@cpp-linter/clang-tidy-win32-x64 19.1.0 → 23.1.3

package/lib/clang/{19 → 23}/include/emmintrin.h

@@ -17,7 +17,6 @@
 #include <xmmintrin.h>
 
 typedef double __m128d __attribute__((__vector_size__(16), __aligned__(16)));
-typedef long long __m128i __attribute__((__vector_size__(16), __aligned__(16)));
 
 typedef double __m128d_u __attribute__((__vector_size__(16), __aligned__(1)));
 typedef long long __m128i_u
@@ -25,14 +24,9 @@ typedef long long __m128i_u
 
 /* Type defines.  */
 typedef double __v2df __attribute__((__vector_size__(16)));
-typedef long long __v2di __attribute__((__vector_size__(16)));
-typedef short __v8hi __attribute__((__vector_size__(16)));
-typedef char __v16qi __attribute__((__vector_size__(16)));
 
 /* Unsigned types */
 typedef unsigned long long __v2du __attribute__((__vector_size__(16)));
-typedef unsigned short __v8hu __attribute__((__vector_size__(16)));
-typedef unsigned char __v16qu __attribute__((__vector_size__(16)));
 
 /* We need an explicitly signed variant for char. Note that this shouldn't
  * appear in the interface though. */
@@ -50,11 +44,23 @@ typedef __bf16 __m128bh __attribute__((__vector_size__(16), __aligned__(16)));
 
 /* Define the default attributes for the functions in this file. */
 #define __DEFAULT_FN_ATTRS                                                     \
-  __attribute__((__always_inline__, __nodebug__,                               \
-                 __target__("sse2,no-evex512"), __min_vector_width__(128)))
-#define __DEFAULT_FN_ATTRS_MMX                                                 \
-  __attribute__((__always_inline__, __nodebug__,                               \
-                 __target__("mmx,sse2,no-evex512"), __min_vector_width__(64)))
+  __attribute__((__always_inline__, __nodebug__, __target__("sse2"),           \
+                 __min_vector_width__(128)))
+
+#if defined(__cplusplus) && (__cplusplus >= 201103L)
+#define __DEFAULT_FN_ATTRS_CONSTEXPR __DEFAULT_FN_ATTRS constexpr
+#else
+#define __DEFAULT_FN_ATTRS_CONSTEXPR __DEFAULT_FN_ATTRS
+#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)
 
 /// Adds lower double-precision values in both operands and returns the
 ///    sum in the lower 64 bits of the result. The upper 64 bits of the result
@@ -71,8 +77,8 @@ typedef __bf16 __m128bh __attribute__((__vector_size__(16), __aligned__(16)));
 /// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
 ///    sum of the lower 64 bits of both operands. The upper 64 bits are copied
 ///    from the upper 64 bits of the first source operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_add_sd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_add_sd(__m128d __a,
+                                                                  __m128d __b) {
   __a[0] += __b[0];
   return __a;
 }
@@ -89,8 +95,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_add_sd(__m128d __a,
 ///    A 128-bit vector of [2 x double] containing one of the source operands.
 /// \returns A 128-bit vector of [2 x double] containing the sums of both
 ///    operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_add_pd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_add_pd(__m128d __a,
+                                                                  __m128d __b) {
   return (__m128d)((__v2df)__a + (__v2df)__b);
 }
 
@@ -111,8 +117,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_add_pd(__m128d __a,
 /// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
 ///    difference of the lower 64 bits of both operands. The upper 64 bits are
 ///    copied from the upper 64 bits of the first source operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_sub_sd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_sub_sd(__m128d __a,
+                                                                  __m128d __b) {
   __a[0] -= __b[0];
   return __a;
 }
@@ -129,8 +135,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_sub_sd(__m128d __a,
 ///    A 128-bit vector of [2 x double] containing the subtrahend.
 /// \returns A 128-bit vector of [2 x double] containing the differences between
 ///    both operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_sub_pd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_sub_pd(__m128d __a,
+                                                                  __m128d __b) {
   return (__m128d)((__v2df)__a - (__v2df)__b);
 }
 
@@ -150,8 +156,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_sub_pd(__m128d __a,
 /// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
 ///    product of the lower 64 bits of both operands. The upper 64 bits are
 ///    copied from the upper 64 bits of the first source operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_mul_sd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_mul_sd(__m128d __a,
+                                                                  __m128d __b) {
   __a[0] *= __b[0];
   return __a;
 }
@@ -168,8 +174,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_mul_sd(__m128d __a,
 ///    A 128-bit vector of [2 x double] containing one of the operands.
 /// \returns A 128-bit vector of [2 x double] containing the products of both
 ///    operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_mul_pd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_mul_pd(__m128d __a,
+                                                                  __m128d __b) {
   return (__m128d)((__v2df)__a * (__v2df)__b);
 }
 
@@ -190,8 +196,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_mul_pd(__m128d __a,
 /// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
 ///    quotient of the lower 64 bits of both operands. The upper 64 bits are
 ///    copied from the upper 64 bits of the first source operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_div_sd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_div_sd(__m128d __a,
+                                                                  __m128d __b) {
   __a[0] /= __b[0];
   return __a;
 }
@@ -209,8 +215,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_div_sd(__m128d __a,
 ///    A 128-bit vector of [2 x double] containing the divisor.
 /// \returns A 128-bit vector of [2 x double] containing the quotients of both
 ///    operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_div_pd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_div_pd(__m128d __a,
+                                                                  __m128d __b) {
   return (__m128d)((__v2df)__a / (__v2df)__b);
 }
 
@@ -235,8 +241,7 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_div_pd(__m128d __a,
 ///    bits are copied from the upper 64 bits of operand \a __a.
 static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_sqrt_sd(__m128d __a,
                                                          __m128d __b) {
-  __m128d __c = __builtin_ia32_sqrtsd((__v2df)__b);
-  return __extension__(__m128d){__c[0], __a[1]};
+  return __extension__(__m128d){__builtin_elementwise_sqrt(__b[0]), __a[1]};
 }
 
 /// Calculates the square root of the each of two values stored in a
@@ -251,7 +256,7 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_sqrt_sd(__m128d __a,
 /// \returns A 128-bit vector of [2 x double] containing the square roots of the
 ///    values in the operand.
 static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_sqrt_pd(__m128d __a) {
-  return __builtin_ia32_sqrtpd((__v2df)__a);
+  return __builtin_elementwise_sqrt(__a);
 }
 
 /// Compares lower 64-bit double-precision values of both operands, and
@@ -274,8 +279,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_sqrt_pd(__m128d __a) {
 /// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
 ///    minimum value between both operands. The upper 64 bits are copied from
 ///    the upper 64 bits of the first source operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_min_sd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_min_sd(__m128d __a,
+                                                                  __m128d __b) {
   return __builtin_ia32_minsd((__v2df)__a, (__v2df)__b);
 }
 
@@ -295,8 +300,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_min_sd(__m128d __a,
 ///    A 128-bit vector of [2 x double] containing one of the operands.
 /// \returns A 128-bit vector of [2 x double] containing the minimum values
 ///    between both operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_min_pd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_min_pd(__m128d __a,
+                                                                  __m128d __b) {
   return __builtin_ia32_minpd((__v2df)__a, (__v2df)__b);
 }
 
@@ -320,8 +325,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_min_pd(__m128d __a,
 /// \returns A 128-bit vector of [2 x double] whose lower 64 bits contain the
 ///    maximum value between both operands. The upper 64 bits are copied from
 ///    the upper 64 bits of the first source operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_max_sd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_max_sd(__m128d __a,
+                                                                  __m128d __b) {
   return __builtin_ia32_maxsd((__v2df)__a, (__v2df)__b);
 }
 
@@ -341,8 +346,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_max_sd(__m128d __a,
 ///    A 128-bit vector of [2 x double] containing one of the operands.
 /// \returns A 128-bit vector of [2 x double] containing the maximum values
 ///    between both operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_max_pd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_max_pd(__m128d __a,
+                                                                  __m128d __b) {
   return __builtin_ia32_maxpd((__v2df)__a, (__v2df)__b);
 }
 
@@ -358,8 +363,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_max_pd(__m128d __a,
 ///    A 128-bit vector of [2 x double] containing one of the source operands.
 /// \returns A 128-bit vector of [2 x double] containing the bitwise AND of the
 ///    values between both operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_and_pd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_and_pd(__m128d __a,
+                                                                  __m128d __b) {
   return (__m128d)((__v2du)__a & (__v2du)__b);
 }
 
