server-skynet-source-3rd-jemalloc/jemalloc/doc/jemalloc.3.in

.\" Copyright (c) 2006-2010 Jason Evans <jasone@canonware.com>.
.\" All rights reserved.
.\" Copyright (c) 2009 Facebook, Inc.  All rights reserved.
.\"
.\" See COPYING for licensing terms provided by the above copyright holders.
.\"
.\" Copyright (c) 1980, 1991, 1993
.\"	The Regents of the University of California.  All rights reserved.
.\"
.\" This code is derived from software contributed to Berkeley by
.\" the American National Standards Committee X3, on Information
.\" Processing Systems.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\"    notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\"    notice, this list of conditions and the following disclaimer in the
.\"    documentation and/or other materials provided with the distribution.
.\" 3. Neither the name of the University nor the names of its contributors
.\"    may be used to endorse or promote products derived from this software
.\"    without specific prior written permission.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\"     @(#)malloc.3	8.1 (Berkeley) 6/4/93
.\" $FreeBSD: head/lib/libc/stdlib/malloc.3 182225 2008-08-27 02:00:53Z jasone $
.\"
.Dd September 30, 2010
.Dt JEMALLOC 3
.Os
.Sh NAME
.Nm @jemalloc_prefix@malloc ,
.Nm @jemalloc_prefix@calloc ,
.Nm @jemalloc_prefix@posix_memalign ,
.Nm @jemalloc_prefix@realloc ,
.Nm @jemalloc_prefix@free ,
.Nm @jemalloc_prefix@malloc_usable_size ,
.Nm @jemalloc_prefix@malloc_stats_print ,
.Nm @jemalloc_prefix@mallctl ,
.Nm @jemalloc_prefix@mallctlnametomib ,
.Nm @jemalloc_prefix@mallctlbymib ,
.Nm @jemalloc_prefix@allocm ,
.Nm @jemalloc_prefix@rallocm ,
.Nm @jemalloc_prefix@sallocm ,
.Nm @jemalloc_prefix@dallocm
.Nd general purpose memory allocation functions
.Sh LIBRARY
.Sy libjemalloc@install_suffix@
.Sh SYNOPSIS
.In stdlib.h
.In jemalloc/jemalloc@install_suffix@.h
.Ss Standard API
.Ft void *
.Fn @jemalloc_prefix@malloc "size_t size"
.Ft void *
.Fn @jemalloc_prefix@calloc "size_t number" "size_t size"
.Ft int
.Fn @jemalloc_prefix@posix_memalign "void **ptr" "size_t alignment" "size_t size"
.Ft void *
.Fn @jemalloc_prefix@realloc "void *ptr" "size_t size"
.Ft void
.Fn @jemalloc_prefix@free "void *ptr"
.Ss Non-standard API
.Ft size_t
.Fn @jemalloc_prefix@malloc_usable_size "const void *ptr"
.Ft void
.Fn @jemalloc_prefix@malloc_stats_print "void (*write_cb)(void *" "const char *)" "void *cbopaque" "const char *opts"
.Ft int
.Fn @jemalloc_prefix@mallctl "const char *name" "void *oldp" "size_t *oldlenp" "void *newp" "size_t newlen"
.Ft int
.Fn @jemalloc_prefix@mallctlnametomib "const char *name" "int *mibp" "size_t *miblenp"
.Ft int
.Fn @jemalloc_prefix@mallctlbymib "const size_t *mib" "size_t miblen" "void *oldp" "size_t *oldlenp" "void *newp" "size_t newlen"
.Ft const char *
.Va @jemalloc_prefix@malloc_options ;
.Ft void
.Fn \*(lp*@jemalloc_prefix@malloc_message\*(rp "void *cbopaque" "const char *s"
.Ss Experimental API
.Ft int
.Fn @jemalloc_prefix@allocm "void **ptr" "size_t *rsize" "size_t size" "int flags"
.Ft int
.Fn @jemalloc_prefix@rallocm "void **ptr" "size_t *rsize" "size_t size" "size_t extra" "int flags"
.Ft int
.Fn @jemalloc_prefix@sallocm "const void *ptr" "size_t *rsize" "int flags"
.Ft int
.Fn @jemalloc_prefix@dallocm "void *ptr" "int flags"
.Sh DESCRIPTION
.Ss Standard API
The
.Fn @jemalloc_prefix@malloc
function allocates
.Fa size
bytes of uninitialized memory.
The allocated space is suitably aligned
@roff_tiny@(after possible pointer coercion)
for storage of any type of object.
.Pp
The
.Fn @jemalloc_prefix@calloc
function allocates space for
.Fa number
objects,
each
.Fa size
bytes in length.
The result is identical to calling
.Fn @jemalloc_prefix@malloc
with an argument of
.Dq "number * size" ,
with the exception that the allocated memory is explicitly initialized
to zero bytes.
.Pp
The
.Fn @jemalloc_prefix@posix_memalign
function allocates
.Fa size
bytes of memory such that the allocation's base address is an even multiple of
.Fa alignment ,
and returns the allocation in the value pointed to by
.Fa ptr .
The requested
.Fa alignment
must be a power of 2 at least as large as
.Fn sizeof "void *" .
.Pp
The
.Fn @jemalloc_prefix@realloc
function changes the size of the previously allocated memory referenced by
.Fa ptr
to
.Fa size
bytes.
The contents of the memory are unchanged up to the lesser of the new and
old sizes.
If the new size is larger,
the contents of the newly allocated portion of the memory are undefined.
Upon success, the memory referenced by
.Fa ptr
is freed and a pointer to the newly allocated memory is returned.
Note that
.Fn @jemalloc_prefix@realloc
may move the memory allocation, resulting in a different return value than
.Fa ptr .
If
.Fa ptr
is
.Dv NULL ,
the
.Fn @jemalloc_prefix@realloc
function behaves identically to
.Fn @jemalloc_prefix@malloc
for the specified size.
.Pp
The
.Fn @jemalloc_prefix@free
function causes the allocated memory referenced by
.Fa ptr
to be made available for future allocations.
If
.Fa ptr
is
.Dv NULL ,
no action occurs.
.Ss Non-standard API
The
.Fn @jemalloc_prefix@malloc_usable_size
function returns the usable size of the allocation pointed to by
.Fa ptr .
The return value may be larger than the size that was requested during
allocation.
The
.Fn @jemalloc_prefix@malloc_usable_size
function is not a mechanism for in-place
.Fn @jemalloc_prefix@realloc ;
rather it is provided solely as a tool for introspection purposes.
Any discrepancy between the requested allocation size and the size reported by
.Fn @jemalloc_prefix@malloc_usable_size
should not be depended on, since such behavior is entirely
implementation-dependent.
.Pp
The
.Fn @jemalloc_prefix@malloc_stats_print
function writes human-readable summary statistics via the
.Fa write_cb
callback function pointer and
.Fa cbopaque
data passed to
.Fn write_cb ,
or
.Fn @jemalloc_prefix@malloc_message
if
.Fa write_cb
is
.Dv NULL .
This function can be called repeatedly.
General information that never changes
during execution can be omitted by specifying
.Dq g
as a character within the
.Fa opts
string.
Note that
.Fn @jemalloc_prefix@malloc_message
uses the
.Fn @jemalloc_prefix@mallctl*
functions internally, so inconsistent statistics can be reported if multiple
threads use these functions simultaneously.
@roff_stats@.Dq m
@roff_stats@and
@roff_stats@.Dq a
@roff_stats@can be specified to omit merged arena and per arena statistics,
@roff_stats@respectively.
@roff_stats@.Dq b
@roff_stats@and
@roff_stats@.Dq l
@roff_stats@can be specified to omit per size class statistics for bins and
@roff_stats@large objects, respectively.
Unrecognized characters are silently ignored.
@roff_tcache@Note that thread caching may prevent some statistics from being
@roff_tcache@completely up to date, since extra locking would be required to
@roff_tcache@merge counters that track thread cache operations.
.Pp
The
.Fn @jemalloc_prefix@mallctl
function provides a general interface for introspecting the memory allocator,
as well as setting modifiable parameters and triggering actions.
The period-separated
.Fa name
argument specifies a location in a tree-structured namespace; see the
.Sx "MALLCTL NAMESPACE"
section for documentation on the tree contents.
To read a value, pass a pointer via
.Fa oldp
to adequate space to contain the value, and a pointer to its length via
.Fa oldlenp ;
otherwise pass
.Dv NULL
and
.Dv NULL .
Similarly, to write a value, pass a pointer to the value via
.Fa newp ,
and its length via
.Fa newlen ;
otherwise pass
.Dv NULL
and 0.
.Pp
The
.Fn @jemalloc_prefix@mallctlnametomib
function provides a way to avoid repeated name lookups for applications that
repeatedly query the same portion of the namespace, by translating a name to a
.Dq Management Information Base
(MIB) that can be passed repeatedly to
.Fn @jemalloc_prefix@mallctlbymib .
Upon successful return from
.Fn @jemalloc_prefix@mallctlnametomib ,
.Fa mibp
contains an array of
.Fa *miblenp
integers, where
.Fa *miblenp
is the lesser of the number of components in
.Fa name
and the input value of
.Fa *miblenp .
Thus it is possible to pass a
.Fa *miblenp
that is smaller than the number of period-separated name components, which
results in a partial MIB that can be used as the basis for constructing a
complete MIB.
For name components that are integers (e.g. the 2 in
.Qq arenas.bin.2.size ) ,
the corresponding MIB component will always be that integer.
Therefore, it is legitimate to construct code like the following:
.Pp
.Bd -literal -offset indent -compact
unsigned nbins, i;
int mib[4];
size_t len, miblen;

len = sizeof(nbins);
@jemalloc_prefix@mallctl("arenas.nbins", &nbins, &len, NULL, 0);

miblen = 4;
@jemalloc_prefix@mallnametomib("arenas.bin.0.size", mib, &miblen);
for (i = 0; i < nbins; i++) {
	size_t bin_size;

	mib[2] = i;
	len = sizeof(bin_size);
	@jemalloc_prefix@mallctlbymib(mib, miblen, &bin_size, &len, NULL, 0);
	/* Do something with bin_size... */
}
.Ed
.Ss Experimental API
The experimental API is subject to change or removal without regard for
backward compatibility.
.Pp
The
.Fn @jemalloc_prefix@allocm ,
.Fn @jemalloc_prefix@rallocm ,
.Fn @jemalloc_prefix@sallocm ,
and
.Fn @jemalloc_prefix@dallocm
functions all have a
.Fa flags
argument that can be used to specify options.
The functions only check the options that are contextually relevant.
Use bitwise or (|) operations to specify one or more of the following:
.Bl -tag -width ".Dv ALLOCM_LG_ALIGN(la)"
.It ALLOCM_LG_ALIGN(la)
Align the memory allocation to start at an address that is a multiple of
(1 <<
.Fa la ) .
This macro does not validate that
.Fa la
is within the valid range.
.It ALLOCM_ALIGN(a)
Align the memory allocation to start at an address that is a multiple of
.Fa a ,
where
.Fa a
is a power of two.
This macro does not validate that
.Fa a
is a power of 2.
.It ALLOCM_ZERO
Initialize newly allocated memory to contain zero bytes.
In the growing reallocation case, the real size prior to reallocation defines
the boundary between untouched bytes and those that are initialized to contain
zero bytes.
If this option is absent, newly allocated memory is uninitialized.
.It ALLOCM_NO_MOVE
For reallocation, fail rather than moving the object.
This constraint can apply to both growth and shrinkage.
.El
.Pp
The
.Fn @jemalloc_prefix@allocm
function allocates at least
.Fa size
bytes of memory, sets
.Fa *ptr
to the base address of the allocation, and sets
.Fa *rsize
to the real size of the allocation if
.Fa rsize
is not
.Dv NULL .
.Pp
The
.Fn @jemalloc_prefix@rallocm
function resizes the allocation at
.Fa *ptr
to be at least
.Fa size
bytes, sets
.Fa *ptr
to the base address of the allocation if it moved, and sets
.Fa *rsize
to the real size of the allocation if
.Fa rsize
is not
.Dv NULL .
If
.Fa extra
is non-zero, an attempt is made to resize the allocation to be at least
.Fa ( size
+
.Fa extra )
bytes, though an inability to allocate the extra byte(s) will not by itself
result in failure.
Behavior is undefined if
.Fa ( size
+
.Fa extra
>
.Dv SIZE_T_MAX ) .
.Pp
The
.Fn @jemalloc_prefix@sallocm
function sets
.Fa *rsize
to the real size of the allocation.
.Pp
The
.Fn @jemalloc_prefix@dallocm
function causes the memory referenced by
.Fa ptr
to be made available for future allocations.
.Sh TUNING
Once, when the first call is made to one of these memory allocation
routines, various flags will be set or reset, which affects the
workings of this allocator implementation.
.Pp
The
.Dq name
of the file referenced by the symbolic link named
.Pa /etc/jemalloc.conf ,
the value of the environment variable
.Ev JEMALLOC_OPTIONS ,
and the string pointed to by the global variable
.Va @jemalloc_prefix@malloc_options
will be interpreted, in that order, from left to right as flags.
.Pp
Each flag is a single letter, optionally prefixed by a non-negative base 10
integer repetition count.
For example,
.Dq 3N
is equivalent to
.Dq NNN .
Some flags control parameter magnitudes, where uppercase increases the
magnitude, and lowercase decreases the magnitude.
Other flags control boolean parameters, where uppercase indicates that a
behavior is set, or on, and lowercase means that a behavior is not set, or off.
