[MPlayer-dev-eng] [PATCH] vda support for mplayer

Xidorn Quan quanxunzhen at gmail.com
Fri Aug 17 12:20:24 CEST 2012


On Fri, Aug 17, 2012 at 5:54 PM, Zongyao Qu <zongyao.qu at gmail.com> wrote:

> Just some technical suggestion.
>
> > Attachment (h264_vda_mplayer_new.patch): application/octet-stream, 13 KiB
> >
>
> +#define IMGFMT_UNWRAP_VDA(fmt) \
> +    { if (IMGFMT_IS_VDA(fmt)) (fmt) = IMGFMT_YUY2; }
>
> This is not a good use of macro, since we can't make sure fmt is a l-value.
>

I'll use

> (IMGFMT_IS_VDA(fmt) ? IMGFMT_IN_VDA : (fmt))
>
instead.


> +static uint32_t get_image(mp_image_t *mpi)
> +{
> +       if (!IMGFMT_IS_VDA(image_format))
> +               return VO_NOTIMPL;
> +
> +       mpi->flags |= MP_IMGFLAG_DIRECT;
> +       mpi->stride[0] = 0;
> +       mpi->stride[1] = 0;
> +       mpi->stride[2] = 0;
> +       mpi->stride[3] = 0;
> +       mpi->planes[0] = 1;
> +       mpi->planes[1] = 0;
> +       mpi->planes[2] = 0;
> +       mpi->planes[3] = 0;
> +       return VO_TRUE;
> +}
> +
>
> It seams kinds of DR support in vo:corevideo, I suggest divide the patch to
>  DR support and
> vda support separately.
>
>
> I guess it would be broken if there is any non-dr filter in the filter
> chain.
>

In fact, it seems that all filter will break the DR in corevideo since the
image_format
would be YUY2 instead of VDA_VLD if there is any filter doesn't fully
support VDA_VLD,
and currently no filter supports it.


More information about the MPlayer-dev-eng mailing list