Home | History | Annotate | Download | only in include 
      1 /*===---- bmi2intrin.h - BMI2 intrinsics -----------------------------------===
      2  *
      3  * Permission is hereby granted, free of charge, to any person obtaining a copy
      4  * of this software and associated documentation files (the "Software"), to deal
      5  * in the Software without restriction, including without limitation the rights
      6  * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
      7  * copies of the Software, and to permit persons to whom the Software is
      8  * furnished to do so, subject to the following conditions:
      9  *
     10  * The above copyright notice and this permission notice shall be included in
     11  * all copies or substantial portions of the Software.
     12  *
     13  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
     14  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
     15  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
     16  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
     17  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
     18  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
     19  * THE SOFTWARE.
     20  *
     21  *===-----------------------------------------------------------------------===
     22  */
     23 
     24 #if !defined __X86INTRIN_H && !defined __IMMINTRIN_H
     25 #error "Never use <bmi2intrin.h> directly; include <x86intrin.h> instead."
     26 #endif
     27 
     28 #ifndef __BMI2__
     29 # error "BMI2 instruction set not enabled"
     30 #endif /* __BMI2__ */
     31 
     32 #ifndef __BMI2INTRIN_H
     33 #define __BMI2INTRIN_H
     34 
     35 static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
     36 _bzhi_u32(unsigned int __X, unsigned int __Y)
     37 {
     38   return __builtin_ia32_bzhi_si(__X, __Y);
     39 }
     40 
     41 static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
     42 _pdep_u32(unsigned int __X, unsigned int __Y)
     43 {
     44   return __builtin_ia32_pdep_si(__X, __Y);
     45 }
     46 
     47 static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
     48 _pext_u32(unsigned int __X, unsigned int __Y)
     49 {
     50   return __builtin_ia32_pext_si(__X, __Y);
     51 }
     52 
     53 #ifdef  __x86_64__
     54 
     55 static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
     56 _bzhi_u64(unsigned long long __X, unsigned long long __Y)
     57 {
     58   return __builtin_ia32_bzhi_di(__X, __Y);
     59 }
     60 
     61 static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
     62 _pdep_u64(unsigned long long __X, unsigned long long __Y)
     63 {
     64   return __builtin_ia32_pdep_di(__X, __Y);
     65 }
     66 
     67 static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
     68 _pext_u64(unsigned long long __X, unsigned long long __Y)
     69 {
     70   return __builtin_ia32_pext_di(__X, __Y);
     71 }
     72 
     73 static __inline__ unsigned long long __attribute__((__always_inline__, __nodebug__))
     74 _mulx_u64 (unsigned long long __X, unsigned long long __Y,
     75 	   unsigned long long *__P)
     76 {
     77   unsigned __int128 __res = (unsigned __int128) __X * __Y;
     78   *__P = (unsigned long long) (__res >> 64);
     79   return (unsigned long long) __res;
     80 }
     81 
     82 #else /* !__x86_64__ */
     83 
     84 static __inline__ unsigned int __attribute__((__always_inline__, __nodebug__))
     85 _mulx_u32 (unsigned int __X, unsigned int __Y, unsigned int *__P)
     86 {
     87   unsigned long long __res = (unsigned long long) __X * __Y;
     88   *__P = (unsigned int) (__res >> 32);
     89   return (unsigned int) __res;
     90 }
     91 
     92 #endif /* !__x86_64__  */
     93 
     94 #endif /* __BMI2INTRIN_H */
     95