2
0
mirror of https://github.com/boostorg/thread.git synced 2026-01-23 18:12:12 +00:00
Files
thread/doc/mutex.html
Beman Dawes b282e06a90 Initial commit
[SVN r10342]
2001-06-15 15:42:45 +00:00

211 lines
4.9 KiB
HTML

<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="keywords" content="threads, BTL, thread library, C++">
<title>Boost.Threads, mutex</title>
</head>
<body bgcolor="#ffffff" link="#0000ff" vlink="#800080">
<table border="0" cellpadding="7" cellspacing="0" width="100%">
<tr>
<td valign="top" width="300">
<h3><IMG height=86 alt="C++ Boost" src="../../../c++boost.gif" width=277></h3>
</td>
<td valign="top">
<h1 align="center">Boost.Threads</h1>
<h2 align="center">mutex / try_mutex / timed_mutex</h2>
</td>
</tr>
</table>
<hr>
<p>The <tt>mutex</tt>, <tt>try_mutex</tt> and <tt>timed_mutex</tt> classes define full featured
<A href="mutex_concept.html">mutex models</a>. These type should be used to synchronize access to
shared resources when recursive locking mechanics need not be employed. Each type adds another
<A href="lock_concept.html">lock concept</a>, including the <A href="basic_lock.html">basic_lock</a>,
<A href="basic_trylock.html">basic_trylock</a> and <A href="basic_timedlock.html">basic_timedlock</a>
types. For the best possible performance you should use the mutex type that supports only the lock
types that you need.</p>
<p>The <tt>mutex</tt>, <tt>try_mutex</tt> and <tt>timed_mutex</tt> use an <tt>Unspecified</tt>
<A href="mutex_concept.html#LockingStrategies">locking strategy</a>, so attempts to recursively lock
them or attempts to unlock them by threads that don't own a lock on them result in <b>undefined behavior</b>.
This strategy allows implementations to be as efficient as possible on any given platform. It is, however,
recommended that implementations include some debugging support to detect misuse when <tt>NDEBUG</tt> is
not defined.</p>
<p>Like all the <b>Boost.Threads</b> <A href="mutex_concept.html">mutex models</a>, the <tt>mutex</tt>,
<tt>try_mutex</tt> and <tt>timed_mutex</tt> leave the
<A href="mutex_concept.html#SchedulingPolicies">scheduling policy</a> as <tt>Unspecified</tt>.
Programmers should assume that threads waiting for a lock on these <tt>mutex types</tt> shall acquire
the lock in a random order, though the specific behavior for a given platform may be different.</p>
<h2>Header</h2>
<pre>
#include <a href="../../../boost/thread/mutex.hpp">&lt;boost/thread/mutex.hpp&gt;</a>
</pre>
<h2>Public Interface</h2>
<hr width="50%" align="left">
<pre>
class mutex : private boost::noncopyable
{
public:
typedef boost::basic_lock&lt;mutex&gt; lock;
mutex();
~mutex();
};
</pre>
<hr width="50%" align="left">
<h3>Constructor</h3>
<pre>
mutex();
</pre>
<p>Constructs a <tt>mutex</tt>.</p>
<h3>Destructor</h3>
<pre>
~mutex();
</pre>
<p>Destructs a <tt>mutex</tt>.</p>
<hr width="50%" align="left">
<pre>
class try_mutex : private boost::noncopyable
{
public:
typedef boost::basic_lock&lt;try_mutex&gt; lock;
typedef boost::basic_trylock&lt;try_mutex&gt; trylock;
try_mutex();
~try_mutex();
};
</pre>
<hr width="50%" align="left">
<h3>Constructor</h3>
<pre>
try_mutex();
</pre>
<p>Constructs a <tt>try_mutex</tt>.</p>
<h3>Destructor</h3>
<pre>
~try_mutex();
</pre>
<p>Destructs a <tt>try_mutex</tt>.</p>
<hr width="50%" align="left">
<pre>
class timed_mutex : private boost::noncopyable
{
public:
typedef boost::basic_lock&lt;timed_mutex&gt; lock;
typedef boost::basic_trylock&lt;timed_mutex&gt; trylock;
typedef boost::basic_timedlock&lt;timed_mutex&gt; timedlock;
timed_mutex();
~timed_mutex();
};
</pre>
<hr width="50%" align="left">
<h3>Constructor</h3>
<pre>
timed_mutex();
</pre>
<p>Constructs a <tt>timed_mutex</tt>.</p>
<h3>Destructor</h3>
<pre>
~timed_mutex();
</pre>
<p>Destructs a <tt>mutex</tt>.</p>
<h2>Example Usage</h2>
<pre>
#include <a href="../../../boost/thread/mutex.hpp">&lt;boost/thread/mutex.hpp&gt;</a>
#include <a href="../../../boost/thread/thread.hpp">&lt;boost/thread/thread.hpp&gt;</a>
#include &lt;iostream&gt;
boost::mutex io_mutex; // The iostreams are not gauranteed to be thread safe!
class counter
{
private:
boost::mutex mutex;
int count;
public:
counter() : count(0) { }
int increment() {
boost::mutex::lock lock(mutex);
return ++count;
}
};
counter c;
void change_count(void*)
{
int i = c.increment();
boost::mutex::lock lock(io_mutex);
std::cout &lt;&lt; &quot;count == &quot; &lt;&lt; i &lt;&lt; std::endl;
}
int main(int, char*[])
{
const int num_threads = 4;
for (int i=0; i &lt; num_threads; ++i)
boost::thread::create(&amp;change_count, 0);
boost::thread::join_all();
return 0;
}
</pre>
<p>The output is:</p>
<pre>
count == 1
count == 2
count == 3
count == 4
</pre>
<hr>
<p><i>Copyright <A href="mailto:williamkempf@hotmail.com">William E. Kempf</a>
2001 all rights reserved.</i></p>
</body>
</html>