From 1f17bd93954d6744fb1667d51d20a2877f2e0737 Mon Sep 17 00:00:00 2001 From: Jason Evans Date: Fri, 5 Nov 2010 15:53:34 -0700 Subject: [PATCH 1/6] Fix mallctlnametomib() documentation. Fix the prototype for mallctlnametomib() in the manual page to correspond to reality. --- jemalloc/doc/jemalloc.3.in | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/jemalloc/doc/jemalloc.3.in b/jemalloc/doc/jemalloc.3.in index 62866642..d0129080 100644 --- a/jemalloc/doc/jemalloc.3.in +++ b/jemalloc/doc/jemalloc.3.in @@ -38,7 +38,7 @@ .\" @(#)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 +.Dd November 5, 2010 .Dt jemalloc 3 .Os .Sh NAME @@ -87,7 +87,7 @@ jemalloc website .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" +.Fn @jemalloc_prefix@mallctlnametomib "const char *name" "size_t *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 * From fc4dcfa2f546354dd536a3a33aa9f10fd3253731 Mon Sep 17 00:00:00 2001 From: Jason Evans Date: Wed, 24 Nov 2010 15:44:21 -0800 Subject: [PATCH 2/6] Push down ctl_mtx. Many mallctl*() end points require no locking, so push the locking down to just the functions that need it. This is of particular import for "thread.allocated" and "thread.deallocated", which are intended as a low-overhead way to introspect per thread allocation activity. --- jemalloc/src/ctl.c | 196 ++++++++++++++++++++++++++++----------------- 1 file changed, 123 insertions(+), 73 deletions(-) diff --git a/jemalloc/src/ctl.c b/jemalloc/src/ctl.c index c83ee4f1..0e100590 100644 --- a/jemalloc/src/ctl.c +++ b/jemalloc/src/ctl.c @@ -4,6 +4,13 @@ /******************************************************************************/ /* Data. */ +/* + * ctl_mtx protects the following: + * - ctl_stats.* + * - opt_prof_active + * - swap_enabled + * - swap_prezeroed + */ static malloc_mutex_t ctl_mtx; static bool ctl_initialized; static uint64_t ctl_epoch; @@ -680,7 +687,9 @@ ctl_refresh(void) static bool ctl_init(void) { + bool ret; + malloc_mutex_lock(&ctl_mtx); if (ctl_initialized == false) { #ifdef JEMALLOC_STATS unsigned i; @@ -692,8 +701,10 @@ ctl_init(void) */ ctl_stats.arenas = (ctl_arena_stats_t *)base_alloc( (narenas + 1) * sizeof(ctl_arena_stats_t)); - if (ctl_stats.arenas == NULL) - return (true); + if (ctl_stats.arenas == NULL) { + ret = true; + goto RETURN; + } memset(ctl_stats.arenas, 0, (narenas + 1) * sizeof(ctl_arena_stats_t)); @@ -704,8 +715,10 @@ ctl_init(void) */ #ifdef JEMALLOC_STATS for (i = 0; i <= narenas; i++) { - if (ctl_arena_init(&ctl_stats.arenas[i])) - return (true); + if (ctl_arena_init(&ctl_stats.arenas[i])) { + ret = true; + goto RETURN; + } } #endif ctl_stats.arenas[narenas].initialized = true; @@ -715,7 +728,10 @@ ctl_init(void) ctl_initialized = true; } - return (false); + ret = false; +RETURN: + malloc_mutex_unlock(&ctl_mtx); + return (ret); } static int @@ -825,8 +841,7 @@ ctl_byname(const char *name, void *oldp, size_t *oldlenp, void *newp, ctl_node_t const *nodes[CTL_MAX_DEPTH]; size_t mib[CTL_MAX_DEPTH]; - malloc_mutex_lock(&ctl_mtx); - if (ctl_init()) { + if (ctl_initialized == false && ctl_init()) { ret = EAGAIN; goto RETURN; } @@ -841,10 +856,9 @@ ctl_byname(const char *name, void *oldp, size_t *oldlenp, void *newp, ret = ENOENT; goto RETURN; } - ret = nodes[depth-1]->ctl(mib, depth, oldp, oldlenp, newp, newlen); + ret = nodes[depth-1]->ctl(mib, depth, oldp, oldlenp, newp, newlen); RETURN: - malloc_mutex_unlock(&ctl_mtx); return(ret); } @@ -853,16 +867,13 @@ ctl_nametomib(const char *name, size_t *mibp, size_t *miblenp) { int ret; - malloc_mutex_lock(&ctl_mtx); - if (ctl_init()) { + if (ctl_initialized == false && ctl_init()) { ret = EAGAIN; goto RETURN; } ret = ctl_lookup(name, NULL, mibp, miblenp); - RETURN: - malloc_mutex_unlock(&ctl_mtx); return(ret); } @@ -874,8 +885,7 @@ ctl_bymib(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp, const ctl_node_t *node; size_t i; - malloc_mutex_lock(&ctl_mtx); - if (ctl_init()) { + if (ctl_initialized == false && ctl_init()) { ret = EAGAIN; goto RETURN; } @@ -912,7 +922,6 @@ ctl_bymib(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp, ret = node->ctl(mib, miblen, oldp, oldlenp, newp, newlen); RETURN: - malloc_mutex_unlock(&ctl_mtx); return(ret); } @@ -975,6 +984,29 @@ ctl_boot(void) #define CTL_RO_GEN(n, v, t) \ static int \ +n##_ctl(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp, \ + void *newp, size_t newlen) \ +{ \ + int ret; \ + t oldval; \ + \ + malloc_mutex_lock(&ctl_mtx); \ + READONLY(); \ + oldval = v; \ + READ(oldval, t); \ + \ + ret = 0; \ +RETURN: \ + malloc_mutex_unlock(&ctl_mtx); \ + return (ret); \ +} + +/* + * ctl_mtx is not acquired, under the assumption that no pertinent data will + * mutate during the call. + */ +#define CTL_RO_NL_GEN(n, v, t) \ +static int \ n##_ctl(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp, \ void *newp, size_t newlen) \ { \ @@ -1024,7 +1056,7 @@ RETURN: \ return (ret); \ } -CTL_RO_GEN(version, JEMALLOC_VERSION, const char *) +CTL_RO_NL_GEN(version, JEMALLOC_VERSION, const char *) static int epoch_ctl(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp, @@ -1033,6 +1065,7 @@ epoch_ctl(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp, int ret; uint64_t newval; + malloc_mutex_lock(&ctl_mtx); newval = 0; WRITE(newval, uint64_t); if (newval != 0) @@ -1041,6 +1074,7 @@ epoch_ctl(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp, ret = 0; RETURN: + malloc_mutex_unlock(&ctl_mtx); return (ret); } @@ -1107,8 +1141,8 @@ RETURN: } #ifdef JEMALLOC_STATS -CTL_RO_GEN(thread_allocated, ALLOCATED_GET(), uint64_t); -CTL_RO_GEN(thread_deallocated, DEALLOCATED_GET(), uint64_t); +CTL_RO_NL_GEN(thread_allocated, ALLOCATED_GET(), uint64_t); +CTL_RO_NL_GEN(thread_deallocated, DEALLOCATED_GET(), uint64_t); #endif /******************************************************************************/ @@ -1205,48 +1239,48 @@ CTL_RO_FALSE_GEN(config_xmalloc) /******************************************************************************/ -CTL_RO_GEN(opt_abort, opt_abort, bool) -CTL_RO_GEN(opt_lg_qspace_max, opt_lg_qspace_max, size_t) -CTL_RO_GEN(opt_lg_cspace_max, opt_lg_cspace_max, size_t) -CTL_RO_GEN(opt_lg_chunk, opt_lg_chunk, size_t) -CTL_RO_GEN(opt_narenas, opt_narenas, size_t) -CTL_RO_GEN(opt_lg_dirty_mult, opt_lg_dirty_mult, ssize_t) -CTL_RO_GEN(opt_stats_print, opt_stats_print, bool) +CTL_RO_NL_GEN(opt_abort, opt_abort, bool) +CTL_RO_NL_GEN(opt_lg_qspace_max, opt_lg_qspace_max, size_t) +CTL_RO_NL_GEN(opt_lg_cspace_max, opt_lg_cspace_max, size_t) +CTL_RO_NL_GEN(opt_lg_chunk, opt_lg_chunk, size_t) +CTL_RO_NL_GEN(opt_narenas, opt_narenas, size_t) +CTL_RO_NL_GEN(opt_lg_dirty_mult, opt_lg_dirty_mult, ssize_t) +CTL_RO_NL_GEN(opt_stats_print, opt_stats_print, bool) #ifdef JEMALLOC_FILL -CTL_RO_GEN(opt_junk, opt_junk, bool) -CTL_RO_GEN(opt_zero, opt_zero, bool) +CTL_RO_NL_GEN(opt_junk, opt_junk, bool) +CTL_RO_NL_GEN(opt_zero, opt_zero, bool) #endif #ifdef JEMALLOC_SYSV -CTL_RO_GEN(opt_sysv, opt_sysv, bool) +CTL_RO_NL_GEN(opt_sysv, opt_sysv, bool) #endif #ifdef JEMALLOC_XMALLOC -CTL_RO_GEN(opt_xmalloc, opt_xmalloc, bool) +CTL_RO_NL_GEN(opt_xmalloc, opt_xmalloc, bool) #endif #ifdef JEMALLOC_TCACHE -CTL_RO_GEN(opt_tcache, opt_tcache, bool) -CTL_RO_GEN(opt_lg_tcache_gc_sweep, opt_lg_tcache_gc_sweep, ssize_t) +CTL_RO_NL_GEN(opt_tcache, opt_tcache, bool) +CTL_RO_NL_GEN(opt_lg_tcache_gc_sweep, opt_lg_tcache_gc_sweep, ssize_t) #endif #ifdef JEMALLOC_PROF -CTL_RO_GEN(opt_prof, opt_prof, bool) -CTL_RO_GEN(opt_prof_prefix, opt_prof_prefix, const char *) -CTL_RO_GEN(opt_prof_active, opt_prof_active, bool) -CTL_RO_GEN(opt_lg_prof_bt_max, opt_lg_prof_bt_max, size_t) -CTL_RO_GEN(opt_lg_prof_sample, opt_lg_prof_sample, size_t) -CTL_RO_GEN(opt_lg_prof_interval, opt_lg_prof_interval, ssize_t) -CTL_RO_GEN(opt_prof_gdump, opt_prof_gdump, bool) -CTL_RO_GEN(opt_prof_leak, opt_prof_leak, bool) -CTL_RO_GEN(opt_prof_accum, opt_prof_accum, bool) -CTL_RO_GEN(opt_lg_prof_tcmax, opt_lg_prof_tcmax, ssize_t) +CTL_RO_NL_GEN(opt_prof, opt_prof, bool) +CTL_RO_NL_GEN(opt_prof_prefix, opt_prof_prefix, const char *) +CTL_RO_GEN(opt_prof_active, opt_prof_active, bool) /* Mutable. */ +CTL_RO_NL_GEN(opt_lg_prof_bt_max, opt_lg_prof_bt_max, size_t) +CTL_RO_NL_GEN(opt_lg_prof_sample, opt_lg_prof_sample, size_t) +CTL_RO_NL_GEN(opt_lg_prof_interval, opt_lg_prof_interval, ssize_t) +CTL_RO_NL_GEN(opt_prof_gdump, opt_prof_gdump, bool) +CTL_RO_NL_GEN(opt_prof_leak, opt_prof_leak, bool) +CTL_RO_NL_GEN(opt_prof_accum, opt_prof_accum, bool) +CTL_RO_NL_GEN(opt_lg_prof_tcmax, opt_lg_prof_tcmax, ssize_t) #endif #ifdef JEMALLOC_SWAP -CTL_RO_GEN(opt_overcommit, opt_overcommit, bool) +CTL_RO_NL_GEN(opt_overcommit, opt_overcommit, bool) #endif /******************************************************************************/ -CTL_RO_GEN(arenas_bin_i_size, arenas[0]->bins[mib[2]].reg_size, size_t) -CTL_RO_GEN(arenas_bin_i_nregs, arenas[0]->bins[mib[2]].nregs, uint32_t) -CTL_RO_GEN(arenas_bin_i_run_size, arenas[0]->bins[mib[2]].run_size, size_t) +CTL_RO_NL_GEN(arenas_bin_i_size, arenas[0]->bins[mib[2]].reg_size, size_t) +CTL_RO_NL_GEN(arenas_bin_i_nregs, arenas[0]->bins[mib[2]].nregs, uint32_t) +CTL_RO_NL_GEN(arenas_bin_i_run_size, arenas[0]->bins[mib[2]].run_size, size_t) const ctl_node_t * arenas_bin_i_index(const size_t *mib, size_t miblen, size_t i) { @@ -1256,7 +1290,7 @@ arenas_bin_i_index(const size_t *mib, size_t miblen, size_t i) return (super_arenas_bin_i_node); } -CTL_RO_GEN(arenas_lrun_i_size, ((mib[2]+1) << PAGE_SHIFT), size_t) +CTL_RO_NL_GEN(arenas_lrun_i_size, ((mib[2]+1) << PAGE_SHIFT), size_t) const ctl_node_t * arenas_lrun_i_index(const size_t *mib, size_t miblen, size_t i) { @@ -1266,7 +1300,7 @@ arenas_lrun_i_index(const size_t *mib, size_t miblen, size_t i) return (super_arenas_lrun_i_node); } -CTL_RO_GEN(arenas_narenas, narenas, unsigned) +CTL_RO_NL_GEN(arenas_narenas, narenas, unsigned) static int arenas_initialized_ctl(const size_t *mib, size_t miblen, void *oldp, @@ -1275,6 +1309,7 @@ arenas_initialized_ctl(const size_t *mib, size_t miblen, void *oldp, int ret; unsigned nread, i; + malloc_mutex_lock(&ctl_mtx); READONLY(); if (*oldlenp != narenas * sizeof(bool)) { ret = EINVAL; @@ -1289,36 +1324,37 @@ arenas_initialized_ctl(const size_t *mib, size_t miblen, void *oldp, ((bool *)oldp)[i] = ctl_stats.arenas[i].initialized; RETURN: + malloc_mutex_unlock(&ctl_mtx); return (ret); } -CTL_RO_GEN(arenas_quantum, QUANTUM, size_t) -CTL_RO_GEN(arenas_cacheline, CACHELINE, size_t) -CTL_RO_GEN(arenas_subpage, SUBPAGE, size_t) -CTL_RO_GEN(arenas_pagesize, PAGE_SIZE, size_t) -CTL_RO_GEN(arenas_chunksize, chunksize, size_t) +CTL_RO_NL_GEN(arenas_quantum, QUANTUM, size_t) +CTL_RO_NL_GEN(arenas_cacheline, CACHELINE, size_t) +CTL_RO_NL_GEN(arenas_subpage, SUBPAGE, size_t) +CTL_RO_NL_GEN(arenas_pagesize, PAGE_SIZE, size_t) +CTL_RO_NL_GEN(arenas_chunksize, chunksize, size_t) #ifdef JEMALLOC_TINY -CTL_RO_GEN(arenas_tspace_min, (1U << LG_TINY_MIN), size_t) -CTL_RO_GEN(arenas_tspace_max, (qspace_min >> 1), size_t) +CTL_RO_NL_GEN(arenas_tspace_min, (1U << LG_TINY_MIN), size_t) +CTL_RO_NL_GEN(arenas_tspace_max, (qspace_min >> 1), size_t) #endif -CTL_RO_GEN(arenas_qspace_min, qspace_min, size_t) -CTL_RO_GEN(arenas_qspace_max, qspace_max, size_t) -CTL_RO_GEN(arenas_cspace_min, cspace_min, size_t) -CTL_RO_GEN(arenas_cspace_max, cspace_max, size_t) -CTL_RO_GEN(arenas_sspace_min, sspace_min, size_t) -CTL_RO_GEN(arenas_sspace_max, sspace_max, size_t) +CTL_RO_NL_GEN(arenas_qspace_min, qspace_min, size_t) +CTL_RO_NL_GEN(arenas_qspace_max, qspace_max, size_t) +CTL_RO_NL_GEN(arenas_cspace_min, cspace_min, size_t) +CTL_RO_NL_GEN(arenas_cspace_max, cspace_max, size_t) +CTL_RO_NL_GEN(arenas_sspace_min, sspace_min, size_t) +CTL_RO_NL_GEN(arenas_sspace_max, sspace_max, size_t) #ifdef JEMALLOC_TCACHE -CTL_RO_GEN(arenas_tcache_max, tcache_maxclass, size_t) +CTL_RO_NL_GEN(arenas_tcache_max, tcache_maxclass, size_t) #endif -CTL_RO_GEN(arenas_ntbins, ntbins, unsigned) -CTL_RO_GEN(arenas_nqbins, nqbins, unsigned) -CTL_RO_GEN(arenas_ncbins, ncbins, unsigned) -CTL_RO_GEN(arenas_nsbins, nsbins, unsigned) -CTL_RO_GEN(arenas_nbins, nbins, unsigned) +CTL_RO_NL_GEN(arenas_ntbins, ntbins, unsigned) +CTL_RO_NL_GEN(arenas_nqbins, nqbins, unsigned) +CTL_RO_NL_GEN(arenas_ncbins, ncbins, unsigned) +CTL_RO_NL_GEN(arenas_nsbins, nsbins, unsigned) +CTL_RO_NL_GEN(arenas_nbins, nbins, unsigned) #ifdef JEMALLOC_TCACHE -CTL_RO_GEN(arenas_nhbins, nhbins, unsigned) +CTL_RO_NL_GEN(arenas_nhbins, nhbins, unsigned) #endif -CTL_RO_GEN(arenas_nlruns, nlclasses, size_t) +CTL_RO_NL_GEN(arenas_nlruns, nlclasses, size_t) static int arenas_purge_ctl(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp, @@ -1368,6 +1404,7 @@ prof_active_ctl(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp, int ret; bool oldval; + malloc_mutex_lock(&ctl_mtx); /* Protect opt_prof_active. */ oldval = opt_prof_active; if (newp != NULL) { /* @@ -1382,6 +1419,7 @@ prof_active_ctl(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp, ret = 0; RETURN: + malloc_mutex_unlock(&ctl_mtx); return (ret); } @@ -1405,7 +1443,7 @@ RETURN: return (ret); } -CTL_RO_GEN(prof_interval, prof_interval, uint64_t) +CTL_RO_NL_GEN(prof_interval, prof_interval, uint64_t) #endif /******************************************************************************/ @@ -1503,10 +1541,18 @@ CTL_RO_GEN(stats_arenas_i_purged, ctl_stats.arenas[mib[2]].astats.purged, const ctl_node_t * stats_arenas_i_index(const size_t *mib, size_t miblen, size_t i) { + const ctl_node_t * ret; - if (ctl_stats.arenas[i].initialized == false) - return (NULL); - return (super_stats_arenas_i_node); + malloc_mutex_lock(&ctl_mtx); + if (ctl_stats.arenas[i].initialized == false) { + ret = NULL; + goto RETURN; + } + + ret = super_stats_arenas_i_node; +RETURN: + malloc_mutex_unlock(&ctl_mtx); + return (ret); } #ifdef JEMALLOC_STATS @@ -1528,6 +1574,7 @@ swap_prezeroed_ctl(const size_t *mib, size_t miblen, void *oldp, { int ret; + malloc_mutex_lock(&ctl_mtx); if (swap_enabled) { READONLY(); } else { @@ -1545,6 +1592,7 @@ swap_prezeroed_ctl(const size_t *mib, size_t miblen, void *oldp, ret = 0; RETURN: + malloc_mutex_unlock(&ctl_mtx); return (ret); } @@ -1556,6 +1604,7 @@ swap_fds_ctl(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp, { int ret; + malloc_mutex_lock(&ctl_mtx); if (swap_enabled) { READONLY(); } else if (newp != NULL) { @@ -1586,6 +1635,7 @@ swap_fds_ctl(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp, ret = 0; RETURN: + malloc_mutex_unlock(&ctl_mtx); return (ret); } #endif From aee7fd2b70050fb434f2c9f52153194de73dc051 Mon Sep 17 00:00:00 2001 From: Jason Evans Date: Wed, 24 Nov 2010 22:00:02 -0800 Subject: [PATCH 3/6] Convert man page from roff to DocBook. Convert the man page source from roff to DocBook, and generate html and roff output. Modify the build system such that the documentation can be built as part of the release process, so that users need not have DocBook tools installed. --- .gitignore | 5 + jemalloc/INSTALL | 16 +- jemalloc/Makefile.in | 45 +- jemalloc/configure.ac | 103 +- jemalloc/doc/html.xsl.in | 4 + jemalloc/doc/jemalloc.3.in | 1688 -------------------------- jemalloc/doc/jemalloc.xml.in | 2223 ++++++++++++++++++++++++++++++++++ jemalloc/doc/manpages.xsl.in | 4 + jemalloc/doc/stylesheet.xsl | 7 + 9 files changed, 2319 insertions(+), 1776 deletions(-) create mode 100644 jemalloc/doc/html.xsl.in delete mode 100644 jemalloc/doc/jemalloc.3.in create mode 100644 jemalloc/doc/jemalloc.xml.in create mode 100644 jemalloc/doc/manpages.xsl.in create mode 100644 jemalloc/doc/stylesheet.xsl diff --git a/.gitignore b/.gitignore index b4681866..f3ee5ae0 100644 --- a/.gitignore +++ b/.gitignore @@ -3,6 +3,10 @@ /jemalloc/config.log /jemalloc/config.status /jemalloc/configure +/jemalloc/doc/html.xsl +/jemalloc/doc/manpages.xsl +/jemalloc/doc/jemalloc.xml +/jemalloc/doc/jemalloc.html /jemalloc/doc/jemalloc.3 /jemalloc/lib/ /jemalloc/Makefile @@ -13,4 +17,5 @@ /jemalloc/src/*.[od] /jemalloc/test/*.[od] /jemalloc/test/*.out +/jemalloc/test/[a-z]* /jemalloc/VERSION diff --git a/jemalloc/INSTALL b/jemalloc/INSTALL index fafd7883..b77ebfd4 100644 --- a/jemalloc/INSTALL +++ b/jemalloc/INSTALL @@ -132,8 +132,11 @@ any of the following arguments (not a definitive list) to 'configure': --disable-tls Disable thread-local storage (TLS), which allows for fast access to thread-local variables via the __thread keyword. If TLS is available, - jemalloc uses it for several purposes. Note that disabling TLS implies - --disable-tcache. + jemalloc uses it for several purposes. + +--with-xslroot= + Specify where to find DocBook XSL stylesheets when building the + documentation. The following environment variables (not a definitive list) impact configure's behavior: @@ -172,7 +175,7 @@ To install only parts of jemalloc, use the following targets: install_bin install_include install_lib - install_man + install_doc To clean up build results to varying degrees, use the following make targets: @@ -232,11 +235,12 @@ directory, issue configuration and build commands: === Documentation ============================================================== -The manual page that the configure script generates can be manually formatted +The manual page is generated in both html and roff formats. Any web browser +can be used to view the html manual. The roff manual page can be formatted prior to installation via any of the following commands: - nroff -man -man-ext -t doc/jemalloc.3 + nroff -man -t doc/jemalloc.3 - groff -man -man-ext -t -Tps doc/jemalloc.3 | ps2pdf - doc/jemalloc.3.pdf + groff -man -t -Tps doc/jemalloc.3 | ps2pdf - doc/jemalloc.3.pdf (cd doc; groff -man -man-ext -t -Thtml jemalloc.3 > jemalloc.3.html) diff --git a/jemalloc/Makefile.in b/jemalloc/Makefile.in index 46eddf46..335f6f6d 100644 --- a/jemalloc/Makefile.in +++ b/jemalloc/Makefile.in @@ -15,6 +15,7 @@ DESTDIR = BINDIR := $(DESTDIR)@BINDIR@ INCLUDEDIR := $(DESTDIR)@INCLUDEDIR@ LIBDIR := $(DESTDIR)@LIBDIR@ +DATADIR := $(DESTDIR)@DATADIR@ MANDIR := $(DESTDIR)@MANDIR@ # Build parameters. @@ -58,15 +59,34 @@ DSOS := @objroot@lib/libjemalloc@install_suffix@.$(SO).$(REV) \ @objroot@lib/libjemalloc@install_suffix@.$(SO) \ @objroot@lib/libjemalloc@install_suffix@_pic.a MAN3 := @objroot@doc/jemalloc@install_suffix@.3 +DOCS_XML := @objroot@doc/jemalloc@install_suffix@.xml +DOCS_HTML := $(DOCS_XML:@objroot@%.xml=@srcroot@%.html) +DOCS_MAN3 := $(DOCS_XML:@objroot@%.xml=@srcroot@%.3) +DOCS := $(DOCS_HTML) $(DOCS_MAN3) CTESTS := @srcroot@test/allocated.c @srcroot@test/allocm.c \ @srcroot@test/posix_memalign.c \ @srcroot@test/rallocm.c @srcroot@test/thread_arena.c -.PHONY: all dist install check clean distclean relclean +.PHONY: all dist doc_html doc_man doc +.PHONY: install_bin install_include install_lib +.PHONY: install_html install_man install_doc install +.PHONY: tests check clean distclean relclean # Default target. all: $(DSOS) +dist: doc + +@srcroot@doc/%.html : @objroot@doc/%.xml @srcroot@doc/stylesheet.xsl @objroot@doc/html.xsl + @XSLTPROC@ -o $@ @objroot@doc/html.xsl $< + +@srcroot@doc/%.3 : @objroot@doc/%.xml @srcroot@doc/stylesheet.xsl @objroot@doc/manpages.xsl + @XSLTPROC@ -o $@ @objroot@doc/manpages.xsl $< + +doc_html: $(DOCS_HTML) +doc_man: $(DOCS_MAN3) +doc: $(DOCS) + # # Include generated dependency files. # @@ -123,14 +143,23 @@ install_lib: $(DSOS) ln -sf libjemalloc@install_suffix@.$(SO).$(REV) $(LIBDIR)/libjemalloc@install_suffix@.