Merge tag 'drm-intel-gt-next-2023-10-19' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

Driver Changes:

Fixes/improvements/new stuff:

- Retry gtt fault when out of fence registers (Ville Syrjälä)
- Determine context valid in OA reports [perf] (Umesh Nerlige Ramappa)

Future platform enablement:

- GuC based TLB invalidation for Meteorlake (Jonathan Cavitt, Prathap Kumar Valsan)
- Don't set PIPE_CONTROL_FLUSH_L3 [mtl] (Vinay Belgaumkar)

Miscellaneous:

- Clean up zero initializers [guc,pxp] (Ville Syrjälä)
- Prevent potential null-ptr-deref in engine_init_common (Nirmoy Das)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/ZTFDFSbd/U7YP+hI@tursulin-desk

21 files changed, 407 insertions, 30 deletions

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_mman.c b/drivers/gpu/drm/i915/gem/i915_gem_mman.c
index aa4d842d4c5a..310654542b42 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_mman.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_mman.c
@@ -235,6 +235,7 @@ static vm_fault_t i915_error_to_vmf_fault(int err)
 	case 0:
 	case -EAGAIN:
 	case -ENOSPC: /* transient failure to evict? */
+	case -ENOBUFS: /* temporarily out of fences? */
 	case -ERESTARTSYS:
 	case -EINTR:
 	case -EBUSY:
diff --git a/drivers/gpu/drm/i915/gt/gen8_engine_cs.c b/drivers/gpu/drm/i915/gt/gen8_engine_cs.c
index ba4c2422b340..86a04afff64b 100644
--- a/drivers/gpu/drm/i915/gt/gen8_engine_cs.c
+++ b/drivers/gpu/drm/i915/gt/gen8_engine_cs.c
@@ -278,7 +278,8 @@ int gen12_emit_flush_rcs(struct i915_request *rq, u32 mode)
 		 * deals with Protected Memory which is not needed for
 		 * AUX CCS invalidation and lead to unwanted side effects.
 		 */
-		if (mode & EMIT_FLUSH)
+		if ((mode & EMIT_FLUSH) &&
+		    GRAPHICS_VER_FULL(rq->i915) < IP_VER(12, 70))
 			bit_group_1 |= PIPE_CONTROL_FLUSH_L3;
 
 		bit_group_1 |= PIPE_CONTROL_TILE_CACHE_FLUSH;
@@ -812,12 +813,14 @@ u32 *gen12_emit_fini_breadcrumb_rcs(struct i915_request *rq, u32 *cs)
 	u32 flags = (PIPE_CONTROL_CS_STALL |
 		     PIPE_CONTROL_TLB_INVALIDATE |
 		     PIPE_CONTROL_TILE_CACHE_FLUSH |
-		     PIPE_CONTROL_FLUSH_L3 |
 		     PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH |
 		     PIPE_CONTROL_DEPTH_CACHE_FLUSH |
 		     PIPE_CONTROL_DC_FLUSH_ENABLE |
 		     PIPE_CONTROL_FLUSH_ENABLE);
 
+	if (GRAPHICS_VER_FULL(rq->i915) < IP_VER(12, 70))
+		flags |= PIPE_CONTROL_FLUSH_L3;
+
 	/* Wa_14016712196 */
 	if (IS_GFX_GT_IP_RANGE(gt, IP_VER(12, 70), IP_VER(12, 71)) || IS_DG2(i915))
 		/* dummy PIPE_CONTROL + depth flush */
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_cs.c b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
index 179d9546865b..4a11219e560e 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine_cs.c
+++ b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
@@ -1491,7 +1491,8 @@ static int engine_init_common(struct intel_engine_cs *engine)
 	return 0;
 
 err_bce_context:
-	intel_engine_destroy_pinned_context(bce);
+	if (bce)
+		intel_engine_destroy_pinned_context(bce);
 err_ce_context:
 	intel_engine_destroy_pinned_context(ce);
 	return ret;
diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt.c b/drivers/gpu/drm/i915/gt/intel_ggtt.c
index 4d7d88b92632..1c93e84278a0 100644
--- a/drivers/gpu/drm/i915/gt/intel_ggtt.c
+++ b/drivers/gpu/drm/i915/gt/intel_ggtt.c
@@ -206,22 +206,36 @@ static void gen8_ggtt_invalidate(struct i915_ggtt *ggtt)
 	intel_uncore_write_fw(uncore, GFX_FLSH_CNTL_GEN6, GFX_FLSH_CNTL_EN);
 }
 
+static void guc_ggtt_ct_invalidate(struct intel_gt *gt)
+{
+	struct intel_uncore *uncore = gt->uncore;
+	intel_wakeref_t wakeref;
+
+	with_intel_runtime_pm_if_active(uncore->rpm, wakeref) {
+		struct intel_guc *guc = &gt->uc.guc;
+
+		intel_guc_invalidate_tlb_guc(guc);
+	}
+}
+
 static void guc_ggtt_invalidate(struct i915_ggtt *ggtt)
 {
 	struct drm_i915_private *i915 = ggtt->vm.i915;
+	struct intel_gt *gt;
 
 	gen8_ggtt_invalidate(ggtt);
 
-	if (GRAPHICS_VER(i915) >= 12) {
-		struct intel_gt *gt;
-
-		list_for_each_entry(gt, &ggtt->gt_list, ggtt_link)
+	list_for_each_entry(gt, &ggtt->gt_list, ggtt_link) {
+		if (intel_guc_tlb_invalidation_is_available(&gt->uc.guc)) {
+			guc_ggtt_ct_invalidate(gt);
+		} else if (GRAPHICS_VER(i915) >= 12) {
 			intel_uncore_write_fw(gt->uncore,
 					      GEN12_GUC_TLB_INV_CR,
 					      GEN12_GUC_TLB_INV_CR_INVALIDATE);
-	} else {
-		intel_uncore_write_fw(ggtt->vm.gt->uncore,
-				      GEN8_GTCR, GEN8_GTCR_INVALIDATE);
+		} else {
+			intel_uncore_write_fw(gt->uncore,
+					      GEN8_GTCR, GEN8_GTCR_INVALIDATE);
+		}
 	}
 }
 
@@ -1243,7 +1257,7 @@ static int gen8_gmch_probe(struct i915_ggtt *ggtt)
 		ggtt->vm.raw_insert_page = gen8_ggtt_insert_page;
 	}
 
-	if (intel_uc_wants_guc(&ggtt->vm.gt->uc))
+	if (intel_uc_wants_guc_submission(&ggtt->vm.gt->uc))
 		ggtt->invalidate = guc_ggtt_invalidate;
 	else
 		ggtt->invalidate = gen8_ggtt_invalidate;
diff --git a/drivers/gpu/drm/i915/gt/intel_tlb.c b/drivers/gpu/drm/i915/gt/intel_tlb.c
index 139608c30d97..4bb13d1890e3 100644
--- a/drivers/gpu/drm/i915/gt/intel_tlb.c
+++ b/drivers/gpu/drm/i915/gt/intel_tlb.c
@@ -12,6 +12,7 @@
 #include "intel_gt_print.h"
 #include "intel_gt_regs.h"
 #include "intel_tlb.h"
+#include "uc/intel_guc.h"
 
 /*
  * HW architecture suggest typical invalidation time at 40us,
@@ -131,11 +132,24 @@ void intel_gt_invalidate_tlb_full(struct intel_gt *gt, u32 seqno)
 		return;
 
 	with_intel_gt_pm_if_awake(gt, wakeref) {
+		struct intel_guc *guc = &gt->uc.guc;
+
 		mutex_lock(&gt->tlb.invalidate_lock);
 		if (tlb_seqno_passed(gt, seqno))
 			goto unlock;
 
-		mmio_invalidate_full(gt);
+		if (HAS_GUC_TLB_INVALIDATION(gt->i915)) {
+			/*
+			 * Only perform GuC TLB invalidation if GuC is ready.
+			 * The only time GuC could not be ready is on GT reset,
+			 * which would clobber all the TLBs anyways, making
+			 * any TLB invalidation path here unnecessary.
+			 */
+			if (intel_guc_is_ready(guc))
+				intel_guc_invalidate_tlb_engines(guc);
+		} else {
+			mmio_invalidate_full(gt);
+		}
 
 		write_seqcount_invalidate(&gt->tlb.seqno);
 unlock:
diff --git a/drivers/gpu/drm/i915/gt/selftest_tlb.c b/drivers/gpu/drm/i915/gt/selftest_tlb.c
index 7e41f69fc818..00b872b6380b 100644
--- a/drivers/gpu/drm/i915/gt/selftest_tlb.c
+++ b/drivers/gpu/drm/i915/gt/selftest_tlb.c
@@ -136,8 +136,15 @@ pte_tlbinv(struct intel_context *ce,
 	i915_request_get(rq);
 	i915_request_add(rq);
 
