Port x264 deblocking code to libavcodec.  This includes SSE2 luma deblocking code and both MMXEXT and SSE2 luma intra deblocking code for H.264 decoding.  This assembly is available under --enable-gpl and speeds decoding of Cathedral by 7%.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16239 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16238 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16237 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

FRExt/HPCAMOLQ_BRCM_B
FRExt/HPCAQ2LQ_BRCM_B
FRExt/HPCVMOLQ_BRCM_B


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16236 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16235 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Add automatic prefix handling to yasm functions.  Does nothing now, but will be useful for porting x264 asm in the future.

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16234 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

3 cpu cycles speed up on pentium dual.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16233 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

no speed change


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16232 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

no speed change


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16231 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Fixes valgrind uninitialized value warnings at the end of decoding H.264
frames.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16230 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

Fixes valgrind invalid read warnings during H.264 extradata parsing.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16229 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16228 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16227 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

The case for 16x16 blocks becomes 10 cpu cycles faster on pentium dual,
i could not find a speed difference in the case of subblocks though.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16226 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16225 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16224 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16223 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16222 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

check.
14 cpu cycles speedup


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16221 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16220 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

decode_residual is about 3.3% faster.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16219 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

No speed change meassureable with START/STOP_TIMER, but this is needed
for future optimizations.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16218 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16217 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16216 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16215 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16214 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16213 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16212 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16211 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16210 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16209 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16208 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

cathedral +0.5% speed
aladin +0.6% speed [note aladin has been cat-ed 10 times to reduce the influence
of init time]
Speedup also verified via START/STOP_TIMER (difference was very significant
for the changed parts)


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16207 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

thus avoiding the calling overhead.
New functions are not yet used.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16206 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

give them a ff_ prefix. See "[PATCH] asf.c: make functions of interest for
MS-RTSP public" thread on ML.



git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16205 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

RTSP stack for RTSP-MS support. This way, they can read input from any given
input source. See discussion in "[PATCH] add gb argument to ASF functions of
interest to MS-RTSP" ML thread.



git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16204 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

0.7% speedup for the cathedral sample.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16203 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16202 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

0.5% overall speedup for the cathedral sample.


git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16201 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b

git-svn-id: file:///var/local/repositories/ffmpeg/trunk@16200 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b