BackMerge v4.19-rc5 into drm-next

Sean Paul requested an -rc5 backmerge from some sun4i fixes.

Signed-off-by: Dave Airlie <airlied@redhat.com>

11 files changed, 214 insertions, 193 deletions

diff --git a/drivers/gpu/drm/i915/gvt/dmabuf.c b/drivers/gpu/drm/i915/gvt/dmabuf.c
index 6e3f56684f4e..51ed99a37803 100644
--- a/drivers/gpu/drm/i915/gvt/dmabuf.c
+++ b/drivers/gpu/drm/i915/gvt/dmabuf.c
@@ -170,20 +170,22 @@ static struct drm_i915_gem_object *vgpu_create_gem(struct drm_device *dev,
 		unsigned int tiling_mode = 0;
 		unsigned int stride = 0;
 
-		switch (info->drm_format_mod << 10) {
-		case PLANE_CTL_TILED_LINEAR:
+		switch (info->drm_format_mod) {
+		case DRM_FORMAT_MOD_LINEAR:
 			tiling_mode = I915_TILING_NONE;
 			break;
-		case PLANE_CTL_TILED_X:
+		case I915_FORMAT_MOD_X_TILED:
 			tiling_mode = I915_TILING_X;
 			stride = info->stride;
 			break;
-		case PLANE_CTL_TILED_Y:
+		case I915_FORMAT_MOD_Y_TILED:
+		case I915_FORMAT_MOD_Yf_TILED:
 			tiling_mode = I915_TILING_Y;
 			stride = info->stride;
 			break;
 		default:
-			gvt_dbg_core("not supported tiling mode\n");
+			gvt_dbg_core("invalid drm_format_mod %llx for tiling\n",
+				     info->drm_format_mod);
 		}
 		obj->tiling_and_stride = tiling_mode | stride;
 	} else {
@@ -222,9 +224,26 @@ static int vgpu_get_plane_info(struct drm_device *dev,
 		info->height = p.height;
 		info->stride = p.stride;
 		info->drm_format = p.drm_format;
-		info->drm_format_mod = p.tiled;
+
+		switch (p.tiled) {
+		case PLANE_CTL_TILED_LINEAR:
+			info->drm_format_mod = DRM_FORMAT_MOD_LINEAR;
+			break;
+		case PLANE_CTL_TILED_X:
+			info->drm_format_mod = I915_FORMAT_MOD_X_TILED;
+			break;
+		case PLANE_CTL_TILED_Y:
+			info->drm_format_mod = I915_FORMAT_MOD_Y_TILED;
+			break;
+		case PLANE_CTL_TILED_YF:
+			info->drm_format_mod = I915_FORMAT_MOD_Yf_TILED;
+			break;
+		default:
+			gvt_vgpu_err("invalid tiling mode: %x\n", p.tiled);
+		}
+
 		info->size = (((p.stride * p.height * p.bpp) / 8) +
-				(PAGE_SIZE - 1)) >> PAGE_SHIFT;
+			      (PAGE_SIZE - 1)) >> PAGE_SHIFT;
 	} else if (plane_id == DRM_PLANE_TYPE_CURSOR) {
 		ret = intel_vgpu_decode_cursor_plane(vgpu, &c);
 		if (ret)
diff --git a/drivers/gpu/drm/i915/gvt/fb_decoder.c b/drivers/gpu/drm/i915/gvt/fb_decoder.c
index face664be3e8..481896fb712a 100644
--- a/drivers/gpu/drm/i915/gvt/fb_decoder.c
+++ b/drivers/gpu/drm/i915/gvt/fb_decoder.c
@@ -220,8 +220,7 @@ int intel_vgpu_decode_primary_plane(struct intel_vgpu *vgpu,
 	if (IS_SKYLAKE(dev_priv)
 		|| IS_KABYLAKE(dev_priv)
 		|| IS_BROXTON(dev_priv)) {
-		plane->tiled = (val & PLANE_CTL_TILED_MASK) >>
-		_PLANE_CTL_TILED_SHIFT;
+		plane->tiled = val & PLANE_CTL_TILED_MASK;
 		fmt = skl_format_to_drm(
 			val & PLANE_CTL_FORMAT_MASK,
 			val & PLANE_CTL_ORDER_RGBX,
@@ -260,7 +259,7 @@ int intel_vgpu_decode_primary_plane(struct intel_vgpu *vgpu,
 		return  -EINVAL;
 	}
 
-	plane->stride = intel_vgpu_get_stride(vgpu, pipe, (plane->tiled << 10),
+	plane->stride = intel_vgpu_get_stride(vgpu, pipe, plane->tiled,
 		(IS_SKYLAKE(dev_priv)
 		|| IS_KABYLAKE(dev_priv)
 		|| IS_BROXTON(dev_priv)) ?
diff --git a/drivers/gpu/drm/i915/gvt/fb_decoder.h b/drivers/gpu/drm/i915/gvt/fb_decoder.h
index cb055f3c81a2..60c155085029 100644
--- a/drivers/gpu/drm/i915/gvt/fb_decoder.h
+++ b/drivers/gpu/drm/i915/gvt/fb_decoder.h
@@ -101,7 +101,7 @@ struct intel_gvt;
 /* color space conversion and gamma correction are not included */
 struct intel_vgpu_primary_plane_format {
 	u8	enabled;	/* plane is enabled */
-	u8	tiled;		/* X-tiled */
+	u32	tiled;		/* tiling mode: linear, X-tiled, Y tiled, etc */
 	u8	bpp;		/* bits per pixel */
 	u32	hw_format;	/* format field in the PRI_CTL register */
 	u32	drm_format;	/* format in DRM definition */
diff --git a/drivers/gpu/drm/i915/gvt/handlers.c b/drivers/gpu/drm/i915/gvt/handlers.c
index d26258786e3f..90f50f67909a 100644
--- a/drivers/gpu/drm/i915/gvt/handlers.c
+++ b/drivers/gpu/drm/i915/gvt/handlers.c
@@ -1297,6 +1297,19 @@ static int power_well_ctl_mmio_write(struct intel_vgpu *vgpu,
 	return 0;
 }
 
+static int gen9_dbuf_ctl_mmio_write(struct intel_vgpu *vgpu,
+		unsigned int offset, void *p_data, unsigned int bytes)
+{
+	write_vreg(vgpu, offset, p_data, bytes);
+
+	if (vgpu_vreg(vgpu, offset) & DBUF_POWER_REQUEST)
+		vgpu_vreg(vgpu, offset) |= DBUF_POWER_STATE;
+	else
+		vgpu_vreg(vgpu, offset) &= ~DBUF_POWER_STATE;
+
+	return 0;
+}
+
 static int fpga_dbg_mmio_write(struct intel_vgpu *vgpu,
 	unsigned int offset, void *p_data, unsigned int bytes)
 {
@@ -1526,9 +1539,15 @@ static int bxt_phy_ctl_family_write(struct intel_vgpu *vgpu,
 	u32 v = *(u32 *)p_data;
 	u32 data = v & COMMON_RESET_DIS ? BXT_PHY_LANE_ENABLED : 0;
 
-	vgpu_vreg(vgpu, _BXT_PHY_CTL_DDI_A) = data;
-	vgpu_vreg(vgpu, _BXT_PHY_CTL_DDI_B) = data;
-	vgpu_vreg(vgpu, _BXT_PHY_CTL_DDI_C) = data;
+	switch (offset) {
+	case _PHY_CTL_FAMILY_EDP:
+		vgpu_vreg(vgpu, _BXT_PHY_CTL_DDI_A) = data;
+		break;
+	case _PHY_CTL_FAMILY_DDI:
+		vgpu_vreg(vgpu, _BXT_PHY_CTL_DDI_B) = data;
+		vgpu_vreg(vgpu, _BXT_PHY_CTL_DDI_C) = data;
+		break;
+	}
 
 	vgpu_vreg(vgpu, offset) = v;
 
@@ -2801,6 +2820,8 @@ static int init_skl_mmio_info(struct intel_gvt *gvt)
 	MMIO_D(HSW_PWR_WELL_CTL1, D_SKL_PLUS);
 	MMIO_DH(HSW_PWR_WELL_CTL2, D_SKL_PLUS, NULL, skl_power_well_ctl_write);
 
+	MMIO_DH(DBUF_CTL, D_SKL_PLUS, NULL, gen9_dbuf_ctl_mmio_write);
+
 	MMIO_D(_MMIO(0xa210), D_SKL_PLUS);
 	MMIO_D(GEN9_MEDIA_PG_IDLE_HYSTERESIS, D_SKL_PLUS);
 	MMIO_D(GEN9_RENDER_PG_IDLE_HYSTERESIS, D_SKL_PLUS);
@@ -2976,8 +2997,6 @@ static int init_skl_mmio_info(struct intel_gvt *gvt)
 		NULL, gen9_trtte_write);
 	MMIO_DH(_MMIO(0x4dfc), D_SKL_PLUS, NULL, gen9_trtt_chicken_write);
 
-	MMIO_D(_MMIO(0x45008), D_SKL_PLUS);
-
 	MMIO_D(_MMIO(0x46430), D_SKL_PLUS);
 
 	MMIO_D(_MMIO(0x46520), D_SKL_PLUS);
@@ -3014,7 +3033,9 @@ static int init_skl_mmio_info(struct intel_gvt *gvt)
 	MMIO_D(_MMIO(0x44500), D_SKL_PLUS);
 	MMIO_DFH(GEN9_CSFE_CHICKEN1_RCS, D_SKL_PLUS, F_CMD_ACCESS, NULL, NULL);
 	MMIO_DFH(GEN8_HDC_CHICKEN1, D_SKL_PLUS, F_MODE_MASK | F_CMD_ACCESS,
-		NULL, NULL);
+		 NULL, NULL);
+	MMIO_DFH(GEN9_WM_CHICKEN3, D_SKL_PLUS, F_MODE_MASK | F_CMD_ACCESS,
+		 NULL, NULL);
 
 	MMIO_D(_MMIO(0x4ab8), D_KBL);
 	MMIO_D(_MMIO(0x2248), D_KBL | D_SKL);
@@ -3178,6 +3199,7 @@ static int init_bxt_mmio_info(struct intel_gvt *gvt)
 	MMIO_D(BXT_DSI_PLL_ENABLE, D_BXT);
 
 	MMIO_D(GEN9_CLKGATE_DIS_0, D_BXT);
+	MMIO_D(GEN9_CLKGATE_DIS_4, D_BXT);
 
 	MMIO_D(HSW_TVIDEO_DIP_GCP(TRANSCODER_A), D_BXT);
 	MMIO_D(HSW_TVIDEO_DIP_GCP(TRANSCODER_B), D_BXT);
diff --git a/drivers/gpu/drm/i915/gvt/kvmgt.c b/drivers/gpu/drm/i915/gvt/kvmgt.c
index 71751be329e3..c1072143da1d 100644
--- a/drivers/gpu/drm/i915/gvt/kvmgt.c
+++ b/drivers/gpu/drm/i915/gvt/kvmgt.c
@@ -32,6 +32,7 @@
 #include <linux/device.h>
 #include <linux/mm.h>
 #include <linux/mmu_context.h>
+#include <linux/sched/mm.h>
 #include <linux/types.h>
 #include <linux/list.h>
 #include <linux/rbtree.h>
@@ -1792,16 +1793,21 @@ static int kvmgt_rw_gpa(unsigned long handle, unsigned long gpa,
 	info = (struct kvmgt_guest_info *)handle;
 	kvm = info->kvm;
 
-	if (kthread)
+	if (kthread) {
+		if (!mmget_not_zero(kvm->mm))
+			return -EFAULT;
 		use_mm(kvm->mm);
+	}
 
 	idx = srcu_read_lock(&kvm->srcu);
 	ret = write ? kvm_write_guest(kvm, gpa, buf, len) :
 		      kvm_read_guest(kvm, gpa, buf, len);
 	srcu_read_unlock(&kvm->srcu, idx);
 
-	if (kthread)
+	if (kthread) {
 		unuse_mm(kvm->mm);
+		mmput(kvm->mm);
+	}
 
 	return ret;
 }
@@ -1827,6 +1833,8 @@ static bool kvmgt_is_valid_gfn(unsigned long handle, unsigned long gfn)
 {
 	struct kvmgt_guest_info *info;
 	struct kvm *kvm;
+	int idx;
+	bool ret;
 
 	if (!handle_valid(handle))
 		return false;
@@ -1834,8 +1842,11 @@ static bool kvmgt_is_valid_gfn(unsigned long handle, unsigned long gfn)
 	info = (struct kvmgt_guest_info *)handle;
 	kvm = info->kvm;
 
-	return kvm_is_visible_gfn(kvm, gfn);
+	idx = srcu_read_lock(&kvm->srcu);
+	ret = kvm_is_visible_gfn(kvm, gfn);
+	srcu_read_unlock(&kvm->srcu, idx);
 
+	return ret;
 }
 
 struct intel_gvt_mpt kvmgt_mpt = {
diff --git a/drivers/gpu/drm/i915/gvt/mmio.c b/drivers/gpu/drm/i915/gvt/mmio.c
index 4db817c21ed8..43f65848ecd6 100644
--- a/drivers/gpu/drm/i915/gvt/mmio.c
+++ b/drivers/gpu/drm/i915/gvt/mmio.c
@@ -245,6 +245,34 @@ void intel_vgpu_reset_mmio(struct intel_vgpu *vgpu, bool dmlr)
 
 		/* set the bit 0:2(Core C-State ) to C0 */
 		vgpu_vreg_t(vgpu, GEN6_GT_CORE_STATUS) = 0;
+
+		if (IS_BROXTON(vgpu->gvt->dev_priv)) {
+			vgpu_vreg_t(vgpu, BXT_P_CR_GT_DISP_PWRON) &=
+				    ~(BIT(0) | BIT(1));
+			vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY0)) &=
+				    ~PHY_POWER_GOOD;
+			vgpu_vreg_t(vgpu, BXT_PORT_CL1CM_DW0(DPIO_PHY1)) &=
+				    ~PHY_POWER_GOOD;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL_FAMILY(DPIO_PHY0)) &=
+				    ~BIT(30);
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL_FAMILY(DPIO_PHY1)) &=
+				    ~BIT(30);
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_A)) &=
+				    ~BXT_PHY_LANE_ENABLED;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_A)) |=
+				    BXT_PHY_CMNLANE_POWERDOWN_ACK |
+				    BXT_PHY_LANE_POWERDOWN_ACK;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_B)) &=
+				    ~BXT_PHY_LANE_ENABLED;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_B)) |=
+				    BXT_PHY_CMNLANE_POWERDOWN_ACK |
+				    BXT_PHY_LANE_POWERDOWN_ACK;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_C)) &=
+				    ~BXT_PHY_LANE_ENABLED;
+			vgpu_vreg_t(vgpu, BXT_PHY_CTL(PORT_C)) |=
+				    BXT_PHY_CMNLANE_POWERDOWN_ACK |
+				    BXT_PHY_LANE_POWERDOWN_ACK;
+		}
 	} else {
 #define GVT_GEN8_MMIO_RESET_OFFSET		(0x44200)
 		/* only reset the engine related, so starting with 0x44200
diff --git a/drivers/gpu/drm/i915/gvt/mmio_context.c b/drivers/gpu/drm/i915/gvt/mmio_context.c
index 7e702c6a32af..10e63eea5492 100644
--- a/drivers/gpu/drm/i915/gvt/mmio_context.c
+++ b/drivers/gpu/drm/i915/gvt/mmio_context.c
@@ -549,11 +549,9 @@ void intel_gvt_switch_mmio(struct intel_vgpu *pre,
 	 * performace for batch mmio read/write, so we need
 	 * handle forcewake mannually.
 	 */
-	intel_runtime_pm_get(dev_priv);
 	intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL);
 	switch_mmio(pre, next, ring_id);
 	intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);
-	intel_runtime_pm_put(dev_priv);
 }
 
 /**
diff --git a/drivers/gpu/drm/i915/gvt/opregion.c b/drivers/gpu/drm/i915/gvt/opregion.c
index 82586c8e434f..276db53f1bf1 100644
--- a/drivers/gpu/drm/i915/gvt/opregion.c
+++ b/drivers/gpu/drm/i915/gvt/opregion.c
@@ -42,8 +42,6 @@
 #define DEVICE_TYPE_EFP3   0x20
 #define DEVICE_TYPE_EFP4   0x10
 
-#define DEV_SIZE	38
-
 struct opregion_header {
 	u8 signature[16];
 	u32 size;
@@ -63,6 +61,10 @@ struct bdb_data_header {
 	u16 size; /* data size */
 } __packed;
 
+/* For supporting windows guest with opregion, here hardcode the emulated
+ * bdb header version as '186', and the corresponding child_device_config
+ * length should be '33' but not '38'.
+ */
 struct efp_child_device_config {
 	u16 handle;
 	u16 device_type;
@@ -109,12 +111,6 @@ struct efp_child_device_config {
 	u8 mipi_bridge_type; /* 171 */
 	u16 device_class_ext;
 	u8 dvo_function;
-	u8 dp_usb_type_c:1; /* 195 */
-	u8 skip6:7;
-	u8 dp_usb_type_c_2x_gpio_index; /* 195 */
-	u16 dp_usb_type_c_2x_gpio_pin; /* 195 */
-	u8 iboost_dp:4; /* 196 */
-	u8 iboost_hdmi:4; /* 196 */
 } __packed;
 
 struct vbt {
@@ -155,7 +151,7 @@ static void virt_vbt_generation(struct vbt *v)
 	v->header.bdb_offset = offsetof(struct vbt, bdb_header);
 
 	strcpy(&v->bdb_header.signature[0], "BIOS_DATA_BLOCK");
-	v->bdb_header.version = 186; /* child_dev_size = 38 */
+	v->bdb_header.version = 186; /* child_dev_size = 33 */
 	v->bdb_header.header_size = sizeof(v->bdb_header);
 
 	v->bdb_header.bdb_size = sizeof(struct vbt) - sizeof(struct vbt_header)
@@ -169,11 +165,13 @@ static void virt_vbt_generation(struct vbt *v)
 
 	/* child device */
 	num_child = 4; /* each port has one child */
+	v->general_definitions.child_dev_size =
+		sizeof(struct efp_child_device_config);
 	v->general_definitions_header.id = BDB_GENERAL_DEFINITIONS;
 	/* size will include child devices */
 	v->general_definitions_header.size =
-		sizeof(struct bdb_general_definitions) + num_child * DEV_SIZE;
-	v->general_definitions.child_dev_size = DEV_SIZE;
+		sizeof(struct bdb_general_definitions) +
+			num_child * v->general_definitions.child_dev_size;
 
 	/* portA */
 	v->child0.handle = DEVICE_TYPE_EFP1;
diff --git a/drivers/gpu/drm/i915/gvt/sched_policy.c b/drivers/gpu/drm/i915/gvt/sched_policy.c
index 09d7bb72b4ff..c32e7d5e8629 100644
--- a/drivers/gpu/drm/i915/gvt/sched_policy.c
+++ b/drivers/gpu/drm/i915/gvt/sched_policy.c
@@ -47,11 +47,15 @@ static bool vgpu_has_pending_workload(struct intel_vgpu *vgpu)
 	return false;
 }
 
+/* We give 2 seconds higher prio for vGPU during start */
+#define GVT_SCHED_VGPU_PRI_TIME  2
+
 struct vgpu_sched_data {
 	struct list_head lru_list;
 	struct intel_vgpu *vgpu;
 	bool active;
-
+	bool pri_sched;
+	ktime_t pri_time;
 	ktime_t sched_in_time;
 	ktime_t sched_time;
 	ktime_t left_ts;
@@ -183,6 +187,14 @@ static struct intel_vgpu *find_busy_vgpu(struct gvt_sched_data *sched_data)
 		if (!vgpu_has_pending_workload(vgpu_data->vgpu))
 			continue;
 
+		if (vgpu_data->pri_sched) {
+			if (ktime_before(ktime_get(), vgpu_data->pri_time)) {
+				vgpu = vgpu_data->vgpu;
+				break;
+			} else
+				vgpu_data->pri_sched = false;
+		}
+
 		/* Return the vGPU only if it has time slice left */
 		if (vgpu_data->left_ts > 0) {
 			vgpu = vgpu_data->vgpu;
@@ -202,6 +214,7 @@ static void tbs_sched_func(struct gvt_sched_data *sched_data)
 	struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler;
 	struct vgpu_sched_data *vgpu_data;
 	struct intel_vgpu *vgpu = NULL;
+
 	/* no active vgpu or has already had a target */
 	if (list_empty(&sched_data->lru_runq_head) || scheduler->next_vgpu)
 		goto out;
@@ -209,12 +222,13 @@ static void tbs_sched_func(struct gvt_sched_data *sched_data)
 	vgpu = find_busy_vgpu(sched_data);
 	if (vgpu) {
 		scheduler->next_vgpu = vgpu;
-
-		/* Move the last used vGPU to the tail of lru_list */
 		vgpu_data = vgpu->sched_data;
-		list_del_init(&vgpu_data->lru_list);
-		list_add_tail(&vgpu_data->lru_list,
-				&sched_data->lru_runq_head);
+		if (!vgpu_data->pri_sched) {
+			/* Move the last used vGPU to the tail of lru_list */
+			list_del_init(&vgpu_data->lru_list);
+			list_add_tail(&vgpu_data->lru_list,
+				      &sched_data->lru_runq_head);
+		}
 	} else {
 		scheduler->next_vgpu = gvt->idle_vgpu;
 	}
@@ -328,11 +342,17 @@ static void tbs_sched_start_schedule(struct intel_vgpu *vgpu)
 {
 	struct gvt_sched_data *sched_data = vgpu->gvt->scheduler.sched_data;
 	struct vgpu_sched_data *vgpu_data = vgpu->sched_data;
+	ktime_t now;
 
 	if (!list_empty(&vgpu_data->lru_list))
 		return;
 
-	list_add_tail(&vgpu_data->lru_list, &sched_data->lru_runq_head);
+	now = ktime_get();
+	vgpu_data->pri_time = ktime_add(now,
+					ktime_set(GVT_SCHED_VGPU_PRI_TIME, 0));
+	vgpu_data->pri_sched = true;
+
+	list_add(&vgpu_data->lru_list, &sched_data->lru_runq_head);
 
 	if (!hrtimer_active(&sched_data->timer))
 		hrtimer_start(&sched_data->timer, ktime_add_ns(ktime_get(),
@@ -426,6 +446,7 @@ void intel_vgpu_stop_schedule(struct intel_vgpu *vgpu)
 		&vgpu->gvt->scheduler;
 	int ring_id;
 	struct vgpu_sched_data *vgpu_data = vgpu->sched_data;
+	struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv;
 
 	if (!vgpu_data->active)
 		return;
@@ -444,6 +465,7 @@ void intel_vgpu_stop_schedule(struct intel_vgpu *vgpu)
 		scheduler->current_vgpu = NULL;
 	}
 
+	intel_runtime_pm_get(dev_priv);
 	spin_lock_bh(&scheduler->mmio_context_lock);
 	for (ring_id = 0; ring_id < I915_NUM_ENGINES; ring_id++) {
 		if (scheduler->engine_owner[ring_id] == vgpu) {
@@ -452,5 +474,6 @@ void intel_vgpu_stop_schedule(struct intel_vgpu *vgpu)
 		}
 	}
 	spin_unlock_bh(&scheduler->mmio_context_lock);
+	intel_runtime_pm_put(dev_priv);
 	mutex_unlock(&vgpu->gvt->sched_lock);
 }
diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c
index a4e8e3cf74fd..c628be05fbfe 100644
--- a/drivers/gpu/drm/i915/gvt/vgpu.c
+++ b/drivers/gpu/drm/i915/gvt/vgpu.c
@@ -281,6 +281,7 @@ void intel_gvt_destroy_vgpu(struct intel_vgpu *vgpu)
 	intel_vgpu_clean_submission(vgpu);
 	intel_vgpu_clean_display(vgpu);
 	intel_vgpu_clean_opregion(vgpu);
+	intel_vgpu_reset_ggtt(vgpu, true);
 	intel_vgpu_clean_gtt(vgpu);
 	intel_gvt_hypervisor_detach_vgpu(vgpu);
 	intel_vgpu_free_resource(vgpu);
diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index c2f10d899329..443dfaefd7a6 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -181,8 +181,9 @@ struct intel_overlay {
 	u32 brightness, contrast, saturation;
 	u32 old_xscale, old_yscale;
 	/* register access */
-	u32 flip_addr;
 	struct drm_i915_gem_object *reg_bo;
+	struct overlay_registers __iomem *regs;
+	u32 flip_addr;
 	/* flip handling */
 	struct i915_gem_active last_flip;
 };
@@ -210,29 +211,6 @@ static void i830_overlay_clock_gating(struct drm_i915_private *dev_priv,
 				  PCI_DEVFN(0, 0), I830_CLOCK_GATE, val);
 }
 
-static struct overlay_registers __iomem *
-intel_overlay_map_regs(struct intel_overlay *overlay)
-{
-	struct drm_i915_private *dev_priv = overlay->i915;
-	struct overlay_registers __iomem *regs;
-
-	if (OVERLAY_NEEDS_PHYSICAL(dev_priv))
-		regs = (struct overlay_registers __iomem *)overlay->reg_bo->phys_handle->vaddr;
-	else
-		regs = io_mapping_map_wc(&dev_priv->ggtt.iomap,
-					 overlay->flip_addr,
-					 PAGE_SIZE);
-
-	return regs;
-}
-
-static void intel_overlay_unmap_regs(struct intel_overlay *overlay,
-				     struct overlay_registers __iomem *regs)
-{
-	if (!OVERLAY_NEEDS_PHYSICAL(overlay->i915))
-		io_mapping_unmap(regs);
-}
-
 static void intel_overlay_submit_request(struct intel_overlay *overlay,
 					 struct i915_request *rq,
 					 i915_gem_retire_fn retire)
@@ -784,13 +762,13 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 				      struct drm_i915_gem_object *new_bo,
 				      struct put_image_params *params)
 {
-	int ret, tmp_width;
-	struct overlay_registers __iomem *regs;
-	bool scale_changed = false;
+	struct overlay_registers __iomem *regs = overlay->regs;
 	struct drm_i915_private *dev_priv = overlay->i915;
 	u32 swidth, swidthsw, sheight, ostride;
 	enum pipe pipe = overlay->crtc->pipe;
+	bool scale_changed = false;
 	struct i915_vma *vma;
+	int ret, tmp_width;
 
 	lockdep_assert_held(&dev_priv->drm.struct_mutex);
 	WARN_ON(!drm_modeset_is_locked(&dev_priv->drm.mode_config.connection_mutex));
@@ -815,30 +793,19 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 
 	if (!overlay->active) {
 		u32 oconfig;
-		regs = intel_overlay_map_regs(overlay);
-		if (!regs) {
-			ret = -ENOMEM;
-			goto out_unpin;
-		}
+
 		oconfig = OCONF_CC_OUT_8BIT;
 		if (IS_GEN4(dev_priv))
 			oconfig |= OCONF_CSC_MODE_BT709;
 		oconfig |= pipe == 0 ?
 			OCONF_PIPE_A : OCONF_PIPE_B;
 		iowrite32(oconfig, &regs->OCONFIG);
-		intel_overlay_unmap_regs(overlay, regs);
 
 		ret = intel_overlay_on(overlay);
 		if (ret != 0)
 			goto out_unpin;
 	}
 
-	regs = intel_overlay_map_regs(overlay);
-	if (!regs) {
-		ret = -ENOMEM;
-		goto out_unpin;
-	}
-
 	iowrite32((params->dst_y << 16) | params->dst_x, &regs->DWINPOS);
 	iowrite32((params->dst_h << 16) | params->dst_w, &regs->DWINSZ);
 
@@ -882,8 +849,6 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 
 	iowrite32(overlay_cmd_reg(params), &regs->OCMD);
 
-	intel_overlay_unmap_regs(overlay, regs);
-
 	ret = intel_overlay_continue(overlay, vma, scale_changed);
 	if (ret)
 		goto out_unpin;
@@ -901,7 +866,6 @@ out_pin_section:
 int intel_overlay_switch_off(struct intel_overlay *overlay)
 {
 	struct drm_i915_private *dev_priv = overlay->i915;
-	struct overlay_registers __iomem *regs;
 	int ret;
 
 	lockdep_assert_held(&dev_priv->drm.struct_mutex);
@@ -918,9 +882,7 @@ int intel_overlay_switch_off(struct intel_overlay *overlay)
 	if (ret != 0)
 		return ret;
 
-	regs = intel_overlay_map_regs(overlay);
-	iowrite32(0, &regs->OCMD);
-	intel_overlay_unmap_regs(overlay, regs);
+	iowrite32(0, &overlay->regs->OCMD);
 
 	return intel_overlay_off(overlay);
 }
@@ -1305,7 +1267,6 @@ int intel_overlay_attrs_ioctl(struct drm_device *dev, void *data,
 	struct drm_intel_overlay_attrs *attrs = data;
 	struct drm_i915_private *dev_priv = to_i915(dev);
 	struct intel_overlay *overlay;
-	struct overlay_registers __iomem *regs;
 	int ret;
 
 	overlay = dev_priv->overlay;
@@ -1345,15 +1306,7 @@ int intel_overlay_attrs_ioctl(struct drm_device *dev, void *data,
 		overlay->contrast   = attrs->contrast;
 		overlay->saturation = attrs->saturation;
 
-		regs = intel_overlay_map_regs(overlay);
-		if (!regs) {
-			ret = -ENOMEM;
-			goto out_unlock;
-		}
-
-		update_reg_attrs(overlay, regs);
-
-		intel_overlay_unmap_regs(overlay, regs);
+		update_reg_attrs(overlay, overlay->regs);
 
 		if (attrs->flags & I915_OVERLAY_UPDATE_GAMMA) {
 			if (IS_GEN2(dev_priv))
@@ -1386,12 +1339,47 @@ out_unlock:
 	return ret;
 }
 
+static int get_registers(struct intel_overlay *overlay, bool use_phys)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int err;
+
+	obj = i915_gem_object_create_stolen(overlay->i915, PAGE_SIZE);
+	if (obj == NULL)
+		obj = i915_gem_object_create_internal(overlay->i915, PAGE_SIZE);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vma = i915_gem_object_ggtt_pin(obj, NULL, 0, 0, PIN_MAPPABLE);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_put_bo;
+	}
+
+	if (use_phys)
+		overlay->flip_addr = sg_dma_address(obj->mm.pages->sgl);
+	else
+		overlay->flip_addr = i915_ggtt_offset(vma);
+	overlay->regs = i915_vma_pin_iomap(vma);
+	i915_vma_unpin(vma);
+
+	if (IS_ERR(overlay->regs)) {
+		err = PTR_ERR(overlay->regs);
+		goto err_put_bo;
+	}
+
+	overlay->reg_bo = obj;
+	return 0;
+
+err_put_bo:
+	i915_gem_object_put(obj);
+	return err;
+}
+
 void intel_setup_overlay(struct drm_i915_private *dev_priv)
 {
 	struct intel_overlay *overlay;
-	struct drm_i915_gem_object *reg_bo;
-	struct overlay_registers __iomem *regs;
-	struct i915_vma *vma = NULL;
 	int ret;
 
 	if (!HAS_OVERLAY(dev_priv))
@@ -1401,46 +1389,8 @@ void intel_setup_overlay(struct drm_i915_private *dev_priv)
 	if (!overlay)
 		return;
 
-	mutex_lock(&dev_priv->drm.struct_mutex);
-	if (WARN_ON(dev_priv->overlay))
-		goto out_free;
-
 	overlay->i915 = dev_priv;
 
-	reg_bo = NULL;
-	if (!OVERLAY_NEEDS_PHYSICAL(dev_priv))
-		reg_bo = i915_gem_object_create_stolen(dev_priv, PAGE_SIZE);
-	if (reg_bo == NULL)
-		reg_bo = i915_gem_object_create(dev_priv, PAGE_SIZE);
-	if (IS_ERR(reg_bo))
-		goto out_free;
-	overlay->reg_bo = reg_bo;
-
-	if (OVERLAY_NEEDS_PHYSICAL(dev_priv)) {
-		ret = i915_gem_object_attach_phys(reg_bo, PAGE_SIZE);
-		if (ret) {
-			DRM_ERROR("failed to attach phys overlay regs\n");
-			goto out_free_bo;
-		}
-		overlay->flip_addr = reg_bo->phys_handle->busaddr;
-	} else {
-		vma = i915_gem_object_ggtt_pin(reg_bo, NULL,
-					       0, PAGE_SIZE, PIN_MAPPABLE);
-		if (IS_ERR(vma)) {
-			DRM_ERROR("failed to pin overlay register bo\n");
-			ret = PTR_ERR(vma);
-			goto out_free_bo;
-		}
-		overlay->flip_addr = i915_ggtt_offset(vma);
-
-		ret = i915_gem_object_set_to_gtt_domain(reg_bo, true);
-		if (ret) {
-			DRM_ERROR("failed to move overlay register bo into the GTT\n");
-			goto out_unpin_bo;
-		}
-	}
-
-	/* init all values */
 	overlay->color_key = 0x0101fe;
 	overlay->color_key_enabled = true;
 	overlay->brightness = -19;
@@ -1449,44 +1399,51 @@ void intel_setup_overlay(struct drm_i915_private *dev_priv)
 
 	init_request_active(&overlay->last_flip, NULL);
 
-	regs = intel_overlay_map_regs(overlay);
-	if (!regs)
-		goto out_unpin_bo;
+	mutex_lock(&dev_priv->drm.struct_mutex);
+
+	ret = get_registers(overlay, OVERLAY_NEEDS_PHYSICAL(dev_priv));
+	if (ret)
+		goto out_free;
+
+	ret = i915_gem_object_set_to_gtt_domain(overlay->reg_bo, true);
+	if (ret)
+		goto out_reg_bo;
 
-	memset_io(regs, 0, sizeof(struct overlay_registers));
-	update_polyphase_filter(regs);
-	update_reg_attrs(overlay, regs);
+	mutex_unlock(&dev_priv->drm.struct_mutex);
 
-	intel_overlay_unmap_regs(overlay, regs);
+	memset_io(overlay->regs, 0, sizeof(struct overlay_registers));
+	update_polyphase_filter(overlay->regs);
+	update_reg_attrs(overlay, overlay->regs);
 
 	dev_priv->overlay = overlay;
-	mutex_unlock(&dev_priv->drm.struct_mutex);
-	DRM_INFO("initialized overlay support\n");
+	DRM_INFO("Initialized overlay support.\n");
 	return;
 
-out_unpin_bo:
-	if (vma)
-		i915_vma_unpin(vma);
-out_free_bo:
-	i915_gem_object_put(reg_bo);
+out_reg_bo:
+	i915_gem_object_put(overlay->reg_bo);
 out_free:
 	mutex_unlock(&dev_priv->drm.struct_mutex);
 	kfree(overlay);
-	return;
 }
 
 void intel_cleanup_overlay(struct drm_i915_private *dev_priv)
 {
-	if (!dev_priv->overlay)
+	struct intel_overlay *overlay;
+
+	overlay = fetch_and_zero(&dev_priv->overlay);
+	if (!overlay)
 		return;
 
-	/* The bo's should be free'd by the generic code already.
+	/*
+	 * The bo's should be free'd by the generic code already.
 	 * Furthermore modesetting teardown happens beforehand so the
-	 * hardware should be off already */
-	WARN_ON(dev_priv->overlay->active);
+	 * hardware should be off already.
+	 */
+	WARN_ON(overlay->active);
+
+	i915_gem_object_put(overlay->reg_bo);
 
-	i915_gem_object_put(dev_priv->overlay->reg_bo);
-	kfree(dev_priv->overlay);
+	kfree(overlay);
 }
 
 #if IS_ENABLED(CONFIG_DRM_I915_CAPTURE_ERROR)
@@ -1498,37 +1455,11 @@ struct intel_overlay_error_state {
 	u32 isr;
 };
 
-static struct overlay_registers __iomem *
-intel_overlay_map_regs_atomic(struct intel_overlay *overlay)
-{
-	struct drm_i915_private *dev_priv = overlay->i915;
-	struct overlay_registers __iomem *regs;
-
-	if (OVERLAY_NEEDS_PHYSICAL(dev_priv))
-		/* Cast to make sparse happy, but it's wc memory anyway, so
-		 * equivalent to the wc io mapping on X86. */
-		regs = (struct overlay_registers __iomem *)
-			overlay->reg_bo->phys_handle->vaddr;
-	else
-		regs = io_mapping_map_atomic_wc(&dev_priv->ggtt.iomap,
-						overlay->flip_addr);
-
-	return regs;
-}
-
-static void intel_overlay_unmap_regs_atomic(struct intel_overlay *overlay,
-					struct overlay_registers __iomem *regs)
-{
-	if (!OVERLAY_NEEDS_PHYSICAL(overlay->i915))
-		io_mapping_unmap_atomic(regs);
-}
-
 struct intel_overlay_error_state *
 intel_overlay_capture_error_state(struct drm_i915_private *dev_priv)
 {
 	struct intel_overlay *overlay = dev_priv->overlay;
 	struct intel_overlay_error_state *error;
-	struct overlay_registers __iomem *regs;
 
 	if (!overlay || !overlay->active)
 		return NULL;
@@ -1541,18 +1472,9 @@ intel_overlay_capture_error_state(struct drm_i915_private *dev_priv)
 	error->isr = I915_READ(ISR);
 	error->base = overlay->flip_addr;
 
-	regs = intel_overlay_map_regs_atomic(overlay);
-	if (!regs)
-		goto err;
-
-	memcpy_fromio(&error->regs, regs, sizeof(struct overlay_registers));
-	intel_overlay_unmap_regs_atomic(overlay, regs);
+	memcpy_fromio(&error->regs, overlay->regs, sizeof(error->regs));
 
 	return error;
-
-err:
-	kfree(error);
-	return NULL;
 }
 
 void