.\" Copyright (c) 2006-2010 Jason Evans . .\" 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 October 24, 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@ .Pp This manual describes jemalloc @jemalloc_version@. More information can be found at the .UR http://\:www.canonware.com/\:jemalloc/ jemalloc website .UE . .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_conf ; .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 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 the memory allocation routines, the allocator initializes its internals based in part on various options that can be specified at compile- or run-time. .Pp The string pointed to by the global variable .Va @jemalloc_prefix@malloc_conf , the .Dq name of the file referenced by the symbolic link named .Pa /etc/@jemalloc_prefix@malloc.conf , and the value of the environment variable .Ev @jemalloc_cprefix@MALLOC_CONF , will be interpreted, in that order, from left to right as options. .Pp An options string is a comma-separated list of option:value pairs. There is one key corresponding to each .Dq opt.* mallctl. For example, .Dq abort:true,narenas:1 sets the .Dq opt.abort and .Dq opt.narenas options. Some options have boolean values (true/false), others have integer values (base 8, 10, or 16, depending on prefix), and yet others have raw string values. .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 @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 frontier and free list to track 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 opt.lg_qspace_max 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 opt.lg_cspace_max 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 opt.lg_chunk 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 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. .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 allbox tab(;); LLL LLL ^LL ^LL ^LL LsL LsL. 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-spaced;[768, 1024, 1280, ..., 3840] Large;[4 KiB, 8 KiB, 12 KiB, ..., 4072 KiB] Huge;[4 MiB, 8 MiB, 12 MiB, ...] .TE .\"----------------------------------------------------------------------------- .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 or 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..* , @roff_stats@ equal to @roff_stats@.Dq arenas.narenas @roff_stats@can be used to access the summation of statistics from all arenas. .Pp Take special note of the .Dq epoch mallctl, which controls refreshing of cached dynamic statistics. .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 .\"----------------------------------------------------------------------------- .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 Abort-on-warning enabled/disabled. If true, most warnings are fatal. The process will call .Xr abort 3 in these cases. This option is @roff_debug@enabled @roff_no_debug@disabled by default. .Ed .\"----------------------------------------------------------------------------- .It Sy "opt.lg_qspace_max (size_t) r-" .Bd -ragged -offset indent -compact Size (log base 2) 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 (2^7). .Ed .\"----------------------------------------------------------------------------- .It Sy "opt.lg_cspace_max (size_t) r-" .Bd -ragged -offset indent -compact Size (log base 2) 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 (2^9). .Ed .\"----------------------------------------------------------------------------- .It Sy "opt.lg_chunk (size_t) r-" .Bd -ragged -offset indent -compact Virtual memory chunk size (log base 2). The default chunk size is 4 MiB (2^22). .Ed .\"----------------------------------------------------------------------------- .It Sy "opt.narenas (size_t) r-" .Bd -ragged -offset indent -compact Maximum number of arenas to use. The default maximum number of arenas is four times the number of CPUs, or one if there is a single CPU. .Ed .\"----------------------------------------------------------------------------- .It Sy "opt.lg_dirty_mult (ssize_t) r-" .Bd -ragged -offset indent -compact Per-arena minimum ratio (log base 2) 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 or a similar system call. 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 (2^5:1); an option value of -1 will disable dirty page purging. .Ed .\"----------------------------------------------------------------------------- .It Sy "opt.stats_print (bool) r-" .Bd -ragged -offset indent -compact Enable/disable statistics printing at exit. If enabled, 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. This option is disabled by default. .Ed .