xref: /redis-3.2.3/deps/jemalloc/doc/jemalloc.html (revision ad4c0b41)
1<html><head><meta http-equiv="Content-Type" content="text/html; charset=ISO-8859-1"><title>JEMALLOC</title><meta name="generator" content="DocBook XSL Stylesheets V1.76.1"></head><body bgcolor="white" text="black" link="#0000FF" vlink="#840084" alink="#0000FF"><div class="refentry" title="JEMALLOC"><a name="id286949398"></a><div class="titlepage"></div><div class="refnamediv"><h2>Name</h2><p>jemalloc &#8212; general purpose memory allocation functions</p></div><div class="refsect1" title="LIBRARY"><a name="library"></a><h2>LIBRARY</h2><p>This manual describes jemalloc 3.0.0-0-gfc9b1dbf69f59d7ecfc4ac68da9847e017e1d046.  More information
2    can be found at the <a class="ulink" href="http://www.canonware.com/jemalloc/" target="_top">jemalloc website</a>.</p></div><div class="refsynopsisdiv" title="SYNOPSIS"><h2>SYNOPSIS</h2><div class="funcsynopsis"><pre class="funcsynopsisinfo">#include &lt;<code class="filename">stdlib.h</code>&gt;
3#include &lt;<code class="filename">jemalloc/jemalloc.h</code>&gt;</pre><div class="refsect2" title="Standard API"><a name="id286949667"></a><h3>Standard API</h3><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">void *<b class="fsfunc">malloc</b>(</code></td><td>size_t <var class="pdparam">size</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">void *<b class="fsfunc">calloc</b>(</code></td><td>size_t <var class="pdparam">number</var>, </td></tr><tr><td>�</td><td>size_t <var class="pdparam">size</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">int <b class="fsfunc">posix_memalign</b>(</code></td><td>void **<var class="pdparam">ptr</var>, </td></tr><tr><td>�</td><td>size_t <var class="pdparam">alignment</var>, </td></tr><tr><td>�</td><td>size_t <var class="pdparam">size</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">void *<b class="fsfunc">aligned_alloc</b>(</code></td><td>size_t <var class="pdparam">alignment</var>, </td></tr><tr><td>�</td><td>size_t <var class="pdparam">size</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">void *<b class="fsfunc">realloc</b>(</code></td><td>void *<var class="pdparam">ptr</var>, </td></tr><tr><td>�</td><td>size_t <var class="pdparam">size</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">void <b class="fsfunc">free</b>(</code></td><td>void *<var class="pdparam">ptr</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div></div><div class="refsect2" title="Non-standard API"><a name="id286949803"></a><h3>Non-standard API</h3><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">size_t <b class="fsfunc">malloc_usable_size</b>(</code></td><td>const void *<var class="pdparam">ptr</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">void <b class="fsfunc">malloc_stats_print</b>(</code></td><td>void <var class="pdparam">(*write_cb)</var>
4            <code>(</code>void *, const char *<code>)</code>
5          , </td></tr><tr><td>�</td><td>void *<var class="pdparam">cbopaque</var>, </td></tr><tr><td>�</td><td>const char *<var class="pdparam">opts</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">int <b class="fsfunc">mallctl</b>(</code></td><td>const char *<var class="pdparam">name</var>, </td></tr><tr><td>�</td><td>void *<var class="pdparam">oldp</var>, </td></tr><tr><td>�</td><td>size_t *<var class="pdparam">oldlenp</var>, </td></tr><tr><td>�</td><td>void *<var class="pdparam">newp</var>, </td></tr><tr><td>�</td><td>size_t <var class="pdparam">newlen</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">int <b class="fsfunc">mallctlnametomib</b>(</code></td><td>const char *<var class="pdparam">name</var>, </td></tr><tr><td>�</td><td>size_t *<var class="pdparam">mibp</var>, </td></tr><tr><td>�</td><td>size_t *<var class="pdparam">miblenp</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">int <b class="fsfunc">mallctlbymib</b>(</code></td><td>const size_t *<var class="pdparam">mib</var>, </td></tr><tr><td>�</td><td>size_t <var class="pdparam">miblen</var>, </td></tr><tr><td>�</td><td>void *<var class="pdparam">oldp</var>, </td></tr><tr><td>�</td><td>size_t *<var class="pdparam">oldlenp</var>, </td></tr><tr><td>�</td><td>void *<var class="pdparam">newp</var>, </td></tr><tr><td>�</td><td>size_t <var class="pdparam">newlen</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">void <b class="fsfunc">(*malloc_message)</b>(</code></td><td>void *<var class="pdparam">cbopaque</var>, </td></tr><tr><td>�</td><td>const char *<var class="pdparam">s</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><p><span class="type">const char *</span><code class="varname">malloc_conf</code>;</p></div><div class="refsect2" title="Experimental API"><a name="id286950044"></a><h3>Experimental API</h3><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">int <b class="fsfunc">allocm</b>(</code></td><td>void **<var class="pdparam">ptr</var>, </td></tr><tr><td>�</td><td>size_t *<var class="pdparam">rsize</var>, </td></tr><tr><td>�</td><td>size_t <var class="pdparam">size</var>, </td></tr><tr><td>�</td><td>int <var class="pdparam">flags</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">int <b class="fsfunc">rallocm</b>(</code></td><td>void **<var class="pdparam">ptr</var>, </td></tr><tr><td>�</td><td>size_t *<var class="pdparam">rsize</var>, </td></tr><tr><td>�</td><td>size_t <var class="pdparam">size</var>, </td></tr><tr><td>�</td><td>size_t <var class="pdparam">extra</var>, </td></tr><tr><td>�</td><td>int <var class="pdparam">flags</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">int <b class="fsfunc">sallocm</b>(</code></td><td>const void *<var class="pdparam">ptr</var>, </td></tr><tr><td>�</td><td>size_t *<var class="pdparam">rsize</var>, </td></tr><tr><td>�</td><td>int <var class="pdparam">flags</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">int <b class="fsfunc">dallocm</b>(</code></td><td>void *<var class="pdparam">ptr</var>, </td></tr><tr><td>�</td><td>int <var class="pdparam">flags</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div><table border="0" summary="Function synopsis" cellspacing="0" cellpadding="0" class="funcprototype-table"><tr><td><code class="funcdef">int <b class="fsfunc">nallocm</b>(</code></td><td>size_t *<var class="pdparam">rsize</var>, </td></tr><tr><td>�</td><td>size_t <var class="pdparam">size</var>, </td></tr><tr><td>�</td><td>int <var class="pdparam">flags</var><code>)</code>;</td></tr></table><div class="funcprototype-spacer">�</div></div></div></div><div class="refsect1" title="DESCRIPTION"><a name="description"></a><h2>DESCRIPTION</h2><div class="refsect2" title="Standard API"><a name="id286950218"></a><h3>Standard API</h3><p>The <code class="function">malloc</code>(<em class="parameter"><code></code></em>) function allocates
6      <em class="parameter"><code>size</code></em> bytes of uninitialized memory.  The allocated
7      space is suitably aligned (after possible pointer coercion) for storage
8      of any type of object.</p><p>The <code class="function">calloc</code>(<em class="parameter"><code></code></em>) function allocates
9      space for <em class="parameter"><code>number</code></em> objects, each
10      <em class="parameter"><code>size</code></em> bytes in length.  The result is identical to
11      calling <code class="function">malloc</code>(<em class="parameter"><code></code></em>) with an argument of
12      <em class="parameter"><code>number</code></em> * <em class="parameter"><code>size</code></em>, with the
13      exception that the allocated memory is explicitly initialized to zero
14      bytes.</p><p>The <code class="function">posix_memalign</code>(<em class="parameter"><code></code></em>) function
15      allocates <em class="parameter"><code>size</code></em> bytes of memory such that the
16      allocation's base address is an even multiple of
17      <em class="parameter"><code>alignment</code></em>, and returns the allocation in the value
18      pointed to by <em class="parameter"><code>ptr</code></em>.  The requested
19      <em class="parameter"><code>alignment</code></em> must be a power of 2 at least as large
20      as <code class="code">sizeof(<span class="type">void *</span>)</code>.</p><p>The <code class="function">aligned_alloc</code>(<em class="parameter"><code></code></em>) function
21      allocates <em class="parameter"><code>size</code></em> bytes of memory such that the
22      allocation's base address is an even multiple of
23      <em class="parameter"><code>alignment</code></em>.  The requested
24      <em class="parameter"><code>alignment</code></em> must be a power of 2.  Behavior is
25      undefined if <em class="parameter"><code>size</code></em> is not an integral multiple of
26      <em class="parameter"><code>alignment</code></em>.</p><p>The <code class="function">realloc</code>(<em class="parameter"><code></code></em>) function changes the
27      size of the previously allocated memory referenced by
28      <em class="parameter"><code>ptr</code></em> to <em class="parameter"><code>size</code></em> bytes.  The
29      contents of the memory are unchanged up to the lesser of the new and old
30      sizes.  If the new size is larger, the contents of the newly allocated
31      portion of the memory are undefined.  Upon success, the memory referenced
32      by <em class="parameter"><code>ptr</code></em> is freed and a pointer to the newly
33      allocated memory is returned.  Note that
34      <code class="function">realloc</code>(<em class="parameter"><code></code></em>) may move the memory allocation,
35      resulting in a different return value than <em class="parameter"><code>ptr</code></em>.
36      If <em class="parameter"><code>ptr</code></em> is <code class="constant">NULL</code>, the
37      <code class="function">realloc</code>(<em class="parameter"><code></code></em>) function behaves identically to
38      <code class="function">malloc</code>(<em class="parameter"><code></code></em>) for the specified size.</p><p>The <code class="function">free</code>(<em class="parameter"><code></code></em>) function causes the
39      allocated memory referenced by <em class="parameter"><code>ptr</code></em> to be made
40      available for future allocations.  If <em class="parameter"><code>ptr</code></em> is
41      <code class="constant">NULL</code>, no action occurs.</p></div><div class="refsect2" title="Non-standard API"><a name="id286950483"></a><h3>Non-standard API</h3><p>The <code class="function">malloc_usable_size</code>(<em class="parameter"><code></code></em>) function
42      returns the usable size of the allocation pointed to by
43      <em class="parameter"><code>ptr</code></em>.  The return value may be larger than the size
44      that was requested during allocation.  The
45      <code class="function">malloc_usable_size</code>(<em class="parameter"><code></code></em>) function is not a
46      mechanism for in-place <code class="function">realloc</code>(<em class="parameter"><code></code></em>); rather
47      it is provided solely as a tool for introspection purposes.  Any
48      discrepancy between the requested allocation size and the size reported
49      by <code class="function">malloc_usable_size</code>(<em class="parameter"><code></code></em>) should not be
50      depended on, since such behavior is entirely implementation-dependent.
51      </p><p>The <code class="function">malloc_stats_print</code>(<em class="parameter"><code></code></em>) function
52      writes human-readable summary statistics via the
53      <em class="parameter"><code>write_cb</code></em> callback function pointer and
54      <em class="parameter"><code>cbopaque</code></em> data passed to
55      <em class="parameter"><code>write_cb</code></em>, or
56      <code class="function">malloc_message</code>(<em class="parameter"><code></code></em>) if
57      <em class="parameter"><code>write_cb</code></em> is <code class="constant">NULL</code>.  This
58      function can be called repeatedly.  General information that never
59      changes during execution can be omitted by specifying "g" as a character
60      within the <em class="parameter"><code>opts</code></em> string.  Note that
61      <code class="function">malloc_message</code>(<em class="parameter"><code></code></em>) uses the
62      <code class="function">mallctl*</code>(<em class="parameter"><code></code></em>) functions internally, so
63      inconsistent statistics can be reported if multiple threads use these
64      functions simultaneously.  If <code class="option">--enable-stats</code> is
65      specified during configuration, &#8220;m&#8221; and &#8220;a&#8221; can
66      be specified to omit merged arena and per arena statistics, respectively;
67      &#8220;b&#8221; and &#8220;l&#8221; can be specified to omit per size
68      class statistics for bins and large objects, respectively.  Unrecognized
69      characters are silently ignored.  Note that thread caching may prevent
70      some statistics from being completely up to date, since extra locking
71      would be required to merge counters that track thread cache operations.
72      </p><p>The <code class="function">mallctl</code>(<em class="parameter"><code></code></em>) function provides a
73      general interface for introspecting the memory allocator, as well as
74      setting modifiable parameters and triggering actions.  The
75      period-separated <em class="parameter"><code>name</code></em> argument specifies a
76      location in a tree-structured namespace; see the <a class="xref" href="#mallctl_namespace" title="MALLCTL NAMESPACE">MALLCTL NAMESPACE</a> section for
77      documentation on the tree contents.  To read a value, pass a pointer via
78      <em class="parameter"><code>oldp</code></em> to adequate space to contain the value, and a
79      pointer to its length via <em class="parameter"><code>oldlenp</code></em>; otherwise pass
80      <code class="constant">NULL</code> and <code class="constant">NULL</code>.  Similarly, to
81      write a value, pass a pointer to the value via
82      <em class="parameter"><code>newp</code></em>, and its length via
83      <em class="parameter"><code>newlen</code></em>; otherwise pass <code class="constant">NULL</code>
84      and <code class="constant">0</code>.</p><p>The <code class="function">mallctlnametomib</code>(<em class="parameter"><code></code></em>) function
85      provides a way to avoid repeated name lookups for applications that
86      repeatedly query the same portion of the namespace, by translating a name
87      to a &#8220;Management Information Base&#8221; (MIB) that can be passed
88      repeatedly to <code class="function">mallctlbymib</code>(<em class="parameter"><code></code></em>).  Upon
89      successful return from <code class="function">mallctlnametomib</code>(<em class="parameter"><code></code></em>),
90      <em class="parameter"><code>mibp</code></em> contains an array of
91      <em class="parameter"><code>*miblenp</code></em> integers, where
92      <em class="parameter"><code>*miblenp</code></em> is the lesser of the number of components
93      in <em class="parameter"><code>name</code></em> and the input value of
94      <em class="parameter"><code>*miblenp</code></em>.  Thus it is possible to pass a
95      <em class="parameter"><code>*miblenp</code></em> that is smaller than the number of
96      period-separated name components, which results in a partial MIB that can
97      be used as the basis for constructing a complete MIB.  For name
98      components that are integers (e.g. the 2 in
99      <a class="link" href="#arenas.bin.i.size">
100    "<code class="mallctl">arenas.bin.2.size</code>"
101  </a>),
102      the corresponding MIB component will always be that integer.  Therefore,
103      it is legitimate to construct code like the following: </p><pre class="programlisting">
104unsigned nbins, i;
105
106int mib[4];
107size_t len, miblen;
108
109len = sizeof(nbins);
110mallctl("arenas.nbins", &amp;nbins, &amp;len, NULL, 0);
111
112miblen = 4;
113mallnametomib("arenas.bin.0.size", mib, &amp;miblen);
114for (i = 0; i &lt; nbins; i++) {
115	size_t bin_size;
116
117	mib[2] = i;
118	len = sizeof(bin_size);
119	mallctlbymib(mib, miblen, &amp;bin_size, &amp;len, NULL, 0);
120	/* Do something with bin_size... */
121}</pre></div><div class="refsect2" title="Experimental API"><a name="id286950724"></a><h3>Experimental API</h3><p>The experimental API is subject to change or removal without regard
122      for backward compatibility.  If <code class="option">--disable-experimental</code>
123      is specified during configuration, the experimental API is
124      omitted.</p><p>The <code class="function">allocm</code>(<em class="parameter"><code></code></em>),
125      <code class="function">rallocm</code>(<em class="parameter"><code></code></em>),
126      <code class="function">sallocm</code>(<em class="parameter"><code></code></em>),
127      <code class="function">dallocm</code>(<em class="parameter"><code></code></em>), and
128      <code class="function">nallocm</code>(<em class="parameter"><code></code></em>) functions all have a
129      <em class="parameter"><code>flags</code></em> argument that can be used to specify
130      options.  The functions only check the options that are contextually
131      relevant.  Use bitwise or (<code class="code">|</code>) operations to
132      specify one or more of the following:
133        </p><div class="variablelist"><dl><dt><span class="term"><code class="constant">ALLOCM_LG_ALIGN(<em class="parameter"><code>la</code></em>)
134            </code></span></dt><dd><p>Align the memory allocation to start at an address
135            that is a multiple of <code class="code">(1 &lt;&lt;
136            <em class="parameter"><code>la</code></em>)</code>.  This macro does not validate
137            that <em class="parameter"><code>la</code></em> is within the valid
138            range.</p></dd><dt><span class="term"><code class="constant">ALLOCM_ALIGN(<em class="parameter"><code>a</code></em>)
139            </code></span></dt><dd><p>Align the memory allocation to start at an address
140            that is a multiple of <em class="parameter"><code>a</code></em>, where
141            <em class="parameter"><code>a</code></em> is a power of two.  This macro does not
142            validate that <em class="parameter"><code>a</code></em> is a power of 2.
143            </p></dd><dt><span class="term"><code class="constant">ALLOCM_ZERO</code></span></dt><dd><p>Initialize newly allocated memory to contain zero
144            bytes.  In the growing reallocation case, the real size prior to
145            reallocation defines the boundary between untouched bytes and those
146            that are initialized to contain zero bytes.  If this option is
147            absent, newly allocated memory is uninitialized.</p></dd><dt><span class="term"><code class="constant">ALLOCM_NO_MOVE</code></span></dt><dd><p>For reallocation, fail rather than moving the
148            object.  This constraint can apply to both growth and
149            shrinkage.</p></dd></dl></div><p>
150      </p><p>The <code class="function">allocm</code>(<em class="parameter"><code></code></em>) function allocates at
151      least <em class="parameter"><code>size</code></em> bytes of memory, sets
152      <em class="parameter"><code>*ptr</code></em> to the base address of the allocation, and
153      sets <em class="parameter"><code>*rsize</code></em> to the real size of the allocation if
154      <em class="parameter"><code>rsize</code></em> is not <code class="constant">NULL</code>.  Behavior
155      is undefined if <em class="parameter"><code>size</code></em> is
156      <code class="constant">0</code>.</p><p>The <code class="function">rallocm</code>(<em class="parameter"><code></code></em>) function resizes the
157      allocation at <em class="parameter"><code>*ptr</code></em> to be at least
158      <em class="parameter"><code>size</code></em> bytes, sets <em class="parameter"><code>*ptr</code></em> to
159      the base address of the allocation if it moved, and sets
160      <em class="parameter"><code>*rsize</code></em> to the real size of the allocation if
161      <em class="parameter"><code>rsize</code></em> is not <code class="constant">NULL</code>.  If
162      <em class="parameter"><code>extra</code></em> is non-zero, an attempt is made to resize
163      the allocation to be at least <code class="code"><em class="parameter"><code>size</code></em> +
164      <em class="parameter"><code>extra</code></em>)</code> bytes, though inability to allocate
165      the extra byte(s) will not by itself result in failure.  Behavior is
166      undefined if <em class="parameter"><code>size</code></em> is <code class="constant">0</code>, or if
167      <code class="code">(<em class="parameter"><code>size</code></em> +
168      <em class="parameter"><code>extra</code></em> &gt;
169      <code class="constant">SIZE_T_MAX</code>)</code>.</p><p>The <code class="function">sallocm</code>(<em class="parameter"><code></code></em>) function sets
170      <em class="parameter"><code>*rsize</code></em> to the real size of the allocation.</p><p>The <code class="function">dallocm</code>(<em class="parameter"><code></code></em>) function causes the
171      memory referenced by <em class="parameter"><code>ptr</code></em> to be made available for
172      future allocations.</p><p>The <code class="function">nallocm</code>(<em class="parameter"><code></code></em>) function allocates no
173      memory, but it performs the same size computation as the
174      <code class="function">allocm</code>(<em class="parameter"><code></code></em>) function, and if
175      <em class="parameter"><code>rsize</code></em> is not <code class="constant">NULL</code> it sets
176      <em class="parameter"><code>*rsize</code></em> to the real size of the allocation that
177      would result from the equivalent <code class="function">allocm</code>(<em class="parameter"><code></code></em>)
178      function call.  Behavior is undefined if
179      <em class="parameter"><code>size</code></em> is <code class="constant">0</code>.</p></div></div><div class="refsect1" title="TUNING"><a name="tuning"></a><h2>TUNING</h2><p>Once, when the first call is made to one of the memory allocation
180    routines, the allocator initializes its internals based in part on various
181    options that can be specified at compile- or run-time.</p><p>The string pointed to by the global variable
182    <code class="varname">malloc_conf</code>, the &#8220;name&#8221; of the file
183    referenced by the symbolic link named <code class="filename">/etc/malloc.conf</code>, and the value of the
184    environment variable <code class="envar">MALLOC_CONF</code>, will be interpreted, in
185    that order, from left to right as options.</p><p>An options string is a comma-separated list of option:value pairs.
186    There is one key corresponding to each <a class="link" href="#opt.abort">
187    "<code class="mallctl">opt.*</code>"
188  </a> mallctl (see the <a class="xref" href="#mallctl_namespace" title="MALLCTL NAMESPACE">MALLCTL NAMESPACE</a> section for options
189    documentation).  For example, <code class="literal">abort:true,narenas:1</code> sets
190    the <a class="link" href="#opt.abort">
191    "<code class="mallctl">opt.abort</code>"
192  </a> and <a class="link" href="#opt.narenas">
193    "<code class="mallctl">opt.narenas</code>"
194  </a> options.  Some
195    options have boolean values (true/false), others have integer values (base
196    8, 10, or 16, depending on prefix), and yet others have raw string
197    values.</p></div><div class="refsect1" title="IMPLEMENTATION NOTES"><a name="implementation_notes"></a><h2>IMPLEMENTATION NOTES</h2><p>Traditionally, allocators have used
198    <span class="citerefentry"><span class="refentrytitle">sbrk</span>(2)</span> to obtain memory, which is
199    suboptimal for several reasons, including race conditions, increased
200    fragmentation, and artificial limitations on maximum usable memory.  If
201    <code class="option">--enable-dss</code> is specified during configuration, this
202    allocator uses both <span class="citerefentry"><span class="refentrytitle">mmap</span>(2)</span> and
203    <span class="citerefentry"><span class="refentrytitle">sbrk</span>(2)</span>, in that order of preference;
204    otherwise only <span class="citerefentry"><span class="refentrytitle">mmap</span>(2)</span> is used.</p><p>This allocator uses multiple arenas in order to reduce lock
205    contention for threaded programs on multi-processor systems.  This works
206    well with regard to threading scalability, but incurs some costs.  There is
207    a small fixed per-arena overhead, and additionally, arenas manage memory
208    completely independently of each other, which means a small fixed increase
209    in overall memory fragmentation.  These overheads are not generally an
210    issue, given the number of arenas normally used.  Note that using
211    substantially more arenas than the default is not likely to improve
212    performance, mainly due to reduced cache performance.  However, it may make
213    sense to reduce the number of arenas if an application does not make much
214    use of the allocation functions.</p><p>In addition to multiple arenas, unless
215    <code class="option">--disable-tcache</code> is specified during configuration, this
216    allocator supports thread-specific caching for small and large objects, in
217    order to make it possible to completely avoid synchronization for most
218    allocation requests.  Such caching allows very fast allocation in the
219    common case, but it increases memory usage and fragmentation, since a
220    bounded number of objects can remain allocated in each thread cache.</p><p>Memory is conceptually broken into equal-sized chunks, where the
221    chunk size is a power of two that is greater than the page size.  Chunks
222    are always aligned to multiples of the chunk size.  This alignment makes it
223    possible to find metadata for user objects very quickly.</p><p>User objects are broken into three categories according to size:
224    small, large, and huge.  Small objects are smaller than one page.  Large
225    objects are smaller than the chunk size.  Huge objects are a multiple of
226    the chunk size.  Small and large objects are managed by arenas; huge
227    objects are managed separately in a single data structure that is shared by
228    all threads.  Huge objects are used by applications infrequently enough
229    that this single data structure is not a scalability issue.</p><p>Each chunk that is managed by an arena tracks its contents as runs of
230    contiguous pages (unused, backing a set of small objects, or backing one
231    large object).  The combination of chunk alignment and chunk page maps
232    makes it possible to determine all metadata regarding small and large
233    allocations in constant time.</p><p>Small objects are managed in groups by page runs.  Each run maintains
234    a frontier and free list to track which regions are in use.  Allocation
235    requests that are no more than half the quantum (8 or 16, depending on
236    architecture) are rounded up to the nearest power of two that is at least
237    <code class="code">sizeof(<span class="type">double</span>)</code>.  All other small
238    object size classes are multiples of the quantum, spaced such that internal
239    fragmentation is limited to approximately 25% for all but the smallest size
240    classes.  Allocation requests that are larger than the maximum small size
241    class, but small enough to fit in an arena-managed chunk (see the <a class="link" href="#opt.lg_chunk">
242    "<code class="mallctl">opt.lg_chunk</code>"
243  </a> option), are
244    rounded up to the nearest run size.  Allocation requests that are too large
245    to fit in an arena-managed chunk are rounded up to the nearest multiple of
246    the chunk size.</p><p>Allocations are packed tightly together, which can be an issue for
247    multi-threaded applications.  If you need to assure that allocations do not
248    suffer from cacheline sharing, round your allocation requests up to the
249    nearest multiple of the cacheline size, or specify cacheline alignment when
250    allocating.</p><p>Assuming 4 MiB chunks, 4 KiB pages, and a 16-byte quantum on a 64-bit
251    system, the size classes in each category are as shown in <a class="xref" href="#size_classes" title="Table�1.�Size classes">Table 1</a>.</p><div class="table"><a name="size_classes"></a><p class="title"><b>Table�1.�Size classes</b></p><div class="table-contents"><table summary="Size classes" border="1"><colgroup><col align="left" class="c1"><col align="right" class="c2"><col align="left" class="c3"></colgroup><thead><tr><th align="left">Category</th><th align="right">Spacing</th><th align="left">Size</th></tr></thead><tbody><tr><td rowspan="7" align="left">Small</td><td align="right">lg</td><td align="left">[8]</td></tr><tr><td align="right">16</td><td align="left">[16, 32, 48, ..., 128]</td></tr><tr><td align="right">32</td><td align="left">[160, 192, 224, 256]</td></tr><tr><td align="right">64</td><td align="left">[320, 384, 448, 512]</td></tr><tr><td align="right">128</td><td align="left">[640, 768, 896, 1024]</td></tr><tr><td align="right">256</td><td align="left">[1280, 1536, 1792, 2048]</td></tr><tr><td align="right">512</td><td align="left">[2560, 3072, 3584]</td></tr><tr><td align="left">Large</td><td align="right">4 KiB</td><td align="left">[4 KiB, 8 KiB, 12 KiB, ..., 4072 KiB]</td></tr><tr><td align="left">Huge</td><td align="right">4 MiB</td><td align="left">[4 MiB, 8 MiB, 12 MiB, ...]</td></tr></tbody></table></div></div><br class="table-break"></div><div class="refsect1" title="MALLCTL NAMESPACE"><a name="mallctl_namespace"></a><h2>MALLCTL NAMESPACE</h2><p>The following names are defined in the namespace accessible via the
252    <code class="function">mallctl*</code>(<em class="parameter"><code></code></em>) functions.  Value types are
253    specified in parentheses, their readable/writable statuses are encoded as
254    <code class="literal">rw</code>, <code class="literal">r-</code>, <code class="literal">-w</code>, or
255    <code class="literal">--</code>, and required build configuration flags follow, if
256    any.  A name element encoded as <code class="literal">&lt;i&gt;</code> or
257    <code class="literal">&lt;j&gt;</code> indicates an integer component, where the
258    integer varies from 0 to some upper value that must be determined via
259    introspection.  In the case of
260    "<code class="mallctl">stats.arenas.&lt;i&gt;.*</code>"
261  ,
262    <code class="literal">&lt;i&gt;</code> equal to <a class="link" href="#arenas.narenas">
263    "<code class="mallctl">arenas.narenas</code>"
264  </a> can be
265    used to access the summation of statistics from all arenas.  Take special
266    note of the <a class="link" href="#epoch">
267    "<code class="mallctl">epoch</code>"
268  </a> mallctl,
269    which controls refreshing of cached dynamic statistics.</p><div class="variablelist"><dl><dt><span class="term">
270
271    "<code class="mallctl">version</code>"
272
273          (<span class="type">const char *</span>)
274          <code class="literal">r-</code>
275        </span></dt><dd><p>Return the jemalloc version string.</p></dd><dt><a name="epoch"></a><span class="term">
276
277    "<code class="mallctl">epoch</code>"
278
279          (<span class="type">uint64_t</span>)
280          <code class="literal">rw</code>
281        </span></dt><dd><p>If a value is passed in, refresh the data from which
282        the <code class="function">mallctl*</code>(<em class="parameter"><code></code></em>) functions report values,
283        and increment the epoch.  Return the current epoch.  This is useful for
284        detecting whether another thread caused a refresh.</p></dd><dt><span class="term">
285
286    "<code class="mallctl">config.debug</code>"
287
288          (<span class="type">bool</span>)
289          <code class="literal">r-</code>
290        </span></dt><dd><p><code class="option">--enable-debug</code> was specified during
291        build configuration.</p></dd><dt><span class="term">
292
293    "<code class="mallctl">config.dss</code>"
294
295          (<span class="type">bool</span>)
296          <code class="literal">r-</code>
297        </span></dt><dd><p><code class="option">--enable-dss</code> was specified during
298        build configuration.</p></dd><dt><span class="term">
299
300    "<code class="mallctl">config.fill</code>"
301
302          (<span class="type">bool</span>)
303          <code class="literal">r-</code>
304        </span></dt><dd><p><code class="option">--enable-fill</code> was specified during
305        build configuration.</p></dd><dt><span class="term">
306
307    "<code class="mallctl">config.lazy_lock</code>"
308
309          (<span class="type">bool</span>)
310          <code class="literal">r-</code>
311        </span></dt><dd><p><code class="option">--enable-lazy-lock</code> was specified
312        during build configuration.</p></dd><dt><span class="term">
313
314    "<code class="mallctl">config.mremap</code>"
315
316          (<span class="type">bool</span>)
317          <code class="literal">r-</code>
318        </span></dt><dd><p><code class="option">--enable-mremap</code> was specified during
319        build configuration.</p></dd><dt><span class="term">
320
321    "<code class="mallctl">config.munmap</code>"
322
323          (<span class="type">bool</span>)
324          <code class="literal">r-</code>
325        </span></dt><dd><p><code class="option">--enable-munmap</code> was specified during
326        build configuration.</p></dd><dt><span class="term">
327
328    "<code class="mallctl">config.prof</code>"
329
330          (<span class="type">bool</span>)
331          <code class="literal">r-</code>
332        </span></dt><dd><p><code class="option">--enable-prof</code> was specified during
333        build configuration.</p></dd><dt><span class="term">
334
335    "<code class="mallctl">config.prof_libgcc</code>"
336
337          (<span class="type">bool</span>)
338          <code class="literal">r-</code>
339        </span></dt><dd><p><code class="option">--disable-prof-libgcc</code> was not
340        specified during build configuration.</p></dd><dt><span class="term">
341
342    "<code class="mallctl">config.prof_libunwind</code>"
343
344          (<span class="type">bool</span>)
345          <code class="literal">r-</code>
346        </span></dt><dd><p><code class="option">--enable-prof-libunwind</code> was specified
347        during build configuration.</p></dd><dt><span class="term">
348
349    "<code class="mallctl">config.stats</code>"
350
351          (<span class="type">bool</span>)
352          <code class="literal">r-</code>
353        </span></dt><dd><p><code class="option">--enable-stats</code> was specified during
354        build configuration.</p></dd><dt><span class="term">
355
356    "<code class="mallctl">config.tcache</code>"
357
358          (<span class="type">bool</span>)
359          <code class="literal">r-</code>
360        </span></dt><dd><p><code class="option">--disable-tcache</code> was not specified
361        during build configuration.</p></dd><dt><span class="term">
362
363    "<code class="mallctl">config.tls</code>"
364
365          (<span class="type">bool</span>)
366          <code class="literal">r-</code>
367        </span></dt><dd><p><code class="option">--disable-tls</code> was not specified during
368        build configuration.</p></dd><dt><span class="term">
369
370    "<code class="mallctl">config.utrace</code>"
371
372          (<span class="type">bool</span>)
373          <code class="literal">r-</code>
374        </span></dt><dd><p><code class="option">--enable-utrace</code> was specified during
375        build configuration.</p></dd><dt><span class="term">
376
377    "<code class="mallctl">config.valgrind</code>"
378
379          (<span class="type">bool</span>)
380          <code class="literal">r-</code>
381        </span></dt><dd><p><code class="option">--enable-valgrind</code> was specified during
382        build configuration.</p></dd><dt><span class="term">
383
384    "<code class="mallctl">config.xmalloc</code>"
385
386          (<span class="type">bool</span>)
387          <code class="literal">r-</code>
388        </span></dt><dd><p><code class="option">--enable-xmalloc</code> was specified during
389        build configuration.</p></dd><dt><a name="opt.abort"></a><span class="term">
390
391    "<code class="mallctl">opt.abort</code>"
392
393          (<span class="type">bool</span>)
394          <code class="literal">r-</code>
395        </span></dt><dd><p>Abort-on-warning enabled/disabled.  If true, most
396        warnings are fatal.  The process will call
397        <span class="citerefentry"><span class="refentrytitle">abort</span>(3)</span> in these cases.  This option is
398        disabled by default unless <code class="option">--enable-debug</code> is
399        specified during configuration, in which case it is enabled by default.
400        </p></dd><dt><a name="opt.lg_chunk"></a><span class="term">
401
402    "<code class="mallctl">opt.lg_chunk</code>"
403
404          (<span class="type">size_t</span>)
405          <code class="literal">r-</code>
406        </span></dt><dd><p>Virtual memory chunk size (log base 2).  The default
407        chunk size is 4 MiB (2^22).</p></dd><dt><a name="opt.narenas"></a><span class="term">
408
409    "<code class="mallctl">opt.narenas</code>"
410
411          (<span class="type">size_t</span>)
412          <code class="literal">r-</code>
413        </span></dt><dd><p>Maximum number of arenas to use.  The default maximum
414        number of arenas is four times the number of CPUs, or one if there is a
415        single CPU.</p></dd><dt><a name="opt.lg_dirty_mult"></a><span class="term">
416
417    "<code class="mallctl">opt.lg_dirty_mult</code>"
418
419          (<span class="type">ssize_t</span>)
420          <code class="literal">r-</code>
421        </span></dt><dd><p>Per-arena minimum ratio (log base 2) of active to dirty
422        pages.  Some dirty unused pages may be allowed to accumulate, within
423        the limit set by the ratio (or one chunk worth of dirty pages,
424        whichever is greater), before informing the kernel about some of those
425        pages via <span class="citerefentry"><span class="refentrytitle">madvise</span>(2)</span> or a similar system call.  This
426        provides the kernel with sufficient information to recycle dirty pages
427        if physical memory becomes scarce and the pages remain unused.  The
428        default minimum ratio is 32:1 (2^5:1); an option value of -1 will
429        disable dirty page purging.</p></dd><dt><a name="opt.stats_print"></a><span class="term">
430
431    "<code class="mallctl">opt.stats_print</code>"
432
433          (<span class="type">bool</span>)
434          <code class="literal">r-</code>
435        </span></dt><dd><p>Enable/disable statistics printing at exit.  If
436        enabled, the <code class="function">malloc_stats_print</code>(<em class="parameter"><code></code></em>)
437        function is called at program exit via an
438        <span class="citerefentry"><span class="refentrytitle">atexit</span>(3)</span> function.  If
439        <code class="option">--enable-stats</code> is specified during configuration, this
440        has the potential to cause deadlock for a multi-threaded process that
441        exits while one or more threads are executing in the memory allocation
442        functions.  Therefore, this option should only be used with care; it is
443        primarily intended as a performance tuning aid during application
444        development.  This option is disabled by default.</p></dd><dt><a name="opt.junk"></a><span class="term">
445
446    "<code class="mallctl">opt.junk</code>"
447
448          (<span class="type">bool</span>)
449          <code class="literal">r-</code>
450          [<code class="option">--enable-fill</code>]
451        </span></dt><dd><p>Junk filling enabled/disabled.  If enabled, each byte
452        of uninitialized allocated memory will be initialized to
453        <code class="literal">0xa5</code>.  All deallocated memory will be initialized to
454        <code class="literal">0x5a</code>.  This is intended for debugging and will
455        impact performance negatively.  This option is disabled by default
456        unless <code class="option">--enable-debug</code> is specified during
457        configuration, in which case it is enabled by default.</p></dd><dt><a name="opt.quarantine"></a><span class="term">
458
459    "<code class="mallctl">opt.quarantine</code>"
460
461          (<span class="type">size_t</span>)
462          <code class="literal">r-</code>
463          [<code class="option">--enable-fill</code>]
464        </span></dt><dd><p>Per thread quarantine size in bytes.  If non-zero, each
465        thread maintains a FIFO object quarantine that stores up to the
466        specified number of bytes of memory.  The quarantined memory is not
467        freed until it is released from quarantine, though it is immediately
468        junk-filled if the <a class="link" href="#opt.junk">
469    "<code class="mallctl">opt.junk</code>"
470  </a> option is
471        enabled.  This feature is of particular use in combination with <a class="ulink" href="http://valgrind.org/" target="_top">Valgrind</a>, which can detect attempts
472        to access quarantined objects.  This is intended for debugging and will
473        impact performance negatively.  The default quarantine size is
474        0.</p></dd><dt><a name="opt.redzone"></a><span class="term">
475
476    "<code class="mallctl">opt.redzone</code>"
477
478          (<span class="type">bool</span>)
479          <code class="literal">r-</code>
480          [<code class="option">--enable-fill</code>]
481        </span></dt><dd><p>Redzones enabled/disabled.  If enabled, small
482        allocations have redzones before and after them.  Furthermore, if the
483        <a class="link" href="#opt.junk">
484    "<code class="mallctl">opt.junk</code>"
485  </a> option is
486        enabled, the redzones are checked for corruption during deallocation.
487        However, the primary intended purpose of this feature is to be used in
488        combination with <a class="ulink" href="http://valgrind.org/" target="_top">Valgrind</a>,
489        which needs redzones in order to do effective buffer overflow/underflow
490        detection.  This option is intended for debugging and will impact
491        performance negatively.  This option is disabled by
492        default.</p></dd><dt><a name="opt.zero"></a><span class="term">
493
494    "<code class="mallctl">opt.zero</code>"
495
496          (<span class="type">bool</span>)
497          <code class="literal">r-</code>
498          [<code class="option">--enable-fill</code>]
499        </span></dt><dd><p>Zero filling enabled/disabled.  If enabled, each byte
500        of uninitialized allocated memory will be initialized to 0.  Note that
501        this initialization only happens once for each byte, so
502        <code class="function">realloc</code>(<em class="parameter"><code></code></em>) and
503        <code class="function">rallocm</code>(<em class="parameter"><code></code></em>) calls do not zero memory that
504        was previously allocated.  This is intended for debugging and will
505        impact performance negatively.  This option is disabled by default.
506        </p></dd><dt><a name="opt.utrace"></a><span class="term">
507
508    "<code class="mallctl">opt.utrace</code>"
509
510          (<span class="type">bool</span>)
511          <code class="literal">r-</code>
512          [<code class="option">--enable-utrace</code>]
513        </span></dt><dd><p>Allocation tracing based on
514        <span class="citerefentry"><span class="refentrytitle">utrace</span>(2)</span> enabled/disabled.  This option
515        is disabled by default.</p></dd><dt><a name="opt.valgrind"></a><span class="term">
516
517    "<code class="mallctl">opt.valgrind</code>"
518
519          (<span class="type">bool</span>)
520          <code class="literal">r-</code>
521          [<code class="option">--enable-valgrind</code>]
522        </span></dt><dd><p><a class="ulink" href="http://valgrind.org/" target="_top">Valgrind</a>
523        support enabled/disabled.  If enabled, several other options are
524        automatically modified during options processing to work well with
525        Valgrind: <a class="link" href="#opt.junk">
526    "<code class="mallctl">opt.junk</code>"
527  </a>
528        and <a class="link" href="#opt.zero">
529    "<code class="mallctl">opt.zero</code>"
530  </a> are set
531        to false, <a class="link" href="#opt.quarantine">
532    "<code class="mallctl">opt.quarantine</code>"
533  </a> is
534        set to 16 MiB, and <a class="link" href="#opt.redzone">
535    "<code class="mallctl">opt.redzone</code>"
536  </a> is set to
537        true.  This option is disabled by default.</p></dd><dt><a name="opt.xmalloc"></a><span class="term">
538
539    "<code class="mallctl">opt.xmalloc</code>"
540
541          (<span class="type">bool</span>)
542          <code class="literal">r-</code>
543          [<code class="option">--enable-xmalloc</code>]
544        </span></dt><dd><p>Abort-on-out-of-memory enabled/disabled.  If enabled,
545        rather than returning failure for any allocation function, display a
546        diagnostic message on <code class="constant">STDERR_FILENO</code> and cause the
547        program to drop core (using
548        <span class="citerefentry"><span class="refentrytitle">abort</span>(3)</span>).  If an application is
549        designed to depend on this behavior, set the option at compile time by
550        including the following in the source code:
551        </p><pre class="programlisting">
552malloc_conf = "xmalloc:true";</pre><p>
553        This option is disabled by default.</p></dd><dt><a name="opt.tcache"></a><span class="term">
554
555    "<code class="mallctl">opt.tcache</code>"
556
557          (<span class="type">bool</span>)
558          <code class="literal">r-</code>
559          [<code class="option">--enable-tcache</code>]
560        </span></dt><dd><p>Thread-specific caching enabled/disabled.  When there
561        are multiple threads, each thread uses a thread-specific cache for
562        objects up to a certain size.  Thread-specific caching allows many
563        allocations to be satisfied without performing any thread
564        synchronization, at the cost of increased memory use.  See the
565        <a class="link" href="#opt.lg_tcache_max">
566    "<code class="mallctl">opt.lg_tcache_max</code>"
567  </a>
568        option for related tuning information.  This option is enabled by
569        default.</p></dd><dt><a name="opt.lg_tcache_max"></a><span class="term">
570
571    "<code class="mallctl">opt.lg_tcache_max</code>"
572
573          (<span class="type">size_t</span>)
574          <code class="literal">r-</code>
575          [<code class="option">--enable-tcache</code>]
576        </span></dt><dd><p>Maximum size class (log base 2) to cache in the
577        thread-specific cache.  At a minimum, all small size classes are
578        cached, and at a maximum all large size classes are cached.  The
579        default maximum is 32 KiB (2^15).</p></dd><dt><a name="opt.prof"></a><span class="term">
580
581    "<code class="mallctl">opt.prof</code>"
582
583          (<span class="type">bool</span>)
584          <code class="literal">r-</code>
585          [<code class="option">--enable-prof</code>]
586        </span></dt><dd><p>Memory profiling enabled/disabled.  If enabled, profile
587        memory allocation activity.  See the <a class="link" href="#opt.prof_active">
588    "<code class="mallctl">opt.prof_active</code>"
589  </a>
590        option for on-the-fly activation/deactivation.  See the <a class="link" href="#opt.lg_prof_sample">
591    "<code class="mallctl">opt.lg_prof_sample</code>"
592  </a>
593        option for probabilistic sampling control.  See the <a class="link" href="#opt.prof_accum">
594    "<code class="mallctl">opt.prof_accum</code>"
595  </a>
596        option for control of cumulative sample reporting.  See the <a class="link" href="#opt.lg_prof_interval">
597    "<code class="mallctl">opt.lg_prof_interval</code>"
598  </a>
599        option for information on interval-triggered profile dumping, the <a class="link" href="#opt.prof_gdump">
600    "<code class="mallctl">opt.prof_gdump</code>"
601  </a>
602        option for information on high-water-triggered profile dumping, and the
603        <a class="link" href="#opt.prof_final">
604    "<code class="mallctl">opt.prof_final</code>"
605  </a>
606        option for final profile dumping.  Profile output is compatible with
607        the included <span class="command"><strong>pprof</strong></span> Perl script, which originates
608        from the <a class="ulink" href="http://code.google.com/p/gperftools/" target="_top">gperftools
609        package</a>.</p></dd><dt><a name="opt.prof_prefix"></a><span class="term">
610
611    "<code class="mallctl">opt.prof_prefix</code>"
612
613          (<span class="type">const char *</span>)
614          <code class="literal">r-</code>
615          [<code class="option">--enable-prof</code>]
616        </span></dt><dd><p>Filename prefix for profile dumps.  If the prefix is
617        set to the empty string, no automatic dumps will occur; this is
618        primarily useful for disabling the automatic final heap dump (which
619        also disables leak reporting, if enabled).  The default prefix is
620        <code class="filename">jeprof</code>.</p></dd><dt><a name="opt.prof_active"></a><span class="term">
621
622    "<code class="mallctl">opt.prof_active</code>"
623
624          (<span class="type">bool</span>)
625          <code class="literal">r-</code>
626          [<code class="option">--enable-prof</code>]
627        </span></dt><dd><p>Profiling activated/deactivated.  This is a secondary
628        control mechanism that makes it possible to start the application with
629        profiling enabled (see the <a class="link" href="#opt.prof">
630    "<code class="mallctl">opt.prof</code>"
631  </a> option) but
632        inactive, then toggle profiling at any time during program execution
633        with the <a class="link" href="#prof.active">
634    "<code class="mallctl">prof.active</code>"
635  </a> mallctl.
636        This option is enabled by default.</p></dd><dt><a name="opt.lg_prof_sample"></a><span class="term">
637
638    "<code class="mallctl">opt.lg_prof_sample</code>"
639
640          (<span class="type">ssize_t</span>)
641          <code class="literal">r-</code>
642          [<code class="option">--enable-prof</code>]
643        </span></dt><dd><p>Average interval (log base 2) between allocation
644        samples, as measured in bytes of allocation activity.  Increasing the
645        sampling interval decreases profile fidelity, but also decreases the
646        computational overhead.  The default sample interval is 512 KiB (2^19
647        B).</p></dd><dt><a name="opt.prof_accum"></a><span class="term">
648
649    "<code class="mallctl">opt.prof_accum</code>"
650
651          (<span class="type">bool</span>)
652          <code class="literal">r-</code>
653          [<code class="option">--enable-prof</code>]
654        </span></dt><dd><p>Reporting of cumulative object/byte counts in profile
655        dumps enabled/disabled.  If this option is enabled, every unique
656        backtrace must be stored for the duration of execution.  Depending on
657        the application, this can impose a large memory overhead, and the
658        cumulative counts are not always of interest.  This option is disabled
659        by default.</p></dd><dt><a name="opt.lg_prof_interval"></a><span class="term">
660
661    "<code class="mallctl">opt.lg_prof_interval</code>"
662
663          (<span class="type">ssize_t</span>)
664          <code class="literal">r-</code>
665          [<code class="option">--enable-prof</code>]
666        </span></dt><dd><p>Average interval (log base 2) between memory profile
667        dumps, as measured in bytes of allocation activity.  The actual
668        interval between dumps may be sporadic because decentralized allocation
669        counters are used to avoid synchronization bottlenecks.  Profiles are
670        dumped to files named according to the pattern
671        <code class="filename">&lt;prefix&gt;.&lt;pid&gt;.&lt;seq&gt;.i&lt;iseq&gt;.heap</code>,
672        where <code class="literal">&lt;prefix&gt;</code> is controlled by the
673        <a class="link" href="#opt.prof_prefix">
674    "<code class="mallctl">opt.prof_prefix</code>"
675  </a>
676        option.  By default, interval-triggered profile dumping is disabled
677        (encoded as -1).
678        </p></dd><dt><a name="opt.prof_gdump"></a><span class="term">
679
680    "<code class="mallctl">opt.prof_gdump</code>"
681
682          (<span class="type">bool</span>)
683          <code class="literal">r-</code>
684          [<code class="option">--enable-prof</code>]
685        </span></dt><dd><p>Trigger a memory profile dump every time the total
686        virtual memory exceeds the previous maximum.  Profiles are dumped to
687        files named according to the pattern
688        <code class="filename">&lt;prefix&gt;.&lt;pid&gt;.&lt;seq&gt;.u&lt;useq&gt;.heap</code>,
689        where <code class="literal">&lt;prefix&gt;</code> is controlled by the <a class="link" href="#opt.prof_prefix">
690    "<code class="mallctl">opt.prof_prefix</code>"
691  </a>
692        option.  This option is disabled by default.</p></dd><dt><a name="opt.prof_final"></a><span class="term">
693
694    "<code class="mallctl">opt.prof_final</code>"
695
696          (<span class="type">bool</span>)
697          <code class="literal">r-</code>
698          [<code class="option">--enable-prof</code>]
699        </span></dt><dd><p>Use an
700        <span class="citerefentry"><span class="refentrytitle">atexit</span>(3)</span> function to dump final memory
701        usage to a file named according to the pattern
702        <code class="filename">&lt;prefix&gt;.&lt;pid&gt;.&lt;seq&gt;.f.heap</code>,
703        where <code class="literal">&lt;prefix&gt;</code> is controlled by the <a class="link" href="#opt.prof_prefix">
704    "<code class="mallctl">opt.prof_prefix</code>"
705  </a>
706        option.  This option is enabled by default.</p></dd><dt><a name="opt.prof_leak"></a><span class="term">
707
708    "<code class="mallctl">opt.prof_leak</code>"
709
710          (<span class="type">bool</span>)
711          <code class="literal">r-</code>
712          [<code class="option">--enable-prof</code>]
713        </span></dt><dd><p>Leak reporting enabled/disabled.  If enabled, use an
714        <span class="citerefentry"><span class="refentrytitle">atexit</span>(3)</span> function to report memory leaks
715        detected by allocation sampling.  See the
716        <a class="link" href="#opt.prof">
717    "<code class="mallctl">opt.prof</code>"
718  </a> option for
719        information on analyzing heap profile output.  This option is disabled
720        by default.</p></dd><dt><span class="term">
721
722    "<code class="mallctl">thread.arena</code>"
723
724          (<span class="type">unsigned</span>)
725          <code class="literal">rw</code>
726        </span></dt><dd><p>Get or set the arena associated with the calling
727        thread.  The arena index must be less than the maximum number of arenas
728        (see the <a class="link" href="#arenas.narenas">
729    "<code class="mallctl">arenas.narenas</code>"
730  </a>
731        mallctl).  If the specified arena was not initialized beforehand (see
732        the <a class="link" href="#arenas.initialized">
733    "<code class="mallctl">arenas.initialized</code>"
734  </a>
735        mallctl), it will be automatically initialized as a side effect of
736        calling this interface.</p></dd><dt><a name="thread.allocated"></a><span class="term">
737
738    "<code class="mallctl">thread.allocated</code>"
739
740          (<span class="type">uint64_t</span>)
741          <code class="literal">r-</code>
742          [<code class="option">--enable-stats</code>]
743        </span></dt><dd><p>Get the total number of bytes ever allocated by the
744        calling thread.  This counter has the potential to wrap around; it is
745        up to the application to appropriately interpret the counter in such
746        cases.</p></dd><dt><span class="term">
747
748    "<code class="mallctl">thread.allocatedp</code>"
749
750          (<span class="type">uint64_t *</span>)
751          <code class="literal">r-</code>
752          [<code class="option">--enable-stats</code>]
753        </span></dt><dd><p>Get a pointer to the the value that is returned by the
754        <a class="link" href="#thread.allocated">
755    "<code class="mallctl">thread.allocated</code>"
756  </a>
757        mallctl.  This is useful for avoiding the overhead of repeated
758        <code class="function">mallctl*</code>(<em class="parameter"><code></code></em>) calls.</p></dd><dt><a name="thread.deallocated"></a><span class="term">
759
760    "<code class="mallctl">thread.deallocated</code>"
761
762          (<span class="type">uint64_t</span>)
763          <code class="literal">r-</code>
764          [<code class="option">--enable-stats</code>]
765        </span></dt><dd><p>Get the total number of bytes ever deallocated by the
766        calling thread.  This counter has the potential to wrap around; it is
767        up to the application to appropriately interpret the counter in such
768        cases.</p></dd><dt><span class="term">
769
770    "<code class="mallctl">thread.deallocatedp</code>"
771
772          (<span class="type">uint64_t *</span>)
773          <code class="literal">r-</code>
774          [<code class="option">--enable-stats</code>]
775        </span></dt><dd><p>Get a pointer to the the value that is returned by the
776        <a class="link" href="#thread.deallocated">
777    "<code class="mallctl">thread.deallocated</code>"
778  </a>
779        mallctl.  This is useful for avoiding the overhead of repeated
780        <code class="function">mallctl*</code>(<em class="parameter"><code></code></em>) calls.</p></dd><dt><span class="term">
781
782    "<code class="mallctl">thread.tcache.enabled</code>"
783
784          (<span class="type">bool</span>)
785          <code class="literal">rw</code>
786          [<code class="option">--enable-tcache</code>]
787        </span></dt><dd><p>Enable/disable calling thread's tcache.  The tcache is
788        implicitly flushed as a side effect of becoming
789        disabled (see
790    "<code class="mallctl">thread.tcache.flush</code>"
791  ).
792        </p></dd><dt><span class="term">
793
794    "<code class="mallctl">thread.tcache.flush</code>"
795
796          (<span class="type">void</span>)
797          <code class="literal">--</code>
798          [<code class="option">--enable-tcache</code>]
799        </span></dt><dd><p>Flush calling thread's tcache.  This interface releases
800        all cached objects and internal data structures associated with the
801        calling thread's thread-specific cache.  Ordinarily, this interface
802        need not be called, since automatic periodic incremental garbage
803        collection occurs, and the thread cache is automatically discarded when
804        a thread exits.  However, garbage collection is triggered by allocation
805        activity, so it is possible for a thread that stops
806        allocating/deallocating to retain its cache indefinitely, in which case
807        the developer may find manual flushing useful.</p></dd><dt><a name="arenas.narenas"></a><span class="term">
808
809    "<code class="mallctl">arenas.narenas</code>"
810
811          (<span class="type">unsigned</span>)
812          <code class="literal">r-</code>
813        </span></dt><dd><p>Maximum number of arenas.</p></dd><dt><a name="arenas.initialized"></a><span class="term">
814
815    "<code class="mallctl">arenas.initialized</code>"
816
817          (<span class="type">bool *</span>)
818          <code class="literal">r-</code>
819        </span></dt><dd><p>An array of <a class="link" href="#arenas.narenas">
820    "<code class="mallctl">arenas.narenas</code>"
821  </a>
822        booleans.  Each boolean indicates whether the corresponding arena is
823        initialized.</p></dd><dt><span class="term">
824
825    "<code class="mallctl">arenas.quantum</code>"
826
827          (<span class="type">size_t</span>)
828          <code class="literal">r-</code>
829        </span></dt><dd><p>Quantum size.</p></dd><dt><span class="term">
830
831    "<code class="mallctl">arenas.page</code>"
832
833          (<span class="type">size_t</span>)
834          <code class="literal">r-</code>
835        </span></dt><dd><p>Page size.</p></dd><dt><span class="term">
836
837    "<code class="mallctl">arenas.tcache_max</code>"
838
839          (<span class="type">size_t</span>)
840          <code class="literal">r-</code>
841          [<code class="option">--enable-tcache</code>]
842        </span></dt><dd><p>Maximum thread-cached size class.</p></dd><dt><span class="term">
843
844    "<code class="mallctl">arenas.nbins</code>"
845
846          (<span class="type">unsigned</span>)
847          <code class="literal">r-</code>
848        </span></dt><dd><p>Number of bin size classes.</p></dd><dt><span class="term">
849
850    "<code class="mallctl">arenas.nhbins</code>"
851
852          (<span class="type">unsigned</span>)
853          <code class="literal">r-</code>
854          [<code class="option">--enable-tcache</code>]
855        </span></dt><dd><p>Total number of thread cache bin size
856        classes.</p></dd><dt><a name="arenas.bin.i.size"></a><span class="term">
857
858    "<code class="mallctl">arenas.bin.&lt;i&gt;.size</code>"
859
860          (<span class="type">size_t</span>)
861          <code class="literal">r-</code>
862        </span></dt><dd><p>Maximum size supported by size class.</p></dd><dt><span class="term">
863
864    "<code class="mallctl">arenas.bin.&lt;i&gt;.nregs</code>"
865
866          (<span class="type">uint32_t</span>)
867          <code class="literal">r-</code>
868        </span></dt><dd><p>Number of regions per page run.</p></dd><dt><span class="term">
869
870    "<code class="mallctl">arenas.bin.&lt;i&gt;.run_size</code>"
871
872          (<span class="type">size_t</span>)
873          <code class="literal">r-</code>
874        </span></dt><dd><p>Number of bytes per page run.</p></dd><dt><span class="term">
875
876    "<code class="mallctl">arenas.nlruns</code>"
877
878          (<span class="type">size_t</span>)
879          <code class="literal">r-</code>
880        </span></dt><dd><p>Total number of large size classes.</p></dd><dt><span class="term">
881
882    "<code class="mallctl">arenas.lrun.&lt;i&gt;.size</code>"
883
884          (<span class="type">size_t</span>)
885          <code class="literal">r-</code>
886        </span></dt><dd><p>Maximum size supported by this large size
887        class.</p></dd><dt><span class="term">
888
889    "<code class="mallctl">arenas.purge</code>"
890
891          (<span class="type">unsigned</span>)
892          <code class="literal">-w</code>
893        </span></dt><dd><p>Purge unused dirty pages for the specified arena, or
894        for all arenas if none is specified.</p></dd><dt><a name="prof.active"></a><span class="term">
895
896    "<code class="mallctl">prof.active</code>"
897
898          (<span class="type">bool</span>)
899          <code class="literal">rw</code>
900          [<code class="option">--enable-prof</code>]
901        </span></dt><dd><p>Control whether sampling is currently active.  See the
902        <a class="link" href="#opt.prof_active">
903    "<code class="mallctl">opt.prof_active</code>"
904  </a>
905        option for additional information.
906        </p></dd><dt><span class="term">
907
908    "<code class="mallctl">prof.dump</code>"
909
910          (<span class="type">const char *</span>)
911          <code class="literal">-w</code>
912          [<code class="option">--enable-prof</code>]
913        </span></dt><dd><p>Dump a memory profile to the specified file, or if NULL
914        is specified, to a file according to the pattern
915        <code class="filename">&lt;prefix&gt;.&lt;pid&gt;.&lt;seq&gt;.m&lt;mseq&gt;.heap</code>,
916        where <code class="literal">&lt;prefix&gt;</code> is controlled by the
917        <a class="link" href="#opt.prof_prefix">
918    "<code class="mallctl">opt.prof_prefix</code>"
919  </a>
920        option.</p></dd><dt><span class="term">
921
922    "<code class="mallctl">prof.interval</code>"
923
924          (<span class="type">uint64_t</span>)
925          <code class="literal">r-</code>
926          [<code class="option">--enable-prof</code>]
927        </span></dt><dd><p>Average number of bytes allocated between
928        inverval-based profile dumps.  See the
929        <a class="link" href="#opt.lg_prof_interval">
930    "<code class="mallctl">opt.lg_prof_interval</code>"
931  </a>
932        option for additional information.</p></dd><dt><a name="stats.cactive"></a><span class="term">
933
934    "<code class="mallctl">stats.cactive</code>"
935
936          (<span class="type">size_t *</span>)
937          <code class="literal">r-</code>
938          [<code class="option">--enable-stats</code>]
939        </span></dt><dd><p>Pointer to a counter that contains an approximate count
940        of the current number of bytes in active pages.  The estimate may be
941        high, but never low, because each arena rounds up to the nearest
942        multiple of the chunk size when computing its contribution to the
943        counter.  Note that the <a class="link" href="#epoch">
944    "<code class="mallctl">epoch</code>"
945  </a> mallctl has no bearing
946        on this counter.  Furthermore, counter consistency is maintained via
947        atomic operations, so it is necessary to use an atomic operation in
948        order to guarantee a consistent read when dereferencing the pointer.
949        </p></dd><dt><a name="stats.allocated"></a><span class="term">
950
951    "<code class="mallctl">stats.allocated</code>"
952
953          (<span class="type">size_t</span>)
954          <code class="literal">r-</code>
955          [<code class="option">--enable-stats</code>]
956        </span></dt><dd><p>Total number of bytes allocated by the
957        application.</p></dd><dt><a name="stats.active"></a><span class="term">
958
959    "<code class="mallctl">stats.active</code>"
960
961          (<span class="type">size_t</span>)
962          <code class="literal">r-</code>
963          [<code class="option">--enable-stats</code>]
964        </span></dt><dd><p>Total number of bytes in active pages allocated by the
965        application.  This is a multiple of the page size, and greater than or
966        equal to <a class="link" href="#stats.allocated">
967    "<code class="mallctl">stats.allocated</code>"
968  </a>.
969        </p></dd><dt><span class="term">
970
971    "<code class="mallctl">stats.mapped</code>"
972
973          (<span class="type">size_t</span>)
974          <code class="literal">r-</code>
975          [<code class="option">--enable-stats</code>]
976        </span></dt><dd><p>Total number of bytes in chunks mapped on behalf of the
977        application.  This is a multiple of the chunk size, and is at least as
978        large as <a class="link" href="#stats.active">
979    "<code class="mallctl">stats.active</code>"
980  </a>.  This
981        does not include inactive chunks.</p></dd><dt><span class="term">
982
983    "<code class="mallctl">stats.chunks.current</code>"
984
985          (<span class="type">size_t</span>)
986          <code class="literal">r-</code>
987          [<code class="option">--enable-stats</code>]
988        </span></dt><dd><p>Total number of chunks actively mapped on behalf of the
989        application.  This does not include inactive chunks.
990        </p></dd><dt><span class="term">
991
992    "<code class="mallctl">stats.chunks.total</code>"
993
994          (<span class="type">uint64_t</span>)
995          <code class="literal">r-</code>
996          [<code class="option">--enable-stats</code>]
997        </span></dt><dd><p>Cumulative number of chunks allocated.</p></dd><dt><span class="term">
998
999    "<code class="mallctl">stats.chunks.high</code>"
1000
1001          (<span class="type">size_t</span>)
1002          <code class="literal">r-</code>
1003          [<code class="option">--enable-stats</code>]
1004        </span></dt><dd><p>Maximum number of active chunks at any time thus far.
1005        </p></dd><dt><span class="term">
1006
1007    "<code class="mallctl">stats.huge.allocated</code>"
1008
1009          (<span class="type">size_t</span>)
1010          <code class="literal">r-</code>
1011          [<code class="option">--enable-stats</code>]
1012        </span></dt><dd><p>Number of bytes currently allocated by huge objects.
1013        </p></dd><dt><span class="term">
1014
1015    "<code class="mallctl">stats.huge.nmalloc</code>"
1016
1017          (<span class="type">uint64_t</span>)
1018          <code class="literal">r-</code>
1019          [<code class="option">--enable-stats</code>]
1020        </span></dt><dd><p>Cumulative number of huge allocation requests.
1021        </p></dd><dt><span class="term">
1022
1023    "<code class="mallctl">stats.huge.ndalloc</code>"
1024
1025          (<span class="type">uint64_t</span>)
1026          <code class="literal">r-</code>
1027          [<code class="option">--enable-stats</code>]
1028        </span></dt><dd><p>Cumulative number of huge deallocation requests.
1029        </p></dd><dt><span class="term">
1030
1031    "<code class="mallctl">stats.arenas.&lt;i&gt;.nthreads</code>"
1032
1033          (<span class="type">unsigned</span>)
1034          <code class="literal">r-</code>
1035        </span></dt><dd><p>Number of threads currently assigned to
1036        arena.</p></dd><dt><span class="term">
1037
1038    "<code class="mallctl">stats.arenas.&lt;i&gt;.pactive</code>"
1039
1040          (<span class="type">size_t</span>)
1041          <code class="literal">r-</code>
1042        </span></dt><dd><p>Number of pages in active runs.</p></dd><dt><span class="term">
1043
1044    "<code class="mallctl">stats.arenas.&lt;i&gt;.pdirty</code>"
1045
1046          (<span class="type">size_t</span>)
1047          <code class="literal">r-</code>
1048        </span></dt><dd><p>Number of pages within unused runs that are potentially
1049        dirty, and for which <code class="function">madvise</code>(<em class="parameter"><code>...</code></em>,
1050        <em class="parameter"><code><code class="constant">MADV_DONTNEED</code></code></em>) or
1051        similar has not been called.</p></dd><dt><span class="term">
1052
1053    "<code class="mallctl">stats.arenas.&lt;i&gt;.mapped</code>"
1054
1055          (<span class="type">size_t</span>)
1056          <code class="literal">r-</code>
1057          [<code class="option">--enable-stats</code>]
1058        </span></dt><dd><p>Number of mapped bytes.</p></dd><dt><span class="term">
1059
1060    "<code class="mallctl">stats.arenas.&lt;i&gt;.npurge</code>"
1061
1062          (<span class="type">uint64_t</span>)
1063          <code class="literal">r-</code>
1064          [<code class="option">--enable-stats</code>]
1065        </span></dt><dd><p>Number of dirty page purge sweeps performed.
1066        </p></dd><dt><span class="term">
1067
1068    "<code class="mallctl">stats.arenas.&lt;i&gt;.nmadvise</code>"
1069
1070          (<span class="type">uint64_t</span>)
1071          <code class="literal">r-</code>
1072          [<code class="option">--enable-stats</code>]
1073        </span></dt><dd><p>Number of <code class="function">madvise</code>(<em class="parameter"><code>...</code></em>,
1074        <em class="parameter"><code><code class="constant">MADV_DONTNEED</code></code></em>) or
1075        similar calls made to purge dirty pages.</p></dd><dt><span class="term">
1076
1077    "<code class="mallctl">stats.arenas.&lt;i&gt;.npurged</code>"
1078
1079          (<span class="type">uint64_t</span>)
1080          <code class="literal">r-</code>
1081          [<code class="option">--enable-stats</code>]
1082        </span></dt><dd><p>Number of pages purged.</p></dd><dt><span class="term">
1083
1084    "<code class="mallctl">stats.arenas.&lt;i&gt;.small.allocated</code>"
1085
1086          (<span class="type">size_t</span>)
1087          <code class="literal">r-</code>
1088          [<code class="option">--enable-stats</code>]
1089        </span></dt><dd><p>Number of bytes currently allocated by small objects.
1090        </p></dd><dt><span class="term">
1091
1092    "<code class="mallctl">stats.arenas.&lt;i&gt;.small.nmalloc</code>"
1093
1094          (<span class="type">uint64_t</span>)
1095          <code class="literal">r-</code>
1096          [<code class="option">--enable-stats</code>]
1097        </span></dt><dd><p>Cumulative number of allocation requests served by
1098        small bins.</p></dd><dt><span class="term">
1099
1100    "<code class="mallctl">stats.arenas.&lt;i&gt;.small.ndalloc</code>"
1101
1102          (<span class="type">uint64_t</span>)
1103          <code class="literal">r-</code>
1104          [<code class="option">--enable-stats</code>]
1105        </span></dt><dd><p>Cumulative number of small objects returned to bins.
1106        </p></dd><dt><span class="term">
1107
1108    "<code class="mallctl">stats.arenas.&lt;i&gt;.small.nrequests</code>"
1109
1110          (<span class="type">uint64_t</span>)
1111          <code class="literal">r-</code>
1112          [<code class="option">--enable-stats</code>]
1113        </span></dt><dd><p>Cumulative number of small allocation requests.
1114        </p></dd><dt><span class="term">
1115
1116    "<code class="mallctl">stats.arenas.&lt;i&gt;.large.allocated</code>"
1117
1118          (<span class="type">size_t</span>)
1119          <code class="literal">r-</code>
1120          [<code class="option">--enable-stats</code>]
1121        </span></dt><dd><p>Number of bytes currently allocated by large objects.
1122        </p></dd><dt><span class="term">
1123
1124    "<code class="mallctl">stats.arenas.&lt;i&gt;.large.nmalloc</code>"
1125
1126          (<span class="type">uint64_t</span>)
1127          <code class="literal">r-</code>
1128          [<code class="option">--enable-stats</code>]
1129        </span></dt><dd><p>Cumulative number of large allocation requests served
1130        directly by the arena.</p></dd><dt><span class="term">
1131
1132    "<code class="mallctl">stats.arenas.&lt;i&gt;.large.ndalloc</code>"
1133
1134          (<span class="type">uint64_t</span>)
1135          <code class="literal">r-</code>
1136          [<code class="option">--enable-stats</code>]
1137        </span></dt><dd><p>Cumulative number of large deallocation requests served
1138        directly by the arena.</p></dd><dt><span class="term">
1139
1140    "<code class="mallctl">stats.arenas.&lt;i&gt;.large.nrequests</code>"
1141
1142          (<span class="type">uint64_t</span>)
1143          <code class="literal">r-</code>
1144          [<code class="option">--enable-stats</code>]
1145        </span></dt><dd><p>Cumulative number of large allocation requests.
1146        </p></dd><dt><span class="term">
1147
1148    "<code class="mallctl">stats.arenas.&lt;i&gt;.bins.&lt;j&gt;.allocated</code>"
1149
1150          (<span class="type">size_t</span>)
1151          <code class="literal">r-</code>
1152          [<code class="option">--enable-stats</code>]
1153        </span></dt><dd><p>Current number of bytes allocated by
1154        bin.</p></dd><dt><span class="term">
1155
1156    "<code class="mallctl">stats.arenas.&lt;i&gt;.bins.&lt;j&gt;.nmalloc</code>"
1157
1158          (<span class="type">uint64_t</span>)
1159          <code class="literal">r-</code>
1160          [<code class="option">--enable-stats</code>]
1161        </span></dt><dd><p>Cumulative number of allocations served by bin.
1162        </p></dd><dt><span class="term">
1163
1164    "<code class="mallctl">stats.arenas.&lt;i&gt;.bins.&lt;j&gt;.ndalloc</code>"
1165
1166          (<span class="type">uint64_t</span>)
1167          <code class="literal">r-</code>
1168          [<code class="option">--enable-stats</code>]
1169        </span></dt><dd><p>Cumulative number of allocations returned to bin.
1170        </p></dd><dt><span class="term">
1171
1172    "<code class="mallctl">stats.arenas.&lt;i&gt;.bins.&lt;j&gt;.nrequests</code>"
1173
1174          (<span class="type">uint64_t</span>)
1175          <code class="literal">r-</code>
1176          [<code class="option">--enable-stats</code>]
1177        </span></dt><dd><p>Cumulative number of allocation
1178        requests.</p></dd><dt><span class="term">
1179
1180    "<code class="mallctl">stats.arenas.&lt;i&gt;.bins.&lt;j&gt;.nfills</code>"
1181
1182          (<span class="type">uint64_t</span>)
1183          <code class="literal">r-</code>
1184          [<code class="option">--enable-stats</code> <code class="option">--enable-tcache</code>]
1185        </span></dt><dd><p>Cumulative number of tcache fills.</p></dd><dt><span class="term">
1186
1187    "<code class="mallctl">stats.arenas.&lt;i&gt;.bins.&lt;j&gt;.nflushes</code>"
1188
1189          (<span class="type">uint64_t</span>)
1190          <code class="literal">r-</code>
1191          [<code class="option">--enable-stats</code> <code class="option">--enable-tcache</code>]
1192        </span></dt><dd><p>Cumulative number of tcache flushes.</p></dd><dt><span class="term">
1193
1194    "<code class="mallctl">stats.arenas.&lt;i&gt;.bins.&lt;j&gt;.nruns</code>"
1195
1196          (<span class="type">uint64_t</span>)
1197          <code class="literal">r-</code>
1198          [<code class="option">--enable-stats</code>]
1199        </span></dt><dd><p>Cumulative number of runs created.</p></dd><dt><span class="term">
1200
1201    "<code class="mallctl">stats.arenas.&lt;i&gt;.bins.&lt;j&gt;.nreruns</code>"
1202
1203          (<span class="type">uint64_t</span>)
1204          <code class="literal">r-</code>
1205          [<code class="option">--enable-stats</code>]
1206        </span></dt><dd><p>Cumulative number of times the current run from which
1207        to allocate changed.</p></dd><dt><span class="term">
1208
1209    "<code class="mallctl">stats.arenas.&lt;i&gt;.bins.&lt;j&gt;.curruns</code>"
1210
1211          (<span class="type">size_t</span>)
1212          <code class="literal">r-</code>
1213          [<code class="option">--enable-stats</code>]
1214        </span></dt><dd><p>Current number of runs.</p></dd><dt><span class="term">
1215
1216    "<code class="mallctl">stats.arenas.&lt;i&gt;.lruns.&lt;j&gt;.nmalloc</code>"
1217
1218          (<span class="type">uint64_t</span>)
1219          <code class="literal">r-</code>
1220          [<code class="option">--enable-stats</code>]
1221        </span></dt><dd><p>Cumulative number of allocation requests for this size
1222        class served directly by the arena.</p></dd><dt><span class="term">
1223
1224    "<code class="mallctl">stats.arenas.&lt;i&gt;.lruns.&lt;j&gt;.ndalloc</code>"
1225
1226          (<span class="type">uint64_t</span>)
1227          <code class="literal">r-</code>
1228          [<code class="option">--enable-stats</code>]
1229        </span></dt><dd><p>Cumulative number of deallocation requests for this
1230        size class served directly by the arena.</p></dd><dt><span class="term">
1231
1232    "<code class="mallctl">stats.arenas.&lt;i&gt;.lruns.&lt;j&gt;.nrequests</code>"
1233
1234          (<span class="type">uint64_t</span>)
1235          <code class="literal">r-</code>
1236          [<code class="option">--enable-stats</code>]
1237        </span></dt><dd><p>Cumulative number of allocation requests for this size
1238        class.</p></dd><dt><span class="term">
1239
1240    "<code class="mallctl">stats.arenas.&lt;i&gt;.lruns.&lt;j&gt;.curruns</code>"
1241
1242          (<span class="type">size_t</span>)
1243          <code class="literal">r-</code>
1244          [<code class="option">--enable-stats</code>]
1245        </span></dt><dd><p>Current number of runs for this size class.
1246        </p></dd></dl></div></div><div class="refsect1" title="DEBUGGING MALLOC PROBLEMS"><a name="debugging_malloc_problems"></a><h2>DEBUGGING MALLOC PROBLEMS</h2><p>When debugging, it is a good idea to configure/build jemalloc with
1247    the <code class="option">--enable-debug</code> and <code class="option">--enable-fill</code>
1248    options, and recompile the program with suitable options and symbols for
1249    debugger support.  When so configured, jemalloc incorporates a wide variety
1250    of run-time assertions that catch application errors such as double-free,
1251    write-after-free, etc.</p><p>Programs often accidentally depend on &#8220;uninitialized&#8221;
1252    memory actually being filled with zero bytes.  Junk filling
1253    (see the <a class="link" href="#opt.junk">
1254    "<code class="mallctl">opt.junk</code>"
1255  </a>
1256    option) tends to expose such bugs in the form of obviously incorrect
1257    results and/or coredumps.  Conversely, zero
1258    filling (see the <a class="link" href="#opt.zero">
1259    "<code class="mallctl">opt.zero</code>"
1260  </a> option) eliminates
1261    the symptoms of such bugs.  Between these two options, it is usually
1262    possible to quickly detect, diagnose, and eliminate such bugs.</p><p>This implementation does not provide much detail about the problems
1263    it detects, because the performance impact for storing such information
1264    would be prohibitive.  However, jemalloc does integrate with the most
1265    excellent <a class="ulink" href="http://valgrind.org/" target="_top">Valgrind</a> tool if the
1266    <code class="option">--enable-valgrind</code> configuration option is enabled and the
1267    <a class="link" href="#opt.valgrind">
1268    "<code class="mallctl">opt.valgrind</code>"
1269  </a> option
1270    is enabled.</p></div><div class="refsect1" title="DIAGNOSTIC MESSAGES"><a name="diagnostic_messages"></a><h2>DIAGNOSTIC MESSAGES</h2><p>If any of the memory allocation/deallocation functions detect an
1271    error or warning condition, a message will be printed to file descriptor
1272    <code class="constant">STDERR_FILENO</code>.  Errors will result in the process
1273    dumping core.  If the <a class="link" href="#opt.abort">
1274    "<code class="mallctl">opt.abort</code>"
1275  </a> option is set, most
1276    warnings are treated as errors.</p><p>The <code class="varname">malloc_message</code> variable allows the programmer
1277    to override the function which emits the text strings forming the errors
1278    and warnings if for some reason the <code class="constant">STDERR_FILENO</code> file
1279    descriptor is not suitable for this.
