blob: e9275a2fa40e7abaae2d7a0a8cf64bc932c12e51 [file] [log] [blame]
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=US-ASCII">
<title>Customizing Boost.Interprocess</title>
<link rel="stylesheet" href="../../../doc/src/boostbook.css" type="text/css">
<meta name="generator" content="DocBook XSL Stylesheets V1.78.1">
<link rel="home" href="../index.html" title="The Boost C++ Libraries BoostBook Documentation Subset">
<link rel="up" href="../interprocess.html" title="Chapter&#160;14.&#160;Boost.Interprocess">
<link rel="prev" href="architecture.html" title="Architecture and internals">
<link rel="next" href="acknowledgements_notes.html" title="Acknowledgements, notes and links">
</head>
<body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF">
<table cellpadding="2" width="100%"><tr>
<td valign="top"><img alt="Boost C++ Libraries" width="277" height="86" src="../../../boost.png"></td>
<td align="center"><a href="../../../index.html">Home</a></td>
<td align="center"><a href="../../../libs/libraries.htm">Libraries</a></td>
<td align="center"><a href="http://www.boost.org/users/people.html">People</a></td>
<td align="center"><a href="http://www.boost.org/users/faq.html">FAQ</a></td>
<td align="center"><a href="../../../more/index.htm">More</a></td>
</tr></table>
<hr>
<div class="spirit-nav">
<a accesskey="p" href="architecture.html"><img src="../../../doc/src/images/prev.png" alt="Prev"></a><a accesskey="u" href="../interprocess.html"><img src="../../../doc/src/images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../../../doc/src/images/home.png" alt="Home"></a><a accesskey="n" href="acknowledgements_notes.html"><img src="../../../doc/src/images/next.png" alt="Next"></a>
</div>
<div class="section">
<div class="titlepage"><div><div><h2 class="title" style="clear: both">
<a name="interprocess.customizing_interprocess"></a><a class="link" href="customizing_interprocess.html" title="Customizing Boost.Interprocess">Customizing Boost.Interprocess</a>
</h2></div></div></div>
<div class="toc"><dl class="toc">
<dt><span class="section"><a href="customizing_interprocess.html#interprocess.customizing_interprocess.custom_interprocess_alloc">Writing
a new shared memory allocation algorithm</a></span></dt>
<dt><span class="section"><a href="customizing_interprocess.html#interprocess.customizing_interprocess.custom_allocators">Building
custom STL compatible allocators for Boost.Interprocess</a></span></dt>
<dt><span class="section"><a href="customizing_interprocess.html#interprocess.customizing_interprocess.custom_indexes">Building
custom indexes</a></span></dt>
</dl></div>
<div class="section">
<div class="titlepage"><div><div><h3 class="title">
<a name="interprocess.customizing_interprocess.custom_interprocess_alloc"></a><a class="link" href="customizing_interprocess.html#interprocess.customizing_interprocess.custom_interprocess_alloc" title="Writing a new shared memory allocation algorithm">Writing
a new shared memory allocation algorithm</a>
</h3></div></div></div>
<p>
If the default algorithm does not satisfy user requirements, it's easy to
provide different algorithms like bitmapping or more advanced segregated
lists to meet requirements. The class implementing the algorithm must be
compatible with shared memory, so it shouldn't have any virtual function
or virtual inheritance or any indirect base class with virtual function or
inheritance.
</p>
<p>
This is the interface to be implemented:
</p>
<pre class="programlisting"><span class="keyword">class</span> <span class="identifier">my_algorithm</span>
<span class="special">{</span>
<span class="keyword">public</span><span class="special">:</span>
<span class="comment">//!The mutex type to be used by the rest of Interprocess framework</span>
<span class="keyword">typedef</span> <span class="identifier">implementation_defined</span> <span class="identifier">mutex_family</span><span class="special">;</span>
<span class="comment">//!The pointer type to be used by the rest of Interprocess framework</span>
<span class="keyword">typedef</span> <span class="identifier">implementation_defined</span> <span class="identifier">void_pointer</span><span class="special">;</span>
<span class="comment">//!Constructor. "size" is the total size of the managed memory segment,</span>
<span class="comment">//!"extra_hdr_bytes" indicates the extra bytes after the sizeof(my_algorithm)</span>
<span class="comment">//!that the allocator should not use at all.</span>
<span class="identifier">my_algorithm</span> <span class="special">(</span><span class="identifier">std</span><span class="special">::</span><span class="identifier">size_t</span> <span class="identifier">size</span><span class="special">,</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">size_t</span> <span class="identifier">extra_hdr_bytes</span><span class="special">);</span>
<span class="comment">//!Obtains the minimum size needed by the algorithm</span>
<span class="keyword">static</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">size_t</span> <span class="identifier">get_min_size</span> <span class="special">(</span><span class="identifier">std</span><span class="special">::</span><span class="identifier">size_t</span> <span class="identifier">extra_hdr_bytes</span><span class="special">);</span>
<span class="comment">//!Allocates bytes, returns 0 if there is not more memory</span>
<span class="keyword">void</span><span class="special">*</span> <span class="identifier">allocate</span> <span class="special">(</span><span class="identifier">std</span><span class="special">::</span><span class="identifier">size_t</span> <span class="identifier">nbytes</span><span class="special">);</span>
<span class="comment">//!Deallocates previously allocated bytes</span>
<span class="keyword">void</span> <span class="identifier">deallocate</span> <span class="special">(</span><span class="keyword">void</span> <span class="special">*</span><span class="identifier">adr</span><span class="special">);</span>
<span class="comment">//!Returns the size of the memory segment</span>
<span class="identifier">std</span><span class="special">::</span><span class="identifier">size_t</span> <span class="identifier">get_size</span><span class="special">()</span> <span class="keyword">const</span><span class="special">;</span>
<span class="comment">//!Increases managed memory in extra_size bytes more</span>
<span class="keyword">void</span> <span class="identifier">grow</span><span class="special">(</span><span class="identifier">std</span><span class="special">::</span><span class="identifier">size_t</span> <span class="identifier">extra_size</span><span class="special">);</span>
<span class="comment">/*...*/</span>
<span class="special">};</span>
</pre>
<p>
Let's see the public typedefs to define:
</p>
<pre class="programlisting"><span class="keyword">typedef</span> <span class="comment">/* . . . */</span> <span class="identifier">void_pointer</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="comment">/* . . . */</span> <span class="identifier">mutex_family</span><span class="special">;</span>
</pre>
<p>
The <code class="computeroutput"><span class="identifier">void_pointer</span></code> typedef
specifies the pointer type to be used in the <span class="bold"><strong>Boost.Interprocess</strong></span>
framework that uses the algorithm. For example, if we define
</p>
<pre class="programlisting"><span class="keyword">typedef</span> <span class="keyword">void</span> <span class="special">*</span> <span class="identifier">void_pointer</span><span class="special">;</span>
</pre>
<p>
all <span class="bold"><strong>Boost.Interprocess</strong></span> framework using this
algorithm will use raw pointers as members. But if we define:
</p>
<pre class="programlisting"><span class="keyword">typedef</span> <span class="identifier">offset_ptr</span><span class="special">&lt;</span><span class="keyword">void</span><span class="special">&gt;</span> <span class="identifier">void_pointer</span><span class="special">;</span>
</pre>
<p>
then all <span class="bold"><strong>Boost.Interprocess</strong></span> framework will
use relative pointers.
</p>
<p>
The <code class="computeroutput"><span class="identifier">mutex_family</span></code> is a structure
containing typedefs for different interprocess_mutex types to be used in
the <span class="bold"><strong>Boost.Interprocess</strong></span> framework. For example
the defined
</p>
<pre class="programlisting"><span class="keyword">struct</span> <span class="identifier">mutex_family</span>
<span class="special">{</span>
<span class="keyword">typedef</span> <span class="identifier">boost</span><span class="special">::</span><span class="identifier">interprocess</span><span class="special">::</span><span class="identifier">interprocess_mutex</span> <span class="identifier">mutex_type</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="identifier">boost</span><span class="special">::</span><span class="identifier">interprocess</span><span class="special">::</span><span class="identifier">interprocess_recursive_mutex</span> <span class="identifier">recursive_mutex_type</span><span class="special">;</span>
<span class="special">};</span>
</pre>
<p>
defines all interprocess_mutex types using boost::interprocess interprocess_mutex
types. The user can specify the desired mutex family.
</p>
<pre class="programlisting"><span class="keyword">typedef</span> <span class="identifier">mutex_family</span> <span class="identifier">mutex_family</span><span class="special">;</span>
</pre>
<p>
The new algorithm (let's call it <span class="bold"><strong>my_algorithm</strong></span>)
must implement all the functions that boost::interprocess::rbtree_best_fit
class offers:
</p>
<div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; ">
<li class="listitem">
<span class="bold"><strong>my_algorithm</strong></span>'s constructor must take
2 arguments:
<div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: circle; ">
<li class="listitem">
<span class="bold"><strong>size</strong></span> indicates the total size
of the managed memory segment, and <span class="bold"><strong>my_algorithm</strong></span>
object will be always constructed a at offset 0 of the memory segment.
</li>
<li class="listitem">
The <span class="bold"><strong>extra_hdr_bytes</strong></span> parameter
indicates the number of bytes after the offset <code class="computeroutput"><span class="keyword">sizeof</span><span class="special">(</span><span class="identifier">my_algorithm</span><span class="special">)</span></code> that <span class="bold"><strong>my_algorithm</strong></span>
can't use at all. This extra bytes will be used to store additional
data that should not be overwritten. So, <span class="bold"><strong>my_algorithm</strong></span>
will be placed at address XXX of the memory segment, and will manage
the <span class="bold"><strong>[XXX + sizeof(my_algorithm) + extra_hdr_bytes,
XXX + size)</strong></span> range of the segment.
</li>
</ul></div>
</li>
<li class="listitem">
The <span class="bold"><strong>get_min_size()</strong></span> function should return
the minimum space the algorithm needs to be valid with the passed <span class="bold"><strong>extra_hdr_bytes</strong></span> parameter. This function will
be used to check if the memory segment is big enough to place the algorithm
there.
</li>
<li class="listitem">
The <span class="bold"><strong>allocate()</strong></span> function must return
0 if there is no more available memory. The memory returned by <span class="bold"><strong>my_algorithm</strong></span> must be aligned to the most restrictive
memory alignment of the system. This function should be executed with
the synchronization capabilities offered by <code class="computeroutput"><span class="keyword">typename</span>
<span class="identifier">mutex_family</span><span class="special">::</span><span class="identifier">mutex_type</span></code> interprocess_mutex. That
means, that if we define <code class="computeroutput"><span class="keyword">typedef</span>
<span class="identifier">mutex_family</span> <span class="identifier">mutex_family</span><span class="special">;</span></code> then this function should offer the
same synchronization as if it was surrounded by an interprocess_mutex
lock/unlock. Normally, this is implemented using a member of type <code class="computeroutput"><span class="identifier">mutex_family</span><span class="special">::</span><span class="identifier">mutex_type</span></code>, but it could be done using
atomic instructions or lock free algorithms.
</li>
<li class="listitem">
The <span class="bold"><strong>deallocate()</strong></span> function must make
the returned buffer available for new allocations. This function should
offer the same synchronization as <code class="computeroutput"><span class="identifier">allocate</span><span class="special">()</span></code>.
</li>
<li class="listitem">
The <span class="bold"><strong>size()</strong></span> function will return the
passed <span class="bold"><strong>size</strong></span> parameter in the constructor.
So, <span class="bold"><strong>my_algorithm</strong></span> should store the size
internally.
</li>
<li class="listitem">
The <span class="bold"><strong>grow()</strong></span> function will expand the
managed memory by <span class="bold"><strong>my_algorithm</strong></span> in <span class="bold"><strong>extra_size</strong></span> bytes. So <span class="bold"><strong>size()</strong></span>
function should return the updated size, and the new managed memory range
will be (if the address where the algorithm is constructed is XXX):
<span class="bold"><strong>[XXX + sizeof(my_algorithm) + extra_hdr_bytes,
XXX + old_size + extra_size)</strong></span>. This function should offer the
same synchronization as <code class="computeroutput"><span class="identifier">allocate</span><span class="special">()</span></code>.
</li>
</ul></div>
<p>
That's it. Now we can create new managed shared memory that uses our new
algorithm:
</p>
<pre class="programlisting"><span class="comment">//Managed memory segment to allocate named (c-string) objects</span>
<span class="comment">//using a user-defined memory allocation algorithm</span>
<span class="identifier">basic_managed_shared_memory</span><span class="special">&lt;</span><span class="keyword">char</span><span class="special">,</span>
<span class="special">,</span><span class="identifier">my_algorithm</span>
<span class="special">,</span><span class="identifier">flat_map_index</span><span class="special">&gt;</span>
<span class="identifier">my_managed_shared_memory</span><span class="special">;</span>
</pre>
</div>
<div class="section">
<div class="titlepage"><div><div><h3 class="title">
<a name="interprocess.customizing_interprocess.custom_allocators"></a><a class="link" href="customizing_interprocess.html#interprocess.customizing_interprocess.custom_allocators" title="Building custom STL compatible allocators for Boost.Interprocess">Building
custom STL compatible allocators for Boost.Interprocess</a>
</h3></div></div></div>
<p>
If provided STL-like allocators don't satisfy user needs, the user can implement
another STL compatible allocator using raw memory allocation and named object
construction functions. The user can this way implement more suitable allocation
schemes on top of basic shared memory allocation schemes, just like more
complex allocators are built on top of new/delete functions.
</p>
<p>
When using a managed memory segment, <span class="bold"><strong>get_segment_manager()</strong></span>
function returns a pointer to the segment manager. With this pointer, the
raw memory allocation and named object construction functions can be called
directly:
</p>
<pre class="programlisting"><span class="comment">//Create the managed shared memory and initialize resources</span>
<span class="identifier">managed_shared_memory</span> <span class="identifier">segment</span>
<span class="special">(</span><span class="identifier">create_only</span>
<span class="special">,</span><span class="string">"/MySharedMemory"</span> <span class="comment">//segment name</span>
<span class="special">,</span><span class="number">65536</span><span class="special">);</span> <span class="comment">//segment size in bytes</span>
<span class="comment">//Obtain the segment manager</span>
<span class="identifier">managed_shared_memory</span><span class="special">::</span><span class="identifier">segment_manager</span> <span class="special">*</span><span class="identifier">segment_mngr</span>
<span class="special">=</span> <span class="identifier">segment</span><span class="special">.</span><span class="identifier">get_segment_manager</span><span class="special">();</span>
<span class="comment">//With the segment manager, now we have access to all allocation functions</span>
<span class="identifier">segment_mngr</span><span class="special">-&gt;</span><span class="identifier">deallocate</span><span class="special">(</span><span class="identifier">segment_mngr</span><span class="special">-&gt;</span><span class="identifier">allocate</span><span class="special">(</span><span class="number">32</span><span class="special">));</span>
<span class="identifier">segment_mngr</span><span class="special">-&gt;</span><span class="identifier">construct</span><span class="special">&lt;</span><span class="keyword">int</span><span class="special">&gt;(</span><span class="string">"My_Int"</span><span class="special">)[</span><span class="number">32</span><span class="special">](</span><span class="number">0</span><span class="special">);</span>
<span class="identifier">segment_mngr</span><span class="special">-&gt;</span><span class="identifier">destroy</span><span class="special">&lt;</span><span class="keyword">int</span><span class="special">&gt;(</span><span class="string">"My_Int"</span><span class="special">);</span>
<span class="comment">//Initialize the custom, managed memory segment compatible</span>
<span class="comment">//allocator with the segment manager.</span>
<span class="comment">//</span>
<span class="comment">//MySTLAllocator uses segment_mngr-&gt;xxx functions to</span>
<span class="comment">//implement its allocation scheme</span>
<span class="identifier">MySTLAllocator</span><span class="special">&lt;</span><span class="keyword">int</span><span class="special">&gt;</span> <span class="identifier">stl_alloc</span><span class="special">(</span><span class="identifier">segment_mngr</span><span class="special">);</span>
<span class="comment">//Alias a new vector type that uses the custom STL compatible allocator</span>
<span class="keyword">typedef</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">vector</span><span class="special">&lt;</span><span class="keyword">int</span><span class="special">,</span> <span class="identifier">MySTLAllocator</span><span class="special">&lt;</span><span class="keyword">int</span><span class="special">&gt;</span> <span class="special">&gt;</span> <span class="identifier">MyVect</span><span class="special">;</span>
<span class="comment">//Construct the vector in shared memory with the allocator as constructor parameter</span>
<span class="identifier">segment</span><span class="special">.</span><span class="identifier">construct</span><span class="special">&lt;</span><span class="identifier">MyVect</span><span class="special">&gt;(</span><span class="string">"MyVect_instance"</span><span class="special">)(</span><span class="identifier">stl_alloc</span><span class="special">);</span>
</pre>
<p>
The user can create new STL compatible allocators that use the segment manager
to access to all memory management/object construction functions. All <span class="bold"><strong>Boost.Interprocess</strong></span>' STL compatible allocators are
based on this approach. <span class="bold"><strong>Remember</strong></span> that to
be compatible with managed memory segments, allocators should define their
<span class="bold"><strong>pointer</strong></span> typedef as the same pointer family
as <code class="computeroutput"><span class="identifier">segment_manager</span><span class="special">::</span><span class="identifier">void_pointer</span></code> typedef. This means that if
<code class="computeroutput"><span class="identifier">segment_manager</span><span class="special">::</span><span class="identifier">void_pointer</span></code> is <code class="computeroutput"><span class="identifier">offset_ptr</span><span class="special">&lt;</span><span class="keyword">void</span><span class="special">&gt;</span></code>,
<code class="computeroutput"><span class="identifier">MySTLAllocator</span><span class="special">&lt;</span><span class="keyword">int</span><span class="special">&gt;</span></code> should
define <code class="computeroutput"><span class="identifier">pointer</span></code> as <code class="computeroutput"><span class="identifier">offset_ptr</span><span class="special">&lt;</span><span class="keyword">int</span><span class="special">&gt;</span></code>. The
reason for this is that allocators are members of containers, and if we want
to put the container in a managed memory segment, the allocator should be
ready for that.
</p>
</div>
<div class="section">
<div class="titlepage"><div><div><h3 class="title">
<a name="interprocess.customizing_interprocess.custom_indexes"></a><a class="link" href="customizing_interprocess.html#interprocess.customizing_interprocess.custom_indexes" title="Building custom indexes">Building
custom indexes</a>
</h3></div></div></div>
<p>
The managed memory segment uses a name/object index to speed up object searching
and creation. Default specializations of managed memory segments (<code class="computeroutput"><span class="identifier">managed_shared_memory</span></code> for example), use
<code class="computeroutput"><span class="identifier">boost</span><span class="special">::</span><span class="identifier">interprocess</span><span class="special">::</span><span class="identifier">flat_map</span></code> as index.
</p>
<p>
However, the index type can be chosen via template parameter, so that the
user can define its own index type if he needs that. To construct a new index
type, the user must create a class with the following guidelines:
</p>
<div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem">
The interface of the index must follow the common public interface of
std::map and std::tr1::unordered_map including public typedefs. The
<code class="computeroutput"><span class="identifier">value_type</span></code> typedef can
be of type:
</li></ul></div>
<pre class="programlisting"><span class="identifier">std</span><span class="special">::</span><span class="identifier">pair</span><span class="special">&lt;</span><span class="identifier">key_type</span><span class="special">,</span> <span class="identifier">mapped_type</span><span class="special">&gt;</span>
</pre>
<p>
or
</p>
<pre class="programlisting"><span class="identifier">std</span><span class="special">::</span><span class="identifier">pair</span><span class="special">&lt;</span><span class="keyword">const</span> <span class="identifier">key_type</span><span class="special">,</span> <span class="identifier">mapped_type</span><span class="special">&gt;</span>
</pre>
<p>
so that ordered arrays or deques can be used as index types. Some known classes
following this basic interface are <code class="computeroutput"><span class="identifier">boost</span><span class="special">::</span><span class="identifier">unordered_map</span></code>,
<code class="computeroutput"><span class="identifier">boost</span><span class="special">::</span><span class="identifier">interprocess</span><span class="special">::</span><span class="identifier">flat_map</span></code> and <code class="computeroutput"><span class="identifier">boost</span><span class="special">::</span><span class="identifier">interprocess</span><span class="special">::</span><span class="identifier">map</span></code>.
</p>
<div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem">
The class must be a class template taking only a traits struct of this
type:
</li></ul></div>
<pre class="programlisting"><span class="keyword">struct</span> <span class="identifier">index_traits</span>
<span class="special">{</span>
<span class="keyword">typedef</span> <span class="comment">/*...*/</span> <span class="identifier">key_type</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="comment">/*...*/</span> <span class="identifier">mapped_type</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="comment">/*...*/</span> <span class="identifier">segment_manager</span><span class="special">;</span>
<span class="special">};</span>
</pre>
<pre class="programlisting"><span class="keyword">template</span> <span class="special">&lt;</span><span class="keyword">class</span> <span class="identifier">IndexTraits</span><span class="special">&gt;</span>
<span class="keyword">class</span> <span class="identifier">my_index_type</span><span class="special">;</span>
</pre>
<p>
The <code class="computeroutput"><span class="identifier">key_type</span></code> typedef of the
passed <code class="computeroutput"><span class="identifier">index_traits</span></code> will
be a specialization of the following class:
</p>
<pre class="programlisting"><span class="comment">//!The key of the named allocation information index. Stores a to</span>
<span class="comment">//!a null string and the length of the string to speed up sorting</span>
<span class="keyword">template</span><span class="special">&lt;...&gt;</span>
<span class="keyword">struct</span> <span class="identifier">index_key</span>
<span class="special">{</span>
<span class="keyword">typedef</span> <span class="comment">/*...*/</span> <span class="identifier">char_type</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="comment">/*...*/</span> <span class="identifier">const_char_ptr_t</span><span class="special">;</span>
<span class="comment">//Pointer to the object's name (null terminated)</span>
<span class="identifier">const_char_ptr_t</span> <span class="identifier">mp_str</span><span class="special">;</span>
<span class="comment">//Length of the name buffer (null NOT included)</span>
<span class="identifier">std</span><span class="special">::</span><span class="identifier">size_t</span> <span class="identifier">m_len</span><span class="special">;</span>
