17 files changed, 328 insertions, 87 deletions

diff --git a/drivers/gpu/drm/i915/gvt/aperture_gm.c b/drivers/gpu/drm/i915/gvt/aperture_gm.c
index fe754022e356..359d37d5c958 100644
--- a/drivers/gpu/drm/i915/gvt/aperture_gm.c
+++ b/drivers/gpu/drm/i915/gvt/aperture_gm.c
@@ -61,10 +61,12 @@ static int alloc_gm(struct intel_vgpu *vgpu, bool high_gm)
 	}
 
 	mutex_lock(&dev_priv->drm.struct_mutex);
+	mmio_hw_access_pre(dev_priv);
 	ret = i915_gem_gtt_insert(&dev_priv->ggtt.vm, node,
 				  size, I915_GTT_PAGE_SIZE,
 				  I915_COLOR_UNEVICTABLE,
 				  start, end, flags);
+	mmio_hw_access_post(dev_priv);
 	mutex_unlock(&dev_priv->drm.struct_mutex);
 	if (ret)
 		gvt_err("fail to alloc %s gm space from host\n",
diff --git a/drivers/gpu/drm/i915/gvt/gtt.c b/drivers/gpu/drm/i915/gvt/gtt.c
index 58e166effa45..c7103dd2d8d5 100644
--- a/drivers/gpu/drm/i915/gvt/gtt.c
+++ b/drivers/gpu/drm/i915/gvt/gtt.c
@@ -2447,10 +2447,11 @@ static void intel_vgpu_destroy_all_ppgtt_mm(struct intel_vgpu *vgpu)
 
 static void intel_vgpu_destroy_ggtt_mm(struct intel_vgpu *vgpu)
 {
-	struct intel_gvt_partial_pte *pos;
+	struct intel_gvt_partial_pte *pos, *next;
 
-	list_for_each_entry(pos,
-			&vgpu->gtt.ggtt_mm->ggtt_mm.partial_pte_list, list) {
+	list_for_each_entry_safe(pos, next,
+				 &vgpu->gtt.ggtt_mm->ggtt_mm.partial_pte_list,
+				 list) {
 		gvt_dbg_mm("partial PTE update on hold 0x%lx : 0x%llx\n",
 			pos->offset, pos->data);
 		kfree(pos);
diff --git a/drivers/gpu/drm/i915/gvt/mmio_context.c b/drivers/gpu/drm/i915/gvt/mmio_context.c
index 36a5147cd01e..d6e02c15ef97 100644
--- a/drivers/gpu/drm/i915/gvt/mmio_context.c
+++ b/drivers/gpu/drm/i915/gvt/mmio_context.c
@@ -158,6 +158,8 @@ static void load_render_mocs(struct drm_i915_private *dev_priv)
 	int ring_id, i;
 
 	for (ring_id = 0; ring_id < ARRAY_SIZE(regs); ring_id++) {
+		if (!HAS_ENGINE(dev_priv, ring_id))
+			continue;
 		offset.reg = regs[ring_id];
 		for (i = 0; i < GEN9_MOCS_SIZE; i++) {
 			gen9_render_mocs.control_table[ring_id][i] =
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 1aaccbe7e1de..d4fac09095f8 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -1268,7 +1268,7 @@ relocate_entry(struct i915_vma *vma,
 		else if (gen >= 4)
 			len = 4;
 		else
-			len = 3;
+			len = 6;
 
 		batch = reloc_gpu(eb, vma, len);
 		if (IS_ERR(batch))
@@ -1309,6 +1309,11 @@ relocate_entry(struct i915_vma *vma,
 			*batch++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL;
 			*batch++ = addr;
 			*batch++ = target_offset;
+
+			/* And again for good measure (blb/pnv) */
+			*batch++ = MI_STORE_DWORD_IMM | MI_MEM_VIRTUAL;
+			*batch++ = addr;
+			*batch++ = target_offset;
 		}
 
 		goto out;
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 47c302543799..07999fe09ad2 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -3413,6 +3413,11 @@ static int gen8_gmch_probe(struct i915_ggtt *ggtt)
 		ggtt->vm.insert_page    = bxt_vtd_ggtt_insert_page__BKL;
 		if (ggtt->vm.clear_range != nop_clear_range)
 			ggtt->vm.clear_range = bxt_vtd_ggtt_clear_range__BKL;
+
+		/* Prevent recursively calling stop_machine() and deadlocks. */
+		dev_info(dev_priv->drm.dev,
+			 "Disabling error capture for VT-d workaround\n");
+		i915_disable_error_state(dev_priv, -ENODEV);
 	}
 
 	ggtt->invalidate = gen6_ggtt_invalidate;
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
index 8762d17b6659..3eb33e000d6f 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.c
+++ b/drivers/gpu/drm/i915/i915_gpu_error.c
@@ -648,6 +648,9 @@ int i915_error_state_to_str(struct drm_i915_error_state_buf *m,
 		return 0;
 	}
 
+	if (IS_ERR(error))
+		return PTR_ERR(error);
+
 	if (*error->error_msg)
 		err_printf(m, "%s\n", error->error_msg);
 	err_printf(m, "Kernel: " UTS_RELEASE "\n");
@@ -1859,6 +1862,7 @@ void i915_capture_error_state(struct drm_i915_private *i915,
 	error = i915_capture_gpu_state(i915);
 	if (!error) {
 		DRM_DEBUG_DRIVER("out of memory, not capturing error state\n");
+		i915_disable_error_state(i915, -ENOMEM);
 		return;
 	}
 
@@ -1914,5 +1918,14 @@ void i915_reset_error_state(struct drm_i915_private *i915)
 	i915->gpu_error.first_error = NULL;
 	spin_unlock_irq(&i915->gpu_error.lock);
 
-	i915_gpu_state_put(error);
+	if (!IS_ERR(error))
+		i915_gpu_state_put(error);
+}
+
+void i915_disable_error_state(struct drm_i915_private *i915, int err)
+{
+	spin_lock_irq(&i915->gpu_error.lock);
+	if (!i915->gpu_error.first_error)
+		i915->gpu_error.first_error = ERR_PTR(err);
+	spin_unlock_irq(&i915->gpu_error.lock);
 }
diff --git a/drivers/gpu/drm/i915/i915_gpu_error.h b/drivers/gpu/drm/i915/i915_gpu_error.h
index 8710fb18ed74..3ec89a504de5 100644
--- a/drivers/gpu/drm/i915/i915_gpu_error.h
+++ b/drivers/gpu/drm/i915/i915_gpu_error.h
@@ -343,6 +343,7 @@ static inline void i915_gpu_state_put(struct i915_gpu_state *gpu)
 
 struct i915_gpu_state *i915_first_error_state(struct drm_i915_private *i915);
 void i915_reset_error_state(struct drm_i915_private *i915);
+void i915_disable_error_state(struct drm_i915_private *i915, int err);
 
 #else
 
@@ -355,13 +356,18 @@ static inline void i915_capture_error_state(struct drm_i915_private *dev_priv,
 static inline struct i915_gpu_state *
 i915_first_error_state(struct drm_i915_private *i915)
 {
-	return NULL;
+	return ERR_PTR(-ENODEV);
 }
 
 static inline void i915_reset_error_state(struct drm_i915_private *i915)
 {
 }
 
+static inline void i915_disable_error_state(struct drm_i915_private *i915,
+					    int err)
+{
+}
+
 #endif /* IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR) */
 
 #endif /* _I915_GPU_ERROR_H_ */
diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c
index 0ef0c6448d53..01fa98299bae 100644
--- a/drivers/gpu/drm/i915/intel_device_info.c
+++ b/drivers/gpu/drm/i915/intel_device_info.c
@@ -474,7 +474,7 @@ static void broadwell_sseu_info_init(struct drm_i915_private *dev_priv)
 			u8 eu_disabled_mask;
 			u32 n_disabled;
 
-			if (!(sseu->subslice_mask[ss] & BIT(ss)))
+			if (!(sseu->subslice_mask[s] & BIT(ss)))
 				/* skip disabled subslice */
 				continue;
 
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 23d8008a93bb..c9878dd1f7cd 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -2890,6 +2890,7 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc,
 	return;
 
 valid_fb:
+	intel_state->base.rotation = plane_config->rotation;
 	intel_fill_fb_ggtt_view(&intel_state->view, fb,
 				intel_state->base.rotation);
 	intel_state->color_plane[0].stride =
@@ -4850,8 +4851,31 @@ static void cpt_verify_modeset(struct drm_device *dev, int pipe)
  * chroma samples for both of the luma samples, and thus we don't
  * actually get the expected MPEG2 chroma siting convention :(
  * The same behaviour is observed on pre-SKL platforms as well.
+ *
+ * Theory behind the formula (note that we ignore sub-pixel
+ * source coordinates):
+ * s = source sample position
+ * d = destination sample position
+ *
+ * Downscaling 4:1:
+ * -0.5
+ * | 0.0
+ * | |     1.5 (initial phase)
+ * | |     |
+ * v v     v
+ * | s | s | s | s |
+ * |       d       |
+ *
+ * Upscaling 1:4:
+ * -0.5
+ * | -0.375 (initial phase)
+ * | |     0.0
+ * | |     |
+ * v v     v
+ * |       s       |
+ * | d | d | d | d |
  */
-u16 skl_scaler_calc_phase(int sub, bool chroma_cosited)
+u16 skl_scaler_calc_phase(int sub, int scale, bool chroma_cosited)
 {
 	int phase = -0x8000;
 	u16 trip = 0;
@@ -4859,6 +4883,15 @@ u16 skl_scaler_calc_phase(int sub, bool chroma_cosited)
 	if (chroma_cosited)
 		phase += (sub - 1) * 0x8000 / sub;
 
+	phase += scale / (2 * sub);
+
+	/*
+	 * Hardware initial phase limited to [-0.5:1.5].
+	 * Since the max hardware scale factor is 3.0, we
+	 * should never actually excdeed 1.0 here.
+	 */
+	WARN_ON(phase < -0x8000 || phase > 0x18000);
+
 	if (phase < 0)
 		phase = 0x10000 + phase;
 	else
@@ -5067,13 +5100,20 @@ static void skylake_pfit_enable(struct intel_crtc *crtc)
 
 	if (crtc->config->pch_pfit.enabled) {
 		u16 uv_rgb_hphase, uv_rgb_vphase;
+		int pfit_w, pfit_h, hscale, vscale;
 		int id;
 
 		if (WARN_ON(crtc->config->scaler_state.scaler_id < 0))
 			return;
 
-		uv_rgb_hphase = skl_scaler_calc_phase(1, false);
-		uv_rgb_vphase = skl_scaler_calc_phase(1, false);
+		pfit_w = (crtc->config->pch_pfit.size >> 16) & 0xFFFF;
+		pfit_h = crtc->config->pch_pfit.size & 0xFFFF;
+
+		hscale = (crtc->config->pipe_src_w << 16) / pfit_w;
+		vscale = (crtc->config->pipe_src_h << 16) / pfit_h;
+
+		uv_rgb_hphase = skl_scaler_calc_phase(1, hscale, false);
+		uv_rgb_vphase = skl_scaler_calc_phase(1, vscale, false);
 
 		id = scaler_state->scaler_id;
 		I915_WRITE(SKL_PS_CTRL(pipe, id), PS_SCALER_EN |
@@ -7843,8 +7883,15 @@ i9xx_get_initial_plane_config(struct intel_crtc *crtc,
 			plane_config->tiling = I915_TILING_X;
 			fb->modifier = I915_FORMAT_MOD_X_TILED;
 		}
+
+		if (val & DISPPLANE_ROTATE_180)
+			plane_config->rotation = DRM_MODE_ROTATE_180;
 	}
 
+	if (IS_CHERRYVIEW(dev_priv) && pipe == PIPE_B &&
+	    val & DISPPLANE_MIRROR)
+		plane_config->rotation |= DRM_MODE_REFLECT_X;
+
 	pixel_format = val & DISPPLANE_PIXFORMAT_MASK;
 	fourcc = i9xx_format_to_fourcc(pixel_format);
 	fb->format = drm_format_info(fourcc);
@@ -8913,6 +8960,29 @@ skylake_get_initial_plane_config(struct intel_crtc *crtc,
 		goto error;
 	}
 
+	/*
+	 * DRM_MODE_ROTATE_ is counter clockwise to stay compatible with Xrandr
+	 * while i915 HW rotation is clockwise, thats why this swapping.
+	 */
+	switch (val & PLANE_CTL_ROTATE_MASK) {
+	case PLANE_CTL_ROTATE_0:
+		plane_config->rotation = DRM_MODE_ROTATE_0;
+		break;
+	case PLANE_CTL_ROTATE_90:
+		plane_config->rotation = DRM_MODE_ROTATE_270;
+		break;
+	case PLANE_CTL_ROTATE_180:
+		plane_config->rotation = DRM_MODE_ROTATE_180;
+		break;
+	case PLANE_CTL_ROTATE_270:
+		plane_config->rotation = DRM_MODE_ROTATE_90;
+		break;
+	}
+
+	if (INTEL_GEN(dev_priv) >= 10 &&
+	    val & PLANE_CTL_FLIP_HORIZONTAL)
+		plane_config->rotation |= DRM_MODE_REFLECT_X;
+
 	base = I915_READ(PLANE_SURF(pipe, plane_id)) & 0xfffff000;
 	plane_config->base = base;
 
@@ -15228,6 +15298,14 @@ retry:
 			ret = drm_atomic_add_affected_planes(state, crtc);
 			if (ret)
 				goto out;
+
+			/*
+			 * FIXME hack to force a LUT update to avoid the
+			 * plane update forcing the pipe gamma on without
+			 * having a proper LUT loaded. Remove once we
+			 * have readout for pipe gamma enable.
+			 */
+			crtc_state->color_mgmt_changed = true;
 		}
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c
index 1b00f8ea145b..a911691dbd0f 100644
--- a/drivers/gpu/drm/i915/intel_dp_mst.c
+++ b/drivers/gpu/drm/i915/intel_dp_mst.c
@@ -452,6 +452,10 @@ static struct drm_connector *intel_dp_add_mst_connector(struct drm_dp_mst_topolo
 	if (!intel_connector)
 		return NULL;
 
