drm/i915: Revert async unpin and nonblocking atomic commit

This reverts the following patches:

d55dbd06bb5e1399aba9ab5227465339d1bbefff drm/i915: Allow nonblocking update of pageflips.
15c86bdb760185e871c7a0f559978328aa500971 drm/i915: Check for unpin correctness.
95c2ccdc82d520f59ae3b6fdc097b63c9b7082bb Reapply "drm/i915: Avoid stalling on pending flips for legacy cursor updates"
a6747b7304a9d66758a196d885dab8bbfa5e7d1f drm/i915: Make unpin async.
03f476e1fcb42fca88fc50b94b0d3adbdbe887f0 drm/i915: Prepare connectors for nonblocking checks.
2099deffef4404f949ba1b68d2b17e0608190bc2 drm/i915: Pass atomic states to fbc update functions.
ee7171af72c39c18b7d7571419a4ac6ca30aea66 drm/i915: Remove reset_counter from intel_crtc.
2ee004f7c59b2e642f0bb2834f847d756f2dd7b7 drm/i915: Remove queue_flip pointer.
b8d2afae557dbb9b9c7bc6f6ec4f5278f3c4c34e drm/i915: Remove use_mmio_flip kernel parameter.
8dd634d922615ec3a9af7976029110ec037f8b50 drm/i915: Remove cs based page flip support.
143f73b3bf48c089b40f58462dd7f7c199fd4f0f drm/i915: Rework intel_crtc_page_flip to be almost atomic, v3.
84fc494b64e8c591be446a966b7447a9db519c88 drm/i915: Add the exclusive fence to plane_state.
6885843ae164e11f6c802209d06921e678a3f3f3 drm/i915: Convert flip_work to a list.
aa420ddd8eeaa5df579894a412289e4d07c2fee9 drm/i915: Allow mmio updates on all platforms, v2.
afee4d8707ab1f21b7668de995be3a5961e83582 Revert "drm/i915: Avoid stalling on pending flips for legacy cursor updates"

"drm/i915: Allow nonblocking update of pageflips" should have been
split up, misses a proper commit message and seems to cause issues in
the legacy page_flip path as demonstrated by kms_flip.

"drm/i915: Make unpin async" doesn't handle the unthrottled cursor
updates correctly, leading to an apparent pin count leak. This is
caught by the WARN_ON in i915_gem_object_do_pin which screams if we
have more than DRM_I915_GEM_OBJECT_MAX_PIN_COUNT pins.

Unfortuantely we can't just revert these two because this patch series
came with a built-in bisect breakage in the form of temporarily
removing the unthrottled cursor update hack for legacy cursor ioctl.
Therefore there's no other option than to revert the entire pile :(

There's one tiny conflict in intel_drv.h due to other patches, nothing
serious.

Normally I'd wait a bit longer with doing a maintainer revert, but
since the minimal set of patches we need to revert (due to the bisect
breakage) is so big, time is running out fast. And very soon
(especially after a few attempts at fixing issues) it'll be really
hard to revert things cleanly.

Lessons learned:
- Not a good idea to rush the review (done by someone fairly new to
  the area) and not make sure domain experts had a chance to read it.

- Patches should be properly split up. I only looked at the two
  patches that should be reverted in detail, but both look like the
  mix up different things in one patch.

- Patches really should have proper commit messages. Especially when
  doing more than one thing, and especially when touching critical and
  tricky core code.

- Building a patch series and r-b stamping it when it has a built-in
  bisect breakage is not a good idea.

- I also think we need to stop building up technical debt by
  postponing atomic igt testcases even longer. I think it's clear that
  there's enough corner cases in this beast that we really need to
  have the testcases _before_ the next step lands.

(cherry picked from commit 5a21b6650a239ebc020912968a44047701104159
from drm-intel-next-queeud)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Patrik Jakobsson <patrik.jakobsson@linux.intel.com>
Cc: John Harrison <John.C.Harrison@Intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Acked-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: Dave Airlie <airlied@redhat.com>
Acked-by: Jani Nikula <jani.nikula@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>

11 files changed, 1283 insertions, 625 deletions

diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 4c6b48dbd6e2..0a4bedb96d65 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -592,52 +592,6 @@ static int i915_gem_gtt_info(struct seq_file *m, void *data)
 	return 0;
 }
 
-static void i915_dump_pageflip(struct seq_file *m,
-			       struct drm_i915_private *dev_priv,
-			       struct intel_crtc *crtc,
-			       struct intel_flip_work *work)
-{
-	const char pipe = pipe_name(crtc->pipe);
-	u32 pending;
-	int i;
-
-	pending = atomic_read(&work->pending);
-	if (pending) {
-		seq_printf(m, "Flip ioctl preparing on pipe %c (plane %c)\n",
-			   pipe, plane_name(crtc->plane));
-	} else {
-		seq_printf(m, "Flip pending (waiting for vsync) on pipe %c (plane %c)\n",
-			   pipe, plane_name(crtc->plane));
-	}
-
-	for (i = 0; i < work->num_planes; i++) {
-		struct intel_plane_state *old_plane_state = work->old_plane_state[i];
-		struct drm_plane *plane = old_plane_state->base.plane;
-		struct drm_i915_gem_request *req = old_plane_state->wait_req;
-		struct intel_engine_cs *engine;
-
-		seq_printf(m, "[PLANE:%i] part of flip.\n", plane->base.id);
-
-		if (!req) {
-			seq_printf(m, "Plane not associated with any engine\n");
-			continue;
-		}
-
-		engine = i915_gem_request_get_engine(req);
-
-		seq_printf(m, "Plane blocked on %s at seqno %x, next seqno %x [current breadcrumb %x], completed? %d\n",
-			   engine->name,
-			   i915_gem_request_get_seqno(req),
-			   dev_priv->next_seqno,
-			   engine->get_seqno(engine),
-			   i915_gem_request_completed(req, true));
-	}
-
-	seq_printf(m, "Flip queued on frame %d, now %d\n",
-		   pending ? work->flip_queued_vblank : -1,
-		   intel_crtc_get_vblank_counter(crtc));
-}
-
 static int i915_gem_pageflip_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = m->private;
@@ -656,13 +610,48 @@ static int i915_gem_pageflip_info(struct seq_file *m, void *data)
 		struct intel_flip_work *work;
 
 		spin_lock_irq(&dev->event_lock);
-		if (list_empty(&crtc->flip_work)) {
+		work = crtc->flip_work;
+		if (work == NULL) {
 			seq_printf(m, "No flip due on pipe %c (plane %c)\n",
 				   pipe, plane);
 		} else {
-			list_for_each_entry(work, &crtc->flip_work, head) {
-				i915_dump_pageflip(m, dev_priv, crtc, work);
-				seq_puts(m, "\n");
+			u32 pending;
+			u32 addr;
+
+			pending = atomic_read(&work->pending);
+			if (pending) {
+				seq_printf(m, "Flip ioctl preparing on pipe %c (plane %c)\n",
+					   pipe, plane);
+			} else {
+				seq_printf(m, "Flip pending (waiting for vsync) on pipe %c (plane %c)\n",
+					   pipe, plane);
+			}
+			if (work->flip_queued_req) {
+				struct intel_engine_cs *engine = i915_gem_request_get_engine(work->flip_queued_req);
+
+				seq_printf(m, "Flip queued on %s at seqno %x, next seqno %x [current breadcrumb %x], completed? %d\n",
+					   engine->name,
+					   i915_gem_request_get_seqno(work->flip_queued_req),
+					   dev_priv->next_seqno,
+					   engine->get_seqno(engine),
+					   i915_gem_request_completed(work->flip_queued_req, true));
+			} else
+				seq_printf(m, "Flip not associated with any ring\n");
+			seq_printf(m, "Flip queued on frame %d, (was ready on frame %d), now %d\n",
+				   work->flip_queued_vblank,
+				   work->flip_ready_vblank,
+				   intel_crtc_get_vblank_counter(crtc));
+			seq_printf(m, "%d prepares\n", atomic_read(&work->pending));
+
+			if (INTEL_INFO(dev)->gen >= 4)
+				addr = I915_HI_DISPBASE(I915_READ(DSPSURF(crtc->plane)));
+			else
+				addr = I915_READ(DSPADDR(crtc->plane));
+			seq_printf(m, "Current scanout address 0x%08x\n", addr);
+
+			if (work->pending_flip_obj) {
+				seq_printf(m, "New framebuffer address 0x%08lx\n", (long)work->gtt_offset);
+				seq_printf(m, "MMIO update completed? %d\n",  addr == work->gtt_offset);
 			}
 		}
 		spin_unlock_irq(&dev->event_lock);
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 255d4c3e3eea..33a8d4e3ec21 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -618,6 +618,11 @@ struct drm_i915_display_funcs {
 	void (*audio_codec_disable)(struct intel_encoder *encoder);
 	void (*fdi_link_train)(struct drm_crtc *crtc);
 	void (*init_clock_gating)(struct drm_device *dev);
+	int (*queue_flip)(struct drm_device *dev, struct drm_crtc *crtc,
+			  struct drm_framebuffer *fb,
+			  struct drm_i915_gem_object *obj,
+			  struct drm_i915_gem_request *req,
+			  uint32_t flags);
 	void (*hpd_irq_setup)(struct drm_i915_private *dev_priv);
 	/* clock updates for mode set */
 	/* cursor updates */
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index 5513c4cb02e0..3242a37fb304 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -136,12 +136,6 @@ static const u32 hpd_bxt[HPD_NUM_PINS] = {
 	POSTING_READ(type##IIR); \
 } while (0)
 
-static void
-intel_finish_page_flip_cs(struct drm_i915_private *dev_priv, unsigned pipe)
-{
-	DRM_DEBUG_KMS("Finished page flip\n");
-}
-
 /*
  * We should clear IMR at preinstall/uninstall, and just check at postinstall.
  */
@@ -1637,11 +1631,16 @@ static void gen6_rps_irq_handler(struct drm_i915_private *dev_priv, u32 pm_iir)
 	}
 }
 
-static void intel_pipe_handle_vblank(struct drm_i915_private *dev_priv,
+static bool intel_pipe_handle_vblank(struct drm_i915_private *dev_priv,
 				     enum pipe pipe)
 {
-	if (drm_handle_vblank(dev_priv->dev, pipe))
+	bool ret;
+
+	ret = drm_handle_vblank(dev_priv->dev, pipe);
+	if (ret)
 		intel_finish_page_flip_mmio(dev_priv, pipe);
+
+	return ret;
 }
 
 static void valleyview_pipestat_irq_ack(struct drm_i915_private *dev_priv,
@@ -1708,8 +1707,9 @@ static void valleyview_pipestat_irq_handler(struct drm_i915_private *dev_priv,
 	enum pipe pipe;
 
 	for_each_pipe(dev_priv, pipe) {
-		if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS)
-			intel_pipe_handle_vblank(dev_priv, pipe);
+		if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS &&
+		    intel_pipe_handle_vblank(dev_priv, pipe))
+			intel_check_page_flip(dev_priv, pipe);
 
 		if (pipe_stats[pipe] & PLANE_FLIP_DONE_INT_STATUS_VLV)
 			intel_finish_page_flip_cs(dev_priv, pipe);
@@ -2155,8 +2155,9 @@ static void ilk_display_irq_handler(struct drm_i915_private *dev_priv,
 		DRM_ERROR("Poison interrupt\n");
 
 	for_each_pipe(dev_priv, pipe) {
-		if (de_iir & DE_PIPE_VBLANK(pipe))
-			intel_pipe_handle_vblank(dev_priv, pipe);
+		if (de_iir & DE_PIPE_VBLANK(pipe) &&
+		    intel_pipe_handle_vblank(dev_priv, pipe))
+			intel_check_page_flip(dev_priv, pipe);
 
 		if (de_iir & DE_PIPE_FIFO_UNDERRUN(pipe))
 			intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe);
@@ -2205,8 +2206,9 @@ static void ivb_display_irq_handler(struct drm_i915_private *dev_priv,
 		intel_opregion_asle_intr(dev_priv);
 
 	for_each_pipe(dev_priv, pipe) {
-		if (de_iir & (DE_PIPE_VBLANK_IVB(pipe)))
-			intel_pipe_handle_vblank(dev_priv, pipe);
+		if (de_iir & (DE_PIPE_VBLANK_IVB(pipe)) &&
+		    intel_pipe_handle_vblank(dev_priv, pipe))
+			intel_check_page_flip(dev_priv, pipe);
 
 		/* plane/pipes map 1:1 on ilk+ */
 		if (de_iir & DE_PLANE_FLIP_DONE_IVB(pipe))
@@ -2405,8 +2407,9 @@ gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl)
 		ret = IRQ_HANDLED;
 		I915_WRITE(GEN8_DE_PIPE_IIR(pipe), iir);
 
-		if (iir & GEN8_PIPE_VBLANK)
-			intel_pipe_handle_vblank(dev_priv, pipe);
+		if (iir & GEN8_PIPE_VBLANK &&
+		    intel_pipe_handle_vblank(dev_priv, pipe))
+			intel_check_page_flip(dev_priv, pipe);
 
 		flip_done = iir;
 		if (INTEL_INFO(dev_priv)->gen >= 9)
@@ -3970,6 +3973,37 @@ static int i8xx_irq_postinstall(struct drm_device *dev)
 	return 0;
 }
 
