Skip to contentSkip to conversation

drm/scheduler: add NULL pointer check for run queue

PATCH In review Junwei Zhang 3 participants 5 messages Jul 16, 2018 4 messages not in the archive

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/scheduler: add NULL pointer check for run queue

To check rq pointer before adding entity into it.
That avoids NULL pointer access in some case.

---
 drivers/gpu/drm/scheduler/gpu_scheduler.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/gpu/drm/scheduler/gpu_scheduler.c +4 -0
index 16bf446..5e5268d 100644
@@ -91,6 +91,10 @@ static void drm_sched_rq_add_entity(struct drm_sched_rq *rq,
 {
 	if (!list_empty(&entity->list))
 		return;
+	if (!rq) {
+		DRM_ERROR("rq is NULL!\n");
+		return;
+	}
 	spin_lock(&rq->lock);
 	list_add_tail(&entity->list, &rq->entities);
 	spin_unlock(&rq->lock);
Signature · 6 lines
1.9.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): 1531709980-21298-1-git-send-email-Jerry.Zhang-5C7GfCeVMHo@public.gmane.org

#2 lore raw

Am 16.07.2018 um 04:59 schrieb Junwei Zhang:

Quoted 14 lines

To check rq pointer before adding entity into it.
That avoids NULL pointer access in some case.

---
  drivers/gpu/drm/scheduler/gpu_scheduler.c | 4 ++++
  1 file changed, 4 insertions(+)
drivers/gpu/drm/scheduler/gpu_scheduler.c +0 -0
index 16bf446..5e5268d 100644

drivers/gpu/drm/scheduler/gpu_scheduler.c static void drm_sched_rq_add_entity(struct drm_sched_rq *rq, · line 97

Quoted 7 lines
drivers/gpu/drm/scheduler/gpu_scheduler.c +4 -0
@@ -91,6 +91,10 @@ static void drm_sched_rq_add_entity(struct drm_sched_rq *rq,
  {
  	if (!list_empty(&entity->list))
  		return;
+	if (!rq) {
+		DRM_ERROR("rq is NULL!\n");
+		return;
+	}

Better put that into drm_sched_entity_push_job(), e.g. something like:

/* first job wakes up scheduler */
if (first) {
    /* Add the entity to the run queue */ spin_lock(&entity->rq_lock);
    if (!entity->rq) {
        DRM_ERROR("Trying to push to killed entity!\n");
....

Regards,
Christian.

drivers/gpu/drm/scheduler/gpu_scheduler.c static void drm_sched_rq_add_entity(struct drm_sched_rq *rq, · line 100

Quoted 3 lines
drivers/gpu/drm/scheduler/gpu_scheduler.c +0 -0
  	spin_lock(&rq->lock);
  	list_add_tail(&entity->list, &rq->entities);
  	spin_unlock(&rq->lock);

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): 045ffcc7-9b9a-2534-9ebe-295e18ce62e7-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org

#4 lore raw
Quoted 10 lines · Christian König

Am 16.07.2018 um 04:59 schrieb Junwei Zhang:

Quoted 9 lines

To check rq pointer before adding entity into it.
That avoids NULL pointer access in some case.

---
  drivers/gpu/drm/scheduler/gpu_scheduler.c | 4 ++++
  1 file changed, 4 insertions(+)
Quoted 14 lines

drivers/gpu/drm/scheduler/gpu_scheduler.c static void drm_sched_rq_add_entity(struct drm_sched_rq *rq, · line 97

Quoted 12 lines
drivers/gpu/drm/scheduler/gpu_scheduler.c +4 -0
index 16bf446..5e5268d 100644
@@ -91,6 +91,10 @@ static void drm_sched_rq_add_entity(struct drm_sched_rq *rq,
  {
      if (!list_empty(&entity->list))
          return;
+    if (!rq) {
+        DRM_ERROR("rq is NULL!\n");
+        return;
+    }

Better put that into drm_sched_entity_push_job(), e.g. something like:

Considered that as well.
Just be afraid of someone else could call it in another place without rq checking in the future.

Regards,
Jerry

Quoted 7 lines

/* first job wakes up scheduler */

if (first) {
     /* Add the entity to the run queue */
     spin_lock(&entity->rq_lock);
     if (!entity->rq) {
         DRM_ERROR("Trying to push to killed entity!\n");
Quoted 9 lines

....

Regards,
Christian.

Quoted 3 lines
drivers/gpu/drm/scheduler/gpu_scheduler.c +0 -0
      spin_lock(&rq->lock);
      list_add_tail(&entity->list, &rq->entities);
      spin_unlock(&rq->lock);

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): [email protected]

#6 lore raw

Am 16.07.2018 um 11:01 schrieb Zhang, Jerry (Junwei):

Quoted 27 lines · Christian König
Quoted 10 lines

Am 16.07.2018 um 04:59 schrieb Junwei Zhang:

Quoted 9 lines

To check rq pointer before adding entity into it.
That avoids NULL pointer access in some case.

---
  drivers/gpu/drm/scheduler/gpu_scheduler.c | 4 ++++

  1 file changed, 4 insertions(+)
Quoted 15 lines

drivers/gpu/drm/scheduler/gpu_scheduler.c static void drm_sched_rq_add_entity(struct · line 98

Quoted 13 lines
drivers/gpu/drm/scheduler/gpu_scheduler.c +4 -0
index 16bf446..5e5268d 100644
@@ -91,6 +91,10 @@ static void drm_sched_rq_add_entity(struct 
 drm_sched_rq *rq,
   {
       if (!list_empty(&entity->list))
           return;
+    if (!rq) {
+        DRM_ERROR("rq is NULL!\n");
+        return;
+    }

Better put that into drm_sched_entity_push_job(), e.g. something like:

Quoted 2 lines

Considered that as well.
Just be afraid of someone else could call it in another place without rq checking in the future.

Well that's exactly the reason why I wanted to have the check in drm_sched_rq_add_entity().

Calling drm_sched_rq_add_entity() will a NULL rq is illegal and that should be avoided in the caller instead of more or less silently dropped in the function.

Regards,
Christian.

Quoted 20 lines

Regards,
Jerry

Quoted 16 lines

/* first job wakes up scheduler */
if (first) {
     /* Add the entity to the run queue */ spin_lock(&entity->rq_lock);
     if (!entity->rq) {
         DRM_ERROR("Trying to push to killed entity!\n");
....

Regards,
Christian.

Quoted 3 lines
drivers/gpu/drm/scheduler/gpu_scheduler.c +0 -0
       spin_lock(&rq->lock);
       list_add_tail(&entity->list, &rq->entities);

      spin_unlock(&rq->lock);

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): e2dc7604-e706-c646-bf98-c058d847695e-5C7GfCeVMHo@public.gmane.org

#8 lore raw
Quoted 27 lines · Christian König

Am 16.07.2018 um 11:01 schrieb Zhang, Jerry (Junwei):

Quoted 26 lines · Christian König
Quoted 24 lines

Am 16.07.2018 um 04:59 schrieb Junwei Zhang:

Quoted 14 lines

To check rq pointer before adding entity into it.
That avoids NULL pointer access in some case.

---
  drivers/gpu/drm/scheduler/gpu_scheduler.c | 4 ++++
  1 file changed, 4 insertions(+)
drivers/gpu/drm/scheduler/gpu_scheduler.c +0 -0
index 16bf446..5e5268d 100644

drivers/gpu/drm/scheduler/gpu_scheduler.c static void drm_sched_rq_add_entity(struct drm_sched_rq *rq, · line 97

Quoted 7 lines
drivers/gpu/drm/scheduler/gpu_scheduler.c +4 -0
@@ -91,6 +91,10 @@ static void drm_sched_rq_add_entity(struct drm_sched_rq *rq,
  {
      if (!list_empty(&entity->list))
          return;
+    if (!rq) {
+        DRM_ERROR("rq is NULL!\n");
+        return;
+    }

Better put that into drm_sched_entity_push_job(), e.g. something like:

Quoted 6 lines
Quoted 2 lines

Considered that as well.
Just be afraid of someone else could call it in another place without rq checking in the future.

Well that's exactly the reason why I wanted to have the check in drm_sched_rq_add_entity().

Calling drm_sched_rq_add_entity() will a NULL rq is illegal and that should be avoided in the caller instead of more or less silently dropped in the function.

Fine, to expose the error in place explicitly.

Regards,
Jerry

Quoted 25 lines

Regards,
Christian.

Quoted 20 lines

Regards,
Jerry

Quoted 16 lines

/* first job wakes up scheduler */

if (first) {
     /* Add the entity to the run queue */
     spin_lock(&entity->rq_lock);
     if (!entity->rq) {
         DRM_ERROR("Trying to push to killed entity!\n");

....

Regards,
Christian.

Quoted 3 lines
drivers/gpu/drm/scheduler/gpu_scheduler.c +0 -0
      spin_lock(&rq->lock);
      list_add_tail(&entity->list, &rq->entities);
      spin_unlock(&rq->lock);
Quoted 4 lines

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx