Subject: Re: [PATCH] drm/ttm: optimize errors checking and free _manager when
 finishing
To: "He, Roger" <Hongbo.He@amd.com>, Xiongwei Song <sxwjean@gmail.com>,
        "airlied@linux.ie" <airlied@linux.ie>
Cc: "dri-devel@lists.freedesktop.org" <dri-devel@lists.freedesktop.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
References: <20171231114024.5026-1-sxwjean@gmail.com>
 <MWHPR1201MB01273E49A430E2F0ED7DF9DFFD190@MWHPR1201MB0127.namprd12.prod.outlook.com>
From: =?UTF-8?Q?Christian_K=c3=b6nig?= <christian.koenig@amd.com>
Message-ID: <aa789821-a730-4b71-f73f-05ff04cf89d3@amd.com>
Date: Tue, 2 Jan 2018 10:29:45 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.5.0
MIME-Version: 1.0
In-Reply-To: <MWHPR1201MB01273E49A430E2F0ED7DF9DFFD190@MWHPR1201MB0127.namprd12.prod.outlook.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Content-Language: en-US
SpamDiagnosticOutput: 1:99
SpamDiagnosticMetadata: NSPM
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 02 Jan 2018 09:29:56.7712 (UTC)
X-MS-Exchange-CrossTenant-Network-Message-Id: bc46e199-2a46-43fc-5045-08d551c36439
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-CrossTenant-Id: 3dd8961f-e488-4e60-8e11-a82d994e183d
X-MS-Exchange-Transport-CrossTenantHeadersStamped: MWHPR12MB1309
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org
Return-Path: <linux-kernel-owner@vger.kernel.org>

Am 02.01.2018 um 04:05 schrieb He, Roger:
>
> -----Original Message-----
> From: Xiongwei Song [mailto:sxwjean@gmail.com]
> Sent: Sunday, December 31, 2017 7:40 PM
> To: Koenig, Christian <Christian.Koenig@amd.com>; He, Roger <Hongbo.He@amd.com>; airlied@linux.ie
> Cc: dri-devel@lists.freedesktop.org; linux-kernel@vger.kernel.org
> Subject: [PATCH] drm/ttm: optimize errors checking and free _manager when finishing
>
> In the function ttm_page_alloc_init, kzalloc call is made for variable _manager, we need to check its return value, it may return NULL.
>
> In the function ttm_page_alloc_fini, we need to call kfree for variable _manager, instead of make _manager NULL directly.
>
> Signed-off-by: Xiongwei Song <sxwjean@gmail.com>
> ---
>   drivers/gpu/drm/ttm/ttm_page_alloc.c | 13 +++++++++++--
>   1 file changed, 11 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/ttm/ttm_page_alloc.c b/drivers/gpu/drm/ttm/ttm_page_alloc.c
> index b5ba6441489f..e20a0b8e352b 100644
> --- a/drivers/gpu/drm/ttm/ttm_page_alloc.c
> +++ b/drivers/gpu/drm/ttm/ttm_page_alloc.c
> @@ -1007,6 +1007,10 @@ int ttm_page_alloc_init(struct ttm_mem_global *glob, unsigned max_pages)
>   	pr_info("Initializing pool allocator\n");
>   
>   	_manager = kzalloc(sizeof(*_manager), GFP_KERNEL);
> +	if (!_manager) {
> +		ret = -ENOMEM;
> +		goto out;
> +	}
> Seems we only need above here for this patch I think.
> The rest is no need, because ttm_pool_kobj_release will kfree _manager.

Yes, agree. Freeing it again in ttm_page_alloc_fini will actually double 
free the memory and probably cause some problems.

Please refine the patch with only the additional error handling.

Regards,
Christian.

>
>
> Thanks
> Roger(Hongbo.He)
>
>   	ttm_page_pool_init_locked(&_manager->wc_pool, GFP_HIGHUSER, "wc", 0);
>   
> @@ -1034,13 +1038,17 @@ int ttm_page_alloc_init(struct ttm_mem_global *glob, unsigned max_pages)
>   				   &glob->kobj, "pool");
>   	if (unlikely(ret != 0)) {
>   		kobject_put(&_manager->kobj);
> -		_manager = NULL;
> -		return ret;
> +		goto out_free_mgr;
>   	}
>   
>   	ttm_pool_mm_shrink_init(_manager);
>   
>   	return 0;
> +out_free_mgr:
> +	kfree(_manager);
> +	_manager = NULL;
> +out:
> +	return ret;
>   }
>   
>   void ttm_page_alloc_fini(void)
> @@ -1055,6 +1063,7 @@ void ttm_page_alloc_fini(void)
>   		ttm_page_pool_free(&_manager->pools[i], FREE_ALL_PAGES, true);
>   
>   	kobject_put(&_manager->kobj);
> +	kfree(_manager);
>   	_manager = NULL;
>   }
>   
> --
> 2.15.1
>