[MPlayer-dev-eng] [PATCH] VIA Unichrome support

[Michael Niedermayer]

Hi

On Saturday 07 May 2005 22:27, Ivor Hewitt wrote:
> Hi,
>
> Attached is the patch to add support for the open source driver for
> unichrome and unichrome-pro chipsets to mplayer updated for current CVS.
>
> As mentioned before this extends the xvmc interface and so requires a
> configure of:-
> --enable-xxmc
>
> and a usage line of:-
> mplayer -vo xvmc -vc ffmpeg12mc

anoter quick review (note iam not the xvmc maintainer, iam just reviewing it 
so that he doesnt miss any issues ...)


the libavcodec part should be split and send to ffmpeg-devel


> @@ -231,6 +231,7 @@ enum PixelFormat {
>      PIX_FMT_UYVY422,   ///< Packed pixel, Cb Y0 Cr Y1 
>      PIX_FMT_UYVY411,   ///< Packed pixel, Cb Y0 Y1 Cr Y2 Y3
>      PIX_FMT_NB,
> +    PIX_FMT_XVMC_MPEG2_VLD,
>  };

PIX_FMT_NB must be last


> -#ifdef HAVE_XVMC
> +#if defined(HAVE_XVMC)|defined(HAVE_XVMC_VLD)

if HAVE_XVMC_VLD is defined then HAVE_XVMC should be too


> +#ifdef HAVE_XVMC_VLD
> +extern int XVMC_decode_slice(MpegEncContext *s, int start_code,
> +                                 uint8_t *buffer, int buf_size);
> +#endif

IMHO we should avoid such uneeded #ifdefs


> +static int length_to_next_start(uint8_t* pbuf_ptr, int buf_size)

isnt that the same as find_start_code()?


[...]
-- 
Michael

"nothing is evil in the beginning. Even Sauron was not so." -- Elrond