Merge tag 'drm-fixes-for-v4.14-rc5' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "Couple of the arm people seem to wake up so this has imx and msm
  fixes, along with a bunch of i915 stable bounds fixes and an amdgpu
  regression fix.

  All seems pretty okay for now"

* tag 'drm-fixes-for-v4.14-rc5' of git://people.freedesktop.org/~airlied/linux:
  drm/msm: fix _NO_IMPLICIT fencing case
  drm/msm: fix error path cleanup
  drm/msm/mdp5: Remove extra pm_runtime_put call in mdp5_crtc_cursor_set()
  drm/msm/dsi: Use correct pm_runtime_put variant during host_init
  drm/msm: fix return value check in _msm_gem_kernel_new()
  drm/msm: use proper memory barriers for updating tail/head
  drm/msm/mdp5: add missing max size for 8x74 v1
  drm/amdgpu: fix placement flags in amdgpu_ttm_bind
  drm/i915/bios: parse DDI ports also for CHV for HDMI DDC pin and DP AUX channel
  gpu: ipu-v3: pre: implement workaround for ERR009624
  gpu: ipu-v3: prg: wait for double buffers to be filled on channel startup
  gpu: ipu-v3: Allow channel burst locking on i.MX6 only
  drm/i915: Read timings from the correct transcoder in intel_crtc_mode_get()
  drm/i915: Order two completing nop_submit_request
  drm/i915: Silence compiler warning for hsw_power_well_enable()
  drm/i915: Use crtc_state_is_legacy_gamma in intel_color_check
  drm/i915/edp: Increase the T12 delay quirk to 1300ms
  drm/i915/edp: Get the Panel Power Off timestamp after panel is off
  sync_file: Return consistent status in SYNC_IOC_FILE_INFO
  drm/atomic: Unref duplicated drm_atomic_state in drm_atomic_helper_resume()

19 files changed, 119 insertions, 50 deletions

diff --git a/drivers/dma-buf/sync_file.c b/drivers/dma-buf/sync_file.c
index 66fb40d0ebdb..03830634e141 100644
--- a/drivers/dma-buf/sync_file.c
+++ b/drivers/dma-buf/sync_file.c
@@ -383,7 +383,7 @@ err_put_fd:
 	return err;
 }
 
-static void sync_fill_fence_info(struct dma_fence *fence,
+static int sync_fill_fence_info(struct dma_fence *fence,
 				 struct sync_fence_info *info)
 {
 	strlcpy(info->obj_name, fence->ops->get_timeline_name(fence),
@@ -399,6 +399,8 @@ static void sync_fill_fence_info(struct dma_fence *fence,
 		test_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT, &fence->flags) ?
 		ktime_to_ns(fence->timestamp) :
 		ktime_set(0, 0);
+
+	return info->status;
 }
 
 static long sync_file_ioctl_fence_info(struct sync_file *sync_file,
@@ -424,8 +426,12 @@ static long sync_file_ioctl_fence_info(struct sync_file *sync_file,
 	 * sync_fence_info and return the actual number of fences on
 	 * info->num_fences.
 	 */
-	if (!info.num_fences)
+	if (!info.num_fences) {
+		info.status = dma_fence_is_signaled(sync_file->fence);
 		goto no_fences;
+	} else {
+		info.status = 1;
+	}
 
 	if (info.num_fences < num_fences)
 		return -EINVAL;
@@ -435,8 +441,10 @@ static long sync_file_ioctl_fence_info(struct sync_file *sync_file,
 	if (!fence_info)
 		return -ENOMEM;
 
-	for (i = 0; i < num_fences; i++)
-		sync_fill_fence_info(fences[i], &fence_info[i]);
+	for (i = 0; i < num_fences; i++) {
+		int status = sync_fill_fence_info(fences[i], &fence_info[i]);
+		info.status = info.status <= 0 ? info.status : status;
+	}
 
 	if (copy_to_user(u64_to_user_ptr(info.sync_fence_info), fence_info,
 			 size)) {
@@ -446,7 +454,6 @@ static long sync_file_ioctl_fence_info(struct sync_file *sync_file,
 
 no_fences:
 	sync_file_get_name(sync_file, info.name, sizeof(info.name));
-	info.status = dma_fence_is_signaled(sync_file->fence);
 	info.num_fences = num_fences;
 
 	if (copy_to_user((void __user *)arg, &info, sizeof(info)))
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 7ef6c28a34d9..bc746131987f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -834,7 +834,7 @@ int amdgpu_ttm_bind(struct ttm_buffer_object *bo, struct ttm_mem_reg *bo_mem)
 	placement.busy_placement = &placements;
 	placements.fpfn = 0;
 	placements.lpfn = adev->mc.gart_size >> PAGE_SHIFT;
-	placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
+	placements.flags = bo->mem.placement | TTM_PL_FLAG_TT;
 
 	r = ttm_bo_mem_space(bo, &placement, &tmp, true, false);
 	if (unlikely(r))
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
index 4e53aae9a1fb..0028591f3f95 100644
--- a/drivers/gpu/drm/drm_atomic_helper.c
+++ b/drivers/gpu/drm/drm_atomic_helper.c
@@ -2960,6 +2960,7 @@ out:
 		drm_modeset_backoff(&ctx);
 	}
 
+	drm_atomic_state_put(state);
 	drm_modeset_drop_locks(&ctx);
 	drm_modeset_acquire_fini(&ctx);
 
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 19404c96eeb1..af289d35b77a 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -3013,10 +3013,15 @@ void i915_gem_reset_finish(struct drm_i915_private *dev_priv)
 
 static void nop_submit_request(struct drm_i915_gem_request *request)
 {
+	unsigned long flags;
+
 	GEM_BUG_ON(!i915_terminally_wedged(&request->i915->gpu_error));
 	dma_fence_set_error(&request->fence, -EIO);
-	i915_gem_request_submit(request);
+
+	spin_lock_irqsave(&request->engine->timeline->lock, flags);
+	__i915_gem_request_submit(request);
 	intel_engine_init_global_seqno(request->engine, request->global_seqno);
+	spin_unlock_irqrestore(&request->engine->timeline->lock, flags);
 }
 
 static void engine_set_wedged(struct intel_engine_cs *engine)
diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c
index 00c6aee0a9a1..5d4cd3d00564 100644
--- a/drivers/gpu/drm/i915/intel_bios.c
+++ b/drivers/gpu/drm/i915/intel_bios.c
@@ -1240,7 +1240,7 @@ static void parse_ddi_ports(struct drm_i915_private *dev_priv,
 {
 	enum port port;
 
-	if (!HAS_DDI(dev_priv))
+	if (!HAS_DDI(dev_priv) && !IS_CHERRYVIEW(dev_priv))
 		return;
 
 	if (!dev_priv->vbt.child_dev_num)
diff --git a/drivers/gpu/drm/i915/intel_color.c b/drivers/gpu/drm/i915/intel_color.c
index ff9ecd211abb..b8315bca852b 100644
--- a/drivers/gpu/drm/i915/intel_color.c
+++ b/drivers/gpu/drm/i915/intel_color.c
@@ -74,7 +74,7 @@
 #define I9XX_CSC_COEFF_1_0		\
 	((7 << 12) | I9XX_CSC_COEFF_FP(CTM_COEFF_1_0, 8))
 
-static bool crtc_state_is_legacy(struct drm_crtc_state *state)
+static bool crtc_state_is_legacy_gamma(struct drm_crtc_state *state)
 {
 	return !state->degamma_lut &&
 		!state->ctm &&
@@ -288,7 +288,7 @@ static void cherryview_load_csc_matrix(struct drm_crtc_state *state)
 	}
 
 	mode = (state->ctm ? CGM_PIPE_MODE_CSC : 0);
-	if (!crtc_state_is_legacy(state)) {
+	if (!crtc_state_is_legacy_gamma(state)) {
 		mode |= (state->degamma_lut ? CGM_PIPE_MODE_DEGAMMA : 0) |
 			(state->gamma_lut ? CGM_PIPE_MODE_GAMMA : 0);
 	}
@@ -469,7 +469,7 @@ static void broadwell_load_luts(struct drm_crtc_state *state)
 	struct intel_crtc_state *intel_state = to_intel_crtc_state(state);
 	enum pipe pipe = to_intel_crtc(state->crtc)->pipe;
 
-	if (crtc_state_is_legacy(state)) {
+	if (crtc_state_is_legacy_gamma(state)) {
 		haswell_load_luts(state);
 		return;
 	}
@@ -529,7 +529,7 @@ static void glk_load_luts(struct drm_crtc_state *state)
 
 	glk_load_degamma_lut(state);
 
-	if (crtc_state_is_legacy(state)) {
+	if (crtc_state_is_legacy_gamma(state)) {
 		haswell_load_luts(state);
 		return;
 	}
@@ -551,7 +551,7 @@ static void cherryview_load_luts(struct drm_crtc_state *state)
 	uint32_t i, lut_size;
 	uint32_t word0, word1;
 
-	if (crtc_state_is_legacy(state)) {
+	if (crtc_state_is_legacy_gamma(state)) {
 		/* Turn off degamma/gamma on CGM block. */
 		I915_WRITE(CGM_PIPE_MODE(pipe),
 			   (state->ctm ? CGM_PIPE_MODE_CSC : 0));
@@ -632,12 +632,10 @@ int intel_color_check(struct drm_crtc *crtc,
 		return 0;
 
 	/*
-	 * We also allow no degamma lut and a gamma lut at the legacy
+	 * We also allow no degamma lut/ctm and a gamma lut at the legacy
 	 * size (256 entries).
 	 */
-	if (!crtc_state->degamma_lut &&
-	    crtc_state->gamma_lut &&
-	    crtc_state->gamma_lut->length == LEGACY_LUT_LENGTH)
+	if (crtc_state_is_legacy_gamma(crtc_state))
 		return 0;
 
 	return -EINVAL;
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 64f7b51ed97c..5c7828c52d12 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -10245,13 +10245,10 @@ struct drm_display_mode *intel_crtc_mode_get(struct drm_device *dev,
 {
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder;
+	enum transcoder cpu_transcoder;
 	struct drm_display_mode *mode;
 	struct intel_crtc_state *pipe_config;
-	int htot = I915_READ(HTOTAL(cpu_transcoder));
-	int hsync = I915_READ(HSYNC(cpu_transcoder));
-	int vtot = I915_READ(VTOTAL(cpu_transcoder));
-	int vsync = I915_READ(VSYNC(cpu_transcoder));
+	u32 htot, hsync, vtot, vsync;
 	enum pipe pipe = intel_crtc->pipe;
 
 	mode = kzalloc(sizeof(*mode), GFP_KERNEL);
@@ -10279,6 +10276,13 @@ struct drm_display_mode *intel_crtc_mode_get(struct drm_device *dev,
 	i9xx_crtc_clock_get(intel_crtc, pipe_config);
 
 	mode->clock = pipe_config->port_clock / pipe_config->pixel_multiplier;
+
+	cpu_transcoder = pipe_config->cpu_transcoder;
+	htot = I915_READ(HTOTAL(cpu_transcoder));
+	hsync = I915_READ(HSYNC(cpu_transcoder));
+	vtot = I915_READ(VTOTAL(cpu_transcoder));
+	vsync = I915_READ(VSYNC(cpu_transcoder));
+
 	mode->hdisplay = (htot & 0xffff) + 1;
 	mode->htotal = ((htot & 0xffff0000) >> 16) + 1;
 	mode->hsync_start = (hsync & 0xffff) + 1;
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 64134947c0aa..203198659ab2 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -2307,8 +2307,8 @@ static void edp_panel_off(struct intel_dp *intel_dp)
 	I915_WRITE(pp_ctrl_reg, pp);
 	POSTING_READ(pp_ctrl_reg);
 
