diff options
| author | Michael Niedermayer <michael@niedermayer.cc> | 2015-07-17 23:20:14 +0200 |
|---|---|---|
| committer | Michael Niedermayer <michael@niedermayer.cc> | 2015-07-17 23:20:14 +0200 |
| commit | 52b6d96268987db14937fcf5f9a53ab9b4184adb (patch) | |
| tree | 7077ef6c23ddd678b9f1498e8ffdde3fbbc5dde4 /libavcodec/x86/bswapdsp.asm | |
| parent | c7c33ab721cfafbbe269c8f1505eb2d9a80d7af4 (diff) | |
| parent | a344e5d094ebcf9a23acf3a27c56cbbbc829db42 (diff) | |
| download | ffmpeg-52b6d96268987db14937fcf5f9a53ab9b4184adb.tar.gz | |
Merge commit 'a344e5d094ebcf9a23acf3a27c56cbbbc829db42'
* commit 'a344e5d094ebcf9a23acf3a27c56cbbbc829db42':
x86: bswapdsp: Don't treat 32-bit integers as 64-bit
Conflicts:
libavcodec/x86/bswapdsp.asm
Merged-by: Michael Niedermayer <michael@niedermayer.cc>
Diffstat (limited to 'libavcodec/x86/bswapdsp.asm')
| -rw-r--r-- | libavcodec/x86/bswapdsp.asm | 21 |
1 files changed, 10 insertions, 11 deletions
diff --git a/libavcodec/x86/bswapdsp.asm b/libavcodec/x86/bswapdsp.asm index ec060c93b6..b2d749cf19 100644 --- a/libavcodec/x86/bswapdsp.asm +++ b/libavcodec/x86/bswapdsp.asm @@ -32,8 +32,8 @@ SECTION_TEXT ; %1 = aligned/unaligned %macro BSWAP_LOOPS 1 - mov r3, r2 - sar r2, 3 + mov r3d, r2d + sar r2d, 3 jz .left4_%1 .loop8_%1: mov%1 m0, [r1 + 0] @@ -61,11 +61,11 @@ SECTION_TEXT %endif add r0, 32 add r1, 32 - dec r2 + dec r2d jnz .loop8_%1 .left4_%1: - mov r2, r3 - and r3, 4 + mov r2d, r3d + test r3d, 4 jz .left mov%1 m0, [r1] %if cpuflag(ssse3) @@ -95,7 +95,7 @@ cglobal bswap32_buf, 3,4,5 mov r3, r1 %endif or r3, r0 - and r3, 15 + test r3, 15 jz .start_align BSWAP_LOOPS u jmp .left @@ -103,8 +103,7 @@ cglobal bswap32_buf, 3,4,5 BSWAP_LOOPS a .left: %if cpuflag(ssse3) - mov r3, r2 - and r2, 2 + test r2d, 2 jz .left1 movq m0, [r1] pshufb m0, m2 @@ -112,13 +111,13 @@ cglobal bswap32_buf, 3,4,5 add r1, 8 add r0, 8 .left1: - and r3, 1 + test r2d, 1 jz .end mov r2d, [r1] bswap r2d mov [r0], r2d %else - and r2, 3 + and r2d, 3 jz .end .loop2: mov r3d, [r1] @@ -126,7 +125,7 @@ cglobal bswap32_buf, 3,4,5 mov [r0], r3d add r1, 4 add r0, 4 - dec r2 + dec r2d jnz .loop2 %endif .end: |