ffmpeg | branch: master | James Almer <[email protected]> | Wed Jan 2 20:48:59 2019 -0300| [9b5bd665e105894919cdcfa0ed9818919538e5f6] | committer: James Almer
x86/af_afir: fix processing the last element ff_fcmul_add_sse3() is now identical to the C version. Reviewed-by: Paul B Mahol <[email protected]> Signed-off-by: James Almer <[email protected]> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=9b5bd665e105894919cdcfa0ed9818919538e5f6 --- libavfilter/x86/af_afir.asm | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/libavfilter/x86/af_afir.asm b/libavfilter/x86/af_afir.asm index 849d85e70f..fcc1f426db 100644 --- a/libavfilter/x86/af_afir.asm +++ b/libavfilter/x86/af_afir.asm @@ -30,7 +30,6 @@ SECTION .text INIT_XMM sse3 cglobal fcmul_add, 4,4,6, sum, t, c, len shl lend, 3 - add lend, mmsize*2 add tq, lenq add cq, lenq add sumq, lenq @@ -57,4 +56,8 @@ ALIGN 16 movaps [sumq + lenq+mmsize], m3 add lenq, mmsize*2 jl .loop - REP_RET + movss xm0, [tq + lenq] + mulss xm0, [cq + lenq] + addss xm0, [sumq + lenq] + movss [sumq + lenq], xm0 + RET _______________________________________________ ffmpeg-cvslog mailing list [email protected] http://ffmpeg.org/mailman/listinfo/ffmpeg-cvslog
