On Tue, Jan 03, 2012 at 04:24:27PM -0800, Jason Garrett-Glaser wrote:
> On Tue, Jan 3, 2012 at 4:22 PM, Diego Biurrun <[email protected]> wrote:
> > On Tue, Jan 03, 2012 at 03:38:48PM -0800, Jason Garrett-Glaser wrote:
> >> On Tue, Jan 3, 2012 at 2:28 PM, Diego Biurrun <[email protected]> wrote:
> >> > ---
> >> >  libavcodec/cabac.h      |   47 
> >> > -----------------------------------------------
> >> >  libavcodec/h264_cabac.c |   47 
> >> > +++++++++++++++++++++++++++++++++++++++++++++++
> >>
> >> What is the purpose of cabac.h then?  In x264, we separate the
> >> implementation of CABAC (cabac.h) from the binarization process
> >> (h264_cabac.c), primarily just for cleanliness and abstraction.  Is
> >> there any reason we shouldn't continue to do this here?
> >
> > Applying the whole set fixes a series of build failures with
> > --disable-optimizations where some x86 cabac code was appearing
> > in places it was unnecessary and not getting optimized away...
> 
> Why does cabac.h (platform-independent CABAC implementation) have
> anything to do with problems in x86 code?

libavcodec/cabac.h includes libavcodec/x86/cabac.h and uses the
optimized functions from there.

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

Reply via email to