1 files changed, 50 insertions, 29 deletions

diff --git a/drivers/gpu/drm/i915/gvt/scheduler.c b/drivers/gpu/drm/i915/gvt/scheduler.c
index bada32b..4f7057d 100644
--- a/drivers/gpu/drm/i915/gvt/scheduler.c
+++ b/drivers/gpu/drm/i915/gvt/scheduler.c
@@ -69,8 +69,7 @@ static int populate_shadow_context(struct intel_vgpu_workload *workload)
 	gvt_dbg_sched("ring id %d workload lrca %x", ring_id,
 			workload->ctx_desc.lrca);
 
-	context_page_num = intel_lr_context_size(
-			gvt->dev_priv->engine[ring_id]);
+	context_page_num = gvt->dev_priv->engine[ring_id]->context_size;
 
 	context_page_num = context_page_num >> PAGE_SHIFT;
 
@@ -139,30 +138,42 @@ static int shadow_context_status_change(struct notifier_block *nb,
 	struct intel_gvt *gvt = container_of(nb, struct intel_gvt,
 				shadow_ctx_notifier_block[req->engine->id]);
 	struct intel_gvt_workload_scheduler *scheduler = &gvt->scheduler;
-	struct intel_vgpu_workload *workload =
-		scheduler->current_workload[req->engine->id];
+	enum intel_engine_id ring_id = req->engine->id;
+	struct intel_vgpu_workload *workload;
+
+	if (!is_gvt_request(req)) {
+		spin_lock_bh(&scheduler->mmio_context_lock);
+		if (action == INTEL_CONTEXT_SCHEDULE_IN &&
+		    scheduler->engine_owner[ring_id]) {
+			/* Switch ring from vGPU to host. */
+			intel_gvt_switch_mmio(scheduler->engine_owner[ring_id],
+					      NULL, ring_id);
+			scheduler->engine_owner[ring_id] = NULL;
+		}
+		spin_unlock_bh(&scheduler->mmio_context_lock);
 
-	if (!is_gvt_request(req) || unlikely(!workload))
+		return NOTIFY_OK;
+	}
+
+	workload = scheduler->current_workload[ring_id];
+	if (unlikely(!workload))
 		return NOTIFY_OK;
 
 	switch (action) {
 	case INTEL_CONTEXT_SCHEDULE_IN:
-		intel_gvt_load_render_mmio(workload->vgpu,
-					   workload->ring_id);
+		spin_lock_bh(&scheduler->mmio_context_lock);
+		if (workload->vgpu != scheduler->engine_owner[ring_id]) {
+			/* Switch ring from host to vGPU or vGPU to vGPU. */
+			intel_gvt_switch_mmio(scheduler->engine_owner[ring_id],
+					      workload->vgpu, ring_id);
+			scheduler->engine_owner[ring_id] = workload->vgpu;
+		} else
+			gvt_dbg_sched("skip ring %d mmio switch for vgpu%d\n",
+				      ring_id, workload->vgpu->id);
+		spin_unlock_bh(&scheduler->mmio_context_lock);
 		atomic_set(&workload->shadow_ctx_active, 1);
 		break;
 	case INTEL_CONTEXT_SCHEDULE_OUT:
-		intel_gvt_restore_render_mmio(workload->vgpu,
-					      workload->ring_id);
-		/* If the status is -EINPROGRESS means this workload
-		 * doesn't meet any issue during dispatching so when
-		 * get the SCHEDULE_OUT set the status to be zero for
-		 * good. If the status is NOT -EINPROGRESS means there
-		 * is something wrong happened during dispatching and
-		 * the status should not be set to zero
-		 */
-		if (workload->status == -EINPROGRESS)
-			workload->status = 0;
 		atomic_set(&workload->shadow_ctx_active, 0);
 		break;
 	default:
@@ -181,6 +192,7 @@ static int dispatch_workload(struct intel_vgpu_workload *workload)
 	struct intel_engine_cs *engine = dev_priv->engine[ring_id];
 	struct drm_i915_gem_request *rq;
 	struct intel_vgpu *vgpu = workload->vgpu;
+	struct intel_ring *ring;
 	int ret;
 
 	gvt_dbg_sched("ring id %d prepare to dispatch workload %p\n",
@@ -199,8 +211,9 @@ static int dispatch_workload(struct intel_vgpu_workload *workload)
 	 * shadow_ctx pages invalid. So gvt need to pin itself. After update
 	 * the guest context, gvt can unpin the shadow_ctx safely.
 	 */
-	ret = engine->context_pin(engine, shadow_ctx);
-	if (ret) {
+	ring = engine->context_pin(engine, shadow_ctx);
+	if (IS_ERR(ring)) {
+		ret = PTR_ERR(ring);
 		gvt_vgpu_err("fail to pin shadow context\n");
 		workload->status = ret;
 		mutex_unlock(&dev_priv->drm.struct_mutex);
@@ -330,8 +343,7 @@ static void update_guest_context(struct intel_vgpu_workload *workload)
 	gvt_dbg_sched("ring id %d workload lrca %x\n", ring_id,
 			workload->ctx_desc.lrca);
 
-	context_page_num = intel_lr_context_size(
-			gvt->dev_priv->engine[ring_id]);
+	context_page_num = gvt->dev_priv->engine[ring_id]->context_size;
 
 	context_page_num = context_page_num >> PAGE_SHIFT;
 
@@ -406,6 +418,18 @@ static void complete_current_workload(struct intel_gvt *gvt, int ring_id)
 		wait_event(workload->shadow_ctx_status_wq,
 			   !atomic_read(&workload->shadow_ctx_active));
 
+		/* If this request caused GPU hang, req->fence.error will
+		 * be set to -EIO. Use -EIO to set workload status so
+		 * that when this request caused GPU hang, didn't trigger
+		 * context switch interrupt to guest.
+		 */
+		if (likely(workload->status == -EINPROGRESS)) {
+			if (workload->req->fence.error == -EIO)
+				workload->status = -EIO;
+			else
+				workload->status = 0;
+		}
+
 		i915_gem_request_put(fetch_and_zero(&workload->req));
 
 		if (!workload->status && !vgpu->resetting) {
@@ -431,6 +455,10 @@ static void complete_current_workload(struct intel_gvt *gvt, int ring_id)
 
 	atomic_dec(&vgpu->running_workload_num);
 	wake_up(&scheduler->workload_complete_wq);
+
+	if (gvt->scheduler.need_reschedule)
+		intel_gvt_request_service(gvt, INTEL_GVT_REQUEST_EVENT_SCHED);
+
 	mutex_unlock(&gvt->lock);
 }
 
@@ -439,8 +467,6 @@ struct workload_thread_param {
 	int ring_id;
 };
 
-static DEFINE_MUTEX(scheduler_mutex);
-
 static int workload_thread(void *priv)
 {
 	struct workload_thread_param *p = (struct workload_thread_param *)priv;
@@ -472,8 +498,6 @@ static int workload_thread(void *priv)
 		if (!workload)
 			break;
 
-		mutex_lock(&scheduler_mutex);
-
 		gvt_dbg_sched("ring id %d next workload %p vgpu %d\n",
 				workload->ring_id, workload,
 				workload->vgpu->id);
@@ -512,9 +536,6 @@ complete:
 					FORCEWAKE_ALL);
 
 		intel_runtime_pm_put(gvt->dev_priv);
-
-		mutex_unlock(&scheduler_mutex);
-
 	}
 	return 0;
 }