1 | /* |
2 | * Copyright (c) 2003, 2019, Oracle and/or its affiliates. All rights reserved. |
3 | * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. |
4 | * |
5 | * This code is free software; you can redistribute it and/or modify it |
6 | * under the terms of the GNU General Public License version 2 only, as |
7 | * published by the Free Software Foundation. |
8 | * |
9 | * This code is distributed in the hope that it will be useful, but WITHOUT |
10 | * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or |
11 | * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
12 | * version 2 for more details (a copy is included in the LICENSE file that |
13 | * accompanied this code). |
14 | * |
15 | * You should have received a copy of the GNU General Public License version |
16 | * 2 along with this work; if not, write to the Free Software Foundation, |
17 | * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. |
18 | * |
19 | * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA |
20 | * or visit www.oracle.com if you need additional information or have any |
21 | * questions. |
22 | * |
23 | */ |
24 | |
25 | #ifndef OS_CPU_LINUX_X86_ORDERACCESS_LINUX_X86_HPP |
26 | #define OS_CPU_LINUX_X86_ORDERACCESS_LINUX_X86_HPP |
27 | |
28 | // Included in orderAccess.hpp header file. |
29 | |
30 | // Compiler version last used for testing: gcc 4.8.2 |
31 | // Please update this information when this file changes |
32 | |
33 | // Implementation of class OrderAccess. |
34 | |
35 | // A compiler barrier, forcing the C++ compiler to invalidate all memory assumptions |
36 | static inline void compiler_barrier() { |
37 | __asm__ volatile ("" : : : "memory" ); |
38 | } |
39 | |
40 | inline void OrderAccess::loadload() { compiler_barrier(); } |
41 | inline void OrderAccess::storestore() { compiler_barrier(); } |
42 | inline void OrderAccess::loadstore() { compiler_barrier(); } |
43 | inline void OrderAccess::storeload() { fence(); } |
44 | |
45 | inline void OrderAccess::acquire() { compiler_barrier(); } |
46 | inline void OrderAccess::release() { compiler_barrier(); } |
47 | |
48 | inline void OrderAccess::fence() { |
49 | // always use locked addl since mfence is sometimes expensive |
50 | #ifdef AMD64 |
51 | __asm__ volatile ("lock; addl $0,0(%%rsp)" : : : "cc" , "memory" ); |
52 | #else |
53 | __asm__ volatile ("lock; addl $0,0(%%esp)" : : : "cc" , "memory" ); |
54 | #endif |
55 | compiler_barrier(); |
56 | } |
57 | |
58 | inline void OrderAccess::cross_modify_fence() { |
59 | int idx = 0; |
60 | #ifdef AMD64 |
61 | __asm__ volatile ("cpuid " : "+a" (idx) : : "ebx" , "ecx" , "edx" , "memory" ); |
62 | #else |
63 | // On some x86 systems EBX is a reserved register that cannot be |
64 | // clobbered, so we must protect it around the CPUID. |
65 | __asm__ volatile ("xchg %%esi, %%ebx; cpuid; xchg %%esi, %%ebx " : "+a" (idx) : : "esi" , "ecx" , "edx" , "memory" ); |
66 | #endif |
67 | } |
68 | |
69 | template<> |
70 | struct OrderAccess::PlatformOrderedStore<1, RELEASE_X_FENCE> |
71 | { |
72 | template <typename T> |
73 | void operator()(T v, volatile T* p) const { |
74 | __asm__ volatile ( "xchgb (%2),%0" |
75 | : "=q" (v) |
76 | : "0" (v), "r" (p) |
77 | : "memory" ); |
78 | } |
79 | }; |
80 | |
81 | template<> |
82 | struct OrderAccess::PlatformOrderedStore<2, RELEASE_X_FENCE> |
83 | { |
84 | template <typename T> |
85 | void operator()(T v, volatile T* p) const { |
86 | __asm__ volatile ( "xchgw (%2),%0" |
87 | : "=r" (v) |
88 | : "0" (v), "r" (p) |
89 | : "memory" ); |
90 | } |
91 | }; |
92 | |
93 | template<> |
94 | struct OrderAccess::PlatformOrderedStore<4, RELEASE_X_FENCE> |
95 | { |
96 | template <typename T> |
97 | void operator()(T v, volatile T* p) const { |
98 | __asm__ volatile ( "xchgl (%2),%0" |
99 | : "=r" (v) |
100 | : "0" (v), "r" (p) |
101 | : "memory" ); |
102 | } |
103 | }; |
104 | |
105 | #ifdef AMD64 |
106 | template<> |
107 | struct OrderAccess::PlatformOrderedStore<8, RELEASE_X_FENCE> |
108 | { |
109 | template <typename T> |
110 | void operator()(T v, volatile T* p) const { |
111 | __asm__ volatile ( "xchgq (%2), %0" |
112 | : "=r" (v) |
113 | : "0" (v), "r" (p) |
114 | : "memory" ); |
115 | } |
116 | }; |
117 | #endif // AMD64 |
118 | |
119 | #endif // OS_CPU_LINUX_X86_ORDERACCESS_LINUX_X86_HPP |
120 | |