Commit Graph

3 Commits

Author SHA1 Message Date
Måns Rullgård bdd19e29df Mark all intreadwrite functions av_always_inline
Originally committed as revision 21278 to svn://svn.ffmpeg.org/ffmpeg/trunk
2010-01-18 01:35:19 +00:00
Måns Rullgård e6956a6e48 ARM: first value loaded in AV_RN64 needs to be early-clobber
Originally committed as revision 19656 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-08-16 15:51:50 +00:00
Måns Rullgård 3c55ce039d ARM asm for AV_RN*()
ARMv6 and later support unaligned loads and stores for single
word/halfword but not double/multiple.  GCC is ignorant of this and
will always use bytewise accesses for unaligned data.  Casting to an
int32_t pointer is dangerous since a load/store double or multiple
instruction might be used (this happens with some code in FFmpeg).
Implementing the AV_[RW]* macros with inline asm using only supported
instructions gives fast and safe unaligned accesses.  ARM RVCT does
the right thing with generic code.

This gives an overall speedup of up to 10%.

Originally committed as revision 18601 to svn://svn.ffmpeg.org/ffmpeg/trunk
2009-04-18 00:00:28 +00:00