[PATCH 1/2] drm/amdgpu: fix a potential deadlock in amdgpu_bo_create_restricted()

[Michel Dänzer]

On 09/02/17 08:35 AM, Samuel Pitoiset wrote:
> When ttm_bo_init() fails, the reservation mutex should be unlocked.
> 
> In debug build, the kernel reported "possible recursive locking
> detected" in this codepath. For debugging purposes, I also added
> a "WARN_ON(ww_mutex_is_locked())" when ttm_bo_init() fails and the
> mutex was locked as expected.
> 
> This should fix (random) GPU hangs. The easy way to reproduce the
> issue is to change the "Super Sampling" option from 1.0 to 2.0 in
> Hitman. It will create a huge buffer, evict a bunch of buffers
> (around ~5k) and deadlock.
> 
> This regression has been introduced pretty recently.
> 
> Fixes: "drm/amdgpu: improve AMDGPU_GEM_CREATE_VRAM_CLEARED handling (v2)"

Please add at least the first ~12 characters of the commit hash
(f45dc74c93241ad0125fbc08c48b2ebe20f2f472) to the Fixes tag.


> @@ -403,8 +403,10 @@ int amdgpu_bo_create_restricted(struct amdgpu_device *adev,
>  			&bo->placement, page_align, !kernel, NULL,
>  			acc_size, sg, resv ? resv : &bo->tbo.ttm_resv,
>  			&amdgpu_ttm_bo_destroy);
> -	if (unlikely(r != 0))
> +	if (unlikely(r != 0)) {
> +		ww_mutex_unlock(&bo->tbo.resv->lock);
>  		return r;
> +	}

This must only be done if (!resv).


-- 
Earthling Michel Dänzer               |               http://www.amd.com
Libre software enthusiast             |             Mesa and X developer