Merge tag 'drm-fixes-2020-03-13' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "It's a bit quieter, probably not as much as it could be.

  There is on large regression fix in here from Lyude for displayport
  bandwidth calculations, there've been reports of multi-monitor in
  docks not working since -rc1 and this has been tested to fix those.

  Otherwise it's a bunch of i915 (with some GVT fixes), a set of amdgpu
  watermark + bios fixes, and an exynos iommu cleanup fix.

  core:
   - DP MST bandwidth regression fix.

  i915:
   - hard lockup fix
   - GVT fixes
   - 32-bit alignment issue fix
   - timeline wait fixes
   - cacheline_retire and free

  amdgpu:
   - Update the display watermark bounding box for navi14
   - Fix fetching vbios directly from rom on vega20/arcturus
   - Navi and renoir watermark fixes

  exynos:
   - iommu object cleanup fix"

`

* tag 'drm-fixes-2020-03-13' of git://anongit.freedesktop.org/drm/drm:
  drm/dp_mst: Rewrite and fix bandwidth limit checks
  drm/dp_mst: Reprobe path resources in CSN handler
  drm/dp_mst: Use full_pbn instead of available_pbn for bandwidth checks
  drm/dp_mst: Rename drm_dp_mst_is_dp_mst_end_device() to be less redundant
  drm/i915: Defer semaphore priority bumping to a workqueue
  drm/i915/gt: Close race between cacheline_retire and free
  drm/i915/execlists: Enable timeslice on partial virtual engine dequeue
  drm/i915: be more solid in checking the alignment
  drm/i915/gvt: Fix dma-buf display blur issue on CFL
  drm/i915: Return early for await_start on same timeline
  drm/i915: Actually emit the await_start
  drm/amdgpu/powerplay: nv1x, renior copy dcn clock settings of watermark to smu during boot up
  drm/exynos: Fix cleanup of IOMMU related objects
  drm/amdgpu: correct ROM_INDEX/DATA offset for VEGA20
  drm/amd/display: update soc bb for nv14
  drm/i915/gvt: Fix emulated vbt size issue
  drm/i915/gvt: Fix unnecessary schedule timer when no vGPU exits

27 files changed, 405 insertions, 133 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c
index 2b488dfb2f21..d8945c31b622 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.c
@@ -89,6 +89,13 @@
 #define HDP_MEM_POWER_CTRL__RC_MEM_POWER_CTRL_EN_MASK	0x00010000L
 #define HDP_MEM_POWER_CTRL__RC_MEM_POWER_LS_EN_MASK		0x00020000L
 #define mmHDP_MEM_POWER_CTRL_BASE_IDX	0
+
+/* for Vega20/arcturus regiter offset change */
+#define	mmROM_INDEX_VG20				0x00e4
+#define	mmROM_INDEX_VG20_BASE_IDX			0
+#define	mmROM_DATA_VG20					0x00e5
+#define	mmROM_DATA_VG20_BASE_IDX			0
+
 /*
  * Indirect registers accessor
  */
@@ -309,6 +316,8 @@ static bool soc15_read_bios_from_rom(struct amdgpu_device *adev,
 {
 	u32 *dw_ptr;
 	u32 i, length_dw;
+	uint32_t rom_index_offset;
+	uint32_t rom_data_offset;
 
 	if (bios == NULL)
 		return false;
@@ -321,11 +330,23 @@ static bool soc15_read_bios_from_rom(struct amdgpu_device *adev,
 	dw_ptr = (u32 *)bios;
 	length_dw = ALIGN(length_bytes, 4) / 4;
 
+	switch (adev->asic_type) {
+	case CHIP_VEGA20:
+	case CHIP_ARCTURUS:
+		rom_index_offset = SOC15_REG_OFFSET(SMUIO, 0, mmROM_INDEX_VG20);
+		rom_data_offset = SOC15_REG_OFFSET(SMUIO, 0, mmROM_DATA_VG20);
+		break;
+	default:
+		rom_index_offset = SOC15_REG_OFFSET(SMUIO, 0, mmROM_INDEX);
+		rom_data_offset = SOC15_REG_OFFSET(SMUIO, 0, mmROM_DATA);
+		break;
+	}
+
 	/* set rom index to 0 */
-	WREG32(SOC15_REG_OFFSET(SMUIO, 0, mmROM_INDEX), 0);
+	WREG32(rom_index_offset, 0);
 	/* read out the rom data */
 	for (i = 0; i < length_dw; i++)
-		dw_ptr[i] = RREG32(SOC15_REG_OFFSET(SMUIO, 0, mmROM_DATA));
+		dw_ptr[i] = RREG32(rom_data_offset);
 
 	return true;
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c
index 85f90f3e24cb..e310d67c399a 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c
@@ -335,6 +335,117 @@ struct _vcs_dpi_soc_bounding_box_st dcn2_0_soc = {
 	.use_urgent_burst_bw = 0
 };
 
+struct _vcs_dpi_soc_bounding_box_st dcn2_0_nv14_soc = {
+	.clock_limits = {
+			{
+				.state = 0,
+				.dcfclk_mhz = 560.0,
+				.fabricclk_mhz = 560.0,
+				.dispclk_mhz = 513.0,
+				.dppclk_mhz = 513.0,
+				.phyclk_mhz = 540.0,
+				.socclk_mhz = 560.0,
+				.dscclk_mhz = 171.0,
+				.dram_speed_mts = 8960.0,
+			},
+			{
+				.state = 1,
+				.dcfclk_mhz = 694.0,
+				.fabricclk_mhz = 694.0,
+				.dispclk_mhz = 642.0,
+				.dppclk_mhz = 642.0,
+				.phyclk_mhz = 600.0,
+				.socclk_mhz = 694.0,
+				.dscclk_mhz = 214.0,
+				.dram_speed_mts = 11104.0,
+			},
+			{
+				.state = 2,
+				.dcfclk_mhz = 875.0,
+				.fabricclk_mhz = 875.0,
+				.dispclk_mhz = 734.0,
+				.dppclk_mhz = 734.0,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 875.0,
+				.dscclk_mhz = 245.0,
+				.dram_speed_mts = 14000.0,
+			},
+			{
+				.state = 3,
+				.dcfclk_mhz = 1000.0,
+				.fabricclk_mhz = 1000.0,
+				.dispclk_mhz = 1100.0,
+				.dppclk_mhz = 1100.0,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 1000.0,
+				.dscclk_mhz = 367.0,
+				.dram_speed_mts = 16000.0,
+			},
+			{
+				.state = 4,
+				.dcfclk_mhz = 1200.0,
+				.fabricclk_mhz = 1200.0,
+				.dispclk_mhz = 1284.0,
+				.dppclk_mhz = 1284.0,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 1200.0,
+				.dscclk_mhz = 428.0,
+				.dram_speed_mts = 16000.0,
+			},
+			/*Extra state, no dispclk ramping*/
+			{
+				.state = 5,
+				.dcfclk_mhz = 1200.0,
+				.fabricclk_mhz = 1200.0,
+				.dispclk_mhz = 1284.0,
+				.dppclk_mhz = 1284.0,
+				.phyclk_mhz = 810.0,
+				.socclk_mhz = 1200.0,
+				.dscclk_mhz = 428.0,
+				.dram_speed_mts = 16000.0,
+			},
+		},
+	.num_states = 5,
+	.sr_exit_time_us = 8.6,
+	.sr_enter_plus_exit_time_us = 10.9,
+	.urgent_latency_us = 4.0,
+	.urgent_latency_pixel_data_only_us = 4.0,
+	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
+	.urgent_latency_vm_data_only_us = 4.0,
+	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
+	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 40.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 40.0,
+	.pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0,
+	.max_avg_sdp_bw_use_normal_percent = 40.0,
+	.max_avg_dram_bw_use_normal_percent = 40.0,
+	.writeback_latency_us = 12.0,
+	.ideal_dram_bw_after_urgent_percent = 40.0,
+	.max_request_size_bytes = 256,
+	.dram_channel_width_bytes = 2,
+	.fabric_datapath_to_dcn_data_return_bytes = 64,
+	.dcn_downspread_percent = 0.5,
+	.downspread_percent = 0.38,
+	.dram_page_open_time_ns = 50.0,
+	.dram_rw_turnaround_time_ns = 17.5,
+	.dram_return_buffer_per_channel_bytes = 8192,
+	.round_trip_ping_latency_dcfclk_cycles = 131,
+	.urgent_out_of_order_return_per_channel_bytes = 256,
+	.channel_interleave_bytes = 256,
+	.num_banks = 8,
+	.num_chans = 8,
+	.vmm_page_size_bytes = 4096,
+	.dram_clock_change_latency_us = 404.0,
+	.dummy_pstate_latency_us = 5.0,
+	.writeback_dram_clock_change_latency_us = 23.0,
+	.return_bus_width_bytes = 64,
+	.dispclk_dppclk_vco_speed_mhz = 3850,
+	.xfc_bus_transport_time_us = 20,
+	.xfc_xbuf_latency_tolerance_us = 4,
+	.use_urgent_burst_bw = 0
+};
+
 struct _vcs_dpi_soc_bounding_box_st dcn2_0_nv12_soc = { 0 };
 
 #ifndef mmDP0_DP_DPHY_INTERNAL_CTRL
@@ -3291,6 +3402,9 @@ void dcn20_patch_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_st
 static struct _vcs_dpi_soc_bounding_box_st *get_asic_rev_soc_bb(
 	uint32_t hw_internal_rev)
 {
+	if (ASICREV_IS_NAVI14_M(hw_internal_rev))
+		return &dcn2_0_nv14_soc;
+
 	if (ASICREV_IS_NAVI12_P(hw_internal_rev))
 		return &dcn2_0_nv12_soc;
 
diff --git a/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c b/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c
index ad8e9b5628e4..96e81c7bc266 100644
--- a/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c
+++ b/drivers/gpu/drm/amd/powerplay/amdgpu_smu.c
@@ -2006,8 +2006,11 @@ int smu_set_watermarks_for_clock_ranges(struct smu_context *smu,
 			smu_feature_is_enabled(smu, SMU_FEATURE_DPM_DCEFCLK_BIT) &&
 			smu_feature_is_enabled(smu, SMU_FEATURE_DPM_SOCCLK_BIT)) {
 		smu_set_watermarks_table(smu, table, clock_ranges);
-		smu->watermarks_bitmap |= WATERMARKS_EXIST;
-		smu->watermarks_bitmap &= ~WATERMARKS_LOADED;
+
+		if (!(smu->watermarks_bitmap & WATERMARKS_EXIST)) {
+			smu->watermarks_bitmap |= WATERMARKS_EXIST;
+			smu->watermarks_bitmap &= ~WATERMARKS_LOADED;
+		}
 	}
 
 	mutex_unlock(&smu->mutex);
diff --git a/drivers/gpu/drm/amd/powerplay/navi10_ppt.c b/drivers/gpu/drm/amd/powerplay/navi10_ppt.c
index 0d73a49166af..aed4d6e60907 100644
--- a/drivers/gpu/drm/amd/powerplay/navi10_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/navi10_ppt.c
@@ -1063,15 +1063,6 @@ static int navi10_display_config_changed(struct smu_context *smu)
 	int ret = 0;
 
 	if ((smu->watermarks_bitmap & WATERMARKS_EXIST) &&
-	    !(smu->watermarks_bitmap & WATERMARKS_LOADED)) {
-		ret = smu_write_watermarks_table(smu);
-		if (ret)
-			return ret;
-
-		smu->watermarks_bitmap |= WATERMARKS_LOADED;
-	}
-
-	if ((smu->watermarks_bitmap & WATERMARKS_EXIST) &&
 	    smu_feature_is_supported(smu, SMU_FEATURE_DPM_DCEFCLK_BIT) &&
 	    smu_feature_is_supported(smu, SMU_FEATURE_DPM_SOCCLK_BIT)) {
 		ret = smu_send_smc_msg_with_param(smu, SMU_MSG_NumOfDisplays,
@@ -1493,6 +1484,7 @@ static int navi10_set_watermarks_table(struct smu_context *smu,
 				       *clock_ranges)
 {
 	int i;
+	int ret = 0;
 	Watermarks_t *table = watermarks;
 
 	if (!table || !clock_ranges)
@@ -1544,6 +1536,18 @@ static int navi10_set_watermarks_table(struct smu_context *smu,
 				clock_ranges->wm_mcif_clocks_ranges[i].wm_set_id;
 	}
 
+	smu->watermarks_bitmap |= WATERMARKS_EXIST;
+
+	/* pass data to smu controller */
+	if (!(smu->watermarks_bitmap & WATERMARKS_LOADED)) {
+		ret = smu_write_watermarks_table(smu);
+		if (ret) {
+			pr_err("Failed to update WMTABLE!");
+			return ret;
+		}
+		smu->watermarks_bitmap |= WATERMARKS_LOADED;
+	}
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/powerplay/renoir_ppt.c b/drivers/gpu/drm/amd/powerplay/renoir_ppt.c
index 568c041c2206..3ad0f4aa3aa3 100644
--- a/drivers/gpu/drm/amd/powerplay/renoir_ppt.c
+++ b/drivers/gpu/drm/amd/powerplay/renoir_ppt.c
@@ -806,9 +806,10 @@ static int renoir_set_watermarks_table(
 				clock_ranges->wm_mcif_clocks_ranges[i].wm_set_id;
 	}
 
+	smu->watermarks_bitmap |= WATERMARKS_EXIST;
+
 	/* pass data to smu controller */
-	if ((smu->watermarks_bitmap & WATERMARKS_EXIST) &&
-			!(smu->watermarks_bitmap & WATERMARKS_LOADED)) {
+	if (!(smu->watermarks_bitmap & WATERMARKS_LOADED)) {
 		ret = smu_write_watermarks_table(smu);
 		if (ret) {
 			pr_err("Failed to update WMTABLE!");
diff --git a/drivers/gpu/drm/drm_dp_mst_topology.c b/drivers/gpu/drm/drm_dp_mst_topology.c
index cce0b1bba591..ed0fea2ac322 100644
--- a/drivers/gpu/drm/drm_dp_mst_topology.c
+++ b/drivers/gpu/drm/drm_dp_mst_topology.c
@@ -1935,7 +1935,7 @@ static u8 drm_dp_calculate_rad(struct drm_dp_mst_port *port,
 	return parent_lct + 1;
 }
 
-static bool drm_dp_mst_is_dp_mst_end_device(u8 pdt, bool mcs)
+static bool drm_dp_mst_is_end_device(u8 pdt, bool mcs)
 {
 	switch (pdt) {
 	case DP_PEER_DEVICE_DP_LEGACY_CONV:
@@ -1965,13 +1965,13 @@ drm_dp_port_set_pdt(struct drm_dp_mst_port *port, u8 new_pdt,
 
 	/* Teardown the old pdt, if there is one */
 	if (port->pdt != DP_PEER_DEVICE_NONE) {
-		if (drm_dp_mst_is_dp_mst_end_device(port->pdt, port->mcs)) {
+		if (drm_dp_mst_is_end_device(port->pdt, port->mcs)) {
 			/*
 			 * If the new PDT would also have an i2c bus,
 			 * don't bother with reregistering it
 			 */
 			if (new_pdt != DP_PEER_DEVICE_NONE &&
-			    drm_dp_mst_is_dp_mst_end_device(new_pdt, new_mcs)) {
+			    drm_dp_mst_is_end_device(new_pdt, new_mcs)) {
 				port->pdt = new_pdt;
 				port->mcs = new_mcs;
 				return 0;
@@ -1991,7 +1991,7 @@ drm_dp_port_set_pdt(struct drm_dp_mst_port *port, u8 new_pdt,
 	port->mcs = new_mcs;
 
 	if (port->pdt != DP_PEER_DEVICE_NONE) {
-		if (drm_dp_mst_is_dp_mst_end_device(port->pdt, port->mcs)) {
+		if (drm_dp_mst_is_end_device(port->pdt, port->mcs)) {
 			/* add i2c over sideband */
 			ret = drm_dp_mst_register_i2c_bus(&port->aux);
 		} else {
@@ -2172,7 +2172,7 @@ drm_dp_mst_port_add_connector(struct drm_dp_mst_branch *mstb,
 	}
 
 	if (port->pdt != DP_PEER_DEVICE_NONE &&
-	    drm_dp_mst_is_dp_mst_end_device(port->pdt, port->mcs)) {
+	    drm_dp_mst_is_end_device(port->pdt, port->mcs)) {
 		port->cached_edid = drm_get_edid(port->connector,
 						 &port->aux.ddc);
 		drm_connector_set_tile_property(port->connector);
@@ -2302,14 +2302,18 @@ drm_dp_mst_handle_link_address_port(struct drm_dp_mst_branch *mstb,
 		mutex_unlock(&mgr->lock);
 	}
 
-	if (old_ddps != port->ddps) {
-		if (port->ddps) {
-			if (!port->input) {
-				drm_dp_send_enum_path_resources(mgr, mstb,
-								port);
-			}
+	/*
+	 * Reprobe PBN caps on both hotplug, and when re-probing the link
+	 * for our parent mstb
+	 */
+	if (old_ddps != port->ddps || !created) {
+		if (port->ddps && !port->input) {
+			ret = drm_dp_send_enum_path_resources(mgr, mstb,
+							      port);
+			if (ret == 1)
+				changed = true;
 		} else {
-			port->available_pbn = 0;
+			port->full_pbn = 0;
 		}
 	}
 
@@ -2401,11 +2405,10 @@ drm_dp_mst_handle_conn_stat(struct drm_dp_mst_branch *mstb,
 	port->ddps = conn_stat->displayport_device_plug_status;
 
 	if (old_ddps != port->ddps) {
-		if (port->ddps) {
-			dowork = true;
-		} else {
-			port->available_pbn = 0;
-		}
+		if (port->ddps && !port->input)
+			drm_dp_send_enum_path_resources(mgr, mstb, port);
+		else
+			port->full_pbn = 0;
 	}
 
 	new_pdt = port->input ? DP_PEER_DEVICE_NONE : conn_stat->peer_device_type;
@@ -2556,13 +2559,6 @@ static int drm_dp_check_and_send_link_address(struct drm_dp_mst_topology_mgr *mg
 		if (port->input || !port->ddps)
 			continue;
 
-		if (!port->available_pbn) {
-			drm_modeset_lock(&mgr->base.lock, NULL);
-			drm_dp_send_enum_path_resources(mgr, mstb, port);
-			drm_modeset_unlock(&mgr->base.lock);
-			changed = true;
-		}
-
 		if (port->mstb)
 			mstb_child = drm_dp_mst_topology_get_mstb_validated(
 			    mgr, port->mstb);
@@ -2990,6 +2986,7 @@ drm_dp_send_enum_path_resources(struct drm_dp_mst_topology_mgr *mgr,
 
 	ret = drm_dp_mst_wait_tx_reply(mstb, txmsg);
 	if (ret > 0) {
+		ret = 0;
 		path_res = &txmsg->reply.u.path_resources;
 
 		if (txmsg->reply.reply_type == DP_SIDEBAND_REPLY_NAK) {
@@ -3002,14 +2999,22 @@ drm_dp_send_enum_path_resources(struct drm_dp_mst_topology_mgr *mgr,
 				      path_res->port_number,
 				      path_res->full_payload_bw_number,
 				      path_res->avail_payload_bw_number);
-			port->available_pbn =
-				path_res->avail_payload_bw_number;
+
+			/*
+			 * If something changed, make sure we send a
+			 * hotplug
+			 */
+			if (port->full_pbn != path_res->full_payload_bw_number ||
+			    port->fec_capable != path_res->fec_capable)
+				ret = 1;
+
+			port->full_pbn = path_res->full_payload_bw_number;
 			port->fec_capable = path_res->fec_capable;
 		}
 	}
 
 	kfree(txmsg);
-	return 0;
+	return ret;
 }
 
 static struct drm_dp_mst_port *drm_dp_get_last_connected_port_to_mstb(struct drm_dp_mst_branch *mstb)
@@ -3596,13 +3601,9 @@ drm_dp_mst_topology_mgr_invalidate_mstb(struct drm_dp_mst_branch *mstb)
 	/* The link address will need to be re-sent on resume */
 	mstb->link_address_sent = false;
 
-	list_for_each_entry(port, &mstb->ports, next) {
-		/* The PBN for each port will also need to be re-probed */
-		port->available_pbn = 0;
-
+	list_for_each_entry(port, &mstb->ports, next)
 		if (port->mstb)
 			drm_dp_mst_topology_mgr_invalidate_mstb(port->mstb);
-	}
 }
 
 /**
@@ -4829,41 +4830,102 @@ static bool drm_dp_mst_port_downstream_of_branch(struct drm_dp_mst_port *port,
 	return false;
 }
 
-static inline
-int drm_dp_mst_atomic_check_bw_limit(struct drm_dp_mst_branch *branch,
-				     struct drm_dp_mst_topology_state *mst_state)
+static int
+drm_dp_mst_atomic_check_port_bw_limit(struct drm_dp_mst_port *port,
+				      struct drm_dp_mst_topology_state *state);
+
+static int
+drm_dp_mst_atomic_check_mstb_bw_limit(struct drm_dp_mst_branch *mstb,
+				      struct drm_dp_mst_topology_state *state)
 {
-	struct drm_dp_mst_port *port;
 	struct drm_dp_vcpi_allocation *vcpi;
-	int pbn_limit = 0, pbn_used = 0;
+	struct drm_dp_mst_port *port;
+	int pbn_used = 0, ret;
+	bool found = false;
 
-	list_for_each_entry(port, &branch->ports, next) {
-		if (port->mstb)
-			if (drm_dp_mst_atomic_check_bw_limit(port->mstb, mst_state))
-				return -ENOSPC;
+	/* Check that we have at least one port in our state that's downstream
+	 * of this branch, otherwise we can skip this branch
+	 */
+	list_for_each_entry(vcpi, &state->vcpis, next) {
+		if (!vcpi->pbn ||
+		    !drm_dp_mst_port_downstream_of_branch(vcpi->port, mstb))
+			continue;
 
-		if (port->available_pbn > 0)
-			pbn_limit = port->available_pbn;
+		found = true;
+		break;
 	}
-	DRM_DEBUG_ATOMIC("[MST BRANCH:%p] branch has %d PBN available\n",
-			 branch, pbn_limit);
+	if (!found)
+		return 0;
 
-	list_for_each_entry(vcpi, &mst_state->vcpis, next) {
-		if (!vcpi->pbn)
-			continue;
+	if (mstb->port_parent)
+		DRM_DEBUG_ATOMIC("[MSTB:%p] [MST PORT:%p] Checking bandwidth limits on [MSTB:%p]\n",
+				 mstb->port_parent->parent, mstb->port_parent,
+				 mstb);
+	else
+		DRM_DEBUG_ATOMIC("[MSTB:%p] Checking bandwidth limits\n",
+				 mstb);
+
+	list_for_each_entry(port, &mstb->ports, next) {
+		ret = drm_dp_mst_atomic_check_port_bw_limit(port, state);
+		if (ret < 0)
+			return ret;
 
-		if (drm_dp_mst_port_downstream_of_branch(vcpi->port, branch))
-			pbn_used += vcpi->pbn;
+		pbn_used += ret;
 	}
-	DRM_DEBUG_ATOMIC("[MST BRANCH:%p] branch used %d PBN\n",
-			 branch, pbn_used);
 
-	if (pbn_used > pbn_limit) {
-		DRM_DEBUG_ATOMIC("[MST BRANCH:%p] No available bandwidth\n",
-				 branch);
+	return pbn_used;
+}
+
+static int
+drm_dp_mst_atomic_check_port_bw_limit(struct drm_dp_mst_port *port,
+				      struct drm_dp_mst_topology_state *state)
+{
+	struct drm_dp_vcpi_allocation *vcpi;
+	int pbn_used = 0;
+
+	if (port->pdt == DP_PEER_DEVICE_NONE)
+		return 0;
+
+	if (drm_dp_mst_is_end_device(port->pdt, port->mcs)) {
+		bool found = false;
+
+		list_for_each_entry(vcpi, &state->vcpis, next) {
+			if (vcpi->port != port)
+				continue;
+			if (!vcpi->pbn)
+				return 0;
+
+			found = true;
+			break;
+		}
+		if (!found)
+			return 0;
+
+		/* This should never happen, as it means we tried to
+		 * set a mode before querying the full_pbn
+		 */
+		if (WARN_ON(!port->full_pbn))
+			return -EINVAL;
+
+		pbn_used = vcpi->pbn;
+	} else {
+		pbn_used = drm_dp_mst_atomic_check_mstb_bw_limit(port->mstb,
+								 state);
+		if (pbn_used <= 0)
+			return pbn_used;
+	}
+
+	if (pbn_used > port->full_pbn) {
+		DRM_DEBUG_ATOMIC("[MSTB:%p] [MST PORT:%p] required PBN of %d exceeds port limit of %d\n",
+				 port->parent, port, pbn_used,
+				 port->full_pbn);
 		return -ENOSPC;
 	}
-	return 0;
+
+	DRM_DEBUG_ATOMIC("[MSTB:%p] [MST PORT:%p] uses %d out of %d PBN\n",
+			 port->parent, port, pbn_used, port->full_pbn);
+
+	return pbn_used;
 }
 
 static inline int
@@ -5061,9 +5123,15 @@ int drm_dp_mst_atomic_check(struct drm_atomic_state *state)
 		ret = drm_dp_mst_atomic_check_vcpi_alloc_limit(mgr, mst_state);
 		if (ret)
 			break;
-		ret = drm_dp_mst_atomic_check_bw_limit(mgr->mst_primary, mst_state);
-		if (ret)
+
+		mutex_lock(&mgr->lock);
+		ret = drm_dp_mst_atomic_check_mstb_bw_limit(mgr->mst_primary,
+							    mst_state);
+		mutex_unlock(&mgr->lock);
+		if (ret < 0)
 			break;
+		else
+			ret = 0;
 	}
 
 	return ret;
diff --git a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c
index 8428ae12dfa5..1f79bc2a881e 100644
--- a/drivers/gpu/drm/exynos/exynos5433_drm_decon.c
+++ b/drivers/gpu/drm/exynos/exynos5433_drm_decon.c
@@ -55,6 +55,7 @@ static const char * const decon_clks_name[] = {
 struct decon_context {
 	struct device			*dev;
 	struct drm_device		*drm_dev;
+	void				*dma_priv;
 	struct exynos_drm_crtc		*crtc;
 	struct exynos_drm_plane		planes[WINDOWS_NR];
 	struct exynos_drm_plane_config	configs[WINDOWS_NR];
@@ -644,7 +645,7 @@ static int decon_bind(struct device *dev, struct device *master, void *data)
 
 	decon_clear_channels(ctx->crtc);
 
-	return exynos_drm_register_dma(drm_dev, dev);
+	return exynos_drm_register_dma(drm_dev, dev, &ctx->dma_priv);
 }
 
 static void decon_unbind(struct device *dev, struct device *master, void *data)
@@ -654,7 +655,7 @@ static void decon_unbind(struct device *dev, struct device *master, void *data)
 	decon_atomic_disable(ctx->crtc);
 
 	/* detach this sub driver from iommu mapping if supported. */
-	exynos_drm_unregister_dma(ctx->drm_dev, ctx->dev);
+	exynos_drm_unregister_dma(ctx->drm_dev, ctx->dev, &ctx->dma_priv);
 }
 
 static const struct component_ops decon_component_ops = {
diff --git a/drivers/gpu/drm/exynos/exynos7_drm_decon.c b/drivers/gpu/drm/exynos/exynos7_drm_decon.c
index ff59c641fa80..1eed3327999f 100644
--- a/drivers/gpu/drm/exynos/exynos7_drm_decon.c
+++ b/drivers/gpu/drm/exynos/exynos7_drm_decon.c
@@ -40,6 +40,7 @@
 struct decon_context {
 	struct device			*dev;
 	struct drm_device		*drm_dev;
+	void				*dma_priv;
 	struct exynos_drm_crtc		*crtc;
 	struct exynos_drm_plane		planes[WINDOWS_NR];
 	struct exynos_drm_plane_config	configs[WINDOWS_NR];
@@ -127,13 +128,13 @@ static int decon_ctx_initialize(struct decon_context *ctx,
 
 	decon_clear_channels(ctx->crtc);
 
-	return exynos_drm_register_dma(drm_dev, ctx->dev);
+	return exynos_drm_register_dma(drm_dev, ctx->dev, &ctx->dma_priv);
 }
 
 static void decon_ctx_remove(struct decon_context *ctx)
 {
 	/* detach this sub driver from iommu mapping if supported. */
-	exynos_drm_unregister_dma(ctx->drm_dev, ctx->dev);
+	exynos_drm_unregister_dma(ctx->drm_dev, ctx->dev, &ctx->dma_priv);
 }
 
 static u32 decon_calc_clkdiv(struct decon_context *ctx,
diff --git a/drivers/gpu/drm/exynos/exynos_drm_dma.c b/drivers/gpu/drm/exynos/exynos_drm_dma.c
index 9ebc02768847..619f81435c1b 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_dma.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_dma.c
@@ -58,7 +58,7 @@ static inline void clear_dma_max_seg_size(struct device *dev)
  * mapping.
  */
 static int drm_iommu_attach_device(struct drm_device *drm_dev,
-				struct device *subdrv_dev)
+				struct device *subdrv_dev, void **dma_priv)
 {
 	struct exynos_drm_private *priv = drm_dev->dev_private;
 	int ret;
@@ -74,7 +74,14 @@ static int drm_iommu_attach_device(struct drm_device *drm_dev,
 		return ret;
 
 	if (IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU)) {
-		if (to_dma_iommu_mapping(subdrv_dev))
+		/*
+		 * Keep the original DMA mapping of the sub-device and
+		 * restore it on Exynos DRM detach, otherwise the DMA
+		 * framework considers it as IOMMU-less during the next
+		 * probe (in case of deferred probe or modular build)
+		 */
+		*dma_priv = to_dma_iommu_mapping(subdrv_dev);
+		if (*dma_priv)
 			arm_iommu_detach_device(subdrv_dev);
 
 		ret = arm_iommu_attach_device(subdrv_dev, priv->mapping);
@@ -98,19 +105,21 @@ static int drm_iommu_attach_device(struct drm_device *drm_dev,
  * mapping
  */
 static void drm_iommu_detach_device(struct drm_device *drm_dev,
-				struct device *subdrv_dev)
+				    struct device *subdrv_dev, void **dma_priv)
 {
 	struct exynos_drm_private *priv = drm_dev->dev_private;
 
-	if (IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU))
+	if (IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU)) {
 		arm_iommu_detach_device(subdrv_dev);
-	else if (IS_ENABLED(CONFIG_IOMMU_DMA))
+		arm_iommu_attach_device(subdrv_dev, *dma_priv);
+	} else if (IS_ENABLED(CONFIG_IOMMU_DMA))
 		iommu_detach_device(priv->mapping, subdrv_dev);
 
 	clear_dma_max_seg_size(subdrv_dev);
 }
 
-int exynos_drm_register_dma(struct drm_device *drm, struct device *dev)
+int exynos_drm_register_dma(struct drm_device *drm, struct device *dev,
+			    void **dma_priv)
 {
 	struct exynos_drm_private *priv = drm->dev_private;
 
@@ -137,13 +146,14 @@ int exynos_drm_register_dma(struct drm_device *drm, struct device *dev)
 		priv->mapping = mapping;
 	}
 
-	return drm_iommu_attach_device(drm, dev);
+	return drm_iommu_attach_device(drm, dev, dma_priv);
 }
 
-void exynos_drm_unregister_dma(struct drm_device *drm, struct device *dev)
+void exynos_drm_unregister_dma(struct drm_device *drm, struct device *dev,
+			       void **dma_priv)
 {
 	if (IS_ENABLED(CONFIG_EXYNOS_IOMMU))
-		drm_iommu_detach_device(drm, dev);
+		drm_iommu_detach_device(drm, dev, dma_priv);
 }
 
 void exynos_drm_cleanup_dma(struct drm_device *drm)
diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.h b/drivers/gpu/drm/exynos/exynos_drm_drv.h
index d4d21d8cfb90..6ae9056e7a18 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_drv.h
+++ b/drivers/gpu/drm/exynos/exynos_drm_drv.h
@@ -223,8 +223,10 @@ static inline bool is_drm_iommu_supported(struct drm_device *drm_dev)
 	return priv->mapping ? true : false;
 }
 
-int exynos_drm_register_dma(struct drm_device *drm, struct device *dev);
-void exynos_drm_unregister_dma(struct drm_device *drm, struct device *dev);
+int exynos_drm_register_dma(struct drm_device *drm, struct device *dev,
+			    void **dma_priv);
+void exynos_drm_unregister_dma(struct drm_device *drm, struct device *dev,
+			       void **dma_priv);
 void exynos_drm_cleanup_dma(struct drm_device *drm);
 
 #ifdef CONFIG_DRM_EXYNOS_DPI
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimc.c b/drivers/gpu/drm/exynos/exynos_drm_fimc.c
index 8ea2e1d77802..29ab8be8604c 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimc.c
@@ -97,6 +97,7 @@ struct fimc_scaler {
 struct fimc_context {
 	struct exynos_drm_ipp ipp;
 	struct drm_device *drm_dev;
+	void		*dma_priv;
 	struct device	*dev;
 	struct exynos_drm_ipp_task	*task;
 	struct exynos_drm_ipp_formats	*formats;
@@ -1133,7 +1134,7 @@ static int fimc_bind(struct device *dev, struct device *master, void *data)
 
 	ctx->drm_dev = drm_dev;
 	ipp->drm_dev = drm_dev;
-	exynos_drm_register_dma(drm_dev, dev);
+	exynos_drm_register_dma(drm_dev, dev, &ctx->dma_priv);
 
 	exynos_drm_ipp_register(dev, ipp, &ipp_funcs,
 			DRM_EXYNOS_IPP_CAP_CROP | DRM_EXYNOS_IPP_CAP_ROTATE |
@@ -1153,7 +1154,7 @@ static void fimc_unbind(struct device *dev, struct device *master,
 	struct exynos_drm_ipp *ipp = &ctx->ipp;
 
 	exynos_drm_ipp_unregister(dev, ipp);
-	exynos_drm_unregister_dma(drm_dev, dev);
+	exynos_drm_unregister_dma(drm_dev, dev, &ctx->dma_priv);
 }
 
 static const struct component_ops fimc_component_ops = {
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fimd.c b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
index 21aec38702fc..bb67cad8371f 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fimd.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fimd.c
@@ -167,6 +167,7 @@ static struct fimd_driver_data exynos5420_fimd_driver_data = {
 struct fimd_context {
 	struct device			*dev;
 	struct drm_device		*drm_dev;
+	void				*dma_priv;
 	struct exynos_drm_crtc		*crtc;
 	struct exynos_drm_plane		planes[WINDOWS_NR];
 	struct exynos_drm_plane_config	configs[WINDOWS_NR];
@@ -1090,7 +1091,7 @@ static int fimd_bind(struct device *dev, struct device *master, void *data)
 	if (is_drm_iommu_supported(drm_dev))
 		fimd_clear_channels(ctx->crtc);
 
-	return exynos_drm_register_dma(drm_dev, dev);
+	return exynos_drm_register_dma(drm_dev, dev, &ctx->dma_priv);
 }
 
 static void fimd_unbind(struct device *dev, struct device *master,
@@ -1100,7 +1101,7 @@ static void fimd_unbind(struct device *dev, struct device *master,
 
 	fimd_atomic_disable(ctx->crtc);
 
-	exynos_drm_unregister_dma(ctx->drm_dev, ctx->dev);
+	exynos_drm_unregister_dma(ctx->drm_dev, ctx->dev, &ctx->dma_priv);
 
 	if (ctx->encoder)
 		exynos_dpi_remove(ctx->encoder);
diff --git a/drivers/gpu/drm/exynos/exynos_drm_g2d.c b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
index 2a3382d43bc9..fcee33a43aca 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_g2d.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_g2d.c
@@ -232,6 +232,7 @@ struct g2d_runqueue_node {
 
 struct g2d_data {
 	struct device			*dev;
+	void				*dma_priv;
 	struct clk			*gate_clk;
 	void __iomem			*regs;
 	int				irq;
@@ -1409,7 +1410,7 @@ static int g2d_bind(struct device *dev, struct device *master, void *data)
 		return ret;
 	}
 
-	ret = exynos_drm_register_dma(drm_dev, dev);
+	ret = exynos_drm_register_dma(drm_dev, dev, &g2d->dma_priv);
 	if (ret < 0) {
 		dev_err(dev, "failed to enable iommu.\n");
 		g2d_fini_cmdlist(g2d);
@@ -1434,7 +1435,7 @@ static void g2d_unbind(struct device *dev, struct device *master, void *data)
 	priv->g2d_dev = NULL;
 
 	cancel_work_sync(&g2d->runqueue_work);
-	exynos_drm_unregister_dma(g2d->drm_dev, dev);
+	exynos_drm_unregister_dma(g2d->drm_dev, dev, &g2d->dma_priv);
 }
 
 static const struct component_ops g2d_component_ops = {
diff --git a/drivers/gpu/drm/exynos/exynos_drm_gsc.c b/drivers/gpu/drm/exynos/exynos_drm_gsc.c
index 88b6fcaa20be..45e9aee8366a 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gsc.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gsc.c
@@ -97,6 +97,7 @@ struct gsc_scaler {
 struct gsc_context {
 	struct exynos_drm_ipp ipp;
 	struct drm_device *drm_dev;
+	void		*dma_priv;
 	struct device	*dev;
 	struct exynos_drm_ipp_task	*task;
 	struct exynos_drm_ipp_formats	*formats;
@@ -1169,7 +1170,7 @@ static int gsc_bind(struct device *dev, struct device *master, void *data)
 
 	ctx->drm_dev = drm_dev;
 	ctx->drm_dev = drm_dev;
-	exynos_drm_register_dma(drm_dev, dev);
+	exynos_drm_register_dma(drm_dev, dev, &ctx->dma_priv);
 
 	exynos_drm_ipp_register(dev, ipp, &ipp_funcs,
 			DRM_EXYNOS_IPP_CAP_CROP | DRM_EXYNOS_IPP_CAP_ROTATE |
@@ -1189,7 +1190,7 @@ static void gsc_unbind(struct device *dev, struct device *master,
 	struct exynos_drm_ipp *ipp = &ctx->ipp;
 
 	exynos_drm_ipp_unregister(dev, ipp);
-	exynos_drm_unregister_dma(drm_dev, dev);
+	exynos_drm_unregister_dma(drm_dev, dev, &ctx->dma_priv);
 }
 
 static const struct component_ops gsc_component_ops = {
diff --git a/drivers/gpu/drm/exynos/exynos_drm_rotator.c b/drivers/gpu/drm/exynos/exynos_drm_rotator.c
index b98482990d1a..dafa87b82052 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_rotator.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_rotator.c
@@ -56,6 +56,7 @@ struct rot_variant {
 struct rot_context {
 	struct exynos_drm_ipp ipp;
 	struct drm_device *drm_dev;
+	void		*dma_priv;
 	struct device	*dev;
 	void __iomem	*regs;
 	struct clk	*clock;
@@ -243,7 +244,7 @@ static int rotator_bind(struct device *dev, struct device *master, void *data)
 
 	rot->drm_dev = drm_dev;
 	ipp->drm_dev = drm_dev;
-	exynos_drm_register_dma(drm_dev, dev);
+	exynos_drm_register_dma(drm_dev, dev, &rot->dma_priv);
 
 	exynos_drm_ipp_register(dev, ipp, &ipp_funcs,
 			   DRM_EXYNOS_IPP_CAP_CROP | DRM_EXYNOS_IPP_CAP_ROTATE,
@@ -261,7 +262,7 @@ static void rotator_unbind(struct device *dev, struct device *master,
 	struct exynos_drm_ipp *ipp = &rot->ipp;
 
 	exynos_drm_ipp_unregister(dev, ipp);
-	exynos_drm_unregister_dma(rot->drm_dev, rot->dev);
+	exynos_drm_unregister_dma(rot->drm_dev, rot->dev, &rot->dma_priv);
 }
 
 static const struct component_ops rotator_component_ops = {
diff --git a/drivers/gpu/drm/exynos/exynos_drm_scaler.c b/drivers/gpu/drm/exynos/exynos_drm_scaler.c
index 497973e9b2c5..93c43c8d914e 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_scaler.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_scaler.c
@@ -39,6 +39,7 @@ struct scaler_data {
 struct scaler_context {
 	struct exynos_drm_ipp		ipp;
 	struct drm_device		*drm_dev;
+	void				*dma_priv;
 	struct device			*dev;
 	void __iomem			*regs;
 	struct clk			*clock[SCALER_MAX_CLK];
@@ -450,7 +451,7 @@ static int scaler_bind(struct device *dev, struct device *master, void *data)
 
 	scaler->drm_dev = drm_dev;
 	ipp->drm_dev = drm_dev;
-	exynos_drm_register_dma(drm_dev, dev);
+	exynos_drm_register_dma(drm_dev, dev, &scaler->dma_priv);
 
 	exynos_drm_ipp_register(dev, ipp, &ipp_funcs,
 			DRM_EXYNOS_IPP_CAP_CROP | DRM_EXYNOS_IPP_CAP_ROTATE |
@@ -470,7 +471,8 @@ static void scaler_unbind(struct device *dev, struct device *master,
 	struct exynos_drm_ipp *ipp = &scaler->ipp;
 
 	exynos_drm_ipp_unregister(dev, ipp);
-	exynos_drm_unregister_dma(scaler->drm_dev, scaler->dev);
+	exynos_drm_unregister_dma(scaler->drm_dev, scaler->dev,
+				  &scaler->dma_priv);
 }
 
 static const struct component_ops scaler_component_ops = {
diff --git a/drivers/gpu/drm/exynos/exynos_mixer.c b/drivers/gpu/drm/exynos/exynos_mixer.c
index 38ae9c32feef..21b726baedea 100644
--- a/drivers/gpu/drm/exynos/exynos_mixer.c
+++ b/drivers/gpu/drm/exynos/exynos_mixer.c
@@ -94,6 +94,7 @@ struct mixer_context {
 	struct platform_device *pdev;
 	struct device		*dev;
 	struct drm_device	*drm_dev;
+	void			*dma_priv;
 	struct exynos_drm_crtc	*crtc;
 	struct exynos_drm_plane	planes[MIXER_WIN_NR];
 	unsigned long		flags;
@@ -894,12 +895,14 @@ static int mixer_initialize(struct mixer_context *mixer_ctx,
 		}
 	}
 
-	return exynos_drm_register_dma(drm_dev, mixer_ctx->dev);
+	return exynos_drm_register_dma(drm_dev, mixer_ctx->dev,
+				       &mixer_ctx->dma_priv);
 }
 
 static void mixer_ctx_remove(struct mixer_context *mixer_ctx)
 {
-	exynos_drm_unregister_dma(mixer_ctx->drm_dev, mixer_ctx->dev);
+	exynos_drm_unregister_dma(mixer_ctx->drm_dev, mixer_ctx->dev,
+				  &mixer_ctx->dma_priv);
 }
 
 static int mixer_enable_vblank(struct exynos_drm_crtc *crtc)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
index 60c984e10c4a..7643a30ba4cd 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c
@@ -423,7 +423,8 @@ eb_validate_vma(struct i915_execbuffer *eb,
 	if (unlikely(entry->flags & eb->invalid_flags))
 		return -EINVAL;
 
-	if (unlikely(entry->alignment && !is_power_of_2(entry->alignment)))
+	if (unlikely(entry->alignment &&
+		     !is_power_of_2_u64(entry->alignment)))
 		return -EINVAL;
 
 	/*
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c
index fe8a59aaa629..940e7f7df69a 100644
--- a/drivers/gpu/drm/i915/gt/intel_lrc.c
+++ b/drivers/gpu/drm/i915/gt/intel_lrc.c
@@ -1679,11 +1679,9 @@ need_timeslice(struct intel_engine_cs *engine, const struct i915_request *rq)
 	if (!intel_engine_has_timeslices(engine))
 		return false;
 
-	if (list_is_last(&rq->sched.link, &engine->active.requests))
-		return false;
-
-	hint = max(rq_prio(list_next_entry(rq, sched.link)),
-		   engine->execlists.queue_priority_hint);
+	hint = engine->execlists.queue_priority_hint;
+	if (!list_is_last(&rq->sched.link, &engine->active.requests))
+		hint = max(hint, rq_prio(list_next_entry(rq, sched.link)));
 
 	return hint >= effective_prio(rq);
 }
@@ -1725,6 +1723,18 @@ static void set_timeslice(struct intel_engine_cs *engine)
 	set_timer_ms(&engine->execlists.timer, active_timeslice(engine));
 }
 
+static void start_timeslice(struct intel_engine_cs *engine)
+{
+	struct intel_engine_execlists *execlists = &engine->execlists;
+
+	execlists->switch_priority_hint = execlists->queue_priority_hint;
+
+	if (timer_pending(&execlists->timer))
+		return;
+
+	set_timer_ms(&execlists->timer, timeslice(engine));
+}
+
 static void record_preemption(struct intel_engine_execlists *execlists)
 {
 	(void)I915_SELFTEST_ONLY(execlists->preempt_hang.count++);
@@ -1888,11 +1898,7 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 				 * Even if ELSP[1] is occupied and not worthy
 				 * of timeslices, our queue might be.
 				 */
-				if (!execlists->timer.expires &&
-				    need_timeslice(engine, last))
-					set_timer_ms(&execlists->timer,
-						     timeslice(engine));
-
+				start_timeslice(engine);
 				return;
 			}
 		}
@@ -1927,7 +1933,8 @@ static void execlists_dequeue(struct intel_engine_cs *engine)
 
 			if (last && !can_merge_rq(last, rq)) {
 				spin_unlock(&ve->base.active.lock);
-				return; /* leave this for another */
+				start_timeslice(engine);
+				return; /* leave this for another sibling */
 			}
 
 			ENGINE_TRACE(engine,
diff --git a/drivers/gpu/drm/i915/gt/intel_timeline.c b/drivers/gpu/drm/i915/gt/intel_timeline.c
index 87716529cd2f..d8d9f1179c2b 100644
--- a/drivers/gpu/drm/i915/gt/intel_timeline.c
+++ b/drivers/gpu/drm/i915/gt/intel_timeline.c
@@ -192,11 +192,15 @@ static void cacheline_release(struct intel_timeline_cacheline *cl)
 
 static void cacheline_free(struct intel_timeline_cacheline *cl)
 {
+	if (!i915_active_acquire_if_busy(&cl->active)) {
+		__idle_cacheline_free(cl);
+		return;
+	}
+
 	GEM_BUG_ON(ptr_test_bit(cl->vaddr, CACHELINE_FREE));
 	cl->vaddr = ptr_set_bit(cl->vaddr, CACHELINE_FREE);
 
-	if (i915_active_is_idle(&cl->active))
-		__idle_cacheline_free(cl);
+	i915_active_release(&cl->active);
 }
 
 int intel_timeline_init(struct intel_timeline *timeline,
diff --git a/drivers/gpu/drm/i915/gvt/display.c b/drivers/gpu/drm/i915/gvt/display.c
index e1c313da6c00..a62bdf9be682 100644
--- a/drivers/gpu/drm/i915/gvt/display.c
+++ b/drivers/gpu/drm/i915/gvt/display.c
@@ -457,7 +457,8 @@ void intel_vgpu_emulate_hotplug(struct intel_vgpu *vgpu, bool connected)
 	struct drm_i915_private *dev_priv = vgpu->gvt->dev_priv;
 
 	/* TODO: add more platforms support */
-	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) {
+	if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv) ||
+		IS_COFFEELAKE(dev_priv)) {
 		if (connected) {
 			vgpu_vreg_t(vgpu, SFUSE_STRAP) |=
 				SFUSE_STRAP_DDID_DETECTED;
diff --git a/drivers/gpu/drm/i915/gvt/opregion.c b/drivers/gpu/drm/i915/gvt/opregion.c
index 867e7629025b..33569b910ed5 100644
--- a/drivers/gpu/drm/i915/gvt/opregion.c
+++ b/drivers/gpu/drm/i915/gvt/opregion.c
@@ -147,15 +147,14 @@ static void virt_vbt_generation(struct vbt *v)
 	/* there's features depending on version! */
 	v->header.version = 155;
 	v->header.header_size = sizeof(v->header);
-	v->header.vbt_size = sizeof(struct vbt) - sizeof(v->header);
+	v->header.vbt_size = sizeof(struct vbt);
 	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 = 33 */
 	v->bdb_header.header_size = sizeof(v->bdb_header);
 
-	v->bdb_header.bdb_size = sizeof(struct vbt) - sizeof(struct vbt_header)
-		- sizeof(struct bdb_header);
+	v->bdb_header.bdb_size = sizeof(struct vbt) - sizeof(struct vbt_header);
 
 	/* general features */
 	v->general_features_header.id = BDB_GENERAL_FEATURES;
diff --git a/drivers/gpu/drm/i915/gvt/vgpu.c b/drivers/gpu/drm/i915/gvt/vgpu.c
index 487af6ea9972..345c2aa3b491 100644
--- a/drivers/gpu/drm/i915/gvt/vgpu.c
+++ b/drivers/gpu/drm/i915/gvt/vgpu.c
@@ -272,10 +272,17 @@ void intel_gvt_destroy_vgpu(struct intel_vgpu *vgpu)
 {
 	struct intel_gvt *gvt = vgpu->gvt;
 
-	mutex_lock(&vgpu->vgpu_lock);
-
 	WARN(vgpu->active, "vGPU is still active!\n");
 
+	/*
+	 * remove idr first so later clean can judge if need to stop
+	 * service if no active vgpu.
+	 */
+	mutex_lock(&gvt->lock);
+	idr_remove(&gvt->vgpu_idr, vgpu->id);
+	mutex_unlock(&gvt->lock);
+
+	mutex_lock(&vgpu->vgpu_lock);
 	intel_gvt_debugfs_remove_vgpu(vgpu);
 	intel_vgpu_clean_sched_policy(vgpu);
 	intel_vgpu_clean_submission(vgpu);
@@ -290,7 +297,6 @@ void intel_gvt_destroy_vgpu(struct intel_vgpu *vgpu)
 	mutex_unlock(&vgpu->vgpu_lock);
 
 	mutex_lock(&gvt->lock);
-	idr_remove(&gvt->vgpu_idr, vgpu->id);
 	if (idr_is_empty(&gvt->vgpu_idr))
 		intel_gvt_clean_irq(gvt);
 	intel_gvt_update_vgpu_types(gvt);
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index dcaa85a91090..a18b2a244706 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -527,19 +527,31 @@ submit_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
 	return NOTIFY_DONE;
 }
 
+static void irq_semaphore_cb(struct irq_work *wrk)
+{
+	struct i915_request *rq =
+		container_of(wrk, typeof(*rq), semaphore_work);
+
+	i915_schedule_bump_priority(rq, I915_PRIORITY_NOSEMAPHORE);
+	i915_request_put(rq);
+}
+
 static int __i915_sw_fence_call
 semaphore_notify(struct i915_sw_fence *fence, enum i915_sw_fence_notify state)
 {
-	struct i915_request *request =
-		container_of(fence, typeof(*request), semaphore);
+	struct i915_request *rq = container_of(fence, typeof(*rq), semaphore);
 
 	switch (state) {
 	case FENCE_COMPLETE:
-		i915_schedule_bump_priority(request, I915_PRIORITY_NOSEMAPHORE);
+		if (!(READ_ONCE(rq->sched.attr.priority) & I915_PRIORITY_NOSEMAPHORE)) {
+			i915_request_get(rq);
+			init_irq_work(&rq->semaphore_work, irq_semaphore_cb);
+			irq_work_queue(&rq->semaphore_work);
+		}
 		break;
 
 	case FENCE_FREE:
-		i915_request_put(request);
+		i915_request_put(rq);
 		break;
 	}
 
@@ -776,8 +788,8 @@ i915_request_await_start(struct i915_request *rq, struct i915_request *signal)
 	struct dma_fence *fence;
 	int err;
 
-	GEM_BUG_ON(i915_request_timeline(rq) ==
-		   rcu_access_pointer(signal->timeline));
+	if (i915_request_timeline(rq) == rcu_access_pointer(signal->timeline))
+		return 0;
 
 	if (i915_request_started(signal))
 		return 0;
@@ -821,7 +833,7 @@ i915_request_await_start(struct i915_request *rq, struct i915_request *signal)
 		return 0;
 
 	err = 0;
-	if (intel_timeline_sync_is_later(i915_request_timeline(rq), fence))
+	if (!intel_timeline_sync_is_later(i915_request_timeline(rq), fence))
 		err = i915_sw_fence_await_dma_fence(&rq->submit,
 						    fence, 0,
 						    I915_FENCE_GFP);
@@ -1318,9 +1330,9 @@ void __i915_request_queue(struct i915_request *rq,
 	 * decide whether to preempt the entire chain so that it is ready to
 	 * run at the earliest possible convenience.
 	 */
-	i915_sw_fence_commit(&rq->semaphore);
 	if (attr && rq->engine->schedule)
 		rq->engine->schedule(rq, attr);
+	i915_sw_fence_commit(&rq->semaphore);
 	i915_sw_fence_commit(&rq->submit);
 }
 
diff --git a/drivers/gpu/drm/i915/i915_request.h b/drivers/gpu/drm/i915/i915_request.h
index f57eadcf3583..fccc339949ec 100644
--- a/drivers/gpu/drm/i915/i915_request.h
+++ b/drivers/gpu/drm/i915/i915_request.h
@@ -26,6 +26,7 @@
 #define I915_REQUEST_H
 
 #include <linux/dma-fence.h>
+#include <linux/irq_work.h>
 #include <linux/lockdep.h>
 
 #include "gem/i915_gem_context_types.h"
@@ -208,6 +209,7 @@ struct i915_request {
 	};
 	struct list_head execute_cb;
 	struct i915_sw_fence semaphore;
+	struct irq_work semaphore_work;
 
 	/*
 	 * A list of everyone we wait upon, and everyone who waits upon us.
diff --git a/drivers/gpu/drm/i915/i915_utils.h b/drivers/gpu/drm/i915/i915_utils.h
index b0ade76bec90..d34141f7dcd8 100644
--- a/drivers/gpu/drm/i915/i915_utils.h
+++ b/drivers/gpu/drm/i915/i915_utils.h
@@ -234,6 +234,11 @@ static inline u64 ptr_to_u64(const void *ptr)
 	__idx;								\
 })
 
+static inline bool is_power_of_2_u64(u64 n)
+{
+	return (n != 0 && ((n & (n - 1)) == 0));
+}
+
 static inline void __list_del_many(struct list_head *head,
 				   struct list_head *first)
 {
diff --git a/include/drm/drm_dp_mst_helper.h b/include/drm/drm_dp_mst_helper.h
index bcb39da9adb4..41725d88d27e 100644
--- a/include/drm/drm_dp_mst_helper.h
+++ b/include/drm/drm_dp_mst_helper.h
@@ -81,7 +81,7 @@ struct drm_dp_vcpi {
  * &drm_dp_mst_topology_mgr.base.lock.
  * @num_sdp_stream_sinks: Number of stream sinks. Protected by
  * &drm_dp_mst_topology_mgr.base.lock.
- * @available_pbn: Available bandwidth for this port. Protected by
+ * @full_pbn: Max possible bandwidth for this port. Protected by
  * &drm_dp_mst_topology_mgr.base.lock.
  * @next: link to next port on this branch device
  * @aux: i2c aux transport to talk to device connected to this port, protected
@@ -126,7 +126,7 @@ struct drm_dp_mst_port {
 	u8 dpcd_rev;
 	u8 num_sdp_streams;
 	u8 num_sdp_stream_sinks;
-	uint16_t available_pbn;
+	uint16_t full_pbn;
 	struct list_head next;
 	/**
 	 * @mstb: the branch device connected to this port, if there is one.