-	intel_dp->panel_power_off_time = ktime_get_boottime();
 	wait_panel_off(intel_dp);
+	intel_dp->panel_power_off_time = ktime_get_boottime();
 
 	/* We got a reference when we enabled the VDD. */
 	intel_display_power_put(dev_priv, intel_dp->aux_power_domain);
@@ -5273,7 +5273,7 @@ intel_dp_init_panel_power_sequencer(struct drm_device *dev,
 	 * seems sufficient to avoid this problem.
 	 */
 	if (dev_priv->quirks & QUIRK_INCREASE_T12_DELAY) {
-		vbt.t11_t12 = max_t(u16, vbt.t11_t12, 900 * 10);
+		vbt.t11_t12 = max_t(u16, vbt.t11_t12, 1300 * 10);
 		DRM_DEBUG_KMS("Increasing T12 panel delay as per the quirk to %d\n",
 			      vbt.t11_t12);
 	}
diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
index b3a087cb0860..49577eba8e7e 100644
--- a/drivers/gpu/drm/i915/intel_runtime_pm.c
+++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
@@ -368,7 +368,7 @@ static void hsw_power_well_enable(struct drm_i915_private *dev_priv,
 {
 	enum i915_power_well_id id = power_well->id;
 	bool wait_fuses = power_well->hsw.has_fuses;
-	enum skl_power_gate pg;
+	enum skl_power_gate uninitialized_var(pg);
 	u32 val;
 
 	if (wait_fuses) {
diff --git a/drivers/gpu/drm/msm/dsi/dsi_host.c b/drivers/gpu/drm/msm/dsi/dsi_host.c
index dbb31a014419..deaf869374ea 100644
--- a/drivers/gpu/drm/msm/dsi/dsi_host.c
+++ b/drivers/gpu/drm/msm/dsi/dsi_host.c
@@ -248,7 +248,7 @@ disable_clks:
 	clk_disable_unprepare(ahb_clk);
 disable_gdsc:
 	regulator_disable(gdsc_reg);
-	pm_runtime_put_autosuspend(dev);
+	pm_runtime_put_sync(dev);
 put_clk:
 	clk_put(ahb_clk);
 put_gdsc:
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.c
index c2bdad88447e..824067d2d427 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.c
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cfg.c
@@ -83,6 +83,8 @@ const struct mdp5_cfg_hw msm8x74v1_config = {
 				  .caps = MDP_LM_CAP_WB },
 			     },
 		.nb_stages = 5,
+		.max_width = 2048,
+		.max_height = 0xFFFF,
 	},
 	.dspp = {
 		.count = 3,
diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c
index 6fcb58ab718c..440977677001 100644
--- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c
+++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c
@@ -804,8 +804,6 @@ static int mdp5_crtc_cursor_set(struct drm_crtc *crtc,
 
 	spin_unlock_irqrestore(&mdp5_crtc->cursor.lock, flags);
 
-	pm_runtime_put_autosuspend(&pdev->dev);
-
 set_cursor:
 	ret = mdp5_ctl_set_cursor(ctl, pipeline, 0, cursor_enable);
 	if (ret) {
diff --git a/drivers/gpu/drm/msm/msm_gem.c b/drivers/gpu/drm/msm/msm_gem.c
index f15821a0d900..ea5bb0e1632c 100644
--- a/drivers/gpu/drm/msm/msm_gem.c
+++ b/drivers/gpu/drm/msm/msm_gem.c
@@ -610,17 +610,6 @@ int msm_gem_sync_object(struct drm_gem_object *obj,
 	struct dma_fence *fence;
 	int i, ret;
 
