[x264-devel] x86: Fix SIGILL in high bit-depth intra_sad_x3_4x4_sse2

[Henrik Gramner]

x264 | branch: master | Henrik Gramner <henrik at gramner.com> | Tue Jul  8 21:15:32 2014 +0200| [6eb483e4ca23f34a6a8fe09f3f2e9c9f192fd76b] | committer: Fiona Glaser

x86: Fix SIGILL in high bit-depth intra_sad_x3_4x4_sse2

An SSE3 instruction was used in an SSE2 function.

> http://git.videolan.org/gitweb.cgi/x264.git/?a=commit;h=6eb483e4ca23f34a6a8fe09f3f2e9c9f192fd76b
---

 common/x86/sad16-a.asm |    5 +++++
 1 file changed, 5 insertions(+)

diff --git a/common/x86/sad16-a.asm b/common/x86/sad16-a.asm
index 5b5e9c8..20028d1 100644
--- a/common/x86/sad16-a.asm
+++ b/common/x86/sad16-a.asm
@@ -546,7 +546,12 @@ SAD_X 4, 16,  8
 
 %macro INTRA_SAD_X3_4x4 0
 cglobal intra_sad_x3_4x4, 3,3,7
+%if cpuflag(ssse3)
     movddup   m0, [r1-1*FDEC_STRIDEB]
+%else
+    movq      m0, [r1-1*FDEC_STRIDEB]
+    punpcklqdq m0, m0
+%endif
     movq      m1, [r0+0*FENC_STRIDEB]
     movq      m2, [r0+2*FENC_STRIDEB]
     pshuflw   m6, m0, q1032