@cpp-linter/clang-tidy-win32-x64 19.1.0 → 21.1.0

package/lib/clang/{19 → 21}/include/xmmintrin.h

@@ -32,12 +32,41 @@ typedef unsigned int __v4su __attribute__((__vector_size__(16)));
 #endif
 
 /* Define the default attributes for the functions in this file. */
+#if defined(__EVEX512__) && !defined(__AVX10_1_512__)
 #define __DEFAULT_FN_ATTRS                                                     \
   __attribute__((__always_inline__, __nodebug__, __target__("sse,no-evex512"), \
                  __min_vector_width__(128)))
-#define __DEFAULT_FN_ATTRS_MMX                                                 \
+#define __DEFAULT_FN_ATTRS_SSE2                                                \
   __attribute__((__always_inline__, __nodebug__,                               \
-                 __target__("mmx,sse,no-evex512"), __min_vector_width__(64)))
+                 __target__("sse2,no-evex512"), __min_vector_width__(128)))
+#else
+#define __DEFAULT_FN_ATTRS                                                     \
+  __attribute__((__always_inline__, __nodebug__, __target__("sse"),            \
+                 __min_vector_width__(128)))
+#define __DEFAULT_FN_ATTRS_SSE2                                                \
+  __attribute__((__always_inline__, __nodebug__, __target__("sse2"),           \
+                 __min_vector_width__(128)))
+#endif
+
+#if defined(__cplusplus) && (__cplusplus >= 201103L)
+#define __DEFAULT_FN_ATTRS_CONSTEXPR __DEFAULT_FN_ATTRS constexpr
+#define __DEFAULT_FN_ATTRS_SSE2_CONSTEXPR __DEFAULT_FN_ATTRS_SSE2 constexpr
+#else
+#define __DEFAULT_FN_ATTRS_CONSTEXPR __DEFAULT_FN_ATTRS
+#define __DEFAULT_FN_ATTRS_SSE2_CONSTEXPR __DEFAULT_FN_ATTRS_SSE2
+#endif
+
+#define __trunc64(x)                                                           \
+  (__m64) __builtin_shufflevector((__v2di)(x), __extension__(__v2di){}, 0)
+#define __zext128(x)                                                           \
+  (__m128i) __builtin_shufflevector((__v2si)(x), __extension__(__v2si){}, 0,   \
+                                    1, 2, 3)
+#define __anyext128(x)                                                         \
+  (__m128i) __builtin_shufflevector((__v2si)(x), __extension__(__v2si){}, 0,   \
+                                    1, -1, -1)
+#define __zeroupper64(x)                                                       \
+  (__m128i) __builtin_shufflevector((__v4si)(x), __extension__(__v4si){}, 0,   \
+                                    1, 4, 5)
 
 /// Adds the 32-bit float values in the low-order bits of the operands.
 ///
@@ -54,9 +83,8 @@ typedef unsigned int __v4su __attribute__((__vector_size__(16)));
 /// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the sum
 ///    of the lower 32 bits of both operands. The upper 96 bits are copied from
 ///    the upper 96 bits of the first source operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_add_ss(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_add_ss(__m128 __a, __m128 __b) {
   __a[0] += __b[0];
   return __a;
 }
@@ -74,9 +102,8 @@ _mm_add_ss(__m128 __a, __m128 __b)
 ///    A 128-bit vector of [4 x float] containing one of the source operands.
 /// \returns A 128-bit vector of [4 x float] containing the sums of both
 ///    operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_add_ps(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_add_ps(__m128 __a, __m128 __b) {
   return (__m128)((__v4sf)__a + (__v4sf)__b);
 }
 
@@ -96,9 +123,8 @@ _mm_add_ps(__m128 __a, __m128 __b)
 /// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
 ///    difference of the lower 32 bits of both operands. The upper 96 bits are
 ///    copied from the upper 96 bits of the first source operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_sub_ss(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_sub_ss(__m128 __a, __m128 __b) {
   __a[0] -= __b[0];
   return __a;
 }
@@ -117,9 +143,8 @@ _mm_sub_ss(__m128 __a, __m128 __b)
 ///    A 128-bit vector of [4 x float] containing the subtrahend.
 /// \returns A 128-bit vector of [4 x float] containing the differences between
 ///    both operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_sub_ps(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_sub_ps(__m128 __a, __m128 __b) {
   return (__m128)((__v4sf)__a - (__v4sf)__b);
 }
 
@@ -139,9 +164,8 @@ _mm_sub_ps(__m128 __a, __m128 __b)
 /// \returns A 128-bit vector of [4 x float] containing the product of the lower
 ///    32 bits of both operands. The upper 96 bits are copied from the upper 96
 ///    bits of the first source operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_mul_ss(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_mul_ss(__m128 __a, __m128 __b) {
   __a[0] *= __b[0];
   return __a;
 }
@@ -159,9 +183,8 @@ _mm_mul_ss(__m128 __a, __m128 __b)
 ///    A 128-bit vector of [4 x float] containing one of the source operands.
 /// \returns A 128-bit vector of [4 x float] containing the products of both
 ///    operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_mul_ps(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_mul_ps(__m128 __a, __m128 __b) {
   return (__m128)((__v4sf)__a * (__v4sf)__b);
 }
 
@@ -181,9 +204,8 @@ _mm_mul_ps(__m128 __a, __m128 __b)
 /// \returns A 128-bit vector of [4 x float] containing the quotients of the
 ///    lower 32 bits of both operands. The upper 96 bits are copied from the
 ///    upper 96 bits of the first source operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_div_ss(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_div_ss(__m128 __a, __m128 __b) {
   __a[0] /= __b[0];
   return __a;
 }
@@ -200,9 +222,8 @@ _mm_div_ss(__m128 __a, __m128 __b)
 ///    A 128-bit vector of [4 x float] containing the divisor.
 /// \returns A 128-bit vector of [4 x float] containing the quotients of both
 ///    operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_div_ps(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_div_ps(__m128 __a, __m128 __b) {
   return (__m128)((__v4sf)__a / (__v4sf)__b);
 }
 
@@ -416,9 +437,8 @@ _mm_max_ps(__m128 __a, __m128 __b)
 ///    A 128-bit vector containing one of the source operands.
 /// \returns A 128-bit vector of [4 x float] containing the bitwise AND of the
 ///    values between both operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_and_ps(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_and_ps(__m128 __a, __m128 __b) {
   return (__m128)((__v4su)__a & (__v4su)__b);
 }
 
@@ -438,9 +458,8 @@ _mm_and_ps(__m128 __a, __m128 __b)
 /// \returns A 128-bit vector of [4 x float] containing the bitwise AND of the
 ///    one's complement of the first operand and the values in the second
 ///    operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_andnot_ps(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_andnot_ps(__m128 __a, __m128 __b) {
   return (__m128)(~(__v4su)__a & (__v4su)__b);
 }
 
@@ -456,9 +475,8 @@ _mm_andnot_ps(__m128 __a, __m128 __b)
 ///    A 128-bit vector of [4 x float] containing one of the source operands.
 /// \returns A 128-bit vector of [4 x float] containing the bitwise OR of the
 ///    values between both operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_or_ps(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_or_ps(__m128 __a, __m128 __b) {
   return (__m128)((__v4su)__a | (__v4su)__b);
 }
 
@@ -475,9 +493,8 @@ _mm_or_ps(__m128 __a, __m128 __b)
 ///    A 128-bit vector of [4 x float] containing one of the source operands.
 /// \returns A 128-bit vector of [4 x float] containing the bitwise exclusive OR
 ///    of the values between both operands.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_xor_ps(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_xor_ps(__m128 __a, __m128 __b) {
   return (__m128)((__v4su)__a ^ (__v4su)__b);
 }
 
@@ -1448,10 +1465,10 @@ _mm_cvtss_si64(__m128 __a)
 /// \param __a
 ///    A 128-bit vector of [4 x float].
 /// \returns A 64-bit integer vector containing the converted values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvtps_pi32(__m128 __a)
 {
-  return (__m64)__builtin_ia32_cvtps2pi((__v4sf)__a);
+  return __trunc64(__builtin_ia32_cvtps2dq((__v4sf)__zeroupper64(__a)));
 }
 
 /// Converts two low-order float values in a 128-bit vector of
@@ -1468,7 +1485,7 @@ _mm_cvtps_pi32(__m128 __a)
 /// \param __a
 ///    A 128-bit vector of [4 x float].
 /// \returns A 64-bit integer vector containing the converted values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvt_ps2pi(__m128 __a)
 {
   return _mm_cvtps_pi32(__a);
@@ -1558,10 +1575,10 @@ _mm_cvttss_si64(__m128 __a)
 /// \param __a
 ///    A 128-bit vector of [4 x float].
 /// \returns A 64-bit integer vector containing the converted values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvttps_pi32(__m128 __a)
 {
-  return (__m64)__builtin_ia32_cvttps2pi((__v4sf)__a);
+  return __trunc64(__builtin_ia32_cvttps2dq((__v4sf)__zeroupper64(__a)));
 }
 
 /// Converts the lower (first) two elements of a 128-bit vector of [4 x float]
@@ -1579,7 +1596,7 @@ _mm_cvttps_pi32(__m128 __a)
 /// \param __a
 ///    A 128-bit vector of [4 x float].
 /// \returns A 64-bit integer vector containing the converted values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvtt_ps2pi(__m128 __a)
 {
   return _mm_cvttps_pi32(__a);
@@ -1601,9 +1618,8 @@ _mm_cvtt_ps2pi(__m128 __a)
 /// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
 ///    converted value of the second operand. The upper 96 bits are copied from
 ///    the upper 96 bits of the first operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cvtsi32_ss(__m128 __a, int __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_cvtsi32_ss(__m128 __a,
+                                                                     int __b) {
   __a[0] = __b;
   return __a;
 }
@@ -1624,9 +1640,8 @@ _mm_cvtsi32_ss(__m128 __a, int __b)
 /// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
 ///    converted value of the second operand. The upper 96 bits are copied from
 ///    the upper 96 bits of the first operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cvt_si2ss(__m128 __a, int __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_cvt_si2ss(__m128 __a,
+                                                                    int __b) {
   return _mm_cvtsi32_ss(__a, __b);
 }
 
@@ -1648,9 +1663,8 @@ _mm_cvt_si2ss(__m128 __a, int __b)
 /// \returns A 128-bit vector of [4 x float] whose lower 32 bits contain the
 ///    converted value of the second operand. The upper 96 bits are copied from
 ///    the upper 96 bits of the first operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_cvtsi64_ss(__m128 __a, long long __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtsi64_ss(__m128 __a, long long __b) {
   __a[0] = __b;
   return __a;
 }
@@ -1674,10 +1688,13 @@ _mm_cvtsi64_ss(__m128 __a, long long __b)
 /// \returns A 128-bit vector of [4 x float] whose lower 64 bits contain the
 ///    converted value of the second operand. The upper 64 bits are copied from
 ///    the upper 64 bits of the first operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvtpi32_ps(__m128 __a, __m64 __b)
 {
-  return __builtin_ia32_cvtpi2ps((__v4sf)__a, (__v2si)__b);
+  return (__m128)__builtin_shufflevector(
+      (__v4sf)__a,
+      __builtin_convertvector((__v4si)__zext128(__b), __v4sf),
+      4, 5, 2, 3);
 }
 
 /// Converts two elements of a 64-bit vector of [2 x i32] into two
@@ -1697,7 +1714,7 @@ _mm_cvtpi32_ps(__m128 __a, __m64 __b)
 /// \returns A 128-bit vector of [4 x float] whose lower 64 bits contain the
 ///    converted value from the second operand. The upper 64 bits are copied
 ///    from the upper 64 bits of the first operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvt_pi2ps(__m128 __a, __m64 __b)
 {
   return _mm_cvtpi32_ps(__a, __b);
@@ -1714,9 +1731,8 @@ _mm_cvt_pi2ps(__m128 __a, __m64 __b)
 ///    A 128-bit vector of [4 x float]. The lower 32 bits of this operand are
 ///    used in the extraction.
 /// \returns A 32-bit float containing the extracted value.
-static __inline__ float __DEFAULT_FN_ATTRS
-_mm_cvtss_f32(__m128 __a)
-{
+static __inline__ float __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtss_f32(__m128 __a) {
   return __a[0];
 }
 
@@ -1907,9 +1923,8 @@ _mm_undefined_ps(void)
 /// \returns An initialized 128-bit floating-point vector of [4 x float]. The
 ///    lower 32 bits contain the value provided in the source operand. The
 ///    upper 96 bits are set to zero.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_set_ss(float __w)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_set_ss(float __w) {
   return __extension__ (__m128){ __w, 0.0f, 0.0f, 0.0f };
 }
 
@@ -1925,9 +1940,8 @@ _mm_set_ss(float __w)
 ///    A single-precision floating-point value used to initialize each vector
 ///    element of the result.
 /// \returns An initialized 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_set1_ps(float __w)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_set1_ps(float __w) {
   return __extension__ (__m128){ __w, __w, __w, __w };
 }
 
@@ -1944,9 +1958,8 @@ _mm_set1_ps(float __w)
 ///    A single-precision floating-point value used to initialize each vector
 ///    element of the result.
 /// \returns An initialized 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_set_ps1(float __w)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_set_ps1(float __w) {
     return _mm_set1_ps(__w);
 }
 
@@ -1971,9 +1984,8 @@ _mm_set_ps1(float __w)
 ///    A single-precision floating-point value used to initialize bits [31:0]
 ///    of the result.
 /// \returns An initialized 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_set_ps(float __z, float __y, float __x, float __w)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_set_ps(float __z, float __y, float __x, float __w) {
   return __extension__ (__m128){ __w, __x, __y, __z };
 }
 
@@ -1999,9 +2011,8 @@ _mm_set_ps(float __z, float __y, float __x, float __w)
 ///    A single-precision floating-point value used to initialize bits [127:96]
 ///    of the result.
 /// \returns An initialized 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_setr_ps(float __z, float __y, float __x, float __w)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_setr_ps(float __z, float __y, float __x, float __w) {
   return __extension__ (__m128){ __z, __y, __x, __w };
 }
 
@@ -2014,9 +2025,8 @@ _mm_setr_ps(float __z, float __y, float __x, float __w)
 ///
 /// \returns An initialized 128-bit floating-point vector of [4 x float] with
 ///    all elements set to zero.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_setzero_ps(void)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_setzero_ps(void) {
   return __extension__ (__m128){ 0.0f, 0.0f, 0.0f, 0.0f };
 }
 
@@ -2188,8 +2198,9 @@ _mm_storer_ps(float *__p, __m128 __a)
 #define _MM_HINT_NTA 0
 
 #ifndef _MSC_VER
-/* FIXME: We have to #define this because "sel" must be a constant integer, and
-   Sema doesn't do any form of constant propagation yet. */
+// If _MSC_VER is defined, we use the builtin variant of _mm_prefetch.
+// Otherwise, we provide this macro, which includes a cast, allowing the user
+// to pass a pointer of any time. The _mm_prefetch accepts char to match MSVC.
 
 /// Loads one cache line of data from the specified address to a location
 ///    closer to the processor.
@@ -2231,10 +2242,10 @@ _mm_storer_ps(float *__p, __m128 __a)
 ///    A pointer to an aligned memory location used to store the register value.
 /// \param __a
 ///    A 64-bit integer containing the value to be stored.
-static __inline__ void __DEFAULT_FN_ATTRS_MMX
+static __inline__ void __DEFAULT_FN_ATTRS
 _mm_stream_pi(void *__p, __m64 __a)
 {
-  __builtin_ia32_movntq((__m64 *)__p, __a);
+  __builtin_nontemporal_store(__a, (__m64 *)__p);
 }
 
 /// Moves packed float values from a 128-bit vector of [4 x float] to a
@@ -2296,7 +2307,7 @@ void _mm_sfence(void);
 ///    3: Bits [63:48] are copied to the destination.
 /// \returns A 16-bit integer containing the extracted 16 bits of packed data.
 #define _mm_extract_pi16(a, n) \
-  ((int)__builtin_ia32_vec_ext_v4hi((__v4hi)a, (int)n))
+  ((int)(unsigned short)__builtin_ia32_vec_ext_v4hi((__v4hi)a, (int)n))
 
 /// Copies data from the 64-bit vector of [4 x i16] to the destination,
 ///    and inserts the lower 16-bits of an integer operand at the 16-bit offset
@@ -2342,10 +2353,10 @@ void _mm_sfence(void);
 /// \param __b
 ///    A 64-bit integer vector containing one of the source operands.
 /// \returns A 64-bit integer vector containing the comparison results.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_max_pi16(__m64 __a, __m64 __b)
 {
-  return (__m64)__builtin_ia32_pmaxsw((__v4hi)__a, (__v4hi)__b);
+  return (__m64)__builtin_elementwise_max((__v4hi)__a, (__v4hi)__b);
 }
 
