Re: [PATCH] drm/sched: Avoid data corruptions

[Christian König]

Am 19.05.21 um 16:14 schrieb Andrey Grodzovsky:

Wait for all dependencies of a job  to complete before
killing it to avoid data corruptions.

Signed-off-by: Andrey Grodzovsky 


Reviewed-by: Christian König 


---
  drivers/gpu/drm/scheduler/sched_entity.c | 5 +
  1 file changed, 5 insertions(+)

diff --git a/drivers/gpu/drm/scheduler/sched_entity.c 
b/drivers/gpu/drm/scheduler/sched_entity.c
index 2e93e881b65f..d5cf61972558 100644
--- a/drivers/gpu/drm/scheduler/sched_entity.c
+++ b/drivers/gpu/drm/scheduler/sched_entity.c
@@ -222,11 +222,16 @@ static void drm_sched_entity_kill_jobs_cb(struct 
dma_fence *f,
  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(>job_queue {

struct drm_sched_fence *s_fence = job->s_fence;
  
+		/* Wait for all dependencies to avoid data corruptions */

+   while ((f = job->sched->ops->dependency(job, entity)))
+   dma_fence_wait(f);
+
drm_sched_fence_scheduled(s_fence);
dma_fence_set_error(_fence->finished, -ESRCH);
  


___
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH] drm/sched: Avoid data corruptions

[Andrey Grodzovsky]

Wait for all dependencies of a job  to complete before
killing it to avoid data corruptions.

Signed-off-by: Andrey Grodzovsky 
---
 drivers/gpu/drm/scheduler/sched_entity.c | 5 +
 1 file changed, 5 insertions(+)

diff --git a/drivers/gpu/drm/scheduler/sched_entity.c 
b/drivers/gpu/drm/scheduler/sched_entity.c
index 2e93e881b65f..d5cf61972558 100644
--- a/drivers/gpu/drm/scheduler/sched_entity.c
+++ b/drivers/gpu/drm/scheduler/sched_entity.c
@@ -222,11 +222,16 @@ static void drm_sched_entity_kill_jobs_cb(struct 
dma_fence *f,
 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(>job_queue {
struct drm_sched_fence *s_fence = job->s_fence;
 
+   /* Wait for all dependencies to avoid data corruptions */
+   while ((f = job->sched->ops->dependency(job, entity)))
+   dma_fence_wait(f);
+
drm_sched_fence_scheduled(s_fence);
dma_fence_set_error(_fence->finished, -ESRCH);
 
-- 
2.25.1

___
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx