relevant. Use bitwise or (<code language="C">|</code>) operations to
specify one or more of the following:
<variablelist>
- <varlistentry>
+ <varlistentry id="MALLOCX_LG_ALIGN">
<term><constant>MALLOCX_LG_ALIGN(<parameter>la</parameter>)
</constant></term>
that <parameter>la</parameter> is within the valid
range.</para></listitem>
</varlistentry>
- <varlistentry>
+ <varlistentry id="MALLOCX_ALIGN">
<term><constant>MALLOCX_ALIGN(<parameter>a</parameter>)
</constant></term>
validate that <parameter>a</parameter> is a power of 2.
</para></listitem>
</varlistentry>
- <varlistentry>
+ <varlistentry id="MALLOCX_ZERO">
<term><constant>MALLOCX_ZERO</constant></term>
<listitem><para>Initialize newly allocated memory to contain zero
that are initialized to contain zero bytes. If this macro is
absent, newly allocated memory is uninitialized.</para></listitem>
</varlistentry>
- <varlistentry>
+ <varlistentry id="MALLOCX_TCACHE">
+ <term><constant>MALLOCX_TCACHE(<parameter>tc</parameter>)
+ </constant></term>
+
+ <listitem><para>Use the thread-specific cache (tcache) specified by
+ the identifier <parameter>tc</parameter>, which must have been
+ acquired via the <link
+ linkend="tcache.create"><mallctl>tcache.create</mallctl></link>
+ mallctl. This macro does not validate that
+ <parameter>tc</parameter> specifies a valid
+ identifier.</para></listitem>
+ </varlistentry>
+ <varlistentry id="MALLOC_TCACHE_NONE">
+ <term><constant>MALLOCX_TCACHE_NONE</constant></term>
+
+ <listitem><para>Do not use a thread-specific cache (tcache). Unless
+ <constant>MALLOCX_TCACHE(<parameter>tc</parameter>)</constant> or
+ <constant>MALLOCX_TCACHE_NONE</constant> is specified, an
+ automatically managed tcache will be used under many circumstances.
+ This macro cannot be used in the same <parameter>flags</parameter>
+ argument as
+ <constant>MALLOCX_TCACHE(<parameter>tc</parameter>)</constant>.</para></listitem>
+ </varlistentry>
+ <varlistentry id="MALLOCX_ARENA">
<term><constant>MALLOCX_ARENA(<parameter>a</parameter>)
</constant></term>
<listitem><para>Use the arena specified by the index
- <parameter>a</parameter> (and by necessity bypass the thread
- cache). This macro has no effect for regions that were allocated
- via an arena other than the one specified. This macro does not
- validate that <parameter>a</parameter> specifies an arena index in
- the valid range.</para></listitem>
+ <parameter>a</parameter>. This macro has no effect for regions that
+ were allocated via an arena other than the one specified. This
+ macro does not validate that <parameter>a</parameter> specifies an
+ arena index in the valid range.</para></listitem>
</varlistentry>
</variablelist>
</para>
<para>The <function>mallocx<parameter/></function> function allocates at
least <parameter>size</parameter> bytes of memory, and returns a pointer
to the base address of the allocation. Behavior is undefined if
- <parameter>size</parameter> is <constant>0</constant>, or if request size
- overflows due to size class and/or alignment constraints.</para>
+ <parameter>size</parameter> is <constant>0</constant>.</para>
<para>The <function>rallocx<parameter/></function> function resizes the
allocation at <parameter>ptr</parameter> to be at least
<parameter>size</parameter> bytes, and returns a pointer to the base
address of the resulting allocation, which may or may not have moved from
its original location. Behavior is undefined if
- <parameter>size</parameter> is <constant>0</constant>, or if request size
- overflows due to size class and/or alignment constraints.</para>
+ <parameter>size</parameter> is <constant>0</constant>.</para>
<para>The <function>xallocx<parameter/></function> function resizes the
allocation at <parameter>ptr</parameter> in place to be at least
memory, but it performs the same size computation as the
<function>mallocx<parameter/></function> function, and returns the real
size of the allocation that would result from the equivalent
- <function>mallocx<parameter/></function> function call. Behavior is
- undefined if <parameter>size</parameter> is <constant>0</constant>, or if
- request size overflows due to size class and/or alignment
- constraints.</para>
+ <function>mallocx<parameter/></function> function call, or
+ <constant>0</constant> if the inputs exceed the maximum supported size
+ class and/or alignment. Behavior is undefined if
+ <parameter>size</parameter> is <constant>0</constant>.</para>
<para>The <function>mallctl<parameter/></function> function provides a
general interface for introspecting the memory allocator, as well as
functions simultaneously. If <option>--enable-stats</option> is
specified during configuration, “m” and “a” can
be specified to omit merged arena and per arena statistics, respectively;
- “b” and “l” can be specified to omit per size
- class statistics for bins and large objects, respectively. Unrecognized
- characters are silently ignored. Note that thread caching may prevent
- some statistics from being completely up to date, since extra locking
- would be required to merge counters that track thread cache operations.
+ “b”, “l”, and “h” can be specified to
+ omit per size class statistics for bins, large objects, and huge objects,
+ respectively. Unrecognized characters are silently ignored. Note that
+ thread caching may prevent some statistics from being completely up to
+ date, since extra locking would be required to merge counters that track
+ thread cache operations.
</para>
<para>The <function>malloc_usable_size<parameter/></function> function
routines, the allocator initializes its internals based in part on various
options that can be specified at compile- or run-time.</para>
- <para>The string pointed to by the global variable
- <varname>malloc_conf</varname>, the “name” of the file
- referenced by the symbolic link named <filename
- class="symlink">/etc/malloc.conf</filename>, and the value of the
+ <para>The string specified via <option>--with-malloc-conf</option>, the
+ string pointed to by the global variable <varname>malloc_conf</varname>, the
+ “name” of the file referenced by the symbolic link named
+ <filename class="symlink">/etc/malloc.conf</filename>, and the value of the
environment variable <envar>MALLOC_CONF</envar>, will be interpreted, in
that order, from left to right as options. Note that
<varname>malloc_conf</varname> may be read before
<function>main<parameter/></function> is entered, so the declaration of
<varname>malloc_conf</varname> should specify an initializer that contains
- the final value to be read by jemalloc. <varname>malloc_conf</varname> is
- a compile-time setting, whereas <filename
- class="symlink">/etc/malloc.conf</filename> and <envar>MALLOC_CONF</envar>
- can be safely set any time prior to program invocation.</para>
+ the final value to be read by jemalloc. <option>--with-malloc-conf</option>
+ and <varname>malloc_conf</varname> are compile-time mechanisms, whereas
+ <filename class="symlink">/etc/malloc.conf</filename> and
+ <envar>MALLOC_CONF</envar> can be safely set any time prior to program
+ invocation.</para>
<para>An options string is a comma-separated list of option:value pairs.
There is one key corresponding to each <link
common case, but it increases memory usage and fragmentation, since a
bounded number of objects can remain allocated in each thread cache.</para>
- <para>Memory is conceptually broken into equal-sized chunks, where the
- chunk size is a power of two that is greater than the page size. Chunks
- are always aligned to multiples of the chunk size. This alignment makes it
- possible to find metadata for user objects very quickly.</para>
-
- <para>User objects are broken into three categories according to size:
- small, large, and huge. Small objects are smaller than one page. Large
- objects are smaller than the chunk size. Huge objects are a multiple of
- the chunk size. Small and large objects are managed entirely by arenas;
- huge objects are additionally aggregated in a single data structure that is
- shared by all threads. Huge objects are typically used by applications
- infrequently enough that this single data structure is not a scalability
- issue.</para>
-
- <para>Each chunk that is managed by an arena tracks its contents as runs of
+ <para>Memory is conceptually broken into equal-sized chunks, where the chunk
+ size is a power of two that is greater than the page size. Chunks are
+ always aligned to multiples of the chunk size. This alignment makes it
+ possible to find metadata for user objects very quickly. User objects are
+ broken into three categories according to size: small, large, and huge.
+ Multiple small and large objects can reside within a single chunk, whereas
+ huge objects each have one or more chunks backing them. Each chunk that
+ contains small and/or large objects tracks its contents as runs of
contiguous pages (unused, backing a set of small objects, or backing one
- large object). The combination of chunk alignment and chunk page maps
- makes it possible to determine all metadata regarding small and large
- allocations in constant time.</para>
+ large object). The combination of chunk alignment and chunk page maps makes
+ it possible to determine all metadata regarding small and large allocations
+ in constant time.</para>
<para>Small objects are managed in groups by page runs. Each run maintains
- a frontier and free list to track which regions are in use. Allocation
- requests that are no more than half the quantum (8 or 16, depending on
- architecture) are rounded up to the nearest power of two that is at least
- <code language="C">sizeof(<type>double</type>)</code>. All other small
- object size classes are multiples of the quantum, spaced such that internal
- fragmentation is limited to approximately 25% for all but the smallest size
- classes. Allocation requests that are larger than the maximum small size
- class, but small enough to fit in an arena-managed chunk (see the <link
- linkend="opt.lg_chunk"><mallctl>opt.lg_chunk</mallctl></link> option), are
- rounded up to the nearest run size. Allocation requests that are too large
- to fit in an arena-managed chunk are rounded up to the nearest multiple of
- the chunk size.</para>
+ a bitmap to track which regions are in use. Allocation requests that are no
+ more than half the quantum (8 or 16, depending on architecture) are rounded
+ up to the nearest power of two that is at least <code
+ language="C">sizeof(<type>double</type>)</code>. All other object size
+ classes are multiples of the quantum, spaced such that there are four size
+ classes for each doubling in size, which limits internal fragmentation to
+ approximately 20% for all but the smallest size classes. Small size classes
+ are smaller than four times the page size, large size classes are smaller
+ than the chunk size (see the <link
+ linkend="opt.lg_chunk"><mallctl>opt.lg_chunk</mallctl></link> option), and
+ huge size classes extend from the chunk size up to one size class less than
+ the full address space size.</para>
<para>Allocations are packed tightly together, which can be an issue for
multi-threaded applications. If you need to assure that allocations do not
nearest multiple of the cacheline size, or specify cacheline alignment when
allocating.</para>
- <para>Assuming 4 MiB chunks, 4 KiB pages, and a 16-byte quantum on a 64-bit
- system, the size classes in each category are as shown in <xref
+ <para>The <function>realloc<parameter/></function>,
+ <function>rallocx<parameter/></function>, and
+ <function>xallocx<parameter/></function> functions may resize allocations
+ without moving them under limited circumstances. Unlike the
+ <function>*allocx<parameter/></function> API, the standard API does not
+ officially round up the usable size of an allocation to the nearest size
+ class, so technically it is necessary to call
+ <function>realloc<parameter/></function> to grow e.g. a 9-byte allocation to
+ 16 bytes, or shrink a 16-byte allocation to 9 bytes. Growth and shrinkage
+ trivially succeeds in place as long as the pre-size and post-size both round
+ up to the same size class. No other API guarantees are made regarding
+ in-place resizing, but the current implementation also tries to resize large
+ and huge allocations in place, as long as the pre-size and post-size are
+ both large or both huge. In such cases shrinkage always succeeds for large
+ size classes, but for huge size classes the chunk allocator must support
+ splitting (see <link
+ linkend="arena.i.chunk_hooks"><mallctl>arena.<i>.chunk_hooks</mallctl></link>).
+ Growth only succeeds if the trailing memory is currently available, and
+ additionally for huge size classes the chunk allocator must support
+ merging.</para>
+
+ <para>Assuming 2 MiB chunks, 4 KiB pages, and a 16-byte quantum on a
+ 64-bit system, the size classes in each category are as shown in <xref
linkend="size_classes" xrefstyle="template:Table %n"/>.</para>
<table xml:id="size_classes" frame="all">
</thead>
<tbody>
<row>
- <entry morerows="6">Small</entry>
+ <entry morerows="8">Small</entry>
<entry>lg</entry>
<entry>[8]</entry>
</row>
<row>
<entry>16</entry>
- <entry>[16, 32, 48, ..., 128]</entry>
+ <entry>[16, 32, 48, 64, 80, 96, 112, 128]</entry>
</row>
<row>
<entry>32</entry>
</row>
<row>
<entry>512</entry>
- <entry>[2560, 3072, 3584]</entry>
+ <entry>[2560, 3072, 3584, 4096]</entry>
+ </row>
+ <row>
+ <entry>1 KiB</entry>
+ <entry>[5 KiB, 6 KiB, 7 KiB, 8 KiB]</entry>
+ </row>
+ <row>
+ <entry>2 KiB</entry>
+ <entry>[10 KiB, 12 KiB, 14 KiB]</entry>
+ </row>
+ <row>
+ <entry morerows="7">Large</entry>
+ <entry>2 KiB</entry>
+ <entry>[16 KiB]</entry>
</row>
<row>
- <entry>Large</entry>
<entry>4 KiB</entry>
- <entry>[4 KiB, 8 KiB, 12 KiB, ..., 4072 KiB]</entry>
+ <entry>[20 KiB, 24 KiB, 28 KiB, 32 KiB]</entry>
+ </row>
+ <row>
+ <entry>8 KiB</entry>
+ <entry>[40 KiB, 48 KiB, 54 KiB, 64 KiB]</entry>
+ </row>
+ <row>
+ <entry>16 KiB</entry>
+ <entry>[80 KiB, 96 KiB, 112 KiB, 128 KiB]</entry>
+ </row>
+ <row>
+ <entry>32 KiB</entry>
+ <entry>[160 KiB, 192 KiB, 224 KiB, 256 KiB]</entry>
+ </row>
+ <row>
+ <entry>64 KiB</entry>
+ <entry>[320 KiB, 384 KiB, 448 KiB, 512 KiB]</entry>
+ </row>
+ <row>
+ <entry>128 KiB</entry>
+ <entry>[640 KiB, 768 KiB, 896 KiB, 1 MiB]</entry>
+ </row>
+ <row>
+ <entry>256 KiB</entry>
+ <entry>[1280 KiB, 1536 KiB, 1792 KiB]</entry>
+ </row>
+ <row>
+ <entry morerows="6">Huge</entry>
+ <entry>256 KiB</entry>
+ <entry>[2 MiB]</entry>
+ </row>
+ <row>
+ <entry>512 KiB</entry>
+ <entry>[2560 KiB, 3 MiB, 3584 KiB, 4 MiB]</entry>
+ </row>
+ <row>
+ <entry>1 MiB</entry>
+ <entry>[5 MiB, 6 MiB, 7 MiB, 8 MiB]</entry>
+ </row>
+ <row>
+ <entry>2 MiB</entry>
+ <entry>[10 MiB, 12 MiB, 14 MiB, 16 MiB]</entry>
</row>
<row>
- <entry>Huge</entry>
<entry>4 MiB</entry>
- <entry>[4 MiB, 8 MiB, 12 MiB, ...]</entry>
+ <entry>[20 MiB, 24 MiB, 28 MiB, 32 MiB]</entry>
+ </row>
+ <row>
+ <entry>8 MiB</entry>
+ <entry>[40 MiB, 48 MiB, 56 MiB, 64 MiB]</entry>
+ </row>
+ <row>
+ <entry>...</entry>
+ <entry>...</entry>
</row>
</tbody>
</tgroup>
detecting whether another thread caused a refresh.</para></listitem>
</varlistentry>
+ <varlistentry id="config.cache_oblivious">
+ <term>
+ <mallctl>config.cache_oblivious</mallctl>
+ (<type>bool</type>)
+ <literal>r-</literal>
+ </term>
+ <listitem><para><option>--enable-cache-oblivious</option> was specified
+ during build configuration.</para></listitem>
+ </varlistentry>
+
<varlistentry id="config.debug">
<term>
<mallctl>config.debug</mallctl>
during build configuration.</para></listitem>
</varlistentry>
+ <varlistentry id="config.malloc_conf">
+ <term>
+ <mallctl>config.malloc_conf</mallctl>
+ (<type>const char *</type>)
+ <literal>r-</literal>
+ </term>
+ <listitem><para>Embedded configure-time-specified run-time options
+ string, empty unless <option>--with-malloc-conf</option> was specified
+ during build configuration.</para></listitem>
+ </varlistentry>
+
<varlistentry id="config.munmap">
<term>
<mallctl>config.munmap</mallctl>
<listitem><para>Virtual memory chunk size (log base 2). If a chunk
size outside the supported size range is specified, the size is
silently clipped to the minimum/maximum supported size. The default
- chunk size is 4 MiB (2^22).