 /// Compares each of the corresponding packed 8-bit unsigned integer
@@ -2361,10 +2372,10 @@ _mm_max_pi16(__m64 __a, __m64 __b)
 /// \param __b
 ///    A 64-bit integer vector containing one of the source operands.
 /// \returns A 64-bit integer vector containing the comparison results.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_max_pu8(__m64 __a, __m64 __b)
 {
-  return (__m64)__builtin_ia32_pmaxub((__v8qi)__a, (__v8qi)__b);
+  return (__m64)__builtin_elementwise_max((__v8qu)__a, (__v8qu)__b);
 }
 
 /// Compares each of the corresponding packed 16-bit integer values of
@@ -2380,10 +2391,10 @@ _mm_max_pu8(__m64 __a, __m64 __b)
 /// \param __b
 ///    A 64-bit integer vector containing one of the source operands.
 /// \returns A 64-bit integer vector containing the comparison results.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_min_pi16(__m64 __a, __m64 __b)
 {
-  return (__m64)__builtin_ia32_pminsw((__v4hi)__a, (__v4hi)__b);
+  return (__m64)__builtin_elementwise_min((__v4hi)__a, (__v4hi)__b);
 }
 
 /// Compares each of the corresponding packed 8-bit unsigned integer
@@ -2399,10 +2410,10 @@ _mm_min_pi16(__m64 __a, __m64 __b)
 /// \param __b
 ///    A 64-bit integer vector containing one of the source operands.
 /// \returns A 64-bit integer vector containing the comparison results.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_min_pu8(__m64 __a, __m64 __b)
 {
-  return (__m64)__builtin_ia32_pminub((__v8qi)__a, (__v8qi)__b);
+  return (__m64)__builtin_elementwise_min((__v8qu)__a, (__v8qu)__b);
 }
 
 /// Takes the most significant bit from each 8-bit element in a 64-bit
@@ -2417,10 +2428,10 @@ _mm_min_pu8(__m64 __a, __m64 __b)
 ///    A 64-bit integer vector containing the values with bits to be extracted.
 /// \returns The most significant bit from each 8-bit element in \a __a,
 ///    written to bits [7:0].
-static __inline__ int __DEFAULT_FN_ATTRS_MMX
+static __inline__ int __DEFAULT_FN_ATTRS_SSE2
 _mm_movemask_pi8(__m64 __a)
 {
-  return __builtin_ia32_pmovmskb((__v8qi)__a);
+  return __builtin_ia32_pmovmskb128((__v16qi)__zext128(__a));
 }
 
 /// Multiplies packed 16-bit unsigned integer values and writes the
@@ -2436,10 +2447,11 @@ _mm_movemask_pi8(__m64 __a)
 /// \param __b
 ///    A 64-bit integer vector containing one of the source operands.
 /// \returns A 64-bit integer vector containing the products of both operands.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_mulhi_pu16(__m64 __a, __m64 __b)
 {
-  return (__m64)__builtin_ia32_pmulhuw((__v4hi)__a, (__v4hi)__b);
+  return __trunc64(__builtin_ia32_pmulhuw128((__v8hi)__anyext128(__a),
+                                             (__v8hi)__anyext128(__b)));
 }
 
 /// Shuffles the 4 16-bit integers from a 64-bit integer vector to the
@@ -2476,8 +2488,10 @@ _mm_mulhi_pu16(__m64 __a, __m64 __b)
 ///    <c>_MM_SHUFFLE(b6, b4, b2, b0)</c> can create an 8-bit mask of the form
 ///    <c>[b6, b4, b2, b0]</c>.
 /// \returns A 64-bit integer vector containing the shuffled values.
-#define _mm_shuffle_pi16(a, n) \
-  ((__m64)__builtin_ia32_pshufw((__v4hi)(__m64)(a), (n)))
+#define _mm_shuffle_pi16(a, n)                                                 \
+  ((__m64)__builtin_shufflevector((__v4hi)(__m64)(a), __extension__(__v4hi){}, \
+                                  (n) & 0x3, ((n) >> 2) & 0x3,                 \
+                                  ((n) >> 4) & 0x3, ((n) >> 6) & 0x3))
 
