[FFmpeg-devel] [FFmpeg-cvslog] kmsgrab: Add more DRM plane formats

Carl Eugen Hoyos ceffmpeg at gmail.com
Wed Sep 20 01:59:23 EEST 2017


2017-09-20 0:42 GMT+02:00 Mark Thompson <sw at jkqxz.net>:
> On 19/09/17 23:23, Carl Eugen Hoyos wrote:
>> 2017-09-20 0:16 GMT+02:00 Carl Eugen Hoyos <git at videolan.org>:
>>> ffmpeg | branch: master | Carl Eugen Hoyos <ceffmpeg at gmail.com> | Fri Sep 15 21:24:48 2017 +0100| [f952edaa73ee8618fcc8c105b57b9032ca0d1cec] | committer: Mark Thompson
>>
>>> +    { AV_PIX_FMT_RGB8,     DRM_FORMAT_RGB332   },
>>
>> Were you able to test this?
>> The definition looks different...
>
> No, I don't have anything with this format, so I was just looking at the
> docs for those.  From your set, I was able to test RGB/BGR 565 LE only.
>
> pixfmt.h:
>   AV_PIX_FMT_BGR8,  ///< packed RGB 3:3:2,  8bpp, (msb)2B 3G 3R(lsb)
>   AV_PIX_FMT_RGB8,  ///< packed RGB 3:3:2,  8bpp, (msb)2R 3G 3B(lsb)

This is consistent with libswscale.

> drm_fourcc.h:
> #define DRM_FORMAT_BGR233       fourcc_code('B', 'G', 'R', '8') /*
> [7:0] B:G:R 2:3:3 */
> #define DRM_FORMAT_RGB332       fourcc_code('R', 'G', 'B', '8') /*
> [7:0] R:G:B 3:3:2 */
>
> Argh.  So I missed that the bits have the same pattern in ffmpeg for
> the two formats vs. the same component-allocation in libdrm.
>
> Do you want to remove that one?  Consider it pre-approved (keeps
> me away from it to avoid messing up any further on this...).

I reverted this (hoping that the drm documentation is correct).

Thank you, Carl Eugen


More information about the ffmpeg-devel mailing list