<span class="comment">//!Constructor of the key</span>
<span class="identifier">index_key</span> <span class="special">(</span><span class="keyword">const</span> <span class="identifier">CharT</span> <span class="special">*</span><span class="identifier">name</span><span class="special">,</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">size_t</span> <span class="identifier">length</span><span class="special">);</span>
<span class="comment">//!Less than function for index ordering</span>
<span class="keyword">bool</span> <span class="keyword">operator</span> <span class="special">&lt;</span> <span class="special">(</span><span class="keyword">const</span> <span class="identifier">index_key</span> <span class="special">&amp;</span> <span class="identifier">right</span><span class="special">)</span> <span class="keyword">const</span><span class="special">;</span>
<span class="comment">//!Equal to function for index ordering</span>
<span class="keyword">bool</span> <span class="keyword">operator</span> <span class="special">==</span> <span class="special">(</span><span class="keyword">const</span> <span class="identifier">index_key</span> <span class="special">&amp;</span> <span class="identifier">right</span><span class="special">)</span> <span class="keyword">const</span><span class="special">;</span>
<span class="special">};</span>
</pre>
<p>
The <code class="computeroutput"><span class="identifier">mapped_type</span></code> is not directly
modified by the customized index but it is needed to define the index type.
The <span class="bold"><strong>segment_manager</strong></span> will be the type of
the segment manager that will manage the index. <code class="computeroutput"><span class="identifier">segment_manager</span></code>
will define interesting internal types like <code class="computeroutput"><span class="identifier">void_pointer</span></code>
or <code class="computeroutput"><span class="identifier">mutex_family</span></code>.
</p>
<div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem">
The constructor of the customized index type must take a pointer to segment_manager
as constructor argument:
</li></ul></div>
<pre class="programlisting"><span class="identifier">constructor</span><span class="special">(</span><span class="identifier">segment_manager</span> <span class="special">*</span><span class="identifier">segment_mngr</span><span class="special">);</span>
</pre>
<div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem">
The index must provide a memory reservation function, that optimizes
the index if the user knows the number of elements to be inserted in
the index:
</li></ul></div>
<pre class="programlisting"><span class="keyword">void</span> <span class="identifier">reserve</span><span class="special">(</span><span class="identifier">std</span><span class="special">::</span><span class="identifier">size_t</span> <span class="identifier">n</span><span class="special">);</span>
</pre>
<p>
For example, the index type <code class="computeroutput"><span class="identifier">flat_map_index</span></code>
based in <code class="computeroutput"><span class="identifier">boost</span><span class="special">::</span><span class="identifier">interprocess</span><span class="special">::</span><span class="identifier">flat_map</span></code> is just defined as:
</p>
<p>
</p>
<pre class="programlisting"><span class="keyword">namespace</span> <span class="identifier">boost</span> <span class="special">{</span> <span class="keyword">namespace</span> <span class="identifier">interprocess</span> <span class="special">{</span>
<span class="preprocessor">#ifndef</span> <span class="identifier">BOOST_INTERPROCESS_DOXYGEN_INVOKED</span>
<span class="comment">//!Helper class to define typedefs from IndexTraits</span>
<span class="keyword">template</span> <span class="special">&lt;</span><span class="keyword">class</span> <span class="identifier">MapConfig</span><span class="special">&gt;</span>
<span class="keyword">struct</span> <span class="identifier">flat_map_index_aux</span>
<span class="special">{</span>
<span class="keyword">typedef</span> <span class="keyword">typename</span> <span class="identifier">MapConfig</span><span class="special">::</span><span class="identifier">key_type</span> <span class="identifier">key_type</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="keyword">typename</span> <span class="identifier">MapConfig</span><span class="special">::</span><span class="identifier">mapped_type</span> <span class="identifier">mapped_type</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="keyword">typename</span> <span class="identifier">MapConfig</span><span class="special">::</span>
