From 4480e081b60fba33f388a551bd97234d29d42cc3 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Nicolai=20H=C3=A4hnle?= Date: Wed, 2 Dec 2015 13:28:25 +0100 Subject: [PATCH] drm/amdgpu: fix race condition in amd_sched_entity_push_job MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit As soon as we leave the spinlock after the job has been added to the job queue, we can no longer rely on the job's data to be available. I have seen a null-pointer dereference due to sched == NULL in amd_sched_wakeup via amd_sched_entity_push_job and amd_sched_ib_submit_kernel_helper. Since the latter initializes sched_job->sched with the address of the ring scheduler, which is guaranteed to be non-NULL, this race appears to be a likely culprit. Signed-off-by: Nicolai Hähnle Bugzilla: https://bugs.freedesktop.org/attachment.cgi?bugid=93079 Reviewed-by: Christian König --- drivers/gpu/drm/amd/scheduler/gpu_scheduler.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c index e13b7a0..5ace1a7 100644 --- a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c +++ b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c @@ -288,6 +288,7 @@ amd_sched_entity_pop_job(struct amd_sched_entity *entity) */ static bool amd_sched_entity_in(struct amd_sched_job *sched_job) { + struct amd_gpu_scheduler *sched = sched_job->sched; struct amd_sched_entity *entity = sched_job->s_entity; bool added, first = false; @@ -302,7 +303,7 @@ static bool amd_sched_entity_in(struct amd_sched_job *sched_job) /* first job wakes up scheduler */ if (first) - amd_sched_wakeup(sched_job->sched); + amd_sched_wakeup(sched); return added; } @@ -318,9 +319,9 @@ void amd_sched_entity_push_job(struct amd_sched_job *sched_job) { struct amd_sched_entity *entity = sched_job->s_entity; + trace_amd_sched_job(sched_job); wait_event(entity->sched->job_scheduled, amd_sched_entity_in(sched_job)); - trace_amd_sched_job(sched_job); } /** -- 2.5.0