diff options
| author | Matthias Räncker <theonetruecamper@gmx.de> | 2018-09-20 20:20:39 +0200 |
|---|---|---|
| committer | Matthias Räncker <theonetruecamper@gmx.de> | 2018-09-20 22:37:42 +0200 |
| commit | a93705f7f9c15cdc2a1e62f6142e99f794923826 (patch) | |
| tree | 8cd6bfae1ce7c998e8c733727bb1701d2fc95b7e /vpx_dsp/x86/variance_sse2.c | |
| parent | 0aa83d61a18fbdd5921247e0401b0fbba443cf35 (diff) | |
| download | libvpx-a93705f7f9c15cdc2a1e62f6142e99f794923826.tar libvpx-a93705f7f9c15cdc2a1e62f6142e99f794923826.tar.gz libvpx-a93705f7f9c15cdc2a1e62f6142e99f794923826.tar.bz2 libvpx-a93705f7f9c15cdc2a1e62f6142e99f794923826.zip | |
sanitizer: fix unaligned load/stores
When built with -fsanitizer=address,undefined a number of tests,
such as ByteAlignmentTest.SwitchByteAlignment or
ByteAlignmentTest.SwitchByteAlignment produce runtime errors about
unaligned 4-byte loads/stores. While normally not really a problem,
this does technically violate the language and it is eays to fix in
a standard conforming way using memcpy which does not produce
inferior code.
Signed-off-by: Matthias Räncker <theonetruecamper@gmx.de>
Change-Id: Ie1e97ab25fe874f864df48b473569f00563181ae
Diffstat (limited to 'vpx_dsp/x86/variance_sse2.c')
| -rw-r--r-- | vpx_dsp/x86/variance_sse2.c | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/vpx_dsp/x86/variance_sse2.c b/vpx_dsp/x86/variance_sse2.c index a2a13a68b..027905255 100644 --- a/vpx_dsp/x86/variance_sse2.c +++ b/vpx_dsp/x86/variance_sse2.c @@ -14,6 +14,7 @@ #include "./vpx_config.h" #include "./vpx_dsp_rtcd.h" #include "vpx_ports/mem.h" +#include "vpx_dsp/x86/mem_sse2.h" static INLINE unsigned int add32x4_sse2(__m128i val) { val = _mm_add_epi32(val, _mm_srli_si128(val, 8)); @@ -35,8 +36,8 @@ unsigned int vpx_get_mb_ss_sse2(const int16_t *src) { } static INLINE __m128i load4x2_sse2(const uint8_t *const p, const int stride) { - const __m128i p0 = _mm_cvtsi32_si128(*(const uint32_t *)(p + 0 * stride)); - const __m128i p1 = _mm_cvtsi32_si128(*(const uint32_t *)(p + 1 * stride)); + const __m128i p0 = _mm_cvtsi32_si128(loadu_uint32(p + 0 * stride)); + const __m128i p1 = _mm_cvtsi32_si128(loadu_uint32(p + 1 * stride)); const __m128i p01 = _mm_unpacklo_epi32(p0, p1); return _mm_unpacklo_epi8(p01, _mm_setzero_si128()); } |