[FFmpeg-devel] [PATCH] x86/intmath: disable sse av_clip functions when using ICC

James Almer jamrial at gmail.com
Thu Jan 21 04:08:58 CET 2016


It seems to miscompile them

Should fix fate-ra-288 and fate-twinvq

Signed-off-by: James Almer <jamrial at gmail.com>
---
The early clobber fix didn't do the trick.
Someone with ICC should to give it a look, but until then, disabling these
functions will solve the failures.

 libavutil/x86/intmath.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/libavutil/x86/intmath.h b/libavutil/x86/intmath.h
index 3e279df..f58b0d0 100644
--- a/libavutil/x86/intmath.h
+++ b/libavutil/x86/intmath.h
@@ -99,7 +99,7 @@ static av_always_inline av_const unsigned av_mod_uintp2_bmi2(unsigned a, unsigne
 
 #endif /* __BMI2__ */
 
-#if defined(__SSE2__)
+#if defined(__SSE2__) && !defined(__INTEL_COMPILER)
 
 #define av_clipd av_clipd_sse2
 static av_always_inline av_const double av_clipd_sse2(double a, double amin, double amax)
@@ -115,7 +115,7 @@ static av_always_inline av_const double av_clipd_sse2(double a, double amin, dou
 
 #endif /* __SSE2__ */
 
-#if defined(__SSE__)
+#if defined(__SSE__) && !defined(__INTEL_COMPILER)
 
 #define av_clipf av_clipf_sse
 static av_always_inline av_const float av_clipf_sse(float a, float amin, float amax)
-- 
2.7.0



More information about the ffmpeg-devel mailing list