drm_sched_entity_flush should only kill the entity if the current process is the last user of the entity. The last_user is only set when adding new job, so entities that had no jobs submitted to them have NULL last_user and would always be killed. Another issue is setting last_user to NULL from drm_sched_entity_flush, which causes subsequent calls to kill the entity.
Signed-off-by: David Rosca <[email protected]> Fixes: 51564e9f06f0 ("drm/amdgpu: Avoid extra evict-restore process.") --- drivers/gpu/drm/scheduler/sched_entity.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/scheduler/sched_entity.c b/drivers/gpu/drm/scheduler/sched_entity.c index 8867b95ab089..a325e4a59990 100644 --- a/drivers/gpu/drm/scheduler/sched_entity.c +++ b/drivers/gpu/drm/scheduler/sched_entity.c @@ -70,6 +70,7 @@ int drm_sched_entity_init(struct drm_sched_entity *entity, entity->guilty = guilty; entity->num_sched_list = num_sched_list; entity->priority = priority; + entity->last_user = current->group_leader; /* * It's perfectly valid to initialize an entity without having a valid * scheduler attached. It's just not valid to use the scheduler before it @@ -278,7 +279,6 @@ static void drm_sched_entity_kill(struct drm_sched_entity *entity) long drm_sched_entity_flush(struct drm_sched_entity *entity, long timeout) { struct drm_gpu_scheduler *sched; - struct task_struct *last_user; long ret = timeout; if (!entity->rq) @@ -301,8 +301,7 @@ long drm_sched_entity_flush(struct drm_sched_entity *entity, long timeout) } /* For killed process disable any more IBs enqueue right now */ - last_user = cmpxchg(&entity->last_user, current->group_leader, NULL); - if ((!last_user || last_user == current->group_leader) && + if (entity->last_user == current->group_leader && (current->flags & PF_EXITING) && (current->exit_code == SIGKILL)) drm_sched_entity_kill(entity); -- 2.43.0
