On Fri, 29 Nov 2019 14:19:50 +0000 Steven Price steven.price@arm.com wrote:
On 29/11/2019 13:59, Boris Brezillon wrote:
If we don't do that, dma_fence_set_error() complains (called from drm_sched_main()).
Fixes: f3ba91228e8e ("drm/panfrost: Add initial panfrost driver") Cc: stable@vger.kernel.org Signed-off-by: Boris Brezillon boris.brezillon@collabora.com
This might be worth doing, but actually it's not Panfrost that is broken it's the callers, see [1] and [2]. So I don't think we want the Fixes/stable tag.
Okay.
[1] https://patchwork.kernel.org/patch/11218399/ [2] https://patchwork.kernel.org/patch/11267073/
drivers/gpu/drm/panfrost/panfrost_job.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/panfrost/panfrost_job.c b/drivers/gpu/drm/panfrost/panfrost_job.c index 21f34d44aac2..cdd9448fbbdd 100644 --- a/drivers/gpu/drm/panfrost/panfrost_job.c +++ b/drivers/gpu/drm/panfrost/panfrost_job.c @@ -328,13 +328,13 @@ static struct dma_fence *panfrost_job_run(struct drm_sched_job *sched_job) struct dma_fence *fence = NULL; if (unlikely(job->base.s_fence->finished.error))
return NULL;
return ERR_PTR(job->base.s_fence->finished.error);
Hm, so we can keep the return NULL here if [1] is applied (the error is preserved), but I'm not sure it's clearer that way.
pfdev->jobs[slot] = job; fence = panfrost_fence_create(pfdev, slot); if (IS_ERR(fence))
return NULL;
return ERR_PTR(-ENOMEM);
This one should be fixed though, otherwise the error is never updated, so I'm wondering if it doesn't deserve a Fixes tag in the end...
Why override the error from panfrost_fence_create? In this case we can just:
return fence;
Indeed.
Steve
if (job->done_fence) dma_fence_put(job->done_fence);