mirror of https://git.ffmpeg.org/ffmpeg.git
x86/vf_transpose: make ff_transpose_8x8_16_sse2 work on x86_32
Reviewed-by: Paul B Mahol <onemda@gmail.com> Signed-off-by: James Almer <jamrial@gmail.com>
This commit is contained in:
parent
2a75006ddc
commit
738bc3e742
|
@ -56,10 +56,7 @@ cglobal transpose_8x8_8, 4,5,8, src, src_linesize, dst, dst_linesize, linesize3
|
|||
movq [dstq + linesize3q], m7
|
||||
RET
|
||||
|
||||
%if ARCH_X86_64
|
||||
|
||||
INIT_XMM sse2
|
||||
cglobal transpose_8x8_16, 4,5,9, src, src_linesize, dst, dst_linesize, linesize3
|
||||
cglobal transpose_8x8_16, 4,5,9, ARCH_X86_32 * 32, src, src_linesize, dst, dst_linesize, linesize3
|
||||
lea linesize3q, [src_linesizeq * 3]
|
||||
movu m0, [srcq + src_linesizeq * 0]
|
||||
movu m1, [srcq + src_linesizeq * 1]
|
||||
|
@ -71,7 +68,11 @@ cglobal transpose_8x8_16, 4,5,9, src, src_linesize, dst, dst_linesize, linesize3
|
|||
movu m6, [srcq + src_linesizeq * 2]
|
||||
movu m7, [srcq + linesize3q]
|
||||
|
||||
%if ARCH_X86_64
|
||||
TRANSPOSE8x8W 0, 1, 2, 3, 4, 5, 6, 7, 8
|
||||
%else
|
||||
TRANSPOSE8x8W 0, 1, 2, 3, 4, 5, 6, 7, [rsp], [rsp + 16]
|
||||
%endif
|
||||
|
||||
lea linesize3q, [dst_linesizeq * 3]
|
||||
movu [dstq + dst_linesizeq * 0], m0
|
||||
|
@ -84,5 +85,3 @@ cglobal transpose_8x8_16, 4,5,9, src, src_linesize, dst, dst_linesize, linesize3
|
|||
movu [dstq + dst_linesizeq * 2], m6
|
||||
movu [dstq + linesize3q], m7
|
||||
RET
|
||||
|
||||
%endif
|
||||
|
|
|
@ -43,7 +43,7 @@ av_cold void ff_transpose_init_x86(TransVtable *v, int pixstep)
|
|||
v->transpose_8x8 = ff_transpose_8x8_8_sse2;
|
||||
}
|
||||
|
||||
if (ARCH_X86_64 && EXTERNAL_SSE2(cpu_flags) && pixstep == 2) {
|
||||
if (EXTERNAL_SSE2(cpu_flags) && pixstep == 2) {
|
||||
v->transpose_8x8 = ff_transpose_8x8_16_sse2;
|
||||
}
|
||||
}
|
||||
|
|
Loading…
Reference in New Issue