[FFmpeg-devel] [PATCH 1/4] x86: xvid: port SSE2 idct to yasm

James Almer jamrial at gmail.com
Wed Mar 11 03:46:36 CET 2015


On 10/03/15 8:11 PM, Christophe Gisquet wrote:
> The main difference consists in renaming properly labels, and
> letting yasm select the gprs for skipping 1D transforms.
> ---
>  libavcodec/x86/Makefile        |   2 +-
>  libavcodec/x86/xvididct.asm    | 379 ++++++++++++++++++++++++++++++++++++++
>  libavcodec/x86/xvididct_init.c |  18 +-
>  libavcodec/x86/xvididct_sse2.c | 406 -----------------------------------------
>  4 files changed, 395 insertions(+), 410 deletions(-)
>  create mode 100644 libavcodec/x86/xvididct.asm
>  delete mode 100644 libavcodec/x86/xvididct_sse2.c
> 
> diff --git a/libavcodec/x86/Makefile b/libavcodec/x86/Makefile
> index 6b9164a..276df44 100644
> --- a/libavcodec/x86/Makefile
> +++ b/libavcodec/x86/Makefile
> @@ -74,7 +74,7 @@ MMX-OBJS-$(CONFIG_IDCTDSP)             += x86/simple_idct.o
>  
>  # decoders/encoders
>  MMX-OBJS-$(CONFIG_MPEG4_DECODER)       += x86/xvididct_mmx.o            \
> -                                          x86/xvididct_sse2.o
> +                                          x86/xvididct.o

Should be YASM-OBJS, and moved to the end of the file.
Also, related to the build failure Michael mentioned for the second patch, this is 
missing an inline -> external change in libavcodec/x86/dct-test.c

As you said the patch is pretty much straightforward, so if fate passes and the output 
is the same then it should be ok.


More information about the ffmpeg-devel mailing list