2 files changed, 40 insertions, 10 deletions

diff --git a/drivers/gpu/drm/scheduler/sched_entity.c b/drivers/gpu/drm/scheduler/sched_entity.c
index 1a5153197fe9..461a7a8129f4 100644
--- a/drivers/gpu/drm/scheduler/sched_entity.c
+++ b/drivers/gpu/drm/scheduler/sched_entity.c
@@ -23,6 +23,7 @@
 
 #include <linux/kthread.h>
 #include <linux/slab.h>
+#include <linux/completion.h>
 
 #include <drm/drm_print.h>
 #include <drm/gpu_scheduler.h>
@@ -68,6 +69,8 @@ int drm_sched_entity_init(struct drm_sched_entity *entity,
 	if (!entity->rq_list)
 		return -ENOMEM;
 
+	init_completion(&entity->entity_idle);
+
 	for (i = 0; i < num_rq_list; ++i)
 		entity->rq_list[i] = rq_list[i];
 
@@ -286,11 +289,12 @@ void drm_sched_entity_fini(struct drm_sched_entity *entity)
 	 */
 	if (spsc_queue_count(&entity->job_queue)) {
 		if (sched) {
-			/* Park the kernel for a moment to make sure it isn't processing
-			 * our enity.
+			/*
+			 * Wait for thread to idle to make sure it isn't processing
+			 * this entity.
 			 */
-			kthread_park(sched->thread);
-			kthread_unpark(sched->thread);
+			wait_for_completion(&entity->entity_idle);
+
 		}
 		if (entity->dependency) {
 			dma_fence_remove_callback(entity->dependency,
diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c
index 40f396b00797..3fad5876a13f 100644
--- a/drivers/gpu/drm/scheduler/sched_main.c
+++ b/drivers/gpu/drm/scheduler/sched_main.c
@@ -47,6 +47,7 @@
 #include <linux/kthread.h>
 #include <linux/wait.h>
 #include <linux/sched.h>
+#include <linux/completion.h>
 #include <uapi/linux/sched/types.h>
 
 #include <drm/drm_print.h>
@@ -134,6 +135,7 @@ drm_sched_rq_select_entity(struct drm_sched_rq *rq)
 		list_for_each_entry_continue(entity, &rq->entities, list) {
 			if (drm_sched_entity_is_ready(entity)) {
 				rq->current_entity = entity;
+				reinit_completion(&entity->entity_idle);
 				spin_unlock(&rq->lock);
 				return entity;
 			}
@@ -144,6 +146,7 @@ drm_sched_rq_select_entity(struct drm_sched_rq *rq)
 
 		if (drm_sched_entity_is_ready(entity)) {
 			rq->current_entity = entity;
+			reinit_completion(&entity->entity_idle);
 			spin_unlock(&rq->lock);
 			return entity;
 		}
@@ -506,6 +509,7 @@ void drm_sched_resubmit_jobs(struct drm_gpu_scheduler *sched)
 	struct drm_sched_job *s_job, *tmp;
 	uint64_t guilty_context;
 	bool found_guilty = false;
+	struct dma_fence *fence;
 
 	list_for_each_entry_safe(s_job, tmp, &sched->ring_mirror_list, node) {
 		struct drm_sched_fence *s_fence = s_job->s_fence;
@@ -519,7 +523,18 @@ void drm_sched_resubmit_jobs(struct drm_gpu_scheduler *sched)
 			dma_fence_set_error(&s_fence->finished, -ECANCELED);
 
 		dma_fence_put(s_job->s_fence->parent);
-		s_job->s_fence->parent = sched->ops->run_job(s_job);
+		fence = sched->ops->run_job(s_job);
+
+		if (IS_ERR_OR_NULL(fence)) {
+			if (IS_ERR(fence))
+				dma_fence_set_error(&s_fence->finished, PTR_ERR(fence));
+
+			s_job->s_fence->parent = NULL;
+		} else {
+			s_job->s_fence->parent = fence;
+		}
+
+
 	}
 }
 EXPORT_SYMBOL(drm_sched_resubmit_jobs);
@@ -662,9 +677,13 @@ drm_sched_get_cleanup_job(struct drm_gpu_scheduler *sched)
 	struct drm_sched_job *job;
 	unsigned long flags;
 
-	/* Don't destroy jobs while the timeout worker is running */
-	if (sched->timeout != MAX_SCHEDULE_TIMEOUT &&
-	    !cancel_delayed_work(&sched->work_tdr))
+	/*
+	 * Don't destroy jobs while the timeout worker is running  OR thread
+	 * is being parked and hence assumed to not touch ring_mirror_list
+	 */
+	if ((sched->timeout != MAX_SCHEDULE_TIMEOUT &&
+	    !cancel_delayed_work(&sched->work_tdr)) ||
+	    __kthread_should_park(sched->thread))
 		return NULL;
 
 	spin_lock_irqsave(&sched->job_list_lock, flags);
@@ -741,6 +760,9 @@ static int drm_sched_main(void *param)
 			continue;
 
 		sched_job = drm_sched_entity_pop_job(entity);
+
+		complete(&entity->entity_idle);
+
 		if (!sched_job)
 			continue;
 
@@ -752,7 +774,7 @@ static int drm_sched_main(void *param)
 		fence = sched->ops->run_job(sched_job);
 		drm_sched_fence_scheduled(s_fence);
 
-		if (fence) {
+		if (!IS_ERR_OR_NULL(fence)) {
 			s_fence->parent = dma_fence_get(fence);
 			r = dma_fence_add_callback(fence, &sched_job->cb,
 						   drm_sched_process_job);
@@ -762,8 +784,12 @@ static int drm_sched_main(void *param)
 				DRM_ERROR("fence add callback failed (%d)\n",
 					  r);
 			dma_fence_put(fence);
-		} else
+		} else {
+			if (IS_ERR(fence))
+				dma_fence_set_error(&s_fence->finished, PTR_ERR(fence));
+
 			drm_sched_process_job(NULL, &sched_job->cb);
+		}
 
 		wake_up(&sched->job_scheduled);
 	}