+/*
+ * Returns true when a page flip has completed.
+ */
+static bool i8xx_handle_vblank(struct drm_i915_private *dev_priv,
+			       int plane, int pipe, u32 iir)
+{
+	u16 flip_pending = DISPLAY_PLANE_FLIP_PENDING(plane);
+
+	if (!intel_pipe_handle_vblank(dev_priv, pipe))
+		return false;
+
+	if ((iir & flip_pending) == 0)
+		goto check_page_flip;
+
+	/* We detect FlipDone by looking for the change in PendingFlip from '1'
+	 * to '0' on the following vblank, i.e. IIR has the Pendingflip
+	 * asserted following the MI_DISPLAY_FLIP, but ISR is deasserted, hence
+	 * the flip is completed (no longer pending). Since this doesn't raise
+	 * an interrupt per se, we watch for the change at vblank.
+	 */
+	if (I915_READ16(ISR) & flip_pending)
+		goto check_page_flip;
+
+	intel_finish_page_flip_cs(dev_priv, pipe);
+	return true;
+
+check_page_flip:
+	intel_check_page_flip(dev_priv, pipe);
+	return false;
+}
+
 static irqreturn_t i8xx_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = arg;
@@ -4022,8 +4056,13 @@ static irqreturn_t i8xx_irq_handler(int irq, void *arg)
 			notify_ring(&dev_priv->engine[RCS]);
 
 		for_each_pipe(dev_priv, pipe) {
-			if (pipe_stats[pipe] & PIPE_VBLANK_INTERRUPT_STATUS)
-				intel_pipe_handle_vblank(dev_priv, pipe);
+			int plane = pipe;
+			if (HAS_FBC(dev_priv))
+				plane = !plane;
+
+			if (pipe_stats[pipe] & PIPE_VBLANK_INTERRUPT_STATUS &&
+			    i8xx_handle_vblank(dev_priv, plane, pipe, iir))
+				flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(plane);
 
 			if (pipe_stats[pipe] & PIPE_CRC_DONE_INTERRUPT_STATUS)
 				i9xx_pipe_crc_irq_handler(dev_priv, pipe);
@@ -4123,6 +4162,37 @@ static int i915_irq_postinstall(struct drm_device *dev)
 	return 0;
 }
 
+/*
+ * Returns true when a page flip has completed.
+ */
+static bool i915_handle_vblank(struct drm_i915_private *dev_priv,
+			       int plane, int pipe, u32 iir)
+{
+	u32 flip_pending = DISPLAY_PLANE_FLIP_PENDING(plane);
+
+	if (!intel_pipe_handle_vblank(dev_priv, pipe))
+		return false;
+
+	if ((iir & flip_pending) == 0)
+		goto check_page_flip;
+
+	/* We detect FlipDone by looking for the change in PendingFlip from '1'
+	 * to '0' on the following vblank, i.e. IIR has the Pendingflip
+	 * asserted following the MI_DISPLAY_FLIP, but ISR is deasserted, hence
+	 * the flip is completed (no longer pending). Since this doesn't raise
+	 * an interrupt per se, we watch for the change at vblank.
+	 */
+	if (I915_READ(ISR) & flip_pending)
+		goto check_page_flip;
+
+	intel_finish_page_flip_cs(dev_priv, pipe);
+	return true;
+
+check_page_flip:
+	intel_check_page_flip(dev_priv, pipe);
+	return false;
+}
+
 static irqreturn_t i915_irq_handler(int irq, void *arg)
 {
 	struct drm_device *dev = arg;
@@ -4183,8 +4253,13 @@ static irqreturn_t i915_irq_handler(int irq, void *arg)
 			notify_ring(&dev_priv->engine[RCS]);
 
 		for_each_pipe(dev_priv, pipe) {
-			if (pipe_stats[pipe] & PIPE_VBLANK_INTERRUPT_STATUS)
-				intel_pipe_handle_vblank(dev_priv, pipe);
+			int plane = pipe;
+			if (HAS_FBC(dev_priv))
+				plane = !plane;
+
+			if (pipe_stats[pipe] & PIPE_VBLANK_INTERRUPT_STATUS &&
+			    i915_handle_vblank(dev_priv, plane, pipe, iir))
+				flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(plane);
 
 			if (pipe_stats[pipe] & PIPE_LEGACY_BLC_EVENT_STATUS)
 				blc_event = true;
@@ -4412,8 +4487,9 @@ static irqreturn_t i965_irq_handler(int irq, void *arg)
 			notify_ring(&dev_priv->engine[VCS]);
 
 		for_each_pipe(dev_priv, pipe) {
-			if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS)
-				intel_pipe_handle_vblank(dev_priv, pipe);
+			if (pipe_stats[pipe] & PIPE_START_VBLANK_INTERRUPT_STATUS &&
+			    i915_handle_vblank(dev_priv, pipe, pipe, iir))
+				flip_mask &= ~DISPLAY_PLANE_FLIP_PENDING(pipe);
 
 			if (pipe_stats[pipe] & PIPE_LEGACY_BLC_EVENT_STATUS)
 				blc_event = true;
diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c
index cd74fb8e9387..383c076919ed 100644
--- a/drivers/gpu/drm/i915/i915_params.c
+++ b/drivers/gpu/drm/i915/i915_params.c
@@ -49,6 +49,7 @@ struct i915_params i915 __read_mostly = {
 	.invert_brightness = 0,
 	.disable_display = 0,
 	.enable_cmd_parser = 1,
+	.use_mmio_flip = 0,
 	.mmio_debug = 0,
 	.verbose_state_checks = 1,
 	.nuclear_pageflip = 0,
@@ -173,6 +174,10 @@ module_param_named_unsafe(enable_cmd_parser, i915.enable_cmd_parser, int, 0600);
 MODULE_PARM_DESC(enable_cmd_parser,
 		 "Enable command parsing (1=enabled [default], 0=disabled)");
 
+module_param_named_unsafe(use_mmio_flip, i915.use_mmio_flip, int, 0600);
+MODULE_PARM_DESC(use_mmio_flip,
+		 "use MMIO flips (-1=never, 0=driver discretion [default], 1=always)");
+
 module_param_named(mmio_debug, i915.mmio_debug, int, 0600);
 MODULE_PARM_DESC(mmio_debug,
 	"Enable the MMIO debug code for the first N failures (default: off). "
diff --git a/drivers/gpu/drm/i915/i915_params.h b/drivers/gpu/drm/i915/i915_params.h
index dd0d0bbcc05b..65e73dd7d970 100644
--- a/drivers/gpu/drm/i915/i915_params.h
+++ b/drivers/gpu/drm/i915/i915_params.h
@@ -46,6 +46,7 @@ struct i915_params {
 	int invert_brightness;
 	int enable_cmd_parser;
 	int guc_log_level;
+	int use_mmio_flip;
 	int mmio_debug;
 	int edp_vswing;
 	unsigned int inject_load_failure;
diff --git a/drivers/gpu/drm/i915/intel_atomic.c b/drivers/gpu/drm/i915/intel_atomic.c
index b4927f6bbeac..50ff90aea721 100644
--- a/drivers/gpu/drm/i915/intel_atomic.c
+++ b/drivers/gpu/drm/i915/intel_atomic.c
@@ -311,17 +311,6 @@ intel_atomic_state_alloc(struct drm_device *dev)
 void intel_atomic_state_clear(struct drm_atomic_state *s)
 {
 	struct intel_atomic_state *state = to_intel_atomic_state(s);
-	int i;
-
-	for (i = 0; i < ARRAY_SIZE(state->work); i++) {
-		struct intel_flip_work *work = state->work[i];
-
-		if (work)
-			intel_free_flip_work(work);
-
-		state->work[i] = NULL;
-	}
-
 	drm_atomic_state_default_clear(&state->base);
 	state->dpll_set = state->modeset = false;
 }
diff --git a/drivers/gpu/drm/i915/intel_atomic_plane.c b/drivers/gpu/drm/i915/intel_atomic_plane.c
index 2ab45f16fa65..7de7721f65bc 100644
--- a/drivers/gpu/drm/i915/intel_atomic_plane.c
+++ b/drivers/gpu/drm/i915/intel_atomic_plane.c
@@ -102,7 +102,6 @@ intel_plane_destroy_state(struct drm_plane *plane,
 			  struct drm_plane_state *state)
 {
 	WARN_ON(state && to_intel_plane_state(state)->wait_req);
-	WARN_ON(state && state->fence);
 	drm_atomic_helper_plane_destroy_state(plane, state);
 }
 
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 60ffbfd53ffa..ae253bb57c25 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -48,6 +48,11 @@
 #include <linux/reservation.h>
 #include <linux/dma-buf.h>
 
+static bool is_mmio_work(struct intel_flip_work *work)
+{
+	return work->mmio_work.func;
+}
+
 /* Primary plane formats for gen <= 3 */
 static const uint32_t i8xx_primary_formats[] = {
 	DRM_FORMAT_C8,
@@ -108,6 +113,8 @@ static void vlv_prepare_pll(struct intel_crtc *crtc,
 			    const struct intel_crtc_state *pipe_config);
 static void chv_prepare_pll(struct intel_crtc *crtc,
 			    const struct intel_crtc_state *pipe_config);
+static void intel_begin_crtc_commit(struct drm_crtc *, struct drm_crtc_state *);
+static void intel_finish_crtc_commit(struct drm_crtc *, struct drm_crtc_state *);
 static void skl_init_scalers(struct drm_device *dev, struct intel_crtc *intel_crtc,
 	struct intel_crtc_state *crtc_state);
 static void skylake_pfit_enable(struct intel_crtc *crtc);
@@ -116,9 +123,6 @@ 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 void intel_modeset_verify_crtc(struct drm_crtc *crtc,
-				      struct drm_crtc_state *old_state,
-				      struct drm_crtc_state *new_state);
 
 struct intel_limit {
 	struct {
@@ -2523,6 +2527,20 @@ out_unref_obj:
 	return false;
 }
 
+/* Update plane->state->fb to match plane->fb after driver-internal updates */
+static void
+update_state_fb(struct drm_plane *plane)
+{
+	if (plane->fb == plane->state->fb)
+		return;
+
+	if (plane->state->fb)
+		drm_framebuffer_unreference(plane->state->fb);
+	plane->state->fb = plane->fb;
+	if (plane->state->fb)
+		drm_framebuffer_reference(plane->state->fb);
+}
+
 static void
 intel_find_initial_plane_obj(struct intel_crtc *intel_crtc,
 			     struct intel_initial_plane_config *plane_config)
@@ -3095,6 +3113,14 @@ intel_pipe_set_base_atomic(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	return -ENODEV;
 }
 
+static void intel_complete_page_flips(struct drm_i915_private *dev_priv)
+{
+	struct intel_crtc *crtc;
+
+	for_each_intel_crtc(dev_priv->dev, crtc)
+		intel_finish_page_flip_cs(dev_priv, crtc->pipe);
+}
+
 static void intel_update_primary_planes(struct drm_device *dev)
 {
 	struct drm_crtc *crtc;
@@ -3135,6 +3161,13 @@ void intel_prepare_reset(struct drm_i915_private *dev_priv)
 
 void intel_finish_reset(struct drm_i915_private *dev_priv)
 {
+	/*
+	 * Flips in the rings will be nuked by the reset,
+	 * so complete all pending flips so that user space
+	 * will get its events and not get stuck.
+	 */
+	intel_complete_page_flips(dev_priv);
+
 	/* no reset support for gen2 */
 	if (IS_GEN2(dev_priv))
 		return;
@@ -3177,7 +3210,20 @@ void intel_finish_reset(struct drm_i915_private *dev_priv)
 
 static bool intel_crtc_has_pending_flip(struct drm_crtc *crtc)
 {
-	return !list_empty_careful(&to_intel_crtc(crtc)->flip_work);
+	struct drm_device *dev = crtc->dev;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	unsigned reset_counter;
+	bool pending;
+
+	reset_counter = i915_reset_counter(&to_i915(dev)->gpu_error);
+	if (intel_crtc->reset_counter != reset_counter)
+		return false;
+
+	spin_lock_irq(&dev->event_lock);
+	pending = to_intel_crtc(crtc)->flip_work != NULL;
+	spin_unlock_irq(&dev->event_lock);
+
+	return pending;
 }
 
 static void intel_update_pipe_config(struct intel_crtc *crtc,
@@ -3753,7 +3799,7 @@ bool intel_has_pending_fb_unpin(struct drm_device *dev)
 		if (atomic_read(&crtc->unpin_work_count) == 0)
 			continue;
 
-		if (!list_empty_careful(&crtc->flip_work))
+		if (crtc->flip_work)
 			intel_wait_for_vblank(dev, crtc->pipe);
 
 		return true;
@@ -3762,30 +3808,23 @@ bool intel_has_pending_fb_unpin(struct drm_device *dev)
 	return false;
 }
 
-static void page_flip_completed(struct intel_crtc *intel_crtc, struct intel_flip_work *work)
+static void page_flip_completed(struct intel_crtc *intel_crtc)
 {
 	struct drm_i915_private *dev_priv = to_i915(intel_crtc->base.dev);
-	struct drm_plane_state *new_plane_state;
-	struct drm_plane *primary = intel_crtc->base.primary;
+	struct intel_flip_work *work = intel_crtc->flip_work;
+
+	intel_crtc->flip_work = NULL;
 
 	if (work->event)
 		drm_crtc_send_vblank_event(&intel_crtc->base, work->event);
 
 	drm_crtc_vblank_put(&intel_crtc->base);
 
-	new_plane_state = &work->old_plane_state[0]->base;
-	if (work->num_planes >= 1 &&
-	    new_plane_state->plane == primary &&
-	    new_plane_state->fb)
-		trace_i915_flip_complete(intel_crtc->plane,
-					 intel_fb_obj(new_plane_state->fb));
-
-	if (work->can_async_unpin) {
-		list_del_init(&work->head);
-		wake_up_all(&dev_priv->pending_flip_queue);
-	}
-
+	wake_up_all(&dev_priv->pending_flip_queue);
 	queue_work(dev_priv->wq, &work->unpin_work);
+
+	trace_i915_flip_complete(intel_crtc->plane,
+				 work->pending_flip_obj);
 }
 
 static int intel_crtc_wait_for_pending_flips(struct drm_crtc *crtc)
@@ -3804,7 +3843,18 @@ static int intel_crtc_wait_for_pending_flips(struct drm_crtc *crtc)
 	if (ret < 0)
 		return ret;
 
