Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [RFC PATCH 2/4] DRM: Add support of AI Processor Unit (APU)
To:     Dave Airlie <airlied@gmail.com>,
        Alexandre Bailon <abailon@baylibre.com>
Cc:     Dave Airlie <airlied@linux.ie>, Daniel Vetter <daniel@ffwll.ch>,
        Rob Herring <robh+dt@kernel.org>,
        Matthias Brugger <matthias.bgg@gmail.com>,
        Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
        Maxime Ripard <mripard@kernel.org>,
        Thomas Zimmermann <tzimmermann@suse.de>, ohad@wizery.com,
        bjorn.andersson@linaro.org,
        Mathieu Poirier <mathieu.poirier@linaro.org>,
        Sumit Semwal <sumit.semwal@linaro.org>,
        dri-devel <dri-devel@lists.freedesktop.org>,
        "open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS" 
        <devicetree@vger.kernel.org>,
        linux-arm-kernel <linux-arm-kernel@lists.infradead.org>,
        "moderated list:ARM/Mediatek SoC support" 
        <linux-mediatek@lists.infradead.org>,
        LKML <linux-kernel@vger.kernel.org>,
        linux-remoteproc@vger.kernel.org,
        Linux Media Mailing List <linux-media@vger.kernel.org>,
        "moderated list:DMA BUFFER SHARING FRAMEWORK" 
        <linaro-mm-sig@lists.linaro.org>, khilman@baylibre.com,
        gpain@baylibre.com
References: <20210917125945.620097-1-abailon@baylibre.com>
 <20210917125945.620097-3-abailon@baylibre.com>
 <CAPM=9tzOADabEgEP1L+yNO4gj2JhNuVDL-Bhpbsz4=UX5feLcg@mail.gmail.com>
From:   =?UTF-8?Q?Christian_K=c3=b6nig?= <christian.koenig@amd.com>
Message-ID: <ffc29919-809f-05c3-6079-62f1e6453f24@amd.com>
Date:   Thu, 23 Sep 2021 08:17:33 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.13.0
In-Reply-To: <CAPM=9tzOADabEgEP1L+yNO4gj2JhNuVDL-Bhpbsz4=UX5feLcg@mail.gmail.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Content-Language: en-US
MIME-Version: 1.0
X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1
X-MS-Exchange-AntiSpam-MessageData-0: =?utf-8?B?bFJEQVNzM2ZMazNJQlFkSTU5QUFDTjNVakNRM2JJR3krMWxzVEFIUWRQM0J2?=
 =?utf-8?B?eVlPNTM1Q004TklzUXp4T2h5TldydXdGZWxhZDFXeC80L2NSTWFFSFhINTNz?=
 =?utf-8?B?RFBQa2JaRnFUbVZGYTJmbVNpVnlhWUs0bUVBbUZNN3VnQXFuRTFoUWwxUlBi?=
 =?utf-8?B?YitGdjVZVUFnMzE3VGVwRUpqN1FSZ1Y5QmFFaEZobi9CZjBpc0VkSlJNajZ4?=
 =?utf-8?B?SkNSVVVnTlRtbXJpaHFOQ1FSWjdlbE9EVjVGMWl5K0xyTTRWZ3pQdktmZjc1?=
 =?utf-8?B?Qm9wZ2xLcGl2RmdWOE9OaVMra09samlRVnRyRGlaK2RkMHB4aGJkdnRtek1E?=
 =?utf-8?B?M2g0TFF6UlBLQU0xZGtIaFRKNnNnaTBOY0wwam9LZFI1RU9hZFdLMFVlZHlo?=
 =?utf-8?B?OEdlc2pTdDJjVFRRMHJqUlBJV3MrMHhIQlpFOElMVWNUemRveFlJWEtZNmsv?=
 =?utf-8?B?TkRSdjJjVElYdEJSK0F5OW5DaDd0MWJ4cWZkQnhLM29HU1JJeFRRVFVkdFJW?=
 =?utf-8?B?akpuZWluSWRxK0NuSzN5VmgzeTIvTXpLZ2E5cTRTb1FOUDdkSWFudkNkSExM?=
 =?utf-8?B?Y2ZnMlRSUGRTcVIxR3JxR2Z0NzdHZWg1ZFlpR3NhU0psOFFSVUZwdVAveEdV?=
 =?utf-8?B?REJKeVlrNFBqOUdZSFZXRDRSQWFob3ZCampaakVpbUZ6OW9xY21UK0xQckFa?=
 =?utf-8?B?N05tb2UrUCtJbjZwYVFScHRZVHcyTC9sVjlydFdMaUlmRzZjamhFalZST09q?=
 =?utf-8?B?TkZFSXRaeUp3dHlLYlZjOUx3ajZFQWI5a21HVVlhZGQ5dE9QaDhhV1hCNUFD?=
 =?utf-8?B?VkhOV3hMRFdISlJJR2czSmVZeE5KaHg4Z0JqZ0JxZGlxMzdWSlVmS1MvbThy?=
 =?utf-8?B?bk9kb3oxTk1oMi8zL1hKZVpQNUlCY1F1WVptdmlZVUhmWXVFVFg5dnZiMUFG?=
 =?utf-8?B?TWZQRysrbVhOblF3bStrM0ZPNjBnS1pBOVAvRmsvSitad2t6QWFPTW9lc3Ba?=
 =?utf-8?B?L0pKTDkxZ2FjWjZTd1g4TUw2dnY1elVEU0pVdTJDckYyZmxBZEo2QW1aNDdV?=
 =?utf-8?B?ZG1XQ3F6SnoxNlg5Y09QUWluUEw4bHpWZXYyV2kvU1NnYUdkYnd6MDdmd0Q0?=
 =?utf-8?B?eUxpclJ4QXVXdHpRUXo1cXU2Sno4NDRDL2tRS1pGTCtxdWhtd3lTbEFreVJ1?=
 =?utf-8?B?UzZRbjY5RTB0RWIyaEZSSkpCL1NpR0pEdzJCME1EVHFZK2U1NDgxaW41MDBr?=
 =?utf-8?B?VHBqdzFHQjhQb2ZMaFdWQmwvU2dSTkdBYTZna3NWaTZrckFjZ1diRHpZYXJW?=
 =?utf-8?B?M3h4NTI0T1JGSGE2eG11WGtKSlpYMm1pamZjUGs3Tzg1aDZjNUNNWW55RCsr?=
 =?utf-8?B?d210bHZZOE92b2tJbThLSzU3THVzS09WTUlYQW91dkNjcjFEaFN4bUdPQVBV?=
 =?utf-8?B?UytlRi9YNld1VlBXcnpHZm16eHc0Qjk3RzNibHR5ajI4bStwWElOSjB5Q0Fh?=
 =?utf-8?B?YXFRTUxjTGUvNm9kalFwajRLL1phbmZzMVVPNHFsY0ZrYjZxcEhsRFdwQmkx?=
 =?utf-8?B?OFJKSFc5MEZxVVhucEd1WlVBbkp1OU1CYy9uejhDY003Ymp5WTdyaE9zemtv?=
 =?utf-8?B?N056UlNRZlZHRnNJcnh1a1d0WElGYjlSeEQ3djBOZE1vV1dETjQ3OG5aYnA0?=
 =?utf-8?B?azhHSkdhSENJZ2ppTnRNSFdCeDRGckh2SUg3VkY2K1VESkZkbGIrOGdHS2Qv?=
 =?utf-8?Q?yQauzGVL4ggD98lZmViwOwXn9R9TDtFxNgTsI/F?=
X-OriginatorOrg: amd.com
X-MS-Exchange-CrossTenant-Network-Message-Id: 5a93b627-b420-487d-f228-08d97e59d945
X-MS-Exchange-CrossTenant-AuthSource: MN2PR12MB3775.namprd12.prod.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Internal
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 23 Sep 2021 06:17:41.3773
 (UTC)
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-CrossTenant-Id: 3dd8961f-e488-4e60-8e11-a82d994e183d
X-MS-Exchange-CrossTenant-MailboxType: HOSTED
X-MS-Exchange-CrossTenant-UserPrincipalName: i8ka79Yy4qjAzt5uimpZkB9KaQQt1SH7Mxl/Xj6gWKZyAoXfTP89tFfbNfWG3lyg
X-MS-Exchange-Transport-CrossTenantHeadersStamped: MN2PR12MB4375
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