@@ -378,8 +383,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_and_pd(__m128d __a,
 /// \returns A 128-bit vector of [2 x double] containing the bitwise AND of the
 ///    values in the second operand and the one's complement of the first
 ///    operand.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_andnot_pd(__m128d __a,
-                                                           __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_andnot_pd(__m128d __a, __m128d __b) {
   return (__m128d)(~(__v2du)__a & (__v2du)__b);
 }
 
@@ -395,8 +400,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_andnot_pd(__m128d __a,
 ///    A 128-bit vector of [2 x double] containing one of the source operands.
 /// \returns A 128-bit vector of [2 x double] containing the bitwise OR of the
 ///    values between both operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_or_pd(__m128d __a,
-                                                       __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_or_pd(__m128d __a,
+                                                                 __m128d __b) {
   return (__m128d)((__v2du)__a | (__v2du)__b);
 }
 
@@ -412,8 +417,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_or_pd(__m128d __a,
 ///    A 128-bit vector of [2 x double] containing one of the source operands.
 /// \returns A 128-bit vector of [2 x double] containing the bitwise XOR of the
 ///    values between both operands.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_xor_pd(__m128d __a,
-                                                        __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_xor_pd(__m128d __a,
+                                                                  __m128d __b) {
   return (__m128d)((__v2du)__a ^ (__v2du)__b);
 }
 
@@ -1273,7 +1278,8 @@ static __inline__ int __DEFAULT_FN_ATTRS _mm_ucomineq_sd(__m128d __a,
 ///    A 128-bit vector of [2 x double].
 /// \returns A 128-bit vector of [4 x float] whose lower 64 bits contain the
 ///    converted values. The upper 64 bits are set to zero.
-static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cvtpd_ps(__m128d __a) {
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtpd_ps(__m128d __a) {
   return __builtin_ia32_cvtpd2ps((__v2df)__a);
 }
 
@@ -1291,7 +1297,8 @@ static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cvtpd_ps(__m128d __a) {
 ///    floating-point elements are converted to double-precision values. The
 ///    upper two elements are unused.
 /// \returns A 128-bit vector of [2 x double] containing the converted values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_cvtps_pd(__m128 __a) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtps_pd(__m128 __a) {
   return (__m128d) __builtin_convertvector(
       __builtin_shufflevector((__v4sf)__a, (__v4sf)__a, 0, 1), __v2df);
 }
@@ -1312,7 +1319,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_cvtps_pd(__m128 __a) {
 ///
 ///    The upper two elements are unused.
 /// \returns A 128-bit vector of [2 x double] containing the converted values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_cvtepi32_pd(__m128i __a) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtepi32_pd(__m128i __a) {
   return (__m128d) __builtin_convertvector(
       __builtin_shufflevector((__v4si)__a, (__v4si)__a, 0, 1), __v2df);
 }
@@ -1335,7 +1343,7 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_cvtepi32_pd(__m128i __a) {
 /// \returns A 128-bit vector of [4 x i32] whose lower 64 bits contain the
 ///    converted values. The upper 64 bits are set to zero.
 static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtpd_epi32(__m128d __a) {
-  return __builtin_ia32_cvtpd2dq((__v2df)__a);
+  return (__m128i)__builtin_ia32_cvtpd2dq((__v2df)__a);
 }
 
 /// Converts the low-order element of a 128-bit vector of [2 x double]
@@ -1376,8 +1384,8 @@ static __inline__ int __DEFAULT_FN_ATTRS _mm_cvtsd_si32(__m128d __a) {
 /// \returns A 128-bit vector of [4 x float]. The lower 32 bits contain the
 ///    converted value from the second parameter. The upper 96 bits are copied
 ///    from the upper 96 bits of the first parameter.
-static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cvtsd_ss(__m128 __a,
-                                                         __m128d __b) {
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtsd_ss(__m128 __a, __m128d __b) {
   return (__m128)__builtin_ia32_cvtsd2ss((__v4sf)__a, (__v2df)__b);
 }
 
@@ -1398,8 +1406,8 @@ static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cvtsd_ss(__m128 __a,
 /// \returns A 128-bit vector of [2 x double]. The lower 64 bits contain the
 ///    converted value from the second parameter. The upper 64 bits are copied
 ///    from the upper 64 bits of the first parameter.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_cvtsi32_sd(__m128d __a,
-                                                            int __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtsi32_sd(__m128d __a, int __b) {
   __a[0] = __b;
   return __a;
 }
@@ -1423,8 +1431,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_cvtsi32_sd(__m128d __a,
 /// \returns A 128-bit vector of [2 x double]. The lower 64 bits contain the
 ///    converted value from the second parameter. The upper 64 bits are copied
 ///    from the upper 64 bits of the first parameter.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_cvtss_sd(__m128d __a,
-                                                          __m128 __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtss_sd(__m128d __a, __m128 __b) {
   __a[0] = __b[0];
   return __a;
 }
@@ -1486,8 +1494,8 @@ static __inline__ int __DEFAULT_FN_ATTRS _mm_cvttsd_si32(__m128d __a) {
 /// \param __a
 ///    A 128-bit vector of [2 x double].
 /// \returns A 64-bit vector of [2 x i32] containing the converted values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_cvtpd_pi32(__m128d __a) {
-  return (__m64)__builtin_ia32_cvtpd2pi((__v2df)__a);
+static __inline__ __m64 __DEFAULT_FN_ATTRS _mm_cvtpd_pi32(__m128d __a) {
+  return __trunc64(__builtin_ia32_cvtpd2dq((__v2df)__a));
 }
 
 /// Converts the two double-precision floating-point elements of a
@@ -1505,8 +1513,8 @@ static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_cvtpd_pi32(__m128d __a) {
 /// \param __a
 ///    A 128-bit vector of [2 x double].
 /// \returns A 64-bit vector of [2 x i32] containing the converted values.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_cvttpd_pi32(__m128d __a) {
-  return (__m64)__builtin_ia32_cvttpd2pi((__v2df)__a);
+static __inline__ __m64 __DEFAULT_FN_ATTRS _mm_cvttpd_pi32(__m128d __a) {
+  return __trunc64(__builtin_ia32_cvttpd2dq((__v2df)__a));
 }
 
 /// Converts the two signed 32-bit integer elements of a 64-bit vector of
@@ -1520,8 +1528,9 @@ static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_cvttpd_pi32(__m128d __a) {
 /// \param __a
 ///    A 64-bit vector of [2 x i32].
 /// \returns A 128-bit vector of [2 x double] containing the converted values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS_MMX _mm_cvtpi32_pd(__m64 __a) {
-  return __builtin_ia32_cvtpi2pd((__v2si)__a);
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtpi32_pd(__m64 __a) {
+  return (__m128d) __builtin_convertvector((__v2si)__a, __v2df);
 }
 
 /// Returns the low-order element of a 128-bit vector of [2 x double] as
@@ -1535,7 +1544,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS_MMX _mm_cvtpi32_pd(__m64 __a) {
 ///    A 128-bit vector of [2 x double]. The lower 64 bits are returned.
 /// \returns A double-precision floating-point value copied from the lower 64
 ///    bits of \a __a.
-static __inline__ double __DEFAULT_FN_ATTRS _mm_cvtsd_f64(__m128d __a) {
+static __inline__ double __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtsd_f64(__m128d __a) {
   return __a[0];
 }
 
@@ -1770,7 +1780,7 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_undefined_pd(void) {
 /// \returns An initialized 128-bit floating-point vector of [2 x double]. The
 ///    lower 64 bits contain the value of the parameter. The upper 64 bits are
 ///    set to zero.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_set_sd(double __w) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_set_sd(double __w) {
   return __extension__(__m128d){__w, 0.0};
 }
 
@@ -1786,7 +1796,7 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_set_sd(double __w) {
 ///    A double-precision floating-point value used to initialize each vector
 ///    element of the result.
 /// \returns An initialized 128-bit floating-point vector of [2 x double].
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_set1_pd(double __w) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_set1_pd(double __w) {
   return __extension__(__m128d){__w, __w};
 }
 
@@ -1802,7 +1812,7 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_set1_pd(double __w) {
 ///    A double-precision floating-point value used to initialize each vector
 ///    element of the result.
 /// \returns An initialized 128-bit floating-point vector of [2 x double].
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_set_pd1(double __w) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_set_pd1(double __w) {
   return _mm_set1_pd(__w);
 }
 