\"----------------------------------------------------------------------------- @roff_fill@.It Sy "opt.junk (bool) r-" @roff_fill@.Bd -ragged -offset indent -compact @roff_fill@Junk filling enabled/disabled. @roff_fill@If enabled, each byte of uninitialized allocated memory will be @roff_fill@initialized to 0xa5. @roff_fill@All deallocated memory will be initialized to 0x5a. @roff_fill@This is intended for debugging and will impact performance @roff_fill@negatively. @roff_fill@This option is @roff_fill@@roff_debug@enabled @roff_fill@@roff_no_debug@disabled @roff_fill@by default. @roff_fill@.Ed .\"----------------------------------------------------------------------------- @roff_fill@.It Sy "opt.zero (bool) r-" @roff_fill@.Bd -ragged -offset indent -compact @roff_fill@Zero filling enabled/disabled. @roff_fill@If enabled, each byte of uninitialized allocated memory will be @roff_fill@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. @roff_fill@This option is disabled by default. @roff_fill@.Ed .\"----------------------------------------------------------------------------- @roff_sysv@.It Sy "opt.sysv (bool) r-" @roff_sysv@.Bd -ragged -offset indent -compact @roff_sysv@If enabled, 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 opt.xmalloc @roff_sysv@@roff_xmalloc@option. @roff_sysv@This option is disabled by default. @roff_sysv@.Ed .\"----------------------------------------------------------------------------- @roff_xmalloc@.It Sy "opt.xmalloc (bool) r-" @roff_xmalloc@.Bd -ragged -offset indent -compact @roff_xmalloc@Abort-on-out-of-memory enabled/disabled. @roff_xmalloc@If enabled, rather than returning failure for any allocation @roff_xmalloc@function, display a 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@If an application is designed to depend on this behavior, set the @roff_xmalloc@option at compile time by including the following in the source @roff_xmalloc@code: @roff_xmalloc@.Bd -literal -offset indent @roff_xmalloc@@jemalloc_prefix@malloc_conf = "xmalloc:true"; @roff_xmalloc@.Ed @roff_xmalloc@.Pp @roff_xmalloc@This option is disabled by default. @roff_xmalloc@.Ed .\"----------------------------------------------------------------------------- @roff_tcache@.It Sy "opt.tcache (bool) r-" @roff_tcache@.Bd -ragged -offset indent -compact @roff_tcache@Thread-specific caching enabled/disabled. @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 opt.lg_tcache_gc_sweep @roff_tcache@and @roff_tcache@.Dq opt.tcache_max @roff_tcache@options for related tuning information. @roff_tcache@This option is enabled by default. @roff_tcache@.Ed .\"----------------------------------------------------------------------------- @roff_tcache@.It Sy "opt.lg_tcache_gc_sweep (ssize_t) r-" @roff_tcache@.Bd -ragged -offset indent -compact @roff_tcache@Approximate interval (log base 2) between full thread-specific @roff_tcache@cache garbage collection sweeps, counted in terms of @roff_tcache@thread-specific cache allocation/deallocation events. @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 (2^13); setting this option to @roff_tcache@-1 will disable garbage collection. @roff_tcache@.Ed .\"----------------------------------------------------------------------------- @roff_tcache@.It Sy "opt.lg_tcache_max (size_t) r-" @roff_tcache@.Bd -ragged -offset indent -compact @roff_tcache@Maximum size class (log base 2) to cache in the thread-specific @roff_tcache@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 (2^15). @roff_tcache@.Ed .\"----------------------------------------------------------------------------- @roff_prof@.It Sy "opt.prof (bool) r-" @roff_prof@.Bd -ragged -offset indent -compact @roff_prof@Memory profiling enabled/disabled. @roff_prof@If enabled, 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 ...f.heap , @roff_prof@where @roff_prof@.Pa @roff_prof@is controlled by the @roff_prof@.Dq opt.prof_prefix @roff_prof@option. @roff_prof@See the @roff_prof@.Dq opt.lg_prof_bt_max @roff_prof@option for backtrace depth control. @roff_prof@See the @roff_prof@.Dq opt.prof_active @roff_prof@option for on-the-fly activation/deactivation. @roff_prof@See the @roff_prof@.Dq opt.lg_prof_sample @roff_prof@option for probabilistic sampling control. @roff_prof@See the @roff_prof@.Dq opt.prof_accum @roff_prof@option for control of cumulative sample reporting. @roff_prof@See the @roff_prof@.Dq opt.lg_prof_tcmax @roff_prof@option for control of per thread backtrace caching. @roff_prof@See the @roff_prof@.Dq opt.lg_prof_interval @roff_prof@option for information on interval-triggered profile dumping, and the @roff_prof@.Dq opt.prof_gdump @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 @roff_prof@.UR http://\:code.google.com/\:p/\:google-perftools/ @roff_prof@google-perftools package @roff_prof@.UE . @roff_prof@.Ed .\"----------------------------------------------------------------------------- @roff_prof@.It Sy "opt.prof_prefix (const char *) r-" @roff_prof@.Bd -ragged -offset indent -compact @roff_prof@Filename prefix for profile dumps. @roff_prof@If the prefix is set to the empty string, no automatic dumps will @roff_prof@occur; this is primarily useful for disabling the automatic final @roff_prof@heap dump (which also disables leak reporting, if enabled). @roff_prof@The default prefix is @roff_prof@.Pa jeprof . @roff_prof@.Ed .\"----------------------------------------------------------------------------- @roff_prof@.It Sy "opt.lg_prof_bt_max (size_t) r-" @roff_prof@.Bd -ragged -offset indent -compact @roff_prof@Maximum backtrace depth (log base 2) when profiling memory @roff_prof@allocation activity. @roff_prof@The default is 128 (2^7). @roff_prof@.Ed .\"----------------------------------------------------------------------------- @roff_prof@.It Sy "opt.prof_active (bool) r-" @roff_prof@.Bd -ragged -offset indent -compact @roff_prof@Profiling activated/deactivated. @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 opt.prof @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@.Ed .\"----------------------------------------------------------------------------- @roff_prof@.It Sy "opt.lg_prof_sample (ssize_t) r-" @roff_prof@.Bd -ragged -offset indent -compact @roff_prof@Average interval (log base 2) 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 1 (2^0) (i.e. all allocations are @roff_prof@sampled). @roff_prof@.Ed .\"----------------------------------------------------------------------------- @roff_prof@.It Sy "opt.prof_accum (bool) r-" @roff_prof@.Bd -ragged -offset indent -compact @roff_prof@Reporting of cumulative object/byte counts in profile dumps @roff_prof@enabled/disabled. @roff_prof@If this option is enabled, every unique backtrace must be stored for @roff_prof@the duration of execution. @roff_prof@Depending on the application, this can impose a large memory @roff_prof@overhead, and the cumulative counts are not always of interest. @roff_prof@See the @roff_prof@.Dq opt.lg_prof_tcmax @roff_prof@option for control of per thread backtrace caching, which has @roff_prof@important interactions. @roff_prof@This option is enabled by default. @roff_prof@.Ed .\"----------------------------------------------------------------------------- @roff_prof@.It Sy "opt.lg_prof_tcmax (ssize_t) r-" @roff_prof@.Bd -ragged -offset indent -compact @roff_prof@Maximum per thread backtrace cache (log base 2) used for heap @roff_prof@profiling. @roff_prof@A backtrace can only be discarded if the @roff_prof@.Dq opt.prof_accum @roff_prof@option is disabled, and no thread caches currently refer to the @roff_prof@backtrace. @roff_prof@Therefore, a backtrace cache limit should be imposed if the @roff_prof@intention is to limit how much memory is used by backtraces. @roff_prof@By default, no limit is imposed (encoded as -1). @roff_prof@.Ed .\"----------------------------------------------------------------------------- @roff_prof@.It Sy "opt.lg_prof_interval (ssize_t) r-" @roff_prof@.Bd -ragged -offset indent -compact @roff_prof@Average interval (log base 2) 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 ...i.heap , @roff_prof@where @roff_prof@.Pa @roff_prof@is controlled by the @roff_prof@.Dq opt.prof_prefix @roff_prof@option. @roff_prof@By default, interval-triggered profile dumping is disabled (encoded @roff_prof@as -1). @roff_prof@.Ed .\"----------------------------------------------------------------------------- @roff_prof@.It Sy "opt.prof_gdump (bool) r-" @roff_prof@.Bd -ragged -offset indent -compact @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 ...u.heap , @roff_prof@where @roff_prof@.Pa @roff_prof@is controlled by the @roff_prof@.Dq opt.prof_prefix @roff_prof@option. @roff_prof@This option is disabled by default. @roff_prof@.Ed .\"----------------------------------------------------------------------------- @roff_prof@.It Sy "opt.prof_leak (bool) r-" @roff_prof@.Bd -ragged -offset indent -compact @roff_prof@Leak reporting enabled/disabled. @roff_prof@If enabled, use an @roff_prof@.Xr atexit 3 @roff_prof@function to report memory leaks detected by allocation sampling. @roff_prof@See the @roff_prof@.Dq opt.lg_prof_bt_max @roff_prof@option for backtrace depth control. @roff_prof@See the @roff_prof@.Dq opt.prof @roff_prof@option for information on analyzing heap profile output. @roff_prof@This option is disabled by default. @roff_prof@.Ed .\"----------------------------------------------------------------------------- .It Sy "opt.overcommit (bool) r-" .Bd -ragged -offset indent -compact @roff_swap@Over-commit enabled/disabled. @roff_swap@If enabled, 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. .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 .\"----------------------------------------------------------------------------- @roff_stats@.It Sy "thread.allocated (uint64_t) r-" @roff_stats@.Bd -ragged -offset indent -compact @roff_stats@Get the total number of bytes ever allocated by the calling thread. @roff_stats@This counter has the potential to wrap around; it is up to the @roff_stats@application to appropriately interpret the counter in such cases. @roff_stats@.Ed .