diff options
| author | Andrea Di Biagio <Andrea_DiBiagio@sn.scee.net> | 2015-10-20 11:19:54 +0000 |
|---|---|---|
| committer | Andrea Di Biagio <Andrea_DiBiagio@sn.scee.net> | 2015-10-20 11:19:54 +0000 |
| commit | 8bb12d0a77160aa38716f77f15ba6f968765686e (patch) | |
| tree | 235a524dcf2adabcb6e7669430a589ae77abfe57 /clang/lib | |
| parent | 505f792cbb67ec38083c8d7c7c1d78775a3c5ce8 (diff) | |
| download | bcm5719-llvm-8bb12d0a77160aa38716f77f15ba6f968765686e.tar.gz bcm5719-llvm-8bb12d0a77160aa38716f77f15ba6f968765686e.zip | |
[x86] Fix maskload/store intrinsic definitions in avxintrin.h
According to the Intel documentation, the mask operand of a maskload and
maskstore intrinsics is always a vector of packed integer/long integer values.
This patch introduces the following two changes:
1. It fixes the avx maskload/store intrinsic definitions in avxintrin.h.
2. It changes BuiltinsX86.def to match the correct gcc definitions for avx
maskload/store (see D13861 for more details).
Differential Revision: http://reviews.llvm.org/D13861
llvm-svn: 250816
Diffstat (limited to 'clang/lib')
| -rw-r--r-- | clang/lib/Headers/avxintrin.h | 32 |
1 files changed, 16 insertions, 16 deletions
diff --git a/clang/lib/Headers/avxintrin.h b/clang/lib/Headers/avxintrin.h index 7e4de9d443c..0041365edff 100644 --- a/clang/lib/Headers/avxintrin.h +++ b/clang/lib/Headers/avxintrin.h @@ -835,53 +835,53 @@ _mm256_storeu_si256(__m256i *__p, __m256i __a) /* Conditional load ops */ static __inline __m128d __DEFAULT_FN_ATTRS -_mm_maskload_pd(double const *__p, __m128d __m) +_mm_maskload_pd(double const *__p, __m128i __m) { - return (__m128d)__builtin_ia32_maskloadpd((const __v2df *)__p, (__v2df)__m); + return (__m128d)__builtin_ia32_maskloadpd((const __v2df *)__p, (__v2di)__m); } static __inline __m256d __DEFAULT_FN_ATTRS -_mm256_maskload_pd(double const *__p, __m256d __m) +_mm256_maskload_pd(double const *__p, __m256i __m) { return (__m256d)__builtin_ia32_maskloadpd256((const __v4df *)__p, - (__v4df)__m); + (__v4di)__m); } static __inline __m128 __DEFAULT_FN_ATTRS -_mm_maskload_ps(float const *__p, __m128 __m) +_mm_maskload_ps(float const *__p, __m128i __m) { - return (__m128)__builtin_ia32_maskloadps((const __v4sf *)__p, (__v4sf)__m); + return (__m128)__builtin_ia32_maskloadps((const __v4sf *)__p, (__v4si)__m); } static __inline __m256 __DEFAULT_FN_ATTRS -_mm256_maskload_ps(float const *__p, __m256 __m) +_mm256_maskload_ps(float const *__p, __m256i __m) { - return (__m256)__builtin_ia32_maskloadps256((const __v8sf *)__p, (__v8sf)__m); + return (__m256)__builtin_ia32_maskloadps256((const __v8sf *)__p, (__v8si)__m); } /* Conditional store ops */ static __inline void __DEFAULT_FN_ATTRS -_mm256_maskstore_ps(float *__p, __m256 __m, __m256 __a) +_mm256_maskstore_ps(float *__p, __m256i __m, __m256 __a) { - __builtin_ia32_maskstoreps256((__v8sf *)__p, (__v8sf)__m, (__v8sf)__a); + __builtin_ia32_maskstoreps256((__v8sf *)__p, (__v8si)__m, (__v8sf)__a); } static __inline void __DEFAULT_FN_ATTRS -_mm_maskstore_pd(double *__p, __m128d __m, __m128d __a) +_mm_maskstore_pd(double *__p, __m128i __m, __m128d __a) { - __builtin_ia32_maskstorepd((__v2df *)__p, (__v2df)__m, (__v2df)__a); + __builtin_ia32_maskstorepd((__v2df *)__p, (__v2di)__m, (__v2df)__a); } static __inline void __DEFAULT_FN_ATTRS -_mm256_maskstore_pd(double *__p, __m256d __m, __m256d __a) +_mm256_maskstore_pd(double *__p, __m256i __m, __m256d __a) { - __builtin_ia32_maskstorepd256((__v4df *)__p, (__v4df)__m, (__v4df)__a); + __builtin_ia32_maskstorepd256((__v4df *)__p, (__v4di)__m, (__v4df)__a); } static __inline void __DEFAULT_FN_ATTRS -_mm_maskstore_ps(float *__p, __m128 __m, __m128 __a) +_mm_maskstore_ps(float *__p, __m128i __m, __m128 __a) { - __builtin_ia32_maskstoreps((__v4sf *)__p, (__v4sf)__m, (__v4sf)__a); + __builtin_ia32_maskstoreps((__v4sf *)__p, (__v4si)__m, (__v4sf)__a); } /* Cacheability support ops */ |
