1 <?xml version=
"1.0" encoding=
"UTF-8" standalone=
"no"?>
2 <!DOCTYPE html PUBLIC
"-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns=
"http://www.w3.org/1999/xhtml"><head><meta http-equiv=
"Content-Type" content=
"text/html; charset=UTF-8" /><title>Chapter
30. Concurrency
</title><meta name=
"generator" content=
"DocBook XSL-NS Stylesheets V1.78.1" /><meta name=
"keywords" content=
"ISO C++, library" /><meta name=
"keywords" content=
"ISO C++, library" /><meta name=
"keywords" content=
"ISO C++, runtime, library" /><link rel=
"home" href=
"../index.html" title=
"The GNU C++ Library" /><link rel=
"up" href=
"extensions.html" title=
"Part III. Extensions" /><link rel=
"prev" href=
"ext_demangling.html" title=
"Chapter 29. Demangling" /><link rel=
"next" href=
"ext_concurrency_impl.html" title=
"Implementation" /></head><body><div class=
"navheader"><table width=
"100%" summary=
"Navigation header"><tr><th colspan=
"3" align=
"center">Chapter
30. Concurrency
</th></tr><tr><td width=
"20%" align=
"left"><a accesskey=
"p" href=
"ext_demangling.html">Prev
</a> </td><th width=
"60%" align=
"center">Part III.
5 </th><td width=
"20%" align=
"right"> <a accesskey=
"n" href=
"ext_concurrency_impl.html">Next
</a></td></tr></table><hr /></div><div class=
"chapter"><div class=
"titlepage"><div><div><h2 class=
"title"><a id=
"manual.ext.concurrency"></a>Chapter
30. Concurrency
</h2></div></div></div><div class=
"toc"><p><strong>Table of Contents
</strong></p><dl class=
"toc"><dt><span class=
"section"><a href=
"ext_concurrency.html#manual.ext.concurrency.design">Design
</a></span></dt><dd><dl><dt><span class=
"section"><a href=
"ext_concurrency.html#manual.ext.concurrency.design.threads">Interface to Locks and Mutexes
</a></span></dt><dt><span class=
"section"><a href=
"ext_concurrency.html#manual.ext.concurrency.design.atomics">Interface to Atomic Functions
</a></span></dt></dl></dd><dt><span class=
"section"><a href=
"ext_concurrency_impl.html">Implementation
</a></span></dt><dd><dl><dt><span class=
"section"><a href=
"ext_concurrency_impl.html#manual.ext.concurrency.impl.atomic_fallbacks">Using Builtin Atomic Functions
</a></span></dt><dt><span class=
"section"><a href=
"ext_concurrency_impl.html#manual.ext.concurrency.impl.thread">Thread Abstraction
</a></span></dt></dl></dd><dt><span class=
"section"><a href=
"ext_concurrency_use.html">Use
</a></span></dt></dl></div><div class=
"section"><div class=
"titlepage"><div><div><h2 class=
"title" style=
"clear: both"><a id=
"manual.ext.concurrency.design"></a>Design
</h2></div></div></div><div class=
"section"><div class=
"titlepage"><div><div><h3 class=
"title"><a id=
"manual.ext.concurrency.design.threads"></a>Interface to Locks and Mutexes
</h3></div></div></div><p>The file
<code class=
"filename"><ext/concurrence.h
></code>
6 contains all the higher-level
7 constructs for playing with threads. In contrast to the atomics layer,
8 the concurrence layer consists largely of types. All types are defined within
<code class=
"code">namespace __gnu_cxx
</code>.
10 These types can be used in a portable manner, regardless of the
11 specific environment. They are carefully designed to provide optimum
12 efficiency and speed, abstracting out underlying thread calls and
13 accesses when compiling for single-threaded situations (even on hosts
14 that support multiple threads.)
15 </p><p>The enumerated type
<code class=
"code">_Lock_policy
</code> details the set of
17 policies:
<code class=
"code">_S_single
</code>,
<code class=
"code">_S_mutex
</code>,
18 and
<code class=
"code">_S_atomic
</code>.
19 </p><div class=
"itemizedlist"><ul class=
"itemizedlist" style=
"list-style-type: disc; "><li class=
"listitem"><p><code class=
"code">_S_single
</code></p><p>Indicates single-threaded code that does not need locking.
20 </p></li><li class=
"listitem"><p><code class=
"code">_S_mutex
</code></p><p>Indicates multi-threaded code using thread-layer abstractions.
21 </p></li><li class=
"listitem"><p><code class=
"code">_S_atomic
</code></p><p>Indicates multi-threaded code using atomic operations.
22 </p></li></ul></div><p>The compile-time constant
<code class=
"code">__default_lock_policy
</code> is set
23 to one of the three values above, depending on characteristics of the
24 host environment and the current compilation flags.
25 </p><p>Two more datatypes make up the rest of the
26 interface:
<code class=
"code">__mutex
</code>, and
<code class=
"code">__scoped_lock
</code>.
27 </p><p>The scoped lock idiom is well-discussed within the C++
28 community. This version takes a
<code class=
"code">__mutex
</code> reference, and
29 locks it during construction of
<code class=
"code">__scoped_lock
</code> and
30 unlocks it during destruction. This is an efficient way of locking
31 critical sections, while retaining exception-safety.
32 These types have been superseded in the ISO C++
2011 standard by the
33 mutex and lock types defined in the header
34 <code class=
"filename"><mutex
></code>.
35 </p></div><div class=
"section"><div class=
"titlepage"><div><div><h3 class=
"title"><a id=
"manual.ext.concurrency.design.atomics"></a>Interface to Atomic Functions
</h3></div></div></div><p>
36 Two functions and one type form the base of atomic support.
37 </p><p>The type
<code class=
"code">_Atomic_word
</code> is a signed integral type
38 supporting atomic operations.
40 The two functions functions are:
41 </p><pre class=
"programlisting">
43 __exchange_and_add_dispatch(volatile _Atomic_word*, int);
46 __atomic_add_dispatch(volatile _Atomic_word*, int);
47 </pre><p>Both of these functions are declared in the header file
48 <ext/atomicity.h
>, and are in
<code class=
"code">namespace __gnu_cxx
</code>.
49 </p><div class=
"itemizedlist"><ul class=
"itemizedlist" style=
"list-style-type: disc; "><li class=
"listitem"><p>
51 __exchange_and_add_dispatch
53 </p><p>Adds the second argument's value to the first argument. Returns the old value.
54 </p></li><li class=
"listitem"><p>
58 </p><p>Adds the second argument's value to the first argument. Has no return value.
59 </p></li></ul></div><p>
60 These functions forward to one of several specialized helper
61 functions, depending on the circumstances. For instance,
64 __exchange_and_add_dispatch
67 Calls through to either of:
68 </p><div class=
"itemizedlist"><ul class=
"itemizedlist" style=
"list-style-type: disc; "><li class=
"listitem"><p><code class=
"code">__exchange_and_add
</code>
69 </p><p>Multi-thread version. Inlined if compiler-generated builtin atomics
70 can be used, otherwise resolved at link time to a non-builtin code
72 </p></li><li class=
"listitem"><p><code class=
"code">__exchange_and_add_single
</code>
73 </p><p>Single threaded version. Inlined.
</p></li></ul></div><p>However, only
<code class=
"code">__exchange_and_add_dispatch
</code>
74 and
<code class=
"code">__atomic_add_dispatch
</code> should be used. These functions
75 can be used in a portable manner, regardless of the specific
76 environment. They are carefully designed to provide optimum efficiency
77 and speed, abstracting out atomic accesses when they are not required
78 (even on hosts that support compiler intrinsics for atomic
81 In addition, there are two macros
84 _GLIBCXX_READ_MEM_BARRIER
88 _GLIBCXX_WRITE_MEM_BARRIER
91 Which expand to the appropriate write and read barrier required by the
92 host hardware and operating system.
93 </p></div></div></div><div class=
"navfooter"><hr /><table width=
"100%" summary=
"Navigation footer"><tr><td width=
"40%" align=
"left"><a accesskey=
"p" href=
"ext_demangling.html">Prev
</a> </td><td width=
"20%" align=
"center"><a accesskey=
"u" href=
"extensions.html">Up
</a></td><td width=
"40%" align=
"right"> <a accesskey=
"n" href=
"ext_concurrency_impl.html">Next
</a></td></tr><tr><td width=
"40%" align=
"left" valign=
"top">Chapter
29. Demangling
</td><td width=
"20%" align=
"center"><a accesskey=
"h" href=
"../index.html">Home
</a></td><td width=
"40%" align=
"right" valign=
"top"> Implementation
</td></tr></table></div></body></html>