* config/i386/tmmintrin.h (_mm_alignr_epi32): Implement as always

        inlined function, not as a macro.
        (_mm_alignr_pi8): Ditto.
        * config/i386/ammintrin.h (_mm_extracti_si64): Ditto.
        (_mm_inserti_si64): Ditto.
        * config/i386/emmintrin.h (_mm_shuffle_pd): Ditto.
        (_mm_extract_epi16): Ditto.
        (_mm_insert_epi16): Ditto.
        (_mm_shufflehi_epi16): Ditto.
        (_mm_shufflelo_epi16): Ditto.
        (_mm_shuffle_epi32): Ditto.
        * config/i386/xmmintrin.h (_mm_set_ss): Use 0.0f for float constant.
        * config/386/mm3dnow.h: Add __attribute__((__always_inline__)) to
        all functions.
        (_m_from_float): Add __extension__ to conversion.  Use 0.0f for
        float constant.
        (_m_to_float): Use C89 compatible assignment.

testsuite/ChangeLog:

        * gcc.target/i386/sse-vect-types.c: Revert 'Use "-msse"
        instead of "-msse2".'
        * gcc.target/i386/sse-12.c: Use "-march=k8 -m3dnow -mssse3 -msse4a"
        instead of "-msse3".  Include only ammintrin.h, tmmintrin.h
        and mm3dnow.h.
        * gcc.target/i386/sse-13.c (__builtin_ia32_extrqi): Redefine
        to test with immediate operand.
        (__builtin_ia32_insertqi): Ditto.
        (__builtin_ia32_palignr128): Ditto.
        (__builtin_ia32_palignr): Ditto.
        (__builtin_ia32_pshufhw): Ditto.
        (__builtin_ia32_pshuflw): Ditto.
        (__builtin_ia32_pshufd): Ditto.
        (__builtin_ia32_vec_set_v8hi): Ditto.
        (__builtin_ia32_vec_ext_v8hi): Ditto.
        (__builtin_ia32_shufpd): Ditto.
        * gcc.target/i386/sse-14.c: Same changes as sse-13.c.



git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@124873 138bc75d-0d04-0410-961f-82ee72b054a4

1 files changed, 25 insertions, 14 deletions

diff --git a/gcc/config/i386/emmintrin.h b/gcc/config/i386/emmintrin.h
index e8ef024c6e8..f878728a04f 100644
--- a/gcc/config/i386/emmintrin.h
+++ b/gcc/config/i386/emmintrin.h
@@ -57,7 +57,7 @@ typedef double __m128d __attribute__ ((__vector_size__ (16), __may_alias__));
 static __inline __m128d __attribute__((__always_inline__))
 _mm_set_sd (double __F)
 {
-  return __extension__ (__m128d){ __F, 0 };
+  return __extension__ (__m128d){ __F, 0.0 };
 }
 
 /* Create a vector with both elements equal to F.  */
@@ -880,7 +880,11 @@ _mm_cvtss_sd (__m128d __A, __m128 __B)
   return (__m128d)__builtin_ia32_cvtss2sd ((__v2df) __A, (__v4sf)__B);
 }
 
-#define _mm_shuffle_pd(__A, __B, __C) ((__m128d)__builtin_ia32_shufpd ((__v2df)__A, (__v2df)__B, (__C)))
+static __inline __m128d __attribute__((__always_inline__))
+_mm_shuffle_pd(__m128d __A, __m128d __B, const int __mask)
+{
+  return (__m128d)__builtin_ia32_shufpd ((__v2df)__A, (__v2df)__B, __mask);
+}
 
 static __inline __m128d __attribute__((__always_inline__))
 _mm_unpackhi_pd (__m128d __A, __m128d __B)
@@ -1137,13 +1141,13 @@ _mm_srai_epi32 (__m128i __A, const int __B)
 static __inline __m128i __attribute__((__always_inline__))
 _mm_srli_si128 (__m128i __A, const int __B)
 {
-  return ((__m128i)__builtin_ia32_psrldqi128 (__A, __B * 8));
+  return (__m128i)__builtin_ia32_psrldqi128 (__A, __B * 8);
 }
 
 static __inline __m128i __attribute__((__always_inline__))
 _mm_slli_si128 (__m128i __A, const int __B)
 {
-  return ((__m128i)__builtin_ia32_pslldqi128 (__A, __B * 8));
+  return (__m128i)__builtin_ia32_pslldqi128 (__A, __B * 8);
 }
 
 static __inline __m128i __attribute__((__always_inline__))
@@ -1290,7 +1294,6 @@ _mm_cmpgt_epi32 (__m128i __A, __m128i __B)
   return (__m128i)__builtin_ia32_pcmpgtd128 ((__v4si)__A, (__v4si)__B);
 }
 
-#if 0
 static __inline int __attribute__((__always_inline__))
 _mm_extract_epi16 (__m128i const __A, int const __N)
 {
@@ -1302,12 +1305,6 @@ _mm_insert_epi16 (__m128i const __A, int const __D, int const __N)
 {
   return (__m128i) __builtin_ia32_vec_set_v8hi ((__v8hi)__A, __D, __N);
 }
-#else
-#define _mm_extract_epi16(A, N) \
-  ((int) __builtin_ia32_vec_ext_v8hi ((__v8hi)(A), (N)))
-#define _mm_insert_epi16(A, D, N) \
-  ((__m128i) __builtin_ia32_vec_set_v8hi ((__v8hi)(A), (D), (N)))
-#endif
 
 static __inline __m128i __attribute__((__always_inline__))
 _mm_max_epi16 (__m128i __A, __m128i __B)
@@ -1345,9 +1342,23 @@ _mm_mulhi_epu16 (__m128i __A, __m128i __B)
   return (__m128i)__builtin_ia32_pmulhuw128 ((__v8hi)__A, (__v8hi)__B);
 }
 
-#define _mm_shufflehi_epi16(__A, __B) ((__m128i)__builtin_ia32_pshufhw ((__v8hi)__A, __B))
-#define _mm_shufflelo_epi16(__A, __B) ((__m128i)__builtin_ia32_pshuflw ((__v8hi)__A, __B))
-#define _mm_shuffle_epi32(__A, __B) ((__m128i)__builtin_ia32_pshufd ((__v4si)__A, __B))
+static __inline __m128i __attribute__((__always_inline__))
+_mm_shufflehi_epi16 (__m128i __A, const int __mask)
+{
+  return (__m128i)__builtin_ia32_pshufhw ((__v8hi)__A, __mask);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_shufflelo_epi16 (__m128i __A, const int __mask)
+{
+  return (__m128i)__builtin_ia32_pshuflw ((__v8hi)__A, __mask);
+}
+
+static __inline __m128i __attribute__((__always_inline__))
+_mm_shuffle_epi32 (__m128i __A, const int __mask)
+{
+  return (__m128i)__builtin_ia32_pshufd ((__v4si)__A, __mask);
+}
 
 static __inline void __attribute__((__always_inline__))
 _mm_maskmoveu_si128 (__m128i __A, __m128i __B, char *__C)