[FFmpeg-devel] [PATCH] Implement PAFF in H.264

[Andreas Öman]

Hi,

Jeff Downs wrote:
> It does not interfere in the sense that the smp code cannot be used.  I 
> believe it to work with it, and I did try it with -threads x, but I only 
> have uniprocessor systems to test with; it could use exercising on 
> multiprocessor/core systems.

If you just want to test that there are no incorrect inter-slice
dependencies you can test with the attached patch and

no "./configure --enable-pthreads" but with "-threads x"

to use the default avctx->execute() (and with this patch
decode the slices in reverse order)

I used it a few times during my work get repeatable errors.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: reverse.patch
Type: text/x-patch
Size: 461 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20070919/04a72ed0/attachment.bin>