[FFmpeg-cvslog] vp9: remove another optimization branch in iadst16 which causes overflows.

Ronald S. Bultje git at videolan.org
Sat May 16 00:16:09 CEST 2015


ffmpeg | branch: release/2.6 | Ronald S. Bultje <rsbultje at gmail.com> | Wed Apr 22 15:48:37 2015 -0400| [b53b6a25e232f6ad5f3d09eb2f06dce8144c1c58] | committer: Michael Niedermayer

vp9: remove another optimization branch in iadst16 which causes overflows.

See sample vp90-2-14-resize-fp-tiles-16-8.webm from the vp9 test vector
set to reproduce the issue.
(cherry picked from commit 3de13d5212b6a0ff146dc0b50e34ee7dbc822eef)

Signed-off-by: Michael Niedermayer <michaelni at gmx.at>

> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=b53b6a25e232f6ad5f3d09eb2f06dce8144c1c58
---

 libavcodec/x86/vp9itxfm.asm |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/libavcodec/x86/vp9itxfm.asm b/libavcodec/x86/vp9itxfm.asm
index a9d45a4..9cf0d78 100644
--- a/libavcodec/x86/vp9itxfm.asm
+++ b/libavcodec/x86/vp9itxfm.asm
@@ -1622,7 +1622,9 @@ VP9_IDCT_IDCT_16x16_ADD_XMM avx
     PSIGNW                  m3, [pw_m1]                     ; m3=out1[w], m7=t10[w]
     SUMSUB_BA                w,  2,  6,  1                  ; m2=out14[w], m6=t11[w]
 
-%if cpuflag(ssse3)
+    ; unfortunately, the code below overflows in some cases, e.g.
+    ; http://downloads.webmproject.org/test_data/libvpx/vp90-2-14-resize-fp-tiles-16-8.webm
+%if 0; cpuflag(ssse3)
     SUMSUB_BA                w,  7,  6,  1
     pmulhrsw                m7, [pw_11585x2]                ; m7=out6[w]
     pmulhrsw                m6, [pw_11585x2]                ; m6=out9[w]



More information about the ffmpeg-cvslog mailing list