2
0
mirror of https://github.com/boostorg/thread.git synced 2026-01-23 18:12:12 +00:00
Files
thread/doc/condition.html
2001-07-19 20:42:31 +00:00

321 lines
9.2 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++">
<link rel="stylesheet" type="text/css" href="styles.css">
<title>Boost.Threads, condition</title>
</head>
<body bgcolor="#FFFFFF" link="#0000FF" vlink="#800080" text="#000000">
<table border="0" cellpadding="7" cellspacing="0" width="100%">
<tr>
<td valign="top" width="300">
<h3><img src="../../../c++boost.gif" alt="C++ Boost" width="277" height="86"></h3>
</td>
<td valign="top">
<h1 align="center">Boost.Threads</h1>
<h2 align="center">condition</h2>
</td>
</tr>
</table>
<hr>
<p><a href="#Introduction">Introduction</a><br>
<a href="#Header">Header</a><br>
<a href="#Synopsis">Synopsis</a><br>
<a href="#Members">Members</a><br>
<a href="#Example">Example</a></p>
<h2><a name="Introduction">Introduction</a></h2>
<p>A <code>condition</code> is a synchronization primitive used to cause a thread to
wait until a particular condition (or time) is met. A <code>condition</code> always works in
conjunction with a <a href="mutex_concept.html">mutex model</a>. The
<a href="mutex_concept.html">mutex model</a> must be locked prior to waiting on the
<code>condition</code>, which is insured by passing a <a href="lock_concept.html">lock model</a>
to the <code>condition</code> class <code>wait</code> functions. While the thread is waiting on
the <code>condition</code> the <a href="mutex_concept.html">mutex model</a> is unlocked. When the
thread returns from a call to one of the <code>condition</code> class <code>wait</code> functions,
the <a href="mutex_concept.html">mutex model</a> is again locked.</p>
<p>The <code>condition</code> type is often used to implement the
<i>Monitor Object</i> pattern. See <a href="bibliography.html#Schmidt-00">[Schmidt
00]</a> and <a href="bibliography.html#Hoare-74">[Hoare 74</a>]. Monitors
are one of the most important patterns for creating reliable multithreaded
programs.</p>
<p>See <a href="definitions.html"> Formal Definitions</a> for definitions of thread
states <a href="definitions.html#state">blocked</a> and
<a href="definitions.html#state">ready</a>. Note that &quot;waiting&quot; is a synonym
for blocked.</p>
<h2><a name="Header">Header</a></h2>
<pre>
#include <a href="../../../boost/thread/condition.hpp">&lt;boost/thread/condition.hpp&gt;</a>
</pre>
<h2><a name="Synopsis">Synopsis</a></h2>
<pre>
namespace boost {
class condition : private <a href="../../utility/utility.htm#Class noncopyable">boost::noncopyable</a>
{
public:
condition();
~condition();
void notify_one();
void notify_all();
template &lt;typename L&gt;
void wait(L&amp; lock);
template &lt;typename L, typename Pr&gt;
void wait(L&amp; lock, Pr pred);
template &lt;typename L&gt;
bool timed_wait(L&amp; lock, const xtime&amp; xt);
template &lt;typename L, typename Pr&gt;
bool timed_wait(L&amp; lock, const xtime&amp; xt, Pr pred);
};
} // namespace boost
</pre>
<h2><a name="Members">Members</a></h2>
<hr>
<h3>Constructor</h3>
<pre>
condition();
</pre>
<p><b>Effects:</b> Constructs a <code>condition</code>.</p>
<hr>
<h3>Destructor</h3>
<pre>
~condition();
</pre>
<p><b>Effects:</b> Destroys <code>*this</code>.</p>
<hr>
<h3>notify_one</h3>
<pre>
void notify_one();
</pre>
<p><b>Effects:</b> If there is a thread waiting on <code>*this</code>, change
that thread's state to ready. Otherwise there is no effect.</p>
<p><b>Notes:</b> If more that one thread is waiting on the condition, it is
unspecified which is made ready.</p>
<hr>
<h3>notify_all</h3>
<pre>
void notify_all();
</pre>
<p><b>Effects:</b> Change the state of all threads waiting on <code>*this</code>
to ready. If there are no waiting threads, <code>notify_all()</code> has no effect.</p>
<hr>
<h3>wait</h3>
<pre>
template &lt;typename L&gt;
void wait(L&amp; lock);
</pre>
<p><b>Requires:</b> <code>L</code> meet the requirements of the <a href="lock_concept.html">Lock
concept</a>.</p>
<p><b>Effects:</b> Releases the lock on the <a href="mutex_concept.html">mutex model</a>
associated with <code>lock</code>, blocks the current thread of execution until readied by a
call to <code>*this-&gt;notify_one()</code> or <code>*this-&gt;notify_all()</code>, and then
reacquires the lock. All effects occur in an atomic fashion.</p>
<p><b>Throws:</b> <code><a href="lock_error.html">lock_error</a></code> if <code>!lock</code></p>
<p><b>Danger:</b> This version should always be used within a loop checking that the state
logically associated with the <code>condition</code> has become true. Without the loop, race
conditions can ensue due to possible "spurious wake ups". The second version encapsulates this
loop idiom internally and is generally the preferred method.</p>
<pre>
template &lt;typename L, typename Pr&gt;
void wait(L&amp; lock, Pr pred);
</pre>
<p><b>Requires:</b> <code>L</code> meet the requirements of the <a href="lock_concept.html">Lock
concept</a>. Return from <code>pred()</code> convertible to bool.</p>
<p><b>Effects:</b> As if:</p>
<code>
while !(pred()) wait(lock)
</code>
<p><b>Throws:</b> <code><a href="lock_error.html">lock_error</a></code> if <code>!lock</code></p>
<hr>
<h3>timed_wait</h3>
<pre>
template &lt;typename L&gt;
bool timed_wait(L&amp; lock, const xtime&amp; xt);
</pre>
<p><b>Requires:</b> <code>L</code> meet the requirements of the <a href="lock_concept.html">Lock
concept</a>.</p>
<p><b>Effects:</b> Releases the lock on the <a href="mutex_concept.html">mutex model</a> associated
with the <code>lock</code>, blocks the current thread of execution until readied by a call to
<code>*this-&gt;notify_one()</code> or <code>*this-&gt;notify_all()</code>, or until <code>xt</code>,
and then reacquires the lock. All effects occur in an atomic fashion.</p>
<p><b>Throws:</b> <code><a href="lock_error.html">lock_error</a></code> if <code>!lock</code></p>
<p><b>Danger:</b> This version should always be used within a loop checking that the state
logically associated with the <code>condition</code> has become true. Without the loop, race
conditions can ensue due to "spurious wake ups". The second version encapsulates this loop idiom
internally and is generally the preferred method.</p>
<p><b>Returns:</b> <code>false</code> if <code>xt</code> expires, otherwise <code>true</code>.</p>
<pre>
template &lt;typename L, typename Pr&gt;
bool timed_wait(L&amp; lock, const xtime&amp; xt, Pr pred);
</pre>
<p><b>Requires: </b>L meet the requirements of the <a href="lock_concept.html">Lock
concept</a>. Return from <code>pred()</code> convertible to bool.</p>
<p><b>Effects: </b>As if:</p>
<code>
while (!pred())
{
if (!timed_wait(lock, xt))
return false;
}
</code>
<p><b>Throws:</b> <code><a href="lock_error.html">lock_error</a></code> if <code>!lock</code></p>
<p><b>Returns:</b> <code>false</code> if <code>xt</code> expires, otherwise <code>true</code>.</p>
<hr>
<h2><a name="Example">Example Usage</a></h2>
<pre>
#include &lt;iostream&gt;
#include &lt;vector&gt;
#include <a href="../../../boost/utility.hpp">&lt;boost/utility.hpp&gt;</a>
#include <a href="../../../boost/thread/condition.hpp">&lt;boost/thread/condition.hpp&gt;</a>
#include <a href="../../../boost/thread/thread.hpp">&lt;boost/thread/thread.hpp&gt;</a>
class bounded_buffer : private boost::noncopyable
{
private:
int begin, end, buffered;
std::vector&lt;int&gt; circular_buf;
boost::condition buffer_not_full, buffer_not_empty;
boost::mutex monitor;
typedef boost::mutex::lock lock;
public:
buffer(int n) : begin(0), end(0), buffered(0), circular_buf(n) { }
void send (int m) {
lock lk(monitor);
while (buffered == circular_buf.size())
buffer_not_full.wait(lk);
circular_buf[end] = m;
end = (end+1) % circular_buf.size();
++buffered;
buffer_not_empty.notify_one();
}
int receive() {
lock lk(monitor);
while (buffered == 0 &amp;&amp; !finished)
buffer_not_empty.wait(lk);
int i = circular_buf[begin];
begin = (begin+1) % circular_buf.size();
--buffered;
buffer_not_full.notify_one();
return i;
}
};
bounded_buffer buf(2);
void sender(void*) {
int n = 0;
while (n &lt; 100) {
buf.send(n);
std::cout &lt;&lt; &quot;sent: &quot; &lt;&lt; n &lt;&lt; std::endl;
++n;
}
buf.send(-1);
}
void receiver(void*) {
int n;
do {
n = buf.receive();
std::cout &lt;&lt; &quot;received: &quot; &lt;&lt; n &lt;&lt; std::endl;
} while (n != -1); // -1 indicates end of buffer
}
int main(int, char*[])
{
boost::thread::create(&amp;sender, 0);
boost::thread::create(&amp;receiver, 0);
boost::thread::join_all();
return 0;
}
</pre>
<p>Typical output (dependent on scheduling policies) is:</p>
<pre>
sent: 0
sent: 1
received: 0
received: 1
sent: 2
sent: 3
received: 2
received: 3
sent: 4
received: 4
</pre>
<hr>
<p>Revised <!--webbot bot="Timestamp" S-Type="EDITED" S-Format="%d %B, %Y" startspan -->19 July, 2001<!--webbot bot="Timestamp" endspan i-checksum="21079" -->
</p>
<p><i>Copyright <a href="mailto:williamkempf@hotmail.com">William E. Kempf</a>
2001 all rights reserved.</i></p>
</body>
</html>