Hi,

On Tue, Jul 31, 2012 at 9:12 PM, Luca Barbato <[email protected]> wrote:
> @@ -2879,10 +2879,9 @@ static void dsputil_init_sse2(DSPContext *c, 
> AVCodecContext *avctx,
>                                int mm_flags)
>  {
>      const int bit_depth      = avctx->bits_per_raw_sample;
> -
> -#if HAVE_INLINE_ASM
>      const int high_bit_depth = bit_depth > 8;
>
> +#if HAVE_INLINE_ASM
>      if (!(mm_flags & AV_CPU_FLAG_SSE2SLOW)) {
>          // these functions are slower than mmx on AMD, but faster on Intel
>          if (!high_bit_depth) {
> @@ -2893,7 +2892,9 @@ static void dsputil_init_sse2(DSPContext *c, 
> AVCodecContext *avctx,
>                  H264_QPEL_FUNCS(0, 0, sse2);
>          }
>      }
> +#endif /* HAVE_INLINE_ASM */

The H264_QPEL_FUNCS() call should go outside HAVE_INLINE_ASM.

> +;*****************************************************************************
> +;* MMX/SSE2/SSSE3-optimized H.264 qpel code
> +;*****************************************************************************
> +;* Copyright (C) 2012 Daniel Kang
> +;*
> +;* Authors: Daniel Kang <[email protected]>

You told me you based this substantially on the existing
implementation; if that's the case, you should acknowledge the
original author(s) of that code also.

I'll do a quick review pass of the assembly itself separately...

Ronald
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to