1280    <code class="function">malloc_message</code>(<em class="parameter"><code></code></em>) takes the
1281    <em class="parameter"><code>cbopaque</code></em> pointer argument that is
1282    <code class="constant">NULL</code> unless overridden by the arguments in a call to
1283    <code class="function">malloc_stats_print</code>(<em class="parameter"><code></code></em>), followed by a string
1284    pointer.  Please note that doing anything which tries to allocate memory in
1285    this function is likely to result in a crash or deadlock.</p><p>All messages are prefixed by
1286    &#8220;<code class="computeroutput">&lt;jemalloc&gt;: </code>&#8221;.</p></div><div class="refsect1" title="RETURN VALUES"><a name="return_values"></a><h2>RETURN VALUES</h2><div class="refsect2" title="Standard API"><a name="id286955289"></a><h3>Standard API</h3><p>The <code class="function">malloc</code>(<em class="parameter"><code></code></em>) and
1287      <code class="function">calloc</code>(<em class="parameter"><code></code></em>) functions return a pointer to the
1288      allocated memory if successful; otherwise a <code class="constant">NULL</code>
1289      pointer is returned and <code class="varname">errno</code> is set to
1290      <span class="errorname">ENOMEM</span>.</p><p>The <code class="function">posix_memalign</code>(<em class="parameter"><code></code></em>) function
1291      returns the value 0 if successful; otherwise it returns an error value.
1292      The <code class="function">posix_memalign</code>(<em class="parameter"><code></code></em>) function will fail
1293      if:
1294        </p><div class="variablelist"><dl><dt><span class="term"><span class="errorname">EINVAL</span></span></dt><dd><p>The <em class="parameter"><code>alignment</code></em> parameter is
1295            not a power of 2 at least as large as
1296            <code class="code">sizeof(<span class="type">void *</span>)</code>.
1297            </p></dd><dt><span class="term"><span class="errorname">ENOMEM</span></span></dt><dd><p>Memory allocation error.</p></dd></dl></div><p>
1298      </p><p>The <code class="function">aligned_alloc</code>(<em class="parameter"><code></code></em>) function returns
1299      a pointer to the allocated memory if successful; otherwise a
1300      <code class="constant">NULL</code> pointer is returned and
1301      <code class="varname">errno</code> is set.  The
1302      <code class="function">aligned_alloc</code>(<em class="parameter"><code></code></em>) function will fail if:
1303        </p><div class="variablelist"><dl><dt><span class="term"><span class="errorname">EINVAL</span></span></dt><dd><p>The <em class="parameter"><code>alignment</code></em> parameter is
1304            not a power of 2.
1305            </p></dd><dt><span class="term"><span class="errorname">ENOMEM</span></span></dt><dd><p>Memory allocation error.</p></dd></dl></div><p>
1306      </p><p>The <code class="function">realloc</code>(<em class="parameter"><code></code></em>) function returns a
1307      pointer, possibly identical to <em class="parameter"><code>ptr</code></em>, to the
1308      allocated memory if successful; otherwise a <code class="constant">NULL</code>
1309      pointer is returned, and <code class="varname">errno</code> is set to
1310      <span class="errorname">ENOMEM</span> if the error was the result of an
1311      allocation failure.  The <code class="function">realloc</code>(<em class="parameter"><code></code></em>)
1312      function always leaves the original buffer intact when an error occurs.
1313      </p><p>The <code class="function">free</code>(<em class="parameter"><code></code></em>) function returns no
1314      value.</p></div><div class="refsect2" title="Non-standard API"><a name="id286955505"></a><h3>Non-standard API</h3><p>The <code class="function">malloc_usable_size</code>(<em class="parameter"><code></code></em>) function
1315      returns the usable size of the allocation pointed to by
1316      <em class="parameter"><code>ptr</code></em>.  </p><p>The <code class="function">mallctl</code>(<em class="parameter"><code></code></em>),
1317      <code class="function">mallctlnametomib</code>(<em class="parameter"><code></code></em>), and
1318      <code class="function">mallctlbymib</code>(<em class="parameter"><code></code></em>) functions return 0 on
1319      success; otherwise they return an error value.  The functions will fail
1320      if:
1321        </p><div class="variablelist"><dl><dt><span class="term"><span class="errorname">EINVAL</span></span></dt><dd><p><em class="parameter"><code>newp</code></em> is not
1322            <code class="constant">NULL</code>, and <em class="parameter"><code>newlen</code></em> is too
1323            large or too small.  Alternatively, <em class="parameter"><code>*oldlenp</code></em>
1324            is too large or too small; in this case as much data as possible
1325            are read despite the error.</p></dd><dt><span class="term"><span class="errorname">ENOMEM</span></span></dt><dd><p><em class="parameter"><code>*oldlenp</code></em> is too short to
1326            hold the requested value.</p></dd><dt><span class="term"><span class="errorname">ENOENT</span></span></dt><dd><p><em class="parameter"><code>name</code></em> or
1327            <em class="parameter"><code>mib</code></em> specifies an unknown/invalid
1328            value.</p></dd><dt><span class="term"><span class="errorname">EPERM</span></span></dt><dd><p>Attempt to read or write void value, or attempt to
1329            write read-only value.</p></dd><dt><span class="term"><span class="errorname">EAGAIN</span></span></dt><dd><p>A memory allocation failure
1330            occurred.</p></dd><dt><span class="term"><span class="errorname">EFAULT</span></span></dt><dd><p>An interface with side effects failed in some way
1331            not directly related to <code class="function">mallctl*</code>(<em class="parameter"><code></code></em>)
1332            read/write processing.</p></dd></dl></div><p>
1333      </p></div><div class="refsect2" title="Experimental API"><a name="id286955658"></a><h3>Experimental API</h3><p>The <code class="function">allocm</code>(<em class="parameter"><code></code></em>),
1334      <code class="function">rallocm</code>(<em class="parameter"><code></code></em>),
1335      <code class="function">sallocm</code>(<em class="parameter"><code></code></em>),
1336      <code class="function">dallocm</code>(<em class="parameter"><code></code></em>), and
1337      <code class="function">nallocm</code>(<em class="parameter"><code></code></em>) functions return
1338      <code class="constant">ALLOCM_SUCCESS</code> on success; otherwise they return an
1339      error value.  The <code class="function">allocm</code>(<em class="parameter"><code></code></em>),
1340      <code class="function">rallocm</code>(<em class="parameter"><code></code></em>), and
1341      <code class="function">nallocm</code>(<em class="parameter"><code></code></em>) functions will fail if:
1342        </p><div class="variablelist"><dl><dt><span class="term"><span class="errorname">ALLOCM_ERR_OOM</span></span></dt><dd><p>Out of memory.  Insufficient contiguous memory was
1343            available to service the allocation request.  The
1344            <code class="function">allocm</code>(<em class="parameter"><code></code></em>) function additionally sets
1345            <em class="parameter"><code>*ptr</code></em> to <code class="constant">NULL</code>, whereas
1346            the <code class="function">rallocm</code>(<em class="parameter"><code></code></em>) function leaves
1347            <code class="constant">*ptr</code> unmodified.</p></dd></dl></div><p>
1348      The <code class="function">rallocm</code>(<em class="parameter"><code></code></em>) function will also
1349      fail if:
1350        </p><div class="variablelist"><dl><dt><span class="term"><span class="errorname">ALLOCM_ERR_NOT_MOVED</span></span></dt><dd><p><code class="constant">ALLOCM_NO_MOVE</code> was specified,
1351            but the reallocation request could not be serviced without moving
1352            the object.</p></dd></dl></div><p>
1353      </p></div></div><div class="refsect1" title="ENVIRONMENT"><a name="environment"></a><h2>ENVIRONMENT</h2><p>The following environment variable affects the execution of the
1354    allocation functions:
1355      </p><div class="variablelist"><dl><dt><span class="term"><code class="envar">MALLOC_CONF</code></span></dt><dd><p>If the environment variable
1356          <code class="envar">MALLOC_CONF</code> is set, the characters it contains
1357          will be interpreted as options.</p></dd></dl></div><p>
1358    </p></div><div class="refsect1" title="EXAMPLES"><a name="examples"></a><h2>EXAMPLES</h2><p>To dump core whenever a problem occurs:
1359      </p><pre class="screen">ln -s 'abort:true' /etc/malloc.conf</pre><p>
1360    </p><p>To specify in the source a chunk size that is 16 MiB:
1361      </p><pre class="programlisting">
1362malloc_conf = "lg_chunk:24";</pre></div><div class="refsect1" title="SEE ALSO"><a name="see_also"></a><h2>SEE ALSO</h2><p><span class="citerefentry"><span class="refentrytitle">madvise</span>(2)</span>,
1363    <span class="citerefentry"><span class="refentrytitle">mmap</span>(2)</span>,
1364    <span class="citerefentry"><span class="refentrytitle">sbrk</span>(2)</span>,
1365    <span class="citerefentry"><span class="refentrytitle">utrace</span>(2)</span>,
1366    <span class="citerefentry"><span class="refentrytitle">alloca</span>(3)</span>,
1367    <span class="citerefentry"><span class="refentrytitle">atexit</span>(3)</span>,
1368    <span class="citerefentry"><span class="refentrytitle">getpagesize</span>(3)</span></p></div><div class="refsect1" title="STANDARDS"><a name="standards"></a><h2>STANDARDS</h2><p>The <code class="function">malloc</code>(<em class="parameter"><code></code></em>),
1369    <code class="function">calloc</code>(<em class="parameter"><code></code></em>),
1370    <code class="function">realloc</code>(<em class="parameter"><code></code></em>), and
1371    <code class="function">free</code>(<em class="parameter"><code></code></em>) functions conform to ISO/IEC
1372    9899:1990 (&#8220;ISO C90&#8221;).</p><p>The <code class="function">posix_memalign</code>(<em class="parameter"><code></code></em>) function conforms
1373    to IEEE Std 1003.1-2001 (&#8220;POSIX.1&#8221;).</p></div></div></body></html>
1374