RepoMirrors/ffmpeg
1
0
Fork 0
mirror of https://git.ffmpeg.org/ffmpeg.git synced 2024-12-18 05:24:42 +00:00
Code Issues Projects Releases Wiki Activity

Attempt to fix asm compilation failure.

Browse Source 
Only tested on gcc 4 & x86_64.

Originally committed as revision 21355 to svn://svn.ffmpeg.org/ffmpeg/trunk
This commit is contained in:
Michael Niedermayer 2010-01-20 19:23:19 +00:00
parent 8d821b0181
commit b4c2ada528
1 changed files with 20 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

41
libavcodec/x86/h264dsp_mmx.c
View File

@ -812,7 +812,7 @@ static void h264_loop_filter_strength_mmx2( int16_t bS[2][4][4], uint8_t nnz[40]
// could do a special case for dir==0 && edges==1, but it only reduces the
// average filter time by 1.2%
for( dir=1; dir>=0; dir-- ) {
const int d_idx = dir ? -8 : -1;
const x86_reg d_idx = dir ? -8 : -1;
const int mask_mv = dir ? mask_mv1 : mask_mv0;
DECLARE_ALIGNED_8(const uint64_t, mask_dir) = dir ? 0 : 0xffffffffffffffffULL;
int b_idx, edge, l;
@ -825,54 +825,53 @@ static void h264_loop_filter_strength_mmx2( int16_t bS[2][4][4], uint8_t nnz[40]
__asm__ volatile("pxor %%mm0, %%mm0 \n\t":);
for( l = bidir; l >= 0; l-- ) {
__asm__ volatile(
"movd %0, %%mm1 \n\t"
"punpckldq %1, %%mm1 \n\t"
"movd (%0), %%mm1 \n\t"
"punpckldq (%0,%1), %%mm1 \n\t"
"punpckldq %%mm1, %%mm2 \n\t"
"pcmpeqb %%mm2, %%mm1 \n\t"
"paddb %%mm6, %%mm1 \n\t"
"punpckhbw %%mm7, %%mm1 \n\t" // ref[b] != ref[bn]
"por %%mm1, %%mm0 \n\t"
"movq %2, %%mm1 \n\t"
"movq 8+1*%2, %%mm2 \n\t"
"psubw %3, %%mm1 \n\t"
"psubw 8+1*%3, %%mm2 \n\t"
"movq (%2), %%mm1 \n\t"
"movq 8(%2), %%mm2 \n\t"
"psubw (%2,%1,4), %%mm1 \n\t"
"psubw 8(%2,%1,4), %%mm2 \n\t"
"packsswb %%mm2, %%mm1 \n\t"
"paddb %%mm5, %%mm1 \n\t"
"pminub %%mm4, %%mm1 \n\t"
"pcmpeqb %%mm4, %%mm1 \n\t" // abs(mv[b] - mv[bn]) >= limit
"por %%mm1, %%mm0 \n\t"
::"m"(ref[l][b_idx]),
"m"(ref[l][b_idx+d_idx]),
"m"(mv[l][b_idx][0]),
"m"(mv[l][b_idx+d_idx][0])
::"r"(ref[l]+b_idx),
"r"(d_idx),
"r"(mv[l]+b_idx)
);
}
if(bidir==1){
__asm__ volatile("pxor %%mm3, %%mm3 \n\t":);
for( l = bidir; l >= 0; l-- ) {
__asm__ volatile(
"movd %0, %%mm1 \n\t"
"punpckldq %1, %%mm1 \n\t"
"movd (%0), %%mm1 \n\t"
"punpckldq (%1), %%mm1 \n\t"
"punpckldq %%mm1, %%mm2 \n\t"
"pcmpeqb %%mm2, %%mm1 \n\t"
"paddb %%mm6, %%mm1 \n\t"
"punpckhbw %%mm7, %%mm1 \n\t" // ref[b] != ref[bn]
"por %%mm1, %%mm3 \n\t"
"movq %2, %%mm1 \n\t"
"movq 8+1*%2, %%mm2 \n\t"
"psubw %3, %%mm1 \n\t"
"psubw 8+1*%3, %%mm2 \n\t"
"movq (%2), %%mm1 \n\t"
"movq 8(%2), %%mm2 \n\t"
"psubw (%3), %%mm1 \n\t"
"psubw 8(%3), %%mm2 \n\t"
"packsswb %%mm2, %%mm1 \n\t"
"paddb %%mm5, %%mm1 \n\t"
"pminub %%mm4, %%mm1 \n\t"
"pcmpeqb %%mm4, %%mm1 \n\t" // abs(mv[b] - mv[bn]) >= limit
"por %%mm1, %%mm3 \n\t"
::"m"(ref[l][b_idx]),
"m"(ref[1-l][b_idx+d_idx]),
"m"(mv[l][b_idx][0]),
"m"(mv[1-l][b_idx+d_idx][0])
::"r"(ref[l]+b_idx),
"r"(ref[1-l]+b_idx+d_idx),
"r"(mv[l][b_idx]),
"r"(mv[1-l][b_idx+d_idx])
);
}
__asm__ volatile(