$(SO) install -m 755 @objroot@lib/libjemalloc@install_suffix@_pic.a $(LIBDIR) -install_man: - install -d $(MANDIR)/man3 - @for m in $(MAN3); do \ - echo "install -m 644 $$m $(MANDIR)/man3"; \ - install -m 644 $$m $(MANDIR)/man3; \ +install_html: + install -d $(DATADIR)/doc/jemalloc@install_suffix@ + @for d in $(DOCS_HTML); do \ + echo "install -m 644 $$d $(DATADIR)/doc/jemalloc@install_suffix@"; \ + install -m 644 $$d $(DATADIR)/doc/jemalloc@install_suffix@; \ done -install: install_bin install_include install_lib install_man +install_man: + install -d $(MANDIR)/man3 + @for d in $(DOCS_MAN3); do \ + echo "install -m 644 $$d $(MANDIR)/man3"; \ + install -m 644 $$d $(MANDIR)/man3; \ +done + +install_doc: install_html install_man + +install: install_bin install_include install_lib install_doc tests: $(CTESTS:@srcroot@%.c=@objroot@%) @@ -182,6 +211,8 @@ distclean: clean relclean: distclean rm -f @objroot@configure rm -f @srcroot@VERSION + rm -f $(DOCS_HTML) + rm -f $(DOCS_MAN3) #=============================================================================== # Re-configuration rules. diff --git a/jemalloc/configure.ac b/jemalloc/configure.ac index 0ed13739..7aea6a87 100644 --- a/jemalloc/configure.ac +++ b/jemalloc/configure.ac @@ -80,6 +80,19 @@ MANDIR=`eval echo $mandir` MANDIR=`eval echo $MANDIR` AC_SUBST([MANDIR]) +dnl Support for building documentation. +AC_PATH_PROG([XSLTPROC], [xsltproc], , [$PATH]) +AC_ARG_WITH([xslroot], + [AS_HELP_STRING([--with-xslroot=], [XSL stylesheet root path])], +if test "x$with_xslroot" = "xno" ; then + XSLROOT="/usr/share/xml/docbook/stylesheet/docbook-xsl" +else + XSLROOT="${with_xslroot}" +fi, + XSLROOT="/usr/share/xml/docbook/stylesheet/docbook-xsl" +) +AC_SUBST([XSLROOT]) + dnl If CFLAGS isn't defined, set CFLAGS to something reasonable. Otherwise, dnl just prevent autoconf from molesting CFLAGS. CFLAGS=$CFLAGS @@ -275,17 +288,26 @@ AC_ARG_WITH([install_suffix], install_suffix="$INSTALL_SUFFIX" AC_SUBST([install_suffix]) -cfgoutputs_in="${srcroot}Makefile.in ${srcroot}doc/jemalloc.3.in" +cfgoutputs_in="${srcroot}Makefile.in" +cfgoutputs_in="${cfgoutputs_in} ${srcroot}doc/html.xsl.in" +cfgoutputs_in="${cfgoutputs_in} ${srcroot}doc/manpages.xsl.in" +cfgoutputs_in="${cfgoutputs_in} ${srcroot}doc/jemalloc.xml.in" cfgoutputs_in="${cfgoutputs_in} ${srcroot}include/jemalloc/jemalloc.h.in" cfgoutputs_in="${cfgoutputs_in} ${srcroot}include/jemalloc/internal/jemalloc_internal.h.in" cfgoutputs_in="${cfgoutputs_in} ${srcroot}test/jemalloc_test.h.in" -cfgoutputs_out="Makefile doc/jemalloc${install_suffix}.3" +cfgoutputs_out="Makefile" +cfgoutputs_out="${cfgoutputs_out} doc/html.xsl" +cfgoutputs_out="${cfgoutputs_out} doc/manpages.xsl" +cfgoutputs_out="${cfgoutputs_out} doc/jemalloc${install_suffix}.xml" cfgoutputs_out="${cfgoutputs_out} include/jemalloc/jemalloc${install_suffix}.h" cfgoutputs_out="${cfgoutputs_out} include/jemalloc/internal/jemalloc_internal.h" cfgoutputs_out="${cfgoutputs_out} test/jemalloc_test.h" -cfgoutputs_tup="Makefile doc/jemalloc${install_suffix}.3:doc/jemalloc.3.in" +cfgoutputs_tup="Makefile" +cfgoutputs_tup="${cfgoutputs_tup} doc/html.xsl:doc/html.xsl.in" +cfgoutputs_tup="${cfgoutputs_tup} doc/manpages.xsl:doc/manpages.xsl.in" +cfgoutputs_tup="${cfgoutputs_tup} doc/jemalloc${install_suffix}.xml:doc/jemalloc.xml.in" cfgoutputs_tup="${cfgoutputs_tup} include/jemalloc/jemalloc${install_suffix}.h:include/jemalloc/jemalloc.h.in" cfgoutputs_tup="${cfgoutputs_tup} include/jemalloc/internal/jemalloc_internal.h" cfgoutputs_tup="${cfgoutputs_tup} test/jemalloc_test.h:test/jemalloc_test.h.in" @@ -329,15 +351,6 @@ if test "x$enable_debug" = "x1" ; then AC_DEFINE([JEMALLOC_IVSALLOC], [ ]) fi AC_SUBST([enable_debug]) -if test "x$enable_debug" = "x0" ; then - roff_debug=".\\\" " - roff_no_debug="" -else - roff_debug="" - roff_no_debug=".\\\" " -fi -AC_SUBST([roff_debug]) -AC_SUBST([roff_no_debug]) dnl Only optimize if not debugging. if test "x$enable_debug" = "x0" -a "x$no_CFLAGS" = "xyes" ; then @@ -369,12 +382,6 @@ if test "x$enable_stats" = "x1" ; then AC_DEFINE([JEMALLOC_STATS], [ ]) fi AC_SUBST([enable_stats]) -if test "x$enable_stats" = "x0" ; then - roff_stats=".\\\" " -else - roff_stats="" -fi -AC_SUBST([roff_stats]) dnl Do not enable profiling by default. AC_ARG_ENABLE([prof], @@ -438,15 +445,6 @@ if test "x$enable_prof" = "x1" ; then fi fi AC_SUBST([enable_prof]) -if test "x$enable_prof" = "x0" ; then - roff_prof=".\\\" " - roff_no_prof="" -else - roff_prof="" - roff_no_prof=".\\\" " -fi -AC_SUBST([roff_prof]) -AC_SUBST([roff_no_prof]) dnl If libunwind isn't enabled, try to use libgcc rather than gcc intrinsics dnl for backtracing. @@ -478,15 +476,6 @@ if test "x$enable_tiny" = "x1" ; then AC_DEFINE([JEMALLOC_TINY], [ ]) fi AC_SUBST([enable_tiny]) -if test "x$enable_tiny" = "x0" ; then - roff_tiny=".\\\" " - roff_no_tiny="" -else - roff_tiny="" - roff_no_tiny=".\\\" " -fi -AC_SUBST([roff_tiny]) -AC_SUBST([roff_no_tiny]) dnl Enable thread-specific caching by default. AC_ARG_ENABLE([tcache], @@ -503,15 +492,6 @@ if test "x$enable_tcache" = "x1" ; then AC_DEFINE([JEMALLOC_TCACHE], [ ]) fi AC_SUBST([enable_tcache]) -if test "x$enable_tcache" = "x0" ; then - roff_tcache=".\\\" " - roff_no_tcache="" -else - roff_tcache="" - roff_no_tcache=".\\\" " -fi -AC_SUBST([roff_tcache]) -AC_SUBST([roff_no_tcache]) dnl Do not enable mmap()ped swap files by default. AC_ARG_ENABLE([swap], @@ -528,12 +508,6 @@ if test "x$enable_swap" = "x1" ; then AC_DEFINE([JEMALLOC_SWAP], [ ]) fi AC_SUBST([enable_swap]) -if test "x$enable_swap" = "x0" ; then - roff_swap=".\\\" " -else - roff_swap="" -fi -AC_SUBST([roff_swap]) dnl Do not enable allocation from DSS by default. AC_ARG_ENABLE([dss], @@ -550,12 +524,6 @@ if test "x$enable_dss" = "x1" ; then AC_DEFINE([JEMALLOC_DSS], [ ]) fi AC_SUBST([enable_dss]) -if test "x$enable_dss" = "x0" ; then - roff_dss=".\\\" " -else - roff_dss="" -fi -AC_SUBST([roff_dss]) dnl Do not support the junk/zero filling option by default. AC_ARG_ENABLE([fill], @@ -572,12 +540,6 @@ if test "x$enable_fill" = "x1" ; then AC_DEFINE([JEMALLOC_FILL], [ ]) fi AC_SUBST([enable_fill]) -if test "x$enable_fill" = "x0" ; then - roff_fill=".\\\" " -else - roff_fill="" -fi -AC_SUBST([roff_fill]) dnl Do not support the xmalloc option by default. AC_ARG_ENABLE([xmalloc], @@ -594,12 +556,6 @@ if test "x$enable_xmalloc" = "x1" ; then AC_DEFINE([JEMALLOC_XMALLOC], [ ]) fi AC_SUBST([enable_xmalloc]) -if test "x$enable_xmalloc" = "x0" ; then - roff_xmalloc=".\\\" " -else - roff_xmalloc="" -fi -AC_SUBST([roff_xmalloc]) dnl Do not support the SYSV option by default. AC_ARG_ENABLE([sysv], @@ -616,12 +572,6 @@ if test "x$enable_sysv" = "x1" ; then AC_DEFINE([JEMALLOC_SYSV], [ ]) fi AC_SUBST([enable_sysv]) -if test "x$enable_sysv" = "x0" ; then - roff_sysv=".\\\" " -else - roff_sysv="" -fi -AC_SUBST([roff_sysv]) dnl Do not determine page shift at run time by default. AC_ARG_ENABLE([dynamic_page_shift], @@ -828,6 +778,9 @@ AC_MSG_RESULT([LDFLAGS : ${LDFLAGS}]) AC_MSG_RESULT([LIBS : ${LIBS}]) AC_MSG_RESULT([RPATH_EXTRA : ${RPATH_EXTRA}]) AC_MSG_RESULT([]) +AC_MSG_RESULT([XSLTPROC : ${XSLTPROC}]) +AC_MSG_RESULT([XSLROOT : ${XSLROOT}]) +AC_MSG_RESULT([]) AC_MSG_RESULT([PREFIX : ${PREFIX}]) AC_MSG_RESULT([BINDIR : ${BINDIR}]) AC_MSG_RESULT([INCLUDEDIR : ${INCLUDEDIR}]) diff --git a/jemalloc/doc/html.xsl.in b/jemalloc/doc/html.xsl.in new file mode 100644 index 00000000..a91d9746 --- /dev/null +++ b/jemalloc/doc/html.xsl.in @@ -0,0 +1,4 @@ + + + + diff --git a/jemalloc/doc/jemalloc.3.in b/jemalloc/doc/jemalloc.3.in deleted file mode 100644 index d0129080..00000000 --- a/jemalloc/doc/jemalloc.3.in +++ /dev/null @@ -1,1688 +0,0 @@ -.\" 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 November 5, 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" "size_t *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 . diff --git a/jemalloc/doc/jemalloc.xml.in b/jemalloc/doc/jemalloc.xml.in new file mode 100644 index 00000000..6951160e --- /dev/null +++ b/jemalloc/doc/jemalloc.xml.in @@ -0,0 +1,2223 @@ + + + + + + + User Manual + jemalloc + @jemalloc_version@ + + + Jason + Evans + Author + + + + + JEMALLOC + 3 + + + jemalloc + jemalloc + + general purpose memory allocation functions + + + LIBRARY + This manual describes jemalloc @jemalloc_version@. More information + can be found at the jemalloc website. + + + SYNOPSIS + + #include <stdlib.h> +#include <jemalloc/jemalloc.h> + + Standard API + + void *malloc + size_t size + + + void *calloc + size_t number + size_t size + + + int posix_memalign + void **ptr + size_t alignment + size_t size + + + void *realloc + void *ptr + size_t size + + + void free + void *ptr + + + + Non-standard API + + size_t malloc_usable_size + const void *ptr + + + void malloc_stats_print + void (*write_cb) + void *, const char * + + void *cbopaque + const char *opts + + + int mallctl + const char *name + void *oldp + size_t *oldlenp + void *newp + size_t newlen + + + int mallctlnametomib + const char *name + size_t *mibp + size_t *miblenp + + + int mallctlbymib + const size_t *mib + size_t miblen + void *oldp + size_t *oldlenp + void *newp + size_t newlen + + + void (*malloc_message) + void *cbopaque + const char *s + + const char *malloc_conf; + + + Experimental API + + int allocm + void **ptr + size_t *rsize + size_t size + int flags + + + int rallocm + void **ptr + size_t *rsize + size_t size + size_t extra + int flags + + + int sallocm + const void *ptr + size_t *rsize + int flags + + + int dallocm + void *ptr + int flags + + + + + + DESCRIPTION + + Standard API + + The malloc function allocates + size bytes of uninitialized memory. The allocated + space is suitably aligned (after possible pointer coercion) for storage + of any type of object. + + The calloc function allocates + space for number objects, each + size bytes in length. The result is identical to + calling malloc with an argument of + number * size, with the + exception that the allocated memory is explicitly initialized to zero + bytes. + + The posix_memalign function + allocates size bytes of memory such that the + allocation's base address is an even multiple of + alignment, and returns the allocation in the value + pointed to by ptr. The requested + alignment must be a power of 2 at least as large + as sizeof(void *). + + The realloc function changes the + size of the previously allocated memory referenced by + ptr to 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 ptr is freed and a pointer to the newly + allocated memory is returned. Note that + realloc may move the memory allocation, + resulting in a different return value than ptr. + If ptr is NULL, the + realloc function behaves identically to + malloc for the specified size. + + The free function causes the + allocated memory referenced by ptr to be made + available for future allocations. If ptr is + NULL, no action occurs. + + + Non-standard API + + The malloc_usable_size function + returns the usable size of the allocation pointed to by + ptr. The return value may be larger than the size + that was requested during allocation. The + malloc_usable_size function is not a + mechanism for in-place realloc; rather + it is provided solely as a tool for introspection purposes. Any + discrepancy between the requested allocation size and the size reported + by malloc_usable_size should not be + depended on, since such behavior is entirely implementation-dependent. + + + The malloc_stats_print function + writes human-readable summary statistics via the + write_cb callback function pointer and + cbopaque data passed to + write_cb, or + malloc_message if + write_cb is NULL. This + function can be called repeatedly. General information that never + changes during execution can be omitted by specifying "g" as a character + within the opts string. Note that + malloc_message uses the + mallctl* functions internally, so + inconsistent statistics can be reported if multiple threads use these + functions simultaneously. If is + specified during configuration, “m” and “a” can + be specified to omit merged arena and per arena statistics, respectively; + “b” and “l” can be specified to omit per size + class statistics for bins and large objects, respectively. Unrecognized + characters are silently ignored. Note that thread caching may prevent + some statistics from being completely up to date, since extra locking + would be required to merge counters that track thread cache operations. + + + The mallctl function provides a + general interface for introspecting the memory allocator, as well as + setting modifiable parameters and triggering actions. The + period-separated name argument specifies a + location in a tree-structured namespace; see the section for + documentation on the tree contents. To read a value, pass a pointer via + oldp to adequate space to contain the value, and a + pointer to its length via oldlenp; otherwise pass + NULL and NULL. Similarly, to + write a value, pass a pointer to the value via + newp, and its length via + newlen; otherwise pass NULL + and 0. + + The 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 “Management Information Base” (MIB) that can be passed + repeatedly to mallctlbymib. Upon + successful return from mallctlnametomib, + mibp contains an array of + *miblenp integers, where + *miblenp is the lesser of the number of components + in name and the input value of + *miblenp. Thus it is possible to pass a + *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 + arenas.bin.2.size), + the corresponding MIB component will always be that integer. Therefore, + it is legitimate to construct code like the following: + + + Experimental API + The experimental API is subject to change or removal without regard + for backward compatibility. + + The allocm, + rallocm, + sallocm, and + dallocm functions all have a + 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: + + + ALLOCM_LG_ALIGN(la) + + + Align the memory allocation to start at an address + that is a multiple of (1 << + la). This macro does not validate + that la is within the valid + range. + + + ALLOCM_ALIGN(a) + + + Align the memory allocation to start at an address + that is a multiple of a, where + a is a power of two. This macro does not + validate that a is a power of 2. + + + + 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. + + + ALLOCM_NO_MOVE + + For reallocation, fail rather than moving the + object. This constraint can apply to both growth and + shrinkage. + + + + + The allocm function allocates at + least size bytes of memory, sets + *ptr to the base address of the allocation, and + sets *rsize to the real size of the allocation if + rsize is not NULL. + + The rallocm function resizes the + allocation at *ptr to be at least + size bytes, sets *ptr to + the base address of the allocation if it moved, and sets + *rsize to the real size of the allocation if + rsize is not NULL. If + extra is non-zero, an attempt is made to resize + the allocation to be at least size + + extra) bytes, though inability to allocate + the extra byte(s) will not by itself result in failure. Behavior is + undefined if (size + + extra > + SIZE_T_MAX). + + The sallocm function sets + *rsize to the real size of the allocation. + + The dallocm function causes the + memory referenced by ptr to be made available for + future allocations. + + + + 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. + + The string pointed to by the global variable + malloc_conf, the “name” of the file + referenced by the symbolic link named /etc/malloc.conf, and the value of the + environment variable MALLOC_CONF, will be interpreted, in + that order, from left to right as options. + + An options string is a comma-separated list of option:value pairs. + There is one key corresponding to each opt.* mallctl (see the section for options + documentation). For example, abort:true,narenas:1 sets + the opt.abort and 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. + + + IMPLEMENTATION NOTES + Traditionally, allocators have used + sbrk + 2 to obtain memory, which is + suboptimal for several reasons, including race conditions, increased + fragmentation, and artificial limitations on maximum usable memory. If + is specified during configuration, this + allocator uses both sbrk + 2 and + mmap + 2, in that order of preference; + otherwise only mmap + 2 is used. + + 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. + + In addition to multiple arenas, unless + is specified during configuration, this + allocator supports thread-specific caching for small and large objects, in + order to make it possible to completely avoid synchronization for most + allocation requests. Such caching allows very fast allocation in the + common case, but it increases memory usage and fragmentation, since a + bounded number of objects can remain allocated in each thread cache. + + Memory is conceptually broken into equal-sized chunks, where the + chunk size is a power of two that is greater than the page size. Chunks + are always aligned to multiples of the chunk size. This alignment makes it + possible to find metadata for user objects very quickly. + + User objects are broken into three categories according to size: + small, large, and huge. 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. + + 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. + + Small objects are managed in groups by page runs. Each run maintains + a frontier and free list to track which regions are in use. Unless + is specified during configuration, + allocation requests that are no more than half the quantum (8 or 16, + depending on architecture) are rounded up to the nearest power of two that + is at least sizeof(void *). + Allocation requests that are more than half the quantum, but no more than + the minimum cacheline-multiple size class (see the opt.lg_qspace_max + option) are rounded up to the nearest multiple of the quantum. Allocation + requests that are more than the minimum cacheline-multiple size class, but + no more than the minimum subpage-multiple size class (see the 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 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. + + 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. + + Assuming 4 MiB chunks, 4 KiB pages, and a 16-byte quantum on a 64-bit + system, the size classes in each category are as shown in . + + + Size classes + + + + + + + Category + Subcategory + Size + + + + + Small + Tiny + [8] + + + Quantum-spaced + [16, 32, 48, ..., 128] + + + Cacheline-spaced + [192, 256, 320, ..., 512] + + + Subpage-spaced + [768, 1024, 1280, ..., 3840] + + + Large + [4 KiB, 8 KiB, 12 KiB, ..., 4072 KiB] + + + Huge + [4 MiB, 8 MiB, 12 MiB, ...] + + + +
+ + + MALLCTL NAMESPACE + The following names are defined in the namespace accessible via the + mallctl* functions. Value types are + specified in parentheses, their readable/writable statuses are encoded as + rw, r-, -w, or + --, and required build configuration flags follow, if + any. 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. In the case of stats.arenas.<i>.