 /// Conditionally copies the values from each 8-bit element in the first
 ///    64-bit integer vector operand to the specified memory location, as
@@ -2502,10 +2516,25 @@ _mm_mulhi_pu16(__m64 __a, __m64 __b)
 ///    A pointer to a 64-bit memory location that will receive the conditionally
 ///    copied integer values. The address of the memory location does not have
 ///    to be aligned.
-static __inline__ void __DEFAULT_FN_ATTRS_MMX
+static __inline__ void __DEFAULT_FN_ATTRS_SSE2
 _mm_maskmove_si64(__m64 __d, __m64 __n, char *__p)
 {
-  __builtin_ia32_maskmovq((__v8qi)__d, (__v8qi)__n, __p);
+  // This is complex, because we need to support the case where __p is pointing
+  // within the last 15 to 8 bytes of a page. In that case, using a 128-bit
+  // write might cause a trap where a 64-bit maskmovq would not. (Memory
+  // locations not selected by the mask bits might still cause traps.)
+  __m128i __d128  = __anyext128(__d);
+  __m128i __n128  = __zext128(__n);
+  if (((__SIZE_TYPE__)__p & 0xfff) >= 4096-15 &&
+      ((__SIZE_TYPE__)__p & 0xfff) <= 4096-8) {
+    // If there's a risk of spurious trap due to a 128-bit write, back up the
+    // pointer by 8 bytes and shift values in registers to match.
+    __p -= 8;
+    __d128 = __builtin_ia32_pslldqi128_byteshift((__v2di)__d128, 8);
+    __n128 = __builtin_ia32_pslldqi128_byteshift((__v2di)__n128, 8);
+  }
+
+  __builtin_ia32_maskmovdqu((__v16qi)__d128, (__v16qi)__n128, __p);
 }
 
 /// Computes the rounded averages of the packed unsigned 8-bit integer
@@ -2521,10 +2550,11 @@ _mm_maskmove_si64(__m64 __d, __m64 __n, char *__p)
 /// \param __b
 ///    A 64-bit integer vector containing one of the source operands.
 /// \returns A 64-bit integer vector containing the averages of both operands.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_avg_pu8(__m64 __a, __m64 __b)
 {
-  return (__m64)__builtin_ia32_pavgb((__v8qi)__a, (__v8qi)__b);
+  return __trunc64(__builtin_ia32_pavgb128((__v16qi)__anyext128(__a),
+                                           (__v16qi)__anyext128(__b)));
 }
 
 /// Computes the rounded averages of the packed unsigned 16-bit integer
@@ -2540,10 +2570,11 @@ _mm_avg_pu8(__m64 __a, __m64 __b)
 /// \param __b
 ///    A 64-bit integer vector containing one of the source operands.
 /// \returns A 64-bit integer vector containing the averages of both operands.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_avg_pu16(__m64 __a, __m64 __b)
 {
-  return (__m64)__builtin_ia32_pavgw((__v4hi)__a, (__v4hi)__b);
+  return __trunc64(__builtin_ia32_pavgw128((__v8hi)__anyext128(__a),
+                                           (__v8hi)__anyext128(__b)));
 }
 
 /// Subtracts the corresponding 8-bit unsigned integer values of the two
@@ -2562,10 +2593,11 @@ _mm_avg_pu16(__m64 __a, __m64 __b)
 /// \returns A 64-bit integer vector whose lower 16 bits contain the sums of the
 ///    sets of absolute differences between both operands. The upper bits are
 ///    cleared.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_sad_pu8(__m64 __a, __m64 __b)
 {
-  return (__m64)__builtin_ia32_psadbw((__v8qi)__a, (__v8qi)__b);
+  return __trunc64(__builtin_ia32_psadbw128((__v16qi)__zext128(__a),
+                                            (__v16qi)__zext128(__b)));
 }
 
 #if defined(__cplusplus)
@@ -2741,9 +2773,8 @@ void _mm_setcsr(unsigned int __i);
 ///    Bits [95:64] are written to bits [63:32] of the destination. \n
 ///    Bits [127:96] are written to bits [127:96] of the destination.
 /// \returns A 128-bit vector of [4 x float] containing the interleaved values.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_unpackhi_ps(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_unpackhi_ps(__m128 __a, __m128 __b) {
   return __builtin_shufflevector((__v4sf)__a, (__v4sf)__b, 2, 6, 3, 7);
 }
 
@@ -2763,9 +2794,8 @@ _mm_unpackhi_ps(__m128 __a, __m128 __b)
 ///    Bits [31:0] are written to bits [63:32] of the destination. \n
 ///    Bits [63:32] are written to bits [127:96] of the destination.
 /// \returns A 128-bit vector of [4 x float] containing the interleaved values.
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_unpacklo_ps(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_unpacklo_ps(__m128 __a, __m128 __b) {
   return __builtin_shufflevector((__v4sf)__a, (__v4sf)__b, 0, 4, 1, 5);
 }
 
@@ -2785,9 +2815,8 @@ _mm_unpacklo_ps(__m128 __a, __m128 __b)
 ///    A 128-bit floating-point vector of [4 x float]. The lower 32 bits are
 ///    written to the lower 32 bits of the result.
 /// \returns A 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_move_ss(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_move_ss(__m128 __a, __m128 __b) {
   __a[0] = __b[0];
   return __a;
 }
@@ -2807,9 +2836,8 @@ _mm_move_ss(__m128 __a, __m128 __b)
 ///    A 128-bit floating-point vector of [4 x float]. The upper 64 bits are
 ///    written to the lower 64 bits of the result.
 /// \returns A 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_movehl_ps(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_movehl_ps(__m128 __a, __m128 __b) {
   return __builtin_shufflevector((__v4sf)__a, (__v4sf)__b, 6, 7, 2, 3);
 }
 
@@ -2828,9 +2856,8 @@ _mm_movehl_ps(__m128 __a, __m128 __b)
 ///    A 128-bit floating-point vector of [4 x float]. The lower 64 bits are
 ///    written to the upper 64 bits of the result.
 /// \returns A 128-bit floating-point vector of [4 x float].
-static __inline__ __m128 __DEFAULT_FN_ATTRS
-_mm_movelh_ps(__m128 __a, __m128 __b)
-{
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_movelh_ps(__m128 __a, __m128 __b) {
   return __builtin_shufflevector((__v4sf)__a, (__v4sf)__b, 0, 1, 4, 5);
 }
 
@@ -2846,22 +2873,10 @@ _mm_movelh_ps(__m128 __a, __m128 __b)
 ///    from the corresponding elements in this operand.
 /// \returns A 128-bit vector of [4 x float] containing the copied and converted
 ///    values from the operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvtpi16_ps(__m64 __a)
 {
-  __m64 __b, __c;
-  __m128 __r;
-
-  __b = _mm_setzero_si64();
-  __b = _mm_cmpgt_pi16(__b, __a);
-  __c = _mm_unpackhi_pi16(__a, __b);
-  __r = _mm_setzero_ps();
-  __r = _mm_cvtpi32_ps(__r, __c);
-  __r = _mm_movelh_ps(__r, __r);
-  __c = _mm_unpacklo_pi16(__a, __b);
-  __r = _mm_cvtpi32_ps(__r, __c);
-
-  return __r;
+  return __builtin_convertvector((__v4hi)__a, __v4sf);
 }
 
 /// Converts a 64-bit vector of 16-bit unsigned integer values into a
@@ -2876,21 +2891,10 @@ _mm_cvtpi16_ps(__m64 __a)
 ///    destination are copied from the corresponding elements in this operand.
 /// \returns A 128-bit vector of [4 x float] containing the copied and converted
 ///    values from the operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvtpu16_ps(__m64 __a)
 {
-  __m64 __b, __c;
-  __m128 __r;
-
-  __b = _mm_setzero_si64();
-  __c = _mm_unpackhi_pi16(__a, __b);
-  __r = _mm_setzero_ps();
-  __r = _mm_cvtpi32_ps(__r, __c);
-  __r = _mm_movelh_ps(__r, __r);
-  __c = _mm_unpacklo_pi16(__a, __b);
-  __r = _mm_cvtpi32_ps(__r, __c);
-
-  return __r;
+  return __builtin_convertvector((__v4hu)__a, __v4sf);
 }
 
 /// Converts the lower four 8-bit values from a 64-bit vector of [8 x i8]
@@ -2905,16 +2909,12 @@ _mm_cvtpu16_ps(__m64 __a)
 ///    from the corresponding lower 4 elements in this operand.
 /// \returns A 128-bit vector of [4 x float] containing the copied and converted
 ///    values from the operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvtpi8_ps(__m64 __a)
 {
-  __m64 __b;
-
-  __b = _mm_setzero_si64();
-  __b = _mm_cmpgt_pi8(__b, __a);
-  __b = _mm_unpacklo_pi8(__a, __b);
-
-  return _mm_cvtpi16_ps(__b);
+  return __builtin_convertvector(
+      __builtin_shufflevector((__v8qs)__a, __extension__ (__v8qs){},
+                              0, 1, 2, 3), __v4sf);
 }
 
 /// Converts the lower four unsigned 8-bit integer values from a 64-bit
@@ -2930,15 +2930,12 @@ _mm_cvtpi8_ps(__m64 __a)
 ///    operand.
 /// \returns A 128-bit vector of [4 x float] containing the copied and converted
 ///    values from the source operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvtpu8_ps(__m64 __a)
 {
-  __m64 __b;
-
-  __b = _mm_setzero_si64();
-  __b = _mm_unpacklo_pi8(__a, __b);
-
-  return _mm_cvtpi16_ps(__b);
+  return __builtin_convertvector(
+      __builtin_shufflevector((__v8qu)__a, __extension__ (__v8qu){},
+                              0, 1, 2, 3), __v4sf);
 }
 
 /// Converts the two 32-bit signed integer values from each 64-bit vector
@@ -2957,16 +2954,12 @@ _mm_cvtpu8_ps(__m64 __a)
 /// \returns A 128-bit vector of [4 x float] whose lower 64 bits contain the
 ///    copied and converted values from the first operand. The upper 64 bits
 ///    contain the copied and converted values from the second operand.
-static __inline__ __m128 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m128 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvtpi32x2_ps(__m64 __a, __m64 __b)
 {
-  __m128 __c;
-
-  __c = _mm_setzero_ps();
-  __c = _mm_cvtpi32_ps(__c, __b);
-  __c = _mm_movelh_ps(__c, __c);
-
-  return _mm_cvtpi32_ps(__c, __a);
+  return __builtin_convertvector(
+      __builtin_shufflevector((__v2si)__a, (__v2si)__b,
+                              0, 1, 2, 3), __v4sf);
 }
 
 /// Converts each single-precision floating-point element of a 128-bit
@@ -2986,16 +2979,11 @@ _mm_cvtpi32x2_ps(__m64 __a, __m64 __b)
 ///    A 128-bit floating-point vector of [4 x float].
 /// \returns A 64-bit integer vector of [4 x i16] containing the converted
 ///    values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvtps_pi16(__m128 __a)
 {
-  __m64 __b, __c;
-
-  __b = _mm_cvtps_pi32(__a);
-  __a = _mm_movehl_ps(__a, __a);
-  __c = _mm_cvtps_pi32(__a);
-
-  return _mm_packs_pi32(__b, __c);
+  return __trunc64(__builtin_ia32_packssdw128(
+      (__v4si)__builtin_ia32_cvtps2dq((__v4sf)__a), (__v4si)_mm_setzero_ps()));
 }
 
 /// Converts each single-precision floating-point element of a 128-bit
@@ -3016,7 +3004,7 @@ _mm_cvtps_pi16(__m128 __a)
 ///    128-bit floating-point vector of [4 x float].
 /// \returns A 64-bit integer vector of [8 x i8]. The lower 32 bits contain the
 ///    converted values and the uppper 32 bits are set to zero.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX
+static __inline__ __m64 __DEFAULT_FN_ATTRS_SSE2
 _mm_cvtps_pi8(__m128 __a)
 {
   __m64 __b, __c;
@@ -3196,8 +3184,14 @@ do { \
 #define _m_psadbw _mm_sad_pu8
 #define _m_ _mm_
 
+#undef __trunc64
+#undef __zext128
+#undef __anyext128
+#undef __zeroupper64
 #undef __DEFAULT_FN_ATTRS
-#undef __DEFAULT_FN_ATTRS_MMX
+#undef __DEFAULT_FN_ATTRS_CONSTEXPR
+#undef __DEFAULT_FN_ATTRS_SSE2
+#undef __DEFAULT_FN_ATTRS_SSE2_CONSTEXPR
 
 /* Ugly hack for backwards-compatibility (compatible with gcc) */
 #if defined(__SSE2__) && !__building_module(_Builtin_intrinsics)