-	WARN(ret == 0, "Stuck page flip\n");
+	if (ret == 0) {
+		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+		struct intel_flip_work *work;
+
+		spin_lock_irq(&dev->event_lock);
+		work = intel_crtc->flip_work;
+		if (work && !is_mmio_work(work)) {
+			WARN_ONCE(1, "Removing stuck page flip\n");
+			page_flip_completed(intel_crtc);
+		}
+		spin_unlock_irq(&dev->event_lock);
+	}
 
 	return 0;
 }
@@ -4535,6 +4585,39 @@ intel_pre_disable_primary_noatomic(struct drm_crtc *crtc)
 	}
 }
 
+static void intel_post_plane_update(struct intel_crtc_state *old_crtc_state)
+{
+	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->base.crtc);
+	struct drm_atomic_state *old_state = old_crtc_state->base.state;
+	struct intel_crtc_state *pipe_config =
+		to_intel_crtc_state(crtc->base.state);
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_plane *primary = crtc->base.primary;
+	struct drm_plane_state *old_pri_state =
+		drm_atomic_get_existing_plane_state(old_state, primary);
+
+	intel_frontbuffer_flip(dev, pipe_config->fb_bits);
+
+	crtc->wm.cxsr_allowed = true;
+
+	if (pipe_config->update_wm_post && pipe_config->base.active)
+		intel_update_watermarks(&crtc->base);
+
+	if (old_pri_state) {
+		struct intel_plane_state *primary_state =
+			to_intel_plane_state(primary->state);
+		struct intel_plane_state *old_primary_state =
+			to_intel_plane_state(old_pri_state);
+
+		intel_fbc_post_update(crtc);
+
+		if (primary_state->visible &&
+		    (needs_modeset(&pipe_config->base) ||
+		     !old_primary_state->visible))
+			intel_post_enable_primary(&crtc->base);
+	}
+}
+
 static void intel_pre_plane_update(struct intel_crtc_state *old_crtc_state)
 {
 	struct intel_crtc *crtc = to_intel_crtc(old_crtc_state->base.crtc);
@@ -4554,7 +4637,7 @@ static void intel_pre_plane_update(struct intel_crtc_state *old_crtc_state)
 		struct intel_plane_state *old_primary_state =
 			to_intel_plane_state(old_pri_state);
 
-		intel_fbc_pre_update(crtc, pipe_config, primary_state);
+		intel_fbc_pre_update(crtc);
 
 		if (old_primary_state->visible &&
 		    (modeset || !primary_state->visible))
@@ -5144,21 +5227,18 @@ modeset_get_crtc_power_domains(struct drm_crtc *crtc,
 	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	enum intel_display_power_domain domain;
-	unsigned long domains, new_domains, old_domains, ms_domain = 0;
+	unsigned long domains, new_domains, old_domains;
 
 	old_domains = intel_crtc->enabled_power_domains;
 	intel_crtc->enabled_power_domains = new_domains =
 		get_crtc_power_domains(crtc, crtc_state);
 
-	if (needs_modeset(&crtc_state->base))
-		ms_domain = BIT(POWER_DOMAIN_MODESET);
-
-	domains = (new_domains & ~old_domains) | ms_domain;
+	domains = new_domains & ~old_domains;
 
 	for_each_power_domain(domain, domains)
 		intel_display_power_get(dev_priv, domain);
 
-	return (old_domains & ~new_domains) | ms_domain;
+	return old_domains & ~new_domains;
 }
 
 static void modeset_put_power_domains(struct drm_i915_private *dev_priv,
@@ -6152,7 +6232,7 @@ static void intel_crtc_disable_noatomic(struct drm_crtc *crtc)
 		return;
 
 	if (to_intel_plane_state(crtc->primary->state)->visible) {
-		WARN_ON(list_empty(&intel_crtc->flip_work));
+		WARN_ON(intel_crtc->flip_work);
 
 		intel_pre_disable_primary_noatomic(crtc);
 
@@ -6204,12 +6284,6 @@ int intel_display_suspend(struct drm_device *dev)
 		DRM_ERROR("Suspending crtc's failed with %i\n", ret);
 	else
 		dev_priv->modeset_restore_state = state;
-
-	/*
-	 * Make sure all unpin_work completes before returning.
-	 */
-	flush_workqueue(dev_priv->wq);
-
 	return ret;
 }
 
@@ -6223,10 +6297,9 @@ void intel_encoder_destroy(struct drm_encoder *encoder)
 
 /* Cross check the actual hw state with our own modeset state tracking (and it's
  * internal consistency). */
-static void intel_connector_verify_state(struct intel_connector *connector,
-					 struct drm_connector_state *conn_state)
+static void intel_connector_verify_state(struct intel_connector *connector)
 {
-	struct drm_crtc *crtc = conn_state->crtc;
+	struct drm_crtc *crtc = connector->base.state->crtc;
 
 	DRM_DEBUG_KMS("[CONNECTOR:%d:%s]\n",
 		      connector->base.base.id,
@@ -6234,6 +6307,7 @@ static void intel_connector_verify_state(struct intel_connector *connector,
 
 	if (connector->get_hw_state(connector)) {
 		struct intel_encoder *encoder = connector->encoder;
+		struct drm_connector_state *conn_state = connector->base.state;
 
 		I915_STATE_WARN(!crtc,
 			 "connector enabled without attached crtc\n");
@@ -6255,7 +6329,7 @@ static void intel_connector_verify_state(struct intel_connector *connector,
 	} else {
 		I915_STATE_WARN(crtc && crtc->state->active,
 			"attached crtc is active, but connector isn't\n");
-		I915_STATE_WARN(!crtc && conn_state->best_encoder,
+		I915_STATE_WARN(!crtc && connector->base.state->best_encoder,
 			"best encoder set without crtc!\n");
 	}
 }
@@ -10750,13 +10824,6 @@ void intel_mark_idle(struct drm_i915_private *dev_priv)
 	intel_runtime_pm_put(dev_priv);
 }
 
-void intel_free_flip_work(struct intel_flip_work *work)
-{
-	kfree(work->old_connector_state);
-	kfree(work->new_connector_state);
-	kfree(work);
-}
-
 static void intel_crtc_destroy(struct drm_crtc *crtc)
 {
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
@@ -10764,278 +10831,885 @@ static void intel_crtc_destroy(struct drm_crtc *crtc)
 	struct intel_flip_work *work;
 
 	spin_lock_irq(&dev->event_lock);
-	while (!list_empty(&intel_crtc->flip_work)) {
-		work = list_first_entry(&intel_crtc->flip_work,
-					struct intel_flip_work, head);
-		list_del_init(&work->head);
-		spin_unlock_irq(&dev->event_lock);
+	work = intel_crtc->flip_work;
+	intel_crtc->flip_work = NULL;
+	spin_unlock_irq(&dev->event_lock);
 
+	if (work) {
 		cancel_work_sync(&work->mmio_work);
 		cancel_work_sync(&work->unpin_work);
-		intel_free_flip_work(work);
-
-		spin_lock_irq(&dev->event_lock);
+		kfree(work);
 	}
-	spin_unlock_irq(&dev->event_lock);
 
 	drm_crtc_cleanup(crtc);
 
 	kfree(intel_crtc);
 }
 
-static void intel_crtc_post_flip_update(struct intel_flip_work *work,
-					struct drm_crtc *crtc)
+static void intel_unpin_work_fn(struct work_struct *__work)
+{
+	struct intel_flip_work *work =
+		container_of(__work, struct intel_flip_work, unpin_work);
+	struct intel_crtc *crtc = to_intel_crtc(work->crtc);
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_plane *primary = crtc->base.primary;
+
+	if (is_mmio_work(work))
+		flush_work(&work->mmio_work);
+
+	mutex_lock(&dev->struct_mutex);
+	intel_unpin_fb_obj(work->old_fb, primary->state->rotation);
+	drm_gem_object_unreference(&work->pending_flip_obj->base);
+
+	if (work->flip_queued_req)
+		i915_gem_request_assign(&work->flip_queued_req, NULL);
+	mutex_unlock(&dev->struct_mutex);
+
+	intel_frontbuffer_flip_complete(dev, to_intel_plane(primary)->frontbuffer_bit);
+	intel_fbc_post_update(crtc);
+	drm_framebuffer_unreference(work->old_fb);
+
+	BUG_ON(atomic_read(&crtc->unpin_work_count) == 0);
+	atomic_dec(&crtc->unpin_work_count);
+
+	kfree(work);
+}
+
+/* Is 'a' after or equal to 'b'? */
+static bool g4x_flip_count_after_eq(u32 a, u32 b)
+{
+	return !((a - b) & 0x80000000);
+}
+
+static bool __pageflip_finished_cs(struct intel_crtc *crtc,
+				   struct intel_flip_work *work)
+{
+	struct drm_device *dev = crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned reset_counter;
+
+	reset_counter = i915_reset_counter(&dev_priv->gpu_error);
+	if (crtc->reset_counter != reset_counter)
+		return true;
+
+	/*
+	 * The relevant registers doen't exist on pre-ctg.
+	 * As the flip done interrupt doesn't trigger for mmio
+	 * flips on gmch platforms, a flip count check isn't
+	 * really needed there. But since ctg has the registers,
+	 * include it in the check anyway.
+	 */
+	if (INTEL_INFO(dev)->gen < 5 && !IS_G4X(dev))
+		return true;
+
+	/*
+	 * BDW signals flip done immediately if the plane
+	 * is disabled, even if the plane enable is already
+	 * armed to occur at the next vblank :(
+	 */
+
+	/*
+	 * A DSPSURFLIVE check isn't enough in case the mmio and CS flips
+	 * used the same base address. In that case the mmio flip might
+	 * have completed, but the CS hasn't even executed the flip yet.
+	 *
+	 * A flip count check isn't enough as the CS might have updated
+	 * the base address just after start of vblank, but before we
+	 * managed to process the interrupt. This means we'd complete the
+	 * CS flip too soon.
+	 *
+	 * Combining both checks should get us a good enough result. It may
+	 * still happen that the CS flip has been executed, but has not
+	 * yet actually completed. But in case the base address is the same
+	 * anyway, we don't really care.
+	 */
+	return (I915_READ(DSPSURFLIVE(crtc->plane)) & ~0xfff) ==
+		crtc->flip_work->gtt_offset &&
+		g4x_flip_count_after_eq(I915_READ(PIPE_FLIPCOUNT_G4X(crtc->pipe)),
+				    crtc->flip_work->flip_count);
+}
+
+static bool
+__pageflip_finished_mmio(struct intel_crtc *crtc,
+			       struct intel_flip_work *work)
+{
+	/*
+	 * MMIO work completes when vblank is different from
+	 * flip_queued_vblank.
+	 *
+	 * Reset counter value doesn't matter, this is handled by
+	 * i915_wait_request finishing early, so no need to handle
+	 * reset here.
+	 */
+	return intel_crtc_get_vblank_counter(crtc) != work->flip_queued_vblank;
+}
+
+
+static bool pageflip_finished(struct intel_crtc *crtc,
+			      struct intel_flip_work *work)
 {
-	struct intel_crtc_state *crtc_state = work->new_crtc_state;
+	if (!atomic_read(&work->pending))
+		return false;
+
+	smp_rmb();
+
+	if (is_mmio_work(work))
+		return __pageflip_finished_mmio(crtc, work);
+	else
+		return __pageflip_finished_cs(crtc, work);
+}
+
+void intel_finish_page_flip_cs(struct drm_i915_private *dev_priv, int pipe)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_flip_work *work;
+	unsigned long flags;
 
-	if (crtc_state->disable_cxsr)
-		intel_crtc->wm.cxsr_allowed = true;
+	/* Ignore early vblank irqs */
+	if (!crtc)
+		return;
 
-	if (crtc_state->update_wm_post && crtc_state->base.active)
-		intel_update_watermarks(crtc);
+	/*
+	 * This is called both by irq handlers and the reset code (to complete
+	 * lost pageflips) so needs the full irqsave spinlocks.
+	 */
+	spin_lock_irqsave(&dev->event_lock, flags);
+	work = intel_crtc->flip_work;
 
-	if (work->num_planes > 0 &&
-	    work->old_plane_state[0]->base.plane == crtc->primary) {
-		struct intel_plane_state *plane_state =
-			work->new_plane_state[0];
+	if (work != NULL &&
+	    !is_mmio_work(work) &&
+	    pageflip_finished(intel_crtc, work))
+		page_flip_completed(intel_crtc);
 
-		if (plane_state->visible &&
-		    (needs_modeset(&crtc_state->base) ||
-		     !work->old_plane_state[0]->visible))
-			intel_post_enable_primary(crtc);
-	}
+	spin_unlock_irqrestore(&dev->event_lock, flags);
 }
 
-static void intel_unpin_work_fn(struct work_struct *__work)
+void intel_finish_page_flip_mmio(struct drm_i915_private *dev_priv, int pipe)
 {
-	struct intel_flip_work *work =
-		container_of(__work, struct intel_flip_work, unpin_work);
-	struct drm_crtc *crtc = work->old_crtc_state->base.crtc;
+	struct drm_device *dev = dev_priv->dev;
+	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int i;
+	struct intel_flip_work *work;
+	unsigned long flags;
 
-	if (work->fb_bits)
-		intel_frontbuffer_flip_complete(dev, work->fb_bits);
+	/* Ignore early vblank irqs */
+	if (!crtc)
+		return;
 
 	/*
-	 * Unless work->can_async_unpin is false, there's no way to ensure
-	 * that work->new_crtc_state contains valid memory during unpin
-	 * because intel_atomic_commit may free it before this runs.
+	 * This is called both by irq handlers and the reset code (to complete
+	 * lost pageflips) so needs the full irqsave spinlocks.
 	 */
-	if (!work->can_async_unpin) {
-		intel_crtc_post_flip_update(work, crtc);
+	spin_lock_irqsave(&dev->event_lock, flags);
+	work = intel_crtc->flip_work;
 