@@ -1820,8 +1830,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_set_pd1(double __w) {
 ///    A double-precision floating-point value used to initialize the lower 64
 ///    bits of the result.
 /// \returns An initialized 128-bit floating-point vector of [2 x double].
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_set_pd(double __w,
-                                                        double __x) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_set_pd(double __w,
+                                                                  double __x) {
   return __extension__(__m128d){__x, __w};
 }
 
@@ -1840,8 +1850,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_set_pd(double __w,
 ///    A double-precision floating-point value used to initialize the upper 64
 ///    bits of the result.
 /// \returns An initialized 128-bit floating-point vector of [2 x double].
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_setr_pd(double __w,
-                                                         double __x) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_setr_pd(double __w,
+                                                                   double __x) {
   return __extension__(__m128d){__w, __x};
 }
 
@@ -1854,7 +1864,7 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_setr_pd(double __w,
 ///
 /// \returns An initialized 128-bit floating-point vector of [2 x double] with
 ///    all elements set to zero.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_setzero_pd(void) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR _mm_setzero_pd(void) {
   return __extension__(__m128d){0.0, 0.0};
 }
 
@@ -1873,8 +1883,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_setzero_pd(void) {
 ///    A 128-bit vector of [2 x double]. The lower 64 bits are written to the
 ///    lower 64 bits of the result.
 /// \returns A 128-bit vector of [2 x double] containing the moved values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_move_sd(__m128d __a,
-                                                         __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_move_sd(__m128d __a, __m128d __b) {
   __a[0] = __b[0];
   return __a;
 }
@@ -2049,8 +2059,8 @@ static __inline__ void __DEFAULT_FN_ATTRS _mm_storel_pd(double *__dp,
 ///    A 128-bit vector of [16 x i8].
 /// \returns A 128-bit vector of [16 x i8] containing the sums of both
 ///    parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_add_epi8(__m128i __a,
-                                                          __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_add_epi8(__m128i __a, __m128i __b) {
   return (__m128i)((__v16qu)__a + (__v16qu)__b);
 }
 
@@ -2070,8 +2080,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_add_epi8(__m128i __a,
 ///    A 128-bit vector of [8 x i16].
 /// \returns A 128-bit vector of [8 x i16] containing the sums of both
 ///    parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_add_epi16(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_add_epi16(__m128i __a, __m128i __b) {
   return (__m128i)((__v8hu)__a + (__v8hu)__b);
 }
 
@@ -2091,8 +2101,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_add_epi16(__m128i __a,
 ///    A 128-bit vector of [4 x i32].
 /// \returns A 128-bit vector of [4 x i32] containing the sums of both
 ///    parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_add_epi32(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_add_epi32(__m128i __a, __m128i __b) {
   return (__m128i)((__v4su)__a + (__v4su)__b);
 }
 
@@ -2108,9 +2118,9 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_add_epi32(__m128i __a,
 /// \param __b
 ///    A 64-bit integer.
 /// \returns A 64-bit integer containing the sum of both parameters.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_add_si64(__m64 __a,
-                                                            __m64 __b) {
-  return (__m64)__builtin_ia32_paddq((__v1di)__a, (__v1di)__b);
+static __inline__ __m64 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_add_si64(__m64 __a,
+                                                                  __m64 __b) {
+  return (__m64)(((__v1du)__a)[0] + ((__v1du)__b)[0]);
 }
 
 /// Adds the corresponding elements of two 128-bit vectors of [2 x i64],
@@ -2129,8 +2139,8 @@ static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_add_si64(__m64 __a,
 ///    A 128-bit vector of [2 x i64].
 /// \returns A 128-bit vector of [2 x i64] containing the sums of both
 ///    parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_add_epi64(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_add_epi64(__m128i __a, __m128i __b) {
   return (__m128i)((__v2du)__a + (__v2du)__b);
 }
 
@@ -2151,8 +2161,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_add_epi64(__m128i __a,
 ///    A 128-bit signed [16 x i8] vector.
 /// \returns A 128-bit signed [16 x i8] vector containing the saturated sums of
 ///    both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_adds_epi8(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_adds_epi8(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_elementwise_add_sat((__v16qs)__a, (__v16qs)__b);
 }
 
@@ -2173,8 +2183,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_adds_epi8(__m128i __a,
 ///    A 128-bit signed [8 x i16] vector.
 /// \returns A 128-bit signed [8 x i16] vector containing the saturated sums of
 ///    both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_adds_epi16(__m128i __a,
-                                                            __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_adds_epi16(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_elementwise_add_sat((__v8hi)__a, (__v8hi)__b);
 }
 
@@ -2195,8 +2205,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_adds_epi16(__m128i __a,
 ///    A 128-bit unsigned [16 x i8] vector.
 /// \returns A 128-bit unsigned [16 x i8] vector containing the saturated sums
 ///    of both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_adds_epu8(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_adds_epu8(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_elementwise_add_sat((__v16qu)__a, (__v16qu)__b);
 }
 
@@ -2217,8 +2227,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_adds_epu8(__m128i __a,
 ///    A 128-bit unsigned [8 x i16] vector.
 /// \returns A 128-bit unsigned [8 x i16] vector containing the saturated sums
 ///    of both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_adds_epu16(__m128i __a,
-                                                            __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_adds_epu16(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_elementwise_add_sat((__v8hu)__a, (__v8hu)__b);
 }
 
@@ -2236,9 +2246,9 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_adds_epu16(__m128i __a,
 ///    A 128-bit unsigned [16 x i8] vector.
 /// \returns A 128-bit unsigned [16 x i8] vector containing the rounded
 ///    averages of both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_avg_epu8(__m128i __a,
-                                                          __m128i __b) {
-  return (__m128i)__builtin_ia32_pavgb128((__v16qi)__a, (__v16qi)__b);
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_avg_epu8(__m128i __a, __m128i __b) {
+  return (__m128i)__builtin_ia32_pavgb128((__v16qu)__a, (__v16qu)__b);
 }
 
 /// Computes the rounded averages of corresponding elements of two
@@ -2255,9 +2265,9 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_avg_epu8(__m128i __a,
 ///    A 128-bit unsigned [8 x i16] vector.
 /// \returns A 128-bit unsigned [8 x i16] vector containing the rounded
 ///    averages of both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_avg_epu16(__m128i __a,
-                                                           __m128i __b) {
-  return (__m128i)__builtin_ia32_pavgw128((__v8hi)__a, (__v8hi)__b);
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_avg_epu16(__m128i __a, __m128i __b) {
+  return (__m128i)__builtin_ia32_pavgw128((__v8hu)__a, (__v8hu)__b);
 }
 
 /// Multiplies the corresponding elements of two 128-bit signed [8 x i16]
@@ -2280,8 +2290,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_avg_epu16(__m128i __a,
 ///    A 128-bit signed [8 x i16] vector.
 /// \returns A 128-bit signed [4 x i32] vector containing the sums of products
 ///    of both parameters.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_madd_epi16(__m128i __a,
-                                                            __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_madd_epi16(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_ia32_pmaddwd128((__v8hi)__a, (__v8hi)__b);
 }
 
@@ -2299,8 +2309,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_madd_epi16(__m128i __a,
 ///    A 128-bit signed [8 x i16] vector.
 /// \returns A 128-bit signed [8 x i16] vector containing the greater value of
 ///    each comparison.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_max_epi16(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_max_epi16(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_elementwise_max((__v8hi)__a, (__v8hi)__b);
 }
 
@@ -2318,8 +2328,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_max_epi16(__m128i __a,
 ///    A 128-bit unsigned [16 x i8] vector.
 /// \returns A 128-bit unsigned [16 x i8] vector containing the greater value of
 ///    each comparison.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_max_epu8(__m128i __a,
-                                                          __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_max_epu8(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_elementwise_max((__v16qu)__a, (__v16qu)__b);
 }
 
@@ -2337,8 +2347,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_max_epu8(__m128i __a,
 ///    A 128-bit signed [8 x i16] vector.
 /// \returns A 128-bit signed [8 x i16] vector containing the smaller value of
 ///    each comparison.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_min_epi16(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_min_epi16(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_elementwise_min((__v8hi)__a, (__v8hi)__b);
 }
 
@@ -2356,8 +2366,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_min_epi16(__m128i __a,
 ///    A 128-bit unsigned [16 x i8] vector.
 /// \returns A 128-bit unsigned [16 x i8] vector containing the smaller value of
 ///    each comparison.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_min_epu8(__m128i __a,
-                                                          __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_min_epu8(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_elementwise_min((__v16qu)__a, (__v16qu)__b);
 }
 
@@ -2375,8 +2385,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_min_epu8(__m128i __a,
 ///    A 128-bit signed [8 x i16] vector.
 /// \returns A 128-bit signed [8 x i16] vector containing the upper 16 bits of
 ///    each of the eight 32-bit products.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_mulhi_epi16(__m128i __a,
-                                                             __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_mulhi_epi16(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_ia32_pmulhw128((__v8hi)__a, (__v8hi)__b);
 }
 
@@ -2394,9 +2404,9 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_mulhi_epi16(__m128i __a,
 ///    A 128-bit unsigned [8 x i16] vector.
 /// \returns A 128-bit unsigned [8 x i16] vector containing the upper 16 bits
 ///    of each of the eight 32-bit products.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_mulhi_epu16(__m128i __a,
-                                                             __m128i __b) {
-  return (__m128i)__builtin_ia32_pmulhuw128((__v8hi)__a, (__v8hi)__b);
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_mulhi_epu16(__m128i __a, __m128i __b) {
+  return (__m128i)__builtin_ia32_pmulhuw128((__v8hu)__a, (__v8hu)__b);
 }
 
 /// Multiplies the corresponding elements of two signed [8 x i16]
@@ -2413,8 +2423,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_mulhi_epu16(__m128i __a,
 ///    A 128-bit signed [8 x i16] vector.
 /// \returns A 128-bit signed [8 x i16] vector containing the lower 16 bits of
 ///    each of the eight 32-bit products.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_mullo_epi16(__m128i __a,
-                                                             __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_mullo_epi16(__m128i __a, __m128i __b) {
   return (__m128i)((__v8hu)__a * (__v8hu)__b);
 }
 
@@ -2431,9 +2441,10 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_mullo_epi16(__m128i __a,
 /// \param __b
 ///    A 64-bit integer containing one of the source operands.
 /// \returns A 64-bit integer vector containing the product of both operands.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_mul_su32(__m64 __a,
-                                                            __m64 __b) {
-  return __builtin_ia32_pmuludq((__v2si)__a, (__v2si)__b);
+static __inline__ __m64 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_mul_su32(__m64 __a,
+                                                                  __m64 __b) {
+  return __trunc64(__builtin_ia32_pmuludq128((__v4si)__zext128(__a),
+                                             (__v4si)__zext128(__b)));
 }
 
 /// Multiplies 32-bit unsigned integer values contained in the lower
@@ -2449,8 +2460,8 @@ static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_mul_su32(__m64 __a,
 /// \param __b
 ///    A [2 x i64] vector containing one of the source operands.
 /// \returns A [2 x i64] vector containing the product of both operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_mul_epu32(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_mul_epu32(__m128i __a, __m128i __b) {
   return __builtin_ia32_pmuludq128((__v4si)__a, (__v4si)__b);
 }
 
@@ -2470,9 +2481,9 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_mul_epu32(__m128i __a,
 ///    A 128-bit integer vector containing one of the source operands.
 /// \returns A [2 x i64] vector containing the sums of the sets of absolute
 ///    differences between both operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sad_epu8(__m128i __a,
-                                                          __m128i __b) {
-  return __builtin_ia32_psadbw128((__v16qi)__a, (__v16qi)__b);
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_sad_epu8(__m128i __a, __m128i __b) {
+  return __builtin_ia32_psadbw128((__v16qu)__a, (__v16qu)__b);
 }
 
 /// Subtracts the corresponding 8-bit integer values in the operands.
@@ -2487,8 +2498,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sad_epu8(__m128i __a,
 ///    A 128-bit integer vector containing the subtrahends.
 /// \returns A 128-bit integer vector containing the differences of the values
 ///    in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sub_epi8(__m128i __a,
-                                                          __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_sub_epi8(__m128i __a, __m128i __b) {
   return (__m128i)((__v16qu)__a - (__v16qu)__b);
 }
 
@@ -2504,8 +2515,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sub_epi8(__m128i __a,
 ///    A 128-bit integer vector containing the subtrahends.
 /// \returns A 128-bit integer vector containing the differences of the values
 ///    in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sub_epi16(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_sub_epi16(__m128i __a, __m128i __b) {
   return (__m128i)((__v8hu)__a - (__v8hu)__b);
 }
 
@@ -2521,8 +2532,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sub_epi16(__m128i __a,
 ///    A 128-bit integer vector containing the subtrahends.
 /// \returns A 128-bit integer vector containing the differences of the values
 ///    in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sub_epi32(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_sub_epi32(__m128i __a, __m128i __b) {
   return (__m128i)((__v4su)__a - (__v4su)__b);
 }
 
@@ -2539,9 +2550,9 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sub_epi32(__m128i __a,
 ///    A 64-bit integer vector containing the subtrahend.
 /// \returns A 64-bit integer vector containing the difference of the values in
 ///    the operands.
-static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_sub_si64(__m64 __a,
-                                                            __m64 __b) {
-  return (__m64)__builtin_ia32_psubq((__v1di)__a, (__v1di)__b);
+static __inline__ __m64 __DEFAULT_FN_ATTRS_CONSTEXPR _mm_sub_si64(__m64 __a,
+                                                                  __m64 __b) {
+  return (__m64)(((__v1du)__a)[0] - ((__v1du)__b)[0]);
 }
 
 /// Subtracts the corresponding elements of two [2 x i64] vectors.
@@ -2556,8 +2567,8 @@ static __inline__ __m64 __DEFAULT_FN_ATTRS_MMX _mm_sub_si64(__m64 __a,
 ///    A 128-bit integer vector containing the subtrahends.
 /// \returns A 128-bit integer vector containing the differences of the values
 ///    in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sub_epi64(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_sub_epi64(__m128i __a, __m128i __b) {
   return (__m128i)((__v2du)__a - (__v2du)__b);
 }
 
@@ -2578,8 +2589,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sub_epi64(__m128i __a,
 ///    A 128-bit integer vector containing the subtrahends.
 /// \returns A 128-bit integer vector containing the differences of the values
 ///    in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_subs_epi8(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_subs_epi8(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_elementwise_sub_sat((__v16qs)__a, (__v16qs)__b);
 }
 
@@ -2600,8 +2611,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_subs_epi8(__m128i __a,
 ///    A 128-bit integer vector containing the subtrahends.
 /// \returns A 128-bit integer vector containing the differences of the values
 ///    in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_subs_epi16(__m128i __a,
-                                                            __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_subs_epi16(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_elementwise_sub_sat((__v8hi)__a, (__v8hi)__b);
 }
 
@@ -2621,8 +2632,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_subs_epi16(__m128i __a,
 ///    A 128-bit integer vector containing the subtrahends.
 /// \returns A 128-bit integer vector containing the unsigned integer
 ///    differences of the values in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_subs_epu8(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_subs_epu8(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_elementwise_sub_sat((__v16qu)__a, (__v16qu)__b);
 }
 
@@ -2642,8 +2653,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_subs_epu8(__m128i __a,
 ///    A 128-bit integer vector containing the subtrahends.
 /// \returns A 128-bit integer vector containing the unsigned integer
 ///    differences of the values in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_subs_epu16(__m128i __a,
-                                                            __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_subs_epu16(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_elementwise_sub_sat((__v8hu)__a, (__v8hu)__b);
 }
 
@@ -2659,8 +2670,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_subs_epu16(__m128i __a,
 ///    A 128-bit integer vector containing one of the source operands.
 /// \returns A 128-bit integer vector containing the bitwise AND of the values
 ///    in both operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_and_si128(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_and_si128(__m128i __a, __m128i __b) {
   return (__m128i)((__v2du)__a & (__v2du)__b);
 }
 
@@ -2678,8 +2689,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_and_si128(__m128i __a,
 ///    A 128-bit vector containing the right source operand.
 /// \returns A 128-bit integer vector containing the bitwise AND of the one's
 ///    complement of the first operand and the values in the second operand.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_andnot_si128(__m128i __a,
-                                                              __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_andnot_si128(__m128i __a, __m128i __b) {
   return (__m128i)(~(__v2du)__a & (__v2du)__b);
 }
 /// Performs a bitwise OR of two 128-bit integer vectors.
@@ -2694,8 +2705,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_andnot_si128(__m128i __a,
 ///    A 128-bit integer vector containing one of the source operands.
 /// \returns A 128-bit integer vector containing the bitwise OR of the values
 ///    in both operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_or_si128(__m128i __a,
-                                                          __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_or_si128(__m128i __a, __m128i __b) {
   return (__m128i)((__v2du)__a | (__v2du)__b);
 }
 
