Commit f3cd1b06 authored by Lucas Stach's avatar Lucas Stach

drm/etnaviv: (re-)protect fence allocation with GPU mutex

The fence allocation needs to be protected by the GPU mutex, otherwise
the fence seqnos of concurrent submits might not match the insertion order
of the jobs in the kernel ring. This breaks the assumption that jobs
complete with monotonically increasing fence seqnos.

Fixes: d9853490 (drm/etnaviv: take GPU lock later in the submit process)
CC: stable@vger.kernel.org #4.9+
Signed-off-by: default avatarLucas Stach <l.stach@pengutronix.de>
parent c1ae3cfa
...@@ -1311,6 +1311,8 @@ int etnaviv_gpu_submit(struct etnaviv_gpu *gpu, ...@@ -1311,6 +1311,8 @@ int etnaviv_gpu_submit(struct etnaviv_gpu *gpu,
goto out_pm_put; goto out_pm_put;
} }
mutex_lock(&gpu->lock);
fence = etnaviv_gpu_fence_alloc(gpu); fence = etnaviv_gpu_fence_alloc(gpu);
if (!fence) { if (!fence) {
event_free(gpu, event); event_free(gpu, event);
...@@ -1318,8 +1320,6 @@ int etnaviv_gpu_submit(struct etnaviv_gpu *gpu, ...@@ -1318,8 +1320,6 @@ int etnaviv_gpu_submit(struct etnaviv_gpu *gpu,
goto out_pm_put; goto out_pm_put;
} }
mutex_lock(&gpu->lock);
gpu->event[event].fence = fence; gpu->event[event].fence = fence;
submit->fence = fence->seqno; submit->fence = fence->seqno;
gpu->active_fence = submit->fence; gpu->active_fence = submit->fence;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment