Revert "drm/sched: Convert the GPU scheduler to variable number ..."

Revert submission 1033322

Reason for revert: Breaks GPU reset
Reverted Changes:
I518055d36:drm/amdgpu: drop setting buffer funcs in sdma442
I3161b3ba3:drm/amdgpu: fix buffer funcs setting order on susp...
If96a48811:drm/amdklc: fix a intree build error.
I370135be6:drm/sched: Don't disturb the entity when in RR-mod...
I051ca6e0b:drm/sched: Move free worker re-queuing out of the ...
I25fb2785c:drm/sched: Rename drm_sched_get_cleanup_job to be ...
I666125051:drm/sched: Rename drm_sched_run_job_queue_if_ready...
I9a32472a9:drm/sched: Rename drm_sched_free_job_queue to be m...
I4c2c9bed1:drm/sched: Re-queue run job worker when drm_sched_...
Ie513a85c7:drm/sched: Drain all entities in DRM sched run job...
I855c12a4e:drm/sched: Split free_job into own work item
I7a116284b:drm/sched: Convert drm scheduler to use a work que...
Iffc1aa4cd:drm/amdgpu: move buffer funcs setting up a level
I3753e9dd4:drm/sched: Convert the GPU scheduler to variable n...

Change-Id: I8fa3facd7111e26a08ff77e6345b4437631e9ac0

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c

@@ -2661,7 +2661,6 @@ static int amdgpu_device_init_schedulers(struct amdgpu_device *adev)
 		}
 
 		r = drm_sched_init(&ring->sched, &amdgpu_sched_ops,
-				   DRM_SCHED_PRIORITY_COUNT,
 				   ring->num_hw_submission, 0,
 				   timeout, adev->reset_domain->wq,
 				   ring->sched_score, ring->name,

drivers/gpu/drm/amd/amdgpu/amdgpu_job.c

@@ -326,8 +326,8 @@ void amdgpu_job_stop_all_jobs_on_sched(struct drm_gpu_scheduler *sched)
 	int i;
 
 	/* Signal all jobs not yet scheduled */
-        for (i = DRM_SCHED_PRIORITY_KERNEL; i < sched->num_rqs; i++) {
-		struct drm_sched_rq *rq = sched->sched_rq[i];
+	for (i = DRM_SCHED_PRIORITY_COUNT - 1; i >= DRM_SCHED_PRIORITY_KERNEL; i--) {
+		struct drm_sched_rq *rq = &sched->sched_rq[i];
 		spin_lock(&rq->lock);
 		list_for_each_entry(s_entity, &rq->entities, list) {
 			while ((s_job = to_drm_sched_job(spsc_queue_pop(&s_entity->job_queue)))) {

drivers/gpu/drm/etnaviv/etnaviv_sched.c

@@ -135,7 +135,6 @@ int etnaviv_sched_init(struct etnaviv_gpu *gpu)
 	int ret;
 
 	ret = drm_sched_init(&gpu->sched, &etnaviv_sched_ops,
-			     DRM_SCHED_PRIORITY_COUNT,
 			     etnaviv_hw_jobs_limit, etnaviv_job_hang_limit,
 			     msecs_to_jiffies(500), NULL, NULL,
 			     dev_name(gpu->dev), gpu->dev);

drivers/gpu/drm/lima/lima_sched.c

@@ -489,7 +489,6 @@ int lima_sched_pipe_init(struct lima_sched_pipe *pipe, const char *name)
 	INIT_WORK(&pipe->recover_work, lima_sched_recover_work);
 
 	return drm_sched_init(&pipe->base, &lima_sched_ops,
-			      DRM_SCHED_PRIORITY_COUNT,
 			      1,
 			      lima_job_hang_limit,
 			      msecs_to_jiffies(timeout), NULL,

drivers/gpu/drm/msm/msm_ringbuffer.c

@@ -95,7 +95,6 @@ struct msm_ringbuffer *msm_ringbuffer_new(struct msm_gpu *gpu, int id,
 	sched_timeout = MAX_SCHEDULE_TIMEOUT;
 
 	ret = drm_sched_init(&ring->sched, &msm_sched_ops,
-			     DRM_SCHED_PRIORITY_COUNT,
 			     num_hw_submissions, 0, sched_timeout,
 			     NULL, NULL, to_msm_bo(ring->bo)->name, gpu->dev->dev);
 	if (ret) {

drivers/gpu/drm/panfrost/panfrost_job.c

@@ -871,7 +871,6 @@ int panfrost_job_init(struct panfrost_device *pfdev)
 
 		ret = drm_sched_init(&js->queue[j].sched,
 				     &panfrost_sched_ops,
-				     DRM_SCHED_PRIORITY_COUNT,
 				     nentries, 0,
 				     msecs_to_jiffies(JOB_TIMEOUT_MS),
 				     pfdev->reset.wq,

drivers/gpu/drm/scheduler/sched_entity.c

@@ -75,20 +75,8 @@ int drm_sched_entity_init(struct drm_sched_entity *entity,
 	RCU_INIT_POINTER(entity->last_scheduled, NULL);
 	RB_CLEAR_NODE(&entity->rb_tree_node);
 
-	if (!sched_list[0]->sched_rq) {
-		/* Warn drivers not to do this and to fix their DRM
-		 * calling order.
-		 */
-		pr_warn("%s: called with uninitialized scheduler\n", __func__);
-	} else if (num_sched_list) {
-		/* The "priority" of an entity cannot exceed the number
-		 * of run-queues of a scheduler.
-		 */
-		if (entity->priority >= sched_list[0]->num_rqs)
-			entity->priority = max_t(u32, sched_list[0]->num_rqs,
-						 DRM_SCHED_PRIORITY_KERNEL);
-		entity->rq = sched_list[0]->sched_rq[entity->priority];
-	}
+	if(num_sched_list)
+		entity->rq = &sched_list[0]->sched_rq[entity->priority];
 
 	init_completion(&entity->entity_idle);
 
@@ -646,7 +634,7 @@ void drm_sched_entity_select_rq(struct drm_sched_entity *entity)
 
 	spin_lock(&entity->rq_lock);
 	sched = drm_sched_pick_best(entity->sched_list, entity->num_sched_list);
-	rq = sched ? sched->sched_rq[entity->priority] : NULL;
+	rq = sched ? &sched->sched_rq[entity->priority] : NULL;
 	if (rq != entity->rq) {
 		drm_sched_rq_remove_entity(entity->rq, entity);
 		entity->rq = rq;

drivers/gpu/drm/scheduler/sched_main.c

@@ -732,14 +732,8 @@ int drm_sched_job_init(struct drm_sched_job *job,
 		       struct drm_sched_entity *entity,
 		       u32 credits, void *owner)
 {
-	if (!entity->rq) {
-		/* This will most likely be followed by missing frames
-		 * or worse--a blank screen--leave a trail in the
-		 * logs, so this can be debugged easier.
-		 */
-		drm_err(job->sched, "%s: entity has no rq!\n", __func__);
+	if (!entity->rq)
 		return -ENOENT;
-	}
 
 	if (unlikely(!credits)) {
 		pr_err("*ERROR* %s: credits cannot be 0!\n", __func__);
@@ -784,7 +778,7 @@ void drm_sched_job_arm(struct drm_sched_job *job)
 	sched = entity->rq->sched;
 
 	job->sched = sched;
-	job->s_priority = entity->priority;
+	job->s_priority = entity->rq - sched->sched_rq;
 	job->id = atomic64_inc_return(&sched->job_id_count);
 
 	drm_sched_fence_init(job->s_fence, job->entity);
@@ -999,11 +993,10 @@ drm_sched_select_entity(struct drm_gpu_scheduler *sched)
 	int i;
 
 	/* Kernel run queue has higher priority than normal run queue*/
-        for (i = DRM_SCHED_PRIORITY_KERNEL; i < sched->num_rqs; i++) {
-                entity = drm_sched_policy == DRM_SCHED_POLICY_FIFO ?
-                        drm_sched_rq_select_entity_fifo(sched, sched->sched_rq[i]) :
-                        drm_sched_rq_select_entity_rr(sched, sched->sched_rq[i]);
-
+	for (i = DRM_SCHED_PRIORITY_COUNT - 1; i >= DRM_SCHED_PRIORITY_KERNEL; i--) {
+		entity = drm_sched_policy == DRM_SCHED_POLICY_FIFO ?
+			drm_sched_rq_select_entity_fifo(sched, &sched->sched_rq[i]) :
+			drm_sched_rq_select_entity_rr(sched, &sched->sched_rq[i]);
 		if (entity)
 			break;
 	}
@@ -1183,9 +1176,7 @@ static int drm_sched_main(void *param)
  *
  * @sched: scheduler instance
  * @ops: backend operations for this scheduler
- * @num_rqs: number of runqueues, one for each priority, up to DRM_SCHED_PRIORITY_COUNT
  * @credit_limit: the number of credits this scheduler can hold from all jobs
- * @num_rqs: number of runqueues, one for each priority, up to DRM_SCHED_PRIORITY_COUNT
  * @hang_limit: number of times to allow a job to hang before dropping it
  * @timeout: timeout value in jiffies for the scheduler
  * @timeout_wq: workqueue to use for timeout work. If NULL, the system_wq is
@@ -1198,12 +1189,11 @@ static int drm_sched_main(void *param)
  */
 int drm_sched_init(struct drm_gpu_scheduler *sched,
 		   const struct drm_sched_backend_ops *ops,
-		   u32 num_rqs, u32 credit_limit, unsigned int hang_limit,
+		   u32 credit_limit, unsigned int hang_limit,
 		   long timeout, struct workqueue_struct *timeout_wq,
 		   atomic_t *score, const char *name, struct device *dev)
 {
 	int i, ret;
-
 	sched->ops = ops;
 	sched->credit_limit = credit_limit;
 	sched->name = name;
@@ -1213,35 +1203,8 @@ int drm_sched_init(struct drm_gpu_scheduler *sched,
 	sched->score = score ? score : &sched->_score;
 	sched->dev = dev;
 
-	if (num_rqs > DRM_SCHED_PRIORITY_COUNT) {
-		/* This is a gross violation--tell drivers what the  problem is.
-		 */
-		drm_err(sched, "%s: num_rqs cannot be greater than DRM_SCHED_PRIORITY_COUNT\n",
-			__func__);
-		return -EINVAL;
-	} else if (sched->sched_rq) {
-		/* Not an error, but warn anyway so drivers can
-		 * fine-tune their DRM calling order, and return all
-		 * is good.
-		 */
-		drm_warn(sched, "%s: scheduler already initialized!\n", __func__);
-		return 0;
-	}
-
-	sched->sched_rq = kmalloc_array(num_rqs, sizeof(*sched->sched_rq),
-					GFP_KERNEL | __GFP_ZERO);
-	if (!sched->sched_rq) {
-		drm_err(sched, "%s: out of memory for sched_rq\n", __func__);
-		return -ENOMEM;
-	}
-	sched->num_rqs = num_rqs;
-	ret = -ENOMEM;
-	for (i = DRM_SCHED_PRIORITY_KERNEL; i < sched->num_rqs; i++) {
-		sched->sched_rq[i] = kzalloc(sizeof(*sched->sched_rq[i]), GFP_KERNEL);
-		if (!sched->sched_rq[i])
-			goto Out_unroll;
-		drm_sched_rq_init(sched, sched->sched_rq[i]);
-	}
+	for (i = DRM_SCHED_PRIORITY_KERNEL; i < DRM_SCHED_PRIORITY_COUNT; i++)
+		drm_sched_rq_init(sched, &sched->sched_rq[i]);
 
 	init_waitqueue_head(&sched->wake_up_worker);
 	init_waitqueue_head(&sched->job_scheduled);
@@ -1252,24 +1215,18 @@ int drm_sched_init(struct drm_gpu_scheduler *sched,
 	atomic_set(&sched->_score, 0);
 	atomic64_set(&sched->job_id_count, 0);
 
-	/* Each scheduler will run on a seperate kernel thread */
-	sched->thread = kthread_run(drm_sched_main, sched, sched->name);
-	if (IS_ERR(sched->thread)) {
-		ret = PTR_ERR(sched->thread);
-		sched->thread = NULL;
-		DRM_DEV_ERROR(sched->dev, "Failed to create scheduler for %s.\n", name);
-		goto Out_unroll;
-	}
+
+        /* Each scheduler will run on a seperate kernel thread */
+        sched->thread = kthread_run(drm_sched_main, sched, sched->name);
+        if (IS_ERR(sched->thread)) {
+                ret = PTR_ERR(sched->thread);
+                sched->thread = NULL;
+                DRM_DEV_ERROR(sched->dev, "Failed to create scheduler for %s.\n", name);
+                return ret;
+        }
 
 	sched->ready = true;
 	return 0;
-Out_unroll:
-	for (--i ; i >= DRM_SCHED_PRIORITY_KERNEL; i--)
-		kfree(sched->sched_rq[i]);
-	kfree(sched->sched_rq);
-	sched->sched_rq = NULL;
-	drm_err(sched, "%s: Failed to setup GPU scheduler--out of memory\n", __func__);
-	return ret;
 }
 EXPORT_SYMBOL(drm_sched_init);
 
@@ -1288,8 +1245,8 @@ void drm_sched_fini(struct drm_gpu_scheduler *sched)
 	if (sched->thread)
 		kthread_stop(sched->thread);
 
-	for (i = DRM_SCHED_PRIORITY_KERNEL; i < sched->num_rqs; i++) {
-		struct drm_sched_rq *rq = sched->sched_rq[i];
+	for (i = DRM_SCHED_PRIORITY_COUNT - 1; i >= DRM_SCHED_PRIORITY_KERNEL; i--) {
+		struct drm_sched_rq *rq = &sched->sched_rq[i];
 
 		spin_lock(&rq->lock);
 		list_for_each_entry(s_entity, &rq->entities, list)
@@ -1300,7 +1257,7 @@ void drm_sched_fini(struct drm_gpu_scheduler *sched)
 			 */
 			s_entity->stopped = true;
 		spin_unlock(&rq->lock);
-		kfree(sched->sched_rq[i]);
+
 	}
 
 	/* Wakeup everyone stuck in drm_sched_entity_flush for this scheduler */
@@ -1310,8 +1267,6 @@ void drm_sched_fini(struct drm_gpu_scheduler *sched)
 	cancel_delayed_work_sync(&sched->work_tdr);
 
 	sched->ready = false;
-	kfree(sched->sched_rq);
-	sched->sched_rq = NULL;
 }
 EXPORT_SYMBOL(drm_sched_fini);
 
@@ -1337,8 +1292,10 @@ void drm_sched_increase_karma(struct drm_sched_job *bad)
 	 */
 	if (bad->s_priority != DRM_SCHED_PRIORITY_KERNEL) {
 		atomic_inc(&bad->karma);
-		for (i = DRM_SCHED_PRIORITY_HIGH; i < sched->num_rqs; i++) {
-			struct drm_sched_rq *rq = sched->sched_rq[i];
+
+		for (i = DRM_SCHED_PRIORITY_KERNEL; i < DRM_SCHED_PRIORITY_COUNT;
+		     i++) {
+			struct drm_sched_rq *rq = &sched->sched_rq[i];
 
 			spin_lock(&rq->lock);
 			list_for_each_entry_safe(entity, tmp, &rq->entities, list) {

drivers/gpu/drm/v3d/v3d_sched.c

@@ -733,7 +733,6 @@ v3d_sched_init(struct v3d_dev *v3d)
 
 	ret = drm_sched_init(&v3d->queue[V3D_BIN].sched,
 			     &v3d_bin_sched_ops,
-			     DRM_SCHED_PRIORITY_COUNT,
 			     hw_jobs_limit, job_hang_limit,
 			     msecs_to_jiffies(hang_limit_ms), NULL,
 			     NULL, "v3d_bin", v3d->drm.dev);
@@ -742,7 +741,6 @@ v3d_sched_init(struct v3d_dev *v3d)
 
 	ret = drm_sched_init(&v3d->queue[V3D_RENDER].sched,
 			     &v3d_render_sched_ops,
-			     DRM_SCHED_PRIORITY_COUNT,
 			     hw_jobs_limit, job_hang_limit,
 			     msecs_to_jiffies(hang_limit_ms), NULL,
 			     NULL, "v3d_render", v3d->drm.dev);
@@ -751,7 +749,6 @@ v3d_sched_init(struct v3d_dev *v3d)
 
 	ret = drm_sched_init(&v3d->queue[V3D_TFU].sched,
 			     &v3d_tfu_sched_ops,
-			     DRM_SCHED_PRIORITY_COUNT,
 			     hw_jobs_limit, job_hang_limit,
 			     msecs_to_jiffies(hang_limit_ms), NULL,
 			     NULL, "v3d_tfu", v3d->drm.dev);
@@ -761,7 +758,6 @@ v3d_sched_init(struct v3d_dev *v3d)
 	if (v3d_has_csd(v3d)) {
 		ret = drm_sched_init(&v3d->queue[V3D_CSD].sched,
 				     &v3d_csd_sched_ops,
-				     DRM_SCHED_PRIORITY_COUNT,
 				     hw_jobs_limit, job_hang_limit,
 				     msecs_to_jiffies(hang_limit_ms), NULL,
 				     NULL, "v3d_csd", v3d->drm.dev);
@@ -770,7 +766,6 @@ v3d_sched_init(struct v3d_dev *v3d)
 
 		ret = drm_sched_init(&v3d->queue[V3D_CACHE_CLEAN].sched,
 				     &v3d_cache_clean_sched_ops,
-				     DRM_SCHED_PRIORITY_COUNT,
 				     hw_jobs_limit, job_hang_limit,
 				     msecs_to_jiffies(hang_limit_ms), NULL,
 				     NULL, "v3d_cache_clean", v3d->drm.dev);

include/drm/gpu_scheduler.h

@@ -490,7 +490,6 @@ struct drm_sched_backend_ops {
  * @credit_count: the current credit count of this scheduler
  * @timeout: the time after which a job is removed from the scheduler.
  * @name: name of the ring for which this scheduler is being used.
- * @num_rqs: Number of run-queues. This is at most DRM_SCHED_PRIORITY_COUNT,
  *           as there's usually one run-queue per priority, but could be less.
  * @sched_rq: An allocated array of run-queues of size @num_rqs;
  * @wake_up_worker: the wait queue on which the scheduler sleeps until a job
@@ -521,8 +520,7 @@ struct drm_gpu_scheduler {
 	atomic_t			credit_count;
 	long				timeout;
 	const char			*name;
-	u32                             num_rqs;
-	struct drm_sched_rq             **sched_rq;
+	struct drm_sched_rq             sched_rq[DRM_SCHED_PRIORITY_COUNT];
 	wait_queue_head_t		wake_up_worker;
 	wait_queue_head_t		job_scheduled;
 	atomic64_t			job_id_count;
@@ -541,7 +539,7 @@ struct drm_gpu_scheduler {
 
 int drm_sched_init(struct drm_gpu_scheduler *sched,
 		   const struct drm_sched_backend_ops *ops,
-		   u32 num_rqs, u32 credit_limit, unsigned int hang_limit,
+		   u32 credit_limit, unsigned int hang_limit,
 		   long timeout, struct workqueue_struct *timeout_wq,
 		   atomic_t *score, const char *name, struct device *dev);