This allows for some simplifications like removing some outer loops
and gives much better performance for thread_count > number of idle CPUs.

Originally committed as revision 20211 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 16072 to svn://svn.ffmpeg.org/ffmpeg/trunk

Originally committed as revision 16051 to svn://svn.ffmpeg.org/ffmpeg/trunk