[FFmpeg-devel] Patch: Inline asm fixes for Intel compiler on Windows

Matt Oliver protogonoi at gmail.com
Sat Mar 22 06:24:27 CET 2014


>
>
>> applied most of it
>
>
> Ill see if I can find another way to get icl to compile while still using
> MANGLE but if memory serves I tried basically every combination I could
> think off and couldnt get it to compile any other way. Are there any issues
> with fixing it the way I did?
>

OK so I couldnt get icl to play nice with that 1 MANGLE line. The only way
ive been able to get it to work is to add round_tab[2] to the asm interface
instead of having it as a direct symbol reference. This should not be a
problem as this is how all the other sad_xxx functions use round_tab (see
x86/motion_est.c lines 351, 366, 408, 423 ). The patch is attached.

Also now that the named reference patch has been applied to ffmpeg then the
corresponding changes to libmpcodec can now be applied (patch 4/4 also
attached). libmpcodes is the only part that hasnt been updated yet with the
new changes (this may also be useful for the lto problem Carl was working
on).
Reimar Döffinger apparently applied the last set of patches (patches 1/4 ->
3/4) to upstream mplayer so if he or someone else can get this applied
upstream (assuming everything looks good of course).
Once that is done then the upstream libmpcodec can be merged into ffmpeg as
the first 3 fixes applied upstream over a month ago still havnt been synced.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Additional-icl-inline-asm-fix.patch
Type: application/octet-stream
Size: 1376 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20140322/955df37e/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 4-4-Fix-libmpcodecs-inline-asm-on-ICL-part-2.patch
Type: application/octet-stream
Size: 3378 bytes
Desc: not available
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20140322/955df37e/attachment-0001.obj>


More information about the ffmpeg-devel mailing list