1 files changed, 154 insertions, 84 deletions

diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c
index bfea608a7106..e2cda28a1af4 100644
--- a/drivers/gpu/drm/scheduler/sched_main.c
+++ b/drivers/gpu/drm/scheduler/sched_main.c
@@ -66,6 +66,7 @@
  * This implies waiting for previously executed jobs.
  */
 
+#include <linux/export.h>
 #include <linux/wait.h>
 #include <linux/sched.h>
 #include <linux/completion.h>
@@ -83,12 +84,6 @@
 #define CREATE_TRACE_POINTS
 #include "gpu_scheduler_trace.h"
 
-#ifdef CONFIG_LOCKDEP
-static struct lockdep_map drm_sched_lockdep_map = {
-	.name = "drm_sched_lockdep_map"
-};
-#endif
-
 int drm_sched_policy = DRM_SCHED_POLICY_FIFO;
 
 /**
@@ -268,38 +263,14 @@ drm_sched_rq_select_entity_rr(struct drm_gpu_scheduler *sched,
 	entity = rq->current_entity;
 	if (entity) {
 		list_for_each_entry_continue(entity, &rq->entities, list) {
-			if (drm_sched_entity_is_ready(entity)) {
-				/* If we can't queue yet, preserve the current
-				 * entity in terms of fairness.
-				 */
-				if (!drm_sched_can_queue(sched, entity)) {
-					spin_unlock(&rq->lock);
-					return ERR_PTR(-ENOSPC);
-				}
-
-				rq->current_entity = entity;
-				reinit_completion(&entity->entity_idle);
-				spin_unlock(&rq->lock);
-				return entity;
-			}
+			if (drm_sched_entity_is_ready(entity))
+				goto found;
 		}
 	}
 
 	list_for_each_entry(entity, &rq->entities, list) {
-		if (drm_sched_entity_is_ready(entity)) {
-			/* If we can't queue yet, preserve the current entity in
-			 * terms of fairness.
-			 */
-			if (!drm_sched_can_queue(sched, entity)) {
-				spin_unlock(&rq->lock);
-				return ERR_PTR(-ENOSPC);
-			}
-
-			rq->current_entity = entity;
-			reinit_completion(&entity->entity_idle);
-			spin_unlock(&rq->lock);
-			return entity;
-		}
+		if (drm_sched_entity_is_ready(entity))
+			goto found;
 
 		if (entity == rq->current_entity)
 			break;
@@ -308,6 +279,22 @@ drm_sched_rq_select_entity_rr(struct drm_gpu_scheduler *sched,
 	spin_unlock(&rq->lock);
 
 	return NULL;
+
+found:
+	if (!drm_sched_can_queue(sched, entity)) {
+		/*
+		 * If scheduler cannot take more jobs signal the caller to not
+		 * consider lower priority queues.
+		 */
+		entity = ERR_PTR(-ENOSPC);
+	} else {
+		rq->current_entity = entity;
+		reinit_completion(&entity->entity_idle);
+	}
+
+	spin_unlock(&rq->lock);
+
+	return entity;
 }
 
 /**
@@ -379,11 +366,16 @@ static void drm_sched_run_free_queue(struct drm_gpu_scheduler *sched)
 {
 	struct drm_sched_job *job;
 
-	spin_lock(&sched->job_list_lock);
 	job = list_first_entry_or_null(&sched->pending_list,
 				       struct drm_sched_job, list);
 	if (job && dma_fence_is_signaled(&job->s_fence->finished))
 		__drm_sched_run_free_queue(sched);
+}
+
+static void drm_sched_run_free_queue_unlocked(struct drm_gpu_scheduler *sched)
+{
+	spin_lock(&sched->job_list_lock);
+	drm_sched_run_free_queue(sched);
 	spin_unlock(&sched->job_list_lock);
 }
 
@@ -391,7 +383,7 @@ static void drm_sched_run_free_queue(struct drm_gpu_scheduler *sched)
  * drm_sched_job_done - complete a job
  * @s_job: pointer to the job which is done
  *
- * Finish the job's fence and wake up the worker thread.
+ * Finish the job's fence and resubmit the work items.
  */
 static void drm_sched_job_done(struct drm_sched_job *s_job, int result)
 {
@@ -401,7 +393,7 @@ static void drm_sched_job_done(struct drm_sched_job *s_job, int result)
 	atomic_sub(s_job->credits, &sched->credit_count);
 	atomic_dec(sched->score);
 
-	trace_drm_sched_process_job(s_fence);
+	trace_drm_sched_job_done(s_fence);
 
 	dma_fence_get(&s_fence->finished);
 	drm_sched_fence_finished(s_fence, result);
@@ -536,11 +528,37 @@ static void drm_sched_job_begin(struct drm_sched_job *s_job)
 	spin_unlock(&sched->job_list_lock);
 }
 
+/**
+ * drm_sched_job_reinsert_on_false_timeout - reinsert the job on a false timeout
+ * @sched: scheduler instance
+ * @job: job to be reinserted on the pending list
+ *
+ * In the case of a "false timeout" - when a timeout occurs but the GPU isn't
+ * hung and is making progress, the scheduler must reinsert the job back into
+ * @sched->pending_list. Otherwise, the job and its resources won't be freed
+ * through the &struct drm_sched_backend_ops.free_job callback.
+ *
+ * This function must be used in "false timeout" cases only.
+ */
+static void drm_sched_job_reinsert_on_false_timeout(struct drm_gpu_scheduler *sched,
+						    struct drm_sched_job *job)
+{
+	spin_lock(&sched->job_list_lock);
+	list_add(&job->list, &sched->pending_list);
+
+	/* After reinserting the job, the scheduler enqueues the free-job work
+	 * again if ready. Otherwise, a signaled job could be added to the
+	 * pending list, but never freed.
+	 */
+	drm_sched_run_free_queue(sched);
+	spin_unlock(&sched->job_list_lock);
+}
+
 static void drm_sched_job_timedout(struct work_struct *work)
 {
 	struct drm_gpu_scheduler *sched;
 	struct drm_sched_job *job;
-	enum drm_gpu_sched_stat status = DRM_GPU_SCHED_STAT_NOMINAL;
+	enum drm_gpu_sched_stat status = DRM_GPU_SCHED_STAT_RESET;
 
 	sched = container_of(work, struct drm_gpu_scheduler, work_tdr.work);
 
@@ -551,9 +569,10 @@ static void drm_sched_job_timedout(struct work_struct *work)
 
 	if (job) {
 		/*
-		 * Remove the bad job so it cannot be freed by concurrent
-		 * drm_sched_cleanup_jobs. It will be reinserted back after sched->thread
-		 * is parked at which point it's safe.
+		 * Remove the bad job so it cannot be freed by a concurrent
+		 * &struct drm_sched_backend_ops.free_job. It will be
+		 * reinserted after the scheduler's work items have been
+		 * cancelled, at which point it's safe.
 		 */
 		list_del_init(&job->list);
 		spin_unlock(&sched->job_list_lock);
@@ -568,6 +587,9 @@ static void drm_sched_job_timedout(struct work_struct *work)
 			job->sched->ops->free_job(job);
 			sched->free_guilty = false;
 		}
+
+		if (status == DRM_GPU_SCHED_STAT_NO_HANG)
+			drm_sched_job_reinsert_on_false_timeout(sched, job);
 	} else {
 		spin_unlock(&sched->job_list_lock);
 	}
@@ -590,6 +612,10 @@ static void drm_sched_job_timedout(struct work_struct *work)
  * This function is typically used for reset recovery (see the docu of
  * drm_sched_backend_ops.timedout_job() for details). Do not call it for
  * scheduler teardown, i.e., before calling drm_sched_fini().
+ *
+ * As it's only used for reset recovery, drivers must not call this function
+ * in their &struct drm_sched_backend_ops.timedout_job callback when they
+ * skip a reset using &enum drm_gpu_sched_stat.DRM_GPU_SCHED_STAT_NO_HANG.
  */
 void drm_sched_stop(struct drm_gpu_scheduler *sched, struct drm_sched_job *bad)
 {
@@ -599,10 +625,10 @@ void drm_sched_stop(struct drm_gpu_scheduler *sched, struct drm_sched_job *bad)
 
 	/*
 	 * Reinsert back the bad job here - now it's safe as
-	 * drm_sched_get_finished_job cannot race against us and release the
+	 * drm_sched_get_finished_job() cannot race against us and release the
 	 * bad job at this point - we parked (waited for) any in progress
-	 * (earlier) cleanups and drm_sched_get_finished_job will not be called
-	 * now until the scheduler thread is unparked.
+	 * (earlier) cleanups and drm_sched_get_finished_job() will not be
+	 * called now until the scheduler's work items are submitted again.
 	 */
 	if (bad && bad->sched == sched)
 		/*
@@ -615,7 +641,8 @@ void drm_sched_stop(struct drm_gpu_scheduler *sched, struct drm_sched_job *bad)
 	 * Iterate the job list from later to  earlier one and either deactive
 	 * their HW callbacks or remove them from pending list if they already
 	 * signaled.
-	 * This iteration is thread safe as sched thread is stopped.
+	 * This iteration is thread safe as the scheduler's work items have been
+	 * cancelled.
 	 */
 	list_for_each_entry_safe_reverse(s_job, tmp, &sched->pending_list,
 					 list) {
@@ -674,15 +701,19 @@ EXPORT_SYMBOL(drm_sched_stop);
  * drm_sched_backend_ops.timedout_job() for details). Do not call it for
  * scheduler startup. The scheduler itself is fully operational after
  * drm_sched_init() succeeded.
+ *
+ * As it's only used for reset recovery, drivers must not call this function
+ * in their &struct drm_sched_backend_ops.timedout_job callback when they
+ * skip a reset using &enum drm_gpu_sched_stat.DRM_GPU_SCHED_STAT_NO_HANG.
  */
 void drm_sched_start(struct drm_gpu_scheduler *sched, int errno)
 {
 	struct drm_sched_job *s_job, *tmp;
 
 	/*
-	 * Locking the list is not required here as the sched thread is parked
-	 * so no new jobs are being inserted or removed. Also concurrent
-	 * GPU recovers can't run in parallel.
+	 * Locking the list is not required here as the scheduler's work items
+	 * are currently not running, so no new jobs are being inserted or
+	 * removed. Also concurrent GPU recovers can't run in parallel.
 	 */
 	list_for_each_entry_safe(s_job, tmp, &sched->pending_list, list) {
 		struct dma_fence *fence = s_job->s_fence->parent;
@@ -764,6 +795,8 @@ EXPORT_SYMBOL(drm_sched_resubmit_jobs);
  * @credits: the number of credits this job contributes to the schedulers
  * credit limit
  * @owner: job owner for debugging
+ * @drm_client_id: &struct drm_file.client_id of the owner (used by trace
+ * events)
  *
  * Refer to drm_sched_entity_push_job() documentation
  * for locking considerations.
@@ -784,7 +817,8 @@ EXPORT_SYMBOL(drm_sched_resubmit_jobs);
  */
 int drm_sched_job_init(struct drm_sched_job *job,
 		       struct drm_sched_entity *entity,
-		       u32 credits, void *owner)
+		       u32 credits, void *owner,
+		       uint64_t drm_client_id)
 {
 	if (!entity->rq) {
 		/* This will most likely be followed by missing frames
@@ -810,7 +844,7 @@ int drm_sched_job_init(struct drm_sched_job *job,
 
 	job->entity = entity;
 	job->credits = credits;
-	job->s_fence = drm_sched_fence_alloc(entity, owner);
+	job->s_fence = drm_sched_fence_alloc(entity, owner, drm_client_id);
 	if (!job->s_fence)
 		return -ENOMEM;
 
@@ -828,11 +862,15 @@ EXPORT_SYMBOL(drm_sched_job_init);
  *
  * This arms a scheduler job for execution. Specifically it initializes the
  * &drm_sched_job.s_fence of @job, so that it can be attached to struct dma_resv
- * or other places that need to track the completion of this job.
+ * or other places that need to track the completion of this job. It also
+ * initializes sequence numbers, which are fundamental for fence ordering.
  *
  * Refer to drm_sched_entity_push_job() documentation for locking
  * considerations.
  *
+ * Once this function was called, you *must* submit @job with
+ * drm_sched_entity_push_job().
+ *
  * This can only be called if drm_sched_job_init() succeeded.
  */
 void drm_sched_job_arm(struct drm_sched_job *job)
@@ -846,7 +884,6 @@ void drm_sched_job_arm(struct drm_sched_job *job)
 
 	job->sched = sched;
 	job->s_priority = entity->priority;
-	job->id = atomic64_inc_return(&sched->job_id_count);
 
 	drm_sched_fence_init(job->s_fence, job->entity);
 }
@@ -1015,13 +1052,14 @@ EXPORT_SYMBOL(drm_sched_job_has_dependency);
  * Cleans up the resources allocated with drm_sched_job_init().
  *
  * Drivers should call this from their error unwind code if @job is aborted
- * before it was submitted to an entity with drm_sched_entity_push_job().
+ * before drm_sched_job_arm() is called.
  *
- * Since calling drm_sched_job_arm() causes the job's fences to be initialized,
- * it is up to the driver to ensure that fences that were exposed to external
- * parties get signaled. drm_sched_job_cleanup() does not ensure this.
+ * drm_sched_job_arm() is a point of no return since it initializes the fences
+ * and their sequence number etc. Once that function has been called, you *must*
+ * submit it with drm_sched_entity_push_job() and cannot simply abort it by
+ * calling drm_sched_job_cleanup().
  *
- * This function must also be called in &struct drm_sched_backend_ops.free_job
+ * This function should be called in the &drm_sched_backend_ops.free_job callback.
  */
 void drm_sched_job_cleanup(struct drm_sched_job *job)
 {
@@ -1029,10 +1067,15 @@ void drm_sched_job_cleanup(struct drm_sched_job *job)
 	unsigned long index;
 
 	if (kref_read(&job->s_fence->finished.refcount)) {
-		/* drm_sched_job_arm() has been called */
+		/* The job has been processed by the scheduler, i.e.,
+		 * drm_sched_job_arm() and drm_sched_entity_push_job() have
+		 * been called.
+		 */
 		dma_fence_put(&job->s_fence->finished);
 	} else {
-		/* aborted job before arming */
+		/* The job was aborted before it has been committed to be run;
+		 * notably, drm_sched_job_arm() has not been called.
+		 */
 		drm_sched_fence_free(job->s_fence);
 	}
 
@@ -1183,7 +1226,7 @@ static void drm_sched_free_job_work(struct work_struct *w)
 	if (job)
 		sched->ops->free_job(job);
 
-	drm_sched_run_free_queue(sched);
+	drm_sched_run_free_queue_unlocked(sched);
 	drm_sched_run_job_queue(sched);
 }
 
@@ -1219,21 +1262,24 @@ static void drm_sched_run_job_work(struct work_struct *w)
 	atomic_add(sched_job->credits, &sched->credit_count);
 	drm_sched_job_begin(sched_job);
 
-	trace_drm_run_job(sched_job, entity);
+	trace_drm_sched_job_run(sched_job, entity);
+	/*
+	 * The run_job() callback must by definition return a fence whose
+	 * refcount has been incremented for the scheduler already.
+	 */
 	fence = sched->ops->run_job(sched_job);
 	complete_all(&entity->entity_idle);
 	drm_sched_fence_scheduled(s_fence, fence);
 
 	if (!IS_ERR_OR_NULL(fence)) {
-		/* Drop for original kref_init of the fence */
-		dma_fence_put(fence);
-
 		r = dma_fence_add_callback(fence, &sched_job->cb,
 					   drm_sched_job_done_cb);
 		if (r == -ENOENT)
 			drm_sched_job_done(sched_job, fence->error);
 		else if (r)
 			DRM_DEV_ERROR(sched->dev, "fence add callback failed (%d)\n", r);
+
+		dma_fence_put(fence);
 	} else {
 		drm_sched_job_done(sched_job, IS_ERR(fence) ?
 				   PTR_ERR(fence) : 0);
@@ -1243,6 +1289,25 @@ static void drm_sched_run_job_work(struct work_struct *w)
 	drm_sched_run_job_queue(sched);
 }
 
+static struct workqueue_struct *drm_sched_alloc_wq(const char *name)
+{
+#if (IS_ENABLED(CONFIG_LOCKDEP))
+	static struct lockdep_map map = {
+		.name = "drm_sched_lockdep_map"
+	};
+
+	/*
+	 * Avoid leaking a lockdep map on each drm sched creation and
+	 * destruction by using a single lockdep map for all drm sched
+	 * allocated submit_wq.
+	 */
+
+	return alloc_ordered_workqueue_lockdep_map(name, WQ_MEM_RECLAIM, &map);
+#else
+	return alloc_ordered_workqueue(name, WQ_MEM_RECLAIM);
+#endif
+}
+
 /**
  * drm_sched_init - Init a gpu scheduler instance
  *
@@ -1283,13 +1348,7 @@ int drm_sched_init(struct drm_gpu_scheduler *sched, const struct drm_sched_init_
 		sched->submit_wq = args->submit_wq;
 		sched->own_submit_wq = false;
 	} else {
-#ifdef CONFIG_LOCKDEP
-		sched->submit_wq = alloc_ordered_workqueue_lockdep_map(args->name,
-								       WQ_MEM_RECLAIM,
-								       &drm_sched_lockdep_map);
-#else
-		sched->submit_wq = alloc_ordered_workqueue(args->name, WQ_MEM_RECLAIM);
-#endif
+		sched->submit_wq = drm_sched_alloc_wq(args->name);
 		if (!sched->submit_wq)
 			return -ENOMEM;
 
@@ -1335,6 +1394,18 @@ Out_check_own:
 }
 EXPORT_SYMBOL(drm_sched_init);
 
+static void drm_sched_cancel_remaining_jobs(struct drm_gpu_scheduler *sched)
+{
+	struct drm_sched_job *job, *tmp;
+
+	/* All other accessors are stopped. No locking necessary. */
+	list_for_each_entry_safe_reverse(job, tmp, &sched->pending_list, list) {
+		sched->ops->cancel_job(job);
+		list_del(&job->list);
+		sched->ops->free_job(job);
+	}
+}
+
 /**
  * drm_sched_fini - Destroy a gpu scheduler
  *
@@ -1342,19 +1413,11 @@ EXPORT_SYMBOL(drm_sched_init);
  *
  * Tears down and cleans up the scheduler.
  *
- * This stops submission of new jobs to the hardware through
- * drm_sched_backend_ops.run_job(). Consequently, drm_sched_backend_ops.free_job()
- * will not be called for all jobs still in drm_gpu_scheduler.pending_list.
- * There is no solution for this currently. Thus, it is up to the driver to make
- * sure that:
- *
- *  a) drm_sched_fini() is only called after for all submitted jobs
- *     drm_sched_backend_ops.free_job() has been called or that
- *  b) the jobs for which drm_sched_backend_ops.free_job() has not been called
- *     after drm_sched_fini() ran are freed manually.
- *
- * FIXME: Take care of the above problem and prevent this function from leaking
- * the jobs in drm_gpu_scheduler.pending_list under any circumstances.
+ * This stops submission of new jobs to the hardware through &struct
+ * drm_sched_backend_ops.run_job. If &struct drm_sched_backend_ops.cancel_job
+ * is implemented, all jobs will be canceled through it and afterwards cleaned
+ * up through &struct drm_sched_backend_ops.free_job. If cancel_job is not
+ * implemented, memory could leak.
  */
 void drm_sched_fini(struct drm_gpu_scheduler *sched)
 {
@@ -1384,11 +1447,18 @@ void drm_sched_fini(struct drm_gpu_scheduler *sched)
 	/* Confirm no work left behind accessing device structures */
 	cancel_delayed_work_sync(&sched->work_tdr);
 
+	/* Avoid memory leaks if supported by the driver. */
+	if (sched->ops->cancel_job)
+		drm_sched_cancel_remaining_jobs(sched);
+
 	if (sched->own_submit_wq)
 		destroy_workqueue(sched->submit_wq);
 	sched->ready = false;
 	kfree(sched->sched_rq);
 	sched->sched_rq = NULL;
+
+	if (!list_empty(&sched->pending_list))
+		dev_warn(sched->dev, "Tearing down scheduler while jobs are pending!\n");
 }
 EXPORT_SYMBOL(drm_sched_fini);