[FFmpeg-devel] H.264 Regressions for x86_32/gcc 2.95.3

Loren Merritt lorenm
Fri Feb 8 06:36:20 CET 2008


On Thu, 7 Feb 2008, Mike Melanson wrote:

> A build of SVN 11876 using gcc 2.95.3 on x86_32 passes the current FATE
> test suite. Either SVN 11877 or 11878 broke 31 of the current H.264
> tests (though it should be noted that there are still a lot of H.264
> tests that pass on the same build). The tests all seem to segfault when
> they fail.
>
> Build 11876/gcc 2.95.3/x86_32:
> http://fate.multimedia.cx/index.php?build_record=3132
>
> Build 11878/gcc 2.95.3/x86_32:
> http://fate.multimedia.cx/index.php?build_record=3148
>
> 11877's log message was, "sse2 h264 motion compensation. not new code,
> just separate out the cases that didn't need ssse3."

Doesn't crash for me (but I don't have gcc 2.95), so I have to ask for all 
the normal bugreport details.
Also, do your cpu and assembler support ssse3? If not, it might be one of 
the previous few commits that broke, and you just weren't using that code.

If I had to guess, it might be the notorious non-support of aligned 
variables.

--Loren Merritt




More information about the ffmpeg-devel mailing list