Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <888841c4-7bd4-8174-7786-033715c995c6@linux.intel.com>
Date:   Wed, 3 May 2023 11:11:32 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.8.0
Subject: Re: [Intel-xe] [RFC PATCH 3/4] drm/ttm: Handle -EAGAIN in
 ttm_resource_alloc as -ENOSPC.
Content-Language: en-US
To:     Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
        dri-devel@lists.freedesktop.org, cgroups@vger.kernel.org,
        intel-xe@lists.freedesktop.org
Cc:     Daniel Vetter <daniel@ffwll.ch>,
        Tvrtko Ursulin <tvrtko.ursulin@intel.com>,
        Thomas Zimmermann <tzimmermann@suse.de>,
        intel-gfx@lists.freedesktop.org, linux-kernel@vger.kernel.org,
        amd-gfx@lists.freedesktop.org, Maxime Ripard <mripard@kernel.org>,
        Zefan Li <lizefan.x@bytedance.com>,
        Johannes Weiner <hannes@cmpxchg.org>,
        Tejun Heo <tj@kernel.org>, David Airlie <airlied@gmail.com>
References: <20230503083500.645848-1-maarten.lankhorst@linux.intel.com>
 <20230503083500.645848-4-maarten.lankhorst@linux.intel.com>
From:   =?UTF-8?Q?Thomas_Hellstr=c3=b6m?= 
        <thomas.hellstrom@linux.intel.com>
In-Reply-To: <20230503083500.645848-4-maarten.lankhorst@linux.intel.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

Hi, Maarten

On 5/3/23 10:34, Maarten Lankhorst wrote:
> This allows the drm cgroup controller to return no space is available..
>
> XXX: This is a hopeless simplification that changes behavior, and
> returns -ENOSPC even if we could evict ourselves from the current
> cgroup.
>
> Ideally, the eviction code becomes cgroup aware, and will force eviction
> from the current cgroup or its parents.
>
> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>

Thinking of the shrinker analogy, do non-cgroup aware shrinkers just 
shrink blindly or do they reject shrinking like this patch when a cgroup 
limit is reached?

/Thomas


> ---
>   drivers/gpu/drm/ttm/ttm_bo.c | 4 +++-
>   1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
> index bd5dae4d1624..e057d5d8f09a 100644
> --- a/drivers/gpu/drm/ttm/ttm_bo.c
> +++ b/drivers/gpu/drm/ttm/ttm_bo.c
> @@ -731,6 +731,8 @@ static int ttm_bo_mem_force_space(struct ttm_buffer_object *bo,
>   		ret = ttm_resource_alloc(bo, place, mem);
>   		if (likely(!ret))
>   			break;
> +		if (ret == -EAGAIN)
> +			return -ENOSPC;
>   		if (unlikely(ret != -ENOSPC))
>   			return ret;
>   		ret = ttm_mem_evict_first(bdev, man, place, ctx,
> @@ -783,7 +785,7 @@ int ttm_bo_mem_space(struct ttm_buffer_object *bo,
>   
>   		type_found = true;
>   		ret = ttm_resource_alloc(bo, place, mem);
> -		if (ret == -ENOSPC)
> +		if (ret == -ENOSPC || ret == -EAGAIN)
>   			continue;
>   		if (unlikely(ret))
>   			goto error;