct scatterlist *
+assert_rotated(struct drm_i915_gem_object *obj,
+	       const struct intel_rotation_info *r, unsigned int n,
+	       struct scatterlist *sg)
+{
+	unsigned int x, y;
+
+	for (x = 0; x < r->plane[n].width; x++) {
+		for (y = 0; y < r->plane[n].height; y++) {
+			unsigned long src_idx;
+			dma_addr_t src;
+
+			if (!sg) {
+				pr_err("Invalid sg table: too short at plane %d, (%d, %d)!\n",
+				       n, x, y);
+				return ERR_PTR(-EINVAL);
+			}
+
+			src_idx = rotated_index(r, n, x, y);
+			src = i915_gem_object_get_dma_address(obj, src_idx);
+
+			if (sg_dma_len(sg) != PAGE_SIZE) {
+				pr_err("Invalid sg.length, found %d, expected %lu for rotated page (%d, %d) [src index %lu]\n",
+				       sg_dma_len(sg), PAGE_SIZE,
+				       x, y, src_idx);
+				return ERR_PTR(-EINVAL);
+			}
+
+			if (sg_dma_address(sg) != src) {
+				pr_err("Invalid address for rotated page (%d, %d) [src index %lu]\n",
+				       x, y, src_idx);
+				return ERR_PTR(-EINVAL);
+			}
+
+			sg = sg_next(sg);
+		}
+	}
+
+	return sg;
+}
+
+static unsigned int rotated_size(const struct intel_rotation_plane_info *a,
+				 const struct intel_rotation_plane_info *b)
+{
+	return a->width * a->height + b->width * b->height;
+}
+
+static int igt_vma_rotate(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_address_space *vm = &i915->ggtt.base;
+	struct drm_i915_gem_object *obj;
+	const struct intel_rotation_plane_info planes[] = {
+		{ .width = 1, .height = 1, .stride = 1 },
+		{ .width = 2, .height = 2, .stride = 2 },
+		{ .width = 4, .height = 4, .stride = 4 },
+		{ .width = 8, .height = 8, .stride = 8 },
+
+		{ .width = 3, .height = 5, .stride = 3 },
+		{ .width = 3, .height = 5, .stride = 4 },
+		{ .width = 3, .height = 5, .stride = 5 },
+
+		{ .width = 5, .height = 3, .stride = 5 },
+		{ .width = 5, .height = 3, .stride = 7 },
+		{ .width = 5, .height = 3, .stride = 9 },
+
+		{ .width = 4, .height = 6, .stride = 6 },
+		{ .width = 6, .height = 4, .stride = 6 },
+		{ }
+	}, *a, *b;
+	const unsigned int max_pages = 64;
+	int err = -ENOMEM;
+
+	/* Create VMA for many different combinations of planes and check
+	 * that the page layout within the rotated VMA match our expectations.
+	 */
+
+	obj = i915_gem_object_create_internal(i915, max_pages * PAGE_SIZE);
+	if (IS_ERR(obj))
+		goto out;
+
+	for (a = planes; a->width; a++) {
+		for (b = planes + ARRAY_SIZE(planes); b-- != planes; ) {
+			struct i915_ggtt_view view;
+			unsigned int n, max_offset;
+
+			max_offset = max(a->stride * a->height,
+					 b->stride * b->height);
+			GEM_BUG_ON(max_offset > max_pages);
+			max_offset = max_pages - max_offset;
+
+			view.type = I915_GGTT_VIEW_ROTATED;
+			view.rotated.plane[0] = *a;
+			view.rotated.plane[1] = *b;
+
+			for_each_prime_number_from(view.rotated.plane[0].offset, 0, max_offset) {
+				for_each_prime_number_from(view.rotated.plane[1].offset, 0, max_offset) {
+					struct scatterlist *sg;
+					struct i915_vma *vma;
+
+					vma = checked_vma_instance(obj, vm, &view);
+					if (IS_ERR(vma)) {
+						err = PTR_ERR(vma);
+						goto out_object;
+					}
+
+					err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+					if (err) {
+						pr_err("Failed to pin VMA, err=%d\n", err);
+						goto out_object;
+					}
+
+					if (vma->size != rotated_size(a, b) * PAGE_SIZE) {
+						pr_err("VMA is wrong size, expected %lu, found %llu\n",
+						       PAGE_SIZE * rotated_size(a, b), vma->size);
+						err = -EINVAL;
+						goto out_object;
+					}
+
+					if (vma->pages->nents != rotated_size(a, b)) {
+						pr_err("sg table is wrong sizeo, expected %u, found %u nents\n",
+						       rotated_size(a, b), vma->pages->nents);
+						err = -EINVAL;
+						goto out_object;
+					}
+
+					if (vma->node.size < vma->size) {
+						pr_err("VMA binding too small, expected %llu, found %llu\n",
+						       vma->size, vma->node.size);
+						err = -EINVAL;
+						goto out_object;
+					}
+
+					if (vma->pages == obj->mm.pages) {
+						pr_err("VMA using unrotated object pages!\n");
+						err = -EINVAL;
+						goto out_object;
+					}
+
+					sg = vma->pages->sgl;
+					for (n = 0; n < ARRAY_SIZE(view.rotated.plane); n++) {
+						sg = assert_rotated(obj, &view.rotated, n, sg);
+						if (IS_ERR(sg)) {
+							pr_err("Inconsistent VMA pages for plane %d: [(%d, %d, %d, %d), (%d, %d, %d, %d)]\n", n,
+							       view.rotated.plane[0].width,
+							       view.rotated.plane[0].height,
+							       view.rotated.plane[0].stride,
+							       view.rotated.plane[0].offset,
+							       view.rotated.plane[1].width,
+							       view.rotated.plane[1].height,
+							       view.rotated.plane[1].stride,
+							       view.rotated.plane[1].offset);
+							err = -EINVAL;
+							goto out_object;
+						}
+					}
+
+					i915_vma_unpin(vma);
+				}
+			}
+		}
+	}
+
+out_object:
+	i915_gem_object_put(obj);
+out:
+	return err;
+}
+
+static bool assert_partial(struct drm_i915_gem_object *obj,
+			   struct i915_vma *vma,
+			   unsigned long offset,
+			   unsigned long size)
+{
+	struct sgt_iter sgt;
+	dma_addr_t dma;
+
+	for_each_sgt_dma(dma, sgt, vma->pages) {
+		dma_addr_t src;
+
+		if (!size) {
+			pr_err("Partial scattergather list too long\n");
+			return false;
+		}
+
+		src = i915_gem_object_get_dma_address(obj, offset);
+		if (src != dma) {
+			pr_err("DMA mismatch for partial page offset %lu\n",
+			       offset);
+			return false;
+		}
+
+		offset++;
+		size--;
+	}
+
+	return true;
+}
+
+static bool assert_pin(struct i915_vma *vma,
+		       struct i915_ggtt_view *view,
+		       u64 size,
+		       const char *name)
+{
+	bool ok = true;
+
+	if (vma->size != size) {
+		pr_err("(%s) VMA is wrong size, expected %llu, found %llu\n",
+		       name, size, vma->size);
+		ok = false;
+	}
+
+	if (vma->node.size < vma->size) {
+		pr_err("(%s) VMA binding too small, expected %llu, found %llu\n",
+		       name, vma->size, vma->node.size);
+		ok = false;
+	}
+
+	if (view && view->type != I915_GGTT_VIEW_NORMAL) {
+		if (memcmp(&vma->ggtt_view, view, sizeof(*view))) {
+			pr_err("(%s) VMA mismatch upon creation!\n",
+			       name);
+			ok = false;
+		}
+
+		if (vma->pages == vma->obj->mm.pages) {
+			pr_err("(%s) VMA using original object pages!\n",
+			       name);
+			ok = false;
+		}
+	} else {
+		if (vma->ggtt_view.type != I915_GGTT_VIEW_NORMAL) {
+			pr_err("Not the normal ggtt view! Found %d\n",
+			       vma->ggtt_view.type);
+			ok = false;
+		}
+
+		if (vma->pages != vma->obj->mm.pages) {
+			pr_err("VMA not using object pages!\n");
+			ok = false;
+		}
+	}
+
+	return ok;
+}
+
+static int igt_vma_partial(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct i915_address_space *vm = &i915->ggtt.base;
+	const unsigned int npages = 1021; /* prime! */
+	struct drm_i915_gem_object *obj;
+	const struct phase {
+		const char *name;
+	} phases[] = {
+		{ "create" },
+		{ "lookup" },
+		{ },
+	}, *p;
+	unsigned int sz, offset;
+	struct i915_vma *vma;
+	int err = -ENOMEM;
+
+	/* Create lots of different VMA for the object and check that
+	 * we are returned the same VMA when we later request the same range.