-		if (dev_priv->display.optimize_watermarks)
-			dev_priv->display.optimize_watermarks(work->new_crtc_state);
-	}
+	if (work != NULL &&
+	    is_mmio_work(work) &&
+	    pageflip_finished(intel_crtc, work))
+		page_flip_completed(intel_crtc);
 
-	if (work->fb_bits & to_intel_plane(crtc->primary)->frontbuffer_bit)
-		intel_fbc_post_update(intel_crtc);
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+}
 
-	if (work->put_power_domains)
-		modeset_put_power_domains(dev_priv, work->put_power_domains);
+static inline void intel_mark_page_flip_active(struct intel_crtc *crtc,
+					       struct intel_flip_work *work)
+{
+	work->flip_queued_vblank = intel_crtc_get_vblank_counter(crtc);
 
-	/* Make sure mmio work is completely finished before freeing all state here. */
-	flush_work(&work->mmio_work);
+	/* Ensure that the work item is consistent when activating it ... */
+	smp_mb__before_atomic();
+	atomic_set(&work->pending, 1);
+}
 
-	if (!work->can_async_unpin &&
-	    (work->new_crtc_state->update_pipe ||
-	     needs_modeset(&work->new_crtc_state->base))) {
-		/* This must be called before work is unpinned for serialization. */
-		intel_modeset_verify_crtc(crtc, &work->old_crtc_state->base,
-					  &work->new_crtc_state->base);
+static int intel_gen2_queue_flip(struct drm_device *dev,
+				 struct drm_crtc *crtc,
+				 struct drm_framebuffer *fb,
+				 struct drm_i915_gem_object *obj,
+				 struct drm_i915_gem_request *req,
+				 uint32_t flags)
+{
+	struct intel_engine_cs *engine = req->engine;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	u32 flip_mask;
+	int ret;
 
-		for (i = 0; i < work->num_new_connectors; i++) {
-			struct drm_connector_state *conn_state =
-				work->new_connector_state[i];
-			struct drm_connector *con = conn_state->connector;
+	ret = intel_ring_begin(req, 6);
+	if (ret)
+		return ret;
 
-			WARN_ON(!con);
+	/* Can't queue multiple flips, so wait for the previous
+	 * one to finish before executing the next.
+	 */
+	if (intel_crtc->plane)
+		flip_mask = MI_WAIT_FOR_PLANE_B_FLIP;
+	else
+		flip_mask = MI_WAIT_FOR_PLANE_A_FLIP;
+	intel_ring_emit(engine, MI_WAIT_FOR_EVENT | flip_mask);
+	intel_ring_emit(engine, MI_NOOP);
+	intel_ring_emit(engine, MI_DISPLAY_FLIP |
+			MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
+	intel_ring_emit(engine, fb->pitches[0]);
+	intel_ring_emit(engine, intel_crtc->flip_work->gtt_offset);
+	intel_ring_emit(engine, 0); /* aux display base address, unused */
 
-			intel_connector_verify_state(to_intel_connector(con),
-						     conn_state);
-		}
-	}
+	return 0;
+}
+
+static int intel_gen3_queue_flip(struct drm_device *dev,
+				 struct drm_crtc *crtc,
+				 struct drm_framebuffer *fb,
+				 struct drm_i915_gem_object *obj,
+				 struct drm_i915_gem_request *req,
+				 uint32_t flags)
+{
+	struct intel_engine_cs *engine = req->engine;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	u32 flip_mask;
+	int ret;
+
+	ret = intel_ring_begin(req, 6);
+	if (ret)
+		return ret;
 
-	for (i = 0; i < work->num_old_connectors; i++) {
-		struct drm_connector_state *old_con_state =
-			work->old_connector_state[i];
-		struct drm_connector *con =
-			old_con_state->connector;
+	if (intel_crtc->plane)
+		flip_mask = MI_WAIT_FOR_PLANE_B_FLIP;
+	else
+		flip_mask = MI_WAIT_FOR_PLANE_A_FLIP;
+	intel_ring_emit(engine, MI_WAIT_FOR_EVENT | flip_mask);
+	intel_ring_emit(engine, MI_NOOP);
+	intel_ring_emit(engine, MI_DISPLAY_FLIP_I915 |
+			MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
+	intel_ring_emit(engine, fb->pitches[0]);
+	intel_ring_emit(engine, intel_crtc->flip_work->gtt_offset);
+	intel_ring_emit(engine, MI_NOOP);
+
+	return 0;
+}
+
+static int intel_gen4_queue_flip(struct drm_device *dev,
+				 struct drm_crtc *crtc,
+				 struct drm_framebuffer *fb,
+				 struct drm_i915_gem_object *obj,
+				 struct drm_i915_gem_request *req,
+				 uint32_t flags)
+{
+	struct intel_engine_cs *engine = req->engine;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	uint32_t pf, pipesrc;
+	int ret;
+
+	ret = intel_ring_begin(req, 4);
+	if (ret)
+		return ret;
+
+	/* 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(engine, MI_DISPLAY_FLIP |
+			MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
+	intel_ring_emit(engine, fb->pitches[0]);
+	intel_ring_emit(engine, intel_crtc->flip_work->gtt_offset |
+			obj->tiling_mode);
+
+	/* XXX Enabling the panel-fitter across page-flip is so far
+	 * untested on non-native modes, so ignore it for now.
+	 * pf = I915_READ(pipe == 0 ? PFA_CTL_1 : PFB_CTL_1) & PF_ENABLE;
+	 */
+	pf = 0;
+	pipesrc = I915_READ(PIPESRC(intel_crtc->pipe)) & 0x0fff0fff;
+	intel_ring_emit(engine, pf | pipesrc);
+
+	return 0;
+}
+
+static int intel_gen6_queue_flip(struct drm_device *dev,
+				 struct drm_crtc *crtc,
+				 struct drm_framebuffer *fb,
+				 struct drm_i915_gem_object *obj,
+				 struct drm_i915_gem_request *req,
+				 uint32_t flags)
+{
+	struct intel_engine_cs *engine = req->engine;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	uint32_t pf, pipesrc;
+	int ret;
+
+	ret = intel_ring_begin(req, 4);
+	if (ret)
+		return ret;
+
+	intel_ring_emit(engine, MI_DISPLAY_FLIP |
+			MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
+	intel_ring_emit(engine, fb->pitches[0] | obj->tiling_mode);
+	intel_ring_emit(engine, intel_crtc->flip_work->gtt_offset);
+
+	/* Contrary to the suggestions in the documentation,
+	 * "Enable Panel Fitter" does not seem to be required when page
+	 * flipping with a non-native mode, and worse causes a normal
+	 * modeset to fail.
+	 * pf = I915_READ(PF_CTL(intel_crtc->pipe)) & PF_ENABLE;
+	 */
+	pf = 0;
+	pipesrc = I915_READ(PIPESRC(intel_crtc->pipe)) & 0x0fff0fff;
+	intel_ring_emit(engine, pf | pipesrc);
+
+	return 0;
+}
+
+static int intel_gen7_queue_flip(struct drm_device *dev,
+				 struct drm_crtc *crtc,
+				 struct drm_framebuffer *fb,
+				 struct drm_i915_gem_object *obj,
+				 struct drm_i915_gem_request *req,
+				 uint32_t flags)
+{
+	struct intel_engine_cs *engine = req->engine;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	uint32_t plane_bit = 0;
+	int len, ret;
+
+	switch (intel_crtc->plane) {
+	case PLANE_A:
+		plane_bit = MI_DISPLAY_FLIP_IVB_PLANE_A;
+		break;
+	case PLANE_B:
+		plane_bit = MI_DISPLAY_FLIP_IVB_PLANE_B;
+		break;
+	case PLANE_C:
+		plane_bit = MI_DISPLAY_FLIP_IVB_PLANE_C;
+		break;
+	default:
+		WARN_ONCE(1, "unknown plane in flip command\n");
+		return -ENODEV;
+	}
 
-		con->funcs->atomic_destroy_state(con, old_con_state);
+	len = 4;
+	if (engine->id == RCS) {
+		len += 6;
+		/*
+		 * On Gen 8, SRM is now taking an extra dword to accommodate
+		 * 48bits addresses, and we need a NOOP for the batch size to
+		 * stay even.
+		 */
+		if (IS_GEN8(dev))
+			len += 2;
 	}
 
-	if (!work->can_async_unpin || !list_empty(&work->head)) {
-		spin_lock_irq(&dev->event_lock);
-		WARN(list_empty(&work->head) != work->can_async_unpin,
-		     "[CRTC:%i] Pin work %p async %i with %i planes, active %i -> %i ms %i\n",
-		     crtc->base.id, work, work->can_async_unpin, work->num_planes,
-		     work->old_crtc_state->base.active, work->new_crtc_state->base.active,
-		     needs_modeset(&work->new_crtc_state->base));
+	/*
+	 * BSpec MI_DISPLAY_FLIP for IVB:
+	 * "The full packet must be contained within the same cache line."
+	 *
+	 * Currently the LRI+SRM+MI_DISPLAY_FLIP all fit within the same
+	 * cacheline, if we ever start emitting more commands before
+	 * the MI_DISPLAY_FLIP we may need to first emit everything else,
+	 * then do the cacheline alignment, and finally emit the
+	 * MI_DISPLAY_FLIP.
+	 */
+	ret = intel_ring_cacheline_align(req);
+	if (ret)
+		return ret;
 
-		if (!list_empty(&work->head))
-			list_del(&work->head);
+	ret = intel_ring_begin(req, len);
+	if (ret)
+		return ret;
 
-		wake_up_all(&dev_priv->pending_flip_queue);
-		spin_unlock_irq(&dev->event_lock);
+	/* 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
+	 * more than one flip event at any time (or ensure that one flip message
+	 * can be sent by waiting for flip-done prior to queueing new flips).
+	 * Experimentation says that BCS works despite DERRMR masking all
+	 * flip-done completion events and that unmasking all planes at once
+	 * for the RCS also doesn't appear to drop events. Setting the DERRMR
+	 * to zero does lead to lockups within MI_DISPLAY_FLIP.
+	 */
+	if (engine->id == RCS) {
+		intel_ring_emit(engine, MI_LOAD_REGISTER_IMM(1));
+		intel_ring_emit_reg(engine, DERRMR);
+		intel_ring_emit(engine, ~(DERRMR_PIPEA_PRI_FLIP_DONE |
+					  DERRMR_PIPEB_PRI_FLIP_DONE |
+					  DERRMR_PIPEC_PRI_FLIP_DONE));
+		if (IS_GEN8(dev))
+			intel_ring_emit(engine, MI_STORE_REGISTER_MEM_GEN8 |
+					      MI_SRM_LRM_GLOBAL_GTT);
+		else
+			intel_ring_emit(engine, MI_STORE_REGISTER_MEM |
+					      MI_SRM_LRM_GLOBAL_GTT);
+		intel_ring_emit_reg(engine, DERRMR);
+		intel_ring_emit(engine, engine->scratch.gtt_offset + 256);
+		if (IS_GEN8(dev)) {
+			intel_ring_emit(engine, 0);
+			intel_ring_emit(engine, MI_NOOP);
+		}
 	}
 
-	/* New crtc_state freed? */
-	if (work->free_new_crtc_state)
-		intel_crtc_destroy_state(crtc, &work->new_crtc_state->base);
+	intel_ring_emit(engine, MI_DISPLAY_FLIP_I915 | plane_bit);
+	intel_ring_emit(engine, (fb->pitches[0] | obj->tiling_mode));
+	intel_ring_emit(engine, intel_crtc->flip_work->gtt_offset);
+	intel_ring_emit(engine, (MI_NOOP));
 
-	intel_crtc_destroy_state(crtc, &work->old_crtc_state->base);
+	return 0;
+}
 
-	for (i = 0; i < work->num_planes; i++) {
-		struct intel_plane_state *old_plane_state =
-			work->old_plane_state[i];
-		struct drm_framebuffer *old_fb = old_plane_state->base.fb;
-		struct drm_plane *plane = old_plane_state->base.plane;
-		struct drm_i915_gem_request *req;
+static bool use_mmio_flip(struct intel_engine_cs *engine,
+			  struct drm_i915_gem_object *obj)
+{
+	/*
+	 * This is not being used for older platforms, because
+	 * non-availability of flip done interrupt forces us to use
+	 * CS flips. Older platforms derive flip done using some clever
+	 * tricks involving the flip_pending status bits and vblank irqs.
+	 * So using MMIO flips there would disrupt this mechanism.
+	 */
 
-		req = old_plane_state->wait_req;
-		old_plane_state->wait_req = NULL;
-		if (req)
-			i915_gem_request_unreference(req);
+	if (engine == NULL)
+		return true;
 
-		fence_put(old_plane_state->base.fence);
-		old_plane_state->base.fence = NULL;
+	if (INTEL_GEN(engine->i915) < 5)
+		return false;
 
