1 /* 2 * Copyright (c) 2016-present, Yann Collet, Facebook, Inc. 3 * All rights reserved. 4 * 5 * This source code is licensed under both the BSD-style license (found in the 6 * LICENSE file in the root directory of this source tree) and the GPLv2 (found 7 * in the COPYING file in the root directory of this source tree). 8 * You may select, at your option, one of the above-listed licenses. 9 */ 10 11 #ifndef ZSTD_COMPILER_H 12 #define ZSTD_COMPILER_H 13 14 /*-******************************************************* 15 * Compiler specifics 16 *********************************************************/ 17 /* force inlining */ 18 #if defined (__GNUC__) || defined(__cplusplus) || defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 */ 19 # define INLINE_KEYWORD inline 20 #else 21 # define INLINE_KEYWORD 22 #endif 23 24 #if defined(__GNUC__) 25 # define FORCE_INLINE_ATTR __attribute__((always_inline)) 26 #elif defined(_MSC_VER) 27 # define FORCE_INLINE_ATTR __forceinline 28 #else 29 # define FORCE_INLINE_ATTR 30 #endif 31 32 /** 33 * FORCE_INLINE_TEMPLATE is used to define C "templates", which take constant 34 * parameters. They must be inlined for the compiler to elimininate the constant 35 * branches. 36 */ 37 #define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR 38 /** 39 * HINT_INLINE is used to help the compiler generate better code. It is *not* 40 * used for "templates", so it can be tweaked based on the compilers 41 * performance. 42 * 43 * gcc-4.8 and gcc-4.9 have been shown to benefit from leaving off the 44 * always_inline attribute. 45 * 46 * clang up to 5.0.0 (trunk) benefit tremendously from the always_inline 47 * attribute. 48 */ 49 #if !defined(__clang__) && defined(__GNUC__) && __GNUC__ >= 4 && __GNUC_MINOR__ >= 8 && __GNUC__ < 5 50 # define HINT_INLINE static INLINE_KEYWORD 51 #else 52 # define HINT_INLINE static INLINE_KEYWORD FORCE_INLINE_ATTR 53 #endif 54 55 /* force no inlining */ 56 #ifdef _MSC_VER 57 # define FORCE_NOINLINE static __declspec(noinline) 58 #else 59 # ifdef __GNUC__ 60 # define FORCE_NOINLINE static __attribute__((__noinline__)) 61 # else 62 # define FORCE_NOINLINE static 63 # endif 64 #endif 65 66 /* target attribute */ 67 #ifndef __has_attribute 68 #define __has_attribute(x) 0 /* Compatibility with non-clang compilers. */ 69 #endif 70 #if defined(__GNUC__) 71 # define TARGET_ATTRIBUTE(target) __attribute__((__target__(target))) 72 #else 73 # define TARGET_ATTRIBUTE(target) 74 #endif 75 76 /* Enable runtime BMI2 dispatch based on the CPU. 77 * Enabled for clang & gcc >=4.8 on x86 when BMI2 isn't enabled by default. 78 */ 79 #ifndef DYNAMIC_BMI2 80 #if ((defined(__clang__) && __has_attribute(__target__)) \ 81 || (defined(__GNUC__) \ 82 && (__GNUC__ >= 5 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 8)))) \ 83 && (defined(__x86_64__) || defined(_M_X86)) \ 84 && !defined(__BMI2__) 85 # define DYNAMIC_BMI2 1 86 #else 87 # define DYNAMIC_BMI2 0 88 #endif 89 #endif 90 91 /* prefetch 92 * can be disabled, by declaring NO_PREFETCH macro 93 * All prefetch invocations use a single default locality 2, 94 * generating instruction prefetcht1, 95 * which, according to Intel, means "load data into L2 cache". 96 * This is a good enough "middle ground" for the time being, 97 * though in theory, it would be better to specialize locality depending on data being prefetched. 98 * Tests could not determine any sensible difference based on locality value. */ 99 #if defined(NO_PREFETCH) 100 # define PREFETCH(ptr) (void)(ptr) /* disabled */ 101 #else 102 # if defined(_MSC_VER) && (defined(_M_X64) || defined(_M_I86)) /* _mm_prefetch() is not defined outside of x86/x64 */ 103 # include <mmintrin.h> /* https://msdn.microsoft.com/fr-fr/library/84szxsww(v=vs.90).aspx */ 104 # define PREFETCH(ptr) _mm_prefetch((const char*)(ptr), _MM_HINT_T1) 105 # elif defined(__GNUC__) && ( (__GNUC__ >= 4) || ( (__GNUC__ == 3) && (__GNUC_MINOR__ >= 1) ) ) 106 # define PREFETCH(ptr) __builtin_prefetch((ptr), 0 /* rw==read */, 2 /* locality */) 107 # else 108 # define PREFETCH(ptr) (void)(ptr) /* disabled */ 109 # endif 110 #endif /* NO_PREFETCH */ 111 112 #define CACHELINE_SIZE 64 113 114 #define PREFETCH_AREA(p, s) { \ 115 const char* const _ptr = (const char*)(p); \ 116 size_t const _size = (size_t)(s); \ 117 size_t _pos; \ 118 for (_pos=0; _pos<_size; _pos+=CACHELINE_SIZE) { \ 119 PREFETCH(_ptr + _pos); \ 120 } \ 121 } 122 123 /* disable warnings */ 124 #ifdef _MSC_VER /* Visual Studio */ 125 # include <intrin.h> /* For Visual 2005 */ 126 # pragma warning(disable : 4100) /* disable: C4100: unreferenced formal parameter */ 127 # pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ 128 # pragma warning(disable : 4204) /* disable: C4204: non-constant aggregate initializer */ 129 # pragma warning(disable : 4214) /* disable: C4214: non-int bitfields */ 130 # pragma warning(disable : 4324) /* disable: C4324: padded structure */ 131 #endif 132 133 #endif /* ZSTD_COMPILER_H */ 134