+	 */
+
+	obj = i915_gem_object_create_internal(i915, npages*PAGE_SIZE);
+	if (IS_ERR(obj))
+		goto out;
+
+	for (p = phases; p->name; p++) { /* exercise both create/lookup */
+		unsigned int count, nvma;
+
+		nvma = 0;
+		for_each_prime_number_from(sz, 1, npages) {
+			for_each_prime_number_from(offset, 0, npages - sz) {
+				struct i915_ggtt_view view;
+
+				view.type = I915_GGTT_VIEW_PARTIAL;
+				view.partial.offset = offset;
+				view.partial.size = sz;
+
+				if (sz == npages)
+					view.type = I915_GGTT_VIEW_NORMAL;
+
+				vma = checked_vma_instance(obj, vm, &view);
+				if (IS_ERR(vma)) {
+					err = PTR_ERR(vma);
+					goto out_object;
+				}
+
+				err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+				if (err)
+					goto out_object;
+
+				if (!assert_pin(vma, &view, sz*PAGE_SIZE, p->name)) {
+					pr_err("(%s) Inconsistent partial pinning for (offset=%d, size=%d)\n",
+					       p->name, offset, sz);
+					err = -EINVAL;
+					goto out_object;
+				}
+
+				if (!assert_partial(obj, vma, offset, sz)) {
+					pr_err("(%s) Inconsistent partial pages for (offset=%d, size=%d)\n",
+					       p->name, offset, sz);
+					err = -EINVAL;
+					goto out_object;
+				}
+
+				i915_vma_unpin(vma);
+				nvma++;
+			}
+		}
+
+		count = 0;
+		list_for_each_entry(vma, &obj->vma_list, obj_link)
+			count++;
+		if (count != nvma) {
+			pr_err("(%s) All partial vma were not recorded on the obj->vma_list: found %u, expected %u\n",
+			       p->name, count, nvma);
+			err = -EINVAL;
+			goto out_object;
+		}
+
+		/* Check that we did create the whole object mapping */
+		vma = checked_vma_instance(obj, vm, NULL);
+		if (IS_ERR(vma)) {
+			err = PTR_ERR(vma);
+			goto out_object;
+		}
+
+		err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL);
+		if (err)
+			goto out_object;
+
+		if (!assert_pin(vma, NULL, obj->base.size, p->name)) {
+			pr_err("(%s) inconsistent full pin\n", p->name);
+			err = -EINVAL;
+			goto out_object;
+		}
+
+		i915_vma_unpin(vma);
+
+		count = 0;
+		list_for_each_entry(vma, &obj->vma_list, obj_link)
+			count++;
+		if (count != nvma) {
+			pr_err("(%s) allocated an extra full vma!\n", p->name);
+			err = -EINVAL;
+			goto out_object;
+		}
+	}
+
+out_object:
+	i915_gem_object_put(obj);
+out:
+	return err;
+}
+
+int i915_vma_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_vma_create),
+		SUBTEST(igt_vma_pin1),
+		SUBTEST(igt_vma_rotate),
+		SUBTEST(igt_vma_partial),
+	};
+	struct drm_i915_private *i915;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	mutex_lock(&i915->drm.struct_mutex);
+	err = i915_subtests(tests, i915);
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	drm_dev_unref(&i915->drm);
+	return err;
+}
+
diff --git a/drivers/gpu/drm/i915/selftests/intel_breadcrumbs.c b/drivers/gpu/drm/i915/selftests/intel_breadcrumbs.c
new file mode 100644
index 000000000000..19860a372d90
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/intel_breadcrumbs.c
@@ -0,0 +1,481 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "../i915_selftest.h"
+#include "i915_random.h"
+
+#include "mock_gem_device.h"
+#include "mock_engine.h"
+
+static int check_rbtree(struct intel_engine_cs *engine,
+			const unsigned long *bitmap,
+			const struct intel_wait *waiters,
+			const int count)
+{
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+	struct rb_node *rb;
+	int n;
+
+	if (&b->irq_wait->node != rb_first(&b->waiters)) {
+		pr_err("First waiter does not match first element of wait-tree\n");
+		return -EINVAL;
+	}
+
+	n = find_first_bit(bitmap, count);
+	for (rb = rb_first(&b->waiters); rb; rb = rb_next(rb)) {
+		struct intel_wait *w = container_of(rb, typeof(*w), node);
+		int idx = w - waiters;
+
+		if (!test_bit(idx, bitmap)) {
+			pr_err("waiter[%d, seqno=%d] removed but still in wait-tree\n",
+			       idx, w->seqno);
+			return -EINVAL;
+		}
+
+		if (n != idx) {
+			pr_err("waiter[%d, seqno=%d] does not match expected next element in tree [%d]\n",
+			       idx, w->seqno, n);
+			return -EINVAL;
+		}
+
+		n = find_next_bit(bitmap, count, n + 1);
+	}
+
+	return 0;
+}
+
+static int check_completion(struct intel_engine_cs *engine,
+			    const unsigned long *bitmap,
+			    const struct intel_wait *waiters,
+			    const int count)
+{
+	int n;
+
+	for (n = 0; n < count; n++) {
+		if (intel_wait_complete(&waiters[n]) != !!test_bit(n, bitmap))
+			continue;
+
+		pr_err("waiter[%d, seqno=%d] is %s, but expected %s\n",
+		       n, waiters[n].seqno,
+		       intel_wait_complete(&waiters[n]) ? "complete" : "active",
+		       test_bit(n, bitmap) ? "active" : "complete");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int check_rbtree_empty(struct intel_engine_cs *engine)
+{
+	struct intel_breadcrumbs *b = &engine->breadcrumbs;
+
+	if (b->irq_wait) {
+		pr_err("Empty breadcrumbs still has a waiter\n");
+		return -EINVAL;
+	}
+
+	if (!RB_EMPTY_ROOT(&b->waiters)) {
+		pr_err("Empty breadcrumbs, but wait-tree not empty\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int igt_random_insert_remove(void *arg)
+{
+	const u32 seqno_bias = 0x1000;
+	I915_RND_STATE(prng);
+	struct intel_engine_cs *engine = arg;
+	struct intel_wait *waiters;
+	const int count = 4096;
+	unsigned int *order;
+	unsigned long *bitmap;
+	int err = -ENOMEM;
+	int n;
+
+	mock_engine_reset(engine);
+
+	waiters = drm_malloc_gfp(count, sizeof(*waiters), GFP_TEMPORARY);
+	if (!waiters)
+		goto out_engines;
+
+	bitmap = kcalloc(DIV_ROUND_UP(count, BITS_PER_LONG), sizeof(*bitmap),
+			 GFP_TEMPORARY);
+	if (!bitmap)
+		goto out_waiters;
+
+	order = i915_random_order(count, &prng);
+	if (!order)
+		goto out_bitmap;
+
+	for (n = 0; n < count; n++)
+		intel_wait_init_for_seqno(&waiters[n], seqno_bias + n);
+
+	err = check_rbtree(engine, bitmap, waiters, count);
+	if (err)
+		goto out_order;
+
+	/* Add and remove waiters into the rbtree in random order. At each
+	 * step, we verify that the rbtree is correctly ordered.
+	 */
+	for (n = 0; n < count; n++) {
+		int i = order[n];
+
+		intel_engine_add_wait(engine, &waiters[i]);
+		__set_bit(i, bitmap);
+
+		err = check_rbtree(engine, bitmap, waiters, count);
+		if (err)
+			goto out_order;
+	}
+
+	i915_random_reorder(order, count, &prng);
+	for (n = 0; n < count; n++) {
+		int i = order[n];
+
+		intel_engine_remove_wait(engine, &waiters[i]);
+		__clear_bit(i, bitmap);
+
+		err = check_rbtree(engine, bitmap, waiters, count);
+		if (err)
+			goto out_order;
+	}
+
+	err = check_rbtree_empty(engine);
+out_order:
+	kfree(order);
+out_bitmap:
+	kfree(bitmap);
+out_waiters:
+	drm_free_large(waiters);
+out_engines:
+	mock_engine_flush(engine);
+	return err;
+}
+
+static int igt_insert_complete(void *arg)
+{
+	const u32 seqno_bias = 0x1000;
+	struct intel_engine_cs *engine = arg;
+	struct intel_wait *waiters;
+	const int count = 4096;
+	unsigned long *bitmap;
+	int err = -ENOMEM;
+	int n, m;
+
+	mock_engine_reset(engine);
+
+	waiters = drm_malloc_gfp(count, sizeof(*waiters), GFP_TEMPORARY);
+	if (!waiters)
+		goto out_engines;
+
+	bitmap = kcalloc(DIV_ROUND_UP(count, BITS_PER_LONG), sizeof(*bitmap),
+			 GFP_TEMPORARY);
+	if (!bitmap)
+		goto out_waiters;
+
+	for (n = 0; n < count; n++) {
+		intel_wait_init_for_seqno(&waiters[n], n + seqno_bias);
+		intel_engine_add_wait(engine, &waiters[n]);
+		__set_bit(n, bitmap);
+	}
+	err = check_rbtree(engine, bitmap, waiters, count);
+	if (err)
+		goto out_bitmap;
+
+	/* On each step, we advance the seqno so that several waiters are then
+	 * complete (we increase the seqno by increasingly larger values to
+	 * retire more and more waiters at once). All retired waiters should
+	 * be woken and removed from the rbtree, and so that we check.
+	 */
+	for (n = 0; n < count; n = m) {
+		int seqno = 2 * n;
+
+		GEM_BUG_ON(find_first_bit(bitmap, count) != n);
+
+		if (intel_wait_complete(&waiters[n])) {
+			pr_err("waiter[%d, seqno=%d] completed too early\n",
+			       n, waiters[n].seqno);
+			err = -EINVAL;
+			goto out_bitmap;
+		}
+
+		/* complete the following waiters */
+		mock_seqno_advance(engine, seqno + seqno_bias);
+		for (m = n; m <= seqno; m++) {
+			if (m == count)
+				break;
+
+			GEM_BUG_ON(!test_bit(m, bitmap));
+			__clear_bit(m, bitmap);
+		}
+
+		intel_engine_remove_wait(engine, &waiters[n]);
+		RB_CLEAR_NODE(&waiters[n].node);
+
+		err = check_rbtree(engine, bitmap, waiters, count);
+		if (err) {
+			pr_err("rbtree corrupt after seqno advance to %d\n",
+			       seqno + seqno_bias);
+			goto out_bitmap;
+		}
+
+		err = check_completion(engine, bitmap, waiters, count);
+		if (err) {
+			pr_err("completions after seqno advance to %d failed\n",
+			       seqno + seqno_bias);
+			goto out_bitmap;
+		}
+	}
+
+	err = check_rbtree_empty(engine);
+out_bitmap:
+	kfree(bitmap);
+out_waiters:
+	drm_free_large(waiters);
+out_engines:
+	mock_engine_flush(engine);
+	return err;
+}
+
+struct igt_wakeup {
+	struct task_struct *tsk;
+	atomic_t *ready, *set, *done;
+	struct intel_engine_cs *engine;
+	unsigned long flags;
+#define STOP 0
+#define IDLE 1
+	wait_queue_head_t *wq;
+	u32 seqno;
+};
+
+static int wait_atomic(atomic_t *p)
+{
+	schedule();
+	return 0;
+}
+
+static int wait_atomic_timeout(atomic_t *p)
+{
+	return schedule_timeout(10 * HZ) ? 0 : -ETIMEDOUT;
+}
+
+static bool wait_for_ready(struct igt_wakeup *w)
+{
+	DEFINE_WAIT(ready);
+
+	set_bit(IDLE, &w->flags);
+	if (atomic_dec_and_test(w->done))
+		wake_up_atomic_t(w->done);
+
+	if (test_bit(STOP, &w->flags))
+		goto out;
+
+	for (;;) {
+		prepare_to_wait(w->wq, &ready, TASK_INTERRUPTIBLE);
+		if (atomic_read(w->ready) == 0)
+			break;
+
+		schedule();
+	}
+	finish_wait(w->wq, &ready);
+
+out:
+	clear_bit(IDLE, &w->flags);
+	if (atomic_dec_and_test(w->set))
+		wake_up_atomic_t(w->set);
+
+	return !test_bit(STOP, &w->flags);
+}
+
+static int igt_wakeup_thread(void *arg)
+{
+	struct igt_wakeup *w = arg;
+	struct intel_wait wait;
+
+	while (wait_for_ready(w)) {
+		GEM_BUG_ON(kthread_should_stop());
+
+		intel_wait_init_for_seqno(&wait, w->seqno);
+		intel_engine_add_wait(w->engine, &wait);
+		for (;;) {
+			set_current_state(TASK_UNINTERRUPTIBLE);
+			if (i915_seqno_passed(intel_engine_get_seqno(w->engine),
+					      w->seqno))
+				break;
+
+			if (test_bit(STOP, &w->flags)) /* emergency escape */
+				break;
+
+			schedule();
+		}
+		intel_engine_remove_wait(w->engine, &wait);
+		__set_current_state(TASK_RUNNING);
+	}
+
+	return 0;
+}
+
+static void igt_wake_all_sync(atomic_t *ready,
+			      atomic_t *set,
+			      atomic_t *done,
+			      wait_queue_head_t *wq,
+			      int count)
+{
+	atomic_set(set, count);
+	atomic_set(ready, 0);
+	wake_up_all(wq);
+
+	wait_on_atomic_t(set, wait_atomic, TASK_UNINTERRUPTIBLE);
+	atomic_set(ready, count);
+	atomic_set(done, count);
+}
+
+static int igt_wakeup(void *arg)
+{
+	I915_RND_STATE(prng);
+	const int state = TASK_UNINTERRUPTIBLE;
+	struct intel_engine_cs *engine = arg;
+	struct igt_wakeup *waiters;
+	DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq);
+	const int count = 4096;
+	const u32 max_seqno = count / 4;
+	atomic_t ready, set, done;
+	int err = -ENOMEM;
+	int n, step;
+
+	mock_engine_reset(engine);
+
+	waiters = drm_malloc_gfp(count, sizeof(*waiters), GFP_TEMPORARY);
+	if (!waiters)
+		goto out_engines;
+
+	/* Create a large number of threads, each waiting on a random seqno.
+	 * Multiple waiters will be waiting for the same seqno.
+	 */
+	atomic_set(&ready, count);
+	for (n = 0; n < count; n++) {
+		waiters[n].wq = &wq;
+		waiters[n].ready = &ready;
+		waiters[n].set = &set;
+		waiters[n].done = &done;
+		waiters[n].engine = engine;
+		waiters[n].flags = BIT(IDLE);
+
+		waiters[n].tsk = kthread_run(igt_wakeup_thread, &waiters[n],
+					     "i915/igt:%d", n);
+		if (IS_ERR(waiters[n].tsk))
+			goto out_waiters;
+
+		get_task_struct(waiters[n].tsk);
+	}
+
+	for (step = 1; step <= max_seqno; step <<= 1) {
+		u32 seqno;
+
+		/* The waiter threads start paused as we assign them a random
+		 * seqno and reset the engine. Once the engine is reset,
+		 * we signal that the threads may begin their wait upon their
+		 * seqno.
+		 */
+		for (n = 0; n < count; n++) {
+			GEM_BUG_ON(!test_bit(IDLE, &waiters[n].flags));
+			waiters[n].seqno =
+				1 + prandom_u32_state(&prng) % max_seqno;
+		}
+		mock_seqno_advance(engine, 0);
+		igt_wake_all_sync(&ready, &set, &done, &wq, count);
+
+		/* Simulate the GPU doing chunks of work, with one or more
+		 * seqno appearing to finish at the same time. A random number
+		 * of threads will be waiting upon the update and hopefully be
+		 * woken.
+		 */
+		for (seqno = 1; seqno <= max_seqno + step; seqno += step) {
+			usleep_range(50, 500);
+			mock_seqno_advance(engine, seqno);
+		}
+		GEM_BUG_ON(intel_engine_get_seqno(engine) < 1 + max_seqno);
+
+		/* With the seqno now beyond any of the waiting threads, they
+		 * should all be woken, see that they are complete and signal
+		 * that they are ready for the next test. We wait until all
+		 * threads are complete and waiting for us (i.e. not a seqno).
+		 */
+		err = wait_on_atomic_t(&done, wait_atomic_timeout, state);
+		if (err) {
+			pr_err("Timed out waiting for %d remaining waiters\n",
+			       atomic_read(&done));
+			break;
+		}
+
+		err = check_rbtree_empty(engine);
+		if (err)
+			break;
+	}
+
+out_waiters:
+	for (n = 0; n < count; n++) {
+		if (IS_ERR(waiters[n].tsk))
+			break;
+
+		set_bit(STOP, &waiters[n].flags);
+	}
+	mock_seqno_advance(engine, INT_MAX); /* wakeup any broken waiters */
+	igt_wake_all_sync(&ready, &set, &done, &wq, n);
+
+	for (n = 0; n < count; n++) {
+		if (IS_ERR(waiters[n].tsk))
+			break;
+
+		kthread_stop(waiters[n].tsk);
+		put_task_struct(waiters[n].tsk);
+	}
+
+	drm_free_large(waiters);
+out_engines:
+	mock_engine_flush(engine);
+	return err;
+}
+
+int intel_breadcrumbs_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_random_insert_remove),
+		SUBTEST(igt_insert_complete),
+		SUBTEST(igt_wakeup),
+	};
+	struct drm_i915_private *i915;
+	int err;
+
+	i915 = mock_gem_device();
+	if (!i915)
+		return -ENOMEM;
+
+	err = i915_subtests(tests, i915->engine[RCS]);
+	drm_dev_unref(&i915->drm);
+
+	return err;
+}
diff --git a/drivers/gpu/drm/i915/selftests/intel_hangcheck.c b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c
new file mode 100644
index 000000000000..d4acee6730e9
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/intel_hangcheck.c
@@ -0,0 +1,537 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "../i915_selftest.h"
+
+struct hang {
+	struct drm_i915_private *i915;
+	struct drm_i915_gem_object *hws;
+	struct drm_i915_gem_object *obj;
+	u32 *seqno;
+	u32 *batch;
+};
+
+static int hang_init(struct hang *h, struct drm_i915_private *i915)
+{
+	void *vaddr;
+	int err;
+
+	memset(h, 0, sizeof(*h));
+	h->i915 = i915;
+
+	h->hws = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(h->hws))
+		return PTR_ERR(h->hws);
+
+	h->obj = i915_gem_object_create_internal(i915, PAGE_SIZE);
+	if (IS_ERR(h->obj)) {
+		err = PTR_ERR(h->obj);
+		goto err_hws;
+	}
+
+	i915_gem_object_set_cache_level(h->hws, I915_CACHE_LLC);
+	vaddr = i915_gem_object_pin_map(h->hws, I915_MAP_WB);
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		goto err_obj;
+	}
+	h->seqno = memset(vaddr, 0xff, PAGE_SIZE);
+
+	vaddr = i915_gem_object_pin_map(h->obj,
+					HAS_LLC(i915) ? I915_MAP_WB : I915_MAP_WC);
+	if (IS_ERR(vaddr)) {
+		err = PTR_ERR(vaddr);
+		goto err_unpin_hws;
+	}
+	h->batch = vaddr;
+
+	return 0;
+
+err_unpin_hws:
+	i915_gem_object_unpin_map(h->hws);
+err_obj:
+	i915_gem_object_put(h->obj);
+err_hws:
+	i915_gem_object_put(h->hws);
+	return err;
+}
+
+static u64 hws_address(const struct i915_vma *hws,
+		       const struct drm_i915_gem_request *rq)
+{
+	return hws->node.start + offset_in_page(sizeof(u32)*rq->fence.context);
+}
+
+static int emit_recurse_batch(struct hang *h,
+			      struct drm_i915_gem_request *rq)
+{
+	struct drm_i915_private *i915 = h->i915;
+	struct i915_address_space *vm = rq->ctx->ppgtt ? &rq->ctx->ppgtt->base : &i915->ggtt.base;
+	struct i915_vma *hws, *vma;
+	unsigned int flags;
+	u32 *batch;
+	int err;
+
+	vma = i915_vma_instance(h->obj, vm, NULL);
+	if (IS_ERR(vma))
+		return PTR_ERR(vma);
+
+	hws = i915_vma_instance(h->hws, vm, NULL);
+	if (IS_ERR(hws))
+		return PTR_ERR(hws);
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER);
+	if (err)
+		return err;
+
+	err = i915_vma_pin(hws, 0, 0, PIN_USER);
+	if (err)
+		goto unpin_vma;
+
+	err = rq->engine->emit_flush(rq, EMIT_INVALIDATE);
+	if (err)
+		goto unpin_hws;
+
+	err = i915_switch_context(rq);
+	if (err)
+		goto unpin_hws;
+
+	i915_vma_move_to_active(vma, rq, 0);
+	if (!i915_gem_object_has_active_reference(vma->obj)) {
+		i915_gem_object_get(vma->obj);
+		i915_gem_object_set_active_reference(vma->obj);
+	}
+
+	i915_vma_move_to_active(hws, rq, 0);
+	if (!i915_gem_object_has_active_reference(hws->obj)) {
+		i915_gem_object_get(hws->obj);
+		i915_gem_object_set_active_reference(hws->obj);
+	}
+
+	batch = h->batch;
+	if (INTEL_GEN(i915) >= 8) {
+		*batch++ = MI_STORE_DWORD_IMM_GEN4;
+		*batch++ = lower_32_bits(hws_address(hws, rq));
+		*batch++ = upper_32_bits(hws_address(hws, rq));
+		*batch++ = rq->fence.seqno;
+		*batch++ = MI_BATCH_BUFFER_START | 1 << 8 | 1;
+		*batch++ = lower_32_bits(vma->node.start);
+		*batch++ = upper_32_bits(vma->node.start);
+	} else if (INTEL_GEN(i915) >= 6) {
+		*batch++ = MI_STORE_DWORD_IMM_GEN4;
+		*batch++ = 0;
+		*batch++ = lower_32_bits(hws_address(hws, rq));
+		*batch++ = rq->fence.seqno;
+		*batch++ = MI_BATCH_BUFFER_START | 1 << 8;
+		*batch++ = lower_32_bits(vma->node.start);
+	} else if (INTEL_GEN(i915) >= 4) {
+		*batch++ = MI_STORE_DWORD_IMM_GEN4 | 1 << 22;
+		*batch++ = 0;
+		*batch++ = lower_32_bits(hws_address(hws, rq));
+		*batch++ = rq->fence.seqno;
+		*batch++ = MI_BATCH_BUFFER_START | 2 << 6;
+		*batch++ = lower_32_bits(vma->node.start);
+	} else {
+		*batch++ = MI_STORE_DWORD_IMM;
+		*batch++ = lower_32_bits(hws_address(hws, rq));
+		*batch++ = rq->fence.seqno;
+		*batch++ = MI_BATCH_BUFFER_START | 2 << 6 | 1;
+		*batch++ = lower_32_bits(vma->node.start);
+	}
+	*batch++ = MI_BATCH_BUFFER_END; /* not reached */
+
+	flags = 0;
+	if (INTEL_GEN(vm->i915) <= 5)
+		flags |= I915_DISPATCH_SECURE;
+
+	err = rq->engine->emit_bb_start(rq, vma->node.start, PAGE_SIZE, flags);
+
+unpin_hws:
+	i915_vma_unpin(hws);
+unpin_vma:
+	i915_vma_unpin(vma);
+	return err;
+}
+
+static struct drm_i915_gem_request *
+hang_create_request(struct hang *h,
+		    struct intel_engine_cs *engine,
+		    struct i915_gem_context *ctx)
+{
+	struct drm_i915_gem_request *rq;
+	int err;
+
+	if (i915_gem_object_is_active(h->obj)) {
+		struct drm_i915_gem_object *obj;
+		void *vaddr;
+
+		obj = i915_gem_object_create_internal(h->i915, PAGE_SIZE);
+		if (IS_ERR(obj))
+			return ERR_CAST(obj);
+
+		vaddr = i915_gem_object_pin_map(obj,
+						HAS_LLC(h->i915) ? I915_MAP_WB : I915_MAP_WC);
+		if (IS_ERR(vaddr)) {
+			i915_gem_object_put(obj);
+			return ERR_CAST(vaddr);
+		}
+
+		i915_gem_object_unpin_map(h->obj);
+		i915_gem_object_put(h->obj);
+
+		h->obj = obj;
+		h->batch = vaddr;
+	}
+
+	rq = i915_gem_request_alloc(engine, ctx);
+	if (IS_ERR(rq))
+		return rq;
+
+	err = emit_recurse_batch(h, rq);
+	if (err) {
+		__i915_add_request(rq, false);
+		return ERR_PTR(err);
+	}
+
+	return rq;
+}
+
+static u32 hws_seqno(const struct hang *h,
+		     const struct drm_i915_gem_request *rq)
+{
+	return READ_ONCE(h->seqno[rq->fence.context % (PAGE_SIZE/sizeof(u32))]);
+}
+
+static void hang_fini(struct hang *h)
+{
+	*h->batch = MI_BATCH_BUFFER_END;
+	wmb();
+
+	i915_gem_object_unpin_map(h->obj);
+	i915_gem_object_put(h->obj);
+
+	i915_gem_object_unpin_map(h->hws);
+	i915_gem_object_put(h->hws);
+
+	i915_gem_wait_for_idle(h->i915, I915_WAIT_LOCKED);
+	i915_gem_retire_requests(h->i915);
+}
+
+static int igt_hang_sanitycheck(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_request *rq;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct hang h;
+	int err;
+
+	/* Basic check that we can execute our hanging batch */
+
+	if (!igt_can_mi_store_dword_imm(i915))
+		return 0;
+
+	mutex_lock(&i915->drm.struct_mutex);
+	err = hang_init(&h, i915);
+	if (err)
+		goto unlock;
+
+	for_each_engine(engine, i915, id) {
+		long timeout;
+
+		rq = hang_create_request(&h, engine, i915->kernel_context);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			pr_err("Failed to create request for %s, err=%d\n",
+			       engine->name, err);
+			goto fini;
+		}
+
+		i915_gem_request_get(rq);
+
+		*h.batch = MI_BATCH_BUFFER_END;
+		__i915_add_request(rq, true);
+
+		timeout = i915_wait_request(rq,
+					    I915_WAIT_LOCKED,
+					    MAX_SCHEDULE_TIMEOUT);
+		i915_gem_request_put(rq);
+
+		if (timeout < 0) {
+			err = timeout;
+			pr_err("Wait for request failed on %s, err=%d\n",
+			       engine->name, err);
+			goto fini;
+		}
+	}
+
+fini:
+	hang_fini(&h);
+unlock:
+	mutex_unlock(&i915->drm.struct_mutex);
+	return err;
+}
+
+static int igt_global_reset(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	unsigned int reset_count;
+	int err = 0;
+
+	/* Check that we can issue a global GPU reset */
+
+	set_bit(I915_RESET_IN_PROGRESS, &i915->gpu_error.flags);
+
+	mutex_lock(&i915->drm.struct_mutex);
+	reset_count = i915_reset_count(&i915->gpu_error);
+
+	i915_reset(i915);
+
+	if (i915_reset_count(&i915->gpu_error) == reset_count) {
+		pr_err("No GPU reset recorded!\n");
+		err = -EINVAL;
+	}
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	GEM_BUG_ON(test_bit(I915_RESET_IN_PROGRESS, &i915->gpu_error.flags));
+	if (i915_terminally_wedged(&i915->gpu_error))
+		err = -EIO;
+
+	return err;
+}
+
+static u32 fake_hangcheck(struct drm_i915_gem_request *rq)
+{
+	u32 reset_count;
+
+	rq->engine->hangcheck.stalled = true;
+	rq->engine->hangcheck.seqno = intel_engine_get_seqno(rq->engine);
+
+	reset_count = i915_reset_count(&rq->i915->gpu_error);
+
+	set_bit(I915_RESET_IN_PROGRESS, &rq->i915->gpu_error.flags);
+	wake_up_all(&rq->i915->gpu_error.wait_queue);
+
+	return reset_count;
+}
+
+static bool wait_for_hang(struct hang *h, struct drm_i915_gem_request *rq)
+{
+	return !(wait_for_us(i915_seqno_passed(hws_seqno(h, rq),
+					       rq->fence.seqno),
+			     10) &&
+		 wait_for(i915_seqno_passed(hws_seqno(h, rq),
+					    rq->fence.seqno),
+			  1000));
+}
+
+static int igt_wait_reset(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct drm_i915_gem_request *rq;
+	unsigned int reset_count;
+	struct hang h;
+	long timeout;
+	int err;
+
+	/* Check that we detect a stuck waiter and issue a reset */
+
+	set_bit(I915_RESET_IN_PROGRESS, &i915->gpu_error.flags);
+
+	mutex_lock(&i915->drm.struct_mutex);
+	err = hang_init(&h, i915);
+	if (err)
+		goto unlock;
+
+	rq = hang_create_request(&h, i915->engine[RCS], i915->kernel_context);
+	if (IS_ERR(rq)) {
+		err = PTR_ERR(rq);
+		goto fini;
+	}
+
+	i915_gem_request_get(rq);
+	__i915_add_request(rq, true);
+
+	if (!wait_for_hang(&h, rq)) {
+		pr_err("Failed to start request %x\n", rq->fence.seqno);
+		err = -EIO;
+		goto out_rq;
+	}
+
+	reset_count = fake_hangcheck(rq);
+
+	timeout = i915_wait_request(rq, I915_WAIT_LOCKED, 10);
+	if (timeout < 0) {
+		pr_err("i915_wait_request failed on a stuck request: err=%ld\n",
+		       timeout);
+		err = timeout;
+		goto out_rq;
+	}
+	GEM_BUG_ON(test_bit(I915_RESET_IN_PROGRESS, &i915->gpu_error.flags));
+
+	if (i915_reset_count(&i915->gpu_error) == reset_count) {
+		pr_err("No GPU reset recorded!\n");
+		err = -EINVAL;
+		goto out_rq;
+	}
+
+out_rq:
+	i915_gem_request_put(rq);
+fini:
+	hang_fini(&h);
+unlock:
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	if (i915_terminally_wedged(&i915->gpu_error))
+		return -EIO;
+
+	return err;
+}
+
+static int igt_reset_queue(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct hang h;
+	int err;
+
+	/* Check that we replay pending requests following a hang */
+
+	if (!igt_can_mi_store_dword_imm(i915))
+		return 0;
+
+	mutex_lock(&i915->drm.struct_mutex);
+	err = hang_init(&h, i915);
+	if (err)
+		goto unlock;
+
+	for_each_engine(engine, i915, id) {
+		struct drm_i915_gem_request *prev;
+		IGT_TIMEOUT(end_time);
+		unsigned int count;
+
+		prev = hang_create_request(&h, engine, i915->kernel_context);
+		if (IS_ERR(prev)) {
+			err = PTR_ERR(prev);
+			goto fini;
+		}
+
+		i915_gem_request_get(prev);
+		__i915_add_request(prev, true);
+
+		count = 0;
+		do {
+			struct drm_i915_gem_request *rq;
+			unsigned int reset_count;
+
+			rq = hang_create_request(&h,
+						 engine,
+						 i915->kernel_context);
+			if (IS_ERR(rq)) {
+				err = PTR_ERR(rq);
+				goto fini;
+			}
+
+			i915_gem_request_get(rq);
+			__i915_add_request(rq, true);
+
+			if (!wait_for_hang(&h, prev)) {
+				pr_err("Failed to start request %x\n",
+				       prev->fence.seqno);
+				i915_gem_request_put(rq);
+				i915_gem_request_put(prev);
+				err = -EIO;
+				goto fini;
+			}
+
+			reset_count = fake_hangcheck(prev);
+
+			i915_reset(i915);
+
+			GEM_BUG_ON(test_bit(I915_RESET_IN_PROGRESS,
+					    &i915->gpu_error.flags));
+			if (prev->fence.error != -EIO) {
+				pr_err("GPU reset not recorded on hanging request [fence.error=%d]!\n",
+				       prev->fence.error);
+				i915_gem_request_put(rq);
+				i915_gem_request_put(prev);
+				err = -EINVAL;
+				goto fini;
+			}
+
+			if (rq->fence.error) {
+				pr_err("Fence error status not zero [%d] after unrelated reset\n",
+				       rq->fence.error);
+				i915_gem_request_put(rq);
+				i915_gem_request_put(prev);
+				err = -EINVAL;
+				goto fini;
+			}
+
+			if (i915_reset_count(&i915->gpu_error) == reset_count) {
+				pr_err("No GPU reset recorded!\n");
+				i915_gem_request_put(rq);
+				i915_gem_request_put(prev);
+				err = -EINVAL;
+				goto fini;
+			}
+
+			i915_gem_request_put(prev);
+			prev = rq;
+			count++;
+		} while (time_before(jiffies, end_time));
+		pr_info("%s: Completed %d resets\n", engine->name, count);
+
+		*h.batch = MI_BATCH_BUFFER_END;
+		wmb();
+
+		i915_gem_request_put(prev);
+	}
+
+fini:
+	hang_fini(&h);
+unlock:
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	if (i915_terminally_wedged(&i915->gpu_error))
+		return -EIO;
+
+	return err;
+}
+
+int intel_hangcheck_live_selftests(struct drm_i915_private *i915)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_hang_sanitycheck),
+		SUBTEST(igt_global_reset),
+		SUBTEST(igt_wait_reset),
+		SUBTEST(igt_reset_queue),
+	};
+
+	if (!intel_has_gpu_reset(i915))
+		return 0;
+
+	return i915_subtests(tests, i915);
+}
diff --git a/drivers/gpu/drm/i915/selftests/intel_uncore.c b/drivers/gpu/drm/i915/selftests/intel_uncore.c
new file mode 100644
index 000000000000..2d0fef2cfca6
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/intel_uncore.c
@@ -0,0 +1,182 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "../i915_selftest.h"
+
+static int intel_fw_table_check(const struct intel_forcewake_range *ranges,
+				unsigned int num_ranges,
+				bool is_watertight)
+{
+	unsigned int i;
+	s32 prev;
+
+	for (i = 0, prev = -1; i < num_ranges; i++, ranges++) {
+		/* Check that the table is watertight */
+		if (is_watertight && (prev + 1) != (s32)ranges->start) {
+			pr_err("%s: entry[%d]:(%x, %x) is not watertight to previous (%x)\n",
+			       __func__, i, ranges->start, ranges->end, prev);
+			return -EINVAL;
+		}
+
+		/* Check that the table never goes backwards */
+		if (prev >= (s32)ranges->start) {
+			pr_err("%s: entry[%d]:(%x, %x) is less than the previous (%x)\n",
+			       __func__, i, ranges->start, ranges->end, prev);
+			return -EINVAL;
+		}
+
+		/* Check that the entry is valid */
+		if (ranges->start >= ranges->end) {
+			pr_err("%s: entry[%d]:(%x, %x) has negative length\n",
+			       __func__, i, ranges->start, ranges->end);
+			return -EINVAL;
+		}
+
+		prev = ranges->end;
+	}
+
+	return 0;
+}
+
+static int intel_shadow_table_check(void)
+{
+	const i915_reg_t *reg = gen8_shadowed_regs;
+	unsigned int i;
+	s32 prev;
+
+	for (i = 0, prev = -1; i < ARRAY_SIZE(gen8_shadowed_regs); i++, reg++) {
+		u32 offset = i915_mmio_reg_offset(*reg);
+
+		if (prev >= (s32)offset) {
+			pr_err("%s: entry[%d]:(%x) is before previous (%x)\n",
+			       __func__, i, offset, prev);
+			return -EINVAL;
+		}
+
+		prev = offset;
+	}
+
+	return 0;
+}
+
+int intel_uncore_mock_selftests(void)
+{
+	struct {
+		const struct intel_forcewake_range *ranges;
+		unsigned int num_ranges;
+		bool is_watertight;
+	} fw[] = {
+		{ __vlv_fw_ranges, ARRAY_SIZE(__vlv_fw_ranges), false },
+		{ __chv_fw_ranges, ARRAY_SIZE(__chv_fw_ranges), false },
+		{ __gen9_fw_ranges, ARRAY_SIZE(__gen9_fw_ranges), true },
+	};
+	int err, i;
+
+	for (i = 0; i < ARRAY_SIZE(fw); i++) {
+		err = intel_fw_table_check(fw[i].ranges,
+					   fw[i].num_ranges,
+					   fw[i].is_watertight);
+		if (err)
+			return err;
+	}
+
+	err = intel_shadow_table_check();
+	if (err)
+		return err;
+
+	return 0;
+}
+
+static int intel_uncore_check_forcewake_domains(struct drm_i915_private *dev_priv)
+{
+#define FW_RANGE 0x40000
+	unsigned long *valid;
+	u32 offset;
+	int err;
+
+	if (!HAS_FPGA_DBG_UNCLAIMED(dev_priv) &&
+	    !IS_VALLEYVIEW(dev_priv) &&
+	    !IS_CHERRYVIEW(dev_priv))
+		return 0;
+
+	if (IS_VALLEYVIEW(dev_priv)) /* XXX system lockup! */
+		return 0;
+
+	if (IS_BROADWELL(dev_priv)) /* XXX random GPU hang afterwards! */
+		return 0;
+
+	valid = kzalloc(BITS_TO_LONGS(FW_RANGE) * sizeof(*valid),
+			GFP_TEMPORARY);
+	if (!valid)
+		return -ENOMEM;
+
+	intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL);
+
+	check_for_unclaimed_mmio(dev_priv);
+	for (offset = 0; offset < FW_RANGE; offset += 4) {
+		i915_reg_t reg = { offset };
+
+		(void)I915_READ_FW(reg);
+		if (!check_for_unclaimed_mmio(dev_priv))
+			set_bit(offset, valid);
+	}
+
+	intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);
+
+	err = 0;
+	for_each_set_bit(offset, valid, FW_RANGE) {
+		i915_reg_t reg = { offset };
+
+		intel_uncore_forcewake_reset(dev_priv, false);
+		check_for_unclaimed_mmio(dev_priv);
+
+		(void)I915_READ(reg);
+		if (check_for_unclaimed_mmio(dev_priv)) {
+			pr_err("Unclaimed mmio read to register 0x%04x\n",
+			       offset);
+			err = -EINVAL;
+		}
+	}
+
+	kfree(valid);
+	return err;
+}
+
+int intel_uncore_live_selftests(struct drm_i915_private *i915)
+{
+	int err;
+
+	/* Confirm the table we load is still valid */
+	err = intel_fw_table_check(i915->uncore.fw_domains_table,
+				   i915->uncore.fw_domains_table_entries,
+				   INTEL_GEN(i915) >= 9);
+	if (err)
+		return err;
+
+	err = intel_uncore_check_forcewake_domains(i915);
+	if (err)
+		return err;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_context.c b/drivers/gpu/drm/i915/selftests/mock_context.c
new file mode 100644
index 000000000000..8d3a90c3f8ac
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_context.c
@@ -0,0 +1,78 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "mock_context.h"
+#include "mock_gtt.h"
+
+struct i915_gem_context *
+mock_context(struct drm_i915_private *i915,
+	     const char *name)
+{
+	struct i915_gem_context *ctx;
+	int ret;
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return NULL;
+
+	kref_init(&ctx->ref);
+	INIT_LIST_HEAD(&ctx->link);
+	ctx->i915 = i915;
+
+	ret = ida_simple_get(&i915->context_hw_ida,
+			     0, MAX_CONTEXT_HW_ID, GFP_KERNEL);
+	if (ret < 0)
+		goto err_free;
+	ctx->hw_id = ret;
+
+	if (name) {
+		ctx->name = kstrdup(name, GFP_KERNEL);
+		if (!ctx->name)
+			goto err_put;
+
+		ctx->ppgtt = mock_ppgtt(i915, name);
+		if (!ctx->ppgtt)
+			goto err_put;
+	}
+
+	return ctx;
+
+err_free:
+	kfree(ctx);
+	return NULL;
+
+err_put:
+	i915_gem_context_set_closed(ctx);
+	i915_gem_context_put(ctx);
+	return NULL;
+}
+
+void mock_context_close(struct i915_gem_context *ctx)
+{
+	i915_gem_context_set_closed(ctx);
+
+	i915_ppgtt_close(&ctx->ppgtt->base);
+
+	i915_gem_context_put(ctx);
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_context.h b/drivers/gpu/drm/i915/selftests/mock_context.h
new file mode 100644
index 000000000000..2427e5c0916a
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_context.h
@@ -0,0 +1,34 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MOCK_CONTEXT_H
+#define __MOCK_CONTEXT_H
+
+struct i915_gem_context *
+mock_context(struct drm_i915_private *i915,
+	     const char *name);
+
+void mock_context_close(struct i915_gem_context *ctx);
+
+#endif /* !__MOCK_CONTEXT_H */
diff --git a/drivers/gpu/drm/i915/selftests/mock_dmabuf.c b/drivers/gpu/drm/i915/selftests/mock_dmabuf.c
new file mode 100644
index 000000000000..99da8f4ef497
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_dmabuf.c
@@ -0,0 +1,176 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "mock_dmabuf.h"
+
+static struct sg_table *mock_map_dma_buf(struct dma_buf_attachment *attachment,
+					 enum dma_data_direction dir)
+{
+	struct mock_dmabuf *mock = to_mock(attachment->dmabuf);
+	struct sg_table *st;
+	struct scatterlist *sg;
+	int i, err;
+
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (!st)
+		return ERR_PTR(-ENOMEM);
+
+	err = sg_alloc_table(st, mock->npages, GFP_KERNEL);
+	if (err)
+		goto err_free;
+
+	sg = st->sgl;
+	for (i = 0; i < mock->npages; i++) {
+		sg_set_page(sg, mock->pages[i], PAGE_SIZE, 0);
+		sg = sg_next(sg);
+	}
+
+	if (!dma_map_sg(attachment->dev, st->sgl, st->nents, dir)) {
+		err = -ENOMEM;
+		goto err_st;
+	}
+
+	return st;
+
+err_st:
+	sg_free_table(st);
+err_free:
+	kfree(st);
+	return ERR_PTR(err);
+}
+
+static void mock_unmap_dma_buf(struct dma_buf_attachment *attachment,
+			       struct sg_table *st,
+			       enum dma_data_direction dir)
+{
+	dma_unmap_sg(attachment->dev, st->sgl, st->nents, dir);
+	sg_free_table(st);
+	kfree(st);
+}
+
+static void mock_dmabuf_release(struct dma_buf *dma_buf)
+{
+	struct mock_dmabuf *mock = to_mock(dma_buf);
+	int i;
+
+	for (i = 0; i < mock->npages; i++)
+		put_page(mock->pages[i]);
+
+	kfree(mock);
+}
+
+static void *mock_dmabuf_vmap(struct dma_buf *dma_buf)
+{
+	struct mock_dmabuf *mock = to_mock(dma_buf);
+
+	return vm_map_ram(mock->pages, mock->npages, 0, PAGE_KERNEL);
+}
+
+static void mock_dmabuf_vunmap(struct dma_buf *dma_buf, void *vaddr)
+{
+	struct mock_dmabuf *mock = to_mock(dma_buf);
+
+	vm_unmap_ram(vaddr, mock->npages);
+}
+
+static void *mock_dmabuf_kmap_atomic(struct dma_buf *dma_buf, unsigned long page_num)
+{
+	struct mock_dmabuf *mock = to_mock(dma_buf);
+
+	return kmap_atomic(mock->pages[page_num]);
+}
+
+static void mock_dmabuf_kunmap_atomic(struct dma_buf *dma_buf, unsigned long page_num, void *addr)
+{
+	kunmap_atomic(addr);
+}
+
+static void *mock_dmabuf_kmap(struct dma_buf *dma_buf, unsigned long page_num)
+{
+	struct mock_dmabuf *mock = to_mock(dma_buf);
+
+	return kmap(mock->pages[page_num]);
+}
+
+static void mock_dmabuf_kunmap(struct dma_buf *dma_buf, unsigned long page_num, void *addr)
+{
+	struct mock_dmabuf *mock = to_mock(dma_buf);
+
+	return kunmap(mock->pages[page_num]);
+}
+
+static int mock_dmabuf_mmap(struct dma_buf *dma_buf, struct vm_area_struct *vma)
+{
+	return -ENODEV;
+}
+
+static const struct dma_buf_ops mock_dmabuf_ops =  {
+	.map_dma_buf = mock_map_dma_buf,
+	.unmap_dma_buf = mock_unmap_dma_buf,
+	.release = mock_dmabuf_release,
+	.kmap = mock_dmabuf_kmap,
+	.kmap_atomic = mock_dmabuf_kmap_atomic,
+	.kunmap = mock_dmabuf_kunmap,
+	.kunmap_atomic = mock_dmabuf_kunmap_atomic,
+	.mmap = mock_dmabuf_mmap,
+	.vmap = mock_dmabuf_vmap,
+	.vunmap = mock_dmabuf_vunmap,
+};
+
+static struct dma_buf *mock_dmabuf(int npages)
+{
+	struct mock_dmabuf *mock;
+	DEFINE_DMA_BUF_EXPORT_INFO(exp_info);
+	struct dma_buf *dmabuf;
+	int i;
+
+	mock = kmalloc(sizeof(*mock) + npages * sizeof(struct page *),
+		       GFP_KERNEL);
+	if (!mock)
+		return ERR_PTR(-ENOMEM);
+
+	mock->npages = npages;
+	for (i = 0; i < npages; i++) {
+		mock->pages[i] = alloc_page(GFP_KERNEL);
+		if (!mock->pages[i])
+			goto err;
+	}
+
+	exp_info.ops = &mock_dmabuf_ops;
+	exp_info.size = npages * PAGE_SIZE;
+	exp_info.flags = O_CLOEXEC;
+	exp_info.priv = mock;
+
+	dmabuf = dma_buf_export(&exp_info);
+	if (IS_ERR(dmabuf))
+		goto err;
+
+	return dmabuf;
+
+err:
+	while (i--)
+		put_page(mock->pages[i]);
+	kfree(mock);
+	return ERR_PTR(-ENOMEM);
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_dmabuf.h b/drivers/gpu/drm/i915/selftests/mock_dmabuf.h
new file mode 100644
index 000000000000..ec80613159b9
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_dmabuf.h
@@ -0,0 +1,41 @@
+
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MOCK_DMABUF_H__
+#define __MOCK_DMABUF_H__
+
+#include <linux/dma-buf.h>
+
+struct mock_dmabuf {
+	int npages;
+	struct page *pages[];
+};
+
+static struct mock_dmabuf *to_mock(struct dma_buf *buf)
+{
+	return buf->priv;
+}
+
+#endif /* !__MOCK_DMABUF_H__ */
diff --git a/drivers/gpu/drm/i915/selftests/mock_drm.c b/drivers/gpu/drm/i915/selftests/mock_drm.c
new file mode 100644
index 000000000000..113dec05c7dc
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_drm.c
@@ -0,0 +1,54 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "mock_drm.h"
+
+static inline struct inode fake_inode(struct drm_i915_private *i915)
+{
+	return (struct inode){ .i_rdev = i915->drm.primary->index };
+}
+
+struct drm_file *mock_file(struct drm_i915_private *i915)
+{
+	struct inode inode = fake_inode(i915);
+	struct file filp = {};
+	struct drm_file *file;
+	int err;
+
+	err = drm_open(&inode, &filp);
+	if (unlikely(err))
+		return ERR_PTR(err);
+
+	file = filp.private_data;
+	file->authenticated = true;
+	return file;
+}
+
+void mock_file_free(struct drm_i915_private *i915, struct drm_file *file)
+{
+	struct inode inode = fake_inode(i915);
+	struct file filp = { .private_data = file };
+
+	drm_release(&inode, &filp);
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_drm.h b/drivers/gpu/drm/i915/selftests/mock_drm.h
new file mode 100644
index 000000000000..b39beee9f8f6
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_drm.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright © 2017 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MOCK_DRM_H
+#define __MOCK_DRM_H
+
+struct drm_file *mock_file(struct drm_i915_private *i915);
+void mock_file_free(struct drm_i915_private *i915, struct drm_file *file);
+
+#endif /* !__MOCK_DRM_H */
diff --git a/drivers/gpu/drm/i915/selftests/mock_engine.c b/drivers/gpu/drm/i915/selftests/mock_engine.c
new file mode 100644
index 000000000000..8d5ba037064c
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_engine.c
@@ -0,0 +1,207 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "mock_engine.h"
+#include "mock_request.h"
+
+static struct mock_request *first_request(struct mock_engine *engine)
+{
+	return list_first_entry_or_null(&engine->hw_queue,
+					struct mock_request,
+					link);
+}
+
+static void hw_delay_complete(unsigned long data)
+{
+	struct mock_engine *engine = (typeof(engine))data;
+	struct mock_request *request;
+
+	spin_lock(&engine->hw_lock);
+
+	request = first_request(engine);
+	if (request) {
+		list_del_init(&request->link);
+		mock_seqno_advance(&engine->base, request->base.global_seqno);
+	}
+
+	request = first_request(engine);
+	if (request)
+		mod_timer(&engine->hw_delay, jiffies + request->delay);
+
+	spin_unlock(&engine->hw_lock);
+}
+
+static int mock_context_pin(struct intel_engine_cs *engine,
+			    struct i915_gem_context *ctx)
+{
+	i915_gem_context_get(ctx);
+	return 0;
+}
+
+static void mock_context_unpin(struct intel_engine_cs *engine,
+			       struct i915_gem_context *ctx)
+{
+	i915_gem_context_put(ctx);
+}
+
+static int mock_request_alloc(struct drm_i915_gem_request *request)
+{
+	struct mock_request *mock = container_of(request, typeof(*mock), base);
+
+	INIT_LIST_HEAD(&mock->link);
+	mock->delay = 0;
+
+	request->ring = request->engine->buffer;
+	return 0;
+}
+
+static int mock_emit_flush(struct drm_i915_gem_request *request,
+			   unsigned int flags)
+{
+	return 0;
+}
+
+static void mock_emit_breadcrumb(struct drm_i915_gem_request *request,
+				 u32 *flags)
+{
+}
+
+static void mock_submit_request(struct drm_i915_gem_request *request)
+{
+	struct mock_request *mock = container_of(request, typeof(*mock), base);
+	struct mock_engine *engine =
+		container_of(request->engine, typeof(*engine), base);
+
+	i915_gem_request_submit(request);
+	GEM_BUG_ON(!request->global_seqno);
+
+	spin_lock_irq(&engine->hw_lock);
+	list_add_tail(&mock->link, &engine->hw_queue);
+	if (mock->link.prev == &engine->hw_queue)
+		mod_timer(&engine->hw_delay, jiffies + mock->delay);
+	spin_unlock_irq(&engine->hw_lock);
+}
+
+static struct intel_ring *mock_ring(struct intel_engine_cs *engine)
+{
+	const unsigned long sz = roundup_pow_of_two(sizeof(struct intel_ring));
+	struct intel_ring *ring;
+
+	ring = kzalloc(sizeof(*ring) + sz, GFP_KERNEL);
+	if (!ring)
+		return NULL;
+
+	ring->engine = engine;
+	ring->size = sz;
+	ring->effective_size = sz;
+	ring->vaddr = (void *)(ring + 1);
+
+	INIT_LIST_HEAD(&ring->request_list);
+	ring->last_retired_head = -1;
+	intel_ring_update_space(ring);
+
+	return ring;
+}
+
+struct intel_engine_cs *mock_engine(struct drm_i915_private *i915,
+				    const char *name)
+{
+	struct mock_engine *engine;
+	static int id;
+
+	engine = kzalloc(sizeof(*engine) + PAGE_SIZE, GFP_KERNEL);
+	if (!engine)
+		return NULL;
+
+	engine->base.buffer = mock_ring(&engine->base);
+	if (!engine->base.buffer) {
+		kfree(engine);
+		return NULL;
+	}
+
+	/* minimal engine setup for requests */
+	engine->base.i915 = i915;
+	engine->base.name = name;
+	engine->base.id = id++;
+	engine->base.status_page.page_addr = (void *)(engine + 1);
+
+	engine->base.context_pin = mock_context_pin;
+	engine->base.context_unpin = mock_context_unpin;
+	engine->base.request_alloc = mock_request_alloc;
+	engine->base.emit_flush = mock_emit_flush;
+	engine->base.emit_breadcrumb = mock_emit_breadcrumb;
+	engine->base.submit_request = mock_submit_request;
+
+	engine->base.timeline =
+		&i915->gt.global_timeline.engine[engine->base.id];
+
+	intel_engine_init_breadcrumbs(&engine->base);
+	engine->base.breadcrumbs.mock = true; /* prevent touching HW for irqs */
+
+	/* fake hw queue */
+	spin_lock_init(&engine->hw_lock);
+	setup_timer(&engine->hw_delay,
+		    hw_delay_complete,
+		    (unsigned long)engine);
+	INIT_LIST_HEAD(&engine->hw_queue);
+
+	return &engine->base;
+}
+
+void mock_engine_flush(struct intel_engine_cs *engine)
+{
+	struct mock_engine *mock =
+		container_of(engine, typeof(*mock), base);
+	struct mock_request *request, *rn;
+
+	del_timer_sync(&mock->hw_delay);
+
+	spin_lock_irq(&mock->hw_lock);
+	list_for_each_entry_safe(request, rn, &mock->hw_queue, link) {
+		list_del_init(&request->link);
+		mock_seqno_advance(&mock->base, request->base.global_seqno);
+	}
+	spin_unlock_irq(&mock->hw_lock);
+}
+
+void mock_engine_reset(struct intel_engine_cs *engine)
+{
+	intel_write_status_page(engine, I915_GEM_HWS_INDEX, 0);
+}
+
+void mock_engine_free(struct intel_engine_cs *engine)
+{
+	struct mock_engine *mock =
+		container_of(engine, typeof(*mock), base);
+
+	GEM_BUG_ON(timer_pending(&mock->hw_delay));
+
+	if (engine->last_retired_context)
+		engine->context_unpin(engine, engine->last_retired_context);
+
+	intel_engine_fini_breadcrumbs(engine);
+
+	kfree(engine->buffer);
+	kfree(engine);
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_engine.h b/drivers/gpu/drm/i915/selftests/mock_engine.h
new file mode 100644
index 000000000000..e5e240216ba3
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_engine.h
@@ -0,0 +1,54 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MOCK_ENGINE_H__
+#define __MOCK_ENGINE_H__
+
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/timer.h>
+
+#include "../intel_ringbuffer.h"
+
+struct mock_engine {
+	struct intel_engine_cs base;
+
+	spinlock_t hw_lock;
+	struct list_head hw_queue;
+	struct timer_list hw_delay;
+};
+
+struct intel_engine_cs *mock_engine(struct drm_i915_private *i915,
+				    const char *name);
+void mock_engine_flush(struct intel_engine_cs *engine);
+void mock_engine_reset(struct intel_engine_cs *engine);
+void mock_engine_free(struct intel_engine_cs *engine);
+
+static inline void mock_seqno_advance(struct intel_engine_cs *engine, u32 seqno)
+{
+	intel_write_status_page(engine, I915_GEM_HWS_INDEX, seqno);
+	intel_engine_wakeup(engine);
+}
+
+#endif /* !__MOCK_ENGINE_H__ */
diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_device.c b/drivers/gpu/drm/i915/selftests/mock_gem_device.c
new file mode 100644
index 000000000000..6a8258eacdcb
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_gem_device.c
@@ -0,0 +1,226 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/pm_runtime.h>
+
+#include "mock_engine.h"
+#include "mock_context.h"
+#include "mock_request.h"
+#include "mock_gem_device.h"
+#include "mock_gem_object.h"
+#include "mock_gtt.h"
+
+void mock_device_flush(struct drm_i915_private *i915)
+{
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	lockdep_assert_held(&i915->drm.struct_mutex);
+
+	for_each_engine(engine, i915, id)
+		mock_engine_flush(engine);
+
+	i915_gem_retire_requests(i915);
+}
+
+static void mock_device_release(struct drm_device *dev)
+{
+	struct drm_i915_private *i915 = to_i915(dev);
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	mutex_lock(&i915->drm.struct_mutex);
+	mock_device_flush(i915);
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	cancel_delayed_work_sync(&i915->gt.retire_work);
+	cancel_delayed_work_sync(&i915->gt.idle_work);
+
+	mutex_lock(&i915->drm.struct_mutex);
+	for_each_engine(engine, i915, id)
+		mock_engine_free(engine);
+	i915_gem_context_fini(i915);
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	drain_workqueue(i915->wq);
+	i915_gem_drain_freed_objects(i915);
+
+	mutex_lock(&i915->drm.struct_mutex);
+	mock_fini_ggtt(i915);
+	i915_gem_timeline_fini(&i915->gt.global_timeline);
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	destroy_workqueue(i915->wq);
+
+	kmem_cache_destroy(i915->dependencies);
+	kmem_cache_destroy(i915->requests);
+	kmem_cache_destroy(i915->vmas);
+	kmem_cache_destroy(i915->objects);
+
+	drm_dev_fini(&i915->drm);
+	put_device(&i915->drm.pdev->dev);
+}
+
+static struct drm_driver mock_driver = {
+	.name = "mock",
+	.driver_features = DRIVER_GEM,
+	.release = mock_device_release,
+
+	.gem_close_object = i915_gem_close_object,
+	.gem_free_object_unlocked = i915_gem_free_object,
+};
+
+static void release_dev(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+
+	kfree(pdev);
+}
+
+static void mock_retire_work_handler(struct work_struct *work)
+{
+}
+
+static void mock_idle_work_handler(struct work_struct *work)
+{
+}
+
+struct drm_i915_private *mock_gem_device(void)
+{
+	struct drm_i915_private *i915;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+	struct pci_dev *pdev;
+	int err;
+
+	pdev = kzalloc(sizeof(*pdev) + sizeof(*i915), GFP_KERNEL);
+	if (!pdev)
+		goto err;
+
+	device_initialize(&pdev->dev);
+	pdev->dev.release = release_dev;
+	dev_set_name(&pdev->dev, "mock");
+	dma_coerce_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(32));
+