*, + <i> equal to arenas.narenas can be + used to access the summation of statistics from all arenas. Take special + note of the epoch mallctl, + which controls refreshing of cached dynamic statistics. + + + + + version + (const char *) + r- + + Return the jemalloc version string. + + + + + epoch + (uint64_t) + rw + + If a value is passed in, refresh the data from which + the mallctl* functions report values, + and increment the epoch. Return the current epoch. This is useful for + detecting whether another thread caused a refresh. + + + + + config.debug + (bool) + r- + + was specified during + build configuration. + + + + + config.dss + (bool) + r- + + was specified during + build configuration. + + + + + config.dynamic_page_shift + (bool) + r- + + was + specified during build configuration. + + + + + config.fill + (bool) + r- + + was specified during + build configuration. + + + + + config.lazy_lock + (bool) + r- + + was specified + during build configuration. + + + + + config.prof + (bool) + r- + + was specified during + build configuration. + + + + + config.prof_libgcc + (bool) + r- + + was not + specified during build configuration. + + + + + config.prof_libunwind + (bool) + r- + + was specified + during build configuration. + + + + + config.stats + (bool) + r- + + was specified during + build configuration. + + + + + config.swap + (bool) + r- + + was specified during + build configuration. + + + + + config.sysv + (bool) + r- + + was specified during + build configuration. + + + + + config.tcache + (bool) + r- + + was not specified + during build configuration. + + + + + config.tiny + (bool) + r- + + was not specified + during build configuration. + + + + + config.tls + (bool) + r- + + was not specified during + build configuration. + + + + + config.xmalloc + (bool) + r- + + was specified during + build configuration. + + + + + opt.abort + (bool) + r- + + Abort-on-warning enabled/disabled. If true, most + warnings are fatal. The process will call + abort + 3 in these cases. This option is + disabled by default unless is + specified during configuration, in which case it is enabled by default. + + + + + + opt.lg_qspace_max + (size_t) + r- + + 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). + + + + + opt.lg_cspace_max + (size_t) + r- + + 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). + + + + + opt.lg_chunk + (size_t) + r- + + Virtual memory chunk size (log base 2). The default + chunk size is 4 MiB (2^22). + + + + + opt.narenas + (size_t) + r- + + 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. + + + + + opt.lg_dirty_mult + (ssize_t) + r- + + 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 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. + + + + + opt.stats_print + (bool) + r- + + Enable/disable statistics printing at exit. If + enabled, the malloc_stats_print + function is called at program exit via an + atexit + 3 function. If + is specified during configuration, this + has the potential to cause deadlock for a multi-threaded process that + exits while one or more threads are executing in the memory allocation + functions. Therefore, this option should only be used with care; it is + primarily intended as a performance tuning aid during application + development. This option is disabled by default. + + + + + opt.junk + (bool) + r- + [] + + Junk filling enabled/disabled. If enabled, each byte + of uninitialized allocated memory will be initialized to + 0xa5. All deallocated memory will be initialized to + 0x5a. This is intended for debugging and will + impact performance negatively. This option is disabled by default + unless is specified during + configuration, in which case it is enabled by default. + + + + + opt.zero + (bool) + r- + [] + + Zero filling enabled/disabled. If enabled, each byte + of uninitialized allocated memory will be initialized to 0. Note that + this initialization only happens once for each byte, so + realloc and + rallocm calls do not zero memory that + was previously allocated. This is intended for debugging and will + impact performance negatively. This option is disabled by default. + + + + + + opt.sysv + (bool) + r- + [] + + If enabled, attempting to allocate zero bytes will + return a NULL pointer instead of a valid pointer. + (The default behavior is to make a minimal allocation and return a + pointer to it.) This option is provided for System V compatibility. + This option is incompatible with the opt.xmalloc option. + This option is disabled by default. + + + + + opt.xmalloc + (bool) + r- + [] + + Abort-on-out-of-memory enabled/disabled. If enabled, + rather than returning failure for any allocation function, display a + diagnostic message on STDERR_FILENO and cause the + program to drop core (using + abort + 3). If an application is + designed to depend on this behavior, set the option at compile time by + including the following in the source code: + + This option is disabled by default. + + + + + opt.tcache + (bool) + r- + [] + + Thread-specific caching enabled/disabled. When there + are multiple threads, each thread uses a thread-specific cache for + objects up to a certain size. Thread-specific caching allows many + allocations to be satisfied without performing any thread + synchronization, at the cost of increased memory use. See the + opt.lg_tcache_gc_sweep + and opt.lg_tcache_max + options for related tuning information. This option is enabled by + default. + + + + + opt.lg_tcache_gc_sweep + (ssize_t) + r- + [] + + Approximate interval (log base 2) between full + thread-specific cache garbage collection sweeps, counted in terms of + thread-specific cache allocation/deallocation events. Garbage + collection is actually performed incrementally, one size class at a + time, in order to avoid large collection pauses. The default sweep + interval is 8192 (2^13); setting this option to -1 will disable garbage + collection. + + + + + opt.lg_tcache_max + (size_t) + r- + [] + + Maximum size class (log base 2) to cache in the + thread-specific cache. At a minimum, all small size classes are + cached, and at a maximum all large size classes are cached. The + default maximum is 32 KiB (2^15). + + + + + opt.prof + (bool) + r- + [] + + Memory profiling enabled/disabled. If enabled, profile + memory allocation activity, and use an + atexit + 3 function to dump final memory + usage to a file named according to the pattern + <prefix>.<pid>.<seq>.f.heap, + where <prefix> is controlled by the opt.prof_prefix + option. See the opt.lg_prof_bt_max + option for backtrace depth control. See the opt.prof_active + option for on-the-fly activation/deactivation. See the opt.lg_prof_sample + option for probabilistic sampling control. See the opt.prof_accum + option for control of cumulative sample reporting. See the opt.lg_prof_tcmax + option for control of per thread backtrace caching. See the opt.lg_prof_interval + option for information on interval-triggered profile dumping, and the + opt.prof_gdump + option for information on high-water-triggered profile dumping. + Profile output is compatible with the included pprof + Perl script, which originates from the google-perftools + package. + + + + + opt.prof_prefix + (const char *) + r- + [] + + Filename prefix for profile dumps. If the prefix is + set to the empty string, no automatic dumps will occur; this is + primarily useful for disabling the automatic final heap dump (which + also disables leak reporting, if enabled). The default prefix is + jeprof. + + + + + opt.lg_prof_bt_max + (size_t) + r- + [] + + Maximum backtrace depth (log base 2) when profiling + memory allocation activity. The default is 128 (2^7). + + + + + opt.prof_active + (bool) + r- + [] + + Profiling activated/deactivated. This is a secondary + control mechanism that makes it possible to start the application with + profiling enabled (see the opt.prof option) but + inactive, then toggle profiling at any time during program execution + with the prof.active mallctl. + This option is enabled by default. + + + + + opt.lg_prof_sample + (ssize_t) + r- + [] + + Average interval (log base 2) between allocation + samples, as measured in bytes of allocation activity. Increasing the + sampling interval decreases profile fidelity, but also decreases the + computational overhead. The default sample interval is 1 (2^0) (i.e. + all allocations are sampled). + + + + + opt.prof_accum + (bool) + r- + [] + + Reporting of cumulative object/byte counts in profile + dumps enabled/disabled. If this option is enabled, every unique + backtrace must be stored for the duration of execution. Depending on + the application, this can impose a large memory overhead, and the + cumulative counts are not always of interest. See the + opt.lg_prof_tcmax + option for control of per thread backtrace caching, which has important + interactions. This option is enabled by default. + + + + + opt.lg_prof_tcmax + (ssize_t) + r- + [] + + Maximum per thread backtrace cache (log base 2) used + for heap profiling. A backtrace can only be discarded if the + opt.prof_accum + option is disabled, and no thread caches currently refer to the + backtrace. Therefore, a backtrace cache limit should be imposed if the + intention is to limit how much memory is used by backtraces. By + default, no limit is imposed (encoded as -1). + + + + + + opt.lg_prof_interval + (ssize_t) + r- + [] + + Average interval (log base 2) between memory profile + dumps, as measured in bytes of allocation activity. The actual + interval between dumps may be sporadic because decentralized allocation + counters are used to avoid synchronization bottlenecks. Profiles are + dumped to files named according to the pattern + <prefix>.<pid>.<seq>.i<iseq>.heap, + where <prefix> is controlled by the + opt.prof_prefix + option. By default, interval-triggered profile dumping is disabled + (encoded as -1). + + + + + + opt.prof_gdump + (bool) + r- + [] + + Trigger a memory profile dump every time the total + virtual memory exceeds the previous maximum. Profiles are dumped to + files named according to the pattern + <prefix>.<pid>.<seq>.u<useq>.heap, + where <prefix> is controlled by the opt.prof_prefix + option. This option is disabled by default. + + + + + opt.prof_leak + (bool) + r- + [] + + Leak reporting enabled/disabled. If enabled, use an + atexit + 3 function to report memory leaks + detected by allocation sampling. See the + opt.lg_prof_bt_max + option for backtrace depth control. See the + opt.prof option for + information on analyzing heap profile output. This option is disabled + by default. + + + + + opt.overcommit + (bool) + r- + [] + + Over-commit enabled/disabled. If enabled, over-commit + memory as a side effect of using anonymous + mmap + 2 or + sbrk + 2 for virtual memory allocation. + In order for overcommit to be disabled, the swap.fds mallctl must have + been successfully written to. This option is enabled by + default. + + + + + tcache.flush + (void) + -- + [] + + Flush calling thread's tcache. This interface releases + all cached objects and internal data structures associated with the + calling thread's thread-specific cache. Ordinarily, this interface + need not be called, since automatic periodic incremental garbage + collection occurs, and the thread cache is automatically discarded when + a thread exits. However, garbage collection is triggered by allocation + activity, so it is possible for a thread that stops + allocating/deallocating to retain its cache indefinitely, in which case + the developer may find manual flushing useful. + + + + + thread.arena + (unsigned) + rw + + Get or set the arena associated with the calling + thread. The arena index must be less than the maximum number of arenas + (see the arenas.narenas + mallctl). If the specified arena was not initialized beforehand (see + the arenas.initialized + mallctl), it will be automatically initialized as a side effect of + calling this interface. + + + + + thread.allocated + (uint64_t) + r- + [] + + Get the total number of bytes ever allocated by the + calling thread. This counter has the potential to wrap around; it is + up to the application to appropriately interpret the counter in such + cases. + + + + + thread.deallocated + (uint64_t) + r- + [] + + Get the total number of bytes ever deallocated by the + calling thread. This counter has the potential to wrap around; it is + up to the application to appropriately interpret the counter in such + cases. + + + + + arenas.narenas + (unsigned) + r- + + Maximum number of arenas. + + + + + arenas.initialized + (bool *) + r- + + An array of arenas.narenas + booleans. Each boolean indicates whether the corresponding arena is + initialized. + + + + + arenas.quantum + (size_t) + r- + + Quantum size. + + + + + arenas.cacheline + (size_t) + r- + + Assumed cacheline size. + + + + + arenas.subpage + (size_t) + r- + + Subpage size class interval. + + + + + arenas.pagesize + (size_t) + r- + + Page size. + + + + + arenas.chunksize + (size_t) + r- + + Chunk size. + + + + + arenas.tspace_min + (size_t) + r- + + Minimum tiny size class. Tiny size classes are powers + of two. + + + + + arenas.tspace_max + (size_t) + r- + + Maximum tiny size class. Tiny size classes are powers + of two. + + + + + arenas.qspace_min + (size_t) + r- + + Minimum quantum-spaced size class. + + + + + arenas.qspace_max + (size_t) + r- + + Maximum quantum-spaced size class. + + + + + arenas.cspace_min + (size_t) + r- + + Minimum cacheline-spaced size class. + + + + + arenas.cspace_max + (size_t) + r- + + Maximum cacheline-spaced size class. + + + + + arenas.sspace_min + (size_t) + r- + + Minimum subpage-spaced size class. + + + + + arenas.sspace_max + (size_t) + r- + + Maximum subpage-spaced size class. + + + + + arenas.tcache_max + (size_t) + r- + [] + + Maximum thread-cached size class. + + + + + arenas.ntbins + (unsigned) + r- + + Number of tiny bin size classes. + + + + + arenas.nqbins + (unsigned) + r- + + Number of quantum-spaced bin size + classes. + + + + + arenas.ncbins + (unsigned) + r- + + Number of cacheline-spaced bin size + classes. + + + + + arenas.nsbins + (unsigned) + r- + + Number of subpage-spaced bin size + classes. + + + + + arenas.nbins + (unsigned) + r- + + Total number of bin size classes. + + + + + arenas.nhbins + (unsigned) + r- + [] + + Total number of thread cache bin size + classes. + + + + + arenas.bin.<i>.size + (size_t) + r- + + Maximum size supported by size class. + + + + + arenas.bin.<i>.nregs + (uint32_t) + r- + + Number of regions per page run. + + + + + arenas.bin.<i>.run_size + (size_t) + r- + + Number of bytes per page run. + + + + + arenas.nlruns + (size_t) + r- + + Total number of large size classes. + + + + + arenas.lrun.<i>.size + (size_t) + r- + + Maximum size supported by this large size + class. + + + + + arenas.purge + (unsigned) + -w + + Purge unused dirty pages for the specified arena, or + for all arenas if none is specified. + + + + + prof.active + (bool) + rw + [] + + Control whether sampling is currently active. See the + opt.prof_active + option for additional information. + + + + + + prof.dump + (const char *) + -w + [] + + Dump a memory profile to the specified file, or if NULL + is specified, to a file according to the pattern + <prefix>.<pid>.<seq>.m<mseq>.heap, + where <prefix> is controlled by the + opt.prof_prefix + option. + + + + + prof.interval + (uint64_t) + r- + [] + + Average number of bytes allocated between + inverval-based profile dumps. See the + opt.lg_prof_interval + option for additional information. + + + + + stats.allocated + (size_t) + r- + [] + + Total number of bytes allocated by the + application. + + + + + stats.active + (size_t) + r- + [] + + Total number of bytes in active pages allocated by the + application. This is a multiple of the page size, and greater than or + equal to stats.allocated. + + + + + + stats.mapped + (size_t) + r- + [] + + Total number of bytes in chunks mapped on behalf of the + application. This is a multiple of the chunk size, and is at least as + large as stats.active. This + does not include inactive chunks backed by swap files. his does not + include inactive chunks embedded in the DSS. + + + + + stats.chunks.current + (size_t) + r- + [] + + Total number of chunks actively mapped on behalf of the + application. This does not include inactive chunks backed by swap + files. This does not include inactive chunks embedded in the DSS. + + + + + + stats.chunks.total + (uint64_t) + r- + [] + + Cumulative number of chunks allocated. + + + + + stats.chunks.high + (size_t) + r- + [] + + Maximum number of active chunks at any time thus far. + + + + + + stats.huge.allocated + (size_t) + r- + [] + + Number of bytes currently allocated by huge objects. + + + + + + stats.huge.nmalloc + (uint64_t) + r- + [] + + Cumulative number of huge allocation requests. + + + + + + stats.huge.ndalloc + (uint64_t) + r- + [] + + Cumulative number of huge deallocation requests. + + + + + + stats.arenas.<i>.pactive + (size_t) + r- + + Number of pages in active runs. + + + + + stats.arenas.<i>.pdirty + (size_t) + r- + + Number of pages within unused runs that are potentially + dirty, and for which madvise... + MADV_DONTNEED or + similar has not been called. + + + + + stats.arenas.<i>.mapped + (size_t) + r- + [] + + Number of mapped bytes. + + + + + stats.arenas.<i>.npurge + (uint64_t) + r- + [] + + Number of dirty page purge sweeps performed. + + + + + + stats.arenas.<i>.nmadvise + (uint64_t) + r- + [] + + Number of madvise... + MADV_DONTNEED or + similar calls made to purge dirty pages. + + + + + stats.arenas.<i>.npurged + (uint64_t) + r- + [] + + Number of pages purged. + + + + + stats.arenas.<i>.small.allocated + (size_t) + r- + [] + + Number of bytes currently allocated by small objects. + + + + + + stats.arenas.<i>.small.nmalloc + (uint64_t) + r- + [] + + Cumulative number of allocation requests served by + small bins. + + + + + stats.arenas.<i>.small.ndalloc + (uint64_t) + r- + [] + + Cumulative number of small objects returned to bins. + + + + + + stats.arenas.<i>.small.nrequests + (uint64_t) + r- + [] + + Cumulative number of small allocation requests. + + + + + + stats.arenas.<i>.large.allocated + (size_t) + r- + [] + + Number of bytes currently allocated by large objects. + + + + + + stats.arenas.<i>.large.nmalloc + (uint64_t) + r- + [] + + Cumulative number of large allocation requests served + directly by the arena. + + + + + stats.arenas.<i>.large.ndalloc + (uint64_t) + r- + [] + + Cumulative number of large deallocation requests served + directly by the arena. + + + + + stats.arenas.<i>.large.nrequests + (uint64_t) + r- + [] + + Cumulative number of large allocation requests. + + + + + + stats.arenas.<i>.bins.<j>.allocated + (size_t) + r- + [] + + Current number of bytes allocated by + bin. + + + + + stats.arenas.<i>.bins.<j>.nmalloc + (uint64_t) + r- + [] + + Cumulative number of allocations served by bin. + + + + + + stats.arenas.<i>.bins.<j>.ndalloc + (uint64_t) + r- + [] + + Cumulative number of allocations returned to bin. + + + + + + stats.arenas.<i>.bins.<j>.nrequests + (uint64_t) + r- + [] + + Cumulative number of allocation + requests. + + + + + stats.arenas.<i>.bins.<j>.nfills + (uint64_t) + r- + [ ] + + Cumulative number of tcache fills. + + + + + stats.arenas.<i>.bins.<j>.nflushes + (uint64_t) + r- + [ ] + + Cumulative number of tcache flushes. + + + + + stats.arenas.<i>.bins.<j>.nruns + (uint64_t) + r- + [] + + Cumulative number of runs created. + + + + + stats.arenas.<i>.bins.<j>.nreruns + (uint64_t) + r- + [] + + Cumulative number of times the current run from which + to allocate changed. + + + + + stats.arenas.<i>.bins.<j>.highruns + (size_t) + r- + [] + + Maximum number of runs at any time thus far. + + + + + + stats.arenas.<i>.bins.<j>.curruns + (size_t) + r- + [] + + Current number of runs. + + + + + stats.arenas.<i>.lruns.<j>.nmalloc + (uint64_t) + r- + [] + + Cumulative number of allocation requests for this size + class served directly by the arena. + + + + + stats.arenas.<i>.lruns.<j>.ndalloc + (uint64_t) + r- + [] + + Cumulative number of deallocation requests for this + size class served directly by the arena. + + + + + stats.arenas.<i>.lruns.<j>.nrequests + (uint64_t) + r- + [] + + Cumulative number of allocation requests for this size + class. + + + + + stats.arenas.<i>.lruns.<j>.highruns + (size_t) + r- + [] + + Maximum number of runs at any time thus far for this + size class. + + + + + stats.arenas.<i>.lruns.