[PATCH] drm/amdgpu: drop redudant sched job cleanup when cs is aborted

[Christian König]

Am 27.04.23 um 03:40 schrieb Guchun Chen:
> Once command submission failed due to userptr invalidation in
> amdgpu_cs_submit, legacy code will perform cleanup of scheduler
> job. However, it's not needed at all, as f7d66fb2ea43 has integrated
> job cleanup stuff into amdgpu_job_free. Otherwise, because of double
> free, a NULL pointer dereference will occur in such scenario.
>
> Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2457
> Fixes: f7d66fb2ea43("drm/amdgpu: cleanup scheduler job initialization v2")
> Signed-off-by: Guchun Chen <guchun.chen at amd.com>

Reviewed-by: Christian König <christian.koenig at amd.com>

> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 13 +++----------
>   1 file changed, 3 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> index 08eced097bd8..2eb2c66843a8 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> @@ -1276,7 +1276,7 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
>   		r = drm_sched_job_add_dependency(&leader->base, fence);
>   		if (r) {
>   			dma_fence_put(fence);
> -			goto error_cleanup;
> +			return r;
>   		}
>   	}
>   
> @@ -1303,7 +1303,8 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
>   	}
>   	if (r) {
>   		r = -EAGAIN;
> -		goto error_unlock;
> +		mutex_unlock(&p->adev->notifier_lock);
> +		return r;
>   	}
>   
>   	p->fence = dma_fence_get(&leader->base.s_fence->finished);
> @@ -1350,14 +1351,6 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
>   	mutex_unlock(&p->adev->notifier_lock);
>   	mutex_unlock(&p->bo_list->bo_list_mutex);
>   	return 0;
> -
> -error_unlock:
> -	mutex_unlock(&p->adev->notifier_lock);
> -
> -error_cleanup:
> -	for (i = 0; i < p->gang_size; ++i)
> -		drm_sched_job_cleanup(&p->jobs[i]->base);
> -	return r;
>   }
>   
>   /* Cleanup the parser structure */