-		if (old_fb &&
-		    (plane->type != DRM_PLANE_TYPE_CURSOR ||
-		     !INTEL_INFO(dev_priv)->cursor_needs_physical)) {
-			mutex_lock(&dev->struct_mutex);
-			intel_unpin_fb_obj(old_fb, old_plane_state->base.rotation);
-			mutex_unlock(&dev->struct_mutex);
-		}
+	if (i915.use_mmio_flip < 0)
+		return false;
+	else if (i915.use_mmio_flip > 0)
+		return true;
+	else if (i915.enable_execlists)
+		return true;
+	else if (obj->base.dma_buf &&
+		 !reservation_object_test_signaled_rcu(obj->base.dma_buf->resv,
+						       false))
+		return true;
+	else
+		return engine != i915_gem_request_get_engine(obj->last_write_req);
+}
+
+static void skl_do_mmio_flip(struct intel_crtc *intel_crtc,
+			     unsigned int rotation,
+			     struct intel_flip_work *work)
+{
+	struct drm_device *dev = intel_crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_framebuffer *fb = intel_crtc->base.primary->fb;
+	const enum pipe pipe = intel_crtc->pipe;
+	u32 ctl, stride, tile_height;
+
+	ctl = I915_READ(PLANE_CTL(pipe, 0));
+	ctl &= ~PLANE_CTL_TILED_MASK;
+	switch (fb->modifier[0]) {
+	case DRM_FORMAT_MOD_NONE:
+		break;
+	case I915_FORMAT_MOD_X_TILED:
+		ctl |= PLANE_CTL_TILED_X;
+		break;
+	case I915_FORMAT_MOD_Y_TILED:
+		ctl |= PLANE_CTL_TILED_Y;
+		break;
+	case I915_FORMAT_MOD_Yf_TILED:
+		ctl |= PLANE_CTL_TILED_YF;
+		break;
+	default:
+		MISSING_CASE(fb->modifier[0]);
+	}
 
-		intel_plane_destroy_state(plane, &old_plane_state->base);
+	/*
+	 * The stride is either expressed as a multiple of 64 bytes chunks for
+	 * linear buffers or in number of tiles for tiled buffers.
+	 */
+	if (intel_rotation_90_or_270(rotation)) {
+		/* stride = Surface height in tiles */
+		tile_height = intel_tile_height(dev_priv, fb->modifier[0], 0);
+		stride = DIV_ROUND_UP(fb->height, tile_height);
+	} else {
+		stride = fb->pitches[0] /
+			intel_fb_stride_alignment(dev_priv, fb->modifier[0],
+						  fb->pixel_format);
 	}
 
-	if (!WARN_ON(atomic_read(&intel_crtc->unpin_work_count) == 0))
-		atomic_dec(&intel_crtc->unpin_work_count);
+	/*
+	 * Both PLANE_CTL and PLANE_STRIDE are not updated on vblank but on
+	 * PLANE_SURF updates, the update is then guaranteed to be atomic.
+	 */
+	I915_WRITE(PLANE_CTL(pipe, 0), ctl);
+	I915_WRITE(PLANE_STRIDE(pipe, 0), stride);
 
-	intel_free_flip_work(work);
+	I915_WRITE(PLANE_SURF(pipe, 0), work->gtt_offset);
+	POSTING_READ(PLANE_SURF(pipe, 0));
 }
 
+static void ilk_do_mmio_flip(struct intel_crtc *intel_crtc,
+			     struct intel_flip_work *work)
+{
+	struct drm_device *dev = intel_crtc->base.dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct intel_framebuffer *intel_fb =
+		to_intel_framebuffer(intel_crtc->base.primary->fb);
+	struct drm_i915_gem_object *obj = intel_fb->obj;
+	i915_reg_t reg = DSPCNTR(intel_crtc->plane);
+	u32 dspcntr;
 
-static bool pageflip_finished(struct intel_crtc *crtc,
-			      struct intel_flip_work *work)
+	dspcntr = I915_READ(reg);
+
+	if (obj->tiling_mode != I915_TILING_NONE)
+		dspcntr |= DISPPLANE_TILED;
+	else
+		dspcntr &= ~DISPPLANE_TILED;
+
+	I915_WRITE(reg, dspcntr);
+
+	I915_WRITE(DSPSURF(intel_crtc->plane), work->gtt_offset);
+	POSTING_READ(DSPSURF(intel_crtc->plane));
+}
+
+static void intel_mmio_flip_work_func(struct work_struct *w)
+{
+	struct intel_flip_work *work =
+		container_of(w, struct intel_flip_work, mmio_work);
+	struct intel_crtc *crtc = to_intel_crtc(work->crtc);
+	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct intel_framebuffer *intel_fb =
+		to_intel_framebuffer(crtc->base.primary->fb);
+	struct drm_i915_gem_object *obj = intel_fb->obj;
+
+	if (work->flip_queued_req)
+		WARN_ON(__i915_wait_request(work->flip_queued_req,
+					    false, NULL,
+					    &dev_priv->rps.mmioflips));
+
+	/* For framebuffer backed by dmabuf, wait for fence */
+	if (obj->base.dma_buf)
+		WARN_ON(reservation_object_wait_timeout_rcu(obj->base.dma_buf->resv,
+							    false, false,
+							    MAX_SCHEDULE_TIMEOUT) < 0);
+
+	intel_pipe_update_start(crtc);
+
+	if (INTEL_GEN(dev_priv) >= 9)
+		skl_do_mmio_flip(crtc, work->rotation, work);
+	else
+		/* use_mmio_flip() retricts MMIO flips to ilk+ */
+		ilk_do_mmio_flip(crtc, work);
+
+	intel_pipe_update_end(crtc, work);
+}
+
+static int intel_default_queue_flip(struct drm_device *dev,
+				    struct drm_crtc *crtc,
+				    struct drm_framebuffer *fb,
+				    struct drm_i915_gem_object *obj,
+				    struct drm_i915_gem_request *req,
+				    uint32_t flags)
 {
+	return -ENODEV;
+}
+
+static bool __pageflip_stall_check_cs(struct drm_i915_private *dev_priv,
+				      struct intel_crtc *intel_crtc,
+				      struct intel_flip_work *work)
+{
+	u32 addr, vblank;
+
 	if (!atomic_read(&work->pending))
 		return false;
 
 	smp_rmb();
 
-	/*
-	 * MMIO work completes when vblank is different from
-	 * flip_queued_vblank.
+	vblank = intel_crtc_get_vblank_counter(intel_crtc);
+	if (work->flip_ready_vblank == 0) {
+		if (work->flip_queued_req &&
+		    !i915_gem_request_completed(work->flip_queued_req, true))
+			return false;
+
+		work->flip_ready_vblank = vblank;
+	}
+
+	if (vblank - work->flip_ready_vblank < 3)
+		return false;
+
+	/* Potential stall - if we see that the flip has happened,
+	 * assume a missed interrupt. */
+	if (INTEL_GEN(dev_priv) >= 4)
+		addr = I915_HI_DISPBASE(I915_READ(DSPSURF(intel_crtc->plane)));
+	else
+		addr = I915_READ(DSPADDR(intel_crtc->plane));
+
+	/* There is a potential issue here with a false positive after a flip
+	 * to the same address. We could address this by checking for a
+	 * non-incrementing frame counter.
 	 */
-	return intel_crtc_get_vblank_counter(crtc) != work->flip_queued_vblank;
+	return addr == work->gtt_offset;
 }
 
-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)
 {
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct intel_flip_work *work;
-	unsigned long flags;
 
-	/* Ignore early vblank irqs */
-	if (!crtc)
-		return;
+	WARN_ON(!in_interrupt());
 
-	/*
-	 * This is called both by irq handlers and the reset code (to complete
-	 * lost pageflips) so needs the full irqsave spinlocks.
-	 */
-	spin_lock_irqsave(&dev->event_lock, flags);
-	while (!list_empty(&intel_crtc->flip_work)) {
-		work = list_first_entry(&intel_crtc->flip_work,
-					struct intel_flip_work,
-					head);
+	if (crtc == NULL)
+		return;
 
-		if (!pageflip_finished(intel_crtc, work) ||
-		    work_busy(&work->unpin_work))
-			break;
+	spin_lock(&dev->event_lock);
+	work = intel_crtc->flip_work;
 
-		page_flip_completed(intel_crtc, work);
+	if (work != NULL && !is_mmio_work(work) &&
+	    __pageflip_stall_check_cs(dev_priv, intel_crtc, work)) {
+		WARN_ONCE(1,
+			  "Kicking stuck page flip: queued at %d, now %d\n",
+			work->flip_queued_vblank, intel_crtc_get_vblank_counter(intel_crtc));
+		page_flip_completed(intel_crtc);
+		work = NULL;
 	}
-	spin_unlock_irqrestore(&dev->event_lock, flags);
+
+	if (work != NULL && !is_mmio_work(work) &&
+	    intel_crtc_get_vblank_counter(intel_crtc) - work->flip_queued_vblank > 1)
+		intel_queue_rps_boost_for_request(work->flip_queued_req);
+	spin_unlock(&dev->event_lock);
 }
 
-static void intel_mmio_flip_work_func(struct work_struct *w)
+static int intel_crtc_page_flip(struct drm_crtc *crtc,
+				struct drm_framebuffer *fb,
+				struct drm_pending_vblank_event *event,
+				uint32_t page_flip_flags)
 {
-	struct intel_flip_work *work =
-		container_of(w, struct intel_flip_work, mmio_work);
-	struct drm_crtc *crtc = work->old_crtc_state->base.crtc;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct intel_crtc_state *crtc_state = work->new_crtc_state;
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_gem_request *req;
-	int i, ret;
+	struct drm_framebuffer *old_fb = crtc->primary->fb;
+	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct drm_plane *primary = crtc->primary;
+	enum pipe pipe = intel_crtc->pipe;
+	struct intel_flip_work *work;
+	struct intel_engine_cs *engine;
+	bool mmio_flip;
+	struct drm_i915_gem_request *request = NULL;
+	int ret;
+
+	/*
+	 * drm_mode_page_flip_ioctl() should already catch this, but double
+	 * check to be safe.  In the future we may enable pageflipping from
+	 * a disabled primary plane.
+	 */
+	if (WARN_ON(intel_fb_obj(old_fb) == NULL))
+		return -EBUSY;
 
-	if (!needs_modeset(&crtc_state->base) && crtc_state->update_pipe) {
-		work->put_power_domains =
-			modeset_get_crtc_power_domains(crtc, crtc_state);
+	/* Can't change pixel format via MI display flips. */
+	if (fb->pixel_format != crtc->primary->fb->pixel_format)
+		return -EINVAL;
+
+	/*
+	 * TILEOFF/LINOFF registers can't be changed via MI display flips.
+	 * Note that pitch changes could also affect these register.
+	 */
+	if (INTEL_INFO(dev)->gen > 3 &&
+	    (fb->offsets[0] != crtc->primary->fb->offsets[0] ||
+	     fb->pitches[0] != crtc->primary->fb->pitches[0]))
+		return -EINVAL;
+
+	if (i915_terminally_wedged(&dev_priv->gpu_error))
+		goto out_hang;
+
+	work = kzalloc(sizeof(*work), GFP_KERNEL);
+	if (work == NULL)
+		return -ENOMEM;
+
+	work->event = event;
+	work->crtc = crtc;
+	work->old_fb = old_fb;
+	INIT_WORK(&work->unpin_work, intel_unpin_work_fn);
+
+	ret = drm_crtc_vblank_get(crtc);
+	if (ret)
+		goto free_work;
+
+	/* We borrow the event spin lock for protecting flip_work */
+	spin_lock_irq(&dev->event_lock);
+	if (intel_crtc->flip_work) {
+		/* Before declaring the flip queue wedged, check if
+		 * the hardware completed the operation behind our backs.
+		 */
+		if (pageflip_finished(intel_crtc, intel_crtc->flip_work)) {
+			DRM_DEBUG_DRIVER("flip queue: previous flip completed, continuing\n");
+			page_flip_completed(intel_crtc);
+		} else {
+			DRM_DEBUG_DRIVER("flip queue: crtc already busy\n");
+			spin_unlock_irq(&dev->event_lock);
+
+			drm_crtc_vblank_put(crtc);
+			kfree(work);
+			return -EBUSY;
+		}
 	}
+	intel_crtc->flip_work = work;
+	spin_unlock_irq(&dev->event_lock);
 
-	for (i = 0; i < work->num_planes; i++) {
-		struct intel_plane_state *old_plane_state = work->old_plane_state[i];
+	if (atomic_read(&intel_crtc->unpin_work_count) >= 2)
+		flush_workqueue(dev_priv->wq);
 
-		/* For framebuffer backed by dmabuf, wait for fence */
-		if (old_plane_state->base.fence)
-			WARN_ON(fence_wait(old_plane_state->base.fence, false) < 0);
+	/* Reference the objects for the scheduled work. */
+	drm_framebuffer_reference(work->old_fb);
+	drm_gem_object_reference(&obj->base);
 
-		req = old_plane_state->wait_req;
-		if (!req)
-			continue;
+	crtc->primary->fb = fb;
+	update_state_fb(crtc->primary);
+	intel_fbc_pre_update(intel_crtc);
 
-		WARN_ON(__i915_wait_request(req, false, NULL,
-					    &dev_priv->rps.mmioflips));
+	work->pending_flip_obj = obj;
+
+	ret = i915_mutex_lock_interruptible(dev);
+	if (ret)
+		goto cleanup;
+
+	intel_crtc->reset_counter = i915_reset_counter(&dev_priv->gpu_error);
+	if (__i915_reset_in_progress_or_wedged(intel_crtc->reset_counter)) {
+		ret = -EIO;
+		goto cleanup;
 	}
 
-	ret = drm_crtc_vblank_get(crtc);
-	I915_STATE_WARN(ret < 0, "enabling vblank failed with %i\n", ret);
+	atomic_inc(&intel_crtc->unpin_work_count);
+
+	if (INTEL_INFO(dev)->gen >= 5 || IS_G4X(dev))
+		work->flip_count = I915_READ(PIPE_FLIPCOUNT_G4X(pipe)) + 1;
 
-	if (work->num_planes &&
-	    work->old_plane_state[0]->base.plane == crtc->primary)
-		intel_fbc_enable(intel_crtc, work->new_crtc_state, work->new_plane_state[0]);
+	if (IS_VALLEYVIEW(dev) || IS_CHERRYVIEW(dev)) {
+		engine = &dev_priv->engine[BCS];
+		if (obj->tiling_mode != intel_fb_obj(work->old_fb)->tiling_mode)
+			/* vlv: DISPLAY_FLIP fails to change tiling */
+			engine = NULL;
+	} else if (IS_IVYBRIDGE(dev) || IS_HASWELL(dev)) {
+		engine = &dev_priv->engine[BCS];
+	} else if (INTEL_INFO(dev)->gen >= 7) {
+		engine = i915_gem_request_get_engine(obj->last_write_req);
+		if (engine == NULL || engine->id != RCS)
+			engine = &dev_priv->engine[BCS];
+	} else {
+		engine = &dev_priv->engine[RCS];
+	}
 
-	intel_frontbuffer_flip_prepare(dev, work->fb_bits);
+	mmio_flip = use_mmio_flip(engine, obj);
 
-	intel_pipe_update_start(intel_crtc);
-	if (!needs_modeset(&crtc_state->base)) {
-		if (crtc_state->base.color_mgmt_changed || crtc_state->update_pipe) {
-			intel_color_set_csc(&crtc_state->base);
-			intel_color_load_luts(&crtc_state->base);
+	/* When using CS flips, we want to emit semaphores between rings.
+	 * However, when using mmio flips we will create a task to do the
+	 * synchronisation, so all we want here is to pin the framebuffer
+	 * into the display plane and skip any waits.
+	 */
+	if (!mmio_flip) {
+		ret = i915_gem_object_sync(obj, engine, &request);
+		if (!ret && !request) {
+			request = i915_gem_request_alloc(engine, NULL);
+			ret = PTR_ERR_OR_ZERO(request);
 		}
 
-		if (crtc_state->update_pipe)
-			intel_update_pipe_config(intel_crtc, work->old_crtc_state);
-		else if (INTEL_INFO(dev)->gen >= 9)
-			skl_detach_scalers(intel_crtc);
+		if (ret)
+			goto cleanup_pending;
 	}
 
-	for (i = 0; i < work->num_planes; i++) {
-		struct intel_plane_state *new_plane_state = work->new_plane_state[i];
-		struct intel_plane *plane = to_intel_plane(new_plane_state->base.plane);
+	ret = intel_pin_and_fence_fb_obj(fb, primary->state->rotation);
+	if (ret)
+		goto cleanup_pending;
 
-		if (new_plane_state->visible)
-			plane->update_plane(&plane->base, crtc_state, new_plane_state);
-		else
-			plane->disable_plane(&plane->base, crtc);
+	work->gtt_offset = intel_plane_obj_offset(to_intel_plane(primary),
+						  obj, 0);
+	work->gtt_offset += intel_crtc->dspaddr_offset;
+	work->rotation = crtc->primary->state->rotation;
+
+	if (mmio_flip) {
+		INIT_WORK(&work->mmio_work, intel_mmio_flip_work_func);
+
+		i915_gem_request_assign(&work->flip_queued_req,
+					obj->last_write_req);
+
+		schedule_work(&work->mmio_work);
+	} else {
+		i915_gem_request_assign(&work->flip_queued_req, request);
+		ret = dev_priv->display.queue_flip(dev, crtc, fb, obj, request,
+						   page_flip_flags);
+		if (ret)
+			goto cleanup_unpin;
+
+		intel_mark_page_flip_active(intel_crtc, work);
+
+		i915_add_request_no_flush(request);
 	}
 
-	intel_pipe_update_end(intel_crtc, work);
+	i915_gem_track_fb(intel_fb_obj(old_fb), obj,
+			  to_intel_plane(primary)->frontbuffer_bit);
+	mutex_unlock(&dev->struct_mutex);
+
+	intel_frontbuffer_flip_prepare(dev,
+				       to_intel_plane(primary)->frontbuffer_bit);
+
+	trace_i915_flip_request(intel_crtc->plane, obj);
+
+	return 0;
+
+cleanup_unpin:
+	intel_unpin_fb_obj(fb, crtc->primary->state->rotation);
+cleanup_pending:
+	if (!IS_ERR_OR_NULL(request))
+		i915_add_request_no_flush(request);
+	atomic_dec(&intel_crtc->unpin_work_count);
+	mutex_unlock(&dev->struct_mutex);
+cleanup:
+	crtc->primary->fb = old_fb;
+	update_state_fb(crtc->primary);
+
+	drm_gem_object_unreference_unlocked(&obj->base);
+	drm_framebuffer_unreference(work->old_fb);
+
+	spin_lock_irq(&dev->event_lock);
+	intel_crtc->flip_work = NULL;
+	spin_unlock_irq(&dev->event_lock);
+
+	drm_crtc_vblank_put(crtc);
+free_work:
+	kfree(work);
+
+	if (ret == -EIO) {
+		struct drm_atomic_state *state;
+		struct drm_plane_state *plane_state;
+
+out_hang:
+		state = drm_atomic_state_alloc(dev);
+		if (!state)
+			return -ENOMEM;
+		state->acquire_ctx = drm_modeset_legacy_acquire_ctx(crtc);
+
+retry:
+		plane_state = drm_atomic_get_plane_state(state, primary);
+		ret = PTR_ERR_OR_ZERO(plane_state);
+		if (!ret) {
+			drm_atomic_set_fb_for_plane(plane_state, fb);
+
+			ret = drm_atomic_set_crtc_for_plane(plane_state, crtc);
+			if (!ret)
+				ret = drm_atomic_commit(state);
+		}
+
+		if (ret == -EDEADLK) {
+			drm_modeset_backoff(state->acquire_ctx);
+			drm_atomic_state_clear(state);
+			goto retry;
+		}
+
+		if (ret)
+			drm_atomic_state_free(state);
+
+		if (ret == 0 && event) {
+			spin_lock_irq(&dev->event_lock);
+			drm_crtc_send_vblank_event(crtc, event);
+			spin_unlock_irq(&dev->event_lock);
+		}
+	}
+	return ret;
 }
 
+
 /**
  * intel_wm_need_update - Check whether watermarks need updating
  * @plane: drm plane
@@ -11312,6 +11986,8 @@ static int intel_crtc_atomic_check(struct drm_crtc *crtc,
 
 static const struct drm_crtc_helper_funcs intel_helper_funcs = {
 	.mode_set_base_atomic = intel_pipe_set_base_atomic,
+	.atomic_begin = intel_begin_crtc_commit,
+	.atomic_flush = intel_finish_crtc_commit,
 	.atomic_check = intel_crtc_atomic_check,
 };
 
@@ -12151,8 +12827,7 @@ verify_connector_state(struct drm_device *dev, struct drm_crtc *crtc)
 		if (state->crtc != crtc)
 			continue;
 
-		intel_connector_verify_state(to_intel_connector(connector),
-					     connector->state);
+		intel_connector_verify_state(to_intel_connector(connector));
 
 		I915_STATE_WARN(state->best_encoder != encoder,
 		     "connector's atomic encoder doesn't match legacy encoder\n");
@@ -12354,7 +13029,12 @@ intel_modeset_verify_crtc(struct drm_crtc *crtc,
 			 struct drm_crtc_state *old_state,
 			 struct drm_crtc_state *new_state)
 {
+	if (!needs_modeset(new_state) &&
+	    !to_intel_crtc_state(new_state)->update_pipe)
+		return;
+
 	verify_wm_state(crtc, new_state);
+	verify_connector_state(crtc->dev, crtc);
 	verify_crtc_state(crtc, old_state, new_state);
 	verify_shared_dpll_state(crtc->dev, crtc, old_state, new_state);
 }
@@ -12698,83 +13378,32 @@ static int intel_atomic_check(struct drm_device *dev,
 	return calc_watermark_data(state);
 }
 
-static bool needs_work(struct drm_crtc_state *crtc_state)
-{
-	/* hw state checker needs to run */
-	if (needs_modeset(crtc_state))
-		return true;
-
-	/* unpin old fb's, possibly vblank update */
-	if (crtc_state->planes_changed)
-		return true;
-
-	/* pipe parameters need to be updated, and hw state checker */
-	if (to_intel_crtc_state(crtc_state)->update_pipe)
-		return true;
-
-	/* vblank event requested? */
-	if (crtc_state->event)
-		return true;
-
-	return false;
-}
-
 static int intel_atomic_prepare_commit(struct drm_device *dev,
 				       struct drm_atomic_state *state,
 				       bool nonblock)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
 	struct drm_plane_state *plane_state;
 	struct drm_crtc_state *crtc_state;
 	struct drm_plane *plane;
 	struct drm_crtc *crtc;
 	int i, ret;
 
-	for_each_crtc_in_state(state, crtc, crtc_state, i) {
-		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-		struct intel_flip_work *work;
-
-		if (!state->legacy_cursor_update) {
-			ret = intel_crtc_wait_for_pending_flips(crtc);
-			if (ret)
-				return ret;
-
-			if (atomic_read(&intel_crtc->unpin_work_count) >= 2)
-				flush_workqueue(dev_priv->wq);
-		}
+	if (nonblock) {
+		DRM_DEBUG_KMS("i915 does not yet support nonblocking commit\n");
+		return -EINVAL;
+	}
 
-		/* test if we need to update something */
-		if (!needs_work(crtc_state))
+	for_each_crtc_in_state(state, crtc, crtc_state, i) {
+		if (state->legacy_cursor_update)
 			continue;
 
-		intel_state->work[i] = work =
-			kzalloc(sizeof(**intel_state->work), GFP_KERNEL);
-
-		if (!work)
-			return -ENOMEM;
-
-		if (needs_modeset(crtc_state) ||
-		    to_intel_crtc_state(crtc_state)->update_pipe) {
-			work->num_old_connectors = hweight32(crtc->state->connector_mask);
-
-			work->old_connector_state = kcalloc(work->num_old_connectors,
-							    sizeof(*work->old_connector_state),
-							    GFP_KERNEL);
-
-			work->num_new_connectors = hweight32(crtc_state->connector_mask);
-			work->new_connector_state = kcalloc(work->num_new_connectors,
-							    sizeof(*work->new_connector_state),
-							    GFP_KERNEL);
-
-			if (!work->old_connector_state || !work->new_connector_state)
-				return -ENOMEM;
-		}
-	}
+		ret = intel_crtc_wait_for_pending_flips(crtc);
+		if (ret)
+			return ret;
 
-	if (intel_state->modeset && nonblock) {
-		DRM_DEBUG_ATOMIC("Nonblock modesets are not yet supported!\n");
-		return -EINVAL;
+		if (atomic_read(&to_intel_crtc(crtc)->unpin_work_count) >= 2)
+			flush_workqueue(dev_priv->wq);
 	}
 
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
@@ -12789,15 +13418,6 @@ static int intel_atomic_prepare_commit(struct drm_device *dev,
 			struct intel_plane_state *intel_plane_state =
 				to_intel_plane_state(plane_state);
 
-			if (plane_state->fence) {
-				long lret = fence_wait(plane_state->fence, true);
-
-				if (lret < 0) {
-					ret = lret;
-					break;
-				}
-			}
-
 			if (!intel_plane_state->wait_req)
 				continue;
 
@@ -12827,157 +13447,69 @@ u32 intel_crtc_get_vblank_counter(struct intel_crtc *crtc)
 	return dev->driver->get_vblank_counter(dev, crtc->pipe);
 }
 
-static void intel_prepare_work(struct drm_crtc *crtc,
-			       struct intel_flip_work *work,
-			       struct drm_atomic_state *state,
-			       struct drm_crtc_state *old_crtc_state)
+static void intel_atomic_wait_for_vblanks(struct drm_device *dev,
+					  struct drm_i915_private *dev_priv,
+					  unsigned crtc_mask)
 {
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct drm_plane_state *old_plane_state;
-	struct drm_plane *plane;
-	int i, j = 0;
+	unsigned last_vblank_count[I915_MAX_PIPES];
+	enum pipe pipe;
+	int ret;
 
-	INIT_WORK(&work->unpin_work, intel_unpin_work_fn);
-	INIT_WORK(&work->mmio_work, intel_mmio_flip_work_func);
-	atomic_inc(&intel_crtc->unpin_work_count);
+	if (!crtc_mask)
+		return;
 
-	for_each_plane_in_state(state, plane, old_plane_state, i) {
-		struct intel_plane_state *old_state = to_intel_plane_state(old_plane_state);
-		struct intel_plane_state *new_state = to_intel_plane_state(plane->state);
+	for_each_pipe(dev_priv, pipe) {
+		struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
 
-		if (old_state->base.crtc != crtc &&
-		    new_state->base.crtc != crtc)
+		if (!((1 << pipe) & crtc_mask))
 			continue;
 
-		if (plane->type == DRM_PLANE_TYPE_PRIMARY) {
-			plane->fb = new_state->base.fb;
-			crtc->x = new_state->base.src_x >> 16;
-			crtc->y = new_state->base.src_y >> 16;
+		ret = drm_crtc_vblank_get(crtc);
+		if (WARN_ON(ret != 0)) {
+			crtc_mask &= ~(1 << pipe);
+			continue;
 		}
 
-		old_state->wait_req = new_state->wait_req;
-		new_state->wait_req = NULL;
-
-		old_state->base.fence = new_state->base.fence;
-		new_state->base.fence = NULL;
-
-		/* remove plane state from the atomic state and move it to work */
-		old_plane_state->state = NULL;
-		state->planes[i] = NULL;
-		state->plane_states[i] = NULL;
-
-		work->old_plane_state[j] = old_state;
-		work->new_plane_state[j++] = new_state;
+		last_vblank_count[pipe] = drm_crtc_vblank_count(crtc);
 	}
 
-	old_crtc_state->state = NULL;
-	state->crtcs[drm_crtc_index(crtc)] = NULL;
-	state->crtc_states[drm_crtc_index(crtc)] = NULL;
-
-	work->old_crtc_state = to_intel_crtc_state(old_crtc_state);
-	work->new_crtc_state = to_intel_crtc_state(crtc->state);
-	work->num_planes = j;
-
-	work->event = crtc->state->event;
-	crtc->state->event = NULL;
-
-	if (needs_modeset(crtc->state) || work->new_crtc_state->update_pipe) {
-		struct drm_connector *conn;
-		struct drm_connector_state *old_conn_state;
-		int k = 0;
-
-		j = 0;
-
-		/*
-		 * intel_unpin_work_fn cannot depend on the connector list
-		 * because it may be freed from underneath it, so add
-		 * them all to the work struct while we're holding locks.
-		 */
-		for_each_connector_in_state(state, conn, old_conn_state, i) {
-			if (old_conn_state->crtc == crtc) {
-				work->old_connector_state[j++] = old_conn_state;
-
-				state->connectors[i] = NULL;
-				state->connector_states[i] = NULL;
-			}
-		}
-
-		/* If another crtc has stolen the connector from state,
-		 * then for_each_connector_in_state is no longer reliable,
-		 * so use drm_for_each_connector here.
-		 */
-		drm_for_each_connector(conn, state->dev)
-			if (conn->state->crtc == crtc)
-				work->new_connector_state[k++] = conn->state;
-
-		WARN(j != work->num_old_connectors, "j = %i, expected %i\n", j, work->num_old_connectors);
-		WARN(k != work->num_new_connectors, "k = %i, expected %i\n", k, work->num_new_connectors);
-	} else if (!work->new_crtc_state->update_wm_post)
-		work->can_async_unpin = true;
-
-	work->fb_bits = work->new_crtc_state->fb_bits;
-}
-
-static void intel_schedule_unpin(struct drm_crtc *crtc,
-				 struct intel_atomic_state *state,
-				 struct intel_flip_work *work)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	to_intel_crtc(crtc)->config = work->new_crtc_state;
-
-	queue_work(dev_priv->wq, &work->unpin_work);
-}
+	for_each_pipe(dev_priv, pipe) {
+		struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+		long lret;
 
-static void intel_schedule_flip(struct drm_crtc *crtc,
-				struct intel_atomic_state *state,
-				struct intel_flip_work *work,
-				bool nonblock)
-{
-	struct intel_crtc_state *crtc_state = work->new_crtc_state;
+		if (!((1 << pipe) & crtc_mask))
+			continue;
 
-	if (crtc_state->base.planes_changed ||
-	    needs_modeset(&crtc_state->base) ||
-	    crtc_state->update_pipe) {
-		if (nonblock)
-			schedule_work(&work->mmio_work);
-		else
-			intel_mmio_flip_work_func(&work->mmio_work);
-	} else {
-		int ret;
+		lret = wait_event_timeout(dev->vblank[pipe].queue,
+				last_vblank_count[pipe] !=
+					drm_crtc_vblank_count(crtc),
+				msecs_to_jiffies(50));
 
-		ret = drm_crtc_vblank_get(crtc);
-		I915_STATE_WARN(ret < 0, "enabling vblank failed with %i\n", ret);
+		WARN(!lret, "pipe %c vblank wait timed out\n", pipe_name(pipe));
 
-		work->flip_queued_vblank = intel_crtc_get_vblank_counter(to_intel_crtc(crtc));
-		smp_mb__before_atomic();
-		atomic_set(&work->pending, 1);
+		drm_crtc_vblank_put(crtc);
 	}
 }
 
-static void intel_schedule_update(struct drm_crtc *crtc,
-				  struct intel_atomic_state *state,
-				  struct intel_flip_work *work,
-				  bool nonblock)
+static bool needs_vblank_wait(struct intel_crtc_state *crtc_state)
 {
-	struct drm_device *dev = crtc->dev;
-	struct intel_crtc_state *pipe_config = work->new_crtc_state;
+	/* fb updated, need to unpin old fb */
+	if (crtc_state->fb_changed)
+		return true;
 
-	if (!pipe_config->base.active && work->can_async_unpin) {
-		INIT_LIST_HEAD(&work->head);
-		intel_schedule_unpin(crtc, state, work);
-		return;
-	}
+	/* wm changes, need vblank before final wm's */
+	if (crtc_state->update_wm_post)
+		return true;
 
-	spin_lock_irq(&dev->event_lock);
-	list_add_tail(&work->head, &to_intel_crtc(crtc)->flip_work);
-	spin_unlock_irq(&dev->event_lock);
+	/*
+	 * 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)
+		return true;
 
-	if (!pipe_config->base.active)
-		intel_schedule_unpin(crtc, state, work);
-	else
-		intel_schedule_flip(crtc, state, work, nonblock);
+	return false;
 }
 
 /**
@@ -13004,7 +13536,11 @@ static int intel_atomic_commit(struct drm_device *dev,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_crtc_state *old_crtc_state;
 	struct drm_crtc *crtc;
+	struct intel_crtc_state *intel_cstate;
 	int ret = 0, i;
+	bool hw_check = intel_state->modeset;
+	unsigned long put_domains[I915_MAX_PIPES] = {};
+	unsigned crtc_vblank_mask = 0;
 
 	ret = intel_atomic_prepare_commit(dev, state, nonblock);
 	if (ret) {
@@ -13022,20 +13558,27 @@ static int intel_atomic_commit(struct drm_device *dev,
 		       sizeof(intel_state->min_pixclk));
 		dev_priv->active_crtcs = intel_state->active_crtcs;
 		dev_priv->atomic_cdclk_freq = intel_state->cdclk;
+
+		intel_display_power_get(dev_priv, POWER_DOMAIN_MODESET);
 	}
 
 	for_each_crtc_in_state(state, crtc, old_crtc_state, i) {
 		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 
+		if (needs_modeset(crtc->state) ||
+		    to_intel_crtc_state(crtc->state)->update_pipe) {
+			hw_check = true;
+
+			put_domains[to_intel_crtc(crtc)->pipe] =
+				modeset_get_crtc_power_domains(crtc,
+					to_intel_crtc_state(crtc->state));
+		}
+
 		if (!needs_modeset(crtc->state))
 			continue;
 
 		intel_pre_plane_update(to_intel_crtc_state(old_crtc_state));
 
-		intel_state->work[i]->put_power_domains =
-			modeset_get_crtc_power_domains(crtc,
-				to_intel_crtc_state(crtc->state));
-
 		if (old_crtc_state->active) {
 			intel_crtc_disable_planes(crtc, old_crtc_state->plane_mask);
 			dev_priv->display.crtc_disable(crtc);
@@ -13071,9 +13614,11 @@ static int intel_atomic_commit(struct drm_device *dev,
 
 	/* Now enable the clocks, plane, pipe, and connectors that we set up. */
 	for_each_crtc_in_state(state, crtc, old_crtc_state, i) {
-		struct intel_flip_work *work = intel_state->work[i];
 		struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 		bool modeset = needs_modeset(crtc->state);
+		struct intel_crtc_state *pipe_config =
+			to_intel_crtc_state(crtc->state);
+		bool update_pipe = !modeset && pipe_config->update_pipe;
 
 		if (modeset && crtc->state->active) {
 			update_scanline_offset(to_intel_crtc(crtc));
@@ -13083,30 +13628,53 @@ static int intel_atomic_commit(struct drm_device *dev,
 		if (!modeset)
 			intel_pre_plane_update(to_intel_crtc_state(old_crtc_state));
 
-		if (!work) {
-			if (!list_empty_careful(&intel_crtc->flip_work)) {
-				spin_lock_irq(&dev->event_lock);
-				if (!list_empty(&intel_crtc->flip_work))
-					work = list_last_entry(&intel_crtc->flip_work,
-							       struct intel_flip_work, head);
-
-				if (work && work->new_crtc_state == to_intel_crtc_state(old_crtc_state)) {
-					work->free_new_crtc_state = true;
-					state->crtc_states[i] = NULL;
-					state->crtcs[i] = NULL;
-				}
-				spin_unlock_irq(&dev->event_lock);
-			}
-			continue;
-		}
+		if (crtc->state->active &&
+		    drm_atomic_get_existing_plane_state(state, crtc->primary))
+			intel_fbc_enable(intel_crtc);
+
+		if (crtc->state->active &&
+		    (crtc->state->planes_changed || update_pipe))
+			drm_atomic_helper_commit_planes_on_crtc(old_crtc_state);
 
-		intel_state->work[i] = NULL;
-		intel_prepare_work(crtc, work, state, old_crtc_state);
-		intel_schedule_update(crtc, intel_state, work, nonblock);
+		if (pipe_config->base.active && needs_vblank_wait(pipe_config))
+			crtc_vblank_mask |= 1 << i;
 	}
 
 	/* FIXME: add subpixel order */
 
+	if (!state->legacy_cursor_update)
+		intel_atomic_wait_for_vblanks(dev, dev_priv, crtc_vblank_mask);
+
+	/*
+	 * Now that the vblank has passed, we can go ahead and program the
+	 * optimal watermarks on platforms that need two-step watermark
+	 * programming.
+	 *
+	 * TODO: Move this (and other cleanup) to an async worker eventually.
+	 */
+	for_each_crtc_in_state(state, crtc, old_crtc_state, i) {
+		intel_cstate = to_intel_crtc_state(crtc->state);
+
+		if (dev_priv->display.optimize_watermarks)
+			dev_priv->display.optimize_watermarks(intel_cstate);
+	}
+
+	for_each_crtc_in_state(state, crtc, old_crtc_state, i) {
+		intel_post_plane_update(to_intel_crtc_state(old_crtc_state));
+
+		if (put_domains[i])
+			modeset_put_power_domains(dev_priv, put_domains[i]);
+
+		intel_modeset_verify_crtc(crtc, old_crtc_state, crtc->state);
+	}
+
+	if (intel_state->modeset)
+		intel_display_power_put(dev_priv, POWER_DOMAIN_MODESET);
+
+	mutex_lock(&dev->struct_mutex);
+	drm_atomic_helper_cleanup_planes(dev, state);
+	mutex_unlock(&dev->struct_mutex);
+
 	drm_atomic_state_free(state);
 
 	/* As one of the primary mmio accessors, KMS has a high likelihood
@@ -13170,38 +13738,11 @@ 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 = drm_atomic_helper_page_flip,
+	.page_flip = intel_crtc_page_flip,
 	.atomic_duplicate_state = intel_crtc_duplicate_state,
 	.atomic_destroy_state = intel_crtc_destroy_state,
 };
 
-static struct fence *intel_get_excl_fence(struct drm_i915_gem_object *obj)
-{
-	struct reservation_object *resv;
-
-
-	if (!obj->base.dma_buf)
-		return NULL;
-
-	resv = obj->base.dma_buf->resv;
-
-	/* For framebuffer backed by dmabuf, wait for fence */
-	while (1) {
-		struct fence *fence_excl, *ret = NULL;
-
-		rcu_read_lock();
-
-		fence_excl = rcu_dereference(resv->fence_excl);
-		if (fence_excl)
-			ret = fence_get_rcu(fence_excl);
-
-		rcu_read_unlock();
-
-		if (ret == fence_excl)
-			return ret;
-	}
-}
-
 /**
  * intel_prepare_plane_fb - Prepare fb for usage on plane
  * @plane: drm plane to prepare for
@@ -13225,20 +13766,11 @@ intel_prepare_plane_fb(struct drm_plane *plane,
 	struct intel_plane *intel_plane = to_intel_plane(plane);
 	struct drm_i915_gem_object *obj = intel_fb_obj(fb);
 	struct drm_i915_gem_object *old_obj = intel_fb_obj(plane->state->fb);
-	struct drm_crtc *crtc = new_state->crtc ?: plane->state->crtc;
 	int ret = 0;
 
 	if (!obj && !old_obj)
 		return 0;
 
-	if (WARN_ON(!new_state->state) || WARN_ON(!crtc) ||
-	    WARN_ON(!to_intel_atomic_state(new_state->state)->work[to_intel_crtc(crtc)->pipe])) {
-		if (WARN_ON(old_obj != obj))
-			return -EINVAL;
-
-		return 0;
-	}
-
 	if (old_obj) {
 		struct drm_crtc_state *crtc_state =
 			drm_atomic_get_existing_crtc_state(new_state->state, plane->state->crtc);
@@ -13263,6 +13795,19 @@ intel_prepare_plane_fb(struct drm_plane *plane,
 		}
 	}
 
+	/* For framebuffer backed by dmabuf, wait for fence */
+	if (obj && obj->base.dma_buf) {
+		long lret;
+
+		lret = reservation_object_wait_timeout_rcu(obj->base.dma_buf->resv,
+							   false, true,
+							   MAX_SCHEDULE_TIMEOUT);
+		if (lret == -ERESTARTSYS)
+			return lret;
+
+		WARN(lret < 0, "waiting returns %li\n", lret);
+	}
+
 	if (!obj) {
 		ret = 0;
 	} else if (plane->type == DRM_PLANE_TYPE_CURSOR &&
@@ -13282,8 +13827,6 @@ intel_prepare_plane_fb(struct drm_plane *plane,
 
 			i915_gem_request_assign(&plane_state->wait_req,
 						obj->last_write_req);
-
-			plane_state->base.fence = intel_get_excl_fence(obj);
 		}
 
 		i915_gem_track_fb(old_obj, obj, intel_plane->frontbuffer_bit);
@@ -13326,9 +13869,6 @@ intel_cleanup_plane_fb(struct drm_plane *plane,
 		i915_gem_track_fb(old_obj, obj, intel_plane->frontbuffer_bit);
 
 	i915_gem_request_assign(&old_intel_state->wait_req, NULL);
-
-	fence_put(old_intel_state->base.fence);
-	old_intel_state->base.fence = NULL;
 }
 
 int
@@ -13388,6 +13928,40 @@ intel_check_primary_plane(struct drm_plane *plane,
 					     &state->visible);
 }
 
