conv_acc: Our code requires SSSE3, not just SSE3
The accelerated convolutional decoder uses SSSE3 instructions such
as PSIGNW (via _mm_sign_epi16) which go beyond what SSE3 offers. So
let's make sure we use the right compiler flag (-mssse3) and also the
right runtime check.
Without this patch, we would use illegal instructions e.g. on Opteron
Gen3 such as Opteron 2427, which are also used as build.opensuse.org
build hosts (build31 through build36) where we wouldn't pass "make
check" as a result.
Change-Id: I2754164384109f2821fd98ffb48f625893f2923d
Fixes: OS#2386
diff --git a/m4/ax_check_simd.m4 b/m4/ax_check_simd.m4
index 8a0ceb7..daca2be 100644
--- a/m4/ax_check_simd.m4
+++ b/m4/ax_check_simd.m4
@@ -19,7 +19,7 @@
#
# And defines:
#
-# HAVE_AVX3 / HAVE_SSE3 / HAVE_SSE4.1
+# HAVE_AVX3 / HAVE_SSSE3 / HAVE_SSE4.1
#
# LICENSE
#
@@ -42,7 +42,7 @@
AC_REQUIRE([AC_CANONICAL_HOST])
AM_CONDITIONAL(HAVE_AVX2, false)
- AM_CONDITIONAL(HAVE_SSE3, false)
+ AM_CONDITIONAL(HAVE_SSSE3, false)
AM_CONDITIONAL(HAVE_SSE4_1, false)
case $host_cpu in
@@ -57,14 +57,14 @@
AC_MSG_WARN([Your compiler does not support AVX2 instructions])
fi
- AX_CHECK_COMPILE_FLAG(-msse3, ax_cv_support_sse3_ext=yes, [])
- if test x"$ax_cv_support_sse3_ext" = x"yes"; then
- SIMD_FLAGS="$SIMD_FLAGS -msse3"
- AC_DEFINE(HAVE_SSE3,,
- [Support SSE3 (Streaming SIMD Extensions 3) instructions])
- AM_CONDITIONAL(HAVE_SSE3, true)
+ AX_CHECK_COMPILE_FLAG(-mssse3, ax_cv_support_ssse3_ext=yes, [])
+ if test x"$ax_cv_support_ssse3_ext" = x"yes"; then
+ SIMD_FLAGS="$SIMD_FLAGS -mssse3"
+ AC_DEFINE(HAVE_SSSE3,,
+ [Support SSSE3 (Supplemental Streaming SIMD Extensions 3) instructions])
+ AM_CONDITIONAL(HAVE_SSSE3, true)
else
- AC_MSG_WARN([Your compiler does not support SSE3 instructions])
+ AC_MSG_WARN([Your compiler does not support SSSE3 instructions])
fi
AX_CHECK_COMPILE_FLAG(-msse4.1, ax_cv_support_sse41_ext=yes, [])