+	pm_runtime_dont_use_autosuspend(&pdev->dev);
+	pm_runtime_get_sync(&pdev->dev);
+
+	i915 = (struct drm_i915_private *)(pdev + 1);
+	pci_set_drvdata(pdev, i915);
+
+	err = drm_dev_init(&i915->drm, &mock_driver, &pdev->dev);
+	if (err) {
+		pr_err("Failed to initialise mock GEM device: err=%d\n", err);
+		goto put_device;
+	}
+	i915->drm.pdev = pdev;
+	i915->drm.dev_private = i915;
+
+	/* Using the global GTT may ask questions about KMS users, so prepare */
+	drm_mode_config_init(&i915->drm);
+
+	mkwrite_device_info(i915)->gen = -1;
+
+	spin_lock_init(&i915->mm.object_stat_lock);
+
+	init_waitqueue_head(&i915->gpu_error.wait_queue);
+	init_waitqueue_head(&i915->gpu_error.reset_queue);
+
+	i915->wq = alloc_ordered_workqueue("mock", 0);
+	if (!i915->wq)
+		goto put_device;
+
+	INIT_WORK(&i915->mm.free_work, __i915_gem_free_work);
+	init_llist_head(&i915->mm.free_list);
+	INIT_LIST_HEAD(&i915->mm.unbound_list);
+	INIT_LIST_HEAD(&i915->mm.bound_list);
+
+	ida_init(&i915->context_hw_ida);
+
+	INIT_DELAYED_WORK(&i915->gt.retire_work, mock_retire_work_handler);
+	INIT_DELAYED_WORK(&i915->gt.idle_work, mock_idle_work_handler);
+
+	i915->gt.awake = true;
+
+	i915->objects = KMEM_CACHE(mock_object, SLAB_HWCACHE_ALIGN);
+	if (!i915->objects)
+		goto err_wq;
+
+	i915->vmas = KMEM_CACHE(i915_vma, SLAB_HWCACHE_ALIGN);
+	if (!i915->vmas)
+		goto err_objects;
+
+	i915->requests = KMEM_CACHE(mock_request,
+				    SLAB_HWCACHE_ALIGN |
+				    SLAB_RECLAIM_ACCOUNT |
+				    SLAB_DESTROY_BY_RCU);
+	if (!i915->requests)
+		goto err_vmas;
+
+	i915->dependencies = KMEM_CACHE(i915_dependency,
+					SLAB_HWCACHE_ALIGN |
+					SLAB_RECLAIM_ACCOUNT);
+	if (!i915->dependencies)
+		goto err_requests;
+
+	mutex_lock(&i915->drm.struct_mutex);
+	INIT_LIST_HEAD(&i915->gt.timelines);
+	err = i915_gem_timeline_init__global(i915);
+	if (err) {
+		mutex_unlock(&i915->drm.struct_mutex);
+		goto err_dependencies;
+	}
+
+	mock_init_ggtt(i915);
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	mkwrite_device_info(i915)->ring_mask = BIT(0);
+	i915->engine[RCS] = mock_engine(i915, "mock");
+	if (!i915->engine[RCS])
+		goto err_dependencies;
+
+	i915->kernel_context = mock_context(i915, NULL);
+	if (!i915->kernel_context)
+		goto err_engine;
+
+	return i915;
+
+err_engine:
+	for_each_engine(engine, i915, id)
+		mock_engine_free(engine);
+err_dependencies:
+	kmem_cache_destroy(i915->dependencies);
+err_requests:
+	kmem_cache_destroy(i915->requests);
+err_vmas:
+	kmem_cache_destroy(i915->vmas);
+err_objects:
+	kmem_cache_destroy(i915->objects);
+err_wq:
+	destroy_workqueue(i915->wq);
+put_device:
+	put_device(&pdev->dev);
+err:
+	return NULL;
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_device.h b/drivers/gpu/drm/i915/selftests/mock_gem_device.h
new file mode 100644
index 000000000000..4cca4d57f52c
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_gem_device.h
@@ -0,0 +1,9 @@
+#ifndef __MOCK_GEM_DEVICE_H__
+#define __MOCK_GEM_DEVICE_H__
+
+struct drm_i915_private;
+
+struct drm_i915_private *mock_gem_device(void);
+void mock_device_flush(struct drm_i915_private *i915);
+
+#endif /* !__MOCK_GEM_DEVICE_H__ */
diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_object.h b/drivers/gpu/drm/i915/selftests/mock_gem_object.h
new file mode 100644
index 000000000000..9fbf67321662
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_gem_object.h
@@ -0,0 +1,8 @@
+#ifndef __MOCK_GEM_OBJECT_H__
+#define __MOCK_GEM_OBJECT_H__
+
+struct mock_object {
+	struct drm_i915_gem_object base;
+};
+
+#endif /* !__MOCK_GEM_OBJECT_H__ */
diff --git a/drivers/gpu/drm/i915/selftests/mock_gtt.c b/drivers/gpu/drm/i915/selftests/mock_gtt.c
new file mode 100644
index 000000000000..a61309c7cb3e
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_gtt.c
@@ -0,0 +1,138 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "mock_gtt.h"
+
+static void mock_insert_page(struct i915_address_space *vm,
+			     dma_addr_t addr,
+			     u64 offset,
+			     enum i915_cache_level level,
+			     u32 flags)
+{
+}
+
+static void mock_insert_entries(struct i915_address_space *vm,
+				struct sg_table *st,
+				u64 start,
+				enum i915_cache_level level, u32 flags)
+{
+}
+
+static int mock_bind_ppgtt(struct i915_vma *vma,
+			   enum i915_cache_level cache_level,
+			   u32 flags)
+{
+	GEM_BUG_ON(flags & I915_VMA_GLOBAL_BIND);
+	vma->pages = vma->obj->mm.pages;
+	vma->flags |= I915_VMA_LOCAL_BIND;
+	return 0;
+}
+
+static void mock_unbind_ppgtt(struct i915_vma *vma)
+{
+}
+
+static void mock_cleanup(struct i915_address_space *vm)
+{
+}
+
+struct i915_hw_ppgtt *
+mock_ppgtt(struct drm_i915_private *i915,
+	   const char *name)
+{
+	struct i915_hw_ppgtt *ppgtt;
+
+	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
+	if (!ppgtt)
+		return NULL;
+
+	kref_init(&ppgtt->ref);
+	ppgtt->base.i915 = i915;
+	ppgtt->base.total = round_down(U64_MAX, PAGE_SIZE);
+	ppgtt->base.file = ERR_PTR(-ENODEV);
+
+	INIT_LIST_HEAD(&ppgtt->base.active_list);
+	INIT_LIST_HEAD(&ppgtt->base.inactive_list);
+	INIT_LIST_HEAD(&ppgtt->base.unbound_list);
+
+	INIT_LIST_HEAD(&ppgtt->base.global_link);
+	drm_mm_init(&ppgtt->base.mm, 0, ppgtt->base.total);
+	i915_gem_timeline_init(i915, &ppgtt->base.timeline, name);
+
+	ppgtt->base.clear_range = nop_clear_range;
+	ppgtt->base.insert_page = mock_insert_page;
+	ppgtt->base.insert_entries = mock_insert_entries;
+	ppgtt->base.bind_vma = mock_bind_ppgtt;
+	ppgtt->base.unbind_vma = mock_unbind_ppgtt;
+	ppgtt->base.cleanup = mock_cleanup;
+
+	return ppgtt;
+}
+
+static int mock_bind_ggtt(struct i915_vma *vma,
+			  enum i915_cache_level cache_level,
+			  u32 flags)
+{
+	int err;
+
+	err = i915_get_ggtt_vma_pages(vma);
+	if (err)
+		return err;
+
+	vma->flags |= I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND;
+	return 0;
+}
+
+static void mock_unbind_ggtt(struct i915_vma *vma)
+{
+}
+
+void mock_init_ggtt(struct drm_i915_private *i915)
+{
+	struct i915_ggtt *ggtt = &i915->ggtt;
+
+	INIT_LIST_HEAD(&i915->vm_list);
+
+	ggtt->base.i915 = i915;
+
+	ggtt->mappable_base = 0;
+	ggtt->mappable_end = 2048 * PAGE_SIZE;
+	ggtt->base.total = 4096 * PAGE_SIZE;
+
+	ggtt->base.clear_range = nop_clear_range;
+	ggtt->base.insert_page = mock_insert_page;
+	ggtt->base.insert_entries = mock_insert_entries;
+	ggtt->base.bind_vma = mock_bind_ggtt;
+	ggtt->base.unbind_vma = mock_unbind_ggtt;
+	ggtt->base.cleanup = mock_cleanup;
+
+	i915_address_space_init(&ggtt->base, i915, "global");
+}
+
+void mock_fini_ggtt(struct drm_i915_private *i915)
+{
+	struct i915_ggtt *ggtt = &i915->ggtt;
+
+	i915_address_space_fini(&ggtt->base);
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_gtt.h b/drivers/gpu/drm/i915/selftests/mock_gtt.h
new file mode 100644
index 000000000000..9a0a833bb545
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_gtt.h
@@ -0,0 +1,35 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MOCK_GTT_H
+#define __MOCK_GTT_H
+
+void mock_init_ggtt(struct drm_i915_private *i915);
+void mock_fini_ggtt(struct drm_i915_private *i915);
+
+struct i915_hw_ppgtt *
+mock_ppgtt(struct drm_i915_private *i915,
+	   const char *name);
+
+#endif /* !__MOCK_GTT_H */
diff --git a/drivers/gpu/drm/i915/selftests/mock_request.c b/drivers/gpu/drm/i915/selftests/mock_request.c
new file mode 100644
index 000000000000..0e8d2e7f8c70
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_request.c
@@ -0,0 +1,63 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include "mock_engine.h"
+#include "mock_request.h"
+
+struct drm_i915_gem_request *
+mock_request(struct intel_engine_cs *engine,
+	     struct i915_gem_context *context,
+	     unsigned long delay)
+{
+	struct drm_i915_gem_request *request;
+	struct mock_request *mock;
+
+	/* NB the i915->requests slab cache is enlarged to fit mock_request */
+	request = i915_gem_request_alloc(engine, context);
+	if (!request)
+		return NULL;
+
+	mock = container_of(request, typeof(*mock), base);
+	mock->delay = delay;
+
+	return &mock->base;
+}
+
+bool mock_cancel_request(struct drm_i915_gem_request *request)
+{
+	struct mock_request *mock = container_of(request, typeof(*mock), base);
+	struct mock_engine *engine =
+		container_of(request->engine, typeof(*engine), base);
+	bool was_queued;
+
+	spin_lock_irq(&engine->hw_lock);
+	was_queued = !list_empty(&mock->link);
+	list_del_init(&mock->link);
+	spin_unlock_irq(&engine->hw_lock);
+
+	if (was_queued)
+		i915_gem_request_unsubmit(request);
+
+	return was_queued;
+}
diff --git a/drivers/gpu/drm/i915/selftests/mock_request.h b/drivers/gpu/drm/i915/selftests/mock_request.h
new file mode 100644
index 000000000000..4dea74c8e96d
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/mock_request.h
@@ -0,0 +1,46 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __MOCK_REQUEST__
+#define __MOCK_REQUEST__
+
+#include <linux/list.h>
+
+#include "../i915_gem_request.h"
+
+struct mock_request {
+	struct drm_i915_gem_request base;
+
+	struct list_head link;
+	unsigned long delay;
+};
+
+struct drm_i915_gem_request *
+mock_request(struct intel_engine_cs *engine,
+	     struct i915_gem_context *context,
+	     unsigned long delay);
+
+bool mock_cancel_request(struct drm_i915_gem_request *request);
+
+#endif /* !__MOCK_REQUEST__ */
diff --git a/drivers/gpu/drm/i915/selftests/scatterlist.c b/drivers/gpu/drm/i915/selftests/scatterlist.c
new file mode 100644
index 000000000000..eb2cda8e2b9f
--- /dev/null
+++ b/drivers/gpu/drm/i915/selftests/scatterlist.c
@@ -0,0 +1,355 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/prime_numbers.h>
+#include <linux/random.h>
+
+#include "../i915_selftest.h"
+
+#define PFN_BIAS (1 << 10)
+
+struct pfn_table {
+	struct sg_table st;
+	unsigned long start, end;
+};
+
+typedef unsigned int (*npages_fn_t)(unsigned long n,
+				    unsigned long count,
+				    struct rnd_state *rnd);
+
+static noinline int expect_pfn_sg(struct pfn_table *pt,
+				  npages_fn_t npages_fn,
+				  struct rnd_state *rnd,
+				  const char *who,
+				  unsigned long timeout)
+{
+	struct scatterlist *sg;
+	unsigned long pfn, n;
+
+	pfn = pt->start;
+	for_each_sg(pt->st.sgl, sg, pt->st.nents, n) {
+		struct page *page = sg_page(sg);
+		unsigned int npages = npages_fn(n, pt->st.nents, rnd);
+
+		if (page_to_pfn(page) != pfn) {
+			pr_err("%s: %s left pages out of order, expected pfn %lu, found pfn %lu (using for_each_sg)\n",
+			       __func__, who, pfn, page_to_pfn(page));
+			return -EINVAL;
+		}
+
+		if (sg->length != npages * PAGE_SIZE) {
+			pr_err("%s: %s copied wrong sg length, expected size %lu, found %u (using for_each_sg)\n",
+			       __func__, who, npages * PAGE_SIZE, sg->length);
+			return -EINVAL;
+		}
+
+		if (igt_timeout(timeout, "%s timed out\n", who))
+			return -EINTR;
+
+		pfn += npages;
+	}
+	if (pfn != pt->end) {
+		pr_err("%s: %s finished on wrong pfn, expected %lu, found %lu\n",
+		       __func__, who, pt->end, pfn);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static noinline int expect_pfn_sg_page_iter(struct pfn_table *pt,
+					    const char *who,
+					    unsigned long timeout)
+{
+	struct sg_page_iter sgiter;
+	unsigned long pfn;
+
+	pfn = pt->start;
+	for_each_sg_page(pt->st.sgl, &sgiter, pt->st.nents, 0) {
+		struct page *page = sg_page_iter_page(&sgiter);
+
+		if (page != pfn_to_page(pfn)) {
+			pr_err("%s: %s left pages out of order, expected pfn %lu, found pfn %lu (using for_each_sg_page)\n",
+			       __func__, who, pfn, page_to_pfn(page));
+			return -EINVAL;
+		}
+
+		if (igt_timeout(timeout, "%s timed out\n", who))
+			return -EINTR;
+
+		pfn++;
+	}
+	if (pfn != pt->end) {
+		pr_err("%s: %s finished on wrong pfn, expected %lu, found %lu\n",
+		       __func__, who, pt->end, pfn);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static noinline int expect_pfn_sgtiter(struct pfn_table *pt,
+				       const char *who,
+				       unsigned long timeout)
+{
+	struct sgt_iter sgt;
+	struct page *page;
+	unsigned long pfn;
+
+	pfn = pt->start;
+	for_each_sgt_page(page, sgt, &pt->st) {
+		if (page != pfn_to_page(pfn)) {
+			pr_err("%s: %s left pages out of order, expected pfn %lu, found pfn %lu (using for_each_sgt_page)\n",
+			       __func__, who, pfn, page_to_pfn(page));
+			return -EINVAL;
+		}
+
+		if (igt_timeout(timeout, "%s timed out\n", who))
+			return -EINTR;
+
+		pfn++;
+	}
+	if (pfn != pt->end) {
+		pr_err("%s: %s finished on wrong pfn, expected %lu, found %lu\n",
+		       __func__, who, pt->end, pfn);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int expect_pfn_sgtable(struct pfn_table *pt,
+			      npages_fn_t npages_fn,
+			      struct rnd_state *rnd,
+			      const char *who,
+			      unsigned long timeout)
+{
+	int err;
+
+	err = expect_pfn_sg(pt, npages_fn, rnd, who, timeout);
+	if (err)
+		return err;
+
+	err = expect_pfn_sg_page_iter(pt, who, timeout);
+	if (err)
+		return err;
+
+	err = expect_pfn_sgtiter(pt, who, timeout);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+static unsigned int one(unsigned long n,
+			unsigned long count,
+			struct rnd_state *rnd)
+{
+	return 1;
+}
+
+static unsigned int grow(unsigned long n,
+			 unsigned long count,
+			 struct rnd_state *rnd)
+{
+	return n + 1;
+}
+
+static unsigned int shrink(unsigned long n,
+			   unsigned long count,
+			   struct rnd_state *rnd)
+{
+	return count - n;
+}
+
+static unsigned int random(unsigned long n,
+			   unsigned long count,
+			   struct rnd_state *rnd)
+{
+	return 1 + (prandom_u32_state(rnd) % 1024);
+}
+
+static int alloc_table(struct pfn_table *pt,
+		       unsigned long count, unsigned long max,
+		       npages_fn_t npages_fn,
+		       struct rnd_state *rnd,
+		       int alloc_error)
+{
+	struct scatterlist *sg;
+	unsigned long n, pfn;
+
+	if (sg_alloc_table(&pt->st, max,
+			   GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN))
+		return alloc_error;
+
+	/* count should be less than 20 to prevent overflowing sg->length */
+	GEM_BUG_ON(overflows_type(count * PAGE_SIZE, sg->length));
+
+	/* Construct a table where each scatterlist contains different number
+	 * of entries. The idea is to check that we can iterate the individual
+	 * pages from inside the coalesced lists.
+	 */
+	pt->start = PFN_BIAS;
+	pfn = pt->start;
+	sg = pt->st.sgl;
+	for (n = 0; n < count; n++) {
+		unsigned long npages = npages_fn(n, count, rnd);
+
+		/* Nobody expects the Sparse Memmap! */
+		if (pfn_to_page(pfn + npages) != pfn_to_page(pfn) + npages) {
+			sg_free_table(&pt->st);
+			return -ENOSPC;
+		}
+
+		if (n)
+			sg = sg_next(sg);
+		sg_set_page(sg, pfn_to_page(pfn), npages * PAGE_SIZE, 0);
+
+		GEM_BUG_ON(page_to_pfn(sg_page(sg)) != pfn);
+		GEM_BUG_ON(sg->length != npages * PAGE_SIZE);
+		GEM_BUG_ON(sg->offset != 0);
+
+		pfn += npages;
+	}
+	sg_mark_end(sg);
+	pt->st.nents = n;
+	pt->end = pfn;
+
+	return 0;
+}
+
+static const npages_fn_t npages_funcs[] = {
+	one,
+	grow,
+	shrink,
+	random,
+	NULL,
+};
+
+static int igt_sg_alloc(void *ignored)
+{
+	IGT_TIMEOUT(end_time);
+	const unsigned long max_order = 20; /* approximating a 4GiB object */
+	struct rnd_state prng;
+	unsigned long prime;
+	int alloc_error = -ENOMEM;
+
+	for_each_prime_number(prime, max_order) {
+		unsigned long size = BIT(prime);
+		int offset;
+
+		for (offset = -1; offset <= 1; offset++) {
+			unsigned long sz = size + offset;
+			const npages_fn_t *npages;
+			struct pfn_table pt;
+			int err;
+
+			for (npages = npages_funcs; *npages; npages++) {
+				prandom_seed_state(&prng,
+						   i915_selftest.random_seed);
+				err = alloc_table(&pt, sz, sz, *npages, &prng,
+						  alloc_error);
+				if (err == -ENOSPC)
+					break;
+				if (err)
+					return err;
+
+				prandom_seed_state(&prng,
+						   i915_selftest.random_seed);
+				err = expect_pfn_sgtable(&pt, *npages, &prng,
+							 "sg_alloc_table",
+							 end_time);
+				sg_free_table(&pt.st);
+				if (err)
+					return err;
+			}
+		}
+
+		/* Test at least one continuation before accepting oom */
+		if (size > SG_MAX_SINGLE_ALLOC)
+			alloc_error = -ENOSPC;
+	}
+
+	return 0;
+}
+
+static int igt_sg_trim(void *ignored)
+{
+	IGT_TIMEOUT(end_time);
+	const unsigned long max = PAGE_SIZE; /* not prime! */
+	struct pfn_table pt;
+	unsigned long prime;
+	int alloc_error = -ENOMEM;
+
+	for_each_prime_number(prime, max) {
+		const npages_fn_t *npages;
+		int err;
+
+		for (npages = npages_funcs; *npages; npages++) {
+			struct rnd_state prng;
+
+			prandom_seed_state(&prng, i915_selftest.random_seed);
+			err = alloc_table(&pt, prime, max, *npages, &prng,
+					  alloc_error);
+			if (err == -ENOSPC)
+				break;
+			if (err)
+				return err;
+
+			if (i915_sg_trim(&pt.st)) {
+				if (pt.st.orig_nents != prime ||
+				    pt.st.nents != prime) {
+					pr_err("i915_sg_trim failed (nents %u, orig_nents %u), expected %lu\n",
+					       pt.st.nents, pt.st.orig_nents, prime);
+					err = -EINVAL;
+				} else {
+					prandom_seed_state(&prng,
+							   i915_selftest.random_seed);
+					err = expect_pfn_sgtable(&pt,
+								 *npages, &prng,
+								 "i915_sg_trim",
+								 end_time);
+				}
+			}
+			sg_free_table(&pt.st);
+			if (err)
+				return err;
+		}
+
+		/* Test at least one continuation before accepting oom */
+		if (prime > SG_MAX_SINGLE_ALLOC)
+			alloc_error = -ENOSPC;
+	}
+
+	return 0;
+}
+
+int scatterlist_mock_selftests(void)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(igt_sg_alloc),
+		SUBTEST(igt_sg_trim),
+	};
+
+	return i915_subtests(tests, NULL);
+}
author	Dave Airlie <airlied@redhat.com>	2017-03-08 12:41:47 +1000
committer	Dave Airlie <airlied@redhat.com>	2017-03-08 12:41:47 +1000
commit	2e16101780e9cc8c4c68566db002e7513a1530eb (patch)
tree	e3fa976f5fe68aea853c3653be13168bb6b43438 /drivers/gpu/drm
parent	b558dfd56a5c1f915327967ecfe1181cf2a7a494 (diff)
parent	505b681539a7e14aeb866515d3ef1a67375839bc (diff)
download	linux-2e16101780e9cc8c4c68566db002e7513a1530eb.tar.gz