[FFmpeg-cvslog] avfilter/x86/vf_limiter: use movu, dst may not be always aligned

Paul B Mahol git at videolan.org
Thu Mar 24 10:42:35 EET 2022


ffmpeg | branch: master | Paul B Mahol <onemda at gmail.com> | Thu Mar 24 09:41:40 2022 +0100| [28d011516b72a2401b2289a4be34c6e8ab611a0b] | committer: Paul B Mahol

avfilter/x86/vf_limiter: use movu, dst may not be always aligned

Happens with pad filter after limiter.

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=28d011516b72a2401b2289a4be34c6e8ab611a0b
---

 libavfilter/x86/vf_limiter.asm | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/libavfilter/x86/vf_limiter.asm b/libavfilter/x86/vf_limiter.asm
index c5b9b0a64d..e1b80eba61 100644
--- a/libavfilter/x86/vf_limiter.asm
+++ b/libavfilter/x86/vf_limiter.asm
@@ -41,7 +41,7 @@ cglobal limiter_8bit, 6, 7, 3, src, dst, slinesize, dlinesize, w, h, x
     .loop:
         movu           m0, [srcq + xq]
         CLIPUB         m0, m1, m2
-        mova    [dstq+xq], m0
+        movu    [dstq+xq], m0
         add            xq, mmsize
     jl .loop
 
@@ -69,7 +69,7 @@ cglobal limiter_16bit, 6, 7, 3, src, dst, slinesize, dlinesize, w, h, x
         movu           m0, [srcq + xq]
         pmaxuw         m0, m1
         pminuw         m0, m2
-        mova    [dstq+xq], m0
+        movu    [dstq+xq], m0
         add            xq, mmsize
     jl .loop
 



More information about the ffmpeg-cvslog mailing list