drm/panfrost: Prevent concurrent resets

If a job times out in slot 0 while a reset is performed because a job
timed out in slot 1, the drm-sched core can get into a deadlock.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Signed-off-by: Rob Herring <robh@kernel.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20190418084305.45021-1-tomeu.vizoso@collabora.com
This commit is contained in:
Tomeu Vizoso
2019-04-18 10:41:48 +02:00
committed by Rob Herring
parent db594ba3fc
commit aa20236784
3 changed files with 6 additions and 0 deletions

View File

@@ -78,6 +78,7 @@ struct panfrost_device {
struct list_head scheduled_jobs;
struct mutex sched_lock;
struct mutex reset_lock;
struct {
struct devfreq *devfreq;