Merge remote-tracking branch 'drm-xe/drm-xe-next' into drm-tip

# Conflicts:
#	drivers/gpu/drm/xe/xe_exec_queue.c

30 files changed, 558 insertions, 273 deletions

diff --git a/drivers/gpu/drm/i915/display/intel_cx0_phy.c b/drivers/gpu/drm/i915/display/intel_cx0_phy.c
index 5ebc90d210d4..e768dc6a15b3 100644
--- a/drivers/gpu/drm/i915/display/intel_cx0_phy.c
+++ b/drivers/gpu/drm/i915/display/intel_cx0_phy.c
@@ -3068,7 +3068,10 @@ int intel_mtl_tbt_calc_port_clock(struct intel_encoder *encoder)
 
 	val = intel_de_read(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port));
 
-	clock = REG_FIELD_GET(XELPDP_DDI_CLOCK_SELECT_MASK, val);
+	if (DISPLAY_VER(display) >= 30)
+		clock = REG_FIELD_GET(XE3_DDI_CLOCK_SELECT_MASK, val);
+	else
+		clock = REG_FIELD_GET(XELPDP_DDI_CLOCK_SELECT_MASK, val);
 
 	drm_WARN_ON(display->drm, !(val & XELPDP_FORWARD_CLOCK_UNGATE));
 	drm_WARN_ON(display->drm, !(val & XELPDP_TBT_CLOCK_REQUEST));
@@ -3083,13 +3086,18 @@ int intel_mtl_tbt_calc_port_clock(struct intel_encoder *encoder)
 		return 540000;
 	case XELPDP_DDI_CLOCK_SELECT_TBT_810:
 		return 810000;
+	case XELPDP_DDI_CLOCK_SELECT_TBT_312_5:
+		return 1000000;
+	case XELPDP_DDI_CLOCK_SELECT_TBT_625:
+		return 2000000;
 	default:
 		MISSING_CASE(clock);
 		return 162000;
 	}
 }
 
-static int intel_mtl_tbt_clock_select(int clock)
+static int intel_mtl_tbt_clock_select(struct intel_display *display,
+				      int clock)
 {
 	switch (clock) {
 	case 162000:
@@ -3100,6 +3108,18 @@ static int intel_mtl_tbt_clock_select(int clock)
 		return XELPDP_DDI_CLOCK_SELECT_TBT_540;
 	case 810000:
 		return XELPDP_DDI_CLOCK_SELECT_TBT_810;
+	case 1000000:
+		if (DISPLAY_VER(display) < 30) {
+			drm_WARN_ON(display->drm, "UHBR10 not supported for the platform\n");
+			return XELPDP_DDI_CLOCK_SELECT_TBT_162;
+		}
+		return XELPDP_DDI_CLOCK_SELECT_TBT_312_5;
+	case 2000000:
+		if (DISPLAY_VER(display) < 30) {
+			drm_WARN_ON(display->drm, "UHBR20 not supported for the platform\n");
+			return XELPDP_DDI_CLOCK_SELECT_TBT_162;
+		}
+		return XELPDP_DDI_CLOCK_SELECT_TBT_625;
 	default:
 		MISSING_CASE(clock);
 		return XELPDP_DDI_CLOCK_SELECT_TBT_162;
@@ -3112,15 +3132,26 @@ static void intel_mtl_tbt_pll_enable(struct intel_encoder *encoder,
 	struct intel_display *display = to_intel_display(encoder);
 	enum phy phy = intel_encoder_to_phy(encoder);
 	u32 val = 0;
+	u32 mask;
 
 	/*
 	 * 1. Program PORT_CLOCK_CTL REGISTER to configure
 	 * clock muxes, gating and SSC
 	 */
-	val |= XELPDP_DDI_CLOCK_SELECT(intel_mtl_tbt_clock_select(crtc_state->port_clock));
+
+	if (DISPLAY_VER(display) >= 30) {
+		mask = XE3_DDI_CLOCK_SELECT_MASK;
+		val |= XE3_DDI_CLOCK_SELECT(intel_mtl_tbt_clock_select(display, crtc_state->port_clock));
+	} else {
+		mask = XELPDP_DDI_CLOCK_SELECT_MASK;
+		val |= XELPDP_DDI_CLOCK_SELECT(intel_mtl_tbt_clock_select(display, crtc_state->port_clock));
+	}
+
+	mask |= XELPDP_FORWARD_CLOCK_UNGATE;
 	val |= XELPDP_FORWARD_CLOCK_UNGATE;
+
 	intel_de_rmw(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port),
-		     XELPDP_DDI_CLOCK_SELECT_MASK | XELPDP_FORWARD_CLOCK_UNGATE, val);
+		     mask, val);
 
 	/* 2. Read back PORT_CLOCK_CTL REGISTER */
 	val = intel_de_read(display, XELPDP_PORT_CLOCK_CTL(display, encoder->port));
diff --git a/drivers/gpu/drm/i915/display/intel_cx0_phy_regs.h b/drivers/gpu/drm/i915/display/intel_cx0_phy_regs.h
index 4dc6e179a774..da154ff26b96 100644
--- a/drivers/gpu/drm/i915/display/intel_cx0_phy_regs.h
+++ b/drivers/gpu/drm/i915/display/intel_cx0_phy_regs.h
@@ -192,7 +192,9 @@
 #define   XELPDP_TBT_CLOCK_REQUEST			REG_BIT(19)
 #define   XELPDP_TBT_CLOCK_ACK				REG_BIT(18)
 #define   XELPDP_DDI_CLOCK_SELECT_MASK			REG_GENMASK(15, 12)
+#define   XE3_DDI_CLOCK_SELECT_MASK			REG_GENMASK(16, 12)
 #define   XELPDP_DDI_CLOCK_SELECT(val)			REG_FIELD_PREP(XELPDP_DDI_CLOCK_SELECT_MASK, val)
+#define   XE3_DDI_CLOCK_SELECT(val)			REG_FIELD_PREP(XE3_DDI_CLOCK_SELECT_MASK, val)
 #define   XELPDP_DDI_CLOCK_SELECT_NONE			0x0
 #define   XELPDP_DDI_CLOCK_SELECT_MAXPCLK		0x8
 #define   XELPDP_DDI_CLOCK_SELECT_DIV18CLK		0x9
@@ -200,6 +202,8 @@
 #define   XELPDP_DDI_CLOCK_SELECT_TBT_270		0xd
 #define   XELPDP_DDI_CLOCK_SELECT_TBT_540		0xe
 #define   XELPDP_DDI_CLOCK_SELECT_TBT_810		0xf
+#define   XELPDP_DDI_CLOCK_SELECT_TBT_312_5		0x18
+#define   XELPDP_DDI_CLOCK_SELECT_TBT_625		0x19
 #define   XELPDP_FORWARD_CLOCK_UNGATE			REG_BIT(10)
 #define   XELPDP_LANE1_PHY_CLOCK_SELECT			REG_BIT(8)
 #define   XELPDP_SSC_ENABLE_PLLA			REG_BIT(1)
diff --git a/drivers/gpu/drm/xe/instructions/xe_mi_commands.h b/drivers/gpu/drm/xe/instructions/xe_mi_commands.h
index 10ec2920d31b..f4ee910f0943 100644
--- a/drivers/gpu/drm/xe/instructions/xe_mi_commands.h
+++ b/drivers/gpu/drm/xe/instructions/xe_mi_commands.h
@@ -33,12 +33,13 @@
 #define MI_TOPOLOGY_FILTER		__MI_INSTR(0xD)
 #define MI_FORCE_WAKEUP			__MI_INSTR(0x1D)
 
-#define MI_STORE_DATA_IMM		__MI_INSTR(0x20)
-#define   MI_SDI_GGTT			REG_BIT(22)
-#define   MI_SDI_LEN_DW			GENMASK(9, 0)
-#define   MI_SDI_NUM_DW(x)		REG_FIELD_PREP(MI_SDI_LEN_DW, (x) + 3 - 2)
-#define   MI_SDI_NUM_QW(x)		(REG_FIELD_PREP(MI_SDI_LEN_DW, 2 * (x) + 3 - 2) | \
-					 REG_BIT(21))
+#define MI_STORE_DATA_IMM			__MI_INSTR(0x20)
+#define   MI_SDI_GGTT				REG_BIT(22)
+#define   MI_FORCE_WRITE_COMPLETION_CHECK	REG_BIT(10)
+#define   MI_SDI_LEN_DW				GENMASK(9, 0)
+#define   MI_SDI_NUM_DW(x)			REG_FIELD_PREP(MI_SDI_LEN_DW, (x) + 3 - 2)
+#define   MI_SDI_NUM_QW(x)			(REG_FIELD_PREP(MI_SDI_LEN_DW, 2 * (x) + 3 - 2) | \
+						 REG_BIT(21))
 
 #define MI_LOAD_REGISTER_IMM		__MI_INSTR(0x22)
 #define   MI_LRI_LRM_CS_MMIO		REG_BIT(19)
diff --git a/drivers/gpu/drm/xe/regs/xe_engine_regs.h b/drivers/gpu/drm/xe/regs/xe_engine_regs.h
index 7c78496e6213..d86219dedde2 100644
--- a/drivers/gpu/drm/xe/regs/xe_engine_regs.h
+++ b/drivers/gpu/drm/xe/regs/xe_engine_regs.h
@@ -83,6 +83,8 @@
 #define RING_IMR(base)				XE_REG((base) + 0xa8)
 #define RING_INT_STATUS_RPT_PTR(base)		XE_REG((base) + 0xac)
 
+#define CS_INT_VEC(base)			XE_REG((base) + 0x1b8)
+
 #define RING_EIR(base)				XE_REG((base) + 0xb0)
 #define RING_EMR(base)				XE_REG((base) + 0xb4)
 #define RING_ESR(base)				XE_REG((base) + 0xb8)
@@ -138,6 +140,7 @@
 
 #define RING_MODE(base)				XE_REG((base) + 0x29c)
 #define   GFX_DISABLE_LEGACY_MODE		REG_BIT(3)
+#define   GFX_MSIX_INTERRUPT_ENABLE		REG_BIT(13)
 
 #define RING_TIMESTAMP(base)			XE_REG((base) + 0x358)
 
diff --git a/drivers/gpu/drm/xe/regs/xe_lrc_layout.h b/drivers/gpu/drm/xe/regs/xe_lrc_layout.h
index 045dfd09db99..57944f90bbf6 100644
--- a/drivers/gpu/drm/xe/regs/xe_lrc_layout.h
+++ b/drivers/gpu/drm/xe/regs/xe_lrc_layout.h
@@ -25,6 +25,9 @@
 #define CTX_INT_SRC_REPORT_REG		(CTX_LRI_INT_REPORT_PTR + 3)
 #define CTX_INT_SRC_REPORT_PTR		(CTX_LRI_INT_REPORT_PTR + 4)
 
+#define CTX_CS_INT_VEC_REG		0x5a
+#define CTX_CS_INT_VEC_DATA		(CTX_CS_INT_VEC_REG + 1)
+
 #define INDIRECT_CTX_RING_HEAD		(0x02 + 1)
 #define INDIRECT_CTX_RING_TAIL		(0x04 + 1)
 #define INDIRECT_CTX_RING_START		(0x06 + 1)
diff --git a/drivers/gpu/drm/xe/tests/xe_bo.c b/drivers/gpu/drm/xe/tests/xe_bo.c
index c9ec7a313c6b..405ff904153e 100644
--- a/drivers/gpu/drm/xe/tests/xe_bo.c
+++ b/drivers/gpu/drm/xe/tests/xe_bo.c
@@ -606,8 +606,6 @@ static void xe_bo_shrink_kunit(struct kunit *test)
 static struct kunit_case xe_bo_tests[] = {
 	KUNIT_CASE_PARAM(xe_ccs_migrate_kunit, xe_pci_live_device_gen_param),
 	KUNIT_CASE_PARAM(xe_bo_evict_kunit, xe_pci_live_device_gen_param),
-	KUNIT_CASE_PARAM_ATTR(xe_bo_shrink_kunit, xe_pci_live_device_gen_param,
-			      {.speed = KUNIT_SPEED_SLOW}),
 	{}
 };
 
@@ -618,3 +616,17 @@ struct kunit_suite xe_bo_test_suite = {
 	.init = xe_kunit_helper_xe_device_live_test_init,
 };
 EXPORT_SYMBOL_IF_KUNIT(xe_bo_test_suite);
+
+static struct kunit_case xe_bo_shrink_test[] = {
+	KUNIT_CASE_PARAM_ATTR(xe_bo_shrink_kunit, xe_pci_live_device_gen_param,
+			      {.speed = KUNIT_SPEED_SLOW}),
+	{}
+};
+
+VISIBLE_IF_KUNIT
+struct kunit_suite xe_bo_shrink_test_suite = {
+	.name = "xe_bo_shrink",
+	.test_cases = xe_bo_shrink_test,
+	.init = xe_kunit_helper_xe_device_live_test_init,
+};
+EXPORT_SYMBOL_IF_KUNIT(xe_bo_shrink_test_suite);
diff --git a/drivers/gpu/drm/xe/tests/xe_live_test_mod.c b/drivers/gpu/drm/xe/tests/xe_live_test_mod.c
index 0d36ab864ec0..81277c77016d 100644
--- a/drivers/gpu/drm/xe/tests/xe_live_test_mod.c
+++ b/drivers/gpu/drm/xe/tests/xe_live_test_mod.c
@@ -6,11 +6,13 @@
 #include <kunit/test.h>
 
 extern struct kunit_suite xe_bo_test_suite;
+extern struct kunit_suite xe_bo_shrink_test_suite;
 extern struct kunit_suite xe_dma_buf_test_suite;
 extern struct kunit_suite xe_migrate_test_suite;
 extern struct kunit_suite xe_mocs_test_suite;
 
 kunit_test_suite(xe_bo_test_suite);
+kunit_test_suite(xe_bo_shrink_test_suite);
 kunit_test_suite(xe_dma_buf_test_suite);
 kunit_test_suite(xe_migrate_test_suite);
 kunit_test_suite(xe_mocs_test_suite);
diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index 56d4ffb650da..bf36e4fb4679 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -325,7 +325,9 @@ struct xe_device *xe_device_create(struct pci_dev *pdev,
 	xe->info.revid = pdev->revision;
 	xe->info.force_execlist = xe_modparam.force_execlist;
 
-	spin_lock_init(&xe->irq.lock);
+	err = xe_irq_init(xe);
+	if (err)
+		goto err;
 
 	init_waitqueue_head(&xe->ufence_wq);
 
diff --git a/drivers/gpu/drm/xe/xe_device.h b/drivers/gpu/drm/xe/xe_device.h
index f1fbfe916867..fc3c2af3fb7f 100644
--- a/drivers/gpu/drm/xe/xe_device.h
+++ b/drivers/gpu/drm/xe/xe_device.h
@@ -157,8 +157,7 @@ static inline bool xe_device_has_sriov(struct xe_device *xe)
 
 static inline bool xe_device_has_msix(struct xe_device *xe)
 {
-	/* TODO: change this when MSI-X support is fully integrated */
-	return false;
+	return xe->irq.msix.nvec > 0;
 }
 
 static inline bool xe_device_has_memirq(struct xe_device *xe)
diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
index ace22e35e769..8a7b15972413 100644
--- a/drivers/gpu/drm/xe/xe_device_types.h
+++ b/drivers/gpu/drm/xe/xe_device_types.h
@@ -348,6 +348,14 @@ struct xe_device {
 
 		/** @irq.enabled: interrupts enabled on this device */
 		atomic_t enabled;
+
+		/** @irq.msix: irq info for platforms that support MSI-X */
+		struct {
+			/** @irq.msix.nvec: number of MSI-X interrupts */
+			u16 nvec;
+			/** @irq.msix.indexes: used to allocate MSI-X indexes */
+			struct xarray indexes;
+		} msix;
 	} irq;
 
 	/** @ttm: ttm device */
diff --git a/drivers/gpu/drm/xe/xe_exec_queue.c b/drivers/gpu/drm/xe/xe_exec_queue.c
index 1b4bb083555f..7e1abbbfba12 100644
--- a/drivers/gpu/drm/xe/xe_exec_queue.c
+++ b/drivers/gpu/drm/xe/xe_exec_queue.c
@@ -17,6 +17,7 @@
 #include "xe_hw_engine_class_sysfs.h"
 #include "xe_hw_engine_group.h"
 #include "xe_hw_fence.h"
+#include "xe_irq.h"
 #include "xe_lrc.h"
 #include "xe_macros.h"
 #include "xe_migrate.h"
@@ -69,6 +70,7 @@ static struct xe_exec_queue *__xe_exec_queue_alloc(struct xe_device *xe,
 	q->gt = gt;
 	q->class = hwe->class;
 	q->width = width;
+	q->msix_vec = XE_IRQ_DEFAULT_MSIX;
 	q->logical_mask = logical_mask;
 	q->fence_irq = &gt->fence_irq[hwe->class];
 	q->ring_ops = gt->ring_ops[hwe->class];
@@ -118,7 +120,7 @@ static int __xe_exec_queue_init(struct xe_exec_queue *q)
 	}
 
 	for (i = 0; i < q->width; ++i) {
-		q->lrc[i] = xe_lrc_create(q->hwe, q->vm, SZ_16K);
+		q->lrc[i] = xe_lrc_create(q->hwe, q->vm, SZ_16K, q->msix_vec);
 		if (IS_ERR(q->lrc[i])) {
 			err = PTR_ERR(q->lrc[i]);
 			goto err_unlock;
@@ -768,25 +770,20 @@ bool xe_exec_queue_is_idle(struct xe_exec_queue *q)
 void xe_exec_queue_update_run_ticks(struct xe_exec_queue *q)
 {
 	struct xe_device *xe = gt_to_xe(q->gt);
-	struct xe_file *xef;
 	struct xe_lrc *lrc;
 	u32 old_ts, new_ts;
 	int idx;
 
 	/*
-	 * Jobs that are run during driver load may use an exec_queue, but are
-	 * not associated with a user xe file, so avoid accumulating busyness
-	 * for kernel specific work.
+	 * Jobs that are executed by kernel doesn't have a corresponding xe_file
+	 * and thus are not accounted.
 	 */
-	if (!q->vm || !q->vm->xef)
+	if (!q->xef)
 		return;
 
 	/* Synchronize with unbind while holding the xe file open */
 	if (!drm_dev_enter(&xe->drm, &idx))
 		return;
-
-	xef = q->vm->xef;
-
 	/*
 	 * Only sample the first LRC. For parallel submission, all of them are
 	 * scheduled together and we compensate that below by multiplying by
@@ -797,7 +794,7 @@ void xe_exec_queue_update_run_ticks(struct xe_exec_queue *q)
 	 */
 	lrc = q->lrc[0];
 	new_ts = xe_lrc_update_timestamp(lrc, &old_ts);
-	xef->run_ticks[q->class] += (new_ts - old_ts) * q->width;
+	q->xef->run_ticks[q->class] += (new_ts - old_ts) * q->width;
 
 	drm_dev_exit(idx);
 }
diff --git a/drivers/gpu/drm/xe/xe_exec_queue_types.h b/drivers/gpu/drm/xe/xe_exec_queue_types.h
index 1158b6062a6c..eec8f9935a58 100644
--- a/drivers/gpu/drm/xe/xe_exec_queue_types.h
+++ b/drivers/gpu/drm/xe/xe_exec_queue_types.h
@@ -63,6 +63,8 @@ struct xe_exec_queue {
 	char name[MAX_FENCE_NAME_LEN];
 	/** @width: width (number BB submitted per exec) of this exec queue */
 	u16 width;
+	/** @msix_vec: MSI-X vector (for platforms that support it) */
+	u16 msix_vec;
 	/** @fence_irq: fence IRQ used to signal job completion */
 	struct xe_hw_fence_irq *fence_irq;
 
diff --git a/drivers/gpu/drm/xe/xe_execlist.c b/drivers/gpu/drm/xe/xe_execlist.c
index a8c416a48812..5ef96deaa881 100644
--- a/drivers/gpu/drm/xe/xe_execlist.c
+++ b/drivers/gpu/drm/xe/xe_execlist.c
@@ -17,6 +17,7 @@
 #include "xe_exec_queue.h"
 #include "xe_gt.h"
 #include "xe_hw_fence.h"
+#include "xe_irq.h"
 #include "xe_lrc.h"
 #include "xe_macros.h"
 #include "xe_mmio.h"
@@ -47,6 +48,7 @@ static void __start_lrc(struct xe_hw_engine *hwe, struct xe_lrc *lrc,
 	struct xe_mmio *mmio = &gt->mmio;
 	struct xe_device *xe = gt_to_xe(gt);
 	u64 lrc_desc;
+	u32 ring_mode = _MASKED_BIT_ENABLE(GFX_DISABLE_LEGACY_MODE);
 
 	lrc_desc = xe_lrc_descriptor(lrc);
 
@@ -80,8 +82,10 @@ static void __start_lrc(struct xe_hw_engine *hwe, struct xe_lrc *lrc,
 	xe_mmio_write32(mmio, RING_HWS_PGA(hwe->mmio_base),
 			xe_bo_ggtt_addr(hwe->hwsp));
 	xe_mmio_read32(mmio, RING_HWS_PGA(hwe->mmio_base));
-	xe_mmio_write32(mmio, RING_MODE(hwe->mmio_base),
-			_MASKED_BIT_ENABLE(GFX_DISABLE_LEGACY_MODE));
+
+	if (xe_device_has_msix(gt_to_xe(hwe->gt)))
+		ring_mode |= _MASKED_BIT_ENABLE(GFX_MSIX_INTERRUPT_ENABLE);
+	xe_mmio_write32(mmio, RING_MODE(hwe->mmio_base), ring_mode);
 
 	xe_mmio_write32(mmio, RING_EXECLIST_SQ_CONTENTS_LO(hwe->mmio_base),
 			lower_32_bits(lrc_desc));
@@ -265,7 +269,7 @@ struct xe_execlist_port *xe_execlist_port_create(struct xe_device *xe,
 
 	port->hwe = hwe;
 
-	port->lrc = xe_lrc_create(hwe, NULL, SZ_16K);
+	port->lrc = xe_lrc_create(hwe, NULL, SZ_16K, XE_IRQ_DEFAULT_MSIX);
 	if (IS_ERR(port->lrc)) {
 		err = PTR_ERR(port->lrc);
 		goto err;
diff --git a/drivers/gpu/drm/xe/xe_gt_idle.c b/drivers/gpu/drm/xe/xe_gt_idle.c
index fd80afeef56a..ffd3ba7f6656 100644
--- a/drivers/gpu/drm/xe/xe_gt_idle.c
+++ b/drivers/gpu/drm/xe/xe_gt_idle.c
@@ -122,10 +122,12 @@ void xe_gt_idle_enable_pg(struct xe_gt *gt)
 	if (!xe_gt_is_media_type(gt))
 		gtidle->powergate_enable |= RENDER_POWERGATE_ENABLE;
 
-	for (i = XE_HW_ENGINE_VCS0, j = 0; i <= XE_HW_ENGINE_VCS7; ++i, ++j) {
-		if ((gt->info.engine_mask & BIT(i)))
-			gtidle->powergate_enable |= (VDN_HCP_POWERGATE_ENABLE(j) |
-						     VDN_MFXVDENC_POWERGATE_ENABLE(j));
+	if (xe->info.platform != XE_DG1) {
+		for (i = XE_HW_ENGINE_VCS0, j = 0; i <= XE_HW_ENGINE_VCS7; ++i, ++j) {
+			if ((gt->info.engine_mask & BIT(i)))
+				gtidle->powergate_enable |= (VDN_HCP_POWERGATE_ENABLE(j) |
+							     VDN_MFXVDENC_POWERGATE_ENABLE(j));
+		}
 	}
 
 	fw_ref = xe_force_wake_get(gt_to_fw(gt), XE_FW_GT);
diff --git a/drivers/gpu/drm/xe/xe_guc.c b/drivers/gpu/drm/xe/xe_guc.c
index 4e2868efb620..408365dfe4ee 100644
--- a/drivers/gpu/drm/xe/xe_guc.c
+++ b/drivers/gpu/drm/xe/xe_guc.c
@@ -147,6 +147,34 @@ static u32 guc_ctl_ads_flags(struct xe_guc *guc)
 	return flags;
 }
 
+static bool needs_wa_dual_queue(struct xe_gt *gt)
+{
+	/*
+	 * The DUAL_QUEUE_WA tells the GuC to not allow concurrent submissions
+	 * on RCS and CCSes with different address spaces, which on DG2 is
+	 * required as a WA for an HW bug.
+	 */
+	if (XE_WA(gt, 22011391025))
+		return true;
+
+	/*
+	 * On newer platforms, the HW has been updated to not allow parallel
+	 * execution of different address spaces, so the RCS/CCS will stall the
+	 * context switch if one of the other RCS/CCSes is busy with a different
+	 * address space. While functionally correct, having a submission
+	 * stalled on the HW limits the GuC ability to shuffle things around and
+	 * can cause complications if the non-stalled submission runs for a long
+	 * time, because the GuC doesn't know that the stalled submission isn't
+	 * actually running and might declare it as hung. Therefore, we enable
+	 * the DUAL_QUEUE_WA on all newer platforms on GTs that have CCS engines
+	 * to move management back to the GuC.
+	 */
+	if (CCS_MASK(gt) && GRAPHICS_VERx100(gt_to_xe(gt)) >= 1270)
+		return true;
+
+	return false;
+}
+
 static u32 guc_ctl_wa_flags(struct xe_guc *guc)
 {
 	struct xe_device *xe = guc_to_xe(guc);
@@ -159,7 +187,7 @@ static u32 guc_ctl_wa_flags(struct xe_guc *guc)
 	if (XE_WA(gt, 14014475959))
 		flags |= GUC_WA_HOLD_CCS_SWITCHOUT;
 
-	if (XE_WA(gt, 22011391025))
+	if (needs_wa_dual_queue(gt))
 		flags |= GUC_WA_DUAL_QUEUE;
 
 	/*
diff --git a/drivers/gpu/drm/xe/xe_hw_engine.c b/drivers/gpu/drm/xe/xe_hw_engine.c
index b19366744148..ac9c666a9652 100644
--- a/drivers/gpu/drm/xe/xe_hw_engine.c
+++ b/drivers/gpu/drm/xe/xe_hw_engine.c
@@ -324,6 +324,7 @@ void xe_hw_engine_enable_ring(struct xe_hw_engine *hwe)
 {
 	u32 ccs_mask =
 		xe_hw_engine_mask_per_class(hwe->gt, XE_ENGINE_CLASS_COMPUTE);
+	u32 ring_mode = _MASKED_BIT_ENABLE(GFX_DISABLE_LEGACY_MODE);
 
 	if (hwe->class == XE_ENGINE_CLASS_COMPUTE && ccs_mask)
 		xe_mmio_write32(&hwe->gt->mmio, RCU_MODE,
@@ -332,8 +333,10 @@ void xe_hw_engine_enable_ring(struct xe_hw_engine *hwe)
 	xe_hw_engine_mmio_write32(hwe, RING_HWSTAM(0), ~0x0);
 	xe_hw_engine_mmio_write32(hwe, RING_HWS_PGA(0),
 				  xe_bo_ggtt_addr(hwe->hwsp));
-	xe_hw_engine_mmio_write32(hwe, RING_MODE(0),
-				  _MASKED_BIT_ENABLE(GFX_DISABLE_LEGACY_MODE));
+
+	if (xe_device_has_msix(gt_to_xe(hwe->gt)))
+		ring_mode |= _MASKED_BIT_ENABLE(GFX_MSIX_INTERRUPT_ENABLE);
+	xe_hw_engine_mmio_write32(hwe, RING_MODE(0), ring_mode);
 	xe_hw_engine_mmio_write32(hwe, RING_MI_MODE(0),
 				  _MASKED_BIT_DISABLE(STOP_RING));
 	xe_hw_engine_mmio_read32(hwe, RING_MI_MODE(0));
@@ -772,7 +775,7 @@ static void check_gsc_availability(struct xe_gt *gt)
 		xe_mmio_write32(&gt->mmio, GUNIT_GSC_INTR_ENABLE, 0);
 		xe_mmio_write32(&gt->mmio, GUNIT_GSC_INTR_MASK, ~0);
 
-		drm_info(&xe->drm, "gsccs disabled due to lack of FW\n");
+		drm_dbg(&xe->drm, "GSC FW not used, disabling gsccs\n");
 	}
 }
 
diff --git a/drivers/gpu/drm/xe/xe_irq.c b/drivers/gpu/drm/xe/xe_irq.c
index 1c509e66694d..32f5a67a917b 100644
--- a/drivers/gpu/drm/xe/xe_irq.c
+++ b/drivers/gpu/drm/xe/xe_irq.c
@@ -10,6 +10,7 @@
 #include <drm/drm_managed.h>
 
 #include "display/xe_display.h"
+#include "regs/xe_guc_regs.h"
 #include "regs/xe_irq_regs.h"
 #include "xe_device.h"
 #include "xe_drv.h"
@@ -29,6 +30,11 @@
 #define IIR(offset)				XE_REG(offset + 0x8)
 #define IER(offset)				XE_REG(offset + 0xc)
 
+static int xe_irq_msix_init(struct xe_device *xe);
+static void xe_irq_msix_free(struct xe_device *xe);
+static int xe_irq_msix_request_irqs(struct xe_device *xe);
+static void xe_irq_msix_synchronize_irq(struct xe_device *xe);
+
 static void assert_iir_is_zero(struct xe_mmio *mmio, struct xe_reg reg)
 {
 	u32 val = xe_mmio_read32(mmio, reg);
@@ -572,6 +578,11 @@ static void xe_irq_reset(struct xe_device *xe)
 	if (IS_SRIOV_VF(xe))
 		return vf_irq_reset(xe);
 
+	if (xe_device_uses_memirq(xe)) {
+		for_each_tile(tile, xe, id)
+			xe_memirq_reset(&tile->memirq);
+	}
+
 	for_each_tile(tile, xe, id) {
 		if (GRAPHICS_VERx100(xe) >= 1210)
 			dg1_irq_reset(tile);
@@ -614,6 +625,14 @@ static void xe_irq_postinstall(struct xe_device *xe)
 	if (IS_SRIOV_VF(xe))
 		return vf_irq_postinstall(xe);
 
+	if (xe_device_uses_memirq(xe)) {
+		struct xe_tile *tile;
+		unsigned int id;
+
+		for_each_tile(tile, xe, id)
+			xe_memirq_postinstall(&tile->memirq);
+	}
+
 	xe_display_irq_postinstall(xe, xe_root_mmio_gt(xe));
 
 	/*
@@ -656,60 +675,83 @@ static irq_handler_t xe_irq_handler(struct xe_device *xe)
 		return xelp_irq_handler;
 }
 
-static void irq_uninstall(void *arg)
+static int xe_irq_msi_request_irqs(struct xe_device *xe)
+{
+	struct pci_dev *pdev = to_pci_dev(xe->drm.dev);
+	irq_handler_t irq_handler;
+	int irq, err;
+
+	irq_handler = xe_irq_handler(xe);
+	if (!irq_handler) {
+		drm_err(&xe->drm, "No supported interrupt handler");
+		return -EINVAL;
+	}
+
+	irq = pci_irq_vector(pdev, 0);
+	err = request_irq(irq, irq_handler, IRQF_SHARED, DRIVER_NAME, xe);
+	if (err < 0) {
+		drm_err(&xe->drm, "Failed to request MSI IRQ %d\n", err);
+		return err;
+	}
+
+	return 0;
+}
+
+static void xe_irq_msi_free(struct xe_device *xe)
 {
-	struct xe_device *xe = arg;
 	struct pci_dev *pdev = to_pci_dev(xe->drm.dev);
 	int irq;
 
+	irq = pci_irq_vector(pdev, 0);
+	free_irq(irq, xe);
+}
+
+static void irq_uninstall(void *arg)
+{
+	struct xe_device *xe = arg;
+
 	if (!atomic_xchg(&xe->irq.enabled, 0))
 		return;
 
 	xe_irq_reset(xe);
 
-	irq = pci_irq_vector(pdev, 0);
-	free_irq(irq, xe);
+	if (xe_device_has_msix(xe))
+		xe_irq_msix_free(xe);
+	else
+		xe_irq_msi_free(xe);
+}
+
+int xe_irq_init(struct xe_device *xe)
+{
+	spin_lock_init(&xe->irq.lock);
+
+	return xe_irq_msix_init(xe);
 }
 
 int xe_irq_install(struct xe_device *xe)
 {
 	struct pci_dev *pdev = to_pci_dev(xe->drm.dev);
-	unsigned int irq_flags = PCI_IRQ_MSIX;
-	irq_handler_t irq_handler;
-	int err, irq, nvec;
-
-	irq_handler = xe_irq_handler(xe);
-	if (!irq_handler) {
-		drm_err(&xe->drm, "No supported interrupt handler");
-		return -EINVAL;
-	}
+	unsigned int irq_flags = PCI_IRQ_MSI;
+	int nvec = 1;
+	int err;
 
 	xe_irq_reset(xe);
 
-	nvec = pci_msix_vec_count(pdev);
-	if (nvec <= 0) {
-		if (nvec == -EINVAL) {
-			/* MSIX capability is not supported in the device, using MSI */
-			irq_flags = PCI_IRQ_MSI;
-			nvec = 1;
-		} else {
-			drm_err(&xe->drm, "MSIX: Failed getting count\n");
-			return nvec;
-		}
+	if (xe_device_has_msix(xe)) {
+		nvec = xe->irq.msix.nvec;
+		irq_flags = PCI_IRQ_MSIX;
 	}
 
 	err = pci_alloc_irq_vectors(pdev, nvec, nvec, irq_flags);
 	if (err < 0) {
-		drm_err(&xe->drm, "MSI/MSIX: Failed to enable support %d\n", err);
+		drm_err(&xe->drm, "Failed to allocate IRQ vectors: %d\n", err);
 		return err;
 	}
 
-	irq = pci_irq_vector(pdev, 0);
-	err = request_irq(irq, irq_handler, IRQF_SHARED, DRIVER_NAME, xe);
-	if (err < 0) {
-		drm_err(&xe->drm, "Failed to request MSI/MSIX IRQ %d\n", err);
+	err = xe_device_has_msix(xe) ? xe_irq_msix_request_irqs(xe) :
+					xe_irq_msi_request_irqs(xe);
+	if (err)
 		return err;
-	}
 
 	atomic_set(&xe->irq.enabled, 1);
 
@@ -722,18 +764,28 @@ int xe_irq_install(struct xe_device *xe)
 	return 0;
 
 free_irq_handler:
-	free_irq(irq, xe);
+	if (xe_device_has_msix(xe))
+		xe_irq_msix_free(xe);
+	else
+		xe_irq_msi_free(xe);
 
 	return err;
 }
 
-void xe_irq_suspend(struct xe_device *xe)
+static void xe_irq_msi_synchronize_irq(struct xe_device *xe)
 {
-	int irq = to_pci_dev(xe->drm.dev)->irq;
+	synchronize_irq(to_pci_dev(xe->drm.dev)->irq);
+}
 
+void xe_irq_suspend(struct xe_device *xe)
+{
 	atomic_set(&xe->irq.enabled, 0); /* no new irqs */
 
-	synchronize_irq(irq); /* flush irqs */
+	/* flush irqs */
+	if (xe_device_has_msix(xe))
+		xe_irq_msix_synchronize_irq(xe);
+	else
+		xe_irq_msi_synchronize_irq(xe);
 	xe_irq_reset(xe); /* turn irqs off */
 }
 
@@ -754,3 +806,198 @@ void xe_irq_resume(struct xe_device *xe)
 	for_each_gt(gt, xe, id)
 		xe_irq_enable_hwe(gt);
 }
+
+/* MSI-X related definitions and functions below. */
+
+enum xe_irq_msix_static {
+	GUC2HOST_MSIX = 0,
+	DEFAULT_MSIX = XE_IRQ_DEFAULT_MSIX,
+	/* Must be last */
+	NUM_OF_STATIC_MSIX,
+};
+
+static int xe_irq_msix_init(struct xe_device *xe)
+{
+	struct pci_dev *pdev = to_pci_dev(xe->drm.dev);
+	int nvec = pci_msix_vec_count(pdev);
+
+	if (nvec == -EINVAL)
+		return 0;  /* MSI */
+
+	if (nvec < 0) {
+		drm_err(&xe->drm, "Failed getting MSI-X vectors count: %d\n", nvec);
+		return nvec;
+	}
+
+	xe->irq.msix.nvec = nvec;
+	xa_init_flags(&xe->irq.msix.indexes, XA_FLAGS_ALLOC);
+	return 0;
+}
+
+static irqreturn_t guc2host_irq_handler(int irq, void *arg)
+{
+	struct xe_device *xe = arg;
+	struct xe_tile *tile;
+	u8 id;
+
+	if (!atomic_read(&xe->irq.enabled))
+		return IRQ_NONE;
+
+	for_each_tile(tile, xe, id)
+		xe_guc_irq_handler(&tile->primary_gt->uc.guc,
+				   GUC_INTR_GUC2HOST);
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t xe_irq_msix_default_hwe_handler(int irq, void *arg)
+{
+	unsigned int tile_id, gt_id;
+	struct xe_device *xe = arg;
+	struct xe_memirq *memirq;
+	struct xe_hw_engine *hwe;
+	enum xe_hw_engine_id id;
+	struct xe_tile *tile;
+	struct xe_gt *gt;
+
+	if (!atomic_read(&xe->irq.enabled))
+		return IRQ_NONE;
+
+	for_each_tile(tile, xe, tile_id) {
+		memirq = &tile->memirq;
+		if (!memirq->bo)
+			continue;
+
+		for_each_gt(gt, xe, gt_id) {
+			if (gt->tile != tile)
+				continue;
+
+			for_each_hw_engine(hwe, gt, id)
+				xe_memirq_hwe_handler(memirq, hwe);
+		}
+	}
+
+	return IRQ_HANDLED;
+}
+
+static int xe_irq_msix_alloc_vector(struct xe_device *xe, void *irq_buf,
+				    bool dynamic_msix, u16 *msix)
+{
+	struct xa_limit limit;
+	int ret;
+	u32 id;
+
+	limit = (dynamic_msix) ? XA_LIMIT(NUM_OF_STATIC_MSIX, xe->irq.msix.nvec - 1) :
+				 XA_LIMIT(*msix, *msix);
+	ret = xa_alloc(&xe->irq.msix.indexes, &id, irq_buf, limit, GFP_KERNEL);
+	if (ret)
+		return ret;
+
+	if (dynamic_msix)
+		*msix = id;
+
+	return 0;
+}
+
+static void xe_irq_msix_release_vector(struct xe_device *xe, u16 msix)
+{
+	xa_erase(&xe->irq.msix.indexes, msix);
+}
+
+static int xe_irq_msix_request_irq_internal(struct xe_device *xe, irq_handler_t handler,
+					    void *irq_buf, const char *name, u16 msix)
+{
+	struct pci_dev *pdev = to_pci_dev(xe->drm.dev);
+	int ret, irq;
+
+	irq = pci_irq_vector(pdev, msix);
+	if (irq < 0)
+		return irq;
+
+	ret = request_irq(irq, handler, IRQF_SHARED, name, irq_buf);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
+
+int xe_irq_msix_request_irq(struct xe_device *xe, irq_handler_t handler, void *irq_buf,
+			    const char *name, bool dynamic_msix, u16 *msix)
+{
+	int ret;
+
+	ret = xe_irq_msix_alloc_vector(xe, irq_buf, dynamic_msix, msix);
+	if (ret)
+		return ret;
+
+	ret = xe_irq_msix_request_irq_internal(xe, handler, irq_buf, name, *msix);
+	if (ret) {
+		drm_err(&xe->drm, "Failed to request IRQ for MSI-X %u\n", *msix);
+		xe_irq_msix_release_vector(xe, *msix);
+		return ret;
+	}
+
+	return 0;
+}
+
+void xe_irq_msix_free_irq(struct xe_device *xe, u16 msix)
+{
+	struct pci_dev *pdev = to_pci_dev(xe->drm.dev);
+	int irq;
+	void *irq_buf;
+
+	irq_buf = xa_load(&xe->irq.msix.indexes, msix);
+	if (!irq_buf)
+		return;
+
+	irq = pci_irq_vector(pdev, msix);
+	if (irq < 0) {
+		drm_err(&xe->drm, "MSI-X %u can't be released, there is no matching IRQ\n", msix);
+		return;
+	}
+
+	free_irq(irq, irq_buf);
+	xe_irq_msix_release_vector(xe, msix);
+}
+
+int xe_irq_msix_request_irqs(struct xe_device *xe)
+{
+	int err;
+	u16 msix;
+
+	msix = GUC2HOST_MSIX;
+	err = xe_irq_msix_request_irq(xe, guc2host_irq_handler, xe,
+				      DRIVER_NAME "-guc2host", false, &msix);
+	if (err)
+		return err;
+
+	msix = DEFAULT_MSIX;
+	err = xe_irq_msix_request_irq(xe, xe_irq_msix_default_hwe_handler, xe,
+				      DRIVER_NAME "-default-msix", false, &msix);
+	if (err) {
+		xe_irq_msix_free_irq(xe, GUC2HOST_MSIX);
+		return err;
+	}
+
+	return 0;
+}
+
+void xe_irq_msix_free(struct xe_device *xe)
+{
+	unsigned long msix;
+	u32 *dummy;
+
+	xa_for_each(&xe->irq.msix.indexes, msix, dummy)
+		xe_irq_msix_free_irq(xe, msix);
+	xa_destroy(&xe->irq.msix.indexes);
+}
+
+void xe_irq_msix_synchronize_irq(struct xe_device *xe)
+{
+	struct pci_dev *pdev = to_pci_dev(xe->drm.dev);
+	unsigned long msix;
+	u32 *dummy;
+
+	xa_for_each(&xe->irq.msix.indexes, msix, dummy)
+		synchronize_irq(pci_irq_vector(pdev, msix));
+}
diff --git a/drivers/gpu/drm/xe/xe_irq.h b/drivers/gpu/drm/xe/xe_irq.h
index 067514e13675..a28bd577ba52 100644
--- a/drivers/gpu/drm/xe/xe_irq.h
+++ b/drivers/gpu/drm/xe/xe_irq.h
@@ -6,13 +6,21 @@
 #ifndef _XE_IRQ_H_
 #define _XE_IRQ_H_
 
+#include <linux/interrupt.h>
+
+#define XE_IRQ_DEFAULT_MSIX 1
+
 struct xe_device;
 struct xe_tile;
 struct xe_gt;
 
+int xe_irq_init(struct xe_device *xe);
 int xe_irq_install(struct xe_device *xe);
 void xe_irq_suspend(struct xe_device *xe);
 void xe_irq_resume(struct xe_device *xe);
 void xe_irq_enable_hwe(struct xe_gt *gt);
+int xe_irq_msix_request_irq(struct xe_device *xe, irq_handler_t handler, void *irq_buf,
+			    const char *name, bool dynamic_msix, u16 *msix);
+void xe_irq_msix_free_irq(struct xe_device *xe, u16 msix);
 
 #endif
diff --git a/drivers/gpu/drm/xe/xe_lrc.c b/drivers/gpu/drm/xe/xe_lrc.c
index 22e58c6e2a35..bbb9ffbf6367 100644
--- a/drivers/gpu/drm/xe/xe_lrc.c
+++ b/drivers/gpu/drm/xe/xe_lrc.c
@@ -584,6 +584,7 @@ static void set_memory_based_intr(u32 *regs, struct xe_hw_engine *hwe)
 {
 	struct xe_memirq *memirq = &gt_to_tile(hwe->gt)->memirq;
 	struct xe_device *xe = gt_to_xe(hwe->gt);
+	u8 num_regs;
 
 	if (!xe_device_uses_memirq(xe))
 		return;
@@ -593,12 +594,18 @@ static void set_memory_based_intr(u32 *regs, struct xe_hw_engine *hwe)
 	regs[CTX_INT_MASK_ENABLE_REG] = RING_IMR(0).addr;
 	regs[CTX_INT_MASK_ENABLE_PTR] = xe_memirq_enable_ptr(memirq);
 
-	regs[CTX_LRI_INT_REPORT_PTR] = MI_LOAD_REGISTER_IMM | MI_LRI_NUM_REGS(2) |
+	num_regs = xe_device_has_msix(xe) ? 3 : 2;
+	regs[CTX_LRI_INT_REPORT_PTR] = MI_LOAD_REGISTER_IMM | MI_LRI_NUM_REGS(num_regs) |
 				       MI_LRI_LRM_CS_MMIO | MI_LRI_FORCE_POSTED;
 	regs[CTX_INT_STATUS_REPORT_REG] = RING_INT_STATUS_RPT_PTR(0).addr;
 	regs[CTX_INT_STATUS_REPORT_PTR] = xe_memirq_status_ptr(memirq, hwe);
 	regs[CTX_INT_SRC_REPORT_REG] = RING_INT_SRC_RPT_PTR(0).addr;
 	regs[CTX_INT_SRC_REPORT_PTR] = xe_memirq_source_ptr(memirq, hwe);
+
+	if (xe_device_has_msix(xe)) {
+		regs[CTX_CS_INT_VEC_REG] = CS_INT_VEC(0).addr;
+		/* CTX_CS_INT_VEC_DATA will be set in xe_lrc_init */
+	}
 }
 
 static int lrc_ring_mi_mode(struct xe_hw_engine *hwe)
@@ -876,7 +883,7 @@ static void xe_lrc_finish(struct xe_lrc *lrc)
 #define PVC_CTX_ACC_CTR_THOLD	(0x2a + 1)
 
 static int xe_lrc_init(struct xe_lrc *lrc, struct xe_hw_engine *hwe,
-		       struct xe_vm *vm, u32 ring_size)
+		       struct xe_vm *vm, u32 ring_size, u16 msix_vec)
 {
 	struct xe_gt *gt = hwe->gt;
 	struct xe_tile *tile = gt_to_tile(gt);
@@ -945,6 +952,14 @@ static int xe_lrc_init(struct xe_lrc *lrc, struct xe_hw_engine *hwe,
 			xe_drm_client_add_bo(vm->xef->client, lrc->bo);
 	}
 
+	if (xe_device_has_msix(xe)) {
+		xe_lrc_write_ctx_reg(lrc, CTX_INT_STATUS_REPORT_PTR,
+				     xe_memirq_status_ptr(&tile->memirq, hwe));
+		xe_lrc_write_ctx_reg(lrc, CTX_INT_SRC_REPORT_PTR,
+				     xe_memirq_source_ptr(&tile->memirq, hwe));
+		xe_lrc_write_ctx_reg(lrc, CTX_CS_INT_VEC_DATA, msix_vec << 16 | msix_vec);
+	}
+
 	if (xe_gt_has_indirect_ring_state(gt)) {
 		xe_lrc_write_ctx_reg(lrc, CTX_INDIRECT_RING_STATE,
 				     __xe_lrc_indirect_ring_ggtt_addr(lrc));
@@ -1005,6 +1020,7 @@ err_lrc_finish:
  * @hwe: Hardware Engine
  * @vm: The VM (address space)
  * @ring_size: LRC ring size
+ * @msix_vec: MSI-X interrupt vector (for platforms that support it)
  *
  * Allocate and initialize the Logical Ring Context (LRC).
  *
@@ -1012,7 +1028,7 @@ err_lrc_finish:
  * upon failure.
  */
 struct xe_lrc *xe_lrc_create(struct xe_hw_engine *hwe, struct xe_vm *vm,
-			     u32 ring_size)
+			     u32 ring_size, u16 msix_vec)
 {
 	struct xe_lrc *lrc;
 	int err;
@@ -1021,7 +1037,7 @@ struct xe_lrc *xe_lrc_create(struct xe_hw_engine *hwe, struct xe_vm *vm,
 	if (!lrc)
 		return ERR_PTR(-ENOMEM);
 
-	err = xe_lrc_init(lrc, hwe, vm, ring_size);
+	err = xe_lrc_init(lrc, hwe, vm, ring_size, msix_vec);
 	if (err) {
 		kfree(lrc);
 		return ERR_PTR(err);
diff --git a/drivers/gpu/drm/xe/xe_lrc.h b/drivers/gpu/drm/xe/xe_lrc.h
index b459dcab8787..4206e6a8b50a 100644
--- a/drivers/gpu/drm/xe/xe_lrc.h
+++ b/drivers/gpu/drm/xe/xe_lrc.h
@@ -42,7 +42,7 @@ struct xe_lrc_snapshot {
 #define LRC_PPHWSP_SCRATCH_ADDR (0x34 * 4)
 
 struct xe_lrc *xe_lrc_create(struct xe_hw_engine *hwe, struct xe_vm *vm,
-			     u32 ring_size);
+			     u32 ring_size, u16 msix_vec);
 void xe_lrc_destroy(struct kref *ref);
 
 /**
diff --git a/drivers/gpu/drm/xe/xe_migrate.c b/drivers/gpu/drm/xe/xe_migrate.c
index 1b97d90aadda..8b32fad67878 100644
--- a/drivers/gpu/drm/xe/xe_migrate.c
+++ b/drivers/gpu/drm/xe/xe_migrate.c
@@ -581,7 +581,9 @@ static void emit_pte(struct xe_migrate *m,
 	while (ptes) {
 		u32 chunk = min(MAX_PTE_PER_SDI, ptes);
 
-		bb->cs[bb->len++] = MI_STORE_DATA_IMM | MI_SDI_NUM_QW(chunk);
+		bb->cs[bb->len++] = MI_STORE_DATA_IMM |
+				    MI_FORCE_WRITE_COMPLETION_CHECK |
+				    MI_SDI_NUM_QW(chunk);
 		bb->cs[bb->len++] = ofs;
 		bb->cs[bb->len++] = 0;
 
@@ -1223,7 +1225,9 @@ static void write_pgtable(struct xe_tile *tile, struct xe_bb *bb, u64 ppgtt_ofs,
 		if (!(bb->len & 1))
 			bb->cs[bb->len++] = MI_NOOP;
 
-		bb->cs[bb->len++] = MI_STORE_DATA_IMM | MI_SDI_NUM_QW(chunk);
+		bb->cs[bb->len++] = MI_STORE_DATA_IMM |
+				    MI_FORCE_WRITE_COMPLETION_CHECK |
+				    MI_SDI_NUM_QW(chunk);
 		bb->cs[bb->len++] = lower_32_bits(addr);
 		bb->cs[bb->len++] = upper_32_bits(addr);
 		if (pt_op->bind)
@@ -1388,7 +1392,8 @@ __xe_migrate_update_pgtables(struct xe_migrate *m,
 			u32 idx = 0;
 
 			bb->cs[bb->len++] = MI_STORE_DATA_IMM |
-				MI_SDI_NUM_QW(chunk);
+					    MI_FORCE_WRITE_COMPLETION_CHECK |
+					    MI_SDI_NUM_QW(chunk);
 			bb->cs[bb->len++] = ofs;
 			bb->cs[bb->len++] = 0; /* upper_32_bits */
 
diff --git a/drivers/gpu/drm/xe/xe_oa.c b/drivers/gpu/drm/xe/xe_oa.c
index ec88b18e9baa..4e00a77289c5 100644
--- a/drivers/gpu/drm/xe/xe_oa.c
+++ b/drivers/gpu/drm/xe/xe_oa.c
@@ -16,7 +16,6 @@
 #include "instructions/xe_mi_commands.h"
 #include "regs/xe_engine_regs.h"
 #include "regs/xe_gt_regs.h"
-#include "regs/xe_lrc_layout.h"
 #include "regs/xe_oa_regs.h"
 #include "xe_assert.h"
 #include "xe_bb.h"
@@ -28,7 +27,6 @@
 #include "xe_gt_mcr.h"
 #include "xe_gt_printk.h"
 #include "xe_guc_pc.h"
-#include "xe_lrc.h"
 #include "xe_macros.h"
 #include "xe_mmio.h"
 #include "xe_oa.h"
@@ -74,12 +72,6 @@ struct xe_oa_config {
 	struct rcu_head rcu;
 };
 
-struct flex {
-	struct xe_reg reg;
-	u32 offset;
-	u32 value;
-};
-
 struct xe_oa_open_param {
 	struct xe_file *xef;
 	u32 oa_unit_id;
@@ -97,6 +89,7 @@ struct xe_oa_open_param {
 	int num_syncs;
 	struct xe_sync_entry *syncs;
 	size_t oa_buffer_size;
+	int wait_num_reports;
 };
 
 struct xe_oa_config_bo {
@@ -241,11 +234,10 @@ static void oa_timestamp_clear(struct xe_oa_stream *stream, u32 *report)
 static bool xe_oa_buffer_check_unlocked(struct xe_oa_stream *stream)
 {
 	u32 gtt_offset = xe_bo_ggtt_addr(stream->oa_buffer.bo);
+	u32 tail, hw_tail, partial_report_size, available;
 	int report_size = stream->oa_buffer.format->size;
-	u32 tail, hw_tail;
 	unsigned long flags;
 	bool pollin;
-	u32 partial_report_size;
 
 	spin_lock_irqsave(&stream->oa_buffer.ptr_lock, flags);
 
@@ -289,8 +281,8 @@ static bool xe_oa_buffer_check_unlocked(struct xe_oa_stream *stream)
 
 	stream->oa_buffer.tail = tail;
 
-	pollin = xe_oa_circ_diff(stream, stream->oa_buffer.tail,
-				 stream->oa_buffer.head) >= report_size;
+	available = xe_oa_circ_diff(stream, stream->oa_buffer.tail, stream->oa_buffer.head);
+	pollin = available >= stream->wait_num_reports * report_size;
 
 	spin_unlock_irqrestore(&stream->oa_buffer.ptr_lock, flags);
 
@@ -605,19 +597,38 @@ static __poll_t xe_oa_poll(struct file *file, poll_table *wait)
 	return ret;
 }
 
+static void xe_oa_lock_vma(struct xe_exec_queue *q)
+{
+	if (q->vm) {
+		down_read(&q->vm->lock);
+		xe_vm_lock(q->vm, false);
+	}
+}
+
+static void xe_oa_unlock_vma(struct xe_exec_queue *q)
+{
+	if (q->vm) {
+		xe_vm_unlock(q->vm);
+		up_read(&q->vm->lock);
+	}
+}
+
 static struct dma_fence *xe_oa_submit_bb(struct xe_oa_stream *stream, enum xe_oa_submit_deps deps,
 					 struct xe_bb *bb)
 {
+	struct xe_exec_queue *q = stream->exec_q ?: stream->k_exec_q;
 	struct xe_sched_job *job;
 	struct dma_fence *fence;
 	int err = 0;
 
-	/* Kernel configuration is issued on stream->k_exec_q, not stream->exec_q */
-	job = xe_bb_create_job(stream->k_exec_q, bb);
+	xe_oa_lock_vma(q);
+
+	job = xe_bb_create_job(q, bb);
 	if (IS_ERR(job)) {
 		err = PTR_ERR(job);
 		goto exit;
 	}
+	job->ggtt = true;
 
 	if (deps == XE_OA_SUBMIT_ADD_DEPS) {
 		for (int i = 0; i < stream->num_syncs && !err; i++)
@@ -632,10 +643,13 @@ static struct dma_fence *xe_oa_submit_bb(struct xe_oa_stream *stream, enum xe_oa
 	fence = dma_fence_get(&job->drm.s_fence->finished);
 	xe_sched_job_push(job);
 
+	xe_oa_unlock_vma(q);
+
 	return fence;
 err_put_job:
 	xe_sched_job_put(job);
 exit:
+	xe_oa_unlock_vma(q);
 	return ERR_PTR(err);
 }
 
@@ -684,63 +698,19 @@ static void xe_oa_free_configs(struct xe_oa_stream *stream)
 	dma_fence_put(stream->last_fence);
 }
 
-static void xe_oa_store_flex(struct xe_oa_stream *stream, struct xe_lrc *lrc,
-			     struct xe_bb *bb, const struct flex *flex, u32 count)
-{
-	u32 offset = xe_bo_ggtt_addr(lrc->bo);
-
-	do {
-		bb->cs[bb->len++] = MI_STORE_DATA_IMM | MI_SDI_GGTT | MI_SDI_NUM_DW(1);
-		bb->cs[bb->len++] = offset + flex->offset * sizeof(u32);
-		bb->cs[bb->len++] = 0;
-		bb->cs[bb->len++] = flex->value;
-
-	} while (flex++, --count);
-}
-
-static int xe_oa_modify_ctx_image(struct xe_oa_stream *stream, struct xe_lrc *lrc,
-				  const struct flex *flex, u32 count)
+static int xe_oa_load_with_lri(struct xe_oa_stream *stream, struct xe_oa_reg *reg_lri, u32 count)
 {
 	struct dma_fence *fence;
 	struct xe_bb *bb;
 	int err;
 
-	bb = xe_bb_new(stream->gt, 4 * count, false);
+	bb = xe_bb_new(stream->gt, 2 * count + 1, false);
 	if (IS_ERR(bb)) {
 		err = PTR_ERR(bb);
 		goto exit;
 	}
 
-	xe_oa_store_flex(stream, lrc, bb, flex, count);
-
-	fence = xe_oa_submit_bb(stream, XE_OA_SUBMIT_NO_DEPS, bb);
-	if (IS_ERR(fence)) {
-		err = PTR_ERR(fence);
-		goto free_bb;
-	}
-	xe_bb_free(bb, fence);
-	dma_fence_put(fence);
-
-	return 0;
-free_bb:
-	xe_bb_free(bb, NULL);
-exit:
-	return err;
-}
-
-static int xe_oa_load_with_lri(struct xe_oa_stream *stream, struct xe_oa_reg *reg_lri)
-{
-	struct dma_fence *fence;
-	struct xe_bb *bb;
-	int err;
-
-	bb = xe_bb_new(stream->gt, 3, false);
-	if (IS_ERR(bb)) {
-		err = PTR_ERR(bb);
-		goto exit;
-	}
-
-	write_cs_mi_lri(bb, reg_lri, 1);
+	write_cs_mi_lri(bb, reg_lri, count);
 
 	fence = xe_oa_submit_bb(stream, XE_OA_SUBMIT_NO_DEPS, bb);
 	if (IS_ERR(fence)) {
@@ -760,71 +730,55 @@ exit:
 static int xe_oa_configure_oar_context(struct xe_oa_stream *stream, bool enable)
 {
 	const struct xe_oa_format *format = stream->oa_buffer.format;
-	struct xe_lrc *lrc = stream->exec_q->lrc[0];
-	u32 regs_offset = xe_lrc_regs_offset(lrc) / sizeof(u32);
 	u32 oacontrol = __format_to_oactrl(format, OAR_OACONTROL_COUNTER_SEL_MASK) |
 		(enable ? OAR_OACONTROL_COUNTER_ENABLE : 0);
 
-	struct flex regs_context[] = {
+	struct xe_oa_reg reg_lri[] = {
 		{
 			OACTXCONTROL(stream->hwe->mmio_base),
-			stream->oa->ctx_oactxctrl_offset[stream->hwe->class] + 1,
 			enable ? OA_COUNTER_RESUME : 0,
 		},
 		{
+			OAR_OACONTROL,
+			oacontrol,
+		},
+		{
 			RING_CONTEXT_CONTROL(stream->hwe->mmio_base),
-			regs_offset + CTX_CONTEXT_CONTROL,
-			_MASKED_BIT_ENABLE(CTX_CTRL_OAC_CONTEXT_ENABLE),
+			_MASKED_FIELD(CTX_CTRL_OAC_CONTEXT_ENABLE,
+				      enable ? CTX_CTRL_OAC_CONTEXT_ENABLE : 0)
 		},
 	};
-	struct xe_oa_reg reg_lri = { OAR_OACONTROL, oacontrol };
-	int err;
-
-	/* Modify stream hwe context image with regs_context */
-	err = xe_oa_modify_ctx_image(stream, stream->exec_q->lrc[0],
-				     regs_context, ARRAY_SIZE(regs_context));
-	if (err)
-		return err;
 
-	/* Apply reg_lri using LRI */
-	return xe_oa_load_with_lri(stream, &reg_lri);
+	return xe_oa_load_with_lri(stream, reg_lri, ARRAY_SIZE(reg_lri));
 }
 
 static int xe_oa_configure_oac_context(struct xe_oa_stream *stream, bool enable)
 {
 	const struct xe_oa_format *format = stream->oa_buffer.format;
-	struct xe_lrc *lrc = stream->exec_q->lrc[0];
-	u32 regs_offset = xe_lrc_regs_offset(lrc) / sizeof(u32);
 	u32 oacontrol = __format_to_oactrl(format, OAR_OACONTROL_COUNTER_SEL_MASK) |
 		(enable ? OAR_OACONTROL_COUNTER_ENABLE : 0);
-	struct flex regs_context[] = {
+	struct xe_oa_reg reg_lri[] = {
 		{
 			OACTXCONTROL(stream->hwe->mmio_base),
-			stream->oa->ctx_oactxctrl_offset[stream->hwe->class] + 1,
 			enable ? OA_COUNTER_RESUME : 0,
 		},
 		{
+			OAC_OACONTROL,
+			oacontrol
+		},
+		{
 			RING_CONTEXT_CONTROL(stream->hwe->mmio_base),
-			regs_offset + CTX_CONTEXT_CONTROL,
-			_MASKED_BIT_ENABLE(CTX_CTRL_OAC_CONTEXT_ENABLE) |
+			_MASKED_FIELD(CTX_CTRL_OAC_CONTEXT_ENABLE,
+				      enable ? CTX_CTRL_OAC_CONTEXT_ENABLE : 0) |
 			_MASKED_FIELD(CTX_CTRL_RUN_ALONE, enable ? CTX_CTRL_RUN_ALONE : 0),
 		},
 	};
-	struct xe_oa_reg reg_lri = { OAC_OACONTROL, oacontrol };
-	int err;
 
 	/* Set ccs select to enable programming of OAC_OACONTROL */
 	xe_mmio_write32(&stream->gt->mmio, __oa_regs(stream)->oa_ctrl,
 			__oa_ccs_select(stream));
 
-	/* Modify stream hwe context image with regs_context */
-	err = xe_oa_modify_ctx_image(stream, stream->exec_q->lrc[0],
-				     regs_context, ARRAY_SIZE(regs_context));
-	if (err)
-		return err;
-
-	/* Apply reg_lri using LRI */
-	return xe_oa_load_with_lri(stream, &reg_lri);
+	return xe_oa_load_with_lri(stream, reg_lri, ARRAY_SIZE(reg_lri));
 }
 
 static int xe_oa_configure_oa_context(struct xe_oa_stream *stream, bool enable)
@@ -1285,6 +1239,17 @@ static int xe_oa_set_prop_oa_buffer_size(struct xe_oa *oa, u64 value,
 	return 0;
 }
 
+static int xe_oa_set_prop_wait_num_reports(struct xe_oa *oa, u64 value,
+					   struct xe_oa_open_param *param)
+{
+	if (!value) {
+		drm_dbg(&oa->xe->drm, "wait_num_reports %llu\n", value);
+		return -EINVAL;
+	}
+	param->wait_num_reports = value;
+	return 0;
+}
+
 static int xe_oa_set_prop_ret_inval(struct xe_oa *oa, u64 value,
 				    struct xe_oa_open_param *param)
 {
@@ -1306,6 +1271,7 @@ static const xe_oa_set_property_fn xe_oa_set_property_funcs_open[] = {
 	[DRM_XE_OA_PROPERTY_NUM_SYNCS] = xe_oa_set_prop_num_syncs,
 	[DRM_XE_OA_PROPERTY_SYNCS] = xe_oa_set_prop_syncs_user,
 	[DRM_XE_OA_PROPERTY_OA_BUFFER_SIZE] = xe_oa_set_prop_oa_buffer_size,
+	[DRM_XE_OA_PROPERTY_WAIT_NUM_REPORTS] = xe_oa_set_prop_wait_num_reports,
 };
 
 static const xe_oa_set_property_fn xe_oa_set_property_funcs_config[] = {
@@ -1321,6 +1287,7 @@ static const xe_oa_set_property_fn xe_oa_set_property_funcs_config[] = {
 	[DRM_XE_OA_PROPERTY_NUM_SYNCS] = xe_oa_set_prop_num_syncs,
 	[DRM_XE_OA_PROPERTY_SYNCS] = xe_oa_set_prop_syncs_user,
 	[DRM_XE_OA_PROPERTY_OA_BUFFER_SIZE] = xe_oa_set_prop_ret_inval,
+	[DRM_XE_OA_PROPERTY_WAIT_NUM_REPORTS] = xe_oa_set_prop_ret_inval,
 };
 
 static int xe_oa_user_ext_set_property(struct xe_oa *oa, enum xe_oa_user_extn_from from,
@@ -1704,81 +1671,6 @@ static const struct file_operations xe_oa_fops = {
 	.mmap		= xe_oa_mmap,
 };
 
-static bool engine_supports_mi_query(struct xe_hw_engine *hwe)
-{
-	return hwe->class == XE_ENGINE_CLASS_RENDER ||
-		hwe->class == XE_ENGINE_CLASS_COMPUTE;
-}
-
-static bool xe_oa_find_reg_in_lri(u32 *state, u32 reg, u32 *offset, u32 end)
-{
-	u32 idx = *offset;
-	u32 len = min(MI_LRI_LEN(state[idx]) + idx, end);
-	bool found = false;
-
-	idx++;
-	for (; idx < len; idx += 2) {
-		if (state[idx] == reg) {
-			found = true;
-			break;
-		}
-	}
-
-	*offset = idx;
-	return found;
-}
-
-#define IS_MI_LRI_CMD(x) (REG_FIELD_GET(MI_OPCODE, (x)) == \
-			  REG_FIELD_GET(MI_OPCODE, MI_LOAD_REGISTER_IMM))
-
-static u32 xe_oa_context_image_offset(struct xe_oa_stream *stream, u32 reg)
-{
-	struct xe_lrc *lrc = stream->exec_q->lrc[0];
-	u32 len = (xe_gt_lrc_size(stream->gt, stream->hwe->class) +
-		   lrc->ring.size) / sizeof(u32);
-	u32 offset = xe_lrc_regs_offset(lrc) / sizeof(u32);
-	u32 *state = (u32 *)lrc->bo->vmap.vaddr;
-
-	if (drm_WARN_ON(&stream->oa->xe->drm, !state))
-		return U32_MAX;
-
-	for (; offset < len; ) {
-		if (IS_MI_LRI_CMD(state[offset])) {
-			/*
-			 * We expect reg-value pairs in MI_LRI command, so
-			 * MI_LRI_LEN() should be even
-			 */
-			drm_WARN_ON(&stream->oa->xe->drm,
-				    MI_LRI_LEN(state[offset]) & 0x1);
-
-			if (xe_oa_find_reg_in_lri(state, reg, &offset, len))
-				break;
-		} else {
-			offset++;
-		}
-	}
-
-	return offset < len ? offset : U32_MAX;
-}
-
-static int xe_oa_set_ctx_ctrl_offset(struct xe_oa_stream *stream)
-{
-	struct xe_reg reg = OACTXCONTROL(stream->hwe->mmio_base);
-	u32 offset = stream->oa->ctx_oactxctrl_offset[stream->hwe->class];
-
-	/* Do this only once. Failure is stored as offset of U32_MAX */
-	if (offset)
-		goto exit;
-
-	offset = xe_oa_context_image_offset(stream, reg.addr);
-	stream->oa->ctx_oactxctrl_offset[stream->hwe->class] = offset;
-
-	drm_dbg(&stream->oa->xe->drm, "%s oa ctx control at 0x%08x dword offset\n",
-		stream->hwe->name, offset);
-exit:
-	return offset && offset != U32_MAX ? 0 : -ENODEV;
-}
-
 static int xe_oa_stream_init(struct xe_oa_stream *stream,
 			     struct xe_oa_open_param *param)
 {
@@ -1797,6 +1689,7 @@ static int xe_oa_stream_init(struct xe_oa_stream *stream,
 	stream->periodic = param->period_exponent > 0;
 	stream->period_exponent = param->period_exponent;
 	stream->no_preempt = param->no_preempt;
+	stream->wait_num_reports = param->wait_num_reports;
 
 	stream->xef = xe_file_get(param->xef);
 	stream->num_syncs = param->num_syncs;
@@ -1815,17 +1708,6 @@ static int xe_oa_stream_init(struct xe_oa_stream *stream,
 	else
 		stream->oa_buffer.circ_size = param->oa_buffer_size;
 
-	if (stream->exec_q && engine_supports_mi_query(stream->hwe)) {
-		/* If we don't find the context offset, just return error */
-		ret = xe_oa_set_ctx_ctrl_offset(stream);
-		if (ret) {
-			drm_err(&stream->oa->xe->drm,
-				"xe_oa_set_ctx_ctrl_offset failed for %s\n",
-				stream->hwe->name);
-			goto exit;
-		}
-	}
-
 	stream->oa_config = xe_oa_get_oa_config(stream->oa, param->metric_set);
 	if (!stream->oa_config) {
 		drm_dbg(&stream->oa->xe->drm, "Invalid OA config id=%i\n", param->metric_set);
@@ -2094,8 +1976,8 @@ int xe_oa_stream_open_ioctl(struct drm_device *dev, u64 data, struct drm_file *f
 		if (XE_IOCTL_DBG(oa->xe, !param.exec_q))
 			return -ENOENT;
 
-		if (param.exec_q->width > 1)
-			drm_dbg(&oa->xe->drm, "exec_q->width > 1, programming only exec_q->lrc[0]\n");
+		if (XE_IOCTL_DBG(oa->xe, param.exec_q->width > 1))
+			return -EOPNOTSUPP;
 	}
 
 	/*
@@ -2156,6 +2038,14 @@ int xe_oa_stream_open_ioctl(struct drm_device *dev, u64 data, struct drm_file *f
 	if (!param.oa_buffer_size)
 		param.oa_buffer_size = DEFAULT_XE_OA_BUFFER_SIZE;
 
+	if (!param.wait_num_reports)
+		param.wait_num_reports = 1;
+	if (param.wait_num_reports > param.oa_buffer_size / f->size) {
+		drm_dbg(&oa->xe->drm, "wait_num_reports %d\n", param.wait_num_reports);
+		ret = -EINVAL;
+		goto err_exec_q;
+	}
+
 	ret = xe_oa_parse_syncs(oa, &param);
 	if (ret)
 		goto err_exec_q;
diff --git a/drivers/gpu/drm/xe/xe_oa_types.h b/drivers/gpu/drm/xe/xe_oa_types.h
index df7793915628..52e33c37d5ee 100644
--- a/drivers/gpu/drm/xe/xe_oa_types.h
+++ b/drivers/gpu/drm/xe/xe_oa_types.h
@@ -138,9 +138,6 @@ struct xe_oa {
 	/** @metrics_idr: List of dynamic configurations (struct xe_oa_config) */
 	struct idr metrics_idr;
 
-	/** @ctx_oactxctrl_offset: offset of OACTXCONTROL register in context image */
-	u32 ctx_oactxctrl_offset[XE_ENGINE_CLASS_MAX];
-
 	/** @oa_formats: tracks all OA formats across platforms */
 	const struct xe_oa_format *oa_formats;
 
@@ -218,6 +215,9 @@ struct xe_oa_stream {
 	/** @pollin: Whether there is data available to read */
 	bool pollin;
 
+	/** @wait_num_reports: Number of reports to wait for before signalling pollin */
+	int wait_num_reports;
+
 	/** @periodic: Whether periodic sampling is currently enabled */
 	bool periodic;
 
diff --git a/drivers/gpu/drm/xe/xe_pm.c b/drivers/gpu/drm/xe/xe_pm.c
index a6761cb769b2..c6e57af0144c 100644
--- a/drivers/gpu/drm/xe/xe_pm.c
+++ b/drivers/gpu/drm/xe/xe_pm.c
@@ -7,6 +7,7 @@
 
 #include <linux/fault-inject.h>
 #include <linux/pm_runtime.h>
+#include <linux/suspend.h>
 
 #include <drm/drm_managed.h>
 #include <drm/ttm/ttm_placement.h>
@@ -607,7 +608,8 @@ static bool xe_pm_suspending_or_resuming(struct xe_device *xe)
 	struct device *dev = xe->drm.dev;
 
 	return dev->power.runtime_status == RPM_SUSPENDING ||
-		dev->power.runtime_status == RPM_RESUMING;
+		dev->power.runtime_status == RPM_RESUMING ||
+		pm_suspend_target_state != PM_SUSPEND_ON;
 #else
 	return false;
 #endif
diff --git a/drivers/gpu/drm/xe/xe_query.c b/drivers/gpu/drm/xe/xe_query.c
index d2a816f71bf2..c059639613f7 100644
--- a/drivers/gpu/drm/xe/xe_query.c
+++ b/drivers/gpu/drm/xe/xe_query.c
@@ -672,7 +672,8 @@ static int query_oa_units(struct xe_device *xe,
 			du->oa_unit_type = u->type;
 			du->oa_timestamp_freq = xe_oa_timestamp_frequency(gt);
 			du->capabilities = DRM_XE_OA_CAPS_BASE | DRM_XE_OA_CAPS_SYNCS |
-					   DRM_XE_OA_CAPS_OA_BUFFER_SIZE;
+					   DRM_XE_OA_CAPS_OA_BUFFER_SIZE |
+					   DRM_XE_OA_CAPS_WAIT_NUM_REPORTS;
 
 			j = 0;
 			for_each_hw_engine(hwe, gt, hwe_id) {
diff --git a/drivers/gpu/drm/xe/xe_ring_ops.c b/drivers/gpu/drm/xe/xe_ring_ops.c
index 0be4f489d3e1..c8ab37fa0d19 100644
--- a/drivers/gpu/drm/xe/xe_ring_ops.c
+++ b/drivers/gpu/drm/xe/xe_ring_ops.c
@@ -72,7 +72,8 @@ static int emit_user_interrupt(u32 *dw, int i)
 
 static int emit_store_imm_ggtt(u32 addr, u32 value, u32 *dw, int i)
 {
-	dw[i++] = MI_STORE_DATA_IMM | MI_SDI_GGTT | MI_SDI_NUM_DW(1);
+	dw[i++] = MI_STORE_DATA_IMM | MI_SDI_GGTT |
+		  MI_FORCE_WRITE_COMPLETION_CHECK | MI_SDI_NUM_DW(1);
 	dw[i++] = addr;
 	dw[i++] = 0;
 	dw[i++] = value;
@@ -162,7 +163,8 @@ static int emit_pipe_invalidate(u32 mask_flags, bool invalidate_tlb, u32 *dw,
 static int emit_store_imm_ppgtt_posted(u64 addr, u64 value,
 				       u32 *dw, int i)
 {
-	dw[i++] = MI_STORE_DATA_IMM | MI_SDI_NUM_QW(1);
+	dw[i++] = MI_STORE_DATA_IMM | MI_FORCE_WRITE_COMPLETION_CHECK |
+		  MI_SDI_NUM_QW(1);
 	dw[i++] = lower_32_bits(addr);
 	dw[i++] = upper_32_bits(addr);
 	dw[i++] = lower_32_bits(value);
@@ -221,7 +223,10 @@ static int emit_pipe_imm_ggtt(u32 addr, u32 value, bool stall_only, u32 *dw,
 
 static u32 get_ppgtt_flag(struct xe_sched_job *job)
 {
-	return job->q->vm ? BIT(8) : 0;
+	if (job->q->vm && !job->ggtt)
+		return BIT(8);
+
+	return 0;
 }
 
 static int emit_copy_timestamp(struct xe_lrc *lrc, u32 *dw, int i)
diff --git a/drivers/gpu/drm/xe/xe_sched_job_types.h b/drivers/gpu/drm/xe/xe_sched_job_types.h
index f13f333f00be..d942b20a9f29 100644
--- a/drivers/gpu/drm/xe/xe_sched_job_types.h
+++ b/drivers/gpu/drm/xe/xe_sched_job_types.h
@@ -56,6 +56,8 @@ struct xe_sched_job {
 	u32 migrate_flush_flags;
 	/** @ring_ops_flush_tlb: The ring ops need to flush TLB before payload. */
 	bool ring_ops_flush_tlb;
+	/** @ggtt: mapped in ggtt. */
+	bool ggtt;
 	/** @ptrs: per instance pointers. */
 	struct xe_job_ptrs ptrs[];
 };
diff --git a/drivers/gpu/drm/xe/xe_trace_bo.h b/drivers/gpu/drm/xe/xe_trace_bo.h
index 1762dd30ba6d..ea50fee50c7d 100644
--- a/drivers/gpu/drm/xe/xe_trace_bo.h
+++ b/drivers/gpu/drm/xe/xe_trace_bo.h
@@ -60,8 +60,8 @@ TRACE_EVENT(xe_bo_move,
 	    TP_STRUCT__entry(
 		     __field(struct xe_bo *, bo)
 		     __field(size_t, size)
-		     __field(u32, new_placement)
-		     __field(u32, old_placement)
+		     __string(new_placement_name, xe_mem_type_to_name[new_placement])
+		     __string(old_placement_name, xe_mem_type_to_name[old_placement])
 		     __string(device_id, __dev_name_bo(bo))
 		     __field(bool, move_lacks_source)
 			),
@@ -69,15 +69,15 @@ TRACE_EVENT(xe_bo_move,
 	    TP_fast_assign(
 		   __entry->bo      = bo;
 		   __entry->size = bo->size;
-		   __entry->new_placement = new_placement;
-		   __entry->old_placement = old_placement;
+		   __assign_str(new_placement_name);
+		   __assign_str(old_placement_name);
 		   __assign_str(device_id);
 		   __entry->move_lacks_source = move_lacks_source;
 		   ),
 	    TP_printk("move_lacks_source:%s, migrate object %p [size %zu] from %s to %s device_id:%s",
 		      __entry->move_lacks_source ? "yes" : "no", __entry->bo, __entry->size,
-		      xe_mem_type_to_name[__entry->old_placement],
-		      xe_mem_type_to_name[__entry->new_placement], __get_str(device_id))
+		      __get_str(old_placement_name),
+		      __get_str(new_placement_name), __get_str(device_id))
 );
 
 DECLARE_EVENT_CLASS(xe_vma,
diff --git a/include/drm/intel/pciids.h b/include/drm/intel/pciids.h
index c6518b0992cf..77c826589ec1 100644
--- a/include/drm/intel/pciids.h
+++ b/include/drm/intel/pciids.h
@@ -858,6 +858,7 @@
 	MACRO__(0xB092, ## __VA_ARGS__), \
 	MACRO__(0xB0A0, ## __VA_ARGS__), \
 	MACRO__(0xB0A1, ## __VA_ARGS__), \
-	MACRO__(0xB0A2, ## __VA_ARGS__)
+	MACRO__(0xB0A2, ## __VA_ARGS__), \
+	MACRO__(0xB0B0, ## __VA_ARGS__)
 
 #endif /* __PCIIDS_H__ */
diff --git a/include/uapi/drm/xe_drm.h b/include/uapi/drm/xe_drm.h
index 0383b52cbd86..f62689ca861a 100644
--- a/include/uapi/drm/xe_drm.h
+++ b/include/uapi/drm/xe_drm.h
@@ -1487,6 +1487,7 @@ struct drm_xe_oa_unit {
 #define DRM_XE_OA_CAPS_BASE		(1 << 0)
 #define DRM_XE_OA_CAPS_SYNCS		(1 << 1)
 #define DRM_XE_OA_CAPS_OA_BUFFER_SIZE	(1 << 2)
+#define DRM_XE_OA_CAPS_WAIT_NUM_REPORTS	(1 << 3)
 
 	/** @oa_timestamp_freq: OA timestamp freq */
 	__u64 oa_timestamp_freq;
@@ -1660,6 +1661,12 @@ enum drm_xe_oa_property_id {
 	 * buffer is allocated by default.
 	 */
 	DRM_XE_OA_PROPERTY_OA_BUFFER_SIZE,
+
+	/**
+	 * @DRM_XE_OA_PROPERTY_WAIT_NUM_REPORTS: Number of reports to wait
+	 * for before unblocking poll or read
+	 */
+	DRM_XE_OA_PROPERTY_WAIT_NUM_REPORTS,
 };
 
 /**