@@ -2711,8 +2722,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_or_si128(__m128i __a,
 ///    A 128-bit integer vector containing one of the source operands.
 /// \returns A 128-bit integer vector containing the bitwise exclusive OR of the
 ///    values in both operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_xor_si128(__m128i __a,
-                                                           __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_xor_si128(__m128i __a, __m128i __b) {
   return (__m128i)((__v2du)__a ^ (__v2du)__b);
 }
 
@@ -2734,11 +2745,11 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_xor_si128(__m128i __a,
 ///    \a a.
 /// \returns A 128-bit integer vector containing the left-shifted value.
 #define _mm_slli_si128(a, imm)                                                 \
-  ((__m128i)__builtin_ia32_pslldqi128_byteshift((__v2di)(__m128i)(a),          \
+  ((__m128i)__builtin_ia32_pslldqi128_byteshift((__v16qi)(__m128i)(a),         \
                                                 (int)(imm)))
 
 #define _mm_bslli_si128(a, imm)                                                \
-  ((__m128i)__builtin_ia32_pslldqi128_byteshift((__v2di)(__m128i)(a),          \
+  ((__m128i)__builtin_ia32_pslldqi128_byteshift((__v16qi)(__m128i)(a),         \
                                                 (int)(imm)))
 
 /// Left-shifts each 16-bit value in the 128-bit integer vector operand
@@ -2754,8 +2765,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_xor_si128(__m128i __a,
 ///    An integer value specifying the number of bits to left-shift each value
 ///    in operand \a __a.
 /// \returns A 128-bit integer vector containing the left-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_slli_epi16(__m128i __a,
-                                                            int __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_slli_epi16(__m128i __a, int __count) {
   return (__m128i)__builtin_ia32_psllwi128((__v8hi)__a, __count);
 }
 
@@ -2772,8 +2783,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_slli_epi16(__m128i __a,
 ///    A 128-bit integer vector in which bits [63:0] specify the number of bits
 ///    to left-shift each value in operand \a __a.
 /// \returns A 128-bit integer vector containing the left-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sll_epi16(__m128i __a,
-                                                           __m128i __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_sll_epi16(__m128i __a, __m128i __count) {
   return (__m128i)__builtin_ia32_psllw128((__v8hi)__a, (__v8hi)__count);
 }
 
@@ -2790,8 +2801,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sll_epi16(__m128i __a,
 ///    An integer value specifying the number of bits to left-shift each value
 ///    in operand \a __a.
 /// \returns A 128-bit integer vector containing the left-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_slli_epi32(__m128i __a,
-                                                            int __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_slli_epi32(__m128i __a, int __count) {
   return (__m128i)__builtin_ia32_pslldi128((__v4si)__a, __count);
 }
 
@@ -2808,8 +2819,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_slli_epi32(__m128i __a,
 ///    A 128-bit integer vector in which bits [63:0] specify the number of bits
 ///    to left-shift each value in operand \a __a.
 /// \returns A 128-bit integer vector containing the left-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sll_epi32(__m128i __a,
-                                                           __m128i __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_sll_epi32(__m128i __a, __m128i __count) {
   return (__m128i)__builtin_ia32_pslld128((__v4si)__a, (__v4si)__count);
 }
 
@@ -2826,8 +2837,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sll_epi32(__m128i __a,
 ///    An integer value specifying the number of bits to left-shift each value
 ///    in operand \a __a.
 /// \returns A 128-bit integer vector containing the left-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_slli_epi64(__m128i __a,
-                                                            int __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_slli_epi64(__m128i __a, int __count) {
   return __builtin_ia32_psllqi128((__v2di)__a, __count);
 }
 
@@ -2844,8 +2855,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_slli_epi64(__m128i __a,
 ///    A 128-bit integer vector in which bits [63:0] specify the number of bits
 ///    to left-shift each value in operand \a __a.
 /// \returns A 128-bit integer vector containing the left-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sll_epi64(__m128i __a,
-                                                           __m128i __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_sll_epi64(__m128i __a, __m128i __count) {
   return __builtin_ia32_psllq128((__v2di)__a, (__v2di)__count);
 }
 
@@ -2863,8 +2874,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sll_epi64(__m128i __a,
 ///    An integer value specifying the number of bits to right-shift each value
 ///    in operand \a __a.
 /// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srai_epi16(__m128i __a,
-                                                            int __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_srai_epi16(__m128i __a, int __count) {
   return (__m128i)__builtin_ia32_psrawi128((__v8hi)__a, __count);
 }
 
@@ -2882,8 +2893,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srai_epi16(__m128i __a,
 ///    A 128-bit integer vector in which bits [63:0] specify the number of bits
 ///    to right-shift each value in operand \a __a.
 /// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sra_epi16(__m128i __a,
-                                                           __m128i __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_sra_epi16(__m128i __a, __m128i __count) {
   return (__m128i)__builtin_ia32_psraw128((__v8hi)__a, (__v8hi)__count);
 }
 
@@ -2901,8 +2912,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sra_epi16(__m128i __a,
 ///    An integer value specifying the number of bits to right-shift each value
 ///    in operand \a __a.
 /// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srai_epi32(__m128i __a,
-                                                            int __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_srai_epi32(__m128i __a, int __count) {
   return (__m128i)__builtin_ia32_psradi128((__v4si)__a, __count);
 }
 
@@ -2920,8 +2931,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srai_epi32(__m128i __a,
 ///    A 128-bit integer vector in which bits [63:0] specify the number of bits
 ///    to right-shift each value in operand \a __a.
 /// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sra_epi32(__m128i __a,
-                                                           __m128i __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_sra_epi32(__m128i __a, __m128i __count) {
   return (__m128i)__builtin_ia32_psrad128((__v4si)__a, (__v4si)__count);
 }
 
@@ -2943,11 +2954,11 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sra_epi32(__m128i __a,
 ///    \a a.
 /// \returns A 128-bit integer vector containing the right-shifted value.
 #define _mm_srli_si128(a, imm)                                                 \
-  ((__m128i)__builtin_ia32_psrldqi128_byteshift((__v2di)(__m128i)(a),          \
+  ((__m128i)__builtin_ia32_psrldqi128_byteshift((__v16qi)(__m128i)(a),         \
                                                 (int)(imm)))
 
 #define _mm_bsrli_si128(a, imm)                                                \
-  ((__m128i)__builtin_ia32_psrldqi128_byteshift((__v2di)(__m128i)(a),          \
+  ((__m128i)__builtin_ia32_psrldqi128_byteshift((__v16qi)(__m128i)(a),         \
                                                 (int)(imm)))
 
 /// Right-shifts each of 16-bit values in the 128-bit integer vector
@@ -2963,8 +2974,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_sra_epi32(__m128i __a,
 ///    An integer value specifying the number of bits to right-shift each value
 ///    in operand \a __a.
 /// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srli_epi16(__m128i __a,
-                                                            int __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_srli_epi16(__m128i __a, int __count) {
   return (__m128i)__builtin_ia32_psrlwi128((__v8hi)__a, __count);
 }
 
@@ -2981,8 +2992,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srli_epi16(__m128i __a,
 ///    A 128-bit integer vector in which bits [63:0] specify the number of bits
 ///    to right-shift each value in operand \a __a.
 /// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srl_epi16(__m128i __a,
-                                                           __m128i __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_srl_epi16(__m128i __a, __m128i __count) {
   return (__m128i)__builtin_ia32_psrlw128((__v8hi)__a, (__v8hi)__count);
 }
 
@@ -2999,8 +3010,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srl_epi16(__m128i __a,
 ///    An integer value specifying the number of bits to right-shift each value
 ///    in operand \a __a.
 /// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srli_epi32(__m128i __a,
-                                                            int __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_srli_epi32(__m128i __a, int __count) {
   return (__m128i)__builtin_ia32_psrldi128((__v4si)__a, __count);
 }
 
@@ -3017,8 +3028,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srli_epi32(__m128i __a,
 ///    A 128-bit integer vector in which bits [63:0] specify the number of bits
 ///    to right-shift each value in operand \a __a.
 /// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srl_epi32(__m128i __a,
-                                                           __m128i __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_srl_epi32(__m128i __a, __m128i __count) {
   return (__m128i)__builtin_ia32_psrld128((__v4si)__a, (__v4si)__count);
 }
 
@@ -3035,8 +3046,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srl_epi32(__m128i __a,
 ///    An integer value specifying the number of bits to right-shift each value
 ///    in operand \a __a.
 /// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srli_epi64(__m128i __a,
-                                                            int __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_srli_epi64(__m128i __a, int __count) {
   return __builtin_ia32_psrlqi128((__v2di)__a, __count);
 }
 
