0
|
1 /* Copyright (C) 2005, 2008, 2009 Free Software Foundation, Inc.
|
|
2 Contributed by Richard Henderson <rth@redhat.com>.
|
|
3
|
|
4 This file is part of the GNU OpenMP Library (libgomp).
|
|
5
|
|
6 Libgomp is free software; you can redistribute it and/or modify it
|
|
7 under the terms of the GNU General Public License as published by
|
|
8 the Free Software Foundation; either version 3, or (at your option)
|
|
9 any later version.
|
|
10
|
|
11 Libgomp is distributed in the hope that it will be useful, but WITHOUT ANY
|
|
12 WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
|
|
13 FOR A PARTICULAR PURPOSE. See the GNU General Public License for
|
|
14 more details.
|
|
15
|
|
16 Under Section 7 of GPL version 3, you are granted additional
|
|
17 permissions described in the GCC Runtime Library Exception, version
|
|
18 3.1, as published by the Free Software Foundation.
|
|
19
|
|
20 You should have received a copy of the GNU General Public License and
|
|
21 a copy of the GCC Runtime Library Exception along with this program;
|
|
22 see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
|
|
23 <http://www.gnu.org/licenses/>. */
|
|
24
|
|
25 /* This is a Linux specific implementation of a barrier synchronization
|
|
26 mechanism for libgomp. This type is private to the library. This
|
|
27 implementation uses atomic instructions and the futex syscall. */
|
|
28
|
|
29 #ifndef GOMP_BARRIER_H
|
|
30 #define GOMP_BARRIER_H 1
|
|
31
|
|
32 #include "mutex.h"
|
|
33
|
|
34 typedef struct
|
|
35 {
|
|
36 /* Make sure total/generation is in a mostly read cacheline, while
|
|
37 awaited in a separate cacheline. */
|
|
38 unsigned total __attribute__((aligned (64)));
|
|
39 unsigned generation;
|
|
40 unsigned awaited __attribute__((aligned (64)));
|
|
41 } gomp_barrier_t;
|
|
42 typedef unsigned int gomp_barrier_state_t;
|
|
43
|
|
44 static inline void gomp_barrier_init (gomp_barrier_t *bar, unsigned count)
|
|
45 {
|
|
46 bar->total = count;
|
|
47 bar->awaited = count;
|
|
48 bar->generation = 0;
|
|
49 }
|
|
50
|
|
51 static inline void gomp_barrier_reinit (gomp_barrier_t *bar, unsigned count)
|
|
52 {
|
|
53 __sync_fetch_and_add (&bar->awaited, count - bar->total);
|
|
54 bar->total = count;
|
|
55 }
|
|
56
|
|
57 static inline void gomp_barrier_destroy (gomp_barrier_t *bar)
|
|
58 {
|
|
59 }
|
|
60
|
|
61 extern void gomp_barrier_wait (gomp_barrier_t *);
|
|
62 extern void gomp_barrier_wait_last (gomp_barrier_t *);
|
|
63 extern void gomp_barrier_wait_end (gomp_barrier_t *, gomp_barrier_state_t);
|
|
64 extern void gomp_team_barrier_wait (gomp_barrier_t *);
|
|
65 extern void gomp_team_barrier_wait_end (gomp_barrier_t *,
|
|
66 gomp_barrier_state_t);
|
|
67 extern void gomp_team_barrier_wake (gomp_barrier_t *, int);
|
|
68
|
|
69 static inline gomp_barrier_state_t
|
|
70 gomp_barrier_wait_start (gomp_barrier_t *bar)
|
|
71 {
|
|
72 unsigned int ret = bar->generation & ~3;
|
|
73 /* Do we need any barrier here or is __sync_add_and_fetch acting
|
|
74 as the needed LoadLoad barrier already? */
|
|
75 ret += __sync_add_and_fetch (&bar->awaited, -1) == 0;
|
|
76 return ret;
|
|
77 }
|
|
78
|
|
79 static inline bool
|
|
80 gomp_barrier_last_thread (gomp_barrier_state_t state)
|
|
81 {
|
|
82 return state & 1;
|
|
83 }
|
|
84
|
|
85 /* All the inlines below must be called with team->task_lock
|
|
86 held. */
|
|
87
|
|
88 static inline void
|
|
89 gomp_team_barrier_set_task_pending (gomp_barrier_t *bar)
|
|
90 {
|
|
91 bar->generation |= 1;
|
|
92 }
|
|
93
|
|
94 static inline void
|
|
95 gomp_team_barrier_clear_task_pending (gomp_barrier_t *bar)
|
|
96 {
|
|
97 bar->generation &= ~1;
|
|
98 }
|
|
99
|
|
100 static inline void
|
|
101 gomp_team_barrier_set_waiting_for_tasks (gomp_barrier_t *bar)
|
|
102 {
|
|
103 bar->generation |= 2;
|
|
104 }
|
|
105
|
|
106 static inline bool
|
|
107 gomp_team_barrier_waiting_for_tasks (gomp_barrier_t *bar)
|
|
108 {
|
|
109 return (bar->generation & 2) != 0;
|
|
110 }
|
|
111
|
|
112 static inline void
|
|
113 gomp_team_barrier_done (gomp_barrier_t *bar, gomp_barrier_state_t state)
|
|
114 {
|
|
115 bar->generation = (state & ~3) + 4;
|
|
116 }
|
|
117
|
|
118 #endif /* GOMP_BARRIER_H */
|