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
|
/*
* Copyright (c) 2002 Frodo
* Portions Copyright (c) by the authors of ffmpeg and xvid
* Copyright (C) 2002-2013 Team XBMC
* http://xbmc.org
*
* This Program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2, or (at your option)
* any later version.
*
* This Program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with XBMC; see the file COPYING. If not, see
* <http://www.gnu.org/licenses/>.
*
*/
#include <stdarg.h>
#include <limits>
#include "Event.h"
void CEvent::addGroup(XbmcThreads::CEventGroup* group)
{
CSingleLock lock(groupListMutex);
if (groups == NULL)
groups = new std::vector<XbmcThreads::CEventGroup*>();
groups->push_back(group);
}
void CEvent::removeGroup(XbmcThreads::CEventGroup* group)
{
CSingleLock lock(groupListMutex);
if (groups)
{
for (std::vector<XbmcThreads::CEventGroup*>::iterator iter = groups->begin(); iter != groups->end(); ++iter)
{
if ((*iter) == group)
{
groups->erase(iter);
break;
}
}
if (groups->size() <= 0)
{
delete groups;
groups = NULL;
}
}
}
// locking is ALWAYS done in this order:
// CEvent::groupListMutex -> CEventGroup::mutex -> CEvent::mutex
void CEvent::Set()
{
// Originally I had this without locking. Thanks to FernetMenta who
// pointed out that this creates a race condition between setting
// checking the signal and calling wait() on the Wait call in the
// CEvent class. This now perfectly matches the boost example here:
// http://www.boost.org/doc/libs/1_41_0/doc/html/thread/synchronization.html#thread.synchronization.condvar_ref
{
CSingleLock slock(mutex);
signaled = true;
}
condVar.notifyAll();
CSingleLock l(groupListMutex);
if (groups)
{
for (std::vector<XbmcThreads::CEventGroup*>::iterator iter = groups->begin();
iter != groups->end(); ++iter)
(*iter)->Set(this);
}
}
namespace XbmcThreads
{
/**
* This will block until any one of the CEvents in the group are
* signaled at which point a pointer to that CEvents will be
* returned.
*/
CEvent* CEventGroup::wait()
{
return wait(std::numeric_limits<unsigned int>::max());
}
/**
* This will block until any one of the CEvents in the group are
* signaled or the timeout is reachec. If an event is signaled then
* it will return a pointer to that CEvent, otherwise it will return
* NULL.
*/
// locking is ALWAYS done in this order:
// CEvent::groupListMutex -> CEventGroup::mutex -> CEvent::mutex
//
// Notice that this method doesn't grab the CEvent::groupListMutex at all. This
// is fine. It just grabs the CEventGroup::mutex and THEN the individual
// CEvent::mutex's
CEvent* CEventGroup::wait(unsigned int milliseconds)
{
CSingleLock lock(mutex); // grab CEventGroup::mutex
numWaits++;
// ==================================================
// This block checks to see if any child events are
// signaled and sets 'signaled' to the first one it
// finds.
// ==================================================
signaled = NULL;
for (std::vector<CEvent*>::iterator iter = events.begin();
signaled == NULL && iter != events.end(); ++iter)
{
CEvent* cur = *iter;
if (cur->signaled)
signaled = cur;
}
// ==================================================
if(!signaled)
{
// both of these release the CEventGroup::mutex
if (milliseconds == std::numeric_limits<unsigned int>::max())
condVar.wait(mutex);
else
condVar.wait(mutex,milliseconds);
} // at this point the CEventGroup::mutex is reacquired
numWaits--;
// signaled should have been set by a call to CEventGroup::Set
CEvent* ret = signaled;
if (numWaits == 0)
{
if (signaled)
// This acquires and releases the CEvent::mutex. This is fine since the
// CEventGroup::mutex is already being held
signaled->WaitMSec(0); // reset the event if needed
signaled = NULL; // clear the signaled if all the waiters are gone
}
return ret;
}
CEventGroup::CEventGroup(int num, CEvent* v1, ...) : signaled(NULL), condVar(actualCv,signaled), numWaits(0)
{
va_list ap;
va_start(ap, v1);
if (v1)
events.push_back(v1);
num--; // account for v1
for (; num > 0; num--)
{
CEvent* const cur = va_arg(ap, CEvent*);
if (cur)
events.push_back(cur);
}
va_end(ap);
// we preping for a wait, so we need to set the group value on
// all of the CEvents.
for (std::vector<CEvent*>::iterator iter = events.begin();
iter != events.end(); ++iter)
(*iter)->addGroup(this);
}
CEventGroup::CEventGroup(CEvent* v1, ...) : signaled(NULL), condVar(actualCv,signaled), numWaits(0)
{
va_list ap;
va_start(ap, v1);
if (v1)
events.push_back(v1);
bool done = false;
while(!done)
{
CEvent* cur = va_arg(ap,CEvent*);
if (cur)
events.push_back(cur);
else
done = true;
}
va_end(ap);
// we preping for a wait, so we need to set the group value on
// all of the CEvents.
for (std::vector<CEvent*>::iterator iter = events.begin();
iter != events.end(); ++iter)
(*iter)->addGroup(this);
}
CEventGroup::~CEventGroup()
{
for (std::vector<CEvent*>::iterator iter = events.begin();
iter != events.end(); ++iter)
(*iter)->removeGroup(this);
}
}
|
