1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
|
#ifndef __ATOMIC_H__
#define __ATOMIC_H__
#include <deque>
#include <Exceptions.h>
namespace Atomic {
#if (__GNUC__ >= 5) || ((__GNUC__ == 4) && ((__GNUC_MINOR__ >= 1)))
// gcc version of the atomic operations
template <class T> T Or(volatile T * ptr, T mask) { __sync_or_and_fetch(ptr, mask); }
template <class T> T And(volatile T * ptr, T mask) { __sync_and_and_fetch(ptr, mask); }
template <class T> T Xor(volatile T * ptr, T mask) { __sync_xor_and_fetch(ptr, mask); }
template <class T> T Nand(volatile T * ptr, T mask) { __sync_nand_and_fetch(ptr, mask); }
template <class T> T Increment(volatile T * ptr, T delta = 1) { __sync_add_and_fetch(ptr, delta); }
template <class T> T Decrement(volatile T * ptr, T delta = 1) { __sync_sub_and_fetch(ptr, delta); }
namespace Prefetch {
template <class T> T Or(volatile T * ptr, T mask) { __sync_fetch_and_or(ptr, mask); }
template <class T> T And(volatile T * ptr, T mask) { __sync_fetch_and_and(ptr, mask); }
template <class T> T Xor(volatile T * ptr, T mask) { __sync_fetch_and_xor(ptr, mask); }
template <class T> T Nand(volatile T * ptr, T mask) { __sync_fetch_and_nand(ptr, mask); }
template <class T> T Increment(volatile T * ptr, T delta = 1) { __sync_fetch_and_add(ptr, delta); }
template <class T> T Decrement(volatile T * ptr, T delta = 1) { __sync_fetch_and_sub(ptr, delta); }
};
template <class T> T CmpXChgVal(volatile T * ptr, const T xch, const T cmp) { return __sync_val_compare_and_swap(ptr, cmp, xch); }
template <class T> bool CmpXChgBool(volatile T * ptr, const T xch, const T cmp) { return __sync_bool_compare_and_swap(ptr, cmp, xch); }
template <class T> T Exchange32(volatile T * ptr, const T exchange) {
#if defined(i386) || defined (__x86_64)
__asm__ __volatile__("lock xchgl %0, (%1)" : "+r"(exchange) : "r"(ptr));
return exchange;
#else
T p;
do { p = *ptr; } while (!__sync_bool_compare_and_swap(ptr, p, exchange));
return p;
#endif
}
template <class T> T Exchange64(volatile T * ptr, const T exchange) {
#if defined(i386) || defined (__x86_64)
__asm__ __volatile__("lock xchgq %0, (%1)" : "+r"(exchange) : "r"(ptr));
return exchange;
#else
T p;
do { p = *ptr; } while (!__sync_bool_compare_and_swap(ptr, p, exchange));
return p;
#endif
}
#else
#ifdef _MSVC
// Visual Studio version of the atomic operations
#error MSVC not yet implemented... and probably never will -.-
#else
#error No known platform for atomic operations.
#endif
#endif
template <class T> T * ExchangePtr(T * volatile * ptr, const T * exchange) {
#if defined (__x86_64)
return Exchange64(ptr, exchange);
#else
return Exchange32(ptr, exchange);
#endif
}
// This is a multiple producers / single consumer queue system. Stolen from the Fugue VM.
template<class T>
class Queue {
private:
struct ptr_t;
struct node_t {
node_t() : next(0), val(0) { }
explicit node_t(T * val) : next(0), val(val) { }
node_t * next;
T * val;
ptr_t * ptr;
};
struct ptr_t {
ptr_t() : next(0), node(0) { }
explicit ptr_t(node_t * node) : next(0), node(node) { }
ptr_t(ptr_t * next, node_t * node) : next(next), node(node) { }
ptr_t * next;
node_t * node;
};
public:
static const int NUMSLOTS = 16384;
Queue() {
unsigned int i;
WHead = OWHead = new node_t;
RHead = ORHead = new node_t;
NodeFreeList = NULL;
for(i = 0; i < NUMSLOTS; i++) {
ptr_t * ptr = new ptr_t(NodeFreeList, &NodeBuffer[i]);
FreeListHead = NodeFreeList = ptr;
NodeBuffer[i].ptr = ptr;
}
}
~Queue() {
unsigned int i;
typename std::deque<node_t *>::iterator itr;
for (itr = PendingReads.begin(); itr != PendingReads.end(); itr++)
delete (*itr)->val;
node_t * n;
n = WHead;
while (n) {
delete n->val;
n = n->next;
}
n = RHead;
while (n) {
delete n->val;
n = n->next;
}
for (i = 0; i < NUMSLOTS; i++)
delete NodeBuffer[i].ptr;
delete ORHead;
delete OWHead;
}
void enqueue(T * val) {
node_t * node = AllocateNode(val);
while (true) {
node->next = WHead;
if (CmpXChgBool(&WHead, node, node->next))
return;
}
}
public:
T * unqueue(T ** pval = 0) {
T * r;
if (!PendingReads.empty()) {
r = PopPendingRead();
if (pval)
*pval = r;
return r;
}
if (!RHead->next)
SwapReadAndWrite();
node_t * n;
n = RHead;
while (RHead->next) {
PendingReads.push_back(n);
n = n->next;
RHead = n;
}
if (PendingReads.empty()) {
if (pval)
*pval = 0;
return 0;
}
r = this->PopPendingRead();
if (pval)
*pval = r;
return r;
}
private:
T * PopPendingRead() {
node_t * node = PendingReads.back();
T * val = node->val;
FreeNode(node);
PendingReads.pop_back();
return val;
}
void SwapReadAndWrite() {
while (true) {
node_t * rh = RHead, * ow = WHead;
if (CmpXChgBool(&WHead, rh, ow)) {
RHead = ow;
return;
}
}
}
node_t * AllocateNode(T * val) throw (GeneralException) {
ptr_t * ptr;
if (!FreeListHead->next)
throw GeneralException("No more free slots in the queue.");
ptr_t * newhead;
do {
ptr = FreeListHead;
newhead = ptr->next;
} while (!CmpXChgBool(&FreeListHead, newhead, ptr));
ptr->node->val = val;
return ptr->node;
}
void FreeNode(node_t * node) {
node->val = 0;
while (true) {
node->ptr->next = FreeListHead;
if (CmpXChgBool(&FreeListHead, node->ptr, node->ptr->next))
break;
}
}
node_t * WHead, * OWHead, * RHead, * ORHead, NodeBuffer[NUMSLOTS];
ptr_t * NodeFreeList, * FreeListHead;
// this should be stored in TLS...
typename std::deque<node_t *> PendingReads;
};
}; // namespace Atomic
#endif
|