-	/* Short sleep to sanitycheck the batch is spinning before we begin */
-	msleep(10);
+	/*
+	 * Short sleep to sanitycheck the batch is spinning before we begin.
+	 * FIXME: Why is GSC so slow?
+	 */
+	if (ce->engine->class == OTHER_CLASS)
+		msleep(200);
+	else
+		msleep(10);
+
 	if (va == vb) {
 		if (!i915_request_completed(rq)) {
 			pr_err("%s(%s): Semaphore sanitycheck failed %llx, with alignment %llx, using PTE size %x (phys %x, sg %x)\n",
diff --git a/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_abi.h b/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_abi.h
index f359bef046e0..33f253410d0c 100644
--- a/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_abi.h
+++ b/drivers/gpu/drm/i915/gt/uc/abi/guc_actions_abi.h
@@ -138,6 +138,8 @@ enum intel_guc_action {
 	INTEL_GUC_ACTION_REGISTER_CONTEXT_MULTI_LRC = 0x4601,
 	INTEL_GUC_ACTION_CLIENT_SOFT_RESET = 0x5507,
 	INTEL_GUC_ACTION_SET_ENG_UTIL_BUFF = 0x550A,
+	INTEL_GUC_ACTION_TLB_INVALIDATION = 0x7000,
+	INTEL_GUC_ACTION_TLB_INVALIDATION_DONE = 0x7001,
 	INTEL_GUC_ACTION_STATE_CAPTURE_NOTIFICATION = 0x8002,
 	INTEL_GUC_ACTION_NOTIFY_FLUSH_LOG_BUFFER_TO_FILE = 0x8003,
 	INTEL_GUC_ACTION_NOTIFY_CRASH_DUMP_POSTED = 0x8004,
@@ -181,4 +183,35 @@ enum intel_guc_state_capture_event_status {
 
 #define INTEL_GUC_STATE_CAPTURE_EVENT_STATUS_MASK      0x000000FF
 
+#define INTEL_GUC_TLB_INVAL_TYPE_MASK	REG_GENMASK(7, 0)
+#define INTEL_GUC_TLB_INVAL_MODE_MASK	REG_GENMASK(11, 8)
+#define INTEL_GUC_TLB_INVAL_FLUSH_CACHE REG_BIT(31)
+
+enum intel_guc_tlb_invalidation_type {
+	INTEL_GUC_TLB_INVAL_ENGINES = 0x0,
+	INTEL_GUC_TLB_INVAL_GUC = 0x3,
+};
+
+/*
+ * 0: Heavy mode of Invalidation:
+ * The pipeline of the engine(s) for which the invalidation is targeted to is
+ * blocked, and all the in-flight transactions are guaranteed to be Globally
+ * Observed before completing the TLB invalidation
+ * 1: Lite mode of Invalidation:
+ * TLBs of the targeted engine(s) are immediately invalidated.
+ * In-flight transactions are NOT guaranteed to be Globally Observed before
+ * completing TLB invalidation.
+ * Light Invalidation Mode is to be used only when
+ * it can be guaranteed (by SW) that the address translations remain invariant
+ * for the in-flight transactions across the TLB invalidation. In other words,
+ * this mode can be used when the TLB invalidation is intended to clear out the
+ * stale cached translations that are no longer in use. Light Invalidation Mode
+ * is much faster than the Heavy Invalidation Mode, as it does not wait for the
+ * in-flight transactions to be GOd.
+ */
+enum intel_guc_tlb_inval_mode {
+	INTEL_GUC_TLB_INVAL_MODE_HEAVY = 0x0,
+	INTEL_GUC_TLB_INVAL_MODE_LITE = 0x1,
+};
+
 #endif /* _ABI_GUC_ACTIONS_ABI_H */
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc.h b/drivers/gpu/drm/i915/gt/uc/intel_guc.h
index 818c8c146fd4..2b6dfe62c8f2 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc.h
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc.h
@@ -79,6 +79,18 @@ struct intel_guc {
 	 */
 	atomic_t outstanding_submission_g2h;
 
+	/** @tlb_lookup: xarray to store all pending TLB invalidation requests */
+	struct xarray tlb_lookup;
+
+	/**
+	 * @serial_slot: id to the initial waiter created in tlb_lookup,
+	 * which is used only when failed to allocate new waiter.
+	 */
+	u32 serial_slot;
+
+	/** @next_seqno: the next id (sequence number) to allocate. */
+	u32 next_seqno;
+
 	/** @interrupts: pointers to GuC interrupt-managing functions. */
 	struct {
 		bool enabled;
@@ -288,6 +300,11 @@ struct intel_guc {
 #endif
 };
 
+struct intel_guc_tlb_wait {
+	struct wait_queue_head wq;
+	bool busy;
+};
+
 /*
  * GuC version number components are only 8-bit, so converting to a 32bit 8.8.8
  * integer works.
@@ -515,4 +532,10 @@ void intel_guc_dump_time_info(struct intel_guc *guc, struct drm_printer *p);
 
 int intel_guc_sched_disable_gucid_threshold_max(struct intel_guc *guc);
 
+bool intel_guc_tlb_invalidation_is_available(struct intel_guc *guc);
+int intel_guc_invalidate_tlb_engines(struct intel_guc *guc);
+int intel_guc_invalidate_tlb_guc(struct intel_guc *guc);
+int intel_guc_tlb_invalidation_done(struct intel_guc *guc,
+				    const u32 *payload, u32 len);
+void wake_up_all_tlb_invalidate(struct intel_guc *guc);
 #endif
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c
index 331cec07c125..a4da0208c883 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c
@@ -1101,8 +1101,8 @@ guc_capture_create_prealloc_nodes(struct intel_guc *guc)
 static int
 guc_capture_extract_reglists(struct intel_guc *guc, struct __guc_capture_bufstate *buf)
 {
-	struct guc_state_capture_group_header_t ghdr = {0};
-	struct guc_state_capture_header_t hdr = {0};
+	struct guc_state_capture_group_header_t ghdr = {};
+	struct guc_state_capture_header_t hdr = {};
 	struct __guc_capture_parsed_output *node = NULL;
 	struct guc_mmio_reg *regs = NULL;
 	int i, numlists, numregs, ret = 0;
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c
index c33210ead1ef..89e314b3756b 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c
@@ -103,6 +103,33 @@ enum { CTB_SEND = 0, CTB_RECV = 1 };
 
 enum { CTB_OWNER_HOST = 0 };
 
+/*
+ * Some H2G commands involve a synchronous response that the driver needs
+ * to wait for. In such cases, a timeout is required to prevent the driver
+ * from waiting forever in the case of an error (either no error response
+ * is defined in the protocol or something has died and requires a reset).
+ * The specific command may be defined as having a time bound response but
+ * the CT is a queue and that time guarantee only starts from the point
+ * when the command reaches the head of the queue and is processed by GuC.
+ *
+ * Ideally there would be a helper to report the progress of a given
+ * command through the CT. However, that would require a significant
+ * amount of work in the CT layer. In the meantime, provide a reasonable
+ * estimation of the worst case latency it should take for the entire
+ * queue to drain. And therefore, how long a caller should wait before
+ * giving up on their request. The current estimate is based on empirical
+ * measurement of a test that fills the buffer with context creation and
+ * destruction requests as they seem to be the slowest operation.
+ */
+long intel_guc_ct_max_queue_time_jiffies(void)
+{
+	/*
+	 * A 4KB buffer full of context destroy commands takes a little
+	 * over a second to process so bump that to 2s to be super safe.
+	 */
+	return (CTB_H2G_BUFFER_SIZE * HZ) / SZ_2K;
+}
+
 static void ct_receive_tasklet_func(struct tasklet_struct *t);
 static void ct_incoming_request_worker_func(struct work_struct *w);
 
@@ -1115,6 +1142,9 @@ static int ct_process_request(struct intel_guc_ct *ct, struct ct_incoming_msg *r
 	case INTEL_GUC_ACTION_NOTIFY_EXCEPTION:
 		ret = intel_guc_crash_process_msg(guc, action);
 		break;
+	case INTEL_GUC_ACTION_TLB_INVALIDATION_DONE:
+		ret = intel_guc_tlb_invalidation_done(guc, payload, len);
+		break;
 	default:
 		ret = -EOPNOTSUPP;
 		break;
@@ -1186,9 +1216,17 @@ static int ct_handle_event(struct intel_guc_ct *ct, struct ct_incoming_msg *requ
 	switch (action) {
 	case INTEL_GUC_ACTION_SCHED_CONTEXT_MODE_DONE:
 	case INTEL_GUC_ACTION_DEREGISTER_CONTEXT_DONE:
+	case INTEL_GUC_ACTION_TLB_INVALIDATION_DONE:
 		g2h_release_space(ct, request->size);
 	}
 
+	/*
+	 * TLB invalidation responses must be handled immediately as processing
+	 * of other G2H notifications may be blocked by an invalidation request.
+	 */
+	if (action == INTEL_GUC_ACTION_TLB_INVALIDATION_DONE)
+		return ct_process_request(ct, request);
+
 	spin_lock_irqsave(&ct->requests.lock, flags);
 	list_add_tail(&request->link, &ct->requests.incoming);
 	spin_unlock_irqrestore(&ct->requests.lock, flags);
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.h
index 58e42901ff49..2c4bb9a941be 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.h
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.h
@@ -104,6 +104,8 @@ struct intel_guc_ct {
 #endif
 };
 
+long intel_guc_ct_max_queue_time_jiffies(void);
+
 void intel_guc_ct_init_early(struct intel_guc_ct *ct);
 int intel_guc_ct_init(struct intel_guc_ct *ct);
 void intel_guc_ct_fini(struct intel_guc_ct *ct);
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_fwif.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_fwif.h
index 123ad75d2eb2..8ae1846431da 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_fwif.h
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_fwif.h
@@ -22,6 +22,7 @@
 /* Payload length only i.e. don't include G2H header length */
 #define G2H_LEN_DW_SCHED_CONTEXT_MODE_SET	2
 #define G2H_LEN_DW_DEREGISTER_CONTEXT		1
+#define G2H_LEN_DW_INVALIDATE_TLB		1
 
 #define GUC_CONTEXT_DISABLE		0
 #define GUC_CONTEXT_ENABLE		1
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
index 2cce5ec1ff00..d37698bd6b91 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
@@ -32,6 +32,7 @@
 
 #include "i915_drv.h"
 #include "i915_reg.h"
+#include "i915_irq.h"
 #include "i915_trace.h"
 
 /**
@@ -1796,6 +1797,20 @@ next_context:
 	intel_context_put(parent);
 }
 
+void wake_up_all_tlb_invalidate(struct intel_guc *guc)
+{
+	struct intel_guc_tlb_wait *wait;
+	unsigned long i;
+
+	if (!intel_guc_tlb_invalidation_is_available(guc))
+		return;
+
+	xa_lock_irq(&guc->tlb_lookup);
+	xa_for_each(&guc->tlb_lookup, i, wait)
+		wake_up(&wait->wq);
+	xa_unlock_irq(&guc->tlb_lookup);
+}
+
 void intel_guc_submission_reset(struct intel_guc *guc, intel_engine_mask_t stalled)
 {
 	struct intel_context *ce;
@@ -1921,6 +1936,12 @@ void intel_guc_submission_cancel_requests(struct intel_guc *guc)
 
 	/* GuC is blown away, drop all references to contexts */
 	xa_destroy(&guc->context_lookup);
+
+	/*
+	 * Wedged GT won't respond to any TLB invalidation request. Simply
+	 * release all the blocked waiters.
+	 */
+	wake_up_all_tlb_invalidate(guc);
 }
 
 void intel_guc_submission_reset_finish(struct intel_guc *guc)
@@ -1943,11 +1964,65 @@ void intel_guc_submission_reset_finish(struct intel_guc *guc)
 	intel_guc_global_policies_update(guc);
 	enable_submission(guc);
 	intel_gt_unpark_heartbeats(guc_to_gt(guc));
+
+	/*
+	 * The full GT reset will have cleared the TLB caches and flushed the
+	 * G2H message queue; we can release all the blocked waiters.
+	 */
+	wake_up_all_tlb_invalidate(guc);
 }
 
 static void destroyed_worker_func(struct work_struct *w);
 static void reset_fail_worker_func(struct work_struct *w);
 
+bool intel_guc_tlb_invalidation_is_available(struct intel_guc *guc)
+{
+	return HAS_GUC_TLB_INVALIDATION(guc_to_gt(guc)->i915) &&
+		intel_guc_is_ready(guc);
+}
+
+static int init_tlb_lookup(struct intel_guc *guc)
+{
+	struct intel_guc_tlb_wait *wait;
+	int err;
+
+	if (!HAS_GUC_TLB_INVALIDATION(guc_to_gt(guc)->i915))
+		return 0;
+
+	xa_init_flags(&guc->tlb_lookup, XA_FLAGS_ALLOC);
+
+	wait = kzalloc(sizeof(*wait), GFP_KERNEL);
+	if (!wait)
+		return -ENOMEM;
+
+	init_waitqueue_head(&wait->wq);
+
+	/* Preallocate a shared id for use under memory pressure. */
+	err = xa_alloc_cyclic_irq(&guc->tlb_lookup, &guc->serial_slot, wait,
+				  xa_limit_32b, &guc->next_seqno, GFP_KERNEL);
+	if (err < 0) {
+		kfree(wait);
+		return err;
+	}
+
+	return 0;
+}
+
+static void fini_tlb_lookup(struct intel_guc *guc)
+{
+	struct intel_guc_tlb_wait *wait;
+
+	if (!HAS_GUC_TLB_INVALIDATION(guc_to_gt(guc)->i915))
+		return;
+
+	wait = xa_load(&guc->tlb_lookup, guc->serial_slot);
+	if (wait && wait->busy)
+		guc_err(guc, "Unexpected busy item in tlb_lookup on fini\n");
+	kfree(wait);
+
+	xa_destroy(&guc->tlb_lookup);
+}
+
 /*
  * Set up the memory resources to be shared with the GuC (via the GGTT)
  * at firmware loading time.
@@ -1966,11 +2041,15 @@ int intel_guc_submission_init(struct intel_guc *guc)
 			return ret;
 	}
 
+	ret = init_tlb_lookup(guc);
+	if (ret)
+		goto destroy_pool;
+
 	guc->submission_state.guc_ids_bitmap =
 		bitmap_zalloc(NUMBER_MULTI_LRC_GUC_ID(guc), GFP_KERNEL);
 	if (!guc->submission_state.guc_ids_bitmap) {
 		ret = -ENOMEM;
-		goto destroy_pool;
+		goto destroy_tlb;
 	}
 
 	guc->timestamp.ping_delay = (POLL_TIME_CLKS / gt->clock_frequency + 1) * HZ;
@@ -1979,9 +2058,10 @@ int intel_guc_submission_init(struct intel_guc *guc)
 
 	return 0;
 
+destroy_tlb:
+	fini_tlb_lookup(guc);
 destroy_pool:
 	guc_lrc_desc_pool_destroy_v69(guc);
-
 	return ret;
 }
 
@@ -1994,6 +2074,7 @@ void intel_guc_submission_fini(struct intel_guc *guc)
 	guc_lrc_desc_pool_destroy_v69(guc);
 	i915_sched_engine_put(guc->sched_engine);
 	bitmap_free(guc->submission_state.guc_ids_bitmap);
+	fini_tlb_lookup(guc);
 	guc->submission_initialized = false;
 }
 
@@ -4624,6 +4705,154 @@ g2h_context_lookup(struct intel_guc *guc, u32 ctx_id)
 	return ce;
 }
 
+static void wait_wake_outstanding_tlb_g2h(struct intel_guc *guc, u32 seqno)
+{
+	struct intel_guc_tlb_wait *wait;
+	unsigned long flags;
+
+	xa_lock_irqsave(&guc->tlb_lookup, flags);
+	wait = xa_load(&guc->tlb_lookup, seqno);
+
+	if (wait)
+		wake_up(&wait->wq);
+	else
+		guc_dbg(guc,
+			"Stale TLB invalidation response with seqno %d\n", seqno);
+
+	xa_unlock_irqrestore(&guc->tlb_lookup, flags);
+}
+
+int intel_guc_tlb_invalidation_done(struct intel_guc *guc,
+				    const u32 *payload, u32 len)
+{
+	if (len < 1)
+		return -EPROTO;
+
+	wait_wake_outstanding_tlb_g2h(guc, payload[0]);
+	return 0;
+}
+
+static long must_wait_woken(struct wait_queue_entry *wq_entry, long timeout)
+{
+	/*
+	 * This is equivalent to wait_woken() with the exception that
+	 * we do not wake up early if the kthread task has been completed.
+	 * As we are called from page reclaim in any task context,
+	 * we may be invoked from stopped kthreads, but we *must*
+	 * complete the wait from the HW.
+	 */
+	do {
+		set_current_state(TASK_UNINTERRUPTIBLE);
+		if (wq_entry->flags & WQ_FLAG_WOKEN)
+			break;
+
+		timeout = schedule_timeout(timeout);
+	} while (timeout);
+
+	/* See wait_woken() and woken_wake_function() */
+	__set_current_state(TASK_RUNNING);
+	smp_store_mb(wq_entry->flags, wq_entry->flags & ~WQ_FLAG_WOKEN);
+
+	return timeout;
+}
+
+static bool intel_gt_is_enabled(const struct intel_gt *gt)
+{
+	/* Check if GT is wedged or suspended */
+	if (intel_gt_is_wedged(gt) || !intel_irqs_enabled(gt->i915))
+		return false;
+	return true;
+}
+
+static int guc_send_invalidate_tlb(struct intel_guc *guc,
+				   enum intel_guc_tlb_invalidation_type type)
+{
+	struct intel_guc_tlb_wait _wq, *wq = &_wq;
+	struct intel_gt *gt = guc_to_gt(guc);
+	DEFINE_WAIT_FUNC(wait, woken_wake_function);
+	int err;
+	u32 seqno;
+	u32 action[] = {
+		INTEL_GUC_ACTION_TLB_INVALIDATION,
+		0,
+		REG_FIELD_PREP(INTEL_GUC_TLB_INVAL_TYPE_MASK, type) |
+			REG_FIELD_PREP(INTEL_GUC_TLB_INVAL_MODE_MASK,
+				       INTEL_GUC_TLB_INVAL_MODE_HEAVY) |
+			INTEL_GUC_TLB_INVAL_FLUSH_CACHE,
+	};
+	u32 size = ARRAY_SIZE(action);
+
+	/*
+	 * Early guard against GT enablement.  TLB invalidation should not be
+	 * attempted if the GT is disabled due to suspend/wedge.
+	 */
+	if (!intel_gt_is_enabled(gt))
+		return -EINVAL;
+
+	init_waitqueue_head(&_wq.wq);
+
+	if (xa_alloc_cyclic_irq(&guc->tlb_lookup, &seqno, wq,
+				xa_limit_32b, &guc->next_seqno,
+				GFP_ATOMIC | __GFP_NOWARN) < 0) {
+		/* Under severe memory pressure? Serialise TLB allocations */
+		xa_lock_irq(&guc->tlb_lookup);
+		wq = xa_load(&guc->tlb_lookup, guc->serial_slot);
+		wait_event_lock_irq(wq->wq,
+				    !READ_ONCE(wq->busy),
+				    guc->tlb_lookup.xa_lock);
+		/*
+		 * Update wq->busy under lock to ensure only one waiter can
+		 * issue the TLB invalidation command using the serial slot at a
+		 * time. The condition is set to true before releasing the lock
+		 * so that other caller continue to wait until woken up again.
+		 */
+		wq->busy = true;
+		xa_unlock_irq(&guc->tlb_lookup);
+
+		seqno = guc->serial_slot;
+	}
+
+	action[1] = seqno;
+
+	add_wait_queue(&wq->wq, &wait);
+
+	/* This is a critical reclaim path and thus we must loop here. */
+	err = intel_guc_send_busy_loop(guc, action, size, G2H_LEN_DW_INVALIDATE_TLB, true);
+	if (err)
+		goto out;
+
+	/*
+	 * Late guard against GT enablement.  It is not an error for the TLB
+	 * invalidation to time out if the GT is disabled during the process
+	 * due to suspend/wedge.  In fact, the TLB invalidation is cancelled
+	 * in this case.
+	 */
+	if (!must_wait_woken(&wait, intel_guc_ct_max_queue_time_jiffies()) &&
+	    intel_gt_is_enabled(gt)) {
+		guc_err(guc,
+			"TLB invalidation response timed out for seqno %u\n", seqno);
+		err = -ETIME;
+	}
+out:
+	remove_wait_queue(&wq->wq, &wait);
+	if (seqno != guc->serial_slot)
+		xa_erase_irq(&guc->tlb_lookup, seqno);
+
+	return err;
+}
+
+/* Send a H2G command to invalidate the TLBs at engine level and beyond. */
+int intel_guc_invalidate_tlb_engines(struct intel_guc *guc)
+{
+	return guc_send_invalidate_tlb(guc, INTEL_GUC_TLB_INVAL_ENGINES);
+}
+
+/* Send a H2G command to invalidate the GuC's internal TLB. */
+int intel_guc_invalidate_tlb_guc(struct intel_guc *guc)
+{
+	return guc_send_invalidate_tlb(guc, INTEL_GUC_TLB_INVAL_GUC);
+}
+
 int intel_guc_deregister_done_process_msg(struct intel_guc *guc,
 					  const u32 *msg,
 					  u32 len)
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc.c b/drivers/gpu/drm/i915/gt/uc/intel_uc.c
index 98b103375b7a..27f6561dd731 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_uc.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc.c
@@ -688,6 +688,8 @@ void intel_uc_suspend(struct intel_uc *uc)
 	/* flush the GSC worker */
 	intel_gsc_uc_flush_work(&uc->gsc);
 
+	wake_up_all_tlb_invalidate(guc);
+
 	if (!intel_guc_is_ready(guc)) {
 		guc->interrupts.enabled = false;
 		return;
@@ -736,6 +738,11 @@ static int __uc_resume(struct intel_uc *uc, bool enable_communication)
 
 	intel_gsc_uc_resume(&uc->gsc);
 
+	if (intel_guc_tlb_invalidation_is_available(guc)) {
+		intel_guc_invalidate_tlb_engines(guc);
+		intel_guc_invalidate_tlb_guc(guc);
+	}
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index cb60fc9cf873..6a2a78c61f21 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -794,6 +794,8 @@ IS_SUBPLATFORM(const struct drm_i915_private *i915,
 #define HAS_GUC_DEPRIVILEGE(i915) \
 	(INTEL_INFO(i915)->has_guc_deprivilege)
 
+#define HAS_GUC_TLB_INVALIDATION(i915)	(INTEL_INFO(i915)->has_guc_tlb_invalidation)
+
 #define HAS_3D_PIPELINE(i915)	(INTEL_INFO(i915)->has_3d_pipeline)
 
 #define HAS_ONE_EU_PER_FUSE_BIT(i915)	(INTEL_INFO(i915)->has_one_eu_per_fuse_bit)
diff --git a/drivers/gpu/drm/i915/i915_pci.c b/drivers/gpu/drm/i915/i915_pci.c
index bb577345bc0e..8b4fdeabb12a 100644
--- a/drivers/gpu/drm/i915/i915_pci.c
+++ b/drivers/gpu/drm/i915/i915_pci.c
@@ -829,6 +829,7 @@ static const struct intel_device_info mtl_info = {
 	.has_flat_ccs = 0,
 	.has_gmd_id = 1,
 	.has_guc_deprivilege = 1,
+	.has_guc_tlb_invalidation = 1,
 	.has_llc = 0,
 	.has_mslice_steering = 0,
 	.has_snoop = 1,
diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c
index 8f7ab64feec0..7178f298b3e6 100644
--- a/drivers/gpu/drm/i915/i915_perf.c
+++ b/drivers/gpu/drm/i915/i915_perf.c
@@ -483,8 +483,7 @@ static void oa_report_id_clear(struct i915_perf_stream *stream, u32 *report)
 static bool oa_report_ctx_invalid(struct i915_perf_stream *stream, void *report)
 {
 	return !(oa_report_id(stream, report) &
-	       stream->perf->gen8_valid_ctx_bit) &&
-	       GRAPHICS_VER(stream->perf->i915) <= 11;
+	       stream->perf->gen8_valid_ctx_bit);
 }
 
 static u64 oa_timestamp(struct i915_perf_stream *stream, void *report)
@@ -5047,6 +5046,7 @@ static void i915_perf_init_info(struct drm_i915_private *i915)
 		perf->gen8_valid_ctx_bit = BIT(16);
 		break;
 	case 12:
+		perf->gen8_valid_ctx_bit = BIT(16);
 		/*
 		 * Calculate offset at runtime in oa_pin_context for gen12 and
 		 * cache the value in perf->ctx_oactxctrl_offset.
diff --git a/drivers/gpu/drm/i915/intel_device_info.h b/drivers/gpu/drm/i915/intel_device_info.h
index 39817490b13f..eba2f0b919c8 100644
--- a/drivers/gpu/drm/i915/intel_device_info.h
+++ b/drivers/gpu/drm/i915/intel_device_info.h
@@ -153,6 +153,7 @@ enum intel_ppgtt_type {
 	func(has_heci_pxp); \
 	func(has_heci_gscfi); \
 	func(has_guc_deprivilege); \
+	func(has_guc_tlb_invalidation); \
 	func(has_l3_ccs_read); \
 	func(has_l3_dpf); \
 	func(has_llc); \
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_gsccs.c b/drivers/gpu/drm/i915/pxp/intel_pxp_gsccs.c
index 27402ecf0457..75df959b0aa0 100644
--- a/drivers/gpu/drm/i915/pxp/intel_pxp_gsccs.c
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_gsccs.c
@@ -209,8 +209,8 @@ int intel_pxp_gsccs_create_session(struct intel_pxp *pxp,
 				   int arb_session_id)
 {
 	struct drm_i915_private *i915 = pxp->ctrl_gt->i915;
-	struct pxp43_create_arb_in msg_in = {0};
-	struct pxp43_create_arb_out msg_out = {0};
+	struct pxp43_create_arb_in msg_in = {};
+	struct pxp43_create_arb_out msg_out = {};
 	int ret;
 
 	msg_in.header.api_version = PXP_APIVER(4, 3);
@@ -247,8 +247,8 @@ int intel_pxp_gsccs_create_session(struct intel_pxp *pxp,
 void intel_pxp_gsccs_end_arb_fw_session(struct intel_pxp *pxp, u32 session_id)
 {
 	struct drm_i915_private *i915 = pxp->ctrl_gt->i915;
-	struct pxp42_inv_stream_key_in msg_in = {0};
-	struct pxp42_inv_stream_key_out msg_out = {0};
+	struct pxp42_inv_stream_key_in msg_in = {};
+	struct pxp42_inv_stream_key_out msg_out = {};
 	int ret = 0;
 
 	/*
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_huc.c b/drivers/gpu/drm/i915/pxp/intel_pxp_huc.c
index 5eedce916942..0e609547bef8 100644
--- a/drivers/gpu/drm/i915/pxp/intel_pxp_huc.c
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_huc.c
@@ -18,8 +18,8 @@ int intel_pxp_huc_load_and_auth(struct intel_pxp *pxp)
 {
 	struct intel_gt *gt;
 	struct intel_huc *huc;
-	struct pxp43_start_huc_auth_in huc_in = {0};
-	struct pxp43_huc_auth_out huc_out = {0};
+	struct pxp43_start_huc_auth_in huc_in = {};
+	struct pxp43_huc_auth_out huc_out = {};
 	dma_addr_t huc_phys_addr;
 	u8 client_id = 0;
 	u8 fence_id = 0;
diff --git a/drivers/gpu/drm/i915/pxp/intel_pxp_tee.c b/drivers/gpu/drm/i915/pxp/intel_pxp_tee.c
index bb58fa9579b8..9327701da1b5 100644
--- a/drivers/gpu/drm/i915/pxp/intel_pxp_tee.c
+++ b/drivers/gpu/drm/i915/pxp/intel_pxp_tee.c
@@ -327,8 +327,8 @@ int intel_pxp_tee_cmd_create_arb_session(struct intel_pxp *pxp,
 					 int arb_session_id)
 {
 	struct drm_i915_private *i915 = pxp->ctrl_gt->i915;
-	struct pxp42_create_arb_in msg_in = {0};
-	struct pxp42_create_arb_out msg_out = {0};
+	struct pxp42_create_arb_in msg_in = {};
+	struct pxp42_create_arb_out msg_out = {};
 	int ret;
 
 	msg_in.header.api_version = PXP_APIVER(4, 2);
@@ -365,8 +365,8 @@ int intel_pxp_tee_cmd_create_arb_session(struct intel_pxp *pxp,
 void intel_pxp_tee_end_arb_fw_session(struct intel_pxp *pxp, u32 session_id)
 {
 	struct drm_i915_private *i915 = pxp->ctrl_gt->i915;
-	struct pxp42_inv_stream_key_in msg_in = {0};
-	struct pxp42_inv_stream_key_out msg_out = {0};
+	struct pxp42_inv_stream_key_in msg_in = {};
+	struct pxp42_inv_stream_key_out msg_out = {};
 	int ret, trials = 0;
 
 try_again: