Merge branch 'dev'

Update ChangeLog for 3.1.0.
Add ctl_mutex proection to arena_i_dss_ctl().
2026-07-22 12:43:16 +00:00 · 2012-10-16 10:40:57 -07:00 · 2012-10-16 10:12:40 -07:00 · 2012-10-15 12:48:59 -07:00 · 2012-10-12 18:26:16 -07:00 · 2012-10-10 09:10:37 -07:00
123 changed files with 19904 additions and 14469 deletions
--- a/.gitignore
+++ b/.gitignore
@@ -1,23 +1,25 @@
-/jemalloc/autom4te.cache/
-/jemalloc/config.stamp
-/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
-/jemalloc/include/jemalloc/internal/jemalloc_internal\.h
-/jemalloc/include/jemalloc/jemalloc\.h
-/jemalloc/include/jemalloc/jemalloc_defs\.h
-/jemalloc/test/jemalloc_test\.h
-/jemalloc/src/*.[od]
-/jemalloc/test/*.[od]
-/jemalloc/test/*.out
-/jemalloc/test/[a-z]*
-!/jemalloc/test/*.c
-!/jemalloc/test/*.exp
-/jemalloc/VERSION
+/autom4te.cache/
+/config.stamp
+/config.log
+/config.status
+/configure
+/doc/html.xsl
+/doc/manpages.xsl
+/doc/jemalloc.xml
+/doc/jemalloc.html
+/doc/jemalloc.3
+/lib/
+/Makefile
+/include/jemalloc/internal/jemalloc_internal\.h
+/include/jemalloc/internal/size_classes\.h
+/include/jemalloc/jemalloc\.h
+/include/jemalloc/jemalloc_defs\.h
+/test/jemalloc_test\.h
+/src/*.[od]
+/test/*.[od]
+/test/*.out
+/test/[a-zA-Z_]*
+!test/*.c
+!test/*.exp
+/VERSION
+/bin/jemalloc.sh
--- a/27
+++ b/27
@@ -0,0 +1,27 @@
+Unless otherwise specified, files in the jemalloc source distribution are
+subject to the following license:
+--------------------------------------------------------------------------------
+Copyright (C) 2002-2012 Jason Evans <jasone@canonware.com>.
+All rights reserved.
+Copyright (C) 2007-2012 Mozilla Foundation.  All rights reserved.
+Copyright (C) 2009-2012 Facebook, Inc.  All rights reserved.
+
+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(s),
+   this list of conditions and the following disclaimer.
+2. Redistributions in binary form must reproduce the above copyright notice(s),
+   this list of conditions and the following disclaimer in the documentation
+   and/or other materials provided with the distribution.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``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 COPYRIGHT HOLDER(S) 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.
+--------------------------------------------------------------------------------
--- a/365
+++ b/365
@@ -0,0 +1,365 @@
+Following are change highlights associated with official releases.  Important
+bug fixes are all mentioned, but internal enhancements are omitted here for
+brevity (even though they are more fun to write about).  Much more detail can be
+found in the git revision history:
+
+    http://www.canonware.com/cgi-bin/gitweb.cgi?p=jemalloc.git
+    git://canonware.com/jemalloc.git
+
+* 3.1.0 (October 16, 2012)
+
+  New features:
+  - Auto-detect whether running inside Valgrind, thus removing the need to
+    manually specify MALLOC_CONF=valgrind:true.
+  - Add the "arenas.extend" mallctl, which allows applications to create
+    manually managed arenas.
+  - Add the ALLOCM_ARENA() flag for {,r,d}allocm().
+  - Add the "opt.dss", "arena.<i>.dss", and "stats.arenas.<i>.dss" mallctls,
+    which provide control over dss/mmap precedence.
+  - Add the "arena.<i>.purge" mallctl, which obsoletes "arenas.purge".
+  - Define LG_QUANTUM for hppa.
+
+  Incompatible changes:
+  - Disable tcache by default if running inside Valgrind, in order to avoid
+    making unallocated objects appear reachable to Valgrind.
+  - Drop const from malloc_usable_size() argument on Linux.
+
+  Bug fixes:
+  - Fix heap profiling crash if sampled object is freed via realloc(p, 0).
+  - Remove const from __*_hook variable declarations, so that glibc can modify
+    them during process forking.
+  - Fix mlockall(2)/madvise(2) interaction.
+  - Fix fork(2)-related deadlocks.
+  - Fix error return value for "thread.tcache.enabled" mallctl.
+
+* 3.0.0 (May 11, 2012)
+
+  Although this version adds some major new features, the primary focus is on
+  internal code cleanup that facilitates maintainability and portability, most
+  of which is not reflected in the ChangeLog.  This is the first release to
+  incorporate substantial contributions from numerous other developers, and the
+  result is a more broadly useful allocator (see the git revision history for
+  contribution details).  Note that the license has been unified, thanks to
+  Facebook granting a license under the same terms as the other copyright
+  holders (see COPYING).
+
+  New features:
+  - Implement Valgrind support, redzones, and quarantine.
+  - Add support for additional platforms:
+    + FreeBSD
+    + Mac OS X Lion
+    + MinGW
+    + Windows (no support yet for replacing the system malloc)
+  - Add support for additional architectures:
+    + MIPS
+    + SH4
+    + Tilera
+  - Add support for cross compiling.
+  - Add nallocm(), which rounds a request size up to the nearest size class
+    without actually allocating.
+  - Implement aligned_alloc() (blame C11).
+  - Add the "thread.tcache.enabled" mallctl.
+  - Add the "opt.prof_final" mallctl.
+  - Update pprof (from gperftools 2.0).
+  - Add the --with-mangling option.
+  - Add the --disable-experimental option.
+  - Add the --disable-munmap option, and make it the default on Linux.
+  - Add the --enable-mremap option, which disables use of mremap(2) by default.
+
+  Incompatible changes:
+  - Enable stats by default.
+  - Enable fill by default.
+  - Disable lazy locking by default.
+  - Rename the "tcache.flush" mallctl to "thread.tcache.flush".
+  - Rename the "arenas.pagesize" mallctl to "arenas.page".
+  - Change the "opt.lg_prof_sample" default from 0 to 19 (1 B to 512 KiB).
+  - Change the "opt.prof_accum" default from true to false.
+
+  Removed features:
+  - Remove the swap feature, including the "config.swap", "swap.avail",
+    "swap.prezeroed", "swap.nfds", and "swap.fds" mallctls.
+  - Remove highruns statistics, including the
+    "stats.arenas.<i>.bins.<j>.highruns" and
+    "stats.arenas.<i>.lruns.<j>.highruns" mallctls.
+  - As part of small size class refactoring, remove the "opt.lg_[qc]space_max",
+    "arenas.cacheline", "arenas.subpage", "arenas.[tqcs]space_{min,max}", and
+    "arenas.[tqcs]bins" mallctls.
+  - Remove the "arenas.chunksize" mallctl.
+  - Remove the "opt.lg_prof_tcmax" option.
+  - Remove the "opt.lg_prof_bt_max" option.
+  - Remove the "opt.lg_tcache_gc_sweep" option.
+  - Remove the --disable-tiny option, including the "config.tiny" mallctl.
+  - Remove the --enable-dynamic-page-shift configure option.
+  - Remove the --enable-sysv configure option.
+
+  Bug fixes:
+  - Fix a statistics-related bug in the "thread.arena" mallctl that could cause
+    invalid statistics and crashes.
+  - Work around TLS deallocation via free() on Linux.  This bug could cause
+    write-after-free memory corruption.
+  - Fix a potential deadlock that could occur during interval- and
+    growth-triggered heap profile dumps.
+  - Fix large calloc() zeroing bugs due to dropping chunk map unzeroed flags.
+  - Fix chunk_alloc_dss() to stop claiming memory is zeroed.  This bug could
+    cause memory corruption and crashes with --enable-dss specified.
+  - Fix fork-related bugs that could cause deadlock in children between fork
+    and exec.
+  - Fix malloc_stats_print() to honor 'b' and 'l' in the opts parameter.
+  - Fix realloc(p, 0) to act like free(p).
+  - Do not enforce minimum alignment in memalign().
+  - Check for NULL pointer in malloc_usable_size().
+  - Fix an off-by-one heap profile statistics bug that could be observed in
+    interval- and growth-triggered heap profiles.
+  - Fix the "epoch" mallctl to update cached stats even if the passed in epoch
+    is 0.
+  - Fix bin->runcur management to fix a layout policy bug.  This bug did not
+    affect correctness.
+  - Fix a bug in choose_arena_hard() that potentially caused more arenas to be
+    initialized than necessary.
+  - Add missing "opt.lg_tcache_max" mallctl implementation.
+  - Use glibc allocator hooks to make mixed allocator usage less likely.
+  - Fix build issues for --disable-tcache.
+  - Don't mangle pthread_create() when --with-private-namespace is specified.
+
+* 2.2.5 (November 14, 2011)
+
+  Bug fixes:
+  - Fix huge_ralloc() race when using mremap(2).  This is a serious bug that
+    could cause memory corruption and/or crashes.
+  - Fix huge_ralloc() to maintain chunk statistics.
+  - Fix malloc_stats_print(..., "a") output.
+
+* 2.2.4 (November 5, 2011)
+
+  Bug fixes:
+  - Initialize arenas_tsd before using it.  This bug existed for 2.2.[0-3], as
+    well as for --disable-tls builds in earlier releases.
+  - Do not assume a 4 KiB page size in test/rallocm.c.
+
+* 2.2.3 (August 31, 2011)
+
+  This version fixes numerous bugs related to heap profiling.
+
+  Bug fixes:
+  - Fix a prof-related race condition.  This bug could cause memory corruption,
+    but only occurred in non-default configurations (prof_accum:false).
+  - Fix off-by-one backtracing issues (make sure that prof_alloc_prep() is
+    excluded from backtraces).
+  - Fix a prof-related bug in realloc() (only triggered by OOM errors).
+  - Fix prof-related bugs in allocm() and rallocm().
+  - Fix prof_tdata_cleanup() for --disable-tls builds.
+  - Fix a relative include path, to fix objdir builds.
+
+* 2.2.2 (July 30, 2011)
+
+  Bug fixes:
+  - Fix a build error for --disable-tcache.
+  - Fix assertions in arena_purge() (for real this time).
+  - Add the --with-private-namespace option.  This is a workaround for symbol
+    conflicts that can inadvertently arise when using static libraries.
+
+* 2.2.1 (March 30, 2011)
+
+  Bug fixes:
+  - Implement atomic operations for x86/x64.  This fixes compilation failures
+    for versions of gcc that are still in wide use.
+  - Fix an assertion in arena_purge().
+
+* 2.2.0 (March 22, 2011)
+
+  This version incorporates several improvements to algorithms and data
+  structures that tend to reduce fragmentation and increase speed.
+
+  New features:
+  - Add the "stats.cactive" mallctl.
+  - Update pprof (from google-perftools 1.7).
+  - Improve backtracing-related configuration logic, and add the
+    --disable-prof-libgcc option.
+
+  Bug fixes:
+  - Change default symbol visibility from "internal", to "hidden", which
+    decreases the overhead of library-internal function calls.
+  - Fix symbol visibility so that it is also set on OS X.
+  - Fix a build dependency regression caused by the introduction of the .pic.o
+    suffix for PIC object files.
+  - Add missing checks for mutex initialization failures.
+  - Don't use libgcc-based backtracing except on x64, where it is known to work.
+  - Fix deadlocks on OS X that were due to memory allocation in
+    pthread_mutex_lock().
+  - Heap profiling-specific fixes:
+    + Fix memory corruption due to integer overflow in small region index
+      computation, when using a small enough sample interval that profiling
+      context pointers are stored in small run headers.
+    + Fix a bootstrap ordering bug that only occurred with TLS disabled.
+    + Fix a rallocm() rsize bug.
+    + Fix error detection bugs for aligned memory allocation.
+
+* 2.1.3 (March 14, 2011)
+
+  Bug fixes:
+  - Fix a cpp logic regression (due to the "thread.{de,}allocatedp" mallctl fix
+    for OS X in 2.1.2).
+  - Fix a "thread.arena" mallctl bug.
+  - Fix a thread cache stats merging bug.
+
+* 2.1.2 (March 2, 2011)
+
+  Bug fixes:
+  - Fix "thread.{de,}allocatedp" mallctl for OS X.
+  - Add missing jemalloc.a to build system.
+
+* 2.1.1 (January 31, 2011)
+
+  Bug fixes:
+  - Fix aligned huge reallocation (affected allocm()).
+  - Fix the ALLOCM_LG_ALIGN macro definition.
+  - Fix a heap dumping deadlock.
+  - Fix a "thread.arena" mallctl bug.
+
+* 2.1.0 (December 3, 2010)
+
+  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 (October 29, 2010)
+
+  Bug fixes:
+  - Fix a race condition in heap profiling that could cause undefined behavior
+    if "opt.prof_accum" were disabled.
+  - Add missing mutex unlocks for some OOM error paths in the heap profiling
+    code.
+  - Fix a compilation error for non-C99 builds.
+
+* 2.0.0 (October 24, 2010)
+
+  This version focuses on the experimental *allocm() API, and on improved
+  run-time configuration/introspection.  Nonetheless, numerous performance
+  improvements are also included.
+
+  New features:
+  - Implement the experimental {,r,s,d}allocm() API, which provides a superset
+    of the functionality available via malloc(), calloc(), posix_memalign(),
+    realloc(), malloc_usable_size(), and free().  These functions can be used to
+    allocate/reallocate aligned zeroed memory, ask for optional extra memory
+    during reallocation, prevent object movement during reallocation, etc.
+  - Replace JEMALLOC_OPTIONS/JEMALLOC_PROF_PREFIX with MALLOC_CONF, which is
+    more human-readable, and more flexible.  For example:
+      JEMALLOC_OPTIONS=AJP
+    is now:
+      MALLOC_CONF=abort:true,fill:true,stats_print:true
+  - Port to Apple OS X.  Sponsored by Mozilla.
+  - Make it possible for the application to control thread-->arena mappings via
+    the "thread.arena" mallctl.
+  - Add compile-time support for all TLS-related functionality via pthreads TSD.
+    This is mainly of interest for OS X, which does not support TLS, but has a
+    TSD implementation with similar performance.
+  - Override memalign() and valloc() if they are provided by the system.
+  - Add the "arenas.purge" mallctl, which can be used to synchronously purge all
+    dirty unused pages.
+  - Make cumulative heap profiling data optional, so that it is possible to
+    limit the amount of memory consumed by heap profiling data structures.
+  - Add per thread allocation counters that can be accessed via the
+    "thread.allocated" and "thread.deallocated" mallctls.
+
+  Incompatible changes:
+  - Remove JEMALLOC_OPTIONS and malloc_options (see MALLOC_CONF above).
+  - Increase default backtrace depth from 4 to 128 for heap profiling.
+  - Disable interval-based profile dumps by default.
+
+  Bug fixes:
+  - Remove bad assertions in fork handler functions.  These assertions could
+    cause aborts for some combinations of configure settings.
+  - Fix strerror_r() usage to deal with non-standard semantics in GNU libc.
+  - Fix leak context reporting.  This bug tended to cause the number of contexts
+    to be underreported (though the reported number of objects and bytes were
+    correct).
+  - Fix a realloc() bug for large in-place growing reallocation.  This bug could
+    cause memory corruption, but it was hard to trigger.
+  - Fix an allocation bug for small allocations that could be triggered if
+    multiple threads raced to create a new run of backing pages.
+  - Enhance the heap profiler to trigger samples based on usable size, rather
+    than request size.
+  - Fix a heap profiling bug due to sometimes losing track of requested object
+    size for sampled objects.
+
+* 1.0.3 (August 12, 2010)
+
+  Bug fixes:
+  - Fix the libunwind-based implementation of stack backtracing (used for heap
+    profiling).  This bug could cause zero-length backtraces to be reported.
+  - Add a missing mutex unlock in library initialization code.  If multiple
+    threads raced to initialize malloc, some of them could end up permanently
+    blocked.
+
+* 1.0.2 (May 11, 2010)
+
+  Bug fixes:
+  - Fix junk filling of large objects, which could cause memory corruption.
+  - Add MAP_NORESERVE support for chunk mapping, because otherwise virtual
+    memory limits could cause swap file configuration to fail.  Contributed by
+    Jordan DeLong.
+
+* 1.0.1 (April 14, 2010)
+
+  Bug fixes:
+  - Fix compilation when --enable-fill is specified.
+  - Fix threads-related profiling bugs that affected accuracy and caused memory
+    to be leaked during thread exit.
+  - Fix dirty page purging race conditions that could cause crashes.
+  - Fix crash in tcache flushing code during thread destruction.
+
+* 1.0.0 (April 11, 2010)
+
+  This release focuses on speed and run-time introspection.  Numerous
+  algorithmic improvements make this release substantially faster than its
+  predecessors.
+
+  New features:
+  - Implement autoconf-based configuration system.
+  - Add mallctl*(), for the purposes of introspection and run-time
+    configuration.
+  - Make it possible for the application to manually flush a thread's cache, via
+    the "tcache.flush" mallctl.
+  - Base maximum dirty page count on proportion of active memory.
+  - Compute various addtional run-time statistics, including per size class
+    statistics for large objects.
+  - Expose malloc_stats_print(), which can be called repeatedly by the
+    application.
+  - Simplify the malloc_message() signature to only take one string argument,
+    and incorporate an opaque data pointer argument for use by the application
+    in combination with malloc_stats_print().
+  - Add support for allocation backed by one or more swap files, and allow the
+    application to disable over-commit if swap files are in use.
+  - Implement allocation profiling and leak checking.
+
+  Removed features:
+  - Remove the dynamic arena rebalancing code, since thread-specific caching
+    reduces its utility.
+
+  Bug fixes:
+  - Modify chunk allocation to work when address space layout randomization
+    (ASLR) is in use.
+  - Fix thread cleanup bugs related to TLS destruction.
+  - Handle 0-size allocation requests in posix_memalign().
+  - Fix a chunk leak.  The leaked chunks were never touched, so this impacted
+    virtual memory usage, but not physical memory usage.
+
+* linux_2008082[78]a (August 27/28, 2008)
+
+  These snapshot releases are the simple result of incorporating Linux-specific
+  support into the FreeBSD malloc sources.
+
+--------------------------------------------------------------------------------
+vim:filetype=text:textwidth=80
--- a/jemalloc/INSTALL
+++ b/jemalloc/INSTALL
@@ -26,6 +26,19 @@ any of the following arguments (not a definitive list) to 'configure':
    Embed one or more library paths, so that libjemalloc can find the libraries
    it is linked to.  This works only on ELF-based systems.

+--with-mangling=<map>
+    Mangle public symbols specified in <map> which is a comma-separated list of
+    name:mangled pairs.
+
+    For example, to use ld's --wrap option as an alternative method for
+    overriding libc's malloc implementation, specify something like:
+
+      --with-mangling=malloc:__wrap_malloc,free:__wrap_free[...]
+
+    Note that mangling happens prior to application of the prefix specified by
+    --with-jemalloc-prefix, and mangled symbols are then ignored when applying
+    the prefix.
+
 --with-jemalloc-prefix=<prefix>
    Prefix all public APIs with <prefix>.  For example, if <prefix> is
    "prefix_", API changes like the following occur:
@@ -42,6 +55,12 @@ any of the following arguments (not a definitive list) to 'configure':
    jemalloc overlays the default malloc zone, but makes no attempt to actually
    replace the "malloc", "calloc", etc. symbols.

+--with-private-namespace=<prefix>
+    Prefix all library-private APIs with <prefix>.  For shared libraries,
+    symbol visibility mechanisms prevent these symbols from being exported, but
+    for static libraries, naming collisions are a real possibility.  By
+    default, the prefix is "" (empty string).
+
 --with-install-suffix=<suffix>
    Append <suffix> to the base name of all installed files, such that multiple
    versions of jemalloc can coexist in the same installation directory.  For
@@ -56,74 +75,78 @@ any of the following arguments (not a definitive list) to 'configure':
    Enable assertions and validation code.  This incurs a substantial
    performance hit, but is very useful during application development.

--enable-stats
-    Enable statistics gathering functionality.  See the "opt.stats_print"
+--disable-stats
+    Disable statistics gathering functionality.  See the "opt.stats_print"
    option documentation for usage details.

 --enable-prof
    Enable heap profiling and leak detection functionality.  See the "opt.prof"
-    option documentation for usage details.
+    option documentation for usage details.  When enabled, there are several
+    approaches to backtracing, and the configure script chooses the first one
+    in the following list that appears to function correctly:

--disable-prof-libgcc
-    Disable the use of libgcc's backtracing functionality.  Ordinarily, libgcc's
-    backtracing functionality is superior to the alternatives, but it may fail
-    to capture backtraces on some systems.
+    + libunwind      (requires --enable-prof-libunwind)
+    + libgcc         (unless --disable-prof-libgcc)
+    + gcc intrinsics (unless --disable-prof-gcc)

 --enable-prof-libunwind
    Use the libunwind library (http://www.nongnu.org/libunwind/) for stack
-    backtracing.  libunwind is quite slow, but it tends to work across a wider
-    variety of system configurations than the default backtracing code, which is
-    based on libgcc functionality or gcc intrinsics.
+    backtracing.
+
+--disable-prof-libgcc
+    Disable the use of libgcc's backtracing functionality.
+
+--disable-prof-gcc
+    Disable the use of gcc intrinsics for backtracing.

 --with-static-libunwind=<libunwind.a>
    Statically link against the specified libunwind.a rather than dynamically
    linking with -lunwind.

--disable-tiny
-    Disable tiny (sub-quantum-sized) object support.  Technically it is not
-    legal for a malloc implementation to allocate objects with less than
-    quantum alignment (8 or 16 bytes, depending on architecture), but in
-    practice it never causes any problems if, for example, 4-byte allocations
-    are 4-byte-aligned.
-
 --disable-tcache
    Disable thread-specific caches for small objects.  Objects are cached and
    released in bulk, thus reducing the total number of mutex operations.  See
    the "opt.tcache" option for usage details.

--enable-swap
-    Enable mmap()ed swap file support.  When this feature is built in, it is
-    possible to specify one or more files that act as backing store.  This
-    effectively allows for per application swap files.
+--enable-mremap
+    Enable huge realloc() via mremap(2).  mremap() is disabled by default
+    because the flavor used is specific to Linux, which has a quirk in its
+    virtual memory allocation algorithm that causes semi-permanent VM map holes
+    under normal jemalloc operation.
+
+--disable-munmap
+    Disable virtual memory deallocation via munmap(2); instead keep track of
+    the virtual memory for later use.  munmap() is disabled by default (i.e.
+    --disable-munmap is implied) on Linux, which has a quirk in its virtual
+    memory allocation algorithm that causes semi-permanent VM map holes under
+    normal jemalloc operation.

 --enable-dss
    Enable support for page allocation/deallocation via sbrk(2), in addition to
    mmap(2).

--enable-fill
-    Enable support for junk/zero filling of memory.  See the "opt.junk"/
-    "opt.zero" option documentation for usage details.
+--disable-fill
+    Disable support for junk/zero filling of memory, quarantine, and redzones.
+    See the "opt.junk", "opt.zero", "opt.quarantine", and "opt.redzone" option
+    documentation for usage details.
+
+--disable-valgrind
+    Disable support for Valgrind.
+
+--disable-experimental
+    Disable support for the experimental API (*allocm()).
+
+--enable-utrace
+    Enable utrace(2)-based allocation tracing.  This feature is not broadly
+    portable (FreeBSD has it, but Linux and OS X do not).

 --enable-xmalloc
    Enable support for optional immediate termination due to out-of-memory
    errors, as is commonly implemented by "xmalloc" wrapper function for malloc.
    See the "opt.xmalloc" option documentation for usage details.

--enable-sysv
-    Enable support for System V semantics, wherein malloc(0) returns NULL
-    rather than a minimal allocation.  See the "opt.sysv" option documentation
-    for usage details.
-
--enable-dynamic-page-shift
-    Under most conditions, the system page size never changes (usually 4KiB or
-    8KiB, depending on architecture and configuration), and unless this option
-    is enabled, jemalloc assumes that page size can safely be determined during
-    configuration and hard-coded.  Enabling dynamic page size determination has
-    a measurable impact on performance, since the compiler is forced to load
-    the page size from memory rather than embedding immediate values.
-
--disable-lazy-lock
-    Disable code that wraps pthread_create() to detect when an application
+--enable-lazy-lock
+    Enable code that wraps pthread_create() to detect when an application
    switches from single-threaded to multi-threaded mode, so that it can avoid
    mutex locking/unlocking operations while in single-threaded mode.  In
    practice, this feature usually has little impact on performance unless
@@ -170,11 +193,24 @@ PATH="?"

 === Advanced compilation =======================================================

+To build only parts of jemalloc, use the following targets:
+
+    build_lib_shared
+    build_lib_static
+    build_lib
+    build_doc_html
+    build_doc_man
+    build_doc
+
 To install only parts of jemalloc, use the following targets:

    install_bin
    install_include
+    install_lib_shared
+    install_lib_static
    install_lib
+    install_doc_html
+    install_doc_man
    install_doc

 To clean up build results to varying degrees, use the following make targets:
@@ -237,10 +273,6 @@ directory, issue configuration and build commands:

 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:
+prior to installation via the following command:

    nroff -man -t doc/jemalloc.3
-
-    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)
--- a/Makefile.in
+++ b/Makefile.in
@@ -0,0 +1,323 @@
+# Clear out all vpaths, then set just one (default vpath) for the main build
+# directory.
+vpath
+vpath % .
+
+# Clear the default suffixes, so that built-in rules are not used.
+.SUFFIXES :
+
+SHELL := /bin/sh
+
+CC := @CC@
+
+# Configuration parameters.
+DESTDIR =
+BINDIR := $(DESTDIR)@BINDIR@
+INCLUDEDIR := $(DESTDIR)@INCLUDEDIR@
+LIBDIR := $(DESTDIR)@LIBDIR@
+DATADIR := $(DESTDIR)@DATADIR@
+MANDIR := $(DESTDIR)@MANDIR@
+srcroot := @srcroot@
+objroot := @objroot@
+abs_srcroot := @abs_srcroot@
+abs_objroot := @abs_objroot@
+
+# Build parameters.
+CPPFLAGS := @CPPFLAGS@ -I$(srcroot)include -I$(objroot)include
+CFLAGS := @CFLAGS@
+LDFLAGS := @LDFLAGS@
+EXTRA_LDFLAGS := @EXTRA_LDFLAGS@
+LIBS := @LIBS@
+RPATH_EXTRA := @RPATH_EXTRA@
+SO := @so@
+IMPORTLIB := @importlib@
+O := @o@
+A := @a@
+EXE := @exe@
+LIBPREFIX := @libprefix@
+REV := @rev@
+install_suffix := @install_suffix@
+ABI := @abi@
+XSLTPROC := @XSLTPROC@
+AUTOCONF := @AUTOCONF@
+_RPATH = @RPATH@
+RPATH = $(if $(1),$(call _RPATH,$(1)))
+cfghdrs_in := @cfghdrs_in@
+cfghdrs_out := @cfghdrs_out@
+cfgoutputs_in := @cfgoutputs_in@
+cfgoutputs_out := @cfgoutputs_out@
+enable_autogen := @enable_autogen@
+enable_experimental := @enable_experimental@
+DSO_LDFLAGS = @DSO_LDFLAGS@
+SOREV = @SOREV@
+PIC_CFLAGS = @PIC_CFLAGS@
+CTARGET = @CTARGET@
+LDTARGET = @LDTARGET@
+MKLIB = @MKLIB@
+CC_MM = @CC_MM@
+
+ifeq (macho, $(ABI))
+TEST_LIBRARY_PATH := DYLD_FALLBACK_LIBRARY_PATH="$(objroot)lib"
+else
+ifeq (pecoff, $(ABI))
+TEST_LIBRARY_PATH := PATH="$(PATH):$(objroot)lib"
+else
+TEST_LIBRARY_PATH :=
+endif
+endif
+
+LIBJEMALLOC := $(LIBPREFIX)jemalloc$(install_suffix)
+
+# Lists of files.
+BINS := $(srcroot)bin/pprof $(objroot)bin/jemalloc.sh
+CHDRS := $(objroot)include/jemalloc/jemalloc$(install_suffix).h \
+	$(objroot)include/jemalloc/jemalloc_defs$(install_suffix).h
+CSRCS := $(srcroot)src/jemalloc.c $(srcroot)src/arena.c $(srcroot)src/atomic.c \
+	$(srcroot)src/base.c $(srcroot)src/bitmap.c $(srcroot)src/chunk.c \
+	$(srcroot)src/chunk_dss.c $(srcroot)src/chunk_mmap.c \
+	$(srcroot)src/ckh.c $(srcroot)src/ctl.c $(srcroot)src/extent.c \
+	$(srcroot)src/hash.c $(srcroot)src/huge.c $(srcroot)src/mb.c \
+	$(srcroot)src/mutex.c $(srcroot)src/prof.c $(srcroot)src/quarantine.c \
+	$(srcroot)src/rtree.c $(srcroot)src/stats.c $(srcroot)src/tcache.c \
+	$(srcroot)src/util.c $(srcroot)src/tsd.c
+ifeq (macho, $(ABI))
+CSRCS += $(srcroot)src/zone.c
+endif
+ifeq ($(IMPORTLIB),$(SO))
+STATIC_LIBS := $(objroot)lib/$(LIBJEMALLOC).$(A)
+endif
+ifdef PIC_CFLAGS
+STATIC_LIBS += $(objroot)lib/$(LIBJEMALLOC)_pic.$(A)
+else
+STATIC_LIBS += $(objroot)lib/$(LIBJEMALLOC)_s.$(A)
+endif
+DSOS := $(objroot)lib/$(LIBJEMALLOC).$(SOREV)
+ifneq ($(SOREV),$(SO))
+DSOS += $(objroot)lib/$(LIBJEMALLOC).$(SO)
+endif
+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/aligned_alloc.c $(srcroot)test/allocated.c \
+	$(srcroot)test/ALLOCM_ARENA.c $(srcroot)test/bitmap.c \
+	$(srcroot)test/mremap.c $(srcroot)test/posix_memalign.c \
+	$(srcroot)test/thread_arena.c $(srcroot)test/thread_tcache_enabled.c
+ifeq ($(enable_experimental), 1)
+CTESTS += $(srcroot)test/allocm.c $(srcroot)test/rallocm.c
+endif
+
+COBJS := $(CSRCS:$(srcroot)%.c=$(objroot)%.$(O))
+CPICOBJS := $(CSRCS:$(srcroot)%.c=$(objroot)%.pic.$(O))
+CTESTOBJS := $(CTESTS:$(srcroot)%.c=$(objroot)%.$(O))
+
+.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
+
+.SECONDARY : $(CTESTOBJS)
+
+# Default target.
+all: build
+
+dist: build_doc
+
+$(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 $<
+
+build_doc_html: $(DOCS_HTML)
+build_doc_man: $(DOCS_MAN3)
+build_doc: $(DOCS)
+
+#
+# Include generated dependency files.
+#
+ifdef CC_MM
+-include $(COBJS:%.$(O)=%.d)
+-include $(CPICOBJS:%.$(O)=%.d)
+-include $(CTESTOBJS:%.$(O)=%.d)
+endif
+
+$(COBJS): $(objroot)src/%.$(O): $(srcroot)src/%.c
+$(CPICOBJS): $(objroot)src/%.pic.$(O): $(srcroot)src/%.c
+$(CPICOBJS): CFLAGS += $(PIC_CFLAGS)
+$(CTESTOBJS): $(objroot)test/%.$(O): $(srcroot)test/%.c
+$(CTESTOBJS): CPPFLAGS += -I$(objroot)test
+ifneq ($(IMPORTLIB),$(SO))
+$(COBJS): CPPFLAGS += -DDLLEXPORT
+endif
+
+ifndef CC_MM
+# Dependencies
+HEADER_DIRS = $(srcroot)include/jemalloc/internal \
+	$(objroot)include/jemalloc $(objroot)include/jemalloc/internal
+HEADERS = $(wildcard $(foreach dir,$(HEADER_DIRS),$(dir)/*.h))
+$(COBJS) $(CPICOBJS) $(CTESTOBJS): $(HEADERS)
+$(CTESTOBJS): $(objroot)test/jemalloc_test.h
+endif
+
+$(COBJS) $(CPICOBJS) $(CTESTOBJS): %.$(O):
+	@mkdir -p $(@D)
+	$(CC) $(CFLAGS) -c $(CPPFLAGS) $(CTARGET) $<
+ifdef CC_MM
+	@$(CC) -MM $(CPPFLAGS) -MT $@ -o $(@:%.$(O)=%.d) $<
+endif
+
+ifneq ($(SOREV),$(SO))
+%.$(SO) : %.$(SOREV)
+	@mkdir -p $(@D)
+	ln -sf $(<F) $@
+endif
+
+$(objroot)lib/$(LIBJEMALLOC).$(SOREV) : $(if $(PIC_CFLAGS),$(CPICOBJS),$(COBJS))
+	@mkdir -p $(@D)
+	$(CC) $(DSO_LDFLAGS) $(call RPATH,$(RPATH_EXTRA)) $(LDTARGET) $+ $(LDFLAGS) $(LIBS) $(EXTRA_LDFLAGS)
+
+$(objroot)lib/$(LIBJEMALLOC)_pic.$(A) : $(CPICOBJS)
+$(objroot)lib/$(LIBJEMALLOC).$(A) : $(COBJS)
+$(objroot)lib/$(LIBJEMALLOC)_s.$(A) : $(COBJS)
+
+$(STATIC_LIBS):
+	@mkdir -p $(@D)
+	$(MKLIB) $+
+
+$(objroot)test/bitmap$(EXE): $(objroot)src/bitmap.$(O)
+
+$(objroot)test/%$(EXE): $(objroot)test/%.$(O) $(objroot)src/util.$(O) $(DSOS)
+	@mkdir -p $(@D)
+	$(CC) $(LDTARGET) $(filter %.$(O),$^) $(call RPATH,$(objroot)lib) $(objroot)lib/$(LIBJEMALLOC).$(IMPORTLIB) $(filter -lpthread,$(LIBS)) $(EXTRA_LDFLAGS)
+
+build_lib_shared: $(DSOS)
+build_lib_static: $(STATIC_LIBS)
+build: build_lib_shared build_lib_static
+
+install_bin:
+	install -d $(BINDIR)
+	@for b in $(BINS); do \
+	echo "install -m 755 $$b $(BINDIR)"; \
+	install -m 755 $$b $(BINDIR); \
+done
+
+install_include:
+	install -d $(INCLUDEDIR)/jemalloc
+	@for h in $(CHDRS); do \
+	echo "install -m 644 $$h $(INCLUDEDIR)/jemalloc"; \
+	install -m 644 $$h $(INCLUDEDIR)/jemalloc; \
+done
+
+install_lib_shared: $(DSOS)
+	install -d $(LIBDIR)
+	install -m 755 $(objroot)lib/$(LIBJEMALLOC).$(SOREV) $(LIBDIR)
+ifneq ($(SOREV),$(SO))
+	ln -sf $(LIBJEMALLOC).$(SOREV) $(LIBDIR)/$(LIBJEMALLOC).$(SO)
+endif
+
+install_lib_static: $(STATIC_LIBS)
+	install -d $(LIBDIR)
+	@for l in $(STATIC_LIBS); do \
+	echo "install -m 755 $$l $(LIBDIR)"; \
+	install -m 755 $$l $(LIBDIR); \
+done
+
+install_lib: install_lib_shared install_lib_static
+
+install_doc_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_doc_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_doc_html install_doc_man
+
+install: install_bin install_include install_lib install_doc
+
+tests: $(CTESTS:$(srcroot)%.c=$(objroot)%$(EXE))
+
+check: tests
+	@mkdir -p $(objroot)test
+	@$(SHELL) -c 'total=0; \
+		failures=0; \
+		echo "========================================="; \
+		for t in $(CTESTS:$(srcroot)%.c=$(objroot)%); do \
+			total=`expr $$total + 1`; \
+			/bin/echo -n "$${t} ... "; \
+			$(TEST_LIBRARY_PATH) $${t}$(EXE) $(abs_srcroot) \
+			  $(abs_objroot) > $(objroot)$${t}.out 2>&1; \
+			if test -e "$(srcroot)$${t}.exp"; then \
+				diff -w -u $(srcroot)$${t}.exp \
+				  $(objroot)$${t}.out >/dev/null 2>&1; \
+				fail=$$?; \
+				if test "$${fail}" -eq "1" ; then \
+					failures=`expr $${failures} + 1`; \
+					echo "*** FAIL ***"; \
+				else \
+					echo "pass"; \
+				fi; \
+			else \
+				echo "*** FAIL *** (.exp file is missing)"; \
+				failures=`expr $${failures} + 1`; \
+			fi; \
+		done; \
+		echo "========================================="; \
+		echo "Failures: $${failures}/$${total}"'
+
+clean:
+	rm -f $(COBJS)
+	rm -f $(CPICOBJS)
+	rm -f $(COBJS:%.$(O)=%.d)
+	rm -f $(CPICOBJS:%.$(O)=%.d)
+	rm -f $(CTESTOBJS:%.$(O)=%$(EXE))
+	rm -f $(CTESTOBJS)
+	rm -f $(CTESTOBJS:%.$(O)=%.d)
+	rm -f $(CTESTOBJS:%.$(O)=%.out)
+	rm -f $(DSOS) $(STATIC_LIBS)
+
+distclean: clean
+	rm -rf $(objroot)autom4te.cache
+	rm -f $(objroot)config.log
+	rm -f $(objroot)config.status
+	rm -f $(objroot)config.stamp
+	rm -f $(cfghdrs_out)
+	rm -f $(cfgoutputs_out)
+
+relclean: distclean
+	rm -f $(objroot)configure
+	rm -f $(srcroot)VERSION
+	rm -f $(DOCS_HTML)
+	rm -f $(DOCS_MAN3)
+
+#===============================================================================
+# Re-configuration rules.
+
+ifeq ($(enable_autogen), 1)
+$(srcroot)configure : $(srcroot)configure.ac
+	cd ./$(srcroot) && $(AUTOCONF)
+
+$(objroot)config.status : $(srcroot)configure
+	./$(objroot)config.status --recheck
+
+$(srcroot)config.stamp.in : $(srcroot)configure.ac
+	echo stamp > $(srcroot)config.stamp.in
+
+$(objroot)config.stamp : $(cfgoutputs_in) $(cfghdrs_in) $(srcroot)configure
+	./$(objroot)config.status
+	@touch $@
+
+# There must be some action in order for make to re-read Makefile when it is
+# out of date.
+$(cfgoutputs_out) $(cfghdrs_out) : $(objroot)config.stamp
+	@true
+endif
--- a/16
+++ b/16
@@ -0,0 +1,16 @@
+jemalloc is a general-purpose scalable concurrent malloc(3) implementation.
+This distribution is a "portable" implementation that currently targets
+FreeBSD, Linux, Apple OS X, and MinGW.  jemalloc is included as the default
+allocator in the FreeBSD and NetBSD operating systems, and it is used by the
+Mozilla Firefox web browser on Microsoft Windows-related platforms.  Depending
+on your needs, one of the other divergent versions may suit your needs better
+than this distribution.
+
+The COPYING file contains copyright and licensing information.
+
+The INSTALL file contains information on how to configure, build, and install
+jemalloc.
+
+The ChangeLog file contains a brief summary of changes for each release.
+
+URL: http://www.canonware.com/jemalloc/
--- a/jemalloc/autogen.sh
+++ b/jemalloc/autogen.sh
--- a/bin/jemalloc.sh.in
+++ b/bin/jemalloc.sh.in
@@ -0,0 +1,9 @@
+#!/bin/sh
+
+prefix=@prefix@
+exec_prefix=@exec_prefix@
+libdir=@libdir@
+
+@LD_PRELOAD_VAR@=${libdir}/libjemalloc.@SOREV@
+export @LD_PRELOAD_VAR@
+exec "$@"
--- a/jemalloc/bin/pprof
+++ b/jemalloc/bin/pprof
--- a/jemalloc/config.guess
+++ b/jemalloc/config.guess
--- a/jemalloc/config.stamp.in
+++ b/jemalloc/config.stamp.in
--- a/jemalloc/config.sub
+++ b/jemalloc/config.sub
@@ -1,9 +1,10 @@
 #! /bin/sh
 # Configuration validation subroutine script.
 #   Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
-#   2000, 2001, 2002, 2003 Free Software Foundation, Inc.
+#   2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010,
+#   2011, 2012 Free Software Foundation, Inc.

-timestamp='2004-02-23'
+timestamp='2012-02-10'

 # This file is (in principle) common to ALL GNU software.
 # The presence of a machine in this file suggests that SOME GNU software
@@ -20,23 +21,25 @@ timestamp='2004-02-23'
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU General Public License
-# along with this program; if not, write to the Free Software
-# Foundation, Inc., 59 Temple Place - Suite 330,
-# Boston, MA 02111-1307, USA.
-
+# along with this program; if not, see <http://www.gnu.org/licenses/>.
+#
 # As a special exception to the GNU General Public License, if you
 # distribute this file as part of a program that contains a
 # configuration script generated by Autoconf, you may include it under
 # the same distribution terms that you use for the rest of that program.

+
 # Please send patches to <config-patches@gnu.org>.  Submit a context
-# diff and a properly formatted ChangeLog entry.
+# diff and a properly formatted GNU ChangeLog entry.
 #
 # Configuration subroutine to validate and canonicalize a configuration type.
 # Supply the specified configuration type as an argument.
 # If it is invalid, we print an error message on stderr and exit with code 1.
 # Otherwise, we print the canonical config type on stdout and succeed.

+# You can get the latest version of this script from:
+# http://git.savannah.gnu.org/gitweb/?p=config.git;a=blob_plain;f=config.sub;hb=HEAD
+
 # This file is supposed to be the same for all GNU packages
 # and recognize all the CPU types, system types and aliases
 # that are meaningful with *any* GNU software.
@@ -70,7 +73,8 @@ Report bugs and patches to <config-patches@gnu.org>."
 version="\
 GNU config.sub ($timestamp)

-Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001
+Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000,
+2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012
 Free Software Foundation, Inc.

 This is free software; see the source for copying conditions.  There is NO
@@ -83,11 +87,11 @@ Try \`$me --help' for more information."
 while test $# -gt 0 ; do
  case $1 in
    --time-stamp | --time* | -t )
-       echo "$timestamp" ; exit 0 ;;
+       echo "$timestamp" ; exit ;;
    --version | -v )
-       echo "$version" ; exit 0 ;;
+       echo "$version" ; exit ;;
    --help | --h* | -h )
-       echo "$usage"; exit 0 ;;
+       echo "$usage"; exit ;;
    -- )     # Stop option processing
       shift; break ;;
    - )	# Use stdin as input.
@@ -99,7 +103,7 @@ while test $# -gt 0 ; do
    *local*)
       # First pass through any local machine types.
       echo $1
-       exit 0;;
+       exit ;;

    * )
       break ;;
@@ -118,11 +122,18 @@ esac
 # Here we must recognize all the valid KERNEL-OS combinations.
 maybe_os=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\2/'`
 case $maybe_os in
-  nto-qnx* | linux-gnu* | linux-dietlibc | linux-uclibc* | uclinux-uclibc* | uclinux-gnu* | \
-  kfreebsd*-gnu* | knetbsd*-gnu* | netbsd*-gnu* | storm-chaos* | os2-emx* | rtmk-nova*)
+  nto-qnx* | linux-gnu* | linux-android* | linux-dietlibc | linux-newlib* | \
+  linux-uclibc* | uclinux-uclibc* | uclinux-gnu* | kfreebsd*-gnu* | \
+  knetbsd*-gnu* | netbsd*-gnu* | \
+  kopensolaris*-gnu* | \
+  storm-chaos* | os2-emx* | rtmk-nova*)
    os=-$maybe_os
    basic_machine=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\1/'`
    ;;
+  android-linux)
+    os=-linux-android
+    basic_machine=`echo $1 | sed 's/^\(.*\)-\([^-]*-[^-]*\)$/\1/'`-unknown
+    ;;
  *)
    basic_machine=`echo $1 | sed 's/-[^-]*$//'`
    if [ $basic_machine != $1 ]
@@ -145,10 +156,13 @@ case $os in
 	-convergent* | -ncr* | -news | -32* | -3600* | -3100* | -hitachi* |\
 	-c[123]* | -convex* | -sun | -crds | -omron* | -dg | -ultra | -tti* | \
 	-harris | -dolphin | -highlevel | -gould | -cbm | -ns | -masscomp | \
-	-apple | -axis)
+	-apple | -axis | -knuth | -cray | -microblaze)
 		os=
 		basic_machine=$1
 		;;
+	-bluegene*)
+		os=-cnk
+		;;
 	-sim | -cisco | -oki | -wec | -winbond)
 		os=
 		basic_machine=$1
@@ -163,13 +177,17 @@ case $os in
 		os=-chorusos
 		basic_machine=$1
 		;;
- 	-chorusrdb)
- 		os=-chorusrdb
+	-chorusrdb)
+		os=-chorusrdb
 		basic_machine=$1
- 		;;
+		;;
 	-hiux*)
 		os=-hiuxwe2
 		;;
+	-sco6)
+		os=-sco5v6
+		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		;;
 	-sco5)
 		os=-sco3.2v5
 		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
@@ -186,6 +204,10 @@ case $os in
 		# Don't forget version if it is 3.2v4 or newer.
 		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
 		;;
+	-sco5v6*)
+		# Don't forget version if it is 3.2v4 or newer.
+		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
+		;;
 	-sco*)
 		os=-sco3.2v2
 		basic_machine=`echo $1 | sed -e 's/86-.*/86-pc/'`
@@ -227,25 +249,36 @@ case $basic_machine in
 	# Some are omitted here because they have special meanings below.
 	1750a | 580 \
 	| a29k \
+	| aarch64 | aarch64_be \
 	| alpha | alphaev[4-8] | alphaev56 | alphaev6[78] | alphapca5[67] \
 	| alpha64 | alpha64ev[4-8] | alpha64ev56 | alpha64ev6[78] | alpha64pca5[67] \
 	| am33_2.0 \
-	| arc | arm | arm[bl]e | arme[lb] | armv[2345] | armv[345][lb] | avr \
+	| arc | arm | arm[bl]e | arme[lb] | armv[2345] | armv[345][lb] | avr | avr32 \
+        | be32 | be64 \
+	| bfin \
 	| c4x | clipper \
 	| d10v | d30v | dlx | dsp16xx \
-	| fr30 | frv \
+	| epiphany \
+	| fido | fr30 | frv \
 	| h8300 | h8500 | hppa | hppa1.[01] | hppa2.0 | hppa2.0[nw] | hppa64 \
+	| hexagon \
 	| i370 | i860 | i960 | ia64 \
 	| ip2k | iq2000 \
-	| m32r | m68000 | m68k | m88k | mcore \
+	| le32 | le64 \
+	| lm32 \
+	| m32c | m32r | m32rle | m68000 | m68k | m88k \
+	| maxq | mb | microblaze | mcore | mep | metag \
 	| mips | mipsbe | mipseb | mipsel | mipsle \
 	| mips16 \
 	| mips64 | mips64el \
-	| mips64vr | mips64vrel \
+	| mips64octeon | mips64octeonel \
 	| mips64orion | mips64orionel \
+	| mips64r5900 | mips64r5900el \
+	| mips64vr | mips64vrel \
 	| mips64vr4100 | mips64vr4100el \
 	| mips64vr4300 | mips64vr4300el \
 	| mips64vr5000 | mips64vr5000el \
+	| mips64vr5900 | mips64vr5900el \
 	| mipsisa32 | mipsisa32el \
 	| mipsisa32r2 | mipsisa32r2el \
 	| mipsisa64 | mipsisa64el \
@@ -254,30 +287,65 @@ case $basic_machine in
 	| mipsisa64sr71k | mipsisa64sr71kel \
 	| mipstx39 | mipstx39el \
 	| mn10200 | mn10300 \
+	| moxie \
+	| mt \
 	| msp430 \
+	| nds32 | nds32le | nds32be \
+	| nios | nios2 \
 	| ns16k | ns32k \
-	| openrisc | or32 \
+	| open8 \
+	| or32 \
 	| pdp10 | pdp11 | pj | pjl \
-	| powerpc | powerpc64 | powerpc64le | powerpcle | ppcbe \
+	| powerpc | powerpc64 | powerpc64le | powerpcle \
 	| pyramid \
-	| sh | sh[1234] | sh[23]e | sh[34]eb | shbe | shle | sh[1234]le | sh3ele \
+	| rl78 | rx \
+	| score \
+	| sh | sh[1234] | sh[24]a | sh[24]aeb | sh[23]e | sh[34]eb | sheb | shbe | shle | sh[1234]le | sh3ele \
 	| sh64 | sh64le \
-	| sparc | sparc64 | sparc86x | sparclet | sparclite | sparcv9 | sparcv9b \
-	| strongarm \
-	| tahoe | thumb | tic4x | tic80 | tron \
-	| v850 | v850e \
+	| sparc | sparc64 | sparc64b | sparc64v | sparc86x | sparclet | sparclite \
+	| sparcv8 | sparcv9 | sparcv9b | sparcv9v \
+	| spu \
+	| tahoe | tic4x | tic54x | tic55x | tic6x | tic80 | tron \
+	| ubicom32 \
+	| v850 | v850e | v850e1 | v850e2 | v850es | v850e2v3 \
 	| we32k \
-	| x86 | xscale | xstormy16 | xtensa \
-	| z8k)
+	| x86 | xc16x | xstormy16 | xtensa \
+	| z8k | z80)
 		basic_machine=$basic_machine-unknown
 		;;
-	m6811 | m68hc11 | m6812 | m68hc12)
-		# Motorola 68HC11/12.
+	c54x)
+		basic_machine=tic54x-unknown
+		;;
+	c55x)
+		basic_machine=tic55x-unknown
+		;;
+	c6x)
+		basic_machine=tic6x-unknown
+		;;
+	m6811 | m68hc11 | m6812 | m68hc12 | m68hcs12x | picochip)
 		basic_machine=$basic_machine-unknown
 		os=-none
 		;;
 	m88110 | m680[12346]0 | m683?2 | m68360 | m5200 | v70 | w65 | z8k)
 		;;
+	ms1)
+		basic_machine=mt-unknown
+		;;
+
+	strongarm | thumb | xscale)
+		basic_machine=arm-unknown
+		;;
+	xgate)
+		basic_machine=$basic_machine-unknown
+		os=-none
+		;;
+	xscaleeb)
+		basic_machine=armeb-unknown
+		;;
+
+	xscaleel)
+		basic_machine=armel-unknown
+		;;

 	# We use `pc' rather than `unknown'
 	# because (1) that's what they normally are, and
@@ -293,32 +361,40 @@ case $basic_machine in
 	# Recognize the basic CPU types with company name.
 	580-* \
 	| a29k-* \
+	| aarch64-* | aarch64_be-* \
 	| alpha-* | alphaev[4-8]-* | alphaev56-* | alphaev6[78]-* \
 	| alpha64-* | alpha64ev[4-8]-* | alpha64ev56-* | alpha64ev6[78]-* \
 	| alphapca5[67]-* | alpha64pca5[67]-* | arc-* \
 	| arm-*  | armbe-* | armle-* | armeb-* | armv*-* \
-	| avr-* \
-	| bs2000-* \
-	| c[123]* | c30-* | [cjt]90-* | c4x-* | c54x-* | c55x-* | c6x-* \
-	| clipper-* | cydra-* \
+	| avr-* | avr32-* \
+	| be32-* | be64-* \
+	| bfin-* | bs2000-* \
+	| c[123]* | c30-* | [cjt]90-* | c4x-* \
+	| clipper-* | craynv-* | cydra-* \
 	| d10v-* | d30v-* | dlx-* \
 	| elxsi-* \
-	| f30[01]-* | f700-* | fr30-* | frv-* | fx80-* \
+	| f30[01]-* | f700-* | fido-* | fr30-* | frv-* | fx80-* \
 	| h8300-* | h8500-* \
 	| hppa-* | hppa1.[01]-* | hppa2.0-* | hppa2.0[nw]-* | hppa64-* \
+	| hexagon-* \
 	| i*86-* | i860-* | i960-* | ia64-* \
 	| ip2k-* | iq2000-* \
-	| m32r-* \
+	| le32-* | le64-* \
+	| lm32-* \
+	| m32c-* | m32r-* | m32rle-* \
 	| m68000-* | m680[012346]0-* | m68360-* | m683?2-* | m68k-* \
-	| m88110-* | m88k-* | mcore-* \
+	| m88110-* | m88k-* | maxq-* | mcore-* | metag-* | microblaze-* \
 	| mips-* | mipsbe-* | mipseb-* | mipsel-* | mipsle-* \
 	| mips16-* \
 	| mips64-* | mips64el-* \
-	| mips64vr-* | mips64vrel-* \
+	| mips64octeon-* | mips64octeonel-* \
 	| mips64orion-* | mips64orionel-* \
+	| mips64r5900-* | mips64r5900el-* \
+	| mips64vr-* | mips64vrel-* \
 	| mips64vr4100-* | mips64vr4100el-* \
 	| mips64vr4300-* | mips64vr4300el-* \
 	| mips64vr5000-* | mips64vr5000el-* \
+	| mips64vr5900-* | mips64vr5900el-* \
 	| mipsisa32-* | mipsisa32el-* \
 	| mipsisa32r2-* | mipsisa32r2el-* \
 	| mipsisa64-* | mipsisa64el-* \
@@ -326,26 +402,39 @@ case $basic_machine in
 	| mipsisa64sb1-* | mipsisa64sb1el-* \
 	| mipsisa64sr71k-* | mipsisa64sr71kel-* \
 	| mipstx39-* | mipstx39el-* \
+	| mmix-* \
+	| mt-* \
 	| msp430-* \
-	| none-* | np1-* | nv1-* | ns16k-* | ns32k-* \
+	| nds32-* | nds32le-* | nds32be-* \
+	| nios-* | nios2-* \
+	| none-* | np1-* | ns16k-* | ns32k-* \
+	| open8-* \
 	| orion-* \
 	| pdp10-* | pdp11-* | pj-* | pjl-* | pn-* | power-* \
-	| powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* | ppcbe-* \
+	| powerpc-* | powerpc64-* | powerpc64le-* | powerpcle-* \
 	| pyramid-* \
-	| romp-* | rs6000-* \
-	| sh-* | sh[1234]-* | sh[23]e-* | sh[34]eb-* | shbe-* \
+	| rl78-* | romp-* | rs6000-* | rx-* \
+	| sh-* | sh[1234]-* | sh[24]a-* | sh[24]aeb-* | sh[23]e-* | sh[34]eb-* | sheb-* | shbe-* \
 	| shle-* | sh[1234]le-* | sh3ele-* | sh64-* | sh64le-* \
-	| sparc-* | sparc64-* | sparc86x-* | sparclet-* | sparclite-* \
-	| sparcv9-* | sparcv9b-* | strongarm-* | sv1-* | sx?-* \
-	| tahoe-* | thumb-* \
+	| sparc-* | sparc64-* | sparc64b-* | sparc64v-* | sparc86x-* | sparclet-* \
+	| sparclite-* \
+	| sparcv8-* | sparcv9-* | sparcv9b-* | sparcv9v-* | sv1-* | sx?-* \
+	| tahoe-* \
 	| tic30-* | tic4x-* | tic54x-* | tic55x-* | tic6x-* | tic80-* \
+	| tile*-* \
 	| tron-* \
-	| v850-* | v850e-* | vax-* \
+	| ubicom32-* \
+	| v850-* | v850e-* | v850e1-* | v850es-* | v850e2-* | v850e2v3-* \
+	| vax-* \
 	| we32k-* \
-	| x86-* | x86_64-* | xps100-* | xscale-* | xstormy16-* \
-	| xtensa-* \
+	| x86-* | x86_64-* | xc16x-* | xps100-* \
+	| xstormy16-* | xtensa*-* \
 	| ymp-* \
-	| z8k-*)
+	| z8k-* | z80-*)
+		;;
+	# Recognize the basic CPU types without company name, with glob match.
+	xtensa*)
+		basic_machine=$basic_machine-unknown
 		;;
 	# Recognize the various machine names and aliases which stand
 	# for a CPU type and a company and sometimes even an OS.
@@ -363,7 +452,7 @@ case $basic_machine in
 		basic_machine=a29k-amd
 		os=-udi
 		;;
-    	abacus)
+	abacus)
 		basic_machine=abacus-unknown
 		;;
 	adobe68k)
@@ -409,6 +498,10 @@ case $basic_machine in
 		basic_machine=m68k-apollo
 		os=-bsd
 		;;
+	aros)
+		basic_machine=i386-pc
+		os=-aros
+		;;
 	aux)
 		basic_machine=m68k-apple
 		os=-aux
@@ -417,10 +510,35 @@ case $basic_machine in
 		basic_machine=ns32k-sequent
 		os=-dynix
 		;;
+	blackfin)
+		basic_machine=bfin-unknown
+		os=-linux
+		;;
+	blackfin-*)
+		basic_machine=bfin-`echo $basic_machine | sed 's/^[^-]*-//'`
+		os=-linux
+		;;
+	bluegene*)
+		basic_machine=powerpc-ibm
+		os=-cnk
+		;;
+	c54x-*)
+		basic_machine=tic54x-`echo $basic_machine | sed 's/^[^-]*-//'`
+		;;
+	c55x-*)
+		basic_machine=tic55x-`echo $basic_machine | sed 's/^[^-]*-//'`
+		;;
+	c6x-*)
+		basic_machine=tic6x-`echo $basic_machine | sed 's/^[^-]*-//'`
+		;;
 	c90)
 		basic_machine=c90-cray
 		os=-unicos
 		;;
+	cegcc)
+		basic_machine=arm-unknown
+		os=-cegcc
+		;;
 	convex-c1)
 		basic_machine=c1-convex
 		os=-bsd
@@ -445,13 +563,20 @@ case $basic_machine in
 		basic_machine=j90-cray
 		os=-unicos
 		;;
-	cr16c)
-		basic_machine=cr16c-unknown
+	craynv)
+		basic_machine=craynv-cray
+		os=-unicosmp
+		;;
+	cr16 | cr16-*)
+		basic_machine=cr16-unknown
 		os=-elf
 		;;
 	crds | unos)
 		basic_machine=m68k-crds
 		;;
+	crisv32 | crisv32-* | etraxfs*)
+		basic_machine=crisv32-axis
+		;;
 	cris | cris-* | etrax*)
 		basic_machine=cris-axis
 		;;
@@ -481,6 +606,14 @@ case $basic_machine in
 		basic_machine=m88k-motorola
 		os=-sysv3
 		;;
+	dicos)
+		basic_machine=i686-pc
+		os=-dicos
+		;;
+	djgpp)
+		basic_machine=i586-pc
+		os=-msdosdjgpp
+		;;
 	dpx20 | dpx20-*)
 		basic_machine=rs6000-bull
 		os=-bosx
@@ -592,7 +725,6 @@ case $basic_machine in
 	i370-ibm* | ibm*)
 		basic_machine=i370-ibm
 		;;
-# I'm not sure what "Sysv32" means.  Should this be sysv3.2?
 	i*86v32)
 		basic_machine=`echo $1 | sed -e 's/86.*/86-pc/'`
 		os=-sysv32
@@ -631,6 +763,14 @@ case $basic_machine in
 		basic_machine=m68k-isi
 		os=-sysv
 		;;
+	m68knommu)
+		basic_machine=m68k-unknown
+		os=-linux
+		;;
+	m68knommu-*)
+		basic_machine=m68k-`echo $basic_machine | sed 's/^[^-]*-//'`
+		os=-linux
+		;;
 	m88k-omron*)
 		basic_machine=m88k-omron
 		;;
@@ -642,10 +782,17 @@ case $basic_machine in
 		basic_machine=ns32k-utek
 		os=-sysv
 		;;
+	microblaze)
+		basic_machine=microblaze-xilinx
+		;;
 	mingw32)
 		basic_machine=i386-pc
 		os=-mingw32
 		;;
+	mingw32ce)
+		basic_machine=arm-unknown
+		os=-mingw32ce
+		;;
 	miniframe)
 		basic_machine=m68000-convergent
 		;;
@@ -659,10 +806,6 @@ case $basic_machine in
 	mips3*)
 		basic_machine=`echo $basic_machine | sed -e 's/mips3/mips64/'`-unknown
 		;;
-	mmix*)
-		basic_machine=mmix-knuth
-		os=-mmixware
-		;;
 	monitor)
 		basic_machine=m68k-rom68k
 		os=-coff
@@ -675,10 +818,21 @@ case $basic_machine in
 		basic_machine=i386-pc
 		os=-msdos
 		;;
+	ms1-*)
+		basic_machine=`echo $basic_machine | sed -e 's/ms1-/mt-/'`
+		;;
+	msys)
+		basic_machine=i386-pc
+		os=-msys
+		;;
 	mvs)
 		basic_machine=i370-ibm
 		os=-mvs
 		;;
+	nacl)
+		basic_machine=le32-unknown
+		os=-nacl
+		;;
 	ncr3000)
 		basic_machine=i486-ncr
 		os=-sysv4
@@ -743,9 +897,11 @@ case $basic_machine in
 	np1)
 		basic_machine=np1-gould
 		;;
-	nv1)
-		basic_machine=nv1-cray
-		os=-unicosmp
+	neo-tandem)
+		basic_machine=neo-tandem
+		;;
+	nse-tandem)
+		basic_machine=nse-tandem
 		;;
 	nsr-tandem)
 		basic_machine=nsr-tandem
@@ -754,9 +910,8 @@ case $basic_machine in
 		basic_machine=hppa1.1-oki
 		os=-proelf
 		;;
-	or32 | or32-*)
+	openrisc | openrisc-*)
 		basic_machine=or32-unknown
-		os=-coff
 		;;
 	os400)
 		basic_machine=powerpc-ibm
@@ -778,6 +933,14 @@ case $basic_machine in
 		basic_machine=i860-intel
 		os=-osf
 		;;
+	parisc)
+		basic_machine=hppa-unknown
+		os=-linux
+		;;
+	parisc-*)
+		basic_machine=hppa-`echo $basic_machine | sed 's/^[^-]*-//'`
+		os=-linux
+		;;
 	pbd)
 		basic_machine=sparc-tti
 		;;
@@ -787,6 +950,12 @@ case $basic_machine in
 	pc532 | pc532-*)
 		basic_machine=ns32k-pc532
 		;;
+	pc98)
+		basic_machine=i386-pc
+		;;
+	pc98-*)
+		basic_machine=i386-`echo $basic_machine | sed 's/^[^-]*-//'`
+		;;
 	pentium | p5 | k5 | k6 | nexgen | viac3)
 		basic_machine=i586-pc
 		;;
@@ -816,9 +985,10 @@ case $basic_machine in
 		;;
 	power)	basic_machine=power-ibm
 		;;
-	ppc)	basic_machine=powerpc-unknown
+	ppc | ppcbe)	basic_machine=powerpc-unknown
 		;;
-	ppc-*)	basic_machine=powerpc-`echo $basic_machine | sed 's/^[^-]*-//'`
+	ppc-* | ppcbe-*)
+		basic_machine=powerpc-`echo $basic_machine | sed 's/^[^-]*-//'`
 		;;
 	ppcle | powerpclittle | ppc-le | powerpc-little)
 		basic_machine=powerpcle-unknown
@@ -843,6 +1013,10 @@ case $basic_machine in
 		basic_machine=i586-unknown
 		os=-pw32
 		;;
+	rdos)
+		basic_machine=i386-pc
+		os=-rdos
+		;;
 	rom68k)
 		basic_machine=m68k-rom68k
 		os=-coff
@@ -869,6 +1043,10 @@ case $basic_machine in
 	sb1el)
 		basic_machine=mipsisa64sb1el-unknown
 		;;
+	sde)
+		basic_machine=mipsisa32-sde
+		os=-elf
+		;;
 	sei)
 		basic_machine=mips-sei
 		os=-seiux
@@ -880,6 +1058,9 @@ case $basic_machine in
 		basic_machine=sh-hitachi
 		os=-hms
 		;;
+	sh5el)
+		basic_machine=sh5le-unknown
+		;;
 	sh64)
 		basic_machine=sh64-unknown
 		;;
@@ -901,6 +1082,9 @@ case $basic_machine in
 		basic_machine=i860-stratus
 		os=-sysv4
 		;;
+	strongarm-* | thumb-*)
+		basic_machine=arm-`echo $basic_machine | sed 's/^[^-]*-//'`
+		;;
 	sun2)
 		basic_machine=m68000-sun
 		;;
@@ -957,17 +1141,9 @@ case $basic_machine in
 		basic_machine=t90-cray
 		os=-unicos
 		;;
-	tic54x | c54x*)
-		basic_machine=tic54x-unknown
-		os=-coff
-		;;
-	tic55x | c55x*)
-		basic_machine=tic55x-unknown
-		os=-coff
-		;;
-	tic6x | c6x*)
-		basic_machine=tic6x-unknown
-		os=-coff
+	tile*)
+		basic_machine=$basic_machine-unknown
+		os=-linux-gnu
 		;;
 	tx39)
 		basic_machine=mipstx39-unknown
@@ -1029,9 +1205,16 @@ case $basic_machine in
 		basic_machine=hppa1.1-winbond
 		os=-proelf
 		;;
+	xbox)
+		basic_machine=i686-pc
+		os=-mingw32
+		;;
 	xps | xps100)
 		basic_machine=xps100-honeywell
 		;;
+	xscale-* | xscalee[bl]-*)
+		basic_machine=`echo $basic_machine | sed 's/^xscale/arm/'`
+		;;
 	ymp)
 		basic_machine=ymp-cray
 		os=-unicos
@@ -1040,6 +1223,10 @@ case $basic_machine in
 		basic_machine=z8k-unknown
 		os=-sim
 		;;
+	z80-*-coff)
+		basic_machine=z80-unknown
+		os=-sim
+		;;
 	none)
 		basic_machine=none-none
 		os=-none
@@ -1059,6 +1246,9 @@ case $basic_machine in
 	romp)
 		basic_machine=romp-ibm
 		;;
+	mmix)
+		basic_machine=mmix-knuth
+		;;
 	rs6000)
 		basic_machine=rs6000-ibm
 		;;
@@ -1075,13 +1265,10 @@ case $basic_machine in
 	we32k)
 		basic_machine=we32k-att
 		;;
-	sh3 | sh4 | sh[34]eb | sh[1234]le | sh[23]ele)
+	sh[1234] | sh[24]a | sh[24]aeb | sh[34]eb | sh[1234]le | sh[23]ele)
 		basic_machine=sh-unknown
 		;;
-	sh64)
-		basic_machine=sh64-unknown
-		;;
-	sparc | sparcv9 | sparcv9b)
+	sparc | sparcv8 | sparcv9 | sparcv9b | sparcv9v)
 		basic_machine=sparc-sun
 		;;
 	cydra)
@@ -1125,9 +1312,12 @@ esac
 if [ x"$os" != x"" ]
 then
 case $os in
-        # First match some system type aliases
-        # that might get confused with valid system types.
+	# First match some system type aliases
+	# that might get confused with valid system types.
 	# -solaris* is a basic system type, with this one exception.
+	-auroraux)
+		os=-auroraux
+		;;
 	-solaris1 | -solaris1.*)
 		os=`echo $os | sed -e 's|solaris1|sunos4|'`
 		;;
@@ -1148,26 +1338,31 @@ case $os in
 	# Each alternative MUST END IN A *, to match a version number.
 	# -sysv* is not here because it comes later, after sysvr4.
 	-gnu* | -bsd* | -mach* | -minix* | -genix* | -ultrix* | -irix* \
-	      | -*vms* | -sco* | -esix* | -isc* | -aix* | -sunos | -sunos[34]*\
-	      | -hpux* | -unos* | -osf* | -luna* | -dgux* | -solaris* | -sym* \
+	      | -*vms* | -sco* | -esix* | -isc* | -aix* | -cnk* | -sunos | -sunos[34]*\
+	      | -hpux* | -unos* | -osf* | -luna* | -dgux* | -auroraux* | -solaris* \
+	      | -sym* | -kopensolaris* \
 	      | -amigaos* | -amigados* | -msdos* | -newsos* | -unicos* | -aof* \
-	      | -aos* \
+	      | -aos* | -aros* \
 	      | -nindy* | -vxsim* | -vxworks* | -ebmon* | -hms* | -mvs* \
 	      | -clix* | -riscos* | -uniplus* | -iris* | -rtu* | -xenix* \
-	      | -hiux* | -386bsd* | -knetbsd* | -mirbsd* | -netbsd* | -openbsd* \
+	      | -hiux* | -386bsd* | -knetbsd* | -mirbsd* | -netbsd* \
+	      | -openbsd* | -solidbsd* \
 	      | -ekkobsd* | -kfreebsd* | -freebsd* | -riscix* | -lynxos* \
 	      | -bosx* | -nextstep* | -cxux* | -aout* | -elf* | -oabi* \
 	      | -ptx* | -coff* | -ecoff* | -winnt* | -domain* | -vsta* \
 	      | -udi* | -eabi* | -lites* | -ieee* | -go32* | -aux* \
-	      | -chorusos* | -chorusrdb* \
-	      | -cygwin* | -pe* | -psos* | -moss* | -proelf* | -rtems* \
-	      | -mingw32* | -linux-gnu* | -linux-uclibc* | -uxpv* | -beos* | -mpeix* | -udk* \
+	      | -chorusos* | -chorusrdb* | -cegcc* \
+	      | -cygwin* | -msys* | -pe* | -psos* | -moss* | -proelf* | -rtems* \
+	      | -mingw32* | -linux-gnu* | -linux-android* \
+	      | -linux-newlib* | -linux-uclibc* \
+	      | -uxpv* | -beos* | -mpeix* | -udk* \
 	      | -interix* | -uwin* | -mks* | -rhapsody* | -darwin* | -opened* \
 	      | -openstep* | -oskit* | -conix* | -pw32* | -nonstopux* \
 	      | -storm-chaos* | -tops10* | -tenex* | -tops20* | -its* \
 	      | -os2* | -vos* | -palmos* | -uclinux* | -nucleus* \
 	      | -morphos* | -superux* | -rtmk* | -rtmk-nova* | -windiss* \
-	      | -powermax* | -dnix* | -nx6 | -nx7 | -sei* | -dragonfly*)
+	      | -powermax* | -dnix* | -nx6 | -nx7 | -sei* | -dragonfly* \
+	      | -skyos* | -haiku* | -rdos* | -toppers* | -drops* | -es*)
 	# Remember, each alternative MUST END IN *, to match a version number.
 		;;
 	-qnx*)
@@ -1185,7 +1380,7 @@ case $os in
 		os=`echo $os | sed -e 's|nto|nto-qnx|'`
 		;;
 	-sim | -es1800* | -hms* | -xray | -os68k* | -none* | -v88r* \
-	      | -windows* | -osx | -abug | -netware* | -os9* | -beos* \
+	      | -windows* | -osx | -abug | -netware* | -os9* | -beos* | -haiku* \
 	      | -macos* | -mpw* | -magic* | -mmixware* | -mon960* | -lnews*)
 		;;
 	-mac*)
@@ -1206,7 +1401,7 @@ case $os in
 	-opened*)
 		os=-openedition
 		;;
-        -os400*)
+	-os400*)
 		os=-os400
 		;;
 	-wince*)
@@ -1255,7 +1450,7 @@ case $os in
 	-sinix*)
 		os=-sysv4
 		;;
-        -tpf*)
+	-tpf*)
 		os=-tpf
 		;;
 	-triton*)
@@ -1294,6 +1489,14 @@ case $os in
 	-kaos*)
 		os=-kaos
 		;;
+	-zvmoe)
+		os=-zvmoe
+		;;
+	-dicos*)
+		os=-dicos
+		;;
+	-nacl*)
+		;;
 	-none)
 		;;
 	*)
@@ -1316,6 +1519,12 @@ else
 # system, and we'll never get to this point.

 case $basic_machine in
+	score-*)
+		os=-elf
+		;;
+	spu-*)
+		os=-elf
+		;;
 	*-acorn)
 		os=-riscix1.2
 		;;
@@ -1325,9 +1534,18 @@ case $basic_machine in
 	arm*-semi)
 		os=-aout
 		;;
-    c4x-* | tic4x-*)
-        os=-coff
-        ;;
+	c4x-* | tic4x-*)
+		os=-coff
+		;;
+	tic54x-*)
+		os=-coff
+		;;
+	tic55x-*)
+		os=-coff
+		;;
+	tic6x-*)
+		os=-coff
+		;;
 	# This must come before the *-dec entry.
 	pdp10-*)
 		os=-tops20
@@ -1346,13 +1564,13 @@ case $basic_machine in
 		;;
 	m68000-sun)
 		os=-sunos3
-		# This also exists in the configure program, but was not the
-		# default.
-		# os=-sunos4
 		;;
 	m68*-cisco)
 		os=-aout
 		;;
+	mep-*)
+		os=-elf
+		;;
 	mips*-cisco)
 		os=-elf
 		;;
@@ -1371,9 +1589,15 @@ case $basic_machine in
 	*-be)
 		os=-beos
 		;;
+	*-haiku)
+		os=-haiku
+		;;
 	*-ibm)
 		os=-aix
 		;;
+	*-knuth)
+		os=-mmixware
+		;;
 	*-wec)
 		os=-proelf
 		;;
@@ -1476,7 +1700,7 @@ case $basic_machine in
 			-sunos*)
 				vendor=sun
 				;;
-			-aix*)
+			-cnk*|-aix*)
 				vendor=ibm
 				;;
 			-beos*)
@@ -1539,7 +1763,7 @@ case $basic_machine in
 esac

 echo $basic_machine$os
-exit 0
+exit

 # Local variables:
 # eval: (add-hook 'write-file-hooks 'time-stamp)
--- a/configure.ac
+++ b/configure.ac
--- a/jemalloc/doc/html.xsl.in
+++ b/jemalloc/doc/html.xsl.in
--- a/jemalloc/doc/jemalloc.xml.in
+++ b/jemalloc/doc/jemalloc.xml.in
--- a/jemalloc/doc/manpages.xsl.in
+++ b/jemalloc/doc/manpages.xsl.in
--- a/jemalloc/doc/stylesheet.xsl
+++ b/jemalloc/doc/stylesheet.xsl
--- a/include/jemalloc/internal/arena.h
+++ b/include/jemalloc/internal/arena.h
@@ -0,0 +1,980 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+/*
+ * RUN_MAX_OVRHD indicates maximum desired run header overhead.  Runs are sized
+ * as small as possible such that this setting is still honored, without
+ * violating other constraints.  The goal is to make runs as small as possible
+ * without exceeding a per run external fragmentation threshold.
+ *
+ * We use binary fixed point math for overhead computations, where the binary
+ * point is implicitly RUN_BFP bits to the left.
+ *
+ * Note that it is possible to set RUN_MAX_OVRHD low enough that it cannot be
+ * honored for some/all object sizes, since when heap profiling is enabled
+ * there is one pointer of header overhead per object (plus a constant).  This
+ * constraint is relaxed (ignored) for runs that are so small that the
+ * per-region overhead is greater than:
+ *
+ *   (RUN_MAX_OVRHD / (reg_interval << (3+RUN_BFP))
+ */
+#define	RUN_BFP			12
+/*                                    \/   Implicit binary fixed point. */
+#define	RUN_MAX_OVRHD		0x0000003dU
+#define	RUN_MAX_OVRHD_RELAX	0x00001800U
+
+/* Maximum number of regions in one run. */
+#define	LG_RUN_MAXREGS		11
+#define	RUN_MAXREGS		(1U << LG_RUN_MAXREGS)
+
+/*
+ * Minimum redzone size.  Redzones may be larger than this if necessary to
+ * preserve region alignment.
+ */
+#define	REDZONE_MINSIZE		16
+
+/*
+ * The minimum ratio of active:dirty pages per arena is computed as:
+ *
+ *   (nactive >> opt_lg_dirty_mult) >= ndirty
+ *
+ * So, supposing that opt_lg_dirty_mult is 5, there can be no less than 32
+ * times as many active pages as dirty pages.
+ */
+#define	LG_DIRTY_MULT_DEFAULT	5
+
+typedef struct arena_chunk_map_s arena_chunk_map_t;
+typedef struct arena_chunk_s arena_chunk_t;
+typedef struct arena_run_s arena_run_t;
+typedef struct arena_bin_info_s arena_bin_info_t;
+typedef struct arena_bin_s arena_bin_t;
+typedef struct arena_s arena_t;
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+/* Each element of the chunk map corresponds to one page within the chunk. */
+struct arena_chunk_map_s {
+#ifndef JEMALLOC_PROF
+	/*
+	 * Overlay prof_ctx in order to allow it to be referenced by dead code.
+	 * Such antics aren't warranted for per arena data structures, but
+	 * chunk map overhead accounts for a percentage of memory, rather than
+	 * being just a fixed cost.
+	 */
+	union {
+#endif
+	union {
+		/*
+		 * Linkage for run trees.  There are two disjoint uses:
+		 *
+		 * 1) arena_t's runs_avail_{clean,dirty} trees.
+		 * 2) arena_run_t conceptually uses this linkage for in-use
+		 *    non-full runs, rather than directly embedding linkage.
+		 */
+		rb_node(arena_chunk_map_t)	rb_link;
+		/*
+		 * List of runs currently in purgatory.  arena_chunk_purge()
+		 * temporarily allocates runs that contain dirty pages while
+		 * purging, so that other threads cannot use the runs while the
+		 * purging thread is operating without the arena lock held.
+		 */
+		ql_elm(arena_chunk_map_t)	ql_link;
+	}				u;
+
+	/* Profile counters, used for large object runs. */
+	prof_ctx_t			*prof_ctx;
+#ifndef JEMALLOC_PROF
+	}; /* union { ... }; */
+#endif
+
+	/*
+	 * Run address (or size) and various flags are stored together.  The bit
+	 * layout looks like (assuming 32-bit system):
+	 *
+	 *   ???????? ???????? ????nnnn nnnndula
+	 *
+	 * ? : Unallocated: Run address for first/last pages, unset for internal
+	 *                  pages.
+	 *     Small: Run page offset.
+	 *     Large: Run size for first page, unset for trailing pages.
+	 * n : binind for small size class, BININD_INVALID for large size class.
+	 * d : dirty?
+	 * u : unzeroed?
+	 * l : large?
+	 * a : allocated?
+	 *
+	 * Following are example bit patterns for the three types of runs.
+	 *
+	 * p : run page offset
+	 * s : run size
+	 * n : binind for size class; large objects set these to BININD_INVALID
+	 *     except for promoted allocations (see prof_promote)
+	 * x : don't care
+	 * - : 0
+	 * + : 1
+	 * [DULA] : bit set
+	 * [dula] : bit unset
+	 *
+	 *   Unallocated (clean):
+	 *     ssssssss ssssssss ssss++++ ++++du-a
+	 *     xxxxxxxx xxxxxxxx xxxxxxxx xxxx-Uxx
+	 *     ssssssss ssssssss ssss++++ ++++dU-a
+	 *
+	 *   Unallocated (dirty):
+	 *     ssssssss ssssssss ssss++++ ++++D--a
+	 *     xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx
+	 *     ssssssss ssssssss ssss++++ ++++D--a
+	 *
+	 *   Small:
+	 *     pppppppp pppppppp ppppnnnn nnnnd--A
+	 *     pppppppp pppppppp ppppnnnn nnnn---A
+	 *     pppppppp pppppppp ppppnnnn nnnnd--A
+	 *
+	 *   Large:
+	 *     ssssssss ssssssss ssss++++ ++++D-LA
+	 *     xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx
+	 *     -------- -------- ----++++ ++++D-LA
+	 *
+	 *   Large (sampled, size <= PAGE):
+	 *     ssssssss ssssssss ssssnnnn nnnnD-LA
+	 *
+	 *   Large (not sampled, size == PAGE):
+	 *     ssssssss ssssssss ssss++++ ++++D-LA
+	 */
+	size_t				bits;
+#define	CHUNK_MAP_BININD_SHIFT	4
+#define	BININD_INVALID		((size_t)0xffU)
+/*     CHUNK_MAP_BININD_MASK == (BININD_INVALID << CHUNK_MAP_BININD_SHIFT) */
+#define	CHUNK_MAP_BININD_MASK	((size_t)0xff0U)
+#define	CHUNK_MAP_BININD_INVALID CHUNK_MAP_BININD_MASK
+#define	CHUNK_MAP_FLAGS_MASK	((size_t)0xcU)
+#define	CHUNK_MAP_DIRTY		((size_t)0x8U)
+#define	CHUNK_MAP_UNZEROED	((size_t)0x4U)
+#define	CHUNK_MAP_LARGE		((size_t)0x2U)
+#define	CHUNK_MAP_ALLOCATED	((size_t)0x1U)
+#define	CHUNK_MAP_KEY		CHUNK_MAP_ALLOCATED
+};
+typedef rb_tree(arena_chunk_map_t) arena_avail_tree_t;
+typedef rb_tree(arena_chunk_map_t) arena_run_tree_t;
+
+/* Arena chunk header. */
+struct arena_chunk_s {
+	/* Arena that owns the chunk. */
+	arena_t		*arena;
+
+	/* Linkage for the arena's chunks_dirty list. */
+	ql_elm(arena_chunk_t) link_dirty;
+
+	/*
+	 * True if the chunk is currently in the chunks_dirty list, due to
+	 * having at some point contained one or more dirty pages.  Removal
+	 * from chunks_dirty is lazy, so (dirtied && ndirty == 0) is possible.
+	 */
+	bool		dirtied;
+
+	/* Number of dirty pages. */
+	size_t		ndirty;
+
+	/*
+	 * Map of pages within chunk that keeps track of free/large/small.  The
+	 * first map_bias entries are omitted, since the chunk header does not
+	 * need to be tracked in the map.  This omission saves a header page
+	 * for common chunk sizes (e.g. 4 MiB).
+	 */
+	arena_chunk_map_t map[1]; /* Dynamically sized. */
+};
+typedef rb_tree(arena_chunk_t) arena_chunk_tree_t;
+
+struct arena_run_s {
+	/* Bin this run is associated with. */
+	arena_bin_t	*bin;
+
+	/* Index of next region that has never been allocated, or nregs. */
+	uint32_t	nextind;
+
+	/* Number of free regions in run. */
+	unsigned	nfree;
+};
+
+/*
+ * Read-only information associated with each element of arena_t's bins array
+ * is stored separately, partly to reduce memory usage (only one copy, rather
+ * than one per arena), but mainly to avoid false cacheline sharing.
+ *
+ * Each run has the following layout:
+ *
+ *               /--------------------\
+ *               | arena_run_t header |
+ *               | ...                |
+ * bitmap_offset | bitmap             |
+ *               | ...                |
+ *   ctx0_offset | ctx map            |
+ *               | ...                |
+ *               |--------------------|
+ *               | redzone            |
+ *   reg0_offset | region 0           |
+ *               | redzone            |
+ *               |--------------------| \
+ *               | redzone            | |
+ *               | region 1           |  > reg_interval
+ *               | redzone            | /
+ *               |--------------------|
+ *               | ...                |
+ *               | ...                |
+ *               | ...                |
+ *               |--------------------|
+ *               | redzone            |
+ *               | region nregs-1     |
+ *               | redzone            |
+ *               |--------------------|
+ *               | alignment pad?     |
+ *               \--------------------/
+ *
+ * reg_interval has at least the same minimum alignment as reg_size; this
+ * preserves the alignment constraint that sa2u() depends on.  Alignment pad is
+ * either 0 or redzone_size; it is present only if needed to align reg0_offset.
+ */
+struct arena_bin_info_s {
+	/* Size of regions in a run for this bin's size class. */
+	size_t		reg_size;
+
+	/* Redzone size. */
+	size_t		redzone_size;
+
+	/* Interval between regions (reg_size + (redzone_size << 1)). */
+	size_t		reg_interval;
+
+	/* Total size of a run for this bin's size class. */
+	size_t		run_size;
+
+	/* Total number of regions in a run for this bin's size class. */
+	uint32_t	nregs;
+
+	/*
+	 * Offset of first bitmap_t element in a run header for this bin's size
+	 * class.
+	 */
+	uint32_t	bitmap_offset;
+
+	/*
+	 * Metadata used to manipulate bitmaps for runs associated with this
+	 * bin.
+	 */
+	bitmap_info_t	bitmap_info;
+
+	/*
+	 * Offset of first (prof_ctx_t *) in a run header for this bin's size
+	 * class, or 0 if (config_prof == false || opt_prof == false).
+	 */
+	uint32_t	ctx0_offset;
+
+	/* Offset of first region in a run for this bin's size class. */
+	uint32_t	reg0_offset;
+};
+
+struct arena_bin_s {
+	/*
+	 * All operations on runcur, runs, and stats require that lock be
+	 * locked.  Run allocation/deallocation are protected by the arena lock,
+	 * which may be acquired while holding one or more bin locks, but not
+	 * vise versa.
+	 */
+	malloc_mutex_t	lock;
+
+	/*
+	 * Current run being used to service allocations of this bin's size
+	 * class.
+	 */
+	arena_run_t	*runcur;
+
+	/*
+	 * Tree of non-full runs.  This tree is used when looking for an
+	 * existing run when runcur is no longer usable.  We choose the
+	 * non-full run that is lowest in memory; this policy tends to keep
+	 * objects packed well, and it can also help reduce the number of
+	 * almost-empty chunks.
+	 */
+	arena_run_tree_t runs;
+
+	/* Bin statistics. */
+	malloc_bin_stats_t stats;
+};
+
+struct arena_s {
+	/* This arena's index within the arenas array. */
+	unsigned		ind;
+
+	/*
+	 * Number of threads currently assigned to this arena.  This field is
+	 * protected by arenas_lock.
+	 */
+	unsigned		nthreads;
+
+	/*
+	 * There are three classes of arena operations from a locking
+	 * perspective:
+	 * 1) Thread asssignment (modifies nthreads) is protected by
+	 *    arenas_lock.
+	 * 2) Bin-related operations are protected by bin locks.
+	 * 3) Chunk- and run-related operations are protected by this mutex.
+	 */
+	malloc_mutex_t		lock;
+
+	arena_stats_t		stats;
+	/*
+	 * List of tcaches for extant threads associated with this arena.
+	 * Stats from these are merged incrementally, and at exit.
+	 */
+	ql_head(tcache_t)	tcache_ql;
+
+	uint64_t		prof_accumbytes;
+
+	dss_prec_t		dss_prec;
+
+	/* List of dirty-page-containing chunks this arena manages. */
+	ql_head(arena_chunk_t)	chunks_dirty;
+
+	/*
+	 * In order to avoid rapid chunk allocation/deallocation when an arena
+	 * oscillates right on the cusp of needing a new chunk, cache the most
+	 * recently freed chunk.  The spare is left in the arena's chunk trees
+	 * until it is deleted.
+	 *
+	 * There is one spare chunk per arena, rather than one spare total, in
+	 * order to avoid interactions between multiple threads that could make
+	 * a single spare inadequate.
+	 */
+	arena_chunk_t		*spare;
+
+	/* Number of pages in active runs. */
+	size_t			nactive;
+
+	/*
+	 * Current count of pages within unused runs that are potentially
+	 * dirty, and for which madvise(... MADV_DONTNEED) has not been called.
+	 * By tracking this, we can institute a limit on how much dirty unused
+	 * memory is mapped for each arena.
+	 */
+	size_t			ndirty;
+
+	/*
+	 * Approximate number of pages being purged.  It is possible for
+	 * multiple threads to purge dirty pages concurrently, and they use
+	 * npurgatory to indicate the total number of pages all threads are
+	 * attempting to purge.
+	 */
+	size_t			npurgatory;
+
+	/*
+	 * Size/address-ordered trees of this arena's available runs.  The trees
+	 * are used for first-best-fit run allocation.  The dirty tree contains
+	 * runs with dirty pages (i.e. very likely to have been touched and
+	 * therefore have associated physical pages), whereas the clean tree
+	 * contains runs with pages that either have no associated physical
+	 * pages, or have pages that the kernel may recycle at any time due to
+	 * previous madvise(2) calls.  The dirty tree is used in preference to
+	 * the clean tree for allocations, because using dirty pages reduces
+	 * the amount of dirty purging necessary to keep the active:dirty page
+	 * ratio below the purge threshold.
+	 */
+	arena_avail_tree_t	runs_avail_clean;
+	arena_avail_tree_t	runs_avail_dirty;
+
+	/* bins is used to store trees of free regions. */
+	arena_bin_t		bins[NBINS];
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+extern ssize_t	opt_lg_dirty_mult;
+/*
+ * small_size2bin is a compact lookup table that rounds request sizes up to
+ * size classes.  In order to reduce cache footprint, the table is compressed,
+ * and all accesses are via the SMALL_SIZE2BIN macro.
+ */
+extern uint8_t const	small_size2bin[];
+#define	SMALL_SIZE2BIN(s)	(small_size2bin[(s-1) >> LG_TINY_MIN])
+
+extern arena_bin_info_t	arena_bin_info[NBINS];
+
+/* Number of large size classes. */
+#define			nlclasses (chunk_npages - map_bias)
+
+void	arena_purge_all(arena_t *arena);
+void	arena_prof_accum(arena_t *arena, uint64_t accumbytes);
+void	arena_tcache_fill_small(arena_t *arena, tcache_bin_t *tbin,
+    size_t binind, uint64_t prof_accumbytes);
+void	arena_alloc_junk_small(void *ptr, arena_bin_info_t *bin_info,
+    bool zero);
+void	arena_dalloc_junk_small(void *ptr, arena_bin_info_t *bin_info);
+void	*arena_malloc_small(arena_t *arena, size_t size, bool zero);
+void	*arena_malloc_large(arena_t *arena, size_t size, bool zero);
+void	*arena_palloc(arena_t *arena, size_t size, size_t alignment, bool zero);
+void	arena_prof_promoted(const void *ptr, size_t size);
+void	arena_dalloc_bin_locked(arena_t *arena, arena_chunk_t *chunk, void *ptr,
+    arena_chunk_map_t *mapelm);
+void	arena_dalloc_bin(arena_t *arena, arena_chunk_t *chunk, void *ptr,
+    size_t pageind, arena_chunk_map_t *mapelm);
+void	arena_dalloc_small(arena_t *arena, arena_chunk_t *chunk, void *ptr,
+    size_t pageind);
+void	arena_dalloc_large_locked(arena_t *arena, arena_chunk_t *chunk,
+    void *ptr);
+void	arena_dalloc_large(arena_t *arena, arena_chunk_t *chunk, void *ptr);
+void	*arena_ralloc_no_move(void *ptr, size_t oldsize, size_t size,
+    size_t extra, bool zero);
+void	*arena_ralloc(arena_t *arena, void *ptr, size_t oldsize, size_t size,
+    size_t extra, size_t alignment, bool zero, bool try_tcache_alloc,
+    bool try_tcache_dalloc);
+dss_prec_t	arena_dss_prec_get(arena_t *arena);
+void	arena_dss_prec_set(arena_t *arena, dss_prec_t dss_prec);
+void	arena_stats_merge(arena_t *arena, const char **dss, size_t *nactive,
+    size_t *ndirty, arena_stats_t *astats, malloc_bin_stats_t *bstats,
+    malloc_large_stats_t *lstats);
+bool	arena_new(arena_t *arena, unsigned ind);
+void	arena_boot(void);
+void	arena_prefork(arena_t *arena);
+void	arena_postfork_parent(arena_t *arena);
+void	arena_postfork_child(arena_t *arena);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+arena_chunk_map_t	*arena_mapp_get(arena_chunk_t *chunk, size_t pageind);
+size_t	*arena_mapbitsp_get(arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_get(arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_unallocated_size_get(arena_chunk_t *chunk,
+    size_t pageind);
+size_t	arena_mapbits_large_size_get(arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_small_runind_get(arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_binind_get(arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_dirty_get(arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_unzeroed_get(arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_large_get(arena_chunk_t *chunk, size_t pageind);
+size_t	arena_mapbits_allocated_get(arena_chunk_t *chunk, size_t pageind);
+void	arena_mapbits_unallocated_set(arena_chunk_t *chunk, size_t pageind,
+    size_t size, size_t flags);
+void	arena_mapbits_unallocated_size_set(arena_chunk_t *chunk, size_t pageind,
+    size_t size);
+void	arena_mapbits_large_set(arena_chunk_t *chunk, size_t pageind,
+    size_t size, size_t flags);
+void	arena_mapbits_large_binind_set(arena_chunk_t *chunk, size_t pageind,
+    size_t binind);
+void	arena_mapbits_small_set(arena_chunk_t *chunk, size_t pageind,
+    size_t runind, size_t binind, size_t flags);
+void	arena_mapbits_unzeroed_set(arena_chunk_t *chunk, size_t pageind,
+    size_t unzeroed);
+size_t	arena_ptr_small_binind_get(const void *ptr, size_t mapbits);
+size_t	arena_bin_index(arena_t *arena, arena_bin_t *bin);
+unsigned	arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info,
+    const void *ptr);
+prof_ctx_t	*arena_prof_ctx_get(const void *ptr);
+void	arena_prof_ctx_set(const void *ptr, prof_ctx_t *ctx);
+void	*arena_malloc(arena_t *arena, size_t size, bool zero, bool try_tcache);
+size_t	arena_salloc(const void *ptr, bool demote);
+void	arena_dalloc(arena_t *arena, arena_chunk_t *chunk, void *ptr,
+    bool try_tcache);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_ARENA_C_))
+#  ifdef JEMALLOC_ARENA_INLINE_A
+JEMALLOC_INLINE arena_chunk_map_t *
+arena_mapp_get(arena_chunk_t *chunk, size_t pageind)
+{
+
+	assert(pageind >= map_bias);
+	assert(pageind < chunk_npages);
+
+	return (&chunk->map[pageind-map_bias]);
+}
+
+JEMALLOC_INLINE size_t *
+arena_mapbitsp_get(arena_chunk_t *chunk, size_t pageind)
+{
+
+	return (&arena_mapp_get(chunk, pageind)->bits);
+}
+
+JEMALLOC_INLINE size_t
+arena_mapbits_get(arena_chunk_t *chunk, size_t pageind)
+{
+
+	return (*arena_mapbitsp_get(chunk, pageind));
+}
+
+JEMALLOC_INLINE size_t
+arena_mapbits_unallocated_size_get(arena_chunk_t *chunk, size_t pageind)
+{
+	size_t mapbits;
+
+	mapbits = arena_mapbits_get(chunk, pageind);
+	assert((mapbits & (CHUNK_MAP_LARGE|CHUNK_MAP_ALLOCATED)) == 0);
+	return (mapbits & ~PAGE_MASK);
+}
+
+JEMALLOC_INLINE size_t
+arena_mapbits_large_size_get(arena_chunk_t *chunk, size_t pageind)
+{
+	size_t mapbits;
+
+	mapbits = arena_mapbits_get(chunk, pageind);
+	assert((mapbits & (CHUNK_MAP_LARGE|CHUNK_MAP_ALLOCATED)) ==
+	    (CHUNK_MAP_LARGE|CHUNK_MAP_ALLOCATED));
+	return (mapbits & ~PAGE_MASK);
+}
+
+JEMALLOC_INLINE size_t
+arena_mapbits_small_runind_get(arena_chunk_t *chunk, size_t pageind)
+{
+	size_t mapbits;
+
+	mapbits = arena_mapbits_get(chunk, pageind);
+	assert((mapbits & (CHUNK_MAP_LARGE|CHUNK_MAP_ALLOCATED)) ==
+	    CHUNK_MAP_ALLOCATED);
+	return (mapbits >> LG_PAGE);
+}
+
+JEMALLOC_INLINE size_t
+arena_mapbits_binind_get(arena_chunk_t *chunk, size_t pageind)
+{
+	size_t mapbits;
+	size_t binind;
+
+	mapbits = arena_mapbits_get(chunk, pageind);
+	binind = (mapbits & CHUNK_MAP_BININD_MASK) >> CHUNK_MAP_BININD_SHIFT;
+	assert(binind < NBINS || binind == BININD_INVALID);
+	return (binind);
+}
+
+JEMALLOC_INLINE size_t
+arena_mapbits_dirty_get(arena_chunk_t *chunk, size_t pageind)
+{
+	size_t mapbits;
+
+	mapbits = arena_mapbits_get(chunk, pageind);
+	return (mapbits & CHUNK_MAP_DIRTY);
+}
+
+JEMALLOC_INLINE size_t
+arena_mapbits_unzeroed_get(arena_chunk_t *chunk, size_t pageind)
+{
+	size_t mapbits;
+
+	mapbits = arena_mapbits_get(chunk, pageind);
+	return (mapbits & CHUNK_MAP_UNZEROED);
+}
+
+JEMALLOC_INLINE size_t
+arena_mapbits_large_get(arena_chunk_t *chunk, size_t pageind)
+{
+	size_t mapbits;
+
+	mapbits = arena_mapbits_get(chunk, pageind);
+	return (mapbits & CHUNK_MAP_LARGE);
+}
+
+JEMALLOC_INLINE size_t
+arena_mapbits_allocated_get(arena_chunk_t *chunk, size_t pageind)
+{
+	size_t mapbits;
+
+	mapbits = arena_mapbits_get(chunk, pageind);
+	return (mapbits & CHUNK_MAP_ALLOCATED);
+}
+
+JEMALLOC_INLINE void
+arena_mapbits_unallocated_set(arena_chunk_t *chunk, size_t pageind, size_t size,
+    size_t flags)
+{
+	size_t *mapbitsp;
+
+	mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	assert((size & PAGE_MASK) == 0);
+	assert((flags & ~CHUNK_MAP_FLAGS_MASK) == 0);
+	assert((flags & (CHUNK_MAP_DIRTY|CHUNK_MAP_UNZEROED)) == flags);
+	*mapbitsp = size | CHUNK_MAP_BININD_INVALID | flags;
+}
+
+JEMALLOC_INLINE void
+arena_mapbits_unallocated_size_set(arena_chunk_t *chunk, size_t pageind,
+    size_t size)
+{
+	size_t *mapbitsp;
+
+	mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	assert((size & PAGE_MASK) == 0);
+	assert((*mapbitsp & (CHUNK_MAP_LARGE|CHUNK_MAP_ALLOCATED)) == 0);
+	*mapbitsp = size | (*mapbitsp & PAGE_MASK);
+}
+
+JEMALLOC_INLINE void
+arena_mapbits_large_set(arena_chunk_t *chunk, size_t pageind, size_t size,
+    size_t flags)
+{
+	size_t *mapbitsp;
+	size_t unzeroed;
+
+	mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	assert((size & PAGE_MASK) == 0);
+	assert((flags & CHUNK_MAP_DIRTY) == flags);
+	unzeroed = *mapbitsp & CHUNK_MAP_UNZEROED; /* Preserve unzeroed. */
+	*mapbitsp = size | CHUNK_MAP_BININD_INVALID | flags | unzeroed |
+	    CHUNK_MAP_LARGE | CHUNK_MAP_ALLOCATED;
+}
+
+JEMALLOC_INLINE void
+arena_mapbits_large_binind_set(arena_chunk_t *chunk, size_t pageind,
+    size_t binind)
+{
+	size_t *mapbitsp;
+
+	assert(binind <= BININD_INVALID);
+	mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	assert(arena_mapbits_large_size_get(chunk, pageind) == PAGE);
+	*mapbitsp = (*mapbitsp & ~CHUNK_MAP_BININD_MASK) | (binind <<
+	    CHUNK_MAP_BININD_SHIFT);
+}
+
+JEMALLOC_INLINE void
+arena_mapbits_small_set(arena_chunk_t *chunk, size_t pageind, size_t runind,
+    size_t binind, size_t flags)
+{
+	size_t *mapbitsp;
+	size_t unzeroed;
+
+	assert(binind < BININD_INVALID);
+	mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	assert(pageind - runind >= map_bias);
+	assert((flags & CHUNK_MAP_DIRTY) == flags);
+	unzeroed = *mapbitsp & CHUNK_MAP_UNZEROED; /* Preserve unzeroed. */
+	*mapbitsp = (runind << LG_PAGE) | (binind << CHUNK_MAP_BININD_SHIFT) |
+	    flags | unzeroed | CHUNK_MAP_ALLOCATED;
+}
+
+JEMALLOC_INLINE void
+arena_mapbits_unzeroed_set(arena_chunk_t *chunk, size_t pageind,
+    size_t unzeroed)
+{
+	size_t *mapbitsp;
+
+	mapbitsp = arena_mapbitsp_get(chunk, pageind);
+	*mapbitsp = (*mapbitsp & ~CHUNK_MAP_UNZEROED) | unzeroed;
+}
+
+JEMALLOC_INLINE size_t
+arena_ptr_small_binind_get(const void *ptr, size_t mapbits)
+{
+	size_t binind;
+
+	binind = (mapbits & CHUNK_MAP_BININD_MASK) >> CHUNK_MAP_BININD_SHIFT;
+
+	if (config_debug) {
+		arena_chunk_t *chunk;
+		arena_t *arena;
+		size_t pageind;
+		size_t actual_mapbits;
+		arena_run_t *run;
+		arena_bin_t *bin;
+		size_t actual_binind;
+		arena_bin_info_t *bin_info;
+
+		assert(binind != BININD_INVALID);
+		assert(binind < NBINS);
+		chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+		arena = chunk->arena;
+		pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> LG_PAGE;
+		actual_mapbits = arena_mapbits_get(chunk, pageind);
+		assert(mapbits == actual_mapbits);
+		assert(arena_mapbits_large_get(chunk, pageind) == 0);
+		assert(arena_mapbits_allocated_get(chunk, pageind) != 0);
+		run = (arena_run_t *)((uintptr_t)chunk + (uintptr_t)((pageind -
+		    (actual_mapbits >> LG_PAGE)) << LG_PAGE));
+		bin = run->bin;
+		actual_binind = bin - arena->bins;
+		assert(binind == actual_binind);
+		bin_info = &arena_bin_info[actual_binind];
+		assert(((uintptr_t)ptr - ((uintptr_t)run +
+		    (uintptr_t)bin_info->reg0_offset)) % bin_info->reg_interval
+		    == 0);
+	}
+
+	return (binind);
+}
+#  endif /* JEMALLOC_ARENA_INLINE_A */
+
+#  ifdef JEMALLOC_ARENA_INLINE_B
+JEMALLOC_INLINE size_t
+arena_bin_index(arena_t *arena, arena_bin_t *bin)
+{
+	size_t binind = bin - arena->bins;
+	assert(binind < NBINS);
+	return (binind);
+}
+
+JEMALLOC_INLINE unsigned
+arena_run_regind(arena_run_t *run, arena_bin_info_t *bin_info, const void *ptr)
+{
+	unsigned shift, diff, regind;
+	size_t interval;
+
+	/*
+	 * Freeing a pointer lower than region zero can cause assertion
+	 * failure.
+	 */
+	assert((uintptr_t)ptr >= (uintptr_t)run +
+	    (uintptr_t)bin_info->reg0_offset);
+
+	/*
+	 * Avoid doing division with a variable divisor if possible.  Using
+	 * actual division here can reduce allocator throughput by over 20%!
+	 */
+	diff = (unsigned)((uintptr_t)ptr - (uintptr_t)run -
+	    bin_info->reg0_offset);
+
+	/* Rescale (factor powers of 2 out of the numerator and denominator). */
+	interval = bin_info->reg_interval;
+	shift = ffs(interval) - 1;
+	diff >>= shift;
+	interval >>= shift;
+
+	if (interval == 1) {
+		/* The divisor was a power of 2. */
+		regind = diff;
+	} else {
+		/*
+		 * To divide by a number D that is not a power of two we
+		 * multiply by (2^21 / D) and then right shift by 21 positions.
+		 *
+		 *   X / D
+		 *
+		 * becomes
+		 *
+		 *   (X * interval_invs[D - 3]) >> SIZE_INV_SHIFT
+		 *
+		 * We can omit the first three elements, because we never
+		 * divide by 0, and 1 and 2 are both powers of two, which are
+		 * handled above.
+		 */
+#define	SIZE_INV_SHIFT	((sizeof(unsigned) << 3) - LG_RUN_MAXREGS)
+#define	SIZE_INV(s)	(((1U << SIZE_INV_SHIFT) / (s)) + 1)
+		static const unsigned interval_invs[] = {
+		    SIZE_INV(3),
+		    SIZE_INV(4), SIZE_INV(5), SIZE_INV(6), SIZE_INV(7),
+		    SIZE_INV(8), SIZE_INV(9), SIZE_INV(10), SIZE_INV(11),
+		    SIZE_INV(12), SIZE_INV(13), SIZE_INV(14), SIZE_INV(15),
+		    SIZE_INV(16), SIZE_INV(17), SIZE_INV(18), SIZE_INV(19),
+		    SIZE_INV(20), SIZE_INV(21), SIZE_INV(22), SIZE_INV(23),
+		    SIZE_INV(24), SIZE_INV(25), SIZE_INV(26), SIZE_INV(27),
+		    SIZE_INV(28), SIZE_INV(29), SIZE_INV(30), SIZE_INV(31)
+		};
+
+		if (interval <= ((sizeof(interval_invs) / sizeof(unsigned)) +
+		    2)) {
+			regind = (diff * interval_invs[interval - 3]) >>
+			    SIZE_INV_SHIFT;
+		} else
+			regind = diff / interval;
+#undef SIZE_INV
+#undef SIZE_INV_SHIFT
+	}
+	assert(diff == regind * interval);
+	assert(regind < bin_info->nregs);
+
+	return (regind);
+}
+
+JEMALLOC_INLINE prof_ctx_t *
+arena_prof_ctx_get(const void *ptr)
+{
+	prof_ctx_t *ret;
+	arena_chunk_t *chunk;
+	size_t pageind, mapbits;
+
+	cassert(config_prof);
+	assert(ptr != NULL);
+	assert(CHUNK_ADDR2BASE(ptr) != ptr);
+
+	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> LG_PAGE;
+	mapbits = arena_mapbits_get(chunk, pageind);
+	assert((mapbits & CHUNK_MAP_ALLOCATED) != 0);
+	if ((mapbits & CHUNK_MAP_LARGE) == 0) {
+		if (prof_promote)
+			ret = (prof_ctx_t *)(uintptr_t)1U;
+		else {
+			arena_run_t *run = (arena_run_t *)((uintptr_t)chunk +
+			    (uintptr_t)((pageind - (mapbits >> LG_PAGE)) <<
+			    LG_PAGE));
+			size_t binind = arena_ptr_small_binind_get(ptr,
+			    mapbits);
+			arena_bin_info_t *bin_info = &arena_bin_info[binind];
+			unsigned regind;
+
+			regind = arena_run_regind(run, bin_info, ptr);
+			ret = *(prof_ctx_t **)((uintptr_t)run +
+			    bin_info->ctx0_offset + (regind *
+			    sizeof(prof_ctx_t *)));
+		}
+	} else
+		ret = arena_mapp_get(chunk, pageind)->prof_ctx;
+
+	return (ret);
+}
+
+JEMALLOC_INLINE void
+arena_prof_ctx_set(const void *ptr, prof_ctx_t *ctx)
+{
+	arena_chunk_t *chunk;
+	size_t pageind, mapbits;
+
+	cassert(config_prof);
+	assert(ptr != NULL);
+	assert(CHUNK_ADDR2BASE(ptr) != ptr);
+
+	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> LG_PAGE;
+	mapbits = arena_mapbits_get(chunk, pageind);
+	assert((mapbits & CHUNK_MAP_ALLOCATED) != 0);
+	if ((mapbits & CHUNK_MAP_LARGE) == 0) {
+		if (prof_promote == false) {
+			arena_run_t *run = (arena_run_t *)((uintptr_t)chunk +
+			    (uintptr_t)((pageind - (mapbits >> LG_PAGE)) <<
+			    LG_PAGE));
+			size_t binind;
+			arena_bin_info_t *bin_info;
+			unsigned regind;
+
+			binind = arena_ptr_small_binind_get(ptr, mapbits);
+			bin_info = &arena_bin_info[binind];
+			regind = arena_run_regind(run, bin_info, ptr);
+
+			*((prof_ctx_t **)((uintptr_t)run + bin_info->ctx0_offset
+			    + (regind * sizeof(prof_ctx_t *)))) = ctx;
+		} else
+			assert((uintptr_t)ctx == (uintptr_t)1U);
+	} else
+		arena_mapp_get(chunk, pageind)->prof_ctx = ctx;
+}
+
+JEMALLOC_INLINE void *
+arena_malloc(arena_t *arena, size_t size, bool zero, bool try_tcache)
+{
+	tcache_t *tcache;
+
+	assert(size != 0);
+	assert(size <= arena_maxclass);
+
+	if (size <= SMALL_MAXCLASS) {
+		if (try_tcache && (tcache = tcache_get(true)) != NULL)
+			return (tcache_alloc_small(tcache, size, zero));
+		else {
+			return (arena_malloc_small(choose_arena(arena), size,
+			    zero));
+		}
+	} else {
+		/*
+		 * Initialize tcache after checking size in order to avoid
+		 * infinite recursion during tcache initialization.
+		 */
+		if (try_tcache && size <= tcache_maxclass && (tcache =
+		    tcache_get(true)) != NULL)
+			return (tcache_alloc_large(tcache, size, zero));
+		else {
+			return (arena_malloc_large(choose_arena(arena), size,
+			    zero));
+		}
+	}
+}
+
+/* Return the size of the allocation pointed to by ptr. */
+JEMALLOC_INLINE size_t
+arena_salloc(const void *ptr, bool demote)
+{
+	size_t ret;
+	arena_chunk_t *chunk;
+	size_t pageind, binind;
+
+	assert(ptr != NULL);
+	assert(CHUNK_ADDR2BASE(ptr) != ptr);
+
+	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> LG_PAGE;
+	assert(arena_mapbits_allocated_get(chunk, pageind) != 0);
+	binind = arena_mapbits_binind_get(chunk, pageind);
+	if (binind == BININD_INVALID || (config_prof && demote == false &&
+	    prof_promote && arena_mapbits_large_get(chunk, pageind) != 0)) {
+		/*
+		 * Large allocation.  In the common case (demote == true), and
+		 * as this is an inline function, most callers will only end up
+		 * looking at binind to determine that ptr is a small
+		 * allocation.
+		 */
+		assert(((uintptr_t)ptr & PAGE_MASK) == 0);
+		ret = arena_mapbits_large_size_get(chunk, pageind);
+		assert(ret != 0);
+		assert(pageind + (ret>>LG_PAGE) <= chunk_npages);
+		assert(ret == PAGE || arena_mapbits_large_size_get(chunk,
+		    pageind+(ret>>LG_PAGE)-1) == 0);
+		assert(binind == arena_mapbits_binind_get(chunk,
+		    pageind+(ret>>LG_PAGE)-1));
+		assert(arena_mapbits_dirty_get(chunk, pageind) ==
+		    arena_mapbits_dirty_get(chunk, pageind+(ret>>LG_PAGE)-1));
+	} else {
+		/*
+		 * Small allocation (possibly promoted to a large object due to
+		 * prof_promote).
+		 */
+		assert(arena_mapbits_large_get(chunk, pageind) != 0 ||
+		    arena_ptr_small_binind_get(ptr, arena_mapbits_get(chunk,
+		    pageind)) == binind);
+		ret = arena_bin_info[binind].reg_size;
+	}
+
+	return (ret);
+}
+
+JEMALLOC_INLINE void
+arena_dalloc(arena_t *arena, arena_chunk_t *chunk, void *ptr, bool try_tcache)
+{
+	size_t pageind, mapbits;
+	tcache_t *tcache;
+
+	assert(arena != NULL);
+	assert(chunk->arena == arena);
+	assert(ptr != NULL);
+	assert(CHUNK_ADDR2BASE(ptr) != ptr);
+
+	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> LG_PAGE;
+	mapbits = arena_mapbits_get(chunk, pageind);
+	assert(arena_mapbits_allocated_get(chunk, pageind) != 0);
+	if ((mapbits & CHUNK_MAP_LARGE) == 0) {
+		/* Small allocation. */
+		if (try_tcache && (tcache = tcache_get(false)) != NULL) {
+			size_t binind;
+
+			binind = arena_ptr_small_binind_get(ptr, mapbits);
+			tcache_dalloc_small(tcache, ptr, binind);
+		} else
+			arena_dalloc_small(arena, chunk, ptr, pageind);
+	} else {
+		size_t size = arena_mapbits_large_size_get(chunk, pageind);
+
+		assert(((uintptr_t)ptr & PAGE_MASK) == 0);
+
+		if (try_tcache && size <= tcache_maxclass && (tcache =
+		    tcache_get(false)) != NULL) {
+			tcache_dalloc_large(tcache, ptr, size);
+		} else
+			arena_dalloc_large(arena, chunk, ptr);
+	}
+}
+#  endif /* JEMALLOC_ARENA_INLINE_B */
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
--- a/include/jemalloc/internal/atomic.h
+++ b/include/jemalloc/internal/atomic.h
@@ -0,0 +1,304 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+#define	atomic_read_uint64(p)	atomic_add_uint64(p, 0)
+#define	atomic_read_uint32(p)	atomic_add_uint32(p, 0)
+#define	atomic_read_z(p)	atomic_add_z(p, 0)
+#define	atomic_read_u(p)	atomic_add_u(p, 0)
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+uint64_t	atomic_add_uint64(uint64_t *p, uint64_t x);
+uint64_t	atomic_sub_uint64(uint64_t *p, uint64_t x);
+uint32_t	atomic_add_uint32(uint32_t *p, uint32_t x);
+uint32_t	atomic_sub_uint32(uint32_t *p, uint32_t x);
+size_t	atomic_add_z(size_t *p, size_t x);
+size_t	atomic_sub_z(size_t *p, size_t x);
+unsigned	atomic_add_u(unsigned *p, unsigned x);
+unsigned	atomic_sub_u(unsigned *p, unsigned x);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_ATOMIC_C_))
+/******************************************************************************/
+/* 64-bit operations. */
+#if (LG_SIZEOF_PTR == 3 || LG_SIZEOF_INT == 3)
+#  ifdef __GCC_HAVE_SYNC_COMPARE_AND_SWAP_8
+JEMALLOC_INLINE uint64_t
+atomic_add_uint64(uint64_t *p, uint64_t x)
+{
+
+	return (__sync_add_and_fetch(p, x));
+}
+
+JEMALLOC_INLINE uint64_t
+atomic_sub_uint64(uint64_t *p, uint64_t x)
+{
+
+	return (__sync_sub_and_fetch(p, x));
+}
+#elif (defined(_MSC_VER))
+JEMALLOC_INLINE uint64_t
+atomic_add_uint64(uint64_t *p, uint64_t x)
+{
+
+	return (InterlockedExchangeAdd64(p, x));
+}
+
+JEMALLOC_INLINE uint64_t
+atomic_sub_uint64(uint64_t *p, uint64_t x)
+{
+
+	return (InterlockedExchangeAdd64(p, -((int64_t)x)));
+}
+#elif (defined(JEMALLOC_OSATOMIC))
+JEMALLOC_INLINE uint64_t
+atomic_add_uint64(uint64_t *p, uint64_t x)
+{
+
+	return (OSAtomicAdd64((int64_t)x, (int64_t *)p));
+}
+
+JEMALLOC_INLINE uint64_t
+atomic_sub_uint64(uint64_t *p, uint64_t x)
+{
+
+	return (OSAtomicAdd64(-((int64_t)x), (int64_t *)p));
+}
+#  elif (defined(__amd64__) || defined(__x86_64__))
+JEMALLOC_INLINE uint64_t
+atomic_add_uint64(uint64_t *p, uint64_t x)
+{
+
+	asm volatile (
+	    "lock; xaddq %0, %1;"
+	    : "+r" (x), "=m" (*p) /* Outputs. */
+	    : "m" (*p) /* Inputs. */
+	    );
+
+	return (x);
+}
+
+JEMALLOC_INLINE uint64_t
+atomic_sub_uint64(uint64_t *p, uint64_t x)
+{
+
+	x = (uint64_t)(-(int64_t)x);
+	asm volatile (
+	    "lock; xaddq %0, %1;"
+	    : "+r" (x), "=m" (*p) /* Outputs. */
+	    : "m" (*p) /* Inputs. */
+	    );
+
+	return (x);
+}
+#  elif (defined(JEMALLOC_ATOMIC9))
+JEMALLOC_INLINE uint64_t
+atomic_add_uint64(uint64_t *p, uint64_t x)
+{
+
+	/*
+	 * atomic_fetchadd_64() doesn't exist, but we only ever use this
+	 * function on LP64 systems, so atomic_fetchadd_long() will do.
+	 */
+	assert(sizeof(uint64_t) == sizeof(unsigned long));
+
+	return (atomic_fetchadd_long(p, (unsigned long)x) + x);
+}
+
+JEMALLOC_INLINE uint64_t
+atomic_sub_uint64(uint64_t *p, uint64_t x)
+{
+
+	assert(sizeof(uint64_t) == sizeof(unsigned long));
+
+	return (atomic_fetchadd_long(p, (unsigned long)(-(long)x)) - x);
+}
+#  elif (defined(JE_FORCE_SYNC_COMPARE_AND_SWAP_8))
+JEMALLOC_INLINE uint64_t
+atomic_add_uint64(uint64_t *p, uint64_t x)
+{
+
+	return (__sync_add_and_fetch(p, x));
+}
+
+JEMALLOC_INLINE uint64_t
+atomic_sub_uint64(uint64_t *p, uint64_t x)
+{
+
+	return (__sync_sub_and_fetch(p, x));
+}
+#  else
+#    error "Missing implementation for 64-bit atomic operations"
+#  endif
+#endif
+
+/******************************************************************************/
+/* 32-bit operations. */
+#ifdef __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4
+JEMALLOC_INLINE uint32_t
+atomic_add_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (__sync_add_and_fetch(p, x));
+}
+
+JEMALLOC_INLINE uint32_t
+atomic_sub_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (__sync_sub_and_fetch(p, x));
+}
+#elif (defined(_MSC_VER))
+JEMALLOC_INLINE uint32_t
+atomic_add_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (InterlockedExchangeAdd(p, x));
+}
+
+JEMALLOC_INLINE uint32_t
+atomic_sub_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (InterlockedExchangeAdd(p, -((int32_t)x)));
+}
+#elif (defined(JEMALLOC_OSATOMIC))
+JEMALLOC_INLINE uint32_t
+atomic_add_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (OSAtomicAdd32((int32_t)x, (int32_t *)p));
+}
+
+JEMALLOC_INLINE uint32_t
+atomic_sub_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (OSAtomicAdd32(-((int32_t)x), (int32_t *)p));
+}
+#elif (defined(__i386__) || defined(__amd64__) || defined(__x86_64__))
+JEMALLOC_INLINE uint32_t
+atomic_add_uint32(uint32_t *p, uint32_t x)
+{
+
+	asm volatile (
+	    "lock; xaddl %0, %1;"
+	    : "+r" (x), "=m" (*p) /* Outputs. */
+	    : "m" (*p) /* Inputs. */
+	    );
+
+	return (x);
+}
+
+JEMALLOC_INLINE uint32_t
+atomic_sub_uint32(uint32_t *p, uint32_t x)
+{
+
+	x = (uint32_t)(-(int32_t)x);
+	asm volatile (
+	    "lock; xaddl %0, %1;"
+	    : "+r" (x), "=m" (*p) /* Outputs. */
+	    : "m" (*p) /* Inputs. */
+	    );
+
+	return (x);
+}
+#elif (defined(JEMALLOC_ATOMIC9))
+JEMALLOC_INLINE uint32_t
+atomic_add_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (atomic_fetchadd_32(p, x) + x);
+}
+
+JEMALLOC_INLINE uint32_t
+atomic_sub_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (atomic_fetchadd_32(p, (uint32_t)(-(int32_t)x)) - x);
+}
+#elif (defined(JE_FORCE_SYNC_COMPARE_AND_SWAP_4))
+JEMALLOC_INLINE uint32_t
+atomic_add_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (__sync_add_and_fetch(p, x));
+}
+
+JEMALLOC_INLINE uint32_t
+atomic_sub_uint32(uint32_t *p, uint32_t x)
+{
+
+	return (__sync_sub_and_fetch(p, x));
+}
+#else
+#  error "Missing implementation for 32-bit atomic operations"
+#endif
+
+/******************************************************************************/
+/* size_t operations. */
+JEMALLOC_INLINE size_t
+atomic_add_z(size_t *p, size_t x)
+{
+
+#if (LG_SIZEOF_PTR == 3)
+	return ((size_t)atomic_add_uint64((uint64_t *)p, (uint64_t)x));
+#elif (LG_SIZEOF_PTR == 2)
+	return ((size_t)atomic_add_uint32((uint32_t *)p, (uint32_t)x));
+#endif
+}
+
+JEMALLOC_INLINE size_t
+atomic_sub_z(size_t *p, size_t x)
+{
+
+#if (LG_SIZEOF_PTR == 3)
+	return ((size_t)atomic_add_uint64((uint64_t *)p,
+	    (uint64_t)-((int64_t)x)));
+#elif (LG_SIZEOF_PTR == 2)
+	return ((size_t)atomic_add_uint32((uint32_t *)p,
+	    (uint32_t)-((int32_t)x)));
+#endif
+}
+
+/******************************************************************************/
+/* unsigned operations. */
+JEMALLOC_INLINE unsigned
+atomic_add_u(unsigned *p, unsigned x)
+{
+
+#if (LG_SIZEOF_INT == 3)
+	return ((unsigned)atomic_add_uint64((uint64_t *)p, (uint64_t)x));
+#elif (LG_SIZEOF_INT == 2)
+	return ((unsigned)atomic_add_uint32((uint32_t *)p, (uint32_t)x));
+#endif
+}
+
+JEMALLOC_INLINE unsigned
+atomic_sub_u(unsigned *p, unsigned x)
+{
+
+#if (LG_SIZEOF_INT == 3)
+	return ((unsigned)atomic_add_uint64((uint64_t *)p,
+	    (uint64_t)-((int64_t)x)));
+#elif (LG_SIZEOF_INT == 2)
+	return ((unsigned)atomic_add_uint32((uint32_t *)p,
+	    (uint32_t)-((int32_t)x)));
+#endif
+}
+/******************************************************************************/
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
--- a/jemalloc/include/jemalloc/internal/base.h
+++ b/jemalloc/include/jemalloc/internal/base.h
@@ -9,12 +9,14 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS

-extern malloc_mutex_t	base_mtx;
-
 void	*base_alloc(size_t size);
+void	*base_calloc(size_t number, size_t size);
 extent_node_t *base_node_alloc(void);
 void	base_node_dealloc(extent_node_t *node);
 bool	base_boot(void);
+void	base_prefork(void);
+void	base_postfork_parent(void);
+void	base_postfork_child(void);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
--- a/include/jemalloc/internal/bitmap.h
+++ b/include/jemalloc/internal/bitmap.h
@@ -0,0 +1,184 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+/* Maximum bitmap bit count is 2^LG_BITMAP_MAXBITS. */
+#define	LG_BITMAP_MAXBITS	LG_RUN_MAXREGS
+
+typedef struct bitmap_level_s bitmap_level_t;
+typedef struct bitmap_info_s bitmap_info_t;
+typedef unsigned long bitmap_t;
+#define	LG_SIZEOF_BITMAP	LG_SIZEOF_LONG
+
+/* Number of bits per group. */
+#define	LG_BITMAP_GROUP_NBITS		(LG_SIZEOF_BITMAP + 3)
+#define	BITMAP_GROUP_NBITS		(ZU(1) << LG_BITMAP_GROUP_NBITS)
+#define	BITMAP_GROUP_NBITS_MASK		(BITMAP_GROUP_NBITS-1)
+
+/* Maximum number of levels possible. */
+#define	BITMAP_MAX_LEVELS						\
+    (LG_BITMAP_MAXBITS / LG_SIZEOF_BITMAP)				\
+    + !!(LG_BITMAP_MAXBITS % LG_SIZEOF_BITMAP)
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+struct bitmap_level_s {
+	/* Offset of this level's groups within the array of groups. */
+	size_t group_offset;
+};
+
+struct bitmap_info_s {
+	/* Logical number of bits in bitmap (stored at bottom level). */
+	size_t nbits;
+
+	/* Number of levels necessary for nbits. */
+	unsigned nlevels;
+
+	/*
+	 * Only the first (nlevels+1) elements are used, and levels are ordered
+	 * bottom to top (e.g. the bottom level is stored in levels[0]).
+	 */
+	bitmap_level_t levels[BITMAP_MAX_LEVELS+1];
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+void	bitmap_info_init(bitmap_info_t *binfo, size_t nbits);
+size_t	bitmap_info_ngroups(const bitmap_info_t *binfo);
+size_t	bitmap_size(size_t nbits);
+void	bitmap_init(bitmap_t *bitmap, const bitmap_info_t *binfo);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+bool	bitmap_full(bitmap_t *bitmap, const bitmap_info_t *binfo);
+bool	bitmap_get(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit);
+void	bitmap_set(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit);
+size_t	bitmap_sfu(bitmap_t *bitmap, const bitmap_info_t *binfo);
+void	bitmap_unset(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_BITMAP_C_))
+JEMALLOC_INLINE bool
+bitmap_full(bitmap_t *bitmap, const bitmap_info_t *binfo)
+{
+	unsigned rgoff = binfo->levels[binfo->nlevels].group_offset - 1;
+	bitmap_t rg = bitmap[rgoff];
+	/* The bitmap is full iff the root group is 0. */
+	return (rg == 0);
+}
+
+JEMALLOC_INLINE bool
+bitmap_get(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
+{
+	size_t goff;
+	bitmap_t g;
+
+	assert(bit < binfo->nbits);
+	goff = bit >> LG_BITMAP_GROUP_NBITS;
+	g = bitmap[goff];
+	return (!(g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK))));
+}
+
+JEMALLOC_INLINE void
+bitmap_set(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
+{
+	size_t goff;
+	bitmap_t *gp;
+	bitmap_t g;
+
+	assert(bit < binfo->nbits);
+	assert(bitmap_get(bitmap, binfo, bit) == false);
+	goff = bit >> LG_BITMAP_GROUP_NBITS;
+	gp = &bitmap[goff];
+	g = *gp;
+	assert(g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK)));
+	g ^= 1LU << (bit & BITMAP_GROUP_NBITS_MASK);
+	*gp = g;
+	assert(bitmap_get(bitmap, binfo, bit));
+	/* Propagate group state transitions up the tree. */
+	if (g == 0) {
+		unsigned i;
+		for (i = 1; i < binfo->nlevels; i++) {
+			bit = goff;
+			goff = bit >> LG_BITMAP_GROUP_NBITS;
+			gp = &bitmap[binfo->levels[i].group_offset + goff];
+			g = *gp;
+			assert(g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK)));
+			g ^= 1LU << (bit & BITMAP_GROUP_NBITS_MASK);
+			*gp = g;
+			if (g != 0)
+				break;
+		}
+	}
+}
+
+/* sfu: set first unset. */
+JEMALLOC_INLINE size_t
+bitmap_sfu(bitmap_t *bitmap, const bitmap_info_t *binfo)
+{
+	size_t bit;
+	bitmap_t g;
+	unsigned i;
+
+	assert(bitmap_full(bitmap, binfo) == false);
+
+	i = binfo->nlevels - 1;
+	g = bitmap[binfo->levels[i].group_offset];
+	bit = ffsl(g) - 1;
+	while (i > 0) {
+		i--;
+		g = bitmap[binfo->levels[i].group_offset + bit];
+		bit = (bit << LG_BITMAP_GROUP_NBITS) + (ffsl(g) - 1);
+	}
+
+	bitmap_set(bitmap, binfo, bit);
+	return (bit);
+}
+
+JEMALLOC_INLINE void
+bitmap_unset(bitmap_t *bitmap, const bitmap_info_t *binfo, size_t bit)
+{
+	size_t goff;
+	bitmap_t *gp;
+	bitmap_t g;
+	bool propagate;
+
+	assert(bit < binfo->nbits);
+	assert(bitmap_get(bitmap, binfo, bit));
+	goff = bit >> LG_BITMAP_GROUP_NBITS;
+	gp = &bitmap[goff];
+	g = *gp;
+	propagate = (g == 0);
+	assert((g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK))) == 0);
+	g ^= 1LU << (bit & BITMAP_GROUP_NBITS_MASK);
+	*gp = g;
+	assert(bitmap_get(bitmap, binfo, bit) == false);
+	/* Propagate group state transitions up the tree. */
+	if (propagate) {
+		unsigned i;
+		for (i = 1; i < binfo->nlevels; i++) {
+			bit = goff;
+			goff = bit >> LG_BITMAP_GROUP_NBITS;
+			gp = &bitmap[binfo->levels[i].group_offset + goff];
+			g = *gp;
+			propagate = (g == 0);
+			assert((g & (1LU << (bit & BITMAP_GROUP_NBITS_MASK)))
+			    == 0);
+			g ^= 1LU << (bit & BITMAP_GROUP_NBITS_MASK);
+			*gp = g;
+			if (propagate == false)
+				break;
+		}
+	}
+}
+
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
--- a/jemalloc/include/jemalloc/internal/chunk.h
+++ b/jemalloc/include/jemalloc/internal/chunk.h
@@ -28,20 +28,14 @@
 #ifdef JEMALLOC_H_EXTERNS

 extern size_t		opt_lg_chunk;
-#ifdef JEMALLOC_SWAP
-extern bool		opt_overcommit;
-#endif
+extern const char	*opt_dss;

-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
 /* Protects stats_chunks; currently not used for any other purpose. */
 extern malloc_mutex_t	chunks_mtx;
 /* Chunk statistics. */
 extern chunk_stats_t	stats_chunks;
-#endif

-#ifdef JEMALLOC_IVSALLOC
 extern rtree_t		*chunks_rtree;
-#endif

 extern size_t		chunksize;
 extern size_t		chunksize_mask; /* (chunksize - 1). */
@@ -49,9 +43,14 @@ extern size_t		chunk_npages;
 extern size_t		map_bias; /* Number of arena chunk header pages. */
 extern size_t		arena_maxclass; /* Max size class for arenas. */

-void	*chunk_alloc(size_t size, bool base, bool *zero);
-void	chunk_dealloc(void *chunk, size_t size);
+void	*chunk_alloc(size_t size, size_t alignment, bool base, bool *zero,
+    dss_prec_t dss_prec);
+void	chunk_unmap(void *chunk, size_t size);
+void	chunk_dealloc(void *chunk, size_t size, bool unmap);
 bool	chunk_boot(void);
+void	chunk_prefork(void);
+void	chunk_postfork_parent(void);
+void	chunk_postfork_child(void);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
@@ -60,6 +59,5 @@ bool	chunk_boot(void);
 #endif /* JEMALLOC_H_INLINES */
 /******************************************************************************/

-#include "jemalloc/internal/chunk_swap.h"
 #include "jemalloc/internal/chunk_dss.h"
 #include "jemalloc/internal/chunk_mmap.h"
--- a/jemalloc/include/jemalloc/internal/chunk_dss.h
+++ b/jemalloc/include/jemalloc/internal/chunk_dss.h
@@ -1,25 +1,34 @@
-#ifdef JEMALLOC_DSS
 /******************************************************************************/
 #ifdef JEMALLOC_H_TYPES

+typedef enum {
+	dss_prec_disabled  = 0,
+	dss_prec_primary   = 1,
+	dss_prec_secondary = 2,
+
+	dss_prec_limit     = 3
+} dss_prec_t ;
+#define	DSS_PREC_DEFAULT	dss_prec_secondary
+#define	DSS_DEFAULT		"secondary"
+
 #endif /* JEMALLOC_H_TYPES */
 /******************************************************************************/
 #ifdef JEMALLOC_H_STRUCTS

+extern const char *dss_prec_names[];
+
 #endif /* JEMALLOC_H_STRUCTS */
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS

-/*
- * Protects sbrk() calls.  This avoids malloc races among threads, though it
- * does not protect against races with threads that call sbrk() directly.
- */
-extern malloc_mutex_t	dss_mtx;
-
-void	*chunk_alloc_dss(size_t size, bool *zero);
+dss_prec_t	chunk_dss_prec_get(void);
+bool	chunk_dss_prec_set(dss_prec_t dss_prec);
+void	*chunk_alloc_dss(size_t size, size_t alignment, bool *zero);
 bool	chunk_in_dss(void *chunk);
-bool	chunk_dealloc_dss(void *chunk, size_t size);
 bool	chunk_dss_boot(void);
+void	chunk_dss_prefork(void);
+void	chunk_dss_postfork_parent(void);
+void	chunk_dss_postfork_child(void);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
@@ -27,4 +36,3 @@ bool	chunk_dss_boot(void);

 #endif /* JEMALLOC_H_INLINES */
 /******************************************************************************/
-#endif /* JEMALLOC_DSS */
--- a/jemalloc/include/jemalloc/internal/chunk_mmap.h
+++ b/jemalloc/include/jemalloc/internal/chunk_mmap.h
@@ -9,11 +9,10 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS

-void	*chunk_alloc_mmap(size_t size);
-void	*chunk_alloc_mmap_noreserve(size_t size);
-void	chunk_dealloc_mmap(void *chunk, size_t size);
+bool	pages_purge(void *addr, size_t length);

-bool	chunk_mmap_boot(void);
+void	*chunk_alloc_mmap(size_t size, size_t alignment, bool *zero);
+bool	chunk_dealloc_mmap(void *chunk, size_t size);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
--- a/jemalloc/include/jemalloc/internal/ckh.h
+++ b/jemalloc/include/jemalloc/internal/ckh.h
@@ -30,11 +30,6 @@ struct ckhc_s {
 };

 struct ckh_s {
-#ifdef JEMALLOC_DEBUG
-#define	CKH_MAGIC	0x3af2489d
-	uint32_t	magic;
-#endif
-
 #ifdef CKH_COUNT
 	/* Counters used to get an idea of performance. */
 	uint64_t	ngrows;
@@ -47,7 +42,7 @@ struct ckh_s {
 	/* Used for pseudo-random number generation. */
 #define	CKH_A		1103515241
 #define	CKH_C		12347
-	uint32_t	prn_state;
+	uint32_t	prng_state;

 	/* Total number of items. */
 	size_t		count;
--- a/jemalloc/include/jemalloc/internal/ctl.h
+++ b/jemalloc/include/jemalloc/internal/ctl.h
@@ -2,6 +2,8 @@
 #ifdef JEMALLOC_H_TYPES

 typedef struct ctl_node_s ctl_node_t;
+typedef struct ctl_named_node_s ctl_named_node_t;
+typedef struct ctl_indexed_node_s ctl_indexed_node_t;
 typedef struct ctl_arena_stats_s ctl_arena_stats_t;
 typedef struct ctl_stats_s ctl_stats_t;

@@ -11,27 +13,29 @@ typedef struct ctl_stats_s ctl_stats_t;

 struct ctl_node_s {
 	bool			named;
-	union {
-		struct {
-			const char	*name;
-			/* If (nchildren == 0), this is a terminal node. */
-			unsigned	nchildren;
-			const	ctl_node_t *children;
-		} named;
-		struct {
-			const ctl_node_t *(*index)(const size_t *, size_t,
-			    size_t);
-		} indexed;
-	} u;
-	int	(*ctl)(const size_t *, size_t, void *, size_t *, void *,
-	    size_t);
+};
+
+struct ctl_named_node_s {
+	struct ctl_node_s	node;
+	const char		*name;
+	/* If (nchildren == 0), this is a terminal node. */
+	unsigned		nchildren;
+	const			ctl_node_t *children;
+	int			(*ctl)(const size_t *, size_t, void *, size_t *,
+	    void *, size_t);
+};
+
+struct ctl_indexed_node_s {
+	struct ctl_node_s	node;
+	const ctl_named_node_t	*(*index)(const size_t *, size_t, size_t);
 };

 struct ctl_arena_stats_s {
 	bool			initialized;
+	unsigned		nthreads;
+	const char		*dss;
 	size_t			pactive;
 	size_t			pdirty;
-#ifdef JEMALLOC_STATS
 	arena_stats_t		astats;

 	/* Aggregate stats for small size classes, based on bin stats. */
@@ -40,13 +44,11 @@ struct ctl_arena_stats_s {
 	uint64_t		ndalloc_small;
 	uint64_t		nrequests_small;

-	malloc_bin_stats_t	*bstats;	/* nbins elements. */
+	malloc_bin_stats_t	bstats[NBINS];
 	malloc_large_stats_t	*lstats;	/* nlclasses elements. */
-#endif
 };

 struct ctl_stats_s {
-#ifdef JEMALLOC_STATS
 	size_t			allocated;
 	size_t			active;
 	size_t			mapped;
@@ -60,11 +62,8 @@ struct ctl_stats_s {
 		uint64_t	nmalloc;	/* huge_nmalloc */
 		uint64_t	ndalloc;	/* huge_ndalloc */
 	} huge;
-#endif
+	unsigned		narenas;
 	ctl_arena_stats_t	*arenas;	/* (narenas + 1) elements. */
-#ifdef JEMALLOC_SWAP
-	size_t			swap_avail;
-#endif
 };

 #endif /* JEMALLOC_H_STRUCTS */
@@ -78,29 +77,30 @@ int	ctl_nametomib(const char *name, size_t *mibp, size_t *miblenp);
 int	ctl_bymib(const size_t *mib, size_t miblen, void *oldp, size_t *oldlenp,
    void *newp, size_t newlen);
 bool	ctl_boot(void);
+void	ctl_prefork(void);
+void	ctl_postfork_parent(void);
+void	ctl_postfork_child(void);

 #define	xmallctl(name, oldp, oldlenp, newp, newlen) do {		\
-	if (JEMALLOC_P(mallctl)(name, oldp, oldlenp, newp, newlen)	\
+	if (je_mallctl(name, oldp, oldlenp, newp, newlen)		\
 	    != 0) {							\
-		malloc_write("<jemalloc>: Failure in xmallctl(\"");	\
-		malloc_write(name);					\
-		malloc_write("\", ...)\n");				\
+		malloc_printf(						\
+		    "<jemalloc>: Failure in xmallctl(\"%s\", ...)\n",	\
+		    name);						\
 		abort();						\
 	}								\
 } while (0)

 #define	xmallctlnametomib(name, mibp, miblenp) do {			\
-	if (JEMALLOC_P(mallctlnametomib)(name, mibp, miblenp) != 0) {	\
-		malloc_write(						\
-		    "<jemalloc>: Failure in xmallctlnametomib(\"");	\
-		malloc_write(name);					\
-		malloc_write("\", ...)\n");				\
+	if (je_mallctlnametomib(name, mibp, miblenp) != 0) {		\
+		malloc_printf("<jemalloc>: Failure in "			\
+		    "xmallctlnametomib(\"%s\", ...)\n", name);		\
 		abort();						\
 	}								\
 } while (0)

 #define	xmallctlbymib(mib, miblen, oldp, oldlenp, newp, newlen) do {	\
-	if (JEMALLOC_P(mallctlbymib)(mib, miblen, oldp, oldlenp, newp,	\
+	if (je_mallctlbymib(mib, miblen, oldp, oldlenp, newp,		\
 	    newlen) != 0) {						\
 		malloc_write(						\
 		    "<jemalloc>: Failure in xmallctlbymib()\n");	\
--- a/jemalloc/include/jemalloc/internal/extent.h
+++ b/jemalloc/include/jemalloc/internal/extent.h
@@ -9,24 +9,23 @@ typedef struct extent_node_s extent_node_t;

 /* Tree of extents. */
 struct extent_node_s {
-#if (defined(JEMALLOC_SWAP) || defined(JEMALLOC_DSS))
 	/* Linkage for the size/address-ordered tree. */
 	rb_node(extent_node_t)	link_szad;
-#endif

 	/* Linkage for the address-ordered tree. */
 	rb_node(extent_node_t)	link_ad;

-#ifdef JEMALLOC_PROF
 	/* Profile counters, used for huge objects. */
 	prof_ctx_t		*prof_ctx;
-#endif

 	/* Pointer to the extent that this tree node is responsible for. */
 	void			*addr;

 	/* Total region size. */
 	size_t			size;
+
+	/* True if zero-filled; used by chunk recycling code. */
+	bool			zeroed;
 };
 typedef rb_tree(extent_node_t) extent_tree_t;

@@ -34,9 +33,7 @@ typedef rb_tree(extent_node_t) extent_tree_t;
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS

-#if (defined(JEMALLOC_SWAP) || defined(JEMALLOC_DSS))
 rb_proto(, extent_tree_szad_, extent_tree_t, extent_node_t)
-#endif

 rb_proto(, extent_tree_ad_, extent_tree_t, extent_node_t)

--- a/jemalloc/include/jemalloc/internal/hash.h
+++ b/jemalloc/include/jemalloc/internal/hash.h
@@ -17,7 +17,7 @@
 uint64_t	hash(const void *key, size_t len, uint64_t seed);
 #endif

-#if (defined(JEMALLOC_ENABLE_INLINE) || defined(HASH_C_))
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_HASH_C_))
 /*
 * The following hash function is based on MurmurHash64A(), placed into the
 * public domain by Austin Appleby.  See http://murmurhash.googlepages.com/ for
@@ -26,7 +26,7 @@ uint64_t	hash(const void *key, size_t len, uint64_t seed);
 JEMALLOC_INLINE uint64_t
 hash(const void *key, size_t len, uint64_t seed)
 {
-	const uint64_t m = 0xc6a4a7935bd1e995;
+	const uint64_t m = UINT64_C(0xc6a4a7935bd1e995);
 	const int r = 47;
 	uint64_t h = seed ^ (len * m);
 	const uint64_t *data = (const uint64_t *)key;
@@ -48,14 +48,14 @@ hash(const void *key, size_t len, uint64_t seed)

 	data2 = (const unsigned char *)data;
 	switch(len & 7) {
-		case 7: h ^= ((uint64_t)(data2[6])) << 48;
-		case 6: h ^= ((uint64_t)(data2[5])) << 40;
-		case 5: h ^= ((uint64_t)(data2[4])) << 32;
-		case 4: h ^= ((uint64_t)(data2[3])) << 24;
-		case 3: h ^= ((uint64_t)(data2[2])) << 16;
-		case 2: h ^= ((uint64_t)(data2[1])) << 8;
-		case 1: h ^= ((uint64_t)(data2[0]));
-			h *= m;
+	case 7: h ^= ((uint64_t)(data2[6])) << 48;
+	case 6: h ^= ((uint64_t)(data2[5])) << 40;
+	case 5: h ^= ((uint64_t)(data2[4])) << 32;
+	case 4: h ^= ((uint64_t)(data2[3])) << 24;
+	case 3: h ^= ((uint64_t)(data2[2])) << 16;
+	case 2: h ^= ((uint64_t)(data2[1])) << 8;
+	case 1: h ^= ((uint64_t)(data2[0]));
+		h *= m;
 	}

 	h ^= h >> r;
--- a/jemalloc/include/jemalloc/internal/huge.h
+++ b/jemalloc/include/jemalloc/internal/huge.h
@@ -9,12 +9,10 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS

-#ifdef JEMALLOC_STATS
 /* Huge allocation statistics. */
 extern uint64_t		huge_nmalloc;
 extern uint64_t		huge_ndalloc;
 extern size_t		huge_allocated;
-#endif

 /* Protects chunk-related data structures. */
 extern malloc_mutex_t	huge_mtx;
@@ -24,14 +22,15 @@ void	*huge_palloc(size_t size, size_t alignment, bool zero);
 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);
+    size_t alignment, bool zero, bool try_tcache_dalloc);
 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);
 void	huge_prof_ctx_set(const void *ptr, prof_ctx_t *ctx);
-#endif
 bool	huge_boot(void);
+void	huge_prefork(void);
+void	huge_postfork_parent(void);
+void	huge_postfork_child(void);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
--- a/include/jemalloc/internal/jemalloc_internal.h.in
+++ b/include/jemalloc/internal/jemalloc_internal.h.in
--- a/jemalloc/include/jemalloc/internal/mb.h
+++ b/jemalloc/include/jemalloc/internal/mb.h
@@ -17,7 +17,7 @@
 void	mb_write(void);
 #endif

-#if (defined(JEMALLOC_ENABLE_INLINE) || defined(MB_C_))
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_MB_C_))
 #ifdef __i386__
 /*
 * According to the Intel Architecture Software Developer's Manual, current
@@ -54,7 +54,7 @@ mb_write(void)
 	    );
 #endif
 }
-#elif (defined(__amd64_) || defined(__x86_64__))
+#elif (defined(__amd64__) || defined(__x86_64__))
 JEMALLOC_INLINE void
 mb_write(void)
 {
@@ -87,6 +87,13 @@ mb_write(void)
 	    : "memory" /* Clobbers. */
 	    );
 }
+#elif defined(__tile__)
+JEMALLOC_INLINE void
+mb_write(void)
+{
+
+	__sync_synchronize();
+}
 #else
 /*
 * This is much slower than a simple memory barrier, but the semantics of mutex
--- a/include/jemalloc/internal/mutex.h
+++ b/include/jemalloc/internal/mutex.h
@@ -0,0 +1,99 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct malloc_mutex_s malloc_mutex_t;
+
+#ifdef _WIN32
+#  define MALLOC_MUTEX_INITIALIZER
+#elif (defined(JEMALLOC_OSSPIN))
+#  define MALLOC_MUTEX_INITIALIZER {0}
+#elif (defined(JEMALLOC_MUTEX_INIT_CB))
+#  define MALLOC_MUTEX_INITIALIZER {PTHREAD_MUTEX_INITIALIZER, NULL}
+#else
+#  if (defined(PTHREAD_MUTEX_ADAPTIVE_NP) &&				\
+       defined(PTHREAD_ADAPTIVE_MUTEX_INITIALIZER_NP))
+#    define MALLOC_MUTEX_TYPE PTHREAD_MUTEX_ADAPTIVE_NP
+#    define MALLOC_MUTEX_INITIALIZER {PTHREAD_ADAPTIVE_MUTEX_INITIALIZER_NP}
+#  else
+#    define MALLOC_MUTEX_TYPE PTHREAD_MUTEX_DEFAULT
+#    define MALLOC_MUTEX_INITIALIZER {PTHREAD_MUTEX_INITIALIZER}
+#  endif
+#endif
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+struct malloc_mutex_s {
+#ifdef _WIN32
+	CRITICAL_SECTION	lock;
+#elif (defined(JEMALLOC_OSSPIN))
+	OSSpinLock		lock;
+#elif (defined(JEMALLOC_MUTEX_INIT_CB))
+	pthread_mutex_t		lock;
+	malloc_mutex_t		*postponed_next;
+#else
+	pthread_mutex_t		lock;
+#endif
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+#ifdef JEMALLOC_LAZY_LOCK
+extern bool isthreaded;
+#else
+#  undef isthreaded /* Undo private_namespace.h definition. */
+#  define isthreaded true
+#endif
+
+bool	malloc_mutex_init(malloc_mutex_t *mutex);
+void	malloc_mutex_prefork(malloc_mutex_t *mutex);
+void	malloc_mutex_postfork_parent(malloc_mutex_t *mutex);
+void	malloc_mutex_postfork_child(malloc_mutex_t *mutex);
+bool	mutex_boot(void);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+void	malloc_mutex_lock(malloc_mutex_t *mutex);
+void	malloc_mutex_unlock(malloc_mutex_t *mutex);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_MUTEX_C_))
+JEMALLOC_INLINE void
+malloc_mutex_lock(malloc_mutex_t *mutex)
+{
+
+	if (isthreaded) {
+#ifdef _WIN32
+		EnterCriticalSection(&mutex->lock);
+#elif (defined(JEMALLOC_OSSPIN))
+		OSSpinLockLock(&mutex->lock);
+#else
+		pthread_mutex_lock(&mutex->lock);
+#endif
+	}
+}
+
+JEMALLOC_INLINE void
+malloc_mutex_unlock(malloc_mutex_t *mutex)
+{
+
+	if (isthreaded) {
+#ifdef _WIN32
+		LeaveCriticalSection(&mutex->lock);
+#elif (defined(JEMALLOC_OSSPIN))
+		OSSpinLockUnlock(&mutex->lock);
+#else
+		pthread_mutex_unlock(&mutex->lock);
+#endif
+	}
+}
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
--- a/include/jemalloc/internal/private_namespace.h
+++ b/include/jemalloc/internal/private_namespace.h
@@ -0,0 +1,367 @@
+#define	a0calloc JEMALLOC_N(a0calloc)
+#define	a0free JEMALLOC_N(a0free)
+#define	a0malloc JEMALLOC_N(a0malloc)
+#define	arena_alloc_junk_small JEMALLOC_N(arena_alloc_junk_small)
+#define	arena_bin_index JEMALLOC_N(arena_bin_index)
+#define	arena_bin_info JEMALLOC_N(arena_bin_info)
+#define	arena_boot JEMALLOC_N(arena_boot)
+#define	arena_dalloc JEMALLOC_N(arena_dalloc)
+#define	arena_dalloc_bin JEMALLOC_N(arena_dalloc_bin)
+#define	arena_dalloc_bin_locked JEMALLOC_N(arena_dalloc_bin_locked)
+#define	arena_dalloc_junk_small JEMALLOC_N(arena_dalloc_junk_small)
+#define	arena_dalloc_large JEMALLOC_N(arena_dalloc_large)
+#define	arena_dalloc_large_locked JEMALLOC_N(arena_dalloc_large_locked)
+#define	arena_dalloc_small JEMALLOC_N(arena_dalloc_small)
+#define	arena_dss_prec_get JEMALLOC_N(arena_dss_prec_get)
+#define	arena_dss_prec_set JEMALLOC_N(arena_dss_prec_set)
+#define	arena_malloc JEMALLOC_N(arena_malloc)
+#define	arena_malloc_large JEMALLOC_N(arena_malloc_large)
+#define	arena_malloc_small JEMALLOC_N(arena_malloc_small)
+#define	arena_mapbits_allocated_get JEMALLOC_N(arena_mapbits_allocated_get)
+#define	arena_mapbits_binind_get JEMALLOC_N(arena_mapbits_binind_get)
+#define	arena_mapbits_dirty_get JEMALLOC_N(arena_mapbits_dirty_get)
+#define	arena_mapbits_get JEMALLOC_N(arena_mapbits_get)
+#define	arena_mapbits_large_binind_set JEMALLOC_N(arena_mapbits_large_binind_set)
+#define	arena_mapbits_large_get JEMALLOC_N(arena_mapbits_large_get)
+#define	arena_mapbits_large_set JEMALLOC_N(arena_mapbits_large_set)
+#define	arena_mapbits_large_size_get JEMALLOC_N(arena_mapbits_large_size_get)
+#define	arena_mapbits_small_runind_get JEMALLOC_N(arena_mapbits_small_runind_get)
+#define	arena_mapbits_small_set JEMALLOC_N(arena_mapbits_small_set)
+#define	arena_mapbits_unallocated_set JEMALLOC_N(arena_mapbits_unallocated_set)
+#define	arena_mapbits_unallocated_size_get JEMALLOC_N(arena_mapbits_unallocated_size_get)
+#define	arena_mapbits_unallocated_size_set JEMALLOC_N(arena_mapbits_unallocated_size_set)
+#define	arena_mapbits_unzeroed_get JEMALLOC_N(arena_mapbits_unzeroed_get)
+#define	arena_mapbits_unzeroed_set JEMALLOC_N(arena_mapbits_unzeroed_set)
+#define	arena_mapbitsp_get JEMALLOC_N(arena_mapbitsp_get)
+#define	arena_mapp_get JEMALLOC_N(arena_mapp_get)
+#define	arena_maxclass JEMALLOC_N(arena_maxclass)
+#define	arena_new JEMALLOC_N(arena_new)
+#define	arena_palloc JEMALLOC_N(arena_palloc)
+#define	arena_postfork_child JEMALLOC_N(arena_postfork_child)
+#define	arena_postfork_parent JEMALLOC_N(arena_postfork_parent)
+#define	arena_prefork JEMALLOC_N(arena_prefork)
+#define	arena_prof_accum JEMALLOC_N(arena_prof_accum)
+#define	arena_prof_ctx_get JEMALLOC_N(arena_prof_ctx_get)
+#define	arena_prof_ctx_set JEMALLOC_N(arena_prof_ctx_set)
+#define	arena_prof_promoted JEMALLOC_N(arena_prof_promoted)
+#define	arena_ptr_small_binind_get JEMALLOC_N(arena_ptr_small_binind_get)
+#define	arena_purge_all JEMALLOC_N(arena_purge_all)
+#define	arena_ralloc JEMALLOC_N(arena_ralloc)
+#define	arena_ralloc_no_move JEMALLOC_N(arena_ralloc_no_move)
+#define	arena_run_regind JEMALLOC_N(arena_run_regind)
+#define	arena_salloc JEMALLOC_N(arena_salloc)
+#define	arena_stats_merge JEMALLOC_N(arena_stats_merge)
+#define	arena_tcache_fill_small JEMALLOC_N(arena_tcache_fill_small)
+#define	arenas JEMALLOC_N(arenas)
+#define	arenas_booted JEMALLOC_N(arenas_booted)
+#define	arenas_cleanup JEMALLOC_N(arenas_cleanup)
+#define	arenas_extend JEMALLOC_N(arenas_extend)
+#define	arenas_initialized JEMALLOC_N(arenas_initialized)
+#define	arenas_lock JEMALLOC_N(arenas_lock)
+#define	arenas_tls JEMALLOC_N(arenas_tls)
+#define	arenas_tsd JEMALLOC_N(arenas_tsd)
+#define	arenas_tsd_boot JEMALLOC_N(arenas_tsd_boot)
+#define	arenas_tsd_cleanup_wrapper JEMALLOC_N(arenas_tsd_cleanup_wrapper)
+#define	arenas_tsd_get JEMALLOC_N(arenas_tsd_get)
+#define	arenas_tsd_set JEMALLOC_N(arenas_tsd_set)
+#define	atomic_add_u JEMALLOC_N(atomic_add_u)
+#define	atomic_add_uint32 JEMALLOC_N(atomic_add_uint32)
+#define	atomic_add_uint64 JEMALLOC_N(atomic_add_uint64)
+#define	atomic_add_z JEMALLOC_N(atomic_add_z)
+#define	atomic_sub_u JEMALLOC_N(atomic_sub_u)
+#define	atomic_sub_uint32 JEMALLOC_N(atomic_sub_uint32)
+#define	atomic_sub_uint64 JEMALLOC_N(atomic_sub_uint64)
+#define	atomic_sub_z JEMALLOC_N(atomic_sub_z)
+#define	base_alloc JEMALLOC_N(base_alloc)
+#define	base_boot JEMALLOC_N(base_boot)
+#define	base_calloc JEMALLOC_N(base_calloc)
+#define	base_node_alloc JEMALLOC_N(base_node_alloc)
+#define	base_node_dealloc JEMALLOC_N(base_node_dealloc)
+#define	base_postfork_child JEMALLOC_N(base_postfork_child)
+#define	base_postfork_parent JEMALLOC_N(base_postfork_parent)
+#define	base_prefork JEMALLOC_N(base_prefork)
+#define	bitmap_full JEMALLOC_N(bitmap_full)
+#define	bitmap_get JEMALLOC_N(bitmap_get)
+#define	bitmap_info_init JEMALLOC_N(bitmap_info_init)
+#define	bitmap_info_ngroups JEMALLOC_N(bitmap_info_ngroups)
+#define	bitmap_init JEMALLOC_N(bitmap_init)
+#define	bitmap_set JEMALLOC_N(bitmap_set)
+#define	bitmap_sfu JEMALLOC_N(bitmap_sfu)
+#define	bitmap_size JEMALLOC_N(bitmap_size)
+#define	bitmap_unset JEMALLOC_N(bitmap_unset)
+#define	bt_init JEMALLOC_N(bt_init)
+#define	buferror JEMALLOC_N(buferror)
+#define	choose_arena JEMALLOC_N(choose_arena)
+#define	choose_arena_hard JEMALLOC_N(choose_arena_hard)
+#define	chunk_alloc JEMALLOC_N(chunk_alloc)
+#define	chunk_alloc_dss JEMALLOC_N(chunk_alloc_dss)
+#define	chunk_alloc_mmap JEMALLOC_N(chunk_alloc_mmap)
+#define	chunk_boot JEMALLOC_N(chunk_boot)
+#define	chunk_dealloc JEMALLOC_N(chunk_dealloc)
+#define	chunk_dealloc_mmap JEMALLOC_N(chunk_dealloc_mmap)
+#define	chunk_dss_boot JEMALLOC_N(chunk_dss_boot)
+#define	chunk_dss_postfork_child JEMALLOC_N(chunk_dss_postfork_child)
+#define	chunk_dss_postfork_parent JEMALLOC_N(chunk_dss_postfork_parent)
+#define	chunk_dss_prec_get JEMALLOC_N(chunk_dss_prec_get)
+#define	chunk_dss_prec_set JEMALLOC_N(chunk_dss_prec_set)
+#define	chunk_dss_prefork JEMALLOC_N(chunk_dss_prefork)
+#define	chunk_in_dss JEMALLOC_N(chunk_in_dss)
+#define	chunk_npages JEMALLOC_N(chunk_npages)
+#define	chunk_postfork_child JEMALLOC_N(chunk_postfork_child)
+#define	chunk_postfork_parent JEMALLOC_N(chunk_postfork_parent)
+#define	chunk_prefork JEMALLOC_N(chunk_prefork)
+#define	chunk_unmap JEMALLOC_N(chunk_unmap)
+#define	chunks_mtx JEMALLOC_N(chunks_mtx)
+#define	chunks_rtree JEMALLOC_N(chunks_rtree)
+#define	chunksize JEMALLOC_N(chunksize)
+#define	chunksize_mask JEMALLOC_N(chunksize_mask)
+#define	ckh_bucket_search JEMALLOC_N(ckh_bucket_search)
+#define	ckh_count JEMALLOC_N(ckh_count)
+#define	ckh_delete JEMALLOC_N(ckh_delete)
+#define	ckh_evict_reloc_insert JEMALLOC_N(ckh_evict_reloc_insert)
+#define	ckh_insert JEMALLOC_N(ckh_insert)
+#define	ckh_isearch JEMALLOC_N(ckh_isearch)
+#define	ckh_iter JEMALLOC_N(ckh_iter)
+#define	ckh_new JEMALLOC_N(ckh_new)
+#define	ckh_pointer_hash JEMALLOC_N(ckh_pointer_hash)
+#define	ckh_pointer_keycomp JEMALLOC_N(ckh_pointer_keycomp)
+#define	ckh_rebuild JEMALLOC_N(ckh_rebuild)
+#define	ckh_remove JEMALLOC_N(ckh_remove)
+#define	ckh_search JEMALLOC_N(ckh_search)
+#define	ckh_string_hash JEMALLOC_N(ckh_string_hash)
+#define	ckh_string_keycomp JEMALLOC_N(ckh_string_keycomp)
+#define	ckh_try_bucket_insert JEMALLOC_N(ckh_try_bucket_insert)
+#define	ckh_try_insert JEMALLOC_N(ckh_try_insert)
+#define	ctl_boot JEMALLOC_N(ctl_boot)
+#define	ctl_bymib JEMALLOC_N(ctl_bymib)
+#define	ctl_byname JEMALLOC_N(ctl_byname)
+#define	ctl_nametomib JEMALLOC_N(ctl_nametomib)
+#define	ctl_postfork_child JEMALLOC_N(ctl_postfork_child)
+#define	ctl_postfork_parent JEMALLOC_N(ctl_postfork_parent)
+#define	ctl_prefork JEMALLOC_N(ctl_prefork)
+#define	dss_prec_names JEMALLOC_N(dss_prec_names)
+#define	extent_tree_ad_first JEMALLOC_N(extent_tree_ad_first)
+#define	extent_tree_ad_insert JEMALLOC_N(extent_tree_ad_insert)
+#define	extent_tree_ad_iter JEMALLOC_N(extent_tree_ad_iter)
+#define	extent_tree_ad_iter_recurse JEMALLOC_N(extent_tree_ad_iter_recurse)
+#define	extent_tree_ad_iter_start JEMALLOC_N(extent_tree_ad_iter_start)
+#define	extent_tree_ad_last JEMALLOC_N(extent_tree_ad_last)
+#define	extent_tree_ad_new JEMALLOC_N(extent_tree_ad_new)
+#define	extent_tree_ad_next JEMALLOC_N(extent_tree_ad_next)
+#define	extent_tree_ad_nsearch JEMALLOC_N(extent_tree_ad_nsearch)
+#define	extent_tree_ad_prev JEMALLOC_N(extent_tree_ad_prev)
+#define	extent_tree_ad_psearch JEMALLOC_N(extent_tree_ad_psearch)
+#define	extent_tree_ad_remove JEMALLOC_N(extent_tree_ad_remove)
+#define	extent_tree_ad_reverse_iter JEMALLOC_N(extent_tree_ad_reverse_iter)
+#define	extent_tree_ad_reverse_iter_recurse JEMALLOC_N(extent_tree_ad_reverse_iter_recurse)
+#define	extent_tree_ad_reverse_iter_start JEMALLOC_N(extent_tree_ad_reverse_iter_start)
+#define	extent_tree_ad_search JEMALLOC_N(extent_tree_ad_search)
+#define	extent_tree_szad_first JEMALLOC_N(extent_tree_szad_first)
+#define	extent_tree_szad_insert JEMALLOC_N(extent_tree_szad_insert)
+#define	extent_tree_szad_iter JEMALLOC_N(extent_tree_szad_iter)
+#define	extent_tree_szad_iter_recurse JEMALLOC_N(extent_tree_szad_iter_recurse)
+#define	extent_tree_szad_iter_start JEMALLOC_N(extent_tree_szad_iter_start)
+#define	extent_tree_szad_last JEMALLOC_N(extent_tree_szad_last)
+#define	extent_tree_szad_new JEMALLOC_N(extent_tree_szad_new)
+#define	extent_tree_szad_next JEMALLOC_N(extent_tree_szad_next)
+#define	extent_tree_szad_nsearch JEMALLOC_N(extent_tree_szad_nsearch)
+#define	extent_tree_szad_prev JEMALLOC_N(extent_tree_szad_prev)
+#define	extent_tree_szad_psearch JEMALLOC_N(extent_tree_szad_psearch)
+#define	extent_tree_szad_remove JEMALLOC_N(extent_tree_szad_remove)
+#define	extent_tree_szad_reverse_iter JEMALLOC_N(extent_tree_szad_reverse_iter)
+#define	extent_tree_szad_reverse_iter_recurse JEMALLOC_N(extent_tree_szad_reverse_iter_recurse)
+#define	extent_tree_szad_reverse_iter_start JEMALLOC_N(extent_tree_szad_reverse_iter_start)
+#define	extent_tree_szad_search JEMALLOC_N(extent_tree_szad_search)
+#define	get_errno JEMALLOC_N(get_errno)
+#define	hash JEMALLOC_N(hash)
+#define	huge_allocated JEMALLOC_N(huge_allocated)
+#define	huge_boot JEMALLOC_N(huge_boot)
+#define	huge_dalloc JEMALLOC_N(huge_dalloc)
+#define	huge_malloc JEMALLOC_N(huge_malloc)
+#define	huge_mtx JEMALLOC_N(huge_mtx)
+#define	huge_ndalloc JEMALLOC_N(huge_ndalloc)
+#define	huge_nmalloc JEMALLOC_N(huge_nmalloc)
+#define	huge_palloc JEMALLOC_N(huge_palloc)
+#define	huge_postfork_child JEMALLOC_N(huge_postfork_child)
+#define	huge_postfork_parent JEMALLOC_N(huge_postfork_parent)
+#define	huge_prefork JEMALLOC_N(huge_prefork)
+#define	huge_prof_ctx_get JEMALLOC_N(huge_prof_ctx_get)
+#define	huge_prof_ctx_set JEMALLOC_N(huge_prof_ctx_set)
+#define	huge_ralloc JEMALLOC_N(huge_ralloc)
+#define	huge_ralloc_no_move JEMALLOC_N(huge_ralloc_no_move)
+#define	huge_salloc JEMALLOC_N(huge_salloc)
+#define	iallocm JEMALLOC_N(iallocm)
+#define	icalloc JEMALLOC_N(icalloc)
+#define	icallocx JEMALLOC_N(icallocx)
+#define	idalloc JEMALLOC_N(idalloc)
+#define	idallocx JEMALLOC_N(idallocx)
+#define	imalloc JEMALLOC_N(imalloc)
+#define	imallocx JEMALLOC_N(imallocx)
+#define	ipalloc JEMALLOC_N(ipalloc)
+#define	ipallocx JEMALLOC_N(ipallocx)
+#define	iqalloc JEMALLOC_N(iqalloc)
+#define	iqallocx JEMALLOC_N(iqallocx)
+#define	iralloc JEMALLOC_N(iralloc)
+#define	irallocx JEMALLOC_N(irallocx)
+#define	isalloc JEMALLOC_N(isalloc)
+#define	isthreaded JEMALLOC_N(isthreaded)
+#define	ivsalloc JEMALLOC_N(ivsalloc)
+#define	jemalloc_postfork_child JEMALLOC_N(jemalloc_postfork_child)
+#define	jemalloc_postfork_parent JEMALLOC_N(jemalloc_postfork_parent)
+#define	jemalloc_prefork JEMALLOC_N(jemalloc_prefork)
+#define	malloc_cprintf JEMALLOC_N(malloc_cprintf)
+#define	malloc_mutex_init JEMALLOC_N(malloc_mutex_init)
+#define	malloc_mutex_lock JEMALLOC_N(malloc_mutex_lock)
+#define	malloc_mutex_postfork_child JEMALLOC_N(malloc_mutex_postfork_child)
+#define	malloc_mutex_postfork_parent JEMALLOC_N(malloc_mutex_postfork_parent)
+#define	malloc_mutex_prefork JEMALLOC_N(malloc_mutex_prefork)
+#define	malloc_mutex_unlock JEMALLOC_N(malloc_mutex_unlock)
+#define	malloc_printf JEMALLOC_N(malloc_printf)
+#define	malloc_snprintf JEMALLOC_N(malloc_snprintf)
+#define	malloc_strtoumax JEMALLOC_N(malloc_strtoumax)
+#define	malloc_tsd_boot JEMALLOC_N(malloc_tsd_boot)
+#define	malloc_tsd_cleanup_register JEMALLOC_N(malloc_tsd_cleanup_register)
+#define	malloc_tsd_dalloc JEMALLOC_N(malloc_tsd_dalloc)
+#define	malloc_tsd_malloc JEMALLOC_N(malloc_tsd_malloc)
+#define	malloc_tsd_no_cleanup JEMALLOC_N(malloc_tsd_no_cleanup)
+#define	malloc_vcprintf JEMALLOC_N(malloc_vcprintf)
+#define	malloc_vsnprintf JEMALLOC_N(malloc_vsnprintf)
+#define	malloc_write JEMALLOC_N(malloc_write)
+#define	map_bias JEMALLOC_N(map_bias)
+#define	mb_write JEMALLOC_N(mb_write)
+#define	mutex_boot JEMALLOC_N(mutex_boot)
+#define	narenas_auto JEMALLOC_N(narenas_auto)
+#define	narenas_total JEMALLOC_N(narenas_total)
+#define	narenas_total_get JEMALLOC_N(narenas_total_get)
+#define	ncpus JEMALLOC_N(ncpus)
+#define	nhbins JEMALLOC_N(nhbins)
+#define	opt_abort JEMALLOC_N(opt_abort)
+#define	opt_junk JEMALLOC_N(opt_junk)
+#define	opt_lg_chunk JEMALLOC_N(opt_lg_chunk)
+#define	opt_lg_dirty_mult JEMALLOC_N(opt_lg_dirty_mult)
+#define	opt_lg_prof_interval JEMALLOC_N(opt_lg_prof_interval)
+#define	opt_lg_prof_sample JEMALLOC_N(opt_lg_prof_sample)
+#define	opt_lg_tcache_max JEMALLOC_N(opt_lg_tcache_max)
+#define	opt_narenas JEMALLOC_N(opt_narenas)
+#define	opt_prof JEMALLOC_N(opt_prof)
+#define	opt_prof_accum JEMALLOC_N(opt_prof_accum)
+#define	opt_prof_active JEMALLOC_N(opt_prof_active)
+#define	opt_prof_final JEMALLOC_N(opt_prof_final)
+#define	opt_prof_gdump JEMALLOC_N(opt_prof_gdump)
+#define	opt_prof_leak JEMALLOC_N(opt_prof_leak)
+#define	opt_prof_prefix JEMALLOC_N(opt_prof_prefix)
+#define	opt_quarantine JEMALLOC_N(opt_quarantine)
+#define	opt_redzone JEMALLOC_N(opt_redzone)
+#define	opt_stats_print JEMALLOC_N(opt_stats_print)
+#define	opt_tcache JEMALLOC_N(opt_tcache)
+#define	opt_utrace JEMALLOC_N(opt_utrace)
+#define	opt_valgrind JEMALLOC_N(opt_valgrind)
+#define	opt_xmalloc JEMALLOC_N(opt_xmalloc)
+#define	opt_zero JEMALLOC_N(opt_zero)
+#define	p2rz JEMALLOC_N(p2rz)
+#define	pages_purge JEMALLOC_N(pages_purge)
+#define	pow2_ceil JEMALLOC_N(pow2_ceil)
+#define	prof_backtrace JEMALLOC_N(prof_backtrace)
+#define	prof_boot0 JEMALLOC_N(prof_boot0)
+#define	prof_boot1 JEMALLOC_N(prof_boot1)
+#define	prof_boot2 JEMALLOC_N(prof_boot2)
+#define	prof_ctx_get JEMALLOC_N(prof_ctx_get)
+#define	prof_ctx_set JEMALLOC_N(prof_ctx_set)
+#define	prof_free JEMALLOC_N(prof_free)
+#define	prof_gdump JEMALLOC_N(prof_gdump)
+#define	prof_idump JEMALLOC_N(prof_idump)
+#define	prof_interval JEMALLOC_N(prof_interval)
+#define	prof_lookup JEMALLOC_N(prof_lookup)
+#define	prof_malloc JEMALLOC_N(prof_malloc)
+#define	prof_mdump JEMALLOC_N(prof_mdump)
+#define	prof_postfork_child JEMALLOC_N(prof_postfork_child)
+#define	prof_postfork_parent JEMALLOC_N(prof_postfork_parent)
+#define	prof_prefork JEMALLOC_N(prof_prefork)
+#define	prof_promote JEMALLOC_N(prof_promote)
+#define	prof_realloc JEMALLOC_N(prof_realloc)
+#define	prof_sample_accum_update JEMALLOC_N(prof_sample_accum_update)
+#define	prof_sample_threshold_update JEMALLOC_N(prof_sample_threshold_update)
+#define	prof_tdata_booted JEMALLOC_N(prof_tdata_booted)
+#define	prof_tdata_cleanup JEMALLOC_N(prof_tdata_cleanup)
+#define	prof_tdata_get JEMALLOC_N(prof_tdata_get)
+#define	prof_tdata_init JEMALLOC_N(prof_tdata_init)
+#define	prof_tdata_initialized JEMALLOC_N(prof_tdata_initialized)
+#define	prof_tdata_tls JEMALLOC_N(prof_tdata_tls)
+#define	prof_tdata_tsd JEMALLOC_N(prof_tdata_tsd)
+#define	prof_tdata_tsd_boot JEMALLOC_N(prof_tdata_tsd_boot)
+#define	prof_tdata_tsd_cleanup_wrapper JEMALLOC_N(prof_tdata_tsd_cleanup_wrapper)
+#define	prof_tdata_tsd_get JEMALLOC_N(prof_tdata_tsd_get)
+#define	prof_tdata_tsd_set JEMALLOC_N(prof_tdata_tsd_set)
+#define	quarantine JEMALLOC_N(quarantine)
+#define	quarantine_boot JEMALLOC_N(quarantine_boot)
+#define	quarantine_tsd_boot JEMALLOC_N(quarantine_tsd_boot)
+#define	quarantine_tsd_cleanup_wrapper JEMALLOC_N(quarantine_tsd_cleanup_wrapper)
+#define	quarantine_tsd_get JEMALLOC_N(quarantine_tsd_get)
+#define	quarantine_tsd_set JEMALLOC_N(quarantine_tsd_set)
+#define	register_zone JEMALLOC_N(register_zone)
+#define	rtree_get JEMALLOC_N(rtree_get)
+#define	rtree_get_locked JEMALLOC_N(rtree_get_locked)
+#define	rtree_new JEMALLOC_N(rtree_new)
+#define	rtree_postfork_child JEMALLOC_N(rtree_postfork_child)
+#define	rtree_postfork_parent JEMALLOC_N(rtree_postfork_parent)
+#define	rtree_prefork JEMALLOC_N(rtree_prefork)
+#define	rtree_set JEMALLOC_N(rtree_set)
+#define	s2u JEMALLOC_N(s2u)
+#define	sa2u JEMALLOC_N(sa2u)
+#define	set_errno JEMALLOC_N(set_errno)
+#define	stats_cactive JEMALLOC_N(stats_cactive)
+#define	stats_cactive_add JEMALLOC_N(stats_cactive_add)
+#define	stats_cactive_get JEMALLOC_N(stats_cactive_get)
+#define	stats_cactive_sub JEMALLOC_N(stats_cactive_sub)
+#define	stats_chunks JEMALLOC_N(stats_chunks)
+#define	stats_print JEMALLOC_N(stats_print)
+#define	tcache_alloc_easy JEMALLOC_N(tcache_alloc_easy)
+#define	tcache_alloc_large JEMALLOC_N(tcache_alloc_large)
+#define	tcache_alloc_small JEMALLOC_N(tcache_alloc_small)
+#define	tcache_alloc_small_hard JEMALLOC_N(tcache_alloc_small_hard)
+#define	tcache_arena_associate JEMALLOC_N(tcache_arena_associate)
+#define	tcache_arena_dissociate JEMALLOC_N(tcache_arena_dissociate)
+#define	tcache_bin_flush_large JEMALLOC_N(tcache_bin_flush_large)
+#define	tcache_bin_flush_small JEMALLOC_N(tcache_bin_flush_small)
+#define	tcache_bin_info JEMALLOC_N(tcache_bin_info)
+#define	tcache_boot0 JEMALLOC_N(tcache_boot0)
+#define	tcache_boot1 JEMALLOC_N(tcache_boot1)
+#define	tcache_booted JEMALLOC_N(tcache_booted)
+#define	tcache_create JEMALLOC_N(tcache_create)
+#define	tcache_dalloc_large JEMALLOC_N(tcache_dalloc_large)
+#define	tcache_dalloc_small JEMALLOC_N(tcache_dalloc_small)
+#define	tcache_destroy JEMALLOC_N(tcache_destroy)
+#define	tcache_enabled_booted JEMALLOC_N(tcache_enabled_booted)
+#define	tcache_enabled_get JEMALLOC_N(tcache_enabled_get)
+#define	tcache_enabled_initialized JEMALLOC_N(tcache_enabled_initialized)
+#define	tcache_enabled_set JEMALLOC_N(tcache_enabled_set)
+#define	tcache_enabled_tls JEMALLOC_N(tcache_enabled_tls)
+#define	tcache_enabled_tsd JEMALLOC_N(tcache_enabled_tsd)
+#define	tcache_enabled_tsd_boot JEMALLOC_N(tcache_enabled_tsd_boot)
+#define	tcache_enabled_tsd_cleanup_wrapper JEMALLOC_N(tcache_enabled_tsd_cleanup_wrapper)
+#define	tcache_enabled_tsd_get JEMALLOC_N(tcache_enabled_tsd_get)
+#define	tcache_enabled_tsd_set JEMALLOC_N(tcache_enabled_tsd_set)
+#define	tcache_event JEMALLOC_N(tcache_event)
+#define	tcache_event_hard JEMALLOC_N(tcache_event_hard)
+#define	tcache_flush JEMALLOC_N(tcache_flush)
+#define	tcache_get JEMALLOC_N(tcache_get)
+#define	tcache_initialized JEMALLOC_N(tcache_initialized)
+#define	tcache_maxclass JEMALLOC_N(tcache_maxclass)
+#define	tcache_salloc JEMALLOC_N(tcache_salloc)
+#define	tcache_stats_merge JEMALLOC_N(tcache_stats_merge)
+#define	tcache_thread_cleanup JEMALLOC_N(tcache_thread_cleanup)
+#define	tcache_tls JEMALLOC_N(tcache_tls)
+#define	tcache_tsd JEMALLOC_N(tcache_tsd)
+#define	tcache_tsd_boot JEMALLOC_N(tcache_tsd_boot)
+#define	tcache_tsd_cleanup_wrapper JEMALLOC_N(tcache_tsd_cleanup_wrapper)
+#define	tcache_tsd_get JEMALLOC_N(tcache_tsd_get)
+#define	tcache_tsd_set JEMALLOC_N(tcache_tsd_set)
+#define	thread_allocated_booted JEMALLOC_N(thread_allocated_booted)
+#define	thread_allocated_initialized JEMALLOC_N(thread_allocated_initialized)
+#define	thread_allocated_tls JEMALLOC_N(thread_allocated_tls)
+#define	thread_allocated_tsd JEMALLOC_N(thread_allocated_tsd)
+#define	thread_allocated_tsd_boot JEMALLOC_N(thread_allocated_tsd_boot)
+#define	thread_allocated_tsd_cleanup_wrapper JEMALLOC_N(thread_allocated_tsd_cleanup_wrapper)
+#define	thread_allocated_tsd_get JEMALLOC_N(thread_allocated_tsd_get)
+#define	thread_allocated_tsd_set JEMALLOC_N(thread_allocated_tsd_set)
+#define	u2rz JEMALLOC_N(u2rz)
--- a/jemalloc/include/jemalloc/internal/prn.h
+++ b/jemalloc/include/jemalloc/internal/prn.h
@@ -4,7 +4,7 @@
 /*
 * Simple linear congruential pseudo-random number generator:
 *
- *   prn(y) = (a*x + c) % m
+ *   prng(y) = (a*x + c) % m
 *
 * where the following constants ensure maximal period:
 *
@@ -25,7 +25,7 @@
 *   uint32_t state      : Seed value.
 *   const uint32_t a, c : See above discussion.
 */
-#define prn32(r, lg_range, state, a, c) do {				\
+#define prng32(r, lg_range, state, a, c) do {				\
 	assert(lg_range > 0);						\
 	assert(lg_range <= 32);						\
 									\
@@ -34,8 +34,8 @@
 	r >>= (32 - lg_range);						\
 } while (false)

-/* Same as prn32(), but 64 bits of pseudo-randomness, using uint64_t. */
-#define prn64(r, lg_range, state, a, c) do {				\
+/* Same as prng32(), but 64 bits of pseudo-randomness, using uint64_t. */
+#define prng64(r, lg_range, state, a, c) do {				\
 	assert(lg_range > 0);						\
 	assert(lg_range <= 64);						\
 									\
--- a/jemalloc/include/jemalloc/internal/prof.h
+++ b/jemalloc/include/jemalloc/internal/prof.h
@@ -1,4 +1,3 @@
-#ifdef JEMALLOC_PROF
 /******************************************************************************/
 #ifdef JEMALLOC_H_TYPES

@@ -10,28 +9,41 @@ typedef struct prof_tdata_s prof_tdata_t;

 /* Option defaults. */
 #define	PROF_PREFIX_DEFAULT		"jeprof"
-#define	LG_PROF_BT_MAX_DEFAULT		7
-#define	LG_PROF_SAMPLE_DEFAULT		0
+#define	LG_PROF_SAMPLE_DEFAULT		19
 #define	LG_PROF_INTERVAL_DEFAULT	-1
-#define	LG_PROF_TCMAX_DEFAULT		-1

 /*
- * Hard limit on stack backtrace depth.  Note that the version of
- * prof_backtrace() that is based on __builtin_return_address() necessarily has
- * a hard-coded number of backtrace frame handlers.
+ * Hard limit on stack backtrace depth.  The version of prof_backtrace() that
+ * is based on __builtin_return_address() necessarily has a hard-coded number
+ * of backtrace frame handlers, and should be kept in sync with this setting.
 */
-#if (defined(JEMALLOC_PROF_LIBGCC) || defined(JEMALLOC_PROF_LIBUNWIND))
-#  define LG_PROF_BT_MAX	((ZU(1) << (LG_SIZEOF_PTR+3)) - 1)
-#else
-#  define LG_PROF_BT_MAX	7 /* >= LG_PROF_BT_MAX_DEFAULT */
-#endif
-#define	PROF_BT_MAX		(1U << LG_PROF_BT_MAX)
+#define	PROF_BT_MAX			128
+
+/* Maximum number of backtraces to store in each per thread LRU cache. */
+#define	PROF_TCMAX			1024

 /* Initial hash table size. */
-#define	PROF_CKH_MINITEMS	64
+#define	PROF_CKH_MINITEMS		64

 /* Size of memory buffer to use when writing dump files. */
-#define	PROF_DUMP_BUF_SIZE	65536
+#define	PROF_DUMP_BUFSIZE		65536
+
+/* Size of stack-allocated buffer used by prof_printf(). */
+#define	PROF_PRINTF_BUFSIZE		128
+
+/*
+ * Number of mutexes shared among all ctx's.  No space is allocated for these
+ * unless profiling is enabled, so it's okay to over-provision.
+ */
+#define	PROF_NCTX_LOCKS			1024
+
+/*
+ * prof_tdata pointers close to NULL are used to encode state information that
+ * is used for cleaning up during thread shutdown.
+ */
+#define	PROF_TDATA_STATE_REINCARNATED	((prof_tdata_t *)(uintptr_t)1)
+#define	PROF_TDATA_STATE_PURGATORY	((prof_tdata_t *)(uintptr_t)2)
+#define	PROF_TDATA_STATE_MAX		PROF_TDATA_STATE_PURGATORY

 #endif /* JEMALLOC_H_TYPES */
 /******************************************************************************/
@@ -109,8 +121,18 @@ struct prof_ctx_s {
 	/* Associated backtrace. */
 	prof_bt_t		*bt;

-	/* Protects cnt_merged and cnts_ql. */
-	malloc_mutex_t		lock;
+	/* Protects nlimbo, cnt_merged, and cnts_ql. */
+	malloc_mutex_t		*lock;
+
+	/*
+	 * Number of threads that currently cause this ctx to be in a state of
+	 * limbo due to one of:
+	 *   - Initializing per thread counters associated with this ctx.
+	 *   - Preparing to destroy this ctx.
+	 * nlimbo must be 1 (single destroyer) in order to safely destroy the
+	 * ctx.
+	 */
+	unsigned		nlimbo;

 	/* Temporary storage for summation during dump. */
 	prof_cnt_t		cnt_summed;
@@ -145,9 +167,14 @@ struct prof_tdata_s {
 	void			**vec;

 	/* Sampling state. */
-	uint64_t		prn_state;
+	uint64_t		prng_state;
 	uint64_t		threshold;
 	uint64_t		accum;
+
+	/* State used to avoid dumping while operating on prof internals. */
+	bool			enq;
+	bool			enq_idump;
+	bool			enq_gdump;
 };

 #endif /* JEMALLOC_H_STRUCTS */
@@ -162,13 +189,12 @@ extern bool	opt_prof;
 * to notice state changes.
 */
 extern bool	opt_prof_active;
-extern size_t	opt_lg_prof_bt_max;   /* Maximum backtrace depth. */
 extern size_t	opt_lg_prof_sample;   /* Mean bytes between samples. */
 extern ssize_t	opt_lg_prof_interval; /* lg(prof_interval). */
 extern bool	opt_prof_gdump;       /* High-water memory dumping. */
+extern bool	opt_prof_final;       /* Final profile dumping. */
 extern bool	opt_prof_leak;        /* Dump leak summary at exit. */
 extern bool	opt_prof_accum;       /* Report cumulative bytes. */
-extern ssize_t	opt_lg_prof_tcmax;    /* lg(max per thread bactrace cache) */
 extern char	opt_prof_prefix[PATH_MAX + 1];

 /*
@@ -186,50 +212,82 @@ extern uint64_t	prof_interval;
 */
 extern bool	prof_promote;

-/* (1U << opt_lg_prof_bt_max). */
-extern unsigned	prof_bt_max;
-
-/* Thread-specific backtrace cache, used to reduce bt2ctx contention. */
-#ifndef NO_TLS
-extern __thread prof_tdata_t	*prof_tdata_tls
-    JEMALLOC_ATTR(tls_model("initial-exec"));
-#  define PROF_TCACHE_GET()	prof_tdata_tls
-#  define PROF_TCACHE_SET(v)	do {					\
-	prof_tdata_tls = (v);						\
-	pthread_setspecific(prof_tdata_tsd, (void *)(v));		\
-} while (0)
-#else
-#  define PROF_TCACHE_GET()						\
-	((prof_tdata_t *)pthread_getspecific(prof_tdata_tsd))
-#  define PROF_TCACHE_SET(v)	do {					\
-	pthread_setspecific(prof_tdata_tsd, (void *)(v));		\
-} while (0)
-#endif
-/*
- * Same contents as b2cnt_tls, but initialized such that the TSD destructor is
- * called when a thread exits, so that prof_tdata_tls contents can be merged,
- * unlinked, and deallocated.
- */
-extern pthread_key_t	prof_tdata_tsd;
-
 void	bt_init(prof_bt_t *bt, void **vec);
-void	prof_backtrace(prof_bt_t *bt, unsigned nignore, unsigned max);
+void	prof_backtrace(prof_bt_t *bt, unsigned nignore);
 prof_thr_cnt_t	*prof_lookup(prof_bt_t *bt);
 void	prof_idump(void);
 bool	prof_mdump(const char *filename);
 void	prof_gdump(void);
 prof_tdata_t	*prof_tdata_init(void);
+void	prof_tdata_cleanup(void *arg);
 void	prof_boot0(void);
 void	prof_boot1(void);
 bool	prof_boot2(void);
+void	prof_prefork(void);
+void	prof_postfork_parent(void);
+void	prof_postfork_child(void);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
 #ifdef JEMALLOC_H_INLINES

+#define	PROF_ALLOC_PREP(nignore, size, ret) do {			\
+	prof_tdata_t *prof_tdata;					\
+	prof_bt_t bt;							\
+									\
+	assert(size == s2u(size));					\
+									\
+	prof_tdata = prof_tdata_get();					\
+	if ((uintptr_t)prof_tdata <= (uintptr_t)PROF_TDATA_STATE_MAX) {	\
+		if (prof_tdata != NULL)					\
+			ret = (prof_thr_cnt_t *)(uintptr_t)1U;		\
+		else							\
+			ret = NULL;					\
+		break;							\
+	}								\
+									\
+	if (opt_prof_active == false) {					\
+		/* Sampling is currently inactive, so avoid sampling. */\
+		ret = (prof_thr_cnt_t *)(uintptr_t)1U;			\
+	} else if (opt_lg_prof_sample == 0) {				\
+		/* Don't bother with sampling logic, since sampling   */\
+		/* interval is 1.                                     */\
+		bt_init(&bt, prof_tdata->vec);				\
+		prof_backtrace(&bt, nignore);				\
+		ret = prof_lookup(&bt);					\
+	} else {							\
+		if (prof_tdata->threshold == 0) {			\
+			/* Initialize.  Seed the prng differently for */\
+			/* each thread.                               */\
+			prof_tdata->prng_state =			\
+			    (uint64_t)(uintptr_t)&size;			\
+			prof_sample_threshold_update(prof_tdata);	\
+		}							\
+									\
+		/* Determine whether to capture a backtrace based on  */\
+		/* whether size is enough for prof_accum to reach     */\
+		/* prof_tdata->threshold.  However, delay updating    */\
+		/* these variables until prof_{m,re}alloc(), because  */\
+		/* we don't know for sure that the allocation will    */\
+		/* succeed.                                           */\
+		/*                                                    */\
+		/* Use subtraction rather than addition to avoid      */\
+		/* potential integer overflow.                        */\
+		if (size >= prof_tdata->threshold -			\
+		    prof_tdata->accum) {				\
+			bt_init(&bt, prof_tdata->vec);			\
+			prof_backtrace(&bt, nignore);			\
+			ret = prof_lookup(&bt);				\
+		} else							\
+			ret = (prof_thr_cnt_t *)(uintptr_t)1U;		\
+	}								\
+} while (0)
+
 #ifndef JEMALLOC_ENABLE_INLINE
+malloc_tsd_protos(JEMALLOC_ATTR(unused), prof_tdata, prof_tdata_t *)
+
+prof_tdata_t	*prof_tdata_get(void);
 void	prof_sample_threshold_update(prof_tdata_t *prof_tdata);
-prof_thr_cnt_t	*prof_alloc_prep(size_t size);
 prof_ctx_t	*prof_ctx_get(const void *ptr);
 void	prof_ctx_set(const void *ptr, prof_ctx_t *ctx);
 bool	prof_sample_accum_update(size_t size);
@@ -240,102 +298,73 @@ void	prof_free(const void *ptr, size_t size);
 #endif

 #if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_PROF_C_))
+/* Thread-specific backtrace cache, used to reduce bt2ctx contention. */
+malloc_tsd_externs(prof_tdata, prof_tdata_t *)
+malloc_tsd_funcs(JEMALLOC_INLINE, prof_tdata, prof_tdata_t *, NULL,
+    prof_tdata_cleanup)
+
+JEMALLOC_INLINE prof_tdata_t *
+prof_tdata_get(void)
+{
+	prof_tdata_t *prof_tdata;
+
+	cassert(config_prof);
+
+	prof_tdata = *prof_tdata_tsd_get();
+	if ((uintptr_t)prof_tdata <= (uintptr_t)PROF_TDATA_STATE_MAX) {
+		if (prof_tdata == NULL)
+			prof_tdata = prof_tdata_init();
+	}
+
+	return (prof_tdata);
+}
+
 JEMALLOC_INLINE void
 prof_sample_threshold_update(prof_tdata_t *prof_tdata)
 {
 	uint64_t r;
 	double u;

+	cassert(config_prof);
+
 	/*
-	 * Compute prof_sample_threshold as a geometrically distributed random
+	 * Compute sample threshold as a geometrically distributed random
 	 * variable with mean (2^opt_lg_prof_sample).
+	 *
+	 *                         __        __
+	 *                         |  log(u)  |                     1
+	 * prof_tdata->threshold = | -------- |, where p = -------------------
+	 *                         | log(1-p) |             opt_lg_prof_sample
+	 *                                                 2
+	 *
+	 * For more information on the math, see:
+	 *
+	 *   Non-Uniform Random Variate Generation
+	 *   Luc Devroye
+	 *   Springer-Verlag, New York, 1986
+	 *   pp 500
+	 *   (http://cg.scs.carleton.ca/~luc/rnbookindex.html)
 	 */
-	prn64(r, 53, prof_tdata->prn_state,
-	    (uint64_t)6364136223846793005LLU, (uint64_t)1442695040888963407LLU);
+	prng64(r, 53, prof_tdata->prng_state,
+	    UINT64_C(6364136223846793005), UINT64_C(1442695040888963407));
 	u = (double)r * (1.0/9007199254740992.0L);
 	prof_tdata->threshold = (uint64_t)(log(u) /
 	    log(1.0 - (1.0 / (double)((uint64_t)1U << opt_lg_prof_sample))))
 	    + (uint64_t)1U;
 }

-JEMALLOC_INLINE prof_thr_cnt_t *
-prof_alloc_prep(size_t size)
-{
-#ifdef JEMALLOC_ENABLE_INLINE
-   /* This function does not have its own stack frame, because it is inlined. */
-#  define NIGNORE 1
-#else
-#  define NIGNORE 2
-#endif
-	prof_thr_cnt_t *ret;
-	prof_tdata_t *prof_tdata;
-	prof_bt_t bt;
-
-	assert(size == s2u(size));
-
-	prof_tdata = PROF_TCACHE_GET();
-	if (prof_tdata == NULL) {
-		prof_tdata = prof_tdata_init();
-		if (prof_tdata == NULL)
-			return (NULL);
-	}
-
-	if (opt_prof_active == false) {
-		/* Sampling is currently inactive, so avoid sampling. */
-		ret = (prof_thr_cnt_t *)(uintptr_t)1U;
-	} else if (opt_lg_prof_sample == 0) {
-		/*
-		 * Don't bother with sampling logic, since sampling interval is
-		 * 1.
-		 */
-		bt_init(&bt, prof_tdata->vec);
-		prof_backtrace(&bt, NIGNORE, prof_bt_max);
-		ret = prof_lookup(&bt);
-	} else {
-		if (prof_tdata->threshold == 0) {
-			/*
-			 * Initialize.  Seed the prng differently for each
-			 * thread.
-			 */
-			prof_tdata->prn_state = (uint64_t)(uintptr_t)&size;
-			prof_sample_threshold_update(prof_tdata);
-		}
-
-		/*
-		 * Determine whether to capture a backtrace based on whether
-		 * size is enough for prof_accum to reach
-		 * prof_tdata->threshold.  However, delay updating these
-		 * variables until prof_{m,re}alloc(), because we don't know
-		 * for sure that the allocation will succeed.
-		 *
-		 * Use subtraction rather than addition to avoid potential
-		 * integer overflow.
-		 */
-		if (size >= prof_tdata->threshold - prof_tdata->accum) {
-			bt_init(&bt, prof_tdata->vec);
-			prof_backtrace(&bt, NIGNORE, prof_bt_max);
-			ret = prof_lookup(&bt);
-		} else
-			ret = (prof_thr_cnt_t *)(uintptr_t)1U;
-	}
-
-	return (ret);
-#undef NIGNORE
-}
-
 JEMALLOC_INLINE prof_ctx_t *
 prof_ctx_get(const void *ptr)
 {
 	prof_ctx_t *ret;
 	arena_chunk_t *chunk;

+	cassert(config_prof);
 	assert(ptr != NULL);

 	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
 	if (chunk != ptr) {
 		/* Region. */
-		assert(chunk->arena->magic == ARENA_MAGIC);
-
 		ret = arena_prof_ctx_get(ptr);
 	} else
 		ret = huge_prof_ctx_get(ptr);
@@ -348,13 +377,12 @@ prof_ctx_set(const void *ptr, prof_ctx_t *ctx)
 {
 	arena_chunk_t *chunk;

+	cassert(config_prof);
 	assert(ptr != NULL);

 	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
 	if (chunk != ptr) {
 		/* Region. */
-		assert(chunk->arena->magic == ARENA_MAGIC);
-
 		arena_prof_ctx_set(ptr, ctx);
 	} else
 		huge_prof_ctx_set(ptr, ctx);
@@ -365,16 +393,18 @@ prof_sample_accum_update(size_t size)
 {
 	prof_tdata_t *prof_tdata;

+	cassert(config_prof);
 	/* Sampling logic is unnecessary if the interval is 1. */
 	assert(opt_lg_prof_sample != 0);

-	prof_tdata = PROF_TCACHE_GET();
-	assert(prof_tdata != NULL);
+	prof_tdata = *prof_tdata_tsd_get();
+	if ((uintptr_t)prof_tdata <= (uintptr_t)PROF_TDATA_STATE_MAX)
+		return (true);

 	/* Take care to avoid integer overflow. */
 	if (size >= prof_tdata->threshold - prof_tdata->accum) {
 		prof_tdata->accum -= (prof_tdata->threshold - size);
-		/* Compute new prof_sample_threshold. */
+		/* Compute new sample threshold. */
 		prof_sample_threshold_update(prof_tdata);
 		while (prof_tdata->accum >= prof_tdata->threshold) {
 			prof_tdata->accum -= prof_tdata->threshold;
@@ -391,8 +421,9 @@ JEMALLOC_INLINE void
 prof_malloc(const void *ptr, size_t size, prof_thr_cnt_t *cnt)
 {

+	cassert(config_prof);
 	assert(ptr != NULL);
-	assert(size == isalloc(ptr));
+	assert(size == isalloc(ptr, true));

 	if (opt_lg_prof_sample != 0) {
 		if (prof_sample_accum_update(size)) {
@@ -401,7 +432,7 @@ prof_malloc(const void *ptr, size_t size, prof_thr_cnt_t *cnt)
 			 * always possible to tell in advance how large an
 			 * object's usable size will be, so there should never
 			 * be a difference between the size passed to
-			 * prof_alloc_prep() and prof_malloc().
+			 * PROF_ALLOC_PREP() and prof_malloc().
 			 */
 			assert((uintptr_t)cnt == (uintptr_t)1U);
 		}
@@ -437,15 +468,16 @@ prof_realloc(const void *ptr, size_t size, prof_thr_cnt_t *cnt,
 {
 	prof_thr_cnt_t *told_cnt;

+	cassert(config_prof);
 	assert(ptr != NULL || (uintptr_t)cnt <= (uintptr_t)1U);

 	if (ptr != NULL) {
-		assert(size == isalloc(ptr));
+		assert(size == isalloc(ptr, true));
 		if (opt_lg_prof_sample != 0) {
 			if (prof_sample_accum_update(size)) {
 				/*
 				 * Don't sample.  The size passed to
-				 * prof_alloc_prep() was larger than what
+				 * PROF_ALLOC_PREP() was larger than what
 				 * actually got allocated, so a backtrace was
 				 * captured for this allocation, even though
 				 * its actual size was insufficient to cross
@@ -463,10 +495,10 @@ prof_realloc(const void *ptr, size_t size, prof_thr_cnt_t *cnt,
 			 * It's too late to propagate OOM for this realloc(),
 			 * so operate directly on old_cnt->ctx->cnt_merged.
 			 */
-			malloc_mutex_lock(&old_ctx->lock);
+			malloc_mutex_lock(old_ctx->lock);
 			old_ctx->cnt_merged.curobjs--;
 			old_ctx->cnt_merged.curbytes -= old_size;
-			malloc_mutex_unlock(&old_ctx->lock);
+			malloc_mutex_unlock(old_ctx->lock);
 			told_cnt = (prof_thr_cnt_t *)(uintptr_t)1U;
 		}
 	} else
@@ -477,7 +509,7 @@ prof_realloc(const void *ptr, size_t size, prof_thr_cnt_t *cnt,
 	if ((uintptr_t)cnt > (uintptr_t)1U) {
 		prof_ctx_set(ptr, cnt->ctx);
 		cnt->epoch++;
-	} else
+	} else if (ptr != NULL)
 		prof_ctx_set(ptr, (prof_ctx_t *)(uintptr_t)1U);
 	/*********/
 	mb_write();
@@ -510,9 +542,12 @@ prof_free(const void *ptr, size_t size)
 {
 	prof_ctx_t *ctx = prof_ctx_get(ptr);

+	cassert(config_prof);
+
 	if ((uintptr_t)ctx > (uintptr_t)1) {
-		assert(size == isalloc(ptr));
-		prof_thr_cnt_t *tcnt = prof_lookup(ctx->bt);
+		prof_thr_cnt_t *tcnt;
+		assert(size == isalloc(ptr, true));
+		tcnt = prof_lookup(ctx->bt);

 		if (tcnt != NULL) {
 			tcnt->epoch++;
@@ -533,10 +568,10 @@ prof_free(const void *ptr, size_t size)
 			 * OOM during free() cannot be propagated, so operate
 			 * directly on cnt->ctx->cnt_merged.
 			 */
-			malloc_mutex_lock(&ctx->lock);
+			malloc_mutex_lock(ctx->lock);
 			ctx->cnt_merged.curobjs--;
 			ctx->cnt_merged.curbytes -= size;
-			malloc_mutex_unlock(&ctx->lock);
+			malloc_mutex_unlock(ctx->lock);
 		}
 	}
 }
@@ -544,4 +579,3 @@ prof_free(const void *ptr, size_t size)

 #endif /* JEMALLOC_H_INLINES */
 /******************************************************************************/
-#endif /* JEMALLOC_PROF */
--- a/jemalloc/include/jemalloc/internal/ql.h
+++ b/jemalloc/include/jemalloc/internal/ql.h
--- a/jemalloc/include/jemalloc/internal/qr.h
+++ b/jemalloc/include/jemalloc/internal/qr.h
--- a/jemalloc/include/jemalloc/internal/zone.h
+++ b/jemalloc/include/jemalloc/internal/zone.h
@@ -1,9 +1,9 @@
-#ifndef JEMALLOC_ZONE
-#  error "This source file is for zones on Darwin (OS X)."
-#endif
 /******************************************************************************/
 #ifdef JEMALLOC_H_TYPES

+/* Default per thread quarantine size if valgrind is enabled. */
+#define	JEMALLOC_VALGRIND_QUARANTINE_DEFAULT	(ZU(1) << 24)
+
 #endif /* JEMALLOC_H_TYPES */
 /******************************************************************************/
 #ifdef JEMALLOC_H_STRUCTS
@@ -12,8 +12,8 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_EXTERNS

-malloc_zone_t *create_zone(void);
-void	szone2ozone(malloc_zone_t *zone);
+void	quarantine(void *ptr);
+bool	quarantine_boot(void);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
@@ -21,3 +21,4 @@ void	szone2ozone(malloc_zone_t *zone);

 #endif /* JEMALLOC_H_INLINES */
 /******************************************************************************/
+
--- a/jemalloc/include/jemalloc/internal/rb.h
+++ b/jemalloc/include/jemalloc/internal/rb.h
@@ -223,88 +223,88 @@ a_prefix##reverse_iter(a_rbt_type *rbtree, a_type *start,		\
 * The following API is generated:
 *
 *   static void
- *   ex_new(ex_t *extree);
+ *   ex_new(ex_t *tree);
 *       Description: Initialize a red-black tree structure.
 *       Args:
- *         extree: Pointer to an uninitialized red-black tree object.
+ *         tree: Pointer to an uninitialized red-black tree object.
 *
 *   static ex_node_t *
- *   ex_first(ex_t *extree);
+ *   ex_first(ex_t *tree);
 *   static ex_node_t *
- *   ex_last(ex_t *extree);
- *       Description: Get the first/last node in extree.
+ *   ex_last(ex_t *tree);
+ *       Description: Get the first/last node in tree.
 *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *       Ret: First/last node in extree, or NULL if extree is empty.
+ *         tree: Pointer to an initialized red-black tree object.
+ *       Ret: First/last node in tree, or NULL if tree is empty.
 *
 *   static ex_node_t *
- *   ex_next(ex_t *extree, ex_node_t *node);
+ *   ex_next(ex_t *tree, ex_node_t *node);
 *   static ex_node_t *
- *   ex_prev(ex_t *extree, ex_node_t *node);
+ *   ex_prev(ex_t *tree, ex_node_t *node);
 *       Description: Get node's successor/predecessor.
 *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *         node : A node in extree.
- *       Ret: node's successor/predecessor in extree, or NULL if node is
+ *         tree: Pointer to an initialized red-black tree object.
+ *         node: A node in tree.
+ *       Ret: node's successor/predecessor in tree, or NULL if node is
 *            last/first.
 *
 *   static ex_node_t *
- *   ex_search(ex_t *extree, ex_node_t *key);
+ *   ex_search(ex_t *tree, ex_node_t *key);
 *       Description: Search for node that matches key.
 *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *         key  : Search key.
- *       Ret: Node in extree that matches key, or NULL if no match.
+ *         tree: Pointer to an initialized red-black tree object.
+ *         key : Search key.
+ *       Ret: Node in tree that matches key, or NULL if no match.
 *
 *   static ex_node_t *
- *   ex_nsearch(ex_t *extree, ex_node_t *key);
+ *   ex_nsearch(ex_t *tree, ex_node_t *key);
 *   static ex_node_t *
- *   ex_psearch(ex_t *extree, ex_node_t *key);
+ *   ex_psearch(ex_t *tree, ex_node_t *key);
 *       Description: Search for node that matches key.  If no match is found,
 *                    return what would be key's successor/predecessor, were
- *                    key in extree.
+ *                    key in tree.
 *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *         key   : Search key.
- *       Ret: Node in extree that matches key, or if no match, hypothetical
- *            node's successor/predecessor (NULL if no successor/predecessor).
+ *         tree: Pointer to an initialized red-black tree object.
+ *         key : Search key.
+ *       Ret: Node in tree that matches key, or if no match, hypothetical node's
+ *            successor/predecessor (NULL if no successor/predecessor).
 *
 *   static void
- *   ex_insert(ex_t *extree, ex_node_t *node);
- *       Description: Insert node into extree.
+ *   ex_insert(ex_t *tree, ex_node_t *node);
+ *       Description: Insert node into tree.
 *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *         node  : Node to be inserted into extree.
+ *         tree: Pointer to an initialized red-black tree object.
+ *         node: Node to be inserted into tree.
 *
 *   static void
- *   ex_remove(ex_t *extree, ex_node_t *node);
- *       Description: Remove node from extree.
+ *   ex_remove(ex_t *tree, ex_node_t *node);
+ *       Description: Remove node from tree.
 *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *         node  : Node in extree to be removed.
+ *         tree: Pointer to an initialized red-black tree object.
+ *         node: Node in tree to be removed.
 *
 *   static ex_node_t *
- *   ex_iter(ex_t *extree, ex_node_t *start, ex_node_t *(*cb)(ex_t *,
+ *   ex_iter(ex_t *tree, ex_node_t *start, ex_node_t *(*cb)(ex_t *,
 *     ex_node_t *, void *), void *arg);
 *   static ex_node_t *
- *   ex_reverse_iter(ex_t *extree, ex_node_t *start, ex_node *(*cb)(ex_t *,
+ *   ex_reverse_iter(ex_t *tree, ex_node_t *start, ex_node *(*cb)(ex_t *,
 *     ex_node_t *, void *), void *arg);
- *       Description: Iterate forward/backward over extree, starting at node.
- *                    If extree is modified, iteration must be immediately
+ *       Description: Iterate forward/backward over tree, starting at node.  If
+ *                    tree is modified, iteration must be immediately
 *                    terminated by the callback function that causes the
 *                    modification.
 *       Args:
- *         extree: Pointer to an initialized red-black tree object.
- *         start : Node at which to start iteration, or NULL to start at
- *                 first/last node.
- *         cb    : Callback function, which is called for each node during
- *                 iteration.  Under normal circumstances the callback function
- *                 should return NULL, which causes iteration to continue.  If a
- *                 callback function returns non-NULL, iteration is immediately
- *                 terminated and the non-NULL return value is returned by the
- *                 iterator.  This is useful for re-starting iteration after
- *                 modifying extree.
- *         arg   : Opaque pointer passed to cb().
+ *         tree : Pointer to an initialized red-black tree object.
+ *         start: Node at which to start iteration, or NULL to start at
+ *                first/last node.
+ *         cb   : Callback function, which is called for each node during
+ *                iteration.  Under normal circumstances the callback function
+ *                should return NULL, which causes iteration to continue.  If a
+ *                callback function returns non-NULL, iteration is immediately
+ *                terminated and the non-NULL return value is returned by the
+ *                iterator.  This is useful for re-starting iteration after
+ *                modifying tree.
+ *         arg  : Opaque pointer passed to cb().
 *       Ret: NULL if iteration completed, or the non-NULL callback return value
 *            that caused termination of the iteration.
 */
--- a/jemalloc/include/jemalloc/internal/rtree.h
+++ b/jemalloc/include/jemalloc/internal/rtree.h
@@ -36,6 +36,9 @@ struct rtree_s {
 #ifdef JEMALLOC_H_EXTERNS

 rtree_t	*rtree_new(unsigned bits);
+void	rtree_prefork(rtree_t *rtree);
+void	rtree_postfork_parent(rtree_t *rtree);
+void	rtree_postfork_child(rtree_t *rtree);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
@@ -49,7 +52,7 @@ void	*rtree_get(rtree_t *rtree, uintptr_t key);
 bool	rtree_set(rtree_t *rtree, uintptr_t key, void *val);
 #endif

-#if (defined(JEMALLOC_ENABLE_INLINE) || defined(RTREE_C_))
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_RTREE_C_))
 #define	RTREE_GET_GENERATE(f)						\
 /* The least significant bits of the key are ignored. */		\
 JEMALLOC_INLINE void *							\
--- a/include/jemalloc/internal/size_classes.sh
+++ b/include/jemalloc/internal/size_classes.sh
@@ -0,0 +1,122 @@
+#!/bin/sh
+
+# The following limits are chosen such that they cover all supported platforms.
+
+# Range of quanta.
+lg_qmin=3
+lg_qmax=4
+
+# The range of tiny size classes is [2^lg_tmin..2^(lg_q-1)].
+lg_tmin=3
+
+# Range of page sizes.
+lg_pmin=12
+lg_pmax=16
+
+pow2() {
+  e=$1
+  pow2_result=1
+  while [ ${e} -gt 0 ] ; do
+    pow2_result=$((${pow2_result} + ${pow2_result}))
+    e=$((${e} - 1))
+  done
+}
+
+cat <<EOF
+/* This file was automatically generated by size_classes.sh. */
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+EOF
+
+lg_q=${lg_qmin}
+while [ ${lg_q} -le ${lg_qmax} ] ; do
+  lg_t=${lg_tmin}
+  while [ ${lg_t} -le ${lg_q} ] ; do
+    lg_p=${lg_pmin}
+    while [ ${lg_p} -le ${lg_pmax} ] ; do
+      echo "#if (LG_TINY_MIN == ${lg_t} && LG_QUANTUM == ${lg_q} && LG_PAGE == ${lg_p})"
+      echo "#define	SIZE_CLASSES_DEFINED"
+      pow2 ${lg_q}; q=${pow2_result}
+      pow2 ${lg_t}; t=${pow2_result}
+      pow2 ${lg_p}; p=${pow2_result}
+      bin=0
+      psz=0
+      sz=${t}
+      delta=$((${sz} - ${psz}))
+      echo "/*  SIZE_CLASS(bin,	delta,	sz) */"
+      echo "#define	SIZE_CLASSES							\\"
+
+      # Tiny size classes.
+      while [ ${sz} -lt ${q} ] ; do
+        echo "    SIZE_CLASS(${bin},	${delta},	${sz})					\\"
+        bin=$((${bin} + 1))
+        psz=${sz}
+        sz=$((${sz} + ${sz}))
+        delta=$((${sz} - ${psz}))
+      done
+      # Quantum-multiple size classes.  For each doubling of sz, as many as 4
+      # size classes exist.  Their spacing is the greater of:
+      # - q
+      # - sz/4, where sz is a power of 2
+      while [ ${sz} -lt ${p} ] ; do
+        if [ ${sz} -ge $((${q} * 4)) ] ; then
+          i=$((${sz} / 4))
+        else
+          i=${q}
+        fi
+        next_2pow=$((${sz} * 2))
+        while [ ${sz} -lt $next_2pow ] ; do
+          echo "    SIZE_CLASS(${bin},	${delta},	${sz})					\\"
+          bin=$((${bin} + 1))
+          psz=${sz}
+          sz=$((${sz} + ${i}))
+          delta=$((${sz} - ${psz}))
+        done
+      done
+      echo
+      echo "#define	NBINS		${bin}"
+      echo "#define	SMALL_MAXCLASS	${psz}"
+      echo "#endif"
+      echo
+      lg_p=$((${lg_p} + 1))
+    done
+    lg_t=$((${lg_t} + 1))
+  done
+  lg_q=$((${lg_q} + 1))
+done
+
+cat <<EOF
+#ifndef SIZE_CLASSES_DEFINED
+#  error "No size class definitions match configuration"
+#endif
+#undef SIZE_CLASSES_DEFINED
+/*
+ * The small_size2bin lookup table uses uint8_t to encode each bin index, so we
+ * cannot support more than 256 small size classes.  Further constrain NBINS to
+ * 255 to support prof_promote, since all small size classes, plus a "not
+ * small" size class must be stored in 8 bits of arena_chunk_map_t's bits
+ * field.
+ */
+#if (NBINS > 255)
+#  error "Too many small size classes"
+#endif
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
+EOF
--- a/jemalloc/include/jemalloc/internal/stats.h
+++ b/jemalloc/include/jemalloc/internal/stats.h
@@ -1,25 +1,16 @@
 /******************************************************************************/
 #ifdef JEMALLOC_H_TYPES

-#define	UMAX2S_BUFSIZE	65
-
-#ifdef JEMALLOC_STATS
 typedef struct tcache_bin_stats_s tcache_bin_stats_t;
 typedef struct malloc_bin_stats_s malloc_bin_stats_t;
 typedef struct malloc_large_stats_s malloc_large_stats_t;
 typedef struct arena_stats_s arena_stats_t;
-#endif
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
 typedef struct chunk_stats_s chunk_stats_t;
-#endif

 #endif /* JEMALLOC_H_TYPES */
 /******************************************************************************/
 #ifdef JEMALLOC_H_STRUCTS

-#ifdef JEMALLOC_STATS
-
-#ifdef JEMALLOC_TCACHE
 struct tcache_bin_stats_s {
 	/*
 	 * Number of allocation requests that corresponded to the size of this
@@ -27,7 +18,6 @@ struct tcache_bin_stats_s {
 	 */
 	uint64_t	nrequests;
 };
-#endif

 struct malloc_bin_stats_s {
 	/*
@@ -52,13 +42,11 @@ struct malloc_bin_stats_s {
 	 */
 	uint64_t	nrequests;

-#ifdef JEMALLOC_TCACHE
 	/* Number of tcache fills from this bin. */
 	uint64_t	nfills;

 	/* Number of tcache flushes to this bin. */
 	uint64_t	nflushes;
-#endif

 	/* Total number of runs created for this bin's size class. */
 	uint64_t	nruns;
@@ -69,9 +57,6 @@ struct malloc_bin_stats_s {
 	 */
 	uint64_t	reruns;

-	/* High-water mark for this bin. */
-	size_t		highruns;
-
 	/* Current number of runs in this bin. */
 	size_t		curruns;
 };
@@ -93,9 +78,6 @@ struct malloc_large_stats_s {
 	 */
 	uint64_t	nrequests;

-	/* High-water mark for this size class. */
-	size_t		highruns;
-
 	/* Current number of runs of this size class. */
 	size_t		curruns;
 };
@@ -127,14 +109,10 @@ struct arena_stats_s {
 	 */
 	malloc_large_stats_t	*lstats;
 };
-#endif /* JEMALLOC_STATS */

-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
 struct chunk_stats_s {
-#  ifdef JEMALLOC_STATS
 	/* Number of chunks that were allocated. */
 	uint64_t	nchunks;
-#  endif

 	/* High-water mark for number of chunks allocated. */
 	size_t		highchunks;
@@ -146,7 +124,6 @@ struct chunk_stats_s {
 	 */
 	size_t		curchunks;
 };
-#endif /* JEMALLOC_STATS */

 #endif /* JEMALLOC_H_STRUCTS */
 /******************************************************************************/
@@ -154,21 +131,43 @@ struct chunk_stats_s {

 extern bool	opt_stats_print;

-char	*u2s(uint64_t x, unsigned base, char *s);
-#ifdef JEMALLOC_STATS
-void malloc_cprintf(void (*write)(void *, const char *), void *cbopaque,
-    const char *format, ...) JEMALLOC_ATTR(format(printf, 3, 4));
-void	malloc_printf(const char *format, ...)
-    JEMALLOC_ATTR(format(printf, 1, 2));
-#endif
+extern size_t	stats_cactive;
+
 void	stats_print(void (*write)(void *, const char *), void *cbopaque,
    const char *opts);

 #endif /* JEMALLOC_H_EXTERNS */
 /******************************************************************************/
-#ifdef JEMALLOC_STATS
 #ifdef JEMALLOC_H_INLINES

+#ifndef JEMALLOC_ENABLE_INLINE
+size_t	stats_cactive_get(void);
+void	stats_cactive_add(size_t size);
+void	stats_cactive_sub(size_t size);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_STATS_C_))
+JEMALLOC_INLINE size_t
+stats_cactive_get(void)
+{
+
+	return (atomic_read_z(&stats_cactive));
+}
+
+JEMALLOC_INLINE void
+stats_cactive_add(size_t size)
+{
+
+	atomic_add_z(&stats_cactive, size);
+}
+
+JEMALLOC_INLINE void
+stats_cactive_sub(size_t size)
+{
+
+	atomic_sub_z(&stats_cactive, size);
+}
+#endif
+
 #endif /* JEMALLOC_H_INLINES */
-#endif /* JEMALLOC_STATS */
 /******************************************************************************/
--- a/include/jemalloc/internal/tcache.h
+++ b/include/jemalloc/internal/tcache.h
@@ -0,0 +1,440 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+typedef struct tcache_bin_info_s tcache_bin_info_t;
+typedef struct tcache_bin_s tcache_bin_t;
+typedef struct tcache_s tcache_t;
+
+/*
+ * tcache pointers close to NULL are used to encode state information that is
+ * used for two purposes: preventing thread caching on a per thread basis and
+ * cleaning up during thread shutdown.
+ */
+#define	TCACHE_STATE_DISABLED		((tcache_t *)(uintptr_t)1)
+#define	TCACHE_STATE_REINCARNATED	((tcache_t *)(uintptr_t)2)
+#define	TCACHE_STATE_PURGATORY		((tcache_t *)(uintptr_t)3)
+#define	TCACHE_STATE_MAX		TCACHE_STATE_PURGATORY
+
+/*
+ * Absolute maximum number of cache slots for each small bin in the thread
+ * cache.  This is an additional constraint beyond that imposed as: twice the
+ * number of regions per run for this size class.
+ *
+ * This constant must be an even number.
+ */
+#define	TCACHE_NSLOTS_SMALL_MAX		200
+
+/* Number of cache slots for large size classes. */
+#define	TCACHE_NSLOTS_LARGE		20
+
+/* (1U << opt_lg_tcache_max) is used to compute tcache_maxclass. */
+#define	LG_TCACHE_MAXCLASS_DEFAULT	15
+
+/*
+ * TCACHE_GC_SWEEP is the approximate number of allocation events between
+ * full GC sweeps.  Integer rounding may cause the actual number to be
+ * slightly higher, since GC is performed incrementally.
+ */
+#define	TCACHE_GC_SWEEP			8192
+
+/* Number of tcache allocation/deallocation events between incremental GCs. */
+#define	TCACHE_GC_INCR							\
+    ((TCACHE_GC_SWEEP / NBINS) + ((TCACHE_GC_SWEEP / NBINS == 0) ? 0 : 1))
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+typedef enum {
+	tcache_enabled_false   = 0, /* Enable cast to/from bool. */
+	tcache_enabled_true    = 1,
+	tcache_enabled_default = 2
+} tcache_enabled_t;
+
+/*
+ * Read-only information associated with each element of tcache_t's tbins array
+ * is stored separately, mainly to reduce memory usage.
+ */
+struct tcache_bin_info_s {
+	unsigned	ncached_max;	/* Upper limit on ncached. */
+};
+
+struct tcache_bin_s {
+	tcache_bin_stats_t tstats;
+	int		low_water;	/* Min # cached since last GC. */
+	unsigned	lg_fill_div;	/* Fill (ncached_max >> lg_fill_div). */
+	unsigned	ncached;	/* # of cached objects. */
+	void		**avail;	/* Stack of available objects. */
+};
+
+struct tcache_s {
+	ql_elm(tcache_t) link;		/* Used for aggregating stats. */
+	uint64_t	prof_accumbytes;/* Cleared after arena_prof_accum() */
+	arena_t		*arena;		/* This thread's arena. */
+	unsigned	ev_cnt;		/* Event count since incremental GC. */
+	unsigned	next_gc_bin;	/* Next bin to GC. */
+	tcache_bin_t	tbins[1];	/* Dynamically sized. */
+	/*
+	 * The pointer stacks associated with tbins follow as a contiguous
+	 * array.  During tcache initialization, the avail pointer in each
+	 * element of tbins is initialized to point to the proper offset within
+	 * this array.
+	 */
+};
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+extern bool	opt_tcache;
+extern ssize_t	opt_lg_tcache_max;
+
+extern tcache_bin_info_t	*tcache_bin_info;
+
+/*
+ * Number of tcache bins.  There are NBINS small-object bins, plus 0 or more
+ * large-object bins.
+ */
+extern size_t			nhbins;
+
+/* Maximum cached size class. */
+extern size_t			tcache_maxclass;
+
+size_t	tcache_salloc(const void *ptr);
+void	tcache_event_hard(tcache_t *tcache);
+void	*tcache_alloc_small_hard(tcache_t *tcache, tcache_bin_t *tbin,
+    size_t binind);
+void	tcache_bin_flush_small(tcache_bin_t *tbin, size_t binind, unsigned rem,
+    tcache_t *tcache);
+void	tcache_bin_flush_large(tcache_bin_t *tbin, size_t binind, unsigned rem,
+    tcache_t *tcache);
+void	tcache_arena_associate(tcache_t *tcache, arena_t *arena);
+void	tcache_arena_dissociate(tcache_t *tcache);
+tcache_t *tcache_create(arena_t *arena);
+void	tcache_destroy(tcache_t *tcache);
+void	tcache_thread_cleanup(void *arg);
+void	tcache_stats_merge(tcache_t *tcache, arena_t *arena);
+bool	tcache_boot0(void);
+bool	tcache_boot1(void);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+malloc_tsd_protos(JEMALLOC_ATTR(unused), tcache, tcache_t *)
+malloc_tsd_protos(JEMALLOC_ATTR(unused), tcache_enabled, tcache_enabled_t)
+
+void	tcache_event(tcache_t *tcache);
+void	tcache_flush(void);
+bool	tcache_enabled_get(void);
+tcache_t *tcache_get(bool create);
+void	tcache_enabled_set(bool enabled);
+void	*tcache_alloc_easy(tcache_bin_t *tbin);
+void	*tcache_alloc_small(tcache_t *tcache, size_t size, bool zero);
+void	*tcache_alloc_large(tcache_t *tcache, size_t size, bool zero);
+void	tcache_dalloc_small(tcache_t *tcache, void *ptr, size_t binind);
+void	tcache_dalloc_large(tcache_t *tcache, void *ptr, size_t size);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_TCACHE_C_))
+/* Map of thread-specific caches. */
+malloc_tsd_externs(tcache, tcache_t *)
+malloc_tsd_funcs(JEMALLOC_INLINE, tcache, tcache_t *, NULL,
+    tcache_thread_cleanup)
+/* Per thread flag that allows thread caches to be disabled. */
+malloc_tsd_externs(tcache_enabled, tcache_enabled_t)
+malloc_tsd_funcs(JEMALLOC_INLINE, tcache_enabled, tcache_enabled_t,
+    tcache_enabled_default, malloc_tsd_no_cleanup)
+
+JEMALLOC_INLINE void
+tcache_flush(void)
+{
+	tcache_t *tcache;
+
+	cassert(config_tcache);
+
+	tcache = *tcache_tsd_get();
+	if ((uintptr_t)tcache <= (uintptr_t)TCACHE_STATE_MAX)
+		return;
+	tcache_destroy(tcache);
+	tcache = NULL;
+	tcache_tsd_set(&tcache);
+}
+
+JEMALLOC_INLINE bool
+tcache_enabled_get(void)
+{
+	tcache_enabled_t tcache_enabled;
+
+	cassert(config_tcache);
+
+	tcache_enabled = *tcache_enabled_tsd_get();
+	if (tcache_enabled == tcache_enabled_default) {
+		tcache_enabled = (tcache_enabled_t)opt_tcache;
+		tcache_enabled_tsd_set(&tcache_enabled);
+	}
+
+	return ((bool)tcache_enabled);
+}
+
+JEMALLOC_INLINE void
+tcache_enabled_set(bool enabled)
+{
+	tcache_enabled_t tcache_enabled;
+	tcache_t *tcache;
+
+	cassert(config_tcache);
+
+	tcache_enabled = (tcache_enabled_t)enabled;
+	tcache_enabled_tsd_set(&tcache_enabled);
+	tcache = *tcache_tsd_get();
+	if (enabled) {
+		if (tcache == TCACHE_STATE_DISABLED) {
+			tcache = NULL;
+			tcache_tsd_set(&tcache);
+		}
+	} else /* disabled */ {
+		if (tcache > TCACHE_STATE_MAX) {
+			tcache_destroy(tcache);
+			tcache = NULL;
+		}
+		if (tcache == NULL) {
+			tcache = TCACHE_STATE_DISABLED;
+			tcache_tsd_set(&tcache);
+		}
+	}
+}
+
+JEMALLOC_INLINE tcache_t *
+tcache_get(bool create)
+{
+	tcache_t *tcache;
+
+	if (config_tcache == false)
+		return (NULL);
+	if (config_lazy_lock && isthreaded == false)
+		return (NULL);
+
+	tcache = *tcache_tsd_get();
+	if ((uintptr_t)tcache <= (uintptr_t)TCACHE_STATE_MAX) {
+		if (tcache == TCACHE_STATE_DISABLED)
+			return (NULL);
+		if (tcache == NULL) {
+			if (create == false) {
+				/*
+				 * Creating a tcache here would cause
+				 * allocation as a side effect of free().
+				 * Ordinarily that would be okay since
+				 * tcache_create() failure is a soft failure
+				 * that doesn't propagate.  However, if TLS
+				 * data are freed via free() as in glibc,
+				 * subtle corruption could result from setting
+				 * a TLS variable after its backing memory is
+				 * freed.
+				 */
+				return (NULL);
+			}
+			if (tcache_enabled_get() == false) {
+				tcache_enabled_set(false); /* Memoize. */
+				return (NULL);
+			}
+			return (tcache_create(choose_arena(NULL)));
+		}
+		if (tcache == TCACHE_STATE_PURGATORY) {
+			/*
+			 * Make a note that an allocator function was called
+			 * after tcache_thread_cleanup() was called.
+			 */
+			tcache = TCACHE_STATE_REINCARNATED;
+			tcache_tsd_set(&tcache);
+			return (NULL);
+		}
+		if (tcache == TCACHE_STATE_REINCARNATED)
+			return (NULL);
+		not_reached();
+	}
+
+	return (tcache);
+}
+
+JEMALLOC_INLINE void
+tcache_event(tcache_t *tcache)
+{
+
+	if (TCACHE_GC_INCR == 0)
+		return;
+
+	tcache->ev_cnt++;
+	assert(tcache->ev_cnt <= TCACHE_GC_INCR);
+	if (tcache->ev_cnt == TCACHE_GC_INCR)
+		tcache_event_hard(tcache);
+}
+
+JEMALLOC_INLINE void *
+tcache_alloc_easy(tcache_bin_t *tbin)
+{
+	void *ret;
+
+	if (tbin->ncached == 0) {
+		tbin->low_water = -1;
+		return (NULL);
+	}
+	tbin->ncached--;
+	if ((int)tbin->ncached < tbin->low_water)
+		tbin->low_water = tbin->ncached;
+	ret = tbin->avail[tbin->ncached];
+	return (ret);
+}
+
+JEMALLOC_INLINE void *
+tcache_alloc_small(tcache_t *tcache, size_t size, bool zero)
+{
+	void *ret;
+	size_t binind;
+	tcache_bin_t *tbin;
+
+	binind = SMALL_SIZE2BIN(size);
+	assert(binind < NBINS);
+	tbin = &tcache->tbins[binind];
+	ret = tcache_alloc_easy(tbin);
+	if (ret == NULL) {
+		ret = tcache_alloc_small_hard(tcache, tbin, binind);
+		if (ret == NULL)
+			return (NULL);
+	}
+	assert(tcache_salloc(ret) == arena_bin_info[binind].reg_size);
+
+	if (zero == false) {
+		if (config_fill) {
+			if (opt_junk) {
+				arena_alloc_junk_small(ret,
+				    &arena_bin_info[binind], false);
+			} else if (opt_zero)
+				memset(ret, 0, size);
+		}
+	} else {
+		if (config_fill && opt_junk) {
+			arena_alloc_junk_small(ret, &arena_bin_info[binind],
+			    true);
+		}
+		VALGRIND_MAKE_MEM_UNDEFINED(ret, size);
+		memset(ret, 0, size);
+	}
+
+	if (config_stats)
+		tbin->tstats.nrequests++;
+	if (config_prof)
+		tcache->prof_accumbytes += arena_bin_info[binind].reg_size;
+	tcache_event(tcache);
+	return (ret);
+}
+
+JEMALLOC_INLINE void *
+tcache_alloc_large(tcache_t *tcache, size_t size, bool zero)
+{
+	void *ret;
+	size_t binind;
+	tcache_bin_t *tbin;
+
+	size = PAGE_CEILING(size);
+	assert(size <= tcache_maxclass);
+	binind = NBINS + (size >> LG_PAGE) - 1;
+	assert(binind < nhbins);
+	tbin = &tcache->tbins[binind];
+	ret = tcache_alloc_easy(tbin);
+	if (ret == NULL) {
+		/*
+		 * Only allocate one large object at a time, because it's quite
+		 * expensive to create one and not use it.
+		 */
+		ret = arena_malloc_large(tcache->arena, size, zero);
+		if (ret == NULL)
+			return (NULL);
+	} else {
+		if (config_prof && prof_promote && size == PAGE) {
+			arena_chunk_t *chunk =
+			    (arena_chunk_t *)CHUNK_ADDR2BASE(ret);
+			size_t pageind = (((uintptr_t)ret - (uintptr_t)chunk) >>
+			    LG_PAGE);
+			arena_mapbits_large_binind_set(chunk, pageind,
+			    BININD_INVALID);
+		}
+		if (zero == false) {
+			if (config_fill) {
+				if (opt_junk)
+					memset(ret, 0xa5, size);
+				else if (opt_zero)
+					memset(ret, 0, size);
+			}
+		} else {
+			VALGRIND_MAKE_MEM_UNDEFINED(ret, size);
+			memset(ret, 0, size);
+		}
+
+		if (config_stats)
+			tbin->tstats.nrequests++;
+		if (config_prof)
+			tcache->prof_accumbytes += size;
+	}
+
+	tcache_event(tcache);
+	return (ret);
+}
+
+JEMALLOC_INLINE void
+tcache_dalloc_small(tcache_t *tcache, void *ptr, size_t binind)
+{
+	tcache_bin_t *tbin;
+	tcache_bin_info_t *tbin_info;
+
+	assert(tcache_salloc(ptr) <= SMALL_MAXCLASS);
+
+	if (config_fill && opt_junk)
+		arena_dalloc_junk_small(ptr, &arena_bin_info[binind]);
+
+	tbin = &tcache->tbins[binind];
+	tbin_info = &tcache_bin_info[binind];
+	if (tbin->ncached == tbin_info->ncached_max) {
+		tcache_bin_flush_small(tbin, binind, (tbin_info->ncached_max >>
+		    1), tcache);
+	}
+	assert(tbin->ncached < tbin_info->ncached_max);
+	tbin->avail[tbin->ncached] = ptr;
+	tbin->ncached++;
+
+	tcache_event(tcache);
+}
+
+JEMALLOC_INLINE void
+tcache_dalloc_large(tcache_t *tcache, void *ptr, size_t size)
+{
+	size_t binind;
+	tcache_bin_t *tbin;
+	tcache_bin_info_t *tbin_info;
+
+	assert((size & PAGE_MASK) == 0);
+	assert(tcache_salloc(ptr) > SMALL_MAXCLASS);
+	assert(tcache_salloc(ptr) <= tcache_maxclass);
+
+	binind = NBINS + (size >> LG_PAGE) - 1;
+
+	if (config_fill && opt_junk)
+		memset(ptr, 0x5a, size);
+
+	tbin = &tcache->tbins[binind];
+	tbin_info = &tcache_bin_info[binind];
+	if (tbin->ncached == tbin_info->ncached_max) {
+		tcache_bin_flush_large(tbin, binind, (tbin_info->ncached_max >>
+		    1), tcache);
+	}
+	assert(tbin->ncached < tbin_info->ncached_max);
+	tbin->avail[tbin->ncached] = ptr;
+	tbin->ncached++;
+
+	tcache_event(tcache);
+}
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
--- a/include/jemalloc/internal/tsd.h
+++ b/include/jemalloc/internal/tsd.h
@@ -0,0 +1,397 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+/* Maximum number of malloc_tsd users with cleanup functions. */
+#define	MALLOC_TSD_CLEANUPS_MAX	8
+
+typedef bool (*malloc_tsd_cleanup_t)(void);
+
+/*
+ * TLS/TSD-agnostic macro-based implementation of thread-specific data.  There
+ * are four macros that support (at least) three use cases: file-private,
+ * library-private, and library-private inlined.  Following is an example
+ * library-private tsd variable:
+ *
+ * In example.h:
+ *   typedef struct {
+ *           int x;
+ *           int y;
+ *   } example_t;
+ *   #define EX_INITIALIZER JEMALLOC_CONCAT({0, 0})
+ *   malloc_tsd_protos(, example, example_t *)
+ *   malloc_tsd_externs(example, example_t *)
+ * In example.c:
+ *   malloc_tsd_data(, example, example_t *, EX_INITIALIZER)
+ *   malloc_tsd_funcs(, example, example_t *, EX_INITIALIZER,
+ *       example_tsd_cleanup)
+ *
+ * The result is a set of generated functions, e.g.:
+ *
+ *   bool example_tsd_boot(void) {...}
+ *   example_t **example_tsd_get() {...}
+ *   void example_tsd_set(example_t **val) {...}
+ *
+ * Note that all of the functions deal in terms of (a_type *) rather than
+ * (a_type)  so that it is possible to support non-pointer types (unlike
+ * pthreads TSD).  example_tsd_cleanup() is passed an (a_type *) pointer that is
+ * cast to (void *).  This means that the cleanup function needs to cast *and*
+ * dereference the function argument, e.g.:
+ *
+ *   void
+ *   example_tsd_cleanup(void *arg)
+ *   {
+ *           example_t *example = *(example_t **)arg;
+ *
+ *           [...]
+ *           if ([want the cleanup function to be called again]) {
+ *                   example_tsd_set(&example);
+ *           }
+ *   }
+ *
+ * If example_tsd_set() is called within example_tsd_cleanup(), it will be
+ * called again.  This is similar to how pthreads TSD destruction works, except
+ * that pthreads only calls the cleanup function again if the value was set to
+ * non-NULL.
+ */
+
+/* malloc_tsd_protos(). */
+#define	malloc_tsd_protos(a_attr, a_name, a_type)			\
+a_attr bool								\
+a_name##_tsd_boot(void);						\
+a_attr a_type *								\
+a_name##_tsd_get(void);							\
+a_attr void								\
+a_name##_tsd_set(a_type *val);
+
+/* malloc_tsd_externs(). */
+#ifdef JEMALLOC_MALLOC_THREAD_CLEANUP
+#define	malloc_tsd_externs(a_name, a_type)				\
+extern __thread a_type	a_name##_tls;					\
+extern __thread bool	a_name##_initialized;				\
+extern bool		a_name##_booted;
+#elif (defined(JEMALLOC_TLS))
+#define	malloc_tsd_externs(a_name, a_type)				\
+extern __thread a_type	a_name##_tls;					\
+extern pthread_key_t	a_name##_tsd;					\
+extern bool		a_name##_booted;
+#elif (defined(_WIN32))
+#define malloc_tsd_externs(a_name, a_type)				\
+extern DWORD		a_name##_tsd;					\
+extern bool		a_name##_booted;
+#else
+#define	malloc_tsd_externs(a_name, a_type)				\
+extern pthread_key_t	a_name##_tsd;					\
+extern bool		a_name##_booted;
+#endif
+
+/* malloc_tsd_data(). */
+#ifdef JEMALLOC_MALLOC_THREAD_CLEANUP
+#define	malloc_tsd_data(a_attr, a_name, a_type, a_initializer)		\
+a_attr __thread a_type JEMALLOC_TLS_MODEL				\
+    a_name##_tls = a_initializer;					\
+a_attr __thread bool JEMALLOC_TLS_MODEL					\
+    a_name##_initialized = false;					\
+a_attr bool		a_name##_booted = false;
+#elif (defined(JEMALLOC_TLS))
+#define	malloc_tsd_data(a_attr, a_name, a_type, a_initializer)		\
+a_attr __thread a_type JEMALLOC_TLS_MODEL				\
+    a_name##_tls = a_initializer;					\
+a_attr pthread_key_t	a_name##_tsd;					\
+a_attr bool		a_name##_booted = false;
+#elif (defined(_WIN32))
+#define	malloc_tsd_data(a_attr, a_name, a_type, a_initializer)		\
+a_attr DWORD		a_name##_tsd;					\
+a_attr bool		a_name##_booted = false;
+#else
+#define	malloc_tsd_data(a_attr, a_name, a_type, a_initializer)		\
+a_attr pthread_key_t	a_name##_tsd;					\
+a_attr bool		a_name##_booted = false;
+#endif
+
+/* malloc_tsd_funcs(). */
+#ifdef JEMALLOC_MALLOC_THREAD_CLEANUP
+#define	malloc_tsd_funcs(a_attr, a_name, a_type, a_initializer,		\
+    a_cleanup)								\
+/* Initialization/cleanup. */						\
+a_attr bool								\
+a_name##_tsd_cleanup_wrapper(void)					\
+{									\
+									\
+	if (a_name##_initialized) {					\
+		a_name##_initialized = false;				\
+		a_cleanup(&a_name##_tls);				\
+	}								\
+	return (a_name##_initialized);					\
+}									\
+a_attr bool								\
+a_name##_tsd_boot(void)							\
+{									\
+									\
+	if (a_cleanup != malloc_tsd_no_cleanup) {			\
+		malloc_tsd_cleanup_register(				\
+		    &a_name##_tsd_cleanup_wrapper);			\
+	}								\
+	a_name##_booted = true;						\
+	return (false);							\
+}									\
+/* Get/set. */								\
+a_attr a_type *								\
+a_name##_tsd_get(void)							\
+{									\
+									\
+	assert(a_name##_booted);					\
+	return (&a_name##_tls);						\
+}									\
+a_attr void								\
+a_name##_tsd_set(a_type *val)						\
+{									\
+									\
+	assert(a_name##_booted);					\
+	a_name##_tls = (*val);						\
+	if (a_cleanup != malloc_tsd_no_cleanup)				\
+		a_name##_initialized = true;				\
+}
+#elif (defined(JEMALLOC_TLS))
+#define	malloc_tsd_funcs(a_attr, a_name, a_type, a_initializer,		\
+    a_cleanup)								\
+/* Initialization/cleanup. */						\
+a_attr bool								\
+a_name##_tsd_boot(void)							\
+{									\
+									\
+	if (a_cleanup != malloc_tsd_no_cleanup) {			\
+		if (pthread_key_create(&a_name##_tsd, a_cleanup) != 0)	\
+			return (true);					\
+	}								\
+	a_name##_booted = true;						\
+	return (false);							\
+}									\
+/* Get/set. */								\
+a_attr a_type *								\
+a_name##_tsd_get(void)							\
+{									\
+									\
+	assert(a_name##_booted);					\
+	return (&a_name##_tls);						\
+}									\
+a_attr void								\
+a_name##_tsd_set(a_type *val)						\
+{									\
+									\
+	assert(a_name##_booted);					\
+	a_name##_tls = (*val);						\
+	if (a_cleanup != malloc_tsd_no_cleanup) {			\
+		if (pthread_setspecific(a_name##_tsd,			\
+		    (void *)(&a_name##_tls))) {				\
+			malloc_write("<jemalloc>: Error"		\
+			    " setting TSD for "#a_name"\n");		\
+			if (opt_abort)					\
+				abort();				\
+		}							\
+	}								\
+}
+#elif (defined(_WIN32))
+#define	malloc_tsd_funcs(a_attr, a_name, a_type, a_initializer,		\
+    a_cleanup)								\
+/* Data structure. */							\
+typedef struct {							\
+	bool	initialized;						\
+	a_type	val;							\
+} a_name##_tsd_wrapper_t;						\
+/* Initialization/cleanup. */						\
+a_attr bool								\
+a_name##_tsd_cleanup_wrapper(void)					\
+{									\
+	a_name##_tsd_wrapper_t *wrapper;				\
+									\
+	wrapper = (a_name##_tsd_wrapper_t *) TlsGetValue(a_name##_tsd);	\
+	if (wrapper == NULL)						\
+		return (false);						\
+	if (a_cleanup != malloc_tsd_no_cleanup &&			\
+	    wrapper->initialized) {					\
+		a_type val = wrapper->val;				\
+		a_type tsd_static_data = a_initializer;			\
+		wrapper->initialized = false;				\
+		wrapper->val = tsd_static_data;				\
+		a_cleanup(&val);					\
+		if (wrapper->initialized) {				\
+			/* Trigger another cleanup round. */		\
+			return (true);					\
+		}							\
+	}								\
+	malloc_tsd_dalloc(wrapper);					\
+	return (false);							\
+}									\
+a_attr bool								\
+a_name##_tsd_boot(void)							\
+{									\
+									\
+	a_name##_tsd = TlsAlloc();					\
+	if (a_name##_tsd == TLS_OUT_OF_INDEXES)				\
+		return (true);						\
+	if (a_cleanup != malloc_tsd_no_cleanup) {			\
+		malloc_tsd_cleanup_register(				\
+		    &a_name##_tsd_cleanup_wrapper);			\
+	}								\
+	a_name##_booted = true;						\
+	return (false);							\
+}									\
+/* Get/set. */								\
+a_attr a_name##_tsd_wrapper_t *						\
+a_name##_tsd_get_wrapper(void)						\
+{									\
+	a_name##_tsd_wrapper_t *wrapper = (a_name##_tsd_wrapper_t *)	\
+	    TlsGetValue(a_name##_tsd);					\
+									\
+	if (wrapper == NULL) {						\
+		wrapper = (a_name##_tsd_wrapper_t *)			\
+		    malloc_tsd_malloc(sizeof(a_name##_tsd_wrapper_t));	\
+		if (wrapper == NULL) {					\
+			malloc_write("<jemalloc>: Error allocating"	\
+			    " TSD for "#a_name"\n");			\
+			abort();					\
+		} else {						\
+			static a_type tsd_static_data = a_initializer;	\
+			wrapper->initialized = false;			\
+			wrapper->val = tsd_static_data;			\
+		}							\
+		if (!TlsSetValue(a_name##_tsd, (void *)wrapper)) {	\
+			malloc_write("<jemalloc>: Error setting"	\
+			    " TSD for "#a_name"\n");			\
+			abort();					\
+		}							\
+	}								\
+	return (wrapper);						\
+}									\
+a_attr a_type *								\
+a_name##_tsd_get(void)							\
+{									\
+	a_name##_tsd_wrapper_t *wrapper;				\
+									\
+	assert(a_name##_booted);					\
+	wrapper = a_name##_tsd_get_wrapper();				\
+	return (&wrapper->val);						\
+}									\
+a_attr void								\
+a_name##_tsd_set(a_type *val)						\
+{									\
+	a_name##_tsd_wrapper_t *wrapper;				\
+									\
+	assert(a_name##_booted);					\
+	wrapper = a_name##_tsd_get_wrapper();				\
+	wrapper->val = *(val);						\
+	if (a_cleanup != malloc_tsd_no_cleanup)				\
+		wrapper->initialized = true;				\
+}
+#else
+#define	malloc_tsd_funcs(a_attr, a_name, a_type, a_initializer,		\
+    a_cleanup)								\
+/* Data structure. */							\
+typedef struct {							\
+	bool	initialized;						\
+	a_type	val;							\
+} a_name##_tsd_wrapper_t;						\
+/* Initialization/cleanup. */						\
+a_attr void								\
+a_name##_tsd_cleanup_wrapper(void *arg)					\
+{									\
+	a_name##_tsd_wrapper_t *wrapper = (a_name##_tsd_wrapper_t *)arg;\
+									\
+	if (a_cleanup != malloc_tsd_no_cleanup &&			\
+	    wrapper->initialized) {					\
+		wrapper->initialized = false;				\
+		a_cleanup(&wrapper->val);				\
+		if (wrapper->initialized) {				\
+			/* Trigger another cleanup round. */		\
+			if (pthread_setspecific(a_name##_tsd,		\
+			    (void *)wrapper)) {				\
+				malloc_write("<jemalloc>: Error"	\
+				    " setting TSD for "#a_name"\n");	\
+				if (opt_abort)				\
+					abort();			\
+			}						\
+			return;						\
+		}							\
+	}								\
+	malloc_tsd_dalloc(wrapper);					\
+}									\
+a_attr bool								\
+a_name##_tsd_boot(void)							\
+{									\
+									\
+	if (pthread_key_create(&a_name##_tsd,				\
+	    a_name##_tsd_cleanup_wrapper) != 0)				\
+		return (true);						\
+	a_name##_booted = true;						\
+	return (false);							\
+}									\
+/* Get/set. */								\
+a_attr a_name##_tsd_wrapper_t *						\
+a_name##_tsd_get_wrapper(void)						\
+{									\
+	a_name##_tsd_wrapper_t *wrapper = (a_name##_tsd_wrapper_t *)	\
+	    pthread_getspecific(a_name##_tsd);				\
+									\
+	if (wrapper == NULL) {						\
+		wrapper = (a_name##_tsd_wrapper_t *)			\
+		    malloc_tsd_malloc(sizeof(a_name##_tsd_wrapper_t));	\
+		if (wrapper == NULL) {					\
+			malloc_write("<jemalloc>: Error allocating"	\
+			    " TSD for "#a_name"\n");			\
+			abort();					\
+		} else {						\
+			static a_type tsd_static_data = a_initializer;	\
+			wrapper->initialized = false;			\
+			wrapper->val = tsd_static_data;			\
+		}							\
+		if (pthread_setspecific(a_name##_tsd,			\
+		    (void *)wrapper)) {					\
+			malloc_write("<jemalloc>: Error setting"	\
+			    " TSD for "#a_name"\n");			\
+			abort();					\
+		}							\
+	}								\
+	return (wrapper);						\
+}									\
+a_attr a_type *								\
+a_name##_tsd_get(void)							\
+{									\
+	a_name##_tsd_wrapper_t *wrapper;				\
+									\
+	assert(a_name##_booted);					\
+	wrapper = a_name##_tsd_get_wrapper();				\
+	return (&wrapper->val);						\
+}									\
+a_attr void								\
+a_name##_tsd_set(a_type *val)						\
+{									\
+	a_name##_tsd_wrapper_t *wrapper;				\
+									\
+	assert(a_name##_booted);					\
+	wrapper = a_name##_tsd_get_wrapper();				\
+	wrapper->val = *(val);						\
+	if (a_cleanup != malloc_tsd_no_cleanup)				\
+		wrapper->initialized = true;				\
+}
+#endif
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+void	*malloc_tsd_malloc(size_t size);
+void	malloc_tsd_dalloc(void *wrapper);
+void	malloc_tsd_no_cleanup(void *);
+void	malloc_tsd_cleanup_register(bool (*f)(void));
+void	malloc_tsd_boot(void);
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
--- a/include/jemalloc/internal/util.h
+++ b/include/jemalloc/internal/util.h
@@ -0,0 +1,160 @@
+/******************************************************************************/
+#ifdef JEMALLOC_H_TYPES
+
+/* Size of stack-allocated buffer passed to buferror(). */
+#define	BUFERROR_BUF		64
+
+/*
+ * Size of stack-allocated buffer used by malloc_{,v,vc}printf().  This must be
+ * large enough for all possible uses within jemalloc.
+ */
+#define	MALLOC_PRINTF_BUFSIZE	4096
+
+/*
+ * Wrap a cpp argument that contains commas such that it isn't broken up into
+ * multiple arguments.
+ */
+#define JEMALLOC_CONCAT(...) __VA_ARGS__
+
+/*
+ * Silence compiler warnings due to uninitialized values.  This is used
+ * wherever the compiler fails to recognize that the variable is never used
+ * uninitialized.
+ */
+#ifdef JEMALLOC_CC_SILENCE
+#  define JEMALLOC_CC_SILENCE_INIT(v) = v
+#else
+#  define JEMALLOC_CC_SILENCE_INIT(v)
+#endif
+
+/*
+ * Define a custom assert() in order to reduce the chances of deadlock during
+ * assertion failure.
+ */
+#ifndef assert
+#define	assert(e) do {							\
+	if (config_debug && !(e)) {					\
+		malloc_printf(						\
+		    "<jemalloc>: %s:%d: Failed assertion: \"%s\"\n",	\
+		    __FILE__, __LINE__, #e);				\
+		abort();						\
+	}								\
+} while (0)
+#endif
+
+/* Use to assert a particular configuration, e.g., cassert(config_debug). */
+#define	cassert(c) do {							\
+	if ((c) == false)						\
+		assert(false);						\
+} while (0)
+
+#ifndef not_reached
+#define	not_reached() do {						\
+	if (config_debug) {						\
+		malloc_printf(						\
+		    "<jemalloc>: %s:%d: Unreachable code reached\n",	\
+		    __FILE__, __LINE__);				\
+		abort();						\
+	}								\
+} while (0)
+#endif
+
+#ifndef not_implemented
+#define	not_implemented() do {						\
+	if (config_debug) {						\
+		malloc_printf("<jemalloc>: %s:%d: Not implemented\n",	\
+		    __FILE__, __LINE__);				\
+		abort();						\
+	}								\
+} while (0)
+#endif
+
+#define	assert_not_implemented(e) do {					\
+	if (config_debug && !(e))					\
+		not_implemented();					\
+} while (0)
+
+#endif /* JEMALLOC_H_TYPES */
+/******************************************************************************/
+#ifdef JEMALLOC_H_STRUCTS
+
+#endif /* JEMALLOC_H_STRUCTS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_EXTERNS
+
+int	buferror(char *buf, size_t buflen);
+uintmax_t	malloc_strtoumax(const char *nptr, char **endptr, int base);
+void	malloc_write(const char *s);
+
+/*
+ * malloc_vsnprintf() supports a subset of snprintf(3) that avoids floating
+ * point math.
+ */
+int	malloc_vsnprintf(char *str, size_t size, const char *format,
+    va_list ap);
+int	malloc_snprintf(char *str, size_t size, const char *format, ...)
+    JEMALLOC_ATTR(format(printf, 3, 4));
+void	malloc_vcprintf(void (*write_cb)(void *, const char *), void *cbopaque,
+    const char *format, va_list ap);
+void malloc_cprintf(void (*write)(void *, const char *), void *cbopaque,
+    const char *format, ...) JEMALLOC_ATTR(format(printf, 3, 4));
+void	malloc_printf(const char *format, ...)
+    JEMALLOC_ATTR(format(printf, 1, 2));
+
+#endif /* JEMALLOC_H_EXTERNS */
+/******************************************************************************/
+#ifdef JEMALLOC_H_INLINES
+
+#ifndef JEMALLOC_ENABLE_INLINE
+size_t	pow2_ceil(size_t x);
+void	malloc_write(const char *s);
+void	set_errno(int errnum);
+int	get_errno(void);
+#endif
+
+#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_UTIL_C_))
+/* Compute the smallest power of 2 that is >= x. */
+JEMALLOC_INLINE size_t
+pow2_ceil(size_t x)
+{
+
+	x--;
+	x |= x >> 1;
+	x |= x >> 2;
+	x |= x >> 4;
+	x |= x >> 8;
+	x |= x >> 16;
+#if (LG_SIZEOF_PTR == 3)
+	x |= x >> 32;
+#endif
+	x++;
+	return (x);
+}
+
+/* Sets error code */
+JEMALLOC_INLINE void
+set_errno(int errnum)
+{
+
+#ifdef _WIN32
+	SetLastError(errnum);
+#else
+	errno = errnum;
+#endif
+}
+
+/* Get last error code */
+JEMALLOC_INLINE int
+get_errno(void)
+{
+
+#ifdef _WIN32
+	return (GetLastError());
+#else
+	return (errno);
+#endif
+}
+#endif
+
+#endif /* JEMALLOC_H_INLINES */
+/******************************************************************************/
--- a/include/jemalloc/jemalloc.h.in
+++ b/include/jemalloc/jemalloc.h.in
@@ -0,0 +1,157 @@
+#ifndef JEMALLOC_H_
+#define	JEMALLOC_H_
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <limits.h>
+#include <strings.h>
+
+#define	JEMALLOC_VERSION "@jemalloc_version@"
+#define	JEMALLOC_VERSION_MAJOR @jemalloc_version_major@
+#define	JEMALLOC_VERSION_MINOR @jemalloc_version_minor@
+#define	JEMALLOC_VERSION_BUGFIX @jemalloc_version_bugfix@
+#define	JEMALLOC_VERSION_NREV @jemalloc_version_nrev@
+#define	JEMALLOC_VERSION_GID "@jemalloc_version_gid@"
+
+#include "jemalloc_defs@install_suffix@.h"
+
+#ifdef JEMALLOC_EXPERIMENTAL
+#define	ALLOCM_LG_ALIGN(la)	(la)
+#if LG_SIZEOF_PTR == 2
+#define	ALLOCM_ALIGN(a)	(ffs(a)-1)
+#else
+#define	ALLOCM_ALIGN(a)	((a < (size_t)INT_MAX) ? ffs(a)-1 : ffs(a>>32)+31)
+#endif
+#define	ALLOCM_ZERO	((int)0x40)
+#define	ALLOCM_NO_MOVE	((int)0x80)
+/* Bias arena index bits so that 0 encodes "ALLOCM_ARENA() unspecified". */
+#define	ALLOCM_ARENA(a)	((int)(((a)+1) << 8))
+
+#define	ALLOCM_SUCCESS		0
+#define	ALLOCM_ERR_OOM		1
+#define	ALLOCM_ERR_NOT_MOVED	2
+#endif
+
+/*
+ * The je_ prefix on the following public symbol declarations is an artifact of
+ * namespace management, and should be omitted in application code unless
+ * JEMALLOC_NO_DEMANGLE is defined (see below).
+ */
+extern JEMALLOC_EXPORT const char	*je_malloc_conf;
+extern JEMALLOC_EXPORT void		(*je_malloc_message)(void *cbopaque,
+    const char *s);
+
+JEMALLOC_EXPORT void	*je_malloc(size_t size) JEMALLOC_ATTR(malloc);
+JEMALLOC_EXPORT void	*je_calloc(size_t num, size_t size)
+    JEMALLOC_ATTR(malloc);
+JEMALLOC_EXPORT int	je_posix_memalign(void **memptr, size_t alignment,
+    size_t size) JEMALLOC_ATTR(nonnull(1));
+JEMALLOC_EXPORT void	*je_aligned_alloc(size_t alignment, size_t size)
+    JEMALLOC_ATTR(malloc);
+JEMALLOC_EXPORT void	*je_realloc(void *ptr, size_t size);
+JEMALLOC_EXPORT void	je_free(void *ptr);
+
+#ifdef JEMALLOC_OVERRIDE_MEMALIGN
+JEMALLOC_EXPORT void *	je_memalign(size_t alignment, size_t size)
+    JEMALLOC_ATTR(malloc);
+#endif
+
+#ifdef JEMALLOC_OVERRIDE_VALLOC
+JEMALLOC_EXPORT void *	je_valloc(size_t size) JEMALLOC_ATTR(malloc);
+#endif
+
+JEMALLOC_EXPORT size_t	je_malloc_usable_size(
+    JEMALLOC_USABLE_SIZE_CONST void *ptr);
+JEMALLOC_EXPORT void	je_malloc_stats_print(void (*write_cb)(void *,
+    const char *), void *je_cbopaque, const char *opts);
+JEMALLOC_EXPORT int	je_mallctl(const char *name, void *oldp,
+    size_t *oldlenp, void *newp, size_t newlen);
+JEMALLOC_EXPORT int	je_mallctlnametomib(const char *name, size_t *mibp,
+    size_t *miblenp);
+JEMALLOC_EXPORT int	je_mallctlbymib(const size_t *mib, size_t miblen,
+    void *oldp, size_t *oldlenp, void *newp, size_t newlen);
+
+#ifdef JEMALLOC_EXPERIMENTAL
+JEMALLOC_EXPORT int	je_allocm(void **ptr, size_t *rsize, size_t size,
+    int flags) JEMALLOC_ATTR(nonnull(1));
+JEMALLOC_EXPORT int	je_rallocm(void **ptr, size_t *rsize, size_t size,
+    size_t extra, int flags) JEMALLOC_ATTR(nonnull(1));
+JEMALLOC_EXPORT int	je_sallocm(const void *ptr, size_t *rsize, int flags)
+    JEMALLOC_ATTR(nonnull(1));
+JEMALLOC_EXPORT int	je_dallocm(void *ptr, int flags)
+    JEMALLOC_ATTR(nonnull(1));
+JEMALLOC_EXPORT int	je_nallocm(size_t *rsize, size_t size, int flags);
+#endif
+
+/*
+ * By default application code must explicitly refer to mangled symbol names,
+ * so that it is possible to use jemalloc in conjunction with another allocator
+ * in the same application.  Define JEMALLOC_MANGLE in order to cause automatic
+ * name mangling that matches the API prefixing that happened as a result of
+ * --with-mangling and/or --with-jemalloc-prefix configuration settings.
+ */
+#ifdef JEMALLOC_MANGLE
+#ifndef JEMALLOC_NO_DEMANGLE
+#define	JEMALLOC_NO_DEMANGLE
+#endif
+#define	malloc_conf je_malloc_conf
+#define	malloc_message je_malloc_message
+#define	malloc je_malloc
+#define	calloc je_calloc
+#define	posix_memalign je_posix_memalign
+#define	aligned_alloc je_aligned_alloc
+#define	realloc je_realloc
+#define	free je_free
+#define	malloc_usable_size je_malloc_usable_size
+#define	malloc_stats_print je_malloc_stats_print
+#define	mallctl je_mallctl
+#define	mallctlnametomib je_mallctlnametomib
+#define	mallctlbymib je_mallctlbymib
+#define	memalign je_memalign
+#define	valloc je_valloc
+#ifdef JEMALLOC_EXPERIMENTAL
+#define	allocm je_allocm
+#define	rallocm je_rallocm
+#define	sallocm je_sallocm
+#define	dallocm je_dallocm
+#define	nallocm je_nallocm
+#endif
+#endif
+
+/*
+ * The je_* macros can be used as stable alternative names for the public
+ * jemalloc API if JEMALLOC_NO_DEMANGLE is defined.  This is primarily meant
+ * for use in jemalloc itself, but it can be used by application code to
+ * provide isolation from the name mangling specified via --with-mangling
+ * and/or --with-jemalloc-prefix.
+ */
+#ifndef JEMALLOC_NO_DEMANGLE
+#undef je_malloc_conf
+#undef je_malloc_message
+#undef je_malloc
+#undef je_calloc
+#undef je_posix_memalign
+#undef je_aligned_alloc
+#undef je_realloc
+#undef je_free
+#undef je_malloc_usable_size
+#undef je_malloc_stats_print
+#undef je_mallctl
+#undef je_mallctlnametomib
+#undef je_mallctlbymib
+#undef je_memalign
+#undef je_valloc
+#ifdef JEMALLOC_EXPERIMENTAL
+#undef je_allocm
+#undef je_rallocm
+#undef je_sallocm
+#undef je_dallocm
+#undef je_nallocm
+#endif
+#endif
+
+#ifdef __cplusplus
+};
+#endif
+#endif /* JEMALLOC_H_ */
--- a/include/jemalloc/jemalloc_defs.h.in
+++ b/include/jemalloc/jemalloc_defs.h.in
@@ -0,0 +1,262 @@
+/*
+ * If JEMALLOC_PREFIX is defined via --with-jemalloc-prefix, it will cause all
+ * public APIs to be prefixed.  This makes it possible, with some care, to use
+ * multiple allocators simultaneously.
+ */
+#undef JEMALLOC_PREFIX
+#undef JEMALLOC_CPREFIX
+
+/*
+ * Name mangling for public symbols is controlled by --with-mangling and
+ * --with-jemalloc-prefix.  With default settings the je_ prefix is stripped by
+ * these macro definitions.
+ */
+#undef je_malloc_conf
+#undef je_malloc_message
+#undef je_malloc
+#undef je_calloc
+#undef je_posix_memalign
+#undef je_aligned_alloc
+#undef je_realloc
+#undef je_free
+#undef je_malloc_usable_size
+#undef je_malloc_stats_print
+#undef je_mallctl
+#undef je_mallctlnametomib
+#undef je_mallctlbymib
+#undef je_memalign
+#undef je_valloc
+#undef je_allocm
+#undef je_rallocm
+#undef je_sallocm
+#undef je_dallocm
+#undef je_nallocm
+
+/*
+ * JEMALLOC_PRIVATE_NAMESPACE is used as a prefix for all library-private APIs.
+ * For shared libraries, symbol visibility mechanisms prevent these symbols
+ * from being exported, but for static libraries, naming collisions are a real
+ * possibility.
+ */
+#undef JEMALLOC_PRIVATE_NAMESPACE
+#undef JEMALLOC_N
+
+/*
+ * Hyper-threaded CPUs may need a special instruction inside spin loops in
+ * order to yield to another virtual CPU.
+ */
+#undef CPU_SPINWAIT
+
+/* Defined if the equivalent of FreeBSD's atomic(9) functions are available. */
+#undef JEMALLOC_ATOMIC9
+
+/*
+ * Defined if OSAtomic*() functions are available, as provided by Darwin, and
+ * documented in the atomic(3) manual page.
+ */
+#undef JEMALLOC_OSATOMIC
+
+/*
+ * Defined if __sync_add_and_fetch(uint32_t *, uint32_t) and
+ * __sync_sub_and_fetch(uint32_t *, uint32_t) are available, despite
+ * __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4 not being defined (which means the
+ * functions are defined in libgcc instead of being inlines)
+ */
+#undef JE_FORCE_SYNC_COMPARE_AND_SWAP_4
+
+/*
+ * Defined if __sync_add_and_fetch(uint64_t *, uint64_t) and
+ * __sync_sub_and_fetch(uint64_t *, uint64_t) are available, despite
+ * __GCC_HAVE_SYNC_COMPARE_AND_SWAP_8 not being defined (which means the
+ * functions are defined in libgcc instead of being inlines)
+ */
+#undef JE_FORCE_SYNC_COMPARE_AND_SWAP_8
+
+/*
+ * Defined if OSSpin*() functions are available, as provided by Darwin, and
+ * documented in the spinlock(3) manual page.
+ */
+#undef JEMALLOC_OSSPIN
+
+/*
+ * Defined if _malloc_thread_cleanup() exists.  At least in the case of
+ * FreeBSD, pthread_key_create() allocates, which if used during malloc
+ * bootstrapping will cause recursion into the pthreads library.  Therefore, if
+ * _malloc_thread_cleanup() exists, use it as the basis for thread cleanup in
+ * malloc_tsd.
+ */
+#undef JEMALLOC_MALLOC_THREAD_CLEANUP
+
+/*
+ * Defined if threaded initialization is known to be safe on this platform.
+ * Among other things, it must be possible to initialize a mutex without
+ * triggering allocation in order for threaded allocation to be safe.
+ */
+#undef JEMALLOC_THREADED_INIT
+
+/*
+ * Defined if the pthreads implementation defines
+ * _pthread_mutex_init_calloc_cb(), in which case the function is used in order
+ * to avoid recursive allocation during mutex initialization.
+ */
+#undef JEMALLOC_MUTEX_INIT_CB
+
+/* Defined if __attribute__((...)) syntax is supported. */
+#undef JEMALLOC_HAVE_ATTR
+#ifdef JEMALLOC_HAVE_ATTR
+#  define JEMALLOC_ATTR(s) __attribute__((s))
+#  define JEMALLOC_EXPORT JEMALLOC_ATTR(visibility("default"))
+#  define JEMALLOC_ALIGNED(s) JEMALLOC_ATTR(aligned(s))
+#  define JEMALLOC_SECTION(s) JEMALLOC_ATTR(section(s))
+#  define JEMALLOC_NOINLINE JEMALLOC_ATTR(noinline)
+#elif _MSC_VER
+#  define JEMALLOC_ATTR(s)
+#  ifdef DLLEXPORT
+#    define JEMALLOC_EXPORT __declspec(dllexport)
+#  else
+#    define JEMALLOC_EXPORT __declspec(dllimport)
+#  endif
+#  define JEMALLOC_ALIGNED(s) __declspec(align(s))
+#  define JEMALLOC_SECTION(s) __declspec(allocate(s))
+#  define JEMALLOC_NOINLINE __declspec(noinline)
+#else
+#  define JEMALLOC_ATTR(s)
+#  define JEMALLOC_EXPORT
+#  define JEMALLOC_ALIGNED(s)
+#  define JEMALLOC_SECTION(s)
+#  define JEMALLOC_NOINLINE
+#endif
+
+/* Defined if sbrk() is supported. */
+#undef JEMALLOC_HAVE_SBRK
+
+/* Non-empty if the tls_model attribute is supported. */
+#undef JEMALLOC_TLS_MODEL
+
+/* JEMALLOC_CC_SILENCE enables code that silences unuseful compiler warnings. */
+#undef JEMALLOC_CC_SILENCE
+
+/*
+ * JEMALLOC_DEBUG enables assertions and other sanity checks, and disables
+ * inline functions.
+ */
+#undef JEMALLOC_DEBUG
+
+/* JEMALLOC_STATS enables statistics calculation. */
+#undef JEMALLOC_STATS
+
+/* JEMALLOC_PROF enables allocation profiling. */
+#undef JEMALLOC_PROF
+
+/* Use libunwind for profile backtracing if defined. */
+#undef JEMALLOC_PROF_LIBUNWIND
+
+/* Use libgcc for profile backtracing if defined. */
+#undef JEMALLOC_PROF_LIBGCC
+
+/* Use gcc intrinsics for profile backtracing if defined. */
+#undef JEMALLOC_PROF_GCC
+
+/*
+ * JEMALLOC_TCACHE enables a thread-specific caching layer for small objects.
+ * This makes it possible to allocate/deallocate objects without any locking
+ * when the cache is in the steady state.
+ */
+#undef JEMALLOC_TCACHE
+
+/*
+ * JEMALLOC_DSS enables use of sbrk(2) to allocate chunks from the data storage
+ * segment (DSS).
+ */
+#undef JEMALLOC_DSS
+
+/* Support memory filling (junk/zero/quarantine/redzone). */
+#undef JEMALLOC_FILL
+
+/* Support the experimental API. */
+#undef JEMALLOC_EXPERIMENTAL
+
+/* Support utrace(2)-based tracing. */
+#undef JEMALLOC_UTRACE
+
+/* Support Valgrind. */
+#undef JEMALLOC_VALGRIND
+
+/* Support optional abort() on OOM. */
+#undef JEMALLOC_XMALLOC
+
+/* Support lazy locking (avoid locking unless a second thread is launched). */
+#undef JEMALLOC_LAZY_LOCK
+
+/* One page is 2^STATIC_PAGE_SHIFT bytes. */
+#undef STATIC_PAGE_SHIFT
+
+/*
+ * If defined, use munmap() to unmap freed chunks, rather than storing them for
+ * later reuse.  This is disabled by default on Linux because common sequences
+ * of mmap()/munmap() calls will cause virtual memory map holes.
+ */
+#undef JEMALLOC_MUNMAP
+
+/*
+ * If defined, use mremap(...MREMAP_FIXED...) for huge realloc().  This is
+ * disabled by default because it is Linux-specific and it will cause virtual
+ * memory map holes, much like munmap(2) does.
+ */
+#undef JEMALLOC_MREMAP
+
+/* TLS is used to map arenas and magazine caches to threads. */
+#undef JEMALLOC_TLS
+
+/*
+ * JEMALLOC_IVSALLOC enables ivsalloc(), which verifies that pointers reside
+ * within jemalloc-owned chunks before dereferencing them.
+ */
+#undef JEMALLOC_IVSALLOC
+
+/*
+ * Define overrides for non-standard allocator-related functions if they
+ * are present on the system.
+ */
+#undef JEMALLOC_OVERRIDE_MEMALIGN
+#undef JEMALLOC_OVERRIDE_VALLOC
+
+/*
+ * At least Linux omits the "const" in:
+ *
+ *   size_t malloc_usable_size(const void *ptr);
+ *
+ * Match the operating system's prototype.
+ */
+#undef JEMALLOC_USABLE_SIZE_CONST
+
+/*
+ * Darwin (OS X) uses zones to work around Mach-O symbol override shortcomings.
+ */
+#undef JEMALLOC_ZONE
+#undef JEMALLOC_ZONE_VERSION
+
+/*
+ * Methods for purging unused pages differ between operating systems.
+ *
+ *   madvise(..., MADV_DONTNEED) : On Linux, this immediately discards pages,
+ *                                 such that new pages will be demand-zeroed if
+ *                                 the address region is later touched.
+ *   madvise(..., MADV_FREE) : On FreeBSD and Darwin, this marks pages as being
+ *                             unused, such that they will be discarded rather
+ *                             than swapped out.
+ */
+#undef JEMALLOC_PURGE_MADVISE_DONTNEED
+#undef JEMALLOC_PURGE_MADVISE_FREE
+
+/* sizeof(void *) == 2^LG_SIZEOF_PTR. */
+#undef LG_SIZEOF_PTR
+
+/* sizeof(int) == 2^LG_SIZEOF_INT. */
+#undef LG_SIZEOF_INT
+
+/* sizeof(long) == 2^LG_SIZEOF_LONG. */
+#undef LG_SIZEOF_LONG
+
+/* sizeof(intmax_t) == 2^LG_SIZEOF_INTMAX_T. */
+#undef LG_SIZEOF_INTMAX_T
--- a/include/msvc_compat/inttypes.h
+++ b/include/msvc_compat/inttypes.h
@@ -0,0 +1,313 @@
+// ISO C9x  compliant inttypes.h for Microsoft Visual Studio
+// Based on ISO/IEC 9899:TC2 Committee draft (May 6, 2005) WG14/N1124 
+// 
+//  Copyright (c) 2006 Alexander Chemeris
+// 
+// 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. The name of the author may be used to endorse or promote products
+//      derived from this software without specific prior written permission.
+// 
+// THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR 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.
+// 
+///////////////////////////////////////////////////////////////////////////////
+
+#ifndef _MSC_VER // [
+#error "Use this header only with Microsoft Visual C++ compilers!"
+#endif // _MSC_VER ]
+
+#ifndef _MSC_INTTYPES_H_ // [
+#define _MSC_INTTYPES_H_
+
+#if _MSC_VER > 1000
+#pragma once
+#endif
+
+#include "stdint.h"
+
+// 7.8 Format conversion of integer types
+
+typedef struct {
+   intmax_t quot;
+   intmax_t rem;
+} imaxdiv_t;
+
+// 7.8.1 Macros for format specifiers
+
+#if !defined(__cplusplus) || defined(__STDC_FORMAT_MACROS) // [   See footnote 185 at page 198
+
+#ifdef _WIN64
+#  define __PRI64_PREFIX        "l"
+#  define __PRIPTR_PREFIX       "l"
+#else
+#  define __PRI64_PREFIX        "ll"
+#  define __PRIPTR_PREFIX
+#endif
+
+// The fprintf macros for signed integers are:
+#define PRId8       "d"
+#define PRIi8       "i"
+#define PRIdLEAST8  "d"
+#define PRIiLEAST8  "i"
+#define PRIdFAST8   "d"
+#define PRIiFAST8   "i"
+
+#define PRId16       "hd"
+#define PRIi16       "hi"
+#define PRIdLEAST16  "hd"
+#define PRIiLEAST16  "hi"
+#define PRIdFAST16   "hd"
+#define PRIiFAST16   "hi"
+
+#define PRId32       "d"
+#define PRIi32       "i"
+#define PRIdLEAST32  "d"
+#define PRIiLEAST32  "i"
+#define PRIdFAST32   "d"
+#define PRIiFAST32   "i"
+
+#define PRId64       __PRI64_PREFIX "d"
+#define PRIi64       __PRI64_PREFIX "i"
+#define PRIdLEAST64  __PRI64_PREFIX "d"
+#define PRIiLEAST64  __PRI64_PREFIX "i"
+#define PRIdFAST64   __PRI64_PREFIX "d"
+#define PRIiFAST64   __PRI64_PREFIX "i"
+
+#define PRIdMAX     __PRI64_PREFIX "d"
+#define PRIiMAX     __PRI64_PREFIX "i"
+
+#define PRIdPTR     __PRIPTR_PREFIX "d"
+#define PRIiPTR     __PRIPTR_PREFIX "i"
+
+// The fprintf macros for unsigned integers are:
+#define PRIo8       "o"
+#define PRIu8       "u"
+#define PRIx8       "x"
+#define PRIX8       "X"
+#define PRIoLEAST8  "o"
+#define PRIuLEAST8  "u"
+#define PRIxLEAST8  "x"
+#define PRIXLEAST8  "X"
+#define PRIoFAST8   "o"
+#define PRIuFAST8   "u"
+#define PRIxFAST8   "x"
+#define PRIXFAST8   "X"
+
+#define PRIo16       "ho"
+#define PRIu16       "hu"
+#define PRIx16       "hx"
+#define PRIX16       "hX"
+#define PRIoLEAST16  "ho"
+#define PRIuLEAST16  "hu"
+#define PRIxLEAST16  "hx"
+#define PRIXLEAST16  "hX"
+#define PRIoFAST16   "ho"
+#define PRIuFAST16   "hu"
+#define PRIxFAST16   "hx"
+#define PRIXFAST16   "hX"
+
+#define PRIo32       "o"
+#define PRIu32       "u"
+#define PRIx32       "x"
+#define PRIX32       "X"
+#define PRIoLEAST32  "o"
+#define PRIuLEAST32  "u"
+#define PRIxLEAST32  "x"
+#define PRIXLEAST32  "X"
+#define PRIoFAST32   "o"
+#define PRIuFAST32   "u"
+#define PRIxFAST32   "x"
+#define PRIXFAST32   "X"
+
+#define PRIo64       __PRI64_PREFIX "o"
+#define PRIu64       __PRI64_PREFIX "u"
+#define PRIx64       __PRI64_PREFIX "x"
+#define PRIX64       __PRI64_PREFIX "X"
+#define PRIoLEAST64  __PRI64_PREFIX "o"
+#define PRIuLEAST64  __PRI64_PREFIX "u"
+#define PRIxLEAST64  __PRI64_PREFIX "x"
+#define PRIXLEAST64  __PRI64_PREFIX "X"
+#define PRIoFAST64   __PRI64_PREFIX "o"
+#define PRIuFAST64   __PRI64_PREFIX "u"
+#define PRIxFAST64   __PRI64_PREFIX "x"
+#define PRIXFAST64   __PRI64_PREFIX "X"
+
+#define PRIoMAX     __PRI64_PREFIX "o"
+#define PRIuMAX     __PRI64_PREFIX "u"
+#define PRIxMAX     __PRI64_PREFIX "x"
+#define PRIXMAX     __PRI64_PREFIX "X"
+
+#define PRIoPTR     __PRIPTR_PREFIX "o"
+#define PRIuPTR     __PRIPTR_PREFIX "u"
+#define PRIxPTR     __PRIPTR_PREFIX "x"
+#define PRIXPTR     __PRIPTR_PREFIX "X"
+
+// The fscanf macros for signed integers are:
+#define SCNd8       "d"
+#define SCNi8       "i"
+#define SCNdLEAST8  "d"
+#define SCNiLEAST8  "i"
+#define SCNdFAST8   "d"
+#define SCNiFAST8   "i"
+
+#define SCNd16       "hd"
+#define SCNi16       "hi"
+#define SCNdLEAST16  "hd"
+#define SCNiLEAST16  "hi"
+#define SCNdFAST16   "hd"
+#define SCNiFAST16   "hi"
+
+#define SCNd32       "ld"
+#define SCNi32       "li"
+#define SCNdLEAST32  "ld"
+#define SCNiLEAST32  "li"
+#define SCNdFAST32   "ld"
+#define SCNiFAST32   "li"
+
+#define SCNd64       "I64d"
+#define SCNi64       "I64i"
+#define SCNdLEAST64  "I64d"
+#define SCNiLEAST64  "I64i"
+#define SCNdFAST64   "I64d"
+#define SCNiFAST64   "I64i"
+
+#define SCNdMAX     "I64d"
+#define SCNiMAX     "I64i"
+
+#ifdef _WIN64 // [
+#  define SCNdPTR     "I64d"
+#  define SCNiPTR     "I64i"
+#else  // _WIN64 ][
+#  define SCNdPTR     "ld"
+#  define SCNiPTR     "li"
+#endif  // _WIN64 ]
+
+// The fscanf macros for unsigned integers are:
+#define SCNo8       "o"
+#define SCNu8       "u"
+#define SCNx8       "x"
+#define SCNX8       "X"
+#define SCNoLEAST8  "o"
+#define SCNuLEAST8  "u"
+#define SCNxLEAST8  "x"
+#define SCNXLEAST8  "X"
+#define SCNoFAST8   "o"
+#define SCNuFAST8   "u"
+#define SCNxFAST8   "x"
+#define SCNXFAST8   "X"
+
+#define SCNo16       "ho"
+#define SCNu16       "hu"
+#define SCNx16       "hx"
+#define SCNX16       "hX"
+#define SCNoLEAST16  "ho"
+#define SCNuLEAST16  "hu"
+#define SCNxLEAST16  "hx"
+#define SCNXLEAST16  "hX"
+#define SCNoFAST16   "ho"
+#define SCNuFAST16   "hu"
+#define SCNxFAST16   "hx"
+#define SCNXFAST16   "hX"
+
+#define SCNo32       "lo"
+#define SCNu32       "lu"
+#define SCNx32       "lx"
+#define SCNX32       "lX"
+#define SCNoLEAST32  "lo"
+#define SCNuLEAST32  "lu"
+#define SCNxLEAST32  "lx"
+#define SCNXLEAST32  "lX"
+#define SCNoFAST32   "lo"
+#define SCNuFAST32   "lu"
+#define SCNxFAST32   "lx"
+#define SCNXFAST32   "lX"
+
+#define SCNo64       "I64o"
+#define SCNu64       "I64u"
+#define SCNx64       "I64x"
+#define SCNX64       "I64X"
+#define SCNoLEAST64  "I64o"
+#define SCNuLEAST64  "I64u"
+#define SCNxLEAST64  "I64x"
+#define SCNXLEAST64  "I64X"
+#define SCNoFAST64   "I64o"
+#define SCNuFAST64   "I64u"
+#define SCNxFAST64   "I64x"
+#define SCNXFAST64   "I64X"
+
+#define SCNoMAX     "I64o"
+#define SCNuMAX     "I64u"
+#define SCNxMAX     "I64x"
+#define SCNXMAX     "I64X"
+
+#ifdef _WIN64 // [
+#  define SCNoPTR     "I64o"
+#  define SCNuPTR     "I64u"
+#  define SCNxPTR     "I64x"
+#  define SCNXPTR     "I64X"
+#else  // _WIN64 ][
+#  define SCNoPTR     "lo"
+#  define SCNuPTR     "lu"
+#  define SCNxPTR     "lx"
+#  define SCNXPTR     "lX"
+#endif  // _WIN64 ]
+
+#endif // __STDC_FORMAT_MACROS ]
+
+// 7.8.2 Functions for greatest-width integer types
+
+// 7.8.2.1 The imaxabs function
+#define imaxabs _abs64
+
+// 7.8.2.2 The imaxdiv function
+
+// This is modified version of div() function from Microsoft's div.c found
+// in %MSVC.NET%\crt\src\div.c
+#ifdef STATIC_IMAXDIV // [
+static
+#else // STATIC_IMAXDIV ][
+_inline
+#endif // STATIC_IMAXDIV ]
+imaxdiv_t __cdecl imaxdiv(intmax_t numer, intmax_t denom)
+{
+   imaxdiv_t result;
+
+   result.quot = numer / denom;
+   result.rem = numer % denom;
+
+   if (numer < 0 && result.rem > 0) {
+      // did division wrong; must fix up
+      ++result.quot;
+      result.rem -= denom;
+   }
+
+   return result;
+}
+
+// 7.8.2.3 The strtoimax and strtoumax functions
+#define strtoimax _strtoi64
+#define strtoumax _strtoui64
+
+// 7.8.2.4 The wcstoimax and wcstoumax functions
+#define wcstoimax _wcstoi64
+#define wcstoumax _wcstoui64
+
+
+#endif // _MSC_INTTYPES_H_ ]
--- a/include/msvc_compat/stdbool.h
+++ b/include/msvc_compat/stdbool.h
@@ -0,0 +1,16 @@
+#ifndef stdbool_h
+#define stdbool_h
+
+#include <wtypes.h>
+
+/* MSVC doesn't define _Bool or bool in C, but does have BOOL */
+/* Note this doesn't pass autoconf's test because (bool) 0.5 != true */
+typedef BOOL _Bool;
+
+#define bool _Bool
+#define true 1
+#define false 0
+
+#define __bool_true_false_are_defined 1
+
+#endif /* stdbool_h */
--- a/include/msvc_compat/stdint.h
+++ b/include/msvc_compat/stdint.h
@@ -0,0 +1,247 @@
+// ISO C9x  compliant stdint.h for Microsoft Visual Studio
+// Based on ISO/IEC 9899:TC2 Committee draft (May 6, 2005) WG14/N1124 
+// 
+//  Copyright (c) 2006-2008 Alexander Chemeris
+// 
+// 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. The name of the author may be used to endorse or promote products
+//      derived from this software without specific prior written permission.
+// 
+// THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR 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.
+// 
+///////////////////////////////////////////////////////////////////////////////
+
+#ifndef _MSC_VER // [
+#error "Use this header only with Microsoft Visual C++ compilers!"
+#endif // _MSC_VER ]
+
+#ifndef _MSC_STDINT_H_ // [
+#define _MSC_STDINT_H_
+
+#if _MSC_VER > 1000
+#pragma once
+#endif
+
+#include <limits.h>
+
+// For Visual Studio 6 in C++ mode and for many Visual Studio versions when
+// compiling for ARM we should wrap <wchar.h> include with 'extern "C++" {}'
+// or compiler give many errors like this:
+//   error C2733: second C linkage of overloaded function 'wmemchr' not allowed
+#ifdef __cplusplus
+extern "C" {
+#endif
+#  include <wchar.h>
+#ifdef __cplusplus
+}
+#endif
+
+// Define _W64 macros to mark types changing their size, like intptr_t.
+#ifndef _W64
+#  if !defined(__midl) && (defined(_X86_) || defined(_M_IX86)) && _MSC_VER >= 1300
+#     define _W64 __w64
+#  else
+#     define _W64
+#  endif
+#endif
+
+
+// 7.18.1 Integer types
+
+// 7.18.1.1 Exact-width integer types
+
+// Visual Studio 6 and Embedded Visual C++ 4 doesn't
+// realize that, e.g. char has the same size as __int8
+// so we give up on __intX for them.
+#if (_MSC_VER < 1300)
+   typedef signed char       int8_t;
+   typedef signed short      int16_t;
+   typedef signed int        int32_t;
+   typedef unsigned char     uint8_t;
+   typedef unsigned short    uint16_t;
+   typedef unsigned int      uint32_t;
+#else
+   typedef signed __int8     int8_t;
+   typedef signed __int16    int16_t;
+   typedef signed __int32    int32_t;
+   typedef unsigned __int8   uint8_t;
+   typedef unsigned __int16  uint16_t;
+   typedef unsigned __int32  uint32_t;
+#endif
+typedef signed __int64       int64_t;
+typedef unsigned __int64     uint64_t;
+
+
+// 7.18.1.2 Minimum-width integer types
+typedef int8_t    int_least8_t;
+typedef int16_t   int_least16_t;
+typedef int32_t   int_least32_t;
+typedef int64_t   int_least64_t;
+typedef uint8_t   uint_least8_t;
+typedef uint16_t  uint_least16_t;
+typedef uint32_t  uint_least32_t;
+typedef uint64_t  uint_least64_t;
+
+// 7.18.1.3 Fastest minimum-width integer types
+typedef int8_t    int_fast8_t;
+typedef int16_t   int_fast16_t;
+typedef int32_t   int_fast32_t;
+typedef int64_t   int_fast64_t;
+typedef uint8_t   uint_fast8_t;
+typedef uint16_t  uint_fast16_t;
+typedef uint32_t  uint_fast32_t;
+typedef uint64_t  uint_fast64_t;
+
+// 7.18.1.4 Integer types capable of holding object pointers
+#ifdef _WIN64 // [
+   typedef signed __int64    intptr_t;
+   typedef unsigned __int64  uintptr_t;
+#else // _WIN64 ][
+   typedef _W64 signed int   intptr_t;
+   typedef _W64 unsigned int uintptr_t;
+#endif // _WIN64 ]
+
+// 7.18.1.5 Greatest-width integer types
+typedef int64_t   intmax_t;
+typedef uint64_t  uintmax_t;
+
+
+// 7.18.2 Limits of specified-width integer types
+
+#if !defined(__cplusplus) || defined(__STDC_LIMIT_MACROS) // [   See footnote 220 at page 257 and footnote 221 at page 259
+
+// 7.18.2.1 Limits of exact-width integer types
+#define INT8_MIN     ((int8_t)_I8_MIN)
+#define INT8_MAX     _I8_MAX
+#define INT16_MIN    ((int16_t)_I16_MIN)
+#define INT16_MAX    _I16_MAX
+#define INT32_MIN    ((int32_t)_I32_MIN)
+#define INT32_MAX    _I32_MAX
+#define INT64_MIN    ((int64_t)_I64_MIN)
+#define INT64_MAX    _I64_MAX
+#define UINT8_MAX    _UI8_MAX
+#define UINT16_MAX   _UI16_MAX
+#define UINT32_MAX   _UI32_MAX
+#define UINT64_MAX   _UI64_MAX
+
+// 7.18.2.2 Limits of minimum-width integer types
+#define INT_LEAST8_MIN    INT8_MIN
+#define INT_LEAST8_MAX    INT8_MAX
+#define INT_LEAST16_MIN   INT16_MIN
+#define INT_LEAST16_MAX   INT16_MAX
+#define INT_LEAST32_MIN   INT32_MIN
+#define INT_LEAST32_MAX   INT32_MAX
+#define INT_LEAST64_MIN   INT64_MIN
+#define INT_LEAST64_MAX   INT64_MAX
+#define UINT_LEAST8_MAX   UINT8_MAX
+#define UINT_LEAST16_MAX  UINT16_MAX
+#define UINT_LEAST32_MAX  UINT32_MAX
+#define UINT_LEAST64_MAX  UINT64_MAX
+
+// 7.18.2.3 Limits of fastest minimum-width integer types
+#define INT_FAST8_MIN    INT8_MIN
+#define INT_FAST8_MAX    INT8_MAX
+#define INT_FAST16_MIN   INT16_MIN
+#define INT_FAST16_MAX   INT16_MAX
+#define INT_FAST32_MIN   INT32_MIN
+#define INT_FAST32_MAX   INT32_MAX
+#define INT_FAST64_MIN   INT64_MIN
+#define INT_FAST64_MAX   INT64_MAX
+#define UINT_FAST8_MAX   UINT8_MAX
+#define UINT_FAST16_MAX  UINT16_MAX
+#define UINT_FAST32_MAX  UINT32_MAX
+#define UINT_FAST64_MAX  UINT64_MAX
+
+// 7.18.2.4 Limits of integer types capable of holding object pointers
+#ifdef _WIN64 // [
+#  define INTPTR_MIN   INT64_MIN
+#  define INTPTR_MAX   INT64_MAX
+#  define UINTPTR_MAX  UINT64_MAX
+#else // _WIN64 ][
+#  define INTPTR_MIN   INT32_MIN
+#  define INTPTR_MAX   INT32_MAX
+#  define UINTPTR_MAX  UINT32_MAX
+#endif // _WIN64 ]
+
+// 7.18.2.5 Limits of greatest-width integer types
+#define INTMAX_MIN   INT64_MIN
+#define INTMAX_MAX   INT64_MAX
+#define UINTMAX_MAX  UINT64_MAX
+
+// 7.18.3 Limits of other integer types
+
+#ifdef _WIN64 // [
+#  define PTRDIFF_MIN  _I64_MIN
+#  define PTRDIFF_MAX  _I64_MAX
+#else  // _WIN64 ][
+#  define PTRDIFF_MIN  _I32_MIN
+#  define PTRDIFF_MAX  _I32_MAX
+#endif  // _WIN64 ]
+
+#define SIG_ATOMIC_MIN  INT_MIN
+#define SIG_ATOMIC_MAX  INT_MAX
+
+#ifndef SIZE_MAX // [
+#  ifdef _WIN64 // [
+#     define SIZE_MAX  _UI64_MAX
+#  else // _WIN64 ][
+#     define SIZE_MAX  _UI32_MAX
+#  endif // _WIN64 ]
+#endif // SIZE_MAX ]
+
+// WCHAR_MIN and WCHAR_MAX are also defined in <wchar.h>
+#ifndef WCHAR_MIN // [
+#  define WCHAR_MIN  0
+#endif  // WCHAR_MIN ]
+#ifndef WCHAR_MAX // [
+#  define WCHAR_MAX  _UI16_MAX
+#endif  // WCHAR_MAX ]
+
+#define WINT_MIN  0
+#define WINT_MAX  _UI16_MAX
+
+#endif // __STDC_LIMIT_MACROS ]
+
+
+// 7.18.4 Limits of other integer types
+
+#if !defined(__cplusplus) || defined(__STDC_CONSTANT_MACROS) // [   See footnote 224 at page 260
+
+// 7.18.4.1 Macros for minimum-width integer constants
+
+#define INT8_C(val)  val##i8
+#define INT16_C(val) val##i16
+#define INT32_C(val) val##i32
+#define INT64_C(val) val##i64
+
+#define UINT8_C(val)  val##ui8
+#define UINT16_C(val) val##ui16
+#define UINT32_C(val) val##ui32
+#define UINT64_C(val) val##ui64
+
+// 7.18.4.2 Macros for greatest-width integer constants
+#define INTMAX_C   INT64_C
+#define UINTMAX_C  UINT64_C
+
+#endif // __STDC_CONSTANT_MACROS ]
+
+
+#endif // _MSC_STDINT_H_ ]
--- a/include/msvc_compat/strings.h
+++ b/include/msvc_compat/strings.h
@@ -0,0 +1,23 @@
+#ifndef strings_h
+#define strings_h
+
+/* MSVC doesn't define ffs/ffsl. This dummy strings.h header is provided
+ * for both */
+#include <intrin.h>
+#pragma intrinsic(_BitScanForward)
+static __forceinline int ffsl(long x)
+{
+	unsigned long i;
+
+	if (_BitScanForward(&i, x))
+		return (i + 1);
+	return (0);
+}
+
+static __forceinline int ffs(int x)
+{
+
+	return (ffsl(x));
+}
+
+#endif
--- a/jemalloc/install-sh
+++ b/jemalloc/install-sh
--- a/jemalloc/COPYING
+++ b/jemalloc/COPYING
@@ -1,51 +0,0 @@
-Unless otherwise specified, files in the jemalloc source distribution are
-subject to the following licenses:
--------------------------------------------------------------------------------
-Copyright (C) 2002-2010 Jason Evans <jasone@canonware.com>.
-All rights reserved.
-Copyright (C) 2007-2010 Mozilla Foundation.  All rights reserved.
-
-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(s),
-   this list of conditions and the following disclaimer.
-2. Redistributions in binary form must reproduce the above copyright notice(s),
-   this list of conditions and the following disclaimer in the documentation
-   and/or other materials provided with the distribution.
-
-THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``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 COPYRIGHT HOLDER(S) 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.
--------------------------------------------------------------------------------
-Copyright (C) 2009-2010 Facebook, Inc.
-All rights reserved.
-
-Redistribution and use in source and binary forms, with or without modification,
-are permitted provided that the following conditions are met:
-* Redistributions of source code must retain the above copyright notice, this
-  list of conditions and the following disclaimer.
-* 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.
-* Neither the name of Facebook, Inc. 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 COPYRIGHT HOLDERS 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 COPYRIGHT OWNER 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.
--------------------------------------------------------------------------------
--- a/jemalloc/ChangeLog
+++ b/jemalloc/ChangeLog
@@ -1,169 +0,0 @@
-Following are change highlights associated with official releases.  Important
-bug fixes are all mentioned, but internal enhancements are omitted here for
-brevity (even though they are more fun to write about).  Much more detail can be
-found in the git revision history:
-
-    http://www.canonware.com/cgi-bin/gitweb.cgi?p=jemalloc.git
-    git://canonware.com/jemalloc.git
-
-* 2.1.2 (March 2, 2011)
-
-  Bug fixes:
-  - Fix "thread.{de,}allocatedp" mallctl for OS X.
-  - Add missing jemalloc.a to build system.
-
-* 2.1.1 (January 31, 2011)
-
-  Bug fixes:
-  - Fix aligned huge reallocation (affected allocm()).
-  - Fix the ALLOCM_LG_ALIGN macro definition.
-  - Fix a heap dumping deadlock.
-  - Fix a "thread.arena" mallctl bug.
-
-* 2.1.0 (December 3, 2010)
-
-  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 (October 29, 2010)
-
-  Bug fixes:
-  - Fix a race condition in heap profiling that could cause undefined behavior
-    if "opt.prof_accum" were disabled.
-  - Add missing mutex unlocks for some OOM error paths in the heap profiling
-    code.
-  - Fix a compilation error for non-C99 builds.
-
-* 2.0.0 (October 24, 2010)
-
-  This version focuses on the experimental *allocm() API, and on improved
-  run-time configuration/introspection.  Nonetheless, numerous performance
-  improvements are also included.
-
-  New features:
-  - Implement the experimental {,r,s,d}allocm() API, which provides a superset
-    of the functionality available via malloc(), calloc(), posix_memalign(),
-    realloc(), malloc_usable_size(), and free().  These functions can be used to
-    allocate/reallocate aligned zeroed memory, ask for optional extra memory
-    during reallocation, prevent object movement during reallocation, etc.
-  - Replace JEMALLOC_OPTIONS/JEMALLOC_PROF_PREFIX with MALLOC_CONF, which is
-    more human-readable, and more flexible.  For example:
-      JEMALLOC_OPTIONS=AJP
-    is now:
-      MALLOC_CONF=abort:true,fill:true,stats_print:true
-  - Port to Apple OS X.  Sponsored by Mozilla.
-  - Make it possible for the application to control thread-->arena mappings via
-    the "thread.arena" mallctl.
-  - Add compile-time support for all TLS-related functionality via pthreads TSD.
-    This is mainly of interest for OS X, which does not support TLS, but has a
-    TSD implementation with similar performance.
-  - Override memalign() and valloc() if they are provided by the system.
-  - Add the "arenas.purge" mallctl, which can be used to synchronously purge all
-    dirty unused pages.
-  - Make cumulative heap profiling data optional, so that it is possible to
-    limit the amount of memory consumed by heap profiling data structures.
-  - Add per thread allocation counters that can be accessed via the
-    "thread.allocated" and "thread.deallocated" mallctls.
-
-  Incompatible changes:
-  - Remove JEMALLOC_OPTIONS and malloc_options (see MALLOC_CONF above).
-  - Increase default backtrace depth from 4 to 128 for heap profiling.
-  - Disable interval-based profile dumps by default.
-
-  Bug fixes:
-  - Remove bad assertions in fork handler functions.  These assertions could
-    cause aborts for some combinations of configure settings.
-  - Fix strerror_r() usage to deal with non-standard semantics in GNU libc.
-  - Fix leak context reporting.  This bug tended to cause the number of contexts
-    to be underreported (though the reported number of objects and bytes were
-    correct).
-  - Fix a realloc() bug for large in-place growing reallocation.  This bug could
-    cause memory corruption, but it was hard to trigger.
-  - Fix an allocation bug for small allocations that could be triggered if
-    multiple threads raced to create a new run of backing pages.
-  - Enhance the heap profiler to trigger samples based on usable size, rather
-    than request size.
-  - Fix a heap profiling bug due to sometimes losing track of requested object
-    size for sampled objects.
-
-* 1.0.3 (August 12, 2010)
-
-  Bug fixes:
-  - Fix the libunwind-based implementation of stack backtracing (used for heap
-    profiling).  This bug could cause zero-length backtraces to be reported.
-  - Add a missing mutex unlock in library initialization code.  If multiple
-    threads raced to initialize malloc, some of them could end up permanently
-    blocked.
-
-* 1.0.2 (May 11, 2010)
-
-  Bug fixes:
-  - Fix junk filling of large objects, which could cause memory corruption.
-  - Add MAP_NORESERVE support for chunk mapping, because otherwise virtual
-    memory limits could cause swap file configuration to fail.  Contributed by
-    Jordan DeLong.
-
-* 1.0.1 (April 14, 2010)
-
-  Bug fixes:
-  - Fix compilation when --enable-fill is specified.
-  - Fix threads-related profiling bugs that affected accuracy and caused memory
-    to be leaked during thread exit.
-  - Fix dirty page purging race conditions that could cause crashes.
-  - Fix crash in tcache flushing code during thread destruction.
-
-* 1.0.0 (April 11, 2010)
-
-  This release focuses on speed and run-time introspection.  Numerous
-  algorithmic improvements make this release substantially faster than its
-  predecessors.
-
-  New features:
-  - Implement autoconf-based configuration system.
-  - Add mallctl*(), for the purposes of introspection and run-time
-    configuration.
-  - Make it possible for the application to manually flush a thread's cache, via
-    the "tcache.flush" mallctl.
-  - Base maximum dirty page count on proportion of active memory.
-  - Compute various addtional run-time statistics, including per size class
-    statistics for large objects.
-  - Expose malloc_stats_print(), which can be called repeatedly by the
-    application.
-  - Simplify the malloc_message() signature to only take one string argument,
-    and incorporate an opaque data pointer argument for use by the application
-    in combination with malloc_stats_print().
-  - Add support for allocation backed by one or more swap files, and allow the
-    application to disable over-commit if swap files are in use.
-  - Implement allocation profiling and leak checking.
-
-  Removed features:
-  - Remove the dynamic arena rebalancing code, since thread-specific caching
-    reduces its utility.
-
-  Bug fixes:
-  - Modify chunk allocation to work when address space layout randomization
-    (ASLR) is in use.
-  - Fix thread cleanup bugs related to TLS destruction.
-  - Handle 0-size allocation requests in posix_memalign().
-  - Fix a chunk leak.  The leaked chunks were never touched, so this impacted
-    virtual memory usage, but not physical memory usage.
-
-* linux_2008082[78]a (August 27/28, 2008)
-
-  These snapshot releases are the simple result of incorporating Linux-specific
-  support into the FreeBSD malloc sources.
-
--------------------------------------------------------------------------------
-vim:filetype=text:textwidth=80
--- a/jemalloc/Makefile.in
+++ b/jemalloc/Makefile.in
@@ -1,253 +0,0 @@
-# Clear out all vpaths, then set just one (default vpath) for the main build
-# directory.
-vpath
-vpath % .
-
-# Clear the default suffixes, so that built-in rules are not used.
-.SUFFIXES :
-
-SHELL := /bin/sh
-
-CC := @CC@
-
-# Configuration parameters.
-DESTDIR =
-BINDIR := $(DESTDIR)@BINDIR@
-INCLUDEDIR := $(DESTDIR)@INCLUDEDIR@
-LIBDIR := $(DESTDIR)@LIBDIR@
-DATADIR := $(DESTDIR)@DATADIR@
-MANDIR := $(DESTDIR)@MANDIR@
-
-# Build parameters.
-CPPFLAGS := @CPPFLAGS@ -I@srcroot@include -I@objroot@include
-CFLAGS := @CFLAGS@
-ifeq (macho, @abi@)
-CFLAGS += -dynamic
-endif
-LDFLAGS := @LDFLAGS@
-LIBS := @LIBS@
-RPATH_EXTRA := @RPATH_EXTRA@
-ifeq (macho, @abi@)
-SO := dylib
-WL_SONAME := dylib_install_name
-else
-SO := so
-WL_SONAME := soname
-endif
-REV := 1
-ifeq (macho, @abi@)
-TEST_LIBRARY_PATH := DYLD_FALLBACK_LIBRARY_PATH=@objroot@lib
-else
-TEST_LIBRARY_PATH :=
-endif
-
-# Lists of files.
-BINS := @srcroot@bin/pprof
-CHDRS := @objroot@include/jemalloc/jemalloc@install_suffix@.h \
-	@objroot@include/jemalloc/jemalloc_defs@install_suffix@.h
-CSRCS := @srcroot@src/jemalloc.c @srcroot@src/arena.c @srcroot@src/base.c \
-	@srcroot@src/chunk.c @srcroot@src/chunk_dss.c \
-	@srcroot@src/chunk_mmap.c @srcroot@src/chunk_swap.c @srcroot@src/ckh.c \
-	@srcroot@src/ctl.c @srcroot@src/extent.c @srcroot@src/hash.c \
-	@srcroot@src/huge.c @srcroot@src/mb.c @srcroot@src/mutex.c \
-	@srcroot@src/prof.c @srcroot@src/rtree.c \
-	@srcroot@src/stats.c @srcroot@src/tcache.c
-ifeq (macho, @abi@)
-CSRCS += @srcroot@src/zone.c
-endif
-STATIC_LIBS :=	@objroot@lib/libjemalloc@install_suffix@.a
-DSOS := @objroot@lib/libjemalloc@install_suffix@.$(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/mremap.c @srcroot@test/posix_memalign.c \
-	@srcroot@test/rallocm.c @srcroot@test/thread_arena.c
-
-.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
-
-.SECONDARY : $(CTESTS:@srcroot@%.c=@objroot@%.o)
-
-# Default target.
-all: $(DSOS) $(STATIC_LIBS)
-
-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.
-#
-include $(CSRCS:@srcroot@%.c=@objroot@%.d)
-
-@objroot@src/%.o: @srcroot@src/%.c
-	@mkdir -p $(@D)
-	$(CC) $(CFLAGS) -c $(CPPFLAGS) -o $@ $<
-	@$(SHELL) -ec "$(CC) -MM $(CPPFLAGS) $< | sed \"s/\($(subst /,\/,$(notdir $(basename $@)))\)\.o\([ :]*\)/$(subst /,\/,$(strip $(dir $@)))\1.o \2/g\" > $(@:%.o=%.d)"
-
-@objroot@src/%.pic.o: @srcroot@src/%.c
-	@mkdir -p $(@D)
-	$(CC) $(CFLAGS) -fPIC -DPIC -c $(CPPFLAGS) -o $@ $<
-	@$(SHELL) -ec "$(CC) -MM $(CPPFLAGS) $< | sed \"s/\($(subst /,\/,$(notdir $(basename $@)))\)\.o\([ :]*\)/$(subst /,\/,$(strip $(dir $@)))\1.o \2/g\" > $(@:%.o=%.d)"
-
-%.$(SO) : %.$(SO).$(REV)
-	@mkdir -p $(@D)
-	ln -sf $(<F) $@
-
-@objroot@lib/libjemalloc@install_suffix@.$(SO).$(REV) : $(CSRCS:@srcroot@%.c=@objroot@%.pic.o)
-	@mkdir -p $(@D)
-	$(CC) -shared -Wl,-$(WL_SONAME),$(@F) $(RPATH_EXTRA:%=@RPATH@%) -o $@ $+ $(LDFLAGS) $(LIBS)
-
-@objroot@lib/libjemalloc@install_suffix@_pic.a : $(CSRCS:@srcroot@%.c=@objroot@%.pic.o)
-	@mkdir -p $(@D)
-	ar crus $@ $+
-
-@objroot@lib/libjemalloc@install_suffix@.a : $(CSRCS:@srcroot@%.c=@objroot@%.o)
-	@mkdir -p $(@D)
-	ar crus $@ $+
-
-@objroot@test/%.o: @srcroot@test/%.c
-	@mkdir -p $(@D)
-	$(CC) $(CFLAGS) -c $(CPPFLAGS) -I@objroot@test -o $@ $<
-	@$(SHELL) -ec "$(CC) -MM $(CPPFLAGS) -I@objroot@test $< | sed \"s/\($(subst /,\/,$(notdir $(basename $@)))\)\.o\([ :]*\)/$(subst /,\/,$(strip $(dir $@)))\1.o \2/g\" > $(@:%.o=%.d)"
-
-@objroot@test/%: @objroot@test/%.o \
-		 @objroot@lib/libjemalloc@install_suffix@.$(SO)
-	@mkdir -p $(@D)
-ifneq (@RPATH@, )
-	$(CC) -o $@ $< @RPATH@@objroot@lib -L@objroot@lib -ljemalloc@install_suffix@
-else
-	$(CC) -o $@ $< -L@objroot@lib -ljemalloc@install_suffix@
-endif
-
-install_bin:
-	install -d $(BINDIR)
-	@for b in $(BINS); do \
-	echo "install -m 755 $$b $(BINDIR)"; \
-	install -m 755 $$b $(BINDIR); \
-done
-
-install_include:
-	install -d $(INCLUDEDIR)/jemalloc
-	@for h in $(CHDRS); do \
-	echo "install -m 644 $$h $(INCLUDEDIR)/jemalloc"; \
-	install -m 644 $$h $(INCLUDEDIR)/jemalloc; \
-done
-
-install_lib: $(DSOS) $(STATIC_LIBS)
-	install -d $(LIBDIR)
-	install -m 755 @objroot@lib/libjemalloc@install_suffix@.$(SO).$(REV) $(LIBDIR)
-	ln -sf libjemalloc@install_suffix@.$(SO).$(REV) $(LIBDIR)/libjemalloc@install_suffix@.$(SO)
-	install -m 755 @objroot@lib/libjemalloc@install_suffix@_pic.a $(LIBDIR)
-	install -m 755 @objroot@lib/libjemalloc@install_suffix@.a $(LIBDIR)
-
-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_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: $(CTESTS:@srcroot@%.c=@objroot@%)
-
-check: tests
-	@mkdir -p @objroot@test
-	@$(SHELL) -c 'total=0; \
-		failures=0; \
-		echo "========================================="; \
-		for t in $(CTESTS:@srcroot@%.c=@objroot@%); do \
-			total=`expr $$total + 1`; \
-			/bin/echo -n "$${t} ... "; \
-			$(TEST_LIBRARY_PATH) $${t} @abs_srcroot@ @abs_objroot@ \
-			  > @objroot@$${t}.out 2>&1; \
-			if test -e "@srcroot@$${t}.exp"; then \
-				diff -u @srcroot@$${t}.exp \
-				  @objroot@$${t}.out >/dev/null 2>&1; \
-				fail=$$?; \
-				if test "$${fail}" -eq "1" ; then \
-					failures=`expr $${failures} + 1`; \
-					echo "*** FAIL ***"; \
-				else \
-					echo "pass"; \
-				fi; \
-			else \
-				echo "*** FAIL *** (.exp file is missing)"; \
-				failures=`expr $${failures} + 1`; \
-			fi; \
-		done; \
-		echo "========================================="; \
-		echo "Failures: $${failures}/$${total}"'
-
-clean:
-	rm -f $(CSRCS:@srcroot@%.c=@objroot@%.o)
-	rm -f $(CSRCS:@srcroot@%.c=@objroot@%.pic.o)
-	rm -f $(CSRCS:@srcroot@%.c=@objroot@%.d)
-	rm -f $(CSRCS:@srcroot@%.c=@objroot@%.pic.d)
-	rm -f $(CTESTS:@srcroot@%.c=@objroot@%)
-	rm -f $(CTESTS:@srcroot@%.c=@objroot@%.o)
-	rm -f $(CTESTS:@srcroot@%.c=@objroot@%.d)
-	rm -f $(CTESTS:@srcroot@%.c=@objroot@%.out)
-	rm -f $(DSOS) $(STATIC_LIBS)
-
-distclean: clean
-	rm -rf @objroot@autom4te.cache
-	rm -f @objroot@config.log
-	rm -f @objroot@config.status
-	rm -f @objroot@config.stamp
-	rm -f @cfghdrs_out@
-	rm -f @cfgoutputs_out@
-
-relclean: distclean
-	rm -f @objroot@configure
-	rm -f @srcroot@VERSION
-	rm -f $(DOCS_HTML)
-	rm -f $(DOCS_MAN3)
-
-#===============================================================================
-# Re-configuration rules.
-
-ifeq (@enable_autogen@, 1)
-@srcroot@configure : @srcroot@configure.ac
-	cd ./@srcroot@ && @AUTOCONF@
-
-@objroot@config.status : @srcroot@configure
-	./@objroot@config.status --recheck
-
-@srcroot@config.stamp.in : @srcroot@configure.ac
-	echo stamp > @srcroot@config.stamp.in
-
-@objroot@config.stamp : @cfgoutputs_in@ @cfghdrs_in@ @srcroot@configure
-	./@objroot@config.status
-	@touch $@
-
-# There must be some action in order for make to re-read Makefile when it is
-# out of date.
-@cfgoutputs_out@ @cfghdrs_out@ : @objroot@config.stamp
-	@true
-endif
--- a/jemalloc/README
+++ b/jemalloc/README
@@ -1,16 +0,0 @@
-jemalloc is a general-purpose scalable concurrent malloc(3) implementation.
-This distribution is a stand-alone "portable" implementation that currently
-targets Linux and Apple OS X.  jemalloc is included as the default allocator in
-the FreeBSD and NetBSD operating systems, and it is used by the Mozilla Firefox
-web browser on Microsoft Windows-related platforms.  Depending on your needs,
-one of the other divergent versions may suit your needs better than this
-distribution.
-
-The COPYING file contains copyright and licensing information.
-
-The INSTALL file contains information on how to configure, build, and install
-jemalloc.
-
-The ChangeLog file contains a brief summary of changes for each release.
-
-URL: http://www.canonware.com/jemalloc/
--- a/jemalloc/configure.ac
+++ b/jemalloc/configure.ac
@@ -1,825 +0,0 @@
-dnl Process this file with autoconf to produce a configure script.
-AC_INIT([Makefile.in])
-
-dnl ============================================================================
-dnl Custom macro definitions.
-
-dnl JE_CFLAGS_APPEND(cflag)
-AC_DEFUN([JE_CFLAGS_APPEND],
-[
-AC_MSG_CHECKING([whether compiler supports $1])
-TCFLAGS="${CFLAGS}"
-if test "x${CFLAGS}" = "x" ; then
-  CFLAGS="$1"
-else
-  CFLAGS="${CFLAGS} $1"
-fi
-AC_RUN_IFELSE([AC_LANG_PROGRAM(
-[[
-]], [[
-    return 0;
-]])],
-              AC_MSG_RESULT([yes]),
-              AC_MSG_RESULT([no])
-              [CFLAGS="${TCFLAGS}"]
-)
-])
-
-dnl JE_COMPILABLE(label, hcode, mcode, rvar)
-AC_DEFUN([JE_COMPILABLE],
-[
-AC_MSG_CHECKING([whether $1 is compilable])
-AC_RUN_IFELSE([AC_LANG_PROGRAM(
-[$2], [$3])],
-              AC_MSG_RESULT([yes])
-              [$4="yes"],
-              AC_MSG_RESULT([no])
-              [$4="no"]
-)
-])
-
-dnl ============================================================================
-
-srcroot=$srcdir
-if test "x${srcroot}" = "x." ; then
-  srcroot=""
-else
-  srcroot="${srcroot}/"
-fi
-AC_SUBST([srcroot])
-abs_srcroot="`cd \"${srcdir}\"; pwd`/"
-AC_SUBST([abs_srcroot])
-
-objroot=""
-AC_SUBST([objroot])
-abs_objroot="`pwd`/"
-AC_SUBST([abs_objroot])
-
-dnl Munge install path variables.
-if test "x$prefix" = "xNONE" ; then
-  prefix="/usr/local"
-fi
-if test "x$exec_prefix" = "xNONE" ; then
-  exec_prefix=$prefix
-fi
-PREFIX=$prefix
-AC_SUBST([PREFIX])
-BINDIR=`eval echo $bindir`
-BINDIR=`eval echo $BINDIR`
-AC_SUBST([BINDIR])
-INCLUDEDIR=`eval echo $includedir`
-INCLUDEDIR=`eval echo $INCLUDEDIR`
-AC_SUBST([INCLUDEDIR])
-LIBDIR=`eval echo $libdir`
-LIBDIR=`eval echo $LIBDIR`
-AC_SUBST([LIBDIR])
-DATADIR=`eval echo $datadir`
-DATADIR=`eval echo $DATADIR`
-AC_SUBST([DATADIR])
-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=<path>], [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=$CFLAGS
-AC_PROG_CC
-if test "x$CFLAGS" = "x" ; then
-  no_CFLAGS="yes"
-  if test "x$GCC" = "xyes" ; then
-    JE_CFLAGS_APPEND([-std=gnu99])
-    JE_CFLAGS_APPEND([-Wall])
-    JE_CFLAGS_APPEND([-pipe])
-    JE_CFLAGS_APPEND([-g3])
-  fi
-fi
-dnl Append EXTRA_CFLAGS to CFLAGS, if defined.
-if test "x$EXTRA_CFLAGS" != "x" ; then
-  JE_CFLAGS_APPEND([$EXTRA_CFLAGS])
-fi
-AC_PROG_CPP
-
-AC_CHECK_SIZEOF([void *])
-if test "x${ac_cv_sizeof_void_p}" = "x8" ; then
-  LG_SIZEOF_PTR=3
-elif test "x${ac_cv_sizeof_void_p}" = "x4" ; then
-  LG_SIZEOF_PTR=2
-else
-  AC_MSG_ERROR([Unsupported pointer size: ${ac_cv_sizeof_void_p}])
-fi
-AC_DEFINE_UNQUOTED([LG_SIZEOF_PTR], [$LG_SIZEOF_PTR])
-
-AC_CHECK_SIZEOF([int])
-if test "x${ac_cv_sizeof_int}" = "x8" ; then
-  LG_SIZEOF_INT=3
-elif test "x${ac_cv_sizeof_int}" = "x4" ; then
-  LG_SIZEOF_INT=2
-else
-  AC_MSG_ERROR([Unsupported int size: ${ac_cv_sizeof_int}])
-fi
-AC_DEFINE_UNQUOTED([LG_SIZEOF_INT], [$LG_SIZEOF_INT])
-
-AC_CANONICAL_HOST
-dnl CPU-specific settings.
-CPU_SPINWAIT=""
-case "${host_cpu}" in
-  i[[345]]86)
-	;;
-  i686)
-	JE_COMPILABLE([__asm__], [], [[__asm__ volatile("pause"); return 0;]],
-	              [asm])
-	if test "x${asm}" = "xyes" ; then
-	    CPU_SPINWAIT='__asm__ volatile("pause")'
-	fi
-	;;
-  x86_64)
-	JE_COMPILABLE([__asm__ syntax], [],
-	              [[__asm__ volatile("pause"); return 0;]], [asm])
-	if test "x${asm}" = "xyes" ; then
-	    CPU_SPINWAIT='__asm__ volatile("pause")'
-	fi
-	;;
-  *)
-	;;
-esac
-AC_DEFINE_UNQUOTED([CPU_SPINWAIT], [$CPU_SPINWAIT])
-
-JE_COMPILABLE([__attribute__ syntax],
-              [static __attribute__((unused)) void foo(void){}],
-              [],
-              [attribute])
-if test "x${attribute}" = "xyes" ; then
-  AC_DEFINE([JEMALLOC_HAVE_ATTR], [ ])
-  if test "x$GCC" = "xyes" -a "${abi}" = "xelf"; then
-    JE_CFLAGS_APPEND([-fvisibility=internal])
-  fi
-fi
-
-dnl Platform-specific settings.  abi and RPATH can probably be determined
-dnl programmatically, but doing so is error-prone, which makes it generally
-dnl not worth the trouble.
-dnl 
-dnl Define cpp macros in CPPFLAGS, rather than doing AC_DEFINE(macro), since the
-dnl definitions need to be seen before any headers are included, which is a pain
-dnl to make happen otherwise.
-case "${host}" in
-  *-*-darwin*)
-	CFLAGS="$CFLAGS -fno-common -no-cpp-precomp"
-	abi="macho"
-	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE])
-	RPATH=""
-	;;
-  *-*-freebsd*)
-	CFLAGS="$CFLAGS"
-	abi="elf"
-	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE])
-	RPATH="-Wl,-rpath,"
-	;;
-  *-*-linux*)
-	CFLAGS="$CFLAGS"
-	CPPFLAGS="$CPPFLAGS -D_GNU_SOURCE"
-	abi="elf"
-	AC_DEFINE([JEMALLOC_PURGE_MADVISE_DONTNEED])
-	RPATH="-Wl,-rpath,"
-	;;
-  *-*-netbsd*)
-	AC_MSG_CHECKING([ABI])
-        AC_COMPILE_IFELSE([AC_LANG_PROGRAM(
-[[#ifdef __ELF__
-/* ELF */
-#else
-#error aout
-#endif
-]])],
-                          [CFLAGS="$CFLAGS"; abi="elf"],
-                          [abi="aout"])
-	AC_MSG_RESULT([$abi])
-	AC_DEFINE([JEMALLOC_PURGE_MADVISE_FREE])
-	RPATH="-Wl,-rpath,"
-	;;
-  *-*-solaris2*)
-	CFLAGS="$CFLAGS"
-	abi="elf"
-	RPATH="-Wl,-R,"
-	dnl Solaris needs this for sigwait().
-	CPPFLAGS="$CPPFLAGS -D_POSIX_PTHREAD_SEMANTICS"
-	LIBS="$LIBS -lposix4 -lsocket -lnsl"
-	;;
-  *)
-	AC_MSG_RESULT([Unsupported operating system: ${host}])
-	abi="elf"
-	RPATH="-Wl,-rpath,"
-	;;
-esac
-AC_SUBST([abi])
-AC_SUBST([RPATH])
-
-JE_COMPILABLE([mremap(...MREMAP_FIXED...)], [
-#define _GNU_SOURCE
-#include <sys/mman.h>
-], [
-void *p = mremap((void *)0, 0, 0, MREMAP_MAYMOVE|MREMAP_FIXED, (void *)0);
-], [mremap_fixed])
-if test "x${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=<rpath>], [Colon-separated rpath (ELF systems only)])],
-if test "x$with_rpath" = "xno" ; then
-  RPATH_EXTRA=
-else
-  RPATH_EXTRA="`echo $with_rpath | tr \":\" \" \"`"
-fi,
-  RPATH_EXTRA=
-)
-AC_SUBST([RPATH_EXTRA])
-
-dnl Disable rules that do automatic regeneration of configure output by default.
-AC_ARG_ENABLE([autogen],
-  [AS_HELP_STRING([--enable-autogen], [Automatically regenerate configure output])],
-if test "x$enable_autogen" = "xno" ; then
-  enable_autogen="0"
-else
-  enable_autogen="1"
-fi
-,
-enable_autogen="0"
-)
-AC_SUBST([enable_autogen])
-
-AC_PROG_INSTALL
-AC_PROG_RANLIB
-AC_PATH_PROG([AR], [ar], , [$PATH])
-AC_PATH_PROG([LD], [ld], , [$PATH])
-AC_PATH_PROG([AUTOCONF], [autoconf], , [$PATH])
-
-dnl Do not prefix public APIs by default.
-AC_ARG_WITH([jemalloc_prefix],
-  [AS_HELP_STRING([--with-jemalloc-prefix=<prefix>], [Prefix to prepend to all public APIs])],
-  [JEMALLOC_PREFIX="$with_jemalloc_prefix"],
-  [if test "x$abi" != "xmacho" ; then
-  JEMALLOC_PREFIX=""
-else
-  JEMALLOC_PREFIX="je_"
-fi]
-)
-if test "x$JEMALLOC_PREFIX" != "x" ; then
-  JEMALLOC_CPREFIX=`echo ${JEMALLOC_PREFIX} | tr "a-z" "A-Z"`
-  AC_DEFINE_UNQUOTED([JEMALLOC_PREFIX], ["$JEMALLOC_PREFIX"])
-  AC_DEFINE_UNQUOTED([JEMALLOC_CPREFIX], ["$JEMALLOC_CPREFIX"])
-  jemalloc_prefix="$JEMALLOC_PREFIX"
-  jemalloc_cprefix="$JEMALLOC_CPREFIX"
-  AC_SUBST([jemalloc_prefix])
-  AC_SUBST([jemalloc_cprefix])
-  AC_DEFINE_UNQUOTED([JEMALLOC_P(string_that_no_one_should_want_to_use_as_a_jemalloc_API_prefix)], [${JEMALLOC_PREFIX}##string_that_no_one_should_want_to_use_as_a_jemalloc_API_prefix])
-fi
-
-dnl Do not add suffix to installed files by default.
-AC_ARG_WITH([install_suffix],
-  [AS_HELP_STRING([--with-install-suffix=<suffix>], [Suffix to append to all installed files])],
-  [INSTALL_SUFFIX="$with_install_suffix"],
-  [INSTALL_SUFFIX=]
-)
-install_suffix="$INSTALL_SUFFIX"
-AC_SUBST([install_suffix])
-
-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} ${srcroot}test/jemalloc_test.h.in"
-
-cfgoutputs_out="Makefile"
-cfgoutputs_out="${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="${cfgoutputs_out} test/jemalloc_test.h"
-
-cfgoutputs_tup="Makefile"
-cfgoutputs_tup="${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="${cfgoutputs_tup} test/jemalloc_test.h:test/jemalloc_test.h.in"
-
-cfghdrs_in="${srcroot}include/jemalloc/jemalloc_defs.h.in"
-
-cfghdrs_out="include/jemalloc/jemalloc_defs${install_suffix}.h"
-
-cfghdrs_tup="include/jemalloc/jemalloc_defs${install_suffix}.h:include/jemalloc/jemalloc_defs.h.in"
-
-dnl Do not silence irrelevant compiler warnings by default, since enabling this
-dnl option incurs a performance penalty.
-AC_ARG_ENABLE([cc-silence],
-  [AS_HELP_STRING([--enable-cc-silence],
-                  [Silence irrelevant compiler warnings])],
-[if test "x$enable_cc_silence" = "xno" ; then
-  enable_cc_silence="0"
-else
-  enable_cc_silence="1"
-fi
-],
-[enable_cc_silence="0"]
-)
-if test "x$enable_cc_silence" = "x1" ; then
-  AC_DEFINE([JEMALLOC_CC_SILENCE])
-fi
-
-dnl Do not compile with debugging by default.
-AC_ARG_ENABLE([debug],
-  [AS_HELP_STRING([--enable-debug], [Build debugging code])],
-[if test "x$enable_debug" = "xno" ; then
-  enable_debug="0"
-else
-  enable_debug="1"
-fi
-],
-[enable_debug="0"]
-)
-if test "x$enable_debug" = "x1" ; then
-  AC_DEFINE([JEMALLOC_DEBUG], [ ])
-  AC_DEFINE([JEMALLOC_IVSALLOC], [ ])
-fi
-AC_SUBST([enable_debug])
-
-dnl Only optimize if not debugging.
-if test "x$enable_debug" = "x0" -a "x$no_CFLAGS" = "xyes" ; then
-  dnl Make sure that an optimization flag was not specified in EXTRA_CFLAGS.
-  optimize="no"
-  echo "$EXTRA_CFLAGS" | grep "\-O" >/dev/null || optimize="yes"
-  if test "x${optimize}" = "xyes" ; then
-    if test "x$GCC" = "xyes" ; then
-      JE_CFLAGS_APPEND([-O3])
-      JE_CFLAGS_APPEND([-funroll-loops])
-    else
-      JE_CFLAGS_APPEND([-O])
-    fi
-  fi
-fi
-
-dnl Do not enable statistics calculation by default.
-AC_ARG_ENABLE([stats],
-  [AS_HELP_STRING([--enable-stats], [Enable statistics calculation/reporting])],
-[if test "x$enable_stats" = "xno" ; then
-  enable_stats="0"
-else
-  enable_stats="1"
-fi
-],
-[enable_stats="0"]
-)
-if test "x$enable_stats" = "x1" ; then
-  AC_DEFINE([JEMALLOC_STATS], [ ])
-fi
-AC_SUBST([enable_stats])
-
-dnl Do not enable profiling by default.
-AC_ARG_ENABLE([prof],
-  [AS_HELP_STRING([--enable-prof], [Enable allocation profiling])],
-[if test "x$enable_prof" = "xno" ; then
-  enable_prof="0"
-else
-  enable_prof="1"
-fi
-],
-[enable_prof="0"]
-)
-AC_ARG_ENABLE([prof-libgcc],
-  [AS_HELP_STRING([--disable-prof-libgcc],
-  [Do not use libgcc for backtracing])],
-[if test "x$enable_prof_libgcc" = "xno" ; then
-  enable_prof_libgcc="0"
-else
-  enable_prof_libgcc="1"
-fi
-],
-[enable_prof_libgcc="1"]
-)
-AC_ARG_ENABLE([prof-libunwind],
-  [AS_HELP_STRING([--enable-prof-libunwind], [Use libunwind for backtracing])],
-[if test "x$enable_prof_libunwind" = "xno" ; then
-  enable_prof_libunwind="0"
-else
-  enable_prof_libunwind="1"
-fi
-],
-[enable_prof_libunwind="0"]
-)
-AC_ARG_WITH([static_libunwind],
-  [AS_HELP_STRING([--with-static-libunwind=<libunwind.a>],
-  [Path to static libunwind library; use rather than dynamically linking])],
-if test "x$with_static_libunwind" = "xno" ; then
-  LUNWIND="-lunwind"
-else
-  if test ! -f "$with_static_libunwind" ; then
-    AC_MSG_ERROR([Static libunwind not found: $with_static_libunwind])
-  fi
-  LUNWIND="$with_static_libunwind"
-fi,
-  LUNWIND="-lunwind"
-)
-if test "x$enable_prof" = "x1" ; then
-  LIBS="$LIBS -lm"
-  AC_DEFINE([JEMALLOC_PROF], [ ])
-  if test "x$enable_prof_libunwind" = "x1" ; then
-    AC_CHECK_HEADERS([libunwind.h], , [enable_prof_libunwind="0"])
-    if test "x$LUNWIND" = "x-lunwind" ; then
-      AC_CHECK_LIB([unwind], [backtrace], [LIBS="$LIBS $LUNWIND"],
-                   [enable_prof_libunwind="0"])
-    else
-      LIBS="$LIBS $LUNWIND"
-    fi
-    if test "x${enable_prof_libunwind}" = "x1" ; then
-      AC_DEFINE([JEMALLOC_PROF_LIBUNWIND], [ ])
-    fi
-  fi
-fi
-AC_SUBST([enable_prof])
-
-dnl If libunwind isn't enabled, try to use libgcc rather than gcc intrinsics
-dnl for backtracing.
-if test "x$enable_prof" = "x1" -a "x$enable_prof_libgcc" = "x1" ; then
-  if test "x$enable_prof_libunwind" = "x0" -a "x$GCC" = "xyes" ; then
-    enable_prof_libgcc="1"
-    AC_CHECK_HEADERS([unwind.h], , [enable_prof_libgcc="0"])
-    AC_CHECK_LIB([gcc], [_Unwind_Backtrace], [LIBS="$LIBS -lgcc"], [enable_prof_libgcc="0"])
-    if test "x${enable_prof_libgcc}" = "x1" ; then
-      AC_DEFINE([JEMALLOC_PROF_LIBGCC], [ ])
-    fi
-  else
-    enable_prof_libgcc="0"
-  fi
-fi
-
-dnl Enable tiny allocations by default.
-AC_ARG_ENABLE([tiny],
-  [AS_HELP_STRING([--disable-tiny], [Disable tiny (sub-quantum) allocations])],
-[if test "x$enable_tiny" = "xno" ; then
-  enable_tiny="0"
-else
-  enable_tiny="1"
-fi
-],
-[enable_tiny="1"]
-)
-if test "x$enable_tiny" = "x1" ; then
-  AC_DEFINE([JEMALLOC_TINY], [ ])
-fi
-AC_SUBST([enable_tiny])
-
-dnl Enable thread-specific caching by default.
-AC_ARG_ENABLE([tcache],
-  [AS_HELP_STRING([--disable-tcache], [Disable per thread caches])],
-[if test "x$enable_tcache" = "xno" ; then
-  enable_tcache="0"
-else
-  enable_tcache="1"
-fi
-],
-[enable_tcache="1"]
-)
-if test "x$enable_tcache" = "x1" ; then
-  AC_DEFINE([JEMALLOC_TCACHE], [ ])
-fi
-AC_SUBST([enable_tcache])
-
-dnl Do not enable mmap()ped swap files by default.
-AC_ARG_ENABLE([swap],
-  [AS_HELP_STRING([--enable-swap], [Enable mmap()ped swap files])],
-[if test "x$enable_swap" = "xno" ; then
-  enable_swap="0"
-else
-  enable_swap="1"
-fi
-],
-[enable_swap="0"]
-)
-if test "x$enable_swap" = "x1" ; then
-  AC_DEFINE([JEMALLOC_SWAP], [ ])
-fi
-AC_SUBST([enable_swap])
-
-dnl Do not enable allocation from DSS by default.
-AC_ARG_ENABLE([dss],
-  [AS_HELP_STRING([--enable-dss], [Enable allocation from DSS])],
-[if test "x$enable_dss" = "xno" ; then
-  enable_dss="0"
-else
-  enable_dss="1"
-fi
-],
-[enable_dss="0"]
-)
-if test "x$enable_dss" = "x1" ; then
-  AC_DEFINE([JEMALLOC_DSS], [ ])
-fi
-AC_SUBST([enable_dss])
-
-dnl Do not support the junk/zero filling option by default.
-AC_ARG_ENABLE([fill],
-  [AS_HELP_STRING([--enable-fill], [Support junk/zero filling option])],
-[if test "x$enable_fill" = "xno" ; then
-  enable_fill="0"
-else
-  enable_fill="1"
-fi
-],
-[enable_fill="0"]
-)
-if test "x$enable_fill" = "x1" ; then
-  AC_DEFINE([JEMALLOC_FILL], [ ])
-fi
-AC_SUBST([enable_fill])
-
-dnl Do not support the xmalloc option by default.
-AC_ARG_ENABLE([xmalloc],
-  [AS_HELP_STRING([--enable-xmalloc], [Support xmalloc option])],
-[if test "x$enable_xmalloc" = "xno" ; then
-  enable_xmalloc="0"
-else
-  enable_xmalloc="1"
-fi
-],
-[enable_xmalloc="0"]
-)
-if test "x$enable_xmalloc" = "x1" ; then
-  AC_DEFINE([JEMALLOC_XMALLOC], [ ])
-fi
-AC_SUBST([enable_xmalloc])
-
-dnl Do not support the SYSV option by default.
-AC_ARG_ENABLE([sysv],
-  [AS_HELP_STRING([--enable-sysv], [Support SYSV semantics option])],
-[if test "x$enable_sysv" = "xno" ; then
-  enable_sysv="0"
-else
-  enable_sysv="1"
-fi
-],
-[enable_sysv="0"]
-)
-if test "x$enable_sysv" = "x1" ; then
-  AC_DEFINE([JEMALLOC_SYSV], [ ])
-fi
-AC_SUBST([enable_sysv])
-
-dnl Do not determine page shift at run time by default.
-AC_ARG_ENABLE([dynamic_page_shift],
-  [AS_HELP_STRING([--enable-dynamic-page-shift],
-  [Determine page size at run time (don't trust configure result)])],
-[if test "x$enable_dynamic_page_shift" = "xno" ; then
-  enable_dynamic_page_shift="0"
-else
-  enable_dynamic_page_shift="1"
-fi
-],
-[enable_dynamic_page_shift="0"]
-)
-if test "x$enable_dynamic_page_shift" = "x1" ; then
-  AC_DEFINE([DYNAMIC_PAGE_SHIFT], [ ])
-fi
-AC_SUBST([enable_dynamic_page_shift])
-
-AC_MSG_CHECKING([STATIC_PAGE_SHIFT])
-AC_RUN_IFELSE([AC_LANG_PROGRAM(
-[[#include <stdio.h>
-#include <unistd.h>
-#include <strings.h>
-]], [[
-    long result;
-    FILE *f;
-
-    result = sysconf(_SC_PAGESIZE);
-    if (result == -1) {
-	return 1;
-    }
-    f = fopen("conftest.out", "w");
-    if (f == NULL) {
-	return 1;
-    }
-    fprintf(f, "%u\n", ffs((int)result) - 1);
-    close(f);
-
-    return 0;
-]])],
-              [STATIC_PAGE_SHIFT=`cat conftest.out`]
-              AC_MSG_RESULT([$STATIC_PAGE_SHIFT])
-              AC_DEFINE_UNQUOTED([STATIC_PAGE_SHIFT], [$STATIC_PAGE_SHIFT]),
-              AC_MSG_RESULT([error]))
-
-dnl ============================================================================
-dnl jemalloc configuration.
-dnl 
-
-dnl Set VERSION if source directory has an embedded git repository.
-if test -d "${srcroot}../.git" ; then
-  git describe --long --abbrev=40 > ${srcroot}VERSION
-fi
-jemalloc_version=`cat ${srcroot}VERSION`
-jemalloc_version_major=`echo ${jemalloc_version} | tr ".g-" " " | awk '{print [$]1}'`
-jemalloc_version_minor=`echo ${jemalloc_version} | tr ".g-" " " | awk '{print [$]2}'`
-jemalloc_version_bugfix=`echo ${jemalloc_version} | tr ".g-" " " | awk '{print [$]3}'`
-jemalloc_version_nrev=`echo ${jemalloc_version} | tr ".g-" " " | awk '{print [$]4}'`
-jemalloc_version_gid=`echo ${jemalloc_version} | tr ".g-" " " | awk '{print [$]5}'`
-AC_SUBST([jemalloc_version])
-AC_SUBST([jemalloc_version_major])
-AC_SUBST([jemalloc_version_minor])
-AC_SUBST([jemalloc_version_bugfix])
-AC_SUBST([jemalloc_version_nrev])
-AC_SUBST([jemalloc_version_gid])
-
-dnl ============================================================================
-dnl Configure pthreads.
-
-AC_CHECK_HEADERS([pthread.h], , [AC_MSG_ERROR([pthread.h is missing])])
-AC_CHECK_LIB([pthread], [pthread_create], [LIBS="$LIBS -lpthread"],
-             [AC_MSG_ERROR([libpthread is missing])])
-
-CPPFLAGS="$CPPFLAGS -D_REENTRANT"
-
-dnl Enable lazy locking by default.
-AC_ARG_ENABLE([lazy_lock],
-  [AS_HELP_STRING([--disable-lazy-lock],
-  [Disable lazy locking (always lock, even when single-threaded)])],
-[if test "x$enable_lazy_lock" = "xno" ; then
-  enable_lazy_lock="0"
-else
-  enable_lazy_lock="1"
-fi
-],
-[enable_lazy_lock="1"]
-)
-if test "x$enable_lazy_lock" = "x1" ; then
-  AC_CHECK_HEADERS([dlfcn.h], , [AC_MSG_ERROR([dlfcn.h is missing])])
-  AC_CHECK_LIB([dl], [dlopen], [LIBS="$LIBS -ldl"],
-               [AC_MSG_ERROR([libdl is missing])])
-  AC_DEFINE([JEMALLOC_LAZY_LOCK], [ ])
-fi
-AC_SUBST([enable_lazy_lock])
-
-AC_ARG_ENABLE([tls],
-  [AS_HELP_STRING([--disable-tls], [Disable thread-local storage (__thread keyword)])],
-if test "x$enable_tls" = "xno" ; then
-  enable_tls="0"
-else
-  enable_tls="1"
-fi
-,
-enable_tls="1"
-)
-if test "x${enable_tls}" = "x1" ; then
-AC_MSG_CHECKING([for TLS])
-AC_RUN_IFELSE([AC_LANG_PROGRAM(
-[[
-    __thread int x;
-]], [[
-    x = 42;
-
-    return 0;
-]])],
-              AC_MSG_RESULT([yes]),
-              AC_MSG_RESULT([no])
-              enable_tls="0")
-fi
-AC_SUBST([enable_tls])
-if test "x${enable_tls}" = "x0" ; then
-  AC_DEFINE_UNQUOTED([NO_TLS], [ ])
-fi
-
-dnl ============================================================================
-dnl Check for allocator-related functions that should be wrapped.
-
-AC_CHECK_FUNC([memalign],
-	      [AC_DEFINE([JEMALLOC_OVERRIDE_MEMALIGN])])
-AC_CHECK_FUNC([valloc],
-	      [AC_DEFINE([JEMALLOC_OVERRIDE_VALLOC])])
-
-dnl ============================================================================
-dnl Darwin-related configuration.
-
-if test "x${abi}" = "xmacho" ; then
-  AC_DEFINE([JEMALLOC_IVSALLOC])
-  AC_DEFINE([JEMALLOC_ZONE])
-
-  dnl The szone version jumped from 3 to 6 between the OS X 10.5.x and 10.6
-  dnl releases.  malloc_zone_t and malloc_introspection_t have new fields in
-  dnl 10.6, which is the only source-level indication of the change.
-  AC_MSG_CHECKING([malloc zone version])
-  AC_TRY_COMPILE([#include <stdlib.h>
-#include <malloc/malloc.h>], [
-	static malloc_zone_t zone;
-	static struct malloc_introspection_t zone_introspect;
-
-	zone.size = NULL;
-	zone.malloc = NULL;
-	zone.calloc = NULL;
-	zone.valloc = NULL;
-	zone.free = NULL;
-	zone.realloc = NULL;
-	zone.destroy = NULL;
-	zone.zone_name = "jemalloc_zone";
-	zone.batch_malloc = NULL;
-	zone.batch_free = NULL;
-	zone.introspect = &zone_introspect;
-	zone.version = 6;
-	zone.memalign = NULL;
-	zone.free_definite_size = NULL;
-
-	zone_introspect.enumerator = NULL;
-	zone_introspect.good_size = NULL;
-	zone_introspect.check = NULL;
-	zone_introspect.print = NULL;
-	zone_introspect.log = NULL;
-	zone_introspect.force_lock = NULL;
-	zone_introspect.force_unlock = NULL;
-	zone_introspect.statistics = NULL;
-	zone_introspect.zone_locked = NULL;
-], [AC_DEFINE_UNQUOTED([JEMALLOC_ZONE_VERSION], [6])
-    AC_MSG_RESULT([6])],
-   [AC_DEFINE_UNQUOTED([JEMALLOC_ZONE_VERSION], [3])
-   AC_MSG_RESULT([3])])
-fi
-
-dnl ============================================================================
-dnl Check for typedefs, structures, and compiler characteristics.
-AC_HEADER_STDBOOL
-
-dnl Process .in files.
-AC_SUBST([cfghdrs_in])
-AC_SUBST([cfghdrs_out])
-AC_CONFIG_HEADERS([$cfghdrs_tup])
-
-dnl ============================================================================
-dnl Generate outputs.
-AC_CONFIG_FILES([$cfgoutputs_tup config.stamp])
-AC_SUBST([cfgoutputs_in])
-AC_SUBST([cfgoutputs_out])
-AC_OUTPUT
-
-dnl ============================================================================
-dnl Print out the results of configuration.
-AC_MSG_RESULT([===============================================================================])
-AC_MSG_RESULT([jemalloc version   : $jemalloc_version])
-AC_MSG_RESULT([])
-AC_MSG_RESULT([CC                 : ${CC}])
-AC_MSG_RESULT([CPPFLAGS           : ${CPPFLAGS}])
-AC_MSG_RESULT([CFLAGS             : ${CFLAGS}])
-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}])
-AC_MSG_RESULT([LIBDIR             : ${LIBDIR}])
-AC_MSG_RESULT([DATADIR            : ${DATADIR}])
-AC_MSG_RESULT([MANDIR             : ${MANDIR}])
-AC_MSG_RESULT([])
-AC_MSG_RESULT([srcroot            : ${srcroot}])
-AC_MSG_RESULT([abs_srcroot        : ${abs_srcroot}])
-AC_MSG_RESULT([objroot            : ${objroot}])
-AC_MSG_RESULT([abs_objroot        : ${abs_objroot}])
-AC_MSG_RESULT([])
-AC_MSG_RESULT([JEMALLOC_PREFIX    : ${JEMALLOC_PREFIX}])
-AC_MSG_RESULT([install_suffix     : ${install_suffix}])
-AC_MSG_RESULT([autogen            : ${enable_autogen}])
-AC_MSG_RESULT([cc-silence         : ${enable_cc_silence}])
-AC_MSG_RESULT([debug              : ${enable_debug}])
-AC_MSG_RESULT([stats              : ${enable_stats}])
-AC_MSG_RESULT([prof               : ${enable_prof}])
-AC_MSG_RESULT([prof-libgcc        : ${enable_prof_libgcc}])
-AC_MSG_RESULT([prof-libunwind     : ${enable_prof_libunwind}])
-AC_MSG_RESULT([tiny               : ${enable_tiny}])
-AC_MSG_RESULT([tcache             : ${enable_tcache}])
-AC_MSG_RESULT([fill               : ${enable_fill}])
-AC_MSG_RESULT([xmalloc            : ${enable_xmalloc}])
-AC_MSG_RESULT([sysv               : ${enable_sysv}])
-AC_MSG_RESULT([swap               : ${enable_swap}])
-AC_MSG_RESULT([dss                : ${enable_dss}])
-AC_MSG_RESULT([dynamic_page_shift : ${enable_dynamic_page_shift}])
-AC_MSG_RESULT([lazy_lock          : ${enable_lazy_lock}])
-AC_MSG_RESULT([tls                : ${enable_tls}])
-AC_MSG_RESULT([===============================================================================])
--- a/jemalloc/include/jemalloc/internal/arena.h
+++ b/jemalloc/include/jemalloc/internal/arena.h
@@ -1,674 +0,0 @@
-/******************************************************************************/
-#ifdef JEMALLOC_H_TYPES
-
-/*
- * Subpages are an artificially designated partitioning of pages.  Their only
- * purpose is to support subpage-spaced size classes.
- *
- * There must be at least 4 subpages per page, due to the way size classes are
- * handled.
- */
-#define	LG_SUBPAGE		8
-#define	SUBPAGE			((size_t)(1U << LG_SUBPAGE))
-#define	SUBPAGE_MASK		(SUBPAGE - 1)
-
-/* Return the smallest subpage multiple that is >= s. */
-#define	SUBPAGE_CEILING(s)						\
-	(((s) + SUBPAGE_MASK) & ~SUBPAGE_MASK)
-
-#ifdef JEMALLOC_TINY
-   /* Smallest size class to support. */
-#  define LG_TINY_MIN		LG_SIZEOF_PTR
-#endif
-
-/*
- * Maximum size class that is a multiple of the quantum, but not (necessarily)
- * a power of 2.  Above this size, allocations are rounded up to the nearest
- * power of 2.
- */
-#define	LG_QSPACE_MAX_DEFAULT	7
-
-/*
- * Maximum size class that is a multiple of the cacheline, but not (necessarily)
- * a power of 2.  Above this size, allocations are rounded up to the nearest
- * power of 2.
- */
-#define	LG_CSPACE_MAX_DEFAULT	9
-
-/*
- * RUN_MAX_OVRHD indicates maximum desired run header overhead.  Runs are sized
- * as small as possible such that this setting is still honored, without
- * violating other constraints.  The goal is to make runs as small as possible
- * without exceeding a per run external fragmentation threshold.
- *
- * We use binary fixed point math for overhead computations, where the binary
- * point is implicitly RUN_BFP bits to the left.
- *
- * Note that it is possible to set RUN_MAX_OVRHD low enough that it cannot be
- * honored for some/all object sizes, since when heap profiling is enabled
- * there is one pointer of header overhead per object (plus a constant).  This
- * constraint is relaxed (ignored) for runs that are so small that the
- * per-region overhead is greater than:
- *
- *   (RUN_MAX_OVRHD / (reg_size << (3+RUN_BFP))
- */
-#define	RUN_BFP			12
-/*                                    \/   Implicit binary fixed point. */
-#define	RUN_MAX_OVRHD		0x0000003dU
-#define	RUN_MAX_OVRHD_RELAX	0x00001800U
-
-/*
- * The minimum ratio of active:dirty pages per arena is computed as:
- *
- *   (nactive >> opt_lg_dirty_mult) >= ndirty
- *
- * So, supposing that opt_lg_dirty_mult is 5, there can be no less than 32
- * times as many active pages as dirty pages.
- */
-#define	LG_DIRTY_MULT_DEFAULT	5
-
-typedef struct arena_chunk_map_s arena_chunk_map_t;
-typedef struct arena_chunk_s arena_chunk_t;
-typedef struct arena_run_s arena_run_t;
-typedef struct arena_bin_s arena_bin_t;
-typedef struct arena_s arena_t;
-
-#endif /* JEMALLOC_H_TYPES */
-/******************************************************************************/
-#ifdef JEMALLOC_H_STRUCTS
-
-/* Each element of the chunk map corresponds to one page within the chunk. */
-struct arena_chunk_map_s {
-	union {
-		/*
-		 * Linkage for run trees.  There are two disjoint uses:
-		 *
-		 * 1) arena_t's runs_avail_{clean,dirty} trees.
-		 * 2) arena_run_t conceptually uses this linkage for in-use
-		 *    non-full runs, rather than directly embedding linkage.
-		 */
-		rb_node(arena_chunk_map_t)	rb_link;
-		/*
-		 * List of runs currently in purgatory.  arena_chunk_purge()
-		 * temporarily allocates runs that contain dirty pages while
-		 * purging, so that other threads cannot use the runs while the
-		 * purging thread is operating without the arena lock held.
-		 */
-		ql_elm(arena_chunk_map_t)	ql_link;
-	}				u;
-
-#ifdef JEMALLOC_PROF
-	/* Profile counters, used for large object runs. */
-	prof_ctx_t			*prof_ctx;
-#endif
-
-	/*
-	 * Run address (or size) and various flags are stored together.  The bit
-	 * layout looks like (assuming 32-bit system):
-	 *
-	 *   ???????? ???????? ????---- ----dula
-	 *
-	 * ? : Unallocated: Run address for first/last pages, unset for internal
-	 *                  pages.
-	 *     Small: Run page offset.
-	 *     Large: Run size for first page, unset for trailing pages.
-	 * - : Unused.
-	 * d : dirty?
-	 * u : unzeroed?
-	 * l : large?
-	 * a : allocated?
-	 *
-	 * Following are example bit patterns for the three types of runs.
-	 *
-	 * p : run page offset
-	 * s : run size
-	 * c : (binind+1) for size class (used only if prof_promote is true)
-	 * x : don't care
-	 * - : 0
-	 * + : 1
-	 * [DULA] : bit set
-	 * [dula] : bit unset
-	 *
-	 *   Unallocated (clean):
-	 *     ssssssss ssssssss ssss---- ----du-a
-	 *     xxxxxxxx xxxxxxxx xxxx---- -----Uxx
-	 *     ssssssss ssssssss ssss---- ----dU-a
-	 *
-	 *   Unallocated (dirty):
-	 *     ssssssss ssssssss ssss---- ----D--a
-	 *     xxxxxxxx xxxxxxxx xxxx---- ----xxxx
-	 *     ssssssss ssssssss ssss---- ----D--a
-	 *
-	 *   Small:
-	 *     pppppppp pppppppp pppp---- ----d--A
-	 *     pppppppp pppppppp pppp---- -------A
-	 *     pppppppp pppppppp pppp---- ----d--A
-	 *
-	 *   Large:
-	 *     ssssssss ssssssss ssss---- ----D-LA
-	 *     xxxxxxxx xxxxxxxx xxxx---- ----xxxx
-	 *     -------- -------- -------- ----D-LA
-	 *
-	 *   Large (sampled, size <= PAGE_SIZE):
-	 *     ssssssss ssssssss sssscccc ccccD-LA
-	 *
-	 *   Large (not sampled, size == PAGE_SIZE):
-	 *     ssssssss ssssssss ssss---- ----D-LA
-	 */
-	size_t				bits;
-#ifdef JEMALLOC_PROF
-#define	CHUNK_MAP_CLASS_SHIFT	4
-#define	CHUNK_MAP_CLASS_MASK	((size_t)0xff0U)
-#endif
-#define	CHUNK_MAP_FLAGS_MASK	((size_t)0xfU)
-#define	CHUNK_MAP_DIRTY		((size_t)0x8U)
-#define	CHUNK_MAP_UNZEROED	((size_t)0x4U)
-#define	CHUNK_MAP_LARGE		((size_t)0x2U)
-#define	CHUNK_MAP_ALLOCATED	((size_t)0x1U)
-#define	CHUNK_MAP_KEY		CHUNK_MAP_ALLOCATED
-};
-typedef rb_tree(arena_chunk_map_t) arena_avail_tree_t;
-typedef rb_tree(arena_chunk_map_t) arena_run_tree_t;
-
-/* Arena chunk header. */
-struct arena_chunk_s {
-	/* Arena that owns the chunk. */
-	arena_t		*arena;
-
-	/* Linkage for the arena's chunks_dirty list. */
-	ql_elm(arena_chunk_t) link_dirty;
-
-	/*
-	 * True if the chunk is currently in the chunks_dirty list, due to
-	 * having at some point contained one or more dirty pages.  Removal
-	 * from chunks_dirty is lazy, so (dirtied && ndirty == 0) is possible.
-	 */
-	bool		dirtied;
-
-	/* Number of dirty pages. */
-	size_t		ndirty;
-
-	/*
-	 * Map of pages within chunk that keeps track of free/large/small.  The
-	 * first map_bias entries are omitted, since the chunk header does not
-	 * need to be tracked in the map.  This omission saves a header page
-	 * for common chunk sizes (e.g. 4 MiB).
-	 */
-	arena_chunk_map_t map[1]; /* Dynamically sized. */
-};
-typedef rb_tree(arena_chunk_t) arena_chunk_tree_t;
-
-struct arena_run_s {
-#ifdef JEMALLOC_DEBUG
-	uint32_t	magic;
-#  define ARENA_RUN_MAGIC 0x384adf93
-#endif
-
-	/* Bin this run is associated with. */
-	arena_bin_t	*bin;
-
-	/* Stack of available freed regions, or NULL. */
-	void		*avail;
-
-	/* Next region that has never been allocated, or run boundary. */
-	void		*next;
-
-	/* Number of free regions in run. */
-	unsigned	nfree;
-};
-
-struct arena_bin_s {
-	/*
-	 * All operations on runcur, runs, and stats require that lock be
-	 * locked.  Run allocation/deallocation are protected by the arena lock,
-	 * which may be acquired while holding one or more bin locks, but not
-	 * vise versa.
-	 */
-	malloc_mutex_t	lock;
-
-	/*
-	 * Current run being used to service allocations of this bin's size
-	 * class.
-	 */
-	arena_run_t	*runcur;
-
-	/*
-	 * Tree of non-full runs.  This tree is used when looking for an
-	 * existing run when runcur is no longer usable.  We choose the
-	 * non-full run that is lowest in memory; this policy tends to keep
-	 * objects packed well, and it can also help reduce the number of
-	 * almost-empty chunks.
-	 */
-	arena_run_tree_t runs;
-
-	/* Size of regions in a run for this bin's size class. */
-	size_t		reg_size;
-
-	/* Total size of a run for this bin's size class. */
-	size_t		run_size;
-
-	/* Total number of regions in a run for this bin's size class. */
-	uint32_t	nregs;
-
-#ifdef JEMALLOC_PROF
-	/*
-	 * Offset of first (prof_ctx_t *) in a run header for this bin's size
-	 * class, or 0 if (opt_prof == false).
-	 */
-	uint32_t	ctx0_offset;
-#endif
-
-	/* Offset of first region in a run for this bin's size class. */
-	uint32_t	reg0_offset;
-
-#ifdef JEMALLOC_STATS
-	/* Bin statistics. */
-	malloc_bin_stats_t stats;
-#endif
-};
-
-struct arena_s {
-#ifdef JEMALLOC_DEBUG
-	uint32_t		magic;
-#  define ARENA_MAGIC 0x947d3d24
-#endif
-
-	/* This arena's index within the arenas array. */
-	unsigned		ind;
-
-	/*
-	 * All non-bin-related operations on this arena require that lock be
-	 * locked.
-	 */
-	malloc_mutex_t		lock;
-
-#ifdef JEMALLOC_STATS
-	arena_stats_t		stats;
-#  ifdef JEMALLOC_TCACHE
-	/*
-	 * List of tcaches for extant threads associated with this arena.
-	 * Stats from these are merged incrementally, and at exit.
-	 */
-	ql_head(tcache_t)	tcache_ql;
-#  endif
-#endif
-
-#ifdef JEMALLOC_PROF
-	uint64_t		prof_accumbytes;
-#endif
-
-	/* List of dirty-page-containing chunks this arena manages. */
-	ql_head(arena_chunk_t)	chunks_dirty;
-
-	/*
-	 * In order to avoid rapid chunk allocation/deallocation when an arena
-	 * oscillates right on the cusp of needing a new chunk, cache the most
-	 * recently freed chunk.  The spare is left in the arena's chunk trees
-	 * until it is deleted.
-	 *
-	 * There is one spare chunk per arena, rather than one spare total, in
-	 * order to avoid interactions between multiple threads that could make
-	 * a single spare inadequate.
-	 */
-	arena_chunk_t		*spare;
-
-	/* Number of pages in active runs. */
-	size_t			nactive;
-
-	/*
-	 * Current count of pages within unused runs that are potentially
-	 * dirty, and for which madvise(... MADV_DONTNEED) has not been called.
-	 * By tracking this, we can institute a limit on how much dirty unused
-	 * memory is mapped for each arena.
-	 */
-	size_t			ndirty;
-
-	/*
-	 * Approximate number of pages being purged.  It is possible for
-	 * multiple threads to purge dirty pages concurrently, and they use
-	 * npurgatory to indicate the total number of pages all threads are
-	 * attempting to purge.
-	 */
-	size_t			npurgatory;
-
-	/*
-	 * Size/address-ordered trees of this arena's available runs.  The trees
-	 * are used for first-best-fit run allocation.  The dirty tree contains
-	 * runs with dirty pages (i.e. very likely to have been touched and
-	 * therefore have associated physical pages), whereas the clean tree
-	 * contains runs with pages that either have no associated physical
-	 * pages, or have pages that the kernel may recycle at any time due to
-	 * previous madvise(2) calls.  The dirty tree is used in preference to
-	 * the clean tree for allocations, because using dirty pages reduces
-	 * the amount of dirty purging necessary to keep the active:dirty page
-	 * ratio below the purge threshold.
-	 */
-	arena_avail_tree_t	runs_avail_clean;
-	arena_avail_tree_t	runs_avail_dirty;
-
-	/*
-	 * bins is used to store trees of free regions of the following sizes,
-	 * assuming a 64-bit system with 16-byte quantum, 4 KiB page size, and
-	 * default MALLOC_CONF.
-	 *
-	 *   bins[i] |   size |
-	 *   --------+--------+
-	 *        0  |      8 |
-	 *   --------+--------+
-	 *        1  |     16 |
-	 *        2  |     32 |
-	 *        3  |     48 |
-	 *           :        :
-	 *        6  |     96 |
-	 *        7  |    112 |
-	 *        8  |    128 |
-	 *   --------+--------+
-	 *        9  |    192 |
-	 *       10  |    256 |
-	 *       11  |    320 |
-	 *       12  |    384 |
-	 *       13  |    448 |
-	 *       14  |    512 |
-	 *   --------+--------+
-	 *       15  |    768 |
-	 *       16  |   1024 |
-	 *       17  |   1280 |
-	 *           :        :
-	 *       25  |   3328 |
-	 *       26  |   3584 |
-	 *       27  |   3840 |
-	 *   --------+--------+
-	 */
-	arena_bin_t		bins[1]; /* Dynamically sized. */
-};
-
-#endif /* JEMALLOC_H_STRUCTS */
-/******************************************************************************/
-#ifdef JEMALLOC_H_EXTERNS
-
-extern size_t	opt_lg_qspace_max;
-extern size_t	opt_lg_cspace_max;
-extern ssize_t		opt_lg_dirty_mult;
-extern uint8_t const	*small_size2bin;
-
-/* Various bin-related settings. */
-#ifdef JEMALLOC_TINY		/* Number of (2^n)-spaced tiny bins. */
-#  define		ntbins	((unsigned)(LG_QUANTUM - LG_TINY_MIN))
-#else
-#  define		ntbins	0
-#endif
-extern unsigned		nqbins; /* Number of quantum-spaced bins. */
-extern unsigned		ncbins; /* Number of cacheline-spaced bins. */
-extern unsigned		nsbins; /* Number of subpage-spaced bins. */
-extern unsigned		nbins;
-#ifdef JEMALLOC_TINY
-#  define		tspace_max	((size_t)(QUANTUM >> 1))
-#endif
-#define			qspace_min	QUANTUM
-extern size_t		qspace_max;
-extern size_t		cspace_min;
-extern size_t		cspace_max;
-extern size_t		sspace_min;
-extern size_t		sspace_max;
-#define			small_maxclass	sspace_max
-
-#define			nlclasses (chunk_npages - map_bias)
-
-void	arena_purge_all(arena_t *arena);
-#ifdef JEMALLOC_PROF
-void	arena_prof_accum(arena_t *arena, uint64_t accumbytes);
-#endif
-#ifdef JEMALLOC_TCACHE
-void	arena_tcache_fill_small(arena_t *arena, tcache_bin_t *tbin,
-    size_t binind
-#  ifdef JEMALLOC_PROF
-    , uint64_t prof_accumbytes
-#  endif
-    );
-#endif
-void	*arena_malloc_small(arena_t *arena, size_t size, bool zero);
-void	*arena_malloc_large(arena_t *arena, size_t size, bool zero);
-void	*arena_malloc(size_t size, bool zero);
-void	*arena_palloc(arena_t *arena, size_t size, size_t alloc_size,
-    size_t alignment, bool zero);
-size_t	arena_salloc(const void *ptr);
-#ifdef JEMALLOC_PROF
-void	arena_prof_promoted(const void *ptr, size_t size);
-size_t	arena_salloc_demote(const void *ptr);
-#endif
-void	arena_dalloc_bin(arena_t *arena, arena_chunk_t *chunk, void *ptr,
-    arena_chunk_map_t *mapelm);
-void	arena_dalloc_large(arena_t *arena, arena_chunk_t *chunk, void *ptr);
-#ifdef JEMALLOC_STATS
-void	arena_stats_merge(arena_t *arena, size_t *nactive, size_t *ndirty,
-    arena_stats_t *astats, malloc_bin_stats_t *bstats,
-    malloc_large_stats_t *lstats);
-#endif
-void	*arena_ralloc_no_move(void *ptr, size_t oldsize, size_t size,
-    size_t extra, bool zero);
-void	*arena_ralloc(void *ptr, size_t oldsize, size_t size, size_t extra,
-    size_t alignment, bool zero);
-bool	arena_new(arena_t *arena, unsigned ind);
-bool	arena_boot(void);
-
-#endif /* JEMALLOC_H_EXTERNS */
-/******************************************************************************/
-#ifdef JEMALLOC_H_INLINES
-
-#ifndef JEMALLOC_ENABLE_INLINE
-unsigned	arena_run_regind(arena_run_t *run, arena_bin_t *bin,
-    const void *ptr, size_t size);
-#  ifdef JEMALLOC_PROF
-prof_ctx_t	*arena_prof_ctx_get(const void *ptr);
-void	arena_prof_ctx_set(const void *ptr, prof_ctx_t *ctx);
-#  endif
-void	arena_dalloc(arena_t *arena, arena_chunk_t *chunk, void *ptr);
-#endif
-
-#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_ARENA_C_))
-JEMALLOC_INLINE unsigned
-arena_run_regind(arena_run_t *run, arena_bin_t *bin, const void *ptr,
-    size_t size)
-{
-	unsigned shift, diff, regind;
-
-	assert(run->magic == ARENA_RUN_MAGIC);
-
-	/*
-	 * Avoid doing division with a variable divisor if possible.  Using
-	 * actual division here can reduce allocator throughput by over 20%!
-	 */
-	diff = (unsigned)((uintptr_t)ptr - (uintptr_t)run - bin->reg0_offset);
-
-	/* Rescale (factor powers of 2 out of the numerator and denominator). */
-	shift = ffs(size) - 1;
-	diff >>= shift;
-	size >>= shift;
-
-	if (size == 1) {
-		/* The divisor was a power of 2. */
-		regind = diff;
-	} else {
-		/*
-		 * To divide by a number D that is not a power of two we
-		 * multiply by (2^21 / D) and then right shift by 21 positions.
-		 *
-		 *   X / D
-		 *
-		 * becomes
-		 *
-		 *   (X * size_invs[D - 3]) >> SIZE_INV_SHIFT
-		 *
-		 * We can omit the first three elements, because we never
-		 * divide by 0, and 1 and 2 are both powers of two, which are
-		 * handled above.
-		 */
-#define	SIZE_INV_SHIFT 21
-#define	SIZE_INV(s) (((1U << SIZE_INV_SHIFT) / (s)) + 1)
-		static const unsigned size_invs[] = {
-		    SIZE_INV(3),
-		    SIZE_INV(4), SIZE_INV(5), SIZE_INV(6), SIZE_INV(7),
-		    SIZE_INV(8), SIZE_INV(9), SIZE_INV(10), SIZE_INV(11),
-		    SIZE_INV(12), SIZE_INV(13), SIZE_INV(14), SIZE_INV(15),
-		    SIZE_INV(16), SIZE_INV(17), SIZE_INV(18), SIZE_INV(19),
-		    SIZE_INV(20), SIZE_INV(21), SIZE_INV(22), SIZE_INV(23),
-		    SIZE_INV(24), SIZE_INV(25), SIZE_INV(26), SIZE_INV(27),
-		    SIZE_INV(28), SIZE_INV(29), SIZE_INV(30), SIZE_INV(31)
-		};
-
-		if (size <= ((sizeof(size_invs) / sizeof(unsigned)) + 2))
-			regind = (diff * size_invs[size - 3]) >> SIZE_INV_SHIFT;
-		else
-			regind = diff / size;
-#undef SIZE_INV
-#undef SIZE_INV_SHIFT
-	}
-	assert(diff == regind * size);
-	assert(regind < bin->nregs);
-
-	return (regind);
-}
-
-#ifdef JEMALLOC_PROF
-JEMALLOC_INLINE prof_ctx_t *
-arena_prof_ctx_get(const void *ptr)
-{
-	prof_ctx_t *ret;
-	arena_chunk_t *chunk;
-	size_t pageind, mapbits;
-
-	assert(ptr != NULL);
-	assert(CHUNK_ADDR2BASE(ptr) != ptr);
-
-	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
-	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> PAGE_SHIFT;
-	mapbits = chunk->map[pageind-map_bias].bits;
-	assert((mapbits & CHUNK_MAP_ALLOCATED) != 0);
-	if ((mapbits & CHUNK_MAP_LARGE) == 0) {
-		if (prof_promote)
-			ret = (prof_ctx_t *)(uintptr_t)1U;
-		else {
-			arena_run_t *run = (arena_run_t *)((uintptr_t)chunk +
-			    (uintptr_t)((pageind - (mapbits >> PAGE_SHIFT)) <<
-			    PAGE_SHIFT));
-			arena_bin_t *bin = run->bin;
-			unsigned regind;
-
-			assert(run->magic == ARENA_RUN_MAGIC);
-			regind = arena_run_regind(run, bin, ptr, bin->reg_size);
-			ret = *(prof_ctx_t **)((uintptr_t)run +
-			    bin->ctx0_offset + (regind *
-			    sizeof(prof_ctx_t *)));
-		}
-	} else
-		ret = chunk->map[pageind-map_bias].prof_ctx;
-
-	return (ret);
-}
-
-JEMALLOC_INLINE void
-arena_prof_ctx_set(const void *ptr, prof_ctx_t *ctx)
-{
-	arena_chunk_t *chunk;
-	size_t pageind, mapbits;
-
-	assert(ptr != NULL);
-	assert(CHUNK_ADDR2BASE(ptr) != ptr);
-
-	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
-	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> PAGE_SHIFT;
-	mapbits = chunk->map[pageind-map_bias].bits;
-	assert((mapbits & CHUNK_MAP_ALLOCATED) != 0);
-	if ((mapbits & CHUNK_MAP_LARGE) == 0) {
-		if (prof_promote == false) {
-			arena_run_t *run = (arena_run_t *)((uintptr_t)chunk +
-			    (uintptr_t)((pageind - (mapbits >> PAGE_SHIFT)) <<
-			    PAGE_SHIFT));
-			arena_bin_t *bin = run->bin;
-			unsigned regind;
-
-			assert(run->magic == ARENA_RUN_MAGIC);
-			regind = arena_run_regind(run, bin, ptr, bin->reg_size);
-
-			*((prof_ctx_t **)((uintptr_t)run + bin->ctx0_offset
-			    + (regind * sizeof(prof_ctx_t *)))) = ctx;
-		} else
-			assert((uintptr_t)ctx == (uintptr_t)1U);
-	} else
-		chunk->map[pageind-map_bias].prof_ctx = ctx;
-}
-#endif
-
-JEMALLOC_INLINE void
-arena_dalloc(arena_t *arena, arena_chunk_t *chunk, void *ptr)
-{
-	size_t pageind;
-	arena_chunk_map_t *mapelm;
-
-	assert(arena != NULL);
-	assert(arena->magic == ARENA_MAGIC);
-	assert(chunk->arena == arena);
-	assert(ptr != NULL);
-	assert(CHUNK_ADDR2BASE(ptr) != ptr);
-
-	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> PAGE_SHIFT;
-	mapelm = &chunk->map[pageind-map_bias];
-	assert((mapelm->bits & CHUNK_MAP_ALLOCATED) != 0);
-	if ((mapelm->bits & CHUNK_MAP_LARGE) == 0) {
-		/* Small allocation. */
-#ifdef JEMALLOC_TCACHE
-		tcache_t *tcache;
-
-		if ((tcache = tcache_get()) != NULL)
-			tcache_dalloc_small(tcache, ptr);
-		else {
-#endif
-			arena_run_t *run;
-			arena_bin_t *bin;
-
-			run = (arena_run_t *)((uintptr_t)chunk +
-			    (uintptr_t)((pageind - (mapelm->bits >>
-			    PAGE_SHIFT)) << PAGE_SHIFT));
-			assert(run->magic == ARENA_RUN_MAGIC);
-			assert(((uintptr_t)ptr - ((uintptr_t)run +
-			    (uintptr_t)run->bin->reg0_offset)) %
-			    run->bin->reg_size == 0);
-			bin = run->bin;
-			malloc_mutex_lock(&bin->lock);
-			arena_dalloc_bin(arena, chunk, ptr, mapelm);
-			malloc_mutex_unlock(&bin->lock);
-#ifdef JEMALLOC_TCACHE
-		}
-#endif
-	} else {
-#ifdef JEMALLOC_TCACHE
-		size_t size = mapelm->bits & ~PAGE_MASK;
-
-		assert(((uintptr_t)ptr & PAGE_MASK) == 0);
-		if (size <= tcache_maxclass) {
-			tcache_t *tcache;
-
-			if ((tcache = tcache_get()) != NULL)
-				tcache_dalloc_large(tcache, ptr, size);
-			else {
-				malloc_mutex_lock(&arena->lock);
-				arena_dalloc_large(arena, chunk, ptr);
-				malloc_mutex_unlock(&arena->lock);
-			}
-		} else {
-			malloc_mutex_lock(&arena->lock);
-			arena_dalloc_large(arena, chunk, ptr);
-			malloc_mutex_unlock(&arena->lock);
-		}
-#else
-		assert(((uintptr_t)ptr & PAGE_MASK) == 0);
-		malloc_mutex_lock(&arena->lock);
-		arena_dalloc_large(arena, chunk, ptr);
-		malloc_mutex_unlock(&arena->lock);
-#endif
-	}
-}
-#endif
-
-#endif /* JEMALLOC_H_INLINES */
-/******************************************************************************/
--- a/jemalloc/include/jemalloc/internal/chunk_swap.h
+++ b/jemalloc/include/jemalloc/internal/chunk_swap.h
@@ -1,34 +0,0 @@
-#ifdef JEMALLOC_SWAP
-/******************************************************************************/
-#ifdef JEMALLOC_H_TYPES
-
-#endif /* JEMALLOC_H_TYPES */
-/******************************************************************************/
-#ifdef JEMALLOC_H_STRUCTS
-
-#endif /* JEMALLOC_H_STRUCTS */
-/******************************************************************************/
-#ifdef JEMALLOC_H_EXTERNS
-
-extern malloc_mutex_t	swap_mtx;
-extern bool		swap_enabled;
-extern bool		swap_prezeroed;
-extern size_t		swap_nfds;
-extern int		*swap_fds;
-#ifdef JEMALLOC_STATS
-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);
-
-#endif /* JEMALLOC_H_EXTERNS */
-/******************************************************************************/
-#ifdef JEMALLOC_H_INLINES
-
-#endif /* JEMALLOC_H_INLINES */
-/******************************************************************************/
-#endif /* JEMALLOC_SWAP */
--- a/jemalloc/include/jemalloc/internal/jemalloc_internal.h.in
+++ b/jemalloc/include/jemalloc/internal/jemalloc_internal.h.in
@@ -1,738 +0,0 @@
-#include <sys/mman.h>
-#include <sys/param.h>
-#include <sys/time.h>
-#include <sys/types.h>
-#include <sys/sysctl.h>
-#include <sys/uio.h>
-
-#include <errno.h>
-#include <limits.h>
-#ifndef SIZE_T_MAX
-#  define SIZE_T_MAX	SIZE_MAX
-#endif
-#include <pthread.h>
-#include <sched.h>
-#include <stdarg.h>
-#include <stdbool.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <stdint.h>
-#include <stddef.h>
-#ifndef offsetof
-#  define offsetof(type, member)	((size_t)&(((type *)NULL)->member))
-#endif
-#include <inttypes.h>
-#include <string.h>
-#include <strings.h>
-#include <ctype.h>
-#include <unistd.h>
-#include <fcntl.h>
-#include <pthread.h>
-#include <math.h>
-
-#define	JEMALLOC_MANGLE
-#include "../jemalloc@install_suffix@.h"
-
-#ifdef JEMALLOC_ZONE
-#include <mach/mach_error.h>
-#include <mach/mach_init.h>
-#include <mach/vm_map.h>
-#include <malloc/malloc.h>
-#endif
-
-#ifdef JEMALLOC_LAZY_LOCK
-#include <dlfcn.h>
-#endif
-
-#define	RB_COMPACT
-#include "jemalloc/internal/rb.h"
-#include "jemalloc/internal/qr.h"
-#include "jemalloc/internal/ql.h"
-
-extern void	(*JEMALLOC_P(malloc_message))(void *wcbopaque, const char *s);
-
-/*
- * Define a custom assert() in order to reduce the chances of deadlock during
- * assertion failure.
- */
-#ifdef JEMALLOC_DEBUG
-#  define assert(e) do {						\
-	if (!(e)) {							\
-		char line_buf[UMAX2S_BUFSIZE];				\
-		malloc_write("<jemalloc>: ");				\
-		malloc_write(__FILE__);					\
-		malloc_write(":");					\
-		malloc_write(u2s(__LINE__, 10, line_buf));		\
-		malloc_write(": Failed assertion: ");			\
-		malloc_write("\"");					\
-		malloc_write(#e);					\
-		malloc_write("\"\n");					\
-		abort();						\
-	}								\
-} while (0)
-#else
-#define assert(e)
-#endif
-
-/*
- * jemalloc can conceptually be broken into components (arena, tcache, etc.),
- * but there are circular dependencies that cannot be broken without
- * substantial performance degradation.  In order to reduce the effect on
- * visual code flow, read the header files in multiple passes, with one of the
- * following cpp variables defined during each pass:
- *
- *   JEMALLOC_H_TYPES   : Preprocessor-defined constants and psuedo-opaque data
- *                        types.
- *   JEMALLOC_H_STRUCTS : Data structures.
- *   JEMALLOC_H_EXTERNS : Extern data declarations and function prototypes.
- *   JEMALLOC_H_INLINES : Inline functions.
- */
-/******************************************************************************/
-#define JEMALLOC_H_TYPES
-
-#define	ALLOCM_LG_ALIGN_MASK	((int)0x3f)
-
-#define	ZU(z)	((size_t)z)
-
-#ifndef __DECONST
-#  define	__DECONST(type, var)	((type)(uintptr_t)(const void *)(var))
-#endif
-
-#ifdef JEMALLOC_DEBUG
-   /* Disable inlining to make debugging easier. */
-#  define JEMALLOC_INLINE
-#  define inline
-#else
-#  define JEMALLOC_ENABLE_INLINE
-#  define JEMALLOC_INLINE static inline
-#endif
-
-/* Size of stack-allocated buffer passed to buferror(). */
-#define	BUFERROR_BUF		64
-
-/* Minimum alignment of allocations is 2^LG_QUANTUM bytes. */
-#ifdef __i386__
-#  define LG_QUANTUM		4
-#endif
-#ifdef __ia64__
-#  define LG_QUANTUM		4
-#endif
-#ifdef __alpha__
-#  define LG_QUANTUM		4
-#endif
-#ifdef __sparc64__
-#  define LG_QUANTUM		4
-#endif
-#if (defined(__amd64__) || defined(__x86_64__))
-#  define LG_QUANTUM		4
-#endif
-#ifdef __arm__
-#  define LG_QUANTUM		3
-#endif
-#ifdef __mips__
-#  define LG_QUANTUM		3
-#endif
-#ifdef __powerpc__
-#  define LG_QUANTUM		4
-#endif
-#ifdef __s390x__
-#  define LG_QUANTUM		4
-#endif
-
-#define	QUANTUM			((size_t)(1U << LG_QUANTUM))
-#define	QUANTUM_MASK		(QUANTUM - 1)
-
-/* Return the smallest quantum multiple that is >= a. */
-#define	QUANTUM_CEILING(a)						\
-	(((a) + QUANTUM_MASK) & ~QUANTUM_MASK)
-
-#define	SIZEOF_PTR		(1U << LG_SIZEOF_PTR)
-
-/*
- * Maximum size of L1 cache line.  This is used to avoid cache line aliasing.
- * In addition, this controls the spacing of cacheline-spaced size classes.
- */
-#define	LG_CACHELINE		6
-#define	CACHELINE		((size_t)(1U << LG_CACHELINE))
-#define	CACHELINE_MASK		(CACHELINE - 1)
-
-/* Return the smallest cacheline multiple that is >= s. */
-#define	CACHELINE_CEILING(s)						\
-	(((s) + CACHELINE_MASK) & ~CACHELINE_MASK)
-
-/*
- * Page size.  STATIC_PAGE_SHIFT is determined by the configure script.  If
- * DYNAMIC_PAGE_SHIFT is enabled, only use the STATIC_PAGE_* macros where
- * compile-time values are required for the purposes of defining data
- * structures.
- */
-#define	STATIC_PAGE_SIZE ((size_t)(1U << STATIC_PAGE_SHIFT))
-#define	STATIC_PAGE_MASK ((size_t)(STATIC_PAGE_SIZE - 1))
-
-#ifdef PAGE_SHIFT
-#  undef PAGE_SHIFT
-#endif
-#ifdef PAGE_SIZE
-#  undef PAGE_SIZE
-#endif
-#ifdef PAGE_MASK
-#  undef PAGE_MASK
-#endif
-
-#ifdef DYNAMIC_PAGE_SHIFT
-#  define PAGE_SHIFT	lg_pagesize
-#  define PAGE_SIZE	pagesize
-#  define PAGE_MASK	pagesize_mask
-#else
-#  define PAGE_SHIFT	STATIC_PAGE_SHIFT
-#  define PAGE_SIZE	STATIC_PAGE_SIZE
-#  define PAGE_MASK	STATIC_PAGE_MASK
-#endif
-
-/* Return the smallest pagesize multiple that is >= s. */
-#define	PAGE_CEILING(s)							\
-	(((s) + PAGE_MASK) & ~PAGE_MASK)
-
-#include "jemalloc/internal/prn.h"
-#include "jemalloc/internal/ckh.h"
-#include "jemalloc/internal/stats.h"
-#include "jemalloc/internal/ctl.h"
-#include "jemalloc/internal/mutex.h"
-#include "jemalloc/internal/mb.h"
-#include "jemalloc/internal/extent.h"
-#include "jemalloc/internal/arena.h"
-#include "jemalloc/internal/base.h"
-#include "jemalloc/internal/chunk.h"
-#include "jemalloc/internal/huge.h"
-#include "jemalloc/internal/rtree.h"
-#include "jemalloc/internal/tcache.h"
-#include "jemalloc/internal/hash.h"
-#ifdef JEMALLOC_ZONE
-#include "jemalloc/internal/zone.h"
-#endif
-#include "jemalloc/internal/prof.h"
-
-#undef JEMALLOC_H_TYPES
-/******************************************************************************/
-#define JEMALLOC_H_STRUCTS
-
-#include "jemalloc/internal/prn.h"
-#include "jemalloc/internal/ckh.h"
-#include "jemalloc/internal/stats.h"
-#include "jemalloc/internal/ctl.h"
-#include "jemalloc/internal/mutex.h"
-#include "jemalloc/internal/mb.h"
-#include "jemalloc/internal/extent.h"
-#include "jemalloc/internal/arena.h"
-#include "jemalloc/internal/base.h"
-#include "jemalloc/internal/chunk.h"
-#include "jemalloc/internal/huge.h"
-#include "jemalloc/internal/rtree.h"
-#include "jemalloc/internal/tcache.h"
-#include "jemalloc/internal/hash.h"
-#ifdef JEMALLOC_ZONE
-#include "jemalloc/internal/zone.h"
-#endif
-#include "jemalloc/internal/prof.h"
-
-#ifdef JEMALLOC_STATS
-typedef struct {
-	uint64_t	allocated;
-	uint64_t	deallocated;
-} thread_allocated_t;
-#endif
-
-#undef JEMALLOC_H_STRUCTS
-/******************************************************************************/
-#define JEMALLOC_H_EXTERNS
-
-extern bool	opt_abort;
-#ifdef JEMALLOC_FILL
-extern bool	opt_junk;
-#endif
-#ifdef JEMALLOC_SYSV
-extern bool	opt_sysv;
-#endif
-#ifdef JEMALLOC_XMALLOC
-extern bool	opt_xmalloc;
-#endif
-#ifdef JEMALLOC_FILL
-extern bool	opt_zero;
-#endif
-extern size_t	opt_narenas;
-
-#ifdef DYNAMIC_PAGE_SHIFT
-extern size_t		pagesize;
-extern size_t		pagesize_mask;
-extern size_t		lg_pagesize;
-#endif
-
-/* Number of CPUs. */
-extern unsigned		ncpus;
-
-extern malloc_mutex_t	arenas_lock; /* Protects arenas initialization. */
-#ifndef NO_TLS
-/*
- * Map of pthread_self() --> arenas[???], used for selecting an arena to use
- * for allocations.
- */
-extern __thread arena_t	*arenas_tls JEMALLOC_ATTR(tls_model("initial-exec"));
-#  define ARENA_GET()	arenas_tls
-#  define ARENA_SET(v)	do {						\
-	arenas_tls = (v);						\
-} while (0)
-#else
-extern pthread_key_t	arenas_tsd;
-#  define ARENA_GET()	((arena_t *)pthread_getspecific(arenas_tsd))
-#  define ARENA_SET(v)	do {						\
-	pthread_setspecific(arenas_tsd, (void *)(v));			\
-} while (0)
-#endif
-
-/*
- * Arenas that are used to service external requests.  Not all elements of the
- * arenas array are necessarily used; arenas are created lazily as needed.
- */
-extern arena_t		**arenas;
-extern unsigned		narenas;
-
-#ifdef JEMALLOC_STATS
-#  ifndef NO_TLS
-extern __thread thread_allocated_t	thread_allocated_tls;
-#    define ALLOCATED_GET() (thread_allocated_tls.allocated)
-#    define ALLOCATEDP_GET() (&thread_allocated_tls.allocated)
-#    define DEALLOCATED_GET() (thread_allocated_tls.deallocated)
-#    define DEALLOCATEDP_GET() (&thread_allocated_tls.deallocated)
-#    define ALLOCATED_ADD(a, d) do {					\
-	thread_allocated_tls.allocated += a;				\
-	thread_allocated_tls.deallocated += d;				\
-} while (0)
-#  else
-extern pthread_key_t	thread_allocated_tsd;
-thread_allocated_t	*thread_allocated_get_hard(void);
-
-#    define ALLOCATED_GET() (thread_allocated_get()->allocated)
-#    define ALLOCATEDP_GET() (&thread_allocated_get()->allocated)
-#    define DEALLOCATED_GET() (thread_allocated_get()->deallocated)
-#    define DEALLOCATEDP_GET() (&thread_allocated_get()->deallocated)
-#    define ALLOCATED_ADD(a, d) do {					\
-	thread_allocated_t *thread_allocated = thread_allocated_get();	\
-	thread_allocated->allocated += (a);				\
-	thread_allocated->deallocated += (d);				\
-} while (0)
-#  endif
-#endif
-
-arena_t	*arenas_extend(unsigned ind);
-arena_t	*choose_arena_hard(void);
-int	buferror(int errnum, char *buf, size_t buflen);
-void	jemalloc_prefork(void);
-void	jemalloc_postfork(void);
-
-#include "jemalloc/internal/prn.h"
-#include "jemalloc/internal/ckh.h"
-#include "jemalloc/internal/stats.h"
-#include "jemalloc/internal/ctl.h"
-#include "jemalloc/internal/mutex.h"
-#include "jemalloc/internal/mb.h"
-#include "jemalloc/internal/extent.h"
-#include "jemalloc/internal/arena.h"
-#include "jemalloc/internal/base.h"
-#include "jemalloc/internal/chunk.h"
-#include "jemalloc/internal/huge.h"
-#include "jemalloc/internal/rtree.h"
-#include "jemalloc/internal/tcache.h"
-#include "jemalloc/internal/hash.h"
-#ifdef JEMALLOC_ZONE
-#include "jemalloc/internal/zone.h"
-#endif
-#include "jemalloc/internal/prof.h"
-
-#undef JEMALLOC_H_EXTERNS
-/******************************************************************************/
-#define JEMALLOC_H_INLINES
-
-#include "jemalloc/internal/prn.h"
-#include "jemalloc/internal/ckh.h"
-#include "jemalloc/internal/stats.h"
-#include "jemalloc/internal/ctl.h"
-#include "jemalloc/internal/mutex.h"
-#include "jemalloc/internal/mb.h"
-#include "jemalloc/internal/extent.h"
-#include "jemalloc/internal/base.h"
-#include "jemalloc/internal/chunk.h"
-#include "jemalloc/internal/huge.h"
-
-#ifndef JEMALLOC_ENABLE_INLINE
-size_t	pow2_ceil(size_t x);
-size_t	s2u(size_t size);
-size_t	sa2u(size_t size, size_t alignment, size_t *run_size_p);
-void	malloc_write(const char *s);
-arena_t	*choose_arena(void);
-#  ifdef NO_TLS
-thread_allocated_t	*thread_allocated_get(void);
-#  endif
-#endif
-
-#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_C_))
-/* Compute the smallest power of 2 that is >= x. */
-JEMALLOC_INLINE size_t
-pow2_ceil(size_t x)
-{
-
-	x--;
-	x |= x >> 1;
-	x |= x >> 2;
-	x |= x >> 4;
-	x |= x >> 8;
-	x |= x >> 16;
-#if (LG_SIZEOF_PTR == 3)
-	x |= x >> 32;
-#endif
-	x++;
-	return (x);
-}
-
-/*
- * Compute usable size that would result from allocating an object with the
- * specified size.
- */
-JEMALLOC_INLINE size_t
-s2u(size_t size)
-{
-
-	if (size <= small_maxclass)
-		return (arenas[0]->bins[small_size2bin[size]].reg_size);
-	if (size <= arena_maxclass)
-		return (PAGE_CEILING(size));
-	return (CHUNK_CEILING(size));
-}
-
-/*
- * Compute usable size that would result from allocating an object with the
- * specified size and alignment.
- */
-JEMALLOC_INLINE size_t
-sa2u(size_t size, size_t alignment, size_t *run_size_p)
-{
-	size_t usize;
-
-	/*
-	 * Round size up to the nearest multiple of alignment.
-	 *
-	 * This done, we can take advantage of the fact that for each small
-	 * size class, every object is aligned at the smallest power of two
-	 * that is non-zero in the base two representation of the size.  For
-	 * example:
-	 *
-	 *   Size |   Base 2 | Minimum alignment
-	 *   -----+----------+------------------
-	 *     96 |  1100000 |  32
-	 *    144 | 10100000 |  32
-	 *    192 | 11000000 |  64
-	 *
-	 * Depending on runtime settings, it is possible that arena_malloc()
-	 * will further round up to a power of two, but that never causes
-	 * correctness issues.
-	 */
-	usize = (size + (alignment - 1)) & (-alignment);
-	/*
-	 * (usize < size) protects against the combination of maximal
-	 * alignment and size greater than maximal alignment.
-	 */
-	if (usize < size) {
-		/* size_t overflow. */
-		return (0);
-	}
-
-	if (usize <= arena_maxclass && alignment <= PAGE_SIZE) {
-		if (usize <= small_maxclass) {
-			return
-			    (arenas[0]->bins[small_size2bin[usize]].reg_size);
-		}
-		return (PAGE_CEILING(usize));
-	} else {
-		size_t run_size;
-
-		/*
-		 * We can't achieve subpage alignment, so round up alignment
-		 * permanently; it makes later calculations simpler.
-		 */
-		alignment = PAGE_CEILING(alignment);
-		usize = PAGE_CEILING(size);
-		/*
-		 * (usize < size) protects against very large sizes within
-		 * PAGE_SIZE of SIZE_T_MAX.
-		 *
-		 * (usize + alignment < usize) protects against the
-		 * combination of maximal alignment and usize large enough
-		 * to cause overflow.  This is similar to the first overflow
-		 * check above, but it needs to be repeated due to the new
-		 * usize value, which may now be *equal* to maximal
-		 * alignment, whereas before we only detected overflow if the
-		 * original size was *greater* than maximal alignment.
-		 */
-		if (usize < size || usize + alignment < usize) {
-			/* size_t overflow. */
-			return (0);
-		}
-
-		/*
-		 * Calculate the size of the over-size run that arena_palloc()
-		 * would need to allocate in order to guarantee the alignment.
-		 */
-		if (usize >= alignment)
-			run_size = usize + alignment - PAGE_SIZE;
-		else {
-			/*
-			 * It is possible that (alignment << 1) will cause
-			 * overflow, but it doesn't matter because we also
-			 * subtract PAGE_SIZE, which in the case of overflow
-			 * leaves us with a very large run_size.  That causes
-			 * the first conditional below to fail, which means
-			 * that the bogus run_size value never gets used for
-			 * anything important.
-			 */
-			run_size = (alignment << 1) - PAGE_SIZE;
-		}
-		if (run_size_p != NULL)
-			*run_size_p = run_size;
-
-		if (run_size <= arena_maxclass)
-			return (PAGE_CEILING(usize));
-		return (CHUNK_CEILING(usize));
-	}
-}
-
-/*
- * Wrapper around malloc_message() that avoids the need for
- * JEMALLOC_P(malloc_message)(...) throughout the code.
- */
-JEMALLOC_INLINE void
-malloc_write(const char *s)
-{
-
-	JEMALLOC_P(malloc_message)(NULL, s);
-}
-
-/*
- * Choose an arena based on a per-thread value (fast-path code, calls slow-path
- * code if necessary).
- */
-JEMALLOC_INLINE arena_t *
-choose_arena(void)
-{
-	arena_t *ret;
-
-	ret = ARENA_GET();
-	if (ret == NULL) {
-		ret = choose_arena_hard();
-		assert(ret != NULL);
-	}
-
-	return (ret);
-}
-
-#ifdef NO_TLS
-JEMALLOC_INLINE thread_allocated_t *
-thread_allocated_get(void)
-{
-	thread_allocated_t *thread_allocated = (thread_allocated_t *)
-	    pthread_getspecific(thread_allocated_tsd);
-
-	if (thread_allocated == NULL)
-		return (thread_allocated_get_hard());
-	return (thread_allocated);
-}
-#endif
-#endif
-
-#include "jemalloc/internal/rtree.h"
-#include "jemalloc/internal/tcache.h"
-#include "jemalloc/internal/arena.h"
-#include "jemalloc/internal/hash.h"
-#ifdef JEMALLOC_ZONE
-#include "jemalloc/internal/zone.h"
-#endif
-
-#ifndef JEMALLOC_ENABLE_INLINE
-void	*imalloc(size_t size);
-void	*icalloc(size_t size);
-void	*ipalloc(size_t size, size_t alignment, bool zero);
-size_t	isalloc(const void *ptr);
-#  ifdef JEMALLOC_IVSALLOC
-size_t	ivsalloc(const void *ptr);
-#  endif
-void	idalloc(void *ptr);
-void	*iralloc(void *ptr, size_t size, size_t extra, size_t alignment,
-    bool zero, bool no_move);
-#endif
-
-#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_C_))
-JEMALLOC_INLINE void *
-imalloc(size_t size)
-{
-
-	assert(size != 0);
-
-	if (size <= arena_maxclass)
-		return (arena_malloc(size, false));
-	else
-		return (huge_malloc(size, false));
-}
-
-JEMALLOC_INLINE void *
-icalloc(size_t size)
-{
-
-	if (size <= arena_maxclass)
-		return (arena_malloc(size, true));
-	else
-		return (huge_malloc(size, true));
-}
-
-JEMALLOC_INLINE void *
-ipalloc(size_t size, size_t alignment, bool zero)
-{
-	void *ret;
-	size_t usize;
-	size_t run_size
-#  ifdef JEMALLOC_CC_SILENCE
-	    = 0
-#  endif
-	    ;
-
-	usize = sa2u(size, alignment, &run_size);
-	if (usize == 0)
-		return (NULL);
-	if (usize <= arena_maxclass && alignment <= PAGE_SIZE)
-		ret = arena_malloc(usize, zero);
-	else if (run_size <= arena_maxclass) {
-		ret = arena_palloc(choose_arena(), usize, run_size, alignment,
-		    zero);
-	} else if (alignment <= chunksize)
-		ret = huge_malloc(usize, zero);
-	else
-		ret = huge_palloc(usize, alignment, zero);
-
-	assert(((uintptr_t)ret & (alignment - 1)) == 0);
-	return (ret);
-}
-
-JEMALLOC_INLINE size_t
-isalloc(const void *ptr)
-{
-	size_t ret;
-	arena_chunk_t *chunk;
-
-	assert(ptr != NULL);
-
-	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
-	if (chunk != ptr) {
-		/* Region. */
-		assert(chunk->arena->magic == ARENA_MAGIC);
-
-#ifdef JEMALLOC_PROF
-		ret = arena_salloc_demote(ptr);
-#else
-		ret = arena_salloc(ptr);
-#endif
-	} else
-		ret = huge_salloc(ptr);
-
-	return (ret);
-}
-
-#ifdef JEMALLOC_IVSALLOC
-JEMALLOC_INLINE size_t
-ivsalloc(const void *ptr)
-{
-
-	/* Return 0 if ptr is not within a chunk managed by jemalloc. */
-	if (rtree_get(chunks_rtree, (uintptr_t)CHUNK_ADDR2BASE(ptr)) == NULL)
-		return (0);
-
-	return (isalloc(ptr));
-}
-#endif
-
-JEMALLOC_INLINE void
-idalloc(void *ptr)
-{
-	arena_chunk_t *chunk;
-
-	assert(ptr != NULL);
-
-	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
-	if (chunk != ptr)
-		arena_dalloc(chunk->arena, chunk, ptr);
-	else
-		huge_dalloc(ptr, true);
-}
-
-JEMALLOC_INLINE void *
-iralloc(void *ptr, size_t size, size_t extra, size_t alignment, bool zero,
-    bool no_move)
-{
-	void *ret;
-	size_t oldsize;
-
-	assert(ptr != NULL);
-	assert(size != 0);
-
-	oldsize = isalloc(ptr);
-
-	if (alignment != 0 && ((uintptr_t)ptr & ((uintptr_t)alignment-1))
-	    != 0) {
-		size_t copysize;
-
-		/*
-		 * Existing object alignment is inadquate; allocate new space
-		 * and copy.
-		 */
-		if (no_move)
-			return (NULL);
-		ret = ipalloc(size + extra, alignment, zero);
-		if (ret == NULL) {
-			if (extra == 0)
-				return (NULL);
-			/* Try again, without extra this time. */
-			ret = ipalloc(size, alignment, zero);
-			if (ret == NULL)
-				return (NULL);
-		}
-		/*
-		 * Copy at most size bytes (not size+extra), since the caller
-		 * has no expectation that the extra bytes will be reliably
-		 * preserved.
-		 */
-		copysize = (size < oldsize) ? size : oldsize;
-		memcpy(ret, ptr, copysize);
-		idalloc(ptr);
-		return (ret);
-	}
-
-	if (no_move) {
-		if (size <= arena_maxclass) {
-			return (arena_ralloc_no_move(ptr, oldsize, size,
-			    extra, zero));
-		} else {
-			return (huge_ralloc_no_move(ptr, oldsize, size,
-			    extra));
-		}
-	} else {
-		if (size + extra <= arena_maxclass) {
-			return (arena_ralloc(ptr, oldsize, size, extra,
-			    alignment, zero));
-		} else {
-			return (huge_ralloc(ptr, oldsize, size, extra,
-			    alignment, zero));
-		}
-	}
-}
-#endif
-
-#include "jemalloc/internal/prof.h"
-
-#undef JEMALLOC_H_INLINES
-/******************************************************************************/
--- a/jemalloc/include/jemalloc/internal/mutex.h
+++ b/jemalloc/include/jemalloc/internal/mutex.h
@@ -1,68 +0,0 @@
-/******************************************************************************/
-#ifdef JEMALLOC_H_TYPES
-
-typedef pthread_mutex_t malloc_mutex_t;
-
-#ifdef PTHREAD_ADAPTIVE_MUTEX_INITIALIZER_NP
-#  define MALLOC_MUTEX_INITIALIZER PTHREAD_ADAPTIVE_MUTEX_INITIALIZER_NP
-#else
-#  define MALLOC_MUTEX_INITIALIZER PTHREAD_MUTEX_INITIALIZER
-#endif
-
-#endif /* JEMALLOC_H_TYPES */
-/******************************************************************************/
-#ifdef JEMALLOC_H_STRUCTS
-
-#endif /* JEMALLOC_H_STRUCTS */
-/******************************************************************************/
-#ifdef JEMALLOC_H_EXTERNS
-
-#ifdef JEMALLOC_LAZY_LOCK
-extern bool isthreaded;
-#else
-#  define isthreaded true
-#endif
-
-bool	malloc_mutex_init(malloc_mutex_t *mutex);
-void	malloc_mutex_destroy(malloc_mutex_t *mutex);
-
-#endif /* JEMALLOC_H_EXTERNS */
-/******************************************************************************/
-#ifdef JEMALLOC_H_INLINES
-
-#ifndef JEMALLOC_ENABLE_INLINE
-void	malloc_mutex_lock(malloc_mutex_t *mutex);
-bool	malloc_mutex_trylock(malloc_mutex_t *mutex);
-void	malloc_mutex_unlock(malloc_mutex_t *mutex);
-#endif
-
-#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_MUTEX_C_))
-JEMALLOC_INLINE void
-malloc_mutex_lock(malloc_mutex_t *mutex)
-{
-
-	if (isthreaded)
-		pthread_mutex_lock(mutex);
-}
-
-JEMALLOC_INLINE bool
-malloc_mutex_trylock(malloc_mutex_t *mutex)
-{
-
-	if (isthreaded)
-		return (pthread_mutex_trylock(mutex) != 0);
-	else
-		return (false);
-}
-
-JEMALLOC_INLINE void
-malloc_mutex_unlock(malloc_mutex_t *mutex)
-{
-
-	if (isthreaded)
-		pthread_mutex_unlock(mutex);
-}
-#endif
-
-#endif /* JEMALLOC_H_INLINES */
-/******************************************************************************/
--- a/jemalloc/include/jemalloc/internal/tcache.h
+++ b/jemalloc/include/jemalloc/internal/tcache.h
@@ -1,394 +0,0 @@
-#ifdef JEMALLOC_TCACHE
-/******************************************************************************/
-#ifdef JEMALLOC_H_TYPES
-
-typedef struct tcache_bin_s tcache_bin_t;
-typedef struct tcache_s tcache_t;
-
-/*
- * Absolute maximum number of cache slots for each small bin in the thread
- * cache.  This is an additional constraint beyond that imposed as: twice the
- * number of regions per run for this size class.
- *
- * This constant must be an even number.
- */
-#define	TCACHE_NSLOTS_SMALL_MAX		200
-
-/* Number of cache slots for large size classes. */
-#define	TCACHE_NSLOTS_LARGE		20
-
-/* (1U << opt_lg_tcache_max) is used to compute tcache_maxclass. */
-#define	LG_TCACHE_MAXCLASS_DEFAULT	15
-
-/*
- * (1U << opt_lg_tcache_gc_sweep) is the approximate number of allocation
- * events between full GC sweeps (-1: disabled).  Integer rounding may cause
- * the actual number to be slightly higher, since GC is performed
- * incrementally.
- */
-#define	LG_TCACHE_GC_SWEEP_DEFAULT	13
-
-#endif /* JEMALLOC_H_TYPES */
-/******************************************************************************/
-#ifdef JEMALLOC_H_STRUCTS
-
-struct tcache_bin_s {
-#  ifdef JEMALLOC_STATS
-	tcache_bin_stats_t tstats;
-#  endif
-	unsigned	low_water;	/* Min # cached since last GC. */
-	unsigned	ncached;	/* # of cached objects. */
-	unsigned	ncached_max;	/* Upper limit on ncached. */
-	void		*avail;		/* Chain of available objects. */
-};
-
-struct tcache_s {
-#  ifdef JEMALLOC_STATS
-	ql_elm(tcache_t) link;		/* Used for aggregating stats. */
-#  endif
-#  ifdef JEMALLOC_PROF
-	uint64_t	prof_accumbytes;/* Cleared after arena_prof_accum() */
-#  endif
-	arena_t		*arena;		/* This thread's arena. */
-	unsigned	ev_cnt;		/* Event count since incremental GC. */
-	unsigned	next_gc_bin;	/* Next bin to GC. */
-	tcache_bin_t	tbins[1];	/* Dynamically sized. */
-};
-
-#endif /* JEMALLOC_H_STRUCTS */
-/******************************************************************************/
-#ifdef JEMALLOC_H_EXTERNS
-
-extern bool	opt_tcache;
-extern ssize_t	opt_lg_tcache_max;
-extern ssize_t	opt_lg_tcache_gc_sweep;
-
-/* Map of thread-specific caches. */
-#ifndef NO_TLS
-extern __thread tcache_t	*tcache_tls
-    JEMALLOC_ATTR(tls_model("initial-exec"));
-#  define TCACHE_GET()	tcache_tls
-#  define TCACHE_SET(v)	do {						\
-	tcache_tls = (tcache_t *)(v);					\
-	pthread_setspecific(tcache_tsd, (void *)(v));			\
-} while (0)
-#else
-#  define TCACHE_GET()	((tcache_t *)pthread_getspecific(tcache_tsd))
-#  define TCACHE_SET(v)	do {						\
-	pthread_setspecific(tcache_tsd, (void *)(v));			\
-} while (0)
-#endif
-extern pthread_key_t		tcache_tsd;
-
-/*
- * Number of tcache bins.  There are nbins small-object bins, plus 0 or more
- * large-object bins.
- */
-extern size_t			nhbins;
-
-/* Maximum cached size class. */
-extern size_t			tcache_maxclass;
-
-/* Number of tcache allocation/deallocation events between incremental GCs. */
-extern unsigned			tcache_gc_incr;
-
-void	tcache_bin_flush_small(tcache_bin_t *tbin, size_t binind, unsigned rem
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-    , tcache_t *tcache
-#endif
-    );
-void	tcache_bin_flush_large(tcache_bin_t *tbin, size_t binind, unsigned rem
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-    , tcache_t *tcache
-#endif
-    );
-tcache_t *tcache_create(arena_t *arena);
-void	*tcache_alloc_small_hard(tcache_t *tcache, tcache_bin_t *tbin,
-    size_t binind);
-void	tcache_destroy(tcache_t *tcache);
-#ifdef JEMALLOC_STATS
-void	tcache_stats_merge(tcache_t *tcache, arena_t *arena);
-#endif
-void	tcache_boot(void);
-
-#endif /* JEMALLOC_H_EXTERNS */
-/******************************************************************************/
-#ifdef JEMALLOC_H_INLINES
-
-#ifndef JEMALLOC_ENABLE_INLINE
-void	tcache_event(tcache_t *tcache);
-tcache_t *tcache_get(void);
-void	*tcache_alloc_easy(tcache_bin_t *tbin);
-void	*tcache_alloc_small(tcache_t *tcache, size_t size, bool zero);
-void	*tcache_alloc_large(tcache_t *tcache, size_t size, bool zero);
-void	tcache_dalloc_small(tcache_t *tcache, void *ptr);
-void	tcache_dalloc_large(tcache_t *tcache, void *ptr, size_t size);
-#endif
-
-#if (defined(JEMALLOC_ENABLE_INLINE) || defined(JEMALLOC_TCACHE_C_))
-JEMALLOC_INLINE tcache_t *
-tcache_get(void)
-{
-	tcache_t *tcache;
-
-	if ((isthreaded & opt_tcache) == false)
-		return (NULL);
-
-	tcache = TCACHE_GET();
-	if ((uintptr_t)tcache <= (uintptr_t)2) {
-		if (tcache == NULL) {
-			tcache = tcache_create(choose_arena());
-			if (tcache == NULL)
-				return (NULL);
-		} else {
-			if (tcache == (void *)(uintptr_t)1) {
-				/*
-				 * Make a note that an allocator function was
-				 * called after the tcache_thread_cleanup() was
-				 * called.
-				 */
-				TCACHE_SET((uintptr_t)2);
-			}
-			return (NULL);
-		}
-	}
-
-	return (tcache);
-}
-
-JEMALLOC_INLINE void
-tcache_event(tcache_t *tcache)
-{
-
-	if (tcache_gc_incr == 0)
-		return;
-
-	tcache->ev_cnt++;
-	assert(tcache->ev_cnt <= tcache_gc_incr);
-	if (tcache->ev_cnt == tcache_gc_incr) {
-		size_t binind = tcache->next_gc_bin;
-		tcache_bin_t *tbin = &tcache->tbins[binind];
-
-		if (tbin->low_water > 0) {
-			/*
-			 * Flush (ceiling) 3/4 of the objects below the low
-			 * water mark.
-			 */
-			if (binind < nbins) {
-				tcache_bin_flush_small(tbin, binind,
-				    tbin->ncached - tbin->low_water +
-				    (tbin->low_water >> 2)
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-				    , tcache
-#endif
-				    );
-			} else {
-				tcache_bin_flush_large(tbin, binind,
-				    tbin->ncached - tbin->low_water +
-				    (tbin->low_water >> 2)
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-				    , tcache
-#endif
-				    );
-			}
-		}
-		tbin->low_water = tbin->ncached;
-
-		tcache->next_gc_bin++;
-		if (tcache->next_gc_bin == nhbins)
-			tcache->next_gc_bin = 0;
-		tcache->ev_cnt = 0;
-	}
-}
-
-JEMALLOC_INLINE void *
-tcache_alloc_easy(tcache_bin_t *tbin)
-{
-	void *ret;
-
-	if (tbin->ncached == 0)
-		return (NULL);
-	tbin->ncached--;
-	if (tbin->ncached < tbin->low_water)
-		tbin->low_water = tbin->ncached;
-	ret = tbin->avail;
-	tbin->avail = *(void **)ret;
-	return (ret);
-}
-
-JEMALLOC_INLINE void *
-tcache_alloc_small(tcache_t *tcache, size_t size, bool zero)
-{
-	void *ret;
-	size_t binind;
-	tcache_bin_t *tbin;
-
-	binind = small_size2bin[size];
-	assert(binind < nbins);
-	tbin = &tcache->tbins[binind];
-	ret = tcache_alloc_easy(tbin);
-	if (ret == NULL) {
-		ret = tcache_alloc_small_hard(tcache, tbin, binind);
-		if (ret == NULL)
-			return (NULL);
-	}
-	assert(arena_salloc(ret) == tcache->arena->bins[binind].reg_size);
-
-	if (zero == false) {
-#ifdef JEMALLOC_FILL
-		if (opt_junk)
-			memset(ret, 0xa5, size);
-		else if (opt_zero)
-			memset(ret, 0, size);
-#endif
-	} else
-		memset(ret, 0, size);
-
-#ifdef JEMALLOC_STATS
-	tbin->tstats.nrequests++;
-#endif
-#ifdef JEMALLOC_PROF
-	tcache->prof_accumbytes += tcache->arena->bins[binind].reg_size;
-#endif
-	tcache_event(tcache);
-	return (ret);
-}
-
-JEMALLOC_INLINE void *
-tcache_alloc_large(tcache_t *tcache, size_t size, bool zero)
-{
-	void *ret;
-	size_t binind;
-	tcache_bin_t *tbin;
-
-	size = PAGE_CEILING(size);
-	assert(size <= tcache_maxclass);
-	binind = nbins + (size >> PAGE_SHIFT) - 1;
-	assert(binind < nhbins);
-	tbin = &tcache->tbins[binind];
-	ret = tcache_alloc_easy(tbin);
-	if (ret == NULL) {
-		/*
-		 * Only allocate one large object at a time, because it's quite
-		 * expensive to create one and not use it.
-		 */
-		ret = arena_malloc_large(tcache->arena, size, zero);
-		if (ret == NULL)
-			return (NULL);
-	} else {
-#ifdef JEMALLOC_PROF
-		arena_chunk_t *chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ret);
-		size_t pageind = (((uintptr_t)ret - (uintptr_t)chunk) >>
-		    PAGE_SHIFT);
-		chunk->map[pageind-map_bias].bits &= ~CHUNK_MAP_CLASS_MASK;
-#endif
-		if (zero == false) {
-#ifdef JEMALLOC_FILL
-			if (opt_junk)
-				memset(ret, 0xa5, size);
-			else if (opt_zero)
-				memset(ret, 0, size);
-#endif
-		} else
-			memset(ret, 0, size);
-
-#ifdef JEMALLOC_STATS
-		tbin->tstats.nrequests++;
-#endif
-#ifdef JEMALLOC_PROF
-		tcache->prof_accumbytes += size;
-#endif
-	}
-
-	tcache_event(tcache);
-	return (ret);
-}
-
-JEMALLOC_INLINE void
-tcache_dalloc_small(tcache_t *tcache, void *ptr)
-{
-	arena_t *arena;
-	arena_chunk_t *chunk;
-	arena_run_t *run;
-	arena_bin_t *bin;
-	tcache_bin_t *tbin;
-	size_t pageind, binind;
-	arena_chunk_map_t *mapelm;
-
-	assert(arena_salloc(ptr) <= small_maxclass);
-
-	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
-	arena = chunk->arena;
-	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> PAGE_SHIFT;
-	mapelm = &chunk->map[pageind-map_bias];
-	run = (arena_run_t *)((uintptr_t)chunk + (uintptr_t)((pageind -
-	    (mapelm->bits >> PAGE_SHIFT)) << PAGE_SHIFT));
-	assert(run->magic == ARENA_RUN_MAGIC);
-	bin = run->bin;
-	binind = ((uintptr_t)bin - (uintptr_t)&arena->bins) /
-	    sizeof(arena_bin_t);
-	assert(binind < nbins);
-
-#ifdef JEMALLOC_FILL
-	if (opt_junk)
-		memset(ptr, 0x5a, bin->reg_size);
-#endif
-
-	tbin = &tcache->tbins[binind];
-	if (tbin->ncached == tbin->ncached_max) {
-		tcache_bin_flush_small(tbin, binind, (tbin->ncached_max >> 1)
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-		    , tcache
-#endif
-		    );
-	}
-	assert(tbin->ncached < tbin->ncached_max);
-	*(void **)ptr = tbin->avail;
-	tbin->avail = ptr;
-	tbin->ncached++;
-
-	tcache_event(tcache);
-}
-
-JEMALLOC_INLINE void
-tcache_dalloc_large(tcache_t *tcache, void *ptr, size_t size)
-{
-	arena_t *arena;
-	arena_chunk_t *chunk;
-	size_t pageind, binind;
-	tcache_bin_t *tbin;
-
-	assert((size & PAGE_MASK) == 0);
-	assert(arena_salloc(ptr) > small_maxclass);
-	assert(arena_salloc(ptr) <= tcache_maxclass);
-
-	chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
-	arena = chunk->arena;
-	pageind = ((uintptr_t)ptr - (uintptr_t)chunk) >> PAGE_SHIFT;
-	binind = nbins + (size >> PAGE_SHIFT) - 1;
-
-#ifdef JEMALLOC_FILL
-	if (opt_junk)
-		memset(ptr, 0x5a, size);
-#endif
-
-	tbin = &tcache->tbins[binind];
-	if (tbin->ncached == tbin->ncached_max) {
-		tcache_bin_flush_large(tbin, binind, (tbin->ncached_max >> 1)
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-		    , tcache
-#endif
-		    );
-	}
-	assert(tbin->ncached < tbin->ncached_max);
-	*(void **)ptr = tbin->avail;
-	tbin->avail = ptr;
-	tbin->ncached++;
-
-	tcache_event(tcache);
-}
-#endif
-
-#endif /* JEMALLOC_H_INLINES */
-/******************************************************************************/
-#endif /* JEMALLOC_TCACHE */
--- a/jemalloc/include/jemalloc/jemalloc.h.in
+++ b/jemalloc/include/jemalloc/jemalloc.h.in
@@ -1,66 +0,0 @@
-#ifndef JEMALLOC_H_
-#define	JEMALLOC_H_
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-#include <limits.h>
-#include <strings.h>
-
-#define	JEMALLOC_VERSION "@jemalloc_version@"
-#define	JEMALLOC_VERSION_MAJOR @jemalloc_version_major@
-#define	JEMALLOC_VERSION_MINOR @jemalloc_version_minor@
-#define	JEMALLOC_VERSION_BUGFIX @jemalloc_version_bugfix@
-#define	JEMALLOC_VERSION_NREV @jemalloc_version_nrev@
-#define	JEMALLOC_VERSION_GID "@jemalloc_version_gid@"
-
-#include "jemalloc_defs@install_suffix@.h"
-#ifndef JEMALLOC_P
-#  define JEMALLOC_P(s) s
-#endif
-
-#define	ALLOCM_LG_ALIGN(la)	(la)
-#if LG_SIZEOF_PTR == 2
-#define	ALLOCM_ALIGN(a)	(ffs(a)-1)
-#else
-#define	ALLOCM_ALIGN(a)	((a < (size_t)INT_MAX) ? ffs(a)-1 : ffs(a>>32)+31)
-#endif
-#define	ALLOCM_ZERO	((int)0x40)
-#define	ALLOCM_NO_MOVE	((int)0x80)
-
-#define	ALLOCM_SUCCESS		0
-#define	ALLOCM_ERR_OOM		1
-#define	ALLOCM_ERR_NOT_MOVED	2
-
-extern const char	*JEMALLOC_P(malloc_conf);
-extern void		(*JEMALLOC_P(malloc_message))(void *, const char *);
-
-void	*JEMALLOC_P(malloc)(size_t size) JEMALLOC_ATTR(malloc);
-void	*JEMALLOC_P(calloc)(size_t num, size_t size) JEMALLOC_ATTR(malloc);
-int	JEMALLOC_P(posix_memalign)(void **memptr, size_t alignment, size_t size)
-    JEMALLOC_ATTR(nonnull(1));
-void	*JEMALLOC_P(realloc)(void *ptr, size_t size);
-void	JEMALLOC_P(free)(void *ptr);
-
-size_t	JEMALLOC_P(malloc_usable_size)(const void *ptr);
-void	JEMALLOC_P(malloc_stats_print)(void (*write_cb)(void *, const char *),
-    void *cbopaque, const char *opts);
-int	JEMALLOC_P(mallctl)(const char *name, void *oldp, size_t *oldlenp,
-    void *newp, size_t newlen);
-int	JEMALLOC_P(mallctlnametomib)(const char *name, size_t *mibp,
-    size_t *miblenp);
-int	JEMALLOC_P(mallctlbymib)(const size_t *mib, size_t miblen, void *oldp,
-    size_t *oldlenp, void *newp, size_t newlen);
-
-int	JEMALLOC_P(allocm)(void **ptr, size_t *rsize, size_t size, int flags)
-    JEMALLOC_ATTR(nonnull(1));
-int	JEMALLOC_P(rallocm)(void **ptr, size_t *rsize, size_t size,
-    size_t extra, int flags) JEMALLOC_ATTR(nonnull(1));
-int	JEMALLOC_P(sallocm)(const void *ptr, size_t *rsize, int flags)
-    JEMALLOC_ATTR(nonnull(1));
-int	JEMALLOC_P(dallocm)(void *ptr, int flags) JEMALLOC_ATTR(nonnull(1));
-
-#ifdef __cplusplus
-};
-#endif
-#endif /* JEMALLOC_H_ */
--- a/jemalloc/include/jemalloc/jemalloc_defs.h.in
+++ b/jemalloc/include/jemalloc/jemalloc_defs.h.in
@@ -1,140 +0,0 @@
-#ifndef JEMALLOC_DEFS_H_
-#define	JEMALLOC_DEFS_H_
-
-/*
- * If JEMALLOC_PREFIX is defined, it will cause all public APIs to be prefixed.
- * This makes it possible, with some care, to use multiple allocators
- * simultaneously.
- *
- * In many cases it is more convenient to manually prefix allocator function
- * calls than to let macros do it automatically, particularly when using
- * multiple allocators simultaneously.  Define JEMALLOC_MANGLE before
- * #include'ing jemalloc.h in order to cause name mangling that corresponds to
- * the API prefixing.
- */
-#undef JEMALLOC_PREFIX
-#undef JEMALLOC_CPREFIX
-#if (defined(JEMALLOC_PREFIX) && defined(JEMALLOC_MANGLE))
-#undef JEMALLOC_P
-#endif
-
-/*
- * Hyper-threaded CPUs may need a special instruction inside spin loops in
- * order to yield to another virtual CPU.
- */
-#undef CPU_SPINWAIT
-
-/* Defined if __attribute__((...)) syntax is supported. */
-#undef JEMALLOC_HAVE_ATTR
-#ifdef JEMALLOC_HAVE_ATTR
-#  define JEMALLOC_ATTR(s) __attribute__((s))
-#else
-#  define JEMALLOC_ATTR(s)
-#endif
-
-/* JEMALLOC_CC_SILENCE enables code that silences unuseful compiler warnings. */
-#undef JEMALLOC_CC_SILENCE
-
-/*
- * JEMALLOC_DEBUG enables assertions and other sanity checks, and disables
- * inline functions.
- */
-#undef JEMALLOC_DEBUG
-
-/* JEMALLOC_STATS enables statistics calculation. */
-#undef JEMALLOC_STATS
-
-/* JEMALLOC_PROF enables allocation profiling. */
-#undef JEMALLOC_PROF
-
-/* Use libunwind for profile backtracing if defined. */
-#undef JEMALLOC_PROF_LIBUNWIND
-
-/* Use libgcc for profile backtracing if defined. */
-#undef JEMALLOC_PROF_LIBGCC
-
-/*
- * JEMALLOC_TINY enables support for tiny objects, which are smaller than one
- * quantum.
- */
-#undef JEMALLOC_TINY
-
-/*
- * JEMALLOC_TCACHE enables a thread-specific caching layer for small objects.
- * This makes it possible to allocate/deallocate objects without any locking
- * when the cache is in the steady state.
- */
-#undef JEMALLOC_TCACHE
-
-/*
- * JEMALLOC_DSS enables use of sbrk(2) to allocate chunks from the data storage
- * segment (DSS).
- */
-#undef JEMALLOC_DSS
-
-/* JEMALLOC_SWAP enables mmap()ed swap file support. */
-#undef JEMALLOC_SWAP
-
-/* Support memory filling (junk/zero). */
-#undef JEMALLOC_FILL
-
-/* Support optional abort() on OOM. */
-#undef JEMALLOC_XMALLOC
-
-/* Support SYSV semantics. */
-#undef JEMALLOC_SYSV
-
-/* Support lazy locking (avoid locking unless a second thread is launched). */
-#undef JEMALLOC_LAZY_LOCK
-
-/* Determine page size at run time if defined. */
-#undef DYNAMIC_PAGE_SHIFT
-
-/* One page is 2^STATIC_PAGE_SHIFT bytes. */
-#undef STATIC_PAGE_SHIFT
-
-/* TLS is used to map arenas and magazine caches to threads. */
-#undef NO_TLS
-
-/*
- * JEMALLOC_IVSALLOC enables ivsalloc(), which verifies that pointers reside
- * within jemalloc-owned chunks before dereferencing them.
- */
-#undef JEMALLOC_IVSALLOC
-
-/*
- * Define overrides for non-standard allocator-related functions if they
- * are present on the system.
- */
-#undef JEMALLOC_OVERRIDE_MEMALIGN
-#undef JEMALLOC_OVERRIDE_VALLOC
-
-/*
- * Darwin (OS X) uses zones to work around Mach-O symbol override shortcomings.
- */
-#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.
- *
- *   madvise(..., MADV_DONTNEED) : On Linux, this immediately discards pages,
- *                                 such that new pages will be demand-zeroed if
- *                                 the address region is later touched.
- *   madvise(..., MADV_FREE) : On FreeBSD and Darwin, this marks pages as being
- *                             unused, such that they will be discarded rather
- *                             than swapped out.
- */
-#undef JEMALLOC_PURGE_MADVISE_DONTNEED
-#undef JEMALLOC_PURGE_MADVISE_FREE
-
-/* sizeof(void *) == 2^LG_SIZEOF_PTR. */
-#undef LG_SIZEOF_PTR
-
-/* sizeof(int) == 2^LG_SIZEOF_INT. */
-#undef LG_SIZEOF_INT
-
-#endif /* JEMALLOC_DEFS_H_ */
--- a/jemalloc/src/arena.c
+++ b/jemalloc/src/arena.c
--- a/jemalloc/src/chunk.c
+++ b/jemalloc/src/chunk.c
@@ -1,171 +0,0 @@
-#define	JEMALLOC_CHUNK_C_
-#include "jemalloc/internal/jemalloc_internal.h"
-
-/******************************************************************************/
-/* Data. */
-
-size_t	opt_lg_chunk = LG_CHUNK_DEFAULT;
-#ifdef JEMALLOC_SWAP
-bool	opt_overcommit = true;
-#endif
-
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-malloc_mutex_t	chunks_mtx;
-chunk_stats_t	stats_chunks;
-#endif
-
-#ifdef JEMALLOC_IVSALLOC
-rtree_t		*chunks_rtree;
-#endif
-
-/* Various chunk-related settings. */
-size_t		chunksize;
-size_t		chunksize_mask; /* (chunksize - 1). */
-size_t		chunk_npages;
-size_t		map_bias;
-size_t		arena_maxclass; /* Max size class for arenas. */
-
-/******************************************************************************/
-
-/*
- * If the caller specifies (*zero == false), it is still possible to receive
- * zeroed memory, in which case *zero is toggled to true.  arena_chunk_alloc()
- * takes advantage of this to avoid demanding zeroed chunks, but taking
- * advantage of them if they are returned.
- */
-void *
-chunk_alloc(size_t size, bool base, bool *zero)
-{
-	void *ret;
-
-	assert(size != 0);
-	assert((size & chunksize_mask) == 0);
-
-#ifdef JEMALLOC_SWAP
-	if (swap_enabled) {
-		ret = chunk_alloc_swap(size, zero);
-		if (ret != NULL)
-			goto RETURN;
-	}
-
-	if (swap_enabled == false || opt_overcommit) {
-#endif
-#ifdef JEMALLOC_DSS
-		ret = chunk_alloc_dss(size, zero);
-		if (ret != NULL)
-			goto RETURN;
-#endif
-		ret = chunk_alloc_mmap(size);
-		if (ret != NULL) {
-			*zero = true;
-			goto RETURN;
-		}
-#ifdef JEMALLOC_SWAP
-	}
-#endif
-
-	/* All strategies for allocation failed. */
-	ret = NULL;
-RETURN:
-#ifdef JEMALLOC_IVSALLOC
-	if (base == false && ret != NULL) {
-		if (rtree_set(chunks_rtree, (uintptr_t)ret, ret)) {
-			chunk_dealloc(ret, size);
-			return (NULL);
-		}
-	}
-#endif
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-	if (ret != NULL) {
-#  ifdef JEMALLOC_PROF
-		bool gdump;
-#  endif
-		malloc_mutex_lock(&chunks_mtx);
-#  ifdef JEMALLOC_STATS
-		stats_chunks.nchunks += (size / chunksize);
-#  endif
-		stats_chunks.curchunks += (size / chunksize);
-		if (stats_chunks.curchunks > stats_chunks.highchunks) {
-			stats_chunks.highchunks = stats_chunks.curchunks;
-#  ifdef JEMALLOC_PROF
-			gdump = true;
-#  endif
-		}
-#  ifdef JEMALLOC_PROF
-		else
-			gdump = false;
-#  endif
-		malloc_mutex_unlock(&chunks_mtx);
-#  ifdef JEMALLOC_PROF
-		if (opt_prof && opt_prof_gdump && gdump)
-			prof_gdump();
-#  endif
-	}
-#endif
-
-	assert(CHUNK_ADDR2BASE(ret) == ret);
-	return (ret);
-}
-
-void
-chunk_dealloc(void *chunk, size_t size)
-{
-
-	assert(chunk != NULL);
-	assert(CHUNK_ADDR2BASE(chunk) == chunk);
-	assert(size != 0);
-	assert((size & chunksize_mask) == 0);
-
-#ifdef JEMALLOC_IVSALLOC
-	rtree_set(chunks_rtree, (uintptr_t)chunk, NULL);
-#endif
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-	malloc_mutex_lock(&chunks_mtx);
-	stats_chunks.curchunks -= (size / chunksize);
-	malloc_mutex_unlock(&chunks_mtx);
-#endif
-
-#ifdef JEMALLOC_SWAP
-	if (swap_enabled && chunk_dealloc_swap(chunk, size) == false)
-		return;
-#endif
-#ifdef JEMALLOC_DSS
-	if (chunk_dealloc_dss(chunk, size) == false)
-		return;
-#endif
-	chunk_dealloc_mmap(chunk, size);
-}
-
-bool
-chunk_boot(void)
-{
-
-	/* Set variables according to the value of opt_lg_chunk. */
-	chunksize = (ZU(1) << opt_lg_chunk);
-	assert(chunksize >= PAGE_SIZE);
-	chunksize_mask = chunksize - 1;
-	chunk_npages = (chunksize >> PAGE_SHIFT);
-
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-	if (malloc_mutex_init(&chunks_mtx))
-		return (true);
-	memset(&stats_chunks, 0, sizeof(chunk_stats_t));
-#endif
-#ifdef JEMALLOC_SWAP
-	if (chunk_swap_boot())
-		return (true);
-#endif
-	if (chunk_mmap_boot())
-		return (true);
-#ifdef JEMALLOC_DSS
-	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);
-}
--- a/jemalloc/src/chunk_dss.c
+++ b/jemalloc/src/chunk_dss.c
@@ -1,284 +0,0 @@
-#define	JEMALLOC_CHUNK_DSS_C_
-#include "jemalloc/internal/jemalloc_internal.h"
-#ifdef JEMALLOC_DSS
-/******************************************************************************/
-/* Data. */
-
-malloc_mutex_t	dss_mtx;
-
-/* Base address of the DSS. */
-static void	*dss_base;
-/* Current end of the DSS, or ((void *)-1) if the DSS is exhausted. */
-static void	*dss_prev;
-/* Current upper limit on DSS addresses. */
-static void	*dss_max;
-
-/*
- * Trees of chunks that were previously allocated (trees differ only in node
- * ordering).  These are used when allocating chunks, in an attempt to re-use
- * address space.  Depending on function, different tree orderings are needed,
- * which is why there are two trees with the same contents.
- */
-static extent_tree_t	dss_chunks_szad;
-static extent_tree_t	dss_chunks_ad;
-
-/******************************************************************************/
-/* Function prototypes for non-inline static functions. */
-
-static void	*chunk_recycle_dss(size_t size, bool *zero);
-static extent_node_t *chunk_dealloc_dss_record(void *chunk, size_t size);
-
-/******************************************************************************/
-
-static void *
-chunk_recycle_dss(size_t size, bool *zero)
-{
-	extent_node_t *node, key;
-
-	key.addr = NULL;
-	key.size = size;
-	malloc_mutex_lock(&dss_mtx);
-	node = extent_tree_szad_nsearch(&dss_chunks_szad, &key);
-	if (node != NULL) {
-		void *ret = node->addr;
-
-		/* Remove node from the tree. */
-		extent_tree_szad_remove(&dss_chunks_szad, node);
-		if (node->size == size) {
-			extent_tree_ad_remove(&dss_chunks_ad, node);
-			base_node_dealloc(node);
-		} else {
-			/*
-			 * Insert the remainder of node's address range as a
-			 * smaller chunk.  Its position within dss_chunks_ad
-			 * does not change.
-			 */
-			assert(node->size > size);
-			node->addr = (void *)((uintptr_t)node->addr + size);
-			node->size -= size;
-			extent_tree_szad_insert(&dss_chunks_szad, node);
-		}
-		malloc_mutex_unlock(&dss_mtx);
-
-		if (*zero)
-			memset(ret, 0, size);
-		return (ret);
-	}
-	malloc_mutex_unlock(&dss_mtx);
-
-	return (NULL);
-}
-
-void *
-chunk_alloc_dss(size_t size, bool *zero)
-{
-	void *ret;
-
-	ret = chunk_recycle_dss(size, zero);
-	if (ret != NULL)
-		return (ret);
-
-	/*
-	 * sbrk() uses a signed increment argument, so take care not to
-	 * interpret a huge allocation request as a negative increment.
-	 */
-	if ((intptr_t)size < 0)
-		return (NULL);
-
-	malloc_mutex_lock(&dss_mtx);
-	if (dss_prev != (void *)-1) {
-		intptr_t incr;
-
-		/*
-		 * The loop is necessary to recover from races with other
-		 * threads that are using the DSS for something other than
-		 * malloc.
-		 */
-		do {
-			/* Get the current end of the DSS. */
-			dss_max = sbrk(0);
-
-			/*
-			 * Calculate how much padding is necessary to
-			 * chunk-align the end of the DSS.
-			 */
-			incr = (intptr_t)size
-			    - (intptr_t)CHUNK_ADDR2OFFSET(dss_max);
-			if (incr == (intptr_t)size)
-				ret = dss_max;
-			else {
-				ret = (void *)((intptr_t)dss_max + incr);
-				incr += size;
-			}
-
-			dss_prev = sbrk(incr);
-			if (dss_prev == dss_max) {
-				/* Success. */
-				dss_max = (void *)((intptr_t)dss_prev + incr);
-				malloc_mutex_unlock(&dss_mtx);
-				*zero = true;
-				return (ret);
-			}
-		} while (dss_prev != (void *)-1);
-	}
-	malloc_mutex_unlock(&dss_mtx);
-
-	return (NULL);
-}
-
-static extent_node_t *
-chunk_dealloc_dss_record(void *chunk, size_t size)
-{
-	extent_node_t *xnode, *node, *prev, key;
-
-	xnode = NULL;
-	while (true) {
-		key.addr = (void *)((uintptr_t)chunk + size);
-		node = extent_tree_ad_nsearch(&dss_chunks_ad, &key);
-		/* Try to coalesce forward. */
-		if (node != NULL && node->addr == key.addr) {
-			/*
-			 * Coalesce chunk with the following address range.
-			 * This does not change the position within
-			 * dss_chunks_ad, so only remove/insert from/into
-			 * dss_chunks_szad.
-			 */
-			extent_tree_szad_remove(&dss_chunks_szad, node);
-			node->addr = chunk;
-			node->size += size;
-			extent_tree_szad_insert(&dss_chunks_szad, node);
-			break;
-		} else if (xnode == NULL) {
-			/*
-			 * It is possible that base_node_alloc() will cause a
-			 * new base chunk to be allocated, so take care not to
-			 * deadlock on dss_mtx, and recover if another thread
-			 * deallocates an adjacent chunk while this one is busy
-			 * allocating xnode.
-			 */
-			malloc_mutex_unlock(&dss_mtx);
-			xnode = base_node_alloc();
-			malloc_mutex_lock(&dss_mtx);
-			if (xnode == NULL)
-				return (NULL);
-		} else {
-			/* Coalescing forward failed, so insert a new node. */
-			node = xnode;
-			xnode = NULL;
-			node->addr = chunk;
-			node->size = size;
-			extent_tree_ad_insert(&dss_chunks_ad, node);
-			extent_tree_szad_insert(&dss_chunks_szad, node);
-			break;
-		}
-	}
-	/* Discard xnode if it ended up unused do to a race. */
-	if (xnode != NULL)
-		base_node_dealloc(xnode);
-
-	/* Try to coalesce backward. */
-	prev = extent_tree_ad_prev(&dss_chunks_ad, node);
-	if (prev != NULL && (void *)((uintptr_t)prev->addr + prev->size) ==
-	    chunk) {
-		/*
-		 * Coalesce chunk with the previous address range.  This does
-		 * not change the position within dss_chunks_ad, so only
-		 * remove/insert node from/into dss_chunks_szad.
-		 */
-		extent_tree_szad_remove(&dss_chunks_szad, prev);
-		extent_tree_ad_remove(&dss_chunks_ad, prev);
-
-		extent_tree_szad_remove(&dss_chunks_szad, node);
-		node->addr = prev->addr;
-		node->size += prev->size;
-		extent_tree_szad_insert(&dss_chunks_szad, node);
-
-		base_node_dealloc(prev);
-	}
-
-	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)
-{
-	bool ret;
-
-	malloc_mutex_lock(&dss_mtx);
-	if ((uintptr_t)chunk >= (uintptr_t)dss_base
-	    && (uintptr_t)chunk < (uintptr_t)dss_max) {
-		extent_node_t *node;
-
-		/* Try to coalesce with other unused chunks. */
-		node = chunk_dealloc_dss_record(chunk, size);
-		if (node != NULL) {
-			chunk = node->addr;
-			size = node->size;
-		}
-
-		/* Get the current end of the DSS. */
-		dss_max = sbrk(0);
-
-		/*
-		 * Try to shrink the DSS if this chunk is at the end of the
-		 * DSS.  The sbrk() call here is subject to a race condition
-		 * with threads that use brk(2) or sbrk(2) directly, but the
-		 * alternative would be to leak memory for the sake of poorly
-		 * designed multi-threaded programs.
-		 */
-		if ((void *)((uintptr_t)chunk + size) == dss_max
-		    && (dss_prev = sbrk(-(intptr_t)size)) == dss_max) {
-			/* Success. */
-			dss_max = (void *)((intptr_t)dss_prev - (intptr_t)size);
-
-			if (node != NULL) {
-				extent_tree_szad_remove(&dss_chunks_szad, node);
-				extent_tree_ad_remove(&dss_chunks_ad, node);
-				base_node_dealloc(node);
-			}
-		} else
-			madvise(chunk, size, MADV_DONTNEED);
-
-		ret = false;
-		goto RETURN;
-	}
-
-	ret = true;
-RETURN:
-	malloc_mutex_unlock(&dss_mtx);
-	return (ret);
-}
-
-bool
-chunk_dss_boot(void)
-{
-
-	if (malloc_mutex_init(&dss_mtx))
-		return (true);
-	dss_base = sbrk(0);
-	dss_prev = dss_base;
-	dss_max = dss_base;
-	extent_tree_szad_new(&dss_chunks_szad);
-	extent_tree_ad_new(&dss_chunks_ad);
-
-	return (false);
-}
-
-/******************************************************************************/
-#endif /* JEMALLOC_DSS */
--- a/jemalloc/src/chunk_mmap.c
+++ b/jemalloc/src/chunk_mmap.c
@@ -1,239 +0,0 @@
-#define	JEMALLOC_CHUNK_MMAP_C_
-#include "jemalloc/internal/jemalloc_internal.h"
-
-/******************************************************************************/
-/* Data. */
-
-/*
- * Used by chunk_alloc_mmap() to decide whether to attempt the fast path and
- * potentially avoid some system calls.
- */
-#ifndef NO_TLS
-static __thread bool	mmap_unaligned_tls
-    JEMALLOC_ATTR(tls_model("initial-exec"));
-#define	MMAP_UNALIGNED_GET()	mmap_unaligned_tls
-#define	MMAP_UNALIGNED_SET(v)	do {					\
-	mmap_unaligned_tls = (v);					\
-} while (0)
-#else
-static pthread_key_t	mmap_unaligned_tsd;
-#define	MMAP_UNALIGNED_GET()	((bool)pthread_getspecific(mmap_unaligned_tsd))
-#define	MMAP_UNALIGNED_SET(v)	do {					\
-	pthread_setspecific(mmap_unaligned_tsd, (void *)(v));		\
-} while (0)
-#endif
-
-/******************************************************************************/
-/* Function prototypes for non-inline static functions. */
-
-static void	*pages_map(void *addr, size_t size, bool noreserve);
-static void	pages_unmap(void *addr, size_t size);
-static void	*chunk_alloc_mmap_slow(size_t size, bool unaligned,
-    bool noreserve);
-static void	*chunk_alloc_mmap_internal(size_t size, bool noreserve);
-
-/******************************************************************************/
-
-static void *
-pages_map(void *addr, size_t size, bool noreserve)
-{
-	void *ret;
-
-	/*
-	 * We don't use MAP_FIXED here, because it can cause the *replacement*
-	 * of existing mappings, and we only want to create new mappings.
-	 */
-	int flags = MAP_PRIVATE | MAP_ANON;
-#ifdef MAP_NORESERVE
-	if (noreserve)
-		flags |= MAP_NORESERVE;
-#endif
-	ret = mmap(addr, size, PROT_READ | PROT_WRITE, flags, -1, 0);
-	assert(ret != NULL);
-
-	if (ret == MAP_FAILED)
-		ret = NULL;
-	else if (addr != NULL && ret != addr) {
-		/*
-		 * We succeeded in mapping memory, but not in the right place.
-		 */
-		if (munmap(ret, size) == -1) {
-			char buf[BUFERROR_BUF];
-
-			buferror(errno, buf, sizeof(buf));
-			malloc_write("<jemalloc>: Error in munmap(): ");
-			malloc_write(buf);
-			malloc_write("\n");
-			if (opt_abort)
-				abort();
-		}
-		ret = NULL;
-	}
-
-	assert(ret == NULL || (addr == NULL && ret != addr)
-	    || (addr != NULL && ret == addr));
-	return (ret);
-}
-
-static void
-pages_unmap(void *addr, size_t size)
-{
-
-	if (munmap(addr, size) == -1) {
-		char buf[BUFERROR_BUF];
-
-		buferror(errno, buf, sizeof(buf));
-		malloc_write("<jemalloc>: Error in munmap(): ");
-		malloc_write(buf);
-		malloc_write("\n");
-		if (opt_abort)
-			abort();
-	}
-}
-
-static void *
-chunk_alloc_mmap_slow(size_t size, bool unaligned, bool noreserve)
-{
-	void *ret;
-	size_t offset;
-
-	/* Beware size_t wrap-around. */
-	if (size + chunksize <= size)
-		return (NULL);
-
-	ret = pages_map(NULL, size + chunksize, noreserve);
-	if (ret == NULL)
-		return (NULL);
-
-	/* Clean up unneeded leading/trailing space. */
-	offset = CHUNK_ADDR2OFFSET(ret);
-	if (offset != 0) {
-		/* Note that mmap() returned an unaligned mapping. */
-		unaligned = true;
-
-		/* Leading space. */
-		pages_unmap(ret, chunksize - offset);
-
-		ret = (void *)((uintptr_t)ret +
-		    (chunksize - offset));
-
-		/* Trailing space. */
-		pages_unmap((void *)((uintptr_t)ret + size),
-		    offset);
-	} else {
-		/* Trailing space only. */
-		pages_unmap((void *)((uintptr_t)ret + size),
-		    chunksize);
-	}
-
-	/*
-	 * If mmap() returned an aligned mapping, reset mmap_unaligned so that
-	 * the next chunk_alloc_mmap() execution tries the fast allocation
-	 * method.
-	 */
-	if (unaligned == false)
-		MMAP_UNALIGNED_SET(false);
-
-	return (ret);
-}
-
-static void *
-chunk_alloc_mmap_internal(size_t size, bool noreserve)
-{
-	void *ret;
-
-	/*
-	 * Ideally, there would be a way to specify alignment to mmap() (like
-	 * NetBSD has), but in the absence of such a feature, we have to work
-	 * hard to efficiently create aligned mappings.  The reliable, but
-	 * slow method is to create a mapping that is over-sized, then trim the
-	 * excess.  However, that always results in at least one call to
-	 * pages_unmap().
-	 *
-	 * A more optimistic approach is to try mapping precisely the right
-	 * amount, then try to append another mapping if alignment is off.  In
-	 * practice, this works out well as long as the application is not
-	 * interleaving mappings via direct mmap() calls.  If we do run into a
-	 * situation where there is an interleaved mapping and we are unable to
-	 * extend an unaligned mapping, our best option is to switch to the
-	 * slow method until mmap() returns another aligned mapping.  This will
-	 * tend to leave a gap in the memory map that is too small to cause
-	 * later problems for the optimistic method.
-	 *
-	 * Another possible confounding factor is address space layout
-	 * randomization (ASLR), which causes mmap(2) to disregard the
-	 * requested address.  mmap_unaligned tracks whether the previous
-	 * chunk_alloc_mmap() execution received any unaligned or relocated
-	 * mappings, and if so, the current execution will immediately fall
-	 * back to the slow method.  However, we keep track of whether the fast
-	 * method would have succeeded, and if so, we make a note to try the
-	 * fast method next time.
-	 */
-
-	if (MMAP_UNALIGNED_GET() == false) {
-		size_t offset;
-
-		ret = pages_map(NULL, size, noreserve);
-		if (ret == NULL)
-			return (NULL);
-
-		offset = CHUNK_ADDR2OFFSET(ret);
-		if (offset != 0) {
-			MMAP_UNALIGNED_SET(true);
-			/* Try to extend chunk boundary. */
-			if (pages_map((void *)((uintptr_t)ret + size),
-			    chunksize - offset, noreserve) == NULL) {
-				/*
-				 * Extension failed.  Clean up, then revert to
-				 * the reliable-but-expensive method.
-				 */
-				pages_unmap(ret, size);
-				ret = chunk_alloc_mmap_slow(size, true,
-				    noreserve);
-			} else {
-				/* Clean up unneeded leading space. */
-				pages_unmap(ret, chunksize - offset);
-				ret = (void *)((uintptr_t)ret + (chunksize -
-				    offset));
-			}
-		}
-	} else
-		ret = chunk_alloc_mmap_slow(size, false, noreserve);
-
-	return (ret);
-}
-
-void *
-chunk_alloc_mmap(size_t size)
-{
-
-	return (chunk_alloc_mmap_internal(size, false));
-}
-
-void *
-chunk_alloc_mmap_noreserve(size_t size)
-{
-
-	return (chunk_alloc_mmap_internal(size, true));
-}
-
-void
-chunk_dealloc_mmap(void *chunk, size_t size)
-{
-
-	pages_unmap(chunk, size);
-}
-
-bool
-chunk_mmap_boot(void)
-{
-
-#ifdef NO_TLS
-	if (pthread_key_create(&mmap_unaligned_tsd, NULL) != 0) {
-		malloc_write("<jemalloc>: Error in pthread_key_create()\n");
-		return (true);
-	}
-#endif
-
-	return (false);
-}
--- a/jemalloc/src/chunk_swap.c
+++ b/jemalloc/src/chunk_swap.c
@@ -1,402 +0,0 @@
-#define	JEMALLOC_CHUNK_SWAP_C_
-#include "jemalloc/internal/jemalloc_internal.h"
-#ifdef JEMALLOC_SWAP
-/******************************************************************************/
-/* Data. */
-
-malloc_mutex_t	swap_mtx;
-bool		swap_enabled;
-bool		swap_prezeroed;
-size_t		swap_nfds;
-int		*swap_fds;
-#ifdef JEMALLOC_STATS
-size_t		swap_avail;
-#endif
-
-/* Base address of the mmap()ed file(s). */
-static void	*swap_base;
-/* Current end of the space in use (<= swap_max). */
-static void	*swap_end;
-/* Absolute upper limit on file-backed addresses. */
-static void	*swap_max;
-
-/*
- * Trees of chunks that were previously allocated (trees differ only in node
- * ordering).  These are used when allocating chunks, in an attempt to re-use
- * address space.  Depending on function, different tree orderings are needed,
- * which is why there are two trees with the same contents.
- */
-static extent_tree_t	swap_chunks_szad;
-static extent_tree_t	swap_chunks_ad;
-
-/******************************************************************************/
-/* Function prototypes for non-inline static functions. */
-
-static void	*chunk_recycle_swap(size_t size, bool *zero);
-static extent_node_t *chunk_dealloc_swap_record(void *chunk, size_t size);
-
-/******************************************************************************/
-
-static void *
-chunk_recycle_swap(size_t size, bool *zero)
-{
-	extent_node_t *node, key;
-
-	key.addr = NULL;
-	key.size = size;
-	malloc_mutex_lock(&swap_mtx);
-	node = extent_tree_szad_nsearch(&swap_chunks_szad, &key);
-	if (node != NULL) {
-		void *ret = node->addr;
-
-		/* Remove node from the tree. */
-		extent_tree_szad_remove(&swap_chunks_szad, node);
-		if (node->size == size) {
-			extent_tree_ad_remove(&swap_chunks_ad, node);
-			base_node_dealloc(node);
-		} else {
-			/*
-			 * Insert the remainder of node's address range as a
-			 * smaller chunk.  Its position within swap_chunks_ad
-			 * does not change.
-			 */
-			assert(node->size > size);
-			node->addr = (void *)((uintptr_t)node->addr + size);
-			node->size -= size;
-			extent_tree_szad_insert(&swap_chunks_szad, node);
-		}
-#ifdef JEMALLOC_STATS
-		swap_avail -= size;
-#endif
-		malloc_mutex_unlock(&swap_mtx);
-
-		if (*zero)
-			memset(ret, 0, size);
-		return (ret);
-	}
-	malloc_mutex_unlock(&swap_mtx);
-
-	return (NULL);
-}
-
-void *
-chunk_alloc_swap(size_t size, bool *zero)
-{
-	void *ret;
-
-	assert(swap_enabled);
-
-	ret = chunk_recycle_swap(size, zero);
-	if (ret != NULL)
-		return (ret);
-
-	malloc_mutex_lock(&swap_mtx);
-	if ((uintptr_t)swap_end + size <= (uintptr_t)swap_max) {
-		ret = swap_end;
-		swap_end = (void *)((uintptr_t)swap_end + size);
-#ifdef JEMALLOC_STATS
-		swap_avail -= size;
-#endif
-		malloc_mutex_unlock(&swap_mtx);
-
-		if (swap_prezeroed)
-			*zero = true;
-		else if (*zero)
-			memset(ret, 0, size);
-	} else {
-		malloc_mutex_unlock(&swap_mtx);
-		return (NULL);
-	}
-
-	return (ret);
-}
-
-static extent_node_t *
-chunk_dealloc_swap_record(void *chunk, size_t size)
-{
-	extent_node_t *xnode, *node, *prev, key;
-
-	xnode = NULL;
-	while (true) {
-		key.addr = (void *)((uintptr_t)chunk + size);
-		node = extent_tree_ad_nsearch(&swap_chunks_ad, &key);
-		/* Try to coalesce forward. */
-		if (node != NULL && node->addr == key.addr) {
-			/*
-			 * Coalesce chunk with the following address range.
-			 * This does not change the position within
-			 * swap_chunks_ad, so only remove/insert from/into
-			 * swap_chunks_szad.
-			 */
-			extent_tree_szad_remove(&swap_chunks_szad, node);
-			node->addr = chunk;
-			node->size += size;
-			extent_tree_szad_insert(&swap_chunks_szad, node);
-			break;
-		} else if (xnode == NULL) {
-			/*
-			 * It is possible that base_node_alloc() will cause a
-			 * new base chunk to be allocated, so take care not to
-			 * deadlock on swap_mtx, and recover if another thread
-			 * deallocates an adjacent chunk while this one is busy
-			 * allocating xnode.
-			 */
-			malloc_mutex_unlock(&swap_mtx);
-			xnode = base_node_alloc();
-			malloc_mutex_lock(&swap_mtx);
-			if (xnode == NULL)
-				return (NULL);
-		} else {
-			/* Coalescing forward failed, so insert a new node. */
-			node = xnode;
-			xnode = NULL;
-			node->addr = chunk;
-			node->size = size;
-			extent_tree_ad_insert(&swap_chunks_ad, node);
-			extent_tree_szad_insert(&swap_chunks_szad, node);
-			break;
-		}
-	}
-	/* Discard xnode if it ended up unused do to a race. */
-	if (xnode != NULL)
-		base_node_dealloc(xnode);
-
-	/* Try to coalesce backward. */
-	prev = extent_tree_ad_prev(&swap_chunks_ad, node);
-	if (prev != NULL && (void *)((uintptr_t)prev->addr + prev->size) ==
-	    chunk) {
-		/*
-		 * Coalesce chunk with the previous address range.  This does
-		 * not change the position within swap_chunks_ad, so only
-		 * remove/insert node from/into swap_chunks_szad.
-		 */
-		extent_tree_szad_remove(&swap_chunks_szad, prev);
-		extent_tree_ad_remove(&swap_chunks_ad, prev);
-
-		extent_tree_szad_remove(&swap_chunks_szad, node);
-		node->addr = prev->addr;
-		node->size += prev->size;
-		extent_tree_szad_insert(&swap_chunks_szad, node);
-
-		base_node_dealloc(prev);
-	}
-
-	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)
-{
-	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) {
-		extent_node_t *node;
-
-		/* Try to coalesce with other unused chunks. */
-		node = chunk_dealloc_swap_record(chunk, size);
-		if (node != NULL) {
-			chunk = node->addr;
-			size = node->size;
-		}
-
-		/*
-		 * Try to shrink the in-use memory if this chunk is at the end
-		 * of the in-use memory.
-		 */
-		if ((void *)((uintptr_t)chunk + size) == swap_end) {
-			swap_end = (void *)((uintptr_t)swap_end - size);
-
-			if (node != NULL) {
-				extent_tree_szad_remove(&swap_chunks_szad,
-				    node);
-				extent_tree_ad_remove(&swap_chunks_ad, node);
-				base_node_dealloc(node);
-			}
-		} else
-			madvise(chunk, size, MADV_DONTNEED);
-
-#ifdef JEMALLOC_STATS
-		swap_avail += size;
-#endif
-		ret = false;
-		goto RETURN;
-	}
-
-	ret = true;
-RETURN:
-	malloc_mutex_unlock(&swap_mtx);
-	return (ret);
-}
-
-bool
-chunk_swap_enable(const int *fds, unsigned nfds, bool prezeroed)
-{
-	bool ret;
-	unsigned i;
-	off_t off;
-	void *vaddr;
-	size_t cumsize, voff;
-	size_t sizes[nfds];
-
-	malloc_mutex_lock(&swap_mtx);
-
-	/* Get file sizes. */
-	for (i = 0, cumsize = 0; i < nfds; i++) {
-		off = lseek(fds[i], 0, SEEK_END);
-		if (off == ((off_t)-1)) {
-			ret = true;
-			goto RETURN;
-		}
-		if (PAGE_CEILING(off) != off) {
-			/* Truncate to a multiple of the page size. */
-			off &= ~PAGE_MASK;
-			if (ftruncate(fds[i], off) != 0) {
-				ret = true;
-				goto RETURN;
-			}
-		}
-		sizes[i] = off;
-		if (cumsize + off < cumsize) {
-			/*
-			 * Cumulative file size is greater than the total
-			 * address space.  Bail out while it's still obvious
-			 * what the problem is.
-			 */
-			ret = true;
-			goto RETURN;
-		}
-		cumsize += off;
-	}
-
-	/* Round down to a multiple of the chunk size. */
-	cumsize &= ~chunksize_mask;
-	if (cumsize == 0) {
-		ret = true;
-		goto RETURN;
-	}
-
-	/*
-	 * Allocate a chunk-aligned region of anonymous memory, which will
-	 * be the final location for the memory-mapped files.
-	 */
-	vaddr = chunk_alloc_mmap_noreserve(cumsize);
-	if (vaddr == NULL) {
-		ret = true;
-		goto RETURN;
-	}
-
-	/* Overlay the files onto the anonymous mapping. */
-	for (i = 0, voff = 0; i < nfds; i++) {
-		void *addr = mmap((void *)((uintptr_t)vaddr + voff), sizes[i],
-		    PROT_READ | PROT_WRITE, MAP_SHARED | MAP_FIXED, fds[i], 0);
-		if (addr == MAP_FAILED) {
-			char buf[BUFERROR_BUF];
-
-
-			buferror(errno, buf, sizeof(buf));
-			malloc_write(
-			    "<jemalloc>: Error in mmap(..., MAP_FIXED, ...): ");
-			malloc_write(buf);
-			malloc_write("\n");
-			if (opt_abort)
-				abort();
-			if (munmap(vaddr, voff) == -1) {
-				buferror(errno, buf, sizeof(buf));
-				malloc_write("<jemalloc>: Error in munmap(): ");
-				malloc_write(buf);
-				malloc_write("\n");
-			}
-			ret = true;
-			goto RETURN;
-		}
-		assert(addr == (void *)((uintptr_t)vaddr + voff));
-
-		/*
-		 * Tell the kernel that the mapping will be accessed randomly,
-		 * and that it should not gratuitously sync pages to the
-		 * filesystem.
-		 */
-#ifdef MADV_RANDOM
-		madvise(addr, sizes[i], MADV_RANDOM);
-#endif
-#ifdef MADV_NOSYNC
-		madvise(addr, sizes[i], MADV_NOSYNC);
-#endif
-
-		voff += sizes[i];
-	}
-
-	swap_prezeroed = prezeroed;
-	swap_base = vaddr;
-	swap_end = swap_base;
-	swap_max = (void *)((uintptr_t)vaddr + cumsize);
-
-	/* Copy the fds array for mallctl purposes. */
-	swap_fds = (int *)base_alloc(nfds * sizeof(int));
-	if (swap_fds == NULL) {
-		ret = true;
-		goto RETURN;
-	}
-	memcpy(swap_fds, fds, nfds * sizeof(int));
-	swap_nfds = nfds;
-
-#ifdef JEMALLOC_STATS
-	swap_avail = cumsize;
-#endif
-
-	swap_enabled = true;
-
-	ret = false;
-RETURN:
-	malloc_mutex_unlock(&swap_mtx);
-	return (ret);
-}
-
-bool
-chunk_swap_boot(void)
-{
-
-	if (malloc_mutex_init(&swap_mtx))
-		return (true);
-
-	swap_enabled = false;
-	swap_prezeroed = false; /* swap.* mallctl's depend on this. */
-	swap_nfds = 0;
-	swap_fds = NULL;
-#ifdef JEMALLOC_STATS
-	swap_avail = 0;
-#endif
-	swap_base = NULL;
-	swap_end = NULL;
-	swap_max = NULL;
-
-	extent_tree_szad_new(&swap_chunks_szad);
-	extent_tree_ad_new(&swap_chunks_ad);
-
-	return (false);
-}
-
-/******************************************************************************/
-#endif /* JEMALLOC_SWAP */
--- a/jemalloc/src/ctl.c
+++ b/jemalloc/src/ctl.c
--- a/jemalloc/src/jemalloc.c
+++ b/jemalloc/src/jemalloc.c
--- a/jemalloc/src/mutex.c
+++ b/jemalloc/src/mutex.c
@@ -1,84 +0,0 @@
-#define	JEMALLOC_MUTEX_C_
-#include "jemalloc/internal/jemalloc_internal.h"
-
-/******************************************************************************/
-/* Data. */
-
-#ifdef JEMALLOC_LAZY_LOCK
-bool isthreaded = false;
-#endif
-
-#ifdef JEMALLOC_LAZY_LOCK
-static void	pthread_create_once(void);
-#endif
-
-/******************************************************************************/
-/*
- * We intercept pthread_create() calls in order to toggle isthreaded if the
- * process goes multi-threaded.
- */
-
-#ifdef JEMALLOC_LAZY_LOCK
-static int (*pthread_create_fptr)(pthread_t *__restrict, const pthread_attr_t *,
-    void *(*)(void *), void *__restrict);
-
-static void
-pthread_create_once(void)
-{
-
-	pthread_create_fptr = dlsym(RTLD_NEXT, "pthread_create");
-	if (pthread_create_fptr == NULL) {
-		malloc_write("<jemalloc>: Error in dlsym(RTLD_NEXT, "
-		    "\"pthread_create\")\n");
-		abort();
-	}
-
-	isthreaded = true;
-}
-
-JEMALLOC_ATTR(visibility("default"))
-int
-pthread_create(pthread_t *__restrict thread,
-    const pthread_attr_t *__restrict attr, void *(*start_routine)(void *),
-    void *__restrict arg)
-{
-	static pthread_once_t once_control = PTHREAD_ONCE_INIT;
-
-	pthread_once(&once_control, pthread_create_once);
-
-	return (pthread_create_fptr(thread, attr, start_routine, arg));
-}
-#endif
-
-/******************************************************************************/
-
-bool
-malloc_mutex_init(malloc_mutex_t *mutex)
-{
-	pthread_mutexattr_t attr;
-
-	if (pthread_mutexattr_init(&attr) != 0)
-		return (true);
-#ifdef PTHREAD_MUTEX_ADAPTIVE_NP
-	pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_ADAPTIVE_NP);
-#else
-	pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_DEFAULT);
-#endif
-	if (pthread_mutex_init(mutex, &attr) != 0) {
-		pthread_mutexattr_destroy(&attr);
-		return (true);
-	}
-	pthread_mutexattr_destroy(&attr);
-
-	return (false);
-}
-
-void
-malloc_mutex_destroy(malloc_mutex_t *mutex)
-{
-
-	if (pthread_mutex_destroy(mutex) != 0) {
-		malloc_write("<jemalloc>: Error in pthread_mutex_destroy()\n");
-		abort();
-	}
-}
--- a/jemalloc/src/tcache.c
+++ b/jemalloc/src/tcache.c
@@ -1,417 +0,0 @@
-#define	JEMALLOC_TCACHE_C_
-#include "jemalloc/internal/jemalloc_internal.h"
-#ifdef JEMALLOC_TCACHE
-/******************************************************************************/
-/* Data. */
-
-bool	opt_tcache = true;
-ssize_t	opt_lg_tcache_max = LG_TCACHE_MAXCLASS_DEFAULT;
-ssize_t	opt_lg_tcache_gc_sweep = LG_TCACHE_GC_SWEEP_DEFAULT;
-
-/* Map of thread-specific caches. */
-#ifndef NO_TLS
-__thread tcache_t	*tcache_tls JEMALLOC_ATTR(tls_model("initial-exec"));
-#endif
-
-/*
- * Same contents as tcache, but initialized such that the TSD destructor is
- * called when a thread exits, so that the cache can be cleaned up.
- */
-pthread_key_t		tcache_tsd;
-
-size_t				nhbins;
-size_t				tcache_maxclass;
-unsigned			tcache_gc_incr;
-
-/******************************************************************************/
-/* Function prototypes for non-inline static functions. */
-
-static void	tcache_thread_cleanup(void *arg);
-
-/******************************************************************************/
-
-void *
-tcache_alloc_small_hard(tcache_t *tcache, tcache_bin_t *tbin, size_t binind)
-{
-	void *ret;
-
-	arena_tcache_fill_small(tcache->arena, tbin, binind
-#ifdef JEMALLOC_PROF
-	    , tcache->prof_accumbytes
-#endif
-	    );
-#ifdef JEMALLOC_PROF
-	tcache->prof_accumbytes = 0;
-#endif
-	ret = tcache_alloc_easy(tbin);
-
-	return (ret);
-}
-
-void
-tcache_bin_flush_small(tcache_bin_t *tbin, size_t binind, unsigned rem
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-    , tcache_t *tcache
-#endif
-    )
-{
-	void *flush, *deferred, *ptr;
-	unsigned i, nflush, ndeferred;
-	bool first_pass;
-
-	assert(binind < nbins);
-	assert(rem <= tbin->ncached);
-	assert(tbin->ncached > 0 || tbin->avail == NULL);
-
-	for (flush = tbin->avail, nflush = tbin->ncached - rem, first_pass =
-	    true; flush != NULL; flush = deferred, nflush = ndeferred) {
-		/* Lock the arena bin associated with the first object. */
-		arena_chunk_t *chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(flush);
-		arena_t *arena = chunk->arena;
-		arena_bin_t *bin = &arena->bins[binind];
-
-#ifdef JEMALLOC_PROF
-		if (arena == tcache->arena) {
-			malloc_mutex_lock(&arena->lock);
-			arena_prof_accum(arena, tcache->prof_accumbytes);
-			malloc_mutex_unlock(&arena->lock);
-			tcache->prof_accumbytes = 0;
-		}
-#endif
-
-		malloc_mutex_lock(&bin->lock);
-#ifdef JEMALLOC_STATS
-		if (arena == tcache->arena) {
-			bin->stats.nflushes++;
-			bin->stats.nrequests += tbin->tstats.nrequests;
-			tbin->tstats.nrequests = 0;
-		}
-#endif
-		deferred = NULL;
-		ndeferred = 0;
-		for (i = 0; i < nflush; i++) {
-			ptr = flush;
-			assert(ptr != NULL);
-			flush = *(void **)ptr;
-			chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
-			if (chunk->arena == arena) {
-				size_t pageind = ((uintptr_t)ptr -
-				    (uintptr_t)chunk) >> PAGE_SHIFT;
-				arena_chunk_map_t *mapelm =
-				    &chunk->map[pageind-map_bias];
-				arena_dalloc_bin(arena, chunk, ptr, mapelm);
-			} else {
-				/*
-				 * This object was allocated via a different
-				 * arena bin than the one that is currently
-				 * locked.  Stash the object, so that it can be
-				 * handled in a future pass.
-				 */
-				*(void **)ptr = deferred;
-				deferred = ptr;
-				ndeferred++;
-			}
-		}
-		malloc_mutex_unlock(&bin->lock);
-
-		if (first_pass) {
-			tbin->avail = flush;
-			first_pass = false;
-		}
-	}
-
-	tbin->ncached = rem;
-	if (tbin->ncached < tbin->low_water)
-		tbin->low_water = tbin->ncached;
-}
-
-void
-tcache_bin_flush_large(tcache_bin_t *tbin, size_t binind, unsigned rem
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-    , tcache_t *tcache
-#endif
-    )
-{
-	void *flush, *deferred, *ptr;
-	unsigned i, nflush, ndeferred;
-	bool first_pass;
-
-	assert(binind < nhbins);
-	assert(rem <= tbin->ncached);
-	assert(tbin->ncached > 0 || tbin->avail == NULL);
-
-	for (flush = tbin->avail, nflush = tbin->ncached - rem, first_pass =
-	    true; flush != NULL; flush = deferred, nflush = ndeferred) {
-		/* Lock the arena associated with the first object. */
-		arena_chunk_t *chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(flush);
-		arena_t *arena = chunk->arena;
-
-		malloc_mutex_lock(&arena->lock);
-#if (defined(JEMALLOC_PROF) || defined(JEMALLOC_STATS))
-		if (arena == tcache->arena) {
-#endif
-#ifdef JEMALLOC_PROF
-			arena_prof_accum(arena, tcache->prof_accumbytes);
-			tcache->prof_accumbytes = 0;
-#endif
-#ifdef JEMALLOC_STATS
-			arena->stats.nrequests_large += tbin->tstats.nrequests;
-			arena->stats.lstats[binind - nbins].nrequests +=
-			    tbin->tstats.nrequests;
-			tbin->tstats.nrequests = 0;
-#endif
-#if (defined(JEMALLOC_PROF) || defined(JEMALLOC_STATS))
-		}
-#endif
-		deferred = NULL;
-		ndeferred = 0;
-		for (i = 0; i < nflush; i++) {
-			ptr = flush;
-			assert(ptr != NULL);
-			flush = *(void **)ptr;
-			chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
-			if (chunk->arena == arena)
-				arena_dalloc_large(arena, chunk, ptr);
-			else {
-				/*
-				 * This object was allocated via a different
-				 * arena than the one that is currently locked.
-				 * Stash the object, so that it can be handled
-				 * in a future pass.
-				 */
-				*(void **)ptr = deferred;
-				deferred = ptr;
-				ndeferred++;
-			}
-		}
-		malloc_mutex_unlock(&arena->lock);
-
-		if (first_pass) {
-			tbin->avail = flush;
-			first_pass = false;
-		}
-	}
-
-	tbin->ncached = rem;
-	if (tbin->ncached < tbin->low_water)
-		tbin->low_water = tbin->ncached;
-}
-
-tcache_t *
-tcache_create(arena_t *arena)
-{
-	tcache_t *tcache;
-	size_t size;
-	unsigned i;
-
-	size = offsetof(tcache_t, tbins) + (sizeof(tcache_bin_t) * nhbins);
-	/*
-	 * Round up to the nearest multiple of the cacheline size, in order to
-	 * avoid the possibility of false cacheline sharing.
-	 *
-	 * That this works relies on the same logic as in ipalloc(), but we
-	 * cannot directly call ipalloc() here due to tcache bootstrapping
-	 * issues.
-	 */
-	size = (size + CACHELINE_MASK) & (-CACHELINE);
-
-	if (size <= small_maxclass)
-		tcache = (tcache_t *)arena_malloc_small(arena, size, true);
-	else
-		tcache = (tcache_t *)icalloc(size);
-
-	if (tcache == NULL)
-		return (NULL);
-
-#ifdef JEMALLOC_STATS
-	/* Link into list of extant tcaches. */
-	malloc_mutex_lock(&arena->lock);
-	ql_elm_new(tcache, link);
-	ql_tail_insert(&arena->tcache_ql, tcache, link);
-	malloc_mutex_unlock(&arena->lock);
-#endif
-
-	tcache->arena = arena;
-	assert((TCACHE_NSLOTS_SMALL_MAX & 1U) == 0);
-	for (i = 0; i < nbins; i++) {
-		if ((arena->bins[i].nregs << 1) <= TCACHE_NSLOTS_SMALL_MAX) {
-			tcache->tbins[i].ncached_max = (arena->bins[i].nregs <<
-			    1);
-		} else
-			tcache->tbins[i].ncached_max = TCACHE_NSLOTS_SMALL_MAX;
-	}
-	for (; i < nhbins; i++)
-		tcache->tbins[i].ncached_max = TCACHE_NSLOTS_LARGE;
-
-	TCACHE_SET(tcache);
-
-	return (tcache);
-}
-
-void
-tcache_destroy(tcache_t *tcache)
-{
-	unsigned i;
-
-#ifdef JEMALLOC_STATS
-	/* Unlink from list of extant tcaches. */
-	malloc_mutex_lock(&tcache->arena->lock);
-	ql_remove(&tcache->arena->tcache_ql, tcache, link);
-	malloc_mutex_unlock(&tcache->arena->lock);
-	tcache_stats_merge(tcache, tcache->arena);
-#endif
-
-	for (i = 0; i < nbins; i++) {
-		tcache_bin_t *tbin = &tcache->tbins[i];
-		tcache_bin_flush_small(tbin, i, 0
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-		    , tcache
-#endif
-		    );
-
-#ifdef JEMALLOC_STATS
-		if (tbin->tstats.nrequests != 0) {
-			arena_t *arena = tcache->arena;
-			arena_bin_t *bin = &arena->bins[i];
-			malloc_mutex_lock(&bin->lock);
-			bin->stats.nrequests += tbin->tstats.nrequests;
-			malloc_mutex_unlock(&bin->lock);
-		}
-#endif
-	}
-
-	for (; i < nhbins; i++) {
-		tcache_bin_t *tbin = &tcache->tbins[i];
-		tcache_bin_flush_large(tbin, i, 0
-#if (defined(JEMALLOC_STATS) || defined(JEMALLOC_PROF))
-		    , tcache
-#endif
-		    );
-
-#ifdef JEMALLOC_STATS
-		if (tbin->tstats.nrequests != 0) {
-			arena_t *arena = tcache->arena;
-			malloc_mutex_lock(&arena->lock);
-			arena->stats.nrequests_large += tbin->tstats.nrequests;
-			arena->stats.lstats[i - nbins].nrequests +=
-			    tbin->tstats.nrequests;
-			malloc_mutex_unlock(&arena->lock);
-		}
-#endif
-	}
-
-#ifdef JEMALLOC_PROF
-	if (tcache->prof_accumbytes > 0) {
-		malloc_mutex_lock(&tcache->arena->lock);
-		arena_prof_accum(tcache->arena, tcache->prof_accumbytes);
-		malloc_mutex_unlock(&tcache->arena->lock);
-	}
-#endif
-
-	if (arena_salloc(tcache) <= small_maxclass) {
-		arena_chunk_t *chunk = CHUNK_ADDR2BASE(tcache);
-		arena_t *arena = chunk->arena;
-		size_t pageind = ((uintptr_t)tcache - (uintptr_t)chunk) >>
-		    PAGE_SHIFT;
-		arena_chunk_map_t *mapelm = &chunk->map[pageind-map_bias];
-		arena_run_t *run = (arena_run_t *)((uintptr_t)chunk +
-		    (uintptr_t)((pageind - (mapelm->bits >> PAGE_SHIFT)) <<
-		    PAGE_SHIFT));
-		arena_bin_t *bin = run->bin;
-
-		malloc_mutex_lock(&bin->lock);
-		arena_dalloc_bin(arena, chunk, tcache, mapelm);
-		malloc_mutex_unlock(&bin->lock);
-	} else
-		idalloc(tcache);
-}
-
-static void
-tcache_thread_cleanup(void *arg)
-{
-	tcache_t *tcache = (tcache_t *)arg;
-
-	if (tcache == (void *)(uintptr_t)1) {
-		/*
-		 * The previous time this destructor was called, we set the key
-		 * to 1 so that other destructors wouldn't cause re-creation of
-		 * the tcache.  This time, do nothing, so that the destructor
-		 * will not be called again.
-		 */
-	} else if (tcache == (void *)(uintptr_t)2) {
-		/*
-		 * Another destructor called an allocator function after this
-		 * destructor was called.  Reset tcache to 1 in order to
-		 * receive another callback.
-		 */
-		TCACHE_SET((uintptr_t)1);
-	} else if (tcache != NULL) {
-		assert(tcache != (void *)(uintptr_t)1);
-		tcache_destroy(tcache);
-		TCACHE_SET((uintptr_t)1);
-	}
-}
-
-#ifdef JEMALLOC_STATS
-void
-tcache_stats_merge(tcache_t *tcache, arena_t *arena)
-{
-	unsigned i;
-
-	/* Merge and reset tcache stats. */
-	for (i = 0; i < nbins; i++) {
-		arena_bin_t *bin = &arena->bins[i];
-		tcache_bin_t *tbin = &tcache->tbins[i];
-		malloc_mutex_lock(&bin->lock);
-		bin->stats.nrequests += tbin->tstats.nrequests;
-		malloc_mutex_unlock(&bin->lock);
-		tbin->tstats.nrequests = 0;
-	}
-
-	for (; i < nhbins; i++) {
-		malloc_large_stats_t *lstats = &arena->stats.lstats[i - nbins];
-		tcache_bin_t *tbin = &tcache->tbins[i];
-		arena->stats.nrequests_large += tbin->tstats.nrequests;
-		lstats->nrequests += tbin->tstats.nrequests;
-		tbin->tstats.nrequests = 0;
-	}
-}
-#endif
-
-void
-tcache_boot(void)
-{
-
-	if (opt_tcache) {
-		/*
-		 * If necessary, clamp opt_lg_tcache_max, now that
-		 * small_maxclass and arena_maxclass are known.
-		 */
-		if (opt_lg_tcache_max < 0 || (1U <<
-		    opt_lg_tcache_max) < small_maxclass)
-			tcache_maxclass = small_maxclass;
-		else if ((1U << opt_lg_tcache_max) > arena_maxclass)
-			tcache_maxclass = arena_maxclass;
-		else
-			tcache_maxclass = (1U << opt_lg_tcache_max);
-
-		nhbins = nbins + (tcache_maxclass >> PAGE_SHIFT);
-
-		/* Compute incremental GC event threshold. */
-		if (opt_lg_tcache_gc_sweep >= 0) {
-			tcache_gc_incr = ((1U << opt_lg_tcache_gc_sweep) /
-			    nbins) + (((1U << opt_lg_tcache_gc_sweep) % nbins ==
-			    0) ? 0 : 1);
-		} else
-			tcache_gc_incr = 0;
-
-		if (pthread_key_create(&tcache_tsd, tcache_thread_cleanup) !=
-		    0) {
-			malloc_write(
-			    "<jemalloc>: Error in pthread_key_create()\n");
-			abort();
-		}
-	}
-}
-/******************************************************************************/
-#endif /* JEMALLOC_TCACHE */
--- a/jemalloc/src/zone.c
+++ b/jemalloc/src/zone.c
@@ -1,354 +0,0 @@
-#include "jemalloc/internal/jemalloc_internal.h"
-#ifndef JEMALLOC_ZONE
-#  error "This source file is for zones on Darwin (OS X)."
-#endif
-
-/******************************************************************************/
-/* Data. */
-
-static malloc_zone_t zone, szone;
-static struct malloc_introspection_t zone_introspect, ozone_introspect;
-
-/******************************************************************************/
-/* Function prototypes for non-inline static functions. */
-
-static size_t	zone_size(malloc_zone_t *zone, void *ptr);
-static void	*zone_malloc(malloc_zone_t *zone, size_t size);
-static void	*zone_calloc(malloc_zone_t *zone, size_t num, size_t size);
-static void	*zone_valloc(malloc_zone_t *zone, size_t size);
-static void	zone_free(malloc_zone_t *zone, void *ptr);
-static void	*zone_realloc(malloc_zone_t *zone, void *ptr, size_t size);
-#if (JEMALLOC_ZONE_VERSION >= 6)
-static void	*zone_memalign(malloc_zone_t *zone, size_t alignment,
-    size_t size);
-static void	zone_free_definite_size(malloc_zone_t *zone, void *ptr,
-    size_t size);
-#endif
-static void	*zone_destroy(malloc_zone_t *zone);
-static size_t	zone_good_size(malloc_zone_t *zone, size_t size);
-static void	zone_force_lock(malloc_zone_t *zone);
-static void	zone_force_unlock(malloc_zone_t *zone);
-static size_t	ozone_size(malloc_zone_t *zone, void *ptr);
-static void	ozone_free(malloc_zone_t *zone, void *ptr);
-static void	*ozone_realloc(malloc_zone_t *zone, void *ptr, size_t size);
-static unsigned	ozone_batch_malloc(malloc_zone_t *zone, size_t size,
-    void **results, unsigned num_requested);
-static void	ozone_batch_free(malloc_zone_t *zone, void **to_be_freed,
-    unsigned num);
-#if (JEMALLOC_ZONE_VERSION >= 6)
-static void	ozone_free_definite_size(malloc_zone_t *zone, void *ptr,
-    size_t size);
-#endif
-static void	ozone_force_lock(malloc_zone_t *zone);
-static void	ozone_force_unlock(malloc_zone_t *zone);
-
-/******************************************************************************/
-/*
- * Functions.
- */
-
-static size_t
-zone_size(malloc_zone_t *zone, void *ptr)
-{
-
-	/*
-	 * There appear to be places within Darwin (such as setenv(3)) that
-	 * cause calls to this function with pointers that *no* zone owns.  If
-	 * we knew that all pointers were owned by *some* zone, we could split
-	 * our zone into two parts, and use one as the default allocator and
-	 * the other as the default deallocator/reallocator.  Since that will
-	 * not work in practice, we must check all pointers to assure that they
-	 * reside within a mapped chunk before determining size.
-	 */
-	return (ivsalloc(ptr));
-}
-
-static void *
-zone_malloc(malloc_zone_t *zone, size_t size)
-{
-
-	return (JEMALLOC_P(malloc)(size));
-}
-
-static void *
-zone_calloc(malloc_zone_t *zone, size_t num, size_t size)
-{
-
-	return (JEMALLOC_P(calloc)(num, size));
-}
-
-static void *
-zone_valloc(malloc_zone_t *zone, size_t size)
-{
-	void *ret = NULL; /* Assignment avoids useless compiler warning. */
-
-	JEMALLOC_P(posix_memalign)(&ret, PAGE_SIZE, size);
-
-	return (ret);
-}
-
-static void
-zone_free(malloc_zone_t *zone, void *ptr)
-{
-
-	JEMALLOC_P(free)(ptr);
-}
-
-static void *
-zone_realloc(malloc_zone_t *zone, void *ptr, size_t size)
-{
-
-	return (JEMALLOC_P(realloc)(ptr, size));
-}
-
-#if (JEMALLOC_ZONE_VERSION >= 6)
-static void *
-zone_memalign(malloc_zone_t *zone, size_t alignment, size_t size)
-{
-	void *ret = NULL; /* Assignment avoids useless compiler warning. */
-
-	JEMALLOC_P(posix_memalign)(&ret, alignment, size);
-
-	return (ret);
-}
-
-static void
-zone_free_definite_size(malloc_zone_t *zone, void *ptr, size_t size)
-{
-
-	assert(ivsalloc(ptr) == size);
-	JEMALLOC_P(free)(ptr);
-}
-#endif
-
-static void *
-zone_destroy(malloc_zone_t *zone)
-{
-
-	/* This function should never be called. */
-	assert(false);
-	return (NULL);
-}
-
-static size_t
-zone_good_size(malloc_zone_t *zone, size_t size)
-{
-	size_t ret;
-	void *p;
-
-	/*
-	 * Actually create an object of the appropriate size, then find out
-	 * how large it could have been without moving up to the next size
-	 * class.
-	 */
-	p = JEMALLOC_P(malloc)(size);
-	if (p != NULL) {
-		ret = isalloc(p);
-		JEMALLOC_P(free)(p);
-	} else
-		ret = size;
-
-	return (ret);
-}
-
-static void
-zone_force_lock(malloc_zone_t *zone)
-{
-
-	if (isthreaded)
-		jemalloc_prefork();
-}
-
-static void
-zone_force_unlock(malloc_zone_t *zone)
-{
-
-	if (isthreaded)
-		jemalloc_postfork();
-}
-
-malloc_zone_t *
-create_zone(void)
-{
-
-	zone.size = (void *)zone_size;
-	zone.malloc = (void *)zone_malloc;
-	zone.calloc = (void *)zone_calloc;
-	zone.valloc = (void *)zone_valloc;
-	zone.free = (void *)zone_free;
-	zone.realloc = (void *)zone_realloc;
-	zone.destroy = (void *)zone_destroy;
-	zone.zone_name = "jemalloc_zone";
-	zone.batch_malloc = NULL;
-	zone.batch_free = NULL;
-	zone.introspect = &zone_introspect;
-	zone.version = JEMALLOC_ZONE_VERSION;
-#if (JEMALLOC_ZONE_VERSION >= 6)
-	zone.memalign = zone_memalign;
-	zone.free_definite_size = zone_free_definite_size;
-#endif
-
-	zone_introspect.enumerator = NULL;
-	zone_introspect.good_size = (void *)zone_good_size;
-	zone_introspect.check = NULL;
-	zone_introspect.print = NULL;
-	zone_introspect.log = NULL;
-	zone_introspect.force_lock = (void *)zone_force_lock;
-	zone_introspect.force_unlock = (void *)zone_force_unlock;
-	zone_introspect.statistics = NULL;
-#if (JEMALLOC_ZONE_VERSION >= 6)
-	zone_introspect.zone_locked = NULL;
-#endif
-
-	return (&zone);
-}
-
-static size_t
-ozone_size(malloc_zone_t *zone, void *ptr)
-{
-	size_t ret;
-
-	ret = ivsalloc(ptr);
-	if (ret == 0)
-		ret = szone.size(zone, ptr);
-
-	return (ret);
-}
-
-static void
-ozone_free(malloc_zone_t *zone, void *ptr)
-{
-
-	if (ivsalloc(ptr) != 0)
-		JEMALLOC_P(free)(ptr);
-	else {
-		size_t size = szone.size(zone, ptr);
-		if (size != 0)
-			(szone.free)(zone, ptr);
-	}
-}
-
-static void *
-ozone_realloc(malloc_zone_t *zone, void *ptr, size_t size)
-{
-	size_t oldsize;
-
-	if (ptr == NULL)
-		return (JEMALLOC_P(malloc)(size));
-
-	oldsize = ivsalloc(ptr);
-	if (oldsize != 0)
-		return (JEMALLOC_P(realloc)(ptr, size));
-	else {
-		oldsize = szone.size(zone, ptr);
-		if (oldsize == 0)
-			return (JEMALLOC_P(malloc)(size));
-		else {
-			void *ret = JEMALLOC_P(malloc)(size);
-			if (ret != NULL) {
-				memcpy(ret, ptr, (oldsize < size) ? oldsize :
-				    size);
-				(szone.free)(zone, ptr);
-			}
-			return (ret);
-		}
-	}
-}
-
-static unsigned
-ozone_batch_malloc(malloc_zone_t *zone, size_t size, void **results,
-    unsigned num_requested)
-{
-
-	/* Don't bother implementing this interface, since it isn't required. */
-	return (0);
-}
-
-static void
-ozone_batch_free(malloc_zone_t *zone, void **to_be_freed, unsigned num)
-{
-	unsigned i;
-
-	for (i = 0; i < num; i++)
-		ozone_free(zone, to_be_freed[i]);
-}
-
-#if (JEMALLOC_ZONE_VERSION >= 6)
-static void
-ozone_free_definite_size(malloc_zone_t *zone, void *ptr, size_t size)
-{
-
-	if (ivsalloc(ptr) != 0) {
-		assert(ivsalloc(ptr) == size);
-		JEMALLOC_P(free)(ptr);
-	} else {
-		assert(size == szone.size(zone, ptr));
-		szone.free_definite_size(zone, ptr, size);
-	}
-}
-#endif
-
-static void
-ozone_force_lock(malloc_zone_t *zone)
-{
-
-	/* jemalloc locking is taken care of by the normal jemalloc zone. */
-	szone.introspect->force_lock(zone);
-}
-
-static void
-ozone_force_unlock(malloc_zone_t *zone)
-{
-
-	/* jemalloc locking is taken care of by the normal jemalloc zone. */
-	szone.introspect->force_unlock(zone);
-}
-
-/*
- * Overlay the default scalable zone (szone) such that existing allocations are
- * drained, and further allocations come from jemalloc.  This is necessary
- * because Core Foundation directly accesses and uses the szone before the
- * jemalloc library is even loaded.
- */
-void
-szone2ozone(malloc_zone_t *zone)
-{
-
-	/*
-	 * Stash a copy of the original szone so that we can call its
-	 * functions as needed.  Note that the internally, the szone stores its
-	 * bookkeeping data structures immediately following the malloc_zone_t
-	 * header, so when calling szone functions, we need to pass a pointer
-	 * to the original zone structure.
-	 */
-	memcpy(&szone, zone, sizeof(malloc_zone_t));
-
-	zone->size = (void *)ozone_size;
-	zone->malloc = (void *)zone_malloc;
-	zone->calloc = (void *)zone_calloc;
-	zone->valloc = (void *)zone_valloc;
-	zone->free = (void *)ozone_free;
-	zone->realloc = (void *)ozone_realloc;
-	zone->destroy = (void *)zone_destroy;
-	zone->zone_name = "jemalloc_ozone";
-	zone->batch_malloc = ozone_batch_malloc;
-	zone->batch_free = ozone_batch_free;
-	zone->introspect = &ozone_introspect;
-	zone->version = JEMALLOC_ZONE_VERSION;
-#if (JEMALLOC_ZONE_VERSION >= 6)
-	zone->memalign = zone_memalign;
-	zone->free_definite_size = ozone_free_definite_size;
-#endif
-
-	ozone_introspect.enumerator = NULL;
-	ozone_introspect.good_size = (void *)zone_good_size;
-	ozone_introspect.check = NULL;
-	ozone_introspect.print = NULL;
-	ozone_introspect.log = NULL;
-	ozone_introspect.force_lock = (void *)ozone_force_lock;
-	ozone_introspect.force_unlock = (void *)ozone_force_unlock;
-	ozone_introspect.statistics = NULL;
-#if (JEMALLOC_ZONE_VERSION >= 6)
-	ozone_introspect.zone_locked = NULL;
-#endif
-}
--- a/jemalloc/test/allocated.c
+++ b/jemalloc/test/allocated.c
@@ -1,142 +0,0 @@
-#include <stdio.h>
-#include <stdlib.h>
-#include <stdint.h>
-#include <stdbool.h>
-#include <pthread.h>
-#include <assert.h>
-#include <errno.h>
-#include <string.h>
-
-#define	JEMALLOC_MANGLE
-#include "jemalloc_test.h"
-
-void *
-thread_start(void *arg)
-{
-	int err;
-	void *p;
-	uint64_t a0, a1, d0, d1;
-	uint64_t *ap0, *ap1, *dp0, *dp1;
-	size_t sz, usize;
-
-	sz = sizeof(a0);
-	if ((err = JEMALLOC_P(mallctl)("thread.allocated", &a0, &sz, NULL,
-	    0))) {
-		if (err == ENOENT) {
-#ifdef JEMALLOC_STATS
-			assert(false);
-#endif
-			goto RETURN;
-		}
-		fprintf(stderr, "%s(): Error in mallctl(): %s\n", __func__,
-		    strerror(err));
-		exit(1);
-	}
-	sz = sizeof(ap0);
-	if ((err = JEMALLOC_P(mallctl)("thread.allocatedp", &ap0, &sz, NULL,
-	    0))) {
-		if (err == ENOENT) {
-#ifdef JEMALLOC_STATS
-			assert(false);
-#endif
-			goto RETURN;
-		}
-		fprintf(stderr, "%s(): Error in mallctl(): %s\n", __func__,
-		    strerror(err));
-		exit(1);
-	}
-	assert(*ap0 == a0);
-
-	sz = sizeof(d0);
-	if ((err = JEMALLOC_P(mallctl)("thread.deallocated", &d0, &sz, NULL,
-	    0))) {
-		if (err == ENOENT) {
-#ifdef JEMALLOC_STATS
-			assert(false);
-#endif
-			goto RETURN;
-		}
-		fprintf(stderr, "%s(): Error in mallctl(): %s\n", __func__,
-		    strerror(err));
-		exit(1);
-	}
-	sz = sizeof(dp0);
-	if ((err = JEMALLOC_P(mallctl)("thread.deallocatedp", &dp0, &sz, NULL,
-	    0))) {
-		if (err == ENOENT) {
-#ifdef JEMALLOC_STATS
-			assert(false);
-#endif
-			goto RETURN;
-		}
-		fprintf(stderr, "%s(): Error in mallctl(): %s\n", __func__,
-		    strerror(err));
-		exit(1);
-	}
-	assert(*dp0 == d0);
-
-	p = JEMALLOC_P(malloc)(1);
-	if (p == NULL) {
-		fprintf(stderr, "%s(): Error in malloc()\n", __func__);
-		exit(1);
-	}
-
-	sz = sizeof(a1);
-	JEMALLOC_P(mallctl)("thread.allocated", &a1, &sz, NULL, 0);
-	sz = sizeof(ap1);
-	JEMALLOC_P(mallctl)("thread.allocatedp", &ap1, &sz, NULL, 0);
-	assert(*ap1 == a1);
-	assert(ap0 == ap1);
-
-	usize = JEMALLOC_P(malloc_usable_size)(p);
-	assert(a0 + usize <= a1);
-
-	JEMALLOC_P(free)(p);
-
-	sz = sizeof(d1);
-	JEMALLOC_P(mallctl)("thread.deallocated", &d1, &sz, NULL, 0);
-	sz = sizeof(dp1);
-	JEMALLOC_P(mallctl)("thread.deallocatedp", &dp1, &sz, NULL, 0);
-	assert(*dp1 == d1);
-	assert(dp0 == dp1);
-
-	assert(d0 + usize <= d1);
-
-RETURN:
-	return (NULL);
-}
-
-int
-main(void)
-{
-	int ret = 0;
-	pthread_t thread;
-
-	fprintf(stderr, "Test begin\n");
-
-	thread_start(NULL);
-
-	if (pthread_create(&thread, NULL, thread_start, NULL)
-	    != 0) {
-		fprintf(stderr, "%s(): Error in pthread_create()\n", __func__);
-		ret = 1;
-		goto RETURN;
-	}
-	pthread_join(thread, (void *)&ret);
-
-	thread_start(NULL);
-
-	if (pthread_create(&thread, NULL, thread_start, NULL)
-	    != 0) {
-		fprintf(stderr, "%s(): Error in pthread_create()\n", __func__);
-		ret = 1;
-		goto RETURN;
-	}
-	pthread_join(thread, (void *)&ret);
-
-	thread_start(NULL);
-
-RETURN:
-	fprintf(stderr, "Test end\n");
-	return (ret);
-}
--- a/jemalloc/test/allocm.c
+++ b/jemalloc/test/allocm.c
@@ -1,133 +0,0 @@
-#include <stdio.h>
-#include <stdlib.h>
-#include <stdint.h>
-
-#define	JEMALLOC_MANGLE
-#include "jemalloc_test.h"
-
-#define CHUNK 0x400000
-/* #define MAXALIGN ((size_t)0x80000000000LLU) */
-#define MAXALIGN ((size_t)0x2000000LLU)
-#define NITER 4
-
-int
-main(void)
-{
-	int r;
-	void *p;
-	size_t sz, alignment, total, tsz;
-	unsigned i;
-	void *ps[NITER];
-
-	fprintf(stderr, "Test begin\n");
-
-	sz = 0;
-	r = JEMALLOC_P(allocm)(&p, &sz, 42, 0);
-	if (r != ALLOCM_SUCCESS) {
-		fprintf(stderr, "Unexpected allocm() error\n");
-		abort();
-	}
-	if (sz < 42)
-		fprintf(stderr, "Real size smaller than expected\n");
-	if (JEMALLOC_P(dallocm)(p, 0) != ALLOCM_SUCCESS)
-		fprintf(stderr, "Unexpected dallocm() error\n");
-
-	r = JEMALLOC_P(allocm)(&p, NULL, 42, 0);
-	if (r != ALLOCM_SUCCESS) {
-		fprintf(stderr, "Unexpected allocm() error\n");
-		abort();
-	}
-	if (JEMALLOC_P(dallocm)(p, 0) != ALLOCM_SUCCESS)
-		fprintf(stderr, "Unexpected dallocm() error\n");
-
-	r = JEMALLOC_P(allocm)(&p, NULL, 42, ALLOCM_ZERO);
-	if (r != ALLOCM_SUCCESS) {
-		fprintf(stderr, "Unexpected allocm() error\n");
-		abort();
-	}
-	if (JEMALLOC_P(dallocm)(p, 0) != ALLOCM_SUCCESS)
-		fprintf(stderr, "Unexpected dallocm() error\n");
-
-#if LG_SIZEOF_PTR == 3
-	alignment = 0x8000000000000000LLU;
-	sz        = 0x8000000000000000LLU;
-#else
-	alignment = 0x80000000LU;
-	sz        = 0x80000000LU;
-#endif
-	r = JEMALLOC_P(allocm)(&p, NULL, sz, ALLOCM_ALIGN(alignment));
-	if (r == ALLOCM_SUCCESS) {
-		fprintf(stderr,
-		    "Expected error for allocm(&p, %zu, 0x%x)\n",
-		    sz, ALLOCM_ALIGN(alignment));
-	}
-
-#if LG_SIZEOF_PTR == 3
-	alignment = 0x4000000000000000LLU;
-	sz        = 0x8400000000000001LLU;
-#else
-	alignment = 0x40000000LU;
-	sz        = 0x84000001LU;
-#endif
-	r = JEMALLOC_P(allocm)(&p, NULL, sz, ALLOCM_ALIGN(alignment));
-	if (r == ALLOCM_SUCCESS) {
-		fprintf(stderr,
-		    "Expected error for allocm(&p, %zu, 0x%x)\n",
-		    sz, ALLOCM_ALIGN(alignment));
-	}
-
-	alignment = 0x10LLU;
-#if LG_SIZEOF_PTR == 3
-	sz   = 0xfffffffffffffff0LLU;
-#else
-	sz   = 0xfffffff0LU;
-#endif
-	r = JEMALLOC_P(allocm)(&p, NULL, sz, ALLOCM_ALIGN(alignment));
-	if (r == ALLOCM_SUCCESS) {
-		fprintf(stderr,
-		    "Expected error for allocm(&p, %zu, 0x%x)\n",
-		    sz, ALLOCM_ALIGN(alignment));
-	}
-
-	for (i = 0; i < NITER; i++)
-		ps[i] = NULL;
-
-	for (alignment = 8;
-	    alignment <= MAXALIGN;
-	    alignment <<= 1) {
-		total = 0;
-		fprintf(stderr, "Alignment: %zu\n", alignment);
-		for (sz = 1;
-		    sz < 3 * alignment && sz < (1U << 31);
-		    sz += (alignment >> (LG_SIZEOF_PTR-1)) - 1) {
-			for (i = 0; i < NITER; i++) {
-				r = JEMALLOC_P(allocm)(&ps[i], NULL, sz,
-				    ALLOCM_ALIGN(alignment) | ALLOCM_ZERO);
-				if (r != ALLOCM_SUCCESS) {
-					fprintf(stderr,
-					    "Error for size %zu (0x%zx): %d\n",
-					    sz, sz, r);
-					exit(1);
-				}
-				if ((uintptr_t)p & (alignment-1)) {
-					fprintf(stderr,
-					    "%p inadequately aligned for"
-					    " alignment: %zu\n", p, alignment);
-				}
-				JEMALLOC_P(sallocm)(ps[i], &tsz, 0);
-				total += tsz;
-				if (total >= (MAXALIGN << 1))
-					break;
-			}
-			for (i = 0; i < NITER; i++) {
-				if (ps[i] != NULL) {
-					JEMALLOC_P(dallocm)(ps[i], 0);
-					ps[i] = NULL;
-				}
-			}
-		}
-	}
-
-	fprintf(stderr, "Test end\n");
-	return (0);
-}
--- a/jemalloc/test/jemalloc_test.h.in
+++ b/jemalloc/test/jemalloc_test.h.in
@@ -1,6 +0,0 @@
-/*
- * This header should be included by tests, rather than directly including
- * jemalloc/jemalloc.h, because --with-install-suffix may cause the header to
- * have a different name.
- */
-#include "jemalloc/jemalloc@install_suffix@.h"
--- a/jemalloc/test/rallocm.c
+++ b/jemalloc/test/rallocm.c
@@ -1,117 +0,0 @@
-#include <stdio.h>
-#include <stdlib.h>
-#include <string.h>
-
-#define	JEMALLOC_MANGLE
-#include "jemalloc_test.h"
-
-int
-main(void)
-{
-	void *p, *q;
-	size_t sz, tsz;
-	int r;
-
-	fprintf(stderr, "Test begin\n");
-
-	r = JEMALLOC_P(allocm)(&p, &sz, 42, 0);
-	if (r != ALLOCM_SUCCESS) {
-		fprintf(stderr, "Unexpected allocm() error\n");
-		abort();
-	}
-
-	q = p;
-	r = JEMALLOC_P(rallocm)(&q, &tsz, sz, 0, ALLOCM_NO_MOVE);
-	if (r != ALLOCM_SUCCESS)
-		fprintf(stderr, "Unexpected rallocm() error\n");
-	if (q != p)
-		fprintf(stderr, "Unexpected object move\n");
-	if (tsz != sz) {
-		fprintf(stderr, "Unexpected size change: %zu --> %zu\n",
-		    sz, tsz);
-	}
-
-	q = p;
-	r = JEMALLOC_P(rallocm)(&q, &tsz, sz, 5, ALLOCM_NO_MOVE);
-	if (r != ALLOCM_SUCCESS)
-		fprintf(stderr, "Unexpected rallocm() error\n");
-	if (q != p)
-		fprintf(stderr, "Unexpected object move\n");
-	if (tsz != sz) {
-		fprintf(stderr, "Unexpected size change: %zu --> %zu\n",
-		    sz, tsz);
-	}
-
-	q = p;
-	r = JEMALLOC_P(rallocm)(&q, &tsz, sz + 5, 0, ALLOCM_NO_MOVE);
-	if (r != ALLOCM_ERR_NOT_MOVED)
-		fprintf(stderr, "Unexpected rallocm() result\n");
-	if (q != p)
-		fprintf(stderr, "Unexpected object move\n");
-	if (tsz != sz) {
-		fprintf(stderr, "Unexpected size change: %zu --> %zu\n",
-		    sz, tsz);
-	}
-
-	q = p;
-	r = JEMALLOC_P(rallocm)(&q, &tsz, sz + 5, 0, 0);
-	if (r != ALLOCM_SUCCESS)
-		fprintf(stderr, "Unexpected rallocm() error\n");
-	if (q == p)
-		fprintf(stderr, "Expected object move\n");
-	if (tsz == sz) {
-		fprintf(stderr, "Expected size change: %zu --> %zu\n",
-		    sz, tsz);
-	}
-	p = q;
-	sz = tsz;
-
-	r = JEMALLOC_P(rallocm)(&q, &tsz, 8192, 0, 0);
-	if (r != ALLOCM_SUCCESS)
-		fprintf(stderr, "Unexpected rallocm() error\n");
-	if (q == p)
-		fprintf(stderr, "Expected object move\n");
-	if (tsz == sz) {
-		fprintf(stderr, "Expected size change: %zu --> %zu\n",
-		    sz, tsz);
-	}
-	p = q;
-	sz = tsz;
-
-	r = JEMALLOC_P(rallocm)(&q, &tsz, 16384, 0, 0);
-	if (r != ALLOCM_SUCCESS)
-		fprintf(stderr, "Unexpected rallocm() error\n");
-	if (tsz == sz) {
-		fprintf(stderr, "Expected size change: %zu --> %zu\n",
-		    sz, tsz);
-	}
-	p = q;
-	sz = tsz;
-
-	r = JEMALLOC_P(rallocm)(&q, &tsz, 8192, 0, ALLOCM_NO_MOVE);
-	if (r != ALLOCM_SUCCESS)
-		fprintf(stderr, "Unexpected rallocm() error\n");
-	if (q != p)
-		fprintf(stderr, "Unexpected object move\n");
-	if (tsz == sz) {
-		fprintf(stderr, "Expected size change: %zu --> %zu\n",
-		    sz, tsz);
-	}
-	sz = tsz;
-
-	r = JEMALLOC_P(rallocm)(&q, &tsz, 16384, 0, ALLOCM_NO_MOVE);
-	if (r != ALLOCM_SUCCESS)
-		fprintf(stderr, "Unexpected rallocm() error\n");
-	if (q != p)
-		fprintf(stderr, "Unexpected object move\n");
-	if (tsz == sz) {
-		fprintf(stderr, "Expected size change: %zu --> %zu\n",
-		    sz, tsz);
-	}
-	sz = tsz;
-
-	JEMALLOC_P(dallocm)(p, 0);
-
-	fprintf(stderr, "Test end\n");
-	return (0);
-}
--- a/jemalloc/test/thread_arena.c
+++ b/jemalloc/test/thread_arena.c
@@ -1,82 +0,0 @@
-#include <stdio.h>
-#include <stdlib.h>
-#include <pthread.h>
-#include <string.h>
-
-#define	JEMALLOC_MANGLE
-#include "jemalloc_test.h"
-
-void *
-thread_start(void *arg)
-{
-	unsigned main_arena_ind = *(unsigned *)arg;
-	void *p;
-	unsigned arena_ind;
-	size_t size;
-	int err;
-
-	p = JEMALLOC_P(malloc)(1);
-	if (p == NULL) {
-		fprintf(stderr, "%s(): Error in malloc()\n", __func__);
-		return (void *)1;
-	}
-
-	size = sizeof(arena_ind);
-	if ((err = JEMALLOC_P(mallctl)("thread.arena", &arena_ind, &size,
-	    &main_arena_ind, sizeof(main_arena_ind)))) {
-		fprintf(stderr, "%s(): Error in mallctl(): %s\n", __func__,
-		    strerror(err));
-		return (void *)1;
-	}
-
-	return (NULL);
-}
-
-int
-main(void)
-{
-	int ret = 0;
-	void *p;
-	unsigned arena_ind;
-	size_t size;
-	int err;
-	pthread_t thread;
-
-	fprintf(stderr, "Test begin\n");
-
-	p = JEMALLOC_P(malloc)(1);
-	if (p == NULL) {
-		fprintf(stderr, "%s(): Error in malloc()\n", __func__);
-		ret = 1;
-		goto RETURN;
-	}
-
-	size = sizeof(arena_ind);
-	if ((err = JEMALLOC_P(mallctl)("thread.arena", &arena_ind, &size, NULL,
-	    0))) {
-		fprintf(stderr, "%s(): Error in mallctl(): %s\n", __func__,
-		    strerror(err));
-		ret = 1;
-		goto RETURN;
-	}
-
-	if (pthread_create(&thread, NULL, thread_start, (void *)&arena_ind)
-	    != 0) {
-		fprintf(stderr, "%s(): Error in pthread_create()\n", __func__);
-		ret = 1;
-		goto RETURN;
-	}
-	pthread_join(thread, (void *)&ret);
-
-	if (pthread_create(&thread, NULL, thread_start, (void *)&arena_ind)
-	    != 0) {
-		fprintf(stderr, "%s(): Error in pthread_create()\n", __func__);
-		ret = 1;
-		goto RETURN;
-	}
-	pthread_join(thread, (void *)&ret);
-
-RETURN:
-	fprintf(stderr, "Test end\n");
-	return (ret);
-}
--- a/src/arena.c
+++ b/src/arena.c
--- a/src/atomic.c
+++ b/src/atomic.c
@@ -0,0 +1,2 @@
+#define	JEMALLOC_ATOMIC_C_
+#include "jemalloc/internal/jemalloc_internal.h"
--- a/jemalloc/src/base.c
+++ b/jemalloc/src/base.c
@@ -4,7 +4,7 @@
 /******************************************************************************/
 /* Data. */

-malloc_mutex_t	base_mtx;
+static malloc_mutex_t	base_mtx;

 /*
 * Current pages that are being used for internal memory allocations.  These
@@ -32,7 +32,8 @@ base_pages_alloc(size_t minsize)
 	assert(minsize != 0);
 	csize = CHUNK_CEILING(minsize);
 	zero = false;
-	base_pages = chunk_alloc(csize, true, &zero);
+	base_pages = chunk_alloc(csize, chunksize, true, &zero,
+	    chunk_dss_prec_get());
 	if (base_pages == NULL)
 		return (true);
 	base_next_addr = base_pages;
@@ -66,6 +67,17 @@ base_alloc(size_t size)
 	return (ret);
 }

+void *
+base_calloc(size_t number, size_t size)
+{
+	void *ret = base_alloc(number * size);
+
+	if (ret != NULL)
+		memset(ret, 0, number * size);
+
+	return (ret);
+}
+
 extent_node_t *
 base_node_alloc(void)
 {
@@ -104,3 +116,24 @@ base_boot(void)

 	return (false);
 }
+
+void
+base_prefork(void)
+{
+
+	malloc_mutex_prefork(&base_mtx);
+}
+
+void
+base_postfork_parent(void)
+{
+
+	malloc_mutex_postfork_parent(&base_mtx);
+}
+
+void
+base_postfork_child(void)
+{
+
+	malloc_mutex_postfork_child(&base_mtx);
+}
--- a/src/bitmap.c
+++ b/src/bitmap.c
@@ -0,0 +1,90 @@
+#define JEMALLOC_BITMAP_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+/* Function prototypes for non-inline static functions. */
+
+static size_t	bits2groups(size_t nbits);
+
+/******************************************************************************/
+
+static size_t
+bits2groups(size_t nbits)
+{
+
+	return ((nbits >> LG_BITMAP_GROUP_NBITS) +
+	    !!(nbits & BITMAP_GROUP_NBITS_MASK));
+}
+
+void
+bitmap_info_init(bitmap_info_t *binfo, size_t nbits)
+{
+	unsigned i;
+	size_t group_count;
+
+	assert(nbits > 0);
+	assert(nbits <= (ZU(1) << LG_BITMAP_MAXBITS));
+
+	/*
+	 * Compute the number of groups necessary to store nbits bits, and
+	 * progressively work upward through the levels until reaching a level
+	 * that requires only one group.
+	 */
+	binfo->levels[0].group_offset = 0;
+	group_count = bits2groups(nbits);
+	for (i = 1; group_count > 1; i++) {
+		assert(i < BITMAP_MAX_LEVELS);
+		binfo->levels[i].group_offset = binfo->levels[i-1].group_offset
+		    + group_count;
+		group_count = bits2groups(group_count);
+	}
+	binfo->levels[i].group_offset = binfo->levels[i-1].group_offset
+	    + group_count;
+	binfo->nlevels = i;
+	binfo->nbits = nbits;
+}
+
+size_t
+bitmap_info_ngroups(const bitmap_info_t *binfo)
+{
+
+	return (binfo->levels[binfo->nlevels].group_offset << LG_SIZEOF_BITMAP);
+}
+
+size_t
+bitmap_size(size_t nbits)
+{
+	bitmap_info_t binfo;
+
+	bitmap_info_init(&binfo, nbits);
+	return (bitmap_info_ngroups(&binfo));
+}
+
+void
+bitmap_init(bitmap_t *bitmap, const bitmap_info_t *binfo)
+{
+	size_t extra;
+	unsigned i;
+
+	/*
+	 * Bits are actually inverted with regard to the external bitmap
+	 * interface, so the bitmap starts out with all 1 bits, except for
+	 * trailing unused bits (if any).  Note that each group uses bit 0 to
+	 * correspond to the first logical bit in the group, so extra bits
+	 * are the most significant bits of the last group.
+	 */
+	memset(bitmap, 0xffU, binfo->levels[binfo->nlevels].group_offset <<
+	    LG_SIZEOF_BITMAP);
+	extra = (BITMAP_GROUP_NBITS - (binfo->nbits & BITMAP_GROUP_NBITS_MASK))
+	    & BITMAP_GROUP_NBITS_MASK;
+	if (extra != 0)
+		bitmap[binfo->levels[1].group_offset - 1] >>= extra;
+	for (i = 1; i < binfo->nlevels; i++) {
+		size_t group_count = binfo->levels[i].group_offset -
+		    binfo->levels[i-1].group_offset;
+		extra = (BITMAP_GROUP_NBITS - (group_count &
+		    BITMAP_GROUP_NBITS_MASK)) & BITMAP_GROUP_NBITS_MASK;
+		if (extra != 0)
+			bitmap[binfo->levels[i+1].group_offset - 1] >>= extra;
+	}
+}
--- a/src/chunk.c
+++ b/src/chunk.c
@@ -0,0 +1,393 @@
+#define	JEMALLOC_CHUNK_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+/* Data. */
+
+const char	*opt_dss = DSS_DEFAULT;
+size_t		opt_lg_chunk = LG_CHUNK_DEFAULT;
+
+malloc_mutex_t	chunks_mtx;
+chunk_stats_t	stats_chunks;
+
+/*
+ * Trees of chunks that were previously allocated (trees differ only in node
+ * ordering).  These are used when allocating chunks, in an attempt to re-use
+ * address space.  Depending on function, different tree orderings are needed,
+ * which is why there are two trees with the same contents.
+ */
+static extent_tree_t	chunks_szad_mmap;
+static extent_tree_t	chunks_ad_mmap;
+static extent_tree_t	chunks_szad_dss;
+static extent_tree_t	chunks_ad_dss;
+
+rtree_t		*chunks_rtree;
+
+/* Various chunk-related settings. */
+size_t		chunksize;
+size_t		chunksize_mask; /* (chunksize - 1). */
+size_t		chunk_npages;
+size_t		map_bias;
+size_t		arena_maxclass; /* Max size class for arenas. */
+
+/******************************************************************************/
+/* Function prototypes for non-inline static functions. */
+
+static void	*chunk_recycle(extent_tree_t *chunks_szad,
+    extent_tree_t *chunks_ad, size_t size, size_t alignment, bool base,
+    bool *zero);
+static void	chunk_record(extent_tree_t *chunks_szad,
+    extent_tree_t *chunks_ad, void *chunk, size_t size);
+
+/******************************************************************************/
+
+static void *
+chunk_recycle(extent_tree_t *chunks_szad, extent_tree_t *chunks_ad, size_t size,
+    size_t alignment, bool base, bool *zero)
+{
+	void *ret;
+	extent_node_t *node;
+	extent_node_t key;
+	size_t alloc_size, leadsize, trailsize;
+	bool zeroed;
+
+	if (base) {
+		/*
+		 * This function may need to call base_node_{,de}alloc(), but
+		 * the current chunk allocation request is on behalf of the
+		 * base allocator.  Avoid deadlock (and if that weren't an
+		 * issue, potential for infinite recursion) by returning NULL.
+		 */
+		return (NULL);
+	}
+
+	alloc_size = size + alignment - chunksize;
+	/* Beware size_t wrap-around. */
+	if (alloc_size < size)
+		return (NULL);
+	key.addr = NULL;
+	key.size = alloc_size;
+	malloc_mutex_lock(&chunks_mtx);
+	node = extent_tree_szad_nsearch(chunks_szad, &key);
+	if (node == NULL) {
+		malloc_mutex_unlock(&chunks_mtx);
+		return (NULL);
+	}
+	leadsize = ALIGNMENT_CEILING((uintptr_t)node->addr, alignment) -
+	    (uintptr_t)node->addr;
+	assert(node->size >= leadsize + size);
+	trailsize = node->size - leadsize - size;
+	ret = (void *)((uintptr_t)node->addr + leadsize);
+	/* Remove node from the tree. */
+	extent_tree_szad_remove(chunks_szad, node);
+	extent_tree_ad_remove(chunks_ad, node);
+	if (leadsize != 0) {
+		/* Insert the leading space as a smaller chunk. */
+		node->size = leadsize;
+		extent_tree_szad_insert(chunks_szad, node);
+		extent_tree_ad_insert(chunks_ad, node);
+		node = NULL;
+	}
+	if (trailsize != 0) {
+		/* Insert the trailing space as a smaller chunk. */
+		if (node == NULL) {
+			/*
+			 * An additional node is required, but
+			 * base_node_alloc() can cause a new base chunk to be
+			 * allocated.  Drop chunks_mtx in order to avoid
+			 * deadlock, and if node allocation fails, deallocate
+			 * the result before returning an error.
+			 */
+			malloc_mutex_unlock(&chunks_mtx);
+			node = base_node_alloc();
+			if (node == NULL) {
+				chunk_dealloc(ret, size, true);
+				return (NULL);
+			}
+			malloc_mutex_lock(&chunks_mtx);
+		}
+		node->addr = (void *)((uintptr_t)(ret) + size);
+		node->size = trailsize;
+		extent_tree_szad_insert(chunks_szad, node);
+		extent_tree_ad_insert(chunks_ad, node);
+		node = NULL;
+	}
+	malloc_mutex_unlock(&chunks_mtx);
+
+	zeroed = false;
+	if (node != NULL) {
+		if (node->zeroed) {
+			zeroed = true;
+			*zero = true;
+		}
+		base_node_dealloc(node);
+	}
+	if (zeroed == false && *zero) {
+		VALGRIND_MAKE_MEM_UNDEFINED(ret, size);
+		memset(ret, 0, size);
+	}
+	return (ret);
+}
+
+/*
+ * If the caller specifies (*zero == false), it is still possible to receive
+ * zeroed memory, in which case *zero is toggled to true.  arena_chunk_alloc()
+ * takes advantage of this to avoid demanding zeroed chunks, but taking
+ * advantage of them if they are returned.
+ */
+void *
+chunk_alloc(size_t size, size_t alignment, bool base, bool *zero,
+    dss_prec_t dss_prec)
+{
+	void *ret;
+
+	assert(size != 0);
+	assert((size & chunksize_mask) == 0);
+	assert(alignment != 0);
+	assert((alignment & chunksize_mask) == 0);
+
+	/*
+	 * Try to recycle an existing mapping.
+	 */
+
+	/* "primary" dss. */
+	if (config_dss && dss_prec == dss_prec_primary && (ret =
+	    chunk_recycle(&chunks_szad_dss, &chunks_ad_dss, size, alignment,
+	    base, zero)) != NULL)
+		goto label_return;
+	/* mmap. */
+	if ((ret = chunk_recycle(&chunks_szad_mmap, &chunks_ad_mmap, size,
+	    alignment, base, zero)) != NULL)
+		goto label_return;
+	/* "secondary" dss. */
+	if (config_dss && dss_prec == dss_prec_secondary && (ret =
+	    chunk_recycle(&chunks_szad_dss, &chunks_ad_dss, size, alignment,
+	    base, zero)) != NULL)
+		goto label_return;
+
+	/*
+	 * Try to allocate a new mapping.
+	 */
+
+	/* "primary" dss. */
+	if (config_dss && dss_prec == dss_prec_primary && (ret =
+	    chunk_alloc_dss(size, alignment, zero)) != NULL)
+		goto label_return;
+	/* mmap. */
+	if ((ret = chunk_alloc_mmap(size, alignment, zero)) != NULL)
+		goto label_return;
+	/* "secondary" dss. */
+	if (config_dss && dss_prec == dss_prec_secondary && (ret =
+	    chunk_alloc_dss(size, alignment, zero)) != NULL)
+		goto label_return;
+
+	/* All strategies for allocation failed. */
+	ret = NULL;
+label_return:
+	if (config_ivsalloc && base == false && ret != NULL) {
+		if (rtree_set(chunks_rtree, (uintptr_t)ret, ret)) {
+			chunk_dealloc(ret, size, true);
+			return (NULL);
+		}
+	}
+	if ((config_stats || config_prof) && ret != NULL) {
+		bool gdump;
+		malloc_mutex_lock(&chunks_mtx);
+		if (config_stats)
+			stats_chunks.nchunks += (size / chunksize);
+		stats_chunks.curchunks += (size / chunksize);
+		if (stats_chunks.curchunks > stats_chunks.highchunks) {
+			stats_chunks.highchunks = stats_chunks.curchunks;
+			if (config_prof)
+				gdump = true;
+		} else if (config_prof)
+			gdump = false;
+		malloc_mutex_unlock(&chunks_mtx);
+		if (config_prof && opt_prof && opt_prof_gdump && gdump)
+			prof_gdump();
+	}
+	if (config_debug && *zero && ret != NULL) {
+		size_t i;
+		size_t *p = (size_t *)(uintptr_t)ret;
+
+		VALGRIND_MAKE_MEM_DEFINED(ret, size);
+		for (i = 0; i < size / sizeof(size_t); i++)
+			assert(p[i] == 0);
+	}
+	assert(CHUNK_ADDR2BASE(ret) == ret);
+	return (ret);
+}
+
+static void
+chunk_record(extent_tree_t *chunks_szad, extent_tree_t *chunks_ad, void *chunk,
+    size_t size)
+{
+	bool unzeroed;
+	extent_node_t *xnode, *node, *prev, key;
+
+	unzeroed = pages_purge(chunk, size);
+
+	/*
+	 * Allocate a node before acquiring chunks_mtx even though it might not
+	 * be needed, because base_node_alloc() may cause a new base chunk to
+	 * be allocated, which could cause deadlock if chunks_mtx were already
+	 * held.
+	 */
+	xnode = base_node_alloc();
+
+	malloc_mutex_lock(&chunks_mtx);
+	key.addr = (void *)((uintptr_t)chunk + size);
+	node = extent_tree_ad_nsearch(chunks_ad, &key);
+	/* Try to coalesce forward. */
+	if (node != NULL && node->addr == key.addr) {
+		/*
+		 * Coalesce chunk with the following address range.  This does
+		 * not change the position within chunks_ad, so only
+		 * remove/insert from/into chunks_szad.
+		 */
+		extent_tree_szad_remove(chunks_szad, node);
+		node->addr = chunk;
+		node->size += size;
+		node->zeroed = (node->zeroed && (unzeroed == false));
+		extent_tree_szad_insert(chunks_szad, node);
+		if (xnode != NULL)
+			base_node_dealloc(xnode);
+	} else {
+		/* Coalescing forward failed, so insert a new node. */
+		if (xnode == NULL) {
+			/*
+			 * base_node_alloc() failed, which is an exceedingly
+			 * unlikely failure.  Leak chunk; its pages have
+			 * already been purged, so this is only a virtual
+			 * memory leak.
+			 */
+			malloc_mutex_unlock(&chunks_mtx);
+			return;
+		}
+		node = xnode;
+		node->addr = chunk;
+		node->size = size;
+		node->zeroed = (unzeroed == false);
+		extent_tree_ad_insert(chunks_ad, node);
+		extent_tree_szad_insert(chunks_szad, node);
+	}
+
+	/* Try to coalesce backward. */
+	prev = extent_tree_ad_prev(chunks_ad, node);
+	if (prev != NULL && (void *)((uintptr_t)prev->addr + prev->size) ==
+	    chunk) {
+		/*
+		 * Coalesce chunk with the previous address range.  This does
+		 * not change the position within chunks_ad, so only
+		 * remove/insert node from/into chunks_szad.
+		 */
+		extent_tree_szad_remove(chunks_szad, prev);
+		extent_tree_ad_remove(chunks_ad, prev);
+
+		extent_tree_szad_remove(chunks_szad, node);
+		node->addr = prev->addr;
+		node->size += prev->size;
+		node->zeroed = (node->zeroed && prev->zeroed);
+		extent_tree_szad_insert(chunks_szad, node);
+
+		base_node_dealloc(prev);
+	}
+	malloc_mutex_unlock(&chunks_mtx);
+}
+
+void
+chunk_unmap(void *chunk, size_t size)
+{
+	assert(chunk != NULL);
+	assert(CHUNK_ADDR2BASE(chunk) == chunk);
+	assert(size != 0);
+	assert((size & chunksize_mask) == 0);
+
+	if (config_dss && chunk_in_dss(chunk))
+		chunk_record(&chunks_szad_dss, &chunks_ad_dss, chunk, size);
+	else if (chunk_dealloc_mmap(chunk, size))
+		chunk_record(&chunks_szad_mmap, &chunks_ad_mmap, chunk, size);
+}
+
+void
+chunk_dealloc(void *chunk, size_t size, bool unmap)
+{
+
+	assert(chunk != NULL);
+	assert(CHUNK_ADDR2BASE(chunk) == chunk);
+	assert(size != 0);
+	assert((size & chunksize_mask) == 0);
+
+	if (config_ivsalloc)
+		rtree_set(chunks_rtree, (uintptr_t)chunk, NULL);
+	if (config_stats || config_prof) {
+		malloc_mutex_lock(&chunks_mtx);
+		assert(stats_chunks.curchunks >= (size / chunksize));
+		stats_chunks.curchunks -= (size / chunksize);
+		malloc_mutex_unlock(&chunks_mtx);
+	}
+
+	if (unmap)
+		chunk_unmap(chunk, size);
+}
+
+bool
+chunk_boot(void)
+{
+
+	/* Set variables according to the value of opt_lg_chunk. */
+	chunksize = (ZU(1) << opt_lg_chunk);
+	assert(chunksize >= PAGE);
+	chunksize_mask = chunksize - 1;
+	chunk_npages = (chunksize >> LG_PAGE);
+
+	if (config_stats || config_prof) {
+		if (malloc_mutex_init(&chunks_mtx))
+			return (true);
+		memset(&stats_chunks, 0, sizeof(chunk_stats_t));
+	}
+	if (config_dss && chunk_dss_boot())
+		return (true);
+	extent_tree_szad_new(&chunks_szad_mmap);
+	extent_tree_ad_new(&chunks_ad_mmap);
+	extent_tree_szad_new(&chunks_szad_dss);
+	extent_tree_ad_new(&chunks_ad_dss);
+	if (config_ivsalloc) {
+		chunks_rtree = rtree_new((ZU(1) << (LG_SIZEOF_PTR+3)) -
+		    opt_lg_chunk);
+		if (chunks_rtree == NULL)
+			return (true);
+	}
+
+	return (false);
+}
+
+void
+chunk_prefork(void)
+{
+
+	malloc_mutex_lock(&chunks_mtx);
+	if (config_ivsalloc)
+		rtree_prefork(chunks_rtree);
+	chunk_dss_prefork();
+}
+
+void
+chunk_postfork_parent(void)
+{
+
+	chunk_dss_postfork_parent();
+	if (config_ivsalloc)
+		rtree_postfork_parent(chunks_rtree);
+	malloc_mutex_postfork_parent(&chunks_mtx);
+}
+
+void
+chunk_postfork_child(void)
+{
+
+	chunk_dss_postfork_child();
+	if (config_ivsalloc)
+		rtree_postfork_child(chunks_rtree);
+	malloc_mutex_postfork_child(&chunks_mtx);
+}
--- a/src/chunk_dss.c
+++ b/src/chunk_dss.c
@@ -0,0 +1,197 @@
+#define	JEMALLOC_CHUNK_DSS_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+/******************************************************************************/
+/* Data. */
+
+const char	*dss_prec_names[] = {
+	"disabled",
+	"primary",
+	"secondary",
+	"N/A"
+};
+
+/* Current dss precedence default, used when creating new arenas. */
+static dss_prec_t	dss_prec_default = DSS_PREC_DEFAULT;
+
+/*
+ * Protects sbrk() calls.  This avoids malloc races among threads, though it
+ * does not protect against races with threads that call sbrk() directly.
+ */
+static malloc_mutex_t	dss_mtx;
+
+/* Base address of the DSS. */
+static void		*dss_base;
+/* Current end of the DSS, or ((void *)-1) if the DSS is exhausted. */
+static void		*dss_prev;
+/* Current upper limit on DSS addresses. */
+static void		*dss_max;
+
+/******************************************************************************/
+
+#ifndef JEMALLOC_HAVE_SBRK
+static void *
+sbrk(intptr_t increment)
+{
+
+	not_implemented();
+
+	return (NULL);
+}
+#endif
+
+dss_prec_t
+chunk_dss_prec_get(void)
+{
+	dss_prec_t ret;
+
+	if (config_dss == false)
+		return (dss_prec_disabled);
+	malloc_mutex_lock(&dss_mtx);
+	ret = dss_prec_default;
+	malloc_mutex_unlock(&dss_mtx);
+	return (ret);
+}
+
+bool
+chunk_dss_prec_set(dss_prec_t dss_prec)
+{
+
+	if (config_dss == false)
+		return (true);
+	malloc_mutex_lock(&dss_mtx);
+	dss_prec_default = dss_prec;
+	malloc_mutex_unlock(&dss_mtx);
+	return (false);
+}
+
+void *
+chunk_alloc_dss(size_t size, size_t alignment, bool *zero)
+{
+	void *ret;
+
+	cassert(config_dss);
+	assert(size > 0 && (size & chunksize_mask) == 0);
+	assert(alignment > 0 && (alignment & chunksize_mask) == 0);
+
+	/*
+	 * sbrk() uses a signed increment argument, so take care not to
+	 * interpret a huge allocation request as a negative increment.
+	 */
+	if ((intptr_t)size < 0)
+		return (NULL);
+
+	malloc_mutex_lock(&dss_mtx);
+	if (dss_prev != (void *)-1) {
+		size_t gap_size, cpad_size;
+		void *cpad, *dss_next;
+		intptr_t incr;
+
+		/*
+		 * The loop is necessary to recover from races with other
+		 * threads that are using the DSS for something other than
+		 * malloc.
+		 */
+		do {
+			/* Get the current end of the DSS. */
+			dss_max = sbrk(0);
+			/*
+			 * Calculate how much padding is necessary to
+			 * chunk-align the end of the DSS.
+			 */
+			gap_size = (chunksize - CHUNK_ADDR2OFFSET(dss_max)) &
+			    chunksize_mask;
+			/*
+			 * Compute how much chunk-aligned pad space (if any) is
+			 * necessary to satisfy alignment.  This space can be
+			 * recycled for later use.
+			 */
+			cpad = (void *)((uintptr_t)dss_max + gap_size);
+			ret = (void *)ALIGNMENT_CEILING((uintptr_t)dss_max,
+			    alignment);
+			cpad_size = (uintptr_t)ret - (uintptr_t)cpad;
+			dss_next = (void *)((uintptr_t)ret + size);
+			if ((uintptr_t)ret < (uintptr_t)dss_max ||
+			    (uintptr_t)dss_next < (uintptr_t)dss_max) {
+				/* Wrap-around. */
+				malloc_mutex_unlock(&dss_mtx);
+				return (NULL);
+			}
+			incr = gap_size + cpad_size + size;
+			dss_prev = sbrk(incr);
+			if (dss_prev == dss_max) {
+				/* Success. */
+				dss_max = dss_next;
+				malloc_mutex_unlock(&dss_mtx);
+				if (cpad_size != 0)
+					chunk_unmap(cpad, cpad_size);
+				if (*zero) {
+					VALGRIND_MAKE_MEM_UNDEFINED(ret, size);
+					memset(ret, 0, size);
+				}
+				return (ret);
+			}
+		} while (dss_prev != (void *)-1);
+	}
+	malloc_mutex_unlock(&dss_mtx);
+
+	return (NULL);
+}
+
+bool
+chunk_in_dss(void *chunk)
+{
+	bool ret;
+
+	cassert(config_dss);
+
+	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_dss_boot(void)
+{
+
+	cassert(config_dss);
+
+	if (malloc_mutex_init(&dss_mtx))
+		return (true);
+	dss_base = sbrk(0);
+	dss_prev = dss_base;
+	dss_max = dss_base;
+
+	return (false);
+}
+
+void
+chunk_dss_prefork(void)
+{
+
+	if (config_dss)
+		malloc_mutex_prefork(&dss_mtx);
+}
+
+void
+chunk_dss_postfork_parent(void)
+{
+
+	if (config_dss)
+		malloc_mutex_postfork_parent(&dss_mtx);
+}
+
+void
+chunk_dss_postfork_child(void)
+{
+
+	if (config_dss)
+		malloc_mutex_postfork_child(&dss_mtx);
+}
+
+/******************************************************************************/
--- a/src/chunk_mmap.c
+++ b/src/chunk_mmap.c
@@ -0,0 +1,210 @@
+#define	JEMALLOC_CHUNK_MMAP_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+/* Function prototypes for non-inline static functions. */
+
+static void	*pages_map(void *addr, size_t size);
+static void	pages_unmap(void *addr, size_t size);
+static void	*chunk_alloc_mmap_slow(size_t size, size_t alignment,
+    bool *zero);
+
+/******************************************************************************/
+
+static void *
+pages_map(void *addr, size_t size)
+{
+	void *ret;
+
+	assert(size != 0);
+
+#ifdef _WIN32
+	/*
+	 * If VirtualAlloc can't allocate at the given address when one is
+	 * given, it fails and returns NULL.
+	 */
+	ret = VirtualAlloc(addr, size, MEM_COMMIT | MEM_RESERVE,
+	    PAGE_READWRITE);
+#else
+	/*
+	 * We don't use MAP_FIXED here, because it can cause the *replacement*
+	 * of existing mappings, and we only want to create new mappings.
+	 */
+	ret = mmap(addr, size, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANON,
+	    -1, 0);
+	assert(ret != NULL);
+
+	if (ret == MAP_FAILED)
+		ret = NULL;
+	else if (addr != NULL && ret != addr) {
+		/*
+		 * We succeeded in mapping memory, but not in the right place.
+		 */
+		if (munmap(ret, size) == -1) {
+			char buf[BUFERROR_BUF];
+
+			buferror(buf, sizeof(buf));
+			malloc_printf("<jemalloc: Error in munmap(): %s\n",
+			    buf);
+			if (opt_abort)
+				abort();
+		}
+		ret = NULL;
+	}
+#endif
+	assert(ret == NULL || (addr == NULL && ret != addr)
+	    || (addr != NULL && ret == addr));
+	return (ret);
+}
+
+static void
+pages_unmap(void *addr, size_t size)
+{
+
+#ifdef _WIN32
+	if (VirtualFree(addr, 0, MEM_RELEASE) == 0)
+#else
+	if (munmap(addr, size) == -1)
+#endif
+	{
+		char buf[BUFERROR_BUF];
+
+		buferror(buf, sizeof(buf));
+		malloc_printf("<jemalloc>: Error in "
+#ifdef _WIN32
+		              "VirtualFree"
+#else
+		              "munmap"
+#endif
+		              "(): %s\n", buf);
+		if (opt_abort)
+			abort();
+	}
+}
+
+static void *
+pages_trim(void *addr, size_t alloc_size, size_t leadsize, size_t size)
+{
+	void *ret = (void *)((uintptr_t)addr + leadsize);
+
+	assert(alloc_size >= leadsize + size);
+#ifdef _WIN32
+	{
+		void *new_addr;
+
+		pages_unmap(addr, alloc_size);
+		new_addr = pages_map(ret, size);
+		if (new_addr == ret)
+			return (ret);
+		if (new_addr)
+			pages_unmap(new_addr, size);
+		return (NULL);
+	}
+#else
+	{
+		size_t trailsize = alloc_size - leadsize - size;
+
+		if (leadsize != 0)
+			pages_unmap(addr, leadsize);
+		if (trailsize != 0)
+			pages_unmap((void *)((uintptr_t)ret + size), trailsize);
+		return (ret);
+	}
+#endif
+}
+
+bool
+pages_purge(void *addr, size_t length)
+{
+	bool unzeroed;
+
+#ifdef _WIN32
+	VirtualAlloc(addr, length, MEM_RESET, PAGE_READWRITE);
+	unzeroed = true;
+#else
+#  ifdef JEMALLOC_PURGE_MADVISE_DONTNEED
+#    define JEMALLOC_MADV_PURGE MADV_DONTNEED
+#    define JEMALLOC_MADV_ZEROS true
+#  elif defined(JEMALLOC_PURGE_MADVISE_FREE)
+#    define JEMALLOC_MADV_PURGE MADV_FREE
+#    define JEMALLOC_MADV_ZEROS false
+#  else
+#    error "No method defined for purging unused dirty pages."
+#  endif
+	int err = madvise(addr, length, JEMALLOC_MADV_PURGE);
+	unzeroed = (JEMALLOC_MADV_ZEROS == false || err != 0);
+#  undef JEMALLOC_MADV_PURGE
+#  undef JEMALLOC_MADV_ZEROS
+#endif
+	return (unzeroed);
+}
+
+static void *
+chunk_alloc_mmap_slow(size_t size, size_t alignment, bool *zero)
+{
+	void *ret, *pages;
+	size_t alloc_size, leadsize;
+
+	alloc_size = size + alignment - PAGE;
+	/* Beware size_t wrap-around. */
+	if (alloc_size < size)
+		return (NULL);
+	do {
+		pages = pages_map(NULL, alloc_size);
+		if (pages == NULL)
+			return (NULL);
+		leadsize = ALIGNMENT_CEILING((uintptr_t)pages, alignment) -
+		    (uintptr_t)pages;
+		ret = pages_trim(pages, alloc_size, leadsize, size);
+	} while (ret == NULL);
+
+	assert(ret != NULL);
+	*zero = true;
+	return (ret);
+}
+
+void *
+chunk_alloc_mmap(size_t size, size_t alignment, bool *zero)
+{
+	void *ret;
+	size_t offset;
+
+	/*
+	 * Ideally, there would be a way to specify alignment to mmap() (like
+	 * NetBSD has), but in the absence of such a feature, we have to work
+	 * hard to efficiently create aligned mappings.  The reliable, but
+	 * slow method is to create a mapping that is over-sized, then trim the
+	 * excess.  However, that always results in one or two calls to
+	 * pages_unmap().
+	 *
+	 * Optimistically try mapping precisely the right amount before falling
+	 * back to the slow method, with the expectation that the optimistic
+	 * approach works most of the time.
+	 */
+
+	assert(alignment != 0);
+	assert((alignment & chunksize_mask) == 0);
+
+	ret = pages_map(NULL, size);
+	if (ret == NULL)
+		return (NULL);
+	offset = ALIGNMENT_ADDR2OFFSET(ret, alignment);
+	if (offset != 0) {
+		pages_unmap(ret, size);
+		return (chunk_alloc_mmap_slow(size, alignment, zero));
+	}
+
+	assert(ret != NULL);
+	*zero = true;
+	return (ret);
+}
+
+bool
+chunk_dealloc_mmap(void *chunk, size_t size)
+{
+
+	if (config_munmap)
+		pages_unmap(chunk, size);
+
+	return (config_munmap == false);
+}
--- a/jemalloc/src/ckh.c
+++ b/jemalloc/src/ckh.c
@@ -34,7 +34,7 @@
 * respectively.
 *
 ******************************************************************************/
-#define	CKH_C_
+#define	JEMALLOC_CKH_C_
 #include "jemalloc/internal/jemalloc_internal.h"

 /******************************************************************************/
@@ -73,7 +73,6 @@ ckh_isearch(ckh_t *ckh, const void *key)
 	size_t hash1, hash2, bucket, cell;

 	assert(ckh != NULL);
-	assert(ckh->magic == CKH_MAGIC);

 	ckh->hash(key, ckh->lg_curbuckets, &hash1, &hash2);

@@ -100,7 +99,7 @@ ckh_try_bucket_insert(ckh_t *ckh, size_t bucket, const void *key,
 	 * Cycle through the cells in the bucket, starting at a random position.
 	 * The randomness avoids worst-case search overhead as buckets fill up.
 	 */
-	prn32(offset, LG_CKH_BUCKET_CELLS, ckh->prn_state, CKH_A, CKH_C);
+	prng32(offset, LG_CKH_BUCKET_CELLS, ckh->prng_state, CKH_A, CKH_C);
 	for (i = 0; i < (ZU(1) << LG_CKH_BUCKET_CELLS); i++) {
 		cell = &ckh->tab[(bucket << LG_CKH_BUCKET_CELLS) +
 		    ((i + offset) & ((ZU(1) << LG_CKH_BUCKET_CELLS) - 1))];
@@ -142,7 +141,7 @@ ckh_evict_reloc_insert(ckh_t *ckh, size_t argbucket, void const **argkey,
 		 * were an item for which both hashes indicated the same
 		 * bucket.
 		 */
-		prn32(i, LG_CKH_BUCKET_CELLS, ckh->prn_state, CKH_A, CKH_C);
+		prng32(i, LG_CKH_BUCKET_CELLS, ckh->prng_state, CKH_A, CKH_C);
 		cell = &ckh->tab[(bucket << LG_CKH_BUCKET_CELLS) + i];
 		assert(cell->key != NULL);

@@ -262,12 +261,18 @@ ckh_grow(ckh_t *ckh)
 	lg_prevbuckets = ckh->lg_curbuckets;
 	lg_curcells = ckh->lg_curbuckets + LG_CKH_BUCKET_CELLS;
 	while (true) {
+		size_t usize;
+
 		lg_curcells++;
-		tab = (ckhc_t *)ipalloc(sizeof(ckhc_t) << lg_curcells,
-		    ZU(1) << LG_CACHELINE, true);
+		usize = sa2u(sizeof(ckhc_t) << lg_curcells, CACHELINE);
+		if (usize == 0) {
+			ret = true;
+			goto label_return;
+		}
+		tab = (ckhc_t *)ipalloc(usize, CACHELINE, true);
 		if (tab == NULL) {
 			ret = true;
-			goto RETURN;
+			goto label_return;
 		}
 		/* Swap in new table. */
 		ttab = ckh->tab;
@@ -287,7 +292,7 @@ ckh_grow(ckh_t *ckh)
 	}

 	ret = false;
-RETURN:
+label_return:
 	return (ret);
 }

@@ -295,7 +300,7 @@ static void
 ckh_shrink(ckh_t *ckh)
 {
 	ckhc_t *tab, *ttab;
-	size_t lg_curcells;
+	size_t lg_curcells, usize;
 	unsigned lg_prevbuckets;

 	/*
@@ -304,8 +309,10 @@ ckh_shrink(ckh_t *ckh)
 	 */
 	lg_prevbuckets = ckh->lg_curbuckets;
 	lg_curcells = ckh->lg_curbuckets + LG_CKH_BUCKET_CELLS - 1;
-	tab = (ckhc_t *)ipalloc(sizeof(ckhc_t) << lg_curcells,
-	    ZU(1) << LG_CACHELINE, true);
+	usize = sa2u(sizeof(ckhc_t) << lg_curcells, CACHELINE);
+	if (usize == 0)
+		return;
+	tab = (ckhc_t *)ipalloc(usize, CACHELINE, true);
 	if (tab == NULL) {
 		/*
 		 * An OOM error isn't worth propagating, since it doesn't
@@ -340,7 +347,7 @@ bool
 ckh_new(ckh_t *ckh, size_t minitems, ckh_hash_t *hash, ckh_keycomp_t *keycomp)
 {
 	bool ret;
-	size_t mincells;
+	size_t mincells, usize;
 	unsigned lg_mincells;

 	assert(minitems > 0);
@@ -354,7 +361,7 @@ ckh_new(ckh_t *ckh, size_t minitems, ckh_hash_t *hash, ckh_keycomp_t *keycomp)
 	ckh->ninserts = 0;
 	ckh->nrelocs = 0;
 #endif
-	ckh->prn_state = 42; /* Value doesn't really matter. */
+	ckh->prng_state = 42; /* Value doesn't really matter. */
 	ckh->count = 0;

 	/*
@@ -375,19 +382,19 @@ ckh_new(ckh_t *ckh, size_t minitems, ckh_hash_t *hash, ckh_keycomp_t *keycomp)
 	ckh->hash = hash;
 	ckh->keycomp = keycomp;

-	ckh->tab = (ckhc_t *)ipalloc(sizeof(ckhc_t) << lg_mincells,
-	    (ZU(1) << LG_CACHELINE), true);
+	usize = sa2u(sizeof(ckhc_t) << lg_mincells, CACHELINE);
+	if (usize == 0) {
+		ret = true;
+		goto label_return;
+	}
+	ckh->tab = (ckhc_t *)ipalloc(usize, CACHELINE, true);
 	if (ckh->tab == NULL) {
 		ret = true;
-		goto RETURN;
+		goto label_return;
 	}

-#ifdef JEMALLOC_DEBUG
-	ckh->magic = CKH_MAGIC;
-#endif
-
 	ret = false;
-RETURN:
+label_return:
 	return (ret);
 }

@@ -396,7 +403,6 @@ ckh_delete(ckh_t *ckh)
 {

 	assert(ckh != NULL);
-	assert(ckh->magic == CKH_MAGIC);

 #ifdef CKH_VERBOSE
 	malloc_printf(
@@ -421,7 +427,6 @@ ckh_count(ckh_t *ckh)
 {

 	assert(ckh != NULL);
-	assert(ckh->magic == CKH_MAGIC);

 	return (ckh->count);
 }
@@ -452,7 +457,6 @@ ckh_insert(ckh_t *ckh, const void *key, const void *data)
 	bool ret;

 	assert(ckh != NULL);
-	assert(ckh->magic == CKH_MAGIC);
 	assert(ckh_search(ckh, key, NULL, NULL));

 #ifdef CKH_COUNT
@@ -462,12 +466,12 @@ ckh_insert(ckh_t *ckh, const void *key, const void *data)
 	while (ckh_try_insert(ckh, &key, &data)) {
 		if (ckh_grow(ckh)) {
 			ret = true;
-			goto RETURN;
+			goto label_return;
 		}
 	}

 	ret = false;
-RETURN:
+label_return:
 	return (ret);
 }

@@ -477,7 +481,6 @@ ckh_remove(ckh_t *ckh, const void *searchkey, void **key, void **data)
 	size_t cell;

 	assert(ckh != NULL);
-	assert(ckh->magic == CKH_MAGIC);

 	cell = ckh_isearch(ckh, searchkey);
 	if (cell != SIZE_T_MAX) {
@@ -509,7 +512,6 @@ ckh_search(ckh_t *ckh, const void *searchkey, void **key, void **data)
 	size_t cell;

 	assert(ckh != NULL);
-	assert(ckh->magic == CKH_MAGIC);

 	cell = ckh_isearch(ckh, searchkey);
 	if (cell != SIZE_T_MAX) {
@@ -533,7 +535,7 @@ ckh_string_hash(const void *key, unsigned minbits, size_t *hash1, size_t *hash2)
 	assert(hash1 != NULL);
 	assert(hash2 != NULL);

-	h = hash(key, strlen((const char *)key), 0x94122f335b332aeaLLU);
+	h = hash(key, strlen((const char *)key), UINT64_C(0x94122f335b332aea));
 	if (minbits <= 32) {
 		/*
 		 * Avoid doing multiple hashes, since a single hash provides
@@ -544,7 +546,7 @@ ckh_string_hash(const void *key, unsigned minbits, size_t *hash1, size_t *hash2)
 	} else {
 		ret1 = h;
 		ret2 = hash(key, strlen((const char *)key),
-		    0x8432a476666bbc13U);
+		    UINT64_C(0x8432a476666bbc13));
 	}

 	*hash1 = ret1;
@@ -581,7 +583,7 @@ ckh_pointer_hash(const void *key, unsigned minbits, size_t *hash1,
 	u.i = 0;
 #endif
 	u.v = key;
-	h = hash(&u.i, sizeof(u.i), 0xd983396e68886082LLU);
+	h = hash(&u.i, sizeof(u.i), UINT64_C(0xd983396e68886082));
 	if (minbits <= 32) {
 		/*
 		 * Avoid doing multiple hashes, since a single hash provides
@@ -592,7 +594,7 @@ ckh_pointer_hash(const void *key, unsigned minbits, size_t *hash1,
 	} else {
 		assert(SIZEOF_PTR == 8);
 		ret1 = h;
-		ret2 = hash(&u.i, sizeof(u.i), 0x5e2be9aff8709a5dLLU);
+		ret2 = hash(&u.i, sizeof(u.i), UINT64_C(0x5e2be9aff8709a5d));
 	}

 	*hash1 = ret1;
--- a/src/ctl.c
+++ b/src/ctl.c
--- a/jemalloc/src/extent.c
+++ b/jemalloc/src/extent.c
@@ -3,7 +3,6 @@

 /******************************************************************************/

-#if (defined(JEMALLOC_SWAP) || defined(JEMALLOC_DSS))
 static inline int
 extent_szad_comp(extent_node_t *a, extent_node_t *b)
 {
@@ -25,7 +24,6 @@ extent_szad_comp(extent_node_t *a, extent_node_t *b)
 /* Generate red-black tree functions. */
 rb_gen(, extent_tree_szad_, extent_tree_t, extent_node_t, link_szad,
    extent_szad_comp)
-#endif

 static inline int
 extent_ad_comp(extent_node_t *a, extent_node_t *b)
--- a/jemalloc/src/hash.c
+++ b/jemalloc/src/hash.c
@@ -1,2 +1,2 @@
-#define	HASH_C_
+#define	JEMALLOC_HASH_C_
 #include "jemalloc/internal/jemalloc_internal.h"
--- a/jemalloc/src/huge.c
+++ b/jemalloc/src/huge.c
@@ -4,11 +4,9 @@
 /******************************************************************************/
 /* Data. */

-#ifdef JEMALLOC_STATS
 uint64_t	huge_nmalloc;
 uint64_t	huge_ndalloc;
 size_t		huge_allocated;
-#endif

 malloc_mutex_t	huge_mtx;

@@ -19,10 +17,18 @@ static extent_tree_t	huge;

 void *
 huge_malloc(size_t size, bool zero)
+{
+
+	return (huge_palloc(size, chunksize, zero));
+}
+
+void *
+huge_palloc(size_t size, size_t alignment, bool zero)
 {
 	void *ret;
 	size_t csize;
 	extent_node_t *node;
+	bool is_zeroed;

 	/* Allocate one or more contiguous chunks for this request. */

@@ -37,7 +43,13 @@ huge_malloc(size_t size, bool zero)
 	if (node == NULL)
 		return (NULL);

-	ret = chunk_alloc(csize, false, &zero);
+	/*
+	 * Copy zero into is_zeroed and pass the copy to chunk_alloc(), so that
+	 * it is possible to make correct junk/zero fill decisions below.
+	 */
+	is_zeroed = zero;
+	ret = chunk_alloc(csize, alignment, false, &is_zeroed,
+	    chunk_dss_prec_get());
 	if (ret == NULL) {
 		base_node_dealloc(node);
 		return (NULL);
@@ -49,104 +61,19 @@ huge_malloc(size_t size, bool zero)

 	malloc_mutex_lock(&huge_mtx);
 	extent_tree_ad_insert(&huge, node);
-#ifdef JEMALLOC_STATS
-	huge_nmalloc++;
-	huge_allocated += csize;
-#endif
+	if (config_stats) {
+		stats_cactive_add(csize);
+		huge_nmalloc++;
+		huge_allocated += csize;
+	}
 	malloc_mutex_unlock(&huge_mtx);

-#ifdef JEMALLOC_FILL
-	if (zero == false) {
+	if (config_fill && zero == false) {
 		if (opt_junk)
 			memset(ret, 0xa5, csize);
-		else if (opt_zero)
+		else if (opt_zero && is_zeroed == false)
 			memset(ret, 0, csize);
 	}
-#endif
-
-	return (ret);
-}
-
-/* Only handles large allocations that require more than chunk alignment. */
-void *
-huge_palloc(size_t size, size_t alignment, bool zero)
-{
-	void *ret;
-	size_t alloc_size, chunk_size, offset;
-	extent_node_t *node;
-
-	/*
-	 * This allocation requires alignment that is even larger than chunk
-	 * alignment.  This means that huge_malloc() isn't good enough.
-	 *
-	 * Allocate almost twice as many chunks as are demanded by the size or
-	 * alignment, in order to assure the alignment can be achieved, then
-	 * unmap leading and trailing chunks.
-	 */
-	assert(alignment > chunksize);
-
-	chunk_size = CHUNK_CEILING(size);
-
-	if (size >= alignment)
-		alloc_size = chunk_size + alignment - chunksize;
-	else
-		alloc_size = (alignment << 1) - chunksize;
-
-	/* Allocate an extent node with which to track the chunk. */
-	node = base_node_alloc();
-	if (node == NULL)
-		return (NULL);
-
-	ret = chunk_alloc(alloc_size, false, &zero);
-	if (ret == NULL) {
-		base_node_dealloc(node);
-		return (NULL);
-	}
-
-	offset = (uintptr_t)ret & (alignment - 1);
-	assert((offset & chunksize_mask) == 0);
-	assert(offset < alloc_size);
-	if (offset == 0) {
-		/* Trim trailing space. */
-		chunk_dealloc((void *)((uintptr_t)ret + chunk_size), alloc_size
-		    - chunk_size);
-	} else {
-		size_t trailsize;
-
-		/* Trim leading space. */
-		chunk_dealloc(ret, alignment - offset);
-
-		ret = (void *)((uintptr_t)ret + (alignment - offset));
-
-		trailsize = alloc_size - (alignment - offset) - chunk_size;
-		if (trailsize != 0) {
-		    /* Trim trailing space. */
-		    assert(trailsize < alloc_size);
-		    chunk_dealloc((void *)((uintptr_t)ret + chunk_size),
-			trailsize);
-		}
-	}
-
-	/* Insert node into huge. */
-	node->addr = ret;
-	node->size = chunk_size;
-
-	malloc_mutex_lock(&huge_mtx);
-	extent_tree_ad_insert(&huge, node);
-#ifdef JEMALLOC_STATS
-	huge_nmalloc++;
-	huge_allocated += chunk_size;
-#endif
-	malloc_mutex_unlock(&huge_mtx);
-
-#ifdef JEMALLOC_FILL
-	if (zero == false) {
-		if (opt_junk)
-			memset(ret, 0xa5, chunk_size);
-		else if (opt_zero)
-			memset(ret, 0, chunk_size);
-	}
-#endif

 	return (ret);
 }
@@ -162,12 +89,10 @@ huge_ralloc_no_move(void *ptr, size_t oldsize, size_t size, size_t extra)
 	    && CHUNK_CEILING(oldsize) >= CHUNK_CEILING(size)
 	    && CHUNK_CEILING(oldsize) <= CHUNK_CEILING(size+extra)) {
 		assert(CHUNK_CEILING(oldsize) == oldsize);
-#ifdef JEMALLOC_FILL
-		if (opt_junk && size < oldsize) {
+		if (config_fill && opt_junk && size < oldsize) {
 			memset((void *)((uintptr_t)ptr + size), 0x5a,
 			    oldsize - size);
 		}
-#endif
 		return (ptr);
 	}

@@ -177,7 +102,7 @@ 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)
+    size_t alignment, bool zero, bool try_tcache_dalloc)
 {
 	void *ret;
 	size_t copysize;
@@ -216,22 +141,22 @@ huge_ralloc(void *ptr, size_t oldsize, size_t size, size_t extra,
 	 */
 	copysize = (size < oldsize) ? size : oldsize;

+#ifdef JEMALLOC_MREMAP
 	/*
 	 * Use mremap(2) if this is a huge-->huge reallocation, and neither the
-	 * source nor the destination are in swap or dss.
+	 * source nor the destination are in 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
-	    ) {
+	if (oldsize >= chunksize && (config_dss == false || (chunk_in_dss(ptr)
+	    == false && chunk_in_dss(ret) == false))) {
 		size_t newsize = huge_salloc(ret);

+		/*
+		 * Remove ptr from the tree of huge allocations before
+		 * performing the remap operation, in order to avoid the
+		 * possibility of another thread acquiring that mapping before
+		 * this one removes it from the tree.
+		 */
+		huge_dalloc(ptr, false);
 		if (mremap(ptr, oldsize, newsize, MREMAP_MAYMOVE|MREMAP_FIXED,
 		    ret) == MAP_FAILED) {
 			/*
@@ -244,21 +169,19 @@ huge_ralloc(void *ptr, size_t oldsize, size_t size, size_t extra,
 			 */
 			char buf[BUFERROR_BUF];

-			buferror(errno, buf, sizeof(buf));
-			malloc_write("<jemalloc>: Error in mremap(): ");
-			malloc_write(buf);
-			malloc_write("\n");
+			buferror(buf, sizeof(buf));
+			malloc_printf("<jemalloc>: Error in mremap(): %s\n",
+			    buf);
 			if (opt_abort)
 				abort();
 			memcpy(ret, ptr, copysize);
-			idalloc(ptr);
-		} else
-			huge_dalloc(ptr, false);
+			chunk_dealloc_mmap(ptr, oldsize);
+		}
 	} else
 #endif
 	{
 		memcpy(ret, ptr, copysize);
-		idalloc(ptr);
+		iqallocx(ptr, try_tcache_dalloc);
 	}
 	return (ret);
 }
@@ -277,23 +200,18 @@ huge_dalloc(void *ptr, bool unmap)
 	assert(node->addr == ptr);
 	extent_tree_ad_remove(&huge, node);

-#ifdef JEMALLOC_STATS
-	huge_ndalloc++;
-	huge_allocated -= node->size;
-#endif
+	if (config_stats) {
+		stats_cactive_sub(node->size);
+		huge_ndalloc++;
+		huge_allocated -= node->size;
+	}

 	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);
-#endif
-#endif
-		chunk_dealloc(node->addr, node->size);
-	}
+	if (unmap && config_fill && config_dss && opt_junk)
+		memset(node->addr, 0x5a, node->size);
+
+	chunk_dealloc(node->addr, node->size, unmap);

 	base_node_dealloc(node);
 }
@@ -318,7 +236,6 @@ huge_salloc(const void *ptr)
 	return (ret);
 }

-#ifdef JEMALLOC_PROF
 prof_ctx_t *
 huge_prof_ctx_get(const void *ptr)
 {
@@ -355,7 +272,6 @@ huge_prof_ctx_set(const void *ptr, prof_ctx_t *ctx)

 	malloc_mutex_unlock(&huge_mtx);
 }
-#endif

 bool
 huge_boot(void)
@@ -366,11 +282,32 @@ huge_boot(void)
 		return (true);
 	extent_tree_ad_new(&huge);

-#ifdef JEMALLOC_STATS
-	huge_nmalloc = 0;
-	huge_ndalloc = 0;
-	huge_allocated = 0;
-#endif
+	if (config_stats) {
+		huge_nmalloc = 0;
+		huge_ndalloc = 0;
+		huge_allocated = 0;
+	}

 	return (false);
 }
+
+void
+huge_prefork(void)
+{
+
+	malloc_mutex_prefork(&huge_mtx);
+}
+
+void
+huge_postfork_parent(void)
+{
+
+	malloc_mutex_postfork_parent(&huge_mtx);
+}
+
+void
+huge_postfork_child(void)
+{
+
+	malloc_mutex_postfork_child(&huge_mtx);
+}
--- a/src/jemalloc.c
+++ b/src/jemalloc.c
--- a/jemalloc/src/mb.c
+++ b/jemalloc/src/mb.c
@@ -1,2 +1,2 @@
-#define	MB_C_
+#define	JEMALLOC_MB_C_
 #include "jemalloc/internal/jemalloc_internal.h"
--- a/src/mutex.c
+++ b/src/mutex.c
@@ -0,0 +1,149 @@
+#define	JEMALLOC_MUTEX_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+#if defined(JEMALLOC_LAZY_LOCK) && !defined(_WIN32)
+#include <dlfcn.h>
+#endif
+
+#ifndef _CRT_SPINCOUNT
+#define _CRT_SPINCOUNT 4000
+#endif
+
+/******************************************************************************/
+/* Data. */
+
+#ifdef JEMALLOC_LAZY_LOCK
+bool isthreaded = false;
+#endif
+#ifdef JEMALLOC_MUTEX_INIT_CB
+static bool		postpone_init = true;
+static malloc_mutex_t	*postponed_mutexes = NULL;
+#endif
+
+#if defined(JEMALLOC_LAZY_LOCK) && !defined(_WIN32)
+static void	pthread_create_once(void);
+#endif
+
+/******************************************************************************/
+/*
+ * We intercept pthread_create() calls in order to toggle isthreaded if the
+ * process goes multi-threaded.
+ */
+
+#if defined(JEMALLOC_LAZY_LOCK) && !defined(_WIN32)
+static int (*pthread_create_fptr)(pthread_t *__restrict, const pthread_attr_t *,
+    void *(*)(void *), void *__restrict);
+
+static void
+pthread_create_once(void)
+{
+
+	pthread_create_fptr = dlsym(RTLD_NEXT, "pthread_create");
+	if (pthread_create_fptr == NULL) {
+		malloc_write("<jemalloc>: Error in dlsym(RTLD_NEXT, "
+		    "\"pthread_create\")\n");
+		abort();
+	}
+
+	isthreaded = true;
+}
+
+JEMALLOC_EXPORT int
+pthread_create(pthread_t *__restrict thread,
+    const pthread_attr_t *__restrict attr, void *(*start_routine)(void *),
+    void *__restrict arg)
+{
+	static pthread_once_t once_control = PTHREAD_ONCE_INIT;
+
+	pthread_once(&once_control, pthread_create_once);
+
+	return (pthread_create_fptr(thread, attr, start_routine, arg));
+}
+#endif
+
+/******************************************************************************/
+
+#ifdef JEMALLOC_MUTEX_INIT_CB
+JEMALLOC_EXPORT int	_pthread_mutex_init_calloc_cb(pthread_mutex_t *mutex,
+    void *(calloc_cb)(size_t, size_t));
+#endif
+
+bool
+malloc_mutex_init(malloc_mutex_t *mutex)
+{
+
+#ifdef _WIN32
+	if (!InitializeCriticalSectionAndSpinCount(&mutex->lock,
+	    _CRT_SPINCOUNT))
+		return (true);
+#elif (defined(JEMALLOC_OSSPIN))
+	mutex->lock = 0;
+#elif (defined(JEMALLOC_MUTEX_INIT_CB))
+	if (postpone_init) {
+		mutex->postponed_next = postponed_mutexes;
+		postponed_mutexes = mutex;
+	} else {
+		if (_pthread_mutex_init_calloc_cb(&mutex->lock, base_calloc) !=
+		    0)
+			return (true);
+	}
+#else
+	pthread_mutexattr_t attr;
+
+	if (pthread_mutexattr_init(&attr) != 0)
+		return (true);
+	pthread_mutexattr_settype(&attr, MALLOC_MUTEX_TYPE);
+	if (pthread_mutex_init(&mutex->lock, &attr) != 0) {
+		pthread_mutexattr_destroy(&attr);
+		return (true);
+	}
+	pthread_mutexattr_destroy(&attr);
+#endif
+	return (false);
+}
+
+void
+malloc_mutex_prefork(malloc_mutex_t *mutex)
+{
+
+	malloc_mutex_lock(mutex);
+}
+
+void
+malloc_mutex_postfork_parent(malloc_mutex_t *mutex)
+{
+
+	malloc_mutex_unlock(mutex);
+}
+
+void
+malloc_mutex_postfork_child(malloc_mutex_t *mutex)
+{
+
+#ifdef JEMALLOC_MUTEX_INIT_CB
+	malloc_mutex_unlock(mutex);
+#else
+	if (malloc_mutex_init(mutex)) {
+		malloc_printf("<jemalloc>: Error re-initializing mutex in "
+		    "child\n");
+		if (opt_abort)
+			abort();
+	}
+#endif
+}
+
+bool
+mutex_boot(void)
+{
+
+#ifdef JEMALLOC_MUTEX_INIT_CB
+	postpone_init = false;
+	while (postponed_mutexes != NULL) {
+		if (_pthread_mutex_init_calloc_cb(&postponed_mutexes->lock,
+		    base_calloc) != 0)
+			return (true);
+		postponed_mutexes = postponed_mutexes->postponed_next;
+	}
+#endif
+	return (false);
+}
--- a/jemalloc/src/prof.c
+++ b/jemalloc/src/prof.c
--- a/src/quarantine.c
+++ b/src/quarantine.c
@@ -0,0 +1,210 @@
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/*
+ * quarantine pointers close to NULL are used to encode state information that
+ * is used for cleaning up during thread shutdown.
+ */
+#define	QUARANTINE_STATE_REINCARNATED	((quarantine_t *)(uintptr_t)1)
+#define	QUARANTINE_STATE_PURGATORY	((quarantine_t *)(uintptr_t)2)
+#define	QUARANTINE_STATE_MAX		QUARANTINE_STATE_PURGATORY
+
+/******************************************************************************/
+/* Data. */
+
+typedef struct quarantine_obj_s quarantine_obj_t;
+typedef struct quarantine_s quarantine_t;
+
+struct quarantine_obj_s {
+	void	*ptr;
+	size_t	usize;
+};
+
+struct quarantine_s {
+	size_t			curbytes;
+	size_t			curobjs;
+	size_t			first;
+#define	LG_MAXOBJS_INIT 10
+	size_t			lg_maxobjs;
+	quarantine_obj_t	objs[1]; /* Dynamically sized ring buffer. */
+};
+
+static void	quarantine_cleanup(void *arg);
+
+malloc_tsd_data(static, quarantine, quarantine_t *, NULL)
+malloc_tsd_funcs(JEMALLOC_INLINE, quarantine, quarantine_t *, NULL,
+    quarantine_cleanup)
+
+/******************************************************************************/
+/* Function prototypes for non-inline static functions. */
+
+static quarantine_t	*quarantine_init(size_t lg_maxobjs);
+static quarantine_t	*quarantine_grow(quarantine_t *quarantine);
+static void	quarantine_drain(quarantine_t *quarantine, size_t upper_bound);
+
+/******************************************************************************/
+
+static quarantine_t *
+quarantine_init(size_t lg_maxobjs)
+{
+	quarantine_t *quarantine;
+
+	quarantine = (quarantine_t *)imalloc(offsetof(quarantine_t, objs) +
+	    ((ZU(1) << lg_maxobjs) * sizeof(quarantine_obj_t)));
+	if (quarantine == NULL)
+		return (NULL);
+	quarantine->curbytes = 0;
+	quarantine->curobjs = 0;
+	quarantine->first = 0;
+	quarantine->lg_maxobjs = lg_maxobjs;
+
+	quarantine_tsd_set(&quarantine);
+
+	return (quarantine);
+}
+
+static quarantine_t *
+quarantine_grow(quarantine_t *quarantine)
+{
+	quarantine_t *ret;
+
+	ret = quarantine_init(quarantine->lg_maxobjs + 1);
+	if (ret == NULL)
+		return (quarantine);
+
+	ret->curbytes = quarantine->curbytes;
+	ret->curobjs = quarantine->curobjs;
+	if (quarantine->first + quarantine->curobjs <= (ZU(1) <<
+	    quarantine->lg_maxobjs)) {
+		/* objs ring buffer data are contiguous. */
+		memcpy(ret->objs, &quarantine->objs[quarantine->first],
+		    quarantine->curobjs * sizeof(quarantine_obj_t));
+	} else {
+		/* objs ring buffer data wrap around. */
+		size_t ncopy_a = (ZU(1) << quarantine->lg_maxobjs) -
+		    quarantine->first;
+		size_t ncopy_b = quarantine->curobjs - ncopy_a;
+
+		memcpy(ret->objs, &quarantine->objs[quarantine->first], ncopy_a
+		    * sizeof(quarantine_obj_t));
+		memcpy(&ret->objs[ncopy_a], quarantine->objs, ncopy_b *
+		    sizeof(quarantine_obj_t));
+	}
+
+	return (ret);
+}
+
+static void
+quarantine_drain(quarantine_t *quarantine, size_t upper_bound)
+{
+
+	while (quarantine->curbytes > upper_bound && quarantine->curobjs > 0) {
+		quarantine_obj_t *obj = &quarantine->objs[quarantine->first];
+		assert(obj->usize == isalloc(obj->ptr, config_prof));
+		idalloc(obj->ptr);
+		quarantine->curbytes -= obj->usize;
+		quarantine->curobjs--;
+		quarantine->first = (quarantine->first + 1) & ((ZU(1) <<
+		    quarantine->lg_maxobjs) - 1);
+	}
+}
+
+void
+quarantine(void *ptr)
+{
+	quarantine_t *quarantine;
+	size_t usize = isalloc(ptr, config_prof);
+
+	cassert(config_fill);
+	assert(opt_quarantine);
+
+	quarantine = *quarantine_tsd_get();
+	if ((uintptr_t)quarantine <= (uintptr_t)QUARANTINE_STATE_MAX) {
+		if (quarantine == NULL) {
+			if ((quarantine = quarantine_init(LG_MAXOBJS_INIT)) ==
+			    NULL) {
+				idalloc(ptr);
+				return;
+			}
+		} else {
+			if (quarantine == QUARANTINE_STATE_PURGATORY) {
+				/*
+				 * Make a note that quarantine() was called
+				 * after quarantine_cleanup() was called.
+				 */
+				quarantine = QUARANTINE_STATE_REINCARNATED;
+				quarantine_tsd_set(&quarantine);
+			}
+			idalloc(ptr);
+			return;
+		}
+	}
+	/*
+	 * Drain one or more objects if the quarantine size limit would be
+	 * exceeded by appending ptr.
+	 */
+	if (quarantine->curbytes + usize > opt_quarantine) {
+		size_t upper_bound = (opt_quarantine >= usize) ? opt_quarantine
+		    - usize : 0;
+		quarantine_drain(quarantine, upper_bound);
+	}
+	/* Grow the quarantine ring buffer if it's full. */
+	if (quarantine->curobjs == (ZU(1) << quarantine->lg_maxobjs))
+		quarantine = quarantine_grow(quarantine);
+	/* quarantine_grow() must free a slot if it fails to grow. */
+	assert(quarantine->curobjs < (ZU(1) << quarantine->lg_maxobjs));
+	/* Append ptr if its size doesn't exceed the quarantine size. */
+	if (quarantine->curbytes + usize <= opt_quarantine) {
+		size_t offset = (quarantine->first + quarantine->curobjs) &
+		    ((ZU(1) << quarantine->lg_maxobjs) - 1);
+		quarantine_obj_t *obj = &quarantine->objs[offset];
+		obj->ptr = ptr;
+		obj->usize = usize;
+		quarantine->curbytes += usize;
+		quarantine->curobjs++;
+		if (opt_junk)
+			memset(ptr, 0x5a, usize);
+	} else {
+		assert(quarantine->curbytes == 0);
+		idalloc(ptr);
+	}
+}
+
+static void
+quarantine_cleanup(void *arg)
+{
+	quarantine_t *quarantine = *(quarantine_t **)arg;
+
+	if (quarantine == QUARANTINE_STATE_REINCARNATED) {
+		/*
+		 * Another destructor deallocated memory after this destructor
+		 * was called.  Reset quarantine to QUARANTINE_STATE_PURGATORY
+		 * in order to receive another callback.
+		 */
+		quarantine = QUARANTINE_STATE_PURGATORY;
+		quarantine_tsd_set(&quarantine);
+	} else if (quarantine == QUARANTINE_STATE_PURGATORY) {
+		/*
+		 * The previous time this destructor was called, we set the key
+		 * to QUARANTINE_STATE_PURGATORY so that other destructors
+		 * wouldn't cause re-creation of the quarantine.  This time, do
+		 * nothing, so that the destructor will not be called again.
+		 */
+	} else if (quarantine != NULL) {
+		quarantine_drain(quarantine, 0);
+		idalloc(quarantine);
+		quarantine = QUARANTINE_STATE_PURGATORY;
+		quarantine_tsd_set(&quarantine);
+	}
+}
+
+bool
+quarantine_boot(void)
+{
+
+	cassert(config_fill);
+
+	if (quarantine_tsd_boot())
+		return (true);
+
+	return (false);
+}
--- a/jemalloc/src/rtree.c
+++ b/jemalloc/src/rtree.c
@@ -1,4 +1,4 @@
-#define	RTREE_C_
+#define	JEMALLOC_RTREE_C_
 #include "jemalloc/internal/jemalloc_internal.h"

 rtree_t *
@@ -20,7 +20,10 @@ rtree_new(unsigned bits)
 	memset(ret, 0, offsetof(rtree_t, level2bits) + (sizeof(unsigned) *
 	    height));

-	malloc_mutex_init(&ret->mutex);
+	if (malloc_mutex_init(&ret->mutex)) {
+		/* Leak the rtree. */
+		return (NULL);
+	}
 	ret->height = height;
 	if (bits_per_level * height > bits)
 		ret->level2bits[0] = bits % bits_per_level;
@@ -41,3 +44,24 @@ rtree_new(unsigned bits)

 	return (ret);
 }
+
+void
+rtree_prefork(rtree_t *rtree)
+{
+
+	malloc_mutex_prefork(&rtree->mutex);
+}
+
+void
+rtree_postfork_parent(rtree_t *rtree)
+{
+
+	malloc_mutex_postfork_parent(&rtree->mutex);
+}
+
+void
+rtree_postfork_child(rtree_t *rtree)
+{
+
+	malloc_mutex_postfork_child(&rtree->mutex);
+}
--- a/jemalloc/src/stats.c
+++ b/jemalloc/src/stats.c
@@ -39,136 +39,40 @@

 bool	opt_stats_print = false;

+size_t	stats_cactive = 0;
+
 /******************************************************************************/
 /* Function prototypes for non-inline static functions. */

-#ifdef JEMALLOC_STATS
-static void	malloc_vcprintf(void (*write_cb)(void *, const char *),
-    void *cbopaque, const char *format, va_list ap);
 static void	stats_arena_bins_print(void (*write_cb)(void *, const char *),
    void *cbopaque, unsigned i);
 static void	stats_arena_lruns_print(void (*write_cb)(void *, const char *),
    void *cbopaque, unsigned i);
 static void	stats_arena_print(void (*write_cb)(void *, const char *),
-    void *cbopaque, unsigned i);
-#endif
+    void *cbopaque, unsigned i, bool bins, bool large);

 /******************************************************************************/

-/*
- * We don't want to depend on vsnprintf() for production builds, since that can
- * cause unnecessary bloat for static binaries.  u2s() provides minimal integer
- * printing functionality, so that malloc_printf() use can be limited to
- * JEMALLOC_STATS code.
- */
-char *
-u2s(uint64_t x, unsigned base, char *s)
-{
-	unsigned i;
-
-	i = UMAX2S_BUFSIZE - 1;
-	s[i] = '\0';
-	switch (base) {
-	case 10:
-		do {
-			i--;
-			s[i] = "0123456789"[x % (uint64_t)10];
-			x /= (uint64_t)10;
-		} while (x > 0);
-		break;
-	case 16:
-		do {
-			i--;
-			s[i] = "0123456789abcdef"[x & 0xf];
-			x >>= 4;
-		} while (x > 0);
-		break;
-	default:
-		do {
-			i--;
-			s[i] = "0123456789abcdefghijklmnopqrstuvwxyz"[x %
-			    (uint64_t)base];
-			x /= (uint64_t)base;
-		} while (x > 0);
-	}
-
-	return (&s[i]);
-}
-
-#ifdef JEMALLOC_STATS
-static void
-malloc_vcprintf(void (*write_cb)(void *, const char *), void *cbopaque,
-    const char *format, va_list ap)
-{
-	char buf[4096];
-
-	if (write_cb == NULL) {
-		/*
-		 * The caller did not provide an alternate write_cb callback
-		 * function, so use the default one.  malloc_write() is an
-		 * inline function, so use malloc_message() directly here.
-		 */
-		write_cb = JEMALLOC_P(malloc_message);
-		cbopaque = NULL;
-	}
-
-	vsnprintf(buf, sizeof(buf), format, ap);
-	write_cb(cbopaque, buf);
-}
-
-/*
- * Print to a callback function in such a way as to (hopefully) avoid memory
- * allocation.
- */
-JEMALLOC_ATTR(format(printf, 3, 4))
-void
-malloc_cprintf(void (*write_cb)(void *, const char *), void *cbopaque,
-    const char *format, ...)
-{
-	va_list ap;
-
-	va_start(ap, format);
-	malloc_vcprintf(write_cb, cbopaque, format, ap);
-	va_end(ap);
-}
-
-/*
- * Print to stderr in such a way as to (hopefully) avoid memory allocation.
- */
-JEMALLOC_ATTR(format(printf, 1, 2))
-void
-malloc_printf(const char *format, ...)
-{
-	va_list ap;
-
-	va_start(ap, format);
-	malloc_vcprintf(NULL, NULL, format, ap);
-	va_end(ap);
-}
-#endif
-
-#ifdef JEMALLOC_STATS
 static void
 stats_arena_bins_print(void (*write_cb)(void *, const char *), void *cbopaque,
    unsigned i)
 {
-	size_t pagesize;
+	size_t page;
 	bool config_tcache;
 	unsigned nbins, j, gap_start;

-	CTL_GET("arenas.pagesize", &pagesize, size_t);
+	CTL_GET("arenas.page", &page, size_t);

 	CTL_GET("config.tcache", &config_tcache, bool);
 	if (config_tcache) {
 		malloc_cprintf(write_cb, cbopaque,
-		    "bins:     bin    size regs pgs    allocated      nmalloc"
+		    "bins:     bin  size regs pgs    allocated      nmalloc"
 		    "      ndalloc    nrequests       nfills     nflushes"
-		    "      newruns       reruns      maxruns      curruns\n");
+		    "      newruns       reruns      curruns\n");
 	} else {
 		malloc_cprintf(write_cb, cbopaque,
-		    "bins:     bin    size regs pgs    allocated      nmalloc"
-		    "      ndalloc      newruns       reruns      maxruns"
-		    "      curruns\n");
+		    "bins:     bin  size regs pgs    allocated      nmalloc"
+		    "      ndalloc      newruns       reruns      curruns\n");
 	}
 	CTL_GET("arenas.nbins", &nbins, unsigned);
 	for (j = 0, gap_start = UINT_MAX; j < nbins; j++) {
@@ -179,12 +83,11 @@ stats_arena_bins_print(void (*write_cb)(void *, const char *), void *cbopaque,
 			if (gap_start == UINT_MAX)
 				gap_start = j;
 		} else {
-			unsigned ntbins_, nqbins, ncbins, nsbins;
 			size_t reg_size, run_size, allocated;
 			uint32_t nregs;
 			uint64_t nmalloc, ndalloc, nrequests, nfills, nflushes;
 			uint64_t reruns;
-			size_t highruns, curruns;
+			size_t curruns;

 			if (gap_start != UINT_MAX) {
 				if (j > gap_start + 1) {
@@ -199,10 +102,6 @@ stats_arena_bins_print(void (*write_cb)(void *, const char *), void *cbopaque,
 				}
 				gap_start = UINT_MAX;
 			}
-			CTL_GET("arenas.ntbins", &ntbins_, unsigned);
-			CTL_GET("arenas.nqbins", &nqbins, unsigned);
-			CTL_GET("arenas.ncbins", &ncbins, unsigned);
-			CTL_GET("arenas.nsbins", &nsbins, unsigned);
 			CTL_J_GET("arenas.bin.0.size", &reg_size, size_t);
 			CTL_J_GET("arenas.bin.0.nregs", &nregs, uint32_t);
 			CTL_J_GET("arenas.bin.0.run_size", &run_size, size_t);
@@ -222,36 +121,25 @@ stats_arena_bins_print(void (*write_cb)(void *, const char *), void *cbopaque,
 			}
 			CTL_IJ_GET("stats.arenas.0.bins.0.nreruns", &reruns,
 			    uint64_t);
-			CTL_IJ_GET("stats.arenas.0.bins.0.highruns", &highruns,
-			    size_t);
 			CTL_IJ_GET("stats.arenas.0.bins.0.curruns", &curruns,
 			    size_t);
 			if (config_tcache) {
 				malloc_cprintf(write_cb, cbopaque,
-				    "%13u %1s %5zu %4u %3zu %12zu %12"PRIu64
+				    "%13u %5zu %4u %3zu %12zu %12"PRIu64
 				    " %12"PRIu64" %12"PRIu64" %12"PRIu64
 				    " %12"PRIu64" %12"PRIu64" %12"PRIu64
-				    " %12zu %12zu\n",
-				    j,
-				    j < ntbins_ ? "T" : j < ntbins_ + nqbins ?
-				    "Q" : j < ntbins_ + nqbins + ncbins ? "C" :
-				    "S",
-				    reg_size, nregs, run_size / pagesize,
+				    " %12zu\n",
+				    j, reg_size, nregs, run_size / page,
 				    allocated, nmalloc, ndalloc, nrequests,
-				    nfills, nflushes, nruns, reruns, highruns,
-				    curruns);
+				    nfills, nflushes, nruns, reruns, curruns);
 			} else {
 				malloc_cprintf(write_cb, cbopaque,
-				    "%13u %1s %5zu %4u %3zu %12zu %12"PRIu64
+				    "%13u %5zu %4u %3zu %12zu %12"PRIu64
 				    " %12"PRIu64" %12"PRIu64" %12"PRIu64
-				    " %12zu %12zu\n",
-				    j,
-				    j < ntbins_ ? "T" : j < ntbins_ + nqbins ?
-				    "Q" : j < ntbins_ + nqbins + ncbins ? "C" :
-				    "S",
-				    reg_size, nregs, run_size / pagesize,
+				    " %12zu\n",
+				    j, reg_size, nregs, run_size / page,
 				    allocated, nmalloc, ndalloc, nruns, reruns,
-				    highruns, curruns);
+				    curruns);
 			}
 		}
 	}
@@ -271,18 +159,18 @@ static void
 stats_arena_lruns_print(void (*write_cb)(void *, const char *), void *cbopaque,
    unsigned i)
 {
-	size_t pagesize, nlruns, j;
+	size_t page, nlruns, j;
 	ssize_t gap_start;

-	CTL_GET("arenas.pagesize", &pagesize, size_t);
+	CTL_GET("arenas.page", &page, size_t);

 	malloc_cprintf(write_cb, cbopaque,
 	    "large:   size pages      nmalloc      ndalloc    nrequests"
-	    "      maxruns      curruns\n");
+	    "      curruns\n");
 	CTL_GET("arenas.nlruns", &nlruns, size_t);
 	for (j = 0, gap_start = -1; j < nlruns; j++) {
 		uint64_t nmalloc, ndalloc, nrequests;
-		size_t run_size, highruns, curruns;
+		size_t run_size, curruns;

 		CTL_IJ_GET("stats.arenas.0.lruns.0.nmalloc", &nmalloc,
 		    uint64_t);
@@ -295,8 +183,6 @@ stats_arena_lruns_print(void (*write_cb)(void *, const char *), void *cbopaque,
 				gap_start = j;
 		} else {
 			CTL_J_GET("arenas.lrun.0.size", &run_size, size_t);
-			CTL_IJ_GET("stats.arenas.0.lruns.0.highruns", &highruns,
-			    size_t);
 			CTL_IJ_GET("stats.arenas.0.lruns.0.curruns", &curruns,
 			    size_t);
 			if (gap_start != -1) {
@@ -306,9 +192,9 @@ stats_arena_lruns_print(void (*write_cb)(void *, const char *), void *cbopaque,
 			}
 			malloc_cprintf(write_cb, cbopaque,
 			    "%13zu %5zu %12"PRIu64" %12"PRIu64" %12"PRIu64
-			    " %12zu %12zu\n",
-			    run_size, run_size / pagesize, nmalloc, ndalloc,
-			    nrequests, highruns, curruns);
+			    " %12zu\n",
+			    run_size, run_size / page, nmalloc, ndalloc,
+			    nrequests, curruns);
 		}
 	}
 	if (gap_start != -1)
@@ -317,17 +203,25 @@ stats_arena_lruns_print(void (*write_cb)(void *, const char *), void *cbopaque,

 static void
 stats_arena_print(void (*write_cb)(void *, const char *), void *cbopaque,
-    unsigned i)
+    unsigned i, bool bins, bool large)
 {
-	size_t pagesize, pactive, pdirty, mapped;
+	unsigned nthreads;
+	const char *dss;
+	size_t page, pactive, pdirty, mapped;
 	uint64_t npurge, nmadvise, purged;
 	size_t small_allocated;
 	uint64_t small_nmalloc, small_ndalloc, small_nrequests;
 	size_t large_allocated;
 	uint64_t large_nmalloc, large_ndalloc, large_nrequests;

-	CTL_GET("arenas.pagesize", &pagesize, size_t);
+	CTL_GET("arenas.page", &page, size_t);

+	CTL_I_GET("stats.arenas.0.nthreads", &nthreads, unsigned);
+	malloc_cprintf(write_cb, cbopaque,
+	    "assigned threads: %u\n", nthreads);
+	CTL_I_GET("stats.arenas.0.dss", &dss, const char *);
+	malloc_cprintf(write_cb, cbopaque, "dss allocation precedence: %s\n",
+	    dss);
 	CTL_I_GET("stats.arenas.0.pactive", &pactive, size_t);
 	CTL_I_GET("stats.arenas.0.pdirty", &pdirty, size_t);
 	CTL_I_GET("stats.arenas.0.npurge", &npurge, uint64_t);
@@ -361,15 +255,15 @@ stats_arena_print(void (*write_cb)(void *, const char *), void *cbopaque,
 	    small_nmalloc + large_nmalloc,
 	    small_ndalloc + large_ndalloc,
 	    small_nrequests + large_nrequests);
-	malloc_cprintf(write_cb, cbopaque, "active:  %12zu\n",
-	    pactive * pagesize );
+	malloc_cprintf(write_cb, cbopaque, "active:  %12zu\n", pactive * page);
 	CTL_I_GET("stats.arenas.0.mapped", &mapped, size_t);
 	malloc_cprintf(write_cb, cbopaque, "mapped:  %12zu\n", mapped);

-	stats_arena_bins_print(write_cb, cbopaque, i);
-	stats_arena_lruns_print(write_cb, cbopaque, i);
+	if (bins)
+		stats_arena_bins_print(write_cb, cbopaque, i);
+	if (large)
+		stats_arena_lruns_print(write_cb, cbopaque, i);
 }
-#endif

 void
 stats_print(void (*write_cb)(void *, const char *), void *cbopaque,
@@ -378,7 +272,6 @@ stats_print(void (*write_cb)(void *, const char *), void *cbopaque,
 	int err;
 	uint64_t epoch;
 	size_t u64sz;
-	char s[UMAX2S_BUFSIZE];
 	bool general = true;
 	bool merged = true;
 	bool unmerged = true;
@@ -394,8 +287,7 @@ stats_print(void (*write_cb)(void *, const char *), void *cbopaque,
 	 * */
 	epoch = 1;
 	u64sz = sizeof(uint64_t);
-	err = JEMALLOC_P(mallctl)("epoch", &epoch, &u64sz, &epoch,
-	    sizeof(uint64_t));
+	err = je_mallctl("epoch", &epoch, &u64sz, &epoch, sizeof(uint64_t));
 	if (err != 0) {
 		if (err == EAGAIN) {
 			malloc_write("<jemalloc>: Memory allocation failure in "
@@ -407,42 +299,33 @@ stats_print(void (*write_cb)(void *, const char *), void *cbopaque,
 		abort();
 	}

-	if (write_cb == NULL) {
-		/*
-		 * The caller did not provide an alternate write_cb callback
-		 * function, so use the default one.  malloc_write() is an
-		 * inline function, so use malloc_message() directly here.
-		 */
-		write_cb = JEMALLOC_P(malloc_message);
-		cbopaque = NULL;
-	}
-
 	if (opts != NULL) {
 		unsigned i;

 		for (i = 0; opts[i] != '\0'; i++) {
 			switch (opts[i]) {
-				case 'g':
-					general = false;
-					break;
-				case 'm':
-					merged = false;
-					break;
-				case 'a':
-					unmerged = false;
-					break;
-				case 'b':
-					bins = false;
-					break;
-				case 'l':
-					large = false;
-					break;
-				default:;
+			case 'g':
+				general = false;
+				break;
+			case 'm':
+				merged = false;
+				break;
+			case 'a':
+				unmerged = false;
+				break;
+			case 'b':
+				bins = false;
+				break;
+			case 'l':
+				large = false;
+				break;
+			default:;
 			}
 		}
 	}

-	write_cb(cbopaque, "___ Begin jemalloc statistics ___\n");
+	malloc_cprintf(write_cb, cbopaque,
+	    "___ Begin jemalloc statistics ___\n");
 	if (general) {
 		int err;
 		const char *cpv;
@@ -457,256 +340,146 @@ stats_print(void (*write_cb)(void *, const char *), void *cbopaque,
 		cpsz = sizeof(const char *);

 		CTL_GET("version", &cpv, const char *);
-		write_cb(cbopaque, "Version: ");
-		write_cb(cbopaque, cpv);
-		write_cb(cbopaque, "\n");
+		malloc_cprintf(write_cb, cbopaque, "Version: %s\n", cpv);
 		CTL_GET("config.debug", &bv, bool);
-		write_cb(cbopaque, "Assertions ");
-		write_cb(cbopaque, bv ? "enabled" : "disabled");
-		write_cb(cbopaque, "\n");
+		malloc_cprintf(write_cb, cbopaque, "Assertions %s\n",
+		    bv ? "enabled" : "disabled");

 #define OPT_WRITE_BOOL(n)						\
-		if ((err = JEMALLOC_P(mallctl)("opt."#n, &bv, &bsz,	\
-		    NULL, 0)) == 0) {					\
-			write_cb(cbopaque, "  opt."#n": ");		\
-			write_cb(cbopaque, bv ? "true" : "false");	\
-			write_cb(cbopaque, "\n");			\
+		if ((err = je_mallctl("opt."#n, &bv, &bsz, NULL, 0))	\
+		    == 0) {						\
+			malloc_cprintf(write_cb, cbopaque,		\
+			    "  opt."#n": %s\n", bv ? "true" : "false");	\
 		}
 #define OPT_WRITE_SIZE_T(n)						\
-		if ((err = JEMALLOC_P(mallctl)("opt."#n, &sv, &ssz,	\
-		    NULL, 0)) == 0) {					\
-			write_cb(cbopaque, "  opt."#n": ");		\
-			write_cb(cbopaque, u2s(sv, 10, s));		\
-			write_cb(cbopaque, "\n");			\
+		if ((err = je_mallctl("opt."#n, &sv, &ssz, NULL, 0))	\
+		    == 0) {						\
+			malloc_cprintf(write_cb, cbopaque,		\
+			"  opt."#n": %zu\n", sv);			\
 		}
 #define OPT_WRITE_SSIZE_T(n)						\
-		if ((err = JEMALLOC_P(mallctl)("opt."#n, &ssv, &sssz,	\
-		    NULL, 0)) == 0) {					\
-			if (ssv >= 0) {					\
-				write_cb(cbopaque, "  opt."#n": ");	\
-				write_cb(cbopaque, u2s(ssv, 10, s));	\
-			} else {					\
-				write_cb(cbopaque, "  opt."#n": -");	\
-				write_cb(cbopaque, u2s(-ssv, 10, s));	\
-			}						\
-			write_cb(cbopaque, "\n");			\
+		if ((err = je_mallctl("opt."#n, &ssv, &sssz, NULL, 0))	\
+		    == 0) {						\
+			malloc_cprintf(write_cb, cbopaque,		\
+			    "  opt."#n": %zd\n", ssv);			\
 		}
 #define OPT_WRITE_CHAR_P(n)						\
-		if ((err = JEMALLOC_P(mallctl)("opt."#n, &cpv, &cpsz,	\
-		    NULL, 0)) == 0) {					\
-			write_cb(cbopaque, "  opt."#n": \"");		\
-			write_cb(cbopaque, cpv);			\
-			write_cb(cbopaque, "\"\n");			\
+		if ((err = je_mallctl("opt."#n, &cpv, &cpsz, NULL, 0))	\
+		    == 0) {						\
+			malloc_cprintf(write_cb, cbopaque,		\
+			    "  opt."#n": \"%s\"\n", cpv);		\
 		}

-		write_cb(cbopaque, "Run-time option settings:\n");
+		malloc_cprintf(write_cb, cbopaque,
+		    "Run-time option settings:\n");
 		OPT_WRITE_BOOL(abort)
-		OPT_WRITE_SIZE_T(lg_qspace_max)
-		OPT_WRITE_SIZE_T(lg_cspace_max)
 		OPT_WRITE_SIZE_T(lg_chunk)
+		OPT_WRITE_CHAR_P(dss)
 		OPT_WRITE_SIZE_T(narenas)
 		OPT_WRITE_SSIZE_T(lg_dirty_mult)
 		OPT_WRITE_BOOL(stats_print)
 		OPT_WRITE_BOOL(junk)
+		OPT_WRITE_SIZE_T(quarantine)
+		OPT_WRITE_BOOL(redzone)
 		OPT_WRITE_BOOL(zero)
-		OPT_WRITE_BOOL(sysv)
+		OPT_WRITE_BOOL(utrace)
+		OPT_WRITE_BOOL(valgrind)
 		OPT_WRITE_BOOL(xmalloc)
 		OPT_WRITE_BOOL(tcache)
-		OPT_WRITE_SSIZE_T(lg_tcache_gc_sweep)
 		OPT_WRITE_SSIZE_T(lg_tcache_max)
 		OPT_WRITE_BOOL(prof)
 		OPT_WRITE_CHAR_P(prof_prefix)
-		OPT_WRITE_SIZE_T(lg_prof_bt_max)
 		OPT_WRITE_BOOL(prof_active)
 		OPT_WRITE_SSIZE_T(lg_prof_sample)
 		OPT_WRITE_BOOL(prof_accum)
-		OPT_WRITE_SSIZE_T(lg_prof_tcmax)
 		OPT_WRITE_SSIZE_T(lg_prof_interval)
 		OPT_WRITE_BOOL(prof_gdump)
+		OPT_WRITE_BOOL(prof_final)
 		OPT_WRITE_BOOL(prof_leak)
-		OPT_WRITE_BOOL(overcommit)

 #undef OPT_WRITE_BOOL
 #undef OPT_WRITE_SIZE_T
 #undef OPT_WRITE_SSIZE_T
 #undef OPT_WRITE_CHAR_P

-		write_cb(cbopaque, "CPUs: ");
-		write_cb(cbopaque, u2s(ncpus, 10, s));
-		write_cb(cbopaque, "\n");
+		malloc_cprintf(write_cb, cbopaque, "CPUs: %u\n", ncpus);

 		CTL_GET("arenas.narenas", &uv, unsigned);
-		write_cb(cbopaque, "Max arenas: ");
-		write_cb(cbopaque, u2s(uv, 10, s));
-		write_cb(cbopaque, "\n");
+		malloc_cprintf(write_cb, cbopaque, "Arenas: %u\n", uv);

-		write_cb(cbopaque, "Pointer size: ");
-		write_cb(cbopaque, u2s(sizeof(void *), 10, s));
-		write_cb(cbopaque, "\n");
+		malloc_cprintf(write_cb, cbopaque, "Pointer size: %zu\n",
+		    sizeof(void *));

 		CTL_GET("arenas.quantum", &sv, size_t);
-		write_cb(cbopaque, "Quantum size: ");
-		write_cb(cbopaque, u2s(sv, 10, s));
-		write_cb(cbopaque, "\n");
+		malloc_cprintf(write_cb, cbopaque, "Quantum size: %zu\n", sv);

-		CTL_GET("arenas.cacheline", &sv, size_t);
-		write_cb(cbopaque, "Cacheline size (assumed): ");
-		write_cb(cbopaque, u2s(sv, 10, s));
-		write_cb(cbopaque, "\n");
-
-		CTL_GET("arenas.subpage", &sv, size_t);
-		write_cb(cbopaque, "Subpage spacing: ");
-		write_cb(cbopaque, u2s(sv, 10, s));
-		write_cb(cbopaque, "\n");
-
-		if ((err = JEMALLOC_P(mallctl)("arenas.tspace_min", &sv, &ssz,
-		    NULL, 0)) == 0) {
-			write_cb(cbopaque, "Tiny 2^n-spaced sizes: [");
-			write_cb(cbopaque, u2s(sv, 10, s));
-			write_cb(cbopaque, "..");
-
-			CTL_GET("arenas.tspace_max", &sv, size_t);
-			write_cb(cbopaque, u2s(sv, 10, s));
-			write_cb(cbopaque, "]\n");
-		}
-
-		CTL_GET("arenas.qspace_min", &sv, size_t);
-		write_cb(cbopaque, "Quantum-spaced sizes: [");
-		write_cb(cbopaque, u2s(sv, 10, s));
-		write_cb(cbopaque, "..");
-		CTL_GET("arenas.qspace_max", &sv, size_t);
-		write_cb(cbopaque, u2s(sv, 10, s));
-		write_cb(cbopaque, "]\n");
-
-		CTL_GET("arenas.cspace_min", &sv, size_t);
-		write_cb(cbopaque, "Cacheline-spaced sizes: [");
-		write_cb(cbopaque, u2s(sv, 10, s));
-		write_cb(cbopaque, "..");
-		CTL_GET("arenas.cspace_max", &sv, size_t);
-		write_cb(cbopaque, u2s(sv, 10, s));
-		write_cb(cbopaque, "]\n");
-
-		CTL_GET("arenas.sspace_min", &sv, size_t);
-		write_cb(cbopaque, "Subpage-spaced sizes: [");
-		write_cb(cbopaque, u2s(sv, 10, s));
-		write_cb(cbopaque, "..");
-		CTL_GET("arenas.sspace_max", &sv, size_t);
-		write_cb(cbopaque, u2s(sv, 10, s));
-		write_cb(cbopaque, "]\n");
+		CTL_GET("arenas.page", &sv, size_t);
+		malloc_cprintf(write_cb, cbopaque, "Page size: %zu\n", sv);

 		CTL_GET("opt.lg_dirty_mult", &ssv, ssize_t);
 		if (ssv >= 0) {
-			write_cb(cbopaque,
-			    "Min active:dirty page ratio per arena: ");
-			write_cb(cbopaque, u2s((1U << ssv), 10, s));
-			write_cb(cbopaque, ":1\n");
+			malloc_cprintf(write_cb, cbopaque,
+			    "Min active:dirty page ratio per arena: %u:1\n",
+			    (1U << ssv));
 		} else {
-			write_cb(cbopaque,
+			malloc_cprintf(write_cb, cbopaque,
 			    "Min active:dirty page ratio per arena: N/A\n");
 		}
-		if ((err = JEMALLOC_P(mallctl)("arenas.tcache_max", &sv,
-		    &ssz, NULL, 0)) == 0) {
-			write_cb(cbopaque,
-			    "Maximum thread-cached size class: ");
-			write_cb(cbopaque, u2s(sv, 10, s));
-			write_cb(cbopaque, "\n");
+		if ((err = je_mallctl("arenas.tcache_max", &sv, &ssz, NULL, 0))
+		    == 0) {
+			malloc_cprintf(write_cb, cbopaque,
+			    "Maximum thread-cached size class: %zu\n", sv);
 		}
-		if ((err = JEMALLOC_P(mallctl)("opt.lg_tcache_gc_sweep", &ssv,
-		    &ssz, NULL, 0)) == 0) {
-			size_t tcache_gc_sweep = (1U << ssv);
-			bool tcache_enabled;
-			CTL_GET("opt.tcache", &tcache_enabled, bool);
-			write_cb(cbopaque, "Thread cache GC sweep interval: ");
-			write_cb(cbopaque, tcache_enabled && ssv >= 0 ?
-			    u2s(tcache_gc_sweep, 10, s) : "N/A");
-			write_cb(cbopaque, "\n");
-		}
-		if ((err = JEMALLOC_P(mallctl)("opt.prof", &bv, &bsz, NULL, 0))
-		   == 0 && bv) {
-			CTL_GET("opt.lg_prof_bt_max", &sv, size_t);
-			write_cb(cbopaque, "Maximum profile backtrace depth: ");
-			write_cb(cbopaque, u2s((1U << sv), 10, s));
-			write_cb(cbopaque, "\n");
-
-			CTL_GET("opt.lg_prof_tcmax", &ssv, ssize_t);
-			write_cb(cbopaque,
-			    "Maximum per thread backtrace cache: ");
-			if (ssv >= 0) {
-				write_cb(cbopaque, u2s((1U << ssv), 10, s));
-				write_cb(cbopaque, " (2^");
-				write_cb(cbopaque, u2s(ssv, 10, s));
-				write_cb(cbopaque, ")\n");
-			} else
-				write_cb(cbopaque, "N/A\n");
-
+		if ((err = je_mallctl("opt.prof", &bv, &bsz, NULL, 0)) == 0 &&
+		    bv) {
 			CTL_GET("opt.lg_prof_sample", &sv, size_t);
-			write_cb(cbopaque, "Average profile sample interval: ");
-			write_cb(cbopaque, u2s((((uint64_t)1U) << sv), 10, s));
-			write_cb(cbopaque, " (2^");
-			write_cb(cbopaque, u2s(sv, 10, s));
-			write_cb(cbopaque, ")\n");
+			malloc_cprintf(write_cb, cbopaque,
+			    "Average profile sample interval: %"PRIu64
+			    " (2^%zu)\n", (((uint64_t)1U) << sv), sv);

 			CTL_GET("opt.lg_prof_interval", &ssv, ssize_t);
-			write_cb(cbopaque, "Average profile dump interval: ");
 			if (ssv >= 0) {
-				write_cb(cbopaque, u2s((((uint64_t)1U) << ssv),
-				    10, s));
-				write_cb(cbopaque, " (2^");
-				write_cb(cbopaque, u2s(ssv, 10, s));
-				write_cb(cbopaque, ")\n");
-			} else
-				write_cb(cbopaque, "N/A\n");
+				malloc_cprintf(write_cb, cbopaque,
+				    "Average profile dump interval: %"PRIu64
+				    " (2^%zd)\n",
+				    (((uint64_t)1U) << ssv), ssv);
+			} else {
+				malloc_cprintf(write_cb, cbopaque,
+				    "Average profile dump interval: N/A\n");
+			}
 		}
-		CTL_GET("arenas.chunksize", &sv, size_t);
-		write_cb(cbopaque, "Chunk size: ");
-		write_cb(cbopaque, u2s(sv, 10, s));
 		CTL_GET("opt.lg_chunk", &sv, size_t);
-		write_cb(cbopaque, " (2^");
-		write_cb(cbopaque, u2s(sv, 10, s));
-		write_cb(cbopaque, ")\n");
+		malloc_cprintf(write_cb, cbopaque, "Chunk size: %zu (2^%zu)\n",
+		    (ZU(1) << sv), sv);
 	}

-#ifdef JEMALLOC_STATS
-	{
-		int err;
-		size_t ssz;
+	if (config_stats) {
+		size_t *cactive;
 		size_t allocated, active, mapped;
-		size_t chunks_current, chunks_high, swap_avail;
+		size_t chunks_current, chunks_high;
 		uint64_t chunks_total;
 		size_t huge_allocated;
 		uint64_t huge_nmalloc, huge_ndalloc;

-		ssz = sizeof(size_t);
-
+		CTL_GET("stats.cactive", &cactive, size_t *);
 		CTL_GET("stats.allocated", &allocated, size_t);
 		CTL_GET("stats.active", &active, size_t);
 		CTL_GET("stats.mapped", &mapped, size_t);
 		malloc_cprintf(write_cb, cbopaque,
-		    "Allocated: %zu, active: %zu, mapped: %zu\n", allocated,
-		    active, mapped);
+		    "Allocated: %zu, active: %zu, mapped: %zu\n",
+		    allocated, active, mapped);
+		malloc_cprintf(write_cb, cbopaque,
+		    "Current active ceiling: %zu\n", atomic_read_z(cactive));

 		/* Print chunk stats. */
 		CTL_GET("stats.chunks.total", &chunks_total, uint64_t);
 		CTL_GET("stats.chunks.high", &chunks_high, size_t);
 		CTL_GET("stats.chunks.current", &chunks_current, size_t);
-		if ((err = JEMALLOC_P(mallctl)("swap.avail", &swap_avail, &ssz,
-		    NULL, 0)) == 0) {
-			size_t lg_chunk;
-
-			malloc_cprintf(write_cb, cbopaque, "chunks: nchunks   "
-			    "highchunks    curchunks   swap_avail\n");
-			CTL_GET("opt.lg_chunk", &lg_chunk, size_t);
-			malloc_cprintf(write_cb, cbopaque,
-			    "  %13"PRIu64"%13zu%13zu%13zu\n",
-			    chunks_total, chunks_high, chunks_current,
-			    swap_avail << lg_chunk);
-		} else {
-			malloc_cprintf(write_cb, cbopaque, "chunks: nchunks   "
-			    "highchunks    curchunks\n");
-			malloc_cprintf(write_cb, cbopaque,
-			    "  %13"PRIu64"%13zu%13zu\n",
-			    chunks_total, chunks_high, chunks_current);
-		}
+		malloc_cprintf(write_cb, cbopaque, "chunks: nchunks   "
+		    "highchunks    curchunks\n");
+		malloc_cprintf(write_cb, cbopaque,
+		    "  %13"PRIu64" %12zu %12zu\n",
+		    chunks_total, chunks_high, chunks_current);

 		/* Print huge stats. */
 		CTL_GET("stats.huge.nmalloc", &huge_nmalloc, uint64_t);
@@ -723,11 +496,11 @@ stats_print(void (*write_cb)(void *, const char *), void *cbopaque,

 			CTL_GET("arenas.narenas", &narenas, unsigned);
 			{
-				bool initialized[narenas];
+				VARIABLE_ARRAY(bool, initialized, narenas);
 				size_t isz;
 				unsigned i, ninitialized;

-				isz = sizeof(initialized);
+				isz = sizeof(bool) * narenas;
 				xmallctl("arenas.initialized", initialized,
 				    &isz, NULL, 0);
 				for (i = ninitialized = 0; i < narenas; i++) {
@@ -735,12 +508,12 @@ stats_print(void (*write_cb)(void *, const char *), void *cbopaque,
 						ninitialized++;
 				}

-				if (ninitialized > 1) {
+				if (ninitialized > 1 || unmerged == false) {
 					/* Print merged arena stats. */
 					malloc_cprintf(write_cb, cbopaque,
 					    "\nMerged arenas stats:\n");
 					stats_arena_print(write_cb, cbopaque,
-					    narenas);
+					    narenas, bins, large);
 				}
 			}
 		}
@@ -752,11 +525,11 @@ stats_print(void (*write_cb)(void *, const char *), void *cbopaque,

 			CTL_GET("arenas.narenas", &narenas, unsigned);
 			{
-				bool initialized[narenas];
+				VARIABLE_ARRAY(bool, initialized, narenas);
 				size_t isz;
 				unsigned i;

-				isz = sizeof(initialized);
+				isz = sizeof(bool) * narenas;
 				xmallctl("arenas.initialized", initialized,
 				    &isz, NULL, 0);

@@ -766,12 +539,11 @@ stats_print(void (*write_cb)(void *, const char *), void *cbopaque,
 						    cbopaque,
 						    "\narenas[%u]:\n", i);
 						stats_arena_print(write_cb,
-						    cbopaque, i);
+						    cbopaque, i, bins, large);
 					}
 				}
 			}
 		}
 	}
-#endif /* #ifdef JEMALLOC_STATS */
-	write_cb(cbopaque, "--- End jemalloc statistics ---\n");
+	malloc_cprintf(write_cb, cbopaque, "--- End jemalloc statistics ---\n");
 }
--- a/src/tcache.c
+++ b/src/tcache.c
@@ -0,0 +1,474 @@
+#define	JEMALLOC_TCACHE_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+/* Data. */
+
+malloc_tsd_data(, tcache, tcache_t *, NULL)
+malloc_tsd_data(, tcache_enabled, tcache_enabled_t, tcache_enabled_default)
+
+bool	opt_tcache = true;
+ssize_t	opt_lg_tcache_max = LG_TCACHE_MAXCLASS_DEFAULT;
+
+tcache_bin_info_t	*tcache_bin_info;
+static unsigned		stack_nelms; /* Total stack elms per tcache. */
+
+size_t			nhbins;
+size_t			tcache_maxclass;
+
+/******************************************************************************/
+
+size_t	tcache_salloc(const void *ptr)
+{
+
+	return (arena_salloc(ptr, false));
+}
+
+void
+tcache_event_hard(tcache_t *tcache)
+{
+	size_t binind = tcache->next_gc_bin;
+	tcache_bin_t *tbin = &tcache->tbins[binind];
+	tcache_bin_info_t *tbin_info = &tcache_bin_info[binind];
+
+	if (tbin->low_water > 0) {
+		/*
+		 * Flush (ceiling) 3/4 of the objects below the low water mark.
+		 */
+		if (binind < NBINS) {
+			tcache_bin_flush_small(tbin, binind, tbin->ncached -
+			    tbin->low_water + (tbin->low_water >> 2), tcache);
+		} else {
+			tcache_bin_flush_large(tbin, binind, tbin->ncached -
+			    tbin->low_water + (tbin->low_water >> 2), tcache);
+		}
+		/*
+		 * Reduce fill count by 2X.  Limit lg_fill_div such that the
+		 * fill count is always at least 1.
+		 */
+		if ((tbin_info->ncached_max >> (tbin->lg_fill_div+1)) >= 1)
+			tbin->lg_fill_div++;
+	} else if (tbin->low_water < 0) {
+		/*
+		 * Increase fill count by 2X.  Make sure lg_fill_div stays
+		 * greater than 0.
+		 */
+		if (tbin->lg_fill_div > 1)
+			tbin->lg_fill_div--;
+	}
+	tbin->low_water = tbin->ncached;
+
+	tcache->next_gc_bin++;
+	if (tcache->next_gc_bin == nhbins)
+		tcache->next_gc_bin = 0;
+	tcache->ev_cnt = 0;
+}
+
+void *
+tcache_alloc_small_hard(tcache_t *tcache, tcache_bin_t *tbin, size_t binind)
+{
+	void *ret;
+
+	arena_tcache_fill_small(tcache->arena, tbin, binind,
+	    config_prof ? tcache->prof_accumbytes : 0);
+	if (config_prof)
+		tcache->prof_accumbytes = 0;
+	ret = tcache_alloc_easy(tbin);
+
+	return (ret);
+}
+
+void
+tcache_bin_flush_small(tcache_bin_t *tbin, size_t binind, unsigned rem,
+    tcache_t *tcache)
+{
+	void *ptr;
+	unsigned i, nflush, ndeferred;
+	bool merged_stats = false;
+
+	assert(binind < NBINS);
+	assert(rem <= tbin->ncached);
+
+	for (nflush = tbin->ncached - rem; nflush > 0; nflush = ndeferred) {
+		/* Lock the arena bin associated with the first object. */
+		arena_chunk_t *chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(
+		    tbin->avail[0]);
+		arena_t *arena = chunk->arena;
+		arena_bin_t *bin = &arena->bins[binind];
+
+		if (config_prof && arena == tcache->arena) {
+			malloc_mutex_lock(&arena->lock);
+			arena_prof_accum(arena, tcache->prof_accumbytes);
+			malloc_mutex_unlock(&arena->lock);
+			tcache->prof_accumbytes = 0;
+		}
+
+		malloc_mutex_lock(&bin->lock);
+		if (config_stats && arena == tcache->arena) {
+			assert(merged_stats == false);
+			merged_stats = true;
+			bin->stats.nflushes++;
+			bin->stats.nrequests += tbin->tstats.nrequests;
+			tbin->tstats.nrequests = 0;
+		}
+		ndeferred = 0;
+		for (i = 0; i < nflush; i++) {
+			ptr = tbin->avail[i];
+			assert(ptr != NULL);
+			chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+			if (chunk->arena == arena) {
+				size_t pageind = ((uintptr_t)ptr -
+				    (uintptr_t)chunk) >> LG_PAGE;
+				arena_chunk_map_t *mapelm =
+				    arena_mapp_get(chunk, pageind);
+				if (config_fill && opt_junk) {
+					arena_alloc_junk_small(ptr,
+					    &arena_bin_info[binind], true);
+				}
+				arena_dalloc_bin_locked(arena, chunk, ptr,
+				    mapelm);
+			} else {
+				/*
+				 * This object was allocated via a different
+				 * arena bin than the one that is currently
+				 * locked.  Stash the object, so that it can be
+				 * handled in a future pass.
+				 */
+				tbin->avail[ndeferred] = ptr;
+				ndeferred++;
+			}
+		}
+		malloc_mutex_unlock(&bin->lock);
+	}
+	if (config_stats && merged_stats == false) {
+		/*
+		 * The flush loop didn't happen to flush to this thread's
+		 * arena, so the stats didn't get merged.  Manually do so now.
+		 */
+		arena_bin_t *bin = &tcache->arena->bins[binind];
+		malloc_mutex_lock(&bin->lock);
+		bin->stats.nflushes++;
+		bin->stats.nrequests += tbin->tstats.nrequests;
+		tbin->tstats.nrequests = 0;
+		malloc_mutex_unlock(&bin->lock);
+	}
+
+	memmove(tbin->avail, &tbin->avail[tbin->ncached - rem],
+	    rem * sizeof(void *));
+	tbin->ncached = rem;
+	if ((int)tbin->ncached < tbin->low_water)
+		tbin->low_water = tbin->ncached;
+}
+
+void
+tcache_bin_flush_large(tcache_bin_t *tbin, size_t binind, unsigned rem,
+    tcache_t *tcache)
+{
+	void *ptr;
+	unsigned i, nflush, ndeferred;
+	bool merged_stats = false;
+
+	assert(binind < nhbins);
+	assert(rem <= tbin->ncached);
+
+	for (nflush = tbin->ncached - rem; nflush > 0; nflush = ndeferred) {
+		/* Lock the arena associated with the first object. */
+		arena_chunk_t *chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(
+		    tbin->avail[0]);
+		arena_t *arena = chunk->arena;
+
+		malloc_mutex_lock(&arena->lock);
+		if ((config_prof || config_stats) && arena == tcache->arena) {
+			if (config_prof) {
+				arena_prof_accum(arena,
+				    tcache->prof_accumbytes);
+				tcache->prof_accumbytes = 0;
+			}
+			if (config_stats) {
+				merged_stats = true;
+				arena->stats.nrequests_large +=
+				    tbin->tstats.nrequests;
+				arena->stats.lstats[binind - NBINS].nrequests +=
+				    tbin->tstats.nrequests;
+				tbin->tstats.nrequests = 0;
+			}
+		}
+		ndeferred = 0;
+		for (i = 0; i < nflush; i++) {
+			ptr = tbin->avail[i];
+			assert(ptr != NULL);
+			chunk = (arena_chunk_t *)CHUNK_ADDR2BASE(ptr);
+			if (chunk->arena == arena)
+				arena_dalloc_large_locked(arena, chunk, ptr);
+			else {
+				/*
+				 * This object was allocated via a different
+				 * arena than the one that is currently locked.
+				 * Stash the object, so that it can be handled
+				 * in a future pass.
+				 */
+				tbin->avail[ndeferred] = ptr;
+				ndeferred++;
+			}
+		}
+		malloc_mutex_unlock(&arena->lock);
+	}
+	if (config_stats && merged_stats == false) {
+		/*
+		 * The flush loop didn't happen to flush to this thread's
+		 * arena, so the stats didn't get merged.  Manually do so now.
+		 */
+		arena_t *arena = tcache->arena;
+		malloc_mutex_lock(&arena->lock);
+		arena->stats.nrequests_large += tbin->tstats.nrequests;
+		arena->stats.lstats[binind - NBINS].nrequests +=
+		    tbin->tstats.nrequests;
+		tbin->tstats.nrequests = 0;
+		malloc_mutex_unlock(&arena->lock);
+	}
+
+	memmove(tbin->avail, &tbin->avail[tbin->ncached - rem],
+	    rem * sizeof(void *));
+	tbin->ncached = rem;
+	if ((int)tbin->ncached < tbin->low_water)
+		tbin->low_water = tbin->ncached;
+}
+
+void
+tcache_arena_associate(tcache_t *tcache, arena_t *arena)
+{
+
+	if (config_stats) {
+		/* Link into list of extant tcaches. */
+		malloc_mutex_lock(&arena->lock);
+		ql_elm_new(tcache, link);
+		ql_tail_insert(&arena->tcache_ql, tcache, link);
+		malloc_mutex_unlock(&arena->lock);
+	}
+	tcache->arena = arena;
+}
+
+void
+tcache_arena_dissociate(tcache_t *tcache)
+{
+
+	if (config_stats) {
+		/* Unlink from list of extant tcaches. */
+		malloc_mutex_lock(&tcache->arena->lock);
+		ql_remove(&tcache->arena->tcache_ql, tcache, link);
+		malloc_mutex_unlock(&tcache->arena->lock);
+		tcache_stats_merge(tcache, tcache->arena);
+	}
+}
+
+tcache_t *
+tcache_create(arena_t *arena)
+{
+	tcache_t *tcache;
+	size_t size, stack_offset;
+	unsigned i;
+
+	size = offsetof(tcache_t, tbins) + (sizeof(tcache_bin_t) * nhbins);
+	/* Naturally align the pointer stacks. */
+	size = PTR_CEILING(size);
+	stack_offset = size;
+	size += stack_nelms * sizeof(void *);
+	/*
+	 * Round up to the nearest multiple of the cacheline size, in order to
+	 * avoid the possibility of false cacheline sharing.
+	 *
+	 * That this works relies on the same logic as in ipalloc(), but we
+	 * cannot directly call ipalloc() here due to tcache bootstrapping
+	 * issues.
+	 */
+	size = (size + CACHELINE_MASK) & (-CACHELINE);
+
+	if (size <= SMALL_MAXCLASS)
+		tcache = (tcache_t *)arena_malloc_small(arena, size, true);
+	else if (size <= tcache_maxclass)
+		tcache = (tcache_t *)arena_malloc_large(arena, size, true);
+	else
+		tcache = (tcache_t *)icallocx(size, false, arena);
+
+	if (tcache == NULL)
+		return (NULL);
+
+	tcache_arena_associate(tcache, arena);
+
+	assert((TCACHE_NSLOTS_SMALL_MAX & 1U) == 0);
+	for (i = 0; i < nhbins; i++) {
+		tcache->tbins[i].lg_fill_div = 1;
+		tcache->tbins[i].avail = (void **)((uintptr_t)tcache +
+		    (uintptr_t)stack_offset);
+		stack_offset += tcache_bin_info[i].ncached_max * sizeof(void *);
+	}
+
+	tcache_tsd_set(&tcache);
+
+	return (tcache);
+}
+
+void
+tcache_destroy(tcache_t *tcache)
+{
+	unsigned i;
+	size_t tcache_size;
+
+	tcache_arena_dissociate(tcache);
+
+	for (i = 0; i < NBINS; i++) {
+		tcache_bin_t *tbin = &tcache->tbins[i];
+		tcache_bin_flush_small(tbin, i, 0, tcache);
+
+		if (config_stats && tbin->tstats.nrequests != 0) {
+			arena_t *arena = tcache->arena;
+			arena_bin_t *bin = &arena->bins[i];
+			malloc_mutex_lock(&bin->lock);
+			bin->stats.nrequests += tbin->tstats.nrequests;
+			malloc_mutex_unlock(&bin->lock);
+		}
+	}
+
+	for (; i < nhbins; i++) {
+		tcache_bin_t *tbin = &tcache->tbins[i];
+		tcache_bin_flush_large(tbin, i, 0, tcache);
+
+		if (config_stats && tbin->tstats.nrequests != 0) {
+			arena_t *arena = tcache->arena;
+			malloc_mutex_lock(&arena->lock);
+			arena->stats.nrequests_large += tbin->tstats.nrequests;
+			arena->stats.lstats[i - NBINS].nrequests +=
+			    tbin->tstats.nrequests;
+			malloc_mutex_unlock(&arena->lock);
+		}
+	}
+
+	if (config_prof && tcache->prof_accumbytes > 0) {
+		malloc_mutex_lock(&tcache->arena->lock);
+		arena_prof_accum(tcache->arena, tcache->prof_accumbytes);
+		malloc_mutex_unlock(&tcache->arena->lock);
+	}
+
+	tcache_size = arena_salloc(tcache, false);
+	if (tcache_size <= SMALL_MAXCLASS) {
+		arena_chunk_t *chunk = CHUNK_ADDR2BASE(tcache);
+		arena_t *arena = chunk->arena;
+		size_t pageind = ((uintptr_t)tcache - (uintptr_t)chunk) >>
+		    LG_PAGE;
+		arena_chunk_map_t *mapelm = arena_mapp_get(chunk, pageind);
+
+		arena_dalloc_bin(arena, chunk, tcache, pageind, mapelm);
+	} else if (tcache_size <= tcache_maxclass) {
+		arena_chunk_t *chunk = CHUNK_ADDR2BASE(tcache);
+		arena_t *arena = chunk->arena;
+
+		arena_dalloc_large(arena, chunk, tcache);
+	} else
+		idallocx(tcache, false);
+}
+
+void
+tcache_thread_cleanup(void *arg)
+{
+	tcache_t *tcache = *(tcache_t **)arg;
+
+	if (tcache == TCACHE_STATE_DISABLED) {
+		/* Do nothing. */
+	} else if (tcache == TCACHE_STATE_REINCARNATED) {
+		/*
+		 * Another destructor called an allocator function after this
+		 * destructor was called.  Reset tcache to
+		 * TCACHE_STATE_PURGATORY in order to receive another callback.
+		 */
+		tcache = TCACHE_STATE_PURGATORY;
+		tcache_tsd_set(&tcache);
+	} else if (tcache == TCACHE_STATE_PURGATORY) {
+		/*
+		 * The previous time this destructor was called, we set the key
+		 * to TCACHE_STATE_PURGATORY so that other destructors wouldn't
+		 * cause re-creation of the tcache.  This time, do nothing, so
+		 * that the destructor will not be called again.
+		 */
+	} else if (tcache != NULL) {
+		assert(tcache != TCACHE_STATE_PURGATORY);
+		tcache_destroy(tcache);
+		tcache = TCACHE_STATE_PURGATORY;
+		tcache_tsd_set(&tcache);
+	}
+}
+
+void
+tcache_stats_merge(tcache_t *tcache, arena_t *arena)
+{
+	unsigned i;
+
+	/* Merge and reset tcache stats. */
+	for (i = 0; i < NBINS; i++) {
+		arena_bin_t *bin = &arena->bins[i];
+		tcache_bin_t *tbin = &tcache->tbins[i];
+		malloc_mutex_lock(&bin->lock);
+		bin->stats.nrequests += tbin->tstats.nrequests;
+		malloc_mutex_unlock(&bin->lock);
+		tbin->tstats.nrequests = 0;
+	}
+
+	for (; i < nhbins; i++) {
+		malloc_large_stats_t *lstats = &arena->stats.lstats[i - NBINS];
+		tcache_bin_t *tbin = &tcache->tbins[i];
+		arena->stats.nrequests_large += tbin->tstats.nrequests;
+		lstats->nrequests += tbin->tstats.nrequests;
+		tbin->tstats.nrequests = 0;
+	}
+}
+
+bool
+tcache_boot0(void)
+{
+	unsigned i;
+
+	/*
+	 * If necessary, clamp opt_lg_tcache_max, now that arena_maxclass is
+	 * known.
+	 */
+	if (opt_lg_tcache_max < 0 || (1U << opt_lg_tcache_max) < SMALL_MAXCLASS)
+		tcache_maxclass = SMALL_MAXCLASS;
+	else if ((1U << opt_lg_tcache_max) > arena_maxclass)
+		tcache_maxclass = arena_maxclass;
+	else
+		tcache_maxclass = (1U << opt_lg_tcache_max);
+
+	nhbins = NBINS + (tcache_maxclass >> LG_PAGE);
+
+	/* Initialize tcache_bin_info. */
+	tcache_bin_info = (tcache_bin_info_t *)base_alloc(nhbins *
+	    sizeof(tcache_bin_info_t));
+	if (tcache_bin_info == NULL)
+		return (true);
+	stack_nelms = 0;
+	for (i = 0; i < NBINS; i++) {
+		if ((arena_bin_info[i].nregs << 1) <= TCACHE_NSLOTS_SMALL_MAX) {
+			tcache_bin_info[i].ncached_max =
+			    (arena_bin_info[i].nregs << 1);
+		} else {
+			tcache_bin_info[i].ncached_max =
+			    TCACHE_NSLOTS_SMALL_MAX;
+		}
+		stack_nelms += tcache_bin_info[i].ncached_max;
+	}
+	for (; i < nhbins; i++) {
+		tcache_bin_info[i].ncached_max = TCACHE_NSLOTS_LARGE;
+		stack_nelms += tcache_bin_info[i].ncached_max;
+	}
+
+	return (false);
+}
+
+bool
+tcache_boot1(void)
+{
+
+	if (tcache_tsd_boot() || tcache_enabled_tsd_boot())
+		return (true);
+
+	return (false);
+}
--- a/src/tsd.c
+++ b/src/tsd.c
@@ -0,0 +1,107 @@
+#define	JEMALLOC_TSD_C_
+#include "jemalloc/internal/jemalloc_internal.h"
+
+/******************************************************************************/
+/* Data. */
+
+static unsigned ncleanups;
+static malloc_tsd_cleanup_t cleanups[MALLOC_TSD_CLEANUPS_MAX];
+
+/******************************************************************************/
+
+void *
+malloc_tsd_malloc(size_t size)
+{
+
+	/* Avoid choose_arena() in order to dodge bootstrapping issues. */
+	return (arena_malloc(arenas[0], size, false, false));
+}
+
+void
+malloc_tsd_dalloc(void *wrapper)
+{
+
+	idalloc(wrapper);
+}
+
+void
+malloc_tsd_no_cleanup(void *arg)
+{
+
+	not_reached();
+}
+
+#if defined(JEMALLOC_MALLOC_THREAD_CLEANUP) || defined(_WIN32)
+#ifndef _WIN32
+JEMALLOC_EXPORT
+#endif
+void
+_malloc_thread_cleanup(void)
+{
+	bool pending[MALLOC_TSD_CLEANUPS_MAX], again;
+	unsigned i;
+
+	for (i = 0; i < ncleanups; i++)
+		pending[i] = true;
+
+	do {
+		again = false;
+		for (i = 0; i < ncleanups; i++) {
+			if (pending[i]) {
+				pending[i] = cleanups[i]();
+				if (pending[i])
+					again = true;
+			}
+		}
+	} while (again);
+}
+#endif
+
+void
+malloc_tsd_cleanup_register(bool (*f)(void))
+{
+
+	assert(ncleanups < MALLOC_TSD_CLEANUPS_MAX);
+	cleanups[ncleanups] = f;
+	ncleanups++;
+}
+
+void
+malloc_tsd_boot(void)
+{
+
+	ncleanups = 0;
+}
+
+#ifdef _WIN32
+static BOOL WINAPI
+_tls_callback(HINSTANCE hinstDLL, DWORD fdwReason, LPVOID lpvReserved)
+{
+
+	switch (fdwReason) {
+#ifdef JEMALLOC_LAZY_LOCK
+	case DLL_THREAD_ATTACH:
+		isthreaded = true;
+		break;
+#endif
+	case DLL_THREAD_DETACH:
+		_malloc_thread_cleanup();
+		break;
+	default:
+		break;
+	}
+	return (true);
+}
+
+#ifdef _MSC_VER
+#  ifdef _M_IX86
+#    pragma comment(linker, "/INCLUDE:__tls_used")
+#  else
+#    pragma comment(linker, "/INCLUDE:_tls_used")
+#  endif
+#  pragma section(".CRT$XLY",long,read)
+#endif
+JEMALLOC_SECTION(".CRT$XLY") JEMALLOC_ATTR(used)
+static const BOOL	(WINAPI *tls_callback)(HINSTANCE hinstDLL,
+    DWORD fdwReason, LPVOID lpvReserved) = _tls_callback;
+#endif
--- a/Show More
+++ b/Show More