@@ -3053,8 +3064,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srli_epi64(__m128i __a,
 ///    A 128-bit integer vector in which bits [63:0] specify the number of bits
 ///    to right-shift each value in operand \a __a.
 /// \returns A 128-bit integer vector containing the right-shifted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srl_epi64(__m128i __a,
-                                                           __m128i __count) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_srl_epi64(__m128i __a, __m128i __count) {
   return __builtin_ia32_psrlq128((__v2di)__a, (__v2di)__count);
 }
 
@@ -3072,8 +3083,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_srl_epi64(__m128i __a,
 /// \param __b
 ///    A 128-bit integer vector.
 /// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpeq_epi8(__m128i __a,
-                                                            __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cmpeq_epi8(__m128i __a, __m128i __b) {
   return (__m128i)((__v16qi)__a == (__v16qi)__b);
 }
 
@@ -3091,8 +3102,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpeq_epi8(__m128i __a,
 /// \param __b
 ///    A 128-bit integer vector.
 /// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpeq_epi16(__m128i __a,
-                                                             __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cmpeq_epi16(__m128i __a, __m128i __b) {
   return (__m128i)((__v8hi)__a == (__v8hi)__b);
 }
 
@@ -3110,8 +3121,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpeq_epi16(__m128i __a,
 /// \param __b
 ///    A 128-bit integer vector.
 /// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpeq_epi32(__m128i __a,
-                                                             __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cmpeq_epi32(__m128i __a, __m128i __b) {
   return (__m128i)((__v4si)__a == (__v4si)__b);
 }
 
@@ -3130,8 +3141,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpeq_epi32(__m128i __a,
 /// \param __b
 ///    A 128-bit integer vector.
 /// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpgt_epi8(__m128i __a,
-                                                            __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cmpgt_epi8(__m128i __a, __m128i __b) {
   /* This function always performs a signed comparison, but __v16qi is a char
      which may be signed or unsigned, so use __v16qs. */
   return (__m128i)((__v16qs)__a > (__v16qs)__b);
@@ -3152,8 +3163,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpgt_epi8(__m128i __a,
 /// \param __b
 ///    A 128-bit integer vector.
 /// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpgt_epi16(__m128i __a,
-                                                             __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cmpgt_epi16(__m128i __a, __m128i __b) {
   return (__m128i)((__v8hi)__a > (__v8hi)__b);
 }
 
@@ -3172,8 +3183,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpgt_epi16(__m128i __a,
 /// \param __b
 ///    A 128-bit integer vector.
 /// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpgt_epi32(__m128i __a,
-                                                             __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cmpgt_epi32(__m128i __a, __m128i __b) {
   return (__m128i)((__v4si)__a > (__v4si)__b);
 }
 
@@ -3192,8 +3203,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmpgt_epi32(__m128i __a,
 /// \param __b
 ///    A 128-bit integer vector.
 /// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmplt_epi8(__m128i __a,
-                                                            __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cmplt_epi8(__m128i __a, __m128i __b) {
   return _mm_cmpgt_epi8(__b, __a);
 }
 
@@ -3212,8 +3223,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmplt_epi8(__m128i __a,
 /// \param __b
 ///    A 128-bit integer vector.
 /// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmplt_epi16(__m128i __a,
-                                                             __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cmplt_epi16(__m128i __a, __m128i __b) {
   return _mm_cmpgt_epi16(__b, __a);
 }
 
@@ -3232,8 +3243,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmplt_epi16(__m128i __a,
 /// \param __b
 ///    A 128-bit integer vector.
 /// \returns A 128-bit integer vector containing the comparison results.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmplt_epi32(__m128i __a,
-                                                             __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cmplt_epi32(__m128i __a, __m128i __b) {
   return _mm_cmpgt_epi32(__b, __a);
 }
 
@@ -3255,8 +3266,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cmplt_epi32(__m128i __a,
 /// \returns A 128-bit vector of [2 x double] 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__ __m128d __DEFAULT_FN_ATTRS _mm_cvtsi64_sd(__m128d __a,
-                                                            long long __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtsi64_sd(__m128d __a, long long __b) {
   __a[0] = __b;
   return __a;
 }
@@ -3310,7 +3321,8 @@ static __inline__ long long __DEFAULT_FN_ATTRS _mm_cvttsd_si64(__m128d __a) {
 /// \param __a
 ///    A 128-bit integer vector.
 /// \returns A 128-bit vector of [4 x float] containing the converted values.
-static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_cvtepi32_ps(__m128i __a) {
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtepi32_ps(__m128i __a) {
   return (__m128) __builtin_convertvector((__v4si)__a, __v4sf);
 }
 
@@ -3361,7 +3373,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvttps_epi32(__m128 __a) {
 /// \param __a
 ///    A 32-bit signed integer operand.
 /// \returns A 128-bit vector of [4 x i32].
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtsi32_si128(int __a) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtsi32_si128(int __a) {
   return __extension__(__m128i)(__v4si){__a, 0, 0, 0};
 }
 
@@ -3376,7 +3389,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtsi32_si128(int __a) {
 /// \param __a
 ///    A 64-bit signed integer operand containing the value to be converted.
 /// \returns A 128-bit vector of [2 x i64] containing the converted value.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtsi64_si128(long long __a) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtsi64_si128(long long __a) {
   return __extension__(__m128i)(__v2di){__a, 0};
 }
 
@@ -3391,7 +3405,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_cvtsi64_si128(long long __a) {
 ///    A vector of [4 x i32]. The least significant 32 bits are moved to the
 ///    destination.
 /// \returns A 32-bit signed integer containing the moved value.
-static __inline__ int __DEFAULT_FN_ATTRS _mm_cvtsi128_si32(__m128i __a) {
+static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtsi128_si32(__m128i __a) {
   __v4si __b = (__v4si)__a;
   return __b[0];
 }
@@ -3407,7 +3422,8 @@ static __inline__ int __DEFAULT_FN_ATTRS _mm_cvtsi128_si32(__m128i __a) {
 ///    A vector of [2 x i64]. The least significant 64 bits are moved to the
 ///    destination.
 /// \returns A 64-bit signed integer containing the moved value.
-static __inline__ long long __DEFAULT_FN_ATTRS _mm_cvtsi128_si64(__m128i __a) {
+static __inline__ long long __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_cvtsi128_si64(__m128i __a) {
   return __a[0];
 }
 
@@ -3494,8 +3510,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_undefined_si128(void) {
 ///    destination vector of [2 x i64].
 /// \returns An initialized 128-bit vector of [2 x i64] containing the values
 ///    provided in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set_epi64x(long long __q1,
-                                                            long long __q0) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_set_epi64x(long long __q1, long long __q0) {
   return __extension__(__m128i)(__v2di){__q0, __q1};
 }
 
@@ -3515,9 +3531,9 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set_epi64x(long long __q1,
 ///    destination vector of [2 x i64].
 /// \returns An initialized 128-bit vector of [2 x i64] containing the values
 ///    provided in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set_epi64(__m64 __q1,
-                                                           __m64 __q0) {
-  return _mm_set_epi64x((long long)__q1, (long long)__q0);
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_set_epi64(__m64 __q1, __m64 __q0) {
+  return _mm_set_epi64x((long long)__q1[0], (long long)__q0[0]);
 }
 
 /// Initializes the 32-bit values in a 128-bit vector of [4 x i32] with
@@ -3542,8 +3558,10 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set_epi64(__m64 __q1,
 ///    vector.
 /// \returns An initialized 128-bit vector of [4 x i32] containing the values
 ///    provided in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set_epi32(int __i3, int __i2,
-                                                           int __i1, int __i0) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR _mm_set_epi32(int __i3,
+                                                                     int __i2,
+                                                                     int __i1,
+                                                                     int __i0) {
   return __extension__(__m128i)(__v4si){__i0, __i1, __i2, __i3};
 }
 
@@ -3581,7 +3599,7 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set_epi32(int __i3, int __i2,
 ///    vector.
 /// \returns An initialized 128-bit vector of [8 x i16] containing the values
 ///    provided in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
 _mm_set_epi16(short __w7, short __w6, short __w5, short __w4, short __w3,
               short __w2, short __w1, short __w0) {
   return __extension__(__m128i)(__v8hi){__w0, __w1, __w2, __w3,
@@ -3630,7 +3648,7 @@ _mm_set_epi16(short __w7, short __w6, short __w5, short __w4, short __w3,
 ///    Initializes bits [7:0] of the destination vector.
 /// \returns An initialized 128-bit vector of [16 x i8] containing the values
 ///    provided in the operands.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
 _mm_set_epi8(char __b15, char __b14, char __b13, char __b12, char __b11,
              char __b10, char __b9, char __b8, char __b7, char __b6, char __b5,
              char __b4, char __b3, char __b2, char __b1, char __b0) {
@@ -3652,7 +3670,8 @@ _mm_set_epi8(char __b15, char __b14, char __b13, char __b12, char __b11,
 ///    vector.
 /// \returns An initialized 128-bit integer vector of [2 x i64] with both
 ///    elements containing the value provided in the operand.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set1_epi64x(long long __q) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_set1_epi64x(long long __q) {
   return _mm_set_epi64x(__q, __q);
 }
 
@@ -3669,7 +3688,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set1_epi64x(long long __q) {
 ///    vector.
 /// \returns An initialized 128-bit vector of [2 x i64] with all elements
 ///    containing the value provided in the operand.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set1_epi64(__m64 __q) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_set1_epi64(__m64 __q) {
   return _mm_set_epi64(__q, __q);
 }
 
@@ -3686,7 +3706,7 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set1_epi64(__m64 __q) {
 ///    vector.
 /// \returns An initialized 128-bit vector of [4 x i32] with all elements
 ///    containing the value provided in the operand.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set1_epi32(int __i) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR _mm_set1_epi32(int __i) {
   return _mm_set_epi32(__i, __i, __i, __i);
 }
 
@@ -3703,7 +3723,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set1_epi32(int __i) {
 ///    vector.
 /// \returns An initialized 128-bit vector of [8 x i16] with all elements
 ///    containing the value provided in the operand.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set1_epi16(short __w) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_set1_epi16(short __w) {
   return _mm_set_epi16(__w, __w, __w, __w, __w, __w, __w, __w);
 }
 
@@ -3720,7 +3741,7 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set1_epi16(short __w) {
 ///    vector.
 /// \returns An initialized 128-bit vector of [16 x i8] with all elements
 ///    containing the value provided in the operand.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set1_epi8(char __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR _mm_set1_epi8(char __b) {
   return _mm_set_epi8(__b, __b, __b, __b, __b, __b, __b, __b, __b, __b, __b,
                       __b, __b, __b, __b, __b);
 }
@@ -3739,8 +3760,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_set1_epi8(char __b) {
 ///    A 64-bit integral value used to initialize the upper 64 bits of the
 ///    result.
 /// \returns An initialized 128-bit integer vector.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_setr_epi64(__m64 __q0,
-                                                            __m64 __q1) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_setr_epi64(__m64 __q0, __m64 __q1) {
   return _mm_set_epi64(__q1, __q0);
 }
 
@@ -3761,9 +3782,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_setr_epi64(__m64 __q0,
 /// \param __i3
 ///    A 32-bit integral value used to initialize bits [127:96] of the result.
 /// \returns An initialized 128-bit integer vector.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_setr_epi32(int __i0, int __i1,
-                                                            int __i2,
-                                                            int __i3) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_setr_epi32(int __i0, int __i1, int __i2, int __i3) {
   return _mm_set_epi32(__i3, __i2, __i1, __i0);
 }
 
@@ -3792,7 +3812,7 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_setr_epi32(int __i0, int __i1,
 /// \param __w7
 ///    A 16-bit integral value used to initialize bits [127:112] of the result.
 /// \returns An initialized 128-bit integer vector.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
 _mm_setr_epi16(short __w0, short __w1, short __w2, short __w3, short __w4,
                short __w5, short __w6, short __w7) {
   return _mm_set_epi16(__w7, __w6, __w5, __w4, __w3, __w2, __w1, __w0);
@@ -3839,7 +3859,7 @@ _mm_setr_epi16(short __w0, short __w1, short __w2, short __w3, short __w4,
 /// \param __b15
 ///    An 8-bit integral value used to initialize bits [127:120] of the result.
 /// \returns An initialized 128-bit integer vector.
-static __inline__ __m128i __DEFAULT_FN_ATTRS
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
 _mm_setr_epi8(char __b0, char __b1, char __b2, char __b3, char __b4, char __b5,
               char __b6, char __b7, char __b8, char __b9, char __b10,
               char __b11, char __b12, char __b13, char __b14, char __b15) {
@@ -3855,7 +3875,7 @@ _mm_setr_epi8(char __b0, char __b1, char __b2, char __b3, char __b4, char __b5,
 ///
 /// \returns An initialized 128-bit integer vector with all elements set to
 ///    zero.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_setzero_si128(void) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR _mm_setzero_si128(void) {
   return __extension__(__m128i)(__v2di){0LL, 0LL};
 }
 
@@ -4139,8 +4159,8 @@ void _mm_mfence(void);
 ///   A 128-bit integer vector of [8 x i16]. The converted [8 x i8] values are
 ///   written to the higher 64 bits of the result.
 /// \returns A 128-bit vector of [16 x i8] containing the converted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_packs_epi16(__m128i __a,
-                                                             __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_packs_epi16(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_ia32_packsswb128((__v8hi)__a, (__v8hi)__b);
 }
 
@@ -4162,8 +4182,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_packs_epi16(__m128i __a,
 ///    A 128-bit integer vector of [4 x i32]. The converted [4 x i16] values
 ///    are written to the higher 64 bits of the result.
 /// \returns A 128-bit vector of [8 x i16] containing the converted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_packs_epi32(__m128i __a,
-                                                             __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_packs_epi32(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_ia32_packssdw128((__v4si)__a, (__v4si)__b);
 }
 
@@ -4185,8 +4205,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_packs_epi32(__m128i __a,
 ///    A 128-bit integer vector of [8 x i16]. The converted [8 x i8] values are
 ///    written to the higher 64 bits of the result.
 /// \returns A 128-bit vector of [16 x i8] containing the converted values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_packus_epi16(__m128i __a,
-                                                              __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_packus_epi16(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_ia32_packuswb128((__v8hi)__a, (__v8hi)__b);
 }
 
@@ -4260,7 +4280,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_packus_epi16(__m128i __a,
 ///    A 128-bit integer vector containing the values with bits to be extracted.
 /// \returns The most significant bits from each 8-bit element in \a __a,
 ///    written to bits [15:0]. The other bits are assigned zeros.
-static __inline__ int __DEFAULT_FN_ATTRS _mm_movemask_epi8(__m128i __a) {
+static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_movemask_epi8(__m128i __a) {
   return __builtin_ia32_pmovmskb128((__v16qi)__a);
 }
 
@@ -4393,8 +4414,8 @@ static __inline__ int __DEFAULT_FN_ATTRS _mm_movemask_epi8(__m128i __a) {
 ///    Bits [119:112] are written to bits [111:104] of the result. \n
 ///    Bits [127:120] are written to bits [127:120] of the result.
 /// \returns A 128-bit vector of [16 x i8] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpackhi_epi8(__m128i __a,
-                                                               __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_unpackhi_epi8(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_shufflevector(
       (__v16qi)__a, (__v16qi)__b, 8, 16 + 8, 9, 16 + 9, 10, 16 + 10, 11,
       16 + 11, 12, 16 + 12, 13, 16 + 13, 14, 16 + 14, 15, 16 + 15);
@@ -4421,8 +4442,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpackhi_epi8(__m128i __a,
 ///    Bits [111:96] are written to bits [95:80] of the result. \n
 ///    Bits [127:112] are written to bits [127:112] of the result.
 /// \returns A 128-bit vector of [8 x i16] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpackhi_epi16(__m128i __a,
-                                                                __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_unpackhi_epi16(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_shufflevector((__v8hi)__a, (__v8hi)__b, 4, 8 + 4, 5,
                                           8 + 5, 6, 8 + 6, 7, 8 + 7);
 }
@@ -4444,8 +4465,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpackhi_epi16(__m128i __a,
 ///    Bits [95:64] are written to bits [64:32] of the destination. \n
 ///    Bits [127:96] are written to bits [127:96] of the destination.
 /// \returns A 128-bit vector of [4 x i32] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpackhi_epi32(__m128i __a,
-                                                                __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_unpackhi_epi32(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_shufflevector((__v4si)__a, (__v4si)__b, 2, 4 + 2, 3,
                                           4 + 3);
 }
@@ -4465,8 +4486,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpackhi_epi32(__m128i __a,
 ///    A 128-bit vector of [2 x i64]. \n
 ///    Bits [127:64] are written to bits [127:64] of the destination.
 /// \returns A 128-bit vector of [2 x i64] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpackhi_epi64(__m128i __a,
-                                                                __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_unpackhi_epi64(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_shufflevector((__v2di)__a, (__v2di)__b, 1, 2 + 1);
 }
 