+static void intel_begin_crtc_commit(struct drm_crtc *crtc,
+				    struct drm_crtc_state *old_crtc_state)
+{
+	struct drm_device *dev = crtc->dev;
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+	struct intel_crtc_state *old_intel_state =
+		to_intel_crtc_state(old_crtc_state);
+	bool modeset = needs_modeset(crtc->state);
+
+	/* Perform vblank evasion around commit operation */
+	intel_pipe_update_start(intel_crtc);
+
+	if (modeset)
+		return;
+
+	if (crtc->state->color_mgmt_changed || to_intel_crtc_state(crtc->state)->update_pipe) {
+		intel_color_set_csc(crtc->state);
+		intel_color_load_luts(crtc->state);
+	}
+
+	if (to_intel_crtc_state(crtc->state)->update_pipe)
+		intel_update_pipe_config(intel_crtc, old_intel_state);
+	else if (INTEL_INFO(dev)->gen >= 9)
+		skl_detach_scalers(intel_crtc);
+}
+
+static void intel_finish_crtc_commit(struct drm_crtc *crtc,
+				     struct drm_crtc_state *old_crtc_state)
+{
+	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
+
+	intel_pipe_update_end(intel_crtc, NULL);
+}
+
 /**
  * intel_plane_destroy - destroy a plane
  * @plane: plane to destroy
@@ -13698,8 +14272,6 @@ static void intel_crtc_init(struct drm_device *dev, int pipe)
 	intel_crtc->base.state = &crtc_state->base;
 	crtc_state->base.crtc = &intel_crtc->base;
 
-	INIT_LIST_HEAD(&intel_crtc->flip_work);
-
 	/* initialize shared scalers */
 	if (INTEL_INFO(dev)->gen >= 9) {
 		if (pipe == PIPE_C)
@@ -14450,6 +15022,34 @@ void intel_init_display_hooks(struct drm_i915_private *dev_priv)
 		dev_priv->display.modeset_calc_cdclk =
 			broxton_modeset_calc_cdclk;
 	}
+
+	switch (INTEL_INFO(dev_priv)->gen) {
+	case 2:
+		dev_priv->display.queue_flip = intel_gen2_queue_flip;
+		break;
+
+	case 3:
+		dev_priv->display.queue_flip = intel_gen3_queue_flip;
+		break;
+
+	case 4:
+	case 5:
+		dev_priv->display.queue_flip = intel_gen4_queue_flip;
+		break;
+
+	case 6:
+		dev_priv->display.queue_flip = intel_gen6_queue_flip;
+		break;
+	case 7:
+	case 8: /* FIXME(BDW): Check that the gen8 RCS flip works. */
+		dev_priv->display.queue_flip = intel_gen7_queue_flip;
+		break;
+	case 9:
+		/* Drop through - unsupported since execlist only. */
+	default:
+		/* Default just returns -ENODEV to indicate unsupported */
+		dev_priv->display.queue_flip = intel_default_queue_flip;
+	}
 }
 
 /*
@@ -15407,9 +16007,9 @@ void intel_modeset_gem_init(struct drm_device *dev)
 			DRM_ERROR("failed to pin boot fb on pipe %d\n",
 				  to_intel_crtc(c)->pipe);
 			drm_framebuffer_unreference(c->primary->fb);
-			drm_framebuffer_unreference(c->primary->state->fb);
-			c->primary->fb = c->primary->state->fb = NULL;
+			c->primary->fb = NULL;
 			c->primary->crtc = c->primary->state->crtc = NULL;
+			update_state_fb(c->primary);
 			c->state->plane_mask &= ~(1 << drm_plane_index(c->primary));
 		}
 	}
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 0741b2d3aa65..97de5e05890a 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -304,8 +304,6 @@ struct intel_atomic_state {
 	unsigned int active_crtcs;
 	unsigned int min_pixclk[I915_MAX_PIPES];
 
-	struct intel_flip_work *work[I915_MAX_PIPES];
-
 	struct intel_shared_dpll_config shared_dpll[I915_NUM_PLLS];
 
 	/*
@@ -643,7 +641,7 @@ struct intel_crtc {
 	unsigned long enabled_power_domains;
 	bool lowfreq_avail;
 	struct intel_overlay *overlay;
-	struct list_head flip_work;
+	struct intel_flip_work *flip_work;
 
 	atomic_t unpin_work_count;
 
@@ -661,6 +659,9 @@ struct intel_crtc {
 
 	struct intel_crtc_state *config;
 
+	/* reset counter value when the last flip was submitted */
+	unsigned int reset_counter;
+
 	/* Access to these should be protected by dev_priv->irq_lock. */
 	bool cpu_fifo_underrun_disabled;
 	bool pch_fifo_underrun_disabled;
@@ -969,28 +970,20 @@ intel_get_crtc_for_plane(struct drm_device *dev, int plane)
 }
 
 struct intel_flip_work {
-	struct list_head head;
-
 	struct work_struct unpin_work;
 	struct work_struct mmio_work;
 
+	struct drm_crtc *crtc;
+	struct drm_framebuffer *old_fb;
+	struct drm_i915_gem_object *pending_flip_obj;
 	struct drm_pending_vblank_event *event;
 	atomic_t pending;
+	u32 flip_count;
+	u32 gtt_offset;
+	struct drm_i915_gem_request *flip_queued_req;
 	u32 flip_queued_vblank;
-
-	unsigned put_power_domains;
-	unsigned num_planes;
-
-	bool can_async_unpin, free_new_crtc_state;
-	unsigned fb_bits;
-
-	unsigned num_old_connectors, num_new_connectors;
-	struct drm_connector_state **old_connector_state;
-	struct drm_connector_state **new_connector_state;
-
-	struct intel_crtc_state *old_crtc_state, *new_crtc_state;
-	struct intel_plane_state *old_plane_state[I915_MAX_PLANES + 1];
-	struct intel_plane_state *new_plane_state[I915_MAX_PLANES + 1];
+	u32 flip_ready_vblank;
+	unsigned int rotation;
 };
 
 struct intel_load_detect_pipe {
@@ -1149,7 +1142,6 @@ unsigned int intel_rotation_info_size(const struct intel_rotation_info *rot_info
 bool intel_has_pending_fb_unpin(struct drm_device *dev);
 void intel_mark_busy(struct drm_i915_private *dev_priv);
 void intel_mark_idle(struct drm_i915_private *dev_priv);
-void intel_free_flip_work(struct intel_flip_work *work);
 void intel_crtc_restore_mode(struct drm_crtc *crtc);
 int intel_display_suspend(struct drm_device *dev);
 void intel_encoder_destroy(struct drm_encoder *encoder);
@@ -1202,8 +1194,9 @@ struct drm_framebuffer *
 __intel_framebuffer_create(struct drm_device *dev,
 			   struct drm_mode_fb_cmd2 *mode_cmd,
 			   struct drm_i915_gem_object *obj);
+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);
 int intel_prepare_plane_fb(struct drm_plane *plane,
 			   const struct drm_plane_state *new_state);
 void intel_cleanup_plane_fb(struct drm_plane *plane,
@@ -1427,15 +1420,11 @@ static inline void intel_fbdev_restore_mode(struct drm_device *dev)
 void intel_fbc_choose_crtc(struct drm_i915_private *dev_priv,
 			   struct drm_atomic_state *state);
 bool intel_fbc_is_active(struct drm_i915_private *dev_priv);
-void intel_fbc_pre_update(struct intel_crtc *crtc,
-			  struct intel_crtc_state *crtc_state,
-			  struct intel_plane_state *plane_state);
+void intel_fbc_pre_update(struct intel_crtc *crtc);
 void intel_fbc_post_update(struct intel_crtc *crtc);
 void intel_fbc_init(struct drm_i915_private *dev_priv);
 void intel_fbc_init_pipe_state(struct drm_i915_private *dev_priv);
-void intel_fbc_enable(struct intel_crtc *crtc,
-		      struct intel_crtc_state *crtc_state,
-		      struct intel_plane_state *plane_state);
+void intel_fbc_enable(struct intel_crtc *crtc);
 void intel_fbc_disable(struct intel_crtc *crtc);
 void intel_fbc_global_disable(struct drm_i915_private *dev_priv);
 void intel_fbc_invalidate(struct drm_i915_private *dev_priv,
diff --git a/drivers/gpu/drm/i915/intel_fbc.c b/drivers/gpu/drm/i915/intel_fbc.c
index d2b0269b2fe4..0dea5fbcd8aa 100644
--- a/drivers/gpu/drm/i915/intel_fbc.c
+++ b/drivers/gpu/drm/i915/intel_fbc.c
@@ -480,10 +480,10 @@ static void intel_fbc_deactivate(struct drm_i915_private *dev_priv)
 		intel_fbc_hw_deactivate(dev_priv);
 }
 
-static bool multiple_pipes_ok(struct intel_crtc *crtc,
-			      struct intel_plane_state *plane_state)
+static bool multiple_pipes_ok(struct intel_crtc *crtc)
 {
-	struct drm_i915_private *dev_priv = to_i915(crtc->base.dev);
+	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
+	struct drm_plane *primary = crtc->base.primary;
 	struct intel_fbc *fbc = &dev_priv->fbc;
 	enum pipe pipe = crtc->pipe;
 
@@ -491,7 +491,9 @@ static bool multiple_pipes_ok(struct intel_crtc *crtc,
 	if (!no_fbc_on_multiple_pipes(dev_priv))
 		return true;
 
-	if (plane_state->visible)
+	WARN_ON(!drm_modeset_is_locked(&primary->mutex));
+
+	if (to_intel_plane_state(primary->state)->visible)
 		fbc->visible_pipes_mask |= (1 << pipe);
 	else
 		fbc->visible_pipes_mask &= ~(1 << pipe);
@@ -706,16 +708,21 @@ static bool intel_fbc_hw_tracking_covers_screen(struct intel_crtc *crtc)
 	return effective_w <= max_w && effective_h <= max_h;
 }
 
-static void intel_fbc_update_state_cache(struct intel_crtc *crtc,
-					 struct intel_crtc_state *crtc_state,
-					 struct intel_plane_state *plane_state)
+static void intel_fbc_update_state_cache(struct intel_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
 	struct intel_fbc *fbc = &dev_priv->fbc;
 	struct intel_fbc_state_cache *cache = &fbc->state_cache;
+	struct intel_crtc_state *crtc_state =
+		to_intel_crtc_state(crtc->base.state);
+	struct intel_plane_state *plane_state =
+		to_intel_plane_state(crtc->base.primary->state);
 	struct drm_framebuffer *fb = plane_state->base.fb;
 	struct drm_i915_gem_object *obj;
 
+	WARN_ON(!drm_modeset_is_locked(&crtc->base.mutex));
+	WARN_ON(!drm_modeset_is_locked(&crtc->base.primary->mutex));
+
 	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 =
@@ -880,9 +887,7 @@ static bool intel_fbc_reg_params_equal(struct intel_fbc_reg_params *params1,
 	return memcmp(params1, params2, sizeof(*params1)) == 0;
 }
 
-void intel_fbc_pre_update(struct intel_crtc *crtc,
-			  struct intel_crtc_state *crtc_state,
-			  struct intel_plane_state *plane_state)
+void intel_fbc_pre_update(struct intel_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
 	struct intel_fbc *fbc = &dev_priv->fbc;
@@ -892,7 +897,7 @@ void intel_fbc_pre_update(struct intel_crtc *crtc,
 
 	mutex_lock(&fbc->lock);
 
-	if (!multiple_pipes_ok(crtc, plane_state)) {
+	if (!multiple_pipes_ok(crtc)) {
 		fbc->no_fbc_reason = "more than one pipe active";
 		goto deactivate;
 	}
@@ -900,7 +905,7 @@ void intel_fbc_pre_update(struct intel_crtc *crtc,
 	if (!fbc->enabled || fbc->crtc != crtc)
 		goto unlock;
 
-	intel_fbc_update_state_cache(crtc, crtc_state, plane_state);
+	intel_fbc_update_state_cache(crtc);
 
 deactivate:
 	intel_fbc_deactivate(dev_priv);
@@ -1084,9 +1089,7 @@ out:
  * intel_fbc_enable multiple times for the same pipe without an
  * intel_fbc_disable in the middle, as long as it is deactivated.
  */
-void intel_fbc_enable(struct intel_crtc *crtc,
-		      struct intel_crtc_state *crtc_state,
-		      struct intel_plane_state *plane_state)
+void intel_fbc_enable(struct intel_crtc *crtc)
 {
 	struct drm_i915_private *dev_priv = crtc->base.dev->dev_private;
 	struct intel_fbc *fbc = &dev_priv->fbc;
@@ -1099,19 +1102,19 @@ void intel_fbc_enable(struct intel_crtc *crtc,
 	if (fbc->enabled) {
 		WARN_ON(fbc->crtc == NULL);
 		if (fbc->crtc == crtc) {
-			WARN_ON(!crtc_state->enable_fbc);
+			WARN_ON(!crtc->config->enable_fbc);
 			WARN_ON(fbc->active);
 		}
 		goto out;
 	}
 
-	if (!crtc_state->enable_fbc)
+	if (!crtc->config->enable_fbc)
 		goto out;
 
 	WARN_ON(fbc->active);
 	WARN_ON(fbc->crtc != NULL);
 
-	intel_fbc_update_state_cache(crtc, crtc_state, plane_state);
+	intel_fbc_update_state_cache(crtc);
 	if (intel_fbc_alloc_cfb(crtc)) {
 		fbc->no_fbc_reason = "not enough stolen memory";
 		goto out;
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index affca6d5ce7e..db10c961e0f4 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -260,7 +260,9 @@ int intel_sanitize_enable_execlists(struct drm_i915_private *dev_priv, int enabl
 	if (enable_execlists == 0)
 		return 0;
 
-	if (HAS_LOGICAL_RING_CONTEXTS(dev_priv) && USES_PPGTT(dev_priv))
+	if (HAS_LOGICAL_RING_CONTEXTS(dev_priv) &&
+	    USES_PPGTT(dev_priv) &&
+	    i915.use_mmio_flip >= 0)
 		return 1;
 
 	return 0;