-	if (!exclusive) {
-		/* NOTE: _reserve_shared() must happen before _add_shared_fence(),
-		 * which makes this a slightly strange place to call it.  OTOH this
-		 * is a convenient can-fail point to hook it in.  (And similar to
-		 * how etnaviv and nouveau handle this.)
-		 */
-		ret = reservation_object_reserve_shared(msm_obj->resv);
-		if (ret)
-			return ret;
-	}
-
 	fobj = reservation_object_get_list(msm_obj->resv);
 	if (!fobj || (fobj->shared_count == 0)) {
 		fence = reservation_object_get_excl(msm_obj->resv);
@@ -1045,10 +1034,10 @@ static void *_msm_gem_kernel_new(struct drm_device *dev, uint32_t size,
 	}
 
 	vaddr = msm_gem_get_vaddr(obj);
-	if (!vaddr) {
+	if (IS_ERR(vaddr)) {
 		msm_gem_put_iova(obj, aspace);
 		drm_gem_object_unreference(obj);
-		return ERR_PTR(-ENOMEM);
+		return ERR_CAST(vaddr);
 	}
 
 	if (bo)
diff --git a/drivers/gpu/drm/msm/msm_gem_submit.c b/drivers/gpu/drm/msm/msm_gem_submit.c
index 5d0a75d4b249..93535cac0676 100644
--- a/drivers/gpu/drm/msm/msm_gem_submit.c
+++ b/drivers/gpu/drm/msm/msm_gem_submit.c
@@ -221,7 +221,7 @@ fail:
 	return ret;
 }
 
-static int submit_fence_sync(struct msm_gem_submit *submit)
+static int submit_fence_sync(struct msm_gem_submit *submit, bool no_implicit)
 {
 	int i, ret = 0;
 
@@ -229,6 +229,20 @@ static int submit_fence_sync(struct msm_gem_submit *submit)
 		struct msm_gem_object *msm_obj = submit->bos[i].obj;
 		bool write = submit->bos[i].flags & MSM_SUBMIT_BO_WRITE;
 
+		if (!write) {
+			/* NOTE: _reserve_shared() must happen before
+			 * _add_shared_fence(), which makes this a slightly
+			 * strange place to call it.  OTOH this is a
+			 * convenient can-fail point to hook it in.
+			 */
+			ret = reservation_object_reserve_shared(msm_obj->resv);
+			if (ret)
+				return ret;
+		}
+
+		if (no_implicit)
+			continue;
+
 		ret = msm_gem_sync_object(&msm_obj->base, submit->gpu->fctx, write);
 		if (ret)
 			break;
@@ -451,11 +465,9 @@ int msm_ioctl_gem_submit(struct drm_device *dev, void *data,
 	if (ret)
 		goto out;
 
-	if (!(args->flags & MSM_SUBMIT_NO_IMPLICIT)) {
-		ret = submit_fence_sync(submit);
-		if (ret)
-			goto out;
-	}
+	ret = submit_fence_sync(submit, !!(args->flags & MSM_SUBMIT_NO_IMPLICIT));
+	if (ret)
+		goto out;
 
 	ret = submit_pin_objects(submit);
 	if (ret)
diff --git a/drivers/gpu/drm/msm/msm_gpu.c b/drivers/gpu/drm/msm/msm_gpu.c
index ffbff27600e0..6a887032c66a 100644
--- a/drivers/gpu/drm/msm/msm_gpu.c
+++ b/drivers/gpu/drm/msm/msm_gpu.c
@@ -718,7 +718,8 @@ void msm_gpu_cleanup(struct msm_gpu *gpu)
 			msm_gem_put_iova(gpu->rb->bo, gpu->aspace);
 		msm_ringbuffer_destroy(gpu->rb);
 	}
-	if (gpu->aspace) {
+
+	if (!IS_ERR_OR_NULL(gpu->aspace)) {
 		gpu->aspace->mmu->funcs->detach(gpu->aspace->mmu,
 			NULL, 0);
 		msm_gem_address_space_put(gpu->aspace);
diff --git a/drivers/gpu/drm/msm/msm_rd.c b/drivers/gpu/drm/msm/msm_rd.c
index 0366b8092f97..ec56794ad039 100644
--- a/drivers/gpu/drm/msm/msm_rd.c
+++ b/drivers/gpu/drm/msm/msm_rd.c
@@ -111,10 +111,14 @@ static void rd_write(struct msm_rd_state *rd, const void *buf, int sz)
 
 		wait_event(rd->fifo_event, circ_space(&rd->fifo) > 0);
 
+		/* Note that smp_load_acquire() is not strictly required
+		 * as CIRC_SPACE_TO_END() does not access the tail more
+		 * than once.
+		 */
 		n = min(sz, circ_space_to_end(&rd->fifo));
 		memcpy(fptr, ptr, n);
 
-		fifo->head = (fifo->head + n) & (BUF_SZ - 1);
+		smp_store_release(&fifo->head, (fifo->head + n) & (BUF_SZ - 1));
 		sz  -= n;
 		ptr += n;
 
@@ -145,13 +149,17 @@ static ssize_t rd_read(struct file *file, char __user *buf,
 	if (ret)
 		goto out;
 
+	/* Note that smp_load_acquire() is not strictly required
+	 * as CIRC_CNT_TO_END() does not access the head more than
+	 * once.
+	 */
 	n = min_t(int, sz, circ_count_to_end(&rd->fifo));
 	if (copy_to_user(buf, fptr, n)) {
 		ret = -EFAULT;
 		goto out;
 	}
 
-	fifo->tail = (fifo->tail + n) & (BUF_SZ - 1);
+	smp_store_release(&fifo->tail, (fifo->tail + n) & (BUF_SZ - 1));
 	*ppos += n;
 
 	wake_up_all(&rd->fifo_event);
diff --git a/drivers/gpu/ipu-v3/ipu-common.c b/drivers/gpu/ipu-v3/ipu-common.c
index 6a573d21d3cc..658fa2d3e40c 100644
--- a/drivers/gpu/ipu-v3/ipu-common.c
+++ b/drivers/gpu/ipu-v3/ipu-common.c
@@ -405,6 +405,14 @@ int ipu_idmac_lock_enable(struct ipuv3_channel *channel, int num_bursts)
 		return -EINVAL;
 	}
 
+	/*
+	 * IPUv3EX / i.MX51 has a different register layout, and on IPUv3M /
+	 * i.MX53 channel arbitration locking doesn't seem to work properly.
+	 * Allow enabling the lock feature on IPUv3H / i.MX6 only.
+	 */
+	if (bursts && ipu->ipu_type != IPUV3H)
+		return -EINVAL;
+
 	for (i = 0; i < ARRAY_SIZE(idmac_lock_en_info); i++) {
 		if (channel->num == idmac_lock_en_info[i].chnum)
 			break;
diff --git a/drivers/gpu/ipu-v3/ipu-pre.c b/drivers/gpu/ipu-v3/ipu-pre.c
index c35f74c83065..c860a7997cb5 100644
--- a/drivers/gpu/ipu-v3/ipu-pre.c
+++ b/drivers/gpu/ipu-v3/ipu-pre.c
@@ -73,6 +73,14 @@
 #define  IPU_PRE_STORE_ENG_CTRL_WR_NUM_BYTES(v)		((v & 0x7) << 1)
 #define  IPU_PRE_STORE_ENG_CTRL_OUTPUT_ACTIVE_BPP(v)	((v & 0x3) << 4)
 
+#define IPU_PRE_STORE_ENG_STATUS			0x120
+#define  IPU_PRE_STORE_ENG_STATUS_STORE_BLOCK_X_MASK	0xffff
+#define  IPU_PRE_STORE_ENG_STATUS_STORE_BLOCK_X_SHIFT	0
+#define  IPU_PRE_STORE_ENG_STATUS_STORE_BLOCK_Y_MASK	0x3fff
+#define  IPU_PRE_STORE_ENG_STATUS_STORE_BLOCK_Y_SHIFT	16
+#define  IPU_PRE_STORE_ENG_STATUS_STORE_FIFO_FULL	(1 << 30)
+#define  IPU_PRE_STORE_ENG_STATUS_STORE_FIELD		(1 << 31)
+
 #define IPU_PRE_STORE_ENG_SIZE				0x130
 #define  IPU_PRE_STORE_ENG_SIZE_INPUT_WIDTH(v)		((v & 0xffff) << 0)
 #define  IPU_PRE_STORE_ENG_SIZE_INPUT_HEIGHT(v)		((v & 0xffff) << 16)
@@ -93,6 +101,7 @@ struct ipu_pre {
 	dma_addr_t		buffer_paddr;
 	void			*buffer_virt;
 	bool			in_use;
+	unsigned int		safe_window_end;
 };
 
 static DEFINE_MUTEX(ipu_pre_list_mutex);
@@ -160,6 +169,9 @@ void ipu_pre_configure(struct ipu_pre *pre, unsigned int width,
 	u32 active_bpp = info->cpp[0] >> 1;
 	u32 val;
 
+	/* calculate safe window for ctrl register updates */
+	pre->safe_window_end = height - 2;
+
 	writel(bufaddr, pre->regs + IPU_PRE_CUR_BUF);
 	writel(bufaddr, pre->regs + IPU_PRE_NEXT_BUF);
 
@@ -199,7 +211,24 @@ void ipu_pre_configure(struct ipu_pre *pre, unsigned int width,
 
 void ipu_pre_update(struct ipu_pre *pre, unsigned int bufaddr)
 {
+	unsigned long timeout = jiffies + msecs_to_jiffies(5);
+	unsigned short current_yblock;
+	u32 val;
+
 	writel(bufaddr, pre->regs + IPU_PRE_NEXT_BUF);
+
+	do {
+		if (time_after(jiffies, timeout)) {
+			dev_warn(pre->dev, "timeout waiting for PRE safe window\n");
+			return;
+		}
+
+		val = readl(pre->regs + IPU_PRE_STORE_ENG_STATUS);
+		current_yblock =
+			(val >> IPU_PRE_STORE_ENG_STATUS_STORE_BLOCK_Y_SHIFT) &
+			IPU_PRE_STORE_ENG_STATUS_STORE_BLOCK_Y_MASK;
+	} while (current_yblock == 0 || current_yblock >= pre->safe_window_end);
+
 	writel(IPU_PRE_CTRL_SDW_UPDATE, pre->regs + IPU_PRE_CTRL_SET);
 }
 
diff --git a/drivers/gpu/ipu-v3/ipu-prg.c b/drivers/gpu/ipu-v3/ipu-prg.c
index ecc9ea44dc50..0013ca9f72c8 100644
--- a/drivers/gpu/ipu-v3/ipu-prg.c
+++ b/drivers/gpu/ipu-v3/ipu-prg.c
@@ -14,6 +14,7 @@
 #include <drm/drm_fourcc.h>
 #include <linux/clk.h>
 #include <linux/err.h>
+#include <linux/iopoll.h>
 #include <linux/mfd/syscon.h>
 #include <linux/mfd/syscon/imx6q-iomuxc-gpr.h>
 #include <linux/module.h>
@@ -329,6 +330,12 @@ int ipu_prg_channel_configure(struct ipuv3_channel *ipu_chan,
 	val = IPU_PRG_REG_UPDATE_REG_UPDATE;
 	writel(val, prg->regs + IPU_PRG_REG_UPDATE);
 
+	/* wait for both double buffers to be filled */
+	readl_poll_timeout(prg->regs + IPU_PRG_STATUS, val,
+			   (val & IPU_PRG_STATUS_BUFFER0_READY(prg_chan)) &&
+			   (val & IPU_PRG_STATUS_BUFFER1_READY(prg_chan)),
+			   5, 1000);
+
 	clk_disable_unprepare(prg->clk_ipg);
 
 	chan->enabled = true;