Am 23.09.21 um 02:58 schrieb Dave Airlie:
> On Sat, 18 Sept 2021 at 07:57, Alexandre Bailon <abailon@baylibre.com> wrote:
>> Some Mediatek SoC provides hardware accelerator for AI / ML.
>> This driver provides the infrastructure to manage memory
>> shared between host CPU and the accelerator, and to submit
>> jobs to the accelerator.
>> The APU itself is managed by remoteproc so this drivers
>> relies on remoteproc to found the APU and get some important data
>> from it. But, the driver is quite generic and it should possible
>> to manage accelerator using another ways.
>> This driver doesn't manage itself the data transmitions.
>> It must be registered by another driver implementing the transmitions.
>>
>> Signed-off-by: Alexandre Bailon <abailon@baylibre.com>
>> [SNIP]

>> Please refer to
>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.kernel.org%2Fdoc%2FDocumentation%2Fioctl%2Fbotching-up-ioctls.rst&amp;data=04%7C01%7Cchristian.koenig%40amd.com%7C53a0ef2630404ddc4d9408d97e2d409c%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637679555123878415%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&amp;sdata=6oVXAAOjQX%2FnDzJxZIAALqjDourHdrdGF6QVQKR58KI%3D&amp;reserved=0
>>
>> here and below in many places.
>>
>> There's a lot of missing padding/alignment here.

There is also the pahole utility which show you nicely where you need 
padding for your IOCTL structures.

For example "pahole drivers/gpu/drm/amd/amdgpu/amdgpu.ko -C 
drm_amdgpu_gem_va" gives you:

struct drm_amdgpu_gem_va {
     __u32                      handle;               /*     0     4 */
     __u32                      _pad;                 /*     4     4 */
     __u32                      operation;            /*     8     4 */
     __u32                      flags;                /*    12     4 */
     __u64                      va_address;           /*    16     8 */
     __u64                      offset_in_bo;         /*    24     8 */
     __u64                      map_size;             /*    32     8 */

     /* size: 40, cachelines: 1, members: 7 */
     /* last cacheline: 40 bytes */
};

And as you can see we have added the _pad field to our IOCTL parameter 
structure to properly align the 64bit members.

Regards,
Christian.

>>
>> I'm trying to find the time to review this stack in full, any writeups
>> on how this is used from userspace would be useful (not just the code
>> repo, but some sort of how do I get at it) it reads as kinda generic
>> (calling it apu), but then has some specifics around device binding.
>>
>> Dave.