<j>.curruns + (size_t) + r- + [] + + Current number of runs for this size class. + + + + + + swap.avail + (size_t) + r- + [] + + Number of swap file bytes that are currently not + associated with any chunk (i.e. mapped, but otherwise completely + unmanaged). + + + + + swap.prezeroed + (bool) + rw + [] + + If true, the allocator assumes that the swap file(s) + contain nothing but nil bytes. If this assumption is violated, + allocator behavior is undefined. This value becomes read-only after + swap.fds is + successfully written to. + + + + + swap.nfds + (size_t) + r- + [] + + Number of file descriptors in use for swap. + + + + + + swap.fds + (int *) + r- + [] + + When written to, the files associated with the + specified file descriptors are contiguously mapped via + mmap + 2. The resulting virtual memory + region is preferred over anonymous + mmap + 2 and + sbrk + 2 memory. Note that if a file's + size is not a multiple of the page size, it is automatically truncated + to the nearest page size multiple. See the + swap.prezeroed + mallctl for specifying that the files are pre-zeroed. + + + + + DEBUGGING MALLOC PROBLEMS + When debugging, it is a good idea to configure/build jemalloc with + the and + options, and recompile the program with suitable options and symbols for + debugger support. When so configured, jemalloc incorporates a wide variety + of run-time assertions that catch application errors such as double-free, + write-after-free, etc. + + Programs often accidentally depend on “uninitialized” + memory actually being filled with zero bytes. Junk filling + (see the opt.junk + option) tends to expose such bugs in the form of obviously incorrect + results and/or coredumps. Conversely, zero + filling (see the opt.zero option) eliminates + the symptoms of such bugs. Between these two options, it is usually + possible to quickly detect, diagnose, and eliminate such bugs. + + 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. + + + DIAGNOSTIC MESSAGES + If any of the memory allocation/deallocation functions detect an + error or warning condition, a message will be printed to file descriptor + STDERR_FILENO. Errors will result in the process + dumping core. If the opt.abort option is set, most + warnings are treated as errors. + + The 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 STDERR_FILENO file + descriptor is not suitable for this. + malloc_message takes the + cbopaque pointer argument that is + NULL unless overridden by the arguments in a call to + 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. + + All messages are prefixed by + “<jemalloc>: ”. + + + RETURN VALUES + + Standard API + The malloc and + calloc functions return a pointer to the + allocated memory if successful; otherwise a NULL + pointer is returned and errno is set to + ENOMEM. + + The posix_memalign function + returns the value 0 if successful; otherwise it returns an error value. + The posix_memalign function will fail + if: + + + EINVAL + + The alignment parameter is + not a power of 2 at least as large as + sizeof(void *). + + + + ENOMEM + + Memory allocation error. + + + + + The realloc function returns a + pointer, possibly identical to ptr, to the + allocated memory if successful; otherwise a NULL + pointer is returned, and errno is set to + ENOMEM if the error was the result of an + allocation failure. The realloc + function always leaves the original buffer intact when an error occurs. + + + The free function returns no + value. + + + Non-standard API + The malloc_usable_size function + returns the usable size of the allocation pointed to by + ptr. + + The mallctl, + mallctlnametomib, and + mallctlbymib functions return 0 on + success; otherwise they return an error value. The functions will fail + if: + + + EINVAL + + newp is not + NULL, and newlen is too + large or too small. Alternatively, *oldlenp + is too large or too small; in this case as much data as possible + are read despite the error. + + + ENOMEM + + *oldlenp is too short to + hold the requested value. + + + ENOENT + + name or + mib specifies an unknown/invalid + value. + + + EPERM + + Attempt to read or write void value, or attempt to + write read-only value. + + + EAGAIN + + A memory allocation failure + occurred. + + + EFAULT + + An interface with side effects failed in some way + not directly related to mallctl* + read/write processing. + + + + + + Experimental API + The allocm, + rallocm, + sallocm, and + dallocm functions return + ALLOCM_SUCCESS on success; otherwise they return an + error value. The allocm and + rallocm functions will fail if: + + + ALLOCM_ERR_OOM + + Out of memory. Insufficient contiguous memory was + available to service the allocation request. The + allocm function additionally sets + *ptr to NULL, whereas + the rallocm function leaves + *ptr unmodified. + + + The rallocm function will also + fail if: + + + ALLOCM_ERR_NOT_MOVED + + ALLOCM_NO_MOVE was specified, + but the reallocation request could not be serviced without moving + the object. + + + + + + + ENVIRONMENT + The following environment variable affects the execution of the + allocation functions: + + + MALLOC_CONF + + If the environment variable + MALLOC_CONF is set, the characters it contains + will be interpreted as options. + + + + + + EXAMPLES + To dump core whenever a problem occurs: + ln -s 'abort:true' /etc/malloc.conf + + To specify in the source a chunk size that is 16 MiB: + + + + SEE ALSO + madvise + 2, + mmap + 2, + sbrk + 2, + alloca + 3, + atexit + 3, + getpagesize + 3 + + + STANDARDS + The malloc, + calloc, + realloc, and + free functions conform to ISO/IEC + 9899:1990 (“ISO C90”). + + The posix_memalign function conforms + to IEEE Std 1003.1-2001 (“POSIX.1”). + + diff --git a/jemalloc/doc/manpages.xsl.in b/jemalloc/doc/manpages.xsl.in new file mode 100644 index 00000000..88b2626b --- /dev/null +++ b/jemalloc/doc/manpages.xsl.in @@ -0,0 +1,4 @@ + + + + diff --git a/jemalloc/doc/stylesheet.xsl b/jemalloc/doc/stylesheet.xsl new file mode 100644 index 00000000..4e334a86 --- /dev/null +++ b/jemalloc/doc/stylesheet.xsl @@ -0,0 +1,7 @@ + + ansi + + + "" + + From cfdc8cfbd626e83d38417bd8c73ac018b611e390 Mon Sep 17 00:00:00 2001 From: Jason Evans Date: Tue, 30 Nov 2010 16:50:58 -0800 Subject: [PATCH 4/6] Use mremap(2) for huge realloc(). If mremap(2) is available and supports MREMAP_FIXED, use it for huge realloc(). Initialize rtree later during bootstrapping, so that --enable-debug --enable-dss works. Fix a minor swap_avail stats bug. --- .gitignore | 2 + jemalloc/Makefile.in | 2 +- jemalloc/configure.ac | 10 +++ .../include/jemalloc/internal/chunk_dss.h | 1 + .../include/jemalloc/internal/chunk_swap.h | 1 + jemalloc/include/jemalloc/internal/huge.h | 2 +- .../jemalloc/internal/jemalloc_internal.h.in | 2 +- jemalloc/include/jemalloc/jemalloc_defs.h.in | 3 + jemalloc/src/chunk.c | 10 +-- jemalloc/src/chunk_dss.c | 16 +++++ jemalloc/src/chunk_swap.c | 24 ++++++- jemalloc/src/huge.c | 57 ++++++++++++++-- jemalloc/test/mremap.c | 67 +++++++++++++++++++ jemalloc/test/mremap.exp | 2 + 14 files changed, 182 insertions(+), 17 deletions(-) create mode 100644 jemalloc/test/mremap.c create mode 100644 jemalloc/test/mremap.exp diff --git a/.gitignore b/.gitignore index f3ee5ae0..d6fa8fd1 100644 --- a/.gitignore +++ b/.gitignore @@ -18,4 +18,6 @@ /jemalloc/test/*.[od] /jemalloc/test/*.out /jemalloc/test/[a-z]* +!/jemalloc/test/*.c +!/jemalloc/test/*.exp /jemalloc/VERSION diff --git a/jemalloc/Makefile.in b/jemalloc/Makefile.in index 335f6f6d..ee674b33 100644 --- a/jemalloc/Makefile.in +++ b/jemalloc/Makefile.in @@ -64,7 +64,7 @@ DOCS_HTML := $(DOCS_XML:@objroot@%.xml=@srcroot@%.html) DOCS_MAN3 := $(DOCS_XML:@objroot@%.xml=@srcroot@%.3) DOCS := $(DOCS_HTML) $(DOCS_MAN3) CTESTS := @srcroot@test/allocated.c @srcroot@test/allocm.c \ - @srcroot@test/posix_memalign.c \ + @srcroot@test/mremap.c @srcroot@test/posix_memalign.c \ @srcroot@test/rallocm.c @srcroot@test/thread_arena.c .PHONY: all dist doc_html doc_man doc diff --git a/jemalloc/configure.ac b/jemalloc/configure.ac index 7aea6a87..46a2bd4f 100644 --- a/jemalloc/configure.ac +++ b/jemalloc/configure.ac @@ -227,6 +227,16 @@ esac AC_SUBST([abi]) AC_SUBST([RPATH]) +JE_COMPILABLE([mremap(...MREMAP_FIXED...)], [ +#define _GNU_SOURCE +#include +], [ +void *p = mremap((void *)0, 0, 0, MREMAP_MAYMOVE|MREMAP_FIXED, (void *)0); +], [mremap_fixed]) +if test "x${mremap_fixed}" = "xyes" ; then + AC_DEFINE([JEMALLOC_MREMAP_FIXED]) +fi + dnl Support optional additions to rpath. AC_ARG_WITH([rpath], [AS_HELP_STRING([--with-rpath=], [Colon-separated rpath (ELF systems only)])], diff --git a/jemalloc/include/jemalloc/internal/chunk_dss.h b/jemalloc/include/jemalloc/internal/chunk_dss.h index 6be4ad1f..6f005222 100644 --- a/jemalloc/include/jemalloc/internal/chunk_dss.h +++ b/jemalloc/include/jemalloc/internal/chunk_dss.h @@ -17,6 +17,7 @@ extern malloc_mutex_t dss_mtx; void *chunk_alloc_dss(size_t size, bool *zero); +bool chunk_in_dss(void *chunk); bool chunk_dealloc_dss(void *chunk, size_t size); bool chunk_dss_boot(void); diff --git a/jemalloc/include/jemalloc/internal/chunk_swap.h b/jemalloc/include/jemalloc/internal/chunk_swap.h index d50cb197..9faa739f 100644 --- a/jemalloc/include/jemalloc/internal/chunk_swap.h +++ b/jemalloc/include/jemalloc/internal/chunk_swap.h @@ -20,6 +20,7 @@ extern size_t swap_avail; #endif void *chunk_alloc_swap(size_t size, bool *zero); +bool chunk_in_swap(void *chunk); bool chunk_dealloc_swap(void *chunk, size_t size); bool chunk_swap_enable(const int *fds, unsigned nfds, bool prezeroed); bool chunk_swap_boot(void); diff --git a/jemalloc/include/jemalloc/internal/huge.h b/jemalloc/include/jemalloc/internal/huge.h index bf231274..66544cf8 100644 --- a/jemalloc/include/jemalloc/internal/huge.h +++ b/jemalloc/include/jemalloc/internal/huge.h @@ -25,7 +25,7 @@ void *huge_ralloc_no_move(void *ptr, size_t oldsize, size_t size, size_t extra); void *huge_ralloc(void *ptr, size_t oldsize, size_t size, size_t extra, size_t alignment, bool zero); -void huge_dalloc(void *ptr); +void huge_dalloc(void *ptr, bool unmap); size_t huge_salloc(const void *ptr); #ifdef JEMALLOC_PROF prof_ctx_t *huge_prof_ctx_get(const void *ptr); diff --git a/jemalloc/include/jemalloc/internal/jemalloc_internal.h.in b/jemalloc/include/jemalloc/internal/jemalloc_internal.h.in index 3d253001..0680b43e 100644 --- a/jemalloc/include/jemalloc/internal/jemalloc_internal.h.in +++ b/jemalloc/include/jemalloc/internal/jemalloc_internal.h.in @@ -666,7 +666,7 @@ idalloc(void *ptr) if (chunk != ptr) arena_dalloc(chunk->arena, chunk, ptr); else - huge_dalloc(ptr); + huge_dalloc(ptr, true); } JEMALLOC_INLINE void * diff --git a/jemalloc/include/jemalloc/jemalloc_defs.h.in b/jemalloc/include/jemalloc/jemalloc_defs.h.in index b8f3f36b..5f46c5c9 100644 --- a/jemalloc/include/jemalloc/jemalloc_defs.h.in +++ b/jemalloc/include/jemalloc/jemalloc_defs.h.in @@ -115,6 +115,9 @@ #undef JEMALLOC_ZONE #undef JEMALLOC_ZONE_VERSION +/* If defined, use mremap(...MREMAP_FIXED...) for huge realloc(). */ +#undef JEMALLOC_MREMAP_FIXED + /* * Methods for purging unused pages differ between operating systems. * diff --git a/jemalloc/src/chunk.c b/jemalloc/src/chunk.c index 00bf50a0..301519e8 100644 --- a/jemalloc/src/chunk.c +++ b/jemalloc/src/chunk.c @@ -146,11 +146,6 @@ chunk_boot(void) chunksize_mask = chunksize - 1; chunk_npages = (chunksize >> PAGE_SHIFT); -#ifdef JEMALLOC_IVSALLOC - chunks_rtree = rtree_new((ZU(1) << (LG_SIZEOF_PTR+3)) - opt_lg_chunk); - if (chunks_rtree == NULL) - return (true); -#endif #if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF)) if (malloc_mutex_init(&chunks_mtx)) return (true); @@ -166,6 +161,11 @@ chunk_boot(void) if (chunk_dss_boot()) return (true); #endif +#ifdef JEMALLOC_IVSALLOC + chunks_rtree = rtree_new((ZU(1) << (LG_SIZEOF_PTR+3)) - opt_lg_chunk); + if (chunks_rtree == NULL) + return (true); +#endif return (false); } diff --git a/jemalloc/src/chunk_dss.c b/jemalloc/src/chunk_dss.c index d9bd63c3..5c0e290e 100644 --- a/jemalloc/src/chunk_dss.c +++ b/jemalloc/src/chunk_dss.c @@ -199,6 +199,22 @@ chunk_dealloc_dss_record(void *chunk, size_t size) return (node); } +bool +chunk_in_dss(void *chunk) +{ + bool ret; + + malloc_mutex_lock(&dss_mtx); + if ((uintptr_t)chunk >= (uintptr_t)dss_base + && (uintptr_t)chunk < (uintptr_t)dss_max) + ret = true; + else + ret = false; + malloc_mutex_unlock(&dss_mtx); + + return (ret); +} + bool chunk_dealloc_dss(void *chunk, size_t size) { diff --git a/jemalloc/src/chunk_swap.c b/jemalloc/src/chunk_swap.c index ee038ba9..cb25ae0d 100644 --- a/jemalloc/src/chunk_swap.c +++ b/jemalloc/src/chunk_swap.c @@ -184,6 +184,24 @@ chunk_dealloc_swap_record(void *chunk, size_t size) return (node); } +bool +chunk_in_swap(void *chunk) +{ + bool ret; + + assert(swap_enabled); + + malloc_mutex_lock(&swap_mtx); + if ((uintptr_t)chunk >= (uintptr_t)swap_base + && (uintptr_t)chunk < (uintptr_t)swap_max) + ret = true; + else + ret = false; + malloc_mutex_unlock(&swap_mtx); + + return (ret); +} + bool chunk_dealloc_swap(void *chunk, size_t size) { @@ -219,15 +237,15 @@ chunk_dealloc_swap(void *chunk, size_t size) } else madvise(chunk, size, MADV_DONTNEED); +#ifdef JEMALLOC_STATS + swap_avail += size; +#endif ret = false; goto RETURN; } ret = true; RETURN: -#ifdef JEMALLOC_STATS - swap_avail += size; -#endif malloc_mutex_unlock(&swap_mtx); return (ret); } diff --git a/jemalloc/src/huge.c b/jemalloc/src/huge.c index a0351975..0aadc433 100644 --- a/jemalloc/src/huge.c +++ b/jemalloc/src/huge.c @@ -215,13 +215,56 @@ huge_ralloc(void *ptr, size_t oldsize, size_t size, size_t extra, * expectation that the extra bytes will be reliably preserved. */ copysize = (size < oldsize) ? size : oldsize; - memcpy(ret, ptr, copysize); - idalloc(ptr); + + /* + * Use mremap(2) if this is a huge-->huge reallocation, and neither the + * source nor the destination are in swap or dss. + */ +#ifdef JEMALLOC_MREMAP_FIXED + if (oldsize >= chunksize +# ifdef JEMALLOC_SWAP + && (swap_enabled == false || (chunk_in_swap(ptr) == false && + chunk_in_swap(ret) == false)) +# endif +# ifdef JEMALLOC_DSS + && chunk_in_dss(ptr) == false && chunk_in_dss(ret) == false +# endif + ) { + size_t newsize = huge_salloc(ret); + + if (mremap(ptr, oldsize, newsize, MREMAP_MAYMOVE|MREMAP_FIXED, + ret) == MAP_FAILED) { + /* + * Assuming no chunk management bugs in the allocator, + * the only documented way an error can occur here is + * if the application changed the map type for a + * portion of the old allocation. This is firmly in + * undefined behavior territory, so write a diagnostic + * message, and optionally abort. + */ + char buf[BUFERROR_BUF]; + + buferror(errno, buf, sizeof(buf)); + malloc_write(": Error in mremap(): "); + malloc_write(buf); + malloc_write("\n"); + if (opt_abort) + abort(); + memcpy(ret, ptr, copysize); + idalloc(ptr); + } else + huge_dalloc(ptr, false); + } else +#endif + { + memcpy(ret, ptr, copysize); + idalloc(ptr); + } return (ret); } void -huge_dalloc(void *ptr) +huge_dalloc(void *ptr, bool unmap) { extent_node_t *node, key; @@ -241,14 +284,16 @@ huge_dalloc(void *ptr) malloc_mutex_unlock(&huge_mtx); + if (unmap) { /* Unmap chunk. */ #ifdef JEMALLOC_FILL #if (defined(JEMALLOC_SWAP) || defined(JEMALLOC_DSS)) - if (opt_junk) - memset(node->addr, 0x5a, node->size); + if (opt_junk) + memset(node->addr, 0x5a, node->size); #endif #endif - chunk_dealloc(node->addr, node->size); + chunk_dealloc(node->addr, node->size); + } base_node_dealloc(node); } diff --git a/jemalloc/test/mremap.c b/jemalloc/test/mremap.c new file mode 100644 index 00000000..146c66f4 --- /dev/null +++ b/jemalloc/test/mremap.c @@ -0,0 +1,67 @@ +#include +#include +#include +#include +#include + +#define JEMALLOC_MANGLE +#include "jemalloc_test.h" + +int +main(void) +{ + int ret, err; + size_t sz, lg_chunk, chunksize, i; + char *p, *q; + + fprintf(stderr, "Test begin\n"); + + sz = sizeof(lg_chunk); + if ((err = JEMALLOC_P(mallctl)("opt.lg_chunk", &lg_chunk, &sz, NULL, + 0))) { + assert(err != ENOENT); + fprintf(stderr, "%s(): Error in mallctl(): %s\n", __func__, + strerror(err)); + ret = 1; + goto RETURN; + } + chunksize = ((size_t)1U) << lg_chunk; + + p = (char *)malloc(chunksize); + if (p == NULL) { + fprintf(stderr, "malloc(%zu) --> %p\n", chunksize, p); + ret = 1; + goto RETURN; + } + memset(p, 'a', chunksize); + + q = (char *)realloc(p, chunksize * 2); + if (q == NULL) { + fprintf(stderr, "realloc(%p, %zu) --> %p\n", p, chunksize * 2, + q); + ret = 1; + goto RETURN; + } + for (i = 0; i < chunksize; i++) { + assert(q[i] == 'a'); + } + + p = q; + + q = (char *)realloc(p, chunksize); + if (q == NULL) { + fprintf(stderr, "realloc(%p, %zu) --> %p\n", p, chunksize, q); + ret = 1; + goto RETURN; + } + for (i = 0; i < chunksize; i++) { + assert(q[i] == 'a'); + } + + free(q); + + ret = 0; +RETURN: + fprintf(stderr, "Test end\n"); + return (ret); +} diff --git a/jemalloc/test/mremap.exp b/jemalloc/test/mremap.exp new file mode 100644 index 00000000..369a88dd --- /dev/null +++ b/jemalloc/test/mremap.exp @@ -0,0 +1,2 @@ +Test begin +Test end From ecf229a39fc253da39ae6baeab9f5c1955786ff6 Mon Sep 17 00:00:00 2001 From: Jason Evans Date: Fri, 3 Dec 2010 15:55:47 -0800 Subject: [PATCH 5/6] Add the "thread.[de]allocatedp" mallctl's. --- jemalloc/doc/jemalloc.xml.in | 30 +++++++++++++++++++++++++++++- jemalloc/src/ctl.c | 8 +++++++- 2 files changed, 36 insertions(+), 2 deletions(-) diff --git a/jemalloc/doc/jemalloc.xml.in b/jemalloc/doc/jemalloc.xml.in index 6951160e..97893c16 100644 --- a/jemalloc/doc/jemalloc.xml.in +++ b/jemalloc/doc/jemalloc.xml.in @@ -1172,7 +1172,7 @@ malloc_conf = "xmalloc:true";]]> calling this interface. - + thread.allocated (uint64_t) @@ -1186,6 +1186,20 @@ malloc_conf = "xmalloc:true";]]> + + thread.allocatedp + (uint64_t *) + r- + [] + + Get a pointer to the the value that is returned by the + thread.allocated + mallctl. This is useful for avoiding the overhead of repeated + mallctl* calls. + + + thread.deallocated (uint64_t) @@ -1198,6 +1212,20 @@ malloc_conf = "xmalloc:true";]]> cases. + + + thread.deallocatedp + (uint64_t *) + r- + [] + + Get a pointer to the the value that is returned by the + thread.deallocated + mallctl. This is useful for avoiding the overhead of repeated + mallctl* calls. + + arenas.narenas diff --git a/jemalloc/src/ctl.c b/jemalloc/src/ctl.c index 0e100590..3c8adab9 100644 --- a/jemalloc/src/ctl.c +++ b/jemalloc/src/ctl.c @@ -51,7 +51,9 @@ CTL_PROTO(tcache_flush) CTL_PROTO(thread_arena) #ifdef JEMALLOC_STATS CTL_PROTO(thread_allocated) +CTL_PROTO(thread_allocatedp) CTL_PROTO(thread_deallocated) +CTL_PROTO(thread_deallocatedp) #endif CTL_PROTO(config_debug) CTL_PROTO(config_dss) @@ -230,7 +232,9 @@ static const ctl_node_t thread_node[] = { #ifdef JEMALLOC_STATS , {NAME("allocated"), CTL(thread_allocated)}, - {NAME("deallocated"), CTL(thread_deallocated)} + {NAME("allocatedp"), CTL(thread_allocatedp)}, + {NAME("deallocated"), CTL(thread_deallocated)}, + {NAME("deallocatedp"), CTL(thread_deallocatedp)} #endif }; @@ -1142,7 +1146,9 @@ RETURN: #ifdef JEMALLOC_STATS CTL_RO_NL_GEN(thread_allocated, ALLOCATED_GET(), uint64_t); +CTL_RO_NL_GEN(thread_allocatedp, &ALLOCATED_GET(), uint64_t *); CTL_RO_NL_GEN(thread_deallocated, DEALLOCATED_GET(), uint64_t); +CTL_RO_NL_GEN(thread_deallocatedp, &DEALLOCATED_GET(), uint64_t *); #endif /******************************************************************************/ From 0e8d3d2cb9b3c9048b43588271a1e3a837ab186e Mon Sep 17 00:00:00 2001 From: Jason Evans Date: Fri, 3 Dec 2010 17:02:16 -0800 Subject: [PATCH 6/6] Updated ChangeLog for 2.1.0. --- jemalloc/ChangeLog | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) diff --git a/jemalloc/ChangeLog b/jemalloc/ChangeLog index 7b7da788..e32a5883 100644 --- a/jemalloc/ChangeLog +++ b/jemalloc/ChangeLog @@ -6,6 +6,23 @@ found in the git revision history: http://www.canonware.com/cgi-bin/gitweb.cgi?p=jemalloc.git git://canonware.com/jemalloc.git +* 2.1.0 + + This version incorporates some optimizations that can't quite be considered + bug fixes. + + New features: + - Use Linux's mremap(2) for huge object reallocation when possible. + - Avoid locking in mallctl*() when possible. + - Add the "thread.[de]allocatedp" mallctl's. + - Convert the manual page source from roff to DocBook, and generate both roff + and HTML manuals. + + Bug fixes: + - Fix a crash due to incorrect bootstrap ordering. This only impacted + --enable-debug --enable-dss configurations. + - Fix a minor statistics bug for mallctl("swap.avail", ...). + * 2.0.1 Bug fixes: