16 files changed, 100 insertions, 17 deletions

diff --git a/drivers/gpu/drm/xe/tests/xe_mocs.c b/drivers/gpu/drm/xe/tests/xe_mocs.c
index ef1e5256c56a..0e502feaca81 100644
--- a/drivers/gpu/drm/xe/tests/xe_mocs.c
+++ b/drivers/gpu/drm/xe/tests/xe_mocs.c
@@ -46,8 +46,11 @@ static void read_l3cc_table(struct xe_gt *gt,
 	unsigned int fw_ref, i;
 	u32 reg_val;
 
-	fw_ref = xe_force_wake_get(gt_to_fw(gt), XE_FW_GT);
-	KUNIT_ASSERT_NE_MSG(test, fw_ref, 0, "Forcewake Failed.\n");
+	fw_ref = xe_force_wake_get(gt_to_fw(gt), XE_FORCEWAKE_ALL);
+	if (!xe_force_wake_ref_has_domain(fw_ref, XE_FORCEWAKE_ALL)) {
+		xe_force_wake_put(gt_to_fw(gt), fw_ref);
+		KUNIT_ASSERT_TRUE_MSG(test, true, "Forcewake Failed.\n");
+	}
 
 	for (i = 0; i < info->num_mocs_regs; i++) {
 		if (!(i & 1)) {
diff --git a/drivers/gpu/drm/xe/xe_eu_stall.c b/drivers/gpu/drm/xe/xe_eu_stall.c
index f2bb9168967c..e2bb156c71fb 100644
--- a/drivers/gpu/drm/xe/xe_eu_stall.c
+++ b/drivers/gpu/drm/xe/xe_eu_stall.c
@@ -52,6 +52,8 @@ struct xe_eu_stall_data_stream {
 
 	struct xe_gt *gt;
 	struct xe_bo *bo;
+	/* Lock to protect data buffer pointers */
+	struct mutex xecore_buf_lock;
 	struct per_xecore_buf *xecore_buf;
 	struct {
 		bool reported_to_user;
@@ -208,6 +210,9 @@ int xe_eu_stall_init(struct xe_gt *gt)
 	struct xe_device *xe = gt_to_xe(gt);
 	int ret;
 
+	if (!xe_eu_stall_supported_on_platform(xe))
+		return 0;
+
 	gt->eu_stall = kzalloc(sizeof(*gt->eu_stall), GFP_KERNEL);
 	if (!gt->eu_stall) {
 		ret = -ENOMEM;
@@ -378,7 +383,7 @@ static bool eu_stall_data_buf_poll(struct xe_eu_stall_data_stream *stream)
 	u16 group, instance;
 	unsigned int xecore;
 
-	mutex_lock(&gt->eu_stall->stream_lock);
+	mutex_lock(&stream->xecore_buf_lock);
 	for_each_dss_steering(xecore, gt, group, instance) {
 		xecore_buf = &stream->xecore_buf[xecore];
 		read_ptr = xecore_buf->read;
@@ -396,7 +401,7 @@ static bool eu_stall_data_buf_poll(struct xe_eu_stall_data_stream *stream)
 			set_bit(xecore, stream->data_drop.mask);
 		xecore_buf->write = write_ptr;
 	}
-	mutex_unlock(&gt->eu_stall->stream_lock);
+	mutex_unlock(&stream->xecore_buf_lock);
 
 	return min_data_present;
 }
@@ -511,11 +516,13 @@ static ssize_t xe_eu_stall_stream_read_locked(struct xe_eu_stall_data_stream *st
 	unsigned int xecore;
 	int ret = 0;
 