@@ -4499,8 +4520,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpackhi_epi64(__m128i __a,
 ///    Bits [55:48] are written to bits [111:104] of the result. \n
 ///    Bits [63:56] are written to bits [127:120] of the result.
 /// \returns A 128-bit vector of [16 x i8] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpacklo_epi8(__m128i __a,
-                                                               __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_unpacklo_epi8(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_shufflevector(
       (__v16qi)__a, (__v16qi)__b, 0, 16 + 0, 1, 16 + 1, 2, 16 + 2, 3, 16 + 3, 4,
       16 + 4, 5, 16 + 5, 6, 16 + 6, 7, 16 + 7);
@@ -4528,8 +4549,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpacklo_epi8(__m128i __a,
 ///    Bits [47:32] are written to bits [95:80] of the result. \n
 ///    Bits [63:48] are written to bits [127:112] of the result.
 /// \returns A 128-bit vector of [8 x i16] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpacklo_epi16(__m128i __a,
-                                                                __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_unpacklo_epi16(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_shufflevector((__v8hi)__a, (__v8hi)__b, 0, 8 + 0, 1,
                                           8 + 1, 2, 8 + 2, 3, 8 + 3);
 }
@@ -4551,8 +4572,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpacklo_epi16(__m128i __a,
 ///    Bits [31:0] are written to bits [64:32] of the destination. \n
 ///    Bits [63:32] are written to bits [127:96] of the destination.
 /// \returns A 128-bit vector of [4 x i32] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpacklo_epi32(__m128i __a,
-                                                                __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_unpacklo_epi32(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_shufflevector((__v4si)__a, (__v4si)__b, 0, 4 + 0, 1,
                                           4 + 1);
 }
@@ -4572,8 +4593,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpacklo_epi32(__m128i __a,
 ///    A 128-bit vector of [2 x i64]. \n
 ///    Bits [63:0] are written to bits [127:64] of the destination. \n
 /// \returns A 128-bit vector of [2 x i64] containing the interleaved values.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpacklo_epi64(__m128i __a,
-                                                                __m128i __b) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_unpacklo_epi64(__m128i __a, __m128i __b) {
   return (__m128i)__builtin_shufflevector((__v2di)__a, (__v2di)__b, 0, 2 + 0);
 }
 
@@ -4588,7 +4609,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_unpacklo_epi64(__m128i __a,
 ///    A 128-bit integer vector operand. The lower 64 bits are moved to the
 ///    destination.
 /// \returns A 64-bit integer containing the lower 64 bits of the parameter.
-static __inline__ __m64 __DEFAULT_FN_ATTRS _mm_movepi64_pi64(__m128i __a) {
+static __inline__ __m64 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_movepi64_pi64(__m128i __a) {
   return (__m64)__a[0];
 }
 
@@ -4603,8 +4625,9 @@ static __inline__ __m64 __DEFAULT_FN_ATTRS _mm_movepi64_pi64(__m128i __a) {
 ///    A 64-bit value.
 /// \returns A 128-bit integer vector. The lower 64 bits contain the value from
 ///    the operand. The upper 64 bits are assigned zeros.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_movpi64_epi64(__m64 __a) {
-  return __extension__(__m128i)(__v2di){(long long)__a, 0};
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_movpi64_epi64(__m64 __a) {
+  return __builtin_shufflevector((__v1di)__a, _mm_setzero_si64(), 0, 1);
 }
 
 /// Moves the lower 64 bits of a 128-bit integer vector to a 128-bit
@@ -4619,7 +4642,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_movpi64_epi64(__m64 __a) {
 ///    destination.
 /// \returns A 128-bit integer vector. The lower 64 bits contain the value from
 ///    the operand. The upper 64 bits are assigned zeros.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_move_epi64(__m128i __a) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_move_epi64(__m128i __a) {
   return __builtin_shufflevector((__v2di)__a, _mm_setzero_si128(), 0, 2);
 }
 
@@ -4638,8 +4662,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_move_epi64(__m128i __a) {
 ///    A 128-bit vector of [2 x double]. \n
 ///    Bits [127:64] are written to bits [127:64] of the destination.
 /// \returns A 128-bit vector of [2 x double] containing the interleaved values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_unpackhi_pd(__m128d __a,
-                                                             __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_unpackhi_pd(__m128d __a, __m128d __b) {
   return __builtin_shufflevector((__v2df)__a, (__v2df)__b, 1, 2 + 1);
 }
 
@@ -4658,8 +4682,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_unpackhi_pd(__m128d __a,
 ///    A 128-bit vector of [2 x double]. \n
 ///    Bits [63:0] are written to bits [127:64] of the destination.
 /// \returns A 128-bit vector of [2 x double] containing the interleaved values.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_unpacklo_pd(__m128d __a,
-                                                             __m128d __b) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_unpacklo_pd(__m128d __a, __m128d __b) {
   return __builtin_shufflevector((__v2df)__a, (__v2df)__b, 0, 2 + 0);
 }
 
@@ -4676,7 +4700,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_unpacklo_pd(__m128d __a,
 ///    be extracted.
 /// \returns The sign bits from each of the double-precision elements in \a __a,
 ///    written to bits [1:0]. The remaining bits are assigned values of zero.
-static __inline__ int __DEFAULT_FN_ATTRS _mm_movemask_pd(__m128d __a) {
+static __inline__ int __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_movemask_pd(__m128d __a) {
   return __builtin_ia32_movmskpd((__v2df)__a);
 }
 
@@ -4722,7 +4747,8 @@ static __inline__ int __DEFAULT_FN_ATTRS _mm_movemask_pd(__m128d __a) {
 ///    A 128-bit floating-point vector of [2 x double].
 /// \returns A 128-bit floating-point vector of [4 x float] containing the same
 ///    bitwise pattern as the parameter.
-static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_castpd_ps(__m128d __a) {
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_castpd_ps(__m128d __a) {
   return (__m128)__a;
 }
 
@@ -4737,7 +4763,8 @@ static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_castpd_ps(__m128d __a) {
 ///    A 128-bit floating-point vector of [2 x double].
 /// \returns A 128-bit integer vector containing the same bitwise pattern as the
 ///    parameter.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_castpd_si128(__m128d __a) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_castpd_si128(__m128d __a) {
   return (__m128i)__a;
 }
 
@@ -4752,7 +4779,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_castpd_si128(__m128d __a) {
 ///    A 128-bit floating-point vector of [4 x float].
 /// \returns A 128-bit floating-point vector of [2 x double] containing the same
 ///    bitwise pattern as the parameter.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_castps_pd(__m128 __a) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_castps_pd(__m128 __a) {
   return (__m128d)__a;
 }
 
@@ -4767,7 +4795,8 @@ static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_castps_pd(__m128 __a) {
 ///    A 128-bit floating-point vector of [4 x float].
 /// \returns A 128-bit integer vector containing the same bitwise pattern as the
 ///    parameter.
-static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_castps_si128(__m128 __a) {
+static __inline__ __m128i __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_castps_si128(__m128 __a) {
   return (__m128i)__a;
 }
 
@@ -4782,7 +4811,8 @@ static __inline__ __m128i __DEFAULT_FN_ATTRS _mm_castps_si128(__m128 __a) {
 ///    A 128-bit integer vector.
 /// \returns A 128-bit floating-point vector of [4 x float] containing the same
 ///    bitwise pattern as the parameter.
-static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_castsi128_ps(__m128i __a) {
+static __inline__ __m128 __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_castsi128_ps(__m128i __a) {
   return (__m128)__a;
 }
 
@@ -4797,7 +4827,8 @@ static __inline__ __m128 __DEFAULT_FN_ATTRS _mm_castsi128_ps(__m128i __a) {
 ///    A 128-bit integer vector.
 /// \returns A 128-bit floating-point vector of [2 x double] containing the same
 ///    bitwise pattern as the parameter.
-static __inline__ __m128d __DEFAULT_FN_ATTRS _mm_castsi128_pd(__m128i __a) {
+static __inline__ __m128d __DEFAULT_FN_ATTRS_CONSTEXPR
+_mm_castsi128_pd(__m128i __a) {
   return (__m128d)__a;
 }
 
@@ -4889,8 +4920,11 @@ void _mm_pause(void);
 #if defined(__cplusplus)
 } // extern "C"
 #endif
+
+#undef __anyext128
+#undef __trunc64
 #undef __DEFAULT_FN_ATTRS
-#undef __DEFAULT_FN_ATTRS_MMX
+#undef __DEFAULT_FN_ATTRS_CONSTEXPR
 
 #define _MM_SHUFFLE2(x, y) (((x) << 1) | (y))