.Bl -tag -width indent
.It A
All warnings (except for the warning about unknown
flags being set) become fatal.
The process will call
.Xr abort 3
in these cases.
@roff_prof@.It B
@roff_prof@Double/halve the maximum backtrace depth when profiling memory
@roff_prof@allocation activity.
@roff_prof@The default is 4.
.It C
Double/halve the size of the maximum size class that is a multiple of the
cacheline size (64).
Above this size, subpage spacing (256 bytes) is used for size classes.
The default value is 512 bytes.
.It D
Halve/double the per-arena minimum ratio of active to dirty pages.
Some dirty unused pages may be allowed to accumulate, within the limit set by
the ratio (or one chunk worth of dirty pages, whichever is greater), before
informing the kernel about some of those pages via
.Xr madvise 2 .
This 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 32:1;
.Ev JEMALLOC_OPTIONS=6D
will disable dirty page purging.
@roff_prof@.It E
@roff_prof@Activate/deactivate profiling.
@roff_prof@This is a secondary control mechanism that makes it possible to
@roff_prof@start the application with profiling enabled (see the
@roff_prof@.Dq F
@roff_prof@option) but inactive, then toggle profiling at any time during
@roff_prof@program execution with the
@roff_prof@.Dq prof.active
@roff_prof@mallctl.
@roff_prof@This option is enabled by default.
@roff_prof@.It F
@roff_prof@Profile memory allocation activity, and use an
@roff_prof@.Xr atexit 3
@roff_prof@function to dump final memory usage to a file named according to
@roff_prof@the pattern
@roff_prof@.Pa <prefix>.<pid>.<seq>.f.heap ,
@roff_prof@where
@roff_prof@.Pa <prefix>
@roff_prof@is controlled by the
@roff_prof@JEMALLOC_PROF_PREFIX
@roff_prof@environment variable.
@roff_prof@See the
@roff_prof@.Dq B
@roff_prof@option for backtrace depth control.
@roff_prof@See the
@roff_prof@.Dq E
@roff_prof@option for on-the-fly activation/deactivation.
@roff_prof@See the
@roff_prof@.Dq S
@roff_prof@option for probabilistic sampling control.
@roff_prof@See the
@roff_prof@.Dq I
@roff_prof@option for information on interval-triggered profile dumping, and the
@roff_prof@.Dq U
@roff_prof@option for information on high-water-triggered profile dumping.
@roff_prof@Profile output is compatible with the included pprof Perl script,
@roff_prof@which originates from the google-perftools package
@roff_prof@(http://code.google.com/p/google-perftools/).
@roff_tcache@.It G
@roff_tcache@Double/halve the approximate interval (counted in terms of
@roff_tcache@thread-specific cache allocation/deallocation events) between full
@roff_tcache@thread-specific cache garbage collection sweeps.
@roff_tcache@Garbage collection is actually performed incrementally, one size
@roff_tcache@class at a time, in order to avoid large collection pauses.
@roff_tcache@The default sweep interval is 8192;
@roff_tcache@.Ev JEMALLOC_OPTIONS=14g
@roff_tcache@will disable garbage collection.
@roff_tcache@.It H
@roff_tcache@Enable/disable thread-specific caching.
@roff_tcache@When there are multiple threads, each thread uses a
@roff_tcache@thread-specific cache for objects up to a certain size.
@roff_tcache@Thread-specific caching allows many allocations to be satisfied
@roff_tcache@without performing any thread synchronization, at the cost of
@roff_tcache@increased memory use.
@roff_tcache@See the
@roff_tcache@.Dq G
@roff_tcache@and
@roff_tcache@.Dq M
@roff_tcache@options for related tuning information.
@roff_tcache@This option is enabled by default.
@roff_prof@.It I
@roff_prof@Double/halve the average interval between memory profile dumps, as
@roff_prof@measured in bytes of allocation activity.
@roff_prof@The actual interval between dumps may be sporadic because
@roff_prof@decentralized allocation counters are used to avoid synchronization
@roff_prof@bottlenecks.
@roff_prof@Profiles are dumped to files named according to the pattern
@roff_prof@.Pa <prefix>.<pid>.<seq>.i<iseq>.heap ,
@roff_prof@where
@roff_prof@.Pa <prefix>
@roff_prof@is controlled by the
@roff_prof@JEMALLOC_PROF_PREFIX
@roff_prof@environment variable.
@roff_prof@By default, interval-triggered profile dumping is disabled.
@roff_prof@This is internally encoded as (1 << -1), and each
@roff_prof@.Dq I
@roff_prof@that is specified increments the shift amount.
@roff_prof@Therefore, e.g.
@roff_prof@.Ev JEMALLOC_OPTIONS=31I
@roff_prof@specifies a dump interval of 1 GiB.
@roff_fill@.It J
@roff_fill@Each byte of new memory allocated by
@roff_fill@.Fn @jemalloc_prefix@malloc
@roff_fill@or
@roff_fill@.Fn @jemalloc_prefix@realloc
@roff_fill@will be initialized to 0xa5.
@roff_fill@All memory returned by
@roff_fill@.Fn @jemalloc_prefix@free
@roff_fill@or
@roff_fill@.Fn @jemalloc_prefix@realloc
@roff_fill@will be initialized to 0x5a.
@roff_fill@This is intended for debugging and will impact performance
@roff_fill@negatively.
.It K
Double/halve the virtual memory chunk size.
The default chunk size is 4 MiB.
@roff_prof@.It L
@roff_prof@Use an
@roff_prof@.Xr atexit 3
@roff_prof@function to report memory leaks.
@roff_prof@See the
@roff_prof@.Dq B
@roff_prof@option for backtrace depth control.
@roff_prof@See the
@roff_prof@.Dq F option for information on analyzing heap profile output.
@roff_prof@This option is disabled by default.
@roff_tcache@.It M
@roff_tcache@Double/halve the maximum size class to cache.
@roff_tcache@At a minimum, all small size classes are cached, and at a maximum
@roff_tcache@all large size classes are cached.
@roff_tcache@The default maximum is 32 KiB.
.It N
Double/halve the number of arenas.
The default number of arenas is four times the number of CPUs, or one if there
is a single CPU.
@roff_swap@.It O
@roff_swap@Over-commit memory as a side effect of using anonymous
@roff_swap@.Xr mmap 2
@roff_swap@@roff_dss@ and
@roff_swap@@roff_dss@.Xr sbrk 2
@roff_swap@for virtual memory allocation.
@roff_swap@In order for overcommit to be disabled, the
@roff_swap@.Dq swap.fds
@roff_swap@mallctl must have been successfully written to.
@roff_swap@This option is enabled by default.
.It P
The
.Fn @jemalloc_prefix@malloc_stats_print
function is called at program exit via an
.Xr atexit 3
function.
@roff_stats@This has the potential to cause deadlock for a multi-threaded
@roff_stats@process that exits while one or more threads are executing in the
@roff_stats@memory allocation functions.
@roff_stats@Therefore, this option should only be used with care; it is
@roff_stats@primarily intended as a performance tuning aid during application
@roff_stats@development.
.It Q
Double/halve the size of the maximum size class that is a multiple of the
quantum (8 or 16 bytes, depending on architecture).
Above this size, cacheline spacing is used for size classes.
The default value is 128 bytes.
@roff_prof@.It S
@roff_prof@Double/halve the average interval between allocation samples, as
@roff_prof@measured in bytes of allocation activity.
@roff_prof@Increasing the sampling interval decreases profile fidelity, but
@roff_prof@also decreases the computational overhead.
@roff_prof@The default sample interval is one (i.e. all allocations are
@roff_prof@sampled).
@roff_prof@.It U
@roff_prof@Trigger a memory profile dump every time the total virtual memory
@roff_prof@exceeds the previous maximum.
@roff_prof@Profiles are dumped to files named according to the pattern
@roff_prof@.Pa <prefix>.<pid>.<seq>.u<useq>.heap ,
@roff_prof@where
@roff_prof@.Pa <prefix>
@roff_prof@is controlled by the
@roff_prof@JEMALLOC_PROF_PREFIX
@roff_prof@environment variable.
@roff_prof@This option is disabled by default.
@roff_sysv@.It V
@roff_sysv@Attempting to allocate zero bytes will return a
@roff_sysv@.Dv NULL
@roff_sysv@pointer instead of a valid pointer.
@roff_sysv@(The default behavior is to make a minimal allocation and return a
@roff_sysv@pointer to it.)
@roff_sysv@This option is provided for System V compatibility.
@roff_sysv@@roff_xmalloc@This option is incompatible with the
@roff_sysv@@roff_xmalloc@.Dq X
@roff_sysv@@roff_xmalloc@option.
@roff_xmalloc@.It X
@roff_xmalloc@Rather than return failure for any allocation function, display a
@roff_xmalloc@diagnostic message on
@roff_xmalloc@.Dv STDERR_FILENO
@roff_xmalloc@and cause the program to drop core (using
@roff_xmalloc@.Xr abort 3 ) .
@roff_xmalloc@This option should be set at compile time by including the
@roff_xmalloc@following in the source code:
@roff_xmalloc@.Bd -literal -offset indent
@roff_xmalloc@@jemalloc_prefix@malloc_options = "X";
@roff_xmalloc@.Ed
@roff_fill@.It Z
@roff_fill@Each byte of new memory allocated by
@roff_fill@.Fn @jemalloc_prefix@malloc
@roff_fill@or
@roff_fill@.Fn @jemalloc_prefix@realloc
@roff_fill@will be initialized to 0.
@roff_fill@Note that this initialization only happens once for each byte, so
@roff_fill@.Fn @jemalloc_prefix@realloc
@roff_fill@calls do not zero memory that was previously allocated.
@roff_fill@This is intended for debugging and will impact performance
@roff_fill@negatively.
.El
.Pp
@roff_fill@The
@roff_fill@.Dq J
@roff_fill@and
@roff_fill@.Dq Z
@roff_fill@options are intended for testing and debugging.
@roff_fill@An application which changes its behavior when these options are used
@roff_fill@is flawed.
.Sh IMPLEMENTATION NOTES
@roff_dss@Traditionally, allocators have used
@roff_dss@.Xr sbrk 2
@roff_dss@to obtain memory, which is suboptimal for several reasons, including
@roff_dss@race conditions, increased fragmentation, and artificial limitations
@roff_dss@on maximum usable memory.
@roff_dss@This allocator uses both
@roff_dss@.Xr sbrk 2
@roff_dss@and
@roff_dss@.Xr mmap 2 ,
@roff_dss@in that order of preference.
.Pp
This allocator uses multiple arenas in order to reduce lock contention for
threaded programs on multi-processor systems.
This works well with regard to threading scalability, but incurs some costs.
There is a small fixed per-arena overhead, and additionally, arenas manage
memory completely independently of each other, which means a small fixed
increase in overall memory fragmentation.
These overheads are not generally an issue, given the number of arenas normally
used.
Note that using substantially more arenas than the default is not likely to
improve performance, mainly due to reduced cache performance.
However, it may make sense to reduce the number of arenas if an application
does not make much use of the allocation functions.
.Pp
@roff_tcache@In addition to multiple arenas, this allocator supports
@roff_tcache@thread-specific caching for small and large objects, in order to
@roff_tcache@make it possible to completely avoid synchronization for most small
@roff_tcache@allocation requests.
@roff_tcache@Such caching allows very fast allocation in the common case, but it
@roff_tcache@increases memory usage and fragmentation, since a bounded number of
@roff_tcache@objects can remain allocated in each thread cache.
@roff_tcache@.Pp
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.
.Pp
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 by arenas; huge objects are managed
separately in a single data structure that is shared by all threads.
Huge objects are used by applications infrequently enough that this single
data structure is not a scalability issue.
.Pp
Each chunk that is managed by an arena 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.
.Pp
Small objects are managed in groups by page runs.
Each run maintains a bitmap that tracks which regions are in use.
@roff_tiny@Allocation requests that are no more than half the quantum (8 or 16,
@roff_tiny@depending on architecture) are rounded up to the nearest power of
@roff_tiny@two.
Allocation requests that are
@roff_tiny@more than half the quantum, but
no more than the minimum cacheline-multiple size class (see the
.Dq Q
option) are rounded up to the nearest multiple of the
@roff_tiny@quantum.
@roff_no_tiny@quantum (8 or 16, depending on architecture).
Allocation requests that are more than the minimum cacheline-multiple size
class, but no more than the minimum subpage-multiple size class (see the
.Dq C
option) are rounded up to the nearest multiple of the cacheline size (64).
Allocation requests that are more than the minimum subpage-multiple size class,
but no more than the maximum subpage-multiple size class are rounded up to the
nearest multiple of the subpage size (256).
Allocation requests that are more than the maximum subpage-multiple size class,
but small enough to fit in an arena-managed chunk (see the
.Dq K
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.
.Pp
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 follows:
.TS
expand allbox tab(;);
LLR
LLR
^LR
^^R
^^R
^^R
^^R
^LR
^^R
^^R
^^R
^^R
^LR
^^R
^^R
^^R
^^R
LsR
^^R
^^R
^^R
^^R
LsR
^^R.
Category;Subcategory;Size
@roff_tiny@Small;Tiny;8
@roff_no_tiny@Small;Tiny;[disabled]
;Quantum-spaced;16
;;32
;;48
;;...
;;128
;Cacheline-spaced;192
;;256
;;320
;;...
;;512
;Sub-page;760
;;1024
;;1280
;;...
;;3840
Large;4 KiB
;;8 KiB
;;12 KiB
;;...
;;4084 KiB
Huge;4 MiB
;;8 MiB
;;12 MiB
;;...
.TE
.Pp
Allocations are packed tightly together, which can be an issue for
multi-threaded applications.
If you need to assure that allocations do not suffer from cacheline sharing,
round your allocation requests up to the nearest multiple of the cacheline
size, or specify cacheline alignment when allocating.
.Sh MALLCTL NAMESPACE
The following names are defined in the namespace accessible via the
.Fn @jemalloc_prefix@mallctl*
functions.
Value types are specified in parentheses, and their readable/writable statuses
are encoded as rw, r-, -w, or --.
A name element encoded as <i> or <j> indicates an integer component, where the
integer varies from 0 to some upper value that must be determined via
introspection.
@roff_stats@In the case of
@roff_stats@.Dq stats.arenas.<i>.* ,
@roff_stats@<i> equal to
@roff_stats@.Dq arenas.narenas
@roff_stats@can be used to access the summation of statistics from all arenas.
.Bl -ohang
.\"-----------------------------------------------------------------------------
.It Sy "version (const char *) r-"
.Bd -ragged -offset indent -compact
Return the jemalloc version string.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "epoch (uint64_t) rw"
.Bd -ragged -offset indent -compact
If a value is passed in, refresh the data from which the
.Fn @jemalloc_prefix@mallctl*
functions report values, and increment the epoch.
Return the current epoch.
This is useful for detecting whether another thread caused a refresh.
.Ed
.\"-----------------------------------------------------------------------------
@roff_tcache@.It Sy "tcache.flush (void) --"
@roff_tcache@.Bd -ragged -offset indent -compact
@roff_tcache@Flush calling thread's tcache.
@roff_tcache@This interface releases all cached objects and internal data
@roff_tcache@structures associated with the calling thread's thread-specific
@roff_tcache@cache.
@roff_tcache@Ordinarily, this interface need not be called, since automatic
@roff_tcache@periodic incremental garbage collection occurs, and the thread
@roff_tcache@cache is automatically discarded when a thread exits.
@roff_tcache@However, garbage collection is triggered by allocation activity,
@roff_tcache@so it is possible for a thread that stops allocating/deallocating
@roff_tcache@to retain its cache indefinitely, in which case the developer may
@roff_tcache@find manual flushing useful.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "thread.arena (unsigned) rw"
.Bd -ragged -offset indent -compact
Get or set the arena associated with the calling thread.
The arena index must be less than the maximum number of arenas (see the
.Dq arenas.narenas
mallctl).
If the specified arena was not initialized beforehand (see the
.Dq arenas.initialized
mallctl), it will be automatically initialized as a side effect of calling this
interface.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.debug (bool) r-"
.Bd -ragged -offset indent -compact
--enable-debug was specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.dss (bool) r-"
.Bd -ragged -offset indent -compact
--enable-dss was specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.dynamic_page_shift (bool) r-"
.Bd -ragged -offset indent -compact
--enable-dynamic-page-shift was specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.fill (bool) r-"
.Bd -ragged -offset indent -compact
--enable-fill was specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.lazy_lock (bool) r-"
.Bd -ragged -offset indent -compact
--enable-lazy-lock was specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.prof (bool) r-"
.Bd -ragged -offset indent -compact
--enable-prof was specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.prof_libgcc (bool) r-"
.Bd -ragged -offset indent -compact
--disable-prof-libgcc was not specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.prof_libunwind (bool) r-"
.Bd -ragged -offset indent -compact
--enable-prof-libunwind was specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.stats (bool) r-"
.Bd -ragged -offset indent -compact
--enable-stats was specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.swap (bool) r-"
.Bd -ragged -offset indent -compact
--enable-swap was specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.sysv (bool) r-"
.Bd -ragged -offset indent -compact
--enable-sysv was specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.tcache (bool) r-"
.Bd -ragged -offset indent -compact
--disable-tcache was not specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.tiny (bool) r-"
.Bd -ragged -offset indent -compact
--disable-tiny was not specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.tls (bool) r-"
.Bd -ragged -offset indent -compact
--disable-tls was not specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "config.xmalloc (bool) r-"
.Bd -ragged -offset indent -compact
--enable-xmalloc was specified during build configuration.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "opt.abort (bool) r-"
.Bd -ragged -offset indent -compact
See the
.Dq A
option.
.Ed
.\"-----------------------------------------------------------------------------
@roff_fill@.It Sy "opt.junk (bool) r-"
@roff_fill@.Bd -ragged -offset indent -compact
@roff_fill@See the
@roff_fill@.Dq J
@roff_fill@option.
@roff_fill@.Ed
.\"-----------------------------------------------------------------------------
@roff_fill@.It Sy "opt.zero (bool) r-"
@roff_fill@.Bd -ragged -offset indent -compact
@roff_fill@See the
@roff_fill@.Dq Z
@roff_fill@option.
@roff_fill@.Ed
.\"-----------------------------------------------------------------------------
@roff_xmalloc@.It Sy "opt.xmalloc (bool) r-"
@roff_xmalloc@.Bd -ragged -offset indent -compact
@roff_xmalloc@See the
@roff_xmalloc@.Dq X
@roff_xmalloc@option.
@roff_xmalloc@.Ed
.\"-----------------------------------------------------------------------------
@roff_tcache@.It Sy "opt.tcache (bool) r-"
@roff_tcache@.Bd -ragged -offset indent -compact
@roff_tcache@See the
@roff_tcache@.Dq H
@roff_tcache@option.
@roff_tcache@.Ed
.\"-----------------------------------------------------------------------------
@roff_tcache@.It Sy "opt.lg_tcache_gc_sweep (ssize_t) r-"
@roff_tcache@.Bd -ragged -offset indent -compact
@roff_tcache@See the
@roff_tcache@.Dq G
@roff_tcache@option.
@roff_tcache@.Ed
.\"-----------------------------------------------------------------------------
.It Sy "opt.stats_print (bool) r-"
.Bd -ragged -offset indent -compact
See the
.Dq P
option.
.Ed
.\"-----------------------------------------------------------------------------
@roff_prof@.It Sy "opt.prof (bool) r-"
@roff_prof@.Bd -ragged -offset indent -compact
@roff_prof@See the
@roff_prof@.Dq F
@roff_prof@option.
@roff_prof@.Ed
.\"-----------------------------------------------------------------------------
@roff_prof@.It Sy "opt.lg_prof_bt_max (size_t) r-"
@roff_prof@.Bd -ragged -offset indent -compact
@roff_prof@See the
@roff_prof@.Dq B
@roff_prof@option.
@roff_prof@.Ed
.\"-----------------------------------------------------------------------------
@roff_prof@.It Sy "opt.lg_prof_interval (ssize_t) r-"
@roff_prof@.Bd -ragged -offset indent -compact
@roff_prof@See the
@roff_prof@.Dq I
@roff_prof@option.
@roff_prof@.Ed
.\"-----------------------------------------------------------------------------
@roff_prof@.It Sy "opt.prof_udump (bool) r-"
@roff_prof@.Bd -ragged -offset indent -compact
@roff_prof@See the
@roff_prof@.Dq U
@roff_prof@option.
@roff_prof@.Ed
.\"-----------------------------------------------------------------------------
@roff_prof@.It Sy "opt.prof_leak (bool) r-"
@roff_prof@.Bd -ragged -offset indent -compact
@roff_prof@See the
@roff_prof@.Dq L
@roff_prof@option.
@roff_prof@.Ed
.\"-----------------------------------------------------------------------------
.It Sy "opt.lg_qspace_max (size_t) r-"
.Bd -ragged -offset indent -compact
See the
.Dq Q
option.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "opt.lg_cspace_max (size_t) r-"
.Bd -ragged -offset indent -compact
See the
.Dq C
option.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "opt.lg_dirty_mult (ssize_t) r-"
.Bd -ragged -offset indent -compact
See the
.Dq D
option.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "opt.lg_chunk (size_t) r-"
.Bd -ragged -offset indent -compact
See the
.Dq K
option.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "opt.overcommit (bool) r-"
.Bd -ragged -offset indent -compact
See the
.Dq O
option.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.narenas (unsigned) r-"
.Bd -ragged -offset indent -compact
Maximum number of arenas.
See the
.Dq N
option.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.initialized (bool *) r-"
.Bd -ragged -offset indent -compact
An array of arenas.narenas booleans.
Each boolean indicates whether the corresponding arena is initialized.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.quantum (size_t) r-"
.Bd -ragged -offset indent -compact
Quantum size.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.cacheline (size_t) r-"
.Bd -ragged -offset indent -compact
Assumed cacheline size.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.subpage (size_t) r-"
.Bd -ragged -offset indent -compact
Subpage size class interval.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.pagesize (size_t) r-"
.Bd -ragged -offset indent -compact
Page size.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.chunksize (size_t) r-"
.Bd -ragged -offset indent -compact
Chunk size.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.tspace_min (size_t) r-"
.Bd -ragged -offset indent -compact
Minimum tiny size class.
Tiny size classes are powers of two.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.tspace_max (size_t) r-"
.Bd -ragged -offset indent -compact
Maximum tiny size class.
Tiny size classes are powers of two.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.qspace_min (size_t) r-"
.Bd -ragged -offset indent -compact
Minimum quantum-spaced size class.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.qspace_max (size_t) r-"
.Bd -ragged -offset indent -compact
Maximum quantum-spaced size class.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.cspace_min (size_t) r-"
.Bd -ragged -offset indent -compact
Minimum cacheline-spaced size class.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.cspace_max (size_t) r-"
.Bd -ragged -offset indent -compact
Maximum cacheline-spaced size class.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.sspace_min (size_t) r-"
.Bd -ragged -offset indent -compact
Minimum subpage-spaced size class.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.sspace_max (size_t) r-"
.Bd -ragged -offset indent -compact
Maximum subpage-spaced size class.
.Ed
.\"-----------------------------------------------------------------------------
@roff_tcache@.It Sy "arenas.tcache_max (size_t) r-"
@roff_tcache@.Bd -ragged -offset indent -compact
@roff_tcache@Maximum thread-cached size class.
@roff_tcache@.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.ntbins (unsigned) r-"
.Bd -ragged -offset indent -compact
Number of tiny bin size classes.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.nqbins (unsigned) r-"
.Bd -ragged -offset indent -compact
Number of quantum-spaced bin size classes.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.ncbins (unsigned) r-"
.Bd -ragged -offset indent -compact
Number of cacheline-spaced bin size classes.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.nsbins (unsigned) r-"
.Bd -ragged -offset indent -compact
Number of subpage-spaced bin size classes.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.nbins (unsigned) r-"
.Bd -ragged -offset indent -compact
Total number of bin size classes.
.Ed
.\"-----------------------------------------------------------------------------
@roff_tcache@.It Sy "arenas.nhbins (unsigned) r-"
@roff_tcache@.Bd -ragged -offset indent -compact
@roff_tcache@Total number of thread cache bin size classes.
@roff_tcache@.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.bin.<i>.size (size_t) r-"
.Bd -ragged -offset indent -compact
Maximum size supported by size class.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.bin.<i>.nregs (uint32_t) r-"
.Bd -ragged -offset indent -compact
Number of regions per page run.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.bin.<i>.run_size (size_t) r-"
.Bd -ragged -offset indent -compact
Number of bytes per page run.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.nlruns (size_t) r-"
.Bd -ragged -offset indent -compact
Total number of large size classes.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.lrun.<i>.size (size_t) r-"
.Bd -ragged -offset indent -compact
Maximum size supported by this large size class.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "arenas.purge (unsigned) -w"
.Bd -ragged -offset indent -compact
Purge unused dirty pages for the specified arena, or for all arenas if none is
specified.
.Ed
.\"-----------------------------------------------------------------------------
@roff_prof@.It Sy "prof.active (bool) rw"
@roff_prof@.Bd -ragged -offset indent -compact
@roff_prof@Control whether sampling is currently active.
@roff_prof@See the
@roff_prof@.Dq E
@roff_prof@option for additional information.
@roff_prof@.Ed
.\"-----------------------------------------------------------------------------
@roff_prof@.It Sy "prof.dump (const char *) -w"
@roff_prof@.Bd -ragged -offset indent -compact
@roff_prof@Dump a memory profile to the specified file, or if NULL is specified,
@roff_prof@to a file according to the pattern
@roff_prof@.Pa <prefix>.<pid>.<seq>.m<mseq>.heap ,
@roff_prof@where
@roff_prof@.Pa <prefix>
@roff_prof@is controlled by the
@roff_prof@JEMALLOC_PROF_PREFIX
@roff_prof@environment variable.
@roff_prof@.Ed
.\"-----------------------------------------------------------------------------
@roff_prof@.It Sy "prof.interval (uint64_t) r-"
@roff_prof@.Bd -ragged -offset indent -compact
@roff_prof@Average number of bytes allocated between inverval-based profile
@roff_prof@dumps.
@roff_prof@See the
@roff_prof@.Dq I
@roff_prof@option for additional information.
@roff_prof@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.allocated (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Total number of bytes allocated by the application.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.active (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Total number of bytes in active pages allocated by the application.
@roff_stats@This is a multiple of the page size, and greater than or equal to
@roff_stats@.Dq stats.allocated .
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.mapped (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Total number of bytes in chunks mapped on behalf of the application.
@roff_stats@This is a multiple of the chunk size, and is at least as large as
@roff_stats@.Dq stats.active .
@roff_stats@@roff_swap@This does not include inactive chunks backed by swap
@roff_stats@@roff_swap@files.
@roff_stats@@roff_dss@This does not include inactive chunks embedded in the DSS.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.chunks.current (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Total number of chunks actively mapped on behalf of the application.
@roff_stats@@roff_swap@This does not include inactive chunks backed by swap
@roff_stats@@roff_swap@files.
@roff_stats@@roff_dss@This does not include inactive chunks embedded in the DSS.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.chunks.total (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of chunks allocated.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.chunks.high (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Maximum number of active chunks at any time thus far.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.huge.allocated (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Number of bytes currently allocated by huge objects.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.huge.nmalloc (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of huge allocation requests.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.huge.ndalloc (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of huge deallocation requests.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
.It Sy "stats.arenas.<i>.pactive (size_t) r-"
.Bd -ragged -offset indent -compact
Number of pages in active runs.
.Ed
.\"-----------------------------------------------------------------------------
.It Sy "stats.arenas.<i>.pdirty (size_t) r-"
.Bd -ragged -offset indent -compact
Number of pages within unused runs that are potentially dirty, and for which
.Fn madvise "..." "MADV_DONTNEED"
has not been called.
.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.mapped (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Number of mapped bytes.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.npurge (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Number of dirty page purge sweeps performed.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.nmadvise (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Number of
@roff_stats@.Fn madvise "..." "MADV_DONTNEED"
@roff_stats@calls made to purge dirty pages.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.npurged (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Number of pages purged.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.small.allocated (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Number of bytes currently allocated by small objects.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.small.nmalloc (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of allocation requests served by small bins.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.small.ndalloc (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of small objects returned to bins.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.small.nrequests (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of small allocation requests.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.large.allocated (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Number of bytes currently allocated by large objects.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.large.nmalloc (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of large allocation requests served directly by
@roff_stats@the arena.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.large.ndalloc (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of large deallocation requests served directly by
@roff_stats@the arena.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.large.nrequests (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of large allocation requests.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.bins.<j>.allocated (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Current number of bytes allocated by bin.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.bins.<j>.nmalloc (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of allocations served by bin.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.bins.<j>.ndalloc (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of allocations returned to bin.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.bins.<j>.nrequests (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of allocation requests.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@@roff_tcache@.It Sy "stats.arenas.<i>.bins.<j>.nfills (uint64_t) r-"
@roff_stats@@roff_tcache@.Bd -ragged -offset indent -compact
@roff_stats@@roff_tcache@Cumulative number of tcache fills.
@roff_stats@@roff_tcache@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@@roff_tcache@.It Sy "stats.arenas.<i>.bins.<j>.nflushes (uint64_t) r-"
@roff_stats@@roff_tcache@.Bd -ragged -offset indent -compact
@roff_stats@@roff_tcache@Cumulative number of tcache flushes.
@roff_stats@@roff_tcache@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.bins.<j>.nruns (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of runs created.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.bins.<j>.nreruns (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of times the current run from which to allocate
@roff_stats@changed.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.bins.<j>.highruns (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Maximum number of runs at any time thus far.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.bins.<j>.curruns (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Current number of runs.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.lruns.<j>.nmalloc (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of allocation requests for this size class served
@roff_stats@directly by the arena.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.lruns.<j>.ndalloc (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of deallocation requests for this size class
@roff_stats@served directly by the arena.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.lruns.<j>.nrequests (uint64_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Cumulative number of allocation requests for this size class.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.lruns.<j>.highruns (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Maximum number of runs at any time thus far for this size class.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@.It Sy "stats.arenas.<i>.lruns.<j>.curruns (size_t) r-"
@roff_stats@.Bd -ragged -offset indent -compact
@roff_stats@Current number of runs for this size class.
@roff_stats@.Ed
.\"-----------------------------------------------------------------------------
@roff_stats@@roff_swap@.It Sy "swap.avail (size_t) r-"
@roff_stats@@roff_swap@.Bd -ragged -offset indent -compact
@roff_stats@@roff_swap@Number of swap file bytes that are currently not
@roff_stats@@roff_swap@associated with any chunk (i.e. mapped, but otherwise
@roff_stats@@roff_swap@completely unmanaged).
@roff_stats@@roff_swap@.Ed
.\"-----------------------------------------------------------------------------
@roff_swap@.It Sy "swap.prezeroed (bool) rw"
@roff_swap@.Bd -ragged -offset indent -compact
@roff_swap@If true, the allocator assumes that the swap file(s) contain nothing
@roff_swap@but nil bytes.
@roff_swap@If this assumption is violated, allocator behavior is undefined.
@roff_swap@This value becomes read-only after
@roff_swap@.Dq swap.fds
@roff_swap@is successfully written to.
@roff_swap@.Ed
.\"-----------------------------------------------------------------------------
@roff_swap@.It Sy "swap.nfds (size_t) r-"
@roff_swap@.Bd -ragged -offset indent -compact
@roff_swap@Number of file descriptors in use for swap.
@roff_swap@.Ed
.\"-----------------------------------------------------------------------------
@roff_swap@.It Sy "swap.fds (int *) r-"
@roff_swap@.Bd -ragged -offset indent -compact
@roff_swap@When written to, the files associated with the specified file
@roff_swap@descriptors are contiguously mapped via
@roff_swap@.Xr mmap 2 .
@roff_swap@The resulting virtual memory region is preferred over anonymous
@roff_swap@.Xr mmap 2
@roff_swap@@roff_dss@and
@roff_swap@@roff_dss@.Xr sbrk 2
@roff_swap@memory.
@roff_swap@Note that if a file's size is not a multiple of the page size, it is
@roff_swap@automatically truncated to the nearest page size multiple.
@roff_swap@See the
@roff_swap@.Dq swap.prezeroed
@roff_swap@interface for specifying that the files are pre-zeroed.
@roff_swap@.Ed
.\"-----------------------------------------------------------------------------
.El
.Sh DEBUGGING MALLOC PROBLEMS
The first thing to do is to set the
.Dq A
option.
This option forces a coredump (if possible) at the first sign of trouble,
rather than the normal policy of trying to continue if at all possible.
.Pp
It is probably also a good idea to recompile the program with suitable
options and symbols for debugger support.
.Pp
@roff_fill@If the program starts to give unusual results, coredump or generally
@roff_fill@behave differently without emitting any of the messages mentioned in
@roff_fill@the next section, it is likely because it depends on the storage
@roff_fill@being filled with zero bytes.
@roff_fill@Try running it with the
@roff_fill@.Dq Z
@roff_fill@option set;
@roff_fill@if that improves the situation, this diagnosis has been confirmed.
@roff_fill@If the program still misbehaves,
@roff_fill@the likely problem is accessing memory outside the allocated area.
@roff_fill@.Pp
@roff_fill@Alternatively, if the symptoms are not easy to reproduce, setting the
@roff_fill@.Dq J
@roff_fill@option may help provoke the problem.
@roff_fill@.Pp
Unfortunately this implementation does not provide much detail about
the problems it detects; the performance impact for storing such information
would be prohibitive.
There are a number of allocator implementations available on the Internet
which focus on detecting and pinpointing problems by trading performance for
extra sanity checks and detailed diagnostics.
.Sh DIAGNOSTIC MESSAGES
If any of the memory allocation/deallocation functions detect an error or
warning condition, a message will be printed to file descriptor
.Dv STDERR_FILENO .
Errors will result in the process dumping core.
If the
.Dq A
option is set, all warnings are treated as errors.
.Pp
The
.Va @jemalloc_prefix@malloc_message
variable allows the programmer to override the function which emits the text
strings forming the errors and warnings if for some reason the
.Dv STDERR_FILENO
file descriptor is not suitable for this.
.Va @jemalloc_prefix@malloc_message
takes the
.Fa cbopaque
pointer argument that is
.Dv NULL
unless overridden by the arguments in a call to
.Fn @jemalloc_prefix@malloc_stats_print ,
followed by a string pointer.
Please note that doing anything which tries to allocate memory in this function
is likely to result in a crash or deadlock.
.Pp
All messages are prefixed by
.Dq <jemalloc>: .
.Sh RETURN VALUES
.Ss Standard API
The
.Fn @jemalloc_prefix@malloc
and
.Fn @jemalloc_prefix@calloc
functions return a pointer to the allocated memory if successful; otherwise
a
.Dv NULL
pointer is returned and
.Va errno
is set to
.Er ENOMEM .
.Pp
The
.Fn @jemalloc_prefix@posix_memalign
function returns the value 0 if successful; otherwise it returns an error value.
The
.Fn @jemalloc_prefix@posix_memalign
function will fail if:
.Bl -tag -width Er
.It Bq Er EINVAL
The
.Fa alignment
parameter is not a power of 2 at least as large as
.Fn sizeof "void *" .
.It Bq Er ENOMEM
Memory allocation error.
.El
.Pp
The
.Fn @jemalloc_prefix@realloc
function returns a pointer, possibly identical to
.Fa ptr ,
to the allocated memory
if successful; otherwise a
.Dv NULL
pointer is returned, and
.Va errno
is set to
.Er ENOMEM
if the error was the result of an allocation failure.
The
.Fn @jemalloc_prefix@realloc
function always leaves the original buffer intact
when an error occurs.
.Pp
The
.Fn @jemalloc_prefix@free
function returns no value.
.Ss Non-standard API
The
.Fn @jemalloc_prefix@malloc_usable_size
function returns the usable size of the allocation pointed to by
.Fa ptr .
.Pp
The
.Fn @jemalloc_prefix@mallctl ,
.Fn @jemalloc_prefix@mallctlnametomib ,
and
.Fn @jemalloc_prefix@mallctlbymib
functions return 0 on success; otherwise they return an error value.
The functions will fail if:
.Bl -tag -width Er
.It Bq Er EINVAL
.Fa newp
is
.Dv non-NULL ,
and
.Fa newlen
is too large or too small.
Alternatively,
.Fa *oldlenp
is too large or too small; in this case as much data as possible are read
despite the error.
.It Bq Er ENOMEM
.Fa *oldlenp
is too short to hold the requested value.
.It Bq Er ENOENT
.Fa name
or
.Fa mib
specifies an unknown/invalid value.
.It Bq Er EPERM
Attempt to read or write void value, or attempt to write read-only value.
.It Bq Er EAGAIN
A memory allocation failure occurred.
.It Bq Er EFAULT
An interface with side effects failed in some way not directly related to
.Fn @jemalloc_prefix@mallctl*
read/write processing.
.El
.Ss Experimental API
The
.Fn @jemalloc_prefix@allocm ,
.Fn @jemalloc_prefix@rallocm ,
.Fn @jemalloc_prefix@sallocm ,
and
.Fn @jemalloc_prefix@dallocm
functions return
.Dv ALLOCM_SUCCESS
on success; otherwise they return an error value.
The
.Fn @jemalloc_prefix@allocm
and
.Fn @jemalloc_prefix@rallocm
functions will fail if:
.Bl -tag -width ".Dv ALLOCM_ERR_OOM"
.It ALLOCM_ERR_OOM
Out of memory.
Insufficient contiguous memory was available to service the allocation request.
The
.Fn @jemalloc_prefix@allocm
function additionally sets
.Fa *ptr
to
.Dv NULL ,
whereas the
.Fn @jemalloc_prefix@rallocm
function leaves
.Fa *ptr
unmodified.
.El
.Pp
The
.Fn @jemalloc_prefix@rallocm
function will also fail if:
.Bl -tag -width ".Dv ALLOCM_ERR_NOT_MOVED"
.It ALLOCM_ERR_NOT_MOVED
.Dv ALLOCM_NO_MOVE
was specified, but the reallocation request could not be serviced without
moving the object.
.El
.Sh ENVIRONMENT
The following environment variables affect the execution of the allocation
functions:
@roff_prof@.Bl -tag -width ".Ev JEMALLOC_PROF_PREFIX"
@roff_no_prof@.Bl -tag -width ".Ev JEMALLOC_OPTIONS"
.It Ev JEMALLOC_OPTIONS
If the environment variable
.Ev JEMALLOC_OPTIONS
is set, the characters it contains will be interpreted as flags to the
allocation functions.
@roff_prof@.It Ev JEMALLOC_PROF_PREFIX
@roff_prof@If the environment variable
@roff_prof@.Ev JEMALLOC_PROF_PREFIX
@roff_prof@is set, use it as the filename prefix for profile dumps; otherwise
@roff_prof@use
@roff_prof@.Pa jeprof
@roff_prof@as the prefix.
.El
.Sh EXAMPLES
To dump core whenever a problem occurs:
.Pp
.Bd -literal -offset indent
ln -s 'A' /etc/jemalloc.conf
.Ed
.Pp
To specify in the source a chunk size that is twice the default:
.Bd -literal -offset indent
@jemalloc_prefix@malloc_options = "K";
.Ed
.Sh SEE ALSO
.Xr madvise 2 ,
.Xr mmap 2 ,
@roff_dss@.Xr sbrk 2 ,
.Xr alloca 3 ,
.Xr atexit 3 ,
.Xr getpagesize 3
.Sh STANDARDS
The
.Fn @jemalloc_prefix@malloc ,
.Fn @jemalloc_prefix@calloc ,
.Fn @jemalloc_prefix@realloc
and
.Fn @jemalloc_prefix@free
functions conform to
.St -isoC .
.Pp
The
.Fn @jemalloc_prefix@posix_memalign
function conforms to
.St -p1003.1-2001 .