+	intel_connector->get_hw_state = intel_dp_mst_get_hw_state;
+	intel_connector->mst_port = intel_dp;
+	intel_connector->port = port;
+
 	connector = &intel_connector->base;
 	ret = drm_connector_init(dev, connector, &intel_dp_mst_connector_funcs,
 				 DRM_MODE_CONNECTOR_DisplayPort);
@@ -462,10 +466,6 @@ static struct drm_connector *intel_dp_add_mst_connector(struct drm_dp_mst_topolo
 
 	drm_connector_helper_add(connector, &intel_dp_mst_connector_helper_funcs);
 
-	intel_connector->get_hw_state = intel_dp_mst_get_hw_state;
-	intel_connector->mst_port = intel_dp;
-	intel_connector->port = port;
-
 	for_each_pipe(dev_priv, pipe) {
 		struct drm_encoder *enc =
 			&intel_dp->mst_encoders[pipe]->base.base;
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index f8dc84b2d2d3..db6fa1d0cbda 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -547,6 +547,7 @@ struct intel_initial_plane_config {
 	unsigned int tiling;
 	int size;
 	u32 base;
+	u8 rotation;
 };
 
 #define SKL_MIN_SRC_W 8
@@ -1646,7 +1647,7 @@ void intel_mode_from_pipe_config(struct drm_display_mode *mode,
 void intel_crtc_arm_fifo_underrun(struct intel_crtc *crtc,
 				  struct intel_crtc_state *crtc_state);
 
-u16 skl_scaler_calc_phase(int sub, bool chroma_center);
+u16 skl_scaler_calc_phase(int sub, int scale, bool chroma_center);
 int skl_update_scaler_crtc(struct intel_crtc_state *crtc_state);
 int skl_max_scale(const struct intel_crtc_state *crtc_state,
 		  u32 pixel_format);
diff --git a/drivers/gpu/drm/i915/intel_hotplug.c b/drivers/gpu/drm/i915/intel_hotplug.c
index 648a13c6043c..9a8018130237 100644
--- a/drivers/gpu/drm/i915/intel_hotplug.c
+++ b/drivers/gpu/drm/i915/intel_hotplug.c
@@ -228,7 +228,9 @@ static void intel_hpd_irq_storm_reenable_work(struct work_struct *work)
 		drm_for_each_connector_iter(connector, &conn_iter) {
 			struct intel_connector *intel_connector = to_intel_connector(connector);
 
-			if (intel_connector->encoder->hpd_pin == pin) {
+			/* Don't check MST ports, they don't have pins */
+			if (!intel_connector->mst_port &&
+			    intel_connector->encoder->hpd_pin == pin) {
 				if (connector->polled != intel_connector->polled)
 					DRM_DEBUG_DRIVER("Reenabling HPD on connector %s\n",
 							 connector->name);
@@ -395,37 +397,54 @@ void intel_hpd_irq_handler(struct drm_i915_private *dev_priv,
 	struct intel_encoder *encoder;
 	bool storm_detected = false;
 	bool queue_dig = false, queue_hp = false;
+	u32 long_hpd_pulse_mask = 0;
+	u32 short_hpd_pulse_mask = 0;
+	enum hpd_pin pin;
 
 	if (!pin_mask)
 		return;
 
 	spin_lock(&dev_priv->irq_lock);
+
+	/*
+	 * Determine whether ->hpd_pulse() exists for each pin, and
+	 * whether we have a short or a long pulse. This is needed
+	 * as each pin may have up to two encoders (HDMI and DP) and
+	 * only the one of them (DP) will have ->hpd_pulse().
+	 */
 	for_each_intel_encoder(&dev_priv->drm, encoder) {
-		enum hpd_pin pin = encoder->hpd_pin;
 		bool has_hpd_pulse = intel_encoder_has_hpd_pulse(encoder);
+		enum port port = encoder->port;
+		bool long_hpd;
 
+		pin = encoder->hpd_pin;
 		if (!(BIT(pin) & pin_mask))
 			continue;
 
-		if (has_hpd_pulse) {
-			bool long_hpd = long_mask & BIT(pin);
-			enum port port = encoder->port;
+		if (!has_hpd_pulse)
+			continue;
 
-			DRM_DEBUG_DRIVER("digital hpd port %c - %s\n", port_name(port),
-					 long_hpd ? "long" : "short");
-			/*
-			 * For long HPD pulses we want to have the digital queue happen,
-			 * but we still want HPD storm detection to function.
-			 */
-			queue_dig = true;
-			if (long_hpd) {
-				dev_priv->hotplug.long_port_mask |= (1 << port);
-			} else {
-				/* for short HPD just trigger the digital queue */
-				dev_priv->hotplug.short_port_mask |= (1 << port);
-				continue;
-			}
+		long_hpd = long_mask & BIT(pin);
+
+		DRM_DEBUG_DRIVER("digital hpd port %c - %s\n", port_name(port),
+				 long_hpd ? "long" : "short");
+		queue_dig = true;
+
+		if (long_hpd) {
+			long_hpd_pulse_mask |= BIT(pin);
+			dev_priv->hotplug.long_port_mask |= BIT(port);
+		} else {
+			short_hpd_pulse_mask |= BIT(pin);
+			dev_priv->hotplug.short_port_mask |= BIT(port);
 		}
+	}
+
+	/* Now process each pin just once */
+	for_each_hpd_pin(pin) {
+		bool long_hpd;
+
+		if (!(BIT(pin) & pin_mask))
+			continue;
 
 		if (dev_priv->hotplug.stats[pin].state == HPD_DISABLED) {
 			/*
@@ -442,11 +461,22 @@ void intel_hpd_irq_handler(struct drm_i915_private *dev_priv,
 		if (dev_priv->hotplug.stats[pin].state != HPD_ENABLED)
 			continue;
 
-		if (!has_hpd_pulse) {
+		/*
+		 * Delegate to ->hpd_pulse() if one of the encoders for this
+		 * pin has it, otherwise let the hotplug_work deal with this
+		 * pin directly.
+		 */
+		if (((short_hpd_pulse_mask | long_hpd_pulse_mask) & BIT(pin))) {
+			long_hpd = long_hpd_pulse_mask & BIT(pin);
+		} else {
 			dev_priv->hotplug.event_bits |= BIT(pin);
+			long_hpd = true;
 			queue_hp = true;
 		}
 
+		if (!long_hpd)
+			continue;
+
 		if (intel_hpd_irq_storm_detect(dev_priv, pin)) {
 			dev_priv->hotplug.event_bits &= ~BIT(pin);
 			storm_detected = true;
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index 43957bb37a42..37c94a54efcb 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -424,7 +424,8 @@ static u64 execlists_update_context(struct i915_request *rq)
 
 	reg_state[CTX_RING_TAIL+1] = intel_ring_set_tail(rq->ring, rq->tail);
 
-	/* True 32b PPGTT with dynamic page allocation: update PDP
+	/*
+	 * True 32b PPGTT with dynamic page allocation: update PDP
 	 * registers and point the unallocated PDPs to scratch page.
 	 * PML4 is allocated during ppgtt init, so this is not needed
 	 * in 48-bit mode.
@@ -432,6 +433,17 @@ static u64 execlists_update_context(struct i915_request *rq)
 	if (ppgtt && !i915_vm_is_48bit(&ppgtt->vm))
 		execlists_update_context_pdps(ppgtt, reg_state);
 
+	/*
+	 * Make sure the context image is complete before we submit it to HW.
+	 *
+	 * Ostensibly, writes (including the WCB) should be flushed prior to
+	 * an uncached write such as our mmio register access, the empirical
+	 * evidence (esp. on Braswell) suggests that the WC write into memory
+	 * may not be visible to the HW prior to the completion of the UC
+	 * register write and that we may begin execution from the context
+	 * before its image is complete leading to invalid PD chasing.
+	 */
+	wmb();
 	return ce->lrc_desc;
 }
 
diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
index 245f0022bcfd..3fe358db1276 100644
--- a/drivers/gpu/drm/i915/intel_pm.c
+++ b/drivers/gpu/drm/i915/intel_pm.c
@@ -2493,6 +2493,9 @@ static uint32_t ilk_compute_pri_wm(const struct intel_crtc_state *cstate,
 	uint32_t method1, method2;
 	int cpp;
 
+	if (mem_value == 0)
+		return U32_MAX;
+
 	if (!intel_wm_plane_visible(cstate, pstate))
 		return 0;
 
@@ -2522,6 +2525,9 @@ static uint32_t ilk_compute_spr_wm(const struct intel_crtc_state *cstate,
 	uint32_t method1, method2;
 	int cpp;
 
+	if (mem_value == 0)
+		return U32_MAX;
+
 	if (!intel_wm_plane_visible(cstate, pstate))
 		return 0;
 
@@ -2545,6 +2551,9 @@ static uint32_t ilk_compute_cur_wm(const struct intel_crtc_state *cstate,
 {
 	int cpp;
 
+	if (mem_value == 0)
+		return U32_MAX;
+
 	if (!intel_wm_plane_visible(cstate, pstate))
 		return 0;
 
@@ -3008,6 +3017,34 @@ static void snb_wm_latency_quirk(struct drm_i915_private *dev_priv)
 	intel_print_wm_latency(dev_priv, "Cursor", dev_priv->wm.cur_latency);
 }
 
+static void snb_wm_lp3_irq_quirk(struct drm_i915_private *dev_priv)
+{
+	/*
+	 * On some SNB machines (Thinkpad X220 Tablet at least)
+	 * LP3 usage can cause vblank interrupts to be lost.
+	 * The DEIIR bit will go high but it looks like the CPU
+	 * never gets interrupted.
+	 *
+	 * It's not clear whether other interrupt source could
+	 * be affected or if this is somehow limited to vblank
+	 * interrupts only. To play it safe we disable LP3
+	 * watermarks entirely.
+	 */
+	if (dev_priv->wm.pri_latency[3] == 0 &&
+	    dev_priv->wm.spr_latency[3] == 0 &&
+	    dev_priv->wm.cur_latency[3] == 0)
+		return;
+
+	dev_priv->wm.pri_latency[3] = 0;
+	dev_priv->wm.spr_latency[3] = 0;
+	dev_priv->wm.cur_latency[3] = 0;
+
+	DRM_DEBUG_KMS("LP3 watermarks disabled due to potential for lost interrupts\n");
+	intel_print_wm_latency(dev_priv, "Primary", dev_priv->wm.pri_latency);
+	intel_print_wm_latency(dev_priv, "Sprite", dev_priv->wm.spr_latency);
+	intel_print_wm_latency(dev_priv, "Cursor", dev_priv->wm.cur_latency);
+}
+
 static void ilk_setup_wm_latency(struct drm_i915_private *dev_priv)
 {
 	intel_read_wm_latency(dev_priv, dev_priv->wm.pri_latency);
@@ -3024,8 +3061,10 @@ static void ilk_setup_wm_latency(struct drm_i915_private *dev_priv)
 	intel_print_wm_latency(dev_priv, "Sprite", dev_priv->wm.spr_latency);
 	intel_print_wm_latency(dev_priv, "Cursor", dev_priv->wm.cur_latency);
 
-	if (IS_GEN6(dev_priv))
+	if (IS_GEN6(dev_priv)) {
 		snb_wm_latency_quirk(dev_priv);
+		snb_wm_lp3_irq_quirk(dev_priv);
+	}
 }
 
 static void skl_setup_wm_latency(struct drm_i915_private *dev_priv)
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index d0ef50bf930a..187bb0ceb4ac 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -91,6 +91,7 @@ static int
 gen4_render_ring_flush(struct i915_request *rq, u32 mode)
 {
 	u32 cmd, *cs;
+	int i;
 
 	/*
 	 * read/write caches:
@@ -127,12 +128,45 @@ gen4_render_ring_flush(struct i915_request *rq, u32 mode)
 			cmd |= MI_INVALIDATE_ISP;
 	}
 
-	cs = intel_ring_begin(rq, 2);
+	i = 2;
+	if (mode & EMIT_INVALIDATE)
+		i += 20;
+
+	cs = intel_ring_begin(rq, i);
 	if (IS_ERR(cs))
 		return PTR_ERR(cs);
 
 	*cs++ = cmd;
-	*cs++ = MI_NOOP;
+
+	/*
+	 * A random delay to let the CS invalidate take effect? Without this
+	 * delay, the GPU relocation path fails as the CS does not see
+	 * the updated contents. Just as important, if we apply the flushes
+	 * to the EMIT_FLUSH branch (i.e. immediately after the relocation
+	 * write and before the invalidate on the next batch), the relocations
+	 * still fail. This implies that is a delay following invalidation
+	 * that is required to reset the caches as opposed to a delay to
+	 * ensure the memory is written.
+	 */
+	if (mode & EMIT_INVALIDATE) {
+		*cs++ = GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE;
+		*cs++ = i915_ggtt_offset(rq->engine->scratch) |
+			PIPE_CONTROL_GLOBAL_GTT;
+		*cs++ = 0;
+		*cs++ = 0;
+
+		for (i = 0; i < 12; i++)
+			*cs++ = MI_FLUSH;
+
+		*cs++ = GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE;
+		*cs++ = i915_ggtt_offset(rq->engine->scratch) |
+			PIPE_CONTROL_GLOBAL_GTT;
+		*cs++ = 0;
+		*cs++ = 0;
+	}
+
+	*cs++ = cmd;
+
 	intel_ring_advance(rq, cs);
 
 	return 0;
diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
index 0fdabce647ab..44e4491a4918 100644
--- a/drivers/gpu/drm/i915/intel_runtime_pm.c
+++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
@@ -2749,6 +2749,12 @@ static const struct i915_power_well_desc icl_power_wells[] = {
 		},
 	},
 	{
+		.name = "DC off",
+		.domains = ICL_DISPLAY_DC_OFF_POWER_DOMAINS,
+		.ops = &gen9_dc_off_power_well_ops,
+		.id = DISP_PW_ID_NONE,
+	},
+	{
 		.name = "power well 2",
 		.domains = ICL_PW_2_POWER_DOMAINS,
 		.ops = &hsw_power_well_ops,
@@ -2760,12 +2766,6 @@ static const struct i915_power_well_desc icl_power_wells[] = {
 		},
 	},
 	{
-		.name = "DC off",
-		.domains = ICL_DISPLAY_DC_OFF_POWER_DOMAINS,
-		.ops = &gen9_dc_off_power_well_ops,
-		.id = DISP_PW_ID_NONE,
-	},
-	{
 		.name = "power well 3",
 		.domains = ICL_PW_3_POWER_DOMAINS,
 		.ops = &hsw_power_well_ops,
@@ -3176,8 +3176,7 @@ static u8 intel_dbuf_max_slices(struct drm_i915_private *dev_priv)
 void icl_dbuf_slices_update(struct drm_i915_private *dev_priv,
 			    u8 req_slices)
 {
-	u8 hw_enabled_slices = dev_priv->wm.skl_hw.ddb.enabled_slices;
-	u32 val;
+	const u8 hw_enabled_slices = dev_priv->wm.skl_hw.ddb.enabled_slices;
 	bool ret;
 
 	if (req_slices > intel_dbuf_max_slices(dev_priv)) {
@@ -3188,7 +3187,6 @@ void icl_dbuf_slices_update(struct drm_i915_private *dev_priv,
 	if (req_slices == hw_enabled_slices || req_slices == 0)
 		return;
 
-	val = I915_READ(DBUF_CTL_S2);
 	if (req_slices > hw_enabled_slices)
 		ret = intel_dbuf_slice_set(dev_priv, DBUF_CTL_S2, true);
 	else
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
index 5fd2f7bf3927..d3090a7537bb 100644
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ b/drivers/gpu/drm/i915/intel_sprite.c
@@ -302,13 +302,65 @@ skl_plane_max_stride(struct intel_plane *plane,
 		return min(8192 * cpp, 32768);
 }
 
+static void
+skl_program_scaler(struct intel_plane *plane,
+		   const struct intel_crtc_state *crtc_state,
+		   const struct intel_plane_state *plane_state)
+{
+	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
+	enum pipe pipe = plane->pipe;
+	int scaler_id = plane_state->scaler_id;
+	const struct intel_scaler *scaler =
+		&crtc_state->scaler_state.scalers[scaler_id];
+	int crtc_x = plane_state->base.dst.x1;
+	int crtc_y = plane_state->base.dst.y1;
+	uint32_t crtc_w = drm_rect_width(&plane_state->base.dst);
+	uint32_t crtc_h = drm_rect_height(&plane_state->base.dst);
+	u16 y_hphase, uv_rgb_hphase;
+	u16 y_vphase, uv_rgb_vphase;
+	int hscale, vscale;
+
+	hscale = drm_rect_calc_hscale(&plane_state->base.src,
+				      &plane_state->base.dst,
+				      0, INT_MAX);
+	vscale = drm_rect_calc_vscale(&plane_state->base.src,
+				      &plane_state->base.dst,
+				      0, INT_MAX);
+
+	/* TODO: handle sub-pixel coordinates */
+	if (plane_state->base.fb->format->format == DRM_FORMAT_NV12) {
+		y_hphase = skl_scaler_calc_phase(1, hscale, false);
+		y_vphase = skl_scaler_calc_phase(1, vscale, false);
+
+		/* MPEG2 chroma siting convention */
+		uv_rgb_hphase = skl_scaler_calc_phase(2, hscale, true);
+		uv_rgb_vphase = skl_scaler_calc_phase(2, vscale, false);
+	} else {
+		/* not used */
+		y_hphase = 0;
+		y_vphase = 0;
+
+		uv_rgb_hphase = skl_scaler_calc_phase(1, hscale, false);
+		uv_rgb_vphase = skl_scaler_calc_phase(1, vscale, false);
+	}
+
+	I915_WRITE_FW(SKL_PS_CTRL(pipe, scaler_id),
+		      PS_SCALER_EN | PS_PLANE_SEL(plane->id) | scaler->mode);
+	I915_WRITE_FW(SKL_PS_PWR_GATE(pipe, scaler_id), 0);
+	I915_WRITE_FW(SKL_PS_VPHASE(pipe, scaler_id),
+		      PS_Y_PHASE(y_vphase) | PS_UV_RGB_PHASE(uv_rgb_vphase));
+	I915_WRITE_FW(SKL_PS_HPHASE(pipe, scaler_id),
+		      PS_Y_PHASE(y_hphase) | PS_UV_RGB_PHASE(uv_rgb_hphase));
+	I915_WRITE_FW(SKL_PS_WIN_POS(pipe, scaler_id), (crtc_x << 16) | crtc_y);
+	I915_WRITE_FW(SKL_PS_WIN_SZ(pipe, scaler_id), (crtc_w << 16) | crtc_h);
+}
+
 void
 skl_update_plane(struct intel_plane *plane,
 		 const struct intel_crtc_state *crtc_state,
 		 const struct intel_plane_state *plane_state)
 {
 	struct drm_i915_private *dev_priv = to_i915(plane->base.dev);
-	const struct drm_framebuffer *fb = plane_state->base.fb;
 	enum plane_id plane_id = plane->id;
 	enum pipe pipe = plane->pipe;
 	u32 plane_ctl = plane_state->ctl;
@@ -318,8 +370,6 @@ skl_update_plane(struct intel_plane *plane,
 	u32 aux_stride = skl_plane_stride(plane_state, 1);
 	int crtc_x = plane_state->base.dst.x1;
 	int crtc_y = plane_state->base.dst.y1;
-	uint32_t crtc_w = drm_rect_width(&plane_state->base.dst);
-	uint32_t crtc_h = drm_rect_height(&plane_state->base.dst);
 	uint32_t x = plane_state->color_plane[0].x;
 	uint32_t y = plane_state->color_plane[0].y;
 	uint32_t src_w = drm_rect_width(&plane_state->base.src) >> 16;
@@ -329,8 +379,6 @@ skl_update_plane(struct intel_plane *plane,
 	/* Sizes are 0 based */
 	src_w--;
 	src_h--;
-	crtc_w--;
-	crtc_h--;
 
 	spin_lock_irqsave(&dev_priv->uncore.lock, irqflags);
 
@@ -353,41 +401,8 @@ skl_update_plane(struct intel_plane *plane,
 		      (plane_state->color_plane[1].y << 16) |
 		      plane_state->color_plane[1].x);
 
-	/* program plane scaler */
 	if (plane_state->scaler_id >= 0) {
-		int scaler_id = plane_state->scaler_id;
-		const struct intel_scaler *scaler =
-			&crtc_state->scaler_state.scalers[scaler_id];
-		u16 y_hphase, uv_rgb_hphase;
-		u16 y_vphase, uv_rgb_vphase;
-
-		/* TODO: handle sub-pixel coordinates */
-		if (fb->format->format == DRM_FORMAT_NV12) {
-			y_hphase = skl_scaler_calc_phase(1, false);
-			y_vphase = skl_scaler_calc_phase(1, false);
-
-			/* MPEG2 chroma siting convention */
-			uv_rgb_hphase = skl_scaler_calc_phase(2, true);
-			uv_rgb_vphase = skl_scaler_calc_phase(2, false);
-		} else {
-			/* not used */
-			y_hphase = 0;
-			y_vphase = 0;
-
-			uv_rgb_hphase = skl_scaler_calc_phase(1, false);
-			uv_rgb_vphase = skl_scaler_calc_phase(1, false);
-		}
-
-		I915_WRITE_FW(SKL_PS_CTRL(pipe, scaler_id),
-			      PS_SCALER_EN | PS_PLANE_SEL(plane_id) | scaler->mode);
-		I915_WRITE_FW(SKL_PS_PWR_GATE(pipe, scaler_id), 0);
-		I915_WRITE_FW(SKL_PS_VPHASE(pipe, scaler_id),
-			      PS_Y_PHASE(y_vphase) | PS_UV_RGB_PHASE(uv_rgb_vphase));
-		I915_WRITE_FW(SKL_PS_HPHASE(pipe, scaler_id),
-			      PS_Y_PHASE(y_hphase) | PS_UV_RGB_PHASE(uv_rgb_hphase));
-		I915_WRITE_FW(SKL_PS_WIN_POS(pipe, scaler_id), (crtc_x << 16) | crtc_y);
-		I915_WRITE_FW(SKL_PS_WIN_SZ(pipe, scaler_id),
-			      ((crtc_w + 1) << 16)|(crtc_h + 1));
+		skl_program_scaler(plane, crtc_state, plane_state);
 
 		I915_WRITE_FW(PLANE_POS(pipe, plane_id), 0);
 	} else {