+ chunk size is 2 MiB (2^21).
</para></listitem>
</varlistentry>
<varlistentry id="opt.narenas">
<term>
<mallctl>opt.narenas</mallctl>
- (<type>size_t</type>)
+ (<type>unsigned</type>)
<literal>r-</literal>
</term>
<listitem><para>Maximum number of arenas to use for automatic
number of CPUs, or one if there is a single CPU.</para></listitem>
</varlistentry>
+ <varlistentry id="opt.purge">
+ <term>
+ <mallctl>opt.purge</mallctl>
+ (<type>const char *</type>)
+ <literal>r-</literal>
+ </term>
+ <listitem><para>Purge mode is “ratio” (default) or
+ “decay”. See <link
+ linkend="opt.lg_dirty_mult"><mallctl>opt.lg_dirty_mult</mallctl></link>
+ for details of the ratio mode. See <link
+ linkend="opt.decay_time"><mallctl>opt.decay_time</mallctl></link> for
+ details of the decay mode.</para></listitem>
+ </varlistentry>
+
<varlistentry id="opt.lg_dirty_mult">
<term>
<mallctl>opt.lg_dirty_mult</mallctl>
provides the kernel with sufficient information to recycle dirty pages
if physical memory becomes scarce and the pages remain unused. The
default minimum ratio is 8:1 (2^3:1); an option value of -1 will
- disable dirty page purging.</para></listitem>
+ disable dirty page purging. See <link
+ linkend="arenas.lg_dirty_mult"><mallctl>arenas.lg_dirty_mult</mallctl></link>
+ and <link
+ linkend="arena.i.lg_dirty_mult"><mallctl>arena.<i>.lg_dirty_mult</mallctl></link>
+ for related dynamic control options.</para></listitem>
+ </varlistentry>
+
+ <varlistentry id="opt.decay_time">
+ <term>
+ <mallctl>opt.decay_time</mallctl>
+ (<type>ssize_t</type>)
+ <literal>r-</literal>
+ </term>
+ <listitem><para>Approximate time in seconds from the creation of a set
+ of unused dirty pages until an equivalent set of unused dirty pages is
+ purged and/or reused. The pages are incrementally purged according to a
+ sigmoidal decay curve that starts and ends with zero purge rate. A
+ decay time of 0 causes all unused dirty pages to be purged immediately
+ upon creation. A decay time of -1 disables purging. The default decay
+ time is 10 seconds. See <link
+ linkend="arenas.decay_time"><mallctl>arenas.decay_time</mallctl></link>
+ and <link
+ linkend="arena.i.decay_time"><mallctl>arena.<i>.decay_time</mallctl></link>
+ for related dynamic control options.
+ </para></listitem>
</varlistentry>
<varlistentry id="opt.stats_print">
<option>--enable-stats</option> is specified during configuration, this
has the potential to cause deadlock for a multi-threaded process that
exits while one or more threads are executing in the memory allocation
- functions. Therefore, this option should only be used with care; it is
- primarily intended as a performance tuning aid during application
+ functions. Furthermore, <function>atexit<parameter/></function> may
+ allocate memory during application initialization and then deadlock
+ internally when jemalloc in turn calls
+ <function>atexit<parameter/></function>, so this option is not
+ univerally usable (though the application can register its own
+ <function>atexit<parameter/></function> function with equivalent
+ functionality). Therefore, this option should only be used with care;
+ it is primarily intended as a performance tuning aid during application
development. This option is disabled by default.</para></listitem>
</varlistentry>
<varlistentry id="opt.junk">
<term>
<mallctl>opt.junk</mallctl>
- (<type>bool</type>)
+ (<type>const char *</type>)
<literal>r-</literal>
[<option>--enable-fill</option>]
</term>
- <listitem><para>Junk filling enabled/disabled. If enabled, each byte
- of uninitialized allocated memory will be initialized to
- <literal>0xa5</literal>. All deallocated memory will be initialized to
- <literal>0x5a</literal>. This is intended for debugging and will
- impact performance negatively. This option is disabled by default
- unless <option>--enable-debug</option> is specified during
- configuration, in which case it is enabled by default unless running
- inside <ulink
+ <listitem><para>Junk filling. If set to "alloc", each byte of
+ uninitialized allocated memory will be initialized to
+ <literal>0xa5</literal>. If set to "free", all deallocated memory will
+ be initialized to <literal>0x5a</literal>. If set to "true", both
+ allocated and deallocated memory will be initialized, and if set to
+ "false", junk filling be disabled entirely. This is intended for
+ debugging and will impact performance negatively. This option is
+ "false" by default unless <option>--enable-debug</option> is specified
+ during configuration, in which case it is "true" by default unless
+ running inside <ulink
url="http://valgrind.org/">Valgrind</ulink>.</para></listitem>
</varlistentry>
<literal>r-</literal>
[<option>--enable-tcache</option>]
</term>
- <listitem><para>Thread-specific caching enabled/disabled. When there
- are multiple threads, each thread uses a thread-specific cache for
- objects up to a certain size. Thread-specific caching allows many
- allocations to be satisfied without performing any thread
- synchronization, at the cost of increased memory use. See the
- <link
+ <listitem><para>Thread-specific caching (tcache) enabled/disabled. When
+ there are multiple threads, each thread uses a tcache for objects up to
+ a certain size. Thread-specific caching allows many allocations to be
+ satisfied without performing any thread synchronization, at the cost of
+ increased memory use. See the <link
linkend="opt.lg_tcache_max"><mallctl>opt.lg_tcache_max</mallctl></link>
option for related tuning information. This option is enabled by
default unless running inside <ulink
[<option>--enable-tcache</option>]
</term>
<listitem><para>Maximum size class (log base 2) to cache in the
- thread-specific cache. At a minimum, all small size classes are
- cached, and at a maximum all large size classes are cached. The
+ thread-specific cache (tcache). At a minimum, all small size classes
+ are cached, and at a maximum all large size classes are cached. The
default maximum is 32 KiB (2^15).</para></listitem>
</varlistentry>
option for information on high-water-triggered profile dumping, and the
<link linkend="opt.prof_final"><mallctl>opt.prof_final</mallctl></link>
option for final profile dumping. Profile output is compatible with
- the included <command>pprof</command> Perl script, which originates
- from the <ulink url="http://code.google.com/p/gperftools/">gperftools
- package</ulink>.</para></listitem>
+ the <command>jeprof</command> command, which is based on the
+ <command>pprof</command> that is developed as part of the <ulink
+ url="http://code.google.com/p/gperftools/">gperftools
+ package</ulink>. See <link linkend="heap_profile_format">HEAP PROFILE
+ FORMAT</link> for heap profile format documentation.</para></listitem>
</varlistentry>
<varlistentry id="opt.prof_prefix">
<term>
<mallctl>opt.prof_active</mallctl>
(<type>bool</type>)
- <literal>rw</literal>
+ <literal>r-</literal>
[<option>--enable-prof</option>]
</term>
<listitem><para>Profiling activated/deactivated. This is a secondary
<literal>r-</literal>
[<option>--enable-prof</option>]
</term>
- <listitem><para>Trigger a memory profile dump every time the total
- virtual memory exceeds the previous maximum. Profiles are dumped to
- files named according to the pattern
- <filename><prefix>.<pid>.<seq>.u<useq>.heap</filename>,
- where <literal><prefix></literal> is controlled by the <link
- linkend="opt.prof_prefix"><mallctl>opt.prof_prefix</mallctl></link>
- option. This option is disabled by default.</para></listitem>
+ <listitem><para>Set the initial state of <link
+ linkend="prof.gdump"><mallctl>prof.gdump</mallctl></link>, which when
+ enabled triggers a memory profile dump every time the total virtual
+ memory exceeds the previous maximum. This option is disabled by
+ default.</para></listitem>
</varlistentry>
<varlistentry id="opt.prof_final">
<filename><prefix>.<pid>.<seq>.f.heap</filename>,
where <literal><prefix></literal> is controlled by the <link
linkend="opt.prof_prefix"><mallctl>opt.prof_prefix</mallctl></link>
- option. This option is enabled by default.</para></listitem>
+ option. Note that <function>atexit<parameter/></function> may allocate
+ memory during application initialization and then deadlock internally
+ when jemalloc in turn calls <function>atexit<parameter/></function>, so
+ this option is not univerally usable (though the application can
+ register its own <function>atexit<parameter/></function> function with
+ equivalent functionality). This option is disabled by
+ default.</para></listitem>
</varlistentry>
<varlistentry id="opt.prof_leak">
<listitem><para>Enable/disable calling thread's tcache. The tcache is
implicitly flushed as a side effect of becoming
disabled (see <link
- lenkend="thread.tcache.flush"><mallctl>thread.tcache.flush</mallctl></link>).
+ linkend="thread.tcache.flush"><mallctl>thread.tcache.flush</mallctl></link>).
</para></listitem>
</varlistentry>
<literal>--</literal>
[<option>--enable-tcache</option>]
</term>
- <listitem><para>Flush calling thread's tcache. This interface releases
- all cached objects and internal data structures associated with the
- calling thread's thread-specific cache. Ordinarily, this interface
+ <listitem><para>Flush calling thread's thread-specific cache (tcache).
+ This interface releases all cached objects and internal data structures
+ associated with the calling thread's tcache. Ordinarily, this interface
need not be called, since automatic periodic incremental garbage
collection occurs, and the thread cache is automatically discarded when
a thread exits. However, garbage collection is triggered by allocation
can cause asynchronous string deallocation. Furthermore, each
invocation of this interface can only read or write; simultaneous
read/write is not supported due to string lifetime limitations. The
- name string must nil-terminated and comprised only of characters in the
- sets recognized
+ name string must be nil-terminated and comprised only of characters in
+ the sets recognized
by <citerefentry><refentrytitle>isgraph</refentrytitle>
<manvolnum>3</manvolnum></citerefentry> and
<citerefentry><refentrytitle>isblank</refentrytitle>
default.</para></listitem>
</varlistentry>
+ <varlistentry id="tcache.create">
+ <term>
+ <mallctl>tcache.create</mallctl>
+ (<type>unsigned</type>)
+ <literal>r-</literal>
+ [<option>--enable-tcache</option>]
+ </term>
+ <listitem><para>Create an explicit thread-specific cache (tcache) and
+ return an identifier that can be passed to the <link
+ linkend="MALLOCX_TCACHE"><constant>MALLOCX_TCACHE(<parameter>tc</parameter>)</constant></link>
+ macro to explicitly use the specified cache rather than the
+ automatically managed one that is used by default. Each explicit cache
+ can be used by only one thread at a time; the application must assure
+ that this constraint holds.
+ </para></listitem>
+ </varlistentry>
+
+ <varlistentry id="tcache.flush">
+ <term>
+ <mallctl>tcache.flush</mallctl>
+ (<type>unsigned</type>)
+ <literal>-w</literal>
+ [<option>--enable-tcache</option>]
+ </term>
+ <listitem><para>Flush the specified thread-specific cache (tcache). The
+ same considerations apply to this interface as to <link
+ linkend="thread.tcache.flush"><mallctl>thread.tcache.flush</mallctl></link>,
+ except that the tcache will never be automatically discarded.
+ </para></listitem>
+ </varlistentry>
+
+ <varlistentry id="tcache.destroy">
+ <term>
+ <mallctl>tcache.destroy</mallctl>
+ (<type>unsigned</type>)
+ <literal>-w</literal>
+ [<option>--enable-tcache</option>]
+ </term>
+ <listitem><para>Flush the specified thread-specific cache (tcache) and
+ make the identifier available for use during a future tcache creation.
+ </para></listitem>
+ </varlistentry>
+
<varlistentry id="arena.i.purge">
<term>
<mallctl>arena.<i>.purge</mallctl>
(<type>void</type>)
<literal>--</literal>
</term>
- <listitem><para>Purge unused dirty pages for arena <i>, or for
+ <listitem><para>Purge all unused dirty pages for arena <i>, or for
all arenas if <i> equals <link
linkend="arenas.narenas"><mallctl>arenas.narenas</mallctl></link>.
</para></listitem>
</varlistentry>
+ <varlistentry id="arena.i.decay">
+ <term>
+ <mallctl>arena.<i>.decay</mallctl>
+ (<type>void</type>)
+ <literal>--</literal>
+ </term>
+ <listitem><para>Trigger decay-based purging of unused dirty pages for
+ arena <i>, or for all arenas if <i> equals <link
+ linkend="arenas.narenas"><mallctl>arenas.narenas</mallctl></link>.
+ The proportion of unused dirty pages to be purged depends on the current
+ time; see <link
+ linkend="opt.decay_time"><mallctl>opt.decay_time</mallctl></link> for
+ details.</para></listitem>
+ </varlistentry>
+
<varlistentry id="arena.i.dss">
<term>
<mallctl>arena.<i>.dss</mallctl>
settings.</para></listitem>
</varlistentry>
- <varlistentry id="arena.i.chunk.alloc">
+ <varlistentry id="arena.i.lg_dirty_mult">
<term>
- <mallctl>arena.<i>.chunk.alloc</mallctl>
- (<type>chunk_alloc_t *</type>)
+ <mallctl>arena.<i>.lg_dirty_mult</mallctl>
+ (<type>ssize_t</type>)
<literal>rw</literal>
</term>
- <listitem><para>Get or set the chunk allocation function for arena
- <i>. If setting, the chunk deallocation function should
- also be set via <link linkend="arena.i.chunk.dalloc">
- <mallctl>arena.<i>.chunk.dalloc</mallctl></link> to a companion
- function that knows how to deallocate the chunks.
+ <listitem><para>Current per-arena minimum ratio (log base 2) of active
+ to dirty pages for arena <i>. Each time this interface is set and
+ the ratio is increased, pages are synchronously purged as necessary to
+ impose the new ratio. See <link
+ linkend="opt.lg_dirty_mult"><mallctl>opt.lg_dirty_mult</mallctl></link>
+ for additional information.</para></listitem>
+ </varlistentry>
+
+ <varlistentry id="arena.i.decay_time">
+ <term>
+ <mallctl>arena.<i>.decay_time</mallctl>
+ (<type>ssize_t</type>)
+ <literal>rw</literal>
+ </term>
+ <listitem><para>Current per-arena approximate time in seconds from the
+ creation of a set of unused dirty pages until an equivalent set of
+ unused dirty pages is purged and/or reused. Each time this interface is
+ set, all currently unused dirty pages are considered to have fully
+ decayed, which causes immediate purging of all unused dirty pages unless
+ the decay time is set to -1 (i.e. purging disabled). See <link
+ linkend="opt.decay_time"><mallctl>opt.decay_time</mallctl></link> for
+ additional information.</para></listitem>
+ </varlistentry>
+
+ <varlistentry id="arena.i.chunk_hooks">
+ <term>
+ <mallctl>arena.<i>.chunk_hooks</mallctl>
+ (<type>chunk_hooks_t</type>)
+ <literal>rw</literal>
+ </term>
+ <listitem><para>Get or set the chunk management hook functions for arena
+ <i>. The functions must be capable of operating on all extant
+ chunks associated with arena <i>, usually by passing unknown
+ chunks to the replaced functions. In practice, it is feasible to
+ control allocation for arenas created via <link
+ linkend="arenas.extend"><mallctl>arenas.extend</mallctl></link> such
+ that all chunks originate from an application-supplied chunk allocator
+ (by setting custom chunk hook functions just after arena creation), but
+ the automatically created arenas may have already created chunks prior
+ to the application having an opportunity to take over chunk
+ allocation.</para>
+
+ <programlisting language="C"><![CDATA[
+typedef struct {
+ chunk_alloc_t *alloc;
+ chunk_dalloc_t *dalloc;
+ chunk_commit_t *commit;
+ chunk_decommit_t *decommit;
+ chunk_purge_t *purge;
+ chunk_split_t *split;
+ chunk_merge_t *merge;
+} chunk_hooks_t;]]></programlisting>
+ <para>The <type>chunk_hooks_t</type> structure comprises function
+ pointers which are described individually below. jemalloc uses these
+ functions to manage chunk lifetime, which starts off with allocation of
+ mapped committed memory, in the simplest case followed by deallocation.
+ However, there are performance and platform reasons to retain chunks for
+ later reuse. Cleanup attempts cascade from deallocation to decommit to
+ purging, which gives the chunk management functions opportunities to
+ reject the most permanent cleanup operations in favor of less permanent
+ (and often less costly) operations. The chunk splitting and merging
+ operations can also be opted out of, but this is mainly intended to
+ support platforms on which virtual memory mappings provided by the
+ operating system kernel do not automatically coalesce and split, e.g.
+ Windows.</para>
+
<funcsynopsis><funcprototype>
<funcdef>typedef void *<function>(chunk_alloc_t)</function></funcdef>
<paramdef>void *<parameter>chunk</parameter></paramdef>
<paramdef>size_t <parameter>size</parameter></paramdef>
<paramdef>size_t <parameter>alignment</parameter></paramdef>
<paramdef>bool *<parameter>zero</parameter></paramdef>
+ <paramdef>bool *<parameter>commit</parameter></paramdef>
<paramdef>unsigned <parameter>arena_ind</parameter></paramdef>
</funcprototype></funcsynopsis>
- A chunk allocation function conforms to the <type>chunk_alloc_t</type>
- type and upon success returns a pointer to <parameter>size</parameter>
- bytes of memory on behalf of arena <parameter>arena_ind</parameter> such
- that the chunk's base address is a multiple of
- <parameter>alignment</parameter>, as well as setting
- <parameter>*zero</parameter> to indicate whether the chunk is zeroed.
- Upon error the function returns <constant>NULL</constant> and leaves
- <parameter>*zero</parameter> unmodified. The
+ <literallayout></literallayout>
+ <para>A chunk allocation function conforms to the
+ <type>chunk_alloc_t</type> type and upon success returns a pointer to
+ <parameter>size</parameter> bytes of mapped memory on behalf of arena
+ <parameter>arena_ind</parameter> such that the chunk's base address is a
+ multiple of <parameter>alignment</parameter>, as well as setting
+ <parameter>*zero</parameter> to indicate whether the chunk is zeroed and
+ <parameter>*commit</parameter> to indicate whether the chunk is
+ committed. Upon error the function returns <constant>NULL</constant>
+ and leaves <parameter>*zero</parameter> and
+ <parameter>*commit</parameter> unmodified. The
<parameter>size</parameter> parameter is always a multiple of the chunk
size. The <parameter>alignment</parameter> parameter is always a power
of two at least as large as the chunk size. Zeroing is mandatory if
- <parameter>*zero</parameter> is true upon function entry. If
- <parameter>chunk</parameter> is not <constant>NULL</constant>, the
- returned pointer must be <parameter>chunk</parameter> or
- <constant>NULL</constant> if it could not be allocated.</para>
-
- <para>Note that replacing the default chunk allocation function makes
- the arena's <link
+ <parameter>*zero</parameter> is true upon function entry. Committing is
+ mandatory if <parameter>*commit</parameter> is true upon function entry.
+ If <parameter>chunk</parameter> is not <constant>NULL</constant>, the
+ returned pointer must be <parameter>chunk</parameter> on success or
+ <constant>NULL</constant> on error. Committed memory may be committed
+ in absolute terms as on a system that does not overcommit, or in
+ implicit terms as on a system that overcommits and satisfies physical
+ memory needs on demand via soft page faults. Note that replacing the
+ default chunk allocation function makes the arena's <link
linkend="arena.i.dss"><mallctl>arena.<i>.dss</mallctl></link>
- setting irrelevant.</para></listitem>
- </varlistentry>
+ setting irrelevant.</para>
- <varlistentry id="arena.i.chunk.dalloc">
- <term>
- <mallctl>arena.<i>.chunk.dalloc</mallctl>
- (<type>chunk_dalloc_t *</type>)
- <literal>rw</literal>
- </term>
- <listitem><para>Get or set the chunk deallocation function for arena
- <i>. If setting, the chunk deallocation function must
- be capable of deallocating all extant chunks associated with arena
- <i>, usually by passing unknown chunks to the deallocation
- function that was replaced. In practice, it is feasible to control
- allocation for arenas created via <link
- linkend="arenas.extend"><mallctl>arenas.extend</mallctl></link> such
- that all chunks originate from an application-supplied chunk allocator
- (by setting custom chunk allocation/deallocation functions just after
- arena creation), but the automatically created arenas may have already
- created chunks prior to the application having an opportunity to take
- over chunk allocation.