+	mutex_lock(&stream->xecore_buf_lock);
 	if (bitmap_weight(stream->data_drop.mask, XE_MAX_DSS_FUSE_BITS)) {
 		if (!stream->data_drop.reported_to_user) {
 			stream->data_drop.reported_to_user = true;
 			xe_gt_dbg(gt, "EU stall data dropped in XeCores: %*pb\n",
 				  XE_MAX_DSS_FUSE_BITS, stream->data_drop.mask);
+			mutex_unlock(&stream->xecore_buf_lock);
 			return -EIO;
 		}
 		stream->data_drop.reported_to_user = false;
@@ -527,6 +534,7 @@ static ssize_t xe_eu_stall_stream_read_locked(struct xe_eu_stall_data_stream *st
 		if (ret || count == total_size)
 			break;
 	}
+	mutex_unlock(&stream->xecore_buf_lock);
 	return total_size ?: (ret ?: -EAGAIN);
 }
 
@@ -583,6 +591,7 @@ static void xe_eu_stall_stream_free(struct xe_eu_stall_data_stream *stream)
 {
 	struct xe_gt *gt = stream->gt;
 
+	mutex_destroy(&stream->xecore_buf_lock);
 	gt->eu_stall->stream = NULL;
 	kfree(stream);
 }
@@ -718,6 +727,7 @@ static int xe_eu_stall_stream_init(struct xe_eu_stall_data_stream *stream,
 	}
 
 	init_waitqueue_head(&stream->poll_wq);
+	mutex_init(&stream->xecore_buf_lock);
 	INIT_DELAYED_WORK(&stream->buf_poll_work, eu_stall_data_buf_poll_work_fn);
 	stream->per_xecore_buf_size = per_xecore_buf_size;
 	stream->sampling_rate_mult = props->sampling_rate_mult;
diff --git a/drivers/gpu/drm/xe/xe_eu_stall.h b/drivers/gpu/drm/xe/xe_eu_stall.h
index ed9d0f233566..d1c76e503799 100644
--- a/drivers/gpu/drm/xe/xe_eu_stall.h
+++ b/drivers/gpu/drm/xe/xe_eu_stall.h
@@ -7,6 +7,7 @@
 #define __XE_EU_STALL_H__
 
 #include "xe_gt_types.h"
+#include "xe_sriov.h"
 
 size_t xe_eu_stall_get_per_xecore_buf_size(void);
 size_t xe_eu_stall_data_record_size(struct xe_device *xe);
@@ -19,6 +20,6 @@ int xe_eu_stall_stream_open(struct drm_device *dev,
 
 static inline bool xe_eu_stall_supported_on_platform(struct xe_device *xe)
 {
-	return xe->info.platform == XE_PVC || GRAPHICS_VER(xe) >= 20;
+	return !IS_SRIOV_VF(xe) && (xe->info.platform == XE_PVC || GRAPHICS_VER(xe) >= 20);
 }
 #endif
diff --git a/drivers/gpu/drm/xe/xe_gsc.c b/drivers/gpu/drm/xe/xe_gsc.c
index fd41113f8572..0bcf97063ff6 100644
--- a/drivers/gpu/drm/xe/xe_gsc.c
+++ b/drivers/gpu/drm/xe/xe_gsc.c
@@ -555,6 +555,28 @@ void xe_gsc_wait_for_worker_completion(struct xe_gsc *gsc)
 		flush_work(&gsc->work);
 }
 
+void xe_gsc_stop_prepare(struct xe_gsc *gsc)
+{
+	struct xe_gt *gt = gsc_to_gt(gsc);
+	int ret;
+
+	if (!xe_uc_fw_is_loadable(&gsc->fw) || xe_uc_fw_is_in_error_state(&gsc->fw))
+		return;
+
+	xe_force_wake_assert_held(gt_to_fw(gt), XE_FW_GSC);
+
+	/*
+	 * If the GSC FW load or the proxy init are interrupted, the only way
+	 * to recover it is to do an FLR and reload the GSC from scratch.
+	 * Therefore, let's wait for the init to complete before stopping
+	 * operations. The proxy init is the last step, so we can just wait on
+	 * that
+	 */
+	ret = xe_gsc_wait_for_proxy_init_done(gsc);
+	if (ret)
+		xe_gt_err(gt, "failed to wait for GSC init completion before uc stop\n");
+}
+
 /*
  * wa_14015076503: if the GSC FW is loaded, we need to alert it before doing a
  * GSC engine reset by writing a notification bit in the GS1 register and then
diff --git a/drivers/gpu/drm/xe/xe_gsc.h b/drivers/gpu/drm/xe/xe_gsc.h
index d99f66c38075..b8b8e0810ad9 100644
--- a/drivers/gpu/drm/xe/xe_gsc.h
+++ b/drivers/gpu/drm/xe/xe_gsc.h
@@ -16,6 +16,7 @@ struct xe_hw_engine;
 int xe_gsc_init(struct xe_gsc *gsc);
 int xe_gsc_init_post_hwconfig(struct xe_gsc *gsc);
 void xe_gsc_wait_for_worker_completion(struct xe_gsc *gsc);
+void xe_gsc_stop_prepare(struct xe_gsc *gsc);
 void xe_gsc_load_start(struct xe_gsc *gsc);
 void xe_gsc_hwe_irq_handler(struct xe_hw_engine *hwe, u16 intr_vec);
 
diff --git a/drivers/gpu/drm/xe/xe_gsc_proxy.c b/drivers/gpu/drm/xe/xe_gsc_proxy.c
index 8cf70b228ff3..d0519cd6704a 100644
--- a/drivers/gpu/drm/xe/xe_gsc_proxy.c
+++ b/drivers/gpu/drm/xe/xe_gsc_proxy.c
@@ -71,6 +71,17 @@ bool xe_gsc_proxy_init_done(struct xe_gsc *gsc)
 	       HECI1_FWSTS1_PROXY_STATE_NORMAL;
 }
 
+int xe_gsc_wait_for_proxy_init_done(struct xe_gsc *gsc)
+{
+	struct xe_gt *gt = gsc_to_gt(gsc);
+
+	/* Proxy init can take up to 500ms, so wait double that for safety */
+	return xe_mmio_wait32(&gt->mmio, HECI_FWSTS1(MTL_GSC_HECI1_BASE),
+			      HECI1_FWSTS1_CURRENT_STATE,
+			      HECI1_FWSTS1_PROXY_STATE_NORMAL,
+			      USEC_PER_SEC, NULL, false);
+}
+
 static void __gsc_proxy_irq_rmw(struct xe_gsc *gsc, u32 clr, u32 set)
 {
 	struct xe_gt *gt = gsc_to_gt(gsc);
diff --git a/drivers/gpu/drm/xe/xe_gsc_proxy.h b/drivers/gpu/drm/xe/xe_gsc_proxy.h
index fdef56995cd4..765602221dbc 100644
--- a/drivers/gpu/drm/xe/xe_gsc_proxy.h
+++ b/drivers/gpu/drm/xe/xe_gsc_proxy.h
@@ -12,6 +12,7 @@ struct xe_gsc;
 
 int xe_gsc_proxy_init(struct xe_gsc *gsc);
 bool xe_gsc_proxy_init_done(struct xe_gsc *gsc);
+int xe_gsc_wait_for_proxy_init_done(struct xe_gsc *gsc);
 int xe_gsc_proxy_start(struct xe_gsc *gsc);
 
 int xe_gsc_proxy_request_handler(struct xe_gsc *gsc);
diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c
index 10a9e3c72b36..66198cf2662c 100644
--- a/drivers/gpu/drm/xe/xe_gt.c
+++ b/drivers/gpu/drm/xe/xe_gt.c
@@ -857,7 +857,7 @@ void xe_gt_suspend_prepare(struct xe_gt *gt)
 
 	fw_ref = xe_force_wake_get(gt_to_fw(gt), XE_FORCEWAKE_ALL);
 
-	xe_uc_stop_prepare(&gt->uc);
+	xe_uc_suspend_prepare(&gt->uc);
 
 	xe_force_wake_put(gt_to_fw(gt), fw_ref);
 }
diff --git a/drivers/gpu/drm/xe/xe_gt_debugfs.c b/drivers/gpu/drm/xe/xe_gt_debugfs.c
index 2d63a69cbfa3..f7005a3643e6 100644
--- a/drivers/gpu/drm/xe/xe_gt_debugfs.c
+++ b/drivers/gpu/drm/xe/xe_gt_debugfs.c
@@ -92,22 +92,23 @@ static int hw_engines(struct xe_gt *gt, struct drm_printer *p)
 	struct xe_hw_engine *hwe;
 	enum xe_hw_engine_id id;
 	unsigned int fw_ref;
+	int ret = 0;
 
 	xe_pm_runtime_get(xe);
 	fw_ref = xe_force_wake_get(gt_to_fw(gt), XE_FORCEWAKE_ALL);
 	if (!xe_force_wake_ref_has_domain(fw_ref, XE_FORCEWAKE_ALL)) {
-		xe_pm_runtime_put(xe);
-		xe_force_wake_put(gt_to_fw(gt), fw_ref);
-		return -ETIMEDOUT;
+		ret = -ETIMEDOUT;
+		goto fw_put;
 	}
 
 	for_each_hw_engine(hwe, gt, id)
 		xe_hw_engine_print(hwe, p);
 
+fw_put:
 	xe_force_wake_put(gt_to_fw(gt), fw_ref);
 	xe_pm_runtime_put(xe);
 
-	return 0;
+	return ret;
 }
 
 static int powergate_info(struct xe_gt *gt, struct drm_printer *p)
diff --git a/drivers/gpu/drm/xe/xe_gt_pagefault.c b/drivers/gpu/drm/xe/xe_gt_pagefault.c
index c5ad9a0a89c2..0c22b3a36655 100644
--- a/drivers/gpu/drm/xe/xe_gt_pagefault.c
+++ b/drivers/gpu/drm/xe/xe_gt_pagefault.c
@@ -435,9 +435,16 @@ static int xe_alloc_pf_queue(struct xe_gt *gt, struct pf_queue *pf_queue)
 	num_eus = bitmap_weight(gt->fuse_topo.eu_mask_per_dss,
 				XE_MAX_EU_FUSE_BITS) * num_dss;
 
-	/* user can issue separate page faults per EU and per CS */
+	/*
+	 * user can issue separate page faults per EU and per CS
+	 *
+	 * XXX: Multiplier required as compute UMD are getting PF queue errors
+	 * without it. Follow on why this multiplier is required.
+	 */
+#define PF_MULTIPLIER	8
 	pf_queue->num_dw =
-		(num_eus + XE_NUM_HW_ENGINES) * PF_MSG_LEN_DW;
+		(num_eus + XE_NUM_HW_ENGINES) * PF_MSG_LEN_DW * PF_MULTIPLIER;
+#undef PF_MULTIPLIER
 
 	pf_queue->gt = gt;
 	pf_queue->data = devm_kcalloc(xe->drm.dev, pf_queue->num_dw,
diff --git a/drivers/gpu/drm/xe/xe_guc_capture.c b/drivers/gpu/drm/xe/xe_guc_capture.c
index f6d523e4c5fe..9095618648bc 100644
--- a/drivers/gpu/drm/xe/xe_guc_capture.c
+++ b/drivers/gpu/drm/xe/xe_guc_capture.c
@@ -359,7 +359,7 @@ static void __fill_ext_reg(struct __guc_mmio_reg_descr *ext,
 
 	ext->reg = XE_REG(extlist->reg.__reg.addr);
 	ext->flags = FIELD_PREP(GUC_REGSET_STEERING_NEEDED, 1);
-	ext->flags = FIELD_PREP(GUC_REGSET_STEERING_GROUP, slice_id);
+	ext->flags |= FIELD_PREP(GUC_REGSET_STEERING_GROUP, slice_id);
 	ext->flags |= FIELD_PREP(GUC_REGSET_STEERING_INSTANCE, subslice_id);
 	ext->regname = extlist->name;
 }
diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index f8c128524d9f..24c578e1170e 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -79,7 +79,7 @@ xe_svm_range_alloc(struct drm_gpusvm *gpusvm)
 
 	range = kzalloc(sizeof(*range), GFP_KERNEL);
 	if (!range)
-		return ERR_PTR(-ENOMEM);
+		return NULL;
 
 	INIT_LIST_HEAD(&range->garbage_collector_link);
 	xe_vm_get(gpusvm_to_vm(gpusvm));
@@ -947,3 +947,15 @@ int xe_devm_add(struct xe_tile *tile, struct xe_vram_region *vr)
 	return 0;
 }
 #endif
+
+/**
+ * xe_svm_flush() - SVM flush
+ * @vm: The VM.
+ *
+ * Flush all SVM actions.
+ */
+void xe_svm_flush(struct xe_vm *vm)
+{
+	if (xe_vm_in_fault_mode(vm))
+		flush_work(&vm->svm.garbage_collector.work);
+}
diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
index e059590e5076..be306fe7aaa4 100644
--- a/drivers/gpu/drm/xe/xe_svm.h
+++ b/drivers/gpu/drm/xe/xe_svm.h
@@ -72,6 +72,9 @@ bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end);
 int xe_svm_bo_evict(struct xe_bo *bo);
 
 void xe_svm_range_debug(struct xe_svm_range *range, const char *operation);
+
+void xe_svm_flush(struct xe_vm *vm);
+
 #else
 static inline bool xe_svm_range_pages_valid(struct xe_svm_range *range)
 {
@@ -124,6 +127,11 @@ static inline
 void xe_svm_range_debug(struct xe_svm_range *range, const char *operation)
 {
 }
+
+static inline void xe_svm_flush(struct xe_vm *vm)
+{
+}
+
 #endif
 
 /**
diff --git a/drivers/gpu/drm/xe/xe_uc.c b/drivers/gpu/drm/xe/xe_uc.c
index c14bd2282044..3a8751a8b92d 100644
--- a/drivers/gpu/drm/xe/xe_uc.c
+++ b/drivers/gpu/drm/xe/xe_uc.c
@@ -244,7 +244,7 @@ void xe_uc_gucrc_disable(struct xe_uc *uc)
 
 void xe_uc_stop_prepare(struct xe_uc *uc)
 {
-	xe_gsc_wait_for_worker_completion(&uc->gsc);
+	xe_gsc_stop_prepare(&uc->gsc);
 	xe_guc_stop_prepare(&uc->guc);
 }
 
@@ -278,6 +278,12 @@ again:
 		goto again;
 }
 
+void xe_uc_suspend_prepare(struct xe_uc *uc)
+{
+	xe_gsc_wait_for_worker_completion(&uc->gsc);
+	xe_guc_stop_prepare(&uc->guc);
+}
+
 int xe_uc_suspend(struct xe_uc *uc)
 {
 	/* GuC submission not enabled, nothing to do */
diff --git a/drivers/gpu/drm/xe/xe_uc.h b/drivers/gpu/drm/xe/xe_uc.h
index 3813c1ede450..c23e6f5e2514 100644
--- a/drivers/gpu/drm/xe/xe_uc.h
+++ b/drivers/gpu/drm/xe/xe_uc.h
@@ -18,6 +18,7 @@ int xe_uc_reset_prepare(struct xe_uc *uc);
 void xe_uc_stop_prepare(struct xe_uc *uc);
 void xe_uc_stop(struct xe_uc *uc);
 int xe_uc_start(struct xe_uc *uc);
+void xe_uc_suspend_prepare(struct xe_uc *uc);
 int xe_uc_suspend(struct xe_uc *uc);
 int xe_uc_sanitize_reset(struct xe_uc *uc);
 void xe_uc_declare_wedged(struct xe_uc *uc);
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index 60303998bd61..367c84b90e9e 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -3312,8 +3312,7 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 	}
 
 	/* Ensure all UNMAPs visible */
-	if (xe_vm_in_fault_mode(vm))
-		flush_work(&vm->svm.garbage_collector.work);
+	xe_svm_flush(vm);
 
 	err = down_write_killable(&vm->lock);
 	if (err)