<span class="identifier">segment_manager_base</span> <span class="identifier">segment_manager_base</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">less</span><span class="special">&lt;</span><span class="identifier">key_type</span><span class="special">&gt;</span> <span class="identifier">key_less</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="identifier">std</span><span class="special">::</span><span class="identifier">pair</span><span class="special">&lt;</span><span class="identifier">key_type</span><span class="special">,</span> <span class="identifier">mapped_type</span><span class="special">&gt;</span> <span class="identifier">value_type</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="identifier">allocator</span><span class="special">&lt;</span><span class="identifier">value_type</span>
<span class="special">,</span><span class="identifier">segment_manager_base</span><span class="special">&gt;</span> <span class="identifier">allocator_type</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="identifier">flat_map</span><span class="special">&lt;</span><span class="identifier">key_type</span><span class="special">,</span> <span class="identifier">mapped_type</span><span class="special">,</span>
<span class="identifier">key_less</span><span class="special">,</span> <span class="identifier">allocator_type</span><span class="special">&gt;</span> <span class="identifier">index_t</span><span class="special">;</span>
<span class="special">};</span>
<span class="preprocessor">#endif</span> <span class="comment">//#ifndef BOOST_INTERPROCESS_DOXYGEN_INVOKED</span>
<span class="comment">//!Index type based in flat_map. Just derives from flat_map and</span>
<span class="comment">//!defines the interface needed by managed memory segments.</span>
<span class="keyword">template</span> <span class="special">&lt;</span><span class="keyword">class</span> <span class="identifier">MapConfig</span><span class="special">&gt;</span>
<span class="keyword">class</span> <span class="identifier">flat_map_index</span>
<span class="comment">//Derive class from flat_map specialization</span>
<span class="special">:</span> <span class="keyword">public</span> <span class="identifier">flat_map_index_aux</span><span class="special">&lt;</span><span class="identifier">MapConfig</span><span class="special">&gt;::</span><span class="identifier">index_t</span>
<span class="special">{</span>
<span class="preprocessor">#if</span> <span class="special">!</span><span class="identifier">defined</span><span class="special">(</span><span class="identifier">BOOST_INTERPROCESS_DOXYGEN_INVOKED</span><span class="special">)</span>
<span class="keyword">typedef</span> <span class="identifier">flat_map_index_aux</span><span class="special">&lt;</span><span class="identifier">MapConfig</span><span class="special">&gt;</span> <span class="identifier">index_aux</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="keyword">typename</span> <span class="identifier">index_aux</span><span class="special">::</span><span class="identifier">index_t</span> <span class="identifier">base_type</span><span class="special">;</span>
<span class="keyword">typedef</span> <span class="keyword">typename</span> <span class="identifier">index_aux</span><span class="special">::</span>
<span class="identifier">segment_manager_base</span> <span class="identifier">segment_manager_base</span><span class="special">;</span>
<span class="preprocessor">#endif</span> <span class="comment">//#ifndef BOOST_INTERPROCESS_DOXYGEN_INVOKED</span>
<span class="keyword">public</span><span class="special">:</span>
<span class="comment">//!Constructor. Takes a pointer to the segment manager. Can throw</span>
<span class="identifier">flat_map_index</span><span class="special">(</span><span class="identifier">segment_manager_base</span> <span class="special">*</span><span class="identifier">segment_mngr</span><span class="special">)</span>
<span class="special">:</span> <span class="identifier">base_type</span><span class="special">(</span><span class="keyword">typename</span> <span class="identifier">index_aux</span><span class="special">::</span><span class="identifier">key_less</span><span class="special">(),</span>
<span class="keyword">typename</span> <span class="identifier">index_aux</span><span class="special">::</span><span class="identifier">allocator_type</span><span class="special">(</span><span class="identifier">segment_mngr</span><span class="special">))</span>
<span class="special">{}</span>
<span class="comment">//!This reserves memory to optimize the insertion of n elements in the index</span>
<span class="keyword">void</span> <span class="identifier">reserve</span><span class="special">(</span><span class="keyword">typename</span> <span class="identifier">segment_manager_base</span><span class="special">::</span><span class="identifier">size_type</span> <span class="identifier">n</span><span class="special">)</span>
<span class="special">{</span> <span class="identifier">base_type</span><span class="special">::</span><span class="identifier">reserve</span><span class="special">(</span><span class="identifier">n</span><span class="special">);</span> <span class="special">}</span>
<span class="comment">//!This frees all unnecessary memory</span>
<span class="keyword">void</span> <span class="identifier">shrink_to_fit</span><span class="special">()</span>
<span class="special">{</span> <span class="identifier">base_type</span><span class="special">::</span><span class="identifier">shrink_to_fit</span><span class="special">();</span> <span class="special">}</span>
<span class="special">};</span>
<span class="special">}}</span> <span class="comment">//namespace boost { namespace interprocess</span>
</pre>
<p>
</p>
<p>
If the user is defining a node container based index (a container whose iterators
are not invalidated when inserting or erasing other elements), <span class="bold"><strong>Boost.Interprocess</strong></span> can optimize named object destruction
when destructing via pointer. <span class="bold"><strong>Boost.Interprocess</strong></span>
can store an iterator next to the object and instead of using the name of
the object to erase the index entry, it uses the iterator, which is a faster
operation. So if you are creating a new node container based index (for example,
a tree), you should define an specialization of <code class="computeroutput"><span class="identifier">boost</span><span class="special">::</span><span class="identifier">interprocess</span><span class="special">::</span><span class="identifier">is_node_index</span><span class="special">&lt;...&gt;</span></code> defined in <code class="computeroutput"><span class="special">&lt;</span><span class="identifier">boost</span><span class="special">/</span><span class="identifier">interprocess</span><span class="special">/</span><span class="identifier">detail</span><span class="special">/</span><span class="identifier">utilities</span><span class="special">.</span><span class="identifier">hpp</span><span class="special">&gt;</span></code>:
</p>
<pre class="programlisting"><span class="comment">//!Trait classes to detect if an index is a node</span>
<span class="comment">//!index. This allows more efficient operations</span>
<span class="comment">//!when deallocating named objects.</span>
<span class="keyword">template</span><span class="special">&lt;</span><span class="keyword">class</span> <span class="identifier">MapConfig</span><span class="special">&gt;</span>
<span class="keyword">struct</span> <span class="identifier">is_node_index</span>
<span class="special">&lt;</span><span class="identifier">my_index</span><span class="special">&lt;</span><span class="identifier">MapConfig</span><span class="special">&gt;</span> <span class="special">&gt;</span>
<span class="special">{</span>
<span class="keyword">static</span> <span class="keyword">const</span> <span class="keyword">bool</span> <span class="identifier">value</span> <span class="special">=</span> <span class="keyword">true</span><span class="special">;</span>
<span class="special">};</span>
</pre>
<p>
Interprocess also defines other index types:
</p>
<div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; ">
<li class="listitem">
<span class="bold"><strong>boost::map_index</strong></span> uses <span class="bold"><strong>boost::interprocess::map</strong></span>
as index type.
</li>
<li class="listitem">
<span class="bold"><strong>boost::null_index</strong></span> that uses an dummy
index type if the user just needs anonymous allocations and wants to
save some space and class instantations.
</li>
</ul></div>
<p>
Defining a new managed memory segment that uses the new index is easy. For
example, a new managed shared memory that uses the new index:
</p>
<pre class="programlisting"><span class="comment">//!Defines a managed shared memory with a c-strings as</span>
<span class="comment">//!a keys, the red-black tree best fit algorithm (with process-shared mutexes</span>
<span class="comment">//!and offset_ptr pointers) as raw shared memory management algorithm</span>
<span class="comment">//!and a custom index</span>
<span class="keyword">typedef</span>
<span class="identifier">basic_managed_shared_memory</span> <span class="special">&lt;</span>
<span class="keyword">char</span><span class="special">,</span>
<span class="identifier">rbtree_best_fit</span><span class="special">&lt;</span><span class="identifier">mutex_family</span><span class="special">&gt;,</span>
<span class="identifier">my_index_type</span>
<span class="special">&gt;</span>
<span class="identifier">my_managed_shared_memory</span><span class="special">;</span>
</pre>
</div>
</div>
<table xmlns:rev="http://www.cs.rpi.edu/~gregod/boost/tools/doc/revision" width="100%"><tr>
<td align="left"></td>
<td align="right"><div class="copyright-footer">Copyright &#169; 2005-2012 Ion Gaztanaga<p>
Distributed under the Boost Software License, Version 1.0. (See accompanying
file LICENSE_1_0.txt or copy at <a href="http://www.boost.org/LICENSE_1_0.txt" target="_top">http://www.boost.org/LICENSE_1_0.txt</a>)
</p>
</div></td>
</tr></table>
<hr>
<div class="spirit-nav">
<a accesskey="p" href="architecture.html"><img src="../../../doc/src/images/prev.png" alt="Prev"></a><a accesskey="u" href="../interprocess.html"><img src="../../../doc/src/images/up.png" alt="Up"></a><a accesskey="h" href="../index.html"><img src="../../../doc/src/images/home.png" alt="Home"></a><a accesskey="n" href="acknowledgements_notes.html"><img src="../../../doc/src/images/next.png" alt="Next"></a>
</div>
</body>
</html>