\"----------------------------------------------------------------------------- @roff_stats@.It Sy "thread.deallocated (uint64_t) r-" @roff_stats@.Bd -ragged -offset indent -compact @roff_stats@Get the total number of bytes ever deallocated by the calling @roff_stats@thread. @roff_stats@This counter has the potential to wrap around; it is up to the @roff_stats@application to appropriately interpret the counter in such cases. @roff_stats@.Ed .\"----------------------------------------------------------------------------- .It Sy "arenas.narenas (unsigned) r-" .Bd -ragged -offset indent -compact Maximum number of arenas. .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..size (size_t) r-" .Bd -ragged -offset indent -compact Maximum size supported by size class. .Ed .\"----------------------------------------------------------------------------- .It Sy "arenas.bin..nregs (uint32_t) r-" .Bd -ragged -offset indent -compact Number of regions per page run. .Ed .\"----------------------------------------------------------------------------- .It Sy "arenas.bin..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..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 opt.prof_active @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 ...m.heap , @roff_prof@where @roff_prof@.Pa @roff_prof@is controlled by the @roff_prof@.Dq opt.prof_prefix @roff_prof@option. @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 opt.lg_prof_interval @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..pactive (size_t) r-" .Bd -ragged -offset indent -compact Number of pages in active runs. .Ed .\"----------------------------------------------------------------------------- .It Sy "stats.arenas..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..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..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..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..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..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..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..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..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..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..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..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..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..bins..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..bins..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..bins..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..bins..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..bins..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..bins..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..bins..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..bins..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..bins..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..bins..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..lruns..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..lruns..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..lruns..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..lruns..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..lruns..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 Start by setting the .Dq opt.abort option, which 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 the program depends on the @roff_fill@storage being filled with zero bytes. @roff_fill@Try running it with the @roff_fill@.Dq opt.zero @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 opt.junk @roff_fill@option may help provoke the problem. @roff_fill@.Pp This implementation does not provide much detail about the problems it detects, because 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 opt.abort option is set, most 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 : . .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 ".Bq Er ALLOCM_ERR_OOM" .It Bq Er 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 ".Bq Er ALLOCM_ERR_NOT_MOVED" .It Bq Er 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 variable affects the execution of the allocation functions: .Bl -tag -width ".Ev @jemalloc_cprefix@MALLOC_CONF" .It Ev @jemalloc_cprefix@MALLOC_CONF If the environment variable .Ev @jemalloc_cprefix@MALLOC_CONF is set, the characters it contains will be interpreted as options. .El .Sh EXAMPLES To dump core whenever a problem occurs: .Pp .Bd -literal -offset indent ln -s 'abort:true' /etc/@jemalloc_prefix@malloc.conf .Ed .Pp To specify in the source a chunk size that is 16 MiB: .Bd -literal -offset indent @jemalloc_prefix@malloc_conf = "lg_chunk:24"; .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 .