<funcsynopsis><funcprototype>
- <funcdef>typedef void <function>(chunk_dalloc_t)</function></funcdef>
+ <funcdef>typedef bool <function>(chunk_dalloc_t)</function></funcdef>
<paramdef>void *<parameter>chunk</parameter></paramdef>
<paramdef>size_t <parameter>size</parameter></paramdef>
+ <paramdef>bool <parameter>committed</parameter></paramdef>
<paramdef>unsigned <parameter>arena_ind</parameter></paramdef>
</funcprototype></funcsynopsis>
+ <literallayout></literallayout>
+ <para>
A chunk deallocation function conforms to the
<type>chunk_dalloc_t</type> type and deallocates a
- <parameter>chunk</parameter> of given <parameter>size</parameter> on
- behalf of arena <parameter>arena_ind</parameter>.</para></listitem>
+ <parameter>chunk</parameter> of given <parameter>size</parameter> with
+ <parameter>committed</parameter>/decommited memory as indicated, on
+ behalf of arena <parameter>arena_ind</parameter>, returning false upon
+ success. If the function returns true, this indicates opt-out from
+ deallocation; the virtual memory mapping associated with the chunk
+ remains mapped, in the same commit state, and available for future use,
+ in which case it will be automatically retained for later reuse.</para>
+
+ <funcsynopsis><funcprototype>
+ <funcdef>typedef bool <function>(chunk_commit_t)</function></funcdef>
+ <paramdef>void *<parameter>chunk</parameter></paramdef>
+ <paramdef>size_t <parameter>size</parameter></paramdef>
+ <paramdef>size_t <parameter>offset</parameter></paramdef>
+ <paramdef>size_t <parameter>length</parameter></paramdef>
+ <paramdef>unsigned <parameter>arena_ind</parameter></paramdef>
+ </funcprototype></funcsynopsis>
+ <literallayout></literallayout>
+ <para>A chunk commit function conforms to the
+ <type>chunk_commit_t</type> type and commits zeroed physical memory to
+ back pages within a <parameter>chunk</parameter> of given
+ <parameter>size</parameter> at <parameter>offset</parameter> bytes,
+ extending for <parameter>length</parameter> on behalf of arena
+ <parameter>arena_ind</parameter>, returning false upon success.
+ Committed memory may be committed in absolute terms as on a system that
+ does not overcommit, or in implicit terms as on a system that
+ overcommits and satisfies physical memory needs on demand via soft page
+ faults. If the function returns true, this indicates insufficient
+ physical memory to satisfy the request.</para>
+
+ <funcsynopsis><funcprototype>
+ <funcdef>typedef bool <function>(chunk_decommit_t)</function></funcdef>
+ <paramdef>void *<parameter>chunk</parameter></paramdef>
+ <paramdef>size_t <parameter>size</parameter></paramdef>
+ <paramdef>size_t <parameter>offset</parameter></paramdef>
+ <paramdef>size_t <parameter>length</parameter></paramdef>
+ <paramdef>unsigned <parameter>arena_ind</parameter></paramdef>
+ </funcprototype></funcsynopsis>
+ <literallayout></literallayout>
+ <para>A chunk decommit function conforms to the
+ <type>chunk_decommit_t</type> type and decommits any physical memory
+ that is backing pages within a <parameter>chunk</parameter> of given
+ <parameter>size</parameter> at <parameter>offset</parameter> bytes,
+ extending for <parameter>length</parameter> on behalf of arena
+ <parameter>arena_ind</parameter>, returning false upon success, in which
+ case the pages will be committed via the chunk commit function before
+ being reused. If the function returns true, this indicates opt-out from
+ decommit; the memory remains committed and available for future use, in
+ which case it will be automatically retained for later reuse.</para>
+
+ <funcsynopsis><funcprototype>
+ <funcdef>typedef bool <function>(chunk_purge_t)</function></funcdef>
+ <paramdef>void *<parameter>chunk</parameter></paramdef>
+ <paramdef>size_t<parameter>size</parameter></paramdef>
+ <paramdef>size_t <parameter>offset</parameter></paramdef>
+ <paramdef>size_t <parameter>length</parameter></paramdef>
+ <paramdef>unsigned <parameter>arena_ind</parameter></paramdef>
+ </funcprototype></funcsynopsis>
+ <literallayout></literallayout>
+ <para>A chunk purge function conforms to the <type>chunk_purge_t</type>
+ type and optionally discards physical pages within the virtual memory
+ mapping associated with <parameter>chunk</parameter> of given
+ <parameter>size</parameter> at <parameter>offset</parameter> bytes,
+ extending for <parameter>length</parameter> on behalf of arena
+ <parameter>arena_ind</parameter>, returning false if pages within the
+ purged virtual memory range will be zero-filled the next time they are
+ accessed.</para>
+
+ <funcsynopsis><funcprototype>
+ <funcdef>typedef bool <function>(chunk_split_t)</function></funcdef>
+ <paramdef>void *<parameter>chunk</parameter></paramdef>
+ <paramdef>size_t <parameter>size</parameter></paramdef>
+ <paramdef>size_t <parameter>size_a</parameter></paramdef>
+ <paramdef>size_t <parameter>size_b</parameter></paramdef>
+ <paramdef>bool <parameter>committed</parameter></paramdef>
+ <paramdef>unsigned <parameter>arena_ind</parameter></paramdef>
+ </funcprototype></funcsynopsis>
+ <literallayout></literallayout>
+ <para>A chunk split function conforms to the <type>chunk_split_t</type>
+ type and optionally splits <parameter>chunk</parameter> of given
+ <parameter>size</parameter> into two adjacent chunks, the first of
+ <parameter>size_a</parameter> bytes, and the second of
+ <parameter>size_b</parameter> bytes, operating on
+ <parameter>committed</parameter>/decommitted memory as indicated, on
+ behalf of arena <parameter>arena_ind</parameter>, returning false upon
+ success. If the function returns true, this indicates that the chunk
+ remains unsplit and therefore should continue to be operated on as a
+ whole.</para>
+
+ <funcsynopsis><funcprototype>
+ <funcdef>typedef bool <function>(chunk_merge_t)</function></funcdef>
+ <paramdef>void *<parameter>chunk_a</parameter></paramdef>
+ <paramdef>size_t <parameter>size_a</parameter></paramdef>
+ <paramdef>void *<parameter>chunk_b</parameter></paramdef>
+ <paramdef>size_t <parameter>size_b</parameter></paramdef>
+ <paramdef>bool <parameter>committed</parameter></paramdef>
+ <paramdef>unsigned <parameter>arena_ind</parameter></paramdef>
+ </funcprototype></funcsynopsis>
+ <literallayout></literallayout>
+ <para>A chunk merge function conforms to the <type>chunk_merge_t</type>
+ type and optionally merges adjacent chunks,
+ <parameter>chunk_a</parameter> of given <parameter>size_a</parameter>
+ and <parameter>chunk_b</parameter> of given
+ <parameter>size_b</parameter> into one contiguous chunk, operating on
+ <parameter>committed</parameter>/decommitted memory as indicated, on
+ behalf of arena <parameter>arena_ind</parameter>, returning false upon
+ success. If the function returns true, this indicates that the chunks
+ remain distinct mappings and therefore should continue to be operated on
+ independently.</para>
+ </listitem>
</varlistentry>
<varlistentry id="arenas.narenas">
initialized.</para></listitem>
</varlistentry>
+ <varlistentry id="arenas.lg_dirty_mult">
+ <term>
