There is no need to synchronize with oustanding retire jobs if the object
has gone idle. Retire jobs only ever change the object state from active to
idle, not the other way around.
The IOVA put race is uncritical, as the GEM_WAIT ioctl itself is holding
a reference to the GEM object, so the retire worker will not pull the
object into the CPU domain, which is the thing we are trying to guard
against with etnaviv_gpu_wait_obj_inactive.
Signed-off-by: Lucas Stach <l.st...@pengutronix.de>
---
drivers/gpu/drm/etnaviv/etnaviv_gpu.c | 11 +++--------
1 file changed, 3 insertions(+), 8 deletions(-)
diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
index 3dffccfcefce..ae152bb78f18 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
@@ -1295,17 +1295,12 @@ int etnaviv_gpu_wait_obj_inactive(struct etnaviv_gpu
*gpu,
ret = wait_event_interruptible_timeout(gpu->fence_event,
!is_active(etnaviv_obj),
remaining);
- if (ret > 0) {
- struct etnaviv_drm_private *priv = gpu->drm->dev_private;
-
- /* Synchronise with the retire worker */
- flush_workqueue(priv->wq);
+ if (ret > 0)
return 0;
- } else if (ret == -ERESTARTSYS) {
+ else if (ret == -ERESTARTSYS)
return -ERESTARTSYS;
- } else {
+ else
return -ETIMEDOUT;
- }
}
int etnaviv_gpu_pm_get_sync(struct etnaviv_gpu *gpu)
--
2.11.0
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
