diff options
| author |   Andrey Grodzovsky <andrey.grodzovsky@amd.com> | 2018-05-15 14:42:20 -0400 |
|---|---|---|
| committer |   Alex Deucher <alexander.deucher@amd.com> | 2018-05-18 16:08:17 -0500 |
| commit | 563e1e664d27292a3b55ca08366dc8c32db52450 (patch) | |
| tree | 73e8668bc916f9ebee6dcfb338ea28d2feb41500 /drivers/gpu/drm/scheduler/gpu_scheduler.c | |
| parent | drm/amdgpu: remove unused member (diff) | |
| download | linux-dev-563e1e664d27292a3b55ca08366dc8c32db52450.tar.xz linux-dev-563e1e664d27292a3b55ca08366dc8c32db52450.zip | |
drm/scheduler: Remove obsolete spinlock.
This spinlock is superfluous, any call to drm_sched_entity_push_job
should already be under a lock together with matching drm_sched_job_init
to match the order of insertion into queue with job's fence seqence
number.
v2:
Improve patch description.
Add functions documentation describing the locking considerations
Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Acked-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Diffstat (limited to 'drivers/gpu/drm/scheduler/gpu_scheduler.c')
| -rw-r--r-- | drivers/gpu/drm/scheduler/gpu_scheduler.c | 15 |
1 files changed, 10 insertions, 5 deletions
diff --git a/drivers/gpu/drm/scheduler/gpu_scheduler.c b/drivers/gpu/drm/scheduler/gpu_scheduler.c index a364fc0b38c3..df1578d6f42e 100644 --- a/drivers/gpu/drm/scheduler/gpu_scheduler.c +++ b/drivers/gpu/drm/scheduler/gpu_scheduler.c @@ -139,7 +139,6 @@ int drm_sched_entity_init(struct drm_gpu_scheduler *sched, entity->last_scheduled = NULL; spin_lock_init(&entity->rq_lock); - spin_lock_init(&entity->queue_lock); spsc_queue_init(&entity->job_queue); atomic_set(&entity->fence_seq, 0); @@ -413,6 +412,10 @@ drm_sched_entity_pop_job(struct drm_sched_entity *entity) * * @sched_job The pointer to job required to submit * + * Note: To guarantee that the order of insertion to queue matches + * the job's fence sequence number this function should be + * called with drm_sched_job_init under common lock. + * * Returns 0 for success, negative error code otherwise. */ void drm_sched_entity_push_job(struct drm_sched_job *sched_job, @@ -423,11 +426,8 @@ void drm_sched_entity_push_job(struct drm_sched_job *sched_job, trace_drm_sched_job(sched_job, entity); - spin_lock(&entity->queue_lock); first = spsc_queue_push(&entity->job_queue, &sched_job->queue_node); - spin_unlock(&entity->queue_lock); - /* first job wakes up scheduler */ if (first) { /* Add the entity to the run queue */ @@ -593,7 +593,12 @@ void drm_sched_job_recovery(struct drm_gpu_scheduler *sched) } EXPORT_SYMBOL(drm_sched_job_recovery); -/* init a sched_job with basic field */ +/** + * Init a sched_job with basic field + * + * Note: Refer to drm_sched_entity_push_job documentation + * for locking considerations. + */ int drm_sched_job_init(struct drm_sched_job *job, struct drm_gpu_scheduler *sched, struct drm_sched_entity *entity, |