2 files changed, 134 insertions, 195 deletions

diff --git a/drivers/gpu/drm/scheduler/sched_entity.c b/drivers/gpu/drm/scheduler/sched_entity.c
index c97bc1149663..fe09e5be79bd 100644
--- a/drivers/gpu/drm/scheduler/sched_entity.c
+++ b/drivers/gpu/drm/scheduler/sched_entity.c
@@ -140,6 +140,73 @@ bool drm_sched_entity_is_ready(struct drm_sched_entity *entity)
 	return true;
 }
 
+static void drm_sched_entity_kill_jobs_work(struct work_struct *wrk)
+{
+	struct drm_sched_job *job = container_of(wrk, typeof(*job), work);
+
+	drm_sched_fence_finished(job->s_fence);
+	WARN_ON(job->s_fence->parent);
+	job->sched->ops->free_job(job);
+}
+
+/* Signal the scheduler finished fence when the entity in question is killed. */
+static void drm_sched_entity_kill_jobs_cb(struct dma_fence *f,
+					  struct dma_fence_cb *cb)
+{
+	struct drm_sched_job *job = container_of(cb, struct drm_sched_job,
+						 finish_cb);
+	int r;
+
+	dma_fence_put(f);
+
+	/* Wait for all dependencies to avoid data corruptions */
+	while (!xa_empty(&job->dependencies)) {
+		f = xa_erase(&job->dependencies, job->last_dependency++);
+		r = dma_fence_add_callback(f, &job->finish_cb,
+					   drm_sched_entity_kill_jobs_cb);
+		if (!r)
+			return;
+
+		dma_fence_put(f);
+	}
+
+	INIT_WORK(&job->work, drm_sched_entity_kill_jobs_work);
+	schedule_work(&job->work);
+}
+
+/* Remove the entity from the scheduler and kill all pending jobs */
+static void drm_sched_entity_kill(struct drm_sched_entity *entity)
+{
+	struct drm_sched_job *job;
+	struct dma_fence *prev;
+
+	if (!entity->rq)
+		return;
+
+	spin_lock(&entity->rq_lock);
+	entity->stopped = true;
+	drm_sched_rq_remove_entity(entity->rq, entity);
+	spin_unlock(&entity->rq_lock);
+
+	/* Make sure this entity is not used by the scheduler at the moment */
+	wait_for_completion(&entity->entity_idle);
+
+	prev = dma_fence_get(entity->last_scheduled);
+	while ((job = to_drm_sched_job(spsc_queue_pop(&entity->job_queue)))) {
+		struct drm_sched_fence *s_fence = job->s_fence;
+
+		dma_fence_set_error(&s_fence->finished, -ESRCH);
+
+		dma_fence_get(&s_fence->finished);
+		if (!prev || dma_fence_add_callback(prev, &job->finish_cb,
+					   drm_sched_entity_kill_jobs_cb))
+			drm_sched_entity_kill_jobs_cb(NULL, &job->finish_cb);
+
+		prev = &s_fence->finished;
+	}
+	dma_fence_put(prev);
+}
+
 /**
  * drm_sched_entity_flush - Flush a context entity
  *
@@ -180,91 +247,13 @@ long drm_sched_entity_flush(struct drm_sched_entity *entity, long timeout)
 	/* For killed process disable any more IBs enqueue right now */
 	last_user = cmpxchg(&entity->last_user, current->group_leader, NULL);
 	if ((!last_user || last_user == current->group_leader) &&
-	    (current->flags & PF_EXITING) && (current->exit_code == SIGKILL)) {
-		spin_lock(&entity->rq_lock);
-		entity->stopped = true;
-		drm_sched_rq_remove_entity(entity->rq, entity);
-		spin_unlock(&entity->rq_lock);
-	}
+	    (current->flags & PF_EXITING) && (current->exit_code == SIGKILL))
+		drm_sched_entity_kill(entity);
 
 	return ret;
 }
 EXPORT_SYMBOL(drm_sched_entity_flush);
 
-static void drm_sched_entity_kill_jobs_work(struct work_struct *wrk)
-{
-	struct drm_sched_job *job = container_of(wrk, typeof(*job), work);
-
-	drm_sched_fence_finished(job->s_fence);
-	WARN_ON(job->s_fence->parent);
-	job->sched->ops->free_job(job);
-}
-
-
-/* Signal the scheduler finished fence when the entity in question is killed. */
-static void drm_sched_entity_kill_jobs_cb(struct dma_fence *f,
-					  struct dma_fence_cb *cb)
-{
-	struct drm_sched_job *job = container_of(cb, struct drm_sched_job,
-						 finish_cb);
-
-	dma_fence_put(f);
-	INIT_WORK(&job->work, drm_sched_entity_kill_jobs_work);
-	schedule_work(&job->work);
-}
-
-static struct dma_fence *
-drm_sched_job_dependency(struct drm_sched_job *job,
-			 struct drm_sched_entity *entity)
-{
-	if (!xa_empty(&job->dependencies))
-		return xa_erase(&job->dependencies, job->last_dependency++);
-
-	if (job->sched->ops->dependency)
-		return job->sched->ops->dependency(job, entity);
-
-	return NULL;
-}
-
-static void drm_sched_entity_kill_jobs(struct drm_sched_entity *entity)
-{
-	struct drm_sched_job *job;
-	struct dma_fence *f;
-	int r;
-
-	while ((job = to_drm_sched_job(spsc_queue_pop(&entity->job_queue)))) {
-		struct drm_sched_fence *s_fence = job->s_fence;
-
-		/* Wait for all dependencies to avoid data corruptions */
-		while ((f = drm_sched_job_dependency(job, entity))) {
-			dma_fence_wait(f, false);
-			dma_fence_put(f);
-		}
-
-		drm_sched_fence_scheduled(s_fence);
-		dma_fence_set_error(&s_fence->finished, -ESRCH);
-
-		/*
-		 * When pipe is hanged by older entity, new entity might
-		 * not even have chance to submit it's first job to HW
-		 * and so entity->last_scheduled will remain NULL
-		 */
-		if (!entity->last_scheduled) {
-			drm_sched_entity_kill_jobs_cb(NULL, &job->finish_cb);
-			continue;
-		}
-
-		dma_fence_get(entity->last_scheduled);
-		r = dma_fence_add_callback(entity->last_scheduled,
-					   &job->finish_cb,
-					   drm_sched_entity_kill_jobs_cb);
-		if (r == -ENOENT)
-			drm_sched_entity_kill_jobs_cb(NULL, &job->finish_cb);
-		else if (r)
-			DRM_ERROR("fence add callback failed (%d)\n", r);
-	}
-}
-
 /**
  * drm_sched_entity_fini - Destroy a context entity
  *
@@ -278,33 +267,17 @@ static void drm_sched_entity_kill_jobs(struct drm_sched_entity *entity)
  */
 void drm_sched_entity_fini(struct drm_sched_entity *entity)
 {
-	struct drm_gpu_scheduler *sched = NULL;
-
-	if (entity->rq) {
-		sched = entity->rq->sched;
-		drm_sched_rq_remove_entity(entity->rq, entity);
-	}
-
-	/* Consumption of existing IBs wasn't completed. Forcefully
-	 * remove them here.
+	/*
+	 * If consumption of existing IBs wasn't completed. Forcefully remove
+	 * them here. Also makes sure that the scheduler won't touch this entity
+	 * any more.
 	 */
-	if (spsc_queue_count(&entity->job_queue)) {
-		if (sched) {
-			/*
-			 * Wait for thread to idle to make sure it isn't processing
-			 * this entity.
-			 */
-			wait_for_completion(&entity->entity_idle);
+	drm_sched_entity_kill(entity);
 
-		}
-		if (entity->dependency) {
-			dma_fence_remove_callback(entity->dependency,
-						  &entity->cb);
-			dma_fence_put(entity->dependency);
-			entity->dependency = NULL;
-		}
-
-		drm_sched_entity_kill_jobs(entity);
+	if (entity->dependency) {
+		dma_fence_remove_callback(entity->dependency, &entity->cb);
+		dma_fence_put(entity->dependency);
+		entity->dependency = NULL;
 	}
 
 	dma_fence_put(entity->last_scheduled);
@@ -417,6 +390,19 @@ static bool drm_sched_entity_add_dependency_cb(struct drm_sched_entity *entity)
 	return false;
 }
 
+static struct dma_fence *
+drm_sched_job_dependency(struct drm_sched_job *job,
+			 struct drm_sched_entity *entity)
+{
+	if (!xa_empty(&job->dependencies))
+		return xa_erase(&job->dependencies, job->last_dependency++);
+
+	if (job->sched->ops->prepare_job)
+		return job->sched->ops->prepare_job(job, entity);
+
+	return NULL;
+}
+
 struct drm_sched_job *drm_sched_entity_pop_job(struct drm_sched_entity *entity)
 {
 	struct drm_sched_job *sched_job;
diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c
index d0ff9e11cb69..31f3a1267be4 100644
--- a/drivers/gpu/drm/scheduler/sched_main.c
+++ b/drivers/gpu/drm/scheduler/sched_main.c
@@ -286,32 +286,6 @@ static void drm_sched_job_done_cb(struct dma_fence *f, struct dma_fence_cb *cb)
 }
 
 /**
- * drm_sched_dependency_optimized - test if the dependency can be optimized
- *
- * @fence: the dependency fence
- * @entity: the entity which depends on the above fence
- *
- * Returns true if the dependency can be optimized and false otherwise
- */
-bool drm_sched_dependency_optimized(struct dma_fence* fence,
-				    struct drm_sched_entity *entity)
-{
-	struct drm_gpu_scheduler *sched = entity->rq->sched;
-	struct drm_sched_fence *s_fence;
-
-	if (!fence || dma_fence_is_signaled(fence))
-		return false;
-	if (fence->context == entity->fence_context)
-		return true;
-	s_fence = to_drm_sched_fence(fence);
-	if (s_fence && s_fence->sched == sched)
-		return true;
-
-	return false;
-}
-EXPORT_SYMBOL(drm_sched_dependency_optimized);
-
-/**
  * drm_sched_start_timeout - start timeout for reset worker
  *
  * @sched: scheduler instance to start the worker for
@@ -443,27 +417,6 @@ static void drm_sched_job_timedout(struct work_struct *work)
 	}
 }
 
- /**
-  * drm_sched_increase_karma - Update sched_entity guilty flag
-  *
-  * @bad: The job guilty of time out
-  *
-  * Increment on every hang caused by the 'bad' job. If this exceeds the hang
-  * limit of the scheduler then the respective sched entity is marked guilty and
-  * jobs from it will not be scheduled further
-  */
-void drm_sched_increase_karma(struct drm_sched_job *bad)
-{
-	drm_sched_increase_karma_ext(bad, 1);
-}
-EXPORT_SYMBOL(drm_sched_increase_karma);
-
-void drm_sched_reset_karma(struct drm_sched_job *bad)
-{
-	drm_sched_increase_karma_ext(bad, 0);
-}
-EXPORT_SYMBOL(drm_sched_reset_karma);
-
 /**
  * drm_sched_stop - stop the scheduler
  *
@@ -605,31 +558,14 @@ EXPORT_SYMBOL(drm_sched_start);
  */
 void drm_sched_resubmit_jobs(struct drm_gpu_scheduler *sched)
 {
-	drm_sched_resubmit_jobs_ext(sched, INT_MAX);
-}
-EXPORT_SYMBOL(drm_sched_resubmit_jobs);
-
-/**
- * drm_sched_resubmit_jobs_ext - helper to relunch certain number of jobs from mirror ring list
- *
- * @sched: scheduler instance
- * @max: job numbers to relaunch
- *
- */
-void drm_sched_resubmit_jobs_ext(struct drm_gpu_scheduler *sched, int max)
-{
 	struct drm_sched_job *s_job, *tmp;
 	uint64_t guilty_context;
 	bool found_guilty = false;
 	struct dma_fence *fence;
-	int i = 0;
 
 	list_for_each_entry_safe(s_job, tmp, &sched->pending_list, list) {
 		struct drm_sched_fence *s_fence = s_job->s_fence;
 
-		if (i >= max)
-			break;
-
 		if (!found_guilty && atomic_read(&s_job->karma) > sched->hang_limit) {
 			found_guilty = true;
 			guilty_context = s_job->s_fence->scheduled.context;
@@ -639,7 +575,6 @@ void drm_sched_resubmit_jobs_ext(struct drm_gpu_scheduler *sched, int max)
 			dma_fence_set_error(&s_fence->finished, -ECANCELED);
 
 		fence = sched->ops->run_job(s_job);
-		i++;
 
 		if (IS_ERR_OR_NULL(fence)) {
 			if (IS_ERR(fence))
@@ -655,7 +590,7 @@ void drm_sched_resubmit_jobs_ext(struct drm_gpu_scheduler *sched, int max)
 		}
 	}
 }
-EXPORT_SYMBOL(drm_sched_resubmit_jobs_ext);
+EXPORT_SYMBOL(drm_sched_resubmit_jobs);
 
 /**
  * drm_sched_job_init - init a scheduler job
@@ -773,32 +708,28 @@ int drm_sched_job_add_dependency(struct drm_sched_job *job,
 EXPORT_SYMBOL(drm_sched_job_add_dependency);
 
 /**
- * drm_sched_job_add_implicit_dependencies - adds implicit dependencies as job
- *   dependencies
+ * drm_sched_job_add_resv_dependencies - add all fences from the resv to the job
  * @job: scheduler job to add the dependencies to
- * @obj: the gem object to add new dependencies from.
- * @write: whether the job might write the object (so we need to depend on
- * shared fences in the reservation object).
+ * @resv: the dma_resv object to get the fences from
+ * @usage: the dma_resv_usage to use to filter the fences
  *
- * This should be called after drm_gem_lock_reservations() on your array of
- * GEM objects used in the job but before updating the reservations with your
- * own fences.
+ * This adds all fences matching the given usage from @resv to @job.
+ * Must be called with the @resv lock held.
  *
  * Returns:
  * 0 on success, or an error on failing to expand the array.
  */
-int drm_sched_job_add_implicit_dependencies(struct drm_sched_job *job,
-					    struct drm_gem_object *obj,
-					    bool write)
+int drm_sched_job_add_resv_dependencies(struct drm_sched_job *job,
+					struct dma_resv *resv,
+					enum dma_resv_usage usage)
 {
 	struct dma_resv_iter cursor;
 	struct dma_fence *fence;
 	int ret;
 
-	dma_resv_assert_held(obj->resv);
+	dma_resv_assert_held(resv);
 
-	dma_resv_for_each_fence(&cursor, obj->resv, dma_resv_usage_rw(write),
-				fence) {
+	dma_resv_for_each_fence(&cursor, resv, usage, fence) {
 		/* Make sure to grab an additional ref on the added fence */
 		dma_fence_get(fence);
 		ret = drm_sched_job_add_dependency(job, fence);
@@ -809,8 +740,31 @@ int drm_sched_job_add_implicit_dependencies(struct drm_sched_job *job,
 	}
 	return 0;
 }
-EXPORT_SYMBOL(drm_sched_job_add_implicit_dependencies);
+EXPORT_SYMBOL(drm_sched_job_add_resv_dependencies);
 
+/**
+ * drm_sched_job_add_implicit_dependencies - adds implicit dependencies as job
+ *   dependencies
+ * @job: scheduler job to add the dependencies to
+ * @obj: the gem object to add new dependencies from.
+ * @write: whether the job might write the object (so we need to depend on
+ * shared fences in the reservation object).
+ *
+ * This should be called after drm_gem_lock_reservations() on your array of
+ * GEM objects used in the job but before updating the reservations with your
+ * own fences.
+ *
+ * Returns:
+ * 0 on success, or an error on failing to expand the array.
+ */
+int drm_sched_job_add_implicit_dependencies(struct drm_sched_job *job,
+					    struct drm_gem_object *obj,
+					    bool write)
+{
+	return drm_sched_job_add_resv_dependencies(job, obj->resv,
+						   dma_resv_usage_rw(write));
+}
+EXPORT_SYMBOL(drm_sched_job_add_implicit_dependencies);
 
 /**
  * drm_sched_job_cleanup - clean up scheduler job resources
@@ -1172,13 +1126,15 @@ void drm_sched_fini(struct drm_gpu_scheduler *sched)
 EXPORT_SYMBOL(drm_sched_fini);
 
 /**
- * drm_sched_increase_karma_ext - Update sched_entity guilty flag
+ * drm_sched_increase_karma - Update sched_entity guilty flag
  *
  * @bad: The job guilty of time out
- * @type: type for increase/reset karma
  *
+ * Increment on every hang caused by the 'bad' job. If this exceeds the hang
+ * limit of the scheduler then the respective sched entity is marked guilty and
+ * jobs from it will not be scheduled further
  */
-void drm_sched_increase_karma_ext(struct drm_sched_job *bad, int type)
+void drm_sched_increase_karma(struct drm_sched_job *bad)
 {
 	int i;
 	struct drm_sched_entity *tmp;
@@ -1190,10 +1146,7 @@ void drm_sched_increase_karma_ext(struct drm_sched_job *bad, int type)
 	 * corrupt but keep in mind that kernel jobs always considered good.
 	 */
 	if (bad->s_priority != DRM_SCHED_PRIORITY_KERNEL) {
-		if (type == 0)
-			atomic_set(&bad->karma, 0);
-		else if (type == 1)
-			atomic_inc(&bad->karma);
+		atomic_inc(&bad->karma);
 
 		for (i = DRM_SCHED_PRIORITY_MIN; i < DRM_SCHED_PRIORITY_KERNEL;
 		     i++) {
@@ -1204,7 +1157,7 @@ void drm_sched_increase_karma_ext(struct drm_sched_job *bad, int type)
 				if (bad->s_fence->scheduled.context ==
 				    entity->fence_context) {
 					if (entity->guilty)
-						atomic_set(entity->guilty, type);
+						atomic_set(entity->guilty, 1);
 					break;
 				}
 			}
@@ -1214,4 +1167,4 @@ void drm_sched_increase_karma_ext(struct drm_sched_job *bad, int type)
 		}
 	}
 }
-EXPORT_SYMBOL(drm_sched_increase_karma_ext);
+EXPORT_SYMBOL(drm_sched_increase_karma);