+ <mallctl>arenas.lg_dirty_mult</mallctl>
+ (<type>ssize_t</type>)
+ <literal>rw</literal>
+ </term>
+ <listitem><para>Current default per-arena minimum ratio (log base 2) of
+ active to dirty pages, used to initialize <link
+ linkend="arena.i.lg_dirty_mult"><mallctl>arena.<i>.lg_dirty_mult</mallctl></link>
+ during arena creation. See <link
+ linkend="opt.lg_dirty_mult"><mallctl>opt.lg_dirty_mult</mallctl></link>
+ for additional information.</para></listitem>
+ </varlistentry>
+
+ <varlistentry id="arenas.decay_time">
+ <term>
+ <mallctl>arenas.decay_time</mallctl>
+ (<type>ssize_t</type>)
+ <literal>rw</literal>
+ </term>
+ <listitem><para>Current default per-arena approximate time in seconds
+ from the creation of a set of unused dirty pages until an equivalent set
+ of unused dirty pages is purged and/or reused, used to initialize <link
+ linkend="arena.i.decay_time"><mallctl>arena.<i>.decay_time</mallctl></link>
+ during arena creation. See <link
+ linkend="opt.decay_time"><mallctl>opt.decay_time</mallctl></link> for
+ additional information.</para></listitem>
+ </varlistentry>
+
<varlistentry id="arenas.quantum">
<term>
<mallctl>arenas.quantum</mallctl>
<varlistentry id="arenas.nlruns">
<term>
<mallctl>arenas.nlruns</mallctl>
- (<type>size_t</type>)
+ (<type>unsigned</type>)
<literal>r-</literal>
</term>
<listitem><para>Total number of large size classes.</para></listitem>
class.</para></listitem>
</varlistentry>
+ <varlistentry id="arenas.nhchunks">
+ <term>
+ <mallctl>arenas.nhchunks</mallctl>
+ (<type>unsigned</type>)
+ <literal>r-</literal>
+ </term>
+ <listitem><para>Total number of huge size classes.</para></listitem>
+ </varlistentry>
+
+ <varlistentry id="arenas.hchunk.i.size">
+ <term>
+ <mallctl>arenas.hchunk.<i>.size</mallctl>
+ (<type>size_t</type>)
+ <literal>r-</literal>
+ </term>
+ <listitem><para>Maximum size supported by this huge size
+ class.</para></listitem>
+ </varlistentry>
+
<varlistentry id="arenas.extend">
<term>
<mallctl>arenas.extend</mallctl>
option.</para></listitem>
</varlistentry>
+ <varlistentry id="prof.gdump">
+ <term>
+ <mallctl>prof.gdump</mallctl>
+ (<type>bool</type>)
+ <literal>rw</literal>
+ [<option>--enable-prof</option>]
+ </term>
+ <listitem><para>When enabled, trigger a memory profile dump every time
+ the total virtual memory exceeds the previous maximum. Profiles are
+ dumped to files named according to the pattern
+ <filename><prefix>.<pid>.<seq>.u<useq>.heap</filename>,
+ where <literal><prefix></literal> is controlled by the <link
+ linkend="opt.prof_prefix"><mallctl>opt.prof_prefix</mallctl></link>
+ option.</para></listitem>
+ </varlistentry>
+
<varlistentry id="prof.reset">
<term>
<mallctl>prof.reset</mallctl>
</term>
<listitem><para>Pointer to a counter that contains an approximate count
of the current number of bytes in active pages. The estimate may be
- high, but never low, because each arena rounds up to the nearest
- multiple of the chunk size when computing its contribution to the
- counter. Note that the <link
+ high, but never low, because each arena rounds up when computing its
+ contribution to the counter. Note that the <link
linkend="epoch"><mallctl>epoch</mallctl></link> mallctl has no bearing
on this counter. Furthermore, counter consistency is maintained via
atomic operations, so it is necessary to use an atomic operation in
equal to <link
linkend="stats.allocated"><mallctl>stats.allocated</mallctl></link>.
This does not include <link linkend="stats.arenas.i.pdirty">
- <mallctl>stats.arenas.<i>.pdirty</mallctl></link> and pages
+ <mallctl>stats.arenas.<i>.pdirty</mallctl></link>, nor pages
entirely devoted to allocator metadata.</para></listitem>
</varlistentry>
- <varlistentry id="stats.mapped">
+ <varlistentry id="stats.metadata">
<term>
- <mallctl>stats.mapped</mallctl>
+ <mallctl>stats.metadata</mallctl>
(<type>size_t</type>)
<literal>r-</literal>
[<option>--enable-stats</option>]
</term>
- <listitem><para>Total number of bytes in chunks mapped on behalf of the
- application. This is a multiple of the chunk size, and is at least as
- large as <link
- linkend="stats.active"><mallctl>stats.active</mallctl></link>. This
- does not include inactive chunks.</para></listitem>
+ <listitem><para>Total number of bytes dedicated to metadata, which
+ comprise base allocations used for bootstrap-sensitive internal
+ allocator data structures, arena chunk headers (see <link
+ linkend="stats.arenas.i.metadata.mapped"><mallctl>stats.arenas.<i>.metadata.mapped</mallctl></link>),
+ and internal allocations (see <link
+ linkend="stats.arenas.i.metadata.allocated"><mallctl>stats.arenas.<i>.metadata.allocated</mallctl></link>).</para></listitem>
</varlistentry>
- <varlistentry id="stats.chunks.current">
+ <varlistentry id="stats.resident">
<term>
- <mallctl>stats.chunks.current</mallctl>
+ <mallctl>stats.resident</mallctl>
(<type>size_t</type>)
<literal>r-</literal>
[<option>--enable-stats</option>]
</term>
- <listitem><para>Total number of chunks actively mapped on behalf of the
- application. This does not include inactive chunks.
- </para></listitem>
+ <listitem><para>Maximum number of bytes in physically resident data
+ pages mapped by the allocator, comprising all pages dedicated to
+ allocator metadata, pages backing active allocations, and unused dirty
+ pages. This is a maximum rather than precise because pages may not
+ actually be physically resident if they correspond to demand-zeroed
+ virtual memory that has not yet been touched. This is a multiple of the
+ page size, and is larger than <link
+ linkend="stats.active"><mallctl>stats.active</mallctl></link>.</para></listitem>
</varlistentry>
- <varlistentry id="stats.chunks.total">
- <term>
- <mallctl>stats.chunks.total</mallctl>
- (<type>uint64_t</type>)
- <literal>r-</literal>
- [<option>--enable-stats</option>]
- </term>
- <listitem><para>Cumulative number of chunks allocated.</para></listitem>
- </varlistentry>
-
- <varlistentry id="stats.chunks.high">
+ <varlistentry id="stats.mapped">
<term>
- <mallctl>stats.chunks.high</mallctl>
+ <mallctl>stats.mapped</mallctl>
(<type>size_t</type>)
<literal>r-</literal>
[<option>--enable-stats</option>]
</term>
- <listitem><para>Maximum number of active chunks at any time thus far.
- </para></listitem>
+ <listitem><para>Total number of bytes in active chunks mapped by the
+ allocator. This is a multiple of the chunk size, and is larger than
+ <link linkend="stats.active"><mallctl>stats.active</mallctl></link>.
+ This does not include inactive chunks, even those that contain unused
+ dirty pages, which means that there is no strict ordering between this
+ and <link
+ linkend="stats.resident"><mallctl>stats.resident</mallctl></link>.</para></listitem>
</varlistentry>
<varlistentry id="stats.arenas.i.dss">
</para></listitem>
</varlistentry>
+ <varlistentry id="stats.arenas.i.lg_dirty_mult">
+ <term>
+ <mallctl>stats.arenas.<i>.lg_dirty_mult</mallctl>
+ (<type>ssize_t</type>)
+ <literal>r-</literal>
+ </term>
+ <listitem><para>Minimum ratio (log base 2) of active to dirty pages.
+ See <link
+ linkend="opt.lg_dirty_mult"><mallctl>opt.lg_dirty_mult</mallctl></link>
+ for details.</para></listitem>
+ </varlistentry>
+
+ <varlistentry id="stats.arenas.i.decay_time">
+ <term>
+ <mallctl>stats.arenas.<i>.decay_time</mallctl>
+ (<type>ssize_t</type>)
+ <literal>r-</literal>
+ </term>
+ <listitem><para>Approximate time in seconds from the creation of a set
+ of unused dirty pages until an equivalent set of unused dirty pages is
+ purged and/or reused. See <link
+ linkend="opt.decay_time"><mallctl>opt.decay_time</mallctl></link>
+ for details.</para></listitem>
+ </varlistentry>
+
<varlistentry id="stats.arenas.i.nthreads">
<term>
<mallctl>stats.arenas.<i>.nthreads</mallctl>
<listitem><para>Number of mapped bytes.</para></listitem>
</varlistentry>
+ <varlistentry id="stats.arenas.i.metadata.mapped">
+ <term>
+ <mallctl>stats.arenas.<i>.metadata.mapped</mallctl>
+ (<type>size_t</type>)
+ <literal>r-</literal>
+ [<option>--enable-stats</option>]
+ </term>
+ <listitem><para>Number of mapped bytes in arena chunk headers, which
+ track the states of the non-metadata pages.</para></listitem>
+ </varlistentry>
+
+ <varlistentry id="stats.arenas.i.metadata.allocated">
+ <term>
+ <mallctl>stats.arenas.<i>.metadata.allocated</mallctl>
+ (<type>size_t</type>)
+ <literal>r-</literal>
+ [<option>--enable-stats</option>]
+ </term>
+ <listitem><para>Number of bytes dedicated to internal allocations.
+ Internal allocations differ from application-originated allocations in
+ that they are for internal use, and that they are omitted from heap
+ profiles. This statistic is reported separately from <link
+ linkend="stats.metadata"><mallctl>stats.metadata</mallctl></link> and
+ <link
+ linkend="stats.arenas.i.metadata.mapped"><mallctl>stats.arenas.<i>.metadata.mapped</mallctl></link>
+ because it overlaps with e.g. the <link
+ linkend="stats.allocated"><mallctl>stats.allocated</mallctl></link> and
+ <link linkend="stats.active"><mallctl>stats.active</mallctl></link>
+ statistics, whereas the other metadata statistics do
+ not.</para></listitem>
+ </varlistentry>
+
<varlistentry id="stats.arenas.i.npurge">
<term>
<mallctl>stats.arenas.<i>.npurge</mallctl>
</para></listitem>
</varlistentry>
- <varlistentry id="stats.arenas.i.bins.j.allocated">
- <term>
- <mallctl>stats.arenas.<i>.bins.<j>.allocated</mallctl>
- (<type>size_t</type>)
- <literal>r-</literal>
- [<option>--enable-stats</option>]
- </term>
- <listitem><para>Current number of bytes allocated by
- bin.</para></listitem>
- </varlistentry>
-
<varlistentry id="stats.arenas.i.bins.j.nmalloc">
<term>
<mallctl>stats.arenas.<i>.bins.<j>.nmalloc</mallctl>
requests.</para></listitem>
</varlistentry>
+ <varlistentry id="stats.arenas.i.bins.j.curregs">
+ <term>
+ <mallctl>stats.arenas.<i>.bins.<j>.curregs</mallctl>
+ (<type>size_t</type>)
+ <literal>r-</literal>
+ [<option>--enable-stats</option>]
+ </term>
+ <listitem><para>Current number of regions for this size
+ class.</para></listitem>
+ </varlistentry>
+
<varlistentry id="stats.arenas.i.bins.j.nfills">
<term>
<mallctl>stats.arenas.<i>.bins.<j>.nfills</mallctl>
<listitem><para>Current number of runs for this size class.
</para></listitem>
</varlistentry>
+
+ <varlistentry id="stats.arenas.i.hchunks.j.nmalloc">
+ <term>
+ <mallctl>stats.arenas.<i>.hchunks.<j>.nmalloc</mallctl>
+ (<type>uint64_t</type>)
+ <literal>r-</literal>
+ [<option>--enable-stats</option>]
+ </term>
+ <listitem><para>Cumulative number of allocation requests for this size
+ class served directly by the arena.</para></listitem>
+ </varlistentry>
+
+ <varlistentry id="stats.arenas.i.hchunks.j.ndalloc">
+ <term>
+ <mallctl>stats.arenas.<i>.hchunks.<j>.ndalloc</mallctl>
+ (<type>uint64_t</type>)
+ <literal>r-</literal>
+ [<option>--enable-stats</option>]
+ </term>
+ <listitem><para>Cumulative number of deallocation requests for this
+ size class served directly by the arena.</para></listitem>
+ </varlistentry>
+
+ <varlistentry id="stats.arenas.i.hchunks.j.nrequests">
+ <term>
+ <mallctl>stats.arenas.<i>.hchunks.<j>.nrequests</mallctl>
+ (<type>uint64_t</type>)
+ <literal>r-</literal>
+ [<option>--enable-stats</option>]
+ </term>
+ <listitem><para>Cumulative number of allocation requests for this size
+ class.</para></listitem>
+ </varlistentry>
+
+ <varlistentry id="stats.arenas.i.hchunks.j.curhchunks">
+ <term>
+ <mallctl>stats.arenas.<i>.hchunks.<j>.curhchunks</mallctl>
+ (<type>size_t</type>)
+ <literal>r-</literal>
+ [<option>--enable-stats</option>]
+ </term>
+ <listitem><para>Current number of huge allocations for this size class.
+ </para></listitem>
+ </varlistentry>
</variablelist>
</refsect1>
+ <refsect1 id="heap_profile_format">
+ <title>HEAP PROFILE FORMAT</title>
+ <para>Although the heap profiling functionality was originally designed to
+ be compatible with the
+ <command>pprof</command> command that is developed as part of the <ulink
+ url="http://code.google.com/p/gperftools/">gperftools
+ package</ulink>, the addition of per thread heap profiling functionality
+ required a different heap profile format. The <command>jeprof</command>
+ command is derived from <command>pprof</command>, with enhancements to
+ support the heap profile format described here.</para>
+
+ <para>In the following hypothetical heap profile, <constant>[...]</constant>
+ indicates elision for the sake of compactness. <programlisting><![CDATA[
+heap_v2/524288
+ t*: 28106: 56637512 [0: 0]
+ [...]
+ t3: 352: 16777344 [0: 0]
+ [...]
+ t99: 17754: 29341640 [0: 0]
+ [...]
+@ 0x5f86da8 0x5f5a1dc [...] 0x29e4d4e 0xa200316 0xabb2988 [...]
+ t*: 13: 6688 [0: 0]
+ t3: 12: 6496 [0: ]
+ t99: 1: 192 [0: 0]
+[...]
+
+MAPPED_LIBRARIES:
+[...]]]></programlisting> The following matches the above heap profile, but most
+tokens are replaced with <constant><description></constant> to indicate
+descriptions of the corresponding fields. <programlisting><![CDATA[
+<heap_profile_format_version>/<mean_sample_interval>
+ <aggregate>: <curobjs>: <curbytes> [<cumobjs>: <cumbytes>]
+ [...]
+ <thread_3_aggregate>: <curobjs>: <curbytes>[<cumobjs>: <cumbytes>]
+ [...]
+ <thread_99_aggregate>: <curobjs>: <curbytes>[<cumobjs>: <cumbytes>]
+ [...]
+@ <top_frame> <frame> [...] <frame> <frame> <frame> [...]
+ <backtrace_aggregate>: <curobjs>: <curbytes> [<cumobjs>: <cumbytes>]
+ <backtrace_thread_3>: <curobjs>: <curbytes> [<cumobjs>: <cumbytes>]
+ <backtrace_thread_99>: <curobjs>: <curbytes> [<cumobjs>: <cumbytes>]
+[...]
+
+MAPPED_LIBRARIES:
+</proc/<pid>/maps>]]></programlisting></para>
+ </refsect1>
+
<refsect1 id="debugging_malloc_problems">
<title>DEBUGGING MALLOC PROBLEMS</title>
<para